diff options
| author | Maxime Ripard <mripard@kernel.org> | 2025-06-03 15:07:39 +0200 |
|---|---|---|
| committer | Maxime Ripard <mripard@kernel.org> | 2025-06-03 15:07:39 +0200 |
| commit | 7b1166dee847d5018c1f3cc781218e806078f752 (patch) | |
| tree | 84a2e30eade0212b89cc05063fb8e9a5793e9d7a /drivers/gpu/drm | |
| parent | b6cbfa872045d7ed307c55aa1cdb0d5125d84282 (diff) | |
| parent | f8bb3ed3197966fb60bedcbdc126d2bd5bc0a77f (diff) | |
| download | linux-next-7b1166dee847d5018c1f3cc781218e806078f752.tar.gz linux-next-7b1166dee847d5018c1f3cc781218e806078f752.zip | |
Merge drm-next-2025-05-28 into drm-misc-next
Christian needs a recent drm-next branch to merge fence patches.
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Diffstat (limited to 'drivers/gpu/drm')
1009 files changed, 35531 insertions, 22601 deletions
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 9488fc01bca3..f094797f3b2b 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -274,6 +274,8 @@ source "drivers/gpu/drm/amd/amdgpu/Kconfig" source "drivers/gpu/drm/nouveau/Kconfig" +source "drivers/gpu/drm/nova/Kconfig" + source "drivers/gpu/drm/i915/Kconfig" source "drivers/gpu/drm/xe/Kconfig" diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index 70510620f29c..4199715670b1 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -177,6 +177,7 @@ obj-$(CONFIG_DRM_VMWGFX)+= vmwgfx/ obj-$(CONFIG_DRM_VGEM) += vgem/ obj-$(CONFIG_DRM_VKMS) += vkms/ obj-$(CONFIG_DRM_NOUVEAU) +=nouveau/ +obj-$(CONFIG_DRM_NOVA) += nova/ obj-$(CONFIG_DRM_EXYNOS) +=exynos/ obj-$(CONFIG_DRM_ROCKCHIP) +=rockchip/ obj-$(CONFIG_DRM_GMA500) += gma500/ diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index aacc810cabb3..87080c06e5fc 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -66,7 +66,7 @@ amdgpu-y += amdgpu_device.o amdgpu_doorbell_mgr.o amdgpu_kms.o \ amdgpu_fw_attestation.o amdgpu_securedisplay.o \ amdgpu_eeprom.o amdgpu_mca.o amdgpu_psp_ta.o amdgpu_lsdma.o \ amdgpu_ring_mux.o amdgpu_xcp.o amdgpu_seq64.o amdgpu_aca.o amdgpu_dev_coredump.o \ - amdgpu_cper.o + amdgpu_cper.o amdgpu_userq_fence.o amdgpu_eviction_fence.o amdgpu-$(CONFIG_PROC_FS) += amdgpu_fdinfo.o @@ -174,7 +174,10 @@ amdgpu-y += \ amdgpu-y += \ amdgpu_mes.o \ mes_v11_0.o \ - mes_v12_0.o + mes_v12_0.o \ + +# add GFX userqueue support +amdgpu-y += mes_userqueue.o # add UVD block amdgpu-y += \ @@ -253,6 +256,8 @@ amdgpu-y += \ # add amdkfd interfaces amdgpu-y += amdgpu_amdkfd.o +# add gfx usermode queue +amdgpu-y += amdgpu_userq.o ifneq ($(CONFIG_HSA_AMD),) AMDKFD_PATH := ../amdkfd diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index ef6e78224fdf..836ea081088a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -113,6 +113,8 @@ #include "amdgpu_xcp.h" #include "amdgpu_seq64.h" #include "amdgpu_reg_state.h" +#include "amdgpu_userq.h" +#include "amdgpu_eviction_fence.h" #if defined(CONFIG_DRM_AMD_ISP) #include "amdgpu_isp.h" #endif @@ -228,7 +230,7 @@ extern int amdgpu_force_asic_type; extern int amdgpu_smartshift_bias; extern int amdgpu_use_xgmi_p2p; extern int amdgpu_mtype_local; -extern bool enforce_isolation; +extern int amdgpu_enforce_isolation; #ifdef CONFIG_HSA_AMD extern int sched_policy; extern bool debug_evictions; @@ -266,8 +268,10 @@ extern int amdgpu_umsch_mm_fwlog; extern int amdgpu_user_partt_mode; extern int amdgpu_agp; +extern int amdgpu_rebar; extern int amdgpu_wbrf; +extern int amdgpu_user_queue; #define AMDGPU_VM_MAX_NUM_CTX 4096 #define AMDGPU_SG_THRESHOLD (256*1024*1024) @@ -488,7 +492,6 @@ struct amdgpu_flip_work { bool async; }; - /* * file private structure */ @@ -501,6 +504,11 @@ struct amdgpu_fpriv { struct mutex bo_list_lock; struct idr bo_list_handles; struct amdgpu_ctx_mgr ctx_mgr; + struct amdgpu_userq_mgr userq_mgr; + + /* Eviction fence infra */ + struct amdgpu_eviction_fence_mgr evf_mgr; + /** GPU partition selection */ uint32_t xcp_id; }; @@ -512,12 +520,62 @@ int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv); */ #define AMDGPU_MAX_WB 1024 /* Reserve at most 1024 WB slots for amdgpu-owned rings. */ +/** + * amdgpu_wb - This struct is used for small GPU memory allocation. + * + * This struct is used to allocate a small amount of GPU memory that can be + * used to shadow certain states into the memory. This is especially useful for + * providing easy CPU access to some states without requiring register access + * (e.g., if some block is power gated, reading register may be problematic). + * + * Note: the term writeback was initially used because many of the amdgpu + * components had some level of writeback memory, and this struct initially + * described those components. + */ struct amdgpu_wb { + + /** + * @wb_obj: + * + * Buffer Object used for the writeback memory. + */ struct amdgpu_bo *wb_obj; + + /** + * @wb: + * + * Pointer to the first writeback slot. In terms of CPU address + * this value can be accessed directly by using the offset as an index. + * For the GPU address, it is necessary to use gpu_addr and the offset. + */ volatile uint32_t *wb; + + /** + * @gpu_addr: + * + * Writeback base address in the GPU. + */ uint64_t gpu_addr; - u32 num_wb; /* Number of wb slots actually reserved for amdgpu. */ + + /** + * @num_wb: + * + * Number of writeback slots reserved for amdgpu. + */ + u32 num_wb; + + /** + * @used: + * + * Track the writeback slot already used. + */ unsigned long used[DIV_ROUND_UP(AMDGPU_MAX_WB, BITS_PER_LONG)]; + + /** + * @lock: + * + * Protects read and write of the used field array. + */ spinlock_t lock; }; @@ -551,6 +609,7 @@ struct amdgpu_allowed_register_entry { * are reset depends on the ASIC. Notably doesn't reset IPs * shared with the CPU on APUs or the memory controllers (so * VRAM is not lost). Not available on all ASICs. + * @AMD_RESET_LINK: Triggers SW-UP link reset on other GPUs * @AMD_RESET_BACO: BACO (Bus Alive, Chip Off) method powers off and on the card * but without powering off the PCI bus. Suitable only for * discrete GPUs. @@ -568,6 +627,7 @@ enum amd_reset_method { AMD_RESET_METHOD_MODE0, AMD_RESET_METHOD_MODE1, AMD_RESET_METHOD_MODE2, + AMD_RESET_METHOD_LINK, AMD_RESET_METHOD_BACO, AMD_RESET_METHOD_PCI, AMD_RESET_METHOD_ON_INIT, @@ -821,6 +881,11 @@ struct amdgpu_mqd_prop { uint32_t hqd_queue_priority; bool allow_tunneling; bool hqd_active; + uint64_t shadow_addr; + uint64_t gds_bkup_addr; + uint64_t csa_addr; + uint64_t fence_address; + bool tmz_queue; }; struct amdgpu_mqd { @@ -829,6 +894,12 @@ struct amdgpu_mqd { struct amdgpu_mqd_prop *p); }; +struct amdgpu_pcie_reset_ctx { + bool in_link_reset; + bool occurs_dpc; + bool audio_suspended; +}; + /* * Custom Init levels could be defined for different situations where a full * initialization of all hardware blocks are not expected. Sample cases are @@ -853,6 +924,14 @@ struct amdgpu_init_level { struct amdgpu_reset_domain; struct amdgpu_fru_info; +enum amdgpu_enforce_isolation_mode { + AMDGPU_ENFORCE_ISOLATION_DISABLE = 0, + AMDGPU_ENFORCE_ISOLATION_ENABLE = 1, + AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY = 2, + AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER = 3, +}; + + /* * Non-zero (true) if the GPU has VRAM. Zero (false) otherwise. */ @@ -1081,6 +1160,13 @@ struct amdgpu_device { bool enable_uni_mes; struct amdgpu_mes mes; struct amdgpu_mqd mqds[AMDGPU_HW_IP_NUM]; + const struct amdgpu_userq_funcs *userq_funcs[AMDGPU_HW_IP_NUM]; + + /* xarray used to retrieve the user queue fence driver reference + * in the EOP interrupt handler to signal the particular user + * queue fence. + */ + struct xarray userq_xa; /* df */ struct amdgpu_df df; @@ -1160,6 +1246,8 @@ struct amdgpu_device { struct pci_saved_state *pci_state; pci_channel_state_t pci_channel_state; + struct amdgpu_pcie_reset_ctx pcie_reset_ctx; + /* Track auto wait count on s_barrier settings */ bool barrier_has_auto_waitcnt; @@ -1193,10 +1281,11 @@ struct amdgpu_device { bool debug_enable_ras_aca; bool debug_exp_resets; bool debug_disable_gpu_ring_reset; + bool debug_vm_userptr; /* Protection for the following isolation structure */ struct mutex enforce_isolation_mutex; - bool enforce_isolation[MAX_XCP]; + enum amdgpu_enforce_isolation_mode enforce_isolation[MAX_XCP]; struct amdgpu_isolation { void *owner; struct dma_fence *spearhead; @@ -1210,6 +1299,10 @@ struct amdgpu_device { * in KFD: VRAM or GTT. */ bool apu_prefer_gtt; + + struct list_head userq_mgr_list; + struct mutex userq_mutex; + bool userq_halt_for_enforce_isolation; }; static inline uint32_t amdgpu_ip_version(const struct amdgpu_device *adev, @@ -1464,6 +1557,7 @@ void amdgpu_device_program_register_sequence(struct amdgpu_device *adev, const u32 array_size); int amdgpu_device_mode1_reset(struct amdgpu_device *adev); +int amdgpu_device_link_reset(struct amdgpu_device *adev); bool amdgpu_device_supports_atpx(struct drm_device *dev); bool amdgpu_device_supports_px(struct drm_device *dev); bool amdgpu_device_supports_boco(struct drm_device *dev); @@ -1614,11 +1708,9 @@ static inline void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_cap #if defined(CONFIG_ACPI) && defined(CONFIG_SUSPEND) bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev); bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev); -void amdgpu_choose_low_power_state(struct amdgpu_device *adev); #else static inline bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) { return false; } static inline bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev) { return false; } -static inline void amdgpu_choose_low_power_state(struct amdgpu_device *adev) { } #endif void amdgpu_register_gpu_instance(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c index b4ad163f42a7..3835f2592914 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c @@ -120,6 +120,9 @@ static void aca_smu_bank_dump(struct amdgpu_device *adev, int idx, int total, st for (i = 0; i < ARRAY_SIZE(aca_regs); i++) RAS_EVENT_LOG(adev, event_id, HW_ERR "ACA[%02d/%02d].%s=0x%016llx\n", idx + 1, total, aca_regs[i].name, bank->regs[aca_regs[i].reg_idx]); + + if (ACA_REG__STATUS__SCRUB(bank->regs[ACA_REG_IDX_STATUS])) + RAS_EVENT_LOG(adev, event_id, HW_ERR "hardware error logged by the scrubber\n"); } static int aca_smu_get_valid_aca_banks(struct amdgpu_device *adev, enum aca_smu_type type, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c index b7f8f2ff143d..707e131f89d2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c @@ -1533,22 +1533,4 @@ bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) #endif /* CONFIG_AMD_PMC */ } -/** - * amdgpu_choose_low_power_state - * - * @adev: amdgpu_device_pointer - * - * Choose the target low power state for the GPU - */ -void amdgpu_choose_low_power_state(struct amdgpu_device *adev) -{ - if (adev->in_runpm) - return; - - if (amdgpu_acpi_is_s0ix_active(adev)) - adev->in_s0ix = true; - else if (amdgpu_acpi_is_s3_active(adev)) - adev->in_s3 = true; -} - #endif /* CONFIG_SUSPEND */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index d2ec4130a316..260165bbe373 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -2559,6 +2559,18 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info, if (ret != -EFAULT) return ret; + /* If applications unmap memory before destroying the userptr + * from the KFD, trigger a segmentation fault in VM debug mode. + */ + if (amdgpu_ttm_adev(bo->tbo.bdev)->debug_vm_userptr) { + pr_err("Pid %d unmapped memory before destroying userptr at GPU addr 0x%llx\n", + pid_nr(process_info->pid), mem->va); + + // Send GPU VM fault to user space + kfd_signal_vm_fault_event_with_userptr(kfd_lookup_process_by_pid(process_info->pid), + mem->va); + } + ret = 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c index eb015bdda8a7..c7d32fb216e4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c @@ -281,6 +281,9 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev, case ATOM_DGPU_VRAM_TYPE_GDDR6: vram_type = AMDGPU_VRAM_TYPE_GDDR6; break; + case ATOM_DGPU_VRAM_TYPE_HBM3E: + vram_type = AMDGPU_VRAM_TYPE_HBM3E; + break; default: vram_type = AMDGPU_VRAM_TYPE_UNKNOWN; break; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c index 68bce6a6d09d..004a6a9d6b9f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c @@ -252,83 +252,22 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, if (!adev->pm.fw) { switch (adev->asic_type) { - case CHIP_TAHITI: - strcpy(fw_name, "radeon/tahiti_smc.bin"); - break; - case CHIP_PITCAIRN: - if ((adev->pdev->revision == 0x81) && - ((adev->pdev->device == 0x6810) || - (adev->pdev->device == 0x6811))) { - info->is_kicker = true; - strcpy(fw_name, "radeon/pitcairn_k_smc.bin"); - } else { - strcpy(fw_name, "radeon/pitcairn_smc.bin"); - } - break; - case CHIP_VERDE: - if (((adev->pdev->device == 0x6820) && - ((adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83))) || - ((adev->pdev->device == 0x6821) && - ((adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0x87))) || - ((adev->pdev->revision == 0x87) && - ((adev->pdev->device == 0x6823) || - (adev->pdev->device == 0x682b)))) { - info->is_kicker = true; - strcpy(fw_name, "radeon/verde_k_smc.bin"); - } else { - strcpy(fw_name, "radeon/verde_smc.bin"); - } - break; - case CHIP_OLAND: - if (((adev->pdev->revision == 0x81) && - ((adev->pdev->device == 0x6600) || - (adev->pdev->device == 0x6604) || - (adev->pdev->device == 0x6605) || - (adev->pdev->device == 0x6610))) || - ((adev->pdev->revision == 0x83) && - (adev->pdev->device == 0x6610))) { - info->is_kicker = true; - strcpy(fw_name, "radeon/oland_k_smc.bin"); - } else { - strcpy(fw_name, "radeon/oland_smc.bin"); - } - break; - case CHIP_HAINAN: - if (((adev->pdev->revision == 0x81) && - (adev->pdev->device == 0x6660)) || - ((adev->pdev->revision == 0x83) && - ((adev->pdev->device == 0x6660) || - (adev->pdev->device == 0x6663) || - (adev->pdev->device == 0x6665) || - (adev->pdev->device == 0x6667)))) { - info->is_kicker = true; - strcpy(fw_name, "radeon/hainan_k_smc.bin"); - } else if ((adev->pdev->revision == 0xc3) && - (adev->pdev->device == 0x6665)) { - info->is_kicker = true; - strcpy(fw_name, "radeon/banks_k_2_smc.bin"); - } else { - strcpy(fw_name, "radeon/hainan_smc.bin"); - } - break; case CHIP_BONAIRE: if ((adev->pdev->revision == 0x80) || (adev->pdev->revision == 0x81) || (adev->pdev->device == 0x665f)) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/bonaire_k_smc.bin"); + strscpy(fw_name, "amdgpu/bonaire_k_smc.bin"); } else { - strcpy(fw_name, "amdgpu/bonaire_smc.bin"); + strscpy(fw_name, "amdgpu/bonaire_smc.bin"); } break; case CHIP_HAWAII: if (adev->pdev->revision == 0x80) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/hawaii_k_smc.bin"); + strscpy(fw_name, "amdgpu/hawaii_k_smc.bin"); } else { - strcpy(fw_name, "amdgpu/hawaii_smc.bin"); + strscpy(fw_name, "amdgpu/hawaii_smc.bin"); } break; case CHIP_TOPAZ: @@ -338,76 +277,76 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, ((adev->pdev->device == 0x6900) && (adev->pdev->revision == 0xD1)) || ((adev->pdev->device == 0x6900) && (adev->pdev->revision == 0xD3))) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/topaz_k_smc.bin"); + strscpy(fw_name, "amdgpu/topaz_k_smc.bin"); } else - strcpy(fw_name, "amdgpu/topaz_smc.bin"); + strscpy(fw_name, "amdgpu/topaz_smc.bin"); break; case CHIP_TONGA: if (((adev->pdev->device == 0x6939) && (adev->pdev->revision == 0xf1)) || ((adev->pdev->device == 0x6938) && (adev->pdev->revision == 0xf1))) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/tonga_k_smc.bin"); + strscpy(fw_name, "amdgpu/tonga_k_smc.bin"); } else - strcpy(fw_name, "amdgpu/tonga_smc.bin"); + strscpy(fw_name, "amdgpu/tonga_smc.bin"); break; case CHIP_FIJI: - strcpy(fw_name, "amdgpu/fiji_smc.bin"); + strscpy(fw_name, "amdgpu/fiji_smc.bin"); break; case CHIP_POLARIS11: if (type == CGS_UCODE_ID_SMU) { if (ASICID_IS_P21(adev->pdev->device, adev->pdev->revision)) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/polaris11_k_smc.bin"); + strscpy(fw_name, "amdgpu/polaris11_k_smc.bin"); } else if (ASICID_IS_P31(adev->pdev->device, adev->pdev->revision)) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/polaris11_k2_smc.bin"); + strscpy(fw_name, "amdgpu/polaris11_k2_smc.bin"); } else { - strcpy(fw_name, "amdgpu/polaris11_smc.bin"); + strscpy(fw_name, "amdgpu/polaris11_smc.bin"); } } else if (type == CGS_UCODE_ID_SMU_SK) { - strcpy(fw_name, "amdgpu/polaris11_smc_sk.bin"); + strscpy(fw_name, "amdgpu/polaris11_smc_sk.bin"); } break; case CHIP_POLARIS10: if (type == CGS_UCODE_ID_SMU) { if (ASICID_IS_P20(adev->pdev->device, adev->pdev->revision)) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/polaris10_k_smc.bin"); + strscpy(fw_name, "amdgpu/polaris10_k_smc.bin"); } else if (ASICID_IS_P30(adev->pdev->device, adev->pdev->revision)) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/polaris10_k2_smc.bin"); + strscpy(fw_name, "amdgpu/polaris10_k2_smc.bin"); } else { - strcpy(fw_name, "amdgpu/polaris10_smc.bin"); + strscpy(fw_name, "amdgpu/polaris10_smc.bin"); } } else if (type == CGS_UCODE_ID_SMU_SK) { - strcpy(fw_name, "amdgpu/polaris10_smc_sk.bin"); + strscpy(fw_name, "amdgpu/polaris10_smc_sk.bin"); } break; case CHIP_POLARIS12: if (ASICID_IS_P23(adev->pdev->device, adev->pdev->revision)) { info->is_kicker = true; - strcpy(fw_name, "amdgpu/polaris12_k_smc.bin"); + strscpy(fw_name, "amdgpu/polaris12_k_smc.bin"); } else { - strcpy(fw_name, "amdgpu/polaris12_smc.bin"); + strscpy(fw_name, "amdgpu/polaris12_smc.bin"); } break; case CHIP_VEGAM: - strcpy(fw_name, "amdgpu/vegam_smc.bin"); + strscpy(fw_name, "amdgpu/vegam_smc.bin"); break; case CHIP_VEGA10: if ((adev->pdev->device == 0x687f) && ((adev->pdev->revision == 0xc0) || (adev->pdev->revision == 0xc1) || (adev->pdev->revision == 0xc3))) - strcpy(fw_name, "amdgpu/vega10_acg_smc.bin"); + strscpy(fw_name, "amdgpu/vega10_acg_smc.bin"); else - strcpy(fw_name, "amdgpu/vega10_smc.bin"); + strscpy(fw_name, "amdgpu/vega10_smc.bin"); break; case CHIP_VEGA12: - strcpy(fw_name, "amdgpu/vega12_smc.bin"); + strscpy(fw_name, "amdgpu/vega12_smc.bin"); break; case CHIP_VEGA20: - strcpy(fw_name, "amdgpu/vega20_smc.bin"); + strscpy(fw_name, "amdgpu/vega20_smc.bin"); break; default: DRM_ERROR("SMC firmware not supported\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c index 360e07a5c7c1..5a234eadae8b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c @@ -549,7 +549,7 @@ int amdgpu_cper_init(struct amdgpu_device *adev) { int r; - if (!amdgpu_aca_is_enabled(adev)) + if (!amdgpu_aca_is_enabled(adev) && !amdgpu_sriov_ras_cper_en(adev)) return 0; r = amdgpu_cper_ring_init(adev); @@ -568,7 +568,7 @@ int amdgpu_cper_init(struct amdgpu_device *adev) int amdgpu_cper_fini(struct amdgpu_device *adev) { - if (!amdgpu_aca_is_enabled(adev)) + if (!amdgpu_aca_is_enabled(adev) && !amdgpu_sriov_ras_cper_en(adev)) return 0; adev->cper.enabled = false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 5a231b997d65..a2adaacf6adb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -297,7 +297,25 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, p->filp->client_id); if (ret) goto free_all_kdata; - p->jobs[i]->enforce_isolation = p->adev->enforce_isolation[fpriv->xcp_id]; + switch (p->adev->enforce_isolation[fpriv->xcp_id]) { + case AMDGPU_ENFORCE_ISOLATION_DISABLE: + default: + p->jobs[i]->enforce_isolation = false; + p->jobs[i]->run_cleaner_shader = false; + break; + case AMDGPU_ENFORCE_ISOLATION_ENABLE: + p->jobs[i]->enforce_isolation = true; + p->jobs[i]->run_cleaner_shader = true; + break; + case AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY: + p->jobs[i]->enforce_isolation = true; + p->jobs[i]->run_cleaner_shader = false; + break; + case AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER: + p->jobs[i]->enforce_isolation = true; + p->jobs[i]->run_cleaner_shader = false; + break; + } } p->gang_leader = p->jobs[p->gang_leader_idx]; @@ -350,6 +368,10 @@ static int amdgpu_cs_p2_ib(struct amdgpu_cs_parser *p, ring = amdgpu_job_ring(job); ib = &job->ibs[job->num_ibs++]; + /* submissions to kernel queues are disabled */ + if (ring->no_user_submission) + return -EINVAL; + /* MM engine doesn't support user fences */ if (p->uf_bo && ring->funcs->no_user_fence) return -EINVAL; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c index cfdf558b48b6..02138aa55793 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c @@ -109,7 +109,7 @@ int amdgpu_unmap_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct drm_exec exec; int r; - drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT, 0); + drm_exec_init(&exec, 0, 0); drm_exec_until_all_locked(&exec) { r = amdgpu_vm_lock_pd(vm, &exec, 0); if (likely(!r)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index a1450f13d963..8e626f50b362 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -2105,6 +2105,7 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev) amdgpu_rap_debugfs_init(adev); amdgpu_securedisplay_debugfs_init(adev); amdgpu_fw_attestation_debugfs_init(adev); + amdgpu_psp_debugfs_init(adev); debugfs_create_file("amdgpu_evict_vram", 0400, root, adev, &amdgpu_evict_vram_fops); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 7f354cd532dc..4d1b54f58495 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -85,6 +85,7 @@ #if IS_ENABLED(CONFIG_X86) #include <asm/intel-family.h> +#include <asm/cpu_device_id.h> #endif MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin"); @@ -1680,6 +1681,9 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev) if (amdgpu_sriov_vf(adev)) return 0; + if (!amdgpu_rebar) + return 0; + /* resizing on Dell G5 SE platforms causes problems with runtime pm */ if ((amdgpu_runtime_pm != 0) && adev->pdev->vendor == PCI_VENDOR_ID_ATI && @@ -1870,6 +1874,35 @@ static bool amdgpu_device_pcie_dynamic_switching_supported(struct amdgpu_device return true; } +static bool amdgpu_device_aspm_support_quirk(struct amdgpu_device *adev) +{ +#if IS_ENABLED(CONFIG_X86) + struct cpuinfo_x86 *c = &cpu_data(0); + + if (!(amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(12, 0, 0) || + amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(12, 0, 1))) + return false; + + if (c->x86 == 6 && + adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN5) { + switch (c->x86_model) { + case VFM_MODEL(INTEL_ALDERLAKE): + case VFM_MODEL(INTEL_ALDERLAKE_L): + case VFM_MODEL(INTEL_RAPTORLAKE): + case VFM_MODEL(INTEL_RAPTORLAKE_P): + case VFM_MODEL(INTEL_RAPTORLAKE_S): + return true; + default: + return false; + } + } else { + return false; + } +#else + return false; +#endif +} + /** * amdgpu_device_should_use_aspm - check if the device should program ASPM * @@ -1894,7 +1927,7 @@ bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev) } if (adev->flags & AMD_IS_APU) return false; - if (!(adev->pm.pp_feature & PP_PCIE_DPM_MASK)) + if (amdgpu_device_aspm_support_quirk(adev)) return false; return pcie_aspm_enabled(adev->pdev); } @@ -2112,8 +2145,31 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev) adev->firmware.load_type = amdgpu_ucode_get_load_type(adev, amdgpu_fw_load_type); - for (i = 0; i < MAX_XCP; i++) - adev->enforce_isolation[i] = !!enforce_isolation; + for (i = 0; i < MAX_XCP; i++) { + switch (amdgpu_enforce_isolation) { + case -1: + case 0: + default: + /* disable */ + adev->enforce_isolation[i] = AMDGPU_ENFORCE_ISOLATION_DISABLE; + break; + case 1: + /* enable */ + adev->enforce_isolation[i] = + AMDGPU_ENFORCE_ISOLATION_ENABLE; + break; + case 2: + /* enable legacy mode */ + adev->enforce_isolation[i] = + AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY; + break; + case 3: + /* enable only process isolation without submitting cleaner shader */ + adev->enforce_isolation[i] = + AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER; + break; + } + } return 0; } @@ -2689,6 +2745,13 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) break; } + /* Check for IP version 9.4.3 with A0 hardware */ + if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) && + !amdgpu_device_get_rev_id(adev)) { + dev_err(adev->dev, "Unsupported A0 hardware\n"); + return -ENODEV; /* device unsupported - no device error */ + } + if (amdgpu_has_atpx() && (amdgpu_is_atpx_hybrid() || amdgpu_has_atpx_dgpu_power_cntl()) && @@ -2701,7 +2764,6 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) adev->has_pr3 = parent ? pci_pr3_present(parent) : false; } - adev->pm.pp_feature = amdgpu_pp_feature_mask; if (amdgpu_sriov_vf(adev) || sched_policy == KFD_SCHED_POLICY_NO_HWS) adev->pm.pp_feature &= ~PP_GFXOFF_MASK; @@ -3172,6 +3234,7 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev) * always assumed to be lost. */ switch (amdgpu_asic_reset_method(adev)) { + case AMD_RESET_METHOD_LINK: case AMD_RESET_METHOD_BACO: case AMD_RESET_METHOD_MODE1: return true; @@ -3455,6 +3518,7 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev) amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE); amdgpu_amdkfd_suspend(adev, false); + amdgpu_userq_suspend(adev); /* Workaround for ASICs need to disable SMC first */ amdgpu_device_smu_fini_early(adev); @@ -4307,9 +4371,10 @@ int amdgpu_device_init(struct amdgpu_device *adev, amdgpu_sync_create(&adev->isolation[i].active); amdgpu_sync_create(&adev->isolation[i].prev); } - mutex_init(&adev->gfx.kfd_sch_mutex); + mutex_init(&adev->gfx.userq_sch_mutex); mutex_init(&adev->gfx.workload_profile_mutex); mutex_init(&adev->vcn.workload_profile_mutex); + mutex_init(&adev->userq_mutex); amdgpu_device_init_apu_flags(adev); @@ -4329,12 +4394,16 @@ int amdgpu_device_init(struct amdgpu_device *adev, spin_lock_init(&adev->virt.rlcg_reg_lock); spin_lock_init(&adev->wb.lock); + xa_init_flags(&adev->userq_xa, XA_FLAGS_LOCK_IRQ); + INIT_LIST_HEAD(&adev->reset_list); INIT_LIST_HEAD(&adev->ras_list); INIT_LIST_HEAD(&adev->pm.od_kobj_list); + INIT_LIST_HEAD(&adev->userq_mgr_list); + INIT_DELAYED_WORK(&adev->delayed_init_work, amdgpu_device_delayed_init_work_handler); INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work, @@ -4907,28 +4976,20 @@ static int amdgpu_device_evict_resources(struct amdgpu_device *adev) * @data: data * * This function is called when the system is about to suspend or hibernate. - * It is used to evict resources from the device before the system goes to - * sleep while there is still access to swap. + * It is used to set the appropriate flags so that eviction can be optimized + * in the pm prepare callback. */ static int amdgpu_device_pm_notifier(struct notifier_block *nb, unsigned long mode, void *data) { struct amdgpu_device *adev = container_of(nb, struct amdgpu_device, pm_nb); - int r; switch (mode) { case PM_HIBERNATION_PREPARE: adev->in_s4 = true; - fallthrough; - case PM_SUSPEND_PREPARE: - r = amdgpu_device_evict_resources(adev); - /* - * This is considered non-fatal at this time because - * amdgpu_device_prepare() will also fatally evict resources. - * See https://gitlab.freedesktop.org/drm/amd/-/issues/3781 - */ - if (r) - drm_warn(adev_to_drm(adev), "Failed to evict resources, freeze active processes if problems occur: %d\n", r); + break; + case PM_POST_HIBERNATION: + adev->in_s4 = false; break; } @@ -4949,15 +5010,13 @@ int amdgpu_device_prepare(struct drm_device *dev) struct amdgpu_device *adev = drm_to_adev(dev); int i, r; - amdgpu_choose_low_power_state(adev); - if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; /* Evict the majority of BOs before starting suspend sequence */ r = amdgpu_device_evict_resources(adev); if (r) - goto unprepare; + return r; flush_delayed_work(&adev->gfx.gfx_off_delay_work); @@ -4968,15 +5027,10 @@ int amdgpu_device_prepare(struct drm_device *dev) continue; r = adev->ip_blocks[i].version->funcs->prepare_suspend(&adev->ip_blocks[i]); if (r) - goto unprepare; + return r; } return 0; - -unprepare: - adev->in_s0ix = adev->in_s3 = adev->in_s4 = false; - - return r; } /** @@ -5018,8 +5072,10 @@ int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients) amdgpu_device_ip_suspend_phase1(adev); - if (!adev->in_s0ix) + if (!adev->in_s0ix) { amdgpu_amdkfd_suspend(adev, adev->in_runpm); + amdgpu_userq_suspend(adev); + } r = amdgpu_device_evict_resources(adev); if (r) @@ -5086,6 +5142,10 @@ int amdgpu_device_resume(struct drm_device *dev, bool notify_clients) r = amdgpu_amdkfd_resume(adev, adev->in_runpm); if (r) goto exit; + + r = amdgpu_userq_resume(adev); + if (r) + goto exit; } r = amdgpu_device_ip_late_init(adev); @@ -5134,9 +5194,6 @@ exit: } adev->in_suspend = false; - if (adev->enable_mes) - amdgpu_mes_self_test(adev); - if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D0)) DRM_WARN("smart shift update failed\n"); @@ -5517,6 +5574,29 @@ mode1_reset_failed: return ret; } +int amdgpu_device_link_reset(struct amdgpu_device *adev) +{ + int ret = 0; + + dev_info(adev->dev, "GPU link reset\n"); + + if (!adev->pcie_reset_ctx.occurs_dpc) + ret = amdgpu_dpm_link_reset(adev); + + if (ret) + goto link_reset_failed; + + ret = amdgpu_psp_wait_for_bootloader(adev); + if (ret) + goto link_reset_failed; + + return 0; + +link_reset_failed: + dev_err(adev->dev, "GPU link reset failed\n"); + return ret; +} + int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, struct amdgpu_reset_context *reset_context) { @@ -5821,6 +5901,7 @@ static void amdgpu_device_set_mp1_state(struct amdgpu_device *adev) switch (amdgpu_asic_reset_method(adev)) { case AMD_RESET_METHOD_MODE1: + case AMD_RESET_METHOD_LINK: adev->mp1_state = PP_MP1_STATE_SHUTDOWN; break; case AMD_RESET_METHOD_MODE2: @@ -5937,94 +6018,42 @@ static int amdgpu_device_health_check(struct list_head *device_list_handle) return ret; } -/** - * amdgpu_device_gpu_recover - reset the asic and recover scheduler - * - * @adev: amdgpu_device pointer - * @job: which job trigger hang - * @reset_context: amdgpu reset context pointer - * - * Attempt to reset the GPU if it has hung (all asics). - * Attempt to do soft-reset or full-reset and reinitialize Asic - * Returns 0 for success or an error on failure. - */ - -int amdgpu_device_gpu_recover(struct amdgpu_device *adev, +static int amdgpu_device_halt_activities(struct amdgpu_device *adev, struct amdgpu_job *job, - struct amdgpu_reset_context *reset_context) + struct amdgpu_reset_context *reset_context, + struct list_head *device_list, + struct amdgpu_hive_info *hive, + bool need_emergency_restart) { - struct list_head device_list, *device_list_handle = NULL; - bool job_signaled = false; - struct amdgpu_hive_info *hive = NULL; + struct list_head *device_list_handle = NULL; struct amdgpu_device *tmp_adev = NULL; int i, r = 0; - bool need_emergency_restart = false; - bool audio_suspended = false; - int retry_limit = AMDGPU_MAX_RETRY_LIMIT; /* - * If it reaches here because of hang/timeout and a RAS error is - * detected at the same time, let RAS recovery take care of it. - */ - if (amdgpu_ras_is_err_state(adev, AMDGPU_RAS_BLOCK__ANY) && - !amdgpu_sriov_vf(adev) && - reset_context->src != AMDGPU_RESET_SRC_RAS) { - dev_dbg(adev->dev, - "Gpu recovery from source: %d yielding to RAS error recovery handling", - reset_context->src); - return 0; - } - /* - * Special case: RAS triggered and full reset isn't supported - */ - need_emergency_restart = amdgpu_ras_need_emergency_restart(adev); - - /* - * Flush RAM to disk so that after reboot - * the user can read log and see why the system rebooted. - */ - if (need_emergency_restart && amdgpu_ras_get_context(adev) && - amdgpu_ras_get_context(adev)->reboot) { - DRM_WARN("Emergency reboot."); - - ksys_sync_helper(); - emergency_restart(); - } - - dev_info(adev->dev, "GPU %s begin!\n", - need_emergency_restart ? "jobs stop":"reset"); - - if (!amdgpu_sriov_vf(adev)) - hive = amdgpu_get_xgmi_hive(adev); - if (hive) - mutex_lock(&hive->hive_lock); - - reset_context->job = job; - reset_context->hive = hive; - /* * Build list of devices to reset. * In case we are in XGMI hive mode, resort the device list * to put adev in the 1st position. */ - INIT_LIST_HEAD(&device_list); if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1) && hive) { list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) { - list_add_tail(&tmp_adev->reset_list, &device_list); + list_add_tail(&tmp_adev->reset_list, device_list); if (adev->shutdown) tmp_adev->shutdown = true; + if (adev->pcie_reset_ctx.occurs_dpc) + tmp_adev->pcie_reset_ctx.in_link_reset = true; } - if (!list_is_first(&adev->reset_list, &device_list)) - list_rotate_to_front(&adev->reset_list, &device_list); - device_list_handle = &device_list; + if (!list_is_first(&adev->reset_list, device_list)) + list_rotate_to_front(&adev->reset_list, device_list); + device_list_handle = device_list; } else { - list_add_tail(&adev->reset_list, &device_list); - device_list_handle = &device_list; + list_add_tail(&adev->reset_list, device_list); + device_list_handle = device_list; } - if (!amdgpu_sriov_vf(adev)) { + if (!amdgpu_sriov_vf(adev) && (!adev->pcie_reset_ctx.occurs_dpc)) { r = amdgpu_device_health_check(device_list_handle); if (r) - goto end_reset; + return r; } /* We need to lock reset domain only once both for XGMI and single device */ @@ -6048,7 +6077,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, * some audio codec errors. */ if (!amdgpu_device_suspend_display_audio(tmp_adev)) - audio_suspended = true; + tmp_adev->pcie_reset_ctx.audio_suspended = true; amdgpu_ras_set_error_query_ready(tmp_adev, false); @@ -6066,6 +6095,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, /* disable ras on ALL IPs */ if (!need_emergency_restart && + (!adev->pcie_reset_ctx.occurs_dpc) && amdgpu_device_ip_need_full_reset(tmp_adev)) amdgpu_ras_suspend(tmp_adev); @@ -6083,24 +6113,24 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, atomic_inc(&tmp_adev->gpu_reset_counter); } - if (need_emergency_restart) - goto skip_sched_resume; + return r; +} - /* - * Must check guilty signal here since after this point all old - * HW fences are force signaled. - * - * job->base holds a reference to parent fence - */ - if (job && dma_fence_is_signaled(&job->hw_fence)) { - job_signaled = true; - dev_info(adev->dev, "Guilty job already signaled, skipping HW reset"); - goto skip_hw_reset; - } +static int amdgpu_device_asic_reset(struct amdgpu_device *adev, + struct list_head *device_list, + struct amdgpu_reset_context *reset_context) +{ + struct amdgpu_device *tmp_adev = NULL; + int retry_limit = AMDGPU_MAX_RETRY_LIMIT; + int r = 0; retry: /* Rest of adevs pre asic reset from XGMI hive. */ - list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + list_for_each_entry(tmp_adev, device_list, reset_list) { + if (adev->pcie_reset_ctx.occurs_dpc) + tmp_adev->no_hw_access = true; r = amdgpu_device_pre_asic_reset(tmp_adev, reset_context); + if (adev->pcie_reset_ctx.occurs_dpc) + tmp_adev->no_hw_access = false; /*TODO Should we stop ?*/ if (r) { dev_err(tmp_adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ", @@ -6112,6 +6142,11 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ /* Actual ASIC resets if needed.*/ /* Host driver will handle XGMI hive reset for SRIOV */ if (amdgpu_sriov_vf(adev)) { + + /* Bail out of reset early */ + if (amdgpu_ras_is_rma(adev)) + return -ENODEV; + if (amdgpu_ras_get_fed_status(adev) || amdgpu_virt_rcvd_ras_interrupt(adev)) { dev_dbg(adev->dev, "Detected RAS error, wait for FLR completion\n"); amdgpu_ras_set_fed(adev, true); @@ -6126,12 +6161,12 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ if (r) adev->asic_reset_res = r; } else { - r = amdgpu_do_asic_reset(device_list_handle, reset_context); + r = amdgpu_do_asic_reset(device_list, reset_context); if (r && r == -EAGAIN) goto retry; } - list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + list_for_each_entry(tmp_adev, device_list, reset_list) { /* * Drop any pending non scheduler resets queued before reset is done. * Any reset scheduled after this point would be valid. Scheduler resets @@ -6141,10 +6176,18 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ amdgpu_device_stop_pending_resets(tmp_adev); } -skip_hw_reset: + return r; +} + +static int amdgpu_device_sched_resume(struct list_head *device_list, + struct amdgpu_reset_context *reset_context, + bool job_signaled) +{ + struct amdgpu_device *tmp_adev = NULL; + int i, r = 0; /* Post ASIC reset for all devs .*/ - list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + list_for_each_entry(tmp_adev, device_list, reset_list) { for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = tmp_adev->rings[i]; @@ -6180,8 +6223,16 @@ skip_hw_reset: } } -skip_sched_resume: - list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + return r; +} + +static void amdgpu_device_gpu_resume(struct amdgpu_device *adev, + struct list_head *device_list, + bool need_emergency_restart) +{ + struct amdgpu_device *tmp_adev = NULL; + + list_for_each_entry(tmp_adev, device_list, reset_list) { /* unlock kfd: SRIOV would do it separately */ if (!need_emergency_restart && !amdgpu_sriov_vf(tmp_adev)) amdgpu_amdkfd_post_reset(tmp_adev); @@ -6192,18 +6243,114 @@ skip_sched_resume: if (!adev->kfd.init_complete) amdgpu_amdkfd_device_init(adev); - if (audio_suspended) + if (tmp_adev->pcie_reset_ctx.audio_suspended) amdgpu_device_resume_display_audio(tmp_adev); amdgpu_device_unset_mp1_state(tmp_adev); amdgpu_ras_set_error_query_ready(tmp_adev, true); + } - tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device, + tmp_adev = list_first_entry(device_list, struct amdgpu_device, reset_list); amdgpu_device_unlock_reset_domain(tmp_adev->reset_domain); +} + + +/** + * amdgpu_device_gpu_recover - reset the asic and recover scheduler + * + * @adev: amdgpu_device pointer + * @job: which job trigger hang + * @reset_context: amdgpu reset context pointer + * + * Attempt to reset the GPU if it has hung (all asics). + * Attempt to do soft-reset or full-reset and reinitialize Asic + * Returns 0 for success or an error on failure. + */ + +int amdgpu_device_gpu_recover(struct amdgpu_device *adev, + struct amdgpu_job *job, + struct amdgpu_reset_context *reset_context) +{ + struct list_head device_list; + bool job_signaled = false; + struct amdgpu_hive_info *hive = NULL; + int r = 0; + bool need_emergency_restart = false; + + /* + * If it reaches here because of hang/timeout and a RAS error is + * detected at the same time, let RAS recovery take care of it. + */ + if (amdgpu_ras_is_err_state(adev, AMDGPU_RAS_BLOCK__ANY) && + !amdgpu_sriov_vf(adev) && + reset_context->src != AMDGPU_RESET_SRC_RAS) { + dev_dbg(adev->dev, + "Gpu recovery from source: %d yielding to RAS error recovery handling", + reset_context->src); + return 0; + } + + /* + * Special case: RAS triggered and full reset isn't supported + */ + need_emergency_restart = amdgpu_ras_need_emergency_restart(adev); + + /* + * Flush RAM to disk so that after reboot + * the user can read log and see why the system rebooted. + */ + if (need_emergency_restart && amdgpu_ras_get_context(adev) && + amdgpu_ras_get_context(adev)->reboot) { + DRM_WARN("Emergency reboot."); + + ksys_sync_helper(); + emergency_restart(); + } + + dev_info(adev->dev, "GPU %s begin!\n", + need_emergency_restart ? "jobs stop":"reset"); + + if (!amdgpu_sriov_vf(adev)) + hive = amdgpu_get_xgmi_hive(adev); + if (hive) + mutex_lock(&hive->hive_lock); + + reset_context->job = job; + reset_context->hive = hive; + INIT_LIST_HEAD(&device_list); + + r = amdgpu_device_halt_activities(adev, job, reset_context, &device_list, + hive, need_emergency_restart); + if (r) + goto end_reset; + + if (need_emergency_restart) + goto skip_sched_resume; + /* + * Must check guilty signal here since after this point all old + * HW fences are force signaled. + * + * job->base holds a reference to parent fence + */ + if (job && dma_fence_is_signaled(&job->hw_fence)) { + job_signaled = true; + dev_info(adev->dev, "Guilty job already signaled, skipping HW reset"); + goto skip_hw_reset; + } + + r = amdgpu_device_asic_reset(adev, &device_list, reset_context); + if (r) + goto end_reset; +skip_hw_reset: + r = amdgpu_device_sched_resume(&device_list, reset_context, job_signaled); + if (r) + goto end_reset; +skip_sched_resume: + amdgpu_device_gpu_resume(adev, &device_list, need_emergency_restart); end_reset: if (hive) { mutex_unlock(&hive->hive_lock); @@ -6587,12 +6734,15 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta { struct drm_device *dev = pci_get_drvdata(pdev); struct amdgpu_device *adev = drm_to_adev(dev); - int i; + struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev); + struct amdgpu_reset_context reset_context; + struct list_head device_list; + int r = 0; - DRM_INFO("PCI error: detected callback, state(%d)!!\n", state); + dev_info(adev->dev, "PCI error: detected callback!!\n"); - if (adev->gmc.xgmi.num_physical_nodes > 1) { - DRM_WARN("No support for XGMI hive yet..."); + if (!amdgpu_dpm_is_link_reset_supported(adev)) { + dev_warn(adev->dev, "No support for XGMI hive yet...\n"); return PCI_ERS_RESULT_DISCONNECT; } @@ -6600,32 +6750,30 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta switch (state) { case pci_channel_io_normal: + dev_info(adev->dev, "pci_channel_io_normal: state(%d)!!\n", state); return PCI_ERS_RESULT_CAN_RECOVER; - /* Fatal error, prepare for slot reset */ case pci_channel_io_frozen: - /* - * Locking adev->reset_domain->sem will prevent any external access - * to GPU during PCI error recovery - */ - amdgpu_device_lock_reset_domain(adev->reset_domain); - amdgpu_device_set_mp1_state(adev); - - /* - * Block any work scheduling as we do for regular GPU reset - * for the duration of the recovery - */ - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - struct amdgpu_ring *ring = adev->rings[i]; - - if (!amdgpu_ring_sched_ready(ring)) - continue; - - drm_sched_stop(&ring->sched, NULL); + /* Fatal error, prepare for slot reset */ + dev_info(adev->dev, "pci_channel_io_frozen: state(%d)!!\n", state); + + if (hive) + mutex_lock(&hive->hive_lock); + adev->pcie_reset_ctx.occurs_dpc = true; + memset(&reset_context, 0, sizeof(reset_context)); + INIT_LIST_HEAD(&device_list); + + r = amdgpu_device_halt_activities(adev, NULL, &reset_context, &device_list, + hive, false); + if (hive) { + mutex_unlock(&hive->hive_lock); + amdgpu_put_xgmi_hive(hive); } - atomic_inc(&adev->gpu_reset_counter); + if (r) + return PCI_ERS_RESULT_DISCONNECT; return PCI_ERS_RESULT_NEED_RESET; case pci_channel_io_perm_failure: /* Permanent error, prepare for device removal */ + dev_info(adev->dev, "pci_channel_io_perm_failure: state(%d)!!\n", state); return PCI_ERS_RESULT_DISCONNECT; } @@ -6638,8 +6786,10 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta */ pci_ers_result_t amdgpu_pci_mmio_enabled(struct pci_dev *pdev) { + struct drm_device *dev = pci_get_drvdata(pdev); + struct amdgpu_device *adev = drm_to_adev(dev); - DRM_INFO("PCI error: mmio enabled callback!!\n"); + dev_info(adev->dev, "PCI error: mmio enabled callback!!\n"); /* TODO - dump whatever for debugging purposes */ @@ -6663,10 +6813,12 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) { struct drm_device *dev = pci_get_drvdata(pdev); struct amdgpu_device *adev = drm_to_adev(dev); - int r, i; struct amdgpu_reset_context reset_context; - u32 memsize; + struct amdgpu_device *tmp_adev; + struct amdgpu_hive_info *hive; struct list_head device_list; + int r = 0, i; + u32 memsize; /* PCI error slot reset should be skipped During RAS recovery */ if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || @@ -6674,15 +6826,12 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) amdgpu_ras_in_recovery(adev)) return PCI_ERS_RESULT_RECOVERED; - DRM_INFO("PCI error: slot reset callback!!\n"); + dev_info(adev->dev, "PCI error: slot reset callback!!\n"); memset(&reset_context, 0, sizeof(reset_context)); - INIT_LIST_HEAD(&device_list); - list_add_tail(&adev->reset_list, &device_list); - /* wait for asic to come out of reset */ - msleep(500); + msleep(700); /* Restore PCI confspace */ amdgpu_device_load_pci_state(pdev); @@ -6703,26 +6852,40 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags); - - adev->no_hw_access = true; - r = amdgpu_device_pre_asic_reset(adev, &reset_context); - adev->no_hw_access = false; - if (r) - goto out; + set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags); + INIT_LIST_HEAD(&device_list); - r = amdgpu_do_asic_reset(&device_list, &reset_context); + hive = amdgpu_get_xgmi_hive(adev); + if (hive) { + mutex_lock(&hive->hive_lock); + reset_context.hive = hive; + list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) { + tmp_adev->pcie_reset_ctx.in_link_reset = true; + list_add_tail(&tmp_adev->reset_list, &device_list); + } + } else { + set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags); + list_add_tail(&adev->reset_list, &device_list); + } + r = amdgpu_device_asic_reset(adev, &device_list, &reset_context); out: if (!r) { if (amdgpu_device_cache_pci_state(adev->pdev)) pci_restore_state(adev->pdev); - - DRM_INFO("PCIe error recovery succeeded\n"); + dev_info(adev->dev, "PCIe error recovery succeeded\n"); } else { - DRM_ERROR("PCIe error recovery failed, err:%d", r); - amdgpu_device_unset_mp1_state(adev); - amdgpu_device_unlock_reset_domain(adev->reset_domain); + dev_err(adev->dev, "PCIe error recovery failed, err:%d\n", r); + if (hive) { + list_for_each_entry(tmp_adev, &device_list, reset_list) + amdgpu_device_unset_mp1_state(tmp_adev); + amdgpu_device_unlock_reset_domain(adev->reset_domain); + } + } + + if (hive) { + mutex_unlock(&hive->hive_lock); + amdgpu_put_xgmi_hive(hive); } return r ? PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_RECOVERED; @@ -6739,26 +6902,36 @@ void amdgpu_pci_resume(struct pci_dev *pdev) { struct drm_device *dev = pci_get_drvdata(pdev); struct amdgpu_device *adev = drm_to_adev(dev); - int i; - + struct list_head device_list; + struct amdgpu_hive_info *hive = NULL; + struct amdgpu_device *tmp_adev = NULL; - DRM_INFO("PCI error: resume callback!!\n"); + dev_info(adev->dev, "PCI error: resume callback!!\n"); /* Only continue execution for the case of pci_channel_io_frozen */ if (adev->pci_channel_state != pci_channel_io_frozen) return; - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - struct amdgpu_ring *ring = adev->rings[i]; + INIT_LIST_HEAD(&device_list); - if (!amdgpu_ring_sched_ready(ring)) - continue; + hive = amdgpu_get_xgmi_hive(adev); + if (hive) { + mutex_lock(&hive->hive_lock); + list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) { + tmp_adev->pcie_reset_ctx.in_link_reset = false; + list_add_tail(&tmp_adev->reset_list, &device_list); + } + } else + list_add_tail(&adev->reset_list, &device_list); - drm_sched_start(&ring->sched, 0); - } + amdgpu_device_sched_resume(&device_list, NULL, NULL); + amdgpu_device_gpu_resume(adev, &device_list, false); + adev->pcie_reset_ctx.occurs_dpc = false; - amdgpu_device_unset_mp1_state(adev); - amdgpu_device_unlock_reset_domain(adev->reset_domain); + if (hive) { + mutex_unlock(&hive->hive_lock); + amdgpu_put_xgmi_hive(hive); + } } bool amdgpu_device_cache_pci_state(struct pci_dev *pdev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 24ee4710f807..4ddd08ce8885 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -51,6 +51,8 @@ #include "amdgpu_reset.h" #include "amdgpu_sched.h" #include "amdgpu_xgmi.h" +#include "amdgpu_userq.h" +#include "amdgpu_userq_fence.h" #include "../amdxcp/amdgpu_xcp_drv.h" /* @@ -123,9 +125,10 @@ * - 3.61.0 - Contains fix for RV/PCO compute queues * - 3.62.0 - Add AMDGPU_IDS_FLAGS_MODE_PF, AMDGPU_IDS_FLAGS_MODE_VF & AMDGPU_IDS_FLAGS_MODE_PT * - 3.63.0 - GFX12 display DCC supports 256B max compressed block size + * - 3.64.0 - Userq IP support query */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 63 +#define KMS_DRIVER_MINOR 64 #define KMS_DRIVER_PATCHLEVEL 0 /* @@ -140,6 +143,7 @@ enum AMDGPU_DEBUG_MASK { AMDGPU_DEBUG_ENABLE_EXP_RESETS = BIT(5), AMDGPU_DEBUG_DISABLE_GPU_RING_RESET = BIT(6), AMDGPU_DEBUG_SMU_POOL = BIT(7), + AMDGPU_DEBUG_VM_USERPTR = BIT(8), }; unsigned int amdgpu_vram_limit = UINT_MAX; @@ -176,7 +180,7 @@ uint amdgpu_pg_mask = 0xffffffff; uint amdgpu_sdma_phase_quantum = 32; char *amdgpu_disable_cu; char *amdgpu_virtual_display; -bool enforce_isolation; +int amdgpu_enforce_isolation = -1; int amdgpu_modeset = -1; /* Specifies the default granularity for SVM, used in buffer @@ -238,6 +242,8 @@ int amdgpu_agp = -1; /* auto */ int amdgpu_wbrf = -1; int amdgpu_damage_clips = -1; /* auto */ int amdgpu_umsch_mm_fwlog; +int amdgpu_rebar = -1; /* auto */ +int amdgpu_user_queue = -1; DECLARE_DYNDBG_CLASSMAP(drm_debug_classes, DD_CLASS_TYPE_DISJOINT_BITS, 0, "DRM_UT_CORE", @@ -1033,11 +1039,13 @@ module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444); /** - * DOC: enforce_isolation (bool) - * enforce process isolation between graphics and compute via using the same reserved vmid. + * DOC: enforce_isolation (int) + * enforce process isolation between graphics and compute. + * (-1 = auto, 0 = disable, 1 = enable, 2 = enable legacy mode, 3 = enable without cleaner shader) */ -module_param(enforce_isolation, bool, 0444); -MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on"); +module_param_named(enforce_isolation, amdgpu_enforce_isolation, int, 0444); +MODULE_PARM_DESC(enforce_isolation, +"enforce process isolation between graphics and compute. (-1 = auto, 0 = disable, 1 = enable, 2 = enable legacy mode, 3 = enable without cleaner shader)"); /** * DOC: modeset (int) @@ -1096,6 +1104,28 @@ MODULE_PARM_DESC(wbrf, "Enable Wifi RFI interference mitigation (0 = disabled, 1 = enabled, -1 = auto(default)"); module_param_named(wbrf, amdgpu_wbrf, int, 0444); +/** + * DOC: rebar (int) + * Allow BAR resizing. Disable this to prevent the driver from attempting + * to resize the BAR if the GPU supports it and there is available MMIO space. + * Note that this just prevents the driver from resizing the BAR. The BIOS + * may have already resized the BAR at boot time. + */ +MODULE_PARM_DESC(rebar, "Resizable BAR (-1 = auto (default), 0 = disable, 1 = enable)"); +module_param_named(rebar, amdgpu_rebar, int, 0444); + +/** + * DOC: user_queue (int) + * Enable user queues on systems that support user queues. Possible values: + * + * - -1 = auto (ASIC specific default) + * - 0 = user queues disabled + * - 1 = user queues enabled and kernel queues enabled (if supported) + * - 2 = user queues enabled and kernel queues disabled + */ +MODULE_PARM_DESC(user_queue, "Enable user queues (-1 = auto (default), 0 = disable, 1 = enable, 2 = enable UQs and disable KQs)"); +module_param_named(user_queue, amdgpu_user_queue, int, 0444); + /* These devices are not supported by amdgpu. * They are supported by the mach64, r128, radeon drivers */ @@ -2244,6 +2274,10 @@ static void amdgpu_init_debug_options(struct amdgpu_device *adev) pr_info("debug: use vram for smu pool\n"); adev->pm.smu_debug_mask |= SMU_DEBUG_POOL_USE_VRAM; } + if (amdgpu_debug_mask & AMDGPU_DEBUG_VM_USERPTR) { + pr_info("debug: VM mode debug for userptr is enabled\n"); + adev->debug_vm_userptr = true; + } } static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags) @@ -2615,13 +2649,8 @@ static int amdgpu_pmops_freeze(struct device *dev) static int amdgpu_pmops_thaw(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); - struct amdgpu_device *adev = drm_to_adev(drm_dev); - int r; - r = amdgpu_device_resume(drm_dev, true); - adev->in_s4 = false; - - return r; + return amdgpu_device_resume(drm_dev, true); } static int amdgpu_pmops_poweroff(struct device *dev) @@ -2634,9 +2663,6 @@ static int amdgpu_pmops_poweroff(struct device *dev) static int amdgpu_pmops_restore(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); - struct amdgpu_device *adev = drm_to_adev(drm_dev); - - adev->in_s4 = false; return amdgpu_device_resume(drm_dev, true); } @@ -2708,6 +2734,29 @@ static int amdgpu_runtime_idle_check_display(struct device *dev) return 0; } +static int amdgpu_runtime_idle_check_userq(struct device *dev) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + struct amdgpu_device *adev = drm_to_adev(drm_dev); + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_mgr *uqm, *tmp; + int queue_id; + int ret = 0; + + mutex_lock(&adev->userq_mutex); + list_for_each_entry_safe(uqm, tmp, &adev->userq_mgr_list, list) { + idr_for_each_entry(&uqm->userq_idr, queue, queue_id) { + ret = -EBUSY; + goto done; + } + } +done: + mutex_unlock(&adev->userq_mutex); + + return ret; +} + static int amdgpu_pmops_runtime_suspend(struct device *dev) { struct pci_dev *pdev = to_pci_dev(dev); @@ -2723,6 +2772,9 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev) ret = amdgpu_runtime_idle_check_display(dev); if (ret) return ret; + ret = amdgpu_runtime_idle_check_userq(dev); + if (ret) + return ret; /* wait for all rings to drain before suspending */ for (i = 0; i < AMDGPU_MAX_RINGS; i++) { @@ -2844,12 +2896,30 @@ static int amdgpu_pmops_runtime_idle(struct device *dev) } ret = amdgpu_runtime_idle_check_display(dev); + if (ret) + goto done; + ret = amdgpu_runtime_idle_check_userq(dev); +done: pm_runtime_mark_last_busy(dev); pm_runtime_autosuspend(dev); return ret; } +static int amdgpu_drm_release(struct inode *inode, struct file *filp) +{ + struct drm_file *file_priv = filp->private_data; + struct amdgpu_fpriv *fpriv = file_priv->driver_priv; + + if (fpriv) { + fpriv->evf_mgr.fd_closing = true; + amdgpu_userq_mgr_fini(&fpriv->userq_mgr); + amdgpu_eviction_fence_destroy(&fpriv->evf_mgr); + } + + return drm_release(inode, filp); +} + long amdgpu_drm_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { @@ -2901,7 +2971,7 @@ static const struct file_operations amdgpu_driver_kms_fops = { .owner = THIS_MODULE, .open = drm_open, .flush = amdgpu_flush, - .release = drm_release, + .release = amdgpu_drm_release, .unlocked_ioctl = amdgpu_drm_ioctl, .mmap = drm_gem_mmap, .poll = drm_poll, @@ -2948,6 +3018,9 @@ const struct drm_ioctl_desc amdgpu_ioctls_kms[] = { DRM_IOCTL_DEF_DRV(AMDGPU_GEM_VA, amdgpu_gem_va_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(AMDGPU_GEM_OP, amdgpu_gem_op_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(AMDGPU_GEM_USERPTR, amdgpu_gem_userptr_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(AMDGPU_USERQ, amdgpu_userq_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_SIGNAL, amdgpu_userq_signal_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_WAIT, amdgpu_userq_wait_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), }; static const struct drm_driver amdgpu_kms_driver = { @@ -3038,6 +3111,10 @@ static int __init amdgpu_init(void) if (r) goto error_fence; + r = amdgpu_userq_fence_slab_init(); + if (r) + goto error_fence; + DRM_INFO("amdgpu kernel modesetting enabled.\n"); amdgpu_register_atpx_handler(); amdgpu_acpi_detect(); @@ -3069,6 +3146,7 @@ static void __exit amdgpu_exit(void) amdgpu_acpi_release(); amdgpu_sync_fini(); amdgpu_fence_slab_fini(); + amdgpu_userq_fence_slab_fini(); mmu_notifier_synchronize(); amdgpu_xcp_drv_release(); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c new file mode 100644 index 000000000000..73b629b5f56f --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c @@ -0,0 +1,233 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright 2024 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#include <linux/sched.h> +#include <drm/drm_exec.h> +#include "amdgpu.h" + +#define work_to_evf_mgr(w, name) container_of(w, struct amdgpu_eviction_fence_mgr, name) +#define evf_mgr_to_fpriv(e) container_of(e, struct amdgpu_fpriv, evf_mgr) + +static const char * +amdgpu_eviction_fence_get_driver_name(struct dma_fence *fence) +{ + return "amdgpu_eviction_fence"; +} + +static const char * +amdgpu_eviction_fence_get_timeline_name(struct dma_fence *f) +{ + struct amdgpu_eviction_fence *ef; + + ef = container_of(f, struct amdgpu_eviction_fence, base); + return ef->timeline_name; +} + +int +amdgpu_eviction_fence_replace_fence(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct drm_exec *exec) +{ + struct amdgpu_eviction_fence *old_ef, *new_ef; + struct drm_gem_object *obj; + unsigned long index; + int ret; + + if (evf_mgr->ev_fence && + !dma_fence_is_signaled(&evf_mgr->ev_fence->base)) + return 0; + /* + * Steps to replace eviction fence: + * * lock all objects in exec (caller) + * * create a new eviction fence + * * update new eviction fence in evf_mgr + * * attach the new eviction fence to BOs + * * release the old fence + * * unlock the objects (caller) + */ + new_ef = amdgpu_eviction_fence_create(evf_mgr); + if (!new_ef) { + DRM_ERROR("Failed to create new eviction fence\n"); + return -ENOMEM; + } + + /* Update the eviction fence now */ + spin_lock(&evf_mgr->ev_fence_lock); + old_ef = evf_mgr->ev_fence; + evf_mgr->ev_fence = new_ef; + spin_unlock(&evf_mgr->ev_fence_lock); + + /* Attach the new fence */ + drm_exec_for_each_locked_object(exec, index, obj) { + struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); + + if (!bo) + continue; + ret = amdgpu_eviction_fence_attach(evf_mgr, bo); + if (ret) { + DRM_ERROR("Failed to attch new eviction fence\n"); + goto free_err; + } + } + + /* Free old fence */ + if (old_ef) + dma_fence_put(&old_ef->base); + return 0; + +free_err: + kfree(new_ef); + return ret; +} + +static void +amdgpu_eviction_fence_suspend_worker(struct work_struct *work) +{ + struct amdgpu_eviction_fence_mgr *evf_mgr = work_to_evf_mgr(work, suspend_work.work); + struct amdgpu_fpriv *fpriv = evf_mgr_to_fpriv(evf_mgr); + struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr; + struct amdgpu_eviction_fence *ev_fence; + + mutex_lock(&uq_mgr->userq_mutex); + ev_fence = evf_mgr->ev_fence; + if (!ev_fence) + goto unlock; + + amdgpu_userq_evict(uq_mgr, ev_fence); + +unlock: + mutex_unlock(&uq_mgr->userq_mutex); +} + +static bool amdgpu_eviction_fence_enable_signaling(struct dma_fence *f) +{ + struct amdgpu_eviction_fence_mgr *evf_mgr; + struct amdgpu_eviction_fence *ev_fence; + + if (!f) + return true; + + ev_fence = to_ev_fence(f); + evf_mgr = ev_fence->evf_mgr; + + schedule_delayed_work(&evf_mgr->suspend_work, 0); + return true; +} + +static const struct dma_fence_ops amdgpu_eviction_fence_ops = { + .use_64bit_seqno = true, + .get_driver_name = amdgpu_eviction_fence_get_driver_name, + .get_timeline_name = amdgpu_eviction_fence_get_timeline_name, + .enable_signaling = amdgpu_eviction_fence_enable_signaling, +}; + +void amdgpu_eviction_fence_signal(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct amdgpu_eviction_fence *ev_fence) +{ + spin_lock(&evf_mgr->ev_fence_lock); + dma_fence_signal(&ev_fence->base); + spin_unlock(&evf_mgr->ev_fence_lock); +} + +struct amdgpu_eviction_fence * +amdgpu_eviction_fence_create(struct amdgpu_eviction_fence_mgr *evf_mgr) +{ + struct amdgpu_eviction_fence *ev_fence; + + ev_fence = kzalloc(sizeof(*ev_fence), GFP_KERNEL); + if (!ev_fence) + return NULL; + + ev_fence->evf_mgr = evf_mgr; + get_task_comm(ev_fence->timeline_name, current); + spin_lock_init(&ev_fence->lock); + dma_fence_init(&ev_fence->base, &amdgpu_eviction_fence_ops, + &ev_fence->lock, evf_mgr->ev_fence_ctx, + atomic_inc_return(&evf_mgr->ev_fence_seq)); + return ev_fence; +} + +void amdgpu_eviction_fence_destroy(struct amdgpu_eviction_fence_mgr *evf_mgr) +{ + struct amdgpu_eviction_fence *ev_fence; + + /* Wait for any pending work to execute */ + flush_delayed_work(&evf_mgr->suspend_work); + + spin_lock(&evf_mgr->ev_fence_lock); + ev_fence = evf_mgr->ev_fence; + spin_unlock(&evf_mgr->ev_fence_lock); + + if (!ev_fence) + return; + + dma_fence_wait(&ev_fence->base, false); + + /* Last unref of ev_fence */ + dma_fence_put(&ev_fence->base); +} + +int amdgpu_eviction_fence_attach(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct amdgpu_bo *bo) +{ + struct amdgpu_eviction_fence *ev_fence; + struct dma_resv *resv = bo->tbo.base.resv; + int ret; + + if (!resv) + return 0; + + ret = dma_resv_reserve_fences(resv, 1); + if (ret) { + DRM_DEBUG_DRIVER("Failed to resv fence space\n"); + return ret; + } + + spin_lock(&evf_mgr->ev_fence_lock); + ev_fence = evf_mgr->ev_fence; + if (ev_fence) + dma_resv_add_fence(resv, &ev_fence->base, DMA_RESV_USAGE_BOOKKEEP); + spin_unlock(&evf_mgr->ev_fence_lock); + + return 0; +} + +void amdgpu_eviction_fence_detach(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct amdgpu_bo *bo) +{ + struct dma_fence *stub = dma_fence_get_stub(); + + dma_resv_replace_fences(bo->tbo.base.resv, evf_mgr->ev_fence_ctx, + stub, DMA_RESV_USAGE_BOOKKEEP); + dma_fence_put(stub); +} + +int amdgpu_eviction_fence_init(struct amdgpu_eviction_fence_mgr *evf_mgr) +{ + /* This needs to be done one time per open */ + atomic_set(&evf_mgr->ev_fence_seq, 0); + evf_mgr->ev_fence_ctx = dma_fence_context_alloc(1); + spin_lock_init(&evf_mgr->ev_fence_lock); + + INIT_DELAYED_WORK(&evf_mgr->suspend_work, amdgpu_eviction_fence_suspend_worker); + return 0; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h new file mode 100644 index 000000000000..fcd867b7147d --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h @@ -0,0 +1,69 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef AMDGPU_EV_FENCE_H_ +#define AMDGPU_EV_FENCE_H_ + +struct amdgpu_eviction_fence { + struct dma_fence base; + spinlock_t lock; + char timeline_name[TASK_COMM_LEN]; + struct amdgpu_eviction_fence_mgr *evf_mgr; +}; + +struct amdgpu_eviction_fence_mgr { + u64 ev_fence_ctx; + atomic_t ev_fence_seq; + spinlock_t ev_fence_lock; + struct amdgpu_eviction_fence *ev_fence; + struct delayed_work suspend_work; + uint8_t fd_closing; +}; + +/* Eviction fence helper functions */ +struct amdgpu_eviction_fence * +amdgpu_eviction_fence_create(struct amdgpu_eviction_fence_mgr *evf_mgr); + +void +amdgpu_eviction_fence_destroy(struct amdgpu_eviction_fence_mgr *evf_mgr); + +int +amdgpu_eviction_fence_attach(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct amdgpu_bo *bo); + +void +amdgpu_eviction_fence_detach(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct amdgpu_bo *bo); + +int +amdgpu_eviction_fence_init(struct amdgpu_eviction_fence_mgr *evf_mgr); + +void +amdgpu_eviction_fence_signal(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct amdgpu_eviction_fence *ev_fence); + +int +amdgpu_eviction_fence_replace_fence(struct amdgpu_eviction_fence_mgr *evf_mgr, + struct drm_exec *exec); +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 69429df09477..2c68118fe9fd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -36,6 +36,7 @@ #include <drm/drm_exec.h> #include <drm/drm_gem_ttm_helper.h> #include <drm/ttm/ttm_tt.h> +#include <drm/drm_syncobj.h> #include "amdgpu.h" #include "amdgpu_display.h" @@ -44,6 +45,114 @@ #include "amdgpu_xgmi.h" #include "amdgpu_vm.h" +static int +amdgpu_gem_add_input_fence(struct drm_file *filp, + uint64_t syncobj_handles_array, + uint32_t num_syncobj_handles) +{ + struct dma_fence *fence; + uint32_t *syncobj_handles; + int ret, i; + + if (!num_syncobj_handles) + return 0; + + syncobj_handles = memdup_user(u64_to_user_ptr(syncobj_handles_array), + sizeof(uint32_t) * num_syncobj_handles); + if (IS_ERR(syncobj_handles)) + return PTR_ERR(syncobj_handles); + + for (i = 0; i < num_syncobj_handles; i++) { + + if (!syncobj_handles[i]) { + ret = -EINVAL; + goto free_memdup; + } + + ret = drm_syncobj_find_fence(filp, syncobj_handles[i], 0, 0, &fence); + if (ret) + goto free_memdup; + + dma_fence_wait(fence, false); + + /* TODO: optimize async handling */ + dma_fence_put(fence); + } + +free_memdup: + kfree(syncobj_handles); + return ret; +} + +static int +amdgpu_gem_update_timeline_node(struct drm_file *filp, + uint32_t syncobj_handle, + uint64_t point, + struct drm_syncobj **syncobj, + struct dma_fence_chain **chain) +{ + if (!syncobj_handle) + return 0; + + /* Find the sync object */ + *syncobj = drm_syncobj_find(filp, syncobj_handle); + if (!*syncobj) + return -ENOENT; + + if (!point) + return 0; + + /* Allocate the chain node */ + *chain = dma_fence_chain_alloc(); + if (!*chain) { + drm_syncobj_put(*syncobj); + return -ENOMEM; + } + + return 0; +} + +static void +amdgpu_gem_update_bo_mapping(struct drm_file *filp, + struct amdgpu_bo_va *bo_va, + uint32_t operation, + uint64_t point, + struct dma_fence *fence, + struct drm_syncobj *syncobj, + struct dma_fence_chain *chain) +{ + struct amdgpu_bo *bo = bo_va ? bo_va->base.bo : NULL; + struct amdgpu_fpriv *fpriv = filp->driver_priv; + struct amdgpu_vm *vm = &fpriv->vm; + struct dma_fence *last_update; + + if (!syncobj) + return; + + /* Find the last update fence */ + switch (operation) { + case AMDGPU_VA_OP_MAP: + case AMDGPU_VA_OP_REPLACE: + if (bo && (bo->tbo.base.resv == vm->root.bo->tbo.base.resv)) + last_update = vm->last_update; + else + last_update = bo_va->last_pt_update; + break; + case AMDGPU_VA_OP_UNMAP: + case AMDGPU_VA_OP_CLEAR: + last_update = fence; + break; + default: + return; + } + + /* Add fence to timeline */ + if (!point) + drm_syncobj_replace_fence(syncobj, last_update); + else + drm_syncobj_add_point(syncobj, chain, last_update, point); +} + static vm_fault_t amdgpu_gem_fault(struct vm_fault *vmf) { struct ttm_buffer_object *bo = vmf->vma->vm_private_data; @@ -184,6 +293,15 @@ static int amdgpu_gem_object_open(struct drm_gem_object *obj, bo_va = amdgpu_vm_bo_add(adev, vm, abo); else ++bo_va->ref_count; + + /* attach gfx eviction fence */ + r = amdgpu_eviction_fence_attach(&fpriv->evf_mgr, abo); + if (r) { + DRM_DEBUG_DRIVER("Failed to attach eviction fence to BO\n"); + amdgpu_bo_unreserve(abo); + return r; + } + amdgpu_bo_unreserve(abo); /* Validate and add eviction fence to DMABuf imports with dynamic @@ -247,6 +365,9 @@ static void amdgpu_gem_object_close(struct drm_gem_object *obj, goto out_unlock; } + if (!amdgpu_vm_is_bo_always_valid(vm, bo)) + amdgpu_eviction_fence_detach(&fpriv->evf_mgr, bo); + bo_va = amdgpu_vm_bo_find(vm, bo); if (!bo_va || --bo_va->ref_count) goto out_unlock; @@ -321,10 +442,6 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, uint32_t handle, initial_domain; int r; - /* reject DOORBELLs until userspace code to use it is available */ - if (args->in.domains & AMDGPU_GEM_DOMAIN_DOORBELL) - return -EINVAL; - /* reject invalid gem flags */ if (flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | AMDGPU_GEM_CREATE_NO_CPU_ACCESS | @@ -638,18 +755,23 @@ out: * * Update the bo_va directly after setting its address. Errors are not * vital here, so they are not reported back to userspace. + * + * Returns resulting fence if freed BO(s) got cleared from the PT. + * otherwise stub fence in case of error. */ -static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - struct amdgpu_bo_va *bo_va, - uint32_t operation) +static struct dma_fence * +amdgpu_gem_va_update_vm(struct amdgpu_device *adev, + struct amdgpu_vm *vm, + struct amdgpu_bo_va *bo_va, + uint32_t operation) { + struct dma_fence *fence = dma_fence_get_stub(); int r; if (!amdgpu_vm_ready(vm)) - return; + return fence; - r = amdgpu_vm_clear_freed(adev, vm, NULL); + r = amdgpu_vm_clear_freed(adev, vm, &fence); if (r) goto error; @@ -665,6 +787,8 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev, error: if (r && r != -ERESTARTSYS) DRM_ERROR("Couldn't update BO_VA (%d)\n", r); + + return fence; } /** @@ -713,6 +837,9 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, struct amdgpu_fpriv *fpriv = filp->driver_priv; struct amdgpu_bo *abo; struct amdgpu_bo_va *bo_va; + struct drm_syncobj *timeline_syncobj = NULL; + struct dma_fence_chain *timeline_chain = NULL; + struct dma_fence *fence; struct drm_exec exec; uint64_t va_flags; uint64_t vm_size; @@ -774,6 +901,12 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, abo = NULL; } + r = amdgpu_gem_add_input_fence(filp, + args->input_fence_syncobj_handles, + args->num_syncobj_handles); + if (r) + goto error_put_gobj; + drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT | DRM_EXEC_IGNORE_DUPLICATES, 0); drm_exec_until_all_locked(&exec) { @@ -802,6 +935,14 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, bo_va = NULL; } + r = amdgpu_gem_update_timeline_node(filp, + args->vm_timeline_syncobj_out, + args->vm_timeline_point, + &timeline_syncobj, + &timeline_chain); + if (r) + goto error; + switch (args->operation) { case AMDGPU_VA_OP_MAP: va_flags = amdgpu_gem_va_map_flags(adev, args->flags); @@ -827,12 +968,24 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, default: break; } - if (!r && !(args->flags & AMDGPU_VM_DELAY_UPDATE) && !adev->debug_vm) - amdgpu_gem_va_update_vm(adev, &fpriv->vm, bo_va, - args->operation); + if (!r && !(args->flags & AMDGPU_VM_DELAY_UPDATE) && !adev->debug_vm) { + fence = amdgpu_gem_va_update_vm(adev, &fpriv->vm, bo_va, + args->operation); + + if (timeline_syncobj) + amdgpu_gem_update_bo_mapping(filp, bo_va, + args->operation, + args->vm_timeline_point, + fence, timeline_syncobj, + timeline_chain); + else + dma_fence_put(fence); + + } error: drm_exec_fini(&exec); +error_put_gobj: drm_gem_object_put(gobj); return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index cf2df7790077..1db1e6ec0184 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c @@ -33,6 +33,7 @@ #include "amdgpu_reset.h" #include "amdgpu_xcp.h" #include "amdgpu_xgmi.h" +#include "nvd.h" /* delay 0.1 second to enable gfx off feature */ #define GFX_OFF_DELAY_ENABLE msecs_to_jiffies(100) @@ -74,14 +75,15 @@ bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, adev->gfx.mec_bitmap[xcc_id].queue_bitmap); } -int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, - int me, int pipe, int queue) +static int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, + int me, int pipe, int queue) { + int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */ int bit = 0; bit += me * adev->gfx.me.num_pipe_per_me - * adev->gfx.me.num_queue_per_pipe; - bit += pipe * adev->gfx.me.num_queue_per_pipe; + * num_queue_per_pipe; + bit += pipe * num_queue_per_pipe; bit += queue; return bit; @@ -238,8 +240,8 @@ void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev) { int i, queue, pipe; bool multipipe_policy = amdgpu_gfx_is_graphics_multipipe_capable(adev); - int max_queues_per_me = adev->gfx.me.num_pipe_per_me * - adev->gfx.me.num_queue_per_pipe; + int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */ + int max_queues_per_me = adev->gfx.me.num_pipe_per_me * num_queue_per_pipe; if (multipipe_policy) { /* policy: amdgpu owns the first queue per pipe at this stage @@ -247,9 +249,9 @@ void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev) for (i = 0; i < max_queues_per_me; i++) { pipe = i % adev->gfx.me.num_pipe_per_me; queue = (i / adev->gfx.me.num_pipe_per_me) % - adev->gfx.me.num_queue_per_pipe; + num_queue_per_pipe; - set_bit(pipe * adev->gfx.me.num_queue_per_pipe + queue, + set_bit(pipe * num_queue_per_pipe + queue, adev->gfx.me.queue_bitmap); } } else { @@ -258,8 +260,9 @@ void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev) } /* update the number of active graphics rings */ - adev->gfx.num_gfx_rings = - bitmap_weight(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES); + if (adev->gfx.num_gfx_rings) + adev->gfx.num_gfx_rings = + bitmap_weight(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES); } static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev, @@ -1351,6 +1354,10 @@ static ssize_t amdgpu_gfx_get_current_compute_partition(struct device *dev, struct amdgpu_device *adev = drm_to_adev(ddev); int mode; + /* Only minimal precaution taken to reject requests while in reset.*/ + if (amdgpu_in_reset(adev)) + return -EPERM; + mode = amdgpu_xcp_query_partition_mode(adev->xcp_mgr, AMDGPU_XCP_FL_NONE); @@ -1394,8 +1401,14 @@ static ssize_t amdgpu_gfx_set_compute_partition(struct device *dev, return -EINVAL; } + /* Don't allow a switch while under reset */ + if (!down_read_trylock(&adev->reset_domain->sem)) + return -EPERM; + ret = amdgpu_xcp_switch_partition_mode(adev->xcp_mgr, mode); + up_read(&adev->reset_domain->sem); + if (ret) return ret; @@ -1466,6 +1479,8 @@ static int amdgpu_gfx_run_cleaner_shader_job(struct amdgpu_ring *ring) goto err; job->enforce_isolation = true; + /* always run the cleaner shader */ + job->run_cleaner_shader = true; ib = &job->ibs[0]; for (i = 0; i <= ring->funcs->align_mask; ++i) @@ -1552,6 +1567,9 @@ static ssize_t amdgpu_gfx_set_run_cleaner_shader(struct device *dev, if (adev->in_suspend && !adev->in_runpm) return -EPERM; + if (adev->gfx.disable_kq) + return -EPERM; + ret = kstrtol(buf, 0, &value); if (ret) @@ -1594,7 +1612,8 @@ static ssize_t amdgpu_gfx_set_run_cleaner_shader(struct device *dev, * Provides the sysfs read interface to get the current settings of the 'enforce_isolation' * feature for each GPU partition. Reading from the 'enforce_isolation' * sysfs file returns the isolation settings for all partitions, where '0' - * indicates disabled and '1' indicates enabled. + * indicates disabled, '1' indicates enabled, and '2' indicates enabled in legacy mode, + * and '3' indicates enabled without cleaner shader. * * Return: The number of bytes read from the sysfs file. */ @@ -1629,9 +1648,12 @@ static ssize_t amdgpu_gfx_get_enforce_isolation(struct device *dev, * @count: The size of the input data * * This function allows control over the 'enforce_isolation' feature, which - * serializes access to the graphics engine. Writing '1' or '0' to the - * 'enforce_isolation' sysfs file enables or disables process isolation for - * each partition. The input should specify the setting for all partitions. + * serializes access to the graphics engine. Writing '0' to disable, '1' to + * enable isolation with cleaner shader, '2' to enable legacy isolation without + * cleaner shader, or '3' to enable process isolation without submitting the + * cleaner shader to the 'enforce_isolation' sysfs file sets the isolation mode + * for each partition. The input should specify the setting for all + * partitions. * * Return: The number of bytes written to the sysfs file. */ @@ -1668,13 +1690,34 @@ static ssize_t amdgpu_gfx_set_enforce_isolation(struct device *dev, return -EINVAL; for (i = 0; i < num_partitions; i++) { - if (partition_values[i] != 0 && partition_values[i] != 1) + if (partition_values[i] != 0 && + partition_values[i] != 1 && + partition_values[i] != 2 && + partition_values[i] != 3) return -EINVAL; } mutex_lock(&adev->enforce_isolation_mutex); - for (i = 0; i < num_partitions; i++) - adev->enforce_isolation[i] = partition_values[i]; + for (i = 0; i < num_partitions; i++) { + switch (partition_values[i]) { + case 0: + default: + adev->enforce_isolation[i] = AMDGPU_ENFORCE_ISOLATION_DISABLE; + break; + case 1: + adev->enforce_isolation[i] = + AMDGPU_ENFORCE_ISOLATION_ENABLE; + break; + case 2: + adev->enforce_isolation[i] = + AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY; + break; + case 3: + adev->enforce_isolation[i] = + AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER; + break; + } + } mutex_unlock(&adev->enforce_isolation_mutex); amdgpu_mes_update_enforce_isolation(adev); @@ -1923,39 +1966,41 @@ void amdgpu_gfx_cleaner_shader_init(struct amdgpu_device *adev, static void amdgpu_gfx_kfd_sch_ctrl(struct amdgpu_device *adev, u32 idx, bool enable) { - mutex_lock(&adev->gfx.kfd_sch_mutex); + mutex_lock(&adev->gfx.userq_sch_mutex); if (enable) { /* If the count is already 0, it means there's an imbalance bug somewhere. * Note that the bug may be in a different caller than the one which triggers the * WARN_ON_ONCE. */ - if (WARN_ON_ONCE(adev->gfx.kfd_sch_req_count[idx] == 0)) { + if (WARN_ON_ONCE(adev->gfx.userq_sch_req_count[idx] == 0)) { dev_err(adev->dev, "Attempted to enable KFD scheduler when reference count is already zero\n"); goto unlock; } - adev->gfx.kfd_sch_req_count[idx]--; + adev->gfx.userq_sch_req_count[idx]--; - if (adev->gfx.kfd_sch_req_count[idx] == 0 && - adev->gfx.kfd_sch_inactive[idx]) { + if (adev->gfx.userq_sch_req_count[idx] == 0 && + adev->gfx.userq_sch_inactive[idx]) { schedule_delayed_work(&adev->gfx.enforce_isolation[idx].work, msecs_to_jiffies(adev->gfx.enforce_isolation_time[idx])); } } else { - if (adev->gfx.kfd_sch_req_count[idx] == 0) { + if (adev->gfx.userq_sch_req_count[idx] == 0) { cancel_delayed_work_sync(&adev->gfx.enforce_isolation[idx].work); - if (!adev->gfx.kfd_sch_inactive[idx]) { - amdgpu_amdkfd_stop_sched(adev, idx); - adev->gfx.kfd_sch_inactive[idx] = true; + if (!adev->gfx.userq_sch_inactive[idx]) { + amdgpu_userq_stop_sched_for_enforce_isolation(adev, idx); + if (adev->kfd.init_complete) + amdgpu_amdkfd_stop_sched(adev, idx); + adev->gfx.userq_sch_inactive[idx] = true; } } - adev->gfx.kfd_sch_req_count[idx]++; + adev->gfx.userq_sch_req_count[idx]++; } unlock: - mutex_unlock(&adev->gfx.kfd_sch_mutex); + mutex_unlock(&adev->gfx.userq_sch_mutex); } /** @@ -2000,12 +2045,13 @@ void amdgpu_gfx_enforce_isolation_handler(struct work_struct *work) msecs_to_jiffies(1)); } else { /* Tell KFD to resume the runqueue */ - if (adev->kfd.init_complete) { - WARN_ON_ONCE(!adev->gfx.kfd_sch_inactive[idx]); - WARN_ON_ONCE(adev->gfx.kfd_sch_req_count[idx]); + WARN_ON_ONCE(!adev->gfx.userq_sch_inactive[idx]); + WARN_ON_ONCE(adev->gfx.userq_sch_req_count[idx]); + + amdgpu_userq_start_sched_for_enforce_isolation(adev, idx); + if (adev->kfd.init_complete) amdgpu_amdkfd_start_sched(adev, idx); - adev->gfx.kfd_sch_inactive[idx] = false; - } + adev->gfx.userq_sch_inactive[idx] = false; } mutex_unlock(&adev->enforce_isolation_mutex); } @@ -2029,7 +2075,7 @@ amdgpu_gfx_enforce_isolation_wait_for_kfd(struct amdgpu_device *adev, bool wait = false; mutex_lock(&adev->enforce_isolation_mutex); - if (adev->enforce_isolation[idx]) { + if (adev->enforce_isolation[idx] == AMDGPU_ENFORCE_ISOLATION_ENABLE) { /* set the initial values if nothing is set */ if (!adev->gfx.enforce_isolation_jiffies[idx]) { adev->gfx.enforce_isolation_jiffies[idx] = jiffies; @@ -2096,7 +2142,7 @@ void amdgpu_gfx_enforce_isolation_ring_begin_use(struct amdgpu_ring *ring) amdgpu_gfx_enforce_isolation_wait_for_kfd(adev, idx); mutex_lock(&adev->enforce_isolation_mutex); - if (adev->enforce_isolation[idx]) { + if (adev->enforce_isolation[idx] == AMDGPU_ENFORCE_ISOLATION_ENABLE) { if (adev->kfd.init_complete) sched_work = true; } @@ -2133,7 +2179,7 @@ void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring) return; mutex_lock(&adev->enforce_isolation_mutex); - if (adev->enforce_isolation[idx]) { + if (adev->enforce_isolation[idx] == AMDGPU_ENFORCE_ISOLATION_ENABLE) { if (adev->kfd.init_complete) sched_work = true; } @@ -2217,6 +2263,74 @@ void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring) schedule_delayed_work(&ring->adev->gfx.idle_work, GFX_PROFILE_IDLE_TIMEOUT); } +/** + * amdgpu_gfx_csb_preamble_start - Set CSB preamble start + * + * @buffer: This is an output variable that gets the PACKET3 preamble setup. + * + * Return: + * return the latest index. + */ +u32 amdgpu_gfx_csb_preamble_start(volatile u32 *buffer) +{ + u32 count = 0; + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); + buffer[count++] = cpu_to_le32(0x80000000); + buffer[count++] = cpu_to_le32(0x80000000); + + return count; +} + +/** + * amdgpu_gfx_csb_data_parser - Parser CS data + * + * @adev: amdgpu_device pointer used to get the CS data and other gfx info. + * @buffer: This is an output variable that gets the PACKET3 preamble end. + * @count: Index to start set the preemble end. + * + * Return: + * return the latest index. + */ +u32 amdgpu_gfx_csb_data_parser(struct amdgpu_device *adev, volatile u32 *buffer, u32 count) +{ + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + u32 i; + + for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) { + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); + buffer[count++] = cpu_to_le32(ext->reg_index - PACKET3_SET_CONTEXT_REG_START); + + for (i = 0; i < ext->reg_count; i++) + buffer[count++] = cpu_to_le32(ext->extent[i]); + } + } + } + + return count; +} + +/** + * amdgpu_gfx_csb_preamble_end - Set CSB preamble end + * + * @buffer: This is an output variable that gets the PACKET3 preamble end. + * @count: Index to start set the preemble end. + */ +void amdgpu_gfx_csb_preamble_end(volatile u32 *buffer, u32 count) +{ + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); + buffer[count++] = cpu_to_le32(0); +} + /* * debugfs for to enable/disable gfx job submission to specific core. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index 87e862188766..08f268dab8f5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -170,10 +170,46 @@ struct amdgpu_kiq { #define AMDGPU_GFX_MAX_SE 4 #define AMDGPU_GFX_MAX_SH_PER_SE 2 +/** + * amdgpu_rb_config - Configure a single Render Backend (RB) + * + * Bad RBs are fused off and there is a harvest register the driver reads to + * determine which RB(s) are fused off so that the driver can configure the + * hardware state so that nothing gets sent to them. There are also user + * harvest registers that the driver can program to disable additional RBs, + * etc., for testing purposes. + */ struct amdgpu_rb_config { + /** + * @rb_backend_disable: + * + * The value captured from register RB_BACKEND_DISABLE indicates if the + * RB backend is disabled or not. + */ uint32_t rb_backend_disable; + + /** + * @user_rb_backend_disable: + * + * The value captured from register USER_RB_BACKEND_DISABLE indicates + * if the User RB backend is disabled or not. + */ uint32_t user_rb_backend_disable; + + /** + * @raster_config: + * + * To set up all of the states, it is necessary to have two registers + * to keep all of the states. This field holds the first register. + */ uint32_t raster_config; + + /** + * @raster_config_1: + * + * To set up all of the states, it is necessary to have two registers + * to keep all of the states. This field holds the second register. + */ uint32_t raster_config_1; }; @@ -221,6 +257,13 @@ struct amdgpu_gfx_config { uint32_t macrotile_mode_array[16]; struct gb_addr_config gb_addr_config_fields; + + /** + * @rb_config: + * + * Matrix that keeps all the Render Backend (color and depth buffer + * handling) configuration on the 3D engine. + */ struct amdgpu_rb_config rb_config[AMDGPU_GFX_MAX_SE][AMDGPU_GFX_MAX_SH_PER_SE]; /* gfx configure feature */ @@ -305,7 +348,8 @@ struct amdgpu_gfx_funcs { void (*init_spm_golden)(struct amdgpu_device *adev); void (*update_perfmon_mgcg)(struct amdgpu_device *adev, bool enable); int (*get_gfx_shadow_info)(struct amdgpu_device *adev, - struct amdgpu_gfx_shadow_info *shadow_info); + struct amdgpu_gfx_shadow_info *shadow_info, + bool skip_check); enum amdgpu_gfx_partition (*query_partition_mode)(struct amdgpu_device *adev); int (*switch_partition_mode)(struct amdgpu_device *adev, @@ -474,9 +518,9 @@ struct amdgpu_gfx { bool enable_cleaner_shader; struct amdgpu_isolation_work enforce_isolation[MAX_XCP]; /* Mutex for synchronizing KFD scheduler operations */ - struct mutex kfd_sch_mutex; - u64 kfd_sch_req_count[MAX_XCP]; - bool kfd_sch_inactive[MAX_XCP]; + struct mutex userq_sch_mutex; + u64 userq_sch_req_count[MAX_XCP]; + bool userq_sch_inactive[MAX_XCP]; unsigned long enforce_isolation_jiffies[MAX_XCP]; unsigned long enforce_isolation_time[MAX_XCP]; @@ -484,6 +528,9 @@ struct amdgpu_gfx { struct delayed_work idle_work; bool workload_profile_active; struct mutex workload_profile_mutex; + + bool disable_kq; + bool disable_uq; }; struct amdgpu_gfx_ras_reg_entry { @@ -503,7 +550,7 @@ struct amdgpu_gfx_ras_mem_id_entry { #define amdgpu_gfx_select_se_sh(adev, se, sh, instance, xcc_id) ((adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance), (xcc_id))) #define amdgpu_gfx_select_me_pipe_q(adev, me, pipe, q, vmid, xcc_id) ((adev)->gfx.funcs->select_me_pipe_q((adev), (me), (pipe), (q), (vmid), (xcc_id))) #define amdgpu_gfx_init_spm_golden(adev) (adev)->gfx.funcs->init_spm_golden((adev)) -#define amdgpu_gfx_get_gfx_shadow_info(adev, si) ((adev)->gfx.funcs->get_gfx_shadow_info((adev), (si))) +#define amdgpu_gfx_get_gfx_shadow_info(adev, si) ((adev)->gfx.funcs->get_gfx_shadow_info((adev), (si), false)) /** * amdgpu_gfx_create_bitmask - create a bitmask @@ -550,8 +597,6 @@ bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring); bool amdgpu_gfx_is_high_priority_graphics_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring); -int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, int me, - int pipe, int queue); bool amdgpu_gfx_is_me_queue_enabled(struct amdgpu_device *adev, int me, int pipe, int queue); void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable); @@ -597,6 +642,9 @@ void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring); void amdgpu_gfx_profile_idle_work_handler(struct work_struct *work); void amdgpu_gfx_profile_ring_begin_use(struct amdgpu_ring *ring); void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring); +u32 amdgpu_gfx_csb_preamble_start(volatile u32 *buffer); +u32 amdgpu_gfx_csb_data_parser(struct amdgpu_device *adev, volatile u32 *buffer, u32 count); +void amdgpu_gfx_csb_preamble_end(volatile u32 *buffer, u32 count); void amdgpu_debugfs_gfx_sched_mask_init(struct amdgpu_device *adev); void amdgpu_debugfs_compute_sched_mask_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index ecb74ccf1d90..6b0fbbb91e57 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -1230,6 +1230,10 @@ static ssize_t current_memory_partition_show( struct amdgpu_device *adev = drm_to_adev(ddev); enum amdgpu_memory_partition mode; + /* Only minimal precaution taken to reject requests while in reset */ + if (amdgpu_in_reset(adev)) + return -EPERM; + mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); if ((mode >= ARRAY_SIZE(nps_desc)) || (BIT(mode) & AMDGPU_ALL_NPS_MASK) != BIT(mode)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h index bd7fc123b8f9..80fa29c26e9e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h @@ -62,6 +62,9 @@ */ #define AMDGPU_GMC_FAULT_TIMEOUT 5000ULL +/* XNACK flags */ +#define AMDGPU_GMC_XNACK_FLAG_CHAIN BIT(0) + struct firmware; enum amdgpu_memory_partition { @@ -301,6 +304,7 @@ struct amdgpu_gmc { struct amdgpu_xgmi xgmi; struct amdgpu_irq_src ecc_irq; int noretry; + uint32_t xnack_flags; uint32_t vmid0_page_table_block_size; uint32_t vmid0_page_table_depth; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c index b6cf801939aa..6e02fb9ac2f6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c @@ -22,6 +22,7 @@ */ #include "amdgpu.h" #include "amdgpu_ras.h" +#include <uapi/linux/kfd_ioctl.h> int amdgpu_hdp_ras_sw_init(struct amdgpu_device *adev) { @@ -46,3 +47,22 @@ int amdgpu_hdp_ras_sw_init(struct amdgpu_device *adev) /* hdp ras follows amdgpu_ras_block_late_init_default for late init */ return 0; } + +void amdgpu_hdp_generic_flush(struct amdgpu_device *adev, + struct amdgpu_ring *ring) +{ + if (!ring || !ring->funcs->emit_wreg) { + WREG32((adev->rmmio_remap.reg_offset + + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> + 2, + 0); + if (adev->nbio.funcs->get_memsize) + adev->nbio.funcs->get_memsize(adev); + } else { + amdgpu_ring_emit_wreg(ring, + (adev->rmmio_remap.reg_offset + + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> + 2, + 0); + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h index 7b8a6152dc8d..4cfd932b7e91 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h @@ -44,4 +44,6 @@ struct amdgpu_hdp { }; int amdgpu_hdp_ras_sw_init(struct amdgpu_device *adev); +void amdgpu_hdp_generic_flush(struct amdgpu_device *adev, + struct amdgpu_ring *ring); #endif /* __AMDGPU_HDP_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 2ea98ec60220..802743efa3b3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -163,12 +163,12 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, init_shadow = false; } - if (!ring->sched.ready && !ring->is_mes_queue) { + if (!ring->sched.ready) { dev_err(adev->dev, "couldn't schedule ib on ring <%s>\n", ring->name); return -EINVAL; } - if (vm && !job->vmid && !ring->is_mes_queue) { + if (vm && !job->vmid) { dev_err(adev->dev, "VM IB without ID\n"); return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c index 4c4e087230ac..5dd78a9cb12d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c @@ -576,8 +576,16 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev) INIT_LIST_HEAD(&id_mgr->ids_lru); id_mgr->reserved_use_count = 0; - /* manage only VMIDs not used by KFD */ - id_mgr->num_ids = adev->vm_manager.first_kfd_vmid; + /* for GC <10, SDMA uses MMHUB so use first_kfd_vmid for both GC and MM */ + if (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 0, 0)) + /* manage only VMIDs not used by KFD */ + id_mgr->num_ids = adev->vm_manager.first_kfd_vmid; + else if (AMDGPU_IS_MMHUB0(i) || + AMDGPU_IS_MMHUB1(i)) + id_mgr->num_ids = 16; + else + /* manage only VMIDs not used by KFD */ + id_mgr->num_ids = adev->vm_manager.first_kfd_vmid; /* skip over VMID 0, since it is the system VM */ for (j = 1; j < id_mgr->num_ids; ++j) { @@ -588,7 +596,7 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev) } /* alloc a default reserved vmid to enforce isolation */ for (i = 0; i < (adev->xcp_mgr ? adev->xcp_mgr->num_xcps : 1); i++) { - if (adev->enforce_isolation[i]) + if (adev->enforce_isolation[i] != AMDGPU_ENFORCE_ISOLATION_DISABLE) amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(i)); } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c index 901f8b12c672..30f16968b578 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c @@ -25,6 +25,7 @@ #include "amdgpu.h" #include "amdgpu_ih.h" +#include "amdgpu_reset.h" /** * amdgpu_ih_ring_init - initialize the IH state @@ -227,13 +228,23 @@ restart_ih: ih->rptr &= ih->ptr_mask; } - amdgpu_ih_set_rptr(adev, ih); + if (!ih->overflow) + amdgpu_ih_set_rptr(adev, ih); + wake_up_all(&ih->wait_process); /* make sure wptr hasn't changed while processing */ wptr = amdgpu_ih_get_wptr(adev, ih); if (wptr != ih->rptr) - goto restart_ih; + if (!ih->overflow) + goto restart_ih; + + if (ih->overflow) + if (amdgpu_sriov_runtime(adev)) + WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, + &adev->virt.flr_work), + "Failed to queue work! at %s", + __func__); return IRQ_HANDLED; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h index b0a88f92cd82..7f7ea046e209 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h @@ -72,6 +72,7 @@ struct amdgpu_ih_ring { /* For waiting on IH processing at checkpoint. */ wait_queue_head_t wait_process; uint64_t processed_timestamp; + bool overflow; }; /* return true if time stamp t2 is after t1 with 48bit wrap around */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index 19ce4da285e8..5517451fc75d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -619,6 +619,10 @@ int amdgpu_irq_get(struct amdgpu_device *adev, struct amdgpu_irq_src *src, int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned int type) { + /* When the threshold is reached,the interrupt source may not be enabled.return -EINVAL */ + if (amdgpu_ras_is_rma(adev)) + return -EINVAL; + if (!adev->irq.installed) return -ENOENT; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h index 5a8bc6342222..529045d60412 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h @@ -78,6 +78,7 @@ struct amdgpu_job { /* enforce isolation */ bool enforce_isolation; + bool run_cleaner_shader; uint32_t num_ibs; struct amdgpu_ib ibs[]; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 27bfe9c8af06..9fbb04aee97b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -45,6 +45,7 @@ #include "amdgpu_ras.h" #include "amdgpu_reset.h" #include "amd_pcie.h" +#include "amdgpu_userq.h" void amdgpu_unregister_gpu_instance(struct amdgpu_device *adev) { @@ -370,6 +371,26 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info, return 0; } +static int amdgpu_userq_metadata_info_gfx(struct amdgpu_device *adev, + struct drm_amdgpu_info *info, + struct drm_amdgpu_info_uq_metadata_gfx *meta) +{ + int ret = -EOPNOTSUPP; + + if (adev->gfx.funcs->get_gfx_shadow_info) { + struct amdgpu_gfx_shadow_info shadow = {}; + + adev->gfx.funcs->get_gfx_shadow_info(adev, &shadow, true); + meta->shadow_size = shadow.shadow_size; + meta->shadow_alignment = shadow.shadow_alignment; + meta->csa_size = shadow.csa_size; + meta->csa_alignment = shadow.csa_alignment; + ret = 0; + } + + return ret; +} + static int amdgpu_hw_ip_info(struct amdgpu_device *adev, struct drm_amdgpu_info *info, struct drm_amdgpu_info_hw_ip *result) @@ -387,7 +408,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, case AMDGPU_HW_IP_GFX: type = AMD_IP_BLOCK_TYPE_GFX; for (i = 0; i < adev->gfx.num_gfx_rings; i++) - if (adev->gfx.gfx_ring[i].sched.ready) + if (adev->gfx.gfx_ring[i].sched.ready && + !adev->gfx.gfx_ring[i].no_user_submission) ++num_rings; ib_start_alignment = 32; ib_size_alignment = 32; @@ -395,7 +417,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, case AMDGPU_HW_IP_COMPUTE: type = AMD_IP_BLOCK_TYPE_GFX; for (i = 0; i < adev->gfx.num_compute_rings; i++) - if (adev->gfx.compute_ring[i].sched.ready) + if (adev->gfx.compute_ring[i].sched.ready && + !adev->gfx.compute_ring[i].no_user_submission) ++num_rings; ib_start_alignment = 32; ib_size_alignment = 32; @@ -403,7 +426,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, case AMDGPU_HW_IP_DMA: type = AMD_IP_BLOCK_TYPE_SDMA; for (i = 0; i < adev->sdma.num_instances; i++) - if (adev->sdma.instance[i].ring.sched.ready) + if (adev->sdma.instance[i].ring.sched.ready && + !adev->sdma.instance[i].ring.no_user_submission) ++num_rings; ib_start_alignment = 256; ib_size_alignment = 4; @@ -414,7 +438,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, if (adev->uvd.harvest_config & (1 << i)) continue; - if (adev->uvd.inst[i].ring.sched.ready) + if (adev->uvd.inst[i].ring.sched.ready && + !adev->uvd.inst[i].ring.no_user_submission) ++num_rings; } ib_start_alignment = 256; @@ -423,7 +448,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, case AMDGPU_HW_IP_VCE: type = AMD_IP_BLOCK_TYPE_VCE; for (i = 0; i < adev->vce.num_rings; i++) - if (adev->vce.ring[i].sched.ready) + if (adev->vce.ring[i].sched.ready && + !adev->vce.ring[i].no_user_submission) ++num_rings; ib_start_alignment = 256; ib_size_alignment = 4; @@ -435,7 +461,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, continue; for (j = 0; j < adev->uvd.num_enc_rings; j++) - if (adev->uvd.inst[i].ring_enc[j].sched.ready) + if (adev->uvd.inst[i].ring_enc[j].sched.ready && + !adev->uvd.inst[i].ring_enc[j].no_user_submission) ++num_rings; } ib_start_alignment = 256; @@ -447,7 +474,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, if (adev->vcn.harvest_config & (1 << i)) continue; - if (adev->vcn.inst[i].ring_dec.sched.ready) + if (adev->vcn.inst[i].ring_dec.sched.ready && + !adev->vcn.inst[i].ring_dec.no_user_submission) ++num_rings; } ib_start_alignment = 256; @@ -460,7 +488,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, continue; for (j = 0; j < adev->vcn.inst[i].num_enc_rings; j++) - if (adev->vcn.inst[i].ring_enc[j].sched.ready) + if (adev->vcn.inst[i].ring_enc[j].sched.ready && + !adev->vcn.inst[i].ring_enc[j].no_user_submission) ++num_rings; } ib_start_alignment = 256; @@ -475,7 +504,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, continue; for (j = 0; j < adev->jpeg.num_jpeg_rings; j++) - if (adev->jpeg.inst[i].ring_dec[j].sched.ready) + if (adev->jpeg.inst[i].ring_dec[j].sched.ready && + !adev->jpeg.inst[i].ring_dec[j].no_user_submission) ++num_rings; } ib_start_alignment = 256; @@ -483,7 +513,8 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, break; case AMDGPU_HW_IP_VPE: type = AMD_IP_BLOCK_TYPE_VPE; - if (adev->vpe.ring.sched.ready) + if (adev->vpe.ring.sched.ready && + !adev->vpe.ring.no_user_submission) ++num_rings; ib_start_alignment = 256; ib_size_alignment = 4; @@ -978,6 +1009,8 @@ out: } } + dev_info->userq_ip_mask = amdgpu_userq_get_supported_ip_mask(adev); + ret = copy_to_user(out, dev_info, min((size_t)size, sizeof(*dev_info))) ? -EFAULT : 0; kfree(dev_info); @@ -1293,6 +1326,22 @@ out: return copy_to_user(out, &gpuvm_fault, min((size_t)size, sizeof(gpuvm_fault))) ? -EFAULT : 0; } + case AMDGPU_INFO_UQ_FW_AREAS: { + struct drm_amdgpu_info_uq_metadata meta_info = {}; + + switch (info->query_hw_ip.type) { + case AMDGPU_HW_IP_GFX: + ret = amdgpu_userq_metadata_info_gfx(adev, info, &meta_info.gfx); + if (ret) + return ret; + + ret = copy_to_user(out, &meta_info, + min((size_t)size, sizeof(meta_info))) ? -EFAULT : 0; + return 0; + default: + return -EINVAL; + } + } default: DRM_DEBUG_KMS("Invalid request %d\n", info->query); return -EINVAL; @@ -1376,6 +1425,14 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) mutex_init(&fpriv->bo_list_lock); idr_init_base(&fpriv->bo_list_handles, 1); + r = amdgpu_userq_mgr_init(&fpriv->userq_mgr, file_priv, adev); + if (r) + DRM_WARN("Can't setup usermode queues, use legacy workload submission only\n"); + + r = amdgpu_eviction_fence_init(&fpriv->evf_mgr); + if (r) + goto error_vm; + amdgpu_ctx_mgr_init(&fpriv->ctx_mgr, adev); file_priv->driver_priv = fpriv; @@ -1445,6 +1502,11 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev, amdgpu_bo_unreserve(pd); } + if (!fpriv->evf_mgr.fd_closing) { + fpriv->evf_mgr.fd_closing = true; + amdgpu_userq_mgr_fini(&fpriv->userq_mgr); + amdgpu_eviction_fence_destroy(&fpriv->evf_mgr); + } amdgpu_ctx_mgr_fini(&fpriv->ctx_mgr); amdgpu_vm_fini(adev, &fpriv->vm); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index fb212f0a1136..2febb63ab232 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -39,42 +39,6 @@ int amdgpu_mes_doorbell_process_slice(struct amdgpu_device *adev) PAGE_SIZE); } -static int amdgpu_mes_kernel_doorbell_get(struct amdgpu_device *adev, - int ip_type, uint64_t *doorbell_index) -{ - unsigned int offset, found; - struct amdgpu_mes *mes = &adev->mes; - - if (ip_type == AMDGPU_RING_TYPE_SDMA) - offset = adev->doorbell_index.sdma_engine[0]; - else - offset = 0; - - found = find_next_zero_bit(mes->doorbell_bitmap, mes->num_mes_dbs, offset); - if (found >= mes->num_mes_dbs) { - DRM_WARN("No doorbell available\n"); - return -ENOSPC; - } - - set_bit(found, mes->doorbell_bitmap); - - /* Get the absolute doorbell index on BAR */ - *doorbell_index = mes->db_start_dw_offset + found * 2; - return 0; -} - -static void amdgpu_mes_kernel_doorbell_free(struct amdgpu_device *adev, - uint32_t doorbell_index) -{ - unsigned int old, rel_index; - struct amdgpu_mes *mes = &adev->mes; - - /* Find the relative index of the doorbell in this object */ - rel_index = (doorbell_index - mes->db_start_dw_offset) / 2; - old = test_and_clear_bit(rel_index, mes->doorbell_bitmap); - WARN_ON(!old); -} - static int amdgpu_mes_doorbell_init(struct amdgpu_device *adev) { int i; @@ -126,7 +90,7 @@ static void amdgpu_mes_doorbell_free(struct amdgpu_device *adev) int amdgpu_mes_init(struct amdgpu_device *adev) { - int i, r; + int i, r, num_pipes; adev->mes.adev = adev; @@ -142,19 +106,52 @@ int amdgpu_mes_init(struct amdgpu_device *adev) adev->mes.total_max_queue = AMDGPU_FENCE_MES_QUEUE_ID_MASK; adev->mes.vmid_mask_mmhub = 0xffffff00; - adev->mes.vmid_mask_gfxhub = 0xffffff00; + adev->mes.vmid_mask_gfxhub = adev->gfx.disable_kq ? 0xfffffffe : 0xffffff00; + + num_pipes = adev->gfx.me.num_pipe_per_me * adev->gfx.me.num_me; + if (num_pipes > AMDGPU_MES_MAX_GFX_PIPES) + dev_warn(adev->dev, "more gfx pipes than supported by MES! (%d vs %d)\n", + num_pipes, AMDGPU_MES_MAX_GFX_PIPES); + + for (i = 0; i < AMDGPU_MES_MAX_GFX_PIPES; i++) { + if (i >= num_pipes) + break; + if (amdgpu_ip_version(adev, GC_HWIP, 0) >= + IP_VERSION(12, 0, 0)) + /* + * GFX V12 has only one GFX pipe, but 8 queues in it. + * GFX pipe 0 queue 0 is being used by Kernel queue. + * Set GFX pipe 0 queue 1-7 for MES scheduling + * mask = 1111 1110b + */ + adev->mes.gfx_hqd_mask[i] = adev->gfx.disable_kq ? 0xFF : 0xFE; + else + /* + * GFX pipe 0 queue 0 is being used by Kernel queue. + * Set GFX pipe 0 queue 1 for MES scheduling + * mask = 10b + */ + adev->mes.gfx_hqd_mask[i] = adev->gfx.disable_kq ? 0x3 : 0x2; + } + + num_pipes = adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_mec; + if (num_pipes > AMDGPU_MES_MAX_COMPUTE_PIPES) + dev_warn(adev->dev, "more compute pipes than supported by MES! (%d vs %d)\n", + num_pipes, AMDGPU_MES_MAX_COMPUTE_PIPES); for (i = 0; i < AMDGPU_MES_MAX_COMPUTE_PIPES; i++) { - if (i >= (adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_mec)) + if (i >= num_pipes) break; - adev->mes.compute_hqd_mask[i] = 0xc; + adev->mes.compute_hqd_mask[i] = adev->gfx.disable_kq ? 0xF : 0xC; } - for (i = 0; i < AMDGPU_MES_MAX_GFX_PIPES; i++) - adev->mes.gfx_hqd_mask[i] = i ? 0 : 0xfffffffe; + num_pipes = adev->sdma.num_instances; + if (num_pipes > AMDGPU_MES_MAX_SDMA_PIPES) + dev_warn(adev->dev, "more SDMA pipes than supported by MES! (%d vs %d)\n", + num_pipes, AMDGPU_MES_MAX_SDMA_PIPES); for (i = 0; i < AMDGPU_MES_MAX_SDMA_PIPES; i++) { - if (i >= adev->sdma.num_instances) + if (i >= num_pipes) break; adev->mes.sdma_hqd_mask[i] = 0xfc; } @@ -240,244 +237,6 @@ void amdgpu_mes_fini(struct amdgpu_device *adev) mutex_destroy(&adev->mes.mutex_hidden); } -static void amdgpu_mes_queue_free_mqd(struct amdgpu_mes_queue *q) -{ - amdgpu_bo_free_kernel(&q->mqd_obj, - &q->mqd_gpu_addr, - &q->mqd_cpu_ptr); -} - -int amdgpu_mes_create_process(struct amdgpu_device *adev, int pasid, - struct amdgpu_vm *vm) -{ - struct amdgpu_mes_process *process; - int r; - - /* allocate the mes process buffer */ - process = kzalloc(sizeof(struct amdgpu_mes_process), GFP_KERNEL); - if (!process) { - DRM_ERROR("no more memory to create mes process\n"); - return -ENOMEM; - } - - /* allocate the process context bo and map it */ - r = amdgpu_bo_create_kernel(adev, AMDGPU_MES_PROC_CTX_SIZE, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, - &process->proc_ctx_bo, - &process->proc_ctx_gpu_addr, - &process->proc_ctx_cpu_ptr); - if (r) { - DRM_ERROR("failed to allocate process context bo\n"); - goto clean_up_memory; - } - memset(process->proc_ctx_cpu_ptr, 0, AMDGPU_MES_PROC_CTX_SIZE); - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - /* add the mes process to idr list */ - r = idr_alloc(&adev->mes.pasid_idr, process, pasid, pasid + 1, - GFP_KERNEL); - if (r < 0) { - DRM_ERROR("failed to lock pasid=%d\n", pasid); - goto clean_up_ctx; - } - - INIT_LIST_HEAD(&process->gang_list); - process->vm = vm; - process->pasid = pasid; - process->process_quantum = adev->mes.default_process_quantum; - process->pd_gpu_addr = amdgpu_bo_gpu_offset(vm->root.bo); - - amdgpu_mes_unlock(&adev->mes); - return 0; - -clean_up_ctx: - amdgpu_mes_unlock(&adev->mes); - amdgpu_bo_free_kernel(&process->proc_ctx_bo, - &process->proc_ctx_gpu_addr, - &process->proc_ctx_cpu_ptr); -clean_up_memory: - kfree(process); - return r; -} - -void amdgpu_mes_destroy_process(struct amdgpu_device *adev, int pasid) -{ - struct amdgpu_mes_process *process; - struct amdgpu_mes_gang *gang, *tmp1; - struct amdgpu_mes_queue *queue, *tmp2; - struct mes_remove_queue_input queue_input; - unsigned long flags; - int r; - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - process = idr_find(&adev->mes.pasid_idr, pasid); - if (!process) { - DRM_WARN("pasid %d doesn't exist\n", pasid); - amdgpu_mes_unlock(&adev->mes); - return; - } - - /* Remove all queues from hardware */ - list_for_each_entry_safe(gang, tmp1, &process->gang_list, list) { - list_for_each_entry_safe(queue, tmp2, &gang->queue_list, list) { - spin_lock_irqsave(&adev->mes.queue_id_lock, flags); - idr_remove(&adev->mes.queue_id_idr, queue->queue_id); - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - - queue_input.doorbell_offset = queue->doorbell_off; - queue_input.gang_context_addr = gang->gang_ctx_gpu_addr; - - r = adev->mes.funcs->remove_hw_queue(&adev->mes, - &queue_input); - if (r) - DRM_WARN("failed to remove hardware queue\n"); - } - - idr_remove(&adev->mes.gang_id_idr, gang->gang_id); - } - - idr_remove(&adev->mes.pasid_idr, pasid); - amdgpu_mes_unlock(&adev->mes); - - /* free all memory allocated by the process */ - list_for_each_entry_safe(gang, tmp1, &process->gang_list, list) { - /* free all queues in the gang */ - list_for_each_entry_safe(queue, tmp2, &gang->queue_list, list) { - amdgpu_mes_queue_free_mqd(queue); - list_del(&queue->list); - kfree(queue); - } - amdgpu_bo_free_kernel(&gang->gang_ctx_bo, - &gang->gang_ctx_gpu_addr, - &gang->gang_ctx_cpu_ptr); - list_del(&gang->list); - kfree(gang); - - } - amdgpu_bo_free_kernel(&process->proc_ctx_bo, - &process->proc_ctx_gpu_addr, - &process->proc_ctx_cpu_ptr); - kfree(process); -} - -int amdgpu_mes_add_gang(struct amdgpu_device *adev, int pasid, - struct amdgpu_mes_gang_properties *gprops, - int *gang_id) -{ - struct amdgpu_mes_process *process; - struct amdgpu_mes_gang *gang; - int r; - - /* allocate the mes gang buffer */ - gang = kzalloc(sizeof(struct amdgpu_mes_gang), GFP_KERNEL); - if (!gang) { - return -ENOMEM; - } - - /* allocate the gang context bo and map it to cpu space */ - r = amdgpu_bo_create_kernel(adev, AMDGPU_MES_GANG_CTX_SIZE, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, - &gang->gang_ctx_bo, - &gang->gang_ctx_gpu_addr, - &gang->gang_ctx_cpu_ptr); - if (r) { - DRM_ERROR("failed to allocate process context bo\n"); - goto clean_up_mem; - } - memset(gang->gang_ctx_cpu_ptr, 0, AMDGPU_MES_GANG_CTX_SIZE); - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - process = idr_find(&adev->mes.pasid_idr, pasid); - if (!process) { - DRM_ERROR("pasid %d doesn't exist\n", pasid); - r = -EINVAL; - goto clean_up_ctx; - } - - /* add the mes gang to idr list */ - r = idr_alloc(&adev->mes.gang_id_idr, gang, 1, 0, - GFP_KERNEL); - if (r < 0) { - DRM_ERROR("failed to allocate idr for gang\n"); - goto clean_up_ctx; - } - - gang->gang_id = r; - *gang_id = r; - - INIT_LIST_HEAD(&gang->queue_list); - gang->process = process; - gang->priority = gprops->priority; - gang->gang_quantum = gprops->gang_quantum ? - gprops->gang_quantum : adev->mes.default_gang_quantum; - gang->global_priority_level = gprops->global_priority_level; - gang->inprocess_gang_priority = gprops->inprocess_gang_priority; - list_add_tail(&gang->list, &process->gang_list); - - amdgpu_mes_unlock(&adev->mes); - return 0; - -clean_up_ctx: - amdgpu_mes_unlock(&adev->mes); - amdgpu_bo_free_kernel(&gang->gang_ctx_bo, - &gang->gang_ctx_gpu_addr, - &gang->gang_ctx_cpu_ptr); -clean_up_mem: - kfree(gang); - return r; -} - -int amdgpu_mes_remove_gang(struct amdgpu_device *adev, int gang_id) -{ - struct amdgpu_mes_gang *gang; - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - gang = idr_find(&adev->mes.gang_id_idr, gang_id); - if (!gang) { - DRM_ERROR("gang id %d doesn't exist\n", gang_id); - amdgpu_mes_unlock(&adev->mes); - return -EINVAL; - } - - if (!list_empty(&gang->queue_list)) { - DRM_ERROR("queue list is not empty\n"); - amdgpu_mes_unlock(&adev->mes); - return -EBUSY; - } - - idr_remove(&adev->mes.gang_id_idr, gang->gang_id); - list_del(&gang->list); - amdgpu_mes_unlock(&adev->mes); - - amdgpu_bo_free_kernel(&gang->gang_ctx_bo, - &gang->gang_ctx_gpu_addr, - &gang->gang_ctx_cpu_ptr); - - kfree(gang); - - return 0; -} - int amdgpu_mes_suspend(struct amdgpu_device *adev) { struct mes_suspend_gang_input input; @@ -526,304 +285,6 @@ int amdgpu_mes_resume(struct amdgpu_device *adev) return r; } -static int amdgpu_mes_queue_alloc_mqd(struct amdgpu_device *adev, - struct amdgpu_mes_queue *q, - struct amdgpu_mes_queue_properties *p) -{ - struct amdgpu_mqd *mqd_mgr = &adev->mqds[p->queue_type]; - u32 mqd_size = mqd_mgr->mqd_size; - int r; - - r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, - &q->mqd_obj, - &q->mqd_gpu_addr, &q->mqd_cpu_ptr); - if (r) { - dev_warn(adev->dev, "failed to create queue mqd bo (%d)", r); - return r; - } - memset(q->mqd_cpu_ptr, 0, mqd_size); - - r = amdgpu_bo_reserve(q->mqd_obj, false); - if (unlikely(r != 0)) - goto clean_up; - - return 0; - -clean_up: - amdgpu_bo_free_kernel(&q->mqd_obj, - &q->mqd_gpu_addr, - &q->mqd_cpu_ptr); - return r; -} - -static void amdgpu_mes_queue_init_mqd(struct amdgpu_device *adev, - struct amdgpu_mes_queue *q, - struct amdgpu_mes_queue_properties *p) -{ - struct amdgpu_mqd *mqd_mgr = &adev->mqds[p->queue_type]; - struct amdgpu_mqd_prop mqd_prop = {0}; - - mqd_prop.mqd_gpu_addr = q->mqd_gpu_addr; - mqd_prop.hqd_base_gpu_addr = p->hqd_base_gpu_addr; - mqd_prop.rptr_gpu_addr = p->rptr_gpu_addr; - mqd_prop.wptr_gpu_addr = p->wptr_gpu_addr; - mqd_prop.queue_size = p->queue_size; - mqd_prop.use_doorbell = true; - mqd_prop.doorbell_index = p->doorbell_off; - mqd_prop.eop_gpu_addr = p->eop_gpu_addr; - mqd_prop.hqd_pipe_priority = p->hqd_pipe_priority; - mqd_prop.hqd_queue_priority = p->hqd_queue_priority; - mqd_prop.hqd_active = false; - - if (p->queue_type == AMDGPU_RING_TYPE_GFX || - p->queue_type == AMDGPU_RING_TYPE_COMPUTE) { - mutex_lock(&adev->srbm_mutex); - amdgpu_gfx_select_me_pipe_q(adev, p->ring->me, p->ring->pipe, 0, 0, 0); - } - - mqd_mgr->init_mqd(adev, q->mqd_cpu_ptr, &mqd_prop); - - if (p->queue_type == AMDGPU_RING_TYPE_GFX || - p->queue_type == AMDGPU_RING_TYPE_COMPUTE) { - amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0, 0); - mutex_unlock(&adev->srbm_mutex); - } - - amdgpu_bo_unreserve(q->mqd_obj); -} - -int amdgpu_mes_add_hw_queue(struct amdgpu_device *adev, int gang_id, - struct amdgpu_mes_queue_properties *qprops, - int *queue_id) -{ - struct amdgpu_mes_queue *queue; - struct amdgpu_mes_gang *gang; - struct mes_add_queue_input queue_input; - unsigned long flags; - int r; - - memset(&queue_input, 0, sizeof(struct mes_add_queue_input)); - - /* allocate the mes queue buffer */ - queue = kzalloc(sizeof(struct amdgpu_mes_queue), GFP_KERNEL); - if (!queue) { - DRM_ERROR("Failed to allocate memory for queue\n"); - return -ENOMEM; - } - - /* Allocate the queue mqd */ - r = amdgpu_mes_queue_alloc_mqd(adev, queue, qprops); - if (r) - goto clean_up_memory; - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - gang = idr_find(&adev->mes.gang_id_idr, gang_id); - if (!gang) { - DRM_ERROR("gang id %d doesn't exist\n", gang_id); - r = -EINVAL; - goto clean_up_mqd; - } - - /* add the mes gang to idr list */ - spin_lock_irqsave(&adev->mes.queue_id_lock, flags); - r = idr_alloc(&adev->mes.queue_id_idr, queue, 1, 0, - GFP_ATOMIC); - if (r < 0) { - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - goto clean_up_mqd; - } - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - *queue_id = queue->queue_id = r; - - /* allocate a doorbell index for the queue */ - r = amdgpu_mes_kernel_doorbell_get(adev, - qprops->queue_type, - &qprops->doorbell_off); - if (r) - goto clean_up_queue_id; - - /* initialize the queue mqd */ - amdgpu_mes_queue_init_mqd(adev, queue, qprops); - - /* add hw queue to mes */ - queue_input.process_id = gang->process->pasid; - - queue_input.page_table_base_addr = - adev->vm_manager.vram_base_offset + gang->process->pd_gpu_addr - - adev->gmc.vram_start; - - queue_input.process_va_start = 0; - queue_input.process_va_end = - (adev->vm_manager.max_pfn - 1) << AMDGPU_GPU_PAGE_SHIFT; - queue_input.process_quantum = gang->process->process_quantum; - queue_input.process_context_addr = gang->process->proc_ctx_gpu_addr; - queue_input.gang_quantum = gang->gang_quantum; - queue_input.gang_context_addr = gang->gang_ctx_gpu_addr; - queue_input.inprocess_gang_priority = gang->inprocess_gang_priority; - queue_input.gang_global_priority_level = gang->global_priority_level; - queue_input.doorbell_offset = qprops->doorbell_off; - queue_input.mqd_addr = queue->mqd_gpu_addr; - queue_input.wptr_addr = qprops->wptr_gpu_addr; - queue_input.wptr_mc_addr = qprops->wptr_mc_addr; - queue_input.queue_type = qprops->queue_type; - queue_input.paging = qprops->paging; - queue_input.is_kfd_process = 0; - - r = adev->mes.funcs->add_hw_queue(&adev->mes, &queue_input); - if (r) { - DRM_ERROR("failed to add hardware queue to MES, doorbell=0x%llx\n", - qprops->doorbell_off); - goto clean_up_doorbell; - } - - DRM_DEBUG("MES hw queue was added, pasid=%d, gang id=%d, " - "queue type=%d, doorbell=0x%llx\n", - gang->process->pasid, gang_id, qprops->queue_type, - qprops->doorbell_off); - - queue->ring = qprops->ring; - queue->doorbell_off = qprops->doorbell_off; - queue->wptr_gpu_addr = qprops->wptr_gpu_addr; - queue->queue_type = qprops->queue_type; - queue->paging = qprops->paging; - queue->gang = gang; - queue->ring->mqd_ptr = queue->mqd_cpu_ptr; - list_add_tail(&queue->list, &gang->queue_list); - - amdgpu_mes_unlock(&adev->mes); - return 0; - -clean_up_doorbell: - amdgpu_mes_kernel_doorbell_free(adev, qprops->doorbell_off); -clean_up_queue_id: - spin_lock_irqsave(&adev->mes.queue_id_lock, flags); - idr_remove(&adev->mes.queue_id_idr, queue->queue_id); - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); -clean_up_mqd: - amdgpu_mes_unlock(&adev->mes); - amdgpu_mes_queue_free_mqd(queue); -clean_up_memory: - kfree(queue); - return r; -} - -int amdgpu_mes_remove_hw_queue(struct amdgpu_device *adev, int queue_id) -{ - unsigned long flags; - struct amdgpu_mes_queue *queue; - struct amdgpu_mes_gang *gang; - struct mes_remove_queue_input queue_input; - int r; - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - /* remove the mes gang from idr list */ - spin_lock_irqsave(&adev->mes.queue_id_lock, flags); - - queue = idr_find(&adev->mes.queue_id_idr, queue_id); - if (!queue) { - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - amdgpu_mes_unlock(&adev->mes); - DRM_ERROR("queue id %d doesn't exist\n", queue_id); - return -EINVAL; - } - - idr_remove(&adev->mes.queue_id_idr, queue_id); - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - - DRM_DEBUG("try to remove queue, doorbell off = 0x%llx\n", - queue->doorbell_off); - - gang = queue->gang; - queue_input.doorbell_offset = queue->doorbell_off; - queue_input.gang_context_addr = gang->gang_ctx_gpu_addr; - - r = adev->mes.funcs->remove_hw_queue(&adev->mes, &queue_input); - if (r) - DRM_ERROR("failed to remove hardware queue, queue id = %d\n", - queue_id); - - list_del(&queue->list); - amdgpu_mes_kernel_doorbell_free(adev, queue->doorbell_off); - amdgpu_mes_unlock(&adev->mes); - - amdgpu_mes_queue_free_mqd(queue); - kfree(queue); - return 0; -} - -int amdgpu_mes_reset_hw_queue(struct amdgpu_device *adev, int queue_id) -{ - unsigned long flags; - struct amdgpu_mes_queue *queue; - struct amdgpu_mes_gang *gang; - struct mes_reset_queue_input queue_input; - int r; - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - - /* remove the mes gang from idr list */ - spin_lock_irqsave(&adev->mes.queue_id_lock, flags); - - queue = idr_find(&adev->mes.queue_id_idr, queue_id); - if (!queue) { - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - amdgpu_mes_unlock(&adev->mes); - DRM_ERROR("queue id %d doesn't exist\n", queue_id); - return -EINVAL; - } - spin_unlock_irqrestore(&adev->mes.queue_id_lock, flags); - - DRM_DEBUG("try to reset queue, doorbell off = 0x%llx\n", - queue->doorbell_off); - - gang = queue->gang; - queue_input.doorbell_offset = queue->doorbell_off; - queue_input.gang_context_addr = gang->gang_ctx_gpu_addr; - - r = adev->mes.funcs->reset_hw_queue(&adev->mes, &queue_input); - if (r) - DRM_ERROR("failed to reset hardware queue, queue id = %d\n", - queue_id); - - amdgpu_mes_unlock(&adev->mes); - - return 0; -} - -int amdgpu_mes_reset_hw_queue_mmio(struct amdgpu_device *adev, int queue_type, - int me_id, int pipe_id, int queue_id, int vmid) -{ - struct mes_reset_queue_input queue_input; - int r; - - queue_input.queue_type = queue_type; - queue_input.use_mmio = true; - queue_input.me_id = me_id; - queue_input.pipe_id = pipe_id; - queue_input.queue_id = queue_id; - queue_input.vmid = vmid; - r = adev->mes.funcs->reset_hw_queue(&adev->mes, &queue_input); - if (r) - DRM_ERROR("failed to reset hardware queue by mmio, queue id = %d\n", - queue_id); - return r; -} - int amdgpu_mes_map_legacy_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring) { @@ -874,7 +335,7 @@ int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev, unsigned int vmid, bool use_mmio) { - struct mes_reset_legacy_queue_input queue_input; + struct mes_reset_queue_input queue_input; int r; memset(&queue_input, 0, sizeof(queue_input)); @@ -888,8 +349,11 @@ int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev, queue_input.wptr_addr = ring->wptr_gpu_addr; queue_input.vmid = vmid; queue_input.use_mmio = use_mmio; + queue_input.is_kq = true; + if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) + queue_input.legacy_gfx = true; - r = adev->mes.funcs->reset_legacy_queue(&adev->mes, &queue_input); + r = adev->mes.funcs->reset_hw_queue(&adev->mes, &queue_input); if (r) DRM_ERROR("failed to reset legacy queue\n"); @@ -905,7 +369,7 @@ uint32_t amdgpu_mes_rreg(struct amdgpu_device *adev, uint32_t reg) uint32_t *read_val_ptr; if (amdgpu_device_wb_get(adev, &addr_offset)) { - DRM_ERROR("critical bug! too many mes readers\n"); + dev_err(adev->dev, "critical bug! too many mes readers\n"); goto error; } read_val_gpu_addr = adev->wb.gpu_addr + (addr_offset * 4); @@ -915,13 +379,13 @@ uint32_t amdgpu_mes_rreg(struct amdgpu_device *adev, uint32_t reg) op_input.read_reg.buffer_addr = read_val_gpu_addr; if (!adev->mes.funcs->misc_op) { - DRM_ERROR("mes rreg is not supported!\n"); + dev_err(adev->dev, "mes rreg is not supported!\n"); goto error; } r = adev->mes.funcs->misc_op(&adev->mes, &op_input); if (r) - DRM_ERROR("failed to read reg (0x%x)\n", reg); + dev_err(adev->dev, "failed to read reg (0x%x)\n", reg); else val = *(read_val_ptr); @@ -942,14 +406,14 @@ int amdgpu_mes_wreg(struct amdgpu_device *adev, op_input.write_reg.reg_value = val; if (!adev->mes.funcs->misc_op) { - DRM_ERROR("mes wreg is not supported!\n"); + dev_err(adev->dev, "mes wreg is not supported!\n"); r = -EINVAL; goto error; } r = adev->mes.funcs->misc_op(&adev->mes, &op_input); if (r) - DRM_ERROR("failed to write reg (0x%x)\n", reg); + dev_err(adev->dev, "failed to write reg (0x%x)\n", reg); error: return r; @@ -969,14 +433,14 @@ int amdgpu_mes_reg_write_reg_wait(struct amdgpu_device *adev, op_input.wrm_reg.mask = mask; if (!adev->mes.funcs->misc_op) { - DRM_ERROR("mes reg_write_reg_wait is not supported!\n"); + dev_err(adev->dev, "mes reg_write_reg_wait is not supported!\n"); r = -EINVAL; goto error; } r = adev->mes.funcs->misc_op(&adev->mes, &op_input); if (r) - DRM_ERROR("failed to reg_write_reg_wait\n"); + dev_err(adev->dev, "failed to reg_write_reg_wait\n"); error: return r; @@ -994,14 +458,14 @@ int amdgpu_mes_reg_wait(struct amdgpu_device *adev, uint32_t reg, op_input.wrm_reg.mask = mask; if (!adev->mes.funcs->misc_op) { - DRM_ERROR("mes reg wait is not supported!\n"); + dev_err(adev->dev, "mes reg wait is not supported!\n"); r = -EINVAL; goto error; } r = adev->mes.funcs->misc_op(&adev->mes, &op_input); if (r) - DRM_ERROR("failed to reg_write_reg_wait\n"); + dev_err(adev->dev, "failed to reg_write_reg_wait\n"); error: return r; @@ -1075,25 +539,6 @@ int amdgpu_mes_flush_shader_debugger(struct amdgpu_device *adev, return r; } -static void -amdgpu_mes_ring_to_queue_props(struct amdgpu_device *adev, - struct amdgpu_ring *ring, - struct amdgpu_mes_queue_properties *props) -{ - props->queue_type = ring->funcs->type; - props->hqd_base_gpu_addr = ring->gpu_addr; - props->rptr_gpu_addr = ring->rptr_gpu_addr; - props->wptr_gpu_addr = ring->wptr_gpu_addr; - props->wptr_mc_addr = - ring->mes_ctx->meta_data_mc_addr + ring->wptr_offs; - props->queue_size = ring->ring_size; - props->eop_gpu_addr = ring->eop_gpu_addr; - props->hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_NORMAL; - props->hqd_queue_priority = AMDGPU_GFX_QUEUE_PRIORITY_MINIMUM; - props->paging = false; - props->ring = ring; -} - #define DEFINE_AMDGPU_MES_CTX_GET_OFFS_ENG(_eng) \ do { \ if (id_offs < AMDGPU_MES_CTX_MAX_OFFS) \ @@ -1130,453 +575,12 @@ int amdgpu_mes_ctx_get_offs(struct amdgpu_ring *ring, unsigned int id_offs) return -EINVAL; } -int amdgpu_mes_add_ring(struct amdgpu_device *adev, int gang_id, - int queue_type, int idx, - struct amdgpu_mes_ctx_data *ctx_data, - struct amdgpu_ring **out) -{ - struct amdgpu_ring *ring; - struct amdgpu_mes_gang *gang; - struct amdgpu_mes_queue_properties qprops = {0}; - int r, queue_id, pasid; - - /* - * Avoid taking any other locks under MES lock to avoid circular - * lock dependencies. - */ - amdgpu_mes_lock(&adev->mes); - gang = idr_find(&adev->mes.gang_id_idr, gang_id); - if (!gang) { - DRM_ERROR("gang id %d doesn't exist\n", gang_id); - amdgpu_mes_unlock(&adev->mes); - return -EINVAL; - } - pasid = gang->process->pasid; - - ring = kzalloc(sizeof(struct amdgpu_ring), GFP_KERNEL); - if (!ring) { - amdgpu_mes_unlock(&adev->mes); - return -ENOMEM; - } - - ring->ring_obj = NULL; - ring->use_doorbell = true; - ring->is_mes_queue = true; - ring->mes_ctx = ctx_data; - ring->idx = idx; - ring->no_scheduler = true; - - if (queue_type == AMDGPU_RING_TYPE_COMPUTE) { - int offset = offsetof(struct amdgpu_mes_ctx_meta_data, - compute[ring->idx].mec_hpd); - ring->eop_gpu_addr = - amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - } - - switch (queue_type) { - case AMDGPU_RING_TYPE_GFX: - ring->funcs = adev->gfx.gfx_ring[0].funcs; - ring->me = adev->gfx.gfx_ring[0].me; - ring->pipe = adev->gfx.gfx_ring[0].pipe; - break; - case AMDGPU_RING_TYPE_COMPUTE: - ring->funcs = adev->gfx.compute_ring[0].funcs; - ring->me = adev->gfx.compute_ring[0].me; - ring->pipe = adev->gfx.compute_ring[0].pipe; - break; - case AMDGPU_RING_TYPE_SDMA: - ring->funcs = adev->sdma.instance[0].ring.funcs; - break; - default: - BUG(); - } - - r = amdgpu_ring_init(adev, ring, 1024, NULL, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); - if (r) { - amdgpu_mes_unlock(&adev->mes); - goto clean_up_memory; - } - - amdgpu_mes_ring_to_queue_props(adev, ring, &qprops); - - dma_fence_wait(gang->process->vm->last_update, false); - dma_fence_wait(ctx_data->meta_data_va->last_pt_update, false); - amdgpu_mes_unlock(&adev->mes); - - r = amdgpu_mes_add_hw_queue(adev, gang_id, &qprops, &queue_id); - if (r) - goto clean_up_ring; - - ring->hw_queue_id = queue_id; - ring->doorbell_index = qprops.doorbell_off; - - if (queue_type == AMDGPU_RING_TYPE_GFX) - sprintf(ring->name, "gfx_%d.%d.%d", pasid, gang_id, queue_id); - else if (queue_type == AMDGPU_RING_TYPE_COMPUTE) - sprintf(ring->name, "compute_%d.%d.%d", pasid, gang_id, - queue_id); - else if (queue_type == AMDGPU_RING_TYPE_SDMA) - sprintf(ring->name, "sdma_%d.%d.%d", pasid, gang_id, - queue_id); - else - BUG(); - - *out = ring; - return 0; - -clean_up_ring: - amdgpu_ring_fini(ring); -clean_up_memory: - kfree(ring); - return r; -} - -void amdgpu_mes_remove_ring(struct amdgpu_device *adev, - struct amdgpu_ring *ring) -{ - if (!ring) - return; - - amdgpu_mes_remove_hw_queue(adev, ring->hw_queue_id); - timer_delete_sync(&ring->fence_drv.fallback_timer); - amdgpu_ring_fini(ring); - kfree(ring); -} - uint32_t amdgpu_mes_get_aggregated_doorbell_index(struct amdgpu_device *adev, enum amdgpu_mes_priority_level prio) { return adev->mes.aggregated_doorbells[prio]; } -int amdgpu_mes_ctx_alloc_meta_data(struct amdgpu_device *adev, - struct amdgpu_mes_ctx_data *ctx_data) -{ - int r; - - r = amdgpu_bo_create_kernel(adev, - sizeof(struct amdgpu_mes_ctx_meta_data), - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, - &ctx_data->meta_data_obj, - &ctx_data->meta_data_mc_addr, - &ctx_data->meta_data_ptr); - if (r) { - dev_warn(adev->dev, "(%d) create CTX bo failed\n", r); - return r; - } - - if (!ctx_data->meta_data_obj) - return -ENOMEM; - - memset(ctx_data->meta_data_ptr, 0, - sizeof(struct amdgpu_mes_ctx_meta_data)); - - return 0; -} - -void amdgpu_mes_ctx_free_meta_data(struct amdgpu_mes_ctx_data *ctx_data) -{ - if (ctx_data->meta_data_obj) - amdgpu_bo_free_kernel(&ctx_data->meta_data_obj, - &ctx_data->meta_data_mc_addr, - &ctx_data->meta_data_ptr); -} - -int amdgpu_mes_ctx_map_meta_data(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - struct amdgpu_mes_ctx_data *ctx_data) -{ - struct amdgpu_bo_va *bo_va; - struct amdgpu_sync sync; - struct drm_exec exec; - int r; - - amdgpu_sync_create(&sync); - - drm_exec_init(&exec, 0, 0); - drm_exec_until_all_locked(&exec) { - r = drm_exec_lock_obj(&exec, - &ctx_data->meta_data_obj->tbo.base); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto error_fini_exec; - - r = amdgpu_vm_lock_pd(vm, &exec, 0); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto error_fini_exec; - } - - bo_va = amdgpu_vm_bo_add(adev, vm, ctx_data->meta_data_obj); - if (!bo_va) { - DRM_ERROR("failed to create bo_va for meta data BO\n"); - r = -ENOMEM; - goto error_fini_exec; - } - - r = amdgpu_vm_bo_map(adev, bo_va, ctx_data->meta_data_gpu_addr, 0, - sizeof(struct amdgpu_mes_ctx_meta_data), - AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE | - AMDGPU_PTE_EXECUTABLE); - - if (r) { - DRM_ERROR("failed to do bo_map on meta data, err=%d\n", r); - goto error_del_bo_va; - } - - r = amdgpu_vm_bo_update(adev, bo_va, false); - if (r) { - DRM_ERROR("failed to do vm_bo_update on meta data\n"); - goto error_del_bo_va; - } - amdgpu_sync_fence(&sync, bo_va->last_pt_update, GFP_KERNEL); - - r = amdgpu_vm_update_pdes(adev, vm, false); - if (r) { - DRM_ERROR("failed to update pdes on meta data\n"); - goto error_del_bo_va; - } - amdgpu_sync_fence(&sync, vm->last_update, GFP_KERNEL); - - amdgpu_sync_wait(&sync, false); - drm_exec_fini(&exec); - - amdgpu_sync_free(&sync); - ctx_data->meta_data_va = bo_va; - return 0; - -error_del_bo_va: - amdgpu_vm_bo_del(adev, bo_va); - -error_fini_exec: - drm_exec_fini(&exec); - amdgpu_sync_free(&sync); - return r; -} - -int amdgpu_mes_ctx_unmap_meta_data(struct amdgpu_device *adev, - struct amdgpu_mes_ctx_data *ctx_data) -{ - struct amdgpu_bo_va *bo_va = ctx_data->meta_data_va; - struct amdgpu_bo *bo = ctx_data->meta_data_obj; - struct amdgpu_vm *vm = bo_va->base.vm; - struct dma_fence *fence; - struct drm_exec exec; - long r; - - drm_exec_init(&exec, 0, 0); - drm_exec_until_all_locked(&exec) { - r = drm_exec_lock_obj(&exec, - &ctx_data->meta_data_obj->tbo.base); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto out_unlock; - - r = amdgpu_vm_lock_pd(vm, &exec, 0); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto out_unlock; - } - - amdgpu_vm_bo_del(adev, bo_va); - if (!amdgpu_vm_ready(vm)) - goto out_unlock; - - r = dma_resv_get_singleton(bo->tbo.base.resv, DMA_RESV_USAGE_BOOKKEEP, - &fence); - if (r) - goto out_unlock; - if (fence) { - amdgpu_bo_fence(bo, fence, true); - fence = NULL; - } - - r = amdgpu_vm_clear_freed(adev, vm, &fence); - if (r || !fence) - goto out_unlock; - - dma_fence_wait(fence, false); - amdgpu_bo_fence(bo, fence, true); - dma_fence_put(fence); - -out_unlock: - if (unlikely(r < 0)) - dev_err(adev->dev, "failed to clear page tables (%ld)\n", r); - drm_exec_fini(&exec); - - return r; -} - -static int amdgpu_mes_test_create_gang_and_queues(struct amdgpu_device *adev, - int pasid, int *gang_id, - int queue_type, int num_queue, - struct amdgpu_ring **added_rings, - struct amdgpu_mes_ctx_data *ctx_data) -{ - struct amdgpu_ring *ring; - struct amdgpu_mes_gang_properties gprops = {0}; - int r, j; - - /* create a gang for the process */ - gprops.priority = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; - gprops.gang_quantum = adev->mes.default_gang_quantum; - gprops.inprocess_gang_priority = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; - gprops.priority_level = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; - gprops.global_priority_level = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; - - r = amdgpu_mes_add_gang(adev, pasid, &gprops, gang_id); - if (r) { - DRM_ERROR("failed to add gang\n"); - return r; - } - - /* create queues for the gang */ - for (j = 0; j < num_queue; j++) { - r = amdgpu_mes_add_ring(adev, *gang_id, queue_type, j, - ctx_data, &ring); - if (r) { - DRM_ERROR("failed to add ring\n"); - break; - } - - DRM_INFO("ring %s was added\n", ring->name); - added_rings[j] = ring; - } - - return 0; -} - -static int amdgpu_mes_test_queues(struct amdgpu_ring **added_rings) -{ - struct amdgpu_ring *ring; - int i, r; - - for (i = 0; i < AMDGPU_MES_CTX_MAX_RINGS; i++) { - ring = added_rings[i]; - if (!ring) - continue; - - r = amdgpu_ring_test_helper(ring); - if (r) - return r; - - r = amdgpu_ring_test_ib(ring, 1000 * 10); - if (r) { - DRM_DEV_ERROR(ring->adev->dev, - "ring %s ib test failed (%d)\n", - ring->name, r); - return r; - } else - DRM_INFO("ring %s ib test pass\n", ring->name); - } - - return 0; -} - -int amdgpu_mes_self_test(struct amdgpu_device *adev) -{ - struct amdgpu_vm *vm = NULL; - struct amdgpu_mes_ctx_data ctx_data = {0}; - struct amdgpu_ring *added_rings[AMDGPU_MES_CTX_MAX_RINGS] = { NULL }; - int gang_ids[3] = {0}; - int queue_types[][2] = { { AMDGPU_RING_TYPE_GFX, 1 }, - { AMDGPU_RING_TYPE_COMPUTE, 1 }, - { AMDGPU_RING_TYPE_SDMA, 1} }; - int i, r, pasid, k = 0; - - pasid = amdgpu_pasid_alloc(16); - if (pasid < 0) { - dev_warn(adev->dev, "No more PASIDs available!"); - pasid = 0; - } - - vm = kzalloc(sizeof(*vm), GFP_KERNEL); - if (!vm) { - r = -ENOMEM; - goto error_pasid; - } - - r = amdgpu_vm_init(adev, vm, -1); - if (r) { - DRM_ERROR("failed to initialize vm\n"); - goto error_pasid; - } - - r = amdgpu_mes_ctx_alloc_meta_data(adev, &ctx_data); - if (r) { - DRM_ERROR("failed to alloc ctx meta data\n"); - goto error_fini; - } - - ctx_data.meta_data_gpu_addr = AMDGPU_VA_RESERVED_BOTTOM; - r = amdgpu_mes_ctx_map_meta_data(adev, vm, &ctx_data); - if (r) { - DRM_ERROR("failed to map ctx meta data\n"); - goto error_vm; - } - - r = amdgpu_mes_create_process(adev, pasid, vm); - if (r) { - DRM_ERROR("failed to create MES process\n"); - goto error_vm; - } - - for (i = 0; i < ARRAY_SIZE(queue_types); i++) { - /* On GFX v10.3, fw hasn't supported to map sdma queue. */ - if (amdgpu_ip_version(adev, GC_HWIP, 0) >= - IP_VERSION(10, 3, 0) && - amdgpu_ip_version(adev, GC_HWIP, 0) < - IP_VERSION(11, 0, 0) && - queue_types[i][0] == AMDGPU_RING_TYPE_SDMA) - continue; - - r = amdgpu_mes_test_create_gang_and_queues(adev, pasid, - &gang_ids[i], - queue_types[i][0], - queue_types[i][1], - &added_rings[k], - &ctx_data); - if (r) - goto error_queues; - - k += queue_types[i][1]; - } - - /* start ring test and ib test for MES queues */ - amdgpu_mes_test_queues(added_rings); - -error_queues: - /* remove all queues */ - for (i = 0; i < ARRAY_SIZE(added_rings); i++) { - if (!added_rings[i]) - continue; - amdgpu_mes_remove_ring(adev, added_rings[i]); - } - - for (i = 0; i < ARRAY_SIZE(gang_ids); i++) { - if (!gang_ids[i]) - continue; - amdgpu_mes_remove_gang(adev, gang_ids[i]); - } - - amdgpu_mes_destroy_process(adev, pasid); - -error_vm: - amdgpu_mes_ctx_unmap_meta_data(adev, &ctx_data); - -error_fini: - amdgpu_vm_fini(adev, vm); - -error_pasid: - if (pasid) - amdgpu_pasid_free(pasid); - - amdgpu_mes_ctx_free_meta_data(&ctx_data); - kfree(vm); - return 0; -} - int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe) { const struct mes_firmware_header_v1_0 *mes_hdr; @@ -1705,7 +709,7 @@ int amdgpu_mes_update_enforce_isolation(struct amdgpu_device *adev) if (adev->enable_mes && adev->gfx.enable_cleaner_shader) { mutex_lock(&adev->enforce_isolation_mutex); for (i = 0; i < (adev->xcp_mgr ? adev->xcp_mgr->num_xcps : 1); i++) { - if (adev->enforce_isolation[i]) + if (adev->enforce_isolation[i] == AMDGPU_ENFORCE_ISOLATION_ENABLE) r |= amdgpu_mes_set_enforce_isolation(adev, i, true); else r |= amdgpu_mes_set_enforce_isolation(adev, i, false); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h index da2c9a8cb3e0..a41f65b4f733 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h @@ -111,8 +111,8 @@ struct amdgpu_mes { uint32_t vmid_mask_gfxhub; uint32_t vmid_mask_mmhub; - uint32_t compute_hqd_mask[AMDGPU_MES_MAX_COMPUTE_PIPES]; uint32_t gfx_hqd_mask[AMDGPU_MES_MAX_GFX_PIPES]; + uint32_t compute_hqd_mask[AMDGPU_MES_MAX_COMPUTE_PIPES]; uint32_t sdma_hqd_mask[AMDGPU_MES_MAX_SDMA_PIPES]; uint32_t aggregated_doorbells[AMDGPU_MES_PRIORITY_NUM_LEVELS]; uint32_t sch_ctx_offs[AMDGPU_MAX_MES_PIPES]; @@ -149,19 +149,6 @@ struct amdgpu_mes { }; -struct amdgpu_mes_process { - int pasid; - struct amdgpu_vm *vm; - uint64_t pd_gpu_addr; - struct amdgpu_bo *proc_ctx_bo; - uint64_t proc_ctx_gpu_addr; - void *proc_ctx_cpu_ptr; - uint64_t process_quantum; - struct list_head gang_list; - uint32_t doorbell_index; - struct mutex doorbell_lock; -}; - struct amdgpu_mes_gang { int gang_id; int priority; @@ -248,18 +235,6 @@ struct mes_remove_queue_input { uint64_t gang_context_addr; }; -struct mes_reset_queue_input { - uint32_t doorbell_offset; - uint64_t gang_context_addr; - bool use_mmio; - uint32_t queue_type; - uint32_t me_id; - uint32_t pipe_id; - uint32_t queue_id; - uint32_t xcc_id; - uint32_t vmid; -}; - struct mes_map_legacy_queue_input { uint32_t queue_type; uint32_t doorbell_offset; @@ -291,7 +266,7 @@ struct mes_resume_gang_input { uint64_t gang_context_addr; }; -struct mes_reset_legacy_queue_input { +struct mes_reset_queue_input { uint32_t queue_type; uint32_t doorbell_offset; bool use_mmio; @@ -301,6 +276,8 @@ struct mes_reset_legacy_queue_input { uint64_t mqd_addr; uint64_t wptr_addr; uint32_t vmid; + bool legacy_gfx; + bool is_kq; }; enum mes_misc_opcode { @@ -388,9 +365,6 @@ struct amdgpu_mes_funcs { int (*misc_op)(struct amdgpu_mes *mes, struct mes_misc_op_input *input); - int (*reset_legacy_queue)(struct amdgpu_mes *mes, - struct mes_reset_legacy_queue_input *input); - int (*reset_hw_queue)(struct amdgpu_mes *mes, struct mes_reset_queue_input *input); }; @@ -404,26 +378,9 @@ int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe); int amdgpu_mes_init(struct amdgpu_device *adev); void amdgpu_mes_fini(struct amdgpu_device *adev); -int amdgpu_mes_create_process(struct amdgpu_device *adev, int pasid, - struct amdgpu_vm *vm); -void amdgpu_mes_destroy_process(struct amdgpu_device *adev, int pasid); - -int amdgpu_mes_add_gang(struct amdgpu_device *adev, int pasid, - struct amdgpu_mes_gang_properties *gprops, - int *gang_id); -int amdgpu_mes_remove_gang(struct amdgpu_device *adev, int gang_id); - int amdgpu_mes_suspend(struct amdgpu_device *adev); int amdgpu_mes_resume(struct amdgpu_device *adev); -int amdgpu_mes_add_hw_queue(struct amdgpu_device *adev, int gang_id, - struct amdgpu_mes_queue_properties *qprops, - int *queue_id); -int amdgpu_mes_remove_hw_queue(struct amdgpu_device *adev, int queue_id); -int amdgpu_mes_reset_hw_queue(struct amdgpu_device *adev, int queue_id); -int amdgpu_mes_reset_hw_queue_mmio(struct amdgpu_device *adev, int queue_type, - int me_id, int pipe_id, int queue_id, int vmid); - int amdgpu_mes_map_legacy_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring); int amdgpu_mes_unmap_legacy_queue(struct amdgpu_device *adev, @@ -451,27 +408,10 @@ int amdgpu_mes_set_shader_debugger(struct amdgpu_device *adev, bool trap_en); int amdgpu_mes_flush_shader_debugger(struct amdgpu_device *adev, uint64_t process_context_addr); -int amdgpu_mes_add_ring(struct amdgpu_device *adev, int gang_id, - int queue_type, int idx, - struct amdgpu_mes_ctx_data *ctx_data, - struct amdgpu_ring **out); -void amdgpu_mes_remove_ring(struct amdgpu_device *adev, - struct amdgpu_ring *ring); uint32_t amdgpu_mes_get_aggregated_doorbell_index(struct amdgpu_device *adev, enum amdgpu_mes_priority_level prio); -int amdgpu_mes_ctx_alloc_meta_data(struct amdgpu_device *adev, - struct amdgpu_mes_ctx_data *ctx_data); -void amdgpu_mes_ctx_free_meta_data(struct amdgpu_mes_ctx_data *ctx_data); -int amdgpu_mes_ctx_map_meta_data(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - struct amdgpu_mes_ctx_data *ctx_data); -int amdgpu_mes_ctx_unmap_meta_data(struct amdgpu_device *adev, - struct amdgpu_mes_ctx_data *ctx_data); - -int amdgpu_mes_self_test(struct amdgpu_device *adev); - int amdgpu_mes_doorbell_process_slice(struct amdgpu_device *adev); /* diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 0b9987781f76..73403744331a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -1044,7 +1044,8 @@ static const char * const amdgpu_vram_names[] = { "GDDR6", "DDR5", "LPDDR4", - "LPDDR5" + "LPDDR5", + "HBM3E" }; /** @@ -1644,7 +1645,11 @@ u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m) amdgpu_bo_print_flag(m, bo, VRAM_CONTIGUOUS); amdgpu_bo_print_flag(m, bo, VM_ALWAYS_VALID); amdgpu_bo_print_flag(m, bo, EXPLICIT_SYNC); - + /* Add the gem obj resv fence dump*/ + if (dma_resv_trylock(bo->tbo.base.resv)) { + dma_resv_describe(bo->tbo.base.resv, m); + dma_resv_unlock(bo->tbo.base.resv); + } seq_puts(m, "\n"); return size; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index df5d5dbd7f0f..e6f0b035e20b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -2214,7 +2214,8 @@ static int psp_securedisplay_initialize(struct psp_context *psp) if (!psp->securedisplay_context.context.bin_desc.size_bytes || !psp->securedisplay_context.context.bin_desc.start_addr) { - dev_info(psp->adev->dev, "SECUREDISPLAY: securedisplay ta ucode is not available\n"); + dev_info(psp->adev->dev, + "SECUREDISPLAY: optional securedisplay ta ucode is not available\n"); return 0; } @@ -4185,6 +4186,110 @@ const struct attribute_group amdgpu_flash_attr_group = { .is_visible = amdgpu_flash_attr_is_visible, }; +#if defined(CONFIG_DEBUG_FS) +static int psp_read_spirom_debugfs_open(struct inode *inode, struct file *filp) +{ + struct amdgpu_device *adev = filp->f_inode->i_private; + struct spirom_bo *bo_triplet; + int ret; + + /* serialize the open() file calling */ + if (!mutex_trylock(&adev->psp.mutex)) + return -EBUSY; + + /* + * make sure only one userpace process is alive for dumping so that + * only one memory buffer of AMD_VBIOS_FILE_MAX_SIZE * 2 is consumed. + * let's say the case where one process try opening the file while + * another one has proceeded to read or release. In this way, eliminate + * the use of mutex for read() or release() callback as well. + */ + if (adev->psp.spirom_dump_trip) { + mutex_unlock(&adev->psp.mutex); + return -EBUSY; + } + + bo_triplet = kzalloc(sizeof(struct spirom_bo), GFP_KERNEL); + if (!bo_triplet) { + mutex_unlock(&adev->psp.mutex); + return -ENOMEM; + } + + ret = amdgpu_bo_create_kernel(adev, AMD_VBIOS_FILE_MAX_SIZE_B * 2, + AMDGPU_GPU_PAGE_SIZE, + AMDGPU_GEM_DOMAIN_GTT, + &bo_triplet->bo, + &bo_triplet->mc_addr, + &bo_triplet->cpu_addr); + if (ret) + goto rel_trip; + + ret = psp_dump_spirom(&adev->psp, bo_triplet->mc_addr); + if (ret) + goto rel_bo; + + adev->psp.spirom_dump_trip = bo_triplet; + mutex_unlock(&adev->psp.mutex); + return 0; +rel_bo: + amdgpu_bo_free_kernel(&bo_triplet->bo, &bo_triplet->mc_addr, + &bo_triplet->cpu_addr); +rel_trip: + kfree(bo_triplet); + mutex_unlock(&adev->psp.mutex); + dev_err(adev->dev, "Trying IFWI dump fails, err = %d\n", ret); + return ret; +} + +static ssize_t psp_read_spirom_debugfs_read(struct file *filp, char __user *buf, size_t size, + loff_t *pos) +{ + struct amdgpu_device *adev = filp->f_inode->i_private; + struct spirom_bo *bo_triplet = adev->psp.spirom_dump_trip; + + if (!bo_triplet) + return -EINVAL; + + return simple_read_from_buffer(buf, + size, + pos, bo_triplet->cpu_addr, + AMD_VBIOS_FILE_MAX_SIZE_B * 2); +} + +static int psp_read_spirom_debugfs_release(struct inode *inode, struct file *filp) +{ + struct amdgpu_device *adev = filp->f_inode->i_private; + struct spirom_bo *bo_triplet = adev->psp.spirom_dump_trip; + + if (bo_triplet) { + amdgpu_bo_free_kernel(&bo_triplet->bo, &bo_triplet->mc_addr, + &bo_triplet->cpu_addr); + kfree(bo_triplet); + } + + adev->psp.spirom_dump_trip = NULL; + return 0; +} + +static const struct file_operations psp_dump_spirom_debugfs_ops = { + .owner = THIS_MODULE, + .open = psp_read_spirom_debugfs_open, + .read = psp_read_spirom_debugfs_read, + .release = psp_read_spirom_debugfs_release, + .llseek = default_llseek, +}; +#endif + +void amdgpu_psp_debugfs_init(struct amdgpu_device *adev) +{ +#if defined(CONFIG_DEBUG_FS) + struct drm_minor *minor = adev_to_drm(adev)->primary; + + debugfs_create_file_size("psp_spirom_dump", 0444, minor->debugfs_root, + adev, &psp_dump_spirom_debugfs_ops, AMD_VBIOS_FILE_MAX_SIZE_B * 2); +#endif +} + const struct amd_ip_funcs psp_ip_funcs = { .name = "psp", .early_init = psp_early_init, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index 8d5acc415d38..428adc7f741d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -39,6 +39,18 @@ #define PSP_TMR_ALIGNMENT 0x100000 #define PSP_FW_NAME_LEN 0x24 +/* VBIOS gfl defines */ +#define MBOX_READY_MASK 0x80000000 +#define MBOX_STATUS_MASK 0x0000FFFF +#define MBOX_COMMAND_MASK 0x00FF0000 +#define MBOX_READY_FLAG 0x80000000 +#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_LO 0x2 +#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_HI 0x3 +#define C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE 0x4 +#define C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_LO 0xf +#define C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_HI 0x10 +#define C2PMSG_CMD_SPI_GET_FLASH_IMAGE 0x11 + extern const struct attribute_group amdgpu_flash_attr_group; enum psp_shared_mem_size { @@ -107,6 +119,7 @@ enum psp_reg_prog_id { PSP_REG_IH_RB_CNTL = 0, /* register IH_RB_CNTL */ PSP_REG_IH_RB_CNTL_RING1 = 1, /* register IH_RB_CNTL_RING1 */ PSP_REG_IH_RB_CNTL_RING2 = 2, /* register IH_RB_CNTL_RING2 */ + PSP_REG_MMHUB_L1_TLB_CNTL = 25, PSP_REG_LAST }; @@ -137,11 +150,14 @@ struct psp_funcs { int (*load_usbc_pd_fw)(struct psp_context *psp, uint64_t fw_pri_mc_addr); int (*read_usbc_pd_fw)(struct psp_context *psp, uint32_t *fw_ver); int (*update_spirom)(struct psp_context *psp, uint64_t fw_pri_mc_addr); + int (*dump_spirom)(struct psp_context *psp, uint64_t fw_pri_mc_addr); int (*vbflash_stat)(struct psp_context *psp); int (*fatal_error_recovery_quirk)(struct psp_context *psp); bool (*get_ras_capability)(struct psp_context *psp); bool (*is_aux_sos_load_required)(struct psp_context *psp); bool (*is_reload_needed)(struct psp_context *psp); + int (*reg_program_no_ring)(struct psp_context *psp, uint32_t val, + enum psp_reg_prog_id id); }; struct ta_funcs { @@ -319,6 +335,14 @@ struct psp_runtime_scpm_entry { enum psp_runtime_scpm_authentication scpm_status; }; +#if defined(CONFIG_DEBUG_FS) +struct spirom_bo { + struct amdgpu_bo *bo; + uint64_t mc_addr; + void *cpu_addr; +}; +#endif + struct psp_context { struct amdgpu_device *adev; struct psp_ring km_ring; @@ -406,6 +430,9 @@ struct psp_context { char *vbflash_tmp_buf; size_t vbflash_image_size; bool vbflash_done; +#if defined(CONFIG_DEBUG_FS) + struct spirom_bo *spirom_dump_trip; +#endif }; struct amdgpu_psp_funcs { @@ -464,6 +491,10 @@ struct amdgpu_psp_funcs { ((psp)->funcs->update_spirom ? \ (psp)->funcs->update_spirom((psp), fw_pri_mc_addr) : -EINVAL) +#define psp_dump_spirom(psp, fw_pri_mc_addr) \ + ((psp)->funcs->dump_spirom ? \ + (psp)->funcs->dump_spirom((psp), fw_pri_mc_addr) : -EINVAL) + #define psp_vbflash_status(psp) \ ((psp)->funcs->vbflash_stat ? \ (psp)->funcs->vbflash_stat((psp)) : -EINVAL) @@ -475,6 +506,10 @@ struct amdgpu_psp_funcs { #define psp_is_aux_sos_load_required(psp) \ ((psp)->funcs->is_aux_sos_load_required ? (psp)->funcs->is_aux_sos_load_required((psp)) : 0) +#define psp_reg_program_no_ring(psp, val, id) \ + ((psp)->funcs->reg_program_no_ring ? \ + (psp)->funcs->reg_program_no_ring((psp), val, id) : -EINVAL) + extern const struct amd_ip_funcs psp_ip_funcs; extern const struct amdgpu_ip_block_version psp_v3_1_ip_block; @@ -569,5 +604,9 @@ bool amdgpu_psp_get_ras_capability(struct psp_context *psp); int psp_config_sq_perfmon(struct psp_context *psp, uint32_t xcp_id, bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable); bool amdgpu_psp_tos_reload_needed(struct amdgpu_device *adev); +int amdgpu_psp_reg_program_no_ring(struct psp_context *psp, uint32_t val, + enum psp_reg_prog_id id); +void amdgpu_psp_debugfs_init(struct amdgpu_device *adev); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 443409d4f4b0..dc07936d2fcb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1498,6 +1498,9 @@ int amdgpu_ras_reset_error_count(struct amdgpu_device *adev, !amdgpu_ras_get_aca_debug_mode(adev)) return -EOPNOTSUPP; + if (amdgpu_sriov_vf(adev)) + return -EOPNOTSUPP; + /* skip ras error reset in gpu reset */ if ((amdgpu_in_reset(adev) || amdgpu_ras_in_recovery(adev)) && ((smu_funcs && smu_funcs->set_debug_mode) || @@ -2161,7 +2164,7 @@ void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev) /* Fatal error events are handled on host side */ if (amdgpu_sriov_vf(adev)) return; - /** + /* * If the current interrupt is caused by a non-fatal RAS error, skip * check for fatal error. For fatal errors, FED status of all devices * in XGMI hive gets set when the first device gets fatal error @@ -2886,6 +2889,7 @@ static int __amdgpu_ras_convert_rec_from_rom(struct amdgpu_device *adev, if (amdgpu_ras_mca2pa_by_idx(adev, bps, err_data)) return -EINVAL; } + return __amdgpu_ras_restore_bad_pages(adev, err_data->err_addr, adev->umc.retire_unit); } @@ -2900,7 +2904,7 @@ int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev, &adev->psp.ras_context.ras->eeprom_control; enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE; int ret = 0; - uint32_t i; + uint32_t i = 0; if (!con || !con->eh_data || !bps || pages <= 0) return 0; @@ -2921,34 +2925,36 @@ int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev, mutex_lock(&con->recovery_lock); if (from_rom) { - for (i = 0; i < pages; i++) { - if (control->ras_num_recs - i >= adev->umc.retire_unit) { - if ((bps[i].address == bps[i + 1].address) && - (bps[i].mem_channel == bps[i + 1].mem_channel)) { - //deal with retire_unit records a time - ret = __amdgpu_ras_convert_rec_array_from_rom(adev, - &bps[i], &err_data, nps); - if (ret) - goto free; - i += (adev->umc.retire_unit - 1); + /* there is no pa recs in V3, so skip pa recs processing */ + if (control->tbl_hdr.version < RAS_TABLE_VER_V3) { + for (i = 0; i < pages; i++) { + if (control->ras_num_recs - i >= adev->umc.retire_unit) { + if ((bps[i].address == bps[i + 1].address) && + (bps[i].mem_channel == bps[i + 1].mem_channel)) { + /* deal with retire_unit records a time */ + ret = __amdgpu_ras_convert_rec_array_from_rom(adev, + &bps[i], &err_data, nps); + if (ret) + control->ras_num_bad_pages -= adev->umc.retire_unit; + i += (adev->umc.retire_unit - 1); + } else { + break; + } } else { break; } - } else { - break; } } for (; i < pages; i++) { ret = __amdgpu_ras_convert_rec_from_rom(adev, &bps[i], &err_data, nps); if (ret) - goto free; + control->ras_num_bad_pages -= adev->umc.retire_unit; } } else { ret = __amdgpu_ras_restore_bad_pages(adev, bps, pages); } -free: if (from_rom) kfree(err_data.err_addr); mutex_unlock(&con->recovery_lock); @@ -3037,21 +3043,28 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev) dev_err(adev->dev, "Failed to load EEPROM table records!"); } else { if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr) { - for (i = 0; i < control->ras_num_recs; i++) { - if ((control->ras_num_recs - i) >= adev->umc.retire_unit) { - if ((bps[i].address == bps[i + 1].address) && - (bps[i].mem_channel == bps[i + 1].mem_channel)) { - control->ras_num_pa_recs += adev->umc.retire_unit; - i += (adev->umc.retire_unit - 1); + /*In V3, there is no pa recs, and some cases(when address==0) may be parsed + as pa recs, so add verion check to avoid it. + */ + if (control->tbl_hdr.version < RAS_TABLE_VER_V3) { + for (i = 0; i < control->ras_num_recs; i++) { + if ((control->ras_num_recs - i) >= adev->umc.retire_unit) { + if ((bps[i].address == bps[i + 1].address) && + (bps[i].mem_channel == bps[i + 1].mem_channel)) { + control->ras_num_pa_recs += adev->umc.retire_unit; + i += (adev->umc.retire_unit - 1); + } else { + control->ras_num_mca_recs += + (control->ras_num_recs - i); + break; + } } else { - control->ras_num_mca_recs += - (control->ras_num_recs - i); + control->ras_num_mca_recs += (control->ras_num_recs - i); break; } - } else { - control->ras_num_mca_recs += (control->ras_num_recs - i); - break; } + } else { + control->ras_num_mca_recs = control->ras_num_recs; } } @@ -3460,6 +3473,10 @@ int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev) if (!adev->umc.ras || !adev->umc.ras->convert_ras_err_addr) control->ras_num_pa_recs = control->ras_num_recs; + if (adev->umc.ras && + adev->umc.ras->get_retire_flip_bits) + adev->umc.ras->get_retire_flip_bits(adev); + if (control->ras_num_recs) { ret = amdgpu_ras_load_bad_pages(adev); if (ret) @@ -3793,10 +3810,12 @@ init_ras_enabled_flag: adev->ras_hw_enabled & amdgpu_ras_mask; /* aca is disabled by default except for psp v13_0_6/v13_0_12/v13_0_14 */ - adev->aca.is_enabled = - (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) || - amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) || - amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)); + if (!amdgpu_sriov_vf(adev)) { + adev->aca.is_enabled = + (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) || + amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) || + amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)); + } /* bad page feature is not applicable to specific app platform */ if (adev->gmc.is_app_apu && @@ -4479,8 +4498,11 @@ void amdgpu_ras_global_ras_isr(struct amdgpu_device *adev) enum ras_event_type type = RAS_EVENT_TYPE_FATAL; u64 event_id; - if (amdgpu_ras_mark_ras_event(adev, type)) + if (amdgpu_ras_mark_ras_event(adev, type)) { + dev_err(adev->dev, + "uncorrectable hardware error (ERREVENT_ATHUB_INTERRUPT) detected!\n"); return; + } event_id = amdgpu_ras_acquire_event_id(adev, type); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index 0ea7cfaf3587..2c58e09e56f9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -418,6 +418,7 @@ static void amdgpu_ras_set_eeprom_table_version(struct amdgpu_ras_eeprom_control hdr->version = RAS_TABLE_VER_V2_1; return; case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 5, 0): hdr->version = RAS_TABLE_VER_V3; return; default: @@ -1392,17 +1393,39 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control) __decode_table_header_from_buf(hdr, buf); - if (hdr->version >= RAS_TABLE_VER_V2_1) { + if (hdr->header != RAS_TABLE_HDR_VAL && + hdr->header != RAS_TABLE_HDR_BAD) { + dev_info(adev->dev, "Creating a new EEPROM table"); + return amdgpu_ras_eeprom_reset_table(control); + } + + switch (hdr->version) { + case RAS_TABLE_VER_V2_1: + case RAS_TABLE_VER_V3: control->ras_num_recs = RAS_NUM_RECS_V2_1(hdr); control->ras_record_offset = RAS_RECORD_START_V2_1; control->ras_max_record_count = RAS_MAX_RECORD_COUNT_V2_1; - } else { + break; + case RAS_TABLE_VER_V1: control->ras_num_recs = RAS_NUM_RECS(hdr); control->ras_record_offset = RAS_RECORD_START; control->ras_max_record_count = RAS_MAX_RECORD_COUNT; + break; + default: + dev_err(adev->dev, + "RAS header invalid, unsupported version: %u", + hdr->version); + return -EINVAL; } - control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset); + if (control->ras_num_recs > control->ras_max_record_count) { + dev_err(adev->dev, + "RAS header invalid, records in header: %u max allowed :%u", + control->ras_num_recs, control->ras_max_record_count); + return -EINVAL; + } + + control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset); control->ras_num_mca_recs = 0; control->ras_num_pa_recs = 0; return 0; @@ -1413,7 +1436,7 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control) struct amdgpu_device *adev = to_amdgpu_device(control); struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr; struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); - int res; + int res = 0; if (!__is_ras_eeprom_supported(adev)) return 0; @@ -1494,10 +1517,6 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control) "User defined threshold is set, runtime service will be halt when threshold is reached\n"); } } - } else { - DRM_INFO("Creating a new EEPROM table"); - - res = amdgpu_ras_eeprom_reset_table(control); } return res < 0 ? res : 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 59acdbfe28d8..426834806fbf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -187,14 +187,10 @@ void amdgpu_ring_undo(struct amdgpu_ring *ring) } #define amdgpu_ring_get_gpu_addr(ring, offset) \ - (ring->is_mes_queue ? \ - (ring->mes_ctx->meta_data_gpu_addr + offset) : \ - (ring->adev->wb.gpu_addr + offset * 4)) + (ring->adev->wb.gpu_addr + offset * 4) #define amdgpu_ring_get_cpu_addr(ring, offset) \ - (ring->is_mes_queue ? \ - (void *)((uint8_t *)(ring->mes_ctx->meta_data_ptr) + offset) : \ - (&ring->adev->wb.wb[offset])) + (&ring->adev->wb.wb[offset]) /** * amdgpu_ring_init - init driver ring struct. @@ -243,57 +239,42 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, ring->sched_score = sched_score; ring->vmid_wait = dma_fence_get_stub(); - if (!ring->is_mes_queue) { - ring->idx = adev->num_rings++; - adev->rings[ring->idx] = ring; - } + ring->idx = adev->num_rings++; + adev->rings[ring->idx] = ring; r = amdgpu_fence_driver_init_ring(ring); if (r) return r; } - if (ring->is_mes_queue) { - ring->rptr_offs = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_RPTR_OFFS); - ring->wptr_offs = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_WPTR_OFFS); - ring->fence_offs = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_FENCE_OFFS); - ring->trail_fence_offs = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_TRAIL_FENCE_OFFS); - ring->cond_exe_offs = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_COND_EXE_OFFS); - } else { - r = amdgpu_device_wb_get(adev, &ring->rptr_offs); - if (r) { - dev_err(adev->dev, "(%d) ring rptr_offs wb alloc failed\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &ring->rptr_offs); + if (r) { + dev_err(adev->dev, "(%d) ring rptr_offs wb alloc failed\n", r); + return r; + } - r = amdgpu_device_wb_get(adev, &ring->wptr_offs); - if (r) { - dev_err(adev->dev, "(%d) ring wptr_offs wb alloc failed\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &ring->wptr_offs); + if (r) { + dev_err(adev->dev, "(%d) ring wptr_offs wb alloc failed\n", r); + return r; + } - r = amdgpu_device_wb_get(adev, &ring->fence_offs); - if (r) { - dev_err(adev->dev, "(%d) ring fence_offs wb alloc failed\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &ring->fence_offs); + if (r) { + dev_err(adev->dev, "(%d) ring fence_offs wb alloc failed\n", r); + return r; + } - r = amdgpu_device_wb_get(adev, &ring->trail_fence_offs); - if (r) { - dev_err(adev->dev, "(%d) ring trail_fence_offs wb alloc failed\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &ring->trail_fence_offs); + if (r) { + dev_err(adev->dev, "(%d) ring trail_fence_offs wb alloc failed\n", r); + return r; + } - r = amdgpu_device_wb_get(adev, &ring->cond_exe_offs); - if (r) { - dev_err(adev->dev, "(%d) ring cond_exec_polling wb alloc failed\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &ring->cond_exe_offs); + if (r) { + dev_err(adev->dev, "(%d) ring cond_exec_polling wb alloc failed\n", r); + return r; } ring->fence_gpu_addr = @@ -353,18 +334,7 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, ring->cached_rptr = 0; /* Allocate ring buffer */ - if (ring->is_mes_queue) { - int offset = 0; - - BUG_ON(ring->ring_size > PAGE_SIZE*4); - - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_RING_OFFS); - ring->gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ring->ring = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - amdgpu_ring_clear_ring(ring); - - } else if (ring->ring_obj == NULL) { + if (ring->ring_obj == NULL) { r = amdgpu_bo_create_kernel(adev, ring->ring_size + ring->funcs->extra_dw, PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, &ring->ring_obj, @@ -401,32 +371,26 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) { /* Not to finish a ring which is not initialized */ - if (!(ring->adev) || - (!ring->is_mes_queue && !(ring->adev->rings[ring->idx]))) + if (!(ring->adev) || !(ring->adev->rings[ring->idx])) return; ring->sched.ready = false; - if (!ring->is_mes_queue) { - amdgpu_device_wb_free(ring->adev, ring->rptr_offs); - amdgpu_device_wb_free(ring->adev, ring->wptr_offs); + amdgpu_device_wb_free(ring->adev, ring->rptr_offs); + amdgpu_device_wb_free(ring->adev, ring->wptr_offs); - amdgpu_device_wb_free(ring->adev, ring->cond_exe_offs); - amdgpu_device_wb_free(ring->adev, ring->fence_offs); + amdgpu_device_wb_free(ring->adev, ring->cond_exe_offs); + amdgpu_device_wb_free(ring->adev, ring->fence_offs); - amdgpu_bo_free_kernel(&ring->ring_obj, - &ring->gpu_addr, - (void **)&ring->ring); - } else { - kfree(ring->fence_drv.fences); - } + amdgpu_bo_free_kernel(&ring->ring_obj, + &ring->gpu_addr, + (void **)&ring->ring); dma_fence_put(ring->vmid_wait); ring->vmid_wait = NULL; ring->me = 0; - if (!ring->is_mes_queue) - ring->adev->rings[ring->idx] = NULL; + ring->adev->rings[ring->idx] = NULL; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index bb2b66385223..b95b47110769 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -164,8 +164,24 @@ void amdgpu_fence_update_start_timestamp(struct amdgpu_ring *ring, uint32_t seq, /* provided by hw blocks that expose a ring buffer for commands */ struct amdgpu_ring_funcs { + /** + * @type: + * + * GFX, Compute, SDMA, UVD, VCE, VCN, VPE, KIQ, MES, UMSCH, and CPER + * use ring buffers. The type field just identifies which component the + * ring buffer is associated with. + */ enum amdgpu_ring_type type; uint32_t align_mask; + + /** + * @nop: + * + * Every block in the amdgpu has no-op instructions (e.g., GFX 10 + * uses PACKET3(PACKET3_NOP, 0x3FFF), VCN 5 uses VCN_ENC_CMD_NO_OP, + * etc). This field receives the specific no-op for the component + * that initializes the ring. + */ u32 nop; bool support_64bit_ptrs; bool no_user_fence; @@ -241,6 +257,9 @@ struct amdgpu_ring_funcs { bool (*is_guilty)(struct amdgpu_ring *ring); }; +/** + * amdgpu_ring - Holds ring information + */ struct amdgpu_ring { struct amdgpu_device *adev; const struct amdgpu_ring_funcs *funcs; @@ -252,13 +271,61 @@ struct amdgpu_ring { unsigned rptr_offs; u64 rptr_gpu_addr; volatile u32 *rptr_cpu_addr; + + /** + * @wptr: + * + * This is part of the Ring buffer implementation and represents the + * write pointer. The wptr determines where the host has written. + */ u64 wptr; + + /** + * @wptr_old: + * + * Before update wptr with the new value, usually the old value is + * stored in the wptr_old. + */ u64 wptr_old; unsigned ring_size; + + /** + * @max_dw: + * + * Maximum number of DWords for ring allocation. This information is + * provided at the ring initialization time, and each IP block can + * specify a specific value. Check places that invoke + * amdgpu_ring_init() to see the maximum size per block. + */ unsigned max_dw; + + /** + * @count_dw: + * + * This value starts with the maximum amount of DWords supported by the + * ring. This value is updated based on the ring manipulation. + */ int count_dw; uint64_t gpu_addr; + + /** + * @ptr_mask: + * + * Some IPs provide support for 64-bit pointers and others for 32-bit + * only; this behavior is component-specific and defined by the field + * support_64bit_ptr. If the IP block supports 64-bits, the mask + * 0xffffffffffffffff is set; otherwise, this value assumes buf_mask. + * Notice that this field is used to keep wptr under a valid range. + */ uint64_t ptr_mask; + + /** + * @buf_mask: + * + * Buffer mask is a value used to keep wptr count under its + * thresholding. Buffer mask initialized during the ring buffer + * initialization time, and it is defined as (ring_size / 4) -1. + */ uint32_t buf_mask; u32 idx; u32 xcc_id; @@ -276,6 +343,13 @@ struct amdgpu_ring { bool use_pollmem; unsigned wptr_offs; u64 wptr_gpu_addr; + + /** + * @wptr_cpu_addr: + * + * This is the CPU address pointer in the writeback slot. This is used + * to commit changes to the GPU. + */ volatile u32 *wptr_cpu_addr; unsigned fence_offs; u64 fence_gpu_addr; @@ -297,20 +371,15 @@ struct amdgpu_ring { struct dma_fence *vmid_wait; bool has_compute_vm_bug; bool no_scheduler; + bool no_user_submission; int hw_prio; unsigned num_hw_submission; atomic_t *sched_score; - /* used for mes */ - bool is_mes_queue; - uint32_t hw_queue_id; - struct amdgpu_mes_ctx_data *mes_ctx; - bool is_sw_ring; unsigned int entry_index; /* store the cached rptr to restore after reset */ uint64_t cached_rptr; - }; #define amdgpu_ring_parse_cs(r, p, job, ib) ((r)->funcs->parse_cs((p), (job), (ib))) @@ -435,15 +504,6 @@ static inline void amdgpu_ring_patch_cond_exec(struct amdgpu_ring *ring, ring->ring[offset] = cur - offset; } -#define amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset) \ - (ring->is_mes_queue && ring->mes_ctx ? \ - (ring->mes_ctx->meta_data_gpu_addr + offset) : 0) - -#define amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset) \ - (ring->is_mes_queue && ring->mes_ctx ? \ - (void *)((uint8_t *)(ring->mes_ctx->meta_data_ptr) + offset) : \ - NULL) - int amdgpu_ring_test_helper(struct amdgpu_ring *ring); void amdgpu_debugfs_ring_init(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h index fce22d3f816b..c210625be220 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h @@ -237,6 +237,20 @@ struct amdgpu_rlc_funcs { void (*unset_safe_mode)(struct amdgpu_device *adev, int xcc_id); int (*init)(struct amdgpu_device *adev); u32 (*get_csb_size)(struct amdgpu_device *adev); + + /** + * @get_csb_buffer: Get the clear state to be put into the hardware. + * + * The parameter adev is used to get the CS data and other gfx info, + * and buffer is the RLC CS pointer + * + * Sometimes, the user space puts a request to clear the state in the + * command buffer; this function provides the clear state that gets put + * into the hardware. Note that the driver programs Clear State + * Indirect Buffer (CSB) explicitly when it sets up the kernel rings, + * and it also provides a pointer to it which is used by the firmware + * to load the clear state in some cases. + */ void (*get_csb_buffer)(struct amdgpu_device *adev, volatile u32 *buffer); int (*get_cp_table_num)(struct amdgpu_device *adev); int (*resume)(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c index 529c9696c2f3..6716ac281c49 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c @@ -26,6 +26,8 @@ #include "amdgpu_sdma.h" #include "amdgpu_ras.h" #include "amdgpu_reset.h" +#include "gc/gc_10_1_0_offset.h" +#include "gc/gc_10_3_0_sh_mask.h" #define AMDGPU_CSA_SDMA_SIZE 64 /* SDMA CSA reside in the 3rd page of CSA */ @@ -76,22 +78,14 @@ uint64_t amdgpu_sdma_get_csa_mc_addr(struct amdgpu_ring *ring, if (amdgpu_sriov_vf(adev) || vmid == 0 || !adev->gfx.mcbp) return 0; - if (ring->is_mes_queue) { - uint32_t offset = 0; + r = amdgpu_sdma_get_index_from_ring(ring, &index); - offset = offsetof(struct amdgpu_mes_ctx_meta_data, - sdma[ring->idx].sdma_meta_data); - csa_mc_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - } else { - r = amdgpu_sdma_get_index_from_ring(ring, &index); - - if (r || index > 31) - csa_mc_addr = 0; - else - csa_mc_addr = amdgpu_csa_vaddr(adev) + - AMDGPU_CSA_SDMA_OFFSET + - index * AMDGPU_CSA_SDMA_SIZE; - } + if (r || index > 31) + csa_mc_addr = 0; + else + csa_mc_addr = amdgpu_csa_vaddr(adev) + + AMDGPU_CSA_SDMA_OFFSET + + index * AMDGPU_CSA_SDMA_SIZE; return csa_mc_addr; } @@ -537,28 +531,38 @@ bool amdgpu_sdma_is_shared_inv_eng(struct amdgpu_device *adev, struct amdgpu_rin return false; } -/** - * amdgpu_sdma_register_on_reset_callbacks - Register SDMA reset callbacks - * @funcs: Pointer to the callback structure containing pre_reset and post_reset functions - * - * This function allows KFD and AMDGPU to register their own callbacks for handling - * pre-reset and post-reset operations for engine reset. These are needed because engine - * reset will stop all queues on that engine. - */ -void amdgpu_sdma_register_on_reset_callbacks(struct amdgpu_device *adev, struct sdma_on_reset_funcs *funcs) +static int amdgpu_sdma_soft_reset(struct amdgpu_device *adev, u32 instance_id) { - if (!funcs) - return; - - /* Ensure the reset_callback_list is initialized */ - if (!adev->sdma.reset_callback_list.next) { - INIT_LIST_HEAD(&adev->sdma.reset_callback_list); + struct amdgpu_sdma_instance *sdma_instance = &adev->sdma.instance[instance_id]; + int r = -EOPNOTSUPP; + + switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) { + case IP_VERSION(4, 4, 2): + case IP_VERSION(4, 4, 4): + case IP_VERSION(4, 4, 5): + /* For SDMA 4.x, use the existing DPM interface for backward compatibility */ + r = amdgpu_dpm_reset_sdma(adev, 1 << instance_id); + break; + case IP_VERSION(5, 0, 0): + case IP_VERSION(5, 0, 1): + case IP_VERSION(5, 0, 2): + case IP_VERSION(5, 0, 5): + case IP_VERSION(5, 2, 0): + case IP_VERSION(5, 2, 2): + case IP_VERSION(5, 2, 4): + case IP_VERSION(5, 2, 5): + case IP_VERSION(5, 2, 6): + case IP_VERSION(5, 2, 3): + case IP_VERSION(5, 2, 1): + case IP_VERSION(5, 2, 7): + if (sdma_instance->funcs->soft_reset_kernel_queue) + r = sdma_instance->funcs->soft_reset_kernel_queue(adev, instance_id); + break; + default: + break; } - /* Initialize the list node in the callback structure */ - INIT_LIST_HEAD(&funcs->list); - /* Add the callback structure to the global list */ - list_add_tail(&funcs->list, &adev->sdma.reset_callback_list); + return r; } /** @@ -566,16 +570,10 @@ void amdgpu_sdma_register_on_reset_callbacks(struct amdgpu_device *adev, struct * @adev: Pointer to the AMDGPU device * @instance_id: ID of the SDMA engine instance to reset * - * This function performs the following steps: - * 1. Calls all registered pre_reset callbacks to allow KFD and AMDGPU to save their state. - * 2. Resets the specified SDMA engine instance. - * 3. Calls all registered post_reset callbacks to allow KFD and AMDGPU to restore their state. - * * Returns: 0 on success, or a negative error code on failure. */ int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id) { - struct sdma_on_reset_funcs *funcs; int ret = 0; struct amdgpu_sdma_instance *sdma_instance = &adev->sdma.instance[instance_id]; struct amdgpu_ring *gfx_ring = &sdma_instance->ring; @@ -597,38 +595,18 @@ int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id) page_sched_stopped = true; } - /* Invoke all registered pre_reset callbacks */ - list_for_each_entry(funcs, &adev->sdma.reset_callback_list, list) { - if (funcs->pre_reset) { - ret = funcs->pre_reset(adev, instance_id); - if (ret) { - dev_err(adev->dev, - "beforeReset callback failed for instance %u: %d\n", - instance_id, ret); - goto exit; - } - } - } + if (sdma_instance->funcs->stop_kernel_queue) + sdma_instance->funcs->stop_kernel_queue(gfx_ring); /* Perform the SDMA reset for the specified instance */ - ret = amdgpu_dpm_reset_sdma(adev, 1 << instance_id); + ret = amdgpu_sdma_soft_reset(adev, instance_id); if (ret) { dev_err(adev->dev, "Failed to reset SDMA instance %u\n", instance_id); goto exit; } - /* Invoke all registered post_reset callbacks */ - list_for_each_entry(funcs, &adev->sdma.reset_callback_list, list) { - if (funcs->post_reset) { - ret = funcs->post_reset(adev, instance_id); - if (ret) { - dev_err(adev->dev, - "afterReset callback failed for instance %u: %d\n", - instance_id, ret); - goto exit; - } - } - } + if (sdma_instance->funcs->start_kernel_queue) + sdma_instance->funcs->start_kernel_queue(gfx_ring); exit: /* Restart the scheduler's work queue for the GFX and page rings diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h index 47d56fd0589f..5605921212f0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h @@ -50,6 +50,12 @@ enum amdgpu_sdma_irq { #define NUM_SDMA(x) hweight32(x) +struct amdgpu_sdma_funcs { + int (*stop_kernel_queue)(struct amdgpu_ring *ring); + int (*start_kernel_queue)(struct amdgpu_ring *ring); + int (*soft_reset_kernel_queue)(struct amdgpu_device *adev, u32 instance_id); +}; + struct amdgpu_sdma_instance { /* SDMA firmware */ const struct firmware *fw; @@ -68,7 +74,7 @@ struct amdgpu_sdma_instance { /* track guilty state of GFX and PAGE queues */ bool gfx_guilty; bool page_guilty; - + const struct amdgpu_sdma_funcs *funcs; }; enum amdgpu_sdma_ras_memory_id { @@ -103,13 +109,6 @@ struct amdgpu_sdma_ras { struct amdgpu_ras_block_object ras_block; }; -struct sdma_on_reset_funcs { - int (*pre_reset)(struct amdgpu_device *adev, uint32_t instance_id); - int (*post_reset)(struct amdgpu_device *adev, uint32_t instance_id); - /* Linked list node to store this structure in a list; */ - struct list_head list; -}; - struct amdgpu_sdma { struct amdgpu_sdma_instance instance[AMDGPU_MAX_SDMA_INSTANCES]; struct amdgpu_irq_src trap_irq; @@ -131,6 +130,8 @@ struct amdgpu_sdma { uint32_t *ip_dump; uint32_t supported_reset; struct list_head reset_callback_list; + bool no_user_submission; + bool disable_uq; }; /* @@ -170,7 +171,6 @@ struct amdgpu_buffer_funcs { uint32_t byte_count); }; -void amdgpu_sdma_register_on_reset_callbacks(struct amdgpu_device *adev, struct sdma_on_reset_funcs *funcs); int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id); #define amdgpu_emit_copy_buffer(adev, ib, s, d, b, t) (adev)->mman.buffer_funcs->emit_copy_buffer((ib), (s), (d), (b), (t)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c index e22cb2b5cd92..3939761be31c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c @@ -45,7 +45,11 @@ */ static inline u64 amdgpu_seq64_get_va_base(struct amdgpu_device *adev) { - return AMDGPU_VA_RESERVED_SEQ64_START(adev); + u64 addr = AMDGPU_VA_RESERVED_SEQ64_START(adev); + + addr = amdgpu_gmc_sign_extend(addr); + + return addr; } /** @@ -63,9 +67,9 @@ static inline u64 amdgpu_seq64_get_va_base(struct amdgpu_device *adev) int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo_va **bo_va) { + u64 seq64_addr, va_flags; struct amdgpu_bo *bo; struct drm_exec exec; - u64 seq64_addr; int r; bo = adev->seq64.sbo; @@ -88,9 +92,11 @@ int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm, goto error; } - seq64_addr = amdgpu_seq64_get_va_base(adev); + seq64_addr = amdgpu_seq64_get_va_base(adev) & AMDGPU_GMC_HOLE_MASK; + + va_flags = amdgpu_gem_va_map_flags(adev, AMDGPU_VM_PAGE_READABLE | AMDGPU_VM_MTYPE_UC); r = amdgpu_vm_bo_map(adev, *bo_va, seq64_addr, 0, AMDGPU_VA_RESERVED_SEQ64_SIZE, - AMDGPU_PTE_READABLE); + va_flags); if (r) { DRM_ERROR("failed to do bo_map on userq sem, err=%d\n", r); amdgpu_vm_bo_del(adev, *bo_va); @@ -156,6 +162,7 @@ error: * * @adev: amdgpu_device pointer * @va: VA to access the seq in process address space + * @gpu_addr: GPU address to access the seq * @cpu_addr: CPU address to access the seq * * Alloc a 64 bit memory from seq64 pool. @@ -163,7 +170,8 @@ error: * Returns: * 0 on success or a negative error code on failure */ -int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va, u64 **cpu_addr) +int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va, + u64 *gpu_addr, u64 **cpu_addr) { unsigned long bit_pos; @@ -172,7 +180,12 @@ int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va, u64 **cpu_addr) return -ENOSPC; __set_bit(bit_pos, adev->seq64.used); + *va = bit_pos * sizeof(u64) + amdgpu_seq64_get_va_base(adev); + + if (gpu_addr) + *gpu_addr = bit_pos * sizeof(u64) + adev->seq64.gpu_addr; + *cpu_addr = bit_pos + adev->seq64.cpu_base_addr; return 0; @@ -233,7 +246,7 @@ int amdgpu_seq64_init(struct amdgpu_device *adev) */ r = amdgpu_bo_create_kernel(adev, AMDGPU_VA_RESERVED_SEQ64_SIZE, PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, - &adev->seq64.sbo, NULL, + &adev->seq64.sbo, &adev->seq64.gpu_addr, (void **)&adev->seq64.cpu_base_addr); if (r) { dev_warn(adev->dev, "(%d) create seq64 failed\n", r); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h index 4203b2ab318d..26a249aaaee1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h @@ -32,13 +32,14 @@ struct amdgpu_seq64 { struct amdgpu_bo *sbo; u32 num_sem; + u64 gpu_addr; u64 *cpu_base_addr; DECLARE_BITMAP(used, AMDGPU_MAX_SEQ64_SLOTS); }; void amdgpu_seq64_fini(struct amdgpu_device *adev); int amdgpu_seq64_init(struct amdgpu_device *adev); -int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *gpu_addr, u64 **cpu_addr); +int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va, u64 *gpu_addr, u64 **cpu_addr); void amdgpu_seq64_free(struct amdgpu_device *adev, u64 gpu_addr); int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo_va **bo_va); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index 5576ed0b508f..d6ae9974c952 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -249,9 +249,8 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync, if (resv == NULL) return -EINVAL; - - /* TODO: Use DMA_RESV_USAGE_READ here */ - dma_resv_for_each_fence(&cursor, resv, DMA_RESV_USAGE_BOOKKEEP, f) { + /* Implicitly sync only to KERNEL, WRITE and READ */ + dma_resv_for_each_fence(&cursor, resv, DMA_RESV_USAGE_READ, f) { dma_fence_chain_for_each(f, f) { struct dma_fence *tmp = dma_fence_chain_contained(f); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 53b71e9d8076..9c5df35f05b7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -2081,6 +2081,8 @@ void amdgpu_ttm_fini(struct amdgpu_device *adev) amdgpu_vram_mgr_fini(adev); amdgpu_gtt_mgr_fini(adev); amdgpu_preempt_mgr_fini(adev); + amdgpu_doorbell_fini(adev); + ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_GDS); ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_GWS); ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_OA); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index 3d9e9fdc10b4..4a72c2bbd49e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -767,6 +767,7 @@ FW_VERSION_ATTR(vcn_fw_version, 0444, vcn.fw_version); FW_VERSION_ATTR(dmcu_fw_version, 0444, dm.dmcu_fw_version); FW_VERSION_ATTR(mes_fw_version, 0444, mes.sched_version & AMDGPU_MES_VERSION_MASK); FW_VERSION_ATTR(mes_kiq_fw_version, 0444, mes.kiq_version & AMDGPU_MES_VERSION_MASK); +FW_VERSION_ATTR(pldm_fw_version, 0444, firmware.pldm_version); static struct attribute *fw_attrs[] = { &dev_attr_vce_fw_version.attr, &dev_attr_uvd_fw_version.attr, @@ -781,7 +782,7 @@ static struct attribute *fw_attrs[] = { &dev_attr_sdma2_fw_version.attr, &dev_attr_vcn_fw_version.attr, &dev_attr_dmcu_fw_version.attr, &dev_attr_imu_fw_version.attr, &dev_attr_mes_fw_version.attr, &dev_attr_mes_kiq_fw_version.attr, - NULL + &dev_attr_pldm_fw_version.attr, NULL }; #define to_dev_attr(x) container_of(x, struct device_attribute, attr) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h index 4eedd92f000b..9e89c3487be5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h @@ -25,6 +25,8 @@ #include "amdgpu_socbb.h" +#define RS64_FW_UC_START_ADDR_LO 0x3000 + struct common_firmware_header { uint32_t size_bytes; /* size of the entire header+image(s) in bytes */ uint32_t header_size_bytes; /* size of just the header in bytes */ @@ -600,6 +602,7 @@ struct amdgpu_firmware { void *fw_buf_ptr; uint64_t fw_buf_mc; + uint32_t pldm_version; }; void amdgpu_ucode_print_mc_hdr(const struct common_firmware_header *hdr); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index 0a1ef95b2866..8c6e55b5b967 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -529,6 +529,7 @@ int amdgpu_umc_lookup_bad_pages_in_a_row(struct amdgpu_device *adev, pfns[i] = err_data.err_addr[i].retired_page; } ret = i; + adev->umc.err_addr_cnt = err_data.err_addr_cnt; out: kfree(err_data.err_addr); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index 857693bcd8d4..29ce6b1d214a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -78,6 +78,18 @@ #define UMC_NPS_SHIFT 40 #define UMC_NPS_MASK 0xffULL +/* three column bits and one row bit in MCA address flip + * in bad page retirement + */ +#define RETIRE_FLIP_BITS_NUM 4 + +struct amdgpu_umc_flip_bits { + uint32_t flip_bits_in_pa[RETIRE_FLIP_BITS_NUM]; + uint32_t flip_row_bit; + uint32_t r13_in_pa; + uint32_t bit_num; +}; + typedef int (*umc_func)(struct amdgpu_device *adev, uint32_t node_inst, uint32_t umc_inst, uint32_t ch_inst, void *data); @@ -100,6 +112,7 @@ struct amdgpu_umc_ras { bool dump_addr); uint32_t (*get_die_id_from_pa)(struct amdgpu_device *adev, uint64_t mca_addr, uint64_t retired_page); + void (*get_retire_flip_bits)(struct amdgpu_device *adev); }; struct amdgpu_umc_funcs { @@ -130,6 +143,10 @@ struct amdgpu_umc { /* active mask for umc node instance */ unsigned long active_mask; + + struct amdgpu_umc_flip_bits flip_bits; + + unsigned long err_addr_cnt; }; int amdgpu_umc_ras_sw_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c new file mode 100644 index 000000000000..295e7186e156 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c @@ -0,0 +1,924 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include <drm/drm_auth.h> +#include <drm/drm_exec.h> +#include <linux/pm_runtime.h> + +#include "amdgpu.h" +#include "amdgpu_vm.h" +#include "amdgpu_userq.h" +#include "amdgpu_userq_fence.h" + +u32 amdgpu_userq_get_supported_ip_mask(struct amdgpu_device *adev) +{ + int i; + u32 userq_ip_mask = 0; + + for (i = 0; i < AMDGPU_HW_IP_NUM; i++) { + if (adev->userq_funcs[i]) + userq_ip_mask |= (1 << i); + } + + return userq_ip_mask; +} + +static int +amdgpu_userq_unmap_helper(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + struct amdgpu_device *adev = uq_mgr->adev; + const struct amdgpu_userq_funcs *userq_funcs = + adev->userq_funcs[queue->queue_type]; + int r = 0; + + if (queue->state == AMDGPU_USERQ_STATE_MAPPED) { + r = userq_funcs->unmap(uq_mgr, queue); + if (r) + queue->state = AMDGPU_USERQ_STATE_HUNG; + else + queue->state = AMDGPU_USERQ_STATE_UNMAPPED; + } + return r; +} + +static int +amdgpu_userq_map_helper(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + struct amdgpu_device *adev = uq_mgr->adev; + const struct amdgpu_userq_funcs *userq_funcs = + adev->userq_funcs[queue->queue_type]; + int r = 0; + + if (queue->state == AMDGPU_USERQ_STATE_UNMAPPED) { + r = userq_funcs->map(uq_mgr, queue); + if (r) { + queue->state = AMDGPU_USERQ_STATE_HUNG; + } else { + queue->state = AMDGPU_USERQ_STATE_MAPPED; + } + } + return r; +} + +static void +amdgpu_userq_wait_for_last_fence(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + struct dma_fence *f = queue->last_fence; + int ret; + + if (f && !dma_fence_is_signaled(f)) { + ret = dma_fence_wait_timeout(f, true, msecs_to_jiffies(100)); + if (ret <= 0) + drm_file_err(uq_mgr->file, "Timed out waiting for fence=%llu:%llu\n", + f->context, f->seqno); + } +} + +static void +amdgpu_userq_cleanup(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue, + int queue_id) +{ + struct amdgpu_device *adev = uq_mgr->adev; + const struct amdgpu_userq_funcs *uq_funcs = adev->userq_funcs[queue->queue_type]; + + uq_funcs->mqd_destroy(uq_mgr, queue); + amdgpu_userq_fence_driver_free(queue); + idr_remove(&uq_mgr->userq_idr, queue_id); + kfree(queue); +} + +int +amdgpu_userq_active(struct amdgpu_userq_mgr *uq_mgr) +{ + struct amdgpu_usermode_queue *queue; + int queue_id; + int ret = 0; + + mutex_lock(&uq_mgr->userq_mutex); + /* Resume all the queues for this process */ + idr_for_each_entry(&uq_mgr->userq_idr, queue, queue_id) + ret += queue->state == AMDGPU_USERQ_STATE_MAPPED; + + mutex_unlock(&uq_mgr->userq_mutex); + return ret; +} + +static struct amdgpu_usermode_queue * +amdgpu_userq_find(struct amdgpu_userq_mgr *uq_mgr, int qid) +{ + return idr_find(&uq_mgr->userq_idr, qid); +} + +void +amdgpu_userq_ensure_ev_fence(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_eviction_fence_mgr *evf_mgr) +{ + struct amdgpu_eviction_fence *ev_fence; + +retry: + /* Flush any pending resume work to create ev_fence */ + flush_delayed_work(&uq_mgr->resume_work); + + mutex_lock(&uq_mgr->userq_mutex); + spin_lock(&evf_mgr->ev_fence_lock); + ev_fence = evf_mgr->ev_fence; + spin_unlock(&evf_mgr->ev_fence_lock); + if (!ev_fence || dma_fence_is_signaled(&ev_fence->base)) { + mutex_unlock(&uq_mgr->userq_mutex); + /* + * Looks like there was no pending resume work, + * add one now to create a valid eviction fence + */ + schedule_delayed_work(&uq_mgr->resume_work, 0); + goto retry; + } +} + +int amdgpu_userq_create_object(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_userq_obj *userq_obj, + int size) +{ + struct amdgpu_device *adev = uq_mgr->adev; + struct amdgpu_bo_param bp; + int r; + + memset(&bp, 0, sizeof(bp)); + bp.byte_align = PAGE_SIZE; + bp.domain = AMDGPU_GEM_DOMAIN_GTT; + bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED; + bp.type = ttm_bo_type_kernel; + bp.size = size; + bp.resv = NULL; + bp.bo_ptr_size = sizeof(struct amdgpu_bo); + + r = amdgpu_bo_create(adev, &bp, &userq_obj->obj); + if (r) { + drm_file_err(uq_mgr->file, "Failed to allocate BO for userqueue (%d)", r); + return r; + } + + r = amdgpu_bo_reserve(userq_obj->obj, true); + if (r) { + drm_file_err(uq_mgr->file, "Failed to reserve BO to map (%d)", r); + goto free_obj; + } + + r = amdgpu_ttm_alloc_gart(&(userq_obj->obj)->tbo); + if (r) { + drm_file_err(uq_mgr->file, "Failed to alloc GART for userqueue object (%d)", r); + goto unresv; + } + + r = amdgpu_bo_kmap(userq_obj->obj, &userq_obj->cpu_ptr); + if (r) { + drm_file_err(uq_mgr->file, "Failed to map BO for userqueue (%d)", r); + goto unresv; + } + + userq_obj->gpu_addr = amdgpu_bo_gpu_offset(userq_obj->obj); + amdgpu_bo_unreserve(userq_obj->obj); + memset(userq_obj->cpu_ptr, 0, size); + return 0; + +unresv: + amdgpu_bo_unreserve(userq_obj->obj); + +free_obj: + amdgpu_bo_unref(&userq_obj->obj); + return r; +} + +void amdgpu_userq_destroy_object(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_userq_obj *userq_obj) +{ + amdgpu_bo_kunmap(userq_obj->obj); + amdgpu_bo_unref(&userq_obj->obj); +} + +uint64_t +amdgpu_userq_get_doorbell_index(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_db_info *db_info, + struct drm_file *filp) +{ + uint64_t index; + struct drm_gem_object *gobj; + struct amdgpu_userq_obj *db_obj = db_info->db_obj; + int r, db_size; + + gobj = drm_gem_object_lookup(filp, db_info->doorbell_handle); + if (gobj == NULL) { + drm_file_err(uq_mgr->file, "Can't find GEM object for doorbell\n"); + return -EINVAL; + } + + db_obj->obj = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); + drm_gem_object_put(gobj); + + r = amdgpu_bo_reserve(db_obj->obj, true); + if (r) { + drm_file_err(uq_mgr->file, "[Usermode queues] Failed to pin doorbell object\n"); + goto unref_bo; + } + + /* Pin the BO before generating the index, unpin in queue destroy */ + r = amdgpu_bo_pin(db_obj->obj, AMDGPU_GEM_DOMAIN_DOORBELL); + if (r) { + drm_file_err(uq_mgr->file, "[Usermode queues] Failed to pin doorbell object\n"); + goto unresv_bo; + } + + switch (db_info->queue_type) { + case AMDGPU_HW_IP_GFX: + case AMDGPU_HW_IP_COMPUTE: + case AMDGPU_HW_IP_DMA: + db_size = sizeof(u64); + break; + + case AMDGPU_HW_IP_VCN_ENC: + db_size = sizeof(u32); + db_info->doorbell_offset += AMDGPU_NAVI10_DOORBELL64_VCN0_1 << 1; + break; + + case AMDGPU_HW_IP_VPE: + db_size = sizeof(u32); + db_info->doorbell_offset += AMDGPU_NAVI10_DOORBELL64_VPE << 1; + break; + + default: + drm_file_err(uq_mgr->file, "[Usermode queues] IP %d not support\n", + db_info->queue_type); + r = -EINVAL; + goto unpin_bo; + } + + index = amdgpu_doorbell_index_on_bar(uq_mgr->adev, db_obj->obj, + db_info->doorbell_offset, db_size); + drm_dbg_driver(adev_to_drm(uq_mgr->adev), + "[Usermode queues] doorbell index=%lld\n", index); + amdgpu_bo_unreserve(db_obj->obj); + return index; + +unpin_bo: + amdgpu_bo_unpin(db_obj->obj); +unresv_bo: + amdgpu_bo_unreserve(db_obj->obj); +unref_bo: + amdgpu_bo_unref(&db_obj->obj); + return r; +} + +static int +amdgpu_userq_destroy(struct drm_file *filp, int queue_id) +{ + struct amdgpu_fpriv *fpriv = filp->driver_priv; + struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr; + struct amdgpu_device *adev = uq_mgr->adev; + struct amdgpu_usermode_queue *queue; + int r = 0; + + cancel_delayed_work_sync(&uq_mgr->resume_work); + mutex_lock(&uq_mgr->userq_mutex); + + queue = amdgpu_userq_find(uq_mgr, queue_id); + if (!queue) { + drm_dbg_driver(adev_to_drm(uq_mgr->adev), "Invalid queue id to destroy\n"); + mutex_unlock(&uq_mgr->userq_mutex); + return -EINVAL; + } + amdgpu_userq_wait_for_last_fence(uq_mgr, queue); + r = amdgpu_bo_reserve(queue->db_obj.obj, true); + if (!r) { + amdgpu_bo_unpin(queue->db_obj.obj); + amdgpu_bo_unreserve(queue->db_obj.obj); + } + amdgpu_bo_unref(&queue->db_obj.obj); + r = amdgpu_userq_unmap_helper(uq_mgr, queue); + amdgpu_userq_cleanup(uq_mgr, queue, queue_id); + mutex_unlock(&uq_mgr->userq_mutex); + + pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + + return r; +} + +static int amdgpu_userq_priority_permit(struct drm_file *filp, + int priority) +{ + if (priority < AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_HIGH) + return 0; + + if (capable(CAP_SYS_NICE)) + return 0; + + if (drm_is_current_master(filp)) + return 0; + + return -EACCES; +} + +static int +amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args) +{ + struct amdgpu_fpriv *fpriv = filp->driver_priv; + struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr; + struct amdgpu_device *adev = uq_mgr->adev; + const struct amdgpu_userq_funcs *uq_funcs; + struct amdgpu_usermode_queue *queue; + struct amdgpu_db_info db_info; + bool skip_map_queue; + uint64_t index; + int qid, r = 0; + int priority = + (args->in.flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_MASK) >> + AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_SHIFT; + + /* Usermode queues are only supported for GFX IP as of now */ + if (args->in.ip_type != AMDGPU_HW_IP_GFX && + args->in.ip_type != AMDGPU_HW_IP_DMA && + args->in.ip_type != AMDGPU_HW_IP_COMPUTE) { + drm_file_err(uq_mgr->file, "Usermode queue doesn't support IP type %u\n", + args->in.ip_type); + return -EINVAL; + } + + r = amdgpu_userq_priority_permit(filp, priority); + if (r) + return r; + + if ((args->in.flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE) && + (args->in.ip_type != AMDGPU_HW_IP_GFX) && + (args->in.ip_type != AMDGPU_HW_IP_COMPUTE) && + !amdgpu_is_tmz(adev)) { + drm_file_err(uq_mgr->file, "Secure only supported on GFX/Compute queues\n"); + return -EINVAL; + } + + r = pm_runtime_get_sync(adev_to_drm(adev)->dev); + if (r < 0) { + drm_file_err(uq_mgr->file, "pm_runtime_get_sync() failed for userqueue create\n"); + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + return r; + } + + /* + * There could be a situation that we are creating a new queue while + * the other queues under this UQ_mgr are suspended. So if there is any + * resume work pending, wait for it to get done. + * + * This will also make sure we have a valid eviction fence ready to be used. + */ + mutex_lock(&adev->userq_mutex); + amdgpu_userq_ensure_ev_fence(&fpriv->userq_mgr, &fpriv->evf_mgr); + + uq_funcs = adev->userq_funcs[args->in.ip_type]; + if (!uq_funcs) { + drm_file_err(uq_mgr->file, "Usermode queue is not supported for this IP (%u)\n", + args->in.ip_type); + r = -EINVAL; + goto unlock; + } + + queue = kzalloc(sizeof(struct amdgpu_usermode_queue), GFP_KERNEL); + if (!queue) { + drm_file_err(uq_mgr->file, "Failed to allocate memory for queue\n"); + r = -ENOMEM; + goto unlock; + } + queue->doorbell_handle = args->in.doorbell_handle; + queue->queue_type = args->in.ip_type; + queue->vm = &fpriv->vm; + queue->priority = priority; + + db_info.queue_type = queue->queue_type; + db_info.doorbell_handle = queue->doorbell_handle; + db_info.db_obj = &queue->db_obj; + db_info.doorbell_offset = args->in.doorbell_offset; + + /* Convert relative doorbell offset into absolute doorbell index */ + index = amdgpu_userq_get_doorbell_index(uq_mgr, &db_info, filp); + if (index == (uint64_t)-EINVAL) { + drm_file_err(uq_mgr->file, "Failed to get doorbell for queue\n"); + kfree(queue); + goto unlock; + } + + queue->doorbell_index = index; + xa_init_flags(&queue->fence_drv_xa, XA_FLAGS_ALLOC); + r = amdgpu_userq_fence_driver_alloc(adev, queue); + if (r) { + drm_file_err(uq_mgr->file, "Failed to alloc fence driver\n"); + goto unlock; + } + + r = uq_funcs->mqd_create(uq_mgr, &args->in, queue); + if (r) { + drm_file_err(uq_mgr->file, "Failed to create Queue\n"); + amdgpu_userq_fence_driver_free(queue); + kfree(queue); + goto unlock; + } + + + qid = idr_alloc(&uq_mgr->userq_idr, queue, 1, AMDGPU_MAX_USERQ_COUNT, GFP_KERNEL); + if (qid < 0) { + drm_file_err(uq_mgr->file, "Failed to allocate a queue id\n"); + amdgpu_userq_fence_driver_free(queue); + uq_funcs->mqd_destroy(uq_mgr, queue); + kfree(queue); + r = -ENOMEM; + goto unlock; + } + + /* don't map the queue if scheduling is halted */ + if (adev->userq_halt_for_enforce_isolation && + ((queue->queue_type == AMDGPU_HW_IP_GFX) || + (queue->queue_type == AMDGPU_HW_IP_COMPUTE))) + skip_map_queue = true; + else + skip_map_queue = false; + if (!skip_map_queue) { + r = amdgpu_userq_map_helper(uq_mgr, queue); + if (r) { + drm_file_err(uq_mgr->file, "Failed to map Queue\n"); + idr_remove(&uq_mgr->userq_idr, qid); + amdgpu_userq_fence_driver_free(queue); + uq_funcs->mqd_destroy(uq_mgr, queue); + kfree(queue); + goto unlock; + } + } + + + args->out.queue_id = qid; + +unlock: + mutex_unlock(&uq_mgr->userq_mutex); + mutex_unlock(&adev->userq_mutex); + + return r; +} + +int amdgpu_userq_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp) +{ + union drm_amdgpu_userq *args = data; + int r; + + switch (args->in.op) { + case AMDGPU_USERQ_OP_CREATE: + if (args->in.flags & ~(AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_MASK | + AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE)) + return -EINVAL; + r = amdgpu_userq_create(filp, args); + if (r) + drm_file_err(filp, "Failed to create usermode queue\n"); + break; + + case AMDGPU_USERQ_OP_FREE: + if (args->in.ip_type || + args->in.doorbell_handle || + args->in.doorbell_offset || + args->in.flags || + args->in.queue_va || + args->in.queue_size || + args->in.rptr_va || + args->in.wptr_va || + args->in.wptr_va || + args->in.mqd || + args->in.mqd_size) + return -EINVAL; + r = amdgpu_userq_destroy(filp, args->in.queue_id); + if (r) + drm_file_err(filp, "Failed to destroy usermode queue\n"); + break; + + default: + drm_dbg_driver(dev, "Invalid user queue op specified: %d\n", args->in.op); + return -EINVAL; + } + + return r; +} + +static int +amdgpu_userq_restore_all(struct amdgpu_userq_mgr *uq_mgr) +{ + struct amdgpu_usermode_queue *queue; + int queue_id; + int ret = 0, r; + + /* Resume all the queues for this process */ + idr_for_each_entry(&uq_mgr->userq_idr, queue, queue_id) { + r = amdgpu_userq_map_helper(uq_mgr, queue); + if (r) + ret = r; + } + + if (ret) + drm_file_err(uq_mgr->file, "Failed to map all the queues\n"); + return ret; +} + +static int +amdgpu_userq_validate_vm_bo(void *_unused, struct amdgpu_bo *bo) +{ + struct ttm_operation_ctx ctx = { false, false }; + int ret; + + amdgpu_bo_placement_from_domain(bo, bo->allowed_domains); + + ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + if (ret) + DRM_ERROR("Fail to validate\n"); + + return ret; +} + +static int +amdgpu_userq_validate_bos(struct amdgpu_userq_mgr *uq_mgr) +{ + struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr); + struct amdgpu_vm *vm = &fpriv->vm; + struct amdgpu_device *adev = uq_mgr->adev; + struct amdgpu_bo_va *bo_va; + struct ww_acquire_ctx *ticket; + struct drm_exec exec; + struct amdgpu_bo *bo; + struct dma_resv *resv; + bool clear, unlock; + int ret = 0; + + drm_exec_init(&exec, DRM_EXEC_IGNORE_DUPLICATES, 0); + drm_exec_until_all_locked(&exec) { + ret = amdgpu_vm_lock_pd(vm, &exec, 2); + drm_exec_retry_on_contention(&exec); + if (unlikely(ret)) { + drm_file_err(uq_mgr->file, "Failed to lock PD\n"); + goto unlock_all; + } + + /* Lock the done list */ + list_for_each_entry(bo_va, &vm->done, base.vm_status) { + bo = bo_va->base.bo; + if (!bo) + continue; + + ret = drm_exec_lock_obj(&exec, &bo->tbo.base); + drm_exec_retry_on_contention(&exec); + if (unlikely(ret)) + goto unlock_all; + } + } + + spin_lock(&vm->status_lock); + while (!list_empty(&vm->moved)) { + bo_va = list_first_entry(&vm->moved, struct amdgpu_bo_va, + base.vm_status); + spin_unlock(&vm->status_lock); + + /* Per VM BOs never need to bo cleared in the page tables */ + ret = amdgpu_vm_bo_update(adev, bo_va, false); + if (ret) + goto unlock_all; + spin_lock(&vm->status_lock); + } + + ticket = &exec.ticket; + while (!list_empty(&vm->invalidated)) { + bo_va = list_first_entry(&vm->invalidated, struct amdgpu_bo_va, + base.vm_status); + resv = bo_va->base.bo->tbo.base.resv; + spin_unlock(&vm->status_lock); + + bo = bo_va->base.bo; + ret = amdgpu_userq_validate_vm_bo(NULL, bo); + if (ret) { + drm_file_err(uq_mgr->file, "Failed to validate BO\n"); + goto unlock_all; + } + + /* Try to reserve the BO to avoid clearing its ptes */ + if (!adev->debug_vm && dma_resv_trylock(resv)) { + clear = false; + unlock = true; + /* The caller is already holding the reservation lock */ + } else if (dma_resv_locking_ctx(resv) == ticket) { + clear = false; + unlock = false; + /* Somebody else is using the BO right now */ + } else { + clear = true; + unlock = false; + } + + ret = amdgpu_vm_bo_update(adev, bo_va, clear); + + if (unlock) + dma_resv_unlock(resv); + if (ret) + goto unlock_all; + + spin_lock(&vm->status_lock); + } + spin_unlock(&vm->status_lock); + + ret = amdgpu_eviction_fence_replace_fence(&fpriv->evf_mgr, &exec); + if (ret) + drm_file_err(uq_mgr->file, "Failed to replace eviction fence\n"); + +unlock_all: + drm_exec_fini(&exec); + return ret; +} + +static void amdgpu_userq_restore_worker(struct work_struct *work) +{ + struct amdgpu_userq_mgr *uq_mgr = work_to_uq_mgr(work, resume_work.work); + struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr); + int ret; + + flush_work(&fpriv->evf_mgr.suspend_work.work); + + mutex_lock(&uq_mgr->userq_mutex); + + ret = amdgpu_userq_validate_bos(uq_mgr); + if (ret) { + drm_file_err(uq_mgr->file, "Failed to validate BOs to restore\n"); + goto unlock; + } + + ret = amdgpu_userq_restore_all(uq_mgr); + if (ret) { + drm_file_err(uq_mgr->file, "Failed to restore all queues\n"); + goto unlock; + } + +unlock: + mutex_unlock(&uq_mgr->userq_mutex); +} + +static int +amdgpu_userq_evict_all(struct amdgpu_userq_mgr *uq_mgr) +{ + struct amdgpu_usermode_queue *queue; + int queue_id; + int ret = 0, r; + + /* Try to unmap all the queues in this process ctx */ + idr_for_each_entry(&uq_mgr->userq_idr, queue, queue_id) { + r = amdgpu_userq_unmap_helper(uq_mgr, queue); + if (r) + ret = r; + } + + if (ret) + drm_file_err(uq_mgr->file, "Couldn't unmap all the queues\n"); + return ret; +} + +static int +amdgpu_userq_wait_for_signal(struct amdgpu_userq_mgr *uq_mgr) +{ + struct amdgpu_usermode_queue *queue; + int queue_id, ret; + + idr_for_each_entry(&uq_mgr->userq_idr, queue, queue_id) { + struct dma_fence *f = queue->last_fence; + + if (!f || dma_fence_is_signaled(f)) + continue; + ret = dma_fence_wait_timeout(f, true, msecs_to_jiffies(100)); + if (ret <= 0) { + drm_file_err(uq_mgr->file, "Timed out waiting for fence=%llu:%llu\n", + f->context, f->seqno); + return -ETIMEDOUT; + } + } + + return 0; +} + +void +amdgpu_userq_evict(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_eviction_fence *ev_fence) +{ + int ret; + struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr); + struct amdgpu_eviction_fence_mgr *evf_mgr = &fpriv->evf_mgr; + + /* Wait for any pending userqueue fence work to finish */ + ret = amdgpu_userq_wait_for_signal(uq_mgr); + if (ret) { + drm_file_err(uq_mgr->file, "Not evicting userqueue, timeout waiting for work\n"); + return; + } + + ret = amdgpu_userq_evict_all(uq_mgr); + if (ret) { + drm_file_err(uq_mgr->file, "Failed to evict userqueue\n"); + return; + } + + /* Signal current eviction fence */ + amdgpu_eviction_fence_signal(evf_mgr, ev_fence); + + if (evf_mgr->fd_closing) { + cancel_delayed_work_sync(&uq_mgr->resume_work); + return; + } + + /* Schedule a resume work */ + schedule_delayed_work(&uq_mgr->resume_work, 0); +} + +int amdgpu_userq_mgr_init(struct amdgpu_userq_mgr *userq_mgr, struct drm_file *file_priv, + struct amdgpu_device *adev) +{ + mutex_init(&userq_mgr->userq_mutex); + idr_init_base(&userq_mgr->userq_idr, 1); + userq_mgr->adev = adev; + userq_mgr->file = file_priv; + + mutex_lock(&adev->userq_mutex); + list_add(&userq_mgr->list, &adev->userq_mgr_list); + mutex_unlock(&adev->userq_mutex); + + INIT_DELAYED_WORK(&userq_mgr->resume_work, amdgpu_userq_restore_worker); + return 0; +} + +void amdgpu_userq_mgr_fini(struct amdgpu_userq_mgr *userq_mgr) +{ + struct amdgpu_device *adev = userq_mgr->adev; + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_mgr *uqm, *tmp; + uint32_t queue_id; + + cancel_delayed_work_sync(&userq_mgr->resume_work); + + mutex_lock(&adev->userq_mutex); + mutex_lock(&userq_mgr->userq_mutex); + idr_for_each_entry(&userq_mgr->userq_idr, queue, queue_id) { + amdgpu_userq_wait_for_last_fence(userq_mgr, queue); + amdgpu_userq_unmap_helper(userq_mgr, queue); + amdgpu_userq_cleanup(userq_mgr, queue, queue_id); + } + + list_for_each_entry_safe(uqm, tmp, &adev->userq_mgr_list, list) { + if (uqm == userq_mgr) { + list_del(&uqm->list); + break; + } + } + idr_destroy(&userq_mgr->userq_idr); + mutex_unlock(&userq_mgr->userq_mutex); + mutex_unlock(&adev->userq_mutex); + mutex_destroy(&userq_mgr->userq_mutex); +} + +int amdgpu_userq_suspend(struct amdgpu_device *adev) +{ + u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev); + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_mgr *uqm, *tmp; + int queue_id; + int ret = 0, r; + + if (!ip_mask) + return 0; + + mutex_lock(&adev->userq_mutex); + list_for_each_entry_safe(uqm, tmp, &adev->userq_mgr_list, list) { + cancel_delayed_work_sync(&uqm->resume_work); + mutex_lock(&uqm->userq_mutex); + idr_for_each_entry(&uqm->userq_idr, queue, queue_id) { + r = amdgpu_userq_unmap_helper(uqm, queue); + if (r) + ret = r; + } + mutex_unlock(&uqm->userq_mutex); + } + mutex_unlock(&adev->userq_mutex); + return ret; +} + +int amdgpu_userq_resume(struct amdgpu_device *adev) +{ + u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev); + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_mgr *uqm, *tmp; + int queue_id; + int ret = 0, r; + + if (!ip_mask) + return 0; + + mutex_lock(&adev->userq_mutex); + list_for_each_entry_safe(uqm, tmp, &adev->userq_mgr_list, list) { + mutex_lock(&uqm->userq_mutex); + idr_for_each_entry(&uqm->userq_idr, queue, queue_id) { + r = amdgpu_userq_map_helper(uqm, queue); + if (r) + ret = r; + } + mutex_unlock(&uqm->userq_mutex); + } + mutex_unlock(&adev->userq_mutex); + return ret; +} + +int amdgpu_userq_stop_sched_for_enforce_isolation(struct amdgpu_device *adev, + u32 idx) +{ + u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev); + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_mgr *uqm, *tmp; + int queue_id; + int ret = 0, r; + + /* only need to stop gfx/compute */ + if (!(ip_mask & ((1 << AMDGPU_HW_IP_GFX) | (1 << AMDGPU_HW_IP_COMPUTE)))) + return 0; + + mutex_lock(&adev->userq_mutex); + if (adev->userq_halt_for_enforce_isolation) + dev_warn(adev->dev, "userq scheduling already stopped!\n"); + adev->userq_halt_for_enforce_isolation = true; + list_for_each_entry_safe(uqm, tmp, &adev->userq_mgr_list, list) { + cancel_delayed_work_sync(&uqm->resume_work); + mutex_lock(&uqm->userq_mutex); + idr_for_each_entry(&uqm->userq_idr, queue, queue_id) { + if (((queue->queue_type == AMDGPU_HW_IP_GFX) || + (queue->queue_type == AMDGPU_HW_IP_COMPUTE)) && + (queue->xcp_id == idx)) { + r = amdgpu_userq_unmap_helper(uqm, queue); + if (r) + ret = r; + } + } + mutex_unlock(&uqm->userq_mutex); + } + mutex_unlock(&adev->userq_mutex); + return ret; +} + +int amdgpu_userq_start_sched_for_enforce_isolation(struct amdgpu_device *adev, + u32 idx) +{ + u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev); + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_mgr *uqm, *tmp; + int queue_id; + int ret = 0, r; + + /* only need to stop gfx/compute */ + if (!(ip_mask & ((1 << AMDGPU_HW_IP_GFX) | (1 << AMDGPU_HW_IP_COMPUTE)))) + return 0; + + mutex_lock(&adev->userq_mutex); + if (!adev->userq_halt_for_enforce_isolation) + dev_warn(adev->dev, "userq scheduling already started!\n"); + adev->userq_halt_for_enforce_isolation = false; + list_for_each_entry_safe(uqm, tmp, &adev->userq_mgr_list, list) { + mutex_lock(&uqm->userq_mutex); + idr_for_each_entry(&uqm->userq_idr, queue, queue_id) { + if (((queue->queue_type == AMDGPU_HW_IP_GFX) || + (queue->queue_type == AMDGPU_HW_IP_COMPUTE)) && + (queue->xcp_id == idx)) { + r = amdgpu_userq_map_helper(uqm, queue); + if (r) + ret = r; + } + } + mutex_unlock(&uqm->userq_mutex); + } + mutex_unlock(&adev->userq_mutex); + return ret; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h new file mode 100644 index 000000000000..ec040c2fd6c9 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h @@ -0,0 +1,135 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef AMDGPU_USERQ_H_ +#define AMDGPU_USERQ_H_ +#include "amdgpu_eviction_fence.h" + +#define AMDGPU_MAX_USERQ_COUNT 512 + +#define to_ev_fence(f) container_of(f, struct amdgpu_eviction_fence, base) +#define uq_mgr_to_fpriv(u) container_of(u, struct amdgpu_fpriv, userq_mgr) +#define work_to_uq_mgr(w, name) container_of(w, struct amdgpu_userq_mgr, name) + +enum amdgpu_userq_state { + AMDGPU_USERQ_STATE_UNMAPPED = 0, + AMDGPU_USERQ_STATE_MAPPED, + AMDGPU_USERQ_STATE_PREEMPTED, + AMDGPU_USERQ_STATE_HUNG, +}; + +struct amdgpu_mqd_prop; + +struct amdgpu_userq_obj { + void *cpu_ptr; + uint64_t gpu_addr; + struct amdgpu_bo *obj; +}; + +struct amdgpu_usermode_queue { + int queue_type; + enum amdgpu_userq_state state; + uint64_t doorbell_handle; + uint64_t doorbell_index; + uint64_t flags; + struct amdgpu_mqd_prop *userq_prop; + struct amdgpu_userq_mgr *userq_mgr; + struct amdgpu_vm *vm; + struct amdgpu_userq_obj mqd; + struct amdgpu_userq_obj db_obj; + struct amdgpu_userq_obj fw_obj; + struct amdgpu_userq_obj wptr_obj; + struct xarray fence_drv_xa; + struct amdgpu_userq_fence_driver *fence_drv; + struct dma_fence *last_fence; + u32 xcp_id; + int priority; +}; + +struct amdgpu_userq_funcs { + int (*mqd_create)(struct amdgpu_userq_mgr *uq_mgr, + struct drm_amdgpu_userq_in *args, + struct amdgpu_usermode_queue *queue); + void (*mqd_destroy)(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *uq); + int (*unmap)(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue); + int (*map)(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue); +}; + +/* Usermode queues for gfx */ +struct amdgpu_userq_mgr { + struct idr userq_idr; + struct mutex userq_mutex; + struct amdgpu_device *adev; + struct delayed_work resume_work; + struct list_head list; + struct drm_file *file; +}; + +struct amdgpu_db_info { + uint64_t doorbell_handle; + uint32_t queue_type; + uint32_t doorbell_offset; + struct amdgpu_userq_obj *db_obj; +}; + +int amdgpu_userq_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); + +int amdgpu_userq_mgr_init(struct amdgpu_userq_mgr *userq_mgr, struct drm_file *file_priv, + struct amdgpu_device *adev); + +void amdgpu_userq_mgr_fini(struct amdgpu_userq_mgr *userq_mgr); + +int amdgpu_userq_create_object(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_userq_obj *userq_obj, + int size); + +void amdgpu_userq_destroy_object(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_userq_obj *userq_obj); + +void amdgpu_userq_evict(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_eviction_fence *ev_fence); + +int amdgpu_userq_active(struct amdgpu_userq_mgr *uq_mgr); + +void amdgpu_userq_ensure_ev_fence(struct amdgpu_userq_mgr *userq_mgr, + struct amdgpu_eviction_fence_mgr *evf_mgr); + +uint64_t amdgpu_userq_get_doorbell_index(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_db_info *db_info, + struct drm_file *filp); + +u32 amdgpu_userq_get_supported_ip_mask(struct amdgpu_device *adev); + +int amdgpu_userq_suspend(struct amdgpu_device *adev); +int amdgpu_userq_resume(struct amdgpu_device *adev); + +int amdgpu_userq_stop_sched_for_enforce_isolation(struct amdgpu_device *adev, + u32 idx); +int amdgpu_userq_start_sched_for_enforce_isolation(struct amdgpu_device *adev, + u32 idx); + +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c new file mode 100644 index 000000000000..fc4d0d42e223 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c @@ -0,0 +1,968 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include <linux/kref.h> +#include <linux/slab.h> +#include <linux/dma-fence-unwrap.h> + +#include <drm/drm_exec.h> +#include <drm/drm_syncobj.h> + +#include "amdgpu.h" +#include "amdgpu_userq_fence.h" + +static const struct dma_fence_ops amdgpu_userq_fence_ops; +static struct kmem_cache *amdgpu_userq_fence_slab; + +int amdgpu_userq_fence_slab_init(void) +{ + amdgpu_userq_fence_slab = kmem_cache_create("amdgpu_userq_fence", + sizeof(struct amdgpu_userq_fence), + 0, + SLAB_HWCACHE_ALIGN, + NULL); + if (!amdgpu_userq_fence_slab) + return -ENOMEM; + + return 0; +} + +void amdgpu_userq_fence_slab_fini(void) +{ + rcu_barrier(); + kmem_cache_destroy(amdgpu_userq_fence_slab); +} + +static inline struct amdgpu_userq_fence *to_amdgpu_userq_fence(struct dma_fence *f) +{ + if (!f || f->ops != &amdgpu_userq_fence_ops) + return NULL; + + return container_of(f, struct amdgpu_userq_fence, base); +} + +static u64 amdgpu_userq_fence_read(struct amdgpu_userq_fence_driver *fence_drv) +{ + return le64_to_cpu(*fence_drv->cpu_addr); +} + +int amdgpu_userq_fence_driver_alloc(struct amdgpu_device *adev, + struct amdgpu_usermode_queue *userq) +{ + struct amdgpu_userq_fence_driver *fence_drv; + unsigned long flags; + int r; + + fence_drv = kzalloc(sizeof(*fence_drv), GFP_KERNEL); + if (!fence_drv) + return -ENOMEM; + + /* Acquire seq64 memory */ + r = amdgpu_seq64_alloc(adev, &fence_drv->va, &fence_drv->gpu_addr, + &fence_drv->cpu_addr); + if (r) + goto free_fence_drv; + + memset(fence_drv->cpu_addr, 0, sizeof(u64)); + + kref_init(&fence_drv->refcount); + INIT_LIST_HEAD(&fence_drv->fences); + spin_lock_init(&fence_drv->fence_list_lock); + + fence_drv->adev = adev; + fence_drv->context = dma_fence_context_alloc(1); + get_task_comm(fence_drv->timeline_name, current); + + xa_lock_irqsave(&adev->userq_xa, flags); + r = xa_err(__xa_store(&adev->userq_xa, userq->doorbell_index, + fence_drv, GFP_KERNEL)); + xa_unlock_irqrestore(&adev->userq_xa, flags); + if (r) + goto free_seq64; + + userq->fence_drv = fence_drv; + + return 0; + +free_seq64: + amdgpu_seq64_free(adev, fence_drv->va); +free_fence_drv: + kfree(fence_drv); + + return r; +} + +static void amdgpu_userq_walk_and_drop_fence_drv(struct xarray *xa) +{ + struct amdgpu_userq_fence_driver *fence_drv; + unsigned long index; + + if (xa_empty(xa)) + return; + + xa_lock(xa); + xa_for_each(xa, index, fence_drv) { + __xa_erase(xa, index); + amdgpu_userq_fence_driver_put(fence_drv); + } + + xa_unlock(xa); +} + +void +amdgpu_userq_fence_driver_free(struct amdgpu_usermode_queue *userq) +{ + amdgpu_userq_walk_and_drop_fence_drv(&userq->fence_drv_xa); + xa_destroy(&userq->fence_drv_xa); + /* Drop the fence_drv reference held by user queue */ + amdgpu_userq_fence_driver_put(userq->fence_drv); +} + +void amdgpu_userq_fence_driver_process(struct amdgpu_userq_fence_driver *fence_drv) +{ + struct amdgpu_userq_fence *userq_fence, *tmp; + struct dma_fence *fence; + u64 rptr; + int i; + + if (!fence_drv) + return; + + rptr = amdgpu_userq_fence_read(fence_drv); + + spin_lock(&fence_drv->fence_list_lock); + list_for_each_entry_safe(userq_fence, tmp, &fence_drv->fences, link) { + fence = &userq_fence->base; + + if (rptr < fence->seqno) + break; + + dma_fence_signal(fence); + + for (i = 0; i < userq_fence->fence_drv_array_count; i++) + amdgpu_userq_fence_driver_put(userq_fence->fence_drv_array[i]); + + list_del(&userq_fence->link); + dma_fence_put(fence); + } + spin_unlock(&fence_drv->fence_list_lock); +} + +void amdgpu_userq_fence_driver_destroy(struct kref *ref) +{ + struct amdgpu_userq_fence_driver *fence_drv = container_of(ref, + struct amdgpu_userq_fence_driver, + refcount); + struct amdgpu_userq_fence_driver *xa_fence_drv; + struct amdgpu_device *adev = fence_drv->adev; + struct amdgpu_userq_fence *fence, *tmp; + struct xarray *xa = &adev->userq_xa; + unsigned long index, flags; + struct dma_fence *f; + + spin_lock_irqsave(&fence_drv->fence_list_lock, flags); + list_for_each_entry_safe(fence, tmp, &fence_drv->fences, link) { + f = &fence->base; + + if (!dma_fence_is_signaled(f)) { + dma_fence_set_error(f, -ECANCELED); + dma_fence_signal(f); + } + + list_del(&fence->link); + dma_fence_put(f); + } + spin_unlock_irqrestore(&fence_drv->fence_list_lock, flags); + + xa_lock_irqsave(xa, flags); + xa_for_each(xa, index, xa_fence_drv) + if (xa_fence_drv == fence_drv) + __xa_erase(xa, index); + xa_unlock_irqrestore(xa, flags); + + /* Free seq64 memory */ + amdgpu_seq64_free(adev, fence_drv->va); + kfree(fence_drv); +} + +void amdgpu_userq_fence_driver_get(struct amdgpu_userq_fence_driver *fence_drv) +{ + kref_get(&fence_drv->refcount); +} + +void amdgpu_userq_fence_driver_put(struct amdgpu_userq_fence_driver *fence_drv) +{ + kref_put(&fence_drv->refcount, amdgpu_userq_fence_driver_destroy); +} + +static int amdgpu_userq_fence_alloc(struct amdgpu_userq_fence **userq_fence) +{ + *userq_fence = kmem_cache_alloc(amdgpu_userq_fence_slab, GFP_ATOMIC); + return *userq_fence ? 0 : -ENOMEM; +} + +static int amdgpu_userq_fence_create(struct amdgpu_usermode_queue *userq, + struct amdgpu_userq_fence *userq_fence, + u64 seq, struct dma_fence **f) +{ + struct amdgpu_userq_fence_driver *fence_drv; + struct dma_fence *fence; + unsigned long flags; + + fence_drv = userq->fence_drv; + if (!fence_drv) + return -EINVAL; + + spin_lock_init(&userq_fence->lock); + INIT_LIST_HEAD(&userq_fence->link); + fence = &userq_fence->base; + userq_fence->fence_drv = fence_drv; + + dma_fence_init(fence, &amdgpu_userq_fence_ops, &userq_fence->lock, + fence_drv->context, seq); + + amdgpu_userq_fence_driver_get(fence_drv); + dma_fence_get(fence); + + if (!xa_empty(&userq->fence_drv_xa)) { + struct amdgpu_userq_fence_driver *stored_fence_drv; + unsigned long index, count = 0; + int i = 0; + + xa_lock(&userq->fence_drv_xa); + xa_for_each(&userq->fence_drv_xa, index, stored_fence_drv) + count++; + + userq_fence->fence_drv_array = + kvmalloc_array(count, + sizeof(struct amdgpu_userq_fence_driver *), + GFP_ATOMIC); + + if (userq_fence->fence_drv_array) { + xa_for_each(&userq->fence_drv_xa, index, stored_fence_drv) { + userq_fence->fence_drv_array[i] = stored_fence_drv; + __xa_erase(&userq->fence_drv_xa, index); + i++; + } + } + + userq_fence->fence_drv_array_count = i; + xa_unlock(&userq->fence_drv_xa); + } else { + userq_fence->fence_drv_array = NULL; + userq_fence->fence_drv_array_count = 0; + } + + /* Check if hardware has already processed the job */ + spin_lock_irqsave(&fence_drv->fence_list_lock, flags); + if (!dma_fence_is_signaled_locked(fence)) + list_add_tail(&userq_fence->link, &fence_drv->fences); + else + dma_fence_put(fence); + + spin_unlock_irqrestore(&fence_drv->fence_list_lock, flags); + + *f = fence; + + return 0; +} + +static const char *amdgpu_userq_fence_get_driver_name(struct dma_fence *f) +{ + return "amdgpu_userq_fence"; +} + +static const char *amdgpu_userq_fence_get_timeline_name(struct dma_fence *f) +{ + struct amdgpu_userq_fence *fence = to_amdgpu_userq_fence(f); + + return fence->fence_drv->timeline_name; +} + +static bool amdgpu_userq_fence_signaled(struct dma_fence *f) +{ + struct amdgpu_userq_fence *fence = to_amdgpu_userq_fence(f); + struct amdgpu_userq_fence_driver *fence_drv = fence->fence_drv; + u64 rptr, wptr; + + rptr = amdgpu_userq_fence_read(fence_drv); + wptr = fence->base.seqno; + + if (rptr >= wptr) + return true; + + return false; +} + +static void amdgpu_userq_fence_free(struct rcu_head *rcu) +{ + struct dma_fence *fence = container_of(rcu, struct dma_fence, rcu); + struct amdgpu_userq_fence *userq_fence = to_amdgpu_userq_fence(fence); + struct amdgpu_userq_fence_driver *fence_drv = userq_fence->fence_drv; + + /* Release the fence driver reference */ + amdgpu_userq_fence_driver_put(fence_drv); + + kvfree(userq_fence->fence_drv_array); + kmem_cache_free(amdgpu_userq_fence_slab, userq_fence); +} + +static void amdgpu_userq_fence_release(struct dma_fence *f) +{ + call_rcu(&f->rcu, amdgpu_userq_fence_free); +} + +static const struct dma_fence_ops amdgpu_userq_fence_ops = { + .use_64bit_seqno = true, + .get_driver_name = amdgpu_userq_fence_get_driver_name, + .get_timeline_name = amdgpu_userq_fence_get_timeline_name, + .signaled = amdgpu_userq_fence_signaled, + .release = amdgpu_userq_fence_release, +}; + +/** + * amdgpu_userq_fence_read_wptr - Read the userq wptr value + * + * @queue: user mode queue structure pointer + * @wptr: write pointer value + * + * Read the wptr value from userq's MQD. The userq signal IOCTL + * creates a dma_fence for the shared buffers that expects the + * RPTR value written to seq64 memory >= WPTR. + * + * Returns wptr value on success, error on failure. + */ +static int amdgpu_userq_fence_read_wptr(struct amdgpu_usermode_queue *queue, + u64 *wptr) +{ + struct amdgpu_bo_va_mapping *mapping; + struct amdgpu_bo *bo; + u64 addr, *ptr; + int r; + + r = amdgpu_bo_reserve(queue->vm->root.bo, false); + if (r) + return r; + + addr = queue->userq_prop->wptr_gpu_addr; + addr &= AMDGPU_GMC_HOLE_MASK; + + mapping = amdgpu_vm_bo_lookup_mapping(queue->vm, addr >> PAGE_SHIFT); + if (!mapping) { + amdgpu_bo_unreserve(queue->vm->root.bo); + DRM_ERROR("Failed to lookup amdgpu_bo_va_mapping\n"); + return -EINVAL; + } + + bo = amdgpu_bo_ref(mapping->bo_va->base.bo); + amdgpu_bo_unreserve(queue->vm->root.bo); + r = amdgpu_bo_reserve(bo, true); + if (r) { + DRM_ERROR("Failed to reserve userqueue wptr bo"); + return r; + } + + r = amdgpu_bo_kmap(bo, (void **)&ptr); + if (r) { + DRM_ERROR("Failed mapping the userqueue wptr bo"); + goto map_error; + } + + *wptr = le64_to_cpu(*ptr); + + amdgpu_bo_kunmap(bo); + amdgpu_bo_unreserve(bo); + amdgpu_bo_unref(&bo); + + return 0; + +map_error: + amdgpu_bo_unreserve(bo); + amdgpu_bo_unref(&bo); + + return r; +} + +static void amdgpu_userq_fence_cleanup(struct dma_fence *fence) +{ + dma_fence_put(fence); +} + +int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp) +{ + struct amdgpu_fpriv *fpriv = filp->driver_priv; + struct amdgpu_userq_mgr *userq_mgr = &fpriv->userq_mgr; + struct drm_amdgpu_userq_signal *args = data; + struct drm_gem_object **gobj_write = NULL; + struct drm_gem_object **gobj_read = NULL; + struct amdgpu_usermode_queue *queue; + struct amdgpu_userq_fence *userq_fence; + struct drm_syncobj **syncobj = NULL; + u32 *bo_handles_write, num_write_bo_handles; + u32 *syncobj_handles, num_syncobj_handles; + u32 *bo_handles_read, num_read_bo_handles; + int r, i, entry, rentry, wentry; + struct dma_fence *fence; + struct drm_exec exec; + u64 wptr; + + num_syncobj_handles = args->num_syncobj_handles; + syncobj_handles = memdup_user(u64_to_user_ptr(args->syncobj_handles), + sizeof(u32) * num_syncobj_handles); + if (IS_ERR(syncobj_handles)) + return PTR_ERR(syncobj_handles); + + /* Array of pointers to the looked up syncobjs */ + syncobj = kmalloc_array(num_syncobj_handles, sizeof(*syncobj), GFP_KERNEL); + if (!syncobj) { + r = -ENOMEM; + goto free_syncobj_handles; + } + + for (entry = 0; entry < num_syncobj_handles; entry++) { + syncobj[entry] = drm_syncobj_find(filp, syncobj_handles[entry]); + if (!syncobj[entry]) { + r = -ENOENT; + goto free_syncobj; + } + } + + num_read_bo_handles = args->num_bo_read_handles; + bo_handles_read = memdup_user(u64_to_user_ptr(args->bo_read_handles), + sizeof(u32) * num_read_bo_handles); + if (IS_ERR(bo_handles_read)) { + r = PTR_ERR(bo_handles_read); + goto free_syncobj; + } + + /* Array of pointers to the GEM read objects */ + gobj_read = kmalloc_array(num_read_bo_handles, sizeof(*gobj_read), GFP_KERNEL); + if (!gobj_read) { + r = -ENOMEM; + goto free_bo_handles_read; + } + + for (rentry = 0; rentry < num_read_bo_handles; rentry++) { + gobj_read[rentry] = drm_gem_object_lookup(filp, bo_handles_read[rentry]); + if (!gobj_read[rentry]) { + r = -ENOENT; + goto put_gobj_read; + } + } + + num_write_bo_handles = args->num_bo_write_handles; + bo_handles_write = memdup_user(u64_to_user_ptr(args->bo_write_handles), + sizeof(u32) * num_write_bo_handles); + if (IS_ERR(bo_handles_write)) { + r = PTR_ERR(bo_handles_write); + goto put_gobj_read; + } + + /* Array of pointers to the GEM write objects */ + gobj_write = kmalloc_array(num_write_bo_handles, sizeof(*gobj_write), GFP_KERNEL); + if (!gobj_write) { + r = -ENOMEM; + goto free_bo_handles_write; + } + + for (wentry = 0; wentry < num_write_bo_handles; wentry++) { + gobj_write[wentry] = drm_gem_object_lookup(filp, bo_handles_write[wentry]); + if (!gobj_write[wentry]) { + r = -ENOENT; + goto put_gobj_write; + } + } + + /* Retrieve the user queue */ + queue = idr_find(&userq_mgr->userq_idr, args->queue_id); + if (!queue) { + r = -ENOENT; + goto put_gobj_write; + } + + r = amdgpu_userq_fence_read_wptr(queue, &wptr); + if (r) + goto put_gobj_write; + + r = amdgpu_userq_fence_alloc(&userq_fence); + if (r) + goto put_gobj_write; + + /* We are here means UQ is active, make sure the eviction fence is valid */ + amdgpu_userq_ensure_ev_fence(&fpriv->userq_mgr, &fpriv->evf_mgr); + + /* Create a new fence */ + r = amdgpu_userq_fence_create(queue, userq_fence, wptr, &fence); + if (r) { + mutex_unlock(&userq_mgr->userq_mutex); + kmem_cache_free(amdgpu_userq_fence_slab, userq_fence); + goto put_gobj_write; + } + + dma_fence_put(queue->last_fence); + queue->last_fence = dma_fence_get(fence); + mutex_unlock(&userq_mgr->userq_mutex); + + drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT, + (num_read_bo_handles + num_write_bo_handles)); + + /* Lock all BOs with retry handling */ + drm_exec_until_all_locked(&exec) { + r = drm_exec_prepare_array(&exec, gobj_read, num_read_bo_handles, 1); + drm_exec_retry_on_contention(&exec); + if (r) { + amdgpu_userq_fence_cleanup(fence); + goto exec_fini; + } + + r = drm_exec_prepare_array(&exec, gobj_write, num_write_bo_handles, 1); + drm_exec_retry_on_contention(&exec); + if (r) { + amdgpu_userq_fence_cleanup(fence); + goto exec_fini; + } + } + + for (i = 0; i < num_read_bo_handles; i++) { + if (!gobj_read || !gobj_read[i]->resv) + continue; + + dma_resv_add_fence(gobj_read[i]->resv, fence, + DMA_RESV_USAGE_READ); + } + + for (i = 0; i < num_write_bo_handles; i++) { + if (!gobj_write || !gobj_write[i]->resv) + continue; + + dma_resv_add_fence(gobj_write[i]->resv, fence, + DMA_RESV_USAGE_WRITE); + } + + /* Add the created fence to syncobj/BO's */ + for (i = 0; i < num_syncobj_handles; i++) + drm_syncobj_replace_fence(syncobj[i], fence); + + /* drop the reference acquired in fence creation function */ + dma_fence_put(fence); + +exec_fini: + drm_exec_fini(&exec); +put_gobj_write: + while (wentry-- > 0) + drm_gem_object_put(gobj_write[wentry]); + kfree(gobj_write); +free_bo_handles_write: + kfree(bo_handles_write); +put_gobj_read: + while (rentry-- > 0) + drm_gem_object_put(gobj_read[rentry]); + kfree(gobj_read); +free_bo_handles_read: + kfree(bo_handles_read); +free_syncobj: + while (entry-- > 0) + if (syncobj[entry]) + drm_syncobj_put(syncobj[entry]); + kfree(syncobj); +free_syncobj_handles: + kfree(syncobj_handles); + + return r; +} + +int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp) +{ + u32 *syncobj_handles, *timeline_points, *timeline_handles, *bo_handles_read, *bo_handles_write; + u32 num_syncobj, num_read_bo_handles, num_write_bo_handles; + struct drm_amdgpu_userq_fence_info *fence_info = NULL; + struct drm_amdgpu_userq_wait *wait_info = data; + struct amdgpu_fpriv *fpriv = filp->driver_priv; + struct amdgpu_userq_mgr *userq_mgr = &fpriv->userq_mgr; + struct amdgpu_usermode_queue *waitq; + struct drm_gem_object **gobj_write; + struct drm_gem_object **gobj_read; + struct dma_fence **fences = NULL; + u16 num_points, num_fences = 0; + int r, i, rentry, wentry, cnt; + struct drm_exec exec; + + num_read_bo_handles = wait_info->num_bo_read_handles; + bo_handles_read = memdup_user(u64_to_user_ptr(wait_info->bo_read_handles), + sizeof(u32) * num_read_bo_handles); + if (IS_ERR(bo_handles_read)) + return PTR_ERR(bo_handles_read); + + num_write_bo_handles = wait_info->num_bo_write_handles; + bo_handles_write = memdup_user(u64_to_user_ptr(wait_info->bo_write_handles), + sizeof(u32) * num_write_bo_handles); + if (IS_ERR(bo_handles_write)) { + r = PTR_ERR(bo_handles_write); + goto free_bo_handles_read; + } + + num_syncobj = wait_info->num_syncobj_handles; + syncobj_handles = memdup_user(u64_to_user_ptr(wait_info->syncobj_handles), + sizeof(u32) * num_syncobj); + if (IS_ERR(syncobj_handles)) { + r = PTR_ERR(syncobj_handles); + goto free_bo_handles_write; + } + + num_points = wait_info->num_syncobj_timeline_handles; + timeline_handles = memdup_user(u64_to_user_ptr(wait_info->syncobj_timeline_handles), + sizeof(u32) * num_points); + if (IS_ERR(timeline_handles)) { + r = PTR_ERR(timeline_handles); + goto free_syncobj_handles; + } + + timeline_points = memdup_user(u64_to_user_ptr(wait_info->syncobj_timeline_points), + sizeof(u32) * num_points); + if (IS_ERR(timeline_points)) { + r = PTR_ERR(timeline_points); + goto free_timeline_handles; + } + + gobj_read = kmalloc_array(num_read_bo_handles, sizeof(*gobj_read), GFP_KERNEL); + if (!gobj_read) { + r = -ENOMEM; + goto free_timeline_points; + } + + for (rentry = 0; rentry < num_read_bo_handles; rentry++) { + gobj_read[rentry] = drm_gem_object_lookup(filp, bo_handles_read[rentry]); + if (!gobj_read[rentry]) { + r = -ENOENT; + goto put_gobj_read; + } + } + + gobj_write = kmalloc_array(num_write_bo_handles, sizeof(*gobj_write), GFP_KERNEL); + if (!gobj_write) { + r = -ENOMEM; + goto put_gobj_read; + } + + for (wentry = 0; wentry < num_write_bo_handles; wentry++) { + gobj_write[wentry] = drm_gem_object_lookup(filp, bo_handles_write[wentry]); + if (!gobj_write[wentry]) { + r = -ENOENT; + goto put_gobj_write; + } + } + + drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT, + (num_read_bo_handles + num_write_bo_handles)); + + /* Lock all BOs with retry handling */ + drm_exec_until_all_locked(&exec) { + r = drm_exec_prepare_array(&exec, gobj_read, num_read_bo_handles, 1); + drm_exec_retry_on_contention(&exec); + if (r) { + drm_exec_fini(&exec); + goto put_gobj_write; + } + + r = drm_exec_prepare_array(&exec, gobj_write, num_write_bo_handles, 1); + drm_exec_retry_on_contention(&exec); + if (r) { + drm_exec_fini(&exec); + goto put_gobj_write; + } + } + + if (!wait_info->num_fences) { + if (num_points) { + struct dma_fence_unwrap iter; + struct dma_fence *fence; + struct dma_fence *f; + + for (i = 0; i < num_points; i++) { + r = drm_syncobj_find_fence(filp, timeline_handles[i], + timeline_points[i], + DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT, + &fence); + if (r) + goto exec_fini; + + dma_fence_unwrap_for_each(f, &iter, fence) + num_fences++; + + dma_fence_put(fence); + } + } + + /* Count syncobj's fence */ + for (i = 0; i < num_syncobj; i++) { + struct dma_fence *fence; + + r = drm_syncobj_find_fence(filp, syncobj_handles[i], + 0, + DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT, + &fence); + if (r) + goto exec_fini; + + num_fences++; + dma_fence_put(fence); + } + + /* Count GEM objects fence */ + for (i = 0; i < num_read_bo_handles; i++) { + struct dma_resv_iter resv_cursor; + struct dma_fence *fence; + + dma_resv_for_each_fence(&resv_cursor, gobj_read[i]->resv, + DMA_RESV_USAGE_READ, fence) + num_fences++; + } + + for (i = 0; i < num_write_bo_handles; i++) { + struct dma_resv_iter resv_cursor; + struct dma_fence *fence; + + dma_resv_for_each_fence(&resv_cursor, gobj_write[i]->resv, + DMA_RESV_USAGE_WRITE, fence) + num_fences++; + } + + /* + * Passing num_fences = 0 means that userspace doesn't want to + * retrieve userq_fence_info. If num_fences = 0 we skip filling + * userq_fence_info and return the actual number of fences on + * args->num_fences. + */ + wait_info->num_fences = num_fences; + } else { + /* Array of fence info */ + fence_info = kmalloc_array(wait_info->num_fences, sizeof(*fence_info), GFP_KERNEL); + if (!fence_info) { + r = -ENOMEM; + goto exec_fini; + } + + /* Array of fences */ + fences = kmalloc_array(wait_info->num_fences, sizeof(*fences), GFP_KERNEL); + if (!fences) { + r = -ENOMEM; + goto free_fence_info; + } + + /* Retrieve GEM read objects fence */ + for (i = 0; i < num_read_bo_handles; i++) { + struct dma_resv_iter resv_cursor; + struct dma_fence *fence; + + dma_resv_for_each_fence(&resv_cursor, gobj_read[i]->resv, + DMA_RESV_USAGE_READ, fence) { + if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) { + r = -EINVAL; + goto free_fences; + } + + fences[num_fences++] = fence; + dma_fence_get(fence); + } + } + + /* Retrieve GEM write objects fence */ + for (i = 0; i < num_write_bo_handles; i++) { + struct dma_resv_iter resv_cursor; + struct dma_fence *fence; + + dma_resv_for_each_fence(&resv_cursor, gobj_write[i]->resv, + DMA_RESV_USAGE_WRITE, fence) { + if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) { + r = -EINVAL; + goto free_fences; + } + + fences[num_fences++] = fence; + dma_fence_get(fence); + } + } + + if (num_points) { + struct dma_fence_unwrap iter; + struct dma_fence *fence; + struct dma_fence *f; + + for (i = 0; i < num_points; i++) { + r = drm_syncobj_find_fence(filp, timeline_handles[i], + timeline_points[i], + DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT, + &fence); + if (r) + goto free_fences; + + dma_fence_unwrap_for_each(f, &iter, fence) { + if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) { + r = -EINVAL; + goto free_fences; + } + + dma_fence_get(f); + fences[num_fences++] = f; + } + + dma_fence_put(fence); + } + } + + /* Retrieve syncobj's fence */ + for (i = 0; i < num_syncobj; i++) { + struct dma_fence *fence; + + r = drm_syncobj_find_fence(filp, syncobj_handles[i], + 0, + DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT, + &fence); + if (r) + goto free_fences; + + if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) { + r = -EINVAL; + goto free_fences; + } + + fences[num_fences++] = fence; + } + + /* + * Keep only the latest fences to reduce the number of values + * given back to userspace. + */ + num_fences = dma_fence_dedup_array(fences, num_fences); + + waitq = idr_find(&userq_mgr->userq_idr, wait_info->waitq_id); + if (!waitq) { + r = -EINVAL; + goto free_fences; + } + + for (i = 0, cnt = 0; i < num_fences; i++) { + struct amdgpu_userq_fence_driver *fence_drv; + struct amdgpu_userq_fence *userq_fence; + u32 index; + + userq_fence = to_amdgpu_userq_fence(fences[i]); + if (!userq_fence) { + /* + * Just waiting on other driver fences should + * be good for now + */ + r = dma_fence_wait(fences[i], true); + if (r) { + dma_fence_put(fences[i]); + goto free_fences; + } + + dma_fence_put(fences[i]); + continue; + } + + fence_drv = userq_fence->fence_drv; + /* + * We need to make sure the user queue release their reference + * to the fence drivers at some point before queue destruction. + * Otherwise, we would gather those references until we don't + * have any more space left and crash. + */ + r = xa_alloc(&waitq->fence_drv_xa, &index, fence_drv, + xa_limit_32b, GFP_KERNEL); + if (r) + goto free_fences; + + amdgpu_userq_fence_driver_get(fence_drv); + + /* Store drm syncobj's gpu va address and value */ + fence_info[cnt].va = fence_drv->va; + fence_info[cnt].value = fences[i]->seqno; + + dma_fence_put(fences[i]); + /* Increment the actual userq fence count */ + cnt++; + } + + wait_info->num_fences = cnt; + /* Copy userq fence info to user space */ + if (copy_to_user(u64_to_user_ptr(wait_info->out_fences), + fence_info, wait_info->num_fences * sizeof(*fence_info))) { + r = -EFAULT; + goto free_fences; + } + + kfree(fences); + kfree(fence_info); + } + + drm_exec_fini(&exec); + for (i = 0; i < num_read_bo_handles; i++) + drm_gem_object_put(gobj_read[i]); + kfree(gobj_read); + + for (i = 0; i < num_write_bo_handles; i++) + drm_gem_object_put(gobj_write[i]); + kfree(gobj_write); + + kfree(timeline_points); + kfree(timeline_handles); + kfree(syncobj_handles); + kfree(bo_handles_write); + kfree(bo_handles_read); + + return 0; + +free_fences: + while (num_fences-- > 0) + dma_fence_put(fences[num_fences]); + kfree(fences); +free_fence_info: + kfree(fence_info); +exec_fini: + drm_exec_fini(&exec); +put_gobj_write: + while (wentry-- > 0) + drm_gem_object_put(gobj_write[wentry]); + kfree(gobj_write); +put_gobj_read: + while (rentry-- > 0) + drm_gem_object_put(gobj_read[rentry]); + kfree(gobj_read); +free_timeline_points: + kfree(timeline_points); +free_timeline_handles: + kfree(timeline_handles); +free_syncobj_handles: + kfree(syncobj_handles); +free_bo_handles_write: + kfree(bo_handles_write); +free_bo_handles_read: + kfree(bo_handles_read); + + return r; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h new file mode 100644 index 000000000000..97a125ab8a78 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h @@ -0,0 +1,76 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __AMDGPU_USERQ_FENCE_H__ +#define __AMDGPU_USERQ_FENCE_H__ + +#include <linux/types.h> + +#include "amdgpu_userq.h" + +struct amdgpu_userq_fence { + struct dma_fence base; + /* + * This lock is necessary to synchronize the + * userqueue dma fence operations. + */ + spinlock_t lock; + struct list_head link; + unsigned long fence_drv_array_count; + struct amdgpu_userq_fence_driver *fence_drv; + struct amdgpu_userq_fence_driver **fence_drv_array; +}; + +struct amdgpu_userq_fence_driver { + struct kref refcount; + u64 va; + u64 gpu_addr; + u64 *cpu_addr; + u64 context; + /* + * This lock is necesaary to synchronize the access + * to the fences list by the fence driver. + */ + spinlock_t fence_list_lock; + struct list_head fences; + struct amdgpu_device *adev; + char timeline_name[TASK_COMM_LEN]; +}; + +int amdgpu_userq_fence_slab_init(void); +void amdgpu_userq_fence_slab_fini(void); + +void amdgpu_userq_fence_driver_get(struct amdgpu_userq_fence_driver *fence_drv); +void amdgpu_userq_fence_driver_put(struct amdgpu_userq_fence_driver *fence_drv); +int amdgpu_userq_fence_driver_alloc(struct amdgpu_device *adev, + struct amdgpu_usermode_queue *userq); +void amdgpu_userq_fence_driver_free(struct amdgpu_usermode_queue *userq); +void amdgpu_userq_fence_driver_process(struct amdgpu_userq_fence_driver *fence_drv); +void amdgpu_userq_fence_driver_destroy(struct kref *ref); +int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp); +int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp); + +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 1991dd3d1056..c8885c3d54b3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -353,9 +353,9 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev, int i) cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); - /* err_event_athub will corrupt VCPU buffer, so we need to + /* err_event_athub and dpc recovery will corrupt VCPU buffer, so we need to * restore fw data and clear buffer in amdgpu_vcn_resume() */ - if (in_ras_intr) + if (in_ras_intr || adev->pcie_reset_ctx.in_link_reset) return 0; return amdgpu_vcn_save_vcpu_bo_inst(adev, i); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index cdcdae7f71ce..83adf81defc7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -66,7 +66,6 @@ #define VCN_ENC_CMD_REG_WAIT 0x0000000c #define VCN_AON_SOC_ADDRESS_2_0 0x1f800 -#define VCN1_AON_SOC_ADDRESS_3_0 0x48000 #define VCN_VID_IP_ADDRESS_2_0 0x0 #define VCN_AON_IP_ADDRESS_2_0 0x30000 diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 0bb8cbe0dcc0..13f0cdeb59c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -1323,6 +1323,9 @@ static int amdgpu_virt_req_ras_err_count_internal(struct amdgpu_device *adev, bo { struct amdgpu_virt *virt = &adev->virt; + if (!virt->ops || !virt->ops->req_ras_err_count) + return -EOPNOTSUPP; + /* Host allows 15 ras telemetry requests per 60 seconds. Afterwhich, the Host * will ignore incoming guest messages. Ratelimit the guest messages to * prevent guest self DOS. @@ -1378,14 +1381,16 @@ amdgpu_virt_write_cpers_to_ring(struct amdgpu_device *adev, used_size = host_telemetry->header.used_size; if (used_size > (AMD_SRIOV_RAS_TELEMETRY_SIZE_KB << 10)) - return 0; + return -EINVAL; cper_dump = kmemdup(&host_telemetry->body.cper_dump, used_size, GFP_KERNEL); if (!cper_dump) return -ENOMEM; - if (checksum != amd_sriov_msg_checksum(cper_dump, used_size, 0, 0)) + if (checksum != amd_sriov_msg_checksum(cper_dump, used_size, 0, 0)) { + ret = -EINVAL; goto out; + } *more = cper_dump->more; @@ -1425,7 +1430,7 @@ static int amdgpu_virt_req_ras_cper_dump_internal(struct amdgpu_device *adev) int ret = 0; uint32_t more = 0; - if (!amdgpu_sriov_ras_cper_en(adev)) + if (!virt->ops || !virt->ops->req_ras_cper_dump) return -EOPNOTSUPP; do { @@ -1434,7 +1439,7 @@ static int amdgpu_virt_req_ras_cper_dump_internal(struct amdgpu_device *adev) adev, virt->fw_reserve.ras_telemetry, &more); else ret = 0; - } while (more); + } while (more && !ret); return ret; } @@ -1444,6 +1449,9 @@ int amdgpu_virt_req_ras_cper_dump(struct amdgpu_device *adev, bool force_update) struct amdgpu_virt *virt = &adev->virt; int ret = 0; + if (!amdgpu_sriov_ras_cper_en(adev)) + return -EOPNOTSUPP; + if ((__ratelimit(&virt->ras.ras_cper_dump_rs) || force_update) && down_read_trylock(&adev->reset_domain->sem)) { mutex_lock(&virt->ras.ras_telemetry_mutex); @@ -1480,3 +1488,16 @@ bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev, return true; } + +/* + * amdgpu_virt_request_bad_pages() - request bad pages + * @adev: amdgpu device. + * Send command to GPU hypervisor to write new bad pages into the shared PF2VF region + */ +void amdgpu_virt_request_bad_pages(struct amdgpu_device *adev) +{ + struct amdgpu_virt *virt = &adev->virt; + + if (virt->ops && virt->ops->req_bad_pages) + virt->ops->req_bad_pages(adev); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h index df03dba67ab8..577c6194db78 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h @@ -97,6 +97,7 @@ struct amdgpu_virt_ops { bool (*rcvd_ras_intr)(struct amdgpu_device *adev); int (*req_ras_err_count)(struct amdgpu_device *adev); int (*req_ras_cper_dump)(struct amdgpu_device *adev, u64 vf_rptr); + int (*req_bad_pages)(struct amdgpu_device *adev); }; /* @@ -146,11 +147,13 @@ enum AMDGIM_FEATURE_FLAG { enum AMDGIM_REG_ACCESS_FLAG { /* Use PSP to program IH_RB_CNTL */ - AMDGIM_FEATURE_IH_REG_PSP_EN = (1 << 0), + AMDGIM_FEATURE_IH_REG_PSP_EN = (1 << 0), /* Use RLC to program MMHUB regs */ - AMDGIM_FEATURE_MMHUB_REG_RLC_EN = (1 << 1), + AMDGIM_FEATURE_MMHUB_REG_RLC_EN = (1 << 1), /* Use RLC to program GC regs */ - AMDGIM_FEATURE_GC_REG_RLC_EN = (1 << 2), + AMDGIM_FEATURE_GC_REG_RLC_EN = (1 << 2), + /* Use PSP to program L1_TLB_CNTL*/ + AMDGIM_FEATURE_L1_TLB_CNTL_PSP_EN = (1 << 3), }; struct amdgim_pf2vf_info_v1 { @@ -260,7 +263,10 @@ struct amdgpu_virt { uint32_t reg_val_offs; struct amdgpu_irq_src ack_irq; struct amdgpu_irq_src rcv_irq; + struct work_struct flr_work; + struct work_struct bad_pages_work; + struct amdgpu_mm_table mm_table; const struct amdgpu_virt_ops *ops; struct amdgpu_vf_error_buffer vf_errors; @@ -330,6 +336,10 @@ struct amdgpu_video_codec_info; (amdgpu_sriov_vf((adev)) && \ ((adev)->virt.reg_access & (AMDGIM_FEATURE_GC_REG_RLC_EN))) +#define amdgpu_sriov_reg_indirect_l1_tlb_cntl(adev) \ +(amdgpu_sriov_vf((adev)) && \ + ((adev)->virt.reg_access & (AMDGIM_FEATURE_L1_TLB_CNTL_PSP_EN))) + #define amdgpu_sriov_rlcg_error_report_enabled(adev) \ (amdgpu_sriov_reg_indirect_mmhub(adev) || amdgpu_sriov_reg_indirect_gc(adev)) @@ -423,4 +433,5 @@ int amdgpu_virt_req_ras_cper_dump(struct amdgpu_device *adev, bool force_update) int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev); bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev, enum amdgpu_ras_block block); +void amdgpu_virt_request_bad_pages(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index ce52b4d75e94..3911c78f8282 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -787,7 +787,8 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, pasid_mapping_needed &= adev->gmc.gmc_funcs->emit_pasid_mapping && ring->funcs->emit_wreg; - cleaner_shader_needed = adev->gfx.enable_cleaner_shader && + cleaner_shader_needed = job->run_cleaner_shader && + adev->gfx.enable_cleaner_shader && ring->funcs->emit_cleaner_shader && job->base.s_fence && &job->base.s_fence->scheduled == isolation->spearhead; @@ -817,7 +818,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, if (spm_update_needed && adev->gfx.rlc.funcs->update_spm_vmid) adev->gfx.rlc.funcs->update_spm_vmid(adev, ring, job->vmid); - if (!ring->is_mes_queue && ring->funcs->emit_gds_switch && + if (ring->funcs->emit_gds_switch && gds_switch_needed) { amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base, job->gds_size, job->gws_base, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c index 23b6f7a4aa4a..b03c3895897b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c @@ -709,10 +709,10 @@ void amdgpu_xcp_cfg_sysfs_fini(struct amdgpu_device *adev) struct amdgpu_xcp_cfg *xcp_cfg; int i; - if (!adev->xcp_mgr) + if (!adev->xcp_mgr || !adev->xcp_mgr->xcp_cfg) return; - xcp_cfg = adev->xcp_mgr->xcp_cfg; + xcp_cfg = adev->xcp_mgr->xcp_cfg; for (i = 0; i < xcp_cfg->num_res; i++) { xcp_res = &xcp_cfg->xcp_res[i]; kobject_put(&xcp_res->kobj); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index 477424472bbe..f51ef4cf16e0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -296,15 +296,27 @@ static const struct amdgpu_pcs_ras_field xgmi3x16_pcs_ras_fields[] = { static u32 xgmi_v6_4_get_link_status(struct amdgpu_device *adev, int global_link_num) { - const u32 smnpcs_xgmi3x16_pcs_state_hist1 = 0x11a00070; - const int xgmi_inst = 2; - u32 link_inst; + const u32 smn_xgmi_6_4_pcs_state_hist1[2] = { 0x11a00070, 0x11b00070 }; + const u32 smn_xgmi_6_4_1_pcs_state_hist1[2] = { 0x12100070, + 0x11b00070 }; + u32 i, n; u64 addr; - link_inst = global_link_num % xgmi_inst; + switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) { + case IP_VERSION(6, 4, 0): + n = ARRAY_SIZE(smn_xgmi_6_4_pcs_state_hist1); + addr = smn_xgmi_6_4_pcs_state_hist1[global_link_num % n]; + break; + case IP_VERSION(6, 4, 1): + n = ARRAY_SIZE(smn_xgmi_6_4_1_pcs_state_hist1); + addr = smn_xgmi_6_4_1_pcs_state_hist1[global_link_num % n]; + break; + default: + return U32_MAX; + } - addr = (smnpcs_xgmi3x16_pcs_state_hist1 | (link_inst << 20)) + - adev->asic_funcs->encode_ext_smn_addressing(global_link_num / xgmi_inst); + i = global_link_num / n; + addr += adev->asic_funcs->encode_ext_smn_addressing(i); return RREG32_PCIE_EXT(addr); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h index d6ac2652f0ac..92ca13097aaa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h @@ -109,10 +109,11 @@ union amd_sriov_msg_feature_flags { union amd_sriov_reg_access_flags { struct { - uint32_t vf_reg_access_ih : 1; - uint32_t vf_reg_access_mmhub : 1; - uint32_t vf_reg_access_gc : 1; - uint32_t reserved : 29; + uint32_t vf_reg_access_ih : 1; + uint32_t vf_reg_access_mmhub : 1; + uint32_t vf_reg_access_gc : 1; + uint32_t vf_reg_access_l1_tlb_cntl : 1; + uint32_t reserved : 28; } flags; uint32_t all; }; @@ -330,6 +331,7 @@ enum amd_sriov_mailbox_request_message { MB_REQ_MSG_RAS_POISON = 202, MB_REQ_RAS_ERROR_COUNT = 203, MB_REQ_RAS_CPER_DUMP = 204, + MB_REQ_RAS_BAD_PAGES = 205, }; /* mailbox message send from host to guest */ @@ -347,6 +349,9 @@ enum amd_sriov_mailbox_response_message { MB_RES_MSG_GPU_RMA = 10, MB_RES_MSG_RAS_ERROR_COUNT_READY = 11, MB_REQ_RAS_CPER_DUMP_READY = 14, + MB_RES_MSG_RAS_BAD_PAGES_READY = 15, + MB_RES_MSG_RAS_BAD_PAGES_NOTIFICATION = 16, + MB_RES_MSG_UNRECOV_ERR_NOTIFICATION = 17, MB_RES_MSG_TEXT_MESSAGE = 255 }; diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c index ae071985f26e..1c083304ae77 100644 --- a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c +++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c @@ -448,53 +448,71 @@ static int __aqua_vanjaram_get_xcp_ip_info(struct amdgpu_xcp_mgr *xcp_mgr, int x return 0; } -static int aqua_vanjaram_get_xcp_res_info(struct amdgpu_xcp_mgr *xcp_mgr, - int mode, - struct amdgpu_xcp_cfg *xcp_cfg) +static int __aqua_vanjaram_get_px_mode_info(struct amdgpu_xcp_mgr *xcp_mgr, + int px_mode, int *num_xcp, + uint16_t *nps_modes) { struct amdgpu_device *adev = xcp_mgr->adev; - int max_res[AMDGPU_XCP_RES_MAX] = {}; - bool res_lt_xcp; - int num_xcp, i; - u16 nps_modes; - if (!(xcp_mgr->supp_xcp_modes & BIT(mode))) + if (!num_xcp || !nps_modes || !(xcp_mgr->supp_xcp_modes & BIT(px_mode))) return -EINVAL; - max_res[AMDGPU_XCP_RES_XCC] = NUM_XCC(adev->gfx.xcc_mask); - max_res[AMDGPU_XCP_RES_DMA] = adev->sdma.num_instances; - max_res[AMDGPU_XCP_RES_DEC] = adev->vcn.num_vcn_inst; - max_res[AMDGPU_XCP_RES_JPEG] = adev->jpeg.num_jpeg_inst; - - switch (mode) { + switch (px_mode) { case AMDGPU_SPX_PARTITION_MODE: - num_xcp = 1; - nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE); + *num_xcp = 1; + *nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE); break; case AMDGPU_DPX_PARTITION_MODE: - num_xcp = 2; - nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | - BIT(AMDGPU_NPS2_PARTITION_MODE); + *num_xcp = 2; + *nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | + BIT(AMDGPU_NPS2_PARTITION_MODE); break; case AMDGPU_TPX_PARTITION_MODE: - num_xcp = 3; - nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | - BIT(AMDGPU_NPS4_PARTITION_MODE); + *num_xcp = 3; + *nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | + BIT(AMDGPU_NPS4_PARTITION_MODE); break; case AMDGPU_QPX_PARTITION_MODE: - num_xcp = 4; - nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | - BIT(AMDGPU_NPS4_PARTITION_MODE); + *num_xcp = 4; + *nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | + BIT(AMDGPU_NPS4_PARTITION_MODE); break; case AMDGPU_CPX_PARTITION_MODE: - num_xcp = NUM_XCC(adev->gfx.xcc_mask); - nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | - BIT(AMDGPU_NPS4_PARTITION_MODE); + *num_xcp = NUM_XCC(adev->gfx.xcc_mask); + *nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | + BIT(AMDGPU_NPS4_PARTITION_MODE); + if (amdgpu_sriov_vf(adev)) + *nps_modes |= BIT(AMDGPU_NPS2_PARTITION_MODE); break; default: return -EINVAL; } + return 0; +} + +static int aqua_vanjaram_get_xcp_res_info(struct amdgpu_xcp_mgr *xcp_mgr, + int mode, + struct amdgpu_xcp_cfg *xcp_cfg) +{ + struct amdgpu_device *adev = xcp_mgr->adev; + int max_res[AMDGPU_XCP_RES_MAX] = {}; + bool res_lt_xcp; + int num_xcp, i, r; + u16 nps_modes; + + if (!(xcp_mgr->supp_xcp_modes & BIT(mode))) + return -EINVAL; + + max_res[AMDGPU_XCP_RES_XCC] = NUM_XCC(adev->gfx.xcc_mask); + max_res[AMDGPU_XCP_RES_DMA] = adev->sdma.num_instances; + max_res[AMDGPU_XCP_RES_DEC] = adev->vcn.num_vcn_inst; + max_res[AMDGPU_XCP_RES_JPEG] = adev->jpeg.num_jpeg_inst; + + r = __aqua_vanjaram_get_px_mode_info(xcp_mgr, mode, &num_xcp, &nps_modes); + if (r) + return r; + xcp_cfg->compatible_nps_modes = (adev->gmc.supported_nps_modes & nps_modes); xcp_cfg->num_res = ARRAY_SIZE(max_res); @@ -543,30 +561,31 @@ static bool __aqua_vanjaram_is_valid_mode(struct amdgpu_xcp_mgr *xcp_mgr, enum amdgpu_gfx_partition mode) { struct amdgpu_device *adev = xcp_mgr->adev; - int num_xcc, num_xccs_per_xcp; + int num_xcc, num_xccs_per_xcp, r; + int num_xcp, nps_mode; + u16 supp_nps_modes; + bool comp_mode; + + nps_mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); + r = __aqua_vanjaram_get_px_mode_info(xcp_mgr, mode, &num_xcp, + &supp_nps_modes); + if (r) + return false; + comp_mode = !!(BIT(nps_mode) & supp_nps_modes); num_xcc = NUM_XCC(adev->gfx.xcc_mask); switch (mode) { case AMDGPU_SPX_PARTITION_MODE: - return adev->gmc.num_mem_partitions == 1 && num_xcc > 0; + return comp_mode && num_xcc > 0; case AMDGPU_DPX_PARTITION_MODE: - return adev->gmc.num_mem_partitions <= 2 && (num_xcc % 4) == 0; + return comp_mode && (num_xcc % 4) == 0; case AMDGPU_TPX_PARTITION_MODE: - return (adev->gmc.num_mem_partitions == 1 || - adev->gmc.num_mem_partitions == 3) && - ((num_xcc % 3) == 0); + return comp_mode && ((num_xcc % 3) == 0); case AMDGPU_QPX_PARTITION_MODE: num_xccs_per_xcp = num_xcc / 4; - return (adev->gmc.num_mem_partitions == 1 || - adev->gmc.num_mem_partitions == 4) && - (num_xccs_per_xcp >= 2); + return comp_mode && (num_xccs_per_xcp >= 2); case AMDGPU_CPX_PARTITION_MODE: - /* (num_xcc > 1) because 1 XCC is considered SPX, not CPX. - * (num_xcc % adev->gmc.num_mem_partitions) == 0 because - * num_compute_partitions can't be less than num_mem_partitions - */ - return ((num_xcc > 1) && - (num_xcc % adev->gmc.num_mem_partitions) == 0); + return comp_mode && (num_xcc > 1); default: return false; } diff --git a/drivers/gpu/drm/amd/amdgpu/atom.c b/drivers/gpu/drm/amd/amdgpu/atom.c index 81d195d366ce..427b073de2fc 100644 --- a/drivers/gpu/drm/amd/amdgpu/atom.c +++ b/drivers/gpu/drm/amd/amdgpu/atom.c @@ -1444,6 +1444,7 @@ static void atom_get_vbios_pn(struct atom_context *ctx) if (vbios_str == NULL) vbios_str += sizeof(BIOS_ATOM_PREFIX) - 1; } + OPTIMIZER_HIDE_VAR(vbios_str); if (vbios_str != NULL && *vbios_str == 0) vbios_str++; diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 508cea965983..9e8715b4739d 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -56,6 +56,8 @@ static void cik_sdma_set_buffer_funcs(struct amdgpu_device *adev); static void cik_sdma_set_vm_pte_funcs(struct amdgpu_device *adev); static int cik_sdma_soft_reset(struct amdgpu_ip_block *ip_block); +u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev); + MODULE_FIRMWARE("amdgpu/bonaire_sdma.bin"); MODULE_FIRMWARE("amdgpu/bonaire_sdma1.bin"); MODULE_FIRMWARE("amdgpu/hawaii_sdma.bin"); @@ -67,9 +69,6 @@ MODULE_FIRMWARE("amdgpu/kabini_sdma1.bin"); MODULE_FIRMWARE("amdgpu/mullins_sdma.bin"); MODULE_FIRMWARE("amdgpu/mullins_sdma1.bin"); -u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev); - - static void cik_sdma_free_microcode(struct amdgpu_device *adev) { int i; @@ -993,14 +992,9 @@ static int cik_sdma_sw_fini(struct amdgpu_ip_block *ip_block) static int cik_sdma_hw_init(struct amdgpu_ip_block *ip_block) { - int r; struct amdgpu_device *adev = ip_block->adev; - r = cik_sdma_start(adev); - if (r) - return r; - - return r; + return cik_sdma_start(adev); } static int cik_sdma_hw_fini(struct amdgpu_ip_block *ip_block) @@ -1040,14 +1034,10 @@ static bool cik_sdma_is_idle(struct amdgpu_ip_block *ip_block) static int cik_sdma_wait_for_idle(struct amdgpu_ip_block *ip_block) { unsigned i; - u32 tmp; struct amdgpu_device *adev = ip_block->adev; for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(mmSRBM_STATUS2) & (SRBM_STATUS2__SDMA_BUSY_MASK | - SRBM_STATUS2__SDMA1_BUSY_MASK); - - if (!tmp) + if (cik_sdma_is_idle(ip_block)) return 0; udelay(1); } diff --git a/drivers/gpu/drm/amd/amdgpu/cikd.h b/drivers/gpu/drm/amd/amdgpu/cikd.h index 279288365940..8aca4f2734f2 100644 --- a/drivers/gpu/drm/amd/amdgpu/cikd.h +++ b/drivers/gpu/drm/amd/amdgpu/cikd.h @@ -60,9 +60,6 @@ #define AUD5_REGISTER_OFFSET (0x179d - 0x1780) #define AUD6_REGISTER_OFFSET (0x17a4 - 0x1780) -#define BONAIRE_GB_ADDR_CONFIG_GOLDEN 0x12010001 -#define HAWAII_GB_ADDR_CONFIG_GOLDEN 0x12011003 - #define PIPEID(x) ((x) << 0) #define MEID(x) ((x) << 2) #define VMID(x) ((x) << 4) diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index df401aded662..bf7c22f81cda 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -3075,7 +3075,7 @@ static int dce_v10_0_set_hpd_irq_state(struct amdgpu_device *adev, u32 tmp; if (hpd >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", hpd); + DRM_DEBUG("invalid hpd %d\n", hpd); return 0; } @@ -3227,7 +3227,7 @@ static void dce_v10_0_hpd_int_ack(struct amdgpu_device *adev, u32 tmp; if (hpd >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", hpd); + DRM_DEBUG("invalid hpd %d\n", hpd); return; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 80f01c3989cd..47e05783c4a0 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -3206,7 +3206,7 @@ static int dce_v11_0_set_hpd_irq_state(struct amdgpu_device *adev, u32 tmp; if (hpd >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", hpd); + DRM_DEBUG("invalid hpd %d\n", hpd); return 0; } @@ -3358,7 +3358,7 @@ static void dce_v11_0_hpd_int_ack(struct amdgpu_device *adev, u32 tmp; if (hpd >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", hpd); + DRM_DEBUG("invalid hpd %d\n", hpd); return; } @@ -3488,8 +3488,7 @@ static const struct amd_ip_funcs dce_v11_0_ip_funcs = { .set_powergating_state = dce_v11_0_set_powergating_state, }; -static void -dce_v11_0_encoder_mode_set(struct drm_encoder *encoder, +static void dce_v11_0_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c index 255c70959343..276c025c4c03 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c @@ -287,7 +287,7 @@ static void dce_v6_0_hpd_int_ack(struct amdgpu_device *adev, u32 tmp; if (hpd >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", hpd); + DRM_DEBUG("invalid hpd %d\n", hpd); return; } @@ -412,7 +412,7 @@ static void dce_v6_0_set_vga_render_state(struct amdgpu_device *adev, { if (!render) WREG32(mmVGA_RENDER_CONTROL, - RREG32(mmVGA_RENDER_CONTROL) & VGA_VSTATUS_CNTL); + RREG32(mmVGA_RENDER_CONTROL) & ~VGA_RENDER_CONTROL__VGA_VSTATUS_CNTL_MASK); } static int dce_v6_0_get_num_crtc(struct amdgpu_device *adev) @@ -1011,16 +1011,16 @@ static void dce_v6_0_program_watermarks(struct amdgpu_device *adev, /* select wm A */ arb_control3 = RREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset); tmp = arb_control3; - tmp &= ~LATENCY_WATERMARK_MASK(3); - tmp |= LATENCY_WATERMARK_MASK(1); + tmp &= ~(3 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT); + tmp |= (1 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT); WREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, tmp); WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset, ((latency_watermark_a << DPG_PIPE_URGENCY_CONTROL__URGENCY_LOW_WATERMARK__SHIFT) | (line_time << DPG_PIPE_URGENCY_CONTROL__URGENCY_HIGH_WATERMARK__SHIFT))); /* select wm B */ tmp = RREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset); - tmp &= ~LATENCY_WATERMARK_MASK(3); - tmp |= LATENCY_WATERMARK_MASK(2); + tmp &= ~(3 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT); + tmp |= (2 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT); WREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, tmp); WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset, ((latency_watermark_b << DPG_PIPE_URGENCY_CONTROL__URGENCY_LOW_WATERMARK__SHIFT) | @@ -1089,7 +1089,7 @@ static u32 dce_v6_0_line_buffer_adjust(struct amdgpu_device *adev, } WREG32(mmDC_LB_MEMORY_SPLIT + amdgpu_crtc->crtc_offset, - DC_LB_MEMORY_CONFIG(tmp)); + (tmp << DC_LB_MEMORY_SPLIT__DC_LB_MEMORY_CONFIG__SHIFT)); WREG32(mmPIPE0_DMIF_BUFFER_CONTROL + pipe_offset, (buffer_alloc << PIPE0_DMIF_BUFFER_CONTROL__DMIF_BUFFERS_ALLOCATED__SHIFT)); @@ -1306,6 +1306,7 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder) struct amdgpu_device *adev = drm_to_adev(dev); struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder); struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv; + u32 offset; struct drm_connector *connector; struct drm_connector_list_iter iter; struct amdgpu_connector *amdgpu_connector = NULL; @@ -1327,6 +1328,11 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder) { ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13, HDMI_AUDIO_CODING_TYPE_WMA_PRO }, }; + if (!dig || !dig->afmt || !dig->afmt->pin) + return; + + offset = dig->afmt->pin->offset; + drm_connector_list_iter_begin(dev, &iter); drm_for_each_connector_iter(connector, &iter) { if (connector->encoder == encoder) { @@ -1348,7 +1354,7 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder) return; for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) { - u32 tmp = 0; + u32 value = 0; u8 stereo_freqs = 0; int max_channels = -1; int j; @@ -1358,12 +1364,12 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder) if (sad->format == eld_reg_to_type[i][1]) { if (sad->channels > max_channels) { - tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0, - MAX_CHANNELS, sad->channels); - tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0, - DESCRIPTOR_BYTE_2, sad->byte2); - tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0, - SUPPORTED_FREQUENCIES, sad->freq); + value = (sad->channels << + AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__MAX_CHANNELS__SHIFT) | + (sad->byte2 << + AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__DESCRIPTOR_BYTE_2__SHIFT) | + (sad->freq << + AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES__SHIFT); max_channels = sad->channels; } @@ -1374,13 +1380,13 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder) } } - tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0, - SUPPORTED_FREQUENCIES_STEREO, stereo_freqs); - WREG32_AUDIO_ENDPT(dig->afmt->pin->offset, eld_reg_to_type[i][0], tmp); + value |= (stereo_freqs << + AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES_STEREO__SHIFT); + + WREG32_AUDIO_ENDPT(offset, eld_reg_to_type[i][0], value); } kfree(sads); - } static void dce_v6_0_audio_enable(struct amdgpu_device *adev, @@ -1886,7 +1892,7 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc, struct amdgpu_bo *abo; uint64_t fb_location, tiling_flags; uint32_t fb_format, fb_pitch_pixels, pipe_config; - u32 fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_NONE); + u32 fb_swap = (GRPH_ENDIAN_NONE << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); u32 viewport_w, viewport_h; int r; bool bypass_lut = false; @@ -1926,76 +1932,76 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc, switch (target_fb->format->format) { case DRM_FORMAT_C8: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_8BPP) | - GRPH_FORMAT(GRPH_FORMAT_INDEXED)); + fb_format = ((GRPH_DEPTH_8BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_INDEXED << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); break; case DRM_FORMAT_XRGB4444: case DRM_FORMAT_ARGB4444: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) | - GRPH_FORMAT(GRPH_FORMAT_ARGB4444)); + fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_ARGB4444 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16); + fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif break; case DRM_FORMAT_XRGB1555: case DRM_FORMAT_ARGB1555: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) | - GRPH_FORMAT(GRPH_FORMAT_ARGB1555)); + fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_ARGB1555 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16); + fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif break; case DRM_FORMAT_BGRX5551: case DRM_FORMAT_BGRA5551: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) | - GRPH_FORMAT(GRPH_FORMAT_BGRA5551)); + fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_BGRA5551 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16); + fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif break; case DRM_FORMAT_RGB565: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) | - GRPH_FORMAT(GRPH_FORMAT_ARGB565)); + fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_ARGB565 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16); + fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif break; case DRM_FORMAT_XRGB8888: case DRM_FORMAT_ARGB8888: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) | - GRPH_FORMAT(GRPH_FORMAT_ARGB8888)); + fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_ARGB8888 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32); + fb_swap = (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif break; case DRM_FORMAT_XRGB2101010: case DRM_FORMAT_ARGB2101010: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) | - GRPH_FORMAT(GRPH_FORMAT_ARGB2101010)); + fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_ARGB2101010 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32); + fb_swap = (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif /* Greater 8 bpc fb needs to bypass hw-lut to retain precision */ bypass_lut = true; break; case DRM_FORMAT_BGRX1010102: case DRM_FORMAT_BGRA1010102: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) | - GRPH_FORMAT(GRPH_FORMAT_BGRA1010102)); + fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_BGRA1010102 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32); + fb_swap = (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif /* Greater 8 bpc fb needs to bypass hw-lut to retain precision */ bypass_lut = true; break; case DRM_FORMAT_XBGR8888: case DRM_FORMAT_ABGR8888: - fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) | - GRPH_FORMAT(GRPH_FORMAT_ARGB8888)); - fb_swap = (GRPH_RED_CROSSBAR(GRPH_RED_SEL_B) | - GRPH_BLUE_CROSSBAR(GRPH_BLUE_SEL_R)); + fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) | + (GRPH_FORMAT_ARGB8888 << GRPH_CONTROL__GRPH_FORMAT__SHIFT)); + fb_swap = ((GRPH_RED_SEL_B << GRPH_SWAP_CNTL__GRPH_RED_CROSSBAR__SHIFT) | + (GRPH_BLUE_SEL_R << GRPH_SWAP_CNTL__GRPH_BLUE_CROSSBAR__SHIFT)); #ifdef __BIG_ENDIAN - fb_swap |= GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32); + fb_swap |= (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT); #endif break; default: @@ -2013,18 +2019,18 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc, tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT); num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS); - fb_format |= GRPH_NUM_BANKS(num_banks); - fb_format |= GRPH_ARRAY_MODE(GRPH_ARRAY_2D_TILED_THIN1); - fb_format |= GRPH_TILE_SPLIT(tile_split); - fb_format |= GRPH_BANK_WIDTH(bankw); - fb_format |= GRPH_BANK_HEIGHT(bankh); - fb_format |= GRPH_MACRO_TILE_ASPECT(mtaspect); + fb_format |= (num_banks << GRPH_CONTROL__GRPH_NUM_BANKS__SHIFT); + fb_format |= (GRPH_ARRAY_2D_TILED_THIN1 << GRPH_CONTROL__GRPH_ARRAY_MODE__SHIFT); + fb_format |= (tile_split << GRPH_CONTROL__GRPH_TILE_SPLIT__SHIFT); + fb_format |= (bankw << GRPH_CONTROL__GRPH_BANK_WIDTH__SHIFT); + fb_format |= (bankh << GRPH_CONTROL__GRPH_BANK_HEIGHT__SHIFT); + fb_format |= (mtaspect << GRPH_CONTROL__GRPH_MACRO_TILE_ASPECT__SHIFT); } else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == ARRAY_1D_TILED_THIN1) { - fb_format |= GRPH_ARRAY_MODE(GRPH_ARRAY_1D_TILED_THIN1); + fb_format |= (GRPH_ARRAY_1D_TILED_THIN1 << GRPH_CONTROL__GRPH_ARRAY_MODE__SHIFT); } pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG); - fb_format |= GRPH_PIPE_CONFIG(pipe_config); + fb_format |= (pipe_config << GRPH_CONTROL__GRPH_PIPE_CONFIG__SHIFT); dce_v6_0_vga_enable(crtc, false); @@ -2040,7 +2046,7 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc, WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset, (u32)fb_location & GRPH_PRIMARY_SURFACE_ADDRESS__GRPH_PRIMARY_SURFACE_ADDRESS_MASK); WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset, - (u32) fb_location & GRPH_PRIMARY_SURFACE_ADDRESS__GRPH_PRIMARY_SURFACE_ADDRESS_MASK); + (u32) fb_location & GRPH_SECONDARY_SURFACE_ADDRESS__GRPH_SECONDARY_SURFACE_ADDRESS_MASK); WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format); WREG32(mmGRPH_SWAP_CNTL + amdgpu_crtc->crtc_offset, fb_swap); @@ -2108,14 +2114,13 @@ static void dce_v6_0_set_interleave(struct drm_crtc *crtc, if (mode->flags & DRM_MODE_FLAG_INTERLACE) WREG32(mmDATA_FORMAT + amdgpu_crtc->crtc_offset, - INTERLEAVE_EN); + DATA_FORMAT__INTERLEAVE_EN_MASK); else WREG32(mmDATA_FORMAT + amdgpu_crtc->crtc_offset, 0); } static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc) { - struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); struct drm_device *dev = crtc->dev; struct amdgpu_device *adev = drm_to_adev(dev); @@ -2125,15 +2130,15 @@ static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc) DRM_DEBUG_KMS("%d\n", amdgpu_crtc->crtc_id); WREG32(mmINPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset, - ((0 << INPUT_CSC_CONTROL__INPUT_CSC_GRPH_MODE__SHIFT) | - (0 << INPUT_CSC_CONTROL__INPUT_CSC_OVL_MODE__SHIFT))); + ((INPUT_CSC_BYPASS << INPUT_CSC_CONTROL__INPUT_CSC_GRPH_MODE__SHIFT) | + (INPUT_CSC_BYPASS << INPUT_CSC_CONTROL__INPUT_CSC_OVL_MODE__SHIFT))); WREG32(mmPRESCALE_GRPH_CONTROL + amdgpu_crtc->crtc_offset, PRESCALE_GRPH_CONTROL__GRPH_PRESCALE_BYPASS_MASK); WREG32(mmPRESCALE_OVL_CONTROL + amdgpu_crtc->crtc_offset, PRESCALE_OVL_CONTROL__OVL_PRESCALE_BYPASS_MASK); WREG32(mmINPUT_GAMMA_CONTROL + amdgpu_crtc->crtc_offset, - ((0 << INPUT_GAMMA_CONTROL__GRPH_INPUT_GAMMA_MODE__SHIFT) | - (0 << INPUT_GAMMA_CONTROL__OVL_INPUT_GAMMA_MODE__SHIFT))); + ((INPUT_GAMMA_USE_LUT << INPUT_GAMMA_CONTROL__GRPH_INPUT_GAMMA_MODE__SHIFT) | + (INPUT_GAMMA_USE_LUT << INPUT_GAMMA_CONTROL__OVL_INPUT_GAMMA_MODE__SHIFT))); WREG32(mmDC_LUT_CONTROL + amdgpu_crtc->crtc_offset, 0); @@ -2160,19 +2165,19 @@ static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc) } WREG32(mmDEGAMMA_CONTROL + amdgpu_crtc->crtc_offset, - ((0 << DEGAMMA_CONTROL__GRPH_DEGAMMA_MODE__SHIFT) | - (0 << DEGAMMA_CONTROL__OVL_DEGAMMA_MODE__SHIFT) | - ICON_DEGAMMA_MODE(0) | - (0 << DEGAMMA_CONTROL__CURSOR_DEGAMMA_MODE__SHIFT))); + ((DEGAMMA_BYPASS << DEGAMMA_CONTROL__GRPH_DEGAMMA_MODE__SHIFT) | + (DEGAMMA_BYPASS << DEGAMMA_CONTROL__OVL_DEGAMMA_MODE__SHIFT) | + (DEGAMMA_BYPASS << DEGAMMA_CONTROL__ICON_DEGAMMA_MODE__SHIFT) | + (DEGAMMA_BYPASS << DEGAMMA_CONTROL__CURSOR_DEGAMMA_MODE__SHIFT))); WREG32(mmGAMUT_REMAP_CONTROL + amdgpu_crtc->crtc_offset, - ((0 << GAMUT_REMAP_CONTROL__GRPH_GAMUT_REMAP_MODE__SHIFT) | - (0 << GAMUT_REMAP_CONTROL__OVL_GAMUT_REMAP_MODE__SHIFT))); + ((GAMUT_REMAP_BYPASS << GAMUT_REMAP_CONTROL__GRPH_GAMUT_REMAP_MODE__SHIFT) | + (GAMUT_REMAP_BYPASS << GAMUT_REMAP_CONTROL__OVL_GAMUT_REMAP_MODE__SHIFT))); WREG32(mmREGAMMA_CONTROL + amdgpu_crtc->crtc_offset, - ((0 << REGAMMA_CONTROL__GRPH_REGAMMA_MODE__SHIFT) | - (0 << REGAMMA_CONTROL__OVL_REGAMMA_MODE__SHIFT))); + ((REGAMMA_BYPASS << REGAMMA_CONTROL__GRPH_REGAMMA_MODE__SHIFT) | + (REGAMMA_BYPASS << REGAMMA_CONTROL__OVL_REGAMMA_MODE__SHIFT))); WREG32(mmOUTPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset, - ((0 << OUTPUT_CSC_CONTROL__OUTPUT_CSC_GRPH_MODE__SHIFT) | - (0 << OUTPUT_CSC_CONTROL__OUTPUT_CSC_OVL_MODE__SHIFT))); + ((OUTPUT_CSC_BYPASS << OUTPUT_CSC_CONTROL__OUTPUT_CSC_GRPH_MODE__SHIFT) | + (OUTPUT_CSC_BYPASS << OUTPUT_CSC_CONTROL__OUTPUT_CSC_OVL_MODE__SHIFT))); /* XXX match this to the depth of the crtc fmt block, move to modeset? */ WREG32(0x1a50 + amdgpu_crtc->crtc_offset, 0); @@ -2267,8 +2272,6 @@ static void dce_v6_0_hide_cursor(struct drm_crtc *crtc) WREG32(mmCUR_CONTROL + amdgpu_crtc->crtc_offset, (CURSOR_24_8_PRE_MULT << CUR_CONTROL__CURSOR_MODE__SHIFT) | (CURSOR_URGENT_1_2 << CUR_CONTROL__CURSOR_URGENT_CONTROL__SHIFT)); - - } static void dce_v6_0_show_cursor(struct drm_crtc *crtc) @@ -2285,7 +2288,6 @@ static void dce_v6_0_show_cursor(struct drm_crtc *crtc) CUR_CONTROL__CURSOR_EN_MASK | (CURSOR_24_8_PRE_MULT << CUR_CONTROL__CURSOR_MODE__SHIFT) | (CURSOR_URGENT_1_2 << CUR_CONTROL__CURSOR_URGENT_CONTROL__SHIFT)); - } static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc, @@ -2596,7 +2598,6 @@ static bool dce_v6_0_crtc_mode_fixup(struct drm_crtc *crtc, const struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { - struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); struct drm_device *dev = crtc->dev; struct drm_encoder *encoder; @@ -2669,7 +2670,7 @@ static void dce_v6_0_panic_flush(struct drm_plane *plane) /* Disable DC tiling */ fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset); - fb_format &= ~GRPH_ARRAY_MODE(0x7); + fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK; WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format); } @@ -2745,7 +2746,6 @@ static int dce_v6_0_early_init(struct amdgpu_ip_block *ip_block) static int dce_v6_0_sw_init(struct amdgpu_ip_block *ip_block) { int r, i; - bool ret; struct amdgpu_device *adev = ip_block->adev; for (i = 0; i < adev->mode_info.num_crtc; i++) { @@ -2789,8 +2789,7 @@ static int dce_v6_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - ret = amdgpu_atombios_get_connector_info_from_object_table(adev); - if (ret) + if (amdgpu_atombios_get_connector_info_from_object_table(adev)) amdgpu_display_print_display_setup(adev_to_drm(adev)); else return -EINVAL; @@ -2986,12 +2985,12 @@ static void dce_v6_0_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev, switch (state) { case AMDGPU_IRQ_STATE_DISABLE: interrupt_mask = RREG32(mmINT_MASK + reg_block); - interrupt_mask &= ~VBLANK_INT_MASK; + interrupt_mask &= ~INT_MASK__VBLANK_INT_MASK; WREG32(mmINT_MASK + reg_block, interrupt_mask); break; case AMDGPU_IRQ_STATE_ENABLE: interrupt_mask = RREG32(mmINT_MASK + reg_block); - interrupt_mask |= VBLANK_INT_MASK; + interrupt_mask |= INT_MASK__VBLANK_INT_MASK; WREG32(mmINT_MASK + reg_block, interrupt_mask); break; default: @@ -3006,28 +3005,28 @@ static void dce_v6_0_set_crtc_vline_interrupt_state(struct amdgpu_device *adev, } -static int dce_v6_0_set_hpd_interrupt_state(struct amdgpu_device *adev, +static int dce_v6_0_set_hpd_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, - unsigned type, + unsigned hpd, enum amdgpu_interrupt_state state) { u32 dc_hpd_int_cntl; - if (type >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", type); + if (hpd >= adev->mode_info.num_hpd) { + DRM_DEBUG("invalid hpd %d\n", hpd); return 0; } switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type]); - dc_hpd_int_cntl &= ~DC_HPDx_INT_EN; - WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type], dc_hpd_int_cntl); + dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]); + dc_hpd_int_cntl &= ~DC_HPD1_INT_CONTROL__DC_HPD1_INT_EN_MASK; + WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], dc_hpd_int_cntl); break; case AMDGPU_IRQ_STATE_ENABLE: - dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type]); - dc_hpd_int_cntl |= DC_HPDx_INT_EN; - WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type], dc_hpd_int_cntl); + dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]); + dc_hpd_int_cntl |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_EN_MASK; + WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], dc_hpd_int_cntl); break; default: break; @@ -3036,7 +3035,7 @@ static int dce_v6_0_set_hpd_interrupt_state(struct amdgpu_device *adev, return 0; } -static int dce_v6_0_set_crtc_interrupt_state(struct amdgpu_device *adev, +static int dce_v6_0_set_crtc_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, enum amdgpu_interrupt_state state) @@ -3096,7 +3095,7 @@ static int dce_v6_0_crtc_irq(struct amdgpu_device *adev, switch (entry->src_data[0]) { case 0: /* vblank */ if (disp_int & interrupt_status_offsets[crtc].vblank) - WREG32(mmVBLANK_STATUS + crtc_offsets[crtc], VBLANK_ACK); + WREG32(mmVBLANK_STATUS + crtc_offsets[crtc], VBLANK_STATUS__VBLANK_ACK_MASK); else DRM_DEBUG("IH: IH event w/o asserted irq bit?\n"); @@ -3107,7 +3106,7 @@ static int dce_v6_0_crtc_irq(struct amdgpu_device *adev, break; case 1: /* vline */ if (disp_int & interrupt_status_offsets[crtc].vline) - WREG32(mmVLINE_STATUS + crtc_offsets[crtc], VLINE_ACK); + WREG32(mmVLINE_STATUS + crtc_offsets[crtc], VLINE_STATUS__VLINE_ACK_MASK); else DRM_DEBUG("IH: IH event w/o asserted irq bit?\n"); @@ -3121,7 +3120,7 @@ static int dce_v6_0_crtc_irq(struct amdgpu_device *adev, return 0; } -static int dce_v6_0_set_pageflip_interrupt_state(struct amdgpu_device *adev, +static int dce_v6_0_set_pageflip_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, enum amdgpu_interrupt_state state) @@ -3172,7 +3171,7 @@ static int dce_v6_0_pageflip_irq(struct amdgpu_device *adev, spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags); works = amdgpu_crtc->pflip_works; - if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED){ + if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED) { DRM_DEBUG_DRIVER("amdgpu_crtc->pflip_status = %d != " "AMDGPU_FLIP_SUBMITTED(%d)\n", amdgpu_crtc->pflip_status, @@ -3249,12 +3248,10 @@ static const struct amd_ip_funcs dce_v6_0_ip_funcs = { .set_powergating_state = dce_v6_0_set_powergating_state, }; -static void -dce_v6_0_encoder_mode_set(struct drm_encoder *encoder, +static void dce_v6_0_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { - struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder); int em = amdgpu_atombios_encoder_get_encoder_mode(encoder); @@ -3274,7 +3271,6 @@ dce_v6_0_encoder_mode_set(struct drm_encoder *encoder, static void dce_v6_0_encoder_prepare(struct drm_encoder *encoder) { - struct amdgpu_device *adev = drm_to_adev(encoder->dev); struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder); struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder); @@ -3314,7 +3310,6 @@ static void dce_v6_0_encoder_prepare(struct drm_encoder *encoder) static void dce_v6_0_encoder_commit(struct drm_encoder *encoder) { - struct drm_device *dev = encoder->dev; struct amdgpu_device *adev = drm_to_adev(dev); @@ -3325,7 +3320,6 @@ static void dce_v6_0_encoder_commit(struct drm_encoder *encoder) static void dce_v6_0_encoder_disable(struct drm_encoder *encoder) { - struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder); struct amdgpu_encoder_atom_dig *dig; int em = amdgpu_atombios_encoder_get_encoder_mode(encoder); @@ -3541,17 +3535,17 @@ static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev) } static const struct amdgpu_irq_src_funcs dce_v6_0_crtc_irq_funcs = { - .set = dce_v6_0_set_crtc_interrupt_state, + .set = dce_v6_0_set_crtc_irq_state, .process = dce_v6_0_crtc_irq, }; static const struct amdgpu_irq_src_funcs dce_v6_0_pageflip_irq_funcs = { - .set = dce_v6_0_set_pageflip_interrupt_state, + .set = dce_v6_0_set_pageflip_irq_state, .process = dce_v6_0_pageflip_irq, }; static const struct amdgpu_irq_src_funcs dce_v6_0_hpd_irq_funcs = { - .set = dce_v6_0_set_hpd_interrupt_state, + .set = dce_v6_0_set_hpd_irq_state, .process = dce_v6_0_hpd_irq, }; diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index 07358546581f..e62ccf9eb73d 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -271,7 +271,7 @@ static void dce_v8_0_hpd_int_ack(struct amdgpu_device *adev, u32 tmp; if (hpd >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", hpd); + DRM_DEBUG("invalid hpd %d\n", hpd); return; } @@ -3021,7 +3021,7 @@ static void dce_v8_0_set_crtc_vline_interrupt_state(struct amdgpu_device *adev, } } -static int dce_v8_0_set_hpd_interrupt_state(struct amdgpu_device *adev, +static int dce_v8_0_set_hpd_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, enum amdgpu_interrupt_state state) @@ -3029,7 +3029,7 @@ static int dce_v8_0_set_hpd_interrupt_state(struct amdgpu_device *adev, u32 dc_hpd_int_cntl; if (type >= adev->mode_info.num_hpd) { - DRM_DEBUG("invalid hdp %d\n", type); + DRM_DEBUG("invalid hpd %d\n", type); return 0; } @@ -3051,7 +3051,7 @@ static int dce_v8_0_set_hpd_interrupt_state(struct amdgpu_device *adev, return 0; } -static int dce_v8_0_set_crtc_interrupt_state(struct amdgpu_device *adev, +static int dce_v8_0_set_crtc_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, enum amdgpu_interrupt_state state) @@ -3136,7 +3136,7 @@ static int dce_v8_0_crtc_irq(struct amdgpu_device *adev, return 0; } -static int dce_v8_0_set_pageflip_interrupt_state(struct amdgpu_device *adev, +static int dce_v8_0_set_pageflip_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, enum amdgpu_interrupt_state state) @@ -3547,17 +3547,17 @@ static void dce_v8_0_set_display_funcs(struct amdgpu_device *adev) } static const struct amdgpu_irq_src_funcs dce_v8_0_crtc_irq_funcs = { - .set = dce_v8_0_set_crtc_interrupt_state, + .set = dce_v8_0_set_crtc_irq_state, .process = dce_v8_0_crtc_irq, }; static const struct amdgpu_irq_src_funcs dce_v8_0_pageflip_irq_funcs = { - .set = dce_v8_0_set_pageflip_interrupt_state, + .set = dce_v8_0_set_pageflip_irq_state, .process = dce_v8_0_pageflip_irq, }; static const struct amdgpu_irq_src_funcs dce_v8_0_hpd_irq_funcs = { - .set = dce_v8_0_set_hpd_interrupt_state, + .set = dce_v8_0_set_hpd_irq_state, .process = dce_v8_0_hpd_irq, }; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index 23e6a05359c2..75ea071744eb 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -368,11 +368,6 @@ static const struct amdgpu_hwip_reg_entry gc_reg_list_10_1[] = { SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_DEBUG_INST_ADDR), SOC15_REG_ENTRY_STR(GC, 0, mmRLC_LX6_CORE_PDEBUG_INST), /* cp header registers */ - SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME2_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), SOC15_REG_ENTRY_STR(GC, 0, mmCP_MES_HEADER_DUMP), /* SE status registers */ SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE0), @@ -421,7 +416,16 @@ static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_10[] = { SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_CNTL_STACK_OFFSET), SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_CNTL_STACK_DW_CNT), SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_WG_STATE_OFFSET), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_STATUS) + SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_STATUS), + /* cp header registers */ + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), }; static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_10[] = { @@ -448,7 +452,32 @@ static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_10[] = { SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_MQD_BASE_ADDR), SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_MQD_BASE_ADDR_HI), SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI) + SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI), + /* gfx header registers */ + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), }; static const struct soc15_reg_golden golden_settings_gc_10_1[] = { @@ -4296,9 +4325,7 @@ static u32 gfx_v10_0_get_csb_size(struct amdgpu_device *adev) static void gfx_v10_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer) { - u32 count = 0, i; - const struct cs_section_def *sect = NULL; - const struct cs_extent_def *ext = NULL; + u32 count = 0; int ctx_reg_offset; if (adev->gfx.rlc.cs_data == NULL) @@ -4306,39 +4333,15 @@ static void gfx_v10_0_get_csb_buffer(struct amdgpu_device *adev, if (buffer == NULL) return; - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + count = amdgpu_gfx_csb_preamble_start(buffer); + count = amdgpu_gfx_csb_data_parser(adev, buffer, count); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); - buffer[count++] = cpu_to_le32(0x80000000); - buffer[count++] = cpu_to_le32(0x80000000); - - for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { - for (ext = sect->section; ext->extent != NULL; ++ext) { - if (sect->id == SECT_CONTEXT) { - buffer[count++] = - cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); - buffer[count++] = cpu_to_le32(ext->reg_index - - PACKET3_SET_CONTEXT_REG_START); - for (i = 0; i < ext->reg_count; i++) - buffer[count++] = cpu_to_le32(ext->extent[i]); - } else { - return; - } - } - } - - ctx_reg_offset = - SOC15_REG_OFFSET(GC, 0, mmPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START; + ctx_reg_offset = SOC15_REG_OFFSET(GC, 0, mmPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START; buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1)); buffer[count++] = cpu_to_le32(ctx_reg_offset); buffer[count++] = cpu_to_le32(adev->gfx.config.pa_sc_tile_steering_override); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); - buffer[count++] = cpu_to_le32(0); + amdgpu_gfx_csb_preamble_end(buffer, count); } static void gfx_v10_0_rlc_fini(struct amdgpu_device *adev) @@ -4752,6 +4755,7 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) int i, j, k, r, ring_id = 0; int xcc_id = 0; struct amdgpu_device *adev = ip_block->adev; + int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */ INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler); @@ -4763,7 +4767,7 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(10, 1, 4): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 1; - adev->gfx.me.num_queue_per_pipe = 1; + adev->gfx.me.num_queue_per_pipe = 8; adev->gfx.mec.num_mec = 2; adev->gfx.mec.num_pipe_per_mec = 4; adev->gfx.mec.num_queue_per_pipe = 8; @@ -4778,7 +4782,7 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(10, 3, 7): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 2; - adev->gfx.me.num_queue_per_pipe = 1; + adev->gfx.me.num_queue_per_pipe = 2; adev->gfx.mec.num_mec = 2; adev->gfx.mec.num_pipe_per_mec = 4; adev->gfx.mec.num_queue_per_pipe = 4; @@ -4800,7 +4804,7 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) adev->gfx.cleaner_shader_size = sizeof(gfx_10_1_10_cleaner_shader_hex); if (adev->gfx.me_fw_version >= 101 && adev->gfx.pfp_fw_version >= 158 && - adev->gfx.mec_fw_version >= 152) { + adev->gfx.mec_fw_version >= 151) { adev->gfx.enable_cleaner_shader = true; r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size); if (r) { @@ -4810,7 +4814,9 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) } break; case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 1): case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 3): case IP_VERSION(10, 3, 4): case IP_VERSION(10, 3, 5): adev->gfx.cleaner_shader_ptr = gfx_10_3_0_cleaner_shader_hex; @@ -4826,6 +4832,34 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) } } break; + case IP_VERSION(10, 3, 6): + adev->gfx.cleaner_shader_ptr = gfx_10_3_0_cleaner_shader_hex; + adev->gfx.cleaner_shader_size = sizeof(gfx_10_3_0_cleaner_shader_hex); + if (adev->gfx.me_fw_version >= 14 && + adev->gfx.pfp_fw_version >= 17 && + adev->gfx.mec_fw_version >= 24) { + adev->gfx.enable_cleaner_shader = true; + r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size); + if (r) { + adev->gfx.enable_cleaner_shader = false; + dev_err(adev->dev, "Failed to initialize cleaner shader\n"); + } + } + break; + case IP_VERSION(10, 3, 7): + adev->gfx.cleaner_shader_ptr = gfx_10_3_0_cleaner_shader_hex; + adev->gfx.cleaner_shader_size = sizeof(gfx_10_3_0_cleaner_shader_hex); + if (adev->gfx.me_fw_version >= 4 && + adev->gfx.pfp_fw_version >= 9 && + adev->gfx.mec_fw_version >= 12) { + adev->gfx.enable_cleaner_shader = true; + r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size); + if (r) { + adev->gfx.enable_cleaner_shader = false; + dev_err(adev->dev, "Failed to initialize cleaner shader\n"); + } + } + break; default: adev->gfx.enable_cleaner_shader = false; break; @@ -4886,7 +4920,7 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) /* set up the gfx ring */ for (i = 0; i < adev->gfx.me.num_me; i++) { - for (j = 0; j < adev->gfx.me.num_queue_per_pipe; j++) { + for (j = 0; j < num_queue_per_pipe; j++) { for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) { if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j)) continue; @@ -9645,9 +9679,14 @@ static void gfx_v10_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printe for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) { drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k); for (reg = 0; reg < reg_count; reg++) { - drm_printf(p, "%-50s \t 0x%08x\n", - gc_cp_reg_list_10[reg].reg_name, - adev->gfx.ip_dump_compute_queues[index + reg]); + if (i && gc_cp_reg_list_10[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP) + drm_printf(p, "%-50s \t 0x%08x\n", + "mmCP_MEC_ME2_HEADER_DUMP", + adev->gfx.ip_dump_compute_queues[index + reg]); + else + drm_printf(p, "%-50s \t 0x%08x\n", + gc_cp_reg_list_10[reg].reg_name, + adev->gfx.ip_dump_compute_queues[index + reg]); } index += reg_count; } @@ -9708,9 +9747,13 @@ static void gfx_v10_ip_dump(struct amdgpu_ip_block *ip_block) nv_grbm_select(adev, adev->gfx.me.num_me + i, j, k, 0); for (reg = 0; reg < reg_count; reg++) { - adev->gfx.ip_dump_compute_queues[index + reg] = - RREG32(SOC15_REG_ENTRY_OFFSET( - gc_cp_reg_list_10[reg])); + if (i && gc_cp_reg_list_10[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP) + adev->gfx.ip_dump_compute_queues[index + reg] = + RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_MEC_ME2_HEADER_DUMP)); + else + adev->gfx.ip_dump_compute_queues[index + reg] = + RREG32(SOC15_REG_ENTRY_OFFSET( + gc_cp_reg_list_10[reg])); } index += reg_count; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 2a5c2a1ae3c7..afd6d59164bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -48,6 +48,8 @@ #include "gfx_v11_0_3.h" #include "nbio_v4_3.h" #include "mes_v11_0.h" +#include "mes_userqueue.h" +#include "amdgpu_userq_fence.h" #define GFX11_NUM_GFX_RINGS 1 #define GFX11_MEC_HPD_SIZE 2048 @@ -177,9 +179,13 @@ static const struct amdgpu_hwip_reg_entry gc_reg_list_11_0[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_INSTR_PNTR), SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STATUS), /* cp header registers */ - SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), /* SE status registers */ SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0), @@ -230,7 +236,16 @@ static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_11[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_OFFSET), SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_DW_CNT), SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_WG_STATE_OFFSET), - SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS) + SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS), + /* cp header registers */ + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), }; static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_11[] = { @@ -259,7 +274,24 @@ static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_11[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO), SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI), SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ), - SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ) + SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ), + /* cp header registers */ + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), }; static const struct soc15_reg_golden golden_settings_gc_11_0[] = { @@ -580,33 +612,18 @@ static int gfx_v11_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t padding, offset; - - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - padding = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, padding); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, padding); - *cpu_ptr = cpu_to_le32(0xCAFEDEAD); - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) - return r; + r = amdgpu_device_wb_get(adev, &index); + if (r) + return r; - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); - cpu_ptr = &adev->wb.wb[index]; + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); + cpu_ptr = &adev->wb.wb[index]; - r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); - goto err1; - } + r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); + goto err1; } ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3); @@ -633,12 +650,10 @@ static int gfx_v11_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) else r = -EINVAL; err2: - if (!ring->is_mes_queue) - amdgpu_ib_free(&ib, NULL); + amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err1: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -833,9 +848,7 @@ static u32 gfx_v11_0_get_csb_size(struct amdgpu_device *adev) static void gfx_v11_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer) { - u32 count = 0, i; - const struct cs_section_def *sect = NULL; - const struct cs_extent_def *ext = NULL; + u32 count = 0; int ctx_reg_offset; if (adev->gfx.rlc.cs_data == NULL) @@ -843,39 +856,15 @@ static void gfx_v11_0_get_csb_buffer(struct amdgpu_device *adev, if (buffer == NULL) return; - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); - buffer[count++] = cpu_to_le32(0x80000000); - buffer[count++] = cpu_to_le32(0x80000000); - - for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { - for (ext = sect->section; ext->extent != NULL; ++ext) { - if (sect->id == SECT_CONTEXT) { - buffer[count++] = - cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); - buffer[count++] = cpu_to_le32(ext->reg_index - - PACKET3_SET_CONTEXT_REG_START); - for (i = 0; i < ext->reg_count; i++) - buffer[count++] = cpu_to_le32(ext->extent[i]); - } else { - return; - } - } - } + count = amdgpu_gfx_csb_preamble_start(buffer); + count = amdgpu_gfx_csb_data_parser(adev, buffer, count); - ctx_reg_offset = - SOC15_REG_OFFSET(GC, 0, regPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START; + ctx_reg_offset = SOC15_REG_OFFSET(GC, 0, regPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START; buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1)); buffer[count++] = cpu_to_le32(ctx_reg_offset); buffer[count++] = cpu_to_le32(adev->gfx.config.pa_sc_tile_steering_override); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); - buffer[count++] = cpu_to_le32(0); + amdgpu_gfx_csb_preamble_end(buffer, count); } static void gfx_v11_0_rlc_fini(struct amdgpu_device *adev) @@ -1056,14 +1045,21 @@ static void gfx_v11_0_select_me_pipe_q(struct amdgpu_device *adev, #define MQD_FWWORKAREA_SIZE 484 #define MQD_FWWORKAREA_ALIGNMENT 256 -static int gfx_v11_0_get_gfx_shadow_info(struct amdgpu_device *adev, +static void gfx_v11_0_get_gfx_shadow_info_nocheck(struct amdgpu_device *adev, struct amdgpu_gfx_shadow_info *shadow_info) { - if (adev->gfx.cp_gfx_shadow) { - shadow_info->shadow_size = MQD_SHADOW_BASE_SIZE; - shadow_info->shadow_alignment = MQD_SHADOW_BASE_ALIGNMENT; - shadow_info->csa_size = MQD_FWWORKAREA_SIZE; - shadow_info->csa_alignment = MQD_FWWORKAREA_ALIGNMENT; + shadow_info->shadow_size = MQD_SHADOW_BASE_SIZE; + shadow_info->shadow_alignment = MQD_SHADOW_BASE_ALIGNMENT; + shadow_info->csa_size = MQD_FWWORKAREA_SIZE; + shadow_info->csa_alignment = MQD_FWWORKAREA_ALIGNMENT; +} + +static int gfx_v11_0_get_gfx_shadow_info(struct amdgpu_device *adev, + struct amdgpu_gfx_shadow_info *shadow_info, + bool skip_check) +{ + if (adev->gfx.cp_gfx_shadow || skip_check) { + gfx_v11_0_get_gfx_shadow_info_nocheck(adev, shadow_info); return 0; } else { memset(shadow_info, 0, sizeof(struct amdgpu_gfx_shadow_info)); @@ -1136,6 +1132,10 @@ static int gfx_v11_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id, ring->ring_obj = NULL; ring->use_doorbell = true; + if (adev->gfx.disable_kq) { + ring->no_scheduler = true; + ring->no_user_submission = true; + } if (!ring_id) ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1; @@ -1568,24 +1568,18 @@ static void gfx_v11_0_alloc_ip_dump(struct amdgpu_device *adev) static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) { - int i, j, k, r, ring_id = 0; + int i, j, k, r, ring_id; int xcc_id = 0; struct amdgpu_device *adev = ip_block->adev; + int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */ INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler); switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 1): case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): - adev->gfx.me.num_me = 1; - adev->gfx.me.num_pipe_per_me = 1; - adev->gfx.me.num_queue_per_pipe = 1; - adev->gfx.mec.num_mec = 1; - adev->gfx.mec.num_pipe_per_mec = 4; - adev->gfx.mec.num_queue_per_pipe = 4; - break; - case IP_VERSION(11, 0, 1): case IP_VERSION(11, 0, 4): case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): @@ -1593,7 +1587,7 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(11, 5, 3): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 1; - adev->gfx.me.num_queue_per_pipe = 1; + adev->gfx.me.num_queue_per_pipe = 2; adev->gfx.mec.num_mec = 1; adev->gfx.mec.num_pipe_per_mec = 4; adev->gfx.mec.num_queue_per_pipe = 4; @@ -1612,6 +1606,35 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(11, 0, 0): case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): + if (!adev->gfx.disable_uq && + adev->gfx.me_fw_version >= 2390 && + adev->gfx.pfp_fw_version >= 2530 && + adev->gfx.mec_fw_version >= 2600 && + adev->mes.fw_version[0] >= 120) { + adev->userq_funcs[AMDGPU_HW_IP_GFX] = &userq_mes_funcs; + adev->userq_funcs[AMDGPU_HW_IP_COMPUTE] = &userq_mes_funcs; + } + break; + case IP_VERSION(11, 0, 1): + case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): + /* add firmware version checks here */ + if (0 && !adev->gfx.disable_uq) { + adev->userq_funcs[AMDGPU_HW_IP_GFX] = &userq_mes_funcs; + adev->userq_funcs[AMDGPU_HW_IP_COMPUTE] = &userq_mes_funcs; + } + break; + default: + break; + } + + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex; adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex); if (adev->gfx.me_fw_version >= 2280 && @@ -1640,6 +1663,34 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) } } break; + case IP_VERSION(11, 5, 2): + adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex; + adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex); + if (adev->gfx.me_fw_version >= 12 && + adev->gfx.pfp_fw_version >= 15 && + adev->gfx.mec_fw_version >= 15) { + adev->gfx.enable_cleaner_shader = true; + r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size); + if (r) { + adev->gfx.enable_cleaner_shader = false; + dev_err(adev->dev, "Failed to initialize cleaner shader\n"); + } + } + break; + case IP_VERSION(11, 5, 3): + adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex; + adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex); + if (adev->gfx.me_fw_version >= 7 && + adev->gfx.pfp_fw_version >= 8 && + adev->gfx.mec_fw_version >= 8) { + adev->gfx.enable_cleaner_shader = true; + r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size); + if (r) { + adev->gfx.enable_cleaner_shader = false; + dev_err(adev->dev, "Failed to initialize cleaner shader\n"); + } + } + break; default: adev->gfx.enable_cleaner_shader = false; break; @@ -1701,37 +1752,42 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - /* set up the gfx ring */ - for (i = 0; i < adev->gfx.me.num_me; i++) { - for (j = 0; j < adev->gfx.me.num_queue_per_pipe; j++) { - for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) { - if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j)) - continue; - - r = gfx_v11_0_gfx_ring_init(adev, ring_id, - i, k, j); - if (r) - return r; - ring_id++; + if (adev->gfx.num_gfx_rings) { + ring_id = 0; + /* set up the gfx ring */ + for (i = 0; i < adev->gfx.me.num_me; i++) { + for (j = 0; j < num_queue_per_pipe; j++) { + for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) { + if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j)) + continue; + + r = gfx_v11_0_gfx_ring_init(adev, ring_id, + i, k, j); + if (r) + return r; + ring_id++; + } } } } - ring_id = 0; - /* set up the compute queues - allocate horizontally across pipes */ - for (i = 0; i < adev->gfx.mec.num_mec; ++i) { - for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { - for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, - k, j)) - continue; + if (adev->gfx.num_compute_rings) { + ring_id = 0; + /* set up the compute queues - allocate horizontally across pipes */ + for (i = 0; i < adev->gfx.mec.num_mec; ++i) { + for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { + for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { + if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, + k, j)) + continue; - r = gfx_v11_0_compute_ring_init(adev, ring_id, - i, k, j); - if (r) - return r; + r = gfx_v11_0_compute_ring_init(adev, ring_id, + i, k, j); + if (r) + return r; - ring_id++; + ring_id++; + } } } } @@ -4061,6 +4117,8 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, #ifdef __BIG_ENDIAN tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, BUF_SWAP, 1); #endif + if (prop->tmz_queue) + tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, TMZ_MATCH, 1); mqd->cp_gfx_hqd_cntl = tmp; /* set up cp_doorbell_control */ @@ -4081,6 +4139,16 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, /* active the queue */ mqd->cp_gfx_hqd_active = 1; + /* set gfx UQ items */ + mqd->shadow_base_lo = lower_32_bits(prop->shadow_addr); + mqd->shadow_base_hi = upper_32_bits(prop->shadow_addr); + mqd->gds_bkup_base_lo = lower_32_bits(prop->gds_bkup_addr); + mqd->gds_bkup_base_hi = upper_32_bits(prop->gds_bkup_addr); + mqd->fw_work_area_base_lo = lower_32_bits(prop->csa_addr); + mqd->fw_work_area_base_hi = upper_32_bits(prop->csa_addr); + mqd->fence_address_lo = lower_32_bits(prop->fence_address); + mqd->fence_address_hi = upper_32_bits(prop->fence_address); + return 0; } @@ -4205,6 +4273,8 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m, prop->allow_tunneling); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1); + if (prop->tmz_queue) + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TMZ, 1); mqd->cp_hqd_pq_control = tmp; /* set the wb address whether it's enabled or not */ @@ -4256,6 +4326,10 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m, mqd->cp_hqd_active = prop->hqd_active; + /* set UQ fenceaddress */ + mqd->fence_address_lo = lower_32_bits(prop->fence_address); + mqd->fence_address_hi = upper_32_bits(prop->fence_address); + return 0; } @@ -4509,11 +4583,23 @@ static int gfx_v11_0_cp_resume(struct amdgpu_device *adev) return r; } - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - r = amdgpu_ring_test_helper(ring); - if (r) - return r; + if (adev->gfx.disable_kq) { + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + ring = &adev->gfx.gfx_ring[i]; + /* we don't want to set ring->ready */ + r = amdgpu_ring_test_ring(ring); + if (r) + return r; + } + if (amdgpu_async_gfx_ring) + amdgpu_gfx_disable_kgq(adev, 0); + } else { + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + ring = &adev->gfx.gfx_ring[i]; + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + } } for (i = 0; i < adev->gfx.num_compute_rings; i++) { @@ -4722,6 +4808,49 @@ static int gfx_v11_0_hw_init(struct amdgpu_ip_block *ip_block) return r; } +static int gfx_v11_0_set_userq_eop_interrupts(struct amdgpu_device *adev, + bool enable) +{ + unsigned int irq_type; + int m, p, r; + + if (adev->userq_funcs[AMDGPU_HW_IP_GFX]) { + for (m = 0; m < adev->gfx.me.num_me; m++) { + for (p = 0; p < adev->gfx.me.num_pipe_per_me; p++) { + irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + p; + if (enable) + r = amdgpu_irq_get(adev, &adev->gfx.eop_irq, + irq_type); + else + r = amdgpu_irq_put(adev, &adev->gfx.eop_irq, + irq_type); + if (r) + return r; + } + } + } + + if (adev->userq_funcs[AMDGPU_HW_IP_COMPUTE]) { + for (m = 0; m < adev->gfx.mec.num_mec; ++m) { + for (p = 0; p < adev->gfx.mec.num_pipe_per_mec; p++) { + irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP + + (m * adev->gfx.mec.num_pipe_per_mec) + + p; + if (enable) + r = amdgpu_irq_get(adev, &adev->gfx.eop_irq, + irq_type); + else + r = amdgpu_irq_put(adev, &adev->gfx.eop_irq, + irq_type); + if (r) + return r; + } + } + } + + return 0; +} + static int gfx_v11_0_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; @@ -4731,9 +4860,11 @@ static int gfx_v11_0_hw_fini(struct amdgpu_ip_block *ip_block) amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0); + gfx_v11_0_set_userq_eop_interrupts(adev, false); if (!adev->no_hw_access) { - if (amdgpu_async_gfx_ring) { + if (amdgpu_async_gfx_ring && + !adev->gfx.disable_kq) { if (amdgpu_gfx_disable_kgq(adev, 0)) DRM_ERROR("KGQ disable failed\n"); } @@ -5059,11 +5190,36 @@ static int gfx_v11_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + switch (amdgpu_user_queue) { + case -1: + case 0: + default: + adev->gfx.disable_kq = false; + adev->gfx.disable_uq = true; + break; + case 1: + adev->gfx.disable_kq = false; + adev->gfx.disable_uq = false; + break; + case 2: + adev->gfx.disable_kq = true; + adev->gfx.disable_uq = false; + break; + } + adev->gfx.funcs = &gfx_v11_0_gfx_funcs; - adev->gfx.num_gfx_rings = GFX11_NUM_GFX_RINGS; - adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), - AMDGPU_MAX_COMPUTE_RINGS); + if (adev->gfx.disable_kq) { + /* We need one GFX ring temporarily to set up + * the clear state. + */ + adev->gfx.num_gfx_rings = 1; + adev->gfx.num_compute_rings = 0; + } else { + adev->gfx.num_gfx_rings = GFX11_NUM_GFX_RINGS; + adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), + AMDGPU_MAX_COMPUTE_RINGS); + } gfx_v11_0_set_kiq_pm4_funcs(adev); gfx_v11_0_set_ring_funcs(adev); @@ -5094,6 +5250,11 @@ static int gfx_v11_0_late_init(struct amdgpu_ip_block *ip_block) r = amdgpu_irq_get(adev, &adev->gfx.bad_op_irq, 0); if (r) return r; + + r = gfx_v11_0_set_userq_eop_interrupts(adev, true); + if (r) + return r; + return 0; } @@ -5691,10 +5852,6 @@ static void gfx_v11_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, (!amdgpu_sriov_vf(ring->adev) && flags & AMDGPU_IB_PREEMPTED) ? true : false); } - if (ring->is_mes_queue) - /* inherit vmid from mqd */ - control |= 0x400000; - amdgpu_ring_write(ring, header); BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ amdgpu_ring_write(ring, @@ -5714,10 +5871,6 @@ static void gfx_v11_0_ring_emit_ib_compute(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24); - if (ring->is_mes_queue) - /* inherit vmid from mqd */ - control |= 0x40000000; - /* Currently, there is a high possibility to get wave ID mismatch * between ME and GDS, leading to a hw deadlock, because ME generates * different wave IDs than the GDS expects. This situation happens @@ -5775,8 +5928,7 @@ static void gfx_v11_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, amdgpu_ring_write(ring, upper_32_bits(addr)); amdgpu_ring_write(ring, lower_32_bits(seq)); amdgpu_ring_write(ring, upper_32_bits(seq)); - amdgpu_ring_write(ring, ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0); + amdgpu_ring_write(ring, 0); } static void gfx_v11_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring) @@ -5804,10 +5956,7 @@ static void gfx_v11_0_ring_invalidate_tlbs(struct amdgpu_ring *ring, static void gfx_v11_0_ring_emit_vm_flush(struct amdgpu_ring *ring, unsigned vmid, uint64_t pd_addr) { - if (ring->is_mes_queue) - gfx_v11_0_ring_invalidate_tlbs(ring, 0, 0, false, 0); - else - amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); + amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); /* compute doesn't have PFP */ if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) { @@ -6036,28 +6185,13 @@ static void gfx_v11_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume) void *de_payload_cpu_addr; int cnt; - if (ring->is_mes_queue) { - offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gfx_meta_data) + - offsetof(struct v10_gfx_meta_data, de_payload); - de_payload_gpu_addr = - amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - de_payload_cpu_addr = - amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gds_backup) + - offsetof(struct v10_gfx_meta_data, de_payload); - gds_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - } else { - offset = offsetof(struct v10_gfx_meta_data, de_payload); - de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset; - de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset; + offset = offsetof(struct v10_gfx_meta_data, de_payload); + de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset; + de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset; - gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) + - AMDGPU_CSA_SIZE - adev->gds.gds_size, - PAGE_SIZE); - } + gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) + + AMDGPU_CSA_SIZE - adev->gds.gds_size, + PAGE_SIZE); de_payload.gds_backup_addrlo = lower_32_bits(gds_addr); de_payload.gds_backup_addrhi = upper_32_bits(gds_addr); @@ -6296,25 +6430,23 @@ static int gfx_v11_0_eop_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) { - int i; + u32 doorbell_offset = entry->src_data[0]; u8 me_id, pipe_id, queue_id; struct amdgpu_ring *ring; - uint32_t mes_queue_id = entry->src_data[0]; + int i; DRM_DEBUG("IH: CP EOP\n"); - if (adev->enable_mes && (mes_queue_id & AMDGPU_FENCE_MES_QUEUE_FLAG)) { - struct amdgpu_mes_queue *queue; - - mes_queue_id &= AMDGPU_FENCE_MES_QUEUE_ID_MASK; + if (adev->enable_mes && doorbell_offset) { + struct amdgpu_userq_fence_driver *fence_drv = NULL; + struct xarray *xa = &adev->userq_xa; + unsigned long flags; - spin_lock(&adev->mes.queue_id_lock); - queue = idr_find(&adev->mes.queue_id_idr, mes_queue_id); - if (queue) { - DRM_DEBUG("process mes queue id = %d\n", mes_queue_id); - amdgpu_fence_process(queue->ring); - } - spin_unlock(&adev->mes.queue_id_lock); + xa_lock_irqsave(xa, flags); + fence_drv = xa_load(xa, doorbell_offset); + if (fence_drv) + amdgpu_userq_fence_driver_process(fence_drv); + xa_unlock_irqrestore(xa, flags); } else { me_id = (entry->ring_id & 0x0c) >> 2; pipe_id = (entry->ring_id & 0x03) >> 0; @@ -6481,27 +6613,29 @@ static void gfx_v11_0_handle_priv_fault(struct amdgpu_device *adev, pipe_id = (entry->ring_id & 0x03) >> 0; queue_id = (entry->ring_id & 0x70) >> 4; - switch (me_id) { - case 0: - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - if (ring->me == me_id && ring->pipe == pipe_id && - ring->queue == queue_id) - drm_sched_fault(&ring->sched); - } - break; - case 1: - case 2: - for (i = 0; i < adev->gfx.num_compute_rings; i++) { - ring = &adev->gfx.compute_ring[i]; - if (ring->me == me_id && ring->pipe == pipe_id && - ring->queue == queue_id) - drm_sched_fault(&ring->sched); + if (!adev->gfx.disable_kq) { + switch (me_id) { + case 0: + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + ring = &adev->gfx.gfx_ring[i]; + if (ring->me == me_id && ring->pipe == pipe_id && + ring->queue == queue_id) + drm_sched_fault(&ring->sched); + } + break; + case 1: + case 2: + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + ring = &adev->gfx.compute_ring[i]; + if (ring->me == me_id && ring->pipe == pipe_id && + ring->queue == queue_id) + drm_sched_fault(&ring->sched); + } + break; + default: + BUG(); + break; } - break; - default: - BUG(); - break; } } @@ -6609,6 +6743,69 @@ static void gfx_v11_0_emit_mem_sync(struct amdgpu_ring *ring) amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */ } +static bool gfx_v11_pipe_reset_support(struct amdgpu_device *adev) +{ + /* Disable the pipe reset until the CPFW fully support it.*/ + dev_warn_once(adev->dev, "The CPFW hasn't support pipe reset yet.\n"); + return false; +} + + +static int gfx_v11_reset_gfx_pipe(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t reset_pipe = 0, clean_pipe = 0; + int r; + + if (!gfx_v11_pipe_reset_support(adev)) + return -EOPNOTSUPP; + + gfx_v11_0_set_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + PFP_PIPE0_RESET, 1); + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + ME_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + PFP_PIPE0_RESET, 0); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + ME_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + PFP_PIPE1_RESET, 1); + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + ME_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + PFP_PIPE1_RESET, 0); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + ME_PIPE1_RESET, 0); + break; + default: + break; + } + + WREG32_SOC15(GC, 0, regCP_ME_CNTL, reset_pipe); + WREG32_SOC15(GC, 0, regCP_ME_CNTL, clean_pipe); + + r = (RREG32(SOC15_REG_OFFSET(GC, 0, regCP_GFX_RS64_INSTR_PNTR1)) << 2) - + RS64_FW_UC_START_ADDR_LO; + soc21_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + gfx_v11_0_unset_safe_mode(adev, 0); + + dev_info(adev->dev, "The ring %s pipe reset to the ME firmware start PC: %s\n", ring->name, + r == 0 ? "successfully" : "failed"); + /* FIXME: Sometimes driver can't cache the ME firmware start PC correctly, + * so the pipe reset status relies on the later gfx ring test result. + */ + return 0; +} + static int gfx_v11_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) { struct amdgpu_device *adev = ring->adev; @@ -6618,8 +6815,13 @@ static int gfx_v11_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) return -EINVAL; r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, false); - if (r) - return r; + if (r) { + + dev_warn(adev->dev, "reset via MES failed and try pipe reset %d\n", r); + r = gfx_v11_reset_gfx_pipe(ring); + if (r) + return r; + } r = gfx_v11_0_kgq_init_queue(ring, true); if (r) { @@ -6636,6 +6838,136 @@ static int gfx_v11_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) return amdgpu_ring_test_ring(ring); } +static int gfx_v11_0_reset_compute_pipe(struct amdgpu_ring *ring) +{ + + struct amdgpu_device *adev = ring->adev; + uint32_t reset_pipe = 0, clean_pipe = 0; + int r; + + if (!gfx_v11_pipe_reset_support(adev)) + return -EOPNOTSUPP; + + gfx_v11_0_set_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + + reset_pipe = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL); + clean_pipe = reset_pipe; + + if (adev->gfx.rs64_enable) { + + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE1_RESET, 0); + break; + case 2: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE2_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE2_RESET, 0); + break; + case 3: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE3_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE3_RESET, 0); + break; + default: + break; + } + WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, reset_pipe); + WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, clean_pipe); + r = (RREG32_SOC15(GC, 0, regCP_MEC_RS64_INSTR_PNTR) << 2) - + RS64_FW_UC_START_ADDR_LO; + } else { + if (ring->me == 1) { + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE1_RESET, 0); + break; + case 2: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE2_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE2_RESET, 0); + break; + case 3: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE3_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE3_RESET, 0); + break; + default: + break; + } + /* mec1 fw pc: CP_MEC1_INSTR_PNTR */ + } else { + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE1_RESET, 0); + break; + case 2: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE2_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE2_RESET, 0); + break; + case 3: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE3_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE3_RESET, 0); + break; + default: + break; + } + /* mec2 fw pc: CP:CP_MEC2_INSTR_PNTR */ + } + WREG32_SOC15(GC, 0, regCP_MEC_CNTL, reset_pipe); + WREG32_SOC15(GC, 0, regCP_MEC_CNTL, clean_pipe); + r = RREG32(SOC15_REG_OFFSET(GC, 0, regCP_MEC1_INSTR_PNTR)); + } + + soc21_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + gfx_v11_0_unset_safe_mode(adev, 0); + + dev_info(adev->dev, "The ring %s pipe resets to MEC FW start PC: %s\n", ring->name, + r == 0 ? "successfully" : "failed"); + /*FIXME:Sometimes driver can't cache the MEC firmware start PC correctly, so the pipe + * reset status relies on the compute ring test result. + */ + return 0; +} + static int gfx_v11_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) { struct amdgpu_device *adev = ring->adev; @@ -6646,8 +6978,10 @@ static int gfx_v11_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, true); if (r) { - dev_err(adev->dev, "reset via MMIO failed %d\n", r); - return r; + dev_warn(adev->dev, "fail(%d) to reset kcq and try pipe reset\n", r); + r = gfx_v11_0_reset_compute_pipe(ring); + if (r) + return r; } r = gfx_v11_0_kcq_init_queue(ring, true); @@ -6693,9 +7027,14 @@ static void gfx_v11_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printe for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) { drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k); for (reg = 0; reg < reg_count; reg++) { - drm_printf(p, "%-50s \t 0x%08x\n", - gc_cp_reg_list_11[reg].reg_name, - adev->gfx.ip_dump_compute_queues[index + reg]); + if (i && gc_cp_reg_list_11[reg].reg_offset == regCP_MEC_ME1_HEADER_DUMP) + drm_printf(p, "%-50s \t 0x%08x\n", + "regCP_MEC_ME2_HEADER_DUMP", + adev->gfx.ip_dump_compute_queues[index + reg]); + else + drm_printf(p, "%-50s \t 0x%08x\n", + gc_cp_reg_list_11[reg].reg_name, + adev->gfx.ip_dump_compute_queues[index + reg]); } index += reg_count; } @@ -6755,9 +7094,16 @@ static void gfx_v11_ip_dump(struct amdgpu_ip_block *ip_block) /* ME0 is for GFX so start from 1 for CP */ soc21_grbm_select(adev, adev->gfx.me.num_me + i, j, k, 0); for (reg = 0; reg < reg_count; reg++) { - adev->gfx.ip_dump_compute_queues[index + reg] = - RREG32(SOC15_REG_ENTRY_OFFSET( - gc_cp_reg_list_11[reg])); + if (i && + gc_cp_reg_list_11[reg].reg_offset == + regCP_MEC_ME1_HEADER_DUMP) + adev->gfx.ip_dump_compute_queues[index + reg] = + RREG32(SOC15_REG_OFFSET(GC, 0, + regCP_MEC_ME2_HEADER_DUMP)); + else + adev->gfx.ip_dump_compute_queues[index + reg] = + RREG32(SOC15_REG_ENTRY_OFFSET( + gc_cp_reg_list_11[reg])); } index += reg_count; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c index 62a257a4a3e9..f09d96bfee16 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c @@ -44,6 +44,8 @@ #include "gfx_v12_0.h" #include "nbif_v6_3_1.h" #include "mes_v12_0.h" +#include "mes_userqueue.h" +#include "amdgpu_userq_fence.h" #define GFX12_NUM_GFX_RINGS 1 #define GFX12_MEC_HPD_SIZE 2048 @@ -133,11 +135,14 @@ static const struct amdgpu_hwip_reg_entry gc_reg_list_12_0[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_RS64_INSTR_PNTR0), SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_RS64_INSTR_PNTR1), SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_RS64_INSTR_PNTR), - /* cp header registers */ - SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP), /* SE status registers */ SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0), @@ -186,7 +191,16 @@ static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_12[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_OFFSET), SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_DW_CNT), SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_WG_STATE_OFFSET), - SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS) + SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS), + /* cp header registers */ + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), }; static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_12[] = { @@ -215,7 +229,24 @@ static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_12[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO), SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI), SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ), - SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ) + SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ), + /* cp header registers */ + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP), }; static const struct soc15_reg_golden golden_settings_gc_12_0_rev0[] = { @@ -475,33 +506,18 @@ static int gfx_v12_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t padding, offset; - - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - padding = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, padding); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, padding); - *cpu_ptr = cpu_to_le32(0xCAFEDEAD); - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) - return r; + r = amdgpu_device_wb_get(adev, &index); + if (r) + return r; - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); - cpu_ptr = &adev->wb.wb[index]; + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); + cpu_ptr = &adev->wb.wb[index]; - r = amdgpu_ib_get(adev, NULL, 16, AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - dev_err(adev->dev, "amdgpu: failed to get ib (%ld).\n", r); - goto err1; - } + r = amdgpu_ib_get(adev, NULL, 16, AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + dev_err(adev->dev, "amdgpu: failed to get ib (%ld).\n", r); + goto err1; } ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3); @@ -528,12 +544,10 @@ static int gfx_v12_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) else r = -EINVAL; err2: - if (!ring->is_mes_queue) - amdgpu_ib_free(&ib, NULL); + amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err1: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -881,6 +895,34 @@ static void gfx_v12_0_select_me_pipe_q(struct amdgpu_device *adev, soc24_grbm_select(adev, me, pipe, q, vm); } +/* all sizes are in bytes */ +#define MQD_SHADOW_BASE_SIZE 73728 +#define MQD_SHADOW_BASE_ALIGNMENT 256 +#define MQD_FWWORKAREA_SIZE 484 +#define MQD_FWWORKAREA_ALIGNMENT 256 + +static void gfx_v12_0_get_gfx_shadow_info_nocheck(struct amdgpu_device *adev, + struct amdgpu_gfx_shadow_info *shadow_info) +{ + shadow_info->shadow_size = MQD_SHADOW_BASE_SIZE; + shadow_info->shadow_alignment = MQD_SHADOW_BASE_ALIGNMENT; + shadow_info->csa_size = MQD_FWWORKAREA_SIZE; + shadow_info->csa_alignment = MQD_FWWORKAREA_ALIGNMENT; +} + +static int gfx_v12_0_get_gfx_shadow_info(struct amdgpu_device *adev, + struct amdgpu_gfx_shadow_info *shadow_info, + bool skip_check) +{ + if (adev->gfx.cp_gfx_shadow || skip_check) { + gfx_v12_0_get_gfx_shadow_info_nocheck(adev, shadow_info); + return 0; + } + + memset(shadow_info, 0, sizeof(struct amdgpu_gfx_shadow_info)); + return -EINVAL; +} + static const struct amdgpu_gfx_funcs gfx_v12_0_gfx_funcs = { .get_gpu_clock_counter = &gfx_v12_0_get_gpu_clock_counter, .select_se_sh = &gfx_v12_0_select_se_sh, @@ -889,6 +931,7 @@ static const struct amdgpu_gfx_funcs gfx_v12_0_gfx_funcs = { .read_wave_vgprs = &gfx_v12_0_read_wave_vgprs, .select_me_pipe_q = &gfx_v12_0_select_me_pipe_q, .update_perfmon_mgcg = &gfx_v12_0_update_perf_clk, + .get_gfx_shadow_info = &gfx_v12_0_get_gfx_shadow_info, }; static int gfx_v12_0_gpu_early_init(struct amdgpu_device *adev) @@ -1346,6 +1389,7 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block) unsigned num_compute_rings; int xcc_id = 0; struct amdgpu_device *adev = ip_block->adev; + int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */ INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler); @@ -1354,7 +1398,7 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(12, 0, 1): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 1; - adev->gfx.me.num_queue_per_pipe = 1; + adev->gfx.me.num_queue_per_pipe = 8; adev->gfx.mec.num_mec = 1; adev->gfx.mec.num_pipe_per_mec = 2; adev->gfx.mec.num_queue_per_pipe = 4; @@ -1372,6 +1416,22 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block) switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(12, 0, 0): case IP_VERSION(12, 0, 1): + if (!adev->gfx.disable_uq && + adev->gfx.me_fw_version >= 2780 && + adev->gfx.pfp_fw_version >= 2840 && + adev->gfx.mec_fw_version >= 3050 && + adev->mes.fw_version[0] >= 123) { + adev->userq_funcs[AMDGPU_HW_IP_GFX] = &userq_mes_funcs; + adev->userq_funcs[AMDGPU_HW_IP_COMPUTE] = &userq_mes_funcs; + } + break; + default: + break; + } + + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): if (adev->gfx.me_fw_version >= 2480 && adev->gfx.pfp_fw_version >= 2530 && adev->gfx.mec_fw_version >= 2680 && @@ -1383,11 +1443,13 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block) break; } - /* recalculate compute rings to use based on hardware configuration */ - num_compute_rings = (adev->gfx.mec.num_pipe_per_mec * - adev->gfx.mec.num_queue_per_pipe) / 2; - adev->gfx.num_compute_rings = min(adev->gfx.num_compute_rings, - num_compute_rings); + if (adev->gfx.num_compute_rings) { + /* recalculate compute rings to use based on hardware configuration */ + num_compute_rings = (adev->gfx.mec.num_pipe_per_mec * + adev->gfx.mec.num_queue_per_pipe) / 2; + adev->gfx.num_compute_rings = min(adev->gfx.num_compute_rings, + num_compute_rings); + } /* EOP Event */ r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP, @@ -1433,37 +1495,41 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - /* set up the gfx ring */ - for (i = 0; i < adev->gfx.me.num_me; i++) { - for (j = 0; j < adev->gfx.me.num_queue_per_pipe; j++) { - for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) { - if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j)) - continue; - - r = gfx_v12_0_gfx_ring_init(adev, ring_id, - i, k, j); - if (r) - return r; - ring_id++; + if (adev->gfx.num_gfx_rings) { + /* set up the gfx ring */ + for (i = 0; i < adev->gfx.me.num_me; i++) { + for (j = 0; j < num_queue_per_pipe; j++) { + for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) { + if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j)) + continue; + + r = gfx_v12_0_gfx_ring_init(adev, ring_id, + i, k, j); + if (r) + return r; + ring_id++; + } } } } - ring_id = 0; - /* set up the compute queues - allocate horizontally across pipes */ - for (i = 0; i < adev->gfx.mec.num_mec; ++i) { - for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { - for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, - 0, i, k, j)) - continue; + if (adev->gfx.num_compute_rings) { + ring_id = 0; + /* set up the compute queues - allocate horizontally across pipes */ + for (i = 0; i < adev->gfx.mec.num_mec; ++i) { + for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { + for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { + if (!amdgpu_gfx_is_mec_queue_enabled(adev, + 0, i, k, j)) + continue; - r = gfx_v12_0_compute_ring_init(adev, ring_id, - i, k, j); - if (r) - return r; + r = gfx_v12_0_compute_ring_init(adev, ring_id, + i, k, j); + if (r) + return r; - ring_id++; + ring_id++; + } } } } @@ -2948,6 +3014,8 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, #ifdef __BIG_ENDIAN tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, BUF_SWAP, 1); #endif + if (prop->tmz_queue) + tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, TMZ_MATCH, 1); mqd->cp_gfx_hqd_cntl = tmp; /* set up cp_doorbell_control */ @@ -2968,6 +3036,14 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, /* active the queue */ mqd->cp_gfx_hqd_active = 1; + /* set gfx UQ items */ + mqd->shadow_base_lo = lower_32_bits(prop->shadow_addr); + mqd->shadow_base_hi = upper_32_bits(prop->shadow_addr); + mqd->fw_work_area_base_lo = lower_32_bits(prop->csa_addr); + mqd->fw_work_area_base_hi = upper_32_bits(prop->csa_addr); + mqd->fence_address_lo = lower_32_bits(prop->fence_address); + mqd->fence_address_hi = upper_32_bits(prop->fence_address); + return 0; } @@ -3091,6 +3167,8 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m, tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, 0); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1); tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1); + if (prop->tmz_queue) + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TMZ, 1); mqd->cp_hqd_pq_control = tmp; /* set the wb address whether it's enabled or not */ @@ -3142,6 +3220,10 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m, mqd->cp_hqd_active = prop->hqd_active; + /* set UQ fenceaddress */ + mqd->fence_address_lo = lower_32_bits(prop->fence_address); + mqd->fence_address_hi = upper_32_bits(prop->fence_address); + return 0; } @@ -3600,6 +3682,49 @@ static int gfx_v12_0_hw_init(struct amdgpu_ip_block *ip_block) return r; } +static int gfx_v12_0_set_userq_eop_interrupts(struct amdgpu_device *adev, + bool enable) +{ + unsigned int irq_type; + int m, p, r; + + if (adev->userq_funcs[AMDGPU_HW_IP_GFX]) { + for (m = 0; m < adev->gfx.me.num_me; m++) { + for (p = 0; p < adev->gfx.me.num_pipe_per_me; p++) { + irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + p; + if (enable) + r = amdgpu_irq_get(adev, &adev->gfx.eop_irq, + irq_type); + else + r = amdgpu_irq_put(adev, &adev->gfx.eop_irq, + irq_type); + if (r) + return r; + } + } + } + + if (adev->userq_funcs[AMDGPU_HW_IP_COMPUTE]) { + for (m = 0; m < adev->gfx.mec.num_mec; ++m) { + for (p = 0; p < adev->gfx.mec.num_pipe_per_mec; p++) { + irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP + + (m * adev->gfx.mec.num_pipe_per_mec) + + p; + if (enable) + r = amdgpu_irq_get(adev, &adev->gfx.eop_irq, + irq_type); + else + r = amdgpu_irq_put(adev, &adev->gfx.eop_irq, + irq_type); + if (r) + return r; + } + } + } + + return 0; +} + static int gfx_v12_0_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; @@ -3610,6 +3735,7 @@ static int gfx_v12_0_hw_fini(struct amdgpu_ip_block *ip_block) amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0); + gfx_v12_0_set_userq_eop_interrupts(adev, false); if (!adev->no_hw_access) { if (amdgpu_async_gfx_ring) { @@ -3698,11 +3824,33 @@ static int gfx_v12_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + switch (amdgpu_user_queue) { + case -1: + case 0: + default: + adev->gfx.disable_kq = false; + adev->gfx.disable_uq = true; + break; + case 1: + adev->gfx.disable_kq = false; + adev->gfx.disable_uq = false; + break; + case 2: + adev->gfx.disable_kq = true; + adev->gfx.disable_uq = false; + break; + } + adev->gfx.funcs = &gfx_v12_0_gfx_funcs; - adev->gfx.num_gfx_rings = GFX12_NUM_GFX_RINGS; - adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), - AMDGPU_MAX_COMPUTE_RINGS); + if (adev->gfx.disable_kq) { + adev->gfx.num_gfx_rings = 0; + adev->gfx.num_compute_rings = 0; + } else { + adev->gfx.num_gfx_rings = GFX12_NUM_GFX_RINGS; + adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), + AMDGPU_MAX_COMPUTE_RINGS); + } gfx_v12_0_set_kiq_pm4_funcs(adev); gfx_v12_0_set_ring_funcs(adev); @@ -3733,6 +3881,10 @@ static int gfx_v12_0_late_init(struct amdgpu_ip_block *ip_block) if (r) return r; + r = gfx_v12_0_set_userq_eop_interrupts(adev, true); + if (r) + return r; + return 0; } @@ -4172,45 +4324,17 @@ static u64 gfx_v12_0_ring_get_wptr_gfx(struct amdgpu_ring *ring) static void gfx_v12_0_ring_set_wptr_gfx(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - uint32_t *wptr_saved; - uint32_t *is_queue_unmap; - uint64_t aggregated_db_index; - uint32_t mqd_size = adev->mqds[AMDGPU_HW_IP_GFX].mqd_size; - uint64_t wptr_tmp; - - if (ring->is_mes_queue) { - wptr_saved = (uint32_t *)(ring->mqd_ptr + mqd_size); - is_queue_unmap = (uint32_t *)(ring->mqd_ptr + mqd_size + - sizeof(uint32_t)); - aggregated_db_index = - amdgpu_mes_get_aggregated_doorbell_index(adev, - ring->hw_prio); - - wptr_tmp = ring->wptr & ring->buf_mask; - atomic64_set((atomic64_t *)ring->wptr_cpu_addr, wptr_tmp); - *wptr_saved = wptr_tmp; - /* assume doorbell always being used by mes mapped queue */ - if (*is_queue_unmap) { - WDOORBELL64(aggregated_db_index, wptr_tmp); - WDOORBELL64(ring->doorbell_index, wptr_tmp); - } else { - WDOORBELL64(ring->doorbell_index, wptr_tmp); - if (*is_queue_unmap) - WDOORBELL64(aggregated_db_index, wptr_tmp); - } + if (ring->use_doorbell) { + /* XXX check if swapping is necessary on BE */ + atomic64_set((atomic64_t *)ring->wptr_cpu_addr, + ring->wptr); + WDOORBELL64(ring->doorbell_index, ring->wptr); } else { - if (ring->use_doorbell) { - /* XXX check if swapping is necessary on BE */ - atomic64_set((atomic64_t *)ring->wptr_cpu_addr, - ring->wptr); - WDOORBELL64(ring->doorbell_index, ring->wptr); - } else { - WREG32_SOC15(GC, 0, regCP_RB0_WPTR, - lower_32_bits(ring->wptr)); - WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI, - upper_32_bits(ring->wptr)); - } + WREG32_SOC15(GC, 0, regCP_RB0_WPTR, + lower_32_bits(ring->wptr)); + WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI, + upper_32_bits(ring->wptr)); } } @@ -4235,42 +4359,14 @@ static u64 gfx_v12_0_ring_get_wptr_compute(struct amdgpu_ring *ring) static void gfx_v12_0_ring_set_wptr_compute(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - uint32_t *wptr_saved; - uint32_t *is_queue_unmap; - uint64_t aggregated_db_index; - uint32_t mqd_size = adev->mqds[AMDGPU_HW_IP_COMPUTE].mqd_size; - uint64_t wptr_tmp; - - if (ring->is_mes_queue) { - wptr_saved = (uint32_t *)(ring->mqd_ptr + mqd_size); - is_queue_unmap = (uint32_t *)(ring->mqd_ptr + mqd_size + - sizeof(uint32_t)); - aggregated_db_index = - amdgpu_mes_get_aggregated_doorbell_index(adev, - ring->hw_prio); - - wptr_tmp = ring->wptr & ring->buf_mask; - atomic64_set((atomic64_t *)ring->wptr_cpu_addr, wptr_tmp); - *wptr_saved = wptr_tmp; - /* assume doorbell always used by mes mapped queue */ - if (*is_queue_unmap) { - WDOORBELL64(aggregated_db_index, wptr_tmp); - WDOORBELL64(ring->doorbell_index, wptr_tmp); - } else { - WDOORBELL64(ring->doorbell_index, wptr_tmp); - if (*is_queue_unmap) - WDOORBELL64(aggregated_db_index, wptr_tmp); - } + /* XXX check if swapping is necessary on BE */ + if (ring->use_doorbell) { + atomic64_set((atomic64_t *)ring->wptr_cpu_addr, + ring->wptr); + WDOORBELL64(ring->doorbell_index, ring->wptr); } else { - /* XXX check if swapping is necessary on BE */ - if (ring->use_doorbell) { - atomic64_set((atomic64_t *)ring->wptr_cpu_addr, - ring->wptr); - WDOORBELL64(ring->doorbell_index, ring->wptr); - } else { - BUG(); /* only DOORBELL method supported on gfx12 now */ - } + BUG(); /* only DOORBELL method supported on gfx12 now */ } } @@ -4317,10 +4413,6 @@ static void gfx_v12_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, control |= ib->length_dw | (vmid << 24); - if (ring->is_mes_queue) - /* inherit vmid from mqd */ - control |= 0x400000; - amdgpu_ring_write(ring, header); BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ amdgpu_ring_write(ring, @@ -4340,10 +4432,6 @@ static void gfx_v12_0_ring_emit_ib_compute(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24); - if (ring->is_mes_queue) - /* inherit vmid from mqd */ - control |= 0x40000000; - amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ amdgpu_ring_write(ring, @@ -4383,8 +4471,7 @@ static void gfx_v12_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, amdgpu_ring_write(ring, upper_32_bits(addr)); amdgpu_ring_write(ring, lower_32_bits(seq)); amdgpu_ring_write(ring, upper_32_bits(seq)); - amdgpu_ring_write(ring, ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0); + amdgpu_ring_write(ring, 0); } static void gfx_v12_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring) @@ -4412,10 +4499,7 @@ static void gfx_v12_0_ring_invalidate_tlbs(struct amdgpu_ring *ring, static void gfx_v12_0_ring_emit_vm_flush(struct amdgpu_ring *ring, unsigned vmid, uint64_t pd_addr) { - if (ring->is_mes_queue) - gfx_v12_0_ring_invalidate_tlbs(ring, 0, 0, false, 0); - else - amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); + amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); /* compute doesn't have PFP */ if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) { @@ -4749,25 +4833,23 @@ static int gfx_v12_0_eop_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) { - int i; + u32 doorbell_offset = entry->src_data[0]; u8 me_id, pipe_id, queue_id; struct amdgpu_ring *ring; - uint32_t mes_queue_id = entry->src_data[0]; + int i; DRM_DEBUG("IH: CP EOP\n"); - if (adev->enable_mes && (mes_queue_id & AMDGPU_FENCE_MES_QUEUE_FLAG)) { - struct amdgpu_mes_queue *queue; + if (adev->enable_mes && doorbell_offset) { + struct amdgpu_userq_fence_driver *fence_drv = NULL; + struct xarray *xa = &adev->userq_xa; + unsigned long flags; - mes_queue_id &= AMDGPU_FENCE_MES_QUEUE_ID_MASK; - - spin_lock(&adev->mes.queue_id_lock); - queue = idr_find(&adev->mes.queue_id_idr, mes_queue_id); - if (queue) { - DRM_DEBUG("process mes queue id = %d\n", mes_queue_id); - amdgpu_fence_process(queue->ring); - } - spin_unlock(&adev->mes.queue_id_lock); + xa_lock_irqsave(xa, flags); + fence_drv = xa_load(xa, doorbell_offset); + if (fence_drv) + amdgpu_userq_fence_driver_process(fence_drv); + xa_unlock_irqrestore(xa, flags); } else { me_id = (entry->ring_id & 0x0c) >> 2; pipe_id = (entry->ring_id & 0x03) >> 0; @@ -4934,27 +5016,29 @@ static void gfx_v12_0_handle_priv_fault(struct amdgpu_device *adev, pipe_id = (entry->ring_id & 0x03) >> 0; queue_id = (entry->ring_id & 0x70) >> 4; - switch (me_id) { - case 0: - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - if (ring->me == me_id && ring->pipe == pipe_id && - ring->queue == queue_id) - drm_sched_fault(&ring->sched); - } - break; - case 1: - case 2: - for (i = 0; i < adev->gfx.num_compute_rings; i++) { - ring = &adev->gfx.compute_ring[i]; - if (ring->me == me_id && ring->pipe == pipe_id && - ring->queue == queue_id) - drm_sched_fault(&ring->sched); + if (!adev->gfx.disable_kq) { + switch (me_id) { + case 0: + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + ring = &adev->gfx.gfx_ring[i]; + if (ring->me == me_id && ring->pipe == pipe_id && + ring->queue == queue_id) + drm_sched_fault(&ring->sched); + } + break; + case 1: + case 2: + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + ring = &adev->gfx.compute_ring[i]; + if (ring->me == me_id && ring->pipe == pipe_id && + ring->queue == queue_id) + drm_sched_fault(&ring->sched); + } + break; + default: + BUG(); + break; } - break; - default: - BUG(); - break; } } @@ -5160,6 +5244,69 @@ static void gfx_v12_ip_dump(struct amdgpu_ip_block *ip_block) amdgpu_gfx_off_ctrl(adev, true); } +static bool gfx_v12_pipe_reset_support(struct amdgpu_device *adev) +{ + /* Disable the pipe reset until the CPFW fully support it.*/ + dev_warn_once(adev->dev, "The CPFW hasn't support pipe reset yet.\n"); + return false; +} + +static int gfx_v12_reset_gfx_pipe(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t reset_pipe = 0, clean_pipe = 0; + int r; + + if (!gfx_v12_pipe_reset_support(adev)) + return -EOPNOTSUPP; + + gfx_v12_0_set_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + PFP_PIPE0_RESET, 1); + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + ME_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + PFP_PIPE0_RESET, 0); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + ME_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + PFP_PIPE1_RESET, 1); + reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL, + ME_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + PFP_PIPE1_RESET, 0); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL, + ME_PIPE1_RESET, 0); + break; + default: + break; + } + + WREG32_SOC15(GC, 0, regCP_ME_CNTL, reset_pipe); + WREG32_SOC15(GC, 0, regCP_ME_CNTL, clean_pipe); + + r = (RREG32(SOC15_REG_OFFSET(GC, 0, regCP_GFX_RS64_INSTR_PNTR1)) << 2) - + RS64_FW_UC_START_ADDR_LO; + soc24_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + gfx_v12_0_unset_safe_mode(adev, 0); + + dev_info(adev->dev, "The ring %s pipe reset: %s\n", ring->name, + r == 0 ? "successfully" : "failed"); + /* Sometimes the ME start pc counter can't cache correctly, so the + * PC check only as a reference and pipe reset result rely on the + * later ring test. + */ + return 0; +} + static int gfx_v12_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) { struct amdgpu_device *adev = ring->adev; @@ -5170,8 +5317,10 @@ static int gfx_v12_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, false); if (r) { - dev_err(adev->dev, "reset via MES failed %d\n", r); - return r; + dev_warn(adev->dev, "reset via MES failed and try pipe reset %d\n", r); + r = gfx_v12_reset_gfx_pipe(ring); + if (r) + return r; } r = gfx_v12_0_kgq_init_queue(ring, true); @@ -5189,6 +5338,89 @@ static int gfx_v12_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) return amdgpu_ring_test_ring(ring); } +static int gfx_v12_0_reset_compute_pipe(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t reset_pipe = 0, clean_pipe = 0; + int r = 0; + + if (!gfx_v12_pipe_reset_support(adev)) + return -EOPNOTSUPP; + + gfx_v12_0_set_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + + reset_pipe = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL); + clean_pipe = reset_pipe; + + if (adev->gfx.rs64_enable) { + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE1_RESET, 0); + break; + case 2: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE2_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE2_RESET, 0); + break; + case 3: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE3_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL, + MEC_PIPE3_RESET, 0); + break; + default: + break; + } + WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, reset_pipe); + WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, clean_pipe); + r = (RREG32_SOC15(GC, 0, regCP_MEC_RS64_INSTR_PNTR) << 2) - + RS64_FW_UC_START_ADDR_LO; + } else { + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE0_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE0_RESET, 0); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE1_RESET, 1); + clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE1_RESET, 0); + break; + default: + break; + } + WREG32_SOC15(GC, 0, regCP_MEC_CNTL, reset_pipe); + WREG32_SOC15(GC, 0, regCP_MEC_CNTL, clean_pipe); + /* Doesn't find the F32 MEC instruction pointer register, and suppose + * the driver won't run into the F32 mode. + */ + } + + soc24_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + gfx_v12_0_unset_safe_mode(adev, 0); + + dev_info(adev->dev, "The ring %s pipe resets: %s\n", ring->name, + r == 0 ? "successfully" : "failed"); + /* Need the ring test to verify the pipe reset result.*/ + return 0; +} + static int gfx_v12_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) { struct amdgpu_device *adev = ring->adev; @@ -5199,8 +5431,10 @@ static int gfx_v12_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, true); if (r) { - dev_err(adev->dev, "reset via MMIO failed %d\n", r); - return r; + dev_warn(adev->dev, "fail(%d) to reset kcq and try pipe reset\n", r); + r = gfx_v12_0_reset_compute_pipe(ring); + if (r) + return r; } r = gfx_v12_0_kcq_init_queue(ring, true); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c index 13fbee46417a..70d7a1f434c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c @@ -53,6 +53,9 @@ #define VERDE_GB_ADDR_CONFIG_GOLDEN 0x12010002 #define HAINAN_GB_ADDR_CONFIG_GOLDEN 0x02010001 +#define GFX6_NUM_GFX_RINGS 1 +#define GFX6_NUM_COMPUTE_RINGS 2 + static void gfx_v6_0_set_ring_funcs(struct amdgpu_device *adev); static void gfx_v6_0_set_irq_funcs(struct amdgpu_device *adev); static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev); @@ -1732,10 +1735,14 @@ static void gfx_v6_0_constants_init(struct amdgpu_device *adev) gfx_v6_0_get_cu_info(adev); gfx_v6_0_config_init(adev); - WREG32(mmCP_QUEUE_THRESHOLDS, ((0x16 << CP_QUEUE_THRESHOLDS__ROQ_IB1_START__SHIFT) | - (0x2b << CP_QUEUE_THRESHOLDS__ROQ_IB2_START__SHIFT))); - WREG32(mmCP_MEQ_THRESHOLDS, (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) | - (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT)); + WREG32(mmCP_QUEUE_THRESHOLDS, + ((0x16 << CP_QUEUE_THRESHOLDS__ROQ_IB1_START__SHIFT) | + (0x2b << CP_QUEUE_THRESHOLDS__ROQ_IB2_START__SHIFT))); + + /* set HW defaults for 3D engine */ + WREG32(mmCP_MEQ_THRESHOLDS, + (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) | + (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT)); sx_debug_1 = RREG32(mmSX_DEBUG_1); WREG32(mmSX_DEBUG_1, sx_debug_1); @@ -2851,44 +2858,21 @@ static u32 gfx_v6_0_get_csb_size(struct amdgpu_device *adev) static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer) { - u32 count = 0, i; - const struct cs_section_def *sect = NULL; - const struct cs_extent_def *ext = NULL; + u32 count = 0; if (adev->gfx.rlc.cs_data == NULL) return; if (buffer == NULL) return; - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); - buffer[count++] = cpu_to_le32(0x80000000); - buffer[count++] = cpu_to_le32(0x80000000); - - for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { - for (ext = sect->section; ext->extent != NULL; ++ext) { - if (sect->id == SECT_CONTEXT) { - buffer[count++] = - cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); - buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000); - for (i = 0; i < ext->reg_count; i++) - buffer[count++] = cpu_to_le32(ext->extent[i]); - } else { - return; - } - } - } + count = amdgpu_gfx_csb_preamble_start(buffer); + count = amdgpu_gfx_csb_data_parser(adev, buffer, count); buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1)); buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); - buffer[count++] = cpu_to_le32(0); + amdgpu_gfx_csb_preamble_end(buffer, count); } static void gfx_v6_0_init_pg(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 8181bd0e4f18..da0534ff1271 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -55,6 +55,9 @@ #define GFX7_NUM_GFX_RINGS 1 #define GFX7_MEC_HPD_SIZE 2048 +#define BONAIRE_GB_ADDR_CONFIG_GOLDEN 0x12010001 +#define HAWAII_GB_ADDR_CONFIG_GOLDEN 0x12011003 + static void gfx_v7_0_set_ring_funcs(struct amdgpu_device *adev); static void gfx_v7_0_set_irq_funcs(struct amdgpu_device *adev); static void gfx_v7_0_set_gds_init(struct amdgpu_device *adev); @@ -3882,67 +3885,22 @@ static u32 gfx_v7_0_get_csb_size(struct amdgpu_device *adev) static void gfx_v7_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer) { - u32 count = 0, i; - const struct cs_section_def *sect = NULL; - const struct cs_extent_def *ext = NULL; + u32 count = 0; if (adev->gfx.rlc.cs_data == NULL) return; if (buffer == NULL) return; - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); - buffer[count++] = cpu_to_le32(0x80000000); - buffer[count++] = cpu_to_le32(0x80000000); - - for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { - for (ext = sect->section; ext->extent != NULL; ++ext) { - if (sect->id == SECT_CONTEXT) { - buffer[count++] = - cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); - buffer[count++] = cpu_to_le32(ext->reg_index - PACKET3_SET_CONTEXT_REG_START); - for (i = 0; i < ext->reg_count; i++) - buffer[count++] = cpu_to_le32(ext->extent[i]); - } else { - return; - } - } - } + count = amdgpu_gfx_csb_preamble_start(buffer); + count = amdgpu_gfx_csb_data_parser(adev, buffer, count); buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 2)); buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); - switch (adev->asic_type) { - case CHIP_BONAIRE: - buffer[count++] = cpu_to_le32(0x16000012); - buffer[count++] = cpu_to_le32(0x00000000); - break; - case CHIP_KAVERI: - buffer[count++] = cpu_to_le32(0x00000000); /* XXX */ - buffer[count++] = cpu_to_le32(0x00000000); - break; - case CHIP_KABINI: - case CHIP_MULLINS: - buffer[count++] = cpu_to_le32(0x00000000); /* XXX */ - buffer[count++] = cpu_to_le32(0x00000000); - break; - case CHIP_HAWAII: - buffer[count++] = cpu_to_le32(0x3a00161a); - buffer[count++] = cpu_to_le32(0x0000002e); - break; - default: - buffer[count++] = cpu_to_le32(0x00000000); - buffer[count++] = cpu_to_le32(0x00000000); - break; - } - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); + buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config); + buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config_1); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); - buffer[count++] = cpu_to_le32(0); + amdgpu_gfx_csb_preamble_end(buffer, count); } static void gfx_v7_0_init_pg(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index bfedd487efc5..5ee2237d8ee8 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -1223,48 +1223,22 @@ out: static void gfx_v8_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer) { - u32 count = 0, i; - const struct cs_section_def *sect = NULL; - const struct cs_extent_def *ext = NULL; + u32 count = 0; if (adev->gfx.rlc.cs_data == NULL) return; if (buffer == NULL) return; - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); - buffer[count++] = cpu_to_le32(0x80000000); - buffer[count++] = cpu_to_le32(0x80000000); - - for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { - for (ext = sect->section; ext->extent != NULL; ++ext) { - if (sect->id == SECT_CONTEXT) { - buffer[count++] = - cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); - buffer[count++] = cpu_to_le32(ext->reg_index - - PACKET3_SET_CONTEXT_REG_START); - for (i = 0; i < ext->reg_count; i++) - buffer[count++] = cpu_to_le32(ext->extent[i]); - } else { - return; - } - } - } + count = amdgpu_gfx_csb_preamble_start(buffer); + count = amdgpu_gfx_csb_data_parser(adev, buffer, count); buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - - PACKET3_SET_CONTEXT_REG_START); + buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config); buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config_1); - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); - buffer[count++] = cpu_to_le32(0); + amdgpu_gfx_csb_preamble_end(buffer, count); } static int gfx_v8_0_cp_jump_table_num(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index d7db4cb907ae..d377a7c57d5e 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -225,17 +225,36 @@ static const struct amdgpu_hwip_reg_entry gc_reg_list_9[] = { SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_SAFE_MODE), SOC15_REG_ENTRY_STR(GC, 0, mmRLC_INT_STAT), SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_GENERAL_6), - /* cp header registers */ - SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME2_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), /* SE status registers */ SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE0), SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE1), SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE2), - SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE3) + SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE3), + /* packet headers */ + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP) }; static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_9[] = { @@ -277,6 +296,14 @@ static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_9[] = { SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_LO), SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_HI), SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_GFX_STATUS), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP) }; enum ta_ras_gfx_subblock { @@ -1624,42 +1651,16 @@ static u32 gfx_v9_0_get_csb_size(struct amdgpu_device *adev) static void gfx_v9_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer) { - u32 count = 0, i; - const struct cs_section_def *sect = NULL; - const struct cs_extent_def *ext = NULL; + u32 count = 0; if (adev->gfx.rlc.cs_data == NULL) return; if (buffer == NULL) return; - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); - buffer[count++] = cpu_to_le32(0x80000000); - buffer[count++] = cpu_to_le32(0x80000000); - - for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) { - for (ext = sect->section; ext->extent != NULL; ++ext) { - if (sect->id == SECT_CONTEXT) { - buffer[count++] = - cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); - buffer[count++] = cpu_to_le32(ext->reg_index - - PACKET3_SET_CONTEXT_REG_START); - for (i = 0; i < ext->reg_count; i++) - buffer[count++] = cpu_to_le32(ext->extent[i]); - } else { - return; - } - } - } - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); - - buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); - buffer[count++] = cpu_to_le32(0); + count = amdgpu_gfx_csb_preamble_start(buffer); + count = amdgpu_gfx_csb_data_parser(adev, buffer, count); + amdgpu_gfx_csb_preamble_end(buffer, count); } static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev) @@ -5441,16 +5442,8 @@ static void gfx_v9_0_ring_patch_ce_meta(struct amdgpu_ring *ring, payload_size = sizeof(struct v9_ce_ib_state); - if (ring->is_mes_queue) { - payload_offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gfx_meta_data) + - offsetof(struct v9_gfx_meta_data, ce_payload); - ce_payload_cpu_addr = - amdgpu_mes_ctx_get_offs_cpu_addr(ring, payload_offset); - } else { - payload_offset = offsetof(struct v9_gfx_meta_data, ce_payload); - ce_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset; - } + payload_offset = offsetof(struct v9_gfx_meta_data, ce_payload); + ce_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset; if (offset + (payload_size >> 2) <= ring->buf_mask + 1) { memcpy((void *)&ring->ring[offset], ce_payload_cpu_addr, payload_size); @@ -5473,16 +5466,8 @@ static void gfx_v9_0_ring_patch_de_meta(struct amdgpu_ring *ring, payload_size = sizeof(struct v9_de_ib_state); - if (ring->is_mes_queue) { - payload_offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gfx_meta_data) + - offsetof(struct v9_gfx_meta_data, de_payload); - de_payload_cpu_addr = - amdgpu_mes_ctx_get_offs_cpu_addr(ring, payload_offset); - } else { - payload_offset = offsetof(struct v9_gfx_meta_data, de_payload); - de_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset; - } + payload_offset = offsetof(struct v9_gfx_meta_data, de_payload); + de_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset; ((struct v9_de_ib_state *)de_payload_cpu_addr)->ib_completion_status = IB_COMPLETION_STATUS_PREEMPTED; @@ -5672,19 +5657,9 @@ static void gfx_v9_0_ring_emit_ce_meta(struct amdgpu_ring *ring, bool resume) cnt = (sizeof(ce_payload) >> 2) + 4 - 2; - if (ring->is_mes_queue) { - offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gfx_meta_data) + - offsetof(struct v9_gfx_meta_data, ce_payload); - ce_payload_gpu_addr = - amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ce_payload_cpu_addr = - amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - } else { - offset = offsetof(struct v9_gfx_meta_data, ce_payload); - ce_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset; - ce_payload_cpu_addr = adev->virt.csa_cpu_addr + offset; - } + offset = offsetof(struct v9_gfx_meta_data, ce_payload); + ce_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset; + ce_payload_cpu_addr = adev->virt.csa_cpu_addr + offset; amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, cnt)); amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(2) | @@ -5770,28 +5745,13 @@ static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume, bo void *de_payload_cpu_addr; int cnt; - if (ring->is_mes_queue) { - offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gfx_meta_data) + - offsetof(struct v9_gfx_meta_data, de_payload); - de_payload_gpu_addr = - amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - de_payload_cpu_addr = - amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = offsetof(struct amdgpu_mes_ctx_meta_data, - gfx[0].gds_backup) + - offsetof(struct v9_gfx_meta_data, de_payload); - gds_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - } else { - offset = offsetof(struct v9_gfx_meta_data, de_payload); - de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset; - de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset; + offset = offsetof(struct v9_gfx_meta_data, de_payload); + de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset; + de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset; - gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) + - AMDGPU_CSA_SIZE - adev->gds.gds_size, - PAGE_SIZE); - } + gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) + + AMDGPU_CSA_SIZE - adev->gds.gds_size, + PAGE_SIZE); if (usegds) { de_payload.gds_backup_addrlo = lower_32_bits(gds_addr); @@ -7339,9 +7299,14 @@ static void gfx_v9_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) { drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k); for (reg = 0; reg < reg_count; reg++) { - drm_printf(p, "%-50s \t 0x%08x\n", - gc_cp_reg_list_9[reg].reg_name, - adev->gfx.ip_dump_compute_queues[index + reg]); + if (i && gc_cp_reg_list_9[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP) + drm_printf(p, "%-50s \t 0x%08x\n", + "mmCP_MEC_ME2_HEADER_DUMP", + adev->gfx.ip_dump_compute_queues[index + reg]); + else + drm_printf(p, "%-50s \t 0x%08x\n", + gc_cp_reg_list_9[reg].reg_name, + adev->gfx.ip_dump_compute_queues[index + reg]); } index += reg_count; } @@ -7378,9 +7343,13 @@ static void gfx_v9_ip_dump(struct amdgpu_ip_block *ip_block) soc15_grbm_select(adev, 1 + i, j, k, 0, 0); for (reg = 0; reg < reg_count; reg++) { - adev->gfx.ip_dump_compute_queues[index + reg] = - RREG32(SOC15_REG_ENTRY_OFFSET( - gc_cp_reg_list_9[reg])); + if (i && gc_cp_reg_list_9[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP) + adev->gfx.ip_dump_compute_queues[index + reg] = + RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_MEC_ME2_HEADER_DUMP)); + else + adev->gfx.ip_dump_compute_queues[index + reg] = + RREG32(SOC15_REG_ENTRY_OFFSET( + gc_cp_reg_list_9[reg])); } index += reg_count; } @@ -7394,8 +7363,14 @@ static void gfx_v9_ip_dump(struct amdgpu_ip_block *ip_block) static void gfx_v9_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring) { + struct amdgpu_device *adev = ring->adev; + /* Emit the cleaner shader */ - amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0)); + if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2)) + amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0)); + else + amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER_9_0, 0)); + amdgpu_ring_write(ring, 0); /* RESERVED field, programmed to zero */ } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c index d81449f9d822..c48cd47b531f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c @@ -1547,7 +1547,7 @@ static void gfx_v9_4_2_log_utc_edc_count(struct amdgpu_device *adev, { uint32_t bank, way, mem; static const char * const vml2_way_str[] = { "BIGK", "4K" }; - static const char * const utcl2_rounter_str[] = { "VMC", "APT" }; + static const char * const utcl2_router_str[] = { "VMC", "APT" }; mem = instance % blk->num_mem_blocks; way = (instance / blk->num_mem_blocks) % blk->num_ways; @@ -1568,7 +1568,7 @@ static void gfx_v9_4_2_log_utc_edc_count(struct amdgpu_device *adev, dev_info( adev->dev, "GFX SubBlock UTCL2_ROUTER_IFIF%d_GROUP0_%s, SED %d, DED %d\n", - bank, utcl2_rounter_str[mem], sec_cnt, ded_cnt); + bank, utcl2_router_str[mem], sec_cnt, ded_cnt); break; case ATC_L2_CACHE_2M: dev_info( diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 53fbf6ca7cdb..c233edf60569 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -105,9 +105,6 @@ static const struct amdgpu_hwip_reg_entry gc_reg_list_9_4_3[] = { SOC15_REG_ENTRY_STR(GC, 0, regRLC_SMU_SAFE_MODE), SOC15_REG_ENTRY_STR(GC, 0, regRLC_INT_STAT), SOC15_REG_ENTRY_STR(GC, 0, regRLC_GPM_GENERAL_6), - /* cp header registers */ - SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), - SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME2_HEADER_DUMP), /* SE status registers */ SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0), SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE1), @@ -154,6 +151,14 @@ static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_9_4_3[] = { SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_LO), SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_HI), SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_GFX_STATUS), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), + SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP), }; struct amdgpu_gfx_ras gfx_v9_4_3_ras; @@ -1148,6 +1153,12 @@ static int gfx_v9_4_3_sw_init(struct amdgpu_ip_block *ip_block) adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_PIPE; } break; + case IP_VERSION(9, 5, 0): + if (adev->gfx.mec_fw_version >= 21) { + adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; + adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_PIPE; + } + break; default: break; } @@ -1262,6 +1273,22 @@ static void gfx_v9_4_3_xcc_init_gds_vmid(struct amdgpu_device *adev, int xcc_id) } } +/* For ASICs that needs xnack chain and MEC version supports, set SG_CONFIG1 + * DISABLE_XNACK_CHECK_IN_RETRY_DISABLE bit and inform KFD to set xnack_chain + * bit in SET_RESOURCES + */ +static void gfx_v9_4_3_xcc_init_sq(struct amdgpu_device *adev, int xcc_id) +{ + uint32_t data; + + if (!(adev->gmc.xnack_flags & AMDGPU_GMC_XNACK_FLAG_CHAIN)) + return; + + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_CONFIG1); + data = REG_SET_FIELD(data, SQ_CONFIG1, DISABLE_XNACK_CHECK_IN_RETRY_DISABLE, 1); + WREG32_SOC15(GC, xcc_id, regSQ_CONFIG1, data); +} + static void gfx_v9_4_3_xcc_constants_init(struct amdgpu_device *adev, int xcc_id) { @@ -1306,6 +1333,7 @@ static void gfx_v9_4_3_xcc_constants_init(struct amdgpu_device *adev, gfx_v9_4_3_xcc_init_compute_vmid(adev, xcc_id); gfx_v9_4_3_xcc_init_gds_vmid(adev, xcc_id); + gfx_v9_4_3_xcc_init_sq(adev, xcc_id); } static void gfx_v9_4_3_constants_init(struct amdgpu_device *adev) @@ -1318,6 +1346,20 @@ static void gfx_v9_4_3_constants_init(struct amdgpu_device *adev) adev->gfx.config.db_debug2 = RREG32_SOC15(GC, GET_INST(GC, 0), regDB_DEBUG2); + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + /* ToDo: GC 9.4.4 */ + case IP_VERSION(9, 4, 3): + if (adev->gfx.mec_fw_version >= 184) + adev->gmc.xnack_flags |= AMDGPU_GMC_XNACK_FLAG_CHAIN; + break; + case IP_VERSION(9, 5, 0): + if (adev->gfx.mec_fw_version >= 23) + adev->gmc.xnack_flags |= AMDGPU_GMC_XNACK_FLAG_CHAIN; + break; + default: + break; + } + for (i = 0; i < num_xcc; i++) gfx_v9_4_3_xcc_constants_init(adev, i); } @@ -3447,9 +3489,7 @@ static int gfx_v9_4_3_unmap_done(struct amdgpu_device *adev, uint32_t me, static bool gfx_v9_4_3_pipe_reset_support(struct amdgpu_device *adev) { - /*TODO: Need check gfx9.4.4 mec fw whether supports pipe reset as well.*/ - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) && - adev->gfx.mec_fw_version >= 0x0000009b) + if (!!(adev->gfx.compute_supported_reset & AMDGPU_RESET_TYPE_PER_PIPE)) return true; else dev_warn_once(adev->dev, "Please use the latest MEC version to see whether support pipe reset\n"); @@ -4558,12 +4598,21 @@ static void gfx_v9_4_3_ip_print(struct amdgpu_ip_block *ip_block, struct drm_pri "\nxcc:%d mec:%d, pipe:%d, queue:%d\n", xcc_id, i, j, k); for (reg = 0; reg < reg_count; reg++) { - drm_printf(p, - "%-50s \t 0x%08x\n", - gc_cp_reg_list_9_4_3[reg].reg_name, - adev->gfx.ip_dump_compute_queues - [xcc_offset + inst_offset + - reg]); + if (i && gc_cp_reg_list_9_4_3[reg].reg_offset == + regCP_MEC_ME1_HEADER_DUMP) + drm_printf(p, + "%-50s \t 0x%08x\n", + "regCP_MEC_ME2_HEADER_DUMP", + adev->gfx.ip_dump_compute_queues + [xcc_offset + inst_offset + + reg]); + else + drm_printf(p, + "%-50s \t 0x%08x\n", + gc_cp_reg_list_9_4_3[reg].reg_name, + adev->gfx.ip_dump_compute_queues + [xcc_offset + inst_offset + + reg]); } inst_offset += reg_count; } @@ -4612,12 +4661,20 @@ static void gfx_v9_4_3_ip_dump(struct amdgpu_ip_block *ip_block) GET_INST(GC, xcc_id)); for (reg = 0; reg < reg_count; reg++) { - adev->gfx.ip_dump_compute_queues - [xcc_offset + - inst_offset + reg] = - RREG32(SOC15_REG_ENTRY_OFFSET_INST( - gc_cp_reg_list_9_4_3[reg], - GET_INST(GC, xcc_id))); + if (i && gc_cp_reg_list_9_4_3[reg].reg_offset == + regCP_MEC_ME1_HEADER_DUMP) + adev->gfx.ip_dump_compute_queues + [xcc_offset + + inst_offset + reg] = + RREG32(SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), + regCP_MEC_ME2_HEADER_DUMP)); + else + adev->gfx.ip_dump_compute_queues + [xcc_offset + + inst_offset + reg] = + RREG32(SOC15_REG_ENTRY_OFFSET_INST( + gc_cp_reg_list_9_4_3[reg], + GET_INST(GC, xcc_id))); } inst_offset += reg_count; } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index 809b3a882d0d..a3e2787501f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -428,10 +428,6 @@ static void gmc_v10_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int struct amdgpu_device *adev = ring->adev; uint32_t reg; - /* MES fw manages IH_VMID_x_LUT updating */ - if (ring->is_mes_queue) - return; - if (ring->vm_hub == AMDGPU_GFXHUB(0)) reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid; else diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c index e74e26b6a4f2..72211409227b 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c @@ -393,10 +393,6 @@ static void gmc_v11_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int struct amdgpu_device *adev = ring->adev; uint32_t reg; - /* MES fw manages IH_VMID_x_LUT updating */ - if (ring->is_mes_queue) - return; - if (ring->vm_hub == AMDGPU_GFXHUB(0)) reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid; else @@ -752,6 +748,18 @@ static int gmc_v11_0_sw_init(struct amdgpu_ip_block *ip_block) adev->gmc.vram_type = vram_type; adev->gmc.vram_vendor = vram_vendor; + /* The mall_size is already calculated as mall_size_per_umc * num_umc. + * However, for gfx1151, which features a 2-to-1 UMC mapping, + * the result must be multiplied by 2 to determine the actual mall size. + */ + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + case IP_VERSION(11, 5, 1): + adev->gmc.mall_size *= 2; + break; + default: + break; + } + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(11, 0, 0): case IP_VERSION(11, 0, 1): @@ -832,7 +840,7 @@ static int gmc_v11_0_sw_init(struct amdgpu_ip_block *ip_block) * amdgpu graphics/compute will use VMIDs 1-7 * amdkfd will use VMIDs 8-15 */ - adev->vm_manager.first_kfd_vmid = 8; + adev->vm_manager.first_kfd_vmid = adev->gfx.disable_kq ? 1 : 8; amdgpu_vm_manager_init(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c index c6f290704d47..b645d3e6a6c8 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c @@ -413,10 +413,6 @@ static void gmc_v12_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid struct amdgpu_device *adev = ring->adev; uint32_t reg; - /* MES fw manages IH_VMID_x_LUT updating */ - if (ring->is_mes_queue) - return; - if (ring->vm_hub == AMDGPU_GFXHUB(0)) reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid; else @@ -820,7 +816,7 @@ static int gmc_v12_0_sw_init(struct amdgpu_ip_block *ip_block) * amdgpu graphics/compute will use VMIDs 1-7 * amdkfd will use VMIDs 8-15 */ - adev->vm_manager.first_kfd_vmid = 8; + adev->vm_manager.first_kfd_vmid = adev->gfx.disable_kq ? 1 : 8; amdgpu_vm_manager_init(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index a992e79d9581..8030fcd64210 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -249,7 +249,7 @@ static void gmc_v6_0_mc_program(struct amdgpu_device *adev) /* disable VGA render */ tmp = RREG32(mmVGA_RENDER_CONTROL); - tmp &= ~VGA_VSTATUS_CNTL; + tmp &= VGA_RENDER_CONTROL__VGA_VSTATUS_CNTL_MASK; WREG32(mmVGA_RENDER_CONTROL, tmp); } /* Update configuration */ @@ -627,17 +627,16 @@ static void gmc_v6_0_vm_decode_fault(struct amdgpu_device *adev, "write" : "read", block, mc_client, mc_id); } -/* static const u32 mc_cg_registers[] = { - MC_HUB_MISC_HUB_CG, - MC_HUB_MISC_SIP_CG, - MC_HUB_MISC_VM_CG, - MC_XPB_CLK_GAT, - ATC_MISC_CG, - MC_CITF_MISC_WR_CG, - MC_CITF_MISC_RD_CG, - MC_CITF_MISC_VM_CG, - VM_L2_CG, + mmMC_HUB_MISC_HUB_CG, + mmMC_HUB_MISC_SIP_CG, + mmMC_HUB_MISC_VM_CG, + mmMC_XPB_CLK_GAT, + mmATC_MISC_CG, + mmMC_CITF_MISC_WR_CG, + mmMC_CITF_MISC_RD_CG, + mmMC_CITF_MISC_VM_CG, + mmVM_L2_CG, }; static const u32 mc_cg_ls_en[] = { @@ -672,7 +671,7 @@ static void gmc_v6_0_enable_mc_ls(struct amdgpu_device *adev, for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { orig = data = RREG32(mc_cg_registers[i]); - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_MC_LS)) + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_LS)) data |= mc_cg_ls_en[i]; else data &= ~mc_cg_ls_en[i]; @@ -689,7 +688,7 @@ static void gmc_v6_0_enable_mc_mgcg(struct amdgpu_device *adev, for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { orig = data = RREG32(mc_cg_registers[i]); - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_MC_MGCG)) + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)) data |= mc_cg_en[i]; else data &= ~mc_cg_en[i]; @@ -705,7 +704,7 @@ static void gmc_v6_0_enable_bif_mgls(struct amdgpu_device *adev, orig = data = RREG32_PCIE(ixPCIE_CNTL2); - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_BIF_LS)) { + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_BIF_LS)) { data = REG_SET_FIELD(data, PCIE_CNTL2, SLV_MEM_LS_EN, 1); data = REG_SET_FIELD(data, PCIE_CNTL2, MST_MEM_LS_EN, 1); data = REG_SET_FIELD(data, PCIE_CNTL2, REPLAY_MEM_LS_EN, 1); @@ -728,7 +727,7 @@ static void gmc_v6_0_enable_hdp_mgcg(struct amdgpu_device *adev, orig = data = RREG32(mmHDP_HOST_PATH_CNTL); - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_HDP_MGCG)) + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_MGCG)) data = REG_SET_FIELD(data, HDP_HOST_PATH_CNTL, CLOCK_GATING_DIS, 0); else data = REG_SET_FIELD(data, HDP_HOST_PATH_CNTL, CLOCK_GATING_DIS, 1); @@ -744,7 +743,7 @@ static void gmc_v6_0_enable_hdp_ls(struct amdgpu_device *adev, orig = data = RREG32(mmHDP_MEM_POWER_LS); - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_HDP_LS)) + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS)) data = REG_SET_FIELD(data, HDP_MEM_POWER_LS, LS_ENABLE, 1); else data = REG_SET_FIELD(data, HDP_MEM_POWER_LS, LS_ENABLE, 0); @@ -752,7 +751,6 @@ static void gmc_v6_0_enable_hdp_ls(struct amdgpu_device *adev, if (orig != data) WREG32(mmHDP_MEM_POWER_LS, data); } -*/ static int gmc_v6_0_convert_vram_type(int mc_seq_vram_type) { @@ -1098,6 +1096,20 @@ static int gmc_v6_0_process_interrupt(struct amdgpu_device *adev, static int gmc_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, enum amd_clockgating_state state) { + struct amdgpu_device *adev = ip_block->adev; + bool gate = false; + + if (state == AMD_CG_STATE_GATE) + gate = true; + + if (!(adev->flags & AMD_IS_APU)) { + gmc_v6_0_enable_mc_mgcg(adev, gate); + gmc_v6_0_enable_mc_ls(adev, gate); + } + gmc_v6_0_enable_bif_mgls(adev, gate); + gmc_v6_0_enable_hdp_mgcg(adev, gate); + gmc_v6_0_enable_hdp_ls(adev, gate); + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 83e39f16044a..a8d5795084fc 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -1157,17 +1157,10 @@ static bool gmc_v7_0_is_idle(struct amdgpu_ip_block *ip_block) static int gmc_v7_0_wait_for_idle(struct amdgpu_ip_block *ip_block) { unsigned int i; - u32 tmp; struct amdgpu_device *adev = ip_block->adev; for (i = 0; i < adev->usec_timeout; i++) { - /* read MC_STATUS */ - tmp = RREG32(mmSRBM_STATUS) & (SRBM_STATUS__MCB_BUSY_MASK | - SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK | - SRBM_STATUS__MCC_BUSY_MASK | - SRBM_STATUS__MCD_BUSY_MASK | - SRBM_STATUS__VMC_BUSY_MASK); - if (!tmp) + if (gmc_v7_0_is_idle(ip_block)) return 0; udelay(1); } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 5effe8327d29..282197f4ffb1 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -1213,10 +1213,7 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev, if (uncached) { mtype = MTYPE_UC; } else if (ext_coherent) { - if (gc_ip_version == IP_VERSION(9, 5, 0) || adev->rev_id) - mtype = is_local ? MTYPE_CC : MTYPE_UC; - else - mtype = MTYPE_UC; + mtype = is_local ? MTYPE_CC : MTYPE_UC; } else if (adev->flags & AMD_IS_APU) { mtype = is_local ? mtype_local : MTYPE_NC; } else { @@ -1336,7 +1333,7 @@ static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev, mtype_local = MTYPE_CC; *flags = AMDGPU_PTE_MTYPE_VG10(*flags, mtype_local); - } else if (adev->rev_id) { + } else { /* MTYPE_UC case */ *flags = AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_CC); } @@ -1505,7 +1502,6 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev) adev->umc.umc_inst_num = UMC_V12_0_UMC_INSTANCE_NUM; adev->umc.node_inst_num /= UMC_V12_0_UMC_INSTANCE_NUM; adev->umc.channel_offs = UMC_V12_0_PER_CHANNEL_OFFSET; - adev->umc.retire_unit = UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL; if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) adev->umc.ras = &umc_v12_0_ras; break; @@ -2075,6 +2071,9 @@ static void gmc_v9_4_3_init_vram_info(struct amdgpu_device *adev) { adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM; adev->gmc.vram_width = 128 * 64; + + if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM3E; } static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block) @@ -2411,13 +2410,6 @@ static int gmc_v9_0_hw_init(struct amdgpu_ip_block *ip_block) adev->gmc.flush_tlb_needs_extra_type_2 = amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 0) && adev->gmc.xgmi.num_physical_nodes; - /* - * TODO: This workaround is badly documented and had a buggy - * implementation. We should probably verify what we do here. - */ - adev->gmc.flush_tlb_needs_extra_type_0 = - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) && - adev->rev_id == 0; /* The sequence of these two function calls matters.*/ gmc_v9_0_init_golden_registers(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c index f1dc13b3ab38..e6c0d86d3486 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c @@ -36,17 +36,6 @@ #define HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK 0x00020000L #define mmHDP_MEM_POWER_CTRL_BASE_IDX 0 -static void hdp_v4_0_flush_hdp(struct amdgpu_device *adev, - struct amdgpu_ring *ring) -{ - if (!ring || !ring->funcs->emit_wreg) { - WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2); - } else { - amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - } -} - static void hdp_v4_0_invalidate_hdp(struct amdgpu_device *adev, struct amdgpu_ring *ring) { @@ -180,7 +169,7 @@ struct amdgpu_hdp_ras hdp_v4_0_ras = { }; const struct amdgpu_hdp_funcs hdp_v4_0_funcs = { - .flush_hdp = hdp_v4_0_flush_hdp, + .flush_hdp = amdgpu_hdp_generic_flush, .invalidate_hdp = hdp_v4_0_invalidate_hdp, .update_clock_gating = hdp_v4_0_update_clock_gating, .get_clock_gating_state = hdp_v4_0_get_clockgating_state, diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c index 43195c079748..8bc001dc9f63 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c @@ -27,17 +27,6 @@ #include "hdp/hdp_5_0_0_sh_mask.h" #include <uapi/linux/kfd_ioctl.h> -static void hdp_v5_0_flush_hdp(struct amdgpu_device *adev, - struct amdgpu_ring *ring) -{ - if (!ring || !ring->funcs->emit_wreg) { - WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2); - } else { - amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - } -} - static void hdp_v5_0_invalidate_hdp(struct amdgpu_device *adev, struct amdgpu_ring *ring) { @@ -217,7 +206,7 @@ static void hdp_v5_0_init_registers(struct amdgpu_device *adev) } const struct amdgpu_hdp_funcs hdp_v5_0_funcs = { - .flush_hdp = hdp_v5_0_flush_hdp, + .flush_hdp = amdgpu_hdp_generic_flush, .invalidate_hdp = hdp_v5_0_invalidate_hdp, .update_clock_gating = hdp_v5_0_update_clock_gating, .get_clock_gating_state = hdp_v5_0_get_clockgating_state, diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c index fcb8dd2876bc..40940b4ab400 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c @@ -33,7 +33,17 @@ static void hdp_v5_2_flush_hdp(struct amdgpu_device *adev, if (!ring || !ring->funcs->emit_wreg) { WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - RREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2); + if (amdgpu_sriov_vf(adev)) { + /* this is fine because SR_IOV doesn't remap the register */ + RREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2); + } else { + /* We just need to read back a register to post the write. + * Reading back the remapped register causes problems on + * some platforms so just read back the memory size register. + */ + if (adev->nbio.funcs->get_memsize) + adev->nbio.funcs->get_memsize(adev); + } } else { amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c index a88d25a06c29..ec20daf4272c 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c @@ -30,17 +30,6 @@ #define regHDP_CLK_CNTL_V6_1 0xd5 #define regHDP_CLK_CNTL_V6_1_BASE_IDX 0 -static void hdp_v6_0_flush_hdp(struct amdgpu_device *adev, - struct amdgpu_ring *ring) -{ - if (!ring || !ring->funcs->emit_wreg) { - WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2); - } else { - amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - } -} - static void hdp_v6_0_update_clock_gating(struct amdgpu_device *adev, bool enable) { @@ -149,7 +138,7 @@ static void hdp_v6_0_get_clockgating_state(struct amdgpu_device *adev, } const struct amdgpu_hdp_funcs hdp_v6_0_funcs = { - .flush_hdp = hdp_v6_0_flush_hdp, + .flush_hdp = amdgpu_hdp_generic_flush, .update_clock_gating = hdp_v6_0_update_clock_gating, .get_clock_gating_state = hdp_v6_0_get_clockgating_state, }; diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c index 49f7eb4fbd11..ed1debc03507 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c @@ -27,17 +27,6 @@ #include "hdp/hdp_7_0_0_sh_mask.h" #include <uapi/linux/kfd_ioctl.h> -static void hdp_v7_0_flush_hdp(struct amdgpu_device *adev, - struct amdgpu_ring *ring) -{ - if (!ring || !ring->funcs->emit_wreg) { - WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2); - } else { - amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0); - } -} - static void hdp_v7_0_update_clock_gating(struct amdgpu_device *adev, bool enable) { @@ -137,7 +126,7 @@ static void hdp_v7_0_get_clockgating_state(struct amdgpu_device *adev, } const struct amdgpu_hdp_funcs hdp_v7_0_funcs = { - .flush_hdp = hdp_v7_0_flush_hdp, + .flush_hdp = amdgpu_hdp_generic_flush, .update_clock_gating = hdp_v7_0_update_clock_gating, .get_clock_gating_state = hdp_v7_0_get_clockgating_state, }; diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c index eb4185dcbd1d..5900b560b7de 100644 --- a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c @@ -349,6 +349,7 @@ static int ih_v6_0_irq_init(struct amdgpu_device *adev) if (ret) return ret; } + ih[i]->overflow = false; } /* update doorbell range for ih ring 0 */ @@ -446,7 +447,10 @@ static u32 ih_v6_0_get_wptr(struct amdgpu_device *adev, wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr); if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) goto out; - wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); + if (!amdgpu_sriov_vf(adev)) + wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); + else + ih->overflow = true; /* When a ring buffer overflow happen start parsing interrupt * from the last not overwritten vector (wptr + 32). Hopefully diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c index 218e16b68f1d..cb94bd71300f 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c @@ -28,11 +28,13 @@ #include "soc15d.h" #include "jpeg_v4_0_3.h" #include "jpeg_v5_0_1.h" +#include "mmsch_v5_0.h" #include "vcn/vcn_5_0_0_offset.h" #include "vcn/vcn_5_0_0_sh_mask.h" #include "ivsrcid/vcn/irqsrcs_vcn_5_0.h" +static int jpeg_v5_0_1_start_sriov(struct amdgpu_device *adev); static void jpeg_v5_0_1_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v5_0_1_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block, @@ -156,21 +158,16 @@ static int jpeg_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block) for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { ring = &adev->jpeg.inst[i].ring_dec[j]; - ring->use_doorbell = false; + ring->use_doorbell = true; ring->vm_hub = AMDGPU_MMHUB0(adev->jpeg.inst[i].aid_id); if (!amdgpu_sriov_vf(adev)) { ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + j + 11 * jpeg_inst; } else { - if (j < 4) - ring->doorbell_index = - (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + - 4 + j + 32 * jpeg_inst; - else - ring->doorbell_index = - (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + - 8 + j + 32 * jpeg_inst; + ring->doorbell_index = + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 2 + j + 32 * jpeg_inst; } sprintf(ring->name, "jpeg_dec_%d.%d", adev->jpeg.inst[i].aid_id, j); r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0, @@ -237,7 +234,10 @@ static int jpeg_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block) int i, j, r, jpeg_inst; if (amdgpu_sriov_vf(adev)) { - /* jpeg_v5_0_1_start_sriov(adev); */ + r = jpeg_v5_0_1_start_sriov(adev); + if (r) + return r; + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { ring = &adev->jpeg.inst[i].ring_dec[j]; @@ -264,7 +264,7 @@ static int jpeg_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block) ring = &adev->jpeg.inst[i].ring_dec[j]; if (ring->use_doorbell) WREG32_SOC15_OFFSET(VCN, GET_INST(VCN, i), regVCN_JPEG_DB_CTRL, - (ring->pipe ? (ring->pipe - 0x15) : 0), + ring->pipe, ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT | VCN_JPEG_DB_CTRL__EN_MASK); @@ -291,8 +291,10 @@ static int jpeg_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block) cancel_delayed_work_sync(&adev->jpeg.idle_work); - if (adev->jpeg.cur_state != AMD_PG_STATE_GATE) - ret = jpeg_v5_0_1_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + if (!amdgpu_sriov_vf(adev)) { + if (adev->jpeg.cur_state != AMD_PG_STATE_GATE) + ret = jpeg_v5_0_1_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + } return ret; } @@ -422,6 +424,119 @@ static void jpeg_v5_0_1_init_jrbc(struct amdgpu_ring *ring) reg_offset); } +static int jpeg_v5_0_1_start_sriov(struct amdgpu_device *adev) +{ + struct amdgpu_ring *ring; + uint64_t ctx_addr; + uint32_t param, resp, expected; + uint32_t tmp, timeout; + + struct amdgpu_mm_table *table = &adev->virt.mm_table; + uint32_t *table_loc; + uint32_t table_size; + uint32_t size, size_dw, item_offset; + uint32_t init_status; + int i, j, jpeg_inst; + + struct mmsch_v5_0_cmd_direct_write + direct_wt = { {0} }; + struct mmsch_v5_0_cmd_end end = { {0} }; + struct mmsch_v5_0_init_header header; + + direct_wt.cmd_header.command_type = + MMSCH_COMMAND__DIRECT_REG_WRITE; + end.cmd_header.command_type = + MMSCH_COMMAND__END; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) { + jpeg_inst = GET_INST(JPEG, i); + + memset(&header, 0, sizeof(struct mmsch_v5_0_init_header)); + header.version = MMSCH_VERSION; + header.total_size = sizeof(struct mmsch_v5_0_init_header) >> 2; + + table_loc = (uint32_t *)table->cpu_addr; + table_loc += header.total_size; + + item_offset = header.total_size; + + for (j = 0; j < adev->jpeg.num_jpeg_rings; j++) { + ring = &adev->jpeg.inst[i].ring_dec[j]; + table_size = 0; + + tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW); + MMSCH_V5_0_INSERT_DIRECT_WT(tmp, lower_32_bits(ring->gpu_addr)); + tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH); + MMSCH_V5_0_INSERT_DIRECT_WT(tmp, upper_32_bits(ring->gpu_addr)); + tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_JRBC_RB_SIZE); + MMSCH_V5_0_INSERT_DIRECT_WT(tmp, ring->ring_size / 4); + + if (j < 5) { + header.mjpegdec0[j].table_offset = item_offset; + header.mjpegdec0[j].init_status = 0; + header.mjpegdec0[j].table_size = table_size; + } else { + header.mjpegdec1[j - 5].table_offset = item_offset; + header.mjpegdec1[j - 5].init_status = 0; + header.mjpegdec1[j - 5].table_size = table_size; + } + header.total_size += table_size; + item_offset += table_size; + } + + MMSCH_V5_0_INSERT_END(); + + /* send init table to MMSCH */ + size = sizeof(struct mmsch_v5_0_init_header); + table_loc = (uint32_t *)table->cpu_addr; + memcpy((void *)table_loc, &header, size); + + ctx_addr = table->gpu_addr; + WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr)); + WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr)); + + tmp = RREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_VMID); + tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK; + tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT); + WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_VMID, tmp); + + size = header.total_size; + WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_SIZE, size); + + WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_RESP, 0); + + param = 0x00000001; + WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_HOST, param); + tmp = 0; + timeout = 1000; + resp = 0; + expected = MMSCH_VF_MAILBOX_RESP__OK; + init_status = + ((struct mmsch_v5_0_init_header *)(table_loc))->mjpegdec0[i].init_status; + while (resp != expected) { + resp = RREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_RESP); + + if (resp != 0) + break; + udelay(10); + tmp = tmp + 10; + if (tmp >= timeout) { + DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\ + " waiting for regMMSCH_VF_MAILBOX_RESP "\ + "(expected=0x%08x, readback=0x%08x)\n", + tmp, expected, resp); + return -EBUSY; + } + } + if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE && + init_status != MMSCH_VF_ENGINE_STATUS__PASS) + DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init status for jpeg: %x\n", + resp, init_status); + + } + return 0; +} + /** * jpeg_v5_0_1_start - start JPEG block * @@ -581,6 +696,11 @@ static int jpeg_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block, struct amdgpu_device *adev = ip_block->adev; int ret; + if (amdgpu_sriov_vf(adev)) { + adev->jpeg.cur_state = AMD_PG_STATE_UNGATE; + return 0; + } + if (state == adev->jpeg.cur_state) return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c new file mode 100644 index 000000000000..d6f50b13e2ba --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright 2024 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#include "amdgpu.h" +#include "amdgpu_gfx.h" +#include "mes_userqueue.h" +#include "amdgpu_userq_fence.h" + +#define AMDGPU_USERQ_PROC_CTX_SZ PAGE_SIZE +#define AMDGPU_USERQ_GANG_CTX_SZ PAGE_SIZE + +static int +mes_userq_map_gtt_bo_to_gart(struct amdgpu_bo *bo) +{ + int ret; + + ret = amdgpu_bo_reserve(bo, true); + if (ret) { + DRM_ERROR("Failed to reserve bo. ret %d\n", ret); + goto err_reserve_bo_failed; + } + + ret = amdgpu_ttm_alloc_gart(&bo->tbo); + if (ret) { + DRM_ERROR("Failed to bind bo to GART. ret %d\n", ret); + goto err_map_bo_gart_failed; + } + + amdgpu_bo_unreserve(bo); + bo = amdgpu_bo_ref(bo); + + return 0; + +err_map_bo_gart_failed: + amdgpu_bo_unreserve(bo); +err_reserve_bo_failed: + return ret; +} + +static int +mes_userq_create_wptr_mapping(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue, + uint64_t wptr) +{ + struct amdgpu_bo_va_mapping *wptr_mapping; + struct amdgpu_vm *wptr_vm; + struct amdgpu_userq_obj *wptr_obj = &queue->wptr_obj; + int ret; + + wptr_vm = queue->vm; + ret = amdgpu_bo_reserve(wptr_vm->root.bo, false); + if (ret) + return ret; + + wptr &= AMDGPU_GMC_HOLE_MASK; + wptr_mapping = amdgpu_vm_bo_lookup_mapping(wptr_vm, wptr >> PAGE_SHIFT); + amdgpu_bo_unreserve(wptr_vm->root.bo); + if (!wptr_mapping) { + DRM_ERROR("Failed to lookup wptr bo\n"); + return -EINVAL; + } + + wptr_obj->obj = wptr_mapping->bo_va->base.bo; + if (wptr_obj->obj->tbo.base.size > PAGE_SIZE) { + DRM_ERROR("Requested GART mapping for wptr bo larger than one page\n"); + return -EINVAL; + } + + ret = mes_userq_map_gtt_bo_to_gart(wptr_obj->obj); + if (ret) { + DRM_ERROR("Failed to map wptr bo to GART\n"); + return ret; + } + + queue->wptr_obj.gpu_addr = amdgpu_bo_gpu_offset_no_check(wptr_obj->obj); + return 0; +} + +static int convert_to_mes_priority(int priority) +{ + switch (priority) { + case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_NORMAL_LOW: + default: + return AMDGPU_MES_PRIORITY_LEVEL_NORMAL; + case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_LOW: + return AMDGPU_MES_PRIORITY_LEVEL_LOW; + case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_NORMAL_HIGH: + return AMDGPU_MES_PRIORITY_LEVEL_MEDIUM; + case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_HIGH: + return AMDGPU_MES_PRIORITY_LEVEL_HIGH; + } +} + +static int mes_userq_map(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + struct amdgpu_device *adev = uq_mgr->adev; + struct amdgpu_userq_obj *ctx = &queue->fw_obj; + struct amdgpu_mqd_prop *userq_props = queue->userq_prop; + struct mes_add_queue_input queue_input; + int r; + + memset(&queue_input, 0x0, sizeof(struct mes_add_queue_input)); + + queue_input.process_va_start = 0; + queue_input.process_va_end = adev->vm_manager.max_pfn - 1; + + /* set process quantum to 10 ms and gang quantum to 1 ms as default */ + queue_input.process_quantum = 100000; + queue_input.gang_quantum = 10000; + queue_input.paging = false; + + queue_input.process_context_addr = ctx->gpu_addr; + queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ; + queue_input.inprocess_gang_priority = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; + queue_input.gang_global_priority_level = convert_to_mes_priority(queue->priority); + + queue_input.process_id = queue->vm->pasid; + queue_input.queue_type = queue->queue_type; + queue_input.mqd_addr = queue->mqd.gpu_addr; + queue_input.wptr_addr = userq_props->wptr_gpu_addr; + queue_input.queue_size = userq_props->queue_size >> 2; + queue_input.doorbell_offset = userq_props->doorbell_index; + queue_input.page_table_base_addr = amdgpu_gmc_pd_addr(queue->vm->root.bo); + queue_input.wptr_mc_addr = queue->wptr_obj.gpu_addr; + + amdgpu_mes_lock(&adev->mes); + r = adev->mes.funcs->add_hw_queue(&adev->mes, &queue_input); + amdgpu_mes_unlock(&adev->mes); + if (r) { + DRM_ERROR("Failed to map queue in HW, err (%d)\n", r); + return r; + } + + DRM_DEBUG_DRIVER("Queue (doorbell:%d) mapped successfully\n", userq_props->doorbell_index); + return 0; +} + +static int mes_userq_unmap(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + struct amdgpu_device *adev = uq_mgr->adev; + struct mes_remove_queue_input queue_input; + struct amdgpu_userq_obj *ctx = &queue->fw_obj; + int r; + + memset(&queue_input, 0x0, sizeof(struct mes_remove_queue_input)); + queue_input.doorbell_offset = queue->doorbell_index; + queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ; + + amdgpu_mes_lock(&adev->mes); + r = adev->mes.funcs->remove_hw_queue(&adev->mes, &queue_input); + amdgpu_mes_unlock(&adev->mes); + if (r) + DRM_ERROR("Failed to unmap queue in HW, err (%d)\n", r); + return r; +} + +static int mes_userq_create_ctx_space(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue, + struct drm_amdgpu_userq_in *mqd_user) +{ + struct amdgpu_userq_obj *ctx = &queue->fw_obj; + int r, size; + + /* + * The FW expects at least one page space allocated for + * process ctx and gang ctx each. Create an object + * for the same. + */ + size = AMDGPU_USERQ_PROC_CTX_SZ + AMDGPU_USERQ_GANG_CTX_SZ; + r = amdgpu_userq_create_object(uq_mgr, ctx, size); + if (r) { + DRM_ERROR("Failed to allocate ctx space bo for userqueue, err:%d\n", r); + return r; + } + + return 0; +} + +static int mes_userq_mqd_create(struct amdgpu_userq_mgr *uq_mgr, + struct drm_amdgpu_userq_in *args_in, + struct amdgpu_usermode_queue *queue) +{ + struct amdgpu_device *adev = uq_mgr->adev; + struct amdgpu_mqd *mqd_hw_default = &adev->mqds[queue->queue_type]; + struct drm_amdgpu_userq_in *mqd_user = args_in; + struct amdgpu_mqd_prop *userq_props; + int r; + + /* Structure to initialize MQD for userqueue using generic MQD init function */ + userq_props = kzalloc(sizeof(struct amdgpu_mqd_prop), GFP_KERNEL); + if (!userq_props) { + DRM_ERROR("Failed to allocate memory for userq_props\n"); + return -ENOMEM; + } + + if (!mqd_user->wptr_va || !mqd_user->rptr_va || + !mqd_user->queue_va || mqd_user->queue_size == 0) { + DRM_ERROR("Invalid MQD parameters for userqueue\n"); + r = -EINVAL; + goto free_props; + } + + r = amdgpu_userq_create_object(uq_mgr, &queue->mqd, mqd_hw_default->mqd_size); + if (r) { + DRM_ERROR("Failed to create MQD object for userqueue\n"); + goto free_props; + } + + /* Initialize the MQD BO with user given values */ + userq_props->wptr_gpu_addr = mqd_user->wptr_va; + userq_props->rptr_gpu_addr = mqd_user->rptr_va; + userq_props->queue_size = mqd_user->queue_size; + userq_props->hqd_base_gpu_addr = mqd_user->queue_va; + userq_props->mqd_gpu_addr = queue->mqd.gpu_addr; + userq_props->use_doorbell = true; + userq_props->doorbell_index = queue->doorbell_index; + userq_props->fence_address = queue->fence_drv->gpu_addr; + + if (queue->queue_type == AMDGPU_HW_IP_COMPUTE) { + struct drm_amdgpu_userq_mqd_compute_gfx11 *compute_mqd; + + if (mqd_user->mqd_size != sizeof(*compute_mqd)) { + DRM_ERROR("Invalid compute IP MQD size\n"); + r = -EINVAL; + goto free_mqd; + } + + compute_mqd = memdup_user(u64_to_user_ptr(mqd_user->mqd), mqd_user->mqd_size); + if (IS_ERR(compute_mqd)) { + DRM_ERROR("Failed to read user MQD\n"); + r = -ENOMEM; + goto free_mqd; + } + + userq_props->eop_gpu_addr = compute_mqd->eop_va; + userq_props->hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_NORMAL; + userq_props->hqd_queue_priority = AMDGPU_GFX_QUEUE_PRIORITY_MINIMUM; + userq_props->hqd_active = false; + userq_props->tmz_queue = + mqd_user->flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE; + kfree(compute_mqd); + } else if (queue->queue_type == AMDGPU_HW_IP_GFX) { + struct drm_amdgpu_userq_mqd_gfx11 *mqd_gfx_v11; + + if (mqd_user->mqd_size != sizeof(*mqd_gfx_v11) || !mqd_user->mqd) { + DRM_ERROR("Invalid GFX MQD\n"); + r = -EINVAL; + goto free_mqd; + } + + mqd_gfx_v11 = memdup_user(u64_to_user_ptr(mqd_user->mqd), mqd_user->mqd_size); + if (IS_ERR(mqd_gfx_v11)) { + DRM_ERROR("Failed to read user MQD\n"); + r = -ENOMEM; + goto free_mqd; + } + + userq_props->shadow_addr = mqd_gfx_v11->shadow_va; + userq_props->csa_addr = mqd_gfx_v11->csa_va; + userq_props->tmz_queue = + mqd_user->flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE; + kfree(mqd_gfx_v11); + } else if (queue->queue_type == AMDGPU_HW_IP_DMA) { + struct drm_amdgpu_userq_mqd_sdma_gfx11 *mqd_sdma_v11; + + if (mqd_user->mqd_size != sizeof(*mqd_sdma_v11) || !mqd_user->mqd) { + DRM_ERROR("Invalid SDMA MQD\n"); + r = -EINVAL; + goto free_mqd; + } + + mqd_sdma_v11 = memdup_user(u64_to_user_ptr(mqd_user->mqd), mqd_user->mqd_size); + if (IS_ERR(mqd_sdma_v11)) { + DRM_ERROR("Failed to read sdma user MQD\n"); + r = -ENOMEM; + goto free_mqd; + } + + userq_props->csa_addr = mqd_sdma_v11->csa_va; + kfree(mqd_sdma_v11); + } + + queue->userq_prop = userq_props; + + r = mqd_hw_default->init_mqd(adev, (void *)queue->mqd.cpu_ptr, userq_props); + if (r) { + DRM_ERROR("Failed to initialize MQD for userqueue\n"); + goto free_mqd; + } + + /* Create BO for FW operations */ + r = mes_userq_create_ctx_space(uq_mgr, queue, mqd_user); + if (r) { + DRM_ERROR("Failed to allocate BO for userqueue (%d)", r); + goto free_mqd; + } + + /* FW expects WPTR BOs to be mapped into GART */ + r = mes_userq_create_wptr_mapping(uq_mgr, queue, userq_props->wptr_gpu_addr); + if (r) { + DRM_ERROR("Failed to create WPTR mapping\n"); + goto free_ctx; + } + + return 0; + +free_ctx: + amdgpu_userq_destroy_object(uq_mgr, &queue->fw_obj); + +free_mqd: + amdgpu_userq_destroy_object(uq_mgr, &queue->mqd); + +free_props: + kfree(userq_props); + + return r; +} + +static void +mes_userq_mqd_destroy(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + amdgpu_userq_destroy_object(uq_mgr, &queue->fw_obj); + kfree(queue->userq_prop); + amdgpu_userq_destroy_object(uq_mgr, &queue->mqd); +} + +const struct amdgpu_userq_funcs userq_mes_funcs = { + .mqd_create = mes_userq_mqd_create, + .mqd_destroy = mes_userq_mqd_destroy, + .unmap = mes_userq_unmap, + .map = mes_userq_map, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ad102.c b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.h index 7ac87ef26aec..090ae8897770 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ad102.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.h @@ -1,5 +1,6 @@ +/* SPDX-License-Identifier: MIT */ /* - * Copyright 2023 Red Hat Inc. + * Copyright 2024 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -18,27 +19,12 @@ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. + * */ -#include "priv.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_engine_func -ad102_ofa = { - .sclass = { - { -1, -1, NVC9FA_VIDEO_OFA }, - {} - } -}; -int -ad102_ofa_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_engine **pengine) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_ofa_new(&ad102_ofa, device, type, inst, pengine); +#ifndef MES_USERQ_H +#define MES_USERQ_H +#include "amdgpu_userq.h" - return -ENODEV; -} +extern const struct amdgpu_userq_funcs userq_mes_funcs; +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index ef9538fbbf53..c9eba537de09 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -287,6 +287,23 @@ static int convert_to_mes_queue_type(int queue_type) return -1; } +static int convert_to_mes_priority_level(int priority_level) +{ + switch (priority_level) { + case AMDGPU_MES_PRIORITY_LEVEL_LOW: + return AMD_PRIORITY_LEVEL_LOW; + case AMDGPU_MES_PRIORITY_LEVEL_NORMAL: + default: + return AMD_PRIORITY_LEVEL_NORMAL; + case AMDGPU_MES_PRIORITY_LEVEL_MEDIUM: + return AMD_PRIORITY_LEVEL_MEDIUM; + case AMDGPU_MES_PRIORITY_LEVEL_HIGH: + return AMD_PRIORITY_LEVEL_HIGH; + case AMDGPU_MES_PRIORITY_LEVEL_REALTIME: + return AMD_PRIORITY_LEVEL_REALTIME; + } +} + static int mes_v11_0_add_hw_queue(struct amdgpu_mes *mes, struct mes_add_queue_input *input) { @@ -310,9 +327,9 @@ static int mes_v11_0_add_hw_queue(struct amdgpu_mes *mes, mes_add_queue_pkt.gang_quantum = input->gang_quantum; mes_add_queue_pkt.gang_context_addr = input->gang_context_addr; mes_add_queue_pkt.inprocess_gang_priority = - input->inprocess_gang_priority; + convert_to_mes_priority_level(input->inprocess_gang_priority); mes_add_queue_pkt.gang_global_priority_level = - input->gang_global_priority_level; + convert_to_mes_priority_level(input->gang_global_priority_level); mes_add_queue_pkt.doorbell_offset = input->doorbell_offset; mes_add_queue_pkt.mqd_addr = input->mqd_addr; @@ -458,31 +475,6 @@ static int mes_v11_0_reset_queue_mmio(struct amdgpu_mes *mes, uint32_t queue_typ return r; } -static int mes_v11_0_reset_hw_queue(struct amdgpu_mes *mes, - struct mes_reset_queue_input *input) -{ - if (input->use_mmio) - return mes_v11_0_reset_queue_mmio(mes, input->queue_type, - input->me_id, input->pipe_id, - input->queue_id, input->vmid); - - union MESAPI__RESET mes_reset_queue_pkt; - - memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt)); - - mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER; - mes_reset_queue_pkt.header.opcode = MES_SCH_API_RESET; - mes_reset_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS; - - mes_reset_queue_pkt.doorbell_offset = input->doorbell_offset; - mes_reset_queue_pkt.gang_context_addr = input->gang_context_addr; - /*mes_reset_queue_pkt.reset_queue_only = 1;*/ - - return mes_v11_0_submit_pkt_and_poll_completion(mes, - &mes_reset_queue_pkt, sizeof(mes_reset_queue_pkt), - offsetof(union MESAPI__REMOVE_QUEUE, api_status)); -} - static int mes_v11_0_map_legacy_queue(struct amdgpu_mes *mes, struct mes_map_legacy_queue_input *input) { @@ -649,7 +641,7 @@ static int mes_v11_0_misc_op(struct amdgpu_mes *mes, break; case MES_MISC_OP_CHANGE_CONFIG: if ((mes->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) < 0x63) { - dev_err(mes->adev->dev, "MES FW versoin must be larger than 0x63 to support limit single process feature.\n"); + dev_err(mes->adev->dev, "MES FW version must be larger than 0x63 to support limit single process feature.\n"); return -EINVAL; } misc_pkt.opcode = MESAPI_MISC__CHANGE_CONFIG; @@ -694,7 +686,8 @@ static int mes_v11_0_set_hw_resources(struct amdgpu_mes *mes) mes->compute_hqd_mask[i]; for (i = 0; i < MAX_GFX_PIPES; i++) - mes_set_hw_res_pkt.gfx_hqd_mask[i] = mes->gfx_hqd_mask[i]; + mes_set_hw_res_pkt.gfx_hqd_mask[i] = + mes->gfx_hqd_mask[i]; for (i = 0; i < MAX_SDMA_PIPES; i++) mes_set_hw_res_pkt.sdma_hqd_mask[i] = mes->sdma_hqd_mask[i]; @@ -723,7 +716,7 @@ static int mes_v11_0_set_hw_resources(struct amdgpu_mes *mes) mes->event_log_gpu_addr; } - if (enforce_isolation) + if (adev->enforce_isolation[0] == AMDGPU_ENFORCE_ISOLATION_ENABLE) mes_set_hw_res_pkt.limit_single_process = 1; return mes_v11_0_submit_pkt_and_poll_completion(mes, @@ -753,8 +746,8 @@ static int mes_v11_0_set_hw_resources_1(struct amdgpu_mes *mes) offsetof(union MESAPI_SET_HW_RESOURCES_1, api_status)); } -static int mes_v11_0_reset_legacy_queue(struct amdgpu_mes *mes, - struct mes_reset_legacy_queue_input *input) +static int mes_v11_0_reset_hw_queue(struct amdgpu_mes *mes, + struct mes_reset_queue_input *input) { union MESAPI__RESET mes_reset_queue_pkt; @@ -772,7 +765,7 @@ static int mes_v11_0_reset_legacy_queue(struct amdgpu_mes *mes, mes_reset_queue_pkt.queue_type = convert_to_mes_queue_type(input->queue_type); - if (mes_reset_queue_pkt.queue_type == MES_QUEUE_TYPE_GFX) { + if (input->legacy_gfx) { mes_reset_queue_pkt.reset_legacy_gfx = 1; mes_reset_queue_pkt.pipe_id_lp = input->pipe_id; mes_reset_queue_pkt.queue_id_lp = input->queue_id; @@ -798,7 +791,6 @@ static const struct amdgpu_mes_funcs mes_v11_0_funcs = { .suspend_gang = mes_v11_0_suspend_gang, .resume_gang = mes_v11_0_resume_gang, .misc_op = mes_v11_0_misc_op, - .reset_legacy_queue = mes_v11_0_reset_legacy_queue, .reset_hw_queue = mes_v11_0_reset_hw_queue, }; @@ -1701,22 +1693,10 @@ static int mes_v11_0_early_init(struct amdgpu_ip_block *ip_block) return 0; } -static int mes_v11_0_late_init(struct amdgpu_ip_block *ip_block) -{ - struct amdgpu_device *adev = ip_block->adev; - - /* it's only intended for use in mes_self_test case, not for s0ix and reset */ - if (!amdgpu_in_reset(adev) && !adev->in_s0ix && !adev->in_suspend && - (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(11, 0, 3))) - amdgpu_mes_self_test(adev); - - return 0; -} - static const struct amd_ip_funcs mes_v11_0_ip_funcs = { .name = "mes_v11_0", .early_init = mes_v11_0_early_init, - .late_init = mes_v11_0_late_init, + .late_init = NULL, .sw_init = mes_v11_0_sw_init, .sw_fini = mes_v11_0_sw_fini, .hw_init = mes_v11_0_hw_init, diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index e6ab617b9a40..b4f17332d466 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -274,6 +274,23 @@ static int convert_to_mes_queue_type(int queue_type) return -1; } +static int convert_to_mes_priority_level(int priority_level) +{ + switch (priority_level) { + case AMDGPU_MES_PRIORITY_LEVEL_LOW: + return AMD_PRIORITY_LEVEL_LOW; + case AMDGPU_MES_PRIORITY_LEVEL_NORMAL: + default: + return AMD_PRIORITY_LEVEL_NORMAL; + case AMDGPU_MES_PRIORITY_LEVEL_MEDIUM: + return AMD_PRIORITY_LEVEL_MEDIUM; + case AMDGPU_MES_PRIORITY_LEVEL_HIGH: + return AMD_PRIORITY_LEVEL_HIGH; + case AMDGPU_MES_PRIORITY_LEVEL_REALTIME: + return AMD_PRIORITY_LEVEL_REALTIME; + } +} + static int mes_v12_0_add_hw_queue(struct amdgpu_mes *mes, struct mes_add_queue_input *input) { @@ -297,9 +314,9 @@ static int mes_v12_0_add_hw_queue(struct amdgpu_mes *mes, mes_add_queue_pkt.gang_quantum = input->gang_quantum; mes_add_queue_pkt.gang_context_addr = input->gang_context_addr; mes_add_queue_pkt.inprocess_gang_priority = - input->inprocess_gang_priority; + convert_to_mes_priority_level(input->inprocess_gang_priority); mes_add_queue_pkt.gang_global_priority_level = - input->gang_global_priority_level; + convert_to_mes_priority_level(input->gang_global_priority_level); mes_add_queue_pkt.doorbell_offset = input->doorbell_offset; mes_add_queue_pkt.mqd_addr = input->mqd_addr; @@ -477,32 +494,6 @@ static int mes_v12_0_reset_queue_mmio(struct amdgpu_mes *mes, uint32_t queue_typ return r; } -static int mes_v12_0_reset_hw_queue(struct amdgpu_mes *mes, - struct mes_reset_queue_input *input) -{ - union MESAPI__RESET mes_reset_queue_pkt; - int pipe; - - memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt)); - - mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER; - mes_reset_queue_pkt.header.opcode = MES_SCH_API_RESET; - mes_reset_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS; - - mes_reset_queue_pkt.doorbell_offset = input->doorbell_offset; - mes_reset_queue_pkt.gang_context_addr = input->gang_context_addr; - /*mes_reset_queue_pkt.reset_queue_only = 1;*/ - - if (mes->adev->enable_uni_mes) - pipe = AMDGPU_MES_KIQ_PIPE; - else - pipe = AMDGPU_MES_SCHED_PIPE; - - return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe, - &mes_reset_queue_pkt, sizeof(mes_reset_queue_pkt), - offsetof(union MESAPI__REMOVE_QUEUE, api_status)); -} - static int mes_v12_0_map_legacy_queue(struct amdgpu_mes *mes, struct mes_map_legacy_queue_input *input) { @@ -762,7 +753,7 @@ static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes, int pipe) pipe * (AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE); } - if (enforce_isolation) + if (adev->enforce_isolation[0] == AMDGPU_ENFORCE_ISOLATION_ENABLE) mes_set_hw_res_pkt.limit_single_process = 1; return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe, @@ -845,8 +836,8 @@ static void mes_v12_0_enable_unmapped_doorbell_handling( WREG32_SOC15(GC, 0, regCP_UNMAPPED_DOORBELL, data); } -static int mes_v12_0_reset_legacy_queue(struct amdgpu_mes *mes, - struct mes_reset_legacy_queue_input *input) +static int mes_v12_0_reset_hw_queue(struct amdgpu_mes *mes, + struct mes_reset_queue_input *input) { union MESAPI__RESET mes_reset_queue_pkt; int pipe; @@ -865,7 +856,7 @@ static int mes_v12_0_reset_legacy_queue(struct amdgpu_mes *mes, mes_reset_queue_pkt.queue_type = convert_to_mes_queue_type(input->queue_type); - if (mes_reset_queue_pkt.queue_type == MES_QUEUE_TYPE_GFX) { + if (input->legacy_gfx) { mes_reset_queue_pkt.reset_legacy_gfx = 1; mes_reset_queue_pkt.pipe_id_lp = input->pipe_id; mes_reset_queue_pkt.queue_id_lp = input->queue_id; @@ -878,7 +869,7 @@ static int mes_v12_0_reset_legacy_queue(struct amdgpu_mes *mes, mes_reset_queue_pkt.doorbell_offset = input->doorbell_offset; } - if (mes->adev->enable_uni_mes) + if (input->is_kq) pipe = AMDGPU_MES_KIQ_PIPE; else pipe = AMDGPU_MES_SCHED_PIPE; @@ -896,7 +887,6 @@ static const struct amdgpu_mes_funcs mes_v12_0_funcs = { .suspend_gang = mes_v12_0_suspend_gang, .resume_gang = mes_v12_0_resume_gang, .misc_op = mes_v12_0_misc_op, - .reset_legacy_queue = mes_v12_0_reset_legacy_queue, .reset_hw_queue = mes_v12_0_reset_hw_queue, }; @@ -1811,21 +1801,10 @@ static int mes_v12_0_early_init(struct amdgpu_ip_block *ip_block) return 0; } -static int mes_v12_0_late_init(struct amdgpu_ip_block *ip_block) -{ - struct amdgpu_device *adev = ip_block->adev; - - /* it's only intended for use in mes_self_test case, not for s0ix and reset */ - if (!amdgpu_in_reset(adev) && !adev->in_s0ix && !adev->in_suspend) - amdgpu_mes_self_test(adev); - - return 0; -} - static const struct amd_ip_funcs mes_v12_0_ip_funcs = { .name = "mes_v12_0", .early_init = mes_v12_0_early_init, - .late_init = mes_v12_0_late_init, + .late_init = NULL, .sw_init = mes_v12_0_sw_init, .sw_fini = mes_v12_0_sw_fini, .hw_init = mes_v12_0_hw_init, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c index 84cde1239ee4..76167fadb292 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c @@ -30,6 +30,7 @@ #include "soc15_common.h" #include "soc15.h" #include "amdgpu_ras.h" +#include "amdgpu_psp.h" #define regVM_L2_CNTL3_DEFAULT 0x80100007 #define regVM_L2_CNTL4_DEFAULT 0x000000c1 @@ -192,10 +193,8 @@ static void mmhub_v1_8_init_tlb_regs(struct amdgpu_device *adev) uint32_t tmp, inst_mask; int i; - /* Setup TLB control */ - inst_mask = adev->aid_mask; - for_each_inst(i, inst_mask) { - tmp = RREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL); + if (amdgpu_sriov_reg_indirect_l1_tlb_cntl(adev)) { + tmp = RREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1); @@ -209,7 +208,26 @@ static void mmhub_v1_8_init_tlb_regs(struct amdgpu_device *adev) MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); - WREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL, tmp); + psp_reg_program_no_ring(&adev->psp, tmp, PSP_REG_MMHUB_L1_TLB_CNTL); + } else { + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + tmp = RREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL); + + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, + 1); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + SYSTEM_ACCESS_MODE, 3); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, 1); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + MTYPE, MTYPE_UC);/* XXX for emulation. */ + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); + + WREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL, tmp); + } } } @@ -221,6 +239,9 @@ static void mmhub_v1_8_init_snoop_override_regs(struct amdgpu_device *adev) uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE - regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE; + if (amdgpu_sriov_vf(adev)) + return; + inst_mask = adev->aid_mask; for_each_inst(i, inst_mask) { for (j = 0; j < 5; j++) { /* DAGB instances */ @@ -454,6 +475,30 @@ static int mmhub_v1_8_gart_enable(struct amdgpu_device *adev) return 0; } +static void mmhub_v1_8_disable_l1_tlb(struct amdgpu_device *adev) +{ + u32 tmp; + u32 i, inst_mask; + + if (amdgpu_sriov_reg_indirect_l1_tlb_cntl(adev)) { + tmp = RREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, 0); + psp_reg_program_no_ring(&adev->psp, tmp, PSP_REG_MMHUB_L1_TLB_CNTL); + } else { + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + tmp = RREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, + 0); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, 0); + WREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL, tmp); + } + } +} + static void mmhub_v1_8_gart_disable(struct amdgpu_device *adev) { struct amdgpu_vmhub *hub; @@ -467,15 +512,6 @@ static void mmhub_v1_8_gart_disable(struct amdgpu_device *adev) for (i = 0; i < 16; i++) WREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT0_CNTL, i * hub->ctx_distance, 0); - - /* Setup TLB control */ - tmp = RREG32_SOC15(MMHUB, j, regMC_VM_MX_L1_TLB_CNTL); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, - 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - ENABLE_ADVANCED_DRIVER_MODEL, 0); - WREG32_SOC15(MMHUB, j, regMC_VM_MX_L1_TLB_CNTL, tmp); - if (!amdgpu_sriov_vf(adev)) { /* Setup L2 cache */ tmp = RREG32_SOC15(MMHUB, j, regVM_L2_CNTL); @@ -485,6 +521,8 @@ static void mmhub_v1_8_gart_disable(struct amdgpu_device *adev) WREG32_SOC15(MMHUB, j, regVM_L2_CNTL3, 0); } } + + mmhub_v1_8_disable_l1_tlb(adev); } /** diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v5_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v5_0.h new file mode 100644 index 000000000000..6f749814929f --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v5_0.h @@ -0,0 +1,144 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __MMSCH_V5_0_H__ +#define __MMSCH_V5_0_H__ + +#include "amdgpu_vcn.h" + +#define MMSCH_VERSION_MAJOR 5 +#define MMSCH_VERSION_MINOR 0 +#define MMSCH_VERSION (MMSCH_VERSION_MAJOR << 16 | MMSCH_VERSION_MINOR) + +#define RB_ENABLED (1 << 0) +#define RB4_ENABLED (1 << 1) + +#define MMSCH_VF_ENGINE_STATUS__PASS 0x1 + +#define MMSCH_VF_MAILBOX_RESP__OK 0x1 +#define MMSCH_VF_MAILBOX_RESP__INCOMPLETE 0x2 +#define MMSCH_VF_MAILBOX_RESP__FAILED 0x3 +#define MMSCH_VF_MAILBOX_RESP__FAILED_SMALL_CTX_SIZE 0x4 +#define MMSCH_VF_MAILBOX_RESP__UNKNOWN_CMD 0x5 + +enum mmsch_v5_0_command_type { + MMSCH_COMMAND__DIRECT_REG_WRITE = 0, + MMSCH_COMMAND__DIRECT_REG_POLLING = 2, + MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE = 3, + MMSCH_COMMAND__INDIRECT_REG_WRITE = 8, + MMSCH_COMMAND__END = 0xf +}; + +struct mmsch_v5_0_table_info { + uint32_t init_status; + uint32_t table_offset; + uint32_t table_size; +}; + +struct mmsch_v5_0_init_header { + uint32_t version; + uint32_t total_size; + struct mmsch_v5_0_table_info vcn0; + struct mmsch_v5_0_table_info mjpegdec0[5]; + struct mmsch_v5_0_table_info mjpegdec1[5]; +}; + +struct mmsch_v5_0_cmd_direct_reg_header { + uint32_t reg_offset : 28; + uint32_t command_type : 4; +}; + +struct mmsch_v5_0_cmd_indirect_reg_header { + uint32_t reg_offset : 20; + uint32_t reg_idx_space : 8; + uint32_t command_type : 4; +}; + +struct mmsch_v5_0_cmd_direct_write { + struct mmsch_v5_0_cmd_direct_reg_header cmd_header; + uint32_t reg_value; +}; + +struct mmsch_v5_0_cmd_direct_read_modify_write { + struct mmsch_v5_0_cmd_direct_reg_header cmd_header; + uint32_t write_data; + uint32_t mask_value; +}; + +struct mmsch_v5_0_cmd_direct_polling { + struct mmsch_v5_0_cmd_direct_reg_header cmd_header; + uint32_t mask_value; + uint32_t wait_value; +}; + +struct mmsch_v5_0_cmd_end { + struct mmsch_v5_0_cmd_direct_reg_header cmd_header; +}; + +struct mmsch_v5_0_cmd_indirect_write { + struct mmsch_v5_0_cmd_indirect_reg_header cmd_header; + uint32_t reg_value; +}; + +#define MMSCH_V5_0_INSERT_DIRECT_RD_MOD_WT(reg, mask, data) { \ + size = sizeof(struct mmsch_v5_0_cmd_direct_read_modify_write); \ + size_dw = size / 4; \ + direct_rd_mod_wt.cmd_header.reg_offset = reg; \ + direct_rd_mod_wt.mask_value = mask; \ + direct_rd_mod_wt.write_data = data; \ + memcpy((void *)table_loc, &direct_rd_mod_wt, size); \ + table_loc += size_dw; \ + table_size += size_dw; \ +} + +#define MMSCH_V5_0_INSERT_DIRECT_WT(reg, value) { \ + size = sizeof(struct mmsch_v5_0_cmd_direct_write); \ + size_dw = size / 4; \ + direct_wt.cmd_header.reg_offset = reg; \ + direct_wt.reg_value = value; \ + memcpy((void *)table_loc, &direct_wt, size); \ + table_loc += size_dw; \ + table_size += size_dw; \ +} + +#define MMSCH_V5_0_INSERT_DIRECT_POLL(reg, mask, wait) { \ + size = sizeof(struct mmsch_v5_0_cmd_direct_polling); \ + size_dw = size / 4; \ + direct_poll.cmd_header.reg_offset = reg; \ + direct_poll.mask_value = mask; \ + direct_poll.wait_value = wait; \ + memcpy((void *)table_loc, &direct_poll, size); \ + table_loc += size_dw; \ + table_size += size_dw; \ +} + +#define MMSCH_V5_0_INSERT_END() { \ + size = sizeof(struct mmsch_v5_0_cmd_end); \ + size_dw = size / 4; \ + memcpy((void *)table_loc, &end, size); \ + table_loc += size_dw; \ + table_size += size_dw; \ +} + +#endif + diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index f5411b798e11..48101a34e049 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -274,6 +274,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) { struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work); struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt); + struct amdgpu_reset_context reset_context = { 0 }; amdgpu_virt_fini_data_exchange(adev); @@ -281,8 +282,6 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) if (amdgpu_device_should_recover_gpu(adev) && (!amdgpu_device_has_job_running(adev) || adev->sdma_timeout == MAX_SCHEDULE_TIMEOUT)) { - struct amdgpu_reset_context reset_context; - memset(&reset_context, 0, sizeof(reset_context)); reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; @@ -293,6 +292,19 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) } } +static void xgpu_ai_mailbox_bad_pages_work(struct work_struct *work) +{ + struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, bad_pages_work); + struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt); + + if (down_read_trylock(&adev->reset_domain->sem)) { + amdgpu_virt_fini_data_exchange(adev); + amdgpu_virt_request_bad_pages(adev); + amdgpu_virt_init_data_exchange(adev); + up_read(&adev->reset_domain->sem); + } +} + static int xgpu_ai_set_mailbox_rcv_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, @@ -312,26 +324,42 @@ static int xgpu_ai_mailbox_rcv_irq(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { enum idh_event event = xgpu_ai_mailbox_peek_msg(adev); + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); switch (event) { - case IDH_FLR_NOTIFICATION: + case IDH_RAS_BAD_PAGES_NOTIFICATION: + xgpu_ai_mailbox_send_ack(adev); + if (amdgpu_sriov_runtime(adev)) + schedule_work(&adev->virt.bad_pages_work); + break; + case IDH_UNRECOV_ERR_NOTIFICATION: + xgpu_ai_mailbox_send_ack(adev); + ras->is_rma = true; + dev_err(adev->dev, "VF is in an unrecoverable state. Runtime Services are halted.\n"); if (amdgpu_sriov_runtime(adev)) WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, - &adev->virt.flr_work), - "Failed to queue work! at %s", - __func__); + &adev->virt.flr_work), + "Failed to queue work! at %s", + __func__); break; - case IDH_QUERY_ALIVE: - xgpu_ai_mailbox_send_ack(adev); - break; - /* READY_TO_ACCESS_GPU is fetched by kernel polling, IRQ can ignore - * it byfar since that polling thread will handle it, - * other msg like flr complete is not handled here. - */ - case IDH_CLR_MSG_BUF: - case IDH_FLR_NOTIFICATION_CMPL: - case IDH_READY_TO_ACCESS_GPU: - default: + case IDH_FLR_NOTIFICATION: + if (amdgpu_sriov_runtime(adev)) + WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, + &adev->virt.flr_work), + "Failed to queue work! at %s", + __func__); + break; + case IDH_QUERY_ALIVE: + xgpu_ai_mailbox_send_ack(adev); + break; + /* READY_TO_ACCESS_GPU is fetched by kernel polling, IRQ can ignore + * it byfar since that polling thread will handle it, + * other msg like flr complete is not handled here. + */ + case IDH_CLR_MSG_BUF: + case IDH_FLR_NOTIFICATION_CMPL: + case IDH_READY_TO_ACCESS_GPU: + default: break; } @@ -387,6 +415,7 @@ int xgpu_ai_mailbox_get_irq(struct amdgpu_device *adev) } INIT_WORK(&adev->virt.flr_work, xgpu_ai_mailbox_flr_work); + INIT_WORK(&adev->virt.bad_pages_work, xgpu_ai_mailbox_bad_pages_work); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h index ed57cbc150af..874b9f8f9804 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h @@ -40,6 +40,7 @@ enum idh_request { IDH_LOG_VF_ERROR = 200, IDH_READY_TO_RESET = 201, IDH_RAS_POISON = 202, + IDH_REQ_RAS_BAD_PAGES = 205, }; enum idh_event { @@ -54,6 +55,9 @@ enum idh_event { IDH_RAS_POISON_READY, IDH_PF_SOFT_FLR_NOTIFICATION, IDH_RAS_ERROR_DETECTED, + IDH_RAS_BAD_PAGES_READY = 15, + IDH_RAS_BAD_PAGES_NOTIFICATION = 16, + IDH_UNRECOV_ERR_NOTIFICATION = 17, IDH_TEXT_MESSAGE = 255, }; diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c index 5aadf24cb202..f6d8597452ed 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c @@ -67,6 +67,8 @@ static int xgpu_nv_mailbox_rcv_msg(struct amdgpu_device *adev, reg = RREG32_NO_KIQ(mmMAILBOX_MSGBUF_RCV_DW0); if (reg == IDH_FAIL) r = -EINVAL; + if (reg == IDH_UNRECOV_ERR_NOTIFICATION) + r = -ENODEV; else if (reg != event) return -ENOENT; @@ -103,6 +105,7 @@ static int xgpu_nv_poll_msg(struct amdgpu_device *adev, enum idh_event event) { int r; uint64_t timeout, now; + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); now = (uint64_t)ktime_to_ms(ktime_get()); timeout = now + NV_MAILBOX_POLL_MSG_TIMEDOUT; @@ -110,8 +113,16 @@ static int xgpu_nv_poll_msg(struct amdgpu_device *adev, enum idh_event event) do { r = xgpu_nv_mailbox_rcv_msg(adev, event); if (!r) { - dev_dbg(adev->dev, "rcv_msg 0x%x after %llu ms\n", event, NV_MAILBOX_POLL_MSG_TIMEDOUT - timeout + now); + dev_dbg(adev->dev, "rcv_msg 0x%x after %llu ms\n", + event, NV_MAILBOX_POLL_MSG_TIMEDOUT - timeout + now); return 0; + } else if (r == -ENODEV) { + if (!amdgpu_ras_is_rma(adev)) { + ras->is_rma = true; + dev_err(adev->dev, "VF is in an unrecoverable state. " + "Runtime Services are halted.\n"); + } + return r; } msleep(10); @@ -166,6 +177,10 @@ static int xgpu_nv_send_access_requests_with_param(struct amdgpu_device *adev, enum idh_event event = -1; send_request: + + if (amdgpu_ras_is_rma(adev)) + return -ENODEV; + xgpu_nv_mailbox_trans_msg(adev, req, data1, data2, data3); switch (req) { @@ -187,6 +202,9 @@ send_request: case IDH_REQ_RAS_CPER_DUMP: event = IDH_RAS_CPER_DUMP_READY; break; + case IDH_REQ_RAS_BAD_PAGES: + event = IDH_RAS_BAD_PAGES_READY; + break; default: break; } @@ -320,6 +338,7 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work) { struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work); struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt); + struct amdgpu_reset_context reset_context = { 0 }; amdgpu_virt_fini_data_exchange(adev); @@ -330,8 +349,6 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work) adev->gfx_timeout == MAX_SCHEDULE_TIMEOUT || adev->compute_timeout == MAX_SCHEDULE_TIMEOUT || adev->video_timeout == MAX_SCHEDULE_TIMEOUT)) { - struct amdgpu_reset_context reset_context; - memset(&reset_context, 0, sizeof(reset_context)); reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; @@ -342,6 +359,19 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work) } } +static void xgpu_nv_mailbox_bad_pages_work(struct work_struct *work) +{ + struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, bad_pages_work); + struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt); + + if (down_read_trylock(&adev->reset_domain->sem)) { + amdgpu_virt_fini_data_exchange(adev); + amdgpu_virt_request_bad_pages(adev); + amdgpu_virt_init_data_exchange(adev); + up_read(&adev->reset_domain->sem); + } +} + static int xgpu_nv_set_mailbox_rcv_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, @@ -364,8 +394,27 @@ static int xgpu_nv_mailbox_rcv_irq(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { enum idh_event event = xgpu_nv_mailbox_peek_msg(adev); + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); switch (event) { + case IDH_RAS_BAD_PAGES_NOTIFICATION: + xgpu_nv_mailbox_send_ack(adev); + if (amdgpu_sriov_runtime(adev)) + schedule_work(&adev->virt.bad_pages_work); + break; + case IDH_UNRECOV_ERR_NOTIFICATION: + xgpu_nv_mailbox_send_ack(adev); + if (!amdgpu_ras_is_rma(adev)) { + ras->is_rma = true; + dev_err(adev->dev, "VF is in an unrecoverable state. Runtime Services are halted.\n"); + } + + if (amdgpu_sriov_runtime(adev)) + WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, + &adev->virt.flr_work), + "Failed to queue work! at %s", + __func__); + break; case IDH_FLR_NOTIFICATION: if (amdgpu_sriov_runtime(adev)) WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, @@ -436,6 +485,7 @@ int xgpu_nv_mailbox_get_irq(struct amdgpu_device *adev) } INIT_WORK(&adev->virt.flr_work, xgpu_nv_mailbox_flr_work); + INIT_WORK(&adev->virt.bad_pages_work, xgpu_nv_mailbox_bad_pages_work); return 0; } @@ -480,6 +530,11 @@ static int xgpu_nv_req_ras_cper_dump(struct amdgpu_device *adev, u64 vf_rptr) adev, IDH_REQ_RAS_CPER_DUMP, vf_rptr_hi, vf_rptr_lo, 0); } +static int xgpu_nv_req_ras_bad_pages(struct amdgpu_device *adev) +{ + return xgpu_nv_send_access_requests(adev, IDH_REQ_RAS_BAD_PAGES); +} + const struct amdgpu_virt_ops xgpu_nv_virt_ops = { .req_full_gpu = xgpu_nv_request_full_gpu_access, .rel_full_gpu = xgpu_nv_release_full_gpu_access, @@ -492,4 +547,5 @@ const struct amdgpu_virt_ops xgpu_nv_virt_ops = { .rcvd_ras_intr = xgpu_nv_rcvd_ras_intr, .req_ras_err_count = xgpu_nv_req_ras_err_count, .req_ras_cper_dump = xgpu_nv_req_ras_cper_dump, + .req_bad_pages = xgpu_nv_req_ras_bad_pages, }; diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h index 72c9fceb9d79..5808689562cc 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h @@ -42,6 +42,7 @@ enum idh_request { IDH_RAS_POISON = 202, IDH_REQ_RAS_ERROR_COUNT = 203, IDH_REQ_RAS_CPER_DUMP = 204, + IDH_REQ_RAS_BAD_PAGES = 205, }; enum idh_event { @@ -58,6 +59,9 @@ enum idh_event { IDH_RAS_ERROR_DETECTED, IDH_RAS_ERROR_COUNT_READY = 11, IDH_RAS_CPER_DUMP_READY = 14, + IDH_RAS_BAD_PAGES_READY = 15, + IDH_RAS_BAD_PAGES_NOTIFICATION = 16, + IDH_UNRECOV_ERR_NOTIFICATION = 17, IDH_TEXT_MESSAGE = 255, }; diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c index f23cb79110d6..a376f072700d 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c @@ -177,8 +177,12 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do { u32 doorbell_range = 0, doorbell_ctrl = 0; u32 aid_id = instance; + u32 range_size; if (use_doorbell) { + range_size = (amdgpu_ip_version(adev, GC_HWIP, 0) == + IP_VERSION(9, 5, 0)) ? + 0xb : 0x9; doorbell_range = REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0, BIF_DOORBELL0_RANGE_OFFSET_ENTRY, @@ -186,7 +190,7 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do doorbell_range = REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0, BIF_DOORBELL0_RANGE_SIZE_ENTRY, - 0x9); + range_size); if (aid_id) doorbell_range = REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0, @@ -204,7 +208,7 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do S2A_DOORBELL_PORT1_RANGE_OFFSET, 0x4); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, - S2A_DOORBELL_PORT1_RANGE_SIZE, 0x9); + S2A_DOORBELL_PORT1_RANGE_SIZE, range_size); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x4); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c index fcd708eae75c..80153f837470 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c @@ -34,9 +34,6 @@ #include "sdma0/sdma0_4_0_offset.h" #include "nbio/nbio_7_4_offset.h" -#include "oss/osssys_4_0_offset.h" -#include "oss/osssys_4_0_sh_mask.h" - MODULE_FIRMWARE("amdgpu/renoir_asd.bin"); MODULE_FIRMWARE("amdgpu/renoir_ta.bin"); MODULE_FIRMWARE("amdgpu/green_sardine_asd.bin"); @@ -99,9 +96,6 @@ static int psp_v12_0_bootloader_load_sysdrv(struct psp_context *psp) WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35, psp_gfxdrv_command_reg); - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); - ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35), 0x80000000, 0x80000000, false); @@ -138,8 +132,6 @@ static int psp_v12_0_bootloader_load_sos(struct psp_context *psp) WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35, psp_gfxdrv_command_reg); - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_81), RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81), 0, true); @@ -147,37 +139,6 @@ static int psp_v12_0_bootloader_load_sos(struct psp_context *psp) return ret; } -static void psp_v12_0_reroute_ih(struct psp_context *psp) -{ - struct amdgpu_device *adev = psp->adev; - uint32_t tmp; - - /* Change IH ring for VMC */ - tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1244b); - tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, CLIENT_TYPE, 1); - tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1); - - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, 3); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, tmp); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_GBR_IH_SET); - - mdelay(20); - psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), - 0x80000000, 0x8000FFFF, false); - - /* Change IH ring for UMC */ - tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1216b); - tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1); - - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, 4); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, tmp); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_GBR_IH_SET); - - mdelay(20); - psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), - 0x80000000, 0x8000FFFF, false); -} - static int psp_v12_0_ring_create(struct psp_context *psp, enum psp_ring_type ring_type) { @@ -186,49 +147,23 @@ static int psp_v12_0_ring_create(struct psp_context *psp, struct psp_ring *ring = &psp->km_ring; struct amdgpu_device *adev = psp->adev; - psp_v12_0_reroute_ih(psp); - - if (amdgpu_sriov_vf(psp->adev)) { - /* Write low address of the ring to C2PMSG_102 */ - psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_102, psp_ring_reg); - /* Write high address of the ring to C2PMSG_103 */ - psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_103, psp_ring_reg); - - /* Write the ring initialization command to C2PMSG_101 */ - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_101, - GFX_CTRL_CMD_ID_INIT_GPCOM_RING); - - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); - - /* Wait for response flag (bit 31) in C2PMSG_101 */ - ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101), - 0x80000000, 0x8000FFFF, false); - - } else { - /* Write low address of the ring to C2PMSG_69 */ - psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, psp_ring_reg); - /* Write high address of the ring to C2PMSG_70 */ - psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, psp_ring_reg); - /* Write size of ring to C2PMSG_71 */ - psp_ring_reg = ring->ring_size; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_71, psp_ring_reg); - /* Write the ring initialization command to C2PMSG_64 */ - psp_ring_reg = ring_type; - psp_ring_reg = psp_ring_reg << 16; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, psp_ring_reg); - - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); - - /* Wait for response flag (bit 31) in C2PMSG_64 */ - ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), - 0x80000000, 0x8000FFFF, false); - } + /* Write low address of the ring to C2PMSG_69 */ + psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, psp_ring_reg); + /* Write high address of the ring to C2PMSG_70 */ + psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, psp_ring_reg); + /* Write size of ring to C2PMSG_71 */ + psp_ring_reg = ring->ring_size; + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_71, psp_ring_reg); + /* Write the ring initialization command to C2PMSG_64 */ + psp_ring_reg = ring_type; + psp_ring_reg = psp_ring_reg << 16; + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, psp_ring_reg); + + /* Wait for response flag (bit 31) in C2PMSG_64 */ + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), + 0x80000000, 0x8000FFFF, false); return ret; } @@ -247,9 +182,6 @@ static int psp_v12_0_ring_stop(struct psp_context *psp, WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_DESTROY_RINGS); - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); - /* Wait for response flag (bit 31) */ if (amdgpu_sriov_vf(adev)) ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101), diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c index afdf8ce3b4c5..df612fd9cc50 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c @@ -71,20 +71,13 @@ MODULE_FIRMWARE("amdgpu/psp_14_0_4_ta.bin"); /* Retry times for vmbx ready wait */ #define PSP_VMBX_POLLING_LIMIT 3000 -/* VBIOS gfl defines */ -#define MBOX_READY_MASK 0x80000000 -#define MBOX_STATUS_MASK 0x0000FFFF -#define MBOX_COMMAND_MASK 0x00FF0000 -#define MBOX_READY_FLAG 0x80000000 -#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_LO 0x2 -#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_HI 0x3 -#define C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE 0x4 - /* memory training timeout define */ #define MEM_TRAIN_SEND_MSG_TIMEOUT_US 3000000 #define regMP1_PUB_SCRATCH0 0x3b10090 +#define PSP13_BL_STATUS_SIZE 100 + static int psp_v13_0_init_microcode(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; @@ -151,6 +144,32 @@ static bool psp_v13_0_is_sos_alive(struct psp_context *psp) return sol_reg != 0x0; } +static void psp_v13_0_bootloader_print_status(struct psp_context *psp, + const char *msg) +{ + struct amdgpu_device *adev = psp->adev; + u32 bl_status_reg; + char bl_status_msg[PSP13_BL_STATUS_SIZE]; + int i, at; + + if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) || + amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) || + amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)) { + at = 0; + for_each_inst(i, adev->aid_mask) { + bl_status_reg = + (SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_92) + << 2) + + adev->asic_funcs->encode_ext_smn_addressing(i); + at += snprintf(bl_status_msg + at, + PSP13_BL_STATUS_SIZE - at, + " status(%02i): 0x%08x", i, + RREG32_PCIE_EXT(bl_status_reg)); + } + dev_info(adev->dev, "%s - %s", msg, bl_status_msg); + } +} + static int psp_v13_0_wait_for_vmbx_ready(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; @@ -196,6 +215,9 @@ static int psp_v13_0_wait_for_bootloader(struct psp_context *psp) if (ret == 0) return 0; + if (retry_loop && !(retry_loop % 10)) + psp_v13_0_bootloader_print_status( + psp, "Waiting for bootloader completion"); } return ret; @@ -710,7 +732,8 @@ static int psp_v13_0_exec_spi_cmd(struct psp_context *psp, int cmd) /* Ring the doorbell */ WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_73, 1); - if (cmd == C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE) + if (cmd == C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE || + cmd == C2PMSG_CMD_SPI_GET_FLASH_IMAGE) ret = psp_wait_for_spirom_update(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115), MBOX_READY_FLAG, MBOX_READY_MASK, PSP_SPIROM_UPDATE_TIMEOUT); else @@ -766,6 +789,37 @@ static int psp_v13_0_update_spirom(struct psp_context *psp, return 0; } +static int psp_v13_0_dump_spirom(struct psp_context *psp, + uint64_t fw_pri_mc_addr) +{ + struct amdgpu_device *adev = psp->adev; + int ret; + + /* Confirm PSP is ready to start */ + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115), + MBOX_READY_FLAG, MBOX_READY_MASK, false); + if (ret) { + dev_err(adev->dev, "PSP Not ready to start processing, ret = %d", ret); + return ret; + } + + WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_116, lower_32_bits(fw_pri_mc_addr)); + + ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_LO); + if (ret) + return ret; + + WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_116, upper_32_bits(fw_pri_mc_addr)); + + ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_HI); + if (ret) + return ret; + + ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_GET_FLASH_IMAGE); + + return ret; +} + static int psp_v13_0_vbflash_status(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; @@ -858,6 +912,25 @@ static bool psp_v13_0_is_reload_needed(struct psp_context *psp) return false; } +static int psp_v13_0_reg_program_no_ring(struct psp_context *psp, uint32_t val, + enum psp_reg_prog_id id) +{ + struct amdgpu_device *adev = psp->adev; + int ret = -EOPNOTSUPP; + + /* PSP will broadcast the value to all instances */ + if (amdgpu_sriov_vf(adev)) { + WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_101, GFX_CTRL_CMD_ID_GBR_IH_SET); + WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_102, id); + WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_103, val); + + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101), + 0x80000000, 0x80000000, false); + } + + return ret; +} + static const struct psp_funcs psp_v13_0_funcs = { .init_microcode = psp_v13_0_init_microcode, .wait_for_bootloader = psp_v13_0_wait_for_bootloader_steady_state, @@ -879,11 +952,13 @@ static const struct psp_funcs psp_v13_0_funcs = { .load_usbc_pd_fw = psp_v13_0_load_usbc_pd_fw, .read_usbc_pd_fw = psp_v13_0_read_usbc_pd_fw, .update_spirom = psp_v13_0_update_spirom, + .dump_spirom = psp_v13_0_dump_spirom, .vbflash_stat = psp_v13_0_vbflash_status, .fatal_error_recovery_quirk = psp_v13_0_fatal_error_recovery_quirk, .get_ras_capability = psp_v13_0_get_ras_capability, .is_aux_sos_load_required = psp_v13_0_is_aux_sos_load_required, .is_reload_needed = psp_v13_0_is_reload_needed, + .reg_program_no_ring = psp_v13_0_reg_program_no_ring, }; void psp_v13_0_set_psp_funcs(struct psp_context *psp) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c index 688a720bbbbd..9c169112a5e7 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c @@ -106,8 +106,9 @@ static void sdma_v4_4_2_set_buffer_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_set_vm_pte_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_set_ras_funcs(struct amdgpu_device *adev); -static void sdma_v4_4_2_set_engine_reset_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_update_reset_mask(struct amdgpu_device *adev); +static int sdma_v4_4_2_stop_queue(struct amdgpu_ring *ring); +static int sdma_v4_4_2_restore_queue(struct amdgpu_ring *ring); static u32 sdma_v4_4_2_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 offset) @@ -1333,6 +1334,11 @@ static bool sdma_v4_4_2_fw_support_paging_queue(struct amdgpu_device *adev) } } +static const struct amdgpu_sdma_funcs sdma_v4_4_2_sdma_funcs = { + .stop_kernel_queue = &sdma_v4_4_2_stop_queue, + .start_kernel_queue = &sdma_v4_4_2_restore_queue, +}; + static int sdma_v4_4_2_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; @@ -1351,8 +1357,6 @@ static int sdma_v4_4_2_early_init(struct amdgpu_ip_block *ip_block) sdma_v4_4_2_set_vm_pte_funcs(adev); sdma_v4_4_2_set_irq_funcs(adev); sdma_v4_4_2_set_ras_funcs(adev); - sdma_v4_4_2_set_engine_reset_funcs(adev); - return 0; } @@ -1447,6 +1451,7 @@ static int sdma_v4_4_2_sw_init(struct amdgpu_ip_block *ip_block) /* Initialize guilty flags for GFX and PAGE queues */ adev->sdma.instance[i].gfx_guilty = false; adev->sdma.instance[i].page_guilty = false; + adev->sdma.instance[i].funcs = &sdma_v4_4_2_sdma_funcs; ring = &adev->sdma.instance[i].ring; ring->ring_obj = NULL; @@ -1678,11 +1683,12 @@ static int sdma_v4_4_2_reset_queue(struct amdgpu_ring *ring, unsigned int vmid) return r; } -static int sdma_v4_4_2_stop_queue(struct amdgpu_device *adev, uint32_t instance_id) +static int sdma_v4_4_2_stop_queue(struct amdgpu_ring *ring) { + struct amdgpu_device *adev = ring->adev; + u32 instance_id = GET_INST(SDMA0, ring->me); u32 inst_mask; uint64_t rptr; - struct amdgpu_ring *ring = &adev->sdma.instance[instance_id].ring; if (amdgpu_sriov_vf(adev)) return -EINVAL; @@ -1715,11 +1721,11 @@ static int sdma_v4_4_2_stop_queue(struct amdgpu_device *adev, uint32_t instance_ return 0; } -static int sdma_v4_4_2_restore_queue(struct amdgpu_device *adev, uint32_t instance_id) +static int sdma_v4_4_2_restore_queue(struct amdgpu_ring *ring) { - int i; + struct amdgpu_device *adev = ring->adev; u32 inst_mask; - struct amdgpu_ring *ring = &adev->sdma.instance[instance_id].ring; + int i; inst_mask = 1 << ring->me; udelay(50); @@ -1739,16 +1745,6 @@ static int sdma_v4_4_2_restore_queue(struct amdgpu_device *adev, uint32_t instan return sdma_v4_4_2_inst_start(adev, inst_mask, true); } -static struct sdma_on_reset_funcs sdma_v4_4_2_engine_reset_funcs = { - .pre_reset = sdma_v4_4_2_stop_queue, - .post_reset = sdma_v4_4_2_restore_queue, -}; - -static void sdma_v4_4_2_set_engine_reset_funcs(struct amdgpu_device *adev) -{ - amdgpu_sdma_register_on_reset_callbacks(adev, &sdma_v4_4_2_engine_reset_funcs); -} - static int sdma_v4_4_2_set_trap_irq_state(struct amdgpu_device *adev, struct amdgpu_irq_src *source, unsigned type, @@ -2373,7 +2369,9 @@ static void sdma_v4_4_2_update_reset_mask(struct amdgpu_device *adev) adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; break; case IP_VERSION(9, 5, 0): - /*TODO: enable the queue reset flag until fw supported */ + if ((adev->gfx.mec_fw_version >= 0xf) && amdgpu_dpm_reset_sdma_is_supported(adev)) + adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; + break; default: break; } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c index 0dce59f4f6e2..9505ae96fbec 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c @@ -112,6 +112,8 @@ static void sdma_v5_0_set_ring_funcs(struct amdgpu_device *adev); static void sdma_v5_0_set_buffer_funcs(struct amdgpu_device *adev); static void sdma_v5_0_set_vm_pte_funcs(struct amdgpu_device *adev); static void sdma_v5_0_set_irq_funcs(struct amdgpu_device *adev); +static int sdma_v5_0_stop_queue(struct amdgpu_ring *ring); +static int sdma_v5_0_restore_queue(struct amdgpu_ring *ring); static const struct soc15_reg_golden golden_settings_sdma_5[] = { SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_CHICKEN_BITS, 0xffbf1f0f, 0x03ab0107), @@ -369,67 +371,36 @@ static uint64_t sdma_v5_0_ring_get_wptr(struct amdgpu_ring *ring) static void sdma_v5_0_ring_set_wptr(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - uint32_t *wptr_saved; - uint32_t *is_queue_unmap; - uint64_t aggregated_db_index; - uint32_t mqd_size = adev->mqds[AMDGPU_HW_IP_DMA].mqd_size; DRM_DEBUG("Setting write pointer\n"); - if (ring->is_mes_queue) { - wptr_saved = (uint32_t *)(ring->mqd_ptr + mqd_size); - is_queue_unmap = (uint32_t *)(ring->mqd_ptr + mqd_size + - sizeof(uint32_t)); - aggregated_db_index = - amdgpu_mes_get_aggregated_doorbell_index(adev, - AMDGPU_MES_PRIORITY_LEVEL_NORMAL); - + if (ring->use_doorbell) { + DRM_DEBUG("Using doorbell -- " + "wptr_offs == 0x%08x " + "lower_32_bits(ring->wptr) << 2 == 0x%08x " + "upper_32_bits(ring->wptr) << 2 == 0x%08x\n", + ring->wptr_offs, + lower_32_bits(ring->wptr << 2), + upper_32_bits(ring->wptr << 2)); + /* XXX check if swapping is necessary on BE */ atomic64_set((atomic64_t *)ring->wptr_cpu_addr, ring->wptr << 2); - *wptr_saved = ring->wptr << 2; - if (*is_queue_unmap) { - WDOORBELL64(aggregated_db_index, ring->wptr << 2); - DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", - ring->doorbell_index, ring->wptr << 2); - WDOORBELL64(ring->doorbell_index, ring->wptr << 2); - } else { - DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", - ring->doorbell_index, ring->wptr << 2); - WDOORBELL64(ring->doorbell_index, ring->wptr << 2); - - if (*is_queue_unmap) - WDOORBELL64(aggregated_db_index, - ring->wptr << 2); - } + DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", + ring->doorbell_index, ring->wptr << 2); + WDOORBELL64(ring->doorbell_index, ring->wptr << 2); } else { - if (ring->use_doorbell) { - DRM_DEBUG("Using doorbell -- " - "wptr_offs == 0x%08x " - "lower_32_bits(ring->wptr) << 2 == 0x%08x " - "upper_32_bits(ring->wptr) << 2 == 0x%08x\n", - ring->wptr_offs, - lower_32_bits(ring->wptr << 2), - upper_32_bits(ring->wptr << 2)); - /* XXX check if swapping is necessary on BE */ - atomic64_set((atomic64_t *)ring->wptr_cpu_addr, - ring->wptr << 2); - DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", - ring->doorbell_index, ring->wptr << 2); - WDOORBELL64(ring->doorbell_index, ring->wptr << 2); - } else { - DRM_DEBUG("Not using doorbell -- " - "mmSDMA%i_GFX_RB_WPTR == 0x%08x " - "mmSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n", - ring->me, - lower_32_bits(ring->wptr << 2), - ring->me, - upper_32_bits(ring->wptr << 2)); - WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, - ring->me, mmSDMA0_GFX_RB_WPTR), - lower_32_bits(ring->wptr << 2)); - WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, - ring->me, mmSDMA0_GFX_RB_WPTR_HI), - upper_32_bits(ring->wptr << 2)); - } + DRM_DEBUG("Not using doorbell -- " + "mmSDMA%i_GFX_RB_WPTR == 0x%08x " + "mmSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n", + ring->me, + lower_32_bits(ring->wptr << 2), + ring->me, + upper_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, + ring->me, mmSDMA0_GFX_RB_WPTR), + lower_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, + ring->me, mmSDMA0_GFX_RB_WPTR_HI), + upper_32_bits(ring->wptr << 2)); } } @@ -575,11 +546,9 @@ static void sdma_v5_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se } if (flags & AMDGPU_FENCE_FLAG_INT) { - uint32_t ctx = ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0; /* generate an interrupt */ amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_TRAP)); - amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(ctx)); + amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0)); } } @@ -588,15 +557,15 @@ static void sdma_v5_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se * sdma_v5_0_gfx_stop - stop the gfx async dma engines * * @adev: amdgpu_device pointer - * + * @inst_mask: mask of dma engine instances to be disabled * Stop the gfx async dma ring buffers (NAVI10). */ -static void sdma_v5_0_gfx_stop(struct amdgpu_device *adev) +static void sdma_v5_0_gfx_stop(struct amdgpu_device *adev, uint32_t inst_mask) { u32 rb_cntl, ib_cntl; int i; - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0); WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); @@ -688,9 +657,11 @@ static void sdma_v5_0_enable(struct amdgpu_device *adev, bool enable) { u32 f32_cntl; int i; + uint32_t inst_mask; + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); if (!enable) { - sdma_v5_0_gfx_stop(adev); + sdma_v5_0_gfx_stop(adev, 1 << inst_mask); sdma_v5_0_rlc_stop(adev); } @@ -1046,33 +1017,22 @@ static int sdma_v5_0_ring_test_ring(struct amdgpu_ring *ring) int r; u32 tmp; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); - return r; - } - - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + return r; } + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_ring_alloc(ring, 20); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1085,10 +1045,7 @@ static int sdma_v5_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) break; if (amdgpu_emu_mode == 1) @@ -1100,8 +1057,7 @@ static int sdma_v5_0_ring_test_ring(struct amdgpu_ring *ring) if (i >= adev->usec_timeout) r = -ETIMEDOUT; - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1124,38 +1080,24 @@ static int sdma_v5_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) long r; u32 tmp = 0; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = (void *)amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); + return r; + } - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ib_get(adev, NULL, 256, - AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); - goto err0; - } + r = amdgpu_ib_get(adev, NULL, 256, + AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); + goto err0; } ib.ptr[0] = SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) | @@ -1183,10 +1125,7 @@ static int sdma_v5_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto err1; } - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) r = 0; @@ -1197,8 +1136,7 @@ err1: amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err0: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1389,6 +1327,36 @@ static void sdma_v5_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring, amdgpu_ring_emit_reg_wait(ring, reg1, mask, mask); } +static int sdma_v5_0_soft_reset_engine(struct amdgpu_device *adev, u32 instance_id) +{ + u32 grbm_soft_reset; + u32 tmp; + + grbm_soft_reset = REG_SET_FIELD(0, + GRBM_SOFT_RESET, SOFT_RESET_SDMA0, + 1); + grbm_soft_reset <<= instance_id; + + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + tmp |= grbm_soft_reset; + DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp); + WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + + udelay(50); + + tmp &= ~grbm_soft_reset; + WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + return 0; +} + +static const struct amdgpu_sdma_funcs sdma_v5_0_sdma_funcs = { + .stop_kernel_queue = &sdma_v5_0_stop_queue, + .start_kernel_queue = &sdma_v5_0_restore_queue, + .soft_reset_kernel_queue = &sdma_v5_0_soft_reset_engine, +}; + static int sdma_v5_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; @@ -1431,6 +1399,7 @@ static int sdma_v5_0_sw_init(struct amdgpu_ip_block *ip_block) return r; for (i = 0; i < adev->sdma.num_instances; i++) { + adev->sdma.instance[i].funcs = &sdma_v5_0_sdma_funcs; ring = &adev->sdma.instance[i].ring; ring->ring_obj = NULL; ring->use_doorbell = true; @@ -1572,32 +1541,25 @@ static int sdma_v5_0_soft_reset(struct amdgpu_ip_block *ip_block) static int sdma_v5_0_reset_queue(struct amdgpu_ring *ring, unsigned int vmid) { struct amdgpu_device *adev = ring->adev; - int i, j, r; - u32 rb_cntl, ib_cntl, f32_cntl, freeze, cntl, preempt, soft_reset, stat1_reg; + u32 inst_id = ring->me; - if (amdgpu_sriov_vf(adev)) - return -EINVAL; + return amdgpu_sdma_reset_engine(adev, inst_id); +} - for (i = 0; i < adev->sdma.num_instances; i++) { - if (ring == &adev->sdma.instance[i].ring) - break; - } +static int sdma_v5_0_stop_queue(struct amdgpu_ring *ring) +{ + u32 f32_cntl, freeze, cntl, stat1_reg; + struct amdgpu_device *adev = ring->adev; + int i, j, r = 0; - if (i == adev->sdma.num_instances) { - DRM_ERROR("sdma instance not found\n"); + if (amdgpu_sriov_vf(adev)) return -EINVAL; - } + i = ring->me; amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* stop queue */ - ib_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL)); - ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 0); - WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl); - - rb_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0); - WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); + sdma_v5_0_gfx_stop(adev, 1 << i); /* engine stop SDMA1_F32_CNTL.HALT to 1 and SDMAx_FREEZE freeze bit to 1 */ freeze = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_FREEZE)); @@ -1628,30 +1590,25 @@ static int sdma_v5_0_reset_queue(struct amdgpu_ring *ring, unsigned int vmid) cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL)); cntl = REG_SET_FIELD(cntl, SDMA0_CNTL, UTC_L1_ENABLE, 0); WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), cntl); +err0: + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + return r; +} - /* soft reset SDMA_GFX_PREEMPT.IB_PREEMPT = 0 mmGRBM_SOFT_RESET.SOFT_RESET_SDMA0/1 = 1 */ - preempt = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_PREEMPT)); - preempt = REG_SET_FIELD(preempt, SDMA0_GFX_PREEMPT, IB_PREEMPT, 0); - WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_PREEMPT), preempt); - - soft_reset = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); - soft_reset |= 1 << GRBM_SOFT_RESET__SOFT_RESET_SDMA0__SHIFT << i; - - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, soft_reset); - - udelay(50); - - soft_reset &= ~(1 << GRBM_SOFT_RESET__SOFT_RESET_SDMA0__SHIFT << i); - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, soft_reset); +static int sdma_v5_0_restore_queue(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + u32 inst_id = ring->me; + u32 freeze; + int r; + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* unfreeze*/ - freeze = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_FREEZE)); + freeze = RREG32(sdma_v5_0_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE)); freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 0); - WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_FREEZE), freeze); + WREG32(sdma_v5_0_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE), freeze); - r = sdma_v5_0_gfx_resume_instance(adev, i, true); - -err0: + r = sdma_v5_0_gfx_resume_instance(adev, inst_id, true); amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c index 2b39a03ff0c1..a6e612b4a892 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c @@ -113,6 +113,8 @@ static void sdma_v5_2_set_ring_funcs(struct amdgpu_device *adev); static void sdma_v5_2_set_buffer_funcs(struct amdgpu_device *adev); static void sdma_v5_2_set_vm_pte_funcs(struct amdgpu_device *adev); static void sdma_v5_2_set_irq_funcs(struct amdgpu_device *adev); +static int sdma_v5_2_stop_queue(struct amdgpu_ring *ring); +static int sdma_v5_2_restore_queue(struct amdgpu_ring *ring); static u32 sdma_v5_2_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 internal_offset) { @@ -394,11 +396,9 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se } if ((flags & AMDGPU_FENCE_FLAG_INT)) { - uint32_t ctx = ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0; /* generate an interrupt */ amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_TRAP)); - amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(ctx)); + amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0)); } } @@ -407,15 +407,15 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se * sdma_v5_2_gfx_stop - stop the gfx async dma engines * * @adev: amdgpu_device pointer - * + * @inst_mask: mask of dma engine instances to be disabled * Stop the gfx async dma ring buffers. */ -static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev) +static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev, uint32_t inst_mask) { u32 rb_cntl, ib_cntl; int i; - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0); WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); @@ -506,9 +506,11 @@ static void sdma_v5_2_enable(struct amdgpu_device *adev, bool enable) { u32 f32_cntl; int i; + uint32_t inst_mask; + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); if (!enable) { - sdma_v5_2_gfx_stop(adev); + sdma_v5_2_gfx_stop(adev, inst_mask); sdma_v5_2_rlc_stop(adev); } @@ -761,37 +763,49 @@ static int sdma_v5_2_load_microcode(struct amdgpu_device *adev) return 0; } -static int sdma_v5_2_soft_reset(struct amdgpu_ip_block *ip_block) +static int sdma_v5_2_soft_reset_engine(struct amdgpu_device *adev, u32 instance_id) { - struct amdgpu_device *adev = ip_block->adev; u32 grbm_soft_reset; u32 tmp; - int i; - for (i = 0; i < adev->sdma.num_instances; i++) { - grbm_soft_reset = REG_SET_FIELD(0, - GRBM_SOFT_RESET, SOFT_RESET_SDMA0, - 1); - grbm_soft_reset <<= i; + grbm_soft_reset = REG_SET_FIELD(0, + GRBM_SOFT_RESET, SOFT_RESET_SDMA0, + 1); + grbm_soft_reset <<= instance_id; - tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); - tmp |= grbm_soft_reset; - DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); - tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + tmp |= grbm_soft_reset; + DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp); + WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); - udelay(50); + udelay(50); + + tmp &= ~grbm_soft_reset; + WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + return 0; +} - tmp &= ~grbm_soft_reset; - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); - tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); +static int sdma_v5_2_soft_reset(struct amdgpu_ip_block *ip_block) +{ + struct amdgpu_device *adev = ip_block->adev; + int i; + for (i = 0; i < adev->sdma.num_instances; i++) { + sdma_v5_2_soft_reset_engine(adev, i); udelay(50); } return 0; } +static const struct amdgpu_sdma_funcs sdma_v5_2_sdma_funcs = { + .stop_kernel_queue = &sdma_v5_2_stop_queue, + .start_kernel_queue = &sdma_v5_2_restore_queue, + .soft_reset_kernel_queue = &sdma_v5_2_soft_reset_engine, +}; + /** * sdma_v5_2_start - setup and start the async dma engines * @@ -903,33 +917,22 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring) int r; u32 tmp; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); - return r; - } - - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + return r; } + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_ring_alloc(ring, 20); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -942,10 +945,7 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) break; if (amdgpu_emu_mode == 1) @@ -957,8 +957,7 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring) if (i >= adev->usec_timeout) r = -ETIMEDOUT; - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -981,37 +980,23 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout) long r; u32 tmp = 0; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = (void *)amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); + return r; + } - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); - goto err0; - } + r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); + goto err0; } ib.ptr[0] = SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) | @@ -1039,10 +1024,7 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto err1; } - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) r = 0; @@ -1053,8 +1035,7 @@ err1: amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err0: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1337,6 +1318,7 @@ static int sdma_v5_2_sw_init(struct amdgpu_ip_block *ip_block) } for (i = 0; i < adev->sdma.num_instances; i++) { + adev->sdma.instance[i].funcs = &sdma_v5_2_sdma_funcs; ring = &adev->sdma.instance[i].ring; ring->ring_obj = NULL; ring->use_doorbell = true; @@ -1472,32 +1454,25 @@ static int sdma_v5_2_wait_for_idle(struct amdgpu_ip_block *ip_block) static int sdma_v5_2_reset_queue(struct amdgpu_ring *ring, unsigned int vmid) { struct amdgpu_device *adev = ring->adev; - int i, j, r; - u32 rb_cntl, ib_cntl, f32_cntl, freeze, cntl, preempt, soft_reset, stat1_reg; + u32 inst_id = ring->me; - if (amdgpu_sriov_vf(adev)) - return -EINVAL; + return amdgpu_sdma_reset_engine(adev, inst_id); +} - for (i = 0; i < adev->sdma.num_instances; i++) { - if (ring == &adev->sdma.instance[i].ring) - break; - } +static int sdma_v5_2_stop_queue(struct amdgpu_ring *ring) +{ + u32 f32_cntl, freeze, cntl, stat1_reg; + struct amdgpu_device *adev = ring->adev; + int i, j, r = 0; - if (i == adev->sdma.num_instances) { - DRM_ERROR("sdma instance not found\n"); + if (amdgpu_sriov_vf(adev)) return -EINVAL; - } + i = ring->me; amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* stop queue */ - ib_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL)); - ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 0); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl); - - rb_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); + sdma_v5_2_gfx_stop(adev, 1 << i); /*engine stop SDMA1_F32_CNTL.HALT to 1 and SDMAx_FREEZE freeze bit to 1 */ freeze = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE)); @@ -1530,31 +1505,26 @@ static int sdma_v5_2_reset_queue(struct amdgpu_ring *ring, unsigned int vmid) cntl = REG_SET_FIELD(cntl, SDMA0_CNTL, UTC_L1_ENABLE, 0); WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), cntl); - /* soft reset SDMA_GFX_PREEMPT.IB_PREEMPT = 0 mmGRBM_SOFT_RESET.SOFT_RESET_SDMA0/1 = 1 */ - preempt = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_PREEMPT)); - preempt = REG_SET_FIELD(preempt, SDMA0_GFX_PREEMPT, IB_PREEMPT, 0); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_PREEMPT), preempt); - - soft_reset = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); - soft_reset |= 1 << GRBM_SOFT_RESET__SOFT_RESET_SDMA0__SHIFT << i; - - - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, soft_reset); - - udelay(50); - - soft_reset &= ~(1 << GRBM_SOFT_RESET__SOFT_RESET_SDMA0__SHIFT << i); +err0: + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + return r; +} - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, soft_reset); +static int sdma_v5_2_restore_queue(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + u32 inst_id = ring->me; + u32 freeze; + int r; + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* unfreeze and unhalt */ - freeze = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE)); + freeze = RREG32(sdma_v5_2_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE)); freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 0); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE), freeze); + WREG32(sdma_v5_2_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE), freeze); - r = sdma_v5_2_gfx_resume_instance(adev, i, true); + r = sdma_v5_2_gfx_resume_instance(adev, inst_id, true); -err0: amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c index c214c3d2149b..da5b5d64f137 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c @@ -43,6 +43,7 @@ #include "sdma_common.h" #include "sdma_v6_0.h" #include "v11_structs.h" +#include "mes_userqueue.h" MODULE_FIRMWARE("amdgpu/sdma_6_0_0.bin"); MODULE_FIRMWARE("amdgpu/sdma_6_0_1.bin"); @@ -376,11 +377,9 @@ static void sdma_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se } if (flags & AMDGPU_FENCE_FLAG_INT) { - uint32_t ctx = ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0; /* generate an interrupt */ amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_TRAP)); - amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(ctx)); + amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0)); } } @@ -891,6 +890,9 @@ static int sdma_v6_0_mqd_init(struct amdgpu_device *adev, void *mqd, m->sdmax_rlcx_rb_aql_cntl = regSDMA0_QUEUE0_RB_AQL_CNTL_DEFAULT; m->sdmax_rlcx_dummy_reg = regSDMA0_QUEUE0_DUMMY_REG_DEFAULT; + m->sdmax_rlcx_csa_addr_lo = lower_32_bits(prop->csa_addr); + m->sdmax_rlcx_csa_addr_hi = upper_32_bits(prop->csa_addr); + return 0; } @@ -917,33 +919,22 @@ static int sdma_v6_0_ring_test_ring(struct amdgpu_ring *ring) int r; u32 tmp; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); - return r; - } - - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + return r; } + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_ring_alloc(ring, 5); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -956,10 +947,7 @@ static int sdma_v6_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) break; if (amdgpu_emu_mode == 1) @@ -971,8 +959,7 @@ static int sdma_v6_0_ring_test_ring(struct amdgpu_ring *ring) if (i >= adev->usec_timeout) r = -ETIMEDOUT; - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -995,37 +982,23 @@ static int sdma_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) long r; u32 tmp = 0; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = (void *)amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); + return r; + } - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); - goto err0; - } + r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); + goto err0; } ib.ptr[0] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) | @@ -1053,10 +1026,7 @@ static int sdma_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto err1; } - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) r = 0; @@ -1067,8 +1037,7 @@ err1: amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err0: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1300,6 +1269,23 @@ static int sdma_v6_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int r; + switch (amdgpu_user_queue) { + case -1: + case 0: + default: + adev->sdma.no_user_submission = false; + adev->sdma.disable_uq = true; + break; + case 1: + adev->sdma.no_user_submission = false; + adev->sdma.disable_uq = false; + break; + case 2: + adev->sdma.no_user_submission = true; + adev->sdma.disable_uq = false; + break; + } + r = amdgpu_sdma_init_microcode(adev, 0, true); if (r) return r; @@ -1334,6 +1320,7 @@ static int sdma_v6_0_sw_init(struct amdgpu_ip_block *ip_block) ring->ring_obj = NULL; ring->use_doorbell = true; ring->me = i; + ring->no_user_submission = adev->sdma.no_user_submission; DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i, ring->use_doorbell?"true":"false"); @@ -1376,6 +1363,10 @@ static int sdma_v6_0_sw_init(struct amdgpu_ip_block *ip_block) else DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n"); + /* add firmware version checks here */ + if (0 && !adev->sdma.disable_uq) + adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs; + r = amdgpu_sdma_sysfs_reset_mask_init(adev); if (r) return r; @@ -1399,11 +1390,39 @@ static int sdma_v6_0_sw_fini(struct amdgpu_ip_block *ip_block) return 0; } +static int sdma_v6_0_set_userq_trap_interrupts(struct amdgpu_device *adev, + bool enable) +{ + unsigned int irq_type; + int i, r; + + if (adev->userq_funcs[AMDGPU_HW_IP_DMA]) { + for (i = 0; i < adev->sdma.num_instances; i++) { + irq_type = AMDGPU_SDMA_IRQ_INSTANCE0 + i; + if (enable) + r = amdgpu_irq_get(adev, &adev->sdma.trap_irq, + irq_type); + else + r = amdgpu_irq_put(adev, &adev->sdma.trap_irq, + irq_type); + if (r) + return r; + } + } + + return 0; +} + static int sdma_v6_0_hw_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int r; - return sdma_v6_0_start(adev); + r = sdma_v6_0_start(adev); + if (r) + return r; + + return sdma_v6_0_set_userq_trap_interrupts(adev, true); } static int sdma_v6_0_hw_fini(struct amdgpu_ip_block *ip_block) @@ -1415,6 +1434,7 @@ static int sdma_v6_0_hw_fini(struct amdgpu_ip_block *ip_block) sdma_v6_0_ctxempty_int_enable(adev, false); sdma_v6_0_enable(adev, false); + sdma_v6_0_set_userq_trap_interrupts(adev, false); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c index b2706221df99..befe013b11a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c @@ -42,6 +42,7 @@ #include "sdma_common.h" #include "sdma_v7_0.h" #include "v12_structs.h" +#include "mes_userqueue.h" MODULE_FIRMWARE("amdgpu/sdma_7_0_0.bin"); MODULE_FIRMWARE("amdgpu/sdma_7_0_1.bin"); @@ -204,66 +205,39 @@ static uint64_t sdma_v7_0_ring_get_wptr(struct amdgpu_ring *ring) static void sdma_v7_0_ring_set_wptr(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - uint32_t *wptr_saved; - uint32_t *is_queue_unmap; - uint64_t aggregated_db_index; - uint32_t mqd_size = adev->mqds[AMDGPU_HW_IP_DMA].mqd_size; DRM_DEBUG("Setting write pointer\n"); - if (ring->is_mes_queue) { - wptr_saved = (uint32_t *)(ring->mqd_ptr + mqd_size); - is_queue_unmap = (uint32_t *)(ring->mqd_ptr + mqd_size + - sizeof(uint32_t)); - aggregated_db_index = - amdgpu_mes_get_aggregated_doorbell_index(adev, - ring->hw_prio); - + if (ring->use_doorbell) { + DRM_DEBUG("Using doorbell -- " + "wptr_offs == 0x%08x " + "lower_32_bits(ring->wptr) << 2 == 0x%08x " + "upper_32_bits(ring->wptr) << 2 == 0x%08x\n", + ring->wptr_offs, + lower_32_bits(ring->wptr << 2), + upper_32_bits(ring->wptr << 2)); + /* XXX check if swapping is necessary on BE */ atomic64_set((atomic64_t *)ring->wptr_cpu_addr, ring->wptr << 2); - *wptr_saved = ring->wptr << 2; - if (*is_queue_unmap) { - WDOORBELL64(aggregated_db_index, ring->wptr << 2); - DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", - ring->doorbell_index, ring->wptr << 2); - WDOORBELL64(ring->doorbell_index, ring->wptr << 2); - } else { - DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", - ring->doorbell_index, ring->wptr << 2); - WDOORBELL64(ring->doorbell_index, ring->wptr << 2); - } + DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", + ring->doorbell_index, ring->wptr << 2); + WDOORBELL64(ring->doorbell_index, ring->wptr << 2); } else { - if (ring->use_doorbell) { - DRM_DEBUG("Using doorbell -- " - "wptr_offs == 0x%08x " - "lower_32_bits(ring->wptr) << 2 == 0x%08x " - "upper_32_bits(ring->wptr) << 2 == 0x%08x\n", - ring->wptr_offs, - lower_32_bits(ring->wptr << 2), - upper_32_bits(ring->wptr << 2)); - /* XXX check if swapping is necessary on BE */ - atomic64_set((atomic64_t *)ring->wptr_cpu_addr, - ring->wptr << 2); - DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", - ring->doorbell_index, ring->wptr << 2); - WDOORBELL64(ring->doorbell_index, ring->wptr << 2); - } else { - DRM_DEBUG("Not using doorbell -- " - "regSDMA%i_GFX_RB_WPTR == 0x%08x " - "regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n", - ring->me, - lower_32_bits(ring->wptr << 2), - ring->me, - upper_32_bits(ring->wptr << 2)); - WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, - ring->me, - regSDMA0_QUEUE0_RB_WPTR), - lower_32_bits(ring->wptr << 2)); - WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, - ring->me, - regSDMA0_QUEUE0_RB_WPTR_HI), - upper_32_bits(ring->wptr << 2)); - } + DRM_DEBUG("Not using doorbell -- " + "regSDMA%i_GFX_RB_WPTR == 0x%08x " + "regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n", + ring->me, + lower_32_bits(ring->wptr << 2), + ring->me, + upper_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, + ring->me, + regSDMA0_QUEUE0_RB_WPTR), + lower_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, + ring->me, + regSDMA0_QUEUE0_RB_WPTR_HI), + upper_32_bits(ring->wptr << 2)); } } @@ -407,11 +381,9 @@ static void sdma_v7_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se } if (flags & AMDGPU_FENCE_FLAG_INT) { - uint32_t ctx = ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0; /* generate an interrupt */ amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_TRAP)); - amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(ctx)); + amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0)); } } @@ -935,6 +907,9 @@ static int sdma_v7_0_mqd_init(struct amdgpu_device *adev, void *mqd, m->sdmax_rlcx_rb_aql_cntl = 0x4000; //regSDMA0_QUEUE0_RB_AQL_CNTL_DEFAULT; m->sdmax_rlcx_dummy_reg = 0xf; //regSDMA0_QUEUE0_DUMMY_REG_DEFAULT; + m->sdmax_rlcx_csa_addr_lo = lower_32_bits(prop->csa_addr); + m->sdmax_rlcx_csa_addr_hi = upper_32_bits(prop->csa_addr); + return 0; } @@ -961,33 +936,22 @@ static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring) int r; u32 tmp; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); - return r; - } - - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + return r; } + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_ring_alloc(ring, 5); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1000,10 +964,7 @@ static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) break; if (amdgpu_emu_mode == 1) @@ -1015,8 +976,7 @@ static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring) if (i >= adev->usec_timeout) r = -ETIMEDOUT; - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1039,37 +999,23 @@ static int sdma_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) long r; u32 tmp = 0; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = (void *)amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); + return r; + } - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); - goto err0; - } + r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); + goto err0; } ib.ptr[0] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) | @@ -1097,10 +1043,7 @@ static int sdma_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto err1; } - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) r = 0; @@ -1111,8 +1054,7 @@ err1: amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err0: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1312,6 +1254,23 @@ static int sdma_v7_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int r; + switch (amdgpu_user_queue) { + case -1: + case 0: + default: + adev->sdma.no_user_submission = false; + adev->sdma.disable_uq = true; + break; + case 1: + adev->sdma.no_user_submission = false; + adev->sdma.disable_uq = false; + break; + case 2: + adev->sdma.no_user_submission = true; + adev->sdma.disable_uq = false; + break; + } + r = amdgpu_sdma_init_microcode(adev, 0, true); if (r) { DRM_ERROR("Failed to init sdma firmware!\n"); @@ -1347,6 +1306,7 @@ static int sdma_v7_0_sw_init(struct amdgpu_ip_block *ip_block) ring->ring_obj = NULL; ring->use_doorbell = true; ring->me = i; + ring->no_user_submission = adev->sdma.no_user_submission; DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i, ring->use_doorbell?"true":"false"); @@ -1378,6 +1338,10 @@ static int sdma_v7_0_sw_init(struct amdgpu_ip_block *ip_block) else DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n"); + /* add firmware version checks here */ + if (0 && !adev->sdma.disable_uq) + adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs; + return r; } @@ -1400,11 +1364,39 @@ static int sdma_v7_0_sw_fini(struct amdgpu_ip_block *ip_block) return 0; } +static int sdma_v7_0_set_userq_trap_interrupts(struct amdgpu_device *adev, + bool enable) +{ + unsigned int irq_type; + int i, r; + + if (adev->userq_funcs[AMDGPU_HW_IP_DMA]) { + for (i = 0; i < adev->sdma.num_instances; i++) { + irq_type = AMDGPU_SDMA_IRQ_INSTANCE0 + i; + if (enable) + r = amdgpu_irq_get(adev, &adev->sdma.trap_irq, + irq_type); + else + r = amdgpu_irq_put(adev, &adev->sdma.trap_irq, + irq_type); + if (r) + return r; + } + } + + return 0; +} + static int sdma_v7_0_hw_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int r; - return sdma_v7_0_start(adev); + r = sdma_v7_0_start(adev); + if (r) + return r; + + return sdma_v7_0_set_userq_trap_interrupts(adev, true); } static int sdma_v7_0_hw_fini(struct amdgpu_ip_block *ip_block) @@ -1416,6 +1408,7 @@ static int sdma_v7_0_hw_fini(struct amdgpu_ip_block *ip_block) sdma_v7_0_ctx_switch_enable(adev, false); sdma_v7_0_enable(adev, false); + sdma_v7_0_set_userq_trap_interrupts(adev, false); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c index 2247f6a94858..e0f139de7991 100644 --- a/drivers/gpu/drm/amd/amdgpu/si.c +++ b/drivers/gpu/drm/amd/amdgpu/si.c @@ -35,6 +35,7 @@ #include "amdgpu_vce.h" #include "atom.h" #include "amd_pcie.h" + #include "si_dpm.h" #include "sid.h" #include "si_ih.h" @@ -44,17 +45,30 @@ #include "dce_v6_0.h" #include "si.h" #include "uvd_v3_1.h" -#include "amdgpu_vkms.h" + +#include "uvd/uvd_4_0_d.h" + +#include "smu/smu_6_0_d.h" +#include "smu/smu_6_0_sh_mask.h" + #include "gca/gfx_6_0_d.h" +#include "gca/gfx_6_0_sh_mask.h" + #include "oss/oss_1_0_d.h" #include "oss/oss_1_0_sh_mask.h" + #include "gmc/gmc_6_0_d.h" +#include"gmc/gmc_6_0_sh_mask.h" + #include "dce/dce_6_0_d.h" -#include "uvd/uvd_4_0_d.h" +#include "dce/dce_6_0_sh_mask.h" + #include "bif/bif_3_0_d.h" #include "bif/bif_3_0_sh_mask.h" +#include "si_enums.h" #include "amdgpu_dm.h" +#include "amdgpu_vkms.h" static const u32 tahiti_golden_registers[] = { @@ -1071,8 +1085,8 @@ static u32 si_smc_rreg(struct amdgpu_device *adev, u32 reg) u32 r; spin_lock_irqsave(&adev->smc_idx_lock, flags); - WREG32(SMC_IND_INDEX_0, (reg)); - r = RREG32(SMC_IND_DATA_0); + WREG32(mmSMC_IND_INDEX_0, (reg)); + r = RREG32(mmSMC_IND_DATA_0); spin_unlock_irqrestore(&adev->smc_idx_lock, flags); return r; } @@ -1082,8 +1096,8 @@ static void si_smc_wreg(struct amdgpu_device *adev, u32 reg, u32 v) unsigned long flags; spin_lock_irqsave(&adev->smc_idx_lock, flags); - WREG32(SMC_IND_INDEX_0, (reg)); - WREG32(SMC_IND_DATA_0, (v)); + WREG32(mmSMC_IND_INDEX_0, (reg)); + WREG32(mmSMC_IND_DATA_0, (v)); spin_unlock_irqrestore(&adev->smc_idx_lock, flags); } @@ -1110,20 +1124,20 @@ static void si_uvd_ctx_wreg(struct amdgpu_device *adev, u32 reg, u32 v) } static struct amdgpu_allowed_register_entry si_allowed_read_registers[] = { - {GRBM_STATUS}, + {mmGRBM_STATUS}, {mmGRBM_STATUS2}, {mmGRBM_STATUS_SE0}, {mmGRBM_STATUS_SE1}, {mmSRBM_STATUS}, {mmSRBM_STATUS2}, - {DMA_STATUS_REG + DMA0_REGISTER_OFFSET}, - {DMA_STATUS_REG + DMA1_REGISTER_OFFSET}, + {mmDMA_STATUS_REG + DMA0_REGISTER_OFFSET}, + {mmDMA_STATUS_REG + DMA1_REGISTER_OFFSET}, {mmCP_STAT}, {mmCP_STALLED_STAT1}, {mmCP_STALLED_STAT2}, {mmCP_STALLED_STAT3}, - {GB_ADDR_CONFIG}, - {MC_ARB_RAMCFG}, + {mmGB_ADDR_CONFIG}, + {mmMC_ARB_RAMCFG}, {mmGB_TILE_MODE0}, {mmGB_TILE_MODE1}, {mmGB_TILE_MODE2}, @@ -1156,7 +1170,7 @@ static struct amdgpu_allowed_register_entry si_allowed_read_registers[] = { {mmGB_TILE_MODE29}, {mmGB_TILE_MODE30}, {mmGB_TILE_MODE31}, - {CC_RB_BACKEND_DISABLE, true}, + {mmCC_RB_BACKEND_DISABLE, true}, {mmGC_USER_RB_BACKEND_DISABLE, true}, {mmPA_SC_RASTER_CONFIG, true}, }; @@ -1264,37 +1278,37 @@ static bool si_read_disabled_bios(struct amdgpu_device *adev) u32 rom_cntl; bool r; - bus_cntl = RREG32(R600_BUS_CNTL); + bus_cntl = RREG32(mmBUS_CNTL); if (adev->mode_info.num_crtc) { - d1vga_control = RREG32(AVIVO_D1VGA_CONTROL); - d2vga_control = RREG32(AVIVO_D2VGA_CONTROL); - vga_render_control = RREG32(VGA_RENDER_CONTROL); + d1vga_control = RREG32(mmD1VGA_CONTROL); + d2vga_control = RREG32(mmD2VGA_CONTROL); + vga_render_control = RREG32(mmVGA_RENDER_CONTROL); } rom_cntl = RREG32(R600_ROM_CNTL); /* enable the rom */ - WREG32(R600_BUS_CNTL, (bus_cntl & ~R600_BIOS_ROM_DIS)); + WREG32(mmBUS_CNTL, (bus_cntl & ~BUS_CNTL__BIOS_ROM_DIS_MASK)); if (adev->mode_info.num_crtc) { /* Disable VGA mode */ - WREG32(AVIVO_D1VGA_CONTROL, - (d1vga_control & ~(AVIVO_DVGA_CONTROL_MODE_ENABLE | - AVIVO_DVGA_CONTROL_TIMING_SELECT))); - WREG32(AVIVO_D2VGA_CONTROL, - (d2vga_control & ~(AVIVO_DVGA_CONTROL_MODE_ENABLE | - AVIVO_DVGA_CONTROL_TIMING_SELECT))); - WREG32(VGA_RENDER_CONTROL, - (vga_render_control & C_000300_VGA_VSTATUS_CNTL)); + WREG32(mmD1VGA_CONTROL, + (d1vga_control & ~(D1VGA_CONTROL__D1VGA_MODE_ENABLE_MASK | + D1VGA_CONTROL__D1VGA_TIMING_SELECT_MASK))); + WREG32(mmD2VGA_CONTROL, + (d2vga_control & ~(D1VGA_CONTROL__D1VGA_MODE_ENABLE_MASK | + D1VGA_CONTROL__D1VGA_TIMING_SELECT_MASK))); + WREG32(mmVGA_RENDER_CONTROL, + (vga_render_control & ~VGA_RENDER_CONTROL__VGA_VSTATUS_CNTL_MASK)); } WREG32(R600_ROM_CNTL, rom_cntl | R600_SCK_OVERWRITE); r = amdgpu_read_bios(adev); /* restore regs */ - WREG32(R600_BUS_CNTL, bus_cntl); + WREG32(mmBUS_CNTL, bus_cntl); if (adev->mode_info.num_crtc) { - WREG32(AVIVO_D1VGA_CONTROL, d1vga_control); - WREG32(AVIVO_D2VGA_CONTROL, d2vga_control); - WREG32(VGA_RENDER_CONTROL, vga_render_control); + WREG32(mmD1VGA_CONTROL, d1vga_control); + WREG32(mmD2VGA_CONTROL, d2vga_control); + WREG32(mmVGA_RENDER_CONTROL, vga_render_control); } WREG32(R600_ROM_CNTL, rom_cntl); return r; @@ -1331,23 +1345,24 @@ static void si_set_clk_bypass_mode(struct amdgpu_device *adev) { u32 tmp, i; - tmp = RREG32(CG_SPLL_FUNC_CNTL); - tmp |= SPLL_BYPASS_EN; - WREG32(CG_SPLL_FUNC_CNTL, tmp); + tmp = RREG32(mmCG_SPLL_FUNC_CNTL); + tmp |= CG_SPLL_FUNC_CNTL__SPLL_BYPASS_EN_MASK; + WREG32(mmCG_SPLL_FUNC_CNTL, tmp); - tmp = RREG32(CG_SPLL_FUNC_CNTL_2); - tmp |= SPLL_CTLREQ_CHG; - WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + tmp = RREG32(mmCG_SPLL_FUNC_CNTL_2); + tmp |= CG_SPLL_FUNC_CNTL_2__SPLL_CTLREQ_CHG_MASK; + WREG32(mmCG_SPLL_FUNC_CNTL_2, tmp); for (i = 0; i < adev->usec_timeout; i++) { - if (RREG32(SPLL_STATUS) & SPLL_CHG_STATUS) + if (RREG32(mmCG_SPLL_STATUS) & CG_SPLL_STATUS__SPLL_CHG_STATUS_MASK) break; udelay(1); } - tmp = RREG32(CG_SPLL_FUNC_CNTL_2); - tmp &= ~(SPLL_CTLREQ_CHG | SCLK_MUX_UPDATE); - WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + tmp = RREG32(mmCG_SPLL_FUNC_CNTL_2); + tmp &= ~(CG_SPLL_FUNC_CNTL_2__SPLL_CTLREQ_CHG_MASK | + CG_SPLL_FUNC_CNTL_2__SCLK_MUX_UPDATE_MASK); + WREG32(mmCG_SPLL_FUNC_CNTL_2, tmp); tmp = RREG32(MPLL_CNTL_MODE); tmp &= ~MPLL_MCLK_SEL; @@ -1358,21 +1373,21 @@ static void si_spll_powerdown(struct amdgpu_device *adev) { u32 tmp; - tmp = RREG32(SPLL_CNTL_MODE); - tmp |= SPLL_SW_DIR_CONTROL; - WREG32(SPLL_CNTL_MODE, tmp); + tmp = RREG32(mmSPLL_CNTL_MODE); + tmp |= SPLL_CNTL_MODE__SPLL_SW_DIR_CONTROL_MASK; + WREG32(mmSPLL_CNTL_MODE, tmp); - tmp = RREG32(CG_SPLL_FUNC_CNTL); - tmp |= SPLL_RESET; - WREG32(CG_SPLL_FUNC_CNTL, tmp); + tmp = RREG32(mmCG_SPLL_FUNC_CNTL); + tmp |= CG_SPLL_FUNC_CNTL__SPLL_RESET_MASK; + WREG32(mmCG_SPLL_FUNC_CNTL, tmp); - tmp = RREG32(CG_SPLL_FUNC_CNTL); - tmp |= SPLL_SLEEP; - WREG32(CG_SPLL_FUNC_CNTL, tmp); + tmp = RREG32(mmCG_SPLL_FUNC_CNTL); + tmp |= CG_SPLL_FUNC_CNTL__SPLL_SLEEP_MASK; + WREG32(mmCG_SPLL_FUNC_CNTL, tmp); - tmp = RREG32(SPLL_CNTL_MODE); - tmp &= ~SPLL_SW_DIR_CONTROL; - WREG32(SPLL_CNTL_MODE, tmp); + tmp = RREG32(mmSPLL_CNTL_MODE); + tmp &= ~SPLL_CNTL_MODE__SPLL_SW_DIR_CONTROL_MASK; + WREG32(mmSPLL_CNTL_MODE, tmp); } static int si_gpu_pci_config_reset(struct amdgpu_device *adev) @@ -1454,14 +1469,14 @@ static void si_vga_set_state(struct amdgpu_device *adev, bool state) { uint32_t temp; - temp = RREG32(CONFIG_CNTL); + temp = RREG32(mmCONFIG_CNTL); if (!state) { temp &= ~(1<<0); temp |= (1<<1); } else { temp &= ~(1<<1); } - WREG32(CONFIG_CNTL, temp); + WREG32(mmCONFIG_CNTL, temp); } static u32 si_get_xclk(struct amdgpu_device *adev) @@ -1469,12 +1484,12 @@ static u32 si_get_xclk(struct amdgpu_device *adev) u32 reference_clock = adev->clock.spll.reference_freq; u32 tmp; - tmp = RREG32(CG_CLKPIN_CNTL_2); - if (tmp & MUX_TCLK_TO_XCLK) + tmp = RREG32(mmCG_CLKPIN_CNTL_2); + if (tmp & CG_CLKPIN_CNTL_2__MUX_TCLK_TO_XCLK_MASK) return TCLK; - tmp = RREG32(CG_CLKPIN_CNTL); - if (tmp & XTALIN_DIVIDE) + tmp = RREG32(mmCG_CLKPIN_CNTL); + if (tmp & CG_CLKPIN_CNTL__XTALIN_DIVIDE_MASK) return reference_clock / 4; return reference_clock; @@ -1519,9 +1534,9 @@ static int si_get_pcie_lanes(struct amdgpu_device *adev) if (adev->flags & AMD_IS_APU) return 0; - link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); + link_width_cntl = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL); - switch ((link_width_cntl & LC_LINK_WIDTH_RD_MASK) >> LC_LINK_WIDTH_RD_SHIFT) { + switch ((link_width_cntl & PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD_MASK) >> PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD__SHIFT) { case LC_LINK_WIDTH_X1: return 1; case LC_LINK_WIDTH_X2: @@ -1568,13 +1583,13 @@ static void si_set_pcie_lanes(struct amdgpu_device *adev, int lanes) return; } - link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); - link_width_cntl &= ~LC_LINK_WIDTH_MASK; - link_width_cntl |= mask << LC_LINK_WIDTH_SHIFT; - link_width_cntl |= (LC_RECONFIG_NOW | - LC_RECONFIG_ARC_MISSING_ESCAPE); + link_width_cntl = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL); + link_width_cntl &= ~PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_MASK; + link_width_cntl |= mask << PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH__SHIFT; + link_width_cntl |= (PCIE_LC_LINK_WIDTH_CNTL__LC_RECONFIG_NOW_MASK | + PCIE_LC_LINK_WIDTH_CNTL__LC_RECONFIG_ARC_MISSING_ESCAPE_MASK); - WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); + WREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); } static void si_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0, @@ -2018,7 +2033,7 @@ static const struct amdgpu_asic_funcs si_asic_funcs = static uint32_t si_get_rev_id(struct amdgpu_device *adev) { - return (RREG32(CC_DRM_ID_STRAPS) & CC_DRM_ID_STRAPS__ATI_REV_ID_MASK) + return (RREG32(mmCC_DRM_ID_STRAPS) & CC_DRM_ID_STRAPS__ATI_REV_ID_MASK) >> CC_DRM_ID_STRAPS__ATI_REV_ID__SHIFT; } @@ -2239,9 +2254,9 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev) CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3))) return; - speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); - current_data_rate = (speed_cntl & LC_CURRENT_DATA_RATE_MASK) >> - LC_CURRENT_DATA_RATE_SHIFT; + speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL); + current_data_rate = (speed_cntl & PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK) >> + PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE__SHIFT; if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) { if (current_data_rate == 2) { DRM_INFO("PCIE gen 3 link speeds already enabled\n"); @@ -2268,17 +2283,17 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev) pcie_capability_set_word(root, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD); pcie_capability_set_word(adev->pdev, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD); - tmp = RREG32_PCIE(PCIE_LC_STATUS1); - max_lw = (tmp & LC_DETECTED_LINK_WIDTH_MASK) >> LC_DETECTED_LINK_WIDTH_SHIFT; - current_lw = (tmp & LC_OPERATING_LINK_WIDTH_MASK) >> LC_OPERATING_LINK_WIDTH_SHIFT; + tmp = RREG32_PCIE(ixPCIE_LC_STATUS1); + max_lw = (tmp & PCIE_LC_STATUS1__LC_DETECTED_LINK_WIDTH_MASK) >> PCIE_LC_STATUS1__LC_DETECTED_LINK_WIDTH__SHIFT; + current_lw = (tmp & PCIE_LC_STATUS1__LC_OPERATING_LINK_WIDTH_MASK) >> PCIE_LC_STATUS1__LC_OPERATING_LINK_WIDTH__SHIFT; if (current_lw < max_lw) { - tmp = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); - if (tmp & LC_RENEGOTIATION_SUPPORT) { - tmp &= ~(LC_LINK_WIDTH_MASK | LC_UPCONFIGURE_DIS); - tmp |= (max_lw << LC_LINK_WIDTH_SHIFT); - tmp |= LC_UPCONFIGURE_SUPPORT | LC_RENEGOTIATE_EN | LC_RECONFIG_NOW; - WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, tmp); + tmp = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL); + if (tmp & PCIE_LC_LINK_WIDTH_CNTL__LC_RENEGOTIATION_SUPPORT_MASK) { + tmp &= ~(PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_MASK | PCIE_LC_LINK_WIDTH_CNTL__LC_UPCONFIGURE_DIS_MASK); + tmp |= (max_lw << PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH__SHIFT); + tmp |= PCIE_LC_LINK_WIDTH_CNTL__LC_UPCONFIGURE_SUPPORT_MASK | PCIE_LC_LINK_WIDTH_CNTL__LC_RENEGOTIATE_EN_MASK | PCIE_LC_LINK_WIDTH_CNTL__LC_RECONFIG_NOW_MASK; + WREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL, tmp); } } @@ -2301,13 +2316,13 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev) PCI_EXP_LNKCTL2, &gpu_cfg2); - tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); - tmp |= LC_SET_QUIESCE; - WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); + tmp = RREG32_PCIE_PORT(ixPCIE_LC_CNTL4); + tmp |= PCIE_LC_CNTL4__LC_SET_QUIESCE_MASK; + WREG32_PCIE_PORT(ixPCIE_LC_CNTL4, tmp); - tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); - tmp |= LC_REDO_EQ; - WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); + tmp = RREG32_PCIE_PORT(ixPCIE_LC_CNTL4); + tmp |= PCIE_LC_CNTL4__LC_REDO_EQ_MASK; + WREG32_PCIE_PORT(ixPCIE_LC_CNTL4, tmp); mdelay(100); @@ -2333,16 +2348,16 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev) (PCI_EXP_LNKCTL2_ENTER_COMP | PCI_EXP_LNKCTL2_TX_MARGIN)); - tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); - tmp &= ~LC_SET_QUIESCE; - WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); + tmp = RREG32_PCIE_PORT(ixPCIE_LC_CNTL4); + tmp &= ~PCIE_LC_CNTL4__LC_SET_QUIESCE_MASK; + WREG32_PCIE_PORT(ixPCIE_LC_CNTL4, tmp); } } } - speed_cntl |= LC_FORCE_EN_SW_SPEED_CHANGE | LC_FORCE_DIS_HW_SPEED_CHANGE; - speed_cntl &= ~LC_FORCE_DIS_SW_SPEED_CHANGE; - WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); + speed_cntl |= PCIE_LC_SPEED_CNTL__LC_FORCE_EN_SW_SPEED_CHANGE_MASK | PCIE_LC_SPEED_CNTL__LC_FORCE_DIS_HW_SPEED_CHANGE_MASK; + speed_cntl &= ~PCIE_LC_SPEED_CNTL__LC_FORCE_DIS_SW_SPEED_CHANGE_MASK; + WREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL, speed_cntl); tmp16 = 0; if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) @@ -2354,13 +2369,13 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev) pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL2, PCI_EXP_LNKCTL2_TLS, tmp16); - speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); - speed_cntl |= LC_INITIATE_LINK_SPEED_CHANGE; - WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); + speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL); + speed_cntl |= PCIE_LC_SPEED_CNTL__LC_INITIATE_LINK_SPEED_CHANGE_MASK; + WREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL, speed_cntl); for (i = 0; i < adev->usec_timeout; i++) { - speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); - if ((speed_cntl & LC_INITIATE_LINK_SPEED_CHANGE) == 0) + speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL); + if ((speed_cntl & PCIE_LC_SPEED_CNTL__LC_INITIATE_LINK_SPEED_CHANGE_MASK) == 0) break; udelay(1); } @@ -2418,121 +2433,121 @@ static void si_program_aspm(struct amdgpu_device *adev) if (!amdgpu_device_should_use_aspm(adev)) return; - orig = data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL); - data &= ~LC_XMIT_N_FTS_MASK; - data |= LC_XMIT_N_FTS(0x24) | LC_XMIT_N_FTS_OVERRIDE_EN; + orig = data = RREG32_PCIE_PORT(ixPCIE_LC_N_FTS_CNTL); + data &= ~PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS_MASK; + data |= (0x24 << PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS__SHIFT) | PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS_OVERRIDE_EN_MASK; if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL, data); + WREG32_PCIE_PORT(ixPCIE_LC_N_FTS_CNTL, data); - orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL3); - data |= LC_GO_TO_RECOVERY; + orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL3); + data |= PCIE_LC_CNTL3__LC_GO_TO_RECOVERY_MASK; if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_CNTL3, data); + WREG32_PCIE_PORT(ixPCIE_LC_CNTL3, data); - orig = data = RREG32_PCIE(PCIE_P_CNTL); - data |= P_IGNORE_EDB_ERR; + orig = data = RREG32_PCIE(ixPCIE_P_CNTL); + data |= PCIE_P_CNTL__P_IGNORE_EDB_ERR_MASK; if (orig != data) - WREG32_PCIE(PCIE_P_CNTL, data); + WREG32_PCIE(ixPCIE_P_CNTL, data); - orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL); - data &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK); - data |= LC_PMI_TO_L1_DIS; + orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL); + data &= ~(PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK | PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK); + data |= PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK; if (!disable_l0s) - data |= LC_L0S_INACTIVITY(7); + data |= (7 << PCIE_LC_CNTL__LC_L0S_INACTIVITY__SHIFT); if (!disable_l1) { - data |= LC_L1_INACTIVITY(7); - data &= ~LC_PMI_TO_L1_DIS; + data |= (7 << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT); + data &= ~PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK; if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_CNTL, data); + WREG32_PCIE_PORT(ixPCIE_LC_CNTL, data); if (!disable_plloff_in_l1) { bool clk_req_support; - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_0); - data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK); - data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7); + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_0); + data &= ~(PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0_MASK | PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0_MASK); + data |= (7 << PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0__SHIFT) | (7 << PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0__SHIFT); if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_0, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_0, data); - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_1); - data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK); - data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7); + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_1); + data &= ~(PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1_MASK | PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1_MASK); + data |= (7 << PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1__SHIFT) | (7 << PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1__SHIFT); if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_1, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_1, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_0); - data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK); - data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7); + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_0); + data &= ~(PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0_MASK | PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0_MASK); + data |= (7 << PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0__SHIFT) | (7 << PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0__SHIFT); if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_0, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_0, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_1); - data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK); - data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7); + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_1); + data &= ~(PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1_MASK | PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1_MASK); + data |= (7 << PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1__SHIFT) | (7 << PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1__SHIFT); if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_1, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_1, data); if ((adev->asic_type != CHIP_OLAND) && (adev->asic_type != CHIP_HAINAN)) { - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_0); - data &= ~PLL_RAMP_UP_TIME_0_MASK; + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_0); + data &= ~PB0_PIF_PWRDOWN_0__PLL_RAMP_UP_TIME_0_MASK; if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_0, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_0, data); - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_1); - data &= ~PLL_RAMP_UP_TIME_1_MASK; + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_1); + data &= ~PB0_PIF_PWRDOWN_1__PLL_RAMP_UP_TIME_1_MASK; if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_1, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_1, data); - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_2); - data &= ~PLL_RAMP_UP_TIME_2_MASK; + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_2); + data &= ~PB0_PIF_PWRDOWN_2__PLL_RAMP_UP_TIME_2_MASK; if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_2, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_2, data); - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_3); - data &= ~PLL_RAMP_UP_TIME_3_MASK; + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_3); + data &= ~PB0_PIF_PWRDOWN_3__PLL_RAMP_UP_TIME_3_MASK; if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_3, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_3, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_0); - data &= ~PLL_RAMP_UP_TIME_0_MASK; + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_0); + data &= ~PB1_PIF_PWRDOWN_0__PLL_RAMP_UP_TIME_0_MASK; if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_0, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_0, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_1); - data &= ~PLL_RAMP_UP_TIME_1_MASK; + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_1); + data &= ~PB1_PIF_PWRDOWN_1__PLL_RAMP_UP_TIME_1_MASK; if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_1, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_1, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_2); - data &= ~PLL_RAMP_UP_TIME_2_MASK; + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_2); + data &= ~PB1_PIF_PWRDOWN_2__PLL_RAMP_UP_TIME_2_MASK; if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_2, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_2, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_3); - data &= ~PLL_RAMP_UP_TIME_3_MASK; + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_3); + data &= ~PB1_PIF_PWRDOWN_3__PLL_RAMP_UP_TIME_3_MASK; if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_3, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_3, data); } - orig = data = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); - data &= ~LC_DYN_LANES_PWR_STATE_MASK; - data |= LC_DYN_LANES_PWR_STATE(3); + orig = data = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL); + data &= ~PCIE_LC_LINK_WIDTH_CNTL__LC_DYN_LANES_PWR_STATE_MASK; + data |= (3 << PCIE_LC_LINK_WIDTH_CNTL__LC_DYN_LANES_PWR_STATE__SHIFT); if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data); + WREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL, data); - orig = data = si_pif_phy0_rreg(adev,PB0_PIF_CNTL); - data &= ~LS2_EXIT_TIME_MASK; + orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_CNTL); + data &= ~PB0_PIF_CNTL__LS2_EXIT_TIME_MASK; if ((adev->asic_type == CHIP_OLAND) || (adev->asic_type == CHIP_HAINAN)) - data |= LS2_EXIT_TIME(5); + data |= (5 << PB0_PIF_CNTL__LS2_EXIT_TIME__SHIFT); if (orig != data) - si_pif_phy0_wreg(adev,PB0_PIF_CNTL, data); + si_pif_phy0_wreg(adev,ixPB0_PIF_CNTL, data); - orig = data = si_pif_phy1_rreg(adev,PB1_PIF_CNTL); - data &= ~LS2_EXIT_TIME_MASK; + orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_CNTL); + data &= ~PB1_PIF_CNTL__LS2_EXIT_TIME_MASK; if ((adev->asic_type == CHIP_OLAND) || (adev->asic_type == CHIP_HAINAN)) - data |= LS2_EXIT_TIME(5); + data |= (5 << PB1_PIF_CNTL__LS2_EXIT_TIME__SHIFT); if (orig != data) - si_pif_phy1_wreg(adev,PB1_PIF_CNTL, data); + si_pif_phy1_wreg(adev,ixPB1_PIF_CNTL, data); if (!disable_clkreq && !pci_is_root_bus(adev->pdev->bus)) { @@ -2548,64 +2563,64 @@ static void si_program_aspm(struct amdgpu_device *adev) } if (clk_req_support) { - orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL2); - data |= LC_ALLOW_PDWN_IN_L1 | LC_ALLOW_PDWN_IN_L23; + orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL2); + data |= PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L1_MASK | PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L23_MASK; if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_CNTL2, data); + WREG32_PCIE_PORT(ixPCIE_LC_CNTL2, data); - orig = data = RREG32(THM_CLK_CNTL); - data &= ~(CMON_CLK_SEL_MASK | TMON_CLK_SEL_MASK); - data |= CMON_CLK_SEL(1) | TMON_CLK_SEL(1); + orig = data = RREG32(mmTHM_CLK_CNTL); + data &= ~(THM_CLK_CNTL__CMON_CLK_SEL_MASK | THM_CLK_CNTL__TMON_CLK_SEL_MASK); + data |= (1 << THM_CLK_CNTL__CMON_CLK_SEL__SHIFT) | (1 << THM_CLK_CNTL__TMON_CLK_SEL__SHIFT); if (orig != data) - WREG32(THM_CLK_CNTL, data); + WREG32(mmTHM_CLK_CNTL, data); - orig = data = RREG32(MISC_CLK_CNTL); - data &= ~(DEEP_SLEEP_CLK_SEL_MASK | ZCLK_SEL_MASK); - data |= DEEP_SLEEP_CLK_SEL(1) | ZCLK_SEL(1); + orig = data = RREG32(mmMISC_CLK_CNTL); + data &= ~(MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL_MASK | MISC_CLK_CNTL__ZCLK_SEL_MASK); + data |= (1 << MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL__SHIFT) | (1 << MISC_CLK_CNTL__ZCLK_SEL__SHIFT); if (orig != data) - WREG32(MISC_CLK_CNTL, data); + WREG32(mmMISC_CLK_CNTL, data); - orig = data = RREG32(CG_CLKPIN_CNTL); - data &= ~BCLK_AS_XCLK; + orig = data = RREG32(mmCG_CLKPIN_CNTL); + data &= ~CG_CLKPIN_CNTL__BCLK_AS_XCLK_MASK; if (orig != data) - WREG32(CG_CLKPIN_CNTL, data); + WREG32(mmCG_CLKPIN_CNTL, data); - orig = data = RREG32(CG_CLKPIN_CNTL_2); - data &= ~FORCE_BIF_REFCLK_EN; + orig = data = RREG32(mmCG_CLKPIN_CNTL_2); + data &= ~CG_CLKPIN_CNTL_2__FORCE_BIF_REFCLK_EN_MASK; if (orig != data) - WREG32(CG_CLKPIN_CNTL_2, data); + WREG32(mmCG_CLKPIN_CNTL_2, data); - orig = data = RREG32(MPLL_BYPASSCLK_SEL); - data &= ~MPLL_CLKOUT_SEL_MASK; - data |= MPLL_CLKOUT_SEL(4); + orig = data = RREG32(mmMPLL_BYPASSCLK_SEL); + data &= ~MPLL_BYPASSCLK_SEL__MPLL_CLKOUT_SEL_MASK; + data |= 4 << MPLL_BYPASSCLK_SEL__MPLL_CLKOUT_SEL__SHIFT; if (orig != data) - WREG32(MPLL_BYPASSCLK_SEL, data); + WREG32(mmMPLL_BYPASSCLK_SEL, data); - orig = data = RREG32(SPLL_CNTL_MODE); - data &= ~SPLL_REFCLK_SEL_MASK; + orig = data = RREG32(mmSPLL_CNTL_MODE); + data &= ~SPLL_CNTL_MODE__SPLL_REFCLK_SEL_MASK; if (orig != data) - WREG32(SPLL_CNTL_MODE, data); + WREG32(mmSPLL_CNTL_MODE, data); } } } else { if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_CNTL, data); + WREG32_PCIE_PORT(ixPCIE_LC_CNTL, data); } - orig = data = RREG32_PCIE(PCIE_CNTL2); - data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | REPLAY_MEM_LS_EN; + orig = data = RREG32_PCIE(ixPCIE_CNTL2); + data |= PCIE_CNTL2__SLV_MEM_LS_EN_MASK | PCIE_CNTL2__MST_MEM_LS_EN_MASK | PCIE_CNTL2__REPLAY_MEM_LS_EN_MASK; if (orig != data) - WREG32_PCIE(PCIE_CNTL2, data); + WREG32_PCIE(ixPCIE_CNTL2, data); if (!disable_l0s) { - data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL); - if((data & LC_N_FTS_MASK) == LC_N_FTS_MASK) { - data = RREG32_PCIE(PCIE_LC_STATUS1); - if ((data & LC_REVERSE_XMIT) && (data & LC_REVERSE_RCVR)) { - orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL); - data &= ~LC_L0S_INACTIVITY_MASK; + data = RREG32_PCIE_PORT(ixPCIE_LC_N_FTS_CNTL); + if((data & PCIE_LC_N_FTS_CNTL__LC_N_FTS_MASK) == PCIE_LC_N_FTS_CNTL__LC_N_FTS_MASK) { + data = RREG32_PCIE(ixPCIE_LC_STATUS1); + if ((data & PCIE_LC_STATUS1__LC_REVERSE_XMIT_MASK) && (data & PCIE_LC_STATUS1__LC_REVERSE_RCVR_MASK)) { + orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL); + data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK; if (orig != data) - WREG32_PCIE_PORT(PCIE_LC_CNTL, data); + WREG32_PCIE_PORT(ixPCIE_LC_CNTL, data); } } } diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c index e2089c8da71b..7f18e4875287 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dma.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c @@ -27,6 +27,8 @@ #include "si.h" #include "sid.h" +#include "oss/oss_1_0_d.h" +#include "oss/oss_1_0_sh_mask.h" const u32 sdma_offsets[SDMA_MAX_INSTANCE] = { DMA0_REGISTER_OFFSET, @@ -38,17 +40,31 @@ static void si_dma_set_buffer_funcs(struct amdgpu_device *adev); static void si_dma_set_vm_pte_funcs(struct amdgpu_device *adev); static void si_dma_set_irq_funcs(struct amdgpu_device *adev); +/** + * si_dma_ring_get_rptr - get the current read pointer + * + * @ring: amdgpu ring pointer + * + * Get the current rptr from the hardware (SI). + */ static uint64_t si_dma_ring_get_rptr(struct amdgpu_ring *ring) { return *ring->rptr_cpu_addr; } +/** + * si_dma_ring_get_wptr - get the current write pointer + * + * @ring: amdgpu ring pointer + * + * Get the current wptr from the hardware (SI). + */ static uint64_t si_dma_ring_get_wptr(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; u32 me = (ring == &adev->sdma.instance[0].ring) ? 0 : 1; - return (RREG32(DMA_RB_WPTR + sdma_offsets[me]) & 0x3fffc) >> 2; + return (RREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[me]) & 0x3fffc) >> 2; } static void si_dma_ring_set_wptr(struct amdgpu_ring *ring) @@ -56,7 +72,7 @@ static void si_dma_ring_set_wptr(struct amdgpu_ring *ring) struct amdgpu_device *adev = ring->adev; u32 me = (ring == &adev->sdma.instance[0].ring) ? 0 : 1; - WREG32(DMA_RB_WPTR + sdma_offsets[me], (ring->wptr << 2) & 0x3fffc); + WREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[me], (ring->wptr << 2) & 0x3fffc); } static void si_dma_ring_emit_ib(struct amdgpu_ring *ring, @@ -117,9 +133,9 @@ static void si_dma_stop(struct amdgpu_device *adev) for (i = 0; i < adev->sdma.num_instances; i++) { /* dma0 */ - rb_cntl = RREG32(DMA_RB_CNTL + sdma_offsets[i]); - rb_cntl &= ~DMA_RB_ENABLE; - WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl); + rb_cntl = RREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i]); + rb_cntl &= ~DMA_GFX_RB_CNTL__RB_ENABLE_MASK; + WREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i], rb_cntl); } } @@ -133,44 +149,44 @@ static int si_dma_start(struct amdgpu_device *adev) for (i = 0; i < adev->sdma.num_instances; i++) { ring = &adev->sdma.instance[i].ring; - WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL + sdma_offsets[i], 0); - WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0); + WREG32(mmDMA_SEM_INCOMPLETE_TIMER_CNTL + sdma_offsets[i], 0); + WREG32(mmDMA_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0); /* Set ring buffer size in dwords */ rb_bufsz = order_base_2(ring->ring_size / 4); rb_cntl = rb_bufsz << 1; #ifdef __BIG_ENDIAN - rb_cntl |= DMA_RB_SWAP_ENABLE | DMA_RPTR_WRITEBACK_SWAP_ENABLE; + rb_cntl |= DMA_GFX_RB_CNTL__RB_SWAP_ENABLE_MASK | DMA_GFX_RB_CNTL__RPTR_WRITEBACK_SWAP_ENABLE_MASK; #endif - WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl); + WREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i], rb_cntl); /* Initialize the ring buffer's read and write pointers */ - WREG32(DMA_RB_RPTR + sdma_offsets[i], 0); - WREG32(DMA_RB_WPTR + sdma_offsets[i], 0); + WREG32(mmDMA_GFX_RB_RPTR + sdma_offsets[i], 0); + WREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[i], 0); rptr_addr = ring->rptr_gpu_addr; - WREG32(DMA_RB_RPTR_ADDR_LO + sdma_offsets[i], lower_32_bits(rptr_addr)); - WREG32(DMA_RB_RPTR_ADDR_HI + sdma_offsets[i], upper_32_bits(rptr_addr) & 0xFF); + WREG32(mmDMA_GFX_RB_RPTR_ADDR_LO + sdma_offsets[i], lower_32_bits(rptr_addr)); + WREG32(mmDMA_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i], upper_32_bits(rptr_addr) & 0xFF); - rb_cntl |= DMA_RPTR_WRITEBACK_ENABLE; + rb_cntl |= DMA_GFX_RB_CNTL__RPTR_WRITEBACK_ENABLE_MASK; - WREG32(DMA_RB_BASE + sdma_offsets[i], ring->gpu_addr >> 8); + WREG32(mmDMA_GFX_RB_BASE + sdma_offsets[i], ring->gpu_addr >> 8); /* enable DMA IBs */ - ib_cntl = DMA_IB_ENABLE | CMD_VMID_FORCE; + ib_cntl = DMA_GFX_IB_CNTL__IB_ENABLE_MASK | DMA_GFX_IB_CNTL__CMD_VMID_FORCE_MASK; #ifdef __BIG_ENDIAN - ib_cntl |= DMA_IB_SWAP_ENABLE; + ib_cntl |= DMA_GFX_IB_CNTL__IB_SWAP_ENABLE_MASK; #endif - WREG32(DMA_IB_CNTL + sdma_offsets[i], ib_cntl); + WREG32(mmDMA_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); - dma_cntl = RREG32(DMA_CNTL + sdma_offsets[i]); - dma_cntl &= ~CTXEMPTY_INT_ENABLE; - WREG32(DMA_CNTL + sdma_offsets[i], dma_cntl); + dma_cntl = RREG32(mmDMA_CNTL + sdma_offsets[i]); + dma_cntl &= ~DMA_CNTL__CTXEMPTY_INT_ENABLE_MASK; + WREG32(mmDMA_CNTL + sdma_offsets[i], dma_cntl); ring->wptr = 0; - WREG32(DMA_RB_WPTR + sdma_offsets[i], ring->wptr << 2); - WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl | DMA_RB_ENABLE); + WREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[i], ring->wptr << 2); + WREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i], rb_cntl | DMA_GFX_RB_CNTL__RB_ENABLE_MASK); r = amdgpu_ring_test_helper(ring); if (r) @@ -461,7 +477,7 @@ static int si_dma_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - adev->sdma.num_instances = 2; + adev->sdma.num_instances = SDMA_MAX_INSTANCE; si_dma_set_ring_funcs(adev); si_dma_set_buffer_funcs(adev); @@ -545,9 +561,9 @@ static bool si_dma_is_idle(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - u32 tmp = RREG32(SRBM_STATUS2); + u32 tmp = RREG32(mmSRBM_STATUS2); - if (tmp & (DMA_BUSY_MASK | DMA1_BUSY_MASK)) + if (tmp & (SRBM_STATUS2__DMA_BUSY_MASK | SRBM_STATUS2__DMA1_BUSY_MASK)) return false; return true; @@ -583,14 +599,14 @@ static int si_dma_set_trap_irq_state(struct amdgpu_device *adev, case AMDGPU_SDMA_IRQ_INSTANCE0: switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - sdma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET); - sdma_cntl &= ~TRAP_ENABLE; - WREG32(DMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl); + sdma_cntl = RREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET); + sdma_cntl &= ~DMA_CNTL__TRAP_ENABLE_MASK; + WREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl); break; case AMDGPU_IRQ_STATE_ENABLE: - sdma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET); - sdma_cntl |= TRAP_ENABLE; - WREG32(DMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl); + sdma_cntl = RREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET); + sdma_cntl |= DMA_CNTL__TRAP_ENABLE_MASK; + WREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl); break; default: break; @@ -599,14 +615,14 @@ static int si_dma_set_trap_irq_state(struct amdgpu_device *adev, case AMDGPU_SDMA_IRQ_INSTANCE1: switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - sdma_cntl = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET); - sdma_cntl &= ~TRAP_ENABLE; - WREG32(DMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl); + sdma_cntl = RREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET); + sdma_cntl &= ~DMA_CNTL__TRAP_ENABLE_MASK; + WREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl); break; case AMDGPU_IRQ_STATE_ENABLE: - sdma_cntl = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET); - sdma_cntl |= TRAP_ENABLE; - WREG32(DMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl); + sdma_cntl = RREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET); + sdma_cntl |= DMA_CNTL__TRAP_ENABLE_MASK; + WREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl); break; default: break; @@ -645,11 +661,11 @@ static int si_dma_set_clockgating_state(struct amdgpu_ip_block *ip_block, offset = DMA0_REGISTER_OFFSET; else offset = DMA1_REGISTER_OFFSET; - orig = data = RREG32(DMA_POWER_CNTL + offset); - data &= ~MEM_POWER_OVERRIDE; + orig = data = RREG32(mmDMA_POWER_CNTL + offset); + data &= ~DMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK; if (data != orig) - WREG32(DMA_POWER_CNTL + offset, data); - WREG32(DMA_CLK_CTRL + offset, 0x00000100); + WREG32(mmDMA_POWER_CNTL + offset, data); + WREG32(mmDMA_CLK_CTRL + offset, 0x00000100); } } else { for (i = 0; i < adev->sdma.num_instances; i++) { @@ -657,15 +673,15 @@ static int si_dma_set_clockgating_state(struct amdgpu_ip_block *ip_block, offset = DMA0_REGISTER_OFFSET; else offset = DMA1_REGISTER_OFFSET; - orig = data = RREG32(DMA_POWER_CNTL + offset); - data |= MEM_POWER_OVERRIDE; + orig = data = RREG32(mmDMA_POWER_CNTL + offset); + data |= DMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK; if (data != orig) - WREG32(DMA_POWER_CNTL + offset, data); + WREG32(mmDMA_POWER_CNTL + offset, data); - orig = data = RREG32(DMA_CLK_CTRL + offset); + orig = data = RREG32(mmDMA_CLK_CTRL + offset); data = 0xff000000; if (data != orig) - WREG32(DMA_CLK_CTRL + offset, data); + WREG32(mmDMA_CLK_CTRL + offset, data); } } @@ -679,11 +695,11 @@ static int si_dma_set_powergating_state(struct amdgpu_ip_block *ip_block, struct amdgpu_device *adev = ip_block->adev; - WREG32(DMA_PGFSM_WRITE, 0x00002000); - WREG32(DMA_PGFSM_CONFIG, 0x100010ff); + WREG32(mmDMA_PGFSM_WRITE, 0x00002000); + WREG32(mmDMA_PGFSM_CONFIG, 0x100010ff); for (tmp = 0; tmp < 5; tmp++) - WREG32(DMA_PGFSM_WRITE, 0); + WREG32(mmDMA_PGFSM_WRITE, 0); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/si_enums.h b/drivers/gpu/drm/amd/amdgpu/si_enums.h index d656ef1fa6e1..6da65778292b 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_enums.h +++ b/drivers/gpu/drm/amd/amdgpu/si_enums.h @@ -23,115 +23,15 @@ #ifndef SI_ENUMS_H #define SI_ENUMS_H -#define VBLANK_INT_MASK (1 << 0) -#define DC_HPDx_INT_EN (1 << 16) -#define VBLANK_ACK (1 << 4) -#define VLINE_ACK (1 << 4) - -#define CURSOR_WIDTH 64 -#define CURSOR_HEIGHT 64 - -#define VGA_VSTATUS_CNTL 0xFFFCFFFF #define PRIORITY_MARK_MASK 0x7fff #define PRIORITY_OFF (1 << 16) #define PRIORITY_ALWAYS_ON (1 << 20) -#define INTERLEAVE_EN (1 << 0) - -#define LATENCY_WATERMARK_MASK(x) ((x) << 16) -#define DC_LB_MEMORY_CONFIG(x) ((x) << 20) -#define ICON_DEGAMMA_MODE(x) (((x) & 0x3) << 8) - -#define GRPH_ENDIAN_SWAP(x) (((x) & 0x3) << 0) -#define GRPH_ENDIAN_NONE 0 -#define GRPH_ENDIAN_8IN16 1 -#define GRPH_ENDIAN_8IN32 2 -#define GRPH_ENDIAN_8IN64 3 -#define GRPH_RED_CROSSBAR(x) (((x) & 0x3) << 4) -#define GRPH_RED_SEL_R 0 -#define GRPH_RED_SEL_G 1 -#define GRPH_RED_SEL_B 2 -#define GRPH_RED_SEL_A 3 -#define GRPH_GREEN_CROSSBAR(x) (((x) & 0x3) << 6) -#define GRPH_GREEN_SEL_G 0 -#define GRPH_GREEN_SEL_B 1 -#define GRPH_GREEN_SEL_A 2 -#define GRPH_GREEN_SEL_R 3 -#define GRPH_BLUE_CROSSBAR(x) (((x) & 0x3) << 8) -#define GRPH_BLUE_SEL_B 0 -#define GRPH_BLUE_SEL_A 1 -#define GRPH_BLUE_SEL_R 2 -#define GRPH_BLUE_SEL_G 3 -#define GRPH_ALPHA_CROSSBAR(x) (((x) & 0x3) << 10) -#define GRPH_ALPHA_SEL_A 0 -#define GRPH_ALPHA_SEL_R 1 -#define GRPH_ALPHA_SEL_G 2 -#define GRPH_ALPHA_SEL_B 3 - -#define GRPH_DEPTH(x) (((x) & 0x3) << 0) -#define GRPH_DEPTH_8BPP 0 -#define GRPH_DEPTH_16BPP 1 -#define GRPH_DEPTH_32BPP 2 - -#define GRPH_FORMAT(x) (((x) & 0x7) << 8) -#define GRPH_FORMAT_INDEXED 0 -#define GRPH_FORMAT_ARGB1555 0 -#define GRPH_FORMAT_ARGB565 1 -#define GRPH_FORMAT_ARGB4444 2 -#define GRPH_FORMAT_AI88 3 -#define GRPH_FORMAT_MONO16 4 -#define GRPH_FORMAT_BGRA5551 5 -#define GRPH_FORMAT_ARGB8888 0 -#define GRPH_FORMAT_ARGB2101010 1 -#define GRPH_FORMAT_32BPP_DIG 2 -#define GRPH_FORMAT_8B_ARGB2101010 3 -#define GRPH_FORMAT_BGRA1010102 4 -#define GRPH_FORMAT_8B_BGRA1010102 5 -#define GRPH_FORMAT_RGB111110 6 -#define GRPH_FORMAT_BGR101111 7 - -#define GRPH_NUM_BANKS(x) (((x) & 0x3) << 2) -#define GRPH_ARRAY_MODE(x) (((x) & 0x7) << 20) -#define GRPH_ARRAY_LINEAR_GENERAL 0 -#define GRPH_ARRAY_LINEAR_ALIGNED 1 -#define GRPH_ARRAY_1D_TILED_THIN1 2 -#define GRPH_ARRAY_2D_TILED_THIN1 4 -#define GRPH_TILE_SPLIT(x) (((x) & 0x7) << 13) -#define GRPH_BANK_WIDTH(x) (((x) & 0x3) << 6) -#define GRPH_BANK_HEIGHT(x) (((x) & 0x3) << 11) -#define GRPH_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 18) -#define GRPH_ARRAY_MODE(x) (((x) & 0x7) << 20) -#define GRPH_PIPE_CONFIG(x) (((x) & 0x1f) << 24) - -#define CURSOR_EN (1 << 0) -#define CURSOR_MODE(x) (((x) & 0x3) << 8) -#define CURSOR_MONO 0 -#define CURSOR_24_1 1 -#define CURSOR_24_8_PRE_MULT 2 -#define CURSOR_24_8_UNPRE_MULT 3 -#define CURSOR_2X_MAGNIFY (1 << 16) -#define CURSOR_FORCE_MC_ON (1 << 20) -#define CURSOR_URGENT_CONTROL(x) (((x) & 0x7) << 24) -#define CURSOR_URGENT_ALWAYS 0 -#define CURSOR_URGENT_1_8 1 -#define CURSOR_URGENT_1_4 2 -#define CURSOR_URGENT_3_8 3 -#define CURSOR_URGENT_1_2 4 -#define CURSOR_UPDATE_PENDING (1 << 0) -#define CURSOR_UPDATE_TAKEN (1 << 1) -#define CURSOR_UPDATE_LOCK (1 << 16) -#define CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24) - - -#define ES_AND_GS_AUTO 3 -#define RADEON_PACKET_TYPE3 3 -#define CE_PARTITION_BASE 3 -#define BUF_SWAP_32BIT (2 << 16) #define GFX_POWER_STATUS (1 << 1) #define GFX_CLOCK_STATUS (1 << 2) #define GFX_LS_STATUS (1 << 3) -#define RLC_BUSY_STATUS (1 << 0) +#define RLC_BUSY_STATUS (1 << 0) #define RLC_PUD(x) ((x) << 0) #define RLC_PUD_MASK (0xff << 0) #define RLC_PDD(x) ((x) << 8) @@ -140,140 +40,8 @@ #define RLC_TTPD_MASK (0xff << 16) #define RLC_MSD(x) ((x) << 24) #define RLC_MSD_MASK (0xff << 24) -#define WRITE_DATA_ENGINE_SEL(x) ((x) << 30) -#define WRITE_DATA_DST_SEL(x) ((x) << 8) -#define EVENT_TYPE(x) ((x) << 0) -#define EVENT_INDEX(x) ((x) << 8) -#define WAIT_REG_MEM_MEM_SPACE(x) ((x) << 4) -#define WAIT_REG_MEM_FUNCTION(x) ((x) << 0) -#define WAIT_REG_MEM_ENGINE(x) ((x) << 8) -#define GFX6_NUM_GFX_RINGS 1 -#define GFX6_NUM_COMPUTE_RINGS 2 #define RLC_SAVE_AND_RESTORE_STARTING_OFFSET 0x90 #define RLC_CLEAR_STATE_DESCRIPTOR_OFFSET 0x3D -#define PACKET3(op, n) ((RADEON_PACKET_TYPE3 << 30) | \ - (((op) & 0xFF) << 8) | \ - ((n) & 0x3FFF) << 16) -#define PACKET3_COMPUTE(op, n) (PACKET3(op, n) | 1 << 1) -#define PACKET3_NOP 0x10 -#define PACKET3_SET_BASE 0x11 -#define PACKET3_BASE_INDEX(x) ((x) << 0) -#define PACKET3_CLEAR_STATE 0x12 -#define PACKET3_INDEX_BUFFER_SIZE 0x13 -#define PACKET3_DISPATCH_DIRECT 0x15 -#define PACKET3_DISPATCH_INDIRECT 0x16 -#define PACKET3_ALLOC_GDS 0x1B -#define PACKET3_WRITE_GDS_RAM 0x1C -#define PACKET3_ATOMIC_GDS 0x1D -#define PACKET3_ATOMIC 0x1E -#define PACKET3_OCCLUSION_QUERY 0x1F -#define PACKET3_SET_PREDICATION 0x20 -#define PACKET3_REG_RMW 0x21 -#define PACKET3_COND_EXEC 0x22 -#define PACKET3_PRED_EXEC 0x23 -#define PACKET3_DRAW_INDIRECT 0x24 -#define PACKET3_DRAW_INDEX_INDIRECT 0x25 -#define PACKET3_INDEX_BASE 0x26 -#define PACKET3_DRAW_INDEX_2 0x27 -#define PACKET3_CONTEXT_CONTROL 0x28 -#define PACKET3_INDEX_TYPE 0x2A -#define PACKET3_DRAW_INDIRECT_MULTI 0x2C -#define PACKET3_DRAW_INDEX_AUTO 0x2D -#define PACKET3_DRAW_INDEX_IMMD 0x2E -#define PACKET3_NUM_INSTANCES 0x2F -#define PACKET3_DRAW_INDEX_MULTI_AUTO 0x30 -#define PACKET3_INDIRECT_BUFFER_CONST 0x31 -#define PACKET3_INDIRECT_BUFFER 0x3F -#define PACKET3_STRMOUT_BUFFER_UPDATE 0x34 -#define PACKET3_DRAW_INDEX_OFFSET_2 0x35 -#define PACKET3_DRAW_INDEX_MULTI_ELEMENT 0x36 -#define PACKET3_WRITE_DATA 0x37 -#define PACKET3_DRAW_INDEX_INDIRECT_MULTI 0x38 -#define PACKET3_MEM_SEMAPHORE 0x39 -#define PACKET3_MPEG_INDEX 0x3A -#define PACKET3_COPY_DW 0x3B -#define PACKET3_WAIT_REG_MEM 0x3C -#define PACKET3_MEM_WRITE 0x3D -#define PACKET3_COPY_DATA 0x40 -#define PACKET3_CP_DMA 0x41 -# define PACKET3_CP_DMA_DST_SEL(x) ((x) << 20) -# define PACKET3_CP_DMA_ENGINE(x) ((x) << 27) -# define PACKET3_CP_DMA_SRC_SEL(x) ((x) << 29) -# define PACKET3_CP_DMA_CP_SYNC (1 << 31) -# define PACKET3_CP_DMA_DIS_WC (1 << 21) -# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 22) -# define PACKET3_CP_DMA_CMD_DST_SWAP(x) ((x) << 24) -# define PACKET3_CP_DMA_CMD_SAS (1 << 26) -# define PACKET3_CP_DMA_CMD_DAS (1 << 27) -# define PACKET3_CP_DMA_CMD_SAIC (1 << 28) -# define PACKET3_CP_DMA_CMD_DAIC (1 << 29) -# define PACKET3_CP_DMA_CMD_RAW_WAIT (1 << 30) -#define PACKET3_PFP_SYNC_ME 0x42 -#define PACKET3_SURFACE_SYNC 0x43 -# define PACKET3_DEST_BASE_0_ENA (1 << 0) -# define PACKET3_DEST_BASE_1_ENA (1 << 1) -# define PACKET3_CB0_DEST_BASE_ENA (1 << 6) -# define PACKET3_CB1_DEST_BASE_ENA (1 << 7) -# define PACKET3_CB2_DEST_BASE_ENA (1 << 8) -# define PACKET3_CB3_DEST_BASE_ENA (1 << 9) -# define PACKET3_CB4_DEST_BASE_ENA (1 << 10) -# define PACKET3_CB5_DEST_BASE_ENA (1 << 11) -# define PACKET3_CB6_DEST_BASE_ENA (1 << 12) -# define PACKET3_CB7_DEST_BASE_ENA (1 << 13) -# define PACKET3_DB_DEST_BASE_ENA (1 << 14) -# define PACKET3_DEST_BASE_2_ENA (1 << 19) -# define PACKET3_DEST_BASE_3_ENA (1 << 21) -# define PACKET3_TCL1_ACTION_ENA (1 << 22) -# define PACKET3_TC_ACTION_ENA (1 << 23) -# define PACKET3_CB_ACTION_ENA (1 << 25) -# define PACKET3_DB_ACTION_ENA (1 << 26) -# define PACKET3_SH_KCACHE_ACTION_ENA (1 << 27) -# define PACKET3_SH_ICACHE_ACTION_ENA (1 << 29) -#define PACKET3_ME_INITIALIZE 0x44 -#define PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16) -#define PACKET3_COND_WRITE 0x45 -#define PACKET3_EVENT_WRITE 0x46 -#define PACKET3_EVENT_WRITE_EOP 0x47 -#define PACKET3_EVENT_WRITE_EOS 0x48 -#define PACKET3_PREAMBLE_CNTL 0x4A -# define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE (2 << 28) -# define PACKET3_PREAMBLE_END_CLEAR_STATE (3 << 28) -#define PACKET3_ONE_REG_WRITE 0x57 -#define PACKET3_LOAD_CONFIG_REG 0x5F -#define PACKET3_LOAD_CONTEXT_REG 0x60 -#define PACKET3_LOAD_SH_REG 0x61 -#define PACKET3_SET_CONFIG_REG 0x68 -#define PACKET3_SET_CONFIG_REG_START 0x00002000 -#define PACKET3_SET_CONFIG_REG_END 0x00002c00 -#define PACKET3_SET_CONTEXT_REG 0x69 -#define PACKET3_SET_CONTEXT_REG_START 0x000a000 -#define PACKET3_SET_CONTEXT_REG_END 0x000a400 -#define PACKET3_SET_CONTEXT_REG_INDIRECT 0x73 -#define PACKET3_SET_RESOURCE_INDIRECT 0x74 -#define PACKET3_SET_SH_REG 0x76 -#define PACKET3_SET_SH_REG_START 0x00002c00 -#define PACKET3_SET_SH_REG_END 0x00003000 -#define PACKET3_SET_SH_REG_OFFSET 0x77 -#define PACKET3_ME_WRITE 0x7A -#define PACKET3_SCRATCH_RAM_WRITE 0x7D -#define PACKET3_SCRATCH_RAM_READ 0x7E -#define PACKET3_CE_WRITE 0x7F -#define PACKET3_LOAD_CONST_RAM 0x80 -#define PACKET3_WRITE_CONST_RAM 0x81 -#define PACKET3_WRITE_CONST_RAM_OFFSET 0x82 -#define PACKET3_DUMP_CONST_RAM 0x83 -#define PACKET3_INCREMENT_CE_COUNTER 0x84 -#define PACKET3_INCREMENT_DE_COUNTER 0x85 -#define PACKET3_WAIT_ON_CE_COUNTER 0x86 -#define PACKET3_WAIT_ON_DE_COUNTER 0x87 -#define PACKET3_WAIT_ON_DE_COUNTER_DIFF 0x88 -#define PACKET3_SET_CE_DE_COUNTERS 0x89 -#define PACKET3_WAIT_ON_AVAIL_BUFFER 0x8A -#define PACKET3_SWITCH_BUFFER 0x8B -#define PACKET3_SEM_WAIT_ON_SIGNAL (0x1 << 12) -#define PACKET3_SEM_SEL_SIGNAL (0x6 << 29) -#define PACKET3_SEM_SEL_WAIT (0x7 << 29) - #endif diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c index 5c38e1fb1dca..1df00f8a2406 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c @@ -27,6 +27,7 @@ #include "amdgpu_ih.h" #include "sid.h" #include "si_ih.h" + #include "oss/oss_1_0_d.h" #include "oss/oss_1_0_sh_mask.h" @@ -213,7 +214,7 @@ static int si_ih_resume(struct amdgpu_ip_block *ip_block) static bool si_ih_is_idle(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - u32 tmp = RREG32(SRBM_STATUS); + u32 tmp = RREG32(mmSRBM_STATUS); if (tmp & SRBM_STATUS__IH_BUSY_MASK) return false; @@ -239,23 +240,23 @@ static int si_ih_soft_reset(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; u32 srbm_soft_reset = 0; - u32 tmp = RREG32(SRBM_STATUS); + u32 tmp = RREG32(mmSRBM_STATUS); if (tmp & SRBM_STATUS__IH_BUSY_MASK) srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_IH_MASK; if (srbm_soft_reset) { - tmp = RREG32(SRBM_SOFT_RESET); + tmp = RREG32(mmSRBM_SOFT_RESET); tmp |= srbm_soft_reset; - dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32(SRBM_SOFT_RESET, tmp); - tmp = RREG32(SRBM_SOFT_RESET); + dev_info(adev->dev, "mmSRBM_SOFT_RESET=0x%08X\n", tmp); + WREG32(mmSRBM_SOFT_RESET, tmp); + tmp = RREG32(mmSRBM_SOFT_RESET); udelay(50); tmp &= ~srbm_soft_reset; - WREG32(SRBM_SOFT_RESET, tmp); - tmp = RREG32(SRBM_SOFT_RESET); + WREG32(mmSRBM_SOFT_RESET, tmp); + tmp = RREG32(mmSRBM_SOFT_RESET); udelay(50); } diff --git a/drivers/gpu/drm/amd/amdgpu/sid.h b/drivers/gpu/drm/amd/amdgpu/sid.h index cbf232f5235b..cbd4f8951cfa 100644 --- a/drivers/gpu/drm/amd/amdgpu/sid.h +++ b/drivers/gpu/drm/amd/amdgpu/sid.h @@ -24,43 +24,12 @@ #ifndef SI_H #define SI_H -#define TAHITI_RB_BITMAP_WIDTH_PER_SH 2 - -#define SI_MAX_SH_GPRS 256 -#define SI_MAX_TEMP_GPRS 16 -#define SI_MAX_SH_THREADS 256 -#define SI_MAX_SH_STACK_ENTRIES 4096 -#define SI_MAX_FRC_EOV_CNT 16384 -#define SI_MAX_BACKENDS 8 -#define SI_MAX_BACKENDS_MASK 0xFF -#define SI_MAX_BACKENDS_PER_SE_MASK 0x0F -#define SI_MAX_SIMDS 12 -#define SI_MAX_SIMDS_MASK 0x0FFF -#define SI_MAX_SIMDS_PER_SE_MASK 0x00FF -#define SI_MAX_PIPES 8 -#define SI_MAX_PIPES_MASK 0xFF -#define SI_MAX_PIPES_PER_SIMD_MASK 0x3F -#define SI_MAX_LDS_NUM 0xFFFF -#define SI_MAX_TCC 16 -#define SI_MAX_TCC_MASK 0xFFFF #define SI_MAX_CTLACKS_ASSERTION_WAIT 100 -/* SMC IND accessor regs */ -#define SMC_IND_INDEX_0 0x80 -#define SMC_IND_DATA_0 0x81 - -#define SMC_IND_ACCESS_CNTL 0x8A -# define AUTO_INCREMENT_IND_0 (1 << 0) -#define SMC_MESSAGE_0 0x8B -#define SMC_RESP_0 0x8C - /* CG IND registers are accessed via SMC indirect space + SMC_CG_IND_START */ #define SMC_CG_IND_START 0xc0030000 #define SMC_CG_IND_END 0xc0040000 -#define CG_CGTT_LOCAL_0 0x400 -#define CG_CGTT_LOCAL_1 0x401 - /* SMC IND registers */ #define SMC_SYSCON_RESET_CNTL 0x80000000 # define RST_REG (1 << 0) @@ -68,9 +37,6 @@ # define CK_DISABLE (1 << 0) # define CKEN (1 << 24) -#define VGA_HDP_CONTROL 0xCA -#define VGA_MEMORY_DISABLE (1 << 4) - #define DCCG_DISP_SLOW_SELECT_REG 0x13F #define DCCG_DISP1_SLOW_SELECT(x) ((x) << 0) #define DCCG_DISP1_SLOW_SELECT_MASK (7 << 0) @@ -79,47 +45,6 @@ #define DCCG_DISP2_SLOW_SELECT_MASK (7 << 4) #define DCCG_DISP2_SLOW_SELECT_SHIFT 4 -#define CG_SPLL_FUNC_CNTL 0x180 -#define SPLL_RESET (1 << 0) -#define SPLL_SLEEP (1 << 1) -#define SPLL_BYPASS_EN (1 << 3) -#define SPLL_REF_DIV(x) ((x) << 4) -#define SPLL_REF_DIV_MASK (0x3f << 4) -#define SPLL_PDIV_A(x) ((x) << 20) -#define SPLL_PDIV_A_MASK (0x7f << 20) -#define SPLL_PDIV_A_SHIFT 20 -#define CG_SPLL_FUNC_CNTL_2 0x181 -#define SCLK_MUX_SEL(x) ((x) << 0) -#define SCLK_MUX_SEL_MASK (0x1ff << 0) -#define SPLL_CTLREQ_CHG (1 << 23) -#define SCLK_MUX_UPDATE (1 << 26) -#define CG_SPLL_FUNC_CNTL_3 0x182 -#define SPLL_FB_DIV(x) ((x) << 0) -#define SPLL_FB_DIV_MASK (0x3ffffff << 0) -#define SPLL_FB_DIV_SHIFT 0 -#define SPLL_DITHEN (1 << 28) -#define CG_SPLL_FUNC_CNTL_4 0x183 - -#define SPLL_STATUS 0x185 -#define SPLL_CHG_STATUS (1 << 1) -#define SPLL_CNTL_MODE 0x186 -#define SPLL_SW_DIR_CONTROL (1 << 0) -# define SPLL_REFCLK_SEL(x) ((x) << 26) -# define SPLL_REFCLK_SEL_MASK (3 << 26) - -#define CG_SPLL_SPREAD_SPECTRUM 0x188 -#define SSEN (1 << 0) -#define CLK_S(x) ((x) << 4) -#define CLK_S_MASK (0xfff << 4) -#define CLK_S_SHIFT 4 -#define CG_SPLL_SPREAD_SPECTRUM_2 0x189 -#define CLK_V(x) ((x) << 0) -#define CLK_V_MASK (0x3ffffff << 0) -#define CLK_V_SHIFT 0 - -#define CG_SPLL_AUTOSCALE_CNTL 0x18b -# define AUTOSCALE_ON_SS_CLEAR (1 << 9) - /* discrete uvd clocks */ #define CG_UPLL_FUNC_CNTL 0x18d # define UPLL_RESET_MASK 0x00000001 @@ -149,317 +74,13 @@ #define CG_UPLL_SPREAD_SPECTRUM 0x194 # define SSEN_MASK 0x00000001 -#define MPLL_BYPASSCLK_SEL 0x197 -# define MPLL_CLKOUT_SEL(x) ((x) << 8) -# define MPLL_CLKOUT_SEL_MASK 0xFF00 - -#define CG_CLKPIN_CNTL 0x198 -# define XTALIN_DIVIDE (1 << 1) -# define BCLK_AS_XCLK (1 << 2) -#define CG_CLKPIN_CNTL_2 0x199 -# define FORCE_BIF_REFCLK_EN (1 << 3) -# define MUX_TCLK_TO_XCLK (1 << 8) - -#define THM_CLK_CNTL 0x19b -# define CMON_CLK_SEL(x) ((x) << 0) -# define CMON_CLK_SEL_MASK 0xFF -# define TMON_CLK_SEL(x) ((x) << 8) -# define TMON_CLK_SEL_MASK 0xFF00 -#define MISC_CLK_CNTL 0x19c -# define DEEP_SLEEP_CLK_SEL(x) ((x) << 0) -# define DEEP_SLEEP_CLK_SEL_MASK 0xFF -# define ZCLK_SEL(x) ((x) << 8) -# define ZCLK_SEL_MASK 0xFF00 - -#define CG_THERMAL_CTRL 0x1c0 -#define DPM_EVENT_SRC(x) ((x) << 0) -#define DPM_EVENT_SRC_MASK (7 << 0) -#define DIG_THERM_DPM(x) ((x) << 14) -#define DIG_THERM_DPM_MASK 0x003FC000 -#define DIG_THERM_DPM_SHIFT 14 -#define CG_THERMAL_STATUS 0x1c1 -#define FDO_PWM_DUTY(x) ((x) << 9) -#define FDO_PWM_DUTY_MASK (0xff << 9) -#define FDO_PWM_DUTY_SHIFT 9 -#define CG_THERMAL_INT 0x1c2 -#define DIG_THERM_INTH(x) ((x) << 8) -#define DIG_THERM_INTH_MASK 0x0000FF00 -#define DIG_THERM_INTH_SHIFT 8 -#define DIG_THERM_INTL(x) ((x) << 16) -#define DIG_THERM_INTL_MASK 0x00FF0000 -#define DIG_THERM_INTL_SHIFT 16 -#define THERM_INT_MASK_HIGH (1 << 24) -#define THERM_INT_MASK_LOW (1 << 25) - -#define CG_MULT_THERMAL_CTRL 0x1c4 -#define TEMP_SEL(x) ((x) << 20) -#define TEMP_SEL_MASK (0xff << 20) -#define TEMP_SEL_SHIFT 20 -#define CG_MULT_THERMAL_STATUS 0x1c5 -#define ASIC_MAX_TEMP(x) ((x) << 0) -#define ASIC_MAX_TEMP_MASK 0x000001ff -#define ASIC_MAX_TEMP_SHIFT 0 -#define CTF_TEMP(x) ((x) << 9) -#define CTF_TEMP_MASK 0x0003fe00 -#define CTF_TEMP_SHIFT 9 - -#define CG_FDO_CTRL0 0x1d5 -#define FDO_STATIC_DUTY(x) ((x) << 0) -#define FDO_STATIC_DUTY_MASK 0x000000FF -#define FDO_STATIC_DUTY_SHIFT 0 -#define CG_FDO_CTRL1 0x1d6 -#define FMAX_DUTY100(x) ((x) << 0) -#define FMAX_DUTY100_MASK 0x000000FF -#define FMAX_DUTY100_SHIFT 0 -#define CG_FDO_CTRL2 0x1d7 -#define TMIN(x) ((x) << 0) -#define TMIN_MASK 0x000000FF -#define TMIN_SHIFT 0 -#define FDO_PWM_MODE(x) ((x) << 11) -#define FDO_PWM_MODE_MASK (7 << 11) -#define FDO_PWM_MODE_SHIFT 11 -#define TACH_PWM_RESP_RATE(x) ((x) << 25) -#define TACH_PWM_RESP_RATE_MASK (0x7f << 25) -#define TACH_PWM_RESP_RATE_SHIFT 25 - -#define CG_TACH_CTRL 0x1dc -# define EDGE_PER_REV(x) ((x) << 0) -# define EDGE_PER_REV_MASK (0x7 << 0) -# define EDGE_PER_REV_SHIFT 0 -# define TARGET_PERIOD(x) ((x) << 3) -# define TARGET_PERIOD_MASK 0xfffffff8 -# define TARGET_PERIOD_SHIFT 3 -#define CG_TACH_STATUS 0x1dd -# define TACH_PERIOD(x) ((x) << 0) -# define TACH_PERIOD_MASK 0xffffffff -# define TACH_PERIOD_SHIFT 0 - -#define GENERAL_PWRMGT 0x1e0 -# define GLOBAL_PWRMGT_EN (1 << 0) -# define STATIC_PM_EN (1 << 1) -# define THERMAL_PROTECTION_DIS (1 << 2) -# define THERMAL_PROTECTION_TYPE (1 << 3) -# define SW_SMIO_INDEX(x) ((x) << 6) -# define SW_SMIO_INDEX_MASK (1 << 6) -# define SW_SMIO_INDEX_SHIFT 6 -# define VOLT_PWRMGT_EN (1 << 10) -# define DYN_SPREAD_SPECTRUM_EN (1 << 23) -#define CG_TPC 0x1e1 -#define SCLK_PWRMGT_CNTL 0x1e2 -# define SCLK_PWRMGT_OFF (1 << 0) -# define SCLK_LOW_D1 (1 << 1) -# define FIR_RESET (1 << 4) -# define FIR_FORCE_TREND_SEL (1 << 5) -# define FIR_TREND_MODE (1 << 6) -# define DYN_GFX_CLK_OFF_EN (1 << 7) -# define GFX_CLK_FORCE_ON (1 << 8) -# define GFX_CLK_REQUEST_OFF (1 << 9) -# define GFX_CLK_FORCE_OFF (1 << 10) -# define GFX_CLK_OFF_ACPI_D1 (1 << 11) -# define GFX_CLK_OFF_ACPI_D2 (1 << 12) -# define GFX_CLK_OFF_ACPI_D3 (1 << 13) -# define DYN_LIGHT_SLEEP_EN (1 << 14) - -#define TARGET_AND_CURRENT_PROFILE_INDEX 0x1e6 -# define CURRENT_STATE_INDEX_MASK (0xf << 4) -# define CURRENT_STATE_INDEX_SHIFT 4 - -#define CG_FTV 0x1ef - -#define CG_FFCT_0 0x1f0 -# define UTC_0(x) ((x) << 0) -# define UTC_0_MASK (0x3ff << 0) -# define DTC_0(x) ((x) << 10) -# define DTC_0_MASK (0x3ff << 10) - -#define CG_BSP 0x1ff -# define BSP(x) ((x) << 0) -# define BSP_MASK (0xffff << 0) -# define BSU(x) ((x) << 16) -# define BSU_MASK (0xf << 16) -#define CG_AT 0x200 -# define CG_R(x) ((x) << 0) -# define CG_R_MASK (0xffff << 0) -# define CG_L(x) ((x) << 16) -# define CG_L_MASK (0xffff << 16) - -#define CG_GIT 0x201 -# define CG_GICST(x) ((x) << 0) -# define CG_GICST_MASK (0xffff << 0) -# define CG_GIPOT(x) ((x) << 16) -# define CG_GIPOT_MASK (0xffff << 16) - -#define CG_SSP 0x203 -# define SST(x) ((x) << 0) -# define SST_MASK (0xffff << 0) -# define SSTU(x) ((x) << 16) -# define SSTU_MASK (0xf << 16) - -#define CG_DISPLAY_GAP_CNTL 0x20a -# define DISP1_GAP(x) ((x) << 0) -# define DISP1_GAP_MASK (3 << 0) -# define DISP2_GAP(x) ((x) << 2) -# define DISP2_GAP_MASK (3 << 2) -# define VBI_TIMER_COUNT(x) ((x) << 4) -# define VBI_TIMER_COUNT_MASK (0x3fff << 4) -# define VBI_TIMER_UNIT(x) ((x) << 20) -# define VBI_TIMER_UNIT_MASK (7 << 20) -# define DISP1_GAP_MCHG(x) ((x) << 24) -# define DISP1_GAP_MCHG_MASK (3 << 24) -# define DISP2_GAP_MCHG(x) ((x) << 26) -# define DISP2_GAP_MCHG_MASK (3 << 26) - -#define CG_ULV_CONTROL 0x21e -#define CG_ULV_PARAMETER 0x21f - -#define SMC_SCRATCH0 0x221 - -#define CG_CAC_CTRL 0x22e -# define CAC_WINDOW(x) ((x) << 0) -# define CAC_WINDOW_MASK 0x00ffffff - -#define DMIF_ADDR_CONFIG 0x2F5 - -#define DMIF_ADDR_CALC 0x300 - -#define PIPE0_DMIF_BUFFER_CONTROL 0x0328 -# define DMIF_BUFFERS_ALLOCATED(x) ((x) << 0) -# define DMIF_BUFFERS_ALLOCATED_COMPLETED (1 << 4) - -#define SRBM_STATUS 0x394 -#define GRBM_RQ_PENDING (1 << 5) -#define VMC_BUSY (1 << 8) -#define MCB_BUSY (1 << 9) -#define MCB_NON_DISPLAY_BUSY (1 << 10) -#define MCC_BUSY (1 << 11) -#define MCD_BUSY (1 << 12) -#define SEM_BUSY (1 << 14) -#define IH_BUSY (1 << 17) - -#define SRBM_SOFT_RESET 0x398 -#define SOFT_RESET_BIF (1 << 1) -#define SOFT_RESET_DC (1 << 5) -#define SOFT_RESET_DMA1 (1 << 6) -#define SOFT_RESET_GRBM (1 << 8) -#define SOFT_RESET_HDP (1 << 9) -#define SOFT_RESET_IH (1 << 10) -#define SOFT_RESET_MC (1 << 11) -#define SOFT_RESET_ROM (1 << 14) -#define SOFT_RESET_SEM (1 << 15) -#define SOFT_RESET_VMC (1 << 17) -#define SOFT_RESET_DMA (1 << 20) -#define SOFT_RESET_TST (1 << 21) -#define SOFT_RESET_REGBB (1 << 22) -#define SOFT_RESET_ORB (1 << 23) - -#define CC_SYS_RB_BACKEND_DISABLE 0x3A0 -#define GC_USER_SYS_RB_BACKEND_DISABLE 0x3A1 - -#define SRBM_READ_ERROR 0x3A6 -#define SRBM_INT_CNTL 0x3A8 -#define SRBM_INT_ACK 0x3AA - -#define SRBM_STATUS2 0x3B1 -#define DMA_BUSY (1 << 5) -#define DMA1_BUSY (1 << 6) - -#define VM_L2_CNTL 0x500 -#define ENABLE_L2_CACHE (1 << 0) -#define ENABLE_L2_FRAGMENT_PROCESSING (1 << 1) -#define L2_CACHE_PTE_ENDIAN_SWAP_MODE(x) ((x) << 2) -#define L2_CACHE_PDE_ENDIAN_SWAP_MODE(x) ((x) << 4) -#define ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE (1 << 9) -#define ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE (1 << 10) -#define EFFECTIVE_L2_QUEUE_SIZE(x) (((x) & 7) << 15) -#define CONTEXT1_IDENTITY_ACCESS_MODE(x) (((x) & 3) << 19) -#define VM_L2_CNTL2 0x501 -#define INVALIDATE_ALL_L1_TLBS (1 << 0) -#define INVALIDATE_L2_CACHE (1 << 1) -#define INVALIDATE_CACHE_MODE(x) ((x) << 26) -#define INVALIDATE_PTE_AND_PDE_CACHES 0 -#define INVALIDATE_ONLY_PTE_CACHES 1 -#define INVALIDATE_ONLY_PDE_CACHES 2 -#define VM_L2_CNTL3 0x502 -#define BANK_SELECT(x) ((x) << 0) -#define L2_CACHE_UPDATE_MODE(x) ((x) << 6) -#define L2_CACHE_BIGK_FRAGMENT_SIZE(x) ((x) << 15) -#define L2_CACHE_BIGK_ASSOCIATIVITY (1 << 20) -#define VM_L2_STATUS 0x503 -#define L2_BUSY (1 << 0) -#define VM_CONTEXT0_CNTL 0x504 -#define ENABLE_CONTEXT (1 << 0) -#define PAGE_TABLE_DEPTH(x) (((x) & 3) << 1) -#define RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 3) -#define RANGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 4) -#define DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 6) -#define DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 7) -#define PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 9) -#define PDE0_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 10) -#define VALID_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 12) -#define VALID_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 13) -#define READ_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 15) -#define READ_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 16) -#define WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 18) -#define WRITE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 19) -#define PAGE_TABLE_BLOCK_SIZE(x) (((x) & 0xF) << 24) -#define VM_CONTEXT1_CNTL 0x505 -#define VM_CONTEXT0_CNTL2 0x50C -#define VM_CONTEXT1_CNTL2 0x50D -#define VM_CONTEXT8_PAGE_TABLE_BASE_ADDR 0x50E -#define VM_CONTEXT9_PAGE_TABLE_BASE_ADDR 0x50F -#define VM_CONTEXT10_PAGE_TABLE_BASE_ADDR 0x510 -#define VM_CONTEXT11_PAGE_TABLE_BASE_ADDR 0x511 -#define VM_CONTEXT12_PAGE_TABLE_BASE_ADDR 0x512 -#define VM_CONTEXT13_PAGE_TABLE_BASE_ADDR 0x513 -#define VM_CONTEXT14_PAGE_TABLE_BASE_ADDR 0x514 -#define VM_CONTEXT15_PAGE_TABLE_BASE_ADDR 0x515 - -#define VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x53f -#define VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x537 -#define PROTECTIONS_MASK (0xf << 0) -#define PROTECTIONS_SHIFT 0 - /* bit 0: range - * bit 1: pde0 - * bit 2: valid - * bit 3: read - * bit 4: write - */ -#define MEMORY_CLIENT_ID_MASK (0xff << 12) -#define MEMORY_CLIENT_ID_SHIFT 12 -#define MEMORY_CLIENT_RW_MASK (1 << 24) -#define MEMORY_CLIENT_RW_SHIFT 24 -#define FAULT_VMID_MASK (0xf << 25) -#define FAULT_VMID_SHIFT 25 - #define VM_INVALIDATE_REQUEST 0x51E #define VM_INVALIDATE_RESPONSE 0x51F -#define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR 0x546 -#define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR 0x547 - -#define VM_CONTEXT0_PAGE_TABLE_BASE_ADDR 0x54F -#define VM_CONTEXT1_PAGE_TABLE_BASE_ADDR 0x550 -#define VM_CONTEXT2_PAGE_TABLE_BASE_ADDR 0x551 -#define VM_CONTEXT3_PAGE_TABLE_BASE_ADDR 0x552 -#define VM_CONTEXT4_PAGE_TABLE_BASE_ADDR 0x553 -#define VM_CONTEXT5_PAGE_TABLE_BASE_ADDR 0x554 -#define VM_CONTEXT6_PAGE_TABLE_BASE_ADDR 0x555 -#define VM_CONTEXT7_PAGE_TABLE_BASE_ADDR 0x556 -#define VM_CONTEXT0_PAGE_TABLE_START_ADDR 0x557 -#define VM_CONTEXT1_PAGE_TABLE_START_ADDR 0x558 - -#define VM_CONTEXT0_PAGE_TABLE_END_ADDR 0x55F -#define VM_CONTEXT1_PAGE_TABLE_END_ADDR 0x560 - #define VM_L2_CG 0x570 #define MC_CG_ENABLE (1 << 18) #define MC_LS_ENABLE (1 << 19) -#define MC_SHARED_CHMAP 0x801 -#define NOOFCHAN_SHIFT 12 -#define NOOFCHAN_MASK 0x0000f000 -#define MC_SHARED_CHREMAP 0x802 - #define MC_VM_FB_LOCATION 0x809 #define MC_VM_AGP_TOP 0x80A #define MC_VM_AGP_BOT 0x80B @@ -491,21 +112,6 @@ #define MC_CITF_MISC_WR_CG 0x993 #define MC_CITF_MISC_VM_CG 0x994 -#define MC_ARB_RAMCFG 0x9D8 -#define NOOFBANK_SHIFT 0 -#define NOOFBANK_MASK 0x00000003 -#define NOOFRANK_SHIFT 2 -#define NOOFRANK_MASK 0x00000004 -#define NOOFROWS_SHIFT 3 -#define NOOFROWS_MASK 0x00000038 -#define NOOFCOLS_SHIFT 6 -#define NOOFCOLS_MASK 0x000000C0 -#define CHANSIZE_SHIFT 8 -#define CHANSIZE_MASK 0x00000100 -#define CHANSIZE_OVERRIDE (1 << 11) -#define NOOFGROUPS_SHIFT 12 -#define NOOFGROUPS_MASK 0x00001000 - #define MC_ARB_DRAM_TIMING 0x9DD #define MC_ARB_DRAM_TIMING2 0x9DE @@ -631,20 +237,6 @@ #define CLKS(x) ((x) << 0) #define CLKS_MASK (0xfff << 0) -#define HDP_HOST_PATH_CNTL 0xB00 -#define CLOCK_GATING_DIS (1 << 23) -#define HDP_NONSURFACE_BASE 0xB01 -#define HDP_NONSURFACE_INFO 0xB02 -#define HDP_NONSURFACE_SIZE 0xB03 - -#define HDP_DEBUG0 0xBCC - -#define HDP_ADDR_CONFIG 0xBD2 -#define HDP_MISC_CNTL 0xBD3 -#define HDP_FLUSH_INVALIDATE_CACHE (1 << 0) -#define HDP_MEM_POWER_LS 0xBD4 -#define HDP_LS_ENABLE (1 << 0) - #define ATC_MISC_CG 0xCD4 #define IH_RB_CNTL 0xF80 @@ -674,8 +266,6 @@ # define MC_WR_CLEAN_CNT(x) ((x) << 20) # define MC_VMID(x) ((x) << 25) -#define CONFIG_MEMSIZE 0x150A - #define INTERRUPT_CNTL 0x151A # define IH_DUMMY_RD_OVERRIDE (1 << 0) # define IH_DUMMY_RD_EN (1 << 1) @@ -683,486 +273,22 @@ # define GEN_IH_INT_EN (1 << 8) #define INTERRUPT_CNTL2 0x151B -#define HDP_MEM_COHERENCY_FLUSH_CNTL 0x1520 - -#define BIF_FB_EN 0x1524 -#define FB_READ_EN (1 << 0) -#define FB_WRITE_EN (1 << 1) - -#define HDP_REG_COHERENCY_FLUSH_CNTL 0x1528 - -/* DCE6 ELD audio interface */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0 0x28 /* LPCM */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR1 0x29 /* AC3 */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR2 0x2A /* MPEG1 */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR3 0x2B /* MP3 */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR4 0x2C /* MPEG2 */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR5 0x2D /* AAC */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR6 0x2E /* DTS */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR7 0x2F /* ATRAC */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR8 0x30 /* one bit audio - leave at 0 (default) */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR9 0x31 /* Dolby Digital */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR10 0x32 /* DTS-HD */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR11 0x33 /* MAT-MLP */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR12 0x34 /* DTS */ -#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13 0x35 /* WMA Pro */ -# define MAX_CHANNELS(x) (((x) & 0x7) << 0) -/* max channels minus one. 7 = 8 channels */ -# define SUPPORTED_FREQUENCIES(x) (((x) & 0xff) << 8) -# define DESCRIPTOR_BYTE_2(x) (((x) & 0xff) << 16) -# define SUPPORTED_FREQUENCIES_STEREO(x) (((x) & 0xff) << 24) /* LPCM only */ -/* SUPPORTED_FREQUENCIES, SUPPORTED_FREQUENCIES_STEREO - * bit0 = 32 kHz - * bit1 = 44.1 kHz - * bit2 = 48 kHz - * bit3 = 88.2 kHz - * bit4 = 96 kHz - * bit5 = 176.4 kHz - * bit6 = 192 kHz - */ - -#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC 0x37 -# define VIDEO_LIPSYNC(x) (((x) & 0xff) << 0) -# define AUDIO_LIPSYNC(x) (((x) & 0xff) << 8) -/* VIDEO_LIPSYNC, AUDIO_LIPSYNC - * 0 = invalid - * x = legal delay value - * 255 = sync not supported - */ -#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_HBR 0x38 -# define HBR_CAPABLE (1 << 0) /* enabled by default */ - -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO0 0x3a -# define MANUFACTURER_ID(x) (((x) & 0xffff) << 0) -# define PRODUCT_ID(x) (((x) & 0xffff) << 16) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO1 0x3b -# define SINK_DESCRIPTION_LEN(x) (((x) & 0xff) << 0) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO2 0x3c -# define PORT_ID0(x) (((x) & 0xffffffff) << 0) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO3 0x3d -# define PORT_ID1(x) (((x) & 0xffffffff) << 0) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO4 0x3e -# define DESCRIPTION0(x) (((x) & 0xff) << 0) -# define DESCRIPTION1(x) (((x) & 0xff) << 8) -# define DESCRIPTION2(x) (((x) & 0xff) << 16) -# define DESCRIPTION3(x) (((x) & 0xff) << 24) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO5 0x3f -# define DESCRIPTION4(x) (((x) & 0xff) << 0) -# define DESCRIPTION5(x) (((x) & 0xff) << 8) -# define DESCRIPTION6(x) (((x) & 0xff) << 16) -# define DESCRIPTION7(x) (((x) & 0xff) << 24) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO6 0x40 -# define DESCRIPTION8(x) (((x) & 0xff) << 0) -# define DESCRIPTION9(x) (((x) & 0xff) << 8) -# define DESCRIPTION10(x) (((x) & 0xff) << 16) -# define DESCRIPTION11(x) (((x) & 0xff) << 24) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO7 0x41 -# define DESCRIPTION12(x) (((x) & 0xff) << 0) -# define DESCRIPTION13(x) (((x) & 0xff) << 8) -# define DESCRIPTION14(x) (((x) & 0xff) << 16) -# define DESCRIPTION15(x) (((x) & 0xff) << 24) -#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO8 0x42 -# define DESCRIPTION16(x) (((x) & 0xff) << 0) -# define DESCRIPTION17(x) (((x) & 0xff) << 8) - -#define AZ_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL 0x54 -# define AUDIO_ENABLED (1 << 31) - -#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT 0x56 -#define PORT_CONNECTIVITY_MASK (3 << 30) -#define PORT_CONNECTIVITY_SHIFT 30 - -#define DC_LB_MEMORY_SPLIT 0x1AC3 -#define DC_LB_MEMORY_CONFIG(x) ((x) << 20) - -#define PRIORITY_A_CNT 0x1AC6 -#define PRIORITY_MARK_MASK 0x7fff -#define PRIORITY_OFF (1 << 16) -#define PRIORITY_ALWAYS_ON (1 << 20) -#define PRIORITY_B_CNT 0x1AC7 - -#define DPG_PIPE_ARBITRATION_CONTROL3 0x1B32 -# define LATENCY_WATERMARK_MASK(x) ((x) << 16) -#define DPG_PIPE_LATENCY_CONTROL 0x1B33 -# define LATENCY_LOW_WATERMARK(x) ((x) << 0) -# define LATENCY_HIGH_WATERMARK(x) ((x) << 16) - -/* 0x6bb8, 0x77b8, 0x103b8, 0x10fb8, 0x11bb8, 0x127b8 */ -#define VLINE_STATUS 0x1AEE -# define VLINE_OCCURRED (1 << 0) -# define VLINE_ACK (1 << 4) -# define VLINE_STAT (1 << 12) -# define VLINE_INTERRUPT (1 << 16) -# define VLINE_INTERRUPT_TYPE (1 << 17) -/* 0x6bbc, 0x77bc, 0x103bc, 0x10fbc, 0x11bbc, 0x127bc */ -#define VBLANK_STATUS 0x1AEF -# define VBLANK_OCCURRED (1 << 0) -# define VBLANK_ACK (1 << 4) -# define VBLANK_STAT (1 << 12) -# define VBLANK_INTERRUPT (1 << 16) -# define VBLANK_INTERRUPT_TYPE (1 << 17) - -/* 0x6b40, 0x7740, 0x10340, 0x10f40, 0x11b40, 0x12740 */ -#define INT_MASK 0x1AD0 -# define VBLANK_INT_MASK (1 << 0) -# define VLINE_INT_MASK (1 << 4) - -#define DISP_INTERRUPT_STATUS 0x183D -# define LB_D1_VLINE_INTERRUPT (1 << 2) -# define LB_D1_VBLANK_INTERRUPT (1 << 3) -# define DC_HPD1_INTERRUPT (1 << 17) -# define DC_HPD1_RX_INTERRUPT (1 << 18) -# define DACA_AUTODETECT_INTERRUPT (1 << 22) -# define DACB_AUTODETECT_INTERRUPT (1 << 23) -# define DC_I2C_SW_DONE_INTERRUPT (1 << 24) -# define DC_I2C_HW_DONE_INTERRUPT (1 << 25) -#define DISP_INTERRUPT_STATUS_CONTINUE 0x183E -# define LB_D2_VLINE_INTERRUPT (1 << 2) -# define LB_D2_VBLANK_INTERRUPT (1 << 3) -# define DC_HPD2_INTERRUPT (1 << 17) -# define DC_HPD2_RX_INTERRUPT (1 << 18) -# define DISP_TIMER_INTERRUPT (1 << 24) -#define DISP_INTERRUPT_STATUS_CONTINUE2 0x183F -# define LB_D3_VLINE_INTERRUPT (1 << 2) -# define LB_D3_VBLANK_INTERRUPT (1 << 3) -# define DC_HPD3_INTERRUPT (1 << 17) -# define DC_HPD3_RX_INTERRUPT (1 << 18) -#define DISP_INTERRUPT_STATUS_CONTINUE3 0x1840 -# define LB_D4_VLINE_INTERRUPT (1 << 2) -# define LB_D4_VBLANK_INTERRUPT (1 << 3) -# define DC_HPD4_INTERRUPT (1 << 17) -# define DC_HPD4_RX_INTERRUPT (1 << 18) -#define DISP_INTERRUPT_STATUS_CONTINUE4 0x1853 -# define LB_D5_VLINE_INTERRUPT (1 << 2) -# define LB_D5_VBLANK_INTERRUPT (1 << 3) -# define DC_HPD5_INTERRUPT (1 << 17) -# define DC_HPD5_RX_INTERRUPT (1 << 18) -#define DISP_INTERRUPT_STATUS_CONTINUE5 0x1854 -# define LB_D6_VLINE_INTERRUPT (1 << 2) -# define LB_D6_VBLANK_INTERRUPT (1 << 3) -# define DC_HPD6_INTERRUPT (1 << 17) -# define DC_HPD6_RX_INTERRUPT (1 << 18) - -/* 0x6858, 0x7458, 0x10058, 0x10c58, 0x11858, 0x12458 */ -#define GRPH_INT_STATUS 0x1A16 -# define GRPH_PFLIP_INT_OCCURRED (1 << 0) -# define GRPH_PFLIP_INT_CLEAR (1 << 8) -/* 0x685c, 0x745c, 0x1005c, 0x10c5c, 0x1185c, 0x1245c */ -#define GRPH_INT_CONTROL 0x1A17 -# define GRPH_PFLIP_INT_MASK (1 << 0) -# define GRPH_PFLIP_INT_TYPE (1 << 8) - -#define DAC_AUTODETECT_INT_CONTROL 0x19F2 - -#define DC_HPD1_INT_STATUS 0x1807 -#define DC_HPD2_INT_STATUS 0x180A -#define DC_HPD3_INT_STATUS 0x180D -#define DC_HPD4_INT_STATUS 0x1810 -#define DC_HPD5_INT_STATUS 0x1813 -#define DC_HPD6_INT_STATUS 0x1816 -# define DC_HPDx_INT_STATUS (1 << 0) -# define DC_HPDx_SENSE (1 << 1) -# define DC_HPDx_RX_INT_STATUS (1 << 8) - -#define DC_HPD1_INT_CONTROL 0x1808 -#define DC_HPD2_INT_CONTROL 0x180B -#define DC_HPD3_INT_CONTROL 0x180E -#define DC_HPD4_INT_CONTROL 0x1811 -#define DC_HPD5_INT_CONTROL 0x1814 -#define DC_HPD6_INT_CONTROL 0x1817 -# define DC_HPDx_INT_ACK (1 << 0) -# define DC_HPDx_INT_POLARITY (1 << 8) -# define DC_HPDx_INT_EN (1 << 16) -# define DC_HPDx_RX_INT_ACK (1 << 20) -# define DC_HPDx_RX_INT_EN (1 << 24) - -#define DC_HPD1_CONTROL 0x1809 -#define DC_HPD2_CONTROL 0x180C -#define DC_HPD3_CONTROL 0x180F -#define DC_HPD4_CONTROL 0x1812 -#define DC_HPD5_CONTROL 0x1815 -#define DC_HPD6_CONTROL 0x1818 -# define DC_HPDx_CONNECTION_TIMER(x) ((x) << 0) -# define DC_HPDx_RX_INT_TIMER(x) ((x) << 16) -# define DC_HPDx_EN (1 << 28) - -#define DPG_PIPE_STUTTER_CONTROL 0x1B35 -# define STUTTER_ENABLE (1 << 0) - -/* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */ -#define CRTC_STATUS_FRAME_COUNT 0x1BA6 - -/* Audio clocks */ -#define DCCG_AUDIO_DTO0_PHASE 0x05b0 -#define DCCG_AUDIO_DTO0_MODULE 0x05b4 -#define DCCG_AUDIO_DTO1_PHASE 0x05c0 -#define DCCG_AUDIO_DTO1_MODULE 0x05c4 - -#define GRBM_CNTL 0x2000 -#define GRBM_READ_TIMEOUT(x) ((x) << 0) - -#define GRBM_STATUS2 0x2002 -#define RLC_RQ_PENDING (1 << 0) -#define RLC_BUSY (1 << 8) -#define TC_BUSY (1 << 9) - -#define GRBM_STATUS 0x2004 -#define CMDFIFO_AVAIL_MASK 0x0000000F -#define RING2_RQ_PENDING (1 << 4) -#define SRBM_RQ_PENDING (1 << 5) -#define RING1_RQ_PENDING (1 << 6) -#define CF_RQ_PENDING (1 << 7) -#define PF_RQ_PENDING (1 << 8) -#define GDS_DMA_RQ_PENDING (1 << 9) -#define GRBM_EE_BUSY (1 << 10) -#define DB_CLEAN (1 << 12) -#define CB_CLEAN (1 << 13) -#define TA_BUSY (1 << 14) -#define GDS_BUSY (1 << 15) -#define VGT_BUSY (1 << 17) -#define IA_BUSY_NO_DMA (1 << 18) -#define IA_BUSY (1 << 19) -#define SX_BUSY (1 << 20) -#define SPI_BUSY (1 << 22) -#define BCI_BUSY (1 << 23) -#define SC_BUSY (1 << 24) -#define PA_BUSY (1 << 25) -#define DB_BUSY (1 << 26) -#define CP_COHERENCY_BUSY (1 << 28) -#define CP_BUSY (1 << 29) -#define CB_BUSY (1 << 30) -#define GUI_ACTIVE (1 << 31) -#define GRBM_STATUS_SE0 0x2005 -#define GRBM_STATUS_SE1 0x2006 -#define SE_DB_CLEAN (1 << 1) -#define SE_CB_CLEAN (1 << 2) -#define SE_BCI_BUSY (1 << 22) -#define SE_VGT_BUSY (1 << 23) -#define SE_PA_BUSY (1 << 24) -#define SE_TA_BUSY (1 << 25) -#define SE_SX_BUSY (1 << 26) -#define SE_SPI_BUSY (1 << 27) -#define SE_SC_BUSY (1 << 29) -#define SE_DB_BUSY (1 << 30) -#define SE_CB_BUSY (1 << 31) - -#define GRBM_INT_CNTL 0x2018 -# define RDERR_INT_ENABLE (1 << 0) -# define GUI_IDLE_INT_ENABLE (1 << 19) - -#define CP_STRMOUT_CNTL 0x213F -#define SCRATCH_REG0 0x2140 -#define SCRATCH_REG1 0x2141 -#define SCRATCH_REG2 0x2142 -#define SCRATCH_REG3 0x2143 -#define SCRATCH_REG4 0x2144 -#define SCRATCH_REG5 0x2145 -#define SCRATCH_REG6 0x2146 -#define SCRATCH_REG7 0x2147 - -#define SCRATCH_UMSK 0x2150 -#define SCRATCH_ADDR 0x2151 - -#define CP_SEM_WAIT_TIMER 0x216F - -#define CP_SEM_INCOMPLETE_TIMER_CNTL 0x2172 - -#define CP_ME_CNTL 0x21B6 -#define CP_CE_HALT (1 << 24) -#define CP_PFP_HALT (1 << 26) -#define CP_ME_HALT (1 << 28) - -#define CP_COHER_CNTL2 0x217A - -#define CP_RB2_RPTR 0x21BE -#define CP_RB1_RPTR 0x21BF -#define CP_RB0_RPTR 0x21C0 -#define CP_RB_WPTR_DELAY 0x21C1 - -#define CP_QUEUE_THRESHOLDS 0x21D8 -#define ROQ_IB1_START(x) ((x) << 0) -#define ROQ_IB2_START(x) ((x) << 8) -#define CP_MEQ_THRESHOLDS 0x21D9 -#define MEQ1_START(x) ((x) << 0) -#define MEQ2_START(x) ((x) << 8) - -#define CP_PERFMON_CNTL 0x21FF - #define VGT_VTX_VECT_EJECT_REG 0x222C - #define VGT_ESGS_RING_SIZE 0x2232 #define VGT_GSVS_RING_SIZE 0x2233 - #define VGT_GS_VERTEX_REUSE 0x2235 - #define VGT_PRIMITIVE_TYPE 0x2256 #define VGT_INDEX_TYPE 0x2257 - #define VGT_NUM_INDICES 0x225C #define VGT_NUM_INSTANCES 0x225D - #define VGT_TF_RING_SIZE 0x2262 - #define VGT_HS_OFFCHIP_PARAM 0x226C - #define VGT_TF_MEMORY_BASE 0x226E -#define PA_CL_ENHANCE 0x2285 -#define CLIP_VTX_REORDER_ENA (1 << 0) -#define NUM_CLIP_SEQ(x) ((x) << 1) - -#define PA_SU_LINE_STIPPLE_VALUE 0x2298 - -#define PA_SC_LINE_STIPPLE_STATE 0x22C4 - -#define PA_SC_FORCE_EOV_MAX_CNTS 0x22C9 -#define FORCE_EOV_MAX_CLK_CNT(x) ((x) << 0) -#define FORCE_EOV_MAX_REZ_CNT(x) ((x) << 16) - -#define PA_SC_FIFO_SIZE 0x22F3 -#define SC_FRONTEND_PRIM_FIFO_SIZE(x) ((x) << 0) -#define SC_BACKEND_PRIM_FIFO_SIZE(x) ((x) << 6) -#define SC_HIZ_TILE_FIFO_SIZE(x) ((x) << 15) -#define SC_EARLYZ_TILE_FIFO_SIZE(x) ((x) << 23) - #define PA_SC_ENHANCE 0x22FC -#define SQ_CONFIG 0x2300 - -#define SQC_CACHES 0x2302 - -#define SQ_POWER_THROTTLE 0x2396 -#define MIN_POWER(x) ((x) << 0) -#define MIN_POWER_MASK (0x3fff << 0) -#define MIN_POWER_SHIFT 0 -#define MAX_POWER(x) ((x) << 16) -#define MAX_POWER_MASK (0x3fff << 16) -#define MAX_POWER_SHIFT 0 -#define SQ_POWER_THROTTLE2 0x2397 -#define MAX_POWER_DELTA(x) ((x) << 0) -#define MAX_POWER_DELTA_MASK (0x3fff << 0) -#define MAX_POWER_DELTA_SHIFT 0 -#define STI_SIZE(x) ((x) << 16) -#define STI_SIZE_MASK (0x3ff << 16) -#define STI_SIZE_SHIFT 16 -#define LTI_RATIO(x) ((x) << 27) -#define LTI_RATIO_MASK (0xf << 27) -#define LTI_RATIO_SHIFT 27 - -#define SX_DEBUG_1 0x2418 - -#define SPI_STATIC_THREAD_MGMT_1 0x2438 -#define SPI_STATIC_THREAD_MGMT_2 0x2439 -#define SPI_STATIC_THREAD_MGMT_3 0x243A -#define SPI_PS_MAX_WAVE_ID 0x243B - -#define SPI_CONFIG_CNTL 0x2440 - -#define SPI_CONFIG_CNTL_1 0x244F -#define VTX_DONE_DELAY(x) ((x) << 0) -#define INTERP_ONE_PRIM_PER_ROW (1 << 4) - -#define CGTS_TCC_DISABLE 0x2452 -#define CGTS_USER_TCC_DISABLE 0x2453 -#define TCC_DISABLE_MASK 0xFFFF0000 -#define TCC_DISABLE_SHIFT 16 -#define CGTS_SM_CTRL_REG 0x2454 -#define OVERRIDE (1 << 21) -#define LS_OVERRIDE (1 << 22) - -#define SPI_LB_CU_MASK 0x24D5 - #define TA_CNTL_AUX 0x2542 -#define CC_RB_BACKEND_DISABLE 0x263D -#define BACKEND_DISABLE(x) ((x) << 16) -#define GB_ADDR_CONFIG 0x263E -#define NUM_PIPES(x) ((x) << 0) -#define NUM_PIPES_MASK 0x00000007 -#define NUM_PIPES_SHIFT 0 -#define PIPE_INTERLEAVE_SIZE(x) ((x) << 4) -#define PIPE_INTERLEAVE_SIZE_MASK 0x00000070 -#define PIPE_INTERLEAVE_SIZE_SHIFT 4 -#define NUM_SHADER_ENGINES(x) ((x) << 12) -#define NUM_SHADER_ENGINES_MASK 0x00003000 -#define NUM_SHADER_ENGINES_SHIFT 12 -#define SHADER_ENGINE_TILE_SIZE(x) ((x) << 16) -#define SHADER_ENGINE_TILE_SIZE_MASK 0x00070000 -#define SHADER_ENGINE_TILE_SIZE_SHIFT 16 -#define NUM_GPUS(x) ((x) << 20) -#define NUM_GPUS_MASK 0x00700000 -#define NUM_GPUS_SHIFT 20 -#define MULTI_GPU_TILE_SIZE(x) ((x) << 24) -#define MULTI_GPU_TILE_SIZE_MASK 0x03000000 -#define MULTI_GPU_TILE_SIZE_SHIFT 24 -#define ROW_SIZE(x) ((x) << 28) -#define ROW_SIZE_MASK 0x30000000 -#define ROW_SIZE_SHIFT 28 - -#define CB_PERFCOUNTER0_SELECT0 0x2688 -#define CB_PERFCOUNTER0_SELECT1 0x2689 -#define CB_PERFCOUNTER1_SELECT0 0x268A -#define CB_PERFCOUNTER1_SELECT1 0x268B -#define CB_PERFCOUNTER2_SELECT0 0x268C -#define CB_PERFCOUNTER2_SELECT1 0x268D -#define CB_PERFCOUNTER3_SELECT0 0x268E -#define CB_PERFCOUNTER3_SELECT1 0x268F - -#define CB_CGTT_SCLK_CTRL 0x2698 - -#define TCP_CHAN_STEER_LO 0x2B03 -#define TCP_CHAN_STEER_HI 0x2B94 - -#define CP_RB0_BASE 0x3040 -#define CP_RB0_CNTL 0x3041 -#define RB_BUFSZ(x) ((x) << 0) -#define RB_BLKSZ(x) ((x) << 8) -#define BUF_SWAP_32BIT (2 << 16) -#define RB_NO_UPDATE (1 << 27) -#define RB_RPTR_WR_ENA (1 << 31) - -#define CP_RB0_RPTR_ADDR 0x3043 -#define CP_RB0_RPTR_ADDR_HI 0x3044 -#define CP_RB0_WPTR 0x3045 - -#define CP_PFP_UCODE_ADDR 0x3054 -#define CP_PFP_UCODE_DATA 0x3055 -#define CP_ME_RAM_RADDR 0x3056 -#define CP_ME_RAM_WADDR 0x3057 -#define CP_ME_RAM_DATA 0x3058 - -#define CP_CE_UCODE_ADDR 0x305A -#define CP_CE_UCODE_DATA 0x305B - -#define CP_RB1_BASE 0x3060 -#define CP_RB1_CNTL 0x3061 -#define CP_RB1_RPTR_ADDR 0x3062 -#define CP_RB1_RPTR_ADDR_HI 0x3063 -#define CP_RB1_WPTR 0x3064 -#define CP_RB2_BASE 0x3065 -#define CP_RB2_CNTL 0x3066 -#define CP_RB2_RPTR_ADDR 0x3067 -#define CP_RB2_RPTR_ADDR_HI 0x3068 -#define CP_RB2_WPTR 0x3069 -#define CP_INT_CNTL_RING0 0x306A -#define CP_INT_CNTL_RING1 0x306B -#define CP_INT_CNTL_RING2 0x306C -# define CNTX_BUSY_INT_ENABLE (1 << 19) -# define CNTX_EMPTY_INT_ENABLE (1 << 20) -# define WAIT_MEM_SEM_INT_ENABLE (1 << 21) -# define TIME_STAMP_INT_ENABLE (1 << 26) -# define CP_RINGID2_INT_ENABLE (1 << 29) -# define CP_RINGID1_INT_ENABLE (1 << 30) -# define CP_RINGID0_INT_ENABLE (1 << 31) -#define CP_INT_STATUS_RING0 0x306D -#define CP_INT_STATUS_RING1 0x306E -#define CP_INT_STATUS_RING2 0x306F -# define WAIT_MEM_SEM_INT_STAT (1 << 21) -# define TIME_STAMP_INT_STAT (1 << 26) -# define CP_RINGID2_INT_STAT (1 << 29) -# define CP_RINGID1_INT_STAT (1 << 30) -# define CP_RINGID0_INT_STAT (1 << 31) - // #define PA_SC_RASTER_CONFIG 0xA0D4 # define RB_XSEL2(x) ((x) << 4) # define RB_XSEL2_MASK (0x3 << 4) @@ -1185,171 +311,14 @@ # define SE_YSEL(x) ((x) << 28) # define SE_YSEL_MASK (0x3 << 28) -/* PIF PHY0 registers idx/data 0x8/0xc */ -#define PB0_PIF_CNTL 0x10 -# define LS2_EXIT_TIME(x) ((x) << 17) -# define LS2_EXIT_TIME_MASK (0x7 << 17) -# define LS2_EXIT_TIME_SHIFT 17 -#define PB0_PIF_PAIRING 0x11 -# define MULTI_PIF (1 << 25) -#define PB0_PIF_PWRDOWN_0 0x12 -# define PLL_POWER_STATE_IN_TXS2_0(x) ((x) << 7) -# define PLL_POWER_STATE_IN_TXS2_0_MASK (0x7 << 7) -# define PLL_POWER_STATE_IN_TXS2_0_SHIFT 7 -# define PLL_POWER_STATE_IN_OFF_0(x) ((x) << 10) -# define PLL_POWER_STATE_IN_OFF_0_MASK (0x7 << 10) -# define PLL_POWER_STATE_IN_OFF_0_SHIFT 10 -# define PLL_RAMP_UP_TIME_0(x) ((x) << 24) -# define PLL_RAMP_UP_TIME_0_MASK (0x7 << 24) -# define PLL_RAMP_UP_TIME_0_SHIFT 24 -#define PB0_PIF_PWRDOWN_1 0x13 -# define PLL_POWER_STATE_IN_TXS2_1(x) ((x) << 7) -# define PLL_POWER_STATE_IN_TXS2_1_MASK (0x7 << 7) -# define PLL_POWER_STATE_IN_TXS2_1_SHIFT 7 -# define PLL_POWER_STATE_IN_OFF_1(x) ((x) << 10) -# define PLL_POWER_STATE_IN_OFF_1_MASK (0x7 << 10) -# define PLL_POWER_STATE_IN_OFF_1_SHIFT 10 -# define PLL_RAMP_UP_TIME_1(x) ((x) << 24) -# define PLL_RAMP_UP_TIME_1_MASK (0x7 << 24) -# define PLL_RAMP_UP_TIME_1_SHIFT 24 - -#define PB0_PIF_PWRDOWN_2 0x17 -# define PLL_POWER_STATE_IN_TXS2_2(x) ((x) << 7) -# define PLL_POWER_STATE_IN_TXS2_2_MASK (0x7 << 7) -# define PLL_POWER_STATE_IN_TXS2_2_SHIFT 7 -# define PLL_POWER_STATE_IN_OFF_2(x) ((x) << 10) -# define PLL_POWER_STATE_IN_OFF_2_MASK (0x7 << 10) -# define PLL_POWER_STATE_IN_OFF_2_SHIFT 10 -# define PLL_RAMP_UP_TIME_2(x) ((x) << 24) -# define PLL_RAMP_UP_TIME_2_MASK (0x7 << 24) -# define PLL_RAMP_UP_TIME_2_SHIFT 24 -#define PB0_PIF_PWRDOWN_3 0x18 -# define PLL_POWER_STATE_IN_TXS2_3(x) ((x) << 7) -# define PLL_POWER_STATE_IN_TXS2_3_MASK (0x7 << 7) -# define PLL_POWER_STATE_IN_TXS2_3_SHIFT 7 -# define PLL_POWER_STATE_IN_OFF_3(x) ((x) << 10) -# define PLL_POWER_STATE_IN_OFF_3_MASK (0x7 << 10) -# define PLL_POWER_STATE_IN_OFF_3_SHIFT 10 -# define PLL_RAMP_UP_TIME_3(x) ((x) << 24) -# define PLL_RAMP_UP_TIME_3_MASK (0x7 << 24) -# define PLL_RAMP_UP_TIME_3_SHIFT 24 -/* PIF PHY1 registers idx/data 0x10/0x14 */ -#define PB1_PIF_CNTL 0x10 -#define PB1_PIF_PAIRING 0x11 -#define PB1_PIF_PWRDOWN_0 0x12 -#define PB1_PIF_PWRDOWN_1 0x13 - -#define PB1_PIF_PWRDOWN_2 0x17 -#define PB1_PIF_PWRDOWN_3 0x18 -/* PCIE registers idx/data 0x30/0x34 */ -#define PCIE_CNTL2 0x1c /* PCIE */ -# define SLV_MEM_LS_EN (1 << 16) -# define SLV_MEM_AGGRESSIVE_LS_EN (1 << 17) -# define MST_MEM_LS_EN (1 << 18) -# define REPLAY_MEM_LS_EN (1 << 19) -#define PCIE_LC_STATUS1 0x28 /* PCIE */ -# define LC_REVERSE_RCVR (1 << 0) -# define LC_REVERSE_XMIT (1 << 1) -# define LC_OPERATING_LINK_WIDTH_MASK (0x7 << 2) -# define LC_OPERATING_LINK_WIDTH_SHIFT 2 -# define LC_DETECTED_LINK_WIDTH_MASK (0x7 << 5) -# define LC_DETECTED_LINK_WIDTH_SHIFT 5 - -#define PCIE_P_CNTL 0x40 /* PCIE */ -# define P_IGNORE_EDB_ERR (1 << 6) - /* PCIE PORT registers idx/data 0x38/0x3c */ -#define PCIE_LC_CNTL 0xa0 -# define LC_L0S_INACTIVITY(x) ((x) << 8) -# define LC_L0S_INACTIVITY_MASK (0xf << 8) -# define LC_L0S_INACTIVITY_SHIFT 8 -# define LC_L1_INACTIVITY(x) ((x) << 12) -# define LC_L1_INACTIVITY_MASK (0xf << 12) -# define LC_L1_INACTIVITY_SHIFT 12 -# define LC_PMI_TO_L1_DIS (1 << 16) -# define LC_ASPM_TO_L1_DIS (1 << 24) -#define PCIE_LC_LINK_WIDTH_CNTL 0xa2 /* PCIE_P */ -# define LC_LINK_WIDTH_SHIFT 0 -# define LC_LINK_WIDTH_MASK 0x7 +// #define PCIE_LC_LINK_WIDTH_CNTL 0xa2 /* PCIE_P */ # define LC_LINK_WIDTH_X0 0 # define LC_LINK_WIDTH_X1 1 # define LC_LINK_WIDTH_X2 2 # define LC_LINK_WIDTH_X4 3 # define LC_LINK_WIDTH_X8 4 # define LC_LINK_WIDTH_X16 6 -# define LC_LINK_WIDTH_RD_SHIFT 4 -# define LC_LINK_WIDTH_RD_MASK 0x70 -# define LC_RECONFIG_ARC_MISSING_ESCAPE (1 << 7) -# define LC_RECONFIG_NOW (1 << 8) -# define LC_RENEGOTIATION_SUPPORT (1 << 9) -# define LC_RENEGOTIATE_EN (1 << 10) -# define LC_SHORT_RECONFIG_EN (1 << 11) -# define LC_UPCONFIGURE_SUPPORT (1 << 12) -# define LC_UPCONFIGURE_DIS (1 << 13) -# define LC_DYN_LANES_PWR_STATE(x) ((x) << 21) -# define LC_DYN_LANES_PWR_STATE_MASK (0x3 << 21) -# define LC_DYN_LANES_PWR_STATE_SHIFT 21 -#define PCIE_LC_N_FTS_CNTL 0xa3 /* PCIE_P */ -# define LC_XMIT_N_FTS(x) ((x) << 0) -# define LC_XMIT_N_FTS_MASK (0xff << 0) -# define LC_XMIT_N_FTS_SHIFT 0 -# define LC_XMIT_N_FTS_OVERRIDE_EN (1 << 8) -# define LC_N_FTS_MASK (0xff << 24) -#define PCIE_LC_SPEED_CNTL 0xa4 /* PCIE_P */ -# define LC_GEN2_EN_STRAP (1 << 0) -# define LC_GEN3_EN_STRAP (1 << 1) -# define LC_TARGET_LINK_SPEED_OVERRIDE_EN (1 << 2) -# define LC_TARGET_LINK_SPEED_OVERRIDE_MASK (0x3 << 3) -# define LC_TARGET_LINK_SPEED_OVERRIDE_SHIFT 3 -# define LC_FORCE_EN_SW_SPEED_CHANGE (1 << 5) -# define LC_FORCE_DIS_SW_SPEED_CHANGE (1 << 6) -# define LC_FORCE_EN_HW_SPEED_CHANGE (1 << 7) -# define LC_FORCE_DIS_HW_SPEED_CHANGE (1 << 8) -# define LC_INITIATE_LINK_SPEED_CHANGE (1 << 9) -# define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK (0x3 << 10) -# define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT 10 -# define LC_CURRENT_DATA_RATE_MASK (0x3 << 13) /* 0/1/2 = gen1/2/3 */ -# define LC_CURRENT_DATA_RATE_SHIFT 13 -# define LC_CLR_FAILED_SPD_CHANGE_CNT (1 << 16) -# define LC_OTHER_SIDE_EVER_SENT_GEN2 (1 << 18) -# define LC_OTHER_SIDE_SUPPORTS_GEN2 (1 << 19) -# define LC_OTHER_SIDE_EVER_SENT_GEN3 (1 << 20) -# define LC_OTHER_SIDE_SUPPORTS_GEN3 (1 << 21) - -#define PCIE_LC_CNTL2 0xb1 -# define LC_ALLOW_PDWN_IN_L1 (1 << 17) -# define LC_ALLOW_PDWN_IN_L23 (1 << 18) - -#define PCIE_LC_CNTL3 0xb5 /* PCIE_P */ -# define LC_GO_TO_RECOVERY (1 << 30) -#define PCIE_LC_CNTL4 0xb6 /* PCIE_P */ -# define LC_REDO_EQ (1 << 5) -# define LC_SET_QUIESCE (1 << 13) - -/* - * UVD - */ -#define UVD_UDEC_ADDR_CONFIG 0x3bd3 -#define UVD_UDEC_DB_ADDR_CONFIG 0x3bd4 -#define UVD_UDEC_DBW_ADDR_CONFIG 0x3bd5 -#define UVD_RBC_RB_RPTR 0x3da4 -#define UVD_RBC_RB_WPTR 0x3da5 -#define UVD_STATUS 0x3daf - -#define UVD_CGC_CTRL 0x3dc2 -# define DCM (1 << 0) -# define CG_DT(x) ((x) << 2) -# define CG_DT_MASK (0xf << 2) -# define CLK_OD(x) ((x) << 6) -# define CLK_OD_MASK (0x1f << 6) - - /* UVD CTX indirect */ -#define UVD_CGC_MEM_CTRL 0xC0 -#define UVD_CGC_CTRL2 0xC1 -# define DYN_OR_EN (1 << 0) -# define DYN_RR_EN (1 << 1) -# define G_DIV_ID(x) ((x) << 2) -# define G_DIV_ID_MASK (0x7 << 2) /* * PM4 @@ -1583,45 +552,7 @@ /* ASYNC DMA - first instance at 0xd000, second at 0xd800 */ #define DMA0_REGISTER_OFFSET 0x0 /* not a register */ #define DMA1_REGISTER_OFFSET 0x200 /* not a register */ - -#define DMA_RB_CNTL 0x3400 -# define DMA_RB_ENABLE (1 << 0) -# define DMA_RB_SIZE(x) ((x) << 1) /* log2 */ -# define DMA_RB_SWAP_ENABLE (1 << 9) /* 8IN32 */ -# define DMA_RPTR_WRITEBACK_ENABLE (1 << 12) -# define DMA_RPTR_WRITEBACK_SWAP_ENABLE (1 << 13) /* 8IN32 */ -# define DMA_RPTR_WRITEBACK_TIMER(x) ((x) << 16) /* log2 */ -#define DMA_RB_BASE 0x3401 -#define DMA_RB_RPTR 0x3402 -#define DMA_RB_WPTR 0x3403 - -#define DMA_RB_RPTR_ADDR_HI 0x3407 -#define DMA_RB_RPTR_ADDR_LO 0x3408 - -#define DMA_IB_CNTL 0x3409 -# define DMA_IB_ENABLE (1 << 0) -# define DMA_IB_SWAP_ENABLE (1 << 4) -# define CMD_VMID_FORCE (1 << 31) -#define DMA_IB_RPTR 0x340a -#define DMA_CNTL 0x340b -# define TRAP_ENABLE (1 << 0) -# define SEM_INCOMPLETE_INT_ENABLE (1 << 1) -# define SEM_WAIT_INT_ENABLE (1 << 2) -# define DATA_SWAP_ENABLE (1 << 3) -# define FENCE_SWAP_ENABLE (1 << 4) -# define CTXEMPTY_INT_ENABLE (1 << 28) -#define DMA_STATUS_REG 0x340d -# define DMA_IDLE (1 << 0) -#define DMA_TILING_CONFIG 0x342e - -#define DMA_POWER_CNTL 0x342f -# define MEM_POWER_OVERRIDE (1 << 8) -#define DMA_CLK_CTRL 0x3430 - -#define DMA_PG 0x3435 -# define PG_CNTL_ENABLE (1 << 0) -#define DMA_PGFSM_CONFIG 0x3436 -#define DMA_PGFSM_WRITE 0x3437 +#define SDMA_MAX_INSTANCE 2 #define DMA_PACKET(cmd, b, t, s, n) ((((cmd) & 0xF) << 28) | \ (((b) & 0x1) << 26) | \ @@ -1650,6 +581,7 @@ #define DMA_PACKET_POLL_REG_MEM 0xe #define DMA_PACKET_NOP 0xf +/* VCE */ #define VCE_STATUS 0x20004 #define VCE_VCPU_CNTL 0x20014 #define VCE_CLK_EN (1 << 0) @@ -1726,378 +658,118 @@ #define CURSOR_WIDTH 64 #define CURSOR_HEIGHT 64 -#define AMDGPU_MM_INDEX 0x0000 -#define AMDGPU_MM_DATA 0x0001 - -#define VERDE_NUM_CRTC 6 -#define BLACKOUT_MODE_MASK 0x00000007 -#define VGA_RENDER_CONTROL 0xC0 -#define R_000300_VGA_RENDER_CONTROL 0xC0 -#define C_000300_VGA_VSTATUS_CNTL 0xFFFCFFFF -#define EVERGREEN_CRTC_STATUS 0x1BA3 -#define EVERGREEN_CRTC_V_BLANK (1 << 0) -#define EVERGREEN_CRTC_STATUS_POSITION 0x1BA4 -/* CRTC blocks at 0x6df0, 0x79f0, 0x105f0, 0x111f0, 0x11df0, 0x129f0 */ -#define EVERGREEN_CRTC_V_BLANK_START_END 0x1b8d -#define EVERGREEN_CRTC_CONTROL 0x1b9c -#define EVERGREEN_CRTC_MASTER_EN (1 << 0) -#define EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE (1 << 24) -#define EVERGREEN_CRTC_BLANK_CONTROL 0x1b9d -#define EVERGREEN_CRTC_BLANK_DATA_EN (1 << 8) -#define EVERGREEN_CRTC_V_BLANK (1 << 0) -#define EVERGREEN_CRTC_STATUS_HV_COUNT 0x1ba8 -#define EVERGREEN_CRTC_UPDATE_LOCK 0x1bb5 -#define EVERGREEN_MASTER_UPDATE_LOCK 0x1bbd -#define EVERGREEN_MASTER_UPDATE_MODE 0x1bbe -#define EVERGREEN_GRPH_UPDATE_LOCK (1 << 16) -#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x1a07 -#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH 0x1a08 -#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS 0x1a04 -#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS 0x1a05 -#define EVERGREEN_GRPH_UPDATE 0x1a11 -#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS 0xc4 -#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH 0xc9 -#define EVERGREEN_GRPH_SURFACE_UPDATE_PENDING (1 << 2) - -#define EVERGREEN_DATA_FORMAT 0x1ac0 -# define EVERGREEN_INTERLEAVE_EN (1 << 0) - -#define R600_D1GRPH_ARRAY_MODE_LINEAR_GENERAL (0 << 20) -#define R600_D1GRPH_ARRAY_MODE_LINEAR_ALIGNED (1 << 20) -#define R600_D1GRPH_ARRAY_MODE_1D_TILED_THIN1 (2 << 20) -#define R600_D1GRPH_ARRAY_MODE_2D_TILED_THIN1 (4 << 20) - -#define R700_D1GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x1a45 -#define R700_D2GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x1845 - -#define R700_D2GRPH_SECONDARY_SURFACE_ADDRESS_HIGH 0x1847 -#define R700_D1GRPH_SECONDARY_SURFACE_ADDRESS_HIGH 0x1a47 - -#define R600_D1GRPH_SWAP_CONTROL 0x1843 -#define R600_D1GRPH_SWAP_ENDIAN_NONE (0 << 0) -#define R600_D1GRPH_SWAP_ENDIAN_16BIT (1 << 0) -#define R600_D1GRPH_SWAP_ENDIAN_32BIT (2 << 0) -#define R600_D1GRPH_SWAP_ENDIAN_64BIT (3 << 0) - -#define AVIVO_D1VGA_CONTROL 0x00cc -# define AVIVO_DVGA_CONTROL_MODE_ENABLE (1 << 0) -# define AVIVO_DVGA_CONTROL_TIMING_SELECT (1 << 8) -# define AVIVO_DVGA_CONTROL_SYNC_POLARITY_SELECT (1 << 9) -# define AVIVO_DVGA_CONTROL_OVERSCAN_TIMING_SELECT (1 << 10) -# define AVIVO_DVGA_CONTROL_OVERSCAN_COLOR_EN (1 << 16) -# define AVIVO_DVGA_CONTROL_ROTATE (1 << 24) -#define AVIVO_D2VGA_CONTROL 0x00ce - -#define R600_BUS_CNTL 0x1508 -# define R600_BIOS_ROM_DIS (1 << 1) + #define R600_ROM_CNTL 0x580 # define R600_SCK_OVERWRITE (1 << 1) # define R600_SCK_PRESCALE_CRYSTAL_CLK_SHIFT 28 # define R600_SCK_PRESCALE_CRYSTAL_CLK_MASK (0xf << 28) -#define FMT_BIT_DEPTH_CONTROL 0x1bf2 -#define FMT_TRUNCATE_EN (1 << 0) -#define FMT_TRUNCATE_DEPTH (1 << 4) -#define FMT_SPATIAL_DITHER_EN (1 << 8) -#define FMT_SPATIAL_DITHER_MODE(x) ((x) << 9) -#define FMT_SPATIAL_DITHER_DEPTH (1 << 12) -#define FMT_FRAME_RANDOM_ENABLE (1 << 13) -#define FMT_RGB_RANDOM_ENABLE (1 << 14) -#define FMT_HIGHPASS_RANDOM_ENABLE (1 << 15) -#define FMT_TEMPORAL_DITHER_EN (1 << 16) -#define FMT_TEMPORAL_DITHER_DEPTH (1 << 20) -#define FMT_TEMPORAL_DITHER_OFFSET(x) ((x) << 21) -#define FMT_TEMPORAL_LEVEL (1 << 24) -#define FMT_TEMPORAL_DITHER_RESET (1 << 25) -#define FMT_25FRC_SEL(x) ((x) << 26) -#define FMT_50FRC_SEL(x) ((x) << 28) -#define FMT_75FRC_SEL(x) ((x) << 30) - -#define EVERGREEN_DC_LUT_CONTROL 0x1a80 -#define EVERGREEN_DC_LUT_BLACK_OFFSET_BLUE 0x1a81 -#define EVERGREEN_DC_LUT_BLACK_OFFSET_GREEN 0x1a82 -#define EVERGREEN_DC_LUT_BLACK_OFFSET_RED 0x1a83 -#define EVERGREEN_DC_LUT_WHITE_OFFSET_BLUE 0x1a84 -#define EVERGREEN_DC_LUT_WHITE_OFFSET_GREEN 0x1a85 -#define EVERGREEN_DC_LUT_WHITE_OFFSET_RED 0x1a86 -#define EVERGREEN_DC_LUT_30_COLOR 0x1a7c -#define EVERGREEN_DC_LUT_RW_INDEX 0x1a79 -#define EVERGREEN_DC_LUT_WRITE_EN_MASK 0x1a7e -#define EVERGREEN_DC_LUT_RW_MODE 0x1a78 - -#define EVERGREEN_GRPH_ENABLE 0x1a00 -#define EVERGREEN_GRPH_CONTROL 0x1a01 -#define EVERGREEN_GRPH_DEPTH(x) (((x) & 0x3) << 0) -#define EVERGREEN_GRPH_DEPTH_8BPP 0 -#define EVERGREEN_GRPH_DEPTH_16BPP 1 -#define EVERGREEN_GRPH_DEPTH_32BPP 2 -#define EVERGREEN_GRPH_NUM_BANKS(x) (((x) & 0x3) << 2) -#define EVERGREEN_ADDR_SURF_2_BANK 0 -#define EVERGREEN_ADDR_SURF_4_BANK 1 -#define EVERGREEN_ADDR_SURF_8_BANK 2 -#define EVERGREEN_ADDR_SURF_16_BANK 3 -#define EVERGREEN_GRPH_Z(x) (((x) & 0x3) << 4) -#define EVERGREEN_GRPH_BANK_WIDTH(x) (((x) & 0x3) << 6) -#define EVERGREEN_ADDR_SURF_BANK_WIDTH_1 0 -#define EVERGREEN_ADDR_SURF_BANK_WIDTH_2 1 -#define EVERGREEN_ADDR_SURF_BANK_WIDTH_4 2 -#define EVERGREEN_ADDR_SURF_BANK_WIDTH_8 3 -#define EVERGREEN_GRPH_FORMAT(x) (((x) & 0x7) << 8) - -#define EVERGREEN_GRPH_FORMAT_INDEXED 0 -#define EVERGREEN_GRPH_FORMAT_ARGB1555 0 -#define EVERGREEN_GRPH_FORMAT_ARGB565 1 -#define EVERGREEN_GRPH_FORMAT_ARGB4444 2 -#define EVERGREEN_GRPH_FORMAT_AI88 3 -#define EVERGREEN_GRPH_FORMAT_MONO16 4 -#define EVERGREEN_GRPH_FORMAT_BGRA5551 5 +#define GRPH_ARRAY_LINEAR_GENERAL 0 +#define GRPH_ARRAY_LINEAR_ALIGNED 1 +#define GRPH_ARRAY_1D_TILED_THIN1 2 +#define GRPH_ARRAY_2D_TILED_THIN1 4 + +#define ES_AND_GS_AUTO 3 +#define BUF_SWAP_32BIT (2 << 16) + +#define GRPH_DEPTH_8BPP 0 +#define GRPH_DEPTH_16BPP 1 +#define GRPH_DEPTH_32BPP 2 + +/* 8 BPP */ +#define GRPH_FORMAT_INDEXED 0 + +/* 16 BPP */ +#define GRPH_FORMAT_ARGB1555 0 +#define GRPH_FORMAT_ARGB565 1 +#define GRPH_FORMAT_ARGB4444 2 +#define GRPH_FORMAT_AI88 3 +#define GRPH_FORMAT_MONO16 4 +#define GRPH_FORMAT_BGRA5551 5 /* 32 BPP */ -#define EVERGREEN_GRPH_FORMAT_ARGB8888 0 -#define EVERGREEN_GRPH_FORMAT_ARGB2101010 1 -#define EVERGREEN_GRPH_FORMAT_32BPP_DIG 2 -#define EVERGREEN_GRPH_FORMAT_8B_ARGB2101010 3 -#define EVERGREEN_GRPH_FORMAT_BGRA1010102 4 -#define EVERGREEN_GRPH_FORMAT_8B_BGRA1010102 5 -#define EVERGREEN_GRPH_FORMAT_RGB111110 6 -#define EVERGREEN_GRPH_FORMAT_BGR101111 7 -#define EVERGREEN_GRPH_BANK_HEIGHT(x) (((x) & 0x3) << 11) -#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_1 0 -#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_2 1 -#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_4 2 -#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_8 3 -#define EVERGREEN_GRPH_TILE_SPLIT(x) (((x) & 0x7) << 13) -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_64B 0 -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_128B 1 -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_256B 2 -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_512B 3 -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_1KB 4 -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_2KB 5 -#define EVERGREEN_ADDR_SURF_TILE_SPLIT_4KB 6 -#define EVERGREEN_GRPH_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 18) -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1 0 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2 1 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4 2 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8 3 -#define EVERGREEN_GRPH_ARRAY_MODE(x) (((x) & 0x7) << 20) -#define EVERGREEN_GRPH_ARRAY_LINEAR_GENERAL 0 -#define EVERGREEN_GRPH_ARRAY_LINEAR_ALIGNED 1 -#define EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1 2 -#define EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1 4 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1 0 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2 1 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4 2 -#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8 3 - -#define EVERGREEN_GRPH_SWAP_CONTROL 0x1a03 -#define EVERGREEN_GRPH_ENDIAN_SWAP(x) (((x) & 0x3) << 0) -# define EVERGREEN_GRPH_ENDIAN_NONE 0 -# define EVERGREEN_GRPH_ENDIAN_8IN16 1 -# define EVERGREEN_GRPH_ENDIAN_8IN32 2 -# define EVERGREEN_GRPH_ENDIAN_8IN64 3 -#define EVERGREEN_GRPH_RED_CROSSBAR(x) (((x) & 0x3) << 4) -# define EVERGREEN_GRPH_RED_SEL_R 0 -# define EVERGREEN_GRPH_RED_SEL_G 1 -# define EVERGREEN_GRPH_RED_SEL_B 2 -# define EVERGREEN_GRPH_RED_SEL_A 3 -#define EVERGREEN_GRPH_GREEN_CROSSBAR(x) (((x) & 0x3) << 6) -# define EVERGREEN_GRPH_GREEN_SEL_G 0 -# define EVERGREEN_GRPH_GREEN_SEL_B 1 -# define EVERGREEN_GRPH_GREEN_SEL_A 2 -# define EVERGREEN_GRPH_GREEN_SEL_R 3 -#define EVERGREEN_GRPH_BLUE_CROSSBAR(x) (((x) & 0x3) << 8) -# define EVERGREEN_GRPH_BLUE_SEL_B 0 -# define EVERGREEN_GRPH_BLUE_SEL_A 1 -# define EVERGREEN_GRPH_BLUE_SEL_R 2 -# define EVERGREEN_GRPH_BLUE_SEL_G 3 -#define EVERGREEN_GRPH_ALPHA_CROSSBAR(x) (((x) & 0x3) << 10) -# define EVERGREEN_GRPH_ALPHA_SEL_A 0 -# define EVERGREEN_GRPH_ALPHA_SEL_R 1 -# define EVERGREEN_GRPH_ALPHA_SEL_G 2 -# define EVERGREEN_GRPH_ALPHA_SEL_B 3 - -#define EVERGREEN_D3VGA_CONTROL 0xf8 -#define EVERGREEN_D4VGA_CONTROL 0xf9 -#define EVERGREEN_D5VGA_CONTROL 0xfa -#define EVERGREEN_D6VGA_CONTROL 0xfb - -#define EVERGREEN_GRPH_SURFACE_ADDRESS_MASK 0xffffff00 - -#define EVERGREEN_GRPH_LUT_10BIT_BYPASS_CONTROL 0x1a02 -#define EVERGREEN_LUT_10BIT_BYPASS_EN (1 << 8) - -#define EVERGREEN_GRPH_PITCH 0x1a06 -#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x1a07 -#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH 0x1a08 -#define EVERGREEN_GRPH_SURFACE_OFFSET_X 0x1a09 -#define EVERGREEN_GRPH_SURFACE_OFFSET_Y 0x1a0a -#define EVERGREEN_GRPH_X_START 0x1a0b -#define EVERGREEN_GRPH_Y_START 0x1a0c -#define EVERGREEN_GRPH_X_END 0x1a0d -#define EVERGREEN_GRPH_Y_END 0x1a0e -#define EVERGREEN_GRPH_UPDATE 0x1a11 -#define EVERGREEN_GRPH_SURFACE_UPDATE_PENDING (1 << 2) -#define EVERGREEN_GRPH_UPDATE_LOCK (1 << 16) -#define EVERGREEN_GRPH_FLIP_CONTROL 0x1a12 -#define EVERGREEN_GRPH_SURFACE_UPDATE_H_RETRACE_EN (1 << 0) - -#define EVERGREEN_VIEWPORT_START 0x1b5c -#define EVERGREEN_VIEWPORT_SIZE 0x1b5d -#define EVERGREEN_DESKTOP_HEIGHT 0x1ac1 - -/* CUR blocks at 0x6998, 0x7598, 0x10198, 0x10d98, 0x11998, 0x12598 */ -#define EVERGREEN_CUR_CONTROL 0x1a66 -# define EVERGREEN_CURSOR_EN (1 << 0) -# define EVERGREEN_CURSOR_MODE(x) (((x) & 0x3) << 8) -# define EVERGREEN_CURSOR_MONO 0 -# define EVERGREEN_CURSOR_24_1 1 -# define EVERGREEN_CURSOR_24_8_PRE_MULT 2 -# define EVERGREEN_CURSOR_24_8_UNPRE_MULT 3 -# define EVERGREEN_CURSOR_2X_MAGNIFY (1 << 16) -# define EVERGREEN_CURSOR_FORCE_MC_ON (1 << 20) -# define EVERGREEN_CURSOR_URGENT_CONTROL(x) (((x) & 0x7) << 24) -# define EVERGREEN_CURSOR_URGENT_ALWAYS 0 -# define EVERGREEN_CURSOR_URGENT_1_8 1 -# define EVERGREEN_CURSOR_URGENT_1_4 2 -# define EVERGREEN_CURSOR_URGENT_3_8 3 -# define EVERGREEN_CURSOR_URGENT_1_2 4 -#define EVERGREEN_CUR_SURFACE_ADDRESS 0x1a67 -# define EVERGREEN_CUR_SURFACE_ADDRESS_MASK 0xfffff000 -#define EVERGREEN_CUR_SIZE 0x1a68 -#define EVERGREEN_CUR_SURFACE_ADDRESS_HIGH 0x1a69 -#define EVERGREEN_CUR_POSITION 0x1a6a -#define EVERGREEN_CUR_HOT_SPOT 0x1a6b -#define EVERGREEN_CUR_COLOR1 0x1a6c -#define EVERGREEN_CUR_COLOR2 0x1a6d -#define EVERGREEN_CUR_UPDATE 0x1a6e -# define EVERGREEN_CURSOR_UPDATE_PENDING (1 << 0) -# define EVERGREEN_CURSOR_UPDATE_TAKEN (1 << 1) -# define EVERGREEN_CURSOR_UPDATE_LOCK (1 << 16) -# define EVERGREEN_CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24) - - -#define NI_INPUT_CSC_CONTROL 0x1a35 -# define NI_INPUT_CSC_GRPH_MODE(x) (((x) & 0x3) << 0) -# define NI_INPUT_CSC_BYPASS 0 -# define NI_INPUT_CSC_PROG_COEFF 1 -# define NI_INPUT_CSC_PROG_SHARED_MATRIXA 2 -# define NI_INPUT_CSC_OVL_MODE(x) (((x) & 0x3) << 4) - -#define NI_OUTPUT_CSC_CONTROL 0x1a3c -# define NI_OUTPUT_CSC_GRPH_MODE(x) (((x) & 0x7) << 0) -# define NI_OUTPUT_CSC_BYPASS 0 -# define NI_OUTPUT_CSC_TV_RGB 1 -# define NI_OUTPUT_CSC_YCBCR_601 2 -# define NI_OUTPUT_CSC_YCBCR_709 3 -# define NI_OUTPUT_CSC_PROG_COEFF 4 -# define NI_OUTPUT_CSC_PROG_SHARED_MATRIXB 5 -# define NI_OUTPUT_CSC_OVL_MODE(x) (((x) & 0x7) << 4) - -#define NI_DEGAMMA_CONTROL 0x1a58 -# define NI_GRPH_DEGAMMA_MODE(x) (((x) & 0x3) << 0) -# define NI_DEGAMMA_BYPASS 0 -# define NI_DEGAMMA_SRGB_24 1 -# define NI_DEGAMMA_XVYCC_222 2 -# define NI_OVL_DEGAMMA_MODE(x) (((x) & 0x3) << 4) -# define NI_ICON_DEGAMMA_MODE(x) (((x) & 0x3) << 8) -# define NI_CURSOR_DEGAMMA_MODE(x) (((x) & 0x3) << 12) - -#define NI_GAMUT_REMAP_CONTROL 0x1a59 -# define NI_GRPH_GAMUT_REMAP_MODE(x) (((x) & 0x3) << 0) -# define NI_GAMUT_REMAP_BYPASS 0 -# define NI_GAMUT_REMAP_PROG_COEFF 1 -# define NI_GAMUT_REMAP_PROG_SHARED_MATRIXA 2 -# define NI_GAMUT_REMAP_PROG_SHARED_MATRIXB 3 -# define NI_OVL_GAMUT_REMAP_MODE(x) (((x) & 0x3) << 4) - -#define NI_REGAMMA_CONTROL 0x1aa0 -# define NI_GRPH_REGAMMA_MODE(x) (((x) & 0x7) << 0) -# define NI_REGAMMA_BYPASS 0 -# define NI_REGAMMA_SRGB_24 1 -# define NI_REGAMMA_XVYCC_222 2 -# define NI_REGAMMA_PROG_A 3 -# define NI_REGAMMA_PROG_B 4 -# define NI_OVL_REGAMMA_MODE(x) (((x) & 0x7) << 4) - - -#define NI_PRESCALE_GRPH_CONTROL 0x1a2d -# define NI_GRPH_PRESCALE_BYPASS (1 << 4) - -#define NI_PRESCALE_OVL_CONTROL 0x1a31 -# define NI_OVL_PRESCALE_BYPASS (1 << 4) - -#define NI_INPUT_GAMMA_CONTROL 0x1a10 -# define NI_GRPH_INPUT_GAMMA_MODE(x) (((x) & 0x3) << 0) -# define NI_INPUT_GAMMA_USE_LUT 0 -# define NI_INPUT_GAMMA_BYPASS 1 -# define NI_INPUT_GAMMA_SRGB_24 2 -# define NI_INPUT_GAMMA_XVYCC_222 3 -# define NI_OVL_INPUT_GAMMA_MODE(x) (((x) & 0x3) << 4) - -#define BLACKOUT_MODE_MASK 0x00000007 -#define VGA_RENDER_CONTROL 0xC0 -#define R_000300_VGA_RENDER_CONTROL 0xC0 -#define C_000300_VGA_VSTATUS_CNTL 0xFFFCFFFF -#define EVERGREEN_CRTC_STATUS 0x1BA3 -#define EVERGREEN_CRTC_V_BLANK (1 << 0) -#define EVERGREEN_CRTC_STATUS_POSITION 0x1BA4 -/* CRTC blocks at 0x6df0, 0x79f0, 0x105f0, 0x111f0, 0x11df0, 0x129f0 */ -#define EVERGREEN_CRTC_V_BLANK_START_END 0x1b8d -#define EVERGREEN_CRTC_CONTROL 0x1b9c -# define EVERGREEN_CRTC_MASTER_EN (1 << 0) -# define EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE (1 << 24) -#define EVERGREEN_CRTC_BLANK_CONTROL 0x1b9d -# define EVERGREEN_CRTC_BLANK_DATA_EN (1 << 8) -# define EVERGREEN_CRTC_V_BLANK (1 << 0) -#define EVERGREEN_CRTC_STATUS_HV_COUNT 0x1ba8 -#define EVERGREEN_CRTC_UPDATE_LOCK 0x1bb5 -#define EVERGREEN_MASTER_UPDATE_LOCK 0x1bbd -#define EVERGREEN_MASTER_UPDATE_MODE 0x1bbe -#define EVERGREEN_GRPH_UPDATE_LOCK (1 << 16) -#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x1a07 -#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH 0x1a08 -#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS 0x1a04 -#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS 0x1a05 -#define EVERGREEN_GRPH_UPDATE 0x1a11 -#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS 0xc4 -#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH 0xc9 -#define EVERGREEN_GRPH_SURFACE_UPDATE_PENDING (1 << 2) - -#define mmVM_CONTEXT1_CNTL__xxRANGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x10 -#define mmVM_CONTEXT1_CNTL__xxRANGE_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x4 -#define mmVM_CONTEXT1_CNTL__xxDUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x80 -#define mmVM_CONTEXT1_CNTL__xxDUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x7 -#define mmVM_CONTEXT1_CNTL__xxPDE0_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x400 -#define mmVM_CONTEXT1_CNTL__xxPDE0_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0xa -#define mmVM_CONTEXT1_CNTL__xxVALID_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x2000 -#define mmVM_CONTEXT1_CNTL__xxVALID_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0xd -#define mmVM_CONTEXT1_CNTL__xxREAD_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x10000 -#define mmVM_CONTEXT1_CNTL__xxREAD_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x10 -#define mmVM_CONTEXT1_CNTL__xxWRITE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x80000 -#define mmVM_CONTEXT1_CNTL__xxWRITE_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x13 - -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxVMID_MASK 0x1e000000 -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxVMID__SHIFT 0x19 -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxPROTECTIONS_MASK 0xff -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxPROTECTIONS__SHIFT 0x0 -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_ID_MASK 0xff000 -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_ID__SHIFT 0xc -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_RW_MASK 0x1000000 -#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_RW__SHIFT 0x18 - -#define mmMC_SHARED_BLACKOUT_CNTL__xxBLACKOUT_MODE_MASK 0x7 -#define mmMC_SHARED_BLACKOUT_CNTL__xxBLACKOUT_MODE__SHIFT 0x0 - -#define mmBIF_FB_EN__xxFB_READ_EN_MASK 0x1 -#define mmBIF_FB_EN__xxFB_READ_EN__SHIFT 0x0 -#define mmBIF_FB_EN__xxFB_WRITE_EN_MASK 0x2 -#define mmBIF_FB_EN__xxFB_WRITE_EN__SHIFT 0x1 - -#define mmSRBM_SOFT_RESET__xxSOFT_RESET_VMC_MASK 0x20000 -#define mmSRBM_SOFT_RESET__xxSOFT_RESET_VMC__SHIFT 0x11 -#define mmSRBM_SOFT_RESET__xxSOFT_RESET_MC_MASK 0x800 -#define mmSRBM_SOFT_RESET__xxSOFT_RESET_MC__SHIFT 0xb +#define GRPH_FORMAT_ARGB8888 0 +#define GRPH_FORMAT_ARGB2101010 1 +#define GRPH_FORMAT_32BPP_DIG 2 +#define GRPH_FORMAT_8B_ARGB2101010 3 +#define GRPH_FORMAT_BGRA1010102 4 +#define GRPH_FORMAT_8B_BGRA1010102 5 +#define GRPH_FORMAT_RGB111110 6 +#define GRPH_FORMAT_BGR101111 7 + +#define GRPH_ENDIAN_NONE 0 +#define GRPH_ENDIAN_8IN16 1 +#define GRPH_ENDIAN_8IN32 2 +#define GRPH_ENDIAN_8IN64 3 +#define GRPH_RED_SEL_R 0 +#define GRPH_RED_SEL_G 1 +#define GRPH_RED_SEL_B 2 +#define GRPH_RED_SEL_A 3 + +#define GRPH_GREEN_SEL_G 0 +#define GRPH_GREEN_SEL_B 1 +#define GRPH_GREEN_SEL_A 2 +#define GRPH_GREEN_SEL_R 3 + +#define GRPH_BLUE_SEL_B 0 +#define GRPH_BLUE_SEL_A 1 +#define GRPH_BLUE_SEL_R 2 +#define GRPH_BLUE_SEL_G 3 + +#define GRPH_ALPHA_SEL_A 0 +#define GRPH_ALPHA_SEL_R 1 +#define GRPH_ALPHA_SEL_G 2 +#define GRPH_ALPHA_SEL_B 3 + +/* CUR_CONTROL */ + #define CURSOR_MONO 0 + #define CURSOR_24_1 1 + #define CURSOR_24_8_PRE_MULT 2 + #define CURSOR_24_8_UNPRE_MULT 3 + #define CURSOR_URGENT_ALWAYS 0 + #define CURSOR_URGENT_1_8 1 + #define CURSOR_URGENT_1_4 2 + #define CURSOR_URGENT_3_8 3 + #define CURSOR_URGENT_1_2 4 + +/* INPUT_CSC_CONTROL */ +# define INPUT_CSC_BYPASS 0 +# define INPUT_CSC_PROG_COEFF 1 +# define INPUT_CSC_PROG_SHARED_MATRIXA 2 + +/* OUTPUT_CSC_CONTROL */ +# define OUTPUT_CSC_BYPASS 0 +# define OUTPUT_CSC_TV_RGB 1 +# define OUTPUT_CSC_YCBCR_601 2 +# define OUTPUT_CSC_YCBCR_709 3 +# define OUTPUT_CSC_PROG_COEFF 4 +# define OUTPUT_CSC_PROG_SHARED_MATRIXB 5 + +/* DEGAMMA_CONTROL */ +# define DEGAMMA_BYPASS 0 +# define DEGAMMA_SRGB_24 1 +# define DEGAMMA_XVYCC_222 2 + +/* GAMUT_REMAP_CONTROL */ +# define GAMUT_REMAP_BYPASS 0 +# define GAMUT_REMAP_PROG_COEFF 1 +# define GAMUT_REMAP_PROG_SHARED_MATRIXA 2 +# define GAMUT_REMAP_PROG_SHARED_MATRIXB 3 + +/* REGAMMA_CONTROL */ +# define REGAMMA_BYPASS 0 +# define REGAMMA_SRGB_24 1 +# define REGAMMA_XVYCC_222 2 +# define REGAMMA_PROG_A 3 +# define REGAMMA_PROG_B 4 + + +/* INPUT_GAMMA_CONTROL */ +# define INPUT_GAMMA_USE_LUT 0 +# define INPUT_GAMMA_BYPASS 1 +# define INPUT_GAMMA_SRGB_24 2 +# define INPUT_GAMMA_XVYCC_222 3 #define MC_SEQ_MISC0__MT__MASK 0xf0000000 #define MC_SEQ_MISC0__MT__GDDR1 0x10000000 @@ -2113,20 +785,9 @@ #define PACKET3_SEM_SEL_SIGNAL (0x6 << 29) #define PACKET3_SEM_SEL_WAIT (0x7 << 29) -#define CONFIG_CNTL 0x1509 -#define CC_DRM_ID_STRAPS 0X1559 #define AMDGPU_PCIE_INDEX 0xc #define AMDGPU_PCIE_DATA 0xd -#define DMA_SEM_INCOMPLETE_TIMER_CNTL 0x3411 -#define DMA_SEM_WAIT_FAIL_TIMER_CNTL 0x3412 -#define DMA_MODE 0x342f -#define DMA_RB_RPTR_ADDR_HI 0x3407 -#define DMA_RB_RPTR_ADDR_LO 0x3408 -#define DMA_BUSY_MASK 0x20 -#define DMA1_BUSY_MASK 0X40 -#define SDMA_MAX_INSTANCE 2 - #define PCIE_BUS_CLK 10000 #define TCLK (PCIE_BUS_CLK / 10) #define PCIE_PORT_INDEX 0xe @@ -2136,8 +797,6 @@ #define EVERGREEN_PIF_PHY1_INDEX 0x10 #define EVERGREEN_PIF_PHY1_DATA 0x14 -#define MC_VM_FB_OFFSET 0x81a - /* Discrete VCE clocks */ #define CG_VCEPLL_FUNC_CNTL 0xc0030600 #define VCEPLL_RESET_MASK 0x00000001 diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 659eab9b90be..c457be3a3c56 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -584,6 +584,8 @@ soc15_asic_reset_method(struct amdgpu_device *adev) * Enable triggering of GPU reset only if specified * by module parameter. */ + if (adev->pcie_reset_ctx.in_link_reset) + return AMD_RESET_METHOD_LINK; if (amdgpu_gpu_recovery == 4 || amdgpu_gpu_recovery == 5) return AMD_RESET_METHOD_MODE2; else if (!(adev->flags & AMD_IS_APU)) @@ -640,6 +642,9 @@ asic_reset: case AMD_RESET_METHOD_MODE2: dev_info(adev->dev, "MODE2 reset\n"); return amdgpu_dpm_mode2_reset(adev); + case AMD_RESET_METHOD_LINK: + dev_info(adev->dev, "Link reset\n"); + return amdgpu_device_link_reset(adev); default: dev_info(adev->dev, "MODE1 reset\n"); return amdgpu_device_mode1_reset(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/soc15d.h b/drivers/gpu/drm/amd/amdgpu/soc15d.h index a5000c171c02..cf93fa477674 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15d.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15d.h @@ -552,6 +552,11 @@ # define PACKET3_QUERY_STATUS_DOORBELL_OFFSET(x) ((x) << 2) # define PACKET3_QUERY_STATUS_ENG_SEL(x) ((x) << 25) +#define PACKET3_RUN_CLEANER_SHADER_9_0 0xD7 +/* 1. header + * 2. RESERVED [31:0] + */ + #define PACKET3_RUN_CLEANER_SHADER 0xD2 /* 1. header * 2. RESERVED [31:0] diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c index 0e404c074975..e590cbdd8de9 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c @@ -174,19 +174,76 @@ static void umc_v12_0_query_ras_error_count(struct amdgpu_device *adev, umc_v12_0_reset_error_count(adev); } +static void umc_v12_0_get_retire_flip_bits(struct amdgpu_device *adev) +{ + enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE; + uint32_t vram_type = adev->gmc.vram_type; + struct amdgpu_umc_flip_bits *flip_bits = &(adev->umc.flip_bits); + + if (adev->gmc.gmc_funcs->query_mem_partition_mode) + nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); + + /* default setting */ + flip_bits->flip_bits_in_pa[0] = UMC_V12_0_PA_C2_BIT; + flip_bits->flip_bits_in_pa[1] = UMC_V12_0_PA_C3_BIT; + flip_bits->flip_bits_in_pa[2] = UMC_V12_0_PA_C4_BIT; + flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R13_BIT; + flip_bits->flip_row_bit = 13; + flip_bits->bit_num = 4; + flip_bits->r13_in_pa = UMC_V12_0_PA_R13_BIT; + + if (nps == AMDGPU_NPS2_PARTITION_MODE) { + flip_bits->flip_bits_in_pa[0] = UMC_V12_0_PA_CH5_BIT; + flip_bits->flip_bits_in_pa[1] = UMC_V12_0_PA_C2_BIT; + flip_bits->flip_bits_in_pa[2] = UMC_V12_0_PA_B1_BIT; + flip_bits->r13_in_pa = UMC_V12_0_PA_R12_BIT; + } else if (nps == AMDGPU_NPS4_PARTITION_MODE) { + flip_bits->flip_bits_in_pa[0] = UMC_V12_0_PA_CH4_BIT; + flip_bits->flip_bits_in_pa[1] = UMC_V12_0_PA_CH5_BIT; + flip_bits->flip_bits_in_pa[2] = UMC_V12_0_PA_B0_BIT; + flip_bits->r13_in_pa = UMC_V12_0_PA_R11_BIT; + } + + switch (vram_type) { + case AMDGPU_VRAM_TYPE_HBM: + /* other nps modes are taken as nps1 */ + if (nps == AMDGPU_NPS2_PARTITION_MODE) + flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R12_BIT; + else if (nps == AMDGPU_NPS4_PARTITION_MODE) + flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R11_BIT; + + break; + case AMDGPU_VRAM_TYPE_HBM3E: + flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R12_BIT; + flip_bits->flip_row_bit = 12; + + if (nps == AMDGPU_NPS2_PARTITION_MODE) + flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R11_BIT; + else if (nps == AMDGPU_NPS4_PARTITION_MODE) + flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R10_BIT; + + break; + default: + dev_warn(adev->dev, + "Unknown HBM type, set RAS retire flip bits to the value in NPS1 mode.\n"); + break; + } + + adev->umc.retire_unit = 0x1 << flip_bits->bit_num; +} + static int umc_v12_0_convert_error_address(struct amdgpu_device *adev, struct ras_err_data *err_data, struct ta_ras_query_address_input *addr_in, struct ta_ras_query_address_output *addr_out, bool dump_addr) { - uint32_t col, col_lower, row, row_lower, bank; + uint32_t col, col_lower, row, row_lower, row_high, bank; uint32_t channel_index = 0, umc_inst = 0; - uint32_t i, loop_bits[UMC_V12_0_RETIRE_LOOP_BITS]; + uint32_t i, bit_num, retire_unit, *flip_bits; uint64_t soc_pa, column, err_addr; struct ta_ras_query_address_output addr_out_tmp; struct ta_ras_query_address_output *paddr_out; - enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE; int ret = 0; if (!addr_out) @@ -211,46 +268,46 @@ static int umc_v12_0_convert_error_address(struct amdgpu_device *adev, umc_inst = addr_in->ma.umc_inst; } - loop_bits[0] = UMC_V12_0_PA_C2_BIT; - loop_bits[1] = UMC_V12_0_PA_C3_BIT; - loop_bits[2] = UMC_V12_0_PA_C4_BIT; - loop_bits[3] = UMC_V12_0_PA_R13_BIT; - - if (adev->gmc.gmc_funcs->query_mem_partition_mode) - nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); - - /* other nps modes are taken as nps1 */ - if (nps == AMDGPU_NPS4_PARTITION_MODE) { - loop_bits[0] = UMC_V12_0_PA_CH4_BIT; - loop_bits[1] = UMC_V12_0_PA_CH5_BIT; - loop_bits[2] = UMC_V12_0_PA_B0_BIT; - loop_bits[3] = UMC_V12_0_PA_R11_BIT; - } + flip_bits = adev->umc.flip_bits.flip_bits_in_pa; + bit_num = adev->umc.flip_bits.bit_num; + retire_unit = adev->umc.retire_unit; soc_pa = paddr_out->pa.pa; channel_index = paddr_out->pa.channel_idx; /* clear loop bits in soc physical address */ - for (i = 0; i < UMC_V12_0_RETIRE_LOOP_BITS; i++) - soc_pa &= ~BIT_ULL(loop_bits[i]); + for (i = 0; i < bit_num; i++) + soc_pa &= ~BIT_ULL(flip_bits[i]); paddr_out->pa.pa = soc_pa; /* get column bit 0 and 1 in mca address */ col_lower = (err_addr >> 1) & 0x3ULL; - /* MA_R13_BIT will be handled later */ + /* extra row bit will be handled later */ row_lower = (err_addr >> UMC_V12_0_MA_R0_BIT) & 0x1fffULL; + row_lower &= ~BIT_ULL(adev->umc.flip_bits.flip_row_bit); + + if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(9, 5, 0)) { + row_high = (soc_pa >> adev->umc.flip_bits.r13_in_pa) & 0x3ULL; + /* it's 2.25GB in each channel, from MCA address to PA + * [R14 R13] is converted if the two bits value are 0x3, + * get them from PA instead of MCA address. + */ + row_lower |= (row_high << 13); + } if (!err_data && !dump_addr) goto out; /* loop for all possibilities of retired bits */ - for (column = 0; column < UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL; column++) { + for (column = 0; column < retire_unit; column++) { soc_pa = paddr_out->pa.pa; - for (i = 0; i < UMC_V12_0_RETIRE_LOOP_BITS; i++) - soc_pa |= (((column >> i) & 0x1ULL) << loop_bits[i]); + for (i = 0; i < bit_num; i++) + soc_pa |= (((column >> i) & 0x1ULL) << flip_bits[i]); col = ((column & 0x7) << 2) | col_lower; - /* add row bit 13 */ - row = ((column >> 3) << 13) | row_lower; + /* handle extra row bit */ + if (bit_num == RETIRE_FLIP_BITS_NUM) + row = ((column >> 3) << adev->umc.flip_bits.flip_row_bit) | + row_lower; if (dump_addr) dev_info(adev->dev, @@ -428,8 +485,12 @@ static int umc_v12_0_aca_bank_parser(struct aca_handle *handle, struct aca_bank bank->regs[ACA_REG_IDX_ADDR]); ext_error_code = ACA_REG__STATUS__ERRORCODEEXT(status); - count = ext_error_code == 0 ? - ACA_REG__MISC0__ERRCNT(bank->regs[ACA_REG_IDX_MISC0]) : 1ULL; + if (umc_v12_0_is_deferred_error(adev, status)) + count = ext_error_code == 0 ? + adev->umc.err_addr_cnt / adev->umc.retire_unit : 1ULL; + else + count = ext_error_code == 0 ? + ACA_REG__MISC0__ERRCNT(bank->regs[ACA_REG_IDX_MISC0]) : 1ULL; return aca_error_cache_log_bank_error(handle, &info, err_type, count); } @@ -469,8 +530,7 @@ static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev, uint64_t err_addr, pa_addr = 0; struct ras_ecc_err *ecc_err; struct ta_ras_query_address_output addr_out; - enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE; - uint32_t shift_bit = UMC_V12_0_PA_C4_BIT; + uint32_t shift_bit = adev->umc.flip_bits.flip_bits_in_pa[2]; int count, ret, i; hwid = REG_GET_FIELD(ipid, MCMP1_IPIDT0, HardwareID); @@ -515,11 +575,6 @@ static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev, ecc_err->pa_pfn = pa_addr >> AMDGPU_GPU_PAGE_SHIFT; ecc_err->channel_idx = addr_out.pa.channel_idx; - if (adev->gmc.gmc_funcs->query_mem_partition_mode) - nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); - if (nps == AMDGPU_NPS4_PARTITION_MODE) - shift_bit = UMC_V12_0_PA_B0_BIT; - /* If converted pa_pfn is 0, use pa C4 pfn. */ if (!ecc_err->pa_pfn) ecc_err->pa_pfn = BIT_ULL(shift_bit) >> AMDGPU_GPU_PAGE_SHIFT; @@ -665,5 +720,6 @@ struct amdgpu_umc_ras umc_v12_0_ras = { .update_ecc_status = umc_v12_0_update_ecc_status, .convert_ras_err_addr = umc_v12_0_convert_error_address, .get_die_id_from_pa = umc_v12_0_get_die_id, + .get_retire_flip_bits = umc_v12_0_get_retire_flip_bits, }; diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h index 9298018d938f..63b7e7254526 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h +++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h @@ -55,8 +55,6 @@ #define UMC_V12_0_NA_MAP_PA_NUM 8 /* R13 bit shift should be considered, double the number */ #define UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL (UMC_V12_0_NA_MAP_PA_NUM * 2) -/* C2, C3, C4, R13, four bits in MCA address are looped in retirement */ -#define UMC_V12_0_RETIRE_LOOP_BITS 4 /* column bits in SOC physical address */ #define UMC_V12_0_PA_C2_BIT 15 @@ -64,13 +62,16 @@ #define UMC_V12_0_PA_C4_BIT 21 /* row bits in SOC physical address */ #define UMC_V12_0_PA_R0_BIT 22 +#define UMC_V12_0_PA_R10_BIT 32 #define UMC_V12_0_PA_R11_BIT 33 +#define UMC_V12_0_PA_R12_BIT 34 #define UMC_V12_0_PA_R13_BIT 35 /* channel bit in SOC physical address */ #define UMC_V12_0_PA_CH4_BIT 12 #define UMC_V12_0_PA_CH5_BIT 13 /* bank bit in SOC physical address */ #define UMC_V12_0_PA_B0_BIT 19 +#define UMC_V12_0_PA_B1_BIT 20 /* row bits in MCA address */ #define UMC_V12_0_MA_R0_BIT 10 diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index 8e7a36f26e9c..b8d835c9e17e 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -39,6 +39,7 @@ #define VCN_VID_SOC_ADDRESS_2_0 0x1fa00 #define VCN1_VID_SOC_ADDRESS_3_0 0x48200 +#define VCN1_AON_SOC_ADDRESS_3_0 0x48000 #define mmUVD_CONTEXT_ID_INTERNAL_OFFSET 0x1fd #define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET 0x503 diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index d716510b8dd6..3eec1b8feaee 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -39,6 +39,7 @@ #define VCN_VID_SOC_ADDRESS_2_0 0x1fa00 #define VCN1_VID_SOC_ADDRESS_3_0 0x48200 +#define VCN1_AON_SOC_ADDRESS_3_0 0x48000 #define mmUVD_CONTEXT_ID_INTERNAL_OFFSET 0x27 #define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET 0x0f diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 22ae1939476f..0b19f0ab4480 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -40,6 +40,7 @@ #define VCN_VID_SOC_ADDRESS_2_0 0x1fa00 #define VCN1_VID_SOC_ADDRESS_3_0 0x48200 +#define VCN1_AON_SOC_ADDRESS_3_0 0x48000 #define mmUVD_CONTEXT_ID_INTERNAL_OFFSET 0x27 #define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET 0x0f diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index c6f6392c1c20..8fff470bce87 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -46,6 +46,7 @@ #define VCN_VID_SOC_ADDRESS_2_0 0x1fb00 #define VCN1_VID_SOC_ADDRESS_3_0 0x48300 +#define VCN1_AON_SOC_ADDRESS_3_0 0x48000 #define VCN_HARVEST_MMSCH 0 @@ -238,9 +239,9 @@ static int vcn_v4_0_sw_init(struct amdgpu_ip_block *ip_block) adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_pause_dpg_mode; } - /* TODO: Add queue reset mask when FW fully supports it */ adev->vcn.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]); + adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; if (amdgpu_sriov_vf(adev)) { r = amdgpu_virt_alloc_mm_table(adev); @@ -614,7 +615,8 @@ static void vcn_v4_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst, /* VCN global tiling registers */ WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( - VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect); + VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG), + adev->gfx.config.gb_addr_config, 0, indirect); } /** @@ -1945,6 +1947,20 @@ static int vcn_v4_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p, return 0; } +static int vcn_v4_0_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me]; + + if (!(adev->vcn.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)) + return -EOPNOTSUPP; + + vcn_v4_0_stop(vinst); + vcn_v4_0_start(vinst); + + return amdgpu_ring_test_helper(ring); +} + static struct amdgpu_ring_funcs vcn_v4_0_unified_ring_vm_funcs = { .type = AMDGPU_RING_TYPE_VCN_ENC, .align_mask = 0x3f, @@ -1974,6 +1990,7 @@ static struct amdgpu_ring_funcs vcn_v4_0_unified_ring_vm_funcs = { .emit_wreg = vcn_v2_0_enc_ring_emit_wreg, .emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = vcn_v4_0_ring_reset, }; /** diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index 3e176b4b7c69..712e1fba33ce 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -45,6 +45,7 @@ #define VCN_VID_SOC_ADDRESS_2_0 0x1fb00 #define VCN1_VID_SOC_ADDRESS_3_0 0x48300 +#define VCN1_AON_SOC_ADDRESS_3_0 0x48000 static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0_3[] = { SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS), @@ -287,6 +288,31 @@ static int vcn_v4_0_3_sw_fini(struct amdgpu_ip_block *ip_block) return 0; } +static int vcn_v4_0_3_hw_init_inst(struct amdgpu_vcn_inst *vinst) +{ + int vcn_inst; + struct amdgpu_device *adev = vinst->adev; + struct amdgpu_ring *ring; + int inst_idx = vinst->inst; + + vcn_inst = GET_INST(VCN, inst_idx); + ring = &adev->vcn.inst[inst_idx].ring_enc[0]; + if (ring->use_doorbell) { + adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 9 * vcn_inst, + adev->vcn.inst[inst_idx].aid_id); + + WREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL, + ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | + VCN_RB1_DB_CTRL__EN_MASK); + + /* Read DB_CTRL to flush the write DB_CTRL command. */ + RREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL); + } + + return 0; +} + /** * vcn_v4_0_3_hw_init - start and test VCN block * @@ -298,7 +324,8 @@ static int vcn_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; struct amdgpu_ring *ring; - int i, r, vcn_inst; + struct amdgpu_vcn_inst *vinst; + int i, r; if (amdgpu_sriov_vf(adev)) { r = vcn_v4_0_3_start_sriov(adev); @@ -321,28 +348,9 @@ static int vcn_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block) for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { struct amdgpu_vcn4_fw_shared *fw_shared; - vcn_inst = GET_INST(VCN, i); ring = &adev->vcn.inst[i].ring_enc[0]; - - if (ring->use_doorbell) { - adev->nbio.funcs->vcn_doorbell_range( - adev, ring->use_doorbell, - (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + - 9 * vcn_inst, - adev->vcn.inst[i].aid_id); - - WREG32_SOC15( - VCN, GET_INST(VCN, ring->me), - regVCN_RB1_DB_CTRL, - ring->doorbell_index - << VCN_RB1_DB_CTRL__OFFSET__SHIFT | - VCN_RB1_DB_CTRL__EN_MASK); - - /* Read DB_CTRL to flush the write DB_CTRL command. */ - RREG32_SOC15( - VCN, GET_INST(VCN, ring->me), - regVCN_RB1_DB_CTRL); - } + vinst = &adev->vcn.inst[i]; + vcn_v4_0_3_hw_init_inst(vinst); /* Re-init fw_shared when RAS fatal error occurred */ fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; @@ -1563,6 +1571,37 @@ static void vcn_v4_0_3_unified_ring_set_wptr(struct amdgpu_ring *ring) } } +static int vcn_v4_0_3_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + int r = 0; + int vcn_inst; + struct amdgpu_device *adev = ring->adev; + struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me]; + + if (amdgpu_sriov_vf(ring->adev)) + return -EOPNOTSUPP; + + if (!(adev->vcn.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)) + return -EOPNOTSUPP; + + vcn_inst = GET_INST(VCN, ring->me); + r = amdgpu_dpm_reset_vcn(adev, 1 << vcn_inst); + + if (r) { + DRM_DEV_ERROR(adev->dev, "VCN reset fail : %d\n", r); + return r; + } + + /* This flag is not set for VF, assumed to be disabled always */ + if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & 0x100) + adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED); + vcn_v4_0_3_hw_init_inst(vinst); + vcn_v4_0_3_start_dpg_mode(vinst, adev->vcn.inst[ring->me].indirect_sram); + r = amdgpu_ring_test_helper(ring); + + return r; +} + static const struct amdgpu_ring_funcs vcn_v4_0_3_unified_ring_vm_funcs = { .type = AMDGPU_RING_TYPE_VCN_ENC, .align_mask = 0x3f, @@ -1591,6 +1630,7 @@ static const struct amdgpu_ring_funcs vcn_v4_0_3_unified_ring_vm_funcs = { .emit_wreg = vcn_v4_0_3_enc_ring_emit_wreg, .emit_reg_wait = vcn_v4_0_3_enc_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = vcn_v4_0_3_ring_reset, }; /** diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index ba603b2246e2..a09f9a2dd471 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -46,6 +46,7 @@ #define VCN_VID_SOC_ADDRESS_2_0 0x1fb00 #define VCN1_VID_SOC_ADDRESS_3_0 (0x48300 + 0x38000) +#define VCN1_AON_SOC_ADDRESS_3_0 (0x48000 + 0x38000) #define VCN_HARVEST_MMSCH 0 @@ -207,6 +208,10 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG); + fw_shared->present_flag_0 |= AMDGPU_FW_SHARED_FLAG_0_DRM_KEY_INJECT; + fw_shared->drm_key_wa.method = + AMDGPU_DRM_KEY_INJECT_WORKAROUND_VCNFW_ASD_HANDSHAKING; + if (amdgpu_vcnfw_log) amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]); @@ -214,6 +219,13 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_5_pause_dpg_mode; } + adev->vcn.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]); + adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; + + r = amdgpu_vcn_sysfs_reset_mask_init(adev); + if (r) + return r; + if (amdgpu_sriov_vf(adev)) { r = amdgpu_virt_alloc_mm_table(adev); if (r) @@ -1022,6 +1034,10 @@ static int vcn_v4_0_5_start_dpg_mode(struct amdgpu_vcn_inst *vinst, ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | VCN_RB1_DB_CTRL__EN_MASK); + /* Keeping one read-back to ensure all register writes are done, otherwise + * it may introduce race conditions */ + RREG32_SOC15(VCN, inst_idx, regVCN_RB1_DB_CTRL); + return 0; } @@ -1204,6 +1220,10 @@ static int vcn_v4_0_5_start(struct amdgpu_vcn_inst *vinst) WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); + /* Keeping one read-back to ensure all register writes are done, otherwise + * it may introduce race conditions */ + RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + return 0; } @@ -1435,6 +1455,20 @@ static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring) } } +static int vcn_v4_0_5_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me]; + + if (!(adev->vcn.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)) + return -EOPNOTSUPP; + + vcn_v4_0_5_stop(vinst); + vcn_v4_0_5_start(vinst); + + return amdgpu_ring_test_helper(ring); +} + static struct amdgpu_ring_funcs vcn_v4_0_5_unified_ring_vm_funcs = { .type = AMDGPU_RING_TYPE_VCN_ENC, .align_mask = 0x3f, @@ -1462,6 +1496,7 @@ static struct amdgpu_ring_funcs vcn_v4_0_5_unified_ring_vm_funcs = { .emit_wreg = vcn_v2_0_enc_ring_emit_wreg, .emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = vcn_v4_0_5_ring_reset, }; /** diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c index d99d05f42f1d..27dcc6f37a73 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c @@ -196,9 +196,9 @@ static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block) adev->vcn.inst[i].pause_dpg_mode = vcn_v5_0_0_pause_dpg_mode; } - /* TODO: Add queue reset mask when FW fully supports it */ adev->vcn.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]); + adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; vcn_v5_0_0_alloc_ip_dump(adev); @@ -533,7 +533,8 @@ static void vcn_v5_0_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst, /* VCN global tiling registers */ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET( - VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect); + VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG), + adev->gfx.config.gb_addr_config, 0, indirect); return; } @@ -1171,6 +1172,20 @@ static void vcn_v5_0_0_unified_ring_set_wptr(struct amdgpu_ring *ring) } } +static int vcn_v5_0_0_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me]; + + if (!(adev->vcn.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)) + return -EOPNOTSUPP; + + vcn_v5_0_0_stop(vinst); + vcn_v5_0_0_start(vinst); + + return amdgpu_ring_test_helper(ring); +} + static const struct amdgpu_ring_funcs vcn_v5_0_0_unified_ring_vm_funcs = { .type = AMDGPU_RING_TYPE_VCN_ENC, .align_mask = 0x3f, @@ -1198,6 +1213,7 @@ static const struct amdgpu_ring_funcs vcn_v5_0_0_unified_ring_vm_funcs = { .emit_wreg = vcn_v2_0_enc_ring_emit_wreg, .emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = vcn_v5_0_0_ring_reset, }; /** diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index e0e84ef7f568..8e843011703c 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -30,6 +30,7 @@ #include "soc15_hw_ip.h" #include "vcn_v2_0.h" #include "vcn_v4_0_3.h" +#include "mmsch_v5_0.h" #include "vcn/vcn_5_0_0_offset.h" #include "vcn/vcn_5_0_0_sh_mask.h" @@ -39,6 +40,7 @@ #include <drm/drm_drv.h> +static int vcn_v5_0_1_start_sriov(struct amdgpu_device *adev); static void vcn_v5_0_1_set_unified_ring_funcs(struct amdgpu_device *adev); static void vcn_v5_0_1_set_irq_funcs(struct amdgpu_device *adev); static int vcn_v5_0_1_set_pg_state(struct amdgpu_vcn_inst *vinst, @@ -126,7 +128,14 @@ static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block) ring = &adev->vcn.inst[i].ring_enc[0]; ring->use_doorbell = true; - ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 9 * vcn_inst; + if (!amdgpu_sriov_vf(adev)) + ring->doorbell_index = + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 11 * vcn_inst; + else + ring->doorbell_index = + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 32 * vcn_inst; ring->vm_hub = AMDGPU_MMHUB0(adev->vcn.inst[i].aid_id); sprintf(ring->name, "vcn_unified_%d", adev->vcn.inst[i].aid_id); @@ -143,6 +152,12 @@ static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block) adev->vcn.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]); + if (amdgpu_sriov_vf(adev)) { + r = amdgpu_virt_alloc_mm_table(adev); + if (r) + return r; + } + vcn_v5_0_0_alloc_ip_dump(adev); return amdgpu_vcn_sysfs_reset_mask_init(adev); @@ -172,6 +187,9 @@ static int vcn_v5_0_1_sw_fini(struct amdgpu_ip_block *ip_block) drm_dev_exit(idx); } + if (amdgpu_sriov_vf(adev)) + amdgpu_virt_free_mm_table(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { r = amdgpu_vcn_suspend(adev, i); if (r) @@ -204,24 +222,38 @@ static int vcn_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block) struct amdgpu_ring *ring; int i, r, vcn_inst; - if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & 0x100) - adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - vcn_inst = GET_INST(VCN, i); - ring = &adev->vcn.inst[i].ring_enc[0]; - - if (ring->use_doorbell) - adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, - ((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + - 9 * vcn_inst), - adev->vcn.inst[i].aid_id); - - /* Re-init fw_shared, if required */ - vcn_v5_0_1_fw_shared_init(adev, i); - - r = amdgpu_ring_test_helper(ring); + if (amdgpu_sriov_vf(adev)) { + r = vcn_v5_0_1_start_sriov(adev); if (r) return r; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + ring = &adev->vcn.inst[i].ring_enc[0]; + ring->wptr = 0; + ring->wptr_old = 0; + vcn_v5_0_1_unified_ring_set_wptr(ring); + ring->sched.ready = true; + } + } else { + if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & 0x100) + adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + vcn_inst = GET_INST(VCN, i); + ring = &adev->vcn.inst[i].ring_enc[0]; + + if (ring->use_doorbell) + adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, + ((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 11 * vcn_inst), + adev->vcn.inst[i].aid_id); + + /* Re-init fw_shared, if required */ + vcn_v5_0_1_fw_shared_init(adev, i); + + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + } } return 0; @@ -663,6 +695,195 @@ static int vcn_v5_0_1_start_dpg_mode(struct amdgpu_vcn_inst *vinst, return 0; } +static int vcn_v5_0_1_start_sriov(struct amdgpu_device *adev) +{ + int i, vcn_inst; + struct amdgpu_ring *ring_enc; + uint64_t cache_addr; + uint64_t rb_enc_addr; + uint64_t ctx_addr; + uint32_t param, resp, expected; + uint32_t offset, cache_size; + uint32_t tmp, timeout; + + struct amdgpu_mm_table *table = &adev->virt.mm_table; + uint32_t *table_loc; + uint32_t table_size; + uint32_t size, size_dw; + uint32_t init_status; + uint32_t enabled_vcn; + + struct mmsch_v5_0_cmd_direct_write + direct_wt = { {0} }; + struct mmsch_v5_0_cmd_direct_read_modify_write + direct_rd_mod_wt = { {0} }; + struct mmsch_v5_0_cmd_end end = { {0} }; + struct mmsch_v5_0_init_header header; + + volatile struct amdgpu_vcn5_fw_shared *fw_shared; + volatile struct amdgpu_fw_shared_rb_setup *rb_setup; + + direct_wt.cmd_header.command_type = + MMSCH_COMMAND__DIRECT_REG_WRITE; + direct_rd_mod_wt.cmd_header.command_type = + MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE; + end.cmd_header.command_type = MMSCH_COMMAND__END; + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + vcn_inst = GET_INST(VCN, i); + + vcn_v5_0_1_fw_shared_init(adev, vcn_inst); + + memset(&header, 0, sizeof(struct mmsch_v5_0_init_header)); + header.version = MMSCH_VERSION; + header.total_size = sizeof(struct mmsch_v5_0_init_header) >> 2; + + table_loc = (uint32_t *)table->cpu_addr; + table_loc += header.total_size; + + table_size = 0; + + MMSCH_V5_0_INSERT_DIRECT_RD_MOD_WT(SOC15_REG_OFFSET(VCN, 0, regUVD_STATUS), + ~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY); + + cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4); + + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), + adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo); + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), + adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi); + + offset = 0; + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_OFFSET0), 0); + } else { + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), + lower_32_bits(adev->vcn.inst[i].gpu_addr)); + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), + upper_32_bits(adev->vcn.inst[i].gpu_addr)); + offset = cache_size; + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_OFFSET0), + AMDGPU_UVD_FIRMWARE_OFFSET >> 3); + } + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_SIZE0), + cache_size); + + cache_addr = adev->vcn.inst[vcn_inst].gpu_addr + offset; + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), lower_32_bits(cache_addr)); + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), upper_32_bits(cache_addr)); + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_OFFSET1), 0); + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE); + + cache_addr = adev->vcn.inst[vcn_inst].gpu_addr + offset + + AMDGPU_VCN_STACK_SIZE; + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW), lower_32_bits(cache_addr)); + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH), upper_32_bits(cache_addr)); + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_OFFSET2), 0); + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE); + + fw_shared = adev->vcn.inst[vcn_inst].fw_shared.cpu_addr; + rb_setup = &fw_shared->rb_setup; + + ring_enc = &adev->vcn.inst[vcn_inst].ring_enc[0]; + ring_enc->wptr = 0; + rb_enc_addr = ring_enc->gpu_addr; + + rb_setup->is_rb_enabled_flags |= RB_ENABLED; + rb_setup->rb_addr_lo = lower_32_bits(rb_enc_addr); + rb_setup->rb_addr_hi = upper_32_bits(rb_enc_addr); + rb_setup->rb_size = ring_enc->ring_size / 4; + fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG); + + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW), + lower_32_bits(adev->vcn.inst[vcn_inst].fw_shared.gpu_addr)); + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH), + upper_32_bits(adev->vcn.inst[vcn_inst].fw_shared.gpu_addr)); + MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0, + regUVD_VCPU_NONCACHE_SIZE0), + AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared))); + MMSCH_V5_0_INSERT_END(); + + header.vcn0.init_status = 0; + header.vcn0.table_offset = header.total_size; + header.vcn0.table_size = table_size; + header.total_size += table_size; + + /* Send init table to mmsch */ + size = sizeof(struct mmsch_v5_0_init_header); + table_loc = (uint32_t *)table->cpu_addr; + memcpy((void *)table_loc, &header, size); + + ctx_addr = table->gpu_addr; + WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr)); + WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr)); + + tmp = RREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_VMID); + tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK; + tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT); + WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_VMID, tmp); + + size = header.total_size; + WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_SIZE, size); + + WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_RESP, 0); + + param = 0x00000001; + WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_HOST, param); + tmp = 0; + timeout = 1000; + resp = 0; + expected = MMSCH_VF_MAILBOX_RESP__OK; + while (resp != expected) { + resp = RREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_RESP); + if (resp != 0) + break; + + udelay(10); + tmp = tmp + 10; + if (tmp >= timeout) { + DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\ + " waiting for regMMSCH_VF_MAILBOX_RESP "\ + "(expected=0x%08x, readback=0x%08x)\n", + tmp, expected, resp); + return -EBUSY; + } + } + + enabled_vcn = amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, 0) ? 1 : 0; + init_status = ((struct mmsch_v5_0_init_header *)(table_loc))->vcn0.init_status; + if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE + && init_status != MMSCH_VF_ENGINE_STATUS__PASS) { + DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init "\ + "status for VCN%x: 0x%x\n", resp, enabled_vcn, init_status); + } + } + + return 0; +} + /** * vcn_v5_0_1_start - VCN start * @@ -1103,8 +1324,18 @@ static int vcn_v5_0_1_set_clockgating_state(struct amdgpu_ip_block *ip_block, static int vcn_v5_0_1_set_pg_state(struct amdgpu_vcn_inst *vinst, enum amd_powergating_state state) { + struct amdgpu_device *adev = vinst->adev; int ret = 0; + /* for SRIOV, guest should not control VCN Power-gating + * MMSCH FW should control Power-gating and clock-gating + * guest should avoid touching CGC and PG + */ + if (amdgpu_sriov_vf(adev)) { + vinst->cur_state = AMD_PG_STATE_UNGATE; + return 0; + } + if (state == vinst->cur_state) return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c index faa0dd75dd6d..85846fd08ce4 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c @@ -350,6 +350,7 @@ static int vega20_ih_irq_init(struct amdgpu_device *adev) if (ret) return ret; } + ih[i]->overflow = false; } if (!amdgpu_sriov_vf(adev)) @@ -437,7 +438,10 @@ static u32 vega20_ih_get_wptr(struct amdgpu_device *adev, if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) goto out; - wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); + if (!amdgpu_sriov_vf(adev)) + wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); + else + ih->overflow = true; /* When a ring buffer overflow happen start parsing interrupt * from the last not overwritten vector (wptr + 32). Hopefully diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c b/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c index 4a5a0a4e00f2..9bde2c64540f 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c @@ -27,6 +27,16 @@ #include "kfd_priv.h" static struct dentry *debugfs_root; +static struct dentry *debugfs_proc; +static struct list_head procs; + +struct debugfs_proc_entry { + struct list_head list; + struct dentry *proc_dentry; + pid_t pid; +}; + +#define MAX_DEBUGFS_FILENAME_LEN 32 static int kfd_debugfs_open(struct inode *inode, struct file *file) { @@ -92,6 +102,8 @@ static const struct file_operations kfd_debugfs_hang_hws_fops = { void kfd_debugfs_init(void) { debugfs_root = debugfs_create_dir("kfd", NULL); + debugfs_proc = debugfs_create_dir("proc", debugfs_root); + INIT_LIST_HEAD(&procs); debugfs_create_file("mqds", S_IFREG | 0444, debugfs_root, kfd_debugfs_mqds_by_process, &kfd_debugfs_fops); @@ -107,5 +119,69 @@ void kfd_debugfs_init(void) void kfd_debugfs_fini(void) { + debugfs_remove_recursive(debugfs_proc); debugfs_remove_recursive(debugfs_root); } + +static ssize_t kfd_debugfs_pasid_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct kfd_process_device *pdd = file_inode(file)->i_private; + char tmp[32]; + int len; + + len = snprintf(tmp, sizeof(tmp), "%u\n", pdd->pasid); + + return simple_read_from_buffer(buf, count, ppos, tmp, len); +} + +static const struct file_operations kfd_debugfs_pasid_fops = { + .owner = THIS_MODULE, + .read = kfd_debugfs_pasid_read, +}; + +void kfd_debugfs_add_process(struct kfd_process *p) +{ + int i; + char name[MAX_DEBUGFS_FILENAME_LEN]; + struct debugfs_proc_entry *entry; + + entry = kzalloc(sizeof(*entry), GFP_KERNEL); + if (!entry) + return; + + list_add(&entry->list, &procs); + entry->pid = p->lead_thread->pid; + snprintf(name, MAX_DEBUGFS_FILENAME_LEN, "%d", + (int)entry->pid); + entry->proc_dentry = debugfs_create_dir(name, debugfs_proc); + + /* Create debugfs files for each GPU: + * - proc/<pid>/pasid_<gpuid> + */ + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + + snprintf(name, MAX_DEBUGFS_FILENAME_LEN, "pasid_%u", + pdd->dev->id); + debugfs_create_file((const char *)name, S_IFREG | 0444, + entry->proc_dentry, pdd, + &kfd_debugfs_pasid_fops); + } +} + +void kfd_debugfs_remove_process(struct kfd_process *p) +{ + struct debugfs_proc_entry *entry, *next; + + mutex_lock(&kfd_processes_mutex); + list_for_each_entry_safe(entry, next, &procs, list) { + if (entry->pid != p->lead_thread->pid) + continue; + + debugfs_remove_recursive(entry->proc_dentry); + list_del(&entry->list); + kfree(entry); + } + mutex_unlock(&kfd_processes_mutex); +} diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index b9c82be6ce13..bf0854bd5555 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -352,11 +352,6 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) f2g = &aldebaran_kfd2kgd; break; case IP_VERSION(9, 4, 3): - gfx_target_version = adev->rev_id >= 1 ? 90402 - : adev->flags & AMD_IS_APU ? 90400 - : 90401; - f2g = &gc_9_4_3_kfd2kgd; - break; case IP_VERSION(9, 4, 4): gfx_target_version = 90402; f2g = &gc_9_4_3_kfd2kgd; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index c610e172a2b8..76359c6a3f3a 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -1576,8 +1576,9 @@ static int allocate_sdma_queue(struct device_queue_manager *dqm, int bit; if (q->properties.type == KFD_QUEUE_TYPE_SDMA) { - if (bitmap_empty(dqm->sdma_bitmap, KFD_MAX_SDMA_QUEUES)) { - dev_err(dev, "No more SDMA queue to allocate\n"); + if (bitmap_empty(dqm->sdma_bitmap, get_num_sdma_queues(dqm))) { + dev_warn(dev, "No more SDMA queue to allocate (%d total queues)\n", + get_num_sdma_queues(dqm)); return -ENOMEM; } @@ -1602,8 +1603,9 @@ static int allocate_sdma_queue(struct device_queue_manager *dqm, q->properties.sdma_queue_id = q->sdma_id / kfd_get_num_sdma_engines(dqm->dev); } else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) { - if (bitmap_empty(dqm->xgmi_sdma_bitmap, KFD_MAX_SDMA_QUEUES)) { - dev_err(dev, "No more XGMI SDMA queue to allocate\n"); + if (bitmap_empty(dqm->xgmi_sdma_bitmap, get_num_xgmi_sdma_queues(dqm))) { + dev_warn(dev, "No more XGMI SDMA queue to allocate (%d total queues)\n", + get_num_xgmi_sdma_queues(dqm)); return -ENOMEM; } if (restore_sdma_id) { @@ -1662,8 +1664,8 @@ static int allocate_sdma_queue(struct device_queue_manager *dqm, } if (!free_bit_found) { - dev_err(dev, "No more SDMA queue to allocate for target ID %i\n", - q->properties.sdma_engine_id); + dev_warn(dev, "No more SDMA queue to allocate for target ID %i (%d total queues)\n", + q->properties.sdma_engine_id, num_queues); return -ENOMEM; } } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_events.c index fecdb6794075..e54e708ed82d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_events.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.c @@ -1177,6 +1177,25 @@ void kfd_signal_hw_exception_event(u32 pasid) kfd_unref_process(p); } +void kfd_signal_vm_fault_event_with_userptr(struct kfd_process *p, uint64_t gpu_va) +{ + struct kfd_process_device *pdd; + struct kfd_hsa_memory_exception_data exception_data; + int i; + + memset(&exception_data, 0, sizeof(exception_data)); + exception_data.va = gpu_va; + exception_data.failure.NotPresent = 1; + + // Send VM seg fault to all kfd process device + for (i = 0; i < p->n_pdds; i++) { + pdd = p->pdds[i]; + exception_data.gpu_id = pdd->user_gpu_id; + kfd_evict_process_device(pdd); + kfd_signal_vm_fault_event(pdd, NULL, &exception_data); + } +} + void kfd_signal_vm_fault_event(struct kfd_process_device *pdd, struct kfd_vm_fault_info *info, struct kfd_hsa_memory_exception_data *data) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c index 37b69fe0ede3..3e1ad8974797 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c @@ -168,14 +168,14 @@ static bool event_interrupt_isr_v10(struct kfd_node *dev, client_id != SOC15_IH_CLIENTID_SE3SH) return false; - pr_debug("client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", - client_id, source_id, vmid, pasid); - pr_debug("%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", - data[0], data[1], data[2], data[3], - data[4], data[5], data[6], data[7]); + dev_dbg(dev->adev->dev, + "client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", + client_id, source_id, vmid, pasid); + dev_dbg(dev->adev->dev, "%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", + data[0], data[1], data[2], data[3], data[4], data[5], data[6], + data[7]); - /* If there is no valid PASID, it's likely a bug */ - if (WARN_ONCE(pasid == 0, "Bug: No PASID in KFD interrupt")) + if (pasid == 0) return 0; /* Interrupt types we care about: various signals and faults. @@ -217,37 +217,66 @@ static void event_interrupt_wq_v10(struct kfd_node *dev, SQ_INTERRUPT_WORD_WAVE_CTXID1, ENCODING); switch (encoding) { case SQ_INTERRUPT_WORD_ENCODING_AUTO: - pr_debug_ratelimited( + dev_dbg_ratelimited( + dev->adev->dev, "sq_intr: auto, se %d, ttrace %d, wlt %d, ttrac_buf0_full %d, ttrac_buf1_full %d, ttrace_utc_err %d\n", - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_AUTO_CTXID1, - SE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, - THREAD_TRACE), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, - WLT), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, - THREAD_TRACE_BUF0_FULL), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, - THREAD_TRACE_BUF1_FULL), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, - THREAD_TRACE_UTC_ERROR)); + REG_GET_FIELD( + context_id1, + SQ_INTERRUPT_WORD_AUTO_CTXID1, + SE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID0, + WLT), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_BUF0_FULL), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_BUF1_FULL), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_UTC_ERROR)); break; case SQ_INTERRUPT_WORD_ENCODING_INST: - pr_debug_ratelimited("sq_intr: inst, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n", - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, - SE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - DATA), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - SA_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - PRIV), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - WAVE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - SIMD_ID), - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, - WGP_ID)); + dev_dbg_ratelimited( + dev->adev->dev, + "sq_intr: inst, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n", + REG_GET_FIELD( + context_id1, + SQ_INTERRUPT_WORD_WAVE_CTXID1, + SE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + DATA), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + SA_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + PRIV), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + WAVE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + SIMD_ID), + REG_GET_FIELD( + context_id1, + SQ_INTERRUPT_WORD_WAVE_CTXID1, + WGP_ID)); if (context_id0 & SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV_MASK) { if (kfd_set_dbg_ev_from_interrupt(dev, pasid, KFD_DEBUG_DOORBELL_ID(context_id0), @@ -259,21 +288,37 @@ static void event_interrupt_wq_v10(struct kfd_node *dev, case SQ_INTERRUPT_WORD_ENCODING_ERROR: sq_intr_err_type = REG_GET_FIELD(context_id0, KFD_CTXID0, ERR_TYPE); - pr_warn_ratelimited("sq_intr: error, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d, err_type %d\n", - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, - SE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - DATA), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - SA_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - PRIV), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - WAVE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, - SIMD_ID), - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, - WGP_ID), + dev_warn_ratelimited( + dev->adev->dev, + "sq_intr: error, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d, err_type %d\n", + REG_GET_FIELD( + context_id1, + SQ_INTERRUPT_WORD_WAVE_CTXID1, + SE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + DATA), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + SA_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + PRIV), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + WAVE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID0, + SIMD_ID), + REG_GET_FIELD( + context_id1, + SQ_INTERRUPT_WORD_WAVE_CTXID1, + WGP_ID), sq_intr_err_type); break; default: diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c index c5f97e6e36ff..2788a52714d1 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c @@ -148,44 +148,69 @@ enum SQ_INTERRUPT_ERROR_TYPE { #define KFD_CTXID0_DOORBELL_ID(ctxid0) ((ctxid0) & \ KFD_CTXID0_DOORBELL_ID_MASK) -static void print_sq_intr_info_auto(uint32_t context_id0, uint32_t context_id1) +static void print_sq_intr_info_auto(struct kfd_node *dev, uint32_t context_id0, + uint32_t context_id1) { - pr_debug_ratelimited( + dev_dbg_ratelimited( + dev->adev->dev, "sq_intr: auto, ttrace %d, wlt %d, ttrace_buf_full %d, reg_tms %d, cmd_tms %d, host_cmd_ovf %d, host_reg_ovf %d, immed_ovf %d, ttrace_utc_err %d\n", - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, THREAD_TRACE), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE), REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, WLT), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, THREAD_TRACE_BUF_FULL), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, REG_TIMESTAMP), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, CMD_TIMESTAMP), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, HOST_CMD_OVERFLOW), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, HOST_REG_OVERFLOW), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, IMMED_OVERFLOW), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, THREAD_TRACE_UTC_ERROR)); + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_BUF_FULL), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + REG_TIMESTAMP), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + CMD_TIMESTAMP), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + HOST_CMD_OVERFLOW), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + HOST_REG_OVERFLOW), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + IMMED_OVERFLOW), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_UTC_ERROR)); } -static void print_sq_intr_info_inst(uint32_t context_id0, uint32_t context_id1) +static void print_sq_intr_info_inst(struct kfd_node *dev, uint32_t context_id0, + uint32_t context_id1) { - pr_debug_ratelimited( + dev_dbg_ratelimited( + dev->adev->dev, "sq_intr: inst, data 0x%08x, sh %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n", REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, DATA), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, SH_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + SH_ID), REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, PRIV), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, WAVE_ID), - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, SIMD_ID), - REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, WGP_ID)); + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + WAVE_ID), + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, + SIMD_ID), + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, + WGP_ID)); } -static void print_sq_intr_info_error(uint32_t context_id0, uint32_t context_id1) +static void print_sq_intr_info_error(struct kfd_node *dev, uint32_t context_id0, + uint32_t context_id1) { - pr_warn_ratelimited( + dev_warn_ratelimited( + dev->adev->dev, "sq_intr: error, detail 0x%08x, type %d, sh %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n", - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, DETAIL), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, TYPE), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, SH_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, PRIV), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, WAVE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1, SIMD_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1, WGP_ID)); + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, + DETAIL), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, + TYPE), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, + SH_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, + PRIV), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, + WAVE_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1, + SIMD_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1, + WGP_ID)); } static void event_interrupt_poison_consumption_v11(struct kfd_node *dev, @@ -255,14 +280,14 @@ static bool event_interrupt_isr_v11(struct kfd_node *dev, (context_id0 & AMDGPU_FENCE_MES_QUEUE_FLAG)) return false; - pr_debug("client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", - client_id, source_id, vmid, pasid); - pr_debug("%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", - data[0], data[1], data[2], data[3], - data[4], data[5], data[6], data[7]); + dev_dbg(dev->adev->dev, + "client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", + client_id, source_id, vmid, pasid); + dev_dbg(dev->adev->dev, "%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", + data[0], data[1], data[2], data[3], data[4], data[5], data[6], + data[7]); - /* If there is no valid PASID, it's likely a bug */ - if (WARN_ONCE(pasid == 0, "Bug: No PASID in KFD interrupt")) + if (pasid == 0) return false; /* Interrupt types we care about: various signals and faults. @@ -353,10 +378,10 @@ static void event_interrupt_wq_v11(struct kfd_node *dev, SQ_INTERRUPT_WORD_WAVE_CTXID1, ENCODING); switch (sq_int_enc) { case SQ_INTERRUPT_WORD_ENCODING_AUTO: - print_sq_intr_info_auto(context_id0, context_id1); + print_sq_intr_info_auto(dev, context_id0, context_id1); break; case SQ_INTERRUPT_WORD_ENCODING_INST: - print_sq_intr_info_inst(context_id0, context_id1); + print_sq_intr_info_inst(dev, context_id0, context_id1); sq_int_priv = REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, PRIV); if (sq_int_priv && (kfd_set_dbg_ev_from_interrupt(dev, pasid, @@ -366,7 +391,7 @@ static void event_interrupt_wq_v11(struct kfd_node *dev, return; break; case SQ_INTERRUPT_WORD_ENCODING_ERROR: - print_sq_intr_info_error(context_id0, context_id1); + print_sq_intr_info_error(dev, context_id0, context_id1); sq_int_errtype = REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0, TYPE); if (sq_int_errtype != SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST && diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c index b8a91bf4ef30..4ceb251312a6 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c @@ -314,11 +314,12 @@ static bool event_interrupt_isr_v9(struct kfd_node *dev, & ~pasid_mask) | pasid); } - pr_debug("client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", - client_id, source_id, vmid, pasid); - pr_debug("%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", - data[0], data[1], data[2], data[3], - data[4], data[5], data[6], data[7]); + dev_dbg(dev->adev->dev, + "client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", + client_id, source_id, vmid, pasid); + dev_dbg(dev->adev->dev, "%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", + data[0], data[1], data[2], data[3], data[4], data[5], data[6], + data[7]); /* If there is no valid PASID, it's likely a bug */ if (WARN_ONCE(pasid == 0, "Bug: No PASID in KFD interrupt")) @@ -379,28 +380,82 @@ static void event_interrupt_wq_v9(struct kfd_node *dev, encoding = REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, ENCODING); switch (encoding) { case SQ_INTERRUPT_WORD_ENCODING_AUTO: - pr_debug_ratelimited( + dev_dbg_ratelimited( + dev->adev->dev, "sq_intr: auto, se %d, ttrace %d, wlt %d, ttrac_buf_full %d, reg_tms %d, cmd_tms %d, host_cmd_ovf %d, host_reg_ovf %d, immed_ovf %d, ttrace_utc_err %d\n", - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, SE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, THREAD_TRACE), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, WLT), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, THREAD_TRACE_BUF_FULL), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, REG_TIMESTAMP), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, CMD_TIMESTAMP), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, HOST_CMD_OVERFLOW), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, HOST_REG_OVERFLOW), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, IMMED_OVERFLOW), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, THREAD_TRACE_UTC_ERROR)); + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + SE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + THREAD_TRACE), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + WLT), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + THREAD_TRACE_BUF_FULL), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + REG_TIMESTAMP), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + CMD_TIMESTAMP), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + HOST_CMD_OVERFLOW), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + HOST_REG_OVERFLOW), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + IMMED_OVERFLOW), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_AUTO_CTXID, + THREAD_TRACE_UTC_ERROR)); break; case SQ_INTERRUPT_WORD_ENCODING_INST: - pr_debug_ratelimited("sq_intr: inst, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, intr_data 0x%x\n", - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, DATA), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SH_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, PRIV), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, WAVE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SIMD_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, CU_ID), + dev_dbg_ratelimited( + dev->adev->dev, + "sq_intr: inst, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, intr_data 0x%x\n", + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + SE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + DATA), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + SH_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + PRIV), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + WAVE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + SIMD_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + CU_ID), sq_int_data); if (context_id0 & SQ_INTERRUPT_WORD_WAVE_CTXID__PRIV_MASK) { if (kfd_set_dbg_ev_from_interrupt(dev, pasid, @@ -412,14 +467,37 @@ static void event_interrupt_wq_v9(struct kfd_node *dev, break; case SQ_INTERRUPT_WORD_ENCODING_ERROR: sq_intr_err = REG_GET_FIELD(sq_int_data, KFD_SQ_INT_DATA, ERR_TYPE); - pr_warn_ratelimited("sq_intr: error, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, err_type %d\n", - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, DATA), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SH_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, PRIV), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, WAVE_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SIMD_ID), - REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, CU_ID), + dev_warn_ratelimited( + dev->adev->dev, + "sq_intr: error, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, err_type %d\n", + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + SE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + DATA), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + SH_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + PRIV), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + WAVE_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + SIMD_ID), + REG_GET_FIELD( + context_id0, + SQ_INTERRUPT_WORD_WAVE_CTXID, + CU_ID), sq_intr_err); if (sq_intr_err != SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST && sq_intr_err != SQ_INTERRUPT_ERROR_TYPE_MEMVIOL) { diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c index 80320a6c8854..97933d2a3803 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c @@ -495,6 +495,10 @@ static void update_mqd_sdma(struct mqd_manager *mm, void *mqd, m->sdma_engine_id = q->sdma_engine_id; m->sdma_queue_id = q->sdma_queue_id; m->sdmax_rlcx_dummy_reg = SDMA_RLC_DUMMY_DEFAULT; + /* Allow context switch so we don't cross-process starve with a massive + * command buffer of long-running SDMA commands + */ + m->sdmax_rlcx_ib_cntl |= SDMA0_GFX_IB_CNTL__SWITCH_INSIDE_IB_MASK; q->is_active = QUEUE_IS_ACTIVE(*q); } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c index 271c567242ab..b1a6eb349bb3 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c @@ -31,6 +31,7 @@ #define OVER_SUBSCRIPTION_PROCESS_COUNT (1 << 0) #define OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT (1 << 1) #define OVER_SUBSCRIPTION_GWS_QUEUE_COUNT (1 << 2) +#define OVER_SUBSCRIPTION_XNACK_CONFLICT (1 << 3) static inline void inc_wptr(unsigned int *wptr, unsigned int increment_bytes, unsigned int buffer_size_bytes) @@ -44,7 +45,8 @@ static inline void inc_wptr(unsigned int *wptr, unsigned int increment_bytes, static void pm_calc_rlib_size(struct packet_manager *pm, unsigned int *rlib_size, - int *over_subscription) + int *over_subscription, + int xnack_conflict) { unsigned int process_count, queue_count, compute_queue_count, gws_queue_count; unsigned int map_queue_size; @@ -73,6 +75,8 @@ static void pm_calc_rlib_size(struct packet_manager *pm, *over_subscription |= OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT; if (gws_queue_count > 1) *over_subscription |= OVER_SUBSCRIPTION_GWS_QUEUE_COUNT; + if (xnack_conflict && (node->adev->gmc.xnack_flags & AMDGPU_GMC_XNACK_FLAG_CHAIN)) + *over_subscription |= OVER_SUBSCRIPTION_XNACK_CONFLICT; if (*over_subscription) dev_dbg(dev, "Over subscribed runlist\n"); @@ -96,7 +100,8 @@ static int pm_allocate_runlist_ib(struct packet_manager *pm, unsigned int **rl_buffer, uint64_t *rl_gpu_buffer, unsigned int *rl_buffer_size, - int *is_over_subscription) + int *is_over_subscription, + int xnack_conflict) { struct kfd_node *node = pm->dqm->dev; struct device *dev = node->adev->dev; @@ -105,7 +110,8 @@ static int pm_allocate_runlist_ib(struct packet_manager *pm, if (WARN_ON(pm->allocated)) return -EINVAL; - pm_calc_rlib_size(pm, rl_buffer_size, is_over_subscription); + pm_calc_rlib_size(pm, rl_buffer_size, is_over_subscription, + xnack_conflict); mutex_lock(&pm->lock); @@ -142,11 +148,27 @@ static int pm_create_runlist_ib(struct packet_manager *pm, struct queue *q; struct kernel_queue *kq; int is_over_subscription; + int xnack_enabled = -1; + bool xnack_conflict = 0; rl_wptr = retval = processes_mapped = 0; + /* Check if processes set different xnack modes */ + list_for_each_entry(cur, queues, list) { + qpd = cur->qpd; + if (xnack_enabled < 0) + /* First process */ + xnack_enabled = qpd->pqm->process->xnack_enabled; + else if (qpd->pqm->process->xnack_enabled != xnack_enabled) { + /* Found a process with a different xnack mode */ + xnack_conflict = 1; + break; + } + } + retval = pm_allocate_runlist_ib(pm, &rl_buffer, rl_gpu_addr, - &alloc_size_bytes, &is_over_subscription); + &alloc_size_bytes, &is_over_subscription, + xnack_conflict); if (retval) return retval; @@ -156,9 +178,13 @@ static int pm_create_runlist_ib(struct packet_manager *pm, dev_dbg(dev, "Building runlist ib process count: %d queues count %d\n", pm->dqm->processes_count, pm->dqm->active_queue_count); +build_runlist_ib: /* build the run list ib packet */ list_for_each_entry(cur, queues, list) { qpd = cur->qpd; + /* group processes with the same xnack mode together */ + if (qpd->pqm->process->xnack_enabled != xnack_enabled) + continue; /* build map process packet */ if (processes_mapped >= pm->dqm->processes_count) { dev_dbg(dev, "Not enough space left in runlist IB\n"); @@ -215,18 +241,26 @@ static int pm_create_runlist_ib(struct packet_manager *pm, alloc_size_bytes); } } + if (xnack_conflict) { + /* pick up processes with the other xnack mode */ + xnack_enabled = !xnack_enabled; + xnack_conflict = 0; + goto build_runlist_ib; + } dev_dbg(dev, "Finished map process and queues to runlist\n"); if (is_over_subscription) { if (!pm->is_over_subscription) - dev_warn(dev, "Runlist is getting oversubscribed due to%s%s%s. Expect reduced ROCm performance.\n", - is_over_subscription & OVER_SUBSCRIPTION_PROCESS_COUNT ? - " too many processes." : "", - is_over_subscription & OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT ? - " too many queues." : "", - is_over_subscription & OVER_SUBSCRIPTION_GWS_QUEUE_COUNT ? - " multiple processes using cooperative launch." : ""); + dev_warn(dev, "Runlist is getting oversubscribed due to%s%s%s%s. Expect reduced ROCm performance.\n", + is_over_subscription & OVER_SUBSCRIPTION_PROCESS_COUNT ? + " too many processes" : "", + is_over_subscription & OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT ? + " too many queues" : "", + is_over_subscription & OVER_SUBSCRIPTION_GWS_QUEUE_COUNT ? + " multiple processes using cooperative launch" : "", + is_over_subscription & OVER_SUBSCRIPTION_XNACK_CONFLICT ? + " xnack on/off processes mixed on gfx9" : ""); retval = pm->pmf->runlist(pm, &rl_buffer[rl_wptr], *rl_gpu_addr, diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c index 2893fd5e5d00..8fa6489b6f5d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c @@ -43,7 +43,7 @@ static int pm_map_process_v9(struct packet_manager *pm, memset(buffer, 0, sizeof(struct pm4_mes_map_process)); packet->header.u32All = pm_build_pm4_header(IT_MAP_PROCESS, sizeof(struct pm4_mes_map_process)); - if (adev->enforce_isolation[kfd->node_id]) + if (adev->enforce_isolation[kfd->node_id] == AMDGPU_ENFORCE_ISOLATION_ENABLE) packet->bitfields2.exec_cleaner_shader = 1; packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0; packet->bitfields2.process_quantum = 10; @@ -102,7 +102,8 @@ static int pm_map_process_aldebaran(struct packet_manager *pm, memset(buffer, 0, sizeof(struct pm4_mes_map_process_aldebaran)); packet->header.u32All = pm_build_pm4_header(IT_MAP_PROCESS, sizeof(struct pm4_mes_map_process_aldebaran)); - if (adev->enforce_isolation[knode->node_id]) + if (adev->enforce_isolation[knode->node_id] == + AMDGPU_ENFORCE_ISOLATION_ENABLE) packet->bitfields2.exec_cleaner_shader = 1; packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0; packet->bitfields2.process_quantum = 10; @@ -165,9 +166,9 @@ static int pm_runlist_v9(struct packet_manager *pm, uint32_t *buffer, * hws_max_conc_proc has been done in * kgd2kfd_device_init(). */ - concurrent_proc_cnt = adev->enforce_isolation[kfd->node_id] ? - 1 : min(pm->dqm->processes_count, - kfd->max_proc_per_quantum); + concurrent_proc_cnt = (adev->enforce_isolation[kfd->node_id] == + AMDGPU_ENFORCE_ISOLATION_ENABLE) ? + 1 : min(pm->dqm->processes_count, kfd->max_proc_per_quantum); packet = (struct pm4_mes_runlist *)buffer; @@ -202,6 +203,8 @@ static int pm_set_resources_v9(struct packet_manager *pm, uint32_t *buffer, queue_type__mes_set_resources__hsa_interface_queue_hiq; packet->bitfields2.vmid_mask = res->vmid_mask; packet->bitfields2.unmap_latency = KFD_UNMAP_LATENCY_MS / 100; + if (pm->dqm->dev->adev->gmc.xnack_flags & AMDGPU_GMC_XNACK_FLAG_CHAIN) + packet->bitfields2.enb_xnack_retry_disable_check = 1; packet->bitfields7.oac_mask = res->oac_mask; packet->bitfields8.gds_heap_base = res->gds_heap_base; packet->bitfields8.gds_heap_size = res->gds_heap_size; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h index cd8611401a66..e356a207d03c 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h @@ -63,7 +63,8 @@ struct pm4_mes_set_resources { struct { uint32_t vmid_mask:16; uint32_t unmap_latency:8; - uint32_t reserved1:5; + uint32_t reserved1:4; + uint32_t enb_xnack_retry_disable_check:1; enum mes_set_resources_queue_type_enum queue_type:3; } bitfields2; uint32_t ordinal2; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index f6aedf69c644..d221c58dccc3 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -1507,6 +1507,8 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p, int kfd_get_num_events(struct kfd_process *p); int kfd_event_destroy(struct kfd_process *p, uint32_t event_id); +void kfd_signal_vm_fault_event_with_userptr(struct kfd_process *p, uint64_t gpu_va); + void kfd_signal_vm_fault_event(struct kfd_process_device *pdd, struct kfd_vm_fault_info *info, struct kfd_hsa_memory_exception_data *data); @@ -1581,10 +1583,15 @@ int kfd_debugfs_hang_hws(struct kfd_node *dev); int pm_debugfs_hang_hws(struct packet_manager *pm); int dqm_debugfs_hang_hws(struct device_queue_manager *dqm); +void kfd_debugfs_add_process(struct kfd_process *p); +void kfd_debugfs_remove_process(struct kfd_process *p); + #else static inline void kfd_debugfs_init(void) {} static inline void kfd_debugfs_fini(void) {} +static inline void kfd_debugfs_add_process(struct kfd_process *p) {} +static inline void kfd_debugfs_remove_process(struct kfd_process *p) {} #endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index 7c0c24732481..722ac1662bdc 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -900,6 +900,8 @@ struct kfd_process *kfd_create_process(struct task_struct *thread) kfd_procfs_add_sysfs_files(process); kfd_procfs_add_sysfs_counters(process); + kfd_debugfs_add_process(process); + init_waitqueue_head(&process->wait_irq_drain); } out: @@ -1054,6 +1056,8 @@ static void kfd_process_destroy_pdds(struct kfd_process *p) for (i = 0; i < p->n_pdds; i++) { struct kfd_process_device *pdd = p->pdds[i]; + kfd_smi_event_process(pdd, false); + pr_debug("Releasing pdd (topology id %d, for pid %d)\n", pdd->dev->id, p->lead_thread->pid); kfd_process_device_destroy_cwsr_dgpu(pdd); @@ -1174,6 +1178,7 @@ static void kfd_process_wq_release(struct work_struct *work) dma_fence_signal(ef); kfd_process_remove_sysfs(p); + kfd_debugfs_remove_process(p); kfd_process_kunmap_signal_bo(p); kfd_process_free_outstanding_kfd_bos(p); @@ -1715,6 +1720,8 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd, pdd->pasid = avm->pasid; pdd->drm_file = drm_file; + kfd_smi_event_process(pdd, true); + return 0; err_get_pasid: diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c index 7eb370b68159..6d5fa57d4a23 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c @@ -451,8 +451,15 @@ int pqm_create_queue(struct process_queue_manager *pqm, } if (retval != 0) { - pr_err("process pid %d DQM create queue type %d failed. ret %d\n", - pqm->process->lead_thread->pid, type, retval); + if ((type == KFD_QUEUE_TYPE_SDMA || + type == KFD_QUEUE_TYPE_SDMA_XGMI || + type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) && + retval == -ENOMEM) + pr_warn("process pid %d DQM create queue type %d failed. ret %d\n", + pqm->process->lead_thread->pid, type, retval); + else + pr_err("process pid %d DQM create queue type %d failed. ret %d\n", + pqm->process->lead_thread->pid, type, retval); goto err_create_queue; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c index 4afff7094caf..a65c67cf56ff 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c @@ -402,7 +402,7 @@ static u32 kfd_get_vgpr_size_per_cu(u32 gfxv) { u32 vgpr_size = 0x40000; - if ((gfxv / 100 * 100) == 90400 || /* GFX_VERSION_AQUA_VANJARAM */ + if (gfxv == 90402 || /* GFX_VERSION_AQUA_VANJARAM */ gfxv == 90010 || /* GFX_VERSION_ALDEBARAN */ gfxv == 90008 || /* GFX_VERSION_ARCTURUS */ gfxv == 90500) @@ -462,7 +462,7 @@ void kfd_queue_ctx_save_restore_size(struct kfd_topology_device *dev) if (gfxv == 80002) /* GFX_VERSION_TONGA */ props->eop_buffer_size = 0x8000; - else if ((gfxv / 100 * 100) == 90400) /* GFX_VERSION_AQUA_VANJARAM */ + else if (gfxv == 90402) /* GFX_VERSION_AQUA_VANJARAM */ props->eop_buffer_size = 4096; else if (gfxv >= 80000) props->eop_buffer_size = 4096; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c index 9b8169761ec5..83d9384ac815 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c @@ -163,10 +163,9 @@ static int kfd_smi_ev_release(struct inode *inode, struct file *filep) static bool kfd_smi_ev_enabled(pid_t pid, struct kfd_smi_client *client, unsigned int event) { - uint64_t all = KFD_SMI_EVENT_MASK_FROM_INDEX(KFD_SMI_EVENT_ALL_PROCESS); uint64_t events = READ_ONCE(client->events); - if (pid && client->pid != pid && !(client->suser && (events & all))) + if (pid && client->pid != pid && !client->suser) return false; return events & KFD_SMI_EVENT_MASK_FROM_INDEX(event); @@ -345,6 +344,27 @@ void kfd_smi_event_unmap_from_gpu(struct kfd_node *node, pid_t pid, pid, address, last - address + 1, node->id, trigger)); } +void kfd_smi_event_process(struct kfd_process_device *pdd, bool start) +{ + struct amdgpu_task_info *task_info; + struct amdgpu_vm *avm; + + if (!pdd->drm_priv) + return; + + avm = drm_priv_to_vm(pdd->drm_priv); + task_info = amdgpu_vm_get_task_info_vm(avm); + + if (task_info) { + kfd_smi_event_add(0, pdd->dev, + start ? KFD_SMI_EVENT_PROCESS_START : + KFD_SMI_EVENT_PROCESS_END, + KFD_EVENT_FMT_PROCESS(task_info->pid, + task_info->task_name)); + amdgpu_vm_put_task_info(task_info); + } +} + int kfd_smi_event_open(struct kfd_node *dev, uint32_t *fd) { struct kfd_smi_client *client; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h index 503bff13d815..bb4d72b57387 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h @@ -53,4 +53,5 @@ void kfd_smi_event_queue_restore_rescheduled(struct mm_struct *mm); void kfd_smi_event_unmap_from_gpu(struct kfd_node *node, pid_t pid, unsigned long address, unsigned long last, uint32_t trigger); +void kfd_smi_event_process(struct kfd_process_device *pdd, bool start); #endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index 100717a98ec1..72be6e152e88 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -1245,8 +1245,7 @@ svm_range_get_pte_flags(struct kfd_node *node, case IP_VERSION(9, 4, 4): case IP_VERSION(9, 5, 0): if (ext_coherent) - mtype_local = (gc_ip_version < IP_VERSION(9, 5, 0) && !node->adev->rev_id) ? - AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_CC; + mtype_local = AMDGPU_VM_MTYPE_CC; else mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC : amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 9bbee484d57c..baa2374acdeb 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -1267,34 +1267,41 @@ static void kfd_set_recommended_sdma_engines(struct kfd_topology_device *to_dev, { struct kfd_node *gpu = outbound_link->gpu; struct amdgpu_device *adev = gpu->adev; - int num_xgmi_nodes = adev->gmc.xgmi.num_physical_nodes; + unsigned int num_xgmi_nodes = adev->gmc.xgmi.num_physical_nodes; + unsigned int num_xgmi_sdma_engines = kfd_get_num_xgmi_sdma_engines(gpu); + unsigned int num_sdma_engines = kfd_get_num_sdma_engines(gpu); + uint32_t sdma_eng_id_mask = (1 << num_sdma_engines) - 1; + uint32_t xgmi_sdma_eng_id_mask = + ((1 << num_xgmi_sdma_engines) - 1) << num_sdma_engines; + bool support_rec_eng = !amdgpu_sriov_vf(adev) && to_dev->gpu && adev->aid_mask && num_xgmi_nodes && gpu->kfd->num_nodes == 1 && - kfd_get_num_xgmi_sdma_engines(gpu) >= 14 && - (!(adev->flags & AMD_IS_APU) && num_xgmi_nodes == 8); + num_xgmi_sdma_engines >= 6 && (!(adev->flags & AMD_IS_APU) && + num_xgmi_nodes == 8); if (support_rec_eng) { int src_socket_id = adev->gmc.xgmi.physical_node_id; int dst_socket_id = to_dev->gpu->adev->gmc.xgmi.physical_node_id; + unsigned int reshift = num_xgmi_sdma_engines == 6 ? 1 : 0; outbound_link->rec_sdma_eng_id_mask = - 1 << rec_sdma_eng_map[src_socket_id][dst_socket_id]; + 1 << (rec_sdma_eng_map[src_socket_id][dst_socket_id] >> reshift); inbound_link->rec_sdma_eng_id_mask = - 1 << rec_sdma_eng_map[dst_socket_id][src_socket_id]; - } else { - int num_sdma_eng = kfd_get_num_sdma_engines(gpu); - int i, eng_offset = 0; + 1 << (rec_sdma_eng_map[dst_socket_id][src_socket_id] >> reshift); - if (outbound_link->iolink_type == CRAT_IOLINK_TYPE_XGMI && - kfd_get_num_xgmi_sdma_engines(gpu) && to_dev->gpu) { - eng_offset = num_sdma_eng; - num_sdma_eng = kfd_get_num_xgmi_sdma_engines(gpu); - } + /* If recommended engine is out of range, need to reset the mask */ + if (outbound_link->rec_sdma_eng_id_mask & sdma_eng_id_mask) + outbound_link->rec_sdma_eng_id_mask = xgmi_sdma_eng_id_mask; + if (inbound_link->rec_sdma_eng_id_mask & sdma_eng_id_mask) + inbound_link->rec_sdma_eng_id_mask = xgmi_sdma_eng_id_mask; - for (i = 0; i < num_sdma_eng; i++) { - outbound_link->rec_sdma_eng_id_mask |= (1 << (i + eng_offset)); - inbound_link->rec_sdma_eng_id_mask |= (1 << (i + eng_offset)); - } + } else { + uint32_t engine_mask = (outbound_link->iolink_type == CRAT_IOLINK_TYPE_XGMI && + num_xgmi_sdma_engines && to_dev->gpu) ? xgmi_sdma_eng_id_mask : + sdma_eng_id_mask; + + outbound_link->rec_sdma_eng_id_mask = engine_mask; + inbound_link->rec_sdma_eng_id_mask = engine_mask; } } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile b/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile index ab2a97e354da..7329b8cc2576 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile @@ -38,6 +38,7 @@ AMDGPUDM = \ amdgpu_dm_pp_smu.o \ amdgpu_dm_psr.o \ amdgpu_dm_replay.o \ + amdgpu_dm_quirks.o \ amdgpu_dm_wb.o ifdef CONFIG_DRM_AMD_DC_FP diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 536f73131c2d..4b4e9241619f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -80,7 +80,6 @@ #include <linux/power_supply.h> #include <linux/firmware.h> #include <linux/component.h> -#include <linux/dmi.h> #include <linux/sort.h> #include <drm/display/drm_dp_mst_helper.h> @@ -115,6 +114,8 @@ #include "modules/inc/mod_freesync.h" #include "modules/power/power_helpers.h" +static_assert(AMDGPU_DMUB_NOTIFICATION_MAX == DMUB_NOTIFICATION_MAX, "AMDGPU_DMUB_NOTIFICATION_MAX mismatch"); + #define FIRMWARE_RENOIR_DMUB "amdgpu/renoir_dmcub.bin" MODULE_FIRMWARE(FIRMWARE_RENOIR_DMUB); #define FIRMWARE_SIENNA_CICHLID_DMUB "amdgpu/sienna_cichlid_dmcub.bin" @@ -280,7 +281,7 @@ static u32 dm_vblank_get_counter(struct amdgpu_device *adev, int crtc) acrtc = adev->mode_info.crtcs[crtc]; if (!acrtc->dm_irq_params.stream) { - DRM_ERROR("dc_stream_state is NULL for crtc '%d'!\n", + drm_err(adev_to_drm(adev), "dc_stream_state is NULL for crtc '%d'!\n", crtc); return 0; } @@ -301,7 +302,7 @@ static int dm_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc, acrtc = adev->mode_info.crtcs[crtc]; if (!acrtc->dm_irq_params.stream) { - DRM_ERROR("dc_stream_state is NULL for crtc '%d'!\n", + drm_err(adev_to_drm(adev), "dc_stream_state is NULL for crtc '%d'!\n", crtc); return 0; } @@ -372,6 +373,8 @@ get_crtc_by_otg_inst(struct amdgpu_device *adev, static inline bool is_dc_timing_adjust_needed(struct dm_crtc_state *old_state, struct dm_crtc_state *new_state) { + if (new_state->stream->adjust.timing_adjust_pending) + return true; if (new_state->freesync_config.state == VRR_STATE_ACTIVE_FIXED) return true; else if (amdgpu_dm_crtc_vrr_active(old_state) != amdgpu_dm_crtc_vrr_active(new_state)) @@ -673,15 +676,21 @@ static void dm_crtc_high_irq(void *interrupt_params) spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags); if (acrtc->dm_irq_params.stream && - acrtc->dm_irq_params.vrr_params.supported && - acrtc->dm_irq_params.freesync_config.state == - VRR_STATE_ACTIVE_VARIABLE) { + acrtc->dm_irq_params.vrr_params.supported) { + bool replay_en = acrtc->dm_irq_params.stream->link->replay_settings.replay_feature_enabled; + bool psr_en = acrtc->dm_irq_params.stream->link->psr_settings.psr_feature_enabled; + bool fs_active_var_en = acrtc->dm_irq_params.freesync_config.state == VRR_STATE_ACTIVE_VARIABLE; + mod_freesync_handle_v_update(adev->dm.freesync_module, acrtc->dm_irq_params.stream, &acrtc->dm_irq_params.vrr_params); - dc_stream_adjust_vmin_vmax(adev->dm.dc, acrtc->dm_irq_params.stream, - &acrtc->dm_irq_params.vrr_params.adjust); + /* update vmin_vmax only if freesync is enabled, or only if PSR and REPLAY are disabled */ + if (fs_active_var_en || (!fs_active_var_en && !replay_en && !psr_en)) { + dc_stream_adjust_vmin_vmax(adev->dm.dc, + acrtc->dm_irq_params.stream, + &acrtc->dm_irq_params.vrr_params.adjust); + } } /* @@ -749,6 +758,29 @@ static void dmub_aux_setconfig_callback(struct amdgpu_device *adev, complete(&adev->dm.dmub_aux_transfer_done); } +static void dmub_aux_fused_io_callback(struct amdgpu_device *adev, + struct dmub_notification *notify) +{ + if (!adev || !notify) { + ASSERT(false); + return; + } + + const struct dmub_cmd_fused_request *req = ¬ify->fused_request; + const uint8_t ddc_line = req->u.aux.ddc_line; + + if (ddc_line >= ARRAY_SIZE(adev->dm.fused_io)) { + ASSERT(false); + return; + } + + struct fused_io_sync *sync = &adev->dm.fused_io[ddc_line]; + + static_assert(sizeof(*req) <= sizeof(sync->reply_data), "Size mismatch"); + memcpy(sync->reply_data, req, sizeof(*req)); + complete(&sync->replied); +} + /** * dmub_hpd_callback - DMUB HPD interrupt processing callback. * @adev: amdgpu_device pointer @@ -772,18 +804,18 @@ static void dmub_hpd_callback(struct amdgpu_device *adev, return; if (notify == NULL) { - DRM_ERROR("DMUB HPD callback notification was NULL"); + drm_err(adev_to_drm(adev), "DMUB HPD callback notification was NULL"); return; } if (notify->link_index > adev->dm.dc->link_count) { - DRM_ERROR("DMUB HPD index (%u)is abnormal", notify->link_index); + drm_err(adev_to_drm(adev), "DMUB HPD index (%u)is abnormal", notify->link_index); return; } /* Skip DMUB HPD IRQ in suspend/resume. We will probe them later. */ if (notify->type == DMUB_NOTIFICATION_HPD && adev->in_suspend) { - DRM_INFO("Skip DMUB HPD IRQ callback in suspend/resume\n"); + drm_info(adev_to_drm(adev), "Skip DMUB HPD IRQ callback in suspend/resume\n"); return; } @@ -800,11 +832,11 @@ static void dmub_hpd_callback(struct amdgpu_device *adev, aconnector = to_amdgpu_dm_connector(connector); if (link && aconnector->dc_link == link) { if (notify->type == DMUB_NOTIFICATION_HPD) - DRM_INFO("DMUB HPD IRQ callback: link_index=%u\n", link_index); + drm_info(adev_to_drm(adev), "DMUB HPD IRQ callback: link_index=%u\n", link_index); else if (notify->type == DMUB_NOTIFICATION_HPD_IRQ) - DRM_INFO("DMUB HPD RX IRQ callback: link_index=%u\n", link_index); + drm_info(adev_to_drm(adev), "DMUB HPD RX IRQ callback: link_index=%u\n", link_index); else - DRM_WARN("DMUB Unknown HPD callback type %d, link_index=%u\n", + drm_warn(adev_to_drm(adev), "DMUB Unknown HPD callback type %d, link_index=%u\n", notify->type, link_index); hpd_aconnector = aconnector; @@ -816,7 +848,7 @@ static void dmub_hpd_callback(struct amdgpu_device *adev, if (hpd_aconnector) { if (notify->type == DMUB_NOTIFICATION_HPD) { if (hpd_aconnector->dc_link->hpd_status == (notify->hpd_status == DP_HPD_PLUG)) - DRM_WARN("DMUB reported hpd status unchanged. link_index=%u\n", link_index); + drm_warn(adev_to_drm(adev), "DMUB reported hpd status unchanged. link_index=%u\n", link_index); handle_hpd_irq_helper(hpd_aconnector); } else if (notify->type == DMUB_NOTIFICATION_HPD_IRQ) { handle_hpd_rx_irq(hpd_aconnector); @@ -835,7 +867,7 @@ static void dmub_hpd_callback(struct amdgpu_device *adev, static void dmub_hpd_sense_callback(struct amdgpu_device *adev, struct dmub_notification *notify) { - DRM_DEBUG_DRIVER("DMUB HPD SENSE callback.\n"); + drm_dbg_driver(adev_to_drm(adev), "DMUB HPD SENSE callback.\n"); } /** @@ -871,7 +903,7 @@ static void dm_handle_hpd_work(struct work_struct *work) dmub_hpd_wrk = container_of(work, struct dmub_hpd_work, handle_hpd_work); if (!dmub_hpd_wrk->dmub_notify) { - DRM_ERROR("dmub_hpd_wrk dmub_notify is NULL"); + drm_err(adev_to_drm(dmub_hpd_wrk->adev), "dmub_hpd_wrk dmub_notify is NULL"); return; } @@ -885,6 +917,30 @@ static void dm_handle_hpd_work(struct work_struct *work) } +static const char *dmub_notification_type_str(enum dmub_notification_type e) +{ + switch (e) { + case DMUB_NOTIFICATION_NO_DATA: + return "NO_DATA"; + case DMUB_NOTIFICATION_AUX_REPLY: + return "AUX_REPLY"; + case DMUB_NOTIFICATION_HPD: + return "HPD"; + case DMUB_NOTIFICATION_HPD_IRQ: + return "HPD_IRQ"; + case DMUB_NOTIFICATION_SET_CONFIG_REPLY: + return "SET_CONFIG_REPLY"; + case DMUB_NOTIFICATION_DPIA_NOTIFICATION: + return "DPIA_NOTIFICATION"; + case DMUB_NOTIFICATION_HPD_SENSE_NOTIFY: + return "HPD_SENSE_NOTIFY"; + case DMUB_NOTIFICATION_FUSED_IO: + return "FUSED_IO"; + default: + return "<unknown>"; + } +} + #define DMUB_TRACE_MAX_READ 64 /** * dm_dmub_outbox1_low_irq() - Handles Outbox interrupt @@ -902,22 +958,13 @@ static void dm_dmub_outbox1_low_irq(void *interrupt_params) struct dmcub_trace_buf_entry entry = { 0 }; u32 count = 0; struct dmub_hpd_work *dmub_hpd_wrk; - static const char *const event_type[] = { - "NO_DATA", - "AUX_REPLY", - "HPD", - "HPD_IRQ", - "SET_CONFIGC_REPLY", - "DPIA_NOTIFICATION", - "HPD_SENSE_NOTIFY", - }; do { if (dc_dmub_srv_get_dmub_outbox0_msg(dm->dc, &entry)) { trace_amdgpu_dmub_trace_high_irq(entry.trace_code, entry.tick_count, entry.param0, entry.param1); - DRM_DEBUG_DRIVER("trace_code:%u, tick_count:%u, param0:%u, param1:%u\n", + drm_dbg_driver(adev_to_drm(adev), "trace_code:%u, tick_count:%u, param0:%u, param1:%u\n", entry.trace_code, entry.tick_count, entry.param0, entry.param1); } else break; @@ -927,7 +974,7 @@ static void dm_dmub_outbox1_low_irq(void *interrupt_params) } while (count <= DMUB_TRACE_MAX_READ); if (count > DMUB_TRACE_MAX_READ) - DRM_DEBUG_DRIVER("Warning : count > DMUB_TRACE_MAX_READ"); + drm_dbg_driver(adev_to_drm(adev), "Warning : count > DMUB_TRACE_MAX_READ"); if (dc_enable_dmub_notifications(adev->dm.dc) && irq_params->irq_src == DC_IRQ_SOURCE_DMCUB_OUTBOX) { @@ -935,25 +982,25 @@ static void dm_dmub_outbox1_low_irq(void *interrupt_params) do { dc_stat_get_dmub_notification(adev->dm.dc, ¬ify); if (notify.type >= ARRAY_SIZE(dm->dmub_thread_offload)) { - DRM_ERROR("DM: notify type %d invalid!", notify.type); + drm_err(adev_to_drm(adev), "DM: notify type %d invalid!", notify.type); continue; } if (!dm->dmub_callback[notify.type]) { - DRM_WARN("DMUB notification skipped due to no handler: type=%s\n", - event_type[notify.type]); + drm_warn(adev_to_drm(adev), "DMUB notification skipped due to no handler: type=%s\n", + dmub_notification_type_str(notify.type)); continue; } if (dm->dmub_thread_offload[notify.type] == true) { dmub_hpd_wrk = kzalloc(sizeof(*dmub_hpd_wrk), GFP_ATOMIC); if (!dmub_hpd_wrk) { - DRM_ERROR("Failed to allocate dmub_hpd_wrk"); + drm_err(adev_to_drm(adev), "Failed to allocate dmub_hpd_wrk"); return; } dmub_hpd_wrk->dmub_notify = kmemdup(¬ify, sizeof(struct dmub_notification), GFP_ATOMIC); if (!dmub_hpd_wrk->dmub_notify) { kfree(dmub_hpd_wrk); - DRM_ERROR("Failed to allocate dmub_hpd_wrk->dmub_notify"); + drm_err(adev_to_drm(adev), "Failed to allocate dmub_hpd_wrk->dmub_notify"); return; } INIT_WORK(&dmub_hpd_wrk->handle_hpd_work, dm_handle_hpd_work); @@ -1011,10 +1058,10 @@ static void amdgpu_dm_fbc_init(struct drm_connector *connector) &compressor->gpu_addr, &compressor->cpu_addr); if (r) - DRM_ERROR("DM: Failed to initialize FBC\n"); + drm_err(adev_to_drm(adev), "DM: Failed to initialize FBC\n"); else { adev->dm.dc->ctx->fbc_gpu_addr = compressor->gpu_addr; - DRM_INFO("DM: FBC alloc %lu\n", max_size*4); + drm_info(adev_to_drm(adev), "DM: FBC alloc %lu\n", max_size*4); } } @@ -1179,13 +1226,13 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev) return 0; if (!fb_info) { - DRM_ERROR("No framebuffer info for DMUB service.\n"); + drm_err(adev_to_drm(adev), "No framebuffer info for DMUB service.\n"); return -EINVAL; } if (!dmub_fw) { /* Firmware required for DMUB support. */ - DRM_ERROR("No firmware provided for DMUB.\n"); + drm_err(adev_to_drm(adev), "No firmware provided for DMUB.\n"); return -EINVAL; } @@ -1195,19 +1242,19 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev) status = dmub_srv_has_hw_support(dmub_srv, &has_hw_support); if (status != DMUB_STATUS_OK) { - DRM_ERROR("Error checking HW support for DMUB: %d\n", status); + drm_err(adev_to_drm(adev), "Error checking HW support for DMUB: %d\n", status); return -EINVAL; } if (!has_hw_support) { - DRM_INFO("DMUB unsupported on ASIC\n"); + drm_info(adev_to_drm(adev), "DMUB unsupported on ASIC\n"); return 0; } /* Reset DMCUB if it was previously running - before we overwrite its memory. */ status = dmub_srv_hw_reset(dmub_srv); if (status != DMUB_STATUS_OK) - DRM_WARN("Error resetting DMUB HW: %d\n", status); + drm_warn(adev_to_drm(adev), "Error resetting DMUB HW: %d\n", status); hdr = (const struct dmcub_firmware_header_v1_0 *)dmub_fw->data; @@ -1290,6 +1337,7 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev) case IP_VERSION(3, 5, 1): case IP_VERSION(3, 6, 0): hw_params.ips_sequential_ono = adev->external_rev_id > 0x10; + hw_params.lower_hbr3_phy_ssc = true; break; default: break; @@ -1297,14 +1345,14 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev) status = dmub_srv_hw_init(dmub_srv, &hw_params); if (status != DMUB_STATUS_OK) { - DRM_ERROR("Error initializing DMUB HW: %d\n", status); + drm_err(adev_to_drm(adev), "Error initializing DMUB HW: %d\n", status); return -EINVAL; } /* Wait for firmware load to finish. */ status = dmub_srv_wait_for_auto_load(dmub_srv, 100000); if (status != DMUB_STATUS_OK) - DRM_WARN("Wait for DMUB auto-load failed: %d\n", status); + drm_warn(adev_to_drm(adev), "Wait for DMUB auto-load failed: %d\n", status); /* Init DMCU and ABM if available. */ if (dmcu && abm) { @@ -1315,11 +1363,11 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev) if (!adev->dm.dc->ctx->dmub_srv) adev->dm.dc->ctx->dmub_srv = dc_dmub_srv_create(adev->dm.dc, dmub_srv); if (!adev->dm.dc->ctx->dmub_srv) { - DRM_ERROR("Couldn't allocate DC DMUB server!\n"); + drm_err(adev_to_drm(adev), "Couldn't allocate DC DMUB server!\n"); return -ENOMEM; } - DRM_INFO("DMUB hardware initialized: version=0x%08X\n", + drm_info(adev_to_drm(adev), "DMUB hardware initialized: version=0x%08X\n", adev->dm.dmcub_fw_version); /* Keeping sanity checks off if @@ -1362,18 +1410,18 @@ static void dm_dmub_hw_resume(struct amdgpu_device *adev) status = dmub_srv_is_hw_init(dmub_srv, &init); if (status != DMUB_STATUS_OK) - DRM_WARN("DMUB hardware init check failed: %d\n", status); + drm_warn(adev_to_drm(adev), "DMUB hardware init check failed: %d\n", status); if (status == DMUB_STATUS_OK && init) { /* Wait for firmware load to finish. */ status = dmub_srv_wait_for_auto_load(dmub_srv, 100000); if (status != DMUB_STATUS_OK) - DRM_WARN("Wait for DMUB auto-load failed: %d\n", status); + drm_warn(adev_to_drm(adev), "Wait for DMUB auto-load failed: %d\n", status); } else { /* Perform the full hardware initialization. */ r = dm_dmub_hw_init(adev); if (r) - DRM_ERROR("DMUB interface failed to initialize: status=%d\n", r); + drm_err(adev_to_drm(adev), "DMUB interface failed to initialize: status=%d\n", r); } } @@ -1483,18 +1531,18 @@ static void dm_handle_hpd_rx_offload_work(struct work_struct *work) offload_work = container_of(work, struct hpd_rx_irq_offload_work, work); aconnector = offload_work->offload_wq->aconnector; + adev = offload_work->adev; if (!aconnector) { - DRM_ERROR("Can't retrieve aconnector in hpd_rx_irq_offload_work"); + drm_err(adev_to_drm(adev), "Can't retrieve aconnector in hpd_rx_irq_offload_work"); goto skip; } - adev = drm_to_adev(aconnector->base.dev); dc_link = aconnector->dc_link; mutex_lock(&aconnector->hpd_lock); if (!dc_link_detect_connection_type(dc_link, &new_connection_type)) - DRM_ERROR("KMS: Failed to detect connector\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n"); mutex_unlock(&aconnector->hpd_lock); if (new_connection_type == dc_connection_none) @@ -1563,8 +1611,9 @@ skip: } -static struct hpd_rx_irq_offload_work_queue *hpd_rx_irq_create_workqueue(struct dc *dc) +static struct hpd_rx_irq_offload_work_queue *hpd_rx_irq_create_workqueue(struct amdgpu_device *adev) { + struct dc *dc = adev->dm.dc; int max_caps = dc->caps.max_links; int i = 0; struct hpd_rx_irq_offload_work_queue *hpd_rx_offload_wq = NULL; @@ -1580,7 +1629,7 @@ static struct hpd_rx_irq_offload_work_queue *hpd_rx_irq_create_workqueue(struct create_singlethread_workqueue("amdgpu_dm_hpd_rx_offload_wq"); if (hpd_rx_offload_wq[i].wq == NULL) { - DRM_ERROR("create amdgpu_dm_hpd_rx_offload_wq fail!"); + drm_err(adev_to_drm(adev), "create amdgpu_dm_hpd_rx_offload_wq fail!"); goto out_err; } @@ -1629,153 +1678,6 @@ static bool dm_should_disable_stutter(struct pci_dev *pdev) return false; } -struct amdgpu_dm_quirks { - bool aux_hpd_discon; - bool support_edp0_on_dp1; -}; - -static struct amdgpu_dm_quirks quirk_entries = { - .aux_hpd_discon = false, - .support_edp0_on_dp1 = false -}; - -static int edp0_on_dp1_callback(const struct dmi_system_id *id) -{ - quirk_entries.support_edp0_on_dp1 = true; - return 0; -} - -static int aux_hpd_discon_callback(const struct dmi_system_id *id) -{ - quirk_entries.aux_hpd_discon = true; - return 0; -} - -static const struct dmi_system_id dmi_quirk_table[] = { - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3660"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3260"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3460"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Tower Plus 7010"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Tower 7010"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex SFF Plus 7010"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex SFF 7010"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Micro Plus 7010"), - }, - }, - { - .callback = aux_hpd_discon_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), - DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Micro 7010"), - }, - }, - { - .callback = edp0_on_dp1_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "HP Elite mt645 G8 Mobile Thin Client"), - }, - }, - { - .callback = edp0_on_dp1_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 645 14 inch G11 Notebook PC"), - }, - }, - { - .callback = edp0_on_dp1_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 665 16 inch G11 Notebook PC"), - }, - }, - { - .callback = edp0_on_dp1_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 445 14 inch G11 Notebook PC"), - }, - }, - { - .callback = edp0_on_dp1_callback, - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 465 16 inch G11 Notebook PC"), - }, - }, - {} - /* TODO: refactor this from a fixed table to a dynamic option */ -}; - -static void retrieve_dmi_info(struct amdgpu_display_manager *dm, struct dc_init_data *init_data) -{ - int dmi_id; - struct drm_device *dev = dm->ddev; - - dm->aux_hpd_discon_quirk = false; - init_data->flags.support_edp0_on_dp1 = false; - - dmi_id = dmi_check_system(dmi_quirk_table); - - if (!dmi_id) - return; - - if (quirk_entries.aux_hpd_discon) { - dm->aux_hpd_discon_quirk = true; - drm_info(dev, "aux_hpd_discon_quirk attached\n"); - } - if (quirk_entries.support_edp0_on_dp1) { - init_data->flags.support_edp0_on_dp1 = true; - drm_info(dev, "support_edp0_on_dp1 attached\n"); - } -} void* dm_allocate_gpu_mem( @@ -1951,7 +1853,7 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) mutex_init(&adev->dm.audio_lock); if (amdgpu_dm_irq_init(adev)) { - DRM_ERROR("amdgpu: failed to initialize DM IRQ support.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize DM IRQ support.\n"); goto error; } @@ -2062,7 +1964,9 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 0, 0)) init_data.num_virtual_links = 1; - retrieve_dmi_info(&adev->dm, &init_data); + retrieve_dmi_info(&adev->dm); + if (adev->dm.edp0_on_dp1_quirk) + init_data.flags.support_edp0_on_dp1 = true; if (adev->dm.bb_from_dmub) init_data.bb_from_dmub = adev->dm.bb_from_dmub; @@ -2073,10 +1977,10 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev->dm.dc = dc_create(&init_data); if (adev->dm.dc) { - DRM_INFO("Display Core v%s initialized on %s\n", DC_VER, + drm_info(adev_to_drm(adev), "Display Core v%s initialized on %s\n", DC_VER, dce_version_to_string(adev->dm.dc->ctx->dce_version)); } else { - DRM_INFO("Display Core failed to initialize with v%s!\n", DC_VER); + drm_info(adev_to_drm(adev), "Display Core failed to initialize with v%s!\n", DC_VER); goto error; } @@ -2110,25 +2014,31 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev->dm.dc->debug.using_dml21 = true; } + if (amdgpu_dc_debug_mask & DC_HDCP_LC_FORCE_FW_ENABLE) + adev->dm.dc->debug.hdcp_lc_force_fw_enable = true; + + if (amdgpu_dc_debug_mask & DC_HDCP_LC_ENABLE_SW_FALLBACK) + adev->dm.dc->debug.hdcp_lc_enable_sw_fallback = true; + adev->dm.dc->debug.visual_confirm = amdgpu_dc_visual_confirm; /* TODO: Remove after DP2 receiver gets proper support of Cable ID feature */ adev->dm.dc->debug.ignore_cable_id = true; if (adev->dm.dc->caps.dp_hdmi21_pcon_support) - DRM_INFO("DP-HDMI FRL PCON supported\n"); + drm_info(adev_to_drm(adev), "DP-HDMI FRL PCON supported\n"); r = dm_dmub_hw_init(adev); if (r) { - DRM_ERROR("DMUB interface failed to initialize: status=%d\n", r); + drm_err(adev_to_drm(adev), "DMUB interface failed to initialize: status=%d\n", r); goto error; } dc_hardware_init(adev->dm.dc); - adev->dm.hpd_rx_offload_wq = hpd_rx_irq_create_workqueue(adev->dm.dc); + adev->dm.hpd_rx_offload_wq = hpd_rx_irq_create_workqueue(adev); if (!adev->dm.hpd_rx_offload_wq) { - DRM_ERROR("amdgpu: failed to create hpd rx offload workqueue.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to create hpd rx offload workqueue.\n"); goto error; } @@ -2143,10 +2053,10 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev->dm.freesync_module = mod_freesync_create(adev->dm.dc); if (!adev->dm.freesync_module) { - DRM_ERROR( + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize freesync_module.\n"); } else - DRM_DEBUG_DRIVER("amdgpu: freesync_module init done %p.\n", + drm_dbg_driver(adev_to_drm(adev), "amdgpu: freesync_module init done %p.\n", adev->dm.freesync_module); amdgpu_dm_init_color_mod(); @@ -2155,7 +2065,7 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev->dm.vblank_control_workqueue = create_singlethread_workqueue("dm_vblank_control_workqueue"); if (!adev->dm.vblank_control_workqueue) - DRM_ERROR("amdgpu: failed to initialize vblank_workqueue.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize vblank_workqueue.\n"); } if (adev->dm.dc->caps.ips_support && @@ -2166,9 +2076,9 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev->dm.hdcp_workqueue = hdcp_create_workqueue(adev, &init_params.cp_psp, adev->dm.dc); if (!adev->dm.hdcp_workqueue) - DRM_ERROR("amdgpu: failed to initialize hdcp_workqueue.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize hdcp_workqueue.\n"); else - DRM_DEBUG_DRIVER("amdgpu: hdcp_workqueue init done %p.\n", adev->dm.hdcp_workqueue); + drm_dbg_driver(adev_to_drm(adev), "amdgpu: hdcp_workqueue init done %p.\n", adev->dm.hdcp_workqueue); dc_init_callbacks(adev->dm.dc, &init_params); } @@ -2176,20 +2086,29 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) init_completion(&adev->dm.dmub_aux_transfer_done); adev->dm.dmub_notify = kzalloc(sizeof(struct dmub_notification), GFP_KERNEL); if (!adev->dm.dmub_notify) { - DRM_INFO("amdgpu: fail to allocate adev->dm.dmub_notify"); + drm_info(adev_to_drm(adev), "amdgpu: fail to allocate adev->dm.dmub_notify"); goto error; } adev->dm.delayed_hpd_wq = create_singlethread_workqueue("amdgpu_dm_hpd_wq"); if (!adev->dm.delayed_hpd_wq) { - DRM_ERROR("amdgpu: failed to create hpd offload workqueue.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to create hpd offload workqueue.\n"); goto error; } amdgpu_dm_outbox_init(adev); if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_AUX_REPLY, dmub_aux_setconfig_callback, false)) { - DRM_ERROR("amdgpu: fail to register dmub aux callback"); + drm_err(adev_to_drm(adev), "amdgpu: fail to register dmub aux callback"); + goto error; + } + + for (size_t i = 0; i < ARRAY_SIZE(adev->dm.fused_io); i++) + init_completion(&adev->dm.fused_io[i].replied); + + if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_FUSED_IO, + dmub_aux_fused_io_callback, false)) { + drm_err(adev_to_drm(adev), "amdgpu: fail to register dmub fused io callback"); goto error; } /* Enable outbox notification only after IRQ handlers are registered and DMUB is alive. @@ -2206,7 +2125,7 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) } if (amdgpu_dm_initialize_drm_device(adev)) { - DRM_ERROR( + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize sw for display support.\n"); goto error; } @@ -2221,7 +2140,7 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev_to_drm(adev)->mode_config.cursor_height = adev->dm.dc->caps.max_cursor_size; if (drm_vblank_init(adev_to_drm(adev), adev->dm.display_indexes_num)) { - DRM_ERROR( + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize sw for display support.\n"); goto error; } @@ -2229,14 +2148,14 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) amdgpu_dm_crtc_secure_display_create_contexts(adev); if (!adev->dm.secure_display_ctx.crtc_ctx) - DRM_ERROR("amdgpu: failed to initialize secure display contexts.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to initialize secure display contexts.\n"); if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(4, 0, 1)) adev->dm.secure_display_ctx.support_mul_roi = true; #endif - DRM_DEBUG_DRIVER("KMS initialized.\n"); + drm_dbg_driver(adev_to_drm(adev), "KMS initialized.\n"); return 0; error: @@ -2409,7 +2328,7 @@ static int load_dmcu_fw(struct amdgpu_device *adev) default: break; } - DRM_ERROR("Unsupported ASIC type: 0x%X\n", adev->asic_type); + drm_err(adev_to_drm(adev), "Unsupported ASIC type: 0x%X\n", adev->asic_type); return -EINVAL; } @@ -2427,7 +2346,7 @@ static int load_dmcu_fw(struct amdgpu_device *adev) return 0; } if (r) { - dev_err(adev->dev, "amdgpu_dm: Can't validate firmware \"%s\"\n", + drm_err(adev_to_drm(adev), "amdgpu_dm: Can't validate firmware \"%s\"\n", fw_name_dmcu); amdgpu_ucode_release(&adev->dm.fw_dmcu); return r; @@ -2552,7 +2471,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) adev->firmware.fw_size += ALIGN(le32_to_cpu(hdr->inst_const_bytes), PAGE_SIZE); - DRM_INFO("Loading DMUB firmware via PSP: version=0x%08X\n", + drm_info(adev_to_drm(adev), "Loading DMUB firmware via PSP: version=0x%08X\n", adev->dm.dmcub_fw_version); } @@ -2561,7 +2480,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) dmub_srv = adev->dm.dmub_srv; if (!dmub_srv) { - DRM_ERROR("Failed to allocate DMUB service!\n"); + drm_err(adev_to_drm(adev), "Failed to allocate DMUB service!\n"); return -ENOMEM; } @@ -2574,7 +2493,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) /* Create the DMUB service. */ status = dmub_srv_create(dmub_srv, &create_params); if (status != DMUB_STATUS_OK) { - DRM_ERROR("Error creating DMUB service: %d\n", status); + drm_err(adev_to_drm(adev), "Error creating DMUB service: %d\n", status); return -EINVAL; } @@ -2599,7 +2518,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) ®ion_info); if (status != DMUB_STATUS_OK) { - DRM_ERROR("Error calculating DMUB region info: %d\n", status); + drm_err(adev_to_drm(adev), "Error calculating DMUB region info: %d\n", status); return -EINVAL; } @@ -2628,14 +2547,14 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) fb_info = adev->dm.dmub_fb_info; if (!fb_info) { - DRM_ERROR( + drm_err(adev_to_drm(adev), "Failed to allocate framebuffer info for DMUB service!\n"); return -ENOMEM; } status = dmub_srv_calc_mem_info(dmub_srv, &memory_params, fb_info); if (status != DMUB_STATUS_OK) { - DRM_ERROR("Error calculating DMUB FB info: %d\n", status); + drm_err(adev_to_drm(adev), "Error calculating DMUB FB info: %d\n", status); return -EINVAL; } @@ -2652,7 +2571,7 @@ static int dm_sw_init(struct amdgpu_ip_block *ip_block) adev->dm.cgs_device = amdgpu_cgs_create_device(adev); if (!adev->dm.cgs_device) { - DRM_ERROR("amdgpu: failed to create cgs device.\n"); + drm_err(adev_to_drm(adev), "amdgpu: failed to create cgs device.\n"); return -EINVAL; } @@ -2958,7 +2877,7 @@ static int amdgpu_dm_smu_write_watermarks_table(struct amdgpu_device *adev) ret = amdgpu_dpm_write_watermarks_table(adev); if (ret) { - DRM_ERROR("Failed to update WMTABLE!\n"); + drm_err(adev_to_drm(adev), "Failed to update WMTABLE!\n"); return ret; } @@ -2976,13 +2895,13 @@ static int dm_oem_i2c_hw_init(struct amdgpu_device *adev) if (oem_ddc_service) { oem_i2c = create_i2c(oem_ddc_service, true); if (!oem_i2c) { - dev_info(adev->dev, "Failed to create oem i2c adapter data\n"); + drm_info(adev_to_drm(adev), "Failed to create oem i2c adapter data\n"); return -ENOMEM; } r = i2c_add_adapter(&oem_i2c->base); if (r) { - dev_info(adev->dev, "Failed to register oem i2c\n"); + drm_info(adev_to_drm(adev), "Failed to register oem i2c\n"); kfree(oem_i2c); return r; } @@ -3025,7 +2944,7 @@ static int dm_hw_init(struct amdgpu_ip_block *ip_block) r = dm_oem_i2c_hw_init(adev); if (r) - dev_info(adev->dev, "Failed to add OEM i2c bus\n"); + drm_info(adev_to_drm(adev), "Failed to add OEM i2c bus\n"); return 0; } @@ -3068,7 +2987,7 @@ static void dm_gpureset_toggle_interrupts(struct amdgpu_device *adev, irq_source = IRQ_TYPE_PFLIP + acrtc->otg_inst; rc = dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY; if (rc) - DRM_WARN("Failed to %s pflip interrupts\n", + drm_warn(adev_to_drm(adev), "Failed to %s pflip interrupts\n", enable ? "enable" : "disable"); if (enable) { @@ -3078,14 +2997,14 @@ static void dm_gpureset_toggle_interrupts(struct amdgpu_device *adev, rc = amdgpu_dm_crtc_set_vupdate_irq(&acrtc->base, false); if (rc) - DRM_WARN("Failed to %sable vupdate interrupt\n", enable ? "en" : "dis"); + drm_warn(adev_to_drm(adev), "Failed to %sable vupdate interrupt\n", enable ? "en" : "dis"); irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst; /* During gpu-reset we disable and then enable vblank irq, so * don't use amdgpu_irq_get/put() to avoid refcount change. */ if (!dc_interrupt_set(adev->dm.dc, irq_source, enable)) - DRM_WARN("Failed to %sable vblank interrupt\n", enable ? "en" : "dis"); + drm_warn(adev_to_drm(adev), "Failed to %sable vblank interrupt\n", enable ? "en" : "dis"); } } @@ -3920,20 +3839,21 @@ static void handle_hpd_irq(void *param) } -static void schedule_hpd_rx_offload_work(struct hpd_rx_irq_offload_work_queue *offload_wq, +static void schedule_hpd_rx_offload_work(struct amdgpu_device *adev, struct hpd_rx_irq_offload_work_queue *offload_wq, union hpd_irq_data hpd_irq_data) { struct hpd_rx_irq_offload_work *offload_work = kzalloc(sizeof(*offload_work), GFP_KERNEL); if (!offload_work) { - DRM_ERROR("Failed to allocate hpd_rx_irq_offload_work.\n"); + drm_err(adev_to_drm(adev), "Failed to allocate hpd_rx_irq_offload_work.\n"); return; } INIT_WORK(&offload_work->work, dm_handle_hpd_rx_offload_work); offload_work->data = hpd_irq_data; offload_work->offload_wq = offload_wq; + offload_work->adev = adev; queue_work(offload_wq->wq, &offload_work->work); DRM_DEBUG_KMS("queue work to handle hpd_rx offload work"); @@ -3975,7 +3895,7 @@ static void handle_hpd_rx_irq(void *param) goto out; if (hpd_irq_data.bytes.device_service_irq.bits.AUTOMATED_TEST) { - schedule_hpd_rx_offload_work(offload_wq, hpd_irq_data); + schedule_hpd_rx_offload_work(adev, offload_wq, hpd_irq_data); goto out; } @@ -3997,7 +3917,7 @@ static void handle_hpd_rx_irq(void *param) spin_unlock(&offload_wq->offload_lock); if (!skip) - schedule_hpd_rx_offload_work(offload_wq, hpd_irq_data); + schedule_hpd_rx_offload_work(adev, offload_wq, hpd_irq_data); goto out; } @@ -4014,7 +3934,7 @@ static void handle_hpd_rx_irq(void *param) spin_unlock(&offload_wq->offload_lock); if (!skip) - schedule_hpd_rx_offload_work(offload_wq, hpd_irq_data); + schedule_hpd_rx_offload_work(adev, offload_wq, hpd_irq_data); goto out; } @@ -4024,7 +3944,7 @@ out: if (result && !is_mst_root_connector) { /* Downstream Port status changed. */ if (!dc_link_detect_connection_type(dc_link, &new_connection_type)) - DRM_ERROR("KMS: Failed to detect connector\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n"); if (aconnector->base.force && new_connection_type == dc_connection_none) { emulated_link_detect(dc_link); @@ -4087,19 +4007,19 @@ static int register_hpd_handlers(struct amdgpu_device *adev) if (dc_is_dmub_outbox_supported(adev->dm.dc)) { if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD, dmub_hpd_callback, true)) { - DRM_ERROR("amdgpu: fail to register dmub hpd callback"); + drm_err(adev_to_drm(adev), "amdgpu: fail to register dmub hpd callback"); return -EINVAL; } if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD_IRQ, dmub_hpd_callback, true)) { - DRM_ERROR("amdgpu: fail to register dmub hpd callback"); + drm_err(adev_to_drm(adev), "amdgpu: fail to register dmub hpd callback"); return -EINVAL; } if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD_SENSE_NOTIFY, dmub_hpd_sense_callback, true)) { - DRM_ERROR("amdgpu: fail to register dmub hpd sense callback"); + drm_err(adev_to_drm(adev), "amdgpu: fail to register dmub hpd sense callback"); return -EINVAL; } } @@ -4120,7 +4040,7 @@ static int register_hpd_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_HPD1 || int_params.irq_source > DC_IRQ_SOURCE_HPD6) { - DRM_ERROR("Failed to register hpd irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register hpd irq!\n"); return -EINVAL; } @@ -4138,7 +4058,7 @@ static int register_hpd_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_HPD1RX || int_params.irq_source > DC_IRQ_SOURCE_HPD6RX) { - DRM_ERROR("Failed to register hpd rx irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register hpd rx irq!\n"); return -EINVAL; } @@ -4180,7 +4100,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev) for (i = 0; i < adev->mode_info.num_crtc; i++) { r = amdgpu_irq_add_id(adev, client_id, i + 1, &adev->crtc_irq); if (r) { - DRM_ERROR("Failed to add crtc irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add crtc irq id!\n"); return r; } @@ -4191,7 +4111,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_VBLANK1 || int_params.irq_source > DC_IRQ_SOURCE_VBLANK6) { - DRM_ERROR("Failed to register vblank irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register vblank irq!\n"); return -EINVAL; } @@ -4210,7 +4130,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev) i <= VISLANDS30_IV_SRCID_D6_GRPH_PFLIP; i += 2) { r = amdgpu_irq_add_id(adev, client_id, i, &adev->pageflip_irq); if (r) { - DRM_ERROR("Failed to add page flip irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add page flip irq id!\n"); return r; } @@ -4221,7 +4141,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_PFLIP_FIRST || int_params.irq_source > DC_IRQ_SOURCE_PFLIP_LAST) { - DRM_ERROR("Failed to register pflip irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register pflip irq!\n"); return -EINVAL; } @@ -4239,7 +4159,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev) r = amdgpu_irq_add_id(adev, client_id, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq); if (r) { - DRM_ERROR("Failed to add hpd irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add hpd irq id!\n"); return r; } @@ -4281,7 +4201,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) for (i = VISLANDS30_IV_SRCID_D1_VERTICAL_INTERRUPT0; i <= VISLANDS30_IV_SRCID_D6_VERTICAL_INTERRUPT0; i++) { r = amdgpu_irq_add_id(adev, client_id, i, &adev->crtc_irq); if (r) { - DRM_ERROR("Failed to add crtc irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add crtc irq id!\n"); return r; } @@ -4292,7 +4212,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_VBLANK1 || int_params.irq_source > DC_IRQ_SOURCE_VBLANK6) { - DRM_ERROR("Failed to register vblank irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register vblank irq!\n"); return -EINVAL; } @@ -4310,7 +4230,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) for (i = VISLANDS30_IV_SRCID_D1_V_UPDATE_INT; i <= VISLANDS30_IV_SRCID_D6_V_UPDATE_INT; i += 2) { r = amdgpu_irq_add_id(adev, client_id, i, &adev->vupdate_irq); if (r) { - DRM_ERROR("Failed to add vupdate irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add vupdate irq id!\n"); return r; } @@ -4321,7 +4241,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_VUPDATE1 || int_params.irq_source > DC_IRQ_SOURCE_VUPDATE6) { - DRM_ERROR("Failed to register vupdate irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register vupdate irq!\n"); return -EINVAL; } @@ -4340,7 +4260,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) i <= VISLANDS30_IV_SRCID_D6_GRPH_PFLIP; i += 2) { r = amdgpu_irq_add_id(adev, client_id, i, &adev->pageflip_irq); if (r) { - DRM_ERROR("Failed to add page flip irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add page flip irq id!\n"); return r; } @@ -4351,7 +4271,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_PFLIP_FIRST || int_params.irq_source > DC_IRQ_SOURCE_PFLIP_LAST) { - DRM_ERROR("Failed to register pflip irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register pflip irq!\n"); return -EINVAL; } @@ -4369,7 +4289,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) r = amdgpu_irq_add_id(adev, client_id, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq); if (r) { - DRM_ERROR("Failed to add hpd irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add hpd irq id!\n"); return r; } @@ -4419,7 +4339,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->crtc_irq); if (r) { - DRM_ERROR("Failed to add crtc irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add crtc irq id!\n"); return r; } @@ -4430,7 +4350,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_VBLANK1 || int_params.irq_source > DC_IRQ_SOURCE_VBLANK6) { - DRM_ERROR("Failed to register vblank irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register vblank irq!\n"); return -EINVAL; } @@ -4451,7 +4371,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) vrtl_int_srcid[i], &adev->vline0_irq); if (r) { - DRM_ERROR("Failed to add vline0 irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add vline0 irq id!\n"); return r; } @@ -4462,7 +4382,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_DC1_VLINE0 || int_params.irq_source > DC_IRQ_SOURCE_DC6_VLINE0) { - DRM_ERROR("Failed to register vline0 irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register vline0 irq!\n"); return -EINVAL; } @@ -4490,7 +4410,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->vupdate_irq); if (r) { - DRM_ERROR("Failed to add vupdate irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add vupdate irq id!\n"); return r; } @@ -4501,7 +4421,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_VUPDATE1 || int_params.irq_source > DC_IRQ_SOURCE_VUPDATE6) { - DRM_ERROR("Failed to register vupdate irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register vupdate irq!\n"); return -EINVAL; } @@ -4521,7 +4441,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) i++) { r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->pageflip_irq); if (r) { - DRM_ERROR("Failed to add page flip irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add page flip irq id!\n"); return r; } @@ -4532,7 +4452,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) if (int_params.irq_source == DC_IRQ_SOURCE_INVALID || int_params.irq_source < DC_IRQ_SOURCE_PFLIP_FIRST || int_params.irq_source > DC_IRQ_SOURCE_PFLIP_LAST) { - DRM_ERROR("Failed to register pflip irq!\n"); + drm_err(adev_to_drm(adev), "Failed to register pflip irq!\n"); return -EINVAL; } @@ -4550,7 +4470,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev) r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, DCN_1_0__SRCID__DC_HPD1_INT, &adev->hpd_irq); if (r) { - DRM_ERROR("Failed to add hpd irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add hpd irq id!\n"); return r; } @@ -4572,7 +4492,7 @@ static int register_outbox_irq_handlers(struct amdgpu_device *adev) r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, DCN_1_0__SRCID__DMCUB_OUTBOX_LOW_PRIORITY_READY_INT, &adev->dmub_outbox_irq); if (r) { - DRM_ERROR("Failed to add outbox irq id!\n"); + drm_err(adev_to_drm(adev), "Failed to add outbox irq id!\n"); return r; } @@ -4804,41 +4724,54 @@ static int get_brightness_range(const struct amdgpu_dm_backlight_caps *caps, return 1; } -static u32 convert_brightness_from_user(const struct amdgpu_dm_backlight_caps *caps, - uint32_t brightness) +static void convert_custom_brightness(const struct amdgpu_dm_backlight_caps *caps, + uint32_t *brightness) { - unsigned int min, max; u8 prev_signal = 0, prev_lum = 0; + int i = 0; - if (!get_brightness_range(caps, &min, &max)) - return brightness; - - for (int i = 0; i < caps->data_points; i++) { - u8 signal, lum; + if (amdgpu_dc_debug_mask & DC_DISABLE_CUSTOM_BRIGHTNESS_CURVE) + return; - if (amdgpu_dc_debug_mask & DC_DISABLE_CUSTOM_BRIGHTNESS_CURVE) - break; + if (!caps->data_points) + return; - signal = caps->luminance_data[i].input_signal; - lum = caps->luminance_data[i].luminance; + /* choose start to run less interpolation steps */ + if (caps->luminance_data[caps->data_points/2].input_signal > *brightness) + i = caps->data_points/2; + do { + u8 signal = caps->luminance_data[i].input_signal; + u8 lum = caps->luminance_data[i].luminance; /* * brightness == signal: luminance is percent numerator * brightness < signal: interpolate between previous and current luminance numerator * brightness > signal: find next data point */ - if (brightness < signal) - lum = prev_lum + DIV_ROUND_CLOSEST((lum - prev_lum) * - (brightness - prev_signal), - signal - prev_signal); - else if (brightness > signal) { + if (*brightness > signal) { prev_signal = signal; prev_lum = lum; + i++; continue; } - brightness = DIV_ROUND_CLOSEST(lum * brightness, 101); - break; - } + if (*brightness < signal) + lum = prev_lum + DIV_ROUND_CLOSEST((lum - prev_lum) * + (*brightness - prev_signal), + signal - prev_signal); + *brightness = DIV_ROUND_CLOSEST(lum * *brightness, 101); + return; + } while (i < caps->data_points); +} + +static u32 convert_brightness_from_user(const struct amdgpu_dm_backlight_caps *caps, + uint32_t brightness) +{ + unsigned int min, max; + + if (!get_brightness_range(caps, &min, &max)) + return brightness; + + convert_custom_brightness(caps, &brightness); // Rescale 0..255 to min..max return min + DIV_ROUND_CLOSEST((max - min) * brightness, @@ -5017,10 +4950,10 @@ amdgpu_dm_register_backlight_device(struct amdgpu_dm_connector *aconnector) dm->brightness[aconnector->bl_idx] = props.brightness; if (IS_ERR(dm->backlight_dev[aconnector->bl_idx])) { - DRM_ERROR("DM: Backlight registration failed!\n"); + drm_err(drm, "DM: Backlight registration failed!\n"); dm->backlight_dev[aconnector->bl_idx] = NULL; } else - DRM_DEBUG_DRIVER("DM: Registered Backlight device: %s\n", bl_name); + drm_dbg_driver(drm, "DM: Registered Backlight device: %s\n", bl_name); } static int initialize_plane(struct amdgpu_display_manager *dm, @@ -5034,7 +4967,7 @@ static int initialize_plane(struct amdgpu_display_manager *dm, plane = kzalloc(sizeof(struct drm_plane), GFP_KERNEL); if (!plane) { - DRM_ERROR("KMS: Failed to allocate plane\n"); + drm_err(adev_to_drm(dm->adev), "KMS: Failed to allocate plane\n"); return -ENOMEM; } plane->type = plane_type; @@ -5052,7 +4985,7 @@ static int initialize_plane(struct amdgpu_display_manager *dm, ret = amdgpu_dm_plane_init(dm, plane, possible_crtcs, plane_cap); if (ret) { - DRM_ERROR("KMS: Failed to initialize plane\n"); + drm_err(adev_to_drm(dm->adev), "KMS: Failed to initialize plane\n"); kfree(plane); return ret; } @@ -5121,14 +5054,14 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) link_cnt = dm->dc->caps.max_links; if (amdgpu_dm_mode_config_init(dm->adev)) { - DRM_ERROR("DM: Failed to initialize mode config\n"); + drm_err(adev_to_drm(adev), "DM: Failed to initialize mode config\n"); return -EINVAL; } /* There is one primary plane per CRTC */ primary_planes = dm->dc->caps.max_streams; if (primary_planes > AMDGPU_MAX_PLANES) { - DRM_ERROR("DM: Plane nums out of 6 planes\n"); + drm_err(adev_to_drm(adev), "DM: Plane nums out of 6 planes\n"); return -EINVAL; } @@ -5141,7 +5074,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) if (initialize_plane(dm, mode_info, i, DRM_PLANE_TYPE_PRIMARY, plane)) { - DRM_ERROR("KMS: Failed to initialize primary plane\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to initialize primary plane\n"); goto fail; } } @@ -5173,14 +5106,14 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) if (initialize_plane(dm, NULL, primary_planes + i, DRM_PLANE_TYPE_OVERLAY, plane)) { - DRM_ERROR("KMS: Failed to initialize overlay plane\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to initialize overlay plane\n"); goto fail; } } for (i = 0; i < dm->dc->caps.max_streams; i++) if (amdgpu_dm_crtc_init(dm, mode_info->planes[i], i)) { - DRM_ERROR("KMS: Failed to initialize crtc\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to initialize crtc\n"); goto fail; } @@ -5200,7 +5133,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) case IP_VERSION(3, 6, 0): case IP_VERSION(4, 0, 1): if (register_outbox_irq_handlers(dm->adev)) { - DRM_ERROR("DM: Failed to initialize IRQ\n"); + drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n"); goto fail; } break; @@ -5250,7 +5183,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) } if (link_cnt > MAX_LINKS) { - DRM_ERROR( + drm_err(adev_to_drm(adev), "KMS: Cannot support more than %d display indexes\n", MAX_LINKS); goto fail; @@ -5266,12 +5199,12 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) struct amdgpu_dm_wb_connector *wbcon = kzalloc(sizeof(*wbcon), GFP_KERNEL); if (!wbcon) { - DRM_ERROR("KMS: Failed to allocate writeback connector\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to allocate writeback connector\n"); continue; } if (amdgpu_dm_wb_connector_init(dm, wbcon, i)) { - DRM_ERROR("KMS: Failed to initialize writeback connector\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to initialize writeback connector\n"); kfree(wbcon); continue; } @@ -5291,12 +5224,12 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) goto fail; if (amdgpu_dm_encoder_init(dm->ddev, aencoder, i)) { - DRM_ERROR("KMS: Failed to initialize encoder\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to initialize encoder\n"); goto fail; } if (amdgpu_dm_connector_init(dm, aconnector, i, aencoder)) { - DRM_ERROR("KMS: Failed to initialize connector\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to initialize connector\n"); goto fail; } @@ -5305,7 +5238,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) aconnector; if (!dc_link_detect_connection_type(link, &new_connection_type)) - DRM_ERROR("KMS: Failed to detect connector\n"); + drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n"); if (aconnector->base.force && new_connection_type == dc_connection_none) { emulated_link_detect(link); @@ -5327,8 +5260,15 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) if (amdgpu_dm_set_replay_caps(link, aconnector)) psr_feature_enabled = false; - if (psr_feature_enabled) + if (psr_feature_enabled) { amdgpu_dm_set_psr_caps(link); + drm_info(adev_to_drm(adev), "PSR support %d, DC PSR ver %d, sink PSR ver %d DPCD caps 0x%x su_y_granularity %d\n", + link->psr_settings.psr_feature_enabled, + link->psr_settings.psr_version, + link->dpcd_caps.psr_info.psr_version, + link->dpcd_caps.psr_info.psr_dpcd_caps.raw, + link->dpcd_caps.psr_info.psr2_su_y_granularity_cap); + } } } amdgpu_set_panel_orientation(&aconnector->base); @@ -5342,7 +5282,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) case CHIP_VERDE: case CHIP_OLAND: if (dce60_register_irq_handlers(dm->adev)) { - DRM_ERROR("DM: Failed to initialize IRQ\n"); + drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n"); goto fail; } break; @@ -5364,7 +5304,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) case CHIP_VEGA12: case CHIP_VEGA20: if (dce110_register_irq_handlers(dm->adev)) { - DRM_ERROR("DM: Failed to initialize IRQ\n"); + drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n"); goto fail; } break; @@ -5392,12 +5332,12 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) case IP_VERSION(3, 6, 0): case IP_VERSION(4, 0, 1): if (dcn10_register_irq_handlers(dm->adev)) { - DRM_ERROR("DM: Failed to initialize IRQ\n"); + drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n"); goto fail; } break; default: - DRM_ERROR("Unsupported DCE IP versions: 0x%X\n", + drm_err(adev_to_drm(adev), "Unsupported DCE IP versions: 0x%X\n", amdgpu_ip_version(adev, DCE_HWIP, 0)); goto fail; } @@ -5558,7 +5498,7 @@ static int dm_early_init(struct amdgpu_ip_block *ip_block) /* if there is no object header, skip DM */ if (!amdgpu_atom_parse_data_header(ctx, index, NULL, NULL, NULL, &data_offset)) { adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; - dev_info(adev->dev, "No object header, skipping DM\n"); + drm_info(adev_to_drm(adev), "No object header, skipping DM\n"); return -ENOENT; } @@ -5670,7 +5610,7 @@ static int dm_early_init(struct amdgpu_ip_block *ip_block) adev->mode_info.num_dig = 4; break; default: - DRM_ERROR("Unsupported DCE IP versions: 0x%x\n", + drm_err(adev_to_drm(adev), "Unsupported DCE IP versions: 0x%x\n", amdgpu_ip_version(adev, DCE_HWIP, 0)); return -EINVAL; } @@ -5819,7 +5759,7 @@ fill_dc_plane_info_and_addr(struct amdgpu_device *adev, plane_info->format = SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616; break; default: - DRM_ERROR( + drm_err(adev_to_drm(adev), "Unsupported screen format %p4cc\n", &fb->format->format); return -EINVAL; @@ -6340,6 +6280,7 @@ static void fill_stream_properties_from_drm_display_mode( struct amdgpu_dm_connector *aconnector = NULL; struct hdmi_vendor_infoframe hv_frame; struct hdmi_avi_infoframe avi_frame; + ssize_t err; if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK) aconnector = to_amdgpu_dm_connector(connector); @@ -6386,9 +6327,17 @@ static void fill_stream_properties_from_drm_display_mode( } if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) { - drm_hdmi_avi_infoframe_from_display_mode(&avi_frame, (struct drm_connector *)connector, mode_in); + err = drm_hdmi_avi_infoframe_from_display_mode(&avi_frame, + (struct drm_connector *)connector, + mode_in); + if (err < 0) + drm_warn_once(connector->dev, "Failed to setup avi infoframe on connector %s: %zd \n", connector->name, err); timing_out->vic = avi_frame.video_code; - drm_hdmi_vendor_infoframe_from_display_mode(&hv_frame, (struct drm_connector *)connector, mode_in); + err = drm_hdmi_vendor_infoframe_from_display_mode(&hv_frame, + (struct drm_connector *)connector, + mode_in); + if (err < 0) + drm_warn_once(connector->dev, "Failed to setup vendor infoframe on connector %s: %zd \n", connector->name, err); timing_out->hdmi_vic = hv_frame.vic; } @@ -6513,7 +6462,7 @@ decide_crtc_timing_for_drm_display_mode(struct drm_display_mode *drm_mode, } static struct dc_sink * -create_fake_sink(struct dc_link *link) +create_fake_sink(struct drm_device *dev, struct dc_link *link) { struct dc_sink_init_data sink_init_data = { 0 }; struct dc_sink *sink = NULL; @@ -6523,7 +6472,7 @@ create_fake_sink(struct dc_link *link) sink = dc_sink_create(&sink_init_data); if (!sink) { - DRM_ERROR("Failed to create sink!\n"); + drm_err(dev, "Failed to create sink!\n"); return NULL; } sink->sink_signal = SIGNAL_TYPE_VIRTUAL; @@ -6656,7 +6605,7 @@ get_highest_refresh_rate_mode(struct amdgpu_dm_connector *aconnector, m_pref = list_first_entry_or_null( &aconnector->base.modes, struct drm_display_mode, head); if (!m_pref) { - DRM_DEBUG_DRIVER("No preferred mode found in EDID\n"); + drm_dbg_driver(aconnector->base.dev, "No preferred mode found in EDID\n"); return NULL; } } @@ -6831,7 +6780,7 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector, dc_link_get_highest_encoding_format(aconnector->dc_link), &stream->timing.dsc_cfg)) { stream->timing.flags.DSC = 1; - DRM_DEBUG_DRIVER("%s: SST_DSC [%s] DSC is selected from SST RX\n", + drm_dbg_driver(drm_connector->dev, "%s: SST_DSC [%s] DSC is selected from SST RX\n", __func__, drm_connector->name); } } else if (sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER) { @@ -6851,7 +6800,7 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector, dc_link_get_highest_encoding_format(aconnector->dc_link), &stream->timing.dsc_cfg)) { stream->timing.flags.DSC = 1; - DRM_DEBUG_DRIVER("%s: SST_DSC [%s] DSC is selected from DP-HDMI PCON\n", + drm_dbg_driver(drm_connector->dev, "%s: SST_DSC [%s] DSC is selected from DP-HDMI PCON\n", __func__, drm_connector->name); } } @@ -6879,6 +6828,7 @@ create_stream_for_sink(struct drm_connector *connector, const struct dc_stream_state *old_stream, int requested_bpc) { + struct drm_device *dev = connector->dev; struct amdgpu_dm_connector *aconnector = NULL; struct drm_display_mode *preferred_mode = NULL; const struct drm_connector_state *con_state = &dm_state->base; @@ -6901,11 +6851,6 @@ create_stream_for_sink(struct drm_connector *connector, drm_mode_init(&mode, drm_mode); memset(&saved_mode, 0, sizeof(saved_mode)); - if (connector == NULL) { - DRM_ERROR("connector is NULL!\n"); - return stream; - } - if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK) { aconnector = NULL; aconnector = to_amdgpu_dm_connector(connector); @@ -6920,7 +6865,7 @@ create_stream_for_sink(struct drm_connector *connector, } if (!aconnector || !aconnector->dc_sink) { - sink = create_fake_sink(link); + sink = create_fake_sink(dev, link); if (!sink) return stream; @@ -6932,7 +6877,7 @@ create_stream_for_sink(struct drm_connector *connector, stream = dc_create_stream_for_sink(sink); if (stream == NULL) { - DRM_ERROR("Failed to create stream for sink!\n"); + drm_err(dev, "Failed to create stream for sink!\n"); goto finish; } @@ -6964,7 +6909,7 @@ create_stream_for_sink(struct drm_connector *connector, * case, we call set mode ourselves to restore the previous mode * and the modelist may not be filled in time. */ - DRM_DEBUG_DRIVER("No preferred mode found\n"); + drm_dbg_driver(dev, "No preferred mode found\n"); } else if (aconnector) { recalculate_timing = amdgpu_freesync_vid_mode && is_freesync_video_mode(&mode, aconnector); @@ -7414,6 +7359,7 @@ static void amdgpu_dm_connector_funcs_force(struct drm_connector *connector) struct dc_sink *dc_em_sink = aconnector->dc_em_sink; const struct drm_edid *drm_edid; struct i2c_adapter *ddc; + struct drm_device *dev = connector->dev; if (dc_link && dc_link->aux_mode) ddc = &aconnector->dm_dp_aux.aux.ddc; @@ -7423,7 +7369,7 @@ static void amdgpu_dm_connector_funcs_force(struct drm_connector *connector) drm_edid = drm_edid_read_ddc(connector, ddc); drm_edid_connector_update(connector, drm_edid); if (!drm_edid) { - DRM_ERROR("No EDID found on connector: %s.\n", connector->name); + drm_err(dev, "No EDID found on connector: %s.\n", connector->name); return; } @@ -7482,7 +7428,7 @@ static void create_eml_sink(struct amdgpu_dm_connector *aconnector) drm_edid = drm_edid_read_ddc(connector, ddc); drm_edid_connector_update(connector, drm_edid); if (!drm_edid) { - DRM_ERROR("No EDID found on connector: %s.\n", connector->name); + drm_err(connector->dev, "No EDID found on connector: %s.\n", connector->name); return; } @@ -7616,7 +7562,7 @@ create_validate_stream_for_sink(struct drm_connector *connector, dm_state, old_stream, requested_bpc); if (stream == NULL) { - DRM_ERROR("Failed to create stream for sink!\n"); + drm_err(adev_to_drm(adev), "Failed to create stream for sink!\n"); break; } @@ -7691,7 +7637,7 @@ enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connec if (dc_sink == NULL && aconnector->base.force != DRM_FORCE_ON_DIGITAL && aconnector->base.force != DRM_FORCE_ON) { - DRM_ERROR("dc_sink is NULL!\n"); + drm_err(connector->dev, "dc_sink is NULL!\n"); goto fail; } @@ -8599,7 +8545,7 @@ static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm, i2c = create_i2c(link->ddc, false); if (!i2c) { - DRM_ERROR("Failed to create i2c adapter data\n"); + drm_err(adev_to_drm(dm->adev), "Failed to create i2c adapter data\n"); return -ENOMEM; } @@ -8607,7 +8553,7 @@ static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm, res = i2c_add_adapter(&i2c->base); if (res) { - DRM_ERROR("Failed to register hw i2c %d\n", link->link_index); + drm_err(adev_to_drm(dm->adev), "Failed to register hw i2c %d\n", link->link_index); goto out_free; } @@ -8621,7 +8567,7 @@ static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm, &i2c->base); if (res) { - DRM_ERROR("connector_init failed\n"); + drm_err(adev_to_drm(dm->adev), "connector_init failed\n"); aconnector->connector_id = -1; goto out_free; } @@ -9111,7 +9057,7 @@ static void amdgpu_dm_handle_vrr_transition(struct dm_crtc_state *old_state, */ WARN_ON(amdgpu_dm_crtc_set_vupdate_irq(new_state->base.crtc, true) != 0); WARN_ON(drm_crtc_vblank_get(new_state->base.crtc) != 0); - DRM_DEBUG_DRIVER("%s: crtc=%u VRR off->on: Get vblank ref\n", + drm_dbg_driver(new_state->base.crtc->dev, "%s: crtc=%u VRR off->on: Get vblank ref\n", __func__, new_state->base.crtc->base.id); } else if (old_vrr_active && !new_vrr_active) { /* Transition VRR active -> inactive: @@ -9119,7 +9065,7 @@ static void amdgpu_dm_handle_vrr_transition(struct dm_crtc_state *old_state, */ WARN_ON(amdgpu_dm_crtc_set_vupdate_irq(new_state->base.crtc, false) != 0); drm_crtc_vblank_put(new_state->base.crtc); - DRM_DEBUG_DRIVER("%s: crtc=%u VRR on->off: Drop vblank ref\n", + drm_dbg_driver(new_state->base.crtc->dev, "%s: crtc=%u VRR on->off: Drop vblank ref\n", __func__, new_state->base.crtc->base.id); } } @@ -9206,13 +9152,13 @@ static void amdgpu_dm_update_cursor(struct drm_plane *plane, if (crtc_state->stream) { if (!dc_stream_set_cursor_attributes(crtc_state->stream, &attributes)) - DRM_ERROR("DC failed to set cursor attributes\n"); + drm_err(adev_to_drm(adev), "DC failed to set cursor attributes\n"); update->cursor_attributes = &crtc_state->stream->cursor_attributes; if (!dc_stream_set_cursor_position(crtc_state->stream, &position)) - DRM_ERROR("DC failed to set cursor position\n"); + drm_err(adev_to_drm(adev), "DC failed to set cursor position\n"); update->cursor_position = &crtc_state->stream->cursor_position; } @@ -9463,7 +9409,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, bundle->surface_updates[planes_count].surface = dc_plane; if (!bundle->surface_updates[planes_count].surface) { - DRM_ERROR("No surface for CRTC: id=%d\n", + drm_err(dev, "No surface for CRTC: id=%d\n", acrtc_attach->crtc_id); continue; } @@ -9979,20 +9925,20 @@ static void dm_set_writeback(struct amdgpu_display_manager *dm, wb_info = kzalloc(sizeof(*wb_info), GFP_KERNEL); if (!wb_info) { - DRM_ERROR("Failed to allocate wb_info\n"); + drm_err(adev_to_drm(adev), "Failed to allocate wb_info\n"); return; } acrtc = to_amdgpu_crtc(wb_conn->encoder.crtc); if (!acrtc) { - DRM_ERROR("no amdgpu_crtc found\n"); + drm_err(adev_to_drm(adev), "no amdgpu_crtc found\n"); kfree(wb_info); return; } afb = to_amdgpu_framebuffer(new_con_state->writeback_job->fb); if (!afb) { - DRM_ERROR("No amdgpu_framebuffer found\n"); + drm_err(adev_to_drm(adev), "No amdgpu_framebuffer found\n"); kfree(wb_info); return; } @@ -10213,7 +10159,7 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) new_con_state->content_protection >= DRM_MODE_CONTENT_PROTECTION_DESIRED) enable_encryption = true; - DRM_INFO("[HDCP_DM] hdcp_update_display enable_encryption = %x\n", enable_encryption); + drm_info(adev_to_drm(adev), "[HDCP_DM] hdcp_update_display enable_encryption = %x\n", enable_encryption); if (aconnector->dc_link) hdcp_update_display( @@ -10305,7 +10251,7 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) */ dummy_updates = kzalloc(sizeof(struct dc_surface_update) * MAX_SURFACES, GFP_ATOMIC); if (!dummy_updates) { - DRM_ERROR("Failed to allocate memory for dummy_updates.\n"); + drm_err(adev_to_drm(adev), "Failed to allocate memory for dummy_updates.\n"); continue; } for (j = 0; j < status->plane_count; j++) @@ -10513,16 +10459,20 @@ static int dm_force_atomic_commit(struct drm_connector *connector) */ conn_state = drm_atomic_get_connector_state(state, connector); - ret = PTR_ERR_OR_ZERO(conn_state); - if (ret) + /* Check for error in getting connector state */ + if (IS_ERR(conn_state)) { + ret = PTR_ERR(conn_state); goto out; + } /* Attach crtc to drm_atomic_state*/ crtc_state = drm_atomic_get_crtc_state(state, &disconnected_acrtc->base); - ret = PTR_ERR_OR_ZERO(crtc_state); - if (ret) + /* Check for error in getting crtc state */ + if (IS_ERR(crtc_state)) { + ret = PTR_ERR(crtc_state); goto out; + } /* force a restore */ crtc_state->mode_changed = true; @@ -10530,9 +10480,11 @@ static int dm_force_atomic_commit(struct drm_connector *connector) /* Attach plane to drm_atomic_state */ plane_state = drm_atomic_get_plane_state(state, plane); - ret = PTR_ERR_OR_ZERO(plane_state); - if (ret) + /* Check for error in getting plane state */ + if (IS_ERR(plane_state)) { + ret = PTR_ERR(plane_state); goto out; + } /* Call commit internally with the state we just constructed */ ret = drm_atomic_commit(state); @@ -10540,7 +10492,7 @@ static int dm_force_atomic_commit(struct drm_connector *connector) out: drm_atomic_state_put(state); if (ret) - DRM_ERROR("Restoring old state failed with %i\n", ret); + drm_err(ddev, "Restoring old state failed with %i\n", ret); return ret; } @@ -10624,7 +10576,7 @@ static int do_aquire_global_lock(struct drm_device *dev, &commit->flip_done, 10*HZ); if (ret == 0) - DRM_ERROR("[CRTC:%d:%s] hw_done or flip_done timed out\n", + drm_err(dev, "[CRTC:%d:%s] hw_done or flip_done timed out\n", crtc->base.id, crtc->name); drm_crtc_commit_put(commit); @@ -10740,6 +10692,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, struct dm_atomic_state *dm_state = NULL; struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state; struct dc_stream_state *new_stream; + struct amdgpu_device *adev = dm->adev; int ret = 0; /* @@ -10769,8 +10722,8 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, drm_old_conn_state = drm_atomic_get_old_connector_state(state, connector); - if (IS_ERR(drm_new_conn_state)) { - ret = PTR_ERR_OR_ZERO(drm_new_conn_state); + if (WARN_ON(!drm_new_conn_state)) { + ret = -EINVAL; goto fail; } @@ -10793,7 +10746,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, */ if (!new_stream) { - DRM_DEBUG_DRIVER("%s: Failed to create new stream for crtc %d\n", + drm_dbg_driver(adev_to_drm(adev), "%s: Failed to create new stream for crtc %d\n", __func__, acrtc->base.base.id); ret = -ENOMEM; goto fail; @@ -10831,7 +10784,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, dc_is_stream_unchanged(new_stream, dm_old_crtc_state->stream) && dc_is_stream_scaling_unchanged(new_stream, dm_old_crtc_state->stream)) { new_crtc_state->mode_changed = false; - DRM_DEBUG_DRIVER("Mode change not required, setting mode_changed to %d", + drm_dbg_driver(adev_to_drm(adev), "Mode change not required, setting mode_changed to %d", new_crtc_state->mode_changed); } } @@ -10869,7 +10822,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, is_timing_unchanged_for_freesync(new_crtc_state, old_crtc_state)) { new_crtc_state->mode_changed = false; - DRM_DEBUG_DRIVER( + drm_dbg_driver(adev_to_drm(adev), "Mode change not required for front porch change, setting mode_changed to %d", new_crtc_state->mode_changed); @@ -10890,7 +10843,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, if (ret) goto fail; - DRM_DEBUG_DRIVER("Disabling DRM crtc: %d\n", + drm_dbg_driver(adev_to_drm(adev), "Disabling DRM crtc: %d\n", crtc->base.id); /* i.e. reset mode */ @@ -11743,7 +11696,7 @@ static bool amdgpu_dm_crtc_mem_type_changed(struct drm_device *dev, old_plane_state = drm_atomic_get_plane_state(state, plane); if (IS_ERR(new_plane_state) || IS_ERR(old_plane_state)) { - DRM_ERROR("Failed to get plane state for plane %s\n", plane->name); + drm_err(dev, "Failed to get plane state for plane %s\n", plane->name); return false; } @@ -12312,7 +12265,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm, res = dc_wake_and_execute_dmub_cmd(dm->dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY); if (!res) { - DRM_ERROR("EDID CEA parser failed\n"); + drm_err(adev_to_drm(dm->adev), "EDID CEA parser failed\n"); return false; } @@ -12320,7 +12273,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm, if (output->type == DMUB_CMD__EDID_CEA_ACK) { if (!output->ack.success) { - DRM_ERROR("EDID CEA ack failed at offset %d\n", + drm_err(adev_to_drm(dm->adev), "EDID CEA ack failed at offset %d\n", output->ack.offset); } } else if (output->type == DMUB_CMD__EDID_CEA_AMD_VSDB) { @@ -12332,7 +12285,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm, vsdb->min_refresh_rate_hz = output->amd_vsdb.min_frame_rate; vsdb->max_refresh_rate_hz = output->amd_vsdb.max_frame_rate; } else { - DRM_WARN("Unknown EDID CEA parser results\n"); + drm_warn(adev_to_drm(dm->adev), "Unknown EDID CEA parser results\n"); return false; } @@ -12548,7 +12501,7 @@ void amdgpu_dm_update_freesync_caps(struct drm_connector *connector, enum adaptive_sync_type as_type = ADAPTIVE_SYNC_TYPE_NONE; if (!connector->state) { - DRM_ERROR("%s - Connector has no state", __func__); + drm_err(adev_to_drm(adev), "%s - Connector has no state", __func__); goto update; } @@ -12733,7 +12686,7 @@ int amdgpu_dm_process_dmub_aux_transfer_sync( } if (!wait_for_completion_timeout(&adev->dm.dmub_aux_transfer_done, 10 * HZ)) { - DRM_ERROR("wait_for_completion_timeout timeout!"); + drm_err(adev_to_drm(adev), "wait_for_completion_timeout timeout!"); *operation_result = AUX_RET_ERROR_TIMEOUT; goto out; } @@ -12743,31 +12696,24 @@ int amdgpu_dm_process_dmub_aux_transfer_sync( * Transient states before tunneling is enabled could * lead to this error. We can ignore this for now. */ - if (p_notify->result != AUX_RET_ERROR_PROTOCOL_ERROR) { - DRM_WARN("DPIA AUX failed on 0x%x(%d), error %d\n", + if (p_notify->result == AUX_RET_ERROR_PROTOCOL_ERROR) { + drm_warn(adev_to_drm(adev), "DPIA AUX failed on 0x%x(%d), error %d\n", payload->address, payload->length, p_notify->result); } - *operation_result = AUX_RET_ERROR_INVALID_REPLY; + *operation_result = p_notify->result; goto out; } + payload->reply[0] = adev->dm.dmub_notify->aux_reply.command & 0xF; + if (adev->dm.dmub_notify->aux_reply.command & 0xF0) + /* The reply is stored in the top nibble of the command. */ + payload->reply[0] = (adev->dm.dmub_notify->aux_reply.command >> 4) & 0xF; - payload->reply[0] = adev->dm.dmub_notify->aux_reply.command; - if (!payload->write && p_notify->aux_reply.length && - (payload->reply[0] == AUX_TRANSACTION_REPLY_AUX_ACK)) { - - if (payload->length != p_notify->aux_reply.length) { - DRM_WARN("invalid read length %d from DPIA AUX 0x%x(%d)!\n", - p_notify->aux_reply.length, - payload->address, payload->length); - *operation_result = AUX_RET_ERROR_INVALID_REPLY; - goto out; - } - + /*write req may receive a byte indicating partially written number as well*/ + if (p_notify->aux_reply.length) memcpy(payload->data, p_notify->aux_reply.data, p_notify->aux_reply.length); - } /* success */ ret = p_notify->aux_reply.length; @@ -12778,6 +12724,79 @@ out: return ret; } +static void abort_fused_io( + struct dc_context *ctx, + const struct dmub_cmd_fused_request *request +) +{ + union dmub_rb_cmd command = { 0 }; + struct dmub_rb_cmd_fused_io *io = &command.fused_io; + + io->header.type = DMUB_CMD__FUSED_IO; + io->header.sub_type = DMUB_CMD__FUSED_IO_ABORT; + io->header.payload_bytes = sizeof(*io) - sizeof(io->header); + io->request = *request; + dm_execute_dmub_cmd(ctx, &command, DM_DMUB_WAIT_TYPE_NO_WAIT); +} + +static bool execute_fused_io( + struct amdgpu_device *dev, + struct dc_context *ctx, + union dmub_rb_cmd *commands, + uint8_t count, + uint32_t timeout_us +) +{ + const uint8_t ddc_line = commands[0].fused_io.request.u.aux.ddc_line; + + if (ddc_line >= ARRAY_SIZE(dev->dm.fused_io)) + return false; + + struct fused_io_sync *sync = &dev->dm.fused_io[ddc_line]; + struct dmub_rb_cmd_fused_io *first = &commands[0].fused_io; + const bool result = dm_execute_dmub_cmd_list(ctx, count, commands, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) + && first->header.ret_status + && first->request.status == FUSED_REQUEST_STATUS_SUCCESS; + + if (!result) + return false; + + while (wait_for_completion_timeout(&sync->replied, usecs_to_jiffies(timeout_us))) { + reinit_completion(&sync->replied); + + struct dmub_cmd_fused_request *reply = (struct dmub_cmd_fused_request *) sync->reply_data; + + static_assert(sizeof(*reply) <= sizeof(sync->reply_data), "Size mismatch"); + + if (reply->identifier == first->request.identifier) { + first->request = *reply; + return true; + } + } + + reinit_completion(&sync->replied); + first->request.status = FUSED_REQUEST_STATUS_TIMEOUT; + abort_fused_io(ctx, &first->request); + return false; +} + +bool amdgpu_dm_execute_fused_io( + struct amdgpu_device *dev, + struct dc_link *link, + union dmub_rb_cmd *commands, + uint8_t count, + uint32_t timeout_us) +{ + struct amdgpu_display_manager *dm = &dev->dm; + + mutex_lock(&dm->dpia_aux_lock); + + const bool result = execute_fused_io(dev, link->ctx, commands, count, timeout_us); + + mutex_unlock(&dm->dpia_aux_lock); + return result; +} + int amdgpu_dm_process_dmub_set_config_sync( struct dc_context *ctx, unsigned int link_index, @@ -12796,7 +12815,7 @@ int amdgpu_dm_process_dmub_set_config_sync( ret = 0; *operation_result = adev->dm.dmub_notify->sc_status; } else { - DRM_ERROR("wait_for_completion_timeout timeout!"); + drm_err(adev_to_drm(adev), "wait_for_completion_timeout timeout!"); ret = -1; *operation_result = SET_CONFIG_UNKNOWN_ERROR; } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h index 385faaca6e26..d7d92f9911e4 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h @@ -50,7 +50,7 @@ #define AMDGPU_DM_MAX_NUM_EDP 2 -#define AMDGPU_DMUB_NOTIFICATION_MAX 7 +#define AMDGPU_DMUB_NOTIFICATION_MAX 8 #define HDMI_AMD_VENDOR_SPECIFIC_DATA_BLOCK_IEEE_REGISTRATION_ID 0x00001A #define AMD_VSDB_VERSION_3_FEATURECAP_REPLAYMODE 0x40 @@ -81,6 +81,7 @@ struct amdgpu_bo; struct dmub_srv; struct dc_plane_state; struct dmub_notification; +struct dmub_cmd_fused_request; struct amd_vsdb_block { unsigned char ieee_id[3]; @@ -276,6 +277,10 @@ struct hpd_rx_irq_offload_work { * @offload_wq: offload work queue that this work is queued to */ struct hpd_rx_irq_offload_work_queue *offload_wq; + /** + * @adev: amdgpu_device pointer + */ + struct amdgpu_device *adev; }; /** @@ -614,6 +619,13 @@ struct amdgpu_display_manager { bool aux_hpd_discon_quirk; /** + * @edp0_on_dp1_quirk: + * + * quirk for platforms that put edp0 on DP1. + */ + bool edp0_on_dp1_quirk; + + /** * @dpia_aux_lock: * * Guards access to DPIA AUX @@ -633,6 +645,16 @@ struct amdgpu_display_manager { * OEM i2c bus */ struct amdgpu_i2c_adapter *oem_i2c; + + /** + * @fused_io: + * + * dmub fused io interface + */ + struct fused_io_sync { + struct completion replied; + char reply_data[0x40]; // Cannot include dmub_cmd here + } fused_io[8]; }; enum dsc_clock_force_state { @@ -1012,6 +1034,14 @@ extern const struct drm_encoder_helper_funcs amdgpu_dm_encoder_helper_funcs; int amdgpu_dm_process_dmub_aux_transfer_sync(struct dc_context *ctx, unsigned int link_index, struct aux_payload *payload, enum aux_return_code_type *operation_result); +bool amdgpu_dm_execute_fused_io( + struct amdgpu_device *dev, + struct dc_link *link, + union dmub_rb_cmd *commands, + uint8_t count, + uint32_t timeout_us +); + int amdgpu_dm_process_dmub_set_config_sync(struct dc_context *ctx, unsigned int link_index, struct set_config_cmd_payload *payload, enum set_config_status *operation_result); @@ -1045,4 +1075,6 @@ void hdmi_cec_set_edid(struct amdgpu_dm_connector *aconnector); void hdmi_cec_unset_edid(struct amdgpu_dm_connector *aconnector); int amdgpu_dm_initialize_hdmi_connector(struct amdgpu_dm_connector *aconnector); +void retrieve_dmi_info(struct amdgpu_display_manager *dm); + #endif /* __AMDGPU_DM_H__ */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c index 8f22ad966543..c16962256514 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c @@ -26,6 +26,7 @@ #include "amdgpu_dm_hdcp.h" #include "amdgpu.h" #include "amdgpu_dm.h" +#include "dc_fused_io.h" #include "dm_helpers.h" #include <drm/display/drm_hdcp_helper.h> #include "hdcp_psp.h" @@ -76,6 +77,34 @@ lp_read_dpcd(void *handle, uint32_t address, uint8_t *data, uint32_t size) return dm_helpers_dp_read_dpcd(link->ctx, link, address, data, size); } +static bool lp_atomic_write_poll_read_i2c( + void *handle, + const struct mod_hdcp_atomic_op_i2c *write, + const struct mod_hdcp_atomic_op_i2c *poll, + struct mod_hdcp_atomic_op_i2c *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +) +{ + struct dc_link *link = handle; + + return dm_atomic_write_poll_read_i2c(link, write, poll, read, poll_timeout_us, poll_mask_msb); +} + +static bool lp_atomic_write_poll_read_aux( + void *handle, + const struct mod_hdcp_atomic_op_aux *write, + const struct mod_hdcp_atomic_op_aux *poll, + struct mod_hdcp_atomic_op_aux *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +) +{ + struct dc_link *link = handle; + + return dm_atomic_write_poll_read_aux(link, write, poll, read, poll_timeout_us, poll_mask_msb); +} + static uint8_t *psp_get_srm(struct psp_context *psp, uint32_t *srm_version, uint32_t *srm_size) { struct ta_hdcp_shared_memory *hdcp_cmd; @@ -732,7 +761,10 @@ struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev, INIT_DELAYED_WORK(&hdcp_work[i].watchdog_timer_dwork, event_watchdog_timer); INIT_DELAYED_WORK(&hdcp_work[i].property_validate_dwork, event_property_validate); - hdcp_work[i].hdcp.config.psp.handle = &adev->psp; + struct mod_hdcp_config *config = &hdcp_work[i].hdcp.config; + struct mod_hdcp_ddc_funcs *ddc_funcs = &config->ddc.funcs; + + config->psp.handle = &adev->psp; if (dc->ctx->dce_version == DCN_VERSION_3_1 || dc->ctx->dce_version == DCN_VERSION_3_14 || dc->ctx->dce_version == DCN_VERSION_3_15 || @@ -740,12 +772,22 @@ struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev, dc->ctx->dce_version == DCN_VERSION_3_51 || dc->ctx->dce_version == DCN_VERSION_3_6 || dc->ctx->dce_version == DCN_VERSION_3_16) - hdcp_work[i].hdcp.config.psp.caps.dtm_v3_supported = 1; - hdcp_work[i].hdcp.config.ddc.handle = dc_get_link_at_index(dc, i); - hdcp_work[i].hdcp.config.ddc.funcs.write_i2c = lp_write_i2c; - hdcp_work[i].hdcp.config.ddc.funcs.read_i2c = lp_read_i2c; - hdcp_work[i].hdcp.config.ddc.funcs.write_dpcd = lp_write_dpcd; - hdcp_work[i].hdcp.config.ddc.funcs.read_dpcd = lp_read_dpcd; + config->psp.caps.dtm_v3_supported = 1; + config->ddc.handle = dc_get_link_at_index(dc, i); + + ddc_funcs->write_i2c = lp_write_i2c; + ddc_funcs->read_i2c = lp_read_i2c; + ddc_funcs->write_dpcd = lp_write_dpcd; + ddc_funcs->read_dpcd = lp_read_dpcd; + + config->debug.lc_enable_sw_fallback = dc->debug.hdcp_lc_enable_sw_fallback; + if (dc->caps.fused_io_supported || dc->debug.hdcp_lc_force_fw_enable) { + ddc_funcs->atomic_write_poll_read_i2c = lp_atomic_write_poll_read_i2c; + ddc_funcs->atomic_write_poll_read_aux = lp_atomic_write_poll_read_aux; + } else { + ddc_funcs->atomic_write_poll_read_i2c = NULL; + ddc_funcs->atomic_write_poll_read_aux = NULL; + } memset(hdcp_work[i].aconnector, 0, sizeof(struct amdgpu_dm_connector *) * diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c index 1395a748d726..d4395b92fb85 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c @@ -630,6 +630,19 @@ bool dm_helpers_submit_i2c( return result; } +bool dm_helpers_execute_fused_io( + struct dc_context *ctx, + struct dc_link *link, + union dmub_rb_cmd *commands, + uint8_t count, + uint32_t timeout_us +) +{ + struct amdgpu_device *dev = ctx->driver_context; + + return amdgpu_dm_execute_fused_io(dev, link, commands, count, timeout_us); +} + static bool execute_synaptics_rc_command(struct drm_dp_aux *aux, bool is_write_cmd, unsigned char cmd, diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index 7ceedf626d23..25e8befbcc47 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -51,6 +51,9 @@ #define PEAK_FACTOR_X1000 1006 +/* + * This function handles both native AUX and I2C-Over-AUX transactions. + */ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { @@ -59,6 +62,7 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, enum aux_return_code_type operation_result; struct amdgpu_device *adev; struct ddc_service *ddc; + uint8_t copy[16]; if (WARN_ON(msg->size > 16)) return -E2BIG; @@ -74,6 +78,11 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, (msg->request & DP_AUX_I2C_WRITE_STATUS_UPDATE) != 0; payload.defer_delay = 0; + if (payload.write) { + memcpy(copy, msg->buffer, msg->size); + payload.data = copy; + } + result = dc_link_aux_transfer_raw(TO_DM_AUX(aux)->ddc_service, &payload, &operation_result); @@ -87,15 +96,25 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, if (adev->dm.aux_hpd_discon_quirk) { if (msg->address == DP_SIDEBAND_MSG_DOWN_REQ_BASE && operation_result == AUX_RET_ERROR_HPD_DISCON) { - result = 0; + result = msg->size; operation_result = AUX_RET_SUCCESS; } } - if (payload.write && result >= 0) - result = msg->size; + /* + * result equals to 0 includes the cases of AUX_DEFER/I2C_DEFER + */ + if (payload.write && result >= 0) { + if (result) { + /*one byte indicating partially written bytes*/ + drm_dbg_dp(adev_to_drm(adev), "amdgpu: AUX partially written\n"); + result = payload.data[0]; + } else if (!payload.reply[0]) + /*I2C_ACK|AUX_ACK*/ + result = msg->size; + } - if (result < 0) + if (result < 0) { switch (operation_result) { case AUX_RET_SUCCESS: break; @@ -114,6 +133,13 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, break; } + drm_dbg_dp(adev_to_drm(adev), "amdgpu: DP AUX transfer fail:%d\n", operation_result); + } + + if (payload.reply[0]) + drm_dbg_dp(adev_to_drm(adev), "amdgpu: AUX reply command not ACK: 0x%02x.", + payload.reply[0]); + return result; } @@ -1713,16 +1739,17 @@ static bool is_dsc_common_config_possible(struct dc_stream_state *stream, struct dc_dsc_bw_range *bw_range) { struct dc_dsc_policy dsc_policy = {0}; + bool is_dsc_possible; dc_dsc_get_policy_for_timing(&stream->timing, 0, &dsc_policy, dc_link_get_highest_encoding_format(stream->link)); - dc_dsc_compute_bandwidth_range(stream->sink->ctx->dc->res_pool->dscs[0], - stream->sink->ctx->dc->debug.dsc_min_slice_height_override, - dsc_policy.min_target_bpp * 16, - dsc_policy.max_target_bpp * 16, - &stream->sink->dsc_caps.dsc_dec_caps, - &stream->timing, dc_link_get_highest_encoding_format(stream->link), bw_range); - - return bw_range->max_target_bpp_x16 && bw_range->min_target_bpp_x16; + is_dsc_possible = dc_dsc_compute_bandwidth_range(stream->sink->ctx->dc->res_pool->dscs[0], + stream->sink->ctx->dc->debug.dsc_min_slice_height_override, + dsc_policy.min_target_bpp * 16, + dsc_policy.max_target_bpp * 16, + &stream->sink->dsc_caps.dsc_dec_caps, + &stream->timing, dc_link_get_highest_encoding_format(stream->link), bw_range); + + return is_dsc_possible; } #endif diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c index 3e0f45f1711c..b7c6e8d13435 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c @@ -948,13 +948,13 @@ static int amdgpu_dm_plane_helper_prepare_fb(struct drm_plane *plane, adev = amdgpu_ttm_adev(rbo->tbo.bdev); r = amdgpu_bo_reserve(rbo, true); if (r) { - dev_err(adev->dev, "fail to reserve bo (%d)\n", r); + drm_err(adev_to_drm(adev), "fail to reserve bo (%d)\n", r); return r; } r = dma_resv_reserve_fences(rbo->tbo.base.resv, 1); if (r) { - dev_err(adev->dev, "reserving fence slot failed (%d)\n", r); + drm_err(adev_to_drm(adev), "reserving fence slot failed (%d)\n", r); goto error_unlock; } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c index e140b7a04d72..f984cb0cb889 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c @@ -87,14 +87,6 @@ void amdgpu_dm_set_psr_caps(struct dc_link *link) link->psr_settings.psr_feature_enabled = true; } - - DRM_INFO("PSR support %d, DC PSR ver %d, sink PSR ver %d DPCD caps 0x%x su_y_granularity %d\n", - link->psr_settings.psr_feature_enabled, - link->psr_settings.psr_version, - link->dpcd_caps.psr_info.psr_version, - link->dpcd_caps.psr_info.psr_dpcd_caps.raw, - link->dpcd_caps.psr_info.psr2_su_y_granularity_cap); - } /* diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_quirks.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_quirks.c new file mode 100644 index 000000000000..1da07ebf9217 --- /dev/null +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_quirks.c @@ -0,0 +1,178 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright 2025 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#include <linux/dmi.h> + +#include "amdgpu.h" +#include "amdgpu_dm.h" + +struct amdgpu_dm_quirks { + bool aux_hpd_discon; + bool support_edp0_on_dp1; +}; + +static struct amdgpu_dm_quirks quirk_entries = { + .aux_hpd_discon = false, + .support_edp0_on_dp1 = false +}; + +static int edp0_on_dp1_callback(const struct dmi_system_id *id) +{ + quirk_entries.support_edp0_on_dp1 = true; + return 0; +} + +static int aux_hpd_discon_callback(const struct dmi_system_id *id) +{ + quirk_entries.aux_hpd_discon = true; + return 0; +} + +static const struct dmi_system_id dmi_quirk_table[] = { + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3660"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3260"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3460"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Tower Plus 7010"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Tower 7010"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex SFF Plus 7010"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex SFF 7010"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Micro Plus 7010"), + }, + }, + { + .callback = aux_hpd_discon_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Micro 7010"), + }, + }, + { + .callback = edp0_on_dp1_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "HP"), + DMI_MATCH(DMI_PRODUCT_NAME, "HP Elite mt645 G8 Mobile Thin Client"), + }, + }, + { + .callback = edp0_on_dp1_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "HP"), + DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 645 14 inch G11 Notebook PC"), + }, + }, + { + .callback = edp0_on_dp1_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "HP"), + DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 665 16 inch G11 Notebook PC"), + }, + }, + { + .callback = edp0_on_dp1_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "HP"), + DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 445 14 inch G11 Notebook PC"), + }, + }, + { + .callback = edp0_on_dp1_callback, + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "HP"), + DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 465 16 inch G11 Notebook PC"), + }, + }, + {} + /* TODO: refactor this from a fixed table to a dynamic option */ +}; + +void retrieve_dmi_info(struct amdgpu_display_manager *dm) +{ + struct drm_device *dev = dm->ddev; + int dmi_id; + + dm->aux_hpd_discon_quirk = false; + dm->edp0_on_dp1_quirk = false; + + dmi_id = dmi_check_system(dmi_quirk_table); + + if (!dmi_id) + return; + + if (quirk_entries.aux_hpd_discon) { + dm->aux_hpd_discon_quirk = true; + drm_info(dev, "aux_hpd_discon_quirk attached\n"); + } + if (quirk_entries.support_edp0_on_dp1) { + dm->edp0_on_dp1_quirk = true; + drm_info(dev, "support_edp0_on_dp1 attached\n"); + } +} diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c index 0d5fefb0f591..d9527c05fc87 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c @@ -102,13 +102,13 @@ static int amdgpu_dm_wb_prepare_job(struct drm_writeback_connector *wb_connector r = amdgpu_bo_reserve(rbo, true); if (r) { - dev_err(adev->dev, "fail to reserve bo (%d)\n", r); + drm_err(adev_to_drm(adev), "fail to reserve bo (%d)\n", r); return r; } r = dma_resv_reserve_fences(rbo->tbo.base.resv, 1); if (r) { - dev_err(adev->dev, "reserving fence slot failed (%d)\n", r); + drm_err(adev_to_drm(adev), "reserving fence slot failed (%d)\n", r); goto error_unlock; } diff --git a/drivers/gpu/drm/amd/display/dc/Makefile b/drivers/gpu/drm/amd/display/dc/Makefile index 3e1f5b689718..3c9ecea7eebc 100644 --- a/drivers/gpu/drm/amd/display/dc/Makefile +++ b/drivers/gpu/drm/amd/display/dc/Makefile @@ -53,31 +53,30 @@ DC_LIBS += hdcp ifdef CONFIG_DRM_AMD_DC_FP DC_LIBS += sspl -DC_SPL_TRANS += dc_spl_translate.o +AMD_DISPLAY_FILES += $(addprefix $(AMDDALPATH)/dc/, dc_spl_translate.o) endif AMD_DC = $(addsuffix /Makefile, $(addprefix $(FULL_AMD_DISPLAY_PATH)/dc/,$(DC_LIBS))) include $(AMD_DC) -DISPLAY_CORE = dc.o dc_stat.o dc_resource.o dc_hw_sequencer.o dc_sink.o \ -dc_surface.o dc_debug.o dc_stream.o dc_link_enc_cfg.o dc_link_exports.o dc_state.o +FILES = +FILES += dc_dmub_srv.o +FILES += dc_edid_parser.o +FILES += dc_fused_io.o +FILES += dc_helper.o +FILES += core/dc.o +FILES += core/dc_debug.o +FILES += core/dc_hw_sequencer.o +FILES += core/dc_link_enc_cfg.o +FILES += core/dc_link_exports.o +FILES += core/dc_resource.o +FILES += core/dc_sink.o +FILES += core/dc_stat.o +FILES += core/dc_state.o +FILES += core/dc_stream.o +FILES += core/dc_surface.o +FILES += core/dc_vm_helper.o + +AMD_DISPLAY_FILES += $(addprefix $(AMDDALPATH)/dc/, $(FILES)) -DISPLAY_CORE += dc_vm_helper.o - -AMD_DISPLAY_CORE = $(addprefix $(AMDDALPATH)/dc/core/,$(DISPLAY_CORE)) - -AMD_DM_REG_UPDATE = $(addprefix $(AMDDALPATH)/dc/,dc_helper.o) - -AMD_DC_SPL_TRANS = $(addprefix $(AMDDALPATH)/dc/,$(DC_SPL_TRANS)) - -AMD_DISPLAY_FILES += $(AMD_DISPLAY_CORE) -AMD_DISPLAY_FILES += $(AMD_DM_REG_UPDATE) - -DC_DMUB += dc_dmub_srv.o -DC_EDID += dc_edid_parser.o -AMD_DISPLAY_DMUB = $(addprefix $(AMDDALPATH)/dc/,$(DC_DMUB)) -AMD_DISPLAY_EDID = $(addprefix $(AMDDALPATH)/dc/,$(DC_EDID)) -AMD_DISPLAY_FILES += $(AMD_DISPLAY_DMUB) $(AMD_DISPLAY_EDID) - -AMD_DISPLAY_FILES += $(AMD_DC_SPL_TRANS) diff --git a/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c b/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c index 88d3f9d7dd55..452206b5095e 100644 --- a/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c +++ b/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c @@ -51,8 +51,6 @@ static inline unsigned long long complete_integer_division_u64( { unsigned long long result; - ASSERT(divisor); - result = div64_u64_rem(dividend, divisor, remainder); return result; @@ -213,9 +211,6 @@ struct fixed31_32 dc_fixpt_recip(struct fixed31_32 arg) * @note * Good idea to use Newton's method */ - - ASSERT(arg.value); - return dc_fixpt_from_fraction( dc_fixpt_one.value, arg.value); diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c index 3bacf470f7c5..67f08495b7e6 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c @@ -2384,10 +2384,10 @@ static enum bp_result get_integrated_info_v8( } /* - * get_integrated_info_v8 + * get_integrated_info_v9 * * @brief - * Get V8 integrated BIOS information + * Get V9 integrated BIOS information * * @param * bios_parser *bp - [in]BIOS parser handler to get master data table diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c index 19897fa52e7e..d82a52319088 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c @@ -142,17 +142,3 @@ int rv1_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_di return actual_dispclk_set_mhz * 1000; } - -int rv1_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr) -{ - int actual_dprefclk_set_mhz = -1; - - actual_dprefclk_set_mhz = rv1_vbios_smu_send_msg_with_param( - clk_mgr, - VBIOSSMC_MSG_SetDprefclkFreq, - khz_to_mhz_ceil(clk_mgr->base.dprefclk_khz)); - - /* TODO: add code for programing DP DTO, currently this is down by command table */ - - return actual_dprefclk_set_mhz * 1000; -} diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h index 083cb3158859..81d7c912549c 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h @@ -27,6 +27,5 @@ #define DAL_DC_DCN10_RV1_CLK_MGR_VBIOS_SMU_H_ int rv1_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz); -int rv1_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr); #endif /* DAL_DC_DCN10_RV1_CLK_MGR_VBIOS_SMU_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c index 23b390245b5d..5a633333dbb5 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c @@ -164,20 +164,6 @@ int rn_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dis return actual_dispclk_set_mhz * 1000; } -int rn_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr) -{ - int actual_dprefclk_set_mhz = -1; - - actual_dprefclk_set_mhz = rn_vbios_smu_send_msg_with_param( - clk_mgr, - VBIOSSMC_MSG_SetDprefclkFreq, - khz_to_mhz_ceil(clk_mgr->base.dprefclk_khz)); - - /* TODO: add code for programing DP DTO, currently this is down by command table */ - - return actual_dprefclk_set_mhz * 1000; -} - int rn_vbios_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz) { int actual_dcfclk_set_mhz = -1; diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h index 1ce19d875358..f76fad87f0e1 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h @@ -30,7 +30,6 @@ enum dcn_pwr_state; int rn_vbios_smu_get_smu_version(struct clk_mgr_internal *clk_mgr); int rn_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz); -int rn_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr); int rn_vbios_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz); int rn_vbios_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz); void rn_vbios_smu_set_phyclk(struct clk_mgr_internal *clk_mgr, int requested_phyclk_khz); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c index 2d14346b680e..478b4d6a3544 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c @@ -49,12 +49,9 @@ static const struct IP_BASE MP0_BASE = { { { { 0x00016000, 0x00DC0000, 0x00E0000 { { 0, 0, 0, 0, 0, 0 } }, { { 0, 0, 0, 0, 0, 0 } }, { { 0, 0, 0, 0, 0, 0 } } } }; -static const struct IP_BASE NBIO_BASE = { { { { 0x00000000, 0x00000014, 0x00000D20, 0x00010400, 0x0241B000, 0x04040000 } }, - { { 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0 } } } }; + +#define CTX clk_mgr->base.ctx +#define IND_REG(offset) offset #define regBIF_BX_PF2_RSMU_INDEX 0x0000 #define regBIF_BX_PF2_RSMU_INDEX_BASE_IDX 1 @@ -67,9 +64,6 @@ static const struct IP_BASE NBIO_BASE = { { { { 0x00000000, 0x00000014, 0x00000D #define FN(reg_name, field) \ FD(reg_name##__##field) -#define REG_NBIO(reg_name) \ - (NBIO_BASE.instance[0].segment[regBIF_BX_PF2_ ## reg_name ## _BASE_IDX] + regBIF_BX_PF2_ ## reg_name) - #undef DC_LOGGER #define DC_LOGGER \ CTX->logger @@ -77,6 +71,13 @@ static const struct IP_BASE NBIO_BASE = { { { { 0x00000000, 0x00000014, 0x00000D #define mmMP1_C2PMSG_3 0x3B1050C +#define reg__MP1_C2PMSG_3_MASK (0xFFFFFFFF) +#define reg__MP1_C2PMSG_3__SHIFT (0) + + +#define data_reg_name__MP1_C2PMSG_3_MASK (0xFFFFFFFF) +#define data_reg_name__MP1_C2PMSG_3__SHIFT (0) + #define VBIOSSMC_MSG_TestMessage 0x01 ///< To check if PMFW is alive and responding. Requirement specified by PMFW team #define VBIOSSMC_MSG_GetPmfwVersion 0x02 ///< Get PMFW version #define VBIOSSMC_MSG_Spare0 0x03 ///< Spare0 @@ -153,12 +154,10 @@ static int dcn315_smu_send_msg_with_param( for (i = 0; i < SMU_REGISTER_WRITE_RETRY_COUNT; i++) { /* Trigger the message transaction by writing the message ID */ - generic_write_indirect_reg(CTX, - REG_NBIO(RSMU_INDEX), REG_NBIO(RSMU_DATA), - mmMP1_C2PMSG_3, msg_id); - read_back_data = generic_read_indirect_reg(CTX, - REG_NBIO(RSMU_INDEX), REG_NBIO(RSMU_DATA), - mmMP1_C2PMSG_3); + IX_REG_SET_SYNC(mmMP1_C2PMSG_3, 0, + MP1_C2PMSG_3, msg_id); + IX_REG_GET_SYNC(mmMP1_C2PMSG_3, + MP1_C2PMSG_3, &read_back_data); if (read_back_data == msg_id) break; udelay(2); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c index 6a6ae618650b..4607eff07253 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c @@ -65,6 +65,7 @@ #define mmCLK1_CLK5_ALLOW_DS 0x16EB1 #define mmCLK5_spll_field_8 0x1B04B +#define mmCLK6_spll_field_8 0x1B24B #define mmDENTIST_DISPCLK_CNTL 0x0124 #define regDENTIST_DISPCLK_CNTL 0x0064 #define regDENTIST_DISPCLK_CNTL_BASE_IDX 1 diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c index 142de8938d7c..bb1ac12a2b09 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c @@ -90,6 +90,7 @@ #define mmCLK1_CLK5_ALLOW_DS 0x16EB1 #define mmCLK5_spll_field_8 0x1B24B +#define mmCLK6_spll_field_8 0x1B24B #define mmDENTIST_DISPCLK_CNTL 0x0124 #define regDENTIST_DISPCLK_CNTL 0x0064 #define regDENTIST_DISPCLK_CNTL_BASE_IDX 1 @@ -116,6 +117,7 @@ #define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER_MASK 0x7F000000L #define CLK5_spll_field_8__spll_ssc_en_MASK 0x00002000L +#define CLK6_spll_field_8__spll_ssc_en_MASK 0x00002000L #define SMU_VER_THRESHOLD 0x5D4A00 //93.74.0 #undef FN @@ -596,7 +598,11 @@ static bool dcn35_is_spll_ssc_enabled(struct clk_mgr *clk_mgr_base) uint32_t ssc_enable; - ssc_enable = REG_READ(CLK5_spll_field_8) & CLK5_spll_field_8__spll_ssc_en_MASK; + if (clk_mgr_base->ctx->dce_version == DCN_VERSION_3_51) { + ssc_enable = REG_READ(CLK6_spll_field_8) & CLK6_spll_field_8__spll_ssc_en_MASK; + } else { + ssc_enable = REG_READ(CLK5_spll_field_8) & CLK5_spll_field_8__spll_ssc_en_MASK; + } return ssc_enable != 0; } diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c index f6f0e6a33001..604d256cb47a 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c @@ -84,8 +84,8 @@ #define VBIOSSMC_MSG_AllowZstatesEntry 0x15 #define VBIOSSMC_MSG_DisallowZstatesEntry 0x16 #define VBIOSSMC_MSG_SetDtbClk 0x17 -#define VBIOSSMC_MSG_DispPsrEntry 0x18 ///< Display PSR entry, DMU -#define VBIOSSMC_MSG_DispPsrExit 0x19 ///< Display PSR exit, DMU +#define VBIOSSMC_MSG_DispIPS2Entry 0x18 ///< Display IPS2 entry, DMU +#define VBIOSSMC_MSG_DispIPS2Exit 0x19 ///< Display IPS2 exit, DMU #define VBIOSSMC_MSG_DisableLSdma 0x1A ///< Disable LSDMA; only sent by VBIOS #define VBIOSSMC_MSG_DpControllerPhyStatus 0x1B ///< Inform PMFW about the pre conditions for turning SLDO2 on/off . bit[0]==1 precondition is met, bit[1-2] are for DPPHY number #define VBIOSSMC_MSG_QueryIPS2Support 0x1C ///< Return 1: support; else not supported @@ -475,7 +475,7 @@ int dcn35_smu_exit_low_power_state(struct clk_mgr_internal *clk_mgr) retv = dcn35_smu_send_msg_with_param( clk_mgr, - VBIOSSMC_MSG_DispPsrExit, + VBIOSSMC_MSG_DispIPS2Exit, 0); smu_print("%s: smu_exit_low_power_state return = %d\n", __func__, retv); return retv; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 28d1353f403d..56d011a1323c 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -36,7 +36,9 @@ #include "resource.h" #include "dc_state.h" #include "dc_state_priv.h" +#include "dc_plane.h" #include "dc_plane_priv.h" +#include "dc_stream_priv.h" #include "gpio_service_interface.h" #include "clk_mgr.h" @@ -439,9 +441,12 @@ bool dc_stream_adjust_vmin_vmax(struct dc *dc, * Don't adjust DRR while there's bandwidth optimizations pending to * avoid conflicting with firmware updates. */ - if (dc->ctx->dce_version > DCE_VERSION_MAX) - if (dc->optimized_required || dc->wm_optimized_required) + if (dc->ctx->dce_version > DCE_VERSION_MAX) { + if (dc->optimized_required || dc->wm_optimized_required) { + stream->adjust.timing_adjust_pending = true; return false; + } + } dc_exit_ips_for_hw_access(dc); @@ -1192,6 +1197,12 @@ static void apply_ctx_interdependent_lock(struct dc *dc, static void dc_update_visual_confirm_color(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx) { + if (dc->debug.visual_confirm & VISUAL_CONFIRM_EXPLICIT) { + memcpy(&pipe_ctx->visual_confirm_color, &pipe_ctx->plane_state->visual_confirm_color, + sizeof(pipe_ctx->visual_confirm_color)); + return; + } + if (dc->ctx->dce_version >= DCN_VERSION_1_0) { memset(&pipe_ctx->visual_confirm_color, 0, sizeof(struct tg_color)); @@ -1225,6 +1236,51 @@ static void dc_update_visual_confirm_color(struct dc *dc, struct dc_state *conte } } +void dc_get_visual_confirm_for_stream( + struct dc *dc, + struct dc_stream_state *stream_state, + struct tg_color *color) +{ + struct dc_stream_status *stream_status = dc_stream_get_status(stream_state); + struct pipe_ctx *pipe_ctx; + int i; + struct dc_plane_state *plane_state = NULL; + + if (!stream_status) + return; + + switch (dc->debug.visual_confirm) { + case VISUAL_CONFIRM_DISABLE: + return; + case VISUAL_CONFIRM_PSR: + case VISUAL_CONFIRM_FAMS: + pipe_ctx = dc_stream_get_pipe_ctx(stream_state); + if (!pipe_ctx) + return; + dc_dmub_srv_get_visual_confirm_color_cmd(dc, pipe_ctx); + memcpy(color, &dc->ctx->dmub_srv->dmub->visual_confirm_color, sizeof(struct tg_color)); + return; + + default: + /* find plane with highest layer_index */ + for (i = 0; i < stream_status->plane_count; i++) { + if (stream_status->plane_states[i]->visible) + plane_state = stream_status->plane_states[i]; + } + if (!plane_state) + return; + /* find pipe that contains plane with highest layer index */ + for (i = 0; i < MAX_PIPES; i++) { + struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + + if (pipe->plane_state == plane_state) { + memcpy(color, &pipe->visual_confirm_color, sizeof(struct tg_color)); + return; + } + } + } +} + static void disable_dangling_plane(struct dc *dc, struct dc_state *context) { int i, j; @@ -2053,6 +2109,18 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c dc->hwss.enable_accelerated_mode(dc, context); } + if (dc->hwseq->funcs.wait_for_pipe_update_if_needed) { + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &context->res_ctx.pipe_ctx[i]; + //Only delay otg master for a given config + if (resource_is_pipe_type(pipe, OTG_MASTER)) { + //dc_commit_state_no_check is always a full update + dc->hwseq->funcs.wait_for_pipe_update_if_needed(dc, pipe, false); + break; + } + } + } + if (context->stream_count > get_seamless_boot_stream_count(context) || context->stream_count == 0) dc->hwss.prepare_bandwidth(dc, context); @@ -2117,6 +2185,14 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c if (dc->hwss.program_front_end_for_ctx) { dc->hwss.interdependent_update_lock(dc, context, true); dc->hwss.program_front_end_for_ctx(dc, context); + + if (dc->hwseq->funcs.set_wait_for_update_needed_for_pipe) { + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &context->res_ctx.pipe_ctx[i]; + dc->hwseq->funcs.set_wait_for_update_needed_for_pipe(dc, pipe); + } + } + dc->hwss.interdependent_update_lock(dc, context, false); dc->hwss.post_unlock_program_front_end(dc, context); } @@ -2258,11 +2334,15 @@ enum dc_status dc_commit_streams(struct dc *dc, struct dc_commit_streams_params for (i = 0; i < params->stream_count; i++) { struct dc_stream_state *stream = params->streams[i]; struct dc_stream_status *status = dc_stream_get_status(stream); + struct dc_sink *sink = stream->sink; /* revalidate streams */ - res = dc_validate_stream(dc, stream); - if (res != DC_OK) - return res; + if (!dc_is_virtual_signal(sink->sink_signal)) { + res = dc_validate_stream(dc, stream); + if (res != DC_OK) + return res; + } + dc_stream_log(dc, stream); @@ -2815,7 +2895,7 @@ static enum surface_update_type check_update_surfaces_for_stream( int i; enum surface_update_type overall_type = UPDATE_TYPE_FAST; - if (dc->idle_optimizations_allowed) + if (dc->idle_optimizations_allowed || dc_can_clear_cursor_limit(dc)) overall_type = UPDATE_TYPE_FULL; if (stream_status == NULL || stream_status->plane_count != surface_count) @@ -3168,7 +3248,8 @@ static void copy_stream_update_to_stream(struct dc *dc, if (update->crtc_timing_adjust) { if (stream->adjust.v_total_min != update->crtc_timing_adjust->v_total_min || - stream->adjust.v_total_max != update->crtc_timing_adjust->v_total_max) + stream->adjust.v_total_max != update->crtc_timing_adjust->v_total_max || + stream->adjust.timing_adjust_pending) update->crtc_timing_adjust->timing_adjust_pending = true; stream->adjust = *update->crtc_timing_adjust; update->crtc_timing_adjust->timing_adjust_pending = false; @@ -3219,7 +3300,7 @@ static void copy_stream_update_to_stream(struct dc *dc, if (dsc_validate_context) { stream->timing.dsc_cfg = *update->dsc_config; stream->timing.flags.DSC = enable_dsc; - if (!dc->res_pool->funcs->validate_bandwidth(dc, dsc_validate_context, true)) { + if (dc->res_pool->funcs->validate_bandwidth(dc, dsc_validate_context, true) != DC_OK) { stream->timing.dsc_cfg = old_dsc_cfg; stream->timing.flags.DSC = old_dsc_enabled; update->dsc_config = NULL; @@ -3248,7 +3329,7 @@ static void backup_planes_and_stream_state( return; for (i = 0; i < status->plane_count; i++) { - scratch->plane_states[i] = *status->plane_states[i]; + dc_plane_copy_config(&scratch->plane_states[i], status->plane_states[i]); } scratch->stream_state = *stream; } @@ -3264,10 +3345,7 @@ static void restore_planes_and_stream_state( return; for (i = 0; i < status->plane_count; i++) { - /* refcount will always be valid, restore everything else */ - struct kref refcount = status->plane_states[i]->refcount; - *status->plane_states[i] = scratch->plane_states[i]; - status->plane_states[i]->refcount = refcount; + dc_plane_copy_config(status->plane_states[i], &scratch->plane_states[i]); } *stream = scratch->stream_state; } @@ -3444,7 +3522,7 @@ static bool update_planes_and_stream_state(struct dc *dc, } if (update_type == UPDATE_TYPE_FULL) { - if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false)) { + if (dc->res_pool->funcs->validate_bandwidth(dc, context, false) != DC_OK) { BREAK_TO_DEBUGGER(); goto fail; } @@ -3998,6 +4076,7 @@ static void commit_planes_for_stream(struct dc *dc, &context->res_ctx, stream); ASSERT(top_pipe_to_program != NULL); + for (i = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i]; @@ -4048,6 +4127,9 @@ static void commit_planes_for_stream(struct dc *dc, dc->hwss.wait_for_dcc_meta_propagation(dc, top_pipe_to_program); } + if (dc->hwseq->funcs.wait_for_pipe_update_if_needed) + dc->hwseq->funcs.wait_for_pipe_update_if_needed(dc, top_pipe_to_program, update_type == UPDATE_TYPE_FAST); + if (should_lock_all_pipes && dc->hwss.interdependent_update_lock) { if (dc->hwss.subvp_pipe_control_lock) dc->hwss.subvp_pipe_control_lock(dc, context, true, should_lock_all_pipes, NULL, subvp_prev_use); @@ -4168,12 +4250,6 @@ static void commit_planes_for_stream(struct dc *dc, if (update_type == UPDATE_TYPE_FAST) continue; - ASSERT(!pipe_ctx->plane_state->triplebuffer_flips); - if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) { - /*turn off triple buffer for full update*/ - dc->hwss.program_triplebuffer( - dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips); - } stream_status = stream_get_status(context, pipe_ctx->stream); @@ -4182,8 +4258,37 @@ static void commit_planes_for_stream(struct dc *dc, dc, pipe_ctx->stream, stream_status->plane_count, context); } } + + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + + if (!pipe_ctx->plane_state) + continue; + + /* Full fe update*/ + if (update_type == UPDATE_TYPE_FAST) + continue; + + ASSERT(!pipe_ctx->plane_state->triplebuffer_flips); + if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) { + /*turn off triple buffer for full update*/ + dc->hwss.program_triplebuffer( + dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips); + } + } + if (dc->hwss.program_front_end_for_ctx && update_type != UPDATE_TYPE_FAST) { dc->hwss.program_front_end_for_ctx(dc, context); + + //Pipe busy until some frame and line # + if (dc->hwseq->funcs.set_wait_for_update_needed_for_pipe && update_type == UPDATE_TYPE_FULL) { + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + + dc->hwseq->funcs.set_wait_for_update_needed_for_pipe(dc, pipe_ctx); + } + } + if (dc->debug.validate_dml_output) { for (i = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *cur_pipe = &context->res_ctx.pipe_ctx[i]; @@ -4523,7 +4628,7 @@ static struct dc_state *create_minimal_transition_state(struct dc *dc, backup_and_set_minimal_pipe_split_policy(dc, base_context, policy); /* commit minimal state */ - if (dc->res_pool->funcs->validate_bandwidth(dc, minimal_transition_context, false)) { + if (dc->res_pool->funcs->validate_bandwidth(dc, minimal_transition_context, false) == DC_OK) { /* prevent underflow and corruption when reconfiguring pipes */ force_vsync_flip_in_minimal_transition_context(minimal_transition_context); } else { @@ -4958,6 +5063,9 @@ static bool full_update_required(struct dc *dc, if (dc->idle_optimizations_allowed) return true; + if (dc_can_clear_cursor_limit(dc)) + return true; + return false; } @@ -5043,7 +5151,7 @@ static bool update_planes_and_stream_v1(struct dc *dc, copy_stream_update_to_stream(dc, context, stream, stream_update); if (update_type >= UPDATE_TYPE_FULL) { - if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false)) { + if (dc->res_pool->funcs->validate_bandwidth(dc, context, false) != DC_OK) { DC_ERROR("Mode validation failed for stream update!\n"); dc_state_release(context); return false; @@ -6187,15 +6295,22 @@ bool dc_abm_save_restore( void dc_query_current_properties(struct dc *dc, struct dc_current_properties *properties) { unsigned int i; - bool subvp_sw_cursor_req = false; + unsigned int max_cursor_size = dc->caps.max_cursor_size; + unsigned int stream_cursor_size; - for (i = 0; i < dc->current_state->stream_count; i++) { - if (check_subvp_sw_cursor_fallback_req(dc, dc->current_state->streams[i]) && !dc->current_state->streams[i]->hw_cursor_req) { - subvp_sw_cursor_req = true; - break; + if (dc->debug.allow_sw_cursor_fallback && dc->res_pool->funcs->get_max_hw_cursor_size) { + for (i = 0; i < dc->current_state->stream_count; i++) { + stream_cursor_size = dc->res_pool->funcs->get_max_hw_cursor_size(dc, + dc->current_state, + dc->current_state->streams[i]); + + if (stream_cursor_size < max_cursor_size) { + max_cursor_size = stream_cursor_size; + } } } - properties->cursor_size_limit = subvp_sw_cursor_req ? 64 : dc->caps.max_cursor_size; + + properties->cursor_size_limit = max_cursor_size; } /** @@ -6261,3 +6376,27 @@ unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context) else return 0; } + +bool dc_is_cursor_limit_pending(struct dc *dc) +{ + uint32_t i; + + for (i = 0; i < dc->current_state->stream_count; i++) { + if (dc_stream_is_cursor_limit_pending(dc, dc->current_state->streams[i])) + return true; + } + + return false; +} + +bool dc_can_clear_cursor_limit(struct dc *dc) +{ + uint32_t i; + + for (i = 0; i < dc->current_state->stream_count; i++) { + if (dc_state_can_clear_stream_cursor_subvp_limit(dc->current_state->streams[i], dc->current_state)) + return true; + } + + return false; +} diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_debug.c b/drivers/gpu/drm/amd/display/dc/core/dc_debug.c index 650e89825968..7551d0a3fe82 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_debug.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_debug.c @@ -266,6 +266,8 @@ char *dc_status_to_str(enum dc_status status) return "Fail dp payload allocation"; case DC_FAIL_DP_LINK_BANDWIDTH: return "Insufficient DP link bandwidth"; + case DC_FAIL_HW_CURSOR_SUPPORT: + return "HW Cursor not supported"; case DC_ERROR_UNEXPECTED: return "Unexpected error"; } diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c index 55b32dfbfdd6..7014b8d000bb 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c @@ -697,7 +697,7 @@ void get_fams2_visual_confirm_color( void hwss_build_fast_sequence(struct dc *dc, struct dc_dmub_cmd *dc_dmub_cmd, unsigned int dmub_cmd_count, - struct block_sequence block_sequence[], + struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE], unsigned int *num_steps, struct pipe_ctx *pipe_ctx, struct dc_stream_status *stream_status, @@ -896,7 +896,7 @@ void hwss_build_fast_sequence(struct dc *dc, } void hwss_execute_sequence(struct dc *dc, - struct block_sequence block_sequence[], + struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE], int num_steps) { unsigned int i; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index 313a32248cd7..3da25bd8b578 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -1342,32 +1342,6 @@ static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx) data->viewport_c.y += src.y / vpc_div; } -static bool is_subvp_high_refresh_candidate(struct dc_stream_state *stream) -{ - uint32_t refresh_rate; - struct dc *dc = stream->ctx->dc; - - refresh_rate = (stream->timing.pix_clk_100hz * (uint64_t)100 + - stream->timing.v_total * stream->timing.h_total - (uint64_t)1); - refresh_rate = div_u64(refresh_rate, stream->timing.v_total); - refresh_rate = div_u64(refresh_rate, stream->timing.h_total); - - /* If there's any stream that fits the SubVP high refresh criteria, - * we must return true. This is because cursor updates are asynchronous - * with full updates, so we could transition into a SubVP config and - * remain in HW cursor mode if there's no cursor update which will - * then cause corruption. - */ - if ((refresh_rate >= 120 && refresh_rate <= 175 && - stream->timing.v_addressable >= 1080 && - stream->timing.v_addressable <= 2160) && - (dc->current_state->stream_count > 1 || - (dc->current_state->stream_count == 1 && !stream->allow_freesync))) - return true; - - return false; -} - static enum controller_dp_test_pattern convert_dp_to_controller_test_pattern( enum dp_test_pattern test_pattern) { @@ -3937,6 +3911,10 @@ enum dc_status resource_map_pool_resources( if (!dc->link_srv->dp_decide_link_settings(stream, &pipe_ctx->link_config.dp_link_settings)) return DC_FAIL_DP_LINK_BANDWIDTH; + + dc->link_srv->dp_decide_tunnel_settings(stream, + &pipe_ctx->link_config.dp_tunnel_settings); + if (dc->link_srv->dp_get_encoding_format( &pipe_ctx->link_config.dp_link_settings) == DP_128b_132b_ENCODING) { pipe_ctx->stream_res.hpo_dp_stream_enc = @@ -4259,6 +4237,11 @@ enum dc_status dc_validate_with_context(struct dc *dc, } } + /* clear subvp cursor limitations */ + for (i = 0; i < context->stream_count; i++) { + dc_state_set_stream_subvp_cursor_limit(context->streams[i], context, false); + } + res = dc_validate_global_state(dc, context, fast_validate); /* calculate pixel rate divider after deciding pxiel clock & odm combine */ @@ -4385,8 +4368,7 @@ enum dc_status dc_validate_global_state( result = resource_build_scaling_params_for_context(dc, new_ctx); if (result == DC_OK) - if (!dc->res_pool->funcs->validate_bandwidth(dc, new_ctx, fast_validate)) - result = DC_FAIL_BANDWIDTH_VALIDATE; + result = dc->res_pool->funcs->validate_bandwidth(dc, new_ctx, fast_validate); return result; } @@ -5538,23 +5520,17 @@ enum dc_status update_dp_encoder_resources_for_test_harness(const struct dc *dc, return DC_OK; } -bool check_subvp_sw_cursor_fallback_req(const struct dc *dc, struct dc_stream_state *stream) +struct dscl_prog_data *resource_get_dscl_prog_data(struct pipe_ctx *pipe_ctx) { - if (!dc->debug.disable_subvp_high_refresh && is_subvp_high_refresh_candidate(stream)) - return true; - if (dc->current_state->stream_count == 1 && stream->timing.v_addressable >= 2880 && - ((stream->timing.pix_clk_100hz * 100) / stream->timing.v_total / stream->timing.h_total) < 120) - return true; - else if (dc->current_state->stream_count > 1 && stream->timing.v_addressable >= 1080 && - ((stream->timing.pix_clk_100hz * 100) / stream->timing.v_total / stream->timing.h_total) < 120) - return true; - - return false; + return &pipe_ctx->plane_res.scl_data.dscl_prog_data; } -struct dscl_prog_data *resource_get_dscl_prog_data(struct pipe_ctx *pipe_ctx) +static bool resource_allocate_mcache(struct dc_state *context, const struct dc_mcache_params *mcache_params) { - return &pipe_ctx->plane_res.scl_data.dscl_prog_data; + if (context->clk_mgr->ctx->dc->res_pool->funcs->program_mcache_pipe_config) + context->clk_mgr->ctx->dc->res_pool->funcs->program_mcache_pipe_config(context, mcache_params); + + return true; } void resource_init_common_dml2_callbacks(struct dc *dc, struct dml2_configuration_options *dml2_options) @@ -5576,6 +5552,7 @@ void resource_init_common_dml2_callbacks(struct dc *dc, struct dml2_configuratio dml2_options->callbacks.get_stream_status = &dc_state_get_stream_status; dml2_options->callbacks.get_stream_from_id = &dc_state_get_stream_from_id; dml2_options->callbacks.get_max_flickerless_instant_vtotal_increase = &dc_stream_get_max_flickerless_instant_vtotal_increase; + dml2_options->callbacks.allocate_mcache = &resource_allocate_mcache; dml2_options->svp_pstate.callbacks.dc = dc; dml2_options->svp_pstate.callbacks.add_phantom_plane = &dc_state_add_phantom_plane; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_state.c b/drivers/gpu/drm/amd/display/dc/core/dc_state.c index 1b2cce127981..4db7383720fd 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_state.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_state.c @@ -22,6 +22,7 @@ * Authors: AMD * */ +#include "dc_types.h" #include "core_types.h" #include "core_status.h" #include "dc_state.h" @@ -812,8 +813,12 @@ enum dc_status dc_state_add_phantom_stream(const struct dc *dc, if (phantom_stream_status) { phantom_stream_status->mall_stream_config.type = SUBVP_PHANTOM; phantom_stream_status->mall_stream_config.paired_stream = main_stream; + phantom_stream_status->mall_stream_config.subvp_limit_cursor_size = false; + phantom_stream_status->mall_stream_config.cursor_size_limit_subvp = false; } + dc_state_set_stream_subvp_cursor_limit(main_stream, state, true); + return res; } @@ -939,13 +944,20 @@ void dc_state_release_phantom_streams_and_planes( const struct dc *dc, struct dc_state *state) { + unsigned int phantom_count; + struct dc_stream_state *phantom_streams[MAX_PHANTOM_PIPES]; + struct dc_plane_state *phantom_planes[MAX_PHANTOM_PIPES]; int i; - for (i = 0; i < state->phantom_stream_count; i++) - dc_state_release_phantom_stream(dc, state, state->phantom_streams[i]); + phantom_count = state->phantom_stream_count; + memcpy(phantom_streams, state->phantom_streams, sizeof(struct dc_stream_state *) * MAX_PHANTOM_PIPES); + for (i = 0; i < phantom_count; i++) + dc_state_release_phantom_stream(dc, state, phantom_streams[i]); - for (i = 0; i < state->phantom_plane_count; i++) - dc_state_release_phantom_plane(dc, state, state->phantom_planes[i]); + phantom_count = state->phantom_plane_count; + memcpy(phantom_planes, state->phantom_planes, sizeof(struct dc_plane_state *) * MAX_PHANTOM_PIPES); + for (i = 0; i < phantom_count; i++) + dc_state_release_phantom_plane(dc, state, phantom_planes[i]); } struct dc_stream_state *dc_state_get_stream_from_id(const struct dc_state *state, unsigned int id) @@ -977,3 +989,94 @@ bool dc_state_is_fams2_in_use( return is_fams2_in_use; } + +void dc_state_set_stream_subvp_cursor_limit(const struct dc_stream_state *stream, + struct dc_state *state, + bool limit) +{ + struct dc_stream_status *stream_status; + + stream_status = dc_state_get_stream_status(state, stream); + + if (stream_status) { + stream_status->mall_stream_config.subvp_limit_cursor_size = limit; + } +} + +bool dc_state_get_stream_subvp_cursor_limit(const struct dc_stream_state *stream, + struct dc_state *state) +{ + bool limit = false; + + struct dc_stream_status *stream_status; + + stream_status = dc_state_get_stream_status(state, stream); + + if (stream_status) { + limit = stream_status->mall_stream_config.subvp_limit_cursor_size; + } + + return limit; +} + +void dc_state_set_stream_cursor_subvp_limit(const struct dc_stream_state *stream, + struct dc_state *state, + bool limit) +{ + struct dc_stream_status *stream_status; + + stream_status = dc_state_get_stream_status(state, stream); + + if (stream_status) { + stream_status->mall_stream_config.cursor_size_limit_subvp = limit; + } +} + +bool dc_state_get_stream_cursor_subvp_limit(const struct dc_stream_state *stream, + struct dc_state *state) +{ + bool limit = false; + + struct dc_stream_status *stream_status; + + stream_status = dc_state_get_stream_status(state, stream); + + if (stream_status) { + limit = stream_status->mall_stream_config.cursor_size_limit_subvp; + } + + return limit; +} + +bool dc_state_can_clear_stream_cursor_subvp_limit(const struct dc_stream_state *stream, + struct dc_state *state) +{ + bool can_clear_limit = false; + + struct dc_stream_status *stream_status; + + stream_status = dc_state_get_stream_status(state, stream); + + if (stream_status) { + can_clear_limit = dc_state_get_stream_cursor_subvp_limit(stream, state) && + (stream_status->mall_stream_config.type == SUBVP_PHANTOM || + stream->hw_cursor_req || + !stream_status->mall_stream_config.subvp_limit_cursor_size || + !stream->cursor_position.enable || + dc_stream_check_cursor_attributes(stream, state, &stream->cursor_attributes)); + } + + return can_clear_limit; +} + +bool dc_state_is_subvp_in_use(struct dc_state *state) +{ + uint32_t i; + + for (i = 0; i < state->stream_count; i++) { + if (dc_state_get_stream_subvp_type(state, state->streams[i]) != SUBVP_NONE) + return true; + } + + return false; +} diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c index 0478dd856d8c..b883fb24fa12 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c @@ -265,13 +265,16 @@ void program_cursor_attributes( } /* - * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address + * dc_stream_check_cursor_attributes() - Check validitity of cursor attributes and surface address */ -bool dc_stream_set_cursor_attributes( - struct dc_stream_state *stream, +bool dc_stream_check_cursor_attributes( + const struct dc_stream_state *stream, + struct dc_state *state, const struct dc_cursor_attributes *attributes) { - struct dc *dc; + const struct dc *dc; + + unsigned int max_cursor_size; if (NULL == stream) { dm_error("DC: dc_stream is NULL!\n"); @@ -289,24 +292,38 @@ bool dc_stream_set_cursor_attributes( dc = stream->ctx->dc; - /* SubVP is not compatible with HW cursor larger than 64 x 64 x 4. - * Therefore, if cursor is greater than 64 x 64 x 4, fallback to SW cursor in the following case: - * 1. If the config is a candidate for SubVP high refresh (both single an dual display configs) - * 2. If not subvp high refresh, for single display cases, if resolution is >= 5K and refresh rate < 120hz - * 3. If not subvp high refresh, for multi display cases, if resolution is >= 4K and refresh rate < 120hz + /* SubVP is not compatible with HW cursor larger than what can fit in cursor SRAM. + * Therefore, if cursor is greater than this, fallback to SW cursor. */ - if (dc->debug.allow_sw_cursor_fallback && - attributes->height * attributes->width * 4 > 16384 && - !stream->hw_cursor_req) { - if (check_subvp_sw_cursor_fallback_req(dc, stream)) + if (dc->debug.allow_sw_cursor_fallback && dc->res_pool->funcs->get_max_hw_cursor_size) { + max_cursor_size = dc->res_pool->funcs->get_max_hw_cursor_size(dc, state, stream); + max_cursor_size = max_cursor_size * max_cursor_size * 4; + + if (attributes->height * attributes->width * 4 > max_cursor_size) { return false; + } } - stream->cursor_attributes = *attributes; - return true; } +/* + * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address + */ +bool dc_stream_set_cursor_attributes( + struct dc_stream_state *stream, + const struct dc_cursor_attributes *attributes) +{ + bool result = false; + + if (dc_stream_check_cursor_attributes(stream, stream->ctx->dc->current_state, attributes)) { + stream->cursor_attributes = *attributes; + result = true; + } + + return result; +} + bool dc_stream_program_cursor_attributes( struct dc_stream_state *stream, const struct dc_cursor_attributes *attributes) @@ -552,6 +569,14 @@ bool dc_stream_fc_disable_writeback(struct dc *dc, return true; } +/** + * dc_stream_remove_writeback() - Disables writeback and removes writeback info. + * @dc: Display core control structure. + * @stream: Display core stream state. + * @dwb_pipe_inst: Display writeback pipe. + * + * Return: returns true on success, false otherwise. + */ bool dc_stream_remove_writeback(struct dc *dc, struct dc_stream_state *stream, uint32_t dwb_pipe_inst) @@ -1109,3 +1134,26 @@ unsigned int dc_stream_get_max_flickerless_instant_vtotal_increase(struct dc_str return dc_stream_get_max_flickerless_instant_vtotal_delta(stream, is_gaming, false); } + +bool dc_stream_is_cursor_limit_pending(struct dc *dc, struct dc_stream_state *stream) +{ + bool is_limit_pending = false; + + if (dc->current_state) + is_limit_pending = dc_state_get_stream_cursor_subvp_limit(stream, dc->current_state); + + return is_limit_pending; +} + +bool dc_stream_can_clear_cursor_limit(struct dc *dc, struct dc_stream_state *stream) +{ + bool can_clear_limit = false; + + if (dc->current_state) + can_clear_limit = dc_state_get_stream_cursor_subvp_limit(stream, dc->current_state) && + (stream->hw_cursor_req || + !stream->cursor_position.enable || + dc_stream_check_cursor_attributes(stream, dc->current_state, &stream->cursor_attributes)); + + return can_clear_limit; +} diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c index e6fcc21bb9bc..922f23557f5d 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c @@ -109,7 +109,8 @@ struct dc_plane_state *dc_create_plane_state(const struct dc *dc) ***************************************************************************** */ const struct dc_plane_status *dc_plane_get_status( - const struct dc_plane_state *plane_state) + const struct dc_plane_state *plane_state, + union dc_plane_status_update_flags flags) { const struct dc_plane_status *plane_status; struct dc *dc; @@ -136,7 +137,7 @@ const struct dc_plane_status *dc_plane_get_status( if (pipe_ctx->plane_state != plane_state) continue; - if (pipe_ctx->plane_state) + if (pipe_ctx->plane_state && flags.bits.address) pipe_ctx->plane_state->status.is_flip_pending = false; break; @@ -151,7 +152,8 @@ const struct dc_plane_status *dc_plane_get_status( if (pipe_ctx->plane_state != plane_state) continue; - dc->hwss.update_pending_status(pipe_ctx); + if (flags.bits.address) + dc->hwss.update_pending_status(pipe_ctx); } return plane_status; @@ -294,3 +296,17 @@ void dc_plane_force_dcc_and_tiling_disable(struct dc_plane_state *plane_state, dc->hwss.clear_surface_dcc_and_tiling(pipe_ctx, plane_state, clear_tiling); } } + +void dc_plane_copy_config(struct dc_plane_state *dst, const struct dc_plane_state *src) +{ + struct kref temp_refcount; + + /* backup persistent info */ + memcpy(&temp_refcount, &dst->refcount, sizeof(struct kref)); + + /* copy all configuration information */ + memcpy(dst, src, sizeof(struct dc_plane_state)); + + /* restore persistent info */ + memcpy(&dst->refcount, &temp_refcount, sizeof(struct kref)); +} diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 7c2ee0526926..1d917be36fc4 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -53,7 +53,7 @@ struct aux_payload; struct set_config_cmd_payload; struct dmub_notification; -#define DC_VER "3.2.325" +#define DC_VER "3.2.334" /** * MAX_SURFACES - representative of the upper bound of surfaces that can be piped to a single CRTC @@ -249,6 +249,7 @@ struct dc_caps { uint32_t i2c_speed_in_khz_hdcp; uint32_t dmdata_alloc_size; unsigned int max_cursor_size; + unsigned int max_buffered_cursor_size; unsigned int max_video_width; /* * max video plane width that can be safely assumed to be always @@ -282,6 +283,7 @@ struct dc_caps { bool edp_dsc_support; bool vbios_lttpr_aware; bool vbios_lttpr_enable; + bool fused_io_supported; uint32_t max_otg_num; uint32_t max_cab_allocation_bytes; uint32_t cache_line_size; @@ -447,6 +449,7 @@ struct dc_config { bool enable_windowed_mpo_odm; bool forceHBR2CP2520; // Used for switching between test patterns TPS4 and CP2520 uint32_t allow_edp_hotplug_detection; + bool skip_riommu_prefetch_wa; bool clamp_min_dcfclk; uint64_t vblank_alignment_dto_params; uint8_t vblank_alignment_max_frame_time_diff; @@ -496,6 +499,7 @@ enum visual_confirm { VISUAL_CONFIRM_HW_CURSOR = 20, VISUAL_CONFIRM_VABC = 21, VISUAL_CONFIRM_DCC = 22, + VISUAL_CONFIRM_EXPLICIT = 0x80000000, }; enum dc_psr_power_opts { @@ -902,6 +906,9 @@ struct dc_debug_options { bool voltage_align_fclk; bool disable_min_fclk; + bool hdcp_lc_force_fw_enable; + bool hdcp_lc_enable_sw_fallback; + bool disable_dfs_bypass; bool disable_dpp_power_gate; bool disable_hubp_power_gate; @@ -1418,6 +1425,171 @@ struct dc_scratch_space { struct dc_stream_state stream_state; }; +/* + * A link contains one or more sinks and their connected status. + * The currently active signal type (HDMI, DP-SST, DP-MST) is also reported. + */ + struct dc_link { + struct dc_sink *remote_sinks[MAX_SINKS_PER_LINK]; + unsigned int sink_count; + struct dc_sink *local_sink; + unsigned int link_index; + enum dc_connection_type type; + enum signal_type connector_signal; + enum dc_irq_source irq_source_hpd; + enum dc_irq_source irq_source_hpd_rx;/* aka DP Short Pulse */ + enum dc_irq_source irq_source_read_request;/* Read Request */ + + bool is_hpd_filter_disabled; + bool dp_ss_off; + + /** + * @link_state_valid: + * + * If there is no link and local sink, this variable should be set to + * false. Otherwise, it should be set to true; usually, the function + * core_link_enable_stream sets this field to true. + */ + bool link_state_valid; + bool aux_access_disabled; + bool sync_lt_in_progress; + bool skip_stream_reenable; + bool is_internal_display; + /** @todo Rename. Flag an endpoint as having a programmable mapping to a DIG encoder. */ + bool is_dig_mapping_flexible; + bool hpd_status; /* HPD status of link without physical HPD pin. */ + bool is_hpd_pending; /* Indicates a new received hpd */ + + /* USB4 DPIA links skip verifying link cap, instead performing the fallback method + * for every link training. This is incompatible with DP LL compliance automation, + * which expects the same link settings to be used every retry on a link loss. + * This flag is used to skip the fallback when link loss occurs during automation. + */ + bool skip_fallback_on_link_loss; + + bool edp_sink_present; + + struct dp_trace dp_trace; + + /* caps is the same as reported_link_cap. link_traing use + * reported_link_cap. Will clean up. TODO + */ + struct dc_link_settings reported_link_cap; + struct dc_link_settings verified_link_cap; + struct dc_link_settings cur_link_settings; + struct dc_lane_settings cur_lane_setting[LANE_COUNT_DP_MAX]; + struct dc_link_settings preferred_link_setting; + /* preferred_training_settings are override values that + * come from DM. DM is responsible for the memory + * management of the override pointers. + */ + struct dc_link_training_overrides preferred_training_settings; + struct dp_audio_test_data audio_test_data; + + uint8_t ddc_hw_inst; + + uint8_t hpd_src; + + uint8_t link_enc_hw_inst; + /* DIG link encoder ID. Used as index in link encoder resource pool. + * For links with fixed mapping to DIG, this is not changed after dc_link + * object creation. + */ + enum engine_id eng_id; + enum engine_id dpia_preferred_eng_id; + + bool test_pattern_enabled; + /* Pending/Current test pattern are only used to perform and track + * FIXED_VS retimer test pattern/lane adjustment override state. + * Pending allows link HWSS to differentiate PHY vs non-PHY pattern, + * to perform specific lane adjust overrides before setting certain + * PHY test patterns. In cases when lane adjust and set test pattern + * calls are not performed atomically (i.e. performing link training), + * pending_test_pattern will be invalid or contain a non-PHY test pattern + * and current_test_pattern will contain required context for any future + * set pattern/set lane adjust to transition between override state(s). + * */ + enum dp_test_pattern current_test_pattern; + enum dp_test_pattern pending_test_pattern; + + union compliance_test_state compliance_test_state; + + void *priv; + + struct ddc_service *ddc; + + enum dp_panel_mode panel_mode; + bool aux_mode; + + /* Private to DC core */ + + const struct dc *dc; + + struct dc_context *ctx; + + struct panel_cntl *panel_cntl; + struct link_encoder *link_enc; + struct graphics_object_id link_id; + /* Endpoint type distinguishes display endpoints which do not have entries + * in the BIOS connector table from those that do. Helps when tracking link + * encoder to display endpoint assignments. + */ + enum display_endpoint_type ep_type; + union ddi_channel_mapping ddi_channel_mapping; + struct connector_device_tag_info device_tag; + struct dpcd_caps dpcd_caps; + uint32_t dongle_max_pix_clk; + unsigned short chip_caps; + unsigned int dpcd_sink_count; + struct hdcp_caps hdcp_caps; + enum edp_revision edp_revision; + union dpcd_sink_ext_caps dpcd_sink_ext_caps; + + struct psr_settings psr_settings; + struct replay_settings replay_settings; + + /* Drive settings read from integrated info table */ + struct dc_lane_settings bios_forced_drive_settings; + + /* Vendor specific LTTPR workaround variables */ + uint8_t vendor_specific_lttpr_link_rate_wa; + bool apply_vendor_specific_lttpr_link_rate_wa; + + /* MST record stream using this link */ + struct link_flags { + bool dp_keep_receiver_powered; + bool dp_skip_DID2; + bool dp_skip_reset_segment; + bool dp_skip_fs_144hz; + bool dp_mot_reset_segment; + /* Some USB4 docks do not handle turning off MST DSC once it has been enabled. */ + bool dpia_mst_dsc_always_on; + /* Forced DPIA into TBT3 compatibility mode. */ + bool dpia_forced_tbt3_mode; + bool dongle_mode_timing_override; + bool blank_stream_on_ocs_change; + bool read_dpcd204h_on_irq_hpd; + bool force_dp_ffe_preset; + } wa_flags; + union dc_dp_ffe_preset forced_dp_ffe_preset; + struct link_mst_stream_allocation_table mst_stream_alloc_table; + + struct dc_link_status link_status; + struct dprx_states dprx_states; + + struct gpio *hpd_gpio; + enum dc_link_fec_state fec_state; + bool link_powered_externally; // Used to bypass hardware sequencing delays when panel is powered down forcibly + + struct dc_panel_config panel_config; + struct phy_state phy_state; + uint32_t phy_transition_bitmask; + // BW ALLOCATON USB4 ONLY + struct dc_dpia_bw_alloc dpia_bw_alloc_config; + bool skip_implict_edp_power_control; + enum backlight_control_type backlight_control_type; +}; + struct dc { struct dc_debug_options debug; struct dc_versions versions; @@ -1485,6 +1657,7 @@ struct dc { struct dc_scratch_space current_state; struct dc_scratch_space new_state; struct dc_stream_state temp_stream; // Used so we don't need to allocate stream on the stack + struct dc_link temp_link; bool pipes_to_unlock_first[MAX_PIPES]; /* Any of the pipes indicated here should be unlocked first */ } scratch; @@ -1651,170 +1824,6 @@ uint32_t dc_bandwidth_in_kbps_from_timing( const enum dc_link_encoding_format link_encoding); /* Link Interfaces */ -/* - * A link contains one or more sinks and their connected status. - * The currently active signal type (HDMI, DP-SST, DP-MST) is also reported. - */ -struct dc_link { - struct dc_sink *remote_sinks[MAX_SINKS_PER_LINK]; - unsigned int sink_count; - struct dc_sink *local_sink; - unsigned int link_index; - enum dc_connection_type type; - enum signal_type connector_signal; - enum dc_irq_source irq_source_hpd; - enum dc_irq_source irq_source_hpd_rx;/* aka DP Short Pulse */ - - bool is_hpd_filter_disabled; - bool dp_ss_off; - - /** - * @link_state_valid: - * - * If there is no link and local sink, this variable should be set to - * false. Otherwise, it should be set to true; usually, the function - * core_link_enable_stream sets this field to true. - */ - bool link_state_valid; - bool aux_access_disabled; - bool sync_lt_in_progress; - bool skip_stream_reenable; - bool is_internal_display; - /** @todo Rename. Flag an endpoint as having a programmable mapping to a DIG encoder. */ - bool is_dig_mapping_flexible; - bool hpd_status; /* HPD status of link without physical HPD pin. */ - bool is_hpd_pending; /* Indicates a new received hpd */ - - /* USB4 DPIA links skip verifying link cap, instead performing the fallback method - * for every link training. This is incompatible with DP LL compliance automation, - * which expects the same link settings to be used every retry on a link loss. - * This flag is used to skip the fallback when link loss occurs during automation. - */ - bool skip_fallback_on_link_loss; - - bool edp_sink_present; - - struct dp_trace dp_trace; - - /* caps is the same as reported_link_cap. link_traing use - * reported_link_cap. Will clean up. TODO - */ - struct dc_link_settings reported_link_cap; - struct dc_link_settings verified_link_cap; - struct dc_link_settings cur_link_settings; - struct dc_lane_settings cur_lane_setting[LANE_COUNT_DP_MAX]; - struct dc_link_settings preferred_link_setting; - /* preferred_training_settings are override values that - * come from DM. DM is responsible for the memory - * management of the override pointers. - */ - struct dc_link_training_overrides preferred_training_settings; - struct dp_audio_test_data audio_test_data; - - uint8_t ddc_hw_inst; - - uint8_t hpd_src; - - uint8_t link_enc_hw_inst; - /* DIG link encoder ID. Used as index in link encoder resource pool. - * For links with fixed mapping to DIG, this is not changed after dc_link - * object creation. - */ - enum engine_id eng_id; - enum engine_id dpia_preferred_eng_id; - - bool test_pattern_enabled; - /* Pending/Current test pattern are only used to perform and track - * FIXED_VS retimer test pattern/lane adjustment override state. - * Pending allows link HWSS to differentiate PHY vs non-PHY pattern, - * to perform specific lane adjust overrides before setting certain - * PHY test patterns. In cases when lane adjust and set test pattern - * calls are not performed atomically (i.e. performing link training), - * pending_test_pattern will be invalid or contain a non-PHY test pattern - * and current_test_pattern will contain required context for any future - * set pattern/set lane adjust to transition between override state(s). - * */ - enum dp_test_pattern current_test_pattern; - enum dp_test_pattern pending_test_pattern; - - union compliance_test_state compliance_test_state; - - void *priv; - - struct ddc_service *ddc; - - enum dp_panel_mode panel_mode; - bool aux_mode; - - /* Private to DC core */ - - const struct dc *dc; - - struct dc_context *ctx; - - struct panel_cntl *panel_cntl; - struct link_encoder *link_enc; - struct graphics_object_id link_id; - /* Endpoint type distinguishes display endpoints which do not have entries - * in the BIOS connector table from those that do. Helps when tracking link - * encoder to display endpoint assignments. - */ - enum display_endpoint_type ep_type; - union ddi_channel_mapping ddi_channel_mapping; - struct connector_device_tag_info device_tag; - struct dpcd_caps dpcd_caps; - uint32_t dongle_max_pix_clk; - unsigned short chip_caps; - unsigned int dpcd_sink_count; - struct hdcp_caps hdcp_caps; - enum edp_revision edp_revision; - union dpcd_sink_ext_caps dpcd_sink_ext_caps; - - struct psr_settings psr_settings; - struct replay_settings replay_settings; - - /* Drive settings read from integrated info table */ - struct dc_lane_settings bios_forced_drive_settings; - - /* Vendor specific LTTPR workaround variables */ - uint8_t vendor_specific_lttpr_link_rate_wa; - bool apply_vendor_specific_lttpr_link_rate_wa; - - /* MST record stream using this link */ - struct link_flags { - bool dp_keep_receiver_powered; - bool dp_skip_DID2; - bool dp_skip_reset_segment; - bool dp_skip_fs_144hz; - bool dp_mot_reset_segment; - /* Some USB4 docks do not handle turning off MST DSC once it has been enabled. */ - bool dpia_mst_dsc_always_on; - /* Forced DPIA into TBT3 compatibility mode. */ - bool dpia_forced_tbt3_mode; - bool dongle_mode_timing_override; - bool blank_stream_on_ocs_change; - bool read_dpcd204h_on_irq_hpd; - bool force_dp_ffe_preset; - } wa_flags; - union dc_dp_ffe_preset forced_dp_ffe_preset; - struct link_mst_stream_allocation_table mst_stream_alloc_table; - - struct dc_link_status link_status; - struct dprx_states dprx_states; - - struct gpio *hpd_gpio; - enum dc_link_fec_state fec_state; - bool link_powered_externally; // Used to bypass hardware sequencing delays when panel is powered down forcibly - - struct dc_panel_config panel_config; - struct phy_state phy_state; - uint32_t phy_transition_bitmask; - // BW ALLOCATON USB4 ONLY - struct dc_dpia_bw_alloc dpia_bw_alloc_config; - bool skip_implict_edp_power_control; - enum backlight_control_type backlight_control_type; -}; - /* Return an enumerated dc_link. * dc_link order is constant and determined at * boot time. They cannot be created or destroyed. @@ -2589,10 +2598,18 @@ unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context); /* DSC Interfaces */ #include "dc_dsc.h" +void dc_get_visual_confirm_for_stream( + struct dc *dc, + struct dc_stream_state *stream_state, + struct tg_color *color); + /* Disable acc mode Interfaces */ void dc_disable_accelerated_mode(struct dc *dc); bool dc_is_timing_changed(struct dc_stream_state *cur_stream, struct dc_stream_state *new_stream); +bool dc_is_cursor_limit_pending(struct dc *dc); +bool dc_can_clear_cursor_limit(struct dc *dc); + #endif /* DC_INTERFACE_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c index 614e03bfd598..afbcf866520e 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c @@ -39,6 +39,7 @@ #define CTX dc_dmub_srv->ctx #define DC_LOGGER CTX->logger +#define GPINT_RETRY_NUM 20 static void dc_dmub_srv_construct(struct dc_dmub_srv *dc_srv, struct dc *dc, struct dmub_srv *dmub) @@ -70,20 +71,28 @@ void dc_dmub_srv_destroy(struct dc_dmub_srv **dmub_srv) } } -void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv) +bool dc_dmub_srv_wait_for_pending(struct dc_dmub_srv *dc_dmub_srv) { - struct dmub_srv *dmub = dc_dmub_srv->dmub; - struct dc_context *dc_ctx = dc_dmub_srv->ctx; + struct dmub_srv *dmub; + struct dc_context *dc_ctx; enum dmub_status status; + if (!dc_dmub_srv || !dc_dmub_srv->dmub) + return false; + + dc_ctx = dc_dmub_srv->ctx; + dmub = dc_dmub_srv->dmub; + do { - status = dmub_srv_wait_for_idle(dmub, 100000); + status = dmub_srv_wait_for_pending(dmub, 100000); } while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK); if (status != DMUB_STATUS_OK) { DC_ERROR("Error waiting for DMUB idle: status=%d\n", status); dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); } + + return status == DMUB_STATUS_OK; } void dc_dmub_srv_clear_inbox0_ack(struct dc_dmub_srv *dc_dmub_srv) @@ -126,7 +135,49 @@ void dc_dmub_srv_send_inbox0_cmd(struct dc_dmub_srv *dc_dmub_srv, } } -bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, +static bool dc_dmub_srv_reg_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, + unsigned int count, + union dmub_rb_cmd *cmd_list) +{ + struct dc_context *dc_ctx; + struct dmub_srv *dmub; + enum dmub_status status = DMUB_STATUS_OK; + int i; + + if (!dc_dmub_srv || !dc_dmub_srv->dmub) + return false; + + dc_ctx = dc_dmub_srv->ctx; + dmub = dc_dmub_srv->dmub; + + for (i = 0 ; i < count; i++) { + /* confirm no messages pending */ + do { + status = dmub_srv_wait_for_idle(dmub, 100000); + } while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK); + + /* queue command */ + if (status == DMUB_STATUS_OK) + status = dmub_srv_reg_cmd_execute(dmub, &cmd_list[i]); + + /* check for errors */ + if (status != DMUB_STATUS_OK) { + break; + } + } + + if (status != DMUB_STATUS_OK) { + if (status != DMUB_STATUS_POWER_STATE_D3) { + DC_ERROR("Error starting DMUB execution: status=%d\n", status); + dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); + } + return false; + } + + return true; +} + +static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list) { @@ -143,20 +194,25 @@ bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, for (i = 0 ; i < count; i++) { // Queue command - status = dmub_srv_cmd_queue(dmub, &cmd_list[i]); + if (!cmd_list[i].cmd_common.header.multi_cmd_pending || + dmub_rb_num_free(&dmub->inbox1.rb) >= count - i) { + status = dmub_srv_fb_cmd_queue(dmub, &cmd_list[i]); + } else { + status = DMUB_STATUS_QUEUE_FULL; + } if (status == DMUB_STATUS_QUEUE_FULL) { /* Execute and wait for queue to become empty again. */ - status = dmub_srv_cmd_execute(dmub); + status = dmub_srv_fb_cmd_execute(dmub); if (status == DMUB_STATUS_POWER_STATE_D3) return false; do { - status = dmub_srv_wait_for_idle(dmub, 100000); + status = dmub_srv_wait_for_inbox_free(dmub, 100000, count - i); } while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK); /* Requeue the command. */ - status = dmub_srv_cmd_queue(dmub, &cmd_list[i]); + status = dmub_srv_fb_cmd_queue(dmub, &cmd_list[i]); } if (status != DMUB_STATUS_OK) { @@ -168,7 +224,7 @@ bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, } } - status = dmub_srv_cmd_execute(dmub); + status = dmub_srv_fb_cmd_execute(dmub); if (status != DMUB_STATUS_OK) { if (status != DMUB_STATUS_POWER_STATE_D3) { DC_ERROR("Error starting DMUB execution: status=%d\n", status); @@ -180,6 +236,26 @@ bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, return true; } +bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv, + unsigned int count, + union dmub_rb_cmd *cmd_list) +{ + bool res = false; + + if (dc_dmub_srv && dc_dmub_srv->dmub) { + if (dc_dmub_srv->dmub->inbox_type == DMUB_CMD_INTERFACE_REG) { + res = dc_dmub_srv_reg_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list); + } else { + res = dc_dmub_srv_fb_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list); + } + + if (res) + res = dmub_srv_update_inbox_status(dc_dmub_srv->dmub) == DMUB_STATUS_OK; + } + + return res; +} + bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv, enum dm_dmub_wait_type wait_type, union dmub_rb_cmd *cmd_list) @@ -202,7 +278,8 @@ bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv, DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status); if (!dmub->debug.timeout_info.timeout_occured) { dmub->debug.timeout_info.timeout_occured = true; - dmub->debug.timeout_info.timeout_cmd = *cmd_list; + if (cmd_list) + dmub->debug.timeout_info.timeout_cmd = *cmd_list; dmub->debug.timeout_info.timestamp = dm_get_timestamp(dc_dmub_srv->ctx); } dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); @@ -210,8 +287,9 @@ bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv, } // Copy data back from ring buffer into command - if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) - dmub_rb_get_return_data(&dmub->inbox1_rb, cmd_list); + if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY && cmd_list) { + dmub_srv_cmd_get_response(dc_dmub_srv->dmub, cmd_list); + } } return true; @@ -224,74 +302,10 @@ bool dc_dmub_srv_cmd_run(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd bool dc_dmub_srv_cmd_run_list(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list, enum dm_dmub_wait_type wait_type) { - struct dc_context *dc_ctx; - struct dmub_srv *dmub; - enum dmub_status status; - int i; - - if (!dc_dmub_srv || !dc_dmub_srv->dmub) - return false; - - dc_ctx = dc_dmub_srv->ctx; - dmub = dc_dmub_srv->dmub; - - for (i = 0 ; i < count; i++) { - // Queue command - status = dmub_srv_cmd_queue(dmub, &cmd_list[i]); - - if (status == DMUB_STATUS_QUEUE_FULL) { - /* Execute and wait for queue to become empty again. */ - status = dmub_srv_cmd_execute(dmub); - if (status == DMUB_STATUS_POWER_STATE_D3) - return false; - - status = dmub_srv_wait_for_idle(dmub, 100000); - if (status != DMUB_STATUS_OK) - return false; - - /* Requeue the command. */ - status = dmub_srv_cmd_queue(dmub, &cmd_list[i]); - } - - if (status != DMUB_STATUS_OK) { - if (status != DMUB_STATUS_POWER_STATE_D3) { - DC_ERROR("Error queueing DMUB command: status=%d\n", status); - dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); - } - return false; - } - } - - status = dmub_srv_cmd_execute(dmub); - if (status != DMUB_STATUS_OK) { - if (status != DMUB_STATUS_POWER_STATE_D3) { - DC_ERROR("Error starting DMUB execution: status=%d\n", status); - dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); - } + if (!dc_dmub_srv_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list)) return false; - } - // Wait for DMUB to process command - if (wait_type != DM_DMUB_WAIT_TYPE_NO_WAIT) { - if (dc_dmub_srv->ctx->dc->debug.disable_timeout) { - do { - status = dmub_srv_wait_for_idle(dmub, 100000); - } while (status != DMUB_STATUS_OK); - } else - status = dmub_srv_wait_for_idle(dmub, 100000); - - if (status != DMUB_STATUS_OK) { - DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status); - dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); - return false; - } - - // Copy data back from ring buffer into command - if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) - dmub_rb_get_return_data(&dmub->inbox1_rb, cmd_list); - } - - return true; + return dc_dmub_srv_wait_for_idle(dc_dmub_srv, wait_type, cmd_list); } bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv) @@ -1243,7 +1257,7 @@ static void dc_dmub_srv_notify_idle(const struct dc *dc, bool allow_idle) ips_fw->signals.bits.ips1_commit, ips_fw->signals.bits.ips2_commit); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dc_dmub_srv_wait_for_idle(dc->ctx->dmub_srv, DM_DMUB_WAIT_TYPE_WAIT, NULL); memset(&new_signals, 0, sizeof(new_signals)); @@ -1355,14 +1369,15 @@ static void dc_dmub_srv_exit_low_power_state(const struct dc *dc) if (!dc->debug.optimize_ips_handshake || !ips_fw->signals.bits.ips2_commit) udelay(dc->debug.ips2_eval_delay_us); - if (ips_fw->signals.bits.ips2_commit) { - DC_LOG_IPS( - "exit IPS2 #1 (ips1_commit=%u ips2_commit=%u)", - ips_fw->signals.bits.ips1_commit, - ips_fw->signals.bits.ips2_commit); + DC_LOG_IPS( + "exit IPS2 #1 (ips1_commit=%u ips2_commit=%u)", + ips_fw->signals.bits.ips1_commit, + ips_fw->signals.bits.ips2_commit); - // Tell PMFW to exit low power state - dc->clk_mgr->funcs->exit_low_power_state(dc->clk_mgr); + // Tell PMFW to exit low power state + dc->clk_mgr->funcs->exit_low_power_state(dc->clk_mgr); + + if (ips_fw->signals.bits.ips2_commit) { DC_LOG_IPS( "wait IPS2 entry delay (ips1_commit=%u ips2_commit=%u)", @@ -1400,7 +1415,7 @@ static void dc_dmub_srv_exit_low_power_state(const struct dc *dc) ips_fw->signals.bits.ips1_commit, ips_fw->signals.bits.ips2_commit); - dmub_srv_sync_inbox1(dc->ctx->dmub_srv->dmub); + dmub_srv_sync_inboxes(dc->ctx->dmub_srv->dmub); } } @@ -1654,7 +1669,8 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc, /* fill in generic command header */ global_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH; global_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG; - global_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header); + global_cmd->header.payload_bytes = + sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header); if (enable) { /* send global configuration parameters */ @@ -1673,11 +1689,13 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc, /* configure command header */ stream_base_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH; stream_base_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG; - stream_base_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header); + stream_base_cmd->header.payload_bytes = + sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header); stream_base_cmd->header.multi_cmd_pending = 1; stream_sub_state_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH; stream_sub_state_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG; - stream_sub_state_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header); + stream_sub_state_cmd->header.payload_bytes = + sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header); stream_sub_state_cmd->header.multi_cmd_pending = 1; /* copy stream static base state */ memcpy(&stream_base_cmd->config, @@ -1723,7 +1741,8 @@ void dc_dmub_srv_fams2_drr_update(struct dc *dc, cmd.fams2_drr_update.dmub_optc_state_req.v_total_mid_frame_num = vtotal_mid_frame_num; cmd.fams2_drr_update.dmub_optc_state_req.program_manual_trigger = program_manual_trigger; - cmd.fams2_drr_update.header.payload_bytes = sizeof(cmd.fams2_drr_update) - sizeof(cmd.fams2_drr_update.header); + cmd.fams2_drr_update.header.payload_bytes = + sizeof(cmd.fams2_drr_update) - sizeof(cmd.fams2_drr_update.header); dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } @@ -1759,7 +1778,8 @@ void dc_dmub_srv_fams2_passthrough_flip( /* build command header */ cmds[num_cmds].fams2_flip.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH; cmds[num_cmds].fams2_flip.header.sub_type = DMUB_CMD__FAMS2_FLIP; - cmds[num_cmds].fams2_flip.header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2_flip); + cmds[num_cmds].fams2_flip.header.payload_bytes = + sizeof(struct dmub_rb_cmd_fams2_flip) - sizeof(struct dmub_cmd_header); /* for chaining multiple commands, all but last command should set to 1 */ cmds[num_cmds].fams2_flip.header.multi_cmd_pending = 1; @@ -1869,11 +1889,14 @@ void dc_dmub_srv_ips_query_residency_info(struct dc_dmub_srv *dc_dmub_srv, struc if (command_code == DMUB_GPINT__INVALID_COMMAND) return; - // send gpint commands and wait for ack - if (!dc_wake_and_execute_gpint(dc_dmub_srv->ctx, DMUB_GPINT__GET_IPS_RESIDENCY_PERCENT, - (uint16_t)(output->ips_mode), - &output->residency_percent, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) - output->residency_percent = 0; + for (i = 0; i < GPINT_RETRY_NUM; i++) { + // false could mean GPINT timeout, in which case we should retry + if (dc_wake_and_execute_gpint(dc_dmub_srv->ctx, DMUB_GPINT__GET_IPS_RESIDENCY_PERCENT, + (uint16_t)(output->ips_mode), &output->residency_percent, + DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) + break; + udelay(100); + } if (!dc_wake_and_execute_gpint(dc_dmub_srv->ctx, DMUB_GPINT__GET_IPS_RESIDENCY_ENTRY_COUNTER, (uint16_t)(output->ips_mode), diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h index a636f4c3f01d..ada5c2fb2db3 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h @@ -58,7 +58,7 @@ struct dc_dmub_srv { bool needs_idle_wake; }; -void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv); +bool dc_dmub_srv_wait_for_pending(struct dc_dmub_srv *dc_dmub_srv); bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv); diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h index 77c87ad57220..0bad8304ccf6 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h @@ -159,6 +159,11 @@ struct dc_link_settings { uint8_t link_rate_set; }; +struct dc_tunnel_settings { + bool should_enable_dp_tunneling; + bool should_use_dp_bw_allocation; +}; + union dc_dp_ffe_preset { struct { uint8_t level : 4; @@ -943,10 +948,20 @@ union dpia_info { uint8_t raw; }; +/* DPCD[0xE0020] USB4_DRIVER_BW_CAPABILITY register. */ +union usb4_driver_bw_cap { + struct { + uint8_t rsvd :7; + uint8_t driver_bw_alloc_support :1; + } bits; + uint8_t raw; +}; + /* DP Tunneling over USB4 */ struct dpcd_usb4_dp_tunneling_info { union dp_tun_cap_support dp_tun_cap; union dpia_info dpia_info; + union usb4_driver_bw_cap driver_bw_cap; uint8_t usb4_driver_id; uint8_t usb4_topology_id[DPCD_USB4_TOPOLOGY_ID_LEN]; }; @@ -1486,5 +1501,11 @@ struct dp_trace { # ifndef DP_TUNNELING_BW_ALLOC_CAP_CHANGED # define DP_TUNNELING_BW_ALLOC_CAP_CHANGED (1 << 3) # endif +# ifndef DPTX_BW_ALLOC_UNMASK_IRQ +# define DPTX_BW_ALLOC_UNMASK_IRQ (1 << 6) +# endif +# ifndef DPTX_BW_ALLOC_MODE_ENABLE +# define DPTX_BW_ALLOC_MODE_ENABLE (1 << 7) +# endif #endif /* DC_DP_TYPES_H */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_fused_io.c b/drivers/gpu/drm/amd/display/dc/dc_fused_io.c new file mode 100644 index 000000000000..fee69642fb93 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dc_fused_io.c @@ -0,0 +1,148 @@ +// SPDX-License-Identifier: MIT +// +// Copyright 2025 Advanced Micro Devices, Inc. + +#include "dc_fused_io.h" + +#include "dm_helpers.h" +#include "gpio.h" + +static bool op_i2c_convert( + union dmub_rb_cmd *cmd, + const struct mod_hdcp_atomic_op_i2c *op, + enum dmub_cmd_fused_request_type type, + uint32_t ddc_line, + bool over_aux +) +{ + struct dmub_cmd_fused_request *req = &cmd->fused_io.request; + struct dmub_cmd_fused_request_location_i2c *loc = &req->u.i2c; + + if (!op || op->size > sizeof(req->buffer)) + return false; + + req->type = type; + loc->is_aux = false; + loc->ddc_line = ddc_line; + loc->over_aux = over_aux; + loc->address = op->address; + loc->offset = op->offset; + loc->length = op->size; + memcpy(req->buffer, op->data, op->size); + + return true; +} + +static bool op_aux_convert( + union dmub_rb_cmd *cmd, + const struct mod_hdcp_atomic_op_aux *op, + enum dmub_cmd_fused_request_type type, + uint32_t ddc_line +) +{ + struct dmub_cmd_fused_request *req = &cmd->fused_io.request; + struct dmub_cmd_fused_request_location_aux *loc = &req->u.aux; + + if (!op || op->size > sizeof(req->buffer)) + return false; + + req->type = type; + loc->is_aux = true; + loc->ddc_line = ddc_line; + loc->address = op->address; + loc->length = op->size; + memcpy(req->buffer, op->data, op->size); + + return true; +} + +static bool atomic_write_poll_read( + struct dc_link *link, + union dmub_rb_cmd commands[3], + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +) +{ + const uint8_t count = 3; + const uint32_t timeout_per_request_us = 10000; + const uint32_t timeout_per_aux_transaction_us = 10000; + uint64_t timeout_us = 0; + + commands[1].fused_io.request.poll_mask_msb = poll_mask_msb; + commands[1].fused_io.request.timeout_us = poll_timeout_us; + + for (uint8_t i = 0; i < count; i++) { + struct dmub_rb_cmd_fused_io *io = &commands[i].fused_io; + + io->header.type = DMUB_CMD__FUSED_IO; + io->header.sub_type = DMUB_CMD__FUSED_IO_EXECUTE; + io->header.multi_cmd_pending = i != count - 1; + io->header.payload_bytes = sizeof(commands[i].fused_io) - sizeof(io->header); + + timeout_us += timeout_per_request_us + io->request.timeout_us; + if (!io->request.timeout_us && io->request.u.aux.is_aux) + timeout_us += timeout_per_aux_transaction_us * (io->request.u.aux.length / 16); + } + + if (!dm_helpers_execute_fused_io(link->ctx, link, commands, count, timeout_us)) + return false; + + return commands[0].fused_io.request.status == FUSED_REQUEST_STATUS_SUCCESS; +} + +bool dm_atomic_write_poll_read_i2c( + struct dc_link *link, + const struct mod_hdcp_atomic_op_i2c *write, + const struct mod_hdcp_atomic_op_i2c *poll, + struct mod_hdcp_atomic_op_i2c *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +) +{ + if (!link) + return false; + + const bool over_aux = false; + const uint32_t ddc_line = link->ddc->ddc_pin->pin_data->en; + + union dmub_rb_cmd commands[3] = { 0 }; + const bool converted = op_i2c_convert(&commands[0], write, FUSED_REQUEST_WRITE, ddc_line, over_aux) + && op_i2c_convert(&commands[1], poll, FUSED_REQUEST_POLL, ddc_line, over_aux) + && op_i2c_convert(&commands[2], read, FUSED_REQUEST_READ, ddc_line, over_aux); + + if (!converted) + return false; + + const bool result = atomic_write_poll_read(link, commands, poll_timeout_us, poll_mask_msb); + + memcpy(read->data, commands[0].fused_io.request.buffer, read->size); + return result; +} + +bool dm_atomic_write_poll_read_aux( + struct dc_link *link, + const struct mod_hdcp_atomic_op_aux *write, + const struct mod_hdcp_atomic_op_aux *poll, + struct mod_hdcp_atomic_op_aux *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +) +{ + if (!link) + return false; + + const uint32_t ddc_line = link->ddc->ddc_pin->pin_data->en; + union dmub_rb_cmd commands[3] = { 0 }; + const bool converted = op_aux_convert(&commands[0], write, FUSED_REQUEST_WRITE, ddc_line) + && op_aux_convert(&commands[1], poll, FUSED_REQUEST_POLL, ddc_line) + && op_aux_convert(&commands[2], read, FUSED_REQUEST_READ, ddc_line); + + if (!converted) + return false; + + const bool result = atomic_write_poll_read(link, commands, poll_timeout_us, poll_mask_msb); + + memcpy(read->data, commands[0].fused_io.request.buffer, read->size); + return result; +} + diff --git a/drivers/gpu/drm/amd/display/dc/dc_fused_io.h b/drivers/gpu/drm/amd/display/dc/dc_fused_io.h new file mode 100644 index 000000000000..c74917240985 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dc_fused_io.h @@ -0,0 +1,31 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright 2025 Advanced Micro Devices, Inc. + */ + +#ifndef __DC_FUSED_IO_H__ +#define __DC_FUSED_IO_H__ + +#include "dc.h" +#include "mod_hdcp.h" + +bool dm_atomic_write_poll_read_i2c( + struct dc_link *link, + const struct mod_hdcp_atomic_op_i2c *write, + const struct mod_hdcp_atomic_op_i2c *poll, + struct mod_hdcp_atomic_op_i2c *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +); + +bool dm_atomic_write_poll_read_aux( + struct dc_link *link, + const struct mod_hdcp_atomic_op_aux *write, + const struct mod_hdcp_atomic_op_aux *poll, + struct mod_hdcp_atomic_op_aux *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb +); + +#endif // __DC_FUSED_IO_H__ + diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c index 8f077e15b4f0..7217de258851 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c @@ -682,13 +682,19 @@ void reg_sequence_wait_done(const struct dc_context *ctx) if (offload && ctx->dc->debug.dmub_offload_enabled && !ctx->dc->debug.dmcub_emulation) { - dc_dmub_srv_wait_idle(ctx->dmub_srv); + dc_dmub_srv_wait_for_idle(ctx->dmub_srv, DM_DMUB_WAIT_TYPE_WAIT, NULL); } } char *dce_version_to_string(const int version) { switch (version) { + case DCE_VERSION_6_0: + return "DCE 6.0"; + case DCE_VERSION_6_1: + return "DCE 6.1"; + case DCE_VERSION_6_4: + return "DCE 6.4"; case DCE_VERSION_8_0: return "DCE 8.0"; case DCE_VERSION_8_1: diff --git a/drivers/gpu/drm/amd/display/dc/dc_plane.h b/drivers/gpu/drm/amd/display/dc/dc_plane.h index e9413685ed4f..14feb843e694 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_plane.h +++ b/drivers/gpu/drm/amd/display/dc/dc_plane.h @@ -28,13 +28,24 @@ #include "dc_hw_types.h" +union dc_plane_status_update_flags { + struct { + uint32_t address : 1; + } bits; + uint32_t raw; +}; + struct dc_plane_state *dc_create_plane_state(const struct dc *dc); const struct dc_plane_status *dc_plane_get_status( - const struct dc_plane_state *plane_state); + const struct dc_plane_state *plane_state, + union dc_plane_status_update_flags flags); void dc_plane_state_retain(struct dc_plane_state *plane_state); void dc_plane_state_release(struct dc_plane_state *plane_state); void dc_plane_force_dcc_and_tiling_disable(struct dc_plane_state *plane_state, bool clear_tiling); + +void dc_plane_copy_config(struct dc_plane_state *dst, const struct dc_plane_state *src); + #endif /* _DC_PLANE_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_state_priv.h b/drivers/gpu/drm/amd/display/dc/dc_state_priv.h index 1a12ef579ff4..1d9bae56ff6a 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_state_priv.h +++ b/drivers/gpu/drm/amd/display/dc/dc_state_priv.h @@ -105,4 +105,24 @@ bool dc_state_is_fams2_in_use( const struct dc *dc, const struct dc_state *state); + +void dc_state_set_stream_subvp_cursor_limit(const struct dc_stream_state *stream, + struct dc_state *state, + bool limit); + +bool dc_state_get_stream_subvp_cursor_limit(const struct dc_stream_state *stream, + struct dc_state *state); + +void dc_state_set_stream_cursor_subvp_limit(const struct dc_stream_state *stream, + struct dc_state *state, + bool limit); + +bool dc_state_get_stream_cursor_subvp_limit(const struct dc_stream_state *stream, + struct dc_state *state); + +bool dc_state_can_clear_stream_cursor_subvp_limit(const struct dc_stream_state *stream, + struct dc_state *state); + +bool dc_state_is_subvp_in_use(struct dc_state *state); + #endif /* _DC_STATE_PRIV_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index e0bfddaa23e3..341d2ffb64b1 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -44,6 +44,8 @@ struct mall_stream_config { */ enum mall_stream_type type; struct dc_stream_state *paired_stream; // master / slave stream + bool subvp_limit_cursor_size; /* stream has/is using subvp limiting hw cursor support */ + bool cursor_size_limit_subvp; /* stream is using hw cursor config preventing subvp */ }; struct dc_stream_status { @@ -503,6 +505,11 @@ void program_cursor_position( struct dc *dc, struct dc_stream_state *stream); +bool dc_stream_check_cursor_attributes( + const struct dc_stream_state *stream, + struct dc_state *state, + const struct dc_cursor_attributes *attributes); + bool dc_stream_set_cursor_attributes( struct dc_stream_state *stream, const struct dc_cursor_attributes *attributes); @@ -579,4 +586,8 @@ void dc_dmub_update_dirty_rect(struct dc *dc, struct dc_stream_state *stream, struct dc_surface_update *srf_updates, struct dc_state *context); + +bool dc_stream_is_cursor_limit_pending(struct dc *dc, struct dc_stream_state *stream); +bool dc_stream_can_clear_cursor_limit(struct dc *dc, struct dc_stream_state *stream); + #endif /* DC_STREAM_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 83ffaae9f439..a4cd0eb39a3a 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -210,6 +210,7 @@ struct dc_edid_caps { bool edid_hdmi; bool hdr_supported; + bool rr_capable; struct dc_panel_patch panel_patch; }; @@ -1089,7 +1090,8 @@ union replay_low_refresh_rate_enable_options { struct { //BIT[0-3]: Replay Low Hz Support control unsigned int ENABLE_LOW_RR_SUPPORT :1; - unsigned int RESERVED_1_3 :3; + unsigned int SKIP_ASIC_CHECK :1; + unsigned int RESERVED_2_3 :2; //BIT[4-15]: Replay Low Hz Enable Scenarios unsigned int ENABLE_STATIC_SCREEN :1; unsigned int ENABLE_FULL_SCREEN_VIDEO :1; @@ -1129,6 +1131,10 @@ struct replay_config { union replay_low_refresh_rate_enable_options low_rr_enable_options; /* Replay coasting vtotal is within low refresh rate range. */ bool low_rr_activated; + /* Replay low refresh rate supported*/ + bool low_rr_supported; + /* Replay Video Conferencing Optimization Enabled */ + bool replay_video_conferencing_optimization_enabled; }; /* Replay feature flags*/ @@ -1249,6 +1255,7 @@ enum dc_cm2_gpu_mem_layout { enum dc_cm2_gpu_mem_pixel_component_order { DC_CM2_GPU_MEM_PIXEL_COMPONENT_ORDER_RGBA, + DC_CM2_GPU_MEM_PIXEL_COMPONENT_ORDER_BGRA }; enum dc_cm2_gpu_mem_format { @@ -1270,7 +1277,8 @@ struct dc_cm2_gpu_mem_format_parameters { enum dc_cm2_gpu_mem_size { DC_CM2_GPU_MEM_SIZE_171717, - DC_CM2_GPU_MEM_SIZE_TRANSFORMED + DC_CM2_GPU_MEM_SIZE_333333, + DC_CM2_GPU_MEM_SIZE_TRANSFORMED, }; struct dc_cm2_gpu_mem_parameters { @@ -1279,6 +1287,7 @@ struct dc_cm2_gpu_mem_parameters { struct dc_cm2_gpu_mem_format_parameters format_params; enum dc_cm2_gpu_mem_pixel_component_order component_order; enum dc_cm2_gpu_mem_size size; + uint16_t bit_depth; }; enum dc_cm2_transfer_func_source { @@ -1302,6 +1311,10 @@ struct dc_cm2_func_luts { const struct dc_3dlut *lut3d_func; struct dc_cm2_gpu_mem_parameters gpu_mem_params; }; + bool rmcm_3dlut_shaper_select; + bool mpc_3dlut_enable; + bool rmcm_3dlut_enable; + bool mpc_mcm_post_blend; } lut3d_data; const struct dc_transfer_func *lut1d_func; }; diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c index b363f5360818..58c84f555c0f 100644 --- a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c +++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c @@ -391,6 +391,7 @@ static void dccg35_set_dppclk_rcg(struct dccg *dccg, struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && enable) return; @@ -411,6 +412,8 @@ static void dccg35_set_dppclk_rcg(struct dccg *dccg, BREAK_TO_DEBUGGER(); break; } + //DC_LOG_DEBUG("%s: inst(%d) DPPCLK rcg_disable: %d\n", __func__, inst, enable ? 0 : 1); + } static void dccg35_set_dpstreamclk_rcg( @@ -1035,6 +1038,7 @@ static void dccg35_enable_dpp_clk_new( DPPCLK0_DTO_MODULO, 0xFF); } + static void dccg35_disable_dpp_clk_new( struct dccg *dccg, int inst) @@ -1112,30 +1116,24 @@ static void dcn35_set_dppclk_enable(struct dccg *dccg, { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + switch (dpp_inst) { case 0: REG_UPDATE(DPPCLK_CTRL, DPPCLK0_EN, enable); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, enable); break; case 1: REG_UPDATE(DPPCLK_CTRL, DPPCLK1_EN, enable); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, enable); break; case 2: REG_UPDATE(DPPCLK_CTRL, DPPCLK2_EN, enable); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, enable); break; case 3: REG_UPDATE(DPPCLK_CTRL, DPPCLK3_EN, enable); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, enable); break; default: break; } + //DC_LOG_DEBUG("%s: dpp_inst(%d) DPPCLK_EN = %d\n", __func__, dpp_inst, enable); } @@ -1163,14 +1161,18 @@ static void dccg35_update_dpp_dto(struct dccg *dccg, int dpp_inst, ASSERT(false); phase = 0xff; } + dccg35_set_dppclk_rcg(dccg, dpp_inst, false); REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0, DPPCLK0_DTO_PHASE, phase, DPPCLK0_DTO_MODULO, modulo); dcn35_set_dppclk_enable(dccg, dpp_inst, true); - } else + } else { dcn35_set_dppclk_enable(dccg, dpp_inst, false); + /*we have this in hwss: disable_plane*/ + //dccg35_set_dppclk_rcg(dccg, dpp_inst, true); + } dccg->pipe_dppclk_khz[dpp_inst] = req_dppclk; } @@ -1182,6 +1184,7 @@ static void dccg35_set_dppclk_root_clock_gating(struct dccg *dccg, if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) return; + switch (dpp_inst) { case 0: REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, enable); @@ -1198,6 +1201,8 @@ static void dccg35_set_dppclk_root_clock_gating(struct dccg *dccg, default: break; } + //DC_LOG_DEBUG("%s: dpp_inst(%d) rcg: %d\n", __func__, dpp_inst, enable); + } static void dccg35_get_pixel_rate_div( @@ -1521,28 +1526,30 @@ static void dccg35_set_physymclk_root_clock_gating( switch (phy_inst) { case 0: REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, - PHYASYMCLK_ROOT_GATE_DISABLE, enable ? 1 : 0); + PHYASYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 1: REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, - PHYBSYMCLK_ROOT_GATE_DISABLE, enable ? 1 : 0); + PHYBSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 2: REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, - PHYCSYMCLK_ROOT_GATE_DISABLE, enable ? 1 : 0); + PHYCSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 3: REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, - PHYDSYMCLK_ROOT_GATE_DISABLE, enable ? 1 : 0); + PHYDSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 4: REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, - PHYESYMCLK_ROOT_GATE_DISABLE, enable ? 1 : 0); + PHYESYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1); break; default: BREAK_TO_DEBUGGER(); return; } + //DC_LOG_DEBUG("%s: dpp_inst(%d) PHYESYMCLK_ROOT_GATE_DISABLE:\n", __func__, phy_inst, enable ? 0 : 1); + } static void dccg35_set_physymclk( @@ -1643,6 +1650,8 @@ static void dccg35_dpp_root_clock_control( return; if (clock_on) { + dccg35_set_dppclk_rcg(dccg, dpp_inst, false); + /* turn off the DTO and leave phase/modulo at max */ dcn35_set_dppclk_enable(dccg, dpp_inst, 1); REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0, @@ -1654,6 +1663,8 @@ static void dccg35_dpp_root_clock_control( REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0, DPPCLK0_DTO_PHASE, 0, DPPCLK0_DTO_MODULO, 1); + /*we have this in hwss: disable_plane*/ + //dccg35_set_dppclk_rcg(dccg, dpp_inst, true); } dccg->dpp_clock_gated[dpp_inst] = !clock_on; @@ -1771,36 +1782,40 @@ static void dccg35_enable_dscclk(struct dccg *dccg, int inst) //Disable DTO switch (inst) { case 0: + if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, 1); + REG_UPDATE_2(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_PHASE, 0, DSCCLK0_DTO_MODULO, 0); REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, 1); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, 1); break; case 1: + if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, 1); + REG_UPDATE_2(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_PHASE, 0, DSCCLK1_DTO_MODULO, 0); REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, 1); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, 1); break; case 2: + if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, 1); + REG_UPDATE_2(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_PHASE, 0, DSCCLK2_DTO_MODULO, 0); REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, 1); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, 1); break; case 3: + if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, 1); + REG_UPDATE_2(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_PHASE, 0, DSCCLK3_DTO_MODULO, 0); REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK3_EN, 1); - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, 1); break; default: BREAK_TO_DEBUGGER(); @@ -1813,9 +1828,6 @@ static void dccg35_disable_dscclk(struct dccg *dccg, { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - return; - switch (inst) { case 0: REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, 0); diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c index 077337698e0a..b4f5b4a6331a 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c @@ -976,11 +976,12 @@ static bool dcn31_program_pix_clk( struct bp_pixel_clock_parameters bp_pc_params = {0}; enum transmitter_color_depth bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24; - // Apply ssed(spread spectrum) dpref clock for edp only. - if (clock_source->ctx->dc->clk_mgr->dp_dto_source_clock_in_khz != 0 - && pix_clk_params->signal_type == SIGNAL_TYPE_EDP - && encoding == DP_8b_10b_ENCODING) + // Apply ssed(spread spectrum) dpref clock for edp and dp + if (clock_source->ctx->dc->clk_mgr->dp_dto_source_clock_in_khz != 0 && + dc_is_dp_signal(pix_clk_params->signal_type) && + encoding == DP_8b_10b_ENCODING) dp_dto_ref_khz = clock_source->ctx->dc->clk_mgr->dp_dto_source_clock_in_khz; + // For these signal types Driver to program DP_DTO without calling VBIOS Command table if (dc_is_dp_signal(pix_clk_params->signal_type) || dc_is_virtual_signal(pix_clk_params->signal_type)) { if (e) { diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h index 0721ae895ae9..94128f7a18b1 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h @@ -257,7 +257,7 @@ bool dce110_clk_src_construct( struct dce110_clk_src *clk_src, struct dc_context *ctx, struct dc_bios *bios, - enum clock_source_id, + enum clock_source_id id, const struct dce110_clk_src_regs *regs, const struct dce110_clk_src_shift *cs_shift, const struct dce110_clk_src_mask *cs_mask); diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c index ccc154b0281c..3b9011ef9b68 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c @@ -28,6 +28,8 @@ #include "dc.h" #include "core_types.h" #include "dmub_cmd.h" +#include "dc_dmub_srv.h" +#include "dmub/dmub_srv.h" #define TO_DMUB_ABM(abm)\ container_of(abm, struct dce_abm, base) diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c index 0d7e7f3b81a1..a641ae04450c 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c @@ -240,7 +240,8 @@ bool dmub_abm_save_restore( cmd.abm_save_restore.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; cmd.abm_save_restore.abm_init_config_data.panel_mask = panel_mask; - cmd.abm_save_restore.header.payload_bytes = sizeof(struct dmub_rb_cmd_abm_save_restore); + cmd.abm_save_restore.header.payload_bytes = + sizeof(struct dmub_rb_cmd_abm_save_restore) - sizeof(struct dmub_cmd_header); dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c index c31e4f26a305..fcd3d86ad517 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c @@ -280,7 +280,9 @@ static void dmub_replay_set_power_opt_and_coasting_vtotal(struct dmub_replay *dm memset(&cmd, 0, sizeof(cmd)); pCmd->header.type = DMUB_CMD__REPLAY; pCmd->header.sub_type = DMUB_CMD__REPLAY_SET_POWER_OPT_AND_COASTING_VTOTAL; - pCmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_replay_set_power_opt_and_coasting_vtotal); + pCmd->header.payload_bytes = + sizeof(struct dmub_rb_cmd_replay_set_power_opt_and_coasting_vtotal) - + sizeof(struct dmub_cmd_header); pCmd->replay_set_power_opt_data.power_opt = power_opt; pCmd->replay_set_power_opt_data.panel_inst = panel_inst; pCmd->replay_set_coasting_vtotal_data.coasting_vtotal = (coasting_vtotal & 0xFFFF); @@ -319,7 +321,8 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub, cmd.replay_set_timing_sync.header.sub_type = DMUB_CMD__REPLAY_SET_TIMING_SYNC_SUPPORTED; cmd.replay_set_timing_sync.header.payload_bytes = - sizeof(struct dmub_rb_cmd_replay_set_timing_sync); + sizeof(struct dmub_rb_cmd_replay_set_timing_sync) - + sizeof(struct dmub_cmd_header); //Cmd Body cmd.replay_set_timing_sync.replay_set_timing_sync_data.panel_inst = cmd_element->sync_data.panel_inst; @@ -331,7 +334,8 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub, cmd.replay_set_frameupdate_timer.header.sub_type = DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER; cmd.replay_set_frameupdate_timer.header.payload_bytes = - sizeof(struct dmub_rb_cmd_replay_set_frameupdate_timer); + sizeof(struct dmub_rb_cmd_replay_set_frameupdate_timer) - + sizeof(struct dmub_cmd_header); //Cmd Body cmd.replay_set_frameupdate_timer.data.panel_inst = cmd_element->panel_inst; @@ -345,7 +349,8 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub, cmd.replay_set_pseudo_vtotal.header.sub_type = DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL; cmd.replay_set_pseudo_vtotal.header.payload_bytes = - sizeof(struct dmub_rb_cmd_replay_set_pseudo_vtotal); + sizeof(struct dmub_rb_cmd_replay_set_pseudo_vtotal) - + sizeof(struct dmub_cmd_header); //Cmd Body cmd.replay_set_pseudo_vtotal.data.panel_inst = cmd_element->pseudo_vtotal_data.panel_inst; @@ -357,7 +362,8 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub, cmd.replay_disabled_adaptive_sync_sdp.header.sub_type = DMUB_CMD__REPLAY_DISABLED_ADAPTIVE_SYNC_SDP; cmd.replay_disabled_adaptive_sync_sdp.header.payload_bytes = - sizeof(struct dmub_rb_cmd_replay_disabled_adaptive_sync_sdp); + sizeof(struct dmub_rb_cmd_replay_disabled_adaptive_sync_sdp) - + sizeof(struct dmub_cmd_header); //Cmd Body cmd.replay_disabled_adaptive_sync_sdp.data.panel_inst = cmd_element->disabled_adaptive_sync_sdp_data.panel_inst; @@ -369,7 +375,8 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub, cmd.replay_set_general_cmd.header.sub_type = DMUB_CMD__REPLAY_SET_GENERAL_CMD; cmd.replay_set_general_cmd.header.payload_bytes = - sizeof(struct dmub_rb_cmd_replay_set_general_cmd); + sizeof(struct dmub_rb_cmd_replay_set_general_cmd) - + sizeof(struct dmub_cmd_header); //Cmd Body cmd.replay_set_general_cmd.data.panel_inst = cmd_element->set_general_cmd_data.panel_inst; diff --git a/drivers/gpu/drm/amd/display/dc/dce60/Makefile b/drivers/gpu/drm/amd/display/dc/dce60/Makefile index eede83ad91fa..824f73eb3326 100644 --- a/drivers/gpu/drm/amd/display/dc/dce60/Makefile +++ b/drivers/gpu/drm/amd/display/dc/dce60/Makefile @@ -25,8 +25,7 @@ CFLAGS_$(AMDDALPATH)/dc/dce60/dce60_resource.o = -Wno-override-init -DCE60 = dce60_timing_generator.o dce60_hw_sequencer.o \ - dce60_resource.o +DCE60 = dce60_timing_generator.o AMD_DAL_DCE60 = $(addprefix $(AMDDALPATH)/dc/dce60/,$(DCE60)) diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c index 003a9330c286..88e7a1fc9a30 100644 --- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c +++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c @@ -105,7 +105,7 @@ static void program_pix_dur(struct timing_generator *tg, uint32_t pix_clk_100hz) dm_write_reg(tg->ctx, addr, value); } -static void program_timing(struct timing_generator *tg, +static void dce80_timing_generator_program_timing(struct timing_generator *tg, const struct dc_crtc_timing *timing, int vready_offset, int vstartup_start, @@ -185,7 +185,7 @@ static void dce80_timing_generator_enable_advanced_request( static const struct timing_generator_funcs dce80_tg_funcs = { .validate_timing = dce110_tg_validate_timing, - .program_timing = program_timing, + .program_timing = dce80_timing_generator_program_timing, .enable_crtc = dce110_timing_generator_enable_crtc, .disable_crtc = dce110_timing_generator_disable_crtc, .is_counter_moving = dce110_timing_generator_is_counter_moving, diff --git a/drivers/gpu/drm/amd/display/dc/dm_helpers.h b/drivers/gpu/drm/amd/display/dc/dm_helpers.h index 5efddd48d5c5..9d160b39e8c5 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_helpers.h +++ b/drivers/gpu/drm/amd/display/dc/dm_helpers.h @@ -153,6 +153,14 @@ bool dm_helpers_submit_i2c( const struct dc_link *link, struct i2c_command *cmd); +bool dm_helpers_execute_fused_io( + struct dc_context *ctx, + struct dc_link *link, + union dmub_rb_cmd *commands, + uint8_t count, + uint32_t timeout_us +); + bool dm_helpers_dp_write_dsc_enable( struct dc_context *ctx, const struct dc_stream_state *stream, diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c index f1fe49401bc0..8d24763938ea 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c @@ -1002,6 +1002,7 @@ static bool CalculatePrefetchSchedule( dst_y_prefetch_equ = VStartup - (Tsetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime - (*DSTYAfterScaler + *DSTXAfterScaler / myPipe->HTotal); + dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH Lsw_oto = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC); Tsw_oto = Lsw_oto * LineTime; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index f567a9023682..ed59c77bc6f6 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -1105,6 +1105,7 @@ static bool CalculatePrefetchSchedule( Tr0_oto_lines = dml_ceil(4.0 * Tr0_oto / LineTime, 1) / 4.0; dst_y_prefetch_oto = Tvm_oto_lines + 2 * Tr0_oto_lines + Lsw_oto; dst_y_prefetch_equ = VStartup - (*TSetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime - (*DSTYAfterScaler + *DSTXAfterScaler / myPipe->HTotal); + dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH dst_y_prefetch_equ = dml_floor(4.0 * (dst_y_prefetch_equ + 0.125), 1) / 4.0; Tpre_rounded = dst_y_prefetch_equ * LineTime; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c index 5865e8fa2d8e..9f3938a50240 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c @@ -1123,6 +1123,7 @@ static bool CalculatePrefetchSchedule( Tr0_oto_lines = dml_ceil(4.0 * Tr0_oto / LineTime, 1) / 4.0; dst_y_prefetch_oto = Tvm_oto_lines + 2 * Tr0_oto_lines + Lsw_oto; dst_y_prefetch_equ = VStartup - (*TSetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime - (*DSTYAfterScaler + *DSTXAfterScaler / myPipe->HTotal); + dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH dst_y_prefetch_equ = dml_floor(4.0 * (dst_y_prefetch_equ + 0.125), 1) / 4.0; Tpre_rounded = dst_y_prefetch_equ * LineTime; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c index 56dda686e299..b0fc1fd20208 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c @@ -627,6 +627,7 @@ static bool dcn32_assign_subvp_pipe(struct dc *dc, */ if (pipe->plane_state && !pipe->top_pipe && !pipe->prev_odm_pipe && !dcn32_is_center_timing(pipe) && !pipe->stream->hw_cursor_req && + !dc_state_get_stream_cursor_subvp_limit(pipe->stream, context) && !(pipe->stream->timing.pix_clk_100hz / 10000 > DCN3_2_MAX_SUBVP_PIXEL_RATE_MHZ) && (!dcn32_is_psr_capable(pipe) || (context->stream_count == 1 && dc->caps.dmub_caps.subvp_psr)) && dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_NONE && diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c index d9159ca55412..92f0a099d089 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c @@ -195,9 +195,9 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .dcn_downspread_percent = 0.5, .gpuvm_min_page_size_bytes = 4096, .hostvm_min_page_size_bytes = 4096, - .do_urgent_latency_adjustment = 1, + .do_urgent_latency_adjustment = 0, .urgent_latency_adjustment_fabric_clock_component_us = 0, - .urgent_latency_adjustment_fabric_clock_reference_mhz = 3000, + .urgent_latency_adjustment_fabric_clock_reference_mhz = 0, }; void dcn35_build_wm_range_table_fpu(struct clk_mgr *clk_mgr) diff --git a/drivers/gpu/drm/amd/display/dc/dml2/Makefile b/drivers/gpu/drm/amd/display/dc/dml2/Makefile index 21fd466dba26..157ecf008d6c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/Makefile +++ b/drivers/gpu/drm/amd/display/dc/dml2/Makefile @@ -99,7 +99,6 @@ CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_standalone_libraries/lib_float_math. CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml21_wrapper.o := $(dml2_ccflags) CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/dml21_translation_helper.o := $(dml2_ccflags) CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/dml21_utils.o := $(dml2_ccflags) -CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/inc/dml2_debug.o := $(dml2_ccflags) CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.o := $(dml2_rcflags) CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.o := $(dml2_rcflags) @@ -117,11 +116,9 @@ CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_standalone_libraries/lib_floa CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml21_wrapper.o := $(dml2_rcflags) CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/dml21_translation_helper.o := $(dml2_rcflags) CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/dml21_utils.o := $(dml2_rcflags) -CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/inc/dml2_debug.o := $(dml2_rcflags) DML21 := src/dml2_top/dml2_top_interfaces.o DML21 += src/dml2_top/dml2_top_soc15.o -DML21 += src/inc/dml2_debug.o DML21 += src/dml2_core/dml2_core_dcn4.o DML21 += src/dml2_core/dml2_core_factory.o DML21 += src/dml2_core/dml2_core_dcn4_calcs.o diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c index 0c8ec30ea672..d47cacfdb695 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c @@ -526,7 +526,8 @@ static void populate_dml21_output_config_from_stream_state(struct dml2_link_outp static void populate_dml21_stream_overrides_from_stream_state( struct dml2_stream_parameters *stream_desc, - struct dc_stream_state *stream) + struct dc_stream_state *stream, + struct dc_stream_status *stream_status) { switch (stream->debug.force_odm_combine_segments) { case 0: @@ -551,7 +552,9 @@ static void populate_dml21_stream_overrides_from_stream_state( if (!stream->ctx->dc->debug.enable_single_display_2to1_odm_policy || stream->debug.force_odm_combine_segments > 0) stream_desc->overrides.disable_dynamic_odm = true; - stream_desc->overrides.disable_subvp = stream->ctx->dc->debug.force_disable_subvp || stream->hw_cursor_req; + stream_desc->overrides.disable_subvp = stream->ctx->dc->debug.force_disable_subvp || + stream->hw_cursor_req || + stream_status->mall_stream_config.cursor_size_limit_subvp; } static enum dml2_swizzle_mode gfx_addr3_to_dml2_swizzle_mode(enum swizzle_mode_addr3_values addr3_mode) @@ -885,6 +888,9 @@ static void populate_dml21_plane_config_from_plane_state(struct dml2_context *dm case DC_CM2_GPU_MEM_SIZE_171717: plane->tdlut.tdlut_width_mode = dml2_tdlut_width_17_cube; break; + case DC_CM2_GPU_MEM_SIZE_333333: + plane->tdlut.tdlut_width_mode = dml2_tdlut_width_33_cube; + break; case DC_CM2_GPU_MEM_SIZE_TRANSFORMED: //plane->tdlut.tdlut_width_mode = dml2_tdlut_width_flatten; // dml2_tdlut_width_flatten undefined break; @@ -910,7 +916,7 @@ static void populate_dml21_plane_config_from_plane_state(struct dml2_context *dm } //TODO : Could be possibly moved to a common helper layer. -static bool dml21_wrapper_get_plane_id(const struct dc_state *context, const struct dc_plane_state *plane, unsigned int *plane_id) +static bool dml21_wrapper_get_plane_id(const struct dc_state *context, unsigned int stream_id, const struct dc_plane_state *plane, unsigned int *plane_id) { int i, j; @@ -918,10 +924,12 @@ static bool dml21_wrapper_get_plane_id(const struct dc_state *context, const str return false; for (i = 0; i < context->stream_count; i++) { - for (j = 0; j < context->stream_status[i].plane_count; j++) { - if (context->stream_status[i].plane_states[j] == plane) { - *plane_id = (i << 16) | j; - return true; + if (context->streams[i]->stream_id == stream_id) { + for (j = 0; j < context->stream_status[i].plane_count; j++) { + if (context->stream_status[i].plane_states[j] == plane) { + *plane_id = (i << 16) | j; + return true; + } } } } @@ -944,14 +952,14 @@ static unsigned int map_stream_to_dml21_display_cfg(const struct dml2_context *d return location; } -static unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx, +unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx, unsigned int stream_id, const struct dc_plane_state *plane, const struct dc_state *context) { unsigned int plane_id; int i = 0; int location = -1; - if (!dml21_wrapper_get_plane_id(context, plane, &plane_id)) { + if (!dml21_wrapper_get_plane_id(context, stream_id, plane, &plane_id)) { ASSERT(false); return -1; } @@ -1021,7 +1029,7 @@ bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_s populate_dml21_timing_config_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].timing, context->streams[stream_index], dml_ctx); adjust_dml21_hblank_timing_config_from_pipe_ctx(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].timing, &context->res_ctx.pipe_ctx[stream_index]); populate_dml21_output_config_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].output, context->streams[stream_index], &context->res_ctx.pipe_ctx[stream_index]); - populate_dml21_stream_overrides_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location], context->streams[stream_index]); + populate_dml21_stream_overrides_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location], context->streams[stream_index], &context->stream_status[stream_index]); dml_dispcfg->stream_descriptors[disp_cfg_stream_location].overrides.hw.twait_budgeting.fclk_pstate = dml2_twait_budgeting_setting_if_needed; dml_dispcfg->stream_descriptors[disp_cfg_stream_location].overrides.hw.twait_budgeting.uclk_pstate = dml2_twait_budgeting_setting_if_needed; @@ -1037,7 +1045,7 @@ bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_s dml_dispcfg->plane_descriptors[disp_cfg_plane_location].stream_index = disp_cfg_stream_location; } else { for (plane_index = 0; plane_index < context->stream_status[stream_index].plane_count; plane_index++) { - disp_cfg_plane_location = map_plane_to_dml21_display_cfg(dml_ctx, context->stream_status[stream_index].plane_states[plane_index], context); + disp_cfg_plane_location = map_plane_to_dml21_display_cfg(dml_ctx, context->streams[stream_index]->stream_id, context->stream_status[stream_index].plane_states[plane_index], context); if (disp_cfg_plane_location < 0) disp_cfg_plane_location = dml_dispcfg->num_planes++; @@ -1048,7 +1056,7 @@ bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_s populate_dml21_plane_config_from_plane_state(dml_ctx, &dml_dispcfg->plane_descriptors[disp_cfg_plane_location], context->stream_status[stream_index].plane_states[plane_index], context, stream_index); dml_dispcfg->plane_descriptors[disp_cfg_plane_location].stream_index = disp_cfg_stream_location; - if (dml21_wrapper_get_plane_id(context, context->stream_status[stream_index].plane_states[plane_index], &dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[disp_cfg_plane_location])) + if (dml21_wrapper_get_plane_id(context, context->streams[stream_index]->stream_id, context->stream_status[stream_index].plane_states[plane_index], &dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[disp_cfg_plane_location])) dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[disp_cfg_plane_location] = true; /* apply forced pstate policy */ diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h index 069b939c672a..73a013be1e48 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h @@ -11,6 +11,7 @@ struct dc_state; struct dcn_watermarks; union dcn_watermark_set; struct pipe_ctx; +struct dc_plane_state; struct dml2_context; struct dml2_configuration_options; @@ -25,4 +26,5 @@ void dml21_extract_watermark_sets(const struct dc *in_dc, union dcn_watermark_se void dml21_map_hw_resources(struct dml2_context *dml_ctx); void dml21_get_pipe_mcache_config(struct dc_state *context, struct pipe_ctx *pipe_ctx, struct dml2_per_plane_programming *pln_prog, struct dml2_pipe_configuration_descriptor *mcache_pipe_config); void dml21_set_dc_p_state_type(struct pipe_ctx *pipe_ctx, struct dml2_per_stream_programming *stream_programming, bool sub_vp_enabled); +unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx, unsigned int stream_id, const struct dc_plane_state *plane, const struct dc_state *context); #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c index 5d16f36ec95c..208d3651b6ba 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c @@ -12,6 +12,8 @@ #include "dml21_translation_helper.h" #include "dml2_dc_resource_mgmt.h" +#define INVALID -1 + static bool dml21_allocate_memory(struct dml2_context **dml_ctx) { *dml_ctx = vzalloc(sizeof(struct dml2_context)); @@ -208,10 +210,40 @@ static void dml21_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_sta } } +static void dml21_prepare_mcache_params(struct dml2_context *dml_ctx, struct dc_state *context, struct dc_mcache_params *mcache_params) +{ + int dc_plane_idx = 0; + int dml_prog_idx, stream_idx, plane_idx; + struct dml2_per_plane_programming *pln_prog = NULL; + + for (stream_idx = 0; stream_idx < context->stream_count; stream_idx++) { + for (plane_idx = 0; plane_idx < context->stream_status[stream_idx].plane_count; plane_idx++) { + dml_prog_idx = map_plane_to_dml21_display_cfg(dml_ctx, context->streams[stream_idx]->stream_id, context->stream_status[stream_idx].plane_states[plane_idx], context); + if (dml_prog_idx == INVALID) { + continue; + } + pln_prog = &dml_ctx->v21.mode_programming.programming->plane_programming[dml_prog_idx]; + mcache_params[dc_plane_idx].valid = pln_prog->mcache_allocation.valid; + mcache_params[dc_plane_idx].num_mcaches_plane0 = pln_prog->mcache_allocation.num_mcaches_plane0; + mcache_params[dc_plane_idx].num_mcaches_plane1 = pln_prog->mcache_allocation.num_mcaches_plane1; + mcache_params[dc_plane_idx].requires_dedicated_mall_mcache = pln_prog->mcache_allocation.requires_dedicated_mall_mcache; + mcache_params[dc_plane_idx].last_slice_sharing.plane0_plane1 = pln_prog->mcache_allocation.last_slice_sharing.plane0_plane1; + memcpy(mcache_params[dc_plane_idx].mcache_x_offsets_plane0, + pln_prog->mcache_allocation.mcache_x_offsets_plane0, + sizeof(int) * (DML2_MAX_MCACHES + 1)); + memcpy(mcache_params[dc_plane_idx].mcache_x_offsets_plane1, + pln_prog->mcache_allocation.mcache_x_offsets_plane1, + sizeof(int) * (DML2_MAX_MCACHES + 1)); + dc_plane_idx++; + } + } +} + static bool dml21_mode_check_and_programming(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx) { bool result = false; struct dml2_build_mode_programming_in_out *mode_programming = &dml_ctx->v21.mode_programming; + struct dc_mcache_params mcache_params[MAX_PLANES] = {0}; memset(&dml_ctx->v21.display_config, 0, sizeof(struct dml2_display_cfg)); memset(&dml_ctx->v21.dml_to_dc_pipe_mapping, 0, sizeof(struct dml2_dml_to_dc_pipe_mapping)); @@ -234,7 +266,9 @@ static bool dml21_mode_check_and_programming(const struct dc *in_dc, struct dc_s if (!result) return false; + DC_FP_START(); result = dml2_build_mode_programming(mode_programming); + DC_FP_END(); if (!result) return false; @@ -244,6 +278,14 @@ static bool dml21_mode_check_and_programming(const struct dc *in_dc, struct dc_s dml2_map_dc_pipes(dml_ctx, context, NULL, &dml_ctx->v21.dml_to_dc_pipe_mapping, in_dc->current_state); /* if subvp phantoms are present, expand them into dc context */ dml21_handle_phantom_streams_planes(in_dc, context, dml_ctx); + + if (in_dc->res_pool->funcs->program_mcache_pipe_config) { + //Prepare mcache params for each plane based on mcache output from DML + dml21_prepare_mcache_params(dml_ctx, context, mcache_params); + + //populate mcache regs to each pipe + dml_ctx->config.callbacks.allocate_mcache(context, mcache_params); + } } /* Copy DML CLK, WM and REG outputs to bandwidth context */ @@ -277,7 +319,9 @@ static bool dml21_check_mode_support(const struct dc *in_dc, struct dc_state *co mode_support->dml2_instance = dml_init->dml2_instance; dml21_map_dc_state_into_dml_display_cfg(in_dc, context, dml_ctx); dml_ctx->v21.mode_programming.dml2_instance->scratch.build_mode_programming_locals.mode_programming_params.programming = dml_ctx->v21.mode_programming.programming; + DC_FP_START(); is_supported = dml2_check_mode_supported(mode_support); + DC_FP_END(); if (!is_supported) return false; @@ -288,16 +332,12 @@ bool dml21_validate(const struct dc *in_dc, struct dc_state *context, struct dml { bool out = false; - DC_FP_START(); - /* Use dml_validate_only for fast_validate path */ if (fast_validate) out = dml21_check_mode_support(in_dc, context, dml_ctx); else out = dml21_mode_check_and_programming(in_dc, context, dml_ctx); - DC_FP_END(); - return out; } diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.h index b2075b8c363b..42e715024bc9 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.h @@ -8,6 +8,7 @@ #include "os_types.h" #include "dml_top_soc_parameter_types.h" +#include "dml_top_display_cfg_types.h" struct dc; struct dc_state; @@ -65,4 +66,67 @@ struct socbb_ip_params_external { struct dml2_ip_capabilities ip_params; struct dml2_soc_bb soc_bb; }; + +/*mcache parameters decided by dml*/ +struct dc_mcache_params { + bool valid; + /* + * For iMALL, dedicated mall mcaches are required (sharing of last + * slice possible), for legacy phantom or phantom without return + * the only mall mcaches need to be valid. + */ + bool requires_dedicated_mall_mcache; + unsigned int num_mcaches_plane0; + unsigned int num_mcaches_plane1; + /* + * Generally, plane0/1 slices must use a disjoint set of caches + * but in some cases the final segement of the two planes can + * use the same cache. If plane0_plane1 is set, then this is + * allowed. + * + * Similarly, the caches allocated to MALL prefetcher are generally + * disjoint, but if mall_prefetch is set, then the final segment + * between the main and the mall pixel requestor can use the same + * cache. + * + * Note that both bits may be set at the same time. + */ + struct { + bool mall_comb_mcache_p0; + bool mall_comb_mcache_p1; + bool plane0_plane1; + } last_slice_sharing; + /* + * A plane is divided into vertical slices of mcaches, + * which wrap on the surface width. + * + * For example, if the surface width is 7680, and split into + * three slices of equal width, the boundary array would contain + * [2560, 5120, 7680] + * + * The assignments are + * 0 = [0 .. 2559] + * 1 = [2560 .. 5119] + * 2 = [5120 .. 7679] + * 0 = [7680 .. INF] + * The final element implicitly is the same as the first, and + * at first seems invalid since it is never referenced (since) + * it is outside the surface. However, its useful when shifting + * (see below). + * + * For any given valid mcache assignment, a shifted version, wrapped + * on the surface width boundary is also assumed to be valid. + * + * For example, shifting [2560, 5120, 7680] by -50 results in + * [2510, 5170, 7630]. + * + * The assignments are now: + * 0 = [0 .. 2509] + * 1 = [2510 .. 5169] + * 2 = [5170 .. 7629] + * 0 = [7630 .. INF] + */ + int mcache_x_offsets_plane0[DML2_MAX_MCACHES + 1]; + int mcache_x_offsets_plane1[DML2_MAX_MCACHES + 1]; +}; #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top.h index a64ec4dcf11a..c047d56527c4 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top.h @@ -43,4 +43,5 @@ bool dml2_build_mode_programming(struct dml2_build_mode_programming_in_out *in_o */ bool dml2_build_mcache_programming(struct dml2_build_mcache_programming_in_out *in_out); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_dchub_registers.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_dchub_registers.h index 25b607e7b726..84c90050668c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_dchub_registers.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_dchub_registers.h @@ -156,6 +156,8 @@ struct dml2_dchub_watermark_regs { uint32_t urgent; uint32_t sr_enter; uint32_t sr_exit; + uint32_t sr_enter_z8; + uint32_t sr_exit_z8; uint32_t uclk_pstate; uint32_t fclk_pstate; uint32_t temp_read_or_ppt; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h index 5e1ab6d97640..255f05de362c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h @@ -166,7 +166,7 @@ struct dml2_surface_cfg { enum dml2_swizzle_mode tiling; struct { - unsigned long pitch; + unsigned long pitch; // In elements, two pixels per element in 422 packed format unsigned long width; unsigned long height; } plane0; @@ -385,6 +385,7 @@ struct dml2_plane_parameters { long reserved_vblank_time_ns; unsigned int max_vactive_det_fill_delay_us; // 0 = no reserved time, +ve = explicit max delay unsigned int gpuvm_min_page_size_kbytes; + unsigned int hostvm_min_page_size_kbytes; enum dml2_svp_mode_override legacy_svp_config; //TODO remove in favor of svp_config diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c index bb863c8c6b39..6ee37386f672 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c @@ -456,10 +456,10 @@ bool core_dcn4_mode_support(struct dml2_core_mode_support_in_out *in_out) in_out->mode_support_result.global.active.urgent_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] * 1000), 1.0); in_out->mode_support_result.global.svp_prefetch.average_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] * 1000), 1.0); in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] * 1000), 1.0); - dml2_printf("DML::%s: in_out->mode_support_result.global.active.urgent_bw_sdp_kbps = %ld\n", __func__, in_out->mode_support_result.global.active.urgent_bw_sdp_kbps); - dml2_printf("DML::%s: in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps = %ld\n", __func__, in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps); - dml2_printf("DML::%s: in_out->mode_support_result.global.active.urgent_bw_dram_kbps = %ld\n", __func__, in_out->mode_support_result.global.active.urgent_bw_dram_kbps); - dml2_printf("DML::%s: in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps = %ld\n", __func__, in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps); + DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.active.urgent_bw_sdp_kbps = %ld\n", __func__, in_out->mode_support_result.global.active.urgent_bw_sdp_kbps); + DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps = %ld\n", __func__, in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps); + DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.active.urgent_bw_dram_kbps = %ld\n", __func__, in_out->mode_support_result.global.active.urgent_bw_dram_kbps); + DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps = %ld\n", __func__, in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps); for (i = 0; i < l->svp_expanded_display_cfg.num_planes; i++) { in_out->mode_support_result.per_plane[i].dppclk_khz = (unsigned int)(core->clean_me_up.mode_lib.ms.RequiredDPPCLK[i] * 1000); @@ -509,7 +509,7 @@ bool core_dcn4_mode_support(struct dml2_core_mode_support_in_out *in_out) stream_index = l->svp_expanded_display_cfg.plane_descriptors[i].stream_index; in_out->mode_support_result.per_stream[stream_index].dscclk_khz = (unsigned int)core->clean_me_up.mode_lib.ms.required_dscclk_freq_mhz[i] * 1000; - dml2_printf("CORE_DCN4::%s: i=%d stream_index=%d, in_out->mode_support_result.per_stream[stream_index].dscclk_khz = %u\n", __func__, i, stream_index, in_out->mode_support_result.per_stream[stream_index].dscclk_khz); + DML_LOG_VERBOSE("CORE_DCN4::%s: i=%d stream_index=%d, in_out->mode_support_result.per_stream[stream_index].dscclk_khz = %u\n", __func__, i, stream_index, in_out->mode_support_result.per_stream[stream_index].dscclk_khz); if (!((stream_bitmask >> stream_index) & 0x1)) { in_out->mode_support_result.cfg_support_info.stream_support_info[stream_index].odms_used = odm_count; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c index 4c504cb0e1c5..c4dad7164d31 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c @@ -54,104 +54,104 @@ static double dml2_core_div_rem(double dividend, unsigned int divisor, unsigned static void dml2_print_mode_support_info(const struct dml2_core_internal_mode_support_info *support, bool fail_only) { - dml2_printf("DML: ===================================== \n"); - dml2_printf("DML: DML_MODE_SUPPORT_INFO_ST\n"); + DML_LOG_VERBOSE("DML: ===================================== \n"); + DML_LOG_VERBOSE("DML: DML_MODE_SUPPORT_INFO_ST\n"); if (!fail_only || support->ScaleRatioAndTapsSupport == 0) - dml2_printf("DML: support: ScaleRatioAndTapsSupport = %d\n", support->ScaleRatioAndTapsSupport); + DML_LOG_VERBOSE("DML: support: ScaleRatioAndTapsSupport = %d\n", support->ScaleRatioAndTapsSupport); if (!fail_only || support->SourceFormatPixelAndScanSupport == 0) - dml2_printf("DML: support: SourceFormatPixelAndScanSupport = %d\n", support->SourceFormatPixelAndScanSupport); + DML_LOG_VERBOSE("DML: support: SourceFormatPixelAndScanSupport = %d\n", support->SourceFormatPixelAndScanSupport); if (!fail_only || support->ViewportSizeSupport == 0) - dml2_printf("DML: support: ViewportSizeSupport = %d\n", support->ViewportSizeSupport); + DML_LOG_VERBOSE("DML: support: ViewportSizeSupport = %d\n", support->ViewportSizeSupport); if (!fail_only || support->LinkRateDoesNotMatchDPVersion == 1) - dml2_printf("DML: support: LinkRateDoesNotMatchDPVersion = %d\n", support->LinkRateDoesNotMatchDPVersion); + DML_LOG_VERBOSE("DML: support: LinkRateDoesNotMatchDPVersion = %d\n", support->LinkRateDoesNotMatchDPVersion); if (!fail_only || support->LinkRateForMultistreamNotIndicated == 1) - dml2_printf("DML: support: LinkRateForMultistreamNotIndicated = %d\n", support->LinkRateForMultistreamNotIndicated); + DML_LOG_VERBOSE("DML: support: LinkRateForMultistreamNotIndicated = %d\n", support->LinkRateForMultistreamNotIndicated); if (!fail_only || support->BPPForMultistreamNotIndicated == 1) - dml2_printf("DML: support: BPPForMultistreamNotIndicated = %d\n", support->BPPForMultistreamNotIndicated); + DML_LOG_VERBOSE("DML: support: BPPForMultistreamNotIndicated = %d\n", support->BPPForMultistreamNotIndicated); if (!fail_only || support->MultistreamWithHDMIOreDP == 1) - dml2_printf("DML: support: MultistreamWithHDMIOreDP = %d\n", support->MultistreamWithHDMIOreDP); + DML_LOG_VERBOSE("DML: support: MultistreamWithHDMIOreDP = %d\n", support->MultistreamWithHDMIOreDP); if (!fail_only || support->ExceededMultistreamSlots == 1) - dml2_printf("DML: support: ExceededMultistreamSlots = %d\n", support->ExceededMultistreamSlots); + DML_LOG_VERBOSE("DML: support: ExceededMultistreamSlots = %d\n", support->ExceededMultistreamSlots); if (!fail_only || support->MSOOrODMSplitWithNonDPLink == 1) - dml2_printf("DML: support: MSOOrODMSplitWithNonDPLink = %d\n", support->MSOOrODMSplitWithNonDPLink); + DML_LOG_VERBOSE("DML: support: MSOOrODMSplitWithNonDPLink = %d\n", support->MSOOrODMSplitWithNonDPLink); if (!fail_only || support->NotEnoughLanesForMSO == 1) - dml2_printf("DML: support: NotEnoughLanesForMSO = %d\n", support->NotEnoughLanesForMSO); + DML_LOG_VERBOSE("DML: support: NotEnoughLanesForMSO = %d\n", support->NotEnoughLanesForMSO); if (!fail_only || support->P2IWith420 == 1) - dml2_printf("DML: support: P2IWith420 = %d\n", support->P2IWith420); + DML_LOG_VERBOSE("DML: support: P2IWith420 = %d\n", support->P2IWith420); if (!fail_only || support->DSC422NativeNotSupported == 1) - dml2_printf("DML: support: DSC422NativeNotSupported = %d\n", support->DSC422NativeNotSupported); + DML_LOG_VERBOSE("DML: support: DSC422NativeNotSupported = %d\n", support->DSC422NativeNotSupported); if (!fail_only || support->DSCSlicesODMModeSupported == 0) - dml2_printf("DML: support: DSCSlicesODMModeSupported = %d\n", support->DSCSlicesODMModeSupported); + DML_LOG_VERBOSE("DML: support: DSCSlicesODMModeSupported = %d\n", support->DSCSlicesODMModeSupported); if (!fail_only || support->NotEnoughDSCUnits == 1) - dml2_printf("DML: support: NotEnoughDSCUnits = %d\n", support->NotEnoughDSCUnits); + DML_LOG_VERBOSE("DML: support: NotEnoughDSCUnits = %d\n", support->NotEnoughDSCUnits); if (!fail_only || support->NotEnoughDSCSlices == 1) - dml2_printf("DML: support: NotEnoughDSCSlices = %d\n", support->NotEnoughDSCSlices); + DML_LOG_VERBOSE("DML: support: NotEnoughDSCSlices = %d\n", support->NotEnoughDSCSlices); if (!fail_only || support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == 1) - dml2_printf("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = %d\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe); + DML_LOG_VERBOSE("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = %d\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe); if (!fail_only || support->InvalidCombinationOfMALLUseForPStateAndStaticScreen == 1) - dml2_printf("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = %d\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen); + DML_LOG_VERBOSE("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = %d\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen); if (!fail_only || support->DSCCLKRequiredMoreThanSupported == 1) - dml2_printf("DML: support: DSCCLKRequiredMoreThanSupported = %d\n", support->DSCCLKRequiredMoreThanSupported); + DML_LOG_VERBOSE("DML: support: DSCCLKRequiredMoreThanSupported = %d\n", support->DSCCLKRequiredMoreThanSupported); if (!fail_only || support->PixelsPerLinePerDSCUnitSupport == 0) - dml2_printf("DML: support: PixelsPerLinePerDSCUnitSupport = %d\n", support->PixelsPerLinePerDSCUnitSupport); + DML_LOG_VERBOSE("DML: support: PixelsPerLinePerDSCUnitSupport = %d\n", support->PixelsPerLinePerDSCUnitSupport); if (!fail_only || support->DTBCLKRequiredMoreThanSupported == 1) - dml2_printf("DML: support: DTBCLKRequiredMoreThanSupported = %d\n", support->DTBCLKRequiredMoreThanSupported); + DML_LOG_VERBOSE("DML: support: DTBCLKRequiredMoreThanSupported = %d\n", support->DTBCLKRequiredMoreThanSupported); if (!fail_only || support->InvalidCombinationOfMALLUseForPState == 1) - dml2_printf("DML: support: InvalidCombinationOfMALLUseForPState = %d\n", support->InvalidCombinationOfMALLUseForPState); + DML_LOG_VERBOSE("DML: support: InvalidCombinationOfMALLUseForPState = %d\n", support->InvalidCombinationOfMALLUseForPState); if (!fail_only || support->ROBSupport == 0) - dml2_printf("DML: support: ROBSupport = %d\n", support->ROBSupport); + DML_LOG_VERBOSE("DML: support: ROBSupport = %d\n", support->ROBSupport); if (!fail_only || support->OutstandingRequestsSupport == 0) - dml2_printf("DML: support: OutstandingRequestsSupport = %d\n", support->OutstandingRequestsSupport); + DML_LOG_VERBOSE("DML: support: OutstandingRequestsSupport = %d\n", support->OutstandingRequestsSupport); if (!fail_only || support->OutstandingRequestsUrgencyAvoidance == 0) - dml2_printf("DML: support: OutstandingRequestsUrgencyAvoidance = %d\n", support->OutstandingRequestsUrgencyAvoidance); + DML_LOG_VERBOSE("DML: support: OutstandingRequestsUrgencyAvoidance = %d\n", support->OutstandingRequestsUrgencyAvoidance); if (!fail_only || support->DISPCLK_DPPCLK_Support == 0) - dml2_printf("DML: support: DISPCLK_DPPCLK_Support = %d\n", support->DISPCLK_DPPCLK_Support); + DML_LOG_VERBOSE("DML: support: DISPCLK_DPPCLK_Support = %d\n", support->DISPCLK_DPPCLK_Support); if (!fail_only || support->TotalAvailablePipesSupport == 0) - dml2_printf("DML: support: TotalAvailablePipesSupport = %d\n", support->TotalAvailablePipesSupport); + DML_LOG_VERBOSE("DML: support: TotalAvailablePipesSupport = %d\n", support->TotalAvailablePipesSupport); if (!fail_only || support->NumberOfOTGSupport == 0) - dml2_printf("DML: support: NumberOfOTGSupport = %d\n", support->NumberOfOTGSupport); + DML_LOG_VERBOSE("DML: support: NumberOfOTGSupport = %d\n", support->NumberOfOTGSupport); if (!fail_only || support->NumberOfHDMIFRLSupport == 0) - dml2_printf("DML: support: NumberOfHDMIFRLSupport = %d\n", support->NumberOfHDMIFRLSupport); + DML_LOG_VERBOSE("DML: support: NumberOfHDMIFRLSupport = %d\n", support->NumberOfHDMIFRLSupport); if (!fail_only || support->NumberOfDP2p0Support == 0) - dml2_printf("DML: support: NumberOfDP2p0Support = %d\n", support->NumberOfDP2p0Support); + DML_LOG_VERBOSE("DML: support: NumberOfDP2p0Support = %d\n", support->NumberOfDP2p0Support); if (!fail_only || support->EnoughWritebackUnits == 0) - dml2_printf("DML: support: EnoughWritebackUnits = %d\n", support->EnoughWritebackUnits); + DML_LOG_VERBOSE("DML: support: EnoughWritebackUnits = %d\n", support->EnoughWritebackUnits); if (!fail_only || support->WritebackScaleRatioAndTapsSupport == 0) - dml2_printf("DML: support: WritebackScaleRatioAndTapsSupport = %d\n", support->WritebackScaleRatioAndTapsSupport); + DML_LOG_VERBOSE("DML: support: WritebackScaleRatioAndTapsSupport = %d\n", support->WritebackScaleRatioAndTapsSupport); if (!fail_only || support->WritebackLatencySupport == 0) - dml2_printf("DML: support: WritebackLatencySupport = %d\n", support->WritebackLatencySupport); + DML_LOG_VERBOSE("DML: support: WritebackLatencySupport = %d\n", support->WritebackLatencySupport); if (!fail_only || support->CursorSupport == 0) - dml2_printf("DML: support: CursorSupport = %d\n", support->CursorSupport); + DML_LOG_VERBOSE("DML: support: CursorSupport = %d\n", support->CursorSupport); if (!fail_only || support->PitchSupport == 0) - dml2_printf("DML: support: PitchSupport = %d\n", support->PitchSupport); + DML_LOG_VERBOSE("DML: support: PitchSupport = %d\n", support->PitchSupport); if (!fail_only || support->ViewportExceedsSurface == 1) - dml2_printf("DML: support: ViewportExceedsSurface = %d\n", support->ViewportExceedsSurface); + DML_LOG_VERBOSE("DML: support: ViewportExceedsSurface = %d\n", support->ViewportExceedsSurface); if (!fail_only || support->PrefetchSupported == 0) - dml2_printf("DML: support: PrefetchSupported = %d\n", support->PrefetchSupported); + DML_LOG_VERBOSE("DML: support: PrefetchSupported = %d\n", support->PrefetchSupported); if (!fail_only || support->EnoughUrgentLatencyHidingSupport == 0) - dml2_printf("DML: support: EnoughUrgentLatencyHidingSupport = %d\n", support->EnoughUrgentLatencyHidingSupport); + DML_LOG_VERBOSE("DML: support: EnoughUrgentLatencyHidingSupport = %d\n", support->EnoughUrgentLatencyHidingSupport); if (!fail_only || support->AvgBandwidthSupport == 0) - dml2_printf("DML: support: AvgBandwidthSupport = %d\n", support->AvgBandwidthSupport); + DML_LOG_VERBOSE("DML: support: AvgBandwidthSupport = %d\n", support->AvgBandwidthSupport); if (!fail_only || support->DynamicMetadataSupported == 0) - dml2_printf("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported); + DML_LOG_VERBOSE("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported); if (!fail_only || support->VRatioInPrefetchSupported == 0) - dml2_printf("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported); + DML_LOG_VERBOSE("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported); if (!fail_only || support->PTEBufferSizeNotExceeded == 0) - dml2_printf("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded); + DML_LOG_VERBOSE("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded); if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 0) - dml2_printf("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded); + DML_LOG_VERBOSE("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded); if (!fail_only || support->ExceededMALLSize == 1) - dml2_printf("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize); + DML_LOG_VERBOSE("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize); if (!fail_only || support->g6_temp_read_support == 0) - dml2_printf("DML: support: g6_temp_read_support = %d\n", support->g6_temp_read_support); + DML_LOG_VERBOSE("DML: support: g6_temp_read_support = %d\n", support->g6_temp_read_support); if (!fail_only || support->ImmediateFlipSupport == 0) - dml2_printf("DML: support: ImmediateFlipSupport = %d\n", support->ImmediateFlipSupport); + DML_LOG_VERBOSE("DML: support: ImmediateFlipSupport = %d\n", support->ImmediateFlipSupport); if (!fail_only || support->LinkCapacitySupport == 0) - dml2_printf("DML: support: LinkCapacitySupport = %d\n", support->LinkCapacitySupport); + DML_LOG_VERBOSE("DML: support: LinkCapacitySupport = %d\n", support->LinkCapacitySupport); if (!fail_only || support->ModeSupport == 0) - dml2_printf("DML: support: ModeSupport = %d\n", support->ModeSupport); - dml2_printf("DML: ===================================== \n"); + DML_LOG_VERBOSE("DML: support: ModeSupport = %d\n", support->ModeSupport); + DML_LOG_VERBOSE("DML: ===================================== \n"); } static void get_stream_output_bpp(double *out_bpp, const struct dml2_display_cfg *display_cfg) @@ -179,11 +179,9 @@ static void get_stream_output_bpp(double *out_bpp, const struct dml2_display_cfg } else { out_bpp[k] = 0; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d bpc=%f\n", __func__, k, bpc); - dml2_printf("DML::%s: k=%d dsc.enable=%d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable); - dml2_printf("DML::%s: k=%d out_bpp=%f\n", __func__, k, out_bpp[k]); -#endif + DML_LOG_VERBOSE("DML::%s: k=%d bpc=%f\n", __func__, k, bpc); + DML_LOG_VERBOSE("DML::%s: k=%d dsc.enable=%d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable); + DML_LOG_VERBOSE("DML::%s: k=%d out_bpp=%f\n", __func__, k, out_bpp[k]); } } @@ -212,9 +210,7 @@ static unsigned int dml_get_num_active_pipes(int unsigned num_planes, const stru num_active_pipes = num_active_pipes + (unsigned int)cfg_support_info->plane_support_info[k].dpps_used; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: num_active_pipes = %d\n", __func__, num_active_pipes); -#endif + DML_LOG_VERBOSE("DML::%s: num_active_pipes = %d\n", __func__, num_active_pipes); return num_active_pipes; } @@ -251,7 +247,7 @@ static bool dml_get_is_phantom_pipe(const struct dml2_display_cfg *display_cfg, unsigned int plane_idx = mode_lib->mp.pipe_plane[pipe_idx]; bool is_phantom = dml_is_phantom_pipe(&display_cfg->plane_descriptors[plane_idx]); - dml2_printf("DML::%s: pipe_idx=%d legacy_svp_config=%0d is_phantom=%d\n", __func__, pipe_idx, display_cfg->plane_descriptors[plane_idx].overrides.legacy_svp_config, is_phantom); + DML_LOG_VERBOSE("DML::%s: pipe_idx=%d legacy_svp_config=%0d is_phantom=%d\n", __func__, pipe_idx, display_cfg->plane_descriptors[plane_idx].overrides.legacy_svp_config, is_phantom); return is_phantom; } @@ -415,19 +411,17 @@ static void CalculateMaxDETAndMinCompressedBufferSize( *nomDETInKByte = (unsigned int)(math_floor2((double)*MaxTotalDETInKByte / (double)MaxNumDPP, ConfigReturnBufferSegmentSizeInKByte)); *MinCompressedBufferSizeInKByte = ConfigReturnBufferSizeInKByte - *MaxTotalDETInKByte; -#if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: is_mrq_present = %u\n", __func__, is_mrq_present); - dml2_printf("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte); - dml2_printf("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, ROBBufferSizeInKByte); - dml2_printf("DML::%s: MaxNumDPP = %u\n", __func__, MaxNumDPP); - dml2_printf("DML::%s: MaxTotalDETInKByte = %u\n", __func__, *MaxTotalDETInKByte); - dml2_printf("DML::%s: nomDETInKByte = %u\n", __func__, *nomDETInKByte); - dml2_printf("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, *MinCompressedBufferSizeInKByte); -#endif + DML_LOG_VERBOSE("DML::%s: is_mrq_present = %u\n", __func__, is_mrq_present); + DML_LOG_VERBOSE("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte); + DML_LOG_VERBOSE("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, ROBBufferSizeInKByte); + DML_LOG_VERBOSE("DML::%s: MaxNumDPP = %u\n", __func__, MaxNumDPP); + DML_LOG_VERBOSE("DML::%s: MaxTotalDETInKByte = %u\n", __func__, *MaxTotalDETInKByte); + DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, *nomDETInKByte); + DML_LOG_VERBOSE("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, *MinCompressedBufferSizeInKByte); if (nomDETInKByteOverrideEnable) { *nomDETInKByte = nomDETInKByteOverrideValue; - dml2_printf("DML::%s: nomDETInKByte = %u (overrided)\n", __func__, *nomDETInKByte); + DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u (overrided)\n", __func__, *nomDETInKByte); } } @@ -502,7 +496,7 @@ static bool dml_is_420(enum dml2_source_format_class source_format) val = 0; break; default: - DML2_ASSERT(0); + DML_ASSERT(0); break; } return val; @@ -535,7 +529,7 @@ static unsigned int dml_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode else if (sw_mode == dml2_gfx11_sw_256kb_r_x) return 262144; else { - DML2_ASSERT(0); + DML_ASSERT(0); return 256; } } @@ -570,8 +564,8 @@ static int unsigned dml_get_gfx_version(enum dml2_swizzle_mode sw_mode) sw_mode == dml2_gfx11_sw_256kb_r_x) { version = 11; } else { - dml2_printf("ERROR: Invalid sw_mode setting! val=%u\n", sw_mode); - DML2_ASSERT(0); + DML_LOG_VERBOSE("ERROR: Invalid sw_mode setting! val=%u\n", sw_mode); + DML_ASSERT(0); } return version; @@ -645,21 +639,19 @@ static void CalculateBytePerPixelAndBlockSizes( *BytePerPixelY = 2; *BytePerPixelC = 4; } else { - dml2_printf("ERROR: DML::%s: SourcePixelFormat = %u not supported!\n", __func__, SourcePixelFormat); - DML2_ASSERT(0); + DML_LOG_VERBOSE("ERROR: DML::%s: SourcePixelFormat = %u not supported!\n", __func__, SourcePixelFormat); + DML_ASSERT(0); } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: SourcePixelFormat = %u\n", __func__, SourcePixelFormat); - dml2_printf("DML::%s: BytePerPixelDETY = %f\n", __func__, *BytePerPixelDETY); - dml2_printf("DML::%s: BytePerPixelDETC = %f\n", __func__, *BytePerPixelDETC); - dml2_printf("DML::%s: BytePerPixelY = %u\n", __func__, *BytePerPixelY); - dml2_printf("DML::%s: BytePerPixelC = %u\n", __func__, *BytePerPixelC); - dml2_printf("DML::%s: pitch_y = %u\n", __func__, pitch_y); - dml2_printf("DML::%s: pitch_c = %u\n", __func__, pitch_c); - dml2_printf("DML::%s: surf_linear128_l = %u\n", __func__, *surf_linear128_l); - dml2_printf("DML::%s: surf_linear128_c = %u\n", __func__, *surf_linear128_c); -#endif + DML_LOG_VERBOSE("DML::%s: SourcePixelFormat = %u\n", __func__, SourcePixelFormat); + DML_LOG_VERBOSE("DML::%s: BytePerPixelDETY = %f\n", __func__, *BytePerPixelDETY); + DML_LOG_VERBOSE("DML::%s: BytePerPixelDETC = %f\n", __func__, *BytePerPixelDETC); + DML_LOG_VERBOSE("DML::%s: BytePerPixelY = %u\n", __func__, *BytePerPixelY); + DML_LOG_VERBOSE("DML::%s: BytePerPixelC = %u\n", __func__, *BytePerPixelC); + DML_LOG_VERBOSE("DML::%s: pitch_y = %u\n", __func__, pitch_y); + DML_LOG_VERBOSE("DML::%s: pitch_c = %u\n", __func__, pitch_c); + DML_LOG_VERBOSE("DML::%s: surf_linear128_l = %u\n", __func__, *surf_linear128_l); + DML_LOG_VERBOSE("DML::%s: surf_linear128_c = %u\n", __func__, *surf_linear128_c); if (dml_get_gfx_version(SurfaceTiling) == 11) { *surf_linear128_l = 0; @@ -703,12 +695,10 @@ static void CalculateBytePerPixelAndBlockSizes( *BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY; *BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: BlockWidth256BytesY = %u\n", __func__, *BlockWidth256BytesY); - dml2_printf("DML::%s: BlockHeight256BytesY = %u\n", __func__, *BlockHeight256BytesY); - dml2_printf("DML::%s: BlockWidth256BytesC = %u\n", __func__, *BlockWidth256BytesC); - dml2_printf("DML::%s: BlockHeight256BytesC = %u\n", __func__, *BlockHeight256BytesC); -#endif + DML_LOG_VERBOSE("DML::%s: BlockWidth256BytesY = %u\n", __func__, *BlockWidth256BytesY); + DML_LOG_VERBOSE("DML::%s: BlockHeight256BytesY = %u\n", __func__, *BlockHeight256BytesY); + DML_LOG_VERBOSE("DML::%s: BlockWidth256BytesC = %u\n", __func__, *BlockWidth256BytesC); + DML_LOG_VERBOSE("DML::%s: BlockHeight256BytesC = %u\n", __func__, *BlockHeight256BytesC); if (dml_get_gfx_version(SurfaceTiling) == 11) { if (SurfaceTiling == dml2_gfx11_sw_linear) { @@ -752,8 +742,8 @@ static void CalculateBytePerPixelAndBlockSizes( } else if (SurfaceTiling == dml2_sw_256kb_2d) { macro_tile_scale = 32; } else { - dml2_printf("ERROR: Invalid SurfaceTiling setting! val=%u\n", SurfaceTiling); - DML2_ASSERT(0); + DML_LOG_VERBOSE("ERROR: Invalid SurfaceTiling setting! val=%u\n", SurfaceTiling); + DML_ASSERT(0); } *MacroTileHeightY = macro_tile_scale * *BlockHeight256BytesY; @@ -766,12 +756,10 @@ static void CalculateBytePerPixelAndBlockSizes( } } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: MacroTileWidthY = %u\n", __func__, *MacroTileWidthY); - dml2_printf("DML::%s: MacroTileHeightY = %u\n", __func__, *MacroTileHeightY); - dml2_printf("DML::%s: MacroTileWidthC = %u\n", __func__, *MacroTileWidthC); - dml2_printf("DML::%s: MacroTileHeightC = %u\n", __func__, *MacroTileHeightC); -#endif + DML_LOG_VERBOSE("DML::%s: MacroTileWidthY = %u\n", __func__, *MacroTileWidthY); + DML_LOG_VERBOSE("DML::%s: MacroTileHeightY = %u\n", __func__, *MacroTileHeightY); + DML_LOG_VERBOSE("DML::%s: MacroTileWidthC = %u\n", __func__, *MacroTileWidthC); + DML_LOG_VERBOSE("DML::%s: MacroTileHeightC = %u\n", __func__, *MacroTileHeightC); } static void CalculateSinglePipeDPPCLKAndSCLThroughput( @@ -860,10 +848,8 @@ static void CalculateSwathWidth( unsigned int surface_width_ub_c; unsigned int surface_height_ub_c; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP); - dml2_printf("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces); -#endif + DML_LOG_VERBOSE("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP); + DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces); for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { if (!dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle)) { @@ -872,11 +858,9 @@ static void CalculateSwathWidth( SwathWidthSingleDPPY[k] = (unsigned int)display_cfg->plane_descriptors[k].composition.viewport.plane0.height; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u ViewportWidth=%u\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.width); - dml2_printf("DML::%s: k=%u ViewportHeight=%u\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.height); - dml2_printf("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u ViewportWidth=%lu\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.width); + DML_LOG_VERBOSE("DML::%s: k=%u ViewportHeight=%lu\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.height); + DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]); MainSurfaceODMMode = ODMMode[k]; @@ -899,13 +883,11 @@ static void CalculateSwathWidth( } } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u HActive=%u\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active); - dml2_printf("DML::%s: k=%u HRatio=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio); - dml2_printf("DML::%s: k=%u MainSurfaceODMMode=%u\n", __func__, k, MainSurfaceODMMode); - dml2_printf("DML::%s: k=%u SwathWidthSingleDPPY=%u\n", __func__, k, SwathWidthSingleDPPY[k]); - dml2_printf("DML::%s: k=%u SwathWidthY=%u\n", __func__, k, SwathWidthY[k]); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u HActive=%lu\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active); + DML_LOG_VERBOSE("DML::%s: k=%u HRatio=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u MainSurfaceODMMode=%u\n", __func__, k, MainSurfaceODMMode); + DML_LOG_VERBOSE("DML::%s: k=%u SwathWidthSingleDPPY=%u\n", __func__, k, SwathWidthSingleDPPY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u SwathWidthY=%u\n", __func__, k, SwathWidthY[k]); if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format)) { SwathWidthC[k] = SwathWidthY[k] / 2; @@ -934,22 +916,20 @@ static void CalculateSwathWidth( surface_width_ub_c = (unsigned int)math_ceil2((double)display_cfg->plane_descriptors[k].surface.plane1.width, req_width_horz_c); surface_height_ub_c = (unsigned int)math_ceil2((double)display_cfg->plane_descriptors[k].surface.plane1.height, Read256BytesBlockHeightC[k]); -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u surface_width_ub_l=%u\n", __func__, k, surface_width_ub_l); - dml2_printf("DML::%s: k=%u surface_height_ub_l=%u\n", __func__, k, surface_height_ub_l); - dml2_printf("DML::%s: k=%u surface_width_ub_c=%u\n", __func__, k, surface_width_ub_c); - dml2_printf("DML::%s: k=%u surface_height_ub_c=%u\n", __func__, k, surface_height_ub_c); - dml2_printf("DML::%s: k=%u req_width_horz_y=%u\n", __func__, k, req_width_horz_y); - dml2_printf("DML::%s: k=%u req_width_horz_c=%u\n", __func__, k, req_width_horz_c); - dml2_printf("DML::%s: k=%u Read256BytesBlockWidthY=%u\n", __func__, k, Read256BytesBlockWidthY[k]); - dml2_printf("DML::%s: k=%u Read256BytesBlockHeightY=%u\n", __func__, k, Read256BytesBlockHeightY[k]); - dml2_printf("DML::%s: k=%u Read256BytesBlockWidthC=%u\n", __func__, k, Read256BytesBlockWidthC[k]); - dml2_printf("DML::%s: k=%u Read256BytesBlockHeightC=%u\n", __func__, k, Read256BytesBlockHeightC[k]); - dml2_printf("DML::%s: k=%u req_width_horz_y=%u\n", __func__, k, req_width_horz_y); - dml2_printf("DML::%s: k=%u req_width_horz_c=%u\n", __func__, k, req_width_horz_c); - dml2_printf("DML::%s: k=%u ViewportStationary=%u\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.stationary); - dml2_printf("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u surface_width_ub_l=%u\n", __func__, k, surface_width_ub_l); + DML_LOG_VERBOSE("DML::%s: k=%u surface_height_ub_l=%u\n", __func__, k, surface_height_ub_l); + DML_LOG_VERBOSE("DML::%s: k=%u surface_width_ub_c=%u\n", __func__, k, surface_width_ub_c); + DML_LOG_VERBOSE("DML::%s: k=%u surface_height_ub_c=%u\n", __func__, k, surface_height_ub_c); + DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_y=%u\n", __func__, k, req_width_horz_y); + DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_c=%u\n", __func__, k, req_width_horz_c); + DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockWidthY=%u\n", __func__, k, Read256BytesBlockWidthY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockHeightY=%u\n", __func__, k, Read256BytesBlockHeightY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockWidthC=%u\n", __func__, k, Read256BytesBlockWidthC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockHeightC=%u\n", __func__, k, Read256BytesBlockHeightC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_y=%u\n", __func__, k, req_width_horz_y); + DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_c=%u\n", __func__, k, req_width_horz_c); + DML_LOG_VERBOSE("DML::%s: k=%u ViewportStationary=%u\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.stationary); + DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]); req_per_swath_ub_l[k] = 0; req_per_swath_ub_c[k] = 0; @@ -995,15 +975,12 @@ static void CalculateSwathWidth( } } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u swath_width_luma_ub=%u\n", __func__, k, swath_width_luma_ub[k]); - dml2_printf("DML::%s: k=%u swath_width_chroma_ub=%u\n", __func__, k, swath_width_chroma_ub[k]); - dml2_printf("DML::%s: k=%u MaximumSwathHeightY=%u\n", __func__, k, MaximumSwathHeightY[k]); - dml2_printf("DML::%s: k=%u MaximumSwathHeightC=%u\n", __func__, k, MaximumSwathHeightC[k]); - dml2_printf("DML::%s: k=%u req_per_swath_ub_l=%u\n", __func__, k, req_per_swath_ub_l[k]); - dml2_printf("DML::%s: k=%u req_per_swath_ub_c=%u\n", __func__, k, req_per_swath_ub_c[k]); -#endif - + DML_LOG_VERBOSE("DML::%s: k=%u swath_width_luma_ub=%u\n", __func__, k, swath_width_luma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u swath_width_chroma_ub=%u\n", __func__, k, swath_width_chroma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightY=%u\n", __func__, k, MaximumSwathHeightY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightC=%u\n", __func__, k, MaximumSwathHeightC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u req_per_swath_ub_l=%u\n", __func__, k, req_per_swath_ub_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u req_per_swath_ub_c=%u\n", __func__, k, req_per_swath_ub_c[k]); } } @@ -1018,13 +995,11 @@ static bool UnboundedRequest(bool unb_req_force_en, bool unb_req_force_val, unsi if (unb_req_force_en) { unb_req_en = unb_req_force_val && unb_req_ok; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: unb_req_force_en = %u\n", __func__, unb_req_force_en); - dml2_printf("DML::%s: unb_req_force_val = %u\n", __func__, unb_req_force_val); - dml2_printf("DML::%s: unb_req_ok = %u\n", __func__, unb_req_ok); - dml2_printf("DML::%s: unb_req_en = %u\n", __func__, unb_req_en); -#endif - return (unb_req_en); + DML_LOG_VERBOSE("DML::%s: unb_req_force_en = %u\n", __func__, unb_req_force_en); + DML_LOG_VERBOSE("DML::%s: unb_req_force_val = %u\n", __func__, unb_req_force_val); + DML_LOG_VERBOSE("DML::%s: unb_req_ok = %u\n", __func__, unb_req_ok); + DML_LOG_VERBOSE("DML::%s: unb_req_en = %u\n", __func__, unb_req_en); + return unb_req_en; } static void CalculateDETBufferSize( @@ -1054,16 +1029,14 @@ static void CalculateDETBufferSize( bool NextPotentialSurfaceToAssignDETPieceFound; bool MinimizeReallocationSuccess = false; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP); - dml2_printf("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte); - dml2_printf("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces); - dml2_printf("DML::%s: UnboundedRequestEnabled = %u\n", __func__, UnboundedRequestEnabled); - dml2_printf("DML::%s: MaxTotalDETInKByte = %u\n", __func__, MaxTotalDETInKByte); - dml2_printf("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte); - dml2_printf("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, MinCompressedBufferSizeInKByte); - dml2_printf("DML::%s: CompressedBufferSegmentSizeInkByte = %u\n", __func__, CompressedBufferSegmentSizeInkByte); -#endif + DML_LOG_VERBOSE("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP); + DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte); + DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces); + DML_LOG_VERBOSE("DML::%s: UnboundedRequestEnabled = %u\n", __func__, UnboundedRequestEnabled); + DML_LOG_VERBOSE("DML::%s: MaxTotalDETInKByte = %u\n", __func__, MaxTotalDETInKByte); + DML_LOG_VERBOSE("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte); + DML_LOG_VERBOSE("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, MinCompressedBufferSizeInKByte); + DML_LOG_VERBOSE("DML::%s: CompressedBufferSegmentSizeInkByte = %u\n", __func__, CompressedBufferSegmentSizeInkByte); // Note: Will use default det size if that fits 2 swaths if (UnboundedRequestEnabled) { @@ -1092,19 +1065,15 @@ static void CalculateDETBufferSize( l->minDET = l->minDET + ConfigReturnBufferSegmentSizeInkByte; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u minDET = %u\n", __func__, k, l->minDET); - dml2_printf("DML::%s: k=%u max_minDET = %u\n", __func__, k, l->max_minDET); - dml2_printf("DML::%s: k=%u minDET_pipe = %u\n", __func__, k, l->minDET_pipe); - dml2_printf("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, full_swath_bytes_l[k]); - dml2_printf("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, full_swath_bytes_c[k]); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u minDET = %u\n", __func__, k, l->minDET); + DML_LOG_VERBOSE("DML::%s: k=%u max_minDET = %u\n", __func__, k, l->max_minDET); + DML_LOG_VERBOSE("DML::%s: k=%u minDET_pipe = %u\n", __func__, k, l->minDET_pipe); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, full_swath_bytes_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, full_swath_bytes_c[k]); if (l->minDET_pipe == 0) { l->minDET_pipe = (unsigned int)(math_max2(128, math_ceil2(((double)full_swath_bytes_l[k] + (double)full_swath_bytes_c[k]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte))); -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u minDET_pipe = %u (assume each plane take half DET)\n", __func__, k, l->minDET_pipe); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u minDET_pipe = %u (assume each plane take half DET)\n", __func__, k, l->minDET_pipe); } if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) { @@ -1117,12 +1086,10 @@ static void CalculateDETBufferSize( l->DETBufferSizePoolInKByte = l->DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * l->minDET_pipe; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]); - dml2_printf("DML::%s: k=%u DETSizeOverride = %u\n", __func__, k, display_cfg->plane_descriptors[k].overrides.det_size_override_kb); - dml2_printf("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]); - dml2_printf("DML::%s: DETBufferSizePoolInKByte = %u\n", __func__, l->DETBufferSizePoolInKByte); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETSizeOverride = %u\n", __func__, k, display_cfg->plane_descriptors[k].overrides.det_size_override_kb); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]); + DML_LOG_VERBOSE("DML::%s: DETBufferSizePoolInKByte = %u\n", __func__, l->DETBufferSizePoolInKByte); } if (display_cfg->minimize_det_reallocation) { @@ -1194,14 +1161,12 @@ static void CalculateDETBufferSize( l->TotalBandwidth = l->TotalBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k]; } } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: --- Before bandwidth adjustment ---\n", __func__); + DML_LOG_VERBOSE("DML::%s: --- Before bandwidth adjustment ---\n", __func__); for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { - dml2_printf("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]); } - dml2_printf("DML::%s: --- DET allocation with bandwidth ---\n", __func__); -#endif - dml2_printf("DML::%s: TotalBandwidth = %f\n", __func__, l->TotalBandwidth); + DML_LOG_VERBOSE("DML::%s: --- DET allocation with bandwidth ---\n", __func__); + DML_LOG_VERBOSE("DML::%s: TotalBandwidth = %f\n", __func__, l->TotalBandwidth); l->BandwidthOfSurfacesNotAssignedDETPiece = l->TotalBandwidth; for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { @@ -1213,10 +1178,8 @@ static void CalculateDETBufferSize( } else { DETPieceAssignedToThisSurfaceAlready[k] = false; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, k, DETPieceAssignedToThisSurfaceAlready[k]); - dml2_printf("DML::%s: k=%u BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, k, l->BandwidthOfSurfacesNotAssignedDETPiece); -#endif + DML_LOG_VERBOSE("DML::%s: k=%u DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, k, DETPieceAssignedToThisSurfaceAlready[k]); + DML_LOG_VERBOSE("DML::%s: k=%u BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, k, l->BandwidthOfSurfacesNotAssignedDETPiece); } for (unsigned int j = 0; j < NumberOfActiveSurfaces; ++j) { @@ -1224,22 +1187,18 @@ static void CalculateDETBufferSize( l->NextSurfaceToAssignDETPiece = 0; for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: j=%u k=%u, ReadBandwidthLuma[k] = %f\n", __func__, j, k, ReadBandwidthLuma[k]); - dml2_printf("DML::%s: j=%u k=%u, ReadBandwidthChroma[k] = %f\n", __func__, j, k, ReadBandwidthChroma[k]); - dml2_printf("DML::%s: j=%u k=%u, ReadBandwidthLuma[Next] = %f\n", __func__, j, k, ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece]); - dml2_printf("DML::%s: j=%u k=%u, ReadBandwidthChroma[Next] = %f\n", __func__, j, k, ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]); - dml2_printf("DML::%s: j=%u k=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, k, l->NextSurfaceToAssignDETPiece); -#endif + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthLuma[k] = %f\n", __func__, j, k, ReadBandwidthLuma[k]); + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthChroma[k] = %f\n", __func__, j, k, ReadBandwidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthLuma[Next] = %f\n", __func__, j, k, ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece]); + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthChroma[Next] = %f\n", __func__, j, k, ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]); + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, k, l->NextSurfaceToAssignDETPiece); if (!DETPieceAssignedToThisSurfaceAlready[k] && (!NextPotentialSurfaceToAssignDETPieceFound || ReadBandwidthLuma[k] + ReadBandwidthChroma[k] < ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece])) { l->NextSurfaceToAssignDETPiece = k; NextPotentialSurfaceToAssignDETPieceFound = true; } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: j=%u k=%u, DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, j, k, DETPieceAssignedToThisSurfaceAlready[k]); - dml2_printf("DML::%s: j=%u k=%u, NextPotentialSurfaceToAssignDETPieceFound = %u\n", __func__, j, k, NextPotentialSurfaceToAssignDETPieceFound); -#endif + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, j, k, DETPieceAssignedToThisSurfaceAlready[k]); + DML_LOG_VERBOSE("DML::%s: j=%u k=%u, NextPotentialSurfaceToAssignDETPieceFound = %u\n", __func__, j, k, NextPotentialSurfaceToAssignDETPieceFound); } if (NextPotentialSurfaceToAssignDETPieceFound) { @@ -1249,20 +1208,16 @@ static void CalculateDETBufferSize( * (ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte, math_floor2((double)l->DETBufferSizePoolInKByte, (ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte))); -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: j=%u, DETBufferSizePoolInKByte = %u\n", __func__, j, l->DETBufferSizePoolInKByte); - dml2_printf("DML::%s: j=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, l->NextSurfaceToAssignDETPiece); - dml2_printf("DML::%s: j=%u, ReadBandwidthLuma[%u] = %f\n", __func__, j, l->NextSurfaceToAssignDETPiece, ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece]); - dml2_printf("DML::%s: j=%u, ReadBandwidthChroma[%u] = %f\n", __func__, j, l->NextSurfaceToAssignDETPiece, ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]); - dml2_printf("DML::%s: j=%u, BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, j, l->BandwidthOfSurfacesNotAssignedDETPiece); - dml2_printf("DML::%s: j=%u, NextDETBufferPieceInKByte = %u\n", __func__, j, l->NextDETBufferPieceInKByte); - dml2_printf("DML::%s: j=%u, DETBufferSizeInKByte[%u] increases from %u ", __func__, j, l->NextSurfaceToAssignDETPiece, DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece]); -#endif + DML_LOG_VERBOSE("DML::%s: j=%u, DETBufferSizePoolInKByte = %u\n", __func__, j, l->DETBufferSizePoolInKByte); + DML_LOG_VERBOSE("DML::%s: j=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, l->NextSurfaceToAssignDETPiece); + DML_LOG_VERBOSE("DML::%s: j=%u, ReadBandwidthLuma[%u] = %f\n", __func__, j, l->NextSurfaceToAssignDETPiece, ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece]); + DML_LOG_VERBOSE("DML::%s: j=%u, ReadBandwidthChroma[%u] = %f\n", __func__, j, l->NextSurfaceToAssignDETPiece, ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]); + DML_LOG_VERBOSE("DML::%s: j=%u, BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, j, l->BandwidthOfSurfacesNotAssignedDETPiece); + DML_LOG_VERBOSE("DML::%s: j=%u, NextDETBufferPieceInKByte = %u\n", __func__, j, l->NextDETBufferPieceInKByte); + DML_LOG_VERBOSE("DML::%s: j=%u, DETBufferSizeInKByte[%u] increases from %u ", __func__, j, l->NextSurfaceToAssignDETPiece, DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece]); DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece] = DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece] + l->NextDETBufferPieceInKByte / (ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]); -#ifdef __DML_VBA_DEBUG__ - dml2_printf("to %u\n", DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece]); -#endif + DML_LOG_VERBOSE("to %u\n", DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece]); l->DETBufferSizePoolInKByte = l->DETBufferSizePoolInKByte - l->NextDETBufferPieceInKByte; DETPieceAssignedToThisSurfaceAlready[l->NextSurfaceToAssignDETPiece] = true; @@ -1274,13 +1229,11 @@ static void CalculateDETBufferSize( } *CompressedBufferSizeInkByte = *CompressedBufferSizeInkByte * CompressedBufferSegmentSizeInkByte / ConfigReturnBufferSegmentSizeInkByte; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: --- After bandwidth adjustment ---\n", __func__); - dml2_printf("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *CompressedBufferSizeInkByte); + DML_LOG_VERBOSE("DML::%s: --- After bandwidth adjustment ---\n", __func__); + DML_LOG_VERBOSE("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *CompressedBufferSizeInkByte); for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { - dml2_printf("DML::%s: k=%u DETBufferSizeInKByte = %u (TotalReadBandWidth=%f)\n", __func__, k, DETBufferSizeInKByte[k], ReadBandwidthLuma[k] + ReadBandwidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u (TotalReadBandWidth=%f)\n", __func__, k, DETBufferSizeInKByte[k], ReadBandwidthLuma[k] + ReadBandwidthChroma[k]); } -#endif } static double CalculateRequiredDispclk( @@ -1510,15 +1463,13 @@ static unsigned int dscceComputeDelay( //pixel delay is group_delay (converted to pixels) + pipeline, however, first group is a special case since it is processed as soon as it arrives (i.e., in 3 cycles regardless of pixel format) pixels = (group_delay - 1) * cycles_per_group + 3 + pipeline_delay; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: bpc: %u\n", __func__, bpc); - dml2_printf("DML::%s: BPP: %f\n", __func__, BPP); - dml2_printf("DML::%s: sliceWidth: %u\n", __func__, sliceWidth); - dml2_printf("DML::%s: numSlices: %u\n", __func__, numSlices); - dml2_printf("DML::%s: pixelFormat: %u\n", __func__, pixelFormat); - dml2_printf("DML::%s: Output: %u\n", __func__, Output); - dml2_printf("DML::%s: pixels: %u\n", __func__, pixels); -#endif + DML_LOG_VERBOSE("DML::%s: bpc: %u\n", __func__, bpc); + DML_LOG_VERBOSE("DML::%s: BPP: %f\n", __func__, BPP); + DML_LOG_VERBOSE("DML::%s: sliceWidth: %u\n", __func__, sliceWidth); + DML_LOG_VERBOSE("DML::%s: numSlices: %u\n", __func__, numSlices); + DML_LOG_VERBOSE("DML::%s: pixelFormat: %u\n", __func__, pixelFormat); + DML_LOG_VERBOSE("DML::%s: Output: %u\n", __func__, Output); + DML_LOG_VERBOSE("DML::%s: pixels: %u\n", __func__, pixels); return pixels; } @@ -1593,10 +1544,8 @@ static unsigned int dscComputeDelay(enum dml2_output_format_class pixelFormat, e // sft Delay = Delay + 1; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: pixelFormat = %u\n", __func__, pixelFormat); - dml2_printf("DML::%s: Delay = %u\n", __func__, Delay); -#endif + DML_LOG_VERBOSE("DML::%s: pixelFormat = %u\n", __func__, pixelFormat); + DML_LOG_VERBOSE("DML::%s: Delay = %u\n", __func__, Delay); return Delay; } @@ -1667,10 +1616,8 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ } meta_surface_bytes = (unsigned int)(p->DCCMetaPitch * vp_height_meta_ub * p->BytePerPixel / 256.0); -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DCCMetaPitch = %u\n", __func__, p->DCCMetaPitch); - dml2_printf("DML::%s: meta_surface_bytes = %u\n", __func__, meta_surface_bytes); -#endif + DML_LOG_VERBOSE("DML::%s: DCCMetaPitch = %u\n", __func__, p->DCCMetaPitch); + DML_LOG_VERBOSE("DML::%s: meta_surface_bytes = %u\n", __func__, meta_surface_bytes); if (p->GPUVMEnable == true) { double meta_vmpg_bytes = 4.0 * 1024.0; *p->meta_pte_bytes_per_frame_ub = (unsigned int)((math_ceil2((double) (meta_surface_bytes - meta_vmpg_bytes) / (8 * meta_vmpg_bytes), 1) + 1) * 64); @@ -1724,25 +1671,23 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ vm_bytes = *p->meta_pte_bytes_per_frame_ub + extra_mpde_bytes + *p->dpde0_bytes_per_frame_ub + extra_dpde_bytes; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DCCEnable = %u\n", __func__, p->DCCEnable); - dml2_printf("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable); - dml2_printf("DML::%s: SwModeLinear = %u\n", __func__, p->SurfaceTiling == dml2_sw_linear); - dml2_printf("DML::%s: BytePerPixel = %u\n", __func__, p->BytePerPixel); - dml2_printf("DML::%s: GPUVMMaxPageTableLevels = %u\n", __func__, p->GPUVMMaxPageTableLevels); - dml2_printf("DML::%s: BlockHeight256Bytes = %u\n", __func__, p->BlockHeight256Bytes); - dml2_printf("DML::%s: BlockWidth256Bytes = %u\n", __func__, p->BlockWidth256Bytes); - dml2_printf("DML::%s: MacroTileHeight = %u\n", __func__, p->MacroTileHeight); - dml2_printf("DML::%s: MacroTileWidth = %u\n", __func__, p->MacroTileWidth); - dml2_printf("DML::%s: meta_pte_bytes_per_frame_ub = %u\n", __func__, *p->meta_pte_bytes_per_frame_ub); - dml2_printf("DML::%s: dpde0_bytes_per_frame_ub = %u\n", __func__, *p->dpde0_bytes_per_frame_ub); - dml2_printf("DML::%s: extra_mpde_bytes = %u\n", __func__, extra_mpde_bytes); - dml2_printf("DML::%s: extra_dpde_bytes = %u\n", __func__, extra_dpde_bytes); - dml2_printf("DML::%s: vm_bytes = %u\n", __func__, vm_bytes); - dml2_printf("DML::%s: ViewportHeight = %u\n", __func__, p->ViewportHeight); - dml2_printf("DML::%s: SwathWidth = %u\n", __func__, p->SwathWidth); - dml2_printf("DML::%s: vp_height_dpte_ub = %u\n", __func__, vp_height_dpte_ub); -#endif + DML_LOG_VERBOSE("DML::%s: DCCEnable = %u\n", __func__, p->DCCEnable); + DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable); + DML_LOG_VERBOSE("DML::%s: SwModeLinear = %u\n", __func__, p->SurfaceTiling == dml2_sw_linear); + DML_LOG_VERBOSE("DML::%s: BytePerPixel = %u\n", __func__, p->BytePerPixel); + DML_LOG_VERBOSE("DML::%s: GPUVMMaxPageTableLevels = %u\n", __func__, p->GPUVMMaxPageTableLevels); + DML_LOG_VERBOSE("DML::%s: BlockHeight256Bytes = %u\n", __func__, p->BlockHeight256Bytes); + DML_LOG_VERBOSE("DML::%s: BlockWidth256Bytes = %u\n", __func__, p->BlockWidth256Bytes); + DML_LOG_VERBOSE("DML::%s: MacroTileHeight = %u\n", __func__, p->MacroTileHeight); + DML_LOG_VERBOSE("DML::%s: MacroTileWidth = %u\n", __func__, p->MacroTileWidth); + DML_LOG_VERBOSE("DML::%s: meta_pte_bytes_per_frame_ub = %u\n", __func__, *p->meta_pte_bytes_per_frame_ub); + DML_LOG_VERBOSE("DML::%s: dpde0_bytes_per_frame_ub = %u\n", __func__, *p->dpde0_bytes_per_frame_ub); + DML_LOG_VERBOSE("DML::%s: extra_mpde_bytes = %u\n", __func__, extra_mpde_bytes); + DML_LOG_VERBOSE("DML::%s: extra_dpde_bytes = %u\n", __func__, extra_dpde_bytes); + DML_LOG_VERBOSE("DML::%s: vm_bytes = %u\n", __func__, vm_bytes); + DML_LOG_VERBOSE("DML::%s: ViewportHeight = %u\n", __func__, p->ViewportHeight); + DML_LOG_VERBOSE("DML::%s: SwathWidth = %u\n", __func__, p->SwathWidth); + DML_LOG_VERBOSE("DML::%s: vp_height_dpte_ub = %u\n", __func__, vp_height_dpte_ub); if (p->SurfaceTiling == dml2_sw_linear) { *p->PixelPTEReqHeight = 1; @@ -1778,22 +1723,20 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ *p->vmpg_width = 1024 * p->GPUVMMinPageSizeKBytes / (p->MacroTileHeight * p->BytePerPixel); if (p->GPUVMEnable == true) { - dml2_printf("DML::%s: GPUVMMinPageSizeKBytes=%u and sw_mode=%u (tile_size=%d) not supported!\n", + DML_LOG_VERBOSE("DML::%s: GPUVMMinPageSizeKBytes=%u and sw_mode=%u (tile_size=%d) not supported!\n", __func__, p->GPUVMMinPageSizeKBytes, p->SurfaceTiling, dml_get_tile_block_size_bytes(p->SurfaceTiling)); - DML2_ASSERT(0); + DML_ASSERT(0); } } -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, p->GPUVMMinPageSizeKBytes); - dml2_printf("DML::%s: PixelPTEReqHeight = %u\n", __func__, *p->PixelPTEReqHeight); - dml2_printf("DML::%s: PixelPTEReqWidth = %u\n", __func__, *p->PixelPTEReqWidth); - dml2_printf("DML::%s: PixelPTEReqWidth_linear = %u\n", __func__, PixelPTEReqWidth_linear); - dml2_printf("DML::%s: PTERequestSize = %u\n", __func__, *p->PTERequestSize); - dml2_printf("DML::%s: Pitch = %u\n", __func__, p->Pitch); - dml2_printf("DML::%s: vmpg_width = %u\n", __func__, *p->vmpg_width); - dml2_printf("DML::%s: vmpg_height = %u\n", __func__, *p->vmpg_height); -#endif + DML_LOG_VERBOSE("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, p->GPUVMMinPageSizeKBytes); + DML_LOG_VERBOSE("DML::%s: PixelPTEReqHeight = %u\n", __func__, *p->PixelPTEReqHeight); + DML_LOG_VERBOSE("DML::%s: PixelPTEReqWidth = %u\n", __func__, *p->PixelPTEReqWidth); + DML_LOG_VERBOSE("DML::%s: PixelPTEReqWidth_linear = %u\n", __func__, PixelPTEReqWidth_linear); + DML_LOG_VERBOSE("DML::%s: PTERequestSize = %u\n", __func__, *p->PTERequestSize); + DML_LOG_VERBOSE("DML::%s: Pitch = %u\n", __func__, p->Pitch); + DML_LOG_VERBOSE("DML::%s: vmpg_width = %u\n", __func__, *p->vmpg_width); + DML_LOG_VERBOSE("DML::%s: vmpg_height = %u\n", __func__, *p->vmpg_height); *p->dpte_row_height_one_row_per_frame = vp_height_dpte_ub; *p->dpte_row_width_ub_one_row_per_frame = (unsigned int)((math_ceil2(((double)p->Pitch * (double)*p->dpte_row_height_one_row_per_frame / (double)*p->PixelPTEReqHeight - 1) / (double)*p->PixelPTEReqWidth, 1) + 1) * (double)*p->PixelPTEReqWidth); @@ -1811,7 +1754,7 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ *p->dpte_row_height_linear = 128; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: dpte_row_width_ub = %u (linear)\n", __func__, *p->dpte_row_width_ub); + DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u (linear)\n", __func__, *p->dpte_row_width_ub); #endif } else if (!dml_is_vertical_rotation(p->RotationAngle)) { @@ -1825,7 +1768,7 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ *p->dpte_row_width_ub = (unsigned int)((math_ceil2((double)(p->SwathWidth - 1) / (double)*p->PixelPTEReqWidth, 1) + 1.0) * *p->PixelPTEReqWidth); } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: dpte_row_width_ub = %u (tiled horz)\n", __func__, *p->dpte_row_width_ub); + DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u (tiled horz)\n", __func__, *p->dpte_row_width_ub); #endif *p->PixelPTEBytesPerRow = *p->dpte_row_width_ub / *p->PixelPTEReqWidth * *p->PTERequestSize; @@ -1840,7 +1783,7 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ *p->PixelPTEBytesPerRow = (unsigned int)((double)*p->dpte_row_width_ub / (double)*p->PixelPTEReqHeight * *p->PTERequestSize); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: dpte_row_width_ub = %u (tiled vert)\n", __func__, *p->dpte_row_width_ub); + DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u (tiled vert)\n", __func__, *p->dpte_row_width_ub); #endif } @@ -1852,18 +1795,18 @@ static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_ *p->PixelPTEBytesPerRowStorage = *p->PixelPTEBytesPerRow; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, p->GPUVMMinPageSizeKBytes); - dml2_printf("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable); - dml2_printf("DML::%s: meta_row_height = %u\n", __func__, *p->meta_row_height); - dml2_printf("DML::%s: dpte_row_height = %u\n", __func__, *p->dpte_row_height); - dml2_printf("DML::%s: dpte_row_height_linear = %u\n", __func__, *p->dpte_row_height_linear); - dml2_printf("DML::%s: dpte_row_width_ub = %u\n", __func__, *p->dpte_row_width_ub); - dml2_printf("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, *p->PixelPTEBytesPerRow); - dml2_printf("DML::%s: PixelPTEBytesPerRowStorage = %u\n", __func__, *p->PixelPTEBytesPerRowStorage); - dml2_printf("DML::%s: PTEBufferSizeInRequests = %u\n", __func__, p->PTEBufferSizeInRequests); - dml2_printf("DML::%s: dpte_row_height_one_row_per_frame = %u\n", __func__, *p->dpte_row_height_one_row_per_frame); - dml2_printf("DML::%s: dpte_row_width_ub_one_row_per_frame = %u\n", __func__, *p->dpte_row_width_ub_one_row_per_frame); - dml2_printf("DML::%s: PixelPTEBytesPerRow_one_row_per_frame = %u\n", __func__, *p->PixelPTEBytesPerRow_one_row_per_frame); + DML_LOG_VERBOSE("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, p->GPUVMMinPageSizeKBytes); + DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable); + DML_LOG_VERBOSE("DML::%s: meta_row_height = %u\n", __func__, *p->meta_row_height); + DML_LOG_VERBOSE("DML::%s: dpte_row_height = %u\n", __func__, *p->dpte_row_height); + DML_LOG_VERBOSE("DML::%s: dpte_row_height_linear = %u\n", __func__, *p->dpte_row_height_linear); + DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u\n", __func__, *p->dpte_row_width_ub); + DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, *p->PixelPTEBytesPerRow); + DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRowStorage = %u\n", __func__, *p->PixelPTEBytesPerRowStorage); + DML_LOG_VERBOSE("DML::%s: PTEBufferSizeInRequests = %u\n", __func__, p->PTEBufferSizeInRequests); + DML_LOG_VERBOSE("DML::%s: dpte_row_height_one_row_per_frame = %u\n", __func__, *p->dpte_row_height_one_row_per_frame); + DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub_one_row_per_frame = %u\n", __func__, *p->dpte_row_width_ub_one_row_per_frame); + DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow_one_row_per_frame = %u\n", __func__, *p->PixelPTEBytesPerRow_one_row_per_frame); #endif return vm_bytes; @@ -1894,12 +1837,12 @@ static unsigned int CalculatePrefetchSourceLines( double numLines = 0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VRatio = %f\n", __func__, VRatio); - dml2_printf("DML::%s: VTaps = %u\n", __func__, VTaps); - dml2_printf("DML::%s: ViewportXStart = %u\n", __func__, ViewportXStart); - dml2_printf("DML::%s: ViewportYStart = %u\n", __func__, ViewportYStart); - dml2_printf("DML::%s: ViewportStationary = %u\n", __func__, ViewportStationary); - dml2_printf("DML::%s: SwathHeight = %u\n", __func__, SwathHeight); + DML_LOG_VERBOSE("DML::%s: VRatio = %f\n", __func__, VRatio); + DML_LOG_VERBOSE("DML::%s: VTaps = %u\n", __func__, VTaps); + DML_LOG_VERBOSE("DML::%s: ViewportXStart = %u\n", __func__, ViewportXStart); + DML_LOG_VERBOSE("DML::%s: ViewportYStart = %u\n", __func__, ViewportYStart); + DML_LOG_VERBOSE("DML::%s: ViewportStationary = %u\n", __func__, ViewportStationary); + DML_LOG_VERBOSE("DML::%s: SwathHeight = %u\n", __func__, SwathHeight); #endif if (ProgressiveToInterlaceUnitInOPP) *VInitPreFill = (unsigned int)(math_floor2((VRatio + (double)VTaps + 1) / 2.0, 1)); @@ -1934,11 +1877,11 @@ static unsigned int CalculatePrefetchSourceLines( numLines = *MaxNumSwath * SwathHeight + MaxPartialSwath; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: vp_start_rot = %u\n", __func__, vp_start_rot); - dml2_printf("DML::%s: VInitPreFill = %u\n", __func__, *VInitPreFill); - dml2_printf("DML::%s: MaxPartialSwath = %u\n", __func__, MaxPartialSwath); - dml2_printf("DML::%s: MaxNumSwath = %u\n", __func__, *MaxNumSwath); - dml2_printf("DML::%s: Prefetch source lines = %3.2f\n", __func__, numLines); + DML_LOG_VERBOSE("DML::%s: vp_start_rot = %u\n", __func__, vp_start_rot); + DML_LOG_VERBOSE("DML::%s: VInitPreFill = %u\n", __func__, *VInitPreFill); + DML_LOG_VERBOSE("DML::%s: MaxPartialSwath = %u\n", __func__, MaxPartialSwath); + DML_LOG_VERBOSE("DML::%s: MaxNumSwath = %u\n", __func__, *MaxNumSwath); + DML_LOG_VERBOSE("DML::%s: Prefetch source lines = %3.2f\n", __func__, numLines); #endif return (unsigned int)(numLines); @@ -2007,8 +1950,8 @@ static void CalculateMALLUseForStaticScreen( if (is_using_mall_for_ss[k]) TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, is_using_mall_for_ss = %u\n", __func__, k, is_using_mall_for_ss[k]); - dml2_printf("DML::%s: k=%u, TotalSurfaceSizeInMALL = %u\n", __func__, k, TotalSurfaceSizeInMALL); + DML_LOG_VERBOSE("DML::%s: k=%u, is_using_mall_for_ss = %u\n", __func__, k, is_using_mall_for_ss[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TotalSurfaceSizeInMALL = %u\n", __func__, k, TotalSurfaceSizeInMALL); #endif } @@ -2022,7 +1965,7 @@ static void CalculateMALLUseForStaticScreen( (!CanAddAnotherSurfaceToMALL || SurfaceSizeInMALL[k] < SurfaceSizeInMALL[SurfaceToAddToMALL])) { CanAddAnotherSurfaceToMALL = true; SurfaceToAddToMALL = k; - dml2_printf("DML::%s: k=%u, UseMALLForStaticScreen = %u (dis, en, optimize)\n", __func__, k, display_cfg->plane_descriptors[k].overrides.refresh_from_mall); + DML_LOG_VERBOSE("DML::%s: k=%u, UseMALLForStaticScreen = %u (dis, en, optimize)\n", __func__, k, display_cfg->plane_descriptors[k].overrides.refresh_from_mall); } } if (CanAddAnotherSurfaceToMALL) { @@ -2030,8 +1973,8 @@ static void CalculateMALLUseForStaticScreen( TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[SurfaceToAddToMALL]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: SurfaceToAddToMALL = %u\n", __func__, SurfaceToAddToMALL); - dml2_printf("DML::%s: TotalSurfaceSizeInMALL = %u\n", __func__, TotalSurfaceSizeInMALL); + DML_LOG_VERBOSE("DML::%s: SurfaceToAddToMALL = %u\n", __func__, SurfaceToAddToMALL); + DML_LOG_VERBOSE("DML::%s: TotalSurfaceSizeInMALL = %u\n", __func__, TotalSurfaceSizeInMALL); #endif } } @@ -2203,15 +2146,15 @@ static void CalculateDCCConfiguration( segment_order_vert_contiguous_chroma = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DCCEnabled = %u\n", __func__, DCCEnabled); - dml2_printf("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte); - dml2_printf("DML::%s: DETBufferSizeForDCC = %u\n", __func__, DETBufferSizeForDCC); - dml2_printf("DML::%s: req128_horz_wc_l = %u\n", __func__, req128_horz_wc_l); - dml2_printf("DML::%s: req128_horz_wc_c = %u\n", __func__, req128_horz_wc_c); - dml2_printf("DML::%s: full_swath_bytes_horz_wc_l = %u\n", __func__, full_swath_bytes_horz_wc_l); - dml2_printf("DML::%s: full_swath_bytes_vert_wc_c = %u\n", __func__, full_swath_bytes_vert_wc_c); - dml2_printf("DML::%s: segment_order_horz_contiguous_luma = %u\n", __func__, segment_order_horz_contiguous_luma); - dml2_printf("DML::%s: segment_order_horz_contiguous_chroma = %u\n", __func__, segment_order_horz_contiguous_chroma); + DML_LOG_VERBOSE("DML::%s: DCCEnabled = %u\n", __func__, DCCEnabled); + DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte); + DML_LOG_VERBOSE("DML::%s: DETBufferSizeForDCC = %u\n", __func__, DETBufferSizeForDCC); + DML_LOG_VERBOSE("DML::%s: req128_horz_wc_l = %u\n", __func__, req128_horz_wc_l); + DML_LOG_VERBOSE("DML::%s: req128_horz_wc_c = %u\n", __func__, req128_horz_wc_c); + DML_LOG_VERBOSE("DML::%s: full_swath_bytes_horz_wc_l = %u\n", __func__, full_swath_bytes_horz_wc_l); + DML_LOG_VERBOSE("DML::%s: full_swath_bytes_vert_wc_c = %u\n", __func__, full_swath_bytes_vert_wc_c); + DML_LOG_VERBOSE("DML::%s: segment_order_horz_contiguous_luma = %u\n", __func__, segment_order_horz_contiguous_luma); + DML_LOG_VERBOSE("DML::%s: segment_order_horz_contiguous_chroma = %u\n", __func__, segment_order_horz_contiguous_chroma); #endif if (DCCProgrammingAssumesScanDirectionUnknown == true) { if (req128_horz_wc_l == 0 && req128_vert_wc_l == 0) { @@ -2301,12 +2244,12 @@ static void CalculateDCCConfiguration( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: MaxUncompressedBlockLuma = %u\n", __func__, *MaxUncompressedBlockLuma); - dml2_printf("DML::%s: MaxCompressedBlockLuma = %u\n", __func__, *MaxCompressedBlockLuma); - dml2_printf("DML::%s: IndependentBlockLuma = %u\n", __func__, *IndependentBlockLuma); - dml2_printf("DML::%s: MaxUncompressedBlockChroma = %u\n", __func__, *MaxUncompressedBlockChroma); - dml2_printf("DML::%s: MaxCompressedBlockChroma = %u\n", __func__, *MaxCompressedBlockChroma); - dml2_printf("DML::%s: IndependentBlockChroma = %u\n", __func__, *IndependentBlockChroma); + DML_LOG_VERBOSE("DML::%s: MaxUncompressedBlockLuma = %u\n", __func__, *MaxUncompressedBlockLuma); + DML_LOG_VERBOSE("DML::%s: MaxCompressedBlockLuma = %u\n", __func__, *MaxCompressedBlockLuma); + DML_LOG_VERBOSE("DML::%s: IndependentBlockLuma = %u\n", __func__, *IndependentBlockLuma); + DML_LOG_VERBOSE("DML::%s: MaxUncompressedBlockChroma = %u\n", __func__, *MaxUncompressedBlockChroma); + DML_LOG_VERBOSE("DML::%s: MaxCompressedBlockChroma = %u\n", __func__, *MaxCompressedBlockChroma); + DML_LOG_VERBOSE("DML::%s: IndependentBlockChroma = %u\n", __func__, *IndependentBlockChroma); #endif } @@ -2326,26 +2269,26 @@ static void calculate_mcache_row_bytes( unsigned int mvmpg_per_mcache; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: num_chans = %u\n", __func__, p->num_chans); - dml2_printf("DML::%s: mem_word_bytes = %u\n", __func__, p->mem_word_bytes); - dml2_printf("DML::%s: mcache_line_size_bytes = %u\n", __func__, p->mcache_line_size_bytes); - dml2_printf("DML::%s: mcache_size_bytes = %u\n", __func__, p->mcache_size_bytes); - dml2_printf("DML::%s: gpuvm_enable = %u\n", __func__, p->gpuvm_enable); - dml2_printf("DML::%s: gpuvm_page_size_kbytes = %u\n", __func__, p->gpuvm_page_size_kbytes); - dml2_printf("DML::%s: vp_stationary = %u\n", __func__, p->vp_stationary); - dml2_printf("DML::%s: tiling_mode = %u\n", __func__, p->tiling_mode); - dml2_printf("DML::%s: vp_start_x = %u\n", __func__, p->vp_start_x); - dml2_printf("DML::%s: vp_start_y = %u\n", __func__, p->vp_start_y); - dml2_printf("DML::%s: full_vp_width = %u\n", __func__, p->full_vp_width); - dml2_printf("DML::%s: full_vp_height = %u\n", __func__, p->full_vp_height); - dml2_printf("DML::%s: blk_width = %u\n", __func__, p->blk_width); - dml2_printf("DML::%s: blk_height = %u\n", __func__, p->blk_height); - dml2_printf("DML::%s: vmpg_width = %u\n", __func__, p->vmpg_width); - dml2_printf("DML::%s: vmpg_height = %u\n", __func__, p->vmpg_height); - dml2_printf("DML::%s: full_swath_bytes = %u\n", __func__, p->full_swath_bytes); -#endif - DML2_ASSERT(p->mcache_line_size_bytes != 0); - DML2_ASSERT(p->mcache_size_bytes != 0); + DML_LOG_VERBOSE("DML::%s: num_chans = %u\n", __func__, p->num_chans); + DML_LOG_VERBOSE("DML::%s: mem_word_bytes = %u\n", __func__, p->mem_word_bytes); + DML_LOG_VERBOSE("DML::%s: mcache_line_size_bytes = %u\n", __func__, p->mcache_line_size_bytes); + DML_LOG_VERBOSE("DML::%s: mcache_size_bytes = %u\n", __func__, p->mcache_size_bytes); + DML_LOG_VERBOSE("DML::%s: gpuvm_enable = %u\n", __func__, p->gpuvm_enable); + DML_LOG_VERBOSE("DML::%s: gpuvm_page_size_kbytes = %u\n", __func__, p->gpuvm_page_size_kbytes); + DML_LOG_VERBOSE("DML::%s: vp_stationary = %u\n", __func__, p->vp_stationary); + DML_LOG_VERBOSE("DML::%s: tiling_mode = %u\n", __func__, p->tiling_mode); + DML_LOG_VERBOSE("DML::%s: vp_start_x = %u\n", __func__, p->vp_start_x); + DML_LOG_VERBOSE("DML::%s: vp_start_y = %u\n", __func__, p->vp_start_y); + DML_LOG_VERBOSE("DML::%s: full_vp_width = %u\n", __func__, p->full_vp_width); + DML_LOG_VERBOSE("DML::%s: full_vp_height = %u\n", __func__, p->full_vp_height); + DML_LOG_VERBOSE("DML::%s: blk_width = %u\n", __func__, p->blk_width); + DML_LOG_VERBOSE("DML::%s: blk_height = %u\n", __func__, p->blk_height); + DML_LOG_VERBOSE("DML::%s: vmpg_width = %u\n", __func__, p->vmpg_width); + DML_LOG_VERBOSE("DML::%s: vmpg_height = %u\n", __func__, p->vmpg_height); + DML_LOG_VERBOSE("DML::%s: full_swath_bytes = %u\n", __func__, p->full_swath_bytes); +#endif + DML_ASSERT(p->mcache_line_size_bytes != 0); + DML_ASSERT(p->mcache_size_bytes != 0); *p->mvmpg_width = 0; *p->mvmpg_height = 0; @@ -2370,8 +2313,8 @@ static void calculate_mcache_row_bytes( *p->mvmpg_width = p->vmpg_width; *p->mvmpg_height = p->vmpg_height; } else if (!((blk_bytes == 65536) && (vmpg_bytes == 4096))) { - dml2_printf("ERROR: DML::%s: Tiling size and vm page size combination not supported\n", __func__); - DML2_ASSERT(0); + DML_LOG_VERBOSE("ERROR: DML::%s: Tiling size and vm page size combination not supported\n", __func__); + DML_ASSERT(0); } } @@ -2439,25 +2382,25 @@ static void calculate_mcache_row_bytes( *p->mvmpg_per_mcache_lb = (unsigned int)math_floor2(mvmpg_per_mcache, 1); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: gpuvm_enable = %u\n", __func__, p->gpuvm_enable); - dml2_printf("DML::%s: vmpg_bytes = %u\n", __func__, vmpg_bytes); - dml2_printf("DML::%s: blk_bytes = %u\n", __func__, blk_bytes); - dml2_printf("DML::%s: meta_per_mvmpg_per_channel = %f\n", __func__, meta_per_mvmpg_per_channel); - dml2_printf("DML::%s: mvmpg_per_row_ub = %u\n", __func__, mvmpg_per_row_ub); - dml2_printf("DML::%s: meta_row_width_ub = %u\n", __func__, *p->meta_row_width_ub); - dml2_printf("DML::%s: mvmpg_width = %u\n", __func__, *p->mvmpg_width); - dml2_printf("DML::%s: mvmpg_height = %u\n", __func__, *p->mvmpg_height); - dml2_printf("DML::%s: dcc_dram_bw_nom_overhead_factor = %f\n", __func__, *p->dcc_dram_bw_nom_overhead_factor); - dml2_printf("DML::%s: dcc_dram_bw_pref_overhead_factor = %f\n", __func__, *p->dcc_dram_bw_pref_overhead_factor); + DML_LOG_VERBOSE("DML::%s: gpuvm_enable = %u\n", __func__, p->gpuvm_enable); + DML_LOG_VERBOSE("DML::%s: vmpg_bytes = %u\n", __func__, vmpg_bytes); + DML_LOG_VERBOSE("DML::%s: blk_bytes = %u\n", __func__, blk_bytes); + DML_LOG_VERBOSE("DML::%s: meta_per_mvmpg_per_channel = %f\n", __func__, meta_per_mvmpg_per_channel); + DML_LOG_VERBOSE("DML::%s: mvmpg_per_row_ub = %u\n", __func__, mvmpg_per_row_ub); + DML_LOG_VERBOSE("DML::%s: meta_row_width_ub = %u\n", __func__, *p->meta_row_width_ub); + DML_LOG_VERBOSE("DML::%s: mvmpg_width = %u\n", __func__, *p->mvmpg_width); + DML_LOG_VERBOSE("DML::%s: mvmpg_height = %u\n", __func__, *p->mvmpg_height); + DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_nom_overhead_factor = %f\n", __func__, *p->dcc_dram_bw_nom_overhead_factor); + DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_pref_overhead_factor = %f\n", __func__, *p->dcc_dram_bw_pref_overhead_factor); #endif } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: mcache_row_bytes = %u\n", __func__, *p->mcache_row_bytes); - dml2_printf("DML::%s: mcache_row_bytes_per_channel = %u\n", __func__, *p->mcache_row_bytes_per_channel); - dml2_printf("DML::%s: num_mcaches = %u\n", __func__, *p->num_mcaches); + DML_LOG_VERBOSE("DML::%s: mcache_row_bytes = %u\n", __func__, *p->mcache_row_bytes); + DML_LOG_VERBOSE("DML::%s: mcache_row_bytes_per_channel = %u\n", __func__, *p->mcache_row_bytes_per_channel); + DML_LOG_VERBOSE("DML::%s: num_mcaches = %u\n", __func__, *p->num_mcaches); #endif - DML2_ASSERT(*p->num_mcaches > 0); + DML_ASSERT(*p->num_mcaches > 0); } static void calculate_mcache_setting( @@ -2523,7 +2466,7 @@ static void calculate_mcache_setting( l->l_p.mvmpg_per_mcache_lb = &l->mvmpg_per_mcache_lb_l; calculate_mcache_row_bytes(scratch, &l->l_p); - DML2_ASSERT(*p->num_mcaches_l > 0); + DML_ASSERT(*p->num_mcaches_l > 0); if (l->is_dual_plane) { l->c_p.num_chans = p->num_chans; @@ -2559,7 +2502,7 @@ static void calculate_mcache_setting( l->c_p.mvmpg_per_mcache_lb = &l->mvmpg_per_mcache_lb_c; calculate_mcache_row_bytes(scratch, &l->c_p); - DML2_ASSERT(*p->num_mcaches_c > 0); + DML_ASSERT(*p->num_mcaches_c > 0); } // Sharing for iMALL access @@ -2598,28 +2541,28 @@ static void calculate_mcache_setting( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: imall_enable = %u\n", __func__, p->imall_enable); - dml2_printf("DML::%s: is_dual_plane = %u\n", __func__, l->is_dual_plane); - dml2_printf("DML::%s: surf_vert = %u\n", __func__, p->surf_vert); - dml2_printf("DML::%s: mvmpg_width_l = %u\n", __func__, l->mvmpg_width_l); - dml2_printf("DML::%s: mvmpg_height_l = %u\n", __func__, l->mvmpg_height_l); - dml2_printf("DML::%s: mcache_remainder_l = %f\n", __func__, l->mcache_remainder_l); - dml2_printf("DML::%s: num_mcaches_l = %u\n", __func__, *p->num_mcaches_l); - dml2_printf("DML::%s: avg_mcache_element_size_l = %u\n", __func__, l->avg_mcache_element_size_l); - dml2_printf("DML::%s: mvmpg_access_width_l = %u\n", __func__, l->mvmpg_access_width_l); - dml2_printf("DML::%s: mall_comb_mcache_l = %u\n", __func__, *p->mall_comb_mcache_l); + DML_LOG_VERBOSE("DML::%s: imall_enable = %u\n", __func__, p->imall_enable); + DML_LOG_VERBOSE("DML::%s: is_dual_plane = %u\n", __func__, l->is_dual_plane); + DML_LOG_VERBOSE("DML::%s: surf_vert = %u\n", __func__, p->surf_vert); + DML_LOG_VERBOSE("DML::%s: mvmpg_width_l = %u\n", __func__, l->mvmpg_width_l); + DML_LOG_VERBOSE("DML::%s: mvmpg_height_l = %u\n", __func__, l->mvmpg_height_l); + DML_LOG_VERBOSE("DML::%s: mcache_remainder_l = %f\n", __func__, l->mcache_remainder_l); + DML_LOG_VERBOSE("DML::%s: num_mcaches_l = %u\n", __func__, *p->num_mcaches_l); + DML_LOG_VERBOSE("DML::%s: avg_mcache_element_size_l = %u\n", __func__, l->avg_mcache_element_size_l); + DML_LOG_VERBOSE("DML::%s: mvmpg_access_width_l = %u\n", __func__, l->mvmpg_access_width_l); + DML_LOG_VERBOSE("DML::%s: mall_comb_mcache_l = %u\n", __func__, *p->mall_comb_mcache_l); if (l->is_dual_plane) { - dml2_printf("DML::%s: mvmpg_width_c = %u\n", __func__, l->mvmpg_width_c); - dml2_printf("DML::%s: mvmpg_height_c = %u\n", __func__, l->mvmpg_height_c); - dml2_printf("DML::%s: mcache_remainder_c = %f\n", __func__, l->mcache_remainder_c); - dml2_printf("DML::%s: luma_time_factor = %f\n", __func__, l->luma_time_factor); - dml2_printf("DML::%s: num_mcaches_c = %u\n", __func__, *p->num_mcaches_c); - dml2_printf("DML::%s: avg_mcache_element_size_c = %u\n", __func__, l->avg_mcache_element_size_c); - dml2_printf("DML::%s: mvmpg_access_width_c = %u\n", __func__, l->mvmpg_access_width_c); - dml2_printf("DML::%s: mall_comb_mcache_c = %u\n", __func__, *p->mall_comb_mcache_c); - dml2_printf("DML::%s: lc_comb_last_mcache_size = %u\n", __func__, l->lc_comb_last_mcache_size); - dml2_printf("DML::%s: lc_comb_mcache = %u\n", __func__, *p->lc_comb_mcache); + DML_LOG_VERBOSE("DML::%s: mvmpg_width_c = %u\n", __func__, l->mvmpg_width_c); + DML_LOG_VERBOSE("DML::%s: mvmpg_height_c = %u\n", __func__, l->mvmpg_height_c); + DML_LOG_VERBOSE("DML::%s: mcache_remainder_c = %f\n", __func__, l->mcache_remainder_c); + DML_LOG_VERBOSE("DML::%s: luma_time_factor = %f\n", __func__, l->luma_time_factor); + DML_LOG_VERBOSE("DML::%s: num_mcaches_c = %u\n", __func__, *p->num_mcaches_c); + DML_LOG_VERBOSE("DML::%s: avg_mcache_element_size_c = %u\n", __func__, l->avg_mcache_element_size_c); + DML_LOG_VERBOSE("DML::%s: mvmpg_access_width_c = %u\n", __func__, l->mvmpg_access_width_c); + DML_LOG_VERBOSE("DML::%s: mall_comb_mcache_c = %u\n", __func__, *p->mall_comb_mcache_c); + DML_LOG_VERBOSE("DML::%s: lc_comb_last_mcache_size = %u\n", __func__, l->lc_comb_last_mcache_size); + DML_LOG_VERBOSE("DML::%s: lc_comb_mcache = %u\n", __func__, *p->lc_comb_mcache); } #endif // calculate split_coordinate @@ -2639,11 +2582,11 @@ static void calculate_mcache_setting( } #ifdef __DML_VBA_DEBUG__ for (n = 0; n < *p->num_mcaches_l; n++) - dml2_printf("DML::%s: mcache_offsets_l[%u] = %u\n", __func__, n, p->mcache_offsets_l[n]); + DML_LOG_VERBOSE("DML::%s: mcache_offsets_l[%u] = %u\n", __func__, n, p->mcache_offsets_l[n]); if (l->is_dual_plane) { for (n = 0; n < *p->num_mcaches_c; n++) - dml2_printf("DML::%s: mcache_offsets_c[%u] = %u\n", __func__, n, p->mcache_offsets_c[n]); + DML_LOG_VERBOSE("DML::%s: mcache_offsets_c[%u] = %u\n", __func__, n, p->mcache_offsets_c[n]); } #endif @@ -2660,10 +2603,10 @@ static void calculate_mcache_setting( #ifdef __DML_VBA_DEBUG__ for (n = 0; n < *p->num_mcaches_l; n++) - dml2_printf("DML::%s: mcache_offsets_l[%u] = %u\n", __func__, n, p->mcache_offsets_l[n]); + DML_LOG_VERBOSE("DML::%s: mcache_offsets_l[%u] = %u\n", __func__, n, p->mcache_offsets_l[n]); for (n = 0; n < *p->num_mcaches_c; n++) - dml2_printf("DML::%s: mcache_offsets_c[%u] = %u\n", __func__, n, p->mcache_offsets_c[n]); + DML_LOG_VERBOSE("DML::%s: mcache_offsets_c[%u] = %u\n", __func__, n, p->mcache_offsets_c[n]); #endif } @@ -2694,8 +2637,8 @@ static void calculate_mall_bw_overhead_factor( mall_prefetch_dram_overhead_factor[k] = 2.0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, mall_prefetch_sdp_overhead_factor = %f\n", __func__, k, mall_prefetch_sdp_overhead_factor[k]); - dml2_printf("DML::%s: k=%u, mall_prefetch_dram_overhead_factor = %f\n", __func__, k, mall_prefetch_dram_overhead_factor[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, mall_prefetch_sdp_overhead_factor = %f\n", __func__, k, mall_prefetch_sdp_overhead_factor[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, mall_prefetch_dram_overhead_factor = %f\n", __func__, k, mall_prefetch_dram_overhead_factor[k]); #endif } } @@ -2772,22 +2715,20 @@ static double dml_get_return_bandwidth_available( else // dml2_core_internal_bw_dram return_bw_mbps = derate_dram_bandwidth; -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: is_avg_bw = %u\n", __func__, is_avg_bw); - dml2_printf("DML::%s: is_hvm_en = %u\n", __func__, is_hvm_en); - dml2_printf("DML::%s: is_hvm_only = %u\n", __func__, is_hvm_only); - dml2_printf("DML::%s: state_type = %s\n", __func__, dml2_core_internal_soc_state_type_str(state_type)); - dml2_printf("DML::%s: bw_type = %s\n", __func__, dml2_core_internal_bw_type_str(bw_type)); - dml2_printf("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz); - dml2_printf("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz); - dml2_printf("DML::%s: ideal_sdp_bandwidth = %f\n", __func__, ideal_sdp_bandwidth); - dml2_printf("DML::%s: ideal_fabric_bandwidth = %f\n", __func__, ideal_fabric_bandwidth); - dml2_printf("DML::%s: ideal_dram_bandwidth = %f\n", __func__, ideal_dram_bandwidth); - dml2_printf("DML::%s: derate_sdp_bandwidth = %f (derate %f)\n", __func__, derate_sdp_bandwidth, derate_sdp_factor); - dml2_printf("DML::%s: derate_fabric_bandwidth = %f (derate %f)\n", __func__, derate_fabric_bandwidth, derate_fabric_factor); - dml2_printf("DML::%s: derate_dram_bandwidth = %f (derate %f)\n", __func__, derate_dram_bandwidth, derate_dram_factor); - dml2_printf("DML::%s: return_bw_mbps = %f\n", __func__, return_bw_mbps); -#endif + DML_LOG_VERBOSE("DML::%s: is_avg_bw = %u\n", __func__, is_avg_bw); + DML_LOG_VERBOSE("DML::%s: is_hvm_en = %u\n", __func__, is_hvm_en); + DML_LOG_VERBOSE("DML::%s: is_hvm_only = %u\n", __func__, is_hvm_only); + DML_LOG_VERBOSE("DML::%s: state_type = %s\n", __func__, dml2_core_internal_soc_state_type_str(state_type)); + DML_LOG_VERBOSE("DML::%s: bw_type = %s\n", __func__, dml2_core_internal_bw_type_str(bw_type)); + DML_LOG_VERBOSE("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz); + DML_LOG_VERBOSE("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz); + DML_LOG_VERBOSE("DML::%s: ideal_sdp_bandwidth = %f\n", __func__, ideal_sdp_bandwidth); + DML_LOG_VERBOSE("DML::%s: ideal_fabric_bandwidth = %f\n", __func__, ideal_fabric_bandwidth); + DML_LOG_VERBOSE("DML::%s: ideal_dram_bandwidth = %f\n", __func__, ideal_dram_bandwidth); + DML_LOG_VERBOSE("DML::%s: derate_sdp_bandwidth = %f (derate %f)\n", __func__, derate_sdp_bandwidth, derate_sdp_factor); + DML_LOG_VERBOSE("DML::%s: derate_fabric_bandwidth = %f (derate %f)\n", __func__, derate_fabric_bandwidth, derate_fabric_factor); + DML_LOG_VERBOSE("DML::%s: derate_dram_bandwidth = %f (derate %f)\n", __func__, derate_dram_bandwidth, derate_dram_factor); + DML_LOG_VERBOSE("DML::%s: return_bw_mbps = %f\n", __func__, return_bw_mbps); return return_bw_mbps; } @@ -2807,9 +2748,9 @@ static noinline_for_stack void calculate_bandwidth_available( { unsigned int n, m; - dml2_printf("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz); - dml2_printf("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz); - dml2_printf("DML::%s: dram_bw_mbps = %f\n", __func__, dram_bw_mbps); + DML_LOG_VERBOSE("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz); + DML_LOG_VERBOSE("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz); + DML_LOG_VERBOSE("DML::%s: dram_bw_mbps = %f\n", __func__, dram_bw_mbps); // Calculate all the bandwidth availabe for (m = 0; m < dml2_core_internal_soc_state_max; m++) { @@ -2828,8 +2769,8 @@ static noinline_for_stack void calculate_bandwidth_available( #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: avg_bandwidth_available[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), avg_bandwidth_available[m][n]); - dml2_printf("DML::%s: urg_bandwidth_available[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), urg_bandwidth_available[m][n]); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_available[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), avg_bandwidth_available[m][n]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), urg_bandwidth_available[m][n]); #endif // urg_bandwidth_available_vm_only is indexed by soc_state @@ -2843,9 +2784,9 @@ static noinline_for_stack void calculate_bandwidth_available( urg_bandwidth_available_min[m] = math_min2(urg_bandwidth_available[m][dml2_core_internal_bw_dram], urg_bandwidth_available[m][dml2_core_internal_bw_sdp]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: avg_bandwidth_available_min[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), avg_bandwidth_available_min[m]); - dml2_printf("DML::%s: urg_bandwidth_available_min[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), urg_bandwidth_available_min[m]); - dml2_printf("DML::%s: urg_bandwidth_available_vm_only[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), urg_bandwidth_available_vm_only[n]); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_available_min[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), avg_bandwidth_available_min[m]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_min[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), urg_bandwidth_available_min[m]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_vm_only[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), urg_bandwidth_available_vm_only[n]); #endif } } @@ -2879,13 +2820,13 @@ static void calculate_avg_bandwidth_required( // SysActive and SVP Prefetch AVG bandwidth Check for (k = 0; k < num_active_planes; ++k) { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: plane %0d\n", __func__, k); - dml2_printf("DML::%s: ReadBandwidthLuma=%f\n", __func__, ReadBandwidthLuma[k]); - dml2_printf("DML::%s: ReadBandwidthChroma=%f\n", __func__, ReadBandwidthChroma[k]); - dml2_printf("DML::%s: dcc_dram_bw_nom_overhead_factor_p0=%f\n", __func__, dcc_dram_bw_nom_overhead_factor_p0[k]); - dml2_printf("DML::%s: dcc_dram_bw_nom_overhead_factor_p1=%f\n", __func__, dcc_dram_bw_nom_overhead_factor_p1[k]); - dml2_printf("DML::%s: mall_prefetch_dram_overhead_factor=%f\n", __func__, mall_prefetch_dram_overhead_factor[k]); - dml2_printf("DML::%s: mall_prefetch_sdp_overhead_factor=%f\n", __func__, mall_prefetch_sdp_overhead_factor[k]); + DML_LOG_VERBOSE("DML::%s: plane %0d\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: ReadBandwidthLuma=%f\n", __func__, ReadBandwidthLuma[k]); + DML_LOG_VERBOSE("DML::%s: ReadBandwidthChroma=%f\n", __func__, ReadBandwidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_nom_overhead_factor_p0=%f\n", __func__, dcc_dram_bw_nom_overhead_factor_p0[k]); + DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_nom_overhead_factor_p1=%f\n", __func__, dcc_dram_bw_nom_overhead_factor_p1[k]); + DML_LOG_VERBOSE("DML::%s: mall_prefetch_dram_overhead_factor=%f\n", __func__, mall_prefetch_dram_overhead_factor[k]); + DML_LOG_VERBOSE("DML::%s: mall_prefetch_sdp_overhead_factor=%f\n", __func__, mall_prefetch_sdp_overhead_factor[k]); #endif sdp_overhead_factor = mall_prefetch_sdp_overhead_factor[k]; @@ -2902,10 +2843,10 @@ static void calculate_avg_bandwidth_required( avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] += dram_overhead_factor_p0 * ReadBandwidthLuma[k] + dram_overhead_factor_p1 * ReadBandwidthChroma[k] + cursor_bw[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_sys_active), dml2_core_internal_bw_type_str(dml2_core_internal_bw_sdp), avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]); - dml2_printf("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_sys_active), dml2_core_internal_bw_type_str(dml2_core_internal_bw_dram), avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]); - dml2_printf("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_svp_prefetch), dml2_core_internal_bw_type_str(dml2_core_internal_bw_sdp), avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]); - dml2_printf("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_svp_prefetch), dml2_core_internal_bw_type_str(dml2_core_internal_bw_dram), avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_sys_active), dml2_core_internal_bw_type_str(dml2_core_internal_bw_sdp), avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_sys_active), dml2_core_internal_bw_type_str(dml2_core_internal_bw_dram), avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_svp_prefetch), dml2_core_internal_bw_type_str(dml2_core_internal_bw_sdp), avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_svp_prefetch), dml2_core_internal_bw_type_str(dml2_core_internal_bw_dram), avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]); #endif } } @@ -3080,10 +3021,10 @@ static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch, &p->MaxNumSwathY[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, vm_bytes_l = %u (before hvm level)\n", __func__, k, s->vm_bytes_l); - dml2_printf("DML::%s: k=%u, vm_bytes_c = %u (before hvm level)\n", __func__, k, s->vm_bytes_c); - dml2_printf("DML::%s: k=%u, meta_row_bytes_per_row_ub_l = %u\n", __func__, k, s->meta_row_bytes_per_row_ub_l[k]); - dml2_printf("DML::%s: k=%u, meta_row_bytes_per_row_ub_c = %u\n", __func__, k, s->meta_row_bytes_per_row_ub_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, vm_bytes_l = %u (before hvm level)\n", __func__, k, s->vm_bytes_l); + DML_LOG_VERBOSE("DML::%s: k=%u, vm_bytes_c = %u (before hvm level)\n", __func__, k, s->vm_bytes_c); + DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_bytes_per_row_ub_l = %u\n", __func__, k, s->meta_row_bytes_per_row_ub_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_bytes_per_row_ub_c = %u\n", __func__, k, s->meta_row_bytes_per_row_ub_c[k]); #endif p->vm_bytes[k] = (s->vm_bytes_l + s->vm_bytes_c) * (1 + 8 * s->HostVMDynamicLevels); p->meta_row_bytes[k] = s->meta_row_bytes_per_row_ub_l[k] + s->meta_row_bytes_per_row_ub_c[k]; @@ -3091,8 +3032,8 @@ static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch, p->meta_row_bytes_per_row_ub_c[k] = s->meta_row_bytes_per_row_ub_c[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, meta_row_bytes = %u\n", __func__, k, p->meta_row_bytes[k]); - dml2_printf("DML::%s: k=%u, vm_bytes = %u (after hvm level)\n", __func__, k, p->vm_bytes[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_bytes = %u\n", __func__, k, p->meta_row_bytes[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, vm_bytes = %u (after hvm level)\n", __func__, k, p->vm_bytes[k]); #endif if (s->PixelPTEBytesPerRowStorageY[k] <= 64 * s->PTEBufferSizeInRequestsForLuma[k] && s->PixelPTEBytesPerRowStorageC[k] <= 64 * s->PTEBufferSizeInRequestsForChroma[k]) { p->PTEBufferSizeNotExceeded[k] = true; @@ -3104,18 +3045,18 @@ static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch, s->PixelPTEBytesPerRowC_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForChroma[k]); #ifdef __DML_VBA_DEBUG__ if (p->PTEBufferSizeNotExceeded[k] == 0 || s->one_row_per_frame_fits_in_buffer[k] == 0) { - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowY = %u (before hvm level)\n", __func__, k, s->PixelPTEBytesPerRowY[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowC = %u (before hvm level)\n", __func__, k, s->PixelPTEBytesPerRowC[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowStorageY = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageY[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowStorageC = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageC[k]); - dml2_printf("DML::%s: k=%u, PTEBufferSizeInRequestsForLuma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForLuma[k]); - dml2_printf("DML::%s: k=%u, PTEBufferSizeInRequestsForChroma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForChroma[k]); - dml2_printf("DML::%s: k=%u, PTEBufferSizeNotExceeded (not one_row_per_frame) = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowY = %u (before hvm level)\n", __func__, k, s->PixelPTEBytesPerRowY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowC = %u (before hvm level)\n", __func__, k, s->PixelPTEBytesPerRowC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowStorageY = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowStorageC = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeInRequestsForLuma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeInRequestsForChroma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeNotExceeded (not one_row_per_frame) = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]); - dml2_printf("DML::%s: k=%u, HostVMDynamicLevels = %u\n", __func__, k, s->HostVMDynamicLevels); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowY_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowY_one_row_per_frame[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowC_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowC_one_row_per_frame[k]); - dml2_printf("DML::%s: k=%u, one_row_per_frame_fits_in_buffer = %u\n", __func__, k, s->one_row_per_frame_fits_in_buffer[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, HostVMDynamicLevels = %u\n", __func__, k, s->HostVMDynamicLevels); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowY_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowY_one_row_per_frame[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowC_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowC_one_row_per_frame[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, one_row_per_frame_fits_in_buffer = %u\n", __func__, k, s->one_row_per_frame_fits_in_buffer[k]); } #endif } @@ -3146,8 +3087,8 @@ static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch, for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) { p->DCCMetaBufferSizeNotExceeded[k] = true; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, SurfaceSizeInMALL = %u\n", __func__, k, p->SurfaceSizeInMALL[k]); - dml2_printf("DML::%s: k=%u, is_using_mall_for_ss = %u\n", __func__, k, p->is_using_mall_for_ss[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, SurfaceSizeInMALL = %u\n", __func__, k, p->SurfaceSizeInMALL[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, is_using_mall_for_ss = %u\n", __func__, k, p->is_using_mall_for_ss[k]); #endif p->use_one_row_for_frame[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->is_using_mall_for_ss[k] || (p->display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_main_pipe) || (dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) || (p->display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes > 64 && dml_is_vertical_rotation(p->myPipe[k].RotationAngle)); @@ -3170,9 +3111,9 @@ static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch, p->DCCMetaBufferSizeNotExceeded[k] = false; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, meta_row_bytes = %d\n", __func__, k, p->meta_row_bytes[k]); - dml2_printf("DML::%s: k=%d, DCCMetaBufferSizeBytes = %d\n", __func__, k, p->DCCMetaBufferSizeBytes); - dml2_printf("DML::%s: k=%d, DCCMetaBufferSizeNotExceeded = %d\n", __func__, k, p->DCCMetaBufferSizeNotExceeded[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, meta_row_bytes = %d\n", __func__, k, p->meta_row_bytes[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, DCCMetaBufferSizeBytes = %d\n", __func__, k, p->DCCMetaBufferSizeBytes); + DML_LOG_VERBOSE("DML::%s: k=%d, DCCMetaBufferSizeNotExceeded = %d\n", __func__, k, p->DCCMetaBufferSizeNotExceeded[k]); #endif } @@ -3209,20 +3150,20 @@ static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch, &p->dpte_row_bw[k], &p->meta_row_bw[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]); - dml2_printf("DML::%s: k=%u, use_one_row_for_frame_flip = %u\n", __func__, k, p->use_one_row_for_frame_flip[k]); - dml2_printf("DML::%s: k=%u, UseMALLForPStateChange = %u\n", __func__, k, p->display_cfg->plane_descriptors[k].overrides.legacy_svp_config); - dml2_printf("DML::%s: k=%u, dpte_row_height_luma = %u\n", __func__, k, p->dpte_row_height_luma[k]); - dml2_printf("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowY = %u (after hvm level)\n", __func__, k, s->PixelPTEBytesPerRowY[k]); - dml2_printf("DML::%s: k=%u, dpte_row_height_chroma = %u\n", __func__, k, p->dpte_row_height_chroma[k]); - dml2_printf("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRowC = %u (after hvm level)\n", __func__, k, s->PixelPTEBytesPerRowC[k]); - dml2_printf("DML::%s: k=%u, PixelPTEBytesPerRow = %u\n", __func__, k, p->PixelPTEBytesPerRow[k]); - dml2_printf("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]); - dml2_printf("DML::%s: k=%u, gpuvm_enable = %u\n", __func__, k, p->display_cfg->gpuvm_enable); - dml2_printf("DML::%s: k=%u, PTE_BUFFER_MODE = %u\n", __func__, k, p->PTE_BUFFER_MODE[k]); - dml2_printf("DML::%s: k=%u, BIGK_FRAGMENT_SIZE = %u\n", __func__, k, p->BIGK_FRAGMENT_SIZE[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, use_one_row_for_frame_flip = %u\n", __func__, k, p->use_one_row_for_frame_flip[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, UseMALLForPStateChange = %u\n", __func__, k, p->display_cfg->plane_descriptors[k].overrides.legacy_svp_config); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_height_luma = %u\n", __func__, k, p->dpte_row_height_luma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowY = %u (after hvm level)\n", __func__, k, s->PixelPTEBytesPerRowY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_height_chroma = %u\n", __func__, k, p->dpte_row_height_chroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowC = %u (after hvm level)\n", __func__, k, s->PixelPTEBytesPerRowC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRow = %u\n", __func__, k, p->PixelPTEBytesPerRow[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, gpuvm_enable = %u\n", __func__, k, p->display_cfg->gpuvm_enable); + DML_LOG_VERBOSE("DML::%s: k=%u, PTE_BUFFER_MODE = %u\n", __func__, k, p->PTE_BUFFER_MODE[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, BIGK_FRAGMENT_SIZE = %u\n", __func__, k, p->BIGK_FRAGMENT_SIZE[k]); #endif } } @@ -3257,19 +3198,19 @@ static double CalculateUrgentLatency( } #ifdef __DML_VBA_DEBUG__ if (qos_type == dml2_qos_param_type_dcn4x) { - dml2_printf("DML::%s: qos_type = %d\n", __func__, qos_type); - dml2_printf("DML::%s: urgent_ramp_uclk_cycles = %d\n", __func__, urgent_ramp_uclk_cycles); - dml2_printf("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz); - dml2_printf("DML::%s: umc_urgent_ramp_latency_margin = %f\n", __func__, umc_urgent_ramp_latency_margin); + DML_LOG_VERBOSE("DML::%s: qos_type = %d\n", __func__, qos_type); + DML_LOG_VERBOSE("DML::%s: urgent_ramp_uclk_cycles = %d\n", __func__, urgent_ramp_uclk_cycles); + DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: umc_urgent_ramp_latency_margin = %f\n", __func__, umc_urgent_ramp_latency_margin); } else { - dml2_printf("DML::%s: UrgentLatencyPixelDataOnly = %f\n", __func__, UrgentLatencyPixelDataOnly); - dml2_printf("DML::%s: UrgentLatencyPixelMixedWithVMData = %f\n", __func__, UrgentLatencyPixelMixedWithVMData); - dml2_printf("DML::%s: UrgentLatencyVMDataOnly = %f\n", __func__, UrgentLatencyVMDataOnly); - dml2_printf("DML::%s: UrgentLatencyAdjustmentFabricClockComponent = %f\n", __func__, UrgentLatencyAdjustmentFabricClockComponent); - dml2_printf("DML::%s: UrgentLatencyAdjustmentFabricClockReference = %f\n", __func__, UrgentLatencyAdjustmentFabricClockReference); + DML_LOG_VERBOSE("DML::%s: UrgentLatencyPixelDataOnly = %f\n", __func__, UrgentLatencyPixelDataOnly); + DML_LOG_VERBOSE("DML::%s: UrgentLatencyPixelMixedWithVMData = %f\n", __func__, UrgentLatencyPixelMixedWithVMData); + DML_LOG_VERBOSE("DML::%s: UrgentLatencyVMDataOnly = %f\n", __func__, UrgentLatencyVMDataOnly); + DML_LOG_VERBOSE("DML::%s: UrgentLatencyAdjustmentFabricClockComponent = %f\n", __func__, UrgentLatencyAdjustmentFabricClockComponent); + DML_LOG_VERBOSE("DML::%s: UrgentLatencyAdjustmentFabricClockReference = %f\n", __func__, UrgentLatencyAdjustmentFabricClockReference); } - dml2_printf("DML::%s: FabricClock = %f\n", __func__, FabricClock); - dml2_printf("DML::%s: UrgentLatency = %f\n", __func__, urgent_latency); + DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, FabricClock); + DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, urgent_latency); #endif return urgent_latency; } @@ -3296,18 +3237,18 @@ static double CalculateTripToMemory( #ifdef __DML_VBA_DEBUG__ if (qos_type == dml2_qos_param_type_dcn4x) { - dml2_printf("DML::%s: qos_type = %d\n", __func__, qos_type); - dml2_printf("DML::%s: max_round_trip_to_furthest_cs_fclk_cycles = %d\n", __func__, max_round_trip_to_furthest_cs_fclk_cycles); - dml2_printf("DML::%s: mall_overhead_fclk_cycles = %d\n", __func__, mall_overhead_fclk_cycles); - dml2_printf("DML::%s: trip_to_memory_uclk_cycles = %d\n", __func__, trip_to_memory_uclk_cycles); - dml2_printf("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz); - dml2_printf("DML::%s: FabricClock = %f\n", __func__, FabricClock); - dml2_printf("DML::%s: fabric_max_transport_latency_margin = %f\n", __func__, fabric_max_transport_latency_margin); - dml2_printf("DML::%s: umc_max_latency_margin = %f\n", __func__, umc_max_latency_margin); + DML_LOG_VERBOSE("DML::%s: qos_type = %d\n", __func__, qos_type); + DML_LOG_VERBOSE("DML::%s: max_round_trip_to_furthest_cs_fclk_cycles = %d\n", __func__, max_round_trip_to_furthest_cs_fclk_cycles); + DML_LOG_VERBOSE("DML::%s: mall_overhead_fclk_cycles = %d\n", __func__, mall_overhead_fclk_cycles); + DML_LOG_VERBOSE("DML::%s: trip_to_memory_uclk_cycles = %d\n", __func__, trip_to_memory_uclk_cycles); + DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, FabricClock); + DML_LOG_VERBOSE("DML::%s: fabric_max_transport_latency_margin = %f\n", __func__, fabric_max_transport_latency_margin); + DML_LOG_VERBOSE("DML::%s: umc_max_latency_margin = %f\n", __func__, umc_max_latency_margin); } else { - dml2_printf("DML::%s: UrgLatency = %f\n", __func__, UrgLatency); + DML_LOG_VERBOSE("DML::%s: UrgLatency = %f\n", __func__, UrgLatency); } - dml2_printf("DML::%s: trip_to_memory_us = %f\n", __func__, trip_to_memory_us); + DML_LOG_VERBOSE("DML::%s: trip_to_memory_us = %f\n", __func__, trip_to_memory_us); #endif @@ -3334,14 +3275,14 @@ static double CalculateMetaTripToMemory( #ifdef __DML_VBA_DEBUG__ if (qos_type == dml2_qos_param_type_dcn4x) { - dml2_printf("DML::%s: qos_type = %d\n", __func__, qos_type); - dml2_printf("DML::%s: meta_trip_to_memory_fclk_cycles = %d\n", __func__, meta_trip_to_memory_fclk_cycles); - dml2_printf("DML::%s: meta_trip_to_memory_uclk_cycles = %d\n", __func__, meta_trip_to_memory_uclk_cycles); - dml2_printf("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: qos_type = %d\n", __func__, qos_type); + DML_LOG_VERBOSE("DML::%s: meta_trip_to_memory_fclk_cycles = %d\n", __func__, meta_trip_to_memory_fclk_cycles); + DML_LOG_VERBOSE("DML::%s: meta_trip_to_memory_uclk_cycles = %d\n", __func__, meta_trip_to_memory_uclk_cycles); + DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz); } else { - dml2_printf("DML::%s: UrgLatency = %f\n", __func__, UrgLatency); + DML_LOG_VERBOSE("DML::%s: UrgLatency = %f\n", __func__, UrgLatency); } - dml2_printf("DML::%s: meta_trip_to_memory_us = %f\n", __func__, meta_trip_to_memory_us); + DML_LOG_VERBOSE("DML::%s: meta_trip_to_memory_us = %f\n", __func__, meta_trip_to_memory_us); #endif @@ -3358,7 +3299,6 @@ static void calculate_cursor_req_attributes( unsigned int *cursor_bytes_per_chunk, unsigned int *cursor_bytes) { - unsigned int cursor_pitch = 0; unsigned int cursor_bytes_per_req = 0; unsigned int cursor_width_bytes = 0; unsigned int cursor_height = 0; @@ -3366,10 +3306,6 @@ static void calculate_cursor_req_attributes( //SW determines the cursor pitch to support the maximum cursor_width that will be used but the following restrictions apply. //- For 2bpp, cursor_pitch = 256 pixels due to min cursor request size of 64B //- For 32 or 64 bpp, cursor_pitch = 64, 128 or 256 pixels depending on the cursor width - if (cursor_bpp == 2) - cursor_pitch = 256; - else - cursor_pitch = (unsigned int)1 << (unsigned int)math_ceil2(math_log((float)cursor_width, 2), 1); //The cursor requestor uses a cursor request size of 64B, 128B, or 256B depending on the cursor_width and cursor_bpp as follows. @@ -3409,8 +3345,8 @@ static void calculate_cursor_req_attributes( *cursor_lines_per_chunk = 1; } else { if (cursor_width > 0) { - dml2_printf("DML::%s: Invalid cursor_bpp = %d\n", __func__, cursor_bpp); - DML2_ASSERT(0); + DML_LOG_VERBOSE("DML::%s: Invalid cursor_bpp = %d\n", __func__, cursor_bpp); + DML_ASSERT(0); } } @@ -3421,15 +3357,15 @@ static void calculate_cursor_req_attributes( cursor_height = cursor_width; *cursor_bytes = *cursor_bytes_per_line * cursor_height; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: cursor_bpp = %d\n", __func__, cursor_bpp); - dml2_printf("DML::%s: cursor_width = %d\n", __func__, cursor_width); - dml2_printf("DML::%s: cursor_width_bytes = %d\n", __func__, cursor_width_bytes); - dml2_printf("DML::%s: cursor_bytes_per_req = %d\n", __func__, cursor_bytes_per_req); - dml2_printf("DML::%s: cursor_lines_per_chunk = %d\n", __func__, *cursor_lines_per_chunk); - dml2_printf("DML::%s: cursor_bytes_per_line = %d\n", __func__, *cursor_bytes_per_line); - dml2_printf("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, *cursor_bytes_per_chunk); - dml2_printf("DML::%s: cursor_bytes = %d\n", __func__, *cursor_bytes); - dml2_printf("DML::%s: cursor_pitch = %d\n", __func__, cursor_pitch); + DML_LOG_VERBOSE("DML::%s: cursor_bpp = %d\n", __func__, cursor_bpp); + DML_LOG_VERBOSE("DML::%s: cursor_width = %d\n", __func__, cursor_width); + DML_LOG_VERBOSE("DML::%s: cursor_width_bytes = %d\n", __func__, cursor_width_bytes); + DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_req = %d\n", __func__, cursor_bytes_per_req); + DML_LOG_VERBOSE("DML::%s: cursor_lines_per_chunk = %d\n", __func__, *cursor_lines_per_chunk); + DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_line = %d\n", __func__, *cursor_bytes_per_line); + DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, *cursor_bytes_per_chunk); + DML_LOG_VERBOSE("DML::%s: cursor_bytes = %d\n", __func__, *cursor_bytes); + DML_LOG_VERBOSE("DML::%s: cursor_pitch = %d\n", __func__, cursor_bpp == 2 ? 256 : (unsigned int)1 << (unsigned int)math_ceil2(math_log((float)cursor_width, 2), 1)); #endif } @@ -3460,13 +3396,13 @@ static void calculate_cursor_urgent_burst_factor( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: LinesInCursorBuffer = %u\n", __func__, LinesInCursorBuffer); - dml2_printf("DML::%s: CursorBufferSizeInTime = %f\n", __func__, CursorBufferSizeInTime); - dml2_printf("DML::%s: CursorBufferSize = %u (kbytes)\n", __func__, CursorBufferSize); - dml2_printf("DML::%s: cursor_bytes_per_chunk = %u\n", __func__, cursor_bytes_per_chunk); - dml2_printf("DML::%s: cursor_lines_per_chunk = %u\n", __func__, cursor_lines_per_chunk); - dml2_printf("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, *UrgentBurstFactorCursor); - dml2_printf("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding); + DML_LOG_VERBOSE("DML::%s: LinesInCursorBuffer = %u\n", __func__, LinesInCursorBuffer); + DML_LOG_VERBOSE("DML::%s: CursorBufferSizeInTime = %f\n", __func__, CursorBufferSizeInTime); + DML_LOG_VERBOSE("DML::%s: CursorBufferSize = %u (kbytes)\n", __func__, CursorBufferSize); + DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_chunk = %u\n", __func__, cursor_bytes_per_chunk); + DML_LOG_VERBOSE("DML::%s: cursor_lines_per_chunk = %u\n", __func__, cursor_lines_per_chunk); + DML_LOG_VERBOSE("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, *UrgentBurstFactorCursor); + DML_LOG_VERBOSE("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding); #endif } @@ -3501,15 +3437,15 @@ static void CalculateUrgentBurstFactor( *UrgentBurstFactorChroma = 0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VRatio = %f\n", __func__, VRatio); - dml2_printf("DML::%s: VRatioC = %f\n", __func__, VRatioC); - dml2_printf("DML::%s: DETBufferSizeY = %d\n", __func__, DETBufferSizeY); - dml2_printf("DML::%s: DETBufferSizeC = %d\n", __func__, DETBufferSizeC); - dml2_printf("DML::%s: BytePerPixelInDETY = %f\n", __func__, BytePerPixelInDETY); - dml2_printf("DML::%s: swath_width_luma_ub = %d\n", __func__, swath_width_luma_ub); - dml2_printf("DML::%s: LineTime = %f\n", __func__, LineTime); + DML_LOG_VERBOSE("DML::%s: VRatio = %f\n", __func__, VRatio); + DML_LOG_VERBOSE("DML::%s: VRatioC = %f\n", __func__, VRatioC); + DML_LOG_VERBOSE("DML::%s: DETBufferSizeY = %d\n", __func__, DETBufferSizeY); + DML_LOG_VERBOSE("DML::%s: DETBufferSizeC = %d\n", __func__, DETBufferSizeC); + DML_LOG_VERBOSE("DML::%s: BytePerPixelInDETY = %f\n", __func__, BytePerPixelInDETY); + DML_LOG_VERBOSE("DML::%s: swath_width_luma_ub = %d\n", __func__, swath_width_luma_ub); + DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, LineTime); #endif - DML2_ASSERT(VRatio > 0); + DML_ASSERT(VRatio > 0); LinesInDETLuma = (dml_is_phantom_pipe(plane_cfg) ? 1024 * 1024 : DETBufferSizeY) / BytePerPixelInDETY / swath_width_luma_ub; @@ -3534,12 +3470,12 @@ static void CalculateUrgentBurstFactor( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: LinesInDETLuma = %f\n", __func__, LinesInDETLuma); - dml2_printf("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency); - dml2_printf("DML::%s: DETBufferSizeInTimeLuma = %f\n", __func__, DETBufferSizeInTimeLuma); - dml2_printf("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, *UrgentBurstFactorLuma); - dml2_printf("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, *UrgentBurstFactorChroma); - dml2_printf("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding); + DML_LOG_VERBOSE("DML::%s: LinesInDETLuma = %f\n", __func__, LinesInDETLuma); + DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency); + DML_LOG_VERBOSE("DML::%s: DETBufferSizeInTimeLuma = %f\n", __func__, DETBufferSizeInTimeLuma); + DML_LOG_VERBOSE("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, *UrgentBurstFactorLuma); + DML_LOG_VERBOSE("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, *UrgentBurstFactorChroma); + DML_LOG_VERBOSE("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding); #endif } @@ -3600,10 +3536,10 @@ static void CalculateDCFCLKDeepSleepTdlut( if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut && tdlut_bytes_to_deliver[k] > 0) { double tdlut_required_deepsleep_dcfclk = (double) tdlut_bytes_to_deliver[k] / 64.0 / prefetch_swath_time_us[k]; - dml2_printf("DML::%s: k=%d, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]); - dml2_printf("DML::%s: k=%d, tdlut_bytes_to_deliver = %d\n", __func__, k, tdlut_bytes_to_deliver[k]); - dml2_printf("DML::%s: k=%d, prefetch_swath_time_us = %f\n", __func__, k, prefetch_swath_time_us[k]); - dml2_printf("DML::%s: k=%d, tdlut_required_deepsleep_dcfclk = %f\n", __func__, k, tdlut_required_deepsleep_dcfclk); + DML_LOG_VERBOSE("DML::%s: k=%d, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, tdlut_bytes_to_deliver = %d\n", __func__, k, tdlut_bytes_to_deliver[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, prefetch_swath_time_us = %f\n", __func__, k, prefetch_swath_time_us[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, tdlut_required_deepsleep_dcfclk = %f\n", __func__, k, tdlut_required_deepsleep_dcfclk); // increase the deepsleep dcfclk to match the original dispclk throughput rate if (tdlut_required_deepsleep_dcfclk > DCFClkDeepSleepPerSurface[k]) { @@ -3613,8 +3549,8 @@ static void CalculateDCFCLKDeepSleepTdlut( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, PixelClock = %f\n", __func__, k, pixel_rate_mhz); - dml2_printf("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelClock = %f\n", __func__, k, pixel_rate_mhz); + DML_LOG_VERBOSE("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]); #endif } @@ -3625,17 +3561,17 @@ static void CalculateDCFCLKDeepSleepTdlut( *DCFClkDeepSleep = math_max2(8.0, __DML2_CALCS_DCFCLK_FACTOR__ * ReadBandwidth / (double)ReturnBusWidth); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: __DML2_CALCS_DCFCLK_FACTOR__ = %f\n", __func__, __DML2_CALCS_DCFCLK_FACTOR__); - dml2_printf("DML::%s: ReadBandwidth = %f\n", __func__, ReadBandwidth); - dml2_printf("DML::%s: ReturnBusWidth = %u\n", __func__, ReturnBusWidth); - dml2_printf("DML::%s: DCFClkDeepSleep = %f\n", __func__, *DCFClkDeepSleep); + DML_LOG_VERBOSE("DML::%s: __DML2_CALCS_DCFCLK_FACTOR__ = %f\n", __func__, __DML2_CALCS_DCFCLK_FACTOR__); + DML_LOG_VERBOSE("DML::%s: ReadBandwidth = %f\n", __func__, ReadBandwidth); + DML_LOG_VERBOSE("DML::%s: ReturnBusWidth = %u\n", __func__, ReturnBusWidth); + DML_LOG_VERBOSE("DML::%s: DCFClkDeepSleep = %f\n", __func__, *DCFClkDeepSleep); #endif for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { *DCFClkDeepSleep = math_max2(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]); } - dml2_printf("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep); + DML_LOG_VERBOSE("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep); } static noinline_for_stack void CalculateDCFCLKDeepSleep( @@ -3731,12 +3667,12 @@ static unsigned int CalculateMaxVStartup( else max_vstartup_lines = vblank_size - (unsigned int)math_max2(1.0, math_ceil2(write_back_delay_us / line_time_us, 1.0)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VBlankNom = %u\n", __func__, timing->vblank_nom); - dml2_printf("DML::%s: vblank_nom_default_us = %u\n", __func__, vblank_nom_default_us); - dml2_printf("DML::%s: line_time_us = %f\n", __func__, line_time_us); - dml2_printf("DML::%s: vblank_actual = %u\n", __func__, vblank_actual); - dml2_printf("DML::%s: vblank_avail = %u\n", __func__, vblank_avail); - dml2_printf("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines); + DML_LOG_VERBOSE("DML::%s: VBlankNom = %lu\n", __func__, timing->vblank_nom); + DML_LOG_VERBOSE("DML::%s: vblank_nom_default_us = %u\n", __func__, vblank_nom_default_us); + DML_LOG_VERBOSE("DML::%s: line_time_us = %f\n", __func__, line_time_us); + DML_LOG_VERBOSE("DML::%s: vblank_actual = %u\n", __func__, vblank_actual); + DML_LOG_VERBOSE("DML::%s: vblank_avail = %u\n", __func__, vblank_avail); + DML_LOG_VERBOSE("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines); #endif max_vstartup_lines = (unsigned int)math_min2(max_vstartup_lines, DML_MAX_VSTARTUP_START); return max_vstartup_lines; @@ -3761,9 +3697,9 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch const long MAXIMUMCOMPRESSION = 4; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: ForceSingleDPP = %u\n", __func__, p->ForceSingleDPP); + DML_LOG_VERBOSE("DML::%s: ForceSingleDPP = %u\n", __func__, p->ForceSingleDPP); for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) { - dml2_printf("DML::%s: DPPPerSurface[%u] = %u\n", __func__, k, p->DPPPerSurface[k]); + DML_LOG_VERBOSE("DML::%s: DPPPerSurface[%u] = %u\n", __func__, k, p->DPPPerSurface[k]); } #endif CalculateSwathWidth( @@ -3797,15 +3733,15 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch p->full_swath_bytes_l[k] = (unsigned int)(p->swath_width_luma_ub[k] * p->BytePerPixDETY[k] * MaximumSwathHeightY[k]); p->full_swath_bytes_c[k] = (unsigned int)(p->swath_width_chroma_ub[k] * p->BytePerPixDETC[k] * MaximumSwathHeightC[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, p->DPPPerSurface[k]); - dml2_printf("DML::%s: k=%u swath_width_luma_ub = %u\n", __func__, k, p->swath_width_luma_ub[k]); - dml2_printf("DML::%s: k=%u BytePerPixDETY = %f\n", __func__, k, p->BytePerPixDETY[k]); - dml2_printf("DML::%s: k=%u MaximumSwathHeightY = %u\n", __func__, k, MaximumSwathHeightY[k]); - dml2_printf("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, p->full_swath_bytes_l[k]); - dml2_printf("DML::%s: k=%u swath_width_chroma_ub = %u\n", __func__, k, p->swath_width_chroma_ub[k]); - dml2_printf("DML::%s: k=%u BytePerPixDETC = %f\n", __func__, k, p->BytePerPixDETC[k]); - dml2_printf("DML::%s: k=%u MaximumSwathHeightC = %u\n", __func__, k, MaximumSwathHeightC[k]); - dml2_printf("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, p->full_swath_bytes_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, p->DPPPerSurface[k]); + DML_LOG_VERBOSE("DML::%s: k=%u swath_width_luma_ub = %u\n", __func__, k, p->swath_width_luma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u BytePerPixDETY = %f\n", __func__, k, p->BytePerPixDETY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightY = %u\n", __func__, k, MaximumSwathHeightY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, p->full_swath_bytes_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u swath_width_chroma_ub = %u\n", __func__, k, p->swath_width_chroma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u BytePerPixDETC = %f\n", __func__, k, p->BytePerPixDETC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightC = %u\n", __func__, k, MaximumSwathHeightC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, p->full_swath_bytes_c[k]); #endif if (p->display_cfg->plane_descriptors[k].pixel_format == dml2_420_10) { p->full_swath_bytes_l[k] = (unsigned int)(math_ceil2((double)p->full_swath_bytes_l[k], 256)); @@ -3848,11 +3784,11 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch p->CompressedBufferSizeInkByte); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: TotalActiveDPP = %u\n", __func__, TotalActiveDPP); - dml2_printf("DML::%s: nomDETInKByte = %u\n", __func__, p->nomDETInKByte); - dml2_printf("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, p->ConfigReturnBufferSizeInKByte); - dml2_printf("DML::%s: UnboundedRequestEnabled = %u\n", __func__, *p->UnboundedRequestEnabled); - dml2_printf("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *p->CompressedBufferSizeInkByte); + DML_LOG_VERBOSE("DML::%s: TotalActiveDPP = %u\n", __func__, TotalActiveDPP); + DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, p->nomDETInKByte); + DML_LOG_VERBOSE("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, p->ConfigReturnBufferSizeInKByte); + DML_LOG_VERBOSE("DML::%s: UnboundedRequestEnabled = %u\n", __func__, *p->UnboundedRequestEnabled); + DML_LOG_VERBOSE("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *p->CompressedBufferSizeInkByte); #endif *p->ViewportSizeSupport = true; @@ -3860,7 +3796,7 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch DETBufferSizeInKByteForSwathCalculation = (dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k]) ? 1024 : p->DETBufferSizeInKByte[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation = %u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation = %u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation); #endif if (p->display_cfg->plane_descriptors[k].surface.tiling == dml2_sw_linear) { p->SwathHeightY[k] = MaximumSwathHeightY[k]; @@ -3917,13 +3853,13 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch if ((p->full_swath_bytes_l[k] / 2 + p->full_swath_bytes_c[k] / 2 > DETBufferSizeInKByteForSwathCalculation * 1024 / 2) || p->SwathWidth[k] > p->MaximumSwathWidthLuma[k] || (p->SwathHeightC[k] > 0 && p->SwathWidthChroma[k] > p->MaximumSwathWidthChroma[k])) { *p->ViewportSizeSupport = false; - dml2_printf("DML::%s: k=%u full_swath_bytes_l=%u\n", __func__, k, p->full_swath_bytes_l[k]); - dml2_printf("DML::%s: k=%u full_swath_bytes_c=%u\n", __func__, k, p->full_swath_bytes_c[k]); - dml2_printf("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation=%u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation); - dml2_printf("DML::%s: k=%u SwathWidth=%u\n", __func__, k, p->SwathWidth[k]); - dml2_printf("DML::%s: k=%u MaximumSwathWidthLuma=%f\n", __func__, k, p->MaximumSwathWidthLuma[k]); - dml2_printf("DML::%s: k=%u SwathWidthChroma=%d\n", __func__, k, p->SwathWidthChroma[k]); - dml2_printf("DML::%s: k=%u MaximumSwathWidthChroma=%f\n", __func__, k, p->MaximumSwathWidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l=%u\n", __func__, k, p->full_swath_bytes_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c=%u\n", __func__, k, p->full_swath_bytes_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation=%u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation); + DML_LOG_VERBOSE("DML::%s: k=%u SwathWidth=%u\n", __func__, k, p->SwathWidth[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthLuma=%f\n", __func__, k, p->MaximumSwathWidthLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u SwathWidthChroma=%d\n", __func__, k, p->SwathWidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthChroma=%f\n", __func__, k, p->MaximumSwathWidthChroma[k]); p->ViewportSizeSupportPerSurface[k] = false; } else { p->ViewportSizeSupportPerSurface[k] = true; @@ -3931,35 +3867,35 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch if (p->SwathHeightC[k] == 0) { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, All DET will be used for plane0\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%u, All DET will be used for plane0\n", __func__, k); #endif p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024; p->DETBufferSizeC[k] = 0; } else if (RoundedUpSwathSizeBytesY[k] <= 1.5 * RoundedUpSwathSizeBytesC[k]) { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, Half DET will be used for plane0, and half for plane1\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%u, Half DET will be used for plane0, and half for plane1\n", __func__, k); #endif p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024 / 2; p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 / 2; } else { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, 2/3 DET will be used for plane0, and 1/3 for plane1\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%u, 2/3 DET will be used for plane0, and 1/3 for plane1\n", __func__, k); #endif p->DETBufferSizeY[k] = (unsigned int)(math_floor2(p->DETBufferSizeInKByte[k] * 1024 * 2 / 3, 1024)); p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 - p->DETBufferSizeY[k]; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]); - dml2_printf("DML::%s: k=%u SwathHeightC = %u\n", __func__, k, p->SwathHeightC[k]); - dml2_printf("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, p->full_swath_bytes_l[k]); - dml2_printf("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, p->full_swath_bytes_c[k]); - dml2_printf("DML::%s: k=%u RoundedUpSwathSizeBytesY = %u\n", __func__, k, RoundedUpSwathSizeBytesY[k]); - dml2_printf("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]); - dml2_printf("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, p->DETBufferSizeInKByte[k]); - dml2_printf("DML::%s: k=%u DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]); - dml2_printf("DML::%s: k=%u DETBufferSizeC = %u\n", __func__, k, p->DETBufferSizeC[k]); - dml2_printf("DML::%s: k=%u ViewportSizeSupportPerSurface = %u\n", __func__, k, p->ViewportSizeSupportPerSurface[k]); + DML_LOG_VERBOSE("DML::%s: k=%u SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u SwathHeightC = %u\n", __func__, k, p->SwathHeightC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, p->full_swath_bytes_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, p->full_swath_bytes_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u RoundedUpSwathSizeBytesY = %u\n", __func__, k, RoundedUpSwathSizeBytesY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, p->DETBufferSizeInKByte[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeC = %u\n", __func__, k, p->DETBufferSizeC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u ViewportSizeSupportPerSurface = %u\n", __func__, k, p->ViewportSizeSupportPerSurface[k]); #endif } @@ -3969,12 +3905,12 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch *p->compbuf_reserved_space_64b = (unsigned int)math_ceil2(math_max2(*p->compbuf_reserved_space_64b, (double)(p->rob_buffer_size_kbytes * 1024 / 64) - (double)(RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest] * TTUFIFODEPTH / (p->mrq_present ? MAXIMUMCOMPRESSION : 1) / 64)), 1.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: RoundedUpSwathSizeBytesY[%d] = %u\n", __func__, SurfaceDoingUnboundedRequest, RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest]); - dml2_printf("DML::%s: rob_buffer_size_kbytes = %u\n", __func__, p->rob_buffer_size_kbytes); + DML_LOG_VERBOSE("DML::%s: RoundedUpSwathSizeBytesY[%d] = %u\n", __func__, SurfaceDoingUnboundedRequest, RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest]); + DML_LOG_VERBOSE("DML::%s: rob_buffer_size_kbytes = %u\n", __func__, p->rob_buffer_size_kbytes); #endif } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: compbuf_reserved_space_64b = %u\n", __func__, *p->compbuf_reserved_space_64b); + DML_LOG_VERBOSE("DML::%s: compbuf_reserved_space_64b = %u\n", __func__, *p->compbuf_reserved_space_64b); #endif *p->hw_debug5 = false; @@ -3989,12 +3925,12 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch + *p->CompressedBufferSizeInkByte * MAXIMUMCOMPRESSION * 1024) > TTUFIFODEPTH * (RoundedUpSwathSizeBytesY[k] + RoundedUpSwathSizeBytesC[k]))) *p->hw_debug5 = true; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u UnboundedRequestEnabled = %u\n", __func__, k, *p->UnboundedRequestEnabled); - dml2_printf("DML::%s: k=%u MAXIMUMCOMPRESSION = %lu\n", __func__, k, MAXIMUMCOMPRESSION); - dml2_printf("DML::%s: k=%u TTUFIFODEPTH = %lu\n", __func__, k, TTUFIFODEPTH); - dml2_printf("DML::%s: k=%u CompressedBufferSizeInkByte = %u\n", __func__, k, *p->CompressedBufferSizeInkByte); - dml2_printf("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]); - dml2_printf("DML::%s: k=%u hw_debug5 = %u\n", __func__, k, *p->hw_debug5); + DML_LOG_VERBOSE("DML::%s: k=%u UnboundedRequestEnabled = %u\n", __func__, k, *p->UnboundedRequestEnabled); + DML_LOG_VERBOSE("DML::%s: k=%u MAXIMUMCOMPRESSION = %lu\n", __func__, k, MAXIMUMCOMPRESSION); + DML_LOG_VERBOSE("DML::%s: k=%u TTUFIFODEPTH = %lu\n", __func__, k, TTUFIFODEPTH); + DML_LOG_VERBOSE("DML::%s: k=%u CompressedBufferSizeInkByte = %u\n", __func__, k, *p->CompressedBufferSizeInkByte); + DML_LOG_VERBOSE("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u hw_debug5 = %u\n", __func__, k, *p->hw_debug5); #endif } #endif @@ -4192,15 +4128,15 @@ static noinline_for_stack void CalculateODMMode( SurfaceRequiredDISPCLKWithODMCombineThreeToOne = CalculateRequiredDispclk(dml2_odm_mode_combine_3to1, PixelClock); SurfaceRequiredDISPCLKWithODMCombineFourToOne = CalculateRequiredDispclk(dml2_odm_mode_combine_4to1, PixelClock); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: ODMUse = %d\n", __func__, ODMUse); - dml2_printf("DML::%s: Output = %d\n", __func__, Output); - dml2_printf("DML::%s: DSCEnable = %d\n", __func__, DSCEnable); - dml2_printf("DML::%s: MaxDispclk = %f\n", __func__, MaxDispclk); - dml2_printf("DML::%s: MaximumPixelsPerLinePerDSCUnit = %d\n", __func__, MaximumPixelsPerLinePerDSCUnit); - dml2_printf("DML::%s: SurfaceRequiredDISPCLKWithoutODMCombine = %f\n", __func__, SurfaceRequiredDISPCLKWithoutODMCombine); - dml2_printf("DML::%s: SurfaceRequiredDISPCLKWithODMCombineTwoToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineTwoToOne); - dml2_printf("DML::%s: SurfaceRequiredDISPCLKWithODMCombineThreeToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineThreeToOne); - dml2_printf("DML::%s: SurfaceRequiredDISPCLKWithODMCombineFourToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineFourToOne); + DML_LOG_VERBOSE("DML::%s: ODMUse = %d\n", __func__, ODMUse); + DML_LOG_VERBOSE("DML::%s: Output = %d\n", __func__, Output); + DML_LOG_VERBOSE("DML::%s: DSCEnable = %d\n", __func__, DSCEnable); + DML_LOG_VERBOSE("DML::%s: MaxDispclk = %f\n", __func__, MaxDispclk); + DML_LOG_VERBOSE("DML::%s: MaximumPixelsPerLinePerDSCUnit = %d\n", __func__, MaximumPixelsPerLinePerDSCUnit); + DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithoutODMCombine = %f\n", __func__, SurfaceRequiredDISPCLKWithoutODMCombine); + DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithODMCombineTwoToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineTwoToOne); + DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithODMCombineThreeToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineThreeToOne); + DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithODMCombineFourToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineFourToOne); #endif if (ODMUse == dml2_odm_mode_auto) DecidedODMMode = DecideODMMode(HActive, @@ -4245,10 +4181,10 @@ static noinline_for_stack void CalculateODMMode( *NumberOfDPP = NumberOfDPPRequired; *RequiredDISPCLKPerSurface = success ? DISPCLKRequired : 0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: ODMMode = %d\n", __func__, *ODMMode); - dml2_printf("DML::%s: NumberOfDPP = %d\n", __func__, *NumberOfDPP); - dml2_printf("DML::%s: TotalAvailablePipesSupport = %d\n", __func__, *TotalAvailablePipesSupport); - dml2_printf("DML::%s: RequiredDISPCLKPerSurface = %f\n", __func__, *RequiredDISPCLKPerSurface); + DML_LOG_VERBOSE("DML::%s: ODMMode = %d\n", __func__, *ODMMode); + DML_LOG_VERBOSE("DML::%s: NumberOfDPP = %d\n", __func__, *NumberOfDPP); + DML_LOG_VERBOSE("DML::%s: TotalAvailablePipesSupport = %d\n", __func__, *TotalAvailablePipesSupport); + DML_LOG_VERBOSE("DML::%s: RequiredDISPCLKPerSurface = %f\n", __func__, *RequiredDISPCLKPerSurface); #endif } @@ -4292,17 +4228,17 @@ static noinline_for_stack void CalculateOutputLink( *OutputRate = dml2_core_internal_output_rate_unknown; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DSCEnable = %u (dis, en, en_if_necessary)\n", __func__, DSCEnable); - dml2_printf("DML::%s: PHYCLK = %f\n", __func__, PHYCLK); - dml2_printf("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd); - dml2_printf("DML::%s: AudioSampleRate = %f\n", __func__, AudioSampleRate); - dml2_printf("DML::%s: HActive = %u\n", __func__, HActive); - dml2_printf("DML::%s: HTotal = %u\n", __func__, HTotal); - dml2_printf("DML::%s: ODMModeNoDSC = %u\n", __func__, ODMModeNoDSC); - dml2_printf("DML::%s: ODMModeDSC = %u\n", __func__, ODMModeDSC); - dml2_printf("DML::%s: ForcedOutputLinkBPP = %f\n", __func__, ForcedOutputLinkBPP); - dml2_printf("DML::%s: Output (encoder) = %u\n", __func__, Output); - dml2_printf("DML::%s: OutputLinkDPRate = %u\n", __func__, OutputLinkDPRate); + DML_LOG_VERBOSE("DML::%s: DSCEnable = %u (dis, en, en_if_necessary)\n", __func__, DSCEnable); + DML_LOG_VERBOSE("DML::%s: PHYCLK = %f\n", __func__, PHYCLK); + DML_LOG_VERBOSE("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd); + DML_LOG_VERBOSE("DML::%s: AudioSampleRate = %f\n", __func__, AudioSampleRate); + DML_LOG_VERBOSE("DML::%s: HActive = %u\n", __func__, HActive); + DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, HTotal); + DML_LOG_VERBOSE("DML::%s: ODMModeNoDSC = %u\n", __func__, ODMModeNoDSC); + DML_LOG_VERBOSE("DML::%s: ODMModeDSC = %u\n", __func__, ODMModeDSC); + DML_LOG_VERBOSE("DML::%s: ForcedOutputLinkBPP = %f\n", __func__, ForcedOutputLinkBPP); + DML_LOG_VERBOSE("DML::%s: Output (encoder) = %u\n", __func__, Output); + DML_LOG_VERBOSE("DML::%s: OutputLinkDPRate = %u\n", __func__, OutputLinkDPRate); #endif { if (Output == dml2_hdmi) { @@ -4487,9 +4423,9 @@ static noinline_for_stack void CalculateOutputLink( } } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: RequiresDSC = %u\n", __func__, *RequiresDSC); - dml2_printf("DML::%s: RequiresFEC = %u\n", __func__, *RequiresFEC); - dml2_printf("DML::%s: OutBpp = %f\n", __func__, *OutBpp); + DML_LOG_VERBOSE("DML::%s: RequiresDSC = %u\n", __func__, *RequiresDSC); + DML_LOG_VERBOSE("DML::%s: RequiresFEC = %u\n", __func__, *RequiresFEC); + DML_LOG_VERBOSE("DML::%s: OutBpp = %f\n", __func__, *OutBpp); #endif } @@ -4571,17 +4507,17 @@ static unsigned int DSCDelayRequirement( DSCDelayRequirement_val = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DSCEnabled= %u\n", __func__, DSCEnabled); - dml2_printf("DML::%s: ODMMode = %u\n", __func__, ODMMode); - dml2_printf("DML::%s: OutputBpp = %f\n", __func__, OutputBpp); - dml2_printf("DML::%s: HActive = %u\n", __func__, HActive); - dml2_printf("DML::%s: HTotal = %u\n", __func__, HTotal); - dml2_printf("DML::%s: PixelClock = %f\n", __func__, PixelClock); - dml2_printf("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd); - dml2_printf("DML::%s: OutputFormat = %u\n", __func__, OutputFormat); - dml2_printf("DML::%s: DSCInputBitPerComponent = %u\n", __func__, DSCInputBitPerComponent); - dml2_printf("DML::%s: NumberOfDSCSlices = %u\n", __func__, NumberOfDSCSlices); - dml2_printf("DML::%s: DSCDelayRequirement_val = %u\n", __func__, DSCDelayRequirement_val); + DML_LOG_VERBOSE("DML::%s: DSCEnabled= %u\n", __func__, DSCEnabled); + DML_LOG_VERBOSE("DML::%s: ODMMode = %u\n", __func__, ODMMode); + DML_LOG_VERBOSE("DML::%s: OutputBpp = %f\n", __func__, OutputBpp); + DML_LOG_VERBOSE("DML::%s: HActive = %u\n", __func__, HActive); + DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, HTotal); + DML_LOG_VERBOSE("DML::%s: PixelClock = %f\n", __func__, PixelClock); + DML_LOG_VERBOSE("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd); + DML_LOG_VERBOSE("DML::%s: OutputFormat = %u\n", __func__, OutputFormat); + DML_LOG_VERBOSE("DML::%s: DSCInputBitPerComponent = %u\n", __func__, DSCInputBitPerComponent); + DML_LOG_VERBOSE("DML::%s: NumberOfDSCSlices = %u\n", __func__, NumberOfDSCSlices); + DML_LOG_VERBOSE("DML::%s: DSCDelayRequirement_val = %u\n", __func__, DSCDelayRequirement_val); #endif return DSCDelayRequirement_val; @@ -4654,10 +4590,10 @@ static void CalculateSurfaceSizeInMall( (TotalSurfaceSizeInMALLForSubVP > MALLAllocatedForDCNInBytes); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: MALLAllocatedForDCN = %u\n", __func__, MALLAllocatedForDCN * 1024 * 1024); - dml2_printf("DML::%s: TotalSurfaceSizeInMALLForSubVP = %u\n", __func__, TotalSurfaceSizeInMALLForSubVP); - dml2_printf("DML::%s: TotalSurfaceSizeInMALLForSS = %u\n", __func__, TotalSurfaceSizeInMALLForSS); - dml2_printf("DML::%s: ExceededMALLSize = %u\n", __func__, *ExceededMALLSize); + DML_LOG_VERBOSE("DML::%s: MALLAllocatedForDCN = %u\n", __func__, MALLAllocatedForDCN * 1024 * 1024); + DML_LOG_VERBOSE("DML::%s: TotalSurfaceSizeInMALLForSubVP = %u\n", __func__, TotalSurfaceSizeInMALLForSubVP); + DML_LOG_VERBOSE("DML::%s: TotalSurfaceSizeInMALLForSS = %u\n", __func__, TotalSurfaceSizeInMALLForSS); + DML_LOG_VERBOSE("DML::%s: ExceededMALLSize = %u\n", __func__, *ExceededMALLSize); #endif } @@ -4674,7 +4610,6 @@ static void calculate_tdlut_setting( unsigned int tdlut_vmpg_per_frame; unsigned int tdlut_pte_req_per_frame; unsigned int tdlut_bytes_per_line; - unsigned int tdlut_delivery_cycles; double tdlut_drain_rate; unsigned int tdlut_mpc_width; unsigned int tdlut_bytes_per_group_simple; @@ -4737,13 +4672,13 @@ static void calculate_tdlut_setting( *p->tdlut_bytes_per_frame = tdlut_bytes_per_line * tdlut_mpc_width * tdlut_mpc_width; *p->tdlut_bytes_per_group = tdlut_bytes_per_line * tdlut_mpc_width; //the delivery cycles is DispClk cycles per line * number of lines * number of slices - tdlut_delivery_cycles = (unsigned int)math_ceil2(tdlut_mpc_width/2.0, 1) * tdlut_mpc_width * tdlut_mpc_width; + //tdlut_delivery_cycles = (unsigned int)math_ceil2(tdlut_mpc_width/2.0, 1) * tdlut_mpc_width * tdlut_mpc_width; tdlut_drain_rate = tdlut_bytes_per_line * p->dispclk_mhz / math_ceil2(tdlut_mpc_width/2.0, 1); } else { //tdlut_addressing_mode = tdlut_simple_linear, 3dlut width should be 4*1229=4916 elements *p->tdlut_bytes_per_frame = (unsigned int)math_ceil2(tdlut_width * tdlut_bpe, 256); *p->tdlut_bytes_per_group = tdlut_bytes_per_group_simple; - tdlut_delivery_cycles = (unsigned int)math_ceil2(tdlut_width/2.0, 1); + //tdlut_delivery_cycles = (unsigned int)math_ceil2(tdlut_width/2.0, 1); tdlut_drain_rate = 2 * tdlut_bpe * p->dispclk_mhz; } @@ -4756,25 +4691,25 @@ static void calculate_tdlut_setting( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: gpuvm_enable = %d\n", __func__, p->gpuvm_enable); - dml2_printf("DML::%s: vmpg_bytes = %d\n", __func__, vmpg_bytes); - dml2_printf("DML::%s: tdlut_vmpg_per_frame = %d\n", __func__, tdlut_vmpg_per_frame); - dml2_printf("DML::%s: tdlut_pte_req_per_frame = %d\n", __func__, tdlut_pte_req_per_frame); + DML_LOG_VERBOSE("DML::%s: gpuvm_enable = %d\n", __func__, p->gpuvm_enable); + DML_LOG_VERBOSE("DML::%s: vmpg_bytes = %d\n", __func__, vmpg_bytes); + DML_LOG_VERBOSE("DML::%s: tdlut_vmpg_per_frame = %d\n", __func__, tdlut_vmpg_per_frame); + DML_LOG_VERBOSE("DML::%s: tdlut_pte_req_per_frame = %d\n", __func__, tdlut_pte_req_per_frame); - dml2_printf("DML::%s: dispclk_mhz = %f\n", __func__, p->dispclk_mhz); - dml2_printf("DML::%s: tdlut_width = %u\n", __func__, tdlut_width); - dml2_printf("DML::%s: tdlut_addressing_mode = %s\n", __func__, (p->tdlut_addressing_mode == dml2_tdlut_sw_linear) ? "sw_linear" : "simple_linear"); - dml2_printf("DML::%s: tdlut_pitch_bytes = %u\n", __func__, tdlut_pitch_bytes); - dml2_printf("DML::%s: tdlut_footprint_bytes = %u\n", __func__, tdlut_footprint_bytes); - dml2_printf("DML::%s: tdlut_bytes_per_frame = %u\n", __func__, *p->tdlut_bytes_per_frame); - dml2_printf("DML::%s: tdlut_bytes_per_line = %u\n", __func__, tdlut_bytes_per_line); - dml2_printf("DML::%s: tdlut_bytes_per_group = %u\n", __func__, *p->tdlut_bytes_per_group); - dml2_printf("DML::%s: tdlut_drain_rate = %f\n", __func__, tdlut_drain_rate); - dml2_printf("DML::%s: tdlut_delivery_cycles = %u\n", __func__, tdlut_delivery_cycles); - dml2_printf("DML::%s: tdlut_opt_time = %f\n", __func__, *p->tdlut_opt_time); - dml2_printf("DML::%s: tdlut_drain_time = %f\n", __func__, *p->tdlut_drain_time); - dml2_printf("DML::%s: tdlut_bytes_to_deliver = %d\n", __func__, *p->tdlut_bytes_to_deliver); - dml2_printf("DML::%s: tdlut_groups_per_2row_ub = %d\n", __func__, *p->tdlut_groups_per_2row_ub); + DML_LOG_VERBOSE("DML::%s: dispclk_mhz = %f\n", __func__, p->dispclk_mhz); + DML_LOG_VERBOSE("DML::%s: tdlut_width = %u\n", __func__, tdlut_width); + DML_LOG_VERBOSE("DML::%s: tdlut_addressing_mode = %s\n", __func__, (p->tdlut_addressing_mode == dml2_tdlut_sw_linear) ? "sw_linear" : "simple_linear"); + DML_LOG_VERBOSE("DML::%s: tdlut_pitch_bytes = %u\n", __func__, tdlut_pitch_bytes); + DML_LOG_VERBOSE("DML::%s: tdlut_footprint_bytes = %u\n", __func__, tdlut_footprint_bytes); + DML_LOG_VERBOSE("DML::%s: tdlut_bytes_per_frame = %u\n", __func__, *p->tdlut_bytes_per_frame); + DML_LOG_VERBOSE("DML::%s: tdlut_bytes_per_line = %u\n", __func__, tdlut_bytes_per_line); + DML_LOG_VERBOSE("DML::%s: tdlut_bytes_per_group = %u\n", __func__, *p->tdlut_bytes_per_group); + DML_LOG_VERBOSE("DML::%s: tdlut_drain_rate = %f\n", __func__, tdlut_drain_rate); + DML_LOG_VERBOSE("DML::%s: tdlut_delivery_cycles = %u\n", __func__, p->tdlut_addressing_mode == dml2_tdlut_sw_linear ? (unsigned int)math_ceil2(tdlut_mpc_width/2.0, 1) * tdlut_mpc_width * tdlut_mpc_width : (unsigned int)math_ceil2(tdlut_width/2.0, 1)); + DML_LOG_VERBOSE("DML::%s: tdlut_opt_time = %f\n", __func__, *p->tdlut_opt_time); + DML_LOG_VERBOSE("DML::%s: tdlut_drain_time = %f\n", __func__, *p->tdlut_drain_time); + DML_LOG_VERBOSE("DML::%s: tdlut_bytes_to_deliver = %d\n", __func__, *p->tdlut_bytes_to_deliver); + DML_LOG_VERBOSE("DML::%s: tdlut_groups_per_2row_ub = %d\n", __func__, *p->tdlut_groups_per_2row_ub); #endif } @@ -4820,10 +4755,10 @@ static void CalculateTarb( *Tarb = extra_bytes / ReturnBW; *Tarb_prefetch = extra_bytes_prefetch / ReturnBW; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: PixelChunkSizeInKByte = %d\n", __func__, PixelChunkSizeInKByte); - dml2_printf("DML::%s: MetaChunkSize = %d\n", __func__, MetaChunkSize); - dml2_printf("DML::%s: extra_bytes = %f\n", __func__, extra_bytes); - dml2_printf("DML::%s: extra_bytes_prefetch = %f\n", __func__, extra_bytes_prefetch); + DML_LOG_VERBOSE("DML::%s: PixelChunkSizeInKByte = %d\n", __func__, PixelChunkSizeInKByte); + DML_LOG_VERBOSE("DML::%s: MetaChunkSize = %d\n", __func__, MetaChunkSize); + DML_LOG_VERBOSE("DML::%s: extra_bytes = %f\n", __func__, extra_bytes); + DML_LOG_VERBOSE("DML::%s: extra_bytes_prefetch = %f\n", __func__, extra_bytes_prefetch); #endif } @@ -4838,10 +4773,10 @@ static double CalculateTWait( TWait = math_max2(reserved_vblank_time_ns/1000.0, g6_temp_read_blackout_us) + t_urg_trip; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: reserved_vblank_time_ns = %d\n", __func__, reserved_vblank_time_ns); - dml2_printf("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency); - dml2_printf("DML::%s: Ttrip = %f\n", __func__, Ttrip); - dml2_printf("DML::%s: TWait = %f\n", __func__, TWait); + DML_LOG_VERBOSE("DML::%s: reserved_vblank_time_ns = %ld\n", __func__, reserved_vblank_time_ns); + DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency); + DML_LOG_VERBOSE("DML::%s: Ttrip = %f\n", __func__, Ttrip); + DML_LOG_VERBOSE("DML::%s: TWait = %f\n", __func__, TWait); #endif return TWait; } @@ -4887,20 +4822,20 @@ static void CalculateVUpdateAndDynamicMetadataParameters( *Tdmsks = *Tdmsks / 2; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DynamicMetadataLinesBeforeActiveRequired = %u\n", __func__, DynamicMetadataLinesBeforeActiveRequired); - dml2_printf("DML::%s: VBlank = %u\n", __func__, VBlank); - dml2_printf("DML::%s: HTotal = %u\n", __func__, HTotal); - dml2_printf("DML::%s: PixelClock = %f\n", __func__, PixelClock); - dml2_printf("DML::%s: Dppclk = %f\n", __func__, Dppclk); - dml2_printf("DML::%s: DCFClkDeepSleep = %f\n", __func__, DCFClkDeepSleep); - dml2_printf("DML::%s: MaxInterDCNTileRepeaters = %u\n", __func__, MaxInterDCNTileRepeaters); - dml2_printf("DML::%s: TotalRepeaterDelayTime = %f\n", __func__, TotalRepeaterDelayTime); + DML_LOG_VERBOSE("DML::%s: DynamicMetadataLinesBeforeActiveRequired = %u\n", __func__, DynamicMetadataLinesBeforeActiveRequired); + DML_LOG_VERBOSE("DML::%s: VBlank = %u\n", __func__, VBlank); + DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, HTotal); + DML_LOG_VERBOSE("DML::%s: PixelClock = %f\n", __func__, PixelClock); + DML_LOG_VERBOSE("DML::%s: Dppclk = %f\n", __func__, Dppclk); + DML_LOG_VERBOSE("DML::%s: DCFClkDeepSleep = %f\n", __func__, DCFClkDeepSleep); + DML_LOG_VERBOSE("DML::%s: MaxInterDCNTileRepeaters = %u\n", __func__, MaxInterDCNTileRepeaters); + DML_LOG_VERBOSE("DML::%s: TotalRepeaterDelayTime = %f\n", __func__, TotalRepeaterDelayTime); - dml2_printf("DML::%s: VUpdateWidthPix = %u\n", __func__, *VUpdateWidthPix); - dml2_printf("DML::%s: VReadyOffsetPix = %u\n", __func__, *VReadyOffsetPix); - dml2_printf("DML::%s: VUpdateOffsetPix = %u\n", __func__, *VUpdateOffsetPix); + DML_LOG_VERBOSE("DML::%s: VUpdateWidthPix = %u\n", __func__, *VUpdateWidthPix); + DML_LOG_VERBOSE("DML::%s: VReadyOffsetPix = %u\n", __func__, *VReadyOffsetPix); + DML_LOG_VERBOSE("DML::%s: VUpdateOffsetPix = %u\n", __func__, *VUpdateOffsetPix); - dml2_printf("DML::%s: Tdmsks = %f\n", __func__, *Tdmsks); + DML_LOG_VERBOSE("DML::%s: Tdmsks = %f\n", __func__, *Tdmsks); #endif } @@ -4962,11 +4897,11 @@ static double get_urgent_bandwidth_required( l->adj_factor_cur_pre = UrgentBurstFactorCursorPre[k]; bool is_phantom = dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]); - bool exclude_this_plane = 0; + bool exclude_this_plane = false; // Exclude phantom pipe in bw calculation for non svp prefetch state if (state_type != dml2_core_internal_soc_state_svp_prefetch && is_phantom) - exclude_this_plane = 1; + exclude_this_plane = true; // The qualified row bandwidth, qual_row_bw, accounts for the regular non-flip row bandwidth when there is no possible immediate flip or HostVM invalidation flip. // The qual_row_bw is zero if HostVM is possible and only non-zero and equal to row_bw(i) if immediate flip is not allowed for that pipe. @@ -4995,12 +4930,12 @@ static double get_urgent_bandwidth_required( surface_peak_required_bw[k] = math_max2(surface_required_bw[k], surface_peak_required_bw[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, max1: vm_row_bw=%f\n", __func__, k, l->vm_row_bw); - dml2_printf("DML::%s: k=%d, max2: flip_and_active_bw=%f\n", __func__, k, l->flip_and_active_bw); - dml2_printf("DML::%s: k=%d, max3: flip_and_prefetch_bw=%f\n", __func__, k, l->flip_and_prefetch_bw); - dml2_printf("DML::%s: k=%d, max4: active_and_excess_bw=%f\n", __func__, k, l->active_and_excess_bw); - dml2_printf("DML::%s: k=%d, surface_required_bw=%f\n", __func__, k, surface_required_bw[k]); - dml2_printf("DML::%s: k=%d, surface_peak_required_bw=%f\n", __func__, k, surface_peak_required_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, max1: vm_row_bw=%f\n", __func__, k, l->vm_row_bw); + DML_LOG_VERBOSE("DML::%s: k=%d, max2: flip_and_active_bw=%f\n", __func__, k, l->flip_and_active_bw); + DML_LOG_VERBOSE("DML::%s: k=%d, max3: flip_and_prefetch_bw=%f\n", __func__, k, l->flip_and_prefetch_bw); + DML_LOG_VERBOSE("DML::%s: k=%d, max4: active_and_excess_bw=%f\n", __func__, k, l->active_and_excess_bw); + DML_LOG_VERBOSE("DML::%s: k=%d, surface_required_bw=%f\n", __func__, k, surface_required_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, surface_peak_required_bw=%f\n", __func__, k, surface_peak_required_bw[k]); #endif } else { surface_required_bw[k] = 0.0; @@ -5009,34 +4944,34 @@ static double get_urgent_bandwidth_required( l->required_bandwidth_mbps += surface_required_bw[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, NumberOfDPP=%d\n", __func__, k, NumberOfDPP[k]); - dml2_printf("DML::%s: k=%d, use_qual_row_bw=%d\n", __func__, k, use_qual_row_bw); - dml2_printf("DML::%s: k=%d, immediate_flip=%d\n", __func__, k, display_cfg->plane_descriptors[k].immediate_flip); - dml2_printf("DML::%s: k=%d, mall_svp_prefetch_factor=%f\n", __func__, k, l->mall_svp_prefetch_factor); - dml2_printf("DML::%s: k=%d, adj_factor_p0=%f\n", __func__, k, l->adj_factor_p0); - dml2_printf("DML::%s: k=%d, adj_factor_p1=%f\n", __func__, k, l->adj_factor_p1); - dml2_printf("DML::%s: k=%d, adj_factor_cur=%f\n", __func__, k, l->adj_factor_cur); + DML_LOG_VERBOSE("DML::%s: k=%d, NumberOfDPP=%d\n", __func__, k, NumberOfDPP[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, use_qual_row_bw=%d\n", __func__, k, use_qual_row_bw); + DML_LOG_VERBOSE("DML::%s: k=%d, immediate_flip=%d\n", __func__, k, display_cfg->plane_descriptors[k].immediate_flip); + DML_LOG_VERBOSE("DML::%s: k=%d, mall_svp_prefetch_factor=%f\n", __func__, k, l->mall_svp_prefetch_factor); + DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p0=%f\n", __func__, k, l->adj_factor_p0); + DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p1=%f\n", __func__, k, l->adj_factor_p1); + DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_cur=%f\n", __func__, k, l->adj_factor_cur); - dml2_printf("DML::%s: k=%d, adj_factor_p0_pre=%f\n", __func__, k, l->adj_factor_p0_pre); - dml2_printf("DML::%s: k=%d, adj_factor_p1_pre=%f\n", __func__, k, l->adj_factor_p1_pre); - dml2_printf("DML::%s: k=%d, adj_factor_cur_pre=%f\n", __func__, k, l->adj_factor_cur_pre); + DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p0_pre=%f\n", __func__, k, l->adj_factor_p0_pre); + DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p1_pre=%f\n", __func__, k, l->adj_factor_p1_pre); + DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_cur_pre=%f\n", __func__, k, l->adj_factor_cur_pre); - dml2_printf("DML::%s: k=%d, per_plane_flip_bw=%f\n", __func__, k, l->per_plane_flip_bw[k]); - dml2_printf("DML::%s: k=%d, prefetch_vmrow_bw=%f\n", __func__, k, prefetch_vmrow_bw[k]); - dml2_printf("DML::%s: k=%d, ReadBandwidthLuma=%f\n", __func__, k, ReadBandwidthLuma[k]); - dml2_printf("DML::%s: k=%d, ReadBandwidthChroma=%f\n", __func__, k, ReadBandwidthChroma[k]); - dml2_printf("DML::%s: k=%d, excess_vactive_fill_bw_l=%f\n", __func__, k, excess_vactive_fill_bw_l[k]); - dml2_printf("DML::%s: k=%d, excess_vactive_fill_bw_c=%f\n", __func__, k, excess_vactive_fill_bw_c[k]); - dml2_printf("DML::%s: k=%d, cursor_bw=%f\n", __func__, k, cursor_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, per_plane_flip_bw=%f\n", __func__, k, l->per_plane_flip_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, prefetch_vmrow_bw=%f\n", __func__, k, prefetch_vmrow_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, ReadBandwidthLuma=%f\n", __func__, k, ReadBandwidthLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, ReadBandwidthChroma=%f\n", __func__, k, ReadBandwidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, excess_vactive_fill_bw_l=%f\n", __func__, k, excess_vactive_fill_bw_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, excess_vactive_fill_bw_c=%f\n", __func__, k, excess_vactive_fill_bw_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, cursor_bw=%f\n", __func__, k, cursor_bw[k]); - dml2_printf("DML::%s: k=%d, meta_row_bw=%f\n", __func__, k, meta_row_bw[k]); - dml2_printf("DML::%s: k=%d, dpte_row_bw=%f\n", __func__, k, dpte_row_bw[k]); - dml2_printf("DML::%s: k=%d, PrefetchBandwidthLuma=%f\n", __func__, k, PrefetchBandwidthLuma[k]); - dml2_printf("DML::%s: k=%d, PrefetchBandwidthChroma=%f\n", __func__, k, PrefetchBandwidthChroma[k]); - dml2_printf("DML::%s: k=%d, prefetch_cursor_bw=%f\n", __func__, k, prefetch_cursor_bw[k]); - dml2_printf("DML::%s: k=%d, required_bandwidth_mbps=%f (total), inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, inc_flip_bw, is_phantom, exclude_this_plane); - dml2_printf("DML::%s: k=%d, required_bandwidth_mbps=%f (total), soc_state=%s, inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, dml2_core_internal_soc_state_type_str(state_type), inc_flip_bw, is_phantom, exclude_this_plane); - dml2_printf("DML::%s: k=%d, required_bandwidth_mbps=%f (total), inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, inc_flip_bw, is_phantom, exclude_this_plane); + DML_LOG_VERBOSE("DML::%s: k=%d, meta_row_bw=%f\n", __func__, k, meta_row_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, dpte_row_bw=%f\n", __func__, k, dpte_row_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, PrefetchBandwidthLuma=%f\n", __func__, k, PrefetchBandwidthLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, PrefetchBandwidthChroma=%f\n", __func__, k, PrefetchBandwidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, prefetch_cursor_bw=%f\n", __func__, k, prefetch_cursor_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, required_bandwidth_mbps=%f (total), inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, inc_flip_bw, is_phantom, exclude_this_plane); + DML_LOG_VERBOSE("DML::%s: k=%d, required_bandwidth_mbps=%f (total), soc_state=%s, inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, dml2_core_internal_soc_state_type_str(state_type), inc_flip_bw, is_phantom, exclude_this_plane); + DML_LOG_VERBOSE("DML::%s: k=%d, required_bandwidth_mbps=%f (total), inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, inc_flip_bw, is_phantom, exclude_this_plane); #endif } @@ -5120,19 +5055,19 @@ static void CalculateExtraLatency( *ExtraLatency_sr = *ExtraLatency_sr + Tarb; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: qos_type=%u\n", __func__, qos_type); - dml2_printf("DML::%s: hostvm_mode=%u\n", __func__, hostvm_mode); - dml2_printf("DML::%s: Tex_trips=%u\n", __func__, Tex_trips); - dml2_printf("DML::%s: max_outstanding_when_urgent_expected=%u\n", __func__, max_outstanding_when_urgent_expected); - dml2_printf("DML::%s: FabricClock=%f\n", __func__, FabricClock); - dml2_printf("DML::%s: DCFCLK=%f\n", __func__, DCFCLK); - dml2_printf("DML::%s: ReturnBW=%f\n", __func__, ReturnBW); - dml2_printf("DML::%s: RoundTripPingLatencyCycles=%u\n", __func__, RoundTripPingLatencyCycles); - dml2_printf("DML::%s: ReorderingBytes=%u\n", __func__, ReorderingBytes); - dml2_printf("DML::%s: Tarb=%f\n", __func__, Tarb); - dml2_printf("DML::%s: ExtraLatency=%f\n", __func__, *ExtraLatency); - dml2_printf("DML::%s: ExtraLatency_sr=%f\n", __func__, *ExtraLatency_sr); - dml2_printf("DML::%s: ExtraLatencyPrefetch=%f\n", __func__, *ExtraLatencyPrefetch); + DML_LOG_VERBOSE("DML::%s: qos_type=%u\n", __func__, qos_type); + DML_LOG_VERBOSE("DML::%s: hostvm_mode=%u\n", __func__, hostvm_mode); + DML_LOG_VERBOSE("DML::%s: Tex_trips=%f\n", __func__, Tex_trips); + DML_LOG_VERBOSE("DML::%s: max_outstanding_when_urgent_expected=%u\n", __func__, max_outstanding_when_urgent_expected); + DML_LOG_VERBOSE("DML::%s: FabricClock=%f\n", __func__, FabricClock); + DML_LOG_VERBOSE("DML::%s: DCFCLK=%f\n", __func__, DCFCLK); + DML_LOG_VERBOSE("DML::%s: ReturnBW=%f\n", __func__, ReturnBW); + DML_LOG_VERBOSE("DML::%s: RoundTripPingLatencyCycles=%u\n", __func__, RoundTripPingLatencyCycles); + DML_LOG_VERBOSE("DML::%s: ReorderingBytes=%u\n", __func__, ReorderingBytes); + DML_LOG_VERBOSE("DML::%s: Tarb=%f\n", __func__, Tarb); + DML_LOG_VERBOSE("DML::%s: ExtraLatency=%f\n", __func__, *ExtraLatency); + DML_LOG_VERBOSE("DML::%s: ExtraLatency_sr=%f\n", __func__, *ExtraLatency_sr); + DML_LOG_VERBOSE("DML::%s: ExtraLatencyPrefetch=%f\n", __func__, *ExtraLatencyPrefetch); #endif } @@ -5199,20 +5134,20 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->HostVMDynamicLevelsTrips = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: dcc_enable = %u\n", __func__, p->dcc_enable); - dml2_printf("DML::%s: mrq_present = %u\n", __func__, p->mrq_present); - dml2_printf("DML::%s: dcc_mrq_enable = %u\n", __func__, dcc_mrq_enable); - dml2_printf("DML::%s: GPUVMEnable = %u\n", __func__, p->display_cfg->gpuvm_enable); - dml2_printf("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->display_cfg->gpuvm_max_page_table_levels); - dml2_printf("DML::%s: DCCEnable = %u\n", __func__, p->myPipe->DCCEnable); - dml2_printf("DML::%s: VStartup = %u\n", __func__, p->VStartup); - dml2_printf("DML::%s: HostVMEnable = %u\n", __func__, p->display_cfg->hostvm_enable); - dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); - dml2_printf("DML::%s: TWait = %f\n", __func__, p->TWait); - dml2_printf("DML::%s: TWait_p = %f\n", __func__, s->TWait_p); - dml2_printf("DML::%s: Ttrip = %f\n", __func__, p->Ttrip); - dml2_printf("DML::%s: myPipe->Dppclk = %f\n", __func__, p->myPipe->Dppclk); - dml2_printf("DML::%s: myPipe->Dispclk = %f\n", __func__, p->myPipe->Dispclk); + DML_LOG_VERBOSE("DML::%s: dcc_enable = %u\n", __func__, p->dcc_enable); + DML_LOG_VERBOSE("DML::%s: mrq_present = %u\n", __func__, p->mrq_present); + DML_LOG_VERBOSE("DML::%s: dcc_mrq_enable = %u\n", __func__, dcc_mrq_enable); + DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, p->display_cfg->gpuvm_enable); + DML_LOG_VERBOSE("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->display_cfg->gpuvm_max_page_table_levels); + DML_LOG_VERBOSE("DML::%s: DCCEnable = %u\n", __func__, p->myPipe->DCCEnable); + DML_LOG_VERBOSE("DML::%s: VStartup = %u\n", __func__, p->VStartup); + DML_LOG_VERBOSE("DML::%s: HostVMEnable = %u\n", __func__, p->display_cfg->hostvm_enable); + DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: TWait = %f\n", __func__, p->TWait); + DML_LOG_VERBOSE("DML::%s: TWait_p = %f\n", __func__, s->TWait_p); + DML_LOG_VERBOSE("DML::%s: Ttrip = %f\n", __func__, p->Ttrip); + DML_LOG_VERBOSE("DML::%s: myPipe->Dppclk = %f\n", __func__, p->myPipe->Dppclk); + DML_LOG_VERBOSE("DML::%s: myPipe->Dispclk = %f\n", __func__, p->myPipe->Dispclk); #endif CalculateVUpdateAndDynamicMetadataParameters( p->MaxInterDCNTileRepeaters, @@ -5258,11 +5193,11 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch if (p->DynamicMetadataEnable == true) { if (p->VStartup * s->LineTime < *p->TSetup + *p->Tdmdl + s->Tdmbf + s->Tdmec + s->Tdmsks) { *p->NotEnoughTimeForDynamicMetadata = true; - dml2_printf("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__); - dml2_printf("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf); - dml2_printf("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec); - dml2_printf("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks); - dml2_printf("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl); + DML_LOG_VERBOSE("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__); + DML_LOG_VERBOSE("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf); + DML_LOG_VERBOSE("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec); + DML_LOG_VERBOSE("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks); + DML_LOG_VERBOSE("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl); } else { *p->NotEnoughTimeForDynamicMetadata = false; } @@ -5288,21 +5223,21 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch ((p->myPipe->ODMMode == dml2_odm_mode_mso_1to4) ? (double)p->myPipe->HActive * 3.0 / 4.0 : 0)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DynamicMetadataVMEnabled = %u\n", __func__, p->DynamicMetadataVMEnabled); - dml2_printf("DML::%s: DPPCycles = %u\n", __func__, s->DPPCycles); - dml2_printf("DML::%s: PixelClock = %f\n", __func__, p->myPipe->PixelClock); - dml2_printf("DML::%s: Dppclk = %f\n", __func__, p->myPipe->Dppclk); - dml2_printf("DML::%s: DISPCLKCycles = %u\n", __func__, s->DISPCLKCycles); - dml2_printf("DML::%s: DISPCLK = %f\n", __func__, p->myPipe->Dispclk); - dml2_printf("DML::%s: DSCDelay = %u\n", __func__, p->DSCDelay); - dml2_printf("DML::%s: ODMMode = %u\n", __func__, p->myPipe->ODMMode); - dml2_printf("DML::%s: DPP_RECOUT_WIDTH = %u\n", __func__, p->DPP_RECOUT_WIDTH); - dml2_printf("DML::%s: DSTXAfterScaler = %u\n", __func__, *p->DSTXAfterScaler); + DML_LOG_VERBOSE("DML::%s: DynamicMetadataVMEnabled = %u\n", __func__, p->DynamicMetadataVMEnabled); + DML_LOG_VERBOSE("DML::%s: DPPCycles = %u\n", __func__, s->DPPCycles); + DML_LOG_VERBOSE("DML::%s: PixelClock = %f\n", __func__, p->myPipe->PixelClock); + DML_LOG_VERBOSE("DML::%s: Dppclk = %f\n", __func__, p->myPipe->Dppclk); + DML_LOG_VERBOSE("DML::%s: DISPCLKCycles = %u\n", __func__, s->DISPCLKCycles); + DML_LOG_VERBOSE("DML::%s: DISPCLK = %f\n", __func__, p->myPipe->Dispclk); + DML_LOG_VERBOSE("DML::%s: DSCDelay = %u\n", __func__, p->DSCDelay); + DML_LOG_VERBOSE("DML::%s: ODMMode = %u\n", __func__, p->myPipe->ODMMode); + DML_LOG_VERBOSE("DML::%s: DPP_RECOUT_WIDTH = %u\n", __func__, p->DPP_RECOUT_WIDTH); + DML_LOG_VERBOSE("DML::%s: DSTXAfterScaler = %u\n", __func__, *p->DSTXAfterScaler); - dml2_printf("DML::%s: setup_for_tdlut = %u\n", __func__, p->setup_for_tdlut); - dml2_printf("DML::%s: tdlut_opt_time = %f\n", __func__, p->tdlut_opt_time); - dml2_printf("DML::%s: tdlut_pte_bytes_per_frame = %u\n", __func__, p->tdlut_pte_bytes_per_frame); - dml2_printf("DML::%s: tdlut_drain_time = %f\n", __func__, p->tdlut_drain_time); + DML_LOG_VERBOSE("DML::%s: setup_for_tdlut = %u\n", __func__, p->setup_for_tdlut); + DML_LOG_VERBOSE("DML::%s: tdlut_opt_time = %f\n", __func__, p->tdlut_opt_time); + DML_LOG_VERBOSE("DML::%s: tdlut_pte_bytes_per_frame = %u\n", __func__, p->tdlut_pte_bytes_per_frame); + DML_LOG_VERBOSE("DML::%s: tdlut_drain_time = %f\n", __func__, p->tdlut_drain_time); #endif if (p->OutputFormat == dml2_420 || (p->myPipe->InterlaceEnable && p->myPipe->ProgressiveToInterlaceUnitInOPP)) @@ -5314,17 +5249,17 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->DSTYAfterScaler = (unsigned int)(math_floor2(s->DSTTotalPixelsAfterScaler / p->myPipe->HTotal, 1)); *p->DSTXAfterScaler = (unsigned int)(s->DSTTotalPixelsAfterScaler - ((double)(*p->DSTYAfterScaler * p->myPipe->HTotal))); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DSTXAfterScaler = %u (final)\n", __func__, *p->DSTXAfterScaler); - dml2_printf("DML::%s: DSTYAfterScaler = %u (final)\n", __func__, *p->DSTYAfterScaler); + DML_LOG_VERBOSE("DML::%s: DSTXAfterScaler = %u (final)\n", __func__, *p->DSTXAfterScaler); + DML_LOG_VERBOSE("DML::%s: DSTYAfterScaler = %u (final)\n", __func__, *p->DSTYAfterScaler); #endif #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Tr0_trips = %f\n", __func__, *p->Tr0_trips); - dml2_printf("DML::%s: Tvm_trips = %f\n", __func__, *p->Tvm_trips); - dml2_printf("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem); - dml2_printf("DML::%s: ExtraLatencyPrefetch = %f\n", __func__, p->ExtraLatencyPrefetch); - dml2_printf("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->display_cfg->gpuvm_max_page_table_levels); - dml2_printf("DML::%s: HostVMDynamicLevelsTrips = %u\n", __func__, s->HostVMDynamicLevelsTrips); + DML_LOG_VERBOSE("DML::%s: Tr0_trips = %f\n", __func__, *p->Tr0_trips); + DML_LOG_VERBOSE("DML::%s: Tvm_trips = %f\n", __func__, *p->Tvm_trips); + DML_LOG_VERBOSE("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem); + DML_LOG_VERBOSE("DML::%s: ExtraLatencyPrefetch = %f\n", __func__, p->ExtraLatencyPrefetch); + DML_LOG_VERBOSE("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->display_cfg->gpuvm_max_page_table_levels); + DML_LOG_VERBOSE("DML::%s: HostVMDynamicLevelsTrips = %u\n", __func__, s->HostVMDynamicLevelsTrips); #endif if (p->display_cfg->gpuvm_enable) { s->Tvm_trips_rounded = math_ceil2(4.0 * *p->Tvm_trips / s->LineTime, 1.0) / 4.0 * s->LineTime; @@ -5402,7 +5337,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch } /* oto prefetch bw should be always be less than total vactive bw */ - //DML2_ASSERT(s->prefetch_bw_oto < s->per_pipe_vactive_sw_bw * p->myPipe->DPPPerSurface); + //DML_ASSERT(s->prefetch_bw_oto < s->per_pipe_vactive_sw_bw * p->myPipe->DPPPerSurface); s->prefetch_bw_oto = math_max2(s->per_pipe_vactive_sw_bw, s->prefetch_bw_oto) * p->mall_prefetch_sdp_overhead_factor; @@ -5421,9 +5356,9 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->RequiredPrefetchBWOTO = s->prefetch_bw_oto; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: vactive_sw_bw_l = %f\n", __func__, p->vactive_sw_bw_l); - dml2_printf("DML::%s: vactive_sw_bw_c = %f\n", __func__, p->vactive_sw_bw_c); - dml2_printf("DML::%s: per_pipe_vactive_sw_bw = %f\n", __func__, s->per_pipe_vactive_sw_bw); + DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_l = %f\n", __func__, p->vactive_sw_bw_l); + DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_c = %f\n", __func__, p->vactive_sw_bw_c); + DML_LOG_VERBOSE("DML::%s: per_pipe_vactive_sw_bw = %f\n", __func__, s->per_pipe_vactive_sw_bw); #endif if (p->display_cfg->gpuvm_enable == true) { @@ -5433,9 +5368,9 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->LineTime / 4.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Tvm_oto max0 = %f\n", __func__, *p->Tvm_trips); - dml2_printf("DML::%s: Tvm_oto max1 = %f\n", __func__, *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto); - dml2_printf("DML::%s: Tvm_oto max2 = %f\n", __func__, s->LineTime / 4.0); + DML_LOG_VERBOSE("DML::%s: Tvm_oto max0 = %f\n", __func__, *p->Tvm_trips); + DML_LOG_VERBOSE("DML::%s: Tvm_oto max1 = %f\n", __func__, *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto); + DML_LOG_VERBOSE("DML::%s: Tvm_oto max2 = %f\n", __func__, s->LineTime / 4.0); #endif } else { s->Tvm_oto = s->Tvm_trips_rounded; @@ -5447,9 +5382,9 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto, s->LineTime / 4.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Tr0_oto max0 = %f\n", __func__, *p->Tr0_trips); - dml2_printf("DML::%s: Tr0_oto max1 = %f\n", __func__, (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto); - dml2_printf("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime / 4); + DML_LOG_VERBOSE("DML::%s: Tr0_oto max0 = %f\n", __func__, *p->Tr0_trips); + DML_LOG_VERBOSE("DML::%s: Tr0_oto max1 = %f\n", __func__, (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto); + DML_LOG_VERBOSE("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime / 4); #endif } else s->Tr0_oto = s->LineTime / 4.0; @@ -5459,11 +5394,11 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto; #ifdef DML_GLOBAL_PREFETCH_CHECK - dml2_printf("DML::%s: impacted_Tpre = %f\n", __func__, p->impacted_dst_y_pre); + DML_LOG_VERBOSE("DML::%s: impacted_Tpre = %f\n", __func__, p->impacted_dst_y_pre); if (p->impacted_dst_y_pre > 0) { - dml2_printf("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto); + DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto); s->dst_y_prefetch_oto = math_max2(s->dst_y_prefetch_oto, p->impacted_dst_y_pre); - dml2_printf("DML::%s: dst_y_prefetch_oto = %f (impacted)\n", __func__, s->dst_y_prefetch_oto); + DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_oto = %f (impacted)\n", __func__, s->dst_y_prefetch_oto); } #endif *p->Tpre_oto = s->dst_y_prefetch_oto * s->LineTime; @@ -5492,72 +5427,71 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->dst_y_prefetch_equ = math_min2(s->dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal); - dml2_printf("DML::%s: min_Lsw_oto = %f\n", __func__, s->min_Lsw_oto); - dml2_printf("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ); - dml2_printf("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw); - dml2_printf("DML::%s: Tno_bw_flip = %f\n", __func__, *p->Tno_bw_flip); - dml2_printf("DML::%s: ExtraLatencyPrefetch = %f\n", __func__, p->ExtraLatencyPrefetch); - dml2_printf("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem); - dml2_printf("DML::%s: mall_prefetch_sdp_overhead_factor = %f\n", __func__, p->mall_prefetch_sdp_overhead_factor); - dml2_printf("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY); - dml2_printf("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); - dml2_printf("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub); - dml2_printf("DML::%s: BytePerPixelC = %u\n", __func__, p->myPipe->BytePerPixelC); - dml2_printf("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC); - dml2_printf("DML::%s: swath_width_chroma_ub = %u\n", __func__, p->swath_width_chroma_ub); - dml2_printf("DML::%s: prefetch_sw_bytes = %f\n", __func__, *p->prefetch_sw_bytes); - dml2_printf("DML::%s: max_Tsw = %f\n", __func__, s->max_Tsw); - dml2_printf("DML::%s: bytes_pp = %f\n", __func__, s->bytes_pp); - dml2_printf("DML::%s: vm_bytes = %u\n", __func__, vm_bytes); - dml2_printf("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow); - dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); - dml2_printf("DML::%s: Tvm_trips = %f\n", __func__, *p->Tvm_trips); - dml2_printf("DML::%s: Tr0_trips = %f\n", __func__, *p->Tr0_trips); - dml2_printf("DML::%s: Tvm_trips_flip = %f\n", __func__, *p->Tvm_trips_flip); - dml2_printf("DML::%s: Tr0_trips_flip = %f\n", __func__, *p->Tr0_trips_flip); - dml2_printf("DML::%s: prefetch_bw_pr = %f\n", __func__, s->prefetch_bw_pr); - dml2_printf("DML::%s: prefetch_bw_oto = %f\n", __func__, s->prefetch_bw_oto); - dml2_printf("DML::%s: Tr0_oto = %f\n", __func__, s->Tr0_oto); - dml2_printf("DML::%s: Tvm_oto = %f\n", __func__, s->Tvm_oto); - dml2_printf("DML::%s: Tvm_oto_lines = %f\n", __func__, s->Tvm_oto_lines); - dml2_printf("DML::%s: Tr0_oto_lines = %f\n", __func__, s->Tr0_oto_lines); - dml2_printf("DML::%s: Lsw_oto = %f\n", __func__, s->Lsw_oto); - dml2_printf("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto); - dml2_printf("DML::%s: dst_y_prefetch_equ = %f\n", __func__, s->dst_y_prefetch_equ); - dml2_printf("DML::%s: tdlut_row_bytes = %d\n", __func__, tdlut_row_bytes); - dml2_printf("DML::%s: meta_row_bytes = %d\n", __func__, p->meta_row_bytes); -#endif - double Tpre = s->dst_y_prefetch_equ * s->LineTime; + DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal); + DML_LOG_VERBOSE("DML::%s: min_Lsw_oto = %f\n", __func__, s->min_Lsw_oto); + DML_LOG_VERBOSE("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ); + DML_LOG_VERBOSE("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw); + DML_LOG_VERBOSE("DML::%s: Tno_bw_flip = %f\n", __func__, *p->Tno_bw_flip); + DML_LOG_VERBOSE("DML::%s: ExtraLatencyPrefetch = %f\n", __func__, p->ExtraLatencyPrefetch); + DML_LOG_VERBOSE("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem); + DML_LOG_VERBOSE("DML::%s: mall_prefetch_sdp_overhead_factor = %f\n", __func__, p->mall_prefetch_sdp_overhead_factor); + DML_LOG_VERBOSE("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY); + DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); + DML_LOG_VERBOSE("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub); + DML_LOG_VERBOSE("DML::%s: BytePerPixelC = %u\n", __func__, p->myPipe->BytePerPixelC); + DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC); + DML_LOG_VERBOSE("DML::%s: swath_width_chroma_ub = %u\n", __func__, p->swath_width_chroma_ub); + DML_LOG_VERBOSE("DML::%s: prefetch_sw_bytes = %f\n", __func__, *p->prefetch_sw_bytes); + DML_LOG_VERBOSE("DML::%s: max_Tsw = %f\n", __func__, s->max_Tsw); + DML_LOG_VERBOSE("DML::%s: bytes_pp = %f\n", __func__, s->bytes_pp); + DML_LOG_VERBOSE("DML::%s: vm_bytes = %u\n", __func__, vm_bytes); + DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow); + DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: Tvm_trips = %f\n", __func__, *p->Tvm_trips); + DML_LOG_VERBOSE("DML::%s: Tr0_trips = %f\n", __func__, *p->Tr0_trips); + DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip = %f\n", __func__, *p->Tvm_trips_flip); + DML_LOG_VERBOSE("DML::%s: Tr0_trips_flip = %f\n", __func__, *p->Tr0_trips_flip); + DML_LOG_VERBOSE("DML::%s: prefetch_bw_pr = %f\n", __func__, s->prefetch_bw_pr); + DML_LOG_VERBOSE("DML::%s: prefetch_bw_oto = %f\n", __func__, s->prefetch_bw_oto); + DML_LOG_VERBOSE("DML::%s: Tr0_oto = %f\n", __func__, s->Tr0_oto); + DML_LOG_VERBOSE("DML::%s: Tvm_oto = %f\n", __func__, s->Tvm_oto); + DML_LOG_VERBOSE("DML::%s: Tvm_oto_lines = %f\n", __func__, s->Tvm_oto_lines); + DML_LOG_VERBOSE("DML::%s: Tr0_oto_lines = %f\n", __func__, s->Tr0_oto_lines); + DML_LOG_VERBOSE("DML::%s: Lsw_oto = %f\n", __func__, s->Lsw_oto); + DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto); + DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_equ = %f\n", __func__, s->dst_y_prefetch_equ); + DML_LOG_VERBOSE("DML::%s: tdlut_row_bytes = %d\n", __func__, tdlut_row_bytes); + DML_LOG_VERBOSE("DML::%s: meta_row_bytes = %d\n", __func__, p->meta_row_bytes); +#endif s->dst_y_prefetch_equ = math_floor2(4.0 * (s->dst_y_prefetch_equ + 0.125), 1) / 4.0; *p->Tpre_rounded = s->dst_y_prefetch_equ * s->LineTime; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, s->dst_y_prefetch_equ); - dml2_printf("DML::%s: LineTime: %f\n", __func__, s->LineTime); - dml2_printf("DML::%s: VStartup: %u\n", __func__, p->VStartup); - dml2_printf("DML::%s: Tvstartup: %fus - time between vstartup and first pixel of active\n", __func__, p->VStartup * s->LineTime); - dml2_printf("DML::%s: TSetup: %fus - time from vstartup to vready\n", __func__, *p->TSetup); - dml2_printf("DML::%s: TCalc: %fus - time for calculations in dchub starting at vready\n", __func__, p->TCalc); - dml2_printf("DML::%s: TWait: %fus - time for fabric to become ready max(pstate exit,cstate enter/exit, urgent latency) after TCalc\n", __func__, p->TWait); - dml2_printf("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf); - dml2_printf("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec); - dml2_printf("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks); - dml2_printf("DML::%s: TWait = %f\n", __func__, p->TWait); - dml2_printf("DML::%s: TWait_p = %f\n", __func__, s->TWait_p); - dml2_printf("DML::%s: Ttrip = %f\n", __func__, p->Ttrip); - dml2_printf("DML::%s: Tex = %f\n", __func__, p->ExtraLatencyPrefetch); - dml2_printf("DML::%s: Tdmdl_vm: %fus - time for vm stages of dmd \n", __func__, *p->Tdmdl_vm); - dml2_printf("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl); - dml2_printf("DML::%s: TWait_p: %fus\n", __func__, s->TWait_p); - dml2_printf("DML::%s: Ttrip: %fus\n", __func__, p->Ttrip); - dml2_printf("DML::%s: DSTXAfterScaler: %u pixels - number of pixel clocks pipeline and buffer delay after scaler \n", __func__, *p->DSTXAfterScaler); - dml2_printf("DML::%s: DSTYAfterScaler: %u lines - number of lines of pipeline and buffer delay after scaler \n", __func__, *p->DSTYAfterScaler); - dml2_printf("DML::%s: vm_bytes: %f (hvm inefficiency scaled)\n", __func__, vm_bytes*p->HostVMInefficiencyFactor); - dml2_printf("DML::%s: row_bytes: %f (hvm inefficiency scaled, 1 row)\n", __func__, p->PixelPTEBytesPerRow*p->HostVMInefficiencyFactor+p->meta_row_bytes+tdlut_row_bytes); - dml2_printf("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw); - dml2_printf("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, Tpre, *p->Tpre_rounded, (*p->Tpre_rounded - Tpre)); - dml2_printf("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips)); + DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, s->dst_y_prefetch_equ); + DML_LOG_VERBOSE("DML::%s: LineTime: %f\n", __func__, s->LineTime); + DML_LOG_VERBOSE("DML::%s: VStartup: %u\n", __func__, p->VStartup); + DML_LOG_VERBOSE("DML::%s: Tvstartup: %fus - time between vstartup and first pixel of active\n", __func__, p->VStartup * s->LineTime); + DML_LOG_VERBOSE("DML::%s: TSetup: %fus - time from vstartup to vready\n", __func__, *p->TSetup); + DML_LOG_VERBOSE("DML::%s: TCalc: %fus - time for calculations in dchub starting at vready\n", __func__, p->TCalc); + DML_LOG_VERBOSE("DML::%s: TWait: %fus - time for fabric to become ready max(pstate exit,cstate enter/exit, urgent latency) after TCalc\n", __func__, p->TWait); + DML_LOG_VERBOSE("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf); + DML_LOG_VERBOSE("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec); + DML_LOG_VERBOSE("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks); + DML_LOG_VERBOSE("DML::%s: TWait = %f\n", __func__, p->TWait); + DML_LOG_VERBOSE("DML::%s: TWait_p = %f\n", __func__, s->TWait_p); + DML_LOG_VERBOSE("DML::%s: Ttrip = %f\n", __func__, p->Ttrip); + DML_LOG_VERBOSE("DML::%s: Tex = %f\n", __func__, p->ExtraLatencyPrefetch); + DML_LOG_VERBOSE("DML::%s: Tdmdl_vm: %fus - time for vm stages of dmd \n", __func__, *p->Tdmdl_vm); + DML_LOG_VERBOSE("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl); + DML_LOG_VERBOSE("DML::%s: TWait_p: %fus\n", __func__, s->TWait_p); + DML_LOG_VERBOSE("DML::%s: Ttrip: %fus\n", __func__, p->Ttrip); + DML_LOG_VERBOSE("DML::%s: DSTXAfterScaler: %u pixels - number of pixel clocks pipeline and buffer delay after scaler \n", __func__, *p->DSTXAfterScaler); + DML_LOG_VERBOSE("DML::%s: DSTYAfterScaler: %u lines - number of lines of pipeline and buffer delay after scaler \n", __func__, *p->DSTYAfterScaler); + DML_LOG_VERBOSE("DML::%s: vm_bytes: %f (hvm inefficiency scaled)\n", __func__, vm_bytes*p->HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: row_bytes: %f (hvm inefficiency scaled, 1 row)\n", __func__, p->PixelPTEBytesPerRow*p->HostVMInefficiencyFactor+p->meta_row_bytes+tdlut_row_bytes); + DML_LOG_VERBOSE("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw); + DML_LOG_VERBOSE("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, (s->dst_y_prefetch_equ * s->LineTime), *p->Tpre_rounded, (*p->Tpre_rounded - (s->dst_y_prefetch_equ * s->LineTime))); + DML_LOG_VERBOSE("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips)); #endif *p->dst_y_per_vm_vblank = 0; @@ -5596,19 +5530,19 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch } else s->prefetch_bw1 = 0; - dml2_printf("DML::%s: prefetch_bw1: %f\n", __func__, s->prefetch_bw1); + DML_LOG_VERBOSE("DML::%s: prefetch_bw1: %f\n", __func__, s->prefetch_bw1); if ((s->Tsw_est1 < s->min_Lsw_equ * s->LineTime) && (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw > 0)) { s->prefetch_bw1 = (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) / (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: vm and 2 rows bytes = %f\n", __func__, (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes))); - dml2_printf("DML::%s: Tpre_rounded = %f\n", __func__, *p->Tpre_rounded); - dml2_printf("DML::%s: minus term = %f\n", __func__, s->min_Lsw_equ * s->LineTime + 0.75 * s->LineTime + *p->Tno_bw); - dml2_printf("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ); - dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime); - dml2_printf("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw); - dml2_printf("DML::%s: Time to fetch vm and 2 rows = %f\n", __func__, (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw)); - dml2_printf("DML::%s: prefetch_bw1: %f (updated)\n", __func__, s->prefetch_bw1); + DML_LOG_VERBOSE("DML::%s: vm and 2 rows bytes = %f\n", __func__, (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes))); + DML_LOG_VERBOSE("DML::%s: Tpre_rounded = %f\n", __func__, *p->Tpre_rounded); + DML_LOG_VERBOSE("DML::%s: minus term = %f\n", __func__, s->min_Lsw_equ * s->LineTime + 0.75 * s->LineTime + *p->Tno_bw); + DML_LOG_VERBOSE("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ); + DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime); + DML_LOG_VERBOSE("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw); + DML_LOG_VERBOSE("DML::%s: Time to fetch vm and 2 rows = %f\n", __func__, (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw)); + DML_LOG_VERBOSE("DML::%s: prefetch_bw1: %f (updated)\n", __func__, s->prefetch_bw1); #endif } @@ -5620,10 +5554,10 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch } else s->prefetch_bw2 = 0; - dml2_printf("DML::%s: prefetch_bw2: %f\n", __func__, s->prefetch_bw2); + DML_LOG_VERBOSE("DML::%s: prefetch_bw2: %f\n", __func__, s->prefetch_bw2); if ((s->Tsw_est2 < s->min_Lsw_equ * s->LineTime) && ((*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime) > 0)) { s->prefetch_bw2 = vm_bytes * p->HostVMInefficiencyFactor / (*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime); - dml2_printf("DML::%s: prefetch_bw2: %f (updated)\n", __func__, s->prefetch_bw2); + DML_LOG_VERBOSE("DML::%s: prefetch_bw2: %f (updated)\n", __func__, s->prefetch_bw2); } // prefetch_bw3: 2*R0 + SW @@ -5634,10 +5568,10 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch } else s->prefetch_bw3 = 0; - dml2_printf("DML::%s: prefetch_bw3: %f\n", __func__, s->prefetch_bw3); + DML_LOG_VERBOSE("DML::%s: prefetch_bw3: %f\n", __func__, s->prefetch_bw3); if ((s->Tsw_est3 < s->min_Lsw_equ * s->LineTime) && ((*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded) > 0)) { s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) / (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded); - dml2_printf("DML::%s: prefetch_bw3: %f (updated)\n", __func__, s->prefetch_bw3); + DML_LOG_VERBOSE("DML::%s: prefetch_bw3: %f (updated)\n", __func__, s->prefetch_bw3); } // prefetch_bw4: SW @@ -5647,17 +5581,17 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->prefetch_bw4 = 0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw); - dml2_printf("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, Tpre, *p->Tpre_rounded, (*p->Tpre_rounded - Tpre)); - dml2_printf("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips)); - dml2_printf("DML::%s: Tr0_trips=%f Tr0_trips_rounded: %f, delta=%f\n", __func__, *p->Tr0_trips, s->Tr0_trips_rounded, (s->Tr0_trips_rounded - *p->Tr0_trips)); - dml2_printf("DML::%s: Tsw_est1: %f\n", __func__, s->Tsw_est1); - dml2_printf("DML::%s: Tsw_est2: %f\n", __func__, s->Tsw_est2); - dml2_printf("DML::%s: Tsw_est3: %f\n", __func__, s->Tsw_est3); - dml2_printf("DML::%s: prefetch_bw1: %f (final)\n", __func__, s->prefetch_bw1); - dml2_printf("DML::%s: prefetch_bw2: %f (final)\n", __func__, s->prefetch_bw2); - dml2_printf("DML::%s: prefetch_bw3: %f (final)\n", __func__, s->prefetch_bw3); - dml2_printf("DML::%s: prefetch_bw4: %f (final)\n", __func__, s->prefetch_bw4); + DML_LOG_VERBOSE("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw); + DML_LOG_VERBOSE("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, s->dst_y_prefetch_equ * s->LineTime, *p->Tpre_rounded, (*p->Tpre_rounded - (s->dst_y_prefetch_equ * s->LineTime))); + DML_LOG_VERBOSE("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips)); + DML_LOG_VERBOSE("DML::%s: Tr0_trips=%f Tr0_trips_rounded: %f, delta=%f\n", __func__, *p->Tr0_trips, s->Tr0_trips_rounded, (s->Tr0_trips_rounded - *p->Tr0_trips)); + DML_LOG_VERBOSE("DML::%s: Tsw_est1: %f\n", __func__, s->Tsw_est1); + DML_LOG_VERBOSE("DML::%s: Tsw_est2: %f\n", __func__, s->Tsw_est2); + DML_LOG_VERBOSE("DML::%s: Tsw_est3: %f\n", __func__, s->Tsw_est3); + DML_LOG_VERBOSE("DML::%s: prefetch_bw1: %f (final)\n", __func__, s->prefetch_bw1); + DML_LOG_VERBOSE("DML::%s: prefetch_bw2: %f (final)\n", __func__, s->prefetch_bw2); + DML_LOG_VERBOSE("DML::%s: prefetch_bw3: %f (final)\n", __func__, s->prefetch_bw3); + DML_LOG_VERBOSE("DML::%s: prefetch_bw4: %f (final)\n", __func__, s->prefetch_bw4); #endif { bool Case1OK = false; @@ -5676,14 +5610,14 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch double total_row_bytes = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes); - dml2_printf("DML::%s: Tvm_trips_rounded = %f\n", __func__, s->Tvm_trips_rounded); - dml2_printf("DML::%s: Tr0_trips_rounded = %f\n", __func__, s->Tr0_trips_rounded); + DML_LOG_VERBOSE("DML::%s: Tvm_trips_rounded = %f\n", __func__, s->Tvm_trips_rounded); + DML_LOG_VERBOSE("DML::%s: Tr0_trips_rounded = %f\n", __func__, s->Tr0_trips_rounded); if (s->prefetch_bw1 > 0) { double vm_transfer_time = *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw1; double row_transfer_time = total_row_bytes / s->prefetch_bw1; - dml2_printf("DML::%s: Case1: vm_transfer_time = %f\n", __func__, vm_transfer_time); - dml2_printf("DML::%s: Case1: row_transfer_time = %f\n", __func__, row_transfer_time); + DML_LOG_VERBOSE("DML::%s: Case1: vm_transfer_time = %f\n", __func__, vm_transfer_time); + DML_LOG_VERBOSE("DML::%s: Case1: row_transfer_time = %f\n", __func__, row_transfer_time); if (vm_transfer_time >= s->Tvm_trips_rounded && row_transfer_time >= s->Tr0_trips_rounded) { Case1OK = true; } @@ -5696,8 +5630,8 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch if (s->prefetch_bw2 > 0) { double vm_transfer_time = *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw2; double row_transfer_time = total_row_bytes / s->prefetch_bw2; - dml2_printf("DML::%s: Case2: vm_transfer_time = %f\n", __func__, vm_transfer_time); - dml2_printf("DML::%s: Case2: row_transfer_time = %f\n", __func__, row_transfer_time); + DML_LOG_VERBOSE("DML::%s: Case2: vm_transfer_time = %f\n", __func__, vm_transfer_time); + DML_LOG_VERBOSE("DML::%s: Case2: row_transfer_time = %f\n", __func__, row_transfer_time); if (vm_transfer_time >= s->Tvm_trips_rounded && row_transfer_time < s->Tr0_trips_rounded) { Case2OK = true; } @@ -5709,8 +5643,8 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch if (s->prefetch_bw3 > 0) { double vm_transfer_time = *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw3; double row_transfer_time = total_row_bytes / s->prefetch_bw3; - dml2_printf("DML::%s: Case3: vm_transfer_time = %f\n", __func__, vm_transfer_time); - dml2_printf("DML::%s: Case3: row_transfer_time = %f\n", __func__, row_transfer_time); + DML_LOG_VERBOSE("DML::%s: Case3: vm_transfer_time = %f\n", __func__, vm_transfer_time); + DML_LOG_VERBOSE("DML::%s: Case3: row_transfer_time = %f\n", __func__, row_transfer_time); if (vm_transfer_time < s->Tvm_trips_rounded && row_transfer_time >= s->Tr0_trips_rounded) { Case3OK = true; } @@ -5730,10 +5664,10 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch p->vm_bytes * p->HostVMInefficiencyFactor / (31 * s->LineTime) - *p->Tno_bw, (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / (15 * s->LineTime)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Case1OK: %u\n", __func__, Case1OK); - dml2_printf("DML::%s: Case2OK: %u\n", __func__, Case2OK); - dml2_printf("DML::%s: Case3OK: %u\n", __func__, Case3OK); - dml2_printf("DML::%s: prefetch_bw_equ: %f\n", __func__, s->prefetch_bw_equ); + DML_LOG_VERBOSE("DML::%s: Case1OK: %u\n", __func__, Case1OK); + DML_LOG_VERBOSE("DML::%s: Case2OK: %u\n", __func__, Case2OK); + DML_LOG_VERBOSE("DML::%s: Case3OK: %u\n", __func__, Case3OK); + DML_LOG_VERBOSE("DML::%s: prefetch_bw_equ: %f\n", __func__, s->prefetch_bw_equ); #endif if (s->prefetch_bw_equ > 0) { @@ -5753,12 +5687,12 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch } else { s->Tvm_equ = 0; s->Tr0_equ = 0; - dml2_printf("DML::%s: prefetch_bw_equ equals 0!\n", __func__); + DML_LOG_VERBOSE("DML::%s: prefetch_bw_equ equals 0!\n", __func__); } } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Tvm_equ = %f\n", __func__, s->Tvm_equ); - dml2_printf("DML::%s: Tr0_equ = %f\n", __func__, s->Tr0_equ); + DML_LOG_VERBOSE("DML::%s: Tvm_equ = %f\n", __func__, s->Tvm_equ); + DML_LOG_VERBOSE("DML::%s: Tr0_equ = %f\n", __func__, s->Tr0_equ); #endif // Use the more stressful prefetch schedule if (s->dst_y_prefetch_oto < s->dst_y_prefetch_equ) { @@ -5769,7 +5703,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0; *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Using oto scheduling for prefetch\n", __func__); + DML_LOG_VERBOSE("DML::%s: Using oto scheduling for prefetch\n", __func__); #endif } else { @@ -5785,7 +5719,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Using equ bw scheduling for prefetch\n", __func__); + DML_LOG_VERBOSE("DML::%s: Using equ bw scheduling for prefetch\n", __func__); #endif } @@ -5797,31 +5731,31 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->prefetch_swath_time_us = (s->LinesToRequestPrefetchPixelData * s->LineTime); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: TimeForFetchingVM = %f\n", __func__, s->TimeForFetchingVM); - dml2_printf("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, s->TimeForFetchingRowInVBlank); - dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime); - dml2_printf("DML::%s: dst_y_prefetch = %f\n", __func__, *p->dst_y_prefetch); - dml2_printf("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank); - dml2_printf("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank); - dml2_printf("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData); - dml2_printf("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); - dml2_printf("DML::%s: prefetch_swath_time_us = %f\n", __func__, *p->prefetch_swath_time_us); + DML_LOG_VERBOSE("DML::%s: TimeForFetchingVM = %f\n", __func__, s->TimeForFetchingVM); + DML_LOG_VERBOSE("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, s->TimeForFetchingRowInVBlank); + DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime); + DML_LOG_VERBOSE("DML::%s: dst_y_prefetch = %f\n", __func__, *p->dst_y_prefetch); + DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank); + DML_LOG_VERBOSE("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData); + DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); + DML_LOG_VERBOSE("DML::%s: prefetch_swath_time_us = %f\n", __func__, *p->prefetch_swath_time_us); - dml2_printf("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, p->cursor_bytes_per_chunk); - dml2_printf("DML::%s: cursor_bytes_per_line = %d\n", __func__, p->cursor_bytes_per_line); - dml2_printf("DML::%s: cursor_prefetch_bytes = %d\n", __func__, s->cursor_prefetch_bytes); - dml2_printf("DML::%s: prefetch_cursor_bw = %f\n", __func__, *p->prefetch_cursor_bw); + DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, p->cursor_bytes_per_chunk); + DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_line = %d\n", __func__, p->cursor_bytes_per_line); + DML_LOG_VERBOSE("DML::%s: cursor_prefetch_bytes = %d\n", __func__, s->cursor_prefetch_bytes); + DML_LOG_VERBOSE("DML::%s: prefetch_cursor_bw = %f\n", __func__, *p->prefetch_cursor_bw); #endif - DML2_ASSERT(*p->dst_y_prefetch < 64); + DML_ASSERT(*p->dst_y_prefetch < 64); unsigned int min_lsw_required = (unsigned int)math_max2(2, p->tdlut_drain_time / s->LineTime); if (s->LinesToRequestPrefetchPixelData >= min_lsw_required && s->prefetch_bw_equ > 0) { *p->VRatioPrefetchY = (double)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData; *p->VRatioPrefetchY = math_max2(*p->VRatioPrefetchY, 1.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY); - dml2_printf("DML::%s: SwathHeightY = %u\n", __func__, p->SwathHeightY); - dml2_printf("DML::%s: VInitPreFillY = %u\n", __func__, p->VInitPreFillY); + DML_LOG_VERBOSE("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY); + DML_LOG_VERBOSE("DML::%s: SwathHeightY = %u\n", __func__, p->SwathHeightY); + DML_LOG_VERBOSE("DML::%s: VInitPreFillY = %u\n", __func__, p->VInitPreFillY); #endif if ((p->SwathHeightY > 4) && (p->VInitPreFillY > 3)) { if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillY - 3.0) / 2.0) { @@ -5829,13 +5763,13 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch (double)p->MaxNumSwathY * p->SwathHeightY / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillY - 3.0) / 2.0)); } else { s->NoTimeToPrefetch = true; - dml2_printf("DML::%s: No time to prefetch!. LinesToRequestPrefetchPixelData=%f VinitPreFillY=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillY); + DML_LOG_VERBOSE("DML::%s: No time to prefetch!. LinesToRequestPrefetchPixelData=%f VinitPreFillY=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillY); *p->VRatioPrefetchY = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY); - dml2_printf("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); - dml2_printf("DML::%s: MaxNumSwathY = %u\n", __func__, p->MaxNumSwathY); + DML_LOG_VERBOSE("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY); + DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); + DML_LOG_VERBOSE("DML::%s: MaxNumSwathY = %u\n", __func__, p->MaxNumSwathY); #endif } @@ -5843,22 +5777,22 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->VRatioPrefetchC = math_max2(*p->VRatioPrefetchC, 1.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC); - dml2_printf("DML::%s: SwathHeightC = %u\n", __func__, p->SwathHeightC); - dml2_printf("DML::%s: VInitPreFillC = %u\n", __func__, p->VInitPreFillC); + DML_LOG_VERBOSE("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC); + DML_LOG_VERBOSE("DML::%s: SwathHeightC = %u\n", __func__, p->SwathHeightC); + DML_LOG_VERBOSE("DML::%s: VInitPreFillC = %u\n", __func__, p->VInitPreFillC); #endif if ((p->SwathHeightC > 4) && (p->VInitPreFillC > 3)) { if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillC - 3.0) / 2.0) { *p->VRatioPrefetchC = math_max2(*p->VRatioPrefetchC, (double)p->MaxNumSwathC * p->SwathHeightC / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillC - 3.0) / 2.0)); } else { s->NoTimeToPrefetch = true; - dml2_printf("DML::%s: No time to prefetch!. LinesToRequestPrefetchPixelData=%f VInitPreFillC=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillC); + DML_LOG_VERBOSE("DML::%s: No time to prefetch!. LinesToRequestPrefetchPixelData=%f VInitPreFillC=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillC); *p->VRatioPrefetchC = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC); - dml2_printf("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC); - dml2_printf("DML::%s: MaxNumSwathC = %u\n", __func__, p->MaxNumSwathC); + DML_LOG_VERBOSE("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC); + DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC); + DML_LOG_VERBOSE("DML::%s: MaxNumSwathC = %u\n", __func__, p->MaxNumSwathC); #endif } @@ -5866,36 +5800,34 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->RequiredPrefetchPixelDataBWChroma = (double)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData * p->myPipe->BytePerPixelC * p->swath_width_chroma_ub / s->LineTime; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY); - dml2_printf("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub); - dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime); - dml2_printf("DML::%s: RequiredPrefetchPixelDataBWLuma = %f\n", __func__, *p->RequiredPrefetchPixelDataBWLuma); - dml2_printf("DML::%s: RequiredPrefetchPixelDataBWChroma = %f\n", __func__, *p->RequiredPrefetchPixelDataBWChroma); + DML_LOG_VERBOSE("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY); + DML_LOG_VERBOSE("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub); + DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime); + DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWLuma = %f\n", __func__, *p->RequiredPrefetchPixelDataBWLuma); + DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWChroma = %f\n", __func__, *p->RequiredPrefetchPixelDataBWChroma); #endif } else { s->NoTimeToPrefetch = true; - dml2_printf("DML::%s: No time to prefetch!, LinesToRequestPrefetchPixelData: %f, should be >= %d\n", __func__, s->LinesToRequestPrefetchPixelData, min_lsw_required); - dml2_printf("DML::%s: No time to prefetch!, prefetch_bw_equ: %f, should be > 0\n", __func__, s->prefetch_bw_equ); + DML_LOG_VERBOSE("DML::%s: No time to prefetch!, LinesToRequestPrefetchPixelData: %f, should be >= %d\n", __func__, s->LinesToRequestPrefetchPixelData, min_lsw_required); + DML_LOG_VERBOSE("DML::%s: No time to prefetch!, prefetch_bw_equ: %f, should be > 0\n", __func__, s->prefetch_bw_equ); *p->VRatioPrefetchY = 0; *p->VRatioPrefetchC = 0; *p->RequiredPrefetchPixelDataBWLuma = 0; *p->RequiredPrefetchPixelDataBWChroma = 0; } - dml2_printf("DML: Tpre: %fus - sum of time to request 2 x data pte, swaths\n", (double)s->LinesToRequestPrefetchPixelData * s->LineTime + 2.0 * s->TimeForFetchingRowInVBlank + s->TimeForFetchingVM); - dml2_printf("DML: Tvm: %fus - time to fetch vm\n", s->TimeForFetchingVM); - dml2_printf("DML: Tr0: %fus - time to fetch first row of data pagetables\n", s->TimeForFetchingRowInVBlank); - dml2_printf("DML: Tsw: %fus = time to fetch enough pixel data and cursor data to feed the scalers init position and detile\n", (double)s->LinesToRequestPrefetchPixelData * s->LineTime); - dml2_printf("DML: To: %fus - time for propagation from scaler to optc\n", (*p->DSTYAfterScaler + ((double)(*p->DSTXAfterScaler) / (double)p->myPipe->HTotal)) * s->LineTime); - dml2_printf("DML: Tvstartup - TSetup - Tcalc - TWait - Tpre - To > 0\n"); - dml2_printf("DML: Tslack(pre): %fus - time left over in schedule\n", p->VStartup * s->LineTime - s->TimeForFetchingVM - 2 * s->TimeForFetchingRowInVBlank - (*p->DSTYAfterScaler + ((double)(*p->DSTXAfterScaler) / (double)p->myPipe->HTotal)) * s->LineTime - p->TWait - p->TCalc - *p->TSetup); - dml2_printf("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %u\n", p->PixelPTEBytesPerRow); + DML_LOG_VERBOSE("DML: Tpre: %fus - sum of time to request 2 x data pte, swaths\n", (double)s->LinesToRequestPrefetchPixelData * s->LineTime + 2.0 * s->TimeForFetchingRowInVBlank + s->TimeForFetchingVM); + DML_LOG_VERBOSE("DML: Tvm: %fus - time to fetch vm\n", s->TimeForFetchingVM); + DML_LOG_VERBOSE("DML: Tr0: %fus - time to fetch first row of data pagetables\n", s->TimeForFetchingRowInVBlank); + DML_LOG_VERBOSE("DML: Tsw: %fus = time to fetch enough pixel data and cursor data to feed the scalers init position and detile\n", (double)s->LinesToRequestPrefetchPixelData * s->LineTime); + DML_LOG_VERBOSE("DML: To: %fus - time for propagation from scaler to optc\n", (*p->DSTYAfterScaler + ((double)(*p->DSTXAfterScaler) / (double)p->myPipe->HTotal)) * s->LineTime); + DML_LOG_VERBOSE("DML: Tvstartup - TSetup - Tcalc - TWait - Tpre - To > 0\n"); + DML_LOG_VERBOSE("DML: Tslack(pre): %fus - time left over in schedule\n", p->VStartup * s->LineTime - s->TimeForFetchingVM - 2 * s->TimeForFetchingRowInVBlank - (*p->DSTYAfterScaler + ((double)(*p->DSTXAfterScaler) / (double)p->myPipe->HTotal)) * s->LineTime - p->TWait - p->TCalc - *p->TSetup); + DML_LOG_VERBOSE("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %u\n", p->PixelPTEBytesPerRow); } else { - dml2_printf("DML::%s: No time to prefetch! dst_y_prefetch_equ = %f (should be > 1)\n", __func__, s->dst_y_prefetch_equ); - dml2_printf("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded (%f) should be >= Tvm_trips_rounded (%f) + 2.0*Tr0_trips_rounded (%f) + min_Tsw_equ (%f)\n", + DML_LOG_VERBOSE("DML::%s: No time to prefetch! dst_y_prefetch_equ = %f (should be > 1)\n", __func__, s->dst_y_prefetch_equ); + DML_LOG_VERBOSE("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded (%f) should be >= Tvm_trips_rounded (%f) + 2.0*Tr0_trips_rounded (%f) + min_Tsw_equ (%f)\n", __func__, min_Lsw_equ_ok, *p->Tpre_rounded, s->Tvm_trips_rounded, 2.0*s->Tr0_trips_rounded, s->min_Lsw_equ*s->LineTime); - dml2_printf("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded+Tvm_trips_rounded+2.0*Tr0_trips_rounded+min_Tsw_equ (%f) should be > \n", - __func__, tpre_gt_req_latency, (s->min_Lsw_equ*s->LineTime + s->Tvm_trips_rounded + 2.0*s->Tr0_trips_rounded), p->Turg, s->trip_to_mem, p->ExtraLatencyPrefetch); s->NoTimeToPrefetch = true; s->TimeForFetchingVM = 0; s->TimeForFetchingRowInVBlank = 0; @@ -5916,18 +5848,18 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch prefetch_vm_bw = 0; } else if (*p->dst_y_per_vm_vblank > 0) { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); - dml2_printf("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank); - dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime); + DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime); #endif prefetch_vm_bw = vm_bytes * p->HostVMInefficiencyFactor / (*p->dst_y_per_vm_vblank * s->LineTime); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw); + DML_LOG_VERBOSE("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw); #endif } else { prefetch_vm_bw = 0; s->NoTimeToPrefetch = true; - dml2_printf("DML::%s: No time to prefetch!. dst_y_per_vm_vblank=%f (should be > 0)\n", __func__, *p->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML::%s: No time to prefetch!. dst_y_per_vm_vblank=%f (should be > 0)\n", __func__, *p->dst_y_per_vm_vblank); } if (p->PixelPTEBytesPerRow == 0 && tdlut_row_bytes == 0) { @@ -5936,14 +5868,14 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch prefetch_row_bw = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + tdlut_row_bytes) / (*p->dst_y_per_row_vblank * s->LineTime); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow); - dml2_printf("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank); - dml2_printf("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw); + DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow); + DML_LOG_VERBOSE("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank); + DML_LOG_VERBOSE("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw); #endif } else { prefetch_row_bw = 0; s->NoTimeToPrefetch = true; - dml2_printf("DML::%s: No time to prefetch!. dst_y_per_row_vblank=%f (should be > 0)\n", __func__, *p->dst_y_per_row_vblank); + DML_LOG_VERBOSE("DML::%s: No time to prefetch!. dst_y_per_row_vblank=%f (should be > 0)\n", __func__, *p->dst_y_per_row_vblank); } *p->prefetch_vmrow_bw = math_max2(prefetch_vm_bw, prefetch_row_bw); @@ -5963,12 +5895,12 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->prefetch_vmrow_bw = 0; } - dml2_printf("DML::%s: dst_y_per_vm_vblank = %f (final)\n", __func__, *p->dst_y_per_vm_vblank); - dml2_printf("DML::%s: dst_y_per_row_vblank = %f (final)\n", __func__, *p->dst_y_per_row_vblank); - dml2_printf("DML::%s: prefetch_vmrow_bw = %f (final)\n", __func__, *p->prefetch_vmrow_bw); - dml2_printf("DML::%s: RequiredPrefetchPixelDataBWLuma = %f (final)\n", __func__, *p->RequiredPrefetchPixelDataBWLuma); - dml2_printf("DML::%s: RequiredPrefetchPixelDataBWChroma = %f (final)\n", __func__, *p->RequiredPrefetchPixelDataBWChroma); - dml2_printf("DML::%s: NoTimeToPrefetch=%d\n", __func__, s->NoTimeToPrefetch); + DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_vblank = %f (final)\n", __func__, *p->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML::%s: dst_y_per_row_vblank = %f (final)\n", __func__, *p->dst_y_per_row_vblank); + DML_LOG_VERBOSE("DML::%s: prefetch_vmrow_bw = %f (final)\n", __func__, *p->prefetch_vmrow_bw); + DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWLuma = %f (final)\n", __func__, *p->RequiredPrefetchPixelDataBWLuma); + DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWChroma = %f (final)\n", __func__, *p->RequiredPrefetchPixelDataBWChroma); + DML_LOG_VERBOSE("DML::%s: NoTimeToPrefetch=%d\n", __func__, s->NoTimeToPrefetch); return s->NoTimeToPrefetch; } @@ -6005,7 +5937,7 @@ static unsigned int find_max_impact_plane(unsigned int this_plane_idx, unsigned } } if (max_idx <= 0) { - DML2_ASSERT(max_idx >= 0); + DML_ASSERT(max_idx >= 0); max_idx = this_plane_idx; } @@ -6037,12 +5969,12 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core // worst case if the rob and cdb is fully hogged s->max_Trpd_dcfclk_cycles = (unsigned int) math_ceil2((p->rob_buffer_size_kbytes*1024 + p->compressed_buffer_size_kbytes*DML_MAX_COMPRESSION_RATIO*1024)/64.0, 1.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: num_active_planes = %d\n", __func__, p->num_active_planes); - dml2_printf("DML::%s: rob_buffer_size_kbytes = %d\n", __func__, p->rob_buffer_size_kbytes); - dml2_printf("DML::%s: compressed_buffer_size_kbytes = %d\n", __func__, p->compressed_buffer_size_kbytes); - dml2_printf("DML::%s: estimated_urg_bandwidth_required_mbps = %f\n", __func__, p->estimated_urg_bandwidth_required_mbps); - dml2_printf("DML::%s: estimated_dcfclk_mhz = %f\n", __func__, p->estimated_dcfclk_mhz); - dml2_printf("DML::%s: max_Trpd_dcfclk_cycles = %u\n", __func__, s->max_Trpd_dcfclk_cycles); + DML_LOG_VERBOSE("DML::%s: num_active_planes = %d\n", __func__, p->num_active_planes); + DML_LOG_VERBOSE("DML::%s: rob_buffer_size_kbytes = %d\n", __func__, p->rob_buffer_size_kbytes); + DML_LOG_VERBOSE("DML::%s: compressed_buffer_size_kbytes = %d\n", __func__, p->compressed_buffer_size_kbytes); + DML_LOG_VERBOSE("DML::%s: estimated_urg_bandwidth_required_mbps = %f\n", __func__, p->estimated_urg_bandwidth_required_mbps); + DML_LOG_VERBOSE("DML::%s: estimated_dcfclk_mhz = %f\n", __func__, p->estimated_dcfclk_mhz); + DML_LOG_VERBOSE("DML::%s: max_Trpd_dcfclk_cycles = %u\n", __func__, s->max_Trpd_dcfclk_cycles); #endif // calculate the return impact from each plane, request is 256B per dcfclk @@ -6063,12 +5995,12 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(p->lb_source_lines_l[i] / p->swath_height_l[i], 1) * s->src_swath_bytes_l[i]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: i=%u pixel_format = %d\n", __func__, i, p->pixel_format[i]); - dml2_printf("DML::%s: i=%u chunk_bytes_l = %d\n", __func__, i, p->chunk_bytes_l); - dml2_printf("DML::%s: i=%u lb_source_lines_l = %d\n", __func__, i, p->lb_source_lines_l[i]); - dml2_printf("DML::%s: i=%u src_detile_buf_size_bytes_l=%d\n", __func__, i, s->src_detile_buf_size_bytes_l[i]); - dml2_printf("DML::%s: i=%u src_swath_bytes_l=%d\n", __func__, i, s->src_swath_bytes_l[i]); - dml2_printf("DML::%s: i=%u burst_bytes_to_fill_det=%d (luma)\n", __func__, i, s->burst_bytes_to_fill_det); + DML_LOG_VERBOSE("DML::%s: i=%u pixel_format = %d\n", __func__, i, p->pixel_format[i]); + DML_LOG_VERBOSE("DML::%s: i=%u chunk_bytes_l = %d\n", __func__, i, p->chunk_bytes_l); + DML_LOG_VERBOSE("DML::%s: i=%u lb_source_lines_l = %d\n", __func__, i, p->lb_source_lines_l[i]); + DML_LOG_VERBOSE("DML::%s: i=%u src_detile_buf_size_bytes_l=%d\n", __func__, i, s->src_detile_buf_size_bytes_l[i]); + DML_LOG_VERBOSE("DML::%s: i=%u src_swath_bytes_l=%d\n", __func__, i, s->src_swath_bytes_l[i]); + DML_LOG_VERBOSE("DML::%s: i=%u burst_bytes_to_fill_det=%d (luma)\n", __func__, i, s->burst_bytes_to_fill_det); #endif if (s->src_swath_bytes_c[i] > 0) { // dual_plane @@ -6079,10 +6011,10 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: i=%u chunk_bytes_c = %d\n", __func__, i, p->chunk_bytes_c); - dml2_printf("DML::%s: i=%u lb_source_lines_c = %d\n", __func__, i, p->lb_source_lines_c[i]); - dml2_printf("DML::%s: i=%u src_detile_buf_size_bytes_c=%d\n", __func__, i, s->src_detile_buf_size_bytes_c[i]); - dml2_printf("DML::%s: i=%u src_swath_bytes_c=%d\n", __func__, i, s->src_swath_bytes_c[i]); + DML_LOG_VERBOSE("DML::%s: i=%u chunk_bytes_c = %d\n", __func__, i, p->chunk_bytes_c); + DML_LOG_VERBOSE("DML::%s: i=%u lb_source_lines_c = %d\n", __func__, i, p->lb_source_lines_c[i]); + DML_LOG_VERBOSE("DML::%s: i=%u src_detile_buf_size_bytes_c=%d\n", __func__, i, s->src_detile_buf_size_bytes_c[i]); + DML_LOG_VERBOSE("DML::%s: i=%u src_swath_bytes_c=%d\n", __func__, i, s->src_swath_bytes_c[i]); #endif } @@ -6090,9 +6022,9 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core s->accumulated_return_path_dcfclk_cycles[i] = (unsigned int) math_ceil2(((DML_MAX_COMPRESSION_RATIO-1) * 64 * p->estimated_dcfclk_mhz) * s->time_to_fill_det_us / 64.0, 1.0); //for 64B per DCFClk #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: i=%u burst_bytes_to_fill_det=%d\n", __func__, i, s->burst_bytes_to_fill_det); - dml2_printf("DML::%s: i=%u time_to_fill_det_us=%f\n", __func__, i, s->time_to_fill_det_us); - dml2_printf("DML::%s: i=%u accumulated_return_path_dcfclk_cycles=%u\n", __func__, i, s->accumulated_return_path_dcfclk_cycles[i]); + DML_LOG_VERBOSE("DML::%s: i=%u burst_bytes_to_fill_det=%d\n", __func__, i, s->burst_bytes_to_fill_det); + DML_LOG_VERBOSE("DML::%s: i=%u time_to_fill_det_us=%f\n", __func__, i, s->time_to_fill_det_us); + DML_LOG_VERBOSE("DML::%s: i=%u accumulated_return_path_dcfclk_cycles=%u\n", __func__, i, s->accumulated_return_path_dcfclk_cycles[i]); #endif // clamping to worst case delay which is one which occupy the full rob+cdb if (s->accumulated_return_path_dcfclk_cycles[i] > s->max_Trpd_dcfclk_cycles) @@ -6109,7 +6041,7 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core p->impacted_dst_y_pre[i] = math_ceil2(p->impacted_dst_y_pre[i] / p->line_time[i], 0.25); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: i=%u impacted_Tpre=%f (k=%u)\n", __func__, i, p->impacted_dst_y_pre[i], k); + DML_LOG_VERBOSE("DML::%s: i=%u impacted_Tpre=%f (k=%u)\n", __func__, i, p->impacted_dst_y_pre[i], k); #endif } @@ -6120,8 +6052,8 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core *p->recalc_prefetch_schedule = 1; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: i=%u Tpre_rounded=%f\n", __func__, i, p->Tpre_rounded[i]); - dml2_printf("DML::%s: i=%u Tpre_oto=%f\n", __func__, i, p->Tpre_oto[i]); + DML_LOG_VERBOSE("DML::%s: i=%u Tpre_rounded=%f\n", __func__, i, p->Tpre_rounded[i]); + DML_LOG_VERBOSE("DML::%s: i=%u Tpre_oto=%f\n", __func__, i, p->Tpre_oto[i]); #endif } } else { @@ -6131,8 +6063,8 @@ static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: prefetch_global_check_passed=%u\n", __func__, s->prefetch_global_check_passed); - dml2_printf("DML::%s: recalc_prefetch_schedule=%u\n", __func__, *p->recalc_prefetch_schedule); + DML_LOG_VERBOSE("DML::%s: prefetch_global_check_passed=%u\n", __func__, s->prefetch_global_check_passed); + DML_LOG_VERBOSE("DML::%s: recalc_prefetch_schedule=%u\n", __func__, *p->recalc_prefetch_schedule); #endif return s->prefetch_global_check_passed; @@ -6150,8 +6082,8 @@ static void calculate_peak_bandwidth_required( memset(l, 0, sizeof(struct dml2_core_shared_calculate_peak_bandwidth_required_locals)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: inc_flip_bw = %d\n", __func__, p->inc_flip_bw); - dml2_printf("DML::%s: NumberOfActiveSurfaces = %d\n", __func__, p->num_active_planes); + DML_LOG_VERBOSE("DML::%s: inc_flip_bw = %d\n", __func__, p->inc_flip_bw); + DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %d\n", __func__, p->num_active_planes); #endif for (unsigned int k = 0; k < p->num_active_planes; ++k) { @@ -6347,12 +6279,12 @@ static void calculate_peak_bandwidth_required( p->surface_peak_required_bw[m][n]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: urg_vactive_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_vactive_bandwidth_required[m][n]); - dml2_printf("DML::%s: urg_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_bandwidth_required[m][n]); - dml2_printf("DML::%s: urg_bandwidth_required_qual[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_bandwidth_required[m][n]); - dml2_printf("DML::%s: non_urg_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->non_urg_bandwidth_required[m][n]); + DML_LOG_VERBOSE("DML::%s: urg_vactive_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_vactive_bandwidth_required[m][n]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_bandwidth_required[m][n]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_qual[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_bandwidth_required[m][n]); + DML_LOG_VERBOSE("DML::%s: non_urg_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->non_urg_bandwidth_required[m][n]); #endif - DML2_ASSERT(p->urg_bandwidth_required[m][n] >= p->non_urg_bandwidth_required[m][n]); + DML_ASSERT(p->urg_bandwidth_required[m][n] >= p->non_urg_bandwidth_required[m][n]); } } } @@ -6414,18 +6346,18 @@ static void check_urgent_bandwidth_support( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: frac_urg_bandwidth_nom_sdp = %f\n", __func__, frac_urg_bandwidth_nom_sdp); - dml2_printf("DML::%s: frac_urg_bandwidth_nom_dram = %f\n", __func__, frac_urg_bandwidth_nom_dram); - dml2_printf("DML::%s: frac_urg_bandwidth_nom = %f\n", __func__, *frac_urg_bandwidth_nom); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_nom_sdp = %f\n", __func__, frac_urg_bandwidth_nom_sdp); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_nom_dram = %f\n", __func__, frac_urg_bandwidth_nom_dram); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_nom = %f\n", __func__, *frac_urg_bandwidth_nom); - dml2_printf("DML::%s: frac_urg_bandwidth_mall_sdp = %f\n", __func__, frac_urg_bandwidth_mall_sdp); - dml2_printf("DML::%s: frac_urg_bandwidth_mall_dram = %f\n", __func__, frac_urg_bandwidth_mall_dram); - dml2_printf("DML::%s: frac_urg_bandwidth_mall = %f\n", __func__, *frac_urg_bandwidth_mall); - dml2_printf("DML::%s: bandwidth_support_ok = %d\n", __func__, *bandwidth_support_ok); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_mall_sdp = %f\n", __func__, frac_urg_bandwidth_mall_sdp); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_mall_dram = %f\n", __func__, frac_urg_bandwidth_mall_dram); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_mall = %f\n", __func__, *frac_urg_bandwidth_mall); + DML_LOG_VERBOSE("DML::%s: bandwidth_support_ok = %d\n", __func__, *bandwidth_support_ok); for (unsigned int m = 0; m < dml2_core_internal_soc_state_max; m++) { for (unsigned int n = 0; n < dml2_core_internal_bw_max; n++) { - dml2_printf("DML::%s: state:%s bw_type:%s urg_bandwidth_available=%f %s urg_bandwidth_required=%f\n", + DML_LOG_VERBOSE("DML::%s: state:%s bw_type:%s urg_bandwidth_available=%f %s urg_bandwidth_required=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), urg_bandwidth_available[m][n], (urg_bandwidth_available[m][n] < urg_bandwidth_required[m][n]) ? "<" : ">=", urg_bandwidth_required[m][n]); } @@ -6446,14 +6378,14 @@ static double get_bandwidth_available_for_immediate_flip(enum dml2_core_internal flip_bw_available_mbps = flip_bw_available_sdp_mbps < flip_bw_available_dram_mbps ? flip_bw_available_sdp_mbps : flip_bw_available_dram_mbps; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: eval_state = %s\n", __func__, dml2_core_internal_soc_state_type_str(eval_state)); - dml2_printf("DML::%s: urg_bandwidth_available_sdp_mbps = %f\n", __func__, urg_bandwidth_available[eval_state][dml2_core_internal_bw_sdp]); - dml2_printf("DML::%s: urg_bandwidth_available_dram_mbps = %f\n", __func__, urg_bandwidth_available[eval_state][dml2_core_internal_bw_dram]); - dml2_printf("DML::%s: urg_bandwidth_required_sdp_mbps = %f\n", __func__, urg_bandwidth_required[eval_state][dml2_core_internal_bw_sdp]); - dml2_printf("DML::%s: urg_bandwidth_required_dram_mbps = %f\n", __func__, urg_bandwidth_required[eval_state][dml2_core_internal_bw_dram]); - dml2_printf("DML::%s: flip_bw_available_sdp_mbps = %f\n", __func__, flip_bw_available_sdp_mbps); - dml2_printf("DML::%s: flip_bw_available_dram_mbps = %f\n", __func__, flip_bw_available_dram_mbps); - dml2_printf("DML::%s: flip_bw_available_mbps = %f\n", __func__, flip_bw_available_mbps); + DML_LOG_VERBOSE("DML::%s: eval_state = %s\n", __func__, dml2_core_internal_soc_state_type_str(eval_state)); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_sdp_mbps = %f\n", __func__, urg_bandwidth_available[eval_state][dml2_core_internal_bw_sdp]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_dram_mbps = %f\n", __func__, urg_bandwidth_available[eval_state][dml2_core_internal_bw_dram]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_sdp_mbps = %f\n", __func__, urg_bandwidth_required[eval_state][dml2_core_internal_bw_sdp]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_dram_mbps = %f\n", __func__, urg_bandwidth_required[eval_state][dml2_core_internal_bw_dram]); + DML_LOG_VERBOSE("DML::%s: flip_bw_available_sdp_mbps = %f\n", __func__, flip_bw_available_sdp_mbps); + DML_LOG_VERBOSE("DML::%s: flip_bw_available_dram_mbps = %f\n", __func__, flip_bw_available_dram_mbps); + DML_LOG_VERBOSE("DML::%s: flip_bw_available_mbps = %f\n", __func__, flip_bw_available_mbps); #endif return flip_bw_available_mbps; @@ -6478,28 +6410,28 @@ static void calculate_immediate_flip_bandwidth_support( *flip_bandwidth_support_ok &= urg_bandwidth_available[eval_state][n] >= urg_bandwidth_required_flip[eval_state][n]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: n = %s\n", __func__, dml2_core_internal_bw_type_str(n)); - dml2_printf("DML::%s: urg_bandwidth_available = %f\n", __func__, urg_bandwidth_available[eval_state][n]); - dml2_printf("DML::%s: non_urg_bandwidth_required_flip = %f\n", __func__, non_urg_bandwidth_required_flip[eval_state][n]); - dml2_printf("DML::%s: urg_bandwidth_required_flip = %f\n", __func__, urg_bandwidth_required_flip[eval_state][n]); - dml2_printf("DML::%s: flip_bandwidth_support_ok = %d\n", __func__, *flip_bandwidth_support_ok); + DML_LOG_VERBOSE("DML::%s: n = %s\n", __func__, dml2_core_internal_bw_type_str(n)); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available = %f\n", __func__, urg_bandwidth_available[eval_state][n]); + DML_LOG_VERBOSE("DML::%s: non_urg_bandwidth_required_flip = %f\n", __func__, non_urg_bandwidth_required_flip[eval_state][n]); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_flip = %f\n", __func__, urg_bandwidth_required_flip[eval_state][n]); + DML_LOG_VERBOSE("DML::%s: flip_bandwidth_support_ok = %d\n", __func__, *flip_bandwidth_support_ok); #endif - DML2_ASSERT(urg_bandwidth_required_flip[eval_state][n] >= non_urg_bandwidth_required_flip[eval_state][n]); + DML_ASSERT(urg_bandwidth_required_flip[eval_state][n] >= non_urg_bandwidth_required_flip[eval_state][n]); } *frac_urg_bandwidth_flip = (frac_urg_bw_flip_sdp > frac_urg_bw_flip_dram) ? frac_urg_bw_flip_sdp : frac_urg_bw_flip_dram; *flip_bandwidth_support_ok &= (*frac_urg_bandwidth_flip <= 1.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: eval_state = %s\n", __func__, dml2_core_internal_soc_state_type_str(eval_state)); - dml2_printf("DML::%s: frac_urg_bw_flip_sdp = %f\n", __func__, frac_urg_bw_flip_sdp); - dml2_printf("DML::%s: frac_urg_bw_flip_dram = %f\n", __func__, frac_urg_bw_flip_dram); - dml2_printf("DML::%s: frac_urg_bandwidth_flip = %f\n", __func__, *frac_urg_bandwidth_flip); - dml2_printf("DML::%s: flip_bandwidth_support_ok = %d\n", __func__, *flip_bandwidth_support_ok); + DML_LOG_VERBOSE("DML::%s: eval_state = %s\n", __func__, dml2_core_internal_soc_state_type_str(eval_state)); + DML_LOG_VERBOSE("DML::%s: frac_urg_bw_flip_sdp = %f\n", __func__, frac_urg_bw_flip_sdp); + DML_LOG_VERBOSE("DML::%s: frac_urg_bw_flip_dram = %f\n", __func__, frac_urg_bw_flip_dram); + DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_flip = %f\n", __func__, *frac_urg_bandwidth_flip); + DML_LOG_VERBOSE("DML::%s: flip_bandwidth_support_ok = %d\n", __func__, *flip_bandwidth_support_ok); for (unsigned int m = 0; m < dml2_core_internal_soc_state_max; m++) { for (unsigned int n = 0; n < dml2_core_internal_bw_max; n++) { - dml2_printf("DML::%s: state:%s bw_type:%s, urg_bandwidth_available=%f %s urg_bandwidth_required=%f\n", + DML_LOG_VERBOSE("DML::%s: state:%s bw_type:%s, urg_bandwidth_available=%f %s urg_bandwidth_required=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), urg_bandwidth_available[m][n], (urg_bandwidth_available[m][n] < urg_bandwidth_required_flip[m][n]) ? "<" : ">=", urg_bandwidth_required_flip[m][n]); } @@ -6549,27 +6481,27 @@ static void CalculateFlipSchedule( l->dpte_row_bytes = DPTEBytesPerRow; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable); - dml2_printf("DML::%s: ip.max_flip_time_us = %d\n", __func__, max_flip_time_us); - dml2_printf("DML::%s: ip.max_flip_time_lines = %d\n", __func__, max_flip_time_lines); - dml2_printf("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip); - dml2_printf("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes); - dml2_printf("DML::%s: use_lb_flip_bw = %u\n", __func__, use_lb_flip_bw); - dml2_printf("DML::%s: iflip_enable = %u\n", __func__, iflip_enable); - dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor); - dml2_printf("DML::%s: LineTime = %f\n", __func__, LineTime); - dml2_printf("DML::%s: Tno_bw_flip = %f\n", __func__, Tno_bw_flip); - dml2_printf("DML::%s: Tvm_trips_flip = %f\n", __func__, Tvm_trips_flip); - dml2_printf("DML::%s: Tr0_trips_flip = %f\n", __func__, Tr0_trips_flip); - dml2_printf("DML::%s: Tvm_trips_flip_rounded = %f\n", __func__, Tvm_trips_flip_rounded); - dml2_printf("DML::%s: Tr0_trips_flip_rounded = %f\n", __func__, Tr0_trips_flip_rounded); - dml2_printf("DML::%s: vm_bytes = %f\n", __func__, vm_bytes); - dml2_printf("DML::%s: DPTEBytesPerRow = %f\n", __func__, DPTEBytesPerRow); - dml2_printf("DML::%s: meta_row_bytes = %d\n", __func__, meta_row_bytes); - dml2_printf("DML::%s: dpte_row_bytes = %f\n", __func__, l->dpte_row_bytes); - dml2_printf("DML::%s: dpte_row_height = %d\n", __func__, dpte_row_height); - dml2_printf("DML::%s: meta_row_height = %d\n", __func__, meta_row_height); - dml2_printf("DML::%s: VRatio = %f\n", __func__, VRatio); + DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable); + DML_LOG_VERBOSE("DML::%s: ip.max_flip_time_us = %d\n", __func__, max_flip_time_us); + DML_LOG_VERBOSE("DML::%s: ip.max_flip_time_lines = %d\n", __func__, max_flip_time_lines); + DML_LOG_VERBOSE("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip); + DML_LOG_VERBOSE("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes); + DML_LOG_VERBOSE("DML::%s: use_lb_flip_bw = %u\n", __func__, use_lb_flip_bw); + DML_LOG_VERBOSE("DML::%s: iflip_enable = %u\n", __func__, iflip_enable); + DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, LineTime); + DML_LOG_VERBOSE("DML::%s: Tno_bw_flip = %f\n", __func__, Tno_bw_flip); + DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip = %f\n", __func__, Tvm_trips_flip); + DML_LOG_VERBOSE("DML::%s: Tr0_trips_flip = %f\n", __func__, Tr0_trips_flip); + DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip_rounded = %f\n", __func__, Tvm_trips_flip_rounded); + DML_LOG_VERBOSE("DML::%s: Tr0_trips_flip_rounded = %f\n", __func__, Tr0_trips_flip_rounded); + DML_LOG_VERBOSE("DML::%s: vm_bytes = %f\n", __func__, vm_bytes); + DML_LOG_VERBOSE("DML::%s: DPTEBytesPerRow = %f\n", __func__, DPTEBytesPerRow); + DML_LOG_VERBOSE("DML::%s: meta_row_bytes = %d\n", __func__, meta_row_bytes); + DML_LOG_VERBOSE("DML::%s: dpte_row_bytes = %f\n", __func__, l->dpte_row_bytes); + DML_LOG_VERBOSE("DML::%s: dpte_row_height = %d\n", __func__, dpte_row_height); + DML_LOG_VERBOSE("DML::%s: meta_row_height = %d\n", __func__, meta_row_height); + DML_LOG_VERBOSE("DML::%s: VRatio = %f\n", __func__, VRatio); #endif if (TotImmediateFlipBytes > 0 && (GPUVMEnable || dcc_mrq_enable)) { @@ -6596,9 +6528,9 @@ static void CalculateFlipSchedule( l->min_row_time = l->min_row_height * LineTime / VRatio; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: min_row_time = %f\n", __func__, l->min_row_time); + DML_LOG_VERBOSE("DML::%s: min_row_time = %f\n", __func__, l->min_row_time); #endif - DML2_ASSERT(l->min_row_time > 0); + DML_ASSERT(l->min_row_time > 0); if (use_lb_flip_bw) { // For mode check, calculation the flip bw requirement with worst case flip time @@ -6619,20 +6551,20 @@ static void CalculateFlipSchedule( l->hvm_scaled_vm_bytes / (l->max_flip_time - Tno_bw_flip - 2 * Tr0_trips_flip_rounded), l->hvm_scaled_row_bytes / (l->max_flip_time - Tvm_trips_flip_rounded)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: max_flip_time = %f\n", __func__, l->max_flip_time); - dml2_printf("DML::%s: total vm bytes (hvm ineff scaled) = %f\n", __func__, l->hvm_scaled_vm_bytes); - dml2_printf("DML::%s: total row bytes (%d row, hvm ineff scaled) = %f\n", __func__, l->num_rows, l->hvm_scaled_row_bytes); - dml2_printf("DML::%s: total vm+row bytes (hvm ineff scaled) = %f\n", __func__, l->hvm_scaled_vm_row_bytes); - dml2_printf("DML::%s: lb_flip_bw for vm and row = %f\n", __func__, l->hvm_scaled_vm_row_bytes / (l->max_flip_time - Tno_bw_flip)); - dml2_printf("DML::%s: lb_flip_bw for vm = %f\n", __func__, l->hvm_scaled_vm_bytes / (l->max_flip_time - Tno_bw_flip - 2 * Tr0_trips_flip_rounded)); - dml2_printf("DML::%s: lb_flip_bw for row = %f\n", __func__, l->hvm_scaled_row_bytes / (l->max_flip_time - Tvm_trips_flip_rounded)); + DML_LOG_VERBOSE("DML::%s: max_flip_time = %f\n", __func__, l->max_flip_time); + DML_LOG_VERBOSE("DML::%s: total vm bytes (hvm ineff scaled) = %f\n", __func__, l->hvm_scaled_vm_bytes); + DML_LOG_VERBOSE("DML::%s: total row bytes (%f row, hvm ineff scaled) = %f\n", __func__, l->num_rows, l->hvm_scaled_row_bytes); + DML_LOG_VERBOSE("DML::%s: total vm+row bytes (hvm ineff scaled) = %f\n", __func__, l->hvm_scaled_vm_row_bytes); + DML_LOG_VERBOSE("DML::%s: lb_flip_bw for vm and row = %f\n", __func__, l->hvm_scaled_vm_row_bytes / (l->max_flip_time - Tno_bw_flip)); + DML_LOG_VERBOSE("DML::%s: lb_flip_bw for vm = %f\n", __func__, l->hvm_scaled_vm_bytes / (l->max_flip_time - Tno_bw_flip - 2 * Tr0_trips_flip_rounded)); + DML_LOG_VERBOSE("DML::%s: lb_flip_bw for row = %f\n", __func__, l->hvm_scaled_row_bytes / (l->max_flip_time - Tvm_trips_flip_rounded)); if (l->lb_flip_bw > 0) { - dml2_printf("DML::%s: mode_support est Tvm_flip = %f (bw-based)\n", __func__, Tno_bw_flip + l->hvm_scaled_vm_bytes / l->lb_flip_bw); - dml2_printf("DML::%s: mode_support est Tr0_flip = %f (bw-based)\n", __func__, l->hvm_scaled_row_bytes / l->lb_flip_bw / l->num_rows); - dml2_printf("DML::%s: mode_support est dst_y_per_vm_flip = %f (bw-based)\n", __func__, Tno_bw_flip + l->hvm_scaled_vm_bytes / l->lb_flip_bw / LineTime); - dml2_printf("DML::%s: mode_support est dst_y_per_row_flip = %f (bw-based)\n", __func__, l->hvm_scaled_row_bytes / l->lb_flip_bw / LineTime / l->num_rows); - dml2_printf("DML::%s: Tvm_trips_flip_rounded + 2*Tr0_trips_flip_rounded = %f\n", __func__, (Tvm_trips_flip_rounded + 2 * Tr0_trips_flip_rounded)); + DML_LOG_VERBOSE("DML::%s: mode_support est Tvm_flip = %f (bw-based)\n", __func__, Tno_bw_flip + l->hvm_scaled_vm_bytes / l->lb_flip_bw); + DML_LOG_VERBOSE("DML::%s: mode_support est Tr0_flip = %f (bw-based)\n", __func__, l->hvm_scaled_row_bytes / l->lb_flip_bw / l->num_rows); + DML_LOG_VERBOSE("DML::%s: mode_support est dst_y_per_vm_flip = %f (bw-based)\n", __func__, Tno_bw_flip + l->hvm_scaled_vm_bytes / l->lb_flip_bw / LineTime); + DML_LOG_VERBOSE("DML::%s: mode_support est dst_y_per_row_flip = %f (bw-based)\n", __func__, l->hvm_scaled_row_bytes / l->lb_flip_bw / LineTime / l->num_rows); + DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip_rounded + 2*Tr0_trips_flip_rounded = %f\n", __func__, (Tvm_trips_flip_rounded + 2 * Tr0_trips_flip_rounded)); } #endif l->lb_flip_bw = math_max3(l->lb_flip_bw, @@ -6640,8 +6572,8 @@ static void CalculateFlipSchedule( (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / (15 * LineTime)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: lb_flip_bw for vm reg limit = %f\n", __func__, l->hvm_scaled_vm_bytes / (31 * LineTime) - Tno_bw_flip); - dml2_printf("DML::%s: lb_flip_bw for row reg limit = %f\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / (15 * LineTime)); + DML_LOG_VERBOSE("DML::%s: lb_flip_bw for vm reg limit = %f\n", __func__, l->hvm_scaled_vm_bytes / (31 * LineTime) - Tno_bw_flip); + DML_LOG_VERBOSE("DML::%s: lb_flip_bw for row reg limit = %f\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / (15 * LineTime)); #endif } @@ -6653,13 +6585,12 @@ static void CalculateFlipSchedule( } else { if (iflip_enable) { l->ImmediateFlipBW = (double)per_pipe_flip_bytes * BandwidthAvailableForImmediateFlip / (double)TotImmediateFlipBytes; // flip_bw(i) - double portion = (double)per_pipe_flip_bytes / (double)TotImmediateFlipBytes; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: per_pipe_flip_bytes = %d\n", __func__, per_pipe_flip_bytes); - dml2_printf("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip); - dml2_printf("DML::%s: ImmediateFlipBW = %f\n", __func__, l->ImmediateFlipBW); - dml2_printf("DML::%s: portion of flip bw = %f\n", __func__, portion); + DML_LOG_VERBOSE("DML::%s: per_pipe_flip_bytes = %d\n", __func__, per_pipe_flip_bytes); + DML_LOG_VERBOSE("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip); + DML_LOG_VERBOSE("DML::%s: ImmediateFlipBW = %f\n", __func__, l->ImmediateFlipBW); + DML_LOG_VERBOSE("DML::%s: portion of flip bw = %f\n", __func__, (double)per_pipe_flip_bytes / (double)TotImmediateFlipBytes); #endif if (l->ImmediateFlipBW == 0) { l->Tvm_flip = 0; @@ -6674,11 +6605,11 @@ static void CalculateFlipSchedule( LineTime / 4.0); } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: total vm bytes (hvm ineff scaled) = %f\n", __func__, vm_bytes * HostVMInefficiencyFactor); - dml2_printf("DML::%s: total row bytes (hvm ineff scaled, one row) = %f\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes)); + DML_LOG_VERBOSE("DML::%s: total vm bytes (hvm ineff scaled) = %f\n", __func__, vm_bytes * HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: total row bytes (hvm ineff scaled, one row) = %f\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes)); - dml2_printf("DML::%s: Tvm_flip = %f (bw-based), Tvm_trips_flip = %f (latency-based)\n", __func__, Tno_bw_flip + vm_bytes * HostVMInefficiencyFactor / l->ImmediateFlipBW, Tvm_trips_flip); - dml2_printf("DML::%s: Tr0_flip = %f (bw-based), Tr0_trips_flip = %f (latency-based)\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / l->ImmediateFlipBW, Tr0_trips_flip); + DML_LOG_VERBOSE("DML::%s: Tvm_flip = %f (bw-based), Tvm_trips_flip = %f (latency-based)\n", __func__, Tno_bw_flip + vm_bytes * HostVMInefficiencyFactor / l->ImmediateFlipBW, Tvm_trips_flip); + DML_LOG_VERBOSE("DML::%s: Tr0_flip = %f (bw-based), Tr0_trips_flip = %f (latency-based)\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / l->ImmediateFlipBW, Tr0_trips_flip); #endif *dst_y_per_vm_flip = math_ceil2(4.0 * (l->Tvm_flip / LineTime), 1.0) / 4.0; *dst_y_per_row_flip = math_ceil2(4.0 * (l->Tr0_flip / LineTime), 1.0) / 4.0; @@ -6711,14 +6642,14 @@ static void CalculateFlipSchedule( #ifdef __DML_VBA_DEBUG__ if (!use_lb_flip_bw) { - dml2_printf("DML::%s: dst_y_per_vm_flip = %f (should be < 32)\n", __func__, *dst_y_per_vm_flip); - dml2_printf("DML::%s: dst_y_per_row_flip = %f (should be < 16)\n", __func__, *dst_y_per_row_flip); - dml2_printf("DML::%s: Tvm_flip = %f (final)\n", __func__, l->Tvm_flip); - dml2_printf("DML::%s: Tr0_flip = %f (final)\n", __func__, l->Tr0_flip); - dml2_printf("DML::%s: Tvm_flip + 2*Tr0_flip = %f (should be <= min_row_time=%f)\n", __func__, l->Tvm_flip + 2 * l->Tr0_flip, l->min_row_time); + DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_flip = %f (should be < 32)\n", __func__, *dst_y_per_vm_flip); + DML_LOG_VERBOSE("DML::%s: dst_y_per_row_flip = %f (should be < 16)\n", __func__, *dst_y_per_row_flip); + DML_LOG_VERBOSE("DML::%s: Tvm_flip = %f (final)\n", __func__, l->Tvm_flip); + DML_LOG_VERBOSE("DML::%s: Tr0_flip = %f (final)\n", __func__, l->Tr0_flip); + DML_LOG_VERBOSE("DML::%s: Tvm_flip + 2*Tr0_flip = %f (should be <= min_row_time=%f)\n", __func__, l->Tvm_flip + 2 * l->Tr0_flip, l->min_row_time); } - dml2_printf("DML::%s: final_flip_bw = %f\n", __func__, *final_flip_bw); - dml2_printf("DML::%s: ImmediateFlipSupportedForPipe = %u\n", __func__, *ImmediateFlipSupportedForPipe); + DML_LOG_VERBOSE("DML::%s: final_flip_bw = %f\n", __func__, *final_flip_bw); + DML_LOG_VERBOSE("DML::%s: ImmediateFlipSupportedForPipe = %u\n", __func__, *ImmediateFlipSupportedForPipe); #endif } @@ -6736,7 +6667,7 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( p->Watermark->UrgentWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark); + DML_LOG_VERBOSE("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark); #endif p->Watermark->USRRetrainingWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency + p->mmSOCParameters.USRRetrainingLatency + p->mmSOCParameters.SMNLatency; @@ -6755,20 +6686,20 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( p->Watermark->temp_read_or_ppt_watermark_us = p->mmSOCParameters.g6_temp_read_blackout_us + p->Watermark->UrgentWatermark; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: UrgentLatency = %f\n", __func__, p->mmSOCParameters.UrgentLatency); - dml2_printf("DML::%s: ExtraLatency = %f\n", __func__, p->mmSOCParameters.ExtraLatency); - dml2_printf("DML::%s: DRAMClockChangeLatency = %f\n", __func__, p->mmSOCParameters.DRAMClockChangeLatency); - dml2_printf("DML::%s: SREnterPlusExitZ8Time = %f\n", __func__, p->mmSOCParameters.SREnterPlusExitZ8Time); - dml2_printf("DML::%s: SREnterPlusExitTime = %f\n", __func__, p->mmSOCParameters.SREnterPlusExitTime); - dml2_printf("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark); - dml2_printf("DML::%s: USRRetrainingWatermark = %f\n", __func__, p->Watermark->USRRetrainingWatermark); - dml2_printf("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, p->Watermark->DRAMClockChangeWatermark); - dml2_printf("DML::%s: FCLKChangeWatermark = %f\n", __func__, p->Watermark->FCLKChangeWatermark); - dml2_printf("DML::%s: StutterExitWatermark = %f\n", __func__, p->Watermark->StutterExitWatermark); - dml2_printf("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->StutterEnterPlusExitWatermark); - dml2_printf("DML::%s: Z8StutterExitWatermark = %f\n", __func__, p->Watermark->Z8StutterExitWatermark); - dml2_printf("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->Z8StutterEnterPlusExitWatermark); - dml2_printf("DML::%s: temp_read_or_ppt_watermark_us = %f\n", __func__, p->Watermark->temp_read_or_ppt_watermark_us); + DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, p->mmSOCParameters.UrgentLatency); + DML_LOG_VERBOSE("DML::%s: ExtraLatency = %f\n", __func__, p->mmSOCParameters.ExtraLatency); + DML_LOG_VERBOSE("DML::%s: DRAMClockChangeLatency = %f\n", __func__, p->mmSOCParameters.DRAMClockChangeLatency); + DML_LOG_VERBOSE("DML::%s: SREnterPlusExitZ8Time = %f\n", __func__, p->mmSOCParameters.SREnterPlusExitZ8Time); + DML_LOG_VERBOSE("DML::%s: SREnterPlusExitTime = %f\n", __func__, p->mmSOCParameters.SREnterPlusExitTime); + DML_LOG_VERBOSE("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark); + DML_LOG_VERBOSE("DML::%s: USRRetrainingWatermark = %f\n", __func__, p->Watermark->USRRetrainingWatermark); + DML_LOG_VERBOSE("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, p->Watermark->DRAMClockChangeWatermark); + DML_LOG_VERBOSE("DML::%s: FCLKChangeWatermark = %f\n", __func__, p->Watermark->FCLKChangeWatermark); + DML_LOG_VERBOSE("DML::%s: StutterExitWatermark = %f\n", __func__, p->Watermark->StutterExitWatermark); + DML_LOG_VERBOSE("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->StutterEnterPlusExitWatermark); + DML_LOG_VERBOSE("DML::%s: Z8StutterExitWatermark = %f\n", __func__, p->Watermark->Z8StutterExitWatermark); + DML_LOG_VERBOSE("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->Z8StutterEnterPlusExitWatermark); + DML_LOG_VERBOSE("DML::%s: temp_read_or_ppt_watermark_us = %f\n", __func__, p->Watermark->temp_read_or_ppt_watermark_us); #endif s->TotalActiveWriteback = 0; @@ -6801,11 +6732,11 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( p->Watermark->WritebackFCLKChangeWatermark = p->Watermark->WritebackFCLKChangeWatermark + p->mmSOCParameters.USRRetrainingLatency; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: WritebackDRAMClockChangeWatermark = %f\n", __func__, p->Watermark->WritebackDRAMClockChangeWatermark); - dml2_printf("DML::%s: WritebackFCLKChangeWatermark = %f\n", __func__, p->Watermark->WritebackFCLKChangeWatermark); - dml2_printf("DML::%s: WritebackUrgentWatermark = %f\n", __func__, p->Watermark->WritebackUrgentWatermark); - dml2_printf("DML::%s: USRRetrainingRequired = %u\n", __func__, p->USRRetrainingRequired); - dml2_printf("DML::%s: USRRetrainingLatency = %f\n", __func__, p->mmSOCParameters.USRRetrainingLatency); + DML_LOG_VERBOSE("DML::%s: WritebackDRAMClockChangeWatermark = %f\n", __func__, p->Watermark->WritebackDRAMClockChangeWatermark); + DML_LOG_VERBOSE("DML::%s: WritebackFCLKChangeWatermark = %f\n", __func__, p->Watermark->WritebackFCLKChangeWatermark); + DML_LOG_VERBOSE("DML::%s: WritebackUrgentWatermark = %f\n", __func__, p->Watermark->WritebackUrgentWatermark); + DML_LOG_VERBOSE("DML::%s: USRRetrainingRequired = %u\n", __func__, p->USRRetrainingRequired); + DML_LOG_VERBOSE("DML::%s: USRRetrainingLatency = %f\n", __func__, p->mmSOCParameters.USRRetrainingLatency); #endif s->TotalPixelBW = 0.0; @@ -6836,11 +6767,11 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( s->LBLatencyHidingSourceLinesC[k] = (unsigned int)(math_min2((double)p->MaxLineBufferLines, math_floor2((double)p->LineBufferSize / LBBitPerPixel / ((double)p->SwathWidthC[k] / math_max2(h_ratio_c, 1.0)), 1)) - (v_taps_c - 1)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, MaxLineBufferLines = %u\n", __func__, k, p->MaxLineBufferLines); - dml2_printf("DML::%s: k=%u, LineBufferSize = %u\n", __func__, k, p->LineBufferSize); - dml2_printf("DML::%s: k=%u, LBBitPerPixel = %u\n", __func__, k, LBBitPerPixel); - dml2_printf("DML::%s: k=%u, HRatio = %f\n", __func__, k, h_ratio); - dml2_printf("DML::%s: k=%u, VTaps = %f\n", __func__, k, v_taps); + DML_LOG_VERBOSE("DML::%s: k=%u, MaxLineBufferLines = %u\n", __func__, k, p->MaxLineBufferLines); + DML_LOG_VERBOSE("DML::%s: k=%u, LineBufferSize = %u\n", __func__, k, p->LineBufferSize); + DML_LOG_VERBOSE("DML::%s: k=%u, LBBitPerPixel = %f\n", __func__, k, LBBitPerPixel); + DML_LOG_VERBOSE("DML::%s: k=%u, HRatio = %f\n", __func__, k, h_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u, VTaps = %f\n", __func__, k, v_taps); #endif s->EffectiveLBLatencyHidingY = s->LBLatencyHidingSourceLinesY[k] / v_ratio * (h_total / pixel_clock_mhz); @@ -6943,16 +6874,16 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( s->sub_vp_lines_l = s->src_y_pstate_l + s->src_y_ahead_l + p->meta_row_height_l[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]); - dml2_printf("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]); - dml2_printf("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]); - dml2_printf("DML::%s: k=%u, SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]); - dml2_printf("DML::%s: k=%u, LBLatencyHidingSourceLinesY = %u\n", __func__, k, s->LBLatencyHidingSourceLinesY[k]); - dml2_printf("DML::%s: k=%u, dst_y_pstate = %u\n", __func__, k, s->dst_y_pstate); - dml2_printf("DML::%s: k=%u, src_y_pstate_l = %u\n", __func__, k, s->src_y_pstate_l); - dml2_printf("DML::%s: k=%u, src_y_ahead_l = %u\n", __func__, k, s->src_y_ahead_l); - dml2_printf("DML::%s: k=%u, meta_row_height_l = %u\n", __func__, k, p->meta_row_height_l[k]); - dml2_printf("DML::%s: k=%u, sub_vp_lines_l = %u\n", __func__, k, s->sub_vp_lines_l); + DML_LOG_VERBOSE("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, LBLatencyHidingSourceLinesY = %u\n", __func__, k, s->LBLatencyHidingSourceLinesY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_pstate = %u\n", __func__, k, s->dst_y_pstate); + DML_LOG_VERBOSE("DML::%s: k=%u, src_y_pstate_l = %u\n", __func__, k, s->src_y_pstate_l); + DML_LOG_VERBOSE("DML::%s: k=%u, src_y_ahead_l = %u\n", __func__, k, s->src_y_ahead_l); + DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_height_l = %u\n", __func__, k, p->meta_row_height_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, sub_vp_lines_l = %u\n", __func__, k, s->sub_vp_lines_l); #endif p->SubViewportLinesNeededInMALL[k] = s->sub_vp_lines_l; @@ -6967,10 +6898,10 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( p->SubViewportLinesNeededInMALL[k] = (unsigned int)(math_max2(s->sub_vp_lines_l, s->sub_vp_lines_c)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, meta_row_height_c = %u\n", __func__, k, p->meta_row_height_c[k]); - dml2_printf("DML::%s: k=%u, src_y_pstate_c = %u\n", __func__, k, s->src_y_pstate_c); - dml2_printf("DML::%s: k=%u, src_y_ahead_c = %u\n", __func__, k, s->src_y_ahead_c); - dml2_printf("DML::%s: k=%u, sub_vp_lines_c = %u\n", __func__, k, s->sub_vp_lines_c); + DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_height_c = %u\n", __func__, k, p->meta_row_height_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, src_y_pstate_c = %u\n", __func__, k, s->src_y_pstate_c); + DML_LOG_VERBOSE("DML::%s: k=%u, src_y_ahead_c = %u\n", __func__, k, s->src_y_ahead_c); + DML_LOG_VERBOSE("DML::%s: k=%u, sub_vp_lines_c = %u\n", __func__, k, s->sub_vp_lines_c); #endif } } @@ -6992,10 +6923,10 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DRAMClockChangeSupport = %u\n", __func__, *p->global_dram_clock_change_supported); - dml2_printf("DML::%s: FCLKChangeSupport = %u\n", __func__, *p->global_fclk_change_supported); - dml2_printf("DML::%s: MaxActiveFCLKChangeLatencySupported = %f\n", __func__, *p->MaxActiveFCLKChangeLatencySupported); - dml2_printf("DML::%s: USRRetrainingSupport = %u\n", __func__, *p->USRRetrainingSupport); + DML_LOG_VERBOSE("DML::%s: DRAMClockChangeSupport = %u\n", __func__, *p->global_dram_clock_change_supported); + DML_LOG_VERBOSE("DML::%s: FCLKChangeSupport = %u\n", __func__, *p->global_fclk_change_supported); + DML_LOG_VERBOSE("DML::%s: MaxActiveFCLKChangeLatencySupported = %f\n", __func__, *p->MaxActiveFCLKChangeLatencySupported); + DML_LOG_VERBOSE("DML::%s: USRRetrainingSupport = %u\n", __func__, *p->USRRetrainingSupport); #endif } @@ -7141,7 +7072,7 @@ static unsigned int get_qos_param_index(unsigned long uclk_freq_khz, const struc unsigned int index = 0; for (i = 0; i < DML_MAX_CLK_TABLE_SIZE; i++) { - dml2_printf("DML::%s: per_uclk_dpm_params[%d].minimum_uclk_khz = %d\n", __func__, i, per_uclk_dpm_params[i].minimum_uclk_khz); + DML_LOG_VERBOSE("DML::%s: per_uclk_dpm_params[%d].minimum_uclk_khz = %ld\n", __func__, i, per_uclk_dpm_params[i].minimum_uclk_khz); if (i == 0) index = 0; @@ -7153,32 +7084,30 @@ static unsigned int get_qos_param_index(unsigned long uclk_freq_khz, const struc break; } } -#if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: uclk_freq_khz = %d\n", __func__, uclk_freq_khz); - dml2_printf("DML::%s: index = %d\n", __func__, index); -#endif + DML_LOG_VERBOSE("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz); + DML_LOG_VERBOSE("DML::%s: index = %d\n", __func__, index); return index; } static unsigned int get_active_min_uclk_dpm_index(unsigned long uclk_freq_khz, const struct dml2_soc_state_table *clk_table) { unsigned int i; - bool clk_entry_found = 0; + bool clk_entry_found = false; for (i = 0; i < clk_table->uclk.num_clk_values; i++) { - dml2_printf("DML::%s: clk_table.uclk.clk_values_khz[%d] = %d\n", __func__, i, clk_table->uclk.clk_values_khz[i]); + DML_LOG_VERBOSE("DML::%s: clk_table.uclk.clk_values_khz[%d] = %ld\n", __func__, i, clk_table->uclk.clk_values_khz[i]); if (uclk_freq_khz == clk_table->uclk.clk_values_khz[i]) { - clk_entry_found = 1; + clk_entry_found = true; break; } } if (!clk_entry_found) - DML2_ASSERT(clk_entry_found); + DML_ASSERT(clk_entry_found); #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz); - dml2_printf("DML::%s: index = %d\n", __func__, i); + DML_LOG_VERBOSE("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz); + DML_LOG_VERBOSE("DML::%s: index = %d\n", __func__, i); #endif return i; } @@ -7218,10 +7147,10 @@ static void calculate_hostvm_inefficiency_factor( if ((*HostVMInefficiencyFactorPrefetch < 4) && (remote_iommu_outstanding_translations < max_outstanding_reqs)) *HostVMInefficiencyFactorPrefetch = 4; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: urg_bandwidth_avail_active_pixel_and_vm = %f\n", __func__, urg_bandwidth_avail_active_pixel_and_vm); - dml2_printf("DML::%s: urg_bandwidth_avail_active_vm_only = %f\n", __func__, urg_bandwidth_avail_active_vm_only); - dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, *HostVMInefficiencyFactor); - dml2_printf("DML::%s: HostVMInefficiencyFactorPrefetch = %f\n", __func__, *HostVMInefficiencyFactorPrefetch); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_avail_active_pixel_and_vm = %f\n", __func__, urg_bandwidth_avail_active_pixel_and_vm); + DML_LOG_VERBOSE("DML::%s: urg_bandwidth_avail_active_vm_only = %f\n", __func__, urg_bandwidth_avail_active_vm_only); + DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, *HostVMInefficiencyFactor); + DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactorPrefetch = %f\n", __func__, *HostVMInefficiencyFactorPrefetch); #endif } } @@ -7335,30 +7264,659 @@ static void calculate_pstate_keepout_dst_lines( } } +static noinline_for_stack void dml_core_ms_prefetch_check(struct dml2_core_internal_display_mode_lib *mode_lib, + const struct dml2_display_cfg *display_cfg) +{ + struct dml2_core_calcs_mode_support_locals *s = &mode_lib->scratch.dml_core_mode_support_locals; + struct dml2_core_calcs_calculate_tdlut_setting_params *calculate_tdlut_setting_params = &mode_lib->scratch.calculate_tdlut_setting_params; + struct dml2_core_calcs_CalculatePrefetchSchedule_params *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params; + struct dml2_core_calcs_calculate_peak_bandwidth_required_params *calculate_peak_bandwidth_params = &mode_lib->scratch.calculate_peak_bandwidth_params; +#ifdef DML_GLOBAL_PREFETCH_CHECK + struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *CheckGlobalPrefetchAdmissibility_params = &mode_lib->scratch.CheckGlobalPrefetchAdmissibility_params; +#endif + struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params; + + double min_return_bw_for_latency; + unsigned int k; + + mode_lib->ms.TimeCalc = 24 / mode_lib->ms.dcfclk_deepsleep; + + calculate_hostvm_inefficiency_factor( + &s->HostVMInefficiencyFactor, + &s->HostVMInefficiencyFactorPrefetch, + + display_cfg->gpuvm_enable, + display_cfg->hostvm_enable, + mode_lib->ip.remote_iommu_outstanding_translations, + mode_lib->soc.max_outstanding_reqs, + mode_lib->ms.support.urg_bandwidth_available_pixel_and_vm[dml2_core_internal_soc_state_sys_active], + mode_lib->ms.support.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_sys_active]); + + mode_lib->ms.Total3dlutActive = 0; + for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) { + if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut) + mode_lib->ms.Total3dlutActive = mode_lib->ms.Total3dlutActive + 1; + + // Calculate tdlut schedule related terms + calculate_tdlut_setting_params->dispclk_mhz = mode_lib->ms.RequiredDISPCLK; + calculate_tdlut_setting_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut; + calculate_tdlut_setting_params->tdlut_width_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_width_mode; + calculate_tdlut_setting_params->tdlut_addressing_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_addressing_mode; + calculate_tdlut_setting_params->cursor_buffer_size = mode_lib->ip.cursor_buffer_size; + calculate_tdlut_setting_params->gpuvm_enable = display_cfg->gpuvm_enable; + calculate_tdlut_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes; + calculate_tdlut_setting_params->tdlut_mpc_width_flag = display_cfg->plane_descriptors[k].tdlut.tdlut_mpc_width_flag; + calculate_tdlut_setting_params->is_gfx11 = dml_get_gfx_version(display_cfg->plane_descriptors[k].surface.tiling); + + // output + calculate_tdlut_setting_params->tdlut_pte_bytes_per_frame = &s->tdlut_pte_bytes_per_frame[k]; + calculate_tdlut_setting_params->tdlut_bytes_per_frame = &s->tdlut_bytes_per_frame[k]; + calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k]; + calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k]; + calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k]; + calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k]; + calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k]; + + calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params); + } + + min_return_bw_for_latency = mode_lib->ms.support.urg_bandwidth_available_min_latency[dml2_core_internal_soc_state_sys_active]; + + if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn3) + s->ReorderingBytes = (unsigned int)(mode_lib->soc.clk_table.dram_config.channel_count * math_max3(mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_only_bytes, + mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes, + mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_vm_only_bytes)); + + CalculateExtraLatency( + display_cfg, + mode_lib->ip.rob_buffer_size_kbytes, + mode_lib->soc.qos_parameters.qos_params.dcn32x.loaded_round_trip_latency_fclk_cycles, + s->ReorderingBytes, + mode_lib->ms.DCFCLK, + mode_lib->ms.FabricClock, + mode_lib->ip.pixel_chunk_size_kbytes, + min_return_bw_for_latency, + mode_lib->ms.num_active_planes, + mode_lib->ms.NoOfDPP, + mode_lib->ms.dpte_group_bytes, + s->tdlut_bytes_per_group, + s->HostVMInefficiencyFactor, + s->HostVMInefficiencyFactorPrefetch, + mode_lib->soc.hostvm_min_page_size_kbytes, + mode_lib->soc.qos_parameters.qos_type, + !(display_cfg->overrides.max_outstanding_when_urgent_expected_disable), + mode_lib->soc.max_outstanding_reqs, + mode_lib->ms.support.request_size_bytes_luma, + mode_lib->ms.support.request_size_bytes_chroma, + mode_lib->ip.meta_chunk_size_kbytes, + mode_lib->ip.dchub_arb_to_ret_delay, + mode_lib->ms.TripToMemory, + mode_lib->ip.hostvm_mode, + + // output + &mode_lib->ms.ExtraLatency, + &mode_lib->ms.ExtraLatency_sr, + &mode_lib->ms.ExtraLatencyPrefetch); + + for (k = 0; k < mode_lib->ms.num_active_planes; k++) + s->impacted_dst_y_pre[k] = 0; + + s->recalc_prefetch_schedule = 0; + s->recalc_prefetch_done = 0; + do { + mode_lib->ms.support.PrefetchSupported = true; + + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); + s->pixel_format[k] = display_cfg->plane_descriptors[k].pixel_format; + + s->lb_source_lines_l[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits, + mode_lib->ms.NoOfDPP[k], + display_cfg->plane_descriptors[k].composition.viewport.plane0.width, + display_cfg->plane_descriptors[k].composition.viewport.plane0.height, + display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio, + display_cfg->plane_descriptors[k].composition.rotation_angle); + + s->lb_source_lines_c[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits, + mode_lib->ms.NoOfDPP[k], + display_cfg->plane_descriptors[k].composition.viewport.plane1.width, + display_cfg->plane_descriptors[k].composition.viewport.plane1.height, + display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio, + display_cfg->plane_descriptors[k].composition.rotation_angle); + + struct dml2_core_internal_DmlPipe *myPipe = &s->myPipe; + + mode_lib->ms.TWait[k] = CalculateTWait( + display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns, + mode_lib->ms.UrgLatency, + mode_lib->ms.TripToMemory, + !dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.drr_config.enabled ? + get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), mode_lib->ms.state_idx) : 0.0); + + myPipe->Dppclk = mode_lib->ms.RequiredDPPCLK[k]; + myPipe->Dispclk = mode_lib->ms.RequiredDISPCLK; + myPipe->PixelClock = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); + myPipe->DCFClkDeepSleep = mode_lib->ms.dcfclk_deepsleep; + myPipe->DPPPerSurface = mode_lib->ms.NoOfDPP[k]; + myPipe->ScalerEnabled = display_cfg->plane_descriptors[k].composition.scaler_info.enabled; + myPipe->VRatio = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio; + myPipe->VRatioChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio; + myPipe->VTaps = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps; + myPipe->VTapsChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps; + myPipe->RotationAngle = display_cfg->plane_descriptors[k].composition.rotation_angle; + myPipe->mirrored = display_cfg->plane_descriptors[k].composition.mirrored; + myPipe->BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k]; + myPipe->BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k]; + myPipe->BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k]; + myPipe->BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k]; + myPipe->InterlaceEnable = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced; + myPipe->NumberOfCursors = display_cfg->plane_descriptors[k].cursor.num_cursors; + myPipe->VBlank = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total - display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_active; + myPipe->HTotal = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total; + myPipe->HActive = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active; + myPipe->DCCEnable = display_cfg->plane_descriptors[k].surface.dcc.enable; + myPipe->ODMMode = mode_lib->ms.ODMMode[k]; + myPipe->SourcePixelFormat = display_cfg->plane_descriptors[k].pixel_format; + myPipe->BytePerPixelY = mode_lib->ms.BytePerPixelY[k]; + myPipe->BytePerPixelC = mode_lib->ms.BytePerPixelC[k]; + myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ip.ptoi_supported; + +#ifdef __DML_VBA_DEBUG__ + DML_LOG_VERBOSE("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: MaximumVStartup = %u\n", __func__, s->MaximumVStartup[k]); +#endif + CalculatePrefetchSchedule_params->display_cfg = display_cfg; + CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactorPrefetch; + CalculatePrefetchSchedule_params->myPipe = myPipe; + CalculatePrefetchSchedule_params->DSCDelay = mode_lib->ms.DSCDelay[k]; + CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ip.dppclk_delay_subtotal + mode_lib->ip.dppclk_delay_cnvc_formatter; + CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ip.dppclk_delay_scl; + CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ip.dppclk_delay_scl_lb_only; + CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ip.dppclk_delay_cnvc_cursor; + CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ip.dispclk_delay_subtotal; + CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (unsigned int)(mode_lib->ms.SwathWidthY[k] / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio); + CalculatePrefetchSchedule_params->OutputFormat = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format; + CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ip.max_inter_dcn_tile_repeaters; + CalculatePrefetchSchedule_params->VStartup = s->MaximumVStartup[k]; + CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->soc.hostvm_min_page_size_kbytes; + CalculatePrefetchSchedule_params->DynamicMetadataEnable = display_cfg->plane_descriptors[k].dynamic_meta_data.enable; + CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ip.dynamic_metadata_vm_enabled; + CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = display_cfg->plane_descriptors[k].dynamic_meta_data.lines_before_active_required; + CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = display_cfg->plane_descriptors[k].dynamic_meta_data.transmitted_bytes; + CalculatePrefetchSchedule_params->UrgentLatency = mode_lib->ms.UrgLatency; + CalculatePrefetchSchedule_params->ExtraLatencyPrefetch = mode_lib->ms.ExtraLatencyPrefetch; + CalculatePrefetchSchedule_params->TCalc = mode_lib->ms.TimeCalc; + CalculatePrefetchSchedule_params->vm_bytes = mode_lib->ms.vm_bytes[k]; + CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow[k]; + CalculatePrefetchSchedule_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesY[k]; + CalculatePrefetchSchedule_params->VInitPreFillY = mode_lib->ms.PrefillY[k]; + CalculatePrefetchSchedule_params->MaxNumSwathY = mode_lib->ms.MaxNumSwathY[k]; + CalculatePrefetchSchedule_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesC[k]; + CalculatePrefetchSchedule_params->VInitPreFillC = mode_lib->ms.PrefillC[k]; + CalculatePrefetchSchedule_params->MaxNumSwathC = mode_lib->ms.MaxNumSwathC[k]; + CalculatePrefetchSchedule_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub[k]; + CalculatePrefetchSchedule_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub[k]; + CalculatePrefetchSchedule_params->SwathHeightY = mode_lib->ms.SwathHeightY[k]; + CalculatePrefetchSchedule_params->SwathHeightC = mode_lib->ms.SwathHeightC[k]; + CalculatePrefetchSchedule_params->TWait = mode_lib->ms.TWait[k]; + CalculatePrefetchSchedule_params->Ttrip = mode_lib->ms.TripToMemory; + CalculatePrefetchSchedule_params->Turg = mode_lib->ms.UrgLatency; + CalculatePrefetchSchedule_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut; + CalculatePrefetchSchedule_params->tdlut_pte_bytes_per_frame = s->tdlut_pte_bytes_per_frame[k]; + CalculatePrefetchSchedule_params->tdlut_bytes_per_frame = s->tdlut_bytes_per_frame[k]; + CalculatePrefetchSchedule_params->tdlut_opt_time = s->tdlut_opt_time[k]; + CalculatePrefetchSchedule_params->tdlut_drain_time = s->tdlut_drain_time[k]; + CalculatePrefetchSchedule_params->num_cursors = (display_cfg->plane_descriptors[k].cursor.cursor_width > 0); + CalculatePrefetchSchedule_params->cursor_bytes_per_chunk = s->cursor_bytes_per_chunk[k]; + CalculatePrefetchSchedule_params->cursor_bytes_per_line = s->cursor_bytes_per_line[k]; + CalculatePrefetchSchedule_params->dcc_enable = display_cfg->plane_descriptors[k].surface.dcc.enable; + CalculatePrefetchSchedule_params->mrq_present = mode_lib->ip.dcn_mrq_present; + CalculatePrefetchSchedule_params->meta_row_bytes = mode_lib->ms.meta_row_bytes[k]; + CalculatePrefetchSchedule_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor[k]; + CalculatePrefetchSchedule_params->impacted_dst_y_pre = s->impacted_dst_y_pre[k]; + CalculatePrefetchSchedule_params->vactive_sw_bw_l = mode_lib->ms.vactive_sw_bw_l[k]; + CalculatePrefetchSchedule_params->vactive_sw_bw_c = mode_lib->ms.vactive_sw_bw_c[k]; + + // output + CalculatePrefetchSchedule_params->DSTXAfterScaler = &s->DSTXAfterScaler[k]; + CalculatePrefetchSchedule_params->DSTYAfterScaler = &s->DSTYAfterScaler[k]; + CalculatePrefetchSchedule_params->dst_y_prefetch = &mode_lib->ms.dst_y_prefetch[k]; + CalculatePrefetchSchedule_params->dst_y_per_vm_vblank = &mode_lib->ms.LinesForVM[k]; + CalculatePrefetchSchedule_params->dst_y_per_row_vblank = &mode_lib->ms.LinesForDPTERow[k]; + CalculatePrefetchSchedule_params->VRatioPrefetchY = &mode_lib->ms.VRatioPreY[k]; + CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->ms.VRatioPreC[k]; + CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWLuma = &mode_lib->ms.RequiredPrefetchPixelDataBWLuma[k]; // prefetch_sw_bw_l + CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWChroma = &mode_lib->ms.RequiredPrefetchPixelDataBWChroma[k]; // prefetch_sw_bw_c + CalculatePrefetchSchedule_params->RequiredPrefetchBWOTO = &mode_lib->ms.RequiredPrefetchBWOTO[k]; + CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->ms.NoTimeForDynamicMetadata[k]; + CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->ms.Tno_bw[k]; + CalculatePrefetchSchedule_params->Tno_bw_flip = &mode_lib->ms.Tno_bw_flip[k]; + CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &mode_lib->ms.prefetch_vmrow_bw[k]; + CalculatePrefetchSchedule_params->Tdmdl_vm = &s->dummy_single[0]; + CalculatePrefetchSchedule_params->Tdmdl = &s->dummy_single[1]; + CalculatePrefetchSchedule_params->TSetup = &s->dummy_single[2]; + CalculatePrefetchSchedule_params->Tvm_trips = &s->Tvm_trips[k]; + CalculatePrefetchSchedule_params->Tr0_trips = &s->Tr0_trips[k]; + CalculatePrefetchSchedule_params->Tvm_trips_flip = &s->Tvm_trips_flip[k]; + CalculatePrefetchSchedule_params->Tr0_trips_flip = &s->Tr0_trips_flip[k]; + CalculatePrefetchSchedule_params->Tvm_trips_flip_rounded = &s->Tvm_trips_flip_rounded[k]; + CalculatePrefetchSchedule_params->Tr0_trips_flip_rounded = &s->Tr0_trips_flip_rounded[k]; + CalculatePrefetchSchedule_params->VUpdateOffsetPix = &s->dummy_integer[0]; + CalculatePrefetchSchedule_params->VUpdateWidthPix = &s->dummy_integer[1]; + CalculatePrefetchSchedule_params->VReadyOffsetPix = &s->dummy_integer[2]; + CalculatePrefetchSchedule_params->prefetch_cursor_bw = &mode_lib->ms.prefetch_cursor_bw[k]; + CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k]; + CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k]; + CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k]; + CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->prefetch_swath_time_us[k]; + + mode_lib->ms.NoTimeForPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params); + + mode_lib->ms.support.PrefetchSupported &= !mode_lib->ms.NoTimeForPrefetch[k]; + DML_LOG_VERBOSE("DML::%s: k=%d, dst_y_per_vm_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML::%s: k=%d, dst_y_per_row_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_row_vblank); + } // for k num_planes + + CalculateDCFCLKDeepSleepTdlut( + display_cfg, + mode_lib->ms.num_active_planes, + mode_lib->ms.BytePerPixelY, + mode_lib->ms.BytePerPixelC, + mode_lib->ms.SwathWidthY, + mode_lib->ms.SwathWidthC, + mode_lib->ms.NoOfDPP, + mode_lib->ms.PSCL_FACTOR, + mode_lib->ms.PSCL_FACTOR_CHROMA, + mode_lib->ms.RequiredDPPCLK, + mode_lib->ms.vactive_sw_bw_l, + mode_lib->ms.vactive_sw_bw_c, + mode_lib->soc.return_bus_width_bytes, + mode_lib->ms.RequiredDISPCLK, + s->tdlut_bytes_to_deliver, + s->prefetch_swath_time_us, + + /* Output */ + &mode_lib->ms.dcfclk_deepsleep); + + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + if (mode_lib->ms.dst_y_prefetch[k] < 2.0 + || mode_lib->ms.LinesForVM[k] >= 32.0 + || mode_lib->ms.LinesForDPTERow[k] >= 16.0 + || mode_lib->ms.NoTimeForPrefetch[k] == true + || s->DSTYAfterScaler[k] > 8) { + mode_lib->ms.support.PrefetchSupported = false; + DML_LOG_VERBOSE("DML::%s: k=%d, dst_y_prefetch=%f (should not be < 2)\n", __func__, k, mode_lib->ms.dst_y_prefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, LinesForVM=%f (should not be >= 32)\n", __func__, k, mode_lib->ms.LinesForVM[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, LinesForDPTERow=%f (should not be >= 16)\n", __func__, k, mode_lib->ms.LinesForDPTERow[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, DSTYAfterScaler=%d (should be <= 8)\n", __func__, k, s->DSTYAfterScaler[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, NoTimeForPrefetch=%d\n", __func__, k, mode_lib->ms.NoTimeForPrefetch[k]); + } + } + + mode_lib->ms.support.DynamicMetadataSupported = true; + for (k = 0; k < mode_lib->ms.num_active_planes; ++k) { + if (mode_lib->ms.NoTimeForDynamicMetadata[k] == true) { + mode_lib->ms.support.DynamicMetadataSupported = false; + } + } + + mode_lib->ms.support.VRatioInPrefetchSupported = true; + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + if (mode_lib->ms.VRatioPreY[k] > __DML2_CALCS_MAX_VRATIO_PRE__ || + mode_lib->ms.VRatioPreC[k] > __DML2_CALCS_MAX_VRATIO_PRE__) { + mode_lib->ms.support.VRatioInPrefetchSupported = false; + DML_LOG_VERBOSE("DML::%s: k=%d VRatioPreY = %f (should be <= %f)\n", __func__, k, mode_lib->ms.VRatioPreY[k], __DML2_CALCS_MAX_VRATIO_PRE__); + DML_LOG_VERBOSE("DML::%s: k=%d VRatioPreC = %f (should be <= %f)\n", __func__, k, mode_lib->ms.VRatioPreC[k], __DML2_CALCS_MAX_VRATIO_PRE__); + DML_LOG_VERBOSE("DML::%s: VRatioInPrefetchSupported = %u\n", __func__, mode_lib->ms.support.VRatioInPrefetchSupported); + } + } + + mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.VRatioInPrefetchSupported; + + // By default, do not recalc prefetch schedule + s->recalc_prefetch_schedule = 0; + + // Only do urg vs prefetch bandwidth check, flip schedule check, power saving feature support check IF the Prefetch Schedule Check is ok + if (mode_lib->ms.support.PrefetchSupported) { + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + // Calculate Urgent burst factor for prefetch +#ifdef __DML_VBA_DEBUG__ + DML_LOG_VERBOSE("DML::%s: k=%d, Calling CalculateUrgentBurstFactor (for prefetch)\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%d, VRatioPreY=%f\n", __func__, k, mode_lib->ms.VRatioPreY[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, VRatioPreC=%f\n", __func__, k, mode_lib->ms.VRatioPreC[k]); +#endif + CalculateUrgentBurstFactor( + &display_cfg->plane_descriptors[k], + mode_lib->ms.swath_width_luma_ub[k], + mode_lib->ms.swath_width_chroma_ub[k], + mode_lib->ms.SwathHeightY[k], + mode_lib->ms.SwathHeightC[k], + s->line_times[k], + mode_lib->ms.UrgLatency, + mode_lib->ms.VRatioPreY[k], + mode_lib->ms.VRatioPreC[k], + mode_lib->ms.BytePerPixelInDETY[k], + mode_lib->ms.BytePerPixelInDETC[k], + mode_lib->ms.DETBufferSizeY[k], + mode_lib->ms.DETBufferSizeC[k], + /* Output */ + &mode_lib->ms.UrgentBurstFactorLumaPre[k], + &mode_lib->ms.UrgentBurstFactorChromaPre[k], + &mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]); + } + + // Calculate urgent bandwidth required, both urg and non urg peak bandwidth + // assume flip bw is 0 at this point + for (k = 0; k < mode_lib->ms.num_active_planes; k++) + mode_lib->ms.final_flip_bw[k] = 0; + + calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = mode_lib->ms.support.urg_vactive_bandwidth_required; + calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required; + calculate_peak_bandwidth_params->urg_bandwidth_required_qual = mode_lib->ms.support.urg_bandwidth_required_qual; + calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required; + calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = mode_lib->ms.surface_avg_vactive_required_bw; + calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw; + + calculate_peak_bandwidth_params->display_cfg = display_cfg; + calculate_peak_bandwidth_params->inc_flip_bw = 0; + calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes; + calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP; + calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0; + calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1; + calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0; + calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1; + calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor; + calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor; + + calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l; + calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c; + calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma; + calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma; + calculate_peak_bandwidth_params->prefetch_bandwidth_oto = mode_lib->ms.RequiredPrefetchBWOTO; + calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l; + calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c; + calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw; + calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw; + calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw; + calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw; + calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw; + calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw; + calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma; + calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma; + calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor; + calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre; + calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre; + calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre; + + calculate_peak_bandwidth_required( + &mode_lib->scratch, + calculate_peak_bandwidth_params); + + // Check urg peak bandwidth against available urg bw + // check at SDP and DRAM, for all soc states (SVP prefetch an Sys Active) + check_urgent_bandwidth_support( + &s->dummy_single[0], // double* frac_urg_bandwidth + &s->dummy_single[1], // double* frac_urg_bandwidth_mall + &mode_lib->ms.support.UrgVactiveBandwidthSupport, + &mode_lib->ms.support.PrefetchBandwidthSupported, + + mode_lib->soc.mall_allocated_for_dcn_mbytes, + mode_lib->ms.support.non_urg_bandwidth_required, + mode_lib->ms.support.urg_vactive_bandwidth_required, + mode_lib->ms.support.urg_bandwidth_required, + mode_lib->ms.support.urg_bandwidth_available); + + mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.PrefetchBandwidthSupported; + DML_LOG_VERBOSE("DML::%s: PrefetchBandwidthSupported=%0d\n", __func__, mode_lib->ms.support.PrefetchBandwidthSupported); + + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + if (mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]) { + mode_lib->ms.support.PrefetchSupported = false; + DML_LOG_VERBOSE("DML::%s: k=%d, NotEnoughUrgentLatencyHidingPre=%d\n", __func__, k, mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]); + } + } + +#ifdef DML_GLOBAL_PREFETCH_CHECK + if (mode_lib->ms.support.PrefetchSupported && mode_lib->ms.num_active_planes > 1 && s->recalc_prefetch_done == 0) { + CheckGlobalPrefetchAdmissibility_params->num_active_planes = mode_lib->ms.num_active_planes; + CheckGlobalPrefetchAdmissibility_params->pixel_format = s->pixel_format; + CheckGlobalPrefetchAdmissibility_params->chunk_bytes_l = mode_lib->ip.pixel_chunk_size_kbytes * 1024; + CheckGlobalPrefetchAdmissibility_params->chunk_bytes_c = mode_lib->ip.pixel_chunk_size_kbytes * 1024; + CheckGlobalPrefetchAdmissibility_params->lb_source_lines_l = s->lb_source_lines_l; + CheckGlobalPrefetchAdmissibility_params->lb_source_lines_c = s->lb_source_lines_c; + CheckGlobalPrefetchAdmissibility_params->swath_height_l = mode_lib->ms.SwathHeightY; + CheckGlobalPrefetchAdmissibility_params->swath_height_c = mode_lib->ms.SwathHeightC; + CheckGlobalPrefetchAdmissibility_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes; + CheckGlobalPrefetchAdmissibility_params->compressed_buffer_size_kbytes = mode_lib->ms.CompressedBufferSizeInkByte; + CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_l = mode_lib->ms.DETBufferSizeY; + CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_c = mode_lib->ms.DETBufferSizeC; + CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_l = s->full_swath_bytes_l; + CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_c = s->full_swath_bytes_c; + CheckGlobalPrefetchAdmissibility_params->prefetch_sw_bytes = s->prefetch_sw_bytes; + CheckGlobalPrefetchAdmissibility_params->Tpre_rounded = s->Tpre_rounded; + CheckGlobalPrefetchAdmissibility_params->Tpre_oto = s->Tpre_oto; + CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = mode_lib->ms.support.urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]; + CheckGlobalPrefetchAdmissibility_params->line_time = s->line_times; + CheckGlobalPrefetchAdmissibility_params->dst_y_prefetch = mode_lib->ms.dst_y_prefetch; + if (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps < 10 * 1024) + CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = 10 * 1024; + + CheckGlobalPrefetchAdmissibility_params->estimated_dcfclk_mhz = (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps / (double) mode_lib->soc.return_bus_width_bytes) / + ((double)mode_lib->soc.qos_parameters.derate_table.system_active_urgent.dcfclk_derate_percent / 100.0); + + // if recalc_prefetch_schedule is set, recalculate the prefetch schedule with the new impacted_Tpre, prefetch should be possible + CheckGlobalPrefetchAdmissibility_params->recalc_prefetch_schedule = &s->recalc_prefetch_schedule; + CheckGlobalPrefetchAdmissibility_params->impacted_dst_y_pre = s->impacted_dst_y_pre; + mode_lib->ms.support.PrefetchSupported = CheckGlobalPrefetchAdmissibility(&mode_lib->scratch, CheckGlobalPrefetchAdmissibility_params); + s->recalc_prefetch_done = 1; + s->recalc_prefetch_schedule = 1; + } +#endif + } // prefetch schedule ok, do urg bw and flip schedule + } while (s->recalc_prefetch_schedule); + + // Flip Schedule + // Both prefetch schedule and BW okay + if (mode_lib->ms.support.PrefetchSupported == true) { + mode_lib->ms.BandwidthAvailableForImmediateFlip = + get_bandwidth_available_for_immediate_flip( + dml2_core_internal_soc_state_sys_active, + mode_lib->ms.support.urg_bandwidth_required_qual, // no flip + mode_lib->ms.support.urg_bandwidth_available); + + mode_lib->ms.TotImmediateFlipBytes = 0; + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + if (display_cfg->plane_descriptors[k].immediate_flip) { + s->per_pipe_flip_bytes[k] = get_pipe_flip_bytes( + s->HostVMInefficiencyFactor, + mode_lib->ms.vm_bytes[k], + mode_lib->ms.DPTEBytesPerRow[k], + mode_lib->ms.meta_row_bytes[k]); + } else { + s->per_pipe_flip_bytes[k] = 0; + } + mode_lib->ms.TotImmediateFlipBytes += s->per_pipe_flip_bytes[k] * mode_lib->ms.NoOfDPP[k]; + + } + + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { + CalculateFlipSchedule( + &mode_lib->scratch, + display_cfg->plane_descriptors[k].immediate_flip, + 1, // use_lb_flip_bw + s->HostVMInefficiencyFactor, + s->Tvm_trips_flip[k], + s->Tr0_trips_flip[k], + s->Tvm_trips_flip_rounded[k], + s->Tr0_trips_flip_rounded[k], + display_cfg->gpuvm_enable, + mode_lib->ms.vm_bytes[k], + mode_lib->ms.DPTEBytesPerRow[k], + mode_lib->ms.BandwidthAvailableForImmediateFlip, + mode_lib->ms.TotImmediateFlipBytes, + display_cfg->plane_descriptors[k].pixel_format, + (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)), + display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio, + display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio, + mode_lib->ms.Tno_bw_flip[k], + mode_lib->ms.dpte_row_height[k], + mode_lib->ms.dpte_row_height_chroma[k], + mode_lib->ms.use_one_row_for_frame_flip[k], + mode_lib->ip.max_flip_time_us, + mode_lib->ip.max_flip_time_lines, + s->per_pipe_flip_bytes[k], + mode_lib->ms.meta_row_bytes[k], + s->meta_row_height_luma[k], + s->meta_row_height_chroma[k], + mode_lib->ip.dcn_mrq_present && display_cfg->plane_descriptors[k].surface.dcc.enable, + + /* Output */ + &mode_lib->ms.dst_y_per_vm_flip[k], + &mode_lib->ms.dst_y_per_row_flip[k], + &mode_lib->ms.final_flip_bw[k], + &mode_lib->ms.ImmediateFlipSupportedForPipe[k]); + } + + calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = s->dummy_bw; + calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required_flip; + calculate_peak_bandwidth_params->urg_bandwidth_required_qual = s->dummy_bw; + calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required_flip; + calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = s->surface_dummy_bw; + calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw; + + calculate_peak_bandwidth_params->display_cfg = display_cfg; + calculate_peak_bandwidth_params->inc_flip_bw = 1; + calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes; + calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP; + calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0; + calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1; + calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0; + calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1; + calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor; + calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor; + + calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l; + calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c; + calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma; + calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma; + calculate_peak_bandwidth_params->prefetch_bandwidth_oto = mode_lib->ms.RequiredPrefetchBWOTO; + calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l; + calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c; + calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw; + calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw; + calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw; + calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw; + calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw; + calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw; + calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma; + calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma; + calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor; + calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre; + calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre; + calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre; + + calculate_peak_bandwidth_required( + &mode_lib->scratch, + calculate_peak_bandwidth_params); + + calculate_immediate_flip_bandwidth_support( + &s->dummy_single[0], // double* frac_urg_bandwidth_flip + &mode_lib->ms.support.ImmediateFlipSupport, + + dml2_core_internal_soc_state_sys_active, + mode_lib->ms.support.urg_bandwidth_required_flip, + mode_lib->ms.support.non_urg_bandwidth_required_flip, + mode_lib->ms.support.urg_bandwidth_available); + + for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) { + if (display_cfg->plane_descriptors[k].immediate_flip == true && mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false) + mode_lib->ms.support.ImmediateFlipSupport = false; + } + + } else { // if prefetch not support, assume iflip is not supported too + mode_lib->ms.support.ImmediateFlipSupport = false; + } + + s->mSOCParameters.UrgentLatency = mode_lib->ms.UrgLatency; + s->mSOCParameters.ExtraLatency = mode_lib->ms.ExtraLatency; + s->mSOCParameters.ExtraLatency_sr = mode_lib->ms.ExtraLatency_sr; + s->mSOCParameters.WritebackLatency = mode_lib->soc.qos_parameters.writeback.base_latency_us; + s->mSOCParameters.DRAMClockChangeLatency = mode_lib->soc.power_management_parameters.dram_clk_change_blackout_us; + s->mSOCParameters.FCLKChangeLatency = mode_lib->soc.power_management_parameters.fclk_change_blackout_us; + s->mSOCParameters.SRExitTime = mode_lib->soc.power_management_parameters.stutter_exit_latency_us; + s->mSOCParameters.SREnterPlusExitTime = mode_lib->soc.power_management_parameters.stutter_enter_plus_exit_latency_us; + s->mSOCParameters.SRExitZ8Time = mode_lib->soc.power_management_parameters.z8_stutter_exit_latency_us; + s->mSOCParameters.SREnterPlusExitZ8Time = mode_lib->soc.power_management_parameters.z8_stutter_enter_plus_exit_latency_us; + s->mSOCParameters.USRRetrainingLatency = 0; + s->mSOCParameters.SMNLatency = 0; + s->mSOCParameters.g6_temp_read_blackout_us = get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), mode_lib->ms.state_idx); + s->mSOCParameters.max_urgent_latency_us = get_max_urgent_latency_us(&mode_lib->soc.qos_parameters.qos_params.dcn4x, mode_lib->ms.uclk_freq_mhz, mode_lib->ms.FabricClock, mode_lib->ms.state_idx); + s->mSOCParameters.df_response_time_us = mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles / mode_lib->ms.FabricClock; + s->mSOCParameters.qos_type = mode_lib->soc.qos_parameters.qos_type; + + CalculateWatermarks_params->display_cfg = display_cfg; + CalculateWatermarks_params->USRRetrainingRequired = false; + CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes; + CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ip.max_line_buffer_lines; + CalculateWatermarks_params->LineBufferSize = mode_lib->ip.line_buffer_size_bits; + CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ip.writeback_interface_buffer_size_kbytes; + CalculateWatermarks_params->DCFCLK = mode_lib->ms.DCFCLK; + CalculateWatermarks_params->SynchronizeTimings = display_cfg->overrides.synchronize_timings; + CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChange = display_cfg->overrides.synchronize_ddr_displays_for_uclk_pstate_change; + CalculateWatermarks_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes; + CalculateWatermarks_params->mmSOCParameters = s->mSOCParameters; + CalculateWatermarks_params->WritebackChunkSize = mode_lib->ip.writeback_chunk_size_kbytes; + CalculateWatermarks_params->SOCCLK = mode_lib->ms.SOCCLK; + CalculateWatermarks_params->DCFClkDeepSleep = mode_lib->ms.dcfclk_deepsleep; + CalculateWatermarks_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeY; + CalculateWatermarks_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeC; + CalculateWatermarks_params->SwathHeightY = mode_lib->ms.SwathHeightY; + CalculateWatermarks_params->SwathHeightC = mode_lib->ms.SwathHeightC; + CalculateWatermarks_params->SwathWidthY = mode_lib->ms.SwathWidthY; + CalculateWatermarks_params->SwathWidthC = mode_lib->ms.SwathWidthC; + CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.NoOfDPP; + CalculateWatermarks_params->BytePerPixelDETY = mode_lib->ms.BytePerPixelInDETY; + CalculateWatermarks_params->BytePerPixelDETC = mode_lib->ms.BytePerPixelInDETC; + CalculateWatermarks_params->DSTXAfterScaler = s->DSTXAfterScaler; + CalculateWatermarks_params->DSTYAfterScaler = s->DSTYAfterScaler; + CalculateWatermarks_params->UnboundedRequestEnabled = mode_lib->ms.UnboundedRequestEnabled; + CalculateWatermarks_params->CompressedBufferSizeInkByte = mode_lib->ms.CompressedBufferSizeInkByte; + CalculateWatermarks_params->meta_row_height_l = s->meta_row_height_luma; + CalculateWatermarks_params->meta_row_height_c = s->meta_row_height_chroma; + + // Output + CalculateWatermarks_params->Watermark = &mode_lib->ms.support.watermarks; // Watermarks *Watermark + CalculateWatermarks_params->DRAMClockChangeSupport = mode_lib->ms.support.DRAMClockChangeSupport; + CalculateWatermarks_params->global_dram_clock_change_supported = &mode_lib->ms.support.global_dram_clock_change_supported; + CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0]; // double *MaxActiveDRAMClockChangeLatencySupported[] + CalculateWatermarks_params->SubViewportLinesNeededInMALL = mode_lib->ms.SubViewportLinesNeededInMALL; // unsigned int SubViewportLinesNeededInMALL[] + CalculateWatermarks_params->FCLKChangeSupport = mode_lib->ms.support.FCLKChangeSupport; + CalculateWatermarks_params->global_fclk_change_supported = &mode_lib->ms.support.global_fclk_change_supported; + CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // double *MaxActiveFCLKChangeLatencySupported + CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport; + CalculateWatermarks_params->g6_temp_read_support = &mode_lib->ms.support.g6_temp_read_support; + CalculateWatermarks_params->VActiveLatencyHidingMargin = mode_lib->ms.VActiveLatencyHidingMargin; + CalculateWatermarks_params->VActiveLatencyHidingUs = mode_lib->ms.VActiveLatencyHidingUs; + + CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(&mode_lib->scratch, CalculateWatermarks_params); + + calculate_pstate_keepout_dst_lines(display_cfg, &mode_lib->ms.support.watermarks, s->dummy_integer_array[0]); + DML_LOG_VERBOSE("DML::%s: Done prefetch calculation\n", __func__); + +} + static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out_params) { struct dml2_core_internal_display_mode_lib *mode_lib = in_out_params->mode_lib; const struct dml2_display_cfg *display_cfg = in_out_params->in_display_cfg; const struct dml2_mcg_min_clock_table *min_clk_table = in_out_params->min_clk_table; -#if defined(__DML_VBA_DEBUG__) - double old_ReadBandwidthLuma; - double old_ReadBandwidthChroma; -#endif double outstanding_latency_us = 0; - double min_return_bw_for_latency; struct dml2_core_calcs_mode_support_locals *s = &mode_lib->scratch.dml_core_mode_support_locals; - struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params; struct dml2_core_calcs_CalculateVMRowAndSwath_params *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params; struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params; - struct dml2_core_calcs_CalculatePrefetchSchedule_params *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params; -#ifdef DML_GLOBAL_PREFETCH_CHECK - struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *CheckGlobalPrefetchAdmissibility_params = &mode_lib->scratch.CheckGlobalPrefetchAdmissibility_params; -#endif - struct dml2_core_calcs_calculate_tdlut_setting_params *calculate_tdlut_setting_params = &mode_lib->scratch.calculate_tdlut_setting_params; struct dml2_core_calcs_calculate_mcache_setting_params *calculate_mcache_setting_params = &mode_lib->scratch.calculate_mcache_setting_params; - struct dml2_core_calcs_calculate_peak_bandwidth_required_params *calculate_peak_bandwidth_params = &mode_lib->scratch.calculate_peak_bandwidth_params; struct dml2_core_calcs_calculate_bytes_to_fetch_required_to_hide_latency_params *calculate_bytes_to_fetch_required_to_hide_latency_params = &mode_lib->scratch.calculate_bytes_to_fetch_required_to_hide_latency_params; unsigned int k, m, n; @@ -7374,9 +7932,9 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.FabricClock = ((double)min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].min_fclk_khz / 1000); mode_lib->ms.MaxDCFCLK = (double)min_clk_table->max_clocks_khz.dcfclk / 1000; mode_lib->ms.MaxFabricClock = (double)min_clk_table->max_clocks_khz.fclk / 1000; - mode_lib->ms.max_dispclk_freq_mhz = (double)min_clk_table->max_clocks_khz.dispclk / 1000; + mode_lib->ms.max_dispclk_freq_mhz = (double)min_clk_table->max_ss_clocks_khz.dispclk / 1000; mode_lib->ms.max_dscclk_freq_mhz = (double)min_clk_table->max_clocks_khz.dscclk / 1000; - mode_lib->ms.max_dppclk_freq_mhz = (double)min_clk_table->max_clocks_khz.dppclk / 1000; + mode_lib->ms.max_dppclk_freq_mhz = (double)min_clk_table->max_ss_clocks_khz.dppclk / 1000; mode_lib->ms.uclk_freq_mhz = dram_bw_kbps_to_uclk_mhz(min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps, &mode_lib->soc.clk_table.dram_config); mode_lib->ms.dram_bw_mbps = ((double)min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps / 1000); mode_lib->ms.max_dram_bw_mbps = ((double)min_clk_table->dram_bw_table.entries[min_clk_table->dram_bw_table.num_entries - 1].pre_derate_dram_bw_kbps / 1000); @@ -7384,25 +7942,25 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.active_min_uclk_dpm_index = get_active_min_uclk_dpm_index((unsigned int) (mode_lib->ms.uclk_freq_mhz * 1000.0), &mode_lib->soc.clk_table); #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: --- START --- \n", __func__); - dml2_printf("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes); - dml2_printf("DML::%s: min_clk_index = %0d\n", __func__, in_out_params->min_clk_index); - dml2_printf("DML::%s: qos_param_index = %0d\n", __func__, mode_lib->ms.qos_param_index); - dml2_printf("DML::%s: SOCCLK = %f\n", __func__, mode_lib->ms.SOCCLK); - dml2_printf("DML::%s: dram_bw_mbps = %f\n", __func__, mode_lib->ms.dram_bw_mbps); - dml2_printf("DML::%s: uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz); - dml2_printf("DML::%s: DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK); - dml2_printf("DML::%s: FabricClock = %f\n", __func__, mode_lib->ms.FabricClock); - dml2_printf("DML::%s: MaxDCFCLK = %f\n", __func__, mode_lib->ms.MaxDCFCLK); - dml2_printf("DML::%s: max_dispclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dispclk_freq_mhz); - dml2_printf("DML::%s: max_dscclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dscclk_freq_mhz); - dml2_printf("DML::%s: max_dppclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dppclk_freq_mhz); - dml2_printf("DML::%s: MaxFabricClock = %f\n", __func__, mode_lib->ms.MaxFabricClock); - dml2_printf("DML::%s: ip.compressed_buffer_segment_size_in_kbytes = %u\n", __func__, mode_lib->ip.compressed_buffer_segment_size_in_kbytes); - dml2_printf("DML::%s: ip.dcn_mrq_present = %u\n", __func__, mode_lib->ip.dcn_mrq_present); + DML_LOG_VERBOSE("DML::%s: --- START --- \n", __func__); + DML_LOG_VERBOSE("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes); + DML_LOG_VERBOSE("DML::%s: min_clk_index = %0d\n", __func__, in_out_params->min_clk_index); + DML_LOG_VERBOSE("DML::%s: qos_param_index = %0d\n", __func__, mode_lib->ms.qos_param_index); + DML_LOG_VERBOSE("DML::%s: SOCCLK = %f\n", __func__, mode_lib->ms.SOCCLK); + DML_LOG_VERBOSE("DML::%s: dram_bw_mbps = %f\n", __func__, mode_lib->ms.dram_bw_mbps); + DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK); + DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, mode_lib->ms.FabricClock); + DML_LOG_VERBOSE("DML::%s: MaxDCFCLK = %f\n", __func__, mode_lib->ms.MaxDCFCLK); + DML_LOG_VERBOSE("DML::%s: max_dispclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dispclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: max_dscclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dscclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: max_dppclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dppclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: MaxFabricClock = %f\n", __func__, mode_lib->ms.MaxFabricClock); + DML_LOG_VERBOSE("DML::%s: ip.compressed_buffer_segment_size_in_kbytes = %u\n", __func__, mode_lib->ip.compressed_buffer_segment_size_in_kbytes); + DML_LOG_VERBOSE("DML::%s: ip.dcn_mrq_present = %u\n", __func__, mode_lib->ip.dcn_mrq_present); for (k = 0; k < mode_lib->ms.num_active_planes; k++) - dml2_printf("DML::%s: plane_%d: reserved_vblank_time_ns = %u\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns); + DML_LOG_VERBOSE("DML::%s: plane_%d: reserved_vblank_time_ns = %lu\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns); #endif CalculateMaxDETAndMinCompressedBufferSize( @@ -7504,12 +8062,10 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out display_cfg->plane_descriptors[k].cursor.cursor_bpp / 8.0 / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)); #ifdef __DML_VBA_DEBUG__ - old_ReadBandwidthLuma = mode_lib->ms.SwathWidthYSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelInDETY[k], 1.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio; - old_ReadBandwidthChroma = mode_lib->ms.SwathWidthYSingleDPP[k] / 2 * math_ceil2(mode_lib->ms.BytePerPixelInDETC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio / 2.0; - dml2_printf("DML::%s: k=%u, old_ReadBandwidthLuma = %f\n", __func__, k, old_ReadBandwidthLuma); - dml2_printf("DML::%s: k=%u, old_ReadBandwidthChroma = %f\n", __func__, k, old_ReadBandwidthChroma); - dml2_printf("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_l[k]); - dml2_printf("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, old_ReadBandwidthLuma = %f\n", __func__, k, mode_lib->ms.SwathWidthYSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelInDETY[k], 1.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u, old_ReadBandwidthChroma = %f\n", __func__, k, mode_lib->ms.SwathWidthYSingleDPP[k] / 2 * math_ceil2(mode_lib->ms.BytePerPixelInDETC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio / 2.0); + DML_LOG_VERBOSE("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_c[k]); #endif } @@ -7629,13 +8185,13 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.MaximumSwathWidthLuma[k] = math_min2(s->MaximumSwathWidthSupportLuma, mode_lib->ms.MaximumSwathWidthInLineBufferLuma); mode_lib->ms.MaximumSwathWidthChroma[k] = math_min2(s->MaximumSwathWidthSupportChroma, mode_lib->ms.MaximumSwathWidthInLineBufferChroma); - dml2_printf("DML::%s: k=%u MaximumSwathWidthLuma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthLuma[k]); - dml2_printf("DML::%s: k=%u MaximumSwathWidthSupportLuma=%u\n", __func__, k, s->MaximumSwathWidthSupportLuma); - dml2_printf("DML::%s: k=%u MaximumSwathWidthInLineBufferLuma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthInLineBufferLuma); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthLuma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthSupportLuma=%u\n", __func__, k, s->MaximumSwathWidthSupportLuma); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthInLineBufferLuma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthInLineBufferLuma); - dml2_printf("DML::%s: k=%u MaximumSwathWidthChroma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthChroma[k]); - dml2_printf("DML::%s: k=%u MaximumSwathWidthSupportChroma=%u\n", __func__, k, s->MaximumSwathWidthSupportChroma); - dml2_printf("DML::%s: k=%u MaximumSwathWidthInLineBufferChroma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthInLineBufferChroma); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthChroma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthSupportChroma=%u\n", __func__, k, s->MaximumSwathWidthSupportChroma); + DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthInLineBufferChroma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthInLineBufferChroma); } /* Cursor Support Check */ @@ -7672,11 +8228,11 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.support.AlignedCPitch[k] > display_cfg->plane_descriptors[k].surface.plane1.pitch) { mode_lib->ms.support.PitchSupport = false; #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: k=%u AlignedYPitch = %d\n", __func__, k, mode_lib->ms.support.AlignedYPitch[k]); - dml2_printf("DML::%s: k=%u PitchY = %d\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.pitch); - dml2_printf("DML::%s: k=%u AlignedCPitch = %d\n", __func__, k, mode_lib->ms.support.AlignedCPitch[k]); - dml2_printf("DML::%s: k=%u PitchC = %d\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane1.pitch); - dml2_printf("DML::%s: k=%u PitchSupport = %d\n", __func__, k, mode_lib->ms.support.PitchSupport); + DML_LOG_VERBOSE("DML::%s: k=%u AlignedYPitch = %d\n", __func__, k, mode_lib->ms.support.AlignedYPitch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u PitchY = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.pitch); + DML_LOG_VERBOSE("DML::%s: k=%u AlignedCPitch = %d\n", __func__, k, mode_lib->ms.support.AlignedCPitch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u PitchC = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane1.pitch); + DML_LOG_VERBOSE("DML::%s: k=%u PitchSupport = %d\n", __func__, k, mode_lib->ms.support.PitchSupport); #endif } @@ -7708,11 +8264,11 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out display_cfg->plane_descriptors[k].composition.viewport.plane0.height > display_cfg->plane_descriptors[k].surface.plane0.height) { mode_lib->ms.support.ViewportExceedsSurface = true; #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: k=%u ViewportWidth = %d\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.width); - dml2_printf("DML::%s: k=%u SurfaceWidthY = %d\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.width); - dml2_printf("DML::%s: k=%u ViewportHeight = %d\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.height); - dml2_printf("DML::%s: k=%u SurfaceHeightY = %d\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.height); - dml2_printf("DML::%s: k=%u ViewportExceedsSurface = %d\n", __func__, k, mode_lib->ms.support.ViewportExceedsSurface); + DML_LOG_VERBOSE("DML::%s: k=%u ViewportWidth = %ld\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.width); + DML_LOG_VERBOSE("DML::%s: k=%u SurfaceWidthY = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.width); + DML_LOG_VERBOSE("DML::%s: k=%u ViewportHeight = %ld\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.height); + DML_LOG_VERBOSE("DML::%s: k=%u SurfaceHeightY = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.height); + DML_LOG_VERBOSE("DML::%s: k=%u ViewportExceedsSurface = %d\n", __func__, k, mode_lib->ms.support.ViewportExceedsSurface); #endif } if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format) || display_cfg->plane_descriptors[k].pixel_format == dml2_rgbe_alpha) { @@ -7894,8 +8450,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.TotalNumberOfActiveDPP = mode_lib->ms.TotalNumberOfActiveDPP + s->NumberOfDPPDSC; } #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: k=%d RequiresDSC = %d\n", __func__, k, mode_lib->ms.RequiresDSC[k]); - dml2_printf("DML::%s: k=%d ODMMode = %d\n", __func__, k, mode_lib->ms.ODMMode[k]); + DML_LOG_VERBOSE("DML::%s: k=%d RequiresDSC = %d\n", __func__, k, mode_lib->ms.RequiresDSC[k]); + DML_LOG_VERBOSE("DML::%s: k=%d ODMMode = %d\n", __func__, k, mode_lib->ms.ODMMode[k]); #endif // ensure the number dsc slices is integer multiple based on ODM mode @@ -7911,9 +8467,9 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.support.DSCSlicesODMModeSupported = ((mode_lib->ms.support.NumberOfDSCSlices[k] % 4) == 0); #if defined(__DML_VBA_DEBUG__) if (!mode_lib->ms.support.DSCSlicesODMModeSupported) { - dml2_printf("DML::%s: k=%d Invalid dsc num_slices and ODM mode setting\n", __func__, k); - dml2_printf("DML::%s: k=%d num_slices = %d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.overrides.num_slices); - dml2_printf("DML::%s: k=%d ODMMode = %d\n", __func__, k, mode_lib->ms.ODMMode[k]); + DML_LOG_VERBOSE("DML::%s: k=%d Invalid dsc num_slices and ODM mode setting\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%d num_slices = %d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.overrides.num_slices); + DML_LOG_VERBOSE("DML::%s: k=%d ODMMode = %d\n", __func__, k, mode_lib->ms.ODMMode[k]); } #endif } else { @@ -7958,7 +8514,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.MPCCombine[k] = false; mode_lib->ms.NoOfDPP[k] = 1; if (!mode_lib->ms.SingleDPPViewportSizeSupportPerSurface[k]) { - dml2_printf("WARNING: DML::%s: MPCC is override to disable but viewport is too large to be supported with single pipe!\n", __func__); + DML_LOG_VERBOSE("WARNING: DML::%s: MPCC is override to disable but viewport is too large to be supported with single pipe!\n", __func__); } } else { if ((mode_lib->ms.MinDPPCLKUsingSingleDPP[k] > mode_lib->ms.max_dppclk_freq_mhz) || !mode_lib->ms.SingleDPPViewportSizeSupportPerSurface[k]) { @@ -7968,7 +8524,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out } } #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: k=%d, NoOfDPP = %d\n", __func__, k, mode_lib->ms.NoOfDPP[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, NoOfDPP = %d\n", __func__, k, mode_lib->ms.NoOfDPP[k]); #endif } @@ -8138,7 +8694,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.audio_sample_rate, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.audio_sample_layout); - if (mode_lib->ms.RequiredDTBCLK[k] > ((double)min_clk_table->max_clocks_khz.dtbclk / 1000)) { + if (mode_lib->ms.RequiredDTBCLK[k] > ((double)min_clk_table->max_ss_clocks_khz.dtbclk / 1000)) { mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = true; } } else { @@ -8167,7 +8723,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out s->DSCFormatFactor = 1; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, RequiresDSC = %u\n", __func__, k, mode_lib->ms.RequiresDSC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, RequiresDSC = %u\n", __func__, k, mode_lib->ms.RequiresDSC[k]); #endif if (mode_lib->ms.RequiresDSC[k] == true) { s->PixelClockBackEndFactor = 3.0; @@ -8185,10 +8741,10 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, PixelClockBackEnd = %f\n", __func__, k, s->PixelClockBackEnd[k]); - dml2_printf("DML::%s: k=%u, required_dscclk_freq_mhz = %f\n", __func__, k, mode_lib->ms.required_dscclk_freq_mhz[k]); - dml2_printf("DML::%s: k=%u, DSCFormatFactor = %u\n", __func__, k, s->DSCFormatFactor); - dml2_printf("DML::%s: k=%u, DSCCLKRequiredMoreThanSupported = %u\n", __func__, k, mode_lib->ms.support.DSCCLKRequiredMoreThanSupported); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelClockBackEnd = %f\n", __func__, k, s->PixelClockBackEnd[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, required_dscclk_freq_mhz = %f\n", __func__, k, mode_lib->ms.required_dscclk_freq_mhz[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, DSCFormatFactor = %u\n", __func__, k, s->DSCFormatFactor); + DML_LOG_VERBOSE("DML::%s: k=%u, DSCCLKRequiredMoreThanSupported = %u\n", __func__, k, mode_lib->ms.support.DSCCLKRequiredMoreThanSupported); #endif } } @@ -8423,13 +8979,13 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.support.DCCMetaBufferSizeNotExceeded = false; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, mode_lib->ms.PTEBufferSizeNotExceeded[k]); - dml2_printf("DML::%s: k=%u, DCCMetaBufferSizeNotExceeded = %u\n", __func__, k, mode_lib->ms.DCCMetaBufferSizeNotExceeded[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, mode_lib->ms.PTEBufferSizeNotExceeded[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, DCCMetaBufferSizeNotExceeded = %u\n", __func__, k, mode_lib->ms.DCCMetaBufferSizeNotExceeded[k]); #endif } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: PTEBufferSizeNotExceeded = %u\n", __func__, mode_lib->ms.support.PTEBufferSizeNotExceeded); - dml2_printf("DML::%s: DCCMetaBufferSizeNotExceeded = %u\n", __func__, mode_lib->ms.support.DCCMetaBufferSizeNotExceeded); + DML_LOG_VERBOSE("DML::%s: PTEBufferSizeNotExceeded = %u\n", __func__, mode_lib->ms.support.PTEBufferSizeNotExceeded); + DML_LOG_VERBOSE("DML::%s: DCCMetaBufferSizeNotExceeded = %u\n", __func__, mode_lib->ms.support.DCCMetaBufferSizeNotExceeded); #endif /* VActive bytes to fetch for UCLK P-State */ @@ -8502,7 +9058,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out for (k = 0; k < mode_lib->ms.num_active_planes; ++k) { double line_time_us = (double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); - bool cursor_not_enough_urgent_latency_hiding = 0; + bool cursor_not_enough_urgent_latency_hiding = false; if (display_cfg->plane_descriptors[k].cursor.num_cursors > 0) { calculate_cursor_req_attributes( @@ -8531,9 +9087,9 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.UrgentBurstFactorCursorPre[k] = mode_lib->ms.UrgentBurstFactorCursor[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, Calling CalculateUrgentBurstFactor\n", __func__, k); - dml2_printf("DML::%s: k=%d, VRatio=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); - dml2_printf("DML::%s: k=%d, VRatioChroma=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%d, Calling CalculateUrgentBurstFactor\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%d, VRatio=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%d, VRatioChroma=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio); #endif CalculateUrgentBurstFactor( @@ -8605,7 +9161,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, MaximumVStartup = %u\n", __func__, k, s->MaximumVStartup[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, MaximumVStartup = %u\n", __func__, k, s->MaximumVStartup[k]); #endif /* Immediate Flip and MALL parameters */ @@ -8654,16 +9210,15 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out (s->SubViewportMALLPStateMethod && s->FullFrameMALLPStateMethod) || s->SubViewportMALLRefreshGreaterThan120Hz; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: SubViewportMALLPStateMethod = %u\n", __func__, s->SubViewportMALLPStateMethod); - dml2_printf("DML::%s: PhantomPipeMALLPStateMethod = %u\n", __func__, s->PhantomPipeMALLPStateMethod); - dml2_printf("DML::%s: FullFrameMALLPStateMethod = %u\n", __func__, s->FullFrameMALLPStateMethod); - dml2_printf("DML::%s: SubViewportMALLRefreshGreaterThan120Hz = %u\n", __func__, s->SubViewportMALLRefreshGreaterThan120Hz); - dml2_printf("DML::%s: InvalidCombinationOfMALLUseForPState = %u\n", __func__, mode_lib->ms.support.InvalidCombinationOfMALLUseForPState); - dml2_printf("DML::%s: in_out_params->min_clk_index = %u\n", __func__, in_out_params->min_clk_index); - dml2_printf("DML::%s: mode_lib->ms.DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK); - dml2_printf("DML::%s: mode_lib->ms.FabricClock = %f\n", __func__, mode_lib->ms.FabricClock); - dml2_printf("DML::%s: mode_lib->ms.uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz); - dml2_printf("DML::%s: urgent latency tolarance = %f\n", __func__, ((mode_lib->ip.rob_buffer_size_kbytes - mode_lib->ip.pixel_chunk_size_kbytes) * 1024 / (mode_lib->ms.DCFCLK * mode_lib->soc.return_bus_width_bytes))); + DML_LOG_VERBOSE("DML::%s: SubViewportMALLPStateMethod = %u\n", __func__, s->SubViewportMALLPStateMethod); + DML_LOG_VERBOSE("DML::%s: PhantomPipeMALLPStateMethod = %u\n", __func__, s->PhantomPipeMALLPStateMethod); + DML_LOG_VERBOSE("DML::%s: FullFrameMALLPStateMethod = %u\n", __func__, s->FullFrameMALLPStateMethod); + DML_LOG_VERBOSE("DML::%s: SubViewportMALLRefreshGreaterThan120Hz = %u\n", __func__, s->SubViewportMALLRefreshGreaterThan120Hz); + DML_LOG_VERBOSE("DML::%s: InvalidCombinationOfMALLUseForPState = %u\n", __func__, mode_lib->ms.support.InvalidCombinationOfMALLUseForPState); + DML_LOG_VERBOSE("DML::%s: in_out_params->min_clk_index = %u\n", __func__, in_out_params->min_clk_index); + DML_LOG_VERBOSE("DML::%s: mode_lib->ms.DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK); + DML_LOG_VERBOSE("DML::%s: mode_lib->ms.FabricClock = %f\n", __func__, mode_lib->ms.FabricClock); + DML_LOG_VERBOSE("DML::%s: mode_lib->ms.uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz); #endif mode_lib->ms.support.OutstandingRequestsSupport = true; @@ -8703,10 +9258,10 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: avg_urgent_latency_us = %f\n", __func__, mode_lib->ms.support.avg_urgent_latency_us); - dml2_printf("DML::%s: avg_non_urgent_latency_us = %f\n", __func__, mode_lib->ms.support.avg_non_urgent_latency_us); - dml2_printf("DML::%s: k=%d, request_size_bytes_luma = %d\n", __func__, k, mode_lib->ms.support.request_size_bytes_luma[k]); - dml2_printf("DML::%s: k=%d, outstanding_latency_us = %f (luma)\n", __func__, k, outstanding_latency_us); + DML_LOG_VERBOSE("DML::%s: avg_urgent_latency_us = %f\n", __func__, mode_lib->ms.support.avg_urgent_latency_us); + DML_LOG_VERBOSE("DML::%s: avg_non_urgent_latency_us = %f\n", __func__, mode_lib->ms.support.avg_non_urgent_latency_us); + DML_LOG_VERBOSE("DML::%s: k=%d, request_size_bytes_luma = %d\n", __func__, k, mode_lib->ms.support.request_size_bytes_luma[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, outstanding_latency_us = %f (luma)\n", __func__, k, outstanding_latency_us); #endif } @@ -8722,8 +9277,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.support.OutstandingRequestsUrgencyAvoidance = false; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, request_size_bytes_chroma = %d\n", __func__, k, mode_lib->ms.support.request_size_bytes_chroma[k]); - dml2_printf("DML::%s: k=%d, outstanding_latency_us = %f (chroma)\n", __func__, k, outstanding_latency_us); + DML_LOG_VERBOSE("DML::%s: k=%d, request_size_bytes_chroma = %d\n", __func__, k, mode_lib->ms.support.request_size_bytes_chroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, outstanding_latency_us = %f (chroma)\n", __func__, k, outstanding_latency_us); #endif } } @@ -8869,7 +9424,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out for (k = 0; k < mode_lib->ms.num_active_planes; ++k) { if (mode_lib->ms.NotEnoughUrgentLatencyHiding[k]) { mode_lib->ms.support.EnoughUrgentLatencyHidingSupport = false; - dml2_printf("DML::%s: k=%u NotEnoughUrgentLatencyHiding set\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: k=%u NotEnoughUrgentLatencyHiding set\n", __func__, k); } } @@ -8878,639 +9433,13 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out if (!mode_lib->ms.support.avg_bandwidth_support_ok[m][n] && (m == dml2_core_internal_soc_state_sys_active || mode_lib->soc.mall_allocated_for_dcn_mbytes > 0)) { mode_lib->ms.support.AvgBandwidthSupport = false; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: avg_bandwidth_support_ok[%s][%s] not ok\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n)); + DML_LOG_VERBOSE("DML::%s: avg_bandwidth_support_ok[%s][%s] not ok\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n)); #endif } } } - /* Prefetch Check */ - { - mode_lib->ms.TimeCalc = 24 / mode_lib->ms.dcfclk_deepsleep; - - calculate_hostvm_inefficiency_factor( - &s->HostVMInefficiencyFactor, - &s->HostVMInefficiencyFactorPrefetch, - - display_cfg->gpuvm_enable, - display_cfg->hostvm_enable, - mode_lib->ip.remote_iommu_outstanding_translations, - mode_lib->soc.max_outstanding_reqs, - mode_lib->ms.support.urg_bandwidth_available_pixel_and_vm[dml2_core_internal_soc_state_sys_active], - mode_lib->ms.support.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_sys_active]); - - mode_lib->ms.Total3dlutActive = 0; - for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) { - if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut) - mode_lib->ms.Total3dlutActive = mode_lib->ms.Total3dlutActive + 1; - - // Calculate tdlut schedule related terms - calculate_tdlut_setting_params->dispclk_mhz = mode_lib->ms.RequiredDISPCLK; - calculate_tdlut_setting_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut; - calculate_tdlut_setting_params->tdlut_width_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_width_mode; - calculate_tdlut_setting_params->tdlut_addressing_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_addressing_mode; - calculate_tdlut_setting_params->cursor_buffer_size = mode_lib->ip.cursor_buffer_size; - calculate_tdlut_setting_params->gpuvm_enable = display_cfg->gpuvm_enable; - calculate_tdlut_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes; - calculate_tdlut_setting_params->tdlut_mpc_width_flag = display_cfg->plane_descriptors[k].tdlut.tdlut_mpc_width_flag; - calculate_tdlut_setting_params->is_gfx11 = dml_get_gfx_version(display_cfg->plane_descriptors[k].surface.tiling); - - // output - calculate_tdlut_setting_params->tdlut_pte_bytes_per_frame = &s->tdlut_pte_bytes_per_frame[k]; - calculate_tdlut_setting_params->tdlut_bytes_per_frame = &s->tdlut_bytes_per_frame[k]; - calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k]; - calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k]; - calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k]; - calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k]; - calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k]; - - calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params); - } - - min_return_bw_for_latency = mode_lib->ms.support.urg_bandwidth_available_min_latency[dml2_core_internal_soc_state_sys_active]; - - if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn3) - s->ReorderingBytes = (unsigned int)(mode_lib->soc.clk_table.dram_config.channel_count * math_max3(mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_only_bytes, - mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes, - mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_vm_only_bytes)); - - CalculateExtraLatency( - display_cfg, - mode_lib->ip.rob_buffer_size_kbytes, - mode_lib->soc.qos_parameters.qos_params.dcn32x.loaded_round_trip_latency_fclk_cycles, - s->ReorderingBytes, - mode_lib->ms.DCFCLK, - mode_lib->ms.FabricClock, - mode_lib->ip.pixel_chunk_size_kbytes, - min_return_bw_for_latency, - mode_lib->ms.num_active_planes, - mode_lib->ms.NoOfDPP, - mode_lib->ms.dpte_group_bytes, - s->tdlut_bytes_per_group, - s->HostVMInefficiencyFactor, - s->HostVMInefficiencyFactorPrefetch, - mode_lib->soc.hostvm_min_page_size_kbytes, - mode_lib->soc.qos_parameters.qos_type, - !(display_cfg->overrides.max_outstanding_when_urgent_expected_disable), - mode_lib->soc.max_outstanding_reqs, - mode_lib->ms.support.request_size_bytes_luma, - mode_lib->ms.support.request_size_bytes_chroma, - mode_lib->ip.meta_chunk_size_kbytes, - mode_lib->ip.dchub_arb_to_ret_delay, - mode_lib->ms.TripToMemory, - mode_lib->ip.hostvm_mode, - - // output - &mode_lib->ms.ExtraLatency, - &mode_lib->ms.ExtraLatency_sr, - &mode_lib->ms.ExtraLatencyPrefetch); - - for (k = 0; k < mode_lib->ms.num_active_planes; k++) - s->impacted_dst_y_pre[k] = 0; - - s->recalc_prefetch_schedule = 0; - s->recalc_prefetch_done = 0; - do { - mode_lib->ms.support.PrefetchSupported = true; - - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); - s->pixel_format[k] = display_cfg->plane_descriptors[k].pixel_format; - - s->lb_source_lines_l[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits, - mode_lib->ms.NoOfDPP[k], - display_cfg->plane_descriptors[k].composition.viewport.plane0.width, - display_cfg->plane_descriptors[k].composition.viewport.plane0.height, - display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio, - display_cfg->plane_descriptors[k].composition.rotation_angle); - - s->lb_source_lines_c[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits, - mode_lib->ms.NoOfDPP[k], - display_cfg->plane_descriptors[k].composition.viewport.plane1.width, - display_cfg->plane_descriptors[k].composition.viewport.plane1.height, - display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio, - display_cfg->plane_descriptors[k].composition.rotation_angle); - - struct dml2_core_internal_DmlPipe *myPipe = &s->myPipe; - - mode_lib->ms.TWait[k] = CalculateTWait( - display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns, - mode_lib->ms.UrgLatency, - mode_lib->ms.TripToMemory, - !dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.drr_config.enabled ? - get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), in_out_params->min_clk_index) : 0.0); - - myPipe->Dppclk = mode_lib->ms.RequiredDPPCLK[k]; - myPipe->Dispclk = mode_lib->ms.RequiredDISPCLK; - myPipe->PixelClock = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); - myPipe->DCFClkDeepSleep = mode_lib->ms.dcfclk_deepsleep; - myPipe->DPPPerSurface = mode_lib->ms.NoOfDPP[k]; - myPipe->ScalerEnabled = display_cfg->plane_descriptors[k].composition.scaler_info.enabled; - myPipe->VRatio = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio; - myPipe->VRatioChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio; - myPipe->VTaps = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps; - myPipe->VTapsChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps; - myPipe->RotationAngle = display_cfg->plane_descriptors[k].composition.rotation_angle; - myPipe->mirrored = display_cfg->plane_descriptors[k].composition.mirrored; - myPipe->BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k]; - myPipe->BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k]; - myPipe->BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k]; - myPipe->BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k]; - myPipe->InterlaceEnable = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced; - myPipe->NumberOfCursors = display_cfg->plane_descriptors[k].cursor.num_cursors; - myPipe->VBlank = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total - display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_active; - myPipe->HTotal = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total; - myPipe->HActive = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active; - myPipe->DCCEnable = display_cfg->plane_descriptors[k].surface.dcc.enable; - myPipe->ODMMode = mode_lib->ms.ODMMode[k]; - myPipe->SourcePixelFormat = display_cfg->plane_descriptors[k].pixel_format; - myPipe->BytePerPixelY = mode_lib->ms.BytePerPixelY[k]; - myPipe->BytePerPixelC = mode_lib->ms.BytePerPixelC[k]; - myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ip.ptoi_supported; - -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k); - dml2_printf("DML::%s: MaximumVStartup = %u\n", __func__, s->MaximumVStartup[k]); -#endif - CalculatePrefetchSchedule_params->display_cfg = display_cfg; - CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactorPrefetch; - CalculatePrefetchSchedule_params->myPipe = myPipe; - CalculatePrefetchSchedule_params->DSCDelay = mode_lib->ms.DSCDelay[k]; - CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ip.dppclk_delay_subtotal + mode_lib->ip.dppclk_delay_cnvc_formatter; - CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ip.dppclk_delay_scl; - CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ip.dppclk_delay_scl_lb_only; - CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ip.dppclk_delay_cnvc_cursor; - CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ip.dispclk_delay_subtotal; - CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (unsigned int)(mode_lib->ms.SwathWidthY[k] / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio); - CalculatePrefetchSchedule_params->OutputFormat = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format; - CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ip.max_inter_dcn_tile_repeaters; - CalculatePrefetchSchedule_params->VStartup = s->MaximumVStartup[k]; - CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->soc.hostvm_min_page_size_kbytes; - CalculatePrefetchSchedule_params->DynamicMetadataEnable = display_cfg->plane_descriptors[k].dynamic_meta_data.enable; - CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ip.dynamic_metadata_vm_enabled; - CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = display_cfg->plane_descriptors[k].dynamic_meta_data.lines_before_active_required; - CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = display_cfg->plane_descriptors[k].dynamic_meta_data.transmitted_bytes; - CalculatePrefetchSchedule_params->UrgentLatency = mode_lib->ms.UrgLatency; - CalculatePrefetchSchedule_params->ExtraLatencyPrefetch = mode_lib->ms.ExtraLatencyPrefetch; - CalculatePrefetchSchedule_params->TCalc = mode_lib->ms.TimeCalc; - CalculatePrefetchSchedule_params->vm_bytes = mode_lib->ms.vm_bytes[k]; - CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow[k]; - CalculatePrefetchSchedule_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesY[k]; - CalculatePrefetchSchedule_params->VInitPreFillY = mode_lib->ms.PrefillY[k]; - CalculatePrefetchSchedule_params->MaxNumSwathY = mode_lib->ms.MaxNumSwathY[k]; - CalculatePrefetchSchedule_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesC[k]; - CalculatePrefetchSchedule_params->VInitPreFillC = mode_lib->ms.PrefillC[k]; - CalculatePrefetchSchedule_params->MaxNumSwathC = mode_lib->ms.MaxNumSwathC[k]; - CalculatePrefetchSchedule_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub[k]; - CalculatePrefetchSchedule_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub[k]; - CalculatePrefetchSchedule_params->SwathHeightY = mode_lib->ms.SwathHeightY[k]; - CalculatePrefetchSchedule_params->SwathHeightC = mode_lib->ms.SwathHeightC[k]; - CalculatePrefetchSchedule_params->TWait = mode_lib->ms.TWait[k]; - CalculatePrefetchSchedule_params->Ttrip = mode_lib->ms.TripToMemory; - CalculatePrefetchSchedule_params->Turg = mode_lib->ms.UrgLatency; - CalculatePrefetchSchedule_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut; - CalculatePrefetchSchedule_params->tdlut_pte_bytes_per_frame = s->tdlut_pte_bytes_per_frame[k]; - CalculatePrefetchSchedule_params->tdlut_bytes_per_frame = s->tdlut_bytes_per_frame[k]; - CalculatePrefetchSchedule_params->tdlut_opt_time = s->tdlut_opt_time[k]; - CalculatePrefetchSchedule_params->tdlut_drain_time = s->tdlut_drain_time[k]; - CalculatePrefetchSchedule_params->num_cursors = (display_cfg->plane_descriptors[k].cursor.cursor_width > 0); - CalculatePrefetchSchedule_params->cursor_bytes_per_chunk = s->cursor_bytes_per_chunk[k]; - CalculatePrefetchSchedule_params->cursor_bytes_per_line = s->cursor_bytes_per_line[k]; - CalculatePrefetchSchedule_params->dcc_enable = display_cfg->plane_descriptors[k].surface.dcc.enable; - CalculatePrefetchSchedule_params->mrq_present = mode_lib->ip.dcn_mrq_present; - CalculatePrefetchSchedule_params->meta_row_bytes = mode_lib->ms.meta_row_bytes[k]; - CalculatePrefetchSchedule_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor[k]; - CalculatePrefetchSchedule_params->impacted_dst_y_pre = s->impacted_dst_y_pre[k]; - CalculatePrefetchSchedule_params->vactive_sw_bw_l = mode_lib->ms.vactive_sw_bw_l[k]; - CalculatePrefetchSchedule_params->vactive_sw_bw_c = mode_lib->ms.vactive_sw_bw_c[k]; - - // output - CalculatePrefetchSchedule_params->DSTXAfterScaler = &s->DSTXAfterScaler[k]; - CalculatePrefetchSchedule_params->DSTYAfterScaler = &s->DSTYAfterScaler[k]; - CalculatePrefetchSchedule_params->dst_y_prefetch = &mode_lib->ms.dst_y_prefetch[k]; - CalculatePrefetchSchedule_params->dst_y_per_vm_vblank = &mode_lib->ms.LinesForVM[k]; - CalculatePrefetchSchedule_params->dst_y_per_row_vblank = &mode_lib->ms.LinesForDPTERow[k]; - CalculatePrefetchSchedule_params->VRatioPrefetchY = &mode_lib->ms.VRatioPreY[k]; - CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->ms.VRatioPreC[k]; - CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWLuma = &mode_lib->ms.RequiredPrefetchPixelDataBWLuma[k]; // prefetch_sw_bw_l - CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWChroma = &mode_lib->ms.RequiredPrefetchPixelDataBWChroma[k]; // prefetch_sw_bw_c - CalculatePrefetchSchedule_params->RequiredPrefetchBWOTO = &mode_lib->ms.RequiredPrefetchBWOTO[k]; - CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->ms.NoTimeForDynamicMetadata[k]; - CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->ms.Tno_bw[k]; - CalculatePrefetchSchedule_params->Tno_bw_flip = &mode_lib->ms.Tno_bw_flip[k]; - CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &mode_lib->ms.prefetch_vmrow_bw[k]; - CalculatePrefetchSchedule_params->Tdmdl_vm = &s->dummy_single[0]; - CalculatePrefetchSchedule_params->Tdmdl = &s->dummy_single[1]; - CalculatePrefetchSchedule_params->TSetup = &s->dummy_single[2]; - CalculatePrefetchSchedule_params->Tvm_trips = &s->Tvm_trips[k]; - CalculatePrefetchSchedule_params->Tr0_trips = &s->Tr0_trips[k]; - CalculatePrefetchSchedule_params->Tvm_trips_flip = &s->Tvm_trips_flip[k]; - CalculatePrefetchSchedule_params->Tr0_trips_flip = &s->Tr0_trips_flip[k]; - CalculatePrefetchSchedule_params->Tvm_trips_flip_rounded = &s->Tvm_trips_flip_rounded[k]; - CalculatePrefetchSchedule_params->Tr0_trips_flip_rounded = &s->Tr0_trips_flip_rounded[k]; - CalculatePrefetchSchedule_params->VUpdateOffsetPix = &s->dummy_integer[0]; - CalculatePrefetchSchedule_params->VUpdateWidthPix = &s->dummy_integer[1]; - CalculatePrefetchSchedule_params->VReadyOffsetPix = &s->dummy_integer[2]; - CalculatePrefetchSchedule_params->prefetch_cursor_bw = &mode_lib->ms.prefetch_cursor_bw[k]; - CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k]; - CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k]; - CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k]; - CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->prefetch_swath_time_us[k]; - - mode_lib->ms.NoTimeForPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params); - - mode_lib->ms.support.PrefetchSupported &= !mode_lib->ms.NoTimeForPrefetch[k]; - dml2_printf("DML::%s: k=%d, dst_y_per_vm_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_vm_vblank); - dml2_printf("DML::%s: k=%d, dst_y_per_row_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_row_vblank); - } // for k num_planes - - CalculateDCFCLKDeepSleepTdlut( - display_cfg, - mode_lib->ms.num_active_planes, - mode_lib->ms.BytePerPixelY, - mode_lib->ms.BytePerPixelC, - mode_lib->ms.SwathWidthY, - mode_lib->ms.SwathWidthC, - mode_lib->ms.NoOfDPP, - mode_lib->ms.PSCL_FACTOR, - mode_lib->ms.PSCL_FACTOR_CHROMA, - mode_lib->ms.RequiredDPPCLK, - mode_lib->ms.vactive_sw_bw_l, - mode_lib->ms.vactive_sw_bw_c, - mode_lib->soc.return_bus_width_bytes, - mode_lib->ms.RequiredDISPCLK, - s->tdlut_bytes_to_deliver, - s->prefetch_swath_time_us, - - /* Output */ - &mode_lib->ms.dcfclk_deepsleep); - - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - if (mode_lib->ms.dst_y_prefetch[k] < 2.0 - || mode_lib->ms.LinesForVM[k] >= 32.0 - || mode_lib->ms.LinesForDPTERow[k] >= 16.0 - || mode_lib->ms.NoTimeForPrefetch[k] == true - || s->DSTYAfterScaler[k] > 8) { - mode_lib->ms.support.PrefetchSupported = false; - dml2_printf("DML::%s: k=%d, dst_y_prefetch=%f (should not be < 2)\n", __func__, k, mode_lib->ms.dst_y_prefetch[k]); - dml2_printf("DML::%s: k=%d, LinesForVM=%f (should not be >= 32)\n", __func__, k, mode_lib->ms.LinesForVM[k]); - dml2_printf("DML::%s: k=%d, LinesForDPTERow=%f (should not be >= 16)\n", __func__, k, mode_lib->ms.LinesForDPTERow[k]); - dml2_printf("DML::%s: k=%d, DSTYAfterScaler=%d (should be <= 8)\n", __func__, k, s->DSTYAfterScaler[k]); - dml2_printf("DML::%s: k=%d, NoTimeForPrefetch=%d\n", __func__, k, mode_lib->ms.NoTimeForPrefetch[k]); - } - } - - mode_lib->ms.support.DynamicMetadataSupported = true; - for (k = 0; k < mode_lib->ms.num_active_planes; ++k) { - if (mode_lib->ms.NoTimeForDynamicMetadata[k] == true) { - mode_lib->ms.support.DynamicMetadataSupported = false; - } - } - - mode_lib->ms.support.VRatioInPrefetchSupported = true; - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - if (mode_lib->ms.VRatioPreY[k] > __DML2_CALCS_MAX_VRATIO_PRE__ || - mode_lib->ms.VRatioPreC[k] > __DML2_CALCS_MAX_VRATIO_PRE__) { - mode_lib->ms.support.VRatioInPrefetchSupported = false; - dml2_printf("DML::%s: k=%d VRatioPreY = %f (should be <= %f)\n", __func__, k, mode_lib->ms.VRatioPreY[k], __DML2_CALCS_MAX_VRATIO_PRE__); - dml2_printf("DML::%s: k=%d VRatioPreC = %f (should be <= %f)\n", __func__, k, mode_lib->ms.VRatioPreC[k], __DML2_CALCS_MAX_VRATIO_PRE__); - dml2_printf("DML::%s: VRatioInPrefetchSupported = %u\n", __func__, mode_lib->ms.support.VRatioInPrefetchSupported); - } - } - - mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.VRatioInPrefetchSupported; - - // By default, do not recalc prefetch schedule - s->recalc_prefetch_schedule = 0; - - // Only do urg vs prefetch bandwidth check, flip schedule check, power saving feature support check IF the Prefetch Schedule Check is ok - if (mode_lib->ms.support.PrefetchSupported) { - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - // Calculate Urgent burst factor for prefetch -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, Calling CalculateUrgentBurstFactor (for prefetch)\n", __func__, k); - dml2_printf("DML::%s: k=%d, VRatioPreY=%f\n", __func__, k, mode_lib->ms.VRatioPreY[k]); - dml2_printf("DML::%s: k=%d, VRatioPreC=%f\n", __func__, k, mode_lib->ms.VRatioPreC[k]); -#endif - CalculateUrgentBurstFactor( - &display_cfg->plane_descriptors[k], - mode_lib->ms.swath_width_luma_ub[k], - mode_lib->ms.swath_width_chroma_ub[k], - mode_lib->ms.SwathHeightY[k], - mode_lib->ms.SwathHeightC[k], - s->line_times[k], - mode_lib->ms.UrgLatency, - mode_lib->ms.VRatioPreY[k], - mode_lib->ms.VRatioPreC[k], - mode_lib->ms.BytePerPixelInDETY[k], - mode_lib->ms.BytePerPixelInDETC[k], - mode_lib->ms.DETBufferSizeY[k], - mode_lib->ms.DETBufferSizeC[k], - /* Output */ - &mode_lib->ms.UrgentBurstFactorLumaPre[k], - &mode_lib->ms.UrgentBurstFactorChromaPre[k], - &mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]); - } - - // Calculate urgent bandwidth required, both urg and non urg peak bandwidth - // assume flip bw is 0 at this point - for (k = 0; k < mode_lib->ms.num_active_planes; k++) - mode_lib->ms.final_flip_bw[k] = 0; - - calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = mode_lib->ms.support.urg_vactive_bandwidth_required; - calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required; - calculate_peak_bandwidth_params->urg_bandwidth_required_qual = mode_lib->ms.support.urg_bandwidth_required_qual; - calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required; - calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = mode_lib->ms.surface_avg_vactive_required_bw; - calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw; - - calculate_peak_bandwidth_params->display_cfg = display_cfg; - calculate_peak_bandwidth_params->inc_flip_bw = 0; - calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes; - calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP; - calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0; - calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1; - calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0; - calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1; - calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor; - calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor; - - calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l; - calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c; - calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma; - calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma; - calculate_peak_bandwidth_params->prefetch_bandwidth_oto = mode_lib->ms.RequiredPrefetchBWOTO; - calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l; - calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c; - calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw; - calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw; - calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw; - calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw; - calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw; - calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw; - calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma; - calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma; - calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor; - calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre; - calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre; - calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre; - - calculate_peak_bandwidth_required( - &mode_lib->scratch, - calculate_peak_bandwidth_params); - - // Check urg peak bandwidth against available urg bw - // check at SDP and DRAM, for all soc states (SVP prefetch an Sys Active) - check_urgent_bandwidth_support( - &s->dummy_single[0], // double* frac_urg_bandwidth - &s->dummy_single[1], // double* frac_urg_bandwidth_mall - &mode_lib->ms.support.UrgVactiveBandwidthSupport, - &mode_lib->ms.support.PrefetchBandwidthSupported, - - mode_lib->soc.mall_allocated_for_dcn_mbytes, - mode_lib->ms.support.non_urg_bandwidth_required, - mode_lib->ms.support.urg_vactive_bandwidth_required, - mode_lib->ms.support.urg_bandwidth_required, - mode_lib->ms.support.urg_bandwidth_available); - - mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.PrefetchBandwidthSupported; - dml2_printf("DML::%s: PrefetchBandwidthSupported=%0d\n", __func__, mode_lib->ms.support.PrefetchBandwidthSupported); - - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - if (mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]) { - mode_lib->ms.support.PrefetchSupported = false; - dml2_printf("DML::%s: k=%d, NotEnoughUrgentLatencyHidingPre=%d\n", __func__, k, mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]); - } - } - -#ifdef DML_GLOBAL_PREFETCH_CHECK - if (mode_lib->ms.support.PrefetchSupported && mode_lib->ms.num_active_planes > 1 && s->recalc_prefetch_done == 0) { - CheckGlobalPrefetchAdmissibility_params->num_active_planes = mode_lib->ms.num_active_planes; - CheckGlobalPrefetchAdmissibility_params->pixel_format = s->pixel_format; - CheckGlobalPrefetchAdmissibility_params->chunk_bytes_l = mode_lib->ip.pixel_chunk_size_kbytes * 1024; - CheckGlobalPrefetchAdmissibility_params->chunk_bytes_c = mode_lib->ip.pixel_chunk_size_kbytes * 1024; - CheckGlobalPrefetchAdmissibility_params->lb_source_lines_l = s->lb_source_lines_l; - CheckGlobalPrefetchAdmissibility_params->lb_source_lines_c = s->lb_source_lines_c; - CheckGlobalPrefetchAdmissibility_params->swath_height_l = mode_lib->ms.SwathHeightY; - CheckGlobalPrefetchAdmissibility_params->swath_height_c = mode_lib->ms.SwathHeightC; - CheckGlobalPrefetchAdmissibility_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes; - CheckGlobalPrefetchAdmissibility_params->compressed_buffer_size_kbytes = mode_lib->ms.CompressedBufferSizeInkByte; - CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_l = mode_lib->ms.DETBufferSizeY; - CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_c = mode_lib->ms.DETBufferSizeC; - CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_l = s->full_swath_bytes_l; - CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_c = s->full_swath_bytes_c; - CheckGlobalPrefetchAdmissibility_params->prefetch_sw_bytes = s->prefetch_sw_bytes; - CheckGlobalPrefetchAdmissibility_params->Tpre_rounded = s->Tpre_rounded; - CheckGlobalPrefetchAdmissibility_params->Tpre_oto = s->Tpre_oto; - CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = mode_lib->ms.support.urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]; - CheckGlobalPrefetchAdmissibility_params->line_time = s->line_times; - CheckGlobalPrefetchAdmissibility_params->dst_y_prefetch = mode_lib->ms.dst_y_prefetch; - if (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps < 10 * 1024) - CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = 10 * 1024; - - CheckGlobalPrefetchAdmissibility_params->estimated_dcfclk_mhz = (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps / (double) mode_lib->soc.return_bus_width_bytes) / - ((double)mode_lib->soc.qos_parameters.derate_table.system_active_urgent.dcfclk_derate_percent / 100.0); - - // if recalc_prefetch_schedule is set, recalculate the prefetch schedule with the new impacted_Tpre, prefetch should be possible - CheckGlobalPrefetchAdmissibility_params->recalc_prefetch_schedule = &s->recalc_prefetch_schedule; - CheckGlobalPrefetchAdmissibility_params->impacted_dst_y_pre = s->impacted_dst_y_pre; - mode_lib->ms.support.PrefetchSupported = CheckGlobalPrefetchAdmissibility(&mode_lib->scratch, CheckGlobalPrefetchAdmissibility_params); - s->recalc_prefetch_done = 1; - s->recalc_prefetch_schedule = 1; - } -#endif - } // prefetch schedule ok, do urg bw and flip schedule - } while (s->recalc_prefetch_schedule); - - // Flip Schedule - // Both prefetch schedule and BW okay - if (mode_lib->ms.support.PrefetchSupported == true) { - mode_lib->ms.BandwidthAvailableForImmediateFlip = - get_bandwidth_available_for_immediate_flip( - dml2_core_internal_soc_state_sys_active, - mode_lib->ms.support.urg_bandwidth_required_qual, // no flip - mode_lib->ms.support.urg_bandwidth_available); - - mode_lib->ms.TotImmediateFlipBytes = 0; - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - if (display_cfg->plane_descriptors[k].immediate_flip) { - s->per_pipe_flip_bytes[k] = get_pipe_flip_bytes( - s->HostVMInefficiencyFactor, - mode_lib->ms.vm_bytes[k], - mode_lib->ms.DPTEBytesPerRow[k], - mode_lib->ms.meta_row_bytes[k]); - } else { - s->per_pipe_flip_bytes[k] = 0; - } - mode_lib->ms.TotImmediateFlipBytes += s->per_pipe_flip_bytes[k] * mode_lib->ms.NoOfDPP[k]; - - } - - for (k = 0; k < mode_lib->ms.num_active_planes; k++) { - CalculateFlipSchedule( - &mode_lib->scratch, - display_cfg->plane_descriptors[k].immediate_flip, - 1, // use_lb_flip_bw - s->HostVMInefficiencyFactor, - s->Tvm_trips_flip[k], - s->Tr0_trips_flip[k], - s->Tvm_trips_flip_rounded[k], - s->Tr0_trips_flip_rounded[k], - display_cfg->gpuvm_enable, - mode_lib->ms.vm_bytes[k], - mode_lib->ms.DPTEBytesPerRow[k], - mode_lib->ms.BandwidthAvailableForImmediateFlip, - mode_lib->ms.TotImmediateFlipBytes, - display_cfg->plane_descriptors[k].pixel_format, - (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)), - display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio, - display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio, - mode_lib->ms.Tno_bw_flip[k], - mode_lib->ms.dpte_row_height[k], - mode_lib->ms.dpte_row_height_chroma[k], - mode_lib->ms.use_one_row_for_frame_flip[k], - mode_lib->ip.max_flip_time_us, - mode_lib->ip.max_flip_time_lines, - s->per_pipe_flip_bytes[k], - mode_lib->ms.meta_row_bytes[k], - s->meta_row_height_luma[k], - s->meta_row_height_chroma[k], - mode_lib->ip.dcn_mrq_present && display_cfg->plane_descriptors[k].surface.dcc.enable, - - /* Output */ - &mode_lib->ms.dst_y_per_vm_flip[k], - &mode_lib->ms.dst_y_per_row_flip[k], - &mode_lib->ms.final_flip_bw[k], - &mode_lib->ms.ImmediateFlipSupportedForPipe[k]); - } - - calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = s->dummy_bw; - calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required_flip; - calculate_peak_bandwidth_params->urg_bandwidth_required_qual = s->dummy_bw; - calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required_flip; - calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = s->surface_dummy_bw; - calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw; - - calculate_peak_bandwidth_params->display_cfg = display_cfg; - calculate_peak_bandwidth_params->inc_flip_bw = 1; - calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes; - calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP; - calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0; - calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1; - calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0; - calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1; - calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor; - calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor; - - calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l; - calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c; - calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma; - calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma; - calculate_peak_bandwidth_params->prefetch_bandwidth_oto = mode_lib->ms.RequiredPrefetchBWOTO; - calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l; - calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c; - calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw; - calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw; - calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw; - calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw; - calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw; - calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw; - calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma; - calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma; - calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor; - calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre; - calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre; - calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre; - - calculate_peak_bandwidth_required( - &mode_lib->scratch, - calculate_peak_bandwidth_params); - - calculate_immediate_flip_bandwidth_support( - &s->dummy_single[0], // double* frac_urg_bandwidth_flip - &mode_lib->ms.support.ImmediateFlipSupport, - - dml2_core_internal_soc_state_sys_active, - mode_lib->ms.support.urg_bandwidth_required_flip, - mode_lib->ms.support.non_urg_bandwidth_required_flip, - mode_lib->ms.support.urg_bandwidth_available); - - for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) { - if (display_cfg->plane_descriptors[k].immediate_flip == true && mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false) - mode_lib->ms.support.ImmediateFlipSupport = false; - } - - } else { // if prefetch not support, assume iflip is not supported too - mode_lib->ms.support.ImmediateFlipSupport = false; - } - - s->mSOCParameters.UrgentLatency = mode_lib->ms.UrgLatency; - s->mSOCParameters.ExtraLatency = mode_lib->ms.ExtraLatency; - s->mSOCParameters.ExtraLatency_sr = mode_lib->ms.ExtraLatency_sr; - s->mSOCParameters.WritebackLatency = mode_lib->soc.qos_parameters.writeback.base_latency_us; - s->mSOCParameters.DRAMClockChangeLatency = mode_lib->soc.power_management_parameters.dram_clk_change_blackout_us; - s->mSOCParameters.FCLKChangeLatency = mode_lib->soc.power_management_parameters.fclk_change_blackout_us; - s->mSOCParameters.SRExitTime = mode_lib->soc.power_management_parameters.stutter_exit_latency_us; - s->mSOCParameters.SREnterPlusExitTime = mode_lib->soc.power_management_parameters.stutter_enter_plus_exit_latency_us; - s->mSOCParameters.SRExitZ8Time = mode_lib->soc.power_management_parameters.z8_stutter_exit_latency_us; - s->mSOCParameters.SREnterPlusExitZ8Time = mode_lib->soc.power_management_parameters.z8_stutter_enter_plus_exit_latency_us; - s->mSOCParameters.USRRetrainingLatency = 0; - s->mSOCParameters.SMNLatency = 0; - s->mSOCParameters.g6_temp_read_blackout_us = get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), in_out_params->min_clk_index); - s->mSOCParameters.max_urgent_latency_us = get_max_urgent_latency_us(&mode_lib->soc.qos_parameters.qos_params.dcn4x, mode_lib->ms.uclk_freq_mhz, mode_lib->ms.FabricClock, in_out_params->min_clk_index); - s->mSOCParameters.df_response_time_us = mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles / mode_lib->ms.FabricClock; - s->mSOCParameters.qos_type = mode_lib->soc.qos_parameters.qos_type; - - CalculateWatermarks_params->display_cfg = display_cfg; - CalculateWatermarks_params->USRRetrainingRequired = false; - CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes; - CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ip.max_line_buffer_lines; - CalculateWatermarks_params->LineBufferSize = mode_lib->ip.line_buffer_size_bits; - CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ip.writeback_interface_buffer_size_kbytes; - CalculateWatermarks_params->DCFCLK = mode_lib->ms.DCFCLK; - CalculateWatermarks_params->SynchronizeTimings = display_cfg->overrides.synchronize_timings; - CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChange = display_cfg->overrides.synchronize_ddr_displays_for_uclk_pstate_change; - CalculateWatermarks_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes; - CalculateWatermarks_params->mmSOCParameters = s->mSOCParameters; - CalculateWatermarks_params->WritebackChunkSize = mode_lib->ip.writeback_chunk_size_kbytes; - CalculateWatermarks_params->SOCCLK = mode_lib->ms.SOCCLK; - CalculateWatermarks_params->DCFClkDeepSleep = mode_lib->ms.dcfclk_deepsleep; - CalculateWatermarks_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeY; - CalculateWatermarks_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeC; - CalculateWatermarks_params->SwathHeightY = mode_lib->ms.SwathHeightY; - CalculateWatermarks_params->SwathHeightC = mode_lib->ms.SwathHeightC; - CalculateWatermarks_params->SwathWidthY = mode_lib->ms.SwathWidthY; - CalculateWatermarks_params->SwathWidthC = mode_lib->ms.SwathWidthC; - CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.NoOfDPP; - CalculateWatermarks_params->BytePerPixelDETY = mode_lib->ms.BytePerPixelInDETY; - CalculateWatermarks_params->BytePerPixelDETC = mode_lib->ms.BytePerPixelInDETC; - CalculateWatermarks_params->DSTXAfterScaler = s->DSTXAfterScaler; - CalculateWatermarks_params->DSTYAfterScaler = s->DSTYAfterScaler; - CalculateWatermarks_params->UnboundedRequestEnabled = mode_lib->ms.UnboundedRequestEnabled; - CalculateWatermarks_params->CompressedBufferSizeInkByte = mode_lib->ms.CompressedBufferSizeInkByte; - CalculateWatermarks_params->meta_row_height_l = s->meta_row_height_luma; - CalculateWatermarks_params->meta_row_height_c = s->meta_row_height_chroma; - - // Output - CalculateWatermarks_params->Watermark = &mode_lib->ms.support.watermarks; // Watermarks *Watermark - CalculateWatermarks_params->DRAMClockChangeSupport = mode_lib->ms.support.DRAMClockChangeSupport; - CalculateWatermarks_params->global_dram_clock_change_supported = &mode_lib->ms.support.global_dram_clock_change_supported; - CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0]; // double *MaxActiveDRAMClockChangeLatencySupported[] - CalculateWatermarks_params->SubViewportLinesNeededInMALL = mode_lib->ms.SubViewportLinesNeededInMALL; // unsigned int SubViewportLinesNeededInMALL[] - CalculateWatermarks_params->FCLKChangeSupport = mode_lib->ms.support.FCLKChangeSupport; - CalculateWatermarks_params->global_fclk_change_supported = &mode_lib->ms.support.global_fclk_change_supported; - CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // double *MaxActiveFCLKChangeLatencySupported - CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport; - CalculateWatermarks_params->g6_temp_read_support = &mode_lib->ms.support.g6_temp_read_support; - CalculateWatermarks_params->VActiveLatencyHidingMargin = mode_lib->ms.VActiveLatencyHidingMargin; - CalculateWatermarks_params->VActiveLatencyHidingUs = mode_lib->ms.VActiveLatencyHidingUs; - - CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(&mode_lib->scratch, CalculateWatermarks_params); - - calculate_pstate_keepout_dst_lines(display_cfg, &mode_lib->ms.support.watermarks, s->dummy_integer_array[0]); - } - dml2_printf("DML::%s: Done prefetch calculation\n", __func__); - // End of Prefetch Check + dml_core_ms_prefetch_check(mode_lib, display_cfg); mode_lib->ms.support.max_urgent_latency_us = s->mSOCParameters.max_urgent_latency_us; @@ -9546,8 +9475,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.dram_change_vactive_det_fill_delay_us); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: max_urgent_latency_us = %f\n", __func__, s->mSOCParameters.max_urgent_latency_us); - dml2_printf("DML::%s: ROBSupport = %u\n", __func__, mode_lib->ms.support.ROBSupport); + DML_LOG_VERBOSE("DML::%s: max_urgent_latency_us = %f\n", __func__, s->mSOCParameters.max_urgent_latency_us); + DML_LOG_VERBOSE("DML::%s: ROBSupport = %u\n", __func__, mode_lib->ms.support.ROBSupport); #endif /*Mode Support, Voltage State and SOC Configuration*/ @@ -9597,17 +9526,17 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out && !mode_lib->ms.support.ExceededMALLSize && mode_lib->ms.support.g6_temp_read_support && ((!display_cfg->hostvm_enable && !s->ImmediateFlipRequired) || mode_lib->ms.support.ImmediateFlipSupport)) { - dml2_printf("DML::%s: mode is supported\n", __func__); + DML_LOG_VERBOSE("DML::%s: mode is supported\n", __func__); mode_lib->ms.support.ModeSupport = true; } else { - dml2_printf("DML::%s: mode is NOT supported\n", __func__); + DML_LOG_VERBOSE("DML::%s: mode is NOT supported\n", __func__); mode_lib->ms.support.ModeSupport = false; } } // Since now the mode_support work on 1 particular power state, so there is only 1 state idx (index 0). - dml2_printf("DML::%s: ModeSupport = %u\n", __func__, mode_lib->ms.support.ModeSupport); - dml2_printf("DML::%s: ImmediateFlipSupport = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport); + DML_LOG_VERBOSE("DML::%s: ModeSupport = %u\n", __func__, mode_lib->ms.support.ModeSupport); + DML_LOG_VERBOSE("DML::%s: ImmediateFlipSupport = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport); for (k = 0; k < mode_lib->ms.num_active_planes; k++) { mode_lib->ms.support.MPCCombineEnable[k] = mode_lib->ms.MPCCombine[k]; @@ -9623,8 +9552,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.support.OutputRate[k] = mode_lib->ms.OutputRate[k]; #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: k=%d, ODMMode = %u\n", __func__, k, mode_lib->ms.support.ODMMode[k]); - dml2_printf("DML::%s: k=%d, DSCEnabled = %u\n", __func__, k, mode_lib->ms.support.DSCEnabled[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, ODMMode = %u\n", __func__, k, mode_lib->ms.support.ODMMode[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, DSCEnabled = %u\n", __func__, k, mode_lib->ms.support.DSCEnabled[k]); #endif } @@ -9632,7 +9561,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out if (!mode_lib->ms.support.ModeSupport) dml2_print_mode_support_info(&mode_lib->ms.support, true); - dml2_printf("DML::%s: --- DONE --- \n", __func__); + DML_LOG_VERBOSE("DML::%s: --- DONE --- \n", __func__); #endif return mode_lib->ms.support.ModeSupport; @@ -9642,18 +9571,18 @@ unsigned int dml2_core_calcs_mode_support_ex(struct dml2_core_calcs_mode_support { unsigned int result; - dml2_printf("DML::%s: ------------- START ----------\n", __func__); + DML_LOG_VERBOSE("DML::%s: ------------- START ----------\n", __func__); result = dml_core_mode_support(in_out_params); if (result) *in_out_params->out_evaluation_info = in_out_params->mode_lib->ms.support; - dml2_printf("DML::%s: is_mode_support = %u (min_clk_index=%d)\n", __func__, result, in_out_params->min_clk_index); + DML_LOG_VERBOSE("DML::%s: is_mode_support = %u (min_clk_index=%d)\n", __func__, result, in_out_params->min_clk_index); for (unsigned int k = 0; k < in_out_params->in_display_cfg->num_planes; k++) - dml2_printf("DML::%s: plane_%d: reserved_vblank_time_ns = %u\n", __func__, k, in_out_params->in_display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns); + DML_LOG_VERBOSE("DML::%s: plane_%d: reserved_vblank_time_ns = %lu\n", __func__, k, in_out_params->in_display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns); - dml2_printf("DML::%s: ------------- DONE ----------\n", __func__); + DML_LOG_VERBOSE("DML::%s: ------------- DONE ----------\n", __func__); return result; } @@ -9687,19 +9616,19 @@ static void CalculatePixelDeliveryTimes( double pixel_clock_mhz = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u : HRatio = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio); - dml2_printf("DML::%s: k=%u : VRatio = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); - dml2_printf("DML::%s: k=%u : HRatioChroma = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio); - dml2_printf("DML::%s: k=%u : VRatioChroma = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio); - dml2_printf("DML::%s: k=%u : VRatioPrefetchY = %f\n", __func__, k, VRatioPrefetchY[k]); - dml2_printf("DML::%s: k=%u : VRatioPrefetchC = %f\n", __func__, k, VRatioPrefetchC[k]); - dml2_printf("DML::%s: k=%u : swath_width_luma_ub = %u\n", __func__, k, swath_width_luma_ub[k]); - dml2_printf("DML::%s: k=%u : swath_width_chroma_ub = %u\n", __func__, k, swath_width_chroma_ub[k]); - dml2_printf("DML::%s: k=%u : PSCL_THROUGHPUT = %f\n", __func__, k, PSCL_THROUGHPUT[k]); - dml2_printf("DML::%s: k=%u : PSCL_THROUGHPUT_CHROMA = %f\n", __func__, k, PSCL_THROUGHPUT_CHROMA[k]); - dml2_printf("DML::%s: k=%u : DPPPerSurface = %u\n", __func__, k, cfg_support_info->plane_support_info[k].dpps_used); - dml2_printf("DML::%s: k=%u : pixel_clock_mhz = %f\n", __func__, k, pixel_clock_mhz); - dml2_printf("DML::%s: k=%u : Dppclk = %f\n", __func__, k, Dppclk[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : HRatio = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u : VRatio = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u : HRatioChroma = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u : VRatioChroma = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u : VRatioPrefetchY = %f\n", __func__, k, VRatioPrefetchY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : VRatioPrefetchC = %f\n", __func__, k, VRatioPrefetchC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : swath_width_luma_ub = %u\n", __func__, k, swath_width_luma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : swath_width_chroma_ub = %u\n", __func__, k, swath_width_chroma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : PSCL_THROUGHPUT = %f\n", __func__, k, PSCL_THROUGHPUT[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : PSCL_THROUGHPUT_CHROMA = %f\n", __func__, k, PSCL_THROUGHPUT_CHROMA[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DPPPerSurface = %u\n", __func__, k, cfg_support_info->plane_support_info[k].dpps_used); + DML_LOG_VERBOSE("DML::%s: k=%u : pixel_clock_mhz = %f\n", __func__, k, pixel_clock_mhz); + DML_LOG_VERBOSE("DML::%s: k=%u : Dppclk = %f\n", __func__, k, Dppclk[k]); #endif if (display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio <= 1) { DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] * cfg_support_info->plane_support_info[k].dpps_used / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio / pixel_clock_mhz; @@ -9733,10 +9662,10 @@ static void CalculatePixelDeliveryTimes( } } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLuma[k]); - dml2_printf("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]); - dml2_printf("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChroma[k]); - dml2_printf("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]); #endif } @@ -9752,12 +9681,12 @@ static void CalculatePixelDeliveryTimes( DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = DisplayPipeLineDeliveryTimeChromaPrefetch[k] / req_per_swath_ub_c[k]; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLuma[k]); - dml2_printf("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]); - dml2_printf("DML::%s: k=%u : req_per_swath_ub_l = %d\n", __func__, k, req_per_swath_ub_l[k]); - dml2_printf("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChroma[k]); - dml2_printf("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]); - dml2_printf("DML::%s: k=%u : req_per_swath_ub_c = %d\n", __func__, k, req_per_swath_ub_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : req_per_swath_ub_l = %d\n", __func__, k, req_per_swath_ub_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u : req_per_swath_ub_c = %d\n", __func__, k, req_per_swath_ub_c[k]); #endif } } @@ -9853,14 +9782,14 @@ static void CalculateMetaAndPTETimes(struct dml2_core_shared_CalculateMetaAndPTE } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d, DST_Y_PER_META_ROW_NOM_L = %f\n", __func__, k, p->DST_Y_PER_META_ROW_NOM_L[k]); - dml2_printf("DML::%s: k=%d, DST_Y_PER_META_ROW_NOM_C = %f\n", __func__, k, p->DST_Y_PER_META_ROW_NOM_C[k]); - dml2_printf("DML::%s: k=%d, TimePerMetaChunkNominal = %f\n", __func__, k, p->TimePerMetaChunkNominal[k]); - dml2_printf("DML::%s: k=%d, TimePerMetaChunkVBlank = %f\n", __func__, k, p->TimePerMetaChunkVBlank[k]); - dml2_printf("DML::%s: k=%d, TimePerMetaChunkFlip = %f\n", __func__, k, p->TimePerMetaChunkFlip[k]); - dml2_printf("DML::%s: k=%d, TimePerChromaMetaChunkNominal = %f\n", __func__, k, p->TimePerChromaMetaChunkNominal[k]); - dml2_printf("DML::%s: k=%d, TimePerChromaMetaChunkVBlank = %f\n", __func__, k, p->TimePerChromaMetaChunkVBlank[k]); - dml2_printf("DML::%s: k=%d, TimePerChromaMetaChunkFlip = %f\n", __func__, k, p->TimePerChromaMetaChunkFlip[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, DST_Y_PER_META_ROW_NOM_L = %f\n", __func__, k, p->DST_Y_PER_META_ROW_NOM_L[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, DST_Y_PER_META_ROW_NOM_C = %f\n", __func__, k, p->DST_Y_PER_META_ROW_NOM_C[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, TimePerMetaChunkNominal = %f\n", __func__, k, p->TimePerMetaChunkNominal[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, TimePerMetaChunkVBlank = %f\n", __func__, k, p->TimePerMetaChunkVBlank[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, TimePerMetaChunkFlip = %f\n", __func__, k, p->TimePerMetaChunkFlip[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, TimePerChromaMetaChunkNominal = %f\n", __func__, k, p->TimePerChromaMetaChunkNominal[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, TimePerChromaMetaChunkVBlank = %f\n", __func__, k, p->TimePerChromaMetaChunkVBlank[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, TimePerChromaMetaChunkFlip = %f\n", __func__, k, p->TimePerChromaMetaChunkFlip[k]); #endif } @@ -9881,7 +9810,7 @@ static void CalculateMetaAndPTETimes(struct dml2_core_shared_CalculateMetaAndPTE else p->time_per_tdlut_group[k] = 0; - dml2_printf("DML::%s: k=%u, time_per_tdlut_group = %f\n", __func__, k, p->time_per_tdlut_group[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_tdlut_group = %f\n", __func__, k, p->time_per_tdlut_group[k]); if (p->display_cfg->gpuvm_enable == true) { if (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) { @@ -9897,14 +9826,14 @@ static void CalculateMetaAndPTETimes(struct dml2_core_shared_CalculateMetaAndPTE if (dpte_groups_per_row_luma_ub <= 2) { dpte_groups_per_row_luma_ub = dpte_groups_per_row_luma_ub + 1; } - dml2_printf("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]); - dml2_printf("DML::%s: k=%u, dpte_group_bytes = %u\n", __func__, k, p->dpte_group_bytes[k]); - dml2_printf("DML::%s: k=%u, PTERequestSizeY = %u\n", __func__, k, p->PTERequestSizeY[k]); - dml2_printf("DML::%s: k=%u, PixelPTEReqWidthY = %u\n", __func__, k, p->PixelPTEReqWidthY[k]); - dml2_printf("DML::%s: k=%u, PixelPTEReqHeightY = %u\n", __func__, k, p->PixelPTEReqHeightY[k]); - dml2_printf("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]); - dml2_printf("DML::%s: k=%u, dpte_group_width_luma = %u\n", __func__, k, dpte_group_width_luma); - dml2_printf("DML::%s: k=%u, dpte_groups_per_row_luma_ub = %u\n", __func__, k, dpte_groups_per_row_luma_ub); + DML_LOG_VERBOSE("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_group_bytes = %u\n", __func__, k, p->dpte_group_bytes[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PTERequestSizeY = %u\n", __func__, k, p->PTERequestSizeY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEReqWidthY = %u\n", __func__, k, p->PixelPTEReqWidthY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEReqHeightY = %u\n", __func__, k, p->PixelPTEReqHeightY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_group_width_luma = %u\n", __func__, k, dpte_group_width_luma); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_groups_per_row_luma_ub = %u\n", __func__, k, dpte_groups_per_row_luma_ub); p->time_per_pte_group_nom_luma[k] = p->DST_Y_PER_PTE_ROW_NOM_L[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_luma_ub; p->time_per_pte_group_vblank_luma[k] = p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_luma_ub; @@ -9928,9 +9857,9 @@ static void CalculateMetaAndPTETimes(struct dml2_core_shared_CalculateMetaAndPTE if (dpte_groups_per_row_chroma_ub <= 2) { dpte_groups_per_row_chroma_ub = dpte_groups_per_row_chroma_ub + 1; } - dml2_printf("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]); - dml2_printf("DML::%s: k=%u, dpte_group_width_chroma = %u\n", __func__, k, dpte_group_width_chroma); - dml2_printf("DML::%s: k=%u, dpte_groups_per_row_chroma_ub = %u\n", __func__, k, dpte_groups_per_row_chroma_ub); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_group_width_chroma = %u\n", __func__, k, dpte_group_width_chroma); + DML_LOG_VERBOSE("DML::%s: k=%u, dpte_groups_per_row_chroma_ub = %u\n", __func__, k, dpte_groups_per_row_chroma_ub); p->time_per_pte_group_nom_chroma[k] = p->DST_Y_PER_PTE_ROW_NOM_C[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_chroma_ub; p->time_per_pte_group_vblank_chroma[k] = p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_chroma_ub; @@ -9945,17 +9874,17 @@ static void CalculateMetaAndPTETimes(struct dml2_core_shared_CalculateMetaAndPTE p->time_per_pte_group_flip_chroma[k] = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, dst_y_per_row_vblank = %f\n", __func__, k, p->dst_y_per_row_vblank[k]); - dml2_printf("DML::%s: k=%u, dst_y_per_row_flip = %f\n", __func__, k, p->dst_y_per_row_flip[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_row_vblank = %f\n", __func__, k, p->dst_y_per_row_vblank[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_row_flip = %f\n", __func__, k, p->dst_y_per_row_flip[k]); - dml2_printf("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_L = %f\n", __func__, k, p->DST_Y_PER_PTE_ROW_NOM_L[k]); - dml2_printf("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_C = %f\n", __func__, k, p->DST_Y_PER_PTE_ROW_NOM_C[k]); - dml2_printf("DML::%s: k=%u, time_per_pte_group_nom_luma = %f\n", __func__, k, p->time_per_pte_group_nom_luma[k]); - dml2_printf("DML::%s: k=%u, time_per_pte_group_vblank_luma = %f\n", __func__, k, p->time_per_pte_group_vblank_luma[k]); - dml2_printf("DML::%s: k=%u, time_per_pte_group_flip_luma = %f\n", __func__, k, p->time_per_pte_group_flip_luma[k]); - dml2_printf("DML::%s: k=%u, time_per_pte_group_nom_chroma = %f\n", __func__, k, p->time_per_pte_group_nom_chroma[k]); - dml2_printf("DML::%s: k=%u, time_per_pte_group_vblank_chroma = %f\n", __func__, k, p->time_per_pte_group_vblank_chroma[k]); - dml2_printf("DML::%s: k=%u, time_per_pte_group_flip_chroma = %f\n", __func__, k, p->time_per_pte_group_flip_chroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_L = %f\n", __func__, k, p->DST_Y_PER_PTE_ROW_NOM_L[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_C = %f\n", __func__, k, p->DST_Y_PER_PTE_ROW_NOM_C[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_nom_luma = %f\n", __func__, k, p->time_per_pte_group_nom_luma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_vblank_luma = %f\n", __func__, k, p->time_per_pte_group_vblank_luma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_flip_luma = %f\n", __func__, k, p->time_per_pte_group_flip_luma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_nom_chroma = %f\n", __func__, k, p->time_per_pte_group_nom_chroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_vblank_chroma = %f\n", __func__, k, p->time_per_pte_group_vblank_chroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_flip_chroma = %f\n", __func__, k, p->time_per_pte_group_flip_chroma[k]); #endif } } // CalculateMetaAndPTETimes @@ -9991,18 +9920,18 @@ static void CalculateVMGroupAndRequestTimes( double line_time; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces); + DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces); #endif for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { double pixel_clock_mhz = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); bool dcc_mrq_enable = display_cfg->plane_descriptors[k].surface.dcc.enable && mrq_present; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, dcc_mrq_enable = %u\n", __func__, k, dcc_mrq_enable); - dml2_printf("DML::%s: k=%u, vm_group_bytes = %u\n", __func__, k, vm_group_bytes[k]); - dml2_printf("DML::%s: k=%u, dpde0_bytes_per_frame_ub_l = %u\n", __func__, k, dpde0_bytes_per_frame_ub_l[k]); - dml2_printf("DML::%s: k=%u, dpde0_bytes_per_frame_ub_c = %u\n", __func__, k, dpde0_bytes_per_frame_ub_c[k]); - dml2_printf("DML::%s: k=%d, meta_pte_bytes_per_frame_ub_l = %d\n", __func__, k, meta_pte_bytes_per_frame_ub_l[k]); - dml2_printf("DML::%s: k=%d, meta_pte_bytes_per_frame_ub_c = %d\n", __func__, k, meta_pte_bytes_per_frame_ub_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dcc_mrq_enable = %u\n", __func__, k, dcc_mrq_enable); + DML_LOG_VERBOSE("DML::%s: k=%u, vm_group_bytes = %u\n", __func__, k, vm_group_bytes[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpde0_bytes_per_frame_ub_l = %u\n", __func__, k, dpde0_bytes_per_frame_ub_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dpde0_bytes_per_frame_ub_c = %u\n", __func__, k, dpde0_bytes_per_frame_ub_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, meta_pte_bytes_per_frame_ub_l = %d\n", __func__, k, meta_pte_bytes_per_frame_ub_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%d, meta_pte_bytes_per_frame_ub_c = %d\n", __func__, k, meta_pte_bytes_per_frame_ub_c[k]); #endif if (display_cfg->gpuvm_enable) { @@ -10071,13 +10000,13 @@ static void CalculateVMGroupAndRequestTimes( else TimePerVMRequestFlip[k] = 0.0; - dml2_printf("DML::%s: k=%u, dst_y_per_vm_vblank = %f\n", __func__, k, dst_y_per_vm_vblank[k]); - dml2_printf("DML::%s: k=%u, dst_y_per_vm_flip = %f\n", __func__, k, dst_y_per_vm_flip[k]); - dml2_printf("DML::%s: k=%u, line_time = %f\n", __func__, k, line_time); - dml2_printf("DML::%s: k=%u, num_group_per_lower_vm_stage_pref = %f\n", __func__, k, num_group_per_lower_vm_stage_pref); - dml2_printf("DML::%s: k=%u, num_group_per_lower_vm_stage_flip = %f\n", __func__, k, num_group_per_lower_vm_stage_flip); - dml2_printf("DML::%s: k=%u, num_req_per_lower_vm_stage_pref = %f\n", __func__, k, num_req_per_lower_vm_stage_pref); - dml2_printf("DML::%s: k=%u, num_req_per_lower_vm_stage_flip = %f\n", __func__, k, num_req_per_lower_vm_stage_flip); + DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_vm_vblank = %f\n", __func__, k, dst_y_per_vm_vblank[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_vm_flip = %f\n", __func__, k, dst_y_per_vm_flip[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, line_time = %f\n", __func__, k, line_time); + DML_LOG_VERBOSE("DML::%s: k=%u, num_group_per_lower_vm_stage_pref = %d\n", __func__, k, num_group_per_lower_vm_stage_pref); + DML_LOG_VERBOSE("DML::%s: k=%u, num_group_per_lower_vm_stage_flip = %d\n", __func__, k, num_group_per_lower_vm_stage_flip); + DML_LOG_VERBOSE("DML::%s: k=%u, num_req_per_lower_vm_stage_pref = %d\n", __func__, k, num_req_per_lower_vm_stage_pref); + DML_LOG_VERBOSE("DML::%s: k=%u, num_req_per_lower_vm_stage_flip = %d\n", __func__, k, num_req_per_lower_vm_stage_flip); if (display_cfg->gpuvm_max_page_table_levels > 2) { TimePerVMGroupVBlank[k] = TimePerVMGroupVBlank[k] / 2; @@ -10094,10 +10023,10 @@ static void CalculateVMGroupAndRequestTimes( } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, TimePerVMGroupVBlank = %f\n", __func__, k, TimePerVMGroupVBlank[k]); - dml2_printf("DML::%s: k=%u, TimePerVMGroupFlip = %f\n", __func__, k, TimePerVMGroupFlip[k]); - dml2_printf("DML::%s: k=%u, TimePerVMRequestVBlank = %f\n", __func__, k, TimePerVMRequestVBlank[k]); - dml2_printf("DML::%s: k=%u, TimePerVMRequestFlip = %f\n", __func__, k, TimePerVMRequestFlip[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMGroupVBlank = %f\n", __func__, k, TimePerVMGroupVBlank[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMGroupFlip = %f\n", __func__, k, TimePerVMGroupFlip[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMRequestVBlank = %f\n", __func__, k, TimePerVMRequestVBlank[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMRequestFlip = %f\n", __func__, k, TimePerVMRequestFlip[k]); #endif } } @@ -10113,7 +10042,6 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc unsigned int SingleVTotal = 0; bool SameTiming = true; bool FoundCriticalSurface = false; - double LastZ8StutterPeriod = 0; memset(l, 0, sizeof(struct dml2_core_calcs_CalculateStutterEfficiency_locals)); @@ -10127,9 +10055,9 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc } l->TotalCompressedReadBandwidth = l->TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] / math_min2(p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane0, l->MaximumEffectiveCompressionLuma); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]); - dml2_printf("DML::%s: k=%u, NetDCCRateLuma = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane0); - dml2_printf("DML::%s: k=%u, MaximumEffectiveCompressionLuma = %f\n", __func__, k, l->MaximumEffectiveCompressionLuma); + DML_LOG_VERBOSE("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, NetDCCRateLuma = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane0); + DML_LOG_VERBOSE("DML::%s: k=%u, MaximumEffectiveCompressionLuma = %f\n", __func__, k, l->MaximumEffectiveCompressionLuma); #endif l->TotalZeroSizeRequestReadBandwidth = l->TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane0; l->TotalZeroSizeCompressedReadBandwidth = l->TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane0 / l->MaximumEffectiveCompressionLuma; @@ -10142,9 +10070,9 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc } l->TotalCompressedReadBandwidth = l->TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] / math_min2(p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane1, l->MaximumEffectiveCompressionChroma); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, p->ReadBandwidthSurfaceChroma[k]); - dml2_printf("DML::%s: k=%u, NetDCCRateChroma = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane1); - dml2_printf("DML::%s: k=%u, MaximumEffectiveCompressionChroma = %f\n", __func__, k, l->MaximumEffectiveCompressionChroma); + DML_LOG_VERBOSE("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, p->ReadBandwidthSurfaceChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, NetDCCRateChroma = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane1); + DML_LOG_VERBOSE("DML::%s: k=%u, MaximumEffectiveCompressionChroma = %f\n", __func__, k, l->MaximumEffectiveCompressionChroma); #endif l->TotalZeroSizeRequestReadBandwidth = l->TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane1; l->TotalZeroSizeCompressedReadBandwidth = l->TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane1 / l->MaximumEffectiveCompressionChroma; @@ -10160,19 +10088,19 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc l->AverageDCCZeroSizeFraction = l->TotalZeroSizeRequestReadBandwidth / p->TotalDataReadBandwidth; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: UnboundedRequestEnabled = %u\n", __func__, p->UnboundedRequestEnabled); - dml2_printf("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, l->TotalCompressedReadBandwidth); - dml2_printf("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, l->TotalZeroSizeRequestReadBandwidth); - dml2_printf("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n", __func__, l->TotalZeroSizeCompressedReadBandwidth); - dml2_printf("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, l->MaximumEffectiveCompressionLuma); - dml2_printf("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, l->MaximumEffectiveCompressionChroma); - dml2_printf("DML::%s: AverageDCCCompressionRate = %f\n", __func__, l->AverageDCCCompressionRate); - dml2_printf("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, l->AverageDCCZeroSizeFraction); + DML_LOG_VERBOSE("DML::%s: UnboundedRequestEnabled = %u\n", __func__, p->UnboundedRequestEnabled); + DML_LOG_VERBOSE("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, l->TotalCompressedReadBandwidth); + DML_LOG_VERBOSE("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, l->TotalZeroSizeRequestReadBandwidth); + DML_LOG_VERBOSE("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n", __func__, l->TotalZeroSizeCompressedReadBandwidth); + DML_LOG_VERBOSE("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, l->MaximumEffectiveCompressionLuma); + DML_LOG_VERBOSE("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, l->MaximumEffectiveCompressionChroma); + DML_LOG_VERBOSE("DML::%s: AverageDCCCompressionRate = %f\n", __func__, l->AverageDCCCompressionRate); + DML_LOG_VERBOSE("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, l->AverageDCCZeroSizeFraction); - dml2_printf("DML::%s: CompbufReservedSpace64B = %u (%f kbytes)\n", __func__, p->CompbufReservedSpace64B, p->CompbufReservedSpace64B * 64 / 1024.0); - dml2_printf("DML::%s: CompbufReservedSpaceZs = %u\n", __func__, p->CompbufReservedSpaceZs); - dml2_printf("DML::%s: CompressedBufferSizeInkByte = %u kbytes\n", __func__, p->CompressedBufferSizeInkByte); - dml2_printf("DML::%s: ROBBufferSizeInKByte = %u kbytes\n", __func__, p->ROBBufferSizeInKByte); + DML_LOG_VERBOSE("DML::%s: CompbufReservedSpace64B = %u (%f kbytes)\n", __func__, p->CompbufReservedSpace64B, p->CompbufReservedSpace64B * 64 / 1024.0); + DML_LOG_VERBOSE("DML::%s: CompbufReservedSpaceZs = %u\n", __func__, p->CompbufReservedSpaceZs); + DML_LOG_VERBOSE("DML::%s: CompressedBufferSizeInkByte = %u kbytes\n", __func__, p->CompressedBufferSizeInkByte); + DML_LOG_VERBOSE("DML::%s: ROBBufferSizeInKByte = %u kbytes\n", __func__, p->ROBBufferSizeInKByte); #endif if (l->AverageDCCZeroSizeFraction == 1) { l->AverageZeroSizeCompressionRate = l->TotalZeroSizeRequestReadBandwidth / l->TotalZeroSizeCompressedReadBandwidth; @@ -10189,10 +10117,10 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate); - dml2_printf("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate + 1 / l->AverageDCCCompressionRate)); - dml2_printf("DML::%s: min 3 = %d\n", __func__, (p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64)); - dml2_printf("DML::%s: min 4 = %f\n", __func__, (p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate)); + DML_LOG_VERBOSE("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate); + DML_LOG_VERBOSE("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate + 1 / l->AverageDCCCompressionRate)); + DML_LOG_VERBOSE("DML::%s: min 3 = %d\n", __func__, (p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64)); + DML_LOG_VERBOSE("DML::%s: min 4 = %f\n", __func__, (p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate)); #endif } else { l->EffectiveCompressedBufferSize = math_min2((double)p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate, @@ -10200,16 +10128,16 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc ((double)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * (p->rob_alloc_compressed ? l->AverageDCCCompressionRate : 1.0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate); - dml2_printf("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 * l->AverageDCCCompressionRate); + DML_LOG_VERBOSE("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate); + DML_LOG_VERBOSE("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 * l->AverageDCCCompressionRate); #endif } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: MetaFIFOSizeInKEntries = %u\n", __func__, p->MetaFIFOSizeInKEntries); - dml2_printf("DML::%s: ZeroSizeBufferEntries = %u\n", __func__, p->ZeroSizeBufferEntries); - dml2_printf("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, l->AverageZeroSizeCompressionRate); - dml2_printf("DML::%s: EffectiveCompressedBufferSize = %f (%f kbytes)\n", __func__, l->EffectiveCompressedBufferSize, l->EffectiveCompressedBufferSize / 1024.0); + DML_LOG_VERBOSE("DML::%s: MetaFIFOSizeInKEntries = %u\n", __func__, p->MetaFIFOSizeInKEntries); + DML_LOG_VERBOSE("DML::%s: ZeroSizeBufferEntries = %u\n", __func__, p->ZeroSizeBufferEntries); + DML_LOG_VERBOSE("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, l->AverageZeroSizeCompressionRate); + DML_LOG_VERBOSE("DML::%s: EffectiveCompressedBufferSize = %f (%f kbytes)\n", __func__, l->EffectiveCompressedBufferSize, l->EffectiveCompressedBufferSize / 1024.0); #endif *p->StutterPeriod = 0; @@ -10220,15 +10148,15 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc l->LinesInDETYRoundedDownToSwath = math_floor2(l->LinesInDETY, p->SwathHeightY[k]); l->DETBufferingTimeY = l->LinesInDETYRoundedDownToSwath * ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, DETBufferSizeY = %u (%u kbytes)\n", __func__, k, p->DETBufferSizeY[k], p->DETBufferSizeY[k] / 1024); - dml2_printf("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]); - dml2_printf("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]); - dml2_printf("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]); - dml2_printf("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, p->TotalDataReadBandwidth); - dml2_printf("DML::%s: k=%u, LinesInDETY = %f\n", __func__, k, l->LinesInDETY); - dml2_printf("DML::%s: k=%u, LinesInDETYRoundedDownToSwath = %f\n", __func__, k, l->LinesInDETYRoundedDownToSwath); - dml2_printf("DML::%s: k=%u, VRatio = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); - dml2_printf("DML::%s: k=%u, DETBufferingTimeY = %f\n", __func__, k, l->DETBufferingTimeY); + DML_LOG_VERBOSE("DML::%s: k=%u, DETBufferSizeY = %u (%u kbytes)\n", __func__, k, p->DETBufferSizeY[k], p->DETBufferSizeY[k] / 1024); + DML_LOG_VERBOSE("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, p->TotalDataReadBandwidth); + DML_LOG_VERBOSE("DML::%s: k=%u, LinesInDETY = %f\n", __func__, k, l->LinesInDETY); + DML_LOG_VERBOSE("DML::%s: k=%u, LinesInDETYRoundedDownToSwath = %f\n", __func__, k, l->LinesInDETYRoundedDownToSwath); + DML_LOG_VERBOSE("DML::%s: k=%u, VRatio = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%u, DETBufferingTimeY = %f\n", __func__, k, l->DETBufferingTimeY); #endif if (!FoundCriticalSurface || l->DETBufferingTimeY < *p->StutterPeriod) { @@ -10248,17 +10176,17 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc l->SinglePipeCriticalSurface = (p->DPPPerSurface[k] == 1); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, FoundCriticalSurface = %u\n", __func__, k, FoundCriticalSurface); - dml2_printf("DML::%s: k=%u, StutterPeriod = %f\n", __func__, k, *p->StutterPeriod); - dml2_printf("DML::%s: k=%u, MinTTUVBlankCriticalSurface = %f\n", __func__, k, l->MinTTUVBlankCriticalSurface); - dml2_printf("DML::%s: k=%u, FrameTimeCriticalSurface= %f\n", __func__, k, l->FrameTimeCriticalSurface); - dml2_printf("DML::%s: k=%u, VActiveTimeCriticalSurface = %f\n", __func__, k, l->VActiveTimeCriticalSurface); - dml2_printf("DML::%s: k=%u, BytePerPixelYCriticalSurface = %u\n", __func__, k, l->BytePerPixelYCriticalSurface); - dml2_printf("DML::%s: k=%u, SwathWidthYCriticalSurface = %f\n", __func__, k, l->SwathWidthYCriticalSurface); - dml2_printf("DML::%s: k=%u, SwathHeightYCriticalSurface = %f\n", __func__, k, l->SwathHeightYCriticalSurface); - dml2_printf("DML::%s: k=%u, BlockWidth256BytesYCriticalSurface = %u\n", __func__, k, l->BlockWidth256BytesYCriticalSurface); - dml2_printf("DML::%s: k=%u, SinglePlaneCriticalSurface = %u\n", __func__, k, l->SinglePlaneCriticalSurface); - dml2_printf("DML::%s: k=%u, SinglePipeCriticalSurface = %u\n", __func__, k, l->SinglePipeCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, FoundCriticalSurface = %u\n", __func__, k, FoundCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, StutterPeriod = %f\n", __func__, k, *p->StutterPeriod); + DML_LOG_VERBOSE("DML::%s: k=%u, MinTTUVBlankCriticalSurface = %f\n", __func__, k, l->MinTTUVBlankCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, FrameTimeCriticalSurface= %f\n", __func__, k, l->FrameTimeCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, VActiveTimeCriticalSurface = %f\n", __func__, k, l->VActiveTimeCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, BytePerPixelYCriticalSurface = %u\n", __func__, k, l->BytePerPixelYCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, SwathWidthYCriticalSurface = %f\n", __func__, k, l->SwathWidthYCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, SwathHeightYCriticalSurface = %f\n", __func__, k, l->SwathHeightYCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, BlockWidth256BytesYCriticalSurface = %u\n", __func__, k, l->BlockWidth256BytesYCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, SinglePlaneCriticalSurface = %u\n", __func__, k, l->SinglePlaneCriticalSurface); + DML_LOG_VERBOSE("DML::%s: k=%u, SinglePipeCriticalSurface = %u\n", __func__, k, l->SinglePipeCriticalSurface); #endif } } @@ -10276,14 +10204,14 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = math_min2(*p->StutterPeriod * p->TotalDataReadBandwidth, l->EffectiveCompressedBufferSize); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: AverageDCCCompressionRate = %f\n", __func__, l->AverageDCCCompressionRate); - dml2_printf("DML::%s: StutterPeriod*TotalDataReadBandwidth = %f (%f kbytes)\n", __func__, *p->StutterPeriod * p->TotalDataReadBandwidth, (*p->StutterPeriod * p->TotalDataReadBandwidth) / 1024.0); - dml2_printf("DML::%s: EffectiveCompressedBufferSize = %f (%f kbytes)\n", __func__, l->EffectiveCompressedBufferSize, l->EffectiveCompressedBufferSize / 1024.0); - dml2_printf("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f (%f kbytes)\n", __func__, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / 1024); - dml2_printf("DML::%s: ReturnBW = %f\n", __func__, p->ReturnBW); - dml2_printf("DML::%s: TotalDataReadBandwidth = %f\n", __func__, p->TotalDataReadBandwidth); - dml2_printf("DML::%s: TotalRowReadBandwidth = %f\n", __func__, l->TotalRowReadBandwidth); - dml2_printf("DML::%s: DCFCLK = %f\n", __func__, p->DCFCLK); + DML_LOG_VERBOSE("DML::%s: AverageDCCCompressionRate = %f\n", __func__, l->AverageDCCCompressionRate); + DML_LOG_VERBOSE("DML::%s: StutterPeriod*TotalDataReadBandwidth = %f (%f kbytes)\n", __func__, *p->StutterPeriod * p->TotalDataReadBandwidth, (*p->StutterPeriod * p->TotalDataReadBandwidth) / 1024.0); + DML_LOG_VERBOSE("DML::%s: EffectiveCompressedBufferSize = %f (%f kbytes)\n", __func__, l->EffectiveCompressedBufferSize, l->EffectiveCompressedBufferSize / 1024.0); + DML_LOG_VERBOSE("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f (%f kbytes)\n", __func__, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / 1024); + DML_LOG_VERBOSE("DML::%s: ReturnBW = %f\n", __func__, p->ReturnBW); + DML_LOG_VERBOSE("DML::%s: TotalDataReadBandwidth = %f\n", __func__, p->TotalDataReadBandwidth); + DML_LOG_VERBOSE("DML::%s: TotalRowReadBandwidth = %f\n", __func__, l->TotalRowReadBandwidth); + DML_LOG_VERBOSE("DML::%s: DCFCLK = %f\n", __func__, p->DCFCLK); #endif l->StutterBurstTime = l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer @@ -10292,10 +10220,10 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc / math_min2(p->DCFCLK * 64, p->ReturnBW * (p->hw_debug5 ? 1 : l->AverageDCCCompressionRate)) + *p->StutterPeriod * l->TotalRowReadBandwidth / p->ReturnBW; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Part 1 = %f\n", __func__, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / p->ReturnBW / (p->hw_debug5 ? 1 : l->AverageDCCCompressionRate)); - dml2_printf("DML::%s: Part 2 = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth - l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64)); - dml2_printf("DML::%s: Part 3 = %f\n", __func__, *p->StutterPeriod * l->TotalRowReadBandwidth / p->ReturnBW); - dml2_printf("DML::%s: StutterBurstTime = %f\n", __func__, l->StutterBurstTime); + DML_LOG_VERBOSE("DML::%s: Part 1 = %f\n", __func__, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / p->ReturnBW / (p->hw_debug5 ? 1 : l->AverageDCCCompressionRate)); + DML_LOG_VERBOSE("DML::%s: Part 2 = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth - l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64)); + DML_LOG_VERBOSE("DML::%s: Part 3 = %f\n", __func__, *p->StutterPeriod * l->TotalRowReadBandwidth / p->ReturnBW); + DML_LOG_VERBOSE("DML::%s: StutterBurstTime = %f\n", __func__, l->StutterBurstTime); #endif l->TotalActiveWriteback = 0; memset(l->stream_visited, 0, DML2_MAX_PLANES * sizeof(bool)); @@ -10324,9 +10252,9 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc if (l->TotalActiveWriteback == 0) { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: SRExitTime = %f\n", __func__, p->SRExitTime); - dml2_printf("DML::%s: SRExitZ8Time = %f\n", __func__, p->SRExitZ8Time); - dml2_printf("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod); + DML_LOG_VERBOSE("DML::%s: SRExitTime = %f\n", __func__, p->SRExitTime); + DML_LOG_VERBOSE("DML::%s: SRExitZ8Time = %f\n", __func__, p->SRExitZ8Time); + DML_LOG_VERBOSE("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod); #endif *p->StutterEfficiencyNotIncludingVBlank = math_max2(0., 1 - (p->SRExitTime + l->StutterBurstTime) / *p->StutterPeriod) * 100; *p->Z8StutterEfficiencyNotIncludingVBlank = math_max2(0., 1 - (p->SRExitZ8Time + l->StutterBurstTime) / *p->StutterPeriod) * 100; @@ -10339,11 +10267,11 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc *p->Z8NumberOfStutterBurstsPerFrame = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: VActiveTimeCriticalSurface = %f\n", __func__, l->VActiveTimeCriticalSurface); - dml2_printf("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank); - dml2_printf("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank); - dml2_printf("DML::%s: NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->NumberOfStutterBurstsPerFrame); - dml2_printf("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame); + DML_LOG_VERBOSE("DML::%s: VActiveTimeCriticalSurface = %f\n", __func__, l->VActiveTimeCriticalSurface); + DML_LOG_VERBOSE("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank); + DML_LOG_VERBOSE("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank); + DML_LOG_VERBOSE("DML::%s: NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->NumberOfStutterBurstsPerFrame); + DML_LOG_VERBOSE("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame); #endif if (*p->StutterEfficiencyNotIncludingVBlank > 0) { @@ -10358,7 +10286,7 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc } if (*p->Z8StutterEfficiencyNotIncludingVBlank > 0) { - LastZ8StutterPeriod = l->VActiveTimeCriticalSurface - (*p->Z8NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod; + //LastZ8StutterPeriod = l->VActiveTimeCriticalSurface - (*p->Z8NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod; if (!((p->SynchronizeTimings || TotalNumberOfActiveOTG == 1) && SameTiming)) { *p->Z8StutterEfficiency = *p->Z8StutterEfficiencyNotIncludingVBlank; } else { @@ -10370,25 +10298,25 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: TotalNumberOfActiveOTG = %u\n", __func__, TotalNumberOfActiveOTG); - dml2_printf("DML::%s: SameTiming = %u\n", __func__, SameTiming); - dml2_printf("DML::%s: SynchronizeTimings = %u\n", __func__, p->SynchronizeTimings); - dml2_printf("DML::%s: LastZ8StutterPeriod = %f\n", __func__, LastZ8StutterPeriod); - dml2_printf("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Z8StutterEnterPlusExitWatermark); - dml2_printf("DML::%s: StutterBurstTime = %f\n", __func__, l->StutterBurstTime); - dml2_printf("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod); - dml2_printf("DML::%s: StutterEfficiency = %f\n", __func__, *p->StutterEfficiency); - dml2_printf("DML::%s: Z8StutterEfficiency = %f\n", __func__, *p->Z8StutterEfficiency); - dml2_printf("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank); - dml2_printf("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame); + DML_LOG_VERBOSE("DML::%s: TotalNumberOfActiveOTG = %u\n", __func__, TotalNumberOfActiveOTG); + DML_LOG_VERBOSE("DML::%s: SameTiming = %u\n", __func__, SameTiming); + DML_LOG_VERBOSE("DML::%s: SynchronizeTimings = %u\n", __func__, p->SynchronizeTimings); + DML_LOG_VERBOSE("DML::%s: LastZ8StutterPeriod = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank > 0 ? l->VActiveTimeCriticalSurface - (*p->Z8NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod : 0); + DML_LOG_VERBOSE("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Z8StutterEnterPlusExitWatermark); + DML_LOG_VERBOSE("DML::%s: StutterBurstTime = %f\n", __func__, l->StutterBurstTime); + DML_LOG_VERBOSE("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod); + DML_LOG_VERBOSE("DML::%s: StutterEfficiency = %f\n", __func__, *p->StutterEfficiency); + DML_LOG_VERBOSE("DML::%s: Z8StutterEfficiency = %f\n", __func__, *p->Z8StutterEfficiency); + DML_LOG_VERBOSE("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank); + DML_LOG_VERBOSE("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame); #endif *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = !(!p->UnboundedRequestEnabled && (p->NumberOfActiveSurfaces == 1) && l->SinglePlaneCriticalSurface && l->SinglePipeCriticalSurface); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: DETBufferSizeYCriticalSurface = %u\n", __func__, l->DETBufferSizeYCriticalSurface); - dml2_printf("DML::%s: PixelChunkSizeInKByte = %u\n", __func__, p->PixelChunkSizeInKByte); - dml2_printf("DML::%s: DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = %u\n", __func__, *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE); + DML_LOG_VERBOSE("DML::%s: DETBufferSizeYCriticalSurface = %u\n", __func__, l->DETBufferSizeYCriticalSurface); + DML_LOG_VERBOSE("DML::%s: PixelChunkSizeInKByte = %u\n", __func__, p->PixelChunkSizeInKByte); + DML_LOG_VERBOSE("DML::%s: DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = %u\n", __func__, *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE); #endif } @@ -10422,7 +10350,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex double max_uclk_mhz = 0; double min_return_latency_in_DCFCLK_cycles = 0; - dml2_printf("DML::%s: --- START --- \n", __func__); + DML_LOG_VERBOSE("DML::%s: --- START --- \n", __func__); memset(&mode_lib->scratch, 0, sizeof(struct dml2_core_internal_scratch)); memset(&mode_lib->mp, 0, sizeof(struct dml2_core_internal_mode_program)); @@ -10444,13 +10372,13 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex for (k = 0; k < s->num_active_planes; ++k) { unsigned int stream_index = display_cfg->plane_descriptors[k].stream_index; - DML2_ASSERT(cfg_support_info->stream_support_info[stream_index].odms_used <= 4); - DML2_ASSERT(cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 4 || + DML_ASSERT(cfg_support_info->stream_support_info[stream_index].odms_used <= 4); + DML_ASSERT(cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 4 || cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 2 || cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 1); if (cfg_support_info->stream_support_info[stream_index].odms_used > 1) - DML2_ASSERT(cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 1); + DML_ASSERT(cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 1); switch (cfg_support_info->stream_support_info[stream_index].odms_used) { case (4): @@ -10476,51 +10404,51 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex for (k = 0; k < s->num_active_planes; ++k) { mode_lib->mp.NoOfDPP[k] = cfg_support_info->plane_support_info[k].dpps_used; mode_lib->mp.Dppclk[k] = programming->plane_programming[k].min_clocks.dcn4x.dppclk_khz / 1000.0; - DML2_ASSERT(mode_lib->mp.Dppclk[k] > 0); + DML_ASSERT(mode_lib->mp.Dppclk[k] > 0); } for (k = 0; k < s->num_active_planes; ++k) { unsigned int stream_index = display_cfg->plane_descriptors[k].stream_index; mode_lib->mp.DSCCLK[k] = programming->stream_programming[stream_index].min_clocks.dcn4x.dscclk_khz / 1000.0; - dml2_printf("DML::%s: k=%d stream_index=%d, mode_lib->mp.DSCCLK = %f\n", __func__, k, stream_index, mode_lib->mp.DSCCLK[k]); + DML_LOG_VERBOSE("DML::%s: k=%d stream_index=%d, mode_lib->mp.DSCCLK = %f\n", __func__, k, stream_index, mode_lib->mp.DSCCLK[k]); } mode_lib->mp.Dispclk = programming->min_clocks.dcn4x.dispclk_khz / 1000.0; mode_lib->mp.DCFCLKDeepSleep = programming->min_clocks.dcn4x.deepsleep_dcfclk_khz / 1000.0; - DML2_ASSERT(mode_lib->mp.Dcfclk > 0); - DML2_ASSERT(mode_lib->mp.FabricClock > 0); - DML2_ASSERT(mode_lib->mp.dram_bw_mbps > 0); - DML2_ASSERT(mode_lib->mp.uclk_freq_mhz > 0); - DML2_ASSERT(mode_lib->mp.GlobalDPPCLK > 0); - DML2_ASSERT(mode_lib->mp.Dispclk > 0); - DML2_ASSERT(mode_lib->mp.DCFCLKDeepSleep > 0); - DML2_ASSERT(s->SOCCLK > 0); - -#ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: num_active_planes = %u\n", __func__, s->num_active_planes); - dml2_printf("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes); - dml2_printf("DML::%s: Dcfclk = %f\n", __func__, mode_lib->mp.Dcfclk); - dml2_printf("DML::%s: FabricClock = %f\n", __func__, mode_lib->mp.FabricClock); - dml2_printf("DML::%s: dram_bw_mbps = %f\n", __func__, mode_lib->mp.dram_bw_mbps); - dml2_printf("DML::%s: uclk_freq_mhz = %f\n", __func__, mode_lib->mp.uclk_freq_mhz); - dml2_printf("DML::%s: Dispclk = %f\n", __func__, mode_lib->mp.Dispclk); + DML_ASSERT(mode_lib->mp.Dcfclk > 0); + DML_ASSERT(mode_lib->mp.FabricClock > 0); + DML_ASSERT(mode_lib->mp.dram_bw_mbps > 0); + DML_ASSERT(mode_lib->mp.uclk_freq_mhz > 0); + DML_ASSERT(mode_lib->mp.GlobalDPPCLK > 0); + DML_ASSERT(mode_lib->mp.Dispclk > 0); + DML_ASSERT(mode_lib->mp.DCFCLKDeepSleep > 0); + DML_ASSERT(s->SOCCLK > 0); + +#ifdef __DML_VBA_DEBUG__ + DML_LOG_VERBOSE("DML::%s: num_active_planes = %u\n", __func__, s->num_active_planes); + DML_LOG_VERBOSE("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes); + DML_LOG_VERBOSE("DML::%s: Dcfclk = %f\n", __func__, mode_lib->mp.Dcfclk); + DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, mode_lib->mp.FabricClock); + DML_LOG_VERBOSE("DML::%s: dram_bw_mbps = %f\n", __func__, mode_lib->mp.dram_bw_mbps); + DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, mode_lib->mp.uclk_freq_mhz); + DML_LOG_VERBOSE("DML::%s: Dispclk = %f\n", __func__, mode_lib->mp.Dispclk); for (k = 0; k < s->num_active_planes; ++k) { - dml2_printf("DML::%s: Dppclk[%0d] = %f\n", __func__, k, mode_lib->mp.Dppclk[k]); - } - dml2_printf("DML::%s: GlobalDPPCLK = %f\n", __func__, mode_lib->mp.GlobalDPPCLK); - dml2_printf("DML::%s: DCFCLKDeepSleep = %f\n", __func__, mode_lib->mp.DCFCLKDeepSleep); - dml2_printf("DML::%s: SOCCLK = %f\n", __func__, s->SOCCLK); - dml2_printf("DML::%s: min_clk_index = %0d\n", __func__, in_out_params->min_clk_index); - dml2_printf("DML::%s: min_clk_table min_fclk_khz = %d\n", __func__, min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].min_fclk_khz); - dml2_printf("DML::%s: min_clk_table uclk_mhz = %f\n", __func__, dram_bw_kbps_to_uclk_mhz(min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps, &mode_lib->soc.clk_table.dram_config)); + DML_LOG_VERBOSE("DML::%s: Dppclk[%0d] = %f\n", __func__, k, mode_lib->mp.Dppclk[k]); + } + DML_LOG_VERBOSE("DML::%s: GlobalDPPCLK = %f\n", __func__, mode_lib->mp.GlobalDPPCLK); + DML_LOG_VERBOSE("DML::%s: DCFCLKDeepSleep = %f\n", __func__, mode_lib->mp.DCFCLKDeepSleep); + DML_LOG_VERBOSE("DML::%s: SOCCLK = %f\n", __func__, s->SOCCLK); + DML_LOG_VERBOSE("DML::%s: min_clk_index = %0d\n", __func__, in_out_params->min_clk_index); + DML_LOG_VERBOSE("DML::%s: min_clk_table min_fclk_khz = %ld\n", __func__, min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].min_fclk_khz); + DML_LOG_VERBOSE("DML::%s: min_clk_table uclk_mhz = %f\n", __func__, dram_bw_kbps_to_uclk_mhz(min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps, &mode_lib->soc.clk_table.dram_config)); for (k = 0; k < mode_lib->mp.num_active_pipes; ++k) { - dml2_printf("DML::%s: pipe=%d is in plane=%d\n", __func__, k, mode_lib->mp.pipe_plane[k]); - dml2_printf("DML::%s: Per-plane DPPPerSurface[%0d] = %d\n", __func__, k, mode_lib->mp.NoOfDPP[k]); + DML_LOG_VERBOSE("DML::%s: pipe=%d is in plane=%d\n", __func__, k, mode_lib->mp.pipe_plane[k]); + DML_LOG_VERBOSE("DML::%s: Per-plane DPPPerSurface[%0d] = %d\n", __func__, k, mode_lib->mp.NoOfDPP[k]); } for (k = 0; k < s->num_active_planes; k++) - dml2_printf("DML::%s: plane_%d: reserved_vblank_time_ns = %u\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns); + DML_LOG_VERBOSE("DML::%s: plane_%d: reserved_vblank_time_ns = %lu\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns); #endif CalculateMaxDETAndMinCompressedBufferSize( @@ -10617,8 +10545,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)); mode_lib->mp.vactive_sw_bw_l[k] = mode_lib->mp.SwathWidthSingleDPPY[k] * mode_lib->mp.BytePerPixelY[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio; mode_lib->mp.vactive_sw_bw_c[k] = mode_lib->mp.SwathWidthSingleDPPC[k] * mode_lib->mp.BytePerPixelC[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio; - dml2_printf("DML::%s: vactive_sw_bw_l[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]); - dml2_printf("DML::%s: vactive_sw_bw_c[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]); + DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_l[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]); + DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_c[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]); } CalculateSwathAndDETConfiguration_params->display_cfg = display_cfg; @@ -11097,7 +11025,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin); for (k = 0; k < s->num_active_planes; ++k) { - bool cursor_not_enough_urgent_latency_hiding = 0; + bool cursor_not_enough_urgent_latency_hiding = false; s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); @@ -11173,8 +11101,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.WritebackDelay[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); - dml2_printf("DML::%s: k=%u WritebackDelay = %f\n", __func__, k, mode_lib->mp.WritebackDelay[k]); + DML_LOG_VERBOSE("DML::%s: k=%u MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); + DML_LOG_VERBOSE("DML::%s: k=%u WritebackDelay = %f\n", __func__, k, mode_lib->mp.WritebackDelay[k]); #endif } @@ -11183,7 +11111,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex s->immediate_flip_required = s->immediate_flip_required || display_cfg->plane_descriptors[k].immediate_flip; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: immediate_flip_required = %u\n", __func__, s->immediate_flip_required); + DML_LOG_VERBOSE("DML::%s: immediate_flip_required = %u\n", __func__, s->immediate_flip_required); #endif if (s->num_active_planes > 1) { @@ -11219,12 +11147,12 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex s->DestinationLineTimesForPrefetchLessThan2 = false; s->VRatioPrefetchMoreThanMax = false; - dml2_printf("DML::%s: Start one iteration of prefetch schedule evaluation\n", __func__); + DML_LOG_VERBOSE("DML::%s: Start one iteration of prefetch schedule evaluation\n", __func__); for (k = 0; k < s->num_active_planes; ++k) { struct dml2_core_internal_DmlPipe *myPipe = &s->myPipe; - dml2_printf("DML::%s: k=%d MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); + DML_LOG_VERBOSE("DML::%s: k=%d MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); mode_lib->mp.TWait[k] = CalculateTWait( display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns, mode_lib->mp.UrgentLatency, @@ -11261,7 +11189,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ip.ptoi_supported; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k); #endif CalculatePrefetchSchedule_params->display_cfg = display_cfg; CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactorPrefetch; @@ -11356,7 +11284,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.impacted_prefetch_margin_us[k] = 0; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%0u NoTimeToPrefetch=%0d\n", __func__, k, mode_lib->mp.NoTimeToPrefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u NoTimeToPrefetch=%0d\n", __func__, k, mode_lib->mp.NoTimeToPrefetch[k]); #endif mode_lib->mp.VStartupMin[k] = s->MaxVStartupLines[k]; } // for k @@ -11366,9 +11294,9 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex if (mode_lib->mp.NoTimeToPrefetch[k] == true || mode_lib->mp.NotEnoughTimeForDynamicMetadata[k] || mode_lib->mp.DSTYAfterScaler[k] > 8) { - dml2_printf("DML::%s: k=%u, NoTimeToPrefetch = %0d\n", __func__, k, mode_lib->mp.NoTimeToPrefetch[k]); - dml2_printf("DML::%s: k=%u, NotEnoughTimeForDynamicMetadata=%u\n", __func__, k, mode_lib->mp.NotEnoughTimeForDynamicMetadata[k]); - dml2_printf("DML::%s: k=%u, DSTYAfterScaler=%u (should be <= 0)\n", __func__, k, mode_lib->mp.DSTYAfterScaler[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, NoTimeToPrefetch = %0d\n", __func__, k, mode_lib->mp.NoTimeToPrefetch[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, NotEnoughTimeForDynamicMetadata=%u\n", __func__, k, mode_lib->mp.NotEnoughTimeForDynamicMetadata[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, DSTYAfterScaler=%u (should be <= 0)\n", __func__, k, mode_lib->mp.DSTYAfterScaler[k]); mode_lib->mp.PrefetchModeSupported = false; } if (mode_lib->mp.dst_y_prefetch[k] < 2) @@ -11377,24 +11305,24 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex if (mode_lib->mp.VRatioPrefetchY[k] > __DML2_CALCS_MAX_VRATIO_PRE__ || mode_lib->mp.VRatioPrefetchC[k] > __DML2_CALCS_MAX_VRATIO_PRE__) { s->VRatioPrefetchMoreThanMax = true; - dml2_printf("DML::%s: k=%d, VRatioPrefetchY=%f (should not be < %f)\n", __func__, k, mode_lib->mp.VRatioPrefetchY[k], __DML2_CALCS_MAX_VRATIO_PRE__); - dml2_printf("DML::%s: k=%d, VRatioPrefetchC=%f (should not be < %f)\n", __func__, k, mode_lib->mp.VRatioPrefetchC[k], __DML2_CALCS_MAX_VRATIO_PRE__); - dml2_printf("DML::%s: VRatioPrefetchMoreThanMax = %u\n", __func__, s->VRatioPrefetchMoreThanMax); + DML_LOG_VERBOSE("DML::%s: k=%d, VRatioPrefetchY=%f (should not be < %f)\n", __func__, k, mode_lib->mp.VRatioPrefetchY[k], __DML2_CALCS_MAX_VRATIO_PRE__); + DML_LOG_VERBOSE("DML::%s: k=%d, VRatioPrefetchC=%f (should not be < %f)\n", __func__, k, mode_lib->mp.VRatioPrefetchC[k], __DML2_CALCS_MAX_VRATIO_PRE__); + DML_LOG_VERBOSE("DML::%s: VRatioPrefetchMoreThanMax = %u\n", __func__, s->VRatioPrefetchMoreThanMax); } if (mode_lib->mp.NotEnoughUrgentLatencyHiding[k]) { - dml2_printf("DML::%s: k=%u, NotEnoughUrgentLatencyHiding = %u\n", __func__, k, mode_lib->mp.NotEnoughUrgentLatencyHiding[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, NotEnoughUrgentLatencyHiding = %u\n", __func__, k, mode_lib->mp.NotEnoughUrgentLatencyHiding[k]); mode_lib->mp.PrefetchModeSupported = false; } } if (s->VRatioPrefetchMoreThanMax == true || s->DestinationLineTimesForPrefetchLessThan2 == true) { - dml2_printf("DML::%s: VRatioPrefetchMoreThanMax = %u\n", __func__, s->VRatioPrefetchMoreThanMax); - dml2_printf("DML::%s: DestinationLineTimesForPrefetchLessThan2 = %u\n", __func__, s->DestinationLineTimesForPrefetchLessThan2); + DML_LOG_VERBOSE("DML::%s: VRatioPrefetchMoreThanMax = %u\n", __func__, s->VRatioPrefetchMoreThanMax); + DML_LOG_VERBOSE("DML::%s: DestinationLineTimesForPrefetchLessThan2 = %u\n", __func__, s->DestinationLineTimesForPrefetchLessThan2); mode_lib->mp.PrefetchModeSupported = false; } - dml2_printf("DML::%s: Prefetch schedule is %sOK at vstartup = %u\n", __func__, + DML_LOG_VERBOSE("DML::%s: Prefetch schedule is %sOK at vstartup = %u\n", __func__, mode_lib->mp.PrefetchModeSupported ? "" : "NOT ", CalculatePrefetchSchedule_params->VStartup); // Prefetch schedule OK, now check prefetch bw @@ -11422,24 +11350,24 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex &mode_lib->mp.NotEnoughUrgentLatencyHidingPre[k]); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%0u DPPPerSurface=%u\n", __func__, k, mode_lib->mp.NoOfDPP[k]); - dml2_printf("DML::%s: k=%0u UrgentBurstFactorLuma=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorLuma[k]); - dml2_printf("DML::%s: k=%0u UrgentBurstFactorChroma=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorChroma[k]); - dml2_printf("DML::%s: k=%0u UrgentBurstFactorLumaPre=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorLumaPre[k]); - dml2_printf("DML::%s: k=%0u UrgentBurstFactorChromaPre=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorChromaPre[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u DPPPerSurface=%u\n", __func__, k, mode_lib->mp.NoOfDPP[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u UrgentBurstFactorLuma=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u UrgentBurstFactorChroma=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u UrgentBurstFactorLumaPre=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorLumaPre[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u UrgentBurstFactorChromaPre=%f\n", __func__, k, mode_lib->mp.UrgentBurstFactorChromaPre[k]); - dml2_printf("DML::%s: k=%0u VRatioPrefetchY=%f\n", __func__, k, mode_lib->mp.VRatioPrefetchY[k]); - dml2_printf("DML::%s: k=%0u VRatioY=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); + DML_LOG_VERBOSE("DML::%s: k=%0u VRatioPrefetchY=%f\n", __func__, k, mode_lib->mp.VRatioPrefetchY[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u VRatioY=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio); - dml2_printf("DML::%s: k=%0u prefetch_vmrow_bw=%f\n", __func__, k, mode_lib->mp.prefetch_vmrow_bw[k]); - dml2_printf("DML::%s: k=%0u vactive_sw_bw_l=%f\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]); - dml2_printf("DML::%s: k=%0u vactive_sw_bw_c=%f\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]); - dml2_printf("DML::%s: k=%0u cursor_bw=%f\n", __func__, k, mode_lib->mp.cursor_bw[k]); - dml2_printf("DML::%s: k=%0u dpte_row_bw=%f\n", __func__, k, mode_lib->mp.dpte_row_bw[k]); - dml2_printf("DML::%s: k=%0u meta_row_bw=%f\n", __func__, k, mode_lib->mp.meta_row_bw[k]); - dml2_printf("DML::%s: k=%0u RequiredPrefetchPixelDataBWLuma=%f\n", __func__, k, mode_lib->mp.RequiredPrefetchPixelDataBWLuma[k]); - dml2_printf("DML::%s: k=%0u RequiredPrefetchPixelDataBWChroma=%f\n", __func__, k, mode_lib->mp.RequiredPrefetchPixelDataBWChroma[k]); - dml2_printf("DML::%s: k=%0u prefetch_cursor_bw=%f\n", __func__, k, mode_lib->mp.prefetch_cursor_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u prefetch_vmrow_bw=%f\n", __func__, k, mode_lib->mp.prefetch_vmrow_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u vactive_sw_bw_l=%f\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u vactive_sw_bw_c=%f\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u cursor_bw=%f\n", __func__, k, mode_lib->mp.cursor_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u dpte_row_bw=%f\n", __func__, k, mode_lib->mp.dpte_row_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u meta_row_bw=%f\n", __func__, k, mode_lib->mp.meta_row_bw[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u RequiredPrefetchPixelDataBWLuma=%f\n", __func__, k, mode_lib->mp.RequiredPrefetchPixelDataBWLuma[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u RequiredPrefetchPixelDataBWChroma=%f\n", __func__, k, mode_lib->mp.RequiredPrefetchPixelDataBWChroma[k]); + DML_LOG_VERBOSE("DML::%s: k=%0u prefetch_cursor_bw=%f\n", __func__, k, mode_lib->mp.prefetch_cursor_bw[k]); #endif } @@ -11503,11 +11431,11 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.urg_bandwidth_available); if (!mode_lib->mp.PrefetchModeSupported) - dml2_printf("DML::%s: Bandwidth not sufficient for prefetch!\n", __func__); + DML_LOG_VERBOSE("DML::%s: Bandwidth not sufficient for prefetch!\n", __func__); for (k = 0; k < s->num_active_planes; ++k) { if (mode_lib->mp.NotEnoughUrgentLatencyHidingPre[k]) { - dml2_printf("DML::%s: k=%u, NotEnoughUrgentLatencyHidingPre = %u\n", __func__, k, mode_lib->mp.NotEnoughUrgentLatencyHidingPre[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, NotEnoughUrgentLatencyHidingPre = %u\n", __func__, k, mode_lib->mp.NotEnoughUrgentLatencyHidingPre[k]); mode_lib->mp.PrefetchModeSupported = false; } } @@ -11533,12 +11461,12 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex } mode_lib->mp.TotImmediateFlipBytes += s->per_pipe_flip_bytes[k] * mode_lib->mp.NoOfDPP[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k = %u\n", __func__, k); - dml2_printf("DML::%s: DPPPerSurface = %u\n", __func__, mode_lib->mp.NoOfDPP[k]); - dml2_printf("DML::%s: vm_bytes = %u\n", __func__, mode_lib->mp.vm_bytes[k]); - dml2_printf("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, mode_lib->mp.PixelPTEBytesPerRow[k]); - dml2_printf("DML::%s: meta_row_bytes = %u\n", __func__, mode_lib->mp.meta_row_bytes[k]); - dml2_printf("DML::%s: TotImmediateFlipBytes = %u\n", __func__, mode_lib->mp.TotImmediateFlipBytes); + DML_LOG_VERBOSE("DML::%s: k = %u\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: DPPPerSurface = %u\n", __func__, mode_lib->mp.NoOfDPP[k]); + DML_LOG_VERBOSE("DML::%s: vm_bytes = %u\n", __func__, mode_lib->mp.vm_bytes[k]); + DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, mode_lib->mp.PixelPTEBytesPerRow[k]); + DML_LOG_VERBOSE("DML::%s: meta_row_bytes = %u\n", __func__, mode_lib->mp.meta_row_bytes[k]); + DML_LOG_VERBOSE("DML::%s: TotImmediateFlipBytes = %u\n", __func__, mode_lib->mp.TotImmediateFlipBytes); #endif } for (k = 0; k < s->num_active_planes; ++k) { @@ -11631,13 +11559,13 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.urg_bandwidth_available); if (!mode_lib->mp.ImmediateFlipSupported) - dml2_printf("DML::%s: Bandwidth not sufficient for flip!", __func__); + DML_LOG_VERBOSE("DML::%s: Bandwidth not sufficient for flip!", __func__); for (k = 0; k < s->num_active_planes; ++k) { if (display_cfg->plane_descriptors[k].immediate_flip && mode_lib->mp.ImmediateFlipSupportedForPipe[k] == false) { mode_lib->mp.ImmediateFlipSupported = false; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Pipe %0d not supporting iflip!\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: Pipe %0d not supporting iflip!\n", __func__, k); #endif } } @@ -11650,28 +11578,28 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.PrefetchAndImmediateFlipSupported = (mode_lib->mp.PrefetchModeSupported == true && (!must_support_iflip || mode_lib->mp.ImmediateFlipSupported)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: PrefetchModeSupported = %u\n", __func__, mode_lib->mp.PrefetchModeSupported); + DML_LOG_VERBOSE("DML::%s: PrefetchModeSupported = %u\n", __func__, mode_lib->mp.PrefetchModeSupported); for (k = 0; k < s->num_active_planes; ++k) - dml2_printf("DML::%s: immediate_flip_required[%u] = %u\n", __func__, k, display_cfg->plane_descriptors[k].immediate_flip); - dml2_printf("DML::%s: HostVMEnable = %u\n", __func__, display_cfg->hostvm_enable); - dml2_printf("DML::%s: ImmediateFlipSupported = %u\n", __func__, mode_lib->mp.ImmediateFlipSupported); - dml2_printf("DML::%s: PrefetchAndImmediateFlipSupported = %u\n", __func__, mode_lib->mp.PrefetchAndImmediateFlipSupported); + DML_LOG_VERBOSE("DML::%s: immediate_flip_required[%u] = %u\n", __func__, k, display_cfg->plane_descriptors[k].immediate_flip); + DML_LOG_VERBOSE("DML::%s: HostVMEnable = %u\n", __func__, display_cfg->hostvm_enable); + DML_LOG_VERBOSE("DML::%s: ImmediateFlipSupported = %u\n", __func__, mode_lib->mp.ImmediateFlipSupported); + DML_LOG_VERBOSE("DML::%s: PrefetchAndImmediateFlipSupported = %u\n", __func__, mode_lib->mp.PrefetchAndImmediateFlipSupported); #endif - dml2_printf("DML::%s: Done one iteration: k=%d, MaxVStartupLines=%u\n", __func__, k, s->MaxVStartupLines[k]); + DML_LOG_VERBOSE("DML::%s: Done one iteration: k=%d, MaxVStartupLines=%u\n", __func__, k, s->MaxVStartupLines[k]); } for (k = 0; k < s->num_active_planes; ++k) - dml2_printf("DML::%s: k=%d MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); + DML_LOG_VERBOSE("DML::%s: k=%d MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); if (!mode_lib->mp.PrefetchAndImmediateFlipSupported) { - dml2_printf("DML::%s: Bad, Prefetch and flip scheduling solution NOT found!\n", __func__); + DML_LOG_VERBOSE("DML::%s: Bad, Prefetch and flip scheduling solution NOT found!\n", __func__); } else { - dml2_printf("DML::%s: Good, Prefetch and flip scheduling solution found\n", __func__); + DML_LOG_VERBOSE("DML::%s: Good, Prefetch and flip scheduling solution found\n", __func__); // DCC Configuration for (k = 0; k < s->num_active_planes; ++k) { #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: Calculate DCC configuration for surface k=%u\n", __func__, k); + DML_LOG_VERBOSE("DML::%s: Calculate DCC configuration for surface k=%u\n", __func__, k); #endif CalculateDCCConfiguration( display_cfg->plane_descriptors[k].surface.dcc.enable, @@ -11780,8 +11708,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex calculate_pstate_keepout_dst_lines(display_cfg, &mode_lib->mp.Watermark, mode_lib->mp.pstate_keepout_dst_lines); - dml2_printf("DML::%s: DEBUG stream_index = %0d\n", __func__, display_cfg->plane_descriptors[0].stream_index); - dml2_printf("DML::%s: DEBUG PixelClock = %d kHz\n", __func__, (display_cfg->stream_descriptors[display_cfg->plane_descriptors[0].stream_index].timing.pixel_clock_khz)); + DML_LOG_VERBOSE("DML::%s: DEBUG stream_index = %0d\n", __func__, display_cfg->plane_descriptors[0].stream_index); + DML_LOG_VERBOSE("DML::%s: DEBUG PixelClock = %ld kHz\n", __func__, (display_cfg->stream_descriptors[display_cfg->plane_descriptors[0].stream_index].timing.pixel_clock_khz)); //Display Pipeline Delivery Time in Prefetch, Groups CalculatePixelDeliveryTimes( @@ -11893,15 +11821,15 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.MinTTUVBlank[k] = mode_lib->mp.TCalc + mode_lib->mp.MinTTUVBlank[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, MinTTUVBlank = %f (before vstartup margin)\n", __func__, k, mode_lib->mp.MinTTUVBlank[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, MinTTUVBlank = %f (before vstartup margin)\n", __func__, k, mode_lib->mp.MinTTUVBlank[k]); #endif s->Tvstartup_margin = (s->MaxVStartupLines[k] - mode_lib->mp.VStartupMin[k]) * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000); mode_lib->mp.MinTTUVBlank[k] = mode_lib->mp.MinTTUVBlank[k] + s->Tvstartup_margin; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, Tvstartup_margin = %f\n", __func__, k, s->Tvstartup_margin); - dml2_printf("DML::%s: k=%u, MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); - dml2_printf("DML::%s: k=%u, MinTTUVBlank = %f\n", __func__, k, mode_lib->mp.MinTTUVBlank[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, Tvstartup_margin = %f\n", __func__, k, s->Tvstartup_margin); + DML_LOG_VERBOSE("DML::%s: k=%u, MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, MinTTUVBlank = %f\n", __func__, k, mode_lib->mp.MinTTUVBlank[k]); #endif mode_lib->mp.Tdmdl[k] = mode_lib->mp.Tdmdl[k] + s->Tvstartup_margin; @@ -11920,9 +11848,9 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex s->blank_lines_remaining = (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total - display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_active) - mode_lib->mp.VStartup[k]; if (s->blank_lines_remaining < 0) { - dml2_printf("ERROR: Vstartup is larger than vblank!?\n"); + DML_LOG_VERBOSE("ERROR: Vstartup is larger than vblank!?\n"); s->blank_lines_remaining = 0; - DML2_ASSERT(0); + DML_ASSERT(0); } mode_lib->mp.MIN_DST_Y_NEXT_START[k] = s->dlg_vblank_start + s->blank_lines_remaining + s->LSetup; @@ -11936,18 +11864,18 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.VREADY_AT_OR_AFTER_VSYNC[k] = false; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, VStartup = %u (max)\n", __func__, k, mode_lib->mp.VStartup[k]); - dml2_printf("DML::%s: k=%u, VStartupMin = %u (max)\n", __func__, k, mode_lib->mp.VStartupMin[k]); - dml2_printf("DML::%s: k=%u, VUpdateOffsetPix = %u\n", __func__, k, mode_lib->mp.VUpdateOffsetPix[k]); - dml2_printf("DML::%s: k=%u, VUpdateWidthPix = %u\n", __func__, k, mode_lib->mp.VUpdateWidthPix[k]); - dml2_printf("DML::%s: k=%u, VReadyOffsetPix = %u\n", __func__, k, mode_lib->mp.VReadyOffsetPix[k]); - dml2_printf("DML::%s: k=%u, HTotal = %u\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total); - dml2_printf("DML::%s: k=%u, VTotal = %u\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total); - dml2_printf("DML::%s: k=%u, VActive = %u\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_active); - dml2_printf("DML::%s: k=%u, VFrontPorch = %u\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_front_porch); - dml2_printf("DML::%s: k=%u, TSetup = %f\n", __func__, k, mode_lib->mp.TSetup[k]); - dml2_printf("DML::%s: k=%u, MIN_DST_Y_NEXT_START = %f\n", __func__, k, mode_lib->mp.MIN_DST_Y_NEXT_START[k]); - dml2_printf("DML::%s: k=%u, VREADY_AT_OR_AFTER_VSYNC = %u\n", __func__, k, mode_lib->mp.VREADY_AT_OR_AFTER_VSYNC[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, VStartup = %u (max)\n", __func__, k, mode_lib->mp.VStartup[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, VStartupMin = %u (max)\n", __func__, k, mode_lib->mp.VStartupMin[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, VUpdateOffsetPix = %u\n", __func__, k, mode_lib->mp.VUpdateOffsetPix[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, VUpdateWidthPix = %u\n", __func__, k, mode_lib->mp.VUpdateWidthPix[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, VReadyOffsetPix = %u\n", __func__, k, mode_lib->mp.VReadyOffsetPix[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, HTotal = %lu\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total); + DML_LOG_VERBOSE("DML::%s: k=%u, VTotal = %lu\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total); + DML_LOG_VERBOSE("DML::%s: k=%u, VActive = %lu\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_active); + DML_LOG_VERBOSE("DML::%s: k=%u, VFrontPorch = %lu\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_front_porch); + DML_LOG_VERBOSE("DML::%s: k=%u, TSetup = %f\n", __func__, k, mode_lib->mp.TSetup[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, MIN_DST_Y_NEXT_START = %f\n", __func__, k, mode_lib->mp.MIN_DST_Y_NEXT_START[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, VREADY_AT_OR_AFTER_VSYNC = %u\n", __func__, k, mode_lib->mp.VREADY_AT_OR_AFTER_VSYNC[k]); #endif } @@ -11969,9 +11897,9 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex for (k = 0; k < s->num_active_planes; ++k) { mode_lib->mp.TotalDataReadBandwidth = mode_lib->mp.TotalDataReadBandwidth + mode_lib->mp.vactive_sw_bw_l[k] + mode_lib->mp.vactive_sw_bw_c[k]; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, mode_lib->mp.TotalDataReadBandwidth); - dml2_printf("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]); - dml2_printf("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, mode_lib->mp.TotalDataReadBandwidth); + DML_LOG_VERBOSE("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]); + DML_LOG_VERBOSE("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]); #endif } @@ -12051,28 +11979,28 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex min_return_latency_in_DCFCLK_cycles = (min_return_uclk_cycles / max_uclk_mhz + min_return_fclk_cycles / max_fclk_mhz) * hard_minimum_dcfclk_mhz; mode_lib->mp.min_return_latency_in_dcfclk = (unsigned int)min_return_latency_in_DCFCLK_cycles; mode_lib->mp.dcfclk_deep_sleep_hysteresis = (unsigned int)math_max2(32, (double)mode_lib->ip.pixel_chunk_size_kbytes * 1024 * 3 / 4 / 64 - min_return_latency_in_DCFCLK_cycles); - DML2_ASSERT(mode_lib->mp.dcfclk_deep_sleep_hysteresis < 256); + DML_ASSERT(mode_lib->mp.dcfclk_deep_sleep_hysteresis < 256); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: max_fclk_mhz = %f\n", __func__, max_fclk_mhz); - dml2_printf("DML::%s: max_uclk_mhz = %f\n", __func__, max_uclk_mhz); - dml2_printf("DML::%s: hard_minimum_dcfclk_mhz = %f\n", __func__, hard_minimum_dcfclk_mhz); - dml2_printf("DML::%s: min_return_uclk_cycles = %d\n", __func__, min_return_uclk_cycles); - dml2_printf("DML::%s: min_return_fclk_cycles = %d\n", __func__, min_return_fclk_cycles); - dml2_printf("DML::%s: min_return_latency_in_DCFCLK_cycles = %f\n", __func__, min_return_latency_in_DCFCLK_cycles); - dml2_printf("DML::%s: dcfclk_deep_sleep_hysteresis = %d \n", __func__, mode_lib->mp.dcfclk_deep_sleep_hysteresis); - dml2_printf("DML::%s: --- END --- \n", __func__); + DML_LOG_VERBOSE("DML::%s: max_fclk_mhz = %f\n", __func__, max_fclk_mhz); + DML_LOG_VERBOSE("DML::%s: max_uclk_mhz = %f\n", __func__, max_uclk_mhz); + DML_LOG_VERBOSE("DML::%s: hard_minimum_dcfclk_mhz = %f\n", __func__, hard_minimum_dcfclk_mhz); + DML_LOG_VERBOSE("DML::%s: min_return_uclk_cycles = %ld\n", __func__, min_return_uclk_cycles); + DML_LOG_VERBOSE("DML::%s: min_return_fclk_cycles = %ld\n", __func__, min_return_fclk_cycles); + DML_LOG_VERBOSE("DML::%s: min_return_latency_in_DCFCLK_cycles = %f\n", __func__, min_return_latency_in_DCFCLK_cycles); + DML_LOG_VERBOSE("DML::%s: dcfclk_deep_sleep_hysteresis = %d \n", __func__, mode_lib->mp.dcfclk_deep_sleep_hysteresis); + DML_LOG_VERBOSE("DML::%s: --- END --- \n", __func__); #endif return (in_out_params->mode_lib->mp.PrefetchAndImmediateFlipSupported); } bool dml2_core_calcs_mode_programming_ex(struct dml2_core_calcs_mode_programming_ex *in_out_params) { - dml2_printf("DML::%s: ------------- START ----------\n", __func__); + DML_LOG_VERBOSE("DML::%s: ------------- START ----------\n", __func__); bool result = dml_core_mode_programming(in_out_params); - dml2_printf("DML::%s: result = %0d\n", __func__, result); - dml2_printf("DML::%s: ------------- DONE ----------\n", __func__); + DML_LOG_VERBOSE("DML::%s: result = %0d\n", __func__, result); + DML_LOG_VERBOSE("DML::%s: ------------- DONE ----------\n", __func__); return result; } @@ -12130,16 +12058,16 @@ void dml2_core_calcs_get_dpte_row_height( unsigned int MacroTileHeight = is_plane1 ? MacroTileHeightC : MacroTileHeightY; unsigned int PTEBufferSizeInRequests = is_plane1 ? mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma : mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML: %s: is_plane1 = %u\n", __func__, is_plane1); - dml2_printf("DML: %s: BytePerPixel = %u\n", __func__, BytePerPixel); - dml2_printf("DML: %s: BlockHeight256Bytes = %u\n", __func__, BlockHeight256Bytes); - dml2_printf("DML: %s: BlockWidth256Bytes = %u\n", __func__, BlockWidth256Bytes); - dml2_printf("DML: %s: MacroTileWidth = %u\n", __func__, MacroTileWidth); - dml2_printf("DML: %s: MacroTileHeight = %u\n", __func__, MacroTileHeight); - dml2_printf("DML: %s: PTEBufferSizeInRequests = %u\n", __func__, PTEBufferSizeInRequests); - dml2_printf("DML: %s: dpte_buffer_size_in_pte_reqs_luma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma); - dml2_printf("DML: %s: dpte_buffer_size_in_pte_reqs_chroma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma); - dml2_printf("DML: %s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes); + DML_LOG_VERBOSE("DML: %s: is_plane1 = %u\n", __func__, is_plane1); + DML_LOG_VERBOSE("DML: %s: BytePerPixel = %u\n", __func__, BytePerPixel); + DML_LOG_VERBOSE("DML: %s: BlockHeight256Bytes = %u\n", __func__, BlockHeight256Bytes); + DML_LOG_VERBOSE("DML: %s: BlockWidth256Bytes = %u\n", __func__, BlockWidth256Bytes); + DML_LOG_VERBOSE("DML: %s: MacroTileWidth = %u\n", __func__, MacroTileWidth); + DML_LOG_VERBOSE("DML: %s: MacroTileHeight = %u\n", __func__, MacroTileHeight); + DML_LOG_VERBOSE("DML: %s: PTEBufferSizeInRequests = %u\n", __func__, PTEBufferSizeInRequests); + DML_LOG_VERBOSE("DML: %s: dpte_buffer_size_in_pte_reqs_luma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma); + DML_LOG_VERBOSE("DML: %s: dpte_buffer_size_in_pte_reqs_chroma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma); + DML_LOG_VERBOSE("DML: %s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes); #endif unsigned int dummy_integer[21]; @@ -12193,16 +12121,16 @@ void dml2_core_calcs_get_dpte_row_height( CalculateVMAndRowBytes(&mode_lib->scratch.calculate_vm_and_row_bytes_params); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML: %s: dpte_row_height = %u\n", __func__, *dpte_row_height); + DML_LOG_VERBOSE("DML: %s: dpte_row_height = %u\n", __func__, *dpte_row_height); #endif } static bool is_dual_plane(enum dml2_source_format_class source_format) { - bool ret_val = 0; + bool ret_val = false; if ((source_format == dml2_420_12) || (source_format == dml2_420_8) || (source_format == dml2_420_10) || (source_format == dml2_rgbe_alpha)) - ret_val = 1; + ret_val = true; return ret_val; } @@ -12220,6 +12148,8 @@ static void rq_dlg_get_wm_regs(const struct dml2_display_cfg *display_cfg, const wm_regs->fclk_pstate = (int unsigned)(mode_lib->mp.Watermark.FCLKChangeWatermark * refclk_freq_in_mhz); wm_regs->sr_enter = (int unsigned)(mode_lib->mp.Watermark.StutterEnterPlusExitWatermark * refclk_freq_in_mhz); wm_regs->sr_exit = (int unsigned)(mode_lib->mp.Watermark.StutterExitWatermark * refclk_freq_in_mhz); + wm_regs->sr_enter_z8 = (int unsigned)(mode_lib->mp.Watermark.Z8StutterEnterPlusExitWatermark * refclk_freq_in_mhz); + wm_regs->sr_exit_z8 = (int unsigned)(mode_lib->mp.Watermark.Z8StutterExitWatermark * refclk_freq_in_mhz); wm_regs->temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us * refclk_freq_in_mhz); wm_regs->uclk_pstate = (int unsigned)(mode_lib->mp.Watermark.DRAMClockChangeWatermark * refclk_freq_in_mhz); wm_regs->urgent = (int unsigned)(mode_lib->mp.Watermark.UrgentWatermark * refclk_freq_in_mhz); @@ -12246,11 +12176,11 @@ void dml2_core_calcs_cursor_dlg_reg(struct dml2_cursor_dlg_regs *cursor_dlg_regs cursor_dlg_regs->dst_x_offset = (unsigned int) ((dst_x_offset > 0) ? dst_x_offset : 0); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML_DLG::%s: cursor_x_position=%d\n", __func__, p->cursor_x_position); - dml2_printf("DML_DLG::%s: dlg_refclk_mhz=%f\n", __func__, p->dlg_refclk_mhz); - dml2_printf("DML_DLG::%s: pixel_rate_mhz=%f\n", __func__, p->pixel_rate_mhz); - dml2_printf("DML_DLG::%s: dst_x_offset=%d\n", __func__, dst_x_offset); - dml2_printf("DML_DLG::%s: dst_x_offset=%d (reg)\n", __func__, cursor_dlg_regs->dst_x_offset); + DML_LOG_VERBOSE("DML_DLG::%s: cursor_x_position=%d\n", __func__, p->cursor_x_position); + DML_LOG_VERBOSE("DML_DLG::%s: dlg_refclk_mhz=%f\n", __func__, p->dlg_refclk_mhz); + DML_LOG_VERBOSE("DML_DLG::%s: pixel_rate_mhz=%f\n", __func__, p->pixel_rate_mhz); + DML_LOG_VERBOSE("DML_DLG::%s: dst_x_offset=%d\n", __func__, dst_x_offset); + DML_LOG_VERBOSE("DML_DLG::%s: dst_x_offset=%d (reg)\n", __func__, cursor_dlg_regs->dst_x_offset); #endif cursor_dlg_regs->chunk_hdl_adjust = 3; @@ -12286,7 +12216,7 @@ static void rq_dlg_get_rq_reg(struct dml2_display_rq_regs *rq_regs, double stored_swath_c_bytes; bool is_phantom_pipe; - dml2_printf("DML_DLG::%s: Calculation for pipe[%d] start\n", __func__, pipe_idx); + DML_LOG_VERBOSE("DML_DLG::%s: Calculation for pipe[%d] start\n", __func__, pipe_idx); pixel_chunk_bytes = (unsigned int)(mode_lib->ip.pixel_chunk_size_kbytes * 1024); min_pixel_chunk_bytes = (unsigned int)(mode_lib->ip.min_pixel_chunk_size_bytes); @@ -12329,19 +12259,19 @@ static void rq_dlg_get_rq_reg(struct dml2_display_rq_regs *rq_regs, if (sw_mode == dml2_sw_linear && display_cfg->gpuvm_enable) { unsigned int p0_pte_row_height_linear = (unsigned int)(dml_get_dpte_row_height_linear_l(mode_lib, pipe_idx)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML_DLG: %s: p0_pte_row_height_linear = %u\n", __func__, p0_pte_row_height_linear); + DML_LOG_VERBOSE("DML_DLG: %s: p0_pte_row_height_linear = %u\n", __func__, p0_pte_row_height_linear); #endif - DML2_ASSERT(p0_pte_row_height_linear >= 8); + DML_ASSERT(p0_pte_row_height_linear >= 8); rq_regs->rq_regs_l.pte_row_height_linear = math_log2_approx(p0_pte_row_height_linear) - 3; if (dual_plane) { unsigned int p1_pte_row_height_linear = (unsigned int)(dml_get_dpte_row_height_linear_c(mode_lib, pipe_idx)); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML_DLG: %s: p1_pte_row_height_linear = %u\n", __func__, p1_pte_row_height_linear); + DML_LOG_VERBOSE("DML_DLG: %s: p1_pte_row_height_linear = %u\n", __func__, p1_pte_row_height_linear); #endif if (sw_mode == dml2_sw_linear) { - DML2_ASSERT(p1_pte_row_height_linear >= 8); + DML_ASSERT(p1_pte_row_height_linear >= 8); } rq_regs->rq_regs_c.pte_row_height_linear = math_log2_approx(p1_pte_row_height_linear) - 3; } @@ -12375,12 +12305,12 @@ static void rq_dlg_get_rq_reg(struct dml2_display_rq_regs *rq_regs, if (stored_swath_l_bytes / stored_swath_c_bytes <= 1.5) { detile_buf_plane1_addr = (unsigned int)(detile_buf_size_in_bytes / 2.0 / 1024.0); // half to chroma #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML_DLG: %s: detile_buf_plane1_addr = %d (1/2 to chroma)\n", __func__, detile_buf_plane1_addr); + DML_LOG_VERBOSE("DML_DLG: %s: detile_buf_plane1_addr = %d (1/2 to chroma)\n", __func__, detile_buf_plane1_addr); #endif } else { detile_buf_plane1_addr = (unsigned int)(dml_round_to_multiple((unsigned int)((2.0 * detile_buf_size_in_bytes) / 3.0), 1024, 0) / 1024.0); // 2/3 to luma #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML_DLG: %s: detile_buf_plane1_addr = %d (1/3 chroma)\n", __func__, detile_buf_plane1_addr); + DML_LOG_VERBOSE("DML_DLG: %s: detile_buf_plane1_addr = %d (1/3 chroma)\n", __func__, detile_buf_plane1_addr); #endif } } @@ -12388,15 +12318,15 @@ static void rq_dlg_get_rq_reg(struct dml2_display_rq_regs *rq_regs, rq_regs->plane1_base_address = detile_buf_plane1_addr; #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML_DLG: %s: is_phantom_pipe = %d\n", __func__, is_phantom_pipe); - dml2_printf("DML_DLG: %s: stored_swath_l_bytes = %f\n", __func__, stored_swath_l_bytes); - dml2_printf("DML_DLG: %s: stored_swath_c_bytes = %f\n", __func__, stored_swath_c_bytes); - dml2_printf("DML_DLG: %s: detile_buf_size_in_bytes = %d\n", __func__, detile_buf_size_in_bytes); - dml2_printf("DML_DLG: %s: detile_buf_plane1_addr = %d\n", __func__, detile_buf_plane1_addr); - dml2_printf("DML_DLG: %s: plane1_base_address = %d\n", __func__, rq_regs->plane1_base_address); + DML_LOG_VERBOSE("DML_DLG: %s: is_phantom_pipe = %d\n", __func__, is_phantom_pipe); + DML_LOG_VERBOSE("DML_DLG: %s: stored_swath_l_bytes = %f\n", __func__, stored_swath_l_bytes); + DML_LOG_VERBOSE("DML_DLG: %s: stored_swath_c_bytes = %f\n", __func__, stored_swath_c_bytes); + DML_LOG_VERBOSE("DML_DLG: %s: detile_buf_size_in_bytes = %d\n", __func__, detile_buf_size_in_bytes); + DML_LOG_VERBOSE("DML_DLG: %s: detile_buf_plane1_addr = %d\n", __func__, detile_buf_plane1_addr); + DML_LOG_VERBOSE("DML_DLG: %s: plane1_base_address = %d\n", __func__, rq_regs->plane1_base_address); #endif - //dml2_printf_rq_regs_st(rq_regs); - dml2_printf("DML_DLG::%s: Calculation for pipe[%d] done\n", __func__, pipe_idx); + //DML_LOG_VERBOSE_rq_regs_st(rq_regs); + DML_LOG_VERBOSE("DML_DLG::%s: Calculation for pipe[%d] done\n", __func__, pipe_idx); } static void rq_dlg_get_dlg_reg( @@ -12411,10 +12341,10 @@ static void rq_dlg_get_dlg_reg( memset(l, 0, sizeof(struct dml2_core_shared_rq_dlg_get_dlg_reg_locals)); - dml2_printf("DML_DLG::%s: Calculation for pipe_idx=%d\n", __func__, pipe_idx); + DML_LOG_VERBOSE("DML_DLG::%s: Calculation for pipe_idx=%d\n", __func__, pipe_idx); l->plane_idx = dml_get_plane_idx(mode_lib, pipe_idx); - DML2_ASSERT(l->plane_idx < DML2_MAX_PLANES); + DML_ASSERT(l->plane_idx < DML2_MAX_PLANES); l->source_format = dml2_444_8; l->odm_mode = dml2_odm_mode_bypass; @@ -12444,18 +12374,18 @@ static void rq_dlg_get_dlg_reg( l->pclk_freq_in_mhz = (double)l->timing->pixel_clock_khz / 1000; l->ref_freq_to_pix_freq = l->refclk_freq_in_mhz / l->pclk_freq_in_mhz; - dml2_printf("DML_DLG::%s: plane_idx = %d\n", __func__, l->plane_idx); - dml2_printf("DML_DLG: %s: htotal = %d\n", __func__, l->htotal); - dml2_printf("DML_DLG: %s: refclk_freq_in_mhz = %3.2f\n", __func__, l->refclk_freq_in_mhz); - dml2_printf("DML_DLG: %s: dlg_ref_clk_mhz = %3.2f\n", __func__, display_cfg->overrides.hw.dlg_ref_clk_mhz); - dml2_printf("DML_DLG: %s: soc.refclk_mhz = %3.2f\n", __func__, mode_lib->soc.dchub_refclk_mhz); - dml2_printf("DML_DLG: %s: pclk_freq_in_mhz = %3.2f\n", __func__, l->pclk_freq_in_mhz); - dml2_printf("DML_DLG: %s: ref_freq_to_pix_freq = %3.2f\n", __func__, l->ref_freq_to_pix_freq); - dml2_printf("DML_DLG: %s: interlaced = %d\n", __func__, l->interlaced); + DML_LOG_VERBOSE("DML_DLG::%s: plane_idx = %d\n", __func__, l->plane_idx); + DML_LOG_VERBOSE("DML_DLG: %s: htotal = %d\n", __func__, l->htotal); + DML_LOG_VERBOSE("DML_DLG: %s: refclk_freq_in_mhz = %3.2f\n", __func__, l->refclk_freq_in_mhz); + DML_LOG_VERBOSE("DML_DLG: %s: dlg_ref_clk_mhz = %3.2f\n", __func__, display_cfg->overrides.hw.dlg_ref_clk_mhz); + DML_LOG_VERBOSE("DML_DLG: %s: soc.refclk_mhz = %d\n", __func__, mode_lib->soc.dchub_refclk_mhz); + DML_LOG_VERBOSE("DML_DLG: %s: pclk_freq_in_mhz = %3.2f\n", __func__, l->pclk_freq_in_mhz); + DML_LOG_VERBOSE("DML_DLG: %s: ref_freq_to_pix_freq = %3.2f\n", __func__, l->ref_freq_to_pix_freq); + DML_LOG_VERBOSE("DML_DLG: %s: interlaced = %d\n", __func__, l->interlaced); - DML2_ASSERT(l->refclk_freq_in_mhz != 0); - DML2_ASSERT(l->pclk_freq_in_mhz != 0); - DML2_ASSERT(l->ref_freq_to_pix_freq < 4.0); + DML_ASSERT(l->refclk_freq_in_mhz != 0); + DML_ASSERT(l->pclk_freq_in_mhz != 0); + DML_ASSERT(l->ref_freq_to_pix_freq < 4.0); // Need to figure out which side of odm combine we're in // Assume the pipe instance under the same plane is in order @@ -12484,14 +12414,14 @@ static void rq_dlg_get_dlg_reg( l->pipe_idx_in_combine = pipe_idx - l->first_pipe_idx_in_plane; // DML assumes the pipes in the same plane will have continuous indexing (i.e. plane 0 use pipe 0, 1, and plane 1 uses pipe 2, 3, etc.) disp_dlg_regs->refcyc_h_blank_end = (unsigned int)(((double)l->hblank_end + (double)l->pipe_idx_in_combine * (double)l->hactive / (double)l->odm_combine_factor) * l->ref_freq_to_pix_freq); - dml2_printf("DML_DLG: %s: pipe_idx = %d\n", __func__, pipe_idx); - dml2_printf("DML_DLG: %s: first_pipe_idx_in_plane = %d\n", __func__, l->first_pipe_idx_in_plane); - dml2_printf("DML_DLG: %s: pipe_idx_in_combine = %d\n", __func__, l->pipe_idx_in_combine); - dml2_printf("DML_DLG: %s: odm_combine_factor = %d\n", __func__, l->odm_combine_factor); + DML_LOG_VERBOSE("DML_DLG: %s: pipe_idx = %d\n", __func__, pipe_idx); + DML_LOG_VERBOSE("DML_DLG: %s: first_pipe_idx_in_plane = %d\n", __func__, l->first_pipe_idx_in_plane); + DML_LOG_VERBOSE("DML_DLG: %s: pipe_idx_in_combine = %d\n", __func__, l->pipe_idx_in_combine); + DML_LOG_VERBOSE("DML_DLG: %s: odm_combine_factor = %d\n", __func__, l->odm_combine_factor); } - dml2_printf("DML_DLG: %s: refcyc_h_blank_end = %d\n", __func__, disp_dlg_regs->refcyc_h_blank_end); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_h_blank_end = %d\n", __func__, disp_dlg_regs->refcyc_h_blank_end); - DML2_ASSERT(disp_dlg_regs->refcyc_h_blank_end < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->refcyc_h_blank_end < (unsigned int)math_pow(2, 13)); disp_dlg_regs->ref_freq_to_pix_freq = (unsigned int)(l->ref_freq_to_pix_freq * math_pow(2, 19)); disp_dlg_regs->refcyc_per_htotal = (unsigned int)(l->ref_freq_to_pix_freq * (double)l->htotal * math_pow(2, 8)); @@ -12500,20 +12430,20 @@ static void rq_dlg_get_dlg_reg( l->min_ttu_vblank = mode_lib->mp.MinTTUVBlank[mode_lib->mp.pipe_plane[pipe_idx]]; l->min_dst_y_next_start = (unsigned int)(mode_lib->mp.MIN_DST_Y_NEXT_START[mode_lib->mp.pipe_plane[pipe_idx]]); - dml2_printf("DML_DLG: %s: min_ttu_vblank (us) = %3.2f\n", __func__, l->min_ttu_vblank); - dml2_printf("DML_DLG: %s: min_dst_y_next_start = %d\n", __func__, l->min_dst_y_next_start); - dml2_printf("DML_DLG: %s: ref_freq_to_pix_freq = %3.2f\n", __func__, l->ref_freq_to_pix_freq); + DML_LOG_VERBOSE("DML_DLG: %s: min_ttu_vblank (us) = %3.2f\n", __func__, l->min_ttu_vblank); + DML_LOG_VERBOSE("DML_DLG: %s: min_dst_y_next_start = %d\n", __func__, l->min_dst_y_next_start); + DML_LOG_VERBOSE("DML_DLG: %s: ref_freq_to_pix_freq = %3.2f\n", __func__, l->ref_freq_to_pix_freq); l->vready_after_vcount0 = (unsigned int)(mode_lib->mp.VREADY_AT_OR_AFTER_VSYNC[mode_lib->mp.pipe_plane[pipe_idx]]); disp_dlg_regs->vready_after_vcount0 = l->vready_after_vcount0; - dml2_printf("DML_DLG: %s: vready_after_vcount0 = %d\n", __func__, disp_dlg_regs->vready_after_vcount0); + DML_LOG_VERBOSE("DML_DLG: %s: vready_after_vcount0 = %d\n", __func__, disp_dlg_regs->vready_after_vcount0); l->dst_x_after_scaler = (unsigned int)(mode_lib->mp.DSTXAfterScaler[mode_lib->mp.pipe_plane[pipe_idx]]); l->dst_y_after_scaler = (unsigned int)(mode_lib->mp.DSTYAfterScaler[mode_lib->mp.pipe_plane[pipe_idx]]); - dml2_printf("DML_DLG: %s: dst_x_after_scaler = %d\n", __func__, l->dst_x_after_scaler); - dml2_printf("DML_DLG: %s: dst_y_after_scaler = %d\n", __func__, l->dst_y_after_scaler); + DML_LOG_VERBOSE("DML_DLG: %s: dst_x_after_scaler = %d\n", __func__, l->dst_x_after_scaler); + DML_LOG_VERBOSE("DML_DLG: %s: dst_y_after_scaler = %d\n", __func__, l->dst_y_after_scaler); l->dst_y_prefetch = mode_lib->mp.dst_y_prefetch[mode_lib->mp.pipe_plane[pipe_idx]]; l->dst_y_per_vm_vblank = mode_lib->mp.dst_y_per_vm_vblank[mode_lib->mp.pipe_plane[pipe_idx]]; @@ -12521,28 +12451,28 @@ static void rq_dlg_get_dlg_reg( l->dst_y_per_vm_flip = mode_lib->mp.dst_y_per_vm_flip[mode_lib->mp.pipe_plane[pipe_idx]]; l->dst_y_per_row_flip = mode_lib->mp.dst_y_per_row_flip[mode_lib->mp.pipe_plane[pipe_idx]]; - dml2_printf("DML_DLG: %s: dst_y_prefetch (after rnd) = %3.2f\n", __func__, l->dst_y_prefetch); - dml2_printf("DML_DLG: %s: dst_y_per_vm_flip = %3.2f\n", __func__, l->dst_y_per_vm_flip); - dml2_printf("DML_DLG: %s: dst_y_per_row_flip = %3.2f\n", __func__, l->dst_y_per_row_flip); - dml2_printf("DML_DLG: %s: dst_y_per_vm_vblank = %3.2f\n", __func__, l->dst_y_per_vm_vblank); - dml2_printf("DML_DLG: %s: dst_y_per_row_vblank = %3.2f\n", __func__, l->dst_y_per_row_vblank); + DML_LOG_VERBOSE("DML_DLG: %s: dst_y_prefetch (after rnd) = %3.2f\n", __func__, l->dst_y_prefetch); + DML_LOG_VERBOSE("DML_DLG: %s: dst_y_per_vm_flip = %3.2f\n", __func__, l->dst_y_per_vm_flip); + DML_LOG_VERBOSE("DML_DLG: %s: dst_y_per_row_flip = %3.2f\n", __func__, l->dst_y_per_row_flip); + DML_LOG_VERBOSE("DML_DLG: %s: dst_y_per_vm_vblank = %3.2f\n", __func__, l->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML_DLG: %s: dst_y_per_row_vblank = %3.2f\n", __func__, l->dst_y_per_row_vblank); if (l->dst_y_prefetch > 0 && l->dst_y_per_vm_vblank > 0 && l->dst_y_per_row_vblank > 0) { - DML2_ASSERT(l->dst_y_prefetch > (l->dst_y_per_vm_vblank + l->dst_y_per_row_vblank)); + DML_ASSERT(l->dst_y_prefetch > (l->dst_y_per_vm_vblank + l->dst_y_per_row_vblank)); } l->vratio_pre_l = mode_lib->mp.VRatioPrefetchY[mode_lib->mp.pipe_plane[pipe_idx]]; l->vratio_pre_c = mode_lib->mp.VRatioPrefetchC[mode_lib->mp.pipe_plane[pipe_idx]]; - dml2_printf("DML_DLG: %s: vratio_pre_l = %3.2f\n", __func__, l->vratio_pre_l); - dml2_printf("DML_DLG: %s: vratio_pre_c = %3.2f\n", __func__, l->vratio_pre_c); + DML_LOG_VERBOSE("DML_DLG: %s: vratio_pre_l = %3.2f\n", __func__, l->vratio_pre_l); + DML_LOG_VERBOSE("DML_DLG: %s: vratio_pre_c = %3.2f\n", __func__, l->vratio_pre_c); // Active l->refcyc_per_line_delivery_pre_l = mode_lib->mp.DisplayPipeLineDeliveryTimeLumaPrefetch[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; l->refcyc_per_line_delivery_l = mode_lib->mp.DisplayPipeLineDeliveryTimeLuma[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; - dml2_printf("DML_DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f\n", __func__, l->refcyc_per_line_delivery_pre_l); - dml2_printf("DML_DLG: %s: refcyc_per_line_delivery_l = %3.2f\n", __func__, l->refcyc_per_line_delivery_l); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f\n", __func__, l->refcyc_per_line_delivery_pre_l); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_line_delivery_l = %3.2f\n", __func__, l->refcyc_per_line_delivery_l); l->refcyc_per_line_delivery_pre_c = 0.0; l->refcyc_per_line_delivery_c = 0.0; @@ -12551,8 +12481,8 @@ static void rq_dlg_get_dlg_reg( l->refcyc_per_line_delivery_pre_c = mode_lib->mp.DisplayPipeLineDeliveryTimeChromaPrefetch[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; l->refcyc_per_line_delivery_c = mode_lib->mp.DisplayPipeLineDeliveryTimeChroma[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; - dml2_printf("DML_DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f\n", __func__, l->refcyc_per_line_delivery_pre_c); - dml2_printf("DML_DLG: %s: refcyc_per_line_delivery_c = %3.2f\n", __func__, l->refcyc_per_line_delivery_c); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f\n", __func__, l->refcyc_per_line_delivery_pre_c); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_line_delivery_c = %3.2f\n", __func__, l->refcyc_per_line_delivery_c); } disp_dlg_regs->refcyc_per_vm_dmdata = (unsigned int)(mode_lib->mp.Tdmdl_vm[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz); @@ -12561,8 +12491,8 @@ static void rq_dlg_get_dlg_reg( l->refcyc_per_req_delivery_pre_l = mode_lib->mp.DisplayPipeRequestDeliveryTimeLumaPrefetch[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; l->refcyc_per_req_delivery_l = mode_lib->mp.DisplayPipeRequestDeliveryTimeLuma[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; - dml2_printf("DML_DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f\n", __func__, l->refcyc_per_req_delivery_pre_l); - dml2_printf("DML_DLG: %s: refcyc_per_req_delivery_l = %3.2f\n", __func__, l->refcyc_per_req_delivery_l); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f\n", __func__, l->refcyc_per_req_delivery_pre_l); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_req_delivery_l = %3.2f\n", __func__, l->refcyc_per_req_delivery_l); l->refcyc_per_req_delivery_pre_c = 0.0; l->refcyc_per_req_delivery_c = 0.0; @@ -12570,16 +12500,16 @@ static void rq_dlg_get_dlg_reg( l->refcyc_per_req_delivery_pre_c = mode_lib->mp.DisplayPipeRequestDeliveryTimeChromaPrefetch[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; l->refcyc_per_req_delivery_c = mode_lib->mp.DisplayPipeRequestDeliveryTimeChroma[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz; - dml2_printf("DML_DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f\n", __func__, l->refcyc_per_req_delivery_pre_c); - dml2_printf("DML_DLG: %s: refcyc_per_req_delivery_c = %3.2f\n", __func__, l->refcyc_per_req_delivery_c); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f\n", __func__, l->refcyc_per_req_delivery_pre_c); + DML_LOG_VERBOSE("DML_DLG: %s: refcyc_per_req_delivery_c = %3.2f\n", __func__, l->refcyc_per_req_delivery_c); } // TTU - Cursor - DML2_ASSERT(display_cfg->plane_descriptors[l->plane_idx].cursor.num_cursors <= 1); + DML_ASSERT(display_cfg->plane_descriptors[l->plane_idx].cursor.num_cursors <= 1); // Assign to register structures disp_dlg_regs->min_dst_y_next_start = (unsigned int)((double)l->min_dst_y_next_start * math_pow(2, 2)); - DML2_ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int)math_pow(2, 18)); + DML_ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int)math_pow(2, 18)); disp_dlg_regs->dst_y_after_scaler = l->dst_y_after_scaler; // in terms of line disp_dlg_regs->refcyc_x_after_scaler = (unsigned int)((double)l->dst_x_after_scaler * l->ref_freq_to_pix_freq); // in terms of refclk @@ -12592,10 +12522,10 @@ static void rq_dlg_get_dlg_reg( disp_dlg_regs->vratio_prefetch = (unsigned int)(l->vratio_pre_l * math_pow(2, 19)); disp_dlg_regs->vratio_prefetch_c = (unsigned int)(l->vratio_pre_c * math_pow(2, 19)); - dml2_printf("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_vblank); - dml2_printf("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_vblank); - dml2_printf("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip); - dml2_printf("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip); + DML_LOG_VERBOSE("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_vblank); + DML_LOG_VERBOSE("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_vblank); + DML_LOG_VERBOSE("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip); + DML_LOG_VERBOSE("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip); disp_dlg_regs->refcyc_per_vm_group_vblank = (unsigned int)(mode_lib->mp.TimePerVMGroupVBlank[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz); disp_dlg_regs->refcyc_per_vm_group_flip = (unsigned int)(mode_lib->mp.TimePerVMGroupFlip[mode_lib->mp.pipe_plane[pipe_idx]] * l->refclk_freq_in_mhz); @@ -12662,11 +12592,11 @@ static void rq_dlg_get_dlg_reg( disp_ttu_regs->qos_ramp_disable_c = 0; disp_ttu_regs->min_ttu_vblank = (unsigned int)(l->min_ttu_vblank * l->refclk_freq_in_mhz); - // CHECK for HW registers' range, DML2_ASSERT or clamp - DML2_ASSERT(l->refcyc_per_req_delivery_pre_l < math_pow(2, 13)); - DML2_ASSERT(l->refcyc_per_req_delivery_l < math_pow(2, 13)); - DML2_ASSERT(l->refcyc_per_req_delivery_pre_c < math_pow(2, 13)); - DML2_ASSERT(l->refcyc_per_req_delivery_c < math_pow(2, 13)); + // CHECK for HW registers' range, DML_ASSERT or clamp + DML_ASSERT(l->refcyc_per_req_delivery_pre_l < math_pow(2, 13)); + DML_ASSERT(l->refcyc_per_req_delivery_l < math_pow(2, 13)); + DML_ASSERT(l->refcyc_per_req_delivery_pre_c < math_pow(2, 13)); + DML_ASSERT(l->refcyc_per_req_delivery_c < math_pow(2, 13)); if (disp_dlg_regs->refcyc_per_vm_group_vblank >= (unsigned int)math_pow(2, 23)) disp_dlg_regs->refcyc_per_vm_group_vblank = (unsigned int)(math_pow(2, 23) - 1); @@ -12680,16 +12610,16 @@ static void rq_dlg_get_dlg_reg( disp_dlg_regs->refcyc_per_vm_req_flip = (unsigned int)(math_pow(2, 23) - 1); - DML2_ASSERT(disp_dlg_regs->dst_y_after_scaler < (unsigned int)8); - DML2_ASSERT(disp_dlg_regs->refcyc_x_after_scaler < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->dst_y_after_scaler < (unsigned int)8); + DML_ASSERT(disp_dlg_regs->refcyc_x_after_scaler < (unsigned int)math_pow(2, 13)); if (disp_dlg_regs->dst_y_per_pte_row_nom_l >= (unsigned int)math_pow(2, 17)) { - dml2_printf("DML_DLG: %s: Warning DST_Y_PER_PTE_ROW_NOM_L %u > register max U15.2 %u, clamp to max\n", __func__, disp_dlg_regs->dst_y_per_pte_row_nom_l, (unsigned int)math_pow(2, 17) - 1); + DML_LOG_VERBOSE("DML_DLG: %s: Warning DST_Y_PER_PTE_ROW_NOM_L %u > register max U15.2 %u, clamp to max\n", __func__, disp_dlg_regs->dst_y_per_pte_row_nom_l, (unsigned int)math_pow(2, 17) - 1); l->dst_y_per_pte_row_nom_l = (unsigned int)math_pow(2, 17) - 1; } if (l->dual_plane) { if (disp_dlg_regs->dst_y_per_pte_row_nom_c >= (unsigned int)math_pow(2, 17)) { - dml2_printf("DML_DLG: %s: Warning DST_Y_PER_PTE_ROW_NOM_C %u > register max U15.2 %u, clamp to max\n", __func__, disp_dlg_regs->dst_y_per_pte_row_nom_c, (unsigned int)math_pow(2, 17) - 1); + DML_LOG_VERBOSE("DML_DLG: %s: Warning DST_Y_PER_PTE_ROW_NOM_C %u > register max U15.2 %u, clamp to max\n", __func__, disp_dlg_regs->dst_y_per_pte_row_nom_c, (unsigned int)math_pow(2, 17) - 1); l->dst_y_per_pte_row_nom_c = (unsigned int)math_pow(2, 17) - 1; } } @@ -12700,20 +12630,20 @@ static void rq_dlg_get_dlg_reg( if (disp_dlg_regs->refcyc_per_pte_group_nom_c >= (unsigned int)math_pow(2, 23)) disp_dlg_regs->refcyc_per_pte_group_nom_c = (unsigned int)(math_pow(2, 23) - 1); } - DML2_ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)math_pow(2, 13)); if (l->dual_plane) { - DML2_ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_c < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_c < (unsigned int)math_pow(2, 13)); } - DML2_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_l < (unsigned int)math_pow(2, 13)); - DML2_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_l < (unsigned int)math_pow(2, 13)); - DML2_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_c < (unsigned int)math_pow(2, 13)); - DML2_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_c < (unsigned int)math_pow(2, 13)); - DML2_ASSERT(disp_ttu_regs->qos_level_low_wm < (unsigned int)math_pow(2, 14)); - DML2_ASSERT(disp_ttu_regs->qos_level_high_wm < (unsigned int)math_pow(2, 14)); - DML2_ASSERT(disp_ttu_regs->min_ttu_vblank < (unsigned int)math_pow(2, 24)); + DML_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_l < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_l < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_c < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_dlg_regs->refcyc_per_line_delivery_c < (unsigned int)math_pow(2, 13)); + DML_ASSERT(disp_ttu_regs->qos_level_low_wm < (unsigned int)math_pow(2, 14)); + DML_ASSERT(disp_ttu_regs->qos_level_high_wm < (unsigned int)math_pow(2, 14)); + DML_ASSERT(disp_ttu_regs->min_ttu_vblank < (unsigned int)math_pow(2, 24)); - dml2_printf("DML_DLG::%s: Calculation for pipe[%d] done\n", __func__, pipe_idx); + DML_LOG_VERBOSE("DML_DLG::%s: Calculation for pipe[%d] done\n", __func__, pipe_idx); } } @@ -12736,11 +12666,11 @@ static void rq_dlg_get_arb_params(const struct dml2_display_cfg *display_cfg, co arb_param->pstate_stall_threshold = (unsigned int)(mode_lib->ip_caps.fams2.max_allow_delay_us * refclk_freq_in_mhz); #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: max_req_outstanding = %d\n", __func__, arb_param->max_req_outstanding); - dml2_printf("DML::%s: sdpif_request_rate_limit = %d\n", __func__, arb_param->sdpif_request_rate_limit); - dml2_printf("DML::%s: compbuf_reserved_space_kbytes = %d\n", __func__, arb_param->compbuf_reserved_space_kbytes); - dml2_printf("DML::%s: allow_sdpif_rate_limit_when_cstate_req = %d\n", __func__, arb_param->allow_sdpif_rate_limit_when_cstate_req); - dml2_printf("DML::%s: dcfclk_deep_sleep_hysteresis = %d\n", __func__, arb_param->dcfclk_deep_sleep_hysteresis); + DML_LOG_VERBOSE("DML::%s: max_req_outstanding = %d\n", __func__, arb_param->max_req_outstanding); + DML_LOG_VERBOSE("DML::%s: sdpif_request_rate_limit = %d\n", __func__, arb_param->sdpif_request_rate_limit); + DML_LOG_VERBOSE("DML::%s: compbuf_reserved_space_kbytes = %d\n", __func__, arb_param->compbuf_reserved_space_kbytes); + DML_LOG_VERBOSE("DML::%s: allow_sdpif_rate_limit_when_cstate_req = %d\n", __func__, arb_param->allow_sdpif_rate_limit_when_cstate_req); + DML_LOG_VERBOSE("DML::%s: dcfclk_deep_sleep_hysteresis = %d\n", __func__, arb_param->dcfclk_deep_sleep_hysteresis); #endif } @@ -13013,10 +12943,10 @@ void dml2_core_calcs_get_stream_support_info(const struct dml2_display_cfg *disp out->vblank_reserved_time_us = display_cfg->plane_descriptors[plane_index].overrides.reserved_vblank_time_ns / 1000; #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: subvp_fw_processing_delay_us = %d\n", __func__, mode_lib->ip.subvp_fw_processing_delay_us); - dml2_printf("DML::%s: subvp_pstate_allow_width_us = %d\n", __func__, mode_lib->ip.subvp_pstate_allow_width_us); - dml2_printf("DML::%s: subvp_swath_height_margin_lines = %d\n", __func__, mode_lib->ip.subvp_swath_height_margin_lines); - dml2_printf("DML::%s: vblank_reserved_time_us = %f\n", __func__, out->vblank_reserved_time_us); + DML_LOG_VERBOSE("DML::%s: subvp_fw_processing_delay_us = %d\n", __func__, mode_lib->ip.subvp_fw_processing_delay_us); + DML_LOG_VERBOSE("DML::%s: subvp_pstate_allow_width_us = %d\n", __func__, mode_lib->ip.subvp_pstate_allow_width_us); + DML_LOG_VERBOSE("DML::%s: subvp_swath_height_margin_lines = %d\n", __func__, mode_lib->ip.subvp_swath_height_margin_lines); + DML_LOG_VERBOSE("DML::%s: vblank_reserved_time_us = %u\n", __func__, out->vblank_reserved_time_us); #endif } diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h index 4e502f0a6d20..bdee6ad7bc59 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h @@ -1078,6 +1078,8 @@ struct dml2_core_calcs_mode_programming_locals { enum dml2_source_format_class pixel_format[DML2_MAX_PLANES]; unsigned int lb_source_lines_l[DML2_MAX_PLANES]; unsigned int lb_source_lines_c[DML2_MAX_PLANES]; + unsigned int num_dsc_slices[DML2_MAX_PLANES]; + bool dsc_enable[DML2_MAX_PLANES]; }; struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals { diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c index 2504d9c2ec34..7a220c0141c2 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c @@ -82,7 +82,7 @@ bool dml2_core_utils_is_420(enum dml2_source_format_class source_format) val = 0; break; default: - DML2_ASSERT(0); + DML_ASSERT(0); break; } return val; @@ -145,7 +145,7 @@ bool dml2_core_utils_is_422_planar(enum dml2_source_format_class source_format) val = 0; break; default: - DML2_ASSERT(0); + DML_ASSERT(0); break; } return val; @@ -208,7 +208,7 @@ bool dml2_core_utils_is_422_packed(enum dml2_source_format_class source_format) val = 1; break; default: - DML2_ASSERT(0); + DML_ASSERT(0); break; } return val; @@ -216,104 +216,104 @@ bool dml2_core_utils_is_422_packed(enum dml2_source_format_class source_format) void dml2_core_utils_print_mode_support_info(const struct dml2_core_internal_mode_support_info *support, bool fail_only) { - dml2_printf("DML: ===================================== \n"); - dml2_printf("DML: DML_MODE_SUPPORT_INFO_ST\n"); + DML_LOG_VERBOSE("DML: ===================================== \n"); + DML_LOG_VERBOSE("DML: DML_MODE_SUPPORT_INFO_ST\n"); if (!fail_only || support->ScaleRatioAndTapsSupport == 0) - dml2_printf("DML: support: ScaleRatioAndTapsSupport = %d\n", support->ScaleRatioAndTapsSupport); + DML_LOG_VERBOSE("DML: support: ScaleRatioAndTapsSupport = %d\n", support->ScaleRatioAndTapsSupport); if (!fail_only || support->SourceFormatPixelAndScanSupport == 0) - dml2_printf("DML: support: SourceFormatPixelAndScanSupport = %d\n", support->SourceFormatPixelAndScanSupport); + DML_LOG_VERBOSE("DML: support: SourceFormatPixelAndScanSupport = %d\n", support->SourceFormatPixelAndScanSupport); if (!fail_only || support->ViewportSizeSupport == 0) - dml2_printf("DML: support: ViewportSizeSupport = %d\n", support->ViewportSizeSupport); + DML_LOG_VERBOSE("DML: support: ViewportSizeSupport = %d\n", support->ViewportSizeSupport); if (!fail_only || support->LinkRateDoesNotMatchDPVersion == 1) - dml2_printf("DML: support: LinkRateDoesNotMatchDPVersion = %d\n", support->LinkRateDoesNotMatchDPVersion); + DML_LOG_VERBOSE("DML: support: LinkRateDoesNotMatchDPVersion = %d\n", support->LinkRateDoesNotMatchDPVersion); if (!fail_only || support->LinkRateForMultistreamNotIndicated == 1) - dml2_printf("DML: support: LinkRateForMultistreamNotIndicated = %d\n", support->LinkRateForMultistreamNotIndicated); + DML_LOG_VERBOSE("DML: support: LinkRateForMultistreamNotIndicated = %d\n", support->LinkRateForMultistreamNotIndicated); if (!fail_only || support->BPPForMultistreamNotIndicated == 1) - dml2_printf("DML: support: BPPForMultistreamNotIndicated = %d\n", support->BPPForMultistreamNotIndicated); + DML_LOG_VERBOSE("DML: support: BPPForMultistreamNotIndicated = %d\n", support->BPPForMultistreamNotIndicated); if (!fail_only || support->MultistreamWithHDMIOreDP == 1) - dml2_printf("DML: support: MultistreamWithHDMIOreDP = %d\n", support->MultistreamWithHDMIOreDP); + DML_LOG_VERBOSE("DML: support: MultistreamWithHDMIOreDP = %d\n", support->MultistreamWithHDMIOreDP); if (!fail_only || support->ExceededMultistreamSlots == 1) - dml2_printf("DML: support: ExceededMultistreamSlots = %d\n", support->ExceededMultistreamSlots); + DML_LOG_VERBOSE("DML: support: ExceededMultistreamSlots = %d\n", support->ExceededMultistreamSlots); if (!fail_only || support->MSOOrODMSplitWithNonDPLink == 1) - dml2_printf("DML: support: MSOOrODMSplitWithNonDPLink = %d\n", support->MSOOrODMSplitWithNonDPLink); + DML_LOG_VERBOSE("DML: support: MSOOrODMSplitWithNonDPLink = %d\n", support->MSOOrODMSplitWithNonDPLink); if (!fail_only || support->NotEnoughLanesForMSO == 1) - dml2_printf("DML: support: NotEnoughLanesForMSO = %d\n", support->NotEnoughLanesForMSO); + DML_LOG_VERBOSE("DML: support: NotEnoughLanesForMSO = %d\n", support->NotEnoughLanesForMSO); if (!fail_only || support->P2IWith420 == 1) - dml2_printf("DML: support: P2IWith420 = %d\n", support->P2IWith420); + DML_LOG_VERBOSE("DML: support: P2IWith420 = %d\n", support->P2IWith420); if (!fail_only || support->DSC422NativeNotSupported == 1) - dml2_printf("DML: support: DSC422NativeNotSupported = %d\n", support->DSC422NativeNotSupported); + DML_LOG_VERBOSE("DML: support: DSC422NativeNotSupported = %d\n", support->DSC422NativeNotSupported); if (!fail_only || support->DSCSlicesODMModeSupported == 0) - dml2_printf("DML: support: DSCSlicesODMModeSupported = %d\n", support->DSCSlicesODMModeSupported); + DML_LOG_VERBOSE("DML: support: DSCSlicesODMModeSupported = %d\n", support->DSCSlicesODMModeSupported); if (!fail_only || support->NotEnoughDSCUnits == 1) - dml2_printf("DML: support: NotEnoughDSCUnits = %d\n", support->NotEnoughDSCUnits); + DML_LOG_VERBOSE("DML: support: NotEnoughDSCUnits = %d\n", support->NotEnoughDSCUnits); if (!fail_only || support->NotEnoughDSCSlices == 1) - dml2_printf("DML: support: NotEnoughDSCSlices = %d\n", support->NotEnoughDSCSlices); + DML_LOG_VERBOSE("DML: support: NotEnoughDSCSlices = %d\n", support->NotEnoughDSCSlices); if (!fail_only || support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == 1) - dml2_printf("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = %d\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe); + DML_LOG_VERBOSE("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = %d\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe); if (!fail_only || support->InvalidCombinationOfMALLUseForPStateAndStaticScreen == 1) - dml2_printf("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = %d\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen); + DML_LOG_VERBOSE("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = %d\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen); if (!fail_only || support->DSCCLKRequiredMoreThanSupported == 1) - dml2_printf("DML: support: DSCCLKRequiredMoreThanSupported = %d\n", support->DSCCLKRequiredMoreThanSupported); + DML_LOG_VERBOSE("DML: support: DSCCLKRequiredMoreThanSupported = %d\n", support->DSCCLKRequiredMoreThanSupported); if (!fail_only || support->PixelsPerLinePerDSCUnitSupport == 0) - dml2_printf("DML: support: PixelsPerLinePerDSCUnitSupport = %d\n", support->PixelsPerLinePerDSCUnitSupport); + DML_LOG_VERBOSE("DML: support: PixelsPerLinePerDSCUnitSupport = %d\n", support->PixelsPerLinePerDSCUnitSupport); if (!fail_only || support->DTBCLKRequiredMoreThanSupported == 1) - dml2_printf("DML: support: DTBCLKRequiredMoreThanSupported = %d\n", support->DTBCLKRequiredMoreThanSupported); + DML_LOG_VERBOSE("DML: support: DTBCLKRequiredMoreThanSupported = %d\n", support->DTBCLKRequiredMoreThanSupported); if (!fail_only || support->InvalidCombinationOfMALLUseForPState == 1) - dml2_printf("DML: support: InvalidCombinationOfMALLUseForPState = %d\n", support->InvalidCombinationOfMALLUseForPState); + DML_LOG_VERBOSE("DML: support: InvalidCombinationOfMALLUseForPState = %d\n", support->InvalidCombinationOfMALLUseForPState); if (!fail_only || support->ROBSupport == 0) - dml2_printf("DML: support: ROBSupport = %d\n", support->ROBSupport); + DML_LOG_VERBOSE("DML: support: ROBSupport = %d\n", support->ROBSupport); if (!fail_only || support->OutstandingRequestsSupport == 0) - dml2_printf("DML: support: OutstandingRequestsSupport = %d\n", support->OutstandingRequestsSupport); + DML_LOG_VERBOSE("DML: support: OutstandingRequestsSupport = %d\n", support->OutstandingRequestsSupport); if (!fail_only || support->OutstandingRequestsUrgencyAvoidance == 0) - dml2_printf("DML: support: OutstandingRequestsUrgencyAvoidance = %d\n", support->OutstandingRequestsUrgencyAvoidance); + DML_LOG_VERBOSE("DML: support: OutstandingRequestsUrgencyAvoidance = %d\n", support->OutstandingRequestsUrgencyAvoidance); if (!fail_only || support->DISPCLK_DPPCLK_Support == 0) - dml2_printf("DML: support: DISPCLK_DPPCLK_Support = %d\n", support->DISPCLK_DPPCLK_Support); + DML_LOG_VERBOSE("DML: support: DISPCLK_DPPCLK_Support = %d\n", support->DISPCLK_DPPCLK_Support); if (!fail_only || support->TotalAvailablePipesSupport == 0) - dml2_printf("DML: support: TotalAvailablePipesSupport = %d\n", support->TotalAvailablePipesSupport); + DML_LOG_VERBOSE("DML: support: TotalAvailablePipesSupport = %d\n", support->TotalAvailablePipesSupport); if (!fail_only || support->NumberOfOTGSupport == 0) - dml2_printf("DML: support: NumberOfOTGSupport = %d\n", support->NumberOfOTGSupport); + DML_LOG_VERBOSE("DML: support: NumberOfOTGSupport = %d\n", support->NumberOfOTGSupport); if (!fail_only || support->NumberOfHDMIFRLSupport == 0) - dml2_printf("DML: support: NumberOfHDMIFRLSupport = %d\n", support->NumberOfHDMIFRLSupport); + DML_LOG_VERBOSE("DML: support: NumberOfHDMIFRLSupport = %d\n", support->NumberOfHDMIFRLSupport); if (!fail_only || support->NumberOfDP2p0Support == 0) - dml2_printf("DML: support: NumberOfDP2p0Support = %d\n", support->NumberOfDP2p0Support); + DML_LOG_VERBOSE("DML: support: NumberOfDP2p0Support = %d\n", support->NumberOfDP2p0Support); if (!fail_only || support->EnoughWritebackUnits == 0) - dml2_printf("DML: support: EnoughWritebackUnits = %d\n", support->EnoughWritebackUnits); + DML_LOG_VERBOSE("DML: support: EnoughWritebackUnits = %d\n", support->EnoughWritebackUnits); if (!fail_only || support->WritebackScaleRatioAndTapsSupport == 0) - dml2_printf("DML: support: WritebackScaleRatioAndTapsSupport = %d\n", support->WritebackScaleRatioAndTapsSupport); + DML_LOG_VERBOSE("DML: support: WritebackScaleRatioAndTapsSupport = %d\n", support->WritebackScaleRatioAndTapsSupport); if (!fail_only || support->WritebackLatencySupport == 0) - dml2_printf("DML: support: WritebackLatencySupport = %d\n", support->WritebackLatencySupport); + DML_LOG_VERBOSE("DML: support: WritebackLatencySupport = %d\n", support->WritebackLatencySupport); if (!fail_only || support->CursorSupport == 0) - dml2_printf("DML: support: CursorSupport = %d\n", support->CursorSupport); + DML_LOG_VERBOSE("DML: support: CursorSupport = %d\n", support->CursorSupport); if (!fail_only || support->PitchSupport == 0) - dml2_printf("DML: support: PitchSupport = %d\n", support->PitchSupport); + DML_LOG_VERBOSE("DML: support: PitchSupport = %d\n", support->PitchSupport); if (!fail_only || support->ViewportExceedsSurface == 1) - dml2_printf("DML: support: ViewportExceedsSurface = %d\n", support->ViewportExceedsSurface); + DML_LOG_VERBOSE("DML: support: ViewportExceedsSurface = %d\n", support->ViewportExceedsSurface); if (!fail_only || support->PrefetchSupported == 0) - dml2_printf("DML: support: PrefetchSupported = %d\n", support->PrefetchSupported); + DML_LOG_VERBOSE("DML: support: PrefetchSupported = %d\n", support->PrefetchSupported); if (!fail_only || support->EnoughUrgentLatencyHidingSupport == 0) - dml2_printf("DML: support: EnoughUrgentLatencyHidingSupport = %d\n", support->EnoughUrgentLatencyHidingSupport); + DML_LOG_VERBOSE("DML: support: EnoughUrgentLatencyHidingSupport = %d\n", support->EnoughUrgentLatencyHidingSupport); if (!fail_only || support->AvgBandwidthSupport == 0) - dml2_printf("DML: support: AvgBandwidthSupport = %d\n", support->AvgBandwidthSupport); + DML_LOG_VERBOSE("DML: support: AvgBandwidthSupport = %d\n", support->AvgBandwidthSupport); if (!fail_only || support->DynamicMetadataSupported == 0) - dml2_printf("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported); + DML_LOG_VERBOSE("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported); if (!fail_only || support->VRatioInPrefetchSupported == 0) - dml2_printf("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported); + DML_LOG_VERBOSE("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported); if (!fail_only || support->PTEBufferSizeNotExceeded == 0) - dml2_printf("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded); + DML_LOG_VERBOSE("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded); if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 0) - dml2_printf("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded); + DML_LOG_VERBOSE("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded); if (!fail_only || support->ExceededMALLSize == 1) - dml2_printf("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize); + DML_LOG_VERBOSE("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize); if (!fail_only || support->g6_temp_read_support == 0) - dml2_printf("DML: support: g6_temp_read_support = %d\n", support->g6_temp_read_support); + DML_LOG_VERBOSE("DML: support: g6_temp_read_support = %d\n", support->g6_temp_read_support); if (!fail_only || support->ImmediateFlipSupport == 0) - dml2_printf("DML: support: ImmediateFlipSupport = %d\n", support->ImmediateFlipSupport); + DML_LOG_VERBOSE("DML: support: ImmediateFlipSupport = %d\n", support->ImmediateFlipSupport); if (!fail_only || support->LinkCapacitySupport == 0) - dml2_printf("DML: support: LinkCapacitySupport = %d\n", support->LinkCapacitySupport); + DML_LOG_VERBOSE("DML: support: LinkCapacitySupport = %d\n", support->LinkCapacitySupport); if (!fail_only || support->ModeSupport == 0) - dml2_printf("DML: support: ModeSupport = %d\n", support->ModeSupport); - dml2_printf("DML: ===================================== \n"); + DML_LOG_VERBOSE("DML: support: ModeSupport = %d\n", support->ModeSupport); + DML_LOG_VERBOSE("DML: ===================================== \n"); } const char *dml2_core_utils_internal_soc_state_type_str(enum dml2_core_internal_soc_state_type dml2_core_internal_soc_state_type) @@ -358,9 +358,9 @@ void dml2_core_utils_get_stream_output_bpp(double *out_bpp, const struct dml2_di out_bpp[k] = 0; } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: k=%d bpc=%f\n", __func__, k, bpc); - dml2_printf("DML::%s: k=%d dsc.enable=%d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable); - dml2_printf("DML::%s: k=%d out_bpp=%f\n", __func__, k, out_bpp[k]); + DML_LOG_VERBOSE("DML::%s: k=%d bpc=%f\n", __func__, k, bpc); + DML_LOG_VERBOSE("DML::%s: k=%d dsc.enable=%d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable); + DML_LOG_VERBOSE("DML::%s: k=%d out_bpp=%f\n", __func__, k, out_bpp[k]); #endif } } @@ -391,7 +391,7 @@ unsigned int dml2_core_util_get_num_active_pipes(int unsigned num_planes, const } #ifdef __DML_VBA_DEBUG__ - dml2_printf("DML::%s: num_active_pipes = %d\n", __func__, num_active_pipes); + DML_LOG_VERBOSE("DML::%s: num_active_pipes = %d\n", __func__, num_active_pipes); #endif return num_active_pipes; } @@ -452,7 +452,7 @@ unsigned int dml2_core_utils_get_tile_block_size_bytes(enum dml2_swizzle_mode sw else if (sw_mode == dml2_gfx11_sw_256kb_r_x) return 262144; else { - DML2_ASSERT(0); + DML_ASSERT(0); return 256; }; } @@ -498,8 +498,8 @@ int unsigned dml2_core_utils_get_gfx_version(enum dml2_swizzle_mode sw_mode) sw_mode == dml2_gfx11_sw_256kb_r_x) version = 11; else { - dml2_printf("ERROR: Invalid sw_mode setting! val=%u\n", sw_mode); - DML2_ASSERT(0); + DML_LOG_VERBOSE("ERROR: Invalid sw_mode setting! val=%u\n", sw_mode); + DML_ASSERT(0); } return version; @@ -511,7 +511,7 @@ unsigned int dml2_core_utils_get_qos_param_index(unsigned long uclk_freq_khz, co unsigned int index = 0; for (i = 0; i < DML_MAX_CLK_TABLE_SIZE; i++) { - dml2_printf("DML::%s: per_uclk_dpm_params[%d].minimum_uclk_khz = %d\n", __func__, i, per_uclk_dpm_params[i].minimum_uclk_khz); + DML_LOG_VERBOSE("DML::%s: per_uclk_dpm_params[%d].minimum_uclk_khz = %ld\n", __func__, i, per_uclk_dpm_params[i].minimum_uclk_khz); if (i == 0) index = 0; @@ -524,8 +524,8 @@ unsigned int dml2_core_utils_get_qos_param_index(unsigned long uclk_freq_khz, co } } #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: uclk_freq_khz = %d\n", __func__, uclk_freq_khz); - dml2_printf("DML::%s: index = %d\n", __func__, index); + DML_LOG_VERBOSE("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz); + DML_LOG_VERBOSE("DML::%s: index = %d\n", __func__, index); #endif return index; } @@ -533,32 +533,32 @@ unsigned int dml2_core_utils_get_qos_param_index(unsigned long uclk_freq_khz, co unsigned int dml2_core_utils_get_active_min_uclk_dpm_index(unsigned long uclk_freq_khz, const struct dml2_soc_state_table *clk_table) { unsigned int i; - bool clk_entry_found = 0; + bool clk_entry_found = false; for (i = 0; i < clk_table->uclk.num_clk_values; i++) { - dml2_printf("DML::%s: clk_table.uclk.clk_values_khz[%d] = %d\n", __func__, i, clk_table->uclk.clk_values_khz[i]); + DML_LOG_VERBOSE("DML::%s: clk_table.uclk.clk_values_khz[%d] = %ld\n", __func__, i, clk_table->uclk.clk_values_khz[i]); if (uclk_freq_khz == clk_table->uclk.clk_values_khz[i]) { - clk_entry_found = 1; + clk_entry_found = true; break; } } if (!clk_entry_found) - DML2_ASSERT(clk_entry_found); + DML_ASSERT(clk_entry_found); #if defined(__DML_VBA_DEBUG__) - dml2_printf("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz); - dml2_printf("DML::%s: index = %d\n", __func__, i); + DML_LOG_VERBOSE("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz); + DML_LOG_VERBOSE("DML::%s: index = %d\n", __func__, i); #endif return i; } bool dml2_core_utils_is_dual_plane(enum dml2_source_format_class source_format) { - bool ret_val = 0; + bool ret_val = false; if (dml2_core_utils_is_420(source_format) || dml2_core_utils_is_422_planar(source_format) || (source_format == dml2_rgbe_alpha)) - ret_val = 1; + ret_val = true; return ret_val; } diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c index 15507926f3a4..f486b090bbfc 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c @@ -754,6 +754,8 @@ bool dpmm_dcn4_map_watermarks(struct dml2_dpmm_map_watermarks_params_in_out *in_ dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].fclk_pstate = (int unsigned)(mode_lib->mp.Watermark.FCLKChangeWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].sr_enter = (int unsigned)(mode_lib->mp.Watermark.StutterEnterPlusExitWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].sr_exit = (int unsigned)(mode_lib->mp.Watermark.StutterExitWatermark * refclk_freq_in_mhz); + dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].sr_enter_z8 = (int unsigned)(mode_lib->mp.Watermark.Z8StutterEnterPlusExitWatermark * refclk_freq_in_mhz); + dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].sr_exit_z8 = (int unsigned)(mode_lib->mp.Watermark.Z8StutterExitWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].uclk_pstate = (int unsigned)(mode_lib->mp.Watermark.DRAMClockChangeWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].urgent = (int unsigned)(mode_lib->mp.Watermark.UrgentWatermark * refclk_freq_in_mhz); @@ -768,6 +770,8 @@ bool dpmm_dcn4_map_watermarks(struct dml2_dpmm_map_watermarks_params_in_out *in_ dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].fclk_pstate = (int unsigned)(mode_lib->mp.Watermark.FCLKChangeWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].sr_enter = (int unsigned)(mode_lib->mp.Watermark.StutterEnterPlusExitWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].sr_exit = (int unsigned)(mode_lib->mp.Watermark.StutterExitWatermark * refclk_freq_in_mhz); + dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].sr_enter_z8 = (int unsigned)(mode_lib->mp.Watermark.Z8StutterEnterPlusExitWatermark * refclk_freq_in_mhz); + dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].sr_exit_z8 = (int unsigned)(mode_lib->mp.Watermark.Z8StutterExitWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].uclk_pstate = (int unsigned)(mode_lib->mp.Watermark.DRAMClockChangeWatermark * refclk_freq_in_mhz); dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].urgent = (int unsigned)(mode_lib->mp.Watermark.UrgentWatermark * refclk_freq_in_mhz); diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_mcg/dml2_mcg_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_mcg/dml2_mcg_dcn4.c index f4b1a7d02d42..a265f254152c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_mcg/dml2_mcg_dcn4.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_mcg/dml2_mcg_dcn4.c @@ -182,6 +182,10 @@ static bool build_min_clock_table(const struct dml2_soc_bb *soc_bb, struct dml2_ min_table->max_clocks_khz.dtbclk = soc_bb->clk_table.dtbclk.clk_values_khz[soc_bb->clk_table.dtbclk.num_clk_values - 1]; min_table->max_clocks_khz.phyclk = soc_bb->clk_table.phyclk.clk_values_khz[soc_bb->clk_table.phyclk.num_clk_values - 1]; + min_table->max_ss_clocks_khz.dispclk = (unsigned int)((double)min_table->max_clocks_khz.dispclk / (1.0 + soc_bb->dcn_downspread_percent / 100.0)); + min_table->max_ss_clocks_khz.dppclk = (unsigned int)((double)min_table->max_clocks_khz.dppclk / (1.0 + soc_bb->dcn_downspread_percent / 100.0)); + min_table->max_ss_clocks_khz.dtbclk = (unsigned int)((double)min_table->max_clocks_khz.dtbclk / (1.0 + soc_bb->dcn_downspread_percent / 100.0)); + min_table->max_clocks_khz.dcfclk = soc_bb->clk_table.dcfclk.clk_values_khz[soc_bb->clk_table.dcfclk.num_clk_values - 1]; min_table->max_clocks_khz.fclk = soc_bb->clk_table.fclk.clk_values_khz[soc_bb->clk_table.fclk.num_clk_values - 1]; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c index f50662b83296..d88b3e0082dd 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c @@ -659,7 +659,7 @@ bool pmo_dcn4_fams2_initialize(struct dml2_pmo_initialize_in_out *in_out) for (i = 1; i <= PMO_DCN4_MAX_DISPLAYS; i++) { switch (i) { case 1: - DML2_ASSERT(base_strategy_list_1_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); + DML_ASSERT(base_strategy_list_1_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); /* populate list */ pmo_dcn4_fams2_expand_base_pstate_strategies( @@ -670,7 +670,7 @@ bool pmo_dcn4_fams2_initialize(struct dml2_pmo_initialize_in_out *in_out) &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]); break; case 2: - DML2_ASSERT(base_strategy_list_2_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); + DML_ASSERT(base_strategy_list_2_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); /* populate list */ pmo_dcn4_fams2_expand_base_pstate_strategies( @@ -681,7 +681,7 @@ bool pmo_dcn4_fams2_initialize(struct dml2_pmo_initialize_in_out *in_out) &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]); break; case 3: - DML2_ASSERT(base_strategy_list_3_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); + DML_ASSERT(base_strategy_list_3_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); /* populate list */ pmo_dcn4_fams2_expand_base_pstate_strategies( @@ -692,7 +692,7 @@ bool pmo_dcn4_fams2_initialize(struct dml2_pmo_initialize_in_out *in_out) &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]); break; case 4: - DML2_ASSERT(base_strategy_list_4_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); + DML_ASSERT(base_strategy_list_4_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES); /* populate list */ pmo_dcn4_fams2_expand_base_pstate_strategies( diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c index dc2ce5e77f57..4a7c4c62111e 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c @@ -761,7 +761,7 @@ bool dml2_top_mcache_calc_mcache_count_and_offsets(struct top_mcache_calc_mcache total_mcaches_required--; } } - dml2_printf("DML_CORE_DCN3::%s: plane_%d, total_mcaches_required=%d\n", __func__, i, total_mcaches_required); + DML_LOG_VERBOSE("DML_CORE_DCN3::%s: plane_%d, total_mcaches_required=%d\n", __func__, i, total_mcaches_required); if (total_mcaches_required > dml->soc_bbox.num_dcc_mcaches) { result = false; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c deleted file mode 100644 index c506667897c4..000000000000 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c +++ /dev/null @@ -1,31 +0,0 @@ -// SPDX-License-Identifier: MIT -// -// Copyright 2024 Advanced Micro Devices, Inc. - -#include "dml2_debug.h" - -int dml2_log_internal(const char *format, ...) -{ - return 0; -} - -int dml2_printf(const char *format, ...) -{ -#ifdef _DEBUG -#ifdef _DEBUG_PRINTS - int result; - va_list args; - va_start(args, format); - - result = vprintf(format, args); - - va_end(args); - - return result; -#else - return 0; -#endif -#else - return 0; -#endif -} diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h index bfe6f236d2e4..b226225103c3 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h @@ -5,55 +5,62 @@ #ifndef __DML2_DEBUG_H__ #define __DML2_DEBUG_H__ -#ifndef DML2_ASSERT -#define DML2_ASSERT(condition) ((void)0) -#endif +#include "os_types.h" +#define DML_ASSERT(condition) ASSERT(condition) +#define DML_LOG_LEVEL_DEFAULT DML_LOG_LEVEL_WARN +#define DML_LOG_INTERNAL(fmt, ...) dm_output_to_console(fmt, ## __VA_ARGS__) -/* - * DML_LOG_FATAL - fatal errors for unrecoverable DML states until a restart. - * DML_LOG_ERROR - unexpected but recoverable failures inside DML - * DML_LOG_WARN - unexpected inputs or events to DML - * DML_LOG_INFO - high level tracing of DML interfaces - * DML_LOG_DEBUG - detailed tracing of DML internal components - * DML_LOG_VERBOSE - detailed tracing of DML calculation procedure - */ -#if !defined(DML_LOG_LEVEL) -#if defined(_DEBUG) && defined(_DEBUG_PRINTS) -/* for backward compatibility with old macros */ -#define DML_LOG_LEVEL 5 -#else -#define DML_LOG_LEVEL 0 -#endif -#endif +/* ASSERT with message output */ +#define DML_ASSERT_MSG(condition, fmt, ...) \ + do { \ + if (!(condition)) { \ + DML_LOG_ERROR("DML ASSERT hit in %s line %d\n", __func__, __LINE__); \ + DML_LOG_ERROR(fmt, ## __VA_ARGS__); \ + DML_ASSERT(condition); \ + } \ + } while (0) + +/* fatal errors for unrecoverable DML states until a full reset */ +#define DML_LOG_LEVEL_FATAL 0 +/* unexpected but recoverable failures inside DML */ +#define DML_LOG_LEVEL_ERROR 1 +/* unexpected inputs or events to DML */ +#define DML_LOG_LEVEL_WARN 2 +/* high level tracing of DML interfaces */ +#define DML_LOG_LEVEL_INFO 3 +/* detailed tracing of DML internal components */ +#define DML_LOG_LEVEL_DEBUG 4 +/* detailed tracing of DML calculation procedure */ +#define DML_LOG_LEVEL_VERBOSE 5 -#define DML_LOG_FATAL(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__) -#if DML_LOG_LEVEL >= 1 -#define DML_LOG_ERROR(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__) +#ifndef DML_LOG_LEVEL +#define DML_LOG_LEVEL DML_LOG_LEVEL_DEFAULT +#endif /* #ifndef DML_LOG_LEVEL */ + +#define DML_LOG_FATAL(fmt, ...) DML_LOG_INTERNAL("[DML FATAL] " fmt, ## __VA_ARGS__) +#if DML_LOG_LEVEL >= DML_LOG_LEVEL_ERROR +#define DML_LOG_ERROR(fmt, ...) DML_LOG_INTERNAL("[DML ERROR] "fmt, ## __VA_ARGS__) #else #define DML_LOG_ERROR(fmt, ...) ((void)0) #endif -#if DML_LOG_LEVEL >= 2 -#define DML_LOG_WARN(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__) +#if DML_LOG_LEVEL >= DML_LOG_LEVEL_WARN +#define DML_LOG_WARN(fmt, ...) DML_LOG_INTERNAL("[DML WARN] "fmt, ## __VA_ARGS__) #else #define DML_LOG_WARN(fmt, ...) ((void)0) #endif -#if DML_LOG_LEVEL >= 3 -#define DML_LOG_INFO(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__) +#if DML_LOG_LEVEL >= DML_LOG_LEVEL_INFO +#define DML_LOG_INFO(fmt, ...) DML_LOG_INTERNAL("[DML INFO] "fmt, ## __VA_ARGS__) #else #define DML_LOG_INFO(fmt, ...) ((void)0) #endif -#if DML_LOG_LEVEL >= 4 -#define DML_LOG_DEBUG(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__) +#if DML_LOG_LEVEL >= DML_LOG_LEVEL_DEBUG +#define DML_LOG_DEBUG(fmt, ...) DML_LOG_INTERNAL("[DML DEBUG] "fmt, ## __VA_ARGS__) #else #define DML_LOG_DEBUG(fmt, ...) ((void)0) #endif -#if DML_LOG_LEVEL >= 5 -#define DML_LOG_VERBOSE(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__) +#if DML_LOG_LEVEL >= DML_LOG_LEVEL_VERBOSE +#define DML_LOG_VERBOSE(fmt, ...) DML_LOG_INTERNAL("[DML VERBOSE] "fmt, ## __VA_ARGS__) #else #define DML_LOG_VERBOSE(fmt, ...) ((void)0) #endif - -int dml2_log_internal(const char *format, ...); -int dml2_printf(const char *format, ...); - -#endif +#endif /* __DML2_DEBUG_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h index d8d01dceacdd..00688b9f1df4 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h @@ -38,6 +38,12 @@ struct dml2_mcg_min_clock_table { } max_clocks_khz; struct { + unsigned int dispclk; + unsigned int dppclk; + unsigned int dtbclk; + } max_ss_clocks_khz; + + struct { unsigned int dprefclk; unsigned int xtalclk; unsigned int pcierefclk; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c index a966abd40788..5f1b49a50049 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c @@ -1082,22 +1082,22 @@ bool dml2_map_dc_pipes(struct dml2_context *ctx, struct dc_state *state, const s if (stream_disp_cfg_index >= disp_cfg_index_max) continue; - if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_bypass) { - scratch.odm_info.odm_factor = 1; - } else if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_combine_2to1) { - scratch.odm_info.odm_factor = 2; - } else if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_combine_4to1) { - scratch.odm_info.odm_factor = 4; - } else { - ASSERT(false); - scratch.odm_info.odm_factor = 1; - } - + if (ctx->architecture == dml2_architecture_20) { + if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_bypass) { + scratch.odm_info.odm_factor = 1; + } else if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_combine_2to1) { + scratch.odm_info.odm_factor = 2; + } else if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_combine_4to1) { + scratch.odm_info.odm_factor = 4; + } else { + ASSERT(false); + scratch.odm_info.odm_factor = 1; + } + } else if (ctx->architecture == dml2_architecture_21) { /* After DML2.1 update, ODM interpretation needs to change and is no longer same as for DML2.0. * This is not an issue with new resource management logic. This block ensure backcompat * with legacy pipe management with updated DML. * */ - if (ctx->architecture == dml2_architecture_21) { if (ODMMode[stream_disp_cfg_index] == 1) { scratch.odm_info.odm_factor = 1; } else if (ODMMode[stream_disp_cfg_index] == 2) { diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c index 2061d43b92e1..5de775fd8fce 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c @@ -896,7 +896,7 @@ static void populate_dummy_dml_surface_cfg(struct dml_surface_cfg_st *out, unsig out->SurfaceWidthC[location] = in->timing.h_addressable; out->SurfaceHeightC[location] = in->timing.v_addressable; out->PitchY[location] = ((out->SurfaceWidthY[location] + 127) / 128) * 128; - out->PitchC[location] = 0; + out->PitchC[location] = 1; out->DCCEnable[location] = false; out->DCCMetaPitchY[location] = 0; out->DCCMetaPitchC[location] = 0; @@ -973,7 +973,9 @@ static void populate_dml_surface_cfg_from_plane_state(enum dml_project_id dml2_p } } -static void get_scaler_data_for_plane(const struct dc_plane_state *in, struct dc_state *context, struct scaler_data *out) +static struct scaler_data *get_scaler_data_for_plane( + const struct dc_plane_state *in, + struct dc_state *context) { int i; struct pipe_ctx *temp_pipe = &context->res_ctx.temp_pipe; @@ -994,7 +996,7 @@ static void get_scaler_data_for_plane(const struct dc_plane_state *in, struct dc } ASSERT(i < MAX_PIPES); - memcpy(out, &temp_pipe->plane_res.scl_data, sizeof(*out)); + return &temp_pipe->plane_res.scl_data; } static void populate_dummy_dml_plane_cfg(struct dml_plane_cfg_st *out, unsigned int location, @@ -1057,11 +1059,7 @@ static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out const struct dc_plane_state *in, struct dc_state *context, const struct soc_bounding_box_st *soc) { - struct scaler_data *scaler_data = kzalloc(sizeof(*scaler_data), GFP_KERNEL); - if (!scaler_data) - return; - - get_scaler_data_for_plane(in, context, scaler_data); + struct scaler_data *scaler_data = get_scaler_data_for_plane(in, context); out->CursorBPP[location] = dml_cur_32bit; out->CursorWidth[location] = 256; @@ -1126,8 +1124,6 @@ static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out out->DynamicMetadataTransmittedBytes[location] = 0; out->NumberOfCursors[location] = 1; - - kfree(scaler_data); } static unsigned int map_stream_to_dml_display_cfg(const struct dml2_context *dml2, diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c index e89571874185..525b7d04bf84 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c @@ -663,7 +663,10 @@ static bool dml2_validate_and_build_resource(const struct dc *in_dc, struct dc_s dml2_copy_clocks_to_dc_state(&out_clks, context); dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.a, &dml2->v20.dml_core_ctx); dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.b, &dml2->v20.dml_core_ctx); - memcpy(&context->bw_ctx.bw.dcn.watermarks.c, &dml2->v20.g6_temp_read_watermark_set, sizeof(context->bw_ctx.bw.dcn.watermarks.c)); + if (context->streams[0]->sink->link->dc->caps.is_apu) + dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.c, &dml2->v20.dml_core_ctx); + else + memcpy(&context->bw_ctx.bw.dcn.watermarks.c, &dml2->v20.g6_temp_read_watermark_set, sizeof(context->bw_ctx.bw.dcn.watermarks.c)); dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.d, &dml2->v20.dml_core_ctx); dml2_extract_writeback_wm(context, &dml2->v20.dml_core_ctx); //copy for deciding zstate use diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h index 785226945699..5100f269368e 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h @@ -40,6 +40,7 @@ struct dc_sink; struct dc_stream_state; struct resource_context; struct display_stream_compressor; +struct dc_mcache_params; // Configuration of the MALL on the SoC struct dml2_soc_mall_info { @@ -107,6 +108,7 @@ struct dml2_dc_callbacks { unsigned int (*get_max_flickerless_instant_vtotal_increase)( struct dc_stream_state *stream, bool is_gaming); + bool (*allocate_mcache)(struct dc_state *context, const struct dc_mcache_params *mcache_params); }; struct dml2_dc_svp_callbacks { diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c index abf439e743f2..2d70586cef40 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c @@ -790,8 +790,7 @@ static bool dpp3_program_blnd_lut(struct dpp *dpp_base, if (params == NULL) { REG_SET(CM_BLNDGAM_CONTROL, 0, CM_BLNDGAM_MODE, 0); - if (dpp_base->ctx->dc->debug.enable_mem_low_power.bits.cm) - dpp3_power_on_blnd_lut(dpp_base, false); + dpp3_power_on_blnd_lut(dpp_base, false); return false; } @@ -1204,8 +1203,7 @@ static bool dpp3_program_shaper(struct dpp *dpp_base, if (params == NULL) { REG_SET(CM_SHAPER_CONTROL, 0, CM_SHAPER_LUT_MODE, 0); - if (dpp_base->ctx->dc->debug.enable_mem_low_power.bits.cm) - dpp3_power_on_shaper(dpp_base, false); + dpp3_power_on_shaper(dpp_base, false); return false; } @@ -1399,8 +1397,7 @@ static bool dpp3_program_3dlut(struct dpp *dpp_base, if (params == NULL) { dpp3_set_3dlut_mode(dpp_base, LUT_BYPASS, false, false); - if (dpp_base->ctx->dc->debug.enable_mem_low_power.bits.cm) - dpp3_power_on_hdr3dlut(dpp_base, false); + dpp3_power_on_hdr3dlut(dpp_base, false); return false; } diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn35/dcn35_dpp.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn35/dcn35_dpp.c index 62b7012cda43..f7a373a3d70a 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn35/dcn35_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn35/dcn35_dpp.c @@ -138,7 +138,7 @@ bool dpp35_construct( dpp->base.funcs = &dcn35_dpp_funcs; // w/a for cursor memory stuck in LS by programming DISPCLK_R_GATE_DISABLE, limit w/a to some ASIC revs - if (dpp->base.ctx->asic_id.hw_internal_rev <= 0x10) + if (dpp->base.ctx->asic_id.hw_internal_rev < 0x40) dpp->dispclk_r_gate_disable = true; return ret; } diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c index 1236e0f9a256..712aff7e17f7 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c @@ -120,10 +120,11 @@ void dpp401_set_cursor_attributes( enum dc_cursor_color_format color_format = cursor_attributes->color_format; int cur_rom_en = 0; - // DCN4 should always do Cursor degamma for Cursor Color modes if (color_format == CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA || color_format == CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA) { - cur_rom_en = 1; + if (cursor_attributes->attribute_flags.bits.ENABLE_CURSOR_DEGAMMA) { + cur_rom_en = 1; + } } REG_UPDATE_3(CURSOR0_CONTROL, diff --git a/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.c b/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.c index 75128fd34306..bd1b9aef6d5c 100644 --- a/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.c +++ b/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.c @@ -57,13 +57,6 @@ static const struct dsc_funcs dcn20_dsc_funcs = { #define DC_LOGGER \ dsc->ctx->logger -enum dsc_bits_per_comp { - DSC_BPC_8 = 8, - DSC_BPC_10 = 10, - DSC_BPC_12 = 12, - DSC_BPC_UNKNOWN -}; - /* API functions (external or via structure->function_pointer) */ void dsc2_construct(struct dcn20_dsc *dsc, diff --git a/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.h b/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.h index 1fb90b52b814..a9c04fc95bd1 100644 --- a/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.h +++ b/drivers/gpu/drm/amd/display/dc/dsc/dcn20/dcn20_dsc.h @@ -457,6 +457,12 @@ type DSCRM_DSC_DOUBLE_BUFFER_REG_UPDATE_PENDING; \ type DSCRM_DSC_FORWARD_EN_STATUS +enum dsc_bits_per_comp { + DSC_BPC_8 = 8, + DSC_BPC_10 = 10, + DSC_BPC_12 = 12, + DSC_BPC_UNKNOWN +}; struct dcn20_dsc_registers { uint32_t DSC_TOP_CONTROL; diff --git a/drivers/gpu/drm/amd/display/dc/dsc/dcn401/dcn401_dsc.c b/drivers/gpu/drm/amd/display/dc/dsc/dcn401/dcn401_dsc.c index 4893b793fec0..4222679fd4c9 100644 --- a/drivers/gpu/drm/amd/display/dc/dsc/dcn401/dcn401_dsc.c +++ b/drivers/gpu/drm/amd/display/dc/dsc/dcn401/dcn401_dsc.c @@ -45,12 +45,6 @@ static const struct dsc_funcs dcn401_dsc_funcs = { #define DC_LOGGER \ dsc->ctx->logger -enum dsc_bits_per_comp { - DSC_BPC_8 = 8, - DSC_BPC_10 = 10, - DSC_BPC_12 = 12, - DSC_BPC_UNKNOWN -}; /* API functions (external or via structure->function_pointer) */ diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c index b099989d9364..942d9f0b6df2 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c +++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c @@ -411,6 +411,20 @@ enum dc_irq_source dal_irq_get_rx_source( } } +enum dc_irq_source dal_irq_get_read_request( + const struct gpio *irq) +{ + enum gpio_id id = dal_gpio_get_id(irq); + + switch (id) { + case GPIO_ID_HPD: + return (enum dc_irq_source)(DC_IRQ_SOURCE_DCI2C_RR_DDC1 + + dal_gpio_get_enum(irq)); + default: + return DC_IRQ_SOURCE_INVALID; + } +} + enum gpio_result dal_irq_setup_hpd_filter( struct gpio *irq, struct gpio_hpd_config *config) diff --git a/drivers/gpu/drm/amd/display/dc/hubbub/dcn21/dcn21_hubbub.c b/drivers/gpu/drm/amd/display/dc/hubbub/dcn21/dcn21_hubbub.c index 2546224b326a..e4496ad203b2 100644 --- a/drivers/gpu/drm/amd/display/dc/hubbub/dcn21/dcn21_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/hubbub/dcn21/dcn21_hubbub.c @@ -132,9 +132,9 @@ int hubbub21_init_dchub(struct hubbub *hubbub, // Init VMID 0 based on PA config dcn20_vmid_setup(&hubbub1->vmid[0], &phys_config); } - - dcn21_dchvm_init(hubbub); - + if (!hubbub1->base.ctx->dc->config.skip_riommu_prefetch_wa) { + dcn21_dchvm_init(hubbub); + } return hubbub1->num_vmid; } diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c index 5ed195377a6c..baed31611477 100644 --- a/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c +++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c @@ -1032,7 +1032,7 @@ static struct hubp_funcs dcn401_hubp_funcs = { .hubp_program_3dlut_fl_tmz_protected = hubp401_program_3dlut_fl_tmz_protected, .hubp_program_3dlut_fl_crossbar = hubp401_program_3dlut_fl_crossbar, .hubp_get_3dlut_fl_done = hubp401_get_3dlut_fl_done, - .hubp_clear_tiling = hubp2_clear_tiling, + .hubp_clear_tiling = hubp401_clear_tiling, }; bool hubp401_construct( diff --git a/drivers/gpu/drm/amd/display/dc/hwss/Makefile b/drivers/gpu/drm/amd/display/dc/hwss/Makefile index 40ecebea1ba0..bee617ca0838 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/Makefile +++ b/drivers/gpu/drm/amd/display/dc/hwss/Makefile @@ -27,6 +27,24 @@ # DCE ############################################################################### +ifdef CONFIG_DRM_AMD_DC_SI +HWSS_DCE60 = dce60_hwseq.o + +AMD_DAL_HWSS_DCE60 = $(addprefix $(AMDDALPATH)/dc/hwss/dce60/,$(HWSS_DCE60)) + +AMD_DISPLAY_FILES += $(AMD_DAL_HWSS_DCE60) +endif + +############################################################################### + +HWSS_DCE80 = dce80_hwseq.o + +AMD_DAL_HWSS_DCE80 = $(addprefix $(AMDDALPATH)/dc/hwss/dce80/,$(HWSS_DCE80)) + +AMD_DISPLAY_FILES += $(AMD_DAL_HWSS_DCE80) + +############################################################################### + HWSS_DCE = dce_hwseq.o AMD_DAL_HWSS_DCE = $(addprefix $(AMDDALPATH)/dc/hwss/dce/,$(HWSS_DCE)) @@ -65,14 +83,6 @@ AMD_DAL_HWSS_DCE120 = $(addprefix $(AMDDALPATH)/dc/hwss/dce120/,$(HWSS_DCE120)) AMD_DISPLAY_FILES += $(AMD_DAL_HWSS_DCE120) -############################################################################### - -HWSS_DCE80 = dce80_hwseq.o - -AMD_DAL_HWSS_DCE80 = $(addprefix $(AMDDALPATH)/dc/hwss/dce80/,$(HWSS_DCE80)) - -AMD_DISPLAY_FILES += $(AMD_DAL_HWSS_DCE80) - ifdef CONFIG_DRM_AMD_DC_FP ############################################################################### # DCN diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c index 5656d10368ad..23bec5d25ed6 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c @@ -2763,12 +2763,12 @@ static void dce110_enable_per_frame_crtc_position_reset( } -static void init_pipes(struct dc *dc, struct dc_state *context) +static void dce110_init_pipes(struct dc *dc, struct dc_state *context) { // Do nothing } -static void init_hw(struct dc *dc) +static void dce110_init_hw(struct dc *dc) { int i; struct dc_bios *bp; @@ -3327,7 +3327,7 @@ void dce110_disable_link_output(struct dc_link *link, static const struct hw_sequencer_funcs dce110_funcs = { .program_gamut_remap = program_gamut_remap, .program_output_csc = program_output_csc, - .init_hw = init_hw, + .init_hw = dce110_init_hw, .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = dce110_apply_ctx_for_surface, .post_unlock_program_front_end = dce110_post_unlock_program_front_end, @@ -3371,7 +3371,7 @@ static const struct hw_sequencer_funcs dce110_funcs = { }; static const struct hwseq_private_funcs dce110_private_funcs = { - .init_pipes = init_pipes, + .init_pipes = dce110_init_pipes, .set_input_transfer_func = dce110_set_input_transfer_func, .set_output_transfer_func = dce110_set_output_transfer_func, .power_down = dce110_power_down, diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/hwss/dce60/dce60_hwseq.c index 44b56490e152..a08e9f9eec17 100644 --- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dce60/dce60_hwseq.c @@ -26,7 +26,7 @@ #include "dm_services.h" #include "dc.h" #include "core_types.h" -#include "dce60_hw_sequencer.h" +#include "dce60_hwseq.h" #include "dce/dce_hwseq.h" #include "dce110/dce110_hwseq.h" diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/hwss/dce60/dce60_hwseq.h index f3b2d8b60d5b..f3b2d8b60d5b 100644 --- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/dce60/dce60_hwseq.h diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c index 912f96323ed6..f9ee55998b6b 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c @@ -94,6 +94,128 @@ static void print_microsec(struct dc_context *dc_ctx, us_x10 % frac); } +/* + * Delay until we passed busy-until-point to which we can + * do necessary locking/programming on consecutive full updates + */ +void dcn10_wait_for_pipe_update_if_needed(struct dc *dc, struct pipe_ctx *pipe_ctx, bool is_surface_update_only) +{ + struct crtc_position position; + struct dc_stream_state *stream = pipe_ctx->stream; + unsigned int vpos, frame_count; + uint32_t vupdate_start, vupdate_end, vblank_start; + unsigned int lines_to_vupdate, us_to_vupdate; + unsigned int us_per_line, us_vupdate; + + if (!pipe_ctx->stream || + !pipe_ctx->stream_res.tg || + !pipe_ctx->stream_res.stream_enc) + return; + + if (pipe_ctx->prev_odm_pipe && + pipe_ctx->stream) + return; + + if (!pipe_ctx->wait_is_required) + return; + + struct timing_generator *tg = pipe_ctx->stream_res.tg; + + if (tg->funcs->is_tg_enabled && !tg->funcs->is_tg_enabled(tg)) + return; + + dc->hwss.calc_vupdate_position(dc, pipe_ctx, &vupdate_start, + &vupdate_end); + + dc->hwss.get_position(&pipe_ctx, 1, &position); + vpos = position.vertical_count; + + frame_count = tg->funcs->get_frame_count(tg); + + if (frame_count - pipe_ctx->wait_frame_count > 2) + return; + + vblank_start = pipe_ctx->pipe_dlg_param.vblank_start; + + if (vpos >= vupdate_start && vupdate_start >= vblank_start) + lines_to_vupdate = stream->timing.v_total - vpos + vupdate_start; + else + lines_to_vupdate = vupdate_start - vpos; + + us_per_line = + stream->timing.h_total * 10000u / stream->timing.pix_clk_100hz; + us_to_vupdate = lines_to_vupdate * us_per_line; + + if (vupdate_end < vupdate_start) + vupdate_end += stream->timing.v_total; + + if (lines_to_vupdate > stream->timing.v_total - vupdate_end + vupdate_start) + us_to_vupdate = 0; + + us_vupdate = (vupdate_end - vupdate_start + 1) * us_per_line; + + if (is_surface_update_only && us_to_vupdate + us_vupdate > 200) { + //surface updates come in at high irql + pipe_ctx->wait_is_required = true; + return; + } + + fsleep(us_to_vupdate + us_vupdate); + + //clear + pipe_ctx->next_vupdate = 0; + pipe_ctx->wait_frame_count = 0; + pipe_ctx->wait_is_required = false; +} + +/* + * On pipe unlock and programming, indicate pipe will be busy + * until some frame and line (vupdate), this is required for consecutive + * full updates, need to wait for updates + * to latch to try and program the next update + */ +void dcn10_set_wait_for_update_needed_for_pipe(struct dc *dc, struct pipe_ctx *pipe_ctx) +{ + uint32_t vupdate_start, vupdate_end; + struct crtc_position position; + unsigned int vpos, cur_frame; + + if (!pipe_ctx->stream || + !pipe_ctx->stream_res.tg || + !pipe_ctx->stream_res.stream_enc) + return; + + dc->hwss.get_position(&pipe_ctx, 1, &position); + vpos = position.vertical_count; + + dc->hwss.calc_vupdate_position(dc, pipe_ctx, &vupdate_start, + &vupdate_end); + + struct timing_generator *tg = pipe_ctx->stream_res.tg; + + struct optc *optc1 = DCN10TG_FROM_TG(tg); + + ASSERT(optc1->max_frame_count != 0); + + if (tg->funcs->is_tg_enabled && !tg->funcs->is_tg_enabled(tg)) + return; + + pipe_ctx->next_vupdate = vupdate_start; + + cur_frame = tg->funcs->get_frame_count(tg); + + if (vpos < vupdate_start) { + pipe_ctx->wait_frame_count = cur_frame; + } else { + if (cur_frame + 1 > optc1->max_frame_count) + pipe_ctx->wait_frame_count = cur_frame + 1 - optc1->max_frame_count; + else + pipe_ctx->wait_frame_count = cur_frame + 1; + } + + pipe_ctx->wait_is_required = true; +} + void dcn10_lock_all_pipes(struct dc *dc, struct dc_state *context, bool lock) @@ -2664,7 +2786,6 @@ void dcn10_update_visual_confirm_color(struct dc *dc, struct mpc *mpc = dc->res_pool->mpc; if (mpc->funcs->set_bg_color) { - memcpy(&pipe_ctx->plane_state->visual_confirm_color, &(pipe_ctx->visual_confirm_color), sizeof(struct tg_color)); mpc->funcs->set_bg_color(mpc, &(pipe_ctx->visual_confirm_color), mpcc_id); } } diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.h index 42ffd1e1299c..57d30ea225f2 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.h @@ -50,6 +50,13 @@ void dcn10_optimize_bandwidth( void dcn10_prepare_bandwidth( struct dc *dc, struct dc_state *context); +void dcn10_wait_for_pipe_update_if_needed( + struct dc *dc, + struct pipe_ctx *pipe_ctx, + bool is_surface_update_only); +void dcn10_set_wait_for_update_needed_for_pipe( + struct dc *dc, + struct pipe_ctx *pipe_ctx); void dcn10_pipe_control_lock( struct dc *dc, struct pipe_ctx *pipe, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c index 846c9c51f2d9..858288c3b1ac 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c @@ -2053,7 +2053,7 @@ void dcn20_program_front_end_for_ctx( for (i = 0; i < dc->res_pool->pipe_count; i++) { pipe = &context->res_ctx.pipe_ctx[i]; - if (!pipe->top_pipe && !pipe->prev_odm_pipe && pipe->plane_state) { + if (pipe->plane_state) { ASSERT(!pipe->plane_state->triplebuffer_flips); /*turn off triple buffer for full update*/ dc->hwss.program_triplebuffer( @@ -2482,7 +2482,7 @@ bool dcn20_update_bandwidth( struct dce_hwseq *hws = dc->hwseq; /* recalculate DML parameters */ - if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false)) + if (dc->res_pool->funcs->validate_bandwidth(dc, context, false) != DC_OK) return false; /* apply updated bandwidth parameters */ diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c index be26c925fdfa..e68f21fd5f0f 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c @@ -84,6 +84,20 @@ static void update_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable) struct dsc_config dsc_cfg; struct dsc_optc_config dsc_optc_cfg = {0}; enum optc_dsc_mode optc_dsc_mode; + struct dcn_dsc_state dsc_state = {0}; + + if (!dsc) { + DC_LOG_DSC("DSC is NULL for tg instance %d:", pipe_ctx->stream_res.tg->inst); + return; + } + + if (dsc->funcs->dsc_read_state) { + dsc->funcs->dsc_read_state(dsc, &dsc_state); + if (!dsc_state.dsc_fw_en) { + DC_LOG_DSC("DSC has been disabled for tg instance %d:", pipe_ctx->stream_res.tg->inst); + return; + } + } /* Enable DSC hw block */ dsc_cfg.pic_width = (stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt; diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c index cd0adf72b223..a0b05b9ef660 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c @@ -1181,6 +1181,7 @@ unsigned int dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsign struct dc_stream_state *stream = pipe_ctx->stream; unsigned int odm_combine_factor = 0; bool two_pix_per_container = false; + struct dce_hwseq *hws = stream->ctx->dc->hwseq; two_pix_per_container = pipe_ctx->stream_res.tg->funcs->is_two_pixels_per_container(&stream->timing); odm_combine_factor = get_odm_config(pipe_ctx, NULL); @@ -1201,7 +1202,8 @@ unsigned int dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsign } else { *k1_div = PIXEL_RATE_DIV_BY_1; *k2_div = PIXEL_RATE_DIV_BY_4; - if ((odm_combine_factor == 2) || dcn32_is_dp_dig_pixel_rate_div_policy(pipe_ctx)) + if ((odm_combine_factor == 2) || (hws->funcs.is_dp_dig_pixel_rate_div_policy && + hws->funcs.is_dp_dig_pixel_rate_div_policy(pipe_ctx))) *k2_div = PIXEL_RATE_DIV_BY_2; } } diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c index 922b8d71cf1a..c814d957305a 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c @@ -241,11 +241,6 @@ void dcn35_init_hw(struct dc *dc) dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub, !dc->res_pool->hubbub->ctx->dc->debug.disable_stutter); } - if (res_pool->dccg->funcs->dccg_root_gate_disable_control) { - for (i = 0; i < res_pool->pipe_count; i++) - res_pool->dccg->funcs->dccg_root_gate_disable_control(res_pool->dccg, i, 0); - } - for (i = 0; i < res_pool->audio_count; i++) { struct audio *audio = res_pool->audios[i]; @@ -901,12 +896,18 @@ void dcn35_init_pipes(struct dc *dc, struct dc_state *context) void dcn35_enable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx, struct dc_state *context) { + struct dpp *dpp = pipe_ctx->plane_res.dpp; + struct dccg *dccg = dc->res_pool->dccg; + + /* enable DCFCLK current DCHUB */ pipe_ctx->plane_res.hubp->funcs->hubp_clk_cntl(pipe_ctx->plane_res.hubp, true); /* initialize HUBP on power up */ pipe_ctx->plane_res.hubp->funcs->hubp_init(pipe_ctx->plane_res.hubp); - + /*make sure DPPCLK is on*/ + dccg->funcs->dccg_root_gate_disable_control(dccg, dpp->inst, true); + dpp->funcs->dpp_dppclk_control(dpp, false, true); /* make sure OPP_PIPE_CLOCK_EN = 1 */ pipe_ctx->stream_res.opp->funcs->opp_pipe_clock_control( pipe_ctx->stream_res.opp, @@ -923,6 +924,7 @@ void dcn35_enable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx, // Program system aperture settings pipe_ctx->plane_res.hubp->funcs->hubp_set_vm_system_aperture_settings(pipe_ctx->plane_res.hubp, &apt); } + //DC_LOG_DEBUG("%s: dpp_inst(%d) =\n", __func__, dpp->inst); if (!pipe_ctx->top_pipe && pipe_ctx->plane_state @@ -938,6 +940,8 @@ void dcn35_plane_atomic_disable(struct dc *dc, struct pipe_ctx *pipe_ctx) { struct hubp *hubp = pipe_ctx->plane_res.hubp; struct dpp *dpp = pipe_ctx->plane_res.dpp; + struct dccg *dccg = dc->res_pool->dccg; + dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, pipe_ctx); @@ -955,7 +959,8 @@ void dcn35_plane_atomic_disable(struct dc *dc, struct pipe_ctx *pipe_ctx) hubp->funcs->hubp_clk_cntl(hubp, false); dpp->funcs->dpp_dppclk_control(dpp, false, false); -/*to do, need to support both case*/ + dccg->funcs->dccg_root_gate_disable_control(dccg, dpp->inst, false); + hubp->power_gated = true; hubp->funcs->hubp_reset(hubp); @@ -967,6 +972,8 @@ void dcn35_plane_atomic_disable(struct dc *dc, struct pipe_ctx *pipe_ctx) pipe_ctx->top_pipe = NULL; pipe_ctx->bottom_pipe = NULL; pipe_ctx->plane_state = NULL; + //DC_LOG_DEBUG("%s: dpp_inst(%d)=\n", __func__, dpp->inst); + } void dcn35_disable_plane(struct dc *dc, struct dc_state *state, struct pipe_ctx *pipe_ctx) @@ -1543,7 +1550,7 @@ static bool should_avoid_empty_tu(struct pipe_ctx *pipe_ctx) struct dc_link_settings *link_settings = &pipe_ctx->link_config.dp_link_settings; const struct dc *dc = pipe_ctx->stream->link->dc; - if (pipe_ctx->stream->link->ep_type != DISPLAY_ENDPOINT_USB4_DPIA) + if (pipe_ctx->link_config.dp_tunnel_settings.should_enable_dp_tunneling == false) return false; // Not necessary for MST configurations diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c index 6a82a865209c..a3ccf805bd16 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c @@ -168,6 +168,8 @@ static const struct hwseq_private_funcs dcn35_private_funcs = { .dsc_pg_control = dcn35_dsc_pg_control, .dsc_pg_status = dcn32_dsc_pg_status, .enable_plane = dcn35_enable_plane, + .wait_for_pipe_update_if_needed = dcn10_wait_for_pipe_update_if_needed, + .set_wait_for_update_needed_for_pipe = dcn10_set_wait_for_update_needed_for_pipe, }; void dcn35_hw_sequencer_construct(struct dc *dc) diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c index 902a96940a01..58f2be2a326b 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c @@ -158,10 +158,12 @@ static const struct hwseq_private_funcs dcn351_private_funcs = { .set_mcm_luts = dcn32_set_mcm_luts, .setup_hpo_hw_control = dcn35_setup_hpo_hw_control, .calculate_dccg_k1_k2_values = dcn32_calculate_dccg_k1_k2_values, - .is_dp_dig_pixel_rate_div_policy = dcn32_is_dp_dig_pixel_rate_div_policy, + .is_dp_dig_pixel_rate_div_policy = dcn35_is_dp_dig_pixel_rate_div_policy, .dsc_pg_control = dcn35_dsc_pg_control, .dsc_pg_status = dcn32_dsc_pg_status, .enable_plane = dcn35_enable_plane, + .wait_for_pipe_update_if_needed = dcn10_wait_for_pipe_update_if_needed, + .set_wait_for_update_needed_for_pipe = dcn10_set_wait_for_update_needed_for_pipe, }; void dcn351_hw_sequencer_construct(struct dc *dc) diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c index 5489f3d431f6..c4177a9a662f 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c @@ -396,6 +396,249 @@ static void dcn401_get_mcm_lut_xable_from_pipe_ctx(struct dc *dc, struct pipe_ct } } +static void dcn401_set_mcm_location_post_blend(struct dc *dc, struct pipe_ctx *pipe_ctx, bool bPostBlend) +{ + struct mpc *mpc = dc->res_pool->mpc; + int mpcc_id = pipe_ctx->plane_res.hubp->inst; + + if (!pipe_ctx->plane_state) + return; + + mpc->funcs->set_movable_cm_location(mpc, MPCC_MOVABLE_CM_LOCATION_BEFORE, mpcc_id); + pipe_ctx->plane_state->mcm_location = (bPostBlend) ? + MPCC_MOVABLE_CM_LOCATION_AFTER : + MPCC_MOVABLE_CM_LOCATION_BEFORE; +} + +static void dc_get_lut_mode( + enum dc_cm2_gpu_mem_layout layout, + enum hubp_3dlut_fl_mode *mode, + enum hubp_3dlut_fl_addressing_mode *addr_mode) +{ + switch (layout) { + case DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_RGB: + *mode = hubp_3dlut_fl_mode_native_1; + *addr_mode = hubp_3dlut_fl_addressing_mode_sw_linear; + break; + case DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_BGR: + *mode = hubp_3dlut_fl_mode_native_2; + *addr_mode = hubp_3dlut_fl_addressing_mode_sw_linear; + break; + case DC_CM2_GPU_MEM_LAYOUT_1D_PACKED_LINEAR: + *mode = hubp_3dlut_fl_mode_transform; + *addr_mode = hubp_3dlut_fl_addressing_mode_simple_linear; + break; + default: + *mode = hubp_3dlut_fl_mode_disable; + *addr_mode = hubp_3dlut_fl_addressing_mode_sw_linear; + break; + } +} + +static void dc_get_lut_format( + enum dc_cm2_gpu_mem_format dc_format, + enum hubp_3dlut_fl_format *format) +{ + switch (dc_format) { + case DC_CM2_GPU_MEM_FORMAT_16161616_UNORM_12MSB: + *format = hubp_3dlut_fl_format_unorm_12msb_bitslice; + break; + case DC_CM2_GPU_MEM_FORMAT_16161616_UNORM_12LSB: + *format = hubp_3dlut_fl_format_unorm_12lsb_bitslice; + break; + case DC_CM2_GPU_MEM_FORMAT_16161616_FLOAT_FP1_5_10: + *format = hubp_3dlut_fl_format_float_fp1_5_10; + break; + } +} + +static void dc_get_lut_xbar( + enum dc_cm2_gpu_mem_pixel_component_order order, + enum hubp_3dlut_fl_crossbar_bit_slice *cr_r, + enum hubp_3dlut_fl_crossbar_bit_slice *y_g, + enum hubp_3dlut_fl_crossbar_bit_slice *cb_b) +{ + switch (order) { + case DC_CM2_GPU_MEM_PIXEL_COMPONENT_ORDER_RGBA: + *cr_r = hubp_3dlut_fl_crossbar_bit_slice_32_47; + *y_g = hubp_3dlut_fl_crossbar_bit_slice_16_31; + *cb_b = hubp_3dlut_fl_crossbar_bit_slice_0_15; + break; + case DC_CM2_GPU_MEM_PIXEL_COMPONENT_ORDER_BGRA: + *cr_r = hubp_3dlut_fl_crossbar_bit_slice_0_15; + *y_g = hubp_3dlut_fl_crossbar_bit_slice_16_31; + *cb_b = hubp_3dlut_fl_crossbar_bit_slice_32_47; + break; + } +} + +static void dc_get_lut_width( + enum dc_cm2_gpu_mem_size size, + enum hubp_3dlut_fl_width *width) +{ + switch (size) { + case DC_CM2_GPU_MEM_SIZE_333333: + *width = hubp_3dlut_fl_width_33; + break; + case DC_CM2_GPU_MEM_SIZE_171717: + *width = hubp_3dlut_fl_width_17; + break; + case DC_CM2_GPU_MEM_SIZE_TRANSFORMED: + *width = hubp_3dlut_fl_width_transformed; + break; + } +} +static bool dc_is_rmcm_3dlut_supported(struct hubp *hubp, struct mpc *mpc) +{ + if (mpc->funcs->rmcm.update_3dlut_fast_load_select && + mpc->funcs->rmcm.program_lut_read_write_control && + hubp->funcs->hubp_program_3dlut_fl_addr && + mpc->funcs->rmcm.program_bit_depth && + hubp->funcs->hubp_program_3dlut_fl_mode && + hubp->funcs->hubp_program_3dlut_fl_addressing_mode && + hubp->funcs->hubp_program_3dlut_fl_format && + hubp->funcs->hubp_update_3dlut_fl_bias_scale && + mpc->funcs->rmcm.program_bias_scale && + hubp->funcs->hubp_program_3dlut_fl_crossbar && + hubp->funcs->hubp_program_3dlut_fl_width && + mpc->funcs->rmcm.update_3dlut_fast_load_select && + mpc->funcs->rmcm.populate_lut && + mpc->funcs->rmcm.program_lut_mode && + hubp->funcs->hubp_enable_3dlut_fl && + mpc->funcs->rmcm.enable_3dlut_fl) + return true; + + return false; +} + +bool dcn401_program_rmcm_luts( + struct hubp *hubp, + struct pipe_ctx *pipe_ctx, + enum dc_cm2_transfer_func_source lut3d_src, + struct dc_cm2_func_luts *mcm_luts, + struct mpc *mpc, + bool lut_bank_a, + int mpcc_id) +{ + struct dpp *dpp_base = pipe_ctx->plane_res.dpp; + union mcm_lut_params m_lut_params; + enum MCM_LUT_XABLE shaper_xable, lut3d_xable = MCM_LUT_DISABLE, lut1d_xable; + enum hubp_3dlut_fl_mode mode; + enum hubp_3dlut_fl_addressing_mode addr_mode; + enum hubp_3dlut_fl_format format = 0; + enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_y_g = 0; + enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_cb_b = 0; + enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_cr_r = 0; + enum hubp_3dlut_fl_width width = 0; + struct dc *dc = hubp->ctx->dc; + + bool bypass_rmcm_3dlut = false; + bool bypass_rmcm_shaper = false; + + dcn401_get_mcm_lut_xable_from_pipe_ctx(dc, pipe_ctx, &shaper_xable, &lut3d_xable, &lut1d_xable); + + /* 3DLUT */ + switch (lut3d_src) { + case DC_CM2_TRANSFER_FUNC_SOURCE_SYSMEM: + memset(&m_lut_params, 0, sizeof(m_lut_params)); + // Don't know what to do in this case. + //case DC_CM2_TRANSFER_FUNC_SOURCE_SYSMEM: + break; + case DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM: + dc_get_lut_width(mcm_luts->lut3d_data.gpu_mem_params.size, &width); + if (!dc_is_rmcm_3dlut_supported(hubp, mpc) || + !mpc->funcs->rmcm.is_config_supported(width)) + return false; + + //0. disable fl on mpc + mpc->funcs->update_3dlut_fast_load_select(mpc, mpcc_id, 0xF); + + //1. power down the block + mpc->funcs->rmcm.power_on_shaper_3dlut(mpc, mpcc_id, false); + + //2. program RMCM + //2a. 3dlut reg programming + mpc->funcs->rmcm.program_lut_read_write_control(mpc, MCM_LUT_3DLUT, lut_bank_a, + (!bypass_rmcm_3dlut) && lut3d_xable != MCM_LUT_DISABLE, mpcc_id); + + hubp->funcs->hubp_program_3dlut_fl_addr(hubp, + mcm_luts->lut3d_data.gpu_mem_params.addr); + + mpc->funcs->rmcm.program_bit_depth(mpc, + mcm_luts->lut3d_data.gpu_mem_params.bit_depth, mpcc_id); + + // setting native or transformed mode, + dc_get_lut_mode(mcm_luts->lut3d_data.gpu_mem_params.layout, &mode, &addr_mode); + + //these program the mcm 3dlut + hubp->funcs->hubp_program_3dlut_fl_mode(hubp, mode); + + hubp->funcs->hubp_program_3dlut_fl_addressing_mode(hubp, addr_mode); + + //seems to be only for the MCM + dc_get_lut_format(mcm_luts->lut3d_data.gpu_mem_params.format_params.format, &format); + hubp->funcs->hubp_program_3dlut_fl_format(hubp, format); + + mpc->funcs->rmcm.program_bias_scale(mpc, + mcm_luts->lut3d_data.gpu_mem_params.format_params.float_params.bias, + mcm_luts->lut3d_data.gpu_mem_params.format_params.float_params.scale, + mpcc_id); + hubp->funcs->hubp_update_3dlut_fl_bias_scale(hubp, + mcm_luts->lut3d_data.gpu_mem_params.format_params.float_params.bias, + mcm_luts->lut3d_data.gpu_mem_params.format_params.float_params.scale); + + dc_get_lut_xbar( + mcm_luts->lut3d_data.gpu_mem_params.component_order, + &crossbar_bit_slice_cr_r, + &crossbar_bit_slice_y_g, + &crossbar_bit_slice_cb_b); + + hubp->funcs->hubp_program_3dlut_fl_crossbar(hubp, + crossbar_bit_slice_cr_r, + crossbar_bit_slice_y_g, + crossbar_bit_slice_cb_b); + + mpc->funcs->rmcm.program_3dlut_size(mpc, width, mpcc_id); + + mpc->funcs->update_3dlut_fast_load_select(mpc, mpcc_id, hubp->inst); + + //2b. shaper reg programming + memset(&m_lut_params, 0, sizeof(m_lut_params)); + + if (mcm_luts->shaper->type == TF_TYPE_HWPWL) { + m_lut_params.pwl = &mcm_luts->shaper->pwl; + } else if (mcm_luts->shaper->type == TF_TYPE_DISTRIBUTED_POINTS) { + ASSERT(false); + cm_helper_translate_curve_to_hw_format( + dc->ctx, + mcm_luts->shaper, + &dpp_base->regamma_params, true); + m_lut_params.pwl = &dpp_base->regamma_params; + } + if (m_lut_params.pwl) { + mpc->funcs->rmcm.populate_lut(mpc, m_lut_params, lut_bank_a, mpcc_id); + mpc->funcs->rmcm.program_lut_mode(mpc, !bypass_rmcm_shaper, lut_bank_a, mpcc_id); + } else { + //RMCM 3dlut won't work without its shaper + return false; + } + + //3. Select the hubp connected to this RMCM + hubp->funcs->hubp_enable_3dlut_fl(hubp, true); + mpc->funcs->rmcm.enable_3dlut_fl(mpc, true, mpcc_id); + + //4. power on the block + if (m_lut_params.pwl) + mpc->funcs->rmcm.power_on_shaper_3dlut(mpc, mpcc_id, true); + + break; + default: + return false; + } + + return true; +} + void dcn401_populate_mcm_luts(struct dc *dc, struct pipe_ctx *pipe_ctx, struct dc_cm2_func_luts mcm_luts, @@ -407,21 +650,39 @@ void dcn401_populate_mcm_luts(struct dc *dc, struct mpc *mpc = dc->res_pool->mpc; union mcm_lut_params m_lut_params; enum dc_cm2_transfer_func_source lut3d_src = mcm_luts.lut3d_data.lut3d_src; - enum hubp_3dlut_fl_format format; + enum hubp_3dlut_fl_format format = 0; enum hubp_3dlut_fl_mode mode; - enum hubp_3dlut_fl_width width; + enum hubp_3dlut_fl_width width = 0; enum hubp_3dlut_fl_addressing_mode addr_mode; - enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_y_g; - enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_cb_b; - enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_cr_r; + enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_y_g = 0; + enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_cb_b = 0; + enum hubp_3dlut_fl_crossbar_bit_slice crossbar_bit_slice_cr_r = 0; enum MCM_LUT_XABLE shaper_xable = MCM_LUT_DISABLE; enum MCM_LUT_XABLE lut3d_xable = MCM_LUT_DISABLE; enum MCM_LUT_XABLE lut1d_xable = MCM_LUT_DISABLE; - bool is_17x17x17 = true; bool rval; dcn401_get_mcm_lut_xable_from_pipe_ctx(dc, pipe_ctx, &shaper_xable, &lut3d_xable, &lut1d_xable); + //MCM - setting its location (Before/After) blender + //set to post blend (true) + dcn401_set_mcm_location_post_blend( + dc, + pipe_ctx, + mcm_luts.lut3d_data.mpc_mcm_post_blend); + + //RMCM - 3dLUT+Shaper + if (mcm_luts.lut3d_data.rmcm_3dlut_enable) { + dcn401_program_rmcm_luts( + hubp, + pipe_ctx, + lut3d_src, + &mcm_luts, + mpc, + lut_bank_a, + mpcc_id); + } + /* 1D LUT */ if (mcm_luts.lut1d_func) { memset(&m_lut_params, 0, sizeof(m_lut_params)); @@ -442,7 +703,7 @@ void dcn401_populate_mcm_luts(struct dc *dc, } /* Shaper */ - if (mcm_luts.shaper) { + if (mcm_luts.shaper && mcm_luts.lut3d_data.mpc_3dlut_enable) { memset(&m_lut_params, 0, sizeof(m_lut_params)); if (mcm_luts.shaper->type == TF_TYPE_HWPWL) m_lut_params.pwl = &mcm_luts.shaper->pwl; @@ -454,11 +715,11 @@ void dcn401_populate_mcm_luts(struct dc *dc, m_lut_params.pwl = rval ? &dpp_base->regamma_params : NULL; } if (m_lut_params.pwl) { - if (mpc->funcs->populate_lut) - mpc->funcs->populate_lut(mpc, MCM_LUT_SHAPER, m_lut_params, lut_bank_a, mpcc_id); + if (mpc->funcs->mcm.populate_lut) + mpc->funcs->mcm.populate_lut(mpc, m_lut_params, lut_bank_a, mpcc_id); + if (mpc->funcs->program_lut_mode) + mpc->funcs->program_lut_mode(mpc, MCM_LUT_SHAPER, MCM_LUT_ENABLE, lut_bank_a, mpcc_id); } - if (mpc->funcs->program_lut_mode) - mpc->funcs->program_lut_mode(mpc, MCM_LUT_SHAPER, shaper_xable, lut_bank_a, mpcc_id); } /* 3DLUT */ @@ -467,6 +728,7 @@ void dcn401_populate_mcm_luts(struct dc *dc, memset(&m_lut_params, 0, sizeof(m_lut_params)); if (hubp->funcs->hubp_enable_3dlut_fl) hubp->funcs->hubp_enable_3dlut_fl(hubp, false); + if (mcm_luts.lut3d_data.lut3d_func && mcm_luts.lut3d_data.lut3d_func->state.bits.initialized) { m_lut_params.lut3d = &mcm_luts.lut3d_data.lut3d_func->lut_3d; if (mpc->funcs->populate_lut) @@ -476,16 +738,35 @@ void dcn401_populate_mcm_luts(struct dc *dc, mpcc_id); } break; - case DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM: + case DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM: + switch (mcm_luts.lut3d_data.gpu_mem_params.size) { + case DC_CM2_GPU_MEM_SIZE_333333: + width = hubp_3dlut_fl_width_33; + break; + case DC_CM2_GPU_MEM_SIZE_171717: + width = hubp_3dlut_fl_width_17; + break; + case DC_CM2_GPU_MEM_SIZE_TRANSFORMED: + width = hubp_3dlut_fl_width_transformed; + break; + } + + //check for support + if (mpc->funcs->mcm.is_config_supported && + !mpc->funcs->mcm.is_config_supported(width)) + break; if (mpc->funcs->program_lut_read_write_control) mpc->funcs->program_lut_read_write_control(mpc, MCM_LUT_3DLUT, lut_bank_a, mpcc_id); if (mpc->funcs->program_lut_mode) mpc->funcs->program_lut_mode(mpc, MCM_LUT_3DLUT, lut3d_xable, lut_bank_a, mpcc_id); - if (mpc->funcs->program_3dlut_size) - mpc->funcs->program_3dlut_size(mpc, is_17x17x17, mpcc_id); + if (hubp->funcs->hubp_program_3dlut_fl_addr) hubp->funcs->hubp_program_3dlut_fl_addr(hubp, mcm_luts.lut3d_data.gpu_mem_params.addr); + + if (mpc->funcs->mcm.program_bit_depth) + mpc->funcs->mcm.program_bit_depth(mpc, mcm_luts.lut3d_data.gpu_mem_params.bit_depth, mpcc_id); + switch (mcm_luts.lut3d_data.gpu_mem_params.layout) { case DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_RGB: mode = hubp_3dlut_fl_mode_native_1; @@ -512,7 +793,6 @@ void dcn401_populate_mcm_luts(struct dc *dc, switch (mcm_luts.lut3d_data.gpu_mem_params.format_params.format) { case DC_CM2_GPU_MEM_FORMAT_16161616_UNORM_12MSB: - default: format = hubp_3dlut_fl_format_unorm_12msb_bitslice; break; case DC_CM2_GPU_MEM_FORMAT_16161616_UNORM_12LSB: @@ -524,37 +804,37 @@ void dcn401_populate_mcm_luts(struct dc *dc, } if (hubp->funcs->hubp_program_3dlut_fl_format) hubp->funcs->hubp_program_3dlut_fl_format(hubp, format); - if (hubp->funcs->hubp_update_3dlut_fl_bias_scale) + if (hubp->funcs->hubp_update_3dlut_fl_bias_scale && + mpc->funcs->mcm.program_bias_scale) { + mpc->funcs->mcm.program_bias_scale(mpc, + mcm_luts.lut3d_data.gpu_mem_params.format_params.float_params.bias, + mcm_luts.lut3d_data.gpu_mem_params.format_params.float_params.scale, + mpcc_id); hubp->funcs->hubp_update_3dlut_fl_bias_scale(hubp, - mcm_luts.lut3d_data.gpu_mem_params.format_params.float_params.bias, - mcm_luts.lut3d_data.gpu_mem_params.format_params.float_params.scale); - - switch (mcm_luts.lut3d_data.gpu_mem_params.component_order) { - case DC_CM2_GPU_MEM_PIXEL_COMPONENT_ORDER_RGBA: - default: - crossbar_bit_slice_cr_r = hubp_3dlut_fl_crossbar_bit_slice_0_15; - crossbar_bit_slice_y_g = hubp_3dlut_fl_crossbar_bit_slice_16_31; - crossbar_bit_slice_cb_b = hubp_3dlut_fl_crossbar_bit_slice_32_47; - break; + mcm_luts.lut3d_data.gpu_mem_params.format_params.float_params.bias, + mcm_luts.lut3d_data.gpu_mem_params.format_params.float_params.scale); } + //navi 4x has a bug and r and blue are swapped and need to be worked around here in + //TODO: need to make a method for get_xbar per asic OR do the workaround in program_crossbar for 4x + dc_get_lut_xbar( + mcm_luts.lut3d_data.gpu_mem_params.component_order, + &crossbar_bit_slice_cr_r, + &crossbar_bit_slice_y_g, + &crossbar_bit_slice_cb_b); + if (hubp->funcs->hubp_program_3dlut_fl_crossbar) hubp->funcs->hubp_program_3dlut_fl_crossbar(hubp, + crossbar_bit_slice_cr_r, crossbar_bit_slice_y_g, - crossbar_bit_slice_cb_b, - crossbar_bit_slice_cr_r); + crossbar_bit_slice_cb_b); + + if (mpc->funcs->mcm.program_lut_read_write_control) + mpc->funcs->mcm.program_lut_read_write_control(mpc, MCM_LUT_3DLUT, lut_bank_a, true, mpcc_id); + + if (mpc->funcs->mcm.program_3dlut_size) + mpc->funcs->mcm.program_3dlut_size(mpc, width, mpcc_id); - switch (mcm_luts.lut3d_data.gpu_mem_params.size) { - case DC_CM2_GPU_MEM_SIZE_171717: - default: - width = hubp_3dlut_fl_width_17; - break; - case DC_CM2_GPU_MEM_SIZE_TRANSFORMED: - width = hubp_3dlut_fl_width_transformed; - break; - } - if (hubp->funcs->hubp_program_3dlut_fl_width) - hubp->funcs->hubp_program_3dlut_fl_width(hubp, width); if (mpc->funcs->update_3dlut_fast_load_select) mpc->funcs->update_3dlut_fast_load_select(mpc, mpcc_id, hubp->inst); @@ -1980,9 +2260,9 @@ void dcn401_program_pipe( dc->res_pool->hubbub, pipe_ctx->plane_res.hubp->inst, pipe_ctx->hubp_regs.det_size); } - if (pipe_ctx->update_flags.raw || - (pipe_ctx->plane_state && pipe_ctx->plane_state->update_flags.raw) || - pipe_ctx->stream->update_flags.raw) + if (pipe_ctx->plane_state && (pipe_ctx->update_flags.raw || + pipe_ctx->plane_state->update_flags.raw || + pipe_ctx->stream->update_flags.raw)) dc->hwss.update_dchubp_dpp(dc, pipe_ctx, context); if (pipe_ctx->plane_state && (pipe_ctx->update_flags.bits.enable || @@ -2081,7 +2361,7 @@ void dcn401_program_front_end_for_ctx( for (i = 0; i < dc->res_pool->pipe_count; i++) { pipe = &context->res_ctx.pipe_ctx[i]; - if (!pipe->top_pipe && !pipe->prev_odm_pipe && pipe->plane_state) { + if (pipe->plane_state) { if (pipe->plane_state->triplebuffer_flips) BREAK_TO_DEBUGGER(); @@ -2371,7 +2651,7 @@ bool dcn401_update_bandwidth( struct dce_hwseq *hws = dc->hwseq; /* recalculate DML parameters */ - if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false)) + if (dc->res_pool->funcs->validate_bandwidth(dc, context, false) != DC_OK) return false; /* apply updated bandwidth parameters */ diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h index 781cf0efccc6..ce65b4f6c672 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h @@ -109,4 +109,12 @@ void dcn401_detect_pipe_changes( void dcn401_plane_atomic_power_down(struct dc *dc, struct dpp *dpp, struct hubp *hubp); +bool dcn401_program_rmcm_luts( + struct hubp *hubp, + struct pipe_ctx *pipe_ctx, + enum dc_cm2_transfer_func_source lut3d_src, + struct dc_cm2_func_luts *mcm_luts, + struct mpc *mpc, + bool lut_bank_a, + int mpcc_id); #endif /* __DC_HWSS_DCN401_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h index c8b5ed834579..3a0795045bc6 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h @@ -195,6 +195,8 @@ enum block_sequence_func { DMUB_SUBVP_SAVE_SURF_ADDR, HUBP_WAIT_FOR_DCC_META_PROP, DMUB_FAMS2_GLOBAL_CONTROL_LOCK_FAST, + /* This must be the last value in this enum, add new ones above */ + HWSS_BLOCK_SEQUENCE_FUNC_COUNT }; struct block_sequence { @@ -202,6 +204,8 @@ struct block_sequence { enum block_sequence_func func; }; +#define MAX_HWSS_BLOCK_SEQUENCE_SIZE (HWSS_BLOCK_SEQUENCE_FUNC_COUNT * MAX_PIPES) + struct hw_sequencer_funcs { void (*hardware_release)(struct dc *dc); /* Embedded Display Related */ @@ -534,13 +538,13 @@ void set_drr_and_clear_adjust_pending( struct drr_params *params); void hwss_execute_sequence(struct dc *dc, - struct block_sequence block_sequence[], + struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE], int num_steps); void hwss_build_fast_sequence(struct dc *dc, struct dc_dmub_cmd *dc_dmub_cmd, unsigned int dmub_cmd_count, - struct block_sequence block_sequence[], + struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE], unsigned int *num_steps, struct pipe_ctx *pipe_ctx, struct dc_stream_status *stream_status, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer_private.h b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer_private.h index 22a5d4a03c98..09bc65c2fa23 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer_private.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer_private.h @@ -183,6 +183,8 @@ struct hwseq_private_funcs { struct dc_cm2_func_luts mcm_luts, bool lut_bank_a); void (*perform_3dlut_wa_unlock)(struct pipe_ctx *pipe_ctx); + void (*wait_for_pipe_update_if_needed)(struct dc *dc, struct pipe_ctx *pipe_ctx, bool is_surface_update_only); + void (*set_wait_for_update_needed_for_pipe)(struct dc *dc, struct pipe_ctx *pipe_ctx); }; struct dce_hwseq { diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_status.h b/drivers/gpu/drm/amd/display/dc/inc/core_status.h index b5afd8c3103d..f3696143590c 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_status.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_status.h @@ -26,6 +26,8 @@ #ifndef _CORE_STATUS_H_ #define _CORE_STATUS_H_ +#include "dc_hw_types.h" + enum dc_status { DC_OK = 1, @@ -56,6 +58,7 @@ enum dc_status { DC_NO_LINK_ENC_RESOURCE = 26, DC_FAIL_DP_PAYLOAD_ALLOCATION = 27, DC_FAIL_DP_LINK_BANDWIDTH = 28, + DC_FAIL_HW_CURSOR_SUPPORT = 29, DC_ERROR_UNEXPECTED = -1 }; diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h index d0021f25f3d8..0cf349cafb3e 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h @@ -65,6 +65,7 @@ struct resource_pool; struct dc_state; struct resource_context; struct clk_bw_params; +struct dc_mcache_params; struct resource_funcs { enum engine_id (*get_preferred_eng_id_dpia)(unsigned int dpia_index); @@ -78,8 +79,7 @@ struct resource_funcs { /* Create a minimal link encoder object with no dc_link object * associated with it. */ struct link_encoder *(*link_enc_create_minimal)(struct dc_context *ctx, enum engine_id eng_id); - - bool (*validate_bandwidth)( + enum dc_status (*validate_bandwidth)( struct dc *dc, struct dc_state *context, bool fast_validate); @@ -218,6 +218,11 @@ struct resource_funcs { int (*get_power_profile)(const struct dc_state *context); unsigned int (*get_det_buffer_size)(const struct dc_state *context); unsigned int (*get_vstartup_for_pipe)(struct pipe_ctx *pipe_ctx); + unsigned int (*get_max_hw_cursor_size)(const struct dc *dc, + struct dc_state *state, + const struct dc_stream_state *stream); + bool (*program_mcache_pipe_config)(struct dc_state *context, + const struct dc_mcache_params *mcache_params); }; struct audio_support{ @@ -382,7 +387,9 @@ struct link_resource { struct link_config { struct dc_link_settings dp_link_settings; + struct dc_tunnel_settings dp_tunnel_settings; }; + union pipe_update_flags { struct { uint32_t enable : 1; @@ -480,6 +487,10 @@ struct pipe_ctx { struct pixel_rate_divider pixel_rate_divider; /* pixels borrowed from hblank to hactive */ uint8_t hblank_borrow; + /* next vupdate */ + uint32_t next_vupdate; + uint32_t wait_frame_count; + bool wait_is_required; }; /* Data used for dynamic link encoder assignment. @@ -507,7 +518,7 @@ struct resource_context { unsigned int hpo_dp_link_enc_to_link_idx[MAX_HPO_DP2_LINK_ENCODERS]; int hpo_dp_link_enc_ref_cnts[MAX_HPO_DP2_LINK_ENCODERS]; bool is_mpc_3dlut_acquired[MAX_PIPES]; - /* solely used for build scalar data in dml2 */ + /* used to build scalar data in dml2 and for edp backlight programming */ struct pipe_ctx temp_pipe; }; @@ -630,7 +641,7 @@ struct dc_state { */ struct bw_context bw_ctx; - struct block_sequence block_sequence[100]; + struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE]; unsigned int block_sequence_steps; struct dc_dmub_cmd dc_dmub_cmd[10]; unsigned int dmub_cmd_count; diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h index 221645c023b5..bac8febad69a 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h @@ -199,6 +199,7 @@ enum dentist_divider_range { CLK_SR_DCN35(CLK1_CLK4_ALLOW_DS), \ CLK_SR_DCN35(CLK1_CLK5_ALLOW_DS), \ CLK_SR_DCN35(CLK5_spll_field_8), \ + CLK_SR_DCN35(CLK6_spll_field_8), \ SR(DENTIST_DISPCLK_CNTL), \ #define CLK_COMMON_MASK_SH_LIST_DCN32(mask_sh) \ @@ -307,7 +308,7 @@ struct clk_mgr_registers { uint32_t CLK1_CLK4_ALLOW_DS; uint32_t CLK1_CLK5_ALLOW_DS; uint32_t CLK5_spll_field_8; - + uint32_t CLK6_spll_field_8; }; struct clk_mgr_shift { diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h b/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h index 3a89cc0cffc1..6e303b81bfb0 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h @@ -967,23 +967,6 @@ struct mpc_funcs { */ void (*update_3dlut_fast_load_select)(struct mpc *mpc, int mpcc_id, int hubp_idx); - /** - * @get_3dlut_fast_load_status: - * - * Get 3D LUT fast load status and reference them with done, soft_underflow and hard_underflow pointers. - * - * Parameters: - * - [in/out] mpc - MPC context. - * - [in] mpcc_id - * - [in/out] done - * - [in/out] soft_underflow - * - [in/out] hard_underflow - * - * Return: - * - * void - */ - void (*get_3dlut_fast_load_status)(struct mpc *mpc, int mpcc_id, uint32_t *done, uint32_t *soft_underflow, uint32_t *hard_underflow); /** * @populate_lut: @@ -1054,6 +1037,35 @@ struct mpc_funcs { * void */ void (*program_3dlut_size)(struct mpc *mpc, bool is_17x17x17, int mpcc_id); + + struct { + void (*program_3dlut_size)(struct mpc *mpc, uint32_t width, int mpcc_id); + void (*program_bias_scale)(struct mpc *mpc, uint16_t bias, uint16_t scale, int mpcc_id); + void (*program_bit_depth)(struct mpc *mpc, uint16_t bit_depth, int mpcc_id); + bool (*is_config_supported)(uint32_t width); + void (*program_lut_read_write_control)(struct mpc *mpc, const enum MCM_LUT_ID id, + bool lut_bank_a, bool enabled, int mpcc_id); + + void (*populate_lut)(struct mpc *mpc, const union mcm_lut_params params, + bool lut_bank_a, int mpcc_id); + } mcm; + + struct { + void (*enable_3dlut_fl)(struct mpc *mpc, bool enable, int mpcc_id); + void (*update_3dlut_fast_load_select)(struct mpc *mpc, int mpcc_id, int hubp_idx); + void (*program_lut_read_write_control)(struct mpc *mpc, const enum MCM_LUT_ID id, + bool lut_bank_a, bool enabled, int mpcc_id); + void (*program_lut_mode)(struct mpc *mpc, const enum MCM_LUT_XABLE xable, + bool lut_bank_a, int mpcc_id); + void (*program_3dlut_size)(struct mpc *mpc, uint32_t width, int mpcc_id); + void (*program_bias_scale)(struct mpc *mpc, uint16_t bias, uint16_t scale, int mpcc_id); + void (*program_bit_depth)(struct mpc *mpc, uint16_t bit_depth, int mpcc_id); + bool (*is_config_supported)(uint32_t width); + + void (*power_on_shaper_3dlut)(struct mpc *mpc, uint32_t mpcc_id, bool power_on); + void (*populate_lut)(struct mpc *mpc, const union mcm_lut_params params, + bool lut_bank_a, int mpcc_id); + } rmcm; }; #endif diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h b/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h index 7f371cbb35cd..0d5a8358a778 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h @@ -68,6 +68,7 @@ struct optc { int pstate_keepout; struct dc_crtc_timing orginal_patched_timing; enum signal_type signal; + uint32_t max_frame_count; }; void optc1_read_otg_state(struct timing_generator *optc, struct dcn_otg_state *s); diff --git a/drivers/gpu/drm/amd/display/dc/inc/link.h b/drivers/gpu/drm/amd/display/dc/inc/link.h index 2948a696ee12..7d16351bba99 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/link.h +++ b/drivers/gpu/drm/amd/display/dc/inc/link.h @@ -207,6 +207,9 @@ struct link_service { bool (*dp_decide_link_settings)( struct dc_stream_state *stream, struct dc_link_settings *link_setting); + void (*dp_decide_tunnel_settings)( + struct dc_stream_state *stream, + struct dc_tunnel_settings *dp_tunnel_setting); enum dp_link_encoding (*mst_decide_link_encoding_format)( const struct dc_link *link); bool (*edp_decide_link_settings)(struct dc_link *link, diff --git a/drivers/gpu/drm/amd/display/dc/inc/reg_helper.h b/drivers/gpu/drm/amd/display/dc/inc/reg_helper.h index a402df225a76..26cb1459b743 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/reg_helper.h +++ b/drivers/gpu/drm/amd/display/dc/inc/reg_helper.h @@ -508,6 +508,10 @@ uint32_t generic_indirect_reg_update_ex(const struct dc_context *ctx, initial_val, \ n, __VA_ARGS__) +#define IX_REG_SET_SYNC(index, init_value, f1, v1) \ + IX_REG_SET_N_SYNC(index, 1, init_value, \ + FN(reg, f1), v1) + #define IX_REG_SET_2_SYNC(index, init_value, f1, v1, f2, v2) \ IX_REG_SET_N_SYNC(index, 2, init_value, \ FN(reg, f1), v1,\ diff --git a/drivers/gpu/drm/amd/display/dc/inc/resource.h b/drivers/gpu/drm/amd/display/dc/inc/resource.h index 9458187b834d..a890f581f4e8 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/resource.h +++ b/drivers/gpu/drm/amd/display/dc/inc/resource.h @@ -32,6 +32,7 @@ #define MEMORY_TYPE_MULTIPLIER_CZ 4 #define MEMORY_TYPE_HBM 2 +#define MAX_MCACHES 8 #define IS_PIPE_SYNCD_VALID(pipe) ((((pipe)->pipe_idx_syncd) & 0x80)?1:0) @@ -65,6 +66,13 @@ struct resource_straps { uint32_t audio_stream_number; }; +struct dc_mcache_allocations { + int global_mcache_ids_plane0[MAX_MCACHES + 1]; + int global_mcache_ids_plane1[MAX_MCACHES + 1]; + int global_mcache_ids_mall_plane0[MAX_MCACHES + 1]; + int global_mcache_ids_mall_plane1[MAX_MCACHES + 1]; +}; + struct resource_create_funcs { void (*read_dce_straps)( struct dc_context *ctx, struct resource_straps *straps); @@ -628,8 +636,6 @@ enum dc_status update_dp_encoder_resources_for_test_harness(const struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx); -bool check_subvp_sw_cursor_fallback_req(const struct dc *dc, struct dc_stream_state *stream); - /* Get hw programming parameters container from pipe context * @pipe_ctx: pipe context * @dscl_prog_data: struct to hold programmable hw reg values diff --git a/drivers/gpu/drm/amd/display/dc/irq/dce120/irq_service_dce120.c b/drivers/gpu/drm/amd/display/dc/irq/dce120/irq_service_dce120.c index 953f4a4dacad..33ce470e4c88 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dce120/irq_service_dce120.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dce120/irq_service_dce120.c @@ -37,36 +37,9 @@ #include "ivsrcid/ivsrcid_vislands30.h" -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dce60/irq_service_dce60.c b/drivers/gpu/drm/amd/display/dc/irq/dce60/irq_service_dce60.c index 2c72074310c7..d777b85e70da 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dce60/irq_service_dce60.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dce60/irq_service_dce60.c @@ -46,36 +46,9 @@ #include "dc_types.h" -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - DC_HPD1_INT_STATUS, - DC_HPD1_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - DC_HPD1_INT_CONTROL, - DC_HPD1_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd1_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { @@ -391,5 +364,3 @@ struct irq_service *dal_irq_service_dce60_create( dce60_irq_construct(irq_service, init_data); return irq_service; } - - diff --git a/drivers/gpu/drm/amd/display/dc/irq/dce80/irq_service_dce80.c b/drivers/gpu/drm/amd/display/dc/irq/dce80/irq_service_dce80.c index 49317934ef4f..3a9163acb49b 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dce80/irq_service_dce80.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dce80/irq_service_dce80.c @@ -37,36 +37,9 @@ #include "dc_types.h" -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - DC_HPD1_INT_STATUS, - DC_HPD1_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - DC_HPD1_INT_CONTROL, - DC_HPD1_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd1_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { @@ -303,5 +276,3 @@ struct irq_service *dal_irq_service_dce80_create( dce80_irq_construct(irq_service, init_data); return irq_service; } - - diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c b/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c index 9ca28565a9d1..4ce9edd16344 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c @@ -129,36 +129,9 @@ static enum dc_irq_source to_dal_irq_source_dcn10(struct irq_service *irq_servic } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn20/irq_service_dcn20.c b/drivers/gpu/drm/amd/display/dc/irq/dcn20/irq_service_dcn20.c index 916f0c974637..5847af0e66cb 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn20/irq_service_dcn20.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn20/irq_service_dcn20.c @@ -130,36 +130,9 @@ static enum dc_irq_source to_dal_irq_source_dcn20( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c b/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c index 1d61d475d36f..6417011d2246 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c @@ -80,36 +80,9 @@ static enum dc_irq_source to_dal_irq_source_dcn201( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c b/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c index 42cdfe6c3538..71d2f065140b 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn21/irq_service_dcn21.c @@ -132,36 +132,9 @@ static enum dc_irq_source to_dal_irq_source_dcn21(struct irq_service *irq_servic return DC_IRQ_SOURCE_INVALID; } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn30/irq_service_dcn30.c b/drivers/gpu/drm/amd/display/dc/irq/dcn30/irq_service_dcn30.c index a443a8abb1ea..2a4080bdcf6b 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn30/irq_service_dcn30.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn30/irq_service_dcn30.c @@ -139,36 +139,9 @@ static enum dc_irq_source to_dal_irq_source_dcn30( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { @@ -447,4 +420,3 @@ struct irq_service *dal_irq_service_dcn30_create( dcn30_irq_construct(irq_service, init_data); return irq_service; } - diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn302/irq_service_dcn302.c b/drivers/gpu/drm/amd/display/dc/irq/dcn302/irq_service_dcn302.c index 8ffc7e2c681a..624f1ac309f8 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn302/irq_service_dcn302.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn302/irq_service_dcn302.c @@ -126,26 +126,9 @@ static enum dc_irq_source to_dal_irq_source_dcn302(struct irq_service *irq_servi } } -static bool hpd_ack(struct irq_service *irq_service, const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = get_reg_field_value(value, HPD0_DC_HPD_INT_STATUS, DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value(value, current_status ? 0 : 1, HPD0_DC_HPD_INT_CONTROL, DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn303/irq_service_dcn303.c b/drivers/gpu/drm/amd/display/dc/irq/dcn303/irq_service_dcn303.c index 262bb8b74b15..137caffae916 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn303/irq_service_dcn303.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn303/irq_service_dcn303.c @@ -77,26 +77,9 @@ static enum dc_irq_source to_dal_irq_source_dcn303(struct irq_service *irq_servi } } -static bool hpd_ack(struct irq_service *irq_service, const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = get_reg_field_value(value, HPD0_DC_HPD_INT_STATUS, DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value(value, current_status ? 0 : 1, HPD0_DC_HPD_INT_CONTROL, DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn31/irq_service_dcn31.c b/drivers/gpu/drm/amd/display/dc/irq/dcn31/irq_service_dcn31.c index 53e78ae7eecf..921cb167d920 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn31/irq_service_dcn31.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn31/irq_service_dcn31.c @@ -128,36 +128,9 @@ static enum dc_irq_source to_dal_irq_source_dcn31(struct irq_service *irq_servic } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn314/irq_service_dcn314.c b/drivers/gpu/drm/amd/display/dc/irq/dcn314/irq_service_dcn314.c index e0563e880432..0118fd6e5db0 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn314/irq_service_dcn314.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn314/irq_service_dcn314.c @@ -130,36 +130,9 @@ static enum dc_irq_source to_dal_irq_source_dcn314(struct irq_service *irq_servi } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn315/irq_service_dcn315.c b/drivers/gpu/drm/amd/display/dc/irq/dcn315/irq_service_dcn315.c index 2ef22299101a..adebfc888618 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn315/irq_service_dcn315.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn315/irq_service_dcn315.c @@ -135,36 +135,9 @@ static enum dc_irq_source to_dal_irq_source_dcn315( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn32/irq_service_dcn32.c b/drivers/gpu/drm/amd/display/dc/irq/dcn32/irq_service_dcn32.c index f0ac0aeeac51..e9e315c75d76 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn32/irq_service_dcn32.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn32/irq_service_dcn32.c @@ -129,36 +129,9 @@ static enum dc_irq_source to_dal_irq_source_dcn32( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { @@ -191,6 +164,16 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { .ack = NULL }; +static struct irq_source_info_funcs vline1_irq_info_funcs = { + .set = NULL, + .ack = NULL +}; + +static struct irq_source_info_funcs vline2_irq_info_funcs = { + .set = NULL, + .ack = NULL +}; + #undef BASE_INNER #define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg @@ -259,6 +242,13 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { .funcs = &pflip_irq_info_funcs\ } +#define vblank_int_entry(reg_num)\ + [DC_IRQ_SOURCE_VBLANK1 + reg_num] = {\ + IRQ_REG_ENTRY(OTG, reg_num,\ + OTG_GLOBAL_SYNC_STATUS, VSTARTUP_INT_EN,\ + OTG_GLOBAL_SYNC_STATUS, VSTARTUP_EVENT_CLEAR),\ + .funcs = &vblank_irq_info_funcs\ + } /* vupdate_no_lock_int_entry maps to DC_IRQ_SOURCE_VUPDATEx, to match semantic * of DCE's DC_IRQ_SOURCE_VUPDATEx. */ @@ -270,14 +260,6 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { .funcs = &vupdate_no_lock_irq_info_funcs\ } -#define vblank_int_entry(reg_num)\ - [DC_IRQ_SOURCE_VBLANK1 + reg_num] = {\ - IRQ_REG_ENTRY(OTG, reg_num,\ - OTG_GLOBAL_SYNC_STATUS, VSTARTUP_INT_EN,\ - OTG_GLOBAL_SYNC_STATUS, VSTARTUP_EVENT_CLEAR),\ - .funcs = &vblank_irq_info_funcs\ -} - #define vline0_int_entry(reg_num)\ [DC_IRQ_SOURCE_DC1_VLINE0 + reg_num] = {\ IRQ_REG_ENTRY(OTG, reg_num,\ @@ -285,6 +267,20 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { OTG_VERTICAL_INTERRUPT0_CONTROL, OTG_VERTICAL_INTERRUPT0_CLEAR),\ .funcs = &vline0_irq_info_funcs\ } +#define vline1_int_entry(reg_num)\ + [DC_IRQ_SOURCE_DC1_VLINE1 + reg_num] = {\ + IRQ_REG_ENTRY(OTG, reg_num,\ + OTG_VERTICAL_INTERRUPT1_CONTROL, OTG_VERTICAL_INTERRUPT1_INT_ENABLE,\ + OTG_VERTICAL_INTERRUPT1_CONTROL, OTG_VERTICAL_INTERRUPT1_CLEAR),\ + .funcs = &vline1_irq_info_funcs\ + } +#define vline2_int_entry(reg_num)\ + [DC_IRQ_SOURCE_DC1_VLINE2 + reg_num] = {\ + IRQ_REG_ENTRY(OTG, reg_num,\ + OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_INT_ENABLE,\ + OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_CLEAR),\ + .funcs = &vline2_irq_info_funcs\ + } #define dmub_outbox_int_entry()\ [DC_IRQ_SOURCE_DMCUB_OUTBOX] = {\ IRQ_REG_ENTRY_DMUB(\ @@ -387,21 +383,29 @@ irq_source_info_dcn32[DAL_IRQ_SOURCES_NUMBER] = { dc_underflow_int_entry(6), [DC_IRQ_SOURCE_DMCU_SCP] = dummy_irq_entry(), [DC_IRQ_SOURCE_VBIOS_SW] = dummy_irq_entry(), - vupdate_no_lock_int_entry(0), - vupdate_no_lock_int_entry(1), - vupdate_no_lock_int_entry(2), - vupdate_no_lock_int_entry(3), vblank_int_entry(0), vblank_int_entry(1), vblank_int_entry(2), vblank_int_entry(3), + [DC_IRQ_SOURCE_DC5_VLINE1] = dummy_irq_entry(), + [DC_IRQ_SOURCE_DC6_VLINE1] = dummy_irq_entry(), + dmub_outbox_int_entry(), + vupdate_no_lock_int_entry(0), + vupdate_no_lock_int_entry(1), + vupdate_no_lock_int_entry(2), + vupdate_no_lock_int_entry(3), vline0_int_entry(0), vline0_int_entry(1), vline0_int_entry(2), vline0_int_entry(3), - [DC_IRQ_SOURCE_DC5_VLINE1] = dummy_irq_entry(), - [DC_IRQ_SOURCE_DC6_VLINE1] = dummy_irq_entry(), - dmub_outbox_int_entry(), + vline1_int_entry(0), + vline1_int_entry(1), + vline1_int_entry(2), + vline1_int_entry(3), + vline2_int_entry(0), + vline2_int_entry(1), + vline2_int_entry(2), + vline2_int_entry(3) }; static const struct irq_service_funcs irq_service_funcs_dcn32 = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn35/irq_service_dcn35.c b/drivers/gpu/drm/amd/display/dc/irq/dcn35/irq_service_dcn35.c index ea8c271171bc..79e5e8c137ca 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn35/irq_service_dcn35.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn35/irq_service_dcn35.c @@ -127,36 +127,9 @@ static enum dc_irq_source to_dal_irq_source_dcn35( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn351/irq_service_dcn351.c b/drivers/gpu/drm/amd/display/dc/irq/dcn351/irq_service_dcn351.c index 7ec8e0de2f01..163b8ee9ebf7 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn351/irq_service_dcn351.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn351/irq_service_dcn351.c @@ -106,36 +106,9 @@ static enum dc_irq_source to_dal_irq_source_dcn351( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn36/irq_service_dcn36.c b/drivers/gpu/drm/amd/display/dc/irq/dcn36/irq_service_dcn36.c index ea958628f8b8..f716ab0fd30e 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn36/irq_service_dcn36.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn36/irq_service_dcn36.c @@ -105,36 +105,9 @@ static enum dc_irq_source to_dal_irq_source_dcn36( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn401/irq_service_dcn401.c b/drivers/gpu/drm/amd/display/dc/irq/dcn401/irq_service_dcn401.c index b43c9524b0de..fd9bb1950c20 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/dcn401/irq_service_dcn401.c +++ b/drivers/gpu/drm/amd/display/dc/irq/dcn401/irq_service_dcn401.c @@ -109,36 +109,9 @@ static enum dc_irq_source to_dal_irq_source_dcn401( } } -static bool hpd_ack( - struct irq_service *irq_service, - const struct irq_source_info *info) -{ - uint32_t addr = info->status_reg; - uint32_t value = dm_read_reg(irq_service->ctx, addr); - uint32_t current_status = - get_reg_field_value( - value, - HPD0_DC_HPD_INT_STATUS, - DC_HPD_SENSE_DELAYED); - - dal_irq_service_ack_generic(irq_service, info); - - value = dm_read_reg(irq_service->ctx, info->enable_reg); - - set_reg_field_value( - value, - current_status ? 0 : 1, - HPD0_DC_HPD_INT_CONTROL, - DC_HPD_INT_POLARITY); - - dm_write_reg(irq_service->ctx, info->enable_reg, value); - - return true; -} - static struct irq_source_info_funcs hpd_irq_info_funcs = { .set = NULL, - .ack = hpd_ack + .ack = hpd0_ack }; static struct irq_source_info_funcs hpd_rx_irq_info_funcs = { @@ -171,6 +144,16 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { .ack = NULL }; +static struct irq_source_info_funcs vline1_irq_info_funcs = { + .set = NULL, + .ack = NULL +}; + +static struct irq_source_info_funcs vline2_irq_info_funcs = { + .set = NULL, + .ack = NULL +}; + #undef BASE_INNER #define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg @@ -239,6 +222,13 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { .funcs = &pflip_irq_info_funcs\ } +#define vblank_int_entry(reg_num)\ + [DC_IRQ_SOURCE_VBLANK1 + reg_num] = {\ + IRQ_REG_ENTRY(OTG, reg_num,\ + OTG_GLOBAL_SYNC_STATUS, VSTARTUP_INT_EN,\ + OTG_GLOBAL_SYNC_STATUS, VSTARTUP_EVENT_CLEAR),\ + .funcs = &vblank_irq_info_funcs\ + } /* vupdate_no_lock_int_entry maps to DC_IRQ_SOURCE_VUPDATEx, to match semantic * of DCE's DC_IRQ_SOURCE_VUPDATEx. */ @@ -250,13 +240,6 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { .funcs = &vupdate_no_lock_irq_info_funcs\ } -#define vblank_int_entry(reg_num)\ - [DC_IRQ_SOURCE_VBLANK1 + reg_num] = {\ - IRQ_REG_ENTRY(OTG, reg_num,\ - OTG_GLOBAL_SYNC_STATUS, VSTARTUP_INT_EN,\ - OTG_GLOBAL_SYNC_STATUS, VSTARTUP_EVENT_CLEAR),\ - .funcs = &vblank_irq_info_funcs\ - } #define vline0_int_entry(reg_num)\ [DC_IRQ_SOURCE_DC1_VLINE0 + reg_num] = {\ IRQ_REG_ENTRY(OTG, reg_num,\ @@ -264,6 +247,20 @@ static struct irq_source_info_funcs vline0_irq_info_funcs = { OTG_VERTICAL_INTERRUPT0_CONTROL, OTG_VERTICAL_INTERRUPT0_CLEAR),\ .funcs = &vline0_irq_info_funcs\ } +#define vline1_int_entry(reg_num)\ + [DC_IRQ_SOURCE_DC1_VLINE1 + reg_num] = {\ + IRQ_REG_ENTRY(OTG, reg_num,\ + OTG_VERTICAL_INTERRUPT1_CONTROL, OTG_VERTICAL_INTERRUPT1_INT_ENABLE,\ + OTG_VERTICAL_INTERRUPT1_CONTROL, OTG_VERTICAL_INTERRUPT1_CLEAR),\ + .funcs = &vline1_irq_info_funcs\ + } +#define vline2_int_entry(reg_num)\ + [DC_IRQ_SOURCE_DC1_VLINE2 + reg_num] = {\ + IRQ_REG_ENTRY(OTG, reg_num,\ + OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_INT_ENABLE,\ + OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_CLEAR),\ + .funcs = &vline2_irq_info_funcs\ + } #define dmub_outbox_int_entry()\ [DC_IRQ_SOURCE_DMCUB_OUTBOX] = {\ IRQ_REG_ENTRY_DMUB(\ @@ -364,21 +361,29 @@ irq_source_info_dcn401[DAL_IRQ_SOURCES_NUMBER] = { dc_underflow_int_entry(6), [DC_IRQ_SOURCE_DMCU_SCP] = dummy_irq_entry(), [DC_IRQ_SOURCE_VBIOS_SW] = dummy_irq_entry(), - vupdate_no_lock_int_entry(0), - vupdate_no_lock_int_entry(1), - vupdate_no_lock_int_entry(2), - vupdate_no_lock_int_entry(3), vblank_int_entry(0), vblank_int_entry(1), vblank_int_entry(2), vblank_int_entry(3), + [DC_IRQ_SOURCE_DC5_VLINE1] = dummy_irq_entry(), + [DC_IRQ_SOURCE_DC6_VLINE1] = dummy_irq_entry(), + dmub_outbox_int_entry(), + vupdate_no_lock_int_entry(0), + vupdate_no_lock_int_entry(1), + vupdate_no_lock_int_entry(2), + vupdate_no_lock_int_entry(3), vline0_int_entry(0), vline0_int_entry(1), vline0_int_entry(2), vline0_int_entry(3), - [DC_IRQ_SOURCE_DC5_VLINE1] = dummy_irq_entry(), - [DC_IRQ_SOURCE_DC6_VLINE1] = dummy_irq_entry(), - dmub_outbox_int_entry(), + vline1_int_entry(0), + vline1_int_entry(1), + vline1_int_entry(2), + vline1_int_entry(3), + vline2_int_entry(0), + vline2_int_entry(1), + vline2_int_entry(2), + vline2_int_entry(3), }; static const struct irq_service_funcs irq_service_funcs_dcn401 = { diff --git a/drivers/gpu/drm/amd/display/dc/irq/irq_service.c b/drivers/gpu/drm/amd/display/dc/irq/irq_service.c index eca3d7ee7e4e..b595a11c5eaf 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/irq_service.c +++ b/drivers/gpu/drm/amd/display/dc/irq/irq_service.c @@ -41,6 +41,16 @@ #include "reg_helper.h" #include "irq_service.h" +//HPD0_DC_HPD_INT_STATUS +#define HPD0_DC_HPD_INT_STATUS__DC_HPD_SENSE_DELAYED_MASK 0x00000010L +#define HPD0_DC_HPD_INT_CONTROL__DC_HPD_INT_POLARITY_MASK 0x00000100L +#define HPD0_DC_HPD_INT_STATUS__DC_HPD_SENSE_DELAYED__SHIFT 0x4 +#define HPD0_DC_HPD_INT_CONTROL__DC_HPD_INT_POLARITY__SHIFT 0x8 +//HPD1_DC_HPD_INT_STATUS +#define DC_HPD1_INT_STATUS__DC_HPD1_SENSE_DELAYED_MASK 0x10 +#define DC_HPD1_INT_STATUS__DC_HPD1_SENSE_DELAYED__SHIFT 0x4 +#define DC_HPD1_INT_CONTROL__DC_HPD1_INT_POLARITY_MASK 0x100 +#define DC_HPD1_INT_CONTROL__DC_HPD1_INT_POLARITY__SHIFT 0x8 #define CTX \ @@ -177,3 +187,57 @@ enum dc_irq_source dal_irq_service_to_irq_source( src_id, ext_id); } + +bool hpd0_ack( + struct irq_service *irq_service, + const struct irq_source_info *info) +{ + uint32_t addr = info->status_reg; + uint32_t value = dm_read_reg(irq_service->ctx, addr); + uint32_t current_status = + get_reg_field_value( + value, + HPD0_DC_HPD_INT_STATUS, + DC_HPD_SENSE_DELAYED); + + dal_irq_service_ack_generic(irq_service, info); + + value = dm_read_reg(irq_service->ctx, info->enable_reg); + + set_reg_field_value( + value, + current_status ? 0 : 1, + HPD0_DC_HPD_INT_CONTROL, + DC_HPD_INT_POLARITY); + + dm_write_reg(irq_service->ctx, info->enable_reg, value); + + return true; +} + +bool hpd1_ack( + struct irq_service *irq_service, + const struct irq_source_info *info) +{ + uint32_t addr = info->status_reg; + uint32_t value = dm_read_reg(irq_service->ctx, addr); + uint32_t current_status = + get_reg_field_value( + value, + DC_HPD1_INT_STATUS, + DC_HPD1_SENSE_DELAYED); + + dal_irq_service_ack_generic(irq_service, info); + + value = dm_read_reg(irq_service->ctx, info->enable_reg); + + set_reg_field_value( + value, + current_status ? 0 : 1, + DC_HPD1_INT_CONTROL, + DC_HPD1_INT_POLARITY); + + dm_write_reg(irq_service->ctx, info->enable_reg, value); + + return true; +} diff --git a/drivers/gpu/drm/amd/display/dc/irq/irq_service.h b/drivers/gpu/drm/amd/display/dc/irq/irq_service.h index b178f85944cd..bbcef3d2fe33 100644 --- a/drivers/gpu/drm/amd/display/dc/irq/irq_service.h +++ b/drivers/gpu/drm/amd/display/dc/irq/irq_service.h @@ -82,4 +82,12 @@ void dal_irq_service_set_generic( const struct irq_source_info *info, bool enable); +bool hpd0_ack( + struct irq_service *irq_service, + const struct irq_source_info *info); + +bool hpd1_ack( + struct irq_service *irq_service, + const struct irq_source_info *info); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/irq_types.h b/drivers/gpu/drm/amd/display/dc/irq_types.h index 110f656d43ae..a2f7b933bebf 100644 --- a/drivers/gpu/drm/amd/display/dc/irq_types.h +++ b/drivers/gpu/drm/amd/display/dc/irq_types.h @@ -161,6 +161,20 @@ enum dc_irq_source { DC_IRQ_SOURCE_DPCX_TX_PHYE, DC_IRQ_SOURCE_DPCX_TX_PHYF, + DC_IRQ_SOURCE_DC1_VLINE2, + DC_IRQ_SOURCE_DC2_VLINE2, + DC_IRQ_SOURCE_DC3_VLINE2, + DC_IRQ_SOURCE_DC4_VLINE2, + DC_IRQ_SOURCE_DC5_VLINE2, + DC_IRQ_SOURCE_DC6_VLINE2, + + DC_IRQ_SOURCE_DCI2C_RR_DDC1, + DC_IRQ_SOURCE_DCI2C_RR_DDC2, + DC_IRQ_SOURCE_DCI2C_RR_DDC3, + DC_IRQ_SOURCE_DCI2C_RR_DDC4, + DC_IRQ_SOURCE_DCI2C_RR_DDC5, + DC_IRQ_SOURCE_DCI2C_RR_DDC6, + DAL_IRQ_SOURCES_NUMBER }; @@ -170,6 +184,8 @@ enum irq_type IRQ_TYPE_VUPDATE = DC_IRQ_SOURCE_VUPDATE1, IRQ_TYPE_VBLANK = DC_IRQ_SOURCE_VBLANK1, IRQ_TYPE_VLINE0 = DC_IRQ_SOURCE_DC1_VLINE0, + IRQ_TYPE_VLINE1 = DC_IRQ_SOURCE_DC1_VLINE1, + IRQ_TYPE_VLINE2 = DC_IRQ_SOURCE_DC1_VLINE2, IRQ_TYPE_DCUNDERFLOW = DC_IRQ_SOURCE_DC1UNDERFLOW, }; diff --git a/drivers/gpu/drm/amd/display/dc/link/link_detection.c b/drivers/gpu/drm/amd/display/dc/link/link_detection.c index cc9191a5c9e6..9655e6fa53a4 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_detection.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_detection.c @@ -611,6 +611,7 @@ static bool detect_dp(struct dc_link *link, link->dpcd_caps.dongle_type = sink_caps->dongle_type; link->dpcd_caps.is_dongle_type_one = sink_caps->is_dongle_type_one; link->dpcd_caps.dpcd_rev.raw = 0; + link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.raw = 0; } return true; @@ -1007,21 +1008,11 @@ static bool detect_link_and_local_sink(struct dc_link *link, link->reported_link_cap.link_rate > LINK_RATE_HIGH3) link->reported_link_cap.link_rate = LINK_RATE_HIGH3; - /* - * If this is DP over USB4 link then we need to: - * - Enable BW ALLOC support on DPtx if applicable - */ - if (dc->config.usb4_bw_alloc_support) { - if (link_dp_dpia_set_dptx_usb4_bw_alloc_support(link)) { - /* update with non reduced link cap if bw allocation mode is supported */ - if (link->dpia_bw_alloc_config.nrd_max_link_rate && - link->dpia_bw_alloc_config.nrd_max_lane_count) { - link->reported_link_cap.link_rate = - link->dpia_bw_alloc_config.nrd_max_link_rate; - link->reported_link_cap.lane_count = - link->dpia_bw_alloc_config.nrd_max_lane_count; - } - } + if (link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling + && link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dpia_bw_alloc + && link->dpcd_caps.usb4_dp_tun_info.driver_bw_cap.bits.driver_bw_alloc_support) { + if (link_dpia_enable_usb4_dp_bw_alloc_mode(link) == false) + link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dpia_bw_alloc = false; } break; } diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c index 268626e73c54..273a3be6d593 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c @@ -148,6 +148,7 @@ void link_blank_dp_stream(struct dc_link *link, bool hw_init) void link_set_all_streams_dpms_off_for_link(struct dc_link *link) { struct pipe_ctx *pipes[MAX_PIPES]; + struct dc_stream_state *streams[MAX_PIPES]; struct dc_state *state = link->dc->current_state; uint8_t count; int i; @@ -160,10 +161,18 @@ void link_set_all_streams_dpms_off_for_link(struct dc_link *link) link_get_master_pipes_with_dpms_on(link, state, &count, pipes); + /* The subsequent call to dc_commit_updates_for_stream for a full update + * will release the current state and swap to a new state. Releasing the + * current state results in the stream pointers in the pipe_ctx structs + * to be zero'd. Hence, cache all streams prior to dc_commit_updates_for_stream. + */ + for (i = 0; i < count; i++) + streams[i] = pipes[i]->stream; + for (i = 0; i < count; i++) { - stream_update.stream = pipes[i]->stream; + stream_update.stream = streams[i]; dc_commit_updates_for_stream(link->ctx->dc, NULL, 0, - pipes[i]->stream, &stream_update, + streams[i], &stream_update, state); } @@ -2365,7 +2374,7 @@ void link_set_dpms_off(struct pipe_ctx *pipe_ctx) update_psp_stream_config(pipe_ctx, true); dc->hwss.blank_stream(pipe_ctx); - if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) + if (pipe_ctx->link_config.dp_tunnel_settings.should_use_dp_bw_allocation) deallocate_usb4_bandwidth(pipe_ctx->stream); if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) @@ -2433,7 +2442,7 @@ void link_set_dpms_off(struct pipe_ctx *pipe_ctx) if (link->connector_signal == SIGNAL_TYPE_EDP && dc->debug.psp_disabled_wa) { /* reset internal save state to default since eDP is off */ enum dp_panel_mode panel_mode = dp_get_panel_mode(pipe_ctx->stream->link); - /* since current psp not loaded, we need to reset it to default*/ + /* since current psp not loaded, we need to reset it to default */ link->panel_mode = panel_mode; } } @@ -2611,7 +2620,7 @@ void link_set_dpms_on( if (dc_is_dp_signal(pipe_ctx->stream->signal)) dp_set_hblank_reduction_on_rx(pipe_ctx); - if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) + if (pipe_ctx->link_config.dp_tunnel_settings.should_use_dp_bw_allocation) allocate_usb4_bandwidth(pipe_ctx->stream); if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) diff --git a/drivers/gpu/drm/amd/display/dc/link/link_factory.c b/drivers/gpu/drm/amd/display/dc/link/link_factory.c index f6b6b19e7481..1a04f4b74585 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_factory.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_factory.c @@ -156,6 +156,7 @@ static void construct_link_service_dp_capability(struct link_service *link_srv) link_srv->dp_get_encoding_format = link_dp_get_encoding_format; link_srv->dp_should_enable_fec = dp_should_enable_fec; link_srv->dp_decide_link_settings = link_decide_link_settings; + link_srv->dp_decide_tunnel_settings = link_decide_dp_tunnel_settings; link_srv->mst_decide_link_encoding_format = mst_decide_link_encoding_format; link_srv->edp_decide_link_settings = edp_decide_link_settings; @@ -464,6 +465,7 @@ static bool construct_phy(struct dc_link *link, link->irq_source_hpd = DC_IRQ_SOURCE_INVALID; link->irq_source_hpd_rx = DC_IRQ_SOURCE_INVALID; + link->irq_source_read_request = DC_IRQ_SOURCE_INVALID; link->link_status.dpcd_caps = &link->dpcd_caps; link->dc = init_params->dc; @@ -514,6 +516,9 @@ static bool construct_phy(struct dc_link *link, case CONNECTOR_ID_HDMI_TYPE_A: link->connector_signal = SIGNAL_TYPE_HDMI_TYPE_A; + if (link->hpd_gpio) + link->irq_source_read_request = + dal_irq_get_read_request(link->hpd_gpio); break; case CONNECTOR_ID_SINGLE_LINK_DVID: case CONNECTOR_ID_SINGLE_LINK_DVII: @@ -653,7 +658,7 @@ static bool construct_phy(struct dc_link *link, } /* Look for device tag that matches connector signal, - * CRT for rgb, LCD for other supported signal tyes + * CRT for rgb, LCD for other supported signal types */ if (!bp_funcs->is_device_id_supported(dc_ctx->dc_bios, link->device_tag.dev_id)) diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c index 21ee0d96c9d4..8f79881ad9f1 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c @@ -158,6 +158,14 @@ uint8_t dp_parse_lttpr_repeater_count(uint8_t lttpr_repeater_count) return 0; // invalid value } +uint32_t dp_get_closest_lttpr_offset(uint8_t lttpr_count) +{ + /* Calculate offset for LTTPR closest to DPTX which is highest in the chain + * Offset is 0 for single LTTPR cases as base LTTPR DPCD addresses target LTTPR 1 + */ + return DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE * (lttpr_count - 1); +} + uint32_t link_bw_kbps_from_raw_frl_link_rate_data(uint8_t bw) { switch (bw) { @@ -2013,11 +2021,9 @@ static bool retrieve_link_cap(struct dc_link *link) sizeof(link->dpcd_caps.max_uncompressed_pixel_rate_cap.raw)); /* Read DP tunneling information. */ - if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) { - status = dpcd_get_tunneling_device_data(link); - if (status != DC_OK) - dm_error("%s: Read DP tunneling device data failed.\n", __func__); - } + status = dpcd_get_tunneling_device_data(link); + if (status != DC_OK) + dm_error("%s: Read DP tunneling device data failed.\n", __func__); retrieve_cable_id(link); dpcd_write_cable_id_to_dprx(link); diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.h b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.h index 0ce0af3ddbeb..940b147cc5d4 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.h +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.h @@ -48,6 +48,9 @@ enum dc_status dp_retrieve_lttpr_cap(struct dc_link *link); /* Convert PHY repeater count read from DPCD uint8_t. */ uint8_t dp_parse_lttpr_repeater_count(uint8_t lttpr_repeater_count); +/* Calculate embedded LTTPR address offset for vendor-specific behaviour */ +uint32_t dp_get_closest_lttpr_offset(uint8_t lttpr_count); + bool dp_is_sink_present(struct dc_link *link); bool dp_is_lttpr_present(struct dc_link *link); diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c index 0d123e647652..22bfdced64ab 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c @@ -62,6 +62,36 @@ enum dc_status dpcd_get_tunneling_device_data(struct dc_link *link) if (status != DC_OK) goto err; + link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.raw = + dpcd_dp_tun_data[DP_TUNNELING_CAPABILITIES_SUPPORT - DP_TUNNELING_CAPABILITIES_SUPPORT]; + + if (link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling == false) + goto err; + + link->dpcd_caps.usb4_dp_tun_info.dpia_info.raw = + dpcd_dp_tun_data[DP_IN_ADAPTER_INFO - DP_TUNNELING_CAPABILITIES_SUPPORT]; + link->dpcd_caps.usb4_dp_tun_info.usb4_driver_id = + dpcd_dp_tun_data[DP_USB4_DRIVER_ID - DP_TUNNELING_CAPABILITIES_SUPPORT]; + + if (link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dpia_bw_alloc) { + status = core_link_read_dpcd(link, USB4_DRIVER_BW_CAPABILITY, + dpcd_dp_tun_data, 1); + + if (status != DC_OK) + goto err; + + link->dpcd_caps.usb4_dp_tun_info.driver_bw_cap.raw = dpcd_dp_tun_data[0]; + } + + DC_LOG_DEBUG("%s: Link[%d] DP tunneling support (RouterId=%d AdapterId=%d) " + "DPIA_BW_Alloc_support=%d " + "CM_BW_Alloc_support=%d ", + __func__, link->link_index, + link->dpcd_caps.usb4_dp_tun_info.usb4_driver_id, + link->dpcd_caps.usb4_dp_tun_info.dpia_info.bits.dpia_num, + link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dpia_bw_alloc, + link->dpcd_caps.usb4_dp_tun_info.driver_bw_cap.bits.driver_bw_alloc_support); + status = core_link_read_dpcd( link, DP_USB4_ROUTER_TOPOLOGY_ID, @@ -71,13 +101,6 @@ enum dc_status dpcd_get_tunneling_device_data(struct dc_link *link) if (status != DC_OK) goto err; - link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.raw = - dpcd_dp_tun_data[DP_TUNNELING_CAPABILITIES_SUPPORT - DP_TUNNELING_CAPABILITIES_SUPPORT]; - link->dpcd_caps.usb4_dp_tun_info.dpia_info.raw = - dpcd_dp_tun_data[DP_IN_ADAPTER_INFO - DP_TUNNELING_CAPABILITIES_SUPPORT]; - link->dpcd_caps.usb4_dp_tun_info.usb4_driver_id = - dpcd_dp_tun_data[DP_USB4_DRIVER_ID - DP_TUNNELING_CAPABILITIES_SUPPORT]; - for (i = 0; i < DPCD_USB4_TOPOLOGY_ID_LEN; i++) link->dpcd_caps.usb4_dp_tun_info.usb4_topology_id[i] = dpcd_topology_data[i]; @@ -92,6 +115,7 @@ bool dpia_query_hpd_status(struct dc_link *link) /* prepare QUERY_HPD command */ cmd.query_hpd.header.type = DMUB_CMD__QUERY_HPD_STATE; + cmd.query_hpd.header.payload_bytes = sizeof(cmd.query_hpd.data); cmd.query_hpd.data.instance = link->link_id.enum_id - ENUM_ID_1; cmd.query_hpd.data.ch_type = AUX_CHANNEL_DPIA; @@ -119,3 +143,20 @@ bool dpia_query_hpd_status(struct dc_link *link) return link->hpd_status; } +void link_decide_dp_tunnel_settings(struct dc_stream_state *stream, + struct dc_tunnel_settings *dp_tunnel_setting) +{ + struct dc_link *link = stream->link; + + memset(dp_tunnel_setting, 0, sizeof(*dp_tunnel_setting)); + + if ((stream->signal == SIGNAL_TYPE_DISPLAY_PORT) || (stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)) { + dp_tunnel_setting->should_enable_dp_tunneling = + link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling; + + if (link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dpia_bw_alloc + && link->dpcd_caps.usb4_dp_tun_info.driver_bw_cap.bits.driver_bw_alloc_support) + dp_tunnel_setting->should_use_dp_bw_allocation = true; + } +} + diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.h b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.h index 363f45a1a964..a61edfc9ca7a 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.h +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.h @@ -38,4 +38,10 @@ enum dc_status dpcd_get_tunneling_device_data(struct dc_link *link); * Returns true if HPD high. */ bool dpia_query_hpd_status(struct dc_link *link); + +/* Decide the DP tunneling settings based on the DPCD capabilities + */ +void link_decide_dp_tunnel_settings(struct dc_stream_state *stream, + struct dc_tunnel_settings *dp_tunnel_setting); + #endif /* __DC_LINK_DPIA_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c index a254ead2f7e8..3af7564a84f1 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c @@ -46,9 +46,10 @@ */ static bool link_dp_is_bw_alloc_available(struct dc_link *link) { - return (link && link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA - && link->hpd_status - && link->dpia_bw_alloc_config.bw_alloc_enabled); + return (link && link->hpd_status + && link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling + && link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dpia_bw_alloc + && link->dpcd_caps.usb4_dp_tun_info.driver_bw_cap.bits.driver_bw_alloc_support); } static void reset_bw_alloc_struct(struct dc_link *link) @@ -141,7 +142,7 @@ static int get_non_reduced_max_lane_count(struct dc_link *link) * granuality, Driver_ID, CM_Group, & populate the BW allocation structs * for host router and dpia */ -static void init_usb4_bw_struct(struct dc_link *link) +static void retrieve_usb4_dp_bw_allocation_info(struct dc_link *link) { reset_bw_alloc_struct(link); @@ -282,49 +283,26 @@ static void link_dpia_send_bw_alloc_request(struct dc_link *link, int req_bw) // ------------------------------------------------------------------ // PUBLIC FUNCTIONS // ------------------------------------------------------------------ -bool link_dp_dpia_set_dptx_usb4_bw_alloc_support(struct dc_link *link) +bool link_dpia_enable_usb4_dp_bw_alloc_mode(struct dc_link *link) { bool ret = false; - uint8_t response = 0, - bw_support_dpia = 0, - bw_support_cm = 0; + uint8_t val; - if (!(link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->hpd_status)) - goto out; + if (link->hpd_status) { + val = DPTX_BW_ALLOC_MODE_ENABLE | DPTX_BW_ALLOC_UNMASK_IRQ; - if (core_link_read_dpcd( - link, - DP_TUNNELING_CAPABILITIES, - &response, - sizeof(uint8_t)) == DC_OK) - bw_support_dpia = (response >> 7) & 1; - - if (core_link_read_dpcd( - link, - USB4_DRIVER_BW_CAPABILITY, - &response, - sizeof(uint8_t)) == DC_OK) - bw_support_cm = (response >> 7) & 1; - - /* Send request acknowledgment to Turn ON DPTX support */ - if (bw_support_cm && bw_support_dpia) { - - response = 0x80; - if (core_link_write_dpcd( - link, - DPTX_BW_ALLOCATION_MODE_CONTROL, - &response, - sizeof(uint8_t)) != DC_OK) { - DC_LOG_DEBUG("%s: FAILURE Enabling DPtx BW Allocation Mode Support for link(%d)\n", - __func__, link->link_index); - } else { - // SUCCESS Enabled DPtx BW Allocation Mode Support - DC_LOG_DEBUG("%s: SUCCESS Enabling DPtx BW Allocation Mode Support for link(%d)\n", - __func__, link->link_index); + if (core_link_write_dpcd(link, DPTX_BW_ALLOCATION_MODE_CONTROL, &val, sizeof(uint8_t)) == DC_OK) { + DC_LOG_DEBUG("%s: link[%d] DPTX BW allocation mode enabled", __func__, link->link_index); + + retrieve_usb4_dp_bw_allocation_info(link); + + if (link->dpia_bw_alloc_config.nrd_max_link_rate && link->dpia_bw_alloc_config.nrd_max_lane_count) { + link->reported_link_cap.link_rate = link->dpia_bw_alloc_config.nrd_max_link_rate; + link->reported_link_cap.lane_count = link->dpia_bw_alloc_config.nrd_max_lane_count; + } - ret = true; - init_usb4_bw_struct(link); link->dpia_bw_alloc_config.bw_alloc_enabled = true; + ret = true; /* * During DP tunnel creation, CM preallocates BW and reduces estimated BW of other @@ -332,11 +310,12 @@ bool link_dp_dpia_set_dptx_usb4_bw_alloc_support(struct dc_link *link) * to make the CM to release preallocation and update estimated BW correctly for * all DPIAs per host router */ + // TODO: Zero allocation can be removed once the MSFT CM fix has been released link_dp_dpia_allocate_usb4_bandwidth_for_stream(link, 0); - } + } else + DC_LOG_DEBUG("%s: link[%d] failed to enable DPTX BW allocation mode", __func__, link->link_index); } -out: return ret; } @@ -378,7 +357,8 @@ void link_dp_dpia_handle_bw_alloc_status(struct dc_link *link, uint8_t status) */ void dpia_handle_usb4_bandwidth_allocation_for_link(struct dc_link *link, int peak_bw) { - if (link && link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->dpia_bw_alloc_config.bw_alloc_enabled) { + if (link && link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling + && link->dpia_bw_alloc_config.bw_alloc_enabled) { //1. Hot Plug if (link->hpd_status && peak_bw > 0) { // If DP over USB4 then we need to check BW allocation @@ -401,7 +381,7 @@ void link_dp_dpia_allocate_usb4_bandwidth_for_stream(struct dc_link *link, int r if (link_dp_is_bw_alloc_available(link)) link_dpia_send_bw_alloc_request(link, req_bw); else - DC_LOG_DEBUG("%s: Not able to send the BW Allocation request", __func__); + DC_LOG_DEBUG("%s: BW Allocation mode not available", __func__); } bool dpia_validate_usb4_bw(struct dc_link **link, int *bw_needed_per_dpia, const unsigned int num_dpias) diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h index 6df9b946b00f..801965b5f9a4 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h @@ -43,13 +43,13 @@ enum bw_type { }; /* - * Enable BW Allocation Mode Support from the DP-Tx side + * Enable USB4 DP BW allocation mode * * @link: pointer to the dc_link struct instance * * return: SUCCESS or FAILURE */ -bool link_dp_dpia_set_dptx_usb4_bw_alloc_support(struct dc_link *link); +bool link_dpia_enable_usb4_dp_bw_alloc_mode(struct dc_link *link); /* * Allocates only what the stream needs for bw, so if: diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c index 5be00e4ce10b..693477413347 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c @@ -229,6 +229,10 @@ static void handle_hpd_irq_replay_sink(struct dc_link *link) link->replay_settings.config.replay_error_status.raw |= replay_error_status.raw; + /* Increment desync error counter if a desync error is detected */ + if (replay_configuration.bits.DESYNC_ERROR_STATUS) + link->replay_settings.replay_desync_error_fail_count++; + if (link->replay_settings.config.force_disable_desync_error_check) return; @@ -240,9 +244,6 @@ static void handle_hpd_irq_replay_sink(struct dc_link *link) &replay_configuration.raw, sizeof(replay_configuration.raw)); - /* Update desync error counter */ - link->replay_settings.replay_desync_error_fail_count++; - /* Acknowledge and clear error bits */ dm_helpers_dp_write_dpcd( link->ctx, @@ -351,7 +352,7 @@ enum dc_status dp_read_hpd_rx_irq_data( irq_data->raw, DP_SINK_STATUS - DP_SINK_COUNT + 1); - if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) { + if (link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling) { retval = core_link_read_dpcd( link, DP_LINK_SERVICE_IRQ_VECTOR_ESI0, &irq_data->bytes.link_service_irq_esi0.raw, 1); @@ -520,7 +521,7 @@ bool dp_handle_hpd_rx_irq(struct dc_link *link, dp_trace_link_loss_increment(link); } - if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) { + if (link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.bits.dp_tunneling) { if (hpd_irq_dpcd_data.bytes.link_service_irq_esi0.bits.DP_LINK_TUNNELING_IRQ) dp_handle_tunneling_irq(link); } diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c index ef358afdfb65..2dc1a660e504 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c @@ -785,7 +785,6 @@ void override_training_settings( lt_settings->lttpr_mode = LTTPR_MODE_NON_LTTPR; dp_get_lttpr_mode_override(link, <_settings->lttpr_mode); - } enum dc_dp_training_pattern decide_cr_training_pattern( diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c index 5a5d48fadbf2..66d0fb1b9b9d 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c @@ -142,6 +142,14 @@ void decide_8b_10b_training_settings( lt_settings->lttpr_mode = dp_decide_8b_10b_lttpr_mode(link); lt_settings->cr_pattern_time = get_cr_training_aux_rd_interval(link, link_setting, lt_settings->lttpr_mode); dp_hw_to_dpcd_lane_settings(lt_settings, lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings); + + /* Some embedded LTTPRs rely on receiving TPS2 before LT to interop reliably with sensitive VGA dongles + * This allows these LTTPRs to minimize freq/phase and skew variation during lock and deskew sequences + */ + if ((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == + AMD_EXT_DISPLAY_PATH_CAPS__DP_EARLY_8B10B_TPS2) { + lt_settings->lttpr_early_tps2 = true; + } } enum lttpr_mode dp_decide_8b_10b_lttpr_mode(struct dc_link *link) @@ -173,6 +181,42 @@ enum lttpr_mode dp_decide_8b_10b_lttpr_mode(struct dc_link *link) return LTTPR_MODE_NON_LTTPR; } +static void set_link_settings_and_perform_early_tps2_retimer_pre_lt_sequence(struct dc_link *link, + const struct link_resource *link_res, + struct link_training_settings *lt_settings, + uint32_t lttpr_count) +{ + /* Vendor-specific LTTPR early TPS2 sequence: + * 1. Output TPS2 + * 2. Wait 400us + * 3. Set link settings as usual + * 4. Write TPS1 to DP_TRAINING_PATTERN_SET_PHY_REPEATERx targeting LTTPR closest to host + * 5. Wait 1ms + * 6. Begin link training as usual + * */ + + uint32_t closest_lttpr_address_offset = dp_get_closest_lttpr_offset(lttpr_count); + + union dpcd_training_pattern dpcd_pattern = {0}; + + dpcd_pattern.v1_4.TRAINING_PATTERN_SET = 1; + dpcd_pattern.v1_4.SCRAMBLING_DISABLE = 1; + + DC_LOG_HW_LINK_TRAINING("%s\n GPU sends TPS2. Wait 400us.\n", __func__); + + dp_set_hw_training_pattern(link, link_res, DP_TRAINING_PATTERN_SEQUENCE_2, DPRX); + + dp_set_hw_lane_settings(link, link_res, lt_settings, DPRX); + + udelay(400); + + dpcd_set_link_settings(link, lt_settings); + + core_link_write_dpcd(link, DP_TRAINING_PATTERN_SET_PHY_REPEATER1 + closest_lttpr_address_offset, &dpcd_pattern.raw, 1); + + udelay(1000); + } + enum link_training_result perform_8b_10b_clock_recovery_sequence( struct dc_link *link, const struct link_resource *link_res, @@ -383,7 +427,7 @@ enum link_training_result dp_perform_8b_10b_link_training( { enum link_training_result status = LINK_TRAINING_SUCCESS; - uint8_t repeater_cnt; + uint8_t repeater_cnt = dp_parse_lttpr_repeater_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt); uint8_t repeater_id; uint8_t lane = 0; @@ -391,14 +435,16 @@ enum link_training_result dp_perform_8b_10b_link_training( start_clock_recovery_pattern_early(link, link_res, lt_settings, DPRX); /* 1. set link rate, lane count and spread. */ - dpcd_set_link_settings(link, lt_settings); + if (lt_settings->lttpr_early_tps2) + set_link_settings_and_perform_early_tps2_retimer_pre_lt_sequence(link, link_res, lt_settings, repeater_cnt); + else + dpcd_set_link_settings(link, lt_settings); if (lt_settings->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) { /* 2. perform link training (set link training done * to false is done as well) */ - repeater_cnt = dp_parse_lttpr_repeater_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt); for (repeater_id = repeater_cnt; (repeater_id > 0 && status == LINK_TRAINING_SUCCESS); repeater_id--) { diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c index 1e4adbc764ea..da74c2b5854f 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c @@ -524,7 +524,7 @@ bool edp_set_backlight_level(const struct dc_link *link, struct dc *dc = link->ctx->dc; uint32_t backlight_pwm_u16_16 = backlight_level_params->backlight_pwm_u16_16; uint32_t frame_ramp = backlight_level_params->frame_ramp; - DC_LOGGER_INIT(link->ctx->logger); + DC_LOG_BACKLIGHT("New Backlight level: %d (0x%X)\n", backlight_pwm_u16_16, backlight_pwm_u16_16); @@ -1022,6 +1022,9 @@ bool edp_setup_replay(struct dc_link *link, const struct dc_stream_state *stream &alpm_config.raw, sizeof(alpm_config.raw)); } + + link->replay_settings.config.replay_video_conferencing_optimization_enabled = false; + return true; } @@ -1130,11 +1133,11 @@ static struct abm *get_abm_from_stream_res(const struct dc_link *link) struct abm *abm = NULL; for (i = 0; i < MAX_PIPES; i++) { - struct pipe_ctx pipe_ctx = dc->current_state->res_ctx.pipe_ctx[i]; - struct dc_stream_state *stream = pipe_ctx.stream; + struct pipe_ctx *pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i]; + struct dc_stream_state *stream = pipe_ctx->stream; if (stream && stream->link == link) { - abm = pipe_ctx.stream_res.abm; + abm = pipe_ctx->stream_res.abm; break; } } diff --git a/drivers/gpu/drm/amd/display/dc/mpc/dcn32/dcn32_mpc.c b/drivers/gpu/drm/amd/display/dc/mpc/dcn32/dcn32_mpc.c index a0e9e9f0441a..b4cea2b8cb2a 100644 --- a/drivers/gpu/drm/amd/display/dc/mpc/dcn32/dcn32_mpc.c +++ b/drivers/gpu/drm/amd/display/dc/mpc/dcn32/dcn32_mpc.c @@ -370,275 +370,279 @@ void mpc32_program_shaper_luta_settings( MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].red.custom_float_y); curve = params->arr_curve_points; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_0_1[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_2_3[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_4_5[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_6_7[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_8_9[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_10_11[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_12_13[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_14_15[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_16_17[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_18_19[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_20_21[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_22_23[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_24_25[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_26_27[mpcc_id], 0, + if (curve) { + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_0_1[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_28_29[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_30_31[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_32_33[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); -} - - -void mpc32_program_shaper_lutb_settings( - struct mpc *mpc, - const struct pwl_params *params, - uint32_t mpcc_id) -{ - const struct gamma_curve *curve; - struct dcn30_mpc *mpc30 = TO_DCN30_MPC(mpc); - - REG_SET_2(MPCC_MCM_SHAPER_RAMB_START_CNTL_B[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_B, params->corner_points[0].blue.custom_float_x, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_SEGMENT_B, 0); - REG_SET_2(MPCC_MCM_SHAPER_RAMB_START_CNTL_G[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_B, params->corner_points[0].green.custom_float_x, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_SEGMENT_B, 0); - REG_SET_2(MPCC_MCM_SHAPER_RAMB_START_CNTL_R[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_B, params->corner_points[0].red.custom_float_x, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_SEGMENT_B, 0); - - REG_SET_2(MPCC_MCM_SHAPER_RAMB_END_CNTL_B[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_B, params->corner_points[1].blue.custom_float_x, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].blue.custom_float_y); - REG_SET_2(MPCC_MCM_SHAPER_RAMB_END_CNTL_G[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_B, params->corner_points[1].green.custom_float_x, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].green.custom_float_y); - REG_SET_2(MPCC_MCM_SHAPER_RAMB_END_CNTL_R[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_B, params->corner_points[1].red.custom_float_x, - MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].red.custom_float_y); - - curve = params->arr_curve_points; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_0_1[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_2_3[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_2_3[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_4_5[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_4_5[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_6_7[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_6_7[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_8_9[mpcc_id], 0, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, - MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_8_9[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_10_11[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_10_11[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_12_13[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_12_13[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_14_15[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_14_15[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_16_17[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_16_17[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_18_19[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_18_19[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_20_21[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_20_21[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_22_23[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_22_23[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_24_25[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_24_25[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_26_27[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_26_27[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_28_29[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_28_29[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_30_31[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_30_31[mpcc_id], 0, + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMA_REGION_32_33[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + } +} + + +void mpc32_program_shaper_lutb_settings( + struct mpc *mpc, + const struct pwl_params *params, + uint32_t mpcc_id) +{ + const struct gamma_curve *curve; + struct dcn30_mpc *mpc30 = TO_DCN30_MPC(mpc); + + REG_SET_2(MPCC_MCM_SHAPER_RAMB_START_CNTL_B[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_B, params->corner_points[0].blue.custom_float_x, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_SEGMENT_B, 0); + REG_SET_2(MPCC_MCM_SHAPER_RAMB_START_CNTL_G[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_B, params->corner_points[0].green.custom_float_x, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_SEGMENT_B, 0); + REG_SET_2(MPCC_MCM_SHAPER_RAMB_START_CNTL_R[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_B, params->corner_points[0].red.custom_float_x, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_START_SEGMENT_B, 0); - curve += 2; - REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_32_33[mpcc_id], 0, + REG_SET_2(MPCC_MCM_SHAPER_RAMB_END_CNTL_B[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_B, params->corner_points[1].blue.custom_float_x, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].blue.custom_float_y); + REG_SET_2(MPCC_MCM_SHAPER_RAMB_END_CNTL_G[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_B, params->corner_points[1].green.custom_float_x, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].green.custom_float_y); + REG_SET_2(MPCC_MCM_SHAPER_RAMB_END_CNTL_R[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_B, params->corner_points[1].red.custom_float_x, + MPCC_MCM_SHAPER_RAMA_EXP_REGION_END_BASE_B, params->corner_points[1].red.custom_float_y); + + curve = params->arr_curve_points; + if (curve) { + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_0_1[mpcc_id], 0, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_2_3[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_4_5[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_6_7[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_8_9[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_10_11[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_12_13[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_14_15[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_16_17[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_18_19[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_20_21[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_22_23[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_24_25[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_26_27[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_28_29[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_30_31[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + + curve += 2; + REG_SET_4(MPCC_MCM_SHAPER_RAMB_REGION_32_33[mpcc_id], 0, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_LUT_OFFSET, curve[0].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION0_NUM_SEGMENTS, curve[0].segments_num, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_LUT_OFFSET, curve[1].offset, + MPCC_MCM_SHAPER_RAMA_EXP_REGION1_NUM_SEGMENTS, curve[1].segments_num); + } } diff --git a/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.c b/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.c index ad67197557ca..98cf0cbd59ba 100644 --- a/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.c +++ b/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.c @@ -47,16 +47,6 @@ void mpc401_update_3dlut_fast_load_select(struct mpc *mpc, int mpcc_id, int hubp REG_SET(MPCC_MCM_3DLUT_FAST_LOAD_SELECT[mpcc_id], 0, MPCC_MCM_3DLUT_FL_SEL, hubp_idx); } -void mpc401_get_3dlut_fast_load_status(struct mpc *mpc, int mpcc_id, uint32_t *done, uint32_t *soft_underflow, uint32_t *hard_underflow) -{ - struct dcn401_mpc *mpc401 = TO_DCN401_MPC(mpc); - - REG_GET_3(MPCC_MCM_3DLUT_FAST_LOAD_STATUS[mpcc_id], - MPCC_MCM_3DLUT_FL_DONE, done, - MPCC_MCM_3DLUT_FL_SOFT_UNDERFLOW, soft_underflow, - MPCC_MCM_3DLUT_FL_HARD_UNDERFLOW, hard_underflow); -} - void mpc401_set_movable_cm_location(struct mpc *mpc, enum mpcc_movable_cm_location location, int mpcc_id) { struct dcn401_mpc *mpc401 = TO_DCN401_MPC(mpc); @@ -618,7 +608,6 @@ static const struct mpc_funcs dcn401_mpc_funcs = { .set_bg_color = mpc1_set_bg_color, .set_movable_cm_location = mpc401_set_movable_cm_location, .update_3dlut_fast_load_select = mpc401_update_3dlut_fast_load_select, - .get_3dlut_fast_load_status = mpc401_get_3dlut_fast_load_status, .populate_lut = mpc401_populate_lut, .program_lut_read_write_control = mpc401_program_lut_read_write_control, .program_lut_mode = mpc401_program_lut_mode, diff --git a/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.h b/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.h index ce6fbcf14d7a..8e35ebc603a9 100644 --- a/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.h +++ b/drivers/gpu/drm/amd/display/dc/mpc/dcn401/dcn401_mpc.h @@ -241,23 +241,9 @@ void mpc401_update_3dlut_fast_load_select( int mpcc_id, int hubp_idx); -void mpc401_get_3dlut_fast_load_status( - struct mpc *mpc, - int mpcc_id, - uint32_t *done, - uint32_t *soft_underflow, - uint32_t *hard_underflow); - void mpc401_update_3dlut_fast_load_select( struct mpc *mpc, int mpcc_id, int hubp_idx); -void mpc401_get_3dlut_fast_load_status( - struct mpc *mpc, - int mpcc_id, - uint32_t *done, - uint32_t *soft_underflow, - uint32_t *hard_underflow); - #endif diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c b/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c index b86fe2b094f8..4cfc6c0fa147 100644 --- a/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c +++ b/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c @@ -507,6 +507,7 @@ void dcn35_timing_generator_init(struct optc *optc1) optc1->min_v_blank_interlace = 5; optc1->min_h_sync_width = 4; optc1->min_v_sync_width = 1; + optc1->max_frame_count = 0xFFFFFF; dcn35_timing_generator_set_fgcg( optc1, CTX->dc->debug.enable_fine_grain_clock_gating.bits.optc); diff --git a/drivers/gpu/drm/amd/display/dc/resource/Makefile b/drivers/gpu/drm/amd/display/dc/resource/Makefile index b8cddef6b3d2..5b42da8b79c2 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/Makefile +++ b/drivers/gpu/drm/amd/display/dc/resource/Makefile @@ -27,6 +27,24 @@ # DCE ############################################################################### +ifdef CONFIG_DRM_AMD_DC_SI +RESOURCE_DCE60 = dce60_resource.o + +AMD_DAL_RESOURCE_DCE60 = $(addprefix $(AMDDALPATH)/dc/resource/dce60/,$(RESOURCE_DCE60)) + +AMD_DISPLAY_FILES += $(AMD_DAL_RESOURCE_DCE60) +endif + +############################################################################### + +RESOURCE_DCE80 = dce80_resource.o + +AMD_DAL_RESOURCE_DCE80 = $(addprefix $(AMDDALPATH)/dc/resource/dce80/,$(RESOURCE_DCE80)) + +AMD_DISPLAY_FILES += $(AMD_DAL_RESOURCE_DCE80) + +############################################################################### + RESOURCE_DCE100 = dce100_resource.o AMD_DAL_RESOURCE_DCE100 = $(addprefix $(AMDDALPATH)/dc/resource/dce100/,$(RESOURCE_DCE100)) @@ -57,14 +75,6 @@ AMD_DAL_RESOURCE_DCE120 = $(addprefix $(AMDDALPATH)/dc/resource/dce120/,$(RESOUR AMD_DISPLAY_FILES += $(AMD_DAL_RESOURCE_DCE120) -############################################################################### - -RESOURCE_DCE80 = dce80_resource.o - -AMD_DAL_RESOURCE_DCE80 = $(addprefix $(AMDDALPATH)/dc/resource/dce80/,$(RESOURCE_DCE80)) - -AMD_DISPLAY_FILES += $(AMD_DAL_RESOURCE_DCE80) - ifdef CONFIG_DRM_AMD_DC_FP ############################################################################### # DCN diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce100/dce100_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce100/dce100_resource.c index e698543ec937..84f73fdb0f95 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dce100/dce100_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dce100/dce100_resource.c @@ -836,7 +836,7 @@ static enum dc_status build_mapped_resource( return DC_OK; } -static bool dce100_validate_bandwidth( +static enum dc_status dce100_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate) @@ -858,7 +858,7 @@ static bool dce100_validate_bandwidth( context->bw_ctx.bw.dce.yclk_khz = 0; } - return true; + return DC_OK; } static bool dce100_validate_surface_sets( @@ -1069,7 +1069,7 @@ static bool dce100_resource_construct( pool->base.timing_generator_count = pool->base.res_cap->num_timing_generator; dc->caps.max_downscale_ratio = 200; dc->caps.i2c_speed_in_khz = 40; - dc->caps.i2c_speed_in_khz = 40; + dc->caps.i2c_speed_in_khz_hdcp = 40; dc->caps.max_cursor_size = 128; dc->caps.min_horizontal_blanking_period = 80; dc->caps.dual_link_dvi = true; diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce110/dce110_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce110/dce110_resource.c index 035c6cfdaee5..f3d5baac11bf 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dce110/dce110_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dce110/dce110_resource.c @@ -960,7 +960,7 @@ static enum dc_status build_mapped_resource( return DC_OK; } -static bool dce110_validate_bandwidth( +static enum dc_status dce110_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate) @@ -1031,7 +1031,7 @@ static bool dce110_validate_bandwidth( context->bw_ctx.bw.dce.yclk_khz, context->bw_ctx.bw.dce.blackout_recovery_time_us); } - return result; + return result ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } static enum dc_status dce110_validate_plane(const struct dc_plane_state *plane_state, diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.c index 480a50967385..4225cae68c10 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.c @@ -883,7 +883,7 @@ static enum dc_status build_mapped_resource( return DC_OK; } -bool dce112_validate_bandwidth( +enum dc_status dce112_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate) @@ -952,7 +952,7 @@ bool dce112_validate_bandwidth( context->bw_ctx.bw.dce.yclk_khz, context->bw_ctx.bw.dce.blackout_recovery_time_us); } - return result; + return result ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } enum dc_status resource_map_phy_clock_resources( diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.h index 1f57ebc6f9b4..6221d749246d 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dce112/dce112_resource.h @@ -42,7 +42,7 @@ enum dc_status dce112_validate_with_context( struct dc_state *context, struct dc_state *old_context); -bool dce112_validate_bandwidth( +enum dc_status dce112_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate); diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce60/dce60_resource.c index 889f314cac65..d9ffdded5ce1 100644 --- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dce60/dce60_resource.c @@ -48,7 +48,7 @@ #include "dce/dce_clock_source.h" #include "dce/dce_audio.h" #include "dce/dce_hwseq.h" -#include "dce60/dce60_hw_sequencer.h" +#include "dce60/dce60_hwseq.h" #include "dce100/dce100_resource.h" #include "dce/dce_panel_cntl.h" @@ -863,7 +863,7 @@ static void dce60_resource_destruct(struct dce110_resource_pool *pool) } } -static bool dce60_validate_bandwidth( +static enum dc_status dce60_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate) @@ -885,7 +885,7 @@ static bool dce60_validate_bandwidth( context->bw_ctx.bw.dce.yclk_khz = 0; } - return true; + return DC_OK; } static bool dce60_validate_surface_sets( diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dce60/dce60_resource.h index 5d653a76b0b0..5d653a76b0b0 100644 --- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dce60/dce60_resource.h diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce80/dce80_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce80/dce80_resource.c index 3d5113f010bb..bd5811f97531 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dce80/dce80_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dce80/dce80_resource.c @@ -869,7 +869,7 @@ static void dce80_resource_destruct(struct dce110_resource_pool *pool) } } -static bool dce80_validate_bandwidth( +static enum dc_status dce80_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate) @@ -891,7 +891,7 @@ static bool dce80_validate_bandwidth( context->bw_ctx.bw.dce.yclk_khz = 0; } - return true; + return DC_OK; } static bool dce80_validate_surface_sets( diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c index e92f14d50adb..be4ade0853e9 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c @@ -23,6 +23,7 @@ * */ +#include "core_status.h" #include "dm_services.h" #include "dc.h" @@ -1125,7 +1126,7 @@ static void dcn10_destroy_resource_pool(struct resource_pool **pool) *pool = NULL; } -static bool dcn10_validate_bandwidth( +static enum dc_status dcn10_validate_bandwidth( struct dc *dc, struct dc_state *context, bool fast_validate) @@ -1136,7 +1137,7 @@ static bool dcn10_validate_bandwidth( voltage_supported = dcn_validate_bandwidth(dc, context, fast_validate); DC_FP_END(); - return voltage_supported; + return voltage_supported ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } static enum dc_status dcn10_validate_plane(const struct dc_plane_state *plane_state, struct dc_caps *caps) @@ -1245,6 +1246,10 @@ struct stream_encoder *dcn10_find_first_free_match_stream_enc_for_link( if (link->ep_type == DISPLAY_ENDPOINT_PHY && pool->stream_enc[i]->id == link->link_enc->preferred_engine) return pool->stream_enc[i]; + + if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && pool->stream_enc[i]->id == + link->dpia_preferred_eng_id) + return pool->stream_enc[i]; } } diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c index e4eca3e32c1b..3405be07f5e3 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c @@ -2124,7 +2124,7 @@ validate_out: return out; } -bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, +enum dc_status dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { bool voltage_supported; @@ -2132,14 +2132,14 @@ bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, pipes = kcalloc(dc->res_pool->pipe_count, sizeof(display_e2e_pipe_params_st), GFP_KERNEL); if (!pipes) - return false; + return DC_FAIL_BANDWIDTH_VALIDATE; DC_FP_START(); voltage_supported = dcn20_validate_bandwidth_fp(dc, context, fast_validate, pipes); DC_FP_END(); kfree(pipes); - return voltage_supported; + return voltage_supported ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } struct pipe_ctx *dcn20_acquire_free_pipe_for_layer( diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.h index 4cee3fa11a7f..c0e062c7407d 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.h @@ -119,7 +119,7 @@ void dcn20_set_mcif_arb_params( struct dc_state *context, display_e2e_pipe_params_st *pipes, int pipe_cnt); -bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); +enum dc_status dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); void dcn20_merge_pipes_for_validate( struct dc *dc, struct dc_state *context); diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c index 4bd5c2278596..9ab01b65b177 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c @@ -923,7 +923,7 @@ validate_out: * with DC_FP_START()/DC_FP_END(). Use the same approach as for * dcn20_validate_bandwidth in dcn20_resource.c. */ -static bool dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context, +static enum dc_status dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { bool voltage_supported; @@ -931,14 +931,14 @@ static bool dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context, pipes = kcalloc(dc->res_pool->pipe_count, sizeof(display_e2e_pipe_params_st), GFP_KERNEL); if (!pipes) - return false; + return DC_FAIL_BANDWIDTH_VALIDATE; DC_FP_START(); voltage_supported = dcn21_validate_bandwidth_fp(dc, context, fast_validate, pipes); DC_FP_END(); kfree(pipes); - return voltage_supported; + return voltage_supported ? DC_OK : DC_NOT_SUPPORTED; } static void dcn21_destroy_resource_pool(struct resource_pool **pool) diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c index f01ced015072..f631ae34e320 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c @@ -1891,8 +1891,6 @@ static int get_refresh_rate(struct dc_state *context) /* check if refresh rate at least 120hz */ timing = &context->streams[0]->timing; - if (timing == NULL) - return 0; h_v_total = timing->h_total * timing->v_total; if (h_v_total == 0) @@ -2037,7 +2035,7 @@ void dcn30_calculate_wm_and_dlg( DC_FP_END(); } -bool dcn30_validate_bandwidth(struct dc *dc, +enum dc_status dcn30_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { @@ -2094,7 +2092,7 @@ validate_out: BW_VAL_TRACE_FINISH(); - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } void dcn30_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params) diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.h index 8e6b8b7368fd..689d9bdace81 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.h @@ -56,7 +56,7 @@ unsigned int dcn30_calc_max_scaled_time( enum mmhubbub_wbif_mode mode, unsigned int urgent_watermark); -bool dcn30_validate_bandwidth(struct dc *dc, struct dc_state *context, +enum dc_status dcn30_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); bool dcn30_internal_validate_bw( struct dc *dc, diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c index dddddbfef85f..7e0af5297dc4 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c @@ -1758,7 +1758,7 @@ dcn31_set_mcif_arb_params(struct dc *dc, DC_FP_END(); } -bool dcn31_validate_bandwidth(struct dc *dc, +enum dc_status dcn31_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { @@ -1813,7 +1813,7 @@ validate_out: BW_VAL_TRACE_FINISH(); - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } static void dcn31_get_panel_config_defaults(struct dc_panel_config *panel_config) diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h index 551ad912f7be..dd82815d7efe 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h @@ -37,7 +37,7 @@ struct dcn31_resource_pool { struct resource_pool base; }; -bool dcn31_validate_bandwidth(struct dc *dc, +enum dc_status dcn31_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); void dcn31_calculate_wm_and_dlg( diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c index 26becc4cb804..d96bc6cb73ad 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c @@ -1694,7 +1694,7 @@ static void dcn314_get_panel_config_defaults(struct dc_panel_config *panel_confi *panel_config = panel_config_defaults; } -bool dcn314_validate_bandwidth(struct dc *dc, +enum dc_status dcn314_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { @@ -1750,7 +1750,7 @@ validate_out: BW_VAL_TRACE_FINISH(); - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } static struct resource_funcs dcn314_res_pool_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.h index 49ffe71018df..f8ba531d6342 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.h @@ -39,7 +39,7 @@ struct dcn314_resource_pool { struct resource_pool base; }; -bool dcn314_validate_bandwidth(struct dc *dc, +enum dc_status dcn314_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c index 2a59cc61ed8c..bb0dae0be5b8 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c @@ -24,6 +24,7 @@ * */ +#include "dc_types.h" #include "dm_services.h" #include "dc.h" @@ -1806,19 +1807,56 @@ validate_out: return out; } -bool dcn32_validate_bandwidth(struct dc *dc, +enum dc_status dcn32_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { - bool out = false; + unsigned int i; + enum dc_status status; + const struct dc_stream_state *stream; + + /* reset cursor limitations on subvp */ + for (i = 0; i < context->stream_count; i++) { + stream = context->streams[i]; + + if (dc_state_can_clear_stream_cursor_subvp_limit(stream, context)) { + dc_state_set_stream_cursor_subvp_limit(stream, context, false); + } + } if (dc->debug.using_dml2) - out = dml2_validate(dc, context, + status = dml2_validate(dc, context, context->power_source == DC_POWER_SOURCE_DC ? context->bw_ctx.dml2_dc_power_source : context->bw_ctx.dml2, - fast_validate); + fast_validate) ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; else - out = dml1_validate(dc, context, fast_validate); - return out; + status = dml1_validate(dc, context, fast_validate) ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; + + if (!fast_validate && status == DC_OK && dc_state_is_subvp_in_use(context)) { + /* check new stream configuration still supports cursor if subvp used */ + for (i = 0; i < context->stream_count; i++) { + stream = context->streams[i]; + + if (dc_state_get_stream_subvp_type(context, stream) != SUBVP_PHANTOM && + stream->cursor_position.enable && + !dc_stream_check_cursor_attributes(stream, context, &stream->cursor_attributes)) { + /* hw cursor cannot be supported with subvp active, so disable subvp for now */ + dc_state_set_stream_cursor_subvp_limit(stream, context, true); + status = DC_FAIL_HW_CURSOR_SUPPORT; + } + }; + } + + if (!fast_validate && status == DC_FAIL_HW_CURSOR_SUPPORT) { + /* attempt to validate again with subvp disabled due to cursor */ + if (dc->debug.using_dml2) + status = dml2_validate(dc, context, + context->power_source == DC_POWER_SOURCE_DC ? context->bw_ctx.dml2_dc_power_source : context->bw_ctx.dml2, + fast_validate) ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; + else + status = dml1_validate(dc, context, fast_validate) ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; + } + + return status; } int dcn32_populate_dml_pipes_from_context( @@ -2042,6 +2080,18 @@ static void dcn32_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw DC_FP_END(); } +unsigned int dcn32_get_max_hw_cursor_size(const struct dc *dc, + struct dc_state *state, + const struct dc_stream_state *stream) +{ + bool limit_cur_to_buf; + + limit_cur_to_buf = dc_state_get_stream_subvp_cursor_limit(stream, state) && + !stream->hw_cursor_req; + + return limit_cur_to_buf ? dc->caps.max_buffered_cursor_size : dc->caps.max_cursor_size; +} + static struct resource_funcs dcn32_res_pool_funcs = { .destroy = dcn32_destroy_resource_pool, .link_enc_create = dcn32_link_encoder_create, @@ -2067,7 +2117,8 @@ static struct resource_funcs dcn32_res_pool_funcs = { .add_phantom_pipes = dcn32_add_phantom_pipes, .build_pipe_pix_clk_params = dcn20_build_pipe_pix_clk_params, .calculate_mall_ways_from_bytes = dcn32_calculate_mall_ways_from_bytes, - .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe + .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe, + .get_max_hw_cursor_size = dcn32_get_max_hw_cursor_size, }; static uint32_t read_pipe_fuses(struct dc_context *ctx) @@ -2114,8 +2165,6 @@ static bool dcn32_resource_construct( #define REG_STRUCT dccg_regs dccg_regs_init(); - DC_FP_START(); - ctx->dc_bios->regs = &bios_regs; pool->base.res_cap = &res_cap_dcn32; @@ -2153,6 +2202,7 @@ static bool dcn32_resource_construct( dc->caps.i2c_speed_in_khz_hdcp = 100; /*1.4 w/a applied by default*/ /* TODO: Bring max_cursor_size back to 256 after subvp cursor corruption is fixed*/ dc->caps.max_cursor_size = 64; + dc->caps.max_buffered_cursor_size = 64; // sqrt(16 * 1024 / 4) dc->caps.min_horizontal_blanking_period = 80; dc->caps.dmdata_alloc_size = 2048; dc->caps.mall_size_per_mem_channel = 4; @@ -2501,14 +2551,10 @@ static bool dcn32_resource_construct( if (ASICREV_IS_GC_11_0_3(dc->ctx->asic_id.hw_internal_rev) && (dc->config.sdpif_request_limit_words_per_umc == 0)) dc->config.sdpif_request_limit_words_per_umc = 16; - DC_FP_END(); - return true; create_fail: - DC_FP_END(); - dcn32_resource_destruct(pool); return false; diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.h index 1aa4ced29291..d60ed77eda80 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.h @@ -98,7 +98,7 @@ void dcn32_add_phantom_pipes(struct dc *dc, unsigned int pipe_cnt, unsigned int index); -bool dcn32_validate_bandwidth(struct dc *dc, +enum dc_status dcn32_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); @@ -188,6 +188,10 @@ void dcn32_override_min_req_dcfclk(struct dc *dc, struct dc_state *context); unsigned int dcn32_calculate_mall_ways_from_bytes(const struct dc *dc, unsigned int total_size_in_mall_bytes); +unsigned int dcn32_get_max_hw_cursor_size(const struct dc *dc, + struct dc_state *state, + const struct dc_stream_state *stream); + /* definitions for run time init of reg offsets */ /* CLK SRC */ diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c index 38d76434683e..7db1f7a5613f 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c @@ -1624,7 +1624,8 @@ static struct resource_funcs dcn321_res_pool_funcs = { .add_phantom_pipes = dcn32_add_phantom_pipes, .build_pipe_pix_clk_params = dcn20_build_pipe_pix_clk_params, .calculate_mall_ways_from_bytes = dcn32_calculate_mall_ways_from_bytes, - .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe + .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe, + .get_max_hw_cursor_size = dcn32_get_max_hw_cursor_size, }; static uint32_t read_pipe_fuses(struct dc_context *ctx) @@ -1709,6 +1710,7 @@ static bool dcn321_resource_construct( dc->caps.i2c_speed_in_khz_hdcp = 100; /*1.4 w/a applied by default*/ /* TODO: Bring max cursor size back to 256 after subvp cursor corruption is fixed*/ dc->caps.max_cursor_size = 64; + dc->caps.max_buffered_cursor_size = 64; // sqrt(16 * 1024 / 4) dc->caps.min_horizontal_blanking_period = 80; dc->caps.dmdata_alloc_size = 2048; dc->caps.mall_size_per_mem_channel = 4; diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c index ffd2b816cd02..72c6cf047db0 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c @@ -1732,7 +1732,7 @@ static void dcn35_get_panel_config_defaults(struct dc_panel_config *panel_config } -static bool dcn35_validate_bandwidth(struct dc *dc, +static enum dc_status dcn35_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { @@ -1743,13 +1743,13 @@ static bool dcn35_validate_bandwidth(struct dc *dc, fast_validate); if (fast_validate) - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; DC_FP_START(); dcn35_decide_zstate_support(dc, context); DC_FP_END(); - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } enum dc_status dcn35_patch_unknown_plane_state(struct dc_plane_state *plane_state) @@ -1903,7 +1903,7 @@ static bool dcn35_resource_construct( dc->caps.max_disp_clock_khz_at_vmin = 650000; /* Sequential ONO is based on ASIC. */ - if (dc->ctx->asic_id.hw_internal_rev > 0x10) + if (dc->ctx->asic_id.hw_internal_rev >= 0x40) dc->caps.sequential_ono = true; /* Use pipe context based otg sync logic */ diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c index 98f5bc1b929e..989a270f7dea 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c @@ -1712,7 +1712,7 @@ static void dcn35_get_panel_config_defaults(struct dc_panel_config *panel_config } -static bool dcn351_validate_bandwidth(struct dc *dc, +static enum dc_status dcn351_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { @@ -1723,13 +1723,13 @@ static bool dcn351_validate_bandwidth(struct dc *dc, fast_validate); if (fast_validate) - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; DC_FP_START(); dcn35_decide_zstate_support(dc, context); DC_FP_END(); - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } static struct resource_funcs dcn351_res_pool_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn36/dcn36_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn36/dcn36_resource.c index b6468573dc33..48e1f234185f 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn36/dcn36_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn36/dcn36_resource.c @@ -1713,7 +1713,7 @@ static void dcn35_get_panel_config_defaults(struct dc_panel_config *panel_config } -static bool dcn35_validate_bandwidth(struct dc *dc, +static enum dc_status dcn35_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { @@ -1724,13 +1724,13 @@ static bool dcn35_validate_bandwidth(struct dc *dc, fast_validate); if (fast_validate) - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; DC_FP_START(); dcn35_decide_zstate_support(dc, context); DC_FP_END(); - return out; + return out ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; } @@ -1876,7 +1876,7 @@ static bool dcn36_resource_construct( dc->caps.max_disp_clock_khz_at_vmin = 650000; /* Sequential ONO is based on ASIC. */ - if (dc->ctx->asic_id.hw_internal_rev > 0x10) + if (dc->ctx->asic_id.hw_internal_rev >= 0x40) dc->caps.sequential_ono = true; /* Use pipe context based otg sync logic */ diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c index 7436dfbdf927..e0e32975ca34 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c @@ -1642,16 +1642,52 @@ enum dc_status dcn401_patch_unknown_plane_state(struct dc_plane_state *plane_sta return DC_OK; } -bool dcn401_validate_bandwidth(struct dc *dc, +enum dc_status dcn401_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate) { - bool out = false; + unsigned int i; + enum dc_status status = DC_OK; + const struct dc_stream_state *stream; + + /* reset cursor limitations on subvp */ + for (i = 0; i < context->stream_count; i++) { + stream = context->streams[i]; + + if (dc_state_can_clear_stream_cursor_subvp_limit(stream, context)) { + dc_state_set_stream_cursor_subvp_limit(stream, context, false); + } + } + if (dc->debug.using_dml2) - out = dml2_validate(dc, context, + status = dml2_validate(dc, context, context->power_source == DC_POWER_SOURCE_DC ? context->bw_ctx.dml2_dc_power_source : context->bw_ctx.dml2, - fast_validate); - return out; + fast_validate) ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; + + if (!fast_validate && status == DC_OK && dc_state_is_subvp_in_use(context)) { + /* check new stream configuration still supports cursor if subvp used */ + for (i = 0; i < context->stream_count; i++) { + stream = context->streams[i]; + + if (dc_state_get_stream_subvp_type(context, stream) != SUBVP_PHANTOM && + stream->cursor_position.enable && + !dc_stream_check_cursor_attributes(stream, context, &stream->cursor_attributes)) { + /* hw cursor cannot be supported with subvp active, so disable subvp for now */ + dc_state_set_stream_cursor_subvp_limit(stream, context, true); + status = DC_FAIL_HW_CURSOR_SUPPORT; + } + }; + } + + if (!fast_validate && status == DC_FAIL_HW_CURSOR_SUPPORT) { + /* attempt to validate again with subvp disabled due to cursor */ + if (dc->debug.using_dml2) + status = dml2_validate(dc, context, + context->power_source == DC_POWER_SOURCE_DC ? context->bw_ctx.dml2_dc_power_source : context->bw_ctx.dml2, + fast_validate) ? DC_OK : DC_FAIL_BANDWIDTH_VALIDATE; + } + + return status; } void dcn401_prepare_mcache_programming(struct dc *dc, @@ -1770,7 +1806,8 @@ static struct resource_funcs dcn401_res_pool_funcs = { .build_pipe_pix_clk_params = dcn401_build_pipe_pix_clk_params, .calculate_mall_ways_from_bytes = dcn32_calculate_mall_ways_from_bytes, .get_power_profile = dcn401_get_power_profile, - .get_vstartup_for_pipe = dcn401_get_vstartup_for_pipe + .get_vstartup_for_pipe = dcn401_get_vstartup_for_pipe, + .get_max_hw_cursor_size = dcn32_get_max_hw_cursor_size }; static uint32_t read_pipe_fuses(struct dc_context *ctx) @@ -1846,8 +1883,9 @@ static bool dcn401_resource_construct( dc->caps.max_downscale_ratio = 600; dc->caps.i2c_speed_in_khz = 95; dc->caps.i2c_speed_in_khz_hdcp = 95; /*1.4 w/a applied by default*/ - /* TODO: Bring max cursor size back to 256 after subvp cursor corruption is fixed*/ + /* used to set cursor pitch, so must be aligned to power of 2 (HW actually supported 78x78) */ dc->caps.max_cursor_size = 64; + dc->caps.max_buffered_cursor_size = 64; dc->caps.cursor_not_scaled = true; dc->caps.min_horizontal_blanking_period = 80; dc->caps.dmdata_alloc_size = 2048; diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h index 4c259745d519..dc52a30991af 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h @@ -22,7 +22,7 @@ struct resource_pool *dcn401_create_resource_pool( enum dc_status dcn401_patch_unknown_plane_state(struct dc_plane_state *plane_state); -bool dcn401_validate_bandwidth(struct dc *dc, +enum dc_status dcn401_validate_bandwidth(struct dc *dc, struct dc_state *context, bool fast_validate); diff --git a/drivers/gpu/drm/amd/display/dc/sspl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/sspl/dc_spl.c index 28348734d900..e0008c5f08ad 100644 --- a/drivers/gpu/drm/amd/display/dc/sspl/dc_spl.c +++ b/drivers/gpu/drm/amd/display/dc/sspl/dc_spl.c @@ -776,7 +776,7 @@ static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in, * Do not bypass UV at 1:1 for cositing to be applied */ if (!enable_isharp) { - if (data->ratios.horz.value == one && data->ratios.vert.value == one) + if (data->ratios.horz.value == one && data->ratios.vert.value == one && !spl_in->basic_out.always_scale) return SCL_MODE_SCALING_420_LUMA_BYPASS; } @@ -884,7 +884,7 @@ static bool spl_get_isharp_en(struct spl_in *spl_in, /* Calculate number of tap with adaptive scaling off */ static void spl_get_taps_non_adaptive_scaler( - struct spl_scratch *spl_scratch, const struct spl_taps *in_taps) + struct spl_scratch *spl_scratch, const struct spl_taps *in_taps, bool always_scale) { bool check_max_downscale = false; @@ -944,15 +944,15 @@ static void spl_get_taps_non_adaptive_scaler( spl_fixpt_from_fraction(6, 1)); SPL_ASSERT(check_max_downscale); - if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz)) + + if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz) && !always_scale) spl_scratch->scl_data.taps.h_taps = 1; - if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert)) + if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert) && !always_scale) spl_scratch->scl_data.taps.v_taps = 1; - if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c)) + if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c) && !always_scale) spl_scratch->scl_data.taps.h_taps_c = 1; - if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c)) + if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c) && !always_scale) spl_scratch->scl_data.taps.v_taps_c = 1; - } /* Calculate optimal number of taps */ @@ -965,13 +965,15 @@ static bool spl_get_optimal_number_of_taps( unsigned int max_taps_y, max_taps_c; unsigned int min_taps_y, min_taps_c; enum lb_memory_config lb_config; - bool skip_easf = false; + bool skip_easf = false; + bool always_scale = spl_in->basic_out.always_scale; bool is_subsampled = spl_is_subsampled_format(spl_in->basic_in.format); + if (spl_scratch->scl_data.viewport.width > spl_scratch->scl_data.h_active && max_downscale_src_width != 0 && spl_scratch->scl_data.viewport.width > max_downscale_src_width) { - spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps); + spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps, always_scale); *enable_easf_v = false; *enable_easf_h = false; *enable_isharp = false; @@ -980,7 +982,7 @@ static bool spl_get_optimal_number_of_taps( /* Disable adaptive scaler and sharpener when integer scaling is enabled */ if (spl_in->scaling_quality.integer_scaling) { - spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps); + spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps, always_scale); *enable_easf_v = false; *enable_easf_h = false; *enable_isharp = false; @@ -996,7 +998,7 @@ static bool spl_get_optimal_number_of_taps( * taps = 4 for upscaling */ if (skip_easf) - spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps); + spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps, always_scale); else { if (spl_is_video_format(spl_in->basic_in.format)) { spl_scratch->scl_data.taps.h_taps = 6; @@ -1297,7 +1299,7 @@ static void spl_set_easf_data(struct spl_scratch *spl_scratch, struct spl_out *s if (enable_easf_v) { dscl_prog_data->easf_v_en = true; dscl_prog_data->easf_v_ring = 0; - dscl_prog_data->easf_v_sharp_factor = 0; + dscl_prog_data->easf_v_sharp_factor = 1; dscl_prog_data->easf_v_bf1_en = 1; // 1-bit, BF1 calculation enable, 0=disable, 1=enable dscl_prog_data->easf_v_bf2_mode = 0xF; // 4-bit, BF2 calculation mode /* 2-bit, BF3 chroma mode correction calculation mode */ @@ -1461,7 +1463,7 @@ static void spl_set_easf_data(struct spl_scratch *spl_scratch, struct spl_out *s if (enable_easf_h) { dscl_prog_data->easf_h_en = true; dscl_prog_data->easf_h_ring = 0; - dscl_prog_data->easf_h_sharp_factor = 0; + dscl_prog_data->easf_h_sharp_factor = 1; dscl_prog_data->easf_h_bf1_en = 1; // 1-bit, BF1 calculation enable, 0=disable, 1=enable dscl_prog_data->easf_h_bf2_mode = @@ -1898,3 +1900,4 @@ bool SPL_NAMESPACE(spl_get_number_of_taps(struct spl_in *spl_in, struct spl_out spl_set_taps_data(dscl_prog_data, data); return res; } + diff --git a/drivers/gpu/drm/amd/display/dc/sspl/dc_spl_types.h b/drivers/gpu/drm/amd/display/dc/sspl/dc_spl_types.h index 1c3949b24611..36a284305a70 100644 --- a/drivers/gpu/drm/amd/display/dc/sspl/dc_spl_types.h +++ b/drivers/gpu/drm/amd/display/dc/sspl/dc_spl_types.h @@ -480,6 +480,10 @@ enum sharpness_setting { SHARPNESS_ZERO, SHARPNESS_CUSTOM }; +enum sharpness_range_source { + SHARPNESS_RANGE_DCN = 0, + SHARPNESS_RANGE_DCN_OVERRIDE +}; struct spl_sharpness_range { int sdr_rgb_min; int sdr_rgb_max; diff --git a/drivers/gpu/drm/amd/display/dc/sspl/spl_fixpt31_32.c b/drivers/gpu/drm/amd/display/dc/sspl/spl_fixpt31_32.c index 52d97918a3bd..ebf0287417e0 100644 --- a/drivers/gpu/drm/amd/display/dc/sspl/spl_fixpt31_32.c +++ b/drivers/gpu/drm/amd/display/dc/sspl/spl_fixpt31_32.c @@ -29,8 +29,6 @@ static inline unsigned long long spl_complete_integer_division_u64( { unsigned long long result; - SPL_ASSERT(divisor); - result = spl_div64_u64_rem(dividend, divisor, remainder); return result; @@ -196,8 +194,6 @@ struct spl_fixed31_32 spl_fixpt_recip(struct spl_fixed31_32 arg) * Good idea to use Newton's method */ - SPL_ASSERT(arg.value); - return spl_fixpt_from_fraction( spl_fixpt_one.value, arg.value); diff --git a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h index 4e0efff92dca..3f3fa1b6a69e 100644 --- a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h +++ b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h @@ -51,8 +51,8 @@ * for the cache windows. * * The call to dmub_srv_hw_init() programs the DMCUB registers to prepare - * for command submission. Commands can be queued via dmub_srv_cmd_queue() - * and executed via dmub_srv_cmd_execute(). + * for command submission. Commands can be queued via dmub_srv_fb_cmd_queue() + * and executed via dmub_srv_fb_cmd_execute(). * * If the queue is full the dmub_srv_wait_for_idle() call can be used to * wait until the queue has been cleared. @@ -142,6 +142,7 @@ enum dmub_notification_type { DMUB_NOTIFICATION_SET_CONFIG_REPLY, DMUB_NOTIFICATION_DPIA_NOTIFICATION, DMUB_NOTIFICATION_HPD_SENSE_NOTIFY, + DMUB_NOTIFICATION_FUSED_IO, DMUB_NOTIFICATION_MAX }; @@ -170,6 +171,13 @@ enum dmub_srv_power_state_type { DMUB_POWER_STATE_D3 = 8 }; +/* enum dmub_inbox_cmd_interface type - defines default interface for host->dmub commands */ +enum dmub_inbox_cmd_interface_type { + DMUB_CMD_INTERFACE_DEFAULT = 0, + DMUB_CMD_INTERFACE_FB = 1, + DMUB_CMD_INTERFACE_REG = 2, +}; + /** * struct dmub_region - dmub hw memory region * @base: base address for region, must be 256 byte aligned @@ -349,6 +357,21 @@ struct dmub_diagnostic_data { uint8_t is_cw6_enabled : 1; }; +struct dmub_srv_inbox { + /* generic status */ + uint64_t num_submitted; + uint64_t num_reported; + union { + /* frame buffer mailbox status */ + struct dmub_rb rb; + /* register mailbox status */ + struct { + bool is_pending; + bool is_multi_pending; + }; + }; +}; + /** * struct dmub_srv_base_funcs - Driver specific base callbacks */ @@ -422,6 +445,8 @@ struct dmub_srv_hw_funcs { uint32_t (*emul_get_inbox1_rptr)(struct dmub_srv *dmub); + uint32_t (*emul_get_inbox1_wptr)(struct dmub_srv *dmub); + void (*emul_set_inbox1_wptr)(struct dmub_srv *dmub, uint32_t wptr_offset); bool (*is_supported)(struct dmub_srv *dmub); @@ -462,18 +487,21 @@ struct dmub_srv_hw_funcs { void (*init_reg_offsets)(struct dmub_srv *dmub, struct dc_context *ctx); void (*subvp_save_surf_addr)(struct dmub_srv *dmub, const struct dc_plane_address *addr, uint8_t subvp_index); + void (*send_reg_inbox0_cmd_msg)(struct dmub_srv *dmub, union dmub_rb_cmd *cmd); uint32_t (*read_reg_inbox0_rsp_int_status)(struct dmub_srv *dmub); void (*read_reg_inbox0_cmd_rsp)(struct dmub_srv *dmub, union dmub_rb_cmd *cmd); void (*write_reg_inbox0_rsp_int_ack)(struct dmub_srv *dmub); + void (*clear_reg_inbox0_rsp_int_ack)(struct dmub_srv *dmub); + void (*enable_reg_inbox0_rsp_int)(struct dmub_srv *dmub, bool enable); + uint32_t (*read_reg_outbox0_rdy_int_status)(struct dmub_srv *dmub); void (*write_reg_outbox0_rdy_int_ack)(struct dmub_srv *dmub); void (*read_reg_outbox0_msg)(struct dmub_srv *dmub, uint32_t *msg); void (*write_reg_outbox0_rsp)(struct dmub_srv *dmub, uint32_t *rsp); uint32_t (*read_reg_outbox0_rsp_int_status)(struct dmub_srv *dmub); - void (*enable_reg_inbox0_rsp_int)(struct dmub_srv *dmub, bool enable); void (*enable_reg_outbox0_rdy_int)(struct dmub_srv *dmub, bool enable); }; @@ -493,6 +521,7 @@ struct dmub_srv_create_params { enum dmub_asic asic; uint32_t fw_version; bool is_virtual; + enum dmub_inbox_cmd_interface_type inbox_type; }; /** @@ -521,8 +550,9 @@ struct dmub_srv { const struct dmub_srv_dcn401_regs *regs_dcn401; struct dmub_srv_base_funcs funcs; struct dmub_srv_hw_funcs hw_funcs; - struct dmub_rb inbox1_rb; + struct dmub_srv_inbox inbox1; uint32_t inbox1_last_wptr; + struct dmub_srv_inbox reg_inbox0; /** * outbox1_rb is accessed without locks (dal & dc) * and to be used only in dmub_srv_stat_get_notification() @@ -542,6 +572,7 @@ struct dmub_srv { struct dmub_fw_meta_info meta_info; struct dmub_feature_caps feature_caps; struct dmub_visual_confirm_color visual_confirm_color; + enum dmub_inbox_cmd_interface_type inbox_type; enum dmub_srv_power_state_type power_state; struct dmub_diagnostic_data debug; @@ -566,11 +597,8 @@ struct dmub_notification { struct aux_reply_data aux_reply; enum dp_hpd_status hpd_status; enum set_config_status sc_status; - /** - * DPIA notification command. - */ - struct dmub_rb_cmd_dpia_notification dpia_notification; struct dmub_rb_cmd_hpd_sense_notify_data hpd_sense_notify; + struct dmub_cmd_fused_request fused_request; }; }; @@ -699,19 +727,7 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub, enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub); /** - * dmub_srv_sync_inbox1() - sync sw state with hw state - * @dmub: the dmub service - * - * Sync sw state with hw state when resume from S0i3 - * - * Return: - * DMUB_STATUS_OK - success - * DMUB_STATUS_INVALID - unspecified error - */ -enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub); - -/** - * dmub_srv_cmd_queue() - queues a command to the DMUB + * dmub_srv_fb_cmd_queue() - queues a command to the DMUB * @dmub: the dmub service * @cmd: the command to queue * @@ -723,11 +739,11 @@ enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub); * DMUB_STATUS_QUEUE_FULL - no remaining room in queue * DMUB_STATUS_INVALID - unspecified error */ -enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub, +enum dmub_status dmub_srv_fb_cmd_queue(struct dmub_srv *dmub, const union dmub_rb_cmd *cmd); /** - * dmub_srv_cmd_execute() - Executes a queued sequence to the dmub + * dmub_srv_fb_cmd_execute() - Executes a queued sequence to the dmub * @dmub: the dmub service * * Begins execution of queued commands on the dmub. @@ -736,7 +752,7 @@ enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub, * DMUB_STATUS_OK - success * DMUB_STATUS_INVALID - unspecified error */ -enum dmub_status dmub_srv_cmd_execute(struct dmub_srv *dmub); +enum dmub_status dmub_srv_fb_cmd_execute(struct dmub_srv *dmub); /** * dmub_srv_wait_for_hw_pwr_up() - Waits for firmware hardware power up is completed @@ -795,6 +811,23 @@ enum dmub_status dmub_srv_wait_for_phy_init(struct dmub_srv *dmub, uint32_t timeout_us); /** + * dmub_srv_wait_for_pending() - Re-entrant wait for messages currently pending + * @dmub: the dmub service + * @timeout_us: the maximum number of microseconds to wait + * + * Waits until the commands queued prior to this call are complete. + * If interfaces remain busy due to additional work being submitted + * concurrently, this function will not continue to wait. + * + * Return: + * DMUB_STATUS_OK - success + * DMUB_STATUS_TIMEOUT - wait for buffer to flush timed out + * DMUB_STATUS_INVALID - unspecified error + */ +enum dmub_status dmub_srv_wait_for_pending(struct dmub_srv *dmub, + uint32_t timeout_us); + +/** * dmub_srv_wait_for_idle() - Waits for the DMUB to be idle * @dmub: the dmub service * @timeout_us: the maximum number of microseconds to wait @@ -892,9 +925,6 @@ enum dmub_status dmub_srv_get_fw_boot_status(struct dmub_srv *dmub, enum dmub_status dmub_srv_get_fw_boot_option(struct dmub_srv *dmub, union dmub_fw_boot_options *option); -enum dmub_status dmub_srv_cmd_with_reply_data(struct dmub_srv *dmub, - union dmub_rb_cmd *cmd); - enum dmub_status dmub_srv_set_skip_panel_power_sequence(struct dmub_srv *dmub, bool skip); @@ -959,26 +989,6 @@ enum dmub_status dmub_srv_clear_inbox0_ack(struct dmub_srv *dmub); void dmub_srv_subvp_save_surf_addr(struct dmub_srv *dmub, const struct dc_plane_address *addr, uint8_t subvp_index); /** - * dmub_srv_send_reg_inbox0_cmd() - send a dmub command and wait for the command - * being processed by DMUB. - * @dmub: The dmub service - * @cmd: The dmub command being sent. If with_replay is true, the function will - * update cmd with replied data. - * @with_reply: true if DMUB reply needs to be copied back to cmd. false if the - * cmd doesn't need to be replied. - * @timeout_us: timeout in microseconds. - * - * Return: - * DMUB_STATUS_OK - success - * DMUB_STATUS_TIMEOUT - DMUB fails to process the command within the timeout - * interval. - */ -enum dmub_status dmub_srv_send_reg_inbox0_cmd( - struct dmub_srv *dmub, - union dmub_rb_cmd *cmd, - bool with_reply, uint32_t timeout_us); - -/** * dmub_srv_set_power_state() - Track DC power state in dmub_srv * @dmub: The dmub service * @power_state: DC power state setting @@ -990,4 +1000,71 @@ enum dmub_status dmub_srv_send_reg_inbox0_cmd( */ void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state); +/** + * dmub_srv_reg_cmd_execute() - Executes provided command to the dmub + * @dmub: the dmub service + * @cmd: the command packet to be executed + * + * Executes a single command for the dmub. + * + * Return: + * DMUB_STATUS_OK - success + * DMUB_STATUS_INVALID - unspecified error + */ +enum dmub_status dmub_srv_reg_cmd_execute(struct dmub_srv *dmub, union dmub_rb_cmd *cmd); + + +/** + * dmub_srv_cmd_get_response() - Copies return data for command into buffer + * @dmub: the dmub service + * @cmd_rsp: response buffer + * + * Copies return data for command into buffer + */ +void dmub_srv_cmd_get_response(struct dmub_srv *dmub, + union dmub_rb_cmd *cmd_rsp); + +/** + * dmub_srv_sync_inboxes() - Sync inbox state + * @dmub: the dmub service + * + * Sync inbox state + * + * Return: + * DMUB_STATUS_OK - success + * DMUB_STATUS_INVALID - unspecified error + */ +enum dmub_status dmub_srv_sync_inboxes(struct dmub_srv *dmub); + +/** + * dmub_srv_wait_for_inbox_free() - Waits for space in the DMUB inbox to free up + * @dmub: the dmub service + * @timeout_us: the maximum number of microseconds to wait + * @num_free_required: number of free entries required + * + * Waits until the DMUB buffer is freed to the specified number. + * The maximum wait time is given in microseconds to prevent spinning + * forever. + * + * Return: + * DMUB_STATUS_OK - success + * DMUB_STATUS_TIMEOUT - wait for buffer to flush timed out + * DMUB_STATUS_INVALID - unspecified error + */ +enum dmub_status dmub_srv_wait_for_inbox_free(struct dmub_srv *dmub, + uint32_t timeout_us, + uint32_t num_free_required); + +/** + * dmub_srv_update_inbox_status() - Updates pending status for inbox & reg inbox0 + * @dmub: the dmub service + * + * Return: + * DMUB_STATUS_OK - success + * DMUB_STATUS_TIMEOUT - wait for buffer to flush timed out + * DMUB_STATUS_HW_FAILURE - issue with HW programming + * DMUB_STATUS_INVALID - unspecified error + */ +enum dmub_status dmub_srv_update_inbox_status(struct dmub_srv *dmub); + #endif /* _DMUB_SRV_H_ */ diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h index 1f5f4e3e49d4..57fa05bddb45 100644 --- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h +++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h @@ -36,6 +36,9 @@ //<DMUB_TYPES>================================================================== /* Basic type definitions. */ +#ifdef __forceinline +#undef __forceinline +#endif #define __forceinline inline /** @@ -547,6 +550,11 @@ union replay_hw_flags { * @is_alpm_initialized: Indicates whether ALPM is initialized */ uint32_t is_alpm_initialized : 1; + + /** + * @alpm_mode: Indicates ALPM mode selected + */ + uint32_t alpm_mode : 2; } bitfields; uint32_t u32All; @@ -739,6 +747,14 @@ enum dmub_ips_disable_type { DMUB_IPS_DISABLE_IPS2_Z10 = 4, DMUB_IPS_DISABLE_DYNAMIC = 5, DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF = 6, + DMUB_IPS_DISABLE_Z8_RETENTION = 7, +}; + +enum dmub_ips_rcg_disable_type { + DMUB_IPS_RCG_ENABLE = 0, + DMUB_IPS0_RCG_DISABLE = 1, + DMUB_IPS1_RCG_DISABLE = 2, + DMUB_IPS_RCG_DISABLE = 3 }; #define DMUB_IPS1_ALLOW_MASK 0x00000001 @@ -817,11 +833,12 @@ enum dmub_shared_state_feature_id { */ union dmub_shared_state_ips_fw_signals { struct { - uint32_t ips1_commit : 1; /**< 1 if in IPS1 */ + uint32_t ips1_commit : 1; /**< 1 if in IPS1 or IPS0 RCG */ uint32_t ips2_commit : 1; /**< 1 if in IPS2 */ uint32_t in_idle : 1; /**< 1 if DMCUB is in idle */ uint32_t detection_required : 1; /**< 1 if detection is required */ - uint32_t reserved_bits : 28; /**< Reversed */ + uint32_t ips1z8_commit: 1; /**< 1 if in IPS1 Z8 Retention */ + uint32_t reserved_bits : 27; /**< Reversed */ } bits; uint32_t all; }; @@ -836,7 +853,10 @@ union dmub_shared_state_ips_driver_signals { uint32_t allow_ips2 : 1; /**< 1 is IPS1 is allowed */ uint32_t allow_z10 : 1; /**< 1 if Z10 is allowed */ uint32_t allow_idle: 1; /**< 1 if driver is allowing idle */ - uint32_t reserved_bits : 27; /**< Reversed bits */ + uint32_t allow_ips0_rcg : 1; /**< 1 is IPS0 RCG is allowed */ + uint32_t allow_ips1_rcg : 1; /**< 1 is IPS1 RCG is allowed */ + uint32_t allow_ips1z8 : 1; /**< 1 is IPS1 Z8 Retention is allowed */ + uint32_t reserved_bits : 24; /**< Reversed bits */ } bits; uint32_t all; }; @@ -865,7 +885,9 @@ struct dmub_shared_state_ips_fw { uint32_t ips1_exit_count; /**< Exit counter for IPS1 */ uint32_t ips2_entry_count; /**< Entry counter for IPS2 */ uint32_t ips2_exit_count; /**< Exit counter for IPS2 */ - uint32_t reserved[55]; /**< Reversed, to be updated when adding new fields. */ + uint32_t ips1_z8ret_entry_count; /**< Entry counter for IPS1 Z8 Retention */ + uint32_t ips1_z8ret_exit_count; /**< Exit counter for IPS1 Z8 Retention */ + uint32_t reserved[53]; /**< Reversed, to be updated when adding new fields. */ }; /* 248-bytes, fixed */ /** @@ -1253,6 +1275,10 @@ enum dmub_gpint_command { * DESC: Setup debug configs. */ DMUB_GPINT__SETUP_DEBUG_MODE = 136, + /** + * DESC: Initiates IPS wake sequence. + */ + DMUB_GPINT__IPS_DEBUG_WAKE = 137, }; /** @@ -2113,6 +2139,11 @@ union dmub_cmd_fams2_config { } stream_v1; //v1 }; +struct dmub_fams2_config_v2 { + struct dmub_cmd_fams2_global_config global; + struct dmub_fams2_stream_static_state_v1 stream_v1[DMUB_MAX_STREAMS]; //v1 +}; + /** * DMUB rb command definition for FAMS2 (merged SubVP, FPO, Legacy) */ @@ -2122,6 +2153,22 @@ struct dmub_rb_cmd_fams2 { }; /** + * Indirect buffer descriptor + */ +struct dmub_ib_data { + union dmub_addr src; // location of indirect buffer in memory + uint16_t size; // indirect buffer size in bytes +}; + +/** + * DMUB rb command definition for commands passed over indirect buffer + */ +struct dmub_rb_cmd_ib { + struct dmub_cmd_header header; + struct dmub_ib_data ib_data; +}; + +/** * enum dmub_cmd_idle_opt_type - Idle optimization command type. */ enum dmub_cmd_idle_opt_type { @@ -2144,6 +2191,11 @@ enum dmub_cmd_idle_opt_type { * DCN hardware notify power state. */ DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE = 3, + + /** + * DCN notify to release HW. + */ + DMUB_CMD__IDLE_OPT_RELEASE_HW = 4, }; /** @@ -2636,7 +2688,11 @@ enum dp_hpd_type { /** * DP HPD short pulse */ - DP_IRQ + DP_IRQ = 1, + /** + * Failure to acquire DP HPD state + */ + DP_NONE_HPD = 2 }; /** @@ -2901,8 +2957,9 @@ enum dmub_cmd_fams_type { */ DMUB_CMD__FAMS_SET_MANUAL_TRIGGER = 3, DMUB_CMD__FAMS2_CONFIG = 4, - DMUB_CMD__FAMS2_DRR_UPDATE = 5, - DMUB_CMD__FAMS2_FLIP = 6, + DMUB_CMD__FAMS2_IB_CONFIG = 5, + DMUB_CMD__FAMS2_DRR_UPDATE = 6, + DMUB_CMD__FAMS2_FLIP = 7, }; /** @@ -3609,6 +3666,12 @@ struct dmub_rb_cmd_psr_set_power_opt { struct dmub_cmd_psr_set_power_opt_data psr_set_power_opt_data; }; +enum dmub_alpm_mode { + ALPM_AUXWAKE = 0, + ALPM_AUXLESS = 1, + ALPM_UNSUPPORTED = 2, +}; + /** * Definition of Replay Residency GPINT command. * Bit[0] - Residency mode for Revision 0 @@ -3742,6 +3805,15 @@ enum dmub_cmd_replay_general_subtype { REPLAY_GENERAL_CMD_SET_LOW_RR_ACTIVATE, }; +struct dmub_alpm_auxless_data { + uint16_t lfps_setup_ns; + uint16_t lfps_period_ns; + uint16_t lfps_silence_ns; + uint16_t lfps_t1_t2_override_us; + short lfps_t1_t2_offset_us; + uint8_t lttpr_count; +}; + /** * Data passed from driver to FW in a DMUB_CMD__REPLAY_COPY_SETTINGS command. */ @@ -3812,6 +3884,10 @@ struct dmub_cmd_replay_copy_settings_data { * Use FSM state for Replay power up/down */ uint8_t use_phy_fsm; + /** + * Use for AUX-less ALPM LFPS wake operation + */ + struct dmub_alpm_auxless_data auxless_alpm_data; }; /** @@ -4360,6 +4436,11 @@ enum dmub_cmd_abm_type { * Get the current ACE curve. */ DMUB_CMD__ABM_GET_ACE_CURVE = 10, + + /** + * Get current histogram data + */ + DMUB_CMD__ABM_GET_HISTOGRAM_DATA = 11, }; struct abm_ace_curve { @@ -4954,6 +5035,20 @@ enum dmub_abm_ace_curve_type { }; /** + * enum dmub_abm_histogram_type - Histogram type. + */ +enum dmub_abm_histogram_type { + /** + * ACE curve as defined by the SW layer. + */ + ABM_HISTOGRAM_TYPE__SW = 0, + /** + * ACE curve as defined by the SW to HW translation interface layer. + */ + ABM_HISTOGRAM_TYPE__SW_IF = 1, +}; + +/** * Definition of a DMUB_CMD__ABM_GET_ACE_CURVE command. */ struct dmub_rb_cmd_abm_get_ace_curve { @@ -4989,6 +5084,41 @@ struct dmub_rb_cmd_abm_get_ace_curve { }; /** + * Definition of a DMUB_CMD__ABM_GET_HISTOGRAM command. + */ +struct dmub_rb_cmd_abm_get_histogram { + /** + * Command header. + */ + struct dmub_cmd_header header; + + /** + * Address where Histogram should be copied. + */ + union dmub_addr dest; + + /** + * Type of Histogram being queried. + */ + enum dmub_abm_histogram_type histogram_type; + + /** + * Indirect buffer length. + */ + uint16_t bytes; + + /** + * eDP panel instance. + */ + uint8_t panel_inst; + + /** + * Explicit padding to 4 byte boundary. + */ + uint8_t pad; +}; + +/** * Definition of a DMUB_CMD__ABM_SAVE_RESTORE command. */ struct dmub_rb_cmd_abm_save_restore { @@ -5389,7 +5519,8 @@ struct dmub_cmd_fused_request { struct dmub_cmd_fused_request_location_i2c { uint8_t is_aux : 1; // False uint8_t ddc_line : 3; - uint8_t _reserved0 : 4; + uint8_t over_aux : 1; + uint8_t _reserved0 : 3; uint8_t address; uint8_t offset; uint8_t length; @@ -5687,6 +5818,11 @@ union dmub_rb_cmd { struct dmub_rb_cmd_abm_get_ace_curve abm_get_ace_curve; /** + * Definition of a DMUB_CMD__ABM_GET_HISTOGRAM command. + */ + struct dmub_rb_cmd_abm_get_histogram abm_get_histogram; + + /** * Definition of a DMUB_CMD__ABM_SET_EVENT command. */ struct dmub_rb_cmd_abm_set_event abm_set_event; @@ -5817,8 +5953,11 @@ union dmub_rb_cmd { * Definition of a DMUB_CMD__PSP_ASSR_ENABLE command. */ struct dmub_rb_cmd_assr_enable assr_enable; + struct dmub_rb_cmd_fams2 fams2_config; + struct dmub_rb_cmd_ib ib_fams2_config; + struct dmub_rb_cmd_fams2_drr_update fams2_drr_update; struct dmub_rb_cmd_fams2_flip fams2_flip; @@ -5934,6 +6073,9 @@ static inline uint32_t dmub_rb_num_free(struct dmub_rb *rb) else data_count = rb->capacity - (rb->rptr - rb->wrpt); + /* +1 because 1 entry is always unusable */ + data_count += DMUB_RB_CMD_SIZE; + return (rb->capacity - data_count) / DMUB_RB_CMD_SIZE; } @@ -5953,6 +6095,7 @@ static inline bool dmub_rb_full(struct dmub_rb *rb) else data_count = rb->capacity - (rb->rptr - rb->wrpt); + /* -1 because 1 entry is always unusable */ return (data_count == (rb->capacity - DMUB_RB_CMD_SIZE)); } diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c index e67f7c4784eb..2575dbc448f7 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c @@ -66,24 +66,20 @@ void dmub_dcn401_reset(struct dmub_srv *dmub) const uint32_t timeout_us = 1 * 1000 * 1000; //1s const uint32_t poll_delay_us = 1; //1us uint32_t i = 0; - uint32_t in_reset, scratch, pwait_mode; + uint32_t enabled, in_reset, scratch, pwait_mode; - REG_GET(DMCUB_CNTL2, DMCUB_SOFT_RESET, &in_reset); + REG_GET(DMCUB_CNTL, + DMCUB_ENABLE, &enabled); + REG_GET(DMCUB_CNTL2, + DMCUB_SOFT_RESET, &in_reset); - if (in_reset == 0) { + if (enabled && in_reset == 0) { cmd.bits.status = 1; cmd.bits.command_code = DMUB_GPINT__STOP_FW; cmd.bits.param = 0; dmub->hw_funcs.set_gpint(dmub, cmd); - for (i = 0; i < timeout_us; i++) { - if (dmub->hw_funcs.is_gpint_acked(dmub, cmd)) - break; - - udelay(poll_delay_us); - } - for (; i < timeout_us; i++) { scratch = dmub->hw_funcs.get_gpint_response(dmub); if (scratch == DMUB_GPINT__STOP_FW_RESPONSE) @@ -517,28 +513,69 @@ void dmub_dcn401_send_reg_inbox0_cmd_msg(struct dmub_srv *dmub, union dmub_rb_cmd *cmd) { uint32_t *dwords = (uint32_t *)cmd; - + int32_t payload_size_bytes = cmd->cmd_common.header.payload_bytes; + uint32_t msg_index; static_assert(sizeof(*cmd) == 64, "DMUB command size mismatch"); - REG_WRITE(DMCUB_REG_INBOX0_MSG0, dwords[0]); - REG_WRITE(DMCUB_REG_INBOX0_MSG1, dwords[1]); - REG_WRITE(DMCUB_REG_INBOX0_MSG2, dwords[2]); - REG_WRITE(DMCUB_REG_INBOX0_MSG3, dwords[3]); - REG_WRITE(DMCUB_REG_INBOX0_MSG4, dwords[4]); - REG_WRITE(DMCUB_REG_INBOX0_MSG5, dwords[5]); - REG_WRITE(DMCUB_REG_INBOX0_MSG6, dwords[6]); - REG_WRITE(DMCUB_REG_INBOX0_MSG7, dwords[7]); - REG_WRITE(DMCUB_REG_INBOX0_MSG8, dwords[8]); - REG_WRITE(DMCUB_REG_INBOX0_MSG9, dwords[9]); - REG_WRITE(DMCUB_REG_INBOX0_MSG10, dwords[10]); - REG_WRITE(DMCUB_REG_INBOX0_MSG11, dwords[11]); - REG_WRITE(DMCUB_REG_INBOX0_MSG12, dwords[12]); - REG_WRITE(DMCUB_REG_INBOX0_MSG13, dwords[13]); - REG_WRITE(DMCUB_REG_INBOX0_MSG14, dwords[14]); + /* read remaining data based on payload size */ + for (msg_index = 0; msg_index < 15; msg_index++) { + if (payload_size_bytes <= msg_index * 4) { + break; + } + + switch (msg_index) { + case 0: + REG_WRITE(DMCUB_REG_INBOX0_MSG0, dwords[msg_index + 1]); + break; + case 1: + REG_WRITE(DMCUB_REG_INBOX0_MSG1, dwords[msg_index + 1]); + break; + case 2: + REG_WRITE(DMCUB_REG_INBOX0_MSG2, dwords[msg_index + 1]); + break; + case 3: + REG_WRITE(DMCUB_REG_INBOX0_MSG3, dwords[msg_index + 1]); + break; + case 4: + REG_WRITE(DMCUB_REG_INBOX0_MSG4, dwords[msg_index + 1]); + break; + case 5: + REG_WRITE(DMCUB_REG_INBOX0_MSG5, dwords[msg_index + 1]); + break; + case 6: + REG_WRITE(DMCUB_REG_INBOX0_MSG6, dwords[msg_index + 1]); + break; + case 7: + REG_WRITE(DMCUB_REG_INBOX0_MSG7, dwords[msg_index + 1]); + break; + case 8: + REG_WRITE(DMCUB_REG_INBOX0_MSG8, dwords[msg_index + 1]); + break; + case 9: + REG_WRITE(DMCUB_REG_INBOX0_MSG9, dwords[msg_index + 1]); + break; + case 10: + REG_WRITE(DMCUB_REG_INBOX0_MSG10, dwords[msg_index + 1]); + break; + case 11: + REG_WRITE(DMCUB_REG_INBOX0_MSG11, dwords[msg_index + 1]); + break; + case 12: + REG_WRITE(DMCUB_REG_INBOX0_MSG12, dwords[msg_index + 1]); + break; + case 13: + REG_WRITE(DMCUB_REG_INBOX0_MSG13, dwords[msg_index + 1]); + break; + case 14: + REG_WRITE(DMCUB_REG_INBOX0_MSG14, dwords[msg_index + 1]); + break; + } + } + /* writing to INBOX RDY register will trigger DMUB REG INBOX0 RDY * interrupt. */ - REG_WRITE(DMCUB_REG_INBOX0_RDY, dwords[15]); + REG_WRITE(DMCUB_REG_INBOX0_RDY, dwords[0]); } uint32_t dmub_dcn401_read_reg_inbox0_rsp_int_status(struct dmub_srv *dmub) @@ -556,30 +593,39 @@ void dmub_dcn401_read_reg_inbox0_cmd_rsp(struct dmub_srv *dmub, static_assert(sizeof(*cmd) == 64, "DMUB command size mismatch"); - dwords[0] = REG_READ(DMCUB_REG_INBOX0_MSG0); - dwords[1] = REG_READ(DMCUB_REG_INBOX0_MSG1); - dwords[2] = REG_READ(DMCUB_REG_INBOX0_MSG2); - dwords[3] = REG_READ(DMCUB_REG_INBOX0_MSG3); - dwords[4] = REG_READ(DMCUB_REG_INBOX0_MSG4); - dwords[5] = REG_READ(DMCUB_REG_INBOX0_MSG5); - dwords[6] = REG_READ(DMCUB_REG_INBOX0_MSG6); - dwords[7] = REG_READ(DMCUB_REG_INBOX0_MSG7); - dwords[8] = REG_READ(DMCUB_REG_INBOX0_MSG8); - dwords[9] = REG_READ(DMCUB_REG_INBOX0_MSG9); - dwords[10] = REG_READ(DMCUB_REG_INBOX0_MSG10); - dwords[11] = REG_READ(DMCUB_REG_INBOX0_MSG11); - dwords[12] = REG_READ(DMCUB_REG_INBOX0_MSG12); - dwords[13] = REG_READ(DMCUB_REG_INBOX0_MSG13); - dwords[14] = REG_READ(DMCUB_REG_INBOX0_MSG14); - dwords[15] = REG_READ(DMCUB_REG_INBOX0_RSP); + dwords[0] = REG_READ(DMCUB_REG_INBOX0_RSP); + dwords[1] = REG_READ(DMCUB_REG_INBOX0_MSG0); + dwords[2] = REG_READ(DMCUB_REG_INBOX0_MSG1); + dwords[3] = REG_READ(DMCUB_REG_INBOX0_MSG2); + dwords[4] = REG_READ(DMCUB_REG_INBOX0_MSG3); + dwords[5] = REG_READ(DMCUB_REG_INBOX0_MSG4); + dwords[6] = REG_READ(DMCUB_REG_INBOX0_MSG5); + dwords[7] = REG_READ(DMCUB_REG_INBOX0_MSG6); + dwords[8] = REG_READ(DMCUB_REG_INBOX0_MSG7); + dwords[9] = REG_READ(DMCUB_REG_INBOX0_MSG8); + dwords[10] = REG_READ(DMCUB_REG_INBOX0_MSG9); + dwords[11] = REG_READ(DMCUB_REG_INBOX0_MSG10); + dwords[12] = REG_READ(DMCUB_REG_INBOX0_MSG11); + dwords[13] = REG_READ(DMCUB_REG_INBOX0_MSG12); + dwords[14] = REG_READ(DMCUB_REG_INBOX0_MSG13); + dwords[15] = REG_READ(DMCUB_REG_INBOX0_MSG14); } void dmub_dcn401_write_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub) { REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_ACK, 1); +} + +void dmub_dcn401_clear_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub) +{ REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_ACK, 0); } +void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable) +{ + REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_EN, enable ? 1:0); +} + void dmub_dcn401_write_reg_outbox0_rdy_int_ack(struct dmub_srv *dmub) { REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_ACK, 1); @@ -604,11 +650,6 @@ uint32_t dmub_dcn401_read_reg_outbox0_rsp_int_status(struct dmub_srv *dmub) return status; } -void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable) -{ - REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_EN, enable ? 1:0); -} - void dmub_dcn401_enable_reg_outbox0_rdy_int(struct dmub_srv *dmub, bool enable) { REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_EN, enable ? 1:0); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h index c35be52676f6..88c3a44d67d9 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h @@ -277,11 +277,13 @@ uint32_t dmub_dcn401_read_reg_inbox0_rsp_int_status(struct dmub_srv *dmub); void dmub_dcn401_read_reg_inbox0_cmd_rsp(struct dmub_srv *dmub, union dmub_rb_cmd *cmd); void dmub_dcn401_write_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub); +void dmub_dcn401_clear_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub); +void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable); + void dmub_dcn401_write_reg_outbox0_rdy_int_ack(struct dmub_srv *dmub); void dmub_dcn401_read_reg_outbox0_msg(struct dmub_srv *dmub, uint32_t *msg); void dmub_dcn401_write_reg_outbox0_rsp(struct dmub_srv *dmub, uint32_t *msg); uint32_t dmub_dcn401_read_reg_outbox0_rsp_int_status(struct dmub_srv *dmub); -void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable); void dmub_dcn401_enable_reg_outbox0_rdy_int(struct dmub_srv *dmub, bool enable); uint32_t dmub_dcn401_read_reg_outbox0_rdy_int_status(struct dmub_srv *dmub); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c index ae8133816b43..acca7943a8c8 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c @@ -157,6 +157,9 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) { struct dmub_srv_hw_funcs *funcs = &dmub->hw_funcs; + /* default to specifying now inbox type */ + enum dmub_inbox_cmd_interface_type default_inbox_type = DMUB_CMD_INTERFACE_DEFAULT; + switch (asic) { case DMUB_ASIC_DCN20: case DMUB_ASIC_DCN21: @@ -395,10 +398,15 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) funcs->get_current_time = dmub_dcn401_get_current_time; funcs->get_diagnostic_data = dmub_dcn401_get_diagnostic_data; + funcs->send_reg_inbox0_cmd_msg = dmub_dcn401_send_reg_inbox0_cmd_msg; funcs->read_reg_inbox0_rsp_int_status = dmub_dcn401_read_reg_inbox0_rsp_int_status; funcs->read_reg_inbox0_cmd_rsp = dmub_dcn401_read_reg_inbox0_cmd_rsp; funcs->write_reg_inbox0_rsp_int_ack = dmub_dcn401_write_reg_inbox0_rsp_int_ack; + funcs->clear_reg_inbox0_rsp_int_ack = dmub_dcn401_clear_reg_inbox0_rsp_int_ack; + funcs->enable_reg_inbox0_rsp_int = dmub_dcn401_enable_reg_inbox0_rsp_int; + default_inbox_type = DMUB_CMD_INTERFACE_FB; // still default to FB for now + funcs->write_reg_outbox0_rdy_int_ack = dmub_dcn401_write_reg_outbox0_rdy_int_ack; funcs->read_reg_outbox0_msg = dmub_dcn401_read_reg_outbox0_msg; funcs->write_reg_outbox0_rsp = dmub_dcn401_write_reg_outbox0_rsp; @@ -411,6 +419,20 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) return false; } + /* set default inbox type if not overriden */ + if (dmub->inbox_type == DMUB_CMD_INTERFACE_DEFAULT) { + if (default_inbox_type != DMUB_CMD_INTERFACE_DEFAULT) { + /* use default inbox type as specified by DCN rev */ + dmub->inbox_type = default_inbox_type; + } else if (funcs->send_reg_inbox0_cmd_msg) { + /* prefer reg as default inbox type if present */ + dmub->inbox_type = DMUB_CMD_INTERFACE_REG; + } else { + /* use fb as fallback */ + dmub->inbox_type = DMUB_CMD_INTERFACE_FB; + } + } + return true; } @@ -426,6 +448,7 @@ enum dmub_status dmub_srv_create(struct dmub_srv *dmub, dmub->asic = params->asic; dmub->fw_version = params->fw_version; dmub->is_virtual = params->is_virtual; + dmub->inbox_type = params->inbox_type; /* Setup asic dependent hardware funcs. */ if (!dmub_srv_hw_setup(dmub, params->asic)) { @@ -695,7 +718,7 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub, inbox1.base = cw4.region.base; inbox1.top = cw4.region.base + DMUB_RB_SIZE; outbox1.base = inbox1.top; - outbox1.top = cw4.region.top; + outbox1.top = inbox1.top + DMUB_RB_SIZE; cw5.offset.quad_part = tracebuff_fb->gpu_addr; cw5.region.base = DMUB_CW5_BASE; @@ -737,7 +760,7 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub, rb_params.ctx = dmub; rb_params.base_address = mail_fb->cpu_addr; rb_params.capacity = DMUB_RB_SIZE; - dmub_rb_init(&dmub->inbox1_rb, &rb_params); + dmub_rb_init(&dmub->inbox1.rb, &rb_params); // Initialize outbox1 ring buffer rb_params.ctx = dmub; @@ -768,27 +791,6 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub, return DMUB_STATUS_OK; } -enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub) -{ - if (!dmub->sw_init) - return DMUB_STATUS_INVALID; - - if (dmub->hw_funcs.get_inbox1_rptr && dmub->hw_funcs.get_inbox1_wptr) { - uint32_t rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); - uint32_t wptr = dmub->hw_funcs.get_inbox1_wptr(dmub); - - if (rptr > dmub->inbox1_rb.capacity || wptr > dmub->inbox1_rb.capacity) { - return DMUB_STATUS_HW_FAILURE; - } else { - dmub->inbox1_rb.rptr = rptr; - dmub->inbox1_rb.wrpt = wptr; - dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt; - } - } - - return DMUB_STATUS_OK; -} - enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub) { if (!dmub->sw_init) @@ -799,8 +801,13 @@ enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub) /* mailboxes have been reset in hw, so reset the sw state as well */ dmub->inbox1_last_wptr = 0; - dmub->inbox1_rb.wrpt = 0; - dmub->inbox1_rb.rptr = 0; + dmub->inbox1.rb.wrpt = 0; + dmub->inbox1.rb.rptr = 0; + dmub->inbox1.num_reported = 0; + dmub->inbox1.num_submitted = 0; + dmub->reg_inbox0.num_reported = 0; + dmub->reg_inbox0.num_submitted = 0; + dmub->reg_inbox0.is_pending = 0; dmub->outbox0_rb.wrpt = 0; dmub->outbox0_rb.rptr = 0; dmub->outbox1_rb.wrpt = 0; @@ -811,7 +818,7 @@ enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub) return DMUB_STATUS_OK; } -enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub, +enum dmub_status dmub_srv_fb_cmd_queue(struct dmub_srv *dmub, const union dmub_rb_cmd *cmd) { if (!dmub->hw_init) @@ -820,18 +827,20 @@ enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub, if (dmub->power_state != DMUB_POWER_STATE_D0) return DMUB_STATUS_POWER_STATE_D3; - if (dmub->inbox1_rb.rptr > dmub->inbox1_rb.capacity || - dmub->inbox1_rb.wrpt > dmub->inbox1_rb.capacity) { + if (dmub->inbox1.rb.rptr > dmub->inbox1.rb.capacity || + dmub->inbox1.rb.wrpt > dmub->inbox1.rb.capacity) { return DMUB_STATUS_HW_FAILURE; } - if (dmub_rb_push_front(&dmub->inbox1_rb, cmd)) + if (dmub_rb_push_front(&dmub->inbox1.rb, cmd)) { + dmub->inbox1.num_submitted++; return DMUB_STATUS_OK; + } return DMUB_STATUS_QUEUE_FULL; } -enum dmub_status dmub_srv_cmd_execute(struct dmub_srv *dmub) +enum dmub_status dmub_srv_fb_cmd_execute(struct dmub_srv *dmub) { struct dmub_rb flush_rb; @@ -846,13 +855,13 @@ enum dmub_status dmub_srv_cmd_execute(struct dmub_srv *dmub) * been flushed to framebuffer memory. Otherwise DMCUB might * read back stale, fully invalid or partially invalid data. */ - flush_rb = dmub->inbox1_rb; + flush_rb = dmub->inbox1.rb; flush_rb.rptr = dmub->inbox1_last_wptr; dmub_rb_flush_pending(&flush_rb); - dmub->hw_funcs.set_inbox1_wptr(dmub, dmub->inbox1_rb.wrpt); + dmub->hw_funcs.set_inbox1_wptr(dmub, dmub->inbox1.rb.wrpt); - dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt; + dmub->inbox1_last_wptr = dmub->inbox1.rb.wrpt; return DMUB_STATUS_OK; } @@ -910,26 +919,84 @@ enum dmub_status dmub_srv_wait_for_auto_load(struct dmub_srv *dmub, return DMUB_STATUS_TIMEOUT; } +static void dmub_srv_update_reg_inbox0_status(struct dmub_srv *dmub) +{ + if (dmub->reg_inbox0.is_pending) { + dmub->reg_inbox0.is_pending = dmub->hw_funcs.read_reg_inbox0_rsp_int_status && + !dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub); + + if (!dmub->reg_inbox0.is_pending) { + /* ack the rsp interrupt */ + if (dmub->hw_funcs.write_reg_inbox0_rsp_int_ack) + dmub->hw_funcs.write_reg_inbox0_rsp_int_ack(dmub); + + /* only update the reported count if commands aren't being batched */ + if (!dmub->reg_inbox0.is_pending && !dmub->reg_inbox0.is_multi_pending) { + dmub->reg_inbox0.num_reported = dmub->reg_inbox0.num_submitted; + } + } + } +} + +enum dmub_status dmub_srv_wait_for_pending(struct dmub_srv *dmub, + uint32_t timeout_us) +{ + uint32_t i; + const uint32_t polling_interval_us = 1; + struct dmub_srv_inbox scratch_reg_inbox0 = dmub->reg_inbox0; + struct dmub_srv_inbox scratch_inbox1 = dmub->inbox1; + const volatile struct dmub_srv_inbox *reg_inbox0 = &dmub->reg_inbox0; + const volatile struct dmub_srv_inbox *inbox1 = &dmub->inbox1; + + if (!dmub->hw_init || + !dmub->hw_funcs.get_inbox1_wptr) + return DMUB_STATUS_INVALID; + + for (i = 0; i <= timeout_us; i += polling_interval_us) { + scratch_inbox1.rb.wrpt = dmub->hw_funcs.get_inbox1_wptr(dmub); + scratch_inbox1.rb.rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); + + scratch_reg_inbox0.is_pending = scratch_reg_inbox0.is_pending && + dmub->hw_funcs.read_reg_inbox0_rsp_int_status && + !dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub); + + if (scratch_inbox1.rb.rptr > dmub->inbox1.rb.capacity) + return DMUB_STATUS_HW_FAILURE; + + /* check current HW state first, but use command submission vs reported as a fallback */ + if ((dmub_rb_empty(&scratch_inbox1.rb) || + inbox1->num_reported >= scratch_inbox1.num_submitted) && + (!scratch_reg_inbox0.is_pending || + reg_inbox0->num_reported >= scratch_reg_inbox0.num_submitted)) + return DMUB_STATUS_OK; + + udelay(polling_interval_us); + } + + return DMUB_STATUS_TIMEOUT; +} + enum dmub_status dmub_srv_wait_for_idle(struct dmub_srv *dmub, uint32_t timeout_us) { - uint32_t i, rptr; + enum dmub_status status; + uint32_t i; + const uint32_t polling_interval_us = 1; if (!dmub->hw_init) return DMUB_STATUS_INVALID; - for (i = 0; i <= timeout_us; ++i) { - rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); + for (i = 0; i < timeout_us; i += polling_interval_us) { + status = dmub_srv_update_inbox_status(dmub); - if (rptr > dmub->inbox1_rb.capacity) - return DMUB_STATUS_HW_FAILURE; + if (status != DMUB_STATUS_OK) + return status; - dmub->inbox1_rb.rptr = rptr; - - if (dmub_rb_empty(&dmub->inbox1_rb)) + /* check for idle */ + if (dmub_rb_empty(&dmub->inbox1.rb) && !dmub->reg_inbox0.is_pending) return DMUB_STATUS_OK; - udelay(1); + udelay(polling_interval_us); } return DMUB_STATUS_TIMEOUT; @@ -1040,35 +1107,6 @@ enum dmub_status dmub_srv_set_skip_panel_power_sequence(struct dmub_srv *dmub, return DMUB_STATUS_OK; } -enum dmub_status dmub_srv_cmd_with_reply_data(struct dmub_srv *dmub, - union dmub_rb_cmd *cmd) -{ - enum dmub_status status = DMUB_STATUS_OK; - - // Queue command - status = dmub_srv_cmd_queue(dmub, cmd); - - if (status != DMUB_STATUS_OK) - return status; - - // Execute command - status = dmub_srv_cmd_execute(dmub); - - if (status != DMUB_STATUS_OK) - return status; - - // Wait for DMUB to process command - status = dmub_srv_wait_for_idle(dmub, 100000); - - if (status != DMUB_STATUS_OK) - return status; - - // Copy data back from ring buffer into command - dmub_rb_get_return_data(&dmub->inbox1_rb, cmd); - - return status; -} - static inline bool dmub_rb_out_trace_buffer_front(struct dmub_rb *rb, void *entry) { @@ -1160,47 +1198,162 @@ void dmub_srv_subvp_save_surf_addr(struct dmub_srv *dmub, const struct dc_plane_ } } +void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state) +{ + if (!dmub || !dmub->hw_init) + return; + + dmub->power_state = dmub_srv_power_state; +} -enum dmub_status dmub_srv_send_reg_inbox0_cmd( - struct dmub_srv *dmub, - union dmub_rb_cmd *cmd, - bool with_reply, uint32_t timeout_us) +enum dmub_status dmub_srv_reg_cmd_execute(struct dmub_srv *dmub, union dmub_rb_cmd *cmd) { - uint32_t rsp_ready = 0; - uint32_t i; + uint32_t num_pending = 0; + + if (!dmub->hw_init) + return DMUB_STATUS_INVALID; + if (dmub->power_state != DMUB_POWER_STATE_D0) + return DMUB_STATUS_POWER_STATE_D3; + + if (!dmub->hw_funcs.send_reg_inbox0_cmd_msg || + !dmub->hw_funcs.clear_reg_inbox0_rsp_int_ack) + return DMUB_STATUS_INVALID; + + if (dmub->reg_inbox0.num_submitted >= dmub->reg_inbox0.num_reported) + num_pending = dmub->reg_inbox0.num_submitted - dmub->reg_inbox0.num_reported; + else + /* num_submitted wrapped */ + num_pending = DMUB_REG_INBOX0_RB_MAX_ENTRY - + (dmub->reg_inbox0.num_reported - dmub->reg_inbox0.num_submitted); + + if (num_pending >= DMUB_REG_INBOX0_RB_MAX_ENTRY) + return DMUB_STATUS_QUEUE_FULL; + + /* clear last rsp ack and send message */ + dmub->hw_funcs.clear_reg_inbox0_rsp_int_ack(dmub); dmub->hw_funcs.send_reg_inbox0_cmd_msg(dmub, cmd); - for (i = 0; i < timeout_us; i++) { - rsp_ready = dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub); - if (rsp_ready) - break; - udelay(1); + dmub->reg_inbox0.num_submitted++; + dmub->reg_inbox0.is_pending = true; + dmub->reg_inbox0.is_multi_pending = cmd->cmd_common.header.multi_cmd_pending; + + return DMUB_STATUS_OK; +} + +void dmub_srv_cmd_get_response(struct dmub_srv *dmub, + union dmub_rb_cmd *cmd_rsp) +{ + if (dmub) { + if (dmub->inbox_type == DMUB_CMD_INTERFACE_REG && + dmub->hw_funcs.read_reg_inbox0_cmd_rsp) { + dmub->hw_funcs.read_reg_inbox0_cmd_rsp(dmub, cmd_rsp); + } else { + dmub_rb_get_return_data(&dmub->inbox1.rb, cmd_rsp); + } } - if (rsp_ready == 0) - return DMUB_STATUS_TIMEOUT; +} + +static enum dmub_status dmub_srv_sync_reg_inbox0(struct dmub_srv *dmub) +{ + if (!dmub || !dmub->sw_init) + return DMUB_STATUS_INVALID; - if (with_reply) - dmub->hw_funcs.read_reg_inbox0_cmd_rsp(dmub, cmd); + dmub->reg_inbox0.is_pending = 0; + dmub->reg_inbox0.is_multi_pending = 0; - dmub->hw_funcs.write_reg_inbox0_rsp_int_ack(dmub); + return DMUB_STATUS_OK; +} - /* wait for rsp int status is cleared to initial state before exit */ - for (; i <= timeout_us; i++) { - rsp_ready = dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub); - if (rsp_ready == 0) - break; - udelay(1); +static enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub) +{ + if (!dmub->sw_init) + return DMUB_STATUS_INVALID; + + if (dmub->hw_funcs.get_inbox1_rptr && dmub->hw_funcs.get_inbox1_wptr) { + uint32_t rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); + uint32_t wptr = dmub->hw_funcs.get_inbox1_wptr(dmub); + + if (rptr > dmub->inbox1.rb.capacity || wptr > dmub->inbox1.rb.capacity) { + return DMUB_STATUS_HW_FAILURE; + } else { + dmub->inbox1.rb.rptr = rptr; + dmub->inbox1.rb.wrpt = wptr; + dmub->inbox1_last_wptr = dmub->inbox1.rb.wrpt; + } } - ASSERT(rsp_ready == 0); return DMUB_STATUS_OK; } -void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state) +enum dmub_status dmub_srv_sync_inboxes(struct dmub_srv *dmub) { - if (!dmub || !dmub->hw_init) - return; + enum dmub_status status; - dmub->power_state = dmub_srv_power_state; + status = dmub_srv_sync_reg_inbox0(dmub); + if (status != DMUB_STATUS_OK) + return status; + + status = dmub_srv_sync_inbox1(dmub); + if (status != DMUB_STATUS_OK) + return status; + + return DMUB_STATUS_OK; +} + +enum dmub_status dmub_srv_wait_for_inbox_free(struct dmub_srv *dmub, + uint32_t timeout_us, + uint32_t num_free_required) +{ + enum dmub_status status; + uint32_t i; + const uint32_t polling_interval_us = 1; + + if (!dmub->hw_init) + return DMUB_STATUS_INVALID; + + for (i = 0; i < timeout_us; i += polling_interval_us) { + status = dmub_srv_update_inbox_status(dmub); + + if (status != DMUB_STATUS_OK) + return status; + + /* check for space in inbox1 */ + if (dmub_rb_num_free(&dmub->inbox1.rb) >= num_free_required) + return DMUB_STATUS_OK; + + udelay(polling_interval_us); + } + + return DMUB_STATUS_TIMEOUT; +} + +enum dmub_status dmub_srv_update_inbox_status(struct dmub_srv *dmub) +{ + uint32_t rptr; + + if (!dmub->hw_init) + return DMUB_STATUS_INVALID; + + if (dmub->power_state != DMUB_POWER_STATE_D0) + return DMUB_STATUS_POWER_STATE_D3; + + /* update inbox1 state */ + rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); + + if (rptr > dmub->inbox1.rb.capacity) + return DMUB_STATUS_HW_FAILURE; + + if (dmub->inbox1.rb.rptr > rptr) { + /* rb wrapped */ + dmub->inbox1.num_reported += (rptr + dmub->inbox1.rb.capacity - dmub->inbox1.rb.rptr) / DMUB_RB_CMD_SIZE; + } else { + dmub->inbox1.num_reported += (rptr - dmub->inbox1.rb.rptr) / DMUB_RB_CMD_SIZE; + } + dmub->inbox1.rb.rptr = rptr; + + /* update reg_inbox0 */ + dmub_srv_update_reg_inbox0_status(dmub); + + return DMUB_STATUS_OK; } diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv_stat.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv_stat.c index cce887cefc01..567c5b1aeb7a 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv_stat.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv_stat.c @@ -95,23 +95,6 @@ enum dmub_status dmub_srv_stat_get_notification(struct dmub_srv *dmub, case DMUB_OUT_CMD__DPIA_NOTIFICATION: notify->type = DMUB_NOTIFICATION_DPIA_NOTIFICATION; notify->link_index = cmd.dpia_notification.payload.header.instance; - - if (cmd.dpia_notification.payload.header.type == DPIA_NOTIFY__BW_ALLOCATION) { - - notify->dpia_notification.payload.data.dpia_bw_alloc.estimated_bw = - cmd.dpia_notification.payload.data.dpia_bw_alloc.estimated_bw; - notify->dpia_notification.payload.data.dpia_bw_alloc.allocated_bw = - cmd.dpia_notification.payload.data.dpia_bw_alloc.allocated_bw; - - if (cmd.dpia_notification.payload.data.dpia_bw_alloc.bits.bw_request_failed) - notify->result = DPIA_BW_REQ_FAILED; - else if (cmd.dpia_notification.payload.data.dpia_bw_alloc.bits.bw_request_succeeded) - notify->result = DPIA_BW_REQ_SUCCESS; - else if (cmd.dpia_notification.payload.data.dpia_bw_alloc.bits.est_bw_changed) - notify->result = DPIA_EST_BW_CHANGED; - else if (cmd.dpia_notification.payload.data.dpia_bw_alloc.bits.bw_alloc_cap_changed) - notify->result = DPIA_BW_ALLOC_CAPS_CHANGED; - } break; case DMUB_OUT_CMD__HPD_SENSE_NOTIFY: notify->type = DMUB_NOTIFICATION_HPD_SENSE_NOTIFY; @@ -119,6 +102,10 @@ enum dmub_status dmub_srv_stat_get_notification(struct dmub_srv *dmub, &cmd.hpd_sense_notify.data, sizeof(cmd.hpd_sense_notify.data)); break; + case DMUB_OUT_CMD__FUSED_IO: + notify->type = DMUB_NOTIFICATION_FUSED_IO; + dmub_memcpy(¬ify->fused_request, &cmd.fused_io.request, sizeof(cmd.fused_io.request)); + break; default: notify->type = DMUB_NOTIFICATION_NO_DATA; break; diff --git a/drivers/gpu/drm/amd/display/include/gpio_service_interface.h b/drivers/gpu/drm/amd/display/include/gpio_service_interface.h index 7e3240e73c1f..63813009a3a6 100644 --- a/drivers/gpu/drm/amd/display/include/gpio_service_interface.h +++ b/drivers/gpu/drm/amd/display/include/gpio_service_interface.h @@ -86,6 +86,9 @@ enum dc_irq_source dal_irq_get_source( enum dc_irq_source dal_irq_get_rx_source( const struct gpio *irq); +enum dc_irq_source dal_irq_get_read_request( + const struct gpio *irq); + enum gpio_result dal_irq_setup_hpd_filter( struct gpio *irq, struct gpio_hpd_config *config); diff --git a/drivers/gpu/drm/amd/display/include/link_service_types.h b/drivers/gpu/drm/amd/display/include/link_service_types.h index 1867aac57cf2..da74ed66c8f9 100644 --- a/drivers/gpu/drm/amd/display/include/link_service_types.h +++ b/drivers/gpu/drm/amd/display/include/link_service_types.h @@ -89,6 +89,8 @@ struct link_training_settings { bool enhanced_framing; enum lttpr_mode lttpr_mode; + bool lttpr_early_tps2; + /* disallow different lanes to have different lane settings */ bool disallow_per_lane_settings; /* dpcd lane settings will always use the same hw lane settings diff --git a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h index 55c7d873175f..a37634942b07 100644 --- a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h +++ b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp.h @@ -386,6 +386,7 @@ enum mod_hdcp_status mod_hdcp_write_repeater_auth_ack(struct mod_hdcp *hdcp); enum mod_hdcp_status mod_hdcp_write_stream_manage(struct mod_hdcp *hdcp); enum mod_hdcp_status mod_hdcp_write_content_type(struct mod_hdcp *hdcp); enum mod_hdcp_status mod_hdcp_clear_cp_irq_status(struct mod_hdcp *hdcp); +enum mod_hdcp_status mod_hdcp_write_poll_read_lc_fw(struct mod_hdcp *hdcp); /* hdcp version helpers */ static inline uint8_t is_dp_hdcp(struct mod_hdcp *hdcp) diff --git a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_execution.c b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_execution.c index 1d41dd58f6bc..bb8ae80b37f8 100644 --- a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_execution.c +++ b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_execution.c @@ -452,21 +452,12 @@ out: return status; } -static enum mod_hdcp_status locality_check(struct mod_hdcp *hdcp, +static enum mod_hdcp_status locality_check_sw(struct mod_hdcp *hdcp, struct mod_hdcp_event_context *event_ctx, struct mod_hdcp_transition_input_hdcp2 *input) { enum mod_hdcp_status status = MOD_HDCP_STATUS_SUCCESS; - if (event_ctx->event != MOD_HDCP_EVENT_CALLBACK) { - event_ctx->unexpected_event = 1; - goto out; - } - - if (!mod_hdcp_execute_and_set(mod_hdcp_hdcp2_prepare_lc_init, - &input->lc_init_prepare, &status, - hdcp, "lc_init_prepare")) - goto out; if (!mod_hdcp_execute_and_set(mod_hdcp_write_lc_init, &input->lc_init_write, &status, hdcp, "lc_init_write")) @@ -482,6 +473,48 @@ static enum mod_hdcp_status locality_check(struct mod_hdcp *hdcp, &input->l_prime_read, &status, hdcp, "l_prime_read")) goto out; +out: + return status; +} + +static enum mod_hdcp_status locality_check_fw(struct mod_hdcp *hdcp, + struct mod_hdcp_event_context *event_ctx, + struct mod_hdcp_transition_input_hdcp2 *input) +{ + enum mod_hdcp_status status = MOD_HDCP_STATUS_SUCCESS; + + if (!mod_hdcp_execute_and_set(mod_hdcp_write_poll_read_lc_fw, + &input->l_prime_read, &status, + hdcp, "l_prime_read")) + goto out; + +out: + return status; +} + +static enum mod_hdcp_status locality_check(struct mod_hdcp *hdcp, + struct mod_hdcp_event_context *event_ctx, + struct mod_hdcp_transition_input_hdcp2 *input) +{ + enum mod_hdcp_status status = MOD_HDCP_STATUS_SUCCESS; + const bool use_fw = hdcp->config.ddc.funcs.atomic_write_poll_read_i2c + && hdcp->config.ddc.funcs.atomic_write_poll_read_aux + && !hdcp->connection.link.adjust.hdcp2.force_sw_locality_check; + + if (event_ctx->event != MOD_HDCP_EVENT_CALLBACK) { + event_ctx->unexpected_event = 1; + goto out; + } + + if (!mod_hdcp_execute_and_set(mod_hdcp_hdcp2_prepare_lc_init, + &input->lc_init_prepare, &status, + hdcp, "lc_init_prepare")) + goto out; + + status = (use_fw ? locality_check_fw : locality_check_sw)(hdcp, event_ctx, input); + if (status != MOD_HDCP_STATUS_SUCCESS) + goto out; + if (!mod_hdcp_execute_and_set(mod_hdcp_hdcp2_validate_l_prime, &input->l_prime_validation, &status, hdcp, "l_prime_validation")) diff --git a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_transition.c b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_transition.c index c5f6c11de7e5..89ffb89e1932 100644 --- a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_transition.c +++ b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp2_transition.c @@ -184,17 +184,28 @@ enum mod_hdcp_status mod_hdcp_hdcp2_transition(struct mod_hdcp *hdcp, callback_in_ms(0, output); set_state_id(hdcp, output, H2_A2_LOCALITY_CHECK); break; - case H2_A2_LOCALITY_CHECK: + case H2_A2_LOCALITY_CHECK: { + const bool use_fw = hdcp->config.ddc.funcs.atomic_write_poll_read_i2c + && !adjust->hdcp2.force_sw_locality_check; + + /* + * 1A-05: consider disconnection after LC init a failure + * 1A-13-1: consider invalid l' a failure + * 1A-13-2: consider l' timeout a failure + */ if (hdcp->state.stay_count > 10 || input->lc_init_prepare != PASS || - input->lc_init_write != PASS || - input->l_prime_available_poll != PASS || - input->l_prime_read != PASS) { - /* - * 1A-05: consider disconnection after LC init a failure - * 1A-13-1: consider invalid l' a failure - * 1A-13-2: consider l' timeout a failure - */ + (!use_fw && input->lc_init_write != PASS) || + (!use_fw && input->l_prime_available_poll != PASS)) { + fail_and_restart_in_ms(0, &status, output); + break; + } else if (input->l_prime_read != PASS) { + if (use_fw && hdcp->config.debug.lc_enable_sw_fallback) { + adjust->hdcp2.force_sw_locality_check = true; + callback_in_ms(0, output); + break; + } + fail_and_restart_in_ms(0, &status, output); break; } else if (input->l_prime_validation != PASS) { @@ -205,6 +216,7 @@ enum mod_hdcp_status mod_hdcp_hdcp2_transition(struct mod_hdcp *hdcp, callback_in_ms(0, output); set_state_id(hdcp, output, H2_A3_EXCHANGE_KS_AND_TEST_FOR_REPEATER); break; + } case H2_A3_EXCHANGE_KS_AND_TEST_FOR_REPEATER: if (input->eks_prepare != PASS || input->eks_write != PASS) { @@ -498,14 +510,25 @@ enum mod_hdcp_status mod_hdcp_hdcp2_dp_transition(struct mod_hdcp *hdcp, callback_in_ms(0, output); set_state_id(hdcp, output, D2_A2_LOCALITY_CHECK); break; - case D2_A2_LOCALITY_CHECK: + case D2_A2_LOCALITY_CHECK: { + const bool use_fw = hdcp->config.ddc.funcs.atomic_write_poll_read_aux + && !adjust->hdcp2.force_sw_locality_check; + if (hdcp->state.stay_count > 10 || input->lc_init_prepare != PASS || - input->lc_init_write != PASS || - input->l_prime_read != PASS) { + (!use_fw && input->lc_init_write != PASS)) { /* 1A-12: consider invalid l' a failure */ fail_and_restart_in_ms(0, &status, output); break; + } else if (input->l_prime_read != PASS) { + if (use_fw && hdcp->config.debug.lc_enable_sw_fallback) { + adjust->hdcp2.force_sw_locality_check = true; + callback_in_ms(0, output); + break; + } + + fail_and_restart_in_ms(0, &status, output); + break; } else if (input->l_prime_validation != PASS) { callback_in_ms(0, output); increment_stay_counter(hdcp); @@ -514,6 +537,7 @@ enum mod_hdcp_status mod_hdcp_hdcp2_dp_transition(struct mod_hdcp *hdcp, callback_in_ms(0, output); set_state_id(hdcp, output, D2_A34_EXCHANGE_KS_AND_TEST_FOR_REPEATER); break; + } case D2_A34_EXCHANGE_KS_AND_TEST_FOR_REPEATER: if (input->eks_prepare != PASS || input->eks_write != PASS) { diff --git a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_ddc.c b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_ddc.c index 6e064e6ae949..2e6408579194 100644 --- a/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_ddc.c +++ b/drivers/gpu/drm/amd/display/modules/hdcp/hdcp_ddc.c @@ -688,3 +688,76 @@ enum mod_hdcp_status mod_hdcp_clear_cp_irq_status(struct mod_hdcp *hdcp) return MOD_HDCP_STATUS_INVALID_OPERATION; } + +static bool write_stall_read_lc_fw_aux(struct mod_hdcp *hdcp) +{ + struct mod_hdcp_message_hdcp2 *hdcp2 = &hdcp->auth.msg.hdcp2; + + struct mod_hdcp_atomic_op_aux write = { + hdcp_dpcd_addrs[MOD_HDCP_MESSAGE_ID_WRITE_LC_INIT], + hdcp2->lc_init + 1, + sizeof(hdcp2->lc_init) - 1, + }; + struct mod_hdcp_atomic_op_aux stall = { 0, NULL, 0, }; + struct mod_hdcp_atomic_op_aux read = { + hdcp_dpcd_addrs[MOD_HDCP_MESSAGE_ID_READ_LC_SEND_L_PRIME], + hdcp2->lc_l_prime + 1, + sizeof(hdcp2->lc_l_prime) - 1, + }; + + hdcp2->lc_l_prime[0] = HDCP_2_2_LC_SEND_LPRIME; + + return hdcp->config.ddc.funcs.atomic_write_poll_read_aux( + hdcp->config.ddc.handle, + &write, + &stall, + &read, + 16 * 1000, + 0 + ); +} + +static bool write_poll_read_lc_fw_i2c(struct mod_hdcp *hdcp) +{ + struct mod_hdcp_message_hdcp2 *hdcp2 = &hdcp->auth.msg.hdcp2; + uint8_t expected_rxstatus[2] = { sizeof(hdcp2->lc_l_prime) }; + + hdcp->buf[0] = hdcp_i2c_offsets[MOD_HDCP_MESSAGE_ID_WRITE_LC_INIT]; + memmove(&hdcp->buf[1], hdcp2->lc_init, sizeof(hdcp2->lc_init)); + + struct mod_hdcp_atomic_op_i2c write = { + HDCP_I2C_ADDR, + 0, + hdcp->buf, + sizeof(hdcp2->lc_init) + 1, + }; + struct mod_hdcp_atomic_op_i2c poll = { + HDCP_I2C_ADDR, + hdcp_i2c_offsets[MOD_HDCP_MESSAGE_ID_READ_RXSTATUS], + expected_rxstatus, + sizeof(expected_rxstatus), + }; + struct mod_hdcp_atomic_op_i2c read = { + HDCP_I2C_ADDR, + hdcp_i2c_offsets[MOD_HDCP_MESSAGE_ID_READ_LC_SEND_L_PRIME], + hdcp2->lc_l_prime, + sizeof(hdcp2->lc_l_prime), + }; + + return hdcp->config.ddc.funcs.atomic_write_poll_read_i2c( + hdcp->config.ddc.handle, + &write, + &poll, + &read, + 20 * 1000, + 6 + ); +} + +enum mod_hdcp_status mod_hdcp_write_poll_read_lc_fw(struct mod_hdcp *hdcp) +{ + const bool success = (is_dp_hdcp(hdcp) ? write_stall_read_lc_fw_aux : write_poll_read_lc_fw_i2c)(hdcp); + + return success ? MOD_HDCP_STATUS_SUCCESS : MOD_HDCP_STATUS_DDC_FAILURE; +} + diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h b/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h index a4d344a4db9e..c42468bb70ac 100644 --- a/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_hdcp.h @@ -133,9 +133,22 @@ enum mod_hdcp_display_disable_option { MOD_HDCP_DISPLAY_DISABLE_ENCRYPTION, }; +struct mod_hdcp_atomic_op_i2c { + uint8_t address; + uint8_t offset; + uint8_t *data; + uint32_t size; +}; + +struct mod_hdcp_atomic_op_aux { + uint32_t address; + uint8_t *data; + uint32_t size; +}; + struct mod_hdcp_ddc { void *handle; - struct { + struct mod_hdcp_ddc_funcs { bool (*read_i2c)(void *handle, uint32_t address, uint8_t offset, @@ -153,6 +166,22 @@ struct mod_hdcp_ddc { uint32_t address, const uint8_t *data, uint32_t size); + bool (*atomic_write_poll_read_i2c)( + void *handle, + const struct mod_hdcp_atomic_op_i2c *write, + const struct mod_hdcp_atomic_op_i2c *poll, + struct mod_hdcp_atomic_op_i2c *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb + ); + bool (*atomic_write_poll_read_aux)( + void *handle, + const struct mod_hdcp_atomic_op_aux *write, + const struct mod_hdcp_atomic_op_aux *poll, + struct mod_hdcp_atomic_op_aux *read, + uint32_t poll_timeout_us, + uint8_t poll_mask_msb + ); } funcs; }; @@ -185,7 +214,8 @@ struct mod_hdcp_link_adjustment_hdcp2 { uint8_t force_type : 2; uint8_t force_no_stored_km : 1; uint8_t increase_h_prime_timeout: 1; - uint8_t reserved : 3; + uint8_t force_sw_locality_check : 1; + uint8_t reserved : 2; }; struct mod_hdcp_link_adjustment { @@ -272,6 +302,10 @@ struct mod_hdcp_display_query { struct mod_hdcp_config { struct mod_hdcp_psp psp; struct mod_hdcp_ddc ddc; + struct { + uint8_t lc_enable_sw_fallback : 1; + uint8_t reserved : 7; + } debug; uint8_t index; }; diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index 4c95b885d1d0..c8eccee9b023 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -366,7 +366,7 @@ enum DC_DEBUG_MASK { DC_HDCP_LC_FORCE_FW_ENABLE = 0x80000, /** - * @DC_HDCP_LC_ENABLE_SW_FALLBACK If set, upon HDCP Locality Check FW + * @DC_HDCP_LC_ENABLE_SW_FALLBACK: If set, upon HDCP Locality Check FW * path failure, retry using legacy SW path. */ DC_HDCP_LC_ENABLE_SW_FALLBACK = 0x100000, diff --git a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_6_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_6_0_sh_mask.h index bd8085ec54ed..2d6a598a6c25 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_6_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_6_0_sh_mask.h @@ -5242,6 +5242,8 @@ #define DEGAMMA_CONTROL__CURSOR_DEGAMMA_MODE__SHIFT 0x0000000c #define DEGAMMA_CONTROL__GRPH_DEGAMMA_MODE_MASK 0x00000003L #define DEGAMMA_CONTROL__GRPH_DEGAMMA_MODE__SHIFT 0x00000000 +#define DEGAMMA_CONTROL__ICON_DEGAMMA_MODE_MASK 0x00000300L +#define DEGAMMA_CONTROL__ICON_DEGAMMA_MODE__SHIFT 0x00000008 #define DEGAMMA_CONTROL__OVL_DEGAMMA_MODE_MASK 0x00000030L #define DEGAMMA_CONTROL__OVL_DEGAMMA_MODE__SHIFT 0x00000004 #define DENORM_CONTROL__DENORM_MODE_MASK 0x00000007L diff --git a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_6_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_6_0_d.h index c75aee25619e..6f44345277af 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_6_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_6_0_d.h @@ -1779,6 +1779,8 @@ #define mmRLC_TTOP_D 0x3105 #define mmRLC_CLEAR_STATE_RESTORE_BASE 0x30C8 #define mmRLC_PG_AO_CU_MASK 0x310B +#define mmSPI_STATIC_THREAD_MGMT_1 0x2438 +#define mmSPI_STATIC_THREAD_MGMT_2 0x2439 #define mmSPI_STATIC_THREAD_MGMT_3 0x243A #endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_d.h index edc8a793a95d..4dd386b98748 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_d.h @@ -234,6 +234,26 @@ #define mmIH_RB_WPTR_ADDR_HI 0x0F84 #define mmIH_RB_WPTR_ADDR_LO 0x0F85 #define mmIH_STATUS 0x0F88 + +#define mmDMA_GFX_RB_CNTL 0x3400 +#define mmDMA_GFX_RB_BASE 0x3401 +#define mmDMA_GFX_RB_RPTR 0x3402 +#define mmDMA_GFX_RB_WPTR 0x3403 +#define mmDMA_GFX_RB_RPTR_ADDR_HI 0x3407 +#define mmDMA_GFX_RB_RPTR_ADDR_LO 0x3408 +#define mmDMA_GFX_IB_CNTL 0x3409 +#define mmDMA_GFX_IB_RPTR 0x340a +#define mmDMA_CNTL 0x340b +#define mmDMA_STATUS_REG 0x340D +#define mmDMA_TILING_CONFIG 0x342E +#define mmDMA_SEM_INCOMPLETE_TIMER_CNTL 0x3411 +#define mmDMA_SEM_WAIT_FAIL_TIMER_CNTL 0x3412 +#define mmDMA_POWER_CNTL 0x342F +#define mmDMA_CLK_CTRL 0x3430 +#define mmDMA_PG 0x3435 +#define mmDMA_PGFSM_CONFIG 0x3436 +#define mmDMA_PGFSM_WRITE 0x3437 + #define mmSEM_MAILBOX 0x0F9B #define mmSEM_MAILBOX_CLIENTCONFIG 0x0F9A #define mmSEM_MAILBOX_CONTROL 0x0F9C @@ -269,7 +289,4 @@ #define mmVCE_CONFIG 0x0F94 #define mmXDMA_MSTR_MEM_OVERFLOW_CNTL 0x03F8 -/* from the old sid.h */ -#define mmDMA_TILING_CONFIG 0x342E - #endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_sh_mask.h index 1c540fe136cb..9f7fc2428b69 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/oss/oss_1_0_sh_mask.h @@ -823,6 +823,43 @@ #define LX3__RESERVED__SHIFT 0x00000000 #define RINGOSC_MASK__MASK_MASK 0x0000ffffL #define RINGOSC_MASK__MASK__SHIFT 0x00000000 + +#define DMA_CNTL__TRAP_ENABLE_MASK 0x00000001L +#define DMA_CNTL__TRAP_ENABLE__SHIFT 0x00000000 +#define DMA_CNTL__SEM_INCOMPLETE_INT_ENABLE_MASK 0x00000002L +#define DMA_CNTL__SEM_INCOMPLETE_INT_ENABLE__SHIFT 0x00000001 +#define DMA_CNTL__SEM_WAIT_INT_ENABLE_MASK 0x00000004L +#define DMA_CNTL__SEM_WAIT_INT_ENABLE__SHIFT 0x00000002 +#define DMA_CNTL__DATA_SWAP_ENABLE_MASK 0x00000008L +#define DMA_CNTL__DATA_SWAP_ENABLE__SHIFT 0x00000003 +#define DMA_CNTL__FENCE_SWAP_ENABLE_MASK 0x00000010L +#define DMA_CNTL__FENCE_SWAP_ENABLE__SHIFT 0x00000004 +#define DMA_CNTL__CTXEMPTY_INT_ENABLE_MASK 0x10000000L +#define DMA_CNTL__CTXEMPTY_INT_ENABLE__SHIFT 0x0000001C +#define DMA_GFX_RB_CNTL__RB_ENABLE_MASK 0x00000001L +#define DMA_GFX_RB_CNTL__RB_ENABLE__SHIFT 0x00000000 +#define DMA_GFX_RB_CNTL__RB_SIZE__SHIFT 0x00000001 +#define DMA_GFX_RB_CNTL__RB_SWAP_ENABLE_MASK 0x00000200L +#define DMA_GFX_RB_CNTL__RB_SWAP_ENABLE__SHIFT 0x00000009 +#define DMA_GFX_RB_CNTL__RPTR_WRITEBACK_ENABLE_MASK 0x00001000L +#define DMA_GFX_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT 0x0000000C +#define DMA_GFX_RB_CNTL__RPTR_WRITEBACK_SWAP_ENABLE_MASK 0x00002000L +#define DMA_GFX_RB_CNTL__RPTR_WRITEBACK_SWAP_ENABLE__SHIFT 0x0000000D +#define DMA_GFX_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT 0x00000010 +#define DMA_GFX_IB_CNTL__IB_ENABLE_MASK 0x00000001L +#define DMA_GFX_IB_CNTL__IB_ENABLE__SHIFT 0x00000000 +#define DMA_GFX_IB_CNTL__IB_SWAP_ENABLE_MASK 0x00000010L +#define DMA_GFX_IB_CNTL__IB_SWAP_ENABLE__SHIFT 0x00000004 +#define DMA_GFX_IB_CNTL__CMD_VMID_FORCE_MASK 0x80000000L +#define DMA_GFX_IB_CNTL__CMD_VMID_FORCE__SHIFT 0x0000001F + +#define DMA_STATUS_REG__IDLE_MASK 0x00000001L +#define DMA_STATUS_REG__IDLE__SHIFT 0x00000000 +#define DMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK 0x00000100L +#define DMA_POWER_CNTL__MEM_POWER_OVERRIDE__SHIFT 0x00000008 +#define DMA_PG__PG_CNTL_ENABLE_MASK 0x00000001L +#define DMA_PG__PG_CNTL_ENABLE__SHIFT 0x00000000 + #define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT0_MASK 0x00000007L #define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT0__SHIFT 0x00000000 #define SEM_MAILBOX_CLIENTCONFIG__CP_CLIENT1_MASK 0x00000038L @@ -1015,6 +1052,10 @@ #define SRBM_STATUS2__VCE_BUSY__SHIFT 0x00000007 #define SRBM_STATUS2__VCE_RQ_PENDING_MASK 0x00000008L #define SRBM_STATUS2__VCE_RQ_PENDING__SHIFT 0x00000003 +#define SRBM_STATUS2__DMA_BUSY_MASK 0x00000020L +#define SRBM_STATUS2__DMA_BUSY__SHIFT 0x00000005 +#define SRBM_STATUS2__DMA1_BUSY_MASK 0x00000040L +#define SRBM_STATUS2__DMA1_BUSY__SHIFT 0x00000006 #define SRBM_STATUS2__XDMA_BUSY_MASK 0x00000100L #define SRBM_STATUS2__XDMA_BUSY__SHIFT 0x00000008 #define SRBM_STATUS2__XSP_BUSY_MASK 0x00000010L diff --git a/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_d.h index 6b10be61efc3..bdef1f743df7 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_d.h @@ -41,7 +41,49 @@ #define ixLCAC_MC5_CNTL 0x012B #define ixLCAC_MC5_OVR_SEL 0x012C #define ixLCAC_MC5_OVR_VAL 0x012D + +#define mmCG_SPLL_FUNC_CNTL 0x0180 +#define mmCG_SPLL_FUNC_CNTL_2 0x0181 +#define mmCG_SPLL_FUNC_CNTL_3 0x0182 +#define mmCG_SPLL_FUNC_CNTL_4 0x0183 +#define mmCG_SPLL_STATUS 0x0185 +#define mmSPLL_CNTL_MODE 0x0186 +#define mmCG_SPLL_SPREAD_SPECTRUM 0x0188 +#define mmCG_SPLL_SPREAD_SPECTRUM_2 0x0189 +#define mmCG_SPLL_AUTOSCALE_CNTL 0x018B +#define mmMPLL_BYPASSCLK_SEL 0x0197 +#define mmCG_CLKPIN_CNTL 0x0198 +#define mmCG_CLKPIN_CNTL_2 0x0199 +#define mmTHM_CLK_CNTL 0x019B +#define mmMISC_CLK_CNTL 0x019C +#define mmCG_THERMAL_CTRL 0x01C0 +#define mmCG_THERMAL_STATUS 0x01C1 +#define mmCG_THERMAL_INT 0x01C2 +#define mmCG_MULT_THERMAL_CTRL 0x01C4 +#define mmCG_MULT_THERMAL_STATUS 0x01C5 +#define mmCG_FDO_CTRL0 0x01D5 +#define mmCG_FDO_CTRL1 0x01D6 +#define mmCG_FDO_CTRL2 0x01D7 +#define mmCG_TACH_CTRL 0x01DC +#define mmCG_TACH_STATUS 0x01DD +#define mmGENERAL_PWRMGT 0x1E0 +#define mmCG_TPC 0x1E1 +#define mmSCLK_PWRMGT_CNTL 0x1E2 +#define mmTARGET_AND_CURRENT_PROFILE_INDEX 0x01E6 +#define mmCG_FTV 0x01EF +#define mmCG_FFCT_0 0x01F0 +#define mmCG_BSP 0x01FF +#define mmCG_AT 0x0200 +#define mmCG_GIT 0x0201 +#define mmCG_SSP 0x0203 +#define mmCG_DISPLAY_GAP_CNTL 0x020A +#define mmCG_ULV_CONTROL 0x021E +#define mmCG_ULV_PARAMETER 0x021F +#define mmSMC_SCRATCH0 0x0221 +#define mmCG_CAC_CTRL 0x022E + #define ixSMC_PC_C 0x80000370 + #define ixTHM_TMON0_DEBUG 0x03F0 #define ixTHM_TMON0_INT_DATA 0x0380 #define ixTHM_TMON0_RDIL0_DATA 0x0300 @@ -110,6 +152,7 @@ #define ixTHM_TMON1_RDIR7_DATA 0x0337 #define ixTHM_TMON1_RDIR8_DATA 0x0338 #define ixTHM_TMON1_RDIR9_DATA 0x0339 + #define mmGPIOPAD_A 0x05E7 #define mmGPIOPAD_EN 0x05E8 #define mmGPIOPAD_EXTERN_TRIG_CNTL 0x05F1 @@ -127,6 +170,7 @@ #define mmGPIOPAD_STRENGTH 0x05E5 #define mmGPIOPAD_SW_INT_STAT 0x05E4 #define mmGPIOPAD_Y 0x05E9 + #define mmSMC_IND_ACCESS_CNTL 0x008A #define mmSMC_IND_DATA_0 0x0081 #define mmSMC_IND_DATA 0x0081 diff --git a/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_sh_mask.h index 7d3925b7266e..67d3c7e13a48 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/smu/smu_6_0_sh_mask.h @@ -23,10 +23,142 @@ #ifndef SMU_6_0_SH_MASK_H #define SMU_6_0_SH_MASK_H -#define CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV_MASK 0x03ffffffL -#define CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV__SHIFT 0x00000000 -#define CG_SPLL_FUNC_CNTL__SPLL_REF_DIV_MASK 0x000003f0L +#define CG_AT__CG_R_MASK 0x0000FFFFL +#define CG_AT__CG_R__SHIFT 0x00000000 +#define CG_AT__CG_L_MASK 0xFFFF0000L +#define CG_AT__CG_L__SHIFT 0x00000010 + +#define CG_BSP__BSP_MASK 0x0000FFFFL +#define CG_BSP__BSP__SHIFT 0x00000000 +#define CG_BSP__BSU_MASK 0x000F0000L +#define CG_BSP__BSU__SHIFT 0x00000010 + +#define CG_CAC_CTRL__CAC_WINDOW_MASK 0x00FFFFFFL +#define CG_CAC_CTRL__CAC_WINDOW__SHIFT 0x00000000 + +#define CG_CLKPIN_CNTL__XTALIN_DIVIDE_MASK 0x00000002L +#define CG_CLKPIN_CNTL__XTALIN_DIVIDE__SHIFT 0x00000001 +#define CG_CLKPIN_CNTL__BCLK_AS_XCLK_MASK 0x00000004L +#define CG_CLKPIN_CNTL__BCLK_AS_XCLK__SHIFT 0x00000002 +#define CG_CLKPIN_CNTL_2__FORCE_BIF_REFCLK_EN_MASK 0x00000008L +#define CG_CLKPIN_CNTL_2__FORCE_BIF_REFCLK_EN__SHIFT 0x00000003 +#define CG_CLKPIN_CNTL_2__MUX_TCLK_TO_XCLK_MASK 0x00000100L +#define CG_CLKPIN_CNTL_2__MUX_TCLK_TO_XCLK__SHIFT 0x00000008 + +#define CG_DISPLAY_GAP_CNTL__DISP1_GAP_MASK 0x00000003L +#define CG_DISPLAY_GAP_CNTL__DISP1_GAP__SHIFT 0x00000000 +#define CG_DISPLAY_GAP_CNTL__DISP2_GAP_MASK 0x0000000CL +#define CG_DISPLAY_GAP_CNTL__DISP2_GAP__SHIFT 0x00000002 +#define CG_DISPLAY_GAP_CNTL__VBI_TIMER_COUNT_MASK 0x0003FFF0L +#define CG_DISPLAY_GAP_CNTL__VBI_TIMER_COUNT__SHIFT 0x00000004 +#define CG_DISPLAY_GAP_CNTL__VBI_TIMER_UNIT_MASK 0x00700000 +#define CG_DISPLAY_GAP_CNTL__VBI_TIMER_UNIT__SHIFT 0x00000014 +#define CG_DISPLAY_GAP_CNTL__DISP1_GAP_MCHG_MASK 0x03000000L +#define CG_DISPLAY_GAP_CNTL__DISP1_GAP_MCHG__SHIFT 0x00000018 +#define CG_DISPLAY_GAP_CNTL__DISP2_GAP_MCHG_MASK 0x0C000000L +#define CG_DISPLAY_GAP_CNTL__DISP2_GAP_MCHG__SHIFT 0x0000001A + +#define CG_FFCT_0__UTC_0_MASK 0x000003FFL +#define CG_FFCT_0__UTC_0__SHIFT 0x00000000 +#define CG_FFCT_0__DTC_0_MASK 0x000FFC00L +#define CG_FFCT_0__DTC_0__SHIFT 0x0000000A + +#define CG_GIT__CG_GICST_MASK 0x0000FFFFL +#define CG_GIT__CG_GICST__SHIFT 0x00000000 +#define CG_GIT__CG_GIPOT_MASK 0xFFFF0000L +#define CG_GIT__CG_GIPOT__SHIFT 0x00000010 + +#define CG_SPLL_FUNC_CNTL__SPLL_RESET_MASK 0x00000001L +#define CG_SPLL_FUNC_CNTL__SPLL_RESET__SHIFT 0x00000000 +#define CG_SPLL_FUNC_CNTL__SPLL_SLEEP_MASK 0x00000002L +#define CG_SPLL_FUNC_CNTL__SPLL_SLEEP__SHIFT 0x00000001 +#define CG_SPLL_FUNC_CNTL__SPLL_BYPASS_EN_MASK 0x00000008L +#define CG_SPLL_FUNC_CNTL__SPLL_BYPASS_EN__SHIFT 0x00000003 +#define CG_SPLL_FUNC_CNTL__SPLL_REF_DIV_MASK 0x000003F0L #define CG_SPLL_FUNC_CNTL__SPLL_REF_DIV__SHIFT 0x00000004 +#define CG_SPLL_FUNC_CNTL__SPLL_PDIV_A_MASK 0x007F00000 +#define CG_SPLL_FUNC_CNTL__SPLL_PDIV_A__SHIFT 0x00000014 +#define CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL_MASK 0x0000001FF +#define CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL__SHIFT 0x00000000 +#define CG_SPLL_FUNC_CNTL_2__SPLL_CTLREQ_CHG_MASK 0x00800000 +#define CG_SPLL_FUNC_CNTL_2__SPLL_CTLREQ_CHG__SHIFT 0x00000017 +#define CG_SPLL_FUNC_CNTL_2__SCLK_MUX_UPDATE_MASK 0x04000000 +#define CG_SPLL_FUNC_CNTL_2__SCLK_MUX_UPDATE__SHIFT 0x0000001A +#define CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV_MASK 0x03FFFFFFL +#define CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV__SHIFT 0x00000000 +#define CG_SPLL_FUNC_CNTL_3__SPLL_DITHEN_MASK 0x10000000L +#define CG_SPLL_FUNC_CNTL_3__SPLL_DITHEN__SHIFT 0x0000001C +#define CG_SPLL_STATUS__SPLL_CHG_STATUS_MASK 0x00000002L +#define CG_SPLL_STATUS__SPLL_CHG_STATUS__SHIFT 0x00000001 +#define CG_SPLL_SPREAD_SPECTRUM__SSEN_MASK 0x00000001L +#define CG_SPLL_SPREAD_SPECTRUM__SSEN__SHIFT 0x00000000 +#define CG_SPLL_SPREAD_SPECTRUM__CLK_S_MASK 0x0000FFF0L +#define CG_SPLL_SPREAD_SPECTRUM__CLK_S__SHIFT 0x00000004 +#define CG_SPLL_SPREAD_SPECTRUM_2__CLK_V_MASK 0x00000200L +#define CG_SPLL_SPREAD_SPECTRUM_2__CLK_V__SHIFT 0x00000000 +#define CG_SPLL_AUTOSCALE_CNTL__AUTOSCALE_ON_SS_CLEAR_MASK 0x03FFFFFFL +#define CG_SPLL_AUTOSCALE_CNTL__AUTOSCALE_ON_SS_CLEAR__SHIFT 0x00000009 + +#define CG_SSP__SST_MASK 0x0000FFFFL +#define CG_SSP__SST__SHIFT 0x00000000 +#define CG_SSP__SSTU_MASK 0x000F0000L +#define CG_SSP__SSTU__SHIFT 0x00000010 + +#define CG_THERMAL_CTRL__DPM_EVENT_SRC_MASK 0x00000007L +#define CG_THERMAL_CTRL__DPM_EVENT_SRC__SHIFT 0x00000000 +#define CG_THERMAL_CTRL__DIG_THERM_DPM_MASK 0x003FC000 +#define CG_THERMAL_CTRL__DIG_THERM_DPM__SHIFT 0x0000000E +#define CG_THERMAL_STATUS__FDO_PWM_DUTY_MASK 0x0001FE00L +#define CG_THERMAL_STATUS__FDO_PWM_DUTY__SHIFT 0x00000009 +#define CG_THERMAL_INT__DIG_THERM_INTH_MASK 0x0000FF00L +#define CG_THERMAL_INT__DIG_THERM_INTH__SHIFT 0x00000008 +#define CG_THERMAL_INT__DIG_THERM_INTL_MASK 0x00FF0000L +#define CG_THERMAL_INT__DIG_THERM_INTL__SHIFT 0x00000010 +#define CG_THERMAL_INT__THERM_INT_MASK_HIGH_MASK 0x01000000L +#define CG_THERMAL_INT__THERM_INT_MASK_HIGH__SHIFT 0x00000018 +#define CG_THERMAL_INT__THERM_INT_MASK_LOW_MASK 0x02000000 +#define CG_THERMAL_INT__THERM_INT_MASK_LOW__SHIFT 0x00000019 + +#define CG_MULT_THERMAL_CTRL__TEMP_SEL_MASK 0x0FF00000L +#define CG_MULT_THERMAL_CTRL__TEMP_SEL__SHIFT 0x00000014 +#define CG_MULT_THERMAL_STATUS__ASIC_MAX_TEMP_MASK 0x000001FFL +#define CG_MULT_THERMAL_STATUS__ASIC_MAX_TEMP__SHIFT 0x00000000 +#define CG_MULT_THERMAL_STATUS__CTF_TEMP_MASK 0x0003fe00L +#define CG_MULT_THERMAL_STATUS__CTF_TEMP__SHIFT 0x00000009 + +#define CG_FDO_CTRL0__FDO_STATIC_DUTY_MASK 0x000000FFL +#define CG_FDO_CTRL0__FDO_STATIC_DUTY__SHIFT 0x00000000 +#define CG_FDO_CTRL1__FMAX_DUTY100_MASK 0x000000FFL +#define CG_FDO_CTRL1__FMAX_DUTY100__SHIFT 0x00000000 +#define CG_FDO_CTRL2__TMIN_MASK 0x000000FFL +#define CG_FDO_CTRL2__TMIN__SHIFT 0x00000000 +#define CG_FDO_CTRL2__FDO_PWM_MODE_MASK 0x00003800L +#define CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT 0x0000000B +#define CG_FDO_CTRL2__TACH_PWM_RESP_RATE_MASK 0xFE000000L +#define CG_FDO_CTRL2__TACH_PWM_RESP_RATE__SHIFT 0x00000019 + +#define CG_TACH_CTRL__EDGE_PER_REV_MASK 0x00000007L +#define CG_TACH_CTRL__EDGE_PER_REV__SHIFT 0x00000000 +#define CG_TACH_CTRL__TARGET_PERIOD_MASK 0xFFFFFFF8L +#define CG_TACH_CTRL__TARGET_PERIOD__SHIFT 0x00000003 +#define CG_TACH_STATUS__TACH_PERIOD_MASK 0xFFFFFFFFL +#define CG_TACH_STATUS__TACH_PERIOD__SHIFT 0x00000000 + +#define GENERAL_PWRMGT__GLOBAL_PWRMGT_EN_MASK 0x00000001L +#define GENERAL_PWRMGT__GLOBAL_PWRMGT_EN__SHIFT 0x00000000 +#define GENERAL_PWRMGT__STATIC_PM_EN_MASK 0x00000002L +#define GENERAL_PWRMGT__STATIC_PM_EN__SHIFT 0x00000001 +#define GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK 0x00000004L +#define GENERAL_PWRMGT__THERMAL_PROTECTION_DIS__SHIFT 0x00000002 +#define GENERAL_PWRMGT__THERMAL_PROTECTION_TYPE_MASK 0x00000008L +#define GENERAL_PWRMGT__THERMAL_PROTECTION_TYPE__SHIFT 0x00000003 +#define GENERAL_PWRMGT__SW_SMIO_INDEX_MASK 0x00000040L +#define GENERAL_PWRMGT__SW_SMIO_INDEX__SHIFT 0x00000006 +#define GENERAL_PWRMGT__VOLT_PWRMGT_EN_MASK 0x00000400L +#define GENERAL_PWRMGT__VOLT_PWRMGT_EN__SHIFT 0x0000000A +#define GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN_MASK 0x00800000L +#define GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN__SHIFT 0x00000017 + #define GPIOPAD_A__GPIO_A_MASK 0x7fffffffL #define GPIOPAD_A__GPIO_A__SHIFT 0x00000000 #define GPIOPAD_EN__GPIO_EN_MASK 0x7fffffffL @@ -195,6 +327,7 @@ #define GPIOPAD_SW_INT_STAT__SW_INT_STAT__SHIFT 0x00000000 #define GPIOPAD_Y__GPIO_Y_MASK 0x7fffffffL #define GPIOPAD_Y__GPIO_Y__SHIFT 0x00000000 + #define LCAC_MC0_CNTL__MC0_ENABLE_MASK 0x00000001L #define LCAC_MC0_CNTL__MC0_ENABLE__SHIFT 0x00000000 #define LCAC_MC0_CNTL__MC0_THRESHOLD_MASK 0x0001fffeL @@ -243,6 +376,37 @@ #define LCAC_MC5_OVR_SEL__MC5_OVR_SEL__SHIFT 0x00000000 #define LCAC_MC5_OVR_VAL__MC5_OVR_VAL_MASK 0xffffffffL #define LCAC_MC5_OVR_VAL__MC5_OVR_VAL__SHIFT 0x00000000 + +#define MPLL_BYPASSCLK_SEL__MPLL_CLKOUT_SEL_MASK 0x0000FF00L +#define MPLL_BYPASSCLK_SEL__MPLL_CLKOUT_SEL__SHIFT 0x00000008 + +#define SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF_MASK 0x00000001L +#define SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF__SHIFT 0x00000000 +#define SCLK_PWRMGT_CNTL__SCLK_LOW_D1_MASK 0x00000002L +#define SCLK_PWRMGT_CNTL__SCLK_LOW_D1__SHIFT 0x00000001 +#define SCLK_PWRMGT_CNTL__FIR_RESET_MASK 0x00000010L +#define SCLK_PWRMGT_CNTL__FIR_RESET__SHIFT 0x00000004 +#define SCLK_PWRMGT_CNTL__FIR_FORCE_TREND_SEL_MASK 0x00000020L +#define SCLK_PWRMGT_CNTL__FIR_FORCE_TREND_SEL__SHIFT 0x00000005 +#define SCLK_PWRMGT_CNTL__FIR_TREND_MODE_MASK 0x00000040L +#define SCLK_PWRMGT_CNTL__FIR_TREND_MODE__SHIFT 0x00000006 +#define SCLK_PWRMGT_CNTL__DYN_GFX_CLK_OFF_EN_MASK 0x00000080L +#define SCLK_PWRMGT_CNTL__DYN_GFX_CLK_OFF_EN__SHIFT 0x00000007 +#define SCLK_PWRMGT_CNTL__GFX_CLK_FORCE_ON_MASK 0x00000100L +#define SCLK_PWRMGT_CNTL__GFX_CLK_FORCE_ON__SHIFT 0x00000008 +#define SCLK_PWRMGT_CNTL__GFX_CLK_REQUEST_OFF_MASK 0x00000200L +#define SCLK_PWRMGT_CNTL__GFX_CLK_REQUEST_OFF__SHIFT 0x00000009 +#define SCLK_PWRMGT_CNTL__GFX_CLK_FORCE_OFF_MASK 0x00000400L +#define SCLK_PWRMGT_CNTL__GFX_CLK_FORCE_OFF__SHIFT 0x0000000A +#define SCLK_PWRMGT_CNTL__GFX_CLK_OFF_ACPI_D1_MASK 0x00000800L +#define SCLK_PWRMGT_CNTL__GFX_CLK_OFF_ACPI_D1__SHIFT 0x0000000B +#define SCLK_PWRMGT_CNTL__GFX_CLK_OFF_ACPI_D2_MASK 0x00001000L +#define SCLK_PWRMGT_CNTL__GFX_CLK_OFF_ACPI_D2__SHIFT 0x0000000C +#define SCLK_PWRMGT_CNTL__GFX_CLK_OFF_ACPI_D3_MASK 0x00002000L +#define SCLK_PWRMGT_CNTL__GFX_CLK_OFF_ACPI_D3__SHIFT 0x0000000D +#define SCLK_PWRMGT_CNTL__DYN_LIGHT_SLEEP_EN_MASK 0x00004000L +#define SCLK_PWRMGT_CNTL__DYN_LIGHT_SLEEP_EN__SHIFT 0x0000000E + #define SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK 0x00000001L #define SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0__SHIFT 0x00000000 #define SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_1_MASK 0x00000100L @@ -285,6 +449,7 @@ #define SMC_RESP_1__SMC_RESP__SHIFT 0x00000000 #define SMC_RESP_2__SMC_RESP_MASK 0xffffffffL #define SMC_RESP_2__SMC_RESP__SHIFT 0x00000000 + #define SPLL_CNTL_MODE__SPLL_CTLREQ_DLY_CNT_MASK 0x000ff000L #define SPLL_CNTL_MODE__SPLL_CTLREQ_DLY_CNT__SHIFT 0x0000000c #define SPLL_CNTL_MODE__SPLL_ENSAT_MASK 0x00000010L @@ -293,6 +458,8 @@ #define SPLL_CNTL_MODE__SPLL_FASTEN__SHIFT 0x00000003 #define SPLL_CNTL_MODE__SPLL_LEGACY_PDIV_MASK 0x00000002L #define SPLL_CNTL_MODE__SPLL_LEGACY_PDIV__SHIFT 0x00000001 +#define SPLL_CNTL_MODE__SPLL_REFCLK_SEL_MASK 0x0C000000L +#define SPLL_CNTL_MODE__SPLL_REFCLK_SEL__SHIFT 0x0000001A #define SPLL_CNTL_MODE__SPLL_RESET_EN_MASK 0x10000000L #define SPLL_CNTL_MODE__SPLL_RESET_EN__SHIFT 0x0000001c #define SPLL_CNTL_MODE__SPLL_SW_DIR_CONTROL_MASK 0x00000001L @@ -303,10 +470,25 @@ #define SPLL_CNTL_MODE__SPLL_TEST__SHIFT 0x00000002 #define SPLL_CNTL_MODE__SPLL_VCO_MODE_MASK 0x60000000L #define SPLL_CNTL_MODE__SPLL_VCO_MODE__SHIFT 0x0000001d + #define TARGET_AND_CURRENT_PROFILE_INDEX_1__CURR_PCIE_INDEX_MASK 0x0f000000L #define TARGET_AND_CURRENT_PROFILE_INDEX_1__CURR_PCIE_INDEX__SHIFT 0x00000018 #define TARGET_AND_CURRENT_PROFILE_INDEX_1__TARG_PCIE_INDEX_MASK 0xf0000000L #define TARGET_AND_CURRENT_PROFILE_INDEX_1__TARG_PCIE_INDEX__SHIFT 0x0000001c + +#define TARGET_AND_CURRENT_PROFILE_INDEX__CURRENT_STATE_INDEX_MASK 0x000000F0L +#define TARGET_AND_CURRENT_PROFILE_INDEX__CURRENT_STATE_INDEX__SHIFT 0x00000004 + +#define THM_CLK_CNTL__CMON_CLK_SEL_MASK 0x000000FFL +#define THM_CLK_CNTL__CMON_CLK_SEL__SHIFT 0x00000000 +#define THM_CLK_CNTL__TMON_CLK_SEL_MASK 0x0000FF00L +#define THM_CLK_CNTL__TMON_CLK_SEL__SHIFT 0x00000008 + +#define MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL_MASK 0x000000FFL +#define MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL__SHIFT 0x00000000 +#define MISC_CLK_CNTL__ZCLK_SEL_MASK 0x0000FF00L +#define MISC_CLK_CNTL__ZCLK_SEL__SHIFT 0x00000008 + #define THM_TMON0_DEBUG__DEBUG_RDI_MASK 0x0000001fL #define THM_TMON0_DEBUG__DEBUG_RDI__SHIFT 0x00000000 #define THM_TMON0_DEBUG__DEBUG_Z_MASK 0x0000ffe0L diff --git a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_offset.h index 14574112c469..c4aaa86a95e2 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_offset.h @@ -1147,6 +1147,22 @@ #define regUVD_DPG_LMA_CTL2_BASE_IDX 1 +// addressBlock: uvd_mmsch_dec +// base address: 0x20d2c +#define regMMSCH_VF_VMID 0x054b +#define regMMSCH_VF_VMID_BASE_IDX 1 +#define regMMSCH_VF_CTX_ADDR_LO 0x054c +#define regMMSCH_VF_CTX_ADDR_LO_BASE_IDX 1 +#define regMMSCH_VF_CTX_ADDR_HI 0x054d +#define regMMSCH_VF_CTX_ADDR_HI_BASE_IDX 1 +#define regMMSCH_VF_CTX_SIZE 0x054e +#define regMMSCH_VF_CTX_SIZE_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_HOST 0x0552 +#define regMMSCH_VF_MAILBOX_HOST_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_RESP 0x0553 +#define regMMSCH_VF_MAILBOX_RESP_BASE_IDX 1 + + // addressBlock: uvd_vcn_umsch_dec // base address: 0x21500 #define regVCN_UMSCH_MES_CNTL 0x0740 diff --git a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_sh_mask.h index 5c119a6b87fb..bd7242e4e9c6 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_5_0_0_sh_mask.h @@ -5929,6 +5929,29 @@ #define UVD_DPG_LMA_CTL2__JPEG_WRITE_PTR_MASK 0x0000FE00L +// addressBlock: uvd_mmsch_dec +//MMSCH_VF_VMID +#define MMSCH_VF_VMID__VF_CTX_VMID__SHIFT 0x0 +#define MMSCH_VF_VMID__VF_GPCOM_VMID__SHIFT 0x5 +#define MMSCH_VF_VMID__VF_CTX_VMID_MASK 0x0000001FL +#define MMSCH_VF_VMID__VF_GPCOM_VMID_MASK 0x000003E0L +//MMSCH_VF_CTX_ADDR_LO +#define MMSCH_VF_CTX_ADDR_LO__VF_CTX_ADDR_LO__SHIFT 0x6 +#define MMSCH_VF_CTX_ADDR_LO__VF_CTX_ADDR_LO_MASK 0xFFFFFFC0L +//MMSCH_VF_CTX_ADDR_HI +#define MMSCH_VF_CTX_ADDR_HI__VF_CTX_ADDR_HI__SHIFT 0x0 +#define MMSCH_VF_CTX_ADDR_HI__VF_CTX_ADDR_HI_MASK 0xFFFFFFFFL +//MMSCH_VF_CTX_SIZE +#define MMSCH_VF_CTX_SIZE__VF_CTX_SIZE__SHIFT 0x0 +#define MMSCH_VF_CTX_SIZE__VF_CTX_SIZE_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_HOST +#define MMSCH_VF_MAILBOX_HOST__DATA__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_HOST__DATA_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_RESP +#define MMSCH_VF_MAILBOX_RESP__RESP__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_RESP__RESP_MASK 0xFFFFFFFFL + + // addressBlock: uvd_vcn_umsch_dec //VCN_UMSCH_MES_CNTL #define VCN_UMSCH_MES_CNTL__PIPE_ID__SHIFT 0x0 diff --git a/drivers/gpu/drm/amd/include/atombios.h b/drivers/gpu/drm/amd/include/atombios.h index b78360a71bc9..52bac19fb404 100644 --- a/drivers/gpu/drm/amd/include/atombios.h +++ b/drivers/gpu/drm/amd/include/atombios.h @@ -4308,7 +4308,7 @@ typedef struct _ATOM_DPCD_INFO // note2: From RV770, the memory is more than 32bit addressable, so we will change // ucTableFormatRevision=1,ucTableContentRevision=4, the strcuture remains // exactly same as 1.1 and 1.2 (1.3 is never in use), but ulStartAddrUsedByFirmware -// (in offset to start of memory address) is KB aligned instead of byte aligend. +// (in offset to start of memory address) is KB aligned instead of byte aligned. // Note3: /* If we change usReserved to "usFBUsedbyDrvInKB", then to VBIOS this usFBUsedbyDrvInKB is a predefined, unchanged constant across VGA or non VGA adapter, diff --git a/drivers/gpu/drm/amd/include/atomfirmware.h b/drivers/gpu/drm/amd/include/atomfirmware.h index 0160d65f3f5e..2d1135bdc4b9 100644 --- a/drivers/gpu/drm/amd/include/atomfirmware.h +++ b/drivers/gpu/drm/amd/include/atomfirmware.h @@ -183,6 +183,7 @@ enum atom_dgpu_vram_type { ATOM_DGPU_VRAM_TYPE_HBM2E = 0x61, ATOM_DGPU_VRAM_TYPE_GDDR6 = 0x70, ATOM_DGPU_VRAM_TYPE_HBM3 = 0x80, + ATOM_DGPU_VRAM_TYPE_HBM3E = 0x81, }; enum atom_dp_vs_preemph_def{ diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h index 21dc956b5f35..0f7542d7074b 100644 --- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h @@ -128,6 +128,7 @@ enum amd_pp_sensors { AMDGPU_PP_SENSOR_CPU_CLK, AMDGPU_PP_SENSOR_VDDNB, AMDGPU_PP_SENSOR_VDDGFX, + AMDGPU_PP_SENSOR_VDDBOARD, AMDGPU_PP_SENSOR_UVD_VCLK, AMDGPU_PP_SENSOR_UVD_DCLK, AMDGPU_PP_SENSOR_VCE_ECCLK, diff --git a/drivers/gpu/drm/amd/include/v11_structs.h b/drivers/gpu/drm/amd/include/v11_structs.h index f8008270f813..3728389fc3be 100644 --- a/drivers/gpu/drm/amd/include/v11_structs.h +++ b/drivers/gpu/drm/amd/include/v11_structs.h @@ -535,8 +535,8 @@ struct v11_gfx_mqd { uint32_t reserved_507; // offset: 507 (0x1FB) uint32_t reserved_508; // offset: 508 (0x1FC) uint32_t reserved_509; // offset: 509 (0x1FD) - uint32_t reserved_510; // offset: 510 (0x1FE) - uint32_t reserved_511; // offset: 511 (0x1FF) + uint32_t fence_address_lo; // offset: 510 (0x1FE) + uint32_t fence_address_hi; // offset: 511 (0x1FF) }; struct v11_sdma_mqd { @@ -1118,8 +1118,8 @@ struct v11_compute_mqd { uint32_t reserved_443; // offset: 443 (0x1BB) uint32_t reserved_444; // offset: 444 (0x1BC) uint32_t reserved_445; // offset: 445 (0x1BD) - uint32_t reserved_446; // offset: 446 (0x1BE) - uint32_t reserved_447; // offset: 447 (0x1BF) + uint32_t fence_address_lo; // offset: 446 (0x1BE) + uint32_t fence_address_hi; // offset: 447 (0x1BF) uint32_t gws_0_val; // offset: 448 (0x1C0) uint32_t gws_1_val; // offset: 449 (0x1C1) uint32_t gws_2_val; // offset: 450 (0x1C2) diff --git a/drivers/gpu/drm/amd/include/v12_structs.h b/drivers/gpu/drm/amd/include/v12_structs.h index 5eabab611b02..03a35f8a65b0 100644 --- a/drivers/gpu/drm/amd/include/v12_structs.h +++ b/drivers/gpu/drm/amd/include/v12_structs.h @@ -535,8 +535,8 @@ struct v12_gfx_mqd { uint32_t reserved_507; // offset: 507 (0x1FB) uint32_t reserved_508; // offset: 508 (0x1FC) uint32_t reserved_509; // offset: 509 (0x1FD) - uint32_t reserved_510; // offset: 510 (0x1FE) - uint32_t reserved_511; // offset: 511 (0x1FF) + uint32_t fence_address_lo; // offset: 510 (0x1FE) + uint32_t fence_address_hi; // offset: 511 (0x1FF) }; struct v12_sdma_mqd { @@ -1118,8 +1118,8 @@ struct v12_compute_mqd { uint32_t reserved_443; // offset: 443 (0x1BB) uint32_t reserved_444; // offset: 444 (0x1BC) uint32_t reserved_445; // offset: 445 (0x1BD) - uint32_t reserved_446; // offset: 446 (0x1BE) - uint32_t reserved_447; // offset: 447 (0x1BF) + uint32_t fence_address_lo; // offset: 446 (0x1BE) + uint32_t fence_address_hi; // offset: 447 (0x1BF) uint32_t gws_0_val; // offset: 448 (0x1C0) uint32_t gws_1_val; // offset: 449 (0x1C1) uint32_t gws_2_val; // offset: 450 (0x1C2) diff --git a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c index 3533d43ed1e7..2148c8db5a59 100644 --- a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c +++ b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c @@ -329,6 +329,34 @@ int amdgpu_dpm_mode1_reset(struct amdgpu_device *adev) return ret; } +bool amdgpu_dpm_is_link_reset_supported(struct amdgpu_device *adev) +{ + struct smu_context *smu = adev->powerplay.pp_handle; + bool support_link_reset = false; + + if (is_support_sw_smu(adev)) { + mutex_lock(&adev->pm.mutex); + support_link_reset = smu_link_reset_is_support(smu); + mutex_unlock(&adev->pm.mutex); + } + + return support_link_reset; +} + +int amdgpu_dpm_link_reset(struct amdgpu_device *adev) +{ + struct smu_context *smu = adev->powerplay.pp_handle; + int ret = -EOPNOTSUPP; + + if (is_support_sw_smu(adev)) { + mutex_lock(&adev->pm.mutex); + ret = smu_link_reset(smu); + mutex_unlock(&adev->pm.mutex); + } + + return ret; +} + int amdgpu_dpm_switch_power_profile(struct amdgpu_device *adev, enum PP_SMC_POWER_PROFILE type, bool en) @@ -780,6 +808,21 @@ int amdgpu_dpm_reset_sdma(struct amdgpu_device *adev, uint32_t inst_mask) return ret; } +int amdgpu_dpm_reset_vcn(struct amdgpu_device *adev, uint32_t inst_mask) +{ + struct smu_context *smu = adev->powerplay.pp_handle; + int ret; + + if (!is_support_sw_smu(adev)) + return -EOPNOTSUPP; + + mutex_lock(&adev->pm.mutex); + ret = smu_reset_vcn(smu, inst_mask); + mutex_unlock(&adev->pm.mutex); + + return ret; +} + int amdgpu_dpm_get_dpm_freq_range(struct amdgpu_device *adev, enum pp_clock_type type, uint32_t *min, diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c index 922def51685b..5537dcf23b5c 100644 --- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c @@ -2944,6 +2944,23 @@ static ssize_t amdgpu_hwmon_show_vddgfx(struct device *dev, return sysfs_emit(buf, "%d\n", vddgfx); } +static ssize_t amdgpu_hwmon_show_vddboard(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + u32 vddboard; + int r; + + /* get the voltage */ + r = amdgpu_hwmon_get_sensor_generic(adev, AMDGPU_PP_SENSOR_VDDBOARD, + (void *)&vddboard); + if (r) + return r; + + return sysfs_emit(buf, "%d\n", vddboard); +} + static ssize_t amdgpu_hwmon_show_vddgfx_label(struct device *dev, struct device_attribute *attr, char *buf) @@ -2951,6 +2968,12 @@ static ssize_t amdgpu_hwmon_show_vddgfx_label(struct device *dev, return sysfs_emit(buf, "vddgfx\n"); } +static ssize_t amdgpu_hwmon_show_vddboard_label(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + return sysfs_emit(buf, "vddboard\n"); +} static ssize_t amdgpu_hwmon_show_vddnb(struct device *dev, struct device_attribute *attr, char *buf) @@ -3294,6 +3317,8 @@ static SENSOR_DEVICE_ATTR(in0_input, S_IRUGO, amdgpu_hwmon_show_vddgfx, NULL, 0) static SENSOR_DEVICE_ATTR(in0_label, S_IRUGO, amdgpu_hwmon_show_vddgfx_label, NULL, 0); static SENSOR_DEVICE_ATTR(in1_input, S_IRUGO, amdgpu_hwmon_show_vddnb, NULL, 0); static SENSOR_DEVICE_ATTR(in1_label, S_IRUGO, amdgpu_hwmon_show_vddnb_label, NULL, 0); +static SENSOR_DEVICE_ATTR(in2_input, S_IRUGO, amdgpu_hwmon_show_vddboard, NULL, 0); +static SENSOR_DEVICE_ATTR(in2_label, S_IRUGO, amdgpu_hwmon_show_vddboard_label, NULL, 0); static SENSOR_DEVICE_ATTR(power1_average, S_IRUGO, amdgpu_hwmon_show_power_avg, NULL, 0); static SENSOR_DEVICE_ATTR(power1_input, S_IRUGO, amdgpu_hwmon_show_power_input, NULL, 0); static SENSOR_DEVICE_ATTR(power1_cap_max, S_IRUGO, amdgpu_hwmon_show_power_cap_max, NULL, 0); @@ -3341,6 +3366,8 @@ static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_in0_label.dev_attr.attr, &sensor_dev_attr_in1_input.dev_attr.attr, &sensor_dev_attr_in1_label.dev_attr.attr, + &sensor_dev_attr_in2_input.dev_attr.attr, + &sensor_dev_attr_in2_label.dev_attr.attr, &sensor_dev_attr_power1_average.dev_attr.attr, &sensor_dev_attr_power1_input.dev_attr.attr, &sensor_dev_attr_power1_cap_max.dev_attr.attr, @@ -3492,6 +3519,13 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, attr == &sensor_dev_attr_in1_label.dev_attr.attr)) return 0; + /* only few boards support vddboard */ + if ((attr == &sensor_dev_attr_in2_input.dev_attr.attr || + attr == &sensor_dev_attr_in2_label.dev_attr.attr) && + amdgpu_hwmon_get_sensor_generic(adev, AMDGPU_PP_SENSOR_VDDBOARD, + (void *)&tmp) == -EOPNOTSUPP) + return 0; + /* no mclk on APUs other than gc 9,4,3*/ if (((adev->flags & AMD_IS_APU) && (gc_ver != IP_VERSION(9, 4, 3))) && (attr == &sensor_dev_attr_freq2_input.dev_attr.attr || diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h index 4c0f7ad14816..2c3c97587dd5 100644 --- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h +++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h @@ -416,11 +416,13 @@ int amdgpu_dpm_pause_power_profile(struct amdgpu_device *adev, int amdgpu_dpm_baco_reset(struct amdgpu_device *adev); int amdgpu_dpm_mode2_reset(struct amdgpu_device *adev); +int amdgpu_dpm_link_reset(struct amdgpu_device *adev); int amdgpu_dpm_enable_gfx_features(struct amdgpu_device *adev); int amdgpu_dpm_is_baco_supported(struct amdgpu_device *adev); bool amdgpu_dpm_is_mode1_reset_supported(struct amdgpu_device *adev); +bool amdgpu_dpm_is_link_reset_supported(struct amdgpu_device *adev); int amdgpu_dpm_mode1_reset(struct amdgpu_device *adev); int amdgpu_dpm_set_mp1_state(struct amdgpu_device *adev, @@ -607,5 +609,6 @@ ssize_t amdgpu_dpm_get_pm_policy_info(struct amdgpu_device *adev, enum pp_pm_policy p_type, char *buf); int amdgpu_dpm_reset_sdma(struct amdgpu_device *adev, uint32_t inst_mask); bool amdgpu_dpm_reset_sdma_is_supported(struct amdgpu_device *adev); +int amdgpu_dpm_reset_vcn(struct amdgpu_device *adev, uint32_t inst_mask); #endif diff --git a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c index 1c25f3023e93..4c0e976004ba 100644 --- a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c +++ b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c @@ -30,16 +30,32 @@ #include "amdgpu_atombios.h" #include "amdgpu_dpm_internal.h" #include "amd_pcie.h" -#include "sid.h" +#include "atom.h" +#include "gfx_v6_0.h" #include "r600_dpm.h" +#include "sid.h" #include "si_dpm.h" -#include "atom.h" #include "../include/pptable.h" #include <linux/math64.h> #include <linux/seq_file.h> #include <linux/firmware.h> #include <legacy_dpm.h> +#include "bif/bif_3_0_d.h" +#include "bif/bif_3_0_sh_mask.h" + +#include "dce/dce_6_0_d.h" +#include "dce/dce_6_0_sh_mask.h" + +#include "gca/gfx_6_0_d.h" +#include "gca/gfx_6_0_sh_mask.h" + +#include"gmc/gmc_6_0_d.h" +#include"gmc/gmc_6_0_sh_mask.h" + +#include "smu/smu_6_0_d.h" +#include "smu/smu_6_0_sh_mask.h" + #define MC_CG_ARB_FREQ_F0 0x0a #define MC_CG_ARB_FREQ_F1 0x0b #define MC_CG_ARB_FREQ_F2 0x0c @@ -2193,7 +2209,7 @@ static u32 si_calculate_cac_wintime(struct amdgpu_device *adev) if (xclk == 0) return 0; - cac_window = RREG32(CG_CAC_CTRL) & CAC_WINDOW_MASK; + cac_window = RREG32(mmCG_CAC_CTRL) & CG_CAC_CTRL__CAC_WINDOW_MASK; cac_window_size = ((cac_window & 0xFFFF0000) >> 16) * (cac_window & 0x0000FFFF); wintime = (cac_window_size * 100) / xclk; @@ -2489,19 +2505,19 @@ static int si_populate_sq_ramping_values(struct amdgpu_device *adev, if (adev->pm.dpm.sq_ramping_threshold == 0) return -EINVAL; - if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER > (MAX_POWER_MASK >> MAX_POWER_SHIFT)) + if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER > (SQ_POWER_THROTTLE__MAX_POWER_MASK >> SQ_POWER_THROTTLE__MAX_POWER__SHIFT)) enable_sq_ramping = false; - if (SISLANDS_DPM2_SQ_RAMP_MIN_POWER > (MIN_POWER_MASK >> MIN_POWER_SHIFT)) + if (SISLANDS_DPM2_SQ_RAMP_MIN_POWER > (SQ_POWER_THROTTLE__MIN_POWER_MASK >> SQ_POWER_THROTTLE__MIN_POWER__SHIFT)) enable_sq_ramping = false; - if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA > (MAX_POWER_DELTA_MASK >> MAX_POWER_DELTA_SHIFT)) + if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA > (SQ_POWER_THROTTLE2__MAX_POWER_DELTA_MASK >> SQ_POWER_THROTTLE2__MAX_POWER_DELTA__SHIFT)) enable_sq_ramping = false; - if (SISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT)) + if (SISLANDS_DPM2_SQ_RAMP_STI_SIZE > (SQ_POWER_THROTTLE2__SHORT_TERM_INTERVAL_SIZE_MASK >> SQ_POWER_THROTTLE2__SHORT_TERM_INTERVAL_SIZE__SHIFT)) enable_sq_ramping = false; - if (SISLANDS_DPM2_SQ_RAMP_LTI_RATIO > (LTI_RATIO_MASK >> LTI_RATIO_SHIFT)) + if (SISLANDS_DPM2_SQ_RAMP_LTI_RATIO > (SQ_POWER_THROTTLE2__LONG_TERM_INTERVAL_RATIO_MASK >> SQ_POWER_THROTTLE2__LONG_TERM_INTERVAL_RATIO__SHIFT)) enable_sq_ramping = false; for (i = 0; i < state->performance_level_count; i++) { @@ -2510,14 +2526,17 @@ static int si_populate_sq_ramping_values(struct amdgpu_device *adev, if ((state->performance_levels[i].sclk >= adev->pm.dpm.sq_ramping_threshold) && enable_sq_ramping) { - sq_power_throttle |= MAX_POWER(SISLANDS_DPM2_SQ_RAMP_MAX_POWER); - sq_power_throttle |= MIN_POWER(SISLANDS_DPM2_SQ_RAMP_MIN_POWER); - sq_power_throttle2 |= MAX_POWER_DELTA(SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA); - sq_power_throttle2 |= STI_SIZE(SISLANDS_DPM2_SQ_RAMP_STI_SIZE); - sq_power_throttle2 |= LTI_RATIO(SISLANDS_DPM2_SQ_RAMP_LTI_RATIO); + sq_power_throttle |= SISLANDS_DPM2_SQ_RAMP_MAX_POWER << SQ_POWER_THROTTLE__MAX_POWER__SHIFT; + sq_power_throttle |= SISLANDS_DPM2_SQ_RAMP_MIN_POWER << SQ_POWER_THROTTLE__MIN_POWER__SHIFT; + sq_power_throttle2 |= SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA << SQ_POWER_THROTTLE2__MAX_POWER_DELTA__SHIFT; + sq_power_throttle2 |= SISLANDS_DPM2_SQ_RAMP_STI_SIZE << SQ_POWER_THROTTLE2__SHORT_TERM_INTERVAL_SIZE__SHIFT; + sq_power_throttle2 |= SISLANDS_DPM2_SQ_RAMP_LTI_RATIO << SQ_POWER_THROTTLE2__LONG_TERM_INTERVAL_RATIO__SHIFT; } else { - sq_power_throttle |= MAX_POWER_MASK | MIN_POWER_MASK; - sq_power_throttle2 |= MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK; + sq_power_throttle |= SQ_POWER_THROTTLE__MAX_POWER_MASK | + SQ_POWER_THROTTLE__MIN_POWER_MASK; + sq_power_throttle2 |= SQ_POWER_THROTTLE2__MAX_POWER_DELTA_MASK | + SQ_POWER_THROTTLE2__SHORT_TERM_INTERVAL_SIZE_MASK | + SQ_POWER_THROTTLE2__LONG_TERM_INTERVAL_RATIO_MASK; } smc_state->levels[i].SQPowerThrottle = cpu_to_be32(sq_power_throttle); @@ -2761,9 +2780,9 @@ static int si_initialize_smc_cac_tables(struct amdgpu_device *adev) if (!cac_tables) return -ENOMEM; - reg = RREG32(CG_CAC_CTRL) & ~CAC_WINDOW_MASK; - reg |= CAC_WINDOW(si_pi->powertune_data->cac_window); - WREG32(CG_CAC_CTRL, reg); + reg = RREG32(mmCG_CAC_CTRL) & ~CG_CAC_CTRL__CAC_WINDOW_MASK; + reg |= (si_pi->powertune_data->cac_window << CG_CAC_CTRL__CAC_WINDOW__SHIFT); + WREG32(mmCG_CAC_CTRL, reg); si_pi->dyn_powertune_data.cac_leakage = adev->pm.dpm.cac_leakage; si_pi->dyn_powertune_data.dc_pwr_value = @@ -2962,10 +2981,10 @@ static int si_init_smc_spll_table(struct amdgpu_device *adev) ret = si_calculate_sclk_params(adev, sclk, &sclk_params); if (ret) break; - p_div = (sclk_params.vCG_SPLL_FUNC_CNTL & SPLL_PDIV_A_MASK) >> SPLL_PDIV_A_SHIFT; - fb_div = (sclk_params.vCG_SPLL_FUNC_CNTL_3 & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT; - clk_s = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM & CLK_S_MASK) >> CLK_S_SHIFT; - clk_v = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM_2 & CLK_V_MASK) >> CLK_V_SHIFT; + p_div = (sclk_params.vCG_SPLL_FUNC_CNTL & CG_SPLL_FUNC_CNTL__SPLL_PDIV_A_MASK) >> CG_SPLL_FUNC_CNTL__SPLL_PDIV_A__SHIFT; + fb_div = (sclk_params.vCG_SPLL_FUNC_CNTL_3 & CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV_MASK) >> CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV__SHIFT; + clk_s = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM & CG_SPLL_SPREAD_SPECTRUM__CLK_S_MASK) >> CG_SPLL_SPREAD_SPECTRUM__CLK_S__SHIFT; + clk_v = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM_2 & CG_SPLL_SPREAD_SPECTRUM_2__CLK_V_MASK) >> CG_SPLL_SPREAD_SPECTRUM_2__CLK_V__SHIFT; fb_div &= ~0x00001FFF; fb_div >>= 1; @@ -3669,10 +3688,10 @@ static bool si_is_special_1gb_platform(struct amdgpu_device *adev) WREG32(MC_SEQ_IO_DEBUG_INDEX, 0xb); width = ((RREG32(MC_SEQ_IO_DEBUG_DATA) >> 1) & 1) ? 16 : 32; - tmp = RREG32(MC_ARB_RAMCFG); - row = ((tmp & NOOFROWS_MASK) >> NOOFROWS_SHIFT) + 10; - column = ((tmp & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT) + 8; - bank = ((tmp & NOOFBANK_MASK) >> NOOFBANK_SHIFT) + 2; + tmp = RREG32(mmMC_ARB_RAMCFG); + row = ((tmp & MC_ARB_RAMCFG__NOOFROWS_MASK) >> MC_ARB_RAMCFG__NOOFROWS__SHIFT) + 10; + column = ((tmp & MC_ARB_RAMCFG__NOOFCOLS_MASK) >> MC_ARB_RAMCFG__NOOFCOLS__SHIFT) + 8; + bank = ((tmp & MC_ARB_RAMCFG__NOOFBANK_MASK) >> MC_ARB_RAMCFG__NOOFBANK__SHIFT) + 2; density = (1 << (row + column - 20 + bank)) * width; @@ -3756,11 +3775,11 @@ static void si_set_dpm_event_sources(struct amdgpu_device *adev, u32 sources) } if (want_thermal_protection) { - WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK); + WREG32_P(mmCG_THERMAL_CTRL, dpm_event_src << CG_THERMAL_CTRL__DPM_EVENT_SRC__SHIFT, ~CG_THERMAL_CTRL__DPM_EVENT_SRC_MASK); if (pi->thermal_protection) - WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS); + WREG32_P(mmGENERAL_PWRMGT, 0, ~GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK); } else { - WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS); + WREG32_P(mmGENERAL_PWRMGT, GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK, ~GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK); } } @@ -3785,20 +3804,20 @@ static void si_enable_auto_throttle_source(struct amdgpu_device *adev, static void si_start_dpm(struct amdgpu_device *adev) { - WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN); + WREG32_P(mmGENERAL_PWRMGT, GENERAL_PWRMGT__GLOBAL_PWRMGT_EN_MASK, ~GENERAL_PWRMGT__GLOBAL_PWRMGT_EN_MASK); } static void si_stop_dpm(struct amdgpu_device *adev) { - WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN); + WREG32_P(mmGENERAL_PWRMGT, 0, ~GENERAL_PWRMGT__GLOBAL_PWRMGT_EN_MASK); } static void si_enable_sclk_control(struct amdgpu_device *adev, bool enable) { if (enable) - WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF); + WREG32_P(mmSCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF_MASK); else - WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF); + WREG32_P(mmSCLK_PWRMGT_CNTL, SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF_MASK, ~SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF_MASK); } @@ -3838,7 +3857,7 @@ static int si_notify_hw_of_powersource(struct amdgpu_device *adev, bool ac_power static PPSMC_Result si_send_msg_to_smc_with_parameter(struct amdgpu_device *adev, PPSMC_Msg msg, u32 parameter) { - WREG32(SMC_SCRATCH0, parameter); + WREG32(mmSMC_SCRATCH0, parameter); return amdgpu_si_send_msg_to_smc(adev, msg); } @@ -4023,12 +4042,12 @@ static void si_read_clock_registers(struct amdgpu_device *adev) { struct si_power_info *si_pi = si_get_pi(adev); - si_pi->clock_registers.cg_spll_func_cntl = RREG32(CG_SPLL_FUNC_CNTL); - si_pi->clock_registers.cg_spll_func_cntl_2 = RREG32(CG_SPLL_FUNC_CNTL_2); - si_pi->clock_registers.cg_spll_func_cntl_3 = RREG32(CG_SPLL_FUNC_CNTL_3); - si_pi->clock_registers.cg_spll_func_cntl_4 = RREG32(CG_SPLL_FUNC_CNTL_4); - si_pi->clock_registers.cg_spll_spread_spectrum = RREG32(CG_SPLL_SPREAD_SPECTRUM); - si_pi->clock_registers.cg_spll_spread_spectrum_2 = RREG32(CG_SPLL_SPREAD_SPECTRUM_2); + si_pi->clock_registers.cg_spll_func_cntl = RREG32(mmCG_SPLL_FUNC_CNTL); + si_pi->clock_registers.cg_spll_func_cntl_2 = RREG32(mmCG_SPLL_FUNC_CNTL_2); + si_pi->clock_registers.cg_spll_func_cntl_3 = RREG32(mmCG_SPLL_FUNC_CNTL_3); + si_pi->clock_registers.cg_spll_func_cntl_4 = RREG32(mmCG_SPLL_FUNC_CNTL_4); + si_pi->clock_registers.cg_spll_spread_spectrum = RREG32(mmCG_SPLL_SPREAD_SPECTRUM); + si_pi->clock_registers.cg_spll_spread_spectrum_2 = RREG32(mmCG_SPLL_SPREAD_SPECTRUM_2); si_pi->clock_registers.dll_cntl = RREG32(DLL_CNTL); si_pi->clock_registers.mclk_pwrmgt_cntl = RREG32(MCLK_PWRMGT_CNTL); si_pi->clock_registers.mpll_ad_func_cntl = RREG32(MPLL_AD_FUNC_CNTL); @@ -4044,14 +4063,14 @@ static void si_enable_thermal_protection(struct amdgpu_device *adev, bool enable) { if (enable) - WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS); + WREG32_P(mmGENERAL_PWRMGT, 0, ~GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK); else - WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS); + WREG32_P(mmGENERAL_PWRMGT, GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK, ~GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK); } static void si_enable_acpi_power_management(struct amdgpu_device *adev) { - WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN); + WREG32_P(mmGENERAL_PWRMGT, GENERAL_PWRMGT__STATIC_PM_EN_MASK, ~GENERAL_PWRMGT__STATIC_PM_EN_MASK); } #if 0 @@ -4132,9 +4151,9 @@ static void si_program_ds_registers(struct amdgpu_device *adev) tmp = 0x1; if (eg_pi->sclk_deep_sleep) { - WREG32_P(MISC_CLK_CNTL, DEEP_SLEEP_CLK_SEL(tmp), ~DEEP_SLEEP_CLK_SEL_MASK); - WREG32_P(CG_SPLL_AUTOSCALE_CNTL, AUTOSCALE_ON_SS_CLEAR, - ~AUTOSCALE_ON_SS_CLEAR); + WREG32_P(mmMISC_CLK_CNTL, (tmp << MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL__SHIFT), ~MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL_MASK); + WREG32_P(mmCG_SPLL_AUTOSCALE_CNTL, CG_SPLL_AUTOSCALE_CNTL__AUTOSCALE_ON_SS_CLEAR_MASK, + ~CG_SPLL_AUTOSCALE_CNTL__AUTOSCALE_ON_SS_CLEAR_MASK); } } @@ -4143,18 +4162,18 @@ static void si_program_display_gap(struct amdgpu_device *adev) u32 tmp, pipe; int i; - tmp = RREG32(CG_DISPLAY_GAP_CNTL) & ~(DISP1_GAP_MASK | DISP2_GAP_MASK); + tmp = RREG32(mmCG_DISPLAY_GAP_CNTL) & ~(CG_DISPLAY_GAP_CNTL__DISP1_GAP_MASK | CG_DISPLAY_GAP_CNTL__DISP2_GAP_MASK); if (adev->pm.dpm.new_active_crtc_count > 0) - tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM); + tmp |= R600_PM_DISPLAY_GAP_VBLANK_OR_WM << CG_DISPLAY_GAP_CNTL__DISP1_GAP__SHIFT; else - tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE); + tmp |= R600_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP1_GAP__SHIFT; if (adev->pm.dpm.new_active_crtc_count > 1) - tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM); + tmp |= R600_PM_DISPLAY_GAP_VBLANK_OR_WM << CG_DISPLAY_GAP_CNTL__DISP2_GAP__SHIFT; else - tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE); + tmp |= R600_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP2_GAP__SHIFT; - WREG32(CG_DISPLAY_GAP_CNTL, tmp); + WREG32(mmCG_DISPLAY_GAP_CNTL, tmp); tmp = RREG32(DCCG_DISP_SLOW_SELECT_REG); pipe = (tmp & DCCG_DISP1_SLOW_SELECT_MASK) >> DCCG_DISP1_SLOW_SELECT_SHIFT; @@ -4189,10 +4208,10 @@ static void si_enable_spread_spectrum(struct amdgpu_device *adev, bool enable) if (enable) { if (pi->sclk_ss) - WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN); + WREG32_P(mmGENERAL_PWRMGT, GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN_MASK, ~GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN_MASK); } else { - WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN); - WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN); + WREG32_P(mmCG_SPLL_SPREAD_SPECTRUM, 0, ~CG_SPLL_SPREAD_SPECTRUM__SSEN_MASK); + WREG32_P(mmGENERAL_PWRMGT, 0, ~GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN_MASK); } } @@ -4214,15 +4233,15 @@ static void si_setup_bsp(struct amdgpu_device *adev) &pi->pbsu); - pi->dsp = BSP(pi->bsp) | BSU(pi->bsu); - pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu); + pi->dsp = (pi->bsp << CG_BSP__BSP__SHIFT) | (pi->bsu << CG_BSP__BSU__SHIFT); + pi->psp = (pi->pbsp << CG_BSP__BSP__SHIFT) | (pi->pbsu << CG_BSP__BSU__SHIFT); - WREG32(CG_BSP, pi->dsp); + WREG32(mmCG_BSP, pi->dsp); } static void si_program_git(struct amdgpu_device *adev) { - WREG32_P(CG_GIT, CG_GICST(R600_GICST_DFLT), ~CG_GICST_MASK); + WREG32_P(mmCG_GIT, R600_GICST_DFLT << CG_GIT__CG_GICST__SHIFT, ~CG_GIT__CG_GICST_MASK); } static void si_program_tp(struct amdgpu_device *adev) @@ -4231,54 +4250,54 @@ static void si_program_tp(struct amdgpu_device *adev) enum r600_td td = R600_TD_DFLT; for (i = 0; i < R600_PM_NUMBER_OF_TC; i++) - WREG32(CG_FFCT_0 + i, (UTC_0(r600_utc[i]) | DTC_0(r600_dtc[i]))); + WREG32(mmCG_FFCT_0 + i, (r600_utc[i] << CG_FFCT_0__UTC_0__SHIFT | r600_dtc[i] << CG_FFCT_0__DTC_0__SHIFT)); if (td == R600_TD_AUTO) - WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL); + WREG32_P(mmSCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_CNTL__FIR_FORCE_TREND_SEL_MASK); else - WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL); + WREG32_P(mmSCLK_PWRMGT_CNTL, SCLK_PWRMGT_CNTL__FIR_FORCE_TREND_SEL_MASK, ~SCLK_PWRMGT_CNTL__FIR_FORCE_TREND_SEL_MASK); if (td == R600_TD_UP) - WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE); + WREG32_P(mmSCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_CNTL__FIR_TREND_MODE_MASK); if (td == R600_TD_DOWN) - WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE); + WREG32_P(mmSCLK_PWRMGT_CNTL, SCLK_PWRMGT_CNTL__FIR_TREND_MODE_MASK, ~SCLK_PWRMGT_CNTL__FIR_TREND_MODE_MASK); } static void si_program_tpp(struct amdgpu_device *adev) { - WREG32(CG_TPC, R600_TPC_DFLT); + WREG32(mmCG_TPC, R600_TPC_DFLT); } static void si_program_sstp(struct amdgpu_device *adev) { - WREG32(CG_SSP, (SSTU(R600_SSTU_DFLT) | SST(R600_SST_DFLT))); + WREG32(mmCG_SSP, (R600_SSTU_DFLT << CG_SSP__SSTU__SHIFT| R600_SST_DFLT << CG_SSP__SST__SHIFT)); } static void si_enable_display_gap(struct amdgpu_device *adev) { - u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL); + u32 tmp = RREG32(mmCG_DISPLAY_GAP_CNTL); - tmp &= ~(DISP1_GAP_MASK | DISP2_GAP_MASK); - tmp |= (DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE) | - DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE)); + tmp &= ~(CG_DISPLAY_GAP_CNTL__DISP1_GAP_MASK | CG_DISPLAY_GAP_CNTL__DISP2_GAP_MASK); + tmp |= (R600_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP1_GAP__SHIFT | + R600_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP2_GAP__SHIFT); - tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK); - tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK) | - DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE)); - WREG32(CG_DISPLAY_GAP_CNTL, tmp); + tmp &= ~(CG_DISPLAY_GAP_CNTL__DISP1_GAP_MCHG_MASK | CG_DISPLAY_GAP_CNTL__DISP2_GAP_MCHG_MASK); + tmp |= (R600_PM_DISPLAY_GAP_VBLANK << CG_DISPLAY_GAP_CNTL__DISP1_GAP_MCHG__SHIFT | + R600_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP2_GAP_MCHG__SHIFT); + WREG32(mmCG_DISPLAY_GAP_CNTL, tmp); } static void si_program_vc(struct amdgpu_device *adev) { struct rv7xx_power_info *pi = rv770_get_pi(adev); - WREG32(CG_FTV, pi->vrc); + WREG32(mmCG_FTV, pi->vrc); } static void si_clear_vc(struct amdgpu_device *adev) { - WREG32(CG_FTV, 0); + WREG32(mmCG_FTV, 0); } static u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock) @@ -4735,7 +4754,7 @@ static u32 si_calculate_memory_refresh_rate(struct amdgpu_device *adev, u32 dram_rows; u32 dram_refresh_rate; u32 mc_arb_rfsh_rate; - u32 tmp = (RREG32(MC_ARB_RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT; + u32 tmp = (RREG32(mmMC_ARB_RAMCFG) & MC_ARB_RAMCFG__NOOFROWS_MASK) >> MC_ARB_RAMCFG__NOOFROWS__SHIFT; if (tmp >= 4) dram_rows = 16384; @@ -4909,7 +4928,7 @@ static int si_populate_smc_initial_state(struct amdgpu_device *adev, si_populate_initial_mvdd_value(adev, &table->initialState.level.mvdd); - reg = CG_R(0xffff) | CG_L(0); + reg = 0xffff << CG_AT__CG_R__SHIFT | 0 << CG_AT__CG_L__SHIFT; table->initialState.level.aT = cpu_to_be32(reg); table->initialState.level.bSP = cpu_to_be32(pi->dsp); table->initialState.level.gen2PCIE = (u8)si_pi->boot_pcie_gen; @@ -4935,10 +4954,13 @@ static int si_populate_smc_initial_state(struct amdgpu_device *adev, table->initialState.level.dpm2.BelowSafeInc = 0; table->initialState.level.dpm2.PwrEfficiencyRatio = 0; - reg = MIN_POWER_MASK | MAX_POWER_MASK; + reg = SQ_POWER_THROTTLE__MIN_POWER_MASK | + SQ_POWER_THROTTLE__MAX_POWER_MASK; table->initialState.level.SQPowerThrottle = cpu_to_be32(reg); - reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK; + reg = SQ_POWER_THROTTLE2__MAX_POWER_DELTA_MASK | + SQ_POWER_THROTTLE2__SHORT_TERM_INTERVAL_SIZE_MASK | + SQ_POWER_THROTTLE2__LONG_TERM_INTERVAL_RATIO_MASK; table->initialState.level.SQPowerThrottle_2 = cpu_to_be32(reg); return 0; @@ -5057,8 +5079,8 @@ static int si_populate_smc_acpi_state(struct amdgpu_device *adev, dll_cntl &= ~(MRDCK0_BYPASS | MRDCK1_BYPASS); - spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK; - spll_func_cntl_2 |= SCLK_MUX_SEL(4); + spll_func_cntl_2 &= ~CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL_MASK; + spll_func_cntl_2 |= 4 << CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL__SHIFT; table->ACPIState.level.mclk.vDLL_CNTL = cpu_to_be32(dll_cntl); @@ -5102,10 +5124,10 @@ static int si_populate_smc_acpi_state(struct amdgpu_device *adev, table->ACPIState.level.dpm2.BelowSafeInc = 0; table->ACPIState.level.dpm2.PwrEfficiencyRatio = 0; - reg = MIN_POWER_MASK | MAX_POWER_MASK; + reg = SQ_POWER_THROTTLE__MIN_POWER_MASK | SQ_POWER_THROTTLE__MAX_POWER_MASK; table->ACPIState.level.SQPowerThrottle = cpu_to_be32(reg); - reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK; + reg = SQ_POWER_THROTTLE2__MAX_POWER_DELTA_MASK | SQ_POWER_THROTTLE2__SHORT_TERM_INTERVAL_SIZE_MASK | SQ_POWER_THROTTLE2__LONG_TERM_INTERVAL_RATIO_MASK; table->ACPIState.level.SQPowerThrottle_2 = cpu_to_be32(reg); return 0; @@ -5250,8 +5272,8 @@ static int si_init_smc_table(struct amdgpu_device *adev) if (ret) return ret; - WREG32(CG_ULV_CONTROL, ulv->cg_ulv_control); - WREG32(CG_ULV_PARAMETER, ulv->cg_ulv_parameter); + WREG32(mmCG_ULV_CONTROL, ulv->cg_ulv_control); + WREG32(mmCG_ULV_PARAMETER, ulv->cg_ulv_parameter); lane_width = amdgpu_get_pcie_lanes(adev); si_write_smc_soft_register(adev, SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width, lane_width); @@ -5294,16 +5316,16 @@ static int si_calculate_sclk_params(struct amdgpu_device *adev, do_div(tmp, reference_clock); fbdiv = (u32) tmp; - spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK); - spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div); - spll_func_cntl |= SPLL_PDIV_A(dividers.post_div); + spll_func_cntl &= ~(CG_SPLL_FUNC_CNTL__SPLL_PDIV_A_MASK | CG_SPLL_FUNC_CNTL__SPLL_REF_DIV_MASK); + spll_func_cntl |= dividers.ref_div << CG_SPLL_FUNC_CNTL__SPLL_REF_DIV__SHIFT; + spll_func_cntl |= dividers.post_div << CG_SPLL_FUNC_CNTL__SPLL_PDIV_A__SHIFT; - spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK; - spll_func_cntl_2 |= SCLK_MUX_SEL(2); + spll_func_cntl_2 &= ~CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL_MASK; + spll_func_cntl_2 |= 2 << CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL__SHIFT; - spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK; - spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv); - spll_func_cntl_3 |= SPLL_DITHEN; + spll_func_cntl_3 &= ~CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV_MASK; + spll_func_cntl_3 |= fbdiv << CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV__SHIFT; + spll_func_cntl_3 |= CG_SPLL_FUNC_CNTL_3__SPLL_DITHEN_MASK; if (pi->sclk_ss) { struct amdgpu_atom_ss ss; @@ -5314,12 +5336,12 @@ static int si_calculate_sclk_params(struct amdgpu_device *adev, u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate); u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000); - cg_spll_spread_spectrum &= ~CLK_S_MASK; - cg_spll_spread_spectrum |= CLK_S(clk_s); - cg_spll_spread_spectrum |= SSEN; + cg_spll_spread_spectrum &= ~CG_SPLL_SPREAD_SPECTRUM__CLK_S_MASK; + cg_spll_spread_spectrum |= clk_s << CG_SPLL_SPREAD_SPECTRUM__CLK_S__SHIFT; + cg_spll_spread_spectrum |= CG_SPLL_SPREAD_SPECTRUM__SSEN_MASK; - cg_spll_spread_spectrum_2 &= ~CLK_V_MASK; - cg_spll_spread_spectrum_2 |= CLK_V(clk_v); + cg_spll_spread_spectrum_2 &= ~CG_SPLL_SPREAD_SPECTRUM_2__CLK_V_MASK; + cg_spll_spread_spectrum_2 |= clk_v << CG_SPLL_SPREAD_SPECTRUM_2__CLK_V__SHIFT; } } @@ -5485,7 +5507,7 @@ static int si_convert_power_level_to_smc(struct amdgpu_device *adev, if (pi->mclk_stutter_mode_threshold && (pl->mclk <= pi->mclk_stutter_mode_threshold) && !eg_pi->uvd_enabled && - (RREG32(DPG_PIPE_STUTTER_CONTROL) & STUTTER_ENABLE) && + (RREG32(mmDPG_PIPE_STUTTER_CONTROL) & DPG_PIPE_STUTTER_CONTROL__STUTTER_ENABLE_MASK) && (adev->pm.dpm.new_active_crtc_count <= 2)) { level->mcFlags |= SISLANDS_SMC_MC_STUTTER_EN; } @@ -5579,7 +5601,7 @@ static int si_populate_smc_t(struct amdgpu_device *adev, return -EINVAL; if (state->performance_level_count < 2) { - a_t = CG_R(0xffff) | CG_L(0); + a_t = 0xffff << CG_AT__CG_R__SHIFT | 0 << CG_AT__CG_L__SHIFT; smc_state->levels[0].aT = cpu_to_be32(a_t); return 0; } @@ -5600,13 +5622,13 @@ static int si_populate_smc_t(struct amdgpu_device *adev, t_l = (i + 1) * 1000 + 50 * R600_AH_DFLT; } - a_t = be32_to_cpu(smc_state->levels[i].aT) & ~CG_R_MASK; - a_t |= CG_R(t_l * pi->bsp / 20000); + a_t = be32_to_cpu(smc_state->levels[i].aT) & ~CG_AT__CG_R_MASK; + a_t |= (t_l * pi->bsp / 20000) << CG_AT__CG_R__SHIFT; smc_state->levels[i].aT = cpu_to_be32(a_t); high_bsp = (i == state->performance_level_count - 2) ? pi->pbsp : pi->bsp; - a_t = CG_R(0xffff) | CG_L(t_h * high_bsp / 20000); + a_t = (0xffff) << CG_AT__CG_R__SHIFT | (t_h * high_bsp / 20000) << CG_AT__CG_L__SHIFT; smc_state->levels[i + 1].aT = cpu_to_be32(a_t); } @@ -6180,9 +6202,9 @@ static int si_upload_mc_reg_table(struct amdgpu_device *adev, static void si_enable_voltage_control(struct amdgpu_device *adev, bool enable) { if (enable) - WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN); + WREG32_P(mmGENERAL_PWRMGT, GENERAL_PWRMGT__VOLT_PWRMGT_EN_MASK, ~GENERAL_PWRMGT__VOLT_PWRMGT_EN_MASK); else - WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN); + WREG32_P(mmGENERAL_PWRMGT, 0, ~GENERAL_PWRMGT__VOLT_PWRMGT_EN_MASK); } static enum si_pcie_gen si_get_maximum_link_speed(struct amdgpu_device *adev, @@ -6204,8 +6226,8 @@ static u16 si_get_current_pcie_speed(struct amdgpu_device *adev) { u32 speed_cntl; - speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & LC_CURRENT_DATA_RATE_MASK; - speed_cntl >>= LC_CURRENT_DATA_RATE_SHIFT; + speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL) & PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK; + speed_cntl >>= PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE__SHIFT; return (u16)speed_cntl; } @@ -6412,21 +6434,21 @@ static void si_dpm_setup_asic(struct amdgpu_device *adev) static int si_thermal_enable_alert(struct amdgpu_device *adev, bool enable) { - u32 thermal_int = RREG32(CG_THERMAL_INT); + u32 thermal_int = RREG32(mmCG_THERMAL_INT); if (enable) { PPSMC_Result result; - thermal_int &= ~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW); - WREG32(CG_THERMAL_INT, thermal_int); + thermal_int &= ~(CG_THERMAL_INT__THERM_INT_MASK_HIGH_MASK | CG_THERMAL_INT__THERM_INT_MASK_LOW_MASK); + WREG32(mmCG_THERMAL_INT, thermal_int); result = amdgpu_si_send_msg_to_smc(adev, PPSMC_MSG_EnableThermalInterrupt); if (result != PPSMC_Result_OK) { DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); return -EINVAL; } } else { - thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW; - WREG32(CG_THERMAL_INT, thermal_int); + thermal_int |= CG_THERMAL_INT__THERM_INT_MASK_HIGH_MASK | CG_THERMAL_INT__THERM_INT_MASK_LOW_MASK; + WREG32(mmCG_THERMAL_INT, thermal_int); } return 0; @@ -6447,9 +6469,9 @@ static int si_thermal_set_temperature_range(struct amdgpu_device *adev, return -EINVAL; } - WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK); - WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK); - WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK); + WREG32_P(mmCG_THERMAL_INT, (high_temp / 1000) << CG_THERMAL_INT__DIG_THERM_INTH__SHIFT, ~CG_THERMAL_INT__DIG_THERM_INTH_MASK); + WREG32_P(mmCG_THERMAL_INT, (low_temp / 1000) << CG_THERMAL_INT__DIG_THERM_INTL__SHIFT, ~CG_THERMAL_INT__DIG_THERM_INTL_MASK); + WREG32_P(mmCG_THERMAL_CTRL, (high_temp / 1000) << CG_THERMAL_CTRL__DIG_THERM_DPM__SHIFT, ~CG_THERMAL_CTRL__DIG_THERM_DPM_MASK); adev->pm.dpm.thermal.min_temp = low_temp; adev->pm.dpm.thermal.max_temp = high_temp; @@ -6463,20 +6485,20 @@ static void si_fan_ctrl_set_static_mode(struct amdgpu_device *adev, u32 mode) u32 tmp; if (si_pi->fan_ctrl_is_in_default_mode) { - tmp = (RREG32(CG_FDO_CTRL2) & FDO_PWM_MODE_MASK) >> FDO_PWM_MODE_SHIFT; + tmp = (RREG32(mmCG_FDO_CTRL2) & CG_FDO_CTRL2__FDO_PWM_MODE_MASK) >> CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT; si_pi->fan_ctrl_default_mode = tmp; - tmp = (RREG32(CG_FDO_CTRL2) & TMIN_MASK) >> TMIN_SHIFT; + tmp = (RREG32(mmCG_FDO_CTRL2) & CG_FDO_CTRL2__TMIN_MASK) >> CG_FDO_CTRL2__TMIN__SHIFT; si_pi->t_min = tmp; si_pi->fan_ctrl_is_in_default_mode = false; } - tmp = RREG32(CG_FDO_CTRL2) & ~TMIN_MASK; - tmp |= TMIN(0); - WREG32(CG_FDO_CTRL2, tmp); + tmp = RREG32(mmCG_FDO_CTRL2) & ~CG_FDO_CTRL2__TMIN_MASK; + tmp |= 0 << CG_FDO_CTRL2__TMIN__SHIFT; + WREG32(mmCG_FDO_CTRL2, tmp); - tmp = RREG32(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK; - tmp |= FDO_PWM_MODE(mode); - WREG32(CG_FDO_CTRL2, tmp); + tmp = RREG32(mmCG_FDO_CTRL2) & ~CG_FDO_CTRL2__FDO_PWM_MODE_MASK; + tmp |= mode << CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT; + WREG32(mmCG_FDO_CTRL2, tmp); } static int si_thermal_setup_fan_table(struct amdgpu_device *adev) @@ -6495,7 +6517,7 @@ static int si_thermal_setup_fan_table(struct amdgpu_device *adev) return 0; } - duty100 = (RREG32(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT; + duty100 = (RREG32(mmCG_FDO_CTRL1) & CG_FDO_CTRL1__FMAX_DUTY100_MASK) >> CG_FDO_CTRL1__FMAX_DUTY100__SHIFT; if (duty100 == 0) { adev->pm.dpm.fan.ucode_fan_control = false; @@ -6531,7 +6553,7 @@ static int si_thermal_setup_fan_table(struct amdgpu_device *adev) reference_clock) / 1600); fan_table.fdo_max = cpu_to_be16((u16)duty100); - tmp = (RREG32(CG_MULT_THERMAL_CTRL) & TEMP_SEL_MASK) >> TEMP_SEL_SHIFT; + tmp = (RREG32(mmCG_MULT_THERMAL_CTRL) & CG_MULT_THERMAL_CTRL__TEMP_SEL_MASK) >> CG_MULT_THERMAL_CTRL__TEMP_SEL__SHIFT; fan_table.temp_src = (uint8_t)tmp; ret = amdgpu_si_copy_bytes_to_smc(adev, @@ -6590,8 +6612,8 @@ static int si_dpm_get_fan_speed_pwm(void *handle, if (adev->pm.no_fan) return -ENOENT; - duty100 = (RREG32(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT; - duty = (RREG32(CG_THERMAL_STATUS) & FDO_PWM_DUTY_MASK) >> FDO_PWM_DUTY_SHIFT; + duty100 = (RREG32(mmCG_FDO_CTRL1) & CG_FDO_CTRL1__FMAX_DUTY100_MASK) >> CG_FDO_CTRL1__FMAX_DUTY100__SHIFT; + duty = (RREG32(mmCG_THERMAL_STATUS) & CG_THERMAL_STATUS__FDO_PWM_DUTY_MASK) >> CG_THERMAL_STATUS__FDO_PWM_DUTY__SHIFT; if (duty100 == 0) return -EINVAL; @@ -6621,7 +6643,7 @@ static int si_dpm_set_fan_speed_pwm(void *handle, if (speed > 255) return -EINVAL; - duty100 = (RREG32(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT; + duty100 = (RREG32(mmCG_FDO_CTRL1) & CG_FDO_CTRL1__FMAX_DUTY100_MASK) >> CG_FDO_CTRL1__FMAX_DUTY100__SHIFT; if (duty100 == 0) return -EINVAL; @@ -6630,9 +6652,9 @@ static int si_dpm_set_fan_speed_pwm(void *handle, do_div(tmp64, 255); duty = (u32)tmp64; - tmp = RREG32(CG_FDO_CTRL0) & ~FDO_STATIC_DUTY_MASK; - tmp |= FDO_STATIC_DUTY(duty); - WREG32(CG_FDO_CTRL0, tmp); + tmp = RREG32(mmCG_FDO_CTRL0) & ~CG_FDO_CTRL0__FDO_STATIC_DUTY_MASK; + tmp |= duty << CG_FDO_CTRL0__FDO_STATIC_DUTY__SHIFT; + WREG32(mmCG_FDO_CTRL0, tmp); return 0; } @@ -6672,8 +6694,8 @@ static int si_dpm_get_fan_control_mode(void *handle, u32 *fan_mode) if (si_pi->fan_is_controlled_by_smc) return 0; - tmp = RREG32(CG_FDO_CTRL2) & FDO_PWM_MODE_MASK; - *fan_mode = (tmp >> FDO_PWM_MODE_SHIFT); + tmp = RREG32(mmCG_FDO_CTRL2) & CG_FDO_CTRL2__FDO_PWM_MODE_MASK; + *fan_mode = (tmp >> CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT); return 0; } @@ -6691,7 +6713,7 @@ static int si_fan_ctrl_get_fan_speed_rpm(struct amdgpu_device *adev, if (adev->pm.fan_pulses_per_revolution == 0) return -ENOENT; - tach_period = (RREG32(CG_TACH_STATUS) & TACH_PERIOD_MASK) >> TACH_PERIOD_SHIFT; + tach_period = (RREG32(mmCG_TACH_STATUS) & CG_TACH_STATUS__TACH_PERIOD_MASK) >> CG_TACH_STATUS__TACH_PERIOD__SHIFT; if (tach_period == 0) return -ENOENT; @@ -6720,9 +6742,9 @@ static int si_fan_ctrl_set_fan_speed_rpm(struct amdgpu_device *adev, si_fan_ctrl_stop_smc_fan_control(adev); tach_period = 60 * xclk * 10000 / (8 * speed); - tmp = RREG32(CG_TACH_CTRL) & ~TARGET_PERIOD_MASK; - tmp |= TARGET_PERIOD(tach_period); - WREG32(CG_TACH_CTRL, tmp); + tmp = RREG32(mmCG_TACH_CTRL) & ~CG_TACH_CTRL__TARGET_PERIOD_MASK; + tmp |= tach_period << CG_TACH_CTRL__TARGET_PERIOD__SHIFT; + WREG32(mmCG_TACH_CTRL, tmp); si_fan_ctrl_set_static_mode(adev, FDO_PWM_MODE_STATIC_RPM); @@ -6736,13 +6758,13 @@ static void si_fan_ctrl_set_default_mode(struct amdgpu_device *adev) u32 tmp; if (!si_pi->fan_ctrl_is_in_default_mode) { - tmp = RREG32(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK; - tmp |= FDO_PWM_MODE(si_pi->fan_ctrl_default_mode); - WREG32(CG_FDO_CTRL2, tmp); + tmp = RREG32(mmCG_FDO_CTRL2) & ~CG_FDO_CTRL2__FDO_PWM_MODE_MASK; + tmp |= si_pi->fan_ctrl_default_mode << CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT; + WREG32(mmCG_FDO_CTRL2, tmp); - tmp = RREG32(CG_FDO_CTRL2) & ~TMIN_MASK; - tmp |= TMIN(si_pi->t_min); - WREG32(CG_FDO_CTRL2, tmp); + tmp = RREG32(mmCG_FDO_CTRL2) & ~CG_FDO_CTRL2__TMIN_MASK; + tmp |= si_pi->t_min << CG_FDO_CTRL2__TMIN__SHIFT; + WREG32(mmCG_FDO_CTRL2, tmp); si_pi->fan_ctrl_is_in_default_mode = true; } } @@ -6760,14 +6782,14 @@ static void si_thermal_initialize(struct amdgpu_device *adev) u32 tmp; if (adev->pm.fan_pulses_per_revolution) { - tmp = RREG32(CG_TACH_CTRL) & ~EDGE_PER_REV_MASK; - tmp |= EDGE_PER_REV(adev->pm.fan_pulses_per_revolution -1); - WREG32(CG_TACH_CTRL, tmp); + tmp = RREG32(mmCG_TACH_CTRL) & ~CG_TACH_CTRL__EDGE_PER_REV_MASK; + tmp |= (adev->pm.fan_pulses_per_revolution -1) << CG_TACH_CTRL__EDGE_PER_REV__SHIFT; + WREG32(mmCG_TACH_CTRL, tmp); } - tmp = RREG32(CG_FDO_CTRL2) & ~TACH_PWM_RESP_RATE_MASK; - tmp |= TACH_PWM_RESP_RATE(0x28); - WREG32(CG_FDO_CTRL2, tmp); + tmp = RREG32(mmCG_FDO_CTRL2) & ~CG_FDO_CTRL2__TACH_PWM_RESP_RATE_MASK; + tmp |= 0x28 << CG_FDO_CTRL2__TACH_PWM_RESP_RATE__SHIFT; + WREG32(mmCG_FDO_CTRL2, tmp); } static int si_thermal_start_thermal_controller(struct amdgpu_device *adev) @@ -7530,8 +7552,8 @@ static void si_dpm_debugfs_print_current_performance_level(void *handle, struct si_ps *ps = si_get_ps(rps); struct rv7xx_pl *pl; u32 current_index = - (RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_INDEX_MASK) >> - CURRENT_STATE_INDEX_SHIFT; + (RREG32(mmTARGET_AND_CURRENT_PROFILE_INDEX) & TARGET_AND_CURRENT_PROFILE_INDEX__CURRENT_STATE_INDEX_MASK) >> + TARGET_AND_CURRENT_PROFILE_INDEX__CURRENT_STATE_INDEX__SHIFT; if (current_index >= ps->performance_level_count) { seq_printf(m, "invalid dpm profile %d\n", current_index); @@ -7554,14 +7576,14 @@ static int si_dpm_set_interrupt_state(struct amdgpu_device *adev, case AMDGPU_THERMAL_IRQ_LOW_TO_HIGH: switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - cg_thermal_int = RREG32_SMC(CG_THERMAL_INT); - cg_thermal_int |= THERM_INT_MASK_HIGH; - WREG32_SMC(CG_THERMAL_INT, cg_thermal_int); + cg_thermal_int = RREG32_SMC(mmCG_THERMAL_INT); + cg_thermal_int |= CG_THERMAL_INT__THERM_INT_MASK_HIGH_MASK; + WREG32_SMC(mmCG_THERMAL_INT, cg_thermal_int); break; case AMDGPU_IRQ_STATE_ENABLE: - cg_thermal_int = RREG32_SMC(CG_THERMAL_INT); - cg_thermal_int &= ~THERM_INT_MASK_HIGH; - WREG32_SMC(CG_THERMAL_INT, cg_thermal_int); + cg_thermal_int = RREG32_SMC(mmCG_THERMAL_INT); + cg_thermal_int &= ~CG_THERMAL_INT__THERM_INT_MASK_HIGH_MASK; + WREG32_SMC(mmCG_THERMAL_INT, cg_thermal_int); break; default: break; @@ -7571,14 +7593,14 @@ static int si_dpm_set_interrupt_state(struct amdgpu_device *adev, case AMDGPU_THERMAL_IRQ_HIGH_TO_LOW: switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - cg_thermal_int = RREG32_SMC(CG_THERMAL_INT); - cg_thermal_int |= THERM_INT_MASK_LOW; - WREG32_SMC(CG_THERMAL_INT, cg_thermal_int); + cg_thermal_int = RREG32_SMC(mmCG_THERMAL_INT); + cg_thermal_int |= CG_THERMAL_INT__THERM_INT_MASK_LOW_MASK; + WREG32_SMC(mmCG_THERMAL_INT, cg_thermal_int); break; case AMDGPU_IRQ_STATE_ENABLE: - cg_thermal_int = RREG32_SMC(CG_THERMAL_INT); - cg_thermal_int &= ~THERM_INT_MASK_LOW; - WREG32_SMC(CG_THERMAL_INT, cg_thermal_int); + cg_thermal_int = RREG32_SMC(mmCG_THERMAL_INT); + cg_thermal_int &= ~CG_THERMAL_INT__THERM_INT_MASK_LOW_MASK; + WREG32_SMC(mmCG_THERMAL_INT, cg_thermal_int); break; default: break; @@ -7883,8 +7905,8 @@ static int si_dpm_get_temp(void *handle) int actual_temp = 0; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - temp = (RREG32(CG_MULT_THERMAL_STATUS) & CTF_TEMP_MASK) >> - CTF_TEMP_SHIFT; + temp = (RREG32(mmCG_MULT_THERMAL_STATUS) & CG_MULT_THERMAL_STATUS__CTF_TEMP_MASK) >> + CG_MULT_THERMAL_STATUS__CTF_TEMP__SHIFT; if (temp & 0x200) actual_temp = 255; @@ -8014,8 +8036,8 @@ static int si_dpm_read_sensor(void *handle, int idx, struct si_ps *ps = si_get_ps(rps); uint32_t sclk, mclk; u32 pl_index = - (RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_INDEX_MASK) >> - CURRENT_STATE_INDEX_SHIFT; + (RREG32(mmTARGET_AND_CURRENT_PROFILE_INDEX) & TARGET_AND_CURRENT_PROFILE_INDEX__CURRENT_STATE_INDEX_MASK) >> + TARGET_AND_CURRENT_PROFILE_INDEX__CURRENT_STATE_INDEX__SHIFT; /* size must be at least 4 bytes for all sensors */ if (*size < 4) diff --git a/drivers/gpu/drm/amd/pm/legacy-dpm/si_smc.c b/drivers/gpu/drm/amd/pm/legacy-dpm/si_smc.c index 8f994ffa9cd1..4e65ab9e931c 100644 --- a/drivers/gpu/drm/amd/pm/legacy-dpm/si_smc.c +++ b/drivers/gpu/drm/amd/pm/legacy-dpm/si_smc.c @@ -30,6 +30,12 @@ #include "amdgpu_ucode.h" #include "sislands_smc.h" +#include "smu/smu_6_0_d.h" +#include "smu/smu_6_0_sh_mask.h" + +#include "gca/gfx_6_0_d.h" +#include "gca/gfx_6_0_sh_mask.h" + static int si_set_smc_sram_address(struct amdgpu_device *adev, u32 smc_address, u32 limit) { @@ -38,8 +44,8 @@ static int si_set_smc_sram_address(struct amdgpu_device *adev, if ((smc_address + 3) > limit) return -EINVAL; - WREG32(SMC_IND_INDEX_0, smc_address); - WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0); + WREG32(mmSMC_IND_INDEX_0, smc_address); + WREG32_P(mmSMC_IND_ACCESS_CNTL, 0, ~SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK); return 0; } @@ -68,7 +74,7 @@ int amdgpu_si_copy_bytes_to_smc(struct amdgpu_device *adev, if (ret) goto done; - WREG32(SMC_IND_DATA_0, data); + WREG32(mmSMC_IND_DATA_0, data); src += 4; byte_count -= 4; @@ -83,7 +89,7 @@ int amdgpu_si_copy_bytes_to_smc(struct amdgpu_device *adev, if (ret) goto done; - original_data = RREG32(SMC_IND_DATA_0); + original_data = RREG32(mmSMC_IND_DATA_0); extra_shift = 8 * (4 - byte_count); while (byte_count > 0) { @@ -99,7 +105,7 @@ int amdgpu_si_copy_bytes_to_smc(struct amdgpu_device *adev, if (ret) goto done; - WREG32(SMC_IND_DATA_0, data); + WREG32(mmSMC_IND_DATA_0, data); } done: @@ -121,10 +127,10 @@ void amdgpu_si_reset_smc(struct amdgpu_device *adev) { u32 tmp; - RREG32(CB_CGTT_SCLK_CTRL); - RREG32(CB_CGTT_SCLK_CTRL); - RREG32(CB_CGTT_SCLK_CTRL); - RREG32(CB_CGTT_SCLK_CTRL); + RREG32(mmCB_CGTT_SCLK_CTRL); + RREG32(mmCB_CGTT_SCLK_CTRL); + RREG32(mmCB_CGTT_SCLK_CTRL); + RREG32(mmCB_CGTT_SCLK_CTRL); tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL) | RST_REG; @@ -170,16 +176,16 @@ PPSMC_Result amdgpu_si_send_msg_to_smc(struct amdgpu_device *adev, if (!amdgpu_si_is_smc_running(adev)) return PPSMC_Result_Failed; - WREG32(SMC_MESSAGE_0, msg); + WREG32(mmSMC_MESSAGE_0, msg); for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(SMC_RESP_0); + tmp = RREG32(mmSMC_RESP_0); if (tmp != 0) break; udelay(1); } - return (PPSMC_Result)RREG32(SMC_RESP_0); + return (PPSMC_Result)RREG32(mmSMC_RESP_0); } PPSMC_Result amdgpu_si_wait_for_smc_inactive(struct amdgpu_device *adev) @@ -225,18 +231,18 @@ int amdgpu_si_load_smc_ucode(struct amdgpu_device *adev, u32 limit) return -EINVAL; spin_lock_irqsave(&adev->smc_idx_lock, flags); - WREG32(SMC_IND_INDEX_0, ucode_start_address); - WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0); + WREG32(mmSMC_IND_INDEX_0, ucode_start_address); + WREG32_P(mmSMC_IND_ACCESS_CNTL, SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK, ~SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK); while (ucode_size >= 4) { /* SMC address space is BE */ data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; - WREG32(SMC_IND_DATA_0, data); + WREG32(mmSMC_IND_DATA_0, data); src += 4; ucode_size -= 4; } - WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0); + WREG32_P(mmSMC_IND_ACCESS_CNTL, 0, ~SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK); spin_unlock_irqrestore(&adev->smc_idx_lock, flags); return 0; @@ -251,7 +257,7 @@ int amdgpu_si_read_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address, spin_lock_irqsave(&adev->smc_idx_lock, flags); ret = si_set_smc_sram_address(adev, smc_address, limit); if (ret == 0) - *value = RREG32(SMC_IND_DATA_0); + *value = RREG32(mmSMC_IND_DATA_0); spin_unlock_irqrestore(&adev->smc_idx_lock, flags); return ret; @@ -266,7 +272,7 @@ int amdgpu_si_write_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address, spin_lock_irqsave(&adev->smc_idx_lock, flags); ret = si_set_smc_sram_address(adev, smc_address, limit); if (ret == 0) - WREG32(SMC_IND_DATA_0, value); + WREG32(mmSMC_IND_DATA_0, value); spin_unlock_irqrestore(&adev->smc_idx_lock, flags); return ret; diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c index 4bd92fd782be..8d40ed0f0e83 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c @@ -143,6 +143,10 @@ int atomctrl_initialize_mc_reg_table( vram_info = (ATOM_VRAM_INFO_HEADER_V2_1 *) smu_atom_get_data_table(hwmgr->adev, GetIndexIntoMasterTable(DATA, VRAM_Info), &size, &frev, &crev); + if (!vram_info) { + pr_err("Could not retrieve the VramInfo table!"); + return -EINVAL; + } if (module_index >= vram_info->ucNumOfVRAMModule) { pr_err("Invalid VramInfo table."); @@ -180,6 +184,10 @@ int atomctrl_initialize_mc_reg_table_v2_2( vram_info = (ATOM_VRAM_INFO_HEADER_V2_2 *) smu_atom_get_data_table(hwmgr->adev, GetIndexIntoMasterTable(DATA, VRAM_Info), &size, &frev, &crev); + if (!vram_info) { + pr_err("Could not retrieve the VramInfo table!"); + return -EINVAL; + } if (module_index >= vram_info->ucNumOfVRAMModule) { pr_err("Invalid VramInfo table."); diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.c index 5a010cd38303..baf51cd82a35 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.c @@ -46,42 +46,6 @@ static int smu7_set_smc_sram_address(struct pp_hwmgr *hwmgr, uint32_t smc_addr, } -int smu7_copy_bytes_from_smc(struct pp_hwmgr *hwmgr, uint32_t smc_start_address, uint32_t *dest, uint32_t byte_count, uint32_t limit) -{ - uint32_t data; - uint32_t addr; - uint8_t *dest_byte; - uint8_t i, data_byte[4] = {0}; - uint32_t *pdata = (uint32_t *)&data_byte; - - PP_ASSERT_WITH_CODE((0 == (3 & smc_start_address)), "SMC address must be 4 byte aligned.", return -EINVAL); - PP_ASSERT_WITH_CODE((limit > (smc_start_address + byte_count)), "SMC address is beyond the SMC RAM area.", return -EINVAL); - - addr = smc_start_address; - - while (byte_count >= 4) { - smu7_read_smc_sram_dword(hwmgr, addr, &data, limit); - - *dest = PP_SMC_TO_HOST_UL(data); - - dest += 1; - byte_count -= 4; - addr += 4; - } - - if (byte_count) { - smu7_read_smc_sram_dword(hwmgr, addr, &data, limit); - *pdata = PP_SMC_TO_HOST_UL(data); - /* Cast dest into byte type in dest_byte. This way, we don't overflow if the allocated memory is not 4-byte aligned. */ - dest_byte = (uint8_t *)dest; - for (i = 0; i < byte_count; i++) - dest_byte[i] = data_byte[i]; - } - - return 0; -} - - int smu7_copy_bytes_to_smc(struct pp_hwmgr *hwmgr, uint32_t smc_start_address, const uint8_t *src, uint32_t byte_count, uint32_t limit) { diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.h b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.h index e7303dc8c260..63e428ceaee4 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.h +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu7_smumgr.h @@ -53,8 +53,6 @@ struct smu7_smumgr { }; -int smu7_copy_bytes_from_smc(struct pp_hwmgr *hwmgr, uint32_t smc_start_address, - uint32_t *dest, uint32_t byte_count, uint32_t limit); int smu7_copy_bytes_to_smc(struct pp_hwmgr *hwmgr, uint32_t smc_start_address, const uint8_t *src, uint32_t byte_count, uint32_t limit); int smu7_program_jump_on_start(struct pp_hwmgr *hwmgr); diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index 46cce1d2aaf3..f24a1d8c77db 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -3432,15 +3432,15 @@ bool smu_mode1_reset_is_support(struct smu_context *smu) return ret; } -bool smu_mode2_reset_is_support(struct smu_context *smu) +bool smu_link_reset_is_support(struct smu_context *smu) { bool ret = false; if (!smu->pm_enabled) return false; - if (smu->ppt_funcs && smu->ppt_funcs->mode2_reset_is_support) - ret = smu->ppt_funcs->mode2_reset_is_support(smu); + if (smu->ppt_funcs && smu->ppt_funcs->link_reset_is_support) + ret = smu->ppt_funcs->link_reset_is_support(smu); return ret; } @@ -3475,6 +3475,19 @@ static int smu_mode2_reset(void *handle) return ret; } +int smu_link_reset(struct smu_context *smu) +{ + int ret = 0; + + if (!smu->pm_enabled) + return -EOPNOTSUPP; + + if (smu->ppt_funcs->link_reset) + ret = smu->ppt_funcs->link_reset(smu); + + return ret; +} + static int smu_enable_gfx_features(void *handle) { struct smu_context *smu = handle; @@ -3975,3 +3988,11 @@ int smu_reset_sdma(struct smu_context *smu, uint32_t inst_mask) return ret; } + +int smu_reset_vcn(struct smu_context *smu, uint32_t inst_mask) +{ + if (smu->ppt_funcs && smu->ppt_funcs->dpm_reset_vcn) + smu->ppt_funcs->dpm_reset_vcn(smu, inst_mask); + + return 0; +} diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h index dd6d0e7aa242..d47e32ae4671 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h @@ -438,9 +438,11 @@ struct mclock_latency_table { }; enum smu_reset_mode { - SMU_RESET_MODE_0, - SMU_RESET_MODE_1, - SMU_RESET_MODE_2, + SMU_RESET_MODE_0, + SMU_RESET_MODE_1, + SMU_RESET_MODE_2, + SMU_RESET_MODE_3, + SMU_RESET_MODE_4, }; enum smu_baco_state { @@ -1229,10 +1231,11 @@ struct pptable_funcs { * @mode1_reset_is_support: Check if GPU supports mode1 reset. */ bool (*mode1_reset_is_support)(struct smu_context *smu); + /** - * @mode2_reset_is_support: Check if GPU supports mode2 reset. + * @link_reset_is_support: Check if GPU supports link reset. */ - bool (*mode2_reset_is_support)(struct smu_context *smu); + bool (*link_reset_is_support)(struct smu_context *smu); /** * @mode1_reset: Perform mode1 reset. @@ -1252,6 +1255,13 @@ struct pptable_funcs { int (*enable_gfx_features)(struct smu_context *smu); /** + * @link_reset: Perform link reset. + * + * The gfx device driver reset + */ + int (*link_reset)(struct smu_context *smu); + + /** * @get_dpm_ultimate_freq: Get the hard frequency range of a clock * domain in MHz. */ @@ -1383,6 +1393,11 @@ struct pptable_funcs { bool (*reset_sdma_is_supported)(struct smu_context *smu); /** + * @reset_vcn: message SMU to soft reset vcn instance. + */ + int (*dpm_reset_vcn)(struct smu_context *smu, uint32_t inst_mask); + + /** * @get_ecc_table: message SMU to get ECC INFO table. */ ssize_t (*get_ecc_info)(struct smu_context *smu, void *table); @@ -1601,8 +1616,9 @@ int smu_get_power_limit(void *handle, enum pp_power_type pp_power_type); bool smu_mode1_reset_is_support(struct smu_context *smu); -bool smu_mode2_reset_is_support(struct smu_context *smu); +bool smu_link_reset_is_support(struct smu_context *smu); int smu_mode1_reset(struct smu_context *smu); +int smu_link_reset(struct smu_context *smu); extern const struct amd_ip_funcs smu_ip_funcs; @@ -1643,6 +1659,7 @@ int smu_send_hbm_bad_channel_flag(struct smu_context *smu, uint32_t size); int smu_send_rma_reason(struct smu_context *smu); int smu_reset_sdma(struct smu_context *smu, uint32_t inst_mask); bool smu_reset_sdma_is_supported(struct smu_context *smu); +int smu_reset_vcn(struct smu_context *smu, uint32_t inst_mask); int smu_set_pm_policy(struct smu_context *smu, enum pp_pm_policy p_type, int level); ssize_t smu_get_pm_policy_info(struct smu_context *smu, diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h index d26f35119a12..3d9e5e967c94 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h @@ -459,4 +459,11 @@ typedef struct __attribute__((packed, aligned(4))) { uint64_t AccGfxclkBelowHostLimit; } VfMetricsTable_t; +#pragma pack(push, 4) +typedef struct { + // Telemetry + uint32_t InputTelemetryVoltageInmV; +} StaticMetricsTable_t; +#pragma pack(pop) + #endif diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h index 288b2576432b..41f268313613 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h @@ -94,7 +94,9 @@ #define PPSMC_MSG_RmaDueToBadPageThreshold 0x43 #define PPSMC_MSG_SetThrottlingPolicy 0x44 #define PPSMC_MSG_ResetSDMA 0x4D -#define PPSMC_Message_Count 0x4E +#define PPSMC_MSG_ResetVCN 0x4E +#define PPSMC_MSG_GetStaticMetricsTable 0x59 +#define PPSMC_Message_Count 0x5A //PPSMC Reset Types for driver msg argument #define PPSMC_RESET_TYPE_DRIVER_MODE_1_RESET 0x1 diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h index c9dee09395e3..eefdaa0b5df6 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h @@ -277,6 +277,7 @@ __SMU_DUMMY_MAP(MALLPowerController), \ __SMU_DUMMY_MAP(MALLPowerState), \ __SMU_DUMMY_MAP(ResetSDMA), \ + __SMU_DUMMY_MAP(ResetVCN), \ __SMU_DUMMY_MAP(GetStaticMetricsTable), #undef __SMU_DUMMY_MAP diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v11_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v11_0.h index ed8304d82831..56ae555bb52a 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v11_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v11_0.h @@ -281,11 +281,6 @@ int smu_v11_0_set_single_dpm_table(struct smu_context *smu, enum smu_clk_type clk_type, struct smu_11_0_dpm_table *single_dpm_table); -int smu_v11_0_get_dpm_level_range(struct smu_context *smu, - enum smu_clk_type clk_type, - uint32_t *min_value, - uint32_t *max_value); - int smu_v11_0_get_current_pcie_link_width_level(struct smu_context *smu); uint16_t smu_v11_0_get_current_pcie_link_width(struct smu_context *smu); diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h index cd03caffe317..4263798d716b 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h @@ -112,6 +112,7 @@ struct smu_13_0_dpm_context { uint32_t workload_policy_mask; uint32_t dcef_min_ds_clk; uint64_t caps; + uint32_t board_volt; }; enum smu_13_0_power_state { @@ -162,8 +163,6 @@ int smu_v13_0_notify_memory_pool_location(struct smu_context *smu); int smu_v13_0_system_features_control(struct smu_context *smu, bool en); -int smu_v13_0_init_display_count(struct smu_context *smu, uint32_t count); - int smu_v13_0_set_allowed_mask(struct smu_context *smu); int smu_v13_0_notify_display_change(struct smu_context *smu); @@ -183,13 +182,6 @@ int smu_v13_0_disable_thermal_alert(struct smu_context *smu); int smu_v13_0_get_gfx_vdd(struct smu_context *smu, uint32_t *value); -int smu_v13_0_set_min_deep_sleep_dcefclk(struct smu_context *smu, uint32_t clk); - -int -smu_v13_0_display_clock_voltage_request(struct smu_context *smu, - struct pp_display_clock_request - *clock_req); - uint32_t smu_v13_0_get_fan_control_mode(struct smu_context *smu); @@ -226,11 +218,6 @@ int smu_v13_0_get_dpm_ultimate_freq(struct smu_context *smu, enum smu_clk_type c int smu_v13_0_set_soft_freq_limited_range(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t min, uint32_t max, bool automatic); -int smu_v13_0_set_hard_freq_limited_range(struct smu_context *smu, - enum smu_clk_type clk_type, - uint32_t min, - uint32_t max); - int smu_v13_0_set_performance_level(struct smu_context *smu, enum amd_dpm_forced_level level); @@ -310,14 +297,6 @@ int smu_v13_0_get_boot_freq_by_index(struct smu_context *smu, uint32_t *value); void smu_v13_0_interrupt_work(struct smu_context *smu); -bool smu_v13_0_12_is_dpm_running(struct smu_context *smu); -int smu_v13_0_12_get_max_metrics_size(void); -int smu_v13_0_12_setup_driver_pptable(struct smu_context *smu); -int smu_v13_0_12_get_smu_metrics_data(struct smu_context *smu, - MetricsMember_t member, - uint32_t *value); -ssize_t smu_v13_0_12_get_gpu_metrics(struct smu_context *smu, void **table); -extern const struct cmn2asic_mapping smu_v13_0_12_feature_mask_map[]; -extern const struct cmn2asic_msg_mapping smu_v13_0_12_message_map[]; +void smu_v13_0_reset_custom_level(struct smu_context *smu); #endif #endif diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c index 453952cdc353..9ad46f545d15 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c @@ -1347,7 +1347,7 @@ static int arcturus_get_power_limit(struct smu_context *smu, *default_power_limit = power_limit; if (max_power_limit) *max_power_limit = power_limit; - /** + /* * No lower bound is imposed on the limit. Any unreasonable limit set * will result in frequent throttling. */ diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c index 19a25fdc2f5b..115e3fa456bc 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c @@ -3089,11 +3089,6 @@ static int sienna_cichlid_stb_get_data_direct(struct smu_context *smu, return 0; } -static bool sienna_cichlid_is_mode2_reset_supported(struct smu_context *smu) -{ - return true; -} - static int sienna_cichlid_mode2_reset(struct smu_context *smu) { int ret = 0, index; @@ -3229,7 +3224,6 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = { .get_default_config_table_settings = sienna_cichlid_get_default_config_table_settings, .set_config_table = sienna_cichlid_set_config_table, .get_unique_id = sienna_cichlid_get_unique_id, - .mode2_reset_is_support = sienna_cichlid_is_mode2_reset_supported, .mode2_reset = sienna_cichlid_mode2_reset, }; diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c index 25fabf336a64..78e4186d06cc 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c @@ -2059,45 +2059,6 @@ int smu_v11_0_set_single_dpm_table(struct smu_context *smu, return 0; } -int smu_v11_0_get_dpm_level_range(struct smu_context *smu, - enum smu_clk_type clk_type, - uint32_t *min_value, - uint32_t *max_value) -{ - uint32_t level_count = 0; - int ret = 0; - - if (!min_value && !max_value) - return -EINVAL; - - if (min_value) { - /* by default, level 0 clock value as min value */ - ret = smu_v11_0_get_dpm_freq_by_index(smu, - clk_type, - 0, - min_value); - if (ret) - return ret; - } - - if (max_value) { - ret = smu_v11_0_get_dpm_level_count(smu, - clk_type, - &level_count); - if (ret) - return ret; - - ret = smu_v11_0_get_dpm_freq_by_index(smu, - clk_type, - level_count - 1, - max_value); - if (ret) - return ret; - } - - return ret; -} - int smu_v11_0_get_current_pcie_link_width_level(struct smu_context *smu) { struct amdgpu_device *adev = smu->adev; diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c index 83163d7c7f00..6de653d2ed62 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c @@ -1270,6 +1270,7 @@ static int aldebaran_set_performance_level(struct smu_context *smu, struct smu_13_0_dpm_table *gfx_table = &dpm_context->dpm_tables.gfx_table; struct smu_umd_pstate_table *pstate_table = &smu->pstate_table; + int r; /* Disable determinism if switching to another mode */ if ((smu_dpm->dpm_level == AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM) && @@ -1282,7 +1283,11 @@ static int aldebaran_set_performance_level(struct smu_context *smu, case AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM: return 0; - + case AMD_DPM_FORCED_LEVEL_AUTO: + r = smu_v13_0_set_performance_level(smu, level); + if (!r) + smu_v13_0_reset_custom_level(smu); + return r; case AMD_DPM_FORCED_LEVEL_HIGH: case AMD_DPM_FORCED_LEVEL_LOW: case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD: @@ -1423,7 +1428,11 @@ static int aldebaran_usr_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_ min_clk = dpm_context->dpm_tables.gfx_table.min; max_clk = dpm_context->dpm_tables.gfx_table.max; - return aldebaran_set_soft_freq_limited_range(smu, SMU_GFXCLK, min_clk, max_clk, false); + ret = aldebaran_set_soft_freq_limited_range( + smu, SMU_GFXCLK, min_clk, max_clk, false); + if (ret) + return ret; + smu_v13_0_reset_custom_level(smu); } break; case PP_OD_COMMIT_DPM_TABLE: @@ -1976,11 +1985,6 @@ static bool aldebaran_is_mode1_reset_supported(struct smu_context *smu) return true; } -static bool aldebaran_is_mode2_reset_supported(struct smu_context *smu) -{ - return true; -} - static int aldebaran_set_mp1_state(struct smu_context *smu, enum pp_mp1_state mp1_state) { @@ -2086,7 +2090,6 @@ static const struct pptable_funcs aldebaran_ppt_funcs = { .set_pp_feature_mask = smu_cmn_set_pp_feature_mask, .get_gpu_metrics = aldebaran_get_gpu_metrics, .mode1_reset_is_support = aldebaran_is_mode1_reset_supported, - .mode2_reset_is_support = aldebaran_is_mode2_reset_supported, .smu_handle_passthrough_sbr = aldebaran_smu_handle_passthrough_sbr, .mode1_reset = aldebaran_mode1_reset, .set_mp1_state = aldebaran_set_mp1_state, diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c index ba5a9012dbd5..a7167668d189 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c @@ -709,18 +709,6 @@ int smu_v13_0_notify_memory_pool_location(struct smu_context *smu) return ret; } -int smu_v13_0_set_min_deep_sleep_dcefclk(struct smu_context *smu, uint32_t clk) -{ - int ret; - - ret = smu_cmn_send_smc_msg_with_param(smu, - SMU_MSG_SetMinDeepSleepDcefclk, clk, NULL); - if (ret) - dev_err(smu->adev->dev, "SMU13 attempt to set divider for DCEFCLK Failed!"); - - return ret; -} - int smu_v13_0_set_driver_table_location(struct smu_context *smu) { struct smu_table *driver_table = &smu->smu_table.driver_table; @@ -761,18 +749,6 @@ int smu_v13_0_set_tool_table_location(struct smu_context *smu) return ret; } -int smu_v13_0_init_display_count(struct smu_context *smu, uint32_t count) -{ - int ret = 0; - - if (!smu->pm_enabled) - return ret; - - ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_NumOfDisplays, count, NULL); - - return ret; -} - int smu_v13_0_set_allowed_mask(struct smu_context *smu) { struct smu_feature *feature = &smu->smu_feature; @@ -1073,56 +1049,6 @@ int smu_v13_0_get_gfx_vdd(struct smu_context *smu, uint32_t *value) } -int -smu_v13_0_display_clock_voltage_request(struct smu_context *smu, - struct pp_display_clock_request - *clock_req) -{ - enum amd_pp_clock_type clk_type = clock_req->clock_type; - int ret = 0; - enum smu_clk_type clk_select = 0; - uint32_t clk_freq = clock_req->clock_freq_in_khz / 1000; - - if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_DPM_DCEFCLK_BIT) || - smu_cmn_feature_is_enabled(smu, SMU_FEATURE_DPM_UCLK_BIT)) { - switch (clk_type) { - case amd_pp_dcef_clock: - clk_select = SMU_DCEFCLK; - break; - case amd_pp_disp_clock: - clk_select = SMU_DISPCLK; - break; - case amd_pp_pixel_clock: - clk_select = SMU_PIXCLK; - break; - case amd_pp_phy_clock: - clk_select = SMU_PHYCLK; - break; - case amd_pp_mem_clock: - clk_select = SMU_UCLK; - break; - default: - dev_info(smu->adev->dev, "[%s] Invalid Clock Type!", __func__); - ret = -EINVAL; - break; - } - - if (ret) - goto failed; - - if (clk_select == SMU_UCLK && smu->disable_uclk_switch) - return 0; - - ret = smu_v13_0_set_hard_freq_limited_range(smu, clk_select, clk_freq, 0); - - if (clk_select == SMU_UCLK) - smu->hard_min_uclk_req_from_dal = clk_freq; - } - -failed: - return ret; -} - uint32_t smu_v13_0_get_fan_control_mode(struct smu_context *smu) { if (!smu_cmn_feature_is_enabled(smu, SMU_FEATURE_FAN_CONTROL_BIT)) @@ -1647,45 +1573,6 @@ out: return ret; } -int smu_v13_0_set_hard_freq_limited_range(struct smu_context *smu, - enum smu_clk_type clk_type, - uint32_t min, - uint32_t max) -{ - int ret = 0, clk_id = 0; - uint32_t param; - - if (min <= 0 && max <= 0) - return -EINVAL; - - if (!smu_cmn_clk_dpm_is_enabled(smu, clk_type)) - return 0; - - clk_id = smu_cmn_to_asic_specific_index(smu, - CMN2ASIC_MAPPING_CLK, - clk_type); - if (clk_id < 0) - return clk_id; - - if (max > 0) { - param = (uint32_t)((clk_id << 16) | (max & 0xffff)); - ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetHardMaxByFreq, - param, NULL); - if (ret) - return ret; - } - - if (min > 0) { - param = (uint32_t)((clk_id << 16) | (min & 0xffff)); - ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetHardMinByFreq, - param, NULL); - if (ret) - return ret; - } - - return ret; -} - int smu_v13_0_set_performance_level(struct smu_context *smu, enum amd_dpm_forced_level level) { @@ -2595,3 +2482,13 @@ int smu_v13_0_set_wbrf_exclusion_ranges(struct smu_context *smu, return ret; } + +void smu_v13_0_reset_custom_level(struct smu_context *smu) +{ + struct smu_umd_pstate_table *pstate_table = &smu->pstate_table; + + pstate_table->uclk_pstate.custom.min = 0; + pstate_table->uclk_pstate.custom.max = 0; + pstate_table->gfxclk_pstate.custom.min = 0; + pstate_table->gfxclk_pstate.custom.max = 0; +} diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c index 238bd71baa6d..533d58e57d05 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c @@ -187,26 +187,6 @@ int smu_v13_0_12_get_max_metrics_size(void) return max(sizeof(StaticMetricsTable_t), sizeof(MetricsTable_t)); } -static int smu_v13_0_12_get_static_metrics_table(struct smu_context *smu) -{ - struct smu_table_context *smu_table = &smu->smu_table; - uint32_t table_size = smu_table->tables[SMU_TABLE_SMU_METRICS].size; - struct smu_table *table = &smu_table->driver_table; - int ret; - - ret = smu_cmn_send_smc_msg(smu, SMU_MSG_GetStaticMetricsTable, NULL); - if (ret) { - dev_info(smu->adev->dev, - "Failed to export static metrics table!\n"); - return ret; - } - - amdgpu_asic_invalidate_hdp(smu->adev, NULL); - memcpy(smu_table->metrics_table, table->cpu_addr, table_size); - - return 0; -} - int smu_v13_0_12_setup_driver_pptable(struct smu_context *smu) { struct smu_table_context *smu_table = &smu->smu_table; @@ -217,7 +197,7 @@ int smu_v13_0_12_setup_driver_pptable(struct smu_context *smu) int ret, i; if (!pptable->Init) { - ret = smu_v13_0_12_get_static_metrics_table(smu); + ret = smu_v13_0_6_get_static_metrics_table(smu); if (ret) return ret; @@ -345,8 +325,8 @@ int smu_v13_0_12_get_smu_metrics_data(struct smu_context *smu, ssize_t smu_v13_0_12_get_gpu_metrics(struct smu_context *smu, void **table) { struct smu_table_context *smu_table = &smu->smu_table; - struct gpu_metrics_v1_7 *gpu_metrics = - (struct gpu_metrics_v1_7 *)smu_table->gpu_metrics_table; + struct gpu_metrics_v1_8 *gpu_metrics = + (struct gpu_metrics_v1_8 *)smu_table->gpu_metrics_table; int ret = 0, xcc_id, inst, i, j, k, idx; struct amdgpu_device *adev = smu->adev; u8 num_jpeg_rings_gpu_metrics; @@ -357,7 +337,7 @@ ssize_t smu_v13_0_12_get_gpu_metrics(struct smu_context *smu, void **table) metrics = kzalloc(sizeof(MetricsTable_t), GFP_KERNEL); memcpy(metrics, smu_table->metrics_table, sizeof(MetricsTable_t)); - smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 7); + smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 8); gpu_metrics->temperature_hotspot = SMUQ10_ROUND(metrics->MaxSocketTemperature); @@ -474,6 +454,16 @@ ssize_t smu_v13_0_12_get_gpu_metrics(struct smu_context *smu, void **table) SMUQ10_ROUND(metrics->GfxBusy[inst]); gpu_metrics->xcp_stats[i].gfx_busy_acc[idx] = SMUQ10_ROUND(metrics->GfxBusyAcc[inst]); + if (smu_v13_0_6_cap_supported(smu, SMU_CAP(HST_LIMIT_METRICS))) { + gpu_metrics->xcp_stats[i].gfx_below_host_limit_ppt_acc[idx] = + SMUQ10_ROUND(metrics->GfxclkBelowHostLimitPptAcc[inst]); + gpu_metrics->xcp_stats[i].gfx_below_host_limit_thm_acc[idx] = + SMUQ10_ROUND(metrics->GfxclkBelowHostLimitThmAcc[inst]); + gpu_metrics->xcp_stats[i].gfx_low_utilization_acc[idx] = + SMUQ10_ROUND(metrics->GfxclkLowUtilizationAcc[inst]); + gpu_metrics->xcp_stats[i].gfx_below_host_limit_total_acc[idx] = + SMUQ10_ROUND(metrics->GfxclkBelowHostLimitTotalAcc[inst]); + } idx++; } } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c index c478b3be37af..7d4ff09be7e8 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c @@ -101,24 +101,6 @@ MODULE_FIRMWARE("amdgpu/smu_13_0_14.bin"); #define MCA_BANK_IPID(_ip, _hwid, _type) \ [AMDGPU_MCA_IP_##_ip] = { .hwid = _hwid, .mcatype = _type, } -#define SMU_CAP(x) SMU_13_0_6_CAPS_##x - -enum smu_v13_0_6_caps { - SMU_CAP(DPM), - SMU_CAP(DPM_POLICY), - SMU_CAP(OTHER_END_METRICS), - SMU_CAP(SET_UCLK_MAX), - SMU_CAP(PCIE_METRICS), - SMU_CAP(MCA_DEBUG_MODE), - SMU_CAP(PER_INST_METRICS), - SMU_CAP(CTF_LIMIT), - SMU_CAP(RMA_MSG), - SMU_CAP(ACA_SYND), - SMU_CAP(SDMA_RESET), - SMU_CAP(STATIC_METRICS), - SMU_CAP(ALL), -}; - struct mca_bank_ipid { enum amdgpu_mca_ip ip; uint16_t hwid; @@ -194,6 +176,8 @@ static const struct cmn2asic_msg_mapping smu_v13_0_6_message_map[SMU_MSG_MAX_COU MSG_MAP(RmaDueToBadPageThreshold, PPSMC_MSG_RmaDueToBadPageThreshold, 0), MSG_MAP(SetThrottlingPolicy, PPSMC_MSG_SetThrottlingPolicy, 0), MSG_MAP(ResetSDMA, PPSMC_MSG_ResetSDMA, 0), + MSG_MAP(ResetVCN, PPSMC_MSG_ResetVCN, 0), + MSG_MAP(GetStaticMetricsTable, PPSMC_MSG_GetStaticMetricsTable, 0), }; // clang-format on @@ -299,8 +283,8 @@ static inline void smu_v13_0_6_cap_clear(struct smu_context *smu, dpm_context->caps &= ~BIT_ULL(cap); } -static inline bool smu_v13_0_6_cap_supported(struct smu_context *smu, - enum smu_v13_0_6_caps cap) +bool smu_v13_0_6_cap_supported(struct smu_context *smu, + enum smu_v13_0_6_caps cap) { struct smu_13_0_dpm_context *dpm_context = smu->smu_dpm.dpm_context; @@ -353,6 +337,9 @@ static void smu_v13_0_12_init_caps(struct smu_context *smu) if (fw_ver >= 0x00561E00) smu_v13_0_6_cap_set(smu, SMU_CAP(STATIC_METRICS)); + + if (fw_ver >= 0x00562500) + smu_v13_0_6_cap_set(smu, SMU_CAP(HST_LIMIT_METRICS)); } static void smu_v13_0_6_init_caps(struct smu_context *smu) @@ -402,6 +389,13 @@ static void smu_v13_0_6_init_caps(struct smu_context *smu) smu_v13_0_6_cap_clear(smu, SMU_CAP(RMA_MSG)); if (fw_ver < 0x00555600) smu_v13_0_6_cap_clear(smu, SMU_CAP(ACA_SYND)); + if ((pgm == 7 && fw_ver >= 0x7550E00) || + (pgm == 0 && fw_ver >= 0x00557E00)) + smu_v13_0_6_cap_set(smu, SMU_CAP(HST_LIMIT_METRICS)); + if (fw_ver >= 0x00557F01) { + smu_v13_0_6_cap_set(smu, SMU_CAP(STATIC_METRICS)); + smu_v13_0_6_cap_set(smu, SMU_CAP(BOARD_VOLTAGE)); + } } if (((pgm == 7) && (fw_ver >= 0x7550700)) || ((pgm == 0) && (fw_ver >= 0x00557900)) || @@ -525,7 +519,7 @@ static int smu_v13_0_6_tables_init(struct smu_context *smu) return -ENOMEM; smu_table->metrics_time = 0; - smu_table->gpu_metrics_table_size = sizeof(struct gpu_metrics_v1_7); + smu_table->gpu_metrics_table_size = sizeof(struct gpu_metrics_v1_8); smu_table->gpu_metrics_table = kzalloc(smu_table->gpu_metrics_table_size, GFP_KERNEL); if (!smu_table->gpu_metrics_table) { @@ -747,9 +741,43 @@ static ssize_t smu_v13_0_6_get_pm_metrics(struct smu_context *smu, return pm_metrics->common_header.structure_size; } +static void smu_v13_0_6_fill_static_metrics_table(struct smu_context *smu, + StaticMetricsTable_t *static_metrics) +{ + struct smu_13_0_dpm_context *dpm_context = smu->smu_dpm.dpm_context; + + if (!static_metrics->InputTelemetryVoltageInmV) { + dev_warn(smu->adev->dev, "Invalid board voltage %d\n", + static_metrics->InputTelemetryVoltageInmV); + } + + dpm_context->board_volt = static_metrics->InputTelemetryVoltageInmV; +} + +int smu_v13_0_6_get_static_metrics_table(struct smu_context *smu) +{ + struct smu_table_context *smu_table = &smu->smu_table; + uint32_t table_size = smu_table->tables[SMU_TABLE_SMU_METRICS].size; + struct smu_table *table = &smu_table->driver_table; + int ret; + + ret = smu_cmn_send_smc_msg(smu, SMU_MSG_GetStaticMetricsTable, NULL); + if (ret) { + dev_info(smu->adev->dev, + "Failed to export static metrics table!\n"); + return ret; + } + + amdgpu_asic_invalidate_hdp(smu->adev, NULL); + memcpy(smu_table->metrics_table, table->cpu_addr, table_size); + + return 0; +} + static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) { struct smu_table_context *smu_table = &smu->smu_table; + StaticMetricsTable_t *static_metrics = (StaticMetricsTable_t *)smu_table->metrics_table; MetricsTableV0_t *metrics_v0 = (MetricsTableV0_t *)smu_table->metrics_table; MetricsTableV1_t *metrics_v1 = (MetricsTableV1_t *)smu_table->metrics_table; MetricsTableV2_t *metrics_v2 = (MetricsTableV2_t *)smu_table->metrics_table; @@ -759,7 +787,8 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) int ret, i, retry = 100; uint32_t table_version; - if (smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) + if (amdgpu_ip_version(smu->adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 12) && + smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) return smu_v13_0_12_setup_driver_pptable(smu); /* Store one-time values in driver PPTable */ @@ -813,6 +842,12 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) GET_METRIC_FIELD(PublicSerialNumber_AID, version)[0]; pptable->Init = true; + if (smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) { + ret = smu_v13_0_6_get_static_metrics_table(smu); + if (ret) + return ret; + smu_v13_0_6_fill_static_metrics_table(smu, static_metrics); + } } return 0; @@ -1142,7 +1177,8 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu, if (ret) return ret; - if (smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) + if (amdgpu_ip_version(smu->adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 12) && + smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) return smu_v13_0_12_get_smu_metrics_data(smu, member, value); /* For clocks with multiple instances, only report the first one */ @@ -1616,6 +1652,7 @@ static int smu_v13_0_6_read_sensor(struct smu_context *smu, enum amd_pp_sensors sensor, void *data, uint32_t *size) { + struct smu_13_0_dpm_context *dpm_context = smu->smu_dpm.dpm_context; int ret = 0; if (amdgpu_ras_intr_triggered()) @@ -1660,6 +1697,15 @@ static int smu_v13_0_6_read_sensor(struct smu_context *smu, ret = smu_v13_0_get_gfx_vdd(smu, (uint32_t *)data); *size = 4; break; + case AMDGPU_PP_SENSOR_VDDBOARD: + if (smu_v13_0_6_cap_supported(smu, SMU_CAP(BOARD_VOLTAGE))) { + *(uint32_t *)data = dpm_context->board_volt; + *size = 4; + break; + } else { + ret = -EOPNOTSUPP; + break; + } case AMDGPU_PP_SENSOR_GPU_AVG_POWER: default: ret = -EOPNOTSUPP; @@ -1927,7 +1973,7 @@ static int smu_v13_0_6_set_performance_level(struct smu_context *smu, return ret; pstate_table->uclk_pstate.curr.max = uclk_table->max; } - pstate_table->uclk_pstate.custom.max = 0; + smu_v13_0_reset_custom_level(smu); return 0; case AMD_DPM_FORCED_LEVEL_MANUAL: @@ -2140,7 +2186,7 @@ static int smu_v13_0_6_usr_edit_dpm_table(struct smu_context *smu, smu, SMU_UCLK, min_clk, max_clk, false); if (ret) return ret; - pstate_table->uclk_pstate.custom.max = 0; + smu_v13_0_reset_custom_level(smu); } break; case PP_OD_COMMIT_DPM_TABLE: @@ -2486,8 +2532,8 @@ static int smu_v13_0_6_get_current_pcie_link_speed(struct smu_context *smu) static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table) { struct smu_table_context *smu_table = &smu->smu_table; - struct gpu_metrics_v1_7 *gpu_metrics = - (struct gpu_metrics_v1_7 *)smu_table->gpu_metrics_table; + struct gpu_metrics_v1_8 *gpu_metrics = + (struct gpu_metrics_v1_8 *)smu_table->gpu_metrics_table; int version = smu_v13_0_6_get_metrics_version(smu); int ret = 0, xcc_id, inst, i, j, k, idx; struct amdgpu_device *adev = smu->adev; @@ -2507,13 +2553,14 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table return ret; } - if (smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) + if (amdgpu_ip_version(smu->adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 12) && + smu_v13_0_6_cap_supported(smu, SMU_CAP(STATIC_METRICS))) return smu_v13_0_12_get_gpu_metrics(smu, table); metrics_v1 = (MetricsTableV1_t *)metrics_v0; metrics_v2 = (MetricsTableV2_t *)metrics_v0; - smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 7); + smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 8); gpu_metrics->temperature_hotspot = SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature, version)); @@ -2666,6 +2713,20 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table gpu_metrics->xcp_stats[i].gfx_busy_acc[idx] = SMUQ10_ROUND(GET_GPU_METRIC_FIELD(GfxBusyAcc, version)[inst]); + if (smu_v13_0_6_cap_supported(smu, SMU_CAP(HST_LIMIT_METRICS))) { + gpu_metrics->xcp_stats[i].gfx_below_host_limit_ppt_acc[idx] = + SMUQ10_ROUND + (metrics_v0->GfxclkBelowHostLimitPptAcc[inst]); + gpu_metrics->xcp_stats[i].gfx_below_host_limit_thm_acc[idx] = + SMUQ10_ROUND + (metrics_v0->GfxclkBelowHostLimitThmAcc[inst]); + gpu_metrics->xcp_stats[i].gfx_low_utilization_acc[idx] = + SMUQ10_ROUND + (metrics_v0->GfxclkLowUtilizationAcc[inst]); + gpu_metrics->xcp_stats[i].gfx_below_host_limit_total_acc[idx] = + SMUQ10_ROUND + (metrics_v0->GfxclkBelowHostLimitTotalAcc[inst]); + } idx++; } } @@ -2844,14 +2905,29 @@ static int smu_v13_0_6_mode1_reset(struct smu_context *smu) return ret; } +static int smu_v13_0_6_link_reset(struct smu_context *smu) +{ + int ret = 0; + + ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_GfxDeviceDriverReset, + SMU_RESET_MODE_4, NULL); + return ret; +} + static bool smu_v13_0_6_is_mode1_reset_supported(struct smu_context *smu) { return true; } -static bool smu_v13_0_6_is_mode2_reset_supported(struct smu_context *smu) +static inline bool smu_v13_0_6_is_link_reset_supported(struct smu_context *smu) { - return true; + struct amdgpu_device *adev = smu->adev; + int var = (adev->pdev->device & 0xF); + + if (var == 0x1) + return true; + + return false; } static int smu_v13_0_6_smu_send_hbm_bad_page_num(struct smu_context *smu, @@ -2924,6 +3000,19 @@ static int smu_v13_0_6_reset_sdma(struct smu_context *smu, uint32_t inst_mask) return ret; } +static int smu_v13_0_6_reset_vcn(struct smu_context *smu, uint32_t inst_mask) +{ + int ret = 0; + + ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_ResetVCN, inst_mask, NULL); + if (ret) + dev_err(smu->adev->dev, + "failed to send ResetVCN event with mask 0x%x\n", + inst_mask); + return ret; +} + + static int mca_smu_set_debug_mode(struct amdgpu_device *adev, bool enable) { struct smu_context *smu = adev->powerplay.pp_handle; @@ -3586,9 +3675,10 @@ static const struct pptable_funcs smu_v13_0_6_ppt_funcs = { .get_pm_metrics = smu_v13_0_6_get_pm_metrics, .get_thermal_temperature_range = smu_v13_0_6_get_thermal_temperature_range, .mode1_reset_is_support = smu_v13_0_6_is_mode1_reset_supported, - .mode2_reset_is_support = smu_v13_0_6_is_mode2_reset_supported, + .link_reset_is_support = smu_v13_0_6_is_link_reset_supported, .mode1_reset = smu_v13_0_6_mode1_reset, .mode2_reset = smu_v13_0_6_mode2_reset, + .link_reset = smu_v13_0_6_link_reset, .wait_for_event = smu_v13_0_wait_for_event, .i2c_init = smu_v13_0_6_i2c_control_init, .i2c_fini = smu_v13_0_6_i2c_control_fini, @@ -3596,6 +3686,7 @@ static const struct pptable_funcs smu_v13_0_6_ppt_funcs = { .send_rma_reason = smu_v13_0_6_send_rma_reason, .reset_sdma = smu_v13_0_6_reset_sdma, .reset_sdma_is_supported = smu_v13_0_6_reset_sdma_is_supported, + .dpm_reset_vcn = smu_v13_0_6_reset_vcn, }; void smu_v13_0_6_set_ppt_funcs(struct smu_context *smu) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.h b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.h index 83745909e564..d151bcd0cca7 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.h +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.h @@ -26,6 +26,7 @@ #define SMU_13_0_6_UMD_PSTATE_GFXCLK_LEVEL 0x2 #define SMU_13_0_6_UMD_PSTATE_SOCCLK_LEVEL 0x4 #define SMU_13_0_6_UMD_PSTATE_MCLK_LEVEL 0x2 +#define SMU_CAP(x) SMU_13_0_6_CAPS_##x typedef enum { /*0*/ METRICS_VERSION_V0 = 0, @@ -51,6 +52,34 @@ struct PPTable_t { bool Init; }; +enum smu_v13_0_6_caps { + SMU_CAP(DPM), + SMU_CAP(DPM_POLICY), + SMU_CAP(OTHER_END_METRICS), + SMU_CAP(SET_UCLK_MAX), + SMU_CAP(PCIE_METRICS), + SMU_CAP(MCA_DEBUG_MODE), + SMU_CAP(PER_INST_METRICS), + SMU_CAP(CTF_LIMIT), + SMU_CAP(RMA_MSG), + SMU_CAP(ACA_SYND), + SMU_CAP(SDMA_RESET), + SMU_CAP(STATIC_METRICS), + SMU_CAP(HST_LIMIT_METRICS), + SMU_CAP(BOARD_VOLTAGE), + SMU_CAP(ALL), +}; + extern void smu_v13_0_6_set_ppt_funcs(struct smu_context *smu); +bool smu_v13_0_6_cap_supported(struct smu_context *smu, enum smu_v13_0_6_caps cap); +int smu_v13_0_6_get_static_metrics_table(struct smu_context *smu); +bool smu_v13_0_12_is_dpm_running(struct smu_context *smu); +int smu_v13_0_12_get_max_metrics_size(void); +int smu_v13_0_12_setup_driver_pptable(struct smu_context *smu); +int smu_v13_0_12_get_smu_metrics_data(struct smu_context *smu, + MetricsMember_t member, uint32_t *value); +ssize_t smu_v13_0_12_get_gpu_metrics(struct smu_context *smu, void **table); +extern const struct cmn2asic_mapping smu_v13_0_12_feature_mask_map[]; +extern const struct cmn2asic_msg_mapping smu_v13_0_12_message_map[]; #endif diff --git a/drivers/gpu/drm/ci/build-igt.sh b/drivers/gpu/drm/ci/build-igt.sh index eddb5f782a5e..caa2f4804ed5 100644 --- a/drivers/gpu/drm/ci/build-igt.sh +++ b/drivers/gpu/drm/ci/build-igt.sh @@ -71,4 +71,4 @@ tar -cf artifacts/igt.tar /igt # Pass needed files to the test stage S3_ARTIFACT_NAME="igt.tar.gz" gzip -c artifacts/igt.tar > ${S3_ARTIFACT_NAME} -ci-fairy s3cp --token-file "${S3_JWT_FILE}" ${S3_ARTIFACT_NAME} https://${PIPELINE_ARTIFACTS_BASE}/${KERNEL_ARCH}/${S3_ARTIFACT_NAME} +s3_upload ${S3_ARTIFACT_NAME} https://${PIPELINE_ARTIFACTS_BASE}/${KERNEL_ARCH}/ diff --git a/drivers/gpu/drm/ci/build.sh b/drivers/gpu/drm/ci/build.sh index 284873e94d8d..6fb74c51abe2 100644 --- a/drivers/gpu/drm/ci/build.sh +++ b/drivers/gpu/drm/ci/build.sh @@ -148,13 +148,13 @@ if [[ "$UPLOAD_TO_MINIO" = "1" ]]; then ls -l "${S3_JWT_FILE}" for f in $FILES_TO_UPLOAD; do - ci-fairy s3cp --token-file "${S3_JWT_FILE}" /kernel/$f \ - https://${PIPELINE_ARTIFACTS_BASE}/${DEBIAN_ARCH}/$f + s3_upload /kernel/$f \ + https://${PIPELINE_ARTIFACTS_BASE}/${DEBIAN_ARCH}/ done S3_ARTIFACT_NAME="kernel-files.tar.zst" tar --zstd -cf $S3_ARTIFACT_NAME install - ci-fairy s3cp --token-file "${S3_JWT_FILE}" ${S3_ARTIFACT_NAME} https://${PIPELINE_ARTIFACTS_BASE}/${DEBIAN_ARCH}/${S3_ARTIFACT_NAME} + s3_upload ${S3_ARTIFACT_NAME} https://${PIPELINE_ARTIFACTS_BASE}/${DEBIAN_ARCH}/ echo "Download vmlinux.xz from https://${PIPELINE_ARTIFACTS_BASE}/${DEBIAN_ARCH}/vmlinux.xz" fi diff --git a/drivers/gpu/drm/ci/gitlab-ci.yml b/drivers/gpu/drm/ci/gitlab-ci.yml index 65adcd97e06b..ba75b3a7eca4 100644 --- a/drivers/gpu/drm/ci/gitlab-ci.yml +++ b/drivers/gpu/drm/ci/gitlab-ci.yml @@ -1,6 +1,6 @@ variables: DRM_CI_PROJECT_PATH: &drm-ci-project-path mesa/mesa - DRM_CI_COMMIT_SHA: &drm-ci-commit-sha 82ab58f6c6f94fa80ca7e1615146f08356e3ba69 + DRM_CI_COMMIT_SHA: &drm-ci-commit-sha f73132f1215a37ce8ffc711a0136c90649aaf128 UPSTREAM_REPO: https://gitlab.freedesktop.org/drm/kernel.git TARGET_BRANCH: drm-next @@ -20,8 +20,10 @@ variables: rm download-git-cache.sh set +o xtrace S3_JWT_FILE: /s3_jwt + S3_JWT_HEADER_FILE: /s3_jwt_header S3_JWT_FILE_SCRIPT: |- echo -n '${S3_JWT}' > '${S3_JWT_FILE}' && + echo -n "Authorization: Bearer ${S3_JWT}" > '${S3_JWT_HEADER_FILE}' && unset CI_JOB_JWT S3_JWT # Unsetting vulnerable env variables S3_HOST: s3.freedesktop.org # This bucket is used to fetch the kernel image @@ -251,7 +253,7 @@ make git archive: - tar -cvzf ../$CI_PROJECT_NAME.tar.gz . # Use id_tokens for JWT auth - - ci-fairy s3cp --token-file "${S3_JWT_FILE}" ../$CI_PROJECT_NAME.tar.gz https://$S3_HOST/${S3_GITCACHE_BUCKET}/$CI_PROJECT_NAMESPACE/$CI_PROJECT_NAME/$CI_PROJECT_NAME.tar.gz + - s3_upload ../$CI_PROJECT_NAME.tar.gz https://$S3_HOST/${S3_GITCACHE_BUCKET}/$CI_PROJECT_NAMESPACE/$CI_PROJECT_NAME/ # Sanity checks of MR settings and commit logs diff --git a/drivers/gpu/drm/ci/image-tags.yml b/drivers/gpu/drm/ci/image-tags.yml index c04ba0e69935..53fe34b86578 100644 --- a/drivers/gpu/drm/ci/image-tags.yml +++ b/drivers/gpu/drm/ci/image-tags.yml @@ -1,5 +1,5 @@ variables: - CONTAINER_TAG: "20250307-mesa-uprev" + CONTAINER_TAG: "20250328-mesa-uprev" DEBIAN_X86_64_BUILD_BASE_IMAGE: "debian/x86_64_build-base" DEBIAN_BASE_TAG: "${CONTAINER_TAG}" diff --git a/drivers/gpu/drm/ci/lava-submit.sh b/drivers/gpu/drm/ci/lava-submit.sh index f22720359b33..a1e8b34fb2d4 100755 --- a/drivers/gpu/drm/ci/lava-submit.sh +++ b/drivers/gpu/drm/ci/lava-submit.sh @@ -54,7 +54,7 @@ cp artifacts/ci-common/init-*.sh results/job-rootfs-overlay/ cp "$SCRIPTS_DIR"/setup-test-env.sh results/job-rootfs-overlay/ tar zcf job-rootfs-overlay.tar.gz -C results/job-rootfs-overlay/ . -ci-fairy s3cp --token-file "${S3_JWT_FILE}" job-rootfs-overlay.tar.gz "https://${JOB_ROOTFS_OVERLAY_PATH}" +s3_upload job-rootfs-overlay.tar.gz "https://${JOB_ARTIFACTS_BASE}" # Prepare env vars for upload. section_switch variables "Environment variables passed through to device:" diff --git a/drivers/gpu/drm/display/drm_dp_helper.c b/drivers/gpu/drm/display/drm_dp_helper.c index 57828f2b7b5a..f2a6559a2710 100644 --- a/drivers/gpu/drm/display/drm_dp_helper.c +++ b/drivers/gpu/drm/display/drm_dp_helper.c @@ -2137,14 +2137,17 @@ static int drm_dp_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, for (i = 0; i < num; i++) { msg.address = msgs[i].addr; - drm_dp_i2c_msg_set_request(&msg, &msgs[i]); - /* Send a bare address packet to start the transaction. - * Zero sized messages specify an address only (bare - * address) transaction. - */ - msg.buffer = NULL; - msg.size = 0; - err = drm_dp_i2c_do_msg(aux, &msg); + + if (!aux->no_zero_sized) { + drm_dp_i2c_msg_set_request(&msg, &msgs[i]); + /* Send a bare address packet to start the transaction. + * Zero sized messages specify an address only (bare + * address) transaction. + */ + msg.buffer = NULL; + msg.size = 0; + err = drm_dp_i2c_do_msg(aux, &msg); + } /* * Reset msg.request in case in case it got @@ -2163,6 +2166,8 @@ static int drm_dp_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, msg.buffer = msgs[i].buf + j; msg.size = min(transfer_size, msgs[i].len - j); + if (j + msg.size == msgs[i].len && aux->no_zero_sized) + msg.request &= ~DP_AUX_I2C_MOT; err = drm_dp_i2c_drain_msg(aux, &msg); /* @@ -2180,15 +2185,17 @@ static int drm_dp_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, } if (err >= 0) err = num; - /* Send a bare address packet to close out the transaction. - * Zero sized messages specify an address only (bare - * address) transaction. - */ - msg.request &= ~DP_AUX_I2C_MOT; - msg.buffer = NULL; - msg.size = 0; - (void)drm_dp_i2c_do_msg(aux, &msg); + if (!aux->no_zero_sized) { + /* Send a bare address packet to close out the transaction. + * Zero sized messages specify an address only (bare + * address) transaction. + */ + msg.request &= ~DP_AUX_I2C_MOT; + msg.buffer = NULL; + msg.size = 0; + (void)drm_dp_i2c_do_msg(aux, &msg); + } return err; } @@ -4393,8 +4400,9 @@ EXPORT_SYMBOL(drm_panel_dp_aux_backlight); #endif /* See DP Standard v2.1 2.6.4.4.1.1, 2.8.4.4, 2.8.7 */ -static int drm_dp_link_symbol_cycles(int lane_count, int pixels, int bpp_x16, - int symbol_size, bool is_mst) +static int drm_dp_link_data_symbol_cycles(int lane_count, int pixels, + int bpp_x16, int symbol_size, + bool is_mst) { int cycles = DIV_ROUND_UP(pixels * bpp_x16, 16 * symbol_size * lane_count); int align = is_mst ? 4 / lane_count : 1; @@ -4402,22 +4410,42 @@ static int drm_dp_link_symbol_cycles(int lane_count, int pixels, int bpp_x16, return ALIGN(cycles, align); } -static int drm_dp_link_dsc_symbol_cycles(int lane_count, int pixels, int slice_count, - int bpp_x16, int symbol_size, bool is_mst) +/** + * drm_dp_link_symbol_cycles - calculate the link symbol count with/without dsc + * @lane_count: DP link lane count + * @pixels: number of pixels in a scanline + * @dsc_slice_count: number of slices for DSC or '0' for non-DSC + * @bpp_x16: bits per pixel in .4 binary fixed format + * @symbol_size: DP symbol size + * @is_mst: %true for MST and %false for SST + * + * Calculate the link symbol cycles for both DSC (@dsc_slice_count !=0) and + * non-DSC case (@dsc_slice_count == 0) and return the count. + */ +int drm_dp_link_symbol_cycles(int lane_count, int pixels, int dsc_slice_count, + int bpp_x16, int symbol_size, bool is_mst) { + int slice_count = dsc_slice_count ? : 1; int slice_pixels = DIV_ROUND_UP(pixels, slice_count); - int slice_data_cycles = drm_dp_link_symbol_cycles(lane_count, slice_pixels, - bpp_x16, symbol_size, is_mst); - int slice_eoc_cycles = is_mst ? 4 / lane_count : 1; + int slice_data_cycles = drm_dp_link_data_symbol_cycles(lane_count, + slice_pixels, + bpp_x16, + symbol_size, + is_mst); + int slice_eoc_cycles = 0; + + if (dsc_slice_count) + slice_eoc_cycles = is_mst ? 4 / lane_count : 1; return slice_count * (slice_data_cycles + slice_eoc_cycles); } +EXPORT_SYMBOL(drm_dp_link_symbol_cycles); /** * drm_dp_bw_overhead - Calculate the BW overhead of a DP link stream * @lane_count: DP link lane count * @hactive: pixel count of the active period in one scanline of the stream - * @dsc_slice_count: DSC slice count if @flags/DRM_DP_LINK_BW_OVERHEAD_DSC is set + * @dsc_slice_count: number of slices for DSC or '0' for non-DSC * @bpp_x16: bits per pixel in .4 binary fixed point * @flags: DRM_DP_OVERHEAD_x flags * @@ -4431,7 +4459,7 @@ static int drm_dp_link_dsc_symbol_cycles(int lane_count, int pixels, int slice_c * as well as the stream's * - @hactive timing * - @bpp_x16 color depth - * - compression mode (@flags / %DRM_DP_OVERHEAD_DSC). + * - compression mode (@dsc_slice_count != 0) * Note that this overhead doesn't account for the 8b/10b, 128b/132b * channel coding efficiency, for that see * @drm_dp_link_bw_channel_coding_efficiency(). @@ -4486,15 +4514,10 @@ int drm_dp_bw_overhead(int lane_count, int hactive, WARN_ON((flags & DRM_DP_BW_OVERHEAD_UHBR) && (flags & DRM_DP_BW_OVERHEAD_FEC)); - if (flags & DRM_DP_BW_OVERHEAD_DSC) - symbol_cycles = drm_dp_link_dsc_symbol_cycles(lane_count, hactive, - dsc_slice_count, - bpp_x16, symbol_size, - is_mst); - else - symbol_cycles = drm_dp_link_symbol_cycles(lane_count, hactive, - bpp_x16, symbol_size, - is_mst); + symbol_cycles = drm_dp_link_symbol_cycles(lane_count, hactive, + dsc_slice_count, + bpp_x16, symbol_size, + is_mst); return DIV_ROUND_UP_ULL(mul_u32_u32(symbol_cycles * symbol_size * lane_count, overhead * 16), diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 3dc7acd56b1d..4589ab4bb1ec 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -830,36 +830,62 @@ void *__devm_drm_dev_alloc(struct device *parent, EXPORT_SYMBOL(__devm_drm_dev_alloc); /** - * drm_dev_alloc - Allocate new DRM device - * @driver: DRM driver to allocate device for + * __drm_dev_alloc - Allocation of a &drm_device instance * @parent: Parent device object + * @driver: DRM driver + * @size: the size of the struct which contains struct drm_device + * @offset: the offset of the &drm_device within the container. * - * This is the deprecated version of devm_drm_dev_alloc(), which does not support - * subclassing through embedding the struct &drm_device in a driver private - * structure, and which does not support automatic cleanup through devres. + * This should *NOT* be by any drivers, but is a dedicated interface for the + * corresponding Rust abstraction. * - * RETURNS: - * Pointer to new DRM device, or ERR_PTR on failure. + * This is the same as devm_drm_dev_alloc(), but without the corresponding + * resource management through the parent device, but not the same as + * drm_dev_alloc(), since the latter is the deprecated version, which does not + * support subclassing. + * + * Returns: A pointer to new DRM device, or an ERR_PTR on failure. */ -struct drm_device *drm_dev_alloc(const struct drm_driver *driver, - struct device *parent) +void *__drm_dev_alloc(struct device *parent, + const struct drm_driver *driver, + size_t size, size_t offset) { - struct drm_device *dev; + void *container; + struct drm_device *drm; int ret; - dev = kzalloc(sizeof(*dev), GFP_KERNEL); - if (!dev) + container = kzalloc(size, GFP_KERNEL); + if (!container) return ERR_PTR(-ENOMEM); - ret = drm_dev_init(dev, driver, parent); + drm = container + offset; + ret = drm_dev_init(drm, driver, parent); if (ret) { - kfree(dev); + kfree(container); return ERR_PTR(ret); } + drmm_add_final_kfree(drm, container); - drmm_add_final_kfree(dev, dev); + return container; +} +EXPORT_SYMBOL(__drm_dev_alloc); - return dev; +/** + * drm_dev_alloc - Allocate new DRM device + * @driver: DRM driver to allocate device for + * @parent: Parent device object + * + * This is the deprecated version of devm_drm_dev_alloc(), which does not support + * subclassing through embedding the struct &drm_device in a driver private + * structure, and which does not support automatic cleanup through devres. + * + * RETURNS: + * Pointer to new DRM device, or ERR_PTR on failure. + */ +struct drm_device *drm_dev_alloc(const struct drm_driver *driver, + struct device *parent) +{ + return __drm_dev_alloc(parent, driver, sizeof(struct drm_device), 0); } EXPORT_SYMBOL(drm_dev_alloc); diff --git a/drivers/gpu/drm/drm_file.c b/drivers/gpu/drm/drm_file.c index cf2463090d3a..246cf845e2c9 100644 --- a/drivers/gpu/drm/drm_file.c +++ b/drivers/gpu/drm/drm_file.c @@ -993,6 +993,40 @@ void drm_show_fdinfo(struct seq_file *m, struct file *f) EXPORT_SYMBOL(drm_show_fdinfo); /** + * drm_file_err - log process name, pid and client_name associated with a drm_file + * @file_priv: context of interest for process name and pid + * @fmt: printf() like format string + * + * Helper function for clients which needs to log process details such + * as name and pid etc along with user logs. + */ +void drm_file_err(struct drm_file *file_priv, const char *fmt, ...) +{ + va_list args; + struct va_format vaf; + struct pid *pid; + struct task_struct *task; + struct drm_device *dev = file_priv->minor->dev; + + va_start(args, fmt); + vaf.fmt = fmt; + vaf.va = &args; + + mutex_lock(&file_priv->client_name_lock); + rcu_read_lock(); + pid = rcu_dereference(file_priv->pid); + task = pid_task(pid, PIDTYPE_TGID); + + drm_err(dev, "comm: %s pid: %d client: %s ... %pV", task ? task->comm : "Unset", + task ? task->pid : 0, file_priv->client_name ?: "Unset", &vaf); + + va_end(args); + rcu_read_unlock(); + mutex_unlock(&file_priv->client_name_lock); +} +EXPORT_SYMBOL(drm_file_err); + +/** * mock_drm_getfile - Create a new struct file for the drm device * @minor: drm minor to wrap (e.g. #drm_device.primary) * @flags: file creation mode (O_RDWR etc) diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index c8fc271b33b7..13d4a16f7d33 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -52,7 +52,6 @@ i915-y += \ i915-y += \ soc/intel_dram.o \ soc/intel_gmch.o \ - soc/intel_pch.o \ soc/intel_rom.o # core library code @@ -282,6 +281,7 @@ i915-y += \ display/intel_modeset_setup.o \ display/intel_modeset_verify.o \ display/intel_overlay.o \ + display/intel_pch.o \ display/intel_pch_display.o \ display/intel_pch_refclk.o \ display/intel_plane_initial.o \ diff --git a/drivers/gpu/drm/i915/display/g4x_dp.c b/drivers/gpu/drm/i915/display/g4x_dp.c index b39aae9165df..e0a98e6fd6d1 100644 --- a/drivers/gpu/drm/i915/display/g4x_dp.c +++ b/drivers/gpu/drm/i915/display/g4x_dp.c @@ -7,9 +7,11 @@ #include <linux/string_helpers.h> +#include <drm/drm_print.h> + #include "g4x_dp.h" -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_audio.h" #include "intel_backlight.h" #include "intel_connector.h" @@ -28,7 +30,6 @@ #include "intel_hotplug.h" #include "intel_pch_display.h" #include "intel_pps.h" -#include "vlv_sideband.h" static const struct dpll g4x_dpll[] = { { .dot = 162000, .p1 = 2, .p2 = 10, .n = 2, .m1 = 23, .m2 = 8, }, @@ -60,14 +61,13 @@ static void g4x_dp_set_clock(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); const struct dpll *divisor = NULL; int i, count = 0; if (display->platform.g4x) { divisor = g4x_dpll; count = ARRAY_SIZE(g4x_dpll); - } else if (HAS_PCH_SPLIT(dev_priv)) { + } else if (HAS_PCH_SPLIT(display)) { divisor = pch_dpll; count = ARRAY_SIZE(pch_dpll); } else if (display->platform.cherryview) { @@ -93,7 +93,6 @@ static void intel_dp_prepare(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); enum port port = encoder->port; struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc); @@ -141,7 +140,7 @@ static void intel_dp_prepare(struct intel_encoder *encoder, intel_dp->DP |= DP_ENHANCED_FRAMING; intel_dp->DP |= DP_PIPE_SEL_IVB(crtc->pipe); - } else if (HAS_PCH_CPT(dev_priv) && port != PORT_A) { + } else if (HAS_PCH_CPT(display) && port != PORT_A) { intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT; intel_de_rmw(display, TRANS_DP_CTL(crtc->pipe), @@ -183,7 +182,7 @@ static void assert_dp_port(struct intel_dp *intel_dp, bool state) static void assert_edp_pll(struct intel_display *display, bool state) { - bool cur_state = intel_de_read(display, DP_A) & DP_PLL_ENABLE; + bool cur_state = intel_de_read(display, DP_A) & EDP_PLL_ENABLE; INTEL_DISPLAY_STATE_WARN(display, cur_state != state, "eDP PLL state assertion failure (expected %s, current %s)\n", @@ -205,12 +204,12 @@ static void ilk_edp_pll_on(struct intel_dp *intel_dp, drm_dbg_kms(display->drm, "enabling eDP PLL for clock %d\n", pipe_config->port_clock); - intel_dp->DP &= ~DP_PLL_FREQ_MASK; + intel_dp->DP &= ~EDP_PLL_FREQ_MASK; if (pipe_config->port_clock == 162000) - intel_dp->DP |= DP_PLL_FREQ_162MHZ; + intel_dp->DP |= EDP_PLL_FREQ_162MHZ; else - intel_dp->DP |= DP_PLL_FREQ_270MHZ; + intel_dp->DP |= EDP_PLL_FREQ_270MHZ; intel_de_write(display, DP_A, intel_dp->DP); intel_de_posting_read(display, DP_A); @@ -225,7 +224,7 @@ static void ilk_edp_pll_on(struct intel_dp *intel_dp, if (display->platform.ironlake) intel_wait_for_vblank_if_active(display, !crtc->pipe); - intel_dp->DP |= DP_PLL_ENABLE; + intel_dp->DP |= EDP_PLL_ENABLE; intel_de_write(display, DP_A, intel_dp->DP); intel_de_posting_read(display, DP_A); @@ -243,7 +242,7 @@ static void ilk_edp_pll_off(struct intel_dp *intel_dp, drm_dbg_kms(display->drm, "disabling eDP PLL\n"); - intel_dp->DP &= ~DP_PLL_ENABLE; + intel_dp->DP &= ~EDP_PLL_ENABLE; intel_de_write(display, DP_A, intel_dp->DP); intel_de_posting_read(display, DP_A); @@ -277,7 +276,6 @@ bool g4x_dp_port_enabled(struct intel_display *display, i915_reg_t dp_reg, enum port port, enum pipe *pipe) { - struct drm_i915_private *dev_priv = to_i915(display->drm); bool ret; u32 val; @@ -287,13 +285,13 @@ bool g4x_dp_port_enabled(struct intel_display *display, /* asserts want to know the pipe even if the port is disabled */ if (display->platform.ivybridge && port == PORT_A) - *pipe = (val & DP_PIPE_SEL_MASK_IVB) >> DP_PIPE_SEL_SHIFT_IVB; - else if (HAS_PCH_CPT(dev_priv) && port != PORT_A) + *pipe = REG_FIELD_GET(DP_PIPE_SEL_MASK_IVB, val); + else if (HAS_PCH_CPT(display) && port != PORT_A) ret &= cpt_dp_port_selected(display, port, pipe); else if (display->platform.cherryview) - *pipe = (val & DP_PIPE_SEL_MASK_CHV) >> DP_PIPE_SEL_SHIFT_CHV; + *pipe = REG_FIELD_GET(DP_PIPE_SEL_MASK_CHV, val); else - *pipe = (val & DP_PIPE_SEL_MASK) >> DP_PIPE_SEL_SHIFT; + *pipe = REG_FIELD_GET(DP_PIPE_SEL_MASK, val); return ret; } @@ -338,7 +336,6 @@ static void intel_dp_get_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); u32 tmp, flags = 0; enum port port = encoder->port; @@ -353,7 +350,7 @@ static void intel_dp_get_config(struct intel_encoder *encoder, pipe_config->has_audio = tmp & DP_AUDIO_OUTPUT_ENABLE && port != PORT_A; - if (HAS_PCH_CPT(dev_priv) && port != PORT_A) { + if (HAS_PCH_CPT(display) && port != PORT_A) { u32 trans_dp = intel_de_read(display, TRANS_DP_CTL(crtc->pipe)); @@ -389,13 +386,12 @@ static void intel_dp_get_config(struct intel_encoder *encoder, if (display->platform.g4x && tmp & DP_COLOR_RANGE_16_235) pipe_config->limited_color_range = true; - pipe_config->lane_count = - ((tmp & DP_PORT_WIDTH_MASK) >> DP_PORT_WIDTH_SHIFT) + 1; + pipe_config->lane_count = REG_FIELD_GET(DP_PORT_WIDTH_MASK, tmp) + 1; g4x_dp_get_m_n(pipe_config); if (port == PORT_A) { - if ((intel_de_read(display, DP_A) & DP_PLL_FREQ_MASK) == DP_PLL_FREQ_162MHZ) + if ((intel_de_read(display, DP_A) & EDP_PLL_FREQ_MASK) == EDP_PLL_FREQ_162MHZ) pipe_config->port_clock = 162000; else pipe_config->port_clock = 270000; @@ -416,7 +412,6 @@ intel_dp_link_down(struct intel_encoder *encoder, const struct intel_crtc_state *old_crtc_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc); enum port port = encoder->port; @@ -429,7 +424,7 @@ intel_dp_link_down(struct intel_encoder *encoder, drm_dbg_kms(display->drm, "\n"); if ((display->platform.ivybridge && port == PORT_A) || - (HAS_PCH_CPT(dev_priv) && port != PORT_A)) { + (HAS_PCH_CPT(display) && port != PORT_A)) { intel_dp->DP &= ~DP_LINK_TRAIN_MASK_CPT; intel_dp->DP |= DP_LINK_TRAIN_PAT_IDLE_CPT; } else { @@ -448,7 +443,7 @@ intel_dp_link_down(struct intel_encoder *encoder, * to transcoder A after disabling it to allow the * matching HDMI port to be enabled on transcoder A. */ - if (HAS_PCH_IBX(dev_priv) && crtc->pipe == PIPE_B && port != PORT_A) { + if (HAS_PCH_IBX(display) && crtc->pipe == PIPE_B && port != PORT_A) { /* * We get CPU/PCH FIFO underruns on the other pipe when * doing the workaround. Sweep them under the rug. @@ -581,16 +576,10 @@ static void chv_post_disable_dp(struct intel_atomic_state *state, const struct intel_crtc_state *old_crtc_state, const struct drm_connector_state *old_conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - intel_dp_link_down(encoder, old_crtc_state); - vlv_dpio_get(dev_priv); - /* Assert data lane reset */ chv_data_lane_soft_reset(encoder, old_crtc_state, true); - - vlv_dpio_put(dev_priv); } static void @@ -1223,10 +1212,10 @@ static int g4x_dp_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state, struct drm_connector_state *conn_state) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); int ret; - if (HAS_PCH_SPLIT(i915) && encoder->port != PORT_A) + if (HAS_PCH_SPLIT(display) && encoder->port != PORT_A) crtc_state->has_pch_encoder = true; ret = intel_dp_compute_config(encoder, crtc_state, conn_state); @@ -1279,7 +1268,6 @@ static const struct drm_encoder_funcs intel_dp_enc_funcs = { bool g4x_dp_init(struct intel_display *display, i915_reg_t output_reg, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); const struct intel_bios_encoder_data *devdata; struct intel_digital_port *dig_port; struct intel_encoder *intel_encoder; @@ -1353,7 +1341,7 @@ bool g4x_dp_init(struct intel_display *display, intel_encoder->audio_disable = g4x_dp_audio_disable; if ((display->platform.ivybridge && port == PORT_A) || - (HAS_PCH_CPT(dev_priv) && port != PORT_A)) + (HAS_PCH_CPT(display) && port != PORT_A)) dig_port->dp.set_link_train = cpt_set_link_train; else dig_port->dp.set_link_train = g4x_set_link_train; @@ -1370,7 +1358,7 @@ bool g4x_dp_init(struct intel_display *display, intel_encoder->set_signal_levels = g4x_set_signal_levels; if (display->platform.valleyview || display->platform.cherryview || - (HAS_PCH_SPLIT(dev_priv) && port != PORT_A)) { + (HAS_PCH_SPLIT(display) && port != PORT_A)) { dig_port->dp.preemph_max = intel_dp_preemph_max_3; dig_port->dp.voltage_max = intel_dp_voltage_max_3; } else { diff --git a/drivers/gpu/drm/i915/display/g4x_hdmi.c b/drivers/gpu/drm/i915/display/g4x_hdmi.c index 3dc2c59a3df0..1d252432d729 100644 --- a/drivers/gpu/drm/i915/display/g4x_hdmi.c +++ b/drivers/gpu/drm/i915/display/g4x_hdmi.c @@ -5,8 +5,9 @@ * HDMI support for G4x,ILK,SNB,IVB,VLV,CHV (HSW+ handled by the DDI code). */ +#include <drm/drm_print.h> + #include "g4x_hdmi.h" -#include "i915_drv.h" #include "i915_reg.h" #include "intel_atomic.h" #include "intel_audio.h" @@ -22,13 +23,11 @@ #include "intel_hdmi.h" #include "intel_hotplug.h" #include "intel_sdvo.h" -#include "vlv_sideband.h" static void intel_hdmi_prepare(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; @@ -37,7 +36,7 @@ static void intel_hdmi_prepare(struct intel_encoder *encoder, intel_dp_dual_mode_set_tmds_output(intel_hdmi, true); hdmi_val = SDVO_ENCODING_HDMI; - if (!HAS_PCH_SPLIT(dev_priv) && crtc_state->limited_color_range) + if (!HAS_PCH_SPLIT(display) && crtc_state->limited_color_range) hdmi_val |= HDMI_COLOR_RANGE_16_235; if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC) hdmi_val |= SDVO_VSYNC_ACTIVE_HIGH; @@ -52,7 +51,7 @@ static void intel_hdmi_prepare(struct intel_encoder *encoder, if (crtc_state->has_hdmi_sink) hdmi_val |= HDMI_MODE_SELECT_HDMI; - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) hdmi_val |= SDVO_PIPE_SEL_CPT(crtc->pipe); else if (display->platform.cherryview) hdmi_val |= SDVO_PIPE_SEL_CHV(crtc->pipe); @@ -134,9 +133,8 @@ static int g4x_hdmi_compute_config(struct intel_encoder *encoder, struct intel_display *display = to_intel_display(encoder); struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); - if (HAS_PCH_SPLIT(i915)) { + if (HAS_PCH_SPLIT(display)) { crtc_state->has_pch_encoder = true; if (!intel_fdi_compute_pipe_bpp(crtc_state)) return -EINVAL; @@ -155,7 +153,6 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder); u32 tmp, flags = 0; int dotclock; @@ -186,7 +183,7 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder, if (tmp & HDMI_AUDIO_ENABLE) pipe_config->has_audio = true; - if (!HAS_PCH_SPLIT(dev_priv) && + if (!HAS_PCH_SPLIT(display) && tmp & HDMI_COLOR_RANGE_16_235) pipe_config->limited_color_range = true; @@ -383,7 +380,6 @@ static void intel_disable_hdmi(struct intel_atomic_state *state, const struct drm_connector_state *old_conn_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder); struct intel_digital_port *dig_port = hdmi_to_dig_port(intel_hdmi); @@ -401,7 +397,7 @@ static void intel_disable_hdmi(struct intel_atomic_state *state, * to transcoder A after disabling it to allow the * matching DP port to be enabled on transcoder A. */ - if (HAS_PCH_IBX(dev_priv) && crtc->pipe == PIPE_B) { + if (HAS_PCH_IBX(display) && crtc->pipe == PIPE_B) { /* * We get CPU/PCH FIFO underruns on the other pipe when * doing the workaround. Sweep them under the rug. @@ -539,15 +535,8 @@ static void chv_hdmi_post_disable(struct intel_atomic_state *state, const struct intel_crtc_state *old_crtc_state, const struct drm_connector_state *old_conn_state) { - struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(display->drm); - - vlv_dpio_get(dev_priv); - /* Assert data lane reset */ chv_data_lane_soft_reset(encoder, old_crtc_state, true); - - vlv_dpio_put(dev_priv); } static void chv_hdmi_pre_enable(struct intel_atomic_state *state, @@ -682,7 +671,6 @@ static bool assert_hdmi_port_valid(struct intel_display *display, enum port port bool g4x_hdmi_init(struct intel_display *display, i915_reg_t hdmi_reg, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); const struct intel_bios_encoder_data *devdata; struct intel_digital_port *dig_port; struct intel_encoder *intel_encoder; @@ -724,7 +712,7 @@ bool g4x_hdmi_init(struct intel_display *display, intel_encoder->hotplug = intel_hdmi_hotplug; intel_encoder->compute_config = g4x_hdmi_compute_config; - if (HAS_PCH_SPLIT(dev_priv)) { + if (HAS_PCH_SPLIT(display)) { intel_encoder->disable = pch_disable_hdmi; intel_encoder->post_disable = pch_post_disable_hdmi; } else { @@ -745,9 +733,9 @@ bool g4x_hdmi_init(struct intel_display *display, intel_encoder->post_disable = vlv_hdmi_post_disable; } else { intel_encoder->pre_enable = intel_hdmi_pre_enable; - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) intel_encoder->enable = cpt_enable_hdmi; - else if (HAS_PCH_IBX(dev_priv)) + else if (HAS_PCH_IBX(display)) intel_encoder->enable = ibx_enable_hdmi; else intel_encoder->enable = g4x_enable_hdmi; diff --git a/drivers/gpu/drm/i915/display/i9xx_plane.c b/drivers/gpu/drm/i915/display/i9xx_plane.c index 75780fefe0ce..a2a6d52be0a5 100644 --- a/drivers/gpu/drm/i915/display/i9xx_plane.c +++ b/drivers/gpu/drm/i915/display/i9xx_plane.c @@ -7,9 +7,10 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_blend.h> #include <drm/drm_fourcc.h> +#include <drm/drm_print.h> -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "i9xx_plane.h" #include "i9xx_plane_regs.h" #include "intel_atomic.h" @@ -631,92 +632,84 @@ static void bdw_primary_enable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); bdw_enable_pipe_irq(display, pipe, GEN8_PIPE_PRIMARY_FLIP_DONE); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void bdw_primary_disable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); bdw_disable_pipe_irq(display, pipe, GEN8_PIPE_PRIMARY_FLIP_DONE); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void ivb_primary_enable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); ilk_enable_display_irq(display, DE_PLANE_FLIP_DONE_IVB(plane->i9xx_plane)); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void ivb_primary_disable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); ilk_disable_display_irq(display, DE_PLANE_FLIP_DONE_IVB(plane->i9xx_plane)); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void ilk_primary_enable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); ilk_enable_display_irq(display, DE_PLANE_FLIP_DONE(plane->i9xx_plane)); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void ilk_primary_disable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); ilk_disable_display_irq(display, DE_PLANE_FLIP_DONE(plane->i9xx_plane)); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void vlv_primary_enable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); i915_enable_pipestat(display, pipe, PLANE_FLIP_DONE_INT_STATUS_VLV); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void vlv_primary_disable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); i915_disable_pipestat(display, pipe, PLANE_FLIP_DONE_INT_STATUS_VLV); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static bool i9xx_plane_can_async_flip(u64 modifier) diff --git a/drivers/gpu/drm/i915/display/i9xx_wm.c b/drivers/gpu/drm/i915/display/i9xx_wm.c index 40751f1547b7..77876ef735b7 100644 --- a/drivers/gpu/drm/i915/display/i9xx_wm.c +++ b/drivers/gpu/drm/i915/display/i9xx_wm.c @@ -4143,10 +4143,8 @@ static const struct intel_wm_funcs nop_funcs = { void i9xx_wm_init(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - /* For FIFO watermark updates */ - if (HAS_PCH_SPLIT(dev_priv)) { + if (HAS_PCH_SPLIT(display)) { ilk_setup_wm_latency(display); display->funcs.wm = &ilk_wm_funcs; } else if (display->platform.valleyview || display->platform.cherryview) { diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c b/drivers/gpu/drm/i915/display/intel_alpm.c index 55f3ae1e68c9..c176bdbc19a3 100644 --- a/drivers/gpu/drm/i915/display/intel_alpm.c +++ b/drivers/gpu/drm/i915/display/intel_alpm.c @@ -5,12 +5,15 @@ #include <linux/debugfs.h> +#include <drm/drm_print.h> + #include "intel_alpm.h" #include "intel_crtc.h" #include "intel_de.h" #include "intel_display_types.h" #include "intel_dp.h" #include "intel_dp_aux.h" +#include "intel_psr.h" #include "intel_psr_regs.h" bool intel_alpm_aux_wake_supported(struct intel_dp *intel_dp) @@ -23,7 +26,7 @@ bool intel_alpm_aux_less_wake_supported(struct intel_dp *intel_dp) return intel_dp->alpm_dpcd & DP_ALPM_AUX_LESS_CAP; } -void intel_alpm_init_dpcd(struct intel_dp *intel_dp) +void intel_alpm_init(struct intel_dp *intel_dp) { u8 dpcd; @@ -31,6 +34,7 @@ void intel_alpm_init_dpcd(struct intel_dp *intel_dp) return; intel_dp->alpm_dpcd = dpcd; + mutex_init(&intel_dp->alpm_parameters.lock); } /* @@ -276,6 +280,14 @@ void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp, int waketime_in_lines, first_sdp_position; int context_latency, guardband; + if (intel_dp->alpm_parameters.lobf_disable_debug) { + drm_dbg_kms(display->drm, "LOBF is disabled by debug flag\n"); + return; + } + + if (intel_dp->alpm_parameters.sink_alpm_error) + return; + if (!intel_dp_is_edp(intel_dp)) return; @@ -288,6 +300,10 @@ void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp, if (crtc_state->has_psr) return; + if (crtc_state->vrr.vmin != crtc_state->vrr.vmax || + crtc_state->vrr.vmin != crtc_state->vrr.flipline) + return; + if (!(intel_alpm_aux_wake_supported(intel_dp) || intel_alpm_aux_less_wake_supported(intel_dp))) return; @@ -316,15 +332,16 @@ static void lnl_alpm_configure(struct intel_dp *intel_dp, enum port port = dp_to_dig_port(intel_dp)->base.port; u32 alpm_ctl; - if (DISPLAY_VER(display) < 20 || - (!intel_dp->psr.sel_update_enabled && !intel_dp_is_edp(intel_dp))) + if (DISPLAY_VER(display) < 20 || (!intel_psr_needs_alpm(intel_dp, crtc_state) && + !crtc_state->has_lobf)) return; + mutex_lock(&intel_dp->alpm_parameters.lock); /* * Panel Replay on eDP is always using ALPM aux less. I.e. no need to * check panel support at this point. */ - if ((intel_dp->psr.panel_replay_enabled && intel_dp_is_edp(intel_dp)) || + if ((crtc_state->has_panel_replay && intel_dp_is_edp(intel_dp)) || (crtc_state->has_lobf && intel_alpm_aux_less_wake_supported(intel_dp))) { alpm_ctl = ALPM_CTL_ALPM_ENABLE | ALPM_CTL_ALPM_AUX_LESS_ENABLE | @@ -353,18 +370,107 @@ static void lnl_alpm_configure(struct intel_dp *intel_dp, ALPM_CTL_EXTENDED_FAST_WAKE_TIME(intel_dp->alpm_parameters.fast_wake_lines); } - if (crtc_state->has_lobf) + if (crtc_state->has_lobf) { alpm_ctl |= ALPM_CTL_LOBF_ENABLE; + drm_dbg_kms(display->drm, "Link off between frames (LOBF) enabled\n"); + } alpm_ctl |= ALPM_CTL_ALPM_ENTRY_CHECK(intel_dp->alpm_parameters.check_entry_lines); intel_de_write(display, ALPM_CTL(display, cpu_transcoder), alpm_ctl); + mutex_unlock(&intel_dp->alpm_parameters.lock); } void intel_alpm_configure(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { lnl_alpm_configure(intel_dp, crtc_state); + intel_dp->alpm_parameters.transcoder = crtc_state->cpu_transcoder; +} + +void intel_alpm_pre_plane_update(struct intel_atomic_state *state, + struct intel_crtc *crtc) +{ + struct intel_display *display = to_intel_display(state); + const struct intel_crtc_state *crtc_state = + intel_atomic_get_new_crtc_state(state, crtc); + const struct intel_crtc_state *old_crtc_state = + intel_atomic_get_old_crtc_state(state, crtc); + enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; + struct intel_encoder *encoder; + + if (DISPLAY_VER(display) < 20) + return; + + if (crtc_state->has_lobf || crtc_state->has_lobf == old_crtc_state->has_lobf) + return; + + for_each_intel_encoder_mask(display->drm, encoder, + crtc_state->uapi.encoder_mask) { + struct intel_dp *intel_dp; + + if (!intel_encoder_is_dp(encoder)) + continue; + + intel_dp = enc_to_intel_dp(encoder); + + if (!intel_dp_is_edp(intel_dp)) + continue; + + if (old_crtc_state->has_lobf) { + mutex_lock(&intel_dp->alpm_parameters.lock); + intel_de_write(display, ALPM_CTL(display, cpu_transcoder), 0); + drm_dbg_kms(display->drm, "Link off between frames (LOBF) disabled\n"); + mutex_unlock(&intel_dp->alpm_parameters.lock); + } + } +} + +void intel_alpm_enable_sink(struct intel_dp *intel_dp, + const struct intel_crtc_state *crtc_state) +{ + u8 val; + + if (!intel_psr_needs_alpm(intel_dp, crtc_state) && !crtc_state->has_lobf) + return; + + val = DP_ALPM_ENABLE | DP_ALPM_LOCK_ERROR_IRQ_HPD_ENABLE; + + if (crtc_state->has_panel_replay || (crtc_state->has_lobf && + intel_alpm_aux_less_wake_supported(intel_dp))) + val |= DP_ALPM_MODE_AUX_LESS; + + drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG, val); +} + +void intel_alpm_post_plane_update(struct intel_atomic_state *state, + struct intel_crtc *crtc) +{ + struct intel_display *display = to_intel_display(state); + const struct intel_crtc_state *crtc_state = + intel_atomic_get_new_crtc_state(state, crtc); + const struct intel_crtc_state *old_crtc_state = + intel_atomic_get_old_crtc_state(state, crtc); + struct intel_encoder *encoder; + + if (crtc_state->has_psr || !crtc_state->has_lobf || + crtc_state->has_lobf == old_crtc_state->has_lobf) + return; + + for_each_intel_encoder_mask(display->drm, encoder, + crtc_state->uapi.encoder_mask) { + struct intel_dp *intel_dp; + + if (!intel_encoder_is_dp(encoder)) + continue; + + intel_dp = enc_to_intel_dp(encoder); + + if (intel_dp_is_edp(intel_dp)) { + intel_alpm_enable_sink(intel_dp, crtc_state); + intel_alpm_configure(intel_dp, crtc_state); + } + } } static int i915_edp_lobf_info_show(struct seq_file *m, void *data) @@ -403,6 +509,32 @@ out: DEFINE_SHOW_ATTRIBUTE(i915_edp_lobf_info); +static int +i915_edp_lobf_debug_get(void *data, u64 *val) +{ + struct intel_connector *connector = data; + struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder); + + *val = intel_dp->alpm_parameters.lobf_disable_debug; + + return 0; +} + +static int +i915_edp_lobf_debug_set(void *data, u64 val) +{ + struct intel_connector *connector = data; + struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder); + + intel_dp->alpm_parameters.lobf_disable_debug = val; + + return 0; +} + +DEFINE_SIMPLE_ATTRIBUTE(i915_edp_lobf_debug_fops, + i915_edp_lobf_debug_get, i915_edp_lobf_debug_set, + "%llu\n"); + void intel_alpm_lobf_debugfs_add(struct intel_connector *connector) { struct intel_display *display = to_intel_display(connector); @@ -412,6 +544,55 @@ void intel_alpm_lobf_debugfs_add(struct intel_connector *connector) connector->base.connector_type != DRM_MODE_CONNECTOR_eDP) return; + debugfs_create_file("i915_edp_lobf_debug", 0644, root, + connector, &i915_edp_lobf_debug_fops); + debugfs_create_file("i915_edp_lobf_info", 0444, root, connector, &i915_edp_lobf_info_fops); } + +void intel_alpm_disable(struct intel_dp *intel_dp) +{ + struct intel_display *display = to_intel_display(intel_dp); + enum transcoder cpu_transcoder = intel_dp->alpm_parameters.transcoder; + + if (DISPLAY_VER(display) < 20 || !intel_dp->alpm_dpcd) + return; + + mutex_lock(&intel_dp->alpm_parameters.lock); + + intel_de_rmw(display, ALPM_CTL(display, cpu_transcoder), + ALPM_CTL_ALPM_ENABLE | ALPM_CTL_LOBF_ENABLE | + ALPM_CTL_ALPM_AUX_LESS_ENABLE, 0); + + intel_de_rmw(display, + PORT_ALPM_CTL(cpu_transcoder), + PORT_ALPM_CTL_ALPM_AUX_LESS_ENABLE, 0); + + drm_dbg_kms(display->drm, "Disabling ALPM\n"); + mutex_unlock(&intel_dp->alpm_parameters.lock); +} + +bool intel_alpm_get_error(struct intel_dp *intel_dp) +{ + struct intel_display *display = to_intel_display(intel_dp); + struct drm_dp_aux *aux = &intel_dp->aux; + u8 val; + int r; + + r = drm_dp_dpcd_readb(aux, DP_RECEIVER_ALPM_STATUS, &val); + if (r != 1) { + drm_err(display->drm, "Error reading ALPM status\n"); + return true; + } + + if (val & DP_ALPM_LOCK_TIMEOUT_ERROR) { + drm_dbg_kms(display->drm, "ALPM lock timeout error\n"); + + /* Clearing error */ + drm_dp_dpcd_writeb(aux, DP_RECEIVER_ALPM_STATUS, val); + return true; + } + + return false; +} diff --git a/drivers/gpu/drm/i915/display/intel_alpm.h b/drivers/gpu/drm/i915/display/intel_alpm.h index 8c409b10dce6..c9fe21e3e72c 100644 --- a/drivers/gpu/drm/i915/display/intel_alpm.h +++ b/drivers/gpu/drm/i915/display/intel_alpm.h @@ -12,8 +12,10 @@ struct intel_dp; struct intel_crtc_state; struct drm_connector_state; struct intel_connector; +struct intel_atomic_state; +struct intel_crtc; -void intel_alpm_init_dpcd(struct intel_dp *intel_dp); +void intel_alpm_init(struct intel_dp *intel_dp); bool intel_alpm_compute_params(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state); void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp, @@ -21,7 +23,15 @@ void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp, struct drm_connector_state *conn_state); void intel_alpm_configure(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state); +void intel_alpm_enable_sink(struct intel_dp *intel_dp, + const struct intel_crtc_state *crtc_state); +void intel_alpm_pre_plane_update(struct intel_atomic_state *state, + struct intel_crtc *crtc); +void intel_alpm_post_plane_update(struct intel_atomic_state *state, + struct intel_crtc *crtc); void intel_alpm_lobf_debugfs_add(struct intel_connector *connector); bool intel_alpm_aux_wake_supported(struct intel_dp *intel_dp); bool intel_alpm_aux_less_wake_supported(struct intel_dp *intel_dp); +void intel_alpm_disable(struct intel_dp *intel_dp); +bool intel_alpm_get_error(struct intel_dp *intel_dp); #endif diff --git a/drivers/gpu/drm/i915/display/intel_audio.c b/drivers/gpu/drm/i915/display/intel_audio.c index ea935a5d94c8..40d8bbd8107d 100644 --- a/drivers/gpu/drm/i915/display/intel_audio.c +++ b/drivers/gpu/drm/i915/display/intel_audio.c @@ -27,9 +27,9 @@ #include <drm/drm_edid.h> #include <drm/drm_eld.h> #include <drm/drm_fixed.h> +#include <drm/drm_print.h> #include <drm/intel/i915_component.h> -#include "i915_drv.h" #include "intel_atomic.h" #include "intel_audio.h" #include "intel_audio_regs.h" @@ -587,19 +587,17 @@ static void ibx_audio_regs_init(struct intel_display *display, enum pipe pipe, struct ibx_audio_regs *regs) { - struct drm_i915_private *i915 = to_i915(display->drm); - if (display->platform.valleyview || display->platform.cherryview) { regs->hdmiw_hdmiedid = VLV_HDMIW_HDMIEDID(pipe); regs->aud_config = VLV_AUD_CFG(pipe); regs->aud_cntl_st = VLV_AUD_CNTL_ST(pipe); regs->aud_cntrl_st2 = VLV_AUD_CNTL_ST2; - } else if (HAS_PCH_CPT(i915)) { + } else if (HAS_PCH_CPT(display)) { regs->hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID(pipe); regs->aud_config = CPT_AUD_CFG(pipe); regs->aud_cntl_st = CPT_AUD_CNTL_ST(pipe); regs->aud_cntrl_st2 = CPT_AUD_CNTRL_ST2; - } else if (HAS_PCH_IBX(i915)) { + } else if (HAS_PCH_IBX(display)) { regs->hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID(pipe); regs->aud_config = IBX_AUD_CFG(pipe); regs->aud_cntl_st = IBX_AUD_CNTL_ST(pipe); @@ -889,12 +887,10 @@ static const struct intel_audio_funcs hsw_audio_funcs = { */ void intel_audio_hooks_init(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - if (display->platform.g4x) display->funcs.audio = &g4x_audio_funcs; else if (display->platform.valleyview || display->platform.cherryview || - HAS_PCH_CPT(i915) || HAS_PCH_IBX(i915)) + HAS_PCH_CPT(display) || HAS_PCH_IBX(display)) display->funcs.audio = &ibx_audio_funcs; else if (display->platform.haswell || DISPLAY_VER(display) >= 8) display->funcs.audio = &hsw_audio_funcs; diff --git a/drivers/gpu/drm/i915/display/intel_backlight.c b/drivers/gpu/drm/i915/display/intel_backlight.c index 4f3fa966c537..5827da586003 100644 --- a/drivers/gpu/drm/i915/display/intel_backlight.c +++ b/drivers/gpu/drm/i915/display/intel_backlight.c @@ -10,8 +10,11 @@ #include <acpi/video.h> -#include "i915_drv.h" +#include <drm/drm_file.h> +#include <drm/drm_print.h> + #include "i915_reg.h" +#include "i915_utils.h" #include "intel_backlight.h" #include "intel_backlight_regs.h" #include "intel_connector.h" @@ -473,7 +476,6 @@ static void lpt_enable_backlight(const struct intel_crtc_state *crtc_state, { struct intel_connector *connector = to_intel_connector(conn_state->connector); struct intel_display *display = to_intel_display(connector); - struct drm_i915_private *i915 = to_i915(connector->base.dev); struct intel_panel *panel = &connector->panel; u32 pch_ctl1, pch_ctl2; @@ -486,7 +488,7 @@ static void lpt_enable_backlight(const struct intel_crtc_state *crtc_state, intel_de_write(display, BLC_PWM_PCH_CTL1, pch_ctl1); } - if (HAS_PCH_LPT(i915)) + if (HAS_PCH_LPT(display)) intel_de_rmw(display, SOUTH_CHICKEN2, LPT_PWM_GRANULARITY, panel->backlight.alternate_pwm_increment ? LPT_PWM_GRANULARITY : 0); @@ -503,7 +505,7 @@ static void lpt_enable_backlight(const struct intel_crtc_state *crtc_state, pch_ctl1 |= BLM_PCH_POLARITY; /* After LPT, override is the default. */ - if (HAS_PCH_LPT(i915)) + if (HAS_PCH_LPT(display)) pch_ctl1 |= BLM_PCH_OVERRIDE_ENABLE; intel_de_write(display, BLC_PWM_PCH_CTL1, pch_ctl1); @@ -1064,7 +1066,7 @@ static u32 spt_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz) */ static u32 lpt_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz) { - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_panel *panel = &connector->panel; u32 mul, clock; @@ -1073,7 +1075,7 @@ static u32 lpt_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz) else mul = 128; - if (HAS_PCH_LPT_H(i915)) + if (HAS_PCH_LPT_H(display)) clock = MHz(135); /* LPT:H */ else clock = MHz(24); /* LPT:LP */ @@ -1230,12 +1232,11 @@ static u32 get_backlight_min_vbt(struct intel_connector *connector) static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unused) { struct intel_display *display = to_intel_display(connector); - struct drm_i915_private *i915 = to_i915(connector->base.dev); struct intel_panel *panel = &connector->panel; u32 cpu_ctl2, pch_ctl1, pch_ctl2, val; bool alt, cpu_mode; - if (HAS_PCH_LPT(i915)) + if (HAS_PCH_LPT(display)) alt = intel_de_read(display, SOUTH_CHICKEN2) & LPT_PWM_GRANULARITY; else alt = intel_de_read(display, SOUTH_CHICKEN1) & SPT_PWM_GRANULARITY; @@ -1259,7 +1260,7 @@ static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unus panel->backlight.pwm_enabled = pch_ctl1 & BLM_PCH_PWM_ENABLE; - cpu_mode = panel->backlight.pwm_enabled && HAS_PCH_LPT(i915) && + cpu_mode = panel->backlight.pwm_enabled && HAS_PCH_LPT(display) && !(pch_ctl1 & BLM_PCH_OVERRIDE_ENABLE) && (cpu_ctl2 & BLM_PWM_ENABLE); @@ -1466,15 +1467,13 @@ bxt_setup_backlight(struct intel_connector *connector, enum pipe unused) static int cnp_num_backlight_controllers(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - if (INTEL_PCH_TYPE(i915) >= PCH_MTL) + if (INTEL_PCH_TYPE(display) >= PCH_MTL) return 2; - if (INTEL_PCH_TYPE(i915) >= PCH_DG1) + if (INTEL_PCH_TYPE(display) >= PCH_DG1) return 1; - if (INTEL_PCH_TYPE(i915) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) return 2; return 1; @@ -1482,14 +1481,12 @@ static int cnp_num_backlight_controllers(struct intel_display *display) static bool cnp_backlight_controller_is_valid(struct intel_display *display, int controller) { - struct drm_i915_private *i915 = to_i915(display->drm); - if (controller < 0 || controller >= cnp_num_backlight_controllers(display)) return false; if (controller == 1 && - INTEL_PCH_TYPE(i915) >= PCH_ICP && - INTEL_PCH_TYPE(i915) <= PCH_ADP) + INTEL_PCH_TYPE(display) >= PCH_ICP && + INTEL_PCH_TYPE(display) <= PCH_ADP) return intel_de_read(display, SOUTH_CHICKEN1) & ICP_SECOND_PPS_IO_SELECT; return true; @@ -1818,7 +1815,6 @@ void intel_backlight_init_funcs(struct intel_panel *panel) struct intel_connector *connector = container_of(panel, struct intel_connector, panel); struct intel_display *display = to_intel_display(connector); - struct drm_i915_private *i915 = to_i915(connector->base.dev); if (connector->base.connector_type == DRM_MODE_CONNECTOR_DSI && intel_dsi_dcs_init_backlight_funcs(connector) == 0) @@ -1826,14 +1822,14 @@ void intel_backlight_init_funcs(struct intel_panel *panel) if (display->platform.geminilake || display->platform.broxton) { panel->backlight.pwm_funcs = &bxt_pwm_funcs; - } else if (INTEL_PCH_TYPE(i915) >= PCH_CNP) { + } else if (INTEL_PCH_TYPE(display) >= PCH_CNP) { panel->backlight.pwm_funcs = &cnp_pwm_funcs; - } else if (INTEL_PCH_TYPE(i915) >= PCH_LPT_H) { - if (HAS_PCH_LPT(i915)) + } else if (INTEL_PCH_TYPE(display) >= PCH_LPT_H) { + if (HAS_PCH_LPT(display)) panel->backlight.pwm_funcs = &lpt_pwm_funcs; else panel->backlight.pwm_funcs = &spt_pwm_funcs; - } else if (HAS_PCH_SPLIT(i915)) { + } else if (HAS_PCH_SPLIT(display)) { panel->backlight.pwm_funcs = &pch_pwm_funcs; } else if (display->platform.valleyview || display->platform.cherryview) { if (connector->base.connector_type == DRM_MODE_CONNECTOR_DSI) { diff --git a/drivers/gpu/drm/i915/display/intel_bios.c b/drivers/gpu/drm/i915/display/intel_bios.c index fabfcf2caa69..ba7b8938b17c 100644 --- a/drivers/gpu/drm/i915/display/intel_bios.c +++ b/drivers/gpu/drm/i915/display/intel_bios.c @@ -2245,28 +2245,27 @@ static const u8 adlp_ddc_pin_map[] = { static u8 map_ddc_pin(struct intel_display *display, u8 vbt_pin) { - struct drm_i915_private *i915 = to_i915(display->drm); const u8 *ddc_pin_map; int i, n_entries; - if (INTEL_PCH_TYPE(i915) >= PCH_MTL || display->platform.alderlake_p) { + if (INTEL_PCH_TYPE(display) >= PCH_MTL || display->platform.alderlake_p) { ddc_pin_map = adlp_ddc_pin_map; n_entries = ARRAY_SIZE(adlp_ddc_pin_map); } else if (display->platform.alderlake_s) { ddc_pin_map = adls_ddc_pin_map; n_entries = ARRAY_SIZE(adls_ddc_pin_map); - } else if (INTEL_PCH_TYPE(i915) >= PCH_DG1) { + } else if (INTEL_PCH_TYPE(display) >= PCH_DG1) { return vbt_pin; - } else if (display->platform.rocketlake && INTEL_PCH_TYPE(i915) == PCH_TGP) { + } else if (display->platform.rocketlake && INTEL_PCH_TYPE(display) == PCH_TGP) { ddc_pin_map = rkl_pch_tgp_ddc_pin_map; n_entries = ARRAY_SIZE(rkl_pch_tgp_ddc_pin_map); - } else if (HAS_PCH_TGP(i915) && DISPLAY_VER(display) == 9) { + } else if (HAS_PCH_TGP(display) && DISPLAY_VER(display) == 9) { ddc_pin_map = gen9bc_tgp_ddc_pin_map; n_entries = ARRAY_SIZE(gen9bc_tgp_ddc_pin_map); - } else if (INTEL_PCH_TYPE(i915) >= PCH_ICP) { + } else if (INTEL_PCH_TYPE(display) >= PCH_ICP) { ddc_pin_map = icp_ddc_pin_map; n_entries = ARRAY_SIZE(icp_ddc_pin_map); - } else if (HAS_PCH_CNP(i915)) { + } else if (HAS_PCH_CNP(display)) { ddc_pin_map = cnp_ddc_pin_map; n_entries = ARRAY_SIZE(cnp_ddc_pin_map); } else { @@ -2865,8 +2864,6 @@ parse_general_definitions(struct intel_display *display) static void init_vbt_defaults(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - display->vbt.crt_ddc_pin = GMBUS_PIN_VGADDC; /* general features */ @@ -2883,7 +2880,7 @@ init_vbt_defaults(struct intel_display *display) * clock for LVDS. */ display->vbt.lvds_ssc_freq = intel_bios_ssc_frequency(display, - !HAS_PCH_SPLIT(i915)); + !HAS_PCH_SPLIT(display)); drm_dbg_kms(display->drm, "Set default to SSC at %d kHz\n", display->vbt.lvds_ssc_freq); } @@ -3126,7 +3123,7 @@ static const struct vbt_header *intel_bios_get_vbt(struct intel_display *display * If the OpRegion does not have VBT, look in SPI flash * through MMIO or PCI mapping */ - if (!vbt && IS_DGFX(i915)) + if (!vbt && display->platform.dgfx) with_intel_display_rpm(display) vbt = oprom_get_vbt(display, intel_rom_spi(i915), sizep, "SPI flash"); diff --git a/drivers/gpu/drm/i915/display/intel_bios.h b/drivers/gpu/drm/i915/display/intel_bios.h index f9841f0498c6..6cd7a011b8c4 100644 --- a/drivers/gpu/drm/i915/display/intel_bios.h +++ b/drivers/gpu/drm/i915/display/intel_bios.h @@ -24,7 +24,7 @@ /* * Please use intel_vbt_defs.h for VBT private data, to hide and abstract away * the VBT from the rest of the driver. Add the parsed, clean data to struct - * intel_vbt_data within struct drm_i915_private. + * intel_vbt_data within struct intel_display. */ #ifndef _INTEL_BIOS_H_ diff --git a/drivers/gpu/drm/i915/display/intel_cdclk.c b/drivers/gpu/drm/i915/display/intel_cdclk.c index 6830950aae3f..b1718b491ffd 100644 --- a/drivers/gpu/drm/i915/display/intel_cdclk.c +++ b/drivers/gpu/drm/i915/display/intel_cdclk.c @@ -3494,7 +3494,6 @@ static int dg1_rawclk(struct intel_display *display) static int cnp_rawclk(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); int divider, fraction; u32 rawclk; @@ -3514,7 +3513,7 @@ static int cnp_rawclk(struct intel_display *display) rawclk |= CNP_RAWCLK_DEN(DIV_ROUND_CLOSEST(numerator * 1000, fraction) - 1); - if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) rawclk |= ICP_RAWCLK_NUM(numerator); } @@ -3553,21 +3552,20 @@ static int i9xx_hrawclk(struct intel_display *display) */ u32 intel_read_rawclk(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 freq; - if (INTEL_PCH_TYPE(dev_priv) >= PCH_MTL) + if (INTEL_PCH_TYPE(display) >= PCH_MTL) /* * MTL always uses a 38.4 MHz rawclk. The bspec tells us * "RAWCLK_FREQ defaults to the values for 38.4 and does * not need to be programmed." */ freq = 38400; - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1) + else if (INTEL_PCH_TYPE(display) >= PCH_DG1) freq = dg1_rawclk(display); - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_CNP) + else if (INTEL_PCH_TYPE(display) >= PCH_CNP) freq = cnp_rawclk(display); - else if (HAS_PCH_SPLIT(dev_priv)) + else if (HAS_PCH_SPLIT(display)) freq = pch_rawclk(display); else if (display->platform.valleyview || display->platform.cherryview) freq = vlv_hrawclk(display); diff --git a/drivers/gpu/drm/i915/display/intel_cmtg.c b/drivers/gpu/drm/i915/display/intel_cmtg.c index 07d7f4e8f60f..82606ebae1de 100644 --- a/drivers/gpu/drm/i915/display/intel_cmtg.c +++ b/drivers/gpu/drm/i915/display/intel_cmtg.c @@ -9,7 +9,6 @@ #include <drm/drm_device.h> #include <drm/drm_print.h> -#include "i915_drv.h" #include "i915_reg.h" #include "intel_crtc.h" #include "intel_cmtg.h" diff --git a/drivers/gpu/drm/i915/display/intel_crt.c b/drivers/gpu/drm/i915/display/intel_crt.c index cca22d2402e8..38b50a779b6b 100644 --- a/drivers/gpu/drm/i915/display/intel_crt.c +++ b/drivers/gpu/drm/i915/display/intel_crt.c @@ -31,9 +31,9 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_crtc.h> #include <drm/drm_edid.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> -#include "i915_drv.h" #include "i915_irq.h" #include "i915_reg.h" #include "intel_connector.h" @@ -91,13 +91,12 @@ static struct intel_crt *intel_attached_crt(struct intel_connector *connector) bool intel_crt_port_enabled(struct intel_display *display, i915_reg_t adpa_reg, enum pipe *pipe) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 val; val = intel_de_read(display, adpa_reg); /* asserts want to know the pipe even if the port is disabled */ - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) *pipe = REG_FIELD_GET(ADPA_PIPE_SEL_MASK_CPT, val); else *pipe = REG_FIELD_GET(ADPA_PIPE_SEL_MASK, val); @@ -177,7 +176,6 @@ static void intel_crt_set_dpms(struct intel_encoder *encoder, int mode) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crt *crt = intel_encoder_to_crt(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; @@ -194,14 +192,14 @@ static void intel_crt_set_dpms(struct intel_encoder *encoder, adpa |= ADPA_VSYNC_ACTIVE_HIGH; /* For CPT allow 3 pipe config, for others just use A or B */ - if (HAS_PCH_LPT(dev_priv)) + if (HAS_PCH_LPT(display)) ; /* Those bits don't exist here */ - else if (HAS_PCH_CPT(dev_priv)) + else if (HAS_PCH_CPT(display)) adpa |= ADPA_PIPE_SEL_CPT(crtc->pipe); else adpa |= ADPA_PIPE_SEL(crtc->pipe); - if (!HAS_PCH_SPLIT(dev_priv)) + if (!HAS_PCH_SPLIT(display)) intel_de_write(display, BCLRPAT(display, crtc->pipe), 0); switch (mode) { @@ -356,7 +354,6 @@ intel_crt_mode_valid(struct drm_connector *connector, const struct drm_display_mode *mode) { struct intel_display *display = to_intel_display(connector->dev); - struct drm_i915_private *dev_priv = to_i915(connector->dev); int max_dotclk = display->cdclk.max_dotclk_freq; enum drm_mode_status status; int max_clock; @@ -368,9 +365,9 @@ intel_crt_mode_valid(struct drm_connector *connector, if (mode->clock < 25000) return MODE_CLOCK_LOW; - if (HAS_PCH_LPT(dev_priv)) + if (HAS_PCH_LPT(display)) max_clock = 180000; - else if (IS_VALLEYVIEW(dev_priv)) + else if (display->platform.valleyview) /* * 270 MHz due to current DPLL limits, * DAC limit supposedly 355 MHz. @@ -387,7 +384,7 @@ intel_crt_mode_valid(struct drm_connector *connector, return MODE_CLOCK_HIGH; /* The FDI receiver on LPT only supports 8bpc and only has 2 lanes. */ - if (HAS_PCH_LPT(dev_priv) && + if (HAS_PCH_LPT(display) && ilk_get_lanes_required(mode->clock, 270000, 24) > 2) return MODE_CLOCK_HIGH; @@ -438,7 +435,6 @@ static int hsw_crt_compute_config(struct intel_encoder *encoder, struct drm_connector_state *conn_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; @@ -457,7 +453,7 @@ static int hsw_crt_compute_config(struct intel_encoder *encoder, crtc_state->output_format = INTEL_OUTPUT_FORMAT_RGB; /* LPT FDI RX only supports 8bpc. */ - if (HAS_PCH_LPT(dev_priv)) { + if (HAS_PCH_LPT(display)) { /* TODO: Check crtc_state->max_link_bpp_x16 instead of bw_constrained */ if (crtc_state->bw_constrained && crtc_state->pipe_bpp < 24) { drm_dbg_kms(display->drm, @@ -482,13 +478,12 @@ static bool ilk_crt_detect_hotplug(struct drm_connector *connector) { struct intel_display *display = to_intel_display(connector->dev); struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector)); - struct drm_i915_private *dev_priv = to_i915(connector->dev); u32 adpa; bool ret; /* The first time through, trigger an explicit detection cycle */ if (crt->force_hotplug_required) { - bool turn_off_dac = HAS_PCH_SPLIT(dev_priv); + bool turn_off_dac = HAS_PCH_SPLIT(display); u32 save_adpa; crt->force_hotplug_required = false; @@ -583,15 +578,14 @@ static bool valleyview_crt_detect_hotplug(struct drm_connector *connector) static bool intel_crt_detect_hotplug(struct drm_connector *connector) { struct intel_display *display = to_intel_display(connector->dev); - struct drm_i915_private *dev_priv = to_i915(connector->dev); u32 stat; bool ret = false; int i, tries = 0; - if (HAS_PCH_SPLIT(dev_priv)) + if (HAS_PCH_SPLIT(display)) return ilk_crt_detect_hotplug(connector); - if (IS_VALLEYVIEW(dev_priv)) + if (display->platform.valleyview) return valleyview_crt_detect_hotplug(connector); /* @@ -599,7 +593,7 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector) * to get a reliable result. */ - if (IS_G45(dev_priv)) + if (display->platform.g45) tries = 2; else tries = 1; @@ -940,7 +934,6 @@ out: static int intel_crt_get_modes(struct drm_connector *connector) { struct intel_display *display = to_intel_display(connector->dev); - struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector)); struct intel_encoder *encoder = &crt->base; intel_wakeref_t wakeref; @@ -953,7 +946,7 @@ static int intel_crt_get_modes(struct drm_connector *connector) wakeref = intel_display_power_get(display, encoder->power_domain); ret = intel_crt_ddc_get_modes(connector, connector->ddc); - if (ret || !IS_G4X(dev_priv)) + if (ret || !display->platform.g4x) goto out; /* Try to probe digital port for output in DVI-I -> VGA mode. */ @@ -1012,16 +1005,15 @@ static const struct drm_encoder_funcs intel_crt_enc_funcs = { void intel_crt_init(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_connector *connector; struct intel_crt *crt; i915_reg_t adpa_reg; u8 ddc_pin; u32 adpa; - if (HAS_PCH_SPLIT(dev_priv)) + if (HAS_PCH_SPLIT(display)) adpa_reg = PCH_ADPA; - else if (IS_VALLEYVIEW(dev_priv)) + else if (display->platform.valleyview) adpa_reg = VLV_ADPA; else adpa_reg = ADPA; @@ -1069,7 +1061,7 @@ void intel_crt_init(struct intel_display *display) crt->base.type = INTEL_OUTPUT_ANALOG; crt->base.cloneable = BIT(INTEL_OUTPUT_DVO) | BIT(INTEL_OUTPUT_HDMI); - if (IS_I830(dev_priv)) + if (display->platform.i830) crt->base.pipe_mask = BIT(PIPE_A); else crt->base.pipe_mask = ~0; @@ -1109,7 +1101,7 @@ void intel_crt_init(struct intel_display *display) intel_ddi_buf_trans_init(&crt->base); } else { - if (HAS_PCH_SPLIT(dev_priv)) { + if (HAS_PCH_SPLIT(display)) { crt->base.compute_config = pch_crt_compute_config; crt->base.disable = pch_disable_crt; crt->base.post_disable = pch_post_disable_crt; @@ -1131,7 +1123,7 @@ void intel_crt_init(struct intel_display *display) * polarity and link reversal bits or not, instead of relying on the * BIOS. */ - if (HAS_PCH_LPT(dev_priv)) { + if (HAS_PCH_LPT(display)) { u32 fdi_config = FDI_RX_POLARITY_REVERSED_LPT | FDI_RX_LINK_REVERSAL_OVERRIDE; diff --git a/drivers/gpu/drm/i915/display/intel_crtc.c b/drivers/gpu/drm/i915/display/intel_crtc.c index 5b2603ef2ff7..29cfc38f12e0 100644 --- a/drivers/gpu/drm/i915/display/intel_crtc.c +++ b/drivers/gpu/drm/i915/display/intel_crtc.c @@ -124,7 +124,7 @@ void intel_crtc_vblank_on(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - crtc->block_dc_for_vblank = intel_psr_needs_block_dc_vblank(crtc_state); + crtc->vblank_psr_notify = intel_psr_needs_vblank_notification(crtc_state); assert_vblank_disabled(&crtc->base); drm_crtc_set_max_vblank_count(&crtc->base, @@ -154,9 +154,9 @@ void intel_crtc_vblank_off(const struct intel_crtc_state *crtc_state) drm_crtc_vblank_off(&crtc->base); assert_vblank_disabled(&crtc->base); - crtc->block_dc_for_vblank = false; + crtc->vblank_psr_notify = false; - flush_work(&display->irq.vblank_dc_work); + flush_work(&display->irq.vblank_notify_work); } struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc) @@ -305,7 +305,6 @@ static const struct drm_crtc_funcs i8xx_crtc_funcs = { int intel_crtc_init(struct intel_display *display, enum pipe pipe) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_plane *primary, *cursor; const struct drm_crtc_funcs *funcs; struct intel_crtc *crtc; @@ -333,7 +332,7 @@ int intel_crtc_init(struct intel_display *display, enum pipe pipe) for_each_sprite(display, pipe, sprite) { struct intel_plane *plane; - if (DISPLAY_VER(dev_priv) >= 9) + if (DISPLAY_VER(display) >= 9) plane = skl_universal_plane_create(display, pipe, PLANE_2 + sprite); else plane = intel_sprite_plane_create(display, pipe, sprite); diff --git a/drivers/gpu/drm/i915/display/intel_cursor.c b/drivers/gpu/drm/i915/display/intel_cursor.c index 3276a5b4a9b0..2fec5ba58373 100644 --- a/drivers/gpu/drm/i915/display/intel_cursor.c +++ b/drivers/gpu/drm/i915/display/intel_cursor.c @@ -9,10 +9,11 @@ #include <drm/drm_blend.h> #include <drm/drm_damage_helper.h> #include <drm/drm_fourcc.h> +#include <drm/drm_print.h> #include <drm/drm_vblank.h> -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_atomic.h" #include "intel_atomic_plane.h" #include "intel_cursor.h" diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy.c b/drivers/gpu/drm/i915/display/intel_cx0_phy.c index 22595766eac5..a82b93cbc81d 100644 --- a/drivers/gpu/drm/i915/display/intel_cx0_phy.c +++ b/drivers/gpu/drm/i915/display/intel_cx0_phy.c @@ -6,8 +6,10 @@ #include <linux/log2.h> #include <linux/math64.h> -#include "i915_drv.h" +#include <drm/drm_print.h> + #include "i915_reg.h" +#include "i915_utils.h" #include "intel_cx0_phy.h" #include "intel_cx0_phy_regs.h" #include "intel_ddi.h" @@ -2761,9 +2763,9 @@ static void intel_program_port_clock_ctl(struct intel_encoder *encoder, val |= XELPDP_FORWARD_CLOCK_UNGATE; if (!is_dp && is_hdmi_frl(port_clock)) - val |= XELPDP_DDI_CLOCK_SELECT(XELPDP_DDI_CLOCK_SELECT_DIV18CLK); + val |= XELPDP_DDI_CLOCK_SELECT_PREP(display, XELPDP_DDI_CLOCK_SELECT_DIV18CLK); else - val |= XELPDP_DDI_CLOCK_SELECT(XELPDP_DDI_CLOCK_SELECT_MAXPCLK); + val |= XELPDP_DDI_CLOCK_SELECT_PREP(display, XELPDP_DDI_CLOCK_SELECT_MAXPCLK); /* TODO: HDMI FRL */ /* DP2.0 10G and 20G rates enable MPLLA*/ @@ -2774,7 +2776,7 @@ static void intel_program_port_clock_ctl(struct intel_encoder *encoder, intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port), XELPDP_LANE1_PHY_CLOCK_SELECT | XELPDP_FORWARD_CLOCK_UNGATE | - XELPDP_DDI_CLOCK_SELECT_MASK | XELPDP_SSC_ENABLE_PLLA | + XELPDP_DDI_CLOCK_SELECT_MASK(display) | XELPDP_SSC_ENABLE_PLLA | XELPDP_SSC_ENABLE_PLLB, val); } @@ -3097,10 +3099,7 @@ int intel_mtl_tbt_calc_port_clock(struct intel_encoder *encoder) val = intel_de_read(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port)); - if (DISPLAY_VER(display) >= 30) - clock = REG_FIELD_GET(XE3_DDI_CLOCK_SELECT_MASK, val); - else - clock = REG_FIELD_GET(XELPDP_DDI_CLOCK_SELECT_MASK, val); + clock = XELPDP_DDI_CLOCK_SELECT_GET(display, val); drm_WARN_ON(display->drm, !(val & XELPDP_FORWARD_CLOCK_UNGATE)); drm_WARN_ON(display->drm, !(val & XELPDP_TBT_CLOCK_REQUEST)); @@ -3168,13 +3167,9 @@ static void intel_mtl_tbt_pll_enable(struct intel_encoder *encoder, * clock muxes, gating and SSC */ - if (DISPLAY_VER(display) >= 30) { - mask = XE3_DDI_CLOCK_SELECT_MASK; - val |= XE3_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(display, crtc_state->port_clock)); - } else { - mask = XELPDP_DDI_CLOCK_SELECT_MASK; - val |= XELPDP_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(display, crtc_state->port_clock)); - } + mask = XELPDP_DDI_CLOCK_SELECT_MASK(display); + val |= XELPDP_DDI_CLOCK_SELECT_PREP(display, + intel_mtl_tbt_clock_select(display, crtc_state->port_clock)); mask |= XELPDP_FORWARD_CLOCK_UNGATE; val |= XELPDP_FORWARD_CLOCK_UNGATE; @@ -3287,7 +3282,7 @@ static void intel_cx0pll_disable(struct intel_encoder *encoder) /* 7. Program PORT_CLOCK_CTL register to disable and gate clocks. */ intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port), - XELPDP_DDI_CLOCK_SELECT_MASK, 0); + XELPDP_DDI_CLOCK_SELECT_MASK(display), 0); intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port), XELPDP_FORWARD_CLOCK_UNGATE, 0); @@ -3336,7 +3331,7 @@ static void intel_mtl_tbt_pll_disable(struct intel_encoder *encoder) * 5. Program PORT CLOCK CTRL register to disable and gate clocks */ intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port), - XELPDP_DDI_CLOCK_SELECT_MASK | + XELPDP_DDI_CLOCK_SELECT_MASK(display) | XELPDP_FORWARD_CLOCK_UNGATE, 0); /* 6. Program DDI_CLK_VALFREQ to 0. */ @@ -3365,7 +3360,7 @@ intel_mtl_port_pll_type(struct intel_encoder *encoder, * handling is done via the standard shared DPLL framework. */ val = intel_de_read(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port)); - clock = REG_FIELD_GET(XELPDP_DDI_CLOCK_SELECT_MASK, val); + clock = XELPDP_DDI_CLOCK_SELECT_GET(display, val); if (clock == XELPDP_DDI_CLOCK_SELECT_MAXPCLK || clock == XELPDP_DDI_CLOCK_SELECT_DIV18CLK) diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h b/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h index 960f7f778fb8..59c22beaf1de 100644 --- a/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h +++ b/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h @@ -192,10 +192,17 @@ #define XELPDP_TBT_CLOCK_REQUEST REG_BIT(19) #define XELPDP_TBT_CLOCK_ACK REG_BIT(18) -#define XELPDP_DDI_CLOCK_SELECT_MASK REG_GENMASK(15, 12) -#define XE3_DDI_CLOCK_SELECT_MASK REG_GENMASK(16, 12) -#define XELPDP_DDI_CLOCK_SELECT(val) REG_FIELD_PREP(XELPDP_DDI_CLOCK_SELECT_MASK, val) -#define XE3_DDI_CLOCK_SELECT(val) REG_FIELD_PREP(XE3_DDI_CLOCK_SELECT_MASK, val) +#define _XELPDP_DDI_CLOCK_SELECT_MASK REG_GENMASK(15, 12) +#define _XE3_DDI_CLOCK_SELECT_MASK REG_GENMASK(16, 12) +#define XELPDP_DDI_CLOCK_SELECT_MASK(display) (DISPLAY_VER(display) >= 30 ? \ + _XE3_DDI_CLOCK_SELECT_MASK : _XELPDP_DDI_CLOCK_SELECT_MASK) +#define XELPDP_DDI_CLOCK_SELECT_PREP(display, val) (DISPLAY_VER(display) >= 30 ? \ + REG_FIELD_PREP(_XE3_DDI_CLOCK_SELECT_MASK, (val)) : \ + REG_FIELD_PREP(_XELPDP_DDI_CLOCK_SELECT_MASK, (val))) +#define XELPDP_DDI_CLOCK_SELECT_GET(display, val) (DISPLAY_VER(display) >= 30 ? \ + REG_FIELD_GET(_XE3_DDI_CLOCK_SELECT_MASK, (val)) : \ + REG_FIELD_GET(_XELPDP_DDI_CLOCK_SELECT_MASK, (val))) + #define XELPDP_DDI_CLOCK_SELECT_NONE 0x0 #define XELPDP_DDI_CLOCK_SELECT_MAXPCLK 0x8 #define XELPDP_DDI_CLOCK_SELECT_DIV18CLK 0x9 diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c index b48ed5df7a96..74132c1d6385 100644 --- a/drivers/gpu/drm/i915/display/intel_ddi.c +++ b/drivers/gpu/drm/i915/display/intel_ddi.c @@ -30,11 +30,13 @@ #include <drm/display/drm_dp_helper.h> #include <drm/display/drm_scdc_helper.h> +#include <drm/drm_print.h> #include <drm/drm_privacy_screen_consumer.h> -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "icl_dsi.h" +#include "intel_alpm.h" #include "intel_audio.h" #include "intel_audio_regs.h" #include "intel_backlight.h" @@ -3553,6 +3555,7 @@ static void intel_ddi_disable_dp(struct intel_atomic_state *state, intel_dp->link.active = false; intel_psr_disable(intel_dp, old_crtc_state); + intel_alpm_disable(intel_dp); intel_edp_backlight_off(old_conn_state); /* Disable the decompression in DP Sink */ intel_dp_sink_disable_decompression(state, @@ -4902,9 +4905,7 @@ static enum hpd_pin tgl_hpd_pin(struct intel_display *display, enum port port) static enum hpd_pin rkl_hpd_pin(struct intel_display *display, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - if (HAS_PCH_TGP(dev_priv)) + if (HAS_PCH_TGP(display)) return tgl_hpd_pin(display, port); if (port >= PORT_TC1) @@ -4923,12 +4924,10 @@ static enum hpd_pin icl_hpd_pin(struct intel_display *display, enum port port) static enum hpd_pin ehl_hpd_pin(struct intel_display *display, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (port == PORT_D) return HPD_PORT_A; - if (HAS_PCH_TGP(dev_priv)) + if (HAS_PCH_TGP(display)) return icl_hpd_pin(display, port); return HPD_PORT_A + port - PORT_A; @@ -4936,9 +4935,7 @@ static enum hpd_pin ehl_hpd_pin(struct intel_display *display, enum port port) static enum hpd_pin skl_hpd_pin(struct intel_display *display, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - if (HAS_PCH_TGP(dev_priv)) + if (HAS_PCH_TGP(display)) return icl_hpd_pin(display, port); return HPD_PORT_A + port - PORT_A; diff --git a/drivers/gpu/drm/i915/display/intel_de.h b/drivers/gpu/drm/i915/display/intel_de.h index 655467a6ba87..54ce3e4f8fd9 100644 --- a/drivers/gpu/drm/i915/display/intel_de.h +++ b/drivers/gpu/drm/i915/display/intel_de.h @@ -6,7 +6,6 @@ #ifndef __INTEL_DE_H__ #define __INTEL_DE_H__ -#include "intel_display_conversion.h" #include "intel_display_core.h" #include "intel_dmc_wl.h" #include "intel_dsb.h" @@ -19,7 +18,7 @@ static inline struct intel_uncore *__to_uncore(struct intel_display *display) } static inline u32 -__intel_de_read(struct intel_display *display, i915_reg_t reg) +intel_de_read(struct intel_display *display, i915_reg_t reg) { u32 val; @@ -31,7 +30,6 @@ __intel_de_read(struct intel_display *display, i915_reg_t reg) return val; } -#define intel_de_read(p,...) __intel_de_read(__to_intel_display(p), __VA_ARGS__) static inline u8 intel_de_read8(struct intel_display *display, i915_reg_t reg) @@ -66,7 +64,7 @@ intel_de_read64_2x32(struct intel_display *display, } static inline void -__intel_de_posting_read(struct intel_display *display, i915_reg_t reg) +intel_de_posting_read(struct intel_display *display, i915_reg_t reg) { intel_dmc_wl_get(display, reg); @@ -74,10 +72,9 @@ __intel_de_posting_read(struct intel_display *display, i915_reg_t reg) intel_dmc_wl_put(display, reg); } -#define intel_de_posting_read(p,...) __intel_de_posting_read(__to_intel_display(p), __VA_ARGS__) static inline void -__intel_de_write(struct intel_display *display, i915_reg_t reg, u32 val) +intel_de_write(struct intel_display *display, i915_reg_t reg, u32 val) { intel_dmc_wl_get(display, reg); @@ -85,7 +82,6 @@ __intel_de_write(struct intel_display *display, i915_reg_t reg, u32 val) intel_dmc_wl_put(display, reg); } -#define intel_de_write(p,...) __intel_de_write(__to_intel_display(p), __VA_ARGS__) static inline u32 __intel_de_rmw_nowl(struct intel_display *display, i915_reg_t reg, @@ -95,8 +91,7 @@ __intel_de_rmw_nowl(struct intel_display *display, i915_reg_t reg, } static inline u32 -__intel_de_rmw(struct intel_display *display, i915_reg_t reg, u32 clear, - u32 set) +intel_de_rmw(struct intel_display *display, i915_reg_t reg, u32 clear, u32 set) { u32 val; @@ -108,7 +103,6 @@ __intel_de_rmw(struct intel_display *display, i915_reg_t reg, u32 clear, return val; } -#define intel_de_rmw(p,...) __intel_de_rmw(__to_intel_display(p), __VA_ARGS__) static inline int __intel_de_wait_for_register_nowl(struct intel_display *display, diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 404dd4a8b7df..6f0a0bc71b06 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -55,6 +55,7 @@ #include "i9xx_plane.h" #include "i9xx_plane_regs.h" #include "i9xx_wm.h" +#include "intel_alpm.h" #include "intel_atomic.h" #include "intel_atomic_plane.h" #include "intel_audio.h" @@ -1043,14 +1044,13 @@ static void intel_post_plane_update(struct intel_atomic_state *state, struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(state); - struct drm_i915_private *dev_priv = to_i915(state->base.dev); const struct intel_crtc_state *old_crtc_state = intel_atomic_get_old_crtc_state(state, crtc); const struct intel_crtc_state *new_crtc_state = intel_atomic_get_new_crtc_state(state, crtc); enum pipe pipe = crtc->pipe; - intel_frontbuffer_flip(dev_priv, new_crtc_state->fb_bits); + intel_frontbuffer_flip(display, new_crtc_state->fb_bits); if (new_crtc_state->update_wm_post && new_crtc_state->hw.active) intel_update_watermarks(display); @@ -1079,6 +1079,8 @@ static void intel_post_plane_update(struct intel_atomic_state *state, if (audio_enabling(old_crtc_state, new_crtc_state)) intel_encoders_audio_enable(state, crtc); + intel_alpm_post_plane_update(state, crtc); + intel_psr_post_plane_update(state, crtc); } @@ -1174,6 +1176,7 @@ static void intel_pre_plane_update(struct intel_atomic_state *state, intel_atomic_get_new_crtc_state(state, crtc); enum pipe pipe = crtc->pipe; + intel_alpm_pre_plane_update(state, crtc); intel_psr_pre_plane_update(state, crtc); if (intel_crtc_vrr_disabling(state, crtc)) { @@ -1281,7 +1284,7 @@ static void intel_pre_plane_update(struct intel_atomic_state *state, static void intel_crtc_disable_planes(struct intel_atomic_state *state, struct intel_crtc *crtc) { - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(state); const struct intel_crtc_state *new_crtc_state = intel_atomic_get_new_crtc_state(state, crtc); unsigned int update_mask = new_crtc_state->update_planes; @@ -1303,7 +1306,7 @@ static void intel_crtc_disable_planes(struct intel_atomic_state *state, fb_bits |= plane->frontbuffer_bit; } - intel_frontbuffer_flip(dev_priv, fb_bits); + intel_frontbuffer_flip(display, fb_bits); } static void intel_encoders_update_prepare(struct intel_atomic_state *state) @@ -1510,7 +1513,6 @@ static void ilk_crtc_enable(struct intel_atomic_state *state, struct intel_display *display = to_intel_display(crtc); const struct intel_crtc_state *new_crtc_state = intel_atomic_get_new_crtc_state(state, crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; if (drm_WARN_ON(display->drm, crtc->active)) @@ -1562,7 +1564,7 @@ static void ilk_crtc_enable(struct intel_atomic_state *state, intel_encoders_enable(state, crtc); - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) intel_wait_for_pipe_scanline_moving(crtc); /* @@ -2404,14 +2406,6 @@ static int intel_crtc_compute_pipe_mode(struct intel_crtc_state *crtc_state) return 0; } -static bool intel_crtc_needs_wa_14015401596(const struct intel_crtc_state *crtc_state) -{ - struct intel_display *display = to_intel_display(crtc_state); - - return intel_vrr_possible(crtc_state) && crtc_state->has_psr && - IS_DISPLAY_VER(display, 13, 14); -} - static int intel_crtc_vblank_delay(const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); @@ -2420,9 +2414,7 @@ static int intel_crtc_vblank_delay(const struct intel_crtc_state *crtc_state) if (!HAS_DSB(display)) return 0; - /* Wa_14015401596 */ - if (intel_crtc_needs_wa_14015401596(crtc_state)) - vblank_delay = max(vblank_delay, 1); + vblank_delay = max(vblank_delay, intel_psr_min_vblank_delay(crtc_state)); return vblank_delay; } @@ -2534,15 +2526,13 @@ intel_link_compute_m_n(u16 bits_per_pixel_x16, int nlanes, void intel_panel_sanitize_ssc(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - /* * There may be no VBT; and if the BIOS enabled SSC we can * just keep using it to avoid unnecessary flicker. Whereas if the * BIOS isn't using it, don't assume it will work even if the VBT * indicates as much. */ - if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_IBX(display) || HAS_PCH_CPT(display)) { bool bios_lvds_use_ssc = intel_de_read(display, PCH_DREF_CONTROL) & DREF_SSC1_ENABLE; @@ -2724,6 +2714,19 @@ static void intel_set_transcoder_timings(const struct intel_crtc_state *crtc_sta intel_de_write(display, TRANS_VTOTAL(display, pipe), VACTIVE(crtc_vdisplay - 1) | VTOTAL(crtc_vtotal - 1)); + + if (DISPLAY_VER(display) >= 30) { + /* + * Address issues for resolutions with high refresh rate that + * have small Hblank, specifically where Hblank is smaller than + * one MTP. Simulations indicate this will address the + * jitter issues that currently causes BS to be immediately + * followed by BE which DPRX devices are unable to handle. + * https://groups.vesa.org/wg/DP/document/20494 + */ + intel_de_write(display, DP_MIN_HBLANK_CTL(cpu_transcoder), + crtc_state->min_hblank); + } } static void intel_set_transcoder_timings_lrr(const struct intel_crtc_state *crtc_state) @@ -2867,6 +2870,10 @@ static void intel_get_transcoder_timings(struct intel_crtc *crtc, adjusted_mode->crtc_vdisplay + intel_de_read(display, TRANS_SET_CONTEXT_LATENCY(display, cpu_transcoder)); + + if (DISPLAY_VER(display) >= 30) + pipe_config->min_hblank = intel_de_read(display, + DP_MIN_HBLANK_CTL(cpu_transcoder)); } static void intel_joiner_adjust_pipe_src(struct intel_crtc_state *crtc_state) @@ -5217,6 +5224,8 @@ intel_pipe_config_compare(const struct intel_crtc_state *current_config, PIPE_CONF_CHECK_I(lane_count); PIPE_CONF_CHECK_X(lane_lat_optim_mask); + PIPE_CONF_CHECK_I(min_hblank); + if (HAS_DOUBLE_BUFFERED_M_N(display)) { if (!fastset || !pipe_config->update_m_n) PIPE_CONF_CHECK_M_N(dp_m_n); @@ -6541,7 +6550,6 @@ static void intel_pipe_fastset(const struct intel_crtc_state *old_crtc_state, { struct intel_display *display = to_intel_display(new_crtc_state); struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); /* * Update pipe size and adjust fitter if needed: the reason for this is @@ -6557,7 +6565,7 @@ static void intel_pipe_fastset(const struct intel_crtc_state *old_crtc_state, if (DISPLAY_VER(display) >= 9) { if (new_crtc_state->pch_pfit.enabled) skl_pfit_enable(new_crtc_state); - } else if (HAS_PCH_SPLIT(dev_priv)) { + } else if (HAS_PCH_SPLIT(display)) { if (new_crtc_state->pch_pfit.enabled) ilk_pfit_enable(new_crtc_state); else if (old_crtc_state->pch_pfit.enabled) @@ -6658,6 +6666,8 @@ static void intel_enable_crtc(struct intel_atomic_state *state, intel_crtc_update_active_timings(pipe_crtc_state, false); } + intel_psr_notify_pipe_change(state, crtc, true); + display->funcs.display->crtc_enable(state, crtc); /* vblanks work again, re-enable pipe CRC. */ @@ -6777,6 +6787,8 @@ static void intel_old_crtc_state_disables(struct intel_atomic_state *state, intel_crtc_joined_pipe_mask(old_crtc_state)) intel_crtc_disable_pipe_crc(pipe_crtc); + intel_psr_notify_pipe_change(state, crtc, false); + display->funcs.display->crtc_disable(state, crtc); for_each_intel_crtc_in_pipe_mask(display->drm, pipe_crtc, @@ -7633,15 +7645,13 @@ static bool ilk_has_edp_a(struct intel_display *display) static bool intel_ddi_crt_present(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (DISPLAY_VER(display) >= 9) return false; if (display->platform.haswell_ult || display->platform.broadwell_ult) return false; - if (HAS_PCH_LPT_H(dev_priv) && + if (HAS_PCH_LPT_H(display) && intel_de_read(display, SFUSE_STRAP) & SFUSE_STRAP_CRT_DISABLED) return false; @@ -7663,7 +7673,6 @@ bool assert_port_valid(struct intel_display *display, enum port port) void intel_setup_outputs(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder; bool dpd_is_edp = false; @@ -7680,7 +7689,7 @@ void intel_setup_outputs(struct intel_display *display) if (display->platform.geminilake || display->platform.broxton) vlv_dsi_init(display); - } else if (HAS_PCH_SPLIT(dev_priv)) { + } else if (HAS_PCH_SPLIT(display)) { int found; /* @@ -8048,13 +8057,11 @@ static const struct intel_display_funcs i9xx_display_funcs = { */ void intel_init_display_hooks(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (DISPLAY_VER(display) >= 9) { display->funcs.display = &skl_display_funcs; } else if (HAS_DDI(display)) { display->funcs.display = &ddi_display_funcs; - } else if (HAS_PCH_SPLIT(dev_priv)) { + } else if (HAS_PCH_SPLIT(display)) { display->funcs.display = &pch_split_display_funcs; } else if (display->platform.cherryview || display->platform.valleyview) { diff --git a/drivers/gpu/drm/i915/display/intel_display_core.h b/drivers/gpu/drm/i915/display/intel_display_core.h index eb6d6f2d0f75..b4937e102360 100644 --- a/drivers/gpu/drm/i915/display/intel_display_core.h +++ b/drivers/gpu/drm/i915/display/intel_display_core.h @@ -21,17 +21,15 @@ #include "intel_display_limits.h" #include "intel_display_params.h" #include "intel_display_power.h" +#include "intel_dmc_wl.h" #include "intel_dpll_mgr.h" #include "intel_fbc.h" #include "intel_global_state.h" #include "intel_gmbus.h" #include "intel_opregion.h" -#include "intel_dmc_wl.h" +#include "intel_pch.h" #include "intel_wm_types.h" -struct task_struct; - -struct drm_i915_private; struct drm_property; struct drm_property_blob; struct i915_audio_component; @@ -52,6 +50,7 @@ struct intel_hotplug_funcs; struct intel_initial_plane_config; struct intel_opregion; struct intel_overlay; +struct task_struct; /* Amount of SAGV/QGV points, BSpec precisely defines this */ #define I915_NUM_QGV_POINTS 8 @@ -180,7 +179,7 @@ struct intel_hotplug { /* * Queuing of hotplug_work, reenable_work and poll_init_work is - * enabled. Protected by drm_i915_private::irq_lock. + * enabled. Protected by intel_display::irq::lock. */ bool detection_work_enabled; @@ -289,6 +288,9 @@ struct intel_display { /* Platform (and subplatform, if any) identification */ struct intel_display_platforms platform; + /* Intel PCH: where the south display engine lives */ + enum intel_pch pch_type; + /* Display functions */ struct { /* Top level crtc-ish functions */ @@ -426,7 +428,7 @@ struct intel_display { * reused when sending message to gsc cs. * this is only populated post Meteorlake */ - struct intel_hdcp_gsc_message *hdcp_message; + struct intel_hdcp_gsc_context *gsc_context; /* Mutex to protect the above hdcp related values. */ struct mutex hdcp_mutex; } hdcp; @@ -454,6 +456,9 @@ struct intel_display { } ips; struct { + /* protects the irq masks */ + spinlock_t lock; + /* * Most platforms treat the display irq block as an always-on * power domain. vlv/chv can disable it at runtime and need @@ -466,9 +471,9 @@ struct intel_display { /* For i915gm/i945gm vblank irq workaround */ u8 vblank_enabled; - int vblank_wa_num_pipes; + int vblank_enable_count; - struct work_struct vblank_dc_work; + struct work_struct vblank_notify_work; u32 de_irq_mask[I915_MAX_PIPES]; u32 pipestat_irq_mask[I915_MAX_PIPES]; @@ -575,6 +580,8 @@ struct intel_display { struct intel_vbt_data vbt; struct intel_dmc_wl wl; struct intel_wm wm; + + struct work_struct psr_dc5_dc6_wa_work; }; #endif /* __INTEL_DISPLAY_CORE_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_display_debugfs.c b/drivers/gpu/drm/i915/display/intel_display_debugfs.c index 4c208fdb9137..8d0a1779dd19 100644 --- a/drivers/gpu/drm/i915/display/intel_display_debugfs.c +++ b/drivers/gpu/drm/i915/display/intel_display_debugfs.c @@ -7,11 +7,12 @@ #include <linux/string_helpers.h> #include <drm/drm_debugfs.h> +#include <drm/drm_drv.h> #include <drm/drm_edid.h> +#include <drm/drm_file.h> #include <drm/drm_fourcc.h> #include "hsw_ips.h" -#include "i915_drv.h" #include "i915_irq.h" #include "i915_reg.h" #include "i9xx_wm_regs.h" @@ -53,10 +54,9 @@ static struct intel_display *node_to_intel_display(struct drm_info_node *node) static int intel_display_caps(struct seq_file *m, void *data) { struct intel_display *display = node_to_intel_display(m->private); - struct drm_i915_private *i915 = to_i915(display->drm); struct drm_printer p = drm_seq_file_printer(m); - drm_printf(&p, "PCH type: %d\n", INTEL_PCH_TYPE(i915)); + drm_printf(&p, "PCH type: %d\n", INTEL_PCH_TYPE(display)); intel_display_device_info_print(DISPLAY_INFO(display), DISPLAY_RUNTIME_INFO(display), &p); @@ -85,7 +85,6 @@ static int i915_frontbuffer_tracking(struct seq_file *m, void *unused) static int i915_sr_status(struct seq_file *m, void *unused) { struct intel_display *display = node_to_intel_display(m->private); - struct drm_i915_private *dev_priv = to_i915(display->drm); intel_wakeref_t wakeref; bool sr_enabled = false; @@ -93,7 +92,7 @@ static int i915_sr_status(struct seq_file *m, void *unused) if (DISPLAY_VER(display) >= 9) /* no global SR status; inspect per-plane WM */; - else if (HAS_PCH_SPLIT(dev_priv)) + else if (HAS_PCH_SPLIT(display)) sr_enabled = intel_de_read(display, WM1_LP_ILK) & WM_LP_ENABLE; else if (display->platform.i965gm || display->platform.g4x || display->platform.i945g || display->platform.i945gm) @@ -558,6 +557,8 @@ static void intel_crtc_info(struct seq_file *m, struct intel_crtc *crtc) seq_printf(m, "\tpipe src=" DRM_RECT_FMT ", dither=%s, bpp=%d\n", DRM_RECT_ARG(&crtc_state->pipe_src), str_yes_no(crtc_state->dither), crtc_state->pipe_bpp); + seq_printf(m, "\tport_clock=%d, lane_count=%d\n", + crtc_state->port_clock, crtc_state->lane_count); intel_scaler_info(m, crtc); diff --git a/drivers/gpu/drm/i915/display/intel_display_device.c b/drivers/gpu/drm/i915/display/intel_display_device.c index 738ae522c8f4..90d714598664 100644 --- a/drivers/gpu/drm/i915/display/intel_display_device.c +++ b/drivers/gpu/drm/i915/display/intel_display_device.c @@ -3,11 +3,13 @@ * Copyright © 2023 Intel Corporation */ -#include <drm/intel/pciids.h> -#include <drm/drm_color_mgmt.h> #include <linux/pci.h> -#include "i915_drv.h" +#include <drm/drm_color_mgmt.h> +#include <drm/drm_drv.h> +#include <drm/drm_print.h> +#include <drm/intel/pciids.h> + #include "i915_reg.h" #include "intel_cx0_phy_regs.h" #include "intel_de.h" @@ -1711,7 +1713,6 @@ void intel_display_device_remove(struct intel_display *display) static void __intel_display_device_info_runtime_init(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_display_runtime_info *display_runtime = DISPLAY_RUNTIME_INFO(display); enum pipe pipe; @@ -1775,7 +1776,7 @@ static void __intel_display_device_info_runtime_init(struct intel_display *displ goto display_fused_off; } - if (IS_DISPLAY_VER(display, 7, 8) && HAS_PCH_SPLIT(i915)) { + if (IS_DISPLAY_VER(display, 7, 8) && HAS_PCH_SPLIT(display)) { u32 fuse_strap = intel_de_read(display, FUSE_STRAP); u32 sfuse_strap = intel_de_read(display, SFUSE_STRAP); @@ -1790,7 +1791,7 @@ static void __intel_display_device_info_runtime_init(struct intel_display *displ */ if (fuse_strap & ILK_INTERNAL_DISPLAY_DISABLE || sfuse_strap & SFUSE_STRAP_DISPLAY_DISABLED || - (HAS_PCH_CPT(i915) && + (HAS_PCH_CPT(display) && !(sfuse_strap & SFUSE_STRAP_FUSE_LOCK))) { drm_info(display->drm, "Display fused off, disabling\n"); diff --git a/drivers/gpu/drm/i915/display/intel_display_driver.c b/drivers/gpu/drm/i915/display/intel_display_driver.c index efee8925987e..411fe7b918a7 100644 --- a/drivers/gpu/drm/i915/display/intel_display_driver.c +++ b/drivers/gpu/drm/i915/display/intel_display_driver.c @@ -54,6 +54,7 @@ #include "intel_plane_initial.h" #include "intel_pmdemand.h" #include "intel_pps.h" +#include "intel_psr.h" #include "intel_quirks.h" #include "intel_vga.h" #include "intel_wm.h" @@ -180,6 +181,9 @@ static void intel_plane_possible_crtcs_init(struct intel_display *display) void intel_display_driver_early_probe(struct intel_display *display) { + /* This must be called before any calls to HAS_PCH_* */ + intel_pch_detect(display); + if (!HAS_DISPLAY(display)) return; @@ -223,6 +227,8 @@ int intel_display_driver_probe_noirq(struct intel_display *display) if (ret) goto cleanup_bios; + intel_psr_dc5_dc6_wa_init(display); + /* FIXME: completely on the wrong abstraction layer */ ret = intel_power_domains_init(display); if (ret < 0) @@ -238,31 +244,45 @@ int intel_display_driver_probe_noirq(struct intel_display *display) intel_dmc_init(display); display->wq.modeset = alloc_ordered_workqueue("i915_modeset", 0); + if (!display->wq.modeset) { + ret = -ENOMEM; + goto cleanup_vga_client_pw_domain_dmc; + } + display->wq.flip = alloc_workqueue("i915_flip", WQ_HIGHPRI | WQ_UNBOUND, WQ_UNBOUND_MAX_ACTIVE); + if (!display->wq.flip) { + ret = -ENOMEM; + goto cleanup_wq_modeset; + } + display->wq.cleanup = alloc_workqueue("i915_cleanup", WQ_HIGHPRI, 0); + if (!display->wq.cleanup) { + ret = -ENOMEM; + goto cleanup_wq_flip; + } intel_mode_config_init(display); ret = intel_cdclk_init(display); if (ret) - goto cleanup_vga_client_pw_domain_dmc; + goto cleanup_wq_cleanup; ret = intel_color_init(display); if (ret) - goto cleanup_vga_client_pw_domain_dmc; + goto cleanup_wq_cleanup; ret = intel_dbuf_init(display); if (ret) - goto cleanup_vga_client_pw_domain_dmc; + goto cleanup_wq_cleanup; ret = intel_bw_init(display); if (ret) - goto cleanup_vga_client_pw_domain_dmc; + goto cleanup_wq_cleanup; ret = intel_pmdemand_init(display); if (ret) - goto cleanup_vga_client_pw_domain_dmc; + goto cleanup_wq_cleanup; intel_init_quirks(display); @@ -270,6 +290,12 @@ int intel_display_driver_probe_noirq(struct intel_display *display) return 0; +cleanup_wq_cleanup: + destroy_workqueue(display->wq.cleanup); +cleanup_wq_flip: + destroy_workqueue(display->wq.flip); +cleanup_wq_modeset: + destroy_workqueue(display->wq.modeset); cleanup_vga_client_pw_domain_dmc: intel_dmc_fini(display); intel_power_domains_driver_remove(display); @@ -415,7 +441,6 @@ bool intel_display_driver_check_access(struct intel_display *display) /* part #2: call after irq install, but before gem init */ int intel_display_driver_probe_nogem(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); enum pipe pipe; int ret; @@ -453,8 +478,6 @@ int intel_display_driver_probe_nogem(struct intel_display *display) intel_hti_init(display); - /* Just disable it once at startup */ - intel_vga_disable(display); intel_setup_outputs(display); ret = intel_dp_tunnel_mgr_init(display); @@ -464,7 +487,7 @@ int intel_display_driver_probe_nogem(struct intel_display *display) intel_display_driver_disable_user_access(display); drm_modeset_lock_all(display->drm); - intel_modeset_setup_hw_state(i915, display->drm->mode_config.acquire_ctx); + intel_modeset_setup_hw_state(display, display->drm->mode_config.acquire_ctx); intel_acpi_assign_connector_fwnodes(display); drm_modeset_unlock_all(display->drm); @@ -525,7 +548,6 @@ int intel_display_driver_probe(struct intel_display *display) void intel_display_driver_register(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); struct drm_printer p = drm_dbg_printer(display->drm, DRM_UT_KMS, "i915 display info:"); @@ -552,7 +574,7 @@ void intel_display_driver_register(struct intel_display *display) drm_kms_helper_poll_init(display->drm); intel_hpd_poll_disable(display); - intel_fbdev_setup(i915); + intel_fbdev_setup(display); intel_display_device_info_print(DISPLAY_INFO(display), DISPLAY_RUNTIME_INFO(display), &p); @@ -687,13 +709,11 @@ __intel_display_driver_resume(struct intel_display *display, struct drm_atomic_state *state, struct drm_modeset_acquire_ctx *ctx) { - struct drm_i915_private *i915 = to_i915(display->drm); struct drm_crtc_state *crtc_state; struct drm_crtc *crtc; int ret, i; - intel_modeset_setup_hw_state(i915, ctx); - intel_vga_redisable(display); + intel_modeset_setup_hw_state(display, ctx); if (!state) return 0; diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.c b/drivers/gpu/drm/i915/display/intel_display_irq.c index d2a35e3630b1..3e73832e5e81 100644 --- a/drivers/gpu/drm/i915/display/intel_display_irq.c +++ b/drivers/gpu/drm/i915/display/intel_display_irq.c @@ -5,7 +5,6 @@ #include <drm/drm_vblank.h> -#include "gt/intel_rps.h" #include "i915_drv.h" #include "i915_irq.h" #include "i915_reg.h" @@ -15,6 +14,7 @@ #include "intel_de.h" #include "intel_display_irq.h" #include "intel_display_rpm.h" +#include "intel_display_rps.h" #include "intel_display_trace.h" #include "intel_display_types.h" #include "intel_dmc_wl.h" @@ -135,7 +135,7 @@ void ilk_update_display_irq(struct intel_display *display, struct drm_i915_private *dev_priv = to_i915(display->drm); u32 new_val; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_WARN_ON(display->drm, enabled_irq_mask & ~interrupt_mask); new_val = dev_priv->irq_mask; @@ -173,7 +173,7 @@ void bdw_update_port_irq(struct intel_display *display, u32 new_val; u32 old_val; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_WARN_ON(display->drm, enabled_irq_mask & ~interrupt_mask); @@ -206,7 +206,7 @@ static void bdw_update_pipe_irq(struct intel_display *display, struct drm_i915_private *dev_priv = to_i915(display->drm); u32 new_val; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_WARN_ON(display->drm, enabled_irq_mask & ~interrupt_mask); @@ -254,7 +254,7 @@ void ibx_display_interrupt_update(struct intel_display *display, drm_WARN_ON(display->drm, enabled_irq_mask & ~interrupt_mask); - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); if (drm_WARN_ON(display->drm, !intel_irqs_enabled(dev_priv))) return; @@ -276,11 +276,10 @@ void ibx_disable_display_interrupt(struct intel_display *display, u32 bits) u32 i915_pipestat_enable_mask(struct intel_display *display, enum pipe pipe) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 status_mask = display->irq.pipestat_irq_mask[pipe]; u32 enable_mask = status_mask << 16; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); if (DISPLAY_VER(display) < 5) goto out; @@ -329,7 +328,7 @@ void i915_enable_pipestat(struct intel_display *display, "pipe %c: status_mask=0x%x\n", pipe_name(pipe), status_mask); - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_WARN_ON(display->drm, !intel_irqs_enabled(dev_priv)); if ((display->irq.pipestat_irq_mask[pipe] & status_mask) == status_mask) @@ -353,7 +352,7 @@ void i915_disable_pipestat(struct intel_display *display, "pipe %c: status_mask=0x%x\n", pipe_name(pipe), status_mask); - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_WARN_ON(display->drm, !intel_irqs_enabled(dev_priv)); if ((display->irq.pipestat_irq_mask[pipe] & status_mask) == 0) @@ -377,28 +376,23 @@ static bool i915_has_legacy_blc_interrupt(struct intel_display *display) return IS_DISPLAY_VER(display, 3, 4) && display->platform.mobile; } -/** - * i915_enable_asle_pipestat - enable ASLE pipestat for OpRegion - * @display: display device - */ -void i915_enable_asle_pipestat(struct intel_display *display) +/* enable ASLE pipestat for OpRegion */ +static void i915_enable_asle_pipestat(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (!intel_opregion_asle_present(display)) return; if (!i915_has_legacy_blc_interrupt(display)) return; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); i915_enable_pipestat(display, PIPE_B, PIPE_LEGACY_BLC_EVENT_STATUS); if (DISPLAY_VER(display) >= 4) i915_enable_pipestat(display, PIPE_A, PIPE_LEGACY_BLC_EVENT_STATUS); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } #if IS_ENABLED(CONFIG_DEBUG_FS) @@ -517,14 +511,13 @@ static void i9xx_pipestat_irq_reset(struct intel_display *display) void i9xx_pipestat_irq_ack(struct intel_display *display, u32 iir, u32 pipe_stats[I915_MAX_PIPES]) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe; - spin_lock(&dev_priv->irq_lock); + spin_lock(&display->irq.lock); if ((display->platform.valleyview || display->platform.cherryview) && !display->irq.vlv_display_irqs_enabled) { - spin_unlock(&dev_priv->irq_lock); + spin_unlock(&display->irq.lock); return; } @@ -579,7 +572,7 @@ void i9xx_pipestat_irq_ack(struct intel_display *display, intel_de_write(display, reg, enable_mask); } } - spin_unlock(&dev_priv->irq_lock); + spin_unlock(&display->irq.lock); } void i915_pipestat_irq_handler(struct intel_display *display, @@ -879,7 +872,6 @@ static void ilk_gtt_fault_irq_handler(struct intel_display *display) void ilk_display_irq_handler(struct intel_display *display, u32 de_iir) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe; u32 hotplug_trigger = de_iir & DE_DP_A_HOTPLUG; @@ -916,7 +908,7 @@ void ilk_display_irq_handler(struct intel_display *display, u32 de_iir) if (de_iir & DE_PCH_EVENT) { u32 pch_iir = intel_de_read(display, SDEIIR); - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) cpt_irq_handler(display, pch_iir); else ibx_irq_handler(display, pch_iir); @@ -926,12 +918,11 @@ void ilk_display_irq_handler(struct intel_display *display, u32 de_iir) } if (DISPLAY_VER(display) == 5 && de_iir & DE_PCU_EVENT) - gen5_rps_irq_handler(&to_gt(dev_priv)->rps); + ilk_display_rps_irq_handler(display); } void ivb_display_irq_handler(struct intel_display *display, u32 de_iir) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe; u32 hotplug_trigger = de_iir & DE_DP_A_HOTPLUG_IVB; @@ -969,7 +960,7 @@ void ivb_display_irq_handler(struct intel_display *display, u32 de_iir) } /* check event from PCH */ - if (!HAS_PCH_NOP(dev_priv) && (de_iir & DE_PCH_EVENT_IVB)) { + if (!HAS_PCH_NOP(display) && (de_iir & DE_PCH_EVENT_IVB)) { u32 pch_iir = intel_de_read(display, SDEIIR); cpt_irq_handler(display, pch_iir); @@ -1311,7 +1302,6 @@ static u32 gen8_de_pipe_flip_done_mask(struct intel_display *display) static void gen8_read_and_ack_pch_irqs(struct intel_display *display, u32 *pch_iir, u32 *pica_iir) { - struct drm_i915_private *i915 = to_i915(display->drm); u32 pica_ier = 0; *pica_iir = 0; @@ -1325,7 +1315,7 @@ static void gen8_read_and_ack_pch_irqs(struct intel_display *display, u32 *pch_i * their flags both in the PICA and SDE IIR. */ if (*pch_iir & SDE_PICAINTERRUPT) { - drm_WARN_ON(display->drm, INTEL_PCH_TYPE(i915) < PCH_MTL); + drm_WARN_ON(display->drm, INTEL_PCH_TYPE(display) < PCH_MTL); pica_ier = intel_de_rmw(display, PICAINTERRUPT_IER, ~0, 0); *pica_iir = intel_de_read(display, PICAINTERRUPT_IIR); @@ -1340,7 +1330,6 @@ static void gen8_read_and_ack_pch_irqs(struct intel_display *display, u32 *pch_i void gen8_de_irq_handler(struct intel_display *display, u32 master_ctl) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 iir; enum pipe pipe; @@ -1465,7 +1454,7 @@ void gen8_de_irq_handler(struct intel_display *display, u32 master_ctl) pipe, fault_errors); } - if (HAS_PCH_SPLIT(dev_priv) && !HAS_PCH_NOP(dev_priv) && + if (HAS_PCH_SPLIT(display) && !HAS_PCH_NOP(display) && master_ctl & GEN8_DE_PCH_IRQ) { u32 pica_iir; @@ -1479,9 +1468,9 @@ void gen8_de_irq_handler(struct intel_display *display, u32 master_ctl) if (pica_iir) xelpdp_pica_irq_handler(display, pica_iir); - if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) icp_irq_handler(display, iir); - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_SPT) + else if (INTEL_PCH_TYPE(display) >= PCH_SPT) spt_irq_handler(display, iir); else cpt_irq_handler(display, iir); @@ -1573,13 +1562,12 @@ void i915gm_irq_cstate_wa(struct intel_display *display, bool enable) int i8xx_enable_vblank(struct drm_crtc *crtc) { struct intel_display *display = to_intel_display(crtc->dev); - struct drm_i915_private *dev_priv = to_i915(crtc->dev); enum pipe pipe = to_intel_crtc(crtc)->pipe; unsigned long irqflags; - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); i915_enable_pipestat(display, pipe, PIPE_VBLANK_INTERRUPT_STATUS); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); return 0; } @@ -1587,13 +1575,12 @@ int i8xx_enable_vblank(struct drm_crtc *crtc) void i8xx_disable_vblank(struct drm_crtc *crtc) { struct intel_display *display = to_intel_display(crtc->dev); - struct drm_i915_private *dev_priv = to_i915(crtc->dev); enum pipe pipe = to_intel_crtc(crtc)->pipe; unsigned long irqflags; - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); i915_disable_pipestat(display, pipe, PIPE_VBLANK_INTERRUPT_STATUS); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); } int i915gm_enable_vblank(struct drm_crtc *crtc) @@ -1617,14 +1604,13 @@ void i915gm_disable_vblank(struct drm_crtc *crtc) int i965_enable_vblank(struct drm_crtc *crtc) { struct intel_display *display = to_intel_display(crtc->dev); - struct drm_i915_private *dev_priv = to_i915(crtc->dev); enum pipe pipe = to_intel_crtc(crtc)->pipe; unsigned long irqflags; - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); i915_enable_pipestat(display, pipe, PIPE_START_VBLANK_INTERRUPT_STATUS); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); return 0; } @@ -1632,28 +1618,26 @@ int i965_enable_vblank(struct drm_crtc *crtc) void i965_disable_vblank(struct drm_crtc *crtc) { struct intel_display *display = to_intel_display(crtc->dev); - struct drm_i915_private *dev_priv = to_i915(crtc->dev); enum pipe pipe = to_intel_crtc(crtc)->pipe; unsigned long irqflags; - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); i915_disable_pipestat(display, pipe, PIPE_START_VBLANK_INTERRUPT_STATUS); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); } int ilk_enable_vblank(struct drm_crtc *crtc) { struct intel_display *display = to_intel_display(crtc->dev); - struct drm_i915_private *dev_priv = to_i915(crtc->dev); enum pipe pipe = to_intel_crtc(crtc)->pipe; unsigned long irqflags; u32 bit = DISPLAY_VER(display) >= 7 ? DE_PIPE_VBLANK_IVB(pipe) : DE_PIPE_VBLANK(pipe); - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); ilk_enable_display_irq(display, bit); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); /* Even though there is no DMC, frame counter can get stuck when * PSR is active as no frames are generated. @@ -1667,15 +1651,14 @@ int ilk_enable_vblank(struct drm_crtc *crtc) void ilk_disable_vblank(struct drm_crtc *crtc) { struct intel_display *display = to_intel_display(crtc->dev); - struct drm_i915_private *dev_priv = to_i915(crtc->dev); enum pipe pipe = to_intel_crtc(crtc)->pipe; unsigned long irqflags; u32 bit = DISPLAY_VER(display) >= 7 ? DE_PIPE_VBLANK_IVB(pipe) : DE_PIPE_VBLANK(pipe); - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); ilk_disable_display_irq(display, bit); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); } static bool gen11_dsi_configure_te(struct intel_crtc *intel_crtc, @@ -1701,39 +1684,31 @@ static bool gen11_dsi_configure_te(struct intel_crtc *intel_crtc, return true; } -static void intel_display_vblank_dc_work(struct work_struct *work) +static void intel_display_vblank_notify_work(struct work_struct *work) { struct intel_display *display = - container_of(work, typeof(*display), irq.vblank_dc_work); - int vblank_wa_num_pipes = READ_ONCE(display->irq.vblank_wa_num_pipes); + container_of(work, typeof(*display), irq.vblank_notify_work); + int vblank_enable_count = READ_ONCE(display->irq.vblank_enable_count); - /* - * NOTE: intel_display_power_set_target_dc_state is used only by PSR - * code for DC3CO handling. DC3CO target state is currently disabled in - * PSR code. If DC3CO is taken into use we need take that into account - * here as well. - */ - intel_display_power_set_target_dc_state(display, vblank_wa_num_pipes ? DC_STATE_DISABLE : - DC_STATE_EN_UPTO_DC6); + intel_psr_notify_vblank_enable_disable(display, vblank_enable_count); } int bdw_enable_vblank(struct drm_crtc *_crtc) { struct intel_crtc *crtc = to_intel_crtc(_crtc); struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; unsigned long irqflags; if (gen11_dsi_configure_te(crtc, true)) return 0; - if (crtc->block_dc_for_vblank && display->irq.vblank_wa_num_pipes++ == 0) - schedule_work(&display->irq.vblank_dc_work); + if (crtc->vblank_psr_notify && display->irq.vblank_enable_count++ == 0) + schedule_work(&display->irq.vblank_notify_work); - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); bdw_enable_pipe_irq(display, pipe, GEN8_PIPE_VBLANK); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); /* Even if there is no DMC, frame counter can get stuck when * PSR is active as no frames are generated, so check only for PSR. @@ -1748,19 +1723,18 @@ void bdw_disable_vblank(struct drm_crtc *_crtc) { struct intel_crtc *crtc = to_intel_crtc(_crtc); struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; unsigned long irqflags; if (gen11_dsi_configure_te(crtc, false)) return; - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + spin_lock_irqsave(&display->irq.lock, irqflags); bdw_disable_pipe_irq(display, pipe, GEN8_PIPE_VBLANK); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + spin_unlock_irqrestore(&display->irq.lock, irqflags); - if (crtc->block_dc_for_vblank && --display->irq.vblank_wa_num_pipes == 0) - schedule_work(&display->irq.vblank_dc_work); + if (crtc->vblank_psr_notify && --display->irq.vblank_enable_count == 0) + schedule_work(&display->irq.vblank_notify_work); } static u32 vlv_dpinvgtt_pipe_fault_mask(enum pipe pipe) @@ -1894,8 +1868,10 @@ static void _vlv_display_irq_reset(struct intel_display *display) void vlv_display_irq_reset(struct intel_display *display) { + spin_lock_irq(&display->irq.lock); if (display->irq.vlv_display_irqs_enabled) _vlv_display_irq_reset(display); + spin_unlock_irq(&display->irq.lock); } void i9xx_display_irq_reset(struct intel_display *display) @@ -1908,22 +1884,48 @@ void i9xx_display_irq_reset(struct intel_display *display) i9xx_pipestat_irq_reset(display); } +void i915_display_irq_postinstall(struct intel_display *display) +{ + /* + * Interrupt setup is already guaranteed to be single-threaded, this is + * just to make the assert_spin_locked check happy. + */ + spin_lock_irq(&display->irq.lock); + i915_enable_pipestat(display, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS); + i915_enable_pipestat(display, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS); + spin_unlock_irq(&display->irq.lock); + + i915_enable_asle_pipestat(display); +} + +void i965_display_irq_postinstall(struct intel_display *display) +{ + /* + * Interrupt setup is already guaranteed to be single-threaded, this is + * just to make the assert_spin_locked check happy. + */ + spin_lock_irq(&display->irq.lock); + i915_enable_pipestat(display, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS); + i915_enable_pipestat(display, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS); + i915_enable_pipestat(display, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS); + spin_unlock_irq(&display->irq.lock); + + i915_enable_asle_pipestat(display); +} + static u32 vlv_error_mask(void) { /* TODO enable other errors too? */ return VLV_ERROR_PAGE_TABLE; } -void vlv_display_irq_postinstall(struct intel_display *display) +static void _vlv_display_irq_postinstall(struct intel_display *display) { struct drm_i915_private *dev_priv = to_i915(display->drm); u32 pipestat_mask; u32 enable_mask; enum pipe pipe; - if (!display->irq.vlv_display_irqs_enabled) - return; - if (display->platform.cherryview) intel_de_write(display, DPINVGTT, DPINVGTT_STATUS_MASK_CHV | @@ -1960,8 +1962,30 @@ void vlv_display_irq_postinstall(struct intel_display *display) intel_display_irq_regs_init(display, VLV_IRQ_REGS, dev_priv->irq_mask, enable_mask); } +void vlv_display_irq_postinstall(struct intel_display *display) +{ + spin_lock_irq(&display->irq.lock); + if (display->irq.vlv_display_irqs_enabled) + _vlv_display_irq_postinstall(display); + spin_unlock_irq(&display->irq.lock); +} + +void ibx_display_irq_reset(struct intel_display *display) +{ + struct drm_i915_private *i915 = to_i915(display->drm); + + if (HAS_PCH_NOP(i915)) + return; + + gen2_irq_reset(to_intel_uncore(display->drm), SDE_IRQ_REGS); + + if (HAS_PCH_CPT(i915) || HAS_PCH_LPT(i915)) + intel_de_write(display, SERR_INT, 0xffffffff); +} + void gen8_display_irq_reset(struct intel_display *display) { + struct drm_i915_private *i915 = to_i915(display->drm); enum pipe pipe; if (!HAS_DISPLAY(display)) @@ -1977,11 +2001,13 @@ void gen8_display_irq_reset(struct intel_display *display) intel_display_irq_regs_reset(display, GEN8_DE_PORT_IRQ_REGS); intel_display_irq_regs_reset(display, GEN8_DE_MISC_IRQ_REGS); + + if (HAS_PCH_SPLIT(i915)) + ibx_display_irq_reset(display); } void gen11_display_irq_reset(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe; u32 trans_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | BIT(TRANSCODER_C) | BIT(TRANSCODER_D); @@ -2026,7 +2052,7 @@ void gen11_display_irq_reset(struct intel_display *display) else intel_display_irq_regs_reset(display, GEN11_DE_HPD_IRQ_REGS); - if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) intel_display_irq_regs_reset(display, SDE_IRQ_REGS); } @@ -2038,10 +2064,10 @@ void gen8_irq_power_well_post_enable(struct intel_display *display, gen8_de_pipe_flip_done_mask(display); enum pipe pipe; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); if (!intel_irqs_enabled(dev_priv)) { - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); return; } @@ -2050,7 +2076,7 @@ void gen8_irq_power_well_post_enable(struct intel_display *display, display->irq.de_irq_mask[pipe], ~display->irq.de_irq_mask[pipe] | extra_ier); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } void gen8_irq_power_well_pre_disable(struct intel_display *display, @@ -2059,17 +2085,17 @@ void gen8_irq_power_well_pre_disable(struct intel_display *display, struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); if (!intel_irqs_enabled(dev_priv)) { - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); return; } for_each_pipe_masked(display, pipe, pipe_mask) intel_display_irq_regs_reset(display, GEN8_DE_PIPE_IRQ_REGS(pipe)); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); /* make sure we're done processing display irqs */ intel_synchronize_irq(dev_priv); @@ -2088,15 +2114,14 @@ void gen8_irq_power_well_pre_disable(struct intel_display *display, */ static void ibx_irq_postinstall(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 mask; - if (HAS_PCH_NOP(dev_priv)) + if (HAS_PCH_NOP(display)) return; - if (HAS_PCH_IBX(dev_priv)) + if (HAS_PCH_IBX(display)) mask = SDE_GMBUS | SDE_AUX_MASK | SDE_POISON; - else if (HAS_PCH_CPT(dev_priv) || HAS_PCH_LPT(dev_priv)) + else if (HAS_PCH_CPT(display) || HAS_PCH_LPT(display)) mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT; else mask = SDE_GMBUS_CPT; @@ -2108,32 +2133,37 @@ void valleyview_enable_display_irqs(struct intel_display *display) { struct drm_i915_private *dev_priv = to_i915(display->drm); - lockdep_assert_held(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); if (display->irq.vlv_display_irqs_enabled) - return; + goto out; display->irq.vlv_display_irqs_enabled = true; if (intel_irqs_enabled(dev_priv)) { _vlv_display_irq_reset(display); - vlv_display_irq_postinstall(display); + _vlv_display_irq_postinstall(display); } + +out: + spin_unlock_irq(&display->irq.lock); } void valleyview_disable_display_irqs(struct intel_display *display) { struct drm_i915_private *dev_priv = to_i915(display->drm); - lockdep_assert_held(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); if (!display->irq.vlv_display_irqs_enabled) - return; + goto out; display->irq.vlv_display_irqs_enabled = false; if (intel_irqs_enabled(dev_priv)) _vlv_display_irq_reset(display); +out: + spin_unlock_irq(&display->irq.lock); } void ilk_de_irq_postinstall(struct intel_display *display) @@ -2184,8 +2214,6 @@ static void icp_irq_postinstall(struct intel_display *display); void gen8_de_irq_postinstall(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - u32 de_pipe_masked = gen8_de_pipe_fault_mask(display) | GEN8_PIPE_CDCLK_CRC_DONE; u32 de_pipe_enables; @@ -2201,9 +2229,9 @@ void gen8_de_irq_postinstall(struct intel_display *display) if (DISPLAY_VER(display) >= 14) mtp_irq_postinstall(display); - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + else if (INTEL_PCH_TYPE(display) >= PCH_ICP) icp_irq_postinstall(display); - else if (HAS_PCH_SPLIT(dev_priv)) + else if (HAS_PCH_SPLIT(display)) ibx_irq_postinstall(display); if (DISPLAY_VER(display) < 11) @@ -2323,9 +2351,40 @@ void dg1_de_irq_postinstall(struct intel_display *display) void intel_display_irq_init(struct intel_display *display) { + spin_lock_init(&display->irq.lock); + display->drm->vblank_disable_immediate = true; intel_hotplug_irq_init(display); - INIT_WORK(&display->irq.vblank_dc_work, intel_display_vblank_dc_work); + INIT_WORK(&display->irq.vblank_notify_work, + intel_display_vblank_notify_work); +} + +struct intel_display_irq_snapshot { + u32 derrmr; +}; + +struct intel_display_irq_snapshot * +intel_display_irq_snapshot_capture(struct intel_display *display) +{ + struct intel_display_irq_snapshot *snapshot; + + snapshot = kzalloc(sizeof(*snapshot), GFP_ATOMIC); + if (!snapshot) + return NULL; + + if (DISPLAY_VER(display) >= 6 && DISPLAY_VER(display) < 20 && !HAS_GMCH(display)) + snapshot->derrmr = intel_de_read(display, DERRMR); + + return snapshot; +} + +void intel_display_irq_snapshot_print(const struct intel_display_irq_snapshot *snapshot, + struct drm_printer *p) +{ + if (!snapshot) + return; + + drm_printf(p, "DERRMR: 0x%08x\n", snapshot->derrmr); } diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.h b/drivers/gpu/drm/i915/display/intel_display_irq.h index f72727768351..c66db3851da4 100644 --- a/drivers/gpu/drm/i915/display/intel_display_irq.h +++ b/drivers/gpu/drm/i915/display/intel_display_irq.h @@ -12,7 +12,9 @@ enum pipe; struct drm_crtc; +struct drm_printer; struct intel_display; +struct intel_display_irq_snapshot; void valleyview_enable_display_irqs(struct intel_display *display); void valleyview_disable_display_irqs(struct intel_display *display); @@ -54,10 +56,13 @@ u32 gen11_gu_misc_irq_ack(struct intel_display *display, const u32 master_ctl); void gen11_gu_misc_irq_handler(struct intel_display *display, const u32 iir); void i9xx_display_irq_reset(struct intel_display *display); +void ibx_display_irq_reset(struct intel_display *display); void vlv_display_irq_reset(struct intel_display *display); void gen8_display_irq_reset(struct intel_display *display); void gen11_display_irq_reset(struct intel_display *display); +void i915_display_irq_postinstall(struct intel_display *display); +void i965_display_irq_postinstall(struct intel_display *display); void vlv_display_irq_postinstall(struct intel_display *display); void ilk_de_irq_postinstall(struct intel_display *display); void gen8_de_irq_postinstall(struct intel_display *display); @@ -67,7 +72,6 @@ void dg1_de_irq_postinstall(struct intel_display *display); u32 i915_pipestat_enable_mask(struct intel_display *display, enum pipe pipe); void i915_enable_pipestat(struct intel_display *display, enum pipe pipe, u32 status_mask); void i915_disable_pipestat(struct intel_display *display, enum pipe pipe, u32 status_mask); -void i915_enable_asle_pipestat(struct intel_display *display); void i9xx_pipestat_irq_ack(struct intel_display *display, u32 iir, u32 pipe_stats[I915_MAX_PIPES]); @@ -82,4 +86,7 @@ void intel_display_irq_init(struct intel_display *display); void i915gm_irq_cstate_wa(struct intel_display *display, bool enable); +struct intel_display_irq_snapshot *intel_display_irq_snapshot_capture(struct intel_display *display); +void intel_display_irq_snapshot_print(const struct intel_display_irq_snapshot *snapshot, struct drm_printer *p); + #endif /* __INTEL_DISPLAY_IRQ_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c index c78315eb44fa..16356523816f 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.c +++ b/drivers/gpu/drm/i915/display/intel_display_power.c @@ -323,6 +323,35 @@ unlock: mutex_unlock(&power_domains->lock); } +/** + * intel_display_power_get_current_dc_state - Set target dc state. + * @display: display device + * + * This function set the "DC off" power well target_dc_state, + * based upon this target_dc_stste, "DC off" power well will + * enable desired DC state. + */ +u32 intel_display_power_get_current_dc_state(struct intel_display *display) +{ + struct i915_power_well *power_well; + struct i915_power_domains *power_domains = &display->power.domains; + u32 current_dc_state = DC_STATE_DISABLE; + + mutex_lock(&power_domains->lock); + power_well = lookup_power_well(display, SKL_DISP_DC_OFF); + + if (drm_WARN_ON(display->drm, !power_well)) + goto unlock; + + current_dc_state = intel_power_well_is_enabled(display, power_well) ? + DC_STATE_DISABLE : power_domains->target_dc_state; + +unlock: + mutex_unlock(&power_domains->lock); + + return current_dc_state; +} + static void __async_put_domains_mask(struct i915_power_domains *power_domains, struct intel_power_domain_mask *mask) { @@ -1365,11 +1394,9 @@ static void hsw_restore_lcpll(struct intel_display *display) */ static void hsw_enable_pc8(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - drm_dbg_kms(display->drm, "Enabling package C8+\n"); - if (HAS_PCH_LPT_LP(dev_priv)) + if (HAS_PCH_LPT_LP(display)) intel_de_rmw(display, SOUTH_DSPCLK_GATE_D, PCH_LP_PARTITION_LEVEL_DISABLE, 0); @@ -1415,14 +1442,13 @@ static void intel_pch_reset_handshake(struct intel_display *display, static void skl_display_core_init(struct intel_display *display, bool resume) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct i915_power_domains *power_domains = &display->power.domains; struct i915_power_well *well; gen9_set_dc_state(display, DC_STATE_DISABLE); /* enable PCH reset handshake */ - intel_pch_reset_handshake(display, !HAS_PCH_NOP(dev_priv)); + intel_pch_reset_handshake(display, !HAS_PCH_NOP(display)); if (!HAS_DISPLAY(display)) return; @@ -1624,20 +1650,19 @@ static void tgl_bw_buddy_init(struct intel_display *display) static void icl_display_core_init(struct intel_display *display, bool resume) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct i915_power_domains *power_domains = &display->power.domains; struct i915_power_well *well; gen9_set_dc_state(display, DC_STATE_DISABLE); /* Wa_14011294188:ehl,jsl,tgl,rkl,adl-s */ - if (INTEL_PCH_TYPE(dev_priv) >= PCH_TGP && - INTEL_PCH_TYPE(dev_priv) < PCH_DG1) + if (INTEL_PCH_TYPE(display) >= PCH_TGP && + INTEL_PCH_TYPE(display) < PCH_DG1) intel_de_rmw(display, SOUTH_DSPCLK_GATE_D, 0, PCH_DPMGUNIT_CLOCK_GATE_DISABLE); /* 1. Enable PCH reset handshake. */ - intel_pch_reset_handshake(display, !HAS_PCH_NOP(dev_priv)); + intel_pch_reset_handshake(display, !HAS_PCH_NOP(display)); if (!HAS_DISPLAY(display)) return; @@ -1908,7 +1933,6 @@ static void intel_power_domains_verify_state(struct intel_display *display); */ void intel_power_domains_init_hw(struct intel_display *display, bool resume) { - struct drm_i915_private *i915 = to_i915(display->drm); struct i915_power_domains *power_domains = &display->power.domains; power_domains->initializing = true; @@ -1932,9 +1956,9 @@ void intel_power_domains_init_hw(struct intel_display *display, bool resume) assert_isp_power_gated(display); } else if (display->platform.broadwell || display->platform.haswell) { hsw_assert_cdclk(display); - intel_pch_reset_handshake(display, !HAS_PCH_NOP(i915)); + intel_pch_reset_handshake(display, !HAS_PCH_NOP(display)); } else if (display->platform.ivybridge) { - intel_pch_reset_handshake(display, !HAS_PCH_NOP(i915)); + intel_pch_reset_handshake(display, !HAS_PCH_NOP(display)); } /* @@ -2229,8 +2253,6 @@ static void intel_power_domains_verify_state(struct intel_display *display) void intel_display_power_suspend_late(struct intel_display *display, bool s2idle) { - struct drm_i915_private *i915 = to_i915(display->drm); - intel_power_domains_suspend(display, s2idle); if (DISPLAY_VER(display) >= 11 || display->platform.geminilake || @@ -2241,14 +2263,12 @@ void intel_display_power_suspend_late(struct intel_display *display, bool s2idle } /* Tweaked Wa_14010685332:cnp,icp,jsp,mcc,tgp,adp */ - if (INTEL_PCH_TYPE(i915) >= PCH_CNP && INTEL_PCH_TYPE(i915) < PCH_DG1) - intel_de_rmw(i915, SOUTH_CHICKEN1, SBCLK_RUN_REFCLK_DIS, SBCLK_RUN_REFCLK_DIS); + if (INTEL_PCH_TYPE(display) >= PCH_CNP && INTEL_PCH_TYPE(display) < PCH_DG1) + intel_de_rmw(display, SOUTH_CHICKEN1, SBCLK_RUN_REFCLK_DIS, SBCLK_RUN_REFCLK_DIS); } void intel_display_power_resume_early(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - if (DISPLAY_VER(display) >= 11 || display->platform.geminilake || display->platform.broxton) { gen9_sanitize_dc_state(display); @@ -2258,8 +2278,8 @@ void intel_display_power_resume_early(struct intel_display *display) } /* Tweaked Wa_14010685332:cnp,icp,jsp,mcc,tgp,adp */ - if (INTEL_PCH_TYPE(i915) >= PCH_CNP && INTEL_PCH_TYPE(i915) < PCH_DG1) - intel_de_rmw(i915, SOUTH_CHICKEN1, SBCLK_RUN_REFCLK_DIS, 0); + if (INTEL_PCH_TYPE(display) >= PCH_CNP && INTEL_PCH_TYPE(display) < PCH_DG1) + intel_de_rmw(display, SOUTH_CHICKEN1, SBCLK_RUN_REFCLK_DIS, 0); intel_power_domains_resume(display); } diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h index 1b53d67f9b60..f8813b0e16df 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.h +++ b/drivers/gpu/drm/i915/display/intel_display_power.h @@ -183,6 +183,7 @@ void intel_display_power_suspend(struct intel_display *display); void intel_display_power_resume(struct intel_display *display); void intel_display_power_set_target_dc_state(struct intel_display *display, u32 state); +u32 intel_display_power_get_current_dc_state(struct intel_display *display); bool intel_display_power_is_enabled(struct intel_display *display, enum intel_display_power_domain domain); diff --git a/drivers/gpu/drm/i915/display/intel_display_power_well.c b/drivers/gpu/drm/i915/display/intel_display_power_well.c index b9b4359751cc..b104bce0e14d 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power_well.c +++ b/drivers/gpu/drm/i915/display/intel_display_power_well.c @@ -25,6 +25,7 @@ #include "intel_hotplug.h" #include "intel_pcode.h" #include "intel_pps.h" +#include "intel_psr.h" #include "intel_tc.h" #include "intel_vga.h" #include "skl_watermark.h" @@ -760,6 +761,9 @@ void gen9_set_dc_state(struct intel_display *display, u32 state) state & ~power_domains->allowed_dc_mask)) state &= power_domains->allowed_dc_mask; + if (!power_domains->initializing) + intel_psr_notify_dc5_dc6(display); + val = intel_de_read(display, DC_STATE_EN); mask = gen9_dc_mask(display); drm_dbg_kms(display->drm, "Setting DC state from %02x to %02x\n", @@ -1208,7 +1212,6 @@ static void vlv_init_display_clock_gating(struct intel_display *display) static void vlv_display_power_well_init(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder; enum pipe pipe; @@ -1232,9 +1235,7 @@ static void vlv_display_power_well_init(struct intel_display *display) vlv_init_display_clock_gating(display); - spin_lock_irq(&dev_priv->irq_lock); valleyview_enable_display_irqs(display); - spin_unlock_irq(&dev_priv->irq_lock); /* * During driver initialization/resume we can avoid restoring the @@ -1252,7 +1253,7 @@ static void vlv_display_power_well_init(struct intel_display *display) intel_crt_reset(&encoder->base); } - intel_vga_redisable_power_on(display); + intel_vga_disable(display); intel_pps_unlock_regs_wa(display); } @@ -1261,9 +1262,7 @@ static void vlv_display_power_well_deinit(struct intel_display *display) { struct drm_i915_private *dev_priv = to_i915(display->drm); - spin_lock_irq(&dev_priv->irq_lock); valleyview_disable_display_irqs(display); - spin_unlock_irq(&dev_priv->irq_lock); /* make sure we're done processing display irqs */ intel_synchronize_irq(dev_priv); diff --git a/drivers/gpu/drm/i915/display/intel_display_rps.c b/drivers/gpu/drm/i915/display/intel_display_rps.c index 4074a1879828..678b24115951 100644 --- a/drivers/gpu/drm/i915/display/intel_display_rps.c +++ b/drivers/gpu/drm/i915/display/intel_display_rps.c @@ -8,6 +8,8 @@ #include "gt/intel_rps.h" #include "i915_drv.h" +#include "i915_reg.h" +#include "intel_display_irq.h" #include "intel_display_rps.h" #include "intel_display_types.h" @@ -81,3 +83,24 @@ void intel_display_rps_mark_interactive(struct intel_display *display, intel_rps_mark_interactive(&to_gt(i915)->rps, interactive); state->rps_interactive = interactive; } + +void ilk_display_rps_enable(struct intel_display *display) +{ + spin_lock(&display->irq.lock); + ilk_enable_display_irq(display, DE_PCU_EVENT); + spin_unlock(&display->irq.lock); +} + +void ilk_display_rps_disable(struct intel_display *display) +{ + spin_lock(&display->irq.lock); + ilk_disable_display_irq(display, DE_PCU_EVENT); + spin_unlock(&display->irq.lock); +} + +void ilk_display_rps_irq_handler(struct intel_display *display) +{ + struct drm_i915_private *i915 = to_i915(display->drm); + + gen5_rps_irq_handler(&to_gt(i915)->rps); +} diff --git a/drivers/gpu/drm/i915/display/intel_display_rps.h b/drivers/gpu/drm/i915/display/intel_display_rps.h index 556891edb2dd..183d154f2c7c 100644 --- a/drivers/gpu/drm/i915/display/intel_display_rps.h +++ b/drivers/gpu/drm/i915/display/intel_display_rps.h @@ -13,10 +13,34 @@ struct drm_crtc; struct intel_atomic_state; struct intel_display; +#ifdef I915 void intel_display_rps_boost_after_vblank(struct drm_crtc *crtc, struct dma_fence *fence); void intel_display_rps_mark_interactive(struct intel_display *display, struct intel_atomic_state *state, bool interactive); +void ilk_display_rps_enable(struct intel_display *display); +void ilk_display_rps_disable(struct intel_display *display); +void ilk_display_rps_irq_handler(struct intel_display *display); +#else +static inline void intel_display_rps_boost_after_vblank(struct drm_crtc *crtc, + struct dma_fence *fence) +{ +} +static inline void intel_display_rps_mark_interactive(struct intel_display *display, + struct intel_atomic_state *state, + bool interactive) +{ +} +static inline void ilk_display_rps_enable(struct intel_display *display) +{ +} +static inline void ilk_display_rps_disable(struct intel_display *display) +{ +} +static inline void ilk_display_rps_irq_handler(struct intel_display *display) +{ +} +#endif #endif /* __INTEL_DISPLAY_RPS_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_display_snapshot.c b/drivers/gpu/drm/i915/display/intel_display_snapshot.c index 25ba043cbb65..66087302fdbc 100644 --- a/drivers/gpu/drm/i915/display/intel_display_snapshot.c +++ b/drivers/gpu/drm/i915/display/intel_display_snapshot.c @@ -7,6 +7,7 @@ #include "intel_display_core.h" #include "intel_display_device.h" +#include "intel_display_irq.h" #include "intel_display_params.h" #include "intel_display_snapshot.h" #include "intel_dmc.h" @@ -20,6 +21,7 @@ struct intel_display_snapshot { struct intel_display_params params; struct intel_overlay_snapshot *overlay; struct intel_dmc_snapshot *dmc; + struct intel_display_irq_snapshot *irq; }; struct intel_display_snapshot *intel_display_snapshot_capture(struct intel_display *display) @@ -38,6 +40,7 @@ struct intel_display_snapshot *intel_display_snapshot_capture(struct intel_displ intel_display_params_copy(&snapshot->params); + snapshot->irq = intel_display_irq_snapshot_capture(display); snapshot->overlay = intel_overlay_snapshot_capture(display); snapshot->dmc = intel_dmc_snapshot_capture(display); @@ -57,6 +60,7 @@ void intel_display_snapshot_print(const struct intel_display_snapshot *snapshot, intel_display_device_info_print(&snapshot->info, &snapshot->runtime_info, p); intel_display_params_dump(&snapshot->params, display->drm->driver->name, p); + intel_display_irq_snapshot_print(snapshot->irq, p); intel_overlay_snapshot_print(snapshot->overlay, p); intel_dmc_snapshot_print(snapshot->dmc, p); } @@ -68,6 +72,7 @@ void intel_display_snapshot_free(struct intel_display_snapshot *snapshot) intel_display_params_free(&snapshot->params); + kfree(snapshot->irq); kfree(snapshot->overlay); kfree(snapshot->dmc); kfree(snapshot); diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 94468a9d2e0d..d6d0440dcee9 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -1114,6 +1114,7 @@ struct intel_crtc_state { bool wm_level_disabled; u32 dc3co_exitline; u16 su_y_granularity; + u8 active_non_psr_pipes; /* * Frequency the dpll for the port should run at. Differs from the @@ -1387,7 +1388,7 @@ struct intel_crtc { /* armed event for DSB based updates */ struct drm_pending_vblank_event *dsb_event; - /* Access to these should be protected by dev_priv->irq_lock. */ + /* Access to these should be protected by display->irq.lock. */ bool cpu_fifo_underrun_disabled; bool pch_fifo_underrun_disabled; @@ -1439,7 +1440,7 @@ struct intel_crtc { struct intel_pipe_crc pipe_crc; #endif - bool block_dc_for_vblank; + bool vblank_psr_notify; }; struct intel_plane_error { @@ -1650,6 +1651,8 @@ struct intel_psr { u8 entry_setup_frames; bool link_ok; + + u8 active_non_psr_pipes; }; struct intel_dp { @@ -1805,12 +1808,16 @@ struct intel_dp { struct { u8 io_wake_lines; u8 fast_wake_lines; + enum transcoder transcoder; + struct mutex lock; /* LNL and beyond */ u8 check_entry_lines; u8 aux_less_wake_lines; u8 silence_period_sym_clocks; u8 lfps_half_cycle_num_of_syms; + bool lobf_disable_debug; + bool sink_alpm_error; } alpm_parameters; u8 alpm_dpcd; diff --git a/drivers/gpu/drm/i915/display/intel_dmc.c b/drivers/gpu/drm/i915/display/intel_dmc.c index 98f80a6c63e8..b58189d24e7e 100644 --- a/drivers/gpu/drm/i915/display/intel_dmc.c +++ b/drivers/gpu/drm/i915/display/intel_dmc.c @@ -173,7 +173,6 @@ MODULE_FIRMWARE(BXT_DMC_PATH); static const char *dmc_firmware_default(struct intel_display *display, u32 *size) { - struct drm_i915_private *i915 = to_i915(display->drm); const char *fw_path = NULL; u32 max_fw_size = 0; @@ -189,39 +188,39 @@ static const char *dmc_firmware_default(struct intel_display *display, u32 *size } else if (DISPLAY_VERx100(display) == 1400) { fw_path = MTL_DMC_PATH; max_fw_size = XELPDP_DMC_MAX_FW_SIZE; - } else if (IS_DG2(i915)) { + } else if (display->platform.dg2) { fw_path = DG2_DMC_PATH; max_fw_size = DISPLAY_VER13_DMC_MAX_FW_SIZE; - } else if (IS_ALDERLAKE_P(i915)) { + } else if (display->platform.alderlake_p) { fw_path = ADLP_DMC_PATH; max_fw_size = DISPLAY_VER13_DMC_MAX_FW_SIZE; - } else if (IS_ALDERLAKE_S(i915)) { + } else if (display->platform.alderlake_s) { fw_path = ADLS_DMC_PATH; max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE; - } else if (IS_DG1(i915)) { + } else if (display->platform.dg1) { fw_path = DG1_DMC_PATH; max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE; - } else if (IS_ROCKETLAKE(i915)) { + } else if (display->platform.rocketlake) { fw_path = RKL_DMC_PATH; max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE; - } else if (IS_TIGERLAKE(i915)) { + } else if (display->platform.tigerlake) { fw_path = TGL_DMC_PATH; max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE; } else if (DISPLAY_VER(display) == 11) { fw_path = ICL_DMC_PATH; max_fw_size = ICL_DMC_MAX_FW_SIZE; - } else if (IS_GEMINILAKE(i915)) { + } else if (display->platform.geminilake) { fw_path = GLK_DMC_PATH; max_fw_size = GLK_DMC_MAX_FW_SIZE; - } else if (IS_KABYLAKE(i915) || - IS_COFFEELAKE(i915) || - IS_COMETLAKE(i915)) { + } else if (display->platform.kabylake || + display->platform.coffeelake || + display->platform.cometlake) { fw_path = KBL_DMC_PATH; max_fw_size = KBL_DMC_MAX_FW_SIZE; - } else if (IS_SKYLAKE(i915)) { + } else if (display->platform.skylake) { fw_path = SKL_DMC_PATH; max_fw_size = SKL_DMC_MAX_FW_SIZE; - } else if (IS_BROXTON(i915)) { + } else if (display->platform.broxton) { fw_path = BXT_DMC_PATH; max_fw_size = BXT_DMC_MAX_FW_SIZE; } @@ -517,6 +516,54 @@ void intel_dmc_disable_pipe(struct intel_display *display, enum pipe pipe) intel_de_rmw(display, PIPEDMC_CONTROL(pipe), PIPEDMC_ENABLE, 0); } +/** + * intel_dmc_block_pkgc() - block PKG C-state + * @display: display instance + * @pipe: pipe which register use to block + * @block: block/unblock + * + * This interface is target for Wa_16025596647 usage. I.e. to set/clear + * PIPEDMC_BLOCK_PKGC_SW_BLOCK_PKGC_ALWAYS bit in PIPEDMC_BLOCK_PKGC_SW register. + */ +void intel_dmc_block_pkgc(struct intel_display *display, enum pipe pipe, + bool block) +{ + intel_de_rmw(display, PIPEDMC_BLOCK_PKGC_SW(pipe), + PIPEDMC_BLOCK_PKGC_SW_BLOCK_PKGC_ALWAYS, block ? + PIPEDMC_BLOCK_PKGC_SW_BLOCK_PKGC_ALWAYS : 0); +} + +/** + * intel_dmc_start_pkgc_exit_at_start_of_undelayed_vblank() - start of PKG + * C-state exit + * @display: display instance + * @pipe: pipe which register use to block + * @enable: enable/disable + * + * This interface is target for Wa_16025596647 usage. I.e. start the package C + * exit at the start of the undelayed vblank + */ +void intel_dmc_start_pkgc_exit_at_start_of_undelayed_vblank(struct intel_display *display, + enum pipe pipe, bool enable) +{ + u32 val; + + if (enable) + val = DMC_EVT_CTL_ENABLE | DMC_EVT_CTL_RECURRING | + REG_FIELD_PREP(DMC_EVT_CTL_TYPE_MASK, + DMC_EVT_CTL_TYPE_EDGE_0_1) | + REG_FIELD_PREP(DMC_EVT_CTL_EVENT_ID_MASK, + DMC_EVT_CTL_EVENT_ID_VBLANK_A); + else + val = REG_FIELD_PREP(DMC_EVT_CTL_EVENT_ID_MASK, + DMC_EVT_CTL_EVENT_ID_FALSE) | + REG_FIELD_PREP(DMC_EVT_CTL_TYPE_MASK, + DMC_EVT_CTL_TYPE_EDGE_0_1); + + intel_de_write(display, MTL_PIPEDMC_EVT_CTL_4(pipe), + val); +} + static bool is_dmc_evt_ctl_reg(struct intel_display *display, enum intel_dmc_id dmc_id, i915_reg_t reg) { @@ -541,8 +588,6 @@ static bool disable_dmc_evt(struct intel_display *display, enum intel_dmc_id dmc_id, i915_reg_t reg, u32 data) { - struct drm_i915_private *i915 = to_i915(display->drm); - if (!is_dmc_evt_ctl_reg(display, dmc_id, reg)) return false; @@ -551,12 +596,12 @@ static bool disable_dmc_evt(struct intel_display *display, return true; /* also disable the flip queue event on the main DMC on TGL */ - if (IS_TIGERLAKE(i915) && + if (display->platform.tigerlake && REG_FIELD_GET(DMC_EVT_CTL_EVENT_ID_MASK, data) == DMC_EVT_CTL_EVENT_ID_CLK_MSEC) return true; /* also disable the HRR event on the main DMC on TGL/ADLS */ - if ((IS_TIGERLAKE(i915) || IS_ALDERLAKE_S(i915)) && + if ((display->platform.tigerlake || display->platform.alderlake_s) && REG_FIELD_GET(DMC_EVT_CTL_EVENT_ID_MASK, data) == DMC_EVT_CTL_EVENT_ID_VBLANK_A) return true; @@ -588,7 +633,6 @@ static u32 dmc_mmiodata(struct intel_display *display, */ void intel_dmc_load_program(struct intel_display *display) { - struct drm_i915_private *i915 __maybe_unused = to_i915(display->drm); struct i915_power_domains *power_domains = &display->power.domains; struct intel_dmc *dmc = display_to_dmc(display); enum intel_dmc_id dmc_id; @@ -1012,9 +1056,7 @@ static void intel_dmc_runtime_pm_put(struct intel_display *display) static const char *dmc_fallback_path(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - if (IS_ALDERLAKE_P(i915)) + if (display->platform.alderlake_p) return ADLP_DMC_FALLBACK_PATH; return NULL; @@ -1279,7 +1321,6 @@ static bool intel_dmc_get_dc6_allowed_count(struct intel_display *display, u32 * static int intel_dmc_debugfs_status_show(struct seq_file *m, void *unused) { struct intel_display *display = m->private; - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_dmc *dmc = display_to_dmc(display); struct ref_tracker *wakeref; i915_reg_t dc5_reg, dc6_reg = INVALID_MMIO_REG; @@ -1299,7 +1340,7 @@ static int intel_dmc_debugfs_status_show(struct seq_file *m, void *unused) seq_printf(m, "Pipe A fw loaded: %s\n", str_yes_no(has_dmc_id_fw(display, DMC_FW_PIPEA))); seq_printf(m, "Pipe B fw needed: %s\n", - str_yes_no(IS_ALDERLAKE_P(i915) || + str_yes_no(display->platform.alderlake_p || DISPLAY_VER(display) >= 14)); seq_printf(m, "Pipe B fw loaded: %s\n", str_yes_no(has_dmc_id_fw(display, DMC_FW_PIPEB))); @@ -1313,7 +1354,7 @@ static int intel_dmc_debugfs_status_show(struct seq_file *m, void *unused) if (DISPLAY_VER(display) >= 12) { i915_reg_t dc3co_reg; - if (IS_DGFX(i915) || DISPLAY_VER(display) >= 14) { + if (display->platform.dgfx || DISPLAY_VER(display) >= 14) { dc3co_reg = DG1_DMC_DEBUG3; dc5_reg = DG1_DMC_DEBUG_DC5_COUNT; } else { @@ -1325,9 +1366,9 @@ static int intel_dmc_debugfs_status_show(struct seq_file *m, void *unused) seq_printf(m, "DC3CO count: %d\n", intel_de_read(display, dc3co_reg)); } else { - dc5_reg = IS_BROXTON(i915) ? BXT_DMC_DC3_DC5_COUNT : + dc5_reg = display->platform.broxton ? BXT_DMC_DC3_DC5_COUNT : SKL_DMC_DC3_DC5_COUNT; - if (!IS_GEMINILAKE(i915) && !IS_BROXTON(i915)) + if (!display->platform.geminilake && !display->platform.broxton) dc6_reg = SKL_DMC_DC5_DC6_COUNT; } diff --git a/drivers/gpu/drm/i915/display/intel_dmc.h b/drivers/gpu/drm/i915/display/intel_dmc.h index c78426eb4cd5..bd1c459b0075 100644 --- a/drivers/gpu/drm/i915/display/intel_dmc.h +++ b/drivers/gpu/drm/i915/display/intel_dmc.h @@ -18,6 +18,10 @@ void intel_dmc_load_program(struct intel_display *display); void intel_dmc_disable_program(struct intel_display *display); void intel_dmc_enable_pipe(struct intel_display *display, enum pipe pipe); void intel_dmc_disable_pipe(struct intel_display *display, enum pipe pipe); +void intel_dmc_block_pkgc(struct intel_display *display, enum pipe pipe, + bool block); +void intel_dmc_start_pkgc_exit_at_start_of_undelayed_vblank(struct intel_display *display, + enum pipe pipe, bool enable); void intel_dmc_fini(struct intel_display *display); void intel_dmc_suspend(struct intel_display *display); void intel_dmc_resume(struct intel_display *display); diff --git a/drivers/gpu/drm/i915/display/intel_dmc_regs.h b/drivers/gpu/drm/i915/display/intel_dmc_regs.h index 1bf446f96a10..e16ea3f16ed8 100644 --- a/drivers/gpu/drm/i915/display/intel_dmc_regs.h +++ b/drivers/gpu/drm/i915/display/intel_dmc_regs.h @@ -21,6 +21,20 @@ #define MTL_PIPEDMC_CONTROL _MMIO(0x45250) #define PIPEDMC_ENABLE_MTL(pipe) REG_BIT(((pipe) - PIPE_A) * 4) +#define _MTL_PIPEDMC_EVT_CTL_4_A 0x5f044 +#define _MTL_PIPEDMC_EVT_CTL_4_B 0x5f444 +#define MTL_PIPEDMC_EVT_CTL_4(pipe) _MMIO_PIPE(pipe, \ + _MTL_PIPEDMC_EVT_CTL_4_A, \ + _MTL_PIPEDMC_EVT_CTL_4_B) + +#define PIPEDMC_BLOCK_PKGC_SW_A 0x5f1d0 +#define PIPEDMC_BLOCK_PKGC_SW_B 0x5F5d0 +#define PIPEDMC_BLOCK_PKGC_SW(pipe) _MMIO_PIPE(pipe, \ + PIPEDMC_BLOCK_PKGC_SW_A, \ + PIPEDMC_BLOCK_PKGC_SW_B) +#define PIPEDMC_BLOCK_PKGC_SW_BLOCK_PKGC_ALWAYS BIT(31) +#define PIPEDMC_BLOCK_PKGC_SW_BLOCK_PKGC_UNTIL_NEXT_FRAMESTART BIT(15) + #define _ADLP_PIPEDMC_REG_MMIO_BASE_A 0x5f000 #define _TGL_PIPEDMC_REG_MMIO_BASE_A 0x92000 diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c index d7a30d0992b7..640c43bf62d4 100644 --- a/drivers/gpu/drm/i915/display/intel_dp.c +++ b/drivers/gpu/drm/i915/display/intel_dp.c @@ -45,12 +45,13 @@ #include <drm/drm_crtc.h> #include <drm/drm_edid.h> #include <drm/drm_fixed.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> #include "g4x_dp.h" -#include "i915_drv.h" #include "i915_irq.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_alpm.h" #include "intel_atomic.h" #include "intel_audio.h" @@ -58,6 +59,7 @@ #include "intel_combo_phy_regs.h" #include "intel_connector.h" #include "intel_crtc.h" +#include "intel_crtc_state_dump.h" #include "intel_cx0_phy.h" #include "intel_ddi.h" #include "intel_de.h" @@ -92,7 +94,6 @@ #include "intel_tc.h" #include "intel_vdsc.h" #include "intel_vrr.h" -#include "intel_crtc_state_dump.h" /* DP DSC throughput values used for slice count calculations KPixels/s */ #define DP_DSC_PEAK_PIXEL_RATE 2720000 @@ -2523,6 +2524,7 @@ intel_dp_dsc_compute_pipe_bpp_limits(struct intel_dp *intel_dp, bool intel_dp_compute_config_limits(struct intel_dp *intel_dp, + struct intel_connector *connector, struct intel_crtc_state *crtc_state, bool respect_downstream_limits, bool dsc, @@ -2576,7 +2578,7 @@ intel_dp_compute_config_limits(struct intel_dp *intel_dp, intel_dp_test_compute_config(intel_dp, crtc_state, limits); return intel_dp_compute_config_link_bpp_limits(intel_dp, - intel_dp->attached_connector, + connector, crtc_state, dsc, limits); @@ -2637,7 +2639,7 @@ intel_dp_compute_link_config(struct intel_encoder *encoder, joiner_needs_dsc = intel_dp_joiner_needs_dsc(display, num_joined_pipes); dsc_needed = joiner_needs_dsc || intel_dp->force_dsc_en || - !intel_dp_compute_config_limits(intel_dp, pipe_config, + !intel_dp_compute_config_limits(intel_dp, connector, pipe_config, respect_downstream_limits, false, &limits); @@ -2671,7 +2673,7 @@ intel_dp_compute_link_config(struct intel_encoder *encoder, str_yes_no(ret), str_yes_no(joiner_needs_dsc), str_yes_no(intel_dp->force_dsc_en)); - if (!intel_dp_compute_config_limits(intel_dp, pipe_config, + if (!intel_dp_compute_config_limits(intel_dp, connector, pipe_config, respect_downstream_limits, true, &limits)) @@ -3104,6 +3106,76 @@ intel_dp_queue_modeset_retry_for_link(struct intel_atomic_state *state, } } +int intel_dp_compute_min_hblank(struct intel_crtc_state *crtc_state, + const struct drm_connector_state *conn_state) +{ + struct intel_display *display = to_intel_display(crtc_state); + const struct drm_display_mode *adjusted_mode = + &crtc_state->hw.adjusted_mode; + struct intel_connector *connector = to_intel_connector(conn_state->connector); + int symbol_size = intel_dp_is_uhbr(crtc_state) ? 32 : 8; + /* + * min symbol cycles is 3(BS,VBID, BE) for 128b/132b and + * 5(BS, VBID, MVID, MAUD, BE) for 8b/10b + */ + int min_sym_cycles = intel_dp_is_uhbr(crtc_state) ? 3 : 5; + bool is_mst = intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST); + int num_joined_pipes = intel_crtc_num_joined_pipes(crtc_state); + int min_hblank; + int max_lane_count = 4; + int hactive_sym_cycles, htotal_sym_cycles; + int dsc_slices = 0; + int link_bpp_x16; + + if (DISPLAY_VER(display) < 30) + return 0; + + /* MIN_HBLANK should be set only for 8b/10b MST or for 128b/132b SST/MST */ + if (!is_mst && !intel_dp_is_uhbr(crtc_state)) + return 0; + + if (crtc_state->dsc.compression_enable) { + dsc_slices = intel_dp_dsc_get_slice_count(connector, + adjusted_mode->crtc_clock, + adjusted_mode->crtc_hdisplay, + num_joined_pipes); + if (!dsc_slices) { + drm_dbg(display->drm, "failed to calculate dsc slice count\n"); + return -EINVAL; + } + } + + if (crtc_state->dsc.compression_enable) + link_bpp_x16 = crtc_state->dsc.compressed_bpp_x16; + else + link_bpp_x16 = fxp_q4_from_int(intel_dp_output_bpp(crtc_state->output_format, + crtc_state->pipe_bpp)); + + /* Calculate min Hblank Link Layer Symbol Cycle Count for 8b/10b MST & 128b/132b */ + hactive_sym_cycles = drm_dp_link_symbol_cycles(max_lane_count, + adjusted_mode->hdisplay, + dsc_slices, + link_bpp_x16, + symbol_size, is_mst); + htotal_sym_cycles = adjusted_mode->htotal * hactive_sym_cycles / + adjusted_mode->hdisplay; + + min_hblank = htotal_sym_cycles - hactive_sym_cycles; + /* minimum Hblank calculation: https://groups.vesa.org/wg/DP/document/20494 */ + min_hblank = max(min_hblank, min_sym_cycles); + + /* + * adjust the BlankingStart/BlankingEnd framing control from + * the calculated value + */ + min_hblank = min_hblank - 2; + + min_hblank = min(10, min_hblank); + crtc_state->min_hblank = min_hblank; + + return 0; +} + int intel_dp_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config, @@ -3203,6 +3275,10 @@ intel_dp_compute_config(struct intel_encoder *encoder, &pipe_config->dp_m_n); } + ret = intel_dp_compute_min_hblank(pipe_config, conn_state); + if (ret) + return ret; + /* FIXME: abstract this better */ if (pipe_config->splitter.enable) pipe_config->dp_m_n.data_m *= pipe_config->splitter.link_count; @@ -4456,6 +4532,23 @@ intel_dp_mst_disconnect(struct intel_dp *intel_dp) static bool intel_dp_get_sink_irq_esi(struct intel_dp *intel_dp, u8 *esi) { + struct intel_display *display = to_intel_display(intel_dp); + + /* + * Display WA for HSD #13013007775: mtl/arl/lnl + * Read the sink count and link service IRQ registers in separate + * transactions to prevent disconnecting the sink on a TBT link + * inadvertently. + */ + if (IS_DISPLAY_VER(display, 14, 20) && !display->platform.battlemage) { + if (drm_dp_dpcd_read(&intel_dp->aux, DP_SINK_COUNT_ESI, esi, 3) != 3) + return false; + + /* DP_SINK_COUNT_ESI + 3 == DP_LINK_SERVICE_IRQ_VECTOR_ESI0 */ + return drm_dp_dpcd_readb(&intel_dp->aux, DP_LINK_SERVICE_IRQ_VECTOR_ESI0, + &esi[3]) == 1; + } + return drm_dp_dpcd_read(&intel_dp->aux, DP_SINK_COUNT_ESI, esi, 4) == 4; } @@ -5392,6 +5485,11 @@ intel_dp_short_pulse(struct intel_dp *intel_dp) intel_psr_short_pulse(intel_dp); + if (intel_alpm_get_error(intel_dp)) { + intel_alpm_disable(intel_dp); + intel_dp->alpm_parameters.sink_alpm_error = true; + } + if (intel_dp_test_short_pulse(intel_dp)) reprobe_needed = true; @@ -5827,20 +5925,21 @@ out_vdd_off: } static void -intel_dp_force(struct drm_connector *connector) +intel_dp_force(struct drm_connector *_connector) { - struct intel_display *display = to_intel_display(connector->dev); - struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_dp *intel_dp = intel_attached_dp(connector); drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s]\n", - connector->base.id, connector->name); + connector->base.base.id, connector->base.name); if (!intel_display_driver_check_access(display)) return; intel_dp_unset_edid(intel_dp); - if (connector->status != connector_status_connected) + if (connector->base.status != connector_status_connected) return; intel_dp_set_edid(intel_dp); @@ -5879,24 +5978,25 @@ static int intel_dp_get_modes(struct drm_connector *_connector) } static int -intel_dp_connector_register(struct drm_connector *connector) +intel_dp_connector_register(struct drm_connector *_connector) { - struct intel_display *display = to_intel_display(connector->dev); - struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_dp *intel_dp = intel_attached_dp(connector); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); int ret; - ret = intel_connector_register(connector); + ret = intel_connector_register(&connector->base); if (ret) return ret; drm_dbg_kms(display->drm, "registering %s bus for %s\n", - intel_dp->aux.name, connector->kdev->kobj.name); + intel_dp->aux.name, connector->base.kdev->kobj.name); - intel_dp->aux.dev = connector->kdev; + intel_dp->aux.dev = connector->base.kdev; ret = drm_dp_aux_register(&intel_dp->aux); if (!ret) - drm_dp_cec_register_connector(&intel_dp->aux, connector); + drm_dp_cec_register_connector(&intel_dp->aux, &connector->base); if (!intel_bios_encoder_is_lspcon(dig_port->base.devdata)) return ret; @@ -5907,20 +6007,21 @@ intel_dp_connector_register(struct drm_connector *connector) */ if (intel_lspcon_init(dig_port)) { if (intel_lspcon_detect_hdr_capability(dig_port)) - drm_connector_attach_hdr_output_metadata_property(connector); + drm_connector_attach_hdr_output_metadata_property(&connector->base); } return ret; } static void -intel_dp_connector_unregister(struct drm_connector *connector) +intel_dp_connector_unregister(struct drm_connector *_connector) { - struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_dp *intel_dp = intel_attached_dp(connector); drm_dp_cec_unregister_connector(&intel_dp->aux); drm_dp_aux_unregister(&intel_dp->aux); - intel_connector_unregister(connector); + intel_connector_unregister(&connector->base); } void intel_dp_connector_sync_state(struct intel_connector *connector, @@ -5981,21 +6082,21 @@ static int intel_modeset_tile_group(struct intel_atomic_state *state, { struct intel_display *display = to_intel_display(state); struct drm_connector_list_iter conn_iter; - struct drm_connector *connector; + struct intel_connector *connector; int ret = 0; drm_connector_list_iter_begin(display->drm, &conn_iter); - drm_for_each_connector_iter(connector, &conn_iter) { + for_each_intel_connector_iter(connector, &conn_iter) { struct drm_connector_state *conn_state; struct intel_crtc_state *crtc_state; struct intel_crtc *crtc; - if (!connector->has_tile || - connector->tile_group->id != tile_group_id) + if (!connector->base.has_tile || + connector->base.tile_group->id != tile_group_id) continue; conn_state = drm_atomic_get_connector_state(&state->base, - connector); + &connector->base); if (IS_ERR(conn_state)) { ret = PTR_ERR(conn_state); break; @@ -6059,10 +6160,11 @@ static int intel_modeset_affected_transcoders(struct intel_atomic_state *state, } static int intel_modeset_synced_crtcs(struct intel_atomic_state *state, - struct drm_connector *connector) + struct drm_connector *_connector) { + struct intel_connector *connector = to_intel_connector(_connector); const struct drm_connector_state *old_conn_state = - drm_atomic_get_old_connector_state(&state->base, connector); + drm_atomic_get_old_connector_state(&state->base, &connector->base); const struct intel_crtc_state *old_crtc_state; struct intel_crtc *crtc; u8 transcoders; @@ -6084,17 +6186,18 @@ static int intel_modeset_synced_crtcs(struct intel_atomic_state *state, transcoders); } -static int intel_dp_connector_atomic_check(struct drm_connector *conn, +static int intel_dp_connector_atomic_check(struct drm_connector *_connector, struct drm_atomic_state *_state) { - struct intel_display *display = to_intel_display(conn->dev); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); struct intel_atomic_state *state = to_intel_atomic_state(_state); - struct drm_connector_state *conn_state = drm_atomic_get_new_connector_state(_state, conn); - struct intel_connector *intel_conn = to_intel_connector(conn); - struct intel_dp *intel_dp = enc_to_intel_dp(intel_conn->encoder); + struct drm_connector_state *conn_state = + drm_atomic_get_new_connector_state(_state, &connector->base); + struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder); int ret; - ret = intel_digital_connector_atomic_check(conn, &state->base); + ret = intel_digital_connector_atomic_check(&connector->base, &state->base); if (ret) return ret; @@ -6104,12 +6207,12 @@ static int intel_dp_connector_atomic_check(struct drm_connector *conn, return ret; } - if (!intel_connector_needs_modeset(state, conn)) + if (!intel_connector_needs_modeset(state, &connector->base)) return 0; ret = intel_dp_tunnel_atomic_check_state(state, intel_dp, - intel_conn); + connector); if (ret) return ret; @@ -6120,26 +6223,26 @@ static int intel_dp_connector_atomic_check(struct drm_connector *conn, if (DISPLAY_VER(display) < 9) return 0; - if (conn->has_tile) { - ret = intel_modeset_tile_group(state, conn->tile_group->id); + if (connector->base.has_tile) { + ret = intel_modeset_tile_group(state, connector->base.tile_group->id); if (ret) return ret; } - return intel_modeset_synced_crtcs(state, conn); + return intel_modeset_synced_crtcs(state, &connector->base); } -static void intel_dp_oob_hotplug_event(struct drm_connector *connector, +static void intel_dp_oob_hotplug_event(struct drm_connector *_connector, enum drm_connector_status hpd_state) { - struct intel_display *display = to_intel_display(connector->dev); - struct intel_encoder *encoder = intel_attached_encoder(to_intel_connector(connector)); - struct drm_i915_private *i915 = to_i915(connector->dev); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_encoder *encoder = intel_attached_encoder(connector); bool hpd_high = hpd_state == connector_status_connected; unsigned int hpd_pin = encoder->hpd_pin; bool need_work = false; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); if (hpd_high != test_bit(hpd_pin, &display->hotplug.oob_hotplug_last_state)) { display->hotplug.event_bits |= BIT(hpd_pin); @@ -6148,7 +6251,7 @@ static void intel_dp_oob_hotplug_event(struct drm_connector *connector, hpd_high); need_work = true; } - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); if (need_work) intel_hpd_schedule_detection(display); @@ -6280,36 +6383,37 @@ intel_dp_has_gamut_metadata_dip(struct intel_encoder *encoder) } static void -intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connector) +intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *_connector) { + struct intel_connector *connector = to_intel_connector(_connector); struct intel_display *display = to_intel_display(intel_dp); enum port port = dp_to_dig_port(intel_dp)->base.port; if (!intel_dp_is_edp(intel_dp)) - drm_connector_attach_dp_subconnector_property(connector); + drm_connector_attach_dp_subconnector_property(&connector->base); if (!display->platform.g4x && port != PORT_A) - intel_attach_force_audio_property(connector); + intel_attach_force_audio_property(&connector->base); - intel_attach_broadcast_rgb_property(connector); + intel_attach_broadcast_rgb_property(&connector->base); if (HAS_GMCH(display)) - drm_connector_attach_max_bpc_property(connector, 6, 10); + drm_connector_attach_max_bpc_property(&connector->base, 6, 10); else if (DISPLAY_VER(display) >= 5) - drm_connector_attach_max_bpc_property(connector, 6, 12); + drm_connector_attach_max_bpc_property(&connector->base, 6, 12); /* Register HDMI colorspace for case of lspcon */ if (intel_bios_encoder_is_lspcon(dp_to_dig_port(intel_dp)->base.devdata)) { - drm_connector_attach_content_type_property(connector); - intel_attach_hdmi_colorspace_property(connector); + drm_connector_attach_content_type_property(&connector->base); + intel_attach_hdmi_colorspace_property(&connector->base); } else { - intel_attach_dp_colorspace_property(connector); + intel_attach_dp_colorspace_property(&connector->base); } if (intel_dp_has_gamut_metadata_dip(&dp_to_dig_port(intel_dp)->base)) - drm_connector_attach_hdr_output_metadata_property(connector); + drm_connector_attach_hdr_output_metadata_property(&connector->base); if (HAS_VRR(display)) - drm_connector_attach_vrr_capable_property(connector); + drm_connector_attach_vrr_capable_property(&connector->base); } static void @@ -6344,7 +6448,6 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, struct intel_connector *connector) { struct intel_display *display = to_intel_display(intel_dp); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct drm_display_mode *fixed_mode; struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; bool has_dpcd; @@ -6361,7 +6464,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, */ if (intel_get_lvds_encoder(display)) { drm_WARN_ON(display->drm, - !(HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv))); + !(HAS_PCH_IBX(display) || HAS_PCH_CPT(display))); drm_info(display->drm, "LVDS was detected, not registering eDP\n"); @@ -6392,7 +6495,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, */ intel_hpd_enable_detection(encoder); - intel_alpm_init_dpcd(intel_dp); + intel_alpm_init(intel_dp); /* Cache DPCD and EDID for edp. */ has_dpcd = intel_edp_init_dpcd(intel_dp, connector); diff --git a/drivers/gpu/drm/i915/display/intel_dp.h b/drivers/gpu/drm/i915/display/intel_dp.h index 9189db4c2594..742ae26ac4a9 100644 --- a/drivers/gpu/drm/i915/display/intel_dp.h +++ b/drivers/gpu/drm/i915/display/intel_dp.h @@ -194,6 +194,7 @@ void intel_dp_wait_source_oui(struct intel_dp *intel_dp); int intel_dp_output_bpp(enum intel_output_format output_format, int bpp); bool intel_dp_compute_config_limits(struct intel_dp *intel_dp, + struct intel_connector *connector, struct intel_crtc_state *crtc_state, bool respect_downstream_limits, bool dsc, @@ -208,5 +209,7 @@ bool intel_dp_has_connector(struct intel_dp *intel_dp, const struct drm_connector_state *conn_state); int intel_dp_dsc_max_src_input_bpc(struct intel_display *display); int intel_dp_dsc_min_src_input_bpc(void); +int intel_dp_compute_min_hblank(struct intel_crtc_state *crtc_state, + const struct drm_connector_state *conn_state); #endif /* __INTEL_DP_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c index 0496061203fb..bf8e8e0cc19c 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c @@ -3,8 +3,10 @@ * Copyright © 2020-2021 Intel Corporation */ -#include "i915_drv.h" +#include <drm/drm_print.h> + #include "i915_reg.h" +#include "i915_utils.h" #include "intel_de.h" #include "intel_display_types.h" #include "intel_dp.h" @@ -111,10 +113,9 @@ static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index) static u32 hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { struct intel_display *display = to_intel_display(intel_dp); - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - if (dig_port->aux_ch != AUX_CH_A && HAS_PCH_LPT_H(i915)) { + if (dig_port->aux_ch != AUX_CH_A && HAS_PCH_LPT_H(display)) { /* Workaround for non-ULT HSW */ switch (index) { case 0: return 63; @@ -177,12 +178,11 @@ static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp, int send_bytes, u32 aux_clock_divider) { - struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(intel_dp); u32 timeout; /* Max timeout value on G4x-BDW: 1.6ms */ - if (IS_BROADWELL(i915)) + if (display->platform.broadwell) timeout = DP_AUX_CH_CTL_TIME_OUT_600us; else timeout = DP_AUX_CH_CTL_TIME_OUT_400us; @@ -786,7 +786,6 @@ void intel_dp_aux_fini(struct intel_dp *intel_dp) void intel_dp_aux_init(struct intel_dp *intel_dp) { struct intel_display *display = to_intel_display(intel_dp); - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct intel_encoder *encoder = &dig_port->base; enum aux_ch aux_ch = dig_port->aux_ch; @@ -801,10 +800,10 @@ void intel_dp_aux_init(struct intel_dp *intel_dp) } else if (DISPLAY_VER(display) >= 9) { intel_dp->aux_ch_ctl_reg = skl_aux_ctl_reg; intel_dp->aux_ch_data_reg = skl_aux_data_reg; - } else if (HAS_PCH_SPLIT(i915)) { + } else if (HAS_PCH_SPLIT(display)) { intel_dp->aux_ch_ctl_reg = ilk_aux_ctl_reg; intel_dp->aux_ch_data_reg = ilk_aux_data_reg; - } else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) { + } else if (display->platform.valleyview || display->platform.cherryview) { intel_dp->aux_ch_ctl_reg = vlv_aux_ctl_reg; intel_dp->aux_ch_data_reg = vlv_aux_data_reg; } else { @@ -814,9 +813,9 @@ void intel_dp_aux_init(struct intel_dp *intel_dp) if (DISPLAY_VER(display) >= 9) intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider; - else if (IS_BROADWELL(i915) || IS_HASWELL(i915)) + else if (display->platform.broadwell || display->platform.haswell) intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider; - else if (HAS_PCH_SPLIT(i915)) + else if (HAS_PCH_SPLIT(display)) intel_dp->get_aux_clock_divider = ilk_get_aux_clock_divider; else intel_dp->get_aux_clock_divider = g4x_get_aux_clock_divider; diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c index 20ab90acb351..271b27c9de51 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c +++ b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c @@ -36,7 +36,6 @@ #include <drm/drm_print.h> -#include "i915_utils.h" #include "intel_backlight.h" #include "intel_display_core.h" #include "intel_display_types.h" @@ -149,7 +148,7 @@ intel_dp_aux_supports_hdr_backlight(struct intel_connector *connector) !(connector->base.hdr_sink_metadata.hdmi_type1.metadata_type & BIT(HDMI_STATIC_METADATA_TYPE1))) { drm_info(display->drm, - "[CONNECTOR:%d:%s] Panel is missing HDR static metadata. Possible support for Intel HDR backlight interface is not used. If your backlight controls don't work try booting with i915.enable_dpcd_backlight=%d. needs this, please file a _new_ bug report on drm/i915, see " FDO_BUG_URL " for details.\n", + "[CONNECTOR:%d:%s] Panel is missing HDR static metadata. Possible support for Intel HDR backlight interface is not used. If your backlight controls don't work try booting with i915.enable_dpcd_backlight=%d.\n", connector->base.base.id, connector->base.name, INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL); return false; diff --git a/drivers/gpu/drm/i915/display/intel_dp_mst.c b/drivers/gpu/drm/i915/display/intel_dp_mst.c index 4c15dcb103aa..c1fd6aceec2c 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_mst.c +++ b/drivers/gpu/drm/i915/display/intel_dp_mst.c @@ -27,10 +27,11 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_edid.h> #include <drm/drm_fixed.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_atomic.h" #include "intel_audio.h" #include "intel_connector.h" @@ -51,6 +52,7 @@ #include "intel_link_bw.h" #include "intel_pfit.h" #include "intel_psr.h" +#include "intel_step.h" #include "intel_vdsc.h" #include "intel_vrr.h" #include "skl_scaler.h" @@ -239,26 +241,6 @@ static int intel_dp_mst_dsc_get_slice_count(const struct intel_connector *connec num_joined_pipes); } -static void intel_dp_mst_compute_min_hblank(struct intel_crtc_state *crtc_state, - int bpp_x16) -{ - struct intel_display *display = to_intel_display(crtc_state); - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; - int symbol_size = intel_dp_is_uhbr(crtc_state) ? 32 : 8; - int hblank; - - if (DISPLAY_VER(display) < 20) - return; - - /* Calculate min Hblank Link Layer Symbol Cycle Count for 8b/10b MST & 128b/132b */ - hblank = DIV_ROUND_UP((DIV_ROUND_UP - (adjusted_mode->htotal - adjusted_mode->hdisplay, 4) * bpp_x16), - symbol_size); - - crtc_state->min_hblank = hblank; -} - int intel_dp_mtp_tu_compute_config(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state, struct drm_connector_state *conn_state, @@ -329,8 +311,6 @@ int intel_dp_mtp_tu_compute_config(struct intel_dp *intel_dp, local_bw_overhead = intel_dp_mst_bw_overhead(crtc_state, false, dsc_slice_count, link_bpp_x16); - intel_dp_mst_compute_min_hblank(crtc_state, link_bpp_x16); - intel_dp_mst_compute_m_n(crtc_state, local_bw_overhead, link_bpp_x16, @@ -619,12 +599,13 @@ adjust_limits_for_dsc_hblank_expansion_quirk(struct intel_dp *intel_dp, static bool mst_stream_compute_config_limits(struct intel_dp *intel_dp, - const struct intel_connector *connector, + struct intel_connector *connector, struct intel_crtc_state *crtc_state, bool dsc, struct link_config_limits *limits) { - if (!intel_dp_compute_config_limits(intel_dp, crtc_state, false, dsc, + if (!intel_dp_compute_config_limits(intel_dp, connector, + crtc_state, false, dsc, limits)) return false; @@ -739,6 +720,10 @@ static int mst_stream_compute_config(struct intel_encoder *encoder, pipe_config->lane_lat_optim_mask = bxt_dpio_phy_calc_lane_lat_optim_mask(pipe_config->lane_count); + ret = intel_dp_compute_min_hblank(pipe_config, conn_state); + if (ret) + return ret; + intel_vrr_compute_config(pipe_config, conn_state); intel_dp_audio_compute_config(encoder, pipe_config, conn_state); @@ -1021,12 +1006,10 @@ static void mst_stream_disable(struct intel_atomic_state *state, const struct intel_crtc_state *old_crtc_state, const struct drm_connector_state *old_conn_state) { - struct intel_display *display = to_intel_display(state); struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder); struct intel_dp *intel_dp = to_primary_dp(encoder); struct intel_connector *connector = to_intel_connector(old_conn_state->connector); - enum transcoder trans = old_crtc_state->cpu_transcoder; if (intel_dp_mst_active_streams(intel_dp) == 1) intel_dp->link.active = false; @@ -1034,9 +1017,6 @@ static void mst_stream_disable(struct intel_atomic_state *state, intel_hdcp_disable(intel_mst->connector); intel_dp_sink_disable_decompression(state, connector, old_crtc_state); - - if (DISPLAY_VER(display) >= 20) - intel_de_write(display, DP_MIN_HBLANK_CTL(trans), 0); } static void mst_stream_post_disable(struct intel_atomic_state *state, @@ -1305,7 +1285,7 @@ static void mst_stream_enable(struct intel_atomic_state *state, enum transcoder trans = pipe_config->cpu_transcoder; bool first_mst_stream = intel_dp_mst_active_streams(intel_dp) == 1; struct intel_crtc *pipe_crtc; - int ret, i, min_hblank; + int ret, i; drm_WARN_ON(display->drm, pipe_config->has_pch_encoder); @@ -1320,29 +1300,6 @@ static void mst_stream_enable(struct intel_atomic_state *state, TRANS_DP2_VFREQ_PIXEL_CLOCK(crtc_clock_hz & 0xffffff)); } - if (DISPLAY_VER(display) >= 20) { - /* - * adjust the BlankingStart/BlankingEnd framing control from - * the calculated value - */ - min_hblank = pipe_config->min_hblank - 2; - - /* Maximum value to be programmed is limited to 0x10 */ - min_hblank = min(0x10, min_hblank); - - /* - * Minimum hblank accepted for 128b/132b would be 5 and for - * 8b/10b would be 3 symbol count - */ - if (intel_dp_is_uhbr(pipe_config)) - min_hblank = max(min_hblank, 5); - else - min_hblank = max(min_hblank, 3); - - intel_de_write(display, DP_MIN_HBLANK_CTL(trans), - min_hblank); - } - enable_bs_jitter_was(pipe_config); intel_ddi_enable_transcoder_func(encoder, pipe_config); diff --git a/drivers/gpu/drm/i915/display/intel_dpio_phy.c b/drivers/gpu/drm/i915/display/intel_dpio_phy.c index 429f89543789..69f242139420 100644 --- a/drivers/gpu/drm/i915/display/intel_dpio_phy.c +++ b/drivers/gpu/drm/i915/display/intel_dpio_phy.c @@ -222,9 +222,7 @@ static const struct bxt_dpio_phy_info glk_dpio_phy_info[] = { static const struct bxt_dpio_phy_info * bxt_get_phy_list(struct intel_display *display, int *count) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - if (IS_GEMINILAKE(dev_priv)) { + if (display->platform.geminilake) { *count = ARRAY_SIZE(glk_dpio_phy_info); return glk_dpio_phy_info; } else { @@ -808,9 +806,9 @@ void chv_set_phy_signal_level(struct intel_encoder *encoder, vlv_dpio_put(dev_priv); } -void chv_data_lane_soft_reset(struct intel_encoder *encoder, - const struct intel_crtc_state *crtc_state, - bool reset) +static void __chv_data_lane_soft_reset(struct intel_encoder *encoder, + const struct intel_crtc_state *crtc_state, + bool reset) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_digital_port *dig_port = enc_to_dig_port(encoder); @@ -853,6 +851,17 @@ void chv_data_lane_soft_reset(struct intel_encoder *encoder, } } +void chv_data_lane_soft_reset(struct intel_encoder *encoder, + const struct intel_crtc_state *crtc_state, + bool reset) +{ + struct drm_i915_private *i915 = to_i915(encoder->base.dev); + + vlv_dpio_get(i915); + __chv_data_lane_soft_reset(encoder, crtc_state, reset); + vlv_dpio_put(i915); +} + void chv_phy_pre_pll_enable(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { @@ -880,7 +889,7 @@ void chv_phy_pre_pll_enable(struct intel_encoder *encoder, vlv_dpio_get(dev_priv); /* Assert data lane reset */ - chv_data_lane_soft_reset(encoder, crtc_state, true); + __chv_data_lane_soft_reset(encoder, crtc_state, true); /* program left/right clock distribution */ if (pipe != PIPE_B) { @@ -1008,7 +1017,7 @@ void chv_phy_pre_encoder_enable(struct intel_encoder *encoder, } /* Deassert data lane reset */ - chv_data_lane_soft_reset(encoder, crtc_state, false); + __chv_data_lane_soft_reset(encoder, crtc_state, false); vlv_dpio_put(dev_priv); } diff --git a/drivers/gpu/drm/i915/display/intel_dpll.c b/drivers/gpu/drm/i915/display/intel_dpll.c index 0481b1365b85..a9e9b98d0bf9 100644 --- a/drivers/gpu/drm/i915/display/intel_dpll.c +++ b/drivers/gpu/drm/i915/display/intel_dpll.c @@ -374,12 +374,11 @@ int chv_calc_dpll_params(int refclk, struct dpll *clock) static int i9xx_pll_refclk(const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); - struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev); const struct i9xx_dpll_hw_state *hw_state = &crtc_state->dpll_hw_state.i9xx; if ((hw_state->dpll & PLL_REF_INPUT_MASK) == PLLB_REF_INPUT_SPREADSPECTRUMIN) return display->vbt.lvds_ssc_freq; - else if (HAS_PCH_SPLIT(i915)) + else if (HAS_PCH_SPLIT(display)) return 120000; else if (DISPLAY_VER(display) != 2) return 96000; @@ -1235,12 +1234,10 @@ static int mtl_crtc_compute_clock(struct intel_atomic_state *state, static int ilk_fb_cb_factor(const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); - struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - struct drm_i915_private *i915 = to_i915(crtc->base.dev); if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) && ((intel_panel_use_ssc(display) && display->vbt.lvds_ssc_freq == 100000) || - (HAS_PCH_IBX(i915) && intel_is_dual_link_lvds(display)))) + (HAS_PCH_IBX(display) && intel_is_dual_link_lvds(display)))) return 25; if (crtc_state->sdvo_tv_clock) @@ -1791,15 +1788,13 @@ int intel_dpll_crtc_get_shared_dpll(struct intel_atomic_state *state, void intel_dpll_init_clock_hook(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (DISPLAY_VER(display) >= 14) display->funcs.dpll = &mtl_dpll_funcs; else if (display->platform.dg2) display->funcs.dpll = &dg2_dpll_funcs; else if (DISPLAY_VER(display) >= 9 || HAS_DDI(display)) display->funcs.dpll = &hsw_dpll_funcs; - else if (HAS_PCH_SPLIT(dev_priv)) + else if (HAS_PCH_SPLIT(display)) display->funcs.dpll = &ilk_dpll_funcs; else if (display->platform.cherryview) display->funcs.dpll = &chv_dpll_funcs; diff --git a/drivers/gpu/drm/i915/display/intel_dpll_mgr.c b/drivers/gpu/drm/i915/display/intel_dpll_mgr.c index 84df41086a89..9da051a3f455 100644 --- a/drivers/gpu/drm/i915/display/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/display/intel_dpll_mgr.c @@ -24,9 +24,11 @@ #include <linux/math.h> #include <linux/string_helpers.h> +#include <drm/drm_print.h> + #include "bxt_dpio_phy_regs.h" -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_cx0_phy.h" #include "intel_de.h" #include "intel_display_types.h" @@ -38,6 +40,7 @@ #include "intel_hti.h" #include "intel_mg_phy_regs.h" #include "intel_pch_refclk.h" +#include "intel_step.h" #include "intel_tc.h" /** @@ -609,13 +612,12 @@ static int ibx_get_dpll(struct intel_atomic_state *state, struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(state); - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); struct intel_shared_dpll *pll; enum intel_dpll_id id; - if (HAS_PCH_IBX(i915)) { + if (HAS_PCH_IBX(display)) { /* Ironlake PCH has a fixed PLL->PCH pipe mapping. */ id = (enum intel_dpll_id) crtc->pipe; pll = intel_get_shared_dpll_by_id(display, id); @@ -4305,7 +4307,6 @@ static const struct intel_dpll_mgr adlp_pll_mgr = { */ void intel_shared_dpll_init(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); const struct intel_dpll_mgr *dpll_mgr = NULL; const struct dpll_info *dpll_info; int i; @@ -4335,7 +4336,7 @@ void intel_shared_dpll_init(struct intel_display *display) dpll_mgr = &skl_pll_mgr; else if (HAS_DDI(display)) dpll_mgr = &hsw_pll_mgr; - else if (HAS_PCH_IBX(i915) || HAS_PCH_CPT(i915)) + else if (HAS_PCH_IBX(display) || HAS_PCH_CPT(display)) dpll_mgr = &pch_pll_mgr; if (!dpll_mgr) diff --git a/drivers/gpu/drm/i915/display/intel_dpt_common.c b/drivers/gpu/drm/i915/display/intel_dpt_common.c index d2dede0a5229..ce5aa0ca0fa5 100644 --- a/drivers/gpu/drm/i915/display/intel_dpt_common.c +++ b/drivers/gpu/drm/i915/display/intel_dpt_common.c @@ -3,7 +3,6 @@ * Copyright © 2023 Intel Corporation */ -#include "i915_drv.h" #include "i915_reg.h" #include "intel_de.h" #include "intel_display_types.h" @@ -12,9 +11,9 @@ void intel_dpt_configure(struct intel_crtc *crtc) { - struct drm_i915_private *i915 = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(crtc); - if (DISPLAY_VER(i915) == 14) { + if (DISPLAY_VER(display) == 14) { enum pipe pipe = crtc->pipe; enum plane_id plane_id; @@ -22,15 +21,15 @@ void intel_dpt_configure(struct intel_crtc *crtc) if (plane_id == PLANE_CURSOR) continue; - intel_de_rmw(i915, PLANE_CHICKEN(pipe, plane_id), + intel_de_rmw(display, PLANE_CHICKEN(pipe, plane_id), PLANE_CHICKEN_DISABLE_DPT, - i915->display.params.enable_dpt ? 0 : + display->params.enable_dpt ? 0 : PLANE_CHICKEN_DISABLE_DPT); } - } else if (DISPLAY_VER(i915) == 13) { - intel_de_rmw(i915, CHICKEN_MISC_2, + } else if (DISPLAY_VER(display) == 13) { + intel_de_rmw(display, CHICKEN_MISC_2, CHICKEN_MISC_DISABLE_DPT, - i915->display.params.enable_dpt ? 0 : + display->params.enable_dpt ? 0 : CHICKEN_MISC_DISABLE_DPT); } } diff --git a/drivers/gpu/drm/i915/display/intel_dsb.c b/drivers/gpu/drm/i915/display/intel_dsb.c index 72fe390c5af2..481488d1fe67 100644 --- a/drivers/gpu/drm/i915/display/intel_dsb.c +++ b/drivers/gpu/drm/i915/display/intel_dsb.c @@ -4,11 +4,12 @@ * */ +#include <drm/drm_print.h> #include <drm/drm_vblank.h> -#include "i915_drv.h" #include "i915_irq.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_crtc.h" #include "intel_de.h" #include "intel_display_rpm.h" diff --git a/drivers/gpu/drm/i915/display/intel_dsi.c b/drivers/gpu/drm/i915/display/intel_dsi.c index 403151175a87..a8f012119165 100644 --- a/drivers/gpu/drm/i915/display/intel_dsi.c +++ b/drivers/gpu/drm/i915/display/intel_dsi.c @@ -4,8 +4,9 @@ */ #include <drm/drm_mipi_dsi.h> +#include <drm/drm_print.h> -#include "i915_drv.h" +#include "intel_display_core.h" #include "intel_dsi.h" #include "intel_panel.h" @@ -116,14 +117,14 @@ struct intel_dsi_host *intel_dsi_host_init(struct intel_dsi *intel_dsi, enum drm_panel_orientation intel_dsi_get_panel_orientation(struct intel_connector *connector) { - struct drm_i915_private *dev_priv = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); enum drm_panel_orientation orientation; orientation = connector->panel.vbt.dsi.orientation; if (orientation != DRM_MODE_PANEL_ORIENTATION_UNKNOWN) return orientation; - orientation = dev_priv->display.vbt.orientation; + orientation = display->vbt.orientation; if (orientation != DRM_MODE_PANEL_ORIENTATION_UNKNOWN) return orientation; diff --git a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c index 4e92504f5c14..29c920983413 100644 --- a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c +++ b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c @@ -31,16 +31,16 @@ #include <linux/pinctrl/machine.h> #include <linux/slab.h> #include <linux/string_helpers.h> - #include <linux/unaligned.h> #include <drm/drm_crtc.h> #include <drm/drm_edid.h> +#include <drm/drm_print.h> #include <video/mipi_display.h> -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_de.h" #include "intel_display_types.h" #include "intel_dsi.h" @@ -321,7 +321,6 @@ enum { static void icl_native_gpio_set_value(struct intel_display *display, int gpio, bool value) { - struct drm_i915_private *dev_priv = to_i915(display->drm); int index; if (drm_WARN_ON(display->drm, DISPLAY_VER(display) == 11 && gpio >= MIPI_RESET_2)) @@ -341,12 +340,12 @@ static void icl_native_gpio_set_value(struct intel_display *display, * The locking protects against concurrent SHOTPLUG_CTL_DDI * modifications in irq setup and handling. */ - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); intel_de_rmw(display, SHOTPLUG_CTL_DDI, SHOTPLUG_CTL_DDI_HPD_ENABLE(index) | SHOTPLUG_CTL_DDI_HPD_OUTPUT_DATA(index), value ? SHOTPLUG_CTL_DDI_HPD_OUTPUT_DATA(index) : 0); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); break; case MIPI_AVDD_EN_1: case MIPI_AVDD_EN_2: diff --git a/drivers/gpu/drm/i915/display/intel_fb.c b/drivers/gpu/drm/i915/display/intel_fb.c index 2b0e0f220442..05393bd60c98 100644 --- a/drivers/gpu/drm/i915/display/intel_fb.c +++ b/drivers/gpu/drm/i915/display/intel_fb.c @@ -560,11 +560,11 @@ static bool plane_has_modifier(struct intel_display *display, return false; if (md->modifier == I915_FORMAT_MOD_4_TILED_BMG_CCS && - (GRAPHICS_VER(i915) < 20 || !IS_DGFX(i915))) + (GRAPHICS_VER(i915) < 20 || !display->platform.dgfx)) return false; if (md->modifier == I915_FORMAT_MOD_4_TILED_LNL_CCS && - (GRAPHICS_VER(i915) < 20 || IS_DGFX(i915))) + (GRAPHICS_VER(i915) < 20 || display->platform.dgfx)) return false; return true; diff --git a/drivers/gpu/drm/i915/display/intel_fbc.c b/drivers/gpu/drm/i915/display/intel_fbc.c index ce5b1e3f1c20..bed2bba20b55 100644 --- a/drivers/gpu/drm/i915/display/intel_fbc.c +++ b/drivers/gpu/drm/i915/display/intel_fbc.c @@ -252,9 +252,12 @@ static u16 intel_fbc_override_cfb_stride(const struct intel_plane_state *plane_s * Gen9 hw miscalculates cfb stride for linear as * PLANE_STRIDE*512 instead of PLANE_STRIDE*64, so * we always need to use the override there. + * + * wa_14022269668 For bmg, always program the FBC_STRIDE before fbc enable */ if (stride != stride_aligned || - (DISPLAY_VER(display) == 9 && fb->modifier == DRM_FORMAT_MOD_LINEAR)) + (DISPLAY_VER(display) == 9 && fb->modifier == DRM_FORMAT_MOD_LINEAR) || + display->platform.battlemage) return stride_aligned * 4 / 64; return 0; diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.c b/drivers/gpu/drm/i915/display/intel_fbdev.c index 369f46286e95..2dc4029d71ed 100644 --- a/drivers/gpu/drm/i915/display/intel_fbdev.c +++ b/drivers/gpu/drm/i915/display/intel_fbdev.c @@ -47,9 +47,9 @@ #include <drm/drm_managed.h> #include <drm/drm_print.h> -#include "i915_drv.h" #include "i915_vma.h" #include "intel_bo.h" +#include "intel_display_core.h" #include "intel_display_rpm.h" #include "intel_display_types.h" #include "intel_fb.h" @@ -66,9 +66,9 @@ struct intel_fbdev { static struct intel_fbdev *to_intel_fbdev(struct drm_fb_helper *fb_helper) { - struct drm_i915_private *i915 = to_i915(fb_helper->client.dev); + struct intel_display *display = to_intel_display(fb_helper->client.dev); - return i915->display.fbdev.fbdev; + return display->fbdev.fbdev; } static struct intel_frontbuffer *to_frontbuffer(struct intel_fbdev *ifbdev) @@ -210,11 +210,9 @@ static const struct drm_fb_helper_funcs intel_fb_helper_funcs = { int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes) { + struct intel_display *display = to_intel_display(helper->dev); struct intel_fbdev *ifbdev = to_intel_fbdev(helper); struct intel_framebuffer *fb = ifbdev->fb; - struct drm_device *dev = helper->dev; - struct drm_i915_private *dev_priv = to_i915(dev); - struct intel_display *display = to_intel_display(dev); struct ref_tracker *wakeref; struct fb_info *info; struct i915_vma *vma; @@ -228,7 +226,7 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, if (fb && (sizes->fb_width > fb->base.width || sizes->fb_height > fb->base.height)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "BIOS fb too small (%dx%d), we require (%dx%d)," " releasing it\n", fb->base.width, fb->base.height, @@ -236,14 +234,14 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, drm_framebuffer_put(&fb->base); fb = NULL; } - if (!fb || drm_WARN_ON(dev, !intel_fb_bo(&fb->base))) { - drm_dbg_kms(&dev_priv->drm, + if (!fb || drm_WARN_ON(display->drm, !intel_fb_bo(&fb->base))) { + drm_dbg_kms(display->drm, "no BIOS fb, allocating a new one\n"); fb = intel_fbdev_fb_alloc(helper, sizes); if (IS_ERR(fb)) return PTR_ERR(fb); } else { - drm_dbg_kms(&dev_priv->drm, "re-using BIOS fb\n"); + drm_dbg_kms(display->drm, "re-using BIOS fb\n"); prealloc = true; sizes->fb_width = fb->base.width; sizes->fb_height = fb->base.height; @@ -267,7 +265,7 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, info = drm_fb_helper_alloc_info(helper); if (IS_ERR(info)) { - drm_err(&dev_priv->drm, "Failed to allocate fb_info (%pe)\n", info); + drm_err(display->drm, "Failed to allocate fb_info (%pe)\n", info); ret = PTR_ERR(info); goto out_unpin; } @@ -279,11 +277,11 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, obj = intel_fb_bo(&fb->base); - ret = intel_fbdev_fb_fill_info(dev_priv, info, obj, vma); + ret = intel_fbdev_fb_fill_info(display, info, obj, vma); if (ret) goto out_unpin; - drm_fb_helper_fill_info(info, dev->fb_helper, sizes); + drm_fb_helper_fill_info(info, display->drm->fb_helper, sizes); /* If the object is shmemfs backed, it will have given us zeroed pages. * If the object is stolen however, it will be full of whatever @@ -294,7 +292,7 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ - drm_dbg_kms(&dev_priv->drm, "allocated %dx%d fb: 0x%08x\n", + drm_dbg_kms(display->drm, "allocated %dx%d fb: 0x%08x\n", fb->base.width, fb->base.height, i915_ggtt_offset(vma)); ifbdev->fb = fb; @@ -322,16 +320,15 @@ out_unlock: * Note we only support a single fb shared across pipes for boot (mostly for * fbcon), so we just find the biggest and use that. */ -static bool intel_fbdev_init_bios(struct drm_device *dev, +static bool intel_fbdev_init_bios(struct intel_display *display, struct intel_fbdev *ifbdev) { - struct drm_i915_private *i915 = to_i915(dev); struct intel_framebuffer *fb = NULL; struct intel_crtc *crtc; unsigned int max_size = 0; /* Find the largest fb */ - for_each_intel_crtc(dev, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_plane *plane = @@ -341,21 +338,21 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, struct drm_gem_object *obj = intel_fb_bo(plane_state->uapi.fb); if (!crtc_state->uapi.active) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[CRTC:%d:%s] not active, skipping\n", crtc->base.base.id, crtc->base.name); continue; } if (!obj) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[PLANE:%d:%s] no fb, skipping\n", plane->base.base.id, plane->base.name); continue; } if (obj->size > max_size) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "found possible fb from [PLANE:%d:%s]\n", plane->base.base.id, plane->base.name); fb = to_intel_framebuffer(plane_state->uapi.fb); @@ -364,13 +361,13 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, } if (!fb) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "no active fbs found, not using BIOS config\n"); goto out; } /* Now make sure all the pipes will fit into it */ - for_each_intel_crtc(dev, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_plane *plane = @@ -378,13 +375,13 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, unsigned int cur_size; if (!crtc_state->uapi.active) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[CRTC:%d:%s] not active, skipping\n", crtc->base.base.id, crtc->base.name); continue; } - drm_dbg_kms(&i915->drm, "checking [PLANE:%d:%s] for BIOS fb\n", + drm_dbg_kms(display->drm, "checking [PLANE:%d:%s] for BIOS fb\n", plane->base.base.id, plane->base.name); /* @@ -395,7 +392,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, cur_size = crtc_state->uapi.adjusted_mode.crtc_hdisplay; cur_size = cur_size * fb->base.format->cpp[0]; if (fb->base.pitches[0] < cur_size) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "fb not wide enough for [PLANE:%d:%s] (%d vs %d)\n", plane->base.base.id, plane->base.name, cur_size, fb->base.pitches[0]); @@ -406,7 +403,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, cur_size = crtc_state->uapi.adjusted_mode.crtc_vdisplay; cur_size = intel_fb_align_height(&fb->base, 0, cur_size); cur_size *= fb->base.pitches[0]; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[CRTC:%d:%s] area: %dx%d, bpp: %d, size: %d\n", crtc->base.base.id, crtc->base.name, crtc_state->uapi.adjusted_mode.crtc_hdisplay, @@ -415,7 +412,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, cur_size); if (cur_size > max_size) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "fb not big enough for [PLANE:%d:%s] (%d vs %d)\n", plane->base.base.id, plane->base.name, cur_size, max_size); @@ -423,14 +420,14 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, break; } - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "fb big enough [PLANE:%d:%s] (%d >= %d)\n", plane->base.base.id, plane->base.name, max_size, cur_size); } if (!fb) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "BIOS fb not suitable for all pipes, not using\n"); goto out; } @@ -440,7 +437,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, drm_framebuffer_get(&ifbdev->fb->base); /* Final pass to check if any active pipes don't have fbs */ - for_each_intel_crtc(dev, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_plane *plane = @@ -451,13 +448,13 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, if (!crtc_state->uapi.active) continue; - drm_WARN(dev, !plane_state->uapi.fb, + drm_WARN(display->drm, !plane_state->uapi.fb, "re-used BIOS config but lost an fb on [PLANE:%d:%s]\n", plane->base.base.id, plane->base.name); } - drm_dbg_kms(&i915->drm, "using BIOS fb for initial console\n"); + drm_dbg_kms(display->drm, "using BIOS fb for initial console\n"); return true; out: @@ -482,26 +479,25 @@ static unsigned int intel_fbdev_color_mode(const struct drm_format_info *info) } } -void intel_fbdev_setup(struct drm_i915_private *i915) +void intel_fbdev_setup(struct intel_display *display) { - struct drm_device *dev = &i915->drm; struct intel_fbdev *ifbdev; unsigned int preferred_bpp = 0; - if (!HAS_DISPLAY(i915)) + if (!HAS_DISPLAY(display)) return; - ifbdev = drmm_kzalloc(dev, sizeof(*ifbdev), GFP_KERNEL); + ifbdev = drmm_kzalloc(display->drm, sizeof(*ifbdev), GFP_KERNEL); if (!ifbdev) return; - i915->display.fbdev.fbdev = ifbdev; - if (intel_fbdev_init_bios(dev, ifbdev)) + display->fbdev.fbdev = ifbdev; + if (intel_fbdev_init_bios(display, ifbdev)) preferred_bpp = intel_fbdev_color_mode(ifbdev->fb->base.format); if (!preferred_bpp) preferred_bpp = 32; - drm_client_setup_with_color_mode(dev, preferred_bpp); + drm_client_setup_with_color_mode(display->drm, preferred_bpp); } struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev) diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.h b/drivers/gpu/drm/i915/display/intel_fbdev.h index 89bad3a2b01a..a15e3e222a0c 100644 --- a/drivers/gpu/drm/i915/display/intel_fbdev.h +++ b/drivers/gpu/drm/i915/display/intel_fbdev.h @@ -10,7 +10,7 @@ struct drm_fb_helper; struct drm_fb_helper_surface_size; -struct drm_i915_private; +struct intel_display; struct intel_fbdev; struct intel_framebuffer; @@ -19,14 +19,14 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes); #define INTEL_FBDEV_DRIVER_OPS \ .fbdev_probe = intel_fbdev_driver_fbdev_probe -void intel_fbdev_setup(struct drm_i915_private *dev_priv); +void intel_fbdev_setup(struct intel_display *display); struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev); struct i915_vma *intel_fbdev_vma_pointer(struct intel_fbdev *fbdev); #else #define INTEL_FBDEV_DRIVER_OPS \ .fbdev_probe = NULL -static inline void intel_fbdev_setup(struct drm_i915_private *dev_priv) +static inline void intel_fbdev_setup(struct intel_display *display) { } static inline struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev) diff --git a/drivers/gpu/drm/i915/display/intel_fbdev_fb.c b/drivers/gpu/drm/i915/display/intel_fbdev_fb.c index 4991c35a2632..5f4cb3328265 100644 --- a/drivers/gpu/drm/i915/display/intel_fbdev_fb.c +++ b/drivers/gpu/drm/i915/display/intel_fbdev_fb.c @@ -15,9 +15,9 @@ struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes) { + struct intel_display *display = to_intel_display(helper->dev); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct drm_framebuffer *fb; - struct drm_device *dev = helper->dev; - struct drm_i915_private *dev_priv = to_i915(dev); struct drm_mode_fb_cmd2 mode_cmd = {}; struct drm_i915_gem_object *obj; int size; @@ -50,14 +50,14 @@ struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper, * * Also skip stolen on MTL as Wa_22018444074 mitigation. */ - if (!(IS_METEORLAKE(dev_priv)) && size * 2 < dev_priv->dsm.usable_size) + if (!display->platform.meteorlake && size * 2 < dev_priv->dsm.usable_size) obj = i915_gem_object_create_stolen(dev_priv, size); if (IS_ERR(obj)) obj = i915_gem_object_create_shmem(dev_priv, size); } if (IS_ERR(obj)) { - drm_err(&dev_priv->drm, "failed to allocate framebuffer (%pe)\n", obj); + drm_err(display->drm, "failed to allocate framebuffer (%pe)\n", obj); return ERR_PTR(-ENOMEM); } @@ -67,9 +67,10 @@ struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper, return to_intel_framebuffer(fb); } -int intel_fbdev_fb_fill_info(struct drm_i915_private *i915, struct fb_info *info, +int intel_fbdev_fb_fill_info(struct intel_display *display, struct fb_info *info, struct drm_gem_object *_obj, struct i915_vma *vma) { + struct drm_i915_private *i915 = to_i915(display->drm); struct drm_i915_gem_object *obj = to_intel_bo(_obj); struct i915_gem_ww_ctx ww; void __iomem *vaddr; @@ -101,7 +102,7 @@ int intel_fbdev_fb_fill_info(struct drm_i915_private *i915, struct fb_info *info vaddr = i915_vma_pin_iomap(vma); if (IS_ERR(vaddr)) { - drm_err(&i915->drm, + drm_err(display->drm, "Failed to remap framebuffer into virtual memory (%pe)\n", vaddr); ret = PTR_ERR(vaddr); continue; diff --git a/drivers/gpu/drm/i915/display/intel_fbdev_fb.h b/drivers/gpu/drm/i915/display/intel_fbdev_fb.h index e502ae375fc0..cb7957272715 100644 --- a/drivers/gpu/drm/i915/display/intel_fbdev_fb.h +++ b/drivers/gpu/drm/i915/display/intel_fbdev_fb.h @@ -9,13 +9,13 @@ struct drm_fb_helper; struct drm_fb_helper_surface_size; struct drm_gem_object; -struct drm_i915_private; struct fb_info; struct i915_vma; +struct intel_display; struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes); -int intel_fbdev_fb_fill_info(struct drm_i915_private *i915, struct fb_info *info, +int intel_fbdev_fb_fill_info(struct intel_display *display, struct fb_info *info, struct drm_gem_object *obj, struct i915_vma *vma); #endif diff --git a/drivers/gpu/drm/i915/display/intel_fdi.c b/drivers/gpu/drm/i915/display/intel_fdi.c index 40deee0769ae..169bbe154b5c 100644 --- a/drivers/gpu/drm/i915/display/intel_fdi.c +++ b/drivers/gpu/drm/i915/display/intel_fdi.c @@ -6,15 +6,16 @@ #include <linux/string_helpers.h> #include <drm/drm_fixed.h> +#include <drm/drm_print.h> -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_atomic.h" #include "intel_crtc.h" #include "intel_ddi.h" #include "intel_de.h" -#include "intel_dp.h" #include "intel_display_types.h" +#include "intel_dp.h" #include "intel_fdi.h" #include "intel_fdi_regs.h" #include "intel_link_bw.h" @@ -464,7 +465,6 @@ static void ivb_update_fdi_bc_bifurcation(const struct intel_crtc_state *crtc_st void intel_fdi_normal_train(struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe = crtc->pipe; i915_reg_t reg; u32 temp; @@ -483,7 +483,7 @@ void intel_fdi_normal_train(struct intel_crtc *crtc) reg = FDI_RX_CTL(pipe); temp = intel_de_read(display, reg); - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT; temp |= FDI_LINK_TRAIN_NORMAL_CPT; } else { @@ -607,7 +607,6 @@ static void gen6_fdi_link_train(struct intel_crtc *crtc, const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe = crtc->pipe; i915_reg_t reg; u32 temp, i, retry; @@ -647,7 +646,7 @@ static void gen6_fdi_link_train(struct intel_crtc *crtc, reg = FDI_RX_CTL(pipe); temp = intel_de_read(display, reg); - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT; temp |= FDI_LINK_TRAIN_PATTERN_1_CPT; } else { @@ -698,7 +697,7 @@ static void gen6_fdi_link_train(struct intel_crtc *crtc, reg = FDI_RX_CTL(pipe); temp = intel_de_read(display, reg); - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT; temp |= FDI_LINK_TRAIN_PATTERN_2_CPT; } else { @@ -1077,7 +1076,6 @@ void ilk_fdi_pll_disable(struct intel_crtc *crtc) void ilk_fdi_disable(struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; i915_reg_t reg; u32 temp; @@ -1096,7 +1094,7 @@ void ilk_fdi_disable(struct intel_crtc *crtc) udelay(100); /* Ironlake workaround, disable clock pointer after downing FDI */ - if (HAS_PCH_IBX(dev_priv)) + if (HAS_PCH_IBX(display)) intel_de_write(display, FDI_RX_CHICKEN(pipe), FDI_RX_PHASE_SYNC_POINTER_OVR); @@ -1106,7 +1104,7 @@ void ilk_fdi_disable(struct intel_crtc *crtc) reg = FDI_RX_CTL(pipe); temp = intel_de_read(display, reg); - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT; temp |= FDI_LINK_TRAIN_PATTERN_1_CPT; } else { diff --git a/drivers/gpu/drm/i915/display/intel_fifo_underrun.c b/drivers/gpu/drm/i915/display/intel_fifo_underrun.c index 451cd26024f7..2a787897b2d3 100644 --- a/drivers/gpu/drm/i915/display/intel_fifo_underrun.c +++ b/drivers/gpu/drm/i915/display/intel_fifo_underrun.c @@ -25,7 +25,8 @@ * */ -#include "i915_drv.h" +#include <drm/drm_print.h> + #include "i915_reg.h" #include "intel_de.h" #include "intel_display_irq.h" @@ -57,11 +58,10 @@ static bool ivb_can_enable_err_int(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_crtc *crtc; enum pipe pipe; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); for_each_pipe(display, pipe) { crtc = intel_crtc_for_pipe(display, pipe); @@ -75,11 +75,10 @@ static bool ivb_can_enable_err_int(struct intel_display *display) static bool cpt_can_enable_serr_int(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe pipe; struct intel_crtc *crtc; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); for_each_pipe(display, pipe) { crtc = intel_crtc_for_pipe(display, pipe); @@ -94,11 +93,10 @@ static bool cpt_can_enable_serr_int(struct intel_display *display) static void i9xx_check_fifo_underruns(struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); i915_reg_t reg = PIPESTAT(display, crtc->pipe); u32 enable_mask; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); if ((intel_de_read(display, reg) & PIPE_FIFO_UNDERRUN_STATUS) == 0) return; @@ -115,10 +113,9 @@ static void i9xx_set_fifo_underrun_reporting(struct intel_display *display, enum pipe pipe, bool enable, bool old) { - struct drm_i915_private *dev_priv = to_i915(display->drm); i915_reg_t reg = PIPESTAT(display, pipe); - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); if (enable) { u32 enable_mask = i915_pipestat_enable_mask(display, pipe); @@ -148,11 +145,10 @@ static void ilk_set_fifo_underrun_reporting(struct intel_display *display, static void ivb_check_fifo_underruns(struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; u32 err_int = intel_de_read(display, GEN7_ERR_INT); - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); if ((err_int & ERR_INT_FIFO_UNDERRUN(pipe)) == 0) return; @@ -213,11 +209,10 @@ static void ibx_set_fifo_underrun_reporting(struct intel_display *display, static void cpt_check_pch_fifo_underruns(struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pch_transcoder = crtc->pipe; u32 serr_int = intel_de_read(display, SERR_INT); - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); if ((serr_int & SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder)) == 0) return; @@ -258,11 +253,10 @@ static void cpt_set_fifo_underrun_reporting(struct intel_display *display, static bool __intel_set_cpu_fifo_underrun_reporting(struct intel_display *display, enum pipe pipe, bool enable) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_crtc *crtc = intel_crtc_for_pipe(display, pipe); bool old; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); old = !crtc->cpu_fifo_underrun_disabled; crtc->cpu_fifo_underrun_disabled = !enable; @@ -298,13 +292,12 @@ static bool __intel_set_cpu_fifo_underrun_reporting(struct intel_display *displa bool intel_set_cpu_fifo_underrun_reporting(struct intel_display *display, enum pipe pipe, bool enable) { - struct drm_i915_private *dev_priv = to_i915(display->drm); unsigned long flags; bool ret; - spin_lock_irqsave(&dev_priv->irq_lock, flags); + spin_lock_irqsave(&display->irq.lock, flags); ret = __intel_set_cpu_fifo_underrun_reporting(display, pipe, enable); - spin_unlock_irqrestore(&dev_priv->irq_lock, flags); + spin_unlock_irqrestore(&display->irq.lock, flags); return ret; } @@ -327,7 +320,6 @@ bool intel_set_pch_fifo_underrun_reporting(struct intel_display *display, enum pipe pch_transcoder, bool enable) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_crtc *crtc = intel_crtc_for_pipe(display, pch_transcoder); unsigned long flags; bool old; @@ -341,12 +333,12 @@ bool intel_set_pch_fifo_underrun_reporting(struct intel_display *display, * crtc on LPT won't cause issues. */ - spin_lock_irqsave(&dev_priv->irq_lock, flags); + spin_lock_irqsave(&display->irq.lock, flags); old = !crtc->pch_fifo_underrun_disabled; crtc->pch_fifo_underrun_disabled = !enable; - if (HAS_PCH_IBX(dev_priv)) + if (HAS_PCH_IBX(display)) ibx_set_fifo_underrun_reporting(display, pch_transcoder, enable); @@ -355,7 +347,7 @@ bool intel_set_pch_fifo_underrun_reporting(struct intel_display *display, pch_transcoder, enable, old); - spin_unlock_irqrestore(&dev_priv->irq_lock, flags); + spin_unlock_irqrestore(&display->irq.lock, flags); return old; } @@ -422,10 +414,9 @@ void intel_pch_fifo_underrun_irq_handler(struct intel_display *display, */ void intel_check_cpu_fifo_underruns(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_crtc *crtc; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); for_each_intel_crtc(display->drm, crtc) { if (crtc->cpu_fifo_underrun_disabled) @@ -437,7 +428,7 @@ void intel_check_cpu_fifo_underruns(struct intel_display *display) ivb_check_fifo_underruns(crtc); } - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } /** @@ -450,20 +441,19 @@ void intel_check_cpu_fifo_underruns(struct intel_display *display) */ void intel_check_pch_fifo_underruns(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_crtc *crtc; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); for_each_intel_crtc(display->drm, crtc) { if (crtc->pch_fifo_underrun_disabled) continue; - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) cpt_check_pch_fifo_underruns(crtc); } - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } void intel_init_fifo_underrun_reporting(struct intel_display *display, diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.c b/drivers/gpu/drm/i915/display/intel_frontbuffer.c index ba2f88ca6117..43be5377ddc1 100644 --- a/drivers/gpu/drm/i915/display/intel_frontbuffer.c +++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c @@ -58,7 +58,6 @@ #include <drm/drm_gem.h> #include "i915_active.h" -#include "i915_drv.h" #include "i915_vma.h" #include "intel_bo.h" #include "intel_display_trace.h" @@ -72,7 +71,7 @@ /** * frontbuffer_flush - flush frontbuffer - * @i915: i915 device + * @display: display device * @frontbuffer_bits: frontbuffer plane tracking bits * @origin: which operation caused the flush * @@ -82,16 +81,14 @@ * * Can be called without any locks held. */ -static void frontbuffer_flush(struct drm_i915_private *i915, +static void frontbuffer_flush(struct intel_display *display, unsigned int frontbuffer_bits, enum fb_op_origin origin) { - struct intel_display *display = &i915->display; - /* Delay flushing when rings are still busy.*/ - spin_lock(&i915->display.fb_tracking.lock); - frontbuffer_bits &= ~i915->display.fb_tracking.busy_bits; - spin_unlock(&i915->display.fb_tracking.lock); + spin_lock(&display->fb_tracking.lock); + frontbuffer_bits &= ~display->fb_tracking.busy_bits; + spin_unlock(&display->fb_tracking.lock); if (!frontbuffer_bits) return; @@ -107,7 +104,7 @@ static void frontbuffer_flush(struct drm_i915_private *i915, /** * intel_frontbuffer_flip_prepare - prepare asynchronous frontbuffer flip - * @i915: i915 device + * @display: display device * @frontbuffer_bits: frontbuffer plane tracking bits * * This function gets called after scheduling a flip on @obj. The actual @@ -117,19 +114,19 @@ static void frontbuffer_flush(struct drm_i915_private *i915, * * Can be called without any locks held. */ -void intel_frontbuffer_flip_prepare(struct drm_i915_private *i915, +void intel_frontbuffer_flip_prepare(struct intel_display *display, unsigned frontbuffer_bits) { - spin_lock(&i915->display.fb_tracking.lock); - i915->display.fb_tracking.flip_bits |= frontbuffer_bits; + spin_lock(&display->fb_tracking.lock); + display->fb_tracking.flip_bits |= frontbuffer_bits; /* Remove stale busy bits due to the old buffer. */ - i915->display.fb_tracking.busy_bits &= ~frontbuffer_bits; - spin_unlock(&i915->display.fb_tracking.lock); + display->fb_tracking.busy_bits &= ~frontbuffer_bits; + spin_unlock(&display->fb_tracking.lock); } /** * intel_frontbuffer_flip_complete - complete asynchronous frontbuffer flip - * @i915: i915 device + * @display: display device * @frontbuffer_bits: frontbuffer plane tracking bits * * This function gets called after the flip has been latched and will complete @@ -137,22 +134,22 @@ void intel_frontbuffer_flip_prepare(struct drm_i915_private *i915, * * Can be called without any locks held. */ -void intel_frontbuffer_flip_complete(struct drm_i915_private *i915, +void intel_frontbuffer_flip_complete(struct intel_display *display, unsigned frontbuffer_bits) { - spin_lock(&i915->display.fb_tracking.lock); + spin_lock(&display->fb_tracking.lock); /* Mask any cancelled flips. */ - frontbuffer_bits &= i915->display.fb_tracking.flip_bits; - i915->display.fb_tracking.flip_bits &= ~frontbuffer_bits; - spin_unlock(&i915->display.fb_tracking.lock); + frontbuffer_bits &= display->fb_tracking.flip_bits; + display->fb_tracking.flip_bits &= ~frontbuffer_bits; + spin_unlock(&display->fb_tracking.lock); if (frontbuffer_bits) - frontbuffer_flush(i915, frontbuffer_bits, ORIGIN_FLIP); + frontbuffer_flush(display, frontbuffer_bits, ORIGIN_FLIP); } /** * intel_frontbuffer_flip - synchronous frontbuffer flip - * @i915: i915 device + * @display: display device * @frontbuffer_bits: frontbuffer plane tracking bits * * This function gets called after scheduling a flip on @obj. This is for @@ -161,15 +158,15 @@ void intel_frontbuffer_flip_complete(struct drm_i915_private *i915, * * Can be called without any locks held. */ -void intel_frontbuffer_flip(struct drm_i915_private *i915, +void intel_frontbuffer_flip(struct intel_display *display, unsigned frontbuffer_bits) { - spin_lock(&i915->display.fb_tracking.lock); + spin_lock(&display->fb_tracking.lock); /* Remove stale busy bits due to the old buffer. */ - i915->display.fb_tracking.busy_bits &= ~frontbuffer_bits; - spin_unlock(&i915->display.fb_tracking.lock); + display->fb_tracking.busy_bits &= ~frontbuffer_bits; + spin_unlock(&display->fb_tracking.lock); - frontbuffer_flush(i915, frontbuffer_bits, ORIGIN_FLIP); + frontbuffer_flush(display, frontbuffer_bits, ORIGIN_FLIP); } void __intel_fb_invalidate(struct intel_frontbuffer *front, @@ -198,7 +195,6 @@ void __intel_fb_flush(struct intel_frontbuffer *front, unsigned int frontbuffer_bits) { struct intel_display *display = to_intel_display(front->obj->dev); - struct drm_i915_private *i915 = to_i915(display->drm); if (origin == ORIGIN_CS) { spin_lock(&display->fb_tracking.lock); @@ -209,7 +205,7 @@ void __intel_fb_flush(struct intel_frontbuffer *front, } if (frontbuffer_bits) - frontbuffer_flush(i915, frontbuffer_bits, origin); + frontbuffer_flush(display, frontbuffer_bits, origin); } static void intel_frontbuffer_flush_work(struct work_struct *work) @@ -280,7 +276,7 @@ static void frontbuffer_release(struct kref *ref) struct intel_frontbuffer * intel_frontbuffer_get(struct drm_gem_object *obj) { - struct drm_i915_private *i915 = to_i915(obj->dev); + struct intel_display *display = to_intel_display(obj->dev); struct intel_frontbuffer *front, *cur; front = intel_bo_get_frontbuffer(obj); @@ -300,9 +296,9 @@ intel_frontbuffer_get(struct drm_gem_object *obj) I915_ACTIVE_RETIRE_SLEEPS); INIT_WORK(&front->flush_work, intel_frontbuffer_flush_work); - spin_lock(&i915->display.fb_tracking.lock); + spin_lock(&display->fb_tracking.lock); cur = intel_bo_set_frontbuffer(obj, front); - spin_unlock(&i915->display.fb_tracking.lock); + spin_unlock(&display->fb_tracking.lock); if (cur != front) kfree(front); return cur; diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.h b/drivers/gpu/drm/i915/display/intel_frontbuffer.h index 6237780a9f68..2fee12eaf9b6 100644 --- a/drivers/gpu/drm/i915/display/intel_frontbuffer.h +++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.h @@ -31,7 +31,7 @@ #include "i915_active_types.h" struct drm_gem_object; -struct drm_i915_private; +struct intel_display; enum fb_op_origin { ORIGIN_CPU = 0, @@ -68,11 +68,11 @@ struct intel_frontbuffer { GENMASK(INTEL_FRONTBUFFER_BITS_PER_PIPE * ((pipe) + 1) - 1, \ INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe)) -void intel_frontbuffer_flip_prepare(struct drm_i915_private *i915, +void intel_frontbuffer_flip_prepare(struct intel_display *display, unsigned frontbuffer_bits); -void intel_frontbuffer_flip_complete(struct drm_i915_private *i915, +void intel_frontbuffer_flip_complete(struct intel_display *display, unsigned frontbuffer_bits); -void intel_frontbuffer_flip(struct drm_i915_private *i915, +void intel_frontbuffer_flip(struct intel_display *display, unsigned frontbuffer_bits); void intel_frontbuffer_put(struct intel_frontbuffer *front); diff --git a/drivers/gpu/drm/i915/display/intel_global_state.c b/drivers/gpu/drm/i915/display/intel_global_state.c index 8a49e2bb37fa..000a898c9480 100644 --- a/drivers/gpu/drm/i915/display/intel_global_state.c +++ b/drivers/gpu/drm/i915/display/intel_global_state.c @@ -3,10 +3,13 @@ * Copyright © 2020 Intel Corporation */ +#include <linux/pci.h> #include <linux/string.h> -#include "i915_drv.h" +#include <drm/drm_print.h> + #include "intel_atomic.h" +#include "intel_display_core.h" #include "intel_display_types.h" #include "intel_global_state.h" diff --git a/drivers/gpu/drm/i915/display/intel_gmbus.c b/drivers/gpu/drm/i915/display/intel_gmbus.c index abf457e68ee9..d55cc77650b7 100644 --- a/drivers/gpu/drm/i915/display/intel_gmbus.c +++ b/drivers/gpu/drm/i915/display/intel_gmbus.c @@ -152,32 +152,31 @@ static const struct gmbus_pin gmbus_pins_mtp[] = { static const struct gmbus_pin *get_gmbus_pin(struct intel_display *display, unsigned int pin) { - struct drm_i915_private *i915 = to_i915(display->drm); const struct gmbus_pin *pins; size_t size; - if (INTEL_PCH_TYPE(i915) >= PCH_MTL) { + if (INTEL_PCH_TYPE(display) >= PCH_MTL) { pins = gmbus_pins_mtp; size = ARRAY_SIZE(gmbus_pins_mtp); - } else if (INTEL_PCH_TYPE(i915) >= PCH_DG2) { + } else if (INTEL_PCH_TYPE(display) >= PCH_DG2) { pins = gmbus_pins_dg2; size = ARRAY_SIZE(gmbus_pins_dg2); - } else if (INTEL_PCH_TYPE(i915) >= PCH_DG1) { + } else if (INTEL_PCH_TYPE(display) >= PCH_DG1) { pins = gmbus_pins_dg1; size = ARRAY_SIZE(gmbus_pins_dg1); - } else if (INTEL_PCH_TYPE(i915) >= PCH_ICP) { + } else if (INTEL_PCH_TYPE(display) >= PCH_ICP) { pins = gmbus_pins_icp; size = ARRAY_SIZE(gmbus_pins_icp); - } else if (HAS_PCH_CNP(i915)) { + } else if (HAS_PCH_CNP(display)) { pins = gmbus_pins_cnp; size = ARRAY_SIZE(gmbus_pins_cnp); - } else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) { + } else if (display->platform.geminilake || display->platform.broxton) { pins = gmbus_pins_bxt; size = ARRAY_SIZE(gmbus_pins_bxt); } else if (DISPLAY_VER(display) == 9) { pins = gmbus_pins_skl; size = ARRAY_SIZE(gmbus_pins_skl); - } else if (IS_BROADWELL(i915)) { + } else if (display->platform.broadwell) { pins = gmbus_pins_bdw; size = ARRAY_SIZE(gmbus_pins_bdw); } else { @@ -240,11 +239,10 @@ static void bxt_gmbus_clock_gating(struct intel_display *display, static u32 get_reserved(struct intel_gmbus *bus) { struct intel_display *display = bus->display; - struct drm_i915_private *i915 = to_i915(display->drm); u32 reserved = 0; /* On most chips, these bits must be preserved in software. */ - if (!IS_I830(i915) && !IS_I845G(i915)) + if (!display->platform.i830 && !display->platform.i845g) reserved = intel_de_read_notrace(display, bus->gpio_reg) & (GPIO_DATA_PULLUP_DISABLE | GPIO_CLOCK_PULLUP_DISABLE); @@ -314,11 +312,10 @@ intel_gpio_pre_xfer(struct i2c_adapter *adapter) { struct intel_gmbus *bus = to_intel_gmbus(adapter); struct intel_display *display = bus->display; - struct drm_i915_private *i915 = to_i915(display->drm); intel_gmbus_reset(display); - if (IS_PINEVIEW(i915)) + if (display->platform.pineview) pnv_gmbus_clock_gating(display, false); set_data(bus, 1); @@ -332,12 +329,11 @@ intel_gpio_post_xfer(struct i2c_adapter *adapter) { struct intel_gmbus *bus = to_intel_gmbus(adapter); struct intel_display *display = bus->display; - struct drm_i915_private *i915 = to_i915(display->drm); set_data(bus, 1); set_clock(bus, 1); - if (IS_PINEVIEW(i915)) + if (display->platform.pineview) pnv_gmbus_clock_gating(display, true); } @@ -630,14 +626,13 @@ do_gmbus_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, int num, { struct intel_gmbus *bus = to_intel_gmbus(adapter); struct intel_display *display = bus->display; - struct drm_i915_private *i915 = to_i915(display->drm); int i = 0, inc, try = 0; int ret = 0; /* Display WA #0868: skl,bxt,kbl,cfl,glk */ - if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) + if (display->platform.geminilake || display->platform.broxton) bxt_gmbus_clock_gating(display, false); - else if (HAS_PCH_SPT(i915) || HAS_PCH_CNP(i915)) + else if (HAS_PCH_SPT(display) || HAS_PCH_CNP(display)) pch_gmbus_clock_gating(display, false); retry: @@ -748,9 +743,9 @@ timeout: out: /* Display WA #0868: skl,bxt,kbl,cfl,glk */ - if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) + if (display->platform.geminilake || display->platform.broxton) bxt_gmbus_clock_gating(display, true); - else if (HAS_PCH_SPT(i915) || HAS_PCH_CNP(i915)) + else if (HAS_PCH_SPT(display) || HAS_PCH_CNP(display)) pch_gmbus_clock_gating(display, true); return ret; @@ -873,12 +868,11 @@ static const struct i2c_lock_operations gmbus_lock_ops = { */ int intel_gmbus_setup(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); struct pci_dev *pdev = to_pci_dev(display->drm->dev); unsigned int pin; int ret; - if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) + if (display->platform.valleyview || display->platform.cherryview) display->gmbus.mmio_base = VLV_DISPLAY_BASE; else if (!HAS_GMCH(display)) /* @@ -925,7 +919,7 @@ int intel_gmbus_setup(struct intel_display *display) bus->reg0 = pin | GMBUS_RATE_100KHZ; /* gmbus seems to be broken on i830 */ - if (IS_I830(i915)) + if (display->platform.i830) bus->force_bit = 1; intel_gpio_setup(bus, GPIO(display, gmbus_pin->gpio)); diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c index 411f17655f89..3e3038f4ee1f 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp.c +++ b/drivers/gpu/drm/i915/display/intel_hdcp.c @@ -27,10 +27,13 @@ #include "intel_dp_mst.h" #include "intel_hdcp.h" #include "intel_hdcp_gsc.h" +#include "intel_hdcp_gsc_message.h" #include "intel_hdcp_regs.h" #include "intel_hdcp_shim.h" #include "intel_pcode.h" +#define USE_HDCP_GSC(__display) (DISPLAY_VER(__display) >= 14) + #define KEY_LOAD_TRIES 5 #define HDCP2_LC_RETRY_CNT 3 @@ -250,8 +253,8 @@ static bool intel_hdcp2_prerequisite(struct intel_connector *connector) return false; /* If MTL+ make sure gsc is loaded and proxy is setup */ - if (intel_hdcp_gsc_cs_required(display)) { - if (!intel_hdcp_gsc_check_status(display)) + if (USE_HDCP_GSC(display)) { + if (!intel_hdcp_gsc_check_status(display->drm)) return false; } @@ -2339,7 +2342,7 @@ static int initialize_hdcp_port_data(struct intel_connector *connector, static bool is_hdcp2_supported(struct intel_display *display) { - if (intel_hdcp_gsc_cs_required(display)) + if (USE_HDCP_GSC(display)) return true; if (!IS_ENABLED(CONFIG_INTEL_MEI_HDCP)) @@ -2363,7 +2366,7 @@ void intel_hdcp_component_init(struct intel_display *display) display->hdcp.comp_added = true; mutex_unlock(&display->hdcp.hdcp_mutex); - if (intel_hdcp_gsc_cs_required(display)) + if (USE_HDCP_GSC(display)) ret = intel_hdcp_gsc_init(display); else ret = component_add_typed(display->drm->dev, &i915_hdcp_ops, @@ -2638,7 +2641,7 @@ void intel_hdcp_component_fini(struct intel_display *display) display->hdcp.comp_added = false; mutex_unlock(&display->hdcp.hdcp_mutex); - if (intel_hdcp_gsc_cs_required(display)) + if (USE_HDCP_GSC(display)) intel_hdcp_gsc_fini(display); else component_del(display->drm->dev, &i915_hdcp_ops); diff --git a/drivers/gpu/drm/i915/display/intel_hdcp_gsc.c b/drivers/gpu/drm/i915/display/intel_hdcp_gsc.c index 55965844d829..6a22862d6be1 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp_gsc.c +++ b/drivers/gpu/drm/i915/display/intel_hdcp_gsc.c @@ -11,27 +11,22 @@ #include "i915_drv.h" #include "i915_utils.h" #include "intel_hdcp_gsc.h" -#include "intel_hdcp_gsc_message.h" -struct intel_hdcp_gsc_message { +struct intel_hdcp_gsc_context { + struct drm_i915_private *i915; struct i915_vma *vma; void *hdcp_cmd_in; void *hdcp_cmd_out; }; -bool intel_hdcp_gsc_cs_required(struct intel_display *display) +bool intel_hdcp_gsc_check_status(struct drm_device *drm) { - return DISPLAY_VER(display) >= 14; -} - -bool intel_hdcp_gsc_check_status(struct intel_display *display) -{ - struct drm_i915_private *i915 = to_i915(display->drm); + struct drm_i915_private *i915 = to_i915(drm); struct intel_gt *gt = i915->media_gt; struct intel_gsc_uc *gsc = gt ? >->uc.gsc : NULL; if (!gsc || !intel_uc_fw_is_running(&gsc->fw)) { - drm_dbg_kms(display->drm, + drm_dbg_kms(&i915->drm, "GSC components required for HDCP2.2 are not ready\n"); return false; } @@ -41,7 +36,7 @@ bool intel_hdcp_gsc_check_status(struct intel_display *display) /*This function helps allocate memory for the command that we will send to gsc cs */ static int intel_hdcp_gsc_initialize_message(struct drm_i915_private *i915, - struct intel_hdcp_gsc_message *hdcp_message) + struct intel_hdcp_gsc_context *gsc_context) { struct intel_gt *gt = i915->media_gt; struct drm_i915_gem_object *obj = NULL; @@ -78,9 +73,10 @@ static int intel_hdcp_gsc_initialize_message(struct drm_i915_private *i915, memset(cmd_in, 0, obj->base.size); - hdcp_message->hdcp_cmd_in = cmd_in; - hdcp_message->hdcp_cmd_out = cmd_out; - hdcp_message->vma = vma; + gsc_context->hdcp_cmd_in = cmd_in; + gsc_context->hdcp_cmd_out = cmd_out; + gsc_context->vma = vma; + gsc_context->i915 = i915; return 0; @@ -91,80 +87,37 @@ out_unpin: return err; } -static const struct i915_hdcp_ops gsc_hdcp_ops = { - .initiate_hdcp2_session = intel_hdcp_gsc_initiate_session, - .verify_receiver_cert_prepare_km = - intel_hdcp_gsc_verify_receiver_cert_prepare_km, - .verify_hprime = intel_hdcp_gsc_verify_hprime, - .store_pairing_info = intel_hdcp_gsc_store_pairing_info, - .initiate_locality_check = intel_hdcp_gsc_initiate_locality_check, - .verify_lprime = intel_hdcp_gsc_verify_lprime, - .get_session_key = intel_hdcp_gsc_get_session_key, - .repeater_check_flow_prepare_ack = - intel_hdcp_gsc_repeater_check_flow_prepare_ack, - .verify_mprime = intel_hdcp_gsc_verify_mprime, - .enable_hdcp_authentication = intel_hdcp_gsc_enable_authentication, - .close_hdcp_session = intel_hdcp_gsc_close_session, -}; - -static int intel_hdcp_gsc_hdcp2_init(struct intel_display *display) +struct intel_hdcp_gsc_context *intel_hdcp_gsc_context_alloc(struct drm_device *drm) { - struct drm_i915_private *i915 = to_i915(display->drm); - struct intel_hdcp_gsc_message *hdcp_message; + struct drm_i915_private *i915 = to_i915(drm); + struct intel_hdcp_gsc_context *gsc_context; int ret; - hdcp_message = kzalloc(sizeof(*hdcp_message), GFP_KERNEL); - - if (!hdcp_message) - return -ENOMEM; + gsc_context = kzalloc(sizeof(*gsc_context), GFP_KERNEL); + if (!gsc_context) + return ERR_PTR(-ENOMEM); /* * NOTE: No need to lock the comp mutex here as it is already * going to be taken before this function called */ - display->hdcp.hdcp_message = hdcp_message; - ret = intel_hdcp_gsc_initialize_message(i915, hdcp_message); - - if (ret) - drm_err(display->drm, "Could not initialize hdcp_message\n"); - - return ret; -} - -static void intel_hdcp_gsc_free_message(struct intel_display *display) -{ - struct intel_hdcp_gsc_message *hdcp_message = - display->hdcp.hdcp_message; + ret = intel_hdcp_gsc_initialize_message(i915, gsc_context); + if (ret) { + drm_err(&i915->drm, "Could not initialize gsc_context\n"); + kfree(gsc_context); + gsc_context = ERR_PTR(ret); + } - hdcp_message->hdcp_cmd_in = NULL; - hdcp_message->hdcp_cmd_out = NULL; - i915_vma_unpin_and_release(&hdcp_message->vma, I915_VMA_RELEASE_MAP); - kfree(hdcp_message); + return gsc_context; } -int intel_hdcp_gsc_init(struct intel_display *display) +void intel_hdcp_gsc_context_free(struct intel_hdcp_gsc_context *gsc_context) { - struct i915_hdcp_arbiter *data; - int ret; - - data = kzalloc(sizeof(struct i915_hdcp_arbiter), GFP_KERNEL); - if (!data) - return -ENOMEM; + if (!gsc_context) + return; - mutex_lock(&display->hdcp.hdcp_mutex); - display->hdcp.arbiter = data; - display->hdcp.arbiter->hdcp_dev = display->drm->dev; - display->hdcp.arbiter->ops = &gsc_hdcp_ops; - ret = intel_hdcp_gsc_hdcp2_init(display); - mutex_unlock(&display->hdcp.hdcp_mutex); - - return ret; -} - -void intel_hdcp_gsc_fini(struct intel_display *display) -{ - intel_hdcp_gsc_free_message(display); - kfree(display->hdcp.arbiter); + i915_vma_unpin_and_release(&gsc_context->vma, I915_VMA_RELEASE_MAP); + kfree(gsc_context); } static int intel_gsc_send_sync(struct drm_i915_private *i915, @@ -211,18 +164,18 @@ static int intel_gsc_send_sync(struct drm_i915_private *i915, /* * This function can now be used for sending requests and will also handle * receipt of reply messages hence no different function of message retrieval - * is required. We will initialize intel_hdcp_gsc_message structure then add + * is required. We will initialize intel_hdcp_gsc_context structure then add * gsc cs memory header as stated in specs after which the normal HDCP payload * will follow */ -ssize_t intel_hdcp_gsc_msg_send(struct drm_i915_private *i915, u8 *msg_in, - size_t msg_in_len, u8 *msg_out, - size_t msg_out_len) +ssize_t intel_hdcp_gsc_msg_send(struct intel_hdcp_gsc_context *gsc_context, + void *msg_in, size_t msg_in_len, + void *msg_out, size_t msg_out_len) { + struct drm_i915_private *i915 = gsc_context->i915; struct intel_gt *gt = i915->media_gt; struct intel_gsc_mtl_header *header_in, *header_out; const size_t max_msg_size = PAGE_SIZE - sizeof(*header_in); - struct intel_hdcp_gsc_message *hdcp_message; u64 addr_in, addr_out, host_session_id; u32 reply_size, msg_size_in, msg_size_out; int ret, tries = 0; @@ -235,10 +188,9 @@ ssize_t intel_hdcp_gsc_msg_send(struct drm_i915_private *i915, u8 *msg_in, msg_size_in = msg_in_len + sizeof(*header_in); msg_size_out = msg_out_len + sizeof(*header_out); - hdcp_message = i915->display.hdcp.hdcp_message; - header_in = hdcp_message->hdcp_cmd_in; - header_out = hdcp_message->hdcp_cmd_out; - addr_in = i915_ggtt_offset(hdcp_message->vma); + header_in = gsc_context->hdcp_cmd_in; + header_out = gsc_context->hdcp_cmd_out; + addr_in = i915_ggtt_offset(gsc_context->vma); addr_out = addr_in + PAGE_SIZE; memset(header_in, 0, msg_size_in); @@ -246,7 +198,7 @@ ssize_t intel_hdcp_gsc_msg_send(struct drm_i915_private *i915, u8 *msg_in, get_random_bytes(&host_session_id, sizeof(u64)); intel_gsc_uc_heci_cmd_emit_mtl_header(header_in, HECI_MEADDRESS_HDCP, msg_size_in, host_session_id); - memcpy(hdcp_message->hdcp_cmd_in + sizeof(*header_in), msg_in, msg_in_len); + memcpy(gsc_context->hdcp_cmd_in + sizeof(*header_in), msg_in, msg_in_len); /* * Keep sending request in case the pending bit is set no need to add @@ -280,7 +232,7 @@ ssize_t intel_hdcp_gsc_msg_send(struct drm_i915_private *i915, u8 *msg_in, reply_size, (u32)msg_out_len); } - memcpy(msg_out, hdcp_message->hdcp_cmd_out + sizeof(*header_out), msg_out_len); + memcpy(msg_out, gsc_context->hdcp_cmd_out + sizeof(*header_out), msg_out_len); err: return ret; diff --git a/drivers/gpu/drm/i915/display/intel_hdcp_gsc.h b/drivers/gpu/drm/i915/display/intel_hdcp_gsc.h index 5695a5e4f609..9305c14aaffe 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp_gsc.h +++ b/drivers/gpu/drm/i915/display/intel_hdcp_gsc.h @@ -6,19 +6,17 @@ #ifndef __INTEL_HDCP_GSC_H__ #define __INTEL_HDCP_GSC_H__ -#include <linux/err.h> #include <linux/types.h> -struct drm_i915_private; -struct intel_display; -struct intel_hdcp_gsc_message; +struct drm_device; +struct intel_hdcp_gsc_context; -bool intel_hdcp_gsc_cs_required(struct intel_display *display); -ssize_t intel_hdcp_gsc_msg_send(struct drm_i915_private *i915, u8 *msg_in, - size_t msg_in_len, u8 *msg_out, - size_t msg_out_len); -int intel_hdcp_gsc_init(struct intel_display *display); -void intel_hdcp_gsc_fini(struct intel_display *display); -bool intel_hdcp_gsc_check_status(struct intel_display *display); +ssize_t intel_hdcp_gsc_msg_send(struct intel_hdcp_gsc_context *gsc_context, + void *msg_in, size_t msg_in_len, + void *msg_out, size_t msg_out_len); +bool intel_hdcp_gsc_check_status(struct drm_device *drm); + +struct intel_hdcp_gsc_context *intel_hdcp_gsc_context_alloc(struct drm_device *drm); +void intel_hdcp_gsc_context_free(struct intel_hdcp_gsc_context *gsc_context); #endif /* __INTEL_HDCP_GCS_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c b/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c index 129104fa9b16..98967bb148e3 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c +++ b/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c @@ -4,20 +4,23 @@ */ #include <linux/err.h> + +#include <drm/drm_print.h> #include <drm/intel/i915_hdcp_interface.h> -#include "i915_drv.h" +#include "intel_display_core.h" #include "intel_display_types.h" +#include "intel_hdcp_gsc.h" #include "intel_hdcp_gsc_message.h" -int +static int intel_hdcp_gsc_initiate_session(struct device *dev, struct hdcp_port_data *data, struct hdcp2_ake_init *ake_data) { struct wired_cmd_initiate_hdcp2_session_in session_init_in = {}; struct wired_cmd_initiate_hdcp2_session_out session_init_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !ake_data) @@ -28,7 +31,7 @@ intel_hdcp_gsc_initiate_session(struct device *dev, struct hdcp_port_data *data, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; session_init_in.header.api_version = HDCP_API_VERSION; session_init_in.header.command_id = WIRED_INITIATE_HDCP2_SESSION; @@ -41,9 +44,9 @@ intel_hdcp_gsc_initiate_session(struct device *dev, struct hdcp_port_data *data, session_init_in.port.attached_transcoder = (u8)data->hdcp_transcoder; session_init_in.protocol = data->protocol; - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&session_init_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &session_init_in, sizeof(session_init_in), - (u8 *)&session_init_out, + &session_init_out, sizeof(session_init_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -64,7 +67,7 @@ intel_hdcp_gsc_initiate_session(struct device *dev, struct hdcp_port_data *data, return 0; } -int +static int intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, struct hdcp_port_data *data, struct hdcp2_ake_send_cert *rx_cert, @@ -75,8 +78,8 @@ intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, { struct wired_cmd_verify_receiver_cert_in verify_rxcert_in = {}; struct wired_cmd_verify_receiver_cert_out verify_rxcert_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !rx_cert || !km_stored || !ek_pub_km || !msg_sz) @@ -87,7 +90,7 @@ intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; verify_rxcert_in.header.api_version = HDCP_API_VERSION; verify_rxcert_in.header.command_id = WIRED_VERIFY_RECEIVER_CERT; @@ -103,9 +106,9 @@ intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, memcpy(verify_rxcert_in.r_rx, &rx_cert->r_rx, HDCP_2_2_RRX_LEN); memcpy(verify_rxcert_in.rx_caps, rx_cert->rx_caps, HDCP_2_2_RXCAPS_LEN); - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&verify_rxcert_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &verify_rxcert_in, sizeof(verify_rxcert_in), - (u8 *)&verify_rxcert_out, + &verify_rxcert_out, sizeof(verify_rxcert_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed: %zd\n", byte); @@ -134,14 +137,14 @@ intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, return 0; } -int +static int intel_hdcp_gsc_verify_hprime(struct device *dev, struct hdcp_port_data *data, struct hdcp2_ake_send_hprime *rx_hprime) { struct wired_cmd_ake_send_hprime_in send_hprime_in = {}; struct wired_cmd_ake_send_hprime_out send_hprime_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !rx_hprime) @@ -152,7 +155,7 @@ intel_hdcp_gsc_verify_hprime(struct device *dev, struct hdcp_port_data *data, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; send_hprime_in.header.api_version = HDCP_API_VERSION; send_hprime_in.header.command_id = WIRED_AKE_SEND_HPRIME; @@ -166,9 +169,9 @@ intel_hdcp_gsc_verify_hprime(struct device *dev, struct hdcp_port_data *data, memcpy(send_hprime_in.h_prime, rx_hprime->h_prime, HDCP_2_2_H_PRIME_LEN); - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&send_hprime_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &send_hprime_in, sizeof(send_hprime_in), - (u8 *)&send_hprime_out, + &send_hprime_out, sizeof(send_hprime_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -184,14 +187,14 @@ intel_hdcp_gsc_verify_hprime(struct device *dev, struct hdcp_port_data *data, return 0; } -int +static int intel_hdcp_gsc_store_pairing_info(struct device *dev, struct hdcp_port_data *data, struct hdcp2_ake_send_pairing_info *pairing_info) { struct wired_cmd_ake_send_pairing_info_in pairing_info_in = {}; struct wired_cmd_ake_send_pairing_info_out pairing_info_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !pairing_info) @@ -202,7 +205,7 @@ intel_hdcp_gsc_store_pairing_info(struct device *dev, struct hdcp_port_data *dat dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; pairing_info_in.header.api_version = HDCP_API_VERSION; pairing_info_in.header.command_id = WIRED_AKE_SEND_PAIRING_INFO; @@ -217,9 +220,9 @@ intel_hdcp_gsc_store_pairing_info(struct device *dev, struct hdcp_port_data *dat memcpy(pairing_info_in.e_kh_km, pairing_info->e_kh_km, HDCP_2_2_E_KH_KM_LEN); - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&pairing_info_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &pairing_info_in, sizeof(pairing_info_in), - (u8 *)&pairing_info_out, + &pairing_info_out, sizeof(pairing_info_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -236,15 +239,15 @@ intel_hdcp_gsc_store_pairing_info(struct device *dev, struct hdcp_port_data *dat return 0; } -int +static int intel_hdcp_gsc_initiate_locality_check(struct device *dev, struct hdcp_port_data *data, struct hdcp2_lc_init *lc_init_data) { struct wired_cmd_init_locality_check_in lc_init_in = {}; struct wired_cmd_init_locality_check_out lc_init_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !lc_init_data) @@ -255,7 +258,7 @@ intel_hdcp_gsc_initiate_locality_check(struct device *dev, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; lc_init_in.header.api_version = HDCP_API_VERSION; lc_init_in.header.command_id = WIRED_INIT_LOCALITY_CHECK; @@ -266,8 +269,8 @@ intel_hdcp_gsc_initiate_locality_check(struct device *dev, lc_init_in.port.physical_port = (u8)data->hdcp_ddi; lc_init_in.port.attached_transcoder = (u8)data->hdcp_transcoder; - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&lc_init_in, sizeof(lc_init_in), - (u8 *)&lc_init_out, sizeof(lc_init_out)); + byte = intel_hdcp_gsc_msg_send(gsc_context, &lc_init_in, sizeof(lc_init_in), + &lc_init_out, sizeof(lc_init_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); return byte; @@ -285,14 +288,14 @@ intel_hdcp_gsc_initiate_locality_check(struct device *dev, return 0; } -int +static int intel_hdcp_gsc_verify_lprime(struct device *dev, struct hdcp_port_data *data, struct hdcp2_lc_send_lprime *rx_lprime) { struct wired_cmd_validate_locality_in verify_lprime_in = {}; struct wired_cmd_validate_locality_out verify_lprime_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !rx_lprime) @@ -303,7 +306,7 @@ intel_hdcp_gsc_verify_lprime(struct device *dev, struct hdcp_port_data *data, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; verify_lprime_in.header.api_version = HDCP_API_VERSION; verify_lprime_in.header.command_id = WIRED_VALIDATE_LOCALITY; @@ -318,9 +321,9 @@ intel_hdcp_gsc_verify_lprime(struct device *dev, struct hdcp_port_data *data, memcpy(verify_lprime_in.l_prime, rx_lprime->l_prime, HDCP_2_2_L_PRIME_LEN); - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&verify_lprime_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &verify_lprime_in, sizeof(verify_lprime_in), - (u8 *)&verify_lprime_out, + &verify_lprime_out, sizeof(verify_lprime_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -337,14 +340,15 @@ intel_hdcp_gsc_verify_lprime(struct device *dev, struct hdcp_port_data *data, return 0; } -int intel_hdcp_gsc_get_session_key(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_ske_send_eks *ske_data) +static int +intel_hdcp_gsc_get_session_key(struct device *dev, + struct hdcp_port_data *data, + struct hdcp2_ske_send_eks *ske_data) { struct wired_cmd_get_session_key_in get_skey_in = {}; struct wired_cmd_get_session_key_out get_skey_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !ske_data) @@ -355,7 +359,7 @@ int intel_hdcp_gsc_get_session_key(struct device *dev, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; get_skey_in.header.api_version = HDCP_API_VERSION; get_skey_in.header.command_id = WIRED_GET_SESSION_KEY; @@ -366,8 +370,8 @@ int intel_hdcp_gsc_get_session_key(struct device *dev, get_skey_in.port.physical_port = (u8)data->hdcp_ddi; get_skey_in.port.attached_transcoder = (u8)data->hdcp_transcoder; - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&get_skey_in, sizeof(get_skey_in), - (u8 *)&get_skey_out, sizeof(get_skey_out)); + byte = intel_hdcp_gsc_msg_send(gsc_context, &get_skey_in, sizeof(get_skey_in), + &get_skey_out, sizeof(get_skey_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); return byte; @@ -387,7 +391,7 @@ int intel_hdcp_gsc_get_session_key(struct device *dev, return 0; } -int +static int intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, struct hdcp_port_data *data, struct hdcp2_rep_send_receiverid_list @@ -397,8 +401,8 @@ intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, { struct wired_cmd_verify_repeater_in verify_repeater_in = {}; struct wired_cmd_verify_repeater_out verify_repeater_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !rep_topology || !rep_send_ack || !data) @@ -409,7 +413,7 @@ intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; verify_repeater_in.header.api_version = HDCP_API_VERSION; verify_repeater_in.header.command_id = WIRED_VERIFY_REPEATER; @@ -430,9 +434,9 @@ intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, memcpy(verify_repeater_in.receiver_ids, rep_topology->receiver_ids, HDCP_2_2_RECEIVER_IDS_MAX_LEN); - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&verify_repeater_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &verify_repeater_in, sizeof(verify_repeater_in), - (u8 *)&verify_repeater_out, + &verify_repeater_out, sizeof(verify_repeater_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -453,14 +457,15 @@ intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, return 0; } -int intel_hdcp_gsc_verify_mprime(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_rep_stream_ready *stream_ready) +static int +intel_hdcp_gsc_verify_mprime(struct device *dev, + struct hdcp_port_data *data, + struct hdcp2_rep_stream_ready *stream_ready) { struct wired_cmd_repeater_auth_stream_req_in *verify_mprime_in; struct wired_cmd_repeater_auth_stream_req_out verify_mprime_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; size_t cmd_size; @@ -472,7 +477,7 @@ int intel_hdcp_gsc_verify_mprime(struct device *dev, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; cmd_size = struct_size(verify_mprime_in, streams, data->k); if (cmd_size == SIZE_MAX) @@ -499,8 +504,8 @@ int intel_hdcp_gsc_verify_mprime(struct device *dev, verify_mprime_in->k = cpu_to_be16(data->k); - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)verify_mprime_in, cmd_size, - (u8 *)&verify_mprime_out, + byte = intel_hdcp_gsc_msg_send(gsc_context, verify_mprime_in, cmd_size, + &verify_mprime_out, sizeof(verify_mprime_out)); kfree(verify_mprime_in); if (byte < 0) { @@ -518,13 +523,13 @@ int intel_hdcp_gsc_verify_mprime(struct device *dev, return 0; } -int intel_hdcp_gsc_enable_authentication(struct device *dev, - struct hdcp_port_data *data) +static int intel_hdcp_gsc_enable_authentication(struct device *dev, + struct hdcp_port_data *data) { struct wired_cmd_enable_auth_in enable_auth_in = {}; struct wired_cmd_enable_auth_out enable_auth_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data) @@ -535,7 +540,7 @@ int intel_hdcp_gsc_enable_authentication(struct device *dev, dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; enable_auth_in.header.api_version = HDCP_API_VERSION; enable_auth_in.header.command_id = WIRED_ENABLE_AUTH; @@ -547,9 +552,9 @@ int intel_hdcp_gsc_enable_authentication(struct device *dev, enable_auth_in.port.attached_transcoder = (u8)data->hdcp_transcoder; enable_auth_in.stream_type = data->streams[0].stream_type; - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&enable_auth_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &enable_auth_in, sizeof(enable_auth_in), - (u8 *)&enable_auth_out, + &enable_auth_out, sizeof(enable_auth_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -565,13 +570,13 @@ int intel_hdcp_gsc_enable_authentication(struct device *dev, return 0; } -int +static int intel_hdcp_gsc_close_session(struct device *dev, struct hdcp_port_data *data) { struct wired_cmd_close_session_in session_close_in = {}; struct wired_cmd_close_session_out session_close_out = {}; + struct intel_hdcp_gsc_context *gsc_context; struct intel_display *display; - struct drm_i915_private *i915; ssize_t byte; if (!dev || !data) @@ -582,7 +587,7 @@ intel_hdcp_gsc_close_session(struct device *dev, struct hdcp_port_data *data) dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } - i915 = to_i915(display->drm); + gsc_context = display->hdcp.gsc_context; session_close_in.header.api_version = HDCP_API_VERSION; session_close_in.header.command_id = WIRED_CLOSE_SESSION; @@ -594,9 +599,9 @@ intel_hdcp_gsc_close_session(struct device *dev, struct hdcp_port_data *data) session_close_in.port.physical_port = (u8)data->hdcp_ddi; session_close_in.port.attached_transcoder = (u8)data->hdcp_transcoder; - byte = intel_hdcp_gsc_msg_send(i915, (u8 *)&session_close_in, + byte = intel_hdcp_gsc_msg_send(gsc_context, &session_close_in, sizeof(session_close_in), - (u8 *)&session_close_out, + &session_close_out, sizeof(session_close_out)); if (byte < 0) { drm_dbg_kms(display->drm, "intel_hdcp_gsc_msg_send failed. %zd\n", byte); @@ -611,3 +616,57 @@ intel_hdcp_gsc_close_session(struct device *dev, struct hdcp_port_data *data) return 0; } + +static const struct i915_hdcp_ops gsc_hdcp_ops = { + .initiate_hdcp2_session = intel_hdcp_gsc_initiate_session, + .verify_receiver_cert_prepare_km = + intel_hdcp_gsc_verify_receiver_cert_prepare_km, + .verify_hprime = intel_hdcp_gsc_verify_hprime, + .store_pairing_info = intel_hdcp_gsc_store_pairing_info, + .initiate_locality_check = intel_hdcp_gsc_initiate_locality_check, + .verify_lprime = intel_hdcp_gsc_verify_lprime, + .get_session_key = intel_hdcp_gsc_get_session_key, + .repeater_check_flow_prepare_ack = + intel_hdcp_gsc_repeater_check_flow_prepare_ack, + .verify_mprime = intel_hdcp_gsc_verify_mprime, + .enable_hdcp_authentication = intel_hdcp_gsc_enable_authentication, + .close_hdcp_session = intel_hdcp_gsc_close_session, +}; + +int intel_hdcp_gsc_init(struct intel_display *display) +{ + struct intel_hdcp_gsc_context *gsc_context; + struct i915_hdcp_arbiter *arbiter; + int ret = 0; + + arbiter = kzalloc(sizeof(*arbiter), GFP_KERNEL); + if (!arbiter) + return -ENOMEM; + + mutex_lock(&display->hdcp.hdcp_mutex); + + gsc_context = intel_hdcp_gsc_context_alloc(display->drm); + if (IS_ERR(gsc_context)) { + ret = PTR_ERR(gsc_context); + kfree(arbiter); + goto out; + } + + display->hdcp.arbiter = arbiter; + display->hdcp.arbiter->hdcp_dev = display->drm->dev; + display->hdcp.arbiter->ops = &gsc_hdcp_ops; + display->hdcp.gsc_context = gsc_context; + +out: + mutex_unlock(&display->hdcp.hdcp_mutex); + + return ret; +} + +void intel_hdcp_gsc_fini(struct intel_display *display) +{ + intel_hdcp_gsc_context_free(display->hdcp.gsc_context); + display->hdcp.gsc_context = NULL; + kfree(display->hdcp.arbiter); + display->hdcp.arbiter = NULL; +} diff --git a/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.h b/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.h index 2d597f27e931..9f54157a4a3e 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.h +++ b/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.h @@ -6,68 +6,9 @@ #ifndef __INTEL_HDCP_GSC_MESSAGE_H__ #define __INTEL_HDCP_GSC_MESSAGE_H__ -#include <linux/types.h> - -struct device; -struct drm_i915_private; -struct hdcp_port_data; -struct hdcp2_ake_init; -struct hdcp2_ake_send_cert; -struct hdcp2_ake_no_stored_km; -struct hdcp2_ake_send_hprime; -struct hdcp2_ake_send_pairing_info; -struct hdcp2_lc_init; -struct hdcp2_lc_send_lprime; -struct hdcp2_ske_send_eks; -struct hdcp2_rep_send_receiverid_list; -struct hdcp2_rep_send_ack; -struct hdcp2_rep_stream_ready; struct intel_display; -ssize_t intel_hdcp_gsc_msg_send(struct drm_i915_private *i915, u8 *msg_in, - size_t msg_in_len, u8 *msg_out, - size_t msg_out_len); -bool intel_hdcp_gsc_check_status(struct intel_display *display); -int -intel_hdcp_gsc_initiate_session(struct device *dev, struct hdcp_port_data *data, - struct hdcp2_ake_init *ake_data); -int -intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_ake_send_cert *rx_cert, - bool *km_stored, - struct hdcp2_ake_no_stored_km - *ek_pub_km, - size_t *msg_sz); -int -intel_hdcp_gsc_verify_hprime(struct device *dev, struct hdcp_port_data *data, - struct hdcp2_ake_send_hprime *rx_hprime); -int -intel_hdcp_gsc_store_pairing_info(struct device *dev, struct hdcp_port_data *data, - struct hdcp2_ake_send_pairing_info *pairing_info); -int -intel_hdcp_gsc_initiate_locality_check(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_lc_init *lc_init_data); -int -intel_hdcp_gsc_verify_lprime(struct device *dev, struct hdcp_port_data *data, - struct hdcp2_lc_send_lprime *rx_lprime); -int intel_hdcp_gsc_get_session_key(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_ske_send_eks *ske_data); -int -intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_rep_send_receiverid_list - *rep_topology, - struct hdcp2_rep_send_ack - *rep_send_ack); -int intel_hdcp_gsc_verify_mprime(struct device *dev, - struct hdcp_port_data *data, - struct hdcp2_rep_stream_ready *stream_ready); -int intel_hdcp_gsc_enable_authentication(struct device *dev, - struct hdcp_port_data *data); -int -intel_hdcp_gsc_close_session(struct device *dev, struct hdcp_port_data *data); +int intel_hdcp_gsc_init(struct intel_display *display); +void intel_hdcp_gsc_fini(struct intel_display *display); #endif /* __INTEL_HDCP_GSC_MESSAGE_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c index f9fa17e1f584..98033471902c 100644 --- a/drivers/gpu/drm/i915/display/intel_hdmi.c +++ b/drivers/gpu/drm/i915/display/intel_hdmi.c @@ -38,14 +38,15 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_crtc.h> #include <drm/drm_edid.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> #include <drm/intel/intel_lpe_audio.h> #include <media/cec-notifier.h> #include "g4x_hdmi.h" -#include "i915_drv.h" #include "i915_reg.h" +#include "i915_utils.h" #include "intel_atomic.h" #include "intel_audio.h" #include "intel_connector.h" @@ -715,7 +716,7 @@ intel_hdmi_compute_avi_infoframe(struct intel_encoder *encoder, struct hdmi_avi_infoframe *frame = &crtc_state->infoframes.avi.avi; const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; - struct drm_connector *connector = conn_state->connector; + struct intel_connector *connector = to_intel_connector(conn_state->connector); int ret; if (!crtc_state->has_infoframe) @@ -724,7 +725,7 @@ intel_hdmi_compute_avi_infoframe(struct intel_encoder *encoder, crtc_state->infoframes.enable |= intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI); - ret = drm_hdmi_avi_infoframe_from_display_mode(frame, connector, + ret = drm_hdmi_avi_infoframe_from_display_mode(frame, &connector->base, adjusted_mode); if (ret) return false; @@ -743,7 +744,7 @@ intel_hdmi_compute_avi_infoframe(struct intel_encoder *encoder, crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB); if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_RGB) { - drm_hdmi_avi_infoframe_quant_range(frame, connector, + drm_hdmi_avi_infoframe_quant_range(frame, &connector->base, adjusted_mode, crtc_state->limited_color_range ? HDMI_QUANTIZATION_RANGE_LIMITED : @@ -769,7 +770,7 @@ intel_hdmi_compute_spd_infoframe(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state, struct drm_connector_state *conn_state) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(crtc_state); struct hdmi_spd_infoframe *frame = &crtc_state->infoframes.spd.spd; int ret; @@ -779,7 +780,7 @@ intel_hdmi_compute_spd_infoframe(struct intel_encoder *encoder, crtc_state->infoframes.enable |= intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_SPD); - if (IS_DGFX(i915)) + if (display->platform.dgfx) ret = hdmi_spd_infoframe_init(frame, "Intel", "Discrete gfx"); else ret = hdmi_spd_infoframe_init(frame, "Intel", "Integrated gfx"); @@ -979,7 +980,6 @@ static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder, const struct drm_connector_state *conn_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg; @@ -989,9 +989,9 @@ static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder, if (HAS_DDI(display)) reg = HSW_TVIDEO_DIP_GCP(display, crtc_state->cpu_transcoder); - else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + else if (display->platform.valleyview || display->platform.cherryview) reg = VLV_TVIDEO_DIP_GCP(crtc->pipe); - else if (HAS_PCH_SPLIT(dev_priv)) + else if (HAS_PCH_SPLIT(display)) reg = TVIDEO_DIP_GCP(crtc->pipe); else return false; @@ -1005,7 +1005,6 @@ void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg; @@ -1015,9 +1014,9 @@ void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder, if (HAS_DDI(display)) reg = HSW_TVIDEO_DIP_GCP(display, crtc_state->cpu_transcoder); - else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + else if (display->platform.valleyview || display->platform.cherryview) reg = VLV_TVIDEO_DIP_GCP(crtc->pipe); - else if (HAS_PCH_SPLIT(dev_priv)) + else if (HAS_PCH_SPLIT(display)) reg = TVIDEO_DIP_GCP(crtc->pipe); else return; @@ -1029,9 +1028,9 @@ static void intel_hdmi_compute_gcp_infoframe(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state, struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); - if (IS_G4X(dev_priv) || !crtc_state->has_infoframe) + if (display->platform.g4x || !crtc_state->has_infoframe) return; crtc_state->infoframes.enable |= @@ -1539,7 +1538,6 @@ int intel_hdmi_hdcp_toggle_signalling(struct intel_digital_port *dig_port, struct intel_display *display = to_intel_display(dig_port); struct intel_hdmi *hdmi = &dig_port->hdmi; struct intel_connector *connector = hdmi->attached_connector; - struct drm_i915_private *dev_priv = to_i915(connector->base.dev); int ret; if (!enable) @@ -1558,7 +1556,7 @@ int intel_hdmi_hdcp_toggle_signalling(struct intel_digital_port *dig_port, * WA: To fix incorrect positioning of the window of * opportunity and enc_en signalling in KABYLAKE. */ - if (IS_KABYLAKE(dev_priv) && enable) + if (display->platform.kabylake && enable) return kbl_repositioning_enc_en_signal(connector, cpu_transcoder); @@ -1570,7 +1568,6 @@ bool intel_hdmi_hdcp_check_link_once(struct intel_digital_port *dig_port, struct intel_connector *connector) { struct intel_display *display = to_intel_display(dig_port); - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); enum port port = dig_port->base.port; enum transcoder cpu_transcoder = connector->hdcp.cpu_transcoder; int ret; @@ -1583,15 +1580,15 @@ bool intel_hdmi_hdcp_check_link_once(struct intel_digital_port *dig_port, if (ret) return false; - intel_de_write(i915, HDCP_RPRIME(i915, cpu_transcoder, port), ri.reg); + intel_de_write(display, HDCP_RPRIME(display, cpu_transcoder, port), ri.reg); /* Wait for Ri prime match */ - if (wait_for((intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder, port)) & + if (wait_for((intel_de_read(display, HDCP_STATUS(display, cpu_transcoder, port)) & (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC)) == (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC), 1)) { drm_dbg_kms(display->drm, "Ri' mismatch detected (%x)\n", - intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder, - port))); + intel_de_read(display, HDCP_STATUS(display, cpu_transcoder, + port))); return false; } return true; @@ -1814,14 +1811,13 @@ static const struct intel_hdcp_shim intel_hdmi_hdcp_shim = { static int intel_hdmi_source_max_tmds_clock(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); int max_tmds_clock, vbt_max_tmds_clock; - if (DISPLAY_VER(display) >= 13 || IS_ALDERLAKE_S(dev_priv)) + if (DISPLAY_VER(display) >= 13 || display->platform.alderlake_s) max_tmds_clock = 600000; else if (DISPLAY_VER(display) >= 10) max_tmds_clock = 594000; - else if (DISPLAY_VER(display) >= 8 || IS_HASWELL(dev_priv)) + else if (DISPLAY_VER(display) >= 8 || display->platform.haswell) max_tmds_clock = 300000; else if (DISPLAY_VER(display) >= 5) max_tmds_clock = 225000; @@ -1880,7 +1876,6 @@ hdmi_port_clock_valid(struct intel_hdmi *hdmi, bool has_hdmi_sink) { struct intel_display *display = to_intel_display(hdmi); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder = &hdmi_to_dig_port(hdmi)->base; if (clock < 25000) @@ -1890,16 +1885,16 @@ hdmi_port_clock_valid(struct intel_hdmi *hdmi, return MODE_CLOCK_HIGH; /* GLK DPLL can't generate 446-480 MHz */ - if (IS_GEMINILAKE(dev_priv) && clock > 446666 && clock < 480000) + if (display->platform.geminilake && clock > 446666 && clock < 480000) return MODE_CLOCK_RANGE; /* BXT/GLK DPLL can't generate 223-240 MHz */ - if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) && + if ((display->platform.geminilake || display->platform.broxton) && clock > 223333 && clock < 240000) return MODE_CLOCK_RANGE; /* CHV DPLL can't generate 216-240 MHz */ - if (IS_CHERRYVIEW(dev_priv) && clock > 216000 && clock < 240000) + if (display->platform.cherryview && clock > 216000 && clock < 240000) return MODE_CLOCK_RANGE; /* ICL+ combo PHY PLL can't generate 500-533.2 MHz */ @@ -1943,11 +1938,12 @@ static bool intel_hdmi_source_bpc_possible(struct intel_display *display, int bp } } -static bool intel_hdmi_sink_bpc_possible(struct drm_connector *connector, +static bool intel_hdmi_sink_bpc_possible(struct drm_connector *_connector, int bpc, bool has_hdmi_sink, enum intel_output_format sink_format) { - const struct drm_display_info *info = &connector->display_info; + struct intel_connector *connector = to_intel_connector(_connector); + const struct drm_display_info *info = &connector->base.display_info; const struct drm_hdmi_info *hdmi = &info->hdmi; switch (bpc) { @@ -1976,12 +1972,13 @@ static bool intel_hdmi_sink_bpc_possible(struct drm_connector *connector, } static enum drm_mode_status -intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock, +intel_hdmi_mode_clock_valid(struct drm_connector *_connector, int clock, bool has_hdmi_sink, enum intel_output_format sink_format) { - struct intel_display *display = to_intel_display(connector->dev); - struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_hdmi *hdmi = intel_attached_hdmi(connector); enum drm_mode_status status = MODE_OK; int bpc; @@ -1996,7 +1993,8 @@ intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock, if (!intel_hdmi_source_bpc_possible(display, bpc)) continue; - if (!intel_hdmi_sink_bpc_possible(connector, bpc, has_hdmi_sink, sink_format)) + if (!intel_hdmi_sink_bpc_possible(&connector->base, bpc, has_hdmi_sink, + sink_format)) continue; status = hdmi_port_clock_valid(hdmi, tmds_clock, true, has_hdmi_sink); @@ -2011,15 +2009,16 @@ intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock, } static enum drm_mode_status -intel_hdmi_mode_valid(struct drm_connector *connector, +intel_hdmi_mode_valid(struct drm_connector *_connector, const struct drm_display_mode *mode) { - struct intel_display *display = to_intel_display(connector->dev); - struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_hdmi *hdmi = intel_attached_hdmi(connector); enum drm_mode_status status; int clock = mode->clock; - int max_dotclk = to_i915(connector->dev)->display.cdclk.max_dotclk_freq; - bool has_hdmi_sink = intel_has_hdmi_sink(hdmi, connector->state); + int max_dotclk = display->cdclk.max_dotclk_freq; + bool has_hdmi_sink = intel_has_hdmi_sink(hdmi, connector->base.state); bool ycbcr_420_only; enum intel_output_format sink_format; @@ -2048,22 +2047,23 @@ intel_hdmi_mode_valid(struct drm_connector *connector, if (clock > 600000) return MODE_CLOCK_HIGH; - ycbcr_420_only = drm_mode_is_420_only(&connector->display_info, mode); + ycbcr_420_only = drm_mode_is_420_only(&connector->base.display_info, mode); if (ycbcr_420_only) sink_format = INTEL_OUTPUT_FORMAT_YCBCR420; else sink_format = INTEL_OUTPUT_FORMAT_RGB; - status = intel_hdmi_mode_clock_valid(connector, clock, has_hdmi_sink, sink_format); + status = intel_hdmi_mode_clock_valid(&connector->base, clock, has_hdmi_sink, sink_format); if (status != MODE_OK) { if (ycbcr_420_only || - !connector->ycbcr_420_allowed || - !drm_mode_is_420_also(&connector->display_info, mode)) + !connector->base.ycbcr_420_allowed || + !drm_mode_is_420_also(&connector->base.display_info, mode)) return status; sink_format = INTEL_OUTPUT_FORMAT_YCBCR420; - status = intel_hdmi_mode_clock_valid(connector, clock, has_hdmi_sink, sink_format); + status = intel_hdmi_mode_clock_valid(&connector->base, clock, has_hdmi_sink, + sink_format); if (status != MODE_OK) return status; } @@ -2074,16 +2074,16 @@ intel_hdmi_mode_valid(struct drm_connector *connector, bool intel_hdmi_bpc_possible(const struct intel_crtc_state *crtc_state, int bpc, bool has_hdmi_sink) { - struct drm_atomic_state *state = crtc_state->uapi.state; - struct drm_connector_state *connector_state; - struct drm_connector *connector; + struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state); + struct intel_digital_connector_state *connector_state; + struct intel_connector *connector; int i; - for_each_new_connector_in_state(state, connector, connector_state, i) { - if (connector_state->crtc != crtc_state->uapi.crtc) + for_each_new_intel_connector_in_state(state, connector, connector_state, i) { + if (connector_state->base.crtc != crtc_state->uapi.crtc) continue; - if (!intel_hdmi_sink_bpc_possible(connector, bpc, has_hdmi_sink, + if (!intel_hdmi_sink_bpc_possible(&connector->base, bpc, has_hdmi_sink, crtc_state->sink_format)) return false; } @@ -2211,7 +2211,7 @@ static bool intel_hdmi_has_audio(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_connector *connector = conn_state->connector; + struct intel_connector *connector = to_intel_connector(conn_state->connector); const struct intel_digital_connector_state *intel_conn_state = to_intel_digital_connector_state(conn_state); @@ -2219,7 +2219,7 @@ static bool intel_hdmi_has_audio(struct intel_encoder *encoder, return false; if (intel_conn_state->force_audio == HDMI_AUDIO_AUTO) - return connector->display_info.has_audio; + return connector->base.display_info.has_audio; else return intel_conn_state->force_audio == HDMI_AUDIO_ON; } @@ -2323,14 +2323,14 @@ int intel_hdmi_compute_config(struct intel_encoder *encoder, { struct intel_display *display = to_intel_display(encoder); struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode; - struct drm_connector *connector = conn_state->connector; - struct drm_scdc *scdc = &connector->display_info.hdmi.scdc; + struct intel_connector *connector = to_intel_connector(conn_state->connector); + struct drm_scdc *scdc = &connector->base.display_info.hdmi.scdc; int ret; if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN) return -EINVAL; - if (!connector->interlace_allowed && + if (!connector->base.interlace_allowed && adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) return -EINVAL; @@ -2425,25 +2425,26 @@ void intel_hdmi_encoder_shutdown(struct intel_encoder *encoder) } static void -intel_hdmi_unset_edid(struct drm_connector *connector) +intel_hdmi_unset_edid(struct drm_connector *_connector) { - struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector); intel_hdmi->dp_dual_mode.type = DRM_DP_DUAL_MODE_NONE; intel_hdmi->dp_dual_mode.max_tmds_clock = 0; - drm_edid_free(to_intel_connector(connector)->detect_edid); - to_intel_connector(connector)->detect_edid = NULL; + drm_edid_free(connector->detect_edid); + connector->detect_edid = NULL; } static void -intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) +intel_hdmi_dp_dual_mode_detect(struct drm_connector *_connector) { - struct intel_display *display = to_intel_display(connector->dev); - struct drm_i915_private *dev_priv = to_i915(connector->dev); - struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector)); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_hdmi *hdmi = intel_attached_hdmi(connector); struct intel_encoder *encoder = &hdmi_to_dig_port(hdmi)->base; - struct i2c_adapter *ddc = connector->ddc; + struct i2c_adapter *ddc = connector->base.ddc; enum drm_dp_dual_mode_type type; type = drm_dp_dual_mode_detect(display->drm, ddc); @@ -2458,7 +2459,7 @@ intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) * if the port is a dual mode capable DP port. */ if (type == DRM_DP_DUAL_MODE_UNKNOWN) { - if (!connector->force && + if (!connector->base.force && intel_bios_encoder_supports_dp_dual_mode(encoder->devdata)) { drm_dbg_kms(display->drm, "Assuming DP dual mode adaptor presence based on VBT\n"); @@ -2481,7 +2482,7 @@ intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) hdmi->dp_dual_mode.max_tmds_clock); /* Older VBTs are often buggy and can't be trusted :( Play it safe. */ - if ((DISPLAY_VER(display) >= 8 || IS_HASWELL(dev_priv)) && + if ((DISPLAY_VER(display) >= 8 || display->platform.haswell) && !intel_bios_encoder_supports_dp_dual_mode(encoder->devdata)) { drm_dbg_kms(display->drm, "Ignoring DP dual mode adaptor max TMDS clock for native HDMI port\n"); @@ -2490,34 +2491,35 @@ intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) } static bool -intel_hdmi_set_edid(struct drm_connector *connector) +intel_hdmi_set_edid(struct drm_connector *_connector) { - struct intel_display *display = to_intel_display(connector->dev); - struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector)); - struct i2c_adapter *ddc = connector->ddc; + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); + struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector); + struct i2c_adapter *ddc = connector->base.ddc; intel_wakeref_t wakeref; const struct drm_edid *drm_edid; bool connected = false; wakeref = intel_display_power_get(display, POWER_DOMAIN_GMBUS); - drm_edid = drm_edid_read_ddc(connector, ddc); + drm_edid = drm_edid_read_ddc(&connector->base, ddc); if (!drm_edid && !intel_gmbus_is_forced_bit(ddc)) { drm_dbg_kms(display->drm, "HDMI GMBUS EDID read failed, retry using GPIO bit-banging\n"); intel_gmbus_force_bit(ddc, true); - drm_edid = drm_edid_read_ddc(connector, ddc); + drm_edid = drm_edid_read_ddc(&connector->base, ddc); intel_gmbus_force_bit(ddc, false); } /* Below we depend on display info having been updated */ - drm_edid_connector_update(connector, drm_edid); + drm_edid_connector_update(&connector->base, drm_edid); - to_intel_connector(connector)->detect_edid = drm_edid; + connector->detect_edid = drm_edid; if (drm_edid_is_digital(drm_edid)) { - intel_hdmi_dp_dual_mode_detect(connector); + intel_hdmi_dp_dual_mode_detect(&connector->base); connected = true; } @@ -2525,28 +2527,29 @@ intel_hdmi_set_edid(struct drm_connector *connector) intel_display_power_put(display, POWER_DOMAIN_GMBUS, wakeref); cec_notifier_set_phys_addr(intel_hdmi->cec_notifier, - connector->display_info.source_physical_address); + connector->base.display_info.source_physical_address); return connected; } static enum drm_connector_status -intel_hdmi_detect(struct drm_connector *connector, bool force) +intel_hdmi_detect(struct drm_connector *_connector, bool force) { - struct intel_display *display = to_intel_display(connector->dev); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); enum drm_connector_status status = connector_status_disconnected; - struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector)); + struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector); struct intel_encoder *encoder = &hdmi_to_dig_port(intel_hdmi)->base; intel_wakeref_t wakeref; drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s]\n", - connector->base.id, connector->name); + connector->base.base.id, connector->base.name); if (!intel_display_device_enabled(display)) return connector_status_disconnected; if (!intel_display_driver_check_access(display)) - return connector->status; + return connector->base.status; wakeref = intel_display_power_get(display, POWER_DOMAIN_GMBUS); @@ -2554,9 +2557,9 @@ intel_hdmi_detect(struct drm_connector *connector, bool force) !intel_digital_port_connected(encoder)) goto out; - intel_hdmi_unset_edid(connector); + intel_hdmi_unset_edid(&connector->base); - if (intel_hdmi_set_edid(connector)) + if (intel_hdmi_set_edid(&connector->base)) status = connector_status_connected; out: @@ -2569,49 +2572,54 @@ out: } static void -intel_hdmi_force(struct drm_connector *connector) +intel_hdmi_force(struct drm_connector *_connector) { - struct intel_display *display = to_intel_display(connector->dev); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s]\n", - connector->base.id, connector->name); + connector->base.base.id, connector->base.name); if (!intel_display_driver_check_access(display)) return; - intel_hdmi_unset_edid(connector); + intel_hdmi_unset_edid(&connector->base); - if (connector->status != connector_status_connected) + if (connector->base.status != connector_status_connected) return; - intel_hdmi_set_edid(connector); + intel_hdmi_set_edid(&connector->base); } -static int intel_hdmi_get_modes(struct drm_connector *connector) +static int intel_hdmi_get_modes(struct drm_connector *_connector) { + struct intel_connector *connector = to_intel_connector(_connector); + /* drm_edid_connector_update() done in ->detect() or ->force() */ - return drm_edid_connector_add_modes(connector); + return drm_edid_connector_add_modes(&connector->base); } static int -intel_hdmi_connector_register(struct drm_connector *connector) +intel_hdmi_connector_register(struct drm_connector *_connector) { + struct intel_connector *connector = to_intel_connector(_connector); int ret; - ret = intel_connector_register(connector); + ret = intel_connector_register(&connector->base); if (ret) return ret; return ret; } -static void intel_hdmi_connector_unregister(struct drm_connector *connector) +static void intel_hdmi_connector_unregister(struct drm_connector *_connector) { - struct cec_notifier *n = intel_attached_hdmi(to_intel_connector(connector))->cec_notifier; + struct intel_connector *connector = to_intel_connector(_connector); + struct cec_notifier *n = intel_attached_hdmi(connector)->cec_notifier; cec_notifier_conn_unregister(n); - intel_connector_unregister(connector); + intel_connector_unregister(&connector->base); } static const struct drm_connector_funcs intel_hdmi_connector_funcs = { @@ -2627,15 +2635,16 @@ static const struct drm_connector_funcs intel_hdmi_connector_funcs = { .atomic_duplicate_state = intel_digital_connector_duplicate_state, }; -static int intel_hdmi_connector_atomic_check(struct drm_connector *connector, +static int intel_hdmi_connector_atomic_check(struct drm_connector *_connector, struct drm_atomic_state *state) { - struct intel_display *display = to_intel_display(connector->dev); + struct intel_connector *connector = to_intel_connector(_connector); + struct intel_display *display = to_intel_display(connector); if (HAS_DDI(display)) - return intel_digital_connector_atomic_check(connector, state); + return intel_digital_connector_atomic_check(&connector->base, state); else - return g4x_hdmi_connector_atomic_check(connector, state); + return g4x_hdmi_connector_atomic_check(&connector->base, state); } static const struct drm_connector_helper_funcs intel_hdmi_connector_helper_funcs = { @@ -2645,22 +2654,23 @@ static const struct drm_connector_helper_funcs intel_hdmi_connector_helper_funcs }; static void -intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *connector) +intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *_connector) { + struct intel_connector *connector = to_intel_connector(_connector); struct intel_display *display = to_intel_display(intel_hdmi); - intel_attach_force_audio_property(connector); - intel_attach_broadcast_rgb_property(connector); - intel_attach_aspect_ratio_property(connector); + intel_attach_force_audio_property(&connector->base); + intel_attach_broadcast_rgb_property(&connector->base); + intel_attach_aspect_ratio_property(&connector->base); - intel_attach_hdmi_colorspace_property(connector); - drm_connector_attach_content_type_property(connector); + intel_attach_hdmi_colorspace_property(&connector->base); + drm_connector_attach_content_type_property(&connector->base); if (DISPLAY_VER(display) >= 10) - drm_connector_attach_hdr_output_metadata_property(connector); + drm_connector_attach_hdr_output_metadata_property(&connector->base); if (!HAS_GMCH(display)) - drm_connector_attach_max_bpc_property(connector, 8, 12); + drm_connector_attach_max_bpc_property(&connector->base, 8, 12); } /* @@ -2682,25 +2692,26 @@ intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *c * True on success, false on failure. */ bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, - struct drm_connector *connector, + struct drm_connector *_connector, bool high_tmds_clock_ratio, bool scrambling) { + struct intel_connector *connector = to_intel_connector(_connector); struct intel_display *display = to_intel_display(encoder); struct drm_scrambling *sink_scrambling = - &connector->display_info.hdmi.scdc.scrambling; + &connector->base.display_info.hdmi.scdc.scrambling; if (!sink_scrambling->supported) return true; drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s] scrambling=%s, TMDS bit clock ratio=1/%d\n", - connector->base.id, connector->name, + connector->base.base.id, connector->base.name, str_yes_no(scrambling), high_tmds_clock_ratio ? 40 : 10); /* Set TMDS bit clock ratio to 1/40 or 1/10, and enable/disable scrambling */ - return drm_scdc_set_high_tmds_clock_ratio(connector, high_tmds_clock_ratio) && - drm_scdc_set_scrambling(connector, scrambling); + return drm_scdc_set_high_tmds_clock_ratio(&connector->base, high_tmds_clock_ratio) && + drm_scdc_set_scrambling(&connector->base, scrambling); } static u8 chv_encoder_to_ddc_pin(struct intel_encoder *encoder) @@ -2811,7 +2822,7 @@ static u8 mcc_encoder_to_ddc_pin(struct intel_encoder *encoder) static u8 rkl_encoder_to_ddc_pin(struct intel_encoder *encoder) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); enum phy phy = intel_encoder_to_phy(encoder); WARN_ON(encoder->port == PORT_C); @@ -2822,7 +2833,7 @@ static u8 rkl_encoder_to_ddc_pin(struct intel_encoder *encoder) * combo outputs. With CMP, the traditional DDI A-D pins are used for * all outputs. */ - if (INTEL_PCH_TYPE(dev_priv) >= PCH_TGP && phy >= PHY_C) + if (INTEL_PCH_TYPE(display) >= PCH_TGP && phy >= PHY_C) return GMBUS_PIN_9_TC1_ICP + phy - PHY_C; return GMBUS_PIN_1_BXT + phy; @@ -2831,7 +2842,6 @@ static u8 rkl_encoder_to_ddc_pin(struct intel_encoder *encoder) static u8 gen9bc_tgp_encoder_to_ddc_pin(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); enum phy phy = intel_encoder_to_phy(encoder); drm_WARN_ON(display->drm, encoder->port == PORT_A); @@ -2842,7 +2852,7 @@ static u8 gen9bc_tgp_encoder_to_ddc_pin(struct intel_encoder *encoder) * combo outputs. With CMP, the traditional DDI A-D pins are used for * all outputs. */ - if (INTEL_PCH_TYPE(i915) >= PCH_TGP && phy >= PHY_C) + if (INTEL_PCH_TYPE(display) >= PCH_TGP && phy >= PHY_C) return GMBUS_PIN_9_TC1_ICP + phy - PHY_C; return GMBUS_PIN_1_BXT + phy; @@ -2895,27 +2905,26 @@ static u8 g4x_encoder_to_ddc_pin(struct intel_encoder *encoder) static u8 intel_hdmi_default_ddc_pin(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); u8 ddc_pin; - if (IS_ALDERLAKE_S(dev_priv)) + if (display->platform.alderlake_s) ddc_pin = adls_encoder_to_ddc_pin(encoder); - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1) + else if (INTEL_PCH_TYPE(display) >= PCH_DG1) ddc_pin = dg1_encoder_to_ddc_pin(encoder); - else if (IS_ROCKETLAKE(dev_priv)) + else if (display->platform.rocketlake) ddc_pin = rkl_encoder_to_ddc_pin(encoder); - else if (DISPLAY_VER(display) == 9 && HAS_PCH_TGP(dev_priv)) + else if (DISPLAY_VER(display) == 9 && HAS_PCH_TGP(display)) ddc_pin = gen9bc_tgp_encoder_to_ddc_pin(encoder); - else if ((IS_JASPERLAKE(dev_priv) || IS_ELKHARTLAKE(dev_priv)) && - HAS_PCH_TGP(dev_priv)) + else if ((display->platform.jasperlake || display->platform.elkhartlake) && + HAS_PCH_TGP(display)) ddc_pin = mcc_encoder_to_ddc_pin(encoder); - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + else if (INTEL_PCH_TYPE(display) >= PCH_ICP) ddc_pin = icl_encoder_to_ddc_pin(encoder); - else if (HAS_PCH_CNP(dev_priv)) + else if (HAS_PCH_CNP(display)) ddc_pin = cnp_encoder_to_ddc_pin(encoder); - else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) + else if (display->platform.geminilake || display->platform.broxton) ddc_pin = bxt_encoder_to_ddc_pin(encoder); - else if (IS_CHERRYVIEW(dev_priv)) + else if (display->platform.cherryview) ddc_pin = chv_encoder_to_ddc_pin(encoder); else ddc_pin = g4x_encoder_to_ddc_pin(encoder); @@ -2989,15 +2998,13 @@ static u8 intel_hdmi_ddc_pin(struct intel_encoder *encoder) void intel_infoframe_init(struct intel_digital_port *dig_port) { struct intel_display *display = to_intel_display(dig_port); - struct drm_i915_private *dev_priv = - to_i915(dig_port->base.base.dev); - if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { + if (display->platform.valleyview || display->platform.cherryview) { dig_port->write_infoframe = vlv_write_infoframe; dig_port->read_infoframe = vlv_read_infoframe; dig_port->set_infoframes = vlv_set_infoframes; dig_port->infoframes_enabled = vlv_infoframes_enabled; - } else if (IS_G4X(dev_priv)) { + } else if (display->platform.g4x) { dig_port->write_infoframe = g4x_write_infoframe; dig_port->read_infoframe = g4x_read_infoframe; dig_port->set_infoframes = g4x_set_infoframes; @@ -3014,7 +3021,7 @@ void intel_infoframe_init(struct intel_digital_port *dig_port) dig_port->set_infoframes = hsw_set_infoframes; dig_port->infoframes_enabled = hsw_infoframes_enabled; } - } else if (HAS_PCH_IBX(dev_priv)) { + } else if (HAS_PCH_IBX(display)) { dig_port->write_infoframe = ibx_write_infoframe; dig_port->read_infoframe = ibx_read_infoframe; dig_port->set_infoframes = ibx_set_infoframes; diff --git a/drivers/gpu/drm/i915/display/intel_hotplug.c b/drivers/gpu/drm/i915/display/intel_hotplug.c index 6885e5a09079..fc5d8928c37e 100644 --- a/drivers/gpu/drm/i915/display/intel_hotplug.c +++ b/drivers/gpu/drm/i915/display/intel_hotplug.c @@ -183,9 +183,7 @@ static bool intel_hpd_irq_storm_detect(struct intel_display *display, static bool detection_work_enabled(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); return display->hotplug.detection_work_enabled; } @@ -195,7 +193,7 @@ mod_delayed_detection_work(struct intel_display *display, struct delayed_work *w { struct drm_i915_private *i915 = to_i915(display->drm); - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); if (!detection_work_enabled(display)) return false; @@ -208,7 +206,7 @@ queue_delayed_detection_work(struct intel_display *display, struct delayed_work { struct drm_i915_private *i915 = to_i915(display->drm); - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); if (!detection_work_enabled(display)) return false; @@ -221,7 +219,7 @@ queue_detection_work(struct intel_display *display, struct work_struct *work) { struct drm_i915_private *i915 = to_i915(display->drm); - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); if (!detection_work_enabled(display)) return false; @@ -232,12 +230,11 @@ queue_detection_work(struct intel_display *display, struct work_struct *work) static void intel_hpd_irq_storm_switch_to_polling(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct drm_connector_list_iter conn_iter; struct intel_connector *connector; bool hpd_disabled = false; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { @@ -276,7 +273,6 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work) { struct intel_display *display = container_of(work, typeof(*display), hotplug.reenable_work.work); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct drm_connector_list_iter conn_iter; struct intel_connector *connector; struct ref_tracker *wakeref; @@ -284,7 +280,7 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work) wakeref = intel_display_rpm_get(display); - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { @@ -308,7 +304,7 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work) intel_hpd_irq_setup(display); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); intel_display_rpm_put(display, wakeref); } @@ -376,9 +372,7 @@ static bool hpd_pin_has_pulse(struct intel_display *display, enum hpd_pin pin) static bool hpd_pin_is_blocked(struct intel_display *display, enum hpd_pin pin) { - struct drm_i915_private *i915 = to_i915(display->drm); - - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); return display->hotplug.stats[pin].blocked_count; } @@ -400,14 +394,13 @@ static void i915_digport_work_func(struct work_struct *work) { struct intel_display *display = container_of(work, struct intel_display, hotplug.dig_port_work); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; u32 long_hpd_pin_mask, short_hpd_pin_mask; struct intel_encoder *encoder; u32 blocked_hpd_pin_mask; u32 old_bits = 0; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); blocked_hpd_pin_mask = get_blocked_hpd_pin_mask(display); long_hpd_pin_mask = hotplug->long_hpd_pin_mask & ~blocked_hpd_pin_mask; @@ -415,7 +408,7 @@ static void i915_digport_work_func(struct work_struct *work) short_hpd_pin_mask = hotplug->short_hpd_pin_mask & ~blocked_hpd_pin_mask; hotplug->short_hpd_pin_mask &= ~short_hpd_pin_mask; - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); for_each_intel_encoder(display->drm, encoder) { struct intel_digital_port *dig_port; @@ -442,11 +435,11 @@ static void i915_digport_work_func(struct work_struct *work) } if (old_bits) { - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); display->hotplug.event_bits |= old_bits; queue_delayed_detection_work(display, &display->hotplug.hotplug_work, 0); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } } @@ -460,17 +453,16 @@ static void i915_digport_work_func(struct work_struct *work) void intel_hpd_trigger_irq(struct intel_digital_port *dig_port) { struct intel_display *display = to_intel_display(dig_port); - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; struct intel_encoder *encoder = &dig_port->base; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); hotplug->short_hpd_pin_mask |= BIT(encoder->hpd_pin); if (!hpd_pin_is_blocked(display, encoder->hpd_pin)) queue_work(hotplug->dp_wq, &hotplug->dig_port_work); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } /* @@ -480,7 +472,6 @@ static void i915_hotplug_work_func(struct work_struct *work) { struct intel_display *display = container_of(work, struct intel_display, hotplug.hotplug_work.work); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; struct drm_connector_list_iter conn_iter; struct intel_connector *connector; @@ -494,7 +485,7 @@ static void i915_hotplug_work_func(struct work_struct *work) mutex_lock(&display->drm->mode_config.mutex); drm_dbg_kms(display->drm, "running encoder hotplug functions\n"); - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); blocked_hpd_pin_mask = get_blocked_hpd_pin_mask(display); hpd_event_bits = hotplug->event_bits & ~blocked_hpd_pin_mask; @@ -505,7 +496,7 @@ static void i915_hotplug_work_func(struct work_struct *work) /* Enable polling for connectors which had HPD IRQ storms */ intel_hpd_irq_storm_switch_to_polling(display); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); /* Skip calling encode hotplug handlers if ignore long HPD set*/ if (display->hotplug.ignore_long_hpd) { @@ -569,13 +560,13 @@ static void i915_hotplug_work_func(struct work_struct *work) /* Remove shared HPD pins that have changed */ retry &= ~changed; if (retry) { - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); display->hotplug.retry_bits |= retry; mod_delayed_detection_work(display, &display->hotplug.hotplug_work, msecs_to_jiffies(HPD_RETRY_DELAY)); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } } @@ -599,7 +590,6 @@ static void i915_hotplug_work_func(struct work_struct *work) void intel_hpd_irq_handler(struct intel_display *display, u32 pin_mask, u32 long_mask) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder; bool storm_detected = false; bool queue_dig = false, queue_hp = false; @@ -610,7 +600,7 @@ void intel_hpd_irq_handler(struct intel_display *display, if (!pin_mask) return; - spin_lock(&dev_priv->irq_lock); + spin_lock(&display->irq.lock); /* * Determine whether ->hpd_pulse() exists for each pin, and @@ -711,7 +701,7 @@ void intel_hpd_irq_handler(struct intel_display *display, queue_delayed_detection_work(display, &display->hotplug.hotplug_work, 0); - spin_unlock(&dev_priv->irq_lock); + spin_unlock(&display->irq.lock); } /** @@ -730,7 +720,6 @@ void intel_hpd_irq_handler(struct intel_display *display, */ void intel_hpd_init(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); int i; if (!HAS_DISPLAY(display)) @@ -745,9 +734,9 @@ void intel_hpd_init(struct intel_display *display) * Interrupt setup is already guaranteed to be single-threaded, this is * just to make the assert_spin_locked checks happy. */ - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); intel_hpd_irq_setup(display); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void i915_hpd_poll_detect_connectors(struct intel_display *display) @@ -797,7 +786,6 @@ static void i915_hpd_poll_init_work(struct work_struct *work) { struct intel_display *display = container_of(work, typeof(*display), hotplug.poll_init_work); - struct drm_i915_private *dev_priv = to_i915(display->drm); struct drm_connector_list_iter conn_iter; struct intel_connector *connector; intel_wakeref_t wakeref; @@ -820,7 +808,7 @@ static void i915_hpd_poll_init_work(struct work_struct *work) cancel_work(&display->hotplug.poll_init_work); } - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { @@ -841,7 +829,7 @@ static void i915_hpd_poll_init_work(struct work_struct *work) } drm_connector_list_iter_end(&conn_iter); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); if (enabled) drm_kms_helper_poll_reschedule(display->drm); @@ -879,8 +867,6 @@ static void i915_hpd_poll_init_work(struct work_struct *work) */ void intel_hpd_poll_enable(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (!HAS_DISPLAY(display) || !intel_display_device_enabled(display)) return; @@ -892,10 +878,10 @@ void intel_hpd_poll_enable(struct intel_display *display) * As well, there's no issue if we race here since we always reschedule * this worker anyway */ - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); queue_detection_work(display, &display->hotplug.poll_init_work); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } /** @@ -919,17 +905,15 @@ void intel_hpd_poll_enable(struct intel_display *display) */ void intel_hpd_poll_disable(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (!HAS_DISPLAY(display)) return; WRITE_ONCE(display->hotplug.poll_enabled, false); - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); queue_detection_work(display, &display->hotplug.poll_init_work); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } void intel_hpd_poll_fini(struct intel_display *display) @@ -981,12 +965,10 @@ static bool cancel_all_detection_work(struct intel_display *display) void intel_hpd_cancel_work(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - if (!HAS_DISPLAY(display)) return; - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); drm_WARN_ON(display->drm, get_blocked_hpd_pin_mask(display)); @@ -995,7 +977,7 @@ void intel_hpd_cancel_work(struct intel_display *display) display->hotplug.event_bits = 0; display->hotplug.retry_bits = 0; - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); cancel_work_sync(&display->hotplug.dig_port_work); @@ -1009,13 +991,12 @@ void intel_hpd_cancel_work(struct intel_display *display) static void queue_work_for_missed_irqs(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; bool queue_hp_work = false; u32 blocked_hpd_pin_mask; enum hpd_pin pin; - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); blocked_hpd_pin_mask = get_blocked_hpd_pin_mask(display); if ((hotplug->event_bits | hotplug->retry_bits) & ~blocked_hpd_pin_mask) @@ -1043,10 +1024,9 @@ static void queue_work_for_missed_irqs(struct intel_display *display) static bool block_hpd_pin(struct intel_display *display, enum hpd_pin pin) { - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); hotplug->stats[pin].blocked_count++; @@ -1055,10 +1035,9 @@ static bool block_hpd_pin(struct intel_display *display, enum hpd_pin pin) static bool unblock_hpd_pin(struct intel_display *display, enum hpd_pin pin) { - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; - lockdep_assert_held(&i915->irq_lock); + lockdep_assert_held(&display->irq.lock); if (drm_WARN_ON(display->drm, hotplug->stats[pin].blocked_count == 0)) return true; @@ -1095,19 +1074,18 @@ static bool unblock_hpd_pin(struct intel_display *display, enum hpd_pin pin) void intel_hpd_block(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; bool do_flush = false; if (encoder->hpd_pin == HPD_NONE) return; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); if (block_hpd_pin(display, encoder->hpd_pin)) do_flush = true; - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); if (do_flush && hpd_pin_has_pulse(display, encoder->hpd_pin)) flush_work(&hotplug->dig_port_work); @@ -1125,17 +1103,16 @@ void intel_hpd_block(struct intel_encoder *encoder) void intel_hpd_unblock(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(display->drm); if (encoder->hpd_pin == HPD_NONE) return; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); if (unblock_hpd_pin(display, encoder->hpd_pin)) queue_work_for_missed_irqs(display); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } /** @@ -1149,14 +1126,13 @@ void intel_hpd_unblock(struct intel_encoder *encoder) void intel_hpd_clear_and_unblock(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; enum hpd_pin pin = encoder->hpd_pin; if (pin == HPD_NONE) return; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); if (unblock_hpd_pin(display, pin)) { hotplug->event_bits &= ~BIT(pin); @@ -1165,39 +1141,34 @@ void intel_hpd_clear_and_unblock(struct intel_encoder *encoder) hotplug->long_hpd_pin_mask &= ~BIT(pin); } - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } void intel_hpd_enable_detection_work(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); display->hotplug.detection_work_enabled = true; queue_work_for_missed_irqs(display); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } void intel_hpd_disable_detection_work(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); display->hotplug.detection_work_enabled = false; - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); cancel_all_detection_work(display); } bool intel_hpd_schedule_detection(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); unsigned long flags; bool ret; - spin_lock_irqsave(&i915->irq_lock, flags); + spin_lock_irqsave(&display->irq.lock, flags); ret = queue_delayed_detection_work(display, &display->hotplug.hotplug_work, 0); - spin_unlock_irqrestore(&i915->irq_lock, flags); + spin_unlock_irqrestore(&display->irq.lock, flags); return ret; } @@ -1228,7 +1199,6 @@ static ssize_t i915_hpd_storm_ctl_write(struct file *file, { struct seq_file *m = file->private_data; struct intel_display *display = m->private; - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; unsigned int new_threshold; int i; @@ -1260,12 +1230,12 @@ static ssize_t i915_hpd_storm_ctl_write(struct file *file, else drm_dbg_kms(display->drm, "Disabling HPD storm detection\n"); - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); hotplug->hpd_storm_threshold = new_threshold; /* Reset the HPD storm stats so we don't accidentally trigger a storm */ for_each_hpd_pin(i) hotplug->stats[i].count = 0; - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); /* Re-enable hpd immediately if we were in an irq storm */ flush_delayed_work(&display->hotplug.reenable_work); @@ -1310,7 +1280,6 @@ static ssize_t i915_hpd_short_storm_ctl_write(struct file *file, { struct seq_file *m = file->private_data; struct intel_display *display = m->private; - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hotplug *hotplug = &display->hotplug; char *newline; char tmp[16]; @@ -1339,12 +1308,12 @@ static ssize_t i915_hpd_short_storm_ctl_write(struct file *file, drm_dbg_kms(display->drm, "%sabling HPD short storm detection\n", new_state ? "En" : "Dis"); - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); hotplug->hpd_short_storm_enabled = new_state; /* Reset the HPD storm stats so we don't accidentally trigger a storm */ for_each_hpd_pin(i) hotplug->stats[i].count = 0; - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); /* Re-enable hpd immediately if we were in an irq storm */ flush_delayed_work(&display->hotplug.reenable_work); diff --git a/drivers/gpu/drm/i915/display/intel_hotplug_irq.c b/drivers/gpu/drm/i915/display/intel_hotplug_irq.c index 2463e61e7802..c024b42369c8 100644 --- a/drivers/gpu/drm/i915/display/intel_hotplug_irq.c +++ b/drivers/gpu/drm/i915/display/intel_hotplug_irq.c @@ -3,8 +3,10 @@ * Copyright © 2023 Intel Corporation */ -#include "i915_drv.h" +#include <drm/drm_print.h> + #include "i915_reg.h" +#include "i915_utils.h" #include "intel_de.h" #include "intel_display_irq.h" #include "intel_display_types.h" @@ -133,7 +135,6 @@ static const u32 hpd_mtp[HPD_NUM_PINS] = { static void intel_hpd_init_pins(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_hotplug *hpd = &display->hotplug; if (HAS_GMCH(display)) { @@ -160,33 +161,31 @@ static void intel_hpd_init_pins(struct intel_display *display) else hpd->hpd = hpd_ilk; - if ((INTEL_PCH_TYPE(dev_priv) < PCH_DG1) && - (!HAS_PCH_SPLIT(dev_priv) || HAS_PCH_NOP(dev_priv))) + if ((INTEL_PCH_TYPE(display) < PCH_DG1) && + (!HAS_PCH_SPLIT(display) || HAS_PCH_NOP(display))) return; - if (INTEL_PCH_TYPE(dev_priv) >= PCH_MTL) + if (INTEL_PCH_TYPE(display) >= PCH_MTL) hpd->pch_hpd = hpd_mtp; - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1) + else if (INTEL_PCH_TYPE(display) >= PCH_DG1) hpd->pch_hpd = hpd_sde_dg1; - else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + else if (INTEL_PCH_TYPE(display) >= PCH_ICP) hpd->pch_hpd = hpd_icp; - else if (HAS_PCH_CNP(dev_priv) || HAS_PCH_SPT(dev_priv)) + else if (HAS_PCH_CNP(display) || HAS_PCH_SPT(display)) hpd->pch_hpd = hpd_spt; - else if (HAS_PCH_LPT(dev_priv) || HAS_PCH_CPT(dev_priv)) + else if (HAS_PCH_LPT(display) || HAS_PCH_CPT(display)) hpd->pch_hpd = hpd_cpt; - else if (HAS_PCH_IBX(dev_priv)) + else if (HAS_PCH_IBX(display)) hpd->pch_hpd = hpd_ibx; else - MISSING_CASE(INTEL_PCH_TYPE(dev_priv)); + MISSING_CASE(INTEL_PCH_TYPE(display)); } /* For display hotplug interrupt */ void i915_hotplug_interrupt_update_locked(struct intel_display *display, u32 mask, u32 bits) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); drm_WARN_ON(display->drm, bits & ~mask); intel_de_rmw(display, PORT_HOTPLUG_EN(display), mask, bits); @@ -208,11 +207,9 @@ void i915_hotplug_interrupt_update(struct intel_display *display, u32 mask, u32 bits) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); i915_hotplug_interrupt_update_locked(display, mask, bits); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } static bool gen11_port_hotplug_long_detect(enum hpd_pin pin, u32 val) @@ -557,7 +554,6 @@ void xelpdp_pica_irq_handler(struct intel_display *display, u32 iir) void icp_irq_handler(struct intel_display *display, u32 pch_iir) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 ddi_hotplug_trigger = pch_iir & SDE_DDI_HOTPLUG_MASK_ICP; u32 tc_hotplug_trigger = pch_iir & SDE_TC_HOTPLUG_MASK_ICP; u32 pin_mask = 0, long_mask = 0; @@ -566,9 +562,9 @@ void icp_irq_handler(struct intel_display *display, u32 pch_iir) u32 dig_hotplug_reg; /* Locking due to DSI native GPIO sequences */ - spin_lock(&dev_priv->irq_lock); + spin_lock(&display->irq.lock); dig_hotplug_reg = intel_de_rmw(display, SHOTPLUG_CTL_DDI, 0, 0); - spin_unlock(&dev_priv->irq_lock); + spin_unlock(&display->irq.lock); intel_get_hpd_pins(display, &pin_mask, &long_mask, ddi_hotplug_trigger, dig_hotplug_reg, @@ -711,7 +707,7 @@ static u32 ibx_hotplug_mask(enum hpd_pin hpd_pin) static u32 ibx_hotplug_enables(struct intel_encoder *encoder) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); switch (encoder->hpd_pin) { case HPD_PORT_A: @@ -719,7 +715,7 @@ static u32 ibx_hotplug_enables(struct intel_encoder *encoder) * When CPU and PCH are on the same package, port A * HPD must be enabled in both north and south. */ - return HAS_PCH_LPT_LP(i915) ? + return HAS_PCH_LPT_LP(display) ? PORTA_HOTPLUG_ENABLE : 0; case HPD_PORT_B: return PORTB_HOTPLUG_ENABLE | @@ -940,18 +936,17 @@ static void gen11_tbt_hpd_enable_detection(struct intel_encoder *encoder) static void gen11_hpd_enable_detection(struct intel_encoder *encoder) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); gen11_tc_hpd_enable_detection(encoder); gen11_tbt_hpd_enable_detection(encoder); - if (INTEL_PCH_TYPE(i915) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) icp_hpd_enable_detection(encoder); } static void gen11_hpd_irq_setup(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 hotplug_irqs, enabled_irqs; enabled_irqs = intel_hpd_enabled_irqs(display, display->hotplug.hpd); @@ -964,7 +959,7 @@ static void gen11_hpd_irq_setup(struct intel_display *display) gen11_tc_hpd_detection_setup(display); gen11_tbt_hpd_detection_setup(display); - if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) icp_hpd_irq_setup(display); } @@ -1138,7 +1133,6 @@ static void xelpdp_hpd_enable_detection(struct intel_encoder *encoder) static void xelpdp_hpd_irq_setup(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); u32 hotplug_irqs, enabled_irqs; enabled_irqs = intel_hpd_enabled_irqs(display, display->hotplug.hpd); @@ -1150,9 +1144,9 @@ static void xelpdp_hpd_irq_setup(struct intel_display *display) xelpdp_pica_hpd_detection_setup(display); - if (INTEL_PCH_TYPE(i915) >= PCH_LNL) + if (INTEL_PCH_TYPE(display) >= PCH_LNL) xe2lpd_sde_hpd_irq_setup(display); - else if (INTEL_PCH_TYPE(i915) >= PCH_MTL) + else if (INTEL_PCH_TYPE(display) >= PCH_MTL) mtp_hpd_irq_setup(display); } @@ -1194,10 +1188,8 @@ static u32 spt_hotplug2_enables(struct intel_encoder *encoder) static void spt_hpd_detection_setup(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - /* Display WA #1179 WaHardHangonHotPlug: cnp */ - if (HAS_PCH_CNP(dev_priv)) { + if (HAS_PCH_CNP(display)) { intel_de_rmw(display, SOUTH_CHICKEN1, CHASSIS_CLK_REQ_DURATION_MASK, CHASSIS_CLK_REQ_DURATION(0xf)); } @@ -1215,10 +1207,9 @@ static void spt_hpd_detection_setup(struct intel_display *display) static void spt_hpd_enable_detection(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); /* Display WA #1179 WaHardHangonHotPlug: cnp */ - if (HAS_PCH_CNP(i915)) { + if (HAS_PCH_CNP(display)) { intel_de_rmw(display, SOUTH_CHICKEN1, CHASSIS_CLK_REQ_DURATION_MASK, CHASSIS_CLK_REQ_DURATION(0xf)); @@ -1235,10 +1226,9 @@ static void spt_hpd_enable_detection(struct intel_encoder *encoder) static void spt_hpd_irq_setup(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 hotplug_irqs, enabled_irqs; - if (INTEL_PCH_TYPE(dev_priv) >= PCH_CNP) + if (INTEL_PCH_TYPE(display) >= PCH_CNP) intel_de_write(display, SHPD_FILTER_CNT, SHPD_FILTER_CNT_500_ADJ); enabled_irqs = intel_hpd_enabled_irqs(display, display->hotplug.pch_hpd); @@ -1402,10 +1392,9 @@ static void i915_hpd_enable_detection(struct intel_encoder *encoder) static void i915_hpd_irq_setup(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 hotplug_en; - lockdep_assert_held(&dev_priv->irq_lock); + lockdep_assert_held(&display->irq.lock); /* * Note HDMI and DP share hotplug bits. Enable bits are the same for all @@ -1474,8 +1463,6 @@ void intel_hpd_irq_setup(struct intel_display *display) void intel_hotplug_irq_init(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - intel_hpd_init_pins(display); intel_hpd_init_early(display); @@ -1484,9 +1471,9 @@ void intel_hotplug_irq_init(struct intel_display *display) if (HAS_HOTPLUG(display)) display->funcs.hotplug = &i915_hpd_funcs; } else { - if (HAS_PCH_DG2(i915)) + if (HAS_PCH_DG2(display)) display->funcs.hotplug = &icp_hpd_funcs; - else if (HAS_PCH_DG1(i915)) + else if (HAS_PCH_DG1(display)) display->funcs.hotplug = &dg1_hpd_funcs; else if (DISPLAY_VER(display) >= 14) display->funcs.hotplug = &xelpdp_hpd_funcs; @@ -1494,9 +1481,9 @@ void intel_hotplug_irq_init(struct intel_display *display) display->funcs.hotplug = &gen11_hpd_funcs; else if (display->platform.geminilake || display->platform.broxton) display->funcs.hotplug = &bxt_hpd_funcs; - else if (INTEL_PCH_TYPE(i915) >= PCH_ICP) + else if (INTEL_PCH_TYPE(display) >= PCH_ICP) display->funcs.hotplug = &icp_hpd_funcs; - else if (INTEL_PCH_TYPE(i915) >= PCH_SPT) + else if (INTEL_PCH_TYPE(display) >= PCH_SPT) display->funcs.hotplug = &spt_hpd_funcs; else display->funcs.hotplug = &ilk_hpd_funcs; diff --git a/drivers/gpu/drm/i915/display/intel_lpe_audio.c b/drivers/gpu/drm/i915/display/intel_lpe_audio.c index 59551c8414c2..666148a14522 100644 --- a/drivers/gpu/drm/i915/display/intel_lpe_audio.c +++ b/drivers/gpu/drm/i915/display/intel_lpe_audio.c @@ -179,7 +179,7 @@ static int lpe_audio_irq_init(struct intel_display *display) handle_simple_irq, "hdmi_lpe_audio_irq_handler"); - return irq_set_chip_data(irq, dev_priv); + return 0; } static bool lpe_audio_detect(struct intel_display *display) diff --git a/drivers/gpu/drm/i915/display/intel_lvds.c b/drivers/gpu/drm/i915/display/intel_lvds.c index 89d26913e253..8ce7c630da52 100644 --- a/drivers/gpu/drm/i915/display/intel_lvds.c +++ b/drivers/gpu/drm/i915/display/intel_lvds.c @@ -37,9 +37,9 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_crtc.h> #include <drm/drm_edid.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> -#include "i915_drv.h" #include "i915_reg.h" #include "intel_atomic.h" #include "intel_backlight.h" @@ -87,13 +87,12 @@ static struct intel_lvds_encoder *to_lvds_encoder(struct intel_encoder *encoder) bool intel_lvds_port_enabled(struct intel_display *display, i915_reg_t lvds_reg, enum pipe *pipe) { - struct drm_i915_private *i915 = to_i915(display->drm); u32 val; val = intel_de_read(display, lvds_reg); /* asserts want to know the pipe even if the port is disabled */ - if (HAS_PCH_CPT(i915)) + if (HAS_PCH_CPT(display)) *pipe = REG_FIELD_GET(LVDS_PIPE_SEL_MASK_CPT, val); else *pipe = REG_FIELD_GET(LVDS_PIPE_SEL_MASK, val); @@ -243,13 +242,12 @@ static void intel_pre_enable_lvds(struct intel_atomic_state *state, { struct intel_display *display = to_intel_display(state); struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; enum pipe pipe = crtc->pipe; u32 temp; - if (HAS_PCH_SPLIT(i915)) { + if (HAS_PCH_SPLIT(display)) { assert_fdi_rx_pll_disabled(display, pipe); assert_shared_dpll_disabled(display, crtc_state->shared_dpll); } else { @@ -261,7 +259,7 @@ static void intel_pre_enable_lvds(struct intel_atomic_state *state, temp = lvds_encoder->init_lvds_val; temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP; - if (HAS_PCH_CPT(i915)) { + if (HAS_PCH_CPT(display)) { temp &= ~LVDS_PIPE_SEL_MASK_CPT; temp |= LVDS_PIPE_SEL_CPT(pipe); } else { @@ -421,7 +419,6 @@ static int intel_lvds_compute_config(struct intel_encoder *encoder, struct drm_connector_state *conn_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(encoder); struct intel_connector *connector = lvds_encoder->attached_connector; struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; @@ -435,7 +432,7 @@ static int intel_lvds_compute_config(struct intel_encoder *encoder, return -EINVAL; } - if (HAS_PCH_SPLIT(i915)) { + if (HAS_PCH_SPLIT(display)) { crtc_state->has_pch_encoder = true; if (!intel_fdi_compute_pipe_bpp(crtc_state)) return -EINVAL; @@ -798,7 +795,6 @@ bool intel_is_dual_link_lvds(struct intel_display *display) static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder) { struct intel_display *display = to_intel_display(&lvds_encoder->base); - struct drm_i915_private *i915 = to_i915(lvds_encoder->base.base.dev); struct intel_connector *connector = lvds_encoder->attached_connector; const struct drm_display_mode *fixed_mode = intel_panel_preferred_fixed_mode(connector); @@ -822,7 +818,7 @@ static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder) * register is uninitialized. */ val = intel_de_read(display, lvds_encoder->reg); - if (HAS_PCH_CPT(i915)) + if (HAS_PCH_CPT(display)) val &= ~(LVDS_DETECTED | LVDS_PIPE_SEL_MASK_CPT); else val &= ~(LVDS_DETECTED | LVDS_PIPE_SEL_MASK); @@ -846,7 +842,6 @@ static void intel_lvds_add_properties(struct drm_connector *connector) */ void intel_lvds_init(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); struct intel_lvds_encoder *lvds_encoder; struct intel_connector *connector; const struct drm_edid *drm_edid; @@ -868,14 +863,14 @@ void intel_lvds_init(struct intel_display *display) return; } - if (HAS_PCH_SPLIT(i915)) + if (HAS_PCH_SPLIT(display)) lvds_reg = PCH_LVDS; else lvds_reg = LVDS; lvds = intel_de_read(display, lvds_reg); - if (HAS_PCH_SPLIT(i915)) { + if (HAS_PCH_SPLIT(display)) { if ((lvds & LVDS_DETECTED) == 0) return; } @@ -915,7 +910,7 @@ void intel_lvds_init(struct intel_display *display) encoder->enable = intel_enable_lvds; encoder->pre_enable = intel_pre_enable_lvds; encoder->compute_config = intel_lvds_compute_config; - if (HAS_PCH_SPLIT(i915)) { + if (HAS_PCH_SPLIT(display)) { encoder->disable = pch_disable_lvds; encoder->post_disable = pch_post_disable_lvds; } else { diff --git a/drivers/gpu/drm/i915/display/intel_modeset_setup.c b/drivers/gpu/drm/i915/display/intel_modeset_setup.c index 9e963bce340f..0325b0c9506d 100644 --- a/drivers/gpu/drm/i915/display/intel_modeset_setup.c +++ b/drivers/gpu/drm/i915/display/intel_modeset_setup.c @@ -6,11 +6,11 @@ * state. */ -#include <drm/drm_atomic_uapi.h> #include <drm/drm_atomic_state_helper.h> +#include <drm/drm_atomic_uapi.h> +#include <drm/drm_print.h> #include <drm/drm_vblank.h> -#include "i915_drv.h" #include "i915_reg.h" #include "i9xx_wm.h" #include "intel_atomic.h" @@ -31,13 +31,14 @@ #include "intel_pmdemand.h" #include "intel_tc.h" #include "intel_vblank.h" +#include "intel_vga.h" #include "intel_wm.h" #include "skl_watermark.h" static void intel_crtc_disable_noatomic_begin(struct intel_crtc *crtc, struct drm_modeset_acquire_ctx *ctx) { - struct drm_i915_private *i915 = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(crtc); struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_plane *plane; @@ -48,7 +49,7 @@ static void intel_crtc_disable_noatomic_begin(struct intel_crtc *crtc, if (!crtc_state->hw.active) return; - for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) { + for_each_intel_plane_on_crtc(display->drm, crtc, plane) { const struct intel_plane_state *plane_state = to_intel_plane_state(plane->base.state); @@ -56,9 +57,9 @@ static void intel_crtc_disable_noatomic_begin(struct intel_crtc *crtc, intel_plane_disable_noatomic(crtc, plane); } - state = drm_atomic_state_alloc(&i915->drm); + state = drm_atomic_state_alloc(display->drm); if (!state) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "failed to disable [CRTC:%d:%s], out of memory", crtc->base.base.id, crtc->base.name); return; @@ -68,7 +69,7 @@ static void intel_crtc_disable_noatomic_begin(struct intel_crtc *crtc, to_intel_atomic_state(state)->internal = true; /* Everything's already locked, -EDEADLK can't happen. */ - for_each_intel_crtc_in_pipe_mask(&i915->drm, temp_crtc, + for_each_intel_crtc_in_pipe_mask(display->drm, temp_crtc, BIT(pipe) | intel_crtc_joiner_secondary_pipes(crtc_state)) { struct intel_crtc_state *temp_crtc_state = @@ -77,14 +78,14 @@ static void intel_crtc_disable_noatomic_begin(struct intel_crtc *crtc, ret = drm_atomic_add_affected_connectors(state, &temp_crtc->base); - drm_WARN_ON(&i915->drm, IS_ERR(temp_crtc_state) || ret); + drm_WARN_ON(display->drm, IS_ERR(temp_crtc_state) || ret); } - i915->display.funcs.display->crtc_disable(to_intel_atomic_state(state), crtc); + display->funcs.display->crtc_disable(to_intel_atomic_state(state), crtc); drm_atomic_state_put(state); - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[CRTC:%d:%s] hw state adjusted, was enabled, now disabled\n", crtc->base.base.id, crtc->base.name); @@ -118,13 +119,12 @@ static void set_encoder_for_connector(struct intel_connector *connector, static void reset_encoder_connector_state(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_pmdemand_state *pmdemand_state = - to_intel_pmdemand_state(i915->display.pmdemand.obj.state); + to_intel_pmdemand_state(display->pmdemand.obj.state); struct intel_connector *connector; struct drm_connector_list_iter conn_iter; - drm_connector_list_iter_begin(&i915->drm, &conn_iter); + drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { if (connector->base.encoder != &encoder->base) continue; @@ -143,10 +143,10 @@ static void reset_encoder_connector_state(struct intel_encoder *encoder) static void reset_crtc_encoder_state(struct intel_crtc *crtc) { - struct drm_i915_private *i915 = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(crtc); struct intel_encoder *encoder; - for_each_encoder_on_crtc(&i915->drm, &crtc->base, encoder) { + for_each_encoder_on_crtc(display->drm, &crtc->base, encoder) { reset_encoder_connector_state(encoder); encoder->base.crtc = NULL; } @@ -183,13 +183,13 @@ static void intel_crtc_disable_noatomic_complete(struct intel_crtc *crtc) * Return all the pipes using a transcoder in @transcoder_mask. * For joiner configs return only the joiner primary. */ -static u8 get_transcoder_pipes(struct drm_i915_private *i915, +static u8 get_transcoder_pipes(struct intel_display *display, u8 transcoder_mask) { struct intel_crtc *temp_crtc; u8 pipes = 0; - for_each_intel_crtc(&i915->drm, temp_crtc) { + for_each_intel_crtc(display->drm, temp_crtc) { struct intel_crtc_state *temp_crtc_state = to_intel_crtc_state(temp_crtc->base.state); @@ -214,7 +214,6 @@ static void get_portsync_pipes(struct intel_crtc *crtc, u8 *master_pipe_mask, u8 *slave_pipes_mask) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *i915 = to_i915(crtc->base.dev); struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_crtc *master_crtc; @@ -233,20 +232,20 @@ static void get_portsync_pipes(struct intel_crtc *crtc, else master_transcoder = crtc_state->master_transcoder; - *master_pipe_mask = get_transcoder_pipes(i915, BIT(master_transcoder)); - drm_WARN_ON(&i915->drm, !is_power_of_2(*master_pipe_mask)); + *master_pipe_mask = get_transcoder_pipes(display, BIT(master_transcoder)); + drm_WARN_ON(display->drm, !is_power_of_2(*master_pipe_mask)); master_crtc = intel_crtc_for_pipe(display, ffs(*master_pipe_mask) - 1); master_crtc_state = to_intel_crtc_state(master_crtc->base.state); - *slave_pipes_mask = get_transcoder_pipes(i915, master_crtc_state->sync_mode_slaves_mask); + *slave_pipes_mask = get_transcoder_pipes(display, master_crtc_state->sync_mode_slaves_mask); } -static u8 get_joiner_secondary_pipes(struct drm_i915_private *i915, u8 primary_pipes_mask) +static u8 get_joiner_secondary_pipes(struct intel_display *display, u8 primary_pipes_mask) { struct intel_crtc *primary_crtc; u8 pipes = 0; - for_each_intel_crtc_in_pipe_mask(&i915->drm, primary_crtc, primary_pipes_mask) { + for_each_intel_crtc_in_pipe_mask(display->drm, primary_crtc, primary_pipes_mask) { struct intel_crtc_state *primary_crtc_state = to_intel_crtc_state(primary_crtc->base.state); @@ -259,45 +258,45 @@ static u8 get_joiner_secondary_pipes(struct drm_i915_private *i915, u8 primary_p static void intel_crtc_disable_noatomic(struct intel_crtc *crtc, struct drm_modeset_acquire_ctx *ctx) { - struct drm_i915_private *i915 = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(crtc); + struct intel_crtc *temp_crtc; u8 portsync_master_mask; u8 portsync_slaves_mask; u8 joiner_secondaries_mask; - struct intel_crtc *temp_crtc; /* TODO: Add support for MST */ get_portsync_pipes(crtc, &portsync_master_mask, &portsync_slaves_mask); - joiner_secondaries_mask = get_joiner_secondary_pipes(i915, + joiner_secondaries_mask = get_joiner_secondary_pipes(display, portsync_master_mask | portsync_slaves_mask); - drm_WARN_ON(&i915->drm, + drm_WARN_ON(display->drm, portsync_master_mask & portsync_slaves_mask || portsync_master_mask & joiner_secondaries_mask || portsync_slaves_mask & joiner_secondaries_mask); - for_each_intel_crtc_in_pipe_mask(&i915->drm, temp_crtc, joiner_secondaries_mask) + for_each_intel_crtc_in_pipe_mask(display->drm, temp_crtc, joiner_secondaries_mask) intel_crtc_disable_noatomic_begin(temp_crtc, ctx); - for_each_intel_crtc_in_pipe_mask(&i915->drm, temp_crtc, portsync_slaves_mask) + for_each_intel_crtc_in_pipe_mask(display->drm, temp_crtc, portsync_slaves_mask) intel_crtc_disable_noatomic_begin(temp_crtc, ctx); - for_each_intel_crtc_in_pipe_mask(&i915->drm, temp_crtc, portsync_master_mask) + for_each_intel_crtc_in_pipe_mask(display->drm, temp_crtc, portsync_master_mask) intel_crtc_disable_noatomic_begin(temp_crtc, ctx); - for_each_intel_crtc_in_pipe_mask(&i915->drm, temp_crtc, + for_each_intel_crtc_in_pipe_mask(display->drm, temp_crtc, joiner_secondaries_mask | portsync_slaves_mask | portsync_master_mask) intel_crtc_disable_noatomic_complete(temp_crtc); } -static void intel_modeset_update_connector_atomic_state(struct drm_i915_private *i915) +static void intel_modeset_update_connector_atomic_state(struct intel_display *display) { struct intel_connector *connector; struct drm_connector_list_iter conn_iter; - drm_connector_list_iter_begin(&i915->drm, &conn_iter); + drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { struct drm_connector_state *conn_state = connector->base.state; struct intel_encoder *encoder = @@ -319,7 +318,7 @@ static void intel_modeset_update_connector_atomic_state(struct drm_i915_private static void intel_crtc_copy_hw_to_uapi_state(struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev); + struct intel_display *display = to_intel_display(crtc_state); if (intel_crtc_is_joiner_secondary(crtc_state)) return; @@ -332,7 +331,7 @@ static void intel_crtc_copy_hw_to_uapi_state(struct intel_crtc_state *crtc_state crtc_state->uapi.adjusted_mode = crtc_state->hw.adjusted_mode; crtc_state->uapi.scaling_filter = crtc_state->hw.scaling_filter; - if (DISPLAY_INFO(i915)->color.degamma_lut_size) { + if (DISPLAY_INFO(display)->color.degamma_lut_size) { /* assume 1:1 mapping */ drm_property_replace_blob(&crtc_state->hw.degamma_lut, crtc_state->pre_csc_lut); @@ -347,7 +346,7 @@ static void intel_crtc_copy_hw_to_uapi_state(struct intel_crtc_state *crtc_state * to gamma_lut as that is the only valid source of LUTs * in the uapi. */ - drm_WARN_ON(&i915->drm, crtc_state->post_csc_lut && + drm_WARN_ON(display->drm, crtc_state->post_csc_lut && crtc_state->pre_csc_lut); drm_property_replace_blob(&crtc_state->hw.degamma_lut, @@ -366,15 +365,14 @@ static void intel_crtc_copy_hw_to_uapi_state(struct intel_crtc_state *crtc_state } static void -intel_sanitize_plane_mapping(struct drm_i915_private *i915) +intel_sanitize_plane_mapping(struct intel_display *display) { - struct intel_display *display = &i915->display; struct intel_crtc *crtc; - if (DISPLAY_VER(i915) >= 4) + if (DISPLAY_VER(display) >= 4) return; - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_plane *plane = to_intel_plane(crtc->base.primary); struct intel_crtc *plane_crtc; @@ -386,7 +384,7 @@ intel_sanitize_plane_mapping(struct drm_i915_private *i915) if (pipe == crtc->pipe) continue; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[PLANE:%d:%s] attached to the wrong pipe, disabling plane\n", plane->base.base.id, plane->base.name); @@ -423,12 +421,12 @@ static bool intel_crtc_needs_link_reset(struct intel_crtc *crtc) static struct intel_connector *intel_encoder_find_connector(struct intel_encoder *encoder) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct drm_connector_list_iter conn_iter; struct intel_connector *connector; struct intel_connector *found_connector = NULL; - drm_connector_list_iter_begin(&i915->drm, &conn_iter); + drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { if (&encoder->base == connector->base.encoder) { found_connector = connector; @@ -466,7 +464,7 @@ static void intel_sanitize_fifo_underrun_reporting(const struct intel_crtc_state static bool intel_sanitize_crtc(struct intel_crtc *crtc, struct drm_modeset_acquire_ctx *ctx) { - struct drm_i915_private *i915 = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(crtc); struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); bool needs_link_reset; @@ -474,7 +472,7 @@ static bool intel_sanitize_crtc(struct intel_crtc *crtc, struct intel_plane *plane; /* Disable everything but the primary plane */ - for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) { + for_each_intel_plane_on_crtc(display->drm, crtc, plane) { const struct intel_plane_state *plane_state = to_intel_plane_state(plane->base.state); @@ -515,7 +513,7 @@ static bool intel_sanitize_crtc(struct intel_crtc *crtc, return true; } -static void intel_sanitize_all_crtcs(struct drm_i915_private *i915, +static void intel_sanitize_all_crtcs(struct intel_display *display, struct drm_modeset_acquire_ctx *ctx) { struct intel_crtc *crtc; @@ -530,7 +528,7 @@ static void intel_sanitize_all_crtcs(struct drm_i915_private *i915, for (;;) { u32 old_mask = crtcs_forced_off; - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { u32 crtc_mask = drm_crtc_mask(&crtc->base); if (crtcs_forced_off & crtc_mask) @@ -543,7 +541,7 @@ static void intel_sanitize_all_crtcs(struct drm_i915_private *i915, break; } - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); @@ -553,7 +551,7 @@ static void intel_sanitize_all_crtcs(struct drm_i915_private *i915, static bool has_bogus_dpll_config(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev); + struct intel_display *display = to_intel_display(crtc_state); /* * Some SNB BIOSen (eg. ASUS K53SV) are known to misprogram @@ -565,7 +563,7 @@ static bool has_bogus_dpll_config(const struct intel_crtc_state *crtc_state) * without several WARNs, but for now let's take the easy * road. */ - return IS_SANDYBRIDGE(i915) && + return display->platform.sandybridge && crtc_state->hw.active && crtc_state->shared_dpll && crtc_state->port_clock == 0; @@ -574,13 +572,12 @@ static bool has_bogus_dpll_config(const struct intel_crtc_state *crtc_state) static void intel_sanitize_encoder(struct intel_encoder *encoder) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_connector *connector; struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc); struct intel_crtc_state *crtc_state = crtc ? to_intel_crtc_state(crtc->base.state) : NULL; struct intel_pmdemand_state *pmdemand_state = - to_intel_pmdemand_state(i915->display.pmdemand.obj.state); + to_intel_pmdemand_state(display->pmdemand.obj.state); /* * We need to check both for a crtc link (meaning that the encoder is @@ -591,7 +588,7 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder) crtc_state->hw.active; if (crtc_state && has_bogus_dpll_config(crtc_state)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "BIOS has misprogrammed the hardware. Disabling pipe %c\n", pipe_name(crtc->pipe)); has_active_crtc = false; @@ -599,7 +596,7 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder) connector = intel_encoder_find_connector(encoder); if (connector && !has_active_crtc) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] has active connectors but no active pipe!\n", encoder->base.base.id, encoder->base.name); @@ -616,7 +613,7 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder) if (crtc_state) { struct drm_encoder *best_encoder; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] manually disabled\n", encoder->base.base.id, encoder->base.name); @@ -650,18 +647,17 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder) /* notify opregion of the sanitized encoder state */ intel_opregion_notify_encoder(encoder, connector && has_active_crtc); - if (HAS_DDI(i915)) + if (HAS_DDI(display)) intel_ddi_sanitize_encoder_pll_mapping(encoder); } /* FIXME read out full plane state for all planes */ -static void readout_plane_state(struct drm_i915_private *i915) +static void readout_plane_state(struct intel_display *display) { - struct intel_display *display = &i915->display; struct intel_plane *plane; struct intel_crtc *crtc; - for_each_intel_plane(&i915->drm, plane) { + for_each_intel_plane(display->drm, plane) { struct intel_plane_state *plane_state = to_intel_plane_state(plane->base.state); struct intel_crtc_state *crtc_state; @@ -675,13 +671,13 @@ static void readout_plane_state(struct drm_i915_private *i915) intel_set_plane_visible(crtc_state, plane_state, visible); - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[PLANE:%d:%s] hw state readout: %s, pipe %c\n", plane->base.base.id, plane->base.name, str_enabled_disabled(visible), pipe_name(pipe)); } - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); @@ -689,18 +685,17 @@ static void readout_plane_state(struct drm_i915_private *i915) } } -static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) +static void intel_modeset_readout_hw_state(struct intel_display *display) { - struct intel_display *display = &i915->display; struct intel_pmdemand_state *pmdemand_state = - to_intel_pmdemand_state(i915->display.pmdemand.obj.state); + to_intel_pmdemand_state(display->pmdemand.obj.state); enum pipe pipe; struct intel_crtc *crtc; struct intel_encoder *encoder; struct intel_connector *connector; struct drm_connector_list_iter conn_iter; - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); @@ -715,15 +710,15 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) crtc->base.enabled = crtc_state->hw.enable; crtc->active = crtc_state->hw.active; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[CRTC:%d:%s] hw state readout: %s\n", crtc->base.base.id, crtc->base.name, str_enabled_disabled(crtc_state->hw.active)); } - readout_plane_state(i915); + readout_plane_state(display); - for_each_intel_encoder(&i915->drm, encoder) { + for_each_intel_encoder(display->drm, encoder) { struct intel_crtc_state *crtc_state = NULL; pipe = 0; @@ -742,7 +737,7 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) /* encoder should read be linked to joiner primary */ WARN_ON(intel_crtc_is_joiner_secondary(crtc_state)); - for_each_intel_crtc_in_pipe_mask(&i915->drm, secondary_crtc, + for_each_intel_crtc_in_pipe_mask(display->drm, secondary_crtc, intel_crtc_joiner_secondary_pipes(crtc_state)) { struct intel_crtc_state *secondary_crtc_state; @@ -765,7 +760,7 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) if (encoder->sync_state) encoder->sync_state(encoder, crtc_state); - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] hw state readout: %s, pipe %c\n", encoder->base.base.id, encoder->base.name, str_enabled_disabled(encoder->base.crtc), @@ -774,7 +769,7 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) intel_dpll_readout_hw_state(display); - drm_connector_list_iter_begin(&i915->drm, &conn_iter); + drm_connector_list_iter_begin(display->drm, &conn_iter); for_each_intel_connector_iter(connector, &conn_iter) { struct intel_crtc_state *crtc_state = NULL; @@ -808,14 +803,14 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) if (connector->sync_state) connector->sync_state(connector, crtc_state); - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s] hw state readout: %s\n", connector->base.base.id, connector->base.name, str_enabled_disabled(connector->base.encoder)); } drm_connector_list_iter_end(&conn_iter); - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_plane *plane; @@ -838,7 +833,7 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) intel_crtc_copy_hw_to_uapi_state(crtc_state); } - for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) { + for_each_intel_plane_on_crtc(display->drm, crtc, plane) { const struct intel_plane_state *plane_state = to_intel_plane_state(plane->base.state); @@ -854,14 +849,14 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) * use plane->min_cdclk() :( */ if (plane_state->uapi.visible && plane->min_cdclk) { - if (crtc_state->double_wide || DISPLAY_VER(i915) >= 10) + if (crtc_state->double_wide || DISPLAY_VER(display) >= 10) crtc_state->min_cdclk[plane->id] = DIV_ROUND_UP(crtc_state->pixel_rate, 2); else crtc_state->min_cdclk[plane->id] = crtc_state->pixel_rate; } - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[PLANE:%d:%s] min_cdclk %d kHz\n", plane->base.base.id, plane->base.name, crtc_state->min_cdclk[plane->id]); @@ -882,11 +877,11 @@ static void intel_modeset_readout_hw_state(struct drm_i915_private *i915) } static void -get_encoder_power_domains(struct drm_i915_private *i915) +get_encoder_power_domains(struct intel_display *display) { struct intel_encoder *encoder; - for_each_intel_encoder(&i915->drm, encoder) { + for_each_intel_encoder(display->drm, encoder) { struct intel_crtc_state *crtc_state; if (!encoder->get_power_domains) @@ -904,47 +899,49 @@ get_encoder_power_domains(struct drm_i915_private *i915) } } -static void intel_early_display_was(struct drm_i915_private *i915) +static void intel_early_display_was(struct intel_display *display) { /* * Display WA #1185 WaDisableDARBFClkGating:glk,icl,ehl,tgl * Also known as Wa_14010480278. */ - if (IS_DISPLAY_VER(i915, 10, 12)) - intel_de_rmw(i915, GEN9_CLKGATE_DIS_0, 0, DARBF_GATING_DIS); + if (IS_DISPLAY_VER(display, 10, 12)) + intel_de_rmw(display, GEN9_CLKGATE_DIS_0, 0, DARBF_GATING_DIS); /* * WaRsPkgCStateDisplayPMReq:hsw * System hang if this isn't done before disabling all planes! */ - if (IS_HASWELL(i915)) - intel_de_rmw(i915, CHICKEN_PAR1_1, 0, FORCE_ARB_IDLE_PLANES); + if (display->platform.haswell) + intel_de_rmw(display, CHICKEN_PAR1_1, 0, FORCE_ARB_IDLE_PLANES); - if (IS_KABYLAKE(i915) || IS_COFFEELAKE(i915) || IS_COMETLAKE(i915)) { + if (display->platform.kabylake || display->platform.coffeelake || + display->platform.cometlake) { /* Display WA #1142:kbl,cfl,cml */ - intel_de_rmw(i915, CHICKEN_PAR1_1, + intel_de_rmw(display, CHICKEN_PAR1_1, KBL_ARB_FILL_SPARE_22, KBL_ARB_FILL_SPARE_22); - intel_de_rmw(i915, CHICKEN_MISC_2, + intel_de_rmw(display, CHICKEN_MISC_2, KBL_ARB_FILL_SPARE_13 | KBL_ARB_FILL_SPARE_14, KBL_ARB_FILL_SPARE_14); } } -void intel_modeset_setup_hw_state(struct drm_i915_private *i915, +void intel_modeset_setup_hw_state(struct intel_display *display, struct drm_modeset_acquire_ctx *ctx) { - struct intel_display *display = &i915->display; struct intel_encoder *encoder; struct intel_crtc *crtc; intel_wakeref_t wakeref; wakeref = intel_display_power_get(display, POWER_DOMAIN_INIT); - intel_early_display_was(i915); - intel_modeset_readout_hw_state(i915); + intel_early_display_was(display); + intel_vga_disable(display); + + intel_modeset_readout_hw_state(display); /* HW state is read out, now we need to sanitize this mess. */ - get_encoder_power_domains(i915); + get_encoder_power_domains(display); intel_pch_sanitize(display); @@ -954,7 +951,7 @@ void intel_modeset_setup_hw_state(struct drm_i915_private *i915, * intel_sanitize_plane_mapping() may need to do vblank * waits, so we need vblank interrupts restored beforehand. */ - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); @@ -968,20 +965,20 @@ void intel_modeset_setup_hw_state(struct drm_i915_private *i915, } } - intel_fbc_sanitize(&i915->display); + intel_fbc_sanitize(display); - intel_sanitize_plane_mapping(i915); + intel_sanitize_plane_mapping(display); - for_each_intel_encoder(&i915->drm, encoder) + for_each_intel_encoder(display->drm, encoder) intel_sanitize_encoder(encoder); /* * Sanitizing CRTCs needs their connector atomic state to be * up-to-date, so ensure that already here. */ - intel_modeset_update_connector_atomic_state(i915); + intel_modeset_update_connector_atomic_state(display); - intel_sanitize_all_crtcs(i915, ctx); + intel_sanitize_all_crtcs(display, ctx); intel_dpll_sanitize_state(display); @@ -990,13 +987,13 @@ void intel_modeset_setup_hw_state(struct drm_i915_private *i915, intel_wm_get_hw_state(display); intel_wm_sanitize(display); - for_each_intel_crtc(&i915->drm, crtc) { + for_each_intel_crtc(display->drm, crtc) { struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_power_domain_mask put_domains; intel_modeset_get_crtc_power_domains(crtc_state, &put_domains); - if (drm_WARN_ON(&i915->drm, !bitmap_empty(put_domains.bits, POWER_DOMAIN_NUM))) + if (drm_WARN_ON(display->drm, !bitmap_empty(put_domains.bits, POWER_DOMAIN_NUM))) intel_modeset_put_crtc_power_domains(crtc, &put_domains); } diff --git a/drivers/gpu/drm/i915/display/intel_modeset_setup.h b/drivers/gpu/drm/i915/display/intel_modeset_setup.h index 3beff67b33d0..f5e6f3ae9572 100644 --- a/drivers/gpu/drm/i915/display/intel_modeset_setup.h +++ b/drivers/gpu/drm/i915/display/intel_modeset_setup.h @@ -6,10 +6,10 @@ #ifndef __INTEL_MODESET_SETUP_H__ #define __INTEL_MODESET_SETUP_H__ -struct drm_i915_private; struct drm_modeset_acquire_ctx; +struct intel_display; -void intel_modeset_setup_hw_state(struct drm_i915_private *i915, +void intel_modeset_setup_hw_state(struct intel_display *display, struct drm_modeset_acquire_ctx *ctx); #endif /* __INTEL_MODESET_SETUP_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_modeset_verify.c b/drivers/gpu/drm/i915/display/intel_modeset_verify.c index a008412fdd04..766a9983665a 100644 --- a/drivers/gpu/drm/i915/display/intel_modeset_verify.c +++ b/drivers/gpu/drm/i915/display/intel_modeset_verify.c @@ -6,13 +6,14 @@ */ #include <drm/drm_atomic_state_helper.h> +#include <drm/drm_print.h> -#include "i915_drv.h" #include "intel_atomic.h" #include "intel_crtc.h" #include "intel_crtc_state_dump.h" #include "intel_cx0_phy.h" #include "intel_display.h" +#include "intel_display_core.h" #include "intel_display_types.h" #include "intel_fdi.h" #include "intel_modeset_verify.h" @@ -28,9 +29,8 @@ static void intel_connector_verify_state(const struct intel_crtc_state *crtc_sta { struct intel_connector *connector = to_intel_connector(conn_state->connector); struct intel_display *display = to_intel_display(connector); - struct drm_i915_private *i915 = to_i915(connector->base.dev); - drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s]\n", + drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s]\n", connector->base.base.id, connector->base.name); if (connector->get_hw_state(connector)) { @@ -91,7 +91,6 @@ verify_connector_state(struct intel_atomic_state *state, static void intel_pipe_config_sanity_check(const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); - struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev); if (crtc_state->has_pch_encoder) { int fdi_dotclock = intel_dotclock_calculate(intel_fdi_link_freq(display, crtc_state), @@ -103,7 +102,7 @@ static void intel_pipe_config_sanity_check(const struct intel_crtc_state *crtc_s * Yell if the encoder disagrees. Allow for slight * rounding differences. */ - drm_WARN(&i915->drm, abs(fdi_dotclock - dotclock) > 1, + drm_WARN(display->drm, abs(fdi_dotclock - dotclock) > 1, "FDI dotclock and encoder dotclock mismatch, fdi: %i, encoder: %i\n", fdi_dotclock, dotclock); } @@ -113,17 +112,16 @@ static void verify_encoder_state(struct intel_atomic_state *state) { struct intel_display *display = to_intel_display(state); - struct drm_i915_private *i915 = to_i915(state->base.dev); struct intel_encoder *encoder; struct drm_connector *connector; const struct drm_connector_state *old_conn_state, *new_conn_state; int i; - for_each_intel_encoder(&i915->drm, encoder) { + for_each_intel_encoder(display->drm, encoder) { bool enabled = false, found = false; enum pipe pipe; - drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s]\n", + drm_dbg_kms(display->drm, "[ENCODER:%d:%s]\n", encoder->base.base.id, encoder->base.name); @@ -166,7 +164,6 @@ verify_crtc_state(struct intel_atomic_state *state, struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(state); - struct drm_i915_private *i915 = to_i915(display->drm); const struct intel_crtc_state *sw_crtc_state = intel_atomic_get_new_crtc_state(state, crtc); struct intel_crtc_state *hw_crtc_state; @@ -185,7 +182,7 @@ verify_crtc_state(struct intel_atomic_state *state, intel_crtc_get_pipe_config(hw_crtc_state); /* we keep both pipes enabled on 830 */ - if (IS_I830(i915) && hw_crtc_state->hw.active) + if (display->platform.i830 && hw_crtc_state->hw.active) hw_crtc_state->hw.active = sw_crtc_state->hw.active; INTEL_DISPLAY_STATE_WARN(display, diff --git a/drivers/gpu/drm/i915/display/intel_overlay.c b/drivers/gpu/drm/i915/display/intel_overlay.c index aff9a3455c1b..12308495afa5 100644 --- a/drivers/gpu/drm/i915/display/intel_overlay.c +++ b/drivers/gpu/drm/i915/display/intel_overlay.c @@ -291,7 +291,6 @@ static void intel_overlay_flip_prepare(struct intel_overlay *overlay, struct i915_vma *vma) { struct intel_display *display = overlay->display; - struct drm_i915_private *i915 = to_i915(display->drm); enum pipe pipe = overlay->crtc->pipe; struct intel_frontbuffer *frontbuffer = NULL; @@ -307,7 +306,7 @@ static void intel_overlay_flip_prepare(struct intel_overlay *overlay, intel_frontbuffer_put(overlay->frontbuffer); overlay->frontbuffer = frontbuffer; - intel_frontbuffer_flip_prepare(i915, INTEL_FRONTBUFFER_OVERLAY(pipe)); + intel_frontbuffer_flip_prepare(display, INTEL_FRONTBUFFER_OVERLAY(pipe)); overlay->old_vma = overlay->vma; if (vma) @@ -359,14 +358,13 @@ static int intel_overlay_continue(struct intel_overlay *overlay, static void intel_overlay_release_old_vma(struct intel_overlay *overlay) { struct intel_display *display = overlay->display; - struct drm_i915_private *i915 = to_i915(display->drm); struct i915_vma *vma; vma = fetch_and_zero(&overlay->old_vma); if (drm_WARN_ON(display->drm, !vma)) return; - intel_frontbuffer_flip_complete(i915, INTEL_FRONTBUFFER_OVERLAY(overlay->crtc->pipe)); + intel_frontbuffer_flip_complete(display, INTEL_FRONTBUFFER_OVERLAY(overlay->crtc->pipe)); i915_vma_unpin(vma); i915_vma_put(vma); diff --git a/drivers/gpu/drm/i915/display/intel_pch.c b/drivers/gpu/drm/i915/display/intel_pch.c new file mode 100644 index 000000000000..469e8a3cfb49 --- /dev/null +++ b/drivers/gpu/drm/i915/display/intel_pch.c @@ -0,0 +1,340 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright 2025 Intel Corporation. + */ + +#include <drm/drm_print.h> + +#include "i915_utils.h" +#include "intel_display_core.h" +#include "intel_pch.h" + +#define INTEL_PCH_DEVICE_ID_MASK 0xff80 +#define INTEL_PCH_IBX_DEVICE_ID_TYPE 0x3b00 +#define INTEL_PCH_CPT_DEVICE_ID_TYPE 0x1c00 +#define INTEL_PCH_PPT_DEVICE_ID_TYPE 0x1e00 +#define INTEL_PCH_LPT_DEVICE_ID_TYPE 0x8c00 +#define INTEL_PCH_LPT_LP_DEVICE_ID_TYPE 0x9c00 +#define INTEL_PCH_WPT_DEVICE_ID_TYPE 0x8c80 +#define INTEL_PCH_WPT_LP_DEVICE_ID_TYPE 0x9c80 +#define INTEL_PCH_SPT_DEVICE_ID_TYPE 0xA100 +#define INTEL_PCH_SPT_LP_DEVICE_ID_TYPE 0x9D00 +#define INTEL_PCH_KBP_DEVICE_ID_TYPE 0xA280 +#define INTEL_PCH_CNP_DEVICE_ID_TYPE 0xA300 +#define INTEL_PCH_CNP_LP_DEVICE_ID_TYPE 0x9D80 +#define INTEL_PCH_CMP_DEVICE_ID_TYPE 0x0280 +#define INTEL_PCH_CMP2_DEVICE_ID_TYPE 0x0680 +#define INTEL_PCH_CMP_V_DEVICE_ID_TYPE 0xA380 +#define INTEL_PCH_ICP_DEVICE_ID_TYPE 0x3480 +#define INTEL_PCH_ICP2_DEVICE_ID_TYPE 0x3880 +#define INTEL_PCH_MCC_DEVICE_ID_TYPE 0x4B00 +#define INTEL_PCH_TGP_DEVICE_ID_TYPE 0xA080 +#define INTEL_PCH_TGP2_DEVICE_ID_TYPE 0x4380 +#define INTEL_PCH_JSP_DEVICE_ID_TYPE 0x4D80 +#define INTEL_PCH_ADP_DEVICE_ID_TYPE 0x7A80 +#define INTEL_PCH_ADP2_DEVICE_ID_TYPE 0x5180 +#define INTEL_PCH_ADP3_DEVICE_ID_TYPE 0x7A00 +#define INTEL_PCH_ADP4_DEVICE_ID_TYPE 0x5480 +#define INTEL_PCH_P2X_DEVICE_ID_TYPE 0x7100 +#define INTEL_PCH_P3X_DEVICE_ID_TYPE 0x7000 +#define INTEL_PCH_QEMU_DEVICE_ID_TYPE 0x2900 /* qemu q35 has 2918 */ + +/* + * Check for platforms where the south display is on the same PCI device or SoC + * die as the north display. The PCH (if it even exists) is not involved in + * display. Return a fake PCH type for south display handling on these + * platforms, without actually detecting the PCH, and PCH_NONE otherwise. + */ +static enum intel_pch intel_pch_fake_for_south_display(struct intel_display *display) +{ + enum intel_pch pch_type = PCH_NONE; + + if (DISPLAY_VER(display) >= 20) + pch_type = PCH_LNL; + else if (display->platform.battlemage || display->platform.meteorlake) + pch_type = PCH_MTL; + else if (display->platform.dg2) + pch_type = PCH_DG2; + else if (display->platform.dg1) + pch_type = PCH_DG1; + + return pch_type; +} + +/* Map PCH device id to PCH type, or PCH_NONE if unknown. */ +static enum intel_pch +intel_pch_type(const struct intel_display *display, unsigned short id) +{ + switch (id) { + case INTEL_PCH_IBX_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Ibex Peak PCH\n"); + drm_WARN_ON(display->drm, DISPLAY_VER(display) != 5); + return PCH_IBX; + case INTEL_PCH_CPT_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found CougarPoint PCH\n"); + drm_WARN_ON(display->drm, + DISPLAY_VER(display) != 6 && + !display->platform.ivybridge); + return PCH_CPT; + case INTEL_PCH_PPT_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found PantherPoint PCH\n"); + drm_WARN_ON(display->drm, + DISPLAY_VER(display) != 6 && + !display->platform.ivybridge); + /* PPT is CPT compatible */ + return PCH_CPT; + case INTEL_PCH_LPT_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found LynxPoint PCH\n"); + drm_WARN_ON(display->drm, + !display->platform.haswell && + !display->platform.broadwell); + drm_WARN_ON(display->drm, + display->platform.haswell_ult || + display->platform.broadwell_ult); + return PCH_LPT_H; + case INTEL_PCH_LPT_LP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found LynxPoint LP PCH\n"); + drm_WARN_ON(display->drm, + !display->platform.haswell && + !display->platform.broadwell); + drm_WARN_ON(display->drm, + !display->platform.haswell_ult && + !display->platform.broadwell_ult); + return PCH_LPT_LP; + case INTEL_PCH_WPT_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found WildcatPoint PCH\n"); + drm_WARN_ON(display->drm, + !display->platform.haswell && + !display->platform.broadwell); + drm_WARN_ON(display->drm, + display->platform.haswell_ult || + display->platform.broadwell_ult); + /* WPT is LPT compatible */ + return PCH_LPT_H; + case INTEL_PCH_WPT_LP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found WildcatPoint LP PCH\n"); + drm_WARN_ON(display->drm, + !display->platform.haswell && + !display->platform.broadwell); + drm_WARN_ON(display->drm, + !display->platform.haswell_ult && + !display->platform.broadwell_ult); + /* WPT is LPT compatible */ + return PCH_LPT_LP; + case INTEL_PCH_SPT_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found SunrisePoint PCH\n"); + drm_WARN_ON(display->drm, + !display->platform.skylake && + !display->platform.kabylake && + !display->platform.coffeelake); + return PCH_SPT; + case INTEL_PCH_SPT_LP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found SunrisePoint LP PCH\n"); + drm_WARN_ON(display->drm, + !display->platform.skylake && + !display->platform.kabylake && + !display->platform.coffeelake && + !display->platform.cometlake); + return PCH_SPT; + case INTEL_PCH_KBP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Kaby Lake PCH (KBP)\n"); + drm_WARN_ON(display->drm, + !display->platform.skylake && + !display->platform.kabylake && + !display->platform.coffeelake && + !display->platform.cometlake); + /* KBP is SPT compatible */ + return PCH_SPT; + case INTEL_PCH_CNP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Cannon Lake PCH (CNP)\n"); + drm_WARN_ON(display->drm, + !display->platform.coffeelake && + !display->platform.cometlake); + return PCH_CNP; + case INTEL_PCH_CNP_LP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, + "Found Cannon Lake LP PCH (CNP-LP)\n"); + drm_WARN_ON(display->drm, + !display->platform.coffeelake && + !display->platform.cometlake); + return PCH_CNP; + case INTEL_PCH_CMP_DEVICE_ID_TYPE: + case INTEL_PCH_CMP2_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Comet Lake PCH (CMP)\n"); + drm_WARN_ON(display->drm, + !display->platform.coffeelake && + !display->platform.cometlake && + !display->platform.rocketlake); + /* CMP is CNP compatible */ + return PCH_CNP; + case INTEL_PCH_CMP_V_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Comet Lake V PCH (CMP-V)\n"); + drm_WARN_ON(display->drm, + !display->platform.coffeelake && + !display->platform.cometlake); + /* CMP-V is based on KBP, which is SPT compatible */ + return PCH_SPT; + case INTEL_PCH_ICP_DEVICE_ID_TYPE: + case INTEL_PCH_ICP2_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Ice Lake PCH\n"); + drm_WARN_ON(display->drm, !display->platform.icelake); + return PCH_ICP; + case INTEL_PCH_MCC_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Mule Creek Canyon PCH\n"); + drm_WARN_ON(display->drm, !(display->platform.jasperlake || + display->platform.elkhartlake)); + /* MCC is TGP compatible */ + return PCH_TGP; + case INTEL_PCH_TGP_DEVICE_ID_TYPE: + case INTEL_PCH_TGP2_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Tiger Lake LP PCH\n"); + drm_WARN_ON(display->drm, !display->platform.tigerlake && + !display->platform.rocketlake && + !display->platform.skylake && + !display->platform.kabylake && + !display->platform.coffeelake && + !display->platform.cometlake); + return PCH_TGP; + case INTEL_PCH_JSP_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Jasper Lake PCH\n"); + drm_WARN_ON(display->drm, !(display->platform.jasperlake || + display->platform.elkhartlake)); + /* JSP is ICP compatible */ + return PCH_ICP; + case INTEL_PCH_ADP_DEVICE_ID_TYPE: + case INTEL_PCH_ADP2_DEVICE_ID_TYPE: + case INTEL_PCH_ADP3_DEVICE_ID_TYPE: + case INTEL_PCH_ADP4_DEVICE_ID_TYPE: + drm_dbg_kms(display->drm, "Found Alder Lake PCH\n"); + drm_WARN_ON(display->drm, !display->platform.alderlake_s && + !display->platform.alderlake_p); + return PCH_ADP; + default: + return PCH_NONE; + } +} + +static bool intel_is_virt_pch(unsigned short id, + unsigned short svendor, unsigned short sdevice) +{ + return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE || + id == INTEL_PCH_P3X_DEVICE_ID_TYPE || + (id == INTEL_PCH_QEMU_DEVICE_ID_TYPE && + svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET && + sdevice == PCI_SUBDEVICE_ID_QEMU)); +} + +static void +intel_virt_detect_pch(const struct intel_display *display, + unsigned short *pch_id, enum intel_pch *pch_type) +{ + unsigned short id = 0; + + /* + * In a virtualized passthrough environment we can be in a + * setup where the ISA bridge is not able to be passed through. + * In this case, a south bridge can be emulated and we have to + * make an educated guess as to which PCH is really there. + */ + + if (display->platform.alderlake_s || display->platform.alderlake_p) + id = INTEL_PCH_ADP_DEVICE_ID_TYPE; + else if (display->platform.tigerlake || display->platform.rocketlake) + id = INTEL_PCH_TGP_DEVICE_ID_TYPE; + else if (display->platform.jasperlake || display->platform.elkhartlake) + id = INTEL_PCH_MCC_DEVICE_ID_TYPE; + else if (display->platform.icelake) + id = INTEL_PCH_ICP_DEVICE_ID_TYPE; + else if (display->platform.coffeelake || + display->platform.cometlake) + id = INTEL_PCH_CNP_DEVICE_ID_TYPE; + else if (display->platform.kabylake || display->platform.skylake) + id = INTEL_PCH_SPT_DEVICE_ID_TYPE; + else if (display->platform.haswell_ult || + display->platform.broadwell_ult) + id = INTEL_PCH_LPT_LP_DEVICE_ID_TYPE; + else if (display->platform.haswell || display->platform.broadwell) + id = INTEL_PCH_LPT_DEVICE_ID_TYPE; + else if (DISPLAY_VER(display) == 6 || display->platform.ivybridge) + id = INTEL_PCH_CPT_DEVICE_ID_TYPE; + else if (DISPLAY_VER(display) == 5) + id = INTEL_PCH_IBX_DEVICE_ID_TYPE; + + if (id) + drm_dbg_kms(display->drm, "Assuming PCH ID %04x\n", id); + else + drm_dbg_kms(display->drm, "Assuming no PCH\n"); + + *pch_type = intel_pch_type(display, id); + + /* Sanity check virtual PCH id */ + if (drm_WARN_ON(display->drm, + id && *pch_type == PCH_NONE)) + id = 0; + + *pch_id = id; +} + +void intel_pch_detect(struct intel_display *display) +{ + struct pci_dev *pch = NULL; + unsigned short id; + enum intel_pch pch_type; + + pch_type = intel_pch_fake_for_south_display(display); + if (pch_type != PCH_NONE) { + display->pch_type = pch_type; + drm_dbg_kms(display->drm, + "PCH not involved in display, using fake PCH type %d for south display\n", + pch_type); + return; + } + + /* + * The reason to probe ISA bridge instead of Dev31:Fun0 is to + * make graphics device passthrough work easy for VMM, that only + * need to expose ISA bridge to let driver know the real hardware + * underneath. This is a requirement from virtualization team. + * + * In some virtualized environments (e.g. XEN), there is irrelevant + * ISA bridge in the system. To work reliably, we should scan through + * all the ISA bridge devices and check for the first match, instead + * of only checking the first one. + */ + while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) { + if (pch->vendor != PCI_VENDOR_ID_INTEL) + continue; + + id = pch->device & INTEL_PCH_DEVICE_ID_MASK; + + pch_type = intel_pch_type(display, id); + if (pch_type != PCH_NONE) { + display->pch_type = pch_type; + break; + } else if (intel_is_virt_pch(id, pch->subsystem_vendor, + pch->subsystem_device)) { + intel_virt_detect_pch(display, &id, &pch_type); + display->pch_type = pch_type; + break; + } + } + + /* + * Use PCH_NOP (PCH but no South Display) for PCH platforms without + * display. + */ + if (pch && !HAS_DISPLAY(display)) { + drm_dbg_kms(display->drm, + "Display disabled, reverting to NOP PCH\n"); + display->pch_type = PCH_NOP; + } else if (!pch) { + if (i915_run_as_guest() && HAS_DISPLAY(display)) { + intel_virt_detect_pch(display, &id, &pch_type); + display->pch_type = pch_type; + } else { + drm_dbg_kms(display->drm, "No PCH found.\n"); + } + } + + pci_dev_put(pch); +} diff --git a/drivers/gpu/drm/i915/display/intel_pch.h b/drivers/gpu/drm/i915/display/intel_pch.h new file mode 100644 index 000000000000..cf4dab1b98bf --- /dev/null +++ b/drivers/gpu/drm/i915/display/intel_pch.h @@ -0,0 +1,58 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright 2025 Intel Corporation. + */ + +#ifndef __INTEL_PCH__ +#define __INTEL_PCH__ + +#include "intel_display_conversion.h" + +struct intel_display; + +/* + * Sorted by south display engine compatibility. + * If the new PCH comes with a south display engine that is not + * inherited from the latest item, please do not add it to the + * end. Instead, add it right after its "parent" PCH. + */ +enum intel_pch { + PCH_NOP = -1, /* PCH without south display */ + PCH_NONE = 0, /* No PCH present */ + PCH_IBX, /* Ibexpeak PCH */ + PCH_CPT, /* Cougarpoint/Pantherpoint PCH */ + PCH_LPT_H, /* Lynxpoint/Wildcatpoint H PCH */ + PCH_LPT_LP, /* Lynxpoint/Wildcatpoint LP PCH */ + PCH_SPT, /* Sunrisepoint/Kaby Lake PCH */ + PCH_CNP, /* Cannon/Comet Lake PCH */ + PCH_ICP, /* Ice Lake/Jasper Lake PCH */ + PCH_TGP, /* Tiger Lake/Mule Creek Canyon PCH */ + PCH_ADP, /* Alder Lake PCH */ + + /* Fake PCHs, functionality handled on the same PCI dev */ + PCH_DG1 = 1024, + PCH_DG2, + PCH_MTL, + PCH_LNL, +}; + +#define INTEL_PCH_TYPE(_display) (__to_intel_display(_display)->pch_type) +#define HAS_PCH_DG2(display) (INTEL_PCH_TYPE(display) == PCH_DG2) +#define HAS_PCH_ADP(display) (INTEL_PCH_TYPE(display) == PCH_ADP) +#define HAS_PCH_DG1(display) (INTEL_PCH_TYPE(display) == PCH_DG1) +#define HAS_PCH_TGP(display) (INTEL_PCH_TYPE(display) == PCH_TGP) +#define HAS_PCH_ICP(display) (INTEL_PCH_TYPE(display) == PCH_ICP) +#define HAS_PCH_CNP(display) (INTEL_PCH_TYPE(display) == PCH_CNP) +#define HAS_PCH_SPT(display) (INTEL_PCH_TYPE(display) == PCH_SPT) +#define HAS_PCH_LPT_H(display) (INTEL_PCH_TYPE(display) == PCH_LPT_H) +#define HAS_PCH_LPT_LP(display) (INTEL_PCH_TYPE(display) == PCH_LPT_LP) +#define HAS_PCH_LPT(display) (INTEL_PCH_TYPE(display) == PCH_LPT_H || \ + INTEL_PCH_TYPE(display) == PCH_LPT_LP) +#define HAS_PCH_CPT(display) (INTEL_PCH_TYPE(display) == PCH_CPT) +#define HAS_PCH_IBX(display) (INTEL_PCH_TYPE(display) == PCH_IBX) +#define HAS_PCH_NOP(display) (INTEL_PCH_TYPE(display) == PCH_NOP) +#define HAS_PCH_SPLIT(display) (INTEL_PCH_TYPE(display) != PCH_NONE) + +void intel_pch_detect(struct intel_display *display); + +#endif /* __INTEL_PCH__ */ diff --git a/drivers/gpu/drm/i915/display/intel_pch_display.c b/drivers/gpu/drm/i915/display/intel_pch_display.c index b909ed18a5b2..1743ebf551cb 100644 --- a/drivers/gpu/drm/i915/display/intel_pch_display.c +++ b/drivers/gpu/drm/i915/display/intel_pch_display.c @@ -3,8 +3,9 @@ * Copyright © 2021 Intel Corporation */ +#include <drm/drm_print.h> + #include "g4x_dp.h" -#include "i915_drv.h" #include "i915_reg.h" #include "intel_crt.h" #include "intel_crt_regs.h" @@ -23,17 +24,15 @@ bool intel_has_pch_trancoder(struct intel_display *display, enum pipe pch_transcoder) { - struct drm_i915_private *i915 = to_i915(display->drm); - - return HAS_PCH_IBX(i915) || HAS_PCH_CPT(i915) || - (HAS_PCH_LPT_H(i915) && pch_transcoder == PIPE_A); + return HAS_PCH_IBX(display) || HAS_PCH_CPT(display) || + (HAS_PCH_LPT_H(display) && pch_transcoder == PIPE_A); } enum pipe intel_crtc_pch_transcoder(struct intel_crtc *crtc) { - struct drm_i915_private *i915 = to_i915(crtc->base.dev); + struct intel_display *display = to_intel_display(crtc); - if (HAS_PCH_LPT(i915)) + if (HAS_PCH_LPT(display)) return PIPE_A; else return crtc->pipe; @@ -43,7 +42,6 @@ static void assert_pch_dp_disabled(struct intel_display *display, enum pipe pipe, enum port port, i915_reg_t dp_reg) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe port_pipe; bool state; @@ -54,7 +52,7 @@ static void assert_pch_dp_disabled(struct intel_display *display, port_name(port), pipe_name(pipe)); INTEL_DISPLAY_STATE_WARN(display, - HAS_PCH_IBX(dev_priv) && !state && port_pipe == PIPE_B, + HAS_PCH_IBX(display) && !state && port_pipe == PIPE_B, "IBX PCH DP %c still using transcoder B\n", port_name(port)); } @@ -63,7 +61,6 @@ static void assert_pch_hdmi_disabled(struct intel_display *display, enum pipe pipe, enum port port, i915_reg_t hdmi_reg) { - struct drm_i915_private *dev_priv = to_i915(display->drm); enum pipe port_pipe; bool state; @@ -74,7 +71,7 @@ static void assert_pch_hdmi_disabled(struct intel_display *display, port_name(port), pipe_name(pipe)); INTEL_DISPLAY_STATE_WARN(display, - HAS_PCH_IBX(dev_priv) && !state && port_pipe == PIPE_B, + HAS_PCH_IBX(display) && !state && port_pipe == PIPE_B, "IBX PCH HDMI %c still using transcoder B\n", port_name(port)); } @@ -249,7 +246,6 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; i915_reg_t reg; u32 val, pipeconf_val; @@ -261,7 +257,7 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state) assert_fdi_tx_enabled(display, pipe); assert_fdi_rx_enabled(display, pipe); - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { reg = TRANS_CHICKEN2(pipe); val = intel_de_read(display, reg); /* @@ -279,7 +275,7 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state) val = intel_de_read(display, reg); pipeconf_val = intel_de_read(display, TRANSCONF(display, pipe)); - if (HAS_PCH_IBX(dev_priv)) { + if (HAS_PCH_IBX(display)) { /* Configure frame start delay to match the CPU */ val &= ~TRANS_FRAME_START_DELAY_MASK; val |= TRANS_FRAME_START_DELAY(crtc_state->framestart_delay - 1); @@ -298,7 +294,7 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state) val &= ~TRANS_INTERLACE_MASK; if ((pipeconf_val & TRANSCONF_INTERLACE_MASK_ILK) == TRANSCONF_INTERLACE_IF_ID_ILK) { - if (HAS_PCH_IBX(dev_priv) && + if (HAS_PCH_IBX(display) && intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO)) val |= TRANS_INTERLACE_LEGACY_VSYNC_IBX; else @@ -316,7 +312,6 @@ static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state) static void ilk_disable_pch_transcoder(struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; i915_reg_t reg; @@ -334,7 +329,7 @@ static void ilk_disable_pch_transcoder(struct intel_crtc *crtc) drm_err(display->drm, "failed to disable transcoder %c\n", pipe_name(pipe)); - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) /* Workaround: Clear the timing override chicken bit again. */ intel_de_rmw(display, TRANS_CHICKEN2(pipe), TRANS_CHICKEN2_TIMING_OVERRIDE, 0); @@ -366,7 +361,6 @@ void ilk_pch_enable(struct intel_atomic_state *state, struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); const struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); enum pipe pipe = crtc->pipe; @@ -381,7 +375,7 @@ void ilk_pch_enable(struct intel_atomic_state *state, * We need to program the right clock selection * before writing the pixel multiplier into the DPLL. */ - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { u32 sel; temp = intel_de_read(display, PCH_DPLL_SEL); @@ -417,7 +411,7 @@ void ilk_pch_enable(struct intel_atomic_state *state, intel_fdi_normal_train(crtc); /* For PCH DP, enable TRANS_DP_CTL */ - if (HAS_PCH_CPT(dev_priv) && + if (HAS_PCH_CPT(display) && intel_crtc_has_dp_encoder(crtc_state)) { const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; @@ -459,14 +453,13 @@ void ilk_pch_post_disable(struct intel_atomic_state *state, struct intel_crtc *crtc) { struct intel_display *display = to_intel_display(crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); const struct intel_crtc_state *old_crtc_state = intel_atomic_get_old_crtc_state(state, crtc); enum pipe pipe = crtc->pipe; ilk_disable_pch_transcoder(crtc); - if (HAS_PCH_CPT(dev_priv)) { + if (HAS_PCH_CPT(display)) { /* disable TRANS_DP_CTL */ intel_de_rmw(display, TRANS_DP_CTL(pipe), TRANS_DP_OUTPUT_ENABLE | TRANS_DP_PORT_SEL_MASK, @@ -503,7 +496,6 @@ void ilk_pch_get_config(struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); struct intel_shared_dpll *pll; enum pipe pipe = crtc->pipe; enum intel_dpll_id pll_id; @@ -522,7 +514,7 @@ void ilk_pch_get_config(struct intel_crtc_state *crtc_state) intel_cpu_transcoder_get_m1_n1(crtc, crtc_state->cpu_transcoder, &crtc_state->fdi_m_n); - if (HAS_PCH_IBX(dev_priv)) { + if (HAS_PCH_IBX(display)) { /* * The pipe->pch transcoder and pch transcoder->pll * mapping is fixed. @@ -646,8 +638,6 @@ void lpt_pch_get_config(struct intel_crtc_state *crtc_state) void intel_pch_sanitize(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - if (HAS_PCH_IBX(i915)) + if (HAS_PCH_IBX(display)) ibx_sanitize_pch_ports(display); } diff --git a/drivers/gpu/drm/i915/display/intel_pch_refclk.c b/drivers/gpu/drm/i915/display/intel_pch_refclk.c index 1307a478861a..693b90e3dfc3 100644 --- a/drivers/gpu/drm/i915/display/intel_pch_refclk.c +++ b/drivers/gpu/drm/i915/display/intel_pch_refclk.c @@ -281,7 +281,7 @@ static void lpt_enable_clkout_dp(struct intel_display *display, if (drm_WARN(display->drm, with_fdi && !with_spread, "FDI requires downspread\n")) with_spread = true; - if (drm_WARN(display->drm, HAS_PCH_LPT_LP(dev_priv) && + if (drm_WARN(display->drm, HAS_PCH_LPT_LP(display) && with_fdi, "LP PCH doesn't have FDI\n")) with_fdi = false; @@ -303,7 +303,7 @@ static void lpt_enable_clkout_dp(struct intel_display *display, lpt_fdi_program_mphy(display); } - reg = HAS_PCH_LPT_LP(dev_priv) ? SBI_GEN0 : SBI_DBUFF0; + reg = HAS_PCH_LPT_LP(display) ? SBI_GEN0 : SBI_DBUFF0; tmp = intel_sbi_read(dev_priv, reg, SBI_ICLK); tmp |= SBI_GEN0_CFG_BUFFENABLE_DISABLE; intel_sbi_write(dev_priv, reg, tmp, SBI_ICLK); @@ -319,7 +319,7 @@ void lpt_disable_clkout_dp(struct intel_display *display) intel_sbi_lock(dev_priv); - reg = HAS_PCH_LPT_LP(dev_priv) ? SBI_GEN0 : SBI_DBUFF0; + reg = HAS_PCH_LPT_LP(display) ? SBI_GEN0 : SBI_DBUFF0; tmp = intel_sbi_read(dev_priv, reg, SBI_ICLK); tmp &= ~SBI_GEN0_CFG_BUFFENABLE_DISABLE; intel_sbi_write(dev_priv, reg, tmp, SBI_ICLK); @@ -498,7 +498,6 @@ static void lpt_init_pch_refclk(struct intel_display *display) static void ilk_init_pch_refclk(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder; struct intel_shared_dpll *pll; int i; @@ -527,7 +526,7 @@ static void ilk_init_pch_refclk(struct intel_display *display) } } - if (HAS_PCH_IBX(dev_priv)) { + if (HAS_PCH_IBX(display)) { has_ck505 = display->vbt.display_clock_mode; can_ssc = has_ck505; } else { @@ -678,10 +677,8 @@ static void ilk_init_pch_refclk(struct intel_display *display) */ void intel_init_pch_refclk(struct intel_display *display) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_IBX(display) || HAS_PCH_CPT(display)) ilk_init_pch_refclk(display); - else if (HAS_PCH_LPT(dev_priv)) + else if (HAS_PCH_LPT(display)) lpt_init_pch_refclk(display); } diff --git a/drivers/gpu/drm/i915/display/intel_plane_initial.c b/drivers/gpu/drm/i915/display/intel_plane_initial.c index b0c4892775ce..c00d9184c586 100644 --- a/drivers/gpu/drm/i915/display/intel_plane_initial.c +++ b/drivers/gpu/drm/i915/display/intel_plane_initial.c @@ -53,9 +53,11 @@ intel_reuse_initial_plane_obj(struct intel_crtc *this, } static enum intel_memory_type -initial_plane_memory_type(struct drm_i915_private *i915) +initial_plane_memory_type(struct intel_display *display) { - if (IS_DGFX(i915)) + struct drm_i915_private *i915 = to_i915(display->drm); + + if (display->platform.dgfx) return INTEL_MEMORY_LOCAL; else if (HAS_LMEMBAR_SMEM_STOLEN(i915)) return INTEL_MEMORY_STOLEN_LOCAL; @@ -75,7 +77,7 @@ initial_plane_phys(struct intel_display *display, dma_addr_t dma_addr; u32 base; - mem_type = initial_plane_memory_type(i915); + mem_type = initial_plane_memory_type(display); mem = intel_memory_region_by_type(i915, mem_type); if (!mem) { drm_dbg_kms(display->drm, diff --git a/drivers/gpu/drm/i915/display/intel_pmdemand.c b/drivers/gpu/drm/i915/display/intel_pmdemand.c index d22b5469672d..1253376c7654 100644 --- a/drivers/gpu/drm/i915/display/intel_pmdemand.c +++ b/drivers/gpu/drm/i915/display/intel_pmdemand.c @@ -478,13 +478,34 @@ static bool intel_pmdemand_req_complete(struct intel_display *display) XELPDP_PMDEMAND_REQ_ENABLE); } -static void intel_pmdemand_wait(struct intel_display *display) +static void intel_pmdemand_poll(struct intel_display *display) { - if (!wait_event_timeout(display->pmdemand.waitqueue, - intel_pmdemand_req_complete(display), - msecs_to_jiffies_timeout(10))) + const unsigned int timeout_ms = 10; + u32 status; + int ret; + + ret = intel_de_wait_custom(display, XELPDP_INITIATE_PMDEMAND_REQUEST(1), + XELPDP_PMDEMAND_REQ_ENABLE, 0, + 50, timeout_ms, &status); + + if (ret == -ETIMEDOUT) drm_err(display->drm, - "timed out waiting for Punit PM Demand Response\n"); + "timed out waiting for Punit PM Demand Response within %ums (status 0x%08x)\n", + timeout_ms, status); +} + +static void intel_pmdemand_wait(struct intel_display *display) +{ + /* Wa_14024400148 For lnl use polling method */ + if (DISPLAY_VER(display) == 20) { + intel_pmdemand_poll(display); + } else { + if (!wait_event_timeout(display->pmdemand.waitqueue, + intel_pmdemand_req_complete(display), + msecs_to_jiffies_timeout(10))) + drm_err(display->drm, + "timed out waiting for Punit PM Demand Response\n"); + } } /* Required to be programmed during Display Init Sequences. */ diff --git a/drivers/gpu/drm/i915/display/intel_pps.c b/drivers/gpu/drm/i915/display/intel_pps.c index 4d4e2b9f5f2d..05e1e5c7e8b7 100644 --- a/drivers/gpu/drm/i915/display/intel_pps.c +++ b/drivers/gpu/drm/i915/display/intel_pps.c @@ -350,21 +350,19 @@ vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp) static int intel_num_pps(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - if (display->platform.valleyview || display->platform.cherryview) return 2; if (display->platform.geminilake || display->platform.broxton) return 2; - if (INTEL_PCH_TYPE(i915) >= PCH_MTL) + if (INTEL_PCH_TYPE(display) >= PCH_MTL) return 2; - if (INTEL_PCH_TYPE(i915) >= PCH_DG1) + if (INTEL_PCH_TYPE(display) >= PCH_DG1) return 1; - if (INTEL_PCH_TYPE(i915) >= PCH_ICP) + if (INTEL_PCH_TYPE(display) >= PCH_ICP) return 2; return 1; @@ -373,11 +371,10 @@ static int intel_num_pps(struct intel_display *display) static bool intel_pps_is_valid(struct intel_dp *intel_dp) { struct intel_display *display = to_intel_display(intel_dp); - struct drm_i915_private *i915 = to_i915(display->drm); if (intel_dp->pps.pps_idx == 1 && - INTEL_PCH_TYPE(i915) >= PCH_ICP && - INTEL_PCH_TYPE(i915) <= PCH_ADP) + INTEL_PCH_TYPE(display) >= PCH_ICP && + INTEL_PCH_TYPE(display) <= PCH_ADP) return intel_de_read(display, SOUTH_CHICKEN1) & ICP_SECOND_PPS_IO_SELECT; return true; @@ -499,7 +496,6 @@ static void intel_pps_get_registers(struct intel_dp *intel_dp, struct pps_registers *regs) { struct intel_display *display = to_intel_display(intel_dp); - struct drm_i915_private *dev_priv = to_i915(display->drm); int pps_idx; memset(regs, 0, sizeof(*regs)); @@ -518,7 +514,7 @@ static void intel_pps_get_registers(struct intel_dp *intel_dp, /* Cycle delay moved from PP_DIVISOR to PP_CONTROL */ if (display->platform.geminilake || display->platform.broxton || - INTEL_PCH_TYPE(dev_priv) >= PCH_CNP) + INTEL_PCH_TYPE(display) >= PCH_CNP) regs->pp_div = INVALID_MMIO_REG; else regs->pp_div = PP_DIVISOR(display, pps_idx); @@ -1591,7 +1587,6 @@ static void pps_init_delays(struct intel_dp *intel_dp) static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd) { struct intel_display *display = to_intel_display(intel_dp); - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 pp_on, pp_off, port_sel = 0; int div = DISPLAY_RUNTIME_INFO(display)->rawclk_freq / 1000; struct pps_registers regs; @@ -1638,7 +1633,7 @@ static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd * power sequencer any more. */ if (display->platform.valleyview || display->platform.cherryview) { port_sel = PANEL_PORT_SELECT_VLV(port); - } else if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) { + } else if (HAS_PCH_IBX(display) || HAS_PCH_CPT(display)) { switch (port) { case PORT_A: port_sel = PANEL_PORT_SELECT_DPA; @@ -1791,9 +1786,7 @@ void intel_pps_unlock_regs_wa(struct intel_display *display) void intel_pps_setup(struct intel_display *display) { - struct drm_i915_private *i915 = to_i915(display->drm); - - if (HAS_PCH_SPLIT(i915) || display->platform.geminilake || display->platform.broxton) + if (HAS_PCH_SPLIT(display) || display->platform.geminilake || display->platform.broxton) display->pps.mmio_base = PCH_PPS_BASE; else if (display->platform.valleyview || display->platform.cherryview) display->pps.mmio_base = VLV_PPS_BASE; @@ -1836,7 +1829,6 @@ void intel_pps_connector_debugfs_add(struct intel_connector *connector) void assert_pps_unlocked(struct intel_display *display, enum pipe pipe) { - struct drm_i915_private *dev_priv = to_i915(display->drm); i915_reg_t pp_reg; u32 val; enum pipe panel_pipe = INVALID_PIPE; @@ -1845,7 +1837,7 @@ void assert_pps_unlocked(struct intel_display *display, enum pipe pipe) if (drm_WARN_ON(display->drm, HAS_DDI(display))) return; - if (HAS_PCH_SPLIT(dev_priv)) { + if (HAS_PCH_SPLIT(display)) { u32 port_sel; pp_reg = PP_CONTROL(display, 0); diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index eef48c014112..430ad4ef7146 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -26,6 +26,7 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_damage_helper.h> #include <drm/drm_debugfs.h> +#include <drm/drm_vblank.h> #include "i915_drv.h" #include "i915_reg.h" @@ -38,6 +39,7 @@ #include "intel_display_irq.h" #include "intel_display_rpm.h" #include "intel_display_types.h" +#include "intel_dmc.h" #include "intel_dp.h" #include "intel_dp_aux.h" #include "intel_frontbuffer.h" @@ -46,6 +48,7 @@ #include "intel_psr_regs.h" #include "intel_snps_phy.h" #include "intel_vblank.h" +#include "intel_vrr.h" #include "skl_universal_plane.h" /** @@ -794,31 +797,10 @@ static void _psr_enable_sink(struct intel_dp *intel_dp, drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, val); } -static void intel_psr_enable_sink_alpm(struct intel_dp *intel_dp, - const struct intel_crtc_state *crtc_state) -{ - u8 val; - - /* - * eDP Panel Replay uses always ALPM - * PSR2 uses ALPM but PSR1 doesn't - */ - if (!intel_dp_is_edp(intel_dp) || (!crtc_state->has_panel_replay && - !crtc_state->has_sel_update)) - return; - - val = DP_ALPM_ENABLE | DP_ALPM_LOCK_ERROR_IRQ_HPD_ENABLE; - - if (crtc_state->has_panel_replay) - val |= DP_ALPM_MODE_AUX_LESS; - - drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG, val); -} - static void intel_psr_enable_sink(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { - intel_psr_enable_sink_alpm(intel_dp, crtc_state); + intel_alpm_enable_sink(intel_dp, crtc_state); crtc_state->has_panel_replay ? _panel_replay_enable_sink(intel_dp, crtc_state) : @@ -905,6 +887,18 @@ static u8 psr_compute_idle_frames(struct intel_dp *intel_dp) return idle_frames; } +static bool is_dc5_dc6_blocked(struct intel_dp *intel_dp) +{ + struct intel_display *display = to_intel_display(intel_dp); + u32 current_dc_state = intel_display_power_get_current_dc_state(display); + struct drm_vblank_crtc *vblank = &display->drm->vblank[intel_dp->psr.pipe]; + + return (current_dc_state != DC_STATE_EN_UPTO_DC5 && + current_dc_state != DC_STATE_EN_UPTO_DC6) || + intel_dp->psr.active_non_psr_pipes || + READ_ONCE(vblank->enabled); +} + static void hsw_activate_psr1(struct intel_dp *intel_dp) { struct intel_display *display = to_intel_display(intel_dp); @@ -933,6 +927,14 @@ static void hsw_activate_psr1(struct intel_dp *intel_dp) intel_de_rmw(display, psr_ctl_reg(display, cpu_transcoder), ~EDP_PSR_RESTORE_PSR_ACTIVE_CTX_MASK, val); + + /* Wa_16025596647 */ + if ((DISPLAY_VER(display) == 20 || + IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) && + is_dc5_dc6_blocked(intel_dp)) + intel_dmc_start_pkgc_exit_at_start_of_undelayed_vblank(display, + intel_dp->psr.pipe, + true); } static u32 intel_psr2_get_tp_time(struct intel_dp *intel_dp) @@ -1014,8 +1016,16 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 val = EDP_PSR2_ENABLE; u32 psr_val = 0; + u8 idle_frames; - val |= EDP_PSR2_IDLE_FRAMES(psr_compute_idle_frames(intel_dp)); + /* Wa_16025596647 */ + if ((DISPLAY_VER(display) == 20 || + IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) && + is_dc5_dc6_blocked(intel_dp)) + idle_frames = 0; + else + idle_frames = psr_compute_idle_frames(intel_dp); + val |= EDP_PSR2_IDLE_FRAMES(idle_frames); if (DISPLAY_VER(display) < 14 && !display->platform.alderlake_p) val |= EDP_SU_TRACK_ENABLE; @@ -1649,6 +1659,9 @@ void intel_psr_compute_config(struct intel_dp *intel_dp, { struct intel_display *display = to_intel_display(intel_dp); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; + struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state); + struct intel_crtc *crtc; + u8 active_pipes = 0; if (!psr_global_enabled(intel_dp)) { drm_dbg_kms(display->drm, "PSR disabled by flag\n"); @@ -1702,6 +1715,24 @@ void intel_psr_compute_config(struct intel_dp *intel_dp, drm_dbg_kms(display->drm, "PSR disabled to workaround PSR FSM hang issue\n"); } + + /* Rest is for Wa_16025596647 */ + if (DISPLAY_VER(display) != 20 && + !IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) + return; + + /* Not needed by Panel Replay */ + if (crtc_state->has_panel_replay) + return; + + /* We ignore possible secondary PSR/Panel Replay capable eDP */ + for_each_intel_crtc(display->drm, crtc) + active_pipes |= crtc->active ? BIT(crtc->pipe) : 0; + + active_pipes = intel_calc_active_pipes(state, active_pipes); + + crtc_state->active_non_psr_pipes = active_pipes & + ~BIT(to_intel_crtc(crtc_state->uapi.crtc)->pipe); } void intel_psr_get_config(struct intel_encoder *encoder, @@ -1893,9 +1924,6 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, intel_dp->psr.psr2_sel_fetch_enabled ? IGNORE_PSR2_HW_TRACKING : 0); - if (intel_dp_is_edp(intel_dp)) - intel_alpm_configure(intel_dp, crtc_state); - /* * Wa_16013835468 * Wa_14015648006 @@ -1930,6 +1958,14 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, intel_de_rmw(display, CLKGATE_DIS_MISC, 0, CLKGATE_DIS_MISC_DMASC_GATING_DIS); } + + /* Wa_16025596647 */ + if ((DISPLAY_VER(display) == 20 || + IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) && + !intel_dp->psr.panel_replay_enabled) + intel_dmc_block_pkgc(display, intel_dp->psr.pipe, true); + + intel_alpm_configure(intel_dp, crtc_state); } static bool psr_interrupt_error_check(struct intel_dp *intel_dp) @@ -1985,6 +2021,7 @@ static void intel_psr_enable_locked(struct intel_dp *intel_dp, intel_dp->psr.psr2_sel_fetch_cff_enabled = false; intel_dp->psr.req_psr2_sdp_prior_scanline = crtc_state->req_psr2_sdp_prior_scanline; + intel_dp->psr.active_non_psr_pipes = crtc_state->active_non_psr_pipes; if (!psr_interrupt_error_check(intel_dp)) return; @@ -1996,8 +2033,9 @@ static void intel_psr_enable_locked(struct intel_dp *intel_dp, intel_dp->psr.sel_update_enabled ? "2" : "1"); /* - * Enabling here only for PSR. Panel Replay enable bit is already - * written at this point. See + * Enabling sink PSR/Panel Replay here only for PSR. Panel Replay enable + * bit is already written at this point. Sink ALPM is enabled here for + * PSR and Panel Replay. See * intel_psr_panel_replay_enable_sink. Modifiers/options: * - Selective Update * - Region Early Transport @@ -2060,6 +2098,12 @@ static void intel_psr_exit(struct intel_dp *intel_dp) drm_WARN_ON(display->drm, !(val & EDP_PSR2_ENABLE)); } else { + if (DISPLAY_VER(display) == 20 || + IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) + intel_dmc_start_pkgc_exit_at_start_of_undelayed_vblank(display, + intel_dp->psr.pipe, + false); + val = intel_de_rmw(display, psr_ctl_reg(display, cpu_transcoder), EDP_PSR_ENABLE, 0); @@ -2133,16 +2177,8 @@ static void intel_psr_disable_locked(struct intel_dp *intel_dp) if (intel_dp_is_edp(intel_dp)) intel_snps_phy_update_psr_power_state(&dp_to_dig_port(intel_dp)->base, false); - /* Panel Replay on eDP is always using ALPM aux less. */ - if (intel_dp->psr.panel_replay_enabled && intel_dp_is_edp(intel_dp)) { - intel_de_rmw(display, ALPM_CTL(display, cpu_transcoder), - ALPM_CTL_ALPM_ENABLE | - ALPM_CTL_ALPM_AUX_LESS_ENABLE, 0); - - intel_de_rmw(display, - PORT_ALPM_CTL(cpu_transcoder), - PORT_ALPM_CTL_ALPM_AUX_LESS_ENABLE, 0); - } + if (intel_dp->psr.panel_replay_enabled && intel_dp_is_edp(intel_dp)) + intel_alpm_disable(intel_dp); /* Disable PSR on Sink */ if (!intel_dp->psr.panel_replay_enabled) { @@ -2153,12 +2189,19 @@ static void intel_psr_disable_locked(struct intel_dp *intel_dp) DP_RECEIVER_ALPM_CONFIG, 0); } + /* Wa_16025596647 */ + if ((DISPLAY_VER(display) == 20 || + IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) && + !intel_dp->psr.panel_replay_enabled) + intel_dmc_block_pkgc(display, intel_dp->psr.pipe, false); + intel_dp->psr.enabled = false; intel_dp->psr.panel_replay_enabled = false; intel_dp->psr.sel_update_enabled = false; intel_dp->psr.psr2_sel_fetch_enabled = false; intel_dp->psr.su_region_et_enabled = false; intel_dp->psr.psr2_sel_fetch_cff_enabled = false; + intel_dp->psr.active_non_psr_pipes = 0; } /** @@ -2254,17 +2297,20 @@ out: } /** - * intel_psr_needs_block_dc_vblank - Check if block dc entry is needed + * intel_psr_needs_vblank_notification - Check if PSR need vblank enable/disable + * notification. * @crtc_state: CRTC status * * We need to block DC6 entry in case of Panel Replay as enabling VBI doesn't * prevent it in case of Panel Replay. Panel Replay switches main link off on * DC entry. This means vblank interrupts are not fired and is a problem if - * user-space is polling for vblank events. + * user-space is polling for vblank events. Also Wa_16025596647 needs + * information when vblank is enabled/disabled. */ -bool intel_psr_needs_block_dc_vblank(const struct intel_crtc_state *crtc_state) +bool intel_psr_needs_vblank_notification(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); + struct intel_display *display = to_intel_display(crtc_state); struct intel_encoder *encoder; for_each_encoder_on_crtc(crtc->base.dev, &crtc->base, encoder) { @@ -2275,8 +2321,15 @@ bool intel_psr_needs_block_dc_vblank(const struct intel_crtc_state *crtc_state) intel_dp = enc_to_intel_dp(encoder); - if (intel_dp_is_edp(intel_dp) && - CAN_PANEL_REPLAY(intel_dp)) + if (!intel_dp_is_edp(intel_dp)) + continue; + + if (CAN_PANEL_REPLAY(intel_dp)) + return true; + + if ((DISPLAY_VER(display) == 20 || + IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) && + CAN_PSR(intel_dp)) return true; } @@ -2304,6 +2357,53 @@ void intel_psr_trigger_frame_change_event(struct intel_dsb *dsb, CURSURFLIVE(display, crtc->pipe), 0); } +/** + * intel_psr_min_vblank_delay - Minimum vblank delay needed by PSR + * @crtc_state: the crtc state + * + * Return minimum vblank delay needed by PSR. + */ +int intel_psr_min_vblank_delay(const struct intel_crtc_state *crtc_state) +{ + struct intel_display *display = to_intel_display(crtc_state); + + if (!crtc_state->has_psr) + return 0; + + /* Wa_14015401596 */ + if (intel_vrr_possible(crtc_state) && IS_DISPLAY_VER(display, 13, 14)) + return 1; + + /* Rest is for SRD_STATUS needed on LunarLake and onwards */ + if (DISPLAY_VER(display) < 20) + return 0; + + /* + * Comment on SRD_STATUS register in Bspec for LunarLake and onwards: + * + * To deterministically capture the transition of the state machine + * going from SRDOFFACK to IDLE, the delayed V. Blank should be at least + * one line after the non-delayed V. Blank. + * + * Legacy TG: TRANS_SET_CONTEXT_LATENCY > 0 + * VRR TG: TRANS_VRR_CTL[ VRR Guardband ] < (TRANS_VRR_VMAX[ VRR Vmax ] + * - TRANS_VTOTAL[ Vertical Active ]) + * + * SRD_STATUS is used only by PSR1 on PantherLake. + * SRD_STATUS is used by PSR1 and Panel Replay DP on LunarLake. + */ + + if (DISPLAY_VER(display) >= 30 && (crtc_state->has_panel_replay || + crtc_state->has_sel_update)) + return 0; + else if (DISPLAY_VER(display) < 30 && (crtc_state->has_sel_update || + intel_crtc_has_type(crtc_state, + INTEL_OUTPUT_EDP))) + return 0; + else + return 1; +} + static u32 man_trk_ctl_enable_bit_get(struct intel_display *display) { return display->platform.alderlake_p || DISPLAY_VER(display) >= 14 ? 0 : @@ -3398,29 +3498,14 @@ static int psr_get_status_and_error_status(struct intel_dp *intel_dp, static void psr_alpm_check(struct intel_dp *intel_dp) { - struct intel_display *display = to_intel_display(intel_dp); - struct drm_dp_aux *aux = &intel_dp->aux; struct intel_psr *psr = &intel_dp->psr; - u8 val; - int r; if (!psr->sel_update_enabled) return; - r = drm_dp_dpcd_readb(aux, DP_RECEIVER_ALPM_STATUS, &val); - if (r != 1) { - drm_err(display->drm, "Error reading ALPM status\n"); - return; - } - - if (val & DP_ALPM_LOCK_TIMEOUT_ERROR) { + if (intel_alpm_get_error(intel_dp)) { intel_psr_disable_locked(intel_dp); psr->sink_not_reliable = true; - drm_dbg_kms(display->drm, - "ALPM lock timeout error, disabling PSR\n"); - - /* Clearing error */ - drm_dp_dpcd_writeb(aux, DP_RECEIVER_ALPM_STATUS, val); } } @@ -3605,6 +3690,168 @@ void intel_psr_unlock(const struct intel_crtc_state *crtc_state) } } +/* Wa_16025596647 */ +static void intel_psr_apply_underrun_on_idle_wa_locked(struct intel_dp *intel_dp) +{ + struct intel_display *display = to_intel_display(intel_dp); + bool dc5_dc6_blocked; + + if (!intel_dp->psr.active) + return; + + dc5_dc6_blocked = is_dc5_dc6_blocked(intel_dp); + + if (intel_dp->psr.sel_update_enabled) + psr2_program_idle_frames(intel_dp, dc5_dc6_blocked ? 0 : + psr_compute_idle_frames(intel_dp)); + else + intel_dmc_start_pkgc_exit_at_start_of_undelayed_vblank(display, + intel_dp->psr.pipe, + dc5_dc6_blocked); +} + +static void psr_dc5_dc6_wa_work(struct work_struct *work) +{ + struct intel_display *display = container_of(work, typeof(*display), + psr_dc5_dc6_wa_work); + struct intel_encoder *encoder; + + for_each_intel_encoder_with_psr(display->drm, encoder) { + struct intel_dp *intel_dp = enc_to_intel_dp(encoder); + + mutex_lock(&intel_dp->psr.lock); + + if (intel_dp->psr.enabled && !intel_dp->psr.panel_replay_enabled) + intel_psr_apply_underrun_on_idle_wa_locked(intel_dp); + + mutex_unlock(&intel_dp->psr.lock); + } +} + +/** + * intel_psr_notify_dc5_dc6 - Notify PSR about enable/disable dc5/dc6 + * @display: intel atomic state + * + * This is targeted for underrun on idle PSR HW bug (Wa_16025596647) to schedule + * psr_dc5_dc6_wa_work used for applying/removing the workaround. + */ +void intel_psr_notify_dc5_dc6(struct intel_display *display) +{ + if (DISPLAY_VER(display) != 20 && + !IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) + return; + + schedule_work(&display->psr_dc5_dc6_wa_work); +} + +/** + * intel_psr_dc5_dc6_wa_init - Init work for underrun on idle PSR HW bug wa + * @display: intel atomic state + * + * This is targeted for underrun on idle PSR HW bug (Wa_16025596647) to init + * psr_dc5_dc6_wa_work used for applying the workaround. + */ +void intel_psr_dc5_dc6_wa_init(struct intel_display *display) +{ + if (DISPLAY_VER(display) != 20 && + !IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) + return; + + INIT_WORK(&display->psr_dc5_dc6_wa_work, psr_dc5_dc6_wa_work); +} + +/** + * intel_psr_notify_pipe_change - Notify PSR about enable/disable of a pipe + * @state: intel atomic state + * @crtc: intel crtc + * @enable: enable/disable + * + * This is targeted for underrun on idle PSR HW bug (Wa_16025596647) to apply + * remove the workaround when pipe is getting enabled/disabled + */ +void intel_psr_notify_pipe_change(struct intel_atomic_state *state, + struct intel_crtc *crtc, bool enable) +{ + struct intel_display *display = to_intel_display(state); + struct intel_encoder *encoder; + + if (DISPLAY_VER(display) != 20 && + !IS_DISPLAY_VERx100_STEP(display, 3000, STEP_A0, STEP_B0)) + return; + + for_each_intel_encoder_with_psr(display->drm, encoder) { + struct intel_dp *intel_dp = enc_to_intel_dp(encoder); + u8 active_non_psr_pipes; + + mutex_lock(&intel_dp->psr.lock); + + if (!intel_dp->psr.enabled || intel_dp->psr.panel_replay_enabled) + goto unlock; + + active_non_psr_pipes = intel_dp->psr.active_non_psr_pipes; + + if (enable) + active_non_psr_pipes |= BIT(crtc->pipe); + else + active_non_psr_pipes &= ~BIT(crtc->pipe); + + if (active_non_psr_pipes == intel_dp->psr.active_non_psr_pipes) + goto unlock; + + if ((enable && intel_dp->psr.active_non_psr_pipes) || + (!enable && !intel_dp->psr.active_non_psr_pipes)) { + intel_dp->psr.active_non_psr_pipes = active_non_psr_pipes; + goto unlock; + } + + intel_dp->psr.active_non_psr_pipes = active_non_psr_pipes; + + intel_psr_apply_underrun_on_idle_wa_locked(intel_dp); +unlock: + mutex_unlock(&intel_dp->psr.lock); + } +} + +/** + * intel_psr_notify_vblank_enable_disable - Notify PSR about enable/disable of vblank + * @display: intel display struct + * @enable: enable/disable + * + * This is targeted for underrun on idle PSR HW bug (Wa_16025596647) to apply + * remove the workaround when vblank is getting enabled/disabled + */ +void intel_psr_notify_vblank_enable_disable(struct intel_display *display, + bool enable) +{ + struct intel_encoder *encoder; + + for_each_intel_encoder_with_psr(display->drm, encoder) { + struct intel_dp *intel_dp = enc_to_intel_dp(encoder); + + mutex_lock(&intel_dp->psr.lock); + if (intel_dp->psr.panel_replay_enabled) { + mutex_unlock(&intel_dp->psr.lock); + break; + } + + if (intel_dp->psr.enabled) + intel_psr_apply_underrun_on_idle_wa_locked(intel_dp); + + mutex_unlock(&intel_dp->psr.lock); + return; + } + + /* + * NOTE: intel_display_power_set_target_dc_state is used + * only by PSR * code for DC3CO handling. DC3CO target + * state is currently disabled in * PSR code. If DC3CO + * is taken into use we need take that into account here + * as well. + */ + intel_display_power_set_target_dc_state(display, enable ? DC_STATE_DISABLE : + DC_STATE_EN_UPTO_DC6); +} + static void psr_source_status(struct intel_dp *intel_dp, struct seq_file *m) { @@ -3977,3 +4224,13 @@ void intel_psr_connector_debugfs_add(struct intel_connector *connector) debugfs_create_file("i915_psr_status", 0444, root, connector, &i915_psr_status_fops); } + +bool intel_psr_needs_alpm(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) +{ + /* + * eDP Panel Replay uses always ALPM + * PSR2 uses ALPM but PSR1 doesn't + */ + return intel_dp_is_edp(intel_dp) && (crtc_state->has_sel_update || + crtc_state->has_panel_replay); +} diff --git a/drivers/gpu/drm/i915/display/intel_psr.h b/drivers/gpu/drm/i915/display/intel_psr.h index a43a374cff55..73c3fa40844b 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.h +++ b/drivers/gpu/drm/i915/display/intel_psr.h @@ -59,7 +59,13 @@ void intel_psr2_program_trans_man_trk_ctl(struct intel_dsb *dsb, const struct intel_crtc_state *crtc_state); void intel_psr_pause(struct intel_dp *intel_dp); void intel_psr_resume(struct intel_dp *intel_dp); -bool intel_psr_needs_block_dc_vblank(const struct intel_crtc_state *crtc_state); +bool intel_psr_needs_vblank_notification(const struct intel_crtc_state *crtc_state); +void intel_psr_notify_pipe_change(struct intel_atomic_state *state, + struct intel_crtc *crtc, bool enable); +void intel_psr_notify_dc5_dc6(struct intel_display *display); +void intel_psr_dc5_dc6_wa_init(struct intel_display *display); +void intel_psr_notify_vblank_enable_disable(struct intel_display *display, + bool enable); bool intel_psr_link_ok(struct intel_dp *intel_dp); void intel_psr_lock(const struct intel_crtc_state *crtc_state); @@ -67,7 +73,9 @@ void intel_psr_unlock(const struct intel_crtc_state *crtc_state); void intel_psr_trigger_frame_change_event(struct intel_dsb *dsb, struct intel_atomic_state *state, struct intel_crtc *crtc); +int intel_psr_min_vblank_delay(const struct intel_crtc_state *crtc_state); void intel_psr_connector_debugfs_add(struct intel_connector *connector); void intel_psr_debugfs_register(struct intel_display *display); +bool intel_psr_needs_alpm(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state); #endif /* __INTEL_PSR_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_sdvo.c b/drivers/gpu/drm/i915/display/intel_sdvo.c index 757b9ce7e3b1..8a38df2c0283 100644 --- a/drivers/gpu/drm/i915/display/intel_sdvo.c +++ b/drivers/gpu/drm/i915/display/intel_sdvo.c @@ -36,9 +36,9 @@ #include <drm/drm_crtc.h> #include <drm/drm_edid.h> #include <drm/drm_eld.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> -#include "i915_drv.h" #include "i915_reg.h" #include "intel_atomic.h" #include "intel_audio.h" @@ -214,18 +214,17 @@ intel_sdvo_create_enhance_property(struct intel_sdvo *intel_sdvo, static void intel_sdvo_write_sdvox(struct intel_sdvo *intel_sdvo, u32 val) { struct intel_display *display = to_intel_display(&intel_sdvo->base); - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 bval = val, cval = val; int i; - if (HAS_PCH_SPLIT(dev_priv)) { + if (HAS_PCH_SPLIT(display)) { intel_de_write(display, intel_sdvo->sdvo_reg, val); intel_de_posting_read(display, intel_sdvo->sdvo_reg); /* * HW workaround, need to write this twice for issue * that may result in first write getting masked. */ - if (HAS_PCH_IBX(dev_priv)) { + if (HAS_PCH_IBX(display)) { intel_de_write(display, intel_sdvo->sdvo_reg, val); intel_de_posting_read(display, intel_sdvo->sdvo_reg); } @@ -1360,14 +1359,13 @@ static int intel_sdvo_compute_config(struct intel_encoder *encoder, struct drm_connector_state *conn_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_sdvo *intel_sdvo = to_sdvo(encoder); struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(conn_state->connector); struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode; struct drm_display_mode *mode = &pipe_config->hw.mode; - if (HAS_PCH_SPLIT(i915)) { + if (HAS_PCH_SPLIT(display)) { pipe_config->has_pch_encoder = true; if (!intel_fdi_compute_pipe_bpp(pipe_config)) return -EINVAL; @@ -1527,7 +1525,6 @@ static void intel_sdvo_pre_enable(struct intel_atomic_state *state, const struct drm_connector_state *conn_state) { struct intel_display *display = to_intel_display(intel_encoder); - struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; const struct intel_sdvo_connector_state *sdvo_state = @@ -1634,7 +1631,7 @@ static void intel_sdvo_pre_enable(struct intel_atomic_state *state, sdvox |= (9 << 19) | SDVO_BORDER_ENABLE; } - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) sdvox |= SDVO_PIPE_SEL_CPT(crtc->pipe); else sdvox |= SDVO_PIPE_SEL(crtc->pipe); @@ -1670,13 +1667,12 @@ static bool intel_sdvo_connector_get_hw_state(struct intel_connector *connector) bool intel_sdvo_port_enabled(struct intel_display *display, i915_reg_t sdvo_reg, enum pipe *pipe) { - struct drm_i915_private *dev_priv = to_i915(display->drm); u32 val; val = intel_de_read(display, sdvo_reg); /* asserts want to know the pipe even if the port is disabled */ - if (HAS_PCH_CPT(dev_priv)) + if (HAS_PCH_CPT(display)) *pipe = (val & SDVO_PIPE_SEL_MASK_CPT) >> SDVO_PIPE_SEL_SHIFT_CPT; else if (display->platform.cherryview) *pipe = (val & SDVO_PIPE_SEL_MASK_CHV) >> SDVO_PIPE_SEL_SHIFT_CHV; @@ -1841,7 +1837,6 @@ static void intel_disable_sdvo(struct intel_atomic_state *state, const struct drm_connector_state *conn_state) { struct intel_display *display = to_intel_display(encoder); - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_sdvo *intel_sdvo = to_sdvo(encoder); struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc); u32 temp; @@ -1861,7 +1856,7 @@ static void intel_disable_sdvo(struct intel_atomic_state *state, * to transcoder A after disabling it to allow the * matching DP port to be enabled on transcoder A. */ - if (HAS_PCH_IBX(dev_priv) && crtc->pipe == PIPE_B) { + if (HAS_PCH_IBX(display) && crtc->pipe == PIPE_B) { /* * We get CPU/PCH FIFO underruns on the other pipe when * doing the workaround. Sweep them under the rug. @@ -3367,9 +3362,7 @@ intel_sdvo_init_ddc_proxy(struct intel_sdvo_ddc *ddc, static bool is_sdvo_port_valid(struct intel_display *display, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); - - if (HAS_PCH_SPLIT(dev_priv)) + if (HAS_PCH_SPLIT(display)) return port == PORT_B; else return port == PORT_B || port == PORT_C; @@ -3384,7 +3377,6 @@ static bool assert_sdvo_port_valid(struct intel_display *display, enum port port bool intel_sdvo_init(struct intel_display *display, i915_reg_t sdvo_reg, enum port port) { - struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *intel_encoder; struct intel_sdvo *intel_sdvo; int i; @@ -3427,7 +3419,7 @@ bool intel_sdvo_init(struct intel_display *display, } intel_encoder->compute_config = intel_sdvo_compute_config; - if (HAS_PCH_SPLIT(dev_priv)) { + if (HAS_PCH_SPLIT(display)) { intel_encoder->disable = pch_disable_sdvo; intel_encoder->post_disable = pch_post_disable_sdvo; } else { diff --git a/drivers/gpu/drm/i915/display/intel_sprite.c b/drivers/gpu/drm/i915/display/intel_sprite.c index 1ad6c8a94b3d..fd92e6b89b43 100644 --- a/drivers/gpu/drm/i915/display/intel_sprite.c +++ b/drivers/gpu/drm/i915/display/intel_sprite.c @@ -36,9 +36,10 @@ #include <drm/drm_blend.h> #include <drm/drm_color_mgmt.h> #include <drm/drm_fourcc.h> +#include <drm/drm_print.h> #include <drm/drm_rect.h> -#include "i915_drv.h" +#include "i915_utils.h" #include "i9xx_plane.h" #include "intel_atomic_plane.h" #include "intel_de.h" diff --git a/drivers/gpu/drm/i915/display/intel_sprite_uapi.c b/drivers/gpu/drm/i915/display/intel_sprite_uapi.c index 1d0b84b464c1..4981cc34da05 100644 --- a/drivers/gpu/drm/i915/display/intel_sprite_uapi.c +++ b/drivers/gpu/drm/i915/display/intel_sprite_uapi.c @@ -3,21 +3,21 @@ * Copyright © 2023 Intel Corporation */ -#include "i915_drv.h" #include "intel_crtc.h" +#include "intel_display_core.h" #include "intel_display_types.h" #include "intel_sprite_uapi.h" -static bool has_dst_key_in_primary_plane(struct drm_i915_private *dev_priv) +static bool has_dst_key_in_primary_plane(struct intel_display *display) { - return DISPLAY_VER(dev_priv) >= 9; + return DISPLAY_VER(display) >= 9; } static void intel_plane_set_ckey(struct intel_plane_state *plane_state, const struct drm_intel_sprite_colorkey *set) { + struct intel_display *display = to_intel_display(plane_state); struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane); - struct drm_i915_private *dev_priv = to_i915(plane->base.dev); struct drm_intel_sprite_colorkey *key = &plane_state->ckey; *key = *set; @@ -34,7 +34,7 @@ static void intel_plane_set_ckey(struct intel_plane_state *plane_state, * On SKL+ we want dst key enabled on * the primary and not on the sprite. */ - if (DISPLAY_VER(dev_priv) >= 9 && plane->id != PLANE_PRIMARY && + if (DISPLAY_VER(display) >= 9 && plane->id != PLANE_PRIMARY && set->flags & I915_SET_COLORKEY_DESTINATION) key->flags = 0; } @@ -43,7 +43,6 @@ int intel_sprite_set_colorkey_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { struct intel_display *display = to_intel_display(dev); - struct drm_i915_private *dev_priv = to_i915(dev); struct drm_intel_sprite_colorkey *set = data; struct drm_plane *plane; struct drm_plane_state *plane_state; @@ -61,7 +60,7 @@ int intel_sprite_set_colorkey_ioctl(struct drm_device *dev, void *data, if ((set->flags & (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) == (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) return -EINVAL; - if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) && + if ((display->platform.valleyview || display->platform.cherryview) && set->flags & I915_SET_COLORKEY_DESTINATION) return -EINVAL; @@ -74,7 +73,7 @@ int intel_sprite_set_colorkey_ioctl(struct drm_device *dev, void *data, * Also multiple planes can't do destination keying on the same * pipe simultaneously. */ - if (DISPLAY_VER(dev_priv) >= 9 && + if (DISPLAY_VER(display) >= 9 && to_intel_plane(plane)->id >= PLANE_3 && set->flags & I915_SET_COLORKEY_DESTINATION) return -EINVAL; @@ -99,7 +98,7 @@ int intel_sprite_set_colorkey_ioctl(struct drm_device *dev, void *data, * On some platforms we have to configure * the dst colorkey on the primary plane. */ - if (!ret && has_dst_key_in_primary_plane(dev_priv)) { + if (!ret && has_dst_key_in_primary_plane(display)) { struct intel_crtc *crtc = intel_crtc_for_pipe(display, to_intel_plane(plane)->pipe); diff --git a/drivers/gpu/drm/i915/display/intel_tv.c b/drivers/gpu/drm/i915/display/intel_tv.c index 2e3f3f0207e8..acf0b3733908 100644 --- a/drivers/gpu/drm/i915/display/intel_tv.c +++ b/drivers/gpu/drm/i915/display/intel_tv.c @@ -33,15 +33,15 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_crtc.h> #include <drm/drm_edid.h> +#include <drm/drm_print.h> #include <drm/drm_probe_helper.h> -#include "i915_drv.h" #include "i915_reg.h" #include "intel_connector.h" #include "intel_crtc.h" #include "intel_de.h" -#include "intel_display_irq.h" #include "intel_display_driver.h" +#include "intel_display_irq.h" #include "intel_display_types.h" #include "intel_dpll.h" #include "intel_hotplug.h" @@ -1585,19 +1585,17 @@ intel_tv_detect_type(struct intel_tv *intel_tv, { struct intel_display *display = to_intel_display(connector->dev); struct intel_crtc *crtc = to_intel_crtc(connector->state->crtc); - struct drm_device *dev = connector->dev; - struct drm_i915_private *dev_priv = to_i915(dev); u32 tv_ctl, save_tv_ctl; u32 tv_dac, save_tv_dac; int type; /* Disable TV interrupts around load detect or we'll recurse */ if (connector->polled & DRM_CONNECTOR_POLL_HPD) { - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); i915_disable_pipestat(display, 0, PIPE_HOTPLUG_INTERRUPT_STATUS | PIPE_HOTPLUG_TV_INTERRUPT_STATUS); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } save_tv_dac = tv_dac = intel_de_read(display, TV_DAC); @@ -1668,11 +1666,11 @@ intel_tv_detect_type(struct intel_tv *intel_tv, /* Restore interrupt config */ if (connector->polled & DRM_CONNECTOR_POLL_HPD) { - spin_lock_irq(&dev_priv->irq_lock); + spin_lock_irq(&display->irq.lock); i915_enable_pipestat(display, 0, PIPE_HOTPLUG_INTERRUPT_STATUS | PIPE_HOTPLUG_TV_INTERRUPT_STATUS); - spin_unlock_irq(&dev_priv->irq_lock); + spin_unlock_irq(&display->irq.lock); } return type; diff --git a/drivers/gpu/drm/i915/display/intel_vga.c b/drivers/gpu/drm/i915/display/intel_vga.c index 684b5d1bc87c..05d140c8032d 100644 --- a/drivers/gpu/drm/i915/display/intel_vga.c +++ b/drivers/gpu/drm/i915/display/intel_vga.c @@ -4,15 +4,20 @@ */ #include <linux/delay.h> +#include <linux/pci.h> #include <linux/vgaarb.h> +#include <drm/drm_device.h> +#include <drm/drm_print.h> #include <video/vga.h> + #include "soc/intel_gmch.h" -#include "i915_drv.h" #include "i915_reg.h" #include "intel_de.h" +#include "intel_display.h" #include "intel_vga.h" +#include "intel_vga_regs.h" static i915_reg_t intel_vga_cntrl_reg(struct intel_display *display) { @@ -24,16 +29,42 @@ static i915_reg_t intel_vga_cntrl_reg(struct intel_display *display) return VGACNTRL; } +static bool has_vga_pipe_sel(struct intel_display *display) +{ + if (display->platform.i845g || + display->platform.i865g) + return false; + + if (display->platform.valleyview || + display->platform.cherryview) + return true; + + return DISPLAY_VER(display) < 7; +} + /* Disable the VGA plane that we never use */ void intel_vga_disable(struct intel_display *display) { struct pci_dev *pdev = to_pci_dev(display->drm->dev); i915_reg_t vga_reg = intel_vga_cntrl_reg(display); + enum pipe pipe; + u32 tmp; u8 sr1; - if (intel_de_read(display, vga_reg) & VGA_DISP_DISABLE) + tmp = intel_de_read(display, vga_reg); + if (tmp & VGA_DISP_DISABLE) return; + if (display->platform.cherryview) + pipe = REG_FIELD_GET(VGA_PIPE_SEL_MASK_CHV, tmp); + else if (has_vga_pipe_sel(display)) + pipe = REG_FIELD_GET(VGA_PIPE_SEL_MASK, tmp); + else + pipe = PIPE_A; + + drm_dbg_kms(display->drm, "Disabling VGA plane on pipe %c\n", + pipe_name(pipe)); + /* WaEnableVGAAccessThroughIOPort:ctg,elk,ilk,snb,ivb,vlv,hsw */ vga_get_uninterruptible(pdev, VGA_RSRC_LEGACY_IO); outb(0x01, VGA_SEQ_I); @@ -46,39 +77,6 @@ void intel_vga_disable(struct intel_display *display) intel_de_posting_read(display, vga_reg); } -void intel_vga_redisable_power_on(struct intel_display *display) -{ - i915_reg_t vga_reg = intel_vga_cntrl_reg(display); - - if (!(intel_de_read(display, vga_reg) & VGA_DISP_DISABLE)) { - drm_dbg_kms(display->drm, - "Something enabled VGA plane, disabling it\n"); - intel_vga_disable(display); - } -} - -void intel_vga_redisable(struct intel_display *display) -{ - intel_wakeref_t wakeref; - - /* - * This function can be called both from intel_modeset_setup_hw_state or - * at a very early point in our resume sequence, where the power well - * structures are not yet restored. Since this function is at a very - * paranoid "someone might have enabled VGA while we were not looking" - * level, just check if the power well is enabled instead of trying to - * follow the "don't touch the power well if we don't need it" policy - * the rest of the driver uses. - */ - wakeref = intel_display_power_get_if_enabled(display, POWER_DOMAIN_VGA); - if (!wakeref) - return; - - intel_vga_redisable_power_on(display); - - intel_display_power_put(display, POWER_DOMAIN_VGA, wakeref); -} - void intel_vga_reset_io_mem(struct intel_display *display) { struct pci_dev *pdev = to_pci_dev(display->drm->dev); diff --git a/drivers/gpu/drm/i915/display/intel_vga.h b/drivers/gpu/drm/i915/display/intel_vga.h index 824dfc32a199..16d699f3b641 100644 --- a/drivers/gpu/drm/i915/display/intel_vga.h +++ b/drivers/gpu/drm/i915/display/intel_vga.h @@ -10,8 +10,6 @@ struct intel_display; void intel_vga_reset_io_mem(struct intel_display *display); void intel_vga_disable(struct intel_display *display); -void intel_vga_redisable(struct intel_display *display); -void intel_vga_redisable_power_on(struct intel_display *display); int intel_vga_register(struct intel_display *display); void intel_vga_unregister(struct intel_display *display); diff --git a/drivers/gpu/drm/i915/display/intel_vga_regs.h b/drivers/gpu/drm/i915/display/intel_vga_regs.h new file mode 100644 index 000000000000..cbacced1a69f --- /dev/null +++ b/drivers/gpu/drm/i915/display/intel_vga_regs.h @@ -0,0 +1,36 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright © 2025 Intel Corporation + */ + +#ifndef __INTEL_VGA_REGS_H__ +#define __INTEL_VGA_REGS_H__ + +#include "intel_display_reg_defs.h" + +#define VGACNTRL _MMIO(0x71400) +#define VLV_VGACNTRL _MMIO(VLV_DISPLAY_BASE + 0x71400) +#define CPU_VGACNTRL _MMIO(0x41000) +#define VGA_DISP_DISABLE REG_BIT(31) +#define VGA_2X_MODE REG_BIT(30) /* pre-ilk */ +#define VGA_PIPE_SEL_MASK REG_BIT(29) /* pre-ivb */ +#define VGA_PIPE_SEL(pipe) REG_FIELD_PREP(VGA_PIPE_SEL_MASK, (pipe)) +#define VGA_PIPE_SEL_MASK_CHV REG_GENMASK(29, 28) /* chv */ +#define VGA_PIPE_SEL_CHV(pipe) REG_FIELD_PREP(VGA_PIPE_SEL_MASK_CHV, (pipe)) +#define VGA_BORDER_ENABLE REG_BIT(26) +#define VGA_PIPE_CSC_ENABLE REG_BIT(24) /* ilk+ */ +#define VGA_CENTERING_ENABLE_MASK REG_GENMASK(25, 24) /* pre-ilk */ +#define VGA_PALETTE_READ_SEL REG_BIT(23) /* pre-ivb */ +#define VGA_PALETTE_A_WRITE_DISABLE REG_BIT(22) /* pre-ivb */ +#define VGA_PALETTE_B_WRITE_DISABLE REG_BIT(21) /* pre-ivb */ +#define VGA_LEGACY_8BIT_PALETTE_ENABLE REG_BIT(20) +#define VGA_PALETTE_BYPASS REG_BIT(19) +#define VGA_NINE_DOT_DISABLE REG_BIT(18) +#define VGA_PALETTE_READ_SEL_HI_CHV REG_BIT(15) /* chv */ +#define VGA_PALETTE_C_WRITE_DISABLE_CHV REG_BIT(14) /* chv */ +#define VGA_ACTIVE_THROTTLING_MASK REG_GENMASK(15, 12) /* ilk+ */ +#define VGA_BLANK_THROTTLING_MASK REG_GENMASK(11, 8) /* ilk+ */ +#define VGA_BLINK_DUTY_CYCLE_MASK REG_GENMASK(7, 6) +#define VGA_VSYNC_BLINK_RATE_MASK REG_GENMASK(5, 0) + +#endif /* __INTEL_VGA_REGS_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_vrr.c b/drivers/gpu/drm/i915/display/intel_vrr.c index 633a66f6b73b..c6565baf815a 100644 --- a/drivers/gpu/drm/i915/display/intel_vrr.c +++ b/drivers/gpu/drm/i915/display/intel_vrr.c @@ -444,7 +444,10 @@ static u32 trans_vrr_ctl(const struct intel_crtc_state *crtc_state) { struct intel_display *display = to_intel_display(crtc_state); - if (DISPLAY_VER(display) >= 13) + if (DISPLAY_VER(display) >= 14) + return VRR_CTL_FLIP_LINE_EN | + XELPD_VRR_CTL_VRR_GUARDBAND(crtc_state->vrr.guardband); + else if (DISPLAY_VER(display) >= 13) return VRR_CTL_IGN_MAX_SHIFT | VRR_CTL_FLIP_LINE_EN | XELPD_VRR_CTL_VRR_GUARDBAND(crtc_state->vrr.guardband); else diff --git a/drivers/gpu/drm/i915/display/skl_scaler.c b/drivers/gpu/drm/i915/display/skl_scaler.c index ee81220a7c88..c855426544cf 100644 --- a/drivers/gpu/drm/i915/display/skl_scaler.c +++ b/drivers/gpu/drm/i915/display/skl_scaler.c @@ -3,8 +3,10 @@ * Copyright © 2020 Intel Corporation */ -#include "i915_drv.h" +#include <drm/drm_print.h> + #include "i915_reg.h" +#include "i915_utils.h" #include "intel_de.h" #include "intel_display_trace.h" #include "intel_display_types.h" diff --git a/drivers/gpu/drm/i915/display/skl_universal_plane.c b/drivers/gpu/drm/i915/display/skl_universal_plane.c index f7c3ee1f8664..c7b336359a5e 100644 --- a/drivers/gpu/drm/i915/display/skl_universal_plane.c +++ b/drivers/gpu/drm/i915/display/skl_universal_plane.c @@ -2693,24 +2693,22 @@ static void skl_plane_enable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); bdw_enable_pipe_irq(display, pipe, GEN9_PIPE_PLANE_FLIP_DONE(plane->id)); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static void skl_plane_disable_flip_done(struct intel_plane *plane) { struct intel_display *display = to_intel_display(plane); - struct drm_i915_private *i915 = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - spin_lock_irq(&i915->irq_lock); + spin_lock_irq(&display->irq.lock); bdw_disable_pipe_irq(display, pipe, GEN9_PIPE_PLANE_FLIP_DONE(plane->id)); - spin_unlock_irq(&i915->irq_lock); + spin_unlock_irq(&display->irq.lock); } static bool skl_plane_has_rc_ccs(struct intel_display *display, diff --git a/drivers/gpu/drm/i915/gem/i915_gem_busy.c b/drivers/gpu/drm/i915/gem/i915_gem_busy.c index ddda468241ef..6e4d0ce3952f 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_busy.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_busy.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_clflush.c b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c index 7d97ea2a653e..c4854c5b4e0f 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_clflush.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_clflush.h b/drivers/gpu/drm/i915/gem/i915_gem_clflush.h index e6c382973129..9d7ee1579900 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_clflush.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_clflush.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c index ab1af978911b..15835952352e 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2011-2012 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.h b/drivers/gpu/drm/i915/gem/i915_gem_context.h index e5b0f66ea1fe..6e682a6a0574 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_context.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h index 67ac2586a0f3..0267c924634b 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2019 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c index 9473050ac842..05e440643aa2 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright 2012 Red Hat Inc */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_domain.c b/drivers/gpu/drm/i915/gem/i915_gem_domain.c index 75a143d996e0..7a0cc51923b3 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_domain.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 7796c4119ef5..ca7e9216934a 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2008,2010 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_internal.c b/drivers/gpu/drm/i915/gem/i915_gem_internal.c index ea7561ae6e13..232b984f60b6 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_internal.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_internal.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h b/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h index 28d6526e32ab..8044d34707b6 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2019 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c index c3dabb857960..f6d37dff320d 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.h b/drivers/gpu/drm/i915/gem/i915_gem_mman.h index 196417fd0f5c..946fb9825eb3 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2019 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c index 356530b599ce..1f38e367c60b 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: MIT /* * Copyright © 2017 Intel Corporation * diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.h b/drivers/gpu/drm/i915/gem/i915_gem_object.h index a5f34542135c..c34f41605b46 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h index 68413c05c812..64600aa8227f 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c b/drivers/gpu/drm/i915/gem/i915_gem_pages.c index 8780aa243105..7f83f8bdc8fb 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_pages.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/i915_gem_phys.c index ef85c6dc9fd5..f9e7cab140f8 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pm.c b/drivers/gpu/drm/i915/gem/i915_gem_pm.c index 900c08337942..f0857c5c96df 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_pm.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_pm.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2019 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pm.h b/drivers/gpu/drm/i915/gem/i915_gem_pm.h index bedf1e95941a..bd5bd2c5e7f9 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_pm.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_pm.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2019 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c index ae3343c81a64..3b81cc3ad5be 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c index aec41f0f098f..b81e67504bbe 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2008-2015 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c index 9d958a6f377e..3380151edfc1 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2008-2012 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_throttle.c b/drivers/gpu/drm/i915/gem/i915_gem_throttle.c index af85d0c28168..8814cbcde5b5 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_throttle.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_throttle.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2014-2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_tiling.c b/drivers/gpu/drm/i915/gem/i915_gem_tiling.c index 5ac23ff3feff..5a296ba3758a 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_tiling.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_tiling.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2008 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c index 09b68713ab32..307a18eede72 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c @@ -1,9 +1,8 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2012-2014 Intel Corporation * - * Based on amdgpu_mn, which bears the following notice: + * Based on amdgpu_mn, which bears the following notice: * * Copyright 2014 Advanced Micro Devices, Inc. * All Rights Reserved. diff --git a/drivers/gpu/drm/i915/gem/i915_gem_wait.c b/drivers/gpu/drm/i915/gem/i915_gem_wait.c index 1f55e62044a4..7127e90c1a8f 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_wait.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_wait.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2016 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gemfs.c b/drivers/gpu/drm/i915/gem/i915_gemfs.c index 46b9a17d6abc..65d84a93c525 100644 --- a/drivers/gpu/drm/i915/gem/i915_gemfs.c +++ b/drivers/gpu/drm/i915/gem/i915_gemfs.c @@ -1,6 +1,5 @@ +// SPDX-License-Identifier: MIT /* - * SPDX-License-Identifier: MIT - * * Copyright © 2017 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/i915_gemfs.h b/drivers/gpu/drm/i915/gem/i915_gemfs.h index 5d835e44c4f6..16d4333c9a4e 100644 --- a/drivers/gpu/drm/i915/gem/i915_gemfs.h +++ b/drivers/gpu/drm/i915/gem/i915_gemfs.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2017 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c index 804f74084bd4..9c3f17e51885 100644 --- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c +++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c @@ -1837,6 +1837,8 @@ static int igt_mmap_revoke(void *arg) int i915_gem_mman_live_selftests(struct drm_i915_private *i915) { + int ret; + bool unuse_mm = false; static const struct i915_subtest tests[] = { SUBTEST(igt_partial_tiling), SUBTEST(igt_smoke_tiling), @@ -1848,5 +1850,15 @@ int i915_gem_mman_live_selftests(struct drm_i915_private *i915) SUBTEST(igt_mmap_gpu), }; - return i915_live_subtests(tests, i915); + if (!current->mm) { + kthread_use_mm(current->active_mm); + unuse_mm = true; + } + + ret = i915_live_subtests(tests, i915); + + if (unuse_mm) + kthread_unuse_mm(current->active_mm); + + return ret; } diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.c b/drivers/gpu/drm/i915/gt/intel_gtt.c index 30b128b1fde7..afbc5c769308 100644 --- a/drivers/gpu/drm/i915/gt/intel_gtt.c +++ b/drivers/gpu/drm/i915/gt/intel_gtt.c @@ -176,7 +176,6 @@ static void clear_vm_list(struct list_head *list) i915_vma_destroy_locked(vma); i915_gem_object_put(obj); } - } } diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index 51847a846002..c481b56fa67d 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -751,7 +751,6 @@ static int lrc_ring_indirect_offset(const struct intel_engine_cs *engine) static int lrc_ring_cmd_buf_cctl(const struct intel_engine_cs *engine) { - if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 55)) /* * Note that the CSFE context has a dummy slot for CMD_BUF_CCTL diff --git a/drivers/gpu/drm/i915/gt/intel_mocs.c b/drivers/gpu/drm/i915/gt/intel_mocs.c index cf41d325712e..5dd8121f4b15 100644 --- a/drivers/gpu/drm/i915/gt/intel_mocs.c +++ b/drivers/gpu/drm/i915/gt/intel_mocs.c @@ -314,7 +314,6 @@ static const struct drm_i915_mocs_entry icl_mocs_table[] = { }; static const struct drm_i915_mocs_entry dg1_mocs_table[] = { - /* UC */ MOCS_ENTRY(1, 0, L3_1_UC), /* WB - L3 */ diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c index 6e9977b2d180..a876a34455f1 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c +++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c @@ -365,7 +365,13 @@ static void reset_prepare(struct intel_engine_cs *engine) ENGINE_READ_FW(engine, RING_HEAD), ENGINE_READ_FW(engine, RING_TAIL), ENGINE_READ_FW(engine, RING_START)); - if (!stop_ring(engine)) { + /* + * Sometimes engine head failed to set to zero even after writing into it. + * Use wait_for_atomic() with 20ms delay to let engine resumes from + * correct RING_HEAD. Experimented different values and determined + * that 20ms works best based on testing. + */ + if (wait_for_atomic((!stop_ring(engine) == 0), 20)) { drm_err(&engine->i915->drm, "failed to set %s head to zero " "ctl %08x head %08x tail %08x start %08x\n", diff --git a/drivers/gpu/drm/i915/gt/intel_rps.c b/drivers/gpu/drm/i915/gt/intel_rps.c index 8731f275fdd9..5abc5fcc2514 100644 --- a/drivers/gpu/drm/i915/gt/intel_rps.c +++ b/drivers/gpu/drm/i915/gt/intel_rps.c @@ -8,7 +8,7 @@ #include <drm/intel/i915_drm.h> #include "display/intel_display.h" -#include "display/intel_display_irq.h" +#include "display/intel_display_rps.h" #include "i915_drv.h" #include "i915_irq.h" #include "i915_reg.h" @@ -608,9 +608,7 @@ static bool gen5_rps_enable(struct intel_rps *rps) rps->ips.last_count2 = intel_uncore_read(uncore, GFXEC); rps->ips.last_time2 = ktime_get_raw_ns(); - spin_lock(&i915->irq_lock); - ilk_enable_display_irq(display, DE_PCU_EVENT); - spin_unlock(&i915->irq_lock); + ilk_display_rps_enable(display); spin_unlock_irq(&mchdev_lock); @@ -628,9 +626,7 @@ static void gen5_rps_disable(struct intel_rps *rps) spin_lock_irq(&mchdev_lock); - spin_lock(&i915->irq_lock); - ilk_disable_display_irq(display, DE_PCU_EVENT); - spin_unlock(&i915->irq_lock); + ilk_display_rps_disable(display); rgvswctl = intel_uncore_read16(uncore, MEMSWCTL); @@ -1003,6 +999,10 @@ void intel_rps_dec_waiters(struct intel_rps *rps) if (rps_uses_slpc(rps)) { slpc = rps_to_slpc(rps); + /* Don't decrement num_waiters for req where increment was skipped */ + if (slpc->power_profile == SLPC_POWER_PROFILES_POWER_SAVING) + return; + intel_guc_slpc_dec_waiters(slpc); } else { atomic_dec(&rps->num_waiters); @@ -1031,11 +1031,15 @@ void intel_rps_boost(struct i915_request *rq) if (slpc->power_profile == SLPC_POWER_PROFILES_POWER_SAVING) return; - if (slpc->min_freq_softlimit >= slpc->boost_freq) - return; - /* Return if old value is non zero */ if (!atomic_fetch_inc(&slpc->num_waiters)) { + /* + * Skip queuing boost work if frequency is already boosted, + * but still increment num_waiters. + */ + if (slpc->min_freq_softlimit >= slpc->boost_freq) + return; + GT_TRACE(rps_to_gt(rps), "boost fence:%llx:%llx\n", rq->fence.context, rq->fence.seqno); queue_work(rps_to_gt(rps)->i915->unordered_wq, diff --git a/drivers/gpu/drm/i915/gt/intel_rps_types.h b/drivers/gpu/drm/i915/gt/intel_rps_types.h index 5135b90a2a40..ece445109305 100644 --- a/drivers/gpu/drm/i915/gt/intel_rps_types.h +++ b/drivers/gpu/drm/i915/gt/intel_rps_types.h @@ -57,7 +57,7 @@ struct intel_rps { /* * work, interrupts_enabled and pm_iir are protected by - * i915->irq_lock + * gt->irq_lock */ struct timer_list timer; struct work_struct work; diff --git a/drivers/gpu/drm/i915/gt/intel_wopcm.h b/drivers/gpu/drm/i915/gt/intel_wopcm.h index 17d6aa86008a..d2038b6de5e7 100644 --- a/drivers/gpu/drm/i915/gt/intel_wopcm.h +++ b/drivers/gpu/drm/i915/gt/intel_wopcm.h @@ -1,6 +1,5 @@ +/* SPDX-License-Identifier: MIT */ /* - * SPDX-License-Identifier: MIT - * * Copyright © 2017-2018 Intel Corporation */ diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 116683ebe074..b37e400f74e5 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -156,7 +156,7 @@ static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) if (IS_ALIGNED(wal->count, grow)) { /* Either uninitialized or full. */ struct i915_wa *list; - list = kmalloc_array(ALIGN(wal->count + 1, grow), sizeof(*wa), + list = kmalloc_array(ALIGN(wal->count + 1, grow), sizeof(*list), GFP_KERNEL); if (!list) { drm_err(&i915->drm, "No space for workaround init!\n"); diff --git a/drivers/gpu/drm/i915/gt/selftest_lrc.c b/drivers/gpu/drm/i915/gt/selftest_lrc.c index 22e750108c5f..23f04f6f8fba 100644 --- a/drivers/gpu/drm/i915/gt/selftest_lrc.c +++ b/drivers/gpu/drm/i915/gt/selftest_lrc.c @@ -7,6 +7,7 @@ #include "gem/i915_gem_internal.h" +#include "i915_drv.h" #include "i915_selftest.h" #include "intel_engine_heartbeat.h" #include "intel_engine_pm.h" @@ -859,6 +860,14 @@ static int live_lrc_timestamp(void *arg) }; /* + * This test was designed to isolate a hardware bug. + * The bug was found and fixed in future generations but + * now the test pollutes our CI on previous generation. + */ + if (GRAPHICS_VER(gt->i915) == 12) + return 0; + + /* * We want to verify that the timestamp is saved and restore across * context switches and is monotonic. * diff --git a/drivers/gpu/drm/i915/gt/selftest_rc6.c b/drivers/gpu/drm/i915/gt/selftest_rc6.c index 908483ab0bc8..41716ed454b7 100644 --- a/drivers/gpu/drm/i915/gt/selftest_rc6.c +++ b/drivers/gpu/drm/i915/gt/selftest_rc6.c @@ -33,15 +33,22 @@ int live_rc6_manual(void *arg) { struct intel_gt *gt = arg; struct intel_rc6 *rc6 = >->rc6; - u64 rc0_power, rc6_power; + struct intel_rps *rps = >->rps; intel_wakeref_t wakeref; + u64 rc0_sample_energy[2]; + u64 rc6_sample_energy[2]; + u64 sleep_time = 1000; + u32 rc0_freq = 0; + u32 rc6_freq = 0; + u64 rc0_power; + u64 rc6_power; bool has_power; + u64 threshold; ktime_t dt; u64 res[2]; int err = 0; - u32 rc0_freq = 0; - u32 rc6_freq = 0; - struct intel_rps *rps = >->rps; + u64 diff; + /* * Our claim is that we can "encourage" the GPU to enter rc6 at will. @@ -60,14 +67,15 @@ int live_rc6_manual(void *arg) /* Force RC6 off for starters */ __intel_rc6_disable(rc6); - msleep(1); /* wakeup is not immediate, takes about 100us on icl */ + /* wakeup is not immediate, takes about 100us on icl */ + usleep_range(1000, 2000); res[0] = rc6_residency(rc6); dt = ktime_get(); - rc0_power = librapl_energy_uJ(); - msleep(1000); - rc0_power = librapl_energy_uJ() - rc0_power; + rc0_sample_energy[0] = librapl_energy_uJ(); + msleep(sleep_time); + rc0_sample_energy[1] = librapl_energy_uJ() - rc0_sample_energy[0]; dt = ktime_sub(ktime_get(), dt); res[1] = rc6_residency(rc6); rc0_freq = intel_rps_read_actual_frequency_fw(rps); @@ -79,11 +87,12 @@ int live_rc6_manual(void *arg) } if (has_power) { - rc0_power = div64_u64(NSEC_PER_SEC * rc0_power, + rc0_power = div64_u64(NSEC_PER_SEC * rc0_sample_energy[1], ktime_to_ns(dt)); + if (!rc0_power) { if (rc0_freq) - pr_debug("No power measured while in RC0! GPU Freq: %u in RC0\n", + pr_debug("No power measured while in RC0! GPU Freq: %uMHz in RC0\n", rc0_freq); else pr_err("No power and freq measured while in RC0\n"); @@ -98,10 +107,10 @@ int live_rc6_manual(void *arg) res[0] = rc6_residency(rc6); intel_uncore_forcewake_flush(rc6_to_uncore(rc6), FORCEWAKE_ALL); dt = ktime_get(); - rc6_power = librapl_energy_uJ(); - msleep(1000); + rc6_sample_energy[0] = librapl_energy_uJ(); + msleep(sleep_time); rc6_freq = intel_rps_read_actual_frequency_fw(rps); - rc6_power = librapl_energy_uJ() - rc6_power; + rc6_sample_energy[1] = librapl_energy_uJ() - rc6_sample_energy[0]; dt = ktime_sub(ktime_get(), dt); res[1] = rc6_residency(rc6); if (res[1] == res[0]) { @@ -113,13 +122,24 @@ int live_rc6_manual(void *arg) } if (has_power) { - rc6_power = div64_u64(NSEC_PER_SEC * rc6_power, + rc6_power = div64_u64(NSEC_PER_SEC * rc6_sample_energy[1], ktime_to_ns(dt)); - pr_info("GPU consumed %llduW in RC0 and %llduW in RC6\n", + pr_info("GPU consumed %lluuW in RC0 and %lluuW in RC6\n", rc0_power, rc6_power); + if (2 * rc6_power > rc0_power) { - pr_err("GPU leaked energy while in RC6! GPU Freq: %u in RC6 and %u in RC0\n", - rc6_freq, rc0_freq); + pr_err("GPU leaked energy while in RC6!\n" + "GPU Freq: %uMHz in RC6 and %uMHz in RC0\n" + "RC0 energy before & after sleep respectively: %lluuJ %lluuJ\n" + "RC6 energy before & after sleep respectively: %lluuJ %lluuJ\n", + rc6_freq, rc0_freq, rc0_sample_energy[0], rc0_sample_energy[1], + rc6_sample_energy[0], rc6_sample_energy[1]); + + diff = res[1] - res[0]; + threshold = (9 * NSEC_PER_MSEC * sleep_time) / 10; + if (diff < threshold) + pr_err("Did not enter RC6 properly, RC6 start residency=%lluns, RC6 end residency=%lluns\n", + res[0], res[1]); err = -EINVAL; goto out_unlock; } diff --git a/drivers/gpu/drm/i915/gt/selftest_tlb.c b/drivers/gpu/drm/i915/gt/selftest_tlb.c index 3941f2d6fa47..69ed946a39e5 100644 --- a/drivers/gpu/drm/i915/gt/selftest_tlb.c +++ b/drivers/gpu/drm/i915/gt/selftest_tlb.c @@ -143,7 +143,7 @@ pte_tlbinv(struct intel_context *ce, if (ce->engine->class == OTHER_CLASS) msleep(200); else - msleep(10); + usleep_range(10000, 20000); if (va == vb) { if (!i915_request_completed(rq)) { diff --git a/drivers/gpu/drm/i915/gt/uc/intel_gsc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_gsc_fw.c index 5dc0ccd07636..d550eb6edfb8 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_gsc_fw.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_gsc_fw.c @@ -230,7 +230,7 @@ int intel_gsc_fw_get_binary_info(struct intel_uc_fw *gsc_fw, const void *data, s gt_info(gt, "Invalid GSC firmware for MTL/ARL, got %d.%d.%d.%d but need 102.x.x.x", gsc->release.major, gsc->release.minor, gsc->release.patch, gsc->release.build); - return -EINVAL; + return -EINVAL; } if (min_ver.major) { diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index e6e9010462e3..1344e6d20a34 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -56,6 +56,7 @@ #include "display/intel_pps_regs.h" #include "display/intel_psr_regs.h" #include "display/intel_sprite_regs.h" +#include "display/intel_vga_regs.h" #include "display/skl_universal_plane_regs.h" #include "display/skl_watermark_regs.h" #include "display/vlv_dsi_pll_regs.h" diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c index f5262b8ad237..273bc43468a0 100644 --- a/drivers/gpu/drm/i915/i915_driver.c +++ b/drivers/gpu/drm/i915/i915_driver.c @@ -62,7 +62,6 @@ #include "display/intel_pch_refclk.h" #include "display/intel_pps.h" #include "display/intel_sprite_uapi.h" -#include "display/intel_vga.h" #include "display/skl_watermark.h" #include "gem/i915_gem_context.h" @@ -235,7 +234,6 @@ static int i915_driver_early_probe(struct drm_i915_private *dev_priv) intel_uncore_mmio_debug_init_early(dev_priv); - spin_lock_init(&dev_priv->irq_lock); spin_lock_init(&dev_priv->gpu_error.lock); intel_sbi_init(dev_priv); @@ -263,9 +261,6 @@ static int i915_driver_early_probe(struct drm_i915_private *dev_priv) i915_gem_init_early(dev_priv); - /* This must be called before any calls to HAS_PCH_* */ - intel_detect_pch(dev_priv); - intel_irq_init(dev_priv); intel_display_driver_early_probe(display); intel_clock_gating_hooks_init(dev_priv); @@ -1205,8 +1200,6 @@ static int i915_drm_resume(struct drm_device *dev) i9xx_display_sr_restore(display); - intel_vga_redisable(display); - intel_gmbus_reset(display); intel_pps_unlock_regs_wa(display); diff --git a/drivers/gpu/drm/i915/i915_driver.h b/drivers/gpu/drm/i915/i915_driver.h index 4b67ad9a61cd..1e95ecb2a163 100644 --- a/drivers/gpu/drm/i915/i915_driver.h +++ b/drivers/gpu/drm/i915/i915_driver.h @@ -15,7 +15,6 @@ struct drm_printer; #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_TIMESTAMP 1695980603 extern const struct dev_pm_ops i915_pm_ops; diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 236c48d282e4..d0e1980dcba2 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -49,8 +49,6 @@ #include "gt/intel_workarounds.h" #include "gt/uc/intel_uc.h" -#include "soc/intel_pch.h" - #include "i915_drm_client.h" #include "i915_gem.h" #include "i915_gpu_error.h" @@ -224,8 +222,6 @@ struct drm_i915_private { }; unsigned int engine_uabi_class_count[I915_LAST_UABI_ENGINE_CLASS + 1]; - /* protects the irq masks */ - spinlock_t irq_lock; bool irqs_enabled; /* LPT/WPT IOSF sideband protection */ @@ -272,9 +268,6 @@ struct drm_i915_private { /* pm private clock gating functions */ const struct drm_i915_clock_gating_funcs *clock_gating_funcs; - /* PCH chipset type */ - enum intel_pch pch_type; - unsigned long gem_quirks; struct i915_gem_mm mm; diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index a6613eed3398..4f785cdbd155 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -726,13 +726,6 @@ static void err_print_gt_info(struct drm_i915_error_state_buf *m, intel_sseu_print_topology(gt->_gt->i915, >->info.sseu, &p); } -static void err_print_gt_display(struct drm_i915_error_state_buf *m, - struct intel_gt_coredump *gt) -{ - err_printf(m, "IER: 0x%08x\n", gt->ier); - err_printf(m, "DERRMR: 0x%08x\n", gt->derrmr); -} - static void err_print_gt_global_nonguc(struct drm_i915_error_state_buf *m, struct intel_gt_coredump *gt) { @@ -878,7 +871,6 @@ static void __err_print_to_sgl(struct drm_i915_error_state_buf *m, if (error->gt->uc && error->gt->uc->guc.is_guc_capture) print_guc_capture = true; - err_print_gt_display(m, error->gt); err_print_gt_global_nonguc(m, error->gt); err_print_gt_fences(m, error->gt); @@ -1767,27 +1759,6 @@ gt_record_uc(struct intel_gt_coredump *gt, return error_uc; } -/* Capture display registers. */ -static void gt_record_display_regs(struct intel_gt_coredump *gt) -{ - struct intel_uncore *uncore = gt->_gt->uncore; - struct drm_i915_private *i915 = uncore->i915; - - if (DISPLAY_VER(i915) >= 6 && DISPLAY_VER(i915) < 20) - gt->derrmr = intel_uncore_read(uncore, DERRMR); - - if (GRAPHICS_VER(i915) >= 8) - gt->ier = intel_uncore_read(uncore, GEN8_DE_MISC_IER); - else if (IS_VALLEYVIEW(i915)) - gt->ier = intel_uncore_read(uncore, VLV_IER); - else if (HAS_PCH_SPLIT(i915)) - gt->ier = intel_uncore_read(uncore, DEIER); - else if (GRAPHICS_VER(i915) == 2) - gt->ier = intel_uncore_read16(uncore, GEN2_IER); - else - gt->ier = intel_uncore_read(uncore, GEN2_IER); -} - /* Capture all other registers that GuC doesn't capture. */ static void gt_record_global_nonguc_regs(struct intel_gt_coredump *gt) { @@ -1821,9 +1792,12 @@ static void gt_record_global_nonguc_regs(struct intel_gt_coredump *gt) gt->gtier[i] = intel_uncore_read(uncore, GEN8_GT_IER(i)); gt->ngtier = 4; - } else if (HAS_PCH_SPLIT(i915)) { + } else if (GRAPHICS_VER(i915) >= 5) { gt->gtier[0] = intel_uncore_read(uncore, GTIER); gt->ngtier = 1; + } else { + gt->gtier[0] = intel_uncore_read(uncore, GEN2_IER); + gt->ngtier = 1; } gt->eir = intel_uncore_read(uncore, EIR); @@ -2043,7 +2017,6 @@ intel_gt_coredump_alloc(struct intel_gt *gt, gfp_t gfp, u32 dump_flags) gc->_gt = gt; gc->awake = intel_gt_pm_is_awake(gt); - gt_record_display_regs(gc); gt_record_global_nonguc_regs(gc); /* @@ -2160,7 +2133,6 @@ i915_gpu_coredump(struct intel_gt *gt, intel_engine_mask_t engine_mask, u32 dump void i915_error_state_store(struct i915_gpu_coredump *error) { struct drm_i915_private *i915; - static bool warned; if (IS_ERR_OR_NULL(error)) return; @@ -2174,16 +2146,8 @@ void i915_error_state_store(struct i915_gpu_coredump *error) i915_gpu_coredump_get(error); - if (!xchg(&warned, true) && - ktime_get_real_seconds() - DRIVER_TIMESTAMP < DAY_AS_SECONDS(180)) { - pr_info("GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.\n"); - pr_info("Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new.\n"); - pr_info("Please see https://drm.pages.freedesktop.org/intel-docs/how-to-file-i915-bugs.html for details.\n"); - pr_info("drm/i915 developers can then reassign to the right component if it's not a kernel issue.\n"); - pr_info("The GPU crash dump is required to analyze GPU hangs, so please always attach it.\n"); - pr_info("GPU crash dump saved to /sys/class/drm/card%d/error\n", - i915->drm.primary->index); - } + drm_info(&i915->drm, "GPU error state saved to /sys/class/drm/card%d/error\n", + i915->drm.primary->index); } /** diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h index 749e1c55613e..182324979278 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.h +++ b/drivers/gpu/drm/i915/i915_gpu_error.h @@ -146,7 +146,6 @@ struct intel_gt_coredump { /* Generic register state */ u32 eir; u32 pgtbl_er; - u32 ier; u32 gtier[6], ngtier; u32 forcewake; u32 error; /* gen6+ */ @@ -164,8 +163,6 @@ struct intel_gt_coredump { u32 clock_frequency; u32 clock_period_ns; - /* Display related */ - u32 derrmr; u32 sfc_done[I915_MAX_SFC]; /* gen12 */ u32 nfence; diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index c1f938a1da44..95042879bec4 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -658,23 +658,9 @@ static irqreturn_t dg1_irq_handler(int irq, void *arg) return IRQ_HANDLED; } -static void ibx_irq_reset(struct drm_i915_private *dev_priv) -{ - struct intel_uncore *uncore = &dev_priv->uncore; - - if (HAS_PCH_NOP(dev_priv)) - return; - - gen2_irq_reset(uncore, SDE_IRQ_REGS); - - if (HAS_PCH_CPT(dev_priv) || HAS_PCH_LPT(dev_priv)) - intel_uncore_write(&dev_priv->uncore, SERR_INT, 0xffffffff); -} - -/* drm_dma.h hooks -*/ static void ilk_irq_reset(struct drm_i915_private *dev_priv) { + struct intel_display *display = &dev_priv->display; struct intel_uncore *uncore = &dev_priv->uncore; gen2_irq_reset(uncore, DE_IRQ_REGS); @@ -690,7 +676,7 @@ static void ilk_irq_reset(struct drm_i915_private *dev_priv) gen5_gt_irq_reset(to_gt(dev_priv)); - ibx_irq_reset(dev_priv); + ibx_display_irq_reset(display); } static void valleyview_irq_reset(struct drm_i915_private *dev_priv) @@ -702,9 +688,7 @@ static void valleyview_irq_reset(struct drm_i915_private *dev_priv) gen5_gt_irq_reset(to_gt(dev_priv)); - spin_lock_irq(&dev_priv->irq_lock); vlv_display_irq_reset(display); - spin_unlock_irq(&dev_priv->irq_lock); } static void gen8_irq_reset(struct drm_i915_private *dev_priv) @@ -717,10 +701,6 @@ static void gen8_irq_reset(struct drm_i915_private *dev_priv) gen8_gt_irq_reset(to_gt(dev_priv)); gen8_display_irq_reset(display); gen2_irq_reset(uncore, GEN8_PCU_IRQ_REGS); - - if (HAS_PCH_SPLIT(dev_priv)) - ibx_irq_reset(dev_priv); - } static void gen11_irq_reset(struct drm_i915_private *dev_priv) @@ -770,9 +750,7 @@ static void cherryview_irq_reset(struct drm_i915_private *dev_priv) gen2_irq_reset(uncore, GEN8_PCU_IRQ_REGS); - spin_lock_irq(&dev_priv->irq_lock); vlv_display_irq_reset(display); - spin_unlock_irq(&dev_priv->irq_lock); } static void ilk_irq_postinstall(struct drm_i915_private *dev_priv) @@ -790,9 +768,7 @@ static void valleyview_irq_postinstall(struct drm_i915_private *dev_priv) gen5_gt_irq_postinstall(to_gt(dev_priv)); - spin_lock_irq(&dev_priv->irq_lock); vlv_display_irq_postinstall(display); - spin_unlock_irq(&dev_priv->irq_lock); intel_uncore_write(&dev_priv->uncore, VLV_MASTER_IER, MASTER_INTERRUPT_ENABLE); intel_uncore_posting_read(&dev_priv->uncore, VLV_MASTER_IER); @@ -849,9 +825,7 @@ static void cherryview_irq_postinstall(struct drm_i915_private *dev_priv) gen8_gt_irq_postinstall(to_gt(dev_priv)); - spin_lock_irq(&dev_priv->irq_lock); vlv_display_irq_postinstall(display); - spin_unlock_irq(&dev_priv->irq_lock); intel_uncore_write(&dev_priv->uncore, GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL); intel_uncore_posting_read(&dev_priv->uncore, GEN8_MASTER_IRQ); @@ -961,14 +935,7 @@ static void i915_irq_postinstall(struct drm_i915_private *dev_priv) gen2_irq_init(uncore, GEN2_IRQ_REGS, dev_priv->irq_mask, enable_mask); - /* Interrupt setup is already guaranteed to be single-threaded, this is - * just to make the assert_spin_locked check happy. */ - spin_lock_irq(&dev_priv->irq_lock); - i915_enable_pipestat(display, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS); - i915_enable_pipestat(display, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS); - spin_unlock_irq(&dev_priv->irq_lock); - - i915_enable_asle_pipestat(display); + i915_display_irq_postinstall(display); } static irqreturn_t i915_irq_handler(int irq, void *arg) @@ -1086,15 +1053,7 @@ static void i965_irq_postinstall(struct drm_i915_private *dev_priv) gen2_irq_init(uncore, GEN2_IRQ_REGS, dev_priv->irq_mask, enable_mask); - /* Interrupt setup is already guaranteed to be single-threaded, this is - * just to make the assert_spin_locked check happy. */ - spin_lock_irq(&dev_priv->irq_lock); - i915_enable_pipestat(display, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS); - i915_enable_pipestat(display, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS); - i915_enable_pipestat(display, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS); - spin_unlock_irq(&dev_priv->irq_lock); - - i915_enable_asle_pipestat(display); + i965_display_irq_postinstall(display); } static irqreturn_t i965_irq_handler(int irq, void *arg) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 49beab8e324d..2e4190da3e0d 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -866,6 +866,7 @@ #define FP_M2_DIV_MASK 0x0000003f #define FP_M2_PINEVIEW_DIV_MASK 0x000000ff #define FP_M2_DIV_SHIFT 0 + #define DPLL_TEST _MMIO(0x606c) #define DPLLB_TEST_SDVO_DIV_1 (0 << 22) #define DPLLB_TEST_SDVO_DIV_2 (1 << 22) @@ -877,11 +878,13 @@ #define DPLLA_TEST_N_BYPASS (1 << 3) #define DPLLA_TEST_M_BYPASS (1 << 2) #define DPLLA_INPUT_BUFFER_ENABLE (1 << 0) + #define D_STATE _MMIO(0x6104) #define DSTATE_GFX_RESET_I830 (1 << 6) #define DSTATE_PLL_D3_OFF (1 << 3) #define DSTATE_GFX_CLOCK_GATING (1 << 1) #define DSTATE_DOT_CLOCK_GATING (1 << 0) + #define DSPCLK_GATE_D(__i915) _MMIO(DISPLAY_MMIO_BASE(__i915) + 0x6200) # define DPUNIT_B_CLOCK_GATE_DISABLE (1 << 30) /* 965 */ # define VSUNIT_CLOCK_GATE_DISABLE (1 << 29) /* 965 */ @@ -1050,7 +1053,6 @@ /* * Overlay regs */ - #define OVADD _MMIO(0x30000) #define DOVSTA _MMIO(0x30008) #define OC_BUF (0x3 << 20) @@ -1106,7 +1108,6 @@ /* * Display engine regs */ - /* Pipe/transcoder A timing regs */ #define _TRANS_HTOTAL_A 0x60000 #define _TRANS_HTOTAL_B 0x61000 @@ -1397,88 +1398,50 @@ #define VLV_DP_B _MMIO(VLV_DISPLAY_BASE + 0x64100) #define VLV_DP_C _MMIO(VLV_DISPLAY_BASE + 0x64200) #define CHV_DP_D _MMIO(VLV_DISPLAY_BASE + 0x64300) -#define DP_PORT_EN (1 << 31) -#define DP_PIPE_SEL_SHIFT 30 -#define DP_PIPE_SEL_MASK (1 << 30) -#define DP_PIPE_SEL(pipe) ((pipe) << 30) -#define DP_PIPE_SEL_SHIFT_IVB 29 -#define DP_PIPE_SEL_MASK_IVB (3 << 29) -#define DP_PIPE_SEL_IVB(pipe) ((pipe) << 29) +#define DP_PORT_EN REG_BIT(31) +#define DP_PIPE_SEL_MASK REG_GENMASK(30, 30) +#define DP_PIPE_SEL(pipe) REG_FIELD_PREP(DP_PIPE_SEL_MASK, (pipe)) +#define DP_PIPE_SEL_MASK_IVB REG_GENMASK(30, 29) +#define DP_PIPE_SEL_IVB(pipe) REG_FIELD_PREP(DP_PIPE_SEL_MASK_IVB, (pipe)) #define DP_PIPE_SEL_SHIFT_CHV 16 -#define DP_PIPE_SEL_MASK_CHV (3 << 16) -#define DP_PIPE_SEL_CHV(pipe) ((pipe) << 16) - -/* Link training mode - select a suitable mode for each stage */ -#define DP_LINK_TRAIN_PAT_1 (0 << 28) -#define DP_LINK_TRAIN_PAT_2 (1 << 28) -#define DP_LINK_TRAIN_PAT_IDLE (2 << 28) -#define DP_LINK_TRAIN_OFF (3 << 28) -#define DP_LINK_TRAIN_MASK (3 << 28) -#define DP_LINK_TRAIN_SHIFT 28 - -/* CPT Link training mode */ -#define DP_LINK_TRAIN_PAT_1_CPT (0 << 8) -#define DP_LINK_TRAIN_PAT_2_CPT (1 << 8) -#define DP_LINK_TRAIN_PAT_IDLE_CPT (2 << 8) -#define DP_LINK_TRAIN_OFF_CPT (3 << 8) -#define DP_LINK_TRAIN_MASK_CPT (7 << 8) -#define DP_LINK_TRAIN_SHIFT_CPT 8 - -/* Signal voltages. These are mostly controlled by the other end */ -#define DP_VOLTAGE_0_4 (0 << 25) -#define DP_VOLTAGE_0_6 (1 << 25) -#define DP_VOLTAGE_0_8 (2 << 25) -#define DP_VOLTAGE_1_2 (3 << 25) -#define DP_VOLTAGE_MASK (7 << 25) -#define DP_VOLTAGE_SHIFT 25 - -/* Signal pre-emphasis levels, like voltages, the other end tells us what - * they want - */ -#define DP_PRE_EMPHASIS_0 (0 << 22) -#define DP_PRE_EMPHASIS_3_5 (1 << 22) -#define DP_PRE_EMPHASIS_6 (2 << 22) -#define DP_PRE_EMPHASIS_9_5 (3 << 22) -#define DP_PRE_EMPHASIS_MASK (7 << 22) -#define DP_PRE_EMPHASIS_SHIFT 22 - -/* How many wires to use. I guess 3 was too hard */ -#define DP_PORT_WIDTH(width) (((width) - 1) << 19) -#define DP_PORT_WIDTH_MASK (7 << 19) -#define DP_PORT_WIDTH_SHIFT 19 - -/* Mystic DPCD version 1.1 special mode */ -#define DP_ENHANCED_FRAMING (1 << 18) - -/* eDP */ -#define DP_PLL_FREQ_270MHZ (0 << 16) -#define DP_PLL_FREQ_162MHZ (1 << 16) -#define DP_PLL_FREQ_MASK (3 << 16) - -/* locked once port is enabled */ -#define DP_PORT_REVERSAL (1 << 15) - -/* eDP */ -#define DP_PLL_ENABLE (1 << 14) - -/* sends the clock on lane 15 of the PEG for debug */ -#define DP_CLOCK_OUTPUT_ENABLE (1 << 13) - -#define DP_SCRAMBLING_DISABLE (1 << 12) -#define DP_SCRAMBLING_DISABLE_IRONLAKE (1 << 7) - -/* limit RGB values to avoid confusing TVs */ -#define DP_COLOR_RANGE_16_235 (1 << 8) - -/* Turn on the audio link */ -#define DP_AUDIO_OUTPUT_ENABLE (1 << 6) - -/* vs and hs sync polarity */ -#define DP_SYNC_VS_HIGH (1 << 4) -#define DP_SYNC_HS_HIGH (1 << 3) - -/* A fantasy */ -#define DP_DETECTED (1 << 2) +#define DP_PIPE_SEL_MASK_CHV REG_GENMASK(17, 16) +#define DP_PIPE_SEL_CHV(pipe) REG_FIELD_PREP(DP_PIPE_SEL_MASK_CHV, (pipe)) +#define DP_LINK_TRAIN_MASK REG_GENMASK(29, 28) +#define DP_LINK_TRAIN_PAT_1 REG_FIELD_PREP(DP_LINK_TRAIN_MASK, 0) +#define DP_LINK_TRAIN_PAT_2 REG_FIELD_PREP(DP_LINK_TRAIN_MASK, 1) +#define DP_LINK_TRAIN_PAT_IDLE REG_FIELD_PREP(DP_LINK_TRAIN_MASK, 2) +#define DP_LINK_TRAIN_OFF REG_FIELD_PREP(DP_LINK_TRAIN_MASK, 3) +#define DP_LINK_TRAIN_MASK_CPT REG_GENMASK(10, 8) +#define DP_LINK_TRAIN_PAT_1_CPT REG_FIELD_PREP(DP_LINK_TRAIN_MASK_CPT, 0) +#define DP_LINK_TRAIN_PAT_2_CPT REG_FIELD_PREP(DP_LINK_TRAIN_MASK_CPT, 1) +#define DP_LINK_TRAIN_PAT_IDLE_CPT REG_FIELD_PREP(DP_LINK_TRAIN_MASK_CPT, 2) +#define DP_LINK_TRAIN_OFF_CPT REG_FIELD_PREP(DP_LINK_TRAIN_MASK_CPT, 3) +#define DP_VOLTAGE_MASK REG_GENMASK(27, 25) +#define DP_VOLTAGE_0_4 REG_FIELD_PREP(DP_VOLTAGE_MASK, 0) +#define DP_VOLTAGE_0_6 REG_FIELD_PREP(DP_VOLTAGE_MASK, 1) +#define DP_VOLTAGE_0_8 REG_FIELD_PREP(DP_VOLTAGE_MASK, 2) +#define DP_VOLTAGE_1_2 REG_FIELD_PREP(DP_VOLTAGE_MASK, 3) +#define DP_PRE_EMPHASIS_MASK REG_GENMASK(24, 22) +#define DP_PRE_EMPHASIS_0 REG_FIELD_PREP(DP_PRE_EMPHASIS_MASK, 0) +#define DP_PRE_EMPHASIS_3_5 REG_FIELD_PREP(DP_PRE_EMPHASIS_MASK, 1) +#define DP_PRE_EMPHASIS_6 REG_FIELD_PREP(DP_PRE_EMPHASIS_MASK, 2) +#define DP_PRE_EMPHASIS_9_5 REG_FIELD_PREP(DP_PRE_EMPHASIS_MASK, 3) +#define DP_PORT_WIDTH_MASK REG_GENMASK(21, 19) +#define DP_PORT_WIDTH(width) REG_FIELD_PREP(DP_PORT_WIDTH_MASK, (width) - 1) +#define DP_ENHANCED_FRAMING REG_BIT(18) +#define EDP_PLL_FREQ_MASK REG_GENMASK(17, 16) +#define EDP_PLL_FREQ_270MHZ REG_FIELD_PREP(EDP_PLL_FREQ_MASK, 0) +#define EDP_PLL_FREQ_162MHZ REG_FIELD_PREP(EDP_PLL_FREQ_MASK, 1) +#define DP_PORT_REVERSAL REG_BIT(15) +#define EDP_PLL_ENABLE REG_BIT(14) +#define DP_CLOCK_OUTPUT_ENABLE REG_BIT(13) +#define DP_SCRAMBLING_DISABLE REG_BIT(12) +#define DP_SCRAMBLING_DISABLE_ILK REG_BIT(7) +#define DP_COLOR_RANGE_16_235 REG_BIT(8) +#define DP_AUDIO_OUTPUT_ENABLE REG_BIT(6) +#define DP_SYNC_VS_HIGH REG_BIT(4) +#define DP_SYNC_HS_HIGH REG_BIT(3) +#define DP_DETECTED REG_BIT(2) /* * Computing GMCH M and N values for the Display Port link @@ -1812,18 +1775,6 @@ #define SWF3(dev_priv, i) _MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x72414 + (i) * 4) #define SWF_ILK(i) _MMIO(0x4F000 + (i) * 4) -/* VBIOS regs */ -#define VGACNTRL _MMIO(0x71400) -# define VGA_DISP_DISABLE (1 << 31) -# define VGA_2X_MODE (1 << 30) -# define VGA_PIPE_B_SELECT (1 << 29) - -#define VLV_VGACNTRL _MMIO(VLV_DISPLAY_BASE + 0x71400) - -/* Ironlake */ - -#define CPU_VGACNTRL _MMIO(0x41000) - #define DIGITAL_PORT_HOTPLUG_CNTRL _MMIO(0x44030) #define DIGITAL_PORTA_HOTPLUG_ENABLE (1 << 4) #define DIGITAL_PORTA_PULSE_DURATION_2ms (0 << 2) /* pre-HSW */ @@ -2784,7 +2735,6 @@ * functionality covered in PCH_PORT_HOTPLUG is split into * SHOTPLUG_CTL_DDI and SHOTPLUG_CTL_TC. */ - #define SHOTPLUG_CTL_DDI _MMIO(0xc4030) #define SHOTPLUG_CTL_DDI_HPD_ENABLE(hpd_pin) (0x8 << (_HPD_PIN_DDI(hpd_pin) * 4)) #define SHOTPLUG_CTL_DDI_HPD_OUTPUT_DATA(hpd_pin) (0x4 << (_HPD_PIN_DDI(hpd_pin) * 4)) @@ -2864,7 +2814,6 @@ #define TRANS_DPLL_ENABLE(pipe) (1 << ((pipe) * 4 + 3)) /* transcoder */ - #define _PCH_TRANS_HTOTAL_A 0xe0000 #define _PCH_TRANS_HTOTAL_B 0xe1000 #define PCH_TRANS_HTOTAL(pipe) _MMIO_PIPE(pipe, _PCH_TRANS_HTOTAL_A, _PCH_TRANS_HTOTAL_B) @@ -3795,7 +3744,6 @@ enum skl_power_gate { /* * SKL Clocks */ - /* CDCLK_CTL */ #define CDCLK_CTL _MMIO(0x46000) #define CDCLK_FREQ_SEL_MASK REG_GENMASK(27, 26) diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h index 609214231ffc..f7fb40cfdb70 100644 --- a/drivers/gpu/drm/i915/i915_utils.h +++ b/drivers/gpu/drm/i915/i915_utils.h @@ -40,8 +40,6 @@ struct drm_i915_private; struct timer_list; -#define FDO_BUG_URL "https://drm.pages.freedesktop.org/intel-docs/how-to-file-i915-bugs.html" - #define MISSING_CASE(x) WARN(1, "Missing case (%s == %ld)\n", \ __stringify(x), (long)(x)) diff --git a/drivers/gpu/drm/i915/intel_gvt_mmio_table.c b/drivers/gpu/drm/i915/intel_gvt_mmio_table.c index 76d84cbb8361..d581a9d2c063 100644 --- a/drivers/gpu/drm/i915/intel_gvt_mmio_table.c +++ b/drivers/gpu/drm/i915/intel_gvt_mmio_table.c @@ -21,6 +21,7 @@ #include "display/intel_pfit_regs.h" #include "display/intel_psr_regs.h" #include "display/intel_sprite_regs.h" +#include "display/intel_vga_regs.h" #include "display/skl_universal_plane_regs.h" #include "display/skl_watermark_regs.h" #include "display/vlv_dsi_pll_regs.h" diff --git a/drivers/gpu/drm/i915/soc/intel_pch.c b/drivers/gpu/drm/i915/soc/intel_pch.c deleted file mode 100644 index 82dc7fbd1a3e..000000000000 --- a/drivers/gpu/drm/i915/soc/intel_pch.c +++ /dev/null @@ -1,316 +0,0 @@ -// SPDX-License-Identifier: MIT -/* - * Copyright 2019 Intel Corporation. - */ - -#include "i915_drv.h" -#include "i915_utils.h" -#include "intel_pch.h" - -#define INTEL_PCH_DEVICE_ID_MASK 0xff80 -#define INTEL_PCH_IBX_DEVICE_ID_TYPE 0x3b00 -#define INTEL_PCH_CPT_DEVICE_ID_TYPE 0x1c00 -#define INTEL_PCH_PPT_DEVICE_ID_TYPE 0x1e00 -#define INTEL_PCH_LPT_DEVICE_ID_TYPE 0x8c00 -#define INTEL_PCH_LPT_LP_DEVICE_ID_TYPE 0x9c00 -#define INTEL_PCH_WPT_DEVICE_ID_TYPE 0x8c80 -#define INTEL_PCH_WPT_LP_DEVICE_ID_TYPE 0x9c80 -#define INTEL_PCH_SPT_DEVICE_ID_TYPE 0xA100 -#define INTEL_PCH_SPT_LP_DEVICE_ID_TYPE 0x9D00 -#define INTEL_PCH_KBP_DEVICE_ID_TYPE 0xA280 -#define INTEL_PCH_CNP_DEVICE_ID_TYPE 0xA300 -#define INTEL_PCH_CNP_LP_DEVICE_ID_TYPE 0x9D80 -#define INTEL_PCH_CMP_DEVICE_ID_TYPE 0x0280 -#define INTEL_PCH_CMP2_DEVICE_ID_TYPE 0x0680 -#define INTEL_PCH_CMP_V_DEVICE_ID_TYPE 0xA380 -#define INTEL_PCH_ICP_DEVICE_ID_TYPE 0x3480 -#define INTEL_PCH_ICP2_DEVICE_ID_TYPE 0x3880 -#define INTEL_PCH_MCC_DEVICE_ID_TYPE 0x4B00 -#define INTEL_PCH_TGP_DEVICE_ID_TYPE 0xA080 -#define INTEL_PCH_TGP2_DEVICE_ID_TYPE 0x4380 -#define INTEL_PCH_JSP_DEVICE_ID_TYPE 0x4D80 -#define INTEL_PCH_ADP_DEVICE_ID_TYPE 0x7A80 -#define INTEL_PCH_ADP2_DEVICE_ID_TYPE 0x5180 -#define INTEL_PCH_ADP3_DEVICE_ID_TYPE 0x7A00 -#define INTEL_PCH_ADP4_DEVICE_ID_TYPE 0x5480 -#define INTEL_PCH_P2X_DEVICE_ID_TYPE 0x7100 -#define INTEL_PCH_P3X_DEVICE_ID_TYPE 0x7000 -#define INTEL_PCH_QEMU_DEVICE_ID_TYPE 0x2900 /* qemu q35 has 2918 */ - -/* Map PCH device id to PCH type, or PCH_NONE if unknown. */ -static enum intel_pch -intel_pch_type(const struct drm_i915_private *dev_priv, unsigned short id) -{ - switch (id) { - case INTEL_PCH_IBX_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Ibex Peak PCH\n"); - drm_WARN_ON(&dev_priv->drm, GRAPHICS_VER(dev_priv) != 5); - return PCH_IBX; - case INTEL_PCH_CPT_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found CougarPoint PCH\n"); - drm_WARN_ON(&dev_priv->drm, - GRAPHICS_VER(dev_priv) != 6 && !IS_IVYBRIDGE(dev_priv)); - return PCH_CPT; - case INTEL_PCH_PPT_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found PantherPoint PCH\n"); - drm_WARN_ON(&dev_priv->drm, - GRAPHICS_VER(dev_priv) != 6 && !IS_IVYBRIDGE(dev_priv)); - /* PPT is CPT compatible */ - return PCH_CPT; - case INTEL_PCH_LPT_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found LynxPoint PCH\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv)); - drm_WARN_ON(&dev_priv->drm, - IS_HASWELL_ULT(dev_priv) || IS_BROADWELL_ULT(dev_priv)); - return PCH_LPT_H; - case INTEL_PCH_LPT_LP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found LynxPoint LP PCH\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv)); - drm_WARN_ON(&dev_priv->drm, - !IS_HASWELL_ULT(dev_priv) && !IS_BROADWELL_ULT(dev_priv)); - return PCH_LPT_LP; - case INTEL_PCH_WPT_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found WildcatPoint PCH\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv)); - drm_WARN_ON(&dev_priv->drm, - IS_HASWELL_ULT(dev_priv) || IS_BROADWELL_ULT(dev_priv)); - /* WPT is LPT compatible */ - return PCH_LPT_H; - case INTEL_PCH_WPT_LP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found WildcatPoint LP PCH\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv)); - drm_WARN_ON(&dev_priv->drm, - !IS_HASWELL_ULT(dev_priv) && !IS_BROADWELL_ULT(dev_priv)); - /* WPT is LPT compatible */ - return PCH_LPT_LP; - case INTEL_PCH_SPT_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found SunrisePoint PCH\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_SKYLAKE(dev_priv) && !IS_KABYLAKE(dev_priv)); - return PCH_SPT; - case INTEL_PCH_SPT_LP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found SunrisePoint LP PCH\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_SKYLAKE(dev_priv) && - !IS_KABYLAKE(dev_priv) && - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv)); - return PCH_SPT; - case INTEL_PCH_KBP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Kaby Lake PCH (KBP)\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_SKYLAKE(dev_priv) && - !IS_KABYLAKE(dev_priv) && - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv)); - /* KBP is SPT compatible */ - return PCH_SPT; - case INTEL_PCH_CNP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Cannon Lake PCH (CNP)\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv)); - return PCH_CNP; - case INTEL_PCH_CNP_LP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, - "Found Cannon Lake LP PCH (CNP-LP)\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv)); - return PCH_CNP; - case INTEL_PCH_CMP_DEVICE_ID_TYPE: - case INTEL_PCH_CMP2_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Comet Lake PCH (CMP)\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv) && - !IS_ROCKETLAKE(dev_priv)); - /* CMP is CNP compatible */ - return PCH_CNP; - case INTEL_PCH_CMP_V_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Comet Lake V PCH (CMP-V)\n"); - drm_WARN_ON(&dev_priv->drm, - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv)); - /* CMP-V is based on KBP, which is SPT compatible */ - return PCH_SPT; - case INTEL_PCH_ICP_DEVICE_ID_TYPE: - case INTEL_PCH_ICP2_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Ice Lake PCH\n"); - drm_WARN_ON(&dev_priv->drm, !IS_ICELAKE(dev_priv)); - return PCH_ICP; - case INTEL_PCH_MCC_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Mule Creek Canyon PCH\n"); - drm_WARN_ON(&dev_priv->drm, !(IS_JASPERLAKE(dev_priv) || - IS_ELKHARTLAKE(dev_priv))); - /* MCC is TGP compatible */ - return PCH_TGP; - case INTEL_PCH_TGP_DEVICE_ID_TYPE: - case INTEL_PCH_TGP2_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Tiger Lake LP PCH\n"); - drm_WARN_ON(&dev_priv->drm, !IS_TIGERLAKE(dev_priv) && - !IS_ROCKETLAKE(dev_priv) && - !IS_SKYLAKE(dev_priv) && - !IS_KABYLAKE(dev_priv) && - !IS_COFFEELAKE(dev_priv) && - !IS_COMETLAKE(dev_priv)); - return PCH_TGP; - case INTEL_PCH_JSP_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Jasper Lake PCH\n"); - drm_WARN_ON(&dev_priv->drm, !(IS_JASPERLAKE(dev_priv) || - IS_ELKHARTLAKE(dev_priv))); - /* JSP is ICP compatible */ - return PCH_ICP; - case INTEL_PCH_ADP_DEVICE_ID_TYPE: - case INTEL_PCH_ADP2_DEVICE_ID_TYPE: - case INTEL_PCH_ADP3_DEVICE_ID_TYPE: - case INTEL_PCH_ADP4_DEVICE_ID_TYPE: - drm_dbg_kms(&dev_priv->drm, "Found Alder Lake PCH\n"); - drm_WARN_ON(&dev_priv->drm, !IS_ALDERLAKE_S(dev_priv) && - !IS_ALDERLAKE_P(dev_priv)); - return PCH_ADP; - default: - return PCH_NONE; - } -} - -static bool intel_is_virt_pch(unsigned short id, - unsigned short svendor, unsigned short sdevice) -{ - return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE || - id == INTEL_PCH_P3X_DEVICE_ID_TYPE || - (id == INTEL_PCH_QEMU_DEVICE_ID_TYPE && - svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET && - sdevice == PCI_SUBDEVICE_ID_QEMU)); -} - -static void -intel_virt_detect_pch(const struct drm_i915_private *dev_priv, - unsigned short *pch_id, enum intel_pch *pch_type) -{ - unsigned short id = 0; - - /* - * In a virtualized passthrough environment we can be in a - * setup where the ISA bridge is not able to be passed through. - * In this case, a south bridge can be emulated and we have to - * make an educated guess as to which PCH is really there. - */ - - if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv)) - id = INTEL_PCH_ADP_DEVICE_ID_TYPE; - else if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv)) - id = INTEL_PCH_TGP_DEVICE_ID_TYPE; - else if (IS_JASPERLAKE(dev_priv) || IS_ELKHARTLAKE(dev_priv)) - id = INTEL_PCH_MCC_DEVICE_ID_TYPE; - else if (IS_ICELAKE(dev_priv)) - id = INTEL_PCH_ICP_DEVICE_ID_TYPE; - else if (IS_COFFEELAKE(dev_priv) || - IS_COMETLAKE(dev_priv)) - id = INTEL_PCH_CNP_DEVICE_ID_TYPE; - else if (IS_KABYLAKE(dev_priv) || IS_SKYLAKE(dev_priv)) - id = INTEL_PCH_SPT_DEVICE_ID_TYPE; - else if (IS_HASWELL_ULT(dev_priv) || IS_BROADWELL_ULT(dev_priv)) - id = INTEL_PCH_LPT_LP_DEVICE_ID_TYPE; - else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) - id = INTEL_PCH_LPT_DEVICE_ID_TYPE; - else if (GRAPHICS_VER(dev_priv) == 6 || IS_IVYBRIDGE(dev_priv)) - id = INTEL_PCH_CPT_DEVICE_ID_TYPE; - else if (GRAPHICS_VER(dev_priv) == 5) - id = INTEL_PCH_IBX_DEVICE_ID_TYPE; - - if (id) - drm_dbg_kms(&dev_priv->drm, "Assuming PCH ID %04x\n", id); - else - drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n"); - - *pch_type = intel_pch_type(dev_priv, id); - - /* Sanity check virtual PCH id */ - if (drm_WARN_ON(&dev_priv->drm, - id && *pch_type == PCH_NONE)) - id = 0; - - *pch_id = id; -} - -void intel_detect_pch(struct drm_i915_private *dev_priv) -{ - struct pci_dev *pch = NULL; - unsigned short id; - enum intel_pch pch_type; - - /* - * South display engine on the same PCI device: just assign the fake - * PCH. - */ - if (DISPLAY_VER(dev_priv) >= 20) { - dev_priv->pch_type = PCH_LNL; - return; - } else if (IS_BATTLEMAGE(dev_priv) || IS_METEORLAKE(dev_priv)) { - /* - * Both north display and south display are on the SoC die. - * The real PCH (if it even exists) is uninvolved in display. - */ - dev_priv->pch_type = PCH_MTL; - return; - } else if (IS_DG2(dev_priv)) { - dev_priv->pch_type = PCH_DG2; - return; - } else if (IS_DG1(dev_priv)) { - dev_priv->pch_type = PCH_DG1; - return; - } - - /* - * The reason to probe ISA bridge instead of Dev31:Fun0 is to - * make graphics device passthrough work easy for VMM, that only - * need to expose ISA bridge to let driver know the real hardware - * underneath. This is a requirement from virtualization team. - * - * In some virtualized environments (e.g. XEN), there is irrelevant - * ISA bridge in the system. To work reliably, we should scan through - * all the ISA bridge devices and check for the first match, instead - * of only checking the first one. - */ - while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) { - if (pch->vendor != PCI_VENDOR_ID_INTEL) - continue; - - id = pch->device & INTEL_PCH_DEVICE_ID_MASK; - - pch_type = intel_pch_type(dev_priv, id); - if (pch_type != PCH_NONE) { - dev_priv->pch_type = pch_type; - break; - } else if (intel_is_virt_pch(id, pch->subsystem_vendor, - pch->subsystem_device)) { - intel_virt_detect_pch(dev_priv, &id, &pch_type); - dev_priv->pch_type = pch_type; - break; - } - } - - /* - * Use PCH_NOP (PCH but no South Display) for PCH platforms without - * display. - */ - if (pch && !HAS_DISPLAY(dev_priv)) { - drm_dbg_kms(&dev_priv->drm, - "Display disabled, reverting to NOP PCH\n"); - dev_priv->pch_type = PCH_NOP; - } else if (!pch) { - if (i915_run_as_guest() && HAS_DISPLAY(dev_priv)) { - intel_virt_detect_pch(dev_priv, &id, &pch_type); - dev_priv->pch_type = pch_type; - } else { - drm_dbg_kms(&dev_priv->drm, "No PCH found.\n"); - } - } - - pci_dev_put(pch); -} diff --git a/drivers/gpu/drm/i915/soc/intel_pch.h b/drivers/gpu/drm/i915/soc/intel_pch.h deleted file mode 100644 index 635aea7a5539..000000000000 --- a/drivers/gpu/drm/i915/soc/intel_pch.h +++ /dev/null @@ -1,56 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -/* - * Copyright 2019 Intel Corporation. - */ - -#ifndef __INTEL_PCH__ -#define __INTEL_PCH__ - -struct drm_i915_private; - -/* - * Sorted by south display engine compatibility. - * If the new PCH comes with a south display engine that is not - * inherited from the latest item, please do not add it to the - * end. Instead, add it right after its "parent" PCH. - */ -enum intel_pch { - PCH_NOP = -1, /* PCH without south display */ - PCH_NONE = 0, /* No PCH present */ - PCH_IBX, /* Ibexpeak PCH */ - PCH_CPT, /* Cougarpoint/Pantherpoint PCH */ - PCH_LPT_H, /* Lynxpoint/Wildcatpoint H PCH */ - PCH_LPT_LP, /* Lynxpoint/Wildcatpoint LP PCH */ - PCH_SPT, /* Sunrisepoint/Kaby Lake PCH */ - PCH_CNP, /* Cannon/Comet Lake PCH */ - PCH_ICP, /* Ice Lake/Jasper Lake PCH */ - PCH_TGP, /* Tiger Lake/Mule Creek Canyon PCH */ - PCH_ADP, /* Alder Lake PCH */ - - /* Fake PCHs, functionality handled on the same PCI dev */ - PCH_DG1 = 1024, - PCH_DG2, - PCH_MTL, - PCH_LNL, -}; - -#define INTEL_PCH_TYPE(dev_priv) ((dev_priv)->pch_type) -#define HAS_PCH_DG2(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_DG2) -#define HAS_PCH_ADP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_ADP) -#define HAS_PCH_DG1(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_DG1) -#define HAS_PCH_TGP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_TGP) -#define HAS_PCH_ICP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_ICP) -#define HAS_PCH_CNP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_CNP) -#define HAS_PCH_SPT(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_SPT) -#define HAS_PCH_LPT_H(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_LPT_H) -#define HAS_PCH_LPT_LP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_LPT_LP) -#define HAS_PCH_LPT(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_LPT_H || \ - INTEL_PCH_TYPE(dev_priv) == PCH_LPT_LP) -#define HAS_PCH_CPT(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_CPT) -#define HAS_PCH_IBX(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_IBX) -#define HAS_PCH_NOP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_NOP) -#define HAS_PCH_SPLIT(dev_priv) (INTEL_PCH_TYPE(dev_priv) != PCH_NONE) - -void intel_detect_pch(struct drm_i915_private *dev_priv); - -#endif /* __INTEL_PCH__ */ diff --git a/drivers/gpu/drm/mediatek/Makefile b/drivers/gpu/drm/mediatek/Makefile index 32a2ed6c0cfe..43afd0a26d14 100644 --- a/drivers/gpu/drm/mediatek/Makefile +++ b/drivers/gpu/drm/mediatek/Makefile @@ -21,10 +21,8 @@ mediatek-drm-y := mtk_crtc.o \ obj-$(CONFIG_DRM_MEDIATEK) += mediatek-drm.o -mediatek-drm-hdmi-objs := mtk_cec.o \ - mtk_hdmi.o \ - mtk_hdmi_ddc.o - -obj-$(CONFIG_DRM_MEDIATEK_HDMI) += mediatek-drm-hdmi.o +obj-$(CONFIG_DRM_MEDIATEK_HDMI) += mtk_cec.o +obj-$(CONFIG_DRM_MEDIATEK_HDMI) += mtk_hdmi.o +obj-$(CONFIG_DRM_MEDIATEK_HDMI) += mtk_hdmi_ddc.o obj-$(CONFIG_DRM_MEDIATEK_DP) += mtk_dp.o diff --git a/drivers/gpu/drm/mediatek/mtk_cec.c b/drivers/gpu/drm/mediatek/mtk_cec.c index b42c0d87eba3..c7be530ca041 100644 --- a/drivers/gpu/drm/mediatek/mtk_cec.c +++ b/drivers/gpu/drm/mediatek/mtk_cec.c @@ -12,7 +12,6 @@ #include <linux/platform_device.h> #include "mtk_cec.h" -#include "mtk_hdmi.h" #include "mtk_drm_drv.h" #define TR_CONFIG 0x00 @@ -102,6 +101,7 @@ void mtk_cec_set_hpd_event(struct device *dev, cec->hpd_event = hpd_event; spin_unlock_irqrestore(&cec->lock, flags); } +EXPORT_SYMBOL_NS_GPL(mtk_cec_set_hpd_event, "DRM_MTK_HDMI_V1"); bool mtk_cec_hpd_high(struct device *dev) { @@ -112,6 +112,7 @@ bool mtk_cec_hpd_high(struct device *dev) return (status & (HDMI_PORD | HDMI_HTPLG)) == (HDMI_PORD | HDMI_HTPLG); } +EXPORT_SYMBOL_NS_GPL(mtk_cec_hpd_high, "DRM_MTK_HDMI_V1"); static void mtk_cec_htplg_irq_init(struct mtk_cec *cec) { @@ -247,3 +248,7 @@ struct platform_driver mtk_cec_driver = { .of_match_table = mtk_cec_of_ids, }, }; +module_platform_driver(mtk_cec_driver); + +MODULE_DESCRIPTION("MediaTek HDMI CEC Driver"); +MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c b/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c index fa0e95dd29a0..fe97bb97e004 100644 --- a/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c +++ b/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c @@ -492,11 +492,6 @@ static const struct of_device_id mtk_ovl_adaptor_comp_dt_ids[] = { { /* sentinel */ } }; -static int compare_of(struct device *dev, void *data) -{ - return dev->of_node == data; -} - static int ovl_adaptor_of_get_ddp_comp_type(struct device_node *node, enum mtk_ovl_adaptor_comp_type *ctype) { @@ -567,7 +562,7 @@ static int ovl_adaptor_comp_init(struct device *dev, struct component_match **ma priv->ovl_adaptor_comp[id] = &comp_pdev->dev; - drm_of_component_match_add(dev, match, compare_of, node); + drm_of_component_match_add(dev, match, component_compare_of, node); dev_dbg(dev, "Adding component match for %pOF\n", node); } diff --git a/drivers/gpu/drm/mediatek/mtk_dpi.c b/drivers/gpu/drm/mediatek/mtk_dpi.c index 9669bf4b80a1..8b58421bc39d 100644 --- a/drivers/gpu/drm/mediatek/mtk_dpi.c +++ b/drivers/gpu/drm/mediatek/mtk_dpi.c @@ -59,7 +59,8 @@ enum mtk_dpi_out_channel_swap { enum mtk_dpi_out_color_format { MTK_DPI_COLOR_FORMAT_RGB, - MTK_DPI_COLOR_FORMAT_YCBCR_422 + MTK_DPI_COLOR_FORMAT_YCBCR_422, + MTK_DPI_COLOR_FORMAT_YCBCR_444 }; struct mtk_dpi { @@ -450,9 +451,16 @@ static void mtk_dpi_config_disable_edge(struct mtk_dpi *dpi) static void mtk_dpi_config_color_format(struct mtk_dpi *dpi, enum mtk_dpi_out_color_format format) { - mtk_dpi_config_channel_swap(dpi, MTK_DPI_OUT_CHANNEL_SWAP_RGB); + mtk_dpi_config_channel_swap(dpi, dpi->channel_swap); - if (format == MTK_DPI_COLOR_FORMAT_YCBCR_422) { + switch (format) { + case MTK_DPI_COLOR_FORMAT_YCBCR_444: + mtk_dpi_config_yuv422_enable(dpi, false); + mtk_dpi_config_csc_enable(dpi, true); + if (dpi->conf->swap_input_support) + mtk_dpi_config_swap_input(dpi, false); + break; + case MTK_DPI_COLOR_FORMAT_YCBCR_422: mtk_dpi_config_yuv422_enable(dpi, true); mtk_dpi_config_csc_enable(dpi, true); @@ -463,11 +471,14 @@ static void mtk_dpi_config_color_format(struct mtk_dpi *dpi, mtk_dpi_mask(dpi, DPI_MATRIX_SET, dpi->mode.hdisplay <= 720 ? MATRIX_SEL_RGB_TO_BT601 : MATRIX_SEL_RGB_TO_JPEG, INT_MATRIX_SEL_MASK); - } else { + break; + default: + case MTK_DPI_COLOR_FORMAT_RGB: mtk_dpi_config_yuv422_enable(dpi, false); mtk_dpi_config_csc_enable(dpi, false); if (dpi->conf->swap_input_support) mtk_dpi_config_swap_input(dpi, false); + break; } } @@ -734,6 +745,65 @@ static u32 *mtk_dpi_bridge_atomic_get_input_bus_fmts(struct drm_bridge *bridge, return input_fmts; } +static unsigned int mtk_dpi_bus_fmt_bit_num(unsigned int out_bus_format) +{ + switch (out_bus_format) { + default: + case MEDIA_BUS_FMT_RGB888_1X24: + case MEDIA_BUS_FMT_BGR888_1X24: + case MEDIA_BUS_FMT_RGB888_2X12_LE: + case MEDIA_BUS_FMT_RGB888_2X12_BE: + case MEDIA_BUS_FMT_YUYV8_1X16: + case MEDIA_BUS_FMT_YUV8_1X24: + return MTK_DPI_OUT_BIT_NUM_8BITS; + case MEDIA_BUS_FMT_RGB101010_1X30: + case MEDIA_BUS_FMT_YUYV10_1X20: + case MEDIA_BUS_FMT_YUV10_1X30: + return MTK_DPI_OUT_BIT_NUM_10BITS; + case MEDIA_BUS_FMT_YUYV12_1X24: + return MTK_DPI_OUT_BIT_NUM_12BITS; + } +} + +static unsigned int mtk_dpi_bus_fmt_channel_swap(unsigned int out_bus_format) +{ + switch (out_bus_format) { + default: + case MEDIA_BUS_FMT_RGB888_1X24: + case MEDIA_BUS_FMT_RGB888_2X12_LE: + case MEDIA_BUS_FMT_RGB888_2X12_BE: + case MEDIA_BUS_FMT_RGB101010_1X30: + case MEDIA_BUS_FMT_YUYV8_1X16: + case MEDIA_BUS_FMT_YUYV10_1X20: + case MEDIA_BUS_FMT_YUYV12_1X24: + return MTK_DPI_OUT_CHANNEL_SWAP_RGB; + case MEDIA_BUS_FMT_BGR888_1X24: + case MEDIA_BUS_FMT_YUV8_1X24: + case MEDIA_BUS_FMT_YUV10_1X30: + return MTK_DPI_OUT_CHANNEL_SWAP_BGR; + } +} + +static unsigned int mtk_dpi_bus_fmt_color_format(unsigned int out_bus_format) +{ + switch (out_bus_format) { + default: + case MEDIA_BUS_FMT_RGB888_1X24: + case MEDIA_BUS_FMT_BGR888_1X24: + case MEDIA_BUS_FMT_RGB888_2X12_LE: + case MEDIA_BUS_FMT_RGB888_2X12_BE: + case MEDIA_BUS_FMT_RGB101010_1X30: + return MTK_DPI_COLOR_FORMAT_RGB; + case MEDIA_BUS_FMT_YUYV8_1X16: + case MEDIA_BUS_FMT_YUYV10_1X20: + case MEDIA_BUS_FMT_YUYV12_1X24: + return MTK_DPI_COLOR_FORMAT_YCBCR_422; + case MEDIA_BUS_FMT_YUV8_1X24: + case MEDIA_BUS_FMT_YUV10_1X30: + return MTK_DPI_COLOR_FORMAT_YCBCR_444; + } +} + static int mtk_dpi_bridge_atomic_check(struct drm_bridge *bridge, struct drm_bridge_state *bridge_state, struct drm_crtc_state *crtc_state, @@ -753,13 +823,10 @@ static int mtk_dpi_bridge_atomic_check(struct drm_bridge *bridge, bridge_state->output_bus_cfg.format); dpi->output_fmt = out_bus_format; - dpi->bit_num = MTK_DPI_OUT_BIT_NUM_8BITS; - dpi->channel_swap = MTK_DPI_OUT_CHANNEL_SWAP_RGB; + dpi->bit_num = mtk_dpi_bus_fmt_bit_num(out_bus_format); + dpi->channel_swap = mtk_dpi_bus_fmt_channel_swap(out_bus_format); dpi->yc_map = MTK_DPI_OUT_YC_MAP_RGB; - if (out_bus_format == MEDIA_BUS_FMT_YUYV8_1X16) - dpi->color_format = MTK_DPI_COLOR_FORMAT_YCBCR_422; - else - dpi->color_format = MTK_DPI_COLOR_FORMAT_RGB; + dpi->color_format = mtk_dpi_bus_fmt_color_format(out_bus_format); return 0; } @@ -1027,9 +1094,29 @@ static const u32 mt8183_output_fmts[] = { MEDIA_BUS_FMT_RGB888_2X12_BE, }; -static const u32 mt8195_output_fmts[] = { +static const u32 mt8195_dpi_output_fmts[] = { + MEDIA_BUS_FMT_BGR888_1X24, + MEDIA_BUS_FMT_RGB888_1X24, + MEDIA_BUS_FMT_RGB888_2X12_LE, + MEDIA_BUS_FMT_RGB888_2X12_BE, + MEDIA_BUS_FMT_RGB101010_1X30, + MEDIA_BUS_FMT_YUYV8_1X16, + MEDIA_BUS_FMT_YUYV10_1X20, + MEDIA_BUS_FMT_YUYV12_1X24, + MEDIA_BUS_FMT_YUV8_1X24, + MEDIA_BUS_FMT_YUV10_1X30, +}; + +static const u32 mt8195_dp_intf_output_fmts[] = { + MEDIA_BUS_FMT_BGR888_1X24, MEDIA_BUS_FMT_RGB888_1X24, + MEDIA_BUS_FMT_RGB888_2X12_LE, + MEDIA_BUS_FMT_RGB888_2X12_BE, + MEDIA_BUS_FMT_RGB101010_1X30, MEDIA_BUS_FMT_YUYV8_1X16, + MEDIA_BUS_FMT_YUYV10_1X20, + MEDIA_BUS_FMT_YUV8_1X24, + MEDIA_BUS_FMT_YUV10_1X30, }; static const struct mtk_dpi_factor dpi_factor_mt2701[] = { @@ -1142,8 +1229,8 @@ static const struct mtk_dpi_conf mt8192_conf = { static const struct mtk_dpi_conf mt8195_conf = { .max_clock_khz = 594000, - .output_fmts = mt8183_output_fmts, - .num_output_fmts = ARRAY_SIZE(mt8183_output_fmts), + .output_fmts = mt8195_dpi_output_fmts, + .num_output_fmts = ARRAY_SIZE(mt8195_dpi_output_fmts), .pixels_per_iter = 1, .is_ck_de_pol = true, .swap_input_support = true, @@ -1162,8 +1249,8 @@ static const struct mtk_dpi_conf mt8195_dpintf_conf = { .dpi_factor = dpi_factor_mt8195_dp_intf, .num_dpi_factor = ARRAY_SIZE(dpi_factor_mt8195_dp_intf), .max_clock_khz = 600000, - .output_fmts = mt8195_output_fmts, - .num_output_fmts = ARRAY_SIZE(mt8195_output_fmts), + .output_fmts = mt8195_dp_intf_output_fmts, + .num_output_fmts = ARRAY_SIZE(mt8195_dp_intf_output_fmts), .pixels_per_iter = 4, .dimension_mask = DPINTF_HPW_MASK, .hvsize_mask = DPINTF_HSIZE_MASK, diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c index 74158b9d6503..7c0c12dde488 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c @@ -470,7 +470,7 @@ static int mtk_drm_kms_init(struct drm_device *drm) ret = drmm_mode_config_init(drm); if (ret) - goto put_mutex_dev; + return ret; drm->mode_config.min_width = 64; drm->mode_config.min_height = 64; @@ -488,8 +488,11 @@ static int mtk_drm_kms_init(struct drm_device *drm) for (i = 0; i < private->data->mmsys_dev_num; i++) { drm->dev_private = private->all_drm_private[i]; ret = component_bind_all(private->all_drm_private[i]->dev, drm); - if (ret) - goto put_mutex_dev; + if (ret) { + while (--i >= 0) + component_unbind_all(private->all_drm_private[i]->dev, drm); + return ret; + } } /* @@ -582,9 +585,6 @@ static int mtk_drm_kms_init(struct drm_device *drm) err_component_unbind: for (i = 0; i < private->data->mmsys_dev_num; i++) component_unbind_all(private->all_drm_private[i]->dev, drm); -put_mutex_dev: - for (i = 0; i < private->data->mmsys_dev_num; i++) - put_device(private->all_drm_private[i]->mutex_dev); return ret; } @@ -655,8 +655,10 @@ static int mtk_drm_bind(struct device *dev) return 0; drm = drm_dev_alloc(&mtk_drm_driver, dev); - if (IS_ERR(drm)) - return PTR_ERR(drm); + if (IS_ERR(drm)) { + ret = PTR_ERR(drm); + goto err_put_dev; + } private->drm_master = true; drm->dev_private = private; @@ -682,18 +684,31 @@ err_free: drm_dev_put(drm); for (i = 0; i < private->data->mmsys_dev_num; i++) private->all_drm_private[i]->drm = NULL; +err_put_dev: + for (i = 0; i < private->data->mmsys_dev_num; i++) { + /* For device_find_child in mtk_drm_get_all_priv() */ + put_device(private->all_drm_private[i]->dev); + } + put_device(private->mutex_dev); return ret; } static void mtk_drm_unbind(struct device *dev) { struct mtk_drm_private *private = dev_get_drvdata(dev); + int i; /* for multi mmsys dev, unregister drm dev in mmsys master */ if (private->drm_master) { drm_dev_unregister(private->drm); mtk_drm_kms_deinit(private->drm); drm_dev_put(private->drm); + + for (i = 0; i < private->data->mmsys_dev_num; i++) { + /* For device_find_child in mtk_drm_get_all_priv() */ + put_device(private->all_drm_private[i]->dev); + } + put_device(private->mutex_dev); } private->mtk_drm_bound = false; private->drm_master = false; diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.c b/drivers/gpu/drm/mediatek/mtk_hdmi.c index 76aecad116ba..6943cdc77dec 100644 --- a/drivers/gpu/drm/mediatek/mtk_hdmi.c +++ b/drivers/gpu/drm/mediatek/mtk_hdmi.c @@ -31,7 +31,6 @@ #include <drm/drm_probe_helper.h> #include "mtk_cec.h" -#include "mtk_hdmi.h" #include "mtk_hdmi_regs.h" #define NCTS_BYTES 7 @@ -165,7 +164,7 @@ struct mtk_hdmi { bool dvi_mode; struct regmap *sys_regmap; unsigned int sys_offset; - void __iomem *regs; + struct regmap *regs; struct platform_device *audio_pdev; struct hdmi_audio_param aud_param; bool audio_enable; @@ -181,50 +180,10 @@ static inline struct mtk_hdmi *hdmi_ctx_from_bridge(struct drm_bridge *b) return container_of(b, struct mtk_hdmi, bridge); } -static u32 mtk_hdmi_read(struct mtk_hdmi *hdmi, u32 offset) -{ - return readl(hdmi->regs + offset); -} - -static void mtk_hdmi_write(struct mtk_hdmi *hdmi, u32 offset, u32 val) -{ - writel(val, hdmi->regs + offset); -} - -static void mtk_hdmi_clear_bits(struct mtk_hdmi *hdmi, u32 offset, u32 bits) -{ - void __iomem *reg = hdmi->regs + offset; - u32 tmp; - - tmp = readl(reg); - tmp &= ~bits; - writel(tmp, reg); -} - -static void mtk_hdmi_set_bits(struct mtk_hdmi *hdmi, u32 offset, u32 bits) -{ - void __iomem *reg = hdmi->regs + offset; - u32 tmp; - - tmp = readl(reg); - tmp |= bits; - writel(tmp, reg); -} - -static void mtk_hdmi_mask(struct mtk_hdmi *hdmi, u32 offset, u32 val, u32 mask) -{ - void __iomem *reg = hdmi->regs + offset; - u32 tmp; - - tmp = readl(reg); - tmp = (tmp & ~mask) | (val & mask); - writel(tmp, reg); -} - static void mtk_hdmi_hw_vid_black(struct mtk_hdmi *hdmi, bool black) { - mtk_hdmi_mask(hdmi, VIDEO_CFG_4, black ? GEN_RGB : NORMAL_PATH, - VIDEO_SOURCE_SEL); + regmap_update_bits(hdmi->regs, VIDEO_SOURCE_SEL, + VIDEO_CFG_4, black ? GEN_RGB : NORMAL_PATH); } static void mtk_hdmi_hw_make_reg_writable(struct mtk_hdmi *hdmi, bool enable) @@ -259,12 +218,12 @@ static void mtk_hdmi_hw_1p4_version_enable(struct mtk_hdmi *hdmi, bool enable) static void mtk_hdmi_hw_aud_mute(struct mtk_hdmi *hdmi) { - mtk_hdmi_set_bits(hdmi, GRL_AUDIO_CFG, AUDIO_ZERO); + regmap_set_bits(hdmi->regs, GRL_AUDIO_CFG, AUDIO_ZERO); } static void mtk_hdmi_hw_aud_unmute(struct mtk_hdmi *hdmi) { - mtk_hdmi_clear_bits(hdmi, GRL_AUDIO_CFG, AUDIO_ZERO); + regmap_clear_bits(hdmi->regs, GRL_AUDIO_CFG, AUDIO_ZERO); } static void mtk_hdmi_hw_reset(struct mtk_hdmi *hdmi) @@ -273,25 +232,25 @@ static void mtk_hdmi_hw_reset(struct mtk_hdmi *hdmi) HDMI_RST, HDMI_RST); regmap_update_bits(hdmi->sys_regmap, hdmi->sys_offset + HDMI_SYS_CFG1C, HDMI_RST, 0); - mtk_hdmi_clear_bits(hdmi, GRL_CFG3, CFG3_CONTROL_PACKET_DELAY); + regmap_clear_bits(hdmi->regs, GRL_CFG3, CFG3_CONTROL_PACKET_DELAY); regmap_update_bits(hdmi->sys_regmap, hdmi->sys_offset + HDMI_SYS_CFG1C, ANLG_ON, ANLG_ON); } static void mtk_hdmi_hw_enable_notice(struct mtk_hdmi *hdmi, bool enable_notice) { - mtk_hdmi_mask(hdmi, GRL_CFG2, enable_notice ? CFG2_NOTICE_EN : 0, - CFG2_NOTICE_EN); + regmap_update_bits(hdmi->regs, GRL_CFG2, CFG2_NOTICE_EN, + enable_notice ? CFG2_NOTICE_EN : 0); } static void mtk_hdmi_hw_write_int_mask(struct mtk_hdmi *hdmi, u32 int_mask) { - mtk_hdmi_write(hdmi, GRL_INT_MASK, int_mask); + regmap_write(hdmi->regs, GRL_INT_MASK, int_mask); } static void mtk_hdmi_hw_enable_dvi_mode(struct mtk_hdmi *hdmi, bool enable) { - mtk_hdmi_mask(hdmi, GRL_CFG1, enable ? CFG1_DVI : 0, CFG1_DVI); + regmap_update_bits(hdmi->regs, GRL_CFG1, CFG1_DVI, enable ? CFG1_DVI : 0); } static void mtk_hdmi_hw_send_info_frame(struct mtk_hdmi *hdmi, u8 *buffer, @@ -337,22 +296,22 @@ static void mtk_hdmi_hw_send_info_frame(struct mtk_hdmi *hdmi, u8 *buffer, dev_err(hdmi->dev, "Unknown infoframe type %d\n", frame_type); return; } - mtk_hdmi_clear_bits(hdmi, ctrl_reg, ctrl_frame_en); - mtk_hdmi_write(hdmi, GRL_INFOFRM_TYPE, frame_type); - mtk_hdmi_write(hdmi, GRL_INFOFRM_VER, frame_ver); - mtk_hdmi_write(hdmi, GRL_INFOFRM_LNG, frame_len); + regmap_clear_bits(hdmi->regs, ctrl_reg, ctrl_frame_en); + regmap_write(hdmi->regs, GRL_INFOFRM_TYPE, frame_type); + regmap_write(hdmi->regs, GRL_INFOFRM_VER, frame_ver); + regmap_write(hdmi->regs, GRL_INFOFRM_LNG, frame_len); - mtk_hdmi_write(hdmi, GRL_IFM_PORT, checksum); + regmap_write(hdmi->regs, GRL_IFM_PORT, checksum); for (i = 0; i < frame_len; i++) - mtk_hdmi_write(hdmi, GRL_IFM_PORT, frame_data[i]); + regmap_write(hdmi->regs, GRL_IFM_PORT, frame_data[i]); - mtk_hdmi_set_bits(hdmi, ctrl_reg, ctrl_frame_en); + regmap_set_bits(hdmi->regs, ctrl_reg, ctrl_frame_en); } static void mtk_hdmi_hw_send_aud_packet(struct mtk_hdmi *hdmi, bool enable) { - mtk_hdmi_mask(hdmi, GRL_SHIFT_R2, enable ? 0 : AUDIO_PACKET_OFF, - AUDIO_PACKET_OFF); + regmap_update_bits(hdmi->regs, AUDIO_PACKET_OFF, + GRL_SHIFT_R2, enable ? 0 : AUDIO_PACKET_OFF); } static void mtk_hdmi_hw_config_sys(struct mtk_hdmi *hdmi) @@ -373,44 +332,44 @@ static void mtk_hdmi_hw_set_deep_color_mode(struct mtk_hdmi *hdmi) static void mtk_hdmi_hw_send_av_mute(struct mtk_hdmi *hdmi) { - mtk_hdmi_clear_bits(hdmi, GRL_CFG4, CTRL_AVMUTE); + regmap_clear_bits(hdmi->regs, GRL_CFG4, CTRL_AVMUTE); usleep_range(2000, 4000); - mtk_hdmi_set_bits(hdmi, GRL_CFG4, CTRL_AVMUTE); + regmap_set_bits(hdmi->regs, GRL_CFG4, CTRL_AVMUTE); } static void mtk_hdmi_hw_send_av_unmute(struct mtk_hdmi *hdmi) { - mtk_hdmi_mask(hdmi, GRL_CFG4, CFG4_AV_UNMUTE_EN, - CFG4_AV_UNMUTE_EN | CFG4_AV_UNMUTE_SET); + regmap_update_bits(hdmi->regs, GRL_CFG4, CFG4_AV_UNMUTE_EN | CFG4_AV_UNMUTE_SET, + CFG4_AV_UNMUTE_EN); usleep_range(2000, 4000); - mtk_hdmi_mask(hdmi, GRL_CFG4, CFG4_AV_UNMUTE_SET, - CFG4_AV_UNMUTE_EN | CFG4_AV_UNMUTE_SET); + regmap_update_bits(hdmi->regs, GRL_CFG4, CFG4_AV_UNMUTE_EN | CFG4_AV_UNMUTE_SET, + CFG4_AV_UNMUTE_SET); } static void mtk_hdmi_hw_ncts_enable(struct mtk_hdmi *hdmi, bool on) { - mtk_hdmi_mask(hdmi, GRL_CTS_CTRL, on ? 0 : CTS_CTRL_SOFT, - CTS_CTRL_SOFT); + regmap_update_bits(hdmi->regs, GRL_CTS_CTRL, CTS_CTRL_SOFT, + on ? 0 : CTS_CTRL_SOFT); } static void mtk_hdmi_hw_ncts_auto_write_enable(struct mtk_hdmi *hdmi, bool enable) { - mtk_hdmi_mask(hdmi, GRL_CTS_CTRL, enable ? NCTS_WRI_ANYTIME : 0, - NCTS_WRI_ANYTIME); + regmap_update_bits(hdmi->regs, GRL_CTS_CTRL, NCTS_WRI_ANYTIME, + enable ? NCTS_WRI_ANYTIME : 0); } static void mtk_hdmi_hw_msic_setting(struct mtk_hdmi *hdmi, struct drm_display_mode *mode) { - mtk_hdmi_clear_bits(hdmi, GRL_CFG4, CFG4_MHL_MODE); + regmap_clear_bits(hdmi->regs, GRL_CFG4, CFG4_MHL_MODE); if (mode->flags & DRM_MODE_FLAG_INTERLACE && mode->clock == 74250 && mode->vdisplay == 1080) - mtk_hdmi_clear_bits(hdmi, GRL_CFG2, CFG2_MHL_DE_SEL); + regmap_clear_bits(hdmi->regs, GRL_CFG2, CFG2_MHL_DE_SEL); else - mtk_hdmi_set_bits(hdmi, GRL_CFG2, CFG2_MHL_DE_SEL); + regmap_set_bits(hdmi->regs, GRL_CFG2, CFG2_MHL_DE_SEL); } static void mtk_hdmi_hw_aud_set_channel_swap(struct mtk_hdmi *hdmi, @@ -438,7 +397,7 @@ static void mtk_hdmi_hw_aud_set_channel_swap(struct mtk_hdmi *hdmi, swap_bit = LFE_CC_SWAP; break; } - mtk_hdmi_mask(hdmi, GRL_CH_SWAP, swap_bit, 0xff); + regmap_update_bits(hdmi->regs, GRL_CH_SWAP, 0xff, swap_bit); } static void mtk_hdmi_hw_aud_set_bit_num(struct mtk_hdmi *hdmi, @@ -459,7 +418,7 @@ static void mtk_hdmi_hw_aud_set_bit_num(struct mtk_hdmi *hdmi, break; } - mtk_hdmi_mask(hdmi, GRL_AOUT_CFG, val, AOUT_BNUM_SEL_MASK); + regmap_update_bits(hdmi->regs, GRL_AOUT_CFG, AOUT_BNUM_SEL_MASK, val); } static void mtk_hdmi_hw_aud_set_i2s_fmt(struct mtk_hdmi *hdmi, @@ -467,7 +426,7 @@ static void mtk_hdmi_hw_aud_set_i2s_fmt(struct mtk_hdmi *hdmi, { u32 val; - val = mtk_hdmi_read(hdmi, GRL_CFG0); + regmap_read(hdmi->regs, GRL_CFG0, &val); val &= ~(CFG0_W_LENGTH_MASK | CFG0_I2S_MODE_MASK); switch (i2s_fmt) { @@ -491,7 +450,7 @@ static void mtk_hdmi_hw_aud_set_i2s_fmt(struct mtk_hdmi *hdmi, val |= CFG0_I2S_MODE_I2S | CFG0_W_LENGTH_16BIT; break; } - mtk_hdmi_write(hdmi, GRL_CFG0, val); + regmap_write(hdmi->regs, GRL_CFG0, val); } static void mtk_hdmi_hw_audio_config(struct mtk_hdmi *hdmi, bool dst) @@ -500,14 +459,14 @@ static void mtk_hdmi_hw_audio_config(struct mtk_hdmi *hdmi, bool dst) u8 val; /* Disable high bitrate, set DST packet normal/double */ - mtk_hdmi_clear_bits(hdmi, GRL_AOUT_CFG, HIGH_BIT_RATE_PACKET_ALIGN); + regmap_clear_bits(hdmi->regs, GRL_AOUT_CFG, HIGH_BIT_RATE_PACKET_ALIGN); if (dst) val = DST_NORMAL_DOUBLE | SACD_DST; else val = 0; - mtk_hdmi_mask(hdmi, GRL_AUDIO_CFG, val, mask); + regmap_update_bits(hdmi->regs, GRL_AUDIO_CFG, mask, val); } static void mtk_hdmi_hw_aud_set_i2s_chan_num(struct mtk_hdmi *hdmi, @@ -548,10 +507,10 @@ static void mtk_hdmi_hw_aud_set_i2s_chan_num(struct mtk_hdmi *hdmi, i2s_uv = I2S_UV_CH_EN(0); } - mtk_hdmi_write(hdmi, GRL_CH_SW0, ch_switch & 0xff); - mtk_hdmi_write(hdmi, GRL_CH_SW1, (ch_switch >> 8) & 0xff); - mtk_hdmi_write(hdmi, GRL_CH_SW2, (ch_switch >> 16) & 0xff); - mtk_hdmi_write(hdmi, GRL_I2S_UV, i2s_uv); + regmap_write(hdmi->regs, GRL_CH_SW0, ch_switch & 0xff); + regmap_write(hdmi->regs, GRL_CH_SW1, (ch_switch >> 8) & 0xff); + regmap_write(hdmi->regs, GRL_CH_SW2, (ch_switch >> 16) & 0xff); + regmap_write(hdmi->regs, GRL_I2S_UV, i2s_uv); } static void mtk_hdmi_hw_aud_set_input_type(struct mtk_hdmi *hdmi, @@ -559,7 +518,7 @@ static void mtk_hdmi_hw_aud_set_input_type(struct mtk_hdmi *hdmi, { u32 val; - val = mtk_hdmi_read(hdmi, GRL_CFG1); + regmap_read(hdmi->regs, GRL_CFG1, &val); if (input_type == HDMI_AUD_INPUT_I2S && (val & CFG1_SPDIF) == CFG1_SPDIF) { val &= ~CFG1_SPDIF; @@ -567,7 +526,7 @@ static void mtk_hdmi_hw_aud_set_input_type(struct mtk_hdmi *hdmi, (val & CFG1_SPDIF) == 0) { val |= CFG1_SPDIF; } - mtk_hdmi_write(hdmi, GRL_CFG1, val); + regmap_write(hdmi->regs, GRL_CFG1, val); } static void mtk_hdmi_hw_aud_set_channel_status(struct mtk_hdmi *hdmi, @@ -576,13 +535,13 @@ static void mtk_hdmi_hw_aud_set_channel_status(struct mtk_hdmi *hdmi, int i; for (i = 0; i < 5; i++) { - mtk_hdmi_write(hdmi, GRL_I2S_C_STA0 + i * 4, channel_status[i]); - mtk_hdmi_write(hdmi, GRL_L_STATUS_0 + i * 4, channel_status[i]); - mtk_hdmi_write(hdmi, GRL_R_STATUS_0 + i * 4, channel_status[i]); + regmap_write(hdmi->regs, GRL_I2S_C_STA0 + i * 4, channel_status[i]); + regmap_write(hdmi->regs, GRL_L_STATUS_0 + i * 4, channel_status[i]); + regmap_write(hdmi->regs, GRL_R_STATUS_0 + i * 4, channel_status[i]); } for (; i < 24; i++) { - mtk_hdmi_write(hdmi, GRL_L_STATUS_0 + i * 4, 0); - mtk_hdmi_write(hdmi, GRL_R_STATUS_0 + i * 4, 0); + regmap_write(hdmi->regs, GRL_L_STATUS_0 + i * 4, 0); + regmap_write(hdmi->regs, GRL_R_STATUS_0 + i * 4, 0); } } @@ -590,13 +549,13 @@ static void mtk_hdmi_hw_aud_src_reenable(struct mtk_hdmi *hdmi) { u32 val; - val = mtk_hdmi_read(hdmi, GRL_MIX_CTRL); + regmap_read(hdmi->regs, GRL_MIX_CTRL, &val); if (val & MIX_CTRL_SRC_EN) { val &= ~MIX_CTRL_SRC_EN; - mtk_hdmi_write(hdmi, GRL_MIX_CTRL, val); + regmap_write(hdmi->regs, GRL_MIX_CTRL, val); usleep_range(255, 512); val |= MIX_CTRL_SRC_EN; - mtk_hdmi_write(hdmi, GRL_MIX_CTRL, val); + regmap_write(hdmi->regs, GRL_MIX_CTRL, val); } } @@ -604,10 +563,10 @@ static void mtk_hdmi_hw_aud_src_disable(struct mtk_hdmi *hdmi) { u32 val; - val = mtk_hdmi_read(hdmi, GRL_MIX_CTRL); + regmap_read(hdmi->regs, GRL_MIX_CTRL, &val); val &= ~MIX_CTRL_SRC_EN; - mtk_hdmi_write(hdmi, GRL_MIX_CTRL, val); - mtk_hdmi_write(hdmi, GRL_SHIFT_L1, 0x00); + regmap_write(hdmi->regs, GRL_MIX_CTRL, val); + regmap_write(hdmi->regs, GRL_SHIFT_L1, 0x00); } static void mtk_hdmi_hw_aud_set_mclk(struct mtk_hdmi *hdmi, @@ -615,7 +574,7 @@ static void mtk_hdmi_hw_aud_set_mclk(struct mtk_hdmi *hdmi, { u32 val; - val = mtk_hdmi_read(hdmi, GRL_CFG5); + regmap_read(hdmi->regs, GRL_CFG5, &val); val &= CFG5_CD_RATIO_MASK; switch (mclk) { @@ -638,7 +597,7 @@ static void mtk_hdmi_hw_aud_set_mclk(struct mtk_hdmi *hdmi, val |= CFG5_FS256; break; } - mtk_hdmi_write(hdmi, GRL_CFG5, val); + regmap_write(hdmi->regs, GRL_CFG5, val); } struct hdmi_acr_n { @@ -716,15 +675,22 @@ static unsigned int hdmi_expected_cts(unsigned int audio_sample_rate, 128 * audio_sample_rate); } +static void mtk_hdmi_get_ncts(unsigned int sample_rate, unsigned int clock, + unsigned int *n, unsigned int *cts) +{ + *n = hdmi_recommended_n(sample_rate, clock); + *cts = hdmi_expected_cts(sample_rate, clock, *n); +} + static void do_hdmi_hw_aud_set_ncts(struct mtk_hdmi *hdmi, unsigned int n, unsigned int cts) { unsigned char val[NCTS_BYTES]; int i; - mtk_hdmi_write(hdmi, GRL_NCTS, 0); - mtk_hdmi_write(hdmi, GRL_NCTS, 0); - mtk_hdmi_write(hdmi, GRL_NCTS, 0); + regmap_write(hdmi->regs, GRL_NCTS, 0); + regmap_write(hdmi->regs, GRL_NCTS, 0); + regmap_write(hdmi->regs, GRL_NCTS, 0); memset(val, 0, sizeof(val)); val[0] = (cts >> 24) & 0xff; @@ -737,7 +703,7 @@ static void do_hdmi_hw_aud_set_ncts(struct mtk_hdmi *hdmi, unsigned int n, val[6] = n & 0xff; for (i = 0; i < NCTS_BYTES; i++) - mtk_hdmi_write(hdmi, GRL_NCTS, val[i]); + regmap_write(hdmi->regs, GRL_NCTS, val[i]); } static void mtk_hdmi_hw_aud_set_ncts(struct mtk_hdmi *hdmi, @@ -746,14 +712,12 @@ static void mtk_hdmi_hw_aud_set_ncts(struct mtk_hdmi *hdmi, { unsigned int n, cts; - n = hdmi_recommended_n(sample_rate, clock); - cts = hdmi_expected_cts(sample_rate, clock, n); + mtk_hdmi_get_ncts(sample_rate, clock, &n, &cts); dev_dbg(hdmi->dev, "%s: sample_rate=%u, clock=%d, cts=%u, n=%u\n", __func__, sample_rate, clock, n, cts); - mtk_hdmi_mask(hdmi, DUMMY_304, AUDIO_I2S_NCTS_SEL_64, - AUDIO_I2S_NCTS_SEL); + regmap_update_bits(hdmi->regs, DUMMY_304, AUDIO_I2S_NCTS_SEL, AUDIO_I2S_NCTS_SEL_64); do_hdmi_hw_aud_set_ncts(hdmi, n, cts); } @@ -873,7 +837,7 @@ static void mtk_hdmi_aud_set_input(struct mtk_hdmi *hdmi) bool dst; mtk_hdmi_hw_aud_set_channel_swap(hdmi, HDMI_AUD_SWAP_LFE_CC); - mtk_hdmi_set_bits(hdmi, GRL_MIX_CTRL, MIX_CTRL_FLAT); + regmap_set_bits(hdmi->regs, GRL_MIX_CTRL, MIX_CTRL_FLAT); if (hdmi->aud_param.aud_input_type == HDMI_AUD_INPUT_SPDIF && hdmi->aud_param.aud_codec == HDMI_AUDIO_CODING_TYPE_DST) { @@ -905,7 +869,7 @@ static int mtk_hdmi_aud_set_src(struct mtk_hdmi *hdmi, mtk_hdmi_hw_ncts_enable(hdmi, false); mtk_hdmi_hw_aud_src_disable(hdmi); - mtk_hdmi_clear_bits(hdmi, GRL_CFG2, CFG2_ACLK_INV); + regmap_clear_bits(hdmi->regs, GRL_CFG2, CFG2_ACLK_INV); if (hdmi->aud_param.aud_input_type == HDMI_AUD_INPUT_I2S) { switch (sample_rate) { @@ -1061,20 +1025,6 @@ static int mtk_hdmi_setup_vendor_specific_infoframe(struct mtk_hdmi *hdmi, return 0; } -static int mtk_hdmi_output_init(struct mtk_hdmi *hdmi) -{ - struct hdmi_audio_param *aud_param = &hdmi->aud_param; - - aud_param->aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; - aud_param->aud_sample_size = HDMI_AUDIO_SAMPLE_SIZE_16; - aud_param->aud_input_type = HDMI_AUD_INPUT_I2S; - aud_param->aud_i2s_fmt = HDMI_I2S_MODE_I2S_24BIT; - aud_param->aud_mclk = HDMI_AUD_MCLK_128FS; - aud_param->aud_input_chan_type = HDMI_AUD_CHAN_TYPE_2_0; - - return 0; -} - static void mtk_hdmi_audio_enable(struct mtk_hdmi *hdmi) { mtk_hdmi_hw_send_aud_packet(hdmi, true); @@ -1087,20 +1037,6 @@ static void mtk_hdmi_audio_disable(struct mtk_hdmi *hdmi) hdmi->audio_enable = false; } -static int mtk_hdmi_audio_set_param(struct mtk_hdmi *hdmi, - struct hdmi_audio_param *param) -{ - if (!hdmi->audio_enable) { - dev_err(hdmi->dev, "hdmi audio is in disable state!\n"); - return -EINVAL; - } - dev_dbg(hdmi->dev, "codec:%d, input:%d, channel:%d, fs:%d\n", - param->aud_codec, param->aud_input_type, - param->aud_input_chan_type, param->codec_params.sample_rate); - memcpy(&hdmi->aud_param, param, sizeof(*param)); - return mtk_hdmi_aud_output_config(hdmi, &hdmi->mode); -} - static int mtk_hdmi_output_set_display_mode(struct mtk_hdmi *hdmi, struct drm_display_mode *mode) { @@ -1408,30 +1344,20 @@ static const struct drm_bridge_funcs mtk_hdmi_bridge_funcs = { .edid_read = mtk_hdmi_bridge_edid_read, }; -static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi, - struct platform_device *pdev) +static int mtk_hdmi_get_cec_dev(struct mtk_hdmi *hdmi, struct device *dev, struct device_node *np) { - struct device *dev = &pdev->dev; - struct device_node *np = dev->of_node; - struct device_node *cec_np, *remote, *i2c_np; struct platform_device *cec_pdev; - struct regmap *regmap; + struct device_node *cec_np; int ret; ret = mtk_hdmi_get_all_clk(hdmi, np); - if (ret) { - if (ret != -EPROBE_DEFER) - dev_err(dev, "Failed to get clocks: %d\n", ret); - - return ret; - } + if (ret) + return dev_err_probe(dev, ret, "Failed to get clocks\n"); /* The CEC module handles HDMI hotplug detection */ cec_np = of_get_compatible_child(np->parent, "mediatek,mt8173-cec"); - if (!cec_np) { - dev_err(dev, "Failed to find CEC node\n"); - return -EINVAL; - } + if (!cec_np) + return dev_err_probe(dev, -EINVAL, "Failed to find CEC node\n"); cec_pdev = of_find_device_by_node(cec_np); if (!cec_pdev) { @@ -1441,82 +1367,77 @@ static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi, return -EPROBE_DEFER; } of_node_put(cec_np); - hdmi->cec_dev = &cec_pdev->dev; /* * The mediatek,syscon-hdmi property contains a phandle link to the * MMSYS_CONFIG device and the register offset of the HDMI_SYS_CFG * registers it contains. */ - regmap = syscon_regmap_lookup_by_phandle(np, "mediatek,syscon-hdmi"); - ret = of_property_read_u32_index(np, "mediatek,syscon-hdmi", 1, - &hdmi->sys_offset); - if (IS_ERR(regmap)) - ret = PTR_ERR(regmap); - if (ret) { - dev_err(dev, - "Failed to get system configuration registers: %d\n", - ret); - goto put_device; - } - hdmi->sys_regmap = regmap; + hdmi->sys_regmap = syscon_regmap_lookup_by_phandle_args(np, "mediatek,syscon-hdmi", + 1, &hdmi->sys_offset); + if (IS_ERR(hdmi->sys_regmap)) + return dev_err_probe(dev, PTR_ERR(hdmi->sys_regmap), + "Failed to get system configuration registers\n"); - hdmi->regs = devm_platform_ioremap_resource(pdev, 0); - if (IS_ERR(hdmi->regs)) { - ret = PTR_ERR(hdmi->regs); - goto put_device; - } + hdmi->cec_dev = &cec_pdev->dev; + return 0; +} + +static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi, + struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct device_node *np = dev->of_node; + struct device_node *remote, *i2c_np; + int ret; + + ret = mtk_hdmi_get_all_clk(hdmi, np); + if (ret) + return dev_err_probe(dev, ret, "Failed to get clocks\n"); + + hdmi->regs = device_node_to_regmap(dev->of_node); + if (IS_ERR(hdmi->regs)) + return PTR_ERR(hdmi->regs); remote = of_graph_get_remote_node(np, 1, 0); - if (!remote) { - ret = -EINVAL; - goto put_device; - } + if (!remote) + return -EINVAL; if (!of_device_is_compatible(remote, "hdmi-connector")) { hdmi->next_bridge = of_drm_find_bridge(remote); if (!hdmi->next_bridge) { dev_err(dev, "Waiting for external bridge\n"); of_node_put(remote); - ret = -EPROBE_DEFER; - goto put_device; + return -EPROBE_DEFER; } } i2c_np = of_parse_phandle(remote, "ddc-i2c-bus", 0); - if (!i2c_np) { - dev_err(dev, "Failed to find ddc-i2c-bus node in %pOF\n", - remote); - of_node_put(remote); - ret = -EINVAL; - goto put_device; - } of_node_put(remote); + if (!i2c_np) + return dev_err_probe(dev, -EINVAL, "No ddc-i2c-bus in connector\n"); hdmi->ddc_adpt = of_find_i2c_adapter_by_node(i2c_np); of_node_put(i2c_np); - if (!hdmi->ddc_adpt) { - dev_err(dev, "Failed to get ddc i2c adapter by node\n"); - ret = -EINVAL; - goto put_device; - } + if (!hdmi->ddc_adpt) + return dev_err_probe(dev, -EINVAL, "Failed to get ddc i2c adapter by node\n"); + + ret = mtk_hdmi_get_cec_dev(hdmi, dev, np); + if (ret) + return ret; return 0; -put_device: - put_device(hdmi->cec_dev); - return ret; } /* * HDMI audio codec callbacks */ -static int mtk_hdmi_audio_hw_params(struct device *dev, void *data, - struct hdmi_codec_daifmt *daifmt, - struct hdmi_codec_params *params) +static int mtk_hdmi_audio_params(struct mtk_hdmi *hdmi, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params) { - struct mtk_hdmi *hdmi = dev_get_drvdata(dev); - struct hdmi_audio_param hdmi_params; + struct hdmi_audio_param aud_params = { 0 }; unsigned int chan = params->cea.channels; dev_dbg(hdmi->dev, "%s: %u Hz, %d bit, %d channels\n", __func__, @@ -1527,16 +1448,16 @@ static int mtk_hdmi_audio_hw_params(struct device *dev, void *data, switch (chan) { case 2: - hdmi_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_2_0; + aud_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_2_0; break; case 4: - hdmi_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_4_0; + aud_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_4_0; break; case 6: - hdmi_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_5_1; + aud_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_5_1; break; case 8: - hdmi_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_7_1; + aud_params.aud_input_chan_type = HDMI_AUD_CHAN_TYPE_7_1; break; default: dev_err(hdmi->dev, "channel[%d] not supported!\n", chan); @@ -1560,27 +1481,45 @@ static int mtk_hdmi_audio_hw_params(struct device *dev, void *data, switch (daifmt->fmt) { case HDMI_I2S: - hdmi_params.aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; - hdmi_params.aud_sample_size = HDMI_AUDIO_SAMPLE_SIZE_16; - hdmi_params.aud_input_type = HDMI_AUD_INPUT_I2S; - hdmi_params.aud_i2s_fmt = HDMI_I2S_MODE_I2S_24BIT; - hdmi_params.aud_mclk = HDMI_AUD_MCLK_128FS; + aud_params.aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; + aud_params.aud_sample_size = HDMI_AUDIO_SAMPLE_SIZE_16; + aud_params.aud_input_type = HDMI_AUD_INPUT_I2S; + aud_params.aud_i2s_fmt = HDMI_I2S_MODE_I2S_24BIT; + aud_params.aud_mclk = HDMI_AUD_MCLK_128FS; break; case HDMI_SPDIF: - hdmi_params.aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; - hdmi_params.aud_sample_size = HDMI_AUDIO_SAMPLE_SIZE_16; - hdmi_params.aud_input_type = HDMI_AUD_INPUT_SPDIF; + aud_params.aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; + aud_params.aud_sample_size = HDMI_AUDIO_SAMPLE_SIZE_16; + aud_params.aud_input_type = HDMI_AUD_INPUT_SPDIF; break; default: dev_err(hdmi->dev, "%s: Invalid DAI format %d\n", __func__, daifmt->fmt); return -EINVAL; } + memcpy(&aud_params.codec_params, params, sizeof(aud_params.codec_params)); + memcpy(&hdmi->aud_param, &aud_params, sizeof(aud_params)); + + dev_dbg(hdmi->dev, "codec:%d, input:%d, channel:%d, fs:%d\n", + aud_params.aud_codec, aud_params.aud_input_type, + aud_params.aud_input_chan_type, aud_params.codec_params.sample_rate); - memcpy(&hdmi_params.codec_params, params, - sizeof(hdmi_params.codec_params)); + return 0; +} + +static int mtk_hdmi_audio_hw_params(struct device *dev, void *data, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params) +{ + struct mtk_hdmi *hdmi = dev_get_drvdata(dev); - mtk_hdmi_audio_set_param(hdmi, &hdmi_params); + if (!hdmi->audio_enable) { + dev_err(hdmi->dev, "hdmi audio is in disable state!\n"); + return -EINVAL; + } + + mtk_hdmi_audio_params(hdmi, daifmt, params); + mtk_hdmi_aud_output_config(hdmi, &hdmi->mode); return 0; } @@ -1626,17 +1565,22 @@ static int mtk_hdmi_audio_get_eld(struct device *dev, void *data, uint8_t *buf, return 0; } -static int mtk_hdmi_audio_hook_plugged_cb(struct device *dev, void *data, - hdmi_codec_plugged_cb fn, +static void mtk_hdmi_audio_set_plugged_cb(struct mtk_hdmi *hdmi, hdmi_codec_plugged_cb fn, struct device *codec_dev) { - struct mtk_hdmi *hdmi = data; - mutex_lock(&hdmi->update_plugged_status_lock); hdmi->plugged_cb = fn; hdmi->codec_dev = codec_dev; mutex_unlock(&hdmi->update_plugged_status_lock); +} + +static int mtk_hdmi_audio_hook_plugged_cb(struct device *dev, void *data, + hdmi_codec_plugged_cb fn, + struct device *codec_dev) +{ + struct mtk_hdmi *hdmi = data; + mtk_hdmi_audio_set_plugged_cb(hdmi, fn, codec_dev); mtk_hdmi_update_plugged_status(hdmi); return 0; @@ -1659,6 +1603,7 @@ static void mtk_hdmi_unregister_audio_driver(void *data) static int mtk_hdmi_register_audio_driver(struct device *dev) { struct mtk_hdmi *hdmi = dev_get_drvdata(dev); + struct hdmi_audio_param *aud_param = &hdmi->aud_param; struct hdmi_codec_pdata codec_data = { .ops = &mtk_hdmi_audio_codec_ops, .max_i2s_channels = 2, @@ -1668,6 +1613,13 @@ static int mtk_hdmi_register_audio_driver(struct device *dev) }; int ret; + aud_param->aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; + aud_param->aud_sample_size = HDMI_AUDIO_SAMPLE_SIZE_16; + aud_param->aud_input_type = HDMI_AUD_INPUT_I2S; + aud_param->aud_i2s_fmt = HDMI_I2S_MODE_I2S_24BIT; + aud_param->aud_mclk = HDMI_AUD_MCLK_128FS; + aud_param->aud_input_chan_type = HDMI_AUD_CHAN_TYPE_2_0; + hdmi->audio_pdev = platform_device_register_data(dev, HDMI_CODEC_DRV_NAME, PLATFORM_DEVID_AUTO, @@ -1710,11 +1662,6 @@ static int mtk_hdmi_probe(struct platform_device *pdev) mutex_init(&hdmi->update_plugged_status_lock); platform_set_drvdata(pdev, hdmi); - ret = mtk_hdmi_output_init(hdmi); - if (ret) - return dev_err_probe(dev, ret, - "Failed to initialize hdmi output\n"); - ret = mtk_hdmi_register_audio_driver(dev); if (ret) return dev_err_probe(dev, ret, @@ -1790,28 +1737,9 @@ static struct platform_driver mtk_hdmi_driver = { .pm = &mtk_hdmi_pm_ops, }, }; - -static struct platform_driver * const mtk_hdmi_drivers[] = { - &mtk_hdmi_ddc_driver, - &mtk_cec_driver, - &mtk_hdmi_driver, -}; - -static int __init mtk_hdmitx_init(void) -{ - return platform_register_drivers(mtk_hdmi_drivers, - ARRAY_SIZE(mtk_hdmi_drivers)); -} - -static void __exit mtk_hdmitx_exit(void) -{ - platform_unregister_drivers(mtk_hdmi_drivers, - ARRAY_SIZE(mtk_hdmi_drivers)); -} - -module_init(mtk_hdmitx_init); -module_exit(mtk_hdmitx_exit); +module_platform_driver(mtk_hdmi_driver); MODULE_AUTHOR("Jie Qiu <jie.qiu@mediatek.com>"); MODULE_DESCRIPTION("MediaTek HDMI Driver"); MODULE_LICENSE("GPL v2"); +MODULE_IMPORT_NS("DRM_MTK_HDMI_V1"); diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.h b/drivers/gpu/drm/mediatek/mtk_hdmi.h deleted file mode 100644 index 472bf141c92b..000000000000 --- a/drivers/gpu/drm/mediatek/mtk_hdmi.h +++ /dev/null @@ -1,14 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0-only */ -/* - * Copyright (c) 2014 MediaTek Inc. - * Author: Jie Qiu <jie.qiu@mediatek.com> - */ -#ifndef _MTK_HDMI_CTRL_H -#define _MTK_HDMI_CTRL_H - -struct platform_driver; - -extern struct platform_driver mtk_cec_driver; -extern struct platform_driver mtk_hdmi_ddc_driver; - -#endif /* _MTK_HDMI_CTRL_H */ diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi_ddc.c b/drivers/gpu/drm/mediatek/mtk_hdmi_ddc.c index 07db68067844..6358e1af69b4 100644 --- a/drivers/gpu/drm/mediatek/mtk_hdmi_ddc.c +++ b/drivers/gpu/drm/mediatek/mtk_hdmi_ddc.c @@ -20,7 +20,6 @@ #include <linux/of_platform.h> #include "mtk_drm_drv.h" -#include "mtk_hdmi.h" #define SIF1_CLOK (288) #define DDC_DDCMCTL0 (0x0) @@ -337,6 +336,7 @@ struct platform_driver mtk_hdmi_ddc_driver = { .of_match_table = mtk_hdmi_ddc_match, }, }; +module_platform_driver(mtk_hdmi_ddc_driver); MODULE_AUTHOR("Jie Qiu <jie.qiu@mediatek.com>"); MODULE_DESCRIPTION("MediaTek HDMI DDC Driver"); diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig index 974bc7c0ea76..7f127e2ae442 100644 --- a/drivers/gpu/drm/msm/Kconfig +++ b/drivers/gpu/drm/msm/Kconfig @@ -104,6 +104,7 @@ config DRM_MSM_DPU config DRM_MSM_DP bool "Enable DisplayPort support in MSM DRM driver" depends on DRM_MSM + select DRM_DISPLAY_HDMI_AUDIO_HELPER select RATIONAL default y help diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile index 5df20cbeafb8..7a2ada6e2d74 100644 --- a/drivers/gpu/drm/msm/Makefile +++ b/drivers/gpu/drm/msm/Makefile @@ -48,7 +48,6 @@ msm-display-$(CONFIG_DRM_MSM_MDP4) += \ disp/mdp4/mdp4_dsi_encoder.o \ disp/mdp4/mdp4_dtv_encoder.o \ disp/mdp4/mdp4_lcdc_encoder.o \ - disp/mdp4/mdp4_lvds_connector.o \ disp/mdp4/mdp4_lvds_pll.o \ disp/mdp4/mdp4_irq.o \ disp/mdp4/mdp4_kms.o \ diff --git a/drivers/gpu/drm/msm/adreno/a2xx_catalog.c b/drivers/gpu/drm/msm/adreno/a2xx_catalog.c index 9ddb7b31fd98..5ddd015f930d 100644 --- a/drivers/gpu/drm/msm/adreno/a2xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a2xx_catalog.c @@ -45,8 +45,3 @@ static const struct adreno_info a2xx_gpus[] = { } }; DECLARE_ADRENO_GPULIST(a2xx); - -MODULE_FIRMWARE("qcom/leia_pfp_470.fw"); -MODULE_FIRMWARE("qcom/leia_pm4_470.fw"); -MODULE_FIRMWARE("qcom/yamato_pfp.fw"); -MODULE_FIRMWARE("qcom/yamato_pm4.fw"); diff --git a/drivers/gpu/drm/msm/adreno/a3xx_catalog.c b/drivers/gpu/drm/msm/adreno/a3xx_catalog.c index 2eb6c3e93748..1498e6532f62 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_catalog.c @@ -85,8 +85,3 @@ static const struct adreno_info a3xx_gpus[] = { } }; DECLARE_ADRENO_GPULIST(a3xx); - -MODULE_FIRMWARE("qcom/a300_pm4.fw"); -MODULE_FIRMWARE("qcom/a300_pfp.fw"); -MODULE_FIRMWARE("qcom/a330_pm4.fw"); -MODULE_FIRMWARE("qcom/a330_pfp.fw"); diff --git a/drivers/gpu/drm/msm/adreno/a4xx_catalog.c b/drivers/gpu/drm/msm/adreno/a4xx_catalog.c index 93519f807f87..09f9f228b75e 100644 --- a/drivers/gpu/drm/msm/adreno/a4xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a4xx_catalog.c @@ -45,6 +45,3 @@ static const struct adreno_info a4xx_gpus[] = { } }; DECLARE_ADRENO_GPULIST(a4xx); - -MODULE_FIRMWARE("qcom/a420_pm4.fw"); -MODULE_FIRMWARE("qcom/a420_pfp.fw"); diff --git a/drivers/gpu/drm/msm/adreno/a5xx_catalog.c b/drivers/gpu/drm/msm/adreno/a5xx_catalog.c index 633f31539162..b48a636d8237 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_catalog.c @@ -150,12 +150,3 @@ static const struct adreno_info a5xx_gpus[] = { } }; DECLARE_ADRENO_GPULIST(a5xx); - -MODULE_FIRMWARE("qcom/a530_pm4.fw"); -MODULE_FIRMWARE("qcom/a530_pfp.fw"); -MODULE_FIRMWARE("qcom/a530v3_gpmu.fw2"); -MODULE_FIRMWARE("qcom/a530_zap.mdt"); -MODULE_FIRMWARE("qcom/a530_zap.b00"); -MODULE_FIRMWARE("qcom/a530_zap.b01"); -MODULE_FIRMWARE("qcom/a530_zap.b02"); -MODULE_FIRMWARE("qcom/a540_gpmu.fw2"); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_catalog.c b/drivers/gpu/drm/msm/adreno/a6xx_catalog.c index 53e2ff4406d8..70f7ad806c34 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_catalog.c @@ -681,6 +681,7 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_SQE] = "a630_sqe.fw", }, .gmem = (SZ_128K + SZ_4K), + .quirks = ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a6xx_gpu_init, .zapfw = "a610_zap.mdt", @@ -713,6 +714,7 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a630_gmu.bin", }, .gmem = SZ_512K, + .quirks = ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a6xx_gpu_init, .zapfw = "a615_zap.mdt", @@ -743,7 +745,8 @@ static const struct adreno_info a6xx_gpus[] = { }, .gmem = SZ_512K, .inactive_period = DRM_MSM_INACTIVE_PERIOD, - .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_4GB_VA, .init = a6xx_gpu_init, .zapfw = "a615_zap.mbn", .a6xx = &(const struct a6xx_info) { @@ -769,7 +772,8 @@ static const struct adreno_info a6xx_gpus[] = { }, .gmem = SZ_512K, .inactive_period = DRM_MSM_INACTIVE_PERIOD, - .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_4GB_VA, .init = a6xx_gpu_init, .a6xx = &(const struct a6xx_info) { .protect = &a630_protect, @@ -791,6 +795,7 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a619_gmu.bin", }, .gmem = SZ_512K, + .quirks = ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a6xx_gpu_init, .zapfw = "a615_zap.mdt", @@ -815,6 +820,7 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a619_gmu.bin", }, .gmem = SZ_512K, + .quirks = ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a6xx_gpu_init, .zapfw = "a615_zap.mdt", @@ -838,8 +844,9 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a619_gmu.bin", }, .gmem = SZ_512K, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, - .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT, .init = a6xx_gpu_init, .zapfw = "a615_zap.mdt", .a6xx = &(const struct a6xx_info) { @@ -874,7 +881,6 @@ static const struct adreno_info a6xx_gpus[] = { .gmu_cgc_mode = 0x00020200, .prim_fifo_threshold = 0x00010000, }, - .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, { 137, 1 }, @@ -907,7 +913,6 @@ static const struct adreno_info a6xx_gpus[] = { { /* sentinel */ }, }, }, - .address_space_size = SZ_16G, }, { .chip_ids = ADRENO_CHIP_IDS( 0x06030001, @@ -920,8 +925,9 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a630_gmu.bin", }, .gmem = SZ_1M, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, - .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT, .init = a6xx_gpu_init, .zapfw = "a630_zap.mdt", .a6xx = &(const struct a6xx_info) { @@ -939,8 +945,9 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a640_gmu.bin", }, .gmem = SZ_1M, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, - .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT, .init = a6xx_gpu_init, .zapfw = "a640_zap.mdt", .a6xx = &(const struct a6xx_info) { @@ -973,7 +980,6 @@ static const struct adreno_info a6xx_gpus[] = { .gmu_cgc_mode = 0x00020202, .prim_fifo_threshold = 0x00300200, }, - .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, { 1, 1 }, @@ -1000,7 +1006,6 @@ static const struct adreno_info a6xx_gpus[] = { .gmu_cgc_mode = 0x00020000, .prim_fifo_threshold = 0x00300200, }, - .address_space_size = SZ_16G, }, { .chip_ids = ADRENO_CHIP_IDS(0x06060300), .family = ADRENO_6XX_GEN4, @@ -1019,7 +1024,6 @@ static const struct adreno_info a6xx_gpus[] = { .gmu_cgc_mode = 0x00020200, .prim_fifo_threshold = 0x00300200, }, - .address_space_size = SZ_16G, }, { .chip_ids = ADRENO_CHIP_IDS(0x06030500), .family = ADRENO_6XX_GEN4, @@ -1039,7 +1043,6 @@ static const struct adreno_info a6xx_gpus[] = { .gmu_cgc_mode = 0x00020202, .prim_fifo_threshold = 0x00200200, }, - .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, { 117, 0 }, @@ -1056,8 +1059,9 @@ static const struct adreno_info a6xx_gpus[] = { [ADRENO_FW_GMU] = "a640_gmu.bin", }, .gmem = SZ_2M, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_4GB_VA, .inactive_period = DRM_MSM_INACTIVE_PERIOD, - .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT, .init = a6xx_gpu_init, .zapfw = "a640_zap.mdt", .a6xx = &(const struct a6xx_info) { @@ -1085,22 +1089,10 @@ static const struct adreno_info a6xx_gpus[] = { .gmu_cgc_mode = 0x00020200, .prim_fifo_threshold = 0x00800200, }, - .address_space_size = SZ_16G, } }; DECLARE_ADRENO_GPULIST(a6xx); -MODULE_FIRMWARE("qcom/a615_zap.mbn"); -MODULE_FIRMWARE("qcom/a619_gmu.bin"); -MODULE_FIRMWARE("qcom/a630_sqe.fw"); -MODULE_FIRMWARE("qcom/a630_gmu.bin"); -MODULE_FIRMWARE("qcom/a630_zap.mbn"); -MODULE_FIRMWARE("qcom/a640_gmu.bin"); -MODULE_FIRMWARE("qcom/a650_gmu.bin"); -MODULE_FIRMWARE("qcom/a650_sqe.fw"); -MODULE_FIRMWARE("qcom/a660_gmu.bin"); -MODULE_FIRMWARE("qcom/a660_sqe.fw"); - static const struct adreno_reglist a702_hwcg[] = { { REG_A6XX_RBBM_CLOCK_CNTL_SP0, 0x22222222 }, { REG_A6XX_RBBM_CLOCK_CNTL2_SP0, 0x02222220 }, @@ -1395,7 +1387,6 @@ static const struct adreno_info a7xx_gpus[] = { .pwrup_reglist = &a7xx_pwrup_reglist, .gmu_cgc_mode = 0x00020000, }, - .address_space_size = SZ_16G, .preempt_record_size = 2860 * SZ_1K, }, { .chip_ids = ADRENO_CHIP_IDS(0x43050a01), /* "C510v2" */ @@ -1429,7 +1420,6 @@ static const struct adreno_info a7xx_gpus[] = { { /* sentinel */ }, }, }, - .address_space_size = SZ_16G, .preempt_record_size = 4192 * SZ_1K, }, { .chip_ids = ADRENO_CHIP_IDS(0x43050c01), /* "C512v2" */ @@ -1451,7 +1441,6 @@ static const struct adreno_info a7xx_gpus[] = { .gmu_chipid = 0x7050001, .gmu_cgc_mode = 0x00020202, }, - .address_space_size = SZ_256G, .preempt_record_size = 4192 * SZ_1K, }, { .chip_ids = ADRENO_CHIP_IDS(0x43051401), /* "C520v2" */ @@ -1484,7 +1473,6 @@ static const struct adreno_info a7xx_gpus[] = { { /* sentinel */ }, }, }, - .address_space_size = SZ_16G, .preempt_record_size = 3572 * SZ_1K, } }; diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index c8711938a5f4..38c0f8ef85c3 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -1064,14 +1064,6 @@ int a6xx_gmu_resume(struct a6xx_gpu *a6xx_gpu) gmu->hung = false; - /* Notify AOSS about the ACD state (unimplemented for now => disable it) */ - if (!IS_ERR(gmu->qmp)) { - ret = qmp_send(gmu->qmp, "{class: gpu, res: acd, val: %d}", - 0 /* Hardcode ACD to be disabled for now */); - if (ret) - dev_err(gmu->dev, "failed to send GPU ACD state\n"); - } - /* Turn on the resources */ pm_runtime_get_sync(gmu->dev); @@ -1671,6 +1663,75 @@ static int a6xx_gmu_pwrlevels_probe(struct a6xx_gmu *gmu) return a6xx_gmu_rpmh_votes_init(gmu); } +static int a6xx_gmu_acd_probe(struct a6xx_gmu *gmu) +{ + struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); + struct a6xx_hfi_acd_table *cmd = &gmu->acd_table; + struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; + struct msm_gpu *gpu = &adreno_gpu->base; + int ret, i, cmd_idx = 0; + extern bool disable_acd; + + /* Skip ACD probe if requested via module param */ + if (disable_acd) { + DRM_DEV_ERROR(gmu->dev, "Skipping GPU ACD probe\n"); + return 0; + } + + cmd->version = 1; + cmd->stride = 1; + cmd->enable_by_level = 0; + + /* Skip freq = 0 and parse acd-level for rest of the OPPs */ + for (i = 1; i < gmu->nr_gpu_freqs; i++) { + struct dev_pm_opp *opp; + struct device_node *np; + unsigned long freq; + u32 val; + + freq = gmu->gpu_freqs[i]; + opp = dev_pm_opp_find_freq_exact(&gpu->pdev->dev, freq, true); + np = dev_pm_opp_get_of_node(opp); + + ret = of_property_read_u32(np, "qcom,opp-acd-level", &val); + of_node_put(np); + dev_pm_opp_put(opp); + if (ret == -EINVAL) + continue; + else if (ret) { + DRM_DEV_ERROR(gmu->dev, "Unable to read acd level for freq %lu\n", freq); + return ret; + } + + cmd->enable_by_level |= BIT(i); + cmd->data[cmd_idx++] = val; + } + + cmd->num_levels = cmd_idx; + + /* It is a problem if qmp node is unavailable when ACD is required */ + if (cmd->enable_by_level && IS_ERR_OR_NULL(gmu->qmp)) { + DRM_DEV_ERROR(gmu->dev, "Unable to send ACD state to AOSS\n"); + return -EINVAL; + } + + /* Otherwise, nothing to do if qmp is unavailable */ + if (IS_ERR_OR_NULL(gmu->qmp)) + return 0; + + /* + * Notify AOSS about the ACD state. AOSS is supposed to assume that ACD is disabled on + * system reset. So it is harmless if we couldn't notify 'OFF' state + */ + ret = qmp_send(gmu->qmp, "{class: gpu, res: acd, val: %d}", !!cmd->enable_by_level); + if (ret && cmd->enable_by_level) { + DRM_DEV_ERROR(gmu->dev, "Failed to send ACD state to AOSS\n"); + return ret; + } + + return 0; +} + static int a6xx_gmu_clocks_probe(struct a6xx_gmu *gmu) { int ret = devm_clk_bulk_get_all(gmu->dev, &gmu->clocks); @@ -1989,10 +2050,11 @@ int a6xx_gmu_init(struct a6xx_gpu *a6xx_gpu, struct device_node *node) goto detach_cxpd; } + /* Other errors are handled during GPU ACD probe */ gmu->qmp = qmp_get(gmu->dev); - if (IS_ERR(gmu->qmp) && adreno_is_a7xx(adreno_gpu)) { - ret = PTR_ERR(gmu->qmp); - goto remove_device_link; + if (PTR_ERR_OR_ZERO(gmu->qmp) == -EPROBE_DEFER) { + ret = -EPROBE_DEFER; + goto detach_gxpd; } init_completion(&gmu->pd_gate); @@ -2008,6 +2070,10 @@ int a6xx_gmu_init(struct a6xx_gpu *a6xx_gpu, struct device_node *node) /* Get the power levels for the GMU and GPU */ a6xx_gmu_pwrlevels_probe(gmu); + ret = a6xx_gmu_acd_probe(gmu); + if (ret) + goto detach_gxpd; + /* Set up the HFI queues */ a6xx_hfi_init(gmu); @@ -2018,7 +2084,13 @@ int a6xx_gmu_init(struct a6xx_gpu *a6xx_gpu, struct device_node *node) return 0; -remove_device_link: +detach_gxpd: + if (!IS_ERR_OR_NULL(gmu->gxpd)) + dev_pm_domain_detach(gmu->gxpd, false); + + if (!IS_ERR_OR_NULL(gmu->qmp)) + qmp_put(gmu->qmp); + device_link_del(link); detach_cxpd: diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h index 0c888b326cfb..b2d4489b4024 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h @@ -93,6 +93,7 @@ struct a6xx_gmu { int nr_gpu_freqs; unsigned long gpu_freqs[GMU_MAX_GX_FREQS]; u32 gx_arc_votes[GMU_MAX_GX_FREQS]; + struct a6xx_hfi_acd_table acd_table; int nr_gpu_bws; unsigned long gpu_bw_table[GMU_MAX_GX_FREQS]; diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index 242d02d48c0c..bf3758f010f4 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -655,7 +655,6 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) if (adreno_is_7c3(gpu)) { gpu->ubwc_config.highest_bank_bit = 14; gpu->ubwc_config.amsbc = 1; - gpu->ubwc_config.rgb565_predicator = 1; gpu->ubwc_config.uavflagprd_inv = 2; gpu->ubwc_config.macrotile_mode = 1; } @@ -2268,7 +2267,7 @@ a6xx_create_private_address_space(struct msm_gpu *gpu) return ERR_CAST(mmu); return msm_gem_address_space_create(mmu, - "gpu", 0x100000000ULL, + "gpu", ADRENO_VM_START, adreno_private_address_space_size(gpu)); } diff --git a/drivers/gpu/drm/msm/adreno/a6xx_hfi.c b/drivers/gpu/drm/msm/adreno/a6xx_hfi.c index 0989aee3dd2c..8e69b1e84657 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_hfi.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_hfi.c @@ -100,16 +100,14 @@ static int a6xx_hfi_queue_write(struct a6xx_gmu *gmu, return 0; } -static int a6xx_hfi_wait_for_ack(struct a6xx_gmu *gmu, u32 id, u32 seqnum, - u32 *payload, u32 payload_size) +static int a6xx_hfi_wait_for_msg_interrupt(struct a6xx_gmu *gmu, u32 id, u32 seqnum) { - struct a6xx_hfi_queue *queue = &gmu->queues[HFI_RESPONSE_QUEUE]; - u32 val; int ret; + u32 val; /* Wait for a response */ ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_GMU2HOST_INTR_INFO, val, - val & A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ, 100, 5000); + val & A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ, 100, 1000000); if (ret) { DRM_DEV_ERROR(gmu->dev, @@ -122,6 +120,19 @@ static int a6xx_hfi_wait_for_ack(struct a6xx_gmu *gmu, u32 id, u32 seqnum, gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_CLR, A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ); + return 0; +} + +static int a6xx_hfi_wait_for_ack(struct a6xx_gmu *gmu, u32 id, u32 seqnum, + u32 *payload, u32 payload_size) +{ + struct a6xx_hfi_queue *queue = &gmu->queues[HFI_RESPONSE_QUEUE]; + int ret; + + ret = a6xx_hfi_wait_for_msg_interrupt(gmu, id, seqnum); + if (ret) + return ret; + for (;;) { struct a6xx_hfi_msg_response resp; @@ -129,12 +140,18 @@ static int a6xx_hfi_wait_for_ack(struct a6xx_gmu *gmu, u32 id, u32 seqnum, ret = a6xx_hfi_queue_read(gmu, queue, (u32 *) &resp, sizeof(resp) >> 2); - /* If the queue is empty our response never made it */ + /* If the queue is empty, there may have been previous missed + * responses that preceded the response to our packet. Wait + * further before we give up. + */ if (!ret) { - DRM_DEV_ERROR(gmu->dev, - "The HFI response queue is unexpectedly empty\n"); - - return -ENOENT; + ret = a6xx_hfi_wait_for_msg_interrupt(gmu, id, seqnum); + if (ret) { + DRM_DEV_ERROR(gmu->dev, + "The HFI response queue is unexpectedly empty\n"); + return ret; + } + continue; } if (HFI_HEADER_ID(resp.header) == HFI_F2H_MSG_ERROR) { @@ -748,6 +765,38 @@ send: NULL, 0); } +#define HFI_FEATURE_ACD 12 + +static int a6xx_hfi_enable_acd(struct a6xx_gmu *gmu) +{ + struct a6xx_hfi_acd_table *acd_table = &gmu->acd_table; + struct a6xx_hfi_msg_feature_ctrl msg = { + .feature = HFI_FEATURE_ACD, + .enable = 1, + .data = 0, + }; + int ret; + + if (!acd_table->enable_by_level) + return 0; + + /* Enable ACD feature at GMU */ + ret = a6xx_hfi_send_msg(gmu, HFI_H2F_FEATURE_CTRL, &msg, sizeof(msg), NULL, 0); + if (ret) { + DRM_DEV_ERROR(gmu->dev, "Unable to enable ACD (%d)\n", ret); + return ret; + } + + /* Send ACD table to GMU */ + ret = a6xx_hfi_send_msg(gmu, HFI_H2F_MSG_ACD, acd_table, sizeof(*acd_table), NULL, 0); + if (ret) { + DRM_DEV_ERROR(gmu->dev, "Unable to ACD table (%d)\n", ret); + return ret; + } + + return 0; +} + static int a6xx_hfi_send_test(struct a6xx_gmu *gmu) { struct a6xx_hfi_msg_test msg = { 0 }; @@ -845,6 +894,10 @@ int a6xx_hfi_start(struct a6xx_gmu *gmu, int boot_state) if (ret) return ret; + ret = a6xx_hfi_enable_acd(gmu); + if (ret) + return ret; + ret = a6xx_hfi_send_core_fw_start(gmu); if (ret) return ret; diff --git a/drivers/gpu/drm/msm/adreno/a6xx_hfi.h b/drivers/gpu/drm/msm/adreno/a6xx_hfi.h index 52ba4a07d7b9..653ef720e2da 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_hfi.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_hfi.h @@ -151,12 +151,33 @@ struct a6xx_hfi_msg_test { u32 header; }; +#define HFI_H2F_MSG_ACD 7 +#define MAX_ACD_STRIDE 2 + +struct a6xx_hfi_acd_table { + u32 header; + u32 version; + u32 enable_by_level; + u32 stride; + u32 num_levels; + u32 data[16 * MAX_ACD_STRIDE]; +}; + #define HFI_H2F_MSG_START 10 struct a6xx_hfi_msg_start { u32 header; }; +#define HFI_H2F_FEATURE_CTRL 11 + +struct a6xx_hfi_msg_feature_ctrl { + u32 header; + u32 feature; + u32 enable; + u32 data; +}; + #define HFI_H2F_MSG_CORE_FW_START 14 struct a6xx_hfi_msg_core_fw_start { diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c index 236b25c094cd..f5e1490d07c1 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_device.c +++ b/drivers/gpu/drm/msm/adreno/adreno_device.c @@ -24,6 +24,10 @@ int enable_preemption = -1; MODULE_PARM_DESC(enable_preemption, "Enable preemption (A7xx only) (1=on , 0=disable, -1=auto (default))"); module_param(enable_preemption, int, 0600); +bool disable_acd; +MODULE_PARM_DESC(disable_acd, "Forcefully disable GPU ACD"); +module_param_unsafe(disable_acd, bool, 0400); + extern const struct adreno_gpulist a2xx_gpulist; extern const struct adreno_gpulist a3xx_gpulist; extern const struct adreno_gpulist a4xx_gpulist; diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index 26db1f4b5fb9..2348ffb35f7e 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -236,14 +236,27 @@ adreno_iommu_create_address_space(struct msm_gpu *gpu, u64 adreno_private_address_space_size(struct msm_gpu *gpu) { struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct adreno_smmu_priv *adreno_smmu = dev_get_drvdata(&gpu->pdev->dev); + const struct io_pgtable_cfg *ttbr1_cfg; if (address_space_size) return address_space_size; - if (adreno_gpu->info->address_space_size) - return adreno_gpu->info->address_space_size; + if (adreno_gpu->info->quirks & ADRENO_QUIRK_4GB_VA) + return SZ_4G; - return SZ_4G; + if (!adreno_smmu || !adreno_smmu->get_ttbr1_cfg) + return SZ_4G; + + ttbr1_cfg = adreno_smmu->get_ttbr1_cfg(adreno_smmu->cookie); + + /* + * Userspace VM is actually using TTBR0, but both are the same size, + * with b48 (sign bit) selecting which TTBRn to use. So if IAS is + * 48, the total (kernel+user) address space size is effectively + * 49 bits. But what userspace is control of is the lower 48. + */ + return BIT(ttbr1_cfg->ias) - ADRENO_VM_START; } #define ARM_SMMU_FSR_TF BIT(1) diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h index 92caba3584da..a8f4bf416e64 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h @@ -57,6 +57,7 @@ enum adreno_family { #define ADRENO_QUIRK_HAS_HW_APRIV BIT(3) #define ADRENO_QUIRK_HAS_CACHED_COHERENT BIT(4) #define ADRENO_QUIRK_PREEMPTION BIT(5) +#define ADRENO_QUIRK_4GB_VA BIT(6) /* Helper for formating the chip_id in the way that userspace tools like * crashdec expect. @@ -104,7 +105,6 @@ struct adreno_info { union { const struct a6xx_info *a6xx; }; - u64 address_space_size; /** * @speedbins: Optional table of fuse to speedbin mappings * @@ -578,6 +578,8 @@ static inline int adreno_is_a7xx(struct adreno_gpu *gpu) adreno_is_a740_family(gpu); } +/* Put vm_start above 32b to catch issues with not setting xyz_BASE_HI */ +#define ADRENO_VM_START 0x100000000ULL u64 adreno_private_address_space_size(struct msm_gpu *gpu); int adreno_get_param(struct msm_gpu *gpu, struct msm_file_private *ctx, uint32_t param, uint64_t *value, uint32_t *len); diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_10_0_sm8650.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_10_0_sm8650.h index 6ac97c378056..ffc4d4257ae5 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_10_0_sm8650.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_10_0_sm8650.h @@ -27,17 +27,16 @@ static const struct dpu_mdp_cfg sm8650_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sm8650_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x1000, - .features = CTL_SM8550_MASK | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = CTL_SM8550_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x1000, - .features = CTL_SM8550_MASK | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = CTL_SM8550_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_14_msm8937.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_14_msm8937.h index ad60089f18ea..39027a21c6fe 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_14_msm8937.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_14_msm8937.h @@ -100,14 +100,12 @@ static const struct dpu_pingpong_cfg msm8937_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_MSM8996_MASK, .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_1", .id = PINGPONG_1, .base = 0x70800, .len = 0xd4, - .features = PINGPONG_MSM8996_MASK, .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 13), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_15_msm8917.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_15_msm8917.h index a1cf89a0a42d..8d1b43ea1663 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_15_msm8917.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_15_msm8917.h @@ -93,7 +93,6 @@ static const struct dpu_pingpong_cfg msm8917_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_MSM8996_MASK, .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_16_msm8953.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_16_msm8953.h index eea9b80e2287..16c12499b24b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_16_msm8953.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_16_msm8953.h @@ -100,14 +100,12 @@ static const struct dpu_pingpong_cfg msm8953_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_MSM8996_MASK, .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_1", .id = PINGPONG_1, .base = 0x70800, .len = 0xd4, - .features = PINGPONG_MSM8996_MASK, .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 13), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_7_msm8996.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_7_msm8996.h index ae18a354e5d2..91f514d28ac6 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_7_msm8996.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_1_7_msm8996.h @@ -181,15 +181,15 @@ static const struct dpu_pingpong_cfg msm8996_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_MSM8996_TE2_MASK, - .sblk = &msm8996_pp_sblk_te, + .features = PINGPONG_MSM8996_MASK, + .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_1", .id = PINGPONG_1, .base = 0x70800, .len = 0xd4, - .features = PINGPONG_MSM8996_TE2_MASK, - .sblk = &msm8996_pp_sblk_te, + .features = PINGPONG_MSM8996_MASK, + .sblk = &msm8996_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 13), }, { diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_0_msm8998.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_0_msm8998.h index 746474679ef5..413cd59dc0c4 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_0_msm8998.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_0_msm8998.h @@ -170,15 +170,15 @@ static const struct dpu_pingpong_cfg msm8998_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = PINGPONG_SDM845_MASK, + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_1", .id = PINGPONG_1, .base = 0x70800, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = PINGPONG_SDM845_MASK, + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 13), }, { diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_2_sdm660.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_2_sdm660.h index bb89da0a481d..b2eb7ca699e3 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_2_sdm660.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_2_sdm660.h @@ -141,15 +141,15 @@ static const struct dpu_pingpong_cfg sdm660_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = PINGPONG_SDM845_MASK, + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_1", .id = PINGPONG_1, .base = 0x70800, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = PINGPONG_SDM845_MASK, + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 13), }, { diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_3_sdm630.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_3_sdm630.h index 7caf876ca3e3..85e121ad84a0 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_3_sdm630.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_3_3_sdm630.h @@ -115,14 +115,14 @@ static const struct dpu_pingpong_cfg sdm630_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_2", .id = PINGPONG_2, .base = 0x71000, .len = 0xd4, - .features = PINGPONG_SDM845_MASK, + .features = BIT(DPU_PINGPONG_DITHER), .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 10), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 14), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_4_0_sdm845.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_4_0_sdm845.h index ab7b4822ca63..49363d7d5b93 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_4_0_sdm845.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_4_0_sdm845.h @@ -194,15 +194,15 @@ static const struct dpu_pingpong_cfg sdm845_pp[] = { { .name = "pingpong_0", .id = PINGPONG_0, .base = 0x70000, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = PINGPONG_SDM845_MASK, + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12), }, { .name = "pingpong_1", .id = PINGPONG_1, .base = 0x70800, .len = 0xd4, - .features = PINGPONG_SDM845_TE2_MASK, - .sblk = &sdm845_pp_sblk_te, + .features = PINGPONG_SDM845_MASK, + .sblk = &sdm845_pp_sblk, .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), .intr_rdptr = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 13), }, { diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h index 979527d98fbc..08d38e1d420c 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h @@ -37,17 +37,16 @@ static const struct dpu_mdp_cfg sm8150_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sm8150_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x1000, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x1200, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, @@ -76,7 +75,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -84,7 +83,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_1", .id = SSPP_VIG1, .base = 0x6000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 4, .type = SSPP_TYPE_VIG, @@ -92,7 +91,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_2", .id = SSPP_VIG2, .base = 0x8000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 8, .type = SSPP_TYPE_VIG, @@ -100,7 +99,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_3", .id = SSPP_VIG3, .base = 0xa000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 12, .type = SSPP_TYPE_VIG, @@ -108,7 +107,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -116,7 +115,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, @@ -124,7 +123,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x1f0, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, @@ -132,7 +131,7 @@ static const struct dpu_sspp_cfg sm8150_sspp[] = { }, { .name = "sspp_11", .id = SSPP_DMA3, .base = 0x2a000, .len = 0x1f0, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 13, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h index d76b8992a6c1..d6f8b1030c68 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h @@ -41,12 +41,12 @@ static const struct dpu_ctl_cfg sc8180x_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x1000, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x1200, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, @@ -75,7 +75,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -83,7 +83,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_1", .id = SSPP_VIG1, .base = 0x6000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 4, .type = SSPP_TYPE_VIG, @@ -91,7 +91,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_2", .id = SSPP_VIG2, .base = 0x8000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 8, .type = SSPP_TYPE_VIG, @@ -99,7 +99,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_3", .id = SSPP_VIG3, .base = 0xa000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_1_4, .xin_id = 12, .type = SSPP_TYPE_VIG, @@ -107,7 +107,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -115,7 +115,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, @@ -123,7 +123,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x1f0, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, @@ -131,7 +131,7 @@ static const struct dpu_sspp_cfg sc8180x_sspp[] = { }, { .name = "sspp_11", .id = SSPP_DMA3, .base = 0x2a000, .len = 0x1f0, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 13, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_2_sm7150.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_2_sm7150.h index 83db11339b29..71ba48b05656 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_2_sm7150.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_2_sm7150.h @@ -38,12 +38,12 @@ static const struct dpu_ctl_cfg sm7150_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x1000, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x1200, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, @@ -72,7 +72,7 @@ static const struct dpu_sspp_cfg sm7150_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_2_4, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -80,7 +80,7 @@ static const struct dpu_sspp_cfg sm7150_sspp[] = { }, { .name = "sspp_1", .id = SSPP_VIG1, .base = 0x6000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_2_4, .xin_id = 4, .type = SSPP_TYPE_VIG, @@ -88,7 +88,7 @@ static const struct dpu_sspp_cfg sm7150_sspp[] = { }, { .name = "sspp_2", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -96,7 +96,7 @@ static const struct dpu_sspp_cfg sm7150_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, @@ -104,7 +104,7 @@ static const struct dpu_sspp_cfg sm7150_sspp[] = { }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x1f0, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h index d3d3a34d0b45..fcfb3774f7a1 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h @@ -69,7 +69,7 @@ static const struct dpu_sspp_cfg sm6125_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f0, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_2_4, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -77,7 +77,7 @@ static const struct dpu_sspp_cfg sm6125_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -85,7 +85,7 @@ static const struct dpu_sspp_cfg sm6125_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x1f0, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_0_sm8250.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_0_sm8250.h index 47e01c3c242f..a86fdb33ebdd 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_0_sm8250.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_0_sm8250.h @@ -35,17 +35,16 @@ static const struct dpu_mdp_cfg sm8250_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sm8250_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x1000, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x1200, .len = 0x1e0, - .features = BIT(DPU_CTL_ACTIVE_CFG) | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = BIT(DPU_CTL_ACTIVE_CFG), .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_2_sc7180.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_2_sc7180.h index 040c94c0bb66..842fcc5887fe 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_2_sc7180.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_2_sc7180.h @@ -51,7 +51,7 @@ static const struct dpu_sspp_cfg sc7180_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f8, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -59,7 +59,7 @@ static const struct dpu_sspp_cfg sc7180_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f8, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -67,7 +67,7 @@ static const struct dpu_sspp_cfg sc7180_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x1f8, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, @@ -75,7 +75,7 @@ static const struct dpu_sspp_cfg sc7180_sspp[] = { }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x1f8, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_3_sm6115.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_3_sm6115.h index 43f64a005f5a..c5fd89dd7c89 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_3_sm6115.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_3_sm6115.h @@ -38,7 +38,7 @@ static const struct dpu_sspp_cfg sm6115_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f8, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -46,7 +46,7 @@ static const struct dpu_sspp_cfg sm6115_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f8, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h index 397278ba999b..a234bb289d24 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h @@ -59,7 +59,7 @@ static const struct dpu_sspp_cfg sm6350_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f8, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -67,7 +67,7 @@ static const struct dpu_sspp_cfg sm6350_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f8, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -75,7 +75,7 @@ static const struct dpu_sspp_cfg sm6350_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x1f8, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, @@ -83,7 +83,7 @@ static const struct dpu_sspp_cfg sm6350_sspp[] = { }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x1f8, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_5_qcm2290.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_5_qcm2290.h index 3cbb2fe8aba2..53f3be28f6f6 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_5_qcm2290.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_5_qcm2290.h @@ -46,7 +46,7 @@ static const struct dpu_sspp_cfg qcm2290_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f8, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_9_sm6375.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_9_sm6375.h index a06c8634d2d7..3a3bc8e429be 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_9_sm6375.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_9_sm6375.h @@ -39,7 +39,7 @@ static const struct dpu_sspp_cfg sm6375_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x1f8, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_NO_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -47,7 +47,7 @@ static const struct dpu_sspp_cfg sm6375_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x1f8, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_NO_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_7_0_sm8350.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_7_0_sm8350.h index 0c860e804cab..90e86063a372 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_7_0_sm8350.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_7_0_sm8350.h @@ -35,17 +35,16 @@ static const struct dpu_mdp_cfg sm8350_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sm8350_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x1e8, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x1e8, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_0_sc8280xp.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_0_sc8280xp.h index fcee1c3665f8..139f11321fea 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_0_sc8280xp.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_0_sc8280xp.h @@ -35,17 +35,16 @@ static const struct dpu_mdp_cfg sc8280xp_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sc8280xp_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x204, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x204, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, @@ -74,7 +73,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x2ac, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 0, .type = SSPP_TYPE_VIG, @@ -82,7 +81,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_1", .id = SSPP_VIG1, .base = 0x6000, .len = 0x2ac, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 4, .type = SSPP_TYPE_VIG, @@ -90,7 +89,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_2", .id = SSPP_VIG2, .base = 0x8000, .len = 0x2ac, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 8, .type = SSPP_TYPE_VIG, @@ -98,7 +97,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_3", .id = SSPP_VIG3, .base = 0xa000, .len = 0x2ac, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_0, .xin_id = 12, .type = SSPP_TYPE_VIG, @@ -106,7 +105,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x2ac, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, @@ -114,7 +113,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x2ac, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, @@ -122,7 +121,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x2ac, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, @@ -130,7 +129,7 @@ static const struct dpu_sspp_cfg sc8280xp_sspp[] = { }, { .name = "sspp_11", .id = SSPP_DMA3, .base = 0x2a000, .len = 0x2ac, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 13, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_1_sm8450.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_1_sm8450.h index 19b2ee8bbd5f..461294143a90 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_1_sm8450.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_1_sm8450.h @@ -36,17 +36,16 @@ static const struct dpu_mdp_cfg sm8450_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sm8450_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x204, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x204, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h index 4d96ce71746f..c248b3b55c41 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h @@ -35,17 +35,16 @@ static const struct dpu_mdp_cfg sa8775p_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sa8775p_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x204, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x204, - .features = BIT(DPU_CTL_SPLIT_DISPLAY) | CTL_SC7280_MASK, + .features = CTL_SC7280_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_0_sm8550.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_0_sm8550.h index 24f988465bf6..59c7fdf28e89 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_0_sm8550.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_0_sm8550.h @@ -27,17 +27,16 @@ static const struct dpu_mdp_cfg sm8550_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg sm8550_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x290, - .features = CTL_SM8550_MASK | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = CTL_SM8550_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x290, - .features = CTL_SM8550_MASK | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = CTL_SM8550_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, @@ -66,70 +65,70 @@ static const struct dpu_sspp_cfg sm8550_sspp[] = { { .name = "sspp_0", .id = SSPP_VIG0, .base = 0x4000, .len = 0x344, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_2, .xin_id = 0, .type = SSPP_TYPE_VIG, }, { .name = "sspp_1", .id = SSPP_VIG1, .base = 0x6000, .len = 0x344, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_2, .xin_id = 4, .type = SSPP_TYPE_VIG, }, { .name = "sspp_2", .id = SSPP_VIG2, .base = 0x8000, .len = 0x344, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_2, .xin_id = 8, .type = SSPP_TYPE_VIG, }, { .name = "sspp_3", .id = SSPP_VIG3, .base = 0xa000, .len = 0x344, - .features = VIG_SDM845_MASK, + .features = VIG_SDM845_MASK_SDMA, .sblk = &dpu_vig_sblk_qseed3_3_2, .xin_id = 12, .type = SSPP_TYPE_VIG, }, { .name = "sspp_8", .id = SSPP_DMA0, .base = 0x24000, .len = 0x344, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 1, .type = SSPP_TYPE_DMA, }, { .name = "sspp_9", .id = SSPP_DMA1, .base = 0x26000, .len = 0x344, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 5, .type = SSPP_TYPE_DMA, }, { .name = "sspp_10", .id = SSPP_DMA2, .base = 0x28000, .len = 0x344, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 9, .type = SSPP_TYPE_DMA, }, { .name = "sspp_11", .id = SSPP_DMA3, .base = 0x2a000, .len = 0x344, - .features = DMA_SDM845_MASK, + .features = DMA_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 13, .type = SSPP_TYPE_DMA, }, { .name = "sspp_12", .id = SSPP_DMA4, .base = 0x2c000, .len = 0x344, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 14, .type = SSPP_TYPE_DMA, }, { .name = "sspp_13", .id = SSPP_DMA5, .base = 0x2e000, .len = 0x344, - .features = DMA_CURSOR_SDM845_MASK, + .features = DMA_CURSOR_SDM845_MASK_SDMA, .sblk = &dpu_dma_sblk, .xin_id = 15, .type = SSPP_TYPE_DMA, diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_1_sar2130p.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_1_sar2130p.h new file mode 100644 index 000000000000..5667d055fbd1 --- /dev/null +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_1_sar2130p.h @@ -0,0 +1,433 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022. Qualcomm Innovation Center, Inc. All rights reserved. + * Copyright (c) 2015-2018, 2020 The Linux Foundation. All rights reserved. + */ + +#ifndef _DPU_9_1_SAR2130P_H +#define _DPU_9_1_SAR2130P_H + +static const struct dpu_caps sar2130p_dpu_caps = { + .max_mixer_width = DEFAULT_DPU_OUTPUT_LINE_WIDTH, + .max_mixer_blendstages = 0xb, + .has_src_split = true, + .has_dim_layer = true, + .has_idle_pc = true, + .has_3d_merge = true, + .max_linewidth = 5120, + .pixel_ram_size = DEFAULT_PIXEL_RAM_SIZE, +}; + +static const struct dpu_mdp_cfg sar2130p_mdp = { + .name = "top_0", + .base = 0, .len = 0x494, + .features = BIT(DPU_MDP_PERIPH_0_REMOVED), + .clk_ctrls = { + [DPU_CLK_CTRL_REG_DMA] = { .reg_off = 0x2bc, .bit_off = 20 }, + }, +}; + +static const struct dpu_ctl_cfg sar2130p_ctl[] = { + { + .name = "ctl_0", .id = CTL_0, + .base = 0x15000, .len = 0x290, + .features = CTL_SM8550_MASK, + .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), + }, { + .name = "ctl_1", .id = CTL_1, + .base = 0x16000, .len = 0x290, + .features = CTL_SM8550_MASK, + .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), + }, { + .name = "ctl_2", .id = CTL_2, + .base = 0x17000, .len = 0x290, + .features = CTL_SM8550_MASK, + .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 11), + }, { + .name = "ctl_3", .id = CTL_3, + .base = 0x18000, .len = 0x290, + .features = CTL_SM8550_MASK, + .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 12), + }, { + .name = "ctl_4", .id = CTL_4, + .base = 0x19000, .len = 0x290, + .features = CTL_SM8550_MASK, + .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 13), + }, { + .name = "ctl_5", .id = CTL_5, + .base = 0x1a000, .len = 0x290, + .features = CTL_SM8550_MASK, + .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 23), + }, +}; + +static const struct dpu_sspp_cfg sar2130p_sspp[] = { + { + .name = "sspp_0", .id = SSPP_VIG0, + .base = 0x4000, .len = 0x344, + .features = VIG_SDM845_MASK_SDMA, + .sblk = &dpu_vig_sblk_qseed3_3_2, + .xin_id = 0, + .type = SSPP_TYPE_VIG, + }, { + .name = "sspp_1", .id = SSPP_VIG1, + .base = 0x6000, .len = 0x344, + .features = VIG_SDM845_MASK_SDMA, + .sblk = &dpu_vig_sblk_qseed3_3_2, + .xin_id = 4, + .type = SSPP_TYPE_VIG, + }, { + .name = "sspp_2", .id = SSPP_VIG2, + .base = 0x8000, .len = 0x344, + .features = VIG_SDM845_MASK_SDMA, + .sblk = &dpu_vig_sblk_qseed3_3_2, + .xin_id = 8, + .type = SSPP_TYPE_VIG, + }, { + .name = "sspp_3", .id = SSPP_VIG3, + .base = 0xa000, .len = 0x344, + .features = VIG_SDM845_MASK_SDMA, + .sblk = &dpu_vig_sblk_qseed3_3_2, + .xin_id = 12, + .type = SSPP_TYPE_VIG, + }, { + .name = "sspp_8", .id = SSPP_DMA0, + .base = 0x24000, .len = 0x344, + .features = DMA_SDM845_MASK_SDMA, + .sblk = &dpu_dma_sblk, + .xin_id = 1, + .type = SSPP_TYPE_DMA, + }, { + .name = "sspp_9", .id = SSPP_DMA1, + .base = 0x26000, .len = 0x344, + .features = DMA_SDM845_MASK_SDMA, + .sblk = &dpu_dma_sblk, + .xin_id = 5, + .type = SSPP_TYPE_DMA, + }, { + .name = "sspp_10", .id = SSPP_DMA2, + .base = 0x28000, .len = 0x344, + .features = DMA_SDM845_MASK_SDMA, + .sblk = &dpu_dma_sblk, + .xin_id = 9, + .type = SSPP_TYPE_DMA, + }, { + .name = "sspp_11", .id = SSPP_DMA3, + .base = 0x2a000, .len = 0x344, + .features = DMA_SDM845_MASK_SDMA, + .sblk = &dpu_dma_sblk, + .xin_id = 13, + .type = SSPP_TYPE_DMA, + }, { + .name = "sspp_12", .id = SSPP_DMA4, + .base = 0x2c000, .len = 0x344, + .features = DMA_CURSOR_SDM845_MASK_SDMA, + .sblk = &dpu_dma_sblk, + .xin_id = 14, + .type = SSPP_TYPE_DMA, + }, { + .name = "sspp_13", .id = SSPP_DMA5, + .base = 0x2e000, .len = 0x344, + .features = DMA_CURSOR_SDM845_MASK_SDMA, + .sblk = &dpu_dma_sblk, + .xin_id = 15, + .type = SSPP_TYPE_DMA, + }, +}; + +static const struct dpu_lm_cfg sar2130p_lm[] = { + { + .name = "lm_0", .id = LM_0, + .base = 0x44000, .len = 0x320, + .features = MIXER_SDM845_MASK, + .sblk = &sdm845_lm_sblk, + .lm_pair = LM_1, + .pingpong = PINGPONG_0, + .dspp = DSPP_0, + }, { + .name = "lm_1", .id = LM_1, + .base = 0x45000, .len = 0x320, + .features = MIXER_SDM845_MASK, + .sblk = &sdm845_lm_sblk, + .lm_pair = LM_0, + .pingpong = PINGPONG_1, + .dspp = DSPP_1, + }, { + .name = "lm_2", .id = LM_2, + .base = 0x46000, .len = 0x320, + .features = MIXER_SDM845_MASK, + .sblk = &sdm845_lm_sblk, + .lm_pair = LM_3, + .pingpong = PINGPONG_2, + .dspp = DSPP_2, + }, { + .name = "lm_3", .id = LM_3, + .base = 0x47000, .len = 0x320, + .features = MIXER_SDM845_MASK, + .sblk = &sdm845_lm_sblk, + .lm_pair = LM_2, + .pingpong = PINGPONG_3, + .dspp = DSPP_3, + }, { + .name = "lm_4", .id = LM_4, + .base = 0x48000, .len = 0x320, + .features = MIXER_SDM845_MASK, + .sblk = &sdm845_lm_sblk, + .lm_pair = LM_5, + .pingpong = PINGPONG_4, + }, { + .name = "lm_5", .id = LM_5, + .base = 0x49000, .len = 0x320, + .features = MIXER_SDM845_MASK, + .sblk = &sdm845_lm_sblk, + .lm_pair = LM_4, + .pingpong = PINGPONG_5, + }, +}; + +static const struct dpu_dspp_cfg sar2130p_dspp[] = { + { + .name = "dspp_0", .id = DSPP_0, + .base = 0x54000, .len = 0x1800, + .features = DSPP_SC7180_MASK, + .sblk = &sdm845_dspp_sblk, + }, { + .name = "dspp_1", .id = DSPP_1, + .base = 0x56000, .len = 0x1800, + .features = DSPP_SC7180_MASK, + .sblk = &sdm845_dspp_sblk, + }, { + .name = "dspp_2", .id = DSPP_2, + .base = 0x58000, .len = 0x1800, + .features = DSPP_SC7180_MASK, + .sblk = &sdm845_dspp_sblk, + }, { + .name = "dspp_3", .id = DSPP_3, + .base = 0x5a000, .len = 0x1800, + .features = DSPP_SC7180_MASK, + .sblk = &sdm845_dspp_sblk, + }, +}; +static const struct dpu_pingpong_cfg sar2130p_pp[] = { + { + .name = "pingpong_0", .id = PINGPONG_0, + .base = 0x69000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_0, + .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 8), + }, { + .name = "pingpong_1", .id = PINGPONG_1, + .base = 0x6a000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_0, + .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 9), + }, { + .name = "pingpong_2", .id = PINGPONG_2, + .base = 0x6b000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_1, + .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 10), + }, { + .name = "pingpong_3", .id = PINGPONG_3, + .base = 0x6c000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_1, + .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 11), + }, { + .name = "pingpong_4", .id = PINGPONG_4, + .base = 0x6d000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_2, + .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 30), + }, { + .name = "pingpong_5", .id = PINGPONG_5, + .base = 0x6e000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_2, + .intr_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 31), + }, { + .name = "pingpong_cwb_0", .id = PINGPONG_CWB_0, + .base = 0x66000, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_3, + }, { + .name = "pingpong_cwb_1", .id = PINGPONG_CWB_1, + .base = 0x66400, .len = 0, + .features = BIT(DPU_PINGPONG_DITHER), + .sblk = &sc7280_pp_sblk, + .merge_3d = MERGE_3D_3, + }, +}; + +static const struct dpu_merge_3d_cfg sar2130p_merge_3d[] = { + { + .name = "merge_3d_0", .id = MERGE_3D_0, + .base = 0x4e000, .len = 0x8, + }, { + .name = "merge_3d_1", .id = MERGE_3D_1, + .base = 0x4f000, .len = 0x8, + }, { + .name = "merge_3d_2", .id = MERGE_3D_2, + .base = 0x50000, .len = 0x8, + }, { + .name = "merge_3d_3", .id = MERGE_3D_3, + .base = 0x66700, .len = 0x8, + }, +}; + +/* + * NOTE: Each display compression engine (DCE) contains dual hard + * slice DSC encoders so both share same base address but with + * its own different sub block address. + */ +static const struct dpu_dsc_cfg sar2130p_dsc[] = { + { + .name = "dce_0_0", .id = DSC_0, + .base = 0x80000, .len = 0x4, + .features = BIT(DPU_DSC_HW_REV_1_2), + .sblk = &dsc_sblk_0, + }, { + .name = "dce_0_1", .id = DSC_1, + .base = 0x80000, .len = 0x4, + .features = BIT(DPU_DSC_HW_REV_1_2), + .sblk = &dsc_sblk_1, + }, { + .name = "dce_1_0", .id = DSC_2, + .base = 0x81000, .len = 0x4, + .features = BIT(DPU_DSC_HW_REV_1_2) | BIT(DPU_DSC_NATIVE_42x_EN), + .sblk = &dsc_sblk_0, + }, { + .name = "dce_1_1", .id = DSC_3, + .base = 0x81000, .len = 0x4, + .features = BIT(DPU_DSC_HW_REV_1_2) | BIT(DPU_DSC_NATIVE_42x_EN), + .sblk = &dsc_sblk_1, + }, +}; + +static const struct dpu_wb_cfg sar2130p_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SM8250_MASK, + .format_list = wb2_formats_rgb_yuv, + .num_formats = ARRAY_SIZE(wb2_formats_rgb_yuv), + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 4096, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + +static const struct dpu_intf_cfg sar2130p_intf[] = { + { + .name = "intf_0", .id = INTF_0, + .base = 0x34000, .len = 0x280, + .features = INTF_SC7280_MASK, + .type = INTF_DP, + .controller_id = MSM_DP_CONTROLLER_0, + .prog_fetch_lines_worst_case = 24, + .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 24), + .intr_vsync = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 25), + }, { + .name = "intf_1", .id = INTF_1, + .base = 0x35000, .len = 0x300, + .features = INTF_SC7280_MASK, + .type = INTF_DSI, + .controller_id = MSM_DSI_CONTROLLER_0, + .prog_fetch_lines_worst_case = 24, + .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 26), + .intr_vsync = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 27), + .intr_tear_rd_ptr = DPU_IRQ_IDX(MDP_INTF1_TEAR_INTR, 2), + }, { + .name = "intf_2", .id = INTF_2, + .base = 0x36000, .len = 0x300, + .features = INTF_SC7280_MASK, + .type = INTF_DSI, + .controller_id = MSM_DSI_CONTROLLER_1, + .prog_fetch_lines_worst_case = 24, + .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 28), + .intr_vsync = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 29), + .intr_tear_rd_ptr = DPU_IRQ_IDX(MDP_INTF2_TEAR_INTR, 2), + }, { + .name = "intf_3", .id = INTF_3, + .base = 0x37000, .len = 0x280, + .features = INTF_SC7280_MASK, + .type = INTF_DP, + .controller_id = MSM_DP_CONTROLLER_1, + .prog_fetch_lines_worst_case = 24, + .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 30), + .intr_vsync = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 31), + }, +}; + +static const struct dpu_perf_cfg sar2130p_perf_data = { + .max_bw_low = 13600000, + .max_bw_high = 18200000, + .min_core_ib = 2500000, + .min_llcc_ib = 0, + .min_dram_ib = 800000, + .min_prefill_lines = 35, + /* FIXME: lut tables */ + .danger_lut_tbl = {0x3ffff, 0x3ffff, 0x0}, + .safe_lut_tbl = {0xfe00, 0xfe00, 0xffff}, + .qos_lut_tbl = { + {.nentry = ARRAY_SIZE(sc7180_qos_linear), + .entries = sc7180_qos_linear + }, + {.nentry = ARRAY_SIZE(sc7180_qos_macrotile), + .entries = sc7180_qos_macrotile + }, + {.nentry = ARRAY_SIZE(sc7180_qos_nrt), + .entries = sc7180_qos_nrt + }, + /* TODO: macrotile-qseed is different from macrotile */ + }, + .cdp_cfg = { + {.rd_enable = 0, .wr_enable = 0}, + {.rd_enable = 0, .wr_enable = 0} + }, + .clk_inefficiency_factor = 105, + .bw_inefficiency_factor = 120, +}; + +static const struct dpu_mdss_version sar2130p_mdss_ver = { + .core_major_ver = 9, + .core_minor_ver = 1, +}; + +const struct dpu_mdss_cfg dpu_sar2130p_cfg = { + .mdss_ver = &sar2130p_mdss_ver, + .caps = &sar2130p_dpu_caps, + .mdp = &sar2130p_mdp, + .cdm = &dpu_cdm_5_x, + .ctl_count = ARRAY_SIZE(sar2130p_ctl), + .ctl = sar2130p_ctl, + .sspp_count = ARRAY_SIZE(sar2130p_sspp), + .sspp = sar2130p_sspp, + .mixer_count = ARRAY_SIZE(sar2130p_lm), + .mixer = sar2130p_lm, + .dspp_count = ARRAY_SIZE(sar2130p_dspp), + .dspp = sar2130p_dspp, + .pingpong_count = ARRAY_SIZE(sar2130p_pp), + .pingpong = sar2130p_pp, + .dsc_count = ARRAY_SIZE(sar2130p_dsc), + .dsc = sar2130p_dsc, + .merge_3d_count = ARRAY_SIZE(sar2130p_merge_3d), + .merge_3d = sar2130p_merge_3d, + .wb_count = ARRAY_SIZE(sar2130p_wb), + .wb = sar2130p_wb, + .intf_count = ARRAY_SIZE(sar2130p_intf), + .intf = sar2130p_intf, + .vbif_count = ARRAY_SIZE(sm8550_vbif), + .vbif = sm8550_vbif, + .perf = &sar2130p_perf_data, +}; + +#endif diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_2_x1e80100.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_2_x1e80100.h index 6417baa84f82..52cc10aec1f9 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_2_x1e80100.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_9_2_x1e80100.h @@ -26,17 +26,16 @@ static const struct dpu_mdp_cfg x1e80100_mdp = { }, }; -/* FIXME: get rid of DPU_CTL_SPLIT_DISPLAY in favour of proper ACTIVE_CTL support */ static const struct dpu_ctl_cfg x1e80100_ctl[] = { { .name = "ctl_0", .id = CTL_0, .base = 0x15000, .len = 0x290, - .features = CTL_SM8550_MASK | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = CTL_SM8550_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 9), }, { .name = "ctl_1", .id = CTL_1, .base = 0x16000, .len = 0x290, - .features = CTL_SM8550_MASK | BIT(DPU_CTL_SPLIT_DISPLAY), + .features = CTL_SM8550_MASK, .intr_start = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR2, 10), }, { .name = "ctl_2", .id = CTL_2, diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c index 0714936d8835..a4b0fe0d9899 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c @@ -445,9 +445,9 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, uint32_t lm_idx; bool bg_alpha_enable = false; - DECLARE_BITMAP(fetch_active, SSPP_MAX); + DECLARE_BITMAP(active_fetch, SSPP_MAX); - memset(fetch_active, 0, sizeof(fetch_active)); + memset(active_fetch, 0, sizeof(active_fetch)); drm_atomic_crtc_for_each_plane(plane, crtc) { state = plane->state; if (!state) @@ -464,7 +464,7 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, if (pstate->stage == DPU_STAGE_BASE && format->alpha_enable) bg_alpha_enable = true; - set_bit(pstate->pipe.sspp->idx, fetch_active); + set_bit(pstate->pipe.sspp->idx, active_fetch); _dpu_crtc_blend_setup_pipe(crtc, plane, mixer, cstate->num_mixers, pstate->stage, @@ -472,7 +472,7 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, &pstate->pipe, 0, stage_cfg); if (pstate->r_pipe.sspp) { - set_bit(pstate->r_pipe.sspp->idx, fetch_active); + set_bit(pstate->r_pipe.sspp->idx, active_fetch); _dpu_crtc_blend_setup_pipe(crtc, plane, mixer, cstate->num_mixers, pstate->stage, @@ -492,8 +492,8 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, } } - if (ctl->ops.set_active_pipes) - ctl->ops.set_active_pipes(ctl, fetch_active); + if (ctl->ops.set_active_fetch_pipes) + ctl->ops.set_active_fetch_pipes(ctl, active_fetch); _dpu_crtc_program_lm_output_roi(crtc); } @@ -519,6 +519,8 @@ static void _dpu_crtc_blend_setup(struct drm_crtc *crtc) if (mixer[i].lm_ctl->ops.clear_all_blendstages) mixer[i].lm_ctl->ops.clear_all_blendstages( mixer[i].lm_ctl); + if (mixer[i].lm_ctl->ops.set_active_fetch_pipes) + mixer[i].lm_ctl->ops.set_active_fetch_pipes(mixer[i].lm_ctl, NULL); } /* initialize stage cfg */ diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c index 862e9e6bf0a5..7020098360e4 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c @@ -1246,7 +1246,11 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc, return; } - phys->hw_ctl = i < num_ctl ? to_dpu_hw_ctl(hw_ctl[i]) : NULL; + /* Use first (and only) CTL if active CTLs are supported */ + if (num_ctl == 1) + phys->hw_ctl = to_dpu_hw_ctl(hw_ctl[0]); + else + phys->hw_ctl = i < num_ctl ? to_dpu_hw_ctl(hw_ctl[i]) : NULL; if (!phys->hw_ctl) { DPU_ERROR_ENC(dpu_enc, "no ctl block assigned at idx: %d\n", i); @@ -2190,6 +2194,9 @@ static void dpu_encoder_helper_reset_mixers(struct dpu_encoder_phys *phys_enc) /* clear all blendstages */ if (ctl->ops.setup_blendstage) ctl->ops.setup_blendstage(ctl, hw_mixer[i]->idx, NULL); + + if (ctl->ops.set_active_fetch_pipes) + ctl->ops.set_active_fetch_pipes(ctl, NULL); } } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c index da9994a79ca2..a0ba55ab3c89 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c @@ -60,6 +60,8 @@ static void _dpu_encoder_phys_cmd_update_intf_cfg( return; intf_cfg.intf = phys_enc->hw_intf->idx; + if (phys_enc->split_role == ENC_ROLE_MASTER) + intf_cfg.intf_master = phys_enc->hw_intf->idx; intf_cfg.intf_mode_sel = DPU_CTL_MODE_SEL_CMD; intf_cfg.stream_sel = cmd_enc->stream_sel; intf_cfg.mode_3d = dpu_encoder_helper_get_3d_blend_mode(phys_enc); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c index abd6600046cb..8a618841e3ea 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c @@ -298,6 +298,8 @@ static void dpu_encoder_phys_vid_setup_timing_engine( if (phys_enc->hw_cdm) intf_cfg.cdm = phys_enc->hw_cdm->idx; intf_cfg.intf = phys_enc->hw_intf->idx; + if (phys_enc->split_role == ENC_ROLE_MASTER) + intf_cfg.intf_master = phys_enc->hw_intf->idx; intf_cfg.intf_mode_sel = DPU_CTL_MODE_SEL_VID; intf_cfg.stream_sel = 0; /* Don't care value for video mode */ intf_cfg.mode_3d = dpu_encoder_helper_get_3d_blend_mode(phys_enc); @@ -372,7 +374,8 @@ static void dpu_encoder_phys_vid_underrun_irq(void *arg) static bool dpu_encoder_phys_vid_needs_single_flush( struct dpu_encoder_phys *phys_enc) { - return phys_enc->split_role != ENC_ROLE_SOLO; + return !(phys_enc->hw_ctl->caps->features & BIT(DPU_CTL_ACTIVE_CFG)) && + phys_enc->split_role != ENC_ROLE_SOLO; } static void dpu_encoder_phys_vid_atomic_mode_set( diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c index 64265ca4656a..c878fe196aeb 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c @@ -34,11 +34,11 @@ #define VIG_MSM8998_MASK \ (VIG_MASK | BIT(DPU_SSPP_SCALER_QSEED3_COMPATIBLE)) -#define VIG_SDM845_MASK \ +#define VIG_SDM845_MASK_NO_SDMA \ (VIG_MASK | BIT(DPU_SSPP_QOS_8LVL) | BIT(DPU_SSPP_SCALER_QSEED3_COMPATIBLE)) #define VIG_SDM845_MASK_SDMA \ - (VIG_SDM845_MASK | BIT(DPU_SSPP_SMART_DMA_V2)) + (VIG_SDM845_MASK_NO_SDMA | BIT(DPU_SSPP_SMART_DMA_V2)) #define VIG_QCM2290_MASK (VIG_BASE_MASK | BIT(DPU_SSPP_QOS_8LVL)) @@ -54,24 +54,24 @@ BIT(DPU_SSPP_CDP) | BIT(DPU_SSPP_EXCL_RECT)) #define VIG_SC7280_MASK \ - (VIG_SDM845_MASK | BIT(DPU_SSPP_INLINE_ROTATION)) + (VIG_SDM845_MASK_NO_SDMA | BIT(DPU_SSPP_INLINE_ROTATION)) #define VIG_SC7280_MASK_SDMA \ (VIG_SC7280_MASK | BIT(DPU_SSPP_SMART_DMA_V2)) -#define DMA_SDM845_MASK \ +#define DMA_SDM845_MASK_NO_SDMA \ (BIT(DPU_SSPP_QOS) | BIT(DPU_SSPP_QOS_8LVL) |\ BIT(DPU_SSPP_TS_PREFILL) | BIT(DPU_SSPP_TS_PREFILL_REC1) |\ BIT(DPU_SSPP_CDP) | BIT(DPU_SSPP_EXCL_RECT)) -#define DMA_CURSOR_SDM845_MASK \ - (DMA_SDM845_MASK | BIT(DPU_SSPP_CURSOR)) +#define DMA_CURSOR_SDM845_MASK_NO_SDMA \ + (DMA_SDM845_MASK_NO_SDMA | BIT(DPU_SSPP_CURSOR)) #define DMA_SDM845_MASK_SDMA \ - (DMA_SDM845_MASK | BIT(DPU_SSPP_SMART_DMA_V2)) + (DMA_SDM845_MASK_NO_SDMA | BIT(DPU_SSPP_SMART_DMA_V2)) #define DMA_CURSOR_SDM845_MASK_SDMA \ - (DMA_CURSOR_SDM845_MASK | BIT(DPU_SSPP_SMART_DMA_V2)) + (DMA_CURSOR_SDM845_MASK_NO_SDMA | BIT(DPU_SSPP_SMART_DMA_V2)) #define DMA_CURSOR_MSM8996_MASK \ (DMA_MSM8996_MASK | BIT(DPU_SSPP_CURSOR)) @@ -98,15 +98,9 @@ #define PINGPONG_MSM8996_MASK \ (BIT(DPU_PINGPONG_DSC)) -#define PINGPONG_MSM8996_TE2_MASK \ - (PINGPONG_MSM8996_MASK | BIT(DPU_PINGPONG_TE2)) - #define PINGPONG_SDM845_MASK \ (BIT(DPU_PINGPONG_DITHER) | BIT(DPU_PINGPONG_DSC)) -#define PINGPONG_SDM845_TE2_MASK \ - (PINGPONG_SDM845_MASK | BIT(DPU_PINGPONG_TE2)) - #define PINGPONG_SM8150_MASK \ (BIT(DPU_PINGPONG_DITHER) | BIT(DPU_PINGPONG_DSC)) @@ -376,8 +370,6 @@ static const struct dpu_sspp_sub_blks dpu_dma_sblk = _DMA_SBLK(); * MIXER sub blocks config *************************************************************/ -/* MSM8998 */ - static const struct dpu_lm_sub_blks msm8998_lm_sblk = { .maxwidth = DEFAULT_DPU_OUTPUT_LINE_WIDTH, .maxblendstages = 7, /* excluding base layer */ @@ -387,8 +379,6 @@ static const struct dpu_lm_sub_blks msm8998_lm_sblk = { }, }; -/* SDM845 */ - static const struct dpu_lm_sub_blks sdm845_lm_sblk = { .maxwidth = DEFAULT_DPU_OUTPUT_LINE_WIDTH, .maxblendstages = 11, /* excluding base layer */ @@ -398,8 +388,6 @@ static const struct dpu_lm_sub_blks sdm845_lm_sblk = { }, }; -/* SC7180 */ - static const struct dpu_lm_sub_blks sc7180_lm_sblk = { .maxwidth = DEFAULT_DPU_OUTPUT_LINE_WIDTH, .maxblendstages = 7, /* excluding base layer */ @@ -408,8 +396,6 @@ static const struct dpu_lm_sub_blks sc7180_lm_sblk = { }, }; -/* QCM2290 */ - static const struct dpu_lm_sub_blks qcm2290_lm_sblk = { .maxwidth = DEFAULT_DPU_LINE_WIDTH, .maxblendstages = 4, /* excluding base layer */ @@ -434,22 +420,11 @@ static const struct dpu_dspp_sub_blks sdm845_dspp_sblk = { /************************************************************* * PINGPONG sub blocks config *************************************************************/ -static const struct dpu_pingpong_sub_blks msm8996_pp_sblk_te = { - .te2 = {.name = "te2", .base = 0x2000, .len = 0x0, - .version = 0x1}, -}; static const struct dpu_pingpong_sub_blks msm8996_pp_sblk = { /* No dither block */ }; -static const struct dpu_pingpong_sub_blks sdm845_pp_sblk_te = { - .te2 = {.name = "te2", .base = 0x2000, .len = 0x0, - .version = 0x1}, - .dither = {.name = "dither", .base = 0x30e0, - .len = 0x20, .version = 0x10000}, -}; - static const struct dpu_pingpong_sub_blks sdm845_pp_sblk = { .dither = {.name = "dither", .base = 0x30e0, .len = 0x20, .version = 0x10000}, @@ -759,7 +734,7 @@ static const struct dpu_qos_lut_entry sc7180_qos_nrt[] = { #include "catalog/dpu_8_4_sa8775p.h" #include "catalog/dpu_9_0_sm8550.h" - +#include "catalog/dpu_9_1_sar2130p.h" #include "catalog/dpu_9_2_x1e80100.h" #include "catalog/dpu_10_0_sm8650.h" diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h index 4cea19e1a203..01dd6e65f777 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h @@ -115,7 +115,6 @@ enum { /** * PINGPONG sub-blocks - * @DPU_PINGPONG_TE2 Additional tear check block for split pipes * @DPU_PINGPONG_SPLIT PP block supports split fifo * @DPU_PINGPONG_SLAVE PP block is a suitable slave for split fifo * @DPU_PINGPONG_DITHER Dither blocks @@ -123,8 +122,7 @@ enum { * @DPU_PINGPONG_MAX */ enum { - DPU_PINGPONG_TE2 = 0x1, - DPU_PINGPONG_SPLIT, + DPU_PINGPONG_SPLIT = 0x1, DPU_PINGPONG_SLAVE, DPU_PINGPONG_DITHER, DPU_PINGPONG_DSC, @@ -404,8 +402,6 @@ struct dpu_dspp_sub_blks { }; struct dpu_pingpong_sub_blks { - struct dpu_pp_blk te; - struct dpu_pp_blk te2; struct dpu_pp_blk dither; }; @@ -841,6 +837,7 @@ extern const struct dpu_mdss_cfg dpu_msm8937_cfg; extern const struct dpu_mdss_cfg dpu_msm8953_cfg; extern const struct dpu_mdss_cfg dpu_msm8996_cfg; extern const struct dpu_mdss_cfg dpu_msm8998_cfg; +extern const struct dpu_mdss_cfg dpu_sar2130p_cfg; extern const struct dpu_mdss_cfg dpu_sdm630_cfg; extern const struct dpu_mdss_cfg dpu_sdm660_cfg; extern const struct dpu_mdss_cfg dpu_sdm845_cfg; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c index 411a7cf088eb..573e42b06ad0 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c @@ -261,6 +261,12 @@ static void dpu_hw_ctl_update_pending_flush_mixer(struct dpu_hw_ctl *ctx, case LM_5: ctx->pending_flush_mask |= BIT(20); break; + case LM_6: + ctx->pending_flush_mask |= BIT(21); + break; + case LM_7: + ctx->pending_flush_mask |= BIT(27); + break; default: break; } @@ -563,6 +569,7 @@ static void dpu_hw_ctl_intf_cfg_v1(struct dpu_hw_ctl *ctx, u32 wb_active = 0; u32 cwb_active = 0; u32 mode_sel = 0; + u32 merge_3d_active = 0; /* CTL_TOP[31:28] carries group_id to collate CTL paths * per VM. Explicitly disable it until VM support is @@ -578,6 +585,7 @@ static void dpu_hw_ctl_intf_cfg_v1(struct dpu_hw_ctl *ctx, wb_active = DPU_REG_READ(c, CTL_WB_ACTIVE); cwb_active = DPU_REG_READ(c, CTL_CWB_ACTIVE); dsc_active = DPU_REG_READ(c, CTL_DSC_ACTIVE); + merge_3d_active = DPU_REG_READ(c, CTL_MERGE_3D_ACTIVE); if (cfg->intf) intf_active |= BIT(cfg->intf - INTF_0); @@ -591,15 +599,18 @@ static void dpu_hw_ctl_intf_cfg_v1(struct dpu_hw_ctl *ctx, if (cfg->dsc) dsc_active |= cfg->dsc; + if (cfg->merge_3d) + merge_3d_active |= BIT(cfg->merge_3d - MERGE_3D_0); + DPU_REG_WRITE(c, CTL_TOP, mode_sel); DPU_REG_WRITE(c, CTL_INTF_ACTIVE, intf_active); DPU_REG_WRITE(c, CTL_WB_ACTIVE, wb_active); DPU_REG_WRITE(c, CTL_CWB_ACTIVE, cwb_active); DPU_REG_WRITE(c, CTL_DSC_ACTIVE, dsc_active); + DPU_REG_WRITE(c, CTL_MERGE_3D_ACTIVE, merge_3d_active); - if (cfg->merge_3d) - DPU_REG_WRITE(c, CTL_MERGE_3D_ACTIVE, - BIT(cfg->merge_3d - MERGE_3D_0)); + if (cfg->intf_master) + DPU_REG_WRITE(c, CTL_INTF_MASTER, BIT(cfg->intf_master - INTF_0)); if (cfg->cdm) DPU_REG_WRITE(c, CTL_CDM_ACTIVE, cfg->cdm); @@ -643,6 +654,7 @@ static void dpu_hw_ctl_reset_intf_cfg_v1(struct dpu_hw_ctl *ctx, { struct dpu_hw_blk_reg_map *c = &ctx->hw; u32 intf_active = 0; + u32 intf_master = 0; u32 wb_active = 0; u32 cwb_active = 0; u32 merge3d_active = 0; @@ -666,10 +678,21 @@ static void dpu_hw_ctl_reset_intf_cfg_v1(struct dpu_hw_ctl *ctx, dpu_hw_ctl_clear_all_blendstages(ctx); + if (ctx->ops.set_active_fetch_pipes) + ctx->ops.set_active_fetch_pipes(ctx, NULL); + if (cfg->intf) { intf_active = DPU_REG_READ(c, CTL_INTF_ACTIVE); intf_active &= ~BIT(cfg->intf - INTF_0); DPU_REG_WRITE(c, CTL_INTF_ACTIVE, intf_active); + + intf_master = DPU_REG_READ(c, CTL_INTF_MASTER); + + /* Unset this intf as master, if it is the current master */ + if (intf_master == BIT(cfg->intf - INTF_0)) { + DPU_DEBUG_DRIVER("Unsetting INTF_%d master\n", cfg->intf - INTF_0); + DPU_REG_WRITE(c, CTL_INTF_MASTER, 0); + } } if (cfg->cwb) { @@ -697,8 +720,8 @@ static void dpu_hw_ctl_reset_intf_cfg_v1(struct dpu_hw_ctl *ctx, } } -static void dpu_hw_ctl_set_fetch_pipe_active(struct dpu_hw_ctl *ctx, - unsigned long *fetch_active) +static void dpu_hw_ctl_set_active_fetch_pipes(struct dpu_hw_ctl *ctx, + unsigned long *fetch_active) { int i; u32 val = 0; @@ -761,7 +784,7 @@ static void _setup_ctl_ops(struct dpu_hw_ctl_ops *ops, ops->update_pending_flush_dspp = dpu_hw_ctl_update_pending_flush_dspp; if (cap & BIT(DPU_CTL_FETCH_ACTIVE)) - ops->set_active_pipes = dpu_hw_ctl_set_fetch_pipe_active; + ops->set_active_fetch_pipes = dpu_hw_ctl_set_active_fetch_pipes; }; /** diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h index 080a9550a0cc..feb09590bc8f 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h @@ -36,6 +36,7 @@ struct dpu_hw_stage_cfg { /** * struct dpu_hw_intf_cfg :Describes how the DPU writes data to output interface * @intf : Interface id + * @intf_master: Master interface id in the dual pipe topology * @mode_3d: 3d mux configuration * @merge_3d: 3d merge block used * @intf_mode_sel: Interface mode, cmd / vid @@ -46,6 +47,7 @@ struct dpu_hw_stage_cfg { */ struct dpu_hw_intf_cfg { enum dpu_intf intf; + enum dpu_intf intf_master; enum dpu_wb wb; enum dpu_3d_blend_mode mode_3d; enum dpu_merge_3d merge_3d; @@ -254,7 +256,7 @@ struct dpu_hw_ctl_ops { void (*setup_blendstage)(struct dpu_hw_ctl *ctx, enum dpu_lm lm, struct dpu_hw_stage_cfg *cfg); - void (*set_active_pipes)(struct dpu_hw_ctl *ctx, + void (*set_active_fetch_pipes)(struct dpu_hw_ctl *ctx, unsigned long *fetch_active); }; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h index 8d820cd1b554..175639c8bfbb 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h @@ -125,6 +125,7 @@ enum dpu_lm { LM_4, LM_5, LM_6, + LM_7, LM_MAX }; @@ -169,6 +170,8 @@ enum dpu_dsc { DSC_3, DSC_4, DSC_5, + DSC_6, + DSC_7, DSC_MAX }; @@ -185,6 +188,8 @@ enum dpu_pingpong { PINGPONG_3, PINGPONG_4, PINGPONG_5, + PINGPONG_6, + PINGPONG_7, PINGPONG_CWB_0, PINGPONG_CWB_1, PINGPONG_CWB_2, @@ -199,6 +204,7 @@ enum dpu_merge_3d { MERGE_3D_2, MERGE_3D_3, MERGE_3D_4, + MERGE_3D_5, MERGE_3D_MAX }; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c index 3305ad0623ca..1fd82b6747e9 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c @@ -1512,6 +1512,7 @@ static const struct of_device_id dpu_dt_match[] = { { .compatible = "qcom,msm8998-dpu", .data = &dpu_msm8998_cfg, }, { .compatible = "qcom,qcm2290-dpu", .data = &dpu_qcm2290_cfg, }, { .compatible = "qcom,sa8775p-dpu", .data = &dpu_sa8775p_cfg, }, + { .compatible = "qcom,sar2130p-dpu", .data = &dpu_sar2130p_cfg, }, { .compatible = "qcom,sdm630-mdp5", .data = &dpu_sdm630_cfg, }, { .compatible = "qcom,sdm660-mdp5", .data = &dpu_sdm660_cfg, }, { .compatible = "qcom,sdm670-dpu", .data = &dpu_sdm670_cfg, }, diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c index e03d6091f736..421138bc3cb7 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c @@ -915,10 +915,9 @@ static int dpu_plane_atomic_check_nosspp(struct drm_plane *plane, return 0; } -static int dpu_plane_is_multirect_parallel_capable(struct dpu_hw_sspp *sspp, - struct dpu_sw_pipe_cfg *pipe_cfg, - const struct msm_format *fmt, - uint32_t max_linewidth) +static int dpu_plane_is_multirect_capable(struct dpu_hw_sspp *sspp, + struct dpu_sw_pipe_cfg *pipe_cfg, + const struct msm_format *fmt) { if (drm_rect_width(&pipe_cfg->src_rect) != drm_rect_width(&pipe_cfg->dst_rect) || drm_rect_height(&pipe_cfg->src_rect) != drm_rect_height(&pipe_cfg->dst_rect)) @@ -930,10 +929,6 @@ static int dpu_plane_is_multirect_parallel_capable(struct dpu_hw_sspp *sspp, if (MSM_FORMAT_IS_YUV(fmt)) return false; - if (MSM_FORMAT_IS_UBWC(fmt) && - drm_rect_width(&pipe_cfg->src_rect) > max_linewidth / 2) - return false; - if (!test_bit(DPU_SSPP_SMART_DMA_V1, &sspp->cap->features) && !test_bit(DPU_SSPP_SMART_DMA_V2, &sspp->cap->features)) return false; @@ -941,6 +936,27 @@ static int dpu_plane_is_multirect_parallel_capable(struct dpu_hw_sspp *sspp, return true; } +static int dpu_plane_is_parallel_capable(struct dpu_sw_pipe_cfg *pipe_cfg, + const struct msm_format *fmt, + uint32_t max_linewidth) +{ + if (MSM_FORMAT_IS_UBWC(fmt) && + drm_rect_width(&pipe_cfg->src_rect) > max_linewidth / 2) + return false; + + return true; +} + +static int dpu_plane_is_multirect_parallel_capable(struct dpu_hw_sspp *sspp, + struct dpu_sw_pipe_cfg *pipe_cfg, + const struct msm_format *fmt, + uint32_t max_linewidth) +{ + return dpu_plane_is_multirect_capable(sspp, pipe_cfg, fmt) && + dpu_plane_is_parallel_capable(pipe_cfg, fmt, max_linewidth); +} + + static int dpu_plane_atomic_check_sspp(struct drm_plane *plane, struct drm_atomic_state *state, const struct drm_crtc_state *crtc_state) @@ -1002,6 +1018,69 @@ static bool dpu_plane_try_multirect_parallel(struct dpu_sw_pipe *pipe, struct dp return true; } +static int dpu_plane_try_multirect_shared(struct dpu_plane_state *pstate, + struct dpu_plane_state *prev_adjacent_pstate, + const struct msm_format *fmt, + uint32_t max_linewidth) +{ + struct dpu_sw_pipe *pipe = &pstate->pipe; + struct dpu_sw_pipe *r_pipe = &pstate->r_pipe; + struct dpu_sw_pipe_cfg *pipe_cfg = &pstate->pipe_cfg; + struct dpu_sw_pipe *prev_pipe = &prev_adjacent_pstate->pipe; + struct dpu_sw_pipe_cfg *prev_pipe_cfg = &prev_adjacent_pstate->pipe_cfg; + const struct msm_format *prev_fmt = msm_framebuffer_format(prev_adjacent_pstate->base.fb); + u16 max_tile_height = 1; + + if (prev_adjacent_pstate->r_pipe.sspp != NULL || + prev_pipe->multirect_mode != DPU_SSPP_MULTIRECT_NONE) + return false; + + if (!dpu_plane_is_multirect_capable(pipe->sspp, pipe_cfg, fmt) || + !dpu_plane_is_multirect_capable(prev_pipe->sspp, prev_pipe_cfg, prev_fmt)) + return false; + + if (MSM_FORMAT_IS_UBWC(fmt)) + max_tile_height = max(max_tile_height, fmt->tile_height); + + if (MSM_FORMAT_IS_UBWC(prev_fmt)) + max_tile_height = max(max_tile_height, prev_fmt->tile_height); + + r_pipe->multirect_index = DPU_SSPP_RECT_SOLO; + r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + + r_pipe->sspp = NULL; + + if (dpu_plane_is_parallel_capable(pipe_cfg, fmt, max_linewidth) && + dpu_plane_is_parallel_capable(prev_pipe_cfg, prev_fmt, max_linewidth) && + (pipe_cfg->dst_rect.x1 >= prev_pipe_cfg->dst_rect.x2 || + prev_pipe_cfg->dst_rect.x1 >= pipe_cfg->dst_rect.x2)) { + pipe->sspp = prev_pipe->sspp; + + pipe->multirect_index = DPU_SSPP_RECT_1; + pipe->multirect_mode = DPU_SSPP_MULTIRECT_PARALLEL; + + prev_pipe->multirect_index = DPU_SSPP_RECT_0; + prev_pipe->multirect_mode = DPU_SSPP_MULTIRECT_PARALLEL; + + return true; + } + + if (pipe_cfg->dst_rect.y1 >= prev_pipe_cfg->dst_rect.y2 + 2 * max_tile_height || + prev_pipe_cfg->dst_rect.y1 >= pipe_cfg->dst_rect.y2 + 2 * max_tile_height) { + pipe->sspp = prev_pipe->sspp; + + pipe->multirect_index = DPU_SSPP_RECT_1; + pipe->multirect_mode = DPU_SSPP_MULTIRECT_TIME_MX; + + prev_pipe->multirect_index = DPU_SSPP_RECT_0; + prev_pipe->multirect_mode = DPU_SSPP_MULTIRECT_TIME_MX; + + return true; + } + + return false; +} + static int dpu_plane_atomic_check(struct drm_plane *plane, struct drm_atomic_state *state) { @@ -1102,13 +1181,14 @@ static int dpu_plane_virtual_atomic_check(struct drm_plane *plane, static int dpu_plane_virtual_assign_resources(struct drm_crtc *crtc, struct dpu_global_state *global_state, struct drm_atomic_state *state, - struct drm_plane_state *plane_state) + struct drm_plane_state *plane_state, + struct drm_plane_state *prev_adjacent_plane_state) { const struct drm_crtc_state *crtc_state = NULL; struct drm_plane *plane = plane_state->plane; struct dpu_kms *dpu_kms = _dpu_plane_get_kms(plane); struct dpu_rm_sspp_requirements reqs; - struct dpu_plane_state *pstate; + struct dpu_plane_state *pstate, *prev_adjacent_pstate; struct dpu_sw_pipe *pipe; struct dpu_sw_pipe *r_pipe; struct dpu_sw_pipe_cfg *pipe_cfg; @@ -1120,6 +1200,8 @@ static int dpu_plane_virtual_assign_resources(struct drm_crtc *crtc, plane_state->crtc); pstate = to_dpu_plane_state(plane_state); + prev_adjacent_pstate = prev_adjacent_plane_state ? + to_dpu_plane_state(prev_adjacent_plane_state) : NULL; pipe = &pstate->pipe; r_pipe = &pstate->r_pipe; pipe_cfg = &pstate->pipe_cfg; @@ -1138,24 +1220,42 @@ static int dpu_plane_virtual_assign_resources(struct drm_crtc *crtc, reqs.rot90 = drm_rotation_90_or_270(plane_state->rotation); - pipe->sspp = dpu_rm_reserve_sspp(&dpu_kms->rm, global_state, crtc, &reqs); - if (!pipe->sspp) - return -ENODEV; + if (drm_rect_width(&r_pipe_cfg->src_rect) == 0) { + if (!prev_adjacent_pstate || + !dpu_plane_try_multirect_shared(pstate, prev_adjacent_pstate, fmt, + dpu_kms->catalog->caps->max_linewidth)) { + pipe->sspp = dpu_rm_reserve_sspp(&dpu_kms->rm, global_state, crtc, &reqs); + if (!pipe->sspp) + return -ENODEV; - if (!dpu_plane_try_multirect_parallel(pipe, pipe_cfg, r_pipe, r_pipe_cfg, - pipe->sspp, - msm_framebuffer_format(plane_state->fb), - dpu_kms->catalog->caps->max_linewidth)) { - /* multirect is not possible, use two SSPP blocks */ - r_pipe->sspp = dpu_rm_reserve_sspp(&dpu_kms->rm, global_state, crtc, &reqs); - if (!r_pipe->sspp) + r_pipe->sspp = NULL; + + pipe->multirect_index = DPU_SSPP_RECT_SOLO; + pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + + r_pipe->multirect_index = DPU_SSPP_RECT_SOLO; + r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + } + } else { + pipe->sspp = dpu_rm_reserve_sspp(&dpu_kms->rm, global_state, crtc, &reqs); + if (!pipe->sspp) return -ENODEV; - pipe->multirect_index = DPU_SSPP_RECT_SOLO; - pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + if (!dpu_plane_try_multirect_parallel(pipe, pipe_cfg, r_pipe, r_pipe_cfg, + pipe->sspp, + msm_framebuffer_format(plane_state->fb), + dpu_kms->catalog->caps->max_linewidth)) { + /* multirect is not possible, use two SSPP blocks */ + r_pipe->sspp = dpu_rm_reserve_sspp(&dpu_kms->rm, global_state, crtc, &reqs); + if (!r_pipe->sspp) + return -ENODEV; - r_pipe->multirect_index = DPU_SSPP_RECT_SOLO; - r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + pipe->multirect_index = DPU_SSPP_RECT_SOLO; + pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + + r_pipe->multirect_index = DPU_SSPP_RECT_SOLO; + r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE; + } } return dpu_plane_atomic_check_sspp(plane, state, crtc_state); @@ -1168,6 +1268,7 @@ int dpu_assign_plane_resources(struct dpu_global_state *global_state, unsigned int num_planes) { unsigned int i; + struct drm_plane_state *prev_adjacent_plane_state = NULL; for (i = 0; i < num_planes; i++) { struct drm_plane_state *plane_state = states[i]; @@ -1177,9 +1278,12 @@ int dpu_assign_plane_resources(struct dpu_global_state *global_state, continue; int ret = dpu_plane_virtual_assign_resources(crtc, global_state, - state, plane_state); + state, plane_state, + prev_adjacent_plane_state); if (ret) - return ret; + break; + + prev_adjacent_plane_state = plane_state; } return 0; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c index 3efbba425ca6..2e296f79cba1 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c @@ -53,6 +53,8 @@ int dpu_rm_init(struct drm_device *dev, /* Clear, setup lists */ memset(rm, 0, sizeof(*rm)); + rm->has_legacy_ctls = (cat->mdss_ver->core_major_ver < 5); + /* Interrogate HW catalog and create tracking items for hw blocks */ for (i = 0; i < cat->mixer_count; i++) { struct dpu_hw_mixer *hw; @@ -434,20 +436,19 @@ static int _dpu_rm_reserve_ctls( int i = 0, j, num_ctls; bool needs_split_display; - /* - * For non-CWB mode, each hw_intf needs its own hw_ctl to program its - * control path. - * - * Hardcode num_ctls to 1 if CWB is enabled because in CWB, both the - * writeback and real-time encoders must be driven by the same control - * path - */ - if (top->cwb_enabled) - num_ctls = 1; - else + if (rm->has_legacy_ctls) { + /* + * TODO: check if there is a need for special handling if + * DPU < 5.0 get CWB support. + */ num_ctls = top->num_intf; - needs_split_display = _dpu_rm_needs_split_display(top); + needs_split_display = _dpu_rm_needs_split_display(top); + } else { + /* use single CTL */ + num_ctls = 1; + needs_split_display = false; + } for (j = 0; j < ARRAY_SIZE(rm->ctl_blks); j++) { const struct dpu_hw_ctl *ctl; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h index a19dbdb1b6f4..aa62966056d4 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h @@ -24,6 +24,7 @@ struct dpu_global_state; * @dspp_blks: array of dspp hardware resources * @hw_sspp: array of sspp hardware resources * @cdm_blk: cdm hardware resource + * @has_legacy_ctls: DPU uses pre-ACTIVE CTL blocks. */ struct dpu_rm { struct dpu_hw_blk *pingpong_blks[PINGPONG_MAX - PINGPONG_0]; @@ -37,6 +38,7 @@ struct dpu_rm { struct dpu_hw_blk *dsc_blks[DSC_MAX - DSC_0]; struct dpu_hw_sspp *hw_sspp[SSPP_MAX - SSPP_NONE]; struct dpu_hw_blk *cdm_blk; + bool has_legacy_ctls; }; struct dpu_rm_sspp_requirements { diff --git a/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.c b/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.c index c469e66cfc11..7e942c1337b3 100644 --- a/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.c +++ b/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.c @@ -6,6 +6,8 @@ #include <linux/delay.h> +#include <drm/drm_bridge.h> +#include <drm/drm_bridge_connector.h> #include <drm/drm_vblank.h> #include "msm_drv.h" @@ -189,7 +191,7 @@ static int mdp4_modeset_init_intf(struct mdp4_kms *mdp4_kms, struct msm_drm_private *priv = dev->dev_private; struct drm_encoder *encoder; struct drm_connector *connector; - struct device_node *panel_node; + struct drm_bridge *next_bridge; int dsi_id; int ret; @@ -199,27 +201,43 @@ static int mdp4_modeset_init_intf(struct mdp4_kms *mdp4_kms, * bail out early if there is no panel node (no need to * initialize LCDC encoder and LVDS connector) */ - panel_node = of_graph_get_remote_node(dev->dev->of_node, 0, 0); - if (!panel_node) - return 0; + next_bridge = devm_drm_of_get_bridge(dev->dev, dev->dev->of_node, 0, 0); + if (IS_ERR(next_bridge)) { + ret = PTR_ERR(next_bridge); + if (ret == -ENODEV) + return 0; + return ret; + } - encoder = mdp4_lcdc_encoder_init(dev, panel_node); + encoder = mdp4_lcdc_encoder_init(dev); if (IS_ERR(encoder)) { DRM_DEV_ERROR(dev->dev, "failed to construct LCDC encoder\n"); - of_node_put(panel_node); return PTR_ERR(encoder); } /* LCDC can be hooked to DMA_P (TODO: Add DMA_S later?) */ encoder->possible_crtcs = 1 << DMA_P; - connector = mdp4_lvds_connector_init(dev, panel_node, encoder); + ret = drm_bridge_attach(encoder, next_bridge, NULL, DRM_BRIDGE_ATTACH_NO_CONNECTOR); + if (ret) { + DRM_DEV_ERROR(dev->dev, "failed to attach LVDS panel/bridge: %d\n", ret); + + return ret; + } + + connector = drm_bridge_connector_init(dev, encoder); if (IS_ERR(connector)) { DRM_DEV_ERROR(dev->dev, "failed to initialize LVDS connector\n"); - of_node_put(panel_node); return PTR_ERR(connector); } + ret = drm_connector_attach_encoder(connector, encoder); + if (ret) { + DRM_DEV_ERROR(dev->dev, "failed to attach LVDS connector: %d\n", ret); + + return ret; + } + break; case DRM_MODE_ENCODER_TMDS: encoder = mdp4_dtv_encoder_init(dev); diff --git a/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.h b/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.h index 94b1ba92785f..f9d988076337 100644 --- a/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.h +++ b/drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.h @@ -191,12 +191,7 @@ struct drm_crtc *mdp4_crtc_init(struct drm_device *dev, long mdp4_dtv_round_pixclk(struct drm_encoder *encoder, unsigned long rate); struct drm_encoder *mdp4_dtv_encoder_init(struct drm_device *dev); -long mdp4_lcdc_round_pixclk(struct drm_encoder *encoder, unsigned long rate); -struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev, - struct device_node *panel_node); - -struct drm_connector *mdp4_lvds_connector_init(struct drm_device *dev, - struct device_node *panel_node, struct drm_encoder *encoder); +struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev); #ifdef CONFIG_DRM_MSM_DSI struct drm_encoder *mdp4_dsi_encoder_init(struct drm_device *dev); @@ -207,13 +202,6 @@ static inline struct drm_encoder *mdp4_dsi_encoder_init(struct drm_device *dev) } #endif -#ifdef CONFIG_COMMON_CLK -struct clk *mpd4_lvds_pll_init(struct drm_device *dev); -#else -static inline struct clk *mpd4_lvds_pll_init(struct drm_device *dev) -{ - return ERR_PTR(-ENODEV); -} -#endif +struct clk *mpd4_get_lcdc_clock(struct drm_device *dev); #endif /* __MDP4_KMS_H__ */ diff --git a/drivers/gpu/drm/msm/disp/mdp4/mdp4_lcdc_encoder.c b/drivers/gpu/drm/msm/disp/mdp4/mdp4_lcdc_encoder.c index 8bbc7fb881d5..06a307c1272d 100644 --- a/drivers/gpu/drm/msm/disp/mdp4/mdp4_lcdc_encoder.c +++ b/drivers/gpu/drm/msm/disp/mdp4/mdp4_lcdc_encoder.c @@ -14,7 +14,6 @@ struct mdp4_lcdc_encoder { struct drm_encoder base; - struct device_node *panel_node; struct drm_panel *panel; struct clk *lcdc_clk; unsigned long int pixclock; @@ -262,19 +261,12 @@ static void mdp4_lcdc_encoder_disable(struct drm_encoder *encoder) struct mdp4_lcdc_encoder *mdp4_lcdc_encoder = to_mdp4_lcdc_encoder(encoder); struct mdp4_kms *mdp4_kms = get_kms(encoder); - struct drm_panel *panel; if (WARN_ON(!mdp4_lcdc_encoder->enabled)) return; mdp4_write(mdp4_kms, REG_MDP4_LCDC_ENABLE, 0); - panel = of_drm_find_panel(mdp4_lcdc_encoder->panel_node); - if (!IS_ERR(panel)) { - drm_panel_disable(panel); - drm_panel_unprepare(panel); - } - /* * Wait for a vsync so we know the ENABLE=0 latched before * the (connector) source of the vsync's gets disabled, @@ -300,7 +292,6 @@ static void mdp4_lcdc_encoder_enable(struct drm_encoder *encoder) to_mdp4_lcdc_encoder(encoder); unsigned long pc = mdp4_lcdc_encoder->pixclock; struct mdp4_kms *mdp4_kms = get_kms(encoder); - struct drm_panel *panel; uint32_t config; int ret; @@ -335,12 +326,6 @@ static void mdp4_lcdc_encoder_enable(struct drm_encoder *encoder) if (ret) DRM_DEV_ERROR(dev->dev, "failed to enable lcdc_clk: %d\n", ret); - panel = of_drm_find_panel(mdp4_lcdc_encoder->panel_node); - if (!IS_ERR(panel)) { - drm_panel_prepare(panel); - drm_panel_enable(panel); - } - setup_phy(encoder); mdp4_write(mdp4_kms, REG_MDP4_LCDC_ENABLE, 1); @@ -348,22 +333,34 @@ static void mdp4_lcdc_encoder_enable(struct drm_encoder *encoder) mdp4_lcdc_encoder->enabled = true; } +static enum drm_mode_status +mdp4_lcdc_encoder_mode_valid(struct drm_encoder *encoder, + const struct drm_display_mode *mode) +{ + struct mdp4_lcdc_encoder *mdp4_lcdc_encoder = + to_mdp4_lcdc_encoder(encoder); + long actual, requested; + + requested = 1000 * mode->clock; + actual = clk_round_rate(mdp4_lcdc_encoder->lcdc_clk, requested); + + DBG("requested=%ld, actual=%ld", requested, actual); + + if (actual != requested) + return MODE_CLOCK_RANGE; + + return MODE_OK; +} + static const struct drm_encoder_helper_funcs mdp4_lcdc_encoder_helper_funcs = { .mode_set = mdp4_lcdc_encoder_mode_set, .disable = mdp4_lcdc_encoder_disable, .enable = mdp4_lcdc_encoder_enable, + .mode_valid = mdp4_lcdc_encoder_mode_valid, }; -long mdp4_lcdc_round_pixclk(struct drm_encoder *encoder, unsigned long rate) -{ - struct mdp4_lcdc_encoder *mdp4_lcdc_encoder = - to_mdp4_lcdc_encoder(encoder); - return clk_round_rate(mdp4_lcdc_encoder->lcdc_clk, rate); -} - /* initialize encoder */ -struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev, - struct device_node *panel_node) +struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev) { struct drm_encoder *encoder; struct mdp4_lcdc_encoder *mdp4_lcdc_encoder; @@ -374,14 +371,11 @@ struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev, if (IS_ERR(mdp4_lcdc_encoder)) return ERR_CAST(mdp4_lcdc_encoder); - mdp4_lcdc_encoder->panel_node = panel_node; - encoder = &mdp4_lcdc_encoder->base; drm_encoder_helper_add(encoder, &mdp4_lcdc_encoder_helper_funcs); - /* TODO: do we need different pll in other cases? */ - mdp4_lcdc_encoder->lcdc_clk = mpd4_lvds_pll_init(dev); + mdp4_lcdc_encoder->lcdc_clk = mpd4_get_lcdc_clock(dev); if (IS_ERR(mdp4_lcdc_encoder->lcdc_clk)) { DRM_DEV_ERROR(dev->dev, "failed to get lvds_clk\n"); return ERR_CAST(mdp4_lcdc_encoder->lcdc_clk); diff --git a/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_connector.c b/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_connector.c deleted file mode 100644 index 52e728181b52..000000000000 --- a/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_connector.c +++ /dev/null @@ -1,121 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0-only -/* - * Copyright (C) 2014 Red Hat - * Author: Rob Clark <robdclark@gmail.com> - * Author: Vinay Simha <vinaysimha@inforcecomputing.com> - */ - -#include "mdp4_kms.h" - -struct mdp4_lvds_connector { - struct drm_connector base; - struct drm_encoder *encoder; - struct device_node *panel_node; - struct drm_panel *panel; -}; -#define to_mdp4_lvds_connector(x) container_of(x, struct mdp4_lvds_connector, base) - -static enum drm_connector_status mdp4_lvds_connector_detect( - struct drm_connector *connector, bool force) -{ - struct mdp4_lvds_connector *mdp4_lvds_connector = - to_mdp4_lvds_connector(connector); - - if (!mdp4_lvds_connector->panel) { - mdp4_lvds_connector->panel = - of_drm_find_panel(mdp4_lvds_connector->panel_node); - if (IS_ERR(mdp4_lvds_connector->panel)) - mdp4_lvds_connector->panel = NULL; - } - - return mdp4_lvds_connector->panel ? - connector_status_connected : - connector_status_disconnected; -} - -static void mdp4_lvds_connector_destroy(struct drm_connector *connector) -{ - struct mdp4_lvds_connector *mdp4_lvds_connector = - to_mdp4_lvds_connector(connector); - - drm_connector_cleanup(connector); - - kfree(mdp4_lvds_connector); -} - -static int mdp4_lvds_connector_get_modes(struct drm_connector *connector) -{ - struct mdp4_lvds_connector *mdp4_lvds_connector = - to_mdp4_lvds_connector(connector); - struct drm_panel *panel = mdp4_lvds_connector->panel; - int ret = 0; - - if (panel) - ret = drm_panel_get_modes(panel, connector); - - return ret; -} - -static enum drm_mode_status -mdp4_lvds_connector_mode_valid(struct drm_connector *connector, - const struct drm_display_mode *mode) -{ - struct mdp4_lvds_connector *mdp4_lvds_connector = - to_mdp4_lvds_connector(connector); - struct drm_encoder *encoder = mdp4_lvds_connector->encoder; - long actual, requested; - - requested = 1000 * mode->clock; - actual = mdp4_lcdc_round_pixclk(encoder, requested); - - DBG("requested=%ld, actual=%ld", requested, actual); - - if (actual != requested) - return MODE_CLOCK_RANGE; - - return MODE_OK; -} - -static const struct drm_connector_funcs mdp4_lvds_connector_funcs = { - .detect = mdp4_lvds_connector_detect, - .fill_modes = drm_helper_probe_single_connector_modes, - .destroy = mdp4_lvds_connector_destroy, - .reset = drm_atomic_helper_connector_reset, - .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, - .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, -}; - -static const struct drm_connector_helper_funcs mdp4_lvds_connector_helper_funcs = { - .get_modes = mdp4_lvds_connector_get_modes, - .mode_valid = mdp4_lvds_connector_mode_valid, -}; - -/* initialize connector */ -struct drm_connector *mdp4_lvds_connector_init(struct drm_device *dev, - struct device_node *panel_node, struct drm_encoder *encoder) -{ - struct drm_connector *connector = NULL; - struct mdp4_lvds_connector *mdp4_lvds_connector; - - mdp4_lvds_connector = kzalloc(sizeof(*mdp4_lvds_connector), GFP_KERNEL); - if (!mdp4_lvds_connector) - return ERR_PTR(-ENOMEM); - - mdp4_lvds_connector->encoder = encoder; - mdp4_lvds_connector->panel_node = panel_node; - - connector = &mdp4_lvds_connector->base; - - drm_connector_init(dev, connector, &mdp4_lvds_connector_funcs, - DRM_MODE_CONNECTOR_LVDS); - drm_connector_helper_add(connector, &mdp4_lvds_connector_helper_funcs); - - connector->polled = 0; - - connector->interlace_allowed = 0; - connector->doublescan_allowed = 0; - - drm_connector_attach_encoder(connector, encoder); - - return connector; -} diff --git a/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_pll.c b/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_pll.c index ab8c0c187fb2..fa2c29470510 100644 --- a/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_pll.c +++ b/drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_pll.c @@ -122,40 +122,59 @@ static const struct clk_ops mpd4_lvds_pll_ops = { .set_rate = mpd4_lvds_pll_set_rate, }; -static const char *mpd4_lvds_pll_parents[] = { - "pxo", +static const struct clk_parent_data mpd4_lvds_pll_parents[] = { + { .fw_name = "pxo", .name = "pxo", }, }; static struct clk_init_data pll_init = { .name = "mpd4_lvds_pll", .ops = &mpd4_lvds_pll_ops, - .parent_names = mpd4_lvds_pll_parents, + .parent_data = mpd4_lvds_pll_parents, .num_parents = ARRAY_SIZE(mpd4_lvds_pll_parents), }; -struct clk *mpd4_lvds_pll_init(struct drm_device *dev) +static struct clk_hw *mpd4_lvds_pll_init(struct drm_device *dev) { struct mdp4_lvds_pll *lvds_pll; - struct clk *clk; int ret; lvds_pll = devm_kzalloc(dev->dev, sizeof(*lvds_pll), GFP_KERNEL); - if (!lvds_pll) { - ret = -ENOMEM; - goto fail; - } + if (!lvds_pll) + return ERR_PTR(-ENOMEM); lvds_pll->dev = dev; lvds_pll->pll_hw.init = &pll_init; - clk = devm_clk_register(dev->dev, &lvds_pll->pll_hw); - if (IS_ERR(clk)) { - ret = PTR_ERR(clk); - goto fail; + ret = devm_clk_hw_register(dev->dev, &lvds_pll->pll_hw); + if (ret) + return ERR_PTR(ret); + + ret = devm_of_clk_add_hw_provider(dev->dev, of_clk_hw_simple_get, &lvds_pll->pll_hw); + if (ret) + return ERR_PTR(ret); + + return &lvds_pll->pll_hw; +} + +struct clk *mpd4_get_lcdc_clock(struct drm_device *dev) +{ + struct clk_hw *hw; + struct clk *clk; + + + /* TODO: do we need different pll in other cases? */ + hw = mpd4_lvds_pll_init(dev); + if (IS_ERR(hw)) { + DRM_DEV_ERROR(dev->dev, "failed to register LVDS PLL\n"); + return ERR_CAST(hw); } - return clk; + clk = devm_clk_get(dev->dev, "lcdc_clk"); + if (clk == ERR_PTR(-ENOENT)) { + drm_warn(dev, "can't get LCDC clock, using PLL directly\n"); -fail: - return ERR_PTR(ret); + return devm_clk_hw_get_clk(dev->dev, hw, "lcdc_clk"); + } + + return clk; } diff --git a/drivers/gpu/drm/msm/dp/dp_audio.c b/drivers/gpu/drm/msm/dp/dp_audio.c index 70fdc9fe228a..f8bfb908f9b4 100644 --- a/drivers/gpu/drm/msm/dp/dp_audio.c +++ b/drivers/gpu/drm/msm/dp/dp_audio.c @@ -13,13 +13,13 @@ #include "dp_catalog.h" #include "dp_audio.h" +#include "dp_drm.h" #include "dp_panel.h" #include "dp_reg.h" #include "dp_display.h" #include "dp_utils.h" struct msm_dp_audio_private { - struct platform_device *audio_pdev; struct platform_device *pdev; struct drm_device *drm_dev; struct msm_dp_catalog *catalog; @@ -160,24 +160,11 @@ static void msm_dp_audio_enable(struct msm_dp_audio_private *audio, bool enable) msm_dp_catalog_audio_enable(catalog, enable); } -static struct msm_dp_audio_private *msm_dp_audio_get_data(struct platform_device *pdev) +static struct msm_dp_audio_private *msm_dp_audio_get_data(struct msm_dp *msm_dp_display) { struct msm_dp_audio *msm_dp_audio; - struct msm_dp *msm_dp_display; - - if (!pdev) { - DRM_ERROR("invalid input\n"); - return ERR_PTR(-ENODEV); - } - - msm_dp_display = platform_get_drvdata(pdev); - if (!msm_dp_display) { - DRM_ERROR("invalid input\n"); - return ERR_PTR(-ENODEV); - } msm_dp_audio = msm_dp_display->msm_dp_audio; - if (!msm_dp_audio) { DRM_ERROR("invalid msm_dp_audio data\n"); return ERR_PTR(-EINVAL); @@ -186,68 +173,16 @@ static struct msm_dp_audio_private *msm_dp_audio_get_data(struct platform_device return container_of(msm_dp_audio, struct msm_dp_audio_private, msm_dp_audio); } -static int msm_dp_audio_hook_plugged_cb(struct device *dev, void *data, - hdmi_codec_plugged_cb fn, - struct device *codec_dev) -{ - - struct platform_device *pdev; - struct msm_dp *msm_dp_display; - - pdev = to_platform_device(dev); - if (!pdev) { - pr_err("invalid input\n"); - return -ENODEV; - } - - msm_dp_display = platform_get_drvdata(pdev); - if (!msm_dp_display) { - pr_err("invalid input\n"); - return -ENODEV; - } - - return msm_dp_display_set_plugged_cb(msm_dp_display, fn, codec_dev); -} - -static int msm_dp_audio_get_eld(struct device *dev, - void *data, uint8_t *buf, size_t len) -{ - struct platform_device *pdev; - struct msm_dp *msm_dp_display; - - pdev = to_platform_device(dev); - - if (!pdev) { - DRM_ERROR("invalid input\n"); - return -ENODEV; - } - - msm_dp_display = platform_get_drvdata(pdev); - if (!msm_dp_display) { - DRM_ERROR("invalid input\n"); - return -ENODEV; - } - - mutex_lock(&msm_dp_display->connector->eld_mutex); - memcpy(buf, msm_dp_display->connector->eld, - min(sizeof(msm_dp_display->connector->eld), len)); - mutex_unlock(&msm_dp_display->connector->eld_mutex); - - return 0; -} - -int msm_dp_audio_hw_params(struct device *dev, - void *data, - struct hdmi_codec_daifmt *daifmt, - struct hdmi_codec_params *params) +int msm_dp_audio_prepare(struct drm_connector *connector, + struct drm_bridge *bridge, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params) { int rc = 0; struct msm_dp_audio_private *audio; - struct platform_device *pdev; struct msm_dp *msm_dp_display; - pdev = to_platform_device(dev); - msm_dp_display = platform_get_drvdata(pdev); + msm_dp_display = to_dp_bridge(bridge)->msm_dp_display; /* * there could be cases where sound card can be opened even @@ -262,7 +197,7 @@ int msm_dp_audio_hw_params(struct device *dev, goto end; } - audio = msm_dp_audio_get_data(pdev); + audio = msm_dp_audio_get_data(msm_dp_display); if (IS_ERR(audio)) { rc = PTR_ERR(audio); goto end; @@ -281,15 +216,14 @@ end: return rc; } -static void msm_dp_audio_shutdown(struct device *dev, void *data) +void msm_dp_audio_shutdown(struct drm_connector *connector, + struct drm_bridge *bridge) { struct msm_dp_audio_private *audio; - struct platform_device *pdev; struct msm_dp *msm_dp_display; - pdev = to_platform_device(dev); - msm_dp_display = platform_get_drvdata(pdev); - audio = msm_dp_audio_get_data(pdev); + msm_dp_display = to_dp_bridge(bridge)->msm_dp_display; + audio = msm_dp_audio_get_data(msm_dp_display); if (IS_ERR(audio)) { DRM_ERROR("failed to get audio data\n"); return; @@ -311,47 +245,6 @@ static void msm_dp_audio_shutdown(struct device *dev, void *data) msm_dp_display_signal_audio_complete(msm_dp_display); } -static const struct hdmi_codec_ops msm_dp_audio_codec_ops = { - .hw_params = msm_dp_audio_hw_params, - .audio_shutdown = msm_dp_audio_shutdown, - .get_eld = msm_dp_audio_get_eld, - .hook_plugged_cb = msm_dp_audio_hook_plugged_cb, -}; - -static struct hdmi_codec_pdata codec_data = { - .ops = &msm_dp_audio_codec_ops, - .max_i2s_channels = 8, - .i2s = 1, -}; - -void msm_dp_unregister_audio_driver(struct device *dev, struct msm_dp_audio *msm_dp_audio) -{ - struct msm_dp_audio_private *audio_priv; - - audio_priv = container_of(msm_dp_audio, struct msm_dp_audio_private, msm_dp_audio); - - if (audio_priv->audio_pdev) { - platform_device_unregister(audio_priv->audio_pdev); - audio_priv->audio_pdev = NULL; - } -} - -int msm_dp_register_audio_driver(struct device *dev, - struct msm_dp_audio *msm_dp_audio) -{ - struct msm_dp_audio_private *audio_priv; - - audio_priv = container_of(msm_dp_audio, - struct msm_dp_audio_private, msm_dp_audio); - - audio_priv->audio_pdev = platform_device_register_data(dev, - HDMI_CODEC_DRV_NAME, - PLATFORM_DEVID_AUTO, - &codec_data, - sizeof(codec_data)); - return PTR_ERR_OR_ZERO(audio_priv->audio_pdev); -} - struct msm_dp_audio *msm_dp_audio_get(struct platform_device *pdev, struct msm_dp_catalog *catalog) { diff --git a/drivers/gpu/drm/msm/dp/dp_audio.h b/drivers/gpu/drm/msm/dp/dp_audio.h index beea34cbab77..58fc14693e48 100644 --- a/drivers/gpu/drm/msm/dp/dp_audio.h +++ b/drivers/gpu/drm/msm/dp/dp_audio.h @@ -36,23 +36,6 @@ struct msm_dp_audio *msm_dp_audio_get(struct platform_device *pdev, struct msm_dp_catalog *catalog); /** - * msm_dp_register_audio_driver() - * - * Registers DP device with hdmi_codec interface. - * - * @dev: DP device instance. - * @msm_dp_audio: an instance of msm_dp_audio module. - * - * - * Returns the error code in case of failure, otherwise - * zero on success. - */ -int msm_dp_register_audio_driver(struct device *dev, - struct msm_dp_audio *msm_dp_audio); - -void msm_dp_unregister_audio_driver(struct device *dev, struct msm_dp_audio *msm_dp_audio); - -/** * msm_dp_audio_put() * * Cleans the msm_dp_audio instance. @@ -61,10 +44,12 @@ void msm_dp_unregister_audio_driver(struct device *dev, struct msm_dp_audio *msm */ void msm_dp_audio_put(struct msm_dp_audio *msm_dp_audio); -int msm_dp_audio_hw_params(struct device *dev, - void *data, - struct hdmi_codec_daifmt *daifmt, - struct hdmi_codec_params *params); +int msm_dp_audio_prepare(struct drm_connector *connector, + struct drm_bridge *bridge, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params); +void msm_dp_audio_shutdown(struct drm_connector *connector, + struct drm_bridge *bridge); #endif /* _DP_AUDIO_H_ */ diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c index 69a26bb5fabd..a50bfafbb4ea 100644 --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c @@ -1034,10 +1034,12 @@ static int msm_dp_ctrl_set_vx_px(struct msm_dp_ctrl_private *ctrl, return 0; } -static int msm_dp_ctrl_update_vx_px(struct msm_dp_ctrl_private *ctrl) +static int msm_dp_ctrl_update_phy_vx_px(struct msm_dp_ctrl_private *ctrl, + enum drm_dp_phy dp_phy) { struct msm_dp_link *link = ctrl->link; - int ret = 0, lane, lane_cnt; + int lane, lane_cnt, reg; + int ret = 0; u8 buf[4]; u32 max_level_reached = 0; u32 voltage_swing_level = link->phy_params.v_level; @@ -1075,8 +1077,13 @@ static int msm_dp_ctrl_update_vx_px(struct msm_dp_ctrl_private *ctrl) drm_dbg_dp(ctrl->drm_dev, "sink: p|v=0x%x\n", voltage_swing_level | pre_emphasis_level); - ret = drm_dp_dpcd_write(ctrl->aux, DP_TRAINING_LANE0_SET, - buf, lane_cnt); + + if (dp_phy == DP_PHY_DPRX) + reg = DP_TRAINING_LANE0_SET; + else + reg = DP_TRAINING_LANE0_SET_PHY_REPEATER(dp_phy); + + ret = drm_dp_dpcd_write(ctrl->aux, reg, buf, lane_cnt); if (ret == lane_cnt) ret = 0; @@ -1084,9 +1091,10 @@ static int msm_dp_ctrl_update_vx_px(struct msm_dp_ctrl_private *ctrl) } static bool msm_dp_ctrl_train_pattern_set(struct msm_dp_ctrl_private *ctrl, - u8 pattern) + u8 pattern, enum drm_dp_phy dp_phy) { u8 buf; + int reg; int ret = 0; drm_dbg_dp(ctrl->drm_dev, "sink: pattern=%x\n", pattern); @@ -1096,17 +1104,26 @@ static bool msm_dp_ctrl_train_pattern_set(struct msm_dp_ctrl_private *ctrl, if (pattern && pattern != DP_TRAINING_PATTERN_4) buf |= DP_LINK_SCRAMBLING_DISABLE; - ret = drm_dp_dpcd_writeb(ctrl->aux, DP_TRAINING_PATTERN_SET, buf); + if (dp_phy == DP_PHY_DPRX) + reg = DP_TRAINING_PATTERN_SET; + else + reg = DP_TRAINING_PATTERN_SET_PHY_REPEATER(dp_phy); + + ret = drm_dp_dpcd_writeb(ctrl->aux, reg, buf); return ret == 1; } static int msm_dp_ctrl_link_train_1(struct msm_dp_ctrl_private *ctrl, - int *training_step) + int *training_step, enum drm_dp_phy dp_phy) { + int delay_us; int tries, old_v_level, ret = 0; u8 link_status[DP_LINK_STATUS_SIZE]; int const maximum_retries = 4; + delay_us = drm_dp_read_clock_recovery_delay(ctrl->aux, + ctrl->panel->dpcd, dp_phy, false); + msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, 0); *training_step = DP_TRAINING_1; @@ -1115,18 +1132,19 @@ static int msm_dp_ctrl_link_train_1(struct msm_dp_ctrl_private *ctrl, if (ret) return ret; msm_dp_ctrl_train_pattern_set(ctrl, DP_TRAINING_PATTERN_1 | - DP_LINK_SCRAMBLING_DISABLE); + DP_LINK_SCRAMBLING_DISABLE, dp_phy); - ret = msm_dp_ctrl_update_vx_px(ctrl); + msm_dp_link_reset_phy_params_vx_px(ctrl->link); + ret = msm_dp_ctrl_update_phy_vx_px(ctrl, dp_phy); if (ret) return ret; tries = 0; old_v_level = ctrl->link->phy_params.v_level; for (tries = 0; tries < maximum_retries; tries++) { - drm_dp_link_train_clock_recovery_delay(ctrl->aux, ctrl->panel->dpcd); + fsleep(delay_us); - ret = drm_dp_dpcd_read_link_status(ctrl->aux, link_status); + ret = drm_dp_dpcd_read_phy_link_status(ctrl->aux, dp_phy, link_status); if (ret) return ret; @@ -1147,7 +1165,7 @@ static int msm_dp_ctrl_link_train_1(struct msm_dp_ctrl_private *ctrl, } msm_dp_link_adjust_levels(ctrl->link, link_status); - ret = msm_dp_ctrl_update_vx_px(ctrl); + ret = msm_dp_ctrl_update_phy_vx_px(ctrl, dp_phy); if (ret) return ret; } @@ -1199,21 +1217,31 @@ static int msm_dp_ctrl_link_lane_down_shift(struct msm_dp_ctrl_private *ctrl) return 0; } -static void msm_dp_ctrl_clear_training_pattern(struct msm_dp_ctrl_private *ctrl) +static void msm_dp_ctrl_clear_training_pattern(struct msm_dp_ctrl_private *ctrl, + enum drm_dp_phy dp_phy) { - msm_dp_ctrl_train_pattern_set(ctrl, DP_TRAINING_PATTERN_DISABLE); - drm_dp_link_train_channel_eq_delay(ctrl->aux, ctrl->panel->dpcd); + int delay_us; + + msm_dp_ctrl_train_pattern_set(ctrl, DP_TRAINING_PATTERN_DISABLE, dp_phy); + + delay_us = drm_dp_read_channel_eq_delay(ctrl->aux, + ctrl->panel->dpcd, dp_phy, false); + fsleep(delay_us); } static int msm_dp_ctrl_link_train_2(struct msm_dp_ctrl_private *ctrl, - int *training_step) + int *training_step, enum drm_dp_phy dp_phy) { + int delay_us; int tries = 0, ret = 0; u8 pattern; u32 state_ctrl_bit; int const maximum_retries = 5; u8 link_status[DP_LINK_STATUS_SIZE]; + delay_us = drm_dp_read_channel_eq_delay(ctrl->aux, + ctrl->panel->dpcd, dp_phy, false); + msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, 0); *training_step = DP_TRAINING_2; @@ -1233,12 +1261,12 @@ static int msm_dp_ctrl_link_train_2(struct msm_dp_ctrl_private *ctrl, if (ret) return ret; - msm_dp_ctrl_train_pattern_set(ctrl, pattern); + msm_dp_ctrl_train_pattern_set(ctrl, pattern, dp_phy); for (tries = 0; tries <= maximum_retries; tries++) { - drm_dp_link_train_channel_eq_delay(ctrl->aux, ctrl->panel->dpcd); + fsleep(delay_us); - ret = drm_dp_dpcd_read_link_status(ctrl->aux, link_status); + ret = drm_dp_dpcd_read_phy_link_status(ctrl->aux, dp_phy, link_status); if (ret) return ret; @@ -1248,7 +1276,7 @@ static int msm_dp_ctrl_link_train_2(struct msm_dp_ctrl_private *ctrl, } msm_dp_link_adjust_levels(ctrl->link, link_status); - ret = msm_dp_ctrl_update_vx_px(ctrl); + ret = msm_dp_ctrl_update_phy_vx_px(ctrl, dp_phy); if (ret) return ret; @@ -1257,9 +1285,32 @@ static int msm_dp_ctrl_link_train_2(struct msm_dp_ctrl_private *ctrl, return -ETIMEDOUT; } +static int msm_dp_ctrl_link_train_1_2(struct msm_dp_ctrl_private *ctrl, + int *training_step, enum drm_dp_phy dp_phy) +{ + int ret; + + ret = msm_dp_ctrl_link_train_1(ctrl, training_step, dp_phy); + if (ret) { + DRM_ERROR("link training #1 on phy %d failed. ret=%d\n", dp_phy, ret); + return ret; + } + drm_dbg_dp(ctrl->drm_dev, "link training #1 on phy %d successful\n", dp_phy); + + ret = msm_dp_ctrl_link_train_2(ctrl, training_step, dp_phy); + if (ret) { + DRM_ERROR("link training #2 on phy %d failed. ret=%d\n", dp_phy, ret); + return ret; + } + drm_dbg_dp(ctrl->drm_dev, "link training #2 on phy %d successful\n", dp_phy); + + return 0; +} + static int msm_dp_ctrl_link_train(struct msm_dp_ctrl_private *ctrl, int *training_step) { + int i; int ret = 0; const u8 *dpcd = ctrl->panel->dpcd; u8 encoding[] = { 0, DP_SET_ANSI_8B10B }; @@ -1272,8 +1323,6 @@ static int msm_dp_ctrl_link_train(struct msm_dp_ctrl_private *ctrl, link_info.rate = ctrl->link->link_params.rate; link_info.capabilities = DP_LINK_CAP_ENHANCED_FRAMING; - msm_dp_link_reset_phy_params_vx_px(ctrl->link); - msm_dp_aux_link_configure(ctrl->aux, &link_info); if (drm_dp_max_downspread(dpcd)) @@ -1288,24 +1337,27 @@ static int msm_dp_ctrl_link_train(struct msm_dp_ctrl_private *ctrl, &assr, 1); } - ret = msm_dp_ctrl_link_train_1(ctrl, training_step); + for (i = ctrl->link->lttpr_count - 1; i >= 0; i--) { + enum drm_dp_phy dp_phy = DP_PHY_LTTPR(i); + + ret = msm_dp_ctrl_link_train_1_2(ctrl, training_step, dp_phy); + msm_dp_ctrl_clear_training_pattern(ctrl, dp_phy); + + if (ret) + break; + } + if (ret) { - DRM_ERROR("link training #1 failed. ret=%d\n", ret); + DRM_ERROR("link training of LTTPR(s) failed. ret=%d\n", ret); goto end; } - /* print success info as this is a result of user initiated action */ - drm_dbg_dp(ctrl->drm_dev, "link training #1 successful\n"); - - ret = msm_dp_ctrl_link_train_2(ctrl, training_step); + ret = msm_dp_ctrl_link_train_1_2(ctrl, training_step, DP_PHY_DPRX); if (ret) { - DRM_ERROR("link training #2 failed. ret=%d\n", ret); + DRM_ERROR("link training on sink failed. ret=%d\n", ret); goto end; } - /* print success info as this is a result of user initiated action */ - drm_dbg_dp(ctrl->drm_dev, "link training #2 successful\n"); - end: msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, 0); @@ -1622,7 +1674,7 @@ static int msm_dp_ctrl_link_maintenance(struct msm_dp_ctrl_private *ctrl) if (ret) goto end; - msm_dp_ctrl_clear_training_pattern(ctrl); + msm_dp_ctrl_clear_training_pattern(ctrl, DP_PHY_DPRX); msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO); @@ -1646,7 +1698,7 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl) return false; } msm_dp_catalog_ctrl_send_phy_pattern(ctrl->catalog, pattern_requested); - msm_dp_ctrl_update_vx_px(ctrl); + msm_dp_ctrl_update_phy_vx_px(ctrl, DP_PHY_DPRX); msm_dp_link_send_test_response(ctrl->link); pattern_sent = msm_dp_catalog_ctrl_read_phy_pattern(ctrl->catalog); @@ -1888,7 +1940,7 @@ int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl) } /* stop link training before start re training */ - msm_dp_ctrl_clear_training_pattern(ctrl); + msm_dp_ctrl_clear_training_pattern(ctrl, DP_PHY_DPRX); } rc = msm_dp_ctrl_reinitialize_mainlink(ctrl); @@ -1912,7 +1964,7 @@ int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl) * link training failed * end txing train pattern here */ - msm_dp_ctrl_clear_training_pattern(ctrl); + msm_dp_ctrl_clear_training_pattern(ctrl, DP_PHY_DPRX); msm_dp_ctrl_deinitialize_mainlink(ctrl); rc = -ECONNRESET; @@ -1983,7 +2035,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train msm_dp_ctrl_link_retrain(ctrl); /* stop txing train pattern to end link training */ - msm_dp_ctrl_clear_training_pattern(ctrl); + msm_dp_ctrl_clear_training_pattern(ctrl, DP_PHY_DPRX); /* * Set up transfer unit values and set controller state to send diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c index bbc47d86ae9e..386c4669c831 100644 --- a/drivers/gpu/drm/msm/dp/dp_display.c +++ b/drivers/gpu/drm/msm/dp/dp_display.c @@ -13,6 +13,7 @@ #include <linux/delay.h> #include <linux/string_choices.h> #include <drm/display/drm_dp_aux_bus.h> +#include <drm/display/drm_hdmi_audio_helper.h> #include <drm/drm_edid.h> #include "msm_drv.h" @@ -288,13 +289,6 @@ static int msm_dp_display_bind(struct device *dev, struct device *master, goto end; } - - rc = msm_dp_register_audio_driver(dev, dp->audio); - if (rc) { - DRM_ERROR("Audio registration Dp failed\n"); - goto end; - } - rc = msm_dp_hpd_event_thread_start(dp); if (rc) { DRM_ERROR("Event thread create failed\n"); @@ -316,7 +310,6 @@ static void msm_dp_display_unbind(struct device *dev, struct device *master, of_dp_aux_depopulate_bus(dp->aux); - msm_dp_unregister_audio_driver(dev, dp->audio); msm_dp_aux_unregister(dp->aux); dp->drm_dev = NULL; dp->aux->drm_dev = NULL; @@ -367,17 +360,21 @@ static int msm_dp_display_send_hpd_notification(struct msm_dp_display_private *d return 0; } -static void msm_dp_display_lttpr_init(struct msm_dp_display_private *dp) +static int msm_dp_display_lttpr_init(struct msm_dp_display_private *dp, u8 *dpcd) { - u8 lttpr_caps[DP_LTTPR_COMMON_CAP_SIZE]; - int rc; + int rc, lttpr_count; - if (drm_dp_read_lttpr_common_caps(dp->aux, dp->panel->dpcd, lttpr_caps)) - return; + if (drm_dp_read_lttpr_common_caps(dp->aux, dpcd, dp->link->lttpr_common_caps)) + return 0; - rc = drm_dp_lttpr_init(dp->aux, drm_dp_lttpr_count(lttpr_caps)); - if (rc) + lttpr_count = drm_dp_lttpr_count(dp->link->lttpr_common_caps); + rc = drm_dp_lttpr_init(dp->aux, lttpr_count); + if (rc) { DRM_ERROR("failed to set LTTPRs transparency mode, rc=%d\n", rc); + return 0; + } + + return lttpr_count; } static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp) @@ -385,12 +382,17 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp) struct drm_connector *connector = dp->msm_dp_display.connector; const struct drm_display_info *info = &connector->display_info; int rc = 0; + u8 dpcd[DP_RECEIVER_CAP_SIZE]; - rc = msm_dp_panel_read_sink_caps(dp->panel, connector); + rc = drm_dp_read_dpcd_caps(dp->aux, dpcd); if (rc) goto end; - msm_dp_display_lttpr_init(dp); + dp->link->lttpr_count = msm_dp_display_lttpr_init(dp, dpcd); + + rc = msm_dp_panel_read_sink_caps(dp->panel, connector); + if (rc) + goto end; msm_dp_link_process_request(dp->link); @@ -626,9 +628,9 @@ static void msm_dp_display_handle_plugged_change(struct msm_dp *msm_dp_display, struct msm_dp_display_private, msm_dp_display); /* notify audio subsystem only if sink supports audio */ - if (msm_dp_display->plugged_cb && msm_dp_display->codec_dev && - dp->audio_supported) - msm_dp_display->plugged_cb(msm_dp_display->codec_dev, plugged); + if (dp->audio_supported) + drm_connector_hdmi_audio_plugged_notify(msm_dp_display->connector, + plugged); } static int msm_dp_hpd_unplug_handle(struct msm_dp_display_private *dp, u32 data) @@ -907,19 +909,6 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp) return 0; } -int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display, - hdmi_codec_plugged_cb fn, struct device *codec_dev) -{ - bool plugged; - - msm_dp_display->plugged_cb = fn; - msm_dp_display->codec_dev = codec_dev; - plugged = msm_dp_display->link_ready; - msm_dp_display_handle_plugged_change(msm_dp_display, plugged); - - return 0; -} - /** * msm_dp_bridge_mode_valid - callback to determine if specified mode is valid * @bridge: Pointer to drm bridge structure diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h index ecbc2d92f546..cc6e2cab36e9 100644 --- a/drivers/gpu/drm/msm/dp/dp_display.h +++ b/drivers/gpu/drm/msm/dp/dp_display.h @@ -7,7 +7,6 @@ #define _DP_DISPLAY_H_ #include "dp_panel.h" -#include <sound/hdmi-codec.h> #include "disp/msm_disp_snapshot.h" #define DP_MAX_PIXEL_CLK_KHZ 675000 @@ -15,7 +14,6 @@ struct msm_dp { struct drm_device *drm_dev; struct platform_device *pdev; - struct device *codec_dev; struct drm_connector *connector; struct drm_bridge *next_bridge; bool link_ready; @@ -25,14 +23,10 @@ struct msm_dp { bool is_edp; bool internal_hpd; - hdmi_codec_plugged_cb plugged_cb; - struct msm_dp_audio *msm_dp_audio; bool psr_supported; }; -int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display, - hdmi_codec_plugged_cb fn, struct device *codec_dev); int msm_dp_display_get_modes(struct msm_dp *msm_dp_display); bool msm_dp_display_check_video_test(struct msm_dp *msm_dp_display); int msm_dp_display_get_test_bpp(struct msm_dp *msm_dp_display); diff --git a/drivers/gpu/drm/msm/dp/dp_drm.c b/drivers/gpu/drm/msm/dp/dp_drm.c index 293f4745f1e2..f222d7ccaa88 100644 --- a/drivers/gpu/drm/msm/dp/dp_drm.c +++ b/drivers/gpu/drm/msm/dp/dp_drm.c @@ -12,6 +12,7 @@ #include "msm_drv.h" #include "msm_kms.h" +#include "dp_audio.h" #include "dp_drm.h" /** @@ -114,6 +115,9 @@ static const struct drm_bridge_funcs msm_dp_bridge_ops = { .hpd_disable = msm_dp_bridge_hpd_disable, .hpd_notify = msm_dp_bridge_hpd_notify, .debugfs_init = msm_dp_bridge_debugfs_init, + + .dp_audio_prepare = msm_dp_audio_prepare, + .dp_audio_shutdown = msm_dp_audio_shutdown, }; static int msm_edp_bridge_atomic_check(struct drm_bridge *drm_bridge, @@ -321,9 +325,13 @@ int msm_dp_bridge_init(struct msm_dp *msm_dp_display, struct drm_device *dev, */ if (!msm_dp_display->is_edp) { bridge->ops = + DRM_BRIDGE_OP_DP_AUDIO | DRM_BRIDGE_OP_DETECT | DRM_BRIDGE_OP_HPD | DRM_BRIDGE_OP_MODES; + bridge->hdmi_audio_dev = &msm_dp_display->pdev->dev; + bridge->hdmi_audio_max_i2s_playback_channels = 8; + bridge->hdmi_audio_dai_port = -1; } rc = devm_drm_bridge_add(dev->dev, bridge); diff --git a/drivers/gpu/drm/msm/dp/dp_link.h b/drivers/gpu/drm/msm/dp/dp_link.h index 8db5d5698a97..ba47c6d19fbf 100644 --- a/drivers/gpu/drm/msm/dp/dp_link.h +++ b/drivers/gpu/drm/msm/dp/dp_link.h @@ -7,6 +7,7 @@ #define _DP_LINK_H_ #include "dp_aux.h" +#include <drm/display/drm_dp_helper.h> #define DS_PORT_STATUS_CHANGED 0x200 #define DP_TEST_BIT_DEPTH_UNKNOWN 0xFFFFFFFF @@ -60,6 +61,9 @@ struct msm_dp_link_phy_params { }; struct msm_dp_link { + u8 lttpr_common_caps[DP_LTTPR_COMMON_CAP_SIZE]; + int lttpr_count; + u32 sink_request; u32 test_response; diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c index 92415bf8aa16..4e8ab75c771b 100644 --- a/drivers/gpu/drm/msm/dp/dp_panel.c +++ b/drivers/gpu/drm/msm/dp/dp_panel.c @@ -47,7 +47,7 @@ static void msm_dp_panel_read_psr_cap(struct msm_dp_panel_private *panel) static int msm_dp_panel_read_dpcd(struct msm_dp_panel *msm_dp_panel) { - int rc; + int rc, max_lttpr_lanes, max_lttpr_rate; struct msm_dp_panel_private *panel; struct msm_dp_link_info *link_info; u8 *dpcd, major, minor; @@ -75,6 +75,16 @@ static int msm_dp_panel_read_dpcd(struct msm_dp_panel *msm_dp_panel) if (link_info->rate > msm_dp_panel->max_dp_link_rate) link_info->rate = msm_dp_panel->max_dp_link_rate; + /* Limit data lanes from LTTPR capabilities, if any */ + max_lttpr_lanes = drm_dp_lttpr_max_lane_count(panel->link->lttpr_common_caps); + if (max_lttpr_lanes && max_lttpr_lanes < link_info->num_lanes) + link_info->num_lanes = max_lttpr_lanes; + + /* Limit link rate from LTTPR capabilities, if any */ + max_lttpr_rate = drm_dp_lttpr_max_link_rate(panel->link->lttpr_common_caps); + if (max_lttpr_rate && max_lttpr_rate < link_info->rate) + link_info->rate = max_lttpr_rate; + drm_dbg_dp(panel->drm_dev, "version: %d.%d\n", major, minor); drm_dbg_dp(panel->drm_dev, "link_rate=%d\n", link_info->rate); drm_dbg_dp(panel->drm_dev, "lane_count=%d\n", link_info->num_lanes); diff --git a/drivers/gpu/drm/msm/dsi/dsi_cfg.c b/drivers/gpu/drm/msm/dsi/dsi_cfg.c index 7754dcec33d0..7675558ae2e5 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_cfg.c +++ b/drivers/gpu/drm/msm/dsi/dsi_cfg.c @@ -221,6 +221,22 @@ static const struct msm_dsi_config sc7280_dsi_cfg = { }, }; +static const struct regulator_bulk_data sa8775p_dsi_regulators[] = { + { .supply = "vdda", .init_load_uA = 8300 }, /* 1.2 V */ + { .supply = "refgen" }, +}; + +static const struct msm_dsi_config sa8775p_dsi_cfg = { + .io_offset = DSI_6G_REG_SHIFT, + .regulator_data = sa8775p_dsi_regulators, + .num_regulators = ARRAY_SIZE(sa8775p_dsi_regulators), + .bus_clk_names = dsi_v2_4_clk_names, + .num_bus_clks = ARRAY_SIZE(dsi_v2_4_clk_names), + .io_start = { + { 0xae94000, 0xae96000 }, + }, +}; + static const struct msm_dsi_host_cfg_ops msm_dsi_v2_host_ops = { .link_clk_set_rate = dsi_link_clk_set_rate_v2, .link_clk_enable = dsi_link_clk_enable_v2, @@ -294,6 +310,8 @@ static const struct msm_dsi_cfg_handler dsi_cfg_handlers[] = { &sdm845_dsi_cfg, &msm_dsi_6g_v2_host_ops}, {MSM_DSI_VER_MAJOR_6G, MSM_DSI_6G_VER_MINOR_V2_5_0, &sc7280_dsi_cfg, &msm_dsi_6g_v2_host_ops}, + {MSM_DSI_VER_MAJOR_6G, MSM_DSI_6G_VER_MINOR_V2_5_1, + &sa8775p_dsi_cfg, &msm_dsi_6g_v2_host_ops}, {MSM_DSI_VER_MAJOR_6G, MSM_DSI_6G_VER_MINOR_V2_6_0, &sdm845_dsi_cfg, &msm_dsi_6g_v2_host_ops}, {MSM_DSI_VER_MAJOR_6G, MSM_DSI_6G_VER_MINOR_V2_7_0, diff --git a/drivers/gpu/drm/msm/dsi/dsi_cfg.h b/drivers/gpu/drm/msm/dsi/dsi_cfg.h index 120cb65164c1..65b0705fac0e 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_cfg.h +++ b/drivers/gpu/drm/msm/dsi/dsi_cfg.h @@ -27,6 +27,7 @@ #define MSM_DSI_6G_VER_MINOR_V2_4_0 0x20040000 #define MSM_DSI_6G_VER_MINOR_V2_4_1 0x20040001 #define MSM_DSI_6G_VER_MINOR_V2_5_0 0x20050000 +#define MSM_DSI_6G_VER_MINOR_V2_5_1 0x20050001 #define MSM_DSI_6G_VER_MINOR_V2_6_0 0x20060000 #define MSM_DSI_6G_VER_MINOR_V2_7_0 0x20070000 #define MSM_DSI_6G_VER_MINOR_V2_8_0 0x20080000 diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c index c0bcc6828963..5973d7325699 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c @@ -581,6 +581,10 @@ static const struct of_device_id dsi_phy_dt_match[] = { .data = &dsi_phy_7nm_cfgs }, { .compatible = "qcom,dsi-phy-7nm-8150", .data = &dsi_phy_7nm_8150_cfgs }, + { .compatible = "qcom,sa8775p-dsi-phy-5nm", + .data = &dsi_phy_5nm_8775p_cfgs }, + { .compatible = "qcom,sar2130p-dsi-phy-5nm", + .data = &dsi_phy_5nm_sar2130p_cfgs }, { .compatible = "qcom,sc7280-dsi-phy-7nm", .data = &dsi_phy_7nm_7280_cfgs }, { .compatible = "qcom,sm6375-dsi-phy-7nm", diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h index 1925418d9999..7ea608f620fe 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h @@ -59,6 +59,8 @@ extern const struct msm_dsi_phy_cfg dsi_phy_7nm_8150_cfgs; extern const struct msm_dsi_phy_cfg dsi_phy_7nm_7280_cfgs; extern const struct msm_dsi_phy_cfg dsi_phy_5nm_8350_cfgs; extern const struct msm_dsi_phy_cfg dsi_phy_5nm_8450_cfgs; +extern const struct msm_dsi_phy_cfg dsi_phy_5nm_8775p_cfgs; +extern const struct msm_dsi_phy_cfg dsi_phy_5nm_sar2130p_cfgs; extern const struct msm_dsi_phy_cfg dsi_phy_4nm_8550_cfgs; extern const struct msm_dsi_phy_cfg dsi_phy_4nm_8650_cfgs; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c index a92decbee5b5..c19890358b74 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c @@ -1147,6 +1147,10 @@ static const struct regulator_bulk_data dsi_phy_7nm_37750uA_regulators[] = { { .supply = "vdds", .init_load_uA = 37550 }, }; +static const struct regulator_bulk_data dsi_phy_7nm_48000uA_regulators[] = { + { .supply = "vdds", .init_load_uA = 48000 }, +}; + static const struct regulator_bulk_data dsi_phy_7nm_98000uA_regulators[] = { { .supply = "vdds", .init_load_uA = 98000 }, }; @@ -1289,6 +1293,52 @@ const struct msm_dsi_phy_cfg dsi_phy_5nm_8450_cfgs = { .quirks = DSI_PHY_7NM_QUIRK_V4_3, }; +const struct msm_dsi_phy_cfg dsi_phy_5nm_8775p_cfgs = { + .has_phy_lane = true, + .regulator_data = dsi_phy_7nm_48000uA_regulators, + .num_regulators = ARRAY_SIZE(dsi_phy_7nm_48000uA_regulators), + .ops = { + .enable = dsi_7nm_phy_enable, + .disable = dsi_7nm_phy_disable, + .pll_init = dsi_pll_7nm_init, + .save_pll_state = dsi_7nm_pll_save_state, + .restore_pll_state = dsi_7nm_pll_restore_state, + .set_continuous_clock = dsi_7nm_set_continuous_clock, + }, + .min_pll_rate = 600000000UL, +#ifdef CONFIG_64BIT + .max_pll_rate = 5000000000UL, +#else + .max_pll_rate = ULONG_MAX, +#endif + .io_start = { 0xae94400, 0xae96400 }, + .num_dsi_phy = 2, + .quirks = DSI_PHY_7NM_QUIRK_V4_2, +}; + +const struct msm_dsi_phy_cfg dsi_phy_5nm_sar2130p_cfgs = { + .has_phy_lane = true, + .regulator_data = dsi_phy_7nm_97800uA_regulators, + .num_regulators = ARRAY_SIZE(dsi_phy_7nm_97800uA_regulators), + .ops = { + .enable = dsi_7nm_phy_enable, + .disable = dsi_7nm_phy_disable, + .pll_init = dsi_pll_7nm_init, + .save_pll_state = dsi_7nm_pll_save_state, + .restore_pll_state = dsi_7nm_pll_restore_state, + .set_continuous_clock = dsi_7nm_set_continuous_clock, + }, + .min_pll_rate = 600000000UL, +#ifdef CONFIG_64BIT + .max_pll_rate = 5000000000UL, +#else + .max_pll_rate = ULONG_MAX, +#endif + .io_start = { 0xae95000, 0xae97000 }, + .num_dsi_phy = 2, + .quirks = DSI_PHY_7NM_QUIRK_V5_2, +}; + const struct msm_dsi_phy_cfg dsi_phy_4nm_8550_cfgs = { .has_phy_lane = true, .regulator_data = dsi_phy_7nm_98400uA_regulators, diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c index 248541ff4492..2fd388b892dc 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi.c @@ -8,6 +8,7 @@ #include <linux/gpio/consumer.h> #include <linux/of_irq.h> #include <linux/of_platform.h> +#include <linux/pinctrl/consumer.h> #include <linux/platform_device.h> #include <drm/drm_bridge_connector.h> @@ -199,12 +200,6 @@ int msm_hdmi_modeset_init(struct hdmi *hdmi, goto fail; } - ret = msm_hdmi_hpd_enable(hdmi->bridge); - if (ret < 0) { - DRM_DEV_ERROR(&hdmi->pdev->dev, "failed to enable HPD: %d\n", ret); - goto fail; - } - return 0; fail: @@ -220,28 +215,24 @@ fail: * The hdmi device: */ -#define HDMI_CFG(item, entry) \ - .item ## _names = item ##_names_ ## entry, \ - .item ## _cnt = ARRAY_SIZE(item ## _names_ ## entry) - -static const char *hpd_reg_names_8960[] = {"core-vdda"}; -static const char *hpd_clk_names_8960[] = {"core", "master_iface", "slave_iface"}; +static const char * const pwr_reg_names_8960[] = {"core-vdda"}; +static const char * const pwr_clk_names_8960[] = {"core", "master_iface", "slave_iface"}; static const struct hdmi_platform_config hdmi_tx_8960_config = { - HDMI_CFG(hpd_reg, 8960), - HDMI_CFG(hpd_clk, 8960), + .pwr_reg_names = pwr_reg_names_8960, + .pwr_reg_cnt = ARRAY_SIZE(pwr_reg_names_8960), + .pwr_clk_names = pwr_clk_names_8960, + .pwr_clk_cnt = ARRAY_SIZE(pwr_clk_names_8960), }; -static const char *pwr_reg_names_8x74[] = {"core-vdda", "core-vcc"}; -static const char *pwr_clk_names_8x74[] = {"extp", "alt_iface"}; -static const char *hpd_clk_names_8x74[] = {"iface", "core", "mdp_core"}; -static unsigned long hpd_clk_freq_8x74[] = {0, 19200000, 0}; +static const char * const pwr_reg_names_8x74[] = {"core-vdda", "core-vcc"}; +static const char * const pwr_clk_names_8x74[] = {"iface", "core", "mdp_core", "alt_iface"}; static const struct hdmi_platform_config hdmi_tx_8974_config = { - HDMI_CFG(pwr_reg, 8x74), - HDMI_CFG(pwr_clk, 8x74), - HDMI_CFG(hpd_clk, 8x74), - .hpd_freq = hpd_clk_freq_8x74, + .pwr_reg_names = pwr_reg_names_8x74, + .pwr_reg_cnt = ARRAY_SIZE(pwr_reg_names_8x74), + .pwr_clk_names = pwr_clk_names_8x74, + .pwr_clk_cnt = ARRAY_SIZE(pwr_clk_names_8x74), }; static int msm_hdmi_bind(struct device *dev, struct device *master, void *data) @@ -264,9 +255,6 @@ static void msm_hdmi_unbind(struct device *dev, struct device *master, struct msm_drm_private *priv = dev_get_drvdata(master); if (priv->hdmi) { - if (priv->hdmi->bridge) - msm_hdmi_hpd_disable(priv->hdmi); - msm_hdmi_destroy(priv->hdmi); priv->hdmi = NULL; } @@ -296,6 +284,7 @@ static int msm_hdmi_dev_probe(struct platform_device *pdev) hdmi->pdev = pdev; hdmi->config = config; spin_lock_init(&hdmi->reg_lock); + mutex_init(&hdmi->state_mutex); ret = drm_of_find_panel_or_bridge(pdev->dev.of_node, 1, 0, NULL, &hdmi->next_bridge); if (ret && ret != -ENODEV) @@ -322,20 +311,6 @@ static int msm_hdmi_dev_probe(struct platform_device *pdev) if (hdmi->irq < 0) return hdmi->irq; - hdmi->hpd_regs = devm_kcalloc(&pdev->dev, - config->hpd_reg_cnt, - sizeof(hdmi->hpd_regs[0]), - GFP_KERNEL); - if (!hdmi->hpd_regs) - return -ENOMEM; - - for (i = 0; i < config->hpd_reg_cnt; i++) - hdmi->hpd_regs[i].supply = config->hpd_reg_names[i]; - - ret = devm_regulator_bulk_get(&pdev->dev, config->hpd_reg_cnt, hdmi->hpd_regs); - if (ret) - return dev_err_probe(dev, ret, "failed to get hpd regulators\n"); - hdmi->pwr_regs = devm_kcalloc(&pdev->dev, config->pwr_reg_cnt, sizeof(hdmi->pwr_regs[0]), @@ -350,25 +325,6 @@ static int msm_hdmi_dev_probe(struct platform_device *pdev) if (ret) return dev_err_probe(dev, ret, "failed to get pwr regulators\n"); - hdmi->hpd_clks = devm_kcalloc(&pdev->dev, - config->hpd_clk_cnt, - sizeof(hdmi->hpd_clks[0]), - GFP_KERNEL); - if (!hdmi->hpd_clks) - return -ENOMEM; - - for (i = 0; i < config->hpd_clk_cnt; i++) { - struct clk *clk; - - clk = msm_clk_get(pdev, config->hpd_clk_names[i]); - if (IS_ERR(clk)) - return dev_err_probe(dev, PTR_ERR(clk), - "failed to get hpd clk: %s\n", - config->hpd_clk_names[i]); - - hdmi->hpd_clks[i] = clk; - } - hdmi->pwr_clks = devm_kcalloc(&pdev->dev, config->pwr_clk_cnt, sizeof(hdmi->pwr_clks[0]), @@ -376,17 +332,17 @@ static int msm_hdmi_dev_probe(struct platform_device *pdev) if (!hdmi->pwr_clks) return -ENOMEM; - for (i = 0; i < config->pwr_clk_cnt; i++) { - struct clk *clk; + for (i = 0; i < config->pwr_clk_cnt; i++) + hdmi->pwr_clks[i].id = config->pwr_clk_names[i]; - clk = msm_clk_get(pdev, config->pwr_clk_names[i]); - if (IS_ERR(clk)) - return dev_err_probe(dev, PTR_ERR(clk), - "failed to get pwr clk: %s\n", - config->pwr_clk_names[i]); + ret = devm_clk_bulk_get(&pdev->dev, config->pwr_clk_cnt, hdmi->pwr_clks); + if (ret) + return ret; - hdmi->pwr_clks[i] = clk; - } + hdmi->extp_clk = devm_clk_get_optional(&pdev->dev, "extp"); + if (IS_ERR(hdmi->extp_clk)) + return dev_err_probe(dev, PTR_ERR(hdmi->extp_clk), + "failed to get extp clock\n"); hdmi->hpd_gpiod = devm_gpiod_get_optional(&pdev->dev, "hpd", GPIOD_IN); /* This will catch e.g. -EPROBE_DEFER */ @@ -432,6 +388,48 @@ static void msm_hdmi_dev_remove(struct platform_device *pdev) msm_hdmi_put_phy(hdmi); } +static int msm_hdmi_runtime_suspend(struct device *dev) +{ + struct hdmi *hdmi = dev_get_drvdata(dev); + const struct hdmi_platform_config *config = hdmi->config; + + clk_bulk_disable_unprepare(config->pwr_clk_cnt, hdmi->pwr_clks); + + pinctrl_pm_select_sleep_state(dev); + + regulator_bulk_disable(config->pwr_reg_cnt, hdmi->pwr_regs); + + return 0; +} + +static int msm_hdmi_runtime_resume(struct device *dev) +{ + struct hdmi *hdmi = dev_get_drvdata(dev); + const struct hdmi_platform_config *config = hdmi->config; + int ret; + + ret = regulator_bulk_enable(config->pwr_reg_cnt, hdmi->pwr_regs); + if (ret) + return ret; + + ret = pinctrl_pm_select_default_state(dev); + if (ret) + goto fail; + + ret = clk_bulk_prepare_enable(config->pwr_clk_cnt, hdmi->pwr_clks); + if (ret) + goto fail; + + return 0; + +fail: + pinctrl_pm_select_sleep_state(dev); + + return ret; +} + +DEFINE_RUNTIME_DEV_PM_OPS(msm_hdmi_pm_ops, msm_hdmi_runtime_suspend, msm_hdmi_runtime_resume, NULL); + static const struct of_device_id msm_hdmi_dt_match[] = { { .compatible = "qcom,hdmi-tx-8998", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8996", .data = &hdmi_tx_8974_config }, @@ -449,6 +447,7 @@ static struct platform_driver msm_hdmi_driver = { .driver = { .name = "hdmi_msm", .of_match_table = msm_hdmi_dt_match, + .pm = &msm_hdmi_pm_ops, }, }; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h index a5f481c39277..d5e572d10d6a 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.h +++ b/drivers/gpu/drm/msm/hdmi/hdmi.h @@ -41,16 +41,17 @@ struct hdmi { /* video state: */ bool power_on; + bool hpd_enabled; + struct mutex state_mutex; /* protects two booleans */ unsigned long int pixclock; void __iomem *mmio; void __iomem *qfprom_mmio; phys_addr_t mmio_phy_addr; - struct regulator_bulk_data *hpd_regs; struct regulator_bulk_data *pwr_regs; - struct clk **hpd_clks; - struct clk **pwr_clks; + struct clk_bulk_data *pwr_clks; + struct clk *extp_clk; struct gpio_desc *hpd_gpiod; @@ -83,21 +84,12 @@ struct hdmi { /* platform config data (ie. from DT, or pdata) */ struct hdmi_platform_config { - /* regulators that need to be on for hpd: */ - const char **hpd_reg_names; - int hpd_reg_cnt; - /* regulators that need to be on for screen pwr: */ - const char **pwr_reg_names; + const char * const *pwr_reg_names; int pwr_reg_cnt; - /* clks that need to be on for hpd: */ - const char **hpd_clk_names; - const long unsigned *hpd_freq; - int hpd_clk_cnt; - - /* clks that need to be on for screen pwr (ie pixel clk): */ - const char **pwr_clk_names; + /* clks that need to be on: */ + const char * const *pwr_clk_names; int pwr_clk_cnt; }; @@ -224,8 +216,8 @@ int msm_hdmi_bridge_init(struct hdmi *hdmi); void msm_hdmi_hpd_irq(struct drm_bridge *bridge); enum drm_connector_status msm_hdmi_bridge_detect( struct drm_bridge *bridge); -int msm_hdmi_hpd_enable(struct drm_bridge *bridge); -void msm_hdmi_hpd_disable(struct hdmi *hdmi); +void msm_hdmi_hpd_enable(struct drm_bridge *bridge); +void msm_hdmi_hpd_disable(struct drm_bridge *bridge); /* * i2c adapter for ddc: diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_audio.c b/drivers/gpu/drm/msm/hdmi/hdmi_audio.c index 8bb975e82c17..b9ec14ef2c20 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_audio.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_audio.c @@ -4,6 +4,7 @@ * Author: Rob Clark <robdclark@gmail.com> */ +#include <drm/display/drm_hdmi_helper.h> #include <drm/display/drm_hdmi_state_helper.h> #include <linux/hdmi.h> @@ -12,71 +13,9 @@ #include "hdmi.h" -/* Supported HDMI Audio sample rates */ -#define MSM_HDMI_SAMPLE_RATE_32KHZ 0 -#define MSM_HDMI_SAMPLE_RATE_44_1KHZ 1 -#define MSM_HDMI_SAMPLE_RATE_48KHZ 2 -#define MSM_HDMI_SAMPLE_RATE_88_2KHZ 3 -#define MSM_HDMI_SAMPLE_RATE_96KHZ 4 -#define MSM_HDMI_SAMPLE_RATE_176_4KHZ 5 -#define MSM_HDMI_SAMPLE_RATE_192KHZ 6 -#define MSM_HDMI_SAMPLE_RATE_MAX 7 - - -struct hdmi_msm_audio_acr { - uint32_t n; /* N parameter for clock regeneration */ - uint32_t cts; /* CTS parameter for clock regeneration */ -}; - -struct hdmi_msm_audio_arcs { - unsigned long int pixclock; - struct hdmi_msm_audio_acr lut[MSM_HDMI_SAMPLE_RATE_MAX]; -}; - -#define HDMI_MSM_AUDIO_ARCS(pclk, ...) { (1000 * (pclk)), __VA_ARGS__ } - -/* Audio constants lookup table for hdmi_msm_audio_acr_setup */ -/* Valid Pixel-Clock rates: 25.2MHz, 27MHz, 27.03MHz, 74.25MHz, 148.5MHz */ -static const struct hdmi_msm_audio_arcs acr_lut[] = { - /* 25.200MHz */ - HDMI_MSM_AUDIO_ARCS(25200, { - {4096, 25200}, {6272, 28000}, {6144, 25200}, {12544, 28000}, - {12288, 25200}, {25088, 28000}, {24576, 25200} }), - /* 27.000MHz */ - HDMI_MSM_AUDIO_ARCS(27000, { - {4096, 27000}, {6272, 30000}, {6144, 27000}, {12544, 30000}, - {12288, 27000}, {25088, 30000}, {24576, 27000} }), - /* 27.027MHz */ - HDMI_MSM_AUDIO_ARCS(27030, { - {4096, 27027}, {6272, 30030}, {6144, 27027}, {12544, 30030}, - {12288, 27027}, {25088, 30030}, {24576, 27027} }), - /* 74.250MHz */ - HDMI_MSM_AUDIO_ARCS(74250, { - {4096, 74250}, {6272, 82500}, {6144, 74250}, {12544, 82500}, - {12288, 74250}, {25088, 82500}, {24576, 74250} }), - /* 148.500MHz */ - HDMI_MSM_AUDIO_ARCS(148500, { - {4096, 148500}, {6272, 165000}, {6144, 148500}, {12544, 165000}, - {12288, 148500}, {25088, 165000}, {24576, 148500} }), -}; - -static const struct hdmi_msm_audio_arcs *get_arcs(unsigned long int pixclock) -{ - int i; - - for (i = 0; i < ARRAY_SIZE(acr_lut); i++) { - const struct hdmi_msm_audio_arcs *arcs = &acr_lut[i]; - if (arcs->pixclock == pixclock) - return arcs; - } - - return NULL; -} - int msm_hdmi_audio_update(struct hdmi *hdmi) { struct hdmi_audio *audio = &hdmi->audio; - const struct hdmi_msm_audio_arcs *arcs = NULL; bool enabled = audio->enabled; uint32_t acr_pkt_ctrl, vbi_pkt_ctrl, aud_pkt_ctrl; uint32_t audio_config; @@ -94,15 +33,6 @@ int msm_hdmi_audio_update(struct hdmi *hdmi) enabled = false; } - if (enabled) { - arcs = get_arcs(hdmi->pixclock); - if (!arcs) { - DBG("disabling audio: unsupported pixclock: %lu", - hdmi->pixclock); - enabled = false; - } - } - /* Read first before writing */ acr_pkt_ctrl = hdmi_read(hdmi, REG_HDMI_ACR_PKT_CTRL); vbi_pkt_ctrl = hdmi_read(hdmi, REG_HDMI_VBI_PKT_CTRL); @@ -116,15 +46,12 @@ int msm_hdmi_audio_update(struct hdmi *hdmi) uint32_t n, cts, multiplier; enum hdmi_acr_cts select; - n = arcs->lut[audio->rate].n; - cts = arcs->lut[audio->rate].cts; + drm_hdmi_acr_get_n_cts(hdmi->pixclock, audio->rate, &n, &cts); - if ((MSM_HDMI_SAMPLE_RATE_192KHZ == audio->rate) || - (MSM_HDMI_SAMPLE_RATE_176_4KHZ == audio->rate)) { + if (audio->rate == 192000 || audio->rate == 176400) { multiplier = 4; n >>= 2; /* divide N by 4 and use multiplier */ - } else if ((MSM_HDMI_SAMPLE_RATE_96KHZ == audio->rate) || - (MSM_HDMI_SAMPLE_RATE_88_2KHZ == audio->rate)) { + } else if (audio->rate == 96000 || audio->rate == 88200) { multiplier = 2; n >>= 1; /* divide N by 2 and use multiplier */ } else { @@ -137,13 +64,11 @@ int msm_hdmi_audio_update(struct hdmi *hdmi) acr_pkt_ctrl |= HDMI_ACR_PKT_CTRL_AUDIO_PRIORITY; acr_pkt_ctrl |= HDMI_ACR_PKT_CTRL_N_MULTIPLIER(multiplier); - if ((MSM_HDMI_SAMPLE_RATE_48KHZ == audio->rate) || - (MSM_HDMI_SAMPLE_RATE_96KHZ == audio->rate) || - (MSM_HDMI_SAMPLE_RATE_192KHZ == audio->rate)) + if (audio->rate == 48000 || audio->rate == 96000 || + audio->rate == 192000) select = ACR_48; - else if ((MSM_HDMI_SAMPLE_RATE_44_1KHZ == audio->rate) || - (MSM_HDMI_SAMPLE_RATE_88_2KHZ == audio->rate) || - (MSM_HDMI_SAMPLE_RATE_176_4KHZ == audio->rate)) + else if (audio->rate == 44100 || audio->rate == 88200 || + audio->rate == 176400) select = ACR_44; else /* default to 32k */ select = ACR_32; @@ -204,7 +129,6 @@ int msm_hdmi_bridge_audio_prepare(struct drm_connector *connector, { struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge); struct hdmi *hdmi = hdmi_bridge->hdmi; - unsigned int rate; int ret; drm_dbg_driver(bridge->dev, "%u Hz, %d bit, %d channels\n", @@ -214,25 +138,12 @@ int msm_hdmi_bridge_audio_prepare(struct drm_connector *connector, switch (params->sample_rate) { case 32000: - rate = MSM_HDMI_SAMPLE_RATE_32KHZ; - break; case 44100: - rate = MSM_HDMI_SAMPLE_RATE_44_1KHZ; - break; case 48000: - rate = MSM_HDMI_SAMPLE_RATE_48KHZ; - break; case 88200: - rate = MSM_HDMI_SAMPLE_RATE_88_2KHZ; - break; case 96000: - rate = MSM_HDMI_SAMPLE_RATE_96KHZ; - break; case 176400: - rate = MSM_HDMI_SAMPLE_RATE_176_4KHZ; - break; case 192000: - rate = MSM_HDMI_SAMPLE_RATE_192KHZ; break; default: drm_err(bridge->dev, "rate[%d] not supported!\n", @@ -245,7 +156,7 @@ int msm_hdmi_bridge_audio_prepare(struct drm_connector *connector, if (ret) return ret; - hdmi->audio.rate = rate; + hdmi->audio.rate = params->sample_rate; hdmi->audio.channels = params->cea.channels; hdmi->audio.enabled = true; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c b/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c index 7f71956806a2..53a7ce8cc7bc 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c @@ -18,52 +18,34 @@ static void msm_hdmi_power_on(struct drm_bridge *bridge) struct drm_device *dev = bridge->dev; struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge); struct hdmi *hdmi = hdmi_bridge->hdmi; - const struct hdmi_platform_config *config = hdmi->config; - int i, ret; - - pm_runtime_get_sync(&hdmi->pdev->dev); + int ret; - ret = regulator_bulk_enable(config->pwr_reg_cnt, hdmi->pwr_regs); - if (ret) - DRM_DEV_ERROR(dev->dev, "failed to enable pwr regulator: %d\n", ret); + pm_runtime_resume_and_get(&hdmi->pdev->dev); - if (config->pwr_clk_cnt > 0) { + if (hdmi->extp_clk) { DBG("pixclock: %lu", hdmi->pixclock); - ret = clk_set_rate(hdmi->pwr_clks[0], hdmi->pixclock); - if (ret) { - DRM_DEV_ERROR(dev->dev, "failed to set pixel clk: %s (%d)\n", - config->pwr_clk_names[0], ret); - } - } + ret = clk_set_rate(hdmi->extp_clk, hdmi->pixclock); + if (ret) + DRM_DEV_ERROR(dev->dev, "failed to set extp clk rate: %d\n", ret); - for (i = 0; i < config->pwr_clk_cnt; i++) { - ret = clk_prepare_enable(hdmi->pwr_clks[i]); - if (ret) { - DRM_DEV_ERROR(dev->dev, "failed to enable pwr clk: %s (%d)\n", - config->pwr_clk_names[i], ret); - } + ret = clk_prepare_enable(hdmi->extp_clk); + if (ret) + DRM_DEV_ERROR(dev->dev, "failed to enable extp clk: %d\n", ret); } } static void power_off(struct drm_bridge *bridge) { - struct drm_device *dev = bridge->dev; struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge); struct hdmi *hdmi = hdmi_bridge->hdmi; - const struct hdmi_platform_config *config = hdmi->config; - int i, ret; /* TODO do we need to wait for final vblank somewhere before * cutting the clocks? */ mdelay(16 + 4); - for (i = 0; i < config->pwr_clk_cnt; i++) - clk_disable_unprepare(hdmi->pwr_clks[i]); - - ret = regulator_bulk_disable(config->pwr_reg_cnt, hdmi->pwr_regs); - if (ret) - DRM_DEV_ERROR(dev->dev, "failed to disable pwr regulator: %d\n", ret); + if (hdmi->extp_clk) + clk_disable_unprepare(hdmi->extp_clk); pm_runtime_put(&hdmi->pdev->dev); } @@ -320,13 +302,16 @@ static void msm_hdmi_bridge_atomic_pre_enable(struct drm_bridge *bridge, msm_hdmi_set_timings(hdmi, &crtc_state->adjusted_mode); + mutex_lock(&hdmi->state_mutex); if (!hdmi->power_on) { msm_hdmi_phy_resource_enable(phy); msm_hdmi_power_on(bridge); hdmi->power_on = true; - if (connector->display_info.is_hdmi) - msm_hdmi_audio_update(hdmi); } + mutex_unlock(&hdmi->state_mutex); + + if (connector->display_info.is_hdmi) + msm_hdmi_audio_update(hdmi); drm_atomic_helper_connector_hdmi_update_infoframes(connector, state); @@ -349,7 +334,10 @@ static void msm_hdmi_bridge_atomic_post_disable(struct drm_bridge *bridge, msm_hdmi_hdcp_off(hdmi->hdcp_ctrl); DBG("power down"); - msm_hdmi_set_mode(hdmi, false); + + /* Keep the HDMI enabled if the HPD is enabled */ + mutex_lock(&hdmi->state_mutex); + msm_hdmi_set_mode(hdmi, hdmi->hpd_enabled); msm_hdmi_phy_powerdown(phy); @@ -360,6 +348,7 @@ static void msm_hdmi_bridge_atomic_post_disable(struct drm_bridge *bridge, msm_hdmi_audio_update(hdmi); msm_hdmi_phy_resource_disable(phy); } + mutex_unlock(&hdmi->state_mutex); } static void msm_hdmi_set_timings(struct hdmi *hdmi, @@ -411,9 +400,6 @@ static void msm_hdmi_set_timings(struct hdmi *hdmi, frame_ctrl |= HDMI_FRAME_CTRL_INTERLACED_EN; DBG("frame_ctrl=%08x", frame_ctrl); hdmi_write(hdmi, REG_HDMI_FRAME_CTRL, frame_ctrl); - - if (hdmi->connector->display_info.is_hdmi) - msm_hdmi_audio_update(hdmi); } static const struct drm_edid *msm_hdmi_bridge_edid_read(struct drm_bridge *bridge, @@ -440,7 +426,6 @@ static enum drm_mode_status msm_hdmi_bridge_tmds_char_rate_valid(const struct dr { struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge); struct hdmi *hdmi = hdmi_bridge->hdmi; - const struct hdmi_platform_config *config = hdmi->config; struct msm_drm_private *priv = bridge->dev->dev_private; struct msm_kms *kms = priv->kms; long actual; @@ -453,8 +438,8 @@ static enum drm_mode_status msm_hdmi_bridge_tmds_char_rate_valid(const struct dr actual = kms->funcs->round_pixclk(kms, tmds_rate, hdmi_bridge->hdmi->encoder); - else if (config->pwr_clk_cnt > 0) - actual = clk_round_rate(hdmi->pwr_clks[0], tmds_rate); + else if (hdmi->extp_clk) + actual = clk_round_rate(hdmi->extp_clk, tmds_rate); else actual = tmds_rate; @@ -474,6 +459,8 @@ static const struct drm_bridge_funcs msm_hdmi_bridge_funcs = { .atomic_post_disable = msm_hdmi_bridge_atomic_post_disable, .edid_read = msm_hdmi_bridge_edid_read, .detect = msm_hdmi_bridge_detect, + .hpd_enable = msm_hdmi_hpd_enable, + .hpd_disable = msm_hdmi_hpd_disable, .hdmi_tmds_char_rate_valid = msm_hdmi_bridge_tmds_char_rate_valid, .hdmi_clear_infoframe = msm_hdmi_bridge_clear_infoframe, .hdmi_write_infoframe = msm_hdmi_bridge_write_infoframe, diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_hpd.c b/drivers/gpu/drm/msm/hdmi/hdmi_hpd.c index 9ce0ffa35417..407e6c449ee0 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_hpd.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_hpd.c @@ -60,68 +60,30 @@ static void msm_hdmi_phy_reset(struct hdmi *hdmi) } } -static void enable_hpd_clocks(struct hdmi *hdmi, bool enable) -{ - const struct hdmi_platform_config *config = hdmi->config; - struct device *dev = &hdmi->pdev->dev; - int i, ret; - - if (enable) { - for (i = 0; i < config->hpd_clk_cnt; i++) { - if (config->hpd_freq && config->hpd_freq[i]) { - ret = clk_set_rate(hdmi->hpd_clks[i], - config->hpd_freq[i]); - if (ret) - dev_warn(dev, - "failed to set clk %s (%d)\n", - config->hpd_clk_names[i], ret); - } - - ret = clk_prepare_enable(hdmi->hpd_clks[i]); - if (ret) { - DRM_DEV_ERROR(dev, - "failed to enable hpd clk: %s (%d)\n", - config->hpd_clk_names[i], ret); - } - } - } else { - for (i = config->hpd_clk_cnt - 1; i >= 0; i--) - clk_disable_unprepare(hdmi->hpd_clks[i]); - } -} - -int msm_hdmi_hpd_enable(struct drm_bridge *bridge) +void msm_hdmi_hpd_enable(struct drm_bridge *bridge) { struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge); struct hdmi *hdmi = hdmi_bridge->hdmi; - const struct hdmi_platform_config *config = hdmi->config; struct device *dev = &hdmi->pdev->dev; uint32_t hpd_ctrl; int ret; unsigned long flags; - ret = regulator_bulk_enable(config->hpd_reg_cnt, hdmi->hpd_regs); - if (ret) { - DRM_DEV_ERROR(dev, "failed to enable hpd regulators: %d\n", ret); - goto fail; - } - - ret = pinctrl_pm_select_default_state(dev); - if (ret) { - DRM_DEV_ERROR(dev, "pinctrl state chg failed: %d\n", ret); - goto fail; - } - if (hdmi->hpd_gpiod) gpiod_set_value_cansleep(hdmi->hpd_gpiod, 1); - pm_runtime_get_sync(dev); - enable_hpd_clocks(hdmi, true); + ret = pm_runtime_resume_and_get(dev); + if (WARN_ON(ret)) + return; + mutex_lock(&hdmi->state_mutex); msm_hdmi_set_mode(hdmi, false); msm_hdmi_phy_reset(hdmi); msm_hdmi_set_mode(hdmi, true); + hdmi->hpd_enabled = true; + mutex_unlock(&hdmi->state_mutex); + hdmi_write(hdmi, REG_HDMI_USEC_REFTIMER, 0x0001001b); /* enable HPD events: */ @@ -140,34 +102,23 @@ int msm_hdmi_hpd_enable(struct drm_bridge *bridge) hdmi_write(hdmi, REG_HDMI_HPD_CTRL, HDMI_HPD_CTRL_ENABLE | hpd_ctrl); spin_unlock_irqrestore(&hdmi->reg_lock, flags); - - return 0; - -fail: - return ret; } -void msm_hdmi_hpd_disable(struct hdmi *hdmi) +void msm_hdmi_hpd_disable(struct drm_bridge *bridge) { - const struct hdmi_platform_config *config = hdmi->config; + struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge); + struct hdmi *hdmi = hdmi_bridge->hdmi; struct device *dev = &hdmi->pdev->dev; - int ret; /* Disable HPD interrupt */ hdmi_write(hdmi, REG_HDMI_HPD_INT_CTRL, 0); - msm_hdmi_set_mode(hdmi, false); + mutex_lock(&hdmi->state_mutex); + hdmi->hpd_enabled = false; + msm_hdmi_set_mode(hdmi, hdmi->power_on); + mutex_unlock(&hdmi->state_mutex); - enable_hpd_clocks(hdmi, false); pm_runtime_put(dev); - - ret = pinctrl_pm_select_sleep_state(dev); - if (ret) - dev_warn(dev, "pinctrl state chg failed: %d\n", ret); - - ret = regulator_bulk_disable(config->hpd_reg_cnt, hdmi->hpd_regs); - if (ret) - dev_warn(dev, "failed to disable hpd regulator: %d\n", ret); } void msm_hdmi_hpd_irq(struct drm_bridge *bridge) @@ -202,14 +153,16 @@ void msm_hdmi_hpd_irq(struct drm_bridge *bridge) static enum drm_connector_status detect_reg(struct hdmi *hdmi) { - uint32_t hpd_int_status; + u32 hpd_int_status = 0; + int ret; - pm_runtime_get_sync(&hdmi->pdev->dev); - enable_hpd_clocks(hdmi, true); + ret = pm_runtime_resume_and_get(&hdmi->pdev->dev); + if (ret) + goto out; hpd_int_status = hdmi_read(hdmi, REG_HDMI_HPD_INT_STATUS); - enable_hpd_clocks(hdmi, false); +out: pm_runtime_put(&hdmi->pdev->dev); return (hpd_int_status & HDMI_HPD_INT_STATUS_CABLE_DETECTED) ? diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_i2c.c b/drivers/gpu/drm/msm/hdmi/hdmi_i2c.c index 7aa500d24240..ebefea4fb408 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_i2c.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_i2c.c @@ -107,11 +107,15 @@ static int msm_hdmi_i2c_xfer(struct i2c_adapter *i2c, if (num == 0) return num; + ret = pm_runtime_resume_and_get(&hdmi->pdev->dev); + if (ret) + return ret; + init_ddc(hdmi_i2c); ret = ddc_clear_irq(hdmi_i2c); if (ret) - return ret; + goto fail; for (i = 0; i < num; i++) { struct i2c_msg *p = &msgs[i]; @@ -169,7 +173,7 @@ static int msm_hdmi_i2c_xfer(struct i2c_adapter *i2c, hdmi_read(hdmi, REG_HDMI_DDC_SW_STATUS), hdmi_read(hdmi, REG_HDMI_DDC_HW_STATUS), hdmi_read(hdmi, REG_HDMI_DDC_INT_CTRL)); - return ret; + goto fail; } ddc_status = hdmi_read(hdmi, REG_HDMI_DDC_SW_STATUS); @@ -202,7 +206,13 @@ static int msm_hdmi_i2c_xfer(struct i2c_adapter *i2c, } } + pm_runtime_put(&hdmi->pdev->dev); + return i; + +fail: + pm_runtime_put(&hdmi->pdev->dev); + return ret; } static u32 msm_hdmi_i2c_func(struct i2c_adapter *adapter) diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy.c index 03120c54ced6..667573f1db7c 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_phy.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy.c @@ -58,7 +58,11 @@ int msm_hdmi_phy_resource_enable(struct hdmi_phy *phy) struct device *dev = &phy->pdev->dev; int i, ret = 0; - pm_runtime_get_sync(dev); + ret = pm_runtime_resume_and_get(dev); + if (ret) { + DRM_DEV_ERROR(dev, "runtime resume failed: %d\n", ret); + return ret; + } ret = regulator_bulk_enable(cfg->num_regs, phy->regs); if (ret) { diff --git a/drivers/gpu/drm/msm/msm_gpu_devfreq.c b/drivers/gpu/drm/msm/msm_gpu_devfreq.c index 6970b0f7f457..2e1d5c343272 100644 --- a/drivers/gpu/drm/msm/msm_gpu_devfreq.c +++ b/drivers/gpu/drm/msm/msm_gpu_devfreq.c @@ -156,6 +156,7 @@ void msm_devfreq_init(struct msm_gpu *gpu) priv->gpu_devfreq_config.downdifferential = 10; mutex_init(&df->lock); + df->suspended = true; ret = dev_pm_qos_add_request(&gpu->pdev->dev, &df->boost_freq, DEV_PM_QOS_MIN_FREQUENCY, 0); diff --git a/drivers/gpu/drm/msm/msm_mdss.c b/drivers/gpu/drm/msm/msm_mdss.c index dcb49fd30402..f706e44231a9 100644 --- a/drivers/gpu/drm/msm/msm_mdss.c +++ b/drivers/gpu/drm/msm/msm_mdss.c @@ -592,6 +592,16 @@ static const struct msm_mdss_data sa8775p_data = { .reg_bus_bw = 74000, }; +static const struct msm_mdss_data sar2130p_data = { + .ubwc_enc_version = UBWC_3_0, /* 4.0.2 in hw */ + .ubwc_dec_version = UBWC_4_3, + .ubwc_swizzle = 6, + .ubwc_bank_spread = true, + .highest_bank_bit = 0, + .macrotile_mode = 1, + .reg_bus_bw = 74000, +}; + static const struct msm_mdss_data sc7180_data = { .ubwc_enc_version = UBWC_2_0, .ubwc_dec_version = UBWC_2_0, @@ -738,6 +748,7 @@ static const struct of_device_id mdss_dt_match[] = { { .compatible = "qcom,msm8998-mdss", .data = &msm8998_data }, { .compatible = "qcom,qcm2290-mdss", .data = &qcm2290_data }, { .compatible = "qcom,sa8775p-mdss", .data = &sa8775p_data }, + { .compatible = "qcom,sar2130p-mdss", .data = &sar2130p_data }, { .compatible = "qcom,sdm670-mdss", .data = &sdm670_data }, { .compatible = "qcom,sdm845-mdss", .data = &sdm845_data }, { .compatible = "qcom,sc7180-mdss", .data = &sc7180_data }, diff --git a/drivers/gpu/drm/msm/msm_ringbuffer.c b/drivers/gpu/drm/msm/msm_ringbuffer.c index c5651c39ac2a..89dce15eed3b 100644 --- a/drivers/gpu/drm/msm/msm_ringbuffer.c +++ b/drivers/gpu/drm/msm/msm_ringbuffer.c @@ -93,7 +93,7 @@ struct msm_ringbuffer *msm_ringbuffer_new(struct msm_gpu *gpu, int id, } msm_gem_object_set_name(ring->bo, "ring%d", id); - args.name = to_msm_bo(ring->bo)->name, + args.name = to_msm_bo(ring->bo)->name; ring->end = ring->start + (MSM_GPU_RINGBUFFER_SZ >> 2); ring->next = ring->start; diff --git a/drivers/gpu/drm/nouveau/Kbuild b/drivers/gpu/drm/nouveau/Kbuild index 7b863355c5c6..385d24530d1e 100644 --- a/drivers/gpu/drm/nouveau/Kbuild +++ b/drivers/gpu/drm/nouveau/Kbuild @@ -2,6 +2,7 @@ ccflags-y += -I $(src)/include ccflags-y += -I $(src)/include/nvkm ccflags-y += -I $(src)/nvkm +ccflags-y += -I $(src)/nvkm/subdev/gsp ccflags-y += -I $(src) # NVKM - HW resource manager @@ -68,5 +69,6 @@ nouveau-y += nv17_fence.o nouveau-y += nv50_fence.o nouveau-y += nv84_fence.o nouveau-y += nvc0_fence.o +nouveau-y += gv100_fence.o obj-$(CONFIG_DRM_NOUVEAU) += nouveau.o diff --git a/drivers/gpu/drm/nouveau/dispnv04/crtc.c b/drivers/gpu/drm/nouveau/dispnv04/crtc.c index 67146f1e8482..c063756eaea3 100644 --- a/drivers/gpu/drm/nouveau/dispnv04/crtc.c +++ b/drivers/gpu/drm/nouveau/dispnv04/crtc.c @@ -768,9 +768,7 @@ static void nv_crtc_destroy(struct drm_crtc *crtc) disp->image[nv_crtc->index] = NULL; } - nouveau_bo_unmap(nv_crtc->cursor.nvbo); - nouveau_bo_unpin(nv_crtc->cursor.nvbo); - nouveau_bo_fini(nv_crtc->cursor.nvbo); + nouveau_bo_unpin_del(&nv_crtc->cursor.nvbo); nvif_event_dtor(&nv_crtc->vblank); nvif_head_dtor(&nv_crtc->head); kfree(nv_crtc); @@ -1303,6 +1301,7 @@ nv04_crtc_vblank_handler(struct nvif_event *event, void *repv, u32 repc) int nv04_crtc_create(struct drm_device *dev, int crtc_num) { + struct nouveau_cli *cli = &nouveau_drm(dev)->client; struct nouveau_display *disp = nouveau_display(dev); struct nouveau_crtc *nv_crtc; struct drm_plane *primary; @@ -1336,20 +1335,9 @@ nv04_crtc_create(struct drm_device *dev, int crtc_num) drm_crtc_helper_add(&nv_crtc->base, &nv04_crtc_helper_funcs); drm_mode_crtc_set_gamma_size(&nv_crtc->base, 256); - ret = nouveau_bo_new(&nouveau_drm(dev)->client, 64*64*4, 0x100, - NOUVEAU_GEM_DOMAIN_VRAM, 0, 0x0000, NULL, NULL, - &nv_crtc->cursor.nvbo); - if (!ret) { - ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, - NOUVEAU_GEM_DOMAIN_VRAM, false); - if (!ret) { - ret = nouveau_bo_map(nv_crtc->cursor.nvbo); - if (ret) - nouveau_bo_unpin(nv_crtc->cursor.nvbo); - } - if (ret) - nouveau_bo_fini(nv_crtc->cursor.nvbo); - } + ret = nouveau_bo_new_map(cli, NOUVEAU_GEM_DOMAIN_VRAM, 64 * 64 * 4, &nv_crtc->cursor.nvbo); + if (ret) + return ret; nv04_cursor_init(nv_crtc); diff --git a/drivers/gpu/drm/nouveau/dispnv50/Kbuild b/drivers/gpu/drm/nouveau/dispnv50/Kbuild index 28be2912ff74..d5049dee4b8c 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/Kbuild +++ b/drivers/gpu/drm/nouveau/dispnv50/Kbuild @@ -9,11 +9,13 @@ nouveau-y += dispnv50/core907d.o nouveau-y += dispnv50/core917d.o nouveau-y += dispnv50/corec37d.o nouveau-y += dispnv50/corec57d.o +nouveau-y += dispnv50/coreca7d.o nouveau-$(CONFIG_DEBUG_FS) += dispnv50/crc.o nouveau-$(CONFIG_DEBUG_FS) += dispnv50/crc907d.o nouveau-$(CONFIG_DEBUG_FS) += dispnv50/crcc37d.o nouveau-$(CONFIG_DEBUG_FS) += dispnv50/crcc57d.o +nouveau-$(CONFIG_DEBUG_FS) += dispnv50/crcca7d.o nouveau-y += dispnv50/dac507d.o nouveau-y += dispnv50/dac907d.o @@ -31,6 +33,7 @@ nouveau-y += dispnv50/head907d.o nouveau-y += dispnv50/head917d.o nouveau-y += dispnv50/headc37d.o nouveau-y += dispnv50/headc57d.o +nouveau-y += dispnv50/headca7d.o nouveau-y += dispnv50/wimm.o nouveau-y += dispnv50/wimmc37b.o @@ -39,6 +42,7 @@ nouveau-y += dispnv50/wndw.o nouveau-y += dispnv50/wndwc37e.o nouveau-y += dispnv50/wndwc57e.o nouveau-y += dispnv50/wndwc67e.o +nouveau-y += dispnv50/wndwca7e.o nouveau-y += dispnv50/base.o nouveau-y += dispnv50/base507c.o diff --git a/drivers/gpu/drm/nouveau/dispnv50/core.c b/drivers/gpu/drm/nouveau/dispnv50/core.c index f045515696cb..c6331bf97582 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/core.c +++ b/drivers/gpu/drm/nouveau/dispnv50/core.c @@ -42,6 +42,7 @@ nv50_core_new(struct nouveau_drm *drm, struct nv50_core **pcore) int version; int (*new)(struct nouveau_drm *, s32, struct nv50_core **); } cores[] = { + { GB202_DISP_CORE_CHANNEL_DMA, 0, coreca7d_new }, { AD102_DISP_CORE_CHANNEL_DMA, 0, corec57d_new }, { GA102_DISP_CORE_CHANNEL_DMA, 0, corec57d_new }, { TU102_DISP_CORE_CHANNEL_DMA, 0, corec57d_new }, diff --git a/drivers/gpu/drm/nouveau/dispnv50/core.h b/drivers/gpu/drm/nouveau/dispnv50/core.h index f75088186fba..aa07a3ad5dfd 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/core.h +++ b/drivers/gpu/drm/nouveau/dispnv50/core.h @@ -7,7 +7,10 @@ struct nv50_core { const struct nv50_core_func *func; + struct nv50_disp *disp; + struct nv50_dmac chan; + bool assign_windows; }; @@ -18,6 +21,7 @@ struct nv50_core_func { int (*init)(struct nv50_core *); void (*ntfy_init)(struct nouveau_bo *, u32 offset); int (*caps_init)(struct nouveau_drm *, struct nv50_disp *); + u32 caps_class; int (*ntfy_wait_done)(struct nouveau_bo *, u32 offset, struct nvif_device *); int (*update)(struct nv50_core *, u32 *interlock, bool ntfy); @@ -70,4 +74,6 @@ int corec37d_wndw_owner(struct nv50_core *); extern const struct nv50_outp_func sorc37d; int corec57d_new(struct nouveau_drm *, s32, struct nv50_core **); + +int coreca7d_new(struct nouveau_drm *, s32, struct nv50_core **); #endif diff --git a/drivers/gpu/drm/nouveau/dispnv50/core507d.c b/drivers/gpu/drm/nouveau/dispnv50/core507d.c index ce2cb78bbdd3..4b947b67a844 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/core507d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/core507d.c @@ -165,6 +165,7 @@ core507d_new_(const struct nv50_core_func *func, struct nouveau_drm *drm, if (!(core = *pcore = kzalloc(sizeof(*core), GFP_KERNEL))) return -ENOMEM; core->func = func; + core->disp = disp; ret = nv50_dmac_create(drm, &oclass, 0, &args, sizeof(args), diff --git a/drivers/gpu/drm/nouveau/dispnv50/corec37d.c b/drivers/gpu/drm/nouveau/dispnv50/corec37d.c index 7f637b8830be..83eec2f091f0 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/corec37d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/corec37d.c @@ -105,7 +105,7 @@ int corec37d_caps_init(struct nouveau_drm *drm, struct nv50_disp *disp) int ret; ret = nvif_object_ctor(&disp->disp->object, "dispCaps", 0, - GV100_DISP_CAPS, NULL, 0, &disp->caps); + disp->core->func->caps_class, NULL, 0, &disp->caps); if (ret) { NV_ERROR(drm, "Failed to init notifier caps region: %d\n", @@ -162,6 +162,7 @@ corec37d = { .init = corec37d_init, .ntfy_init = corec37d_ntfy_init, .caps_init = corec37d_caps_init, + .caps_class = GV100_DISP_CAPS, .ntfy_wait_done = corec37d_ntfy_wait_done, .update = corec37d_update, .wndw.owner = corec37d_wndw_owner, diff --git a/drivers/gpu/drm/nouveau/dispnv50/corec57d.c b/drivers/gpu/drm/nouveau/dispnv50/corec57d.c index 421d0d57e1d8..39be576eadcb 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/corec57d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/corec57d.c @@ -22,6 +22,7 @@ #include "core.h" #include "head.h" +#include <nvif/class.h> #include <nvif/pushc37b.h> #include <nvhw/class/clc57d.h> @@ -63,6 +64,7 @@ corec57d = { .init = corec57d_init, .ntfy_init = corec37d_ntfy_init, .caps_init = corec37d_caps_init, + .caps_class = GV100_DISP_CAPS, .ntfy_wait_done = corec37d_ntfy_wait_done, .update = corec37d_update, .wndw.owner = corec37d_wndw_owner, diff --git a/drivers/gpu/drm/nouveau/dispnv50/coreca7d.c b/drivers/gpu/drm/nouveau/dispnv50/coreca7d.c new file mode 100644 index 000000000000..171727be400e --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/coreca7d.c @@ -0,0 +1,122 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "core.h" +#include "head.h" + +#include <nvif/class.h> +#include <nvif/pushc97b.h> + +#include <nvhw/class/clca7d.h> + +#include <nouveau_bo.h> + +static int +coreca7d_update(struct nv50_core *core, u32 *interlock, bool ntfy) +{ + const u64 ntfy_addr = core->disp->sync->offset + NV50_DISP_CORE_NTFY; + const u32 ntfy_hi = upper_32_bits(ntfy_addr); + const u32 ntfy_lo = lower_32_bits(ntfy_addr); + struct nvif_push *push = &core->chan.push; + int ret; + + ret = PUSH_WAIT(push, 5 + (ntfy ? 5 + 2 : 0)); + if (ret) + return ret; + + if (ntfy) { + PUSH_MTHD(push, NVCA7D, SET_SURFACE_ADDRESS_HI_NOTIFIER, ntfy_hi, + + SET_SURFACE_ADDRESS_LO_NOTIFIER, + NVVAL(NVCA7D, SET_SURFACE_ADDRESS_LO_NOTIFIER, ADDRESS_LO, ntfy_lo >> 4) | + NVDEF(NVCA7D, SET_SURFACE_ADDRESS_LO_NOTIFIER, TARGET, PHYSICAL_NVM) | + NVDEF(NVCA7D, SET_SURFACE_ADDRESS_LO_NOTIFIER, ENABLE, ENABLE)); + + PUSH_MTHD(push, NVCA7D, SET_NOTIFIER_CONTROL, + NVDEF(NVCA7D, SET_NOTIFIER_CONTROL, MODE, WRITE) | + NVDEF(NVCA7D, SET_NOTIFIER_CONTROL, NOTIFY, ENABLE)); + } + + PUSH_MTHD(push, NVCA7D, SET_INTERLOCK_FLAGS, interlock[NV50_DISP_INTERLOCK_CURS], + SET_WINDOW_INTERLOCK_FLAGS, interlock[NV50_DISP_INTERLOCK_WNDW]); + + PUSH_MTHD(push, NVCA7D, UPDATE, + NVDEF(NVCA7D, UPDATE, RELEASE_ELV, TRUE) | + NVDEF(NVCA7D, UPDATE, SPECIAL_HANDLING, NONE) | + NVDEF(NVCA7D, UPDATE, INHIBIT_INTERRUPTS, FALSE)); + + if (ntfy) { + PUSH_MTHD(push, NVCA7D, SET_NOTIFIER_CONTROL, + NVDEF(NVCA7D, SET_NOTIFIER_CONTROL, NOTIFY, DISABLE)); + } + + return PUSH_KICK(push); +} + +static int +coreca7d_init(struct nv50_core *core) +{ + struct nvif_push *push = &core->chan.push; + const u32 windows = 8, heads = 4; + int ret, i; + + ret = PUSH_WAIT(push, windows * 6 + heads * 6); + if (ret) + return ret; + + for (i = 0; i < windows; i++) { + PUSH_MTHD(push, NVCA7D, WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS(i), + NVDEF(NVCA7D, WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS, RGB_PACKED1BPP, TRUE) | + NVDEF(NVCA7D, WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS, RGB_PACKED2BPP, TRUE) | + NVDEF(NVCA7D, WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS, RGB_PACKED4BPP, TRUE) | + NVDEF(NVCA7D, WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS, RGB_PACKED8BPP, TRUE), + + WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS(i), 0x00000000); + + PUSH_MTHD(push, NVCA7D, WINDOW_SET_WINDOW_USAGE_BOUNDS(i), + NVVAL(NVCA7D, WINDOW_SET_WINDOW_USAGE_BOUNDS, MAX_PIXELS_FETCHED_PER_LINE, 0x7fff) | + NVDEF(NVCA7D, WINDOW_SET_WINDOW_USAGE_BOUNDS, ILUT_ALLOWED, TRUE) | + NVDEF(NVCA7D, WINDOW_SET_WINDOW_USAGE_BOUNDS, INPUT_SCALER_TAPS, TAPS_2) | + NVDEF(NVCA7D, WINDOW_SET_WINDOW_USAGE_BOUNDS, UPSCALING_ALLOWED, FALSE), + + WINDOW_SET_PHYSICAL(i), BIT(i)); + } + + for (i = 0; i < heads; i++) { + PUSH_MTHD(push, NVCA7D, HEAD_SET_HEAD_USAGE_BOUNDS(i), + NVDEF(NVCA7D, HEAD_SET_HEAD_USAGE_BOUNDS, CURSOR, USAGE_W256_H256) | + NVDEF(NVCA7D, HEAD_SET_HEAD_USAGE_BOUNDS, OLUT_ALLOWED, TRUE) | + NVDEF(NVCA7D, HEAD_SET_HEAD_USAGE_BOUNDS, OUTPUT_SCALER_TAPS, TAPS_2) | + NVDEF(NVCA7D, HEAD_SET_HEAD_USAGE_BOUNDS, UPSCALING_ALLOWED, TRUE)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_TILE_MASK(i), BIT(i)); + + PUSH_MTHD(push, NVCA7D, TILE_SET_TILE_SIZE(i), 0); + } + + core->assign_windows = true; + return PUSH_KICK(push); +} + +static const struct nv50_core_func +coreca7d = { + .init = coreca7d_init, + .ntfy_init = corec37d_ntfy_init, + .caps_init = corec37d_caps_init, + .caps_class = GB202_DISP_CAPS, + .ntfy_wait_done = corec37d_ntfy_wait_done, + .update = coreca7d_update, + .wndw.owner = corec37d_wndw_owner, + .head = &headca7d, + .sor = &sorc37d, +#if IS_ENABLED(CONFIG_DEBUG_FS) + .crc = &crcca7d, +#endif +}; + +int +coreca7d_new(struct nouveau_drm *drm, s32 oclass, struct nv50_core **pcore) +{ + return core507d_new_(&coreca7d, drm, oclass, pcore); +} diff --git a/drivers/gpu/drm/nouveau/dispnv50/crc.c b/drivers/gpu/drm/nouveau/dispnv50/crc.c index 5936b6b3b15d..deb6af40ef32 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/crc.c +++ b/drivers/gpu/drm/nouveau/dispnv50/crc.c @@ -509,6 +509,10 @@ nv50_crc_ctx_init(struct nv50_head *head, struct nvif_mmu *mmu, if (ret) return ret; + /* No CTXDMAs on Blackwell. */ + if (core->chan.base.user.oclass >= GB202_DISP_CORE_CHANNEL_DMA) + return 0; + ret = nvif_object_ctor(&core->chan.base.user, "kmsCrcNtfyCtxDma", NV50_DISP_HANDLE_CRC_CTX(head, idx), NV_DMA_IN_MEMORY, diff --git a/drivers/gpu/drm/nouveau/dispnv50/crc.h b/drivers/gpu/drm/nouveau/dispnv50/crc.h index 4823f1fde2dd..75a2009e8193 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/crc.h +++ b/drivers/gpu/drm/nouveau/dispnv50/crc.h @@ -94,6 +94,7 @@ void nv50_crc_atomic_clr(struct nv50_head *); extern const struct nv50_crc_func crc907d; extern const struct nv50_crc_func crcc37d; extern const struct nv50_crc_func crcc57d; +extern const struct nv50_crc_func crcca7d; #else /* IS_ENABLED(CONFIG_DEBUG_FS) */ struct nv50_crc {}; diff --git a/drivers/gpu/drm/nouveau/dispnv50/crcca7d.c b/drivers/gpu/drm/nouveau/dispnv50/crcca7d.c new file mode 100644 index 000000000000..912f59aebe87 --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/crcca7d.c @@ -0,0 +1,98 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "crcc37d.h" +#include "core.h" +#include "head.h" + +#include <nvif/pushc97b.h> + +#include <nvhw/class/clca7d.h> + +static int +crcca7d_set_ctx(struct nv50_head *head, struct nv50_crc_notifier_ctx *ctx) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, ctx ? 3 : 2); + if (ret) + return ret; + + if (ctx) { + const u32 crc_hi = upper_32_bits(ctx->mem.addr); + const u32 crc_lo = lower_32_bits(ctx->mem.addr); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_HI_CRC(i), crc_hi, + + HEAD_SET_SURFACE_ADDRESS_LO_CRC(i), + NVVAL(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CRC, ADDRESS_LO, crc_lo >> 4) | + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CRC, TARGET, PHYSICAL_NVM) | + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CRC, ENABLE, ENABLE)); + } else { + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CRC(i), + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CRC, ENABLE, DISABLE)); + } + + return 0; +} + +static int +crcca7d_set_src(struct nv50_head *head, int or, enum nv50_crc_source_type source, + struct nv50_crc_notifier_ctx *ctx) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int primary_crc, ret; + + if (!source) { + ret = PUSH_WAIT(push, 1); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_CRC_CONTROL(i), 0); + + return crcca7d_set_ctx(head, NULL); + } + + switch (source) { + case NV50_CRC_SOURCE_TYPE_SOR: + primary_crc = NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR(or); + break; + case NV50_CRC_SOURCE_TYPE_SF: + primary_crc = NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SF; + break; + default: + break; + } + + ret = crcca7d_set_ctx(head, ctx); + if (ret) + return ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_CRC_CONTROL(i), + NVDEF(NVCA7D, HEAD_SET_CRC_CONTROL, CONTROLLING_CHANNEL, CORE) | + NVDEF(NVCA7D, HEAD_SET_CRC_CONTROL, EXPECT_BUFFER_COLLAPSE, FALSE) | + NVVAL(NVCA7D, HEAD_SET_CRC_CONTROL, PRIMARY_CRC, primary_crc) | + NVDEF(NVCA7D, HEAD_SET_CRC_CONTROL, SECONDARY_CRC, NONE) | + NVDEF(NVCA7D, HEAD_SET_CRC_CONTROL, CRC_DURING_SNOOZE, DISABLE)); + + return 0; +} + +const struct nv50_crc_func +crcca7d = { + .set_src = crcca7d_set_src, + .set_ctx = crcca7d_set_ctx, + .get_entry = crcc37d_get_entry, + .ctx_finished = crcc37d_ctx_finished, + .flip_threshold = CRCC37D_FLIP_THRESHOLD, + .num_entries = CRCC37D_MAX_ENTRIES, + .notifier_len = sizeof(struct crcc37d_notifier), +}; diff --git a/drivers/gpu/drm/nouveau/dispnv50/curs.c b/drivers/gpu/drm/nouveau/dispnv50/curs.c index 31d8b2e4791d..557bd05240fa 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/curs.c +++ b/drivers/gpu/drm/nouveau/dispnv50/curs.c @@ -31,6 +31,7 @@ nv50_curs_new(struct nouveau_drm *drm, int head, struct nv50_wndw **pwndw) int version; int (*new)(struct nouveau_drm *, int, s32, struct nv50_wndw **); } curses[] = { + { GB202_DISP_CURSOR, 0, cursc37a_new }, { GA102_DISP_CURSOR, 0, cursc37a_new }, { TU102_DISP_CURSOR, 0, cursc37a_new }, { GV100_DISP_CURSOR, 0, cursc37a_new }, diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c index 9bed728cb00e..e5d37eee4301 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/disp.c +++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c @@ -279,6 +279,16 @@ nv50_dmac_create(struct nouveau_drm *drm, if (syncbuf < 0) return 0; + /* No CTXDMAs on Blackwell. */ + if (disp->oclass >= GB202_DISP) { + /* "handle != NULL_HANDLE" is used to determine enable status + * in a number of places, so fill in some fake object handles. + */ + dmac->sync.handle = NV50_DISP_HANDLE_SYNCBUF; + dmac->vram.handle = NV50_DISP_HANDLE_VRAM; + return 0; + } + ret = nvif_object_ctor(&dmac->base.user, "kmsSyncCtxDma", NV50_DISP_HANDLE_SYNCBUF, NV_DMA_IN_MEMORY, &(struct nv_dma_v0) { @@ -2808,10 +2818,7 @@ nv50_display_destroy(struct drm_device *dev) nvif_object_dtor(&disp->caps); nv50_core_del(&disp->core); - nouveau_bo_unmap(disp->sync); - if (disp->sync) - nouveau_bo_unpin(disp->sync); - nouveau_bo_fini(disp->sync); + nouveau_bo_unpin_del(&disp->sync); nouveau_display(dev)->priv = NULL; kfree(disp); @@ -2843,20 +2850,7 @@ nv50_display_create(struct drm_device *dev) dev->mode_config.normalize_zpos = true; /* small shared memory area we use for notifiers and semaphores */ - ret = nouveau_bo_new(&drm->client, 4096, 0x1000, - NOUVEAU_GEM_DOMAIN_VRAM, - 0, 0x0000, NULL, NULL, &disp->sync); - if (!ret) { - ret = nouveau_bo_pin(disp->sync, NOUVEAU_GEM_DOMAIN_VRAM, true); - if (!ret) { - ret = nouveau_bo_map(disp->sync); - if (ret) - nouveau_bo_unpin(disp->sync); - } - if (ret) - nouveau_bo_fini(disp->sync); - } - + ret = nouveau_bo_new_map(&drm->client, NOUVEAU_GEM_DOMAIN_VRAM, PAGE_SIZE, &disp->sync); if (ret) goto out; diff --git a/drivers/gpu/drm/nouveau/dispnv50/head.c b/drivers/gpu/drm/nouveau/dispnv50/head.c index d7c74cc43ba5..3dd742b4f823 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/head.c +++ b/drivers/gpu/drm/nouveau/dispnv50/head.c @@ -577,6 +577,7 @@ nv50_head_create(struct drm_device *dev, int index) return ERR_PTR(-ENOMEM); head->func = disp->core->func->head; + head->disp = disp; head->base.index = index; if (disp->disp->object.oclass < GF110_DISP) diff --git a/drivers/gpu/drm/nouveau/dispnv50/head.h b/drivers/gpu/drm/nouveau/dispnv50/head.h index e9d17037ffcf..8bd2fcb1eff5 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/head.h +++ b/drivers/gpu/drm/nouveau/dispnv50/head.h @@ -13,6 +13,8 @@ struct nv50_head { const struct nv50_head_func *func; + struct nv50_disp *disp; + struct nouveau_crtc base; struct nv50_crc crc; struct nv50_lut olut; @@ -98,4 +100,7 @@ int headc37d_dither(struct nv50_head *, struct nv50_head_atom *); void headc37d_static_wndw_map(struct nv50_head *, struct nv50_head_atom *); extern const struct nv50_head_func headc57d; +bool headc57d_olut(struct nv50_head *, struct nv50_head_atom *, int size); + +extern const struct nv50_head_func headca7d; #endif diff --git a/drivers/gpu/drm/nouveau/dispnv50/headc57d.c b/drivers/gpu/drm/nouveau/dispnv50/headc57d.c index fde4087e7691..3f8ba495de8f 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/headc57d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/headc57d.c @@ -182,7 +182,7 @@ headc57d_olut_load(struct drm_color_lut *in, int size, void __iomem *mem) writew(readw(mem - 4), mem + 4); } -static bool +bool headc57d_olut(struct nv50_head *head, struct nv50_head_atom *asyh, int size) { if (size != 0 && size != 256 && size != 1024) diff --git a/drivers/gpu/drm/nouveau/dispnv50/headca7d.c b/drivers/gpu/drm/nouveau/dispnv50/headca7d.c new file mode 100644 index 000000000000..eeaeb15aa664 --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/headca7d.c @@ -0,0 +1,297 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "head.h" +#include "atom.h" +#include "core.h" + +#include <nvif/pushc97b.h> + +#include <nvhw/class/clca7d.h> + +static int +headca7d_display_id(struct nv50_head *head, u32 display_id) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_DISPLAY_ID(i, 0), display_id); + + return 0; +} + +static int +headca7d_or(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + u8 depth; + int ret; + + switch (asyh->or.depth) { + case 6: + depth = NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_30_444; + break; + case 5: + depth = NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_24_444; + break; + case 2: + depth = NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_18_444; + break; + case 0: + depth = NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_24_444; + break; + default: + WARN_ON(1); + return -EINVAL; + } + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE(i), + NVVAL(NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE, CRC_MODE, asyh->or.crc_raster) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE, HSYNC_POLARITY, asyh->or.nhsync) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE, VSYNC_POLARITY, asyh->or.nvsync) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE, PIXEL_DEPTH, depth) | + NVDEF(NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE, COLOR_SPACE_OVERRIDE, DISABLE) | + NVDEF(NVCA7D, HEAD_SET_CONTROL_OUTPUT_RESOURCE, EXT_PACKET_WIN, NONE)); + + return 0; +} + +static int +headca7d_procamp(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_PROCAMP(i), + NVDEF(NVCA7D, HEAD_SET_PROCAMP, COLOR_SPACE, RGB) | + NVDEF(NVCA7D, HEAD_SET_PROCAMP, CHROMA_LPF, DISABLE) | + NVDEF(NVCA7D, HEAD_SET_PROCAMP, DYNAMIC_RANGE, VESA)); + + return 0; +} + +static int +headca7d_dither(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_DITHER_CONTROL(i), + NVVAL(NVCA7D, HEAD_SET_DITHER_CONTROL, ENABLE, asyh->dither.enable) | + NVVAL(NVCA7D, HEAD_SET_DITHER_CONTROL, BITS, asyh->dither.bits) | + NVDEF(NVCA7D, HEAD_SET_DITHER_CONTROL, OFFSET_ENABLE, DISABLE) | + NVVAL(NVCA7D, HEAD_SET_DITHER_CONTROL, MODE, asyh->dither.mode) | + NVVAL(NVCA7D, HEAD_SET_DITHER_CONTROL, PHASE, 0)); + + return 0; +} + +static int +headca7d_curs_clr(struct nv50_head *head) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 4); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_CONTROL_CURSOR(i), + NVDEF(NVCA7D, HEAD_SET_CONTROL_CURSOR, ENABLE, DISABLE) | + NVDEF(NVCA7D, HEAD_SET_CONTROL_CURSOR, FORMAT, A8R8G8B8)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CURSOR(i, 0), + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CURSOR, ENABLE, DISABLE)); + + return 0; +} + +static int +headca7d_curs_set(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const u32 curs_hi = upper_32_bits(asyh->curs.offset); + const u32 curs_lo = lower_32_bits(asyh->curs.offset); + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 7); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_HI_CURSOR(i, 0), curs_hi); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CURSOR(i, 0), + NVVAL(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CURSOR, ADDRESS_LO, curs_lo >> 4) | + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CURSOR, TARGET, PHYSICAL_NVM) | + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_CURSOR, ENABLE, ENABLE)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_CONTROL_CURSOR(i), + NVDEF(NVCA7D, HEAD_SET_CONTROL_CURSOR, ENABLE, ENABLE) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_CURSOR, FORMAT, asyh->curs.format) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_CURSOR, SIZE, asyh->curs.layout) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_CURSOR, HOT_SPOT_X, 0) | + NVVAL(NVCA7D, HEAD_SET_CONTROL_CURSOR, HOT_SPOT_Y, 0), + + HEAD_SET_CONTROL_CURSOR_COMPOSITION(i), + NVVAL(NVCA7D, HEAD_SET_CONTROL_CURSOR_COMPOSITION, K1, 0xff) | + NVDEF(NVCA7D, HEAD_SET_CONTROL_CURSOR_COMPOSITION, CURSOR_COLOR_FACTOR_SELECT, + K1) | + NVDEF(NVCA7D, HEAD_SET_CONTROL_CURSOR_COMPOSITION, VIEWPORT_COLOR_FACTOR_SELECT, + NEG_K1_TIMES_SRC) | + NVDEF(NVCA7D, HEAD_SET_CONTROL_CURSOR_COMPOSITION, MODE, BLEND)); + + return 0; +} + +static int +headca7d_olut_clr(struct nv50_head *head) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_OLUT(i), + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_OLUT, ENABLE, DISABLE)); + + return 0; +} + +static int +headca7d_olut_set(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const u32 olut_hi = upper_32_bits(asyh->olut.offset); + const u32 olut_lo = lower_32_bits(asyh->olut.offset); + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 6); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_SURFACE_ADDRESS_HI_OLUT(i), olut_hi, + + HEAD_SET_SURFACE_ADDRESS_LO_OLUT(i), + NVVAL(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_OLUT, ADDRESS_LO, olut_lo >> 4) | + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_OLUT, TARGET, PHYSICAL_NVM) | + NVDEF(NVCA7D, HEAD_SET_SURFACE_ADDRESS_LO_OLUT, ENABLE, ENABLE)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_OLUT_CONTROL(i), + NVVAL(NVCA7D, HEAD_SET_OLUT_CONTROL, INTERPOLATE, asyh->olut.output_mode) | + NVDEF(NVCA7D, HEAD_SET_OLUT_CONTROL, MIRROR, DISABLE) | + NVVAL(NVCA7D, HEAD_SET_OLUT_CONTROL, MODE, asyh->olut.mode) | + NVVAL(NVCA7D, HEAD_SET_OLUT_CONTROL, SIZE, asyh->olut.size), + + HEAD_SET_OLUT_FP_NORM_SCALE(i), 0xffffffff); + + return 0; +} + +static int +headca7d_mode(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + struct nv50_head_mode *m = &asyh->mode; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 11); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_RASTER_SIZE(i), + NVVAL(NVCA7D, HEAD_SET_RASTER_SIZE, WIDTH, m->h.active) | + NVVAL(NVCA7D, HEAD_SET_RASTER_SIZE, HEIGHT, m->v.active), + + HEAD_SET_RASTER_SYNC_END(i), + NVVAL(NVCA7D, HEAD_SET_RASTER_SYNC_END, X, m->h.synce) | + NVVAL(NVCA7D, HEAD_SET_RASTER_SYNC_END, Y, m->v.synce), + + HEAD_SET_RASTER_BLANK_END(i), + NVVAL(NVCA7D, HEAD_SET_RASTER_BLANK_END, X, m->h.blanke) | + NVVAL(NVCA7D, HEAD_SET_RASTER_BLANK_END, Y, m->v.blanke), + + HEAD_SET_RASTER_BLANK_START(i), + NVVAL(NVCA7D, HEAD_SET_RASTER_BLANK_START, X, m->h.blanks) | + NVVAL(NVCA7D, HEAD_SET_RASTER_BLANK_START, Y, m->v.blanks)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_CONTROL(i), + NVDEF(NVCA7D, HEAD_SET_CONTROL, STRUCTURE, PROGRESSIVE)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_PIXEL_CLOCK_FREQUENCY(i), + NVVAL(NVCA7D, HEAD_SET_PIXEL_CLOCK_FREQUENCY, HERTZ, m->clock * 1000)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX(i), + NVVAL(NVCA7D, HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX, HERTZ, m->clock * 1000)); + + return 0; +} + +static int +headca7d_view(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nvif_push *push = &head->disp->core->chan.push; + const int i = head->base.index; + int ret; + + ret = PUSH_WAIT(push, 4); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7D, HEAD_SET_VIEWPORT_SIZE_IN(i), + NVVAL(NVCA7D, HEAD_SET_VIEWPORT_SIZE_IN, WIDTH, asyh->view.iW) | + NVVAL(NVCA7D, HEAD_SET_VIEWPORT_SIZE_IN, HEIGHT, asyh->view.iH)); + + PUSH_MTHD(push, NVCA7D, HEAD_SET_VIEWPORT_SIZE_OUT(i), + NVVAL(NVCA7D, HEAD_SET_VIEWPORT_SIZE_OUT, WIDTH, asyh->view.oW) | + NVVAL(NVCA7D, HEAD_SET_VIEWPORT_SIZE_OUT, HEIGHT, asyh->view.oH)); + return 0; +} + +const struct nv50_head_func +headca7d = { + .view = headca7d_view, + .mode = headca7d_mode, + .olut = headc57d_olut, + .ilut_check = head907d_ilut_check, + .olut_identity = true, + .olut_size = 1024, + .olut_set = headca7d_olut_set, + .olut_clr = headca7d_olut_clr, + .curs_layout = head917d_curs_layout, + .curs_format = headc37d_curs_format, + .curs_set = headca7d_curs_set, + .curs_clr = headca7d_curs_clr, + .dither = headca7d_dither, + .procamp = headca7d_procamp, + .or = headca7d_or, + .static_wndw_map = headc37d_static_wndw_map, + .display_id = headca7d_display_id, +}; diff --git a/drivers/gpu/drm/nouveau/dispnv50/wimm.c b/drivers/gpu/drm/nouveau/dispnv50/wimm.c index 566fbddfc8d7..53c9ab6c138b 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wimm.c +++ b/drivers/gpu/drm/nouveau/dispnv50/wimm.c @@ -31,6 +31,7 @@ nv50_wimm_init(struct nouveau_drm *drm, struct nv50_wndw *wndw) int version; int (*init)(struct nouveau_drm *, s32, struct nv50_wndw *); } wimms[] = { + { GB202_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init }, { GA102_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init }, { TU102_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init }, { GV100_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init }, diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.c b/drivers/gpu/drm/nouveau/dispnv50/wndw.c index f6be426dd525..11d5b923d6e7 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wndw.c +++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.c @@ -556,14 +556,24 @@ nv50_wndw_prepare_fb(struct drm_plane *plane, struct drm_plane_state *state) return ret; if (wndw->ctxdma.parent) { - ctxdma = nv50_wndw_ctxdma_new(wndw, fb); - if (IS_ERR(ctxdma)) { - nouveau_bo_unpin(nvbo); - return PTR_ERR(ctxdma); + if (wndw->wndw.base.user.oclass < GB202_DISP_WINDOW_CHANNEL_DMA) { + ctxdma = nv50_wndw_ctxdma_new(wndw, fb); + if (IS_ERR(ctxdma)) { + nouveau_bo_unpin(nvbo); + return PTR_ERR(ctxdma); + } + + if (asyw->visible) + asyw->image.handle[0] = ctxdma->object.handle; + } else { + /* No CTXDMAs on Blackwell. */ + if (asyw->visible) { + /* "handle != NULL_HANDLE" is used to determine enable status + * in a number of places, so fill in a fake object handle. + */ + asyw->image.handle[0] = NV50_DISP_HANDLE_WNDW_CTX(0); + } } - - if (asyw->visible) - asyw->image.handle[0] = ctxdma->object.handle; } ret = drm_gem_plane_helper_prepare_fb(plane, state); @@ -901,6 +911,7 @@ nv50_wndw_new(struct nouveau_drm *drm, enum drm_plane_type type, int index, int (*new)(struct nouveau_drm *, enum drm_plane_type, int, s32, struct nv50_wndw **); } wndws[] = { + { GB202_DISP_WINDOW_CHANNEL_DMA, 0, wndwca7e_new }, { GA102_DISP_WINDOW_CHANNEL_DMA, 0, wndwc67e_new }, { TU102_DISP_WINDOW_CHANNEL_DMA, 0, wndwc57e_new }, { GV100_DISP_WINDOW_CHANNEL_DMA, 0, wndwc37e_new }, diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.h b/drivers/gpu/drm/nouveau/dispnv50/wndw.h index 76a6ae5d5652..90d100514bef 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wndw.h +++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.h @@ -134,6 +134,9 @@ int wndwc57e_csc_clr(struct nv50_wndw *); int wndwc67e_new(struct nouveau_drm *, enum drm_plane_type, int, s32, struct nv50_wndw **); +int wndwca7e_new(struct nouveau_drm *, enum drm_plane_type, int, s32, + struct nv50_wndw **); + int nv50_wndw_new(struct nouveau_drm *, enum drm_plane_type, int index, struct nv50_wndw **); #endif diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c b/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c index 50a7b97d37a2..554c4f91f8be 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c +++ b/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c @@ -25,6 +25,7 @@ #include <drm/drm_atomic_helper.h> #include <nouveau_bo.h> +#include <nvif/class.h> #include <nvif/if0014.h> #include <nvif/pushc37b.h> diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndwca7e.c b/drivers/gpu/drm/nouveau/dispnv50/wndwca7e.c new file mode 100644 index 000000000000..0d8e9a9d1a57 --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/wndwca7e.c @@ -0,0 +1,209 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "wndw.h" +#include "atom.h" + +#include <nvif/pushc97b.h> + +#include <nvhw/class/clca7e.h> + +#include <nouveau_bo.h> + +static int +wndwca7e_image_clr(struct nv50_wndw *wndw) +{ + struct nvif_push *push = &wndw->wndw.push; + int ret; + + ret = PUSH_WAIT(push, 4); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7E, SET_PRESENT_CONTROL, + NVVAL(NVCA7E, SET_PRESENT_CONTROL, MIN_PRESENT_INTERVAL, 0) | + NVDEF(NVCA7E, SET_PRESENT_CONTROL, BEGIN_MODE, NON_TEARING)); + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_LO_ISO(0), + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_ISO, ENABLE, DISABLE)); + + return 0; +} + +static int +wndwca7e_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) +{ + const u32 iso0_hi = upper_32_bits(asyw->image.offset[0]); + const u32 iso0_lo = lower_32_bits(asyw->image.offset[0]); + struct nvif_push *push = &wndw->wndw.push; + int ret, kind; + + if (asyw->image.kind) + kind = NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_KIND_BLOCKLINEAR; + else + kind = NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_KIND_PITCH; + + ret = PUSH_WAIT(push, 17); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_HI_ISO(0), iso0_hi); + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_LO_ISO(0), + NVVAL(NVCA7E, SET_SURFACE_ADDRESS_LO_ISO, ADDRESS_LO, iso0_lo >> 4) | + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_ISO, TARGET, PHYSICAL_NVM) | + NVVAL(NVCA7E, SET_SURFACE_ADDRESS_LO_ISO, KIND, kind) | + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_ISO, ENABLE, ENABLE)); + + PUSH_MTHD(push, NVCA7E, SET_PRESENT_CONTROL, + NVVAL(NVCA7E, SET_PRESENT_CONTROL, MIN_PRESENT_INTERVAL, asyw->image.interval) | + NVVAL(NVCA7E, SET_PRESENT_CONTROL, BEGIN_MODE, asyw->image.mode) | + NVDEF(NVCA7E, SET_PRESENT_CONTROL, TIMESTAMP_MODE, DISABLE)); + + PUSH_MTHD(push, NVCA7E, SET_SIZE, + NVVAL(NVCA7E, SET_SIZE, WIDTH, asyw->image.w) | + NVVAL(NVCA7E, SET_SIZE, HEIGHT, asyw->image.h), + + SET_STORAGE, + NVVAL(NVCA7E, SET_STORAGE, BLOCK_HEIGHT, asyw->image.blockh), + + SET_PARAMS, + NVVAL(NVCA7E, SET_PARAMS, FORMAT, asyw->image.format) | + NVDEF(NVCA7E, SET_PARAMS, CLAMP_BEFORE_BLEND, DISABLE) | + NVDEF(NVCA7E, SET_PARAMS, SWAP_UV, DISABLE) | + NVDEF(NVCA7E, SET_PARAMS, FMT_ROUNDING_MODE, ROUND_TO_NEAREST), + + SET_PLANAR_STORAGE(0), + NVVAL(NVCA7E, SET_PLANAR_STORAGE, PITCH, asyw->image.blocks[0]) | + NVVAL(NVCA7E, SET_PLANAR_STORAGE, PITCH, asyw->image.pitch[0] >> 6)); + + PUSH_MTHD(push, NVCA7E, SET_POINT_IN(0), + NVVAL(NVCA7E, SET_POINT_IN, X, asyw->state.src_x >> 16) | + NVVAL(NVCA7E, SET_POINT_IN, Y, asyw->state.src_y >> 16)); + + PUSH_MTHD(push, NVCA7E, SET_SIZE_IN, + NVVAL(NVCA7E, SET_SIZE_IN, WIDTH, asyw->state.src_w >> 16) | + NVVAL(NVCA7E, SET_SIZE_IN, HEIGHT, asyw->state.src_h >> 16)); + + PUSH_MTHD(push, NVCA7E, SET_SIZE_OUT, + NVVAL(NVCA7E, SET_SIZE_OUT, WIDTH, asyw->state.crtc_w) | + NVVAL(NVCA7E, SET_SIZE_OUT, HEIGHT, asyw->state.crtc_h)); + + return 0; +} + +static int +wndwca7e_ilut_clr(struct nv50_wndw *wndw) +{ + struct nvif_push *push = &wndw->wndw.push; + int ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_LO_ILUT, + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_ILUT, ENABLE, DISABLE)); + + return 0; +} + +static int +wndwca7e_ilut_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) +{ + const u32 ilut_hi = upper_32_bits(asyw->xlut.i.offset); + const u32 ilut_lo = lower_32_bits(asyw->xlut.i.offset); + struct nvif_push *push = &wndw->wndw.push; + int ret; + + ret = PUSH_WAIT(push, 5); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_HI_ILUT, ilut_hi, + + SET_SURFACE_ADDRESS_LO_ILUT, + NVVAL(NVCA7E, SET_SURFACE_ADDRESS_LO_ILUT, ADDRESS_LO, ilut_lo >> 4) | + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_ILUT, TARGET, PHYSICAL_NVM) | + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_ILUT, ENABLE, ENABLE)); + + PUSH_MTHD(push, NVCA7E, SET_ILUT_CONTROL, + NVVAL(NVCA7E, SET_ILUT_CONTROL, SIZE, asyw->xlut.i.size) | + NVVAL(NVCA7E, SET_ILUT_CONTROL, MODE, asyw->xlut.i.mode) | + NVVAL(NVCA7E, SET_ILUT_CONTROL, INTERPOLATE, asyw->xlut.i.output_mode)); + + return 0; +} + +static int +wndwca7e_ntfy_clr(struct nv50_wndw *wndw) +{ + struct nvif_push *push = &wndw->wndw.push; + int ret; + + ret = PUSH_WAIT(push, 2); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_LO_NOTIFIER, + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_NOTIFIER, ENABLE, DISABLE)); + + return 0; +} + +static int +wndwca7e_ntfy_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) +{ + struct nv50_disp *disp = nv50_disp(wndw->plane.dev); + const u64 ntfy_addr = disp->sync->offset + asyw->ntfy.offset; + const u32 ntfy_hi = upper_32_bits(ntfy_addr); + const u32 ntfy_lo = lower_32_bits(ntfy_addr); + struct nvif_push *push = &wndw->wndw.push; + int ret; + + ret = PUSH_WAIT(push, 5); + if (ret) + return ret; + + PUSH_MTHD(push, NVCA7E, SET_SURFACE_ADDRESS_HI_NOTIFIER, ntfy_hi, + + SET_SURFACE_ADDRESS_LO_NOTIFIER, + NVVAL(NVCA7E, SET_SURFACE_ADDRESS_LO_NOTIFIER, ADDRESS_LO, ntfy_lo >> 4) | + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_NOTIFIER, TARGET, PHYSICAL_NVM) | + NVDEF(NVCA7E, SET_SURFACE_ADDRESS_LO_NOTIFIER, ENABLE, ENABLE)); + + PUSH_MTHD(push, NVCA7E, SET_NOTIFIER_CONTROL, + NVVAL(NVCA7E, SET_NOTIFIER_CONTROL, MODE, asyw->ntfy.awaken)); + + return 0; +} + +static const struct nv50_wndw_func +wndwca7e = { + .acquire = wndwc37e_acquire, + .release = wndwc37e_release, + .ntfy_set = wndwca7e_ntfy_set, + .ntfy_clr = wndwca7e_ntfy_clr, + .ntfy_reset = corec37d_ntfy_init, + .ntfy_wait_begun = base507c_ntfy_wait_begun, + .ilut = wndwc57e_ilut, + .ilut_identity = true, + .ilut_size = 1024, + .xlut_set = wndwca7e_ilut_set, + .xlut_clr = wndwca7e_ilut_clr, + .csc = base907c_csc, + .csc_set = wndwc57e_csc_set, + .csc_clr = wndwc57e_csc_clr, + .image_set = wndwca7e_image_set, + .image_clr = wndwca7e_image_clr, + .blend_set = wndwc37e_blend_set, + .update = wndwc37e_update, +}; + +int +wndwca7e_new(struct nouveau_drm *drm, enum drm_plane_type type, int index, + s32 oclass, struct nv50_wndw **pwndw) +{ + return wndwc37e_new_(&wndwca7e, drm, type, index, oclass, BIT(index >> 1), pwndw); +} diff --git a/drivers/gpu/drm/nouveau/gv100_fence.c b/drivers/gpu/drm/nouveau/gv100_fence.c new file mode 100644 index 000000000000..cccdeca72002 --- /dev/null +++ b/drivers/gpu/drm/nouveau/gv100_fence.c @@ -0,0 +1,93 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "nouveau_drv.h" +#include "nouveau_dma.h" +#include "nouveau_fence.h" + +#include "nv50_display.h" + +#include <nvif/push906f.h> + +#include <nvhw/class/clc36f.h> + +static int +gv100_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence) +{ + struct nvif_push *push = &chan->chan.push; + int ret; + + ret = PUSH_WAIT(push, 8); + if (ret) + return ret; + + PUSH_MTHD(push, NVC36F, SEM_ADDR_LO, lower_32_bits(virtual), + SEM_ADDR_HI, upper_32_bits(virtual), + SEM_PAYLOAD_LO, sequence); + + PUSH_MTHD(push, NVC36F, SEM_EXECUTE, + NVDEF(NVC36F, SEM_EXECUTE, OPERATION, RELEASE) | + NVDEF(NVC36F, SEM_EXECUTE, RELEASE_WFI, EN) | + NVDEF(NVC36F, SEM_EXECUTE, PAYLOAD_SIZE, 32BIT) | + NVDEF(NVC36F, SEM_EXECUTE, RELEASE_TIMESTAMP, DIS)); + + PUSH_MTHD(push, NVC36F, NON_STALL_INTERRUPT, 0); + + PUSH_KICK(push); + return 0; +} + +static int +gv100_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence) +{ + struct nvif_push *push = &chan->chan.push; + int ret; + + ret = PUSH_WAIT(push, 6); + if (ret) + return ret; + + PUSH_MTHD(push, NVC36F, SEM_ADDR_LO, lower_32_bits(virtual), + SEM_ADDR_HI, upper_32_bits(virtual), + SEM_PAYLOAD_LO, sequence); + + PUSH_MTHD(push, NVC36F, SEM_EXECUTE, + NVDEF(NVC36F, SEM_EXECUTE, OPERATION, ACQ_CIRC_GEQ) | + NVDEF(NVC36F, SEM_EXECUTE, ACQUIRE_SWITCH_TSG, EN) | + NVDEF(NVC36F, SEM_EXECUTE, PAYLOAD_SIZE, 32BIT)); + + PUSH_KICK(push); + return 0; +} + +static int +gv100_fence_context_new(struct nouveau_channel *chan) +{ + struct nv84_fence_chan *fctx; + int ret; + + ret = nv84_fence_context_new(chan); + if (ret) + return ret; + + fctx = chan->fence; + fctx->base.emit32 = gv100_fence_emit32; + fctx->base.sync32 = gv100_fence_sync32; + return 0; +} + +int +gv100_fence_create(struct nouveau_drm *drm) +{ + struct nv84_fence_priv *priv; + int ret; + + ret = nv84_fence_create(drm); + if (ret) + return ret; + + priv = drm->fence; + priv->base.context_new = gv100_fence_context_new; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h b/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h new file mode 100644 index 000000000000..8735dda4c8a7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h @@ -0,0 +1,52 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef _clc36f_h_ +#define _clc36f_h_ + +#define NVC36F_NON_STALL_INTERRUPT (0x00000020) +#define NVC36F_NON_STALL_INTERRUPT_HANDLE 31:0 +#define NVC36F_SEM_ADDR_LO (0x0000005c) +#define NVC36F_SEM_ADDR_LO_OFFSET 31:2 +#define NVC36F_SEM_ADDR_HI (0x00000060) +#define NVC36F_SEM_ADDR_HI_OFFSET 7:0 +#define NVC36F_SEM_PAYLOAD_LO (0x00000064) +#define NVC36F_SEM_PAYLOAD_LO_PAYLOAD 31:0 +#define NVC36F_SEM_PAYLOAD_HI (0x00000068) +#define NVC36F_SEM_PAYLOAD_HI_PAYLOAD 31:0 +#define NVC36F_SEM_EXECUTE (0x0000006c) +#define NVC36F_SEM_EXECUTE_OPERATION 2:0 +#define NVC36F_SEM_EXECUTE_OPERATION_ACQUIRE 0x00000000 +#define NVC36F_SEM_EXECUTE_OPERATION_RELEASE 0x00000001 +#define NVC36F_SEM_EXECUTE_OPERATION_ACQ_STRICT_GEQ 0x00000002 +#define NVC36F_SEM_EXECUTE_OPERATION_ACQ_CIRC_GEQ 0x00000003 +#define NVC36F_SEM_EXECUTE_OPERATION_ACQ_AND 0x00000004 +#define NVC36F_SEM_EXECUTE_OPERATION_ACQ_NOR 0x00000005 +#define NVC36F_SEM_EXECUTE_OPERATION_REDUCTION 0x00000006 +#define NVC36F_SEM_EXECUTE_ACQUIRE_SWITCH_TSG 12:12 +#define NVC36F_SEM_EXECUTE_ACQUIRE_SWITCH_TSG_DIS 0x00000000 +#define NVC36F_SEM_EXECUTE_ACQUIRE_SWITCH_TSG_EN 0x00000001 +#define NVC36F_SEM_EXECUTE_RELEASE_WFI 20:20 +#define NVC36F_SEM_EXECUTE_RELEASE_WFI_DIS 0x00000000 +#define NVC36F_SEM_EXECUTE_RELEASE_WFI_EN 0x00000001 +#define NVC36F_SEM_EXECUTE_PAYLOAD_SIZE 24:24 +#define NVC36F_SEM_EXECUTE_PAYLOAD_SIZE_32BIT 0x00000000 +#define NVC36F_SEM_EXECUTE_PAYLOAD_SIZE_64BIT 0x00000001 +#define NVC36F_SEM_EXECUTE_RELEASE_TIMESTAMP 25:25 +#define NVC36F_SEM_EXECUTE_RELEASE_TIMESTAMP_DIS 0x00000000 +#define NVC36F_SEM_EXECUTE_RELEASE_TIMESTAMP_EN 0x00000001 +#define NVC36F_SEM_EXECUTE_REDUCTION 30:27 +#define NVC36F_SEM_EXECUTE_REDUCTION_IMIN 0x00000000 +#define NVC36F_SEM_EXECUTE_REDUCTION_IMAX 0x00000001 +#define NVC36F_SEM_EXECUTE_REDUCTION_IXOR 0x00000002 +#define NVC36F_SEM_EXECUTE_REDUCTION_IAND 0x00000003 +#define NVC36F_SEM_EXECUTE_REDUCTION_IOR 0x00000004 +#define NVC36F_SEM_EXECUTE_REDUCTION_IADD 0x00000005 +#define NVC36F_SEM_EXECUTE_REDUCTION_INC 0x00000006 +#define NVC36F_SEM_EXECUTE_REDUCTION_DEC 0x00000007 +#define NVC36F_SEM_EXECUTE_REDUCTION_FORMAT 31:31 +#define NVC36F_SEM_EXECUTE_REDUCTION_FORMAT_SIGNED 0x00000000 +#define NVC36F_SEM_EXECUTE_REDUCTION_FORMAT_UNSIGNED 0x00000001 + +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvhw/class/clc97b.h b/drivers/gpu/drm/nouveau/include/nvhw/class/clc97b.h new file mode 100644 index 000000000000..092aebe9551c --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/class/clc97b.h @@ -0,0 +1,22 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef _clc97b_h_ +#define _clc97b_h_ + +// dma opcode instructions +#define NVC97B_DMA +#define NVC97B_DMA_OPCODE 31:29 +#define NVC97B_DMA_OPCODE_METHOD 0x00000000 +#define NVC97B_DMA_OPCODE_JUMP 0x00000001 +#define NVC97B_DMA_OPCODE_NONINC_METHOD 0x00000002 +#define NVC97B_DMA_OPCODE_SET_SUBDEVICE_MASK 0x00000003 +#define NVC97B_DMA_METHOD_COUNT 27:18 +#define NVC97B_DMA_METHOD_OFFSET 15:2 +#define NVC97B_DMA_DATA 31:0 +#define NVC97B_DMA_DATA_NOP 0x00000000 +#define NVC97B_DMA_JUMP_OFFSET 15:2 +#define NVC97B_DMA_SET_SUBDEVICE_MASK_VALUE 11:0 + +#endif // _clc97b_h diff --git a/drivers/gpu/drm/nouveau/include/nvhw/class/clca7d.h b/drivers/gpu/drm/nouveau/include/nvhw/class/clca7d.h new file mode 100644 index 000000000000..0fec6fc21d44 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/class/clca7d.h @@ -0,0 +1,868 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef _clca7d_h_ +#define _clca7d_h_ + +// class methods +#define NVCA7D_UPDATE (0x00000200) +#define NVCA7D_UPDATE_SPECIAL_HANDLING 21:20 +#define NVCA7D_UPDATE_SPECIAL_HANDLING_NONE (0x00000000) +#define NVCA7D_UPDATE_SPECIAL_HANDLING_INTERRUPT_RM (0x00000001) +#define NVCA7D_UPDATE_SPECIAL_HANDLING_MODE_SWITCH (0x00000002) +#define NVCA7D_UPDATE_SPECIAL_HANDLING_REASON 19:12 +#define NVCA7D_UPDATE_INHIBIT_INTERRUPTS 24:24 +#define NVCA7D_UPDATE_INHIBIT_INTERRUPTS_FALSE (0x00000000) +#define NVCA7D_UPDATE_INHIBIT_INTERRUPTS_TRUE (0x00000001) +#define NVCA7D_UPDATE_RELEASE_ELV 0:0 +#define NVCA7D_UPDATE_RELEASE_ELV_FALSE (0x00000000) +#define NVCA7D_UPDATE_RELEASE_ELV_TRUE (0x00000001) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN 8:4 +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_NONE (0x00000000) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN(i) (0x00000001 +(i)) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN__SIZE_1 16 +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_0 (0x00000001) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_1 (0x00000002) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_2 (0x00000003) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_3 (0x00000004) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_4 (0x00000005) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_5 (0x00000006) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_6 (0x00000007) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_7 (0x00000008) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_8 (0x00000009) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_9 (0x0000000A) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_A (0x0000000B) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_B (0x0000000C) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_C (0x0000000D) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_D (0x0000000E) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_E (0x0000000F) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_LOCK_PIN_F (0x00000010) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_FLIP_LOCK_0 (0x00000014) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_FLIP_LOCK_1 (0x00000015) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_FLIP_LOCK_2 (0x00000016) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_FLIP_LOCK_3 (0x00000017) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK(i) (0x00000018 +(i)) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK__SIZE_1 8 +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_0 (0x00000018) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_1 (0x00000019) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_2 (0x0000001A) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_3 (0x0000001B) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_4 (0x0000001C) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_5 (0x0000001D) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_6 (0x0000001E) +#define NVCA7D_UPDATE_FLIP_LOCK_PIN_INTERNAL_SCAN_LOCK_7 (0x0000001F) +#define NVCA7D_UPDATE_FORCE_FULLSCREEN 28:28 +#define NVCA7D_UPDATE_FORCE_FULLSCREEN_FALSE (0x00000000) +#define NVCA7D_UPDATE_FORCE_FULLSCREEN_TRUE (0x00000001) +#define NVCA7D_SET_NOTIFIER_CONTROL (0x0000020C) +#define NVCA7D_SET_NOTIFIER_CONTROL_MODE 0:0 +#define NVCA7D_SET_NOTIFIER_CONTROL_MODE_WRITE (0x00000000) +#define NVCA7D_SET_NOTIFIER_CONTROL_MODE_WRITE_AWAKEN (0x00000001) +#define NVCA7D_SET_NOTIFIER_CONTROL_NOTIFY 12:12 +#define NVCA7D_SET_NOTIFIER_CONTROL_NOTIFY_DISABLE (0x00000000) +#define NVCA7D_SET_NOTIFIER_CONTROL_NOTIFY_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS (0x00000218) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR(i) ((i)+0):((i)+0) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR__SIZE_1 8 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR0 0:0 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR0_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR0_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR1 1:1 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR1_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR1_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR2 2:2 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR2_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR2_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR3 3:3 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR3_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR3_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR4 4:4 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR4_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR4_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR5 5:5 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR5_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR5_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR6 6:6 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR6_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR6_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR7 7:7 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR7_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CURSOR7_ENABLE (0x00000001) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CORE 16:16 +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CORE_DISABLE (0x00000000) +#define NVCA7D_SET_INTERLOCK_FLAGS_INTERLOCK_WITH_CORE_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS (0x0000021C) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW(i) ((i)+0):((i)+0) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW__SIZE_1 32 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW0 0:0 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW0_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW0_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW1 1:1 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW1_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW1_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW2 2:2 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW2_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW2_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW3 3:3 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW3_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW3_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW4 4:4 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW4_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW4_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW5 5:5 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW5_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW5_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW6 6:6 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW6_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW6_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW7 7:7 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW7_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW7_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW8 8:8 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW8_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW8_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW9 9:9 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW9_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW9_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW10 10:10 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW10_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW10_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW11 11:11 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW11_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW11_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW12 12:12 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW12_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW12_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW13 13:13 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW13_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW13_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW14 14:14 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW14_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW14_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW15 15:15 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW15_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW15_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW16 16:16 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW16_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW16_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW17 17:17 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW17_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW17_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW18 18:18 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW18_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW18_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW19 19:19 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW19_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW19_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW20 20:20 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW20_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW20_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW21 21:21 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW21_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW21_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW22 22:22 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW22_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW22_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW23 23:23 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW23_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW23_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW24 24:24 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW24_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW24_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW25 25:25 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW25_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW25_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW26 26:26 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW26_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW26_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW27 27:27 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW27_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW27_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW28 28:28 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW28_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW28_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW29 29:29 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW29_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW29_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW30 30:30 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW30_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW30_ENABLE (0x00000001) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW31 31:31 +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW31_DISABLE (0x00000000) +#define NVCA7D_SET_WINDOW_INTERLOCK_FLAGS_INTERLOCK_WITH_WINDOW31_ENABLE (0x00000001) +#define NVCA7D_SET_SURFACE_ADDRESS_HI_NOTIFIER (0x00000260) +#define NVCA7D_SET_SURFACE_ADDRESS_HI_NOTIFIER_ADDRESS_HI 31:0 +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER (0x00000264) +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_ADDRESS_LO 31:4 +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET 3:2 +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_IOVA (0x00000000) +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_ENABLE 0:0 +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_ENABLE_DISABLE (0x00000000) +#define NVCA7D_SET_SURFACE_ADDRESS_LO_NOTIFIER_ENABLE_ENABLE (0x00000001) + +#define NVCA7D_SOR_SET_CONTROL(a) (0x00000300 + (a)*0x00000020) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK 7:0 +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_NONE (0x00000000) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD0 (0x00000001) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD1 (0x00000002) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD2 (0x00000004) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD3 (0x00000008) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD4 (0x00000010) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD5 (0x00000020) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD6 (0x00000040) +#define NVCA7D_SOR_SET_CONTROL_OWNER_MASK_HEAD7 (0x00000080) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL 11:8 +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_LVDS_CUSTOM (0x00000000) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_SINGLE_TMDS_A (0x00000001) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_SINGLE_TMDS_B (0x00000002) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_DUAL_TMDS (0x00000005) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_DP_A (0x00000008) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_DP_B (0x00000009) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_HDMI_FRL (0x0000000C) +#define NVCA7D_SOR_SET_CONTROL_PROTOCOL_CUSTOM (0x0000000F) +#define NVCA7D_SOR_SET_CONTROL_DE_SYNC_POLARITY 16:16 +#define NVCA7D_SOR_SET_CONTROL_DE_SYNC_POLARITY_POSITIVE_TRUE (0x00000000) +#define NVCA7D_SOR_SET_CONTROL_DE_SYNC_POLARITY_NEGATIVE_TRUE (0x00000001) +#define NVCA7D_SOR_SET_CONTROL_PIXEL_REPLICATE_MODE 21:20 +#define NVCA7D_SOR_SET_CONTROL_PIXEL_REPLICATE_MODE_OFF (0x00000000) +#define NVCA7D_SOR_SET_CONTROL_PIXEL_REPLICATE_MODE_X2 (0x00000001) +#define NVCA7D_SOR_SET_CONTROL_PIXEL_REPLICATE_MODE_X4 (0x00000002) + +#define NVCA7D_WINDOW_SET_CONTROL(a) (0x00001000 + (a)*0x00000080) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER 3:0 +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD(i) (0x00000000 +(i)) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD__SIZE_1 8 +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD0 (0x00000000) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD1 (0x00000001) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD2 (0x00000002) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD3 (0x00000003) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD4 (0x00000004) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD5 (0x00000005) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD6 (0x00000006) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_HEAD7 (0x00000007) +#define NVCA7D_WINDOW_SET_CONTROL_OWNER_NONE (0x0000000F) +#define NVCA7D_WINDOW_SET_CONTROL_HIDE 8:8 +#define NVCA7D_WINDOW_SET_CONTROL_HIDE_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_CONTROL_HIDE_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_CONTROL_DISABLE_PHYSICAL_FLIPS 9:9 +#define NVCA7D_WINDOW_SET_CONTROL_DISABLE_PHYSICAL_FLIPS_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_CONTROL_DISABLE_PHYSICAL_FLIPS_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_CONTROL_ALLOW_SUPERFRAME 10:10 +#define NVCA7D_WINDOW_SET_CONTROL_ALLOW_SUPERFRAME_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_CONTROL_ALLOW_SUPERFRAME_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS(a) (0x00001004 + (a)*0x00000080) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED1BPP 0:0 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED1BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED1BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED2BPP 1:1 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED2BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED2BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED4BPP 2:2 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED4BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED4BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED8BPP 3:3 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED8BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_RGB_PACKED8BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PACKED422 4:4 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PACKED422_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PACKED422_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PLANAR420 5:5 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PLANAR444 6:6 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR420 7:7 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422 8:8 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422R 9:9 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422R_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422R_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR444 10:10 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR420 11:11 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR444 12:12 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR420 13:13 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422 14:14 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422R 15:15 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422R_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422R_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR444 16:16 +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS(a) (0x00001008 + (a)*0x00000080) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED1BPP 0:0 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED1BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED1BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED2BPP 1:1 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED2BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED2BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED4BPP 2:2 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED4BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED4BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED8BPP 3:3 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED8BPP_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_RGB_PACKED8BPP_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PACKED422 4:4 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PACKED422_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PACKED422_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PLANAR420 5:5 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PLANAR444 6:6 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR420 7:7 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422 8:8 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422R 9:9 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422R_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR422R_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR444 10:10 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_YUV_SEMI_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR420 11:11 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR444 12:12 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR420 13:13 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR420_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR420_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422 14:14 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422R 15:15 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422R_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR422R_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR444 16:16 +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR444_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_ROTATED_FORMAT_USAGE_BOUNDS_EXT_YUV_SEMI_PLANAR444_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS(a) (0x00001010 + (a)*0x00000080) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_MAX_PIXELS_FETCHED_PER_LINE 14:0 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_ILUT_ALLOWED 16:16 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_ILUT_ALLOWED_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_ILUT_ALLOWED_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_TMO_LUT_ALLOWED 28:28 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_TMO_LUT_ALLOWED_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_TMO_LUT_ALLOWED_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_INPUT_SCALER_TAPS 22:20 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_INPUT_SCALER_TAPS_TAPS_2 (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_INPUT_SCALER_TAPS_TAPS_5 (0x00000004) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_UPSCALING_ALLOWED 24:24 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_UPSCALING_ALLOWED_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_UPSCALING_ALLOWED_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_OVERFETCH_ENABLED 30:30 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_OVERFETCH_ENABLED_FALSE (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_OVERFETCH_ENABLED_TRUE (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_LAYOUT 26:25 +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_LAYOUT_PITCH_BLOCKLINEAR (0x00000000) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_LAYOUT_PITCH (0x00000001) +#define NVCA7D_WINDOW_SET_WINDOW_USAGE_BOUNDS_LAYOUT_BLOCKLINEAR (0x00000002) +#define NVCA7D_WINDOW_SET_PHYSICAL(a) (0x00001014 + (a)*0x00000080) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW 31:0 +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_NONE (0x00000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW0 (0x00000001) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW1 (0x00000002) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW2 (0x00000004) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW3 (0x00000008) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW4 (0x00000010) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW5 (0x00000020) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW6 (0x00000040) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW7 (0x00000080) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW8 (0x00000100) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW9 (0x00000200) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW10 (0x00000400) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW11 (0x00000800) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW12 (0x00001000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW13 (0x00002000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW14 (0x00004000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW15 (0x00008000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW16 (0x00010000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW17 (0x00020000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW18 (0x00040000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW19 (0x00080000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW20 (0x00100000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW21 (0x00200000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW22 (0x00400000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW23 (0x00800000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW24 (0x01000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW25 (0x02000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW26 (0x04000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW27 (0x08000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW28 (0x10000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW29 (0x20000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW30 (0x40000000) +#define NVCA7D_WINDOW_SET_PHYSICAL_WINDOW_WINDOW31 (0x80000000) + +#define NVCA7D_HEAD_SET_PROCAMP(a) (0x00002000 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_PROCAMP_COLOR_SPACE 1:0 +#define NVCA7D_HEAD_SET_PROCAMP_COLOR_SPACE_RGB (0x00000000) +#define NVCA7D_HEAD_SET_PROCAMP_COLOR_SPACE_YUV_601 (0x00000001) +#define NVCA7D_HEAD_SET_PROCAMP_COLOR_SPACE_YUV_709 (0x00000002) +#define NVCA7D_HEAD_SET_PROCAMP_COLOR_SPACE_YUV_2020 (0x00000003) +#define NVCA7D_HEAD_SET_PROCAMP_CHROMA_LPF 3:3 +#define NVCA7D_HEAD_SET_PROCAMP_CHROMA_LPF_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_PROCAMP_CHROMA_LPF_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_PROCAMP_CHROMA_DOWN_V 4:4 +#define NVCA7D_HEAD_SET_PROCAMP_CHROMA_DOWN_V_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_PROCAMP_CHROMA_DOWN_V_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_PROCAMP_DYNAMIC_RANGE 28:28 +#define NVCA7D_HEAD_SET_PROCAMP_DYNAMIC_RANGE_VESA (0x00000000) +#define NVCA7D_HEAD_SET_PROCAMP_DYNAMIC_RANGE_CEA (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE(a) (0x00002004 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_CRC_MODE 1:0 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_CRC_MODE_ACTIVE_RASTER (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_CRC_MODE_COMPLETE_RASTER (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_CRC_MODE_NON_ACTIVE_RASTER (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_HSYNC_POLARITY 2:2 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_HSYNC_POLARITY_POSITIVE_TRUE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_HSYNC_POLARITY_NEGATIVE_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_VSYNC_POLARITY 3:3 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_VSYNC_POLARITY_POSITIVE_TRUE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_VSYNC_POLARITY_NEGATIVE_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH 7:4 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_16_422 (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_18_444 (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_20_422 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_24_422 (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_24_444 (0x00000004) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_30_444 (0x00000005) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_32_422 (0x00000006) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_36_444 (0x00000007) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_48_444 (0x00000008) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_16_444 (0x00000009) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_PIXEL_DEPTH_BPP_18_444NP (0x0000000A) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_COLOR_SPACE_OVERRIDE 24:24 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_COLOR_SPACE_OVERRIDE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_COLOR_SPACE_OVERRIDE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_COLOR_SPACE_FLAG 23:12 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN 31:26 +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN0 (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN1 (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN2 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN3 (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN4 (0x00000004) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN5 (0x00000005) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN6 (0x00000006) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN7 (0x00000007) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN8 (0x00000008) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN9 (0x00000009) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN10 (0x0000000A) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN11 (0x0000000B) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN12 (0x0000000C) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN13 (0x0000000D) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN14 (0x0000000E) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN15 (0x0000000F) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN16 (0x00000010) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN17 (0x00000011) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN18 (0x00000012) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN19 (0x00000013) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN20 (0x00000014) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN21 (0x00000015) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN22 (0x00000016) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN23 (0x00000017) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN24 (0x00000018) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN25 (0x00000019) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN26 (0x0000001A) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN27 (0x0000001B) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN28 (0x0000001C) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN29 (0x0000001D) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN30 (0x0000001E) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_WIN31 (0x0000001F) +#define NVCA7D_HEAD_SET_CONTROL_OUTPUT_RESOURCE_EXT_PACKET_WIN_NONE (0x0000003F) +#define NVCA7D_HEAD_SET_CONTROL(a) (0x00002008 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_CONTROL_STRUCTURE 1:0 +#define NVCA7D_HEAD_SET_CONTROL_STRUCTURE_PROGRESSIVE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_STEREO3D_STRUCTURE 2:2 +#define NVCA7D_HEAD_SET_CONTROL_STEREO3D_STRUCTURE_NORMAL (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_STEREO3D_STRUCTURE_FRAME_PACKED (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_YUV420PACKER 3:3 +#define NVCA7D_HEAD_SET_CONTROL_YUV420PACKER_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_YUV420PACKER_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_MODE 11:10 +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_MODE_NO_LOCK (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_MODE_FRAME_LOCK (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_MODE_RASTER_LOCK (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN 8:4 +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_NONE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN(i) (0x00000001 +(i)) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN__SIZE_1 16 +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_0 (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_1 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_2 (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_3 (0x00000004) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_4 (0x00000005) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_5 (0x00000006) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_6 (0x00000007) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_7 (0x00000008) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_8 (0x00000009) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_9 (0x0000000A) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_A (0x0000000B) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_B (0x0000000C) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_C (0x0000000D) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_D (0x0000000E) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_E (0x0000000F) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_LOCK_PIN_F (0x00000010) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_FLIP_LOCK_0 (0x00000014) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_FLIP_LOCK_1 (0x00000015) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_FLIP_LOCK_2 (0x00000016) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_FLIP_LOCK_3 (0x00000017) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK(i) (0x00000018 +(i)) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK__SIZE_1 8 +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_0 (0x00000018) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_1 (0x00000019) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_2 (0x0000001A) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_3 (0x0000001B) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_4 (0x0000001C) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_5 (0x0000001D) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_6 (0x0000001E) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCK_PIN_INTERNAL_SCAN_LOCK_7 (0x0000001F) +#define NVCA7D_HEAD_SET_CONTROL_SINK_LOCKOUT_WINDOW 15:12 +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_MODE 23:22 +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_MODE_NO_LOCK (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_MODE_FRAME_LOCK (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_MODE_RASTER_LOCK (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN 20:16 +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_NONE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN(i) (0x00000001 +(i)) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN__SIZE_1 16 +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_0 (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_1 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_2 (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_3 (0x00000004) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_4 (0x00000005) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_5 (0x00000006) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_6 (0x00000007) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_7 (0x00000008) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_8 (0x00000009) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_9 (0x0000000A) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_A (0x0000000B) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_B (0x0000000C) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_C (0x0000000D) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_D (0x0000000E) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_E (0x0000000F) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_LOCK_PIN_F (0x00000010) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_FLIP_LOCK_0 (0x00000014) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_FLIP_LOCK_1 (0x00000015) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_FLIP_LOCK_2 (0x00000016) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_FLIP_LOCK_3 (0x00000017) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK(i) (0x00000018 +(i)) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK__SIZE_1 8 +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_0 (0x00000018) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_1 (0x00000019) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_2 (0x0000001A) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_3 (0x0000001B) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_4 (0x0000001C) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_5 (0x0000001D) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_6 (0x0000001E) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_LOCK_PIN_INTERNAL_SCAN_LOCK_7 (0x0000001F) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN 28:24 +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_NONE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN(i) (0x00000001 +(i)) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN__SIZE_1 16 +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_0 (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_1 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_2 (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_3 (0x00000004) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_4 (0x00000005) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_5 (0x00000006) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_6 (0x00000007) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_7 (0x00000008) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_8 (0x00000009) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_9 (0x0000000A) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_A (0x0000000B) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_B (0x0000000C) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_C (0x0000000D) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_D (0x0000000E) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_E (0x0000000F) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_LOCK_PIN_F (0x00000010) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_FLIP_LOCK_0 (0x00000014) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_FLIP_LOCK_1 (0x00000015) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_FLIP_LOCK_2 (0x00000016) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_FLIP_LOCK_3 (0x00000017) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK(i) (0x00000018 +(i)) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK__SIZE_1 8 +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_0 (0x00000018) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_1 (0x00000019) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_2 (0x0000001A) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_3 (0x0000001B) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_4 (0x0000001C) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_5 (0x0000001D) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_6 (0x0000001E) +#define NVCA7D_HEAD_SET_CONTROL_STEREO_PIN_INTERNAL_SCAN_LOCK_7 (0x0000001F) +#define NVCA7D_HEAD_SET_CONTROL_SINK_STEREO_LOCK_MODE 30:30 +#define NVCA7D_HEAD_SET_CONTROL_SINK_STEREO_LOCK_MODE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_SINK_STEREO_LOCK_MODE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_STEREO_LOCK_MODE 31:31 +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_STEREO_LOCK_MODE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_SOURCE_STEREO_LOCK_MODE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY(a) (0x0000200C + (a)*0x00000800) +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_HERTZ 30:0 +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_ADJ1000DIV1001 31:31 +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_ADJ1000DIV1001_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_ADJ1000DIV1001_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_DITHER_CONTROL(a) (0x00002018 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_ENABLE 0:0 +#define NVCA7D_HEAD_SET_DITHER_CONTROL_ENABLE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_ENABLE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_BITS 5:4 +#define NVCA7D_HEAD_SET_DITHER_CONTROL_BITS_TO_6_BITS (0x00000000) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_BITS_TO_8_BITS (0x00000001) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_BITS_TO_10_BITS (0x00000002) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_BITS_TO_12_BITS (0x00000003) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_OFFSET_ENABLE 2:2 +#define NVCA7D_HEAD_SET_DITHER_CONTROL_OFFSET_ENABLE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_OFFSET_ENABLE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE 10:8 +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE_DYNAMIC_ERR_ACC (0x00000000) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE_STATIC_ERR_ACC (0x00000001) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE_DYNAMIC_2X2 (0x00000002) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE_STATIC_2X2 (0x00000003) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE_TEMPORAL (0x00000004) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_MODE_ROUND (0x00000005) +#define NVCA7D_HEAD_SET_DITHER_CONTROL_PHASE 13:12 +#define NVCA7D_HEAD_SET_DISPLAY_ID(a,b) (0x00002020 + (a)*0x00000800 + (b)*0x00000004) +#define NVCA7D_HEAD_SET_DISPLAY_ID_CODE 31:0 +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX(a) (0x00002028 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX_HERTZ 30:0 +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX_ADJ1000DIV1001 31:31 +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX_ADJ1000DIV1001_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_PIXEL_CLOCK_FREQUENCY_MAX_ADJ1000DIV1001_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS(a) (0x00002030 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_CURSOR 2:0 +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_CURSOR_USAGE_NONE (0x00000000) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_CURSOR_USAGE_W32_H32 (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_CURSOR_USAGE_W64_H64 (0x00000002) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_CURSOR_USAGE_W128_H128 (0x00000003) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_CURSOR_USAGE_W256_H256 (0x00000004) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OLUT_ALLOWED 4:4 +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OLUT_ALLOWED_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OLUT_ALLOWED_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_LTM_ALLOWED 5:5 +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_LTM_ALLOWED_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_LTM_ALLOWED_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OUTPUT_SCALER_TAPS 14:12 +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OUTPUT_SCALER_TAPS_TAPS_2 (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OUTPUT_SCALER_TAPS_TAPS_5 (0x00000004) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_UPSCALING_ALLOWED 8:8 +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_UPSCALING_ALLOWED_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_UPSCALING_ALLOWED_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OVERFETCH_ENABLED 16:16 +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OVERFETCH_ENABLED_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_OVERFETCH_ENABLED_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_HEAD_USAGE_BOUNDS_ELV_START 31:17 +#define NVCA7D_HEAD_SET_VIEWPORT_SIZE_IN(a) (0x0000204C + (a)*0x00000800) +#define NVCA7D_HEAD_SET_VIEWPORT_SIZE_IN_WIDTH 14:0 +#define NVCA7D_HEAD_SET_VIEWPORT_SIZE_IN_HEIGHT 30:16 +#define NVCA7D_HEAD_SET_VIEWPORT_SIZE_OUT(a) (0x00002058 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_VIEWPORT_SIZE_OUT_WIDTH 14:0 +#define NVCA7D_HEAD_SET_VIEWPORT_SIZE_OUT_HEIGHT 30:16 +#define NVCA7D_HEAD_SET_TILE_MASK(a) (0x00002060 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE 7:0 +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_NONE (0x00000000) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE0 (0x00000001) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE1 (0x00000002) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE2 (0x00000004) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE3 (0x00000008) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE4 (0x00000010) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE5 (0x00000020) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE6 (0x00000040) +#define NVCA7D_HEAD_SET_TILE_MASK_TILE_TILE7 (0x00000080) +#define NVCA7D_HEAD_SET_RASTER_SIZE(a) (0x00002064 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_RASTER_SIZE_WIDTH 15:0 +#define NVCA7D_HEAD_SET_RASTER_SIZE_HEIGHT 31:16 +#define NVCA7D_HEAD_SET_RASTER_SYNC_END(a) (0x00002068 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_RASTER_SYNC_END_X 14:0 +#define NVCA7D_HEAD_SET_RASTER_SYNC_END_Y 30:16 +#define NVCA7D_HEAD_SET_RASTER_BLANK_END(a) (0x0000206C + (a)*0x00000800) +#define NVCA7D_HEAD_SET_RASTER_BLANK_END_X 14:0 +#define NVCA7D_HEAD_SET_RASTER_BLANK_END_Y 30:16 +#define NVCA7D_HEAD_SET_RASTER_BLANK_START(a) (0x00002070 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_RASTER_BLANK_START_X 14:0 +#define NVCA7D_HEAD_SET_RASTER_BLANK_START_Y 30:16 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR(a) (0x0000209C + (a)*0x00000800) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_ENABLE 31:31 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_ENABLE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_ENABLE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_FORMAT 7:0 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_FORMAT_A1R5G5B5 (0x000000E9) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_FORMAT_A8R8G8B8 (0x000000CF) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_SIZE 9:8 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_SIZE_W32_H32 (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_SIZE_W64_H64 (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_SIZE_W128_H128 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_SIZE_W256_H256 (0x00000003) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_HOT_SPOT_X 19:12 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_HOT_SPOT_Y 27:20 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION(a) (0x000020A0 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_K1 7:0 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_CURSOR_COLOR_FACTOR_SELECT 11:8 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_CURSOR_COLOR_FACTOR_SELECT_K1 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_CURSOR_COLOR_FACTOR_SELECT_K1_TIMES_SRC (0x00000005) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_VIEWPORT_COLOR_FACTOR_SELECT 15:12 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_VIEWPORT_COLOR_FACTOR_SELECT_ZERO (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_VIEWPORT_COLOR_FACTOR_SELECT_K1 (0x00000002) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_VIEWPORT_COLOR_FACTOR_SELECT_NEG_K1_TIMES_SRC (0x00000007) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_MODE 16:16 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_MODE_BLEND (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_MODE_XOR (0x00000001) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_BYPASS 20:20 +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_BYPASS_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_CONTROL_CURSOR_COMPOSITION_BYPASS_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_HI_CRC(a) (0x00002150 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_HI_CRC_ADDRESS_HI 31:0 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC(a) (0x00002154 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_ADDRESS_LO 31:4 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_TARGET 3:2 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_TARGET_IOVA (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_ENABLE 0:0 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_ENABLE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CRC_ENABLE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_HI_OLUT(a) (0x00002158 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_HI_OLUT_ADDRESS_HI 31:0 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT(a) (0x0000215C + (a)*0x00000800) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_ADDRESS_LO 31:4 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_TARGET 3:2 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_TARGET_IOVA (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_ENABLE 0:0 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_ENABLE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_OLUT_ENABLE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_HI_CURSOR(a,b) (0x00002170 + (a)*0x00000800 + (b)*0x00000004) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_HI_CURSOR_ADDRESS_HI 31:0 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR(a,b) (0x00002178 + (a)*0x00000800 + (b)*0x00000004) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_ADDRESS_LO 31:4 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_TARGET 3:2 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_TARGET_IOVA (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_KIND 1:1 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_KIND_PITCH (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_KIND_BLOCKLINEAR (0x00000001) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_ENABLE 0:0 +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_ENABLE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_SURFACE_ADDRESS_LO_CURSOR_ENABLE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_CRC_CONTROL(a) (0x00002184 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL 5:0 +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_0 (0x00000000) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_1 (0x00000001) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_2 (0x00000002) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_3 (0x00000003) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_4 (0x00000004) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_5 (0x00000005) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_6 (0x00000006) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_7 (0x00000007) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_8 (0x00000008) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_9 (0x00000009) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_10 (0x0000000A) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_11 (0x0000000B) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_12 (0x0000000C) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_13 (0x0000000D) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_14 (0x0000000E) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_15 (0x0000000F) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_16 (0x00000010) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_17 (0x00000011) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_18 (0x00000012) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_19 (0x00000013) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_20 (0x00000014) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_21 (0x00000015) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_22 (0x00000016) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_23 (0x00000017) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_24 (0x00000018) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_25 (0x00000019) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_26 (0x0000001A) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_27 (0x0000001B) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_28 (0x0000001C) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_29 (0x0000001D) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_30 (0x0000001E) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_WIN_31 (0x0000001F) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CONTROLLING_CHANNEL_CORE (0x00000020) +#define NVCA7D_HEAD_SET_CRC_CONTROL_EXPECT_BUFFER_COLLAPSE 8:8 +#define NVCA7D_HEAD_SET_CRC_CONTROL_EXPECT_BUFFER_COLLAPSE_FALSE (0x00000000) +#define NVCA7D_HEAD_SET_CRC_CONTROL_EXPECT_BUFFER_COLLAPSE_TRUE (0x00000001) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC 19:12 +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_NONE (0x00000000) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SF (0x00000030) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR(i) (0x00000050 +(i)) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR__SIZE_1 8 +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR0 (0x00000050) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR1 (0x00000051) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR2 (0x00000052) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR3 (0x00000053) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR4 (0x00000054) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR5 (0x00000055) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR6 (0x00000056) +#define NVCA7D_HEAD_SET_CRC_CONTROL_PRIMARY_CRC_SOR7 (0x00000057) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC 27:20 +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_NONE (0x00000000) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SF (0x00000030) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR(i) (0x00000050 +(i)) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR__SIZE_1 8 +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR0 (0x00000050) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR1 (0x00000051) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR2 (0x00000052) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR3 (0x00000053) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR4 (0x00000054) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR5 (0x00000055) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR6 (0x00000056) +#define NVCA7D_HEAD_SET_CRC_CONTROL_SECONDARY_CRC_SOR7 (0x00000057) +#define NVCA7D_HEAD_SET_CRC_CONTROL_CRC_DURING_SNOOZE 9:9 +#define NVCA7D_HEAD_SET_CRC_CONTROL_CRC_DURING_SNOOZE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_OLUT_CONTROL(a) (0x00002280 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_INTERPOLATE 0:0 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_INTERPOLATE_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_INTERPOLATE_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MIRROR 1:1 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MIRROR_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MIRROR_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MODE 3:2 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MODE_SEGMENTED (0x00000000) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MODE_DIRECT8 (0x00000001) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_MODE_DIRECT10 (0x00000002) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_SIZE 18:8 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_DIRECT_ROUND 4:4 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_DIRECT_ROUND_DISABLE (0x00000000) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_DIRECT_ROUND_ENABLE (0x00000001) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_LEVEL 25:20 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_SEGMENT_SIZE_BITS 5:5 +#define NVCA7D_HEAD_SET_OLUT_CONTROL_SEGMENT_SIZE_BITS_SIZE_3BITS (0x00000000) +#define NVCA7D_HEAD_SET_OLUT_CONTROL_SEGMENT_SIZE_BITS_SIZE_4BITS (0x00000001) +#define NVCA7D_HEAD_SET_OLUT_FP_NORM_SCALE(a) (0x00002284 + (a)*0x00000800) +#define NVCA7D_HEAD_SET_OLUT_FP_NORM_SCALE_VALUE 31:0 + +#define NVCA7D_TILE_SET_TILE_SIZE(a) (0x00006000 + (a)*0x00000200) +#define NVCA7D_TILE_SET_TILE_SIZE_START 14:0 +#define NVCA7D_TILE_SET_TILE_SIZE_WIDTH 30:16 + +#endif // _clca7d_h diff --git a/drivers/gpu/drm/nouveau/include/nvhw/class/clca7e.h b/drivers/gpu/drm/nouveau/include/nvhw/class/clca7e.h new file mode 100644 index 000000000000..ebfb2e48a4f4 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/class/clca7e.h @@ -0,0 +1,137 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef _clca7e_h_ +#define _clca7e_h_ + +// class methods +#define NVCA7E_SET_NOTIFIER_CONTROL (0x00000220) +#define NVCA7E_SET_NOTIFIER_CONTROL_MODE 0:0 +#define NVCA7E_SET_NOTIFIER_CONTROL_MODE_WRITE (0x00000000) +#define NVCA7E_SET_NOTIFIER_CONTROL_MODE_WRITE_AWAKEN (0x00000001) +#define NVCA7E_SET_SIZE (0x00000224) +#define NVCA7E_SET_SIZE_WIDTH 15:0 +#define NVCA7E_SET_SIZE_HEIGHT 31:16 +#define NVCA7E_SET_STORAGE (0x00000228) +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT 3:0 +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT_NVD_BLOCK_HEIGHT_ONE_GOB (0x00000000) +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT_NVD_BLOCK_HEIGHT_TWO_GOBS (0x00000001) +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT_NVD_BLOCK_HEIGHT_FOUR_GOBS (0x00000002) +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT_NVD_BLOCK_HEIGHT_EIGHT_GOBS (0x00000003) +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT_NVD_BLOCK_HEIGHT_SIXTEEN_GOBS (0x00000004) +#define NVCA7E_SET_STORAGE_BLOCK_HEIGHT_NVD_BLOCK_HEIGHT_THIRTYTWO_GOBS (0x00000005) +#define NVCA7E_SET_PARAMS (0x0000022C) +#define NVCA7E_SET_PARAMS_FORMAT 7:0 +#define NVCA7E_SET_PARAMS_FORMAT_I8 (0x0000001E) +#define NVCA7E_SET_PARAMS_FORMAT_R4G4B4A4 (0x0000002F) +#define NVCA7E_SET_PARAMS_FORMAT_R5G6B5 (0x000000E8) +#define NVCA7E_SET_PARAMS_FORMAT_A1R5G5B5 (0x000000E9) +#define NVCA7E_SET_PARAMS_FORMAT_R5G5B5A1 (0x0000002E) +#define NVCA7E_SET_PARAMS_FORMAT_A8R8G8B8 (0x000000CF) +#define NVCA7E_SET_PARAMS_FORMAT_X8R8G8B8 (0x000000E6) +#define NVCA7E_SET_PARAMS_FORMAT_A8B8G8R8 (0x000000D5) +#define NVCA7E_SET_PARAMS_FORMAT_X8B8G8R8 (0x000000F9) +#define NVCA7E_SET_PARAMS_FORMAT_A2R10G10B10 (0x000000DF) +#define NVCA7E_SET_PARAMS_FORMAT_A2B10G10R10 (0x000000D1) +#define NVCA7E_SET_PARAMS_FORMAT_R16_G16_B16_A16_NVBIAS (0x00000023) +#define NVCA7E_SET_PARAMS_FORMAT_R16_G16_B16_A16 (0x000000C6) +#define NVCA7E_SET_PARAMS_FORMAT_RF16_GF16_BF16_AF16 (0x000000CA) +#define NVCA7E_SET_PARAMS_FORMAT_Y8_U8__Y8_V8_N422 (0x00000028) +#define NVCA7E_SET_PARAMS_FORMAT_U8_Y8__V8_Y8_N422 (0x00000029) +#define NVCA7E_SET_PARAMS_FORMAT_Y8___U8V8_N444 (0x00000035) +#define NVCA7E_SET_PARAMS_FORMAT_Y8___U8V8_N422 (0x00000036) +#define NVCA7E_SET_PARAMS_FORMAT_Y8___V8U8_N420 (0x00000038) +#define NVCA7E_SET_PARAMS_FORMAT_Y8___U8___V8_N444 (0x0000003A) +#define NVCA7E_SET_PARAMS_FORMAT_Y8___U8___V8_N420 (0x0000003B) +#define NVCA7E_SET_PARAMS_FORMAT_Y10___U10V10_N444 (0x00000055) +#define NVCA7E_SET_PARAMS_FORMAT_Y10___U10V10_N422 (0x00000056) +#define NVCA7E_SET_PARAMS_FORMAT_Y10___V10U10_N420 (0x00000058) +#define NVCA7E_SET_PARAMS_FORMAT_Y12___U12V12_N444 (0x00000075) +#define NVCA7E_SET_PARAMS_FORMAT_Y12___U12V12_N422 (0x00000076) +#define NVCA7E_SET_PARAMS_FORMAT_Y12___V12U12_N420 (0x00000078) +#define NVCA7E_SET_PARAMS_CLAMP_BEFORE_BLEND 18:18 +#define NVCA7E_SET_PARAMS_CLAMP_BEFORE_BLEND_DISABLE (0x00000000) +#define NVCA7E_SET_PARAMS_CLAMP_BEFORE_BLEND_ENABLE (0x00000001) +#define NVCA7E_SET_PARAMS_SWAP_UV 19:19 +#define NVCA7E_SET_PARAMS_SWAP_UV_DISABLE (0x00000000) +#define NVCA7E_SET_PARAMS_SWAP_UV_ENABLE (0x00000001) +#define NVCA7E_SET_PARAMS_FMT_ROUNDING_MODE 22:22 +#define NVCA7E_SET_PARAMS_FMT_ROUNDING_MODE_ROUND_TO_NEAREST (0x00000000) +#define NVCA7E_SET_PARAMS_FMT_ROUNDING_MODE_ROUND_DOWN (0x00000001) +#define NVCA7E_SET_PLANAR_STORAGE(b) (0x00000230 + (b)*0x00000004) +#define NVCA7E_SET_PLANAR_STORAGE_PITCH 12:0 +#define NVCA7E_SET_POINT_IN(b) (0x00000290 + (b)*0x00000004) +#define NVCA7E_SET_POINT_IN_X 15:0 +#define NVCA7E_SET_POINT_IN_Y 31:16 +#define NVCA7E_SET_SIZE_IN (0x00000298) +#define NVCA7E_SET_SIZE_IN_WIDTH 15:0 +#define NVCA7E_SET_SIZE_IN_HEIGHT 31:16 +#define NVCA7E_SET_SIZE_OUT (0x000002A4) +#define NVCA7E_SET_SIZE_OUT_WIDTH 15:0 +#define NVCA7E_SET_SIZE_OUT_HEIGHT 31:16 +#define NVCA7E_SET_PRESENT_CONTROL (0x00000308) +#define NVCA7E_SET_PRESENT_CONTROL_MIN_PRESENT_INTERVAL 3:0 +#define NVCA7E_SET_PRESENT_CONTROL_BEGIN_MODE 6:4 +#define NVCA7E_SET_PRESENT_CONTROL_BEGIN_MODE_NON_TEARING (0x00000000) +#define NVCA7E_SET_PRESENT_CONTROL_BEGIN_MODE_IMMEDIATE (0x00000001) +#define NVCA7E_SET_PRESENT_CONTROL_TIMESTAMP_MODE 8:8 +#define NVCA7E_SET_PRESENT_CONTROL_TIMESTAMP_MODE_DISABLE (0x00000000) +#define NVCA7E_SET_PRESENT_CONTROL_TIMESTAMP_MODE_ENABLE (0x00000001) +#define NVCA7E_SET_PRESENT_CONTROL_STEREO_MODE 13:12 +#define NVCA7E_SET_PRESENT_CONTROL_STEREO_MODE_MONO (0x00000000) +#define NVCA7E_SET_PRESENT_CONTROL_STEREO_MODE_PAIR_FLIP (0x00000001) +#define NVCA7E_SET_PRESENT_CONTROL_STEREO_MODE_AT_ANY_FRAME (0x00000002) +#define NVCA7E_SET_ILUT_CONTROL (0x00000440) +#define NVCA7E_SET_ILUT_CONTROL_INTERPOLATE 0:0 +#define NVCA7E_SET_ILUT_CONTROL_INTERPOLATE_DISABLE (0x00000000) +#define NVCA7E_SET_ILUT_CONTROL_INTERPOLATE_ENABLE (0x00000001) +#define NVCA7E_SET_ILUT_CONTROL_MIRROR 1:1 +#define NVCA7E_SET_ILUT_CONTROL_MIRROR_DISABLE (0x00000000) +#define NVCA7E_SET_ILUT_CONTROL_MIRROR_ENABLE (0x00000001) +#define NVCA7E_SET_ILUT_CONTROL_MODE 3:2 +#define NVCA7E_SET_ILUT_CONTROL_MODE_SEGMENTED (0x00000000) +#define NVCA7E_SET_ILUT_CONTROL_MODE_DIRECT8 (0x00000001) +#define NVCA7E_SET_ILUT_CONTROL_MODE_DIRECT10 (0x00000002) +#define NVCA7E_SET_ILUT_CONTROL_SIZE 18:8 +#define NVCA7E_SET_SURFACE_ADDRESS_HI_NOTIFIER (0x00000650) +#define NVCA7E_SET_SURFACE_ADDRESS_HI_NOTIFIER_ADDRESS_HI 31:0 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER (0x00000654) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_ADDRESS_LO 31:4 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET 3:2 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_IOVA (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_ENABLE 0:0 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_ENABLE_DISABLE (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_NOTIFIER_ENABLE_ENABLE (0x00000001) +#define NVCA7E_SET_SURFACE_ADDRESS_HI_ISO(b) (0x00000658 + (b)*0x00000004) +#define NVCA7E_SET_SURFACE_ADDRESS_HI_ISO_ADDRESS_HI 31:0 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO(b) (0x00000670 + (b)*0x00000004) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_ADDRESS_LO 31:4 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_TARGET 3:2 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_TARGET_IOVA (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_KIND 1:1 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_KIND_PITCH (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_KIND_BLOCKLINEAR (0x00000001) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_ENABLE 0:0 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_ENABLE_DISABLE (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ISO_ENABLE_ENABLE (0x00000001) +#define NVCA7E_SET_SURFACE_ADDRESS_HI_ILUT (0x00000688) +#define NVCA7E_SET_SURFACE_ADDRESS_HI_ILUT_ADDRESS_HI 31:0 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT (0x0000068C) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_ADDRESS_LO 31:4 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_TARGET 3:2 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_TARGET_IOVA (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_TARGET_PHYSICAL_NVM (0x00000001) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_TARGET_PHYSICAL_PCI (0x00000002) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_TARGET_PHYSICAL_PCI_COHERENT (0x00000003) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_ENABLE 0:0 +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_ENABLE_DISABLE (0x00000000) +#define NVCA7E_SET_SURFACE_ADDRESS_LO_ILUT_ENABLE_ENABLE (0x00000001) + +#endif // _clca7e_h diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gb100/dev_hshub_base.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb100/dev_hshub_base.h new file mode 100644 index 000000000000..c9d74bd95e0b --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb100/dev_hshub_base.h @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gb100_dev_hshub_base_h__ +#define __gb100_dev_hshub_base_h__ + +#define NV_PFB_HSHUB0 0x00870fff:0x00870000 + +#define NV_PFB_HSHUB 0x00000FFF:0x00000000 /* RW--D */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_LO 0x00000E50 /* RW-4R */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR 31:0 /* RWIVF */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR_INIT 0x00000000 /* RWI-V */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR_MASK 0xFFFFFF00 /* ----V */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_HI 0x00000E54 /* RW-4R */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR 31:0 /* RWIVF */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_INIT 0x00000000 /* RWI-V */ +#define NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_MASK 0x000FFFFF /* ----V */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_LO 0x000006C0 /* RW-4R */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR 31:0 /* RWIVF */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR_INIT 0x00000000 /* RWI-V */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR_MASK 0xFFFFFF00 /* ----V */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_HI 0x000006C4 /* RW-4R */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR 31:0 /* RWIVF */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_INIT 0x00000000 /* RWI-V */ +#define NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_MASK 0x000FFFFF /* ----V */ + +#endif // __gb100_dev_hshub_base_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gb10b/dev_fbhub.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb10b/dev_fbhub.h new file mode 100644 index 000000000000..4d0bb8e14298 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb10b/dev_fbhub.h @@ -0,0 +1,18 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gb10b_dev_fb_h__ +#define __gb10b_dev_fb_h__ + +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_LO 0x008a1d58 /* RW-4R */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR 31:0 /* RWIVF */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR_INIT 0x00000000 /* RWI-V */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR_MASK 0xffffff00 /* RW--V */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_HI 0x008a1d5c /* RW-4R */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR 31:0 /* RWIVF */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_INIT 0x00000000 /* RWI-V */ +#define NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_MASK 0x000fffff /* RW--V */ + +#endif // __gb10b_dev_fb_h__ + diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gb202/dev_ce.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb202/dev_ce.h new file mode 100644 index 000000000000..b09f04b31738 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb202/dev_ce.h @@ -0,0 +1,12 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gb202_dev_ce_h__ +#define __gb202_dev_ce_h__ + +#define NV_CE_GRCE_MASK 0x001040d8 /* C--4R */ +#define NV_CE_GRCE_MASK_VALUE 9:0 /* C--VF */ +#define NV_CE_GRCE_MASK_VALUE_INIT 0x00f /* C---V */ + +#endif // __gb202_dev_ce_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gb202/dev_therm.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb202/dev_therm.h new file mode 100644 index 000000000000..ed359cb528fb --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gb202/dev_therm.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gb202_dev_therm_h__ +#define __gb202_dev_therm_h__ + +#define NV_THERM_I2CS_SCRATCH 0x00ad00bc /* RW-4R */ +#define NV_THERM_I2CS_SCRATCH_DATA 31:0 /* RWIVF */ +#define NV_THERM_I2CS_SCRATCH_DATA_INIT 0x00000000 /* RWI-V */ +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE NV_THERM_I2CS_SCRATCH +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS 31:0 +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS_SUCCESS 0x000000FF +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS_FAILED 0x00000000 + +#endif // __gb202_dev_therm_h__ + diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_falcon_v4.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_falcon_v4.h new file mode 100644 index 000000000000..52171b412aa1 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_falcon_v4.h @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_falcon_v4_h__ +#define __gh100_dev_falcon_v4_h__ + +#define NV_PFALCON_FALCON_MAILBOX0 0x00000040 /* RW-4R */ +#define NV_PFALCON_FALCON_MAILBOX0_DATA 31:0 /* RWIVF */ +#define NV_PFALCON_FALCON_MAILBOX0_DATA_INIT 0x00000000 /* RWI-V */ +#define NV_PFALCON_FALCON_MAILBOX1 0x00000044 /* RW-4R */ +#define NV_PFALCON_FALCON_MAILBOX1_DATA 31:0 /* RWIVF */ +#define NV_PFALCON_FALCON_MAILBOX1_DATA_INIT 0x00000000 /* RWI-V */ + +#define NV_PFALCON_FALCON_HWCFG2 0x000000f4 /* R--4R */ +#define NV_PFALCON_FALCON_HWCFG2_RISCV_BR_PRIV_LOCKDOWN 13:13 /* R--VF */ +#define NV_PFALCON_FALCON_HWCFG2_RISCV_BR_PRIV_LOCKDOWN_LOCK 0x00000001 /* R---V */ +#define NV_PFALCON_FALCON_HWCFG2_RISCV_BR_PRIV_LOCKDOWN_UNLOCK 0x00000000 /* R---V */ + +#endif // __gh100_dev_falcon_v4_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_fb.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_fb.h new file mode 100644 index 000000000000..819f09465952 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_fb.h @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_fb_h_ +#define __gh100_dev_fb_h_ + +#define NV_PFB_NISO_FLUSH_SYSMEM_ADDR_SHIFT 8 /* */ +#define NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_LO 0x00100A34 /* RW-4R */ +#define NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_LO_ADR 31:0 /* RWIVF */ +#define NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_HI 0x00100A38 /* RW-4R */ +#define NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR 31:0 /* RWIVF */ +#define NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_HI_ADR_MASK 0x000FFFFF /* ----V */ + +#endif // __gh100_dev_fb_h_ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_fsp_pri.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_fsp_pri.h new file mode 100644 index 000000000000..e9507242cae5 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_fsp_pri.h @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_fsp_pri_h__ +#define __gh100_dev_fsp_pri_h__ + +#define NV_PFSP 0x8F3FFF:0x8F0000 /* RW--D */ + +#define NV_PFSP_MSGQ_HEAD(i) (0x008F2c80+(i)*8) /* RW-4A */ +#define NV_PFSP_MSGQ_HEAD__SIZE_1 8 /* */ +#define NV_PFSP_MSGQ_HEAD_VAL 31:0 /* RWIUF */ +#define NV_PFSP_MSGQ_HEAD_VAL_INIT 0x00000000 /* RWI-V */ +#define NV_PFSP_MSGQ_TAIL(i) (0x008F2c84+(i)*8) /* RW-4A */ +#define NV_PFSP_MSGQ_TAIL__SIZE_1 8 /* */ +#define NV_PFSP_MSGQ_TAIL_VAL 31:0 /* RWIUF */ +#define NV_PFSP_MSGQ_TAIL_VAL_INIT 0x00000000 /* RWI-V */ + +#define NV_PFSP_QUEUE_HEAD(i) (0x008F2c00+(i)*8) /* RW-4A */ +#define NV_PFSP_QUEUE_HEAD__SIZE_1 8 /* */ +#define NV_PFSP_QUEUE_HEAD_ADDRESS 31:0 /* RWIVF */ +#define NV_PFSP_QUEUE_HEAD_ADDRESS_INIT 0x00000000 /* RWI-V */ +#define NV_PFSP_QUEUE_TAIL(i) (0x008F2c04+(i)*8) /* RW-4A */ +#define NV_PFSP_QUEUE_TAIL__SIZE_1 8 /* */ +#define NV_PFSP_QUEUE_TAIL_ADDRESS 31:0 /* RWIVF */ +#define NV_PFSP_QUEUE_TAIL_ADDRESS_INIT 0x00000000 /* RWI-V */ + +#endif // __gh100_dev_fsp_pri_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_mmu.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_mmu.h new file mode 100644 index 000000000000..6707e0e3b96b --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_mmu.h @@ -0,0 +1,173 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_mmu_h__ +#define __gh100_dev_mmu_h__ + +#define NV_MMU_PTE /* ----G */ +#define NV_MMU_PTE_APERTURE (1*32+2):(1*32+1) /* RWXVF */ +#define NV_MMU_PTE_APERTURE_VIDEO_MEMORY 0x00000000 /* RW--V */ +#define NV_MMU_PTE_APERTURE_PEER_MEMORY 0x00000001 /* RW--V */ +#define NV_MMU_PTE_APERTURE_SYSTEM_COHERENT_MEMORY 0x00000002 /* RW--V */ +#define NV_MMU_PTE_APERTURE_SYSTEM_NON_COHERENT_MEMORY 0x00000003 /* RW--V */ +#define NV_MMU_PTE_KIND (1*32+7):(1*32+4) /* RWXVF */ +#define NV_MMU_PTE_KIND_INVALID 0x07 /* R---V */ +#define NV_MMU_PTE_KIND_PITCH 0x00 /* R---V */ +#define NV_MMU_PTE_KIND_GENERIC_MEMORY 0x6 /* R---V */ +#define NV_MMU_PTE_KIND_Z16 0x1 /* R---V */ +#define NV_MMU_PTE_KIND_S8 0x2 /* R---V */ +#define NV_MMU_PTE_KIND_S8Z24 0x3 /* R---V */ +#define NV_MMU_PTE_KIND_ZF32_X24S8 0x4 /* R---V */ +#define NV_MMU_PTE_KIND_Z24S8 0x5 /* R---V */ +#define NV_MMU_PTE_KIND_GENERIC_MEMORY_COMPRESSIBLE 0x8 /* R---V */ +#define NV_MMU_PTE_KIND_GENERIC_MEMORY_COMPRESSIBLE_DISABLE_PLC 0x9 /* R---V */ +#define NV_MMU_PTE_KIND_S8_COMPRESSIBLE_DISABLE_PLC 0xA /* R---V */ +#define NV_MMU_PTE_KIND_Z16_COMPRESSIBLE_DISABLE_PLC 0xB /* R---V */ +#define NV_MMU_PTE_KIND_S8Z24_COMPRESSIBLE_DISABLE_PLC 0xC /* R---V */ +#define NV_MMU_PTE_KIND_ZF32_X24S8_COMPRESSIBLE_DISABLE_PLC 0xD /* R---V */ +#define NV_MMU_PTE_KIND_Z24S8_COMPRESSIBLE_DISABLE_PLC 0xE /* R---V */ +#define NV_MMU_PTE_KIND_SMSKED_MESSAGE 0xF /* R---V */ + +#define NV_MMU_VER3_PDE /* ----G */ +#define NV_MMU_VER3_PDE_IS_PTE 0:0 /* RWXVF */ +#define NV_MMU_VER3_PDE_IS_PTE_TRUE 0x1 /* RW--V */ +#define NV_MMU_VER3_PDE_IS_PTE_FALSE 0x0 /* RW--V */ +#define NV_MMU_VER3_PDE_VALID 0:0 /* RWXVF */ +#define NV_MMU_VER3_PDE_VALID_TRUE 0x1 /* RW--V */ +#define NV_MMU_VER3_PDE_VALID_FALSE 0x0 /* RW--V */ +#define NV_MMU_VER3_PDE_APERTURE 2:1 /* RWXVF */ +#define NV_MMU_VER3_PDE_APERTURE_INVALID 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PDE_APERTURE_VIDEO_MEMORY 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PDE_APERTURE_SYSTEM_COHERENT_MEMORY 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PDE_APERTURE_SYSTEM_NON_COHERENT_MEMORY 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF 5:3 /* RWXVF */ +#define NV_MMU_VER3_PDE_PCF_VALID_CACHED_ATS_ALLOWED__OR__INVALID_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_CACHED_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_INVALID_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_UNCACHED_ATS_ALLOWED__OR__SPARSE_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_UNCACHED_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_SPARSE_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_CACHED_ATS_NOT_ALLOWED__OR__INVALID_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_CACHED_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_INVALID_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_UNCACHED_ATS_NOT_ALLOWED__OR__SPARSE_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_VALID_UNCACHED_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PDE_PCF_SPARSE_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PDE_ADDRESS 51:12 /* RWXVF */ +#define NV_MMU_VER3_PDE_ADDRESS_SHIFT 0x0000000c /* */ +#define NV_MMU_VER3_PDE__SIZE 8 + +#define NV_MMU_VER3_DUAL_PDE /* ----G */ +#define NV_MMU_VER3_DUAL_PDE_IS_PTE 0:0 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_IS_PTE_TRUE 0x1 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_IS_PTE_FALSE 0x0 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_VALID 0:0 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_VALID_TRUE 0x1 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_VALID_FALSE 0x0 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_BIG 2:1 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_BIG_INVALID 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_BIG_VIDEO_MEMORY 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_BIG_SYSTEM_COHERENT_MEMORY 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_BIG_SYSTEM_NON_COHERENT_MEMORY 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG 5:3 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_CACHED_ATS_ALLOWED__OR__INVALID_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_CACHED_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_INVALID_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_UNCACHED_ATS_ALLOWED__OR__SPARSE_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_UNCACHED_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_SPARSE_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_CACHED_ATS_NOT_ALLOWED__OR__INVALID_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_CACHED_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_INVALID_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_UNCACHED_ATS_NOT_ALLOWED__OR__SPARSE_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_VALID_UNCACHED_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_BIG_SPARSE_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_ADDRESS_BIG 51:8 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_SMALL 66:65 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_SMALL_INVALID 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_SMALL_VIDEO_MEMORY 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_SMALL_SYSTEM_COHERENT_MEMORY 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_APERTURE_SMALL_SYSTEM_NON_COHERENT_MEMORY 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL 69:67 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_CACHED_ATS_ALLOWED__OR__INVALID_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_CACHED_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_INVALID_ATS_ALLOWED 0x00000000 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_UNCACHED_ATS_ALLOWED__OR__SPARSE_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_UNCACHED_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_SPARSE_ATS_ALLOWED 0x00000001 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_CACHED_ATS_NOT_ALLOWED__OR__INVALID_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_CACHED_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_INVALID_ATS_NOT_ALLOWED 0x00000002 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_UNCACHED_ATS_NOT_ALLOWED__OR__SPARSE_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_VALID_UNCACHED_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_PCF_SMALL_SPARSE_ATS_NOT_ALLOWED 0x00000003 /* RW--V */ +#define NV_MMU_VER3_DUAL_PDE_ADDRESS_SMALL 115:76 /* RWXVF */ +#define NV_MMU_VER3_DUAL_PDE_ADDRESS_SHIFT 0x0000000c /* */ +#define NV_MMU_VER3_DUAL_PDE_ADDRESS_BIG_SHIFT 8 /* */ +#define NV_MMU_VER3_DUAL_PDE__SIZE 16 + +#define NV_MMU_VER3_PTE /* ----G */ +#define NV_MMU_VER3_PTE_VALID 0:0 /* RWXVF */ +#define NV_MMU_VER3_PTE_VALID_TRUE 0x1 /* RW--V */ +#define NV_MMU_VER3_PTE_VALID_FALSE 0x0 /* RW--V */ +#define NV_MMU_VER3_PTE_APERTURE 2:1 /* RWXVF */ +#define NV_MMU_VER3_PTE_APERTURE_VIDEO_MEMORY 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PTE_APERTURE_PEER_MEMORY 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PTE_APERTURE_SYSTEM_COHERENT_MEMORY 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PTE_APERTURE_SYSTEM_NON_COHERENT_MEMORY 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF 7:3 /* RWXVF */ +#define NV_MMU_VER3_PTE_PCF_INVALID 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_SPARSE 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_MAPPING_NOWHERE 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_NO_VALID_4KB_PAGE 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_ATOMIC_CACHED_ACE 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_ATOMIC_UNCACHED_ACE 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_ATOMIC_CACHED_ACE 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_ATOMIC_UNCACHED_ACE 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_ATOMIC_CACHED_ACE 0x00000004 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_ATOMIC_UNCACHED_ACE 0x00000005 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_ATOMIC_CACHED_ACE 0x00000006 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_ATOMIC_UNCACHED_ACE 0x00000007 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_NO_ATOMIC_CACHED_ACE 0x00000008 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_NO_ATOMIC_UNCACHED_ACE 0x00000009 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_NO_ATOMIC_CACHED_ACE 0x0000000A /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_NO_ATOMIC_UNCACHED_ACE 0x0000000B /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_NO_ATOMIC_CACHED_ACE 0x0000000C /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_NO_ATOMIC_UNCACHED_ACE 0x0000000D /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_NO_ATOMIC_CACHED_ACE 0x0000000E /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_NO_ATOMIC_UNCACHED_ACE 0x0000000F /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_ATOMIC_CACHED_ACD 0x00000010 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_ATOMIC_UNCACHED_ACD 0x00000011 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_ATOMIC_CACHED_ACD 0x00000012 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_ATOMIC_UNCACHED_ACD 0x00000013 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_ATOMIC_CACHED_ACD 0x00000014 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_ATOMIC_UNCACHED_ACD 0x00000015 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_ATOMIC_CACHED_ACD 0x00000016 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_ATOMIC_UNCACHED_ACD 0x00000017 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_NO_ATOMIC_CACHED_ACD 0x00000018 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RW_NO_ATOMIC_UNCACHED_ACD 0x00000019 /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_NO_ATOMIC_CACHED_ACD 0x0000001A /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_NO_ATOMIC_UNCACHED_ACD 0x0000001B /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_NO_ATOMIC_CACHED_ACD 0x0000001C /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_REGULAR_RO_NO_ATOMIC_UNCACHED_ACD 0x0000001D /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_NO_ATOMIC_CACHED_ACD 0x0000001E /* RW--V */ +#define NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_NO_ATOMIC_UNCACHED_ACD 0x0000001F /* RW--V */ +#define NV_MMU_VER3_PTE_KIND 11:8 /* RWXVF */ +#define NV_MMU_VER3_PTE_ADDRESS 51:12 /* RWXVF */ +#define NV_MMU_VER3_PTE_ADDRESS_SYS 51:12 /* RWXVF */ +#define NV_MMU_VER3_PTE_ADDRESS_PEER 51:12 /* RWXVF */ +#define NV_MMU_VER3_PTE_ADDRESS_VID 39:12 /* RWXVF */ +#define NV_MMU_VER3_PTE_PEER_ID 63:(64-3) /* RWXVF */ +#define NV_MMU_VER3_PTE_PEER_ID_0 0x00000000 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_1 0x00000001 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_2 0x00000002 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_3 0x00000003 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_4 0x00000004 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_5 0x00000005 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_6 0x00000006 /* RW--V */ +#define NV_MMU_VER3_PTE_PEER_ID_7 0x00000007 /* RW--V */ +#define NV_MMU_VER3_PTE_ADDRESS_SHIFT 0x0000000c /* */ +#define NV_MMU_VER3_PTE__SIZE 8 + +#endif // __gh100_dev_mmu_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_riscv_pri.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_riscv_pri.h new file mode 100644 index 000000000000..8ff4663168d2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_riscv_pri.h @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_riscv_pri_h__ +#define __gh100_dev_riscv_pri_h__ + +#define NV_PRISCV_RISCV_CPUCTL 0x00000388 /* RW-4R */ +#define NV_PRISCV_RISCV_CPUCTL_HALTED 4:4 /* R-IVF */ +#define NV_PRISCV_RISCV_CPUCTL_HALTED_INIT 0x00000001 /* R-I-V */ +#define NV_PRISCV_RISCV_CPUCTL_HALTED_TRUE 0x00000001 /* R---V */ +#define NV_PRISCV_RISCV_CPUCTL_HALTED_FALSE 0x00000000 /* R---V */ + +#endif // __gh100_dev_riscv_pri_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_therm.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_therm.h new file mode 100644 index 000000000000..49b4816cb00b --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_therm.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_therm_h__ +#define __gh100_dev_therm_h__ + +#define NV_THERM_I2CS_SCRATCH 0x000200bc /* RW-4R */ +#define NV_THERM_I2CS_SCRATCH_DATA 31:0 /* RWIVF */ +#define NV_THERM_I2CS_SCRATCH_DATA_INIT 0x00000000 /* RWI-V */ + +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE NV_THERM_I2CS_SCRATCH +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS 31:0 +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS_SUCCESS 0x000000FF +#define NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS_FAILED 0x00000000 + +#endif // __gh100_dev_therm_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_xtl_ep_pri.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_xtl_ep_pri.h new file mode 100644 index 000000000000..12b49e9894a2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/dev_xtl_ep_pri.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_dev_xtl_ep_pri_h__ +#define __gh100_dev_xtl_ep_pri_h__ + +#define NV_EP_PCFGM 0x92FFF:0x92000 /* RW--D */ + +#endif // __gh100_dev_xtl_ep_pri_h__ diff --git a/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/pri_nv_xal_ep.h b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/pri_nv_xal_ep.h new file mode 100644 index 000000000000..1a891bd33fa3 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvhw/ref/gh100/pri_nv_xal_ep.h @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __gh100_pri_nv_xal_ep_h__ +#define __gh100_pri_nv_xal_ep_h__ + +#define NV_XAL_EP_BAR0_WINDOW_BASE_SHIFT 0x000010 +#define NV_XAL_EP_BAR0_WINDOW_BASE 21:0 +#define NV_XAL_EP_BAR0_WINDOW 0x0010fd40 + +#endif // __gh100_pri_nv_xal_ep_h__ + diff --git a/drivers/gpu/drm/nouveau/include/nvif/chan.h b/drivers/gpu/drm/nouveau/include/nvif/chan.h new file mode 100644 index 000000000000..c329a29068d5 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvif/chan.h @@ -0,0 +1,76 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVIF_CHAN_H__ +#define __NVIF_CHAN_H__ +#include "push.h" + +struct nvif_chan { + const struct nvif_chan_func { + struct { + u32 (*read_get)(struct nvif_chan *); + } push; + + struct { + u32 (*read_get)(struct nvif_chan *); + void (*push)(struct nvif_chan *, bool main, u64 addr, u32 size, + bool no_prefetch); + void (*kick)(struct nvif_chan *); + int (*post)(struct nvif_chan *, u32 gpptr, u32 pbptr); + u32 post_size; + } gpfifo; + + struct { + int (*release)(struct nvif_chan *, u64 addr, u32 data); + } sem; + } *func; + + struct { + struct nvif_map map; + } userd; + + struct { + struct nvif_map map; + u32 cur; + u32 max; + int free; + } gpfifo; + + struct { + struct nvif_map map; + u64 addr; + } sema; + + struct nvif_push push; + + struct nvif_user *usermode; + u32 doorbell_token; +}; + +int nvif_chan_dma_wait(struct nvif_chan *, u32 push_nr); + +void nvif_chan_gpfifo_ctor(const struct nvif_chan_func *, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, struct nvif_chan *); +int nvif_chan_gpfifo_wait(struct nvif_chan *, u32 gpfifo_nr, u32 push_nr); +void nvif_chan_gpfifo_push(struct nvif_chan *, u64 addr, u32 size, bool no_prefetch); +int nvif_chan_gpfifo_post(struct nvif_chan *); + +void nvif_chan506f_gpfifo_push(struct nvif_chan *, bool main, u64 addr, u32 size, bool no_prefetch); +void nvif_chan506f_gpfifo_kick(struct nvif_chan *); + +int nvif_chan906f_ctor_(const struct nvif_chan_func *, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, void *sema, u64 sema_addr, + struct nvif_chan *); +u32 nvif_chan906f_read_get(struct nvif_chan *); +u32 nvif_chan906f_gpfifo_read_get(struct nvif_chan *); +int nvif_chan906f_gpfifo_post(struct nvif_chan *, u32 gpptr, u32 pbptr); + +int nvif_chan506f_ctor(struct nvif_chan *, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size); +int nvif_chan906f_ctor(struct nvif_chan *, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, void *sema, u64 sema_addr); +int nvif_chanc36f_ctor(struct nvif_chan *, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, void *sema, u64 sema_addr, + struct nvif_user *usermode, u32 doorbell_token); +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h index ea937fa7bc55..ea8267e0d8da 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h +++ b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h @@ -29,6 +29,8 @@ struct nv_device_info_v0 { #define NV_DEVICE_INFO_V0_TURING 0x0c #define NV_DEVICE_INFO_V0_AMPERE 0x0d #define NV_DEVICE_INFO_V0_ADA 0x0e +#define NV_DEVICE_INFO_V0_HOPPER 0x0f +#define NV_DEVICE_INFO_V0_BLACKWELL 0x10 __u8 family; __u8 pad06[2]; __u64 ram_size; diff --git a/drivers/gpu/drm/nouveau/include/nvif/class.h b/drivers/gpu/drm/nouveau/include/nvif/class.h index 824e052dcc25..ff6823cb2cd8 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/class.h +++ b/drivers/gpu/drm/nouveau/include/nvif/class.h @@ -57,12 +57,15 @@ #define KEPLER_INLINE_TO_MEMORY_A 0x0000a040 #define KEPLER_INLINE_TO_MEMORY_B 0x0000a140 +#define BLACKWELL_INLINE_TO_MEMORY_A 0x0000cd40 #define NV04_DISP /* cl0046.h */ 0x00000046 #define VOLTA_USERMODE_A 0x0000c361 #define TURING_USERMODE_A 0x0000c461 #define AMPERE_USERMODE_A 0x0000c561 +#define HOPPER_USERMODE_A 0x0000c661 +#define BLACKWELL_USERMODE_A 0x0000c761 #define MAXWELL_FAULT_BUFFER_A /* clb069.h */ 0x0000b069 #define VOLTA_FAULT_BUFFER_A /* clb069.h */ 0x0000c369 @@ -85,6 +88,9 @@ #define TURING_CHANNEL_GPFIFO_A /* if0020.h */ 0x0000c46f #define AMPERE_CHANNEL_GPFIFO_A /* if0020.h */ 0x0000c56f #define AMPERE_CHANNEL_GPFIFO_B /* if0020.h */ 0x0000c76f +#define HOPPER_CHANNEL_GPFIFO_A 0x0000c86f +#define BLACKWELL_CHANNEL_GPFIFO_A 0x0000c96f +#define BLACKWELL_CHANNEL_GPFIFO_B 0x0000ca6f #define NV50_DISP /* if0010.h */ 0x00005070 #define G82_DISP /* if0010.h */ 0x00008270 @@ -102,8 +108,10 @@ #define TU102_DISP /* if0010.h */ 0x0000c570 #define GA102_DISP /* if0010.h */ 0x0000c670 #define AD102_DISP /* if0010.h */ 0x0000c770 +#define GB202_DISP 0x0000ca70 #define GV100_DISP_CAPS 0x0000c373 +#define GB202_DISP_CAPS 0x0000ca73 #define NV31_MPEG 0x00003174 #define G82_MPEG 0x00008274 @@ -118,6 +126,7 @@ #define GV100_DISP_CURSOR /* if0014.h */ 0x0000c37a #define TU102_DISP_CURSOR /* if0014.h */ 0x0000c57a #define GA102_DISP_CURSOR /* if0014.h */ 0x0000c67a +#define GB202_DISP_CURSOR 0x0000ca7a #define NV50_DISP_OVERLAY /* if0014.h */ 0x0000507b #define G82_DISP_OVERLAY /* if0014.h */ 0x0000827b @@ -128,6 +137,7 @@ #define GV100_DISP_WINDOW_IMM_CHANNEL_DMA /* if0014.h */ 0x0000c37b #define TU102_DISP_WINDOW_IMM_CHANNEL_DMA /* if0014.h */ 0x0000c57b #define GA102_DISP_WINDOW_IMM_CHANNEL_DMA /* if0014.h */ 0x0000c67b +#define GB202_DISP_WINDOW_IMM_CHANNEL_DMA 0x0000ca7b #define NV50_DISP_BASE_CHANNEL_DMA /* if0014.h */ 0x0000507c #define G82_DISP_BASE_CHANNEL_DMA /* if0014.h */ 0x0000827c @@ -153,6 +163,7 @@ #define TU102_DISP_CORE_CHANNEL_DMA /* if0014.h */ 0x0000c57d #define GA102_DISP_CORE_CHANNEL_DMA /* if0014.h */ 0x0000c67d #define AD102_DISP_CORE_CHANNEL_DMA /* if0014.h */ 0x0000c77d +#define GB202_DISP_CORE_CHANNEL_DMA 0x0000ca7d #define NV50_DISP_OVERLAY_CHANNEL_DMA /* if0014.h */ 0x0000507e #define G82_DISP_OVERLAY_CHANNEL_DMA /* if0014.h */ 0x0000827e @@ -164,6 +175,7 @@ #define GV100_DISP_WINDOW_CHANNEL_DMA /* if0014.h */ 0x0000c37e #define TU102_DISP_WINDOW_CHANNEL_DMA /* if0014.h */ 0x0000c57e #define GA102_DISP_WINDOW_CHANNEL_DMA /* if0014.h */ 0x0000c67e +#define GB202_DISP_WINDOW_CHANNEL_DMA 0x0000ca7e #define NV50_TESLA 0x00005097 #define G82_TESLA 0x00008297 @@ -189,16 +201,25 @@ #define TURING_A /* cl9097.h */ 0x0000c597 +#define AMPERE_A 0x0000c697 #define AMPERE_B /* cl9097.h */ 0x0000c797 #define ADA_A /* cl9097.h */ 0x0000c997 +#define HOPPER_A 0x0000cb97 + +#define BLACKWELL_A 0x0000cd97 +#define BLACKWELL_B 0x0000ce97 + #define NV74_BSP 0x000074b0 +#define NVB8B0_VIDEO_DECODER 0x0000b8b0 #define NVC4B0_VIDEO_DECODER 0x0000c4b0 #define NVC6B0_VIDEO_DECODER 0x0000c6b0 #define NVC7B0_VIDEO_DECODER 0x0000c7b0 #define NVC9B0_VIDEO_DECODER 0x0000c9b0 +#define NVCDB0_VIDEO_DECODER 0x0000cdb0 +#define NVCFB0_VIDEO_DECODER 0x0000cfb0 #define GT212_MSVLD 0x000085b1 #define IGT21A_MSVLD 0x000086b1 @@ -227,10 +248,14 @@ #define TURING_DMA_COPY_A 0x0000c5b5 #define AMPERE_DMA_COPY_A 0x0000c6b5 #define AMPERE_DMA_COPY_B 0x0000c7b5 +#define HOPPER_DMA_COPY_A 0x0000c8b5 +#define BLACKWELL_DMA_COPY_A 0x0000c9b5 +#define BLACKWELL_DMA_COPY_B 0x0000cab5 #define NVC4B7_VIDEO_ENCODER 0x0000c4b7 #define NVC7B7_VIDEO_ENCODER 0x0000c7b7 #define NVC9B7_VIDEO_ENCODER 0x0000c9b7 +#define NVCFB7_VIDEO_ENCODER 0x0000cfb7 #define FERMI_DECOMPRESS 0x000090b8 @@ -246,15 +271,25 @@ #define PASCAL_COMPUTE_B 0x0000c1c0 #define VOLTA_COMPUTE_A 0x0000c3c0 #define TURING_COMPUTE_A 0x0000c5c0 +#define AMPERE_COMPUTE_A 0x0000c6c0 #define AMPERE_COMPUTE_B 0x0000c7c0 #define ADA_COMPUTE_A 0x0000c9c0 +#define HOPPER_COMPUTE_A 0x0000cbc0 +#define BLACKWELL_COMPUTE_A 0x0000cdc0 +#define BLACKWELL_COMPUTE_B 0x0000cec0 #define NV74_CIPHER 0x000074c1 +#define NVB8D1_VIDEO_NVJPG 0x0000b8d1 #define NVC4D1_VIDEO_NVJPG 0x0000c4d1 #define NVC9D1_VIDEO_NVJPG 0x0000c9d1 +#define NVCDD1_VIDEO_NVJPG 0x0000cdd1 +#define NVCFD1_VIDEO_NVJPG 0x0000cfd1 +#define NVB8FA_VIDEO_OFA 0x0000b8fa #define NVC6FA_VIDEO_OFA 0x0000c6fa #define NVC7FA_VIDEO_OFA 0x0000c7fa #define NVC9FA_VIDEO_OFA 0x0000c9fa +#define NVCDFA_VIDEO_OFA 0x0000cdfa +#define NVCFFA_VIDEO_OFA 0x0000cffa #endif diff --git a/drivers/gpu/drm/nouveau/include/nvif/object.h b/drivers/gpu/drm/nouveau/include/nvif/object.h index 8d205b6af46a..1b32dc701f61 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/object.h +++ b/drivers/gpu/drm/nouveau/include/nvif/object.h @@ -16,7 +16,7 @@ struct nvif_object { u32 handle; s32 oclass; void *priv; /*XXX: hack */ - struct { + struct nvif_map { void __iomem *ptr; u64 size; } map; diff --git a/drivers/gpu/drm/nouveau/include/nvif/push.h b/drivers/gpu/drm/nouveau/include/nvif/push.h index 6d3a8a3d2087..a493fababe3c 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/push.h +++ b/drivers/gpu/drm/nouveau/include/nvif/push.h @@ -31,6 +31,12 @@ struct nvif_push { void (*kick)(struct nvif_push *push); struct nvif_mem mem; + u64 addr; + + struct { + u32 get; + u32 max; + } hw; u32 *bgn; u32 *cur; @@ -41,7 +47,7 @@ struct nvif_push { static inline __must_check int PUSH_WAIT(struct nvif_push *push, u32 size) { - if (push->cur + size >= push->end) { + if (push->cur + size > push->end) { int ret = push->wait(push, size); if (ret) return ret; @@ -55,7 +61,11 @@ PUSH_WAIT(struct nvif_push *push, u32 size) static inline int PUSH_KICK(struct nvif_push *push) { - push->kick(push); + if (push->cur != push->bgn) { + push->kick(push); + push->bgn = push->cur; + } + return 0; } diff --git a/drivers/gpu/drm/nouveau/include/nvif/push906f.h b/drivers/gpu/drm/nouveau/include/nvif/push906f.h index cc2866bc8b0a..79df71de98d2 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/push906f.h +++ b/drivers/gpu/drm/nouveau/include/nvif/push906f.h @@ -7,6 +7,7 @@ #ifndef PUSH906F_SUBC // Host methods #define PUSH906F_SUBC_NV906F 0 +#define PUSH906F_SUBC_NVC36F 0 // Twod #define PUSH906F_SUBC_NV902D 3 diff --git a/drivers/gpu/drm/nouveau/include/nvif/pushc97b.h b/drivers/gpu/drm/nouveau/include/nvif/pushc97b.h new file mode 100644 index 000000000000..c8d6b6319134 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvif/pushc97b.h @@ -0,0 +1,18 @@ +#ifndef __NVIF_PUSHC97B_H__ +#define __NVIF_PUSHC97B_H__ +#include <nvif/push.h> + +#include <nvhw/class/clc97b.h> + +#define PUSH_HDR(p,m,c) do { \ + PUSH_ASSERT(!((m) & ~DRF_SMASK(NVC97B_DMA_METHOD_OFFSET)), "mthd"); \ + PUSH_ASSERT(!((c) & ~DRF_MASK(NVC97B_DMA_METHOD_COUNT)), "size"); \ + PUSH_DATA__((p), NVDEF(NVC97B, DMA, OPCODE, METHOD) | \ + NVVAL(NVC97B, DMA, METHOD_COUNT, (c)) | \ + NVVAL(NVC97B, DMA, METHOD_OFFSET, (m) >> 2), \ + " mthd 0x%04x size %d - %s", (u32)(m), (u32)(c), __func__); \ +} while(0) + +#define PUSH_MTHD_HDR(p,s,m,c) PUSH_HDR(p,m,c) +#define PUSH_MTHD_INC 4:4 +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h index 46afb877a296..99579e7b9376 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h @@ -46,7 +46,10 @@ struct nvkm_device { GV100 = 0x140, TU100 = 0x160, GA100 = 0x170, + GH100 = 0x180, AD100 = 0x190, + GB10x = 0x1a0, + GB20x = 0x1b0, } card_type; u32 chipset; u8 chiprev; @@ -77,6 +80,13 @@ struct nvkm_device { struct nvkm_subdev *nvkm_device_subdev(struct nvkm_device *, int type, int inst); struct nvkm_engine *nvkm_device_engine(struct nvkm_device *, int type, int inst); +enum nvkm_bar_id { + NVKM_BAR_INVALID = 0, + NVKM_BAR0_PRI, + NVKM_BAR1_FB, + NVKM_BAR2_INST, +}; + struct nvkm_device_func { struct nvkm_device_pci *(*pci)(struct nvkm_device *); struct nvkm_device_tegra *(*tegra)(struct nvkm_device *); @@ -85,8 +95,8 @@ struct nvkm_device_func { int (*init)(struct nvkm_device *); void (*fini)(struct nvkm_device *, bool suspend); int (*irq)(struct nvkm_device *); - resource_size_t (*resource_addr)(struct nvkm_device *, unsigned bar); - resource_size_t (*resource_size)(struct nvkm_device *, unsigned bar); + resource_size_t (*resource_addr)(struct nvkm_device *, enum nvkm_bar_id); + resource_size_t (*resource_size)(struct nvkm_device *, enum nvkm_bar_id); bool cpu_coherent; }; @@ -124,6 +134,9 @@ struct nvkm_device *nvkm_device_find(u64 name); _temp; \ }) +#define NVKM_RD32_(p,o,dr) nvkm_rd32((p), (o) + (dr)) +#define NVKM_RD32(p,A...) DRF_RV(NVKM_RD32_, (p), 0, ##A) + void nvkm_device_del(struct nvkm_device **); struct nvkm_device_oclass { diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h b/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h index 9d2a1abf64f9..d92ffd17b729 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/core/layout.h @@ -1,4 +1,5 @@ /* SPDX-License-Identifier: MIT */ +NVKM_LAYOUT_ONCE(NVKM_SUBDEV_FSP , struct nvkm_fsp , fsp) NVKM_LAYOUT_ONCE(NVKM_SUBDEV_GSP , struct nvkm_gsp , gsp) NVKM_LAYOUT_ONCE(NVKM_SUBDEV_TOP , struct nvkm_top , top) NVKM_LAYOUT_ONCE(NVKM_SUBDEV_VFN , struct nvkm_vfn , vfn) @@ -29,7 +30,7 @@ NVKM_LAYOUT_INST(NVKM_SUBDEV_IOCTRL , struct nvkm_subdev , ioctrl, 3) NVKM_LAYOUT_ONCE(NVKM_SUBDEV_FLA , struct nvkm_subdev , fla) NVKM_LAYOUT_ONCE(NVKM_ENGINE_BSP , struct nvkm_engine , bsp) -NVKM_LAYOUT_INST(NVKM_ENGINE_CE , struct nvkm_engine , ce, 10) +NVKM_LAYOUT_INST(NVKM_ENGINE_CE , struct nvkm_engine , ce, 20) NVKM_LAYOUT_ONCE(NVKM_ENGINE_CIPHER , struct nvkm_engine , cipher) NVKM_LAYOUT_ONCE(NVKM_ENGINE_DISP , struct nvkm_disp , disp) NVKM_LAYOUT_ONCE(NVKM_ENGINE_DMAOBJ , struct nvkm_dma , dma) @@ -43,9 +44,9 @@ NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSPDEC , struct nvkm_engine , mspdec) NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSPPP , struct nvkm_engine , msppp) NVKM_LAYOUT_ONCE(NVKM_ENGINE_MSVLD , struct nvkm_engine , msvld) NVKM_LAYOUT_INST(NVKM_ENGINE_NVDEC , struct nvkm_nvdec , nvdec, 8) -NVKM_LAYOUT_INST(NVKM_ENGINE_NVENC , struct nvkm_nvenc , nvenc, 3) +NVKM_LAYOUT_INST(NVKM_ENGINE_NVENC , struct nvkm_nvenc , nvenc, 4) NVKM_LAYOUT_INST(NVKM_ENGINE_NVJPG , struct nvkm_engine , nvjpg, 8) -NVKM_LAYOUT_ONCE(NVKM_ENGINE_OFA , struct nvkm_engine , ofa) +NVKM_LAYOUT_INST(NVKM_ENGINE_OFA , struct nvkm_engine , ofa, 2) NVKM_LAYOUT_ONCE(NVKM_ENGINE_SEC , struct nvkm_engine , sec) NVKM_LAYOUT_ONCE(NVKM_ENGINE_SEC2 , struct nvkm_sec2 , sec2) NVKM_LAYOUT_ONCE(NVKM_ENGINE_SW , struct nvkm_sw , sw) diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h index 3e8db8280e2a..7903d7470d19 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h @@ -87,5 +87,4 @@ int gp102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct int gv100_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **); int tu102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **); int ga102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **); -int ad102_disp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_disp **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h index be508f65b280..96c16cfccf16 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h @@ -78,9 +78,6 @@ struct nvkm_fifo { struct { struct nvkm_memory *mem; struct nvkm_vma *bar1; - - struct mutex mutex; - struct list_head list; } userd; struct { diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h index 8145796ffc61..a2333cfe6955 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h @@ -55,5 +55,4 @@ int gp10b_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct n int gv100_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **); int tu102_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **); int ga102_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **); -int ad102_gr_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_gr **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h index 8d2e170883e1..ca83caa55157 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvdec.h @@ -13,7 +13,5 @@ struct nvkm_nvdec { int gm107_nvdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvdec **); int tu102_nvdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvdec **); -int ga100_nvdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvdec **); int ga102_nvdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvdec **); -int ad102_nvdec_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvdec **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h index 018c58fc32ba..1f6eef13f872 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvenc.h @@ -13,6 +13,4 @@ struct nvkm_nvenc { int gm107_nvenc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvenc **); int tu102_nvenc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvenc **); -int ga102_nvenc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvenc **); -int ad102_nvenc_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_nvenc **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvjpg.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/nvjpg.h deleted file mode 100644 index 80b7933a789e..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/nvjpg.h +++ /dev/null @@ -1,8 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -#ifndef __NVKM_NVJPG_H__ -#define __NVKM_NVJPG_H__ -#include <core/engine.h> - -int ga100_nvjpg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **); -int ad102_nvjpg_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **); -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/ofa.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/ofa.h deleted file mode 100644 index e72e2115333b..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/ofa.h +++ /dev/null @@ -1,9 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -#ifndef __NVKM_OFA_H__ -#define __NVKM_OFA_H__ -#include <core/engine.h> - -int ga100_ofa_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **); -int ga102_ofa_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **); -int ad102_ofa_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_engine **); -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h index 5b798a1a313d..e0d777a933e1 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h @@ -102,6 +102,9 @@ int gv100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct n int tu102_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **); int ga100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **); int ga102_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **); +int gh100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **); +int gb100_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **); +int gb202_fb_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fb **); #include <subdev/bios.h> #include <subdev/bios/ramcfg.h> diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fsp.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fsp.h new file mode 100644 index 000000000000..8a3dbb1cbb46 --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fsp.h @@ -0,0 +1,24 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVKM_FSP_H__ +#define __NVKM_FSP_H__ +#include <core/subdev.h> +#include <core/falcon.h> + +struct nvkm_fsp { + const struct nvkm_fsp_func *func; + struct nvkm_subdev subdev; + + struct nvkm_falcon falcon; +}; + +bool nvkm_fsp_verify_gsp_fmc(struct nvkm_fsp *, u32 hash_size, u32 pkey_size, u32 sig_size); +int nvkm_fsp_boot_gsp_fmc(struct nvkm_fsp *, u64 args_addr, u32 rsvd_size, bool resume, + u64 img_addr, const u8 *hash, const u8 *pkey, const u8 *sig); + +int gh100_fsp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fsp **); +int gb100_fsp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fsp **); +int gb202_fsp_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fsp **); +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h index 1c12854a8550..226c7ec56b8e 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/gsp.h @@ -17,6 +17,9 @@ struct nvkm_gsp_mem { dma_addr_t addr; }; +int nvkm_gsp_mem_ctor(struct nvkm_gsp *, size_t size, struct nvkm_gsp_mem *); +void nvkm_gsp_mem_dtor(struct nvkm_gsp_mem *); + struct nvkm_gsp_radix3 { struct nvkm_gsp_mem lvl0; struct nvkm_gsp_mem lvl1; @@ -65,6 +68,9 @@ struct nvkm_gsp { const struct firmware *load; const struct firmware *unload; } booter; + + const struct firmware *fmc; + const struct firmware *bl; const struct firmware *rm; } fws; @@ -112,6 +118,15 @@ struct nvkm_gsp { struct { struct nvkm_gsp_mem fw; + u8 *hash; + u8 *pkey; + u8 *sig; + + struct nvkm_gsp_mem args; + } fmc; + + struct { + struct nvkm_gsp_mem fw; u32 code_offset; u32 data_offset; u32 manifest_offset; @@ -130,6 +145,7 @@ struct nvkm_gsp { struct sg_table sgt; struct nvkm_gsp_radix3 radix3; struct nvkm_gsp_mem meta; + struct sg_table fbsr; } sr; struct { @@ -209,32 +225,7 @@ struct nvkm_gsp { u8 tpcs; } gr; - const struct nvkm_gsp_rm { - void *(*rpc_get)(struct nvkm_gsp *, u32 fn, u32 argc); - void *(*rpc_push)(struct nvkm_gsp *gsp, void *argv, - enum nvkm_gsp_rpc_reply_policy policy, u32 repc); - void (*rpc_done)(struct nvkm_gsp *gsp, void *repv); - - void *(*rm_ctrl_get)(struct nvkm_gsp_object *, u32 cmd, u32 argc); - int (*rm_ctrl_push)(struct nvkm_gsp_object *, void **argv, u32 repc); - void (*rm_ctrl_done)(struct nvkm_gsp_object *, void *repv); - - void *(*rm_alloc_get)(struct nvkm_gsp_object *, u32 oclass, u32 argc); - void *(*rm_alloc_push)(struct nvkm_gsp_object *, void *argv); - void (*rm_alloc_done)(struct nvkm_gsp_object *, void *repv); - - int (*rm_free)(struct nvkm_gsp_object *); - - int (*client_ctor)(struct nvkm_gsp *, struct nvkm_gsp_client *); - void (*client_dtor)(struct nvkm_gsp_client *); - - int (*device_ctor)(struct nvkm_gsp_client *, struct nvkm_gsp_device *); - void (*device_dtor)(struct nvkm_gsp_device *); - - int (*event_ctor)(struct nvkm_gsp_device *, u32 handle, u32 id, - nvkm_gsp_event_func, struct nvkm_gsp_event *); - void (*event_dtor)(struct nvkm_gsp_event *); - } *rm; + struct nvkm_rm *rm; struct { struct mutex mutex; @@ -272,17 +263,19 @@ nvkm_gsp_rm(struct nvkm_gsp *gsp) return gsp && (gsp->fws.rm || gsp->fw.img); } +#include <rm/rm.h> + static inline void * nvkm_gsp_rpc_get(struct nvkm_gsp *gsp, u32 fn, u32 argc) { - return gsp->rm->rpc_get(gsp, fn, argc); + return gsp->rm->api->rpc->get(gsp, fn, argc); } static inline void * nvkm_gsp_rpc_push(struct nvkm_gsp *gsp, void *argv, enum nvkm_gsp_rpc_reply_policy policy, u32 repc) { - return gsp->rm->rpc_push(gsp, argv, policy, repc); + return gsp->rm->api->rpc->push(gsp, argv, policy, repc); } static inline void * @@ -311,19 +304,19 @@ nvkm_gsp_rpc_wr(struct nvkm_gsp *gsp, void *argv, static inline void nvkm_gsp_rpc_done(struct nvkm_gsp *gsp, void *repv) { - gsp->rm->rpc_done(gsp, repv); + gsp->rm->api->rpc->done(gsp, repv); } static inline void * nvkm_gsp_rm_ctrl_get(struct nvkm_gsp_object *object, u32 cmd, u32 argc) { - return object->client->gsp->rm->rm_ctrl_get(object, cmd, argc); + return object->client->gsp->rm->api->ctrl->get(object, cmd, argc); } static inline int nvkm_gsp_rm_ctrl_push(struct nvkm_gsp_object *object, void *argv, u32 repc) { - return object->client->gsp->rm->rm_ctrl_push(object, argv, repc); + return object->client->gsp->rm->api->ctrl->push(object, argv, repc); } static inline void * @@ -354,7 +347,7 @@ nvkm_gsp_rm_ctrl_wr(struct nvkm_gsp_object *object, void *argv) static inline void nvkm_gsp_rm_ctrl_done(struct nvkm_gsp_object *object, void *repv) { - object->client->gsp->rm->rm_ctrl_done(object, repv); + object->client->gsp->rm->api->ctrl->done(object, repv); } static inline void * @@ -369,7 +362,7 @@ nvkm_gsp_rm_alloc_get(struct nvkm_gsp_object *parent, u32 handle, u32 oclass, u3 object->parent = parent; object->handle = handle; - argv = gsp->rm->rm_alloc_get(object, oclass, argc); + argv = gsp->rm->api->alloc->get(object, oclass, argc); if (IS_ERR_OR_NULL(argv)) { object->client = NULL; return argv; @@ -381,7 +374,7 @@ nvkm_gsp_rm_alloc_get(struct nvkm_gsp_object *parent, u32 handle, u32 oclass, u3 static inline void * nvkm_gsp_rm_alloc_push(struct nvkm_gsp_object *object, void *argv) { - void *repv = object->client->gsp->rm->rm_alloc_push(object, argv); + void *repv = object->client->gsp->rm->api->alloc->push(object, argv); if (IS_ERR(repv)) object->client = NULL; @@ -403,7 +396,7 @@ nvkm_gsp_rm_alloc_wr(struct nvkm_gsp_object *object, void *argv) static inline void nvkm_gsp_rm_alloc_done(struct nvkm_gsp_object *object, void *repv) { - object->client->gsp->rm->rm_alloc_done(object, repv); + object->client->gsp->rm->api->alloc->done(object, repv); } static inline int @@ -421,39 +414,29 @@ nvkm_gsp_rm_alloc(struct nvkm_gsp_object *parent, u32 handle, u32 oclass, u32 ar static inline int nvkm_gsp_rm_free(struct nvkm_gsp_object *object) { - if (object->client) - return object->client->gsp->rm->rm_free(object); + if (object->client) { + int ret = object->client->gsp->rm->api->alloc->free(object); + object->client = NULL; + return ret; + } return 0; } -static inline int -nvkm_gsp_client_ctor(struct nvkm_gsp *gsp, struct nvkm_gsp_client *client) -{ - if (WARN_ON(!gsp->rm)) - return -ENOSYS; - - return gsp->rm->client_ctor(gsp, client); -} - -static inline void -nvkm_gsp_client_dtor(struct nvkm_gsp_client *client) -{ - if (client->gsp) - client->gsp->rm->client_dtor(client); -} +int nvkm_gsp_client_ctor(struct nvkm_gsp *, struct nvkm_gsp_client *); +void nvkm_gsp_client_dtor(struct nvkm_gsp_client *); static inline int nvkm_gsp_device_ctor(struct nvkm_gsp_client *client, struct nvkm_gsp_device *device) { - return client->gsp->rm->device_ctor(client, device); + return client->gsp->rm->api->device->ctor(client, device); } static inline void nvkm_gsp_device_dtor(struct nvkm_gsp_device *device) { if (device->object.client) - device->object.client->gsp->rm->device_dtor(device); + device->object.client->gsp->rm->api->device->dtor(device); } static inline int @@ -485,7 +468,9 @@ static inline int nvkm_gsp_device_event_ctor(struct nvkm_gsp_device *device, u32 handle, u32 id, nvkm_gsp_event_func func, struct nvkm_gsp_event *event) { - return device->object.client->gsp->rm->event_ctor(device, handle, id, func, event); + struct nvkm_rm *rm = device->object.client->gsp->rm; + + return rm->api->device->event.ctor(device, handle, id, func, event); } static inline void @@ -494,7 +479,7 @@ nvkm_gsp_event_dtor(struct nvkm_gsp_event *event) struct nvkm_gsp_device *device = event->device; if (device) - device->object.client->gsp->rm->event_dtor(event); + device->object.client->gsp->rm->api->device->event.dtor(event); } int nvkm_gsp_intr_stall(struct nvkm_gsp *, enum nvkm_subdev_type, int); @@ -505,5 +490,8 @@ int tu102_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_ int tu116_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); int ga100_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); int ga102_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); +int gh100_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); int ad102_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); +int gb100_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); +int gb202_gsp_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h index e10cbd9203ec..db835cf7b8ac 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/instmem.h @@ -24,11 +24,6 @@ struct nvkm_instmem { struct nvkm_ramht *ramht; struct nvkm_memory *ramro; struct nvkm_memory *ramfc; - - struct { - struct sg_table fbsr; - bool fbsr_valid; - } rm; }; u32 nvkm_instmem_rd32(struct nvkm_instmem *, u32 addr); @@ -41,4 +36,5 @@ int nv04_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nv int nv40_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **); int nv50_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **); int gk20a_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **); +int gh100_instmem_new(struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h index 935b1cacd528..abcb0dbcde70 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h @@ -8,7 +8,7 @@ struct nvkm_vma { struct list_head head; struct rb_node tree; u64 addr; - u64 size:50; + u64 size; bool mapref:1; /* PTs (de)referenced on (un)map (vs pre-allocated). */ bool sparse:1; /* Unmapped PDEs/PTEs will not trigger MMU faults. */ #define NVKM_VMA_PAGE_NONE 7 @@ -73,6 +73,7 @@ struct nvkm_vmm { struct nvkm_gsp_object object; struct nvkm_vma *rsvd; + bool external; } rm; }; @@ -165,4 +166,5 @@ int gp100_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct int gp10b_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **); int gv100_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **); int tu102_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **); +int gh100_mmu_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_mmu **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h index 3c103101d5fc..112b674ed9c8 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h @@ -50,6 +50,7 @@ int gf100_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct int gf106_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **); int gk104_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **); int gp100_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **); +int gh100_pci_new(struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_pci **); /* pcie functions */ int nvkm_pcie_set_link(struct nvkm_pci *, enum nvkm_pcie_speed, u8 width); diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0000.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0000.h deleted file mode 100644 index 7a3fc023072d..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0000.h +++ /dev/null @@ -1,38 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl0000_h__ -#define __src_common_sdk_nvidia_inc_class_cl0000_h__ -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvlimits.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2001-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV01_ROOT (0x0U) /* finn: Evaluated from "NV0000_ALLOC_PARAMETERS_MESSAGE_ID" */ - -typedef struct NV0000_ALLOC_PARAMETERS { - NvHandle hClient; /* CORERM-2934: hClient must remain the first member until all allocations use these params */ - NvU32 processID; - char processName[NV_PROC_NAME_MAX_LENGTH]; -} NV0000_ALLOC_PARAMETERS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0005.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0005.h deleted file mode 100644 index e4de36d63666..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0005.h +++ /dev/null @@ -1,38 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl0005_h__ -#define __src_common_sdk_nvidia_inc_class_cl0005_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2001-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NV0005_ALLOC_PARAMETERS { - NvHandle hParentClient; - NvHandle hSrcResource; - - NvV32 hClass; - NvV32 notifyIndex; - NV_DECLARE_ALIGNED(NvP64 data, 8); -} NV0005_ALLOC_PARAMETERS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0080.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0080.h deleted file mode 100644 index 8868118e47d6..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0080.h +++ /dev/null @@ -1,43 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl0080_h__ -#define __src_common_sdk_nvidia_inc_class_cl0080_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2001-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV01_DEVICE_0 (0x80U) /* finn: Evaluated from "NV0080_ALLOC_PARAMETERS_MESSAGE_ID" */ - -typedef struct NV0080_ALLOC_PARAMETERS { - NvU32 deviceId; - NvHandle hClientShare; - NvHandle hTargetClient; - NvHandle hTargetDevice; - NvV32 flags; - NV_DECLARE_ALIGNED(NvU64 vaSpaceSize, 8); - NV_DECLARE_ALIGNED(NvU64 vaStartInternal, 8); - NV_DECLARE_ALIGNED(NvU64 vaLimitInternal, 8); - NvV32 vaMode; -} NV0080_ALLOC_PARAMETERS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080.h deleted file mode 100644 index 9040ea5608a0..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080.h +++ /dev/null @@ -1,35 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl2080_h__ -#define __src_common_sdk_nvidia_inc_class_cl2080_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2002-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV20_SUBDEVICE_0 (0x2080U) /* finn: Evaluated from "NV2080_ALLOC_PARAMETERS_MESSAGE_ID" */ - -typedef struct NV2080_ALLOC_PARAMETERS { - NvU32 subDeviceId; -} NV2080_ALLOC_PARAMETERS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080_notification.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080_notification.h deleted file mode 100644 index ba659d6477d3..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080_notification.h +++ /dev/null @@ -1,62 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl2080_notification_h__ -#define __src_common_sdk_nvidia_inc_class_cl2080_notification_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV2080_NOTIFIERS_HOTPLUG (1) - -#define NV2080_NOTIFIERS_DP_IRQ (7) - -#define NV2080_ENGINE_TYPE_GRAPHICS (0x00000001) -#define NV2080_ENGINE_TYPE_GR0 NV2080_ENGINE_TYPE_GRAPHICS - -#define NV2080_ENGINE_TYPE_COPY0 (0x00000009) - -#define NV2080_ENGINE_TYPE_BSP (0x00000013) -#define NV2080_ENGINE_TYPE_NVDEC0 NV2080_ENGINE_TYPE_BSP - -#define NV2080_ENGINE_TYPE_MSENC (0x0000001b) -#define NV2080_ENGINE_TYPE_NVENC0 NV2080_ENGINE_TYPE_MSENC /* Mutually exclusive alias */ - -#define NV2080_ENGINE_TYPE_SW (0x00000022) - -#define NV2080_ENGINE_TYPE_SEC2 (0x00000026) - -#define NV2080_ENGINE_TYPE_NVJPG (0x0000002b) -#define NV2080_ENGINE_TYPE_NVJPEG0 NV2080_ENGINE_TYPE_NVJPG - -#define NV2080_ENGINE_TYPE_OFA (0x00000033) - -typedef struct { - NvU32 plugDisplayMask; - NvU32 unplugDisplayMask; -} Nv2080HotplugNotification; - -typedef struct Nv2080DpIrqNotificationRec { - NvU32 displayId; -} Nv2080DpIrqNotification; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl84a0.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl84a0.h deleted file mode 100644 index 9eb780a1ac72..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl84a0.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl84a0_h__ -#define __src_common_sdk_nvidia_inc_class_cl84a0_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2001-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV01_MEMORY_LIST_SYSTEM (0x00000081) - -#define NV01_MEMORY_LIST_FBMEM (0x00000082) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl90f1.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl90f1.h deleted file mode 100644 index f1d21776e395..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/cl90f1.h +++ /dev/null @@ -1,31 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_cl90f1_h__ -#define __src_common_sdk_nvidia_inc_class_cl90f1_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2011 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define FERMI_VASPACE_A (0x000090f1) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/clc0b5sw.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/clc0b5sw.h deleted file mode 100644 index b8f32576cfaa..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/class/clc0b5sw.h +++ /dev/null @@ -1,34 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_class_clc0b5sw_h__ -#define __src_common_sdk_nvidia_inc_class_clc0b5sw_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2014-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NVC0B5_ALLOCATION_PARAMETERS { - NvU32 version; - NvU32 engineType; -} NVC0B5_ALLOCATION_PARAMETERS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073common.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073common.h deleted file mode 100644 index 58b3ba7badf1..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073common.h +++ /dev/null @@ -1,39 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073common_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073common_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NV0073_CTRL_CMD_DSC_CAP_PARAMS { - NvBool bDscSupported; - NvU32 encoderColorFormatMask; - NvU32 lineBufferSizeKB; - NvU32 rateBufferSizeKB; - NvU32 bitsPerPixelPrecision; - NvU32 maxNumHztSlices; - NvU32 lineBufferBitDepth; -} NV0073_CTRL_CMD_DSC_CAP_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dfp.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dfp.h deleted file mode 100644 index 596f2ea8344e..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dfp.h +++ /dev/null @@ -1,166 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073dfp_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073dfp_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2005-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV0073_CTRL_CMD_DFP_GET_INFO (0x731140U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_GET_INFO_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_DFP_GET_INFO_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 flags; - NvU32 flags2; -} NV0073_CTRL_DFP_GET_INFO_PARAMS; - -#define NV0073_CTRL_DFP_FLAGS_SIGNAL 2:0 -#define NV0073_CTRL_DFP_FLAGS_SIGNAL_TMDS (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_SIGNAL_LVDS (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_SIGNAL_SDI (0x00000002U) -#define NV0073_CTRL_DFP_FLAGS_SIGNAL_DISPLAYPORT (0x00000003U) -#define NV0073_CTRL_DFP_FLAGS_SIGNAL_DSI (0x00000004U) -#define NV0073_CTRL_DFP_FLAGS_SIGNAL_WRBK (0x00000005U) -#define NV0073_CTRL_DFP_FLAGS_LANE 5:3 -#define NV0073_CTRL_DFP_FLAGS_LANE_NONE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_LANE_SINGLE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_LANE_DUAL (0x00000002U) -#define NV0073_CTRL_DFP_FLAGS_LANE_QUAD (0x00000003U) -#define NV0073_CTRL_DFP_FLAGS_LANE_OCT (0x00000004U) -#define NV0073_CTRL_DFP_FLAGS_LIMIT 6:6 -#define NV0073_CTRL_DFP_FLAGS_LIMIT_DISABLE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_LIMIT_60HZ_RR (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER 7:7 -#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER_NORMAL (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER_DISABLE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE 8:8 -#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE 9:9 -#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE 10:10 -#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE 11:11 -#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE 12:12 -#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED 14:14 -#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT 15:15 -#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT 16:16 -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT_NONE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT_PREFER_RBR (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW 19:17 -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_1_62GBPS (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_2_70GBPS (0x00000002U) -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_5_40GBPS (0x00000003U) -#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_8_10GBPS (0x00000004U) -#define NV0073_CTRL_DFP_FLAGS_LINK 21:20 -#define NV0073_CTRL_DFP_FLAGS_LINK_NONE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_LINK_SINGLE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_LINK_DUAL (0x00000002U) -#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID 22:22 -#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID 24:23 -#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_NONE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_A (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_B (0x00000002U) -#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_GANGED (0x00000003U) -#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED 25:25 -#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_FLAGS_DP_PHY_REPEATER_COUNT 29:26 -#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE 30:30 -#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE_TRUE (0x00000001U) - -#define NV0073_CTRL_CMD_DFP_SET_ELD_AUDIO_CAPS (0x731144U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_SET_ELD_AUDIO_CAP_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_ELD_BUFFER 96U - -typedef struct NV0073_CTRL_DFP_SET_ELD_AUDIO_CAP_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 numELDSize; - NvU8 bufferELD[NV0073_CTRL_DFP_ELD_AUDIO_CAPS_ELD_BUFFER]; - NvU32 maxFreqSupported; - NvU32 ctrl; - NvU32 deviceEntry; -} NV0073_CTRL_DFP_SET_ELD_AUDIO_CAP_PARAMS; - -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_PD 0:0 -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_PD_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_PD_TRUE (0x00000001U) -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_ELDV 1:1 -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_ELDV_FALSE (0x00000000U) -#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_ELDV_TRUE (0x00000001U) - -#define NV0073_CTRL_CMD_DFP_SET_AUDIO_ENABLE (0x731150U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_SET_AUDIO_ENABLE_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_DFP_SET_AUDIO_ENABLE_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvBool enable; -} NV0073_CTRL_DFP_SET_AUDIO_ENABLE_PARAMS; - -typedef NvU32 NV0073_CTRL_DFP_ASSIGN_SOR_LINKCONFIG; - -typedef struct NV0073_CTRL_DFP_ASSIGN_SOR_INFO { - NvU32 displayMask; - NvU32 sorType; -} NV0073_CTRL_DFP_ASSIGN_SOR_INFO; - -#define NV0073_CTRL_CMD_DFP_ASSIGN_SOR (0x731152U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_ASSIGN_SOR_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_CMD_DFP_ASSIGN_SOR_MAX_SORS 4U - -typedef struct NV0073_CTRL_DFP_ASSIGN_SOR_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU8 sorExcludeMask; - NvU32 slaveDisplayId; - NV0073_CTRL_DFP_ASSIGN_SOR_LINKCONFIG forceSublinkConfig; - NvBool bIs2Head1Or; - NvU32 sorAssignList[NV0073_CTRL_CMD_DFP_ASSIGN_SOR_MAX_SORS]; - NV0073_CTRL_DFP_ASSIGN_SOR_INFO sorAssignListWithTag[NV0073_CTRL_CMD_DFP_ASSIGN_SOR_MAX_SORS]; - NvU8 reservedSorMask; - NvU32 flags; -} NV0073_CTRL_DFP_ASSIGN_SOR_PARAMS; - -#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_AUDIO 0:0 -#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_AUDIO_OPTIMAL (0x00000001U) -#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_AUDIO_DEFAULT (0x00000000U) -#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_ACTIVE_SOR_NOT_AUDIO_CAPABLE 1:1 -#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_ACTIVE_SOR_NOT_AUDIO_CAPABLE_NO (0x00000000U) -#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_ACTIVE_SOR_NOT_AUDIO_CAPABLE_YES (0x00000001U) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dp.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dp.h deleted file mode 100644 index bae4b1997736..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dp.h +++ /dev/null @@ -1,335 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073dp_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073dp_h__ -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073common.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2005-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV0073_CTRL_CMD_DP_AUXCH_CTRL (0x731341U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_AUXCH_CTRL_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_DP_AUXCH_MAX_DATA_SIZE 16U - -typedef struct NV0073_CTRL_DP_AUXCH_CTRL_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvBool bAddrOnly; - NvU32 cmd; - NvU32 addr; - NvU8 data[NV0073_CTRL_DP_AUXCH_MAX_DATA_SIZE]; - NvU32 size; - NvU32 replyType; - NvU32 retryTimeMs; -} NV0073_CTRL_DP_AUXCH_CTRL_PARAMS; - -#define NV0073_CTRL_DP_AUXCH_CMD_TYPE 3:3 -#define NV0073_CTRL_DP_AUXCH_CMD_TYPE_I2C (0x00000000U) -#define NV0073_CTRL_DP_AUXCH_CMD_TYPE_AUX (0x00000001U) -#define NV0073_CTRL_DP_AUXCH_CMD_I2C_MOT 2:2 -#define NV0073_CTRL_DP_AUXCH_CMD_I2C_MOT_FALSE (0x00000000U) -#define NV0073_CTRL_DP_AUXCH_CMD_I2C_MOT_TRUE (0x00000001U) -#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE 1:0 -#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE_WRITE (0x00000000U) -#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE_READ (0x00000001U) -#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE_WRITE_STATUS (0x00000002U) - -#define NV0073_CTRL_CMD_DP_CTRL (0x731343U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_CTRL_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_DP_CTRL_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 cmd; - NvU32 data; - NvU32 err; - NvU32 retryTimeMs; - NvU32 eightLaneDpcdBaseAddr; -} NV0073_CTRL_DP_CTRL_PARAMS; - -#define NV0073_CTRL_DP_CMD_SET_LANE_COUNT 0:0 -#define NV0073_CTRL_DP_CMD_SET_LANE_COUNT_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_SET_LANE_COUNT_TRUE (0x00000001U) -#define NV0073_CTRL_DP_CMD_SET_LINK_BW 1:1 -#define NV0073_CTRL_DP_CMD_SET_LINK_BW_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_SET_LINK_BW_TRUE (0x00000001U) -#define NV0073_CTRL_DP_CMD_DISABLE_DOWNSPREAD 2:2 -#define NV0073_CTRL_DP_CMD_DISABLE_DOWNSPREAD_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_DISABLE_DOWNSPREAD_TRUE (0x00000001U) -#define NV0073_CTRL_DP_CMD_UNUSED 3:3 -#define NV0073_CTRL_DP_CMD_SET_FORMAT_MODE 4:4 -#define NV0073_CTRL_DP_CMD_SET_FORMAT_MODE_SINGLE_STREAM (0x00000000U) -#define NV0073_CTRL_DP_CMD_SET_FORMAT_MODE_MULTI_STREAM (0x00000001U) -#define NV0073_CTRL_DP_CMD_FAST_LINK_TRAINING 5:5 -#define NV0073_CTRL_DP_CMD_FAST_LINK_TRAINING_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_FAST_LINK_TRAINING_YES (0x00000001U) -#define NV0073_CTRL_DP_CMD_NO_LINK_TRAINING 6:6 -#define NV0073_CTRL_DP_CMD_NO_LINK_TRAINING_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_NO_LINK_TRAINING_YES (0x00000001U) -#define NV0073_CTRL_DP_CMD_SET_ENHANCED_FRAMING 7:7 -#define NV0073_CTRL_DP_CMD_SET_ENHANCED_FRAMING_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_SET_ENHANCED_FRAMING_TRUE (0x00000001U) -#define NV0073_CTRL_DP_CMD_USE_DOWNSPREAD_SETTING 8:8 -#define NV0073_CTRL_DP_CMD_USE_DOWNSPREAD_SETTING_DEFAULT (0x00000000U) -#define NV0073_CTRL_DP_CMD_USE_DOWNSPREAD_SETTING_FORCE (0x00000001U) -#define NV0073_CTRL_DP_CMD_SKIP_HW_PROGRAMMING 9:9 -#define NV0073_CTRL_DP_CMD_SKIP_HW_PROGRAMMING_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_SKIP_HW_PROGRAMMING_YES (0x00000001U) -#define NV0073_CTRL_DP_CMD_POST_LT_ADJ_REQ_GRANTED 10:10 -#define NV0073_CTRL_DP_CMD_POST_LT_ADJ_REQ_GRANTED_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_POST_LT_ADJ_REQ_GRANTED_YES (0x00000001U) -#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING 12:11 -#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING_DONOT_TOGGLE_TRANSMISSION (0x00000001U) -#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING_TOGGLE_TRANSMISSION_ON (0x00000002U) -#define NV0073_CTRL_DP_CMD_TRAIN_PHY_REPEATER 13:13 -#define NV0073_CTRL_DP_CMD_TRAIN_PHY_REPEATER_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_TRAIN_PHY_REPEATER_YES (0x00000001U) -#define NV0073_CTRL_DP_CMD_FALLBACK_CONFIG 14:14 -#define NV0073_CTRL_DP_CMD_FALLBACK_CONFIG_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_FALLBACK_CONFIG_TRUE (0x00000001U) -#define NV0073_CTRL_DP_CMD_ENABLE_FEC 15:15 -#define NV0073_CTRL_DP_CMD_ENABLE_FEC_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_ENABLE_FEC_TRUE (0x00000001U) - -#define NV0073_CTRL_DP_CMD_BANDWIDTH_TEST 29:29 -#define NV0073_CTRL_DP_CMD_BANDWIDTH_TEST_NO (0x00000000U) -#define NV0073_CTRL_DP_CMD_BANDWIDTH_TEST_YES (0x00000001U) -#define NV0073_CTRL_DP_CMD_LINK_CONFIG_CHECK_DISABLE 30:30 -#define NV0073_CTRL_DP_CMD_LINK_CONFIG_CHECK_DISABLE_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_LINK_CONFIG_CHECK_DISABLE_TRUE (0x00000001U) -#define NV0073_CTRL_DP_CMD_DISABLE_LINK_CONFIG 31:31 -#define NV0073_CTRL_DP_CMD_DISABLE_LINK_CONFIG_FALSE (0x00000000U) -#define NV0073_CTRL_DP_CMD_DISABLE_LINK_CONFIG_TRUE (0x00000001U) - -#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT 4:0 -#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_0 (0x00000000U) -#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_1 (0x00000001U) -#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_2 (0x00000002U) -#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_4 (0x00000004U) -#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_8 (0x00000008U) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW 15:8 -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_1_62GBPS (0x00000006U) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_2_16GBPS (0x00000008U) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_2_43GBPS (0x00000009U) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_2_70GBPS (0x0000000AU) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_3_24GBPS (0x0000000CU) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_4_32GBPS (0x00000010U) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_5_40GBPS (0x00000014U) -#define NV0073_CTRL_DP_DATA_SET_LINK_BW_8_10GBPS (0x0000001EU) -#define NV0073_CTRL_DP_DATA_SET_ENHANCED_FRAMING 18:18 -#define NV0073_CTRL_DP_DATA_SET_ENHANCED_FRAMING_NO (0x00000000U) -#define NV0073_CTRL_DP_DATA_SET_ENHANCED_FRAMING_YES (0x00000001U) -#define NV0073_CTRL_DP_DATA_TARGET 22:19 -#define NV0073_CTRL_DP_DATA_TARGET_SINK (0x00000000U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_0 (0x00000001U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_1 (0x00000002U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_2 (0x00000003U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_3 (0x00000004U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_4 (0x00000005U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_5 (0x00000006U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_6 (0x00000007U) -#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_7 (0x00000008U) - -#define NV0073_CTRL_MAX_LANES 8U - -typedef struct NV0073_CTRL_DP_LANE_DATA_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 numLanes; - NvU32 data[NV0073_CTRL_MAX_LANES]; -} NV0073_CTRL_DP_LANE_DATA_PARAMS; - -#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS 1:0 -#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_NONE (0x00000000U) -#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_LEVEL1 (0x00000001U) -#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_LEVEL2 (0x00000002U) -#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_LEVEL3 (0x00000003U) -#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT 3:2 -#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL0 (0x00000000U) -#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL1 (0x00000001U) -#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL2 (0x00000002U) -#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL3 (0x00000003U) - -#define NV0073_CTRL_CMD_DP_SET_LANE_DATA (0x731346U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_SET_LANE_DATA_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_CMD_DP_SET_AUDIO_MUTESTREAM (0x731359U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 mute; -} NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS; - -#define NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID (0x73135bU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 preferredDisplayId; - - NvBool force; - NvBool useBFM; - - NvU32 displayIdAssigned; - NvU32 allDisplayMask; -} NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID_PARAMS; - -#define NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID (0x73135cU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; -} NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID_PARAMS; - -#define NV0073_CTRL_CMD_DP_CONFIG_STREAM (0x731362U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS { - NvU32 subDeviceInstance; - NvU32 head; - NvU32 sorIndex; - NvU32 dpLink; - - NvBool bEnableOverride; - NvBool bMST; - NvU32 singleHeadMultistreamMode; - NvU32 hBlankSym; - NvU32 vBlankSym; - NvU32 colorFormat; - NvBool bEnableTwoHeadOneOr; - - struct { - NvU32 slotStart; - NvU32 slotEnd; - NvU32 PBN; - NvU32 Timeslice; - NvBool sendACT; // deprecated -Use NV0073_CTRL_CMD_DP_SEND_ACT - NvU32 singleHeadMSTPipeline; - NvBool bEnableAudioOverRightPanel; - } MST; - - struct { - NvBool bEnhancedFraming; - NvU32 tuSize; - NvU32 waterMark; - NvU32 actualPclkHz; // deprecated -Use MvidWarParams - NvU32 linkClkFreqHz; // deprecated -Use MvidWarParams - NvBool bEnableAudioOverRightPanel; - struct { - NvU32 activeCnt; - NvU32 activeFrac; - NvU32 activePolarity; - NvBool mvidWarEnabled; - struct { - NvU32 actualPclkHz; - NvU32 linkClkFreqHz; - } MvidWarParams; - } Legacy; - } SST; -} NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS; - -#define NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT (0x731365U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT_PARAMS { - NvU32 subDeviceInstance; -} NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT_PARAMS; - -#define NV0073_CTRL_CMD_DP_GET_CAPS (0x731369U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID (0x69U) - -typedef struct NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS { - NvU32 subDeviceInstance; - NvU32 sorIndex; - NvU32 maxLinkRate; - NvU32 dpVersionsSupported; - NvU32 UHBRSupported; - NvBool bIsMultistreamSupported; - NvBool bIsSCEnabled; - NvBool bHasIncreasedWatermarkLimits; - NvBool bIsPC2Disabled; - NvBool isSingleHeadMSTSupported; - NvBool bFECSupported; - NvBool bIsTrainPhyRepeater; - NvBool bOverrideLinkBw; - NV0073_CTRL_CMD_DSC_CAP_PARAMS DSC; -} NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS; - -#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2 0:0 -#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2_NO (0x00000000U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2_YES (0x00000001U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4 1:1 -#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4_NO (0x00000000U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4_YES (0x00000001U) - -#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE 2:0 -#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_NONE (0x00000000U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_1_62 (0x00000001U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_2_70 (0x00000002U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_5_40 (0x00000003U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_8_10 (0x00000004U) - -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_RGB (0x00000001U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_444 (0x00000002U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_NATIVE_422 (0x00000004U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_NATIVE_420 (0x00000008U) - -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_16 (0x00000001U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_8 (0x00000002U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_4 (0x00000003U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_2 (0x00000004U) -#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1 (0x00000005U) - -#define NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES (0x731377U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES 8U - -typedef struct NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS { - // In - NvU32 subDeviceInstance; - NvU32 displayId; - NvU16 linkRateTbl[NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES]; - - // Out - NvU8 linkBwTbl[NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES]; - NvU8 linkBwCount; -} NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS; - -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE 3:0 -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_MONITOR_ENABLE_BEGIN (0x00000000U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_MONITOR_ENABLE_CHALLENGE (0x00000001U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_MONITOR_ENABLE_CHECK (0x00000002U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_DRIVER_ENABLE_BEGIN (0x00000003U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_DRIVER_ENABLE_CHALLENGE (0x00000004U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_DRIVER_ENABLE_CHECK (0x00000005U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_RESET_MONITOR (0x00000006U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_INIT_PUBLIC_INFO (0x00000007U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_GET_PUBLIC_INFO (0x00000008U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_STATUS_CHECK (0x00000009U) - -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_OK (0x00000000U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_PENDING (0x80000001U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_READ_ERROR (0x80000002U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_WRITE_ERROR (0x80000003U) -#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_DEVICE_ERROR (0x80000004U) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073specific.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073specific.h deleted file mode 100644 index 954958dcf834..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073specific.h +++ /dev/null @@ -1,216 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073specific_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073specific_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 1993-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV0073_CTRL_CMD_SPECIFIC_GET_EDID_V2 (0x730245U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_SPECIFIC_GET_EDID_MAX_EDID_BYTES 2048U - -typedef struct NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 bufferSize; - NvU32 flags; - NvU8 edidBuffer[NV0073_CTRL_SPECIFIC_GET_EDID_MAX_EDID_BYTES]; -} NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_GET_CONNECTOR_DATA (0x730250U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_MAX_CONNECTORS 4U - -typedef struct NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 flags; - NvU32 DDCPartners; - NvU32 count; - struct { - NvU32 index; - NvU32 type; - NvU32 location; - } data[NV0073_CTRL_MAX_CONNECTORS]; - NvU32 platform; -} NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_ENABLE (0x730273U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_HDMI_ENABLE_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SPECIFIC_SET_HDMI_ENABLE_PARAMS { - NvU8 subDeviceInstance; - NvU32 displayId; - NvU8 enable; -} NV0073_CTRL_SPECIFIC_SET_HDMI_ENABLE_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM (0x730275U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM_PARAMS { - NvU8 subDeviceInstance; - NvU32 displayId; - NvU8 mute; -} NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_GET_ALL_HEAD_MASK (0x730287U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_ALL_HEAD_MASK_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SPECIFIC_GET_ALL_HEAD_MASK_PARAMS { - NvU32 subDeviceInstance; - NvU32 headMask; -} NV0073_CTRL_SPECIFIC_GET_ALL_HEAD_MASK_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_SET_OD_PACKET (0x730288U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_OD_PACKET_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_SET_OD_MAX_PACKET_SIZE 36U - -typedef struct NV0073_CTRL_SPECIFIC_SET_OD_PACKET_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 transmitControl; - NvU32 packetSize; - NvU32 targetHead; - NvBool bUsePsrHeadforSdp; - NvU8 aPacket[NV0073_CTRL_SET_OD_MAX_PACKET_SIZE]; -} NV0073_CTRL_SPECIFIC_SET_OD_PACKET_PARAMS; - -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ENABLE 0:0 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ENABLE_NO (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ENABLE_YES (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_OTHER_FRAME 1:1 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_OTHER_FRAME_DISABLE (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_OTHER_FRAME_ENABLE (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SINGLE_FRAME 2:2 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SINGLE_FRAME_DISABLE (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SINGLE_FRAME_ENABLE (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ON_HBLANK 3:3 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ON_HBLANK_DISABLE (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ON_HBLANK_ENABLE (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_IMMEDIATE 4:4 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_IMMEDIATE_DISABLE (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_IMMEDIATE_ENABLE (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_VIDEO_FMT 5:5 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_VIDEO_FMT_SW_CONTROLLED (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_VIDEO_FMT_HW_CONTROLLED (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_STEREO_POLARITY 6:6 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_STEREO_POLARITY_FALSE (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_STEREO_POLARITY_TRUE (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_SELF_REFRESH_SETTING 7:7 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_SELF_REFRESH_SETTING_FALSE (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_SELF_REFRESH_SETTING_TRUE (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_GEN_INFOFRAME_MODE 9:8 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_GEN_INFOFRAME_MODE_INFOFRAME0 (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_GEN_INFOFRAME_MODE_INFOFRAME1 (0x0000001U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_RESERVED_LEGACY_MODE 31:31 -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_RESERVED_LEGACY_MODE_NO (0x0000000U) -#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_RESERVED_LEGACY_MODE_YES (0x0000001U) - -#define NV0073_CTRL_CMD_SPECIFIC_OR_GET_INFO (0x73028bU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 index; - NvU32 type; - NvU32 protocol; - NvU32 ditherType; - NvU32 ditherAlgo; - NvU32 location; - NvU32 rootPortId; - NvU32 dcbIndex; - NV_DECLARE_ALIGNED(NvU64 vbiosAddress, 8); - NvBool bIsLitByVbios; - NvBool bIsDispDynamic; -} NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS; - -#define NV0073_CTRL_SPECIFIC_OR_TYPE_NONE (0x00000000U) -#define NV0073_CTRL_SPECIFIC_OR_TYPE_DAC (0x00000001U) -#define NV0073_CTRL_SPECIFIC_OR_TYPE_SOR (0x00000002U) -#define NV0073_CTRL_SPECIFIC_OR_TYPE_PIOR (0x00000003U) - -#define NV0073_CTRL_SPECIFIC_OR_TYPE_DSI (0x00000005U) - -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_DAC_RGB_CRT (0x00000000U) - -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_LVDS_CUSTOM (0x00000000U) -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_SINGLE_TMDS_A (0x00000001U) -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_SINGLE_TMDS_B (0x00000002U) -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DUAL_TMDS (0x00000005U) -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DP_A (0x00000008U) -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DP_B (0x00000009U) -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DSI (0x00000010U) - -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_DSI (0x00000011U) - -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_PIOR_EXT_TMDS_ENC (0x00000000U) - -#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_UNKNOWN (0xFFFFFFFFU) - -#define NV0073_CTRL_CMD_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS (0x730291U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 brightness; - NvBool bUncalibrated; -} NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS (0x730292U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS_PARAMS_MESSAGE_ID" */ - -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS (0x730293U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayId; - NvU32 caps; -} NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS; - -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_GT_340MHZ_CLOCK_SUPPORTED 0:0 -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_GT_340MHZ_CLOCK_SUPPORTED_FALSE (0x00000000U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_GT_340MHZ_CLOCK_SUPPORTED_TRUE (0x00000001U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_LTE_340MHZ_SCRAMBLING_SUPPORTED 1:1 -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_LTE_340MHZ_SCRAMBLING_SUPPORTED_FALSE (0x00000000U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_LTE_340MHZ_SCRAMBLING_SUPPORTED_TRUE (0x00000001U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_SCDC_SUPPORTED 2:2 -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_SCDC_SUPPORTED_FALSE (0x00000000U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_SCDC_SUPPORTED_TRUE (0x00000001U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED 5:3 -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_NONE (0x00000000U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_3LANES_3G (0x00000001U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_3LANES_6G (0x00000002U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_6G (0x00000003U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_8G (0x00000004U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_10G (0x00000005U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_12G (0x00000006U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_12_SUPPORTED 6:6 -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_12_SUPPORTED_FALSE (0x00000000U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_12_SUPPORTED_TRUE (0x00000001U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED 9:7 -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_NONE (0x00000000U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_3LANES_3G (0x00000001U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_3LANES_6G (0x00000002U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_6G (0x00000003U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_8G (0x00000004U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_10G (0x00000005U) -#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_12G (0x00000006U) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073system.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073system.h deleted file mode 100644 index d69cef3c01fd..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073system.h +++ /dev/null @@ -1,65 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073system_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0073_ctrl0073system_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2005-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV0073_CTRL_CMD_SYSTEM_GET_NUM_HEADS (0x730102U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_NUM_HEADS_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SYSTEM_GET_NUM_HEADS_PARAMS { - NvU32 subDeviceInstance; - NvU32 flags; - NvU32 numHeads; -} NV0073_CTRL_SYSTEM_GET_NUM_HEADS_PARAMS; - -#define NV0073_CTRL_CMD_SYSTEM_GET_SUPPORTED (0x730120U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS { - NvU32 subDeviceInstance; - NvU32 displayMask; - NvU32 displayMaskDDC; -} NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS; - -#define NV0073_CTRL_CMD_SYSTEM_GET_CONNECT_STATE (0x730122U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS { - NvU32 subDeviceInstance; - NvU32 flags; - NvU32 displayMask; - NvU32 retryTimeMs; -} NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS; - -#define NV0073_CTRL_CMD_SYSTEM_GET_ACTIVE (0x730126U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS_MESSAGE_ID" */ - -typedef struct NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS { - NvU32 subDeviceInstance; - NvU32 head; - NvU32 flags; - NvU32 displayId; -} NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS; - -#define NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS (16U) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gpu.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gpu.h deleted file mode 100644 index 3db099e62364..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gpu.h +++ /dev/null @@ -1,48 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0080_ctrl0080gpu_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0080_ctrl0080gpu_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2004-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS { - NvU32 totalVFs; - NvU32 firstVfOffset; - NvU32 vfFeatureMask; - NV_DECLARE_ALIGNED(NvU64 FirstVFBar0Address, 8); - NV_DECLARE_ALIGNED(NvU64 FirstVFBar1Address, 8); - NV_DECLARE_ALIGNED(NvU64 FirstVFBar2Address, 8); - NV_DECLARE_ALIGNED(NvU64 bar0Size, 8); - NV_DECLARE_ALIGNED(NvU64 bar1Size, 8); - NV_DECLARE_ALIGNED(NvU64 bar2Size, 8); - NvBool b64bitBar0; - NvBool b64bitBar1; - NvBool b64bitBar2; - NvBool bSriovEnabled; - NvBool bSriovHeavyEnabled; - NvBool bEmulateVFBar0TlbInvalidationRegister; - NvBool bClientRmAllocatedCtxBuffer; -} NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gr.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gr.h deleted file mode 100644 index ed01df925573..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gr.h +++ /dev/null @@ -1,31 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0080_ctrl0080gr_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0080_ctrl0080gr_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2004-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV0080_CTRL_GR_CAPS_TBL_SIZE 23 - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080bios.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080bios.h deleted file mode 100644 index b5b7631de99b..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080bios.h +++ /dev/null @@ -1,40 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080bios_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080bios_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2005-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS { - NvU32 BoardID; - char chipSKU[4]; - char chipSKUMod[2]; - char project[5]; - char projectSKU[5]; - char CDP[6]; - char projectSKUMod[2]; - NvU32 businessCycle; -} NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080ce.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080ce.h deleted file mode 100644 index fe912d2bd183..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080ce.h +++ /dev/null @@ -1,35 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080ce_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080ce_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2014-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NV2080_CTRL_CE_GET_FAULT_METHOD_BUFFER_SIZE_PARAMS { - NvU32 size; -} NV2080_CTRL_CE_GET_FAULT_METHOD_BUFFER_SIZE_PARAMS; - -#define NV2080_CTRL_CMD_CE_GET_FAULT_METHOD_BUFFER_SIZE (0x20802a08) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_CE_INTERFACE_ID << 8) | NV2080_CTRL_CE_GET_FAULT_METHOD_BUFFER_SIZE_PARAMS_MESSAGE_ID" */ - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080event.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080event.h deleted file mode 100644 index 87bc4ff92ce1..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080event.h +++ /dev/null @@ -1,41 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080event_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080event_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2006-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV2080_CTRL_CMD_EVENT_SET_NOTIFICATION (0x20800301) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_EVENT_INTERFACE_ID << 8) | NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS { - NvU32 event; - NvU32 action; - NvBool bNotifyState; - NvU32 info32; - NvU16 info16; -} NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS; - -#define NV2080_CTRL_EVENT_SET_NOTIFICATION_ACTION_REPEAT (0x00000002) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fb.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fb.h deleted file mode 100644 index 68c81f9f803c..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fb.h +++ /dev/null @@ -1,51 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080fb_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080fb_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2006-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MEM_TYPES 17U - -typedef NvBool NV2080_CTRL_CMD_FB_GET_FB_REGION_SURFACE_MEM_TYPE_FLAG[NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MEM_TYPES]; - -typedef struct NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO { - NV_DECLARE_ALIGNED(NvU64 base, 8); - NV_DECLARE_ALIGNED(NvU64 limit, 8); - NV_DECLARE_ALIGNED(NvU64 reserved, 8); - NvU32 performance; - NvBool supportCompressed; - NvBool supportISO; - NvBool bProtected; - NV2080_CTRL_CMD_FB_GET_FB_REGION_SURFACE_MEM_TYPE_FLAG blackList; -} NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO; - -#define NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MAX_ENTRIES 16U - -typedef struct NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS { - NvU32 numFBRegions; - NV_DECLARE_ALIGNED(NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO fbRegion[NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MAX_ENTRIES], 8); -} NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fifo.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fifo.h deleted file mode 100644 index bc0f63699b06..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fifo.h +++ /dev/null @@ -1,52 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080fifo_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080fifo_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2006-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV2080_CTRL_CMD_FIFO_GET_DEVICE_INFO_TABLE (0x20801112) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_FIFO_INTERFACE_ID << 8) | NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS_MESSAGE_ID" */ - -#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_MAX_ENTRIES 32 -#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_DATA_TYPES 16 -#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_PBDMA 2 -#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_NAME_LEN 16 - -typedef struct NV2080_CTRL_FIFO_DEVICE_ENTRY { - NvU32 engineData[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_DATA_TYPES]; - NvU32 pbdmaIds[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_PBDMA]; - NvU32 pbdmaFaultIds[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_PBDMA]; - NvU32 numPbdmas; - char engineName[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_NAME_LEN]; -} NV2080_CTRL_FIFO_DEVICE_ENTRY; - -typedef struct NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS { - NvU32 baseIndex; - NvU32 numEntries; - NvBool bMore; - // C form: NV2080_CTRL_FIFO_DEVICE_ENTRY entries[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_MAX_ENTRIES]; - NV2080_CTRL_FIFO_DEVICE_ENTRY entries[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_MAX_ENTRIES]; -} NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h deleted file mode 100644 index 29d7a1052142..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h +++ /dev/null @@ -1,100 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080gpu_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080gpu_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2006-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV2080_GPU_MAX_NAME_STRING_LENGTH (0x0000040U) - -#define NV2080_CTRL_GPU_SET_POWER_STATE_GPU_LEVEL_0 (0x00000000U) - -#define NV2080_CTRL_GPU_SET_POWER_STATE_GPU_LEVEL_3 (0x00000003U) - -typedef struct NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY { - NV_DECLARE_ALIGNED(NvU64 gpuPhysAddr, 8); - NV_DECLARE_ALIGNED(NvU64 gpuVirtAddr, 8); - NV_DECLARE_ALIGNED(NvU64 size, 8); - NvU32 physAttr; - NvU16 bufferId; - NvU8 bInitialize; - NvU8 bNonmapped; -} NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY; - -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_MAIN 0U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PM 1U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PATCH 2U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_BUFFER_BUNDLE_CB 3U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PAGEPOOL 4U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_ATTRIBUTE_CB 5U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_RTV_CB_GLOBAL 6U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_GFXP_POOL 7U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_GFXP_CTRL_BLK 8U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_FECS_EVENT 9U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PRIV_ACCESS_MAP 10U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_UNRESTRICTED_PRIV_ACCESS_MAP 11U -#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_GLOBAL_PRIV_ACCESS_MAP 12U - -#define NV2080_CTRL_GPU_PROMOTE_CONTEXT_MAX_ENTRIES 16U - -#define NV2080_CTRL_CMD_GPU_PROMOTE_CTX (0x2080012bU) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_GPU_INTERFACE_ID << 8) | NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS { - NvU32 engineType; - NvHandle hClient; - NvU32 ChID; - NvHandle hChanClient; - NvHandle hObject; - NvHandle hVirtMemory; - NV_DECLARE_ALIGNED(NvU64 virtAddress, 8); - NV_DECLARE_ALIGNED(NvU64 size, 8); - NvU32 entryCount; - // C form: NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY promoteEntry[NV2080_CTRL_GPU_PROMOTE_CONTEXT_MAX_ENTRIES]; - NV_DECLARE_ALIGNED(NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY promoteEntry[NV2080_CTRL_GPU_PROMOTE_CONTEXT_MAX_ENTRIES], 8); -} NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS; - -typedef struct NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS { - NvU32 gpcMask; -} NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS; - -typedef struct NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS { - NvU32 gpcId; - NvU32 tpcMask; -} NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS; - -typedef struct NV2080_CTRL_GPU_GET_FERMI_ZCULL_INFO_PARAMS { - NvU32 gpcId; - NvU32 zcullMask; -} NV2080_CTRL_GPU_GET_FERMI_ZCULL_INFO_PARAMS; - -#define NV2080_GPU_MAX_GID_LENGTH (0x000000100ULL) - -typedef struct NV2080_CTRL_GPU_GET_GID_INFO_PARAMS { - NvU32 index; - NvU32 flags; - NvU32 length; - NvU8 data[NV2080_GPU_MAX_GID_LENGTH]; -} NV2080_CTRL_GPU_GET_GID_INFO_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gr.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gr.h deleted file mode 100644 index 59f8895bc5d7..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gr.h +++ /dev/null @@ -1,41 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080gr_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080gr_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2006-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef enum NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS { - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_MAIN = 0, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_SPILL = 1, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_PAGEPOOL = 2, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_BETACB = 3, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_RTV = 4, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL = 5, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL_CONTROL = 6, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL_CONTROL_CPU = 7, - NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_END = 8, -} NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h deleted file mode 100644 index e11b2dbe5288..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h +++ /dev/null @@ -1,162 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080internal_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl2080_ctrl2080internal_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2020-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_GET_STATIC_INFO (0x20800a01) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS { - NvU32 feHwSysCap; - NvU32 windowPresentMask; - NvBool bFbRemapperEnabled; - NvU32 numHeads; - NvBool bPrimaryVga; - NvU32 i2cPort; - NvU32 internalDispActiveMask; -} NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS; - -#define NV2080_CTRL_INTERNAL_GR_MAX_ENGINES 8 - -#define NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT 0x19 - -typedef struct NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO { - NvU32 size; - NvU32 alignment; -} NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO; - -typedef struct NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO { - NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO engine[NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT]; -} NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO; - -typedef struct NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS { - NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO engineContextBuffersInfo[NV2080_CTRL_INTERNAL_GR_MAX_ENGINES]; -} NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO (0x20800a32) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_CONSTRUCTED_FALCON_INFO { - NvU32 engDesc; - NvU32 ctxAttr; - NvU32 ctxBufferSize; - NvU32 addrSpaceList; - NvU32 registerBase; -} NV2080_CTRL_INTERNAL_CONSTRUCTED_FALCON_INFO; -#define NV2080_CTRL_CMD_INTERNAL_MAX_CONSTRUCTED_FALCONS 0x40 - -#define NV2080_CTRL_CMD_INTERNAL_GET_CONSTRUCTED_FALCON_INFO (0x20800a42) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_GET_CONSTRUCTED_FALCON_INFO_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_GET_CONSTRUCTED_FALCON_INFO_PARAMS { - NvU32 numConstructedFalcons; - NV2080_CTRL_INTERNAL_CONSTRUCTED_FALCON_INFO constructedFalconsTable[NV2080_CTRL_CMD_INTERNAL_MAX_CONSTRUCTED_FALCONS]; -} NV2080_CTRL_INTERNAL_GET_CONSTRUCTED_FALCON_INFO_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_WRITE_INST_MEM (0x20800a49) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS { - NV_DECLARE_ALIGNED(NvU64 instMemPhysAddr, 8); - NV_DECLARE_ALIGNED(NvU64 instMemSize, 8); - NvU32 instMemAddrSpace; - NvU32 instMemCpuCacheAttr; -} NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER (0x20800a58) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS { - NvU32 addressSpace; - NV_DECLARE_ALIGNED(NvU64 physicalAddr, 8); - NV_DECLARE_ALIGNED(NvU64 limit, 8); - NvU32 cacheSnoop; - NvU32 hclass; - NvU32 channelInstance; - NvBool valid; -} NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_INTR_GET_KERNEL_TABLE (0x20800a5c) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS_MESSAGE_ID" */ - -#define NV2080_CTRL_INTERNAL_INTR_MAX_TABLE_SIZE 128 - -typedef enum NV2080_INTR_CATEGORY { - NV2080_INTR_CATEGORY_DEFAULT = 0, - NV2080_INTR_CATEGORY_ESCHED_DRIVEN_ENGINE = 1, - NV2080_INTR_CATEGORY_ESCHED_DRIVEN_ENGINE_NOTIFICATION = 2, - NV2080_INTR_CATEGORY_RUNLIST = 3, - NV2080_INTR_CATEGORY_RUNLIST_NOTIFICATION = 4, - NV2080_INTR_CATEGORY_UVM_OWNED = 5, - NV2080_INTR_CATEGORY_UVM_SHARED = 6, - NV2080_INTR_CATEGORY_ENUM_COUNT = 7, -} NV2080_INTR_CATEGORY; - -typedef struct NV2080_INTR_CATEGORY_SUBTREE_MAP { - NvU8 subtreeStart; - NvU8 subtreeEnd; -} NV2080_INTR_CATEGORY_SUBTREE_MAP; - -typedef struct NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_ENTRY { - NvU16 engineIdx; - NvU32 pmcIntrMask; - NvU32 vectorStall; - NvU32 vectorNonStall; -} NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_ENTRY; - -typedef struct NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS { - NvU32 tableLen; - NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_ENTRY table[NV2080_CTRL_INTERNAL_INTR_MAX_TABLE_SIZE]; - NV2080_INTR_CATEGORY_SUBTREE_MAP subtreeMap[NV2080_INTR_CATEGORY_ENUM_COUNT]; -} NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_FBSR_INIT (0x20800ac2) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS { - NvU32 fbsrType; - NvU32 numRegions; - NvHandle hClient; - NvHandle hSysMem; - NV_DECLARE_ALIGNED(NvU64 gspFbAllocsSysOffset, 8); - NvBool bEnteringGcoffState; -} NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_FBSR_SEND_REGION_INFO (0x20800ac3) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_FBSR_SEND_REGION_INFO_PARAMS_MESSAGE_ID" */ - -typedef struct NV2080_CTRL_INTERNAL_FBSR_SEND_REGION_INFO_PARAMS { - NvU32 fbsrType; - NvHandle hClient; - NvHandle hVidMem; - NV_DECLARE_ALIGNED(NvU64 vidOffset, 8); - NV_DECLARE_ALIGNED(NvU64 sysOffset, 8); - NV_DECLARE_ALIGNED(NvU64 size, 8); -} NV2080_CTRL_INTERNAL_FBSR_SEND_REGION_INFO_PARAMS; - -#define NV2080_CTRL_CMD_INTERNAL_INIT_BRIGHTC_STATE_LOAD (0x20800ac6) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_INIT_BRIGHTC_STATE_LOAD_PARAMS_MESSAGE_ID" */ - -#define NV2080_CTRL_ACPI_DSM_READ_SIZE (0x1000) /* finn: Evaluated from "(4 * 1024)" */ - -typedef struct NV2080_CTRL_INTERNAL_INIT_BRIGHTC_STATE_LOAD_PARAMS { - NvU32 status; - NvU16 backLightDataSize; - NvU8 backLightData[NV2080_CTRL_ACPI_DSM_READ_SIZE]; -} NV2080_CTRL_INTERNAL_INIT_BRIGHTC_STATE_LOAD_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl90f1.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl90f1.h deleted file mode 100644 index 977e59818533..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl90f1.h +++ /dev/null @@ -1,95 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl90f1_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl90f1_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2014-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define GMMU_FMT_MAX_LEVELS 6U - -#define NV90F1_CTRL_CMD_VASPACE_COPY_SERVER_RESERVED_PDES (0x90f10106U) /* finn: Evaluated from "(FINN_FERMI_VASPACE_A_VASPACE_INTERFACE_ID << 8) | NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS_MESSAGE_ID" */ - -typedef struct NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS { - /*! - * [in] GPU sub-device handle - this API only supports unicast. - * Pass 0 to use subDeviceId instead. - */ - NvHandle hSubDevice; - - /*! - * [in] GPU sub-device ID. Ignored if hSubDevice is non-zero. - */ - NvU32 subDeviceId; - - /*! - * [in] Page size (VA coverage) of the level to reserve. - * This need not be a leaf (page table) page size - it can be - * the coverage of an arbitrary level (including root page directory). - */ - NV_DECLARE_ALIGNED(NvU64 pageSize, 8); - - /*! - * [in] First GPU virtual address of the range to reserve. - * This must be aligned to pageSize. - */ - NV_DECLARE_ALIGNED(NvU64 virtAddrLo, 8); - - /*! - * [in] Last GPU virtual address of the range to reserve. - * This (+1) must be aligned to pageSize. - */ - NV_DECLARE_ALIGNED(NvU64 virtAddrHi, 8); - - /*! - * [in] Number of PDE levels to copy. - */ - NvU32 numLevelsToCopy; - - /*! - * [in] Per-level information. - */ - struct { - /*! - * Physical address of this page level instance. - */ - NV_DECLARE_ALIGNED(NvU64 physAddress, 8); - - /*! - * Size in bytes allocated for this level instance. - */ - NV_DECLARE_ALIGNED(NvU64 size, 8); - - /*! - * Aperture in which this page level instance resides. - */ - NvU32 aperture; - - /*! - * Page shift corresponding to the level - */ - NvU8 pageShift; - } levels[GMMU_FMT_MAX_LEVELS]; -} NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrla06f/ctrla06fgpfifo.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrla06f/ctrla06fgpfifo.h deleted file mode 100644 index 684045796232..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrla06f/ctrla06fgpfifo.h +++ /dev/null @@ -1,42 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrla06f_ctrla06fgpfifo_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrla06f_ctrla06fgpfifo_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2007-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NVA06F_CTRL_CMD_GPFIFO_SCHEDULE (0xa06f0103) /* finn: Evaluated from "(FINN_KEPLER_CHANNEL_GPFIFO_A_GPFIFO_INTERFACE_ID << 8) | NVA06F_CTRL_GPFIFO_SCHEDULE_PARAMS_MESSAGE_ID" */ - -typedef struct NVA06F_CTRL_GPFIFO_SCHEDULE_PARAMS { - NvBool bEnable; - NvBool bSkipSubmit; -} NVA06F_CTRL_GPFIFO_SCHEDULE_PARAMS; - -#define NVA06F_CTRL_CMD_BIND (0xa06f0104) /* finn: Evaluated from "(FINN_KEPLER_CHANNEL_GPFIFO_A_GPFIFO_INTERFACE_ID << 8) | NVA06F_CTRL_BIND_PARAMS_MESSAGE_ID" */ - -typedef struct NVA06F_CTRL_BIND_PARAMS { - NvU32 engineType; -} NVA06F_CTRL_BIND_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/nvlimits.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/nvlimits.h deleted file mode 100644 index 5c5a004a8031..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/nvlimits.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_nvlimits_h__ -#define __src_common_sdk_nvidia_inc_nvlimits_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2017 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV_MAX_SUBDEVICES 8 - -#define NV_PROC_NAME_MAX_LENGTH 100U - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h deleted file mode 100644 index 51b5591c603e..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h +++ /dev/null @@ -1,148 +0,0 @@ -#ifndef __src_common_sdk_nvidia_inc_nvos_h__ -#define __src_common_sdk_nvidia_inc_nvos_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NVOS02_FLAGS_PHYSICALITY 7:4 -#define NVOS02_FLAGS_PHYSICALITY_CONTIGUOUS (0x00000000) -#define NVOS02_FLAGS_PHYSICALITY_NONCONTIGUOUS (0x00000001) -#define NVOS02_FLAGS_LOCATION 11:8 -#define NVOS02_FLAGS_LOCATION_PCI (0x00000000) -#define NVOS02_FLAGS_LOCATION_AGP (0x00000001) -#define NVOS02_FLAGS_LOCATION_VIDMEM (0x00000002) -#define NVOS02_FLAGS_COHERENCY 15:12 -#define NVOS02_FLAGS_COHERENCY_UNCACHED (0x00000000) -#define NVOS02_FLAGS_COHERENCY_CACHED (0x00000001) -#define NVOS02_FLAGS_COHERENCY_WRITE_COMBINE (0x00000002) -#define NVOS02_FLAGS_COHERENCY_WRITE_THROUGH (0x00000003) -#define NVOS02_FLAGS_COHERENCY_WRITE_PROTECT (0x00000004) -#define NVOS02_FLAGS_COHERENCY_WRITE_BACK (0x00000005) -#define NVOS02_FLAGS_ALLOC 17:16 -#define NVOS02_FLAGS_ALLOC_NONE (0x00000001) -#define NVOS02_FLAGS_GPU_CACHEABLE 18:18 -#define NVOS02_FLAGS_GPU_CACHEABLE_NO (0x00000000) -#define NVOS02_FLAGS_GPU_CACHEABLE_YES (0x00000001) - -#define NVOS02_FLAGS_KERNEL_MAPPING 19:19 -#define NVOS02_FLAGS_KERNEL_MAPPING_NO_MAP (0x00000000) -#define NVOS02_FLAGS_KERNEL_MAPPING_MAP (0x00000001) -#define NVOS02_FLAGS_ALLOC_NISO_DISPLAY 20:20 -#define NVOS02_FLAGS_ALLOC_NISO_DISPLAY_NO (0x00000000) -#define NVOS02_FLAGS_ALLOC_NISO_DISPLAY_YES (0x00000001) - -#define NVOS02_FLAGS_ALLOC_USER_READ_ONLY 21:21 -#define NVOS02_FLAGS_ALLOC_USER_READ_ONLY_NO (0x00000000) -#define NVOS02_FLAGS_ALLOC_USER_READ_ONLY_YES (0x00000001) - -#define NVOS02_FLAGS_ALLOC_DEVICE_READ_ONLY 22:22 -#define NVOS02_FLAGS_ALLOC_DEVICE_READ_ONLY_NO (0x00000000) -#define NVOS02_FLAGS_ALLOC_DEVICE_READ_ONLY_YES (0x00000001) - -#define NVOS02_FLAGS_PEER_MAP_OVERRIDE 23:23 -#define NVOS02_FLAGS_PEER_MAP_OVERRIDE_DEFAULT (0x00000000) -#define NVOS02_FLAGS_PEER_MAP_OVERRIDE_REQUIRED (0x00000001) - -#define NVOS02_FLAGS_ALLOC_TYPE_SYNCPOINT 24:24 -#define NVOS02_FLAGS_ALLOC_TYPE_SYNCPOINT_APERTURE (0x00000001) - -#define NVOS02_FLAGS_MEMORY_PROTECTION 26:25 -#define NVOS02_FLAGS_MEMORY_PROTECTION_DEFAULT (0x00000000) -#define NVOS02_FLAGS_MEMORY_PROTECTION_PROTECTED (0x00000001) -#define NVOS02_FLAGS_MEMORY_PROTECTION_UNPROTECTED (0x00000002) - -#define NVOS02_FLAGS_MAPPING 31:30 -#define NVOS02_FLAGS_MAPPING_DEFAULT (0x00000000) -#define NVOS02_FLAGS_MAPPING_NO_MAP (0x00000001) -#define NVOS02_FLAGS_MAPPING_NEVER_MAP (0x00000002) - -#define NV01_EVENT_CLIENT_RM (0x04000000) - -typedef struct -{ - NvV32 channelInstance; // One of the n channel instances of a given channel type. - // Note that core channel has only one instance - // while all others have two (one per head). - NvHandle hObjectBuffer; // ctx dma handle for DMA push buffer - NvHandle hObjectNotify; // ctx dma handle for an area (of type NvNotification defined in sdk/nvidia/inc/nvtypes.h) where RM can write errors/notifications - NvU32 offset; // Initial offset for put/get, usually zero. - NvP64 pControl NV_ALIGN_BYTES(8); // pControl gives virt addr of UDISP GET/PUT regs - - NvU32 flags; -#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB 1:1 -#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB_YES 0x00000000 -#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB_NO 0x00000001 - -} NV50VAIO_CHANNELDMA_ALLOCATION_PARAMETERS; - -typedef struct -{ - NvV32 channelInstance; // One of the n channel instances of a given channel type. - // All PIO channels have two instances (one per head). - NvHandle hObjectNotify; // ctx dma handle for an area (of type NvNotification defined in sdk/nvidia/inc/nvtypes.h) where RM can write errors. - NvP64 pControl NV_ALIGN_BYTES(8); // pControl gives virt addr of control region for PIO channel -} NV50VAIO_CHANNELPIO_ALLOCATION_PARAMETERS; - -typedef struct -{ - NvU32 size; - NvU32 prohibitMultipleInstances; - NvU32 engineInstance; // Select NVDEC0 or NVDEC1 or NVDEC2 -} NV_BSP_ALLOCATION_PARAMETERS; - -typedef struct -{ - NvU32 size; - NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of MSENC? - NvU32 engineInstance; // Select MSENC/NVENC0 or NVENC1 or NVENC2 -} NV_MSENC_ALLOCATION_PARAMETERS; - -typedef struct -{ - NvU32 size; - NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of NVJPG? - NvU32 engineInstance; -} NV_NVJPG_ALLOCATION_PARAMETERS; - -typedef struct -{ - NvU32 size; - NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of OFA? -} NV_OFA_ALLOCATION_PARAMETERS; - -typedef struct -{ - NvU32 index; - NvV32 flags; - NvU64 vaSize NV_ALIGN_BYTES(8); - NvU64 vaStartInternal NV_ALIGN_BYTES(8); - NvU64 vaLimitInternal NV_ALIGN_BYTES(8); - NvU32 bigPageSize; - NvU64 vaBase NV_ALIGN_BYTES(8); -} NV_VASPACE_ALLOCATION_PARAMETERS; - -#define NV_VASPACE_ALLOCATION_INDEX_GPU_NEW 0x00 //<! Create new VASpace, by default - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/shared/msgq/inc/msgq/msgq_priv.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/shared/msgq/inc/msgq/msgq_priv.h deleted file mode 100644 index 0e32e71e123f..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/shared/msgq/inc/msgq/msgq_priv.h +++ /dev/null @@ -1,97 +0,0 @@ -#ifndef __src_common_shared_msgq_inc_msgq_msgq_priv_h__ -#define __src_common_shared_msgq_inc_msgq_msgq_priv_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2018-2019 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -/** - * msgqTxHeader -- TX queue data structure - * @version: the version of this structure, must be 0 - * @size: the size of the entire queue, including this header - * @msgSize: the padded size of queue element, 16 is minimum - * @msgCount: the number of elements in this queue - * @writePtr: head index of this queue - * @flags: 1 = swap the RX pointers - * @rxHdrOff: offset of readPtr in this structure - * @entryOff: offset of beginning of queue (msgqRxHeader), relative to - * beginning of this structure - * - * The command queue is a queue of RPCs that are sent from the driver to the - * GSP. The status queue is a queue of messages/responses from GSP-RM to the - * driver. Although the driver allocates memory for both queues, the command - * queue is owned by the driver and the status queue is owned by GSP-RM. In - * addition, the headers of the two queues must not share the same 4K page. - * - * Each queue is prefixed with this data structure. The idea is that a queue - * and its header are written to only by their owner. That is, only the - * driver writes to the command queue and command queue header, and only the - * GSP writes to the status (receive) queue and its header. - * - * This is enforced by the concept of "swapping" the RX pointers. This is - * why the 'flags' field must be set to 1. 'rxHdrOff' is how the GSP knows - * where the where the tail pointer of its status queue. - * - * When the driver writes a new RPC to the command queue, it updates writePtr. - * When it reads a new message from the status queue, it updates readPtr. In - * this way, the GSP knows when a new command is in the queue (it polls - * writePtr) and it knows how much free space is in the status queue (it - * checks readPtr). The driver never cares about how much free space is in - * the status queue. - * - * As usual, producers write to the head pointer, and consumers read from the - * tail pointer. When head == tail, the queue is empty. - * - * So to summarize: - * command.writePtr = head of command queue - * command.readPtr = tail of status queue - * status.writePtr = head of status queue - * status.readPtr = tail of command queue - */ -typedef struct -{ - NvU32 version; // queue version - NvU32 size; // bytes, page aligned - NvU32 msgSize; // entry size, bytes, must be power-of-2, 16 is minimum - NvU32 msgCount; // number of entries in queue - NvU32 writePtr; // message id of next slot - NvU32 flags; // if set it means "i want to swap RX" - NvU32 rxHdrOff; // Offset of msgqRxHeader from start of backing store. - NvU32 entryOff; // Offset of entries from start of backing store. -} msgqTxHeader; - -/** - * msgqRxHeader - RX queue data structure - * @readPtr: tail index of the other queue - * - * Although this is a separate struct, it could easily be merged into - * msgqTxHeader. msgqTxHeader.rxHdrOff is simply the offset of readPtr - * from the beginning of msgqTxHeader. - */ -typedef struct -{ - NvU32 readPtr; // message id of last message read -} msgqRxHeader; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/uproc/os/common/include/libos_init_args.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/uproc/os/common/include/libos_init_args.h deleted file mode 100644 index 83cf1b2c15a3..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/uproc/os/common/include/libos_init_args.h +++ /dev/null @@ -1,52 +0,0 @@ -#ifndef __src_common_uproc_os_common_include_libos_init_args_h__ -#define __src_common_uproc_os_common_include_libos_init_args_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2018-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef NvU64 LibosAddress; - -typedef enum { - LIBOS_MEMORY_REGION_NONE, - LIBOS_MEMORY_REGION_CONTIGUOUS, - LIBOS_MEMORY_REGION_RADIX3 -} LibosMemoryRegionKind; - -typedef enum { - LIBOS_MEMORY_REGION_LOC_NONE, - LIBOS_MEMORY_REGION_LOC_SYSMEM, - LIBOS_MEMORY_REGION_LOC_FB -} LibosMemoryRegionLoc; - -typedef struct -{ - LibosAddress id8; // Id tag. - LibosAddress pa; // Physical address. - LibosAddress size; // Size of memory area. - NvU8 kind; // See LibosMemoryRegionKind above. - NvU8 loc; // See LibosMemoryRegionLoc above. -} LibosMemoryRegionInitArgument; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_sr_meta.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_sr_meta.h deleted file mode 100644 index 73213bdfcbda..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_sr_meta.h +++ /dev/null @@ -1,79 +0,0 @@ -#ifndef __src_nvidia_arch_nvalloc_common_inc_gsp_gsp_fw_sr_meta_h__ -#define __src_nvidia_arch_nvalloc_common_inc_gsp_gsp_fw_sr_meta_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2022-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define GSP_FW_SR_META_MAGIC 0x8a3bb9e6c6c39d93ULL -#define GSP_FW_SR_META_REVISION 2 - -typedef struct -{ - // - // Magic - // Use for verification by Booter - // - NvU64 magic; // = GSP_FW_SR_META_MAGIC; - - // - // Revision number - // Bumped up when we change this interface so it is not backward compatible. - // Bumped up when we revoke GSP-RM ucode - // - NvU64 revision; // = GSP_FW_SR_META_MAGIC_REVISION; - - // - // ---- Members regarding data in SYSMEM ---------------------------- - // Consumed by Booter for DMA - // - NvU64 sysmemAddrOfSuspendResumeData; - NvU64 sizeOfSuspendResumeData; - - // ---- Members for crypto ops across S/R --------------------------- - - // - // HMAC over the entire GspFwSRMeta structure (including padding) - // with the hmac field itself zeroed. - // - NvU8 hmac[32]; - - // Hash over GspFwWprMeta structure - NvU8 wprMetaHash[32]; - - // Hash over GspFwHeapFreeList structure. All zeros signifies no free list. - NvU8 heapFreeListHash[32]; - - // Hash over data in WPR2 (skipping over free heap chunks; see Booter for details) - NvU8 dataHash[32]; - - // - // Pad structure to exactly 256 bytes (1 DMA chunk). - // Padding initialized to zero. - // - NvU32 padding[24]; - -} GspFwSRMeta; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_wpr_meta.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_wpr_meta.h deleted file mode 100644 index a2e141e4b459..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_wpr_meta.h +++ /dev/null @@ -1,170 +0,0 @@ -#ifndef __src_nvidia_arch_nvalloc_common_inc_gsp_gsp_fw_wpr_meta_h__ -#define __src_nvidia_arch_nvalloc_common_inc_gsp_gsp_fw_wpr_meta_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2021-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct -{ - // Magic - // BL to use for verification (i.e. Booter locked it in WPR2) - NvU64 magic; // = 0xdc3aae21371a60b3; - - // Revision number of Booter-BL-Sequencer handoff interface - // Bumped up when we change this interface so it is not backward compatible. - // Bumped up when we revoke GSP-RM ucode - NvU64 revision; // = 1; - - // ---- Members regarding data in SYSMEM ---------------------------- - // Consumed by Booter for DMA - - NvU64 sysmemAddrOfRadix3Elf; - NvU64 sizeOfRadix3Elf; - - NvU64 sysmemAddrOfBootloader; - NvU64 sizeOfBootloader; - - // Offsets inside bootloader image needed by Booter - NvU64 bootloaderCodeOffset; - NvU64 bootloaderDataOffset; - NvU64 bootloaderManifestOffset; - - union - { - // Used only at initial boot - struct - { - NvU64 sysmemAddrOfSignature; - NvU64 sizeOfSignature; - }; - - // - // Used at suspend/resume to read GspFwHeapFreeList - // Offset relative to GspFwWprMeta FBMEM PA (gspFwWprStart) - // - struct - { - NvU32 gspFwHeapFreeListWprOffset; - NvU32 unused0; - NvU64 unused1; - }; - }; - - // ---- Members describing FB layout -------------------------------- - NvU64 gspFwRsvdStart; - - NvU64 nonWprHeapOffset; - NvU64 nonWprHeapSize; - - NvU64 gspFwWprStart; - - // GSP-RM to use to setup heap. - NvU64 gspFwHeapOffset; - NvU64 gspFwHeapSize; - - // BL to use to find ELF for jump - NvU64 gspFwOffset; - // Size is sizeOfRadix3Elf above. - - NvU64 bootBinOffset; - // Size is sizeOfBootloader above. - - NvU64 frtsOffset; - NvU64 frtsSize; - - NvU64 gspFwWprEnd; - - // GSP-RM to use for fbRegionInfo? - NvU64 fbSize; - - // ---- Other members ----------------------------------------------- - - // GSP-RM to use for fbRegionInfo? - NvU64 vgaWorkspaceOffset; - NvU64 vgaWorkspaceSize; - - // Boot count. Used to determine whether to load the firmware image. - NvU64 bootCount; - - // TODO: the partitionRpc* fields below do not really belong in this - // structure. The values are patched in by the partition bootstrapper - // when GSP-RM is booted in a partition, and this structure was a - // convenient place for the bootstrapper to access them. These should - // be moved to a different comm. mechanism between the bootstrapper - // and the GSP-RM tasks. - - union - { - struct - { - // Shared partition RPC memory (physical address) - NvU64 partitionRpcAddr; - - // Offsets relative to partitionRpcAddr - NvU16 partitionRpcRequestOffset; - NvU16 partitionRpcReplyOffset; - - // Code section and dataSection offset and size. - NvU32 elfCodeOffset; - NvU32 elfDataOffset; - NvU32 elfCodeSize; - NvU32 elfDataSize; - - // Used during GSP-RM resume to check for revocation - NvU32 lsUcodeVersion; - }; - - struct - { - // Pad for the partitionRpc* fields, plus 4 bytes - NvU32 partitionRpcPadding[4]; - - // CrashCat (contiguous) buffer size/location - occupies same bytes as the - // elf(Code|Data)(Offset|Size) fields above. - // TODO: move to GSP_FMC_INIT_PARAMS - NvU64 sysmemAddrOfCrashReportQueue; - NvU32 sizeOfCrashReportQueue; - - // Pad for the lsUcodeVersion field - NvU32 lsUcodeVersionPadding[1]; - }; - }; - - // Number of VF partitions allocating sub-heaps from the WPR heap - // Used during boot to ensure the heap is adequately sized - NvU8 gspFwHeapVfPartitionCount; - - // Pad structure to exactly 256 bytes. Can replace padding with additional - // fields without incrementing revision. Padding initialized to 0. - NvU8 padding[7]; - - // BL to use for verification (i.e. Booter says OK to boot) - NvU64 verified; // 0x0 -> unverified, 0xa0a0a0a0a0a0a0a0 -> verified -} GspFwWprMeta; - -#define GSP_FW_WPR_META_REVISION 1 -#define GSP_FW_WPR_META_MAGIC 0xdc3aae21371a60b3ULL - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmRiscvUcode.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmRiscvUcode.h deleted file mode 100644 index 4eff473e8990..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmRiscvUcode.h +++ /dev/null @@ -1,82 +0,0 @@ -#ifndef __src_nvidia_arch_nvalloc_common_inc_rmRiscvUcode_h__ -#define __src_nvidia_arch_nvalloc_common_inc_rmRiscvUcode_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2018-2019 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct { - // - // Version 1 - // Version 2 - // Version 3 = for Partition boot - // Version 4 = for eb riscv boot - // Version 5 = Support signing entire RISC-V image as "code" in code section for hopper and later. - // - NvU32 version; // structure version - NvU32 bootloaderOffset; - NvU32 bootloaderSize; - NvU32 bootloaderParamOffset; - NvU32 bootloaderParamSize; - NvU32 riscvElfOffset; - NvU32 riscvElfSize; - NvU32 appVersion; // Changelist number associated with the image - // - // Manifest contains information about Monitor and it is - // input to BR - // - NvU32 manifestOffset; - NvU32 manifestSize; - // - // Monitor Data offset within RISCV image and size - // - NvU32 monitorDataOffset; - NvU32 monitorDataSize; - // - // Monitor Code offset withtin RISCV image and size - // - NvU32 monitorCodeOffset; - NvU32 monitorCodeSize; - NvU32 bIsMonitorEnabled; - // - // Swbrom Code offset within RISCV image and size - // - NvU32 swbromCodeOffset; - NvU32 swbromCodeSize; - // - // Swbrom Data offset within RISCV image and size - // - NvU32 swbromDataOffset; - NvU32 swbromDataSize; - // - // Total size of FB carveout (image and reserved space). - // - NvU32 fbReservedSize; - // - // Indicates whether the entire RISC-V image is signed as "code" in code section. - // - NvU32 bSignedAsCode; -} RM_RISCV_UCODE_DESC; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmgspseq.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmgspseq.h deleted file mode 100644 index 341ab0dbeaf2..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmgspseq.h +++ /dev/null @@ -1,100 +0,0 @@ -#ifndef __src_nvidia_arch_nvalloc_common_inc_rmgspseq_h__ -#define __src_nvidia_arch_nvalloc_common_inc_rmgspseq_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2019-2020 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef enum GSP_SEQ_BUF_OPCODE -{ - GSP_SEQ_BUF_OPCODE_REG_WRITE = 0, - GSP_SEQ_BUF_OPCODE_REG_MODIFY, - GSP_SEQ_BUF_OPCODE_REG_POLL, - GSP_SEQ_BUF_OPCODE_DELAY_US, - GSP_SEQ_BUF_OPCODE_REG_STORE, - GSP_SEQ_BUF_OPCODE_CORE_RESET, - GSP_SEQ_BUF_OPCODE_CORE_START, - GSP_SEQ_BUF_OPCODE_CORE_WAIT_FOR_HALT, - GSP_SEQ_BUF_OPCODE_CORE_RESUME, -} GSP_SEQ_BUF_OPCODE; - -#define GSP_SEQUENCER_PAYLOAD_SIZE_DWORDS(opcode) \ - ((opcode == GSP_SEQ_BUF_OPCODE_REG_WRITE) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_WRITE) / sizeof(NvU32)) : \ - (opcode == GSP_SEQ_BUF_OPCODE_REG_MODIFY) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_MODIFY) / sizeof(NvU32)) : \ - (opcode == GSP_SEQ_BUF_OPCODE_REG_POLL) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_POLL) / sizeof(NvU32)) : \ - (opcode == GSP_SEQ_BUF_OPCODE_DELAY_US) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_DELAY_US) / sizeof(NvU32)) : \ - (opcode == GSP_SEQ_BUF_OPCODE_REG_STORE) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_STORE) / sizeof(NvU32)) : \ - /* GSP_SEQ_BUF_OPCODE_CORE_RESET */ \ - /* GSP_SEQ_BUF_OPCODE_CORE_START */ \ - /* GSP_SEQ_BUF_OPCODE_CORE_WAIT_FOR_HALT */ \ - /* GSP_SEQ_BUF_OPCODE_CORE_RESUME */ \ - 0) - -typedef struct -{ - NvU32 addr; - NvU32 val; -} GSP_SEQ_BUF_PAYLOAD_REG_WRITE; - -typedef struct -{ - NvU32 addr; - NvU32 mask; - NvU32 val; -} GSP_SEQ_BUF_PAYLOAD_REG_MODIFY; - -typedef struct -{ - NvU32 addr; - NvU32 mask; - NvU32 val; - NvU32 timeout; - NvU32 error; -} GSP_SEQ_BUF_PAYLOAD_REG_POLL; - -typedef struct -{ - NvU32 val; -} GSP_SEQ_BUF_PAYLOAD_DELAY_US; - -typedef struct -{ - NvU32 addr; - NvU32 index; -} GSP_SEQ_BUF_PAYLOAD_REG_STORE; - -typedef struct GSP_SEQUENCER_BUFFER_CMD -{ - GSP_SEQ_BUF_OPCODE opCode; - union - { - GSP_SEQ_BUF_PAYLOAD_REG_WRITE regWrite; - GSP_SEQ_BUF_PAYLOAD_REG_MODIFY regModify; - GSP_SEQ_BUF_PAYLOAD_REG_POLL regPoll; - GSP_SEQ_BUF_PAYLOAD_DELAY_US delayUs; - GSP_SEQ_BUF_PAYLOAD_REG_STORE regStore; - } payload; -} GSP_SEQUENCER_BUFFER_CMD; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_allclasses.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_allclasses.h deleted file mode 100644 index 3144e9beac61..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_allclasses.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef __src_nvidia_generated_g_allclasses_h__ -#define __src_nvidia_generated_g_allclasses_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2021-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV01_EVENT_KERNEL_CALLBACK_EX (0x0000007e) - -#define NV04_DISPLAY_COMMON (0x00000073) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_chipset_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_chipset_nvoc.h deleted file mode 100644 index 6b8921138c7d..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_chipset_nvoc.h +++ /dev/null @@ -1,38 +0,0 @@ -#ifndef __src_nvidia_generated_g_chipset_nvoc_h__ -#define __src_nvidia_generated_g_chipset_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 1993-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct -{ - NvU16 deviceID; // deviceID - NvU16 vendorID; // vendorID - NvU16 subdeviceID; // subsystem deviceID - NvU16 subvendorID; // subsystem vendorID - NvU8 revisionID; // revision ID -} BUSINFO; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_fbsr_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_fbsr_nvoc.h deleted file mode 100644 index a5128f00225b..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_fbsr_nvoc.h +++ /dev/null @@ -1,31 +0,0 @@ -#ifndef __src_nvidia_generated_g_fbsr_nvoc_h__ -#define __src_nvidia_generated_g_fbsr_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2009-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define FBSR_TYPE_DMA 4 // Copy using DMA. Fastest. - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_gpu_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_gpu_nvoc.h deleted file mode 100644 index 5641a21cacca..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_gpu_nvoc.h +++ /dev/null @@ -1,35 +0,0 @@ -#ifndef __src_nvidia_generated_g_gpu_nvoc_h__ -#define __src_nvidia_generated_g_gpu_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2004-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef enum -{ - COMPUTE_BRANDING_TYPE_NONE, - COMPUTE_BRANDING_TYPE_TESLA, -} COMPUTE_BRANDING_TYPE; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_kernel_channel_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_kernel_channel_nvoc.h deleted file mode 100644 index b5ad55f854dc..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_kernel_channel_nvoc.h +++ /dev/null @@ -1,62 +0,0 @@ -#ifndef __src_nvidia_generated_g_kernel_channel_nvoc_h__ -#define __src_nvidia_generated_g_kernel_channel_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2020-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef enum { - /*! - * Initial state as passed in NV_CHANNEL_ALLOC_PARAMS by - * kernel CPU-RM clients. - */ - ERROR_NOTIFIER_TYPE_UNKNOWN = 0, - /*! @brief Error notifier is explicitly not set. - * - * The corresponding hErrorContext or hEccErrorContext must be - * NV01_NULL_OBJECT. - */ - ERROR_NOTIFIER_TYPE_NONE, - /*! @brief Error notifier is a ContextDma */ - ERROR_NOTIFIER_TYPE_CTXDMA, - /*! @brief Error notifier is a NvNotification array in sysmem/vidmem */ - ERROR_NOTIFIER_TYPE_MEMORY -} ErrorNotifierType; - -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE 1:0 -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_USER 0x0 -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_ADMIN 0x1 -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_KERNEL 0x2 -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE 3:2 -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_UNKNOWN ERROR_NOTIFIER_TYPE_UNKNOWN -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_NONE ERROR_NOTIFIER_TYPE_NONE -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_CTXDMA ERROR_NOTIFIER_TYPE_CTXDMA -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_MEMORY ERROR_NOTIFIER_TYPE_MEMORY -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE 5:4 -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_UNKNOWN ERROR_NOTIFIER_TYPE_UNKNOWN -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_NONE ERROR_NOTIFIER_TYPE_NONE -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_CTXDMA ERROR_NOTIFIER_TYPE_CTXDMA -#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_MEMORY ERROR_NOTIFIER_TYPE_MEMORY - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_kernel_fifo_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_kernel_fifo_nvoc.h deleted file mode 100644 index 946954ac5b3d..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_kernel_fifo_nvoc.h +++ /dev/null @@ -1,119 +0,0 @@ -#ifndef __src_nvidia_generated_g_kernel_fifo_nvoc_h__ -#define __src_nvidia_generated_g_kernel_fifo_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2021-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef enum -{ - /* ************************************************************************* - * Bug 3820969 - * THINK BEFORE CHANGING ENUM ORDER HERE. - * VGPU-guest uses this same ordering. Because this enum is not versioned, - * changing the order here WILL BREAK old-guest-on-newer-host compatibility. - * ************************************************************************/ - - // *ENG_XYZ, e.g.: ENG_GR, ENG_CE etc., - ENGINE_INFO_TYPE_ENG_DESC = 0, - - // HW engine ID - ENGINE_INFO_TYPE_FIFO_TAG, - - // RM_ENGINE_TYPE_* - ENGINE_INFO_TYPE_RM_ENGINE_TYPE, - - // - // runlist id (meaning varies by GPU) - // Valid only for Esched-driven engines - // - ENGINE_INFO_TYPE_RUNLIST, - - // NV_PFIFO_INTR_MMU_FAULT_ENG_ID_* - ENGINE_INFO_TYPE_MMU_FAULT_ID, - - // ROBUST_CHANNEL_* - ENGINE_INFO_TYPE_RC_MASK, - - // Reset Bit Position. On Ampere, only valid if not _INVALID - ENGINE_INFO_TYPE_RESET, - - // Interrupt Bit Position - ENGINE_INFO_TYPE_INTR, - - // log2(MC_ENGINE_*) - ENGINE_INFO_TYPE_MC, - - // The DEV_TYPE_ENUM for this engine - ENGINE_INFO_TYPE_DEV_TYPE_ENUM, - - // The particular instance of this engine type - ENGINE_INFO_TYPE_INSTANCE_ID, - - // - // The base address for this engine's NV_RUNLIST. Valid only on Ampere+ - // Valid only for Esched-driven engines - // - ENGINE_INFO_TYPE_RUNLIST_PRI_BASE, - - // - // If this entry is a host-driven engine. - // Update _isEngineInfoTypeValidForOnlyHostDriven when adding any new entry. - // - ENGINE_INFO_TYPE_IS_HOST_DRIVEN_ENGINE, - - // - // The index into the per-engine NV_RUNLIST registers. Valid only on Ampere+ - // Valid only for Esched-driven engines - // - ENGINE_INFO_TYPE_RUNLIST_ENGINE_ID, - - // - // The base address for this engine's NV_CHRAM registers. Valid only on - // Ampere+ - // - // Valid only for Esched-driven engines - // - ENGINE_INFO_TYPE_CHRAM_PRI_BASE, - - // This entry added to copy data at RMCTRL_EXPORT() call for Kernel RM - ENGINE_INFO_TYPE_KERNEL_RM_MAX, - // Used for iterating the engine info table by the index passed. - ENGINE_INFO_TYPE_INVALID = ENGINE_INFO_TYPE_KERNEL_RM_MAX, - - // Size of FIFO_ENGINE_LIST.engineData - ENGINE_INFO_TYPE_ENGINE_DATA_ARRAY_SIZE = ENGINE_INFO_TYPE_INVALID, - - // Input-only parameter for kfifoEngineInfoXlate. - ENGINE_INFO_TYPE_PBDMA_ID - - /* ************************************************************************* - * Bug 3820969 - * THINK BEFORE CHANGING ENUM ORDER HERE. - * VGPU-guest uses this same ordering. Because this enum is not versioned, - * changing the order here WILL BREAK old-guest-on-newer-host compatibility. - * ************************************************************************/ -} ENGINE_INFO_TYPE; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_mem_desc_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_mem_desc_nvoc.h deleted file mode 100644 index daabaee41c87..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_mem_desc_nvoc.h +++ /dev/null @@ -1,32 +0,0 @@ -#ifndef __src_nvidia_generated_g_mem_desc_nvoc_h__ -#define __src_nvidia_generated_g_mem_desc_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define ADDR_SYSMEM 1 // System memory (PCI) -#define ADDR_FBMEM 2 // Frame buffer memory space - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_os_nvoc.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_os_nvoc.h deleted file mode 100644 index 10121218f4d3..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_os_nvoc.h +++ /dev/null @@ -1,44 +0,0 @@ -#ifndef __src_nvidia_generated_g_os_nvoc_h__ -#define __src_nvidia_generated_g_os_nvoc_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct PACKED_REGISTRY_ENTRY -{ - NvU32 nameOffset; - NvU8 type; - NvU32 data; - NvU32 length; -} PACKED_REGISTRY_ENTRY; - -typedef struct PACKED_REGISTRY_TABLE -{ - NvU32 size; - NvU32 numEntries; - PACKED_REGISTRY_ENTRY entries[] __counted_by(numEntries); -} PACKED_REGISTRY_TABLE; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_rpc-structures.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_rpc-structures.h deleted file mode 100644 index 8d925e24faea..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_rpc-structures.h +++ /dev/null @@ -1,124 +0,0 @@ -#ifndef __src_nvidia_generated_g_rpc_structures_h__ -#define __src_nvidia_generated_g_rpc_structures_h__ -#include <nvrm/535.113.01/nvidia/generated/g_sdk-structures.h> -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/sdk-structures.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2008-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct rpc_alloc_memory_v13_01 -{ - NvHandle hClient; - NvHandle hDevice; - NvHandle hMemory; - NvU32 hClass; - NvU32 flags; - NvU32 pteAdjust; - NvU32 format; - NvU64 length NV_ALIGN_BYTES(8); - NvU32 pageCount; - struct pte_desc pteDesc; -} rpc_alloc_memory_v13_01; - -typedef struct rpc_free_v03_00 -{ - NVOS00_PARAMETERS_v03_00 params; -} rpc_free_v03_00; - -typedef struct rpc_unloading_guest_driver_v1F_07 -{ - NvBool bInPMTransition; - NvBool bGc6Entering; - NvU32 newLevel; -} rpc_unloading_guest_driver_v1F_07; - -typedef struct rpc_update_bar_pde_v15_00 -{ - UpdateBarPde_v15_00 info; -} rpc_update_bar_pde_v15_00; - -typedef struct rpc_gsp_rm_alloc_v03_00 -{ - NvHandle hClient; - NvHandle hParent; - NvHandle hObject; - NvU32 hClass; - NvU32 status; - NvU32 paramsSize; - NvU32 flags; - NvU8 reserved[4]; - NvU8 params[]; -} rpc_gsp_rm_alloc_v03_00; - -typedef struct rpc_gsp_rm_control_v03_00 -{ - NvHandle hClient; - NvHandle hObject; - NvU32 cmd; - NvU32 status; - NvU32 paramsSize; - NvU32 flags; - NvU8 params[]; -} rpc_gsp_rm_control_v03_00; - -typedef struct rpc_run_cpu_sequencer_v17_00 -{ - NvU32 bufferSizeDWord; - NvU32 cmdIndex; - NvU32 regSaveArea[8]; - NvU32 commandBuffer[]; -} rpc_run_cpu_sequencer_v17_00; - -typedef struct rpc_post_event_v17_00 -{ - NvHandle hClient; - NvHandle hEvent; - NvU32 notifyIndex; - NvU32 data; - NvU16 info16; - NvU32 status; - NvU32 eventDataSize; - NvBool bNotifyList; - NvU8 eventData[]; -} rpc_post_event_v17_00; - -typedef struct rpc_rc_triggered_v17_02 -{ - NvU32 nv2080EngineType; - NvU32 chid; - NvU32 exceptType; - NvU32 scope; - NvU16 partitionAttributionId; -} rpc_rc_triggered_v17_02; - -typedef struct rpc_os_error_log_v17_00 -{ - NvU32 exceptType; - NvU32 runlistId; - NvU32 chid; - char errString[0x100]; -} rpc_os_error_log_v17_00; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_sdk-structures.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_sdk-structures.h deleted file mode 100644 index e9fed4140468..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/generated/g_sdk-structures.h +++ /dev/null @@ -1,45 +0,0 @@ -#ifndef __src_nvidia_generated_g_sdk_structures_h__ -#define __src_nvidia_generated_g_sdk_structures_h__ -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_headers.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2008-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct NVOS00_PARAMETERS_v03_00 -{ - NvHandle hRoot; - NvHandle hObjectParent; - NvHandle hObjectOld; - NvV32 status; -} NVOS00_PARAMETERS_v03_00; - -typedef struct UpdateBarPde_v15_00 -{ - NV_RPC_UPDATE_PDE_BAR_TYPE barType; - NvU64 entryValue NV_ALIGN_BYTES(8); - NvU64 entryLevelShift NV_ALIGN_BYTES(8); -} UpdateBarPde_v15_00; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_acpi_data.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_acpi_data.h deleted file mode 100644 index af50b11ec3b4..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_acpi_data.h +++ /dev/null @@ -1,74 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_gpu_acpi_data_h__ -#define __src_nvidia_inc_kernel_gpu_gpu_acpi_data_h__ -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073system.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct DOD_METHOD_DATA -{ - NV_STATUS status; - NvU32 acpiIdListLen; - NvU32 acpiIdList[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; -} DOD_METHOD_DATA; - -typedef struct JT_METHOD_DATA -{ - NV_STATUS status; - NvU32 jtCaps; - NvU16 jtRevId; - NvBool bSBIOSCaps; -} JT_METHOD_DATA; - -typedef struct MUX_METHOD_DATA_ELEMENT -{ - NvU32 acpiId; - NvU32 mode; - NV_STATUS status; -} MUX_METHOD_DATA_ELEMENT; - -typedef struct MUX_METHOD_DATA -{ - NvU32 tableLen; - MUX_METHOD_DATA_ELEMENT acpiIdMuxModeTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; - MUX_METHOD_DATA_ELEMENT acpiIdMuxPartTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; -} MUX_METHOD_DATA; - -typedef struct CAPS_METHOD_DATA -{ - NV_STATUS status; - NvU32 optimusCaps; -} CAPS_METHOD_DATA; - -typedef struct ACPI_METHOD_DATA -{ - NvBool bValid; - DOD_METHOD_DATA dodMethodData; - JT_METHOD_DATA jtMethodData; - MUX_METHOD_DATA muxMethodData; - CAPS_METHOD_DATA capsMethodData; -} ACPI_METHOD_DATA; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_engine_type.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_engine_type.h deleted file mode 100644 index e3160c60036d..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_engine_type.h +++ /dev/null @@ -1,86 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_gpu_engine_type_h__ -#define __src_nvidia_inc_kernel_gpu_gpu_engine_type_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2021-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef enum -{ - RM_ENGINE_TYPE_NULL = (0x00000000), - RM_ENGINE_TYPE_GR0 = (0x00000001), - RM_ENGINE_TYPE_GR1 = (0x00000002), - RM_ENGINE_TYPE_GR2 = (0x00000003), - RM_ENGINE_TYPE_GR3 = (0x00000004), - RM_ENGINE_TYPE_GR4 = (0x00000005), - RM_ENGINE_TYPE_GR5 = (0x00000006), - RM_ENGINE_TYPE_GR6 = (0x00000007), - RM_ENGINE_TYPE_GR7 = (0x00000008), - RM_ENGINE_TYPE_COPY0 = (0x00000009), - RM_ENGINE_TYPE_COPY1 = (0x0000000a), - RM_ENGINE_TYPE_COPY2 = (0x0000000b), - RM_ENGINE_TYPE_COPY3 = (0x0000000c), - RM_ENGINE_TYPE_COPY4 = (0x0000000d), - RM_ENGINE_TYPE_COPY5 = (0x0000000e), - RM_ENGINE_TYPE_COPY6 = (0x0000000f), - RM_ENGINE_TYPE_COPY7 = (0x00000010), - RM_ENGINE_TYPE_COPY8 = (0x00000011), - RM_ENGINE_TYPE_COPY9 = (0x00000012), - RM_ENGINE_TYPE_NVDEC0 = (0x0000001d), - RM_ENGINE_TYPE_NVDEC1 = (0x0000001e), - RM_ENGINE_TYPE_NVDEC2 = (0x0000001f), - RM_ENGINE_TYPE_NVDEC3 = (0x00000020), - RM_ENGINE_TYPE_NVDEC4 = (0x00000021), - RM_ENGINE_TYPE_NVDEC5 = (0x00000022), - RM_ENGINE_TYPE_NVDEC6 = (0x00000023), - RM_ENGINE_TYPE_NVDEC7 = (0x00000024), - RM_ENGINE_TYPE_NVENC0 = (0x00000025), - RM_ENGINE_TYPE_NVENC1 = (0x00000026), - RM_ENGINE_TYPE_NVENC2 = (0x00000027), - RM_ENGINE_TYPE_VP = (0x00000028), - RM_ENGINE_TYPE_ME = (0x00000029), - RM_ENGINE_TYPE_PPP = (0x0000002a), - RM_ENGINE_TYPE_MPEG = (0x0000002b), - RM_ENGINE_TYPE_SW = (0x0000002c), - RM_ENGINE_TYPE_TSEC = (0x0000002d), - RM_ENGINE_TYPE_VIC = (0x0000002e), - RM_ENGINE_TYPE_MP = (0x0000002f), - RM_ENGINE_TYPE_SEC2 = (0x00000030), - RM_ENGINE_TYPE_HOST = (0x00000031), - RM_ENGINE_TYPE_DPU = (0x00000032), - RM_ENGINE_TYPE_PMU = (0x00000033), - RM_ENGINE_TYPE_FBFLCN = (0x00000034), - RM_ENGINE_TYPE_NVJPEG0 = (0x00000035), - RM_ENGINE_TYPE_NVJPEG1 = (0x00000036), - RM_ENGINE_TYPE_NVJPEG2 = (0x00000037), - RM_ENGINE_TYPE_NVJPEG3 = (0x00000038), - RM_ENGINE_TYPE_NVJPEG4 = (0x00000039), - RM_ENGINE_TYPE_NVJPEG5 = (0x0000003a), - RM_ENGINE_TYPE_NVJPEG6 = (0x0000003b), - RM_ENGINE_TYPE_NVJPEG7 = (0x0000003c), - RM_ENGINE_TYPE_OFA = (0x0000003d), - RM_ENGINE_TYPE_LAST = (0x0000003e), -} RM_ENGINE_TYPE; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_fw_heap.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_fw_heap.h deleted file mode 100644 index 3abec59f0cc4..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_fw_heap.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_gsp_gsp_fw_heap_h__ -#define __src_nvidia_inc_kernel_gpu_gsp_gsp_fw_heap_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2022-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define GSP_FW_HEAP_PARAM_SIZE_PER_GB_FB (96 << 10) // All architectures - -#define GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE ((48 << 10) * 2048) // Support 2048 channels - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_init_args.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_init_args.h deleted file mode 100644 index 4033a6f85a76..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_init_args.h +++ /dev/null @@ -1,57 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_gsp_gsp_init_args_h__ -#define __src_nvidia_inc_kernel_gpu_gsp_gsp_init_args_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2020-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct { - RmPhysAddr sharedMemPhysAddr; - NvU32 pageTableEntryCount; - NvLength cmdQueueOffset; - NvLength statQueueOffset; - NvLength locklessCmdQueueOffset; - NvLength locklessStatQueueOffset; -} MESSAGE_QUEUE_INIT_ARGUMENTS; - -typedef struct { - NvU32 oldLevel; - NvU32 flags; - NvBool bInPMTransition; -} GSP_SR_INIT_ARGUMENTS; - -typedef struct -{ - MESSAGE_QUEUE_INIT_ARGUMENTS messageQueueInitArguments; - GSP_SR_INIT_ARGUMENTS srInitArguments; - NvU32 gpuInstance; - - struct - { - NvU64 pa; - NvU64 size; - } profilerArgs; -} GSP_ARGUMENTS_CACHED; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_static_config.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_static_config.h deleted file mode 100644 index eeab25a5e290..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_static_config.h +++ /dev/null @@ -1,174 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_gsp_gsp_static_config_h__ -#define __src_nvidia_inc_kernel_gpu_gsp_gsp_static_config_h__ -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gpu.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080gr.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080bios.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fb.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gr.h> -#include <nvrm/535.113.01/nvidia/generated/g_chipset_nvoc.h> -#include <nvrm/535.113.01/nvidia/generated/g_gpu_nvoc.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_acpi_data.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/nvbitmask.h> -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_headers.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2019-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -typedef struct GSP_VF_INFO -{ - NvU32 totalVFs; - NvU32 firstVFOffset; - NvU64 FirstVFBar0Address; - NvU64 FirstVFBar1Address; - NvU64 FirstVFBar2Address; - NvBool b64bitBar0; - NvBool b64bitBar1; - NvBool b64bitBar2; -} GSP_VF_INFO; - -typedef struct GspSMInfo_t -{ - NvU32 version; - NvU32 regBankCount; - NvU32 regBankRegCount; - NvU32 maxWarpsPerSM; - NvU32 maxThreadsPerWarp; - NvU32 geomGsObufEntries; - NvU32 geomXbufEntries; - NvU32 maxSPPerSM; - NvU32 rtCoreCount; -} GspSMInfo; - -typedef struct GspStaticConfigInfo_t -{ - NvU8 grCapsBits[NV0080_CTRL_GR_CAPS_TBL_SIZE]; - NV2080_CTRL_GPU_GET_GID_INFO_PARAMS gidInfo; - NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS gpcInfo; - NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS tpcInfo[MAX_GPC_COUNT]; - NV2080_CTRL_GPU_GET_FERMI_ZCULL_INFO_PARAMS zcullInfo[MAX_GPC_COUNT]; - NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS SKUInfo; - NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS fbRegionInfoParams; - COMPUTE_BRANDING_TYPE computeBranding; - - NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS sriovCaps; - NvU32 sriovMaxGfid; - - NvU32 engineCaps[NVGPU_ENGINE_CAPS_MASK_ARRAY_MAX]; - - GspSMInfo SM_info; - - NvBool poisonFuseEnabled; - - NvU64 fb_length; - NvU32 fbio_mask; - NvU32 fb_bus_width; - NvU32 fb_ram_type; - NvU32 fbp_mask; - NvU32 l2_cache_size; - - NvU32 gfxpBufferSize[NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL]; - NvU32 gfxpBufferAlignment[NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL]; - - NvU8 gpuNameString[NV2080_GPU_MAX_NAME_STRING_LENGTH]; - NvU8 gpuShortNameString[NV2080_GPU_MAX_NAME_STRING_LENGTH]; - NvU16 gpuNameString_Unicode[NV2080_GPU_MAX_NAME_STRING_LENGTH]; - NvBool bGpuInternalSku; - NvBool bIsQuadroGeneric; - NvBool bIsQuadroAd; - NvBool bIsNvidiaNvs; - NvBool bIsVgx; - NvBool bGeforceSmb; - NvBool bIsTitan; - NvBool bIsTesla; - NvBool bIsMobile; - NvBool bIsGc6Rtd3Allowed; - NvBool bIsGcOffRtd3Allowed; - NvBool bIsGcoffLegacyAllowed; - - NvU64 bar1PdeBase; - NvU64 bar2PdeBase; - - NvBool bVbiosValid; - NvU32 vbiosSubVendor; - NvU32 vbiosSubDevice; - - NvBool bPageRetirementSupported; - - NvBool bSplitVasBetweenServerClientRm; - - NvBool bClRootportNeedsNosnoopWAR; - - VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS displaylessMaxHeads; - VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS displaylessMaxResolution; - NvU64 displaylessMaxPixels; - - // Client handle for internal RMAPI control. - NvHandle hInternalClient; - - // Device handle for internal RMAPI control. - NvHandle hInternalDevice; - - // Subdevice handle for internal RMAPI control. - NvHandle hInternalSubdevice; - - NvBool bSelfHostedMode; - NvBool bAtsSupported; - - NvBool bIsGpuUefi; -} GspStaticConfigInfo; - -typedef struct GspSystemInfo -{ - NvU64 gpuPhysAddr; - NvU64 gpuPhysFbAddr; - NvU64 gpuPhysInstAddr; - NvU64 nvDomainBusDeviceFunc; - NvU64 simAccessBufPhysAddr; - NvU64 pcieAtomicsOpMask; - NvU64 consoleMemSize; - NvU64 maxUserVa; - NvU32 pciConfigMirrorBase; - NvU32 pciConfigMirrorSize; - NvU8 oorArch; - NvU64 clPdbProperties; - NvU32 Chipset; - NvBool bGpuBehindBridge; - NvBool bMnocAvailable; - NvBool bUpstreamL0sUnsupported; - NvBool bUpstreamL1Unsupported; - NvBool bUpstreamL1PorSupported; - NvBool bUpstreamL1PorMobileOnly; - NvU8 upstreamAddressValid; - BUSINFO FHBBusInfo; - BUSINFO chipsetIDInfo; - ACPI_METHOD_DATA acpiMethodData; - NvU32 hypervisorType; - NvBool bIsPassthru; - NvU64 sysTimerOffsetNs; - GSP_VF_INFO gspVFInfo; -} GspSystemInfo; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/intr/engine_idx.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/intr/engine_idx.h deleted file mode 100644 index bd5e01f9814b..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/intr/engine_idx.h +++ /dev/null @@ -1,57 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_intr_engine_idx_h__ -#define __src_nvidia_inc_kernel_gpu_intr_engine_idx_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define MC_ENGINE_IDX_DISP 2 - -#define MC_ENGINE_IDX_CE0 15 - -#define MC_ENGINE_IDX_CE9 24 - -#define MC_ENGINE_IDX_MSENC 38 - -#define MC_ENGINE_IDX_MSENC2 40 - -#define MC_ENGINE_IDX_GSP 49 -#define MC_ENGINE_IDX_NVJPG 50 -#define MC_ENGINE_IDX_NVJPEG MC_ENGINE_IDX_NVJPG -#define MC_ENGINE_IDX_NVJPEG0 MC_ENGINE_IDX_NVJPEG - -#define MC_ENGINE_IDX_NVJPEG7 57 - -#define MC_ENGINE_IDX_BSP 64 -#define MC_ENGINE_IDX_NVDEC MC_ENGINE_IDX_BSP -#define MC_ENGINE_IDX_NVDEC0 MC_ENGINE_IDX_NVDEC - -#define MC_ENGINE_IDX_NVDEC7 71 - -#define MC_ENGINE_IDX_OFA0 80 - -#define MC_ENGINE_IDX_GR 82 -#define MC_ENGINE_IDX_GR0 MC_ENGINE_IDX_GR - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/nvbitmask.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/nvbitmask.h deleted file mode 100644 index 366447a368bf..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/gpu/nvbitmask.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_gpu_nvbitmask_h__ -#define __src_nvidia_inc_kernel_gpu_nvbitmask_h__ -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_engine_type.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2021-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NVGPU_ENGINE_CAPS_MASK_BITS 32 -#define NVGPU_ENGINE_CAPS_MASK_ARRAY_MAX ((RM_ENGINE_TYPE_LAST-1)/NVGPU_ENGINE_CAPS_MASK_BITS + 1) - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/os/nv_memory_type.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/os/nv_memory_type.h deleted file mode 100644 index 4a850dad4776..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/inc/kernel/os/nv_memory_type.h +++ /dev/null @@ -1,31 +0,0 @@ -#ifndef __src_nvidia_inc_kernel_os_nv_memory_type_h__ -#define __src_nvidia_inc_kernel_os_nv_memory_type_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2020 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define NV_MEMORY_WRITECOMBINED 2 - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_headers.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_headers.h deleted file mode 100644 index f14b23852456..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_headers.h +++ /dev/null @@ -1,51 +0,0 @@ -#ifndef __src_nvidia_kernel_inc_vgpu_rpc_headers_h__ -#define __src_nvidia_kernel_inc_vgpu_rpc_headers_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2017-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#define MAX_GPC_COUNT 32 - -typedef enum -{ - NV_RPC_UPDATE_PDE_BAR_1, - NV_RPC_UPDATE_PDE_BAR_2, - NV_RPC_UPDATE_PDE_BAR_INVALID, -} NV_RPC_UPDATE_PDE_BAR_TYPE; - -typedef struct VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS -{ - NvU32 headIndex; - NvU32 maxHResolution; - NvU32 maxVResolution; -} VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS; - -typedef struct VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS -{ - NvU32 numHeads; - NvU32 maxNumHeads; -} VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/sdk-structures.h b/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/sdk-structures.h deleted file mode 100644 index 7801af232dff..000000000000 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/sdk-structures.h +++ /dev/null @@ -1,40 +0,0 @@ -#ifndef __src_nvidia_kernel_inc_vgpu_sdk_structures_h__ -#define __src_nvidia_kernel_inc_vgpu_sdk_structures_h__ - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2020-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -struct pte_desc -{ - NvU32 idr:2; - NvU32 reserved1:14; - NvU32 length:16; - union { - NvU64 pte; // PTE when IDR==0; PDE when IDR > 0 - NvU64 pde; // PTE when IDR==0; PDE when IDR > 0 - } pte_pde[] NV_ALIGN_BYTES(8); // PTE when IDR==0; PDE when IDR > 0 -}; - -#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/nvtypes.h b/drivers/gpu/drm/nouveau/include/nvrm/nvtypes.h index e6833df1ccc7..af11648ad9c8 100644 --- a/drivers/gpu/drm/nouveau/include/nvrm/nvtypes.h +++ b/drivers/gpu/drm/nouveau/include/nvrm/nvtypes.h @@ -21,4 +21,6 @@ typedef NvU64 NvLength; typedef NvU64 RmPhysAddr; typedef NvU32 NV_STATUS; + +typedef union {} rpc_generic_union; #endif diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c index 2a0617e5fe2a..a3ba07fc48a0 100644 --- a/drivers/gpu/drm/nouveau/nouveau_abi16.c +++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c @@ -315,7 +315,7 @@ nouveau_abi16_ioctl_getparam(ABI16_IOCTL_ARGS) break; } case NOUVEAU_GETPARAM_VRAM_BAR_SIZE: - getparam->value = nvkm_device->func->resource_size(nvkm_device, 1); + getparam->value = nvkm_device->func->resource_size(nvkm_device, NVKM_BAR1_FB); break; case NOUVEAU_GETPARAM_VRAM_USED: { struct ttm_resource_manager *vram_mgr = ttm_manager_type(&drm->ttm.bdev, TTM_PL_VRAM); @@ -416,7 +416,7 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS) */ if (nouveau_cli_uvmm(cli)) { ret = nouveau_sched_create(&chan->sched, drm, drm->sched_wq, - chan->chan->dma.ib_max); + chan->chan->chan.gpfifo.max); if (ret) goto done; } diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 2016c1e7242f..b96f0555ca14 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -401,6 +401,83 @@ nouveau_bo_new(struct nouveau_cli *cli, u64 size, int align, return 0; } +void +nouveau_bo_unpin_del(struct nouveau_bo **pnvbo) +{ + struct nouveau_bo *nvbo = *pnvbo; + + if (!nvbo) + return; + + nouveau_bo_unmap(nvbo); + nouveau_bo_unpin(nvbo); + nouveau_bo_fini(nvbo); + + *pnvbo = NULL; +} + +int +nouveau_bo_new_pin(struct nouveau_cli *cli, u32 domain, u32 size, struct nouveau_bo **pnvbo) +{ + struct nouveau_bo *nvbo; + int ret; + + ret = nouveau_bo_new(cli, size, 0, domain, 0, 0, NULL, NULL, &nvbo); + if (ret) + return ret; + + ret = nouveau_bo_pin(nvbo, domain, false); + if (ret) { + nouveau_bo_fini(nvbo); + return ret; + } + + *pnvbo = nvbo; + return 0; +} + +int +nouveau_bo_new_map(struct nouveau_cli *cli, u32 domain, u32 size, struct nouveau_bo **pnvbo) +{ + struct nouveau_bo *nvbo; + int ret; + + ret = nouveau_bo_new_pin(cli, domain, size, &nvbo); + if (ret) + return ret; + + ret = nouveau_bo_map(nvbo); + if (ret) { + nouveau_bo_unpin_del(&nvbo); + return ret; + } + + *pnvbo = nvbo; + return 0; +} + +int +nouveau_bo_new_map_gpu(struct nouveau_cli *cli, u32 domain, u32 size, + struct nouveau_bo **pnvbo, struct nouveau_vma **pvma) +{ + struct nouveau_vmm *vmm = nouveau_cli_vmm(cli); + struct nouveau_bo *nvbo; + int ret; + + ret = nouveau_bo_new_map(cli, domain, size, &nvbo); + if (ret) + return ret; + + ret = nouveau_vma_new(nvbo, vmm, pvma); + if (ret) { + nouveau_bo_unpin_del(&nvbo); + return ret; + } + + *pnvbo = nvbo; + return 0; +} + static void set_placement_range(struct nouveau_bo *nvbo, uint32_t domain) { @@ -923,6 +1000,9 @@ nouveau_bo_move_init(struct nouveau_drm *drm) struct ttm_resource *, struct ttm_resource *); int (*init)(struct nouveau_channel *, u32 handle); } _methods[] = { + { "COPY", 4, 0xcab5, nve0_bo_move_copy, nve0_bo_move_init }, + { "COPY", 4, 0xc9b5, nve0_bo_move_copy, nve0_bo_move_init }, + { "COPY", 4, 0xc8b5, nve0_bo_move_copy, nve0_bo_move_init }, { "COPY", 4, 0xc7b5, nve0_bo_move_copy, nve0_bo_move_init }, { "GRCE", 0, 0xc7b5, nve0_bo_move_copy, nvc0_bo_move_init }, { "COPY", 4, 0xc6b5, nve0_bo_move_copy, nve0_bo_move_init }, @@ -1204,7 +1284,7 @@ retry: fallthrough; /* tiled memory */ case TTM_PL_VRAM: reg->bus.offset = (reg->start << PAGE_SHIFT) + - device->func->resource_addr(device, 1); + device->func->resource_addr(device, NVKM_BAR1_FB); reg->bus.is_iomem = true; /* Some BARs do not support being ioremapped WC */ @@ -1295,7 +1375,7 @@ vm_fault_t nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo) struct nouveau_drm *drm = nouveau_bdev(bo->bdev); struct nouveau_bo *nvbo = nouveau_bo(bo); struct nvkm_device *device = nvxx_device(drm); - u32 mappable = device->func->resource_size(device, 1) >> PAGE_SHIFT; + u32 mappable = device->func->resource_size(device, NVKM_BAR1_FB) >> PAGE_SHIFT; int i, ret; /* as long as the bo isn't in vram, and isn't tiled, we've got diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h index 596a63a50a20..d59fd12268b9 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.h +++ b/drivers/gpu/drm/nouveau/nouveau_bo.h @@ -9,6 +9,7 @@ struct nouveau_channel; struct nouveau_cli; struct nouveau_drm; struct nouveau_fence; +struct nouveau_vma; struct nouveau_bo { struct ttm_buffer_object bo; @@ -89,6 +90,12 @@ void nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo); void nouveau_bo_add_io_reserve_lru(struct ttm_buffer_object *bo); void nouveau_bo_del_io_reserve_lru(struct ttm_buffer_object *bo); +int nouveau_bo_new_pin(struct nouveau_cli *, u32 domain, u32 size, struct nouveau_bo **); +int nouveau_bo_new_map(struct nouveau_cli *, u32 domain, u32 size, struct nouveau_bo **); +int nouveau_bo_new_map_gpu(struct nouveau_cli *, u32 domain, u32 size, + struct nouveau_bo **, struct nouveau_vma **); +void nouveau_bo_unpin_del(struct nouveau_bo **); + /* TODO: submit equivalent to TTM generic API upstream? */ static inline void __iomem * nvbo_kmap_obj_iovirtual(struct nouveau_bo *nvbo) diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c index 1286a664f688..b1e92b1f7a26 100644 --- a/drivers/gpu/drm/nouveau/nouveau_chan.c +++ b/drivers/gpu/drm/nouveau/nouveau_chan.c @@ -103,12 +103,11 @@ nouveau_channel_del(struct nouveau_channel **pchan) nvif_event_dtor(&chan->kill); nvif_object_dtor(&chan->user); nvif_mem_dtor(&chan->mem_userd); + nouveau_vma_del(&chan->sema.vma); + nouveau_bo_unpin_del(&chan->sema.bo); nvif_object_dtor(&chan->push.ctxdma); nouveau_vma_del(&chan->push.vma); - nouveau_bo_unmap(chan->push.buffer); - if (chan->push.buffer && chan->push.buffer->bo.pin_count) - nouveau_bo_unpin(chan->push.buffer); - nouveau_bo_fini(chan->push.buffer); + nouveau_bo_unpin_del(&chan->push.buffer); kfree(chan); } *pchan = NULL; @@ -163,14 +162,7 @@ nouveau_channel_prep(struct nouveau_cli *cli, if (nouveau_vram_pushbuf) target = NOUVEAU_GEM_DOMAIN_VRAM; - ret = nouveau_bo_new(cli, size, 0, target, 0, 0, NULL, NULL, - &chan->push.buffer); - if (ret == 0) { - ret = nouveau_bo_pin(chan->push.buffer, target, false); - if (ret == 0) - ret = nouveau_bo_map(chan->push.buffer); - } - + ret = nouveau_bo_new_map(cli, target, size, &chan->push.buffer); if (ret) { nouveau_channel_del(pchan); return ret; @@ -199,8 +191,10 @@ nouveau_channel_prep(struct nouveau_cli *cli, chan->push.addr = chan->push.vma->addr; - if (device->info.family >= NV_DEVICE_INFO_V0_FERMI) - return 0; + if (device->info.family >= NV_DEVICE_INFO_V0_FERMI) { + return nouveau_bo_new_map_gpu(cli, NOUVEAU_GEM_DOMAIN_GART, PAGE_SIZE, + &chan->sema.bo, &chan->sema.vma); + } args.target = NV_DMA_V0_TARGET_VM; args.access = NV_DMA_V0_ACCESS_VM; @@ -209,13 +203,15 @@ nouveau_channel_prep(struct nouveau_cli *cli, } else if (chan->push.buffer->bo.resource->mem_type == TTM_PL_VRAM) { if (device->info.family == NV_DEVICE_INFO_V0_TNT) { + struct nvkm_device *nvkm_device = nvxx_device(drm); + /* nv04 vram pushbuf hack, retarget to its location in * the framebuffer bar rather than direct vram access.. * nfi why this exists, it came from the -nv ddx. */ args.target = NV_DMA_V0_TARGET_PCI; args.access = NV_DMA_V0_ACCESS_RDWR; - args.start = nvxx_device(drm)->func->resource_addr(nvxx_device(drm), 1); + args.start = nvkm_device->func->resource_addr(nvkm_device, NVKM_BAR1_FB); args.limit = args.start + device->info.ram_user - 1; } else { args.target = NV_DMA_V0_TARGET_VRAM; @@ -253,21 +249,24 @@ nouveau_channel_ctor(struct nouveau_cli *cli, bool priv, u64 runm, struct nouveau_channel **pchan) { const struct nvif_mclass hosts[] = { - { AMPERE_CHANNEL_GPFIFO_B, 0 }, - { AMPERE_CHANNEL_GPFIFO_A, 0 }, - { TURING_CHANNEL_GPFIFO_A, 0 }, - { VOLTA_CHANNEL_GPFIFO_A, 0 }, - { PASCAL_CHANNEL_GPFIFO_A, 0 }, - { MAXWELL_CHANNEL_GPFIFO_A, 0 }, - { KEPLER_CHANNEL_GPFIFO_B, 0 }, - { KEPLER_CHANNEL_GPFIFO_A, 0 }, - { FERMI_CHANNEL_GPFIFO , 0 }, - { G82_CHANNEL_GPFIFO , 0 }, - { NV50_CHANNEL_GPFIFO , 0 }, - { NV40_CHANNEL_DMA , 0 }, - { NV17_CHANNEL_DMA , 0 }, - { NV10_CHANNEL_DMA , 0 }, - { NV03_CHANNEL_DMA , 0 }, + { BLACKWELL_CHANNEL_GPFIFO_B, 0 }, + { BLACKWELL_CHANNEL_GPFIFO_A, 0 }, + { HOPPER_CHANNEL_GPFIFO_A, 0 }, + { AMPERE_CHANNEL_GPFIFO_B, 0 }, + { AMPERE_CHANNEL_GPFIFO_A, 0 }, + { TURING_CHANNEL_GPFIFO_A, 0 }, + { VOLTA_CHANNEL_GPFIFO_A, 0 }, + { PASCAL_CHANNEL_GPFIFO_A, 0 }, + { MAXWELL_CHANNEL_GPFIFO_A, 0 }, + { KEPLER_CHANNEL_GPFIFO_B, 0 }, + { KEPLER_CHANNEL_GPFIFO_A, 0 }, + { FERMI_CHANNEL_GPFIFO , 0 }, + { G82_CHANNEL_GPFIFO , 0 }, + { NV50_CHANNEL_GPFIFO , 0 }, + { NV40_CHANNEL_DMA , 0 }, + { NV17_CHANNEL_DMA , 0 }, + { NV10_CHANNEL_DMA , 0 }, + { NV03_CHANNEL_DMA , 0 }, {} }; DEFINE_RAW_FLEX(struct nvif_chan_v0, args, name, TASK_COMM_LEN + 16); @@ -431,25 +430,33 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart) } /* initialise dma tracking parameters */ - switch (chan->user.oclass) { - case NV03_CHANNEL_DMA: - case NV10_CHANNEL_DMA: - case NV17_CHANNEL_DMA: - case NV40_CHANNEL_DMA: + if (chan->user.oclass < NV50_CHANNEL_GPFIFO) { chan->user_put = 0x40; chan->user_get = 0x44; chan->dma.max = (0x10000 / 4) - 2; - break; - default: - chan->user_put = 0x40; - chan->user_get = 0x44; - chan->user_get_hi = 0x60; - chan->dma.ib_base = 0x10000 / 4; - chan->dma.ib_max = NV50_DMA_IB_MAX; - chan->dma.ib_put = 0; - chan->dma.ib_free = chan->dma.ib_max - chan->dma.ib_put; - chan->dma.max = chan->dma.ib_base; - break; + } else + if (chan->user.oclass < FERMI_CHANNEL_GPFIFO) { + ret = nvif_chan506f_ctor(&chan->chan, chan->userd->map.ptr, + (u8*)chan->push.buffer->kmap.virtual + 0x10000, 0x2000, + chan->push.buffer->kmap.virtual, chan->push.addr, 0x10000); + if (ret) + return ret; + } else + if (chan->user.oclass < VOLTA_CHANNEL_GPFIFO_A) { + ret = nvif_chan906f_ctor(&chan->chan, chan->userd->map.ptr, + (u8*)chan->push.buffer->kmap.virtual + 0x10000, 0x2000, + chan->push.buffer->kmap.virtual, chan->push.addr, 0x10000, + chan->sema.bo->kmap.virtual, chan->sema.vma->addr); + if (ret) + return ret; + } else { + ret = nvif_chanc36f_ctor(&chan->chan, chan->userd->map.ptr, + (u8*)chan->push.buffer->kmap.virtual + 0x10000, 0x2000, + chan->push.buffer->kmap.virtual, chan->push.addr, 0x10000, + chan->sema.bo->kmap.virtual, chan->sema.vma->addr, + &drm->client.device.user, chan->token); + if (ret) + return ret; } chan->dma.put = 0; diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.h b/drivers/gpu/drm/nouveau/nouveau_chan.h index 016f668c0bc1..561877725aac 100644 --- a/drivers/gpu/drm/nouveau/nouveau_chan.h +++ b/drivers/gpu/drm/nouveau/nouveau_chan.h @@ -3,13 +3,11 @@ #define __NOUVEAU_CHAN_H__ #include <nvif/object.h> #include <nvif/event.h> -#include <nvif/push.h> +#include <nvif/chan.h> struct nvif_device; struct nouveau_channel { - struct { - struct nvif_push push; - } chan; + struct nvif_chan chan; struct nouveau_cli *cli; struct nouveau_vmm *vmm; @@ -41,15 +39,15 @@ struct nouveau_channel { int free; int cur; int put; - int ib_base; - int ib_max; - int ib_free; - int ib_put; } dma; - u32 user_get_hi; u32 user_get; u32 user_put; + struct { + struct nouveau_bo *bo; + struct nouveau_vma *vma; + } sema; + struct nvif_object user; struct nvif_object blit; diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index 1b10c6c12f46..63621b1510f6 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -1401,6 +1401,8 @@ nouveau_connector_create(struct drm_device *dev, int index) nv_connector->aux.drm_dev = dev; nv_connector->aux.transfer = nouveau_connector_aux_xfer; nv_connector->aux.name = connector->name; + if (disp->disp.object.oclass >= GB202_DISP) + nv_connector->aux.no_zero_sized = true; drm_dp_aux_init(&nv_connector->aux); break; default: diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c index a1f329ef0641..017a803121d4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dma.c +++ b/drivers/gpu/drm/nouveau/nouveau_dma.c @@ -43,8 +43,6 @@ READ_GET(struct nouveau_channel *chan, uint64_t *prev_get, int *timeout) uint64_t val; val = nvif_rd32(chan->userd, chan->user_get); - if (chan->user_get_hi) - val |= (uint64_t)nvif_rd32(chan->userd, chan->user_get_hi) << 32; /* reset counter as long as GET is still advancing, this is * to avoid misdetecting a GPU lockup if the GPU happens to @@ -68,111 +66,12 @@ READ_GET(struct nouveau_channel *chan, uint64_t *prev_get, int *timeout) return (val - chan->push.addr) >> 2; } -void -nv50_dma_push(struct nouveau_channel *chan, u64 offset, u32 length, - bool no_prefetch) -{ - struct nvif_user *user = &chan->cli->drm->client.device.user; - struct nouveau_bo *pb = chan->push.buffer; - int ip = (chan->dma.ib_put * 2) + chan->dma.ib_base; - - BUG_ON(chan->dma.ib_free < 1); - WARN_ON(length > NV50_DMA_PUSH_MAX_LENGTH); - - nouveau_bo_wr32(pb, ip++, lower_32_bits(offset)); - nouveau_bo_wr32(pb, ip++, upper_32_bits(offset) | length << 8 | - (no_prefetch ? (1 << 31) : 0)); - - chan->dma.ib_put = (chan->dma.ib_put + 1) & chan->dma.ib_max; - - mb(); - /* Flush writes. */ - nouveau_bo_rd32(pb, 0); - - nvif_wr32(chan->userd, 0x8c, chan->dma.ib_put); - if (user->func && user->func->doorbell) - user->func->doorbell(user, chan->token); - chan->dma.ib_free--; -} - -static int -nv50_dma_push_wait(struct nouveau_channel *chan, int count) -{ - uint32_t cnt = 0, prev_get = 0; - - while (chan->dma.ib_free < count) { - uint32_t get = nvif_rd32(chan->userd, 0x88); - if (get != prev_get) { - prev_get = get; - cnt = 0; - } - - if ((++cnt & 0xff) == 0) { - udelay(1); - if (cnt > 100000) - return -EBUSY; - } - - chan->dma.ib_free = get - chan->dma.ib_put; - if (chan->dma.ib_free <= 0) - chan->dma.ib_free += chan->dma.ib_max; - } - - return 0; -} - -static int -nv50_dma_wait(struct nouveau_channel *chan, int slots, int count) -{ - uint64_t prev_get = 0; - int ret, cnt = 0; - - ret = nv50_dma_push_wait(chan, slots + 1); - if (unlikely(ret)) - return ret; - - while (chan->dma.free < count) { - int get = READ_GET(chan, &prev_get, &cnt); - if (unlikely(get < 0)) { - if (get == -EINVAL) - continue; - - return get; - } - - if (get <= chan->dma.cur) { - chan->dma.free = chan->dma.max - chan->dma.cur; - if (chan->dma.free >= count) - break; - - FIRE_RING(chan); - do { - get = READ_GET(chan, &prev_get, &cnt); - if (unlikely(get < 0)) { - if (get == -EINVAL) - continue; - return get; - } - } while (get == 0); - chan->dma.cur = 0; - chan->dma.put = 0; - } - - chan->dma.free = get - chan->dma.cur - 1; - } - - return 0; -} - int -nouveau_dma_wait(struct nouveau_channel *chan, int slots, int size) +nouveau_dma_wait(struct nouveau_channel *chan, int size) { uint64_t prev_get = 0; int cnt = 0, get; - if (chan->dma.ib_max) - return nv50_dma_wait(chan, slots, size); - while (chan->dma.free < size) { get = READ_GET(chan, &prev_get, &cnt); if (unlikely(get == -EBUSY)) diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.h b/drivers/gpu/drm/nouveau/nouveau_dma.h index c52cda82353e..0e27b76d1e1c 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dma.h +++ b/drivers/gpu/drm/nouveau/nouveau_dma.h @@ -30,9 +30,7 @@ #include "nouveau_bo.h" #include "nouveau_chan.h" -int nouveau_dma_wait(struct nouveau_channel *, int slots, int size); -void nv50_dma_push(struct nouveau_channel *, u64 addr, u32 length, - bool no_prefetch); +int nouveau_dma_wait(struct nouveau_channel *, int size); /* * There's a hw race condition where you can't jump to your PUT offset, @@ -67,7 +65,7 @@ RING_SPACE(struct nouveau_channel *chan, int size) { int ret; - ret = nouveau_dma_wait(chan, 1, size); + ret = nouveau_dma_wait(chan, size); if (ret) return ret; @@ -94,12 +92,7 @@ FIRE_RING(struct nouveau_channel *chan) return; chan->accel_done = true; - if (chan->dma.ib_max) { - nv50_dma_push(chan, chan->push.addr + (chan->dma.put << 2), - (chan->dma.cur - chan->dma.put) << 2, false); - } else { - WRITE_PUT(chan->dma.cur); - } + WRITE_PUT(chan->dma.cur); chan->dma.put = chan->dma.cur; } diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index 61d0f411ef84..ca4932a150e3 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -256,20 +256,15 @@ nouveau_dmem_chunk_alloc(struct nouveau_drm *drm, struct page **ppage) chunk->pagemap.ops = &nouveau_dmem_pagemap_ops; chunk->pagemap.owner = drm->dev; - ret = nouveau_bo_new(&drm->client, DMEM_CHUNK_SIZE, 0, - NOUVEAU_GEM_DOMAIN_VRAM, 0, 0, NULL, NULL, - &chunk->bo); + ret = nouveau_bo_new_pin(&drm->client, NOUVEAU_GEM_DOMAIN_VRAM, DMEM_CHUNK_SIZE, + &chunk->bo); if (ret) goto out_release; - ret = nouveau_bo_pin(chunk->bo, NOUVEAU_GEM_DOMAIN_VRAM, false); - if (ret) - goto out_bo_free; - ptr = memremap_pages(&chunk->pagemap, numa_node_id()); if (IS_ERR(ptr)) { ret = PTR_ERR(ptr); - goto out_bo_unpin; + goto out_bo_free; } mutex_lock(&drm->dmem->mutex); @@ -292,10 +287,8 @@ nouveau_dmem_chunk_alloc(struct nouveau_drm *drm, struct page **ppage) return 0; -out_bo_unpin: - nouveau_bo_unpin(chunk->bo); out_bo_free: - nouveau_bo_fini(chunk->bo); + nouveau_bo_unpin_del(&chunk->bo); out_release: release_mem_region(chunk->pagemap.range.start, range_len(&chunk->pagemap.range)); out_free: @@ -426,8 +419,7 @@ nouveau_dmem_fini(struct nouveau_drm *drm) list_for_each_entry_safe(chunk, tmp, &drm->dmem->chunks, list) { nouveau_dmem_evict_chunk(chunk); - nouveau_bo_unpin(chunk->bo); - nouveau_bo_fini(chunk->bo); + nouveau_bo_unpin_del(&chunk->bo); WARN_ON(chunk->callocated); list_del(&chunk->list); memunmap_pages(&chunk->pagemap); diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index c69139701056..0c82a63cd49d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -503,11 +503,16 @@ nouveau_accel_init(struct nouveau_drm *drm) case KEPLER_CHANNEL_GPFIFO_B: case MAXWELL_CHANNEL_GPFIFO_A: case PASCAL_CHANNEL_GPFIFO_A: + ret = nvc0_fence_create(drm); + break; case VOLTA_CHANNEL_GPFIFO_A: case TURING_CHANNEL_GPFIFO_A: case AMPERE_CHANNEL_GPFIFO_A: case AMPERE_CHANNEL_GPFIFO_B: - ret = nvc0_fence_create(drm); + case HOPPER_CHANNEL_GPFIFO_A: + case BLACKWELL_CHANNEL_GPFIFO_A: + case BLACKWELL_CHANNEL_GPFIFO_B: + ret = gv100_fence_create(drm); break; default: break; diff --git a/drivers/gpu/drm/nouveau/nouveau_exec.c b/drivers/gpu/drm/nouveau/nouveau_exec.c index a0b5f1b16e8b..41b7c608c905 100644 --- a/drivers/gpu/drm/nouveau/nouveau_exec.c +++ b/drivers/gpu/drm/nouveau/nouveau_exec.c @@ -10,6 +10,8 @@ #include "nouveau_sched.h" #include "nouveau_uvmm.h" +#include <nvif/class.h> + /** * DOC: Overview * @@ -131,7 +133,7 @@ nouveau_exec_job_run(struct nouveau_job *job) struct nouveau_fence *fence = exec_job->fence; int i, ret; - ret = nouveau_dma_wait(chan, exec_job->push.count + 1, 16); + ret = nvif_chan_gpfifo_wait(&chan->chan, exec_job->push.count + 1, 16); if (ret) { NV_PRINTK(err, job->cli, "nv50cal_space: %d\n", ret); return ERR_PTR(ret); @@ -141,9 +143,11 @@ nouveau_exec_job_run(struct nouveau_job *job) struct drm_nouveau_exec_push *p = &exec_job->push.s[i]; bool no_prefetch = p->flags & DRM_NOUVEAU_EXEC_PUSH_NO_PREFETCH; - nv50_dma_push(chan, p->va, p->va_len, no_prefetch); + nvif_chan_gpfifo_push(&chan->chan, p->va, p->va_len, no_prefetch); } + nvif_chan_gpfifo_post(&chan->chan); + ret = nouveau_fence_emit(fence); if (ret) { nouveau_fence_unref(&exec_job->fence); @@ -375,10 +379,10 @@ nouveau_exec_ioctl_exec(struct drm_device *dev, if (unlikely(atomic_read(&chan->killed))) return nouveau_abi16_put(abi16, -ENODEV); - if (!chan->dma.ib_max) + if (chan->user.oclass < NV50_CHANNEL_GPFIFO) return nouveau_abi16_put(abi16, -ENOSYS); - push_max = nouveau_exec_push_max_from_ib_max(chan->dma.ib_max); + push_max = nouveau_exec_push_max_from_ib_max(chan->chan.gpfifo.max); if (unlikely(req->push_count > push_max)) { NV_PRINTK(err, cli, "pushbuf push count exceeds limit: %d max %d\n", req->push_count, push_max); diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h index 8bc065acfe35..6a983dd9f7b9 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.h +++ b/drivers/gpu/drm/nouveau/nouveau_fence.h @@ -83,6 +83,7 @@ void nv17_fence_resume(struct nouveau_drm *drm); int nv50_fence_create(struct nouveau_drm *); int nv84_fence_create(struct nouveau_drm *); int nvc0_fence_create(struct nouveau_drm *); +int gv100_fence_create(struct nouveau_drm *); struct nv84_fence_chan { struct nouveau_fence_chan base; diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index 67e3c99de73a..690e10fbf0bd 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -850,8 +850,8 @@ revalidate: } } - if (chan->dma.ib_max) { - ret = nouveau_dma_wait(chan, req->nr_push + 1, 16); + if (chan->user.oclass >= NV50_CHANNEL_GPFIFO) { + ret = nvif_chan_gpfifo_wait(&chan->chan, req->nr_push + 1, 16); if (ret) { NV_PRINTK(err, cli, "nv50cal_space: %d\n", ret); goto out; @@ -864,8 +864,10 @@ revalidate: u32 length = push[i].length & ~NOUVEAU_GEM_PUSHBUF_NO_PREFETCH; bool no_prefetch = push[i].length & NOUVEAU_GEM_PUSHBUF_NO_PREFETCH; - nv50_dma_push(chan, addr, length, no_prefetch); + nvif_chan_gpfifo_push(&chan->chan, addr, length, no_prefetch); } + + nvif_chan_gpfifo_post(&chan->chan); } else if (drm->client.device.info.chipset >= 0x25) { ret = PUSH_WAIT(&chan->chan.push, req->nr_push * 2); @@ -958,7 +960,7 @@ out_prevalid: u_free(push); out_next: - if (chan->dma.ib_max) { + if (chan->user.oclass >= NV50_CHANNEL_GPFIFO) { req->suffix0 = 0x00000000; req->suffix1 = 0x00000000; } else diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c index e244927eb5d4..7d2436e5d50d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_ttm.c +++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c @@ -312,8 +312,8 @@ nouveau_ttm_init(struct nouveau_drm *drm) /* VRAM init */ drm->gem.vram_available = drm->client.device.info.ram_user; - arch_io_reserve_memtype_wc(device->func->resource_addr(device, 1), - device->func->resource_size(device, 1)); + arch_io_reserve_memtype_wc(device->func->resource_addr(device, NVKM_BAR1_FB), + device->func->resource_size(device, NVKM_BAR1_FB)); ret = nouveau_ttm_init_vram(drm); if (ret) { @@ -321,8 +321,8 @@ nouveau_ttm_init(struct nouveau_drm *drm) return ret; } - drm->ttm.mtrr = arch_phys_wc_add(device->func->resource_addr(device, 1), - device->func->resource_size(device, 1)); + drm->ttm.mtrr = arch_phys_wc_add(device->func->resource_addr(device, NVKM_BAR1_FB), + device->func->resource_size(device, NVKM_BAR1_FB)); /* GART init */ if (!drm->agp.bridge) { @@ -357,7 +357,7 @@ nouveau_ttm_fini(struct nouveau_drm *drm) arch_phys_wc_del(drm->ttm.mtrr); drm->ttm.mtrr = 0; - arch_io_free_memtype_wc(device->func->resource_addr(device, 1), - device->func->resource_size(device, 1)); + arch_io_free_memtype_wc(device->func->resource_addr(device, NVKM_BAR1_FB), + device->func->resource_size(device, NVKM_BAR1_FB)); } diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c index 8c73f40e3bda..40ee95340814 100644 --- a/drivers/gpu/drm/nouveau/nv10_fence.c +++ b/drivers/gpu/drm/nouveau/nv10_fence.c @@ -85,10 +85,8 @@ void nv10_fence_destroy(struct nouveau_drm *drm) { struct nv10_fence_priv *priv = drm->fence; - nouveau_bo_unmap(priv->bo); - if (priv->bo) - nouveau_bo_unpin(priv->bo); - nouveau_bo_fini(priv->bo); + + nouveau_bo_unpin_del(&priv->bo); drm->fence = NULL; kfree(priv); } diff --git a/drivers/gpu/drm/nouveau/nv17_fence.c b/drivers/gpu/drm/nouveau/nv17_fence.c index d09bfd11369f..1b0c0aa3c305 100644 --- a/drivers/gpu/drm/nouveau/nv17_fence.c +++ b/drivers/gpu/drm/nouveau/nv17_fence.c @@ -130,20 +130,7 @@ nv17_fence_create(struct nouveau_drm *drm) priv->base.context_del = nv10_fence_context_del; spin_lock_init(&priv->lock); - ret = nouveau_bo_new(&drm->client, 4096, 0x1000, - NOUVEAU_GEM_DOMAIN_VRAM, - 0, 0x0000, NULL, NULL, &priv->bo); - if (!ret) { - ret = nouveau_bo_pin(priv->bo, NOUVEAU_GEM_DOMAIN_VRAM, false); - if (!ret) { - ret = nouveau_bo_map(priv->bo); - if (ret) - nouveau_bo_unpin(priv->bo); - } - if (ret) - nouveau_bo_fini(priv->bo); - } - + ret = nouveau_bo_new_map(&drm->client, NOUVEAU_GEM_DOMAIN_VRAM, PAGE_SIZE, &priv->bo); if (ret) { nv10_fence_destroy(drm); return ret; diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c index 62e28dddf87c..e1f0e8adf313 100644 --- a/drivers/gpu/drm/nouveau/nv50_fence.c +++ b/drivers/gpu/drm/nouveau/nv50_fence.c @@ -81,20 +81,7 @@ nv50_fence_create(struct nouveau_drm *drm) priv->base.context_del = nv10_fence_context_del; spin_lock_init(&priv->lock); - ret = nouveau_bo_new(&drm->client, 4096, 0x1000, - NOUVEAU_GEM_DOMAIN_VRAM, - 0, 0x0000, NULL, NULL, &priv->bo); - if (!ret) { - ret = nouveau_bo_pin(priv->bo, NOUVEAU_GEM_DOMAIN_VRAM, false); - if (!ret) { - ret = nouveau_bo_map(priv->bo); - if (ret) - nouveau_bo_unpin(priv->bo); - } - if (ret) - nouveau_bo_fini(priv->bo); - } - + ret = nouveau_bo_new_map(&drm->client, NOUVEAU_GEM_DOMAIN_VRAM, PAGE_SIZE, &priv->bo); if (ret) { nv10_fence_destroy(drm); return ret; diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c index aa7dd0c5d917..1765b2cedaf9 100644 --- a/drivers/gpu/drm/nouveau/nv84_fence.c +++ b/drivers/gpu/drm/nouveau/nv84_fence.c @@ -185,10 +185,8 @@ static void nv84_fence_destroy(struct nouveau_drm *drm) { struct nv84_fence_priv *priv = drm->fence; - nouveau_bo_unmap(priv->bo); - if (priv->bo) - nouveau_bo_unpin(priv->bo); - nouveau_bo_fini(priv->bo); + + nouveau_bo_unpin_del(&priv->bo); drm->fence = NULL; kfree(priv); } @@ -222,19 +220,8 @@ nv84_fence_create(struct nouveau_drm *drm) * will lose CPU/GPU coherency! */ NOUVEAU_GEM_DOMAIN_GART | NOUVEAU_GEM_DOMAIN_COHERENT; - ret = nouveau_bo_new(&drm->client, 16 * drm->chan_total, 0, - domain, 0, 0, NULL, NULL, &priv->bo); - if (ret == 0) { - ret = nouveau_bo_pin(priv->bo, domain, false); - if (ret == 0) { - ret = nouveau_bo_map(priv->bo); - if (ret) - nouveau_bo_unpin(priv->bo); - } - if (ret) - nouveau_bo_fini(priv->bo); - } + ret = nouveau_bo_new_map(&drm->client, domain, 16 * drm->chan_total, &priv->bo); if (ret) nv84_fence_destroy(drm); return ret; diff --git a/drivers/gpu/drm/nouveau/nvif/Kbuild b/drivers/gpu/drm/nouveau/nvif/Kbuild index b7963a39dd91..198889c20ce1 100644 --- a/drivers/gpu/drm/nouveau/nvif/Kbuild +++ b/drivers/gpu/drm/nouveau/nvif/Kbuild @@ -14,6 +14,12 @@ nvif-y += nvif/outp.o nvif-y += nvif/timer.o nvif-y += nvif/vmm.o +# Channel classes +nvif-y += nvif/chan.o +nvif-y += nvif/chan506f.o +nvif-y += nvif/chan906f.o +nvif-y += nvif/chanc36f.o + # Usermode classes nvif-y += nvif/user.o nvif-y += nvif/userc361.o diff --git a/drivers/gpu/drm/nouveau/nvif/chan.c b/drivers/gpu/drm/nouveau/nvif/chan.c new file mode 100644 index 000000000000..baa10227d51a --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvif/chan.c @@ -0,0 +1,156 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <nvif/chan.h> + +static void +nvif_chan_gpfifo_push_kick(struct nvif_push *push) +{ + struct nvif_chan *chan = container_of(push, typeof(*chan), push); + u32 put = push->bgn - (u32 *)chan->push.mem.object.map.ptr; + u32 cnt; + + if (chan->func->gpfifo.post) { + if (push->end - push->cur < chan->func->gpfifo.post_size) + push->end = push->cur + chan->func->gpfifo.post_size; + + WARN_ON(nvif_chan_gpfifo_post(chan)); + } + + cnt = push->cur - push->bgn; + + chan->func->gpfifo.push(chan, true, chan->push.addr + (put << 2), cnt << 2, false); + chan->func->gpfifo.kick(chan); +} + +static int +nvif_chan_gpfifo_push_wait(struct nvif_push *push, u32 push_nr) +{ + struct nvif_chan *chan = container_of(push, typeof(*chan), push); + + return nvif_chan_gpfifo_wait(chan, 1, push_nr); +} + +int +nvif_chan_gpfifo_post(struct nvif_chan *chan) +{ + const u32 *map = chan->push.mem.object.map.ptr; + const u32 pbptr = (chan->push.cur - map) + chan->func->gpfifo.post_size; + const u32 gpptr = (chan->gpfifo.cur + 1) & chan->gpfifo.max; + + return chan->func->gpfifo.post(chan, gpptr, pbptr); +} + +void +nvif_chan_gpfifo_push(struct nvif_chan *chan, u64 addr, u32 size, bool no_prefetch) +{ + chan->func->gpfifo.push(chan, false, addr, size, no_prefetch); +} + +int +nvif_chan_gpfifo_wait(struct nvif_chan *chan, u32 gpfifo_nr, u32 push_nr) +{ + struct nvif_push *push = &chan->push; + int ret = 0, time = 1000000; + + if (gpfifo_nr) { + /* Account for pushbuf space needed by nvif_chan_gpfifo_post(), + * if used after pushing userspace GPFIFO entries. + */ + if (chan->func->gpfifo.post) + push_nr += chan->func->gpfifo.post_size; + } + + /* Account for the GPFIFO entry needed to submit pushbuf. */ + if (push_nr) + gpfifo_nr++; + + /* Wait for space in main push buffer. */ + if (push->cur + push_nr > push->end) { + ret = nvif_chan_dma_wait(chan, push_nr); + if (ret) + return ret; + } + + /* Wait for GPFIFO space. */ + while (chan->gpfifo.free < gpfifo_nr) { + chan->gpfifo.free = chan->func->gpfifo.read_get(chan) - chan->gpfifo.cur - 1; + if (chan->gpfifo.free < 0) + chan->gpfifo.free += chan->gpfifo.max + 1; + + if (chan->gpfifo.free < gpfifo_nr) { + if (!time--) + return -ETIMEDOUT; + udelay(1); + } + } + + return 0; +} + +void +nvif_chan_gpfifo_ctor(const struct nvif_chan_func *func, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, struct nvif_chan *chan) +{ + chan->func = func; + + chan->userd.map.ptr = userd; + + chan->gpfifo.map.ptr = gpfifo; + chan->gpfifo.max = (gpfifo_size >> 3) - 1; + chan->gpfifo.free = chan->gpfifo.max; + + chan->push.mem.object.map.ptr = push; + chan->push.wait = nvif_chan_gpfifo_push_wait; + chan->push.kick = nvif_chan_gpfifo_push_kick; + chan->push.addr = push_addr; + chan->push.hw.max = push_size >> 2; + chan->push.bgn = chan->push.cur = chan->push.end = push; +} + +int +nvif_chan_dma_wait(struct nvif_chan *chan, u32 nr) +{ + struct nvif_push *push = &chan->push; + u32 cur = push->cur - (u32 *)push->mem.object.map.ptr; + u32 free, time = 1000000; + + nr += chan->func->gpfifo.post_size; + + do { + u32 get = chan->func->push.read_get(chan); + + if (get <= cur) { + free = push->hw.max - cur; + if (free >= nr) + break; + + PUSH_KICK(push); + + while (get == 0) { + get = chan->func->push.read_get(chan); + if (get == 0) { + if (!time--) + return -ETIMEDOUT; + udelay(1); + } + } + + cur = 0; + } + + free = get - cur - 1; + + if (free < nr) { + if (!time--) + return -ETIMEDOUT; + udelay(1); + } + } while (free < nr); + + push->bgn = (u32 *)push->mem.object.map.ptr + cur; + push->cur = push->bgn; + push->end = push->bgn + free - chan->func->gpfifo.post_size; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvif/chan506f.c b/drivers/gpu/drm/nouveau/nvif/chan506f.c new file mode 100644 index 000000000000..d3900887c4a7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvif/chan506f.c @@ -0,0 +1,72 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <nvif/chan.h> + +void +nvif_chan506f_gpfifo_kick(struct nvif_chan *chan) +{ + wmb(); + nvif_wr32(&chan->userd, 0x8c, chan->gpfifo.cur); +} + +void +nvif_chan506f_gpfifo_push(struct nvif_chan *chan, bool main, u64 addr, u32 size, bool no_prefetch) +{ + u32 gpptr = chan->gpfifo.cur << 3; + + if (WARN_ON(!chan->gpfifo.free)) + return; + + nvif_wr32(&chan->gpfifo, gpptr + 0, lower_32_bits(addr)); + nvif_wr32(&chan->gpfifo, gpptr + 4, upper_32_bits(addr) | + (main ? 0 : BIT(9)) | + (size >> 2) << 10 | + (no_prefetch ? BIT(31) : 0)); + + chan->gpfifo.cur = (chan->gpfifo.cur + 1) & chan->gpfifo.max; + chan->gpfifo.free--; + if (!chan->gpfifo.free) + chan->push.end = chan->push.cur; +} + +static u32 +nvif_chan506f_gpfifo_read_get(struct nvif_chan *chan) +{ + return nvif_rd32(&chan->userd, 0x88); +} + +static u32 +nvif_chan506f_read_get(struct nvif_chan *chan) +{ + u32 tlgetlo = nvif_rd32(&chan->userd, 0x58); + u32 tlgethi = nvif_rd32(&chan->userd, 0x5c); + struct nvif_push *push = &chan->push; + + /* Update cached GET pointer if TOP_LEVEL_GET is valid. */ + if (tlgethi & BIT(31)) { + u64 tlget = ((u64)(tlgethi & 0xff) << 32) | tlgetlo; + + push->hw.get = (tlget - push->addr) >> 2; + } + + return push->hw.get; +} + +static const struct nvif_chan_func +nvif_chan506f = { + .push.read_get = nvif_chan506f_read_get, + .gpfifo.read_get = nvif_chan506f_gpfifo_read_get, + .gpfifo.push = nvif_chan506f_gpfifo_push, + .gpfifo.kick = nvif_chan506f_gpfifo_kick, +}; + +int +nvif_chan506f_ctor(struct nvif_chan *chan, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size) +{ + nvif_chan_gpfifo_ctor(&nvif_chan506f, userd, gpfifo, gpfifo_size, + push, push_addr, push_size, chan); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvif/chan906f.c b/drivers/gpu/drm/nouveau/nvif/chan906f.c new file mode 100644 index 000000000000..c9cfb85179b0 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvif/chan906f.c @@ -0,0 +1,93 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <nvif/chan.h> +#include <nvif/user.h> +#include <nvif/push906f.h> + +#include <nvhw/class/cl906f.h> + +/* Limits GPFIFO size to 1MiB, and "main" push buffer size to 64KiB. */ +#define NVIF_CHAN906F_PBPTR_BITS 15 +#define NVIF_CHAN906F_PBPTR_MASK ((1 << NVIF_CHAN906F_PBPTR_BITS) - 1) + +#define NVIF_CHAN906F_GPPTR_SHIFT NVIF_CHAN906F_PBPTR_BITS +#define NVIF_CHAN906F_GPPTR_BITS (32 - NVIF_CHAN906F_PBPTR_BITS) +#define NVIF_CHAN906F_GPPTR_MASK ((1 << NVIF_CHAN906F_GPPTR_BITS) - 1) + +#define NVIF_CHAN906F_SEM_RELEASE_SIZE 5 + +static int +nvif_chan906f_sem_release(struct nvif_chan *chan, u64 addr, u32 data) +{ + struct nvif_push *push = &chan->push; + int ret; + + ret = PUSH_WAIT(push, NVIF_CHAN906F_SEM_RELEASE_SIZE); + if (ret) + return ret; + + PUSH_MTHD(push, NV906F, SEMAPHOREA, + NVVAL(NV906F, SEMAPHOREA, OFFSET_UPPER, upper_32_bits(addr)), + + SEMAPHOREB, lower_32_bits(addr), + + SEMAPHOREC, data, + + SEMAPHORED, + NVDEF(NV906F, SEMAPHORED, OPERATION, RELEASE) | + NVDEF(NV906F, SEMAPHORED, RELEASE_WFI, DIS) | + NVDEF(NV906F, SEMAPHORED, RELEASE_SIZE, 16BYTE)); + + return 0; +} + +int +nvif_chan906f_gpfifo_post(struct nvif_chan *chan, u32 gpptr, u32 pbptr) +{ + return chan->func->sem.release(chan, chan->sema.addr, + (gpptr << NVIF_CHAN906F_GPPTR_SHIFT) | pbptr); +} + +u32 +nvif_chan906f_gpfifo_read_get(struct nvif_chan *chan) +{ + return nvif_rd32(&chan->sema, 0) >> NVIF_CHAN906F_GPPTR_SHIFT; +} + +u32 +nvif_chan906f_read_get(struct nvif_chan *chan) +{ + return nvif_rd32(&chan->sema, 0) & NVIF_CHAN906F_PBPTR_MASK; +} + +static const struct nvif_chan_func +nvif_chan906f = { + .push.read_get = nvif_chan906f_read_get, + .gpfifo.read_get = nvif_chan906f_gpfifo_read_get, + .gpfifo.push = nvif_chan506f_gpfifo_push, + .gpfifo.kick = nvif_chan506f_gpfifo_kick, + .gpfifo.post = nvif_chan906f_gpfifo_post, + .gpfifo.post_size = NVIF_CHAN906F_SEM_RELEASE_SIZE, + .sem.release = nvif_chan906f_sem_release, +}; + +int +nvif_chan906f_ctor_(const struct nvif_chan_func *func, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, void *sema, u64 sema_addr, + struct nvif_chan *chan) +{ + nvif_chan_gpfifo_ctor(func, userd, gpfifo, gpfifo_size, push, push_addr, push_size, chan); + chan->sema.map.ptr = sema; + chan->sema.addr = sema_addr; + return 0; +} + +int +nvif_chan906f_ctor(struct nvif_chan *chan, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, void *sema, u64 sema_addr) +{ + return nvif_chan906f_ctor_(&nvif_chan906f, userd, gpfifo, gpfifo_size, + push, push_addr, push_size, sema, sema_addr, chan); +} diff --git a/drivers/gpu/drm/nouveau/nvif/chanc36f.c b/drivers/gpu/drm/nouveau/nvif/chanc36f.c new file mode 100644 index 000000000000..ca02b939c3fd --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvif/chanc36f.c @@ -0,0 +1,77 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <nvif/chan.h> +#include <nvif/user.h> + +#include <nvif/push906f.h> +#include <nvhw/class/clc36f.h> + +static void +nvif_chanc36f_gpfifo_kick(struct nvif_chan *chan) +{ + struct nvif_user *usermode = chan->usermode; + + nvif_wr32(&chan->userd, 0x8c, chan->gpfifo.cur); + + wmb(); /* ensure CPU writes are flushed to BAR1 */ + nvif_rd32(&chan->userd, 0); /* ensure BAR1 writes are flushed to vidmem */ + + usermode->func->doorbell(usermode, chan->doorbell_token); +} + +#define NVIF_CHANC36F_SEM_RELEASE_SIZE 6 + +static int +nvif_chanc36f_sem_release(struct nvif_chan *chan, u64 addr, u32 data) +{ + struct nvif_push *push = &chan->push; + int ret; + + ret = PUSH_WAIT(push, NVIF_CHANC36F_SEM_RELEASE_SIZE); + if (ret) + return ret; + + PUSH_MTHD(push, NVC36F, SEM_ADDR_LO, lower_32_bits(addr), + + SEM_ADDR_HI, upper_32_bits(addr), + + SEM_PAYLOAD_LO, data); + + PUSH_MTHD(push, NVC36F, SEM_EXECUTE, + NVDEF(NVC36F, SEM_EXECUTE, OPERATION, RELEASE) | + NVDEF(NVC36F, SEM_EXECUTE, RELEASE_WFI, DIS) | + NVDEF(NVC36F, SEM_EXECUTE, PAYLOAD_SIZE, 32BIT) | + NVDEF(NVC36F, SEM_EXECUTE, RELEASE_TIMESTAMP, DIS)); + + return 0; +} + +static const struct nvif_chan_func +nvif_chanc36f = { + .push.read_get = nvif_chan906f_read_get, + .gpfifo.read_get = nvif_chan906f_gpfifo_read_get, + .gpfifo.push = nvif_chan506f_gpfifo_push, + .gpfifo.kick = nvif_chanc36f_gpfifo_kick, + .gpfifo.post = nvif_chan906f_gpfifo_post, + .gpfifo.post_size = NVIF_CHANC36F_SEM_RELEASE_SIZE, + .sem.release = nvif_chanc36f_sem_release, +}; + +int +nvif_chanc36f_ctor(struct nvif_chan *chan, void *userd, void *gpfifo, u32 gpfifo_size, + void *push, u64 push_addr, u32 push_size, void *sema, u64 sema_addr, + struct nvif_user *usermode, u32 doorbell_token) +{ + int ret; + + ret = nvif_chan906f_ctor_(&nvif_chanc36f, userd, gpfifo, gpfifo_size, + push, push_addr, push_size, sema, sema_addr, chan); + if (ret) + return ret; + + chan->usermode = usermode; + chan->doorbell_token = doorbell_token; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvif/disp.c b/drivers/gpu/drm/nouveau/nvif/disp.c index 14da22fa3b5b..fa42146252da 100644 --- a/drivers/gpu/drm/nouveau/nvif/disp.c +++ b/drivers/gpu/drm/nouveau/nvif/disp.c @@ -36,6 +36,7 @@ int nvif_disp_ctor(struct nvif_device *device, const char *name, s32 oclass, struct nvif_disp *disp) { static const struct nvif_mclass disps[] = { + { GB202_DISP, 0 }, { AD102_DISP, 0 }, { GA102_DISP, 0 }, { TU102_DISP, 0 }, diff --git a/drivers/gpu/drm/nouveau/nvif/user.c b/drivers/gpu/drm/nouveau/nvif/user.c index b648a5e036af..53f03fa1c9c2 100644 --- a/drivers/gpu/drm/nouveau/nvif/user.c +++ b/drivers/gpu/drm/nouveau/nvif/user.c @@ -41,9 +41,11 @@ nvif_user_ctor(struct nvif_device *device, const char *name) int version; const struct nvif_user_func *func; } users[] = { - { AMPERE_USERMODE_A, -1, &nvif_userc361 }, - { TURING_USERMODE_A, -1, &nvif_userc361 }, - { VOLTA_USERMODE_A, -1, &nvif_userc361 }, + { BLACKWELL_USERMODE_A, -1, &nvif_userc361 }, + { HOPPER_USERMODE_A, -1, &nvif_userc361 }, + { AMPERE_USERMODE_A, -1, &nvif_userc361 }, + { TURING_USERMODE_A, -1, &nvif_userc361 }, + { VOLTA_USERMODE_A, -1, &nvif_userc361 }, {} }; int cid, ret; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/Kbuild index 2e48b0816670..ddcf8782d6b6 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/Kbuild @@ -17,8 +17,6 @@ include $(src)/nvkm/engine/msppp/Kbuild include $(src)/nvkm/engine/msvld/Kbuild include $(src)/nvkm/engine/nvenc/Kbuild include $(src)/nvkm/engine/nvdec/Kbuild -include $(src)/nvkm/engine/nvjpg/Kbuild -include $(src)/nvkm/engine/ofa/Kbuild include $(src)/nvkm/engine/sec/Kbuild include $(src)/nvkm/engine/sec2/Kbuild include $(src)/nvkm/engine/sw/Kbuild diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild index 165d61fc5d6c..9754bac65df7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild @@ -10,5 +10,4 @@ nvkm-y += nvkm/engine/ce/gv100.o nvkm-y += nvkm/engine/ce/tu102.o nvkm-y += nvkm/engine/ce/ga100.o nvkm-y += nvkm/engine/ce/ga102.o - -nvkm-y += nvkm/engine/ce/r535.o +nvkm-y += nvkm/engine/ce/gb202.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga100.c index 9427a592bd16..1c0c60138706 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga100.c @@ -90,7 +90,7 @@ ga100_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine) { if (nvkm_gsp_rm(device->gsp)) - return r535_ce_new(&ga100_ce, device, type, inst, pengine); + return -ENODEV; return nvkm_engine_new_(&ga100_ce, device, type, inst, true, pengine); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga102.c index ce56ede7c2e9..9359c5e7aa3a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/ga102.c @@ -44,7 +44,7 @@ ga102_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine) { if (nvkm_gsp_rm(device->gsp)) - return r535_ce_new(&ga102_ce, device, type, inst, pengine); + return -ENODEV; return nvkm_engine_new_(&ga102_ce, device, type, inst, true, pengine); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gb202.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gb202.c new file mode 100644 index 000000000000..37c3c619c71b --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gb202.c @@ -0,0 +1,16 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gb202/dev_ce.h> + +u32 +gb202_ce_grce_mask(struct nvkm_device *device) +{ + u32 data = nvkm_rd32(device, NV_CE_GRCE_MASK); + + return NVVAL_GET(data, NV_CE, GRCE_MASK, VALUE); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h index 806a76a72249..34fd2657134b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h @@ -16,4 +16,6 @@ int ga100_ce_oneinit(struct nvkm_engine *); int ga100_ce_init(struct nvkm_engine *); int ga100_ce_fini(struct nvkm_engine *, bool); int ga100_ce_nonstall(struct nvkm_engine *); + +u32 gb202_ce_grce_mask(struct nvkm_device *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/r535.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/r535.c deleted file mode 100644 index bd0d435dbbd3..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/r535.c +++ /dev/null @@ -1,108 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <core/object.h> -#include <subdev/gsp.h> -#include <engine/fifo.h> - -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080_notification.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/clc0b5sw.h> - -struct r535_ce_obj { - struct nvkm_object object; - struct nvkm_gsp_object rm; -}; - -static void * -r535_ce_obj_dtor(struct nvkm_object *object) -{ - struct r535_ce_obj *obj = container_of(object, typeof(*obj), object); - - nvkm_gsp_rm_free(&obj->rm); - return obj; -} - -static const struct nvkm_object_func -r535_ce_obj = { - .dtor = r535_ce_obj_dtor, -}; - -static int -r535_ce_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, - struct nvkm_object **pobject) -{ - struct nvkm_chan *chan = nvkm_uchan_chan(oclass->parent); - struct r535_ce_obj *obj; - NVC0B5_ALLOCATION_PARAMETERS *args; - - if (!(obj = kzalloc(sizeof(*obj), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_ce_obj, oclass, &obj->object); - *pobject = &obj->object; - - args = nvkm_gsp_rm_alloc_get(&chan->rm.object, oclass->handle, oclass->base.oclass, - sizeof(*args), &obj->rm); - if (WARN_ON(IS_ERR(args))) - return PTR_ERR(args); - - args->version = 1; - args->engineType = NV2080_ENGINE_TYPE_COPY0 + oclass->engine->subdev.inst; - - return nvkm_gsp_rm_alloc_wr(&obj->rm, args); -} - -static void * -r535_ce_dtor(struct nvkm_engine *engine) -{ - kfree(engine->func); - return engine; -} - -int -r535_ce_new(const struct nvkm_engine_func *hw, struct nvkm_device *device, - enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine) -{ - struct nvkm_engine_func *rm; - int nclass, ret; - - for (nclass = 0; hw->sclass[nclass].oclass; nclass++); - - if (!(rm = kzalloc(sizeof(*rm) + (nclass + 1) * sizeof(rm->sclass[0]), GFP_KERNEL))) - return -ENOMEM; - - rm->dtor = r535_ce_dtor; - for (int i = 0; i < nclass; i++) { - rm->sclass[i].minver = hw->sclass[i].minver; - rm->sclass[i].maxver = hw->sclass[i].maxver; - rm->sclass[i].oclass = hw->sclass[i].oclass; - rm->sclass[i].ctor = r535_ce_obj_ctor; - } - - ret = nvkm_engine_new_(rm, device, type, inst, true, pengine); - if (ret) - kfree(rm); - - return ret; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c index 7c8647dcb349..67d0545cf902 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu102.c @@ -40,7 +40,7 @@ tu102_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine) { if (nvkm_gsp_rm(device->gsp)) - return r535_ce_new(&tu102_ce, device, type, inst, pengine); + return -ENODEV; return nvkm_engine_new_(&tu102_ce, device, type, inst, true, pengine); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c index 9093d89b16f3..3375a59ebf1a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c @@ -2529,9 +2529,6 @@ nv170_chipset = { .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x000003ff, ga100_ce_new }, .fifo = { 0x00000001, ga100_fifo_new }, - .nvdec = { 0x0000001f, ga100_nvdec_new }, - .nvjpg = { 0x00000001, ga100_nvjpg_new }, - .ofa = { 0x00000001, ga100_ofa_new }, }; static const struct nvkm_device_chip @@ -2561,8 +2558,6 @@ nv172_chipset = { .fifo = { 0x00000001, ga102_fifo_new }, .gr = { 0x00000001, ga102_gr_new }, .nvdec = { 0x00000003, ga102_nvdec_new }, - .nvenc = { 0x00000001, ga102_nvenc_new }, - .ofa = { 0x00000001, ga102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2593,8 +2588,6 @@ nv173_chipset = { .fifo = { 0x00000001, ga102_fifo_new }, .gr = { 0x00000001, ga102_gr_new }, .nvdec = { 0x00000003, ga102_nvdec_new }, - .nvenc = { 0x00000001, ga102_nvenc_new }, - .ofa = { 0x00000001, ga102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2625,8 +2618,6 @@ nv174_chipset = { .fifo = { 0x00000001, ga102_fifo_new }, .gr = { 0x00000001, ga102_gr_new }, .nvdec = { 0x00000003, ga102_nvdec_new }, - .nvenc = { 0x00000001, ga102_nvenc_new }, - .ofa = { 0x00000001, ga102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2657,8 +2648,6 @@ nv176_chipset = { .fifo = { 0x00000001, ga102_fifo_new }, .gr = { 0x00000001, ga102_gr_new }, .nvdec = { 0x00000003, ga102_nvdec_new }, - .nvenc = { 0x00000001, ga102_nvenc_new }, - .ofa = { 0x00000001, ga102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2689,12 +2678,26 @@ nv177_chipset = { .fifo = { 0x00000001, ga102_fifo_new }, .gr = { 0x00000001, ga102_gr_new }, .nvdec = { 0x00000003, ga102_nvdec_new }, - .nvenc = { 0x00000001, ga102_nvenc_new }, - .ofa = { 0x00000001, ga102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; static const struct nvkm_device_chip +nv180_chipset = { + .name = "GH100", + .bar = { 0x00000001, tu102_bar_new }, + .fault = { 0x00000001, tu102_fault_new }, + .fb = { 0x00000001, gh100_fb_new }, + .fsp = { 0x00000001, gh100_fsp_new }, + .gsp = { 0x00000001, gh100_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip nv192_chipset = { .name = "AD102", .bar = { 0x00000001, tu102_bar_new }, @@ -2709,14 +2712,9 @@ nv192_chipset = { .timer = { 0x00000001, gk20a_timer_new }, .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x0000001f, ga102_ce_new }, - .disp = { 0x00000001, ad102_disp_new }, + .disp = { 0x00000001, ga102_disp_new }, .dma = { 0x00000001, gv100_dma_new }, .fifo = { 0x00000001, ga102_fifo_new }, - .gr = { 0x00000001, ad102_gr_new }, - .nvdec = { 0x0000000f, ad102_nvdec_new }, - .nvenc = { 0x00000007, ad102_nvenc_new }, - .nvjpg = { 0x0000000f, ad102_nvjpg_new }, - .ofa = { 0x00000001, ad102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2735,14 +2733,9 @@ nv193_chipset = { .timer = { 0x00000001, gk20a_timer_new }, .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x0000001f, ga102_ce_new }, - .disp = { 0x00000001, ad102_disp_new }, + .disp = { 0x00000001, ga102_disp_new }, .dma = { 0x00000001, gv100_dma_new }, .fifo = { 0x00000001, ga102_fifo_new }, - .gr = { 0x00000001, ad102_gr_new }, - .nvdec = { 0x0000000f, ad102_nvdec_new }, - .nvenc = { 0x00000007, ad102_nvenc_new }, - .nvjpg = { 0x0000000f, ad102_nvjpg_new }, - .ofa = { 0x00000001, ad102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2761,14 +2754,9 @@ nv194_chipset = { .timer = { 0x00000001, gk20a_timer_new }, .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x0000001f, ga102_ce_new }, - .disp = { 0x00000001, ad102_disp_new }, + .disp = { 0x00000001, ga102_disp_new }, .dma = { 0x00000001, gv100_dma_new }, .fifo = { 0x00000001, ga102_fifo_new }, - .gr = { 0x00000001, ad102_gr_new }, - .nvdec = { 0x0000000f, ad102_nvdec_new }, - .nvenc = { 0x00000007, ad102_nvenc_new }, - .nvjpg = { 0x0000000f, ad102_nvjpg_new }, - .ofa = { 0x00000001, ad102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2787,14 +2775,9 @@ nv196_chipset = { .timer = { 0x00000001, gk20a_timer_new }, .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x0000001f, ga102_ce_new }, - .disp = { 0x00000001, ad102_disp_new }, + .disp = { 0x00000001, ga102_disp_new }, .dma = { 0x00000001, gv100_dma_new }, .fifo = { 0x00000001, ga102_fifo_new }, - .gr = { 0x00000001, ad102_gr_new }, - .nvdec = { 0x0000000f, ad102_nvdec_new }, - .nvenc = { 0x00000007, ad102_nvenc_new }, - .nvjpg = { 0x0000000f, ad102_nvjpg_new }, - .ofa = { 0x00000001, ad102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; @@ -2813,17 +2796,122 @@ nv197_chipset = { .timer = { 0x00000001, gk20a_timer_new }, .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x0000001f, ga102_ce_new }, - .disp = { 0x00000001, ad102_disp_new }, + .disp = { 0x00000001, ga102_disp_new }, .dma = { 0x00000001, gv100_dma_new }, .fifo = { 0x00000001, ga102_fifo_new }, - .gr = { 0x00000001, ad102_gr_new }, - .nvdec = { 0x0000000f, ad102_nvdec_new }, - .nvenc = { 0x00000007, ad102_nvenc_new }, - .nvjpg = { 0x0000000f, ad102_nvjpg_new }, - .ofa = { 0x00000001, ad102_ofa_new }, .sec2 = { 0x00000001, ga102_sec2_new }, }; +static const struct nvkm_device_chip +nv1a0_chipset = { + .name = "GB100", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb100_fb_new }, + .fsp = { 0x00000001, gb100_fsp_new }, + .gsp = { 0x00000001, gb100_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip +nv1a2_chipset = { + .name = "GB102", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb100_fb_new }, + .fsp = { 0x00000001, gb100_fsp_new }, + .gsp = { 0x00000001, gb100_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip +nv1b2_chipset = { + .name = "GB202", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb202_fb_new }, + .fsp = { 0x00000001, gb202_fsp_new }, + .gsp = { 0x00000001, gb202_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .disp = { 0x00000001, ga102_disp_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip +nv1b3_chipset = { + .name = "GB203", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb202_fb_new }, + .fsp = { 0x00000001, gb202_fsp_new }, + .gsp = { 0x00000001, gb202_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .disp = { 0x00000001, ga102_disp_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip +nv1b5_chipset = { + .name = "GB205", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb202_fb_new }, + .fsp = { 0x00000001, gb202_fsp_new }, + .gsp = { 0x00000001, gb202_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .disp = { 0x00000001, ga102_disp_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip +nv1b6_chipset = { + .name = "GB206", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb202_fb_new }, + .fsp = { 0x00000001, gb202_fsp_new }, + .gsp = { 0x00000001, gb202_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .disp = { 0x00000001, ga102_disp_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + +static const struct nvkm_device_chip +nv1b7_chipset = { + .name = "GB207", + .bar = { 0x00000001, tu102_bar_new }, + .fb = { 0x00000001, gb202_fb_new }, + .fsp = { 0x00000001, gb202_fsp_new }, + .gsp = { 0x00000001, gb202_gsp_new }, + .imem = { 0x00000001, gh100_instmem_new }, + .mmu = { 0x00000001, gh100_mmu_new }, + .pci = { 0x00000001, gh100_pci_new }, + .timer = { 0x00000001, gk20a_timer_new }, + .vfn = { 0x00000001, ga100_vfn_new }, + .disp = { 0x00000001, ga102_disp_new }, + .fifo = { 0x00000001, ga102_fifo_new }, +}; + struct nvkm_subdev * nvkm_device_subdev(struct nvkm_device *device, int type, int inst) { @@ -3065,8 +3153,8 @@ nvkm_device_ctor(const struct nvkm_device_func *func, device->debug = nvkm_dbgopt(device->dbgopt, "device"); INIT_LIST_HEAD(&device->subdev); - mmio_base = device->func->resource_addr(device, 0); - mmio_size = device->func->resource_size(device, 0); + mmio_base = device->func->resource_addr(device, NVKM_BAR0_PRI); + mmio_size = device->func->resource_size(device, NVKM_BAR0_PRI); device->pri = ioremap(mmio_base, mmio_size); if (device->pri == NULL) { @@ -3139,7 +3227,10 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x140: device->card_type = GV100; break; case 0x160: device->card_type = TU100; break; case 0x170: device->card_type = GA100; break; + case 0x180: device->card_type = GH100; break; case 0x190: device->card_type = AD100; break; + case 0x1a0: device->card_type = GB10x; break; + case 0x1b0: device->card_type = GB20x; break; default: break; } @@ -3242,11 +3333,19 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x174: device->chip = &nv174_chipset; break; case 0x176: device->chip = &nv176_chipset; break; case 0x177: device->chip = &nv177_chipset; break; + case 0x180: device->chip = &nv180_chipset; break; case 0x192: device->chip = &nv192_chipset; break; case 0x193: device->chip = &nv193_chipset; break; case 0x194: device->chip = &nv194_chipset; break; case 0x196: device->chip = &nv196_chipset; break; case 0x197: device->chip = &nv197_chipset; break; + case 0x1a0: device->chip = &nv1a0_chipset; break; + case 0x1a2: device->chip = &nv1a2_chipset; break; + case 0x1b2: device->chip = &nv1b2_chipset; break; + case 0x1b3: device->chip = &nv1b3_chipset; break; + case 0x1b5: device->chip = &nv1b5_chipset; break; + case 0x1b6: device->chip = &nv1b6_chipset; break; + case 0x1b7: device->chip = &nv1b7_chipset; break; default: if (nvkm_boolopt(device->cfgopt, "NvEnableUnsupportedChipsets", false)) { switch (device->chipset) { diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c index 3ff6436007fa..8f0261a0d618 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c @@ -1560,18 +1560,42 @@ nvkm_device_pci(struct nvkm_device *device) return container_of(device, struct nvkm_device_pci, device); } +static int +nvkm_device_pci_resource_idx(struct nvkm_device_pci *pdev, enum nvkm_bar_id bar) +{ + int idx = 0; + + if (bar == NVKM_BAR0_PRI) + return idx; + + idx += (pci_resource_flags(pdev->pdev, idx) & IORESOURCE_MEM_64) ? 2 : 1; + if (bar == NVKM_BAR1_FB) + return idx; + + idx += (pci_resource_flags(pdev->pdev, idx) & IORESOURCE_MEM_64) ? 2 : 1; + if (bar == NVKM_BAR2_INST) + return idx; + + WARN_ON(1); + return -1; +} + static resource_size_t -nvkm_device_pci_resource_addr(struct nvkm_device *device, unsigned bar) +nvkm_device_pci_resource_addr(struct nvkm_device *device, enum nvkm_bar_id bar) { struct nvkm_device_pci *pdev = nvkm_device_pci(device); - return pci_resource_start(pdev->pdev, bar); + int idx = nvkm_device_pci_resource_idx(pdev, bar); + + return idx >= 0 ? pci_resource_start(pdev->pdev, idx) : 0; } static resource_size_t -nvkm_device_pci_resource_size(struct nvkm_device *device, unsigned bar) +nvkm_device_pci_resource_size(struct nvkm_device *device, enum nvkm_bar_id bar) { struct nvkm_device_pci *pdev = nvkm_device_pci(device); - return pci_resource_len(pdev->pdev, bar); + int idx = nvkm_device_pci_resource_idx(pdev, bar); + + return idx >= 0 ? pci_resource_len(pdev->pdev, idx) : 0; } static int diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h index e42b18820a95..75ee7506d443 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/priv.h @@ -11,6 +11,7 @@ #include <subdev/devinit.h> #include <subdev/fault.h> #include <subdev/fb.h> +#include <subdev/fsp.h> #include <subdev/fuse.h> #include <subdev/gpio.h> #include <subdev/gsp.h> @@ -43,8 +44,6 @@ #include <engine/msvld.h> #include <engine/nvenc.h> #include <engine/nvdec.h> -#include <engine/nvjpg.h> -#include <engine/ofa.h> #include <engine/sec.h> #include <engine/sec2.h> #include <engine/sw.h> diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c index 78a83f904bbd..114e50ca1827 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c @@ -186,21 +186,31 @@ nvkm_device_tegra(struct nvkm_device *device) } static struct resource * -nvkm_device_tegra_resource(struct nvkm_device *device, unsigned bar) +nvkm_device_tegra_resource(struct nvkm_device *device, enum nvkm_bar_id bar) { struct nvkm_device_tegra *tdev = nvkm_device_tegra(device); - return platform_get_resource(tdev->pdev, IORESOURCE_MEM, bar); + int idx; + + switch (bar) { + case NVKM_BAR0_PRI: idx = 0; break; + case NVKM_BAR1_FB : idx = 1; break; + default: + WARN_ON(1); + return NULL; + } + + return platform_get_resource(tdev->pdev, IORESOURCE_MEM, idx); } static resource_size_t -nvkm_device_tegra_resource_addr(struct nvkm_device *device, unsigned bar) +nvkm_device_tegra_resource_addr(struct nvkm_device *device, enum nvkm_bar_id bar) { struct resource *res = nvkm_device_tegra_resource(device, bar); return res ? res->start : 0; } static resource_size_t -nvkm_device_tegra_resource_size(struct nvkm_device *device, unsigned bar) +nvkm_device_tegra_resource_size(struct nvkm_device *device, enum nvkm_bar_id bar) { struct resource *res = nvkm_device_tegra_resource(device, bar); return res ? resource_size(res) : 0; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c index d7f75b3a43c8..58191b7a0494 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c @@ -148,6 +148,9 @@ nvkm_udevice_info(struct nvkm_udevice *udev, void *data, u32 size) case TU100: args->v0.family = NV_DEVICE_INFO_V0_TURING; break; case GA100: args->v0.family = NV_DEVICE_INFO_V0_AMPERE; break; case AD100: args->v0.family = NV_DEVICE_INFO_V0_ADA; break; + case GH100: args->v0.family = NV_DEVICE_INFO_V0_HOPPER; break; + case GB10x: args->v0.family = NV_DEVICE_INFO_V0_BLACKWELL; break; + case GB20x: args->v0.family = NV_DEVICE_INFO_V0_BLACKWELL; break; default: args->v0.family = 0; break; @@ -209,8 +212,8 @@ nvkm_udevice_map(struct nvkm_object *object, void *argv, u32 argc, struct nvkm_udevice *udev = nvkm_udevice(object); struct nvkm_device *device = udev->device; *type = NVKM_OBJECT_MAP_IO; - *addr = device->func->resource_addr(device, 0); - *size = device->func->resource_size(device, 0); + *addr = device->func->resource_addr(device, NVKM_BAR0_PRI); + *size = device->func->resource_size(device, NVKM_BAR0_PRI); return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild index e346e924fee8..e1aecd3fe96c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild @@ -27,9 +27,6 @@ nvkm-y += nvkm/engine/disp/gp102.o nvkm-y += nvkm/engine/disp/gv100.o nvkm-y += nvkm/engine/disp/tu102.o nvkm-y += nvkm/engine/disp/ga102.o -nvkm-y += nvkm/engine/disp/ad102.o - -nvkm-y += nvkm/engine/disp/r535.o nvkm-y += nvkm/engine/disp/udisp.o nvkm-y += nvkm/engine/disp/uconn.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ad102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ad102.c deleted file mode 100644 index 7f300a79aa29..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ad102.c +++ /dev/null @@ -1,52 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" -#include "chan.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_disp_func -ad102_disp = { - .uevent = &gv100_disp_chan_uevent, - .ramht_size = 0x2000, - .root = { 0, 0,AD102_DISP }, - .user = { - {{-1,-1,GV100_DISP_CAPS }, gv100_disp_caps_new }, - {{ 0, 0,GA102_DISP_CURSOR }, nvkm_disp_chan_new, &gv100_disp_curs }, - {{ 0, 0,GA102_DISP_WINDOW_IMM_CHANNEL_DMA}, nvkm_disp_wndw_new, &gv100_disp_wimm }, - {{ 0, 0,AD102_DISP_CORE_CHANNEL_DMA }, nvkm_disp_core_new, &gv100_disp_core }, - {{ 0, 0,GA102_DISP_WINDOW_CHANNEL_DMA }, nvkm_disp_wndw_new, &gv100_disp_wndw }, - {} - }, -}; - -int -ad102_disp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_disp **pdisp) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_disp_new(&ad102_disp, device, type, inst, pdisp); - - return -ENODEV; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/chan.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/chan.c index 4e43ee383c34..9b84e357d354 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/chan.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/chan.c @@ -49,7 +49,7 @@ nvkm_disp_chan_map(struct nvkm_object *object, void *argv, u32 argc, { struct nvkm_disp_chan *chan = nvkm_disp_chan(object); struct nvkm_device *device = chan->disp->engine.subdev.device; - const u64 base = device->func->resource_addr(device, 0); + const u64 base = device->func->resource_addr(device, NVKM_BAR0_PRI); *type = NVKM_OBJECT_MAP_IO; *addr = base + chan->func->user(chan, size); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c index cfa3698d3a2f..614921166fba 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c @@ -805,7 +805,7 @@ gv100_disp_caps_map(struct nvkm_object *object, void *argv, u32 argc, struct gv100_disp_caps *caps = gv100_disp_caps(object); struct nvkm_device *device = caps->disp->engine.subdev.device; *type = NVKM_OBJECT_MAP_IO; - *addr = 0x640000 + device->func->resource_addr(device, 0); + *addr = 0x640000 + device->func->resource_addr(device, NVKM_BAR0_PRI); *size = 0x1000; return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild index aff92848abfe..376e9c3bcb1a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild @@ -25,8 +25,7 @@ nvkm-y += nvkm/engine/fifo/gv100.o nvkm-y += nvkm/engine/fifo/tu102.o nvkm-y += nvkm/engine/fifo/ga100.o nvkm-y += nvkm/engine/fifo/ga102.o - -nvkm-y += nvkm/engine/fifo/r535.o +nvkm-y += nvkm/engine/fifo/gb202.o nvkm-y += nvkm/engine/fifo/ucgrp.o nvkm-y += nvkm/engine/fifo/uchan.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c index 22443fe4a39f..fdffa0391b31 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c @@ -303,7 +303,7 @@ nvkm_fifo_oneinit(struct nvkm_engine *engine) } /* Allocate USERD + BAR1 polling area. */ - if (fifo->func->chan.func->userd->bar == 1) { + if (fifo->func->chan.func->userd->bar == NVKM_BAR1_FB) { struct nvkm_vmm *bar1 = nvkm_bar_bar1_vmm(device); ret = nvkm_memory_new(device, NVKM_MEM_TARGET_INST, fifo->chid->nr * @@ -349,8 +349,6 @@ nvkm_fifo_dtor(struct nvkm_engine *engine) nvkm_chid_unref(&fifo->cgid); nvkm_chid_unref(&fifo->chid); - mutex_destroy(&fifo->userd.mutex); - nvkm_event_fini(&fifo->nonstall.event); mutex_destroy(&fifo->mutex); @@ -391,8 +389,5 @@ nvkm_fifo_new_(const struct nvkm_fifo_func *func, struct nvkm_device *device, spin_lock_init(&fifo->lock); mutex_init(&fifo->mutex); - INIT_LIST_HEAD(&fifo->userd.list); - mutex_init(&fifo->userd.mutex); - return nvkm_engine_ctor(&nvkm_fifo, device, type, inst, true, &fifo->engine); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c index f5cd7f7c48b4..e5bbd8563007 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.c @@ -275,11 +275,7 @@ nvkm_chan_del(struct nvkm_chan **pchan) nvkm_gpuobj_del(&chan->ramfc); if (chan->cgrp) { - if (!chan->func->id_put) - nvkm_chid_put(chan->cgrp->runl->chid, chan->id, &chan->cgrp->lock); - else - chan->func->id_put(chan); - + nvkm_chid_put(chan->cgrp->runl->chid, chan->id, &chan->cgrp->lock); nvkm_cgrp_unref(&chan->cgrp); } @@ -359,14 +355,14 @@ nvkm_chan_new_(const struct nvkm_chan_func *func, struct nvkm_runl *runl, int ru /* Validate arguments against class requirements. */ if ((runq && runq >= runl->func->runqs) || (!func->inst->vmm != !vmm) || - ((func->userd->bar < 0) == !userd) || + (!func->userd->bar == !userd) || (!func->ramfc->ctxdma != !dmaobj) || ((func->ramfc->devm < devm) && devm != BIT(0)) || (!func->ramfc->priv && priv)) { RUNL_DEBUG(runl, "args runq:%d:%d vmm:%d:%p userd:%d:%p " "push:%d:%p devm:%08x:%08x priv:%d:%d", runl->func->runqs, runq, func->inst->vmm, vmm, - func->userd->bar < 0, userd, func->ramfc->ctxdma, dmaobj, + func->userd->bar, userd, func->ramfc->ctxdma, dmaobj, func->ramfc->devm, devm, func->ramfc->priv, priv); return -EINVAL; } @@ -441,30 +437,26 @@ nvkm_chan_new_(const struct nvkm_chan_func *func, struct nvkm_runl *runl, int ru } /* Allocate channel ID. */ - if (!chan->func->id_get) { - chan->id = nvkm_chid_get(runl->chid, chan); - if (chan->id >= 0) { - if (func->userd->bar < 0) { - if (ouserd + chan->func->userd->size >= - nvkm_memory_size(userd)) { - RUNL_DEBUG(runl, "ouserd %llx", ouserd); - return -EINVAL; - } - - ret = nvkm_memory_kmap(userd, &chan->userd.mem); - if (ret) { - RUNL_DEBUG(runl, "userd %d", ret); - return ret; - } - - chan->userd.base = ouserd; - } else { - chan->userd.mem = nvkm_memory_ref(fifo->userd.mem); - chan->userd.base = chan->id * chan->func->userd->size; + chan->id = nvkm_chid_get(runl->chid, chan); + if (chan->id >= 0) { + if (!func->userd->bar) { + if (ouserd + chan->func->userd->size >= + nvkm_memory_size(userd)) { + RUNL_DEBUG(runl, "ouserd %llx", ouserd); + return -EINVAL; + } + + ret = nvkm_memory_kmap(userd, &chan->userd.mem); + if (ret) { + RUNL_DEBUG(runl, "userd %d", ret); + return ret; } + + chan->userd.base = ouserd; + } else { + chan->userd.mem = nvkm_memory_ref(fifo->userd.mem); + chan->userd.base = chan->id * chan->func->userd->size; } - } else { - chan->id = chan->func->id_get(chan, userd, ouserd); } if (chan->id < 0) { diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h index 013682a709d5..445db5dfd1e4 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h @@ -17,9 +17,6 @@ struct nvkm_cctx { }; struct nvkm_chan_func { - int (*id_get)(struct nvkm_chan *, struct nvkm_memory *userd, u64 ouserd); - void (*id_put)(struct nvkm_chan *); - const struct nvkm_chan_func_inst { u32 size; bool zero; @@ -27,7 +24,7 @@ struct nvkm_chan_func { } *inst; const struct nvkm_chan_func_userd { - int bar; + enum nvkm_bar_id bar; u32 base; u32 size; void (*clear)(struct nvkm_chan *); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gb202.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gb202.c new file mode 100644 index 000000000000..b469e8afeb0b --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gb202.c @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" +#include "cgrp.h" +#include "chan.h" +#include "runl.h" + +u32 +gb202_chan_doorbell_handle(struct nvkm_chan *chan) +{ + return BIT(30) | (chan->cgrp->runl->id << 16) | chan->id; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c index 6c94451d0faa..e4a4fad2eafc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c @@ -133,7 +133,7 @@ gf100_chan_userd_clear(struct nvkm_chan *chan) static const struct nvkm_chan_func_userd gf100_chan_userd = { - .bar = 1, + .bar = NVKM_BAR1_FB, .size = 0x1000, .clear = gf100_chan_userd_clear, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c index d8a4d773a58c..5655eda52a7b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c @@ -113,7 +113,7 @@ gk104_chan_ramfc = { const struct nvkm_chan_func_userd gk104_chan_userd = { - .bar = 1, + .bar = NVKM_BAR1_FB, .size = 0x200, .clear = gf100_chan_userd_clear, }; @@ -745,7 +745,7 @@ gk104_fifo_init(struct nvkm_fifo *fifo) { struct nvkm_device *device = fifo->engine.subdev.device; - if (fifo->func->chan.func->userd->bar == 1) + if (fifo->func->chan.func->userd->bar == NVKM_BAR1_FB) nvkm_wr32(device, 0x002254, 0x10000000 | fifo->userd.bar1->addr >> 12); nvkm_wr32(device, 0x002100, 0xffffffff); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c index 33066c8cdc64..d7f046c03cfd 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c @@ -70,7 +70,6 @@ gv100_chan_ramfc = { const struct nvkm_chan_func_userd gv100_chan_userd = { - .bar = -1, .size = 0x200, .clear = gf100_chan_userd_clear, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c index 674faf002b20..c4b8e567d86f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv04.c @@ -154,7 +154,7 @@ nv04_chan_ramfc = { const struct nvkm_chan_func_userd nv04_chan_userd = { - .bar = 0, + .bar = NVKM_BAR0_PRI, .base = 0x800000, .size = 0x010000, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c index e50a94b6d7f8..084ca5561ee1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv40.c @@ -93,7 +93,7 @@ nv40_chan_ramfc = { static const struct nvkm_chan_func_userd nv40_chan_userd = { - .bar = 0, + .bar = NVKM_BAR0_PRI, .base = 0xc00000, .size = 0x001000, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c index 954b5f3a7d57..7bf77661157d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/nv50.c @@ -124,7 +124,7 @@ nv50_chan_ramfc = { const struct nvkm_chan_func_userd nv50_chan_userd = { - .bar = 0, + .bar = NVKM_BAR0_PRI, .base = 0xc00000, .size = 0x002000, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h index a0f3277605a5..5e81ae195329 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h @@ -6,6 +6,7 @@ #include <core/enum.h> struct nvkm_cctx; struct nvkm_cgrp; +struct nvkm_chan; struct nvkm_engn; struct nvkm_memory; struct nvkm_runl; @@ -195,6 +196,7 @@ extern const struct nvkm_chan_func_ramfc gv100_chan_ramfc; void tu102_fifo_intr_ctxsw_timeout_info(struct nvkm_engn *, u32 info); extern const struct nvkm_fifo_func_mmu_fault tu102_fifo_mmu_fault; +u32 tu102_chan_doorbell_handle(struct nvkm_chan *); int ga100_fifo_runl_ctor(struct nvkm_fifo *); int ga100_fifo_nonstall_ctor(struct nvkm_fifo *); @@ -206,6 +208,8 @@ extern const struct nvkm_engn_func ga100_engn_ce; extern const struct nvkm_cgrp_func ga100_cgrp; extern const struct nvkm_chan_func ga100_chan; +u32 gb202_chan_doorbell_handle(struct nvkm_chan *); + int nvkm_uchan_new(struct nvkm_fifo *, struct nvkm_cgrp *, const struct nvkm_oclass *, void *argv, u32 argc, struct nvkm_object **); int nvkm_ucgrp_new(struct nvkm_fifo *, const struct nvkm_oclass *, void *argv, u32 argc, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c index 1d39a6840a40..c5a03298e88c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu102.c @@ -31,7 +31,7 @@ #include <nvif/class.h> -static u32 +u32 tu102_chan_doorbell_handle(struct nvkm_chan *chan) { return (chan->cgrp->runl->id << 16) | chan->id; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/uchan.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/uchan.c index 9e56bcc166ed..52420a1edca5 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/uchan.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/uchan.c @@ -258,7 +258,7 @@ nvkm_uchan_map(struct nvkm_object *object, void *argv, u32 argc, struct nvkm_chan *chan = nvkm_uchan(object)->chan; struct nvkm_device *device = chan->cgrp->runl->fifo->engine.subdev.device; - if (chan->func->userd->bar < 0) + if (!chan->func->userd->bar) return -ENOSYS; *type = NVKM_OBJECT_MAP_IO; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild index 1555f8c40b4f..b5418f05ccd8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild @@ -41,9 +41,6 @@ nvkm-y += nvkm/engine/gr/gp10b.o nvkm-y += nvkm/engine/gr/gv100.o nvkm-y += nvkm/engine/gr/tu102.o nvkm-y += nvkm/engine/gr/ga102.o -nvkm-y += nvkm/engine/gr/ad102.o - -nvkm-y += nvkm/engine/gr/r535.o nvkm-y += nvkm/engine/gr/ctxnv40.o nvkm-y += nvkm/engine/gr/ctxnv50.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ga102.c index d285c597aff9..2b51f1d0c281 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ga102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ga102.c @@ -352,7 +352,7 @@ int ga102_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr) { if (nvkm_gsp_rm(device->gsp)) - return r535_gr_new(&ga102_gr, device, type, inst, pgr); + return -ENODEV; return gf100_gr_new_(ga102_gr_fwif, device, type, inst, pgr); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h index b0e0c9305034..54f686ba39ac 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h @@ -445,6 +445,4 @@ void gp108_gr_acr_bld_patch(struct nvkm_acr *, u32, s64); int gf100_gr_new_(const struct gf100_gr_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gr **); -int r535_gr_new(const struct gf100_gr_func *, struct nvkm_device *, enum nvkm_subdev_type, int, - struct nvkm_gr **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c index 02a8c62a0a32..13407fafe947 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv20.c @@ -297,7 +297,7 @@ nv20_gr_init(struct nvkm_gr *base) nvkm_wr32(device, NV10_PGRAPH_SURFACE, tmp); /* begin RAM config */ - vramsz = device->func->resource_size(device, 1) - 1; + vramsz = device->func->resource_size(device, NVKM_BAR1_FB) - 1; nvkm_wr32(device, 0x4009A4, nvkm_rd32(device, 0x100200)); nvkm_wr32(device, 0x4009A8, nvkm_rd32(device, 0x100204)); nvkm_wr32(device, NV10_PGRAPH_RDI_INDEX, 0x00EA0000); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c index a5e1f02791b4..b609b0150ba1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/nv40.c @@ -386,7 +386,7 @@ nv40_gr_init(struct nvkm_gr *base) } /* begin RAM config */ - vramsz = device->func->resource_size(device, 1) - 1; + vramsz = device->func->resource_size(device, NVKM_BAR1_FB) - 1; switch (device->chipset) { case 0x40: nvkm_wr32(device, 0x4009A4, nvkm_rd32(device, 0x100200)); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/r535.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/r535.c deleted file mode 100644 index f4bed3eb1ec2..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/r535.c +++ /dev/null @@ -1,508 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "gf100.h" - -#include <core/memory.h> -#include <subdev/gsp.h> -#include <subdev/mmu/vmm.h> -#include <engine/fifo/priv.h> - -#include <nvif/if900d.h> - -#include <nvhw/drf.h> - -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/alloc/alloc_channel.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080fifo.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h> -#include <nvrm/535.113.01/nvidia/generated/g_kernel_channel_nvoc.h> - -#define r535_gr(p) container_of((p), struct r535_gr, base) - -#define R515_GR_MAX_CTXBUFS 9 - -struct r535_gr { - struct nvkm_gr base; - - struct { - u16 bufferId; - u32 size; - u8 page; - u8 align; - bool global; - bool init; - bool ro; - } ctxbuf[R515_GR_MAX_CTXBUFS]; - int ctxbuf_nr; - - struct nvkm_memory *ctxbuf_mem[R515_GR_MAX_CTXBUFS]; -}; - -struct r535_gr_chan { - struct nvkm_object object; - struct r535_gr *gr; - - struct nvkm_vmm *vmm; - struct nvkm_chan *chan; - - struct nvkm_memory *mem[R515_GR_MAX_CTXBUFS]; - struct nvkm_vma *vma[R515_GR_MAX_CTXBUFS]; -}; - -struct r535_gr_obj { - struct nvkm_object object; - struct nvkm_gsp_object rm; -}; - -static void * -r535_gr_obj_dtor(struct nvkm_object *object) -{ - struct r535_gr_obj *obj = container_of(object, typeof(*obj), object); - - nvkm_gsp_rm_free(&obj->rm); - return obj; -} - -static const struct nvkm_object_func -r535_gr_obj = { - .dtor = r535_gr_obj_dtor, -}; - -static int -r535_gr_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, - struct nvkm_object **pobject) -{ - struct r535_gr_chan *chan = container_of(oclass->parent, typeof(*chan), object); - struct r535_gr_obj *obj; - - if (!(obj = kzalloc(sizeof(*obj), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_gr_obj, oclass, &obj->object); - *pobject = &obj->object; - - return nvkm_gsp_rm_alloc(&chan->chan->rm.object, oclass->handle, oclass->base.oclass, 0, - &obj->rm); -} - -static void * -r535_gr_chan_dtor(struct nvkm_object *object) -{ - struct r535_gr_chan *grc = container_of(object, typeof(*grc), object); - struct r535_gr *gr = grc->gr; - - for (int i = 0; i < gr->ctxbuf_nr; i++) { - nvkm_vmm_put(grc->vmm, &grc->vma[i]); - nvkm_memory_unref(&grc->mem[i]); - } - - nvkm_vmm_unref(&grc->vmm); - return grc; -} - -static const struct nvkm_object_func -r535_gr_chan = { - .dtor = r535_gr_chan_dtor, -}; - -static int -r535_gr_promote_ctx(struct r535_gr *gr, bool golden, struct nvkm_vmm *vmm, - struct nvkm_memory **pmem, struct nvkm_vma **pvma, - struct nvkm_gsp_object *chan) -{ - struct nvkm_subdev *subdev = &gr->base.engine.subdev; - struct nvkm_device *device = subdev->device; - NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS *ctrl; - - ctrl = nvkm_gsp_rm_ctrl_get(&vmm->rm.device.subdevice, - NV2080_CTRL_CMD_GPU_PROMOTE_CTX, sizeof(*ctrl)); - if (WARN_ON(IS_ERR(ctrl))) - return PTR_ERR(ctrl); - - ctrl->engineType = 1; - ctrl->hChanClient = vmm->rm.client.object.handle; - ctrl->hObject = chan->handle; - - for (int i = 0; i < gr->ctxbuf_nr; i++) { - NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY *entry = - &ctrl->promoteEntry[ctrl->entryCount]; - const bool alloc = golden || !gr->ctxbuf[i].global; - int ret; - - entry->bufferId = gr->ctxbuf[i].bufferId; - entry->bInitialize = gr->ctxbuf[i].init && alloc; - - if (alloc) { - ret = nvkm_memory_new(device, gr->ctxbuf[i].init ? - NVKM_MEM_TARGET_INST : NVKM_MEM_TARGET_INST_SR_LOST, - gr->ctxbuf[i].size, 1 << gr->ctxbuf[i].page, - gr->ctxbuf[i].init, &pmem[i]); - if (WARN_ON(ret)) - return ret; - - if (gr->ctxbuf[i].bufferId == - NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PRIV_ACCESS_MAP) - entry->bNonmapped = 1; - } else { - if (gr->ctxbuf[i].bufferId == - NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_UNRESTRICTED_PRIV_ACCESS_MAP) - continue; - - pmem[i] = nvkm_memory_ref(gr->ctxbuf_mem[i]); - } - - if (!entry->bNonmapped) { - struct gf100_vmm_map_v0 args = { - .priv = 1, - .ro = gr->ctxbuf[i].ro, - }; - - mutex_lock(&vmm->mutex.vmm); - ret = nvkm_vmm_get_locked(vmm, false, true, false, 0, gr->ctxbuf[i].align, - nvkm_memory_size(pmem[i]), &pvma[i]); - mutex_unlock(&vmm->mutex.vmm); - if (ret) - return ret; - - ret = nvkm_memory_map(pmem[i], 0, vmm, pvma[i], &args, sizeof(args)); - if (ret) - return ret; - - entry->gpuVirtAddr = pvma[i]->addr; - } - - if (entry->bInitialize) { - entry->gpuPhysAddr = nvkm_memory_addr(pmem[i]); - entry->size = gr->ctxbuf[i].size; - entry->physAttr = 4; - } - - nvkm_debug(subdev, - "promote %02d: pa %016llx/%08x sz %016llx va %016llx init:%d nm:%d\n", - entry->bufferId, entry->gpuPhysAddr, entry->physAttr, entry->size, - entry->gpuVirtAddr, entry->bInitialize, entry->bNonmapped); - - ctrl->entryCount++; - } - - return nvkm_gsp_rm_ctrl_wr(&vmm->rm.device.subdevice, ctrl); -} - -static int -r535_gr_chan_new(struct nvkm_gr *base, struct nvkm_chan *chan, const struct nvkm_oclass *oclass, - struct nvkm_object **pobject) -{ - struct r535_gr *gr = r535_gr(base); - struct r535_gr_chan *grc; - int ret; - - if (!(grc = kzalloc(sizeof(*grc), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_gr_chan, oclass, &grc->object); - grc->gr = gr; - grc->vmm = nvkm_vmm_ref(chan->vmm); - grc->chan = chan; - *pobject = &grc->object; - - ret = r535_gr_promote_ctx(gr, false, grc->vmm, grc->mem, grc->vma, &chan->rm.object); - if (ret) - return ret; - - return 0; -} - -static u64 -r535_gr_units(struct nvkm_gr *gr) -{ - struct nvkm_gsp *gsp = gr->engine.subdev.device->gsp; - - return (gsp->gr.tpcs << 8) | gsp->gr.gpcs; -} - -static int -r535_gr_oneinit(struct nvkm_gr *base) -{ - NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS *info; - struct r535_gr *gr = container_of(base, typeof(*gr), base); - struct nvkm_subdev *subdev = &gr->base.engine.subdev; - struct nvkm_device *device = subdev->device; - struct nvkm_gsp *gsp = device->gsp; - struct nvkm_mmu *mmu = device->mmu; - struct { - struct nvkm_memory *inst; - struct nvkm_vmm *vmm; - struct nvkm_gsp_object chan; - struct nvkm_vma *vma[R515_GR_MAX_CTXBUFS]; - } golden = {}; - int ret; - - /* Allocate a channel to use for golden context init. */ - ret = nvkm_memory_new(device, NVKM_MEM_TARGET_INST, 0x12000, 0, true, &golden.inst); - if (ret) - goto done; - - ret = nvkm_vmm_new(device, 0x1000, 0, NULL, 0, NULL, "grGoldenVmm", &golden.vmm); - if (ret) - goto done; - - ret = mmu->func->promote_vmm(golden.vmm); - if (ret) - goto done; - - { - NV_CHANNELGPFIFO_ALLOCATION_PARAMETERS *args; - - args = nvkm_gsp_rm_alloc_get(&golden.vmm->rm.device.object, 0xf1f00000, - device->fifo->func->chan.user.oclass, - sizeof(*args), &golden.chan); - if (IS_ERR(args)) { - ret = PTR_ERR(args); - goto done; - } - - args->gpFifoOffset = 0; - args->gpFifoEntries = 0x1000 / 8; - args->flags = - NVDEF(NVOS04, FLAGS, CHANNEL_TYPE, PHYSICAL) | - NVDEF(NVOS04, FLAGS, VPR, FALSE) | - NVDEF(NVOS04, FLAGS, CHANNEL_SKIP_MAP_REFCOUNTING, FALSE) | - NVVAL(NVOS04, FLAGS, GROUP_CHANNEL_RUNQUEUE, 0) | - NVDEF(NVOS04, FLAGS, PRIVILEGED_CHANNEL, TRUE) | - NVDEF(NVOS04, FLAGS, DELAY_CHANNEL_SCHEDULING, FALSE) | - NVDEF(NVOS04, FLAGS, CHANNEL_DENY_PHYSICAL_MODE_CE, FALSE) | - NVVAL(NVOS04, FLAGS, CHANNEL_USERD_INDEX_VALUE, 0) | - NVDEF(NVOS04, FLAGS, CHANNEL_USERD_INDEX_FIXED, FALSE) | - NVVAL(NVOS04, FLAGS, CHANNEL_USERD_INDEX_PAGE_VALUE, 0) | - NVDEF(NVOS04, FLAGS, CHANNEL_USERD_INDEX_PAGE_FIXED, TRUE) | - NVDEF(NVOS04, FLAGS, CHANNEL_DENY_AUTH_LEVEL_PRIV, FALSE) | - NVDEF(NVOS04, FLAGS, CHANNEL_SKIP_SCRUBBER, FALSE) | - NVDEF(NVOS04, FLAGS, CHANNEL_CLIENT_MAP_FIFO, FALSE) | - NVDEF(NVOS04, FLAGS, SET_EVICT_LAST_CE_PREFETCH_CHANNEL, FALSE) | - NVDEF(NVOS04, FLAGS, CHANNEL_VGPU_PLUGIN_CONTEXT, FALSE) | - NVDEF(NVOS04, FLAGS, CHANNEL_PBDMA_ACQUIRE_TIMEOUT, FALSE) | - NVDEF(NVOS04, FLAGS, GROUP_CHANNEL_THREAD, DEFAULT) | - NVDEF(NVOS04, FLAGS, MAP_CHANNEL, FALSE) | - NVDEF(NVOS04, FLAGS, SKIP_CTXBUFFER_ALLOC, FALSE); - args->hVASpace = golden.vmm->rm.object.handle; - args->engineType = 1; - args->instanceMem.base = nvkm_memory_addr(golden.inst); - args->instanceMem.size = 0x1000; - args->instanceMem.addressSpace = 2; - args->instanceMem.cacheAttrib = 1; - args->ramfcMem.base = nvkm_memory_addr(golden.inst); - args->ramfcMem.size = 0x200; - args->ramfcMem.addressSpace = 2; - args->ramfcMem.cacheAttrib = 1; - args->userdMem.base = nvkm_memory_addr(golden.inst) + 0x1000; - args->userdMem.size = 0x200; - args->userdMem.addressSpace = 2; - args->userdMem.cacheAttrib = 1; - args->mthdbufMem.base = nvkm_memory_addr(golden.inst) + 0x2000; - args->mthdbufMem.size = 0x5000; - args->mthdbufMem.addressSpace = 2; - args->mthdbufMem.cacheAttrib = 1; - args->internalFlags = - NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, PRIVILEGE, ADMIN) | - NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, ERROR_NOTIFIER_TYPE, NONE) | - NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, ECC_ERROR_NOTIFIER_TYPE, NONE); - - ret = nvkm_gsp_rm_alloc_wr(&golden.chan, args); - if (ret) - goto done; - } - - /* Fetch context buffer info from RM and allocate each of them here to use - * during golden context init (or later as a global context buffer). - * - * Also build the information that'll be used to create channel contexts. - */ - info = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, - NV2080_CTRL_CMD_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO, - sizeof(*info)); - if (WARN_ON(IS_ERR(info))) { - ret = PTR_ERR(info); - goto done; - } - - for (int i = 0; i < ARRAY_SIZE(info->engineContextBuffersInfo[0].engine); i++) { - static const struct { - u32 id0; /* NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID */ - u32 id1; /* NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID */ - bool global; - bool init; - bool ro; - } map[] = { -#define _A(n,N,G,I,R) { .id0 = NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_##n, \ - .id1 = NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_##N, \ - .global = (G), .init = (I), .ro = (R) } -#define _B(N,G,I,R) _A(GRAPHICS_##N, N, (G), (I), (R)) - /* global init ro */ - _A( GRAPHICS, MAIN, false, true, false), - _B( PATCH, false, true, false), - _A( GRAPHICS_BUNDLE_CB, BUFFER_BUNDLE_CB, true, false, false), - _B( PAGEPOOL, true, false, false), - _B( ATTRIBUTE_CB, true, false, false), - _B( RTV_CB_GLOBAL, true, false, false), - _B( FECS_EVENT, true, true, false), - _B( PRIV_ACCESS_MAP, true, true, true), -#undef _B -#undef _A - }; - u32 size = info->engineContextBuffersInfo[0].engine[i].size; - u8 align, page; - int id; - - for (id = 0; id < ARRAY_SIZE(map); id++) { - if (map[id].id0 == i) - break; - } - - nvkm_debug(subdev, "%02x: size:0x%08x %s\n", i, - size, (id < ARRAY_SIZE(map)) ? "*" : ""); - if (id >= ARRAY_SIZE(map)) - continue; - - if (map[id].id1 == NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_MAIN) - size = ALIGN(size, 0x1000) + 64 * 0x1000; /* per-subctx headers */ - - if (size >= 1 << 21) page = 21; - else if (size >= 1 << 16) page = 16; - else page = 12; - - if (map[id].id1 == NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_ATTRIBUTE_CB) - align = order_base_2(size); - else - align = page; - - if (WARN_ON(gr->ctxbuf_nr == ARRAY_SIZE(gr->ctxbuf))) - continue; - - gr->ctxbuf[gr->ctxbuf_nr].bufferId = map[id].id1; - gr->ctxbuf[gr->ctxbuf_nr].size = size; - gr->ctxbuf[gr->ctxbuf_nr].page = page; - gr->ctxbuf[gr->ctxbuf_nr].align = align; - gr->ctxbuf[gr->ctxbuf_nr].global = map[id].global; - gr->ctxbuf[gr->ctxbuf_nr].init = map[id].init; - gr->ctxbuf[gr->ctxbuf_nr].ro = map[id].ro; - gr->ctxbuf_nr++; - - if (map[id].id1 == NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PRIV_ACCESS_MAP) { - if (WARN_ON(gr->ctxbuf_nr == ARRAY_SIZE(gr->ctxbuf))) - continue; - - gr->ctxbuf[gr->ctxbuf_nr] = gr->ctxbuf[gr->ctxbuf_nr - 1]; - gr->ctxbuf[gr->ctxbuf_nr].bufferId = - NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_UNRESTRICTED_PRIV_ACCESS_MAP; - gr->ctxbuf_nr++; - } - } - - nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, info); - - /* Promote golden context to RM. */ - ret = r535_gr_promote_ctx(gr, true, golden.vmm, gr->ctxbuf_mem, golden.vma, &golden.chan); - if (ret) - goto done; - - /* Allocate 3D class on channel to trigger golden context init in RM. */ - { - int i; - - for (i = 0; gr->base.func->sclass[i].ctor; i++) { - if ((gr->base.func->sclass[i].oclass & 0xff) == 0x97) { - struct nvkm_gsp_object threed; - - ret = nvkm_gsp_rm_alloc(&golden.chan, 0x97000000, - gr->base.func->sclass[i].oclass, 0, - &threed); - if (ret) - goto done; - - nvkm_gsp_rm_free(&threed); - break; - } - } - - if (WARN_ON(!gr->base.func->sclass[i].ctor)) { - ret = -EINVAL; - goto done; - } - } - -done: - nvkm_gsp_rm_free(&golden.chan); - for (int i = gr->ctxbuf_nr - 1; i >= 0; i--) - nvkm_vmm_put(golden.vmm, &golden.vma[i]); - nvkm_vmm_unref(&golden.vmm); - nvkm_memory_unref(&golden.inst); - return ret; - -} - -static void * -r535_gr_dtor(struct nvkm_gr *base) -{ - struct r535_gr *gr = r535_gr(base); - - while (gr->ctxbuf_nr) - nvkm_memory_unref(&gr->ctxbuf_mem[--gr->ctxbuf_nr]); - - kfree(gr->base.func); - return gr; -} - -int -r535_gr_new(const struct gf100_gr_func *hw, - struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr) -{ - struct nvkm_gr_func *rm; - struct r535_gr *gr; - int nclass; - - for (nclass = 0; hw->sclass[nclass].oclass; nclass++); - - if (!(rm = kzalloc(sizeof(*rm) + (nclass + 1) * sizeof(rm->sclass[0]), GFP_KERNEL))) - return -ENOMEM; - - rm->dtor = r535_gr_dtor; - rm->oneinit = r535_gr_oneinit; - rm->units = r535_gr_units; - rm->chan_new = r535_gr_chan_new; - - for (int i = 0; i < nclass; i++) { - rm->sclass[i].minver = hw->sclass[i].minver; - rm->sclass[i].maxver = hw->sclass[i].maxver; - rm->sclass[i].oclass = hw->sclass[i].oclass; - rm->sclass[i].ctor = r535_gr_obj_ctor; - } - - if (!(gr = kzalloc(sizeof(*gr), GFP_KERNEL))) { - kfree(rm); - return -ENOMEM; - } - - *pgr = &gr->base; - - return nvkm_gr_ctor(rm, device, type, inst, true, &gr->base); -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c index b7a458e9040a..bda8054c6b59 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/tu102.c @@ -219,7 +219,7 @@ int tu102_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr) { if (nvkm_gsp_rm(device->gsp)) - return r535_gr_new(&tu102_gr, device, type, inst, pgr); + return -ENODEV; return gf100_gr_new_(tu102_gr_fwif, device, type, inst, pgr); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/Kbuild index 2b0e923cb755..37b0cdc760c7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/Kbuild @@ -2,8 +2,4 @@ nvkm-y += nvkm/engine/nvdec/base.o nvkm-y += nvkm/engine/nvdec/gm107.o nvkm-y += nvkm/engine/nvdec/tu102.o -nvkm-y += nvkm/engine/nvdec/ga100.o nvkm-y += nvkm/engine/nvdec/ga102.o -nvkm-y += nvkm/engine/nvdec/ad102.o - -nvkm-y += nvkm/engine/nvdec/r535.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ga102.c index 022a9c824304..eea6368adae2 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ga102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ga102.c @@ -23,16 +23,6 @@ #include <subdev/gsp.h> -#include <nvif/class.h> - -static const struct nvkm_engine_func -ga102_nvdec_gsp = { - .sclass = { - { -1, -1, NVC7B0_VIDEO_DECODER }, - {} - } -}; - static const struct nvkm_falcon_func ga102_nvdec_flcn = { .disable = gm200_flcn_disable, @@ -67,7 +57,7 @@ ga102_nvdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst struct nvkm_nvdec **pnvdec) { if (nvkm_gsp_rm(device->gsp)) - return r535_nvdec_new(&ga102_nvdec_gsp, device, type, inst, pnvdec); + return -ENODEV; return nvkm_nvdec_new_(ga102_nvdec_fwif, device, type, inst, 0x848000, pnvdec); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h index f506ae83bfd7..f8d43e913093 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/priv.h @@ -20,7 +20,4 @@ extern const struct nvkm_nvdec_fwif gm107_nvdec_fwif[]; int nvkm_nvdec_new_(const struct nvkm_nvdec_fwif *fwif, struct nvkm_device *, enum nvkm_subdev_type, int, u32 addr, struct nvkm_nvdec **); - -int r535_nvdec_new(const struct nvkm_engine_func *, struct nvkm_device *, - enum nvkm_subdev_type, int, struct nvkm_nvdec **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/r535.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/r535.c deleted file mode 100644 index 75a24f3e6617..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/r535.c +++ /dev/null @@ -1,110 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <core/object.h> -#include <subdev/gsp.h> -#include <engine/fifo.h> - -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> - -struct r535_nvdec_obj { - struct nvkm_object object; - struct nvkm_gsp_object rm; -}; - -static void * -r535_nvdec_obj_dtor(struct nvkm_object *object) -{ - struct r535_nvdec_obj *obj = container_of(object, typeof(*obj), object); - - nvkm_gsp_rm_free(&obj->rm); - return obj; -} - -static const struct nvkm_object_func -r535_nvdec_obj = { - .dtor = r535_nvdec_obj_dtor, -}; - -static int -r535_nvdec_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, - struct nvkm_object **pobject) -{ - struct nvkm_chan *chan = nvkm_uchan_chan(oclass->parent); - struct r535_nvdec_obj *obj; - NV_BSP_ALLOCATION_PARAMETERS *args; - - if (!(obj = kzalloc(sizeof(*obj), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_nvdec_obj, oclass, &obj->object); - *pobject = &obj->object; - - args = nvkm_gsp_rm_alloc_get(&chan->rm.object, oclass->handle, oclass->base.oclass, - sizeof(*args), &obj->rm); - if (WARN_ON(IS_ERR(args))) - return PTR_ERR(args); - - args->size = sizeof(*args); - args->engineInstance = oclass->engine->subdev.inst; - - return nvkm_gsp_rm_alloc_wr(&obj->rm, args); -} - -static void * -r535_nvdec_dtor(struct nvkm_engine *engine) -{ - struct nvkm_nvdec *nvdec = nvkm_nvdec(engine); - - kfree(nvdec->engine.func); - return nvdec; -} - -int -r535_nvdec_new(const struct nvkm_engine_func *hw, struct nvkm_device *device, - enum nvkm_subdev_type type, int inst, struct nvkm_nvdec **pnvdec) -{ - struct nvkm_engine_func *rm; - int nclass; - - for (nclass = 0; hw->sclass[nclass].oclass; nclass++); - - if (!(rm = kzalloc(sizeof(*rm) + (nclass + 1) * sizeof(rm->sclass[0]), GFP_KERNEL))) - return -ENOMEM; - - rm->dtor = r535_nvdec_dtor; - for (int i = 0; i < nclass; i++) { - rm->sclass[i].minver = hw->sclass[i].minver; - rm->sclass[i].maxver = hw->sclass[i].maxver; - rm->sclass[i].oclass = hw->sclass[i].oclass; - rm->sclass[i].ctor = r535_nvdec_obj_ctor; - } - - if (!(*pnvdec = kzalloc(sizeof(**pnvdec), GFP_KERNEL))) { - kfree(rm); - return -ENOMEM; - } - - return nvkm_engine_ctor(rm, device, type, inst, true, &(*pnvdec)->engine); -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/tu102.c index 808c8e010b9e..fe95b6e22f21 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/tu102.c @@ -23,22 +23,12 @@ #include <subdev/gsp.h> -#include <nvif/class.h> - -static const struct nvkm_engine_func -tu102_nvdec = { - .sclass = { - { -1, -1, NVC4B0_VIDEO_DECODER }, - {} - } -}; - int tu102_nvdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_nvdec **pnvdec) { if (nvkm_gsp_rm(device->gsp)) - return r535_nvdec_new(&tu102_nvdec, device, type, inst, pnvdec); + return -ENODEV; return nvkm_nvdec_new_(gm107_nvdec_fwif, device, type, inst, 0, pnvdec); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/Kbuild index 2c1495b730f3..6dcb20d1d156 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/Kbuild @@ -2,7 +2,3 @@ nvkm-y += nvkm/engine/nvenc/base.o nvkm-y += nvkm/engine/nvenc/gm107.o nvkm-y += nvkm/engine/nvenc/tu102.o -nvkm-y += nvkm/engine/nvenc/ga102.o -nvkm-y += nvkm/engine/nvenc/ad102.o - -nvkm-y += nvkm/engine/nvenc/r535.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/ga102.c deleted file mode 100644 index 6463ab8e5871..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/ga102.c +++ /dev/null @@ -1,44 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_engine_func -ga102_nvenc = { - .sclass = { - { -1, -1, NVC7B7_VIDEO_ENCODER }, - {} - } -}; - -int -ga102_nvenc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_nvenc **pnvenc) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_nvenc_new(&ga102_nvenc, device, type, inst, pnvenc); - - return -ENODEV; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h index 7917affc6505..b097e3f2867b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/priv.h @@ -18,7 +18,4 @@ extern const struct nvkm_nvenc_fwif gm107_nvenc_fwif[]; int nvkm_nvenc_new_(const struct nvkm_nvenc_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_nvenc **pnvenc); - -int r535_nvenc_new(const struct nvkm_engine_func *, struct nvkm_device *, - enum nvkm_subdev_type, int, struct nvkm_nvenc **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/r535.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/r535.c deleted file mode 100644 index c8a2a9196ce5..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/r535.c +++ /dev/null @@ -1,110 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <core/object.h> -#include <subdev/gsp.h> -#include <engine/fifo.h> - -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> - -struct r535_nvenc_obj { - struct nvkm_object object; - struct nvkm_gsp_object rm; -}; - -static void * -r535_nvenc_obj_dtor(struct nvkm_object *object) -{ - struct r535_nvenc_obj *obj = container_of(object, typeof(*obj), object); - - nvkm_gsp_rm_free(&obj->rm); - return obj; -} - -static const struct nvkm_object_func -r535_nvenc_obj = { - .dtor = r535_nvenc_obj_dtor, -}; - -static int -r535_nvenc_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, - struct nvkm_object **pobject) -{ - struct nvkm_chan *chan = nvkm_uchan_chan(oclass->parent); - struct r535_nvenc_obj *obj; - NV_MSENC_ALLOCATION_PARAMETERS *args; - - if (!(obj = kzalloc(sizeof(*obj), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_nvenc_obj, oclass, &obj->object); - *pobject = &obj->object; - - args = nvkm_gsp_rm_alloc_get(&chan->rm.object, oclass->handle, oclass->base.oclass, - sizeof(*args), &obj->rm); - if (WARN_ON(IS_ERR(args))) - return PTR_ERR(args); - - args->size = sizeof(*args); - args->engineInstance = oclass->engine->subdev.inst; - - return nvkm_gsp_rm_alloc_wr(&obj->rm, args); -} - -static void * -r535_nvenc_dtor(struct nvkm_engine *engine) -{ - struct nvkm_nvenc *nvenc = nvkm_nvenc(engine); - - kfree(nvenc->engine.func); - return nvenc; -} - -int -r535_nvenc_new(const struct nvkm_engine_func *hw, struct nvkm_device *device, - enum nvkm_subdev_type type, int inst, struct nvkm_nvenc **pnvenc) -{ - struct nvkm_engine_func *rm; - int nclass; - - for (nclass = 0; hw->sclass[nclass].oclass; nclass++); - - if (!(rm = kzalloc(sizeof(*rm) + (nclass + 1) * sizeof(rm->sclass[0]), GFP_KERNEL))) - return -ENOMEM; - - rm->dtor = r535_nvenc_dtor; - for (int i = 0; i < nclass; i++) { - rm->sclass[i].minver = hw->sclass[i].minver; - rm->sclass[i].maxver = hw->sclass[i].maxver; - rm->sclass[i].oclass = hw->sclass[i].oclass; - rm->sclass[i].ctor = r535_nvenc_obj_ctor; - } - - if (!(*pnvenc = kzalloc(sizeof(**pnvenc), GFP_KERNEL))) { - kfree(rm); - return -ENOMEM; - } - - return nvkm_engine_ctor(rm, device, type, inst, true, &(*pnvenc)->engine); -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/tu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/tu102.c index 933864423bb3..8a436b398749 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/tu102.c @@ -23,22 +23,12 @@ #include <subdev/gsp.h> -#include <nvif/class.h> - -static const struct nvkm_engine_func -tu102_nvenc = { - .sclass = { - { -1, -1, NVC4B7_VIDEO_ENCODER }, - {} - } -}; - int tu102_nvenc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_nvenc **pnvenc) { if (nvkm_gsp_rm(device->gsp)) - return r535_nvenc_new(&tu102_nvenc, device, type, inst, pnvenc); + return -ENODEV; return nvkm_nvenc_new_(gm107_nvenc_fwif, device, type, inst, pnvenc); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/Kbuild deleted file mode 100644 index 1408f664add6..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/Kbuild +++ /dev/null @@ -1,5 +0,0 @@ -# SPDX-License-Identifier: MIT -nvkm-y += nvkm/engine/nvjpg/ga100.o -nvkm-y += nvkm/engine/nvjpg/ad102.o - -nvkm-y += nvkm/engine/nvjpg/r535.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/ad102.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/ad102.c deleted file mode 100644 index 62705dc6494c..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/ad102.c +++ /dev/null @@ -1,44 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_engine_func -ad102_nvjpg = { - .sclass = { - { -1, -1, NVC9D1_VIDEO_NVJPG }, - {} - } -}; - -int -ad102_nvjpg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_engine **pengine) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_nvjpg_new(&ad102_nvjpg, device, type, inst, pengine); - - return -ENODEV; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/ga100.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/ga100.c deleted file mode 100644 index f550eb07da5a..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/ga100.c +++ /dev/null @@ -1,44 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_engine_func -ga100_nvjpg = { - .sclass = { - { -1, -1, NVC4D1_VIDEO_NVJPG }, - {} - } -}; - -int -ga100_nvjpg_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_engine **pengine) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_nvjpg_new(&ga100_nvjpg, device, type, inst, pengine); - - return -ENODEV; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/priv.h deleted file mode 100644 index 1e80cf70033a..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/priv.h +++ /dev/null @@ -1,8 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -#ifndef __NVKM_NVJPG_PRIV_H__ -#define __NVKM_NVJPG_PRIV_H__ -#include <engine/nvjpg.h> - -int r535_nvjpg_new(const struct nvkm_engine_func *, struct nvkm_device *, - enum nvkm_subdev_type, int, struct nvkm_engine **); -#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/r535.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/r535.c deleted file mode 100644 index 1babddc4eb80..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvjpg/r535.c +++ /dev/null @@ -1,107 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <core/object.h> -#include <subdev/gsp.h> -#include <engine/fifo.h> - -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> - -struct r535_nvjpg_obj { - struct nvkm_object object; - struct nvkm_gsp_object rm; -}; - -static void * -r535_nvjpg_obj_dtor(struct nvkm_object *object) -{ - struct r535_nvjpg_obj *obj = container_of(object, typeof(*obj), object); - - nvkm_gsp_rm_free(&obj->rm); - return obj; -} - -static const struct nvkm_object_func -r535_nvjpg_obj = { - .dtor = r535_nvjpg_obj_dtor, -}; - -static int -r535_nvjpg_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, - struct nvkm_object **pobject) -{ - struct nvkm_chan *chan = nvkm_uchan_chan(oclass->parent); - struct r535_nvjpg_obj *obj; - NV_NVJPG_ALLOCATION_PARAMETERS *args; - - if (!(obj = kzalloc(sizeof(*obj), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_nvjpg_obj, oclass, &obj->object); - *pobject = &obj->object; - - args = nvkm_gsp_rm_alloc_get(&chan->rm.object, oclass->handle, oclass->base.oclass, - sizeof(*args), &obj->rm); - if (WARN_ON(IS_ERR(args))) - return PTR_ERR(args); - - args->size = sizeof(*args); - args->engineInstance = oclass->engine->subdev.inst; - - return nvkm_gsp_rm_alloc_wr(&obj->rm, args); -} - -static void * -r535_nvjpg_dtor(struct nvkm_engine *engine) -{ - kfree(engine->func); - return engine; -} - -int -r535_nvjpg_new(const struct nvkm_engine_func *hw, struct nvkm_device *device, - enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine) -{ - struct nvkm_engine_func *rm; - int nclass, ret; - - for (nclass = 0; hw->sclass[nclass].oclass; nclass++); - - if (!(rm = kzalloc(sizeof(*rm) + (nclass + 1) * sizeof(rm->sclass[0]), GFP_KERNEL))) - return -ENOMEM; - - rm->dtor = r535_nvjpg_dtor; - for (int i = 0; i < nclass; i++) { - rm->sclass[i].minver = hw->sclass[i].minver; - rm->sclass[i].maxver = hw->sclass[i].maxver; - rm->sclass[i].oclass = hw->sclass[i].oclass; - rm->sclass[i].ctor = r535_nvjpg_obj_ctor; - } - - ret = nvkm_engine_new_(rm, device, type, inst, true, pengine); - if (ret) - kfree(rm); - - return ret; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/ofa/Kbuild deleted file mode 100644 index 99f1713d7e51..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/Kbuild +++ /dev/null @@ -1,6 +0,0 @@ -# SPDX-License-Identifier: MIT -nvkm-y += nvkm/engine/ofa/ga100.o -nvkm-y += nvkm/engine/ofa/ga102.o -nvkm-y += nvkm/engine/ofa/ad102.o - -nvkm-y += nvkm/engine/ofa/r535.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ga100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ga100.c deleted file mode 100644 index ef474f61a1b5..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ga100.c +++ /dev/null @@ -1,44 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_engine_func -ga100_ofa = { - .sclass = { - { -1, -1, NVC6FA_VIDEO_OFA }, - {} - } -}; - -int -ga100_ofa_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_engine **pengine) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_ofa_new(&ga100_ofa, device, type, inst, pengine); - - return -ENODEV; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ga102.c deleted file mode 100644 index bea255529993..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/ga102.c +++ /dev/null @@ -1,44 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <subdev/gsp.h> - -#include <nvif/class.h> - -static const struct nvkm_engine_func -ga102_ofa = { - .sclass = { - { -1, -1, NVC7FA_VIDEO_OFA }, - {} - } -}; - -int -ga102_ofa_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_engine **pengine) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_ofa_new(&ga102_ofa, device, type, inst, pengine); - - return -ENODEV; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/ofa/priv.h deleted file mode 100644 index caf29e6bddb4..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/priv.h +++ /dev/null @@ -1,8 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -#ifndef __NVKM_OFA_PRIV_H__ -#define __NVKM_OFA_PRIV_H__ -#include <engine/ofa.h> - -int r535_ofa_new(const struct nvkm_engine_func *, struct nvkm_device *, - enum nvkm_subdev_type, int, struct nvkm_engine **); -#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/r535.c b/drivers/gpu/drm/nouveau/nvkm/engine/ofa/r535.c deleted file mode 100644 index 438dc692eefe..000000000000 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ofa/r535.c +++ /dev/null @@ -1,107 +0,0 @@ -/* - * Copyright 2023 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ -#include "priv.h" - -#include <core/object.h> -#include <subdev/gsp.h> -#include <subdev/mmu.h> -#include <engine/fifo.h> - -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> - -struct r535_ofa_obj { - struct nvkm_object object; - struct nvkm_gsp_object rm; -}; - -static void * -r535_ofa_obj_dtor(struct nvkm_object *object) -{ - struct r535_ofa_obj *obj = container_of(object, typeof(*obj), object); - - nvkm_gsp_rm_free(&obj->rm); - return obj; -} - -static const struct nvkm_object_func -r535_ofa_obj = { - .dtor = r535_ofa_obj_dtor, -}; - -static int -r535_ofa_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, - struct nvkm_object **pobject) -{ - struct nvkm_chan *chan = nvkm_uchan_chan(oclass->parent); - struct r535_ofa_obj *obj; - NV_OFA_ALLOCATION_PARAMETERS *args; - - if (!(obj = kzalloc(sizeof(*obj), GFP_KERNEL))) - return -ENOMEM; - - nvkm_object_ctor(&r535_ofa_obj, oclass, &obj->object); - *pobject = &obj->object; - - args = nvkm_gsp_rm_alloc_get(&chan->rm.object, oclass->handle, oclass->base.oclass, - sizeof(*args), &obj->rm); - if (WARN_ON(IS_ERR(args))) - return PTR_ERR(args); - - args->size = sizeof(*args); - - return nvkm_gsp_rm_alloc_wr(&obj->rm, args); -} - -static void * -r535_ofa_dtor(struct nvkm_engine *engine) -{ - kfree(engine->func); - return engine; -} - -int -r535_ofa_new(const struct nvkm_engine_func *hw, struct nvkm_device *device, - enum nvkm_subdev_type type, int inst, struct nvkm_engine **pengine) -{ - struct nvkm_engine_func *rm; - int nclass, ret; - - for (nclass = 0; hw->sclass[nclass].oclass; nclass++); - - if (!(rm = kzalloc(sizeof(*rm) + (nclass + 1) * sizeof(rm->sclass[0]), GFP_KERNEL))) - return -ENOMEM; - - rm->dtor = r535_ofa_dtor; - for (int i = 0; i < nclass; i++) { - rm->sclass[i].minver = hw->sclass[i].minver; - rm->sclass[i].maxver = hw->sclass[i].maxver; - rm->sclass[i].oclass = hw->sclass[i].oclass; - rm->sclass[i].ctor = r535_ofa_obj_ctor; - } - - ret = nvkm_engine_new_(rm, device, type, inst, true, pengine); - if (ret) - kfree(rm); - - return ret; -} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild index 4c2f6fc4ef58..c19ea4ea9bd3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/Kbuild @@ -9,6 +9,7 @@ include $(src)/nvkm/subdev/fault/Kbuild include $(src)/nvkm/subdev/fb/Kbuild include $(src)/nvkm/subdev/fuse/Kbuild include $(src)/nvkm/subdev/gpio/Kbuild +include $(src)/nvkm/subdev/fsp/Kbuild include $(src)/nvkm/subdev/gsp/Kbuild include $(src)/nvkm/subdev/i2c/Kbuild include $(src)/nvkm/subdev/iccsense/Kbuild diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild index 9754c6872543..8faee3317a74 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild @@ -7,5 +7,3 @@ nvkm-y += nvkm/subdev/bar/gk20a.o nvkm-y += nvkm/subdev/bar/gm107.o nvkm-y += nvkm/subdev/bar/gm20b.o nvkm-y += nvkm/subdev/bar/tu102.o - -nvkm-y += nvkm/subdev/bar/r535.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c index 51070b7dda85..e5e60915029c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c @@ -82,7 +82,7 @@ gf100_bar_bar2_init(struct nvkm_bar *base) static int gf100_bar_oneinit_bar(struct gf100_bar *bar, struct gf100_barN *bar_vm, - struct lock_class_key *key, int bar_nr) + struct lock_class_key *key, enum nvkm_bar_id bar_id) { struct nvkm_device *device = bar->base.subdev.device; resource_size_t bar_len; @@ -93,14 +93,14 @@ gf100_bar_oneinit_bar(struct gf100_bar *bar, struct gf100_barN *bar_vm, if (ret) return ret; - bar_len = device->func->resource_size(device, bar_nr); + bar_len = device->func->resource_size(device, bar_id); if (!bar_len) return -ENOMEM; - if (bar_nr == 3 && bar->bar2_halve) + if (bar_id == NVKM_BAR2_INST && bar->bar2_halve) bar_len >>= 1; ret = nvkm_vmm_new(device, 0, bar_len, NULL, 0, key, - (bar_nr == 3) ? "bar2" : "bar1", &bar_vm->vmm); + (bar_id == NVKM_BAR2_INST) ? "bar2" : "bar1", &bar_vm->vmm); if (ret) return ret; @@ -110,7 +110,7 @@ gf100_bar_oneinit_bar(struct gf100_bar *bar, struct gf100_barN *bar_vm, /* * Bootstrap page table lookup. */ - if (bar_nr == 3) { + if (bar_id == NVKM_BAR2_INST) { ret = nvkm_vmm_boot(bar_vm->vmm); if (ret) return ret; @@ -129,7 +129,7 @@ gf100_bar_oneinit(struct nvkm_bar *base) /* BAR2 */ if (bar->base.func->bar2.init) { - ret = gf100_bar_oneinit_bar(bar, &bar->bar[0], &bar2_lock, 3); + ret = gf100_bar_oneinit_bar(bar, &bar->bar[0], &bar2_lock, NVKM_BAR2_INST); if (ret) return ret; @@ -138,7 +138,7 @@ gf100_bar_oneinit(struct nvkm_bar *base) } /* BAR1 */ - ret = gf100_bar_oneinit_bar(bar, &bar->bar[1], &bar1_lock, 1); + ret = gf100_bar_oneinit_bar(bar, &bar->bar[1], &bar1_lock, NVKM_BAR1_FB); if (ret) return ret; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c index 27d8a1be43e4..6a881becb02c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/nv50.c @@ -127,7 +127,7 @@ nv50_bar_oneinit(struct nvkm_bar *base) /* BAR2 */ start = 0x0100000000ULL; - size = device->func->resource_size(device, 3); + size = device->func->resource_size(device, NVKM_BAR2_INST); if (!size) return -ENOMEM; limit = start + size; @@ -167,7 +167,7 @@ nv50_bar_oneinit(struct nvkm_bar *base) /* BAR1 */ start = 0x0000000000ULL; - size = device->func->resource_size(device, 1); + size = device->func->resource_size(device, NVKM_BAR1_FB); if (!size) return -ENOMEM; limit = start + size; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/fbmem.h b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/fbmem.h index 6c5bbff12eb4..b918e22df5a8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/fbmem.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/fbmem.h @@ -47,8 +47,8 @@ static inline struct io_mapping * fbmem_init(struct nvkm_device *dev) { - return io_mapping_create_wc(dev->func->resource_addr(dev, 1), - dev->func->resource_size(dev, 1)); + return io_mapping_create_wc(dev->func->resource_addr(dev, NVKM_BAR1_FB), + dev->func->resource_size(dev, NVKM_BAR1_FB)); } static inline void diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/user.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/user.c index c123e5893d76..cd2fbc0472d8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/user.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/user.c @@ -50,7 +50,7 @@ nvkm_ufault_map(struct nvkm_object *object, void *argv, u32 argc, struct nvkm_fault_buffer *buffer = nvkm_fault_buffer(object); struct nvkm_device *device = buffer->fault->subdev.device; *type = NVKM_OBJECT_MAP_IO; - *addr = device->func->resource_addr(device, 3) + buffer->addr; + *addr = device->func->resource_addr(device, NVKM_BAR2_INST) + buffer->addr; *size = nvkm_memory_size(buffer->mem); return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild index d1611ad3bf81..8d8a5382d1b1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild @@ -35,6 +35,9 @@ nvkm-y += nvkm/subdev/fb/gv100.o nvkm-y += nvkm/subdev/fb/tu102.o nvkm-y += nvkm/subdev/fb/ga100.o nvkm-y += nvkm/subdev/fb/ga102.o +nvkm-y += nvkm/subdev/fb/gh100.o +nvkm-y += nvkm/subdev/fb/gb100.o +nvkm-y += nvkm/subdev/fb/gb202.o nvkm-y += nvkm/subdev/fb/r535.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c index 25f82b372bca..2819780050d8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ga102.c @@ -25,7 +25,7 @@ #include <subdev/gsp.h> #include <engine/nvdec.h> -static u64 +u64 ga102_fb_vidmem_size(struct nvkm_fb *fb) { return (u64)nvkm_rd32(fb->subdev.device, 0x1183a4) << 20; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gb100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gb100.c new file mode 100644 index 000000000000..1c78c8853617 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gb100.c @@ -0,0 +1,34 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gb100/dev_hshub_base.h> + +static void +gb100_fb_sysmem_flush_page_init(struct nvkm_fb *fb) +{ + const u32 addr_hi = upper_32_bits(fb->sysmem.flush_page_addr); + const u32 addr_lo = lower_32_bits(fb->sysmem.flush_page_addr); + const u32 hshub = DRF_LO(NV_PFB_HSHUB0); + struct nvkm_device *device = fb->subdev.device; + + nvkm_wr32(device, hshub + NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_HI, addr_hi); + nvkm_wr32(device, hshub + NV_PFB_HSHUB_PCIE_FLUSH_SYSMEM_ADDR_LO, addr_lo); + nvkm_wr32(device, hshub + NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_HI, addr_hi); + nvkm_wr32(device, hshub + NV_PFB_HSHUB_EG_PCIE_FLUSH_SYSMEM_ADDR_LO, addr_lo); +} + +static const struct nvkm_fb_func +gb100_fb = { + .sysmem.flush_page_init = gb100_fb_sysmem_flush_page_init, + .vidmem.size = ga102_fb_vidmem_size, +}; + +int +gb100_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb) +{ + return r535_fb_new(&gb100_fb, device, type, inst, pfb); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gb202.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gb202.c new file mode 100644 index 000000000000..848505026d02 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gb202.c @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gb10b/dev_fbhub.h> + +static void +gb202_fb_sysmem_flush_page_init(struct nvkm_fb *fb) +{ + struct nvkm_device *device = fb->subdev.device; + const u64 addr = fb->sysmem.flush_page_addr; + + nvkm_wr32(device, NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_HI, upper_32_bits(addr)); + nvkm_wr32(device, NV_PFB_FBHUB0_PCIE_FLUSH_SYSMEM_ADDR_LO, lower_32_bits(addr)); +} + +static const struct nvkm_fb_func +gb202_fb = { + .sysmem.flush_page_init = gb202_fb_sysmem_flush_page_init, + .vidmem.size = ga102_fb_vidmem_size, +}; + +int +gb202_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb) +{ + return r535_fb_new(&gb202_fb, device, type, inst, pfb); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gh100.c new file mode 100644 index 000000000000..2d8c51f882d5 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gh100.c @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gh100/dev_fb.h> + +static void +gh100_fb_sysmem_flush_page_init(struct nvkm_fb *fb) +{ + const u64 addr = fb->sysmem.flush_page_addr >> NV_PFB_NISO_FLUSH_SYSMEM_ADDR_SHIFT; + struct nvkm_device *device = fb->subdev.device; + + nvkm_wr32(device, NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_HI, upper_32_bits(addr)); + nvkm_wr32(device, NV_PFB_FBHUB_PCIE_FLUSH_SYSMEM_ADDR_LO, lower_32_bits(addr)); +} + +static const struct nvkm_fb_func +gh100_fb = { + .sysmem.flush_page_init = gh100_fb_sysmem_flush_page_init, + .vidmem.size = ga102_fb_vidmem_size, +}; + +int +gh100_fb_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fb **pfb) +{ + return r535_fb_new(&gh100_fb, device, type, inst, pfb); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h index 35c55dfba23d..ebe996503ab2 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h @@ -98,4 +98,6 @@ int gp102_fb_vpr_scrub(struct nvkm_fb *); int gv100_fb_init_page(struct nvkm_fb *); bool tu102_fb_vpr_scrub_required(struct nvkm_fb *); + +u64 ga102_fb_vidmem_size(struct nvkm_fb *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/Kbuild new file mode 100644 index 000000000000..1a9ded3a86f8 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/Kbuild @@ -0,0 +1,8 @@ +# SPDX-License-Identifier: MIT +# +# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + +nvkm-y += nvkm/subdev/fsp/base.o +nvkm-y += nvkm/subdev/fsp/gh100.o +nvkm-y += nvkm/subdev/fsp/gb100.o +nvkm-y += nvkm/subdev/fsp/gb202.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/base.c new file mode 100644 index 000000000000..e366a980baa9 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/base.c @@ -0,0 +1,66 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +int +nvkm_fsp_boot_gsp_fmc(struct nvkm_fsp *fsp, u64 args_addr, u32 rsvd_size, bool resume, + u64 img_addr, const u8 *hash, const u8 *pkey, const u8 *sig) +{ + return fsp->func->cot.boot_gsp_fmc(fsp, args_addr, rsvd_size, resume, + img_addr, hash, pkey, sig); +} + +bool +nvkm_fsp_verify_gsp_fmc(struct nvkm_fsp *fsp, u32 hash_size, u32 pkey_size, u32 sig_size) +{ + return hash_size == fsp->func->cot.size_hash && + pkey_size == fsp->func->cot.size_pkey && + sig_size == fsp->func->cot.size_sig; +} + +static int +nvkm_fsp_preinit(struct nvkm_subdev *subdev) +{ + struct nvkm_fsp *fsp = nvkm_fsp(subdev); + + return fsp->func->wait_secure_boot(fsp); +} + +static void * +nvkm_fsp_dtor(struct nvkm_subdev *subdev) +{ + struct nvkm_fsp *fsp = nvkm_fsp(subdev); + + nvkm_falcon_dtor(&fsp->falcon); + return fsp; +} + +static const struct nvkm_falcon_func +nvkm_fsp_flcn = { + .emem_pio = &gp102_flcn_emem_pio, +}; + +static const struct nvkm_subdev_func +nvkm_fsp = { + .dtor = nvkm_fsp_dtor, + .preinit = nvkm_fsp_preinit, +}; + +int +nvkm_fsp_new_(const struct nvkm_fsp_func *func, + struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_fsp **pfsp) +{ + struct nvkm_fsp *fsp; + + fsp = *pfsp = kzalloc(sizeof(*fsp), GFP_KERNEL); + if (!fsp) + return -ENOMEM; + + fsp->func = func; + nvkm_subdev_ctor(&nvkm_fsp, device, type, inst, &fsp->subdev); + + return nvkm_falcon_ctor(&nvkm_fsp_flcn, &fsp->subdev, "fsp", 0x8f2000, &fsp->falcon); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gb100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gb100.c new file mode 100644 index 000000000000..e06636bf54b6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gb100.c @@ -0,0 +1,24 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +static const struct nvkm_fsp_func +gb100_fsp = { + .wait_secure_boot = gh100_fsp_wait_secure_boot, + .cot = { + .version = 2, + .size_hash = 48, + .size_pkey = 97, + .size_sig = 96, + .boot_gsp_fmc = gh100_fsp_boot_gsp_fmc, + }, +}; + +int +gb100_fsp_new(struct nvkm_device *device, + enum nvkm_subdev_type type, int inst, struct nvkm_fsp **pfsp) +{ + return nvkm_fsp_new_(&gb100_fsp, device, type, inst, pfsp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gb202.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gb202.c new file mode 100644 index 000000000000..3438aac6383e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gb202.c @@ -0,0 +1,45 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gb202/dev_therm.h> + +static int +gb202_fsp_wait_secure_boot(struct nvkm_fsp *fsp) +{ + struct nvkm_device *device = fsp->subdev.device; + unsigned timeout_ms = 4000; + + do { + u32 status = NVKM_RD32(device, NV_THERM, I2CS_SCRATCH, FSP_BOOT_COMPLETE_STATUS); + + if (status == NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS_SUCCESS) + return 0; + + usleep_range(1000, 2000); + } while (timeout_ms--); + + return -ETIMEDOUT; +} + +static const struct nvkm_fsp_func +gb202_fsp = { + .wait_secure_boot = gb202_fsp_wait_secure_boot, + .cot = { + .version = 2, + .size_hash = 48, + .size_pkey = 97, + .size_sig = 96, + .boot_gsp_fmc = gh100_fsp_boot_gsp_fmc, + }, +}; + +int +gb202_fsp_new(struct nvkm_device *device, + enum nvkm_subdev_type type, int inst, struct nvkm_fsp **pfsp) +{ + return nvkm_fsp_new_(&gb202_fsp, device, type, inst, pfsp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gh100.c new file mode 100644 index 000000000000..2815be4bf5de --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/gh100.c @@ -0,0 +1,275 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gh100/dev_fsp_pri.h> +#include <nvhw/ref/gh100/dev_therm.h> + +#include <nvrm/nvtypes.h> + +#define MCTP_HEADER_VERSION 3:0 +#define MCTP_HEADER_RSVD 7:4 + +#define MCTP_HEADER_DEID 15:8 +#define MCTP_HEADER_SEID 23:16 + +#define MCTP_HEADER_TAG 26:24 +#define MCTP_HEADER_TO 27:27 +#define MCTP_HEADER_SEQ 29:28 +#define MCTP_HEADER_EOM 30:30 +#define MCTP_HEADER_SOM 31:31 + +#define MCTP_MSG_HEADER_TYPE 6:0 +#define MCTP_MSG_HEADER_IC 7:7 + +#define MCTP_MSG_HEADER_VENDOR_ID 23:8 +#define MCTP_MSG_HEADER_NVDM_TYPE 31:24 + +#define MCTP_MSG_HEADER_TYPE_VENDOR_PCI 0x7e +#define MCTP_MSG_HEADER_VENDOR_ID_NV 0x10de + +#define NVDM_TYPE_COT 0x14 +#define NVDM_TYPE_FSP_RESPONSE 0x15 + +#pragma pack(1) +typedef struct nvdm_payload_cot +{ + NvU16 version; + NvU16 size; + NvU64 gspFmcSysmemOffset; + NvU64 frtsSysmemOffset; + NvU32 frtsSysmemSize; + + // Note this is an offset from the end of FB + NvU64 frtsVidmemOffset; + NvU32 frtsVidmemSize; + + // Authentication related fields + NvU32 hash384[12]; + NvU32 publicKey[96]; + NvU32 signature[96]; + + NvU64 gspBootArgsSysmemOffset; +} NVDM_PAYLOAD_COT; +#pragma pack() + +#pragma pack(1) +typedef struct +{ + NvU32 taskId; + NvU32 commandNvdmType; + NvU32 errorCode; +} NVDM_PAYLOAD_COMMAND_RESPONSE; +#pragma pack() + +static u32 +gh100_fsp_poll(struct nvkm_fsp *fsp) +{ + struct nvkm_device *device = fsp->subdev.device; + u32 head, tail; + + head = nvkm_rd32(device, NV_PFSP_MSGQ_HEAD(0)); + tail = nvkm_rd32(device, NV_PFSP_MSGQ_TAIL(0)); + + if (head == tail) + return 0; + + return (tail - head) + sizeof(u32); /* TAIL points at last DWORD written. */ +} + +static int +gh100_fsp_recv(struct nvkm_fsp *fsp, u8 *packet, u32 max_packet_size) +{ + struct nvkm_device *device = fsp->subdev.device; + u32 packet_size; + int ret; + + packet_size = gh100_fsp_poll(fsp); + if (!packet_size || WARN_ON(packet_size % 4 || packet_size > max_packet_size)) + return -EINVAL; + + ret = nvkm_falcon_pio_rd(&fsp->falcon, 0, EMEM, 0, packet, 0, packet_size); + if (ret) + return ret; + + nvkm_wr32(device, NV_PFSP_MSGQ_TAIL(0), 0); + nvkm_wr32(device, NV_PFSP_MSGQ_HEAD(0), 0); + + return packet_size; +} + +static int +gh100_fsp_wait(struct nvkm_fsp *fsp) +{ + int time = 1000; + + do { + if (gh100_fsp_poll(fsp)) + return 0; + + usleep_range(1000, 2000); + } while(time--); + + return -ETIMEDOUT; +} + +static int +gh100_fsp_send(struct nvkm_fsp *fsp, const u8 *packet, u32 packet_size) +{ + struct nvkm_device *device = fsp->subdev.device; + int time = 1000, ret; + + if (WARN_ON(packet_size % sizeof(u32))) + return -EINVAL; + + /* Ensure any previously sent message has been consumed. */ + do { + u32 head = nvkm_rd32(device, NV_PFSP_QUEUE_HEAD(0)); + u32 tail = nvkm_rd32(device, NV_PFSP_QUEUE_TAIL(0)); + + if (tail == head) + break; + + usleep_range(1000, 2000); + } while(time--); + + if (time < 0) + return -ETIMEDOUT; + + /* Write message to EMEM. */ + ret = nvkm_falcon_pio_wr(&fsp->falcon, packet, 0, 0, EMEM, 0, packet_size, 0, false); + if (ret) + return ret; + + /* Update queue pointers - TAIL points at last DWORD written. */ + nvkm_wr32(device, NV_PFSP_QUEUE_TAIL(0), packet_size - sizeof(u32)); + nvkm_wr32(device, NV_PFSP_QUEUE_HEAD(0), 0); + return 0; +} + +static int +gh100_fsp_send_sync(struct nvkm_fsp *fsp, u8 nvdm_type, const u8 *packet, u32 packet_size) +{ + struct nvkm_subdev *subdev = &fsp->subdev; + struct { + u32 mctp_header; + u32 nvdm_header; + NVDM_PAYLOAD_COMMAND_RESPONSE response; + } reply; + int ret; + + ret = gh100_fsp_send(fsp, packet, packet_size); + if (ret) + return ret; + + ret = gh100_fsp_wait(fsp); + if (ret) + return ret; + + ret = gh100_fsp_recv(fsp, (u8 *)&reply, sizeof(reply)); + if (ret < 0) + return ret; + + if (NVVAL_TEST(reply.mctp_header, MCTP, HEADER, SOM, !=, 1) || + NVVAL_TEST(reply.mctp_header, MCTP, HEADER, EOM, !=, 1)) { + nvkm_error(subdev, "unexpected MCTP header in reply: 0x%08x\n", reply.mctp_header); + return -EIO; + } + + if (NVDEF_TEST(reply.nvdm_header, MCTP, MSG_HEADER, TYPE, !=, VENDOR_PCI) || + NVDEF_TEST(reply.nvdm_header, MCTP, MSG_HEADER, VENDOR_ID, !=, NV) || + NVVAL_TEST(reply.nvdm_header, MCTP, MSG_HEADER, NVDM_TYPE, !=, NVDM_TYPE_FSP_RESPONSE)) { + nvkm_error(subdev, "unexpected NVDM header in reply: 0x%08x\n", reply.nvdm_header); + return -EIO; + } + + if (reply.response.commandNvdmType != nvdm_type) { + nvkm_error(subdev, "expected NVDM type 0x%02x in reply, got 0x%02x\n", + nvdm_type, reply.response.commandNvdmType); + return -EIO; + } + + if (reply.response.errorCode) { + nvkm_error(subdev, "NVDM command 0x%02x failed with error 0x%08x\n", + nvdm_type, reply.response.errorCode); + return -EIO; + } + + return 0; +} + +int +gh100_fsp_boot_gsp_fmc(struct nvkm_fsp *fsp, u64 args_addr, u32 rsvd_size, bool resume, + u64 img_addr, const u8 *hash, const u8 *pkey, const u8 *sig) +{ + struct { + u32 mctp_header; + u32 nvdm_header; + NVDM_PAYLOAD_COT cot; + } msg = {}; + + msg.mctp_header = NVVAL(MCTP, HEADER, SOM, 1) | + NVVAL(MCTP, HEADER, EOM, 1) | + NVVAL(MCTP, HEADER, SEID, 0) | + NVVAL(MCTP, HEADER, SEQ, 0); + + msg.nvdm_header = NVDEF(MCTP, MSG_HEADER, TYPE, VENDOR_PCI) | + NVDEF(MCTP, MSG_HEADER, VENDOR_ID, NV) | + NVVAL(MCTP, MSG_HEADER, NVDM_TYPE, NVDM_TYPE_COT); + + msg.cot.version = fsp->func->cot.version; + msg.cot.size = sizeof(msg.cot); + msg.cot.gspFmcSysmemOffset = img_addr; + if (!resume) { + msg.cot.frtsVidmemOffset = ALIGN(rsvd_size, 0x200000); + msg.cot.frtsVidmemSize = 0x100000; + } + + memcpy(msg.cot.hash384, hash, fsp->func->cot.size_hash); + memcpy(msg.cot.publicKey, pkey, fsp->func->cot.size_pkey); + memcpy(msg.cot.signature, sig, fsp->func->cot.size_sig); + + msg.cot.gspBootArgsSysmemOffset = args_addr; + + return gh100_fsp_send_sync(fsp, NVDM_TYPE_COT, (const u8 *)&msg, sizeof(msg)); +} + +int +gh100_fsp_wait_secure_boot(struct nvkm_fsp *fsp) +{ + struct nvkm_device *device = fsp->subdev.device; + unsigned timeout_ms = 4000; + + do { + u32 status = NVKM_RD32(device, NV_THERM, I2CS_SCRATCH, FSP_BOOT_COMPLETE_STATUS); + + if (status == NV_THERM_I2CS_SCRATCH_FSP_BOOT_COMPLETE_STATUS_SUCCESS) + return 0; + + usleep_range(1000, 2000); + } while (timeout_ms--); + + return -ETIMEDOUT; +} + +static const struct nvkm_fsp_func +gh100_fsp = { + .wait_secure_boot = gh100_fsp_wait_secure_boot, + .cot = { + .version = 1, + .size_hash = 48, + .size_pkey = 384, + .size_sig = 384, + .boot_gsp_fmc = gh100_fsp_boot_gsp_fmc, + }, +}; + +int +gh100_fsp_new(struct nvkm_device *device, + enum nvkm_subdev_type type, int inst, struct nvkm_fsp **pfsp) +{ + return nvkm_fsp_new_(&gh100_fsp, device, type, inst, pfsp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/priv.h new file mode 100644 index 000000000000..f0b2c605c33d --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fsp/priv.h @@ -0,0 +1,29 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVKM_FSP_PRIV_H__ +#define __NVKM_FSP_PRIV_H__ +#define nvkm_fsp(p) container_of((p), struct nvkm_fsp, subdev) +#include <subdev/fsp.h> + +struct nvkm_fsp_func { + int (*wait_secure_boot)(struct nvkm_fsp *); + + struct { + u32 version; + u32 size_hash; + u32 size_pkey; + u32 size_sig; + int (*boot_gsp_fmc)(struct nvkm_fsp *, u64 args_addr, u32 rsvd_size, bool resume, + u64 img_addr, const u8 *hash, const u8 *pkey, const u8 *sig); + } cot; +}; + +int nvkm_fsp_new_(const struct nvkm_fsp_func *, + struct nvkm_device *, enum nvkm_subdev_type, int inst, struct nvkm_fsp **); + +int gh100_fsp_wait_secure_boot(struct nvkm_fsp *); +int gh100_fsp_boot_gsp_fmc(struct nvkm_fsp *, u64 args_addr, u32 rsvd_size, bool resume, + u64 img_addr, const u8 *hash, const u8 *pkey, const u8 *sig); +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/Kbuild index 16bf2f1bb780..e9c948b67bbd 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/Kbuild @@ -7,6 +7,9 @@ nvkm-y += nvkm/subdev/gsp/tu102.o nvkm-y += nvkm/subdev/gsp/tu116.o nvkm-y += nvkm/subdev/gsp/ga100.o nvkm-y += nvkm/subdev/gsp/ga102.o +nvkm-y += nvkm/subdev/gsp/gh100.o nvkm-y += nvkm/subdev/gsp/ad102.o +nvkm-y += nvkm/subdev/gsp/gb100.o +nvkm-y += nvkm/subdev/gsp/gb202.o -nvkm-y += nvkm/subdev/gsp/r535.o +include $(src)/nvkm/subdev/gsp/rm/Kbuild diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ad102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ad102.c index c849c6299c52..eb765da0876e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ad102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ad102.c @@ -22,30 +22,27 @@ #include "priv.h" static const struct nvkm_gsp_func -ad102_gsp_r535_113_01 = { +ad102_gsp = { .flcn = &ga102_gsp_flcn, .fwsec = &ga102_gsp_fwsec, .sig_section = ".fwsignature_ad10x", - .wpr_heap.os_carveout_size = 20 << 20, - .wpr_heap.base_size = 8 << 20, - .wpr_heap.min_size = 84 << 20, - .booter.ctor = ga102_gsp_booter_ctor, .dtor = r535_gsp_dtor, .oneinit = tu102_gsp_oneinit, - .init = r535_gsp_init, - .fini = r535_gsp_fini, + .init = tu102_gsp_init, + .fini = tu102_gsp_fini, .reset = ga102_gsp_reset, - .rm = &r535_gsp_rm, + .rm.gpu = &ad10x_gpu, }; static struct nvkm_gsp_fwif ad102_gsps[] = { - { 0, r535_gsp_load, &ad102_gsp_r535_113_01, "535.113.01", true }, + { 1, tu102_gsp_load, &ad102_gsp, &r570_rm_ga102, "570.144", true }, + { 0, tu102_gsp_load, &ad102_gsp, &r535_rm_ga102, "535.113.01", true }, {} }; @@ -55,3 +52,15 @@ ad102_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, { return nvkm_gsp_new_(ad102_gsps, device, type, inst, pgsp); } + +NVKM_GSP_FIRMWARE_BOOTER(ad102, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ad103, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ad104, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ad106, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ad107, 535.113.01); + +NVKM_GSP_FIRMWARE_BOOTER(ad102, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ad103, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ad104, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ad106, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ad107, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c index da1bebb896f7..d23243a83a4c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/base.c @@ -52,7 +52,7 @@ nvkm_gsp_fini(struct nvkm_subdev *subdev, bool suspend) { struct nvkm_gsp *gsp = nvkm_gsp(subdev); - if (!gsp->func->fini) + if (!gsp->func->fini || !gsp->running) return 0; return gsp->func->fini(gsp, suspend); @@ -80,6 +80,21 @@ nvkm_gsp_oneinit(struct nvkm_subdev *subdev) return gsp->func->oneinit(gsp); } +void +nvkm_gsp_dtor_fws(struct nvkm_gsp *gsp) +{ + nvkm_firmware_put(gsp->fws.fmc); + gsp->fws.fmc = NULL; + nvkm_firmware_put(gsp->fws.bl); + gsp->fws.bl = NULL; + nvkm_firmware_put(gsp->fws.booter.unload); + gsp->fws.booter.unload = NULL; + nvkm_firmware_put(gsp->fws.booter.load); + gsp->fws.booter.load = NULL; + nvkm_firmware_put(gsp->fws.rm); + gsp->fws.rm = NULL; +} + static void * nvkm_gsp_dtor(struct nvkm_subdev *subdev) { @@ -89,6 +104,7 @@ nvkm_gsp_dtor(struct nvkm_subdev *subdev) gsp->func->dtor(gsp); nvkm_falcon_dtor(&gsp->falcon); + kfree(gsp->rm); return gsp; } @@ -101,6 +117,16 @@ nvkm_gsp = { }; int +nvkm_gsp_load_fw(struct nvkm_gsp *gsp, const char *name, const char *ver, + const struct firmware **pfw) +{ + char fwname[64]; + + snprintf(fwname, sizeof(fwname), "gsp/%s-%s", name, ver); + return nvkm_firmware_get(&gsp->subdev, fwname, 0, pfw); +} + +int nvkm_gsp_new_(const struct nvkm_gsp_fwif *fwif, struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gsp **pgsp) { @@ -116,7 +142,19 @@ nvkm_gsp_new_(const struct nvkm_gsp_fwif *fwif, struct nvkm_device *device, return PTR_ERR(fwif); gsp->func = fwif->func; - gsp->rm = gsp->func->rm; + + if (fwif->rm) { + nvkm_info(&gsp->subdev, "RM version: %s\n", fwif->ver); + + gsp->rm = kzalloc(sizeof(*gsp->rm), GFP_KERNEL); + if (!gsp->rm) + return -ENOMEM; + + gsp->rm->device = device; + gsp->rm->gpu = fwif->func->rm.gpu; + gsp->rm->wpr = fwif->rm->wpr; + gsp->rm->api = fwif->rm->api; + } return nvkm_falcon_ctor(gsp->func->flcn, &gsp->subdev, gsp->subdev.name, 0x110000, &gsp->falcon); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c index 223f68b532ef..d201e8697226 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c @@ -39,29 +39,27 @@ ga100_gsp_flcn = { }; static const struct nvkm_gsp_func -ga100_gsp_r535_113_01 = { +ga100_gsp = { .flcn = &ga100_gsp_flcn, .fwsec = &tu102_gsp_fwsec, .sig_section = ".fwsignature_ga100", - .wpr_heap.base_size = 8 << 20, - .wpr_heap.min_size = 64 << 20, - .booter.ctor = tu102_gsp_booter_ctor, .dtor = r535_gsp_dtor, .oneinit = tu102_gsp_oneinit, - .init = r535_gsp_init, - .fini = r535_gsp_fini, + .init = tu102_gsp_init, + .fini = tu102_gsp_fini, .reset = tu102_gsp_reset, - .rm = &r535_gsp_rm, + .rm.gpu = &ga100_gpu, }; static struct nvkm_gsp_fwif ga100_gsps[] = { - { 0, r535_gsp_load, &ga100_gsp_r535_113_01, "535.113.01" }, + { 1, tu102_gsp_load, &ga100_gsp, &r570_rm_tu102, "570.144" }, + { 0, tu102_gsp_load, &ga100_gsp, &r535_rm_tu102, "535.113.01" }, { -1, gv100_gsp_nofw, &gv100_gsp }, {} }; @@ -72,3 +70,6 @@ ga100_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, { return nvkm_gsp_new_(ga100_gsps, device, type, inst, pgsp); } + +NVKM_GSP_FIRMWARE_BOOTER(ga100, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ga100, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga102.c index 4c4b4168a266..917f7e2f6c46 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga102.c @@ -150,25 +150,21 @@ ga102_gsp_flcn = { }; static const struct nvkm_gsp_func -ga102_gsp_r535_113_01 = { +ga102_gsp_r535 = { .flcn = &ga102_gsp_flcn, .fwsec = &ga102_gsp_fwsec, .sig_section = ".fwsignature_ga10x", - .wpr_heap.os_carveout_size = 20 << 20, - .wpr_heap.base_size = 8 << 20, - .wpr_heap.min_size = 84 << 20, - .booter.ctor = ga102_gsp_booter_ctor, .dtor = r535_gsp_dtor, .oneinit = tu102_gsp_oneinit, - .init = r535_gsp_init, - .fini = r535_gsp_fini, + .init = tu102_gsp_init, + .fini = tu102_gsp_fini, .reset = ga102_gsp_reset, - .rm = &r535_gsp_rm, + .rm.gpu = &ga1xx_gpu, }; static const struct nvkm_gsp_func @@ -178,7 +174,8 @@ ga102_gsp = { static struct nvkm_gsp_fwif ga102_gsps[] = { - { 0, r535_gsp_load, &ga102_gsp_r535_113_01, "535.113.01" }, + { 1, tu102_gsp_load, &ga102_gsp_r535, &r570_rm_ga102, "570.144" }, + { 0, tu102_gsp_load, &ga102_gsp_r535, &r535_rm_ga102, "535.113.01" }, { -1, gv100_gsp_nofw, &ga102_gsp }, {} }; @@ -189,3 +186,15 @@ ga102_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, { return nvkm_gsp_new_(ga102_gsps, device, type, inst, pgsp); } + +NVKM_GSP_FIRMWARE_BOOTER(ga102, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ga103, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ga104, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ga106, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(ga107, 535.113.01); + +NVKM_GSP_FIRMWARE_BOOTER(ga102, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ga103, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ga104, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ga106, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(ga107, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gb100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gb100.c new file mode 100644 index 000000000000..12a3f2c1ed82 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gb100.c @@ -0,0 +1,35 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +static const struct nvkm_gsp_func +gb100_gsp = { + .flcn = &ga102_gsp_flcn, + + .sig_section = ".fwsignature_gb10x", + + .dtor = r535_gsp_dtor, + .oneinit = gh100_gsp_oneinit, + .init = gh100_gsp_init, + .fini = gh100_gsp_fini, + + .rm.gpu = &gb10x_gpu, +}; + +static struct nvkm_gsp_fwif +gb100_gsps[] = { + { 0, gh100_gsp_load, &gb100_gsp, &r570_rm_gb10x, "570.144", true }, + {} +}; + +int +gb100_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_gsp **pgsp) +{ + return nvkm_gsp_new_(gb100_gsps, device, type, inst, pgsp); +} + +NVKM_GSP_FIRMWARE_FMC(gb100, 570.144); +NVKM_GSP_FIRMWARE_FMC(gb102, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gb202.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gb202.c new file mode 100644 index 000000000000..c1d718172ddf --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gb202.c @@ -0,0 +1,38 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +static const struct nvkm_gsp_func +gb202_gsp = { + .flcn = &ga102_gsp_flcn, + + .sig_section = ".fwsignature_gb20x", + + .dtor = r535_gsp_dtor, + .oneinit = gh100_gsp_oneinit, + .init = gh100_gsp_init, + .fini = gh100_gsp_fini, + + .rm.gpu = &gb20x_gpu, +}; + +static struct nvkm_gsp_fwif +gb202_gsps[] = { + { 0, gh100_gsp_load, &gb202_gsp, &r570_rm_gb20x, "570.144", true }, + {} +}; + +int +gb202_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_gsp **pgsp) +{ + return nvkm_gsp_new_(gb202_gsps, device, type, inst, pgsp); +} + +NVKM_GSP_FIRMWARE_FMC(gb202, 570.144); +NVKM_GSP_FIRMWARE_FMC(gb203, 570.144); +NVKM_GSP_FIRMWARE_FMC(gb205, 570.144); +NVKM_GSP_FIRMWARE_FMC(gb206, 570.144); +NVKM_GSP_FIRMWARE_FMC(gb207, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gh100.c new file mode 100644 index 000000000000..ce31e8248807 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/gh100.c @@ -0,0 +1,358 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <linux/elf.h> +#include <linux/crc32.h> + +#include <subdev/fb.h> +#include <subdev/fsp.h> + +#include <rm/r570/nvrm/gsp.h> + +#include <nvhw/drf.h> +#include <nvhw/ref/gh100/dev_falcon_v4.h> +#include <nvhw/ref/gh100/dev_riscv_pri.h> + +int +gh100_gsp_fini(struct nvkm_gsp *gsp, bool suspend) +{ + struct nvkm_falcon *falcon = &gsp->falcon; + int ret, time = 4000; + + /* Shutdown RM. */ + ret = r535_gsp_fini(gsp, suspend); + if (ret && suspend) + return ret; + + /* Wait for RISC-V to halt. */ + do { + u32 data = nvkm_falcon_rd32(falcon, falcon->addr2 + NV_PRISCV_RISCV_CPUCTL); + + if (NVVAL_GET(data, NV_PRISCV, RISCV_CPUCTL, HALTED)) + return 0; + + usleep_range(1000, 2000); + } while(time--); + + return -ETIMEDOUT; +} + +static bool +gh100_gsp_lockdown_released(struct nvkm_gsp *gsp, u32 *mbox0) +{ + u32 data; + + /* Wait for GSP access via BAR0 to be allowed. */ + *mbox0 = nvkm_falcon_rd32(&gsp->falcon, NV_PFALCON_FALCON_MAILBOX0); + + if (*mbox0 && (*mbox0 & 0xffffff00) == 0xbadf4100) + return false; + + /* Check if an error code has been reported. */ + if (*mbox0) { + u32 mbox1 = nvkm_falcon_rd32(&gsp->falcon, NV_PFALCON_FALCON_MAILBOX1); + + /* Any value that's not GSP_FMC_BOOT_PARAMS addr is an error. */ + if ((((u64)mbox1 << 32) | *mbox0) != gsp->fmc.args.addr) + return true; + } + + /* Check if lockdown has been released. */ + data = nvkm_falcon_rd32(&gsp->falcon, NV_PFALCON_FALCON_HWCFG2); + return !NVVAL_GET(data, NV_PFALCON, FALCON_HWCFG2, RISCV_BR_PRIV_LOCKDOWN); +} + +int +gh100_gsp_init(struct nvkm_gsp *gsp) +{ + struct nvkm_subdev *subdev = &gsp->subdev; + struct nvkm_device *device = subdev->device; + const bool resume = gsp->sr.meta.data != NULL; + struct nvkm_gsp_mem *meta; + GSP_FMC_BOOT_PARAMS *args; + int ret, time = 4000; + u32 rsvd_size; + u32 mbox0; + + if (!resume) { + ret = nvkm_gsp_mem_ctor(gsp, sizeof(*args), &gsp->fmc.args); + if (ret) + return ret; + + meta = &gsp->wpr_meta; + } else { + gsp->rm->api->gsp->set_rmargs(gsp, true); + meta = &gsp->sr.meta; + } + + args = gsp->fmc.args.data; + + args->bootGspRmParams.gspRmDescOffset = meta->addr; + args->bootGspRmParams.gspRmDescSize = meta->size; + args->bootGspRmParams.target = GSP_DMA_TARGET_COHERENT_SYSTEM; + args->bootGspRmParams.bIsGspRmBoot = 1; + + args->gspRmParams.target = GSP_DMA_TARGET_NONCOHERENT_SYSTEM; + args->gspRmParams.bootArgsOffset = gsp->libos.addr; + + rsvd_size = gsp->fb.heap.size; + if (gsp->rm->wpr->rsvd_size_pmu) + rsvd_size = ALIGN(rsvd_size + gsp->rm->wpr->rsvd_size_pmu, 0x200000); + + ret = nvkm_fsp_boot_gsp_fmc(device->fsp, gsp->fmc.args.addr, rsvd_size, resume, + gsp->fmc.fw.addr, gsp->fmc.hash, gsp->fmc.pkey, gsp->fmc.sig); + if (ret) + return ret; + + do { + if (gh100_gsp_lockdown_released(gsp, &mbox0)) + break; + + usleep_range(1000, 2000); + } while(time--); + + if (time < 0) { + nvkm_error(subdev, "GSP-FMC boot timed out\n"); + return -ETIMEDOUT; + } + + if (mbox0) { + nvkm_error(subdev, "GSP-FMC boot failed (mbox: 0x%08x)\n", mbox0); + return -EIO; + } + + return r535_gsp_init(gsp); +} + +static int +gh100_gsp_wpr_meta_init(struct nvkm_gsp *gsp) +{ + GspFwWprMeta *meta; + int ret; + + ret = nvkm_gsp_mem_ctor(gsp, sizeof(*meta), &gsp->wpr_meta); + if (ret) + return ret; + + gsp->fb.size = nvkm_fb_vidmem_size(gsp->subdev.device); + gsp->fb.bios.vga_workspace.size = 128 * 1024; + gsp->fb.heap.size = gsp->rm->wpr->heap_size_non_wpr; + + meta = gsp->wpr_meta.data; + + meta->magic = GSP_FW_WPR_META_MAGIC; + meta->revision = GSP_FW_WPR_META_REVISION; + + meta->sizeOfRadix3Elf = gsp->fw.len; + meta->sysmemAddrOfRadix3Elf = gsp->radix3.lvl0.addr; + + meta->sizeOfBootloader = gsp->boot.fw.size; + meta->sysmemAddrOfBootloader = gsp->boot.fw.addr; + meta->bootloaderCodeOffset = gsp->boot.code_offset; + meta->bootloaderDataOffset = gsp->boot.data_offset; + meta->bootloaderManifestOffset = gsp->boot.manifest_offset; + + meta->sysmemAddrOfSignature = gsp->sig.addr; + meta->sizeOfSignature = gsp->sig.size; + + meta->nonWprHeapSize = gsp->fb.heap.size; + meta->gspFwHeapSize = tu102_gsp_wpr_heap_size(gsp); + meta->frtsSize = 0x100000; + meta->vgaWorkspaceSize = gsp->fb.bios.vga_workspace.size; + meta->pmuReservedSize = gsp->rm->wpr->rsvd_size_pmu; + return 0; +} + +/* The sh_flags value for the binary blobs in the ELF image */ +#define FMC_SHF_FLAGS (SHF_MASKPROC | SHF_MASKOS | SHF_OS_NONCONFORMING | SHF_ALLOC) + +#define ELF_HDR_SIZE ((u8)sizeof(struct elf32_hdr)) +#define ELF_SHDR_SIZE ((u8)sizeof(struct elf32_shdr)) + +/* The FMC ELF header must be exactly this */ +static const u8 elf_header[] = { + 0x7f, 'E', 'L', 'F', 1, 1, 1, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + + 0, 0, 0, 0, 1, 0, 0, 0, /* e_type, e_machine, e_version */ + 0, 0, 0, 0, 0, 0, 0, 0, /* e_entry, e_phoff */ + + ELF_HDR_SIZE, 0, 0, 0, 0, 0, 0, 0, /* e_shoff, e_flags */ + ELF_HDR_SIZE, 0, 0, 0, /* e_ehsize, e_phentsize */ + 0, 0, ELF_SHDR_SIZE, 0, /* e_phnum, e_shentsize */ + + 6, 0, 1, 0, /* e_shnum, e_shstrndx */ +}; + +/** + * elf_validate_sections - validate each section in the FMC ELF image + * @elf: ELF image + * @length: size of the entire ELF image + */ +static bool +elf_validate_sections(const void *elf, size_t length) +{ + const struct elf32_hdr *ehdr = elf; + const struct elf32_shdr *shdr = elf + ehdr->e_shoff; + + /* The offset of the first section */ + Elf32_Off section_begin = ehdr->e_shoff + ehdr->e_shnum * ehdr->e_shentsize; + + if (section_begin > length) + return false; + + /* The first section header is the null section, so skip it */ + for (unsigned int i = 1; i < ehdr->e_shnum; i++) { + if (i == ehdr->e_shstrndx) { + if (shdr[i].sh_type != SHT_STRTAB) + return false; + if (shdr[i].sh_flags != SHF_STRINGS) + return false; + } else { + if (shdr[i].sh_type != SHT_PROGBITS) + return false; + if (shdr[i].sh_flags != FMC_SHF_FLAGS) + return false; + } + + /* Ensure that each section is inside the image */ + if (shdr[i].sh_offset < section_begin || + (u64)shdr[i].sh_offset + shdr[i].sh_size > length) + return false; + + /* Non-zero sh_info is a CRC */ + if (shdr[i].sh_info) { + /* The kernel's CRC32 needs a pre- and post-xor to match standard CRCs */ + u32 crc32 = crc32_le(~0, elf + shdr[i].sh_offset, shdr[i].sh_size) ^ ~0; + + if (shdr[i].sh_info != crc32) + return false; + } + } + + return true; +} + +/** + * elf_section - return a pointer to the data for a given section + * @elf: ELF image + * @name: section name to search for + * @len: pointer to returned length of found section + */ +static const void * +elf_section(const void *elf, const char *name, unsigned int *len) +{ + const struct elf32_hdr *ehdr = elf; + const struct elf32_shdr *shdr = elf + ehdr->e_shoff; + const char *names = elf + shdr[ehdr->e_shstrndx].sh_offset; + + for (unsigned int i = 1; i < ehdr->e_shnum; i++) { + if (!strcmp(&names[shdr[i].sh_name], name)) { + *len = shdr[i].sh_size; + return elf + shdr[i].sh_offset; + } + } + + return NULL; +} + +int +gh100_gsp_oneinit(struct nvkm_gsp *gsp) +{ + struct nvkm_subdev *subdev = &gsp->subdev; + struct nvkm_device *device = subdev->device; + struct nvkm_fsp *fsp = device->fsp; + const void *fw = gsp->fws.fmc->data; + const void *hash, *sig, *pkey, *img; + unsigned int img_len = 0, hash_len = 0, pkey_len = 0, sig_len = 0; + int ret; + + if (gsp->fws.fmc->size < ELF_HDR_SIZE || + memcmp(fw, elf_header, sizeof(elf_header)) || + !elf_validate_sections(fw, gsp->fws.fmc->size)) { + nvkm_error(subdev, "fmc firmware image is invalid\n"); + return -ENODATA; + } + + hash = elf_section(fw, "hash", &hash_len); + sig = elf_section(fw, "signature", &sig_len); + pkey = elf_section(fw, "publickey", &pkey_len); + img = elf_section(fw, "image", &img_len); + + if (!hash || !sig || !pkey || !img) { + nvkm_error(subdev, "fmc firmware image is invalid\n"); + return -ENODATA; + } + + if (!nvkm_fsp_verify_gsp_fmc(fsp, hash_len, pkey_len, sig_len)) + return -EINVAL; + + /* Load GSP-FMC FW into memory. */ + ret = nvkm_gsp_mem_ctor(gsp, img_len, &gsp->fmc.fw); + if (ret) + return ret; + + memcpy(gsp->fmc.fw.data, img, img_len); + + gsp->fmc.hash = kmemdup(hash, hash_len, GFP_KERNEL); + gsp->fmc.pkey = kmemdup(pkey, pkey_len, GFP_KERNEL); + gsp->fmc.sig = kmemdup(sig, sig_len, GFP_KERNEL); + if (!gsp->fmc.hash || !gsp->fmc.pkey || !gsp->fmc.sig) + return -ENOMEM; + + ret = r535_gsp_oneinit(gsp); + if (ret) + return ret; + + return gh100_gsp_wpr_meta_init(gsp); +} + +static const struct nvkm_gsp_func +gh100_gsp = { + .flcn = &ga102_gsp_flcn, + + .sig_section = ".fwsignature_gh100", + + .dtor = r535_gsp_dtor, + .oneinit = gh100_gsp_oneinit, + .init = gh100_gsp_init, + .fini = gh100_gsp_fini, + + .rm.gpu = &gh100_gpu, +}; + +int +gh100_gsp_load(struct nvkm_gsp *gsp, int ver, const struct nvkm_gsp_fwif *fwif) +{ + int ret; + + ret = tu102_gsp_load_rm(gsp, fwif); + if (ret) + goto done; + + ret = nvkm_gsp_load_fw(gsp, "fmc", fwif->ver, &gsp->fws.fmc); + +done: + if (ret) + nvkm_gsp_dtor_fws(gsp); + + return ret; +} + +static struct nvkm_gsp_fwif +gh100_gsps[] = { + { 0, gh100_gsp_load, &gh100_gsp, &r570_rm_gh100, "570.144", true }, + {} +}; + +int +gh100_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_gsp **pgsp) +{ + return nvkm_gsp_new_(gh100_gsps, device, type, inst, pgsp); +} + +NVKM_GSP_FIRMWARE_FMC(gh100, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h index 9f4a62375a27..4f14e85fc69e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/priv.h @@ -2,6 +2,7 @@ #ifndef __NVKM_GSP_PRIV_H__ #define __NVKM_GSP_PRIV_H__ #include <subdev/gsp.h> +#include <rm/gpu.h> enum nvkm_acr_lsf_id; int nvkm_gsp_fwsec_frts(struct nvkm_gsp *); @@ -11,12 +12,32 @@ struct nvkm_gsp_fwif { int version; int (*load)(struct nvkm_gsp *, int ver, const struct nvkm_gsp_fwif *); const struct nvkm_gsp_func *func; + const struct nvkm_rm_impl *rm; const char *ver; bool enable; }; +int nvkm_gsp_load_fw(struct nvkm_gsp *, const char *name, const char *ver, + const struct firmware **); +void nvkm_gsp_dtor_fws(struct nvkm_gsp *); + int gv100_gsp_nofw(struct nvkm_gsp *, int, const struct nvkm_gsp_fwif *); -int r535_gsp_load(struct nvkm_gsp *, int, const struct nvkm_gsp_fwif *); + +int tu102_gsp_load(struct nvkm_gsp *, int, const struct nvkm_gsp_fwif *); +int tu102_gsp_load_rm(struct nvkm_gsp *, const struct nvkm_gsp_fwif *); + +int gh100_gsp_load(struct nvkm_gsp *, int, const struct nvkm_gsp_fwif *); + +#define NVKM_GSP_FIRMWARE_BOOTER(chip,vers) \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/booter_load-"#vers".bin"); \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/booter_unload-"#vers".bin"); \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/bootloader-"#vers".bin"); \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/gsp-"#vers".bin") + +#define NVKM_GSP_FIRMWARE_FMC(chip,vers) \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/fmc-"#vers".bin"); \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/bootloader-"#vers".bin"); \ +MODULE_FIRMWARE("nvidia/"#chip"/gsp/gsp-"#vers".bin") struct nvkm_gsp_func { const struct nvkm_falcon_func *flcn; @@ -25,12 +46,6 @@ struct nvkm_gsp_func { char *sig_section; struct { - u32 os_carveout_size; - u32 base_size; - u64 min_size; - } wpr_heap; - - struct { int (*ctor)(struct nvkm_gsp *, const char *name, const struct firmware *, struct nvkm_falcon *, struct nvkm_falcon_fw *); } booter; @@ -41,7 +56,9 @@ struct nvkm_gsp_func { int (*fini)(struct nvkm_gsp *, bool suspend); int (*reset)(struct nvkm_gsp *); - const struct nvkm_gsp_rm *rm; + struct { + const struct nvkm_rm_gpu *gpu; + } rm; }; extern const struct nvkm_falcon_func tu102_gsp_flcn; @@ -49,7 +66,10 @@ extern const struct nvkm_falcon_fw_func tu102_gsp_fwsec; int tu102_gsp_booter_ctor(struct nvkm_gsp *, const char *, const struct firmware *, struct nvkm_falcon *, struct nvkm_falcon_fw *); int tu102_gsp_oneinit(struct nvkm_gsp *); +int tu102_gsp_init(struct nvkm_gsp *); +int tu102_gsp_fini(struct nvkm_gsp *, bool suspend); int tu102_gsp_reset(struct nvkm_gsp *); +u64 tu102_gsp_wpr_heap_size(struct nvkm_gsp *); extern const struct nvkm_falcon_func ga102_gsp_flcn; extern const struct nvkm_falcon_fw_func ga102_gsp_fwsec; @@ -57,11 +77,14 @@ int ga102_gsp_booter_ctor(struct nvkm_gsp *, const char *, const struct firmware struct nvkm_falcon *, struct nvkm_falcon_fw *); int ga102_gsp_reset(struct nvkm_gsp *); +int gh100_gsp_oneinit(struct nvkm_gsp *); +int gh100_gsp_init(struct nvkm_gsp *); +int gh100_gsp_fini(struct nvkm_gsp *, bool suspend); + void r535_gsp_dtor(struct nvkm_gsp *); int r535_gsp_oneinit(struct nvkm_gsp *); int r535_gsp_init(struct nvkm_gsp *); int r535_gsp_fini(struct nvkm_gsp *, bool suspend); -extern const struct nvkm_gsp_rm r535_gsp_rm; int nvkm_gsp_new_(const struct nvkm_gsp_fwif *, struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_gsp **); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/Kbuild new file mode 100644 index 000000000000..04037394a2da --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/Kbuild @@ -0,0 +1,19 @@ +# SPDX-License-Identifier: MIT +# +# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. +nvkm-y += nvkm/subdev/gsp/rm/client.o +nvkm-y += nvkm/subdev/gsp/rm/engine.o +nvkm-y += nvkm/subdev/gsp/rm/gr.o +nvkm-y += nvkm/subdev/gsp/rm/nvdec.o +nvkm-y += nvkm/subdev/gsp/rm/nvenc.o + +nvkm-y += nvkm/subdev/gsp/rm/tu1xx.o +nvkm-y += nvkm/subdev/gsp/rm/ga100.o +nvkm-y += nvkm/subdev/gsp/rm/ga1xx.o +nvkm-y += nvkm/subdev/gsp/rm/ad10x.o +nvkm-y += nvkm/subdev/gsp/rm/gh100.o +nvkm-y += nvkm/subdev/gsp/rm/gb10x.o +nvkm-y += nvkm/subdev/gsp/rm/gb20x.o + +include $(src)/nvkm/subdev/gsp/rm/r535/Kbuild +include $(src)/nvkm/subdev/gsp/rm/r570/Kbuild diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ad10x.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ad10x.c new file mode 100644 index 000000000000..e1ce6355c35f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ad10x.c @@ -0,0 +1,39 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +ad10x_gpu = { + .disp.class = { + .root = AD102_DISP, + .caps = GV100_DISP_CAPS, + .core = AD102_DISP_CORE_CHANNEL_DMA, + .wndw = GA102_DISP_WINDOW_CHANNEL_DMA, + .wimm = GA102_DISP_WINDOW_IMM_CHANNEL_DMA, + .curs = GA102_DISP_CURSOR, + }, + + .usermode.class = AMPERE_USERMODE_A, + + .fifo.chan = { + .class = AMPERE_CHANNEL_GPFIFO_A, + .doorbell_handle = tu102_chan_doorbell_handle, + }, + + .ce.class = AMPERE_DMA_COPY_B, + .gr.class = { + .i2m = KEPLER_INLINE_TO_MEMORY_B, + .twod = FERMI_TWOD_A, + .threed = ADA_A, + .compute = ADA_COMPUTE_A, + }, + .nvdec.class = NVC9B0_VIDEO_DECODER, + .nvenc.class = NVC9B7_VIDEO_ENCODER, + .ofa.class = NVC9FA_VIDEO_OFA, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/client.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/client.c new file mode 100644 index 000000000000..72d3e3ca84c2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/client.c @@ -0,0 +1,49 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "rm.h" + +void +nvkm_gsp_client_dtor(struct nvkm_gsp_client *client) +{ + const unsigned int id = client->object.handle - NVKM_RM_CLIENT(0); + struct nvkm_gsp *gsp = client->gsp; + + if (!gsp) + return; + + if (client->object.client) + nvkm_gsp_rm_free(&client->object); + + mutex_lock(&gsp->client_id.mutex); + idr_remove(&gsp->client_id.idr, id); + mutex_unlock(&gsp->client_id.mutex); + + client->gsp = NULL; +} + +int +nvkm_gsp_client_ctor(struct nvkm_gsp *gsp, struct nvkm_gsp_client *client) +{ + int id, ret; + + if (WARN_ON(!gsp->rm)) + return -ENOSYS; + + mutex_lock(&gsp->client_id.mutex); + id = idr_alloc(&gsp->client_id.idr, client, 0, NVKM_RM_CLIENT_MASK + 1, GFP_KERNEL); + mutex_unlock(&gsp->client_id.mutex); + if (id < 0) + return id; + + client->gsp = gsp; + client->object.client = client; + INIT_LIST_HEAD(&client->events); + + ret = gsp->rm->api->client->ctor(client, NVKM_RM_CLIENT(id)); + if (ret) + nvkm_gsp_client_dtor(client); + + return ret; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/engine.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/engine.c new file mode 100644 index 000000000000..3b0e83b2f57f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/engine.c @@ -0,0 +1,189 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "engine.h" +#include "gpu.h" + +#include <core/object.h> +#include <engine/fifo/chan.h> + +struct nvkm_rm_engine { + struct nvkm_engine engine; + + struct nvkm_engine_func func; +}; + +struct nvkm_rm_engine_obj { + struct nvkm_object object; + struct nvkm_gsp_object rm; +}; + +static void* +nvkm_rm_engine_obj_dtor(struct nvkm_object *object) +{ + struct nvkm_rm_engine_obj *obj = container_of(object, typeof(*obj), object); + + nvkm_gsp_rm_free(&obj->rm); + return obj; +} + +static const struct nvkm_object_func +nvkm_rm_engine_obj = { + .dtor = nvkm_rm_engine_obj_dtor, +}; + +int +nvkm_rm_engine_obj_new(struct nvkm_gsp_object *chan, int chid, const struct nvkm_oclass *oclass, + struct nvkm_object **pobject) +{ + struct nvkm_rm *rm = chan->client->gsp->rm; + const int inst = oclass->engine->subdev.inst; + const u32 class = oclass->base.oclass; + const u32 handle = oclass->handle; + struct nvkm_rm_engine_obj *obj; + int ret; + + obj = kzalloc(sizeof(*obj), GFP_KERNEL); + if (!obj) + return -ENOMEM; + + switch (oclass->engine->subdev.type) { + case NVKM_ENGINE_CE: + ret = rm->api->ce->alloc(chan, handle, class, inst, &obj->rm); + break; + case NVKM_ENGINE_GR: + ret = nvkm_gsp_rm_alloc(chan, handle, class, 0, &obj->rm); + break; + case NVKM_ENGINE_NVDEC: + ret = rm->api->nvdec->alloc(chan, handle, class, inst, &obj->rm); + break; + case NVKM_ENGINE_NVENC: + ret = rm->api->nvenc->alloc(chan, handle, class, inst, &obj->rm); + break; + case NVKM_ENGINE_NVJPG: + ret = rm->api->nvjpg->alloc(chan, handle, class, inst, &obj->rm); + break; + case NVKM_ENGINE_OFA: + ret = rm->api->ofa->alloc(chan, handle, class, inst, &obj->rm); + break; + default: + ret = -EINVAL; + WARN_ON(1); + break; + } + + if (ret) { + kfree(obj); + return ret; + } + + nvkm_object_ctor(&nvkm_rm_engine_obj, oclass, &obj->object); + *pobject = &obj->object; + return 0; +} + +static int +nvkm_rm_engine_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, + struct nvkm_object **pobject) +{ + struct nvkm_chan *chan = nvkm_uchan_chan(oclass->parent); + + return nvkm_rm_engine_obj_new(&chan->rm.object, chan->id, oclass, pobject); +} + +static void * +nvkm_rm_engine_dtor(struct nvkm_engine *engine) +{ + kfree(engine->func); + return engine; +} + +int +nvkm_rm_engine_ctor(void *(*dtor)(struct nvkm_engine *), struct nvkm_rm *rm, + enum nvkm_subdev_type type, int inst, + const u32 *class, int nclass, struct nvkm_engine *engine) +{ + struct nvkm_engine_func *func; + + func = kzalloc(struct_size(func, sclass, nclass + 1), GFP_KERNEL); + if (!func) + return -ENOMEM; + + func->dtor = dtor; + + for (int i = 0; i < nclass; i++) { + func->sclass[i].oclass = class[i]; + func->sclass[i].minver = -1; + func->sclass[i].maxver = 0; + func->sclass[i].ctor = nvkm_rm_engine_obj_ctor; + } + + nvkm_engine_ctor(func, rm->device, type, inst, true, engine); + return 0; +} + +static int +nvkm_rm_engine_new_(struct nvkm_rm *rm, enum nvkm_subdev_type type, int inst, u32 class, + struct nvkm_engine **pengine) +{ + struct nvkm_engine *engine; + int ret; + + engine = kzalloc(sizeof(*engine), GFP_KERNEL); + if (!engine) + return -ENOMEM; + + ret = nvkm_rm_engine_ctor(nvkm_rm_engine_dtor, rm, type, inst, &class, 1, engine); + if (ret) { + kfree(engine); + return ret; + } + + *pengine = engine; + return 0; +} + +int +nvkm_rm_engine_new(struct nvkm_rm *rm, enum nvkm_subdev_type type, int inst) +{ + const struct nvkm_rm_gpu *gpu = rm->gpu; + struct nvkm_device *device = rm->device; + + switch (type) { + case NVKM_ENGINE_CE: + if (WARN_ON(inst >= ARRAY_SIZE(device->ce))) + return -EINVAL; + + return nvkm_rm_engine_new_(rm, type, inst, gpu->ce.class, &device->ce[inst]); + case NVKM_ENGINE_GR: + if (inst != 0) + return -ENODEV; /* MiG not supported, just ignore. */ + + return nvkm_rm_gr_new(rm); + case NVKM_ENGINE_NVDEC: + if (WARN_ON(inst >= ARRAY_SIZE(device->nvdec))) + return -EINVAL; + + return nvkm_rm_nvdec_new(rm, inst); + case NVKM_ENGINE_NVENC: + if (WARN_ON(inst >= ARRAY_SIZE(device->nvenc))) + return -EINVAL; + + return nvkm_rm_nvenc_new(rm, inst); + case NVKM_ENGINE_NVJPG: + if (WARN_ON(inst >= ARRAY_SIZE(device->nvjpg))) + return -EINVAL; + + return nvkm_rm_engine_new_(rm, type, inst, gpu->nvjpg.class, &device->nvjpg[inst]); + case NVKM_ENGINE_OFA: + if (WARN_ON(inst >= ARRAY_SIZE(device->ofa))) + return -EINVAL; + + return nvkm_rm_engine_new_(rm, type, inst, gpu->ofa.class, &device->ofa[inst]); + default: + break; + } + + return -ENODEV; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/engine.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/engine.h new file mode 100644 index 000000000000..5b8c9c3901d4 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/engine.h @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVKM_RM_ENGINE_H__ +#define __NVKM_RM_ENGINE_H__ +#include "gpu.h" + +int nvkm_rm_engine_ctor(void *(*dtor)(struct nvkm_engine *), struct nvkm_rm *, + enum nvkm_subdev_type type, int inst, + const u32 *class, int nclass, struct nvkm_engine *); +int nvkm_rm_engine_new(struct nvkm_rm *, enum nvkm_subdev_type, int inst); + +int nvkm_rm_engine_obj_new(struct nvkm_gsp_object *chan, int chid, const struct nvkm_oclass *, + struct nvkm_object **); + +int nvkm_rm_gr_new(struct nvkm_rm *); +int nvkm_rm_nvdec_new(struct nvkm_rm *, int inst); +int nvkm_rm_nvenc_new(struct nvkm_rm *, int inst); +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ga100.c new file mode 100644 index 000000000000..a48c6134075d --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ga100.c @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +ga100_gpu = { + .usermode.class = AMPERE_USERMODE_A, + + .fifo.chan = { + .class = AMPERE_CHANNEL_GPFIFO_A, + .doorbell_handle = tu102_chan_doorbell_handle, + }, + + .ce.class = AMPERE_DMA_COPY_A, + .gr.class = { + .i2m = KEPLER_INLINE_TO_MEMORY_B, + .twod = FERMI_TWOD_A, + .threed = AMPERE_A, + .compute = AMPERE_COMPUTE_A, + }, + .nvdec.class = NVC6B0_VIDEO_DECODER, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ga1xx.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ga1xx.c new file mode 100644 index 000000000000..50536ad7f85d --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/ga1xx.c @@ -0,0 +1,39 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +ga1xx_gpu = { + .disp.class = { + .root = GA102_DISP, + .caps = GV100_DISP_CAPS, + .core = GA102_DISP_CORE_CHANNEL_DMA, + .wndw = GA102_DISP_WINDOW_CHANNEL_DMA, + .wimm = GA102_DISP_WINDOW_IMM_CHANNEL_DMA, + .curs = GA102_DISP_CURSOR, + }, + + .usermode.class = AMPERE_USERMODE_A, + + .fifo.chan = { + .class = AMPERE_CHANNEL_GPFIFO_A, + .doorbell_handle = tu102_chan_doorbell_handle, + }, + + .ce.class = AMPERE_DMA_COPY_B, + .gr.class = { + .i2m = KEPLER_INLINE_TO_MEMORY_B, + .twod = FERMI_TWOD_A, + .threed = AMPERE_B, + .compute = AMPERE_COMPUTE_B, + }, + .nvdec.class = NVC7B0_VIDEO_DECODER, + .nvenc.class = NVC7B7_VIDEO_ENCODER, + .ofa.class = NVC7FA_VIDEO_OFA, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gb10x.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gb10x.c new file mode 100644 index 000000000000..2f517dcd721a --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gb10x.c @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +gb10x_gpu = { + .usermode.class = HOPPER_USERMODE_A, + + .fifo.chan = { + .class = BLACKWELL_CHANNEL_GPFIFO_A, + .doorbell_handle = tu102_chan_doorbell_handle, + }, + + .ce.class = BLACKWELL_DMA_COPY_A, + .gr.class = { + .i2m = BLACKWELL_INLINE_TO_MEMORY_A, + .twod = FERMI_TWOD_A, + .threed = BLACKWELL_A, + .compute = BLACKWELL_COMPUTE_A, + }, + .nvdec.class = NVCDB0_VIDEO_DECODER, + .nvjpg.class = NVCDD1_VIDEO_NVJPG, + .ofa.class = NVCDFA_VIDEO_OFA, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gb20x.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gb20x.c new file mode 100644 index 000000000000..950471d9996e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gb20x.c @@ -0,0 +1,44 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/ce/priv.h> +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +gb20x_gpu = { + .disp.class = { + .root = GB202_DISP, + .caps = GB202_DISP_CAPS, + .core = GB202_DISP_CORE_CHANNEL_DMA, + .wndw = GB202_DISP_WINDOW_CHANNEL_DMA, + .wimm = GB202_DISP_WINDOW_IMM_CHANNEL_DMA, + .curs = GB202_DISP_CURSOR, + }, + + .usermode.class = BLACKWELL_USERMODE_A, + + .fifo.chan = { + .class = BLACKWELL_CHANNEL_GPFIFO_B, + .doorbell_handle = gb202_chan_doorbell_handle, + }, + + .ce = { + .class = BLACKWELL_DMA_COPY_B, + .grce_mask = gb202_ce_grce_mask, + }, + .gr.class = { + .i2m = BLACKWELL_INLINE_TO_MEMORY_A, + .twod = FERMI_TWOD_A, + .threed = BLACKWELL_B, + .compute = BLACKWELL_COMPUTE_B, + }, + .nvdec.class = NVCFB0_VIDEO_DECODER, + .nvenc.class = NVCFB7_VIDEO_ENCODER, + .nvjpg.class = NVCFD1_VIDEO_NVJPG, + .ofa.class = NVCFFA_VIDEO_OFA, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gh100.c new file mode 100644 index 000000000000..49e2c54e1aa8 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gh100.c @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +gh100_gpu = { + .usermode.class = HOPPER_USERMODE_A, + + .fifo.chan = { + .class = HOPPER_CHANNEL_GPFIFO_A, + .doorbell_handle = tu102_chan_doorbell_handle, + }, + + .ce.class = HOPPER_DMA_COPY_A, + .gr.class = { + .i2m = KEPLER_INLINE_TO_MEMORY_B, + .twod = FERMI_TWOD_A, + .threed = HOPPER_A, + .compute = HOPPER_COMPUTE_A, + }, + .nvdec.class = NVB8B0_VIDEO_DECODER, + .nvjpg.class = NVB8D1_VIDEO_NVJPG, + .ofa.class = NVB8FA_VIDEO_OFA, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gpu.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gpu.h new file mode 100644 index 000000000000..46a6325641b7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gpu.h @@ -0,0 +1,70 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVKM_RM_GPU_H__ +#define __NVKM_RM_GPU_H__ +#include "rm.h" + +struct nvkm_rm_gpu { + struct { + struct { + u32 root; + u32 caps; + u32 core; + u32 wndw; + u32 wimm; + u32 curs; + } class; + } disp; + + struct { + u32 class; + } usermode; + + struct { + struct { + u32 class; + u32 (*doorbell_handle)(struct nvkm_chan *); + } chan; + } fifo; + + struct { + u32 class; + u32 (*grce_mask)(struct nvkm_device *); + } ce; + + struct { + struct { + u32 i2m; + u32 twod; + u32 threed; + u32 compute; + } class; + } gr; + + struct { + u32 class; + } nvdec; + + struct { + u32 class; + } nvenc; + + struct { + u32 class; + } nvjpg; + + struct { + u32 class; + } ofa; +}; + +extern const struct nvkm_rm_gpu tu1xx_gpu; +extern const struct nvkm_rm_gpu ga100_gpu; +extern const struct nvkm_rm_gpu ga1xx_gpu; +extern const struct nvkm_rm_gpu ad10x_gpu; +extern const struct nvkm_rm_gpu gh100_gpu; +extern const struct nvkm_rm_gpu gb10x_gpu; +extern const struct nvkm_rm_gpu gb20x_gpu; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gr.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gr.c new file mode 100644 index 000000000000..f40b8fcc2bcb --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gr.c @@ -0,0 +1,87 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gr.h" + +#include <engine/fifo.h> +#include <engine/gr/priv.h> + +static int +nvkm_rm_gr_obj_ctor(const struct nvkm_oclass *oclass, void *argv, u32 argc, + struct nvkm_object **pobject) +{ + struct r535_gr_chan *chan = container_of(oclass->parent, typeof(*chan), object); + + return nvkm_rm_engine_obj_new(&chan->chan->rm.object, chan->chan->id, oclass, pobject); +} + +static int +nvkm_rm_gr_fini(struct nvkm_gr *base, bool suspend) +{ + struct nvkm_rm *rm = base->engine.subdev.device->gsp->rm; + struct r535_gr *gr = container_of(base, typeof(*gr), base); + + if (rm->api->gr->scrubber.fini) + rm->api->gr->scrubber.fini(gr); + + return 0; +} + +static int +nvkm_rm_gr_init(struct nvkm_gr *base) +{ + struct nvkm_rm *rm = base->engine.subdev.device->gsp->rm; + struct r535_gr *gr = container_of(base, typeof(*gr), base); + int ret; + + if (rm->api->gr->scrubber.init) { + ret = rm->api->gr->scrubber.init(gr); + if (ret) + return ret; + } + + return 0; +} + +int +nvkm_rm_gr_new(struct nvkm_rm *rm) +{ + const u32 classes[] = { + rm->gpu->gr.class.i2m, + rm->gpu->gr.class.twod, + rm->gpu->gr.class.threed, + rm->gpu->gr.class.compute, + }; + struct nvkm_gr_func *func; + struct r535_gr *gr; + + func = kzalloc(struct_size(func, sclass, ARRAY_SIZE(classes) + 1), GFP_KERNEL); + if (!func) + return -ENOMEM; + + func->dtor = r535_gr_dtor; + func->oneinit = r535_gr_oneinit; + func->init = nvkm_rm_gr_init; + func->fini = nvkm_rm_gr_fini; + func->units = r535_gr_units; + func->chan_new = r535_gr_chan_new; + + for (int i = 0; i < ARRAY_SIZE(classes); i++) { + func->sclass[i].oclass = classes[i]; + func->sclass[i].minver = -1; + func->sclass[i].maxver = 0; + func->sclass[i].ctor = nvkm_rm_gr_obj_ctor; + } + + gr = kzalloc(sizeof(*gr), GFP_KERNEL); + if (!gr) { + kfree(func); + return -ENOMEM; + } + + nvkm_gr_ctor(func, rm->device, NVKM_ENGINE_GR, 0, true, &gr->base); + gr->scrubber.chid = -1; + rm->device->gr = &gr->base; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gr.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gr.h new file mode 100644 index 000000000000..24980f23aab9 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/gr.h @@ -0,0 +1,55 @@ +/* SPDX-License-Identifier: MIT */ +#ifndef __NVKM_RM_GR_H__ +#define __NVKM_RM_GR_H__ +#include "engine.h" + +#include <core/object.h> +#include <engine/gr.h> + +#define R515_GR_MAX_CTXBUFS 9 + +struct r535_gr_chan { + struct nvkm_object object; + struct r535_gr *gr; + + struct nvkm_vmm *vmm; + struct nvkm_chan *chan; + + struct nvkm_memory *mem[R515_GR_MAX_CTXBUFS]; + struct nvkm_vma *vma[R515_GR_MAX_CTXBUFS]; +}; + +struct r535_gr { + struct nvkm_gr base; + + struct { + u16 bufferId; + u32 size; + u8 page; + u8 align; + bool global; + bool init; + bool ro; + } ctxbuf[R515_GR_MAX_CTXBUFS]; + int ctxbuf_nr; + + struct nvkm_memory *ctxbuf_mem[R515_GR_MAX_CTXBUFS]; + + struct { + int chid; + struct nvkm_memory *inst; + struct nvkm_vmm *vmm; + struct nvkm_gsp_object chan; + struct nvkm_gsp_object threed; + struct { + struct nvkm_memory *mem[R515_GR_MAX_CTXBUFS]; + struct nvkm_vma *vma[R515_GR_MAX_CTXBUFS]; + } ctxbuf; + bool enabled; + } scrubber; +}; + +struct NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO; +void r535_gr_get_ctxbuf_info(struct r535_gr *, int i, + struct NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO *); +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/handles.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/handles.h new file mode 100644 index 000000000000..3bdb5ad320d7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/handles.h @@ -0,0 +1,18 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVKM_RM_HANDLES_H__ +#define __NVKM_RM_HANDLES_H__ + +/* RMAPI handles for various objects allocated from GSP-RM with RM_ALLOC. */ + +#define NVKM_RM_CLIENT(id) (0xc1d00000 | (id)) +#define NVKM_RM_CLIENT_MASK 0x0000ffff +#define NVKM_RM_DEVICE 0xde1d0000 +#define NVKM_RM_SUBDEVICE 0x5d1d0000 +#define NVKM_RM_DISP 0x00730000 +#define NVKM_RM_VASPACE 0x90f10000 +#define NVKM_RM_CHAN(chid) (0xf1f00000 | (chid)) +#define NVKM_RM_THREED 0x97000000 +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/nvdec.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/nvdec.c new file mode 100644 index 000000000000..d9fbfc377864 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/nvdec.c @@ -0,0 +1,33 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "engine.h" +#include <engine/nvdec.h> + +static void * +nvkm_rm_nvdec_dtor(struct nvkm_engine *engine) +{ + return container_of(engine, struct nvkm_nvdec, engine); +} + +int +nvkm_rm_nvdec_new(struct nvkm_rm *rm, int inst) +{ + struct nvkm_nvdec *nvdec; + int ret; + + nvdec = kzalloc(sizeof(*nvdec), GFP_KERNEL); + if (!nvdec) + return -ENOMEM; + + ret = nvkm_rm_engine_ctor(nvkm_rm_nvdec_dtor, rm, NVKM_ENGINE_NVDEC, inst, + &rm->gpu->nvdec.class, 1, &nvdec->engine); + if (ret) { + kfree(nvdec); + return ret; + } + + rm->device->nvdec[inst] = nvdec; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/nvenc.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/nvenc.c new file mode 100644 index 000000000000..6dfa7b789e07 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/nvenc.c @@ -0,0 +1,33 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "engine.h" +#include <engine/nvenc.h> + +static void * +nvkm_rm_nvenc_dtor(struct nvkm_engine *engine) +{ + return container_of(engine, struct nvkm_nvenc, engine); +} + +int +nvkm_rm_nvenc_new(struct nvkm_rm *rm, int inst) +{ + struct nvkm_nvenc *nvenc; + int ret; + + nvenc = kzalloc(sizeof(*nvenc), GFP_KERNEL); + if (!nvenc) + return -ENOMEM; + + ret = nvkm_rm_engine_ctor(nvkm_rm_nvenc_dtor, rm, NVKM_ENGINE_NVENC, inst, + &rm->gpu->nvenc.class, 1, &nvenc->engine); + if (ret) { + kfree(nvenc); + return ret; + } + + rm->device->nvenc[inst] = nvenc; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/Kbuild new file mode 100644 index 000000000000..a5f6b2abfd33 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/Kbuild @@ -0,0 +1,25 @@ +# SPDX-License-Identifier: MIT +# +# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + +nvkm-y += nvkm/subdev/gsp/rm/r535/rm.o +nvkm-y += nvkm/subdev/gsp/rm/r535/gsp.o +nvkm-y += nvkm/subdev/gsp/rm/r535/rpc.o +nvkm-y += nvkm/subdev/gsp/rm/r535/ctrl.o +nvkm-y += nvkm/subdev/gsp/rm/r535/alloc.o +nvkm-y += nvkm/subdev/gsp/rm/r535/client.o +nvkm-y += nvkm/subdev/gsp/rm/r535/device.o + +nvkm-y += nvkm/subdev/gsp/rm/r535/bar.o +nvkm-y += nvkm/subdev/gsp/rm/r535/fbsr.o +nvkm-y += nvkm/subdev/gsp/rm/r535/vmm.o + +nvkm-y += nvkm/subdev/gsp/rm/r535/disp.o + +nvkm-y += nvkm/subdev/gsp/rm/r535/fifo.o +nvkm-y += nvkm/subdev/gsp/rm/r535/ce.o +nvkm-y += nvkm/subdev/gsp/rm/r535/gr.o +nvkm-y += nvkm/subdev/gsp/rm/r535/nvdec.o +nvkm-y += nvkm/subdev/gsp/rm/r535/nvenc.o +nvkm-y += nvkm/subdev/gsp/rm/r535/nvjpg.o +nvkm-y += nvkm/subdev/gsp/rm/r535/ofa.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/alloc.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/alloc.c new file mode 100644 index 000000000000..46e3a29f2ad7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/alloc.c @@ -0,0 +1,112 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/rpc.h> + +#include "nvrm/alloc.h" +#include "nvrm/rpcfn.h" + +static int +r535_gsp_rpc_rm_free(struct nvkm_gsp_object *object) +{ + struct nvkm_gsp_client *client = object->client; + struct nvkm_gsp *gsp = client->gsp; + rpc_free_v03_00 *rpc; + + nvkm_debug(&gsp->subdev, "cli:0x%08x obj:0x%08x free\n", + client->object.handle, object->handle); + + rpc = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_FREE, sizeof(*rpc)); + if (WARN_ON(IS_ERR_OR_NULL(rpc))) + return -EIO; + + rpc->params.hRoot = client->object.handle; + rpc->params.hObjectParent = 0; + rpc->params.hObjectOld = object->handle; + return nvkm_gsp_rpc_wr(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV); +} + +static void +r535_gsp_rpc_rm_alloc_done(struct nvkm_gsp_object *object, void *params) +{ + rpc_gsp_rm_alloc_v03_00 *rpc = to_payload_hdr(params, rpc); + + nvkm_gsp_rpc_done(object->client->gsp, rpc); +} + +static void * +r535_gsp_rpc_rm_alloc_push(struct nvkm_gsp_object *object, void *params) +{ + rpc_gsp_rm_alloc_v03_00 *rpc = to_payload_hdr(params, rpc); + struct nvkm_gsp *gsp = object->client->gsp; + void *ret = NULL; + + rpc = nvkm_gsp_rpc_push(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV, sizeof(*rpc)); + if (IS_ERR_OR_NULL(rpc)) + return rpc; + + if (rpc->status) { + ret = ERR_PTR(r535_rpc_status_to_errno(rpc->status)); + if (PTR_ERR(ret) != -EAGAIN && PTR_ERR(ret) != -EBUSY) + nvkm_error(&gsp->subdev, "RM_ALLOC: 0x%x\n", rpc->status); + } + + nvkm_gsp_rpc_done(gsp, rpc); + + return ret; +} + +static void * +r535_gsp_rpc_rm_alloc_get(struct nvkm_gsp_object *object, u32 oclass, + u32 params_size) +{ + struct nvkm_gsp_client *client = object->client; + struct nvkm_gsp *gsp = client->gsp; + rpc_gsp_rm_alloc_v03_00 *rpc; + + nvkm_debug(&gsp->subdev, "cli:0x%08x obj:0x%08x new obj:0x%08x\n", + client->object.handle, object->parent->handle, + object->handle); + + nvkm_debug(&gsp->subdev, "cls:0x%08x params_size:%d\n", oclass, + params_size); + + rpc = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_GSP_RM_ALLOC, + sizeof(*rpc) + params_size); + if (IS_ERR(rpc)) + return rpc; + + rpc->hClient = client->object.handle; + rpc->hParent = object->parent->handle; + rpc->hObject = object->handle; + rpc->hClass = oclass; + rpc->status = 0; + rpc->paramsSize = params_size; + return rpc->params; +} + +const struct nvkm_rm_api_alloc +r535_alloc = { + .get = r535_gsp_rpc_rm_alloc_get, + .push = r535_gsp_rpc_rm_alloc_push, + .done = r535_gsp_rpc_rm_alloc_done, + .free = r535_gsp_rpc_rm_free, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/bar.c index 90186f98065c..d06bf95b9a4a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/bar.c @@ -19,7 +19,7 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "gf100.h" +#include <subdev/bar/gf100.h> #include <core/mm.h> #include <subdev/fb.h> @@ -27,14 +27,20 @@ #include <subdev/instmem.h> #include <subdev/mmu/vmm.h> -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/nvidia/generated/g_rpc-structures.h> -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_global_enums.h> -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_headers.h> +#include "nvrm/bar.h" +#include "nvrm/rpcfn.h" static void r535_bar_flush(struct nvkm_bar *bar) { + /* Use NV_PFLUSH in resume path - needed on R570 to flush writes before + * BAR2 page tables have been restored. + */ + if (unlikely(!bar->bar2)) { + g84_bar_flush(bar); + return; + } + ioread32_native(bar->flushBAR2); } @@ -44,7 +50,7 @@ r535_bar_bar2_wait(struct nvkm_bar *base) } static int -r535_bar_bar2_update_pde(struct nvkm_gsp *gsp, u64 addr) +r535_bar_bar2_update_pde(struct nvkm_gsp *gsp, u8 page_shift, u64 pdbe) { rpc_update_bar_pde_v15_00 *rpc; @@ -53,8 +59,8 @@ r535_bar_bar2_update_pde(struct nvkm_gsp *gsp, u64 addr) return -EIO; rpc->info.barType = NV_RPC_UPDATE_PDE_BAR_2; - rpc->info.entryValue = addr ? ((addr >> 4) | 2) : 0; /* PD3 entry format! */ - rpc->info.entryLevelShift = 47; //XXX: probably fetch this from mmu! + rpc->info.entryValue = pdbe; + rpc->info.entryLevelShift = page_shift; return nvkm_gsp_rpc_wr(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV); } @@ -62,12 +68,13 @@ r535_bar_bar2_update_pde(struct nvkm_gsp *gsp, u64 addr) static void r535_bar_bar2_fini(struct nvkm_bar *bar) { + struct nvkm_vmm *vmm = gf100_bar(bar)->bar[0].vmm; struct nvkm_gsp *gsp = bar->subdev.device->gsp; bar->flushBAR2 = bar->flushBAR2PhysMode; nvkm_done(bar->flushFBZero); - WARN_ON(r535_bar_bar2_update_pde(gsp, 0)); + WARN_ON(r535_bar_bar2_update_pde(gsp, vmm->func->page[0].shift, 0)); } static void @@ -76,8 +83,18 @@ r535_bar_bar2_init(struct nvkm_bar *bar) struct nvkm_device *device = bar->subdev.device; struct nvkm_vmm *vmm = gf100_bar(bar)->bar[0].vmm; struct nvkm_gsp *gsp = device->gsp; - - WARN_ON(r535_bar_bar2_update_pde(gsp, vmm->pd->pde[0]->pt[0]->addr)); + struct nvkm_memory *pdb = vmm->pd->pt[0]->memory; + u32 pdb_offset = vmm->pd->pt[0]->base; + u32 pdbe_lo, pdbe_hi; + u64 pdbe; + + nvkm_kmap(pdb); + pdbe_lo = nvkm_ro32(pdb, pdb_offset + 0); + pdbe_hi = nvkm_ro32(pdb, pdb_offset + 4); + pdbe = ((u64)pdbe_hi << 32) | pdbe_lo; + nvkm_done(pdb); + + WARN_ON(r535_bar_bar2_update_pde(gsp, vmm->func->page[0].shift, pdbe)); vmm->rm.bar2_pdb = gsp->bar.rm_bar2_pdb; if (!bar->flushFBZero) { @@ -174,7 +191,7 @@ r535_bar_new_(const struct nvkm_bar_func *hw, struct nvkm_device *device, } *pbar = bar; - bar->flushBAR2PhysMode = ioremap(device->func->resource_addr(device, 3), PAGE_SIZE); + bar->flushBAR2PhysMode = ioremap(device->func->resource_addr(device, NVKM_BAR2_INST), PAGE_SIZE); if (!bar->flushBAR2PhysMode) return -ENOMEM; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ce.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ce.c new file mode 100644 index 000000000000..2d1ce9db2dcf --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ce.c @@ -0,0 +1,46 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/engine.h> + +#include "nvrm/ce.h" +#include "nvrm/engine.h" + +static int +r535_ce_alloc(struct nvkm_gsp_object *chan, u32 handle, u32 class, int inst, + struct nvkm_gsp_object *ce) +{ + NVC0B5_ALLOCATION_PARAMETERS *args; + + args = nvkm_gsp_rm_alloc_get(chan, handle, class, sizeof(*args), ce); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); + + args->version = 1; + args->engineType = NV2080_ENGINE_TYPE_COPY0 + inst; + + return nvkm_gsp_rm_alloc_wr(ce, args); +} + +const struct nvkm_rm_api_engine +r535_ce = { + .alloc = r535_ce_alloc, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/client.c index 932934227b9c..ec71f683e609 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ga100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/client.c @@ -19,26 +19,27 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "priv.h" +#include <rm/rm.h> -#include <subdev/gsp.h> +#include "nvrm/client.h" -#include <nvif/class.h> +static int +r535_gsp_client_ctor(struct nvkm_gsp_client *client, u32 handle) +{ + NV0000_ALLOC_PARAMETERS *args; -static const struct nvkm_engine_func -ga100_nvdec = { - .sclass = { - { -1, -1, NVC6B0_VIDEO_DECODER }, - {} - } -}; + args = nvkm_gsp_rm_alloc_get(&client->object, handle, NV01_ROOT, sizeof(*args), + &client->object); + if (IS_ERR(args)) + return PTR_ERR(args); -int -ga100_nvdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_nvdec **pnvdec) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_nvdec_new(&ga100_nvdec, device, type, inst, pnvdec); + args->hClient = client->object.handle; + args->processID = ~0; - return -ENODEV; + return nvkm_gsp_rm_alloc_wr(&client->object, args); } + +const struct nvkm_rm_api_client +r535_client = { + .ctor = r535_gsp_client_ctor, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ctrl.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ctrl.c new file mode 100644 index 000000000000..70b9ee911c5e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ctrl.c @@ -0,0 +1,93 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/rpc.h> + +#include "nvrm/ctrl.h" +#include "nvrm/rpcfn.h" + +static void +r535_gsp_rpc_rm_ctrl_done(struct nvkm_gsp_object *object, void *params) +{ + rpc_gsp_rm_control_v03_00 *rpc = to_payload_hdr(params, rpc); + + if (!params) + return; + nvkm_gsp_rpc_done(object->client->gsp, rpc); +} + +static int +r535_gsp_rpc_rm_ctrl_push(struct nvkm_gsp_object *object, void **params, u32 repc) +{ + rpc_gsp_rm_control_v03_00 *rpc = to_payload_hdr((*params), rpc); + struct nvkm_gsp *gsp = object->client->gsp; + int ret = 0; + + rpc = nvkm_gsp_rpc_push(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV, repc); + if (IS_ERR_OR_NULL(rpc)) { + *params = NULL; + return PTR_ERR(rpc); + } + + if (rpc->status) { + ret = r535_rpc_status_to_errno(rpc->status); + if (ret != -EAGAIN && ret != -EBUSY) + nvkm_error(&gsp->subdev, "cli:0x%08x obj:0x%08x ctrl cmd:0x%08x failed: 0x%08x\n", + object->client->object.handle, object->handle, rpc->cmd, rpc->status); + } + + if (repc) + *params = rpc->params; + else + nvkm_gsp_rpc_done(gsp, rpc); + + return ret; +} + +static void * +r535_gsp_rpc_rm_ctrl_get(struct nvkm_gsp_object *object, u32 cmd, u32 params_size) +{ + struct nvkm_gsp_client *client = object->client; + struct nvkm_gsp *gsp = client->gsp; + rpc_gsp_rm_control_v03_00 *rpc; + + nvkm_debug(&gsp->subdev, "cli:0x%08x obj:0x%08x ctrl cmd:0x%08x params_size:%d\n", + client->object.handle, object->handle, cmd, params_size); + + rpc = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_GSP_RM_CONTROL, + sizeof(*rpc) + params_size); + if (IS_ERR(rpc)) + return rpc; + + rpc->hClient = client->object.handle; + rpc->hObject = object->handle; + rpc->cmd = cmd; + rpc->status = 0; + rpc->paramsSize = params_size; + return rpc->params; +} + +const struct nvkm_rm_api_ctrl +r535_ctrl = { + .get = r535_gsp_rpc_rm_ctrl_get, + .push = r535_gsp_rpc_rm_ctrl_push, + .done = r535_gsp_rpc_rm_ctrl_done, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/device.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/device.c new file mode 100644 index 000000000000..f830e12a8f6e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/device.c @@ -0,0 +1,148 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/rm.h> + +#include "nvrm/device.h" +#include "nvrm/event.h" + +static void +r535_gsp_event_dtor(struct nvkm_gsp_event *event) +{ + struct nvkm_gsp_device *device = event->device; + struct nvkm_gsp_client *client = device->object.client; + struct nvkm_gsp *gsp = client->gsp; + + mutex_lock(&gsp->client_id.mutex); + if (event->func) { + list_del(&event->head); + event->func = NULL; + } + mutex_unlock(&gsp->client_id.mutex); + + nvkm_gsp_rm_free(&event->object); + event->device = NULL; +} + +static int +r535_gsp_device_event_get(struct nvkm_gsp_event *event) +{ + struct nvkm_gsp_device *device = event->device; + NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&device->subdevice, + NV2080_CTRL_CMD_EVENT_SET_NOTIFICATION, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->event = event->id; + ctrl->action = NV2080_CTRL_EVENT_SET_NOTIFICATION_ACTION_REPEAT; + return nvkm_gsp_rm_ctrl_wr(&device->subdevice, ctrl); +} + +static int +r535_gsp_device_event_ctor(struct nvkm_gsp_device *device, u32 handle, u32 id, + nvkm_gsp_event_func func, struct nvkm_gsp_event *event) +{ + struct nvkm_gsp_client *client = device->object.client; + struct nvkm_gsp *gsp = client->gsp; + NV0005_ALLOC_PARAMETERS *args; + int ret; + + args = nvkm_gsp_rm_alloc_get(&device->subdevice, handle, + NV01_EVENT_KERNEL_CALLBACK_EX, sizeof(*args), + &event->object); + if (IS_ERR(args)) + return PTR_ERR(args); + + args->hParentClient = client->object.handle; + args->hSrcResource = 0; + args->hClass = NV01_EVENT_KERNEL_CALLBACK_EX; + args->notifyIndex = NV01_EVENT_CLIENT_RM | id; + args->data = NULL; + + ret = nvkm_gsp_rm_alloc_wr(&event->object, args); + if (ret) + return ret; + + event->device = device; + event->id = id; + + ret = r535_gsp_device_event_get(event); + if (ret) { + nvkm_gsp_event_dtor(event); + return ret; + } + + mutex_lock(&gsp->client_id.mutex); + event->func = func; + list_add(&event->head, &client->events); + mutex_unlock(&gsp->client_id.mutex); + return 0; +} + +static void +r535_gsp_device_dtor(struct nvkm_gsp_device *device) +{ + nvkm_gsp_rm_free(&device->subdevice); + nvkm_gsp_rm_free(&device->object); +} + +static int +r535_gsp_subdevice_ctor(struct nvkm_gsp_device *device) +{ + NV2080_ALLOC_PARAMETERS *args; + + return nvkm_gsp_rm_alloc(&device->object, NVKM_RM_SUBDEVICE, NV20_SUBDEVICE_0, + sizeof(*args), &device->subdevice); +} + +static int +r535_gsp_device_ctor(struct nvkm_gsp_client *client, struct nvkm_gsp_device *device) +{ + NV0080_ALLOC_PARAMETERS *args; + int ret; + + args = nvkm_gsp_rm_alloc_get(&client->object, NVKM_RM_DEVICE, NV01_DEVICE_0, sizeof(*args), + &device->object); + if (IS_ERR(args)) + return PTR_ERR(args); + + args->hClientShare = client->object.handle; + + ret = nvkm_gsp_rm_alloc_wr(&device->object, args); + if (ret) + return ret; + + ret = r535_gsp_subdevice_ctor(device); + if (ret) + nvkm_gsp_rm_free(&device->object); + + return ret; +} + +const struct nvkm_rm_api_device +r535_device = { + .ctor = r535_gsp_device_ctor, + .dtor = r535_gsp_device_dtor, + .event.ctor = r535_gsp_device_event_ctor, + .event.dtor = r535_gsp_event_dtor, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/disp.c index 43ad467d09db..6e63df816d85 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/disp.c @@ -19,13 +19,13 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "priv.h" -#include "chan.h" -#include "conn.h" -#include "dp.h" -#include "head.h" -#include "ior.h" -#include "outp.h" +#include <engine/disp/priv.h> +#include <engine/disp/chan.h> +#include <engine/disp/conn.h> +#include <engine/disp/dp.h> +#include <engine/disp/head.h> +#include <engine/disp/ior.h> +#include <engine/disp/outp.h> #include <core/ramht.h> #include <subdev/bios.h> @@ -34,19 +34,11 @@ #include <subdev/mmu.h> #include <subdev/vfn.h> +#include <rm/gpu.h> + #include <nvhw/drf.h> -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080_notification.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dfp.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073dp.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073specific.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0073/ctrl0073system.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> -#include <nvrm/535.113.01/nvidia/generated/g_allclasses.h> -#include <nvrm/535.113.01/nvidia/generated/g_mem_desc_nvoc.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/os/nv_memory_type.h> +#include "nvrm/disp.h" #include <linux/acpi.h> @@ -78,9 +70,9 @@ r535_chan_fini(struct nvkm_disp_chan *chan) } static int -r535_chan_push(struct nvkm_disp_chan *chan) +r535_disp_chan_set_pushbuf(struct nvkm_disp *disp, s32 oclass, int inst, struct nvkm_memory *memory) { - struct nvkm_gsp *gsp = chan->disp->engine.subdev.device->gsp; + struct nvkm_gsp *gsp = disp->rm.objcom.client->gsp; NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS *ctrl; ctrl = nvkm_gsp_rm_ctrl_get(&gsp->internal.device.subdevice, @@ -89,8 +81,8 @@ r535_chan_push(struct nvkm_disp_chan *chan) if (IS_ERR(ctrl)) return PTR_ERR(ctrl); - if (chan->memory) { - switch (nvkm_memory_target(chan->memory)) { + if (memory) { + switch (nvkm_memory_target(memory)) { case NVKM_MEM_TARGET_NCOH: ctrl->addressSpace = ADDR_SYSMEM; ctrl->cacheSnoop = 0; @@ -107,13 +99,13 @@ r535_chan_push(struct nvkm_disp_chan *chan) return -EINVAL; } - ctrl->physicalAddr = nvkm_memory_addr(chan->memory); - ctrl->limit = nvkm_memory_size(chan->memory) - 1; + ctrl->physicalAddr = nvkm_memory_addr(memory); + ctrl->limit = nvkm_memory_size(memory) - 1; } - ctrl->hclass = chan->object.oclass; - ctrl->channelInstance = chan->head; - ctrl->valid = ((chan->object.oclass & 0xff) != 0x7a) ? 1 : 0; + ctrl->hclass = oclass; + ctrl->channelInstance = inst; + ctrl->valid = ((oclass & 0xff) != 0x7a) ? 1 : 0; return nvkm_gsp_rm_ctrl_wr(&gsp->internal.device.subdevice, ctrl); } @@ -121,10 +113,11 @@ r535_chan_push(struct nvkm_disp_chan *chan) static int r535_curs_init(struct nvkm_disp_chan *chan) { + const struct nvkm_rm_api *rmapi = chan->disp->rm.objcom.client->gsp->rm->api; NV50VAIO_CHANNELPIO_ALLOCATION_PARAMETERS *args; int ret; - ret = r535_chan_push(chan); + ret = rmapi->disp->chan.set_pushbuf(chan->disp, chan->object.oclass, chan->head, NULL); if (ret) return ret; @@ -172,25 +165,34 @@ r535_dmac_fini(struct nvkm_disp_chan *chan) } static int -r535_dmac_init(struct nvkm_disp_chan *chan) +r535_dmac_alloc(struct nvkm_disp *disp, u32 oclass, int inst, u32 put_offset, + struct nvkm_gsp_object *dmac) { NV50VAIO_CHANNELDMA_ALLOCATION_PARAMETERS *args; - int ret; - - ret = r535_chan_push(chan); - if (ret) - return ret; - args = nvkm_gsp_rm_alloc_get(&chan->disp->rm.object, - (chan->object.oclass << 16) | chan->head, - chan->object.oclass, sizeof(*args), &chan->rm.object); + args = nvkm_gsp_rm_alloc_get(&disp->rm.object, (oclass << 16) | inst, oclass, + sizeof(*args), dmac); if (IS_ERR(args)) return PTR_ERR(args); - args->channelInstance = chan->head; - args->offset = chan->suspend_put; + args->channelInstance = inst; + args->offset = put_offset; - return nvkm_gsp_rm_alloc_wr(&chan->rm.object, args); + return nvkm_gsp_rm_alloc_wr(dmac, args); +} + +static int +r535_dmac_init(struct nvkm_disp_chan *chan) +{ + const struct nvkm_rm_api *rmapi = chan->disp->rm.objcom.client->gsp->rm->api; + int ret; + + ret = rmapi->disp->chan.set_pushbuf(chan->disp, chan->object.oclass, chan->head, chan->memory); + if (ret) + return ret; + + return rmapi->disp->chan.dmac_alloc(chan->disp, chan->object.oclass, chan->head, + chan->suspend_put, &chan->rm.object); } static int @@ -260,47 +262,47 @@ r535_core = { }; static int -r535_sor_bl_set(struct nvkm_ior *sor, int lvl) +r535_bl_ctrl(struct nvkm_disp *disp, unsigned display_id, bool set, int *pval) { - struct nvkm_disp *disp = sor->disp; + u32 cmd = set ? NV0073_CTRL_CMD_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS : + NV0073_CTRL_CMD_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS; NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS *ctrl; + int ret; - ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, - NV0073_CTRL_CMD_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS, - sizeof(*ctrl)); + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, cmd, sizeof(*ctrl)); if (IS_ERR(ctrl)) return PTR_ERR(ctrl); - ctrl->displayId = BIT(sor->asy.outp->index); - ctrl->brightness = lvl; + ctrl->displayId = BIT(display_id); + ctrl->brightness = *pval; - return nvkm_gsp_rm_ctrl_wr(&disp->rm.objcom, ctrl); + ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); + if (ret) + return ret; + + *pval = ctrl->brightness; + + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; } static int -r535_sor_bl_get(struct nvkm_ior *sor) +r535_sor_bl_set(struct nvkm_ior *sor, int lvl) { struct nvkm_disp *disp = sor->disp; - NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS *ctrl; - int ret, lvl; - - ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, - NV0073_CTRL_CMD_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS, - sizeof(*ctrl)); - if (IS_ERR(ctrl)) - return PTR_ERR(ctrl); + const struct nvkm_rm_api *rmapi = disp->engine.subdev.device->gsp->rm->api; - ctrl->displayId = BIT(sor->asy.outp->index); + return rmapi->disp->bl_ctrl(disp, sor->asy.outp->index, true, &lvl); +} - ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); - if (ret) { - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); - return ret; - } +static int +r535_sor_bl_get(struct nvkm_ior *sor) +{ + struct nvkm_disp *disp = sor->disp; + const struct nvkm_rm_api *rmapi = disp->engine.subdev.device->gsp->rm->api; + int lvl, ret = rmapi->disp->bl_ctrl(disp, sor->asy.outp->index, false, &lvl); - lvl = ctrl->brightness; - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); - return lvl; + return (ret == 0) ? lvl : ret; } static const struct nvkm_ior_func_bl @@ -730,7 +732,7 @@ r535_outp_acquire(struct nvkm_outp *outp, bool hda) } static int -r535_disp_head_displayid(struct nvkm_disp *disp, int head, u32 *displayid) +r535_disp_get_active(struct nvkm_disp *disp, unsigned head, u32 *displayid) { NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS *ctrl; int ret; @@ -763,7 +765,9 @@ r535_outp_inherit(struct nvkm_outp *outp) int ret; list_for_each_entry(head, &disp->heads, head) { - ret = r535_disp_head_displayid(disp, head->id, &displayid); + const struct nvkm_rm_api *rmapi = disp->rm.objcom.client->gsp->rm->api; + + ret = rmapi->disp->get_active(disp, head->id, &displayid); if (WARN_ON(ret)) return NULL; @@ -858,10 +862,9 @@ r535_outp_dfp_get_info(struct nvkm_outp *outp) } static int -r535_outp_detect(struct nvkm_outp *outp) +r535_disp_get_connect_state(struct nvkm_disp *disp, unsigned display_id) { NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS *ctrl; - struct nvkm_disp *disp = outp->disp; int ret; ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, @@ -870,23 +873,29 @@ r535_outp_detect(struct nvkm_outp *outp) return PTR_ERR(ctrl); ctrl->subDeviceInstance = 0; - ctrl->displayMask = BIT(outp->index); + ctrl->displayMask = BIT(display_id); ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); - if (ret) { - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); - return ret; - } + if (ret == 0 && (ctrl->displayMask & BIT(display_id))) + ret = 1; + + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return ret; +} - if (ctrl->displayMask & BIT(outp->index)) { +static int +r535_outp_detect(struct nvkm_outp *outp) +{ + const struct nvkm_rm_api *rmapi = outp->disp->rm.objcom.client->gsp->rm->api; + int ret; + + ret = rmapi->disp->get_connect_state(outp->disp, outp->index); + if (ret == 1) { ret = r535_outp_dfp_get_info(outp); if (ret == 0) ret = 1; - } else { - ret = 0; } - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); return ret; } @@ -1029,15 +1038,11 @@ r535_dp_train(struct nvkm_outp *outp, bool retrain) } static int -r535_dp_rates(struct nvkm_outp *outp) +r535_dp_set_indexed_link_rates(struct nvkm_outp *outp) { NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS *ctrl; struct nvkm_disp *disp = outp->disp; - if (outp->conn->info.type != DCB_CONNECTOR_eDP || - !outp->dp.rates || outp->dp.rate[0].dpcd < 0) - return 0; - if (WARN_ON(outp->dp.rates > ARRAY_SIZE(ctrl->linkRateTbl))) return -EINVAL; @@ -1054,6 +1059,18 @@ r535_dp_rates(struct nvkm_outp *outp) } static int +r535_dp_rates(struct nvkm_outp *outp) +{ + struct nvkm_rm *rm = outp->disp->rm.objcom.client->gsp->rm; + + if (outp->conn->info.type != DCB_CONNECTOR_eDP || + !outp->dp.rates || outp->dp.rate[0].dpcd < 0) + return 0; + + return rm->api->disp->dp.set_indexed_link_rates(outp); +} + +static int r535_dp_aux_xfer(struct nvkm_outp *outp, u8 type, u32 addr, u8 *data, u8 *psize) { struct nvkm_disp *disp = outp->disp; @@ -1151,6 +1168,49 @@ r535_dp = { }; static int +r535_dp_get_caps(struct nvkm_disp *disp, int *plink_bw, bool *pmst, bool *pwm) +{ + NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS *ctrl; + int ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, + NV0073_CTRL_CMD_DP_GET_CAPS, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->sorIndex = ~0; + + ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); + if (ret) { + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return ret; + } + + switch (NVVAL_GET(ctrl->maxLinkRate, NV0073_CTRL_CMD, DP_GET_CAPS, MAX_LINK_RATE)) { + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_1_62: + *plink_bw = 0x06; + break; + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_2_70: + *plink_bw = 0x0a; + break; + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_5_40: + *plink_bw = 0x14; + break; + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_8_10: + *plink_bw = 0x1e; + break; + default: + *plink_bw = 0x00; + break; + } + + *pmst = ctrl->bIsMultistreamSupported; + *pwm = ctrl->bHasIncreasedWatermarkLimits; + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; +} + +static int r535_tmds_edid_get(struct nvkm_outp *outp, u8 *data, u16 *psize) { NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS *ctrl; @@ -1194,6 +1254,7 @@ r535_tmds = { static int r535_outp_new(struct nvkm_disp *disp, u32 id) { + const struct nvkm_rm_api *rmapi = disp->rm.objcom.client->gsp->rm->api; NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS *ctrl; enum nvkm_ior_proto proto; struct dcb_output dcbE = {}; @@ -1278,43 +1339,11 @@ r535_outp_new(struct nvkm_disp *disp, u32 id) if (ret) return ret; } else { - NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS *ctrl; bool mst, wm; - ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, - NV0073_CTRL_CMD_DP_GET_CAPS, sizeof(*ctrl)); - if (IS_ERR(ctrl)) - return PTR_ERR(ctrl); - - ctrl->sorIndex = ~0; - - ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); - if (ret) { - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + ret = rmapi->disp->dp.get_caps(disp, &dcbE.dpconf.link_bw, &mst, &wm); + if (ret) return ret; - } - - switch (NVVAL_GET(ctrl->maxLinkRate, NV0073_CTRL_CMD, DP_GET_CAPS, MAX_LINK_RATE)) { - case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_1_62: - dcbE.dpconf.link_bw = 0x06; - break; - case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_2_70: - dcbE.dpconf.link_bw = 0x0a; - break; - case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_5_40: - dcbE.dpconf.link_bw = 0x14; - break; - case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_8_10: - dcbE.dpconf.link_bw = 0x1e; - break; - default: - dcbE.dpconf.link_bw = 0x00; - break; - } - - mst = ctrl->bIsMultistreamSupported; - wm = ctrl->bHasIncreasedWatermarkLimits; - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); if (WARN_ON(!dcbE.dpconf.link_bw)) return -EINVAL; @@ -1441,11 +1470,47 @@ r535_disp_init(struct nvkm_disp *disp) } static int +r535_disp_get_supported(struct nvkm_disp *disp, unsigned long *pmask) +{ + NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_rd(&disp->rm.objcom, + NV0073_CTRL_CMD_SYSTEM_GET_SUPPORTED, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + *pmask = ctrl->displayMask; + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; +} + +static int +r535_disp_get_static_info(struct nvkm_disp *disp) +{ + NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS *ctrl; + struct nvkm_gsp *gsp = disp->rm.objcom.client->gsp; + + ctrl = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_DISPLAY_GET_STATIC_INFO, + sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + disp->wndw.mask = ctrl->windowPresentMask; + disp->wndw.nr = fls(disp->wndw.mask); + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); + return 0; +} + +static int r535_disp_oneinit(struct nvkm_disp *disp) { struct nvkm_device *device = disp->engine.subdev.device; struct nvkm_gsp *gsp = device->gsp; + const struct nvkm_rm_api *rmapi = gsp->rm->api; NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS *ctrl; + unsigned long mask; int ret, i; /* RAMIN. */ @@ -1476,24 +1541,14 @@ r535_disp_oneinit(struct nvkm_disp *disp) if (ret) return ret; - ret = nvkm_gsp_rm_alloc(&disp->rm.device.object, 0x00730000, NV04_DISPLAY_COMMON, 0, + ret = nvkm_gsp_rm_alloc(&disp->rm.device.object, NVKM_RM_DISP, NV04_DISPLAY_COMMON, 0, &disp->rm.objcom); if (ret) return ret; - { - NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS *ctrl; - - ctrl = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, - NV2080_CTRL_CMD_INTERNAL_DISPLAY_GET_STATIC_INFO, - sizeof(*ctrl)); - if (IS_ERR(ctrl)) - return PTR_ERR(ctrl); - - disp->wndw.mask = ctrl->windowPresentMask; - disp->wndw.nr = fls(disp->wndw.mask); - nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); - } + ret = rmapi->disp->get_static_info(disp); + if (ret) + return ret; /* */ { @@ -1622,25 +1677,14 @@ r535_disp_oneinit(struct nvkm_disp *disp) return ret; } - /* */ - { - NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS *ctrl; - unsigned long mask; - int i; - - ctrl = nvkm_gsp_rm_ctrl_rd(&disp->rm.objcom, - NV0073_CTRL_CMD_SYSTEM_GET_SUPPORTED, sizeof(*ctrl)); - if (IS_ERR(ctrl)) - return PTR_ERR(ctrl); - - mask = ctrl->displayMask; - nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + ret = rmapi->disp->get_supported(disp, &mask); + if (ret) + return ret; - for_each_set_bit(i, &mask, 32) { - ret = r535_outp_new(disp, i); - if (ret) - return ret; - } + for_each_set_bit(i, &mask, 32) { + ret = r535_outp_new(disp, i); + if (ret) + return ret; } ret = nvkm_event_init(&r535_disp_event, &gsp->subdev, 3, 32, &disp->rm.event); @@ -1686,6 +1730,7 @@ int r535_disp_new(const struct nvkm_disp_func *hw, struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_disp **pdisp) { + const struct nvkm_rm_gpu *gpu = device->gsp->rm->gpu; struct nvkm_disp_func *rm; int ret; @@ -1701,20 +1746,26 @@ r535_disp_new(const struct nvkm_disp_func *hw, struct nvkm_device *device, rm->sor.new = r535_sor_new; rm->ramht_size = hw->ramht_size; - rm->root = hw->root; + rm->root.oclass = gpu->disp.class.root; - for (int i = 0; hw->user[i].ctor; i++) { - switch (hw->user[i].base.oclass & 0xff) { - case 0x73: rm->user[i] = hw->user[i]; break; - case 0x7d: rm->user[i] = hw->user[i]; rm->user[i].chan = &r535_core; break; - case 0x7e: rm->user[i] = hw->user[i]; rm->user[i].chan = &r535_wndw; break; - case 0x7b: rm->user[i] = hw->user[i]; rm->user[i].chan = &r535_wimm; break; - case 0x7a: rm->user[i] = hw->user[i]; rm->user[i].chan = &r535_curs; break; - default: - WARN_ON(1); - continue; - } - } + rm->user[0].base.oclass = gpu->disp.class.caps; + rm->user[0].ctor = gv100_disp_caps_new; + + rm->user[1].base.oclass = gpu->disp.class.core; + rm->user[1].ctor = nvkm_disp_core_new; + rm->user[1].chan = &r535_core; + + rm->user[2].base.oclass = gpu->disp.class.wndw; + rm->user[2].ctor = nvkm_disp_wndw_new; + rm->user[2].chan = &r535_wndw; + + rm->user[3].base.oclass = gpu->disp.class.wimm; + rm->user[3].ctor = nvkm_disp_wndw_new; + rm->user[3].chan = &r535_wimm; + + rm->user[4].base.oclass = gpu->disp.class.curs; + rm->user[4].ctor = nvkm_disp_chan_new; + rm->user[4].chan = &r535_curs; ret = nvkm_disp_new_(rm, device, type, inst, pdisp); if (ret) @@ -1723,3 +1774,20 @@ r535_disp_new(const struct nvkm_disp_func *hw, struct nvkm_device *device, mutex_init(&(*pdisp)->super.mutex); //XXX return ret; } + +const struct nvkm_rm_api_disp +r535_disp = { + .get_static_info = r535_disp_get_static_info, + .get_supported = r535_disp_get_supported, + .get_connect_state = r535_disp_get_connect_state, + .get_active = r535_disp_get_active, + .bl_ctrl = r535_bl_ctrl, + .dp = { + .get_caps = r535_dp_get_caps, + .set_indexed_link_rates = r535_dp_set_indexed_link_rates, + }, + .chan = { + .set_pushbuf = r535_disp_chan_set_pushbuf, + .dmac_alloc = r535_dmac_alloc, + } +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fbsr.c index 35ba1798ee6e..150e22fde2ac 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fbsr.c @@ -19,19 +19,13 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "priv.h" - +#include <subdev/instmem/priv.h> #include <subdev/gsp.h> #include <nvhw/drf.h> -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl84a0.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> -#include <nvrm/535.113.01/nvidia/generated/g_fbsr_nvoc.h> -#include <nvrm/535.113.01/nvidia/generated/g_rpc-structures.h> -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_global_enums.h> +#include "nvrm/fbsr.h" +#include "nvrm/rpcfn.h" struct fbsr_item { const char *type; @@ -54,9 +48,9 @@ struct fbsr { u64 sys_offset; }; -static int -fbsr_memlist(struct nvkm_gsp_device *device, u32 handle, enum nvkm_memory_target aper, - u64 phys, u64 size, struct sg_table *sgt, struct nvkm_gsp_object *object) +int +r535_fbsr_memlist(struct nvkm_gsp_device *device, u32 handle, enum nvkm_memory_target aper, + u64 phys, u64 size, struct sg_table *sgt, struct nvkm_gsp_object *object) { struct nvkm_gsp_client *client = device->object.client; struct nvkm_gsp *gsp = client->gsp; @@ -123,8 +117,8 @@ fbsr_send(struct fbsr *fbsr, struct fbsr_item *item) struct nvkm_gsp_object memlist; int ret; - ret = fbsr_memlist(&fbsr->device, fbsr->hmemory, NVKM_MEM_TARGET_VRAM, - item->addr, item->size, NULL, &memlist); + ret = r535_fbsr_memlist(&fbsr->device, fbsr->hmemory, NVKM_MEM_TARGET_VRAM, + item->addr, item->size, NULL, &memlist); if (ret) return ret; @@ -161,8 +155,8 @@ fbsr_init(struct fbsr *fbsr, struct sg_table *sgt, u64 items_size) struct nvkm_gsp_object memlist; int ret; - ret = fbsr_memlist(&fbsr->device, fbsr->hmemory, NVKM_MEM_TARGET_HOST, - 0, fbsr->size, sgt, &memlist); + ret = r535_fbsr_memlist(&fbsr->device, fbsr->hmemory, NVKM_MEM_TARGET_HOST, + 0, fbsr->size, sgt, &memlist); if (ret) return ret; @@ -206,22 +200,19 @@ fbsr_inst(struct fbsr *fbsr, const char *type, struct nvkm_memory *memory) return fbsr_vram(fbsr, type, nvkm_memory_addr(memory), nvkm_memory_size(memory)); } -static void -r535_instmem_resume(struct nvkm_instmem *imem) +void +r535_fbsr_resume(struct nvkm_gsp *gsp) { /* RM has restored VRAM contents already, so just need to free the sysmem buffer. */ - if (imem->rm.fbsr_valid) { - nvkm_gsp_sg_free(imem->subdev.device, &imem->rm.fbsr); - imem->rm.fbsr_valid = false; - } + nvkm_gsp_sg_free(gsp->subdev.device, &gsp->sr.fbsr); } static int -r535_instmem_suspend(struct nvkm_instmem *imem) +r535_fbsr_suspend(struct nvkm_gsp *gsp) { - struct nvkm_subdev *subdev = &imem->subdev; + struct nvkm_subdev *subdev = &gsp->subdev; struct nvkm_device *device = subdev->device; - struct nvkm_gsp *gsp = device->gsp; + struct nvkm_instmem *imem = device->imem; struct nvkm_instobj *iobj; struct fbsr fbsr = {}; struct fbsr_item *item, *temp; @@ -262,7 +253,7 @@ r535_instmem_suspend(struct nvkm_instmem *imem) fbsr.size += gsp->fb.bios.vga_workspace.size; nvkm_debug(subdev, "fbsr: size: 0x%llx bytes\n", fbsr.size); - ret = nvkm_gsp_sg(gsp->subdev.device, fbsr.size, &imem->rm.fbsr); + ret = nvkm_gsp_sg(gsp->subdev.device, fbsr.size, &gsp->sr.fbsr); if (ret) goto done; @@ -271,7 +262,7 @@ r535_instmem_suspend(struct nvkm_instmem *imem) if (ret) goto done_sgt; - ret = fbsr_init(&fbsr, &imem->rm.fbsr, items_size); + ret = fbsr_init(&fbsr, &gsp->sr.fbsr, items_size); if (WARN_ON(ret)) goto done_sgt; @@ -282,12 +273,10 @@ r535_instmem_suspend(struct nvkm_instmem *imem) goto done_sgt; } - imem->rm.fbsr_valid = true; - /* Cleanup everything except the sysmem backup, which will be removed after resume. */ done_sgt: if (ret) /* ... unless we failed already. */ - nvkm_gsp_sg_free(device, &imem->rm.fbsr); + nvkm_gsp_sg_free(device, &gsp->sr.fbsr); done: list_for_each_entry_safe(item, temp, &fbsr.items, head) { list_del(&item->head); @@ -299,6 +288,12 @@ done: return ret; } +const struct nvkm_rm_api_fbsr +r535_fbsr = { + .suspend = r535_fbsr_suspend, + .resume = r535_fbsr_resume, +}; + static void * r535_instmem_dtor(struct nvkm_instmem *imem) { @@ -319,11 +314,10 @@ r535_instmem_new(const struct nvkm_instmem_func *hw, rm->dtor = r535_instmem_dtor; rm->fini = hw->fini; - rm->suspend = r535_instmem_suspend; - rm->resume = r535_instmem_resume; rm->memory_new = hw->memory_new; rm->memory_wrap = hw->memory_wrap; rm->zero = false; + rm->set_bar0_window_addr = hw->set_bar0_window_addr; ret = nv50_instmem_new_(rm, device, type, inst, pinstmem); if (ret) diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fifo.c index 3454c7d29502..1ac5628c5140 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fifo.c @@ -19,11 +19,11 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "priv.h" -#include "cgrp.h" -#include "chan.h" -#include "chid.h" -#include "runl.h" +#include <engine/fifo/priv.h> +#include <engine/fifo/cgrp.h> +#include <engine/fifo/chan.h> +#include <engine/fifo/chid.h> +#include <engine/fifo/runl.h> #include <core/gpuobj.h> #include <subdev/gsp.h> @@ -31,24 +31,19 @@ #include <subdev/vfn.h> #include <engine/gr.h> +#include <rm/engine.h> + #include <nvhw/drf.h> -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/alloc/alloc_channel.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080_notification.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080ce.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080fifo.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrla06f/ctrla06fgpfifo.h> -#include <nvrm/535.113.01/nvidia/generated/g_kernel_channel_nvoc.h> -#include <nvrm/535.113.01/nvidia/generated/g_kernel_fifo_nvoc.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/gpu_engine_type.h> +#include "nvrm/fifo.h" +#include "nvrm/engine.h" static u32 r535_chan_doorbell_handle(struct nvkm_chan *chan) { - return (chan->cgrp->runl->id << 16) | chan->id; + struct nvkm_gsp *gsp = chan->rm.object.client->gsp; + + return gsp->rm->gpu->fifo.chan.doorbell_handle(chan); } static void @@ -77,50 +72,29 @@ r535_chan_ramfc_clear(struct nvkm_chan *chan) #define CHID_PER_USERD 8 static int -r535_chan_ramfc_write(struct nvkm_chan *chan, u64 offset, u64 length, u32 devm, bool priv) +r535_chan_alloc(struct nvkm_gsp_device *device, u32 handle, u32 nv2080_engine_type, u8 runq, + bool priv, int chid, u64 inst_addr, u64 userd_addr, u64 mthdbuf_addr, + struct nvkm_vmm *vmm, u64 gpfifo_offset, u32 gpfifo_length, + struct nvkm_gsp_object *chan) { - struct nvkm_fifo *fifo = chan->cgrp->runl->fifo; - struct nvkm_engn *engn; - struct nvkm_device *device = fifo->engine.subdev.device; + struct nvkm_gsp *gsp = device->object.client->gsp; + struct nvkm_fifo *fifo = gsp->subdev.device->fifo; + const int userd_p = chid / CHID_PER_USERD; + const int userd_i = chid % CHID_PER_USERD; NV_CHANNELGPFIFO_ALLOCATION_PARAMETERS *args; - const int userd_p = chan->id / CHID_PER_USERD; - const int userd_i = chan->id % CHID_PER_USERD; - u32 eT = ~0; - int ret; - - if (unlikely(device->gr && !device->gr->engine.subdev.oneinit)) { - ret = nvkm_subdev_oneinit(&device->gr->engine.subdev); - if (ret) - return ret; - } - - nvkm_runl_foreach_engn(engn, chan->cgrp->runl) { - eT = engn->id; - break; - } - - if (WARN_ON(eT == ~0)) - return -EINVAL; - chan->rm.mthdbuf.ptr = dma_alloc_coherent(fifo->engine.subdev.device->dev, - fifo->rm.mthdbuf_size, - &chan->rm.mthdbuf.addr, GFP_KERNEL); - if (!chan->rm.mthdbuf.ptr) - return -ENOMEM; - - args = nvkm_gsp_rm_alloc_get(&chan->vmm->rm.device.object, 0xf1f00000 | chan->id, - fifo->func->chan.user.oclass, sizeof(*args), - &chan->rm.object); + args = nvkm_gsp_rm_alloc_get(&device->object, handle, + fifo->func->chan.user.oclass, sizeof(*args), chan); if (WARN_ON(IS_ERR(args))) return PTR_ERR(args); - args->gpFifoOffset = offset; - args->gpFifoEntries = length / 8; + args->gpFifoOffset = gpfifo_offset; + args->gpFifoEntries = gpfifo_length / 8; args->flags = NVDEF(NVOS04, FLAGS, CHANNEL_TYPE, PHYSICAL); args->flags |= NVDEF(NVOS04, FLAGS, VPR, FALSE); args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_SKIP_MAP_REFCOUNTING, FALSE); - args->flags |= NVVAL(NVOS04, FLAGS, GROUP_CHANNEL_RUNQUEUE, chan->runq); + args->flags |= NVVAL(NVOS04, FLAGS, GROUP_CHANNEL_RUNQUEUE, runq); if (!priv) args->flags |= NVDEF(NVOS04, FLAGS, PRIVILEGED_CHANNEL, FALSE); else @@ -143,25 +117,25 @@ r535_chan_ramfc_write(struct nvkm_chan *chan, u64 offset, u64 length, u32 devm, args->flags |= NVDEF(NVOS04, FLAGS, MAP_CHANNEL, FALSE); args->flags |= NVDEF(NVOS04, FLAGS, SKIP_CTXBUFFER_ALLOC, FALSE); - args->hVASpace = chan->vmm->rm.object.handle; - args->engineType = eT; + args->hVASpace = vmm->rm.object.handle; + args->engineType = nv2080_engine_type; - args->instanceMem.base = chan->inst->addr; - args->instanceMem.size = chan->inst->size; + args->instanceMem.base = inst_addr; + args->instanceMem.size = fifo->func->chan.func->inst->size; args->instanceMem.addressSpace = 2; args->instanceMem.cacheAttrib = 1; - args->userdMem.base = nvkm_memory_addr(chan->userd.mem) + chan->userd.base; + args->userdMem.base = userd_addr; args->userdMem.size = fifo->func->chan.func->userd->size; args->userdMem.addressSpace = 2; args->userdMem.cacheAttrib = 1; - args->ramfcMem.base = chan->inst->addr + 0; + args->ramfcMem.base = inst_addr; args->ramfcMem.size = 0x200; args->ramfcMem.addressSpace = 2; args->ramfcMem.cacheAttrib = 1; - args->mthdbufMem.base = chan->rm.mthdbuf.addr; + args->mthdbufMem.base = mthdbuf_addr; args->mthdbufMem.size = fifo->rm.mthdbuf_size; args->mthdbufMem.addressSpace = 1; args->mthdbufMem.cacheAttrib = 0; @@ -173,7 +147,44 @@ r535_chan_ramfc_write(struct nvkm_chan *chan, u64 offset, u64 length, u32 devm, args->internalFlags |= NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, ERROR_NOTIFIER_TYPE, NONE); args->internalFlags |= NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, ECC_ERROR_NOTIFIER_TYPE, NONE); - ret = nvkm_gsp_rm_alloc_wr(&chan->rm.object, args); + return nvkm_gsp_rm_alloc_wr(chan, args); +} + +static int +r535_chan_ramfc_write(struct nvkm_chan *chan, u64 offset, u64 length, u32 devm, bool priv) +{ + struct nvkm_fifo *fifo = chan->cgrp->runl->fifo; + struct nvkm_engn *engn; + struct nvkm_device *device = fifo->engine.subdev.device; + const struct nvkm_rm_api *rmapi = device->gsp->rm->api; + u32 eT = ~0; + int ret; + + if (unlikely(device->gr && !device->gr->engine.subdev.oneinit)) { + ret = nvkm_subdev_oneinit(&device->gr->engine.subdev); + if (ret) + return ret; + } + + nvkm_runl_foreach_engn(engn, chan->cgrp->runl) { + eT = engn->id; + break; + } + + if (WARN_ON(eT == ~0)) + return -EINVAL; + + chan->rm.mthdbuf.ptr = dma_alloc_coherent(fifo->engine.subdev.device->dev, + fifo->rm.mthdbuf_size, + &chan->rm.mthdbuf.addr, GFP_KERNEL); + if (!chan->rm.mthdbuf.ptr) + return -ENOMEM; + + ret = rmapi->fifo->chan.alloc(&chan->vmm->rm.device, NVKM_RM_CHAN(chan->id), + eT, chan->runq, priv, chan->id, chan->inst->addr, + nvkm_memory_addr(chan->userd.mem) + chan->userd.base, + chan->rm.mthdbuf.addr, chan->vmm, offset, length, + &chan->rm.object); if (ret) return ret; @@ -215,123 +226,8 @@ r535_chan_ramfc = { .priv = true, }; -struct r535_chan_userd { - struct nvkm_memory *mem; - struct nvkm_memory *map; - int chid; - u32 used; - - struct list_head head; -} *userd; - -static void -r535_chan_id_put(struct nvkm_chan *chan) -{ - struct nvkm_runl *runl = chan->cgrp->runl; - struct nvkm_fifo *fifo = runl->fifo; - struct r535_chan_userd *userd; - - mutex_lock(&fifo->userd.mutex); - list_for_each_entry(userd, &fifo->userd.list, head) { - if (userd->map == chan->userd.mem) { - u32 chid = chan->userd.base / chan->func->userd->size; - - userd->used &= ~BIT(chid); - if (!userd->used) { - nvkm_memory_unref(&userd->map); - nvkm_memory_unref(&userd->mem); - nvkm_chid_put(runl->chid, userd->chid, &chan->cgrp->lock); - list_del(&userd->head); - kfree(userd); - } - - break; - } - } - mutex_unlock(&fifo->userd.mutex); - -} - -static int -r535_chan_id_get_locked(struct nvkm_chan *chan, struct nvkm_memory *muserd, u64 ouserd) -{ - const u32 userd_size = CHID_PER_USERD * chan->func->userd->size; - struct nvkm_runl *runl = chan->cgrp->runl; - struct nvkm_fifo *fifo = runl->fifo; - struct r535_chan_userd *userd; - u32 chid; - int ret; - - if (ouserd + chan->func->userd->size >= userd_size || - (ouserd & (chan->func->userd->size - 1))) { - RUNL_DEBUG(runl, "ouserd %llx", ouserd); - return -EINVAL; - } - - chid = div_u64(ouserd, chan->func->userd->size); - - list_for_each_entry(userd, &fifo->userd.list, head) { - if (userd->mem == muserd) { - if (userd->used & BIT(chid)) - return -EBUSY; - break; - } - } - - if (&userd->head == &fifo->userd.list) { - if (nvkm_memory_size(muserd) < userd_size) { - RUNL_DEBUG(runl, "userd too small"); - return -EINVAL; - } - - userd = kzalloc(sizeof(*userd), GFP_KERNEL); - if (!userd) - return -ENOMEM; - - userd->chid = nvkm_chid_get(runl->chid, chan); - if (userd->chid < 0) { - ret = userd->chid; - kfree(userd); - return ret; - } - - userd->mem = nvkm_memory_ref(muserd); - - ret = nvkm_memory_kmap(userd->mem, &userd->map); - if (ret) { - nvkm_chid_put(runl->chid, userd->chid, &chan->cgrp->lock); - kfree(userd); - return ret; - } - - - list_add(&userd->head, &fifo->userd.list); - } - - userd->used |= BIT(chid); - - chan->userd.mem = nvkm_memory_ref(userd->map); - chan->userd.base = ouserd; - - return (userd->chid * CHID_PER_USERD) + chid; -} - -static int -r535_chan_id_get(struct nvkm_chan *chan, struct nvkm_memory *muserd, u64 ouserd) -{ - struct nvkm_fifo *fifo = chan->cgrp->runl->fifo; - int ret; - - mutex_lock(&fifo->userd.mutex); - ret = r535_chan_id_get_locked(chan, muserd, ouserd); - mutex_unlock(&fifo->userd.mutex); - return ret; -} - static const struct nvkm_chan_func r535_chan = { - .id_get = r535_chan_id_get, - .id_put = r535_chan_id_put, .inst = &gf100_chan_inst, .userd = &gv100_chan_userd, .ramfc = &r535_chan_ramfc, @@ -340,10 +236,6 @@ r535_chan = { .doorbell_handle = r535_chan_doorbell_handle, }; -static const struct nvkm_cgrp_func -r535_cgrp = { -}; - static int r535_engn_nonstall(struct nvkm_engn *engn) { @@ -356,7 +248,7 @@ r535_engn_nonstall(struct nvkm_engn *engn) } static const struct nvkm_engn_func -r535_ce = { +r535_engn_ce = { .nonstall = r535_engn_nonstall, }; @@ -376,7 +268,7 @@ r535_gr_ctor(struct nvkm_engn *engn, struct nvkm_vctx *vctx, struct nvkm_chan *c } static const struct nvkm_engn_func -r535_gr = { +r535_engn_gr = { .nonstall = r535_engn_nonstall, .ctor2 = r535_gr_ctor, }; @@ -449,57 +341,86 @@ r535_runl = { .allow = r535_runl_allow, }; -static int -r535_fifo_2080_type(enum nvkm_subdev_type type, int inst) +void +r535_fifo_rc_chid(struct nvkm_fifo *fifo, int chid) { - switch (type) { - case NVKM_ENGINE_GR: return NV2080_ENGINE_TYPE_GR0; - case NVKM_ENGINE_CE: return NV2080_ENGINE_TYPE_COPY0 + inst; - case NVKM_ENGINE_SEC2: return NV2080_ENGINE_TYPE_SEC2; - case NVKM_ENGINE_NVDEC: return NV2080_ENGINE_TYPE_NVDEC0 + inst; - case NVKM_ENGINE_NVENC: return NV2080_ENGINE_TYPE_NVENC0 + inst; - case NVKM_ENGINE_NVJPG: return NV2080_ENGINE_TYPE_NVJPEG0 + inst; - case NVKM_ENGINE_OFA: return NV2080_ENGINE_TYPE_OFA; - case NVKM_ENGINE_SW: return NV2080_ENGINE_TYPE_SW; - default: - break; + struct nvkm_chan *chan; + unsigned long flags; + + chan = nvkm_chan_get_chid(&fifo->engine, chid, &flags); + if (!chan) { + nvkm_error(&fifo->engine.subdev, "rc: chid %d not found!\n", chid); + return; } - WARN_ON(1); - return -EINVAL; + nvkm_chan_error(chan, false); + nvkm_chan_put(&chan, flags); } static int -r535_fifo_engn_type(RM_ENGINE_TYPE rm, enum nvkm_subdev_type *ptype) +r535_fifo_rc_triggered(void *priv, u32 fn, void *repv, u32 repc) { + rpc_rc_triggered_v17_02 *msg = repv; + struct nvkm_gsp *gsp = priv; + + if (WARN_ON(repc < sizeof(*msg))) + return -EINVAL; + + nvkm_error(&gsp->subdev, "rc: engn:%08x chid:%d type:%d scope:%d part:%d\n", + msg->nv2080EngineType, msg->chid, msg->exceptType, msg->scope, + msg->partitionAttributionId); + + r535_fifo_rc_chid(gsp->subdev.device->fifo, msg->chid); + return 0; +} + +static int +r535_fifo_xlat_rm_engine_type(u32 rm, enum nvkm_subdev_type *ptype, int *p2080) +{ +#define RM_ENGINE_TYPE(RM,NVKM,INST) \ + RM_ENGINE_TYPE_##RM: \ + *ptype = NVKM_ENGINE_##NVKM; \ + *p2080 = NV2080_ENGINE_TYPE_##RM; \ + return INST + switch (rm) { - case RM_ENGINE_TYPE_GR0: - *ptype = NVKM_ENGINE_GR; - return 0; - case RM_ENGINE_TYPE_COPY0...RM_ENGINE_TYPE_COPY9: - *ptype = NVKM_ENGINE_CE; - return rm - RM_ENGINE_TYPE_COPY0; - case RM_ENGINE_TYPE_NVDEC0...RM_ENGINE_TYPE_NVDEC7: - *ptype = NVKM_ENGINE_NVDEC; - return rm - RM_ENGINE_TYPE_NVDEC0; - case RM_ENGINE_TYPE_NVENC0...RM_ENGINE_TYPE_NVENC2: - *ptype = NVKM_ENGINE_NVENC; - return rm - RM_ENGINE_TYPE_NVENC0; - case RM_ENGINE_TYPE_SW: - *ptype = NVKM_ENGINE_SW; - return 0; - case RM_ENGINE_TYPE_SEC2: - *ptype = NVKM_ENGINE_SEC2; - return 0; - case RM_ENGINE_TYPE_NVJPEG0...RM_ENGINE_TYPE_NVJPEG7: - *ptype = NVKM_ENGINE_NVJPG; - return rm - RM_ENGINE_TYPE_NVJPEG0; - case RM_ENGINE_TYPE_OFA: - *ptype = NVKM_ENGINE_OFA; - return 0; + case RM_ENGINE_TYPE( GR0, GR, 0); + case RM_ENGINE_TYPE( COPY0, CE, 0); + case RM_ENGINE_TYPE( COPY1, CE, 1); + case RM_ENGINE_TYPE( COPY2, CE, 2); + case RM_ENGINE_TYPE( COPY3, CE, 3); + case RM_ENGINE_TYPE( COPY4, CE, 4); + case RM_ENGINE_TYPE( COPY5, CE, 5); + case RM_ENGINE_TYPE( COPY6, CE, 6); + case RM_ENGINE_TYPE( COPY7, CE, 7); + case RM_ENGINE_TYPE( COPY8, CE, 8); + case RM_ENGINE_TYPE( COPY9, CE, 9); + case RM_ENGINE_TYPE( NVDEC0, NVDEC, 0); + case RM_ENGINE_TYPE( NVDEC1, NVDEC, 1); + case RM_ENGINE_TYPE( NVDEC2, NVDEC, 2); + case RM_ENGINE_TYPE( NVDEC3, NVDEC, 3); + case RM_ENGINE_TYPE( NVDEC4, NVDEC, 4); + case RM_ENGINE_TYPE( NVDEC5, NVDEC, 5); + case RM_ENGINE_TYPE( NVDEC6, NVDEC, 6); + case RM_ENGINE_TYPE( NVDEC7, NVDEC, 7); + case RM_ENGINE_TYPE( NVENC0, NVENC, 0); + case RM_ENGINE_TYPE( NVENC1, NVENC, 1); + case RM_ENGINE_TYPE( NVENC2, NVENC, 2); + case RM_ENGINE_TYPE(NVJPEG0, NVJPG, 0); + case RM_ENGINE_TYPE(NVJPEG1, NVJPG, 1); + case RM_ENGINE_TYPE(NVJPEG2, NVJPG, 2); + case RM_ENGINE_TYPE(NVJPEG3, NVJPG, 3); + case RM_ENGINE_TYPE(NVJPEG4, NVJPG, 4); + case RM_ENGINE_TYPE(NVJPEG5, NVJPG, 5); + case RM_ENGINE_TYPE(NVJPEG6, NVJPG, 6); + case RM_ENGINE_TYPE(NVJPEG7, NVJPG, 7); + case RM_ENGINE_TYPE( SW, SW, 0); + case RM_ENGINE_TYPE( SEC2, SEC2, 0); + case RM_ENGINE_TYPE( OFA, OFA, 0); default: return -EINVAL; } +#undef RM_ENGINE_TYPE } static int @@ -536,16 +457,19 @@ static int r535_fifo_runl_ctor(struct nvkm_fifo *fifo) { struct nvkm_subdev *subdev = &fifo->engine.subdev; - struct nvkm_gsp *gsp = subdev->device->gsp; + struct nvkm_device *device = subdev->device; + struct nvkm_gsp *gsp = device->gsp; + struct nvkm_rm *rm = gsp->rm; struct nvkm_runl *runl; struct nvkm_engn *engn; - u32 cgids = 2048; u32 chids = 2048; + u32 first = rm->api->fifo->rsvd_chids; + u32 count = chids - first; int ret; NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS *ctrl; - if ((ret = nvkm_chid_new(&nvkm_chan_event, subdev, cgids, 0, cgids, &fifo->cgid)) || - (ret = nvkm_chid_new(&nvkm_chan_event, subdev, chids, 0, chids, &fifo->chid))) + if ((ret = nvkm_chid_new(&nvkm_chan_event, subdev, chids, first, count, &fifo->cgid)) || + (ret = nvkm_chid_new(&nvkm_chan_event, subdev, chids, first, count, &fifo->chid))) return ret; ctrl = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, @@ -576,25 +500,43 @@ r535_fifo_runl_ctor(struct nvkm_fifo *fifo) if (!runl) continue; - inst = r535_fifo_engn_type(rmid, &type); + inst = rm->api->fifo->xlat_rm_engine_type(rmid, &type, &nv2080); if (inst < 0) { nvkm_warn(subdev, "RM_ENGINE_TYPE 0x%x\n", rmid); nvkm_runl_del(runl); continue; } - nv2080 = r535_fifo_2080_type(type, inst); - if (nv2080 < 0) { + /* Skip SW engine - there's currently no support for NV SW classes. */ + if (type == NVKM_ENGINE_SW) + continue; + + /* Skip lone GRCEs (ones not paired with GR on a runlist), as they + * don't appear to function as async copy engines. + */ + if (type == NVKM_ENGINE_CE && + rm->gpu->ce.grce_mask && + (rm->gpu->ce.grce_mask(device) & BIT(inst)) && + !nvkm_runl_find_engn(engn, runl, engn->engine->subdev.type == NVKM_ENGINE_GR)) { + RUNL_DEBUG(runl, "skip LCE %d - GRCE without GR", inst); + nvkm_runl_del(runl); + continue; + } + + ret = nvkm_rm_engine_new(gsp->rm, type, inst); + if (ret) { nvkm_runl_del(runl); continue; } + engn = NULL; + switch (type) { case NVKM_ENGINE_CE: - engn = nvkm_runl_add(runl, nv2080, &r535_ce, type, inst); + engn = nvkm_runl_add(runl, nv2080, &r535_engn_ce, type, inst); break; case NVKM_ENGINE_GR: - engn = nvkm_runl_add(runl, nv2080, &r535_gr, type, inst); + engn = nvkm_runl_add(runl, nv2080, &r535_engn_gr, type, inst); break; case NVKM_ENGINE_NVDEC: case NVKM_ENGINE_NVENC: @@ -633,7 +575,7 @@ r535_fifo_runl_ctor(struct nvkm_fifo *fifo) nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); } - return r535_fifo_ectx_size(fifo); + return rm->api->fifo->ectx_size(fifo); } static void @@ -646,6 +588,7 @@ int r535_fifo_new(const struct nvkm_fifo_func *hw, struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_fifo **pfifo) { + const struct nvkm_rm_gpu *gpu = device->gsp->rm->gpu; struct nvkm_fifo_func *rm; if (!(rm = kzalloc(sizeof(*rm), GFP_KERNEL))) @@ -654,12 +597,20 @@ r535_fifo_new(const struct nvkm_fifo_func *hw, struct nvkm_device *device, rm->dtor = r535_fifo_dtor; rm->runl_ctor = r535_fifo_runl_ctor; rm->runl = &r535_runl; - rm->cgrp = hw->cgrp; - rm->cgrp.func = &r535_cgrp; - rm->chan = hw->chan; + rm->chan.user.oclass = gpu->fifo.chan.class; rm->chan.func = &r535_chan; rm->nonstall = &ga100_fifo_nonstall; rm->nonstall_ctor = ga100_fifo_nonstall_ctor; return nvkm_fifo_new_(rm, device, type, inst, pfifo); } + +const struct nvkm_rm_api_fifo +r535_fifo = { + .xlat_rm_engine_type = r535_fifo_xlat_rm_engine_type, + .ectx_size = r535_fifo_ectx_size, + .rc_triggered = r535_fifo_rc_triggered, + .chan = { + .alloc = r535_chan_alloc, + }, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/gr.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/gr.c new file mode 100644 index 000000000000..ddb57d5e73d6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/gr.c @@ -0,0 +1,356 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/gr.h> + +#include <core/memory.h> +#include <subdev/gsp.h> +#include <subdev/mmu/vmm.h> +#include <engine/fifo/priv.h> +#include <engine/gr/priv.h> + +#include <nvif/if900d.h> + +#include <nvhw/drf.h> + +#include "nvrm/gr.h" +#include "nvrm/vmm.h" + +#define r535_gr(p) container_of((p), struct r535_gr, base) + +static void * +r535_gr_chan_dtor(struct nvkm_object *object) +{ + struct r535_gr_chan *grc = container_of(object, typeof(*grc), object); + struct r535_gr *gr = grc->gr; + + for (int i = 0; i < gr->ctxbuf_nr; i++) { + nvkm_vmm_put(grc->vmm, &grc->vma[i]); + nvkm_memory_unref(&grc->mem[i]); + } + + nvkm_vmm_unref(&grc->vmm); + return grc; +} + +static const struct nvkm_object_func +r535_gr_chan = { + .dtor = r535_gr_chan_dtor, +}; + +int +r535_gr_promote_ctx(struct r535_gr *gr, bool golden, struct nvkm_vmm *vmm, + struct nvkm_memory **pmem, struct nvkm_vma **pvma, + struct nvkm_gsp_object *chan) +{ + struct nvkm_subdev *subdev = &gr->base.engine.subdev; + struct nvkm_device *device = subdev->device; + NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&vmm->rm.device.subdevice, + NV2080_CTRL_CMD_GPU_PROMOTE_CTX, sizeof(*ctrl)); + if (WARN_ON(IS_ERR(ctrl))) + return PTR_ERR(ctrl); + + ctrl->engineType = 1; + ctrl->hChanClient = vmm->rm.client.object.handle; + ctrl->hObject = chan->handle; + + for (int i = 0; i < gr->ctxbuf_nr; i++) { + NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY *entry = + &ctrl->promoteEntry[ctrl->entryCount]; + const bool alloc = golden || !gr->ctxbuf[i].global; + int ret; + + entry->bufferId = gr->ctxbuf[i].bufferId; + entry->bInitialize = gr->ctxbuf[i].init && alloc; + + if (alloc) { + ret = nvkm_memory_new(device, gr->ctxbuf[i].init ? + NVKM_MEM_TARGET_INST : NVKM_MEM_TARGET_INST_SR_LOST, + gr->ctxbuf[i].size, 1 << gr->ctxbuf[i].page, + gr->ctxbuf[i].init, &pmem[i]); + if (WARN_ON(ret)) + return ret; + + if (gr->ctxbuf[i].bufferId == + NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PRIV_ACCESS_MAP) + entry->bNonmapped = 1; + } else { + if (gr->ctxbuf[i].bufferId == + NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_UNRESTRICTED_PRIV_ACCESS_MAP) + continue; + + pmem[i] = nvkm_memory_ref(gr->ctxbuf_mem[i]); + } + + if (!entry->bNonmapped) { + struct gf100_vmm_map_v0 args = { + .priv = 1, + .ro = gr->ctxbuf[i].ro, + }; + + mutex_lock(&vmm->mutex.vmm); + ret = nvkm_vmm_get_locked(vmm, false, true, false, 0, gr->ctxbuf[i].align, + nvkm_memory_size(pmem[i]), &pvma[i]); + mutex_unlock(&vmm->mutex.vmm); + if (ret) + return ret; + + ret = nvkm_memory_map(pmem[i], 0, vmm, pvma[i], &args, sizeof(args)); + if (ret) + return ret; + + entry->gpuVirtAddr = pvma[i]->addr; + } + + if (entry->bInitialize) { + entry->gpuPhysAddr = nvkm_memory_addr(pmem[i]); + entry->size = gr->ctxbuf[i].size; + entry->physAttr = 4; + } + + nvkm_debug(subdev, + "promote %02d: pa %016llx/%08x sz %016llx va %016llx init:%d nm:%d\n", + entry->bufferId, entry->gpuPhysAddr, entry->physAttr, entry->size, + entry->gpuVirtAddr, entry->bInitialize, entry->bNonmapped); + + ctrl->entryCount++; + } + + return nvkm_gsp_rm_ctrl_wr(&vmm->rm.device.subdevice, ctrl); +} + +int +r535_gr_chan_new(struct nvkm_gr *base, struct nvkm_chan *chan, const struct nvkm_oclass *oclass, + struct nvkm_object **pobject) +{ + struct r535_gr *gr = r535_gr(base); + struct r535_gr_chan *grc; + int ret; + + if (!(grc = kzalloc(sizeof(*grc), GFP_KERNEL))) + return -ENOMEM; + + nvkm_object_ctor(&r535_gr_chan, oclass, &grc->object); + grc->gr = gr; + grc->vmm = nvkm_vmm_ref(chan->vmm); + grc->chan = chan; + *pobject = &grc->object; + + ret = r535_gr_promote_ctx(gr, false, grc->vmm, grc->mem, grc->vma, &chan->rm.object); + if (ret) + return ret; + + return 0; +} + +u64 +r535_gr_units(struct nvkm_gr *gr) +{ + struct nvkm_gsp *gsp = gr->engine.subdev.device->gsp; + + return (gsp->gr.tpcs << 8) | gsp->gr.gpcs; +} + +void +r535_gr_get_ctxbuf_info(struct r535_gr *gr, int i, + struct NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO *info) +{ + struct nvkm_subdev *subdev = &gr->base.engine.subdev; + static const struct { + u32 id0; /* NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID */ + u32 id1; /* NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID */ + bool global; + bool init; + bool ro; + } map[] = { +#define _A(n,N,G,I,R) { .id0 = NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_##n, \ + .id1 = NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_##N, \ + .global = (G), .init = (I), .ro = (R) } +#define _B(N,G,I,R) _A(GRAPHICS_##N, N, (G), (I), (R)) + /* global init ro */ + _A( GRAPHICS, MAIN, false, true, false), + _B( PATCH, false, true, false), + _A( GRAPHICS_BUNDLE_CB, BUFFER_BUNDLE_CB, true, false, false), + _B( PAGEPOOL, true, false, false), + _B( ATTRIBUTE_CB, true, false, false), + _B( RTV_CB_GLOBAL, true, false, false), + _B( FECS_EVENT, true, true, false), + _B( PRIV_ACCESS_MAP, true, true, true), +#undef _B +#undef _A + }; + u32 size = info->size; + u8 align, page; + int id; + + for (id = 0; id < ARRAY_SIZE(map); id++) { + if (map[id].id0 == i) + break; + } + + nvkm_debug(subdev, "%02x: size:0x%08x %s\n", i, + size, (id < ARRAY_SIZE(map)) ? "*" : ""); + if (id >= ARRAY_SIZE(map)) + return; + + if (map[id].id1 == NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_MAIN) + size = ALIGN(size, 0x1000) + 64 * 0x1000; /* per-subctx headers */ + + if (size >= 1 << 21) page = 21; + else if (size >= 1 << 16) page = 16; + else page = 12; + + if (map[id].id1 == NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_ATTRIBUTE_CB) + align = order_base_2(size); + else + align = page; + + if (WARN_ON(gr->ctxbuf_nr == ARRAY_SIZE(gr->ctxbuf))) + return; + + gr->ctxbuf[gr->ctxbuf_nr].bufferId = map[id].id1; + gr->ctxbuf[gr->ctxbuf_nr].size = size; + gr->ctxbuf[gr->ctxbuf_nr].page = page; + gr->ctxbuf[gr->ctxbuf_nr].align = align; + gr->ctxbuf[gr->ctxbuf_nr].global = map[id].global; + gr->ctxbuf[gr->ctxbuf_nr].init = map[id].init; + gr->ctxbuf[gr->ctxbuf_nr].ro = map[id].ro; + gr->ctxbuf_nr++; + + if (map[id].id1 == NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PRIV_ACCESS_MAP) { + if (WARN_ON(gr->ctxbuf_nr == ARRAY_SIZE(gr->ctxbuf))) + return; + + gr->ctxbuf[gr->ctxbuf_nr] = gr->ctxbuf[gr->ctxbuf_nr - 1]; + gr->ctxbuf[gr->ctxbuf_nr].bufferId = + NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_UNRESTRICTED_PRIV_ACCESS_MAP; + gr->ctxbuf_nr++; + } +} + +static int +r535_gr_get_ctxbufs_info(struct r535_gr *gr) +{ + NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS *info; + struct nvkm_subdev *subdev = &gr->base.engine.subdev; + struct nvkm_gsp *gsp = subdev->device->gsp; + + info = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO, + sizeof(*info)); + if (WARN_ON(IS_ERR(info))) + return PTR_ERR(info); + + for (int i = 0; i < ARRAY_SIZE(info->engineContextBuffersInfo[0].engine); i++) + r535_gr_get_ctxbuf_info(gr, i, &info->engineContextBuffersInfo[0].engine[i]); + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, info); + return 0; +} + +int +r535_gr_oneinit(struct nvkm_gr *base) +{ + struct r535_gr *gr = container_of(base, typeof(*gr), base); + struct nvkm_subdev *subdev = &gr->base.engine.subdev; + struct nvkm_device *device = subdev->device; + struct nvkm_gsp *gsp = device->gsp; + struct nvkm_rm *rm = gsp->rm; + struct { + struct nvkm_memory *inst; + struct nvkm_vmm *vmm; + struct nvkm_gsp_object chan; + struct nvkm_vma *vma[R515_GR_MAX_CTXBUFS]; + } golden = {}; + struct nvkm_gsp_object threed; + int ret; + + /* Allocate a channel to use for golden context init. */ + ret = nvkm_memory_new(device, NVKM_MEM_TARGET_INST, 0x12000, 0, true, &golden.inst); + if (ret) + goto done; + + ret = nvkm_vmm_new(device, 0x1000, 0, NULL, 0, NULL, "grGoldenVmm", &golden.vmm); + if (ret) + goto done; + + ret = r535_mmu_vaspace_new(golden.vmm, NVKM_RM_VASPACE, false); + if (ret) + goto done; + + ret = rm->api->fifo->chan.alloc(&golden.vmm->rm.device, NVKM_RM_CHAN(0), + 1, 0, true, rm->api->fifo->rsvd_chids, + nvkm_memory_addr(golden.inst), + nvkm_memory_addr(golden.inst) + 0x1000, + nvkm_memory_addr(golden.inst) + 0x2000, + golden.vmm, 0, 0x1000, &golden.chan); + if (ret) + goto done; + + /* Fetch context buffer info from RM and allocate each of them here to use + * during golden context init (or later as a global context buffer). + * + * Also build the information that'll be used to create channel contexts. + */ + ret = rm->api->gr->get_ctxbufs_info(gr); + if (ret) + goto done; + + /* Promote golden context to RM. */ + ret = r535_gr_promote_ctx(gr, true, golden.vmm, gr->ctxbuf_mem, golden.vma, &golden.chan); + if (ret) + goto done; + + /* Allocate 3D class on channel to trigger golden context init in RM. */ + ret = nvkm_gsp_rm_alloc(&golden.chan, NVKM_RM_THREED, rm->gpu->gr.class.threed, 0, &threed); + if (ret) + goto done; + + /* There's no need to keep the golden channel around, as RM caches the context. */ + nvkm_gsp_rm_free(&threed); +done: + nvkm_gsp_rm_free(&golden.chan); + for (int i = gr->ctxbuf_nr - 1; i >= 0; i--) + nvkm_vmm_put(golden.vmm, &golden.vma[i]); + nvkm_vmm_unref(&golden.vmm); + nvkm_memory_unref(&golden.inst); + return ret; + +} + +void * +r535_gr_dtor(struct nvkm_gr *base) +{ + struct r535_gr *gr = r535_gr(base); + + while (gr->ctxbuf_nr) + nvkm_memory_unref(&gr->ctxbuf_mem[--gr->ctxbuf_nr]); + + kfree(gr->base.func); + return gr; +} + +const struct nvkm_rm_api_gr +r535_gr = { + .get_ctxbufs_info = r535_gr_get_ctxbufs_info, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/gsp.c index 969f6b921fdb..baf42339f93e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/gsp.c @@ -19,9 +19,12 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ +#include <rm/rpc.h> + #include "priv.h" #include <core/pci.h> +#include <subdev/pci/priv.h> #include <subdev/timer.h> #include <subdev/vfn.h> #include <engine/fifo/chan.h> @@ -30,29 +33,11 @@ #include <nvfw/fw.h> -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0000.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0005.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl0080.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl2080.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080event.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080gpu.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl2080/ctrl2080internal.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> -#include <nvrm/535.113.01/common/shared/msgq/inc/msgq/msgq_priv.h> -#include <nvrm/535.113.01/common/uproc/os/common/include/libos_init_args.h> -#include <nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_sr_meta.h> -#include <nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/gsp/gsp_fw_wpr_meta.h> -#include <nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmRiscvUcode.h> -#include <nvrm/535.113.01/nvidia/arch/nvalloc/common/inc/rmgspseq.h> -#include <nvrm/535.113.01/nvidia/generated/g_allclasses.h> -#include <nvrm/535.113.01/nvidia/generated/g_os_nvoc.h> -#include <nvrm/535.113.01/nvidia/generated/g_rpc-structures.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_fw_heap.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_init_args.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/gsp/gsp_static_config.h> -#include <nvrm/535.113.01/nvidia/inc/kernel/gpu/intr/engine_idx.h> -#include <nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_global_enums.h> +#include "nvrm/gsp.h" +#include "nvrm/rpcfn.h" +#include "nvrm/msgfn.h" +#include "nvrm/event.h" +#include "nvrm/fifo.h" #include <linux/acpi.h> #include <linux/ctype.h> @@ -60,996 +45,6 @@ extern struct dentry *nouveau_debugfs_root; -#define GSP_MSG_MIN_SIZE GSP_PAGE_SIZE -#define GSP_MSG_MAX_SIZE (GSP_MSG_MIN_SIZE * 16) - -/** - * DOC: GSP message queue element - * - * https://github.com/NVIDIA/open-gpu-kernel-modules/blob/535/src/nvidia/inc/kernel/gpu/gsp/message_queue_priv.h - * - * The GSP command queue and status queue are message queues for the - * communication between software and GSP. The software submits the GSP - * RPC via the GSP command queue, GSP writes the status of the submitted - * RPC in the status queue. - * - * A GSP message queue element consists of three parts: - * - * - message element header (struct r535_gsp_msg), which mostly maintains - * the metadata for queuing the element. - * - * - RPC message header (struct nvfw_gsp_rpc), which maintains the info - * of the RPC. E.g., the RPC function number. - * - * - The payload, where the RPC message stays. E.g. the params of a - * specific RPC function. Some RPC functions also have their headers - * in the payload. E.g. rm_alloc, rm_control. - * - * The memory layout of a GSP message element can be illustrated below:: - * - * +------------------------+ - * | Message Element Header | - * | (r535_gsp_msg) | - * | | - * | (r535_gsp_msg.data) | - * | | | - * |----------V-------------| - * | GSP RPC Header | - * | (nvfw_gsp_rpc) | - * | | - * | (nvfw_gsp_rpc.data) | - * | | | - * |----------V-------------| - * | Payload | - * | | - * | header(optional) | - * | params | - * +------------------------+ - * - * The max size of a message queue element is 16 pages (including the - * headers). When a GSP message to be sent is larger than 16 pages, the - * message should be split into multiple elements and sent accordingly. - * - * In the bunch of the split elements, the first element has the expected - * function number, while the rest of the elements are sent with the - * function number NV_VGPU_MSG_FUNCTION_CONTINUATION_RECORD. - * - * GSP consumes the elements from the cmdq and always writes the result - * back to the msgq. The result is also formed as split elements. - * - * Terminology: - * - * - gsp_msg(msg): GSP message element (element header + GSP RPC header + - * payload) - * - gsp_rpc(rpc): GSP RPC (RPC header + payload) - * - gsp_rpc_buf: buffer for (GSP RPC header + payload) - * - gsp_rpc_len: size of (GSP RPC header + payload) - * - params_size: size of params in the payload - * - payload_size: size of (header if exists + params) in the payload - */ - -struct r535_gsp_msg { - u8 auth_tag_buffer[16]; - u8 aad_buffer[16]; - u32 checksum; - u32 sequence; - u32 elem_count; - u32 pad; - u8 data[]; -}; - -struct nvfw_gsp_rpc { - u32 header_version; - u32 signature; - u32 length; - u32 function; - u32 rpc_result; - u32 rpc_result_private; - u32 sequence; - union { - u32 spare; - u32 cpuRmGfid; - }; - u8 data[]; -}; - -#define GSP_MSG_HDR_SIZE offsetof(struct r535_gsp_msg, data) - -#define to_gsp_hdr(p, header) \ - container_of((void *)p, typeof(*header), data) - -#define to_payload_hdr(p, header) \ - container_of((void *)p, typeof(*header), params) - -static int -r535_rpc_status_to_errno(uint32_t rpc_status) -{ - switch (rpc_status) { - case 0x55: /* NV_ERR_NOT_READY */ - case 0x66: /* NV_ERR_TIMEOUT_RETRY */ - return -EBUSY; - case 0x51: /* NV_ERR_NO_MEMORY */ - return -ENOMEM; - default: - return -EINVAL; - } -} - -static int -r535_gsp_msgq_wait(struct nvkm_gsp *gsp, u32 gsp_rpc_len, int *ptime) -{ - u32 size, rptr = *gsp->msgq.rptr; - int used; - - size = DIV_ROUND_UP(GSP_MSG_HDR_SIZE + gsp_rpc_len, - GSP_PAGE_SIZE); - if (WARN_ON(!size || size >= gsp->msgq.cnt)) - return -EINVAL; - - do { - u32 wptr = *gsp->msgq.wptr; - - used = wptr + gsp->msgq.cnt - rptr; - if (used >= gsp->msgq.cnt) - used -= gsp->msgq.cnt; - if (used >= size) - break; - - usleep_range(1, 2); - } while (--(*ptime)); - - if (WARN_ON(!*ptime)) - return -ETIMEDOUT; - - return used; -} - -static struct r535_gsp_msg * -r535_gsp_msgq_get_entry(struct nvkm_gsp *gsp) -{ - u32 rptr = *gsp->msgq.rptr; - - /* Skip the first page, which is the message queue info */ - return (void *)((u8 *)gsp->shm.msgq.ptr + GSP_PAGE_SIZE + - rptr * GSP_PAGE_SIZE); -} - -/** - * DOC: Receive a GSP message queue element - * - * Receiving a GSP message queue element from the message queue consists of - * the following steps: - * - * - Peek the element from the queue: r535_gsp_msgq_peek(). - * Peek the first page of the element to determine the total size of the - * message before allocating the proper memory. - * - * - Allocate memory for the message. - * Once the total size of the message is determined from the GSP message - * queue element, the caller of r535_gsp_msgq_recv() allocates the - * required memory. - * - * - Receive the message: r535_gsp_msgq_recv(). - * Copy the message into the allocated memory. Advance the read pointer. - * If the message is a large GSP message, r535_gsp_msgq_recv() calls - * r535_gsp_msgq_recv_one_elem() repeatedly to receive continuation parts - * until the complete message is received. - * r535_gsp_msgq_recv() assembles the payloads of cotinuation parts into - * the return of the large GSP message. - * - * - Free the allocated memory: r535_gsp_msg_done(). - * The user is responsible for freeing the memory allocated for the GSP - * message pages after they have been processed. - */ -static void * -r535_gsp_msgq_peek(struct nvkm_gsp *gsp, u32 gsp_rpc_len, int *retries) -{ - struct r535_gsp_msg *mqe; - int ret; - - ret = r535_gsp_msgq_wait(gsp, gsp_rpc_len, retries); - if (ret < 0) - return ERR_PTR(ret); - - mqe = r535_gsp_msgq_get_entry(gsp); - - return mqe->data; -} - -struct r535_gsp_msg_info { - int *retries; - u32 gsp_rpc_len; - void *gsp_rpc_buf; - bool continuation; -}; - -static void -r535_gsp_msg_dump(struct nvkm_gsp *gsp, struct nvfw_gsp_rpc *msg, int lvl); - -static void * -r535_gsp_msgq_recv_one_elem(struct nvkm_gsp *gsp, - struct r535_gsp_msg_info *info) -{ - u8 *buf = info->gsp_rpc_buf; - u32 rptr = *gsp->msgq.rptr; - struct r535_gsp_msg *mqe; - u32 size, expected, len; - int ret; - - expected = info->gsp_rpc_len; - - ret = r535_gsp_msgq_wait(gsp, expected, info->retries); - if (ret < 0) - return ERR_PTR(ret); - - mqe = r535_gsp_msgq_get_entry(gsp); - - if (info->continuation) { - struct nvfw_gsp_rpc *rpc = (struct nvfw_gsp_rpc *)mqe->data; - - if (rpc->function != NV_VGPU_MSG_FUNCTION_CONTINUATION_RECORD) { - nvkm_error(&gsp->subdev, - "Not a continuation of a large RPC\n"); - r535_gsp_msg_dump(gsp, rpc, NV_DBG_ERROR); - return ERR_PTR(-EIO); - } - } - - size = ALIGN(expected + GSP_MSG_HDR_SIZE, GSP_PAGE_SIZE); - - len = ((gsp->msgq.cnt - rptr) * GSP_PAGE_SIZE) - sizeof(*mqe); - len = min_t(u32, expected, len); - - if (info->continuation) - memcpy(buf, mqe->data + sizeof(struct nvfw_gsp_rpc), - len - sizeof(struct nvfw_gsp_rpc)); - else - memcpy(buf, mqe->data, len); - - expected -= len; - - if (expected) { - mqe = (void *)((u8 *)gsp->shm.msgq.ptr + 0x1000 + 0 * 0x1000); - memcpy(buf + len, mqe, expected); - } - - rptr = (rptr + DIV_ROUND_UP(size, GSP_PAGE_SIZE)) % gsp->msgq.cnt; - - mb(); - (*gsp->msgq.rptr) = rptr; - return buf; -} - -static void * -r535_gsp_msgq_recv(struct nvkm_gsp *gsp, u32 gsp_rpc_len, int *retries) -{ - struct r535_gsp_msg *mqe; - const u32 max_rpc_size = GSP_MSG_MAX_SIZE - sizeof(*mqe); - struct nvfw_gsp_rpc *rpc; - struct r535_gsp_msg_info info = {0}; - u32 expected = gsp_rpc_len; - void *buf; - - mqe = r535_gsp_msgq_get_entry(gsp); - rpc = (struct nvfw_gsp_rpc *)mqe->data; - - if (WARN_ON(rpc->length > max_rpc_size)) - return NULL; - - buf = kvmalloc(max_t(u32, rpc->length, expected), GFP_KERNEL); - if (!buf) - return ERR_PTR(-ENOMEM); - - info.gsp_rpc_buf = buf; - info.retries = retries; - info.gsp_rpc_len = rpc->length; - - buf = r535_gsp_msgq_recv_one_elem(gsp, &info); - if (IS_ERR(buf)) { - kvfree(info.gsp_rpc_buf); - info.gsp_rpc_buf = NULL; - return buf; - } - - if (expected <= max_rpc_size) - return buf; - - info.gsp_rpc_buf += info.gsp_rpc_len; - expected -= info.gsp_rpc_len; - - while (expected) { - u32 size; - - rpc = r535_gsp_msgq_peek(gsp, sizeof(*rpc), info.retries); - if (IS_ERR_OR_NULL(rpc)) { - kfree(buf); - return rpc; - } - - info.gsp_rpc_len = rpc->length; - info.continuation = true; - - rpc = r535_gsp_msgq_recv_one_elem(gsp, &info); - if (IS_ERR_OR_NULL(rpc)) { - kfree(buf); - return rpc; - } - - size = info.gsp_rpc_len - sizeof(*rpc); - expected -= size; - info.gsp_rpc_buf += size; - } - - rpc = buf; - rpc->length = gsp_rpc_len; - return buf; -} - -static int -r535_gsp_cmdq_push(struct nvkm_gsp *gsp, void *rpc) -{ - struct r535_gsp_msg *msg = to_gsp_hdr(rpc, msg); - struct r535_gsp_msg *cqe; - u32 gsp_rpc_len = msg->checksum; - u64 *ptr = (void *)msg; - u64 *end; - u64 csum = 0; - int free, time = 1000000; - u32 wptr, size, step, len; - u32 off = 0; - - len = ALIGN(GSP_MSG_HDR_SIZE + gsp_rpc_len, GSP_PAGE_SIZE); - - end = (u64 *)((char *)ptr + len); - msg->pad = 0; - msg->checksum = 0; - msg->sequence = gsp->cmdq.seq++; - msg->elem_count = DIV_ROUND_UP(len, 0x1000); - - while (ptr < end) - csum ^= *ptr++; - - msg->checksum = upper_32_bits(csum) ^ lower_32_bits(csum); - - wptr = *gsp->cmdq.wptr; - do { - do { - free = *gsp->cmdq.rptr + gsp->cmdq.cnt - wptr - 1; - if (free >= gsp->cmdq.cnt) - free -= gsp->cmdq.cnt; - if (free >= 1) - break; - - usleep_range(1, 2); - } while(--time); - - if (WARN_ON(!time)) { - kvfree(msg); - return -ETIMEDOUT; - } - - cqe = (void *)((u8 *)gsp->shm.cmdq.ptr + 0x1000 + wptr * 0x1000); - step = min_t(u32, free, (gsp->cmdq.cnt - wptr)); - size = min_t(u32, len, step * GSP_PAGE_SIZE); - - memcpy(cqe, (u8 *)msg + off, size); - - wptr += DIV_ROUND_UP(size, 0x1000); - if (wptr == gsp->cmdq.cnt) - wptr = 0; - - off += size; - len -= size; - } while (len); - - nvkm_trace(&gsp->subdev, "cmdq: wptr %d\n", wptr); - wmb(); - (*gsp->cmdq.wptr) = wptr; - mb(); - - nvkm_falcon_wr32(&gsp->falcon, 0xc00, 0x00000000); - - kvfree(msg); - return 0; -} - -static void * -r535_gsp_cmdq_get(struct nvkm_gsp *gsp, u32 gsp_rpc_len) -{ - struct r535_gsp_msg *msg; - u32 size = GSP_MSG_HDR_SIZE + gsp_rpc_len; - - size = ALIGN(size, GSP_MSG_MIN_SIZE); - msg = kvzalloc(size, GFP_KERNEL); - if (!msg) - return ERR_PTR(-ENOMEM); - - msg->checksum = gsp_rpc_len; - return msg->data; -} - -static void -r535_gsp_msg_done(struct nvkm_gsp *gsp, struct nvfw_gsp_rpc *msg) -{ - kvfree(msg); -} - -static void -r535_gsp_msg_dump(struct nvkm_gsp *gsp, struct nvfw_gsp_rpc *msg, int lvl) -{ - if (gsp->subdev.debug >= lvl) { - nvkm_printk__(&gsp->subdev, lvl, info, - "msg fn:%d len:0x%x/0x%zx res:0x%x resp:0x%x\n", - msg->function, msg->length, msg->length - sizeof(*msg), - msg->rpc_result, msg->rpc_result_private); - print_hex_dump(KERN_INFO, "msg: ", DUMP_PREFIX_OFFSET, 16, 1, - msg->data, msg->length - sizeof(*msg), true); - } -} - -static struct nvfw_gsp_rpc * -r535_gsp_msg_recv(struct nvkm_gsp *gsp, int fn, u32 gsp_rpc_len) -{ - struct nvkm_subdev *subdev = &gsp->subdev; - struct nvfw_gsp_rpc *rpc; - int retries = 4000000, i; - -retry: - rpc = r535_gsp_msgq_peek(gsp, sizeof(*rpc), &retries); - if (IS_ERR_OR_NULL(rpc)) - return rpc; - - rpc = r535_gsp_msgq_recv(gsp, gsp_rpc_len, &retries); - if (IS_ERR_OR_NULL(rpc)) - return rpc; - - if (rpc->rpc_result) { - r535_gsp_msg_dump(gsp, rpc, NV_DBG_ERROR); - r535_gsp_msg_done(gsp, rpc); - return ERR_PTR(-EINVAL); - } - - r535_gsp_msg_dump(gsp, rpc, NV_DBG_TRACE); - - if (fn && rpc->function == fn) { - if (gsp_rpc_len) { - if (rpc->length < gsp_rpc_len) { - nvkm_error(subdev, "rpc len %d < %d\n", - rpc->length, gsp_rpc_len); - r535_gsp_msg_dump(gsp, rpc, NV_DBG_ERROR); - r535_gsp_msg_done(gsp, rpc); - return ERR_PTR(-EIO); - } - - return rpc; - } - - r535_gsp_msg_done(gsp, rpc); - return NULL; - } - - for (i = 0; i < gsp->msgq.ntfy_nr; i++) { - struct nvkm_gsp_msgq_ntfy *ntfy = &gsp->msgq.ntfy[i]; - - if (ntfy->fn == rpc->function) { - if (ntfy->func) - ntfy->func(ntfy->priv, ntfy->fn, rpc->data, - rpc->length - sizeof(*rpc)); - break; - } - } - - if (i == gsp->msgq.ntfy_nr) - r535_gsp_msg_dump(gsp, rpc, NV_DBG_WARN); - - r535_gsp_msg_done(gsp, rpc); - if (fn) - goto retry; - - if (*gsp->msgq.rptr != *gsp->msgq.wptr) - goto retry; - - return NULL; -} - -static int -r535_gsp_msg_ntfy_add(struct nvkm_gsp *gsp, u32 fn, nvkm_gsp_msg_ntfy_func func, void *priv) -{ - int ret = 0; - - mutex_lock(&gsp->msgq.mutex); - if (WARN_ON(gsp->msgq.ntfy_nr >= ARRAY_SIZE(gsp->msgq.ntfy))) { - ret = -ENOSPC; - } else { - gsp->msgq.ntfy[gsp->msgq.ntfy_nr].fn = fn; - gsp->msgq.ntfy[gsp->msgq.ntfy_nr].func = func; - gsp->msgq.ntfy[gsp->msgq.ntfy_nr].priv = priv; - gsp->msgq.ntfy_nr++; - } - mutex_unlock(&gsp->msgq.mutex); - return ret; -} - -static int -r535_gsp_rpc_poll(struct nvkm_gsp *gsp, u32 fn) -{ - void *repv; - - mutex_lock(&gsp->cmdq.mutex); - repv = r535_gsp_msg_recv(gsp, fn, 0); - mutex_unlock(&gsp->cmdq.mutex); - if (IS_ERR(repv)) - return PTR_ERR(repv); - - return 0; -} - -static void * -r535_gsp_rpc_handle_reply(struct nvkm_gsp *gsp, u32 fn, - enum nvkm_gsp_rpc_reply_policy policy, - u32 gsp_rpc_len) -{ - struct nvfw_gsp_rpc *reply; - void *repv = NULL; - - switch (policy) { - case NVKM_GSP_RPC_REPLY_NOWAIT: - break; - case NVKM_GSP_RPC_REPLY_RECV: - reply = r535_gsp_msg_recv(gsp, fn, gsp_rpc_len); - if (!IS_ERR_OR_NULL(reply)) - repv = reply->data; - else - repv = reply; - break; - case NVKM_GSP_RPC_REPLY_POLL: - repv = r535_gsp_msg_recv(gsp, fn, 0); - break; - } - - return repv; -} - -static void * -r535_gsp_rpc_send(struct nvkm_gsp *gsp, void *payload, - enum nvkm_gsp_rpc_reply_policy policy, u32 gsp_rpc_len) -{ - struct nvfw_gsp_rpc *rpc = to_gsp_hdr(payload, rpc); - u32 fn = rpc->function; - int ret; - - if (gsp->subdev.debug >= NV_DBG_TRACE) { - nvkm_trace(&gsp->subdev, "rpc fn:%d len:0x%x/0x%zx\n", rpc->function, - rpc->length, rpc->length - sizeof(*rpc)); - print_hex_dump(KERN_INFO, "rpc: ", DUMP_PREFIX_OFFSET, 16, 1, - rpc->data, rpc->length - sizeof(*rpc), true); - } - - ret = r535_gsp_cmdq_push(gsp, rpc); - if (ret) - return ERR_PTR(ret); - - return r535_gsp_rpc_handle_reply(gsp, fn, policy, gsp_rpc_len); -} - -static void -r535_gsp_event_dtor(struct nvkm_gsp_event *event) -{ - struct nvkm_gsp_device *device = event->device; - struct nvkm_gsp_client *client = device->object.client; - struct nvkm_gsp *gsp = client->gsp; - - mutex_lock(&gsp->client_id.mutex); - if (event->func) { - list_del(&event->head); - event->func = NULL; - } - mutex_unlock(&gsp->client_id.mutex); - - nvkm_gsp_rm_free(&event->object); - event->device = NULL; -} - -static int -r535_gsp_device_event_get(struct nvkm_gsp_event *event) -{ - struct nvkm_gsp_device *device = event->device; - NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS *ctrl; - - ctrl = nvkm_gsp_rm_ctrl_get(&device->subdevice, - NV2080_CTRL_CMD_EVENT_SET_NOTIFICATION, sizeof(*ctrl)); - if (IS_ERR(ctrl)) - return PTR_ERR(ctrl); - - ctrl->event = event->id; - ctrl->action = NV2080_CTRL_EVENT_SET_NOTIFICATION_ACTION_REPEAT; - return nvkm_gsp_rm_ctrl_wr(&device->subdevice, ctrl); -} - -static int -r535_gsp_device_event_ctor(struct nvkm_gsp_device *device, u32 handle, u32 id, - nvkm_gsp_event_func func, struct nvkm_gsp_event *event) -{ - struct nvkm_gsp_client *client = device->object.client; - struct nvkm_gsp *gsp = client->gsp; - NV0005_ALLOC_PARAMETERS *args; - int ret; - - args = nvkm_gsp_rm_alloc_get(&device->subdevice, handle, - NV01_EVENT_KERNEL_CALLBACK_EX, sizeof(*args), - &event->object); - if (IS_ERR(args)) - return PTR_ERR(args); - - args->hParentClient = client->object.handle; - args->hSrcResource = 0; - args->hClass = NV01_EVENT_KERNEL_CALLBACK_EX; - args->notifyIndex = NV01_EVENT_CLIENT_RM | id; - args->data = NULL; - - ret = nvkm_gsp_rm_alloc_wr(&event->object, args); - if (ret) - return ret; - - event->device = device; - event->id = id; - - ret = r535_gsp_device_event_get(event); - if (ret) { - nvkm_gsp_event_dtor(event); - return ret; - } - - mutex_lock(&gsp->client_id.mutex); - event->func = func; - list_add(&event->head, &client->events); - mutex_unlock(&gsp->client_id.mutex); - return 0; -} - -static void -r535_gsp_device_dtor(struct nvkm_gsp_device *device) -{ - nvkm_gsp_rm_free(&device->subdevice); - nvkm_gsp_rm_free(&device->object); -} - -static int -r535_gsp_subdevice_ctor(struct nvkm_gsp_device *device) -{ - NV2080_ALLOC_PARAMETERS *args; - - return nvkm_gsp_rm_alloc(&device->object, 0x5d1d0000, NV20_SUBDEVICE_0, sizeof(*args), - &device->subdevice); -} - -static int -r535_gsp_device_ctor(struct nvkm_gsp_client *client, struct nvkm_gsp_device *device) -{ - NV0080_ALLOC_PARAMETERS *args; - int ret; - - args = nvkm_gsp_rm_alloc_get(&client->object, 0xde1d0000, NV01_DEVICE_0, sizeof(*args), - &device->object); - if (IS_ERR(args)) - return PTR_ERR(args); - - args->hClientShare = client->object.handle; - - ret = nvkm_gsp_rm_alloc_wr(&device->object, args); - if (ret) - return ret; - - ret = r535_gsp_subdevice_ctor(device); - if (ret) - nvkm_gsp_rm_free(&device->object); - - return ret; -} - -static void -r535_gsp_client_dtor(struct nvkm_gsp_client *client) -{ - struct nvkm_gsp *gsp = client->gsp; - - nvkm_gsp_rm_free(&client->object); - - mutex_lock(&gsp->client_id.mutex); - idr_remove(&gsp->client_id.idr, client->object.handle & 0xffff); - mutex_unlock(&gsp->client_id.mutex); - - client->gsp = NULL; -} - -static int -r535_gsp_client_ctor(struct nvkm_gsp *gsp, struct nvkm_gsp_client *client) -{ - NV0000_ALLOC_PARAMETERS *args; - int ret; - - mutex_lock(&gsp->client_id.mutex); - ret = idr_alloc(&gsp->client_id.idr, client, 0, 0xffff + 1, GFP_KERNEL); - mutex_unlock(&gsp->client_id.mutex); - if (ret < 0) - return ret; - - client->gsp = gsp; - client->object.client = client; - INIT_LIST_HEAD(&client->events); - - args = nvkm_gsp_rm_alloc_get(&client->object, 0xc1d00000 | ret, NV01_ROOT, sizeof(*args), - &client->object); - if (IS_ERR(args)) { - r535_gsp_client_dtor(client); - return ret; - } - - args->hClient = client->object.handle; - args->processID = ~0; - - ret = nvkm_gsp_rm_alloc_wr(&client->object, args); - if (ret) { - r535_gsp_client_dtor(client); - return ret; - } - - return 0; -} - -static int -r535_gsp_rpc_rm_free(struct nvkm_gsp_object *object) -{ - struct nvkm_gsp_client *client = object->client; - struct nvkm_gsp *gsp = client->gsp; - rpc_free_v03_00 *rpc; - - nvkm_debug(&gsp->subdev, "cli:0x%08x obj:0x%08x free\n", - client->object.handle, object->handle); - - rpc = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_FREE, sizeof(*rpc)); - if (WARN_ON(IS_ERR_OR_NULL(rpc))) - return -EIO; - - rpc->params.hRoot = client->object.handle; - rpc->params.hObjectParent = 0; - rpc->params.hObjectOld = object->handle; - return nvkm_gsp_rpc_wr(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV); -} - -static void -r535_gsp_rpc_rm_alloc_done(struct nvkm_gsp_object *object, void *params) -{ - rpc_gsp_rm_alloc_v03_00 *rpc = to_payload_hdr(params, rpc); - - nvkm_gsp_rpc_done(object->client->gsp, rpc); -} - -static void * -r535_gsp_rpc_rm_alloc_push(struct nvkm_gsp_object *object, void *params) -{ - rpc_gsp_rm_alloc_v03_00 *rpc = to_payload_hdr(params, rpc); - struct nvkm_gsp *gsp = object->client->gsp; - void *ret = NULL; - - rpc = nvkm_gsp_rpc_push(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV, sizeof(*rpc)); - if (IS_ERR_OR_NULL(rpc)) - return rpc; - - if (rpc->status) { - ret = ERR_PTR(r535_rpc_status_to_errno(rpc->status)); - if (PTR_ERR(ret) != -EAGAIN && PTR_ERR(ret) != -EBUSY) - nvkm_error(&gsp->subdev, "RM_ALLOC: 0x%x\n", rpc->status); - } - - nvkm_gsp_rpc_done(gsp, rpc); - - return ret; -} - -static void * -r535_gsp_rpc_rm_alloc_get(struct nvkm_gsp_object *object, u32 oclass, - u32 params_size) -{ - struct nvkm_gsp_client *client = object->client; - struct nvkm_gsp *gsp = client->gsp; - rpc_gsp_rm_alloc_v03_00 *rpc; - - nvkm_debug(&gsp->subdev, "cli:0x%08x obj:0x%08x new obj:0x%08x\n", - client->object.handle, object->parent->handle, - object->handle); - - nvkm_debug(&gsp->subdev, "cls:0x%08x params_size:%d\n", oclass, - params_size); - - rpc = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_GSP_RM_ALLOC, - sizeof(*rpc) + params_size); - if (IS_ERR(rpc)) - return rpc; - - rpc->hClient = client->object.handle; - rpc->hParent = object->parent->handle; - rpc->hObject = object->handle; - rpc->hClass = oclass; - rpc->status = 0; - rpc->paramsSize = params_size; - return rpc->params; -} - -static void -r535_gsp_rpc_rm_ctrl_done(struct nvkm_gsp_object *object, void *params) -{ - rpc_gsp_rm_control_v03_00 *rpc = to_payload_hdr(params, rpc); - - if (!params) - return; - nvkm_gsp_rpc_done(object->client->gsp, rpc); -} - -static int -r535_gsp_rpc_rm_ctrl_push(struct nvkm_gsp_object *object, void **params, u32 repc) -{ - rpc_gsp_rm_control_v03_00 *rpc = to_payload_hdr((*params), rpc); - struct nvkm_gsp *gsp = object->client->gsp; - int ret = 0; - - rpc = nvkm_gsp_rpc_push(gsp, rpc, NVKM_GSP_RPC_REPLY_RECV, repc); - if (IS_ERR_OR_NULL(rpc)) { - *params = NULL; - return PTR_ERR(rpc); - } - - if (rpc->status) { - ret = r535_rpc_status_to_errno(rpc->status); - if (ret != -EAGAIN && ret != -EBUSY) - nvkm_error(&gsp->subdev, "cli:0x%08x obj:0x%08x ctrl cmd:0x%08x failed: 0x%08x\n", - object->client->object.handle, object->handle, rpc->cmd, rpc->status); - } - - if (repc) - *params = rpc->params; - else - nvkm_gsp_rpc_done(gsp, rpc); - - return ret; -} - -static void * -r535_gsp_rpc_rm_ctrl_get(struct nvkm_gsp_object *object, u32 cmd, u32 params_size) -{ - struct nvkm_gsp_client *client = object->client; - struct nvkm_gsp *gsp = client->gsp; - rpc_gsp_rm_control_v03_00 *rpc; - - nvkm_debug(&gsp->subdev, "cli:0x%08x obj:0x%08x ctrl cmd:0x%08x params_size:%d\n", - client->object.handle, object->handle, cmd, params_size); - - rpc = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_GSP_RM_CONTROL, - sizeof(*rpc) + params_size); - if (IS_ERR(rpc)) - return rpc; - - rpc->hClient = client->object.handle; - rpc->hObject = object->handle; - rpc->cmd = cmd; - rpc->status = 0; - rpc->paramsSize = params_size; - return rpc->params; -} - -static void -r535_gsp_rpc_done(struct nvkm_gsp *gsp, void *repv) -{ - struct nvfw_gsp_rpc *rpc = container_of(repv, typeof(*rpc), data); - - r535_gsp_msg_done(gsp, rpc); -} - -static void * -r535_gsp_rpc_get(struct nvkm_gsp *gsp, u32 fn, u32 payload_size) -{ - struct nvfw_gsp_rpc *rpc; - - rpc = r535_gsp_cmdq_get(gsp, ALIGN(sizeof(*rpc) + payload_size, - sizeof(u64))); - if (IS_ERR(rpc)) - return ERR_CAST(rpc); - - rpc->header_version = 0x03000000; - rpc->signature = ('C' << 24) | ('P' << 16) | ('R' << 8) | 'V'; - rpc->function = fn; - rpc->rpc_result = 0xffffffff; - rpc->rpc_result_private = 0xffffffff; - rpc->length = sizeof(*rpc) + payload_size; - return rpc->data; -} - -static void * -r535_gsp_rpc_push(struct nvkm_gsp *gsp, void *payload, - enum nvkm_gsp_rpc_reply_policy policy, u32 gsp_rpc_len) -{ - struct nvfw_gsp_rpc *rpc = to_gsp_hdr(payload, rpc); - struct r535_gsp_msg *msg = to_gsp_hdr(rpc, msg); - const u32 max_rpc_size = GSP_MSG_MAX_SIZE - sizeof(*msg); - const u32 max_payload_size = max_rpc_size - sizeof(*rpc); - u32 payload_size = rpc->length - sizeof(*rpc); - void *repv; - - mutex_lock(&gsp->cmdq.mutex); - if (payload_size > max_payload_size) { - const u32 fn = rpc->function; - u32 remain_payload_size = payload_size; - - /* Adjust length, and send initial RPC. */ - rpc->length = sizeof(*rpc) + max_payload_size; - msg->checksum = rpc->length; - - repv = r535_gsp_rpc_send(gsp, payload, NVKM_GSP_RPC_REPLY_NOWAIT, 0); - if (IS_ERR(repv)) - goto done; - - payload += max_payload_size; - remain_payload_size -= max_payload_size; - - /* Remaining chunks sent as CONTINUATION_RECORD RPCs. */ - while (remain_payload_size) { - u32 size = min(remain_payload_size, - max_payload_size); - void *next; - - next = r535_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_CONTINUATION_RECORD, size); - if (IS_ERR(next)) { - repv = next; - goto done; - } - - memcpy(next, payload, size); - - repv = r535_gsp_rpc_send(gsp, next, NVKM_GSP_RPC_REPLY_NOWAIT, 0); - if (IS_ERR(repv)) - goto done; - - payload += size; - remain_payload_size -= size; - } - - /* Wait for reply. */ - repv = r535_gsp_rpc_handle_reply(gsp, fn, policy, payload_size + - sizeof(*rpc)); - } else { - repv = r535_gsp_rpc_send(gsp, payload, policy, gsp_rpc_len); - } - -done: - mutex_unlock(&gsp->cmdq.mutex); - return repv; -} - -const struct nvkm_gsp_rm -r535_gsp_rm = { - .rpc_get = r535_gsp_rpc_get, - .rpc_push = r535_gsp_rpc_push, - .rpc_done = r535_gsp_rpc_done, - - .rm_ctrl_get = r535_gsp_rpc_rm_ctrl_get, - .rm_ctrl_push = r535_gsp_rpc_rm_ctrl_push, - .rm_ctrl_done = r535_gsp_rpc_rm_ctrl_done, - - .rm_alloc_get = r535_gsp_rpc_rm_alloc_get, - .rm_alloc_push = r535_gsp_rpc_rm_alloc_push, - .rm_alloc_done = r535_gsp_rpc_rm_alloc_done, - - .rm_free = r535_gsp_rpc_rm_free, - - .client_ctor = r535_gsp_client_ctor, - .client_dtor = r535_gsp_client_dtor, - - .device_ctor = r535_gsp_device_ctor, - .device_dtor = r535_gsp_device_dtor, - - .event_ctor = r535_gsp_device_event_ctor, - .event_dtor = r535_gsp_event_dtor, -}; - static void r535_gsp_msgq_work(struct work_struct *work) { @@ -1092,10 +87,52 @@ r535_gsp_intr(struct nvkm_inth *inth) return IRQ_HANDLED; } +static bool +r535_gsp_xlat_mc_engine_idx(u32 mc_engine_idx, enum nvkm_subdev_type *ptype, int *pinst) +{ + switch (mc_engine_idx) { + case MC_ENGINE_IDX_GSP: + *ptype = NVKM_SUBDEV_GSP; + *pinst = 0; + return true; + case MC_ENGINE_IDX_DISP: + *ptype = NVKM_ENGINE_DISP; + *pinst = 0; + return true; + case MC_ENGINE_IDX_CE0 ... MC_ENGINE_IDX_CE9: + *ptype = NVKM_ENGINE_CE; + *pinst = mc_engine_idx - MC_ENGINE_IDX_CE0; + return true; + case MC_ENGINE_IDX_GR0: + *ptype = NVKM_ENGINE_GR; + *pinst = 0; + return true; + case MC_ENGINE_IDX_NVDEC0 ... MC_ENGINE_IDX_NVDEC7: + *ptype = NVKM_ENGINE_NVDEC; + *pinst = mc_engine_idx - MC_ENGINE_IDX_NVDEC0; + return true; + case MC_ENGINE_IDX_MSENC ... MC_ENGINE_IDX_MSENC2: + *ptype = NVKM_ENGINE_NVENC; + *pinst = mc_engine_idx - MC_ENGINE_IDX_MSENC; + return true; + case MC_ENGINE_IDX_NVJPEG0 ... MC_ENGINE_IDX_NVJPEG7: + *ptype = NVKM_ENGINE_NVJPG; + *pinst = mc_engine_idx - MC_ENGINE_IDX_NVJPEG0; + return true; + case MC_ENGINE_IDX_OFA0: + *ptype = NVKM_ENGINE_OFA; + *pinst = 0; + return true; + default: + return false; + } +} + static int r535_gsp_intr_get_table(struct nvkm_gsp *gsp) { NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS *ctrl; + const struct nvkm_rm_api *rmapi = gsp->rm->api; int ret = 0; ctrl = nvkm_gsp_rm_ctrl_get(&gsp->internal.device.subdevice, @@ -1118,42 +155,8 @@ r535_gsp_intr_get_table(struct nvkm_gsp *gsp) ctrl->table[i].engineIdx, ctrl->table[i].pmcIntrMask, ctrl->table[i].vectorStall, ctrl->table[i].vectorNonStall); - switch (ctrl->table[i].engineIdx) { - case MC_ENGINE_IDX_GSP: - type = NVKM_SUBDEV_GSP; - inst = 0; - break; - case MC_ENGINE_IDX_DISP: - type = NVKM_ENGINE_DISP; - inst = 0; - break; - case MC_ENGINE_IDX_CE0 ... MC_ENGINE_IDX_CE9: - type = NVKM_ENGINE_CE; - inst = ctrl->table[i].engineIdx - MC_ENGINE_IDX_CE0; - break; - case MC_ENGINE_IDX_GR0: - type = NVKM_ENGINE_GR; - inst = 0; - break; - case MC_ENGINE_IDX_NVDEC0 ... MC_ENGINE_IDX_NVDEC7: - type = NVKM_ENGINE_NVDEC; - inst = ctrl->table[i].engineIdx - MC_ENGINE_IDX_NVDEC0; - break; - case MC_ENGINE_IDX_MSENC ... MC_ENGINE_IDX_MSENC2: - type = NVKM_ENGINE_NVENC; - inst = ctrl->table[i].engineIdx - MC_ENGINE_IDX_MSENC; - break; - case MC_ENGINE_IDX_NVJPEG0 ... MC_ENGINE_IDX_NVJPEG7: - type = NVKM_ENGINE_NVJPG; - inst = ctrl->table[i].engineIdx - MC_ENGINE_IDX_NVJPEG0; - break; - case MC_ENGINE_IDX_OFA0: - type = NVKM_ENGINE_OFA; - inst = 0; - break; - default: + if (!rmapi->gsp->xlat_mc_engine_idx(ctrl->table[i].engineIdx, &type, &inst)) continue; - } if (WARN_ON(gsp->intr_nr == ARRAY_SIZE(gsp->intr))) { ret = -ENOSPC; @@ -1171,35 +174,14 @@ r535_gsp_intr_get_table(struct nvkm_gsp *gsp) return ret; } -static int -r535_gsp_rpc_get_gsp_static_info(struct nvkm_gsp *gsp) +void +r535_gsp_get_static_info_fb(struct nvkm_gsp *gsp, + const struct NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS *info) { - GspStaticConfigInfo *rpc; int last_usable = -1; - rpc = nvkm_gsp_rpc_rd(gsp, NV_VGPU_MSG_FUNCTION_GET_GSP_STATIC_INFO, sizeof(*rpc)); - if (IS_ERR(rpc)) - return PTR_ERR(rpc); - - gsp->internal.client.object.client = &gsp->internal.client; - gsp->internal.client.object.parent = NULL; - gsp->internal.client.object.handle = rpc->hInternalClient; - gsp->internal.client.gsp = gsp; - - gsp->internal.device.object.client = &gsp->internal.client; - gsp->internal.device.object.parent = &gsp->internal.client.object; - gsp->internal.device.object.handle = rpc->hInternalDevice; - - gsp->internal.device.subdevice.client = &gsp->internal.client; - gsp->internal.device.subdevice.parent = &gsp->internal.device.object; - gsp->internal.device.subdevice.handle = rpc->hInternalSubdevice; - - gsp->bar.rm_bar1_pdb = rpc->bar1PdeBase; - gsp->bar.rm_bar2_pdb = rpc->bar2PdeBase; - - for (int i = 0; i < rpc->fbRegionInfoParams.numFBRegions; i++) { - NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO *reg = - &rpc->fbRegionInfoParams.fbRegion[i]; + for (int i = 0; i < info->numFBRegions; i++) { + const NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO *reg = &info->fbRegion[i]; nvkm_debug(&gsp->subdev, "fb region %d: " "%016llx-%016llx rsvd:%016llx perf:%08x comp:%d iso:%d prot:%d\n", i, @@ -1221,10 +203,38 @@ r535_gsp_rpc_get_gsp_static_info(struct nvkm_gsp *gsp) } if (last_usable >= 0) { - u32 rsvd_base = rpc->fbRegionInfoParams.fbRegion[last_usable].limit + 1; + u32 rsvd_base = info->fbRegion[last_usable].limit + 1; gsp->fb.rsvd_size = gsp->fb.heap.addr - rsvd_base; } +} + +static int +r535_gsp_get_static_info(struct nvkm_gsp *gsp) +{ + GspStaticConfigInfo *rpc; + + rpc = nvkm_gsp_rpc_rd(gsp, NV_VGPU_MSG_FUNCTION_GET_GSP_STATIC_INFO, sizeof(*rpc)); + if (IS_ERR(rpc)) + return PTR_ERR(rpc); + + gsp->internal.client.object.client = &gsp->internal.client; + gsp->internal.client.object.parent = NULL; + gsp->internal.client.object.handle = rpc->hInternalClient; + gsp->internal.client.gsp = gsp; + + gsp->internal.device.object.client = &gsp->internal.client; + gsp->internal.device.object.parent = &gsp->internal.client.object; + gsp->internal.device.object.handle = rpc->hInternalDevice; + + gsp->internal.device.subdevice.client = &gsp->internal.client; + gsp->internal.device.subdevice.parent = &gsp->internal.device.object; + gsp->internal.device.subdevice.handle = rpc->hInternalSubdevice; + + gsp->bar.rm_bar1_pdb = rpc->bar1PdeBase; + gsp->bar.rm_bar2_pdb = rpc->bar2PdeBase; + + r535_gsp_get_static_info_fb(gsp, &rpc->fbRegionInfoParams); for (int gpc = 0; gpc < ARRAY_SIZE(rpc->tpcInfo); gpc++) { if (rpc->gpcInfo.gpcMask & BIT(gpc)) { @@ -1237,7 +247,7 @@ r535_gsp_rpc_get_gsp_static_info(struct nvkm_gsp *gsp) return 0; } -static void +void nvkm_gsp_mem_dtor(struct nvkm_gsp_mem *mem) { if (mem->data) { @@ -1266,7 +276,7 @@ nvkm_gsp_mem_dtor(struct nvkm_gsp_mem *mem) * so we take a device reference to ensure its lifetime. The reference is * dropped in the destructor. */ -static int +int nvkm_gsp_mem_ctor(struct nvkm_gsp *gsp, size_t size, struct nvkm_gsp_mem *mem) { mem->data = dma_alloc_coherent(gsp->subdev.device->dev, size, &mem->addr, GFP_KERNEL); @@ -1283,9 +293,10 @@ static int r535_gsp_postinit(struct nvkm_gsp *gsp) { struct nvkm_device *device = gsp->subdev.device; + const struct nvkm_rm_api *rmapi = gsp->rm->api; int ret; - ret = r535_gsp_rpc_get_gsp_static_info(gsp); + ret = rmapi->gsp->get_static_info(gsp); if (WARN_ON(ret)) return ret; @@ -1698,7 +709,7 @@ fail: } #if defined(CONFIG_ACPI) && defined(CONFIG_X86) -static void +void r535_gsp_acpi_caps(acpi_handle handle, CAPS_METHOD_DATA *caps) { const guid_t NVOP_DSM_GUID = @@ -1732,7 +743,7 @@ r535_gsp_acpi_caps(acpi_handle handle, CAPS_METHOD_DATA *caps) kfree(argv4.buffer.pointer); } -static void +void r535_gsp_acpi_jt(acpi_handle handle, JT_METHOD_DATA *jt) { const guid_t JT_DSM_GUID = @@ -1824,7 +835,7 @@ r535_gsp_acpi_mux(acpi_handle handle, DOD_METHOD_DATA *dod, MUX_METHOD_DATA *mux } } -static void +void r535_gsp_acpi_dod(acpi_handle handle, DOD_METHOD_DATA *dod) { acpi_status status; @@ -1877,7 +888,7 @@ r535_gsp_acpi_info(struct nvkm_gsp *gsp, ACPI_METHOD_DATA *acpi) } static int -r535_gsp_rpc_set_system_info(struct nvkm_gsp *gsp) +r535_gsp_set_system_info(struct nvkm_gsp *gsp) { struct nvkm_device *device = gsp->subdev.device; struct nvkm_device_pci *pdev = container_of(device, typeof(*pdev), device); @@ -1890,13 +901,13 @@ r535_gsp_rpc_set_system_info(struct nvkm_gsp *gsp) if (IS_ERR(info)) return PTR_ERR(info); - info->gpuPhysAddr = device->func->resource_addr(device, 0); - info->gpuPhysFbAddr = device->func->resource_addr(device, 1); - info->gpuPhysInstAddr = device->func->resource_addr(device, 3); + info->gpuPhysAddr = device->func->resource_addr(device, NVKM_BAR0_PRI); + info->gpuPhysFbAddr = device->func->resource_addr(device, NVKM_BAR1_FB); + info->gpuPhysInstAddr = device->func->resource_addr(device, NVKM_BAR2_INST); info->nvDomainBusDeviceFunc = pci_dev_id(pdev->pdev); info->maxUserVa = TASK_SIZE; - info->pciConfigMirrorBase = 0x088000; - info->pciConfigMirrorSize = 0x001000; + info->pciConfigMirrorBase = device->pci->func->cfg.addr; + info->pciConfigMirrorSize = device->pci->func->cfg.size; r535_gsp_acpi_info(gsp, &info->acpiMethodData); return nvkm_gsp_rpc_wr(gsp, info, NVKM_GSP_RPC_REPLY_NOWAIT); @@ -1917,33 +928,6 @@ r535_gsp_msg_os_error_log(void *priv, u32 fn, void *repv, u32 repc) } static int -r535_gsp_msg_rc_triggered(void *priv, u32 fn, void *repv, u32 repc) -{ - rpc_rc_triggered_v17_02 *msg = repv; - struct nvkm_gsp *gsp = priv; - struct nvkm_subdev *subdev = &gsp->subdev; - struct nvkm_chan *chan; - unsigned long flags; - - if (WARN_ON(repc < sizeof(*msg))) - return -EINVAL; - - nvkm_error(subdev, "rc engn:%08x chid:%d type:%d scope:%d part:%d\n", - msg->nv2080EngineType, msg->chid, msg->exceptType, msg->scope, - msg->partitionAttributionId); - - chan = nvkm_chan_get_chid(&subdev->device->fifo->engine, msg->chid / 8, &flags); - if (!chan) { - nvkm_error(subdev, "rc chid:%d not found!\n", msg->chid); - return 0; - } - - nvkm_chan_error(chan, false); - nvkm_chan_put(&chan, flags); - return 0; -} - -static int r535_gsp_msg_mmu_fault_queued(void *priv, u32 fn, void *repv, u32 repc) { struct nvkm_gsp *gsp = priv; @@ -2136,97 +1120,6 @@ r535_gsp_msg_run_cpu_sequencer(void *priv, u32 fn, void *repv, u32 repc) } static int -r535_gsp_booter_unload(struct nvkm_gsp *gsp, u32 mbox0, u32 mbox1) -{ - struct nvkm_subdev *subdev = &gsp->subdev; - struct nvkm_device *device = subdev->device; - u32 wpr2_hi; - int ret; - - wpr2_hi = nvkm_rd32(device, 0x1fa828); - if (!wpr2_hi) { - nvkm_debug(subdev, "WPR2 not set - skipping booter unload\n"); - return 0; - } - - ret = nvkm_falcon_fw_boot(&gsp->booter.unload, &gsp->subdev, true, &mbox0, &mbox1, 0, 0); - if (WARN_ON(ret)) - return ret; - - wpr2_hi = nvkm_rd32(device, 0x1fa828); - if (WARN_ON(wpr2_hi)) - return -EIO; - - return 0; -} - -static int -r535_gsp_booter_load(struct nvkm_gsp *gsp, u32 mbox0, u32 mbox1) -{ - int ret; - - ret = nvkm_falcon_fw_boot(&gsp->booter.load, &gsp->subdev, true, &mbox0, &mbox1, 0, 0); - if (ret) - return ret; - - nvkm_falcon_wr32(&gsp->falcon, 0x080, gsp->boot.app_version); - - if (WARN_ON(!nvkm_falcon_riscv_active(&gsp->falcon))) - return -EIO; - - return 0; -} - -static int -r535_gsp_wpr_meta_init(struct nvkm_gsp *gsp) -{ - GspFwWprMeta *meta; - int ret; - - ret = nvkm_gsp_mem_ctor(gsp, 0x1000, &gsp->wpr_meta); - if (ret) - return ret; - - meta = gsp->wpr_meta.data; - - meta->magic = GSP_FW_WPR_META_MAGIC; - meta->revision = GSP_FW_WPR_META_REVISION; - - meta->sysmemAddrOfRadix3Elf = gsp->radix3.lvl0.addr; - meta->sizeOfRadix3Elf = gsp->fb.wpr2.elf.size; - - meta->sysmemAddrOfBootloader = gsp->boot.fw.addr; - meta->sizeOfBootloader = gsp->boot.fw.size; - meta->bootloaderCodeOffset = gsp->boot.code_offset; - meta->bootloaderDataOffset = gsp->boot.data_offset; - meta->bootloaderManifestOffset = gsp->boot.manifest_offset; - - meta->sysmemAddrOfSignature = gsp->sig.addr; - meta->sizeOfSignature = gsp->sig.size; - - meta->gspFwRsvdStart = gsp->fb.heap.addr; - meta->nonWprHeapOffset = gsp->fb.heap.addr; - meta->nonWprHeapSize = gsp->fb.heap.size; - meta->gspFwWprStart = gsp->fb.wpr2.addr; - meta->gspFwHeapOffset = gsp->fb.wpr2.heap.addr; - meta->gspFwHeapSize = gsp->fb.wpr2.heap.size; - meta->gspFwOffset = gsp->fb.wpr2.elf.addr; - meta->bootBinOffset = gsp->fb.wpr2.boot.addr; - meta->frtsOffset = gsp->fb.wpr2.frts.addr; - meta->frtsSize = gsp->fb.wpr2.frts.size; - meta->gspFwWprEnd = ALIGN_DOWN(gsp->fb.bios.vga_workspace.addr, 0x20000); - meta->fbSize = gsp->fb.size; - meta->vgaWorkspaceOffset = gsp->fb.bios.vga_workspace.addr; - meta->vgaWorkspaceSize = gsp->fb.bios.vga_workspace.size; - meta->bootCount = 0; - meta->partitionRpcAddr = 0; - meta->partitionRpcRequestOffset = 0; - meta->partitionRpcReplyOffset = 0; - meta->verified = 0; - return 0; -} - -static int r535_gsp_shared_init(struct nvkm_gsp *gsp) { struct { @@ -2277,23 +1170,11 @@ r535_gsp_shared_init(struct nvkm_gsp *gsp) return 0; } -static int -r535_gsp_rmargs_init(struct nvkm_gsp *gsp, bool resume) +static void +r535_gsp_set_rmargs(struct nvkm_gsp *gsp, bool resume) { - GSP_ARGUMENTS_CACHED *args; - int ret; - - if (!resume) { - ret = r535_gsp_shared_init(gsp); - if (ret) - return ret; + GSP_ARGUMENTS_CACHED *args = gsp->rmargs.data; - ret = nvkm_gsp_mem_ctor(gsp, 0x1000, &gsp->rmargs); - if (ret) - return ret; - } - - args = gsp->rmargs.data; args->messageQueueInitArguments.sharedMemPhysAddr = gsp->shm.mem.addr; args->messageQueueInitArguments.pageTableEntryCount = gsp->shm.ptes.nr; args->messageQueueInitArguments.cmdQueueOffset = @@ -2310,7 +1191,24 @@ r535_gsp_rmargs_init(struct nvkm_gsp *gsp, bool resume) args->srInitArguments.flags = 0; args->srInitArguments.bInPMTransition = 1; } +} +static int +r535_gsp_rmargs_init(struct nvkm_gsp *gsp, bool resume) +{ + int ret; + + if (!resume) { + ret = r535_gsp_shared_init(gsp); + if (ret) + return ret; + + ret = nvkm_gsp_mem_ctor(gsp, 0x1000, &gsp->rmargs); + if (ret) + return ret; + } + + gsp->rm->api->gsp->set_rmargs(gsp, resume); return 0; } @@ -2803,18 +1701,22 @@ lvl1_fail: return ret; } +static u32 +r535_gsp_sr_data_size(struct nvkm_gsp *gsp) +{ + GspFwWprMeta *meta = gsp->wpr_meta.data; + + return meta->gspFwWprEnd - meta->gspFwWprStart; +} + int r535_gsp_fini(struct nvkm_gsp *gsp, bool suspend) { - u32 mbox0 = 0xff, mbox1 = 0xff; + struct nvkm_rm *rm = gsp->rm; int ret; - if (!gsp->running) - return 0; - if (suspend) { - GspFwWprMeta *meta = gsp->wpr_meta.data; - u64 len = meta->gspFwWprEnd - meta->gspFwWprStart; + u32 len = rm->api->gsp->sr_data_size(gsp); GspFwSRMeta *sr; ret = nvkm_gsp_sg(gsp->subdev.device, len, &gsp->sr.sgt); @@ -2835,8 +1737,13 @@ r535_gsp_fini(struct nvkm_gsp *gsp, bool suspend) sr->sysmemAddrOfSuspendResumeData = gsp->sr.radix3.lvl0.addr; sr->sizeOfSuspendResumeData = len; - mbox0 = lower_32_bits(gsp->sr.meta.addr); - mbox1 = upper_32_bits(gsp->sr.meta.addr); + ret = rm->api->fbsr->suspend(gsp); + if (ret) { + nvkm_gsp_mem_dtor(&gsp->sr.meta); + nvkm_gsp_radix3_dtor(gsp, &gsp->sr.radix3); + nvkm_gsp_sg_free(gsp->subdev.device, &gsp->sr.sgt); + return ret; + } } ret = r535_gsp_rpc_unloading_guest_driver(gsp, suspend); @@ -2844,18 +1751,10 @@ r535_gsp_fini(struct nvkm_gsp *gsp, bool suspend) return ret; nvkm_msec(gsp->subdev.device, 2000, - if (nvkm_falcon_rd32(&gsp->falcon, 0x040) & 0x80000000) + if (nvkm_falcon_rd32(&gsp->falcon, 0x040) == 0x80000000) break; ); - nvkm_falcon_reset(&gsp->falcon); - - ret = nvkm_gsp_fwsec_sb(gsp); - WARN_ON(ret); - - ret = r535_gsp_booter_unload(gsp, mbox0, mbox1); - WARN_ON(ret); - gsp->running = false; return 0; } @@ -2863,23 +1762,12 @@ r535_gsp_fini(struct nvkm_gsp *gsp, bool suspend) int r535_gsp_init(struct nvkm_gsp *gsp) { - u32 mbox0, mbox1; int ret; - if (!gsp->sr.meta.data) { - mbox0 = lower_32_bits(gsp->wpr_meta.addr); - mbox1 = upper_32_bits(gsp->wpr_meta.addr); - } else { - r535_gsp_rmargs_init(gsp, true); - - mbox0 = lower_32_bits(gsp->sr.meta.addr); - mbox1 = upper_32_bits(gsp->sr.meta.addr); - } + nvkm_falcon_wr32(&gsp->falcon, 0x080, gsp->boot.app_version); - /* Execute booter to handle (eventually...) booting GSP-RM. */ - ret = r535_gsp_booter_load(gsp, mbox0, mbox1); - if (WARN_ON(ret)) - goto done; + if (WARN_ON(!nvkm_falcon_riscv_active(&gsp->falcon))) + return -EIO; ret = r535_gsp_rpc_poll(gsp, NV_VGPU_MSG_EVENT_GSP_INIT_DONE); if (ret) @@ -2889,6 +1777,8 @@ r535_gsp_init(struct nvkm_gsp *gsp) done: if (gsp->sr.meta.data) { + gsp->rm->api->fbsr->resume(gsp); + nvkm_gsp_mem_dtor(&gsp->sr.meta); nvkm_gsp_radix3_dtor(gsp, &gsp->sr.radix3); nvkm_gsp_sg_free(gsp->subdev.device, &gsp->sr.sgt); @@ -2950,19 +1840,6 @@ r535_gsp_elf_section(struct nvkm_gsp *gsp, const char *name, const u8 **pdata, u return -ENOENT; } -static void -r535_gsp_dtor_fws(struct nvkm_gsp *gsp) -{ - nvkm_firmware_put(gsp->fws.bl); - gsp->fws.bl = NULL; - nvkm_firmware_put(gsp->fws.booter.unload); - gsp->fws.booter.unload = NULL; - nvkm_firmware_put(gsp->fws.booter.load); - gsp->fws.booter.load = NULL; - nvkm_firmware_put(gsp->fws.rm); - gsp->fws.rm = NULL; -} - #ifdef CONFIG_DEBUG_FS struct r535_gsp_log { @@ -3196,10 +2073,16 @@ r535_gsp_dtor(struct nvkm_gsp *gsp) nvkm_falcon_fw_dtor(&gsp->booter.unload); nvkm_falcon_fw_dtor(&gsp->booter.load); + nvkm_gsp_mem_dtor(&gsp->fmc.args); + kfree(gsp->fmc.sig); + kfree(gsp->fmc.pkey); + kfree(gsp->fmc.hash); + nvkm_gsp_mem_dtor(&gsp->fmc.fw); + mutex_destroy(&gsp->msgq.mutex); mutex_destroy(&gsp->cmdq.mutex); - r535_gsp_dtor_fws(gsp); + nvkm_gsp_dtor_fws(gsp); nvkm_gsp_mem_dtor(&gsp->rmargs); nvkm_gsp_mem_dtor(&gsp->wpr_meta); @@ -3212,10 +2095,17 @@ r535_gsp_dtor(struct nvkm_gsp *gsp) nvkm_gsp_mem_dtor(&gsp->logrm); } +static void +r535_gsp_drop_send_user_shared_data(struct nvkm_gsp *gsp) +{ + r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_GSP_SEND_USER_SHARED_DATA, NULL, NULL); +} + int r535_gsp_oneinit(struct nvkm_gsp *gsp) { struct nvkm_device *device = gsp->subdev.device; + const struct nvkm_rm_api *rmapi = gsp->rm->api; const u8 *data; u64 size; int ret; @@ -3223,16 +2113,6 @@ r535_gsp_oneinit(struct nvkm_gsp *gsp) mutex_init(&gsp->cmdq.mutex); mutex_init(&gsp->msgq.mutex); - ret = gsp->func->booter.ctor(gsp, "booter-load", gsp->fws.booter.load, - &device->sec2->falcon, &gsp->booter.load); - if (ret) - return ret; - - ret = gsp->func->booter.ctor(gsp, "booter-unload", gsp->fws.booter.unload, - &device->sec2->falcon, &gsp->booter.unload); - if (ret) - return ret; - /* Load GSP firmware from ELF image into DMA-accessible memory. */ ret = r535_gsp_elf_section(gsp, ".fwimage", &data, &size); if (ret) @@ -3261,65 +2141,29 @@ r535_gsp_oneinit(struct nvkm_gsp *gsp) r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_GSP_RUN_CPU_SEQUENCER, r535_gsp_msg_run_cpu_sequencer, gsp); r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_POST_EVENT, r535_gsp_msg_post_event, gsp); - r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_RC_TRIGGERED, - r535_gsp_msg_rc_triggered, gsp); + r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_RC_TRIGGERED, rmapi->fifo->rc_triggered, gsp); r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_MMU_FAULT_QUEUED, r535_gsp_msg_mmu_fault_queued, gsp); r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_OS_ERROR_LOG, r535_gsp_msg_os_error_log, gsp); r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_PERF_BRIDGELESS_INFO_UPDATE, NULL, NULL); r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_UCODE_LIBOS_PRINT, NULL, NULL); - r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_GSP_SEND_USER_SHARED_DATA, NULL, NULL); + if (rmapi->gsp->drop_send_user_shared_data) + rmapi->gsp->drop_send_user_shared_data(gsp); + if (rmapi->gsp->drop_post_nocat_record) + rmapi->gsp->drop_post_nocat_record(gsp); + ret = r535_gsp_rm_boot_ctor(gsp); if (ret) return ret; /* Release FW images - we've copied them to DMA buffers now. */ - r535_gsp_dtor_fws(gsp); - - /* Calculate FB layout. */ - gsp->fb.wpr2.frts.size = 0x100000; - gsp->fb.wpr2.frts.addr = ALIGN_DOWN(gsp->fb.bios.addr, 0x20000) - gsp->fb.wpr2.frts.size; - - gsp->fb.wpr2.boot.size = gsp->boot.fw.size; - gsp->fb.wpr2.boot.addr = ALIGN_DOWN(gsp->fb.wpr2.frts.addr - gsp->fb.wpr2.boot.size, 0x1000); - - gsp->fb.wpr2.elf.size = gsp->fw.len; - gsp->fb.wpr2.elf.addr = ALIGN_DOWN(gsp->fb.wpr2.boot.addr - gsp->fb.wpr2.elf.size, 0x10000); - - { - u32 fb_size_gb = DIV_ROUND_UP_ULL(gsp->fb.size, 1 << 30); - - gsp->fb.wpr2.heap.size = - gsp->func->wpr_heap.os_carveout_size + - gsp->func->wpr_heap.base_size + - ALIGN(GSP_FW_HEAP_PARAM_SIZE_PER_GB_FB * fb_size_gb, 1 << 20) + - ALIGN(GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE, 1 << 20); - - gsp->fb.wpr2.heap.size = max(gsp->fb.wpr2.heap.size, gsp->func->wpr_heap.min_size); - } - - gsp->fb.wpr2.heap.addr = ALIGN_DOWN(gsp->fb.wpr2.elf.addr - gsp->fb.wpr2.heap.size, 0x100000); - gsp->fb.wpr2.heap.size = ALIGN_DOWN(gsp->fb.wpr2.elf.addr - gsp->fb.wpr2.heap.addr, 0x100000); - - gsp->fb.wpr2.addr = ALIGN_DOWN(gsp->fb.wpr2.heap.addr - sizeof(GspFwWprMeta), 0x100000); - gsp->fb.wpr2.size = gsp->fb.wpr2.frts.addr + gsp->fb.wpr2.frts.size - gsp->fb.wpr2.addr; - - gsp->fb.heap.size = 0x100000; - gsp->fb.heap.addr = gsp->fb.wpr2.addr - gsp->fb.heap.size; - - ret = nvkm_gsp_fwsec_frts(gsp); - if (WARN_ON(ret)) - return ret; + nvkm_gsp_dtor_fws(gsp); ret = r535_gsp_libos_init(gsp); if (WARN_ON(ret)) return ret; - ret = r535_gsp_wpr_meta_init(gsp); - if (WARN_ON(ret)) - return ret; - - ret = r535_gsp_rpc_set_system_info(gsp); + ret = rmapi->gsp->set_system_info(gsp); if (WARN_ON(ret)) return ret; @@ -3327,76 +2171,17 @@ r535_gsp_oneinit(struct nvkm_gsp *gsp) if (WARN_ON(ret)) return ret; - /* Reset GSP into RISC-V mode. */ - ret = gsp->func->reset(gsp); - if (WARN_ON(ret)) - return ret; - - nvkm_falcon_wr32(&gsp->falcon, 0x040, lower_32_bits(gsp->libos.addr)); - nvkm_falcon_wr32(&gsp->falcon, 0x044, upper_32_bits(gsp->libos.addr)); - mutex_init(&gsp->client_id.mutex); idr_init(&gsp->client_id.idr); return 0; } -static int -r535_gsp_load_fw(struct nvkm_gsp *gsp, const char *name, const char *ver, - const struct firmware **pfw) -{ - char fwname[64]; - - snprintf(fwname, sizeof(fwname), "gsp/%s-%s", name, ver); - return nvkm_firmware_get(&gsp->subdev, fwname, 0, pfw); -} - -int -r535_gsp_load(struct nvkm_gsp *gsp, int ver, const struct nvkm_gsp_fwif *fwif) -{ - struct nvkm_subdev *subdev = &gsp->subdev; - int ret; - bool enable_gsp = fwif->enable; - -#if IS_ENABLED(CONFIG_DRM_NOUVEAU_GSP_DEFAULT) - enable_gsp = true; -#endif - if (!nvkm_boolopt(subdev->device->cfgopt, "NvGspRm", enable_gsp)) - return -EINVAL; - - if ((ret = r535_gsp_load_fw(gsp, "gsp", fwif->ver, &gsp->fws.rm)) || - (ret = r535_gsp_load_fw(gsp, "booter_load", fwif->ver, &gsp->fws.booter.load)) || - (ret = r535_gsp_load_fw(gsp, "booter_unload", fwif->ver, &gsp->fws.booter.unload)) || - (ret = r535_gsp_load_fw(gsp, "bootloader", fwif->ver, &gsp->fws.bl))) { - r535_gsp_dtor_fws(gsp); - return ret; - } - - return 0; -} - -#define NVKM_GSP_FIRMWARE(chip) \ -MODULE_FIRMWARE("nvidia/"#chip"/gsp/booter_load-535.113.01.bin"); \ -MODULE_FIRMWARE("nvidia/"#chip"/gsp/booter_unload-535.113.01.bin"); \ -MODULE_FIRMWARE("nvidia/"#chip"/gsp/bootloader-535.113.01.bin"); \ -MODULE_FIRMWARE("nvidia/"#chip"/gsp/gsp-535.113.01.bin") - -NVKM_GSP_FIRMWARE(tu102); -NVKM_GSP_FIRMWARE(tu104); -NVKM_GSP_FIRMWARE(tu106); - -NVKM_GSP_FIRMWARE(tu116); -NVKM_GSP_FIRMWARE(tu117); - -NVKM_GSP_FIRMWARE(ga100); - -NVKM_GSP_FIRMWARE(ga102); -NVKM_GSP_FIRMWARE(ga103); -NVKM_GSP_FIRMWARE(ga104); -NVKM_GSP_FIRMWARE(ga106); -NVKM_GSP_FIRMWARE(ga107); - -NVKM_GSP_FIRMWARE(ad102); -NVKM_GSP_FIRMWARE(ad103); -NVKM_GSP_FIRMWARE(ad104); -NVKM_GSP_FIRMWARE(ad106); -NVKM_GSP_FIRMWARE(ad107); +const struct nvkm_rm_api_gsp +r535_gsp = { + .set_rmargs = r535_gsp_set_rmargs, + .set_system_info = r535_gsp_set_system_info, + .get_static_info = r535_gsp_get_static_info, + .xlat_mc_engine_idx = r535_gsp_xlat_mc_engine_idx, + .drop_send_user_shared_data = r535_gsp_drop_send_user_shared_data, + .sr_data_size = r535_gsp_sr_data_size, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/ad102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvdec.c index 1b4619ff9e8e..a8c42ec0367b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/ad102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvdec.c @@ -19,26 +19,27 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "priv.h" +#include <rm/engine.h> -#include <subdev/gsp.h> +#include "nvrm/nvdec.h" -#include <nvif/class.h> +static int +r535_nvdec_alloc(struct nvkm_gsp_object *chan, u32 handle, u32 class, int inst, + struct nvkm_gsp_object *nvdec) +{ + NV_BSP_ALLOCATION_PARAMETERS *args; -static const struct nvkm_engine_func -ad102_nvenc = { - .sclass = { - { -1, -1, NVC9B7_VIDEO_ENCODER }, - {} - } -}; + args = nvkm_gsp_rm_alloc_get(chan, handle, class, sizeof(*args), nvdec); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); -int -ad102_nvenc_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_nvenc **pnvenc) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_nvenc_new(&ad102_nvenc, device, type, inst, pnvenc); + args->size = sizeof(*args); + args->engineInstance = inst; - return -ENODEV; + return nvkm_gsp_rm_alloc_wr(nvdec, args); } + +const struct nvkm_rm_api_engine +r535_nvdec = { + .alloc = r535_nvdec_alloc, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ad102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvenc.c index 7bfa6240d283..acb3ce8bb9de 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ad102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvenc.c @@ -19,28 +19,27 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "gf100.h" +#include <rm/engine.h> -#include <subdev/gsp.h> +#include "nvrm/nvenc.h" -#include <nvif/class.h> +static int +r535_nvenc_alloc(struct nvkm_gsp_object *chan, u32 handle, u32 class, int inst, + struct nvkm_gsp_object *nvenc) +{ + NV_MSENC_ALLOCATION_PARAMETERS *args; -static const struct gf100_gr_func -ad102_gr = { - .sclass = { - { -1, -1, FERMI_TWOD_A }, - { -1, -1, KEPLER_INLINE_TO_MEMORY_B }, - { -1, -1, ADA_A }, - { -1, -1, ADA_COMPUTE_A }, - {} - } -}; + args = nvkm_gsp_rm_alloc_get(chan, handle, class, sizeof(*args), nvenc); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); -int -ad102_gr_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, struct nvkm_gr **pgr) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_gr_new(&ad102_gr, device, type, inst, pgr); + args->size = sizeof(*args); + args->engineInstance = inst; - return -ENODEV; + return nvkm_gsp_rm_alloc_wr(nvenc, args); } + +const struct nvkm_rm_api_engine +r535_nvenc = { + .alloc = r535_nvenc_alloc, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvjpg.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvjpg.c new file mode 100644 index 000000000000..fbc4080ad8d8 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvjpg.c @@ -0,0 +1,45 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/engine.h> + +#include "nvrm/nvjpg.h" + +static int +r535_nvjpg_alloc(struct nvkm_gsp_object *chan, u32 handle, u32 class, int inst, + struct nvkm_gsp_object *nvjpg) +{ + NV_NVJPG_ALLOCATION_PARAMETERS *args; + + args = nvkm_gsp_rm_alloc_get(chan, handle, class, sizeof(*args), nvjpg); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); + + args->size = sizeof(*args); + args->engineInstance = inst; + + return nvkm_gsp_rm_alloc_wr(nvjpg, args); +} + +const struct nvkm_rm_api_engine +r535_nvjpg = { + .alloc = r535_nvjpg_alloc, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/alloc.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/alloc.h new file mode 100644 index 000000000000..cbc7e611fbda --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/alloc.h @@ -0,0 +1,36 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_ALLOC_H__ +#define __NVRM_ALLOC_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct rpc_gsp_rm_alloc_v03_00 +{ + NvHandle hClient; + NvHandle hParent; + NvHandle hObject; + NvU32 hClass; + NvU32 status; + NvU32 paramsSize; + NvU32 flags; + NvU8 reserved[4]; + NvU8 params[]; +} rpc_gsp_rm_alloc_v03_00; + +typedef struct NVOS00_PARAMETERS_v03_00 +{ + NvHandle hRoot; + NvHandle hObjectParent; + NvHandle hObjectOld; + NvV32 status; +} NVOS00_PARAMETERS_v03_00; + +typedef struct rpc_free_v03_00 +{ + NVOS00_PARAMETERS_v03_00 params; +} rpc_free_v03_00; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/bar.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/bar.h new file mode 100644 index 000000000000..60b0b08491ee --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/bar.h @@ -0,0 +1,29 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_BAR_H__ +#define __NVRM_BAR_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef enum +{ + NV_RPC_UPDATE_PDE_BAR_1, + NV_RPC_UPDATE_PDE_BAR_2, + NV_RPC_UPDATE_PDE_BAR_INVALID, +} NV_RPC_UPDATE_PDE_BAR_TYPE; + +typedef struct UpdateBarPde_v15_00 +{ + NV_RPC_UPDATE_PDE_BAR_TYPE barType; + NvU64 entryValue NV_ALIGN_BYTES(8); + NvU64 entryLevelShift NV_ALIGN_BYTES(8); +} UpdateBarPde_v15_00; + +typedef struct rpc_update_bar_pde_v15_00 +{ + UpdateBarPde_v15_00 info; +} rpc_update_bar_pde_v15_00; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ce.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ce.h new file mode 100644 index 000000000000..90b0325203d2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ce.h @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_CE_H__ +#define __NVRM_CE_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct NVC0B5_ALLOCATION_PARAMETERS { + NvU32 version; + NvU32 engineType; +} NVC0B5_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/client.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/client.h new file mode 100644 index 000000000000..df0e63c0cb6b --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/client.h @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_CLIENT_H__ +#define __NVRM_CLIENT_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV01_ROOT (0x0U) /* finn: Evaluated from "NV0000_ALLOC_PARAMETERS_MESSAGE_ID" */ + +#define NV_PROC_NAME_MAX_LENGTH 100U + +typedef struct NV0000_ALLOC_PARAMETERS { + NvHandle hClient; /* CORERM-2934: hClient must remain the first member until all allocations use these params */ + NvU32 processID; + char processName[NV_PROC_NAME_MAX_LENGTH]; +} NV0000_ALLOC_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ctrl.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ctrl.h new file mode 100644 index 000000000000..77f10acd82c9 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ctrl.h @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_CTRL_H__ +#define __NVRM_CTRL_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct rpc_gsp_rm_control_v03_00 +{ + NvHandle hClient; + NvHandle hObject; + NvU32 cmd; + NvU32 status; + NvU32 paramsSize; + NvU32 flags; + NvU8 params[]; +} rpc_gsp_rm_control_v03_00; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/device.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/device.h new file mode 100644 index 000000000000..3933b9ad61ce --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/device.h @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_DEVICE_H__ +#define __NVRM_DEVICE_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV01_DEVICE_0 (0x80U) /* finn: Evaluated from "NV0080_ALLOC_PARAMETERS_MESSAGE_ID" */ + +typedef struct NV0080_ALLOC_PARAMETERS { + NvU32 deviceId; + NvHandle hClientShare; + NvHandle hTargetClient; + NvHandle hTargetDevice; + NvV32 flags; + NV_DECLARE_ALIGNED(NvU64 vaSpaceSize, 8); + NV_DECLARE_ALIGNED(NvU64 vaStartInternal, 8); + NV_DECLARE_ALIGNED(NvU64 vaLimitInternal, 8); + NvV32 vaMode; +} NV0080_ALLOC_PARAMETERS; + +#define NV20_SUBDEVICE_0 (0x2080U) /* finn: Evaluated from "NV2080_ALLOC_PARAMETERS_MESSAGE_ID" */ + +typedef struct NV2080_ALLOC_PARAMETERS { + NvU32 subDeviceId; +} NV2080_ALLOC_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/disp.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/disp.h new file mode 100644 index 000000000000..7b7539639540 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/disp.h @@ -0,0 +1,741 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_DISP_H__ +#define __NVRM_DISP_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_WRITE_INST_MEM (0x20800a49) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS { + NV_DECLARE_ALIGNED(NvU64 instMemPhysAddr, 8); + NV_DECLARE_ALIGNED(NvU64 instMemSize, 8); + NvU32 instMemAddrSpace; + NvU32 instMemCpuCacheAttr; +} NV2080_CTRL_INTERNAL_DISPLAY_WRITE_INST_MEM_PARAMS; + +#define NV_MEMORY_WRITECOMBINED 2 + +#define NV04_DISPLAY_COMMON (0x00000073) + +#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_GET_STATIC_INFO (0x20800a01) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS { + NvU32 feHwSysCap; + NvU32 windowPresentMask; + NvBool bFbRemapperEnabled; + NvU32 numHeads; + NvBool bPrimaryVga; + NvU32 i2cPort; + NvU32 internalDispActiveMask; +} NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS; + +#define NV2080_CTRL_ACPI_DSM_READ_SIZE (0x1000) /* finn: Evaluated from "(4 * 1024)" */ + +#define NV2080_CTRL_CMD_INTERNAL_INIT_BRIGHTC_STATE_LOAD (0x20800ac6) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_INIT_BRIGHTC_STATE_LOAD_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_INIT_BRIGHTC_STATE_LOAD_PARAMS { + NvU32 status; + NvU16 backLightDataSize; + NvU8 backLightData[NV2080_CTRL_ACPI_DSM_READ_SIZE]; +} NV2080_CTRL_INTERNAL_INIT_BRIGHTC_STATE_LOAD_PARAMS; + +typedef struct NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT_PARAMS { + NvU32 subDeviceInstance; +} NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT_PARAMS; + +#define NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT (0x731365U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_SET_MANUAL_DISPLAYPORT_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_CMD_SYSTEM_GET_NUM_HEADS (0x730102U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_NUM_HEADS_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_NUM_HEADS_PARAMS { + NvU32 subDeviceInstance; + NvU32 flags; + NvU32 numHeads; +} NV0073_CTRL_SYSTEM_GET_NUM_HEADS_PARAMS; + +#define NV0073_CTRL_CMD_SPECIFIC_GET_ALL_HEAD_MASK (0x730287U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_ALL_HEAD_MASK_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_GET_ALL_HEAD_MASK_PARAMS { + NvU32 subDeviceInstance; + NvU32 headMask; +} NV0073_CTRL_SPECIFIC_GET_ALL_HEAD_MASK_PARAMS; + +#define NV0073_CTRL_CMD_SYSTEM_GET_SUPPORTED (0x730120U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayMask; + NvU32 displayMaskDDC; +} NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS; + +#define NV0073_CTRL_MAX_CONNECTORS 4U + +#define NV0073_CTRL_CMD_SPECIFIC_GET_CONNECTOR_DATA (0x730250U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 flags; + NvU32 DDCPartners; + NvU32 count; + struct { + NvU32 index; + NvU32 type; + NvU32 location; + } data[NV0073_CTRL_MAX_CONNECTORS]; + NvU32 platform; +} NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS; + +#define NV0073_CTRL_CMD_SPECIFIC_OR_GET_INFO (0x73028bU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 index; + NvU32 type; + NvU32 protocol; + NvU32 ditherType; + NvU32 ditherAlgo; + NvU32 location; + NvU32 rootPortId; + NvU32 dcbIndex; + NV_DECLARE_ALIGNED(NvU64 vbiosAddress, 8); + NvBool bIsLitByVbios; + NvBool bIsDispDynamic; +} NV0073_CTRL_SPECIFIC_OR_GET_INFO_PARAMS; + +#define NV0073_CTRL_SPECIFIC_OR_TYPE_NONE (0x00000000U) +#define NV0073_CTRL_SPECIFIC_OR_TYPE_DAC (0x00000001U) +#define NV0073_CTRL_SPECIFIC_OR_TYPE_SOR (0x00000002U) +#define NV0073_CTRL_SPECIFIC_OR_TYPE_PIOR (0x00000003U) +#define NV0073_CTRL_SPECIFIC_OR_TYPE_DSI (0x00000005U) + +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_DAC_RGB_CRT (0x00000000U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_LVDS_CUSTOM (0x00000000U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_SINGLE_TMDS_A (0x00000001U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_SINGLE_TMDS_B (0x00000002U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DUAL_TMDS (0x00000005U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DP_A (0x00000008U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DP_B (0x00000009U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_SOR_DSI (0x00000010U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_DSI (0x00000011U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_PIOR_EXT_TMDS_ENC (0x00000000U) +#define NV0073_CTRL_SPECIFIC_OR_PROTOCOL_UNKNOWN (0xFFFFFFFFU) + +#define NV0073_CTRL_CMD_DP_GET_CAPS (0x731369U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_CMD_DSC_CAP_PARAMS { + NvBool bDscSupported; + NvU32 encoderColorFormatMask; + NvU32 lineBufferSizeKB; + NvU32 rateBufferSizeKB; + NvU32 bitsPerPixelPrecision; + NvU32 maxNumHztSlices; + NvU32 lineBufferBitDepth; +} NV0073_CTRL_CMD_DSC_CAP_PARAMS; + +typedef struct NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS { + NvU32 subDeviceInstance; + NvU32 sorIndex; + NvU32 maxLinkRate; + NvU32 dpVersionsSupported; + NvU32 UHBRSupported; + NvBool bIsMultistreamSupported; + NvBool bIsSCEnabled; + NvBool bHasIncreasedWatermarkLimits; + NvBool bIsPC2Disabled; + NvBool isSingleHeadMSTSupported; + NvBool bFECSupported; + NvBool bIsTrainPhyRepeater; + NvBool bOverrideLinkBw; + NV0073_CTRL_CMD_DSC_CAP_PARAMS DSC; +} NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS; + +#define NV0073_CTRL_CMD_DP_GET_CAPS (0x731369U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID" */ +#define NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID (0x69U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2 0:0 +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4 1:1 +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE 2:0 +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_NONE (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_1_62 (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_2_70 (0x00000002U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_5_40 (0x00000003U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_8_10 (0x00000004U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_RGB (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_444 (0x00000002U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_NATIVE_422 (0x00000004U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_NATIVE_420 (0x00000008U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_16 (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_8 (0x00000002U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_4 (0x00000003U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_2 (0x00000004U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1 (0x00000005U) + +#define NV2080_NOTIFIERS_HOTPLUG (1) + +typedef struct { + NvU32 plugDisplayMask; + NvU32 unplugDisplayMask; +} Nv2080HotplugNotification; + +#define NV2080_NOTIFIERS_DP_IRQ (7) + +typedef struct Nv2080DpIrqNotificationRec { + NvU32 displayId; +} Nv2080DpIrqNotification; + +#define NV0073_CTRL_CMD_SYSTEM_GET_CONNECT_STATE (0x730122U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS { + NvU32 subDeviceInstance; + NvU32 flags; + NvU32 displayMask; + NvU32 retryTimeMs; +} NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS; + +#define NV0073_CTRL_CMD_DFP_GET_INFO (0x731140U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_GET_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DFP_GET_INFO_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 flags; + NvU32 flags2; +} NV0073_CTRL_DFP_GET_INFO_PARAMS; + +#define NV0073_CTRL_DFP_FLAGS_SIGNAL 2:0 +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_TMDS (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_LVDS (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_SDI (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_DISPLAYPORT (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_DSI (0x00000004U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_WRBK (0x00000005U) +#define NV0073_CTRL_DFP_FLAGS_LANE 5:3 +#define NV0073_CTRL_DFP_FLAGS_LANE_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_LANE_SINGLE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_LANE_DUAL (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_LANE_QUAD (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_LANE_OCT (0x00000004U) +#define NV0073_CTRL_DFP_FLAGS_LIMIT 6:6 +#define NV0073_CTRL_DFP_FLAGS_LIMIT_DISABLE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_LIMIT_60HZ_RR (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER 7:7 +#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER_NORMAL (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER_DISABLE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE 8:8 +#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE 9:9 +#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE 10:10 +#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE 11:11 +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE 12:12 +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED 14:14 +#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT 15:15 +#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT 16:16 +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT_PREFER_RBR (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW 19:17 +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_1_62GBPS (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_2_70GBPS (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_5_40GBPS (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_8_10GBPS (0x00000004U) +#define NV0073_CTRL_DFP_FLAGS_LINK 21:20 +#define NV0073_CTRL_DFP_FLAGS_LINK_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_LINK_SINGLE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_LINK_DUAL (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID 22:22 +#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID 24:23 +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_A (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_B (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_GANGED (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED 25:25 +#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_PHY_REPEATER_COUNT 29:26 +#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE 30:30 +#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE_TRUE (0x00000001U) + +#define NV0073_CTRL_CMD_SYSTEM_GET_ACTIVE (0x730126U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS { + NvU32 subDeviceInstance; + NvU32 head; + NvU32 flags; + NvU32 displayId; +} NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS; + +typedef NvU32 NV0073_CTRL_DFP_ASSIGN_SOR_LINKCONFIG; + +typedef struct NV0073_CTRL_DFP_ASSIGN_SOR_INFO { + NvU32 displayMask; + NvU32 sorType; +} NV0073_CTRL_DFP_ASSIGN_SOR_INFO; + +#define NV0073_CTRL_CMD_DFP_ASSIGN_SOR_MAX_SORS 4U + +#define NV0073_CTRL_CMD_DFP_ASSIGN_SOR (0x731152U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_ASSIGN_SOR_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DFP_ASSIGN_SOR_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU8 sorExcludeMask; + NvU32 slaveDisplayId; + NV0073_CTRL_DFP_ASSIGN_SOR_LINKCONFIG forceSublinkConfig; + NvBool bIs2Head1Or; + NvU32 sorAssignList[NV0073_CTRL_CMD_DFP_ASSIGN_SOR_MAX_SORS]; + NV0073_CTRL_DFP_ASSIGN_SOR_INFO sorAssignListWithTag[NV0073_CTRL_CMD_DFP_ASSIGN_SOR_MAX_SORS]; + NvU8 reservedSorMask; + NvU32 flags; +} NV0073_CTRL_DFP_ASSIGN_SOR_PARAMS; + +#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_AUDIO 0:0 +#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_AUDIO_OPTIMAL (0x00000001U) +#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_AUDIO_DEFAULT (0x00000000U) +#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_ACTIVE_SOR_NOT_AUDIO_CAPABLE 1:1 +#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_ACTIVE_SOR_NOT_AUDIO_CAPABLE_NO (0x00000000U) +#define NV0073_CTRL_DFP_ASSIGN_SOR_FLAGS_ACTIVE_SOR_NOT_AUDIO_CAPABLE_YES (0x00000001U) + +#define NV0073_CTRL_CMD_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS (0x730292U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_CMD_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS (0x730291U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 brightness; + NvBool bUncalibrated; +} NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS; + +#define NV0073_CTRL_CMD_DFP_SET_ELD_AUDIO_CAPS (0x731144U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_SET_ELD_AUDIO_CAP_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_ELD_BUFFER 96U + +typedef struct NV0073_CTRL_DFP_SET_ELD_AUDIO_CAP_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 numELDSize; + NvU8 bufferELD[NV0073_CTRL_DFP_ELD_AUDIO_CAPS_ELD_BUFFER]; + NvU32 maxFreqSupported; + NvU32 ctrl; + NvU32 deviceEntry; +} NV0073_CTRL_DFP_SET_ELD_AUDIO_CAP_PARAMS; + +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_PD 0:0 +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_PD_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_PD_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_ELDV 1:1 +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_ELDV_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_ELD_AUDIO_CAPS_CTRL_ELDV_TRUE (0x00000001U) + +#define NV0073_CTRL_SPECIFIC_GET_EDID_MAX_EDID_BYTES 2048U + +#define NV0073_CTRL_CMD_SPECIFIC_GET_EDID_V2 (0x730245U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 bufferSize; + NvU32 flags; + NvU8 edidBuffer[NV0073_CTRL_SPECIFIC_GET_EDID_MAX_EDID_BYTES]; +} NV0073_CTRL_SPECIFIC_GET_EDID_V2_PARAMS; + +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_ENABLE (0x730273U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_HDMI_ENABLE_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_SET_HDMI_ENABLE_PARAMS { + NvU8 subDeviceInstance; + NvU32 displayId; + NvU8 enable; +} NV0073_CTRL_SPECIFIC_SET_HDMI_ENABLE_PARAMS; + +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS (0x730293U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 caps; +} NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS; + +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS (0x730293U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_HDMI_SINK_CAPS_PARAMS_MESSAGE_ID" */ +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_GT_340MHZ_CLOCK_SUPPORTED 0:0 +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_GT_340MHZ_CLOCK_SUPPORTED_FALSE (0x00000000U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_GT_340MHZ_CLOCK_SUPPORTED_TRUE (0x00000001U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_LTE_340MHZ_SCRAMBLING_SUPPORTED 1:1 +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_LTE_340MHZ_SCRAMBLING_SUPPORTED_FALSE (0x00000000U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_LTE_340MHZ_SCRAMBLING_SUPPORTED_TRUE (0x00000001U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_SCDC_SUPPORTED 2:2 +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_SCDC_SUPPORTED_FALSE (0x00000000U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_SCDC_SUPPORTED_TRUE (0x00000001U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED 5:3 +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_NONE (0x00000000U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_3LANES_3G (0x00000001U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_3LANES_6G (0x00000002U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_6G (0x00000003U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_8G (0x00000004U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_10G (0x00000005U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_MAX_FRL_RATE_SUPPORTED_4LANES_12G (0x00000006U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_12_SUPPORTED 6:6 +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_12_SUPPORTED_FALSE (0x00000000U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_12_SUPPORTED_TRUE (0x00000001U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED 9:7 +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_NONE (0x00000000U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_3LANES_3G (0x00000001U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_3LANES_6G (0x00000002U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_6G (0x00000003U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_8G (0x00000004U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_10G (0x00000005U) +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_SINK_CAPS_DSC_MAX_FRL_RATE_SUPPORTED_4LANES_12G (0x00000006U) + +#define NV0073_CTRL_SET_OD_MAX_PACKET_SIZE 36U + +#define NV0073_CTRL_CMD_SPECIFIC_SET_OD_PACKET (0x730288U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_OD_PACKET_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_SET_OD_PACKET_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 transmitControl; + NvU32 packetSize; + NvU32 targetHead; + NvBool bUsePsrHeadforSdp; + NvU8 aPacket[NV0073_CTRL_SET_OD_MAX_PACKET_SIZE]; +} NV0073_CTRL_SPECIFIC_SET_OD_PACKET_PARAMS; + +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ENABLE 0:0 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ENABLE_NO (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ENABLE_YES (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_OTHER_FRAME 1:1 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_OTHER_FRAME_DISABLE (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_OTHER_FRAME_ENABLE (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SINGLE_FRAME 2:2 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SINGLE_FRAME_DISABLE (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SINGLE_FRAME_ENABLE (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ON_HBLANK 3:3 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ON_HBLANK_DISABLE (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_ON_HBLANK_ENABLE (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_IMMEDIATE 4:4 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_IMMEDIATE_DISABLE (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_IMMEDIATE_ENABLE (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_VIDEO_FMT 5:5 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_VIDEO_FMT_SW_CONTROLLED (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_VIDEO_FMT_HW_CONTROLLED (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_STEREO_POLARITY 6:6 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_STEREO_POLARITY_FALSE (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_STEREO_POLARITY_TRUE (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_SELF_REFRESH_SETTING 7:7 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_SELF_REFRESH_SETTING_FALSE (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_SET_SELF_REFRESH_SETTING_TRUE (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_GEN_INFOFRAME_MODE 9:8 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_GEN_INFOFRAME_MODE_INFOFRAME0 (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_GEN_INFOFRAME_MODE_INFOFRAME1 (0x0000001U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_RESERVED_LEGACY_MODE 31:31 +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_RESERVED_LEGACY_MODE_NO (0x0000000U) +#define NV0073_CTRL_SPECIFIC_SET_OD_PACKET_TRANSMIT_CONTROL_RESERVED_LEGACY_MODE_YES (0x0000001U) + +#define NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM (0x730275U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM_PARAMS { + NvU8 subDeviceInstance; + NvU32 displayId; + NvU8 mute; +} NV0073_CTRL_CMD_SPECIFIC_SET_HDMI_AUDIO_MUTESTREAM_PARAMS; + +#define NV0073_CTRL_DP_AUXCH_MAX_DATA_SIZE 16U + +#define NV0073_CTRL_CMD_DP_AUXCH_CTRL (0x731341U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_AUXCH_CTRL_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DP_AUXCH_CTRL_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvBool bAddrOnly; + NvU32 cmd; + NvU32 addr; + NvU8 data[NV0073_CTRL_DP_AUXCH_MAX_DATA_SIZE]; + NvU32 size; + NvU32 replyType; + NvU32 retryTimeMs; +} NV0073_CTRL_DP_AUXCH_CTRL_PARAMS; + +#define NV0073_CTRL_DP_AUXCH_CMD_TYPE 3:3 +#define NV0073_CTRL_DP_AUXCH_CMD_TYPE_I2C (0x00000000U) +#define NV0073_CTRL_DP_AUXCH_CMD_TYPE_AUX (0x00000001U) +#define NV0073_CTRL_DP_AUXCH_CMD_I2C_MOT 2:2 +#define NV0073_CTRL_DP_AUXCH_CMD_I2C_MOT_FALSE (0x00000000U) +#define NV0073_CTRL_DP_AUXCH_CMD_I2C_MOT_TRUE (0x00000001U) +#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE 1:0 +#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE_WRITE (0x00000000U) +#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE_READ (0x00000001U) +#define NV0073_CTRL_DP_AUXCH_CMD_REQ_TYPE_WRITE_STATUS (0x00000002U) + +#define NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES (0x731377U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES 8U + +typedef struct NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS { + // In + NvU32 subDeviceInstance; + NvU32 displayId; + NvU16 linkRateTbl[NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES]; + + // Out + NvU8 linkBwTbl[NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES]; + NvU8 linkBwCount; +} NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS; + +#define NV0073_CTRL_CMD_DP_CTRL (0x731343U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_CTRL_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DP_CTRL_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 cmd; + NvU32 data; + NvU32 err; + NvU32 retryTimeMs; + NvU32 eightLaneDpcdBaseAddr; +} NV0073_CTRL_DP_CTRL_PARAMS; + +#define NV0073_CTRL_DP_CMD_SET_LANE_COUNT 0:0 +#define NV0073_CTRL_DP_CMD_SET_LANE_COUNT_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_SET_LANE_COUNT_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_SET_LINK_BW 1:1 +#define NV0073_CTRL_DP_CMD_SET_LINK_BW_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_SET_LINK_BW_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_DISABLE_DOWNSPREAD 2:2 +#define NV0073_CTRL_DP_CMD_DISABLE_DOWNSPREAD_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_DISABLE_DOWNSPREAD_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_UNUSED 3:3 +#define NV0073_CTRL_DP_CMD_SET_FORMAT_MODE 4:4 +#define NV0073_CTRL_DP_CMD_SET_FORMAT_MODE_SINGLE_STREAM (0x00000000U) +#define NV0073_CTRL_DP_CMD_SET_FORMAT_MODE_MULTI_STREAM (0x00000001U) +#define NV0073_CTRL_DP_CMD_FAST_LINK_TRAINING 5:5 +#define NV0073_CTRL_DP_CMD_FAST_LINK_TRAINING_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_FAST_LINK_TRAINING_YES (0x00000001U) +#define NV0073_CTRL_DP_CMD_NO_LINK_TRAINING 6:6 +#define NV0073_CTRL_DP_CMD_NO_LINK_TRAINING_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_NO_LINK_TRAINING_YES (0x00000001U) +#define NV0073_CTRL_DP_CMD_SET_ENHANCED_FRAMING 7:7 +#define NV0073_CTRL_DP_CMD_SET_ENHANCED_FRAMING_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_SET_ENHANCED_FRAMING_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_USE_DOWNSPREAD_SETTING 8:8 +#define NV0073_CTRL_DP_CMD_USE_DOWNSPREAD_SETTING_DEFAULT (0x00000000U) +#define NV0073_CTRL_DP_CMD_USE_DOWNSPREAD_SETTING_FORCE (0x00000001U) +#define NV0073_CTRL_DP_CMD_SKIP_HW_PROGRAMMING 9:9 +#define NV0073_CTRL_DP_CMD_SKIP_HW_PROGRAMMING_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_SKIP_HW_PROGRAMMING_YES (0x00000001U) +#define NV0073_CTRL_DP_CMD_POST_LT_ADJ_REQ_GRANTED 10:10 +#define NV0073_CTRL_DP_CMD_POST_LT_ADJ_REQ_GRANTED_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_POST_LT_ADJ_REQ_GRANTED_YES (0x00000001U) +#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING 12:11 +#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING_DONOT_TOGGLE_TRANSMISSION (0x00000001U) +#define NV0073_CTRL_DP_CMD_FAKE_LINK_TRAINING_TOGGLE_TRANSMISSION_ON (0x00000002U) +#define NV0073_CTRL_DP_CMD_TRAIN_PHY_REPEATER 13:13 +#define NV0073_CTRL_DP_CMD_TRAIN_PHY_REPEATER_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_TRAIN_PHY_REPEATER_YES (0x00000001U) +#define NV0073_CTRL_DP_CMD_FALLBACK_CONFIG 14:14 +#define NV0073_CTRL_DP_CMD_FALLBACK_CONFIG_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_FALLBACK_CONFIG_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_ENABLE_FEC 15:15 +#define NV0073_CTRL_DP_CMD_ENABLE_FEC_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_ENABLE_FEC_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_BANDWIDTH_TEST 29:29 +#define NV0073_CTRL_DP_CMD_BANDWIDTH_TEST_NO (0x00000000U) +#define NV0073_CTRL_DP_CMD_BANDWIDTH_TEST_YES (0x00000001U) +#define NV0073_CTRL_DP_CMD_LINK_CONFIG_CHECK_DISABLE 30:30 +#define NV0073_CTRL_DP_CMD_LINK_CONFIG_CHECK_DISABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_LINK_CONFIG_CHECK_DISABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_DISABLE_LINK_CONFIG 31:31 +#define NV0073_CTRL_DP_CMD_DISABLE_LINK_CONFIG_FALSE (0x00000000U) +#define NV0073_CTRL_DP_CMD_DISABLE_LINK_CONFIG_TRUE (0x00000001U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE 3:0 +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_MONITOR_ENABLE_BEGIN (0x00000000U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_MONITOR_ENABLE_CHALLENGE (0x00000001U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_MONITOR_ENABLE_CHECK (0x00000002U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_DRIVER_ENABLE_BEGIN (0x00000003U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_DRIVER_ENABLE_CHALLENGE (0x00000004U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_DRIVER_ENABLE_CHECK (0x00000005U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_RESET_MONITOR (0x00000006U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_INIT_PUBLIC_INFO (0x00000007U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_GET_PUBLIC_INFO (0x00000008U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_CMD_STAGE_STATUS_CHECK (0x00000009U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_OK (0x00000000U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_PENDING (0x80000001U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_READ_ERROR (0x80000002U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_WRITE_ERROR (0x80000003U) +#define NV0073_CTRL_DP_CMD_ENABLE_VRR_STATUS_DEVICE_ERROR (0x80000004U) + +#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT 4:0 +#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_0 (0x00000000U) +#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_1 (0x00000001U) +#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_2 (0x00000002U) +#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_4 (0x00000004U) +#define NV0073_CTRL_DP_DATA_SET_LANE_COUNT_8 (0x00000008U) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW 15:8 +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_1_62GBPS (0x00000006U) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_2_16GBPS (0x00000008U) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_2_43GBPS (0x00000009U) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_2_70GBPS (0x0000000AU) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_3_24GBPS (0x0000000CU) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_4_32GBPS (0x00000010U) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_5_40GBPS (0x00000014U) +#define NV0073_CTRL_DP_DATA_SET_LINK_BW_8_10GBPS (0x0000001EU) +#define NV0073_CTRL_DP_DATA_SET_ENHANCED_FRAMING 18:18 +#define NV0073_CTRL_DP_DATA_SET_ENHANCED_FRAMING_NO (0x00000000U) +#define NV0073_CTRL_DP_DATA_SET_ENHANCED_FRAMING_YES (0x00000001U) +#define NV0073_CTRL_DP_DATA_TARGET 22:19 +#define NV0073_CTRL_DP_DATA_TARGET_SINK (0x00000000U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_0 (0x00000001U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_1 (0x00000002U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_2 (0x00000003U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_3 (0x00000004U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_4 (0x00000005U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_5 (0x00000006U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_6 (0x00000007U) +#define NV0073_CTRL_DP_DATA_TARGET_PHY_REPEATER_7 (0x00000008U) + +#define NV0073_CTRL_CMD_DP_SET_LANE_DATA (0x731346U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_SET_LANE_DATA_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_MAX_LANES 8U + +typedef struct NV0073_CTRL_DP_LANE_DATA_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 numLanes; + NvU32 data[NV0073_CTRL_MAX_LANES]; +} NV0073_CTRL_DP_LANE_DATA_PARAMS; + +#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS 1:0 +#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_NONE (0x00000000U) +#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_LEVEL1 (0x00000001U) +#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_LEVEL2 (0x00000002U) +#define NV0073_CTRL_DP_LANE_DATA_PREEMPHASIS_LEVEL3 (0x00000003U) +#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT 3:2 +#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL0 (0x00000000U) +#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL1 (0x00000001U) +#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL2 (0x00000002U) +#define NV0073_CTRL_DP_LANE_DATA_DRIVECURRENT_LEVEL3 (0x00000003U) + +#define NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID (0x73135bU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 preferredDisplayId; + + NvBool force; + NvBool useBFM; + + NvU32 displayIdAssigned; + NvU32 allDisplayMask; +} NV0073_CTRL_CMD_DP_TOPOLOGY_ALLOCATE_DISPLAYID_PARAMS; + +#define NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID (0x73135cU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; +} NV0073_CTRL_CMD_DP_TOPOLOGY_FREE_DISPLAYID_PARAMS; + +#define NV0073_CTRL_CMD_DP_CONFIG_STREAM (0x731362U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS { + NvU32 subDeviceInstance; + NvU32 head; + NvU32 sorIndex; + NvU32 dpLink; + + NvBool bEnableOverride; + NvBool bMST; + NvU32 singleHeadMultistreamMode; + NvU32 hBlankSym; + NvU32 vBlankSym; + NvU32 colorFormat; + NvBool bEnableTwoHeadOneOr; + + struct { + NvU32 slotStart; + NvU32 slotEnd; + NvU32 PBN; + NvU32 Timeslice; + NvBool sendACT; // deprecated -Use NV0073_CTRL_CMD_DP_SEND_ACT + NvU32 singleHeadMSTPipeline; + NvBool bEnableAudioOverRightPanel; + } MST; + + struct { + NvBool bEnhancedFraming; + NvU32 tuSize; + NvU32 waterMark; + NvU32 actualPclkHz; // deprecated -Use MvidWarParams + NvU32 linkClkFreqHz; // deprecated -Use MvidWarParams + NvBool bEnableAudioOverRightPanel; + struct { + NvU32 activeCnt; + NvU32 activeFrac; + NvU32 activePolarity; + NvBool mvidWarEnabled; + struct { + NvU32 actualPclkHz; + NvU32 linkClkFreqHz; + } MvidWarParams; + } Legacy; + } SST; +} NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS; + +#define NV0073_CTRL_CMD_DFP_SET_AUDIO_ENABLE (0x731150U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_SET_AUDIO_ENABLE_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DFP_SET_AUDIO_ENABLE_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvBool enable; +} NV0073_CTRL_DFP_SET_AUDIO_ENABLE_PARAMS; + +#define NV0073_CTRL_CMD_DP_SET_AUDIO_MUTESTREAM (0x731359U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 mute; +} NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS; + +#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER (0x20800a58) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS { + NvU32 addressSpace; + NV_DECLARE_ALIGNED(NvU64 physicalAddr, 8); + NV_DECLARE_ALIGNED(NvU64 limit, 8); + NvU32 cacheSnoop; + NvU32 hclass; + NvU32 channelInstance; + NvBool valid; +} NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS; + +#define ADDR_SYSMEM (1) // System memory (PCI) + +#define ADDR_FBMEM 2 // Frame buffer memory space + +typedef struct +{ + NvV32 channelInstance; // One of the n channel instances of a given channel type. + // All PIO channels have two instances (one per head). + NvHandle hObjectNotify; // ctx dma handle for an area (of type NvNotification defined in sdk/nvidia/inc/nvtypes.h) where RM can write errors. + NvP64 pControl NV_ALIGN_BYTES(8); // pControl gives virt addr of control region for PIO channel +} NV50VAIO_CHANNELPIO_ALLOCATION_PARAMETERS; + +typedef struct +{ + NvV32 channelInstance; // One of the n channel instances of a given channel type. + // Note that core channel has only one instance + // while all others have two (one per head). + NvHandle hObjectBuffer; // ctx dma handle for DMA push buffer + NvHandle hObjectNotify; // ctx dma handle for an area (of type NvNotification defined in sdk/nvidia/inc/nvtypes.h) where RM can write errors/notifications + NvU32 offset; // Initial offset for put/get, usually zero. + NvP64 pControl NV_ALIGN_BYTES(8); // pControl gives virt addr of UDISP GET/PUT regs + + NvU32 flags; +#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB 1:1 +#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB_YES 0x00000000 +#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB_NO 0x00000001 + +} NV50VAIO_CHANNELDMA_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/engine.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/engine.h new file mode 100644 index 000000000000..b26dfc8f8087 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/engine.h @@ -0,0 +1,260 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_ENGINE_H__ +#define __NVRM_ENGINE_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define MC_ENGINE_IDX_NULL 0 // This must be 0 +#define MC_ENGINE_IDX_TMR 1 +#define MC_ENGINE_IDX_DISP 2 +#define MC_ENGINE_IDX_FB 3 +#define MC_ENGINE_IDX_FIFO 4 +#define MC_ENGINE_IDX_VIDEO 5 +#define MC_ENGINE_IDX_MD 6 +#define MC_ENGINE_IDX_BUS 7 +#define MC_ENGINE_IDX_PMGR 8 +#define MC_ENGINE_IDX_VP2 9 +#define MC_ENGINE_IDX_CIPHER 10 +#define MC_ENGINE_IDX_BIF 11 +#define MC_ENGINE_IDX_PPP 12 +#define MC_ENGINE_IDX_PRIVRING 13 +#define MC_ENGINE_IDX_PMU 14 +#define MC_ENGINE_IDX_CE0 15 +#define MC_ENGINE_IDX_CE1 16 +#define MC_ENGINE_IDX_CE2 17 +#define MC_ENGINE_IDX_CE3 18 +#define MC_ENGINE_IDX_CE4 19 +#define MC_ENGINE_IDX_CE5 20 +#define MC_ENGINE_IDX_CE6 21 +#define MC_ENGINE_IDX_CE7 22 +#define MC_ENGINE_IDX_CE8 23 +#define MC_ENGINE_IDX_CE9 24 +#define MC_ENGINE_IDX_CE_MAX MC_ENGINE_IDX_CE9 +#define MC_ENGINE_IDX_VIC 35 +#define MC_ENGINE_IDX_ISOHUB 36 +#define MC_ENGINE_IDX_VGPU 37 +#define MC_ENGINE_IDX_MSENC 38 +#define MC_ENGINE_IDX_MSENC1 39 +#define MC_ENGINE_IDX_MSENC2 40 +#define MC_ENGINE_IDX_C2C 41 +#define MC_ENGINE_IDX_LTC 42 +#define MC_ENGINE_IDX_FBHUB 43 +#define MC_ENGINE_IDX_HDACODEC 44 +#define MC_ENGINE_IDX_GMMU 45 +#define MC_ENGINE_IDX_SEC2 46 +#define MC_ENGINE_IDX_FSP 47 +#define MC_ENGINE_IDX_NVLINK 48 +#define MC_ENGINE_IDX_GSP 49 +#define MC_ENGINE_IDX_NVJPG 50 +#define MC_ENGINE_IDX_NVJPEG MC_ENGINE_IDX_NVJPG +#define MC_ENGINE_IDX_NVJPEG0 MC_ENGINE_IDX_NVJPEG +#define MC_ENGINE_IDX_NVJPEG1 51 +#define MC_ENGINE_IDX_NVJPEG2 52 +#define MC_ENGINE_IDX_NVJPEG3 53 +#define MC_ENGINE_IDX_NVJPEG4 54 +#define MC_ENGINE_IDX_NVJPEG5 55 +#define MC_ENGINE_IDX_NVJPEG6 56 +#define MC_ENGINE_IDX_NVJPEG7 57 +#define MC_ENGINE_IDX_REPLAYABLE_FAULT 58 +#define MC_ENGINE_IDX_ACCESS_CNTR 59 +#define MC_ENGINE_IDX_NON_REPLAYABLE_FAULT 60 +#define MC_ENGINE_IDX_REPLAYABLE_FAULT_ERROR 61 +#define MC_ENGINE_IDX_NON_REPLAYABLE_FAULT_ERROR 62 +#define MC_ENGINE_IDX_INFO_FAULT 63 +#define MC_ENGINE_IDX_BSP 64 +#define MC_ENGINE_IDX_NVDEC MC_ENGINE_IDX_BSP +#define MC_ENGINE_IDX_NVDEC0 MC_ENGINE_IDX_NVDEC +#define MC_ENGINE_IDX_NVDEC1 65 +#define MC_ENGINE_IDX_NVDEC2 66 +#define MC_ENGINE_IDX_NVDEC3 67 +#define MC_ENGINE_IDX_NVDEC4 68 +#define MC_ENGINE_IDX_NVDEC5 69 +#define MC_ENGINE_IDX_NVDEC6 70 +#define MC_ENGINE_IDX_NVDEC7 71 +#define MC_ENGINE_IDX_CPU_DOORBELL 72 +#define MC_ENGINE_IDX_PRIV_DOORBELL 73 +#define MC_ENGINE_IDX_MMU_ECC_ERROR 74 +#define MC_ENGINE_IDX_BLG 75 +#define MC_ENGINE_IDX_PERFMON 76 +#define MC_ENGINE_IDX_BUF_RESET 77 +#define MC_ENGINE_IDX_XBAR 78 +#define MC_ENGINE_IDX_ZPW 79 +#define MC_ENGINE_IDX_OFA0 80 +#define MC_ENGINE_IDX_TEGRA 81 +#define MC_ENGINE_IDX_GR 82 +#define MC_ENGINE_IDX_GR0 MC_ENGINE_IDX_GR +#define MC_ENGINE_IDX_GR1 83 +#define MC_ENGINE_IDX_GR2 84 +#define MC_ENGINE_IDX_GR3 85 +#define MC_ENGINE_IDX_GR4 86 +#define MC_ENGINE_IDX_GR5 87 +#define MC_ENGINE_IDX_GR6 88 +#define MC_ENGINE_IDX_GR7 89 +#define MC_ENGINE_IDX_ESCHED 90 +#define MC_ENGINE_IDX_ESCHED__SIZE 64 +#define MC_ENGINE_IDX_GR_FECS_LOG 154 +#define MC_ENGINE_IDX_GR0_FECS_LOG MC_ENGINE_IDX_GR_FECS_LOG +#define MC_ENGINE_IDX_GR1_FECS_LOG 155 +#define MC_ENGINE_IDX_GR2_FECS_LOG 156 +#define MC_ENGINE_IDX_GR3_FECS_LOG 157 +#define MC_ENGINE_IDX_GR4_FECS_LOG 158 +#define MC_ENGINE_IDX_GR5_FECS_LOG 159 +#define MC_ENGINE_IDX_GR6_FECS_LOG 160 +#define MC_ENGINE_IDX_GR7_FECS_LOG 161 +#define MC_ENGINE_IDX_TMR_SWRL 162 +#define MC_ENGINE_IDX_DISP_GSP 163 +#define MC_ENGINE_IDX_REPLAYABLE_FAULT_CPU 164 +#define MC_ENGINE_IDX_NON_REPLAYABLE_FAULT_CPU 165 +#define MC_ENGINE_IDX_PXUC 166 +#define MC_ENGINE_IDX_MAX 167 // This must be kept as the max bit if +#define MC_ENGINE_IDX_INVALID 0xFFFFFFFF +#define MC_ENGINE_IDX_GRn(x) (MC_ENGINE_IDX_GR0 + (x)) +#define MC_ENGINE_IDX_GRn_FECS_LOG(x) (MC_ENGINE_IDX_GR0_FECS_LOG + (x)) +#define MC_ENGINE_IDX_CE(x) (MC_ENGINE_IDX_CE0 + (x)) +#define MC_ENGINE_IDX_MSENCn(x) (MC_ENGINE_IDX_MSENC + (x)) +#define MC_ENGINE_IDX_NVDECn(x) (MC_ENGINE_IDX_NVDEC + (x)) +#define MC_ENGINE_IDX_NVJPEGn(x) (MC_ENGINE_IDX_NVJPEG + (x)) +#define MC_ENGINE_IDX_ESCHEDn(x) (MC_ENGINE_IDX_ESCHED + (x)) + +typedef enum +{ + RM_ENGINE_TYPE_NULL = (0x00000000), + RM_ENGINE_TYPE_GR0 = (0x00000001), + RM_ENGINE_TYPE_GR1 = (0x00000002), + RM_ENGINE_TYPE_GR2 = (0x00000003), + RM_ENGINE_TYPE_GR3 = (0x00000004), + RM_ENGINE_TYPE_GR4 = (0x00000005), + RM_ENGINE_TYPE_GR5 = (0x00000006), + RM_ENGINE_TYPE_GR6 = (0x00000007), + RM_ENGINE_TYPE_GR7 = (0x00000008), + RM_ENGINE_TYPE_COPY0 = (0x00000009), + RM_ENGINE_TYPE_COPY1 = (0x0000000a), + RM_ENGINE_TYPE_COPY2 = (0x0000000b), + RM_ENGINE_TYPE_COPY3 = (0x0000000c), + RM_ENGINE_TYPE_COPY4 = (0x0000000d), + RM_ENGINE_TYPE_COPY5 = (0x0000000e), + RM_ENGINE_TYPE_COPY6 = (0x0000000f), + RM_ENGINE_TYPE_COPY7 = (0x00000010), + RM_ENGINE_TYPE_COPY8 = (0x00000011), + RM_ENGINE_TYPE_COPY9 = (0x00000012), + RM_ENGINE_TYPE_NVDEC0 = (0x0000001d), + RM_ENGINE_TYPE_NVDEC1 = (0x0000001e), + RM_ENGINE_TYPE_NVDEC2 = (0x0000001f), + RM_ENGINE_TYPE_NVDEC3 = (0x00000020), + RM_ENGINE_TYPE_NVDEC4 = (0x00000021), + RM_ENGINE_TYPE_NVDEC5 = (0x00000022), + RM_ENGINE_TYPE_NVDEC6 = (0x00000023), + RM_ENGINE_TYPE_NVDEC7 = (0x00000024), + RM_ENGINE_TYPE_NVENC0 = (0x00000025), + RM_ENGINE_TYPE_NVENC1 = (0x00000026), + RM_ENGINE_TYPE_NVENC2 = (0x00000027), + RM_ENGINE_TYPE_VP = (0x00000028), + RM_ENGINE_TYPE_ME = (0x00000029), + RM_ENGINE_TYPE_PPP = (0x0000002a), + RM_ENGINE_TYPE_MPEG = (0x0000002b), + RM_ENGINE_TYPE_SW = (0x0000002c), + RM_ENGINE_TYPE_TSEC = (0x0000002d), + RM_ENGINE_TYPE_VIC = (0x0000002e), + RM_ENGINE_TYPE_MP = (0x0000002f), + RM_ENGINE_TYPE_SEC2 = (0x00000030), + RM_ENGINE_TYPE_HOST = (0x00000031), + RM_ENGINE_TYPE_DPU = (0x00000032), + RM_ENGINE_TYPE_PMU = (0x00000033), + RM_ENGINE_TYPE_FBFLCN = (0x00000034), + RM_ENGINE_TYPE_NVJPEG0 = (0x00000035), + RM_ENGINE_TYPE_NVJPEG1 = (0x00000036), + RM_ENGINE_TYPE_NVJPEG2 = (0x00000037), + RM_ENGINE_TYPE_NVJPEG3 = (0x00000038), + RM_ENGINE_TYPE_NVJPEG4 = (0x00000039), + RM_ENGINE_TYPE_NVJPEG5 = (0x0000003a), + RM_ENGINE_TYPE_NVJPEG6 = (0x0000003b), + RM_ENGINE_TYPE_NVJPEG7 = (0x0000003c), + RM_ENGINE_TYPE_OFA = (0x0000003d), + RM_ENGINE_TYPE_LAST = (0x0000003e), +} RM_ENGINE_TYPE; + +#define NV2080_ENGINE_TYPE_NULL (0x00000000) +#define NV2080_ENGINE_TYPE_GRAPHICS (0x00000001) +#define NV2080_ENGINE_TYPE_GR0 NV2080_ENGINE_TYPE_GRAPHICS +#define NV2080_ENGINE_TYPE_GR1 (0x00000002) +#define NV2080_ENGINE_TYPE_GR2 (0x00000003) +#define NV2080_ENGINE_TYPE_GR3 (0x00000004) +#define NV2080_ENGINE_TYPE_GR4 (0x00000005) +#define NV2080_ENGINE_TYPE_GR5 (0x00000006) +#define NV2080_ENGINE_TYPE_GR6 (0x00000007) +#define NV2080_ENGINE_TYPE_GR7 (0x00000008) +#define NV2080_ENGINE_TYPE_COPY0 (0x00000009) +#define NV2080_ENGINE_TYPE_COPY1 (0x0000000a) +#define NV2080_ENGINE_TYPE_COPY2 (0x0000000b) +#define NV2080_ENGINE_TYPE_COPY3 (0x0000000c) +#define NV2080_ENGINE_TYPE_COPY4 (0x0000000d) +#define NV2080_ENGINE_TYPE_COPY5 (0x0000000e) +#define NV2080_ENGINE_TYPE_COPY6 (0x0000000f) +#define NV2080_ENGINE_TYPE_COPY7 (0x00000010) +#define NV2080_ENGINE_TYPE_COPY8 (0x00000011) +#define NV2080_ENGINE_TYPE_COPY9 (0x00000012) +#define NV2080_ENGINE_TYPE_BSP (0x00000013) +#define NV2080_ENGINE_TYPE_NVDEC0 NV2080_ENGINE_TYPE_BSP +#define NV2080_ENGINE_TYPE_NVDEC1 (0x00000014) +#define NV2080_ENGINE_TYPE_NVDEC2 (0x00000015) +#define NV2080_ENGINE_TYPE_NVDEC3 (0x00000016) +#define NV2080_ENGINE_TYPE_NVDEC4 (0x00000017) +#define NV2080_ENGINE_TYPE_NVDEC5 (0x00000018) +#define NV2080_ENGINE_TYPE_NVDEC6 (0x00000019) +#define NV2080_ENGINE_TYPE_NVDEC7 (0x0000001a) +#define NV2080_ENGINE_TYPE_MSENC (0x0000001b) +#define NV2080_ENGINE_TYPE_NVENC0 NV2080_ENGINE_TYPE_MSENC /* Mutually exclusive alias */ +#define NV2080_ENGINE_TYPE_NVENC1 (0x0000001c) +#define NV2080_ENGINE_TYPE_NVENC2 (0x0000001d) +#define NV2080_ENGINE_TYPE_VP (0x0000001e) +#define NV2080_ENGINE_TYPE_ME (0x0000001f) +#define NV2080_ENGINE_TYPE_PPP (0x00000020) +#define NV2080_ENGINE_TYPE_MPEG (0x00000021) +#define NV2080_ENGINE_TYPE_SW (0x00000022) +#define NV2080_ENGINE_TYPE_CIPHER (0x00000023) +#define NV2080_ENGINE_TYPE_TSEC NV2080_ENGINE_TYPE_CIPHER +#define NV2080_ENGINE_TYPE_VIC (0x00000024) +#define NV2080_ENGINE_TYPE_MP (0x00000025) +#define NV2080_ENGINE_TYPE_SEC2 (0x00000026) +#define NV2080_ENGINE_TYPE_HOST (0x00000027) +#define NV2080_ENGINE_TYPE_DPU (0x00000028) +#define NV2080_ENGINE_TYPE_PMU (0x00000029) +#define NV2080_ENGINE_TYPE_FBFLCN (0x0000002a) +#define NV2080_ENGINE_TYPE_NVJPG (0x0000002b) +#define NV2080_ENGINE_TYPE_NVJPEG0 NV2080_ENGINE_TYPE_NVJPG +#define NV2080_ENGINE_TYPE_NVJPEG1 (0x0000002c) +#define NV2080_ENGINE_TYPE_NVJPEG2 (0x0000002d) +#define NV2080_ENGINE_TYPE_NVJPEG3 (0x0000002e) +#define NV2080_ENGINE_TYPE_NVJPEG4 (0x0000002f) +#define NV2080_ENGINE_TYPE_NVJPEG5 (0x00000030) +#define NV2080_ENGINE_TYPE_NVJPEG6 (0x00000031) +#define NV2080_ENGINE_TYPE_NVJPEG7 (0x00000032) +#define NV2080_ENGINE_TYPE_OFA (0x00000033) +#define NV2080_ENGINE_TYPE_LAST (0x0000003e) +#define NV2080_ENGINE_TYPE_ALLENGINES (0xffffffff) +#define NV2080_ENGINE_TYPE_COPY_SIZE 10 +#define NV2080_ENGINE_TYPE_NVENC_SIZE 3 +#define NV2080_ENGINE_TYPE_NVJPEG_SIZE 8 +#define NV2080_ENGINE_TYPE_NVDEC_SIZE 8 +#define NV2080_ENGINE_TYPE_GR_SIZE 8 +#define NV2080_ENGINE_TYPE_COPY(i) (NV2080_ENGINE_TYPE_COPY0+(i)) +#define NV2080_ENGINE_TYPE_IS_COPY(i) (((i) >= NV2080_ENGINE_TYPE_COPY0) && ((i) <= NV2080_ENGINE_TYPE_COPY9)) +#define NV2080_ENGINE_TYPE_COPY_IDX(i) ((i) - NV2080_ENGINE_TYPE_COPY0) +#define NV2080_ENGINE_TYPE_NVENC(i) (NV2080_ENGINE_TYPE_NVENC0+(i)) +#define NV2080_ENGINE_TYPE_IS_NVENC(i) (((i) >= NV2080_ENGINE_TYPE_NVENC0) && ((i) < NV2080_ENGINE_TYPE_NVENC(NV2080_ENGINE_TYPE_NVENC_SIZE))) +#define NV2080_ENGINE_TYPE_NVENC_IDX(i) ((i) - NV2080_ENGINE_TYPE_NVENC0) +#define NV2080_ENGINE_TYPE_NVDEC(i) (NV2080_ENGINE_TYPE_NVDEC0+(i)) +#define NV2080_ENGINE_TYPE_IS_NVDEC(i) (((i) >= NV2080_ENGINE_TYPE_NVDEC0) && ((i) < NV2080_ENGINE_TYPE_NVDEC(NV2080_ENGINE_TYPE_NVDEC_SIZE))) +#define NV2080_ENGINE_TYPE_NVDEC_IDX(i) ((i) - NV2080_ENGINE_TYPE_NVDEC0) +#define NV2080_ENGINE_TYPE_NVJPEG(i) (NV2080_ENGINE_TYPE_NVJPEG0+(i)) +#define NV2080_ENGINE_TYPE_IS_NVJPEG(i) (((i) >= NV2080_ENGINE_TYPE_NVJPEG0) && ((i) < NV2080_ENGINE_TYPE_NVJPEG(NV2080_ENGINE_TYPE_NVJPEG_SIZE))) +#define NV2080_ENGINE_TYPE_NVJPEG_IDX(i) ((i) - NV2080_ENGINE_TYPE_NVJPEG0) +#define NV2080_ENGINE_TYPE_GR(i) (NV2080_ENGINE_TYPE_GR0 + (i)) +#define NV2080_ENGINE_TYPE_IS_GR(i) (((i) >= NV2080_ENGINE_TYPE_GR0) && ((i) < NV2080_ENGINE_TYPE_GR(NV2080_ENGINE_TYPE_GR_SIZE))) +#define NV2080_ENGINE_TYPE_GR_IDX(i) ((i) - NV2080_ENGINE_TYPE_GR0) +#define NV2080_ENGINE_TYPE_IS_VALID(i) (((i) > (NV2080_ENGINE_TYPE_NULL)) && ((i) < (NV2080_ENGINE_TYPE_LAST))) +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/event.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/event.h new file mode 100644 index 000000000000..057f7220c225 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/event.h @@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_EVENT_H__ +#define __NVRM_EVENT_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV01_EVENT_KERNEL_CALLBACK_EX (0x0000007e) + +typedef struct NV0005_ALLOC_PARAMETERS { + NvHandle hParentClient; + NvHandle hSrcResource; + + NvV32 hClass; + NvV32 notifyIndex; + NV_DECLARE_ALIGNED(NvP64 data, 8); +} NV0005_ALLOC_PARAMETERS; + +#define NV01_EVENT_CLIENT_RM (0x04000000) + +#define NV2080_CTRL_CMD_EVENT_SET_NOTIFICATION (0x20800301) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_EVENT_INTERFACE_ID << 8) | NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS { + NvU32 event; + NvU32 action; + NvBool bNotifyState; + NvU32 info32; + NvU16 info16; +} NV2080_CTRL_EVENT_SET_NOTIFICATION_PARAMS; + +#define NV2080_CTRL_EVENT_SET_NOTIFICATION_ACTION_REPEAT (0x00000002) + +typedef struct rpc_post_event_v17_00 +{ + NvHandle hClient; + NvHandle hEvent; + NvU32 notifyIndex; + NvU32 data; + NvU16 info16; + NvU32 status; + NvU32 eventDataSize; + NvBool bNotifyList; + NvU8 eventData[]; +} rpc_post_event_v17_00; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/fbsr.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/fbsr.h new file mode 100644 index 000000000000..28786ef013a2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/fbsr.h @@ -0,0 +1,106 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_FBSR_H__ +#define __NVRM_FBSR_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV01_MEMORY_LIST_FBMEM (0x00000082) + +#define NV01_MEMORY_LIST_SYSTEM (0x00000081) + +#define NVOS02_FLAGS_PHYSICALITY 7:4 +#define NVOS02_FLAGS_PHYSICALITY_CONTIGUOUS (0x00000000) +#define NVOS02_FLAGS_PHYSICALITY_NONCONTIGUOUS (0x00000001) +#define NVOS02_FLAGS_LOCATION 11:8 +#define NVOS02_FLAGS_LOCATION_PCI (0x00000000) +#define NVOS02_FLAGS_LOCATION_AGP (0x00000001) +#define NVOS02_FLAGS_LOCATION_VIDMEM (0x00000002) +#define NVOS02_FLAGS_COHERENCY 15:12 +#define NVOS02_FLAGS_COHERENCY_UNCACHED (0x00000000) +#define NVOS02_FLAGS_COHERENCY_CACHED (0x00000001) +#define NVOS02_FLAGS_COHERENCY_WRITE_COMBINE (0x00000002) +#define NVOS02_FLAGS_COHERENCY_WRITE_THROUGH (0x00000003) +#define NVOS02_FLAGS_COHERENCY_WRITE_PROTECT (0x00000004) +#define NVOS02_FLAGS_COHERENCY_WRITE_BACK (0x00000005) +#define NVOS02_FLAGS_ALLOC 17:16 +#define NVOS02_FLAGS_ALLOC_NONE (0x00000001) +#define NVOS02_FLAGS_GPU_CACHEABLE 18:18 +#define NVOS02_FLAGS_GPU_CACHEABLE_NO (0x00000000) +#define NVOS02_FLAGS_GPU_CACHEABLE_YES (0x00000001) +#define NVOS02_FLAGS_KERNEL_MAPPING 19:19 +#define NVOS02_FLAGS_KERNEL_MAPPING_NO_MAP (0x00000000) +#define NVOS02_FLAGS_KERNEL_MAPPING_MAP (0x00000001) +#define NVOS02_FLAGS_ALLOC_NISO_DISPLAY 20:20 +#define NVOS02_FLAGS_ALLOC_NISO_DISPLAY_NO (0x00000000) +#define NVOS02_FLAGS_ALLOC_NISO_DISPLAY_YES (0x00000001) +#define NVOS02_FLAGS_ALLOC_USER_READ_ONLY 21:21 +#define NVOS02_FLAGS_ALLOC_USER_READ_ONLY_NO (0x00000000) +#define NVOS02_FLAGS_ALLOC_USER_READ_ONLY_YES (0x00000001) +#define NVOS02_FLAGS_ALLOC_DEVICE_READ_ONLY 22:22 +#define NVOS02_FLAGS_ALLOC_DEVICE_READ_ONLY_NO (0x00000000) +#define NVOS02_FLAGS_ALLOC_DEVICE_READ_ONLY_YES (0x00000001) +#define NVOS02_FLAGS_PEER_MAP_OVERRIDE 23:23 +#define NVOS02_FLAGS_PEER_MAP_OVERRIDE_DEFAULT (0x00000000) +#define NVOS02_FLAGS_PEER_MAP_OVERRIDE_REQUIRED (0x00000001) +#define NVOS02_FLAGS_ALLOC_TYPE_SYNCPOINT 24:24 +#define NVOS02_FLAGS_ALLOC_TYPE_SYNCPOINT_APERTURE (0x00000001) +#define NVOS02_FLAGS_MEMORY_PROTECTION 26:25 +#define NVOS02_FLAGS_MEMORY_PROTECTION_DEFAULT (0x00000000) +#define NVOS02_FLAGS_MEMORY_PROTECTION_PROTECTED (0x00000001) +#define NVOS02_FLAGS_MEMORY_PROTECTION_UNPROTECTED (0x00000002) +#define NVOS02_FLAGS_MAPPING 31:30 +#define NVOS02_FLAGS_MAPPING_DEFAULT (0x00000000) +#define NVOS02_FLAGS_MAPPING_NO_MAP (0x00000001) +#define NVOS02_FLAGS_MAPPING_NEVER_MAP (0x00000002) + +struct pte_desc +{ + NvU32 idr:2; + NvU32 reserved1:14; + NvU32 length:16; + union { + NvU64 pte; // PTE when IDR==0; PDE when IDR > 0 + NvU64 pde; // PTE when IDR==0; PDE when IDR > 0 + } pte_pde[] NV_ALIGN_BYTES(8); // PTE when IDR==0; PDE when IDR > 0 +}; + +typedef struct rpc_alloc_memory_v13_01 +{ + NvHandle hClient; + NvHandle hDevice; + NvHandle hMemory; + NvU32 hClass; + NvU32 flags; + NvU32 pteAdjust; + NvU32 format; + NvU64 length NV_ALIGN_BYTES(8); + NvU32 pageCount; + struct pte_desc pteDesc; +} rpc_alloc_memory_v13_01; + +#define FBSR_TYPE_DMA 4 // Copy using DMA. Fastest. + +#define NV2080_CTRL_CMD_INTERNAL_FBSR_INIT (0x20800ac2) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS { + NvU32 fbsrType; + NvU32 numRegions; + NvHandle hClient; + NvHandle hSysMem; + NV_DECLARE_ALIGNED(NvU64 gspFbAllocsSysOffset, 8); + NvBool bEnteringGcoffState; +} NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS; + +#define NV2080_CTRL_CMD_INTERNAL_FBSR_SEND_REGION_INFO (0x20800ac3) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_FBSR_SEND_REGION_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_FBSR_SEND_REGION_INFO_PARAMS { + NvU32 fbsrType; + NvHandle hClient; + NvHandle hVidMem; + NV_DECLARE_ALIGNED(NvU64 vidOffset, 8); + NV_DECLARE_ALIGNED(NvU64 sysOffset, 8); + NV_DECLARE_ALIGNED(NvU64 size, 8); +} NV2080_CTRL_INTERNAL_FBSR_SEND_REGION_INFO_PARAMS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/fifo.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/fifo.h new file mode 100644 index 000000000000..325fdd8b6090 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/fifo.h @@ -0,0 +1,350 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_FIFO_H__ +#define __NVRM_FIFO_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_MAX_ENTRIES 32 + +#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_DATA_TYPES 16 + +#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_PBDMA 2 + +#define NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_NAME_LEN 16 + +typedef struct NV2080_CTRL_FIFO_DEVICE_ENTRY { + NvU32 engineData[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_DATA_TYPES]; + NvU32 pbdmaIds[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_PBDMA]; + NvU32 pbdmaFaultIds[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_PBDMA]; + NvU32 numPbdmas; + char engineName[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_ENGINE_MAX_NAME_LEN]; +} NV2080_CTRL_FIFO_DEVICE_ENTRY; + +#define NV2080_CTRL_CMD_FIFO_GET_DEVICE_INFO_TABLE (0x20801112) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_FIFO_INTERFACE_ID << 8) | NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS { + NvU32 baseIndex; + NvU32 numEntries; + NvBool bMore; + // C form: NV2080_CTRL_FIFO_DEVICE_ENTRY entries[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_MAX_ENTRIES]; + NV2080_CTRL_FIFO_DEVICE_ENTRY entries[NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_MAX_ENTRIES]; +} NV2080_CTRL_FIFO_GET_DEVICE_INFO_TABLE_PARAMS; + +typedef enum +{ + /* ************************************************************************* + * Bug 3820969 + * THINK BEFORE CHANGING ENUM ORDER HERE. + * VGPU-guest uses this same ordering. Because this enum is not versioned, + * changing the order here WILL BREAK old-guest-on-newer-host compatibility. + * ************************************************************************/ + + // *ENG_XYZ, e.g.: ENG_GR, ENG_CE etc., + ENGINE_INFO_TYPE_ENG_DESC = 0, + + // HW engine ID + ENGINE_INFO_TYPE_FIFO_TAG, + + // RM_ENGINE_TYPE_* + ENGINE_INFO_TYPE_RM_ENGINE_TYPE, + + // + // runlist id (meaning varies by GPU) + // Valid only for Esched-driven engines + // + ENGINE_INFO_TYPE_RUNLIST, + + // NV_PFIFO_INTR_MMU_FAULT_ENG_ID_* + ENGINE_INFO_TYPE_MMU_FAULT_ID, + + // ROBUST_CHANNEL_* + ENGINE_INFO_TYPE_RC_MASK, + + // Reset Bit Position. On Ampere, only valid if not _INVALID + ENGINE_INFO_TYPE_RESET, + + // Interrupt Bit Position + ENGINE_INFO_TYPE_INTR, + + // log2(MC_ENGINE_*) + ENGINE_INFO_TYPE_MC, + + // The DEV_TYPE_ENUM for this engine + ENGINE_INFO_TYPE_DEV_TYPE_ENUM, + + // The particular instance of this engine type + ENGINE_INFO_TYPE_INSTANCE_ID, + + // + // The base address for this engine's NV_RUNLIST. Valid only on Ampere+ + // Valid only for Esched-driven engines + // + ENGINE_INFO_TYPE_RUNLIST_PRI_BASE, + + // + // If this entry is a host-driven engine. + // Update _isEngineInfoTypeValidForOnlyHostDriven when adding any new entry. + // + ENGINE_INFO_TYPE_IS_HOST_DRIVEN_ENGINE, + + // + // The index into the per-engine NV_RUNLIST registers. Valid only on Ampere+ + // Valid only for Esched-driven engines + // + ENGINE_INFO_TYPE_RUNLIST_ENGINE_ID, + + // + // The base address for this engine's NV_CHRAM registers. Valid only on + // Ampere+ + // + // Valid only for Esched-driven engines + // + ENGINE_INFO_TYPE_CHRAM_PRI_BASE, + + // This entry added to copy data at RMCTRL_EXPORT() call for Kernel RM + ENGINE_INFO_TYPE_KERNEL_RM_MAX, + // Used for iterating the engine info table by the index passed. + ENGINE_INFO_TYPE_INVALID = ENGINE_INFO_TYPE_KERNEL_RM_MAX, + + // Size of FIFO_ENGINE_LIST.engineData + ENGINE_INFO_TYPE_ENGINE_DATA_ARRAY_SIZE = ENGINE_INFO_TYPE_INVALID, + + // Input-only parameter for kfifoEngineInfoXlate. + ENGINE_INFO_TYPE_PBDMA_ID + + /* ************************************************************************* + * Bug 3820969 + * THINK BEFORE CHANGING ENUM ORDER HERE. + * VGPU-guest uses this same ordering. Because this enum is not versioned, + * changing the order here WILL BREAK old-guest-on-newer-host compatibility. + * ************************************************************************/ +} ENGINE_INFO_TYPE; + +#define NV2080_CTRL_CMD_CE_GET_FAULT_METHOD_BUFFER_SIZE (0x20802a08) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_CE_INTERFACE_ID << 8) | NV2080_CTRL_CE_GET_FAULT_METHOD_BUFFER_SIZE_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_CE_GET_FAULT_METHOD_BUFFER_SIZE_PARAMS { + NvU32 size; +} NV2080_CTRL_CE_GET_FAULT_METHOD_BUFFER_SIZE_PARAMS; + +#define NV2080_CTRL_CMD_INTERNAL_MAX_CONSTRUCTED_FALCONS 0x40 + +typedef struct NV2080_CTRL_INTERNAL_CONSTRUCTED_FALCON_INFO { + NvU32 engDesc; + NvU32 ctxAttr; + NvU32 ctxBufferSize; + NvU32 addrSpaceList; + NvU32 registerBase; +} NV2080_CTRL_INTERNAL_CONSTRUCTED_FALCON_INFO; + +#define NV2080_CTRL_CMD_INTERNAL_GET_CONSTRUCTED_FALCON_INFO (0x20800a42) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_GET_CONSTRUCTED_FALCON_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_GET_CONSTRUCTED_FALCON_INFO_PARAMS { + NvU32 numConstructedFalcons; + NV2080_CTRL_INTERNAL_CONSTRUCTED_FALCON_INFO constructedFalconsTable[NV2080_CTRL_CMD_INTERNAL_MAX_CONSTRUCTED_FALCONS]; +} NV2080_CTRL_INTERNAL_GET_CONSTRUCTED_FALCON_INFO_PARAMS; + +#define NV_MAX_SUBDEVICES 8 + +typedef struct NV_MEMORY_DESC_PARAMS { + NV_DECLARE_ALIGNED(NvU64 base, 8); + NV_DECLARE_ALIGNED(NvU64 size, 8); + NvU32 addressSpace; + NvU32 cacheAttrib; +} NV_MEMORY_DESC_PARAMS; + +#define CC_CHAN_ALLOC_IV_SIZE_DWORD 3U + +#define CC_CHAN_ALLOC_NONCE_SIZE_DWORD 8U + +typedef struct NV_CHANNEL_ALLOC_PARAMS { + + NvHandle hObjectError; // error context DMA + NvHandle hObjectBuffer; // no longer used + NV_DECLARE_ALIGNED(NvU64 gpFifoOffset, 8); // offset to beginning of GP FIFO + NvU32 gpFifoEntries; // number of GP FIFO entries + + NvU32 flags; + + + NvHandle hContextShare; // context share handle + NvHandle hVASpace; // VASpace for the channel + + // handle to UserD memory object for channel, ignored if hUserdMemory[0]=0 + NvHandle hUserdMemory[NV_MAX_SUBDEVICES]; + + // offset to beginning of UserD within hUserdMemory[x] + NV_DECLARE_ALIGNED(NvU64 userdOffset[NV_MAX_SUBDEVICES], 8); + + // engine type(NV2080_ENGINE_TYPE_*) with which this channel is associated + NvU32 engineType; + // Channel identifier that is unique for the duration of a RM session + NvU32 cid; + // One-hot encoded bitmask to match SET_SUBDEVICE_MASK methods + NvU32 subDeviceId; + NvHandle hObjectEccError; // ECC error context DMA + + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS instanceMem, 8); + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS userdMem, 8); + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS ramfcMem, 8); + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS mthdbufMem, 8); + + NvHandle hPhysChannelGroup; // reserved + NvU32 internalFlags; // reserved + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS errorNotifierMem, 8); // reserved + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS eccErrorNotifierMem, 8); // reserved + NvU32 ProcessID; // reserved + NvU32 SubProcessID; // reserved + + // IV used for CPU-side encryption / GPU-side decryption. + NvU32 encryptIv[CC_CHAN_ALLOC_IV_SIZE_DWORD]; // reserved + // IV used for CPU-side decryption / GPU-side encryption. + NvU32 decryptIv[CC_CHAN_ALLOC_IV_SIZE_DWORD]; // reserved + // Nonce used CPU-side signing / GPU-side signature verification. + NvU32 hmacNonce[CC_CHAN_ALLOC_NONCE_SIZE_DWORD]; // reserved +} NV_CHANNEL_ALLOC_PARAMS; + +typedef NV_CHANNEL_ALLOC_PARAMS NV_CHANNELGPFIFO_ALLOCATION_PARAMETERS; + +#define NVOS04_FLAGS_CHANNEL_TYPE 1:0 +#define NVOS04_FLAGS_CHANNEL_TYPE_PHYSICAL 0x00000000 +#define NVOS04_FLAGS_CHANNEL_TYPE_VIRTUAL 0x00000001 // OBSOLETE +#define NVOS04_FLAGS_CHANNEL_TYPE_PHYSICAL_FOR_VIRTUAL 0x00000002 // OBSOLETE +#define NVOS04_FLAGS_VPR 2:2 +#define NVOS04_FLAGS_VPR_FALSE 0x00000000 +#define NVOS04_FLAGS_VPR_TRUE 0x00000001 +#define NVOS04_FLAGS_CC_SECURE 2:2 +#define NVOS04_FLAGS_CC_SECURE_FALSE 0x00000000 +#define NVOS04_FLAGS_CC_SECURE_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_SKIP_MAP_REFCOUNTING 3:3 +#define NVOS04_FLAGS_CHANNEL_SKIP_MAP_REFCOUNTING_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_SKIP_MAP_REFCOUNTING_TRUE 0x00000001 +#define NVOS04_FLAGS_GROUP_CHANNEL_RUNQUEUE 4:4 +#define NVOS04_FLAGS_GROUP_CHANNEL_RUNQUEUE_DEFAULT 0x00000000 +#define NVOS04_FLAGS_GROUP_CHANNEL_RUNQUEUE_ONE 0x00000001 +#define NVOS04_FLAGS_PRIVILEGED_CHANNEL 5:5 +#define NVOS04_FLAGS_PRIVILEGED_CHANNEL_FALSE 0x00000000 +#define NVOS04_FLAGS_PRIVILEGED_CHANNEL_TRUE 0x00000001 +#define NVOS04_FLAGS_DELAY_CHANNEL_SCHEDULING 6:6 +#define NVOS04_FLAGS_DELAY_CHANNEL_SCHEDULING_FALSE 0x00000000 +#define NVOS04_FLAGS_DELAY_CHANNEL_SCHEDULING_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_DENY_PHYSICAL_MODE_CE 7:7 +#define NVOS04_FLAGS_CHANNEL_DENY_PHYSICAL_MODE_CE_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_DENY_PHYSICAL_MODE_CE_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_VALUE 10:8 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_FIXED 11:11 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_FIXED_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_FIXED_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_VALUE 20:12 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_FIXED 21:21 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_FIXED_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_FIXED_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_DENY_AUTH_LEVEL_PRIV 22:22 +#define NVOS04_FLAGS_CHANNEL_DENY_AUTH_LEVEL_PRIV_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_DENY_AUTH_LEVEL_PRIV_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_SKIP_SCRUBBER 23:23 +#define NVOS04_FLAGS_CHANNEL_SKIP_SCRUBBER_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_SKIP_SCRUBBER_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_CLIENT_MAP_FIFO 24:24 +#define NVOS04_FLAGS_CHANNEL_CLIENT_MAP_FIFO_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_CLIENT_MAP_FIFO_TRUE 0x00000001 +#define NVOS04_FLAGS_SET_EVICT_LAST_CE_PREFETCH_CHANNEL 25:25 +#define NVOS04_FLAGS_SET_EVICT_LAST_CE_PREFETCH_CHANNEL_FALSE 0x00000000 +#define NVOS04_FLAGS_SET_EVICT_LAST_CE_PREFETCH_CHANNEL_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_VGPU_PLUGIN_CONTEXT 26:26 +#define NVOS04_FLAGS_CHANNEL_VGPU_PLUGIN_CONTEXT_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_VGPU_PLUGIN_CONTEXT_TRUE 0x00000001 +#define NVOS04_FLAGS_CHANNEL_PBDMA_ACQUIRE_TIMEOUT 27:27 +#define NVOS04_FLAGS_CHANNEL_PBDMA_ACQUIRE_TIMEOUT_FALSE 0x00000000 +#define NVOS04_FLAGS_CHANNEL_PBDMA_ACQUIRE_TIMEOUT_TRUE 0x00000001 +#define NVOS04_FLAGS_GROUP_CHANNEL_THREAD 29:28 +#define NVOS04_FLAGS_GROUP_CHANNEL_THREAD_DEFAULT 0x00000000 +#define NVOS04_FLAGS_GROUP_CHANNEL_THREAD_ONE 0x00000001 +#define NVOS04_FLAGS_GROUP_CHANNEL_THREAD_TWO 0x00000002 +#define NVOS04_FLAGS_MAP_CHANNEL 30:30 +#define NVOS04_FLAGS_MAP_CHANNEL_FALSE 0x00000000 +#define NVOS04_FLAGS_MAP_CHANNEL_TRUE 0x00000001 +#define NVOS04_FLAGS_SKIP_CTXBUFFER_ALLOC 31:31 +#define NVOS04_FLAGS_SKIP_CTXBUFFER_ALLOC_FALSE 0x00000000 +#define NVOS04_FLAGS_SKIP_CTXBUFFER_ALLOC_TRUE 0x00000001 + +typedef enum { + /*! + * Initial state as passed in NV_CHANNEL_ALLOC_PARAMS by + * kernel CPU-RM clients. + */ + ERROR_NOTIFIER_TYPE_UNKNOWN = 0, + /*! @brief Error notifier is explicitly not set. + * + * The corresponding hErrorContext or hEccErrorContext must be + * NV01_NULL_OBJECT. + */ + ERROR_NOTIFIER_TYPE_NONE, + /*! @brief Error notifier is a ContextDma */ + ERROR_NOTIFIER_TYPE_CTXDMA, + /*! @brief Error notifier is a NvNotification array in sysmem/vidmem */ + ERROR_NOTIFIER_TYPE_MEMORY +} ErrorNotifierType; + +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE 1:0 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_USER 0x0 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_ADMIN 0x1 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_KERNEL 0x2 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE 3:2 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_UNKNOWN ERROR_NOTIFIER_TYPE_UNKNOWN +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_NONE ERROR_NOTIFIER_TYPE_NONE +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_CTXDMA ERROR_NOTIFIER_TYPE_CTXDMA +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_MEMORY ERROR_NOTIFIER_TYPE_MEMORY +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE 5:4 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_UNKNOWN ERROR_NOTIFIER_TYPE_UNKNOWN +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_NONE ERROR_NOTIFIER_TYPE_NONE +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_CTXDMA ERROR_NOTIFIER_TYPE_CTXDMA +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_MEMORY ERROR_NOTIFIER_TYPE_MEMORY + +#define NVA06F_CTRL_CMD_BIND (0xa06f0104) /* finn: Evaluated from "(FINN_KEPLER_CHANNEL_GPFIFO_A_GPFIFO_INTERFACE_ID << 8) | NVA06F_CTRL_BIND_PARAMS_MESSAGE_ID" */ +typedef struct NVA06F_CTRL_BIND_PARAMS { + NvU32 engineType; +} NVA06F_CTRL_BIND_PARAMS; + +#define NVA06F_CTRL_CMD_GPFIFO_SCHEDULE (0xa06f0103) /* finn: Evaluated from "(FINN_KEPLER_CHANNEL_GPFIFO_A_GPFIFO_INTERFACE_ID << 8) | NVA06F_CTRL_GPFIFO_SCHEDULE_PARAMS_MESSAGE_ID" */ +typedef struct NVA06F_CTRL_GPFIFO_SCHEDULE_PARAMS { + NvBool bEnable; + NvBool bSkipSubmit; +} NVA06F_CTRL_GPFIFO_SCHEDULE_PARAMS; + +#define NV2080_CTRL_GPU_PROMOTE_CONTEXT_MAX_ENTRIES 16U + +typedef struct NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY { + NV_DECLARE_ALIGNED(NvU64 gpuPhysAddr, 8); + NV_DECLARE_ALIGNED(NvU64 gpuVirtAddr, 8); + NV_DECLARE_ALIGNED(NvU64 size, 8); + NvU32 physAttr; + NvU16 bufferId; + NvU8 bInitialize; + NvU8 bNonmapped; +} NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY; + +#define NV2080_CTRL_CMD_GPU_PROMOTE_CTX (0x2080012bU) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_GPU_INTERFACE_ID << 8) | NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS { + NvU32 engineType; + NvHandle hClient; + NvU32 ChID; + NvHandle hChanClient; + NvHandle hObject; + NvHandle hVirtMemory; + NV_DECLARE_ALIGNED(NvU64 virtAddress, 8); + NV_DECLARE_ALIGNED(NvU64 size, 8); + NvU32 entryCount; + // C form: NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY promoteEntry[NV2080_CTRL_GPU_PROMOTE_CONTEXT_MAX_ENTRIES]; + NV_DECLARE_ALIGNED(NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ENTRY promoteEntry[NV2080_CTRL_GPU_PROMOTE_CONTEXT_MAX_ENTRIES], 8); +} NV2080_CTRL_GPU_PROMOTE_CTX_PARAMS; + +typedef struct rpc_rc_triggered_v17_02 +{ + NvU32 nv2080EngineType; + NvU32 chid; + NvU32 exceptType; + NvU32 scope; + NvU16 partitionAttributionId; +} rpc_rc_triggered_v17_02; +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080fifo.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/gr.h index 6acb3f73242d..82c5ec727bb4 100644 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl0080/ctrl0080fifo.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/gr.h @@ -1,30 +1,31 @@ -#ifndef __src_common_sdk_nvidia_inc_ctrl_ctrl0080_ctrl0080fifo_h__ -#define __src_common_sdk_nvidia_inc_ctrl_ctrl0080_ctrl0080fifo_h__ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_GR_H__ +#define __NVRM_GR_H__ +#include <nvrm/nvtypes.h> /* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ -/* - * SPDX-FileCopyrightText: Copyright (c) 2006-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ +#define NV2080_CTRL_CMD_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO (0x20800a32) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO_PARAMS_MESSAGE_ID" */ + +#define NV2080_CTRL_INTERNAL_GR_MAX_ENGINES 8 + +#define NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT 0x19 + +typedef struct NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO { + NvU32 size; + NvU32 alignment; +} NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO; + +typedef struct NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO { + NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO engine[NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT]; +} NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO; + +typedef struct NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS { + NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO engineContextBuffersInfo[NV2080_CTRL_INTERNAL_GR_MAX_ENGINES]; +} NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS; #define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID 4:0 #define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS (0x00000000) @@ -54,4 +55,19 @@ #define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PRIV_ACCESS_MAP (0x00000018) #define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT (0x00000019) +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_MAIN 0U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PM 1U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PATCH 2U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_BUFFER_BUNDLE_CB 3U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PAGEPOOL 4U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_ATTRIBUTE_CB 5U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_RTV_CB_GLOBAL 6U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_GFXP_POOL 7U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_GFXP_CTRL_BLK 8U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_FECS_EVENT 9U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_PRIV_ACCESS_MAP 10U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_UNRESTRICTED_PRIV_ACCESS_MAP 11U +#define NV2080_CTRL_GPU_PROMOTE_CTX_BUFFER_ID_GLOBAL_PRIV_ACCESS_MAP 12U + +#include "fifo.h" #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/gsp.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/gsp.h new file mode 100644 index 000000000000..b6683a5bf870 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/gsp.h @@ -0,0 +1,825 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_GSP_H__ +#define __NVRM_GSP_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MAX_ENTRIES 16U + +#define NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MEM_TYPES 17U + +typedef NvBool NV2080_CTRL_CMD_FB_GET_FB_REGION_SURFACE_MEM_TYPE_FLAG[NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MEM_TYPES]; + +typedef struct NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO { + NV_DECLARE_ALIGNED(NvU64 base, 8); + NV_DECLARE_ALIGNED(NvU64 limit, 8); + NV_DECLARE_ALIGNED(NvU64 reserved, 8); + NvU32 performance; + NvBool supportCompressed; + NvBool supportISO; + NvBool bProtected; + NV2080_CTRL_CMD_FB_GET_FB_REGION_SURFACE_MEM_TYPE_FLAG blackList; +} NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO; + +typedef struct NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS { + NvU32 numFBRegions; + NV_DECLARE_ALIGNED(NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO fbRegion[NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MAX_ENTRIES], 8); +} NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS; + +#define NV0080_CTRL_GR_CAPS_TBL_SIZE 23 + +#define NV2080_GPU_MAX_GID_LENGTH (0x000000100ULL) + +typedef struct NV2080_CTRL_GPU_GET_GID_INFO_PARAMS { + NvU32 index; + NvU32 flags; + NvU32 length; + NvU8 data[NV2080_GPU_MAX_GID_LENGTH]; +} NV2080_CTRL_GPU_GET_GID_INFO_PARAMS; + +typedef struct NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS { + NvU32 gpcMask; +} NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS; + +typedef struct NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS { + NvU32 gpcId; + NvU32 tpcMask; +} NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS; + +typedef struct NV2080_CTRL_GPU_GET_FERMI_ZCULL_INFO_PARAMS { + NvU32 gpcId; + NvU32 zcullMask; +} NV2080_CTRL_GPU_GET_FERMI_ZCULL_INFO_PARAMS; + +typedef struct NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS { + NvU32 BoardID; + char chipSKU[4]; + char chipSKUMod[2]; + char project[5]; + char projectSKU[5]; + char CDP[6]; + char projectSKUMod[2]; + NvU32 businessCycle; +} NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS; + +typedef enum +{ + COMPUTE_BRANDING_TYPE_NONE, + COMPUTE_BRANDING_TYPE_TESLA, +} COMPUTE_BRANDING_TYPE; + +#define MAX_GPC_COUNT 32 + +typedef struct NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS { + NvU32 totalVFs; + NvU32 firstVfOffset; + NvU32 vfFeatureMask; + NV_DECLARE_ALIGNED(NvU64 FirstVFBar0Address, 8); + NV_DECLARE_ALIGNED(NvU64 FirstVFBar1Address, 8); + NV_DECLARE_ALIGNED(NvU64 FirstVFBar2Address, 8); + NV_DECLARE_ALIGNED(NvU64 bar0Size, 8); + NV_DECLARE_ALIGNED(NvU64 bar1Size, 8); + NV_DECLARE_ALIGNED(NvU64 bar2Size, 8); + NvBool b64bitBar0; + NvBool b64bitBar1; + NvBool b64bitBar2; + NvBool bSriovEnabled; + NvBool bSriovHeavyEnabled; + NvBool bEmulateVFBar0TlbInvalidationRegister; + NvBool bClientRmAllocatedCtxBuffer; +} NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS; + +#include "engine.h" + +#define NVGPU_ENGINE_CAPS_MASK_BITS 32 + +#define NVGPU_ENGINE_CAPS_MASK_ARRAY_MAX ((RM_ENGINE_TYPE_LAST-1)/NVGPU_ENGINE_CAPS_MASK_BITS + 1) + +typedef struct GspSMInfo_t +{ + NvU32 version; + NvU32 regBankCount; + NvU32 regBankRegCount; + NvU32 maxWarpsPerSM; + NvU32 maxThreadsPerWarp; + NvU32 geomGsObufEntries; + NvU32 geomXbufEntries; + NvU32 maxSPPerSM; + NvU32 rtCoreCount; +} GspSMInfo; + +typedef enum NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS { + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_MAIN = 0, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_SPILL = 1, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_PAGEPOOL = 2, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_BETACB = 3, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_RTV = 4, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL = 5, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL_CONTROL = 6, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL_CONTROL_CPU = 7, + NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_END = 8, +} NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS; + +#define NV2080_GPU_MAX_NAME_STRING_LENGTH (0x0000040U) + +typedef struct VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS +{ + NvU32 numHeads; + NvU32 maxNumHeads; +} VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS; + +typedef struct VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS +{ + NvU32 headIndex; + NvU32 maxHResolution; + NvU32 maxVResolution; +} VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS; + +typedef struct GspStaticConfigInfo_t +{ + NvU8 grCapsBits[NV0080_CTRL_GR_CAPS_TBL_SIZE]; + NV2080_CTRL_GPU_GET_GID_INFO_PARAMS gidInfo; + NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS gpcInfo; + NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS tpcInfo[MAX_GPC_COUNT]; + NV2080_CTRL_GPU_GET_FERMI_ZCULL_INFO_PARAMS zcullInfo[MAX_GPC_COUNT]; + NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS SKUInfo; + NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS fbRegionInfoParams; + COMPUTE_BRANDING_TYPE computeBranding; + + NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS sriovCaps; + NvU32 sriovMaxGfid; + + NvU32 engineCaps[NVGPU_ENGINE_CAPS_MASK_ARRAY_MAX]; + + GspSMInfo SM_info; + + NvBool poisonFuseEnabled; + + NvU64 fb_length; + NvU32 fbio_mask; + NvU32 fb_bus_width; + NvU32 fb_ram_type; + NvU32 fbp_mask; + NvU32 l2_cache_size; + + NvU32 gfxpBufferSize[NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL]; + NvU32 gfxpBufferAlignment[NV2080_CTRL_CMD_GR_CTXSW_PREEMPTION_BIND_BUFFERS_CONTEXT_POOL]; + + NvU8 gpuNameString[NV2080_GPU_MAX_NAME_STRING_LENGTH]; + NvU8 gpuShortNameString[NV2080_GPU_MAX_NAME_STRING_LENGTH]; + NvU16 gpuNameString_Unicode[NV2080_GPU_MAX_NAME_STRING_LENGTH]; + NvBool bGpuInternalSku; + NvBool bIsQuadroGeneric; + NvBool bIsQuadroAd; + NvBool bIsNvidiaNvs; + NvBool bIsVgx; + NvBool bGeforceSmb; + NvBool bIsTitan; + NvBool bIsTesla; + NvBool bIsMobile; + NvBool bIsGc6Rtd3Allowed; + NvBool bIsGcOffRtd3Allowed; + NvBool bIsGcoffLegacyAllowed; + + NvU64 bar1PdeBase; + NvU64 bar2PdeBase; + + NvBool bVbiosValid; + NvU32 vbiosSubVendor; + NvU32 vbiosSubDevice; + + NvBool bPageRetirementSupported; + + NvBool bSplitVasBetweenServerClientRm; + + NvBool bClRootportNeedsNosnoopWAR; + + VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS displaylessMaxHeads; + VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS displaylessMaxResolution; + NvU64 displaylessMaxPixels; + + // Client handle for internal RMAPI control. + NvHandle hInternalClient; + + // Device handle for internal RMAPI control. + NvHandle hInternalDevice; + + // Subdevice handle for internal RMAPI control. + NvHandle hInternalSubdevice; + + NvBool bSelfHostedMode; + NvBool bAtsSupported; + + NvBool bIsGpuUefi; +} GspStaticConfigInfo; + +typedef struct rpc_unloading_guest_driver_v1F_07 +{ + NvBool bInPMTransition; + NvBool bGc6Entering; + NvU32 newLevel; +} rpc_unloading_guest_driver_v1F_07; + +typedef struct PACKED_REGISTRY_ENTRY +{ + NvU32 nameOffset; + NvU8 type; + NvU32 data; + NvU32 length; +} PACKED_REGISTRY_ENTRY; + +typedef struct PACKED_REGISTRY_TABLE +{ + NvU32 size; + NvU32 numEntries; + PACKED_REGISTRY_ENTRY entries[] __counted_by(numEntries); +} PACKED_REGISTRY_TABLE; + +typedef struct +{ + NvU16 deviceID; // deviceID + NvU16 vendorID; // vendorID + NvU16 subdeviceID; // subsystem deviceID + NvU16 subvendorID; // subsystem vendorID + NvU8 revisionID; // revision ID +} BUSINFO; + +#define NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS (16U) + +typedef struct DOD_METHOD_DATA +{ + NV_STATUS status; + NvU32 acpiIdListLen; + NvU32 acpiIdList[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; +} DOD_METHOD_DATA; + +typedef struct JT_METHOD_DATA +{ + NV_STATUS status; + NvU32 jtCaps; + NvU16 jtRevId; + NvBool bSBIOSCaps; +} JT_METHOD_DATA; + +typedef struct MUX_METHOD_DATA_ELEMENT +{ + NvU32 acpiId; + NvU32 mode; + NV_STATUS status; +} MUX_METHOD_DATA_ELEMENT; + +#define NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS (16U) + +typedef struct MUX_METHOD_DATA +{ + NvU32 tableLen; + MUX_METHOD_DATA_ELEMENT acpiIdMuxModeTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; + MUX_METHOD_DATA_ELEMENT acpiIdMuxPartTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; +} MUX_METHOD_DATA; + +typedef struct CAPS_METHOD_DATA +{ + NV_STATUS status; + NvU32 optimusCaps; +} CAPS_METHOD_DATA; + +typedef struct ACPI_METHOD_DATA +{ + NvBool bValid; + DOD_METHOD_DATA dodMethodData; + JT_METHOD_DATA jtMethodData; + MUX_METHOD_DATA muxMethodData; + CAPS_METHOD_DATA capsMethodData; +} ACPI_METHOD_DATA; + +typedef struct GSP_VF_INFO +{ + NvU32 totalVFs; + NvU32 firstVFOffset; + NvU64 FirstVFBar0Address; + NvU64 FirstVFBar1Address; + NvU64 FirstVFBar2Address; + NvBool b64bitBar0; + NvBool b64bitBar1; + NvBool b64bitBar2; +} GSP_VF_INFO; + +typedef struct GspSystemInfo +{ + NvU64 gpuPhysAddr; + NvU64 gpuPhysFbAddr; + NvU64 gpuPhysInstAddr; + NvU64 nvDomainBusDeviceFunc; + NvU64 simAccessBufPhysAddr; + NvU64 pcieAtomicsOpMask; + NvU64 consoleMemSize; + NvU64 maxUserVa; + NvU32 pciConfigMirrorBase; + NvU32 pciConfigMirrorSize; + NvU8 oorArch; + NvU64 clPdbProperties; + NvU32 Chipset; + NvBool bGpuBehindBridge; + NvBool bMnocAvailable; + NvBool bUpstreamL0sUnsupported; + NvBool bUpstreamL1Unsupported; + NvBool bUpstreamL1PorSupported; + NvBool bUpstreamL1PorMobileOnly; + NvU8 upstreamAddressValid; + BUSINFO FHBBusInfo; + BUSINFO chipsetIDInfo; + ACPI_METHOD_DATA acpiMethodData; + NvU32 hypervisorType; + NvBool bIsPassthru; + NvU64 sysTimerOffsetNs; + GSP_VF_INFO gspVFInfo; +} GspSystemInfo; + +typedef struct rpc_os_error_log_v17_00 +{ + NvU32 exceptType; + NvU32 runlistId; + NvU32 chid; + char errString[0x100]; +} rpc_os_error_log_v17_00; + +typedef struct rpc_run_cpu_sequencer_v17_00 +{ + NvU32 bufferSizeDWord; + NvU32 cmdIndex; + NvU32 regSaveArea[8]; + NvU32 commandBuffer[]; +} rpc_run_cpu_sequencer_v17_00; + +typedef enum GSP_SEQ_BUF_OPCODE +{ + GSP_SEQ_BUF_OPCODE_REG_WRITE = 0, + GSP_SEQ_BUF_OPCODE_REG_MODIFY, + GSP_SEQ_BUF_OPCODE_REG_POLL, + GSP_SEQ_BUF_OPCODE_DELAY_US, + GSP_SEQ_BUF_OPCODE_REG_STORE, + GSP_SEQ_BUF_OPCODE_CORE_RESET, + GSP_SEQ_BUF_OPCODE_CORE_START, + GSP_SEQ_BUF_OPCODE_CORE_WAIT_FOR_HALT, + GSP_SEQ_BUF_OPCODE_CORE_RESUME, +} GSP_SEQ_BUF_OPCODE; + +typedef struct +{ + NvU32 addr; + NvU32 val; +} GSP_SEQ_BUF_PAYLOAD_REG_WRITE; + +typedef struct +{ + NvU32 addr; + NvU32 mask; + NvU32 val; +} GSP_SEQ_BUF_PAYLOAD_REG_MODIFY; + +typedef struct +{ + NvU32 addr; + NvU32 mask; + NvU32 val; + NvU32 timeout; + NvU32 error; +} GSP_SEQ_BUF_PAYLOAD_REG_POLL; + +typedef struct +{ + NvU32 val; +} GSP_SEQ_BUF_PAYLOAD_DELAY_US; + +typedef struct +{ + NvU32 addr; + NvU32 index; +} GSP_SEQ_BUF_PAYLOAD_REG_STORE; + +typedef struct GSP_SEQUENCER_BUFFER_CMD +{ + GSP_SEQ_BUF_OPCODE opCode; + union + { + GSP_SEQ_BUF_PAYLOAD_REG_WRITE regWrite; + GSP_SEQ_BUF_PAYLOAD_REG_MODIFY regModify; + GSP_SEQ_BUF_PAYLOAD_REG_POLL regPoll; + GSP_SEQ_BUF_PAYLOAD_DELAY_US delayUs; + GSP_SEQ_BUF_PAYLOAD_REG_STORE regStore; + } payload; +} GSP_SEQUENCER_BUFFER_CMD; + +typedef struct +{ + // Magic + // BL to use for verification (i.e. Booter locked it in WPR2) + NvU64 magic; // = 0xdc3aae21371a60b3; + + // Revision number of Booter-BL-Sequencer handoff interface + // Bumped up when we change this interface so it is not backward compatible. + // Bumped up when we revoke GSP-RM ucode + NvU64 revision; // = 1; + + // ---- Members regarding data in SYSMEM ---------------------------- + // Consumed by Booter for DMA + + NvU64 sysmemAddrOfRadix3Elf; + NvU64 sizeOfRadix3Elf; + + NvU64 sysmemAddrOfBootloader; + NvU64 sizeOfBootloader; + + // Offsets inside bootloader image needed by Booter + NvU64 bootloaderCodeOffset; + NvU64 bootloaderDataOffset; + NvU64 bootloaderManifestOffset; + + union + { + // Used only at initial boot + struct + { + NvU64 sysmemAddrOfSignature; + NvU64 sizeOfSignature; + }; + + // + // Used at suspend/resume to read GspFwHeapFreeList + // Offset relative to GspFwWprMeta FBMEM PA (gspFwWprStart) + // + struct + { + NvU32 gspFwHeapFreeListWprOffset; + NvU32 unused0; + NvU64 unused1; + }; + }; + + // ---- Members describing FB layout -------------------------------- + NvU64 gspFwRsvdStart; + + NvU64 nonWprHeapOffset; + NvU64 nonWprHeapSize; + + NvU64 gspFwWprStart; + + // GSP-RM to use to setup heap. + NvU64 gspFwHeapOffset; + NvU64 gspFwHeapSize; + + // BL to use to find ELF for jump + NvU64 gspFwOffset; + // Size is sizeOfRadix3Elf above. + + NvU64 bootBinOffset; + // Size is sizeOfBootloader above. + + NvU64 frtsOffset; + NvU64 frtsSize; + + NvU64 gspFwWprEnd; + + // GSP-RM to use for fbRegionInfo? + NvU64 fbSize; + + // ---- Other members ----------------------------------------------- + + // GSP-RM to use for fbRegionInfo? + NvU64 vgaWorkspaceOffset; + NvU64 vgaWorkspaceSize; + + // Boot count. Used to determine whether to load the firmware image. + NvU64 bootCount; + + // This union is organized the way it is to start at an 8-byte boundary and achieve natural + // packing of the internal struct fields. + union + { + struct + { + // TODO: the partitionRpc* fields below do not really belong in this + // structure. The values are patched in by the partition bootstrapper + // when GSP-RM is booted in a partition, and this structure was a + // convenient place for the bootstrapper to access them. These should + // be moved to a different comm. mechanism between the bootstrapper + // and the GSP-RM tasks. + + // Shared partition RPC memory (physical address) + NvU64 partitionRpcAddr; + + // Offsets relative to partitionRpcAddr + NvU16 partitionRpcRequestOffset; + NvU16 partitionRpcReplyOffset; + + // Code section and dataSection offset and size. + NvU32 elfCodeOffset; + NvU32 elfDataOffset; + NvU32 elfCodeSize; + NvU32 elfDataSize; + + // Used during GSP-RM resume to check for revocation + NvU32 lsUcodeVersion; + }; + + struct + { + // Pad for the partitionRpc* fields, plus 4 bytes + NvU32 partitionRpcPadding[4]; + + // CrashCat (contiguous) buffer size/location - occupies same bytes as the + // elf(Code|Data)(Offset|Size) fields above. + // TODO: move to GSP_FMC_INIT_PARAMS + NvU64 sysmemAddrOfCrashReportQueue; + NvU32 sizeOfCrashReportQueue; + + // Pad for the lsUcodeVersion field + NvU32 lsUcodeVersionPadding[1]; + }; + }; + + // Number of VF partitions allocating sub-heaps from the WPR heap + // Used during boot to ensure the heap is adequately sized + NvU8 gspFwHeapVfPartitionCount; + + // Pad structure to exactly 256 bytes. Can replace padding with additional + // fields without incrementing revision. Padding initialized to 0. + NvU8 padding[7]; + + // BL to use for verification (i.e. Booter says OK to boot) + NvU64 verified; // 0x0 -> unverified, 0xa0a0a0a0a0a0a0a0 -> verified +} GspFwWprMeta; + +#define GSP_FW_WPR_META_MAGIC 0xdc3aae21371a60b3ULL + +#define GSP_FW_WPR_META_REVISION 1 + +typedef struct +{ + NvU32 version; // queue version + NvU32 size; // bytes, page aligned + NvU32 msgSize; // entry size, bytes, must be power-of-2, 16 is minimum + NvU32 msgCount; // number of entries in queue + NvU32 writePtr; // message id of next slot + NvU32 flags; // if set it means "i want to swap RX" + NvU32 rxHdrOff; // Offset of msgqRxHeader from start of backing store. + NvU32 entryOff; // Offset of entries from start of backing store. +} msgqTxHeader; + +typedef struct +{ + NvU32 readPtr; // message id of last message read +} msgqRxHeader; + +typedef struct { + RmPhysAddr sharedMemPhysAddr; + NvU32 pageTableEntryCount; + NvLength cmdQueueOffset; + NvLength statQueueOffset; + NvLength locklessCmdQueueOffset; + NvLength locklessStatQueueOffset; +} MESSAGE_QUEUE_INIT_ARGUMENTS; + +typedef struct { + NvU32 oldLevel; + NvU32 flags; + NvBool bInPMTransition; +} GSP_SR_INIT_ARGUMENTS; + +typedef struct +{ + MESSAGE_QUEUE_INIT_ARGUMENTS messageQueueInitArguments; + GSP_SR_INIT_ARGUMENTS srInitArguments; + NvU32 gpuInstance; + + struct + { + NvU64 pa; + NvU64 size; + } profilerArgs; +} GSP_ARGUMENTS_CACHED; + +#define NV2080_CTRL_GPU_SET_POWER_STATE_GPU_LEVEL_0 (0x00000000U) + +#define NV2080_CTRL_GPU_SET_POWER_STATE_GPU_LEVEL_3 (0x00000003U) + +typedef NvU64 LibosAddress; + +typedef struct +{ + LibosAddress id8; // Id tag. + LibosAddress pa; // Physical address. + LibosAddress size; // Size of memory area. + NvU8 kind; // See LibosMemoryRegionKind above. + NvU8 loc; // See LibosMemoryRegionLoc above. +} LibosMemoryRegionInitArgument; + +typedef enum { + LIBOS_MEMORY_REGION_NONE, + LIBOS_MEMORY_REGION_CONTIGUOUS, + LIBOS_MEMORY_REGION_RADIX3 +} LibosMemoryRegionKind; + +typedef enum { + LIBOS_MEMORY_REGION_LOC_NONE, + LIBOS_MEMORY_REGION_LOC_SYSMEM, + LIBOS_MEMORY_REGION_LOC_FB +} LibosMemoryRegionLoc; + +typedef struct +{ + // + // Magic + // Use for verification by Booter + // + NvU64 magic; // = GSP_FW_SR_META_MAGIC; + + // + // Revision number + // Bumped up when we change this interface so it is not backward compatible. + // Bumped up when we revoke GSP-RM ucode + // + NvU64 revision; // = GSP_FW_SR_META_MAGIC_REVISION; + + // + // ---- Members regarding data in SYSMEM ---------------------------- + // Consumed by Booter for DMA + // + NvU64 sysmemAddrOfSuspendResumeData; + NvU64 sizeOfSuspendResumeData; + + // ---- Members for crypto ops across S/R --------------------------- + + // + // HMAC over the entire GspFwSRMeta structure (including padding) + // with the hmac field itself zeroed. + // + NvU8 hmac[32]; + + // Hash over GspFwWprMeta structure + NvU8 wprMetaHash[32]; + + // Hash over GspFwHeapFreeList structure. All zeros signifies no free list. + NvU8 heapFreeListHash[32]; + + // Hash over data in WPR2 (skipping over free heap chunks; see Booter for details) + NvU8 dataHash[32]; + + // + // Pad structure to exactly 256 bytes (1 DMA chunk). + // Padding initialized to zero. + // + NvU32 padding[24]; + +} GspFwSRMeta; + +#define GSP_FW_SR_META_MAGIC 0x8a3bb9e6c6c39d93ULL + +#define GSP_FW_SR_META_REVISION 2 + +#define GSP_SEQUENCER_PAYLOAD_SIZE_DWORDS(opcode) \ + ((opcode == GSP_SEQ_BUF_OPCODE_REG_WRITE) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_WRITE) / sizeof(NvU32)) : \ + (opcode == GSP_SEQ_BUF_OPCODE_REG_MODIFY) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_MODIFY) / sizeof(NvU32)) : \ + (opcode == GSP_SEQ_BUF_OPCODE_REG_POLL) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_POLL) / sizeof(NvU32)) : \ + (opcode == GSP_SEQ_BUF_OPCODE_DELAY_US) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_DELAY_US) / sizeof(NvU32)) : \ + (opcode == GSP_SEQ_BUF_OPCODE_REG_STORE) ? (sizeof(GSP_SEQ_BUF_PAYLOAD_REG_STORE) / sizeof(NvU32)) : \ + /* GSP_SEQ_BUF_OPCODE_CORE_RESET */ \ + /* GSP_SEQ_BUF_OPCODE_CORE_START */ \ + /* GSP_SEQ_BUF_OPCODE_CORE_WAIT_FOR_HALT */ \ + /* GSP_SEQ_BUF_OPCODE_CORE_RESUME */ \ + 0) + +typedef struct { + // + // Version 1 + // Version 2 + // Version 3 = for Partition boot + // Version 4 = for eb riscv boot + // Version 5 = Support signing entire RISC-V image as "code" in code section for hopper and later. + // + NvU32 version; // structure version + NvU32 bootloaderOffset; + NvU32 bootloaderSize; + NvU32 bootloaderParamOffset; + NvU32 bootloaderParamSize; + NvU32 riscvElfOffset; + NvU32 riscvElfSize; + NvU32 appVersion; // Changelist number associated with the image + // + // Manifest contains information about Monitor and it is + // input to BR + // + NvU32 manifestOffset; + NvU32 manifestSize; + // + // Monitor Data offset within RISCV image and size + // + NvU32 monitorDataOffset; + NvU32 monitorDataSize; + // + // Monitor Code offset withtin RISCV image and size + // + NvU32 monitorCodeOffset; + NvU32 monitorCodeSize; + NvU32 bIsMonitorEnabled; + // + // Swbrom Code offset within RISCV image and size + // + NvU32 swbromCodeOffset; + NvU32 swbromCodeSize; + // + // Swbrom Data offset within RISCV image and size + // + NvU32 swbromDataOffset; + NvU32 swbromDataSize; + // + // Total size of FB carveout (image and reserved space). + // + NvU32 fbReservedSize; + // + // Indicates whether the entire RISC-V image is signed as "code" in code section. + // + NvU32 bSignedAsCode; +} RM_RISCV_UCODE_DESC; + +typedef struct NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_ENTRY { + NvU16 engineIdx; + NvU32 pmcIntrMask; + NvU32 vectorStall; + NvU32 vectorNonStall; +} NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_ENTRY; + +typedef struct NV2080_INTR_CATEGORY_SUBTREE_MAP { + NvU8 subtreeStart; + NvU8 subtreeEnd; +} NV2080_INTR_CATEGORY_SUBTREE_MAP; + +#define NV2080_CTRL_INTERNAL_INTR_MAX_TABLE_SIZE 128 + +typedef enum NV2080_INTR_CATEGORY { + NV2080_INTR_CATEGORY_DEFAULT = 0, + NV2080_INTR_CATEGORY_ESCHED_DRIVEN_ENGINE = 1, + NV2080_INTR_CATEGORY_ESCHED_DRIVEN_ENGINE_NOTIFICATION = 2, + NV2080_INTR_CATEGORY_RUNLIST = 3, + NV2080_INTR_CATEGORY_RUNLIST_NOTIFICATION = 4, + NV2080_INTR_CATEGORY_UVM_OWNED = 5, + NV2080_INTR_CATEGORY_UVM_SHARED = 6, + NV2080_INTR_CATEGORY_ENUM_COUNT = 7, +} NV2080_INTR_CATEGORY; + +#define NV2080_CTRL_CMD_INTERNAL_INTR_GET_KERNEL_TABLE (0x20800a5c) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS { + NvU32 tableLen; + NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_ENTRY table[NV2080_CTRL_INTERNAL_INTR_MAX_TABLE_SIZE]; + NV2080_INTR_CATEGORY_SUBTREE_MAP subtreeMap[NV2080_INTR_CATEGORY_ENUM_COUNT]; +} NV2080_CTRL_INTERNAL_INTR_GET_KERNEL_TABLE_PARAMS; + +#define GSP_FW_HEAP_PARAM_SIZE_PER_GB_FB (96 << 10) // All architectures + +#define GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE ((48 << 10) * 2048) // Support 2048 channels + +typedef union rpc_message_rpc_union_field_v03_00 +{ + NvU32 spare; + NvU32 cpuRmGfid; +} rpc_message_rpc_union_field_v03_00; + +typedef rpc_message_rpc_union_field_v03_00 rpc_message_rpc_union_field_v; + +typedef struct rpc_message_header_v03_00 +{ + NvU32 header_version; + NvU32 signature; + NvU32 length; + NvU32 function; + NvU32 rpc_result; + NvU32 rpc_result_private; + NvU32 sequence; + rpc_message_rpc_union_field_v u; + rpc_generic_union rpc_message_data[]; +} rpc_message_header_v03_00; + +typedef rpc_message_header_v03_00 rpc_message_header_v; + +typedef struct GSP_MSG_QUEUE_ELEMENT +{ + NvU8 authTagBuffer[16]; // Authentication tag buffer. + NvU8 aadBuffer[16]; // AAD buffer. + NvU32 checkSum; // Set to value needed to make checksum always zero. + NvU32 seqNum; // Sequence number maintained by the message queue. + NvU32 elemCount; // Number of message queue elements this message has. + NV_DECLARE_ALIGNED(rpc_message_header_v rpc, 8); +} GSP_MSG_QUEUE_ELEMENT; + +#define GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS2 (0 << 20) // No FB heap usage +#define GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3 (20 << 20) + +#define GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X (8 << 20) // Turing thru Ada + +#define GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS2_MIN_MB (64u) +#define GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB (84u) +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/msgfn.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/msgfn.h new file mode 100644 index 000000000000..642c13aec325 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/msgfn.h @@ -0,0 +1,53 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_MSGFN_H__ +#define __NVRM_MSGFN_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#ifndef E +# define E(RPC) NV_VGPU_MSG_EVENT_##RPC, +# define DEFINING_E_IN_RPC_GLOBAL_ENUMS_H +enum { +#endif + E(FIRST_EVENT = 0x1000) // 0x1000 + E(GSP_INIT_DONE) // 0x1001 + E(GSP_RUN_CPU_SEQUENCER) // 0x1002 + E(POST_EVENT) // 0x1003 + E(RC_TRIGGERED) // 0x1004 + E(MMU_FAULT_QUEUED) // 0x1005 + E(OS_ERROR_LOG) // 0x1006 + E(RG_LINE_INTR) // 0x1007 + E(GPUACCT_PERFMON_UTIL_SAMPLES) // 0x1008 + E(SIM_READ) // 0x1009 + E(SIM_WRITE) // 0x100a + E(SEMAPHORE_SCHEDULE_CALLBACK) // 0x100b + E(UCODE_LIBOS_PRINT) // 0x100c + E(VGPU_GSP_PLUGIN_TRIGGERED) // 0x100d + E(PERF_GPU_BOOST_SYNC_LIMITS_CALLBACK) // 0x100e + E(PERF_BRIDGELESS_INFO_UPDATE) // 0x100f + E(VGPU_CONFIG) // 0x1010 + E(DISPLAY_MODESET) // 0x1011 + E(EXTDEV_INTR_SERVICE) // 0x1012 + E(NVLINK_INBAND_RECEIVED_DATA_256) // 0x1013 + E(NVLINK_INBAND_RECEIVED_DATA_512) // 0x1014 + E(NVLINK_INBAND_RECEIVED_DATA_1024) // 0x1015 + E(NVLINK_INBAND_RECEIVED_DATA_2048) // 0x1016 + E(NVLINK_INBAND_RECEIVED_DATA_4096) // 0x1017 + E(TIMED_SEMAPHORE_RELEASE) // 0x1018 + E(NVLINK_IS_GPU_DEGRADED) // 0x1019 + E(PFM_REQ_HNDLR_STATE_SYNC_CALLBACK) // 0x101a + E(GSP_SEND_USER_SHARED_DATA) // 0x101b + E(NVLINK_FAULT_UP) // 0x101c + E(GSP_LOCKDOWN_NOTICE) // 0x101d + E(MIG_CI_CONFIG_UPDATE) // 0x101e + E(NUM_EVENTS) // END +#ifdef DEFINING_E_IN_RPC_GLOBAL_ENUMS_H +}; +# undef E +# undef DEFINING_E_IN_RPC_GLOBAL_ENUMS_H +#endif +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvdec.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvdec.h new file mode 100644 index 000000000000..3a04e702677f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvdec.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_NVDEC_H__ +#define __NVRM_NVDEC_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct +{ + NvU32 size; + NvU32 prohibitMultipleInstances; + NvU32 engineInstance; // Select NVDEC0 or NVDEC1 or NVDEC2 +} NV_BSP_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvenc.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvenc.h new file mode 100644 index 000000000000..203c1d5304d9 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvenc.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_NVENC_H__ +#define __NVRM_NVENC_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct +{ + NvU32 size; + NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of MSENC? + NvU32 engineInstance; // Select MSENC/NVENC0 or NVENC1 or NVENC2 +} NV_MSENC_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvjpg.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvjpg.h new file mode 100644 index 000000000000..71fc53889ec7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/nvjpg.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_NVJPG_H__ +#define __NVRM_NVJPG_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct +{ + NvU32 size; + NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of NVJPG? + NvU32 engineInstance; +} NV_NVJPG_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ofa.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ofa.h new file mode 100644 index 000000000000..49d81c7673d2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/ofa.h @@ -0,0 +1,16 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_OFA_H__ +#define __NVRM_OFA_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +typedef struct +{ + NvU32 size; + NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of OFA? +} NV_OFA_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_global_enums.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/rpcfn.h index 73c57f235f6a..2a037acc6b1e 100644 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/nvidia/kernel/inc/vgpu/rpc_global_enums.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/rpcfn.h @@ -1,5 +1,10 @@ -#ifndef __src_nvidia_kernel_inc_vgpu_rpc_global_enums_h__ -#define __src_nvidia_kernel_inc_vgpu_rpc_global_enums_h__ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_RPCFN_H__ +#define __NVRM_RPCFN_H__ +#include <nvrm/nvtypes.h> /* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ @@ -209,54 +214,12 @@ enum { X(RM, CTRL_SET_HS_CREDITS) // 198 X(RM, CTRL_PM_AREA_PC_SAMPLER) // 199 X(RM, INVALIDATE_TLB) // 200 + X(RM, RESERVED_201) // 201 + X(RM, ECC_NOTIFIER_WRITE_ACK) // 202 X(RM, NUM_FUNCTIONS) //END #ifdef DEFINING_X_IN_RPC_GLOBAL_ENUMS_H }; # undef X # undef DEFINING_X_IN_RPC_GLOBAL_ENUMS_H #endif - -#ifndef E -# define E(RPC) NV_VGPU_MSG_EVENT_##RPC, -# define DEFINING_E_IN_RPC_GLOBAL_ENUMS_H -enum { -#endif - E(FIRST_EVENT = 0x1000) // 0x1000 - E(GSP_INIT_DONE) // 0x1001 - E(GSP_RUN_CPU_SEQUENCER) // 0x1002 - E(POST_EVENT) // 0x1003 - E(RC_TRIGGERED) // 0x1004 - E(MMU_FAULT_QUEUED) // 0x1005 - E(OS_ERROR_LOG) // 0x1006 - E(RG_LINE_INTR) // 0x1007 - E(GPUACCT_PERFMON_UTIL_SAMPLES) // 0x1008 - E(SIM_READ) // 0x1009 - E(SIM_WRITE) // 0x100a - E(SEMAPHORE_SCHEDULE_CALLBACK) // 0x100b - E(UCODE_LIBOS_PRINT) // 0x100c - E(VGPU_GSP_PLUGIN_TRIGGERED) // 0x100d - E(PERF_GPU_BOOST_SYNC_LIMITS_CALLBACK) // 0x100e - E(PERF_BRIDGELESS_INFO_UPDATE) // 0x100f - E(VGPU_CONFIG) // 0x1010 - E(DISPLAY_MODESET) // 0x1011 - E(EXTDEV_INTR_SERVICE) // 0x1012 - E(NVLINK_INBAND_RECEIVED_DATA_256) // 0x1013 - E(NVLINK_INBAND_RECEIVED_DATA_512) // 0x1014 - E(NVLINK_INBAND_RECEIVED_DATA_1024) // 0x1015 - E(NVLINK_INBAND_RECEIVED_DATA_2048) // 0x1016 - E(NVLINK_INBAND_RECEIVED_DATA_4096) // 0x1017 - E(TIMED_SEMAPHORE_RELEASE) // 0x1018 - E(NVLINK_IS_GPU_DEGRADED) // 0x1019 - E(PFM_REQ_HNDLR_STATE_SYNC_CALLBACK) // 0x101a - E(GSP_SEND_USER_SHARED_DATA) // 0x101b - E(NVLINK_FAULT_UP) // 0x101c - E(GSP_LOCKDOWN_NOTICE) // 0x101d - E(MIG_CI_CONFIG_UPDATE) // 0x101e - E(NUM_EVENTS) // END -#ifdef DEFINING_E_IN_RPC_GLOBAL_ENUMS_H -}; -# undef E -# undef DEFINING_E_IN_RPC_GLOBAL_ENUMS_H -#endif - #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/vmm.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/vmm.h new file mode 100644 index 000000000000..f6ec04efd119 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/nvrm/vmm.h @@ -0,0 +1,132 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_VMM_H__ +#define __NVRM_VMM_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ + +#define FERMI_VASPACE_A (0x000090f1) + +typedef struct +{ + NvU32 index; + NvV32 flags; + NvU64 vaSize NV_ALIGN_BYTES(8); + NvU64 vaStartInternal NV_ALIGN_BYTES(8); + NvU64 vaLimitInternal NV_ALIGN_BYTES(8); + NvU32 bigPageSize; + NvU64 vaBase NV_ALIGN_BYTES(8); +} NV_VASPACE_ALLOCATION_PARAMETERS; + +#define NV_VASPACE_ALLOCATION_INDEX_GPU_NEW 0x00 //<! Create new VASpace, by default + +#define NV_VASPACE_ALLOCATION_FLAGS_IS_EXTERNALLY_OWNED BIT(3) + +#define SPLIT_VAS_SERVER_RM_MANAGED_VA_START 0x100000000ULL // 4GB +#define SPLIT_VAS_SERVER_RM_MANAGED_VA_SIZE 0x20000000ULL // 512MB + +#define GMMU_FMT_MAX_LEVELS 6U + +#define NV90F1_CTRL_CMD_VASPACE_COPY_SERVER_RESERVED_PDES (0x90f10106U) /* finn: Evaluated from "(FINN_FERMI_VASPACE_A_VASPACE_INTERFACE_ID << 8) | NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS_MESSAGE_ID" */ +typedef struct NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS { + /*! + * [in] GPU sub-device handle - this API only supports unicast. + * Pass 0 to use subDeviceId instead. + */ + NvHandle hSubDevice; + + /*! + * [in] GPU sub-device ID. Ignored if hSubDevice is non-zero. + */ + NvU32 subDeviceId; + + /*! + * [in] Page size (VA coverage) of the level to reserve. + * This need not be a leaf (page table) page size - it can be + * the coverage of an arbitrary level (including root page directory). + */ + NV_DECLARE_ALIGNED(NvU64 pageSize, 8); + + /*! + * [in] First GPU virtual address of the range to reserve. + * This must be aligned to pageSize. + */ + NV_DECLARE_ALIGNED(NvU64 virtAddrLo, 8); + + /*! + * [in] Last GPU virtual address of the range to reserve. + * This (+1) must be aligned to pageSize. + */ + NV_DECLARE_ALIGNED(NvU64 virtAddrHi, 8); + + /*! + * [in] Number of PDE levels to copy. + */ + NvU32 numLevelsToCopy; + + /*! + * [in] Per-level information. + */ + struct { + /*! + * Physical address of this page level instance. + */ + NV_DECLARE_ALIGNED(NvU64 physAddress, 8); + + /*! + * Size in bytes allocated for this level instance. + */ + NV_DECLARE_ALIGNED(NvU64 size, 8); + + /*! + * Aperture in which this page level instance resides. + */ + NvU32 aperture; + + /*! + * Page shift corresponding to the level + */ + NvU8 pageShift; + } levels[GMMU_FMT_MAX_LEVELS]; +} NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS; + +#define NV0080_CTRL_CMD_DMA_SET_PAGE_DIRECTORY (0x801813U) /* finn: Evaluated from "(FINN_NV01_DEVICE_0_DMA_INTERFACE_ +ID << 8) | NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_PARAMS_MESSAGE_ID" */ + +typedef struct NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_PARAMS { + NV_DECLARE_ALIGNED(NvU64 physAddress, 8); + NvU32 numEntries; + NvU32 flags; + NvHandle hVASpace; + NvU32 chId; + NvU32 subDeviceId; // ID+1, 0 for BC + NvU32 pasid; +} NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_PARAMS; + +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_APERTURE 1:0 +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_APERTURE_VIDMEM (0x00000000U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_APERTURE_SYSMEM_COH (0x00000001U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_APERTURE_SYSMEM_NONCOH (0x00000002U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_PRESERVE_PDES 2:2 +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_PRESERVE_PDES_FALSE (0x00000000U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_PRESERVE_PDES_TRUE (0x00000001U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_ALL_CHANNELS 3:3 +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_ALL_CHANNELS_FALSE (0x00000000U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_ALL_CHANNELS_TRUE (0x00000001U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_IGNORE_CHANNEL_BUSY 4:4 +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_IGNORE_CHANNEL_BUSY_FALSE (0x00000000U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_IGNORE_CHANNEL_BUSY_TRUE (0x00000001U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_EXTEND_VASPACE 5:5 +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_EXTEND_VASPACE_FALSE (0x00000000U) +#define NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_FLAGS_EXTEND_VASPACE_TRUE (0x00000001U) + +#define NV0080_CTRL_CMD_DMA_UNSET_PAGE_DIRECTORY (0x801814U) /* finn: Evaluated from "(FINN_NV01_DEVICE_0_DMA_INTERFACE_ID << 8) | NV0080_CTRL_DMA_UNSET_PAGE_DIRECTORY_PARAMS_MESSAGE_ID" */ + +typedef struct NV0080_CTRL_DMA_UNSET_PAGE_DIRECTORY_PARAMS { + NvHandle hVASpace; + NvU32 subDeviceId; // ID+1, 0 for BC +} NV0080_CTRL_DMA_UNSET_PAGE_DIRECTORY_PARAMS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ad102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ofa.c index d72b3aae9a2b..2156808cba4f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/nvdec/ad102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/ofa.c @@ -19,26 +19,26 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "priv.h" +#include <rm/engine.h> -#include <subdev/gsp.h> +#include "nvrm/ofa.h" -#include <nvif/class.h> +static int +r535_ofa_alloc(struct nvkm_gsp_object *chan, u32 handle, u32 class, int inst, + struct nvkm_gsp_object *ofa) +{ + NV_OFA_ALLOCATION_PARAMETERS *args; -static const struct nvkm_engine_func -ad102_nvdec = { - .sclass = { - { -1, -1, NVC9B0_VIDEO_DECODER }, - {} - } -}; + args = nvkm_gsp_rm_alloc_get(chan, handle, class, sizeof(*args), ofa); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); -int -ad102_nvdec_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, - struct nvkm_nvdec **pnvdec) -{ - if (nvkm_gsp_rm(device->gsp)) - return r535_nvdec_new(&ad102_nvdec, device, type, inst, pnvdec); + args->size = sizeof(*args); - return -ENODEV; + return nvkm_gsp_rm_alloc_wr(ofa, args); } + +const struct nvkm_rm_api_engine +r535_ofa = { + .alloc = r535_ofa_alloc, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/rm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/rm.c new file mode 100644 index 000000000000..a4190676e1ad --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/rm.c @@ -0,0 +1,52 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/rm.h> + +#include "nvrm/gsp.h" + +static const struct nvkm_rm_wpr +r535_wpr_libos2 = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS2, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS2_MIN_MB, +}; + +static const struct nvkm_rm_wpr +r535_wpr_libos3 = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB, +}; + +static const struct nvkm_rm_api +r535_api = { + .gsp = &r535_gsp, + .rpc = &r535_rpc, + .ctrl = &r535_ctrl, + .alloc = &r535_alloc, + .client = &r535_client, + .device = &r535_device, + .fbsr = &r535_fbsr, + .disp = &r535_disp, + .fifo = &r535_fifo, + .ce = &r535_ce, + .gr = &r535_gr, + .nvdec = &r535_nvdec, + .nvenc = &r535_nvenc, + .nvjpg = &r535_nvjpg, + .ofa = &r535_ofa, +}; + +const struct nvkm_rm_impl +r535_rm_tu102 = { + .wpr = &r535_wpr_libos2, + .api = &r535_api, +}; + +const struct nvkm_rm_impl +r535_rm_ga102 = { + .wpr = &r535_wpr_libos3, + .api = &r535_api, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/rpc.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/rpc.c new file mode 100644 index 000000000000..5acb98d137bd --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/rpc.c @@ -0,0 +1,691 @@ +/* + * Copyright 2023 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <rm/rpc.h> + +#include "nvrm/rpcfn.h" + +#define GSP_MSG_MIN_SIZE GSP_PAGE_SIZE +#define GSP_MSG_MAX_SIZE (GSP_MSG_MIN_SIZE * 16) + +/** + * DOC: GSP message queue element + * + * https://github.com/NVIDIA/open-gpu-kernel-modules/blob/535/src/nvidia/inc/kernel/gpu/gsp/message_queue_priv.h + * + * The GSP command queue and status queue are message queues for the + * communication between software and GSP. The software submits the GSP + * RPC via the GSP command queue, GSP writes the status of the submitted + * RPC in the status queue. + * + * A GSP message queue element consists of three parts: + * + * - message element header (struct r535_gsp_msg), which mostly maintains + * the metadata for queuing the element. + * + * - RPC message header (struct nvfw_gsp_rpc), which maintains the info + * of the RPC. E.g., the RPC function number. + * + * - The payload, where the RPC message stays. E.g. the params of a + * specific RPC function. Some RPC functions also have their headers + * in the payload. E.g. rm_alloc, rm_control. + * + * The memory layout of a GSP message element can be illustrated below:: + * + * +------------------------+ + * | Message Element Header | + * | (r535_gsp_msg) | + * | | + * | (r535_gsp_msg.data) | + * | | | + * |----------V-------------| + * | GSP RPC Header | + * | (nvfw_gsp_rpc) | + * | | + * | (nvfw_gsp_rpc.data) | + * | | | + * |----------V-------------| + * | Payload | + * | | + * | header(optional) | + * | params | + * +------------------------+ + * + * The max size of a message queue element is 16 pages (including the + * headers). When a GSP message to be sent is larger than 16 pages, the + * message should be split into multiple elements and sent accordingly. + * + * In the bunch of the split elements, the first element has the expected + * function number, while the rest of the elements are sent with the + * function number NV_VGPU_MSG_FUNCTION_CONTINUATION_RECORD. + * + * GSP consumes the elements from the cmdq and always writes the result + * back to the msgq. The result is also formed as split elements. + * + * Terminology: + * + * - gsp_msg(msg): GSP message element (element header + GSP RPC header + + * payload) + * - gsp_rpc(rpc): GSP RPC (RPC header + payload) + * - gsp_rpc_buf: buffer for (GSP RPC header + payload) + * - gsp_rpc_len: size of (GSP RPC header + payload) + * - params_size: size of params in the payload + * - payload_size: size of (header if exists + params) in the payload + */ + +struct r535_gsp_msg { + u8 auth_tag_buffer[16]; + u8 aad_buffer[16]; + u32 checksum; + u32 sequence; + u32 elem_count; + u32 pad; + u8 data[]; +}; + +struct nvfw_gsp_rpc { + u32 header_version; + u32 signature; + u32 length; + u32 function; + u32 rpc_result; + u32 rpc_result_private; + u32 sequence; + union { + u32 spare; + u32 cpuRmGfid; + }; + u8 data[]; +}; + +#define GSP_MSG_HDR_SIZE offsetof(struct r535_gsp_msg, data) + +#define to_gsp_hdr(p, header) \ + container_of((void *)p, typeof(*header), data) + +#define to_payload_hdr(p, header) \ + container_of((void *)p, typeof(*header), params) + +int +r535_rpc_status_to_errno(uint32_t rpc_status) +{ + switch (rpc_status) { + case 0x55: /* NV_ERR_NOT_READY */ + case 0x66: /* NV_ERR_TIMEOUT_RETRY */ + return -EBUSY; + case 0x51: /* NV_ERR_NO_MEMORY */ + return -ENOMEM; + default: + return -EINVAL; + } +} + +static int +r535_gsp_msgq_wait(struct nvkm_gsp *gsp, u32 gsp_rpc_len, int *ptime) +{ + u32 size, rptr = *gsp->msgq.rptr; + int used; + + size = DIV_ROUND_UP(GSP_MSG_HDR_SIZE + gsp_rpc_len, + GSP_PAGE_SIZE); + if (WARN_ON(!size || size >= gsp->msgq.cnt)) + return -EINVAL; + + do { + u32 wptr = *gsp->msgq.wptr; + + used = wptr + gsp->msgq.cnt - rptr; + if (used >= gsp->msgq.cnt) + used -= gsp->msgq.cnt; + if (used >= size) + break; + + usleep_range(1, 2); + } while (--(*ptime)); + + if (WARN_ON(!*ptime)) + return -ETIMEDOUT; + + return used; +} + +static struct r535_gsp_msg * +r535_gsp_msgq_get_entry(struct nvkm_gsp *gsp) +{ + u32 rptr = *gsp->msgq.rptr; + + /* Skip the first page, which is the message queue info */ + return (void *)((u8 *)gsp->shm.msgq.ptr + GSP_PAGE_SIZE + + rptr * GSP_PAGE_SIZE); +} + +/** + * DOC: Receive a GSP message queue element + * + * Receiving a GSP message queue element from the message queue consists of + * the following steps: + * + * - Peek the element from the queue: r535_gsp_msgq_peek(). + * Peek the first page of the element to determine the total size of the + * message before allocating the proper memory. + * + * - Allocate memory for the message. + * Once the total size of the message is determined from the GSP message + * queue element, the caller of r535_gsp_msgq_recv() allocates the + * required memory. + * + * - Receive the message: r535_gsp_msgq_recv(). + * Copy the message into the allocated memory. Advance the read pointer. + * If the message is a large GSP message, r535_gsp_msgq_recv() calls + * r535_gsp_msgq_recv_one_elem() repeatedly to receive continuation parts + * until the complete message is received. + * r535_gsp_msgq_recv() assembles the payloads of cotinuation parts into + * the return of the large GSP message. + * + * - Free the allocated memory: r535_gsp_msg_done(). + * The user is responsible for freeing the memory allocated for the GSP + * message pages after they have been processed. + */ +static void * +r535_gsp_msgq_peek(struct nvkm_gsp *gsp, u32 gsp_rpc_len, int *retries) +{ + struct r535_gsp_msg *mqe; + int ret; + + ret = r535_gsp_msgq_wait(gsp, gsp_rpc_len, retries); + if (ret < 0) + return ERR_PTR(ret); + + mqe = r535_gsp_msgq_get_entry(gsp); + + return mqe->data; +} + +struct r535_gsp_msg_info { + int *retries; + u32 gsp_rpc_len; + void *gsp_rpc_buf; + bool continuation; +}; + +static void +r535_gsp_msg_dump(struct nvkm_gsp *gsp, struct nvfw_gsp_rpc *msg, int lvl); + +static void * +r535_gsp_msgq_recv_one_elem(struct nvkm_gsp *gsp, + struct r535_gsp_msg_info *info) +{ + u8 *buf = info->gsp_rpc_buf; + u32 rptr = *gsp->msgq.rptr; + struct r535_gsp_msg *mqe; + u32 size, expected, len; + int ret; + + expected = info->gsp_rpc_len; + + ret = r535_gsp_msgq_wait(gsp, expected, info->retries); + if (ret < 0) + return ERR_PTR(ret); + + mqe = r535_gsp_msgq_get_entry(gsp); + + if (info->continuation) { + struct nvfw_gsp_rpc *rpc = (struct nvfw_gsp_rpc *)mqe->data; + + if (rpc->function != NV_VGPU_MSG_FUNCTION_CONTINUATION_RECORD) { + nvkm_error(&gsp->subdev, + "Not a continuation of a large RPC\n"); + r535_gsp_msg_dump(gsp, rpc, NV_DBG_ERROR); + return ERR_PTR(-EIO); + } + } + + size = ALIGN(expected + GSP_MSG_HDR_SIZE, GSP_PAGE_SIZE); + + len = ((gsp->msgq.cnt - rptr) * GSP_PAGE_SIZE) - sizeof(*mqe); + len = min_t(u32, expected, len); + + if (info->continuation) + memcpy(buf, mqe->data + sizeof(struct nvfw_gsp_rpc), + len - sizeof(struct nvfw_gsp_rpc)); + else + memcpy(buf, mqe->data, len); + + expected -= len; + + if (expected) { + mqe = (void *)((u8 *)gsp->shm.msgq.ptr + 0x1000 + 0 * 0x1000); + memcpy(buf + len, mqe, expected); + } + + rptr = (rptr + DIV_ROUND_UP(size, GSP_PAGE_SIZE)) % gsp->msgq.cnt; + + mb(); + (*gsp->msgq.rptr) = rptr; + return buf; +} + +static void * +r535_gsp_msgq_recv(struct nvkm_gsp *gsp, u32 gsp_rpc_len, int *retries) +{ + struct r535_gsp_msg *mqe; + const u32 max_rpc_size = GSP_MSG_MAX_SIZE - sizeof(*mqe); + struct nvfw_gsp_rpc *rpc; + struct r535_gsp_msg_info info = {0}; + u32 expected = gsp_rpc_len; + void *buf; + + mqe = r535_gsp_msgq_get_entry(gsp); + rpc = (struct nvfw_gsp_rpc *)mqe->data; + + if (WARN_ON(rpc->length > max_rpc_size)) + return NULL; + + buf = kvmalloc(max_t(u32, rpc->length, expected), GFP_KERNEL); + if (!buf) + return ERR_PTR(-ENOMEM); + + info.gsp_rpc_buf = buf; + info.retries = retries; + info.gsp_rpc_len = rpc->length; + + buf = r535_gsp_msgq_recv_one_elem(gsp, &info); + if (IS_ERR(buf)) { + kvfree(info.gsp_rpc_buf); + info.gsp_rpc_buf = NULL; + return buf; + } + + if (expected <= max_rpc_size) + return buf; + + info.gsp_rpc_buf += info.gsp_rpc_len; + expected -= info.gsp_rpc_len; + + while (expected) { + u32 size; + + rpc = r535_gsp_msgq_peek(gsp, sizeof(*rpc), info.retries); + if (IS_ERR_OR_NULL(rpc)) { + kfree(buf); + return rpc; + } + + info.gsp_rpc_len = rpc->length; + info.continuation = true; + + rpc = r535_gsp_msgq_recv_one_elem(gsp, &info); + if (IS_ERR_OR_NULL(rpc)) { + kfree(buf); + return rpc; + } + + size = info.gsp_rpc_len - sizeof(*rpc); + expected -= size; + info.gsp_rpc_buf += size; + } + + rpc = buf; + rpc->length = gsp_rpc_len; + return buf; +} + +static int +r535_gsp_cmdq_push(struct nvkm_gsp *gsp, void *rpc) +{ + struct r535_gsp_msg *msg = to_gsp_hdr(rpc, msg); + struct r535_gsp_msg *cqe; + u32 gsp_rpc_len = msg->checksum; + u64 *ptr = (void *)msg; + u64 *end; + u64 csum = 0; + int free, time = 1000000; + u32 wptr, size, step, len; + u32 off = 0; + + len = ALIGN(GSP_MSG_HDR_SIZE + gsp_rpc_len, GSP_PAGE_SIZE); + + end = (u64 *)((char *)ptr + len); + msg->pad = 0; + msg->checksum = 0; + msg->sequence = gsp->cmdq.seq++; + msg->elem_count = DIV_ROUND_UP(len, 0x1000); + + while (ptr < end) + csum ^= *ptr++; + + msg->checksum = upper_32_bits(csum) ^ lower_32_bits(csum); + + wptr = *gsp->cmdq.wptr; + do { + do { + free = *gsp->cmdq.rptr + gsp->cmdq.cnt - wptr - 1; + if (free >= gsp->cmdq.cnt) + free -= gsp->cmdq.cnt; + if (free >= 1) + break; + + usleep_range(1, 2); + } while(--time); + + if (WARN_ON(!time)) { + kvfree(msg); + return -ETIMEDOUT; + } + + cqe = (void *)((u8 *)gsp->shm.cmdq.ptr + 0x1000 + wptr * 0x1000); + step = min_t(u32, free, (gsp->cmdq.cnt - wptr)); + size = min_t(u32, len, step * GSP_PAGE_SIZE); + + memcpy(cqe, (u8 *)msg + off, size); + + wptr += DIV_ROUND_UP(size, 0x1000); + if (wptr == gsp->cmdq.cnt) + wptr = 0; + + off += size; + len -= size; + } while (len); + + nvkm_trace(&gsp->subdev, "cmdq: wptr %d\n", wptr); + wmb(); + (*gsp->cmdq.wptr) = wptr; + mb(); + + nvkm_falcon_wr32(&gsp->falcon, 0xc00, 0x00000000); + + kvfree(msg); + return 0; +} + +static void * +r535_gsp_cmdq_get(struct nvkm_gsp *gsp, u32 gsp_rpc_len) +{ + struct r535_gsp_msg *msg; + u32 size = GSP_MSG_HDR_SIZE + gsp_rpc_len; + + size = ALIGN(size, GSP_MSG_MIN_SIZE); + msg = kvzalloc(size, GFP_KERNEL); + if (!msg) + return ERR_PTR(-ENOMEM); + + msg->checksum = gsp_rpc_len; + return msg->data; +} + +static void +r535_gsp_msg_done(struct nvkm_gsp *gsp, struct nvfw_gsp_rpc *msg) +{ + kvfree(msg); +} + +static void +r535_gsp_msg_dump(struct nvkm_gsp *gsp, struct nvfw_gsp_rpc *msg, int lvl) +{ + if (gsp->subdev.debug >= lvl) { + nvkm_printk__(&gsp->subdev, lvl, info, + "msg fn:%d len:0x%x/0x%zx res:0x%x resp:0x%x\n", + msg->function, msg->length, msg->length - sizeof(*msg), + msg->rpc_result, msg->rpc_result_private); + print_hex_dump(KERN_INFO, "msg: ", DUMP_PREFIX_OFFSET, 16, 1, + msg->data, msg->length - sizeof(*msg), true); + } +} + +struct nvfw_gsp_rpc * +r535_gsp_msg_recv(struct nvkm_gsp *gsp, int fn, u32 gsp_rpc_len) +{ + struct nvkm_subdev *subdev = &gsp->subdev; + struct nvfw_gsp_rpc *rpc; + int retries = 4000000, i; + +retry: + rpc = r535_gsp_msgq_peek(gsp, sizeof(*rpc), &retries); + if (IS_ERR_OR_NULL(rpc)) + return rpc; + + rpc = r535_gsp_msgq_recv(gsp, gsp_rpc_len, &retries); + if (IS_ERR_OR_NULL(rpc)) + return rpc; + + if (rpc->rpc_result) { + r535_gsp_msg_dump(gsp, rpc, NV_DBG_ERROR); + r535_gsp_msg_done(gsp, rpc); + return ERR_PTR(-EINVAL); + } + + r535_gsp_msg_dump(gsp, rpc, NV_DBG_TRACE); + + if (fn && rpc->function == fn) { + if (gsp_rpc_len) { + if (rpc->length < gsp_rpc_len) { + nvkm_error(subdev, "rpc len %d < %d\n", + rpc->length, gsp_rpc_len); + r535_gsp_msg_dump(gsp, rpc, NV_DBG_ERROR); + r535_gsp_msg_done(gsp, rpc); + return ERR_PTR(-EIO); + } + + return rpc; + } + + r535_gsp_msg_done(gsp, rpc); + return NULL; + } + + for (i = 0; i < gsp->msgq.ntfy_nr; i++) { + struct nvkm_gsp_msgq_ntfy *ntfy = &gsp->msgq.ntfy[i]; + + if (ntfy->fn == rpc->function) { + if (ntfy->func) + ntfy->func(ntfy->priv, ntfy->fn, rpc->data, + rpc->length - sizeof(*rpc)); + break; + } + } + + if (i == gsp->msgq.ntfy_nr) + r535_gsp_msg_dump(gsp, rpc, NV_DBG_WARN); + + r535_gsp_msg_done(gsp, rpc); + if (fn) + goto retry; + + if (*gsp->msgq.rptr != *gsp->msgq.wptr) + goto retry; + + return NULL; +} + +int +r535_gsp_msg_ntfy_add(struct nvkm_gsp *gsp, u32 fn, nvkm_gsp_msg_ntfy_func func, void *priv) +{ + int ret = 0; + + mutex_lock(&gsp->msgq.mutex); + if (WARN_ON(gsp->msgq.ntfy_nr >= ARRAY_SIZE(gsp->msgq.ntfy))) { + ret = -ENOSPC; + } else { + gsp->msgq.ntfy[gsp->msgq.ntfy_nr].fn = fn; + gsp->msgq.ntfy[gsp->msgq.ntfy_nr].func = func; + gsp->msgq.ntfy[gsp->msgq.ntfy_nr].priv = priv; + gsp->msgq.ntfy_nr++; + } + mutex_unlock(&gsp->msgq.mutex); + return ret; +} + +int +r535_gsp_rpc_poll(struct nvkm_gsp *gsp, u32 fn) +{ + void *repv; + + mutex_lock(&gsp->cmdq.mutex); + repv = r535_gsp_msg_recv(gsp, fn, 0); + mutex_unlock(&gsp->cmdq.mutex); + if (IS_ERR(repv)) + return PTR_ERR(repv); + + return 0; +} + +static void * +r535_gsp_rpc_handle_reply(struct nvkm_gsp *gsp, u32 fn, + enum nvkm_gsp_rpc_reply_policy policy, + u32 gsp_rpc_len) +{ + struct nvfw_gsp_rpc *reply; + void *repv = NULL; + + switch (policy) { + case NVKM_GSP_RPC_REPLY_NOWAIT: + break; + case NVKM_GSP_RPC_REPLY_RECV: + reply = r535_gsp_msg_recv(gsp, fn, gsp_rpc_len); + if (!IS_ERR_OR_NULL(reply)) + repv = reply->data; + else + repv = reply; + break; + case NVKM_GSP_RPC_REPLY_POLL: + repv = r535_gsp_msg_recv(gsp, fn, 0); + break; + } + + return repv; +} + +static void * +r535_gsp_rpc_send(struct nvkm_gsp *gsp, void *payload, + enum nvkm_gsp_rpc_reply_policy policy, u32 gsp_rpc_len) +{ + struct nvfw_gsp_rpc *rpc = to_gsp_hdr(payload, rpc); + u32 fn = rpc->function; + int ret; + + if (gsp->subdev.debug >= NV_DBG_TRACE) { + nvkm_trace(&gsp->subdev, "rpc fn:%d len:0x%x/0x%zx\n", rpc->function, + rpc->length, rpc->length - sizeof(*rpc)); + print_hex_dump(KERN_INFO, "rpc: ", DUMP_PREFIX_OFFSET, 16, 1, + rpc->data, rpc->length - sizeof(*rpc), true); + } + + ret = r535_gsp_cmdq_push(gsp, rpc); + if (ret) + return ERR_PTR(ret); + + return r535_gsp_rpc_handle_reply(gsp, fn, policy, gsp_rpc_len); +} + +static void +r535_gsp_rpc_done(struct nvkm_gsp *gsp, void *repv) +{ + struct nvfw_gsp_rpc *rpc = container_of(repv, typeof(*rpc), data); + + r535_gsp_msg_done(gsp, rpc); +} + +static void * +r535_gsp_rpc_get(struct nvkm_gsp *gsp, u32 fn, u32 payload_size) +{ + struct nvfw_gsp_rpc *rpc; + + rpc = r535_gsp_cmdq_get(gsp, ALIGN(sizeof(*rpc) + payload_size, + sizeof(u64))); + if (IS_ERR(rpc)) + return ERR_CAST(rpc); + + rpc->header_version = 0x03000000; + rpc->signature = ('C' << 24) | ('P' << 16) | ('R' << 8) | 'V'; + rpc->function = fn; + rpc->rpc_result = 0xffffffff; + rpc->rpc_result_private = 0xffffffff; + rpc->length = sizeof(*rpc) + payload_size; + return rpc->data; +} + +static void * +r535_gsp_rpc_push(struct nvkm_gsp *gsp, void *payload, + enum nvkm_gsp_rpc_reply_policy policy, u32 gsp_rpc_len) +{ + struct nvfw_gsp_rpc *rpc = to_gsp_hdr(payload, rpc); + struct r535_gsp_msg *msg = to_gsp_hdr(rpc, msg); + const u32 max_rpc_size = GSP_MSG_MAX_SIZE - sizeof(*msg); + const u32 max_payload_size = max_rpc_size - sizeof(*rpc); + u32 payload_size = rpc->length - sizeof(*rpc); + void *repv; + + mutex_lock(&gsp->cmdq.mutex); + if (payload_size > max_payload_size) { + const u32 fn = rpc->function; + u32 remain_payload_size = payload_size; + + /* Adjust length, and send initial RPC. */ + rpc->length = sizeof(*rpc) + max_payload_size; + msg->checksum = rpc->length; + + repv = r535_gsp_rpc_send(gsp, payload, NVKM_GSP_RPC_REPLY_NOWAIT, 0); + if (IS_ERR(repv)) + goto done; + + payload += max_payload_size; + remain_payload_size -= max_payload_size; + + /* Remaining chunks sent as CONTINUATION_RECORD RPCs. */ + while (remain_payload_size) { + u32 size = min(remain_payload_size, + max_payload_size); + void *next; + + next = r535_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_CONTINUATION_RECORD, size); + if (IS_ERR(next)) { + repv = next; + goto done; + } + + memcpy(next, payload, size); + + repv = r535_gsp_rpc_send(gsp, next, NVKM_GSP_RPC_REPLY_NOWAIT, 0); + if (IS_ERR(repv)) + goto done; + + payload += size; + remain_payload_size -= size; + } + + /* Wait for reply. */ + repv = r535_gsp_rpc_handle_reply(gsp, fn, policy, payload_size + + sizeof(*rpc)); + } else { + repv = r535_gsp_rpc_send(gsp, payload, policy, gsp_rpc_len); + } + +done: + mutex_unlock(&gsp->cmdq.mutex); + return repv; +} + +const struct nvkm_rm_api_rpc +r535_rpc = { + .get = r535_gsp_rpc_get, + .push = r535_gsp_rpc_push, + .done = r535_gsp_rpc_done, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/vmm.c index d3e95453f25d..52f2e5f14517 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/vmm.c @@ -19,15 +19,38 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#include "vmm.h" +#include <subdev/mmu/vmm.h> -#include <nvrm/nvtypes.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/class/cl90f1.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/ctrl/ctrl90f1.h> -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvos.h> +#include <nvhw/drf.h> +#include "nvrm/vmm.h" -static int -r535_mmu_promote_vmm(struct nvkm_vmm *vmm) +void +r535_mmu_vaspace_del(struct nvkm_vmm *vmm) +{ + if (vmm->rm.external) { + NV0080_CTRL_DMA_UNSET_PAGE_DIRECTORY_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&vmm->rm.device.object, + NV0080_CTRL_CMD_DMA_UNSET_PAGE_DIRECTORY, + sizeof(*ctrl)); + if (!IS_ERR(ctrl)) { + ctrl->hVASpace = vmm->rm.object.handle; + + WARN_ON(nvkm_gsp_rm_ctrl_wr(&vmm->rm.device.object, ctrl)); + } + + vmm->rm.external = false; + } + + nvkm_gsp_rm_free(&vmm->rm.object); + nvkm_gsp_device_dtor(&vmm->rm.device); + nvkm_gsp_client_dtor(&vmm->rm.client); + + nvkm_vmm_put(vmm, &vmm->rm.rsvd); +} + +int +r535_mmu_vaspace_new(struct nvkm_vmm *vmm, u32 handle, bool external) { NV_VASPACE_ALLOCATION_PARAMETERS *args; int ret; @@ -37,58 +60,103 @@ r535_mmu_promote_vmm(struct nvkm_vmm *vmm) if (ret) return ret; - args = nvkm_gsp_rm_alloc_get(&vmm->rm.device.object, 0x90f10000, FERMI_VASPACE_A, + args = nvkm_gsp_rm_alloc_get(&vmm->rm.device.object, handle, FERMI_VASPACE_A, sizeof(*args), &vmm->rm.object); if (IS_ERR(args)) return PTR_ERR(args); args->index = NV_VASPACE_ALLOCATION_INDEX_GPU_NEW; + if (external) + args->flags = NV_VASPACE_ALLOCATION_FLAGS_IS_EXTERNALLY_OWNED; ret = nvkm_gsp_rm_alloc_wr(&vmm->rm.object, args); if (ret) return ret; - { + if (!external) { NV90F1_CTRL_VASPACE_COPY_SERVER_RESERVED_PDES_PARAMS *ctrl; + u8 page_shift = 29; /* 512MiB */ + const u64 page_size = BIT_ULL(page_shift); + const struct nvkm_vmm_page *page; + const struct nvkm_vmm_desc *desc; + struct nvkm_vmm_pt *pd = vmm->pd; + + for (page = vmm->func->page; page->shift; page++) { + if (page->shift == page_shift) + break; + } + + if (WARN_ON(!page->shift)) + return -EINVAL; mutex_lock(&vmm->mutex.vmm); - ret = nvkm_vmm_get_locked(vmm, true, false, false, 0x1d, 32, 0x20000000, + ret = nvkm_vmm_get_locked(vmm, true, false, false, page_shift, 32, page_size, &vmm->rm.rsvd); mutex_unlock(&vmm->mutex.vmm); if (ret) return ret; + /* Some parts of RM expect the server-reserved area to be in a specific location. */ + if (WARN_ON(vmm->rm.rsvd->addr != SPLIT_VAS_SERVER_RM_MANAGED_VA_START || + vmm->rm.rsvd->size != SPLIT_VAS_SERVER_RM_MANAGED_VA_SIZE)) + return -EINVAL; + ctrl = nvkm_gsp_rm_ctrl_get(&vmm->rm.object, NV90F1_CTRL_CMD_VASPACE_COPY_SERVER_RESERVED_PDES, sizeof(*ctrl)); if (IS_ERR(ctrl)) return PTR_ERR(ctrl); - ctrl->pageSize = 0x20000000; + ctrl->pageSize = page_size; ctrl->virtAddrLo = vmm->rm.rsvd->addr; ctrl->virtAddrHi = vmm->rm.rsvd->addr + vmm->rm.rsvd->size - 1; - ctrl->numLevelsToCopy = vmm->pd->pde[0]->pde[0] ? 3 : 2; - ctrl->levels[0].physAddress = vmm->pd->pt[0]->addr; - ctrl->levels[0].size = 0x20; - ctrl->levels[0].aperture = 1; - ctrl->levels[0].pageShift = 0x2f; - ctrl->levels[1].physAddress = vmm->pd->pde[0]->pt[0]->addr; - ctrl->levels[1].size = 0x1000; - ctrl->levels[1].aperture = 1; - ctrl->levels[1].pageShift = 0x26; - if (vmm->pd->pde[0]->pde[0]) { - ctrl->levels[2].physAddress = vmm->pd->pde[0]->pde[0]->pt[0]->addr; - ctrl->levels[2].size = 0x1000; - ctrl->levels[2].aperture = 1; - ctrl->levels[2].pageShift = 0x1d; + + for (desc = page->desc; desc->bits; desc++) { + ctrl->numLevelsToCopy++; + page_shift += desc->bits; + } + desc--; + + for (int i = 0; i < ctrl->numLevelsToCopy; i++, desc--) { + page_shift -= desc->bits; + + ctrl->levels[i].physAddress = pd->pt[0]->addr; + ctrl->levels[i].size = (1 << desc->bits) * desc->size; + ctrl->levels[i].aperture = 1; + ctrl->levels[i].pageShift = page_shift; + + pd = pd->pde[0]; } ret = nvkm_gsp_rm_ctrl_wr(&vmm->rm.object, ctrl); + } else { + NV0080_CTRL_DMA_SET_PAGE_DIRECTORY_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&vmm->rm.device.object, + NV0080_CTRL_CMD_DMA_SET_PAGE_DIRECTORY, + sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->physAddress = vmm->pd->pt[0]->addr; + ctrl->numEntries = 1 << vmm->func->page[0].desc->bits; + ctrl->flags = NVDEF(NV0080_CTRL_DMA_SET_PAGE_DIRECTORY, FLAGS, APERTURE, VIDMEM); + ctrl->hVASpace = vmm->rm.object.handle; + + ret = nvkm_gsp_rm_ctrl_wr(&vmm->rm.device.object, ctrl); + if (ret == 0) + vmm->rm.external = true; } return ret; } +static int +r535_mmu_promote_vmm(struct nvkm_vmm *vmm) +{ + return r535_mmu_vaspace_new(vmm, NVKM_RM_VASPACE, true); +} + static void r535_mmu_dtor(struct nvkm_mmu *mmu) { diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/Kbuild new file mode 100644 index 000000000000..5db0e7009e1f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/Kbuild @@ -0,0 +1,9 @@ +# SPDX-License-Identifier: MIT +nvkm-y += nvkm/subdev/gsp/rm/r570/rm.o +nvkm-y += nvkm/subdev/gsp/rm/r570/gsp.o +nvkm-y += nvkm/subdev/gsp/rm/r570/client.o +nvkm-y += nvkm/subdev/gsp/rm/r570/fbsr.o +nvkm-y += nvkm/subdev/gsp/rm/r570/disp.o +nvkm-y += nvkm/subdev/gsp/rm/r570/fifo.o +nvkm-y += nvkm/subdev/gsp/rm/r570/gr.o +nvkm-y += nvkm/subdev/gsp/rm/r570/ofa.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/client.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/client.c new file mode 100644 index 000000000000..87e6240662ed --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/client.c @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/rm.h> + +#include "nvrm/client.h" + +static int +r570_gsp_client_ctor(struct nvkm_gsp_client *client, u32 handle) +{ + NV0000_ALLOC_PARAMETERS *args; + + args = nvkm_gsp_rm_alloc_get(&client->object, handle, NV01_ROOT, sizeof(*args), + &client->object); + if (IS_ERR(args)) + return PTR_ERR(args); + + args->hClient = client->object.handle; + args->processID = ~0; + + return nvkm_gsp_rm_alloc_wr(&client->object, args); +} + +const struct nvkm_rm_api_client +r570_client = { + .ctor = r570_gsp_client_ctor, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/disp.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/disp.c new file mode 100644 index 000000000000..a96e31c2d80b --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/disp.c @@ -0,0 +1,263 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/rm.h> + +#include <engine/disp.h> +#include <engine/disp/outp.h> + +#include "nvhw/drf.h" + +#include "nvrm/disp.h" + +static int +r570_dmac_alloc(struct nvkm_disp *disp, u32 oclass, int inst, u32 put_offset, + struct nvkm_gsp_object *dmac) +{ + NV50VAIO_CHANNELDMA_ALLOCATION_PARAMETERS *args; + + args = nvkm_gsp_rm_alloc_get(&disp->rm.object, (oclass << 16) | inst, oclass, + sizeof(*args), dmac); + if (IS_ERR(args)) + return PTR_ERR(args); + + args->channelInstance = inst; + args->offset = put_offset; + args->subDeviceId = BIT(0); + + return nvkm_gsp_rm_alloc_wr(dmac, args); +} + +static int +r570_disp_chan_set_pushbuf(struct nvkm_disp *disp, s32 oclass, int inst, struct nvkm_memory *memory) +{ + struct nvkm_gsp *gsp = disp->rm.objcom.client->gsp; + NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER, + sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + if (memory) { + switch (nvkm_memory_target(memory)) { + case NVKM_MEM_TARGET_NCOH: + ctrl->addressSpace = ADDR_SYSMEM; + ctrl->cacheSnoop = 0; + ctrl->pbTargetAperture = PHYS_PCI; + break; + case NVKM_MEM_TARGET_HOST: + ctrl->addressSpace = ADDR_SYSMEM; + ctrl->cacheSnoop = 1; + ctrl->pbTargetAperture = PHYS_PCI_COHERENT; + break; + case NVKM_MEM_TARGET_VRAM: + ctrl->addressSpace = ADDR_FBMEM; + ctrl->pbTargetAperture = PHYS_NVM; + break; + default: + WARN_ON(1); + return -EINVAL; + } + + ctrl->physicalAddr = nvkm_memory_addr(memory); + ctrl->limit = nvkm_memory_size(memory) - 1; + } + + ctrl->hclass = oclass; + ctrl->channelInstance = inst; + ctrl->valid = ((oclass & 0xff) != 0x7a) ? 1 : 0; + ctrl->subDeviceId = BIT(0); + + return nvkm_gsp_rm_ctrl_wr(&gsp->internal.device.subdevice, ctrl); +} + +static int +r570_dp_set_indexed_link_rates(struct nvkm_outp *outp) +{ + NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS *ctrl; + struct nvkm_disp *disp = outp->disp; + + if (WARN_ON(outp->dp.rates > ARRAY_SIZE(ctrl->linkRateTbl))) + return -EINVAL; + + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, + NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->displayId = BIT(outp->index); + for (int i = 0; i < outp->dp.rates; i++) + ctrl->linkRateTbl[outp->dp.rate[i].dpcd] = outp->dp.rate[i].rate * 10 / 200; + + return nvkm_gsp_rm_ctrl_wr(&disp->rm.objcom, ctrl); +} + +static int +r570_dp_get_caps(struct nvkm_disp *disp, int *plink_bw, bool *pmst, bool *pwm) +{ + NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS *ctrl; + int ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, + NV0073_CTRL_CMD_DP_GET_CAPS, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->sorIndex = ~0; + + ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); + if (ret) { + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return ret; + } + + switch (NVVAL_GET(ctrl->maxLinkRate, NV0073_CTRL_CMD, DP_GET_CAPS, MAX_LINK_RATE)) { + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_1_62: + *plink_bw = 0x06; + break; + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_2_70: + *plink_bw = 0x0a; + break; + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_5_40: + *plink_bw = 0x14; + break; + case NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_8_10: + *plink_bw = 0x1e; + break; + default: + *plink_bw = 0x00; + break; + } + + *pmst = ctrl->bIsMultistreamSupported; + *pwm = ctrl->bHasIncreasedWatermarkLimits; + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; +} + +static int +r570_bl_ctrl(struct nvkm_disp *disp, unsigned display_id, bool set, int *pval) +{ + u32 cmd = set ? NV0073_CTRL_CMD_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS : + NV0073_CTRL_CMD_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS; + NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS *ctrl; + int ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, cmd, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->displayId = BIT(display_id); + ctrl->brightness = *pval; + ctrl->brightnessType = NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_TYPE_PERCENT100; + + ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); + if (ret) + return ret; + + *pval = ctrl->brightness; + + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; +} + +static int +r570_disp_get_active(struct nvkm_disp *disp, unsigned head, u32 *displayid) +{ + NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS *ctrl; + int ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, + NV0073_CTRL_CMD_SYSTEM_GET_ACTIVE, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->subDeviceInstance = 0; + ctrl->head = head; + + ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); + if (ret) { + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return ret; + } + + *displayid = ctrl->displayId; + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; +} +static int +r570_disp_get_connect_state(struct nvkm_disp *disp, unsigned display_id) +{ + NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS *ctrl; + int ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&disp->rm.objcom, + NV0073_CTRL_CMD_SYSTEM_GET_CONNECT_STATE, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->subDeviceInstance = 0; + ctrl->displayMask = BIT(display_id); + + ret = nvkm_gsp_rm_ctrl_push(&disp->rm.objcom, &ctrl, sizeof(*ctrl)); + if (ret == 0 && (ctrl->displayMask & BIT(display_id))) + ret = 1; + + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return ret; +} + +static int +r570_disp_get_supported(struct nvkm_disp *disp, unsigned long *pmask) +{ + NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_rd(&disp->rm.objcom, + NV0073_CTRL_CMD_SYSTEM_GET_SUPPORTED, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + *pmask = ctrl->displayMask; + + nvkm_gsp_rm_ctrl_done(&disp->rm.objcom, ctrl); + return 0; +} + +static int +r570_disp_get_static_info(struct nvkm_disp *disp) +{ + NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS *ctrl; + struct nvkm_gsp *gsp = disp->engine.subdev.device->gsp; + + ctrl = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_DISPLAY_GET_STATIC_INFO, + sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + disp->wndw.mask = ctrl->windowPresentMask; + disp->wndw.nr = fls(disp->wndw.mask); + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); + return 0; +} + +const struct nvkm_rm_api_disp +r570_disp = { + .get_static_info = r570_disp_get_static_info, + .get_supported = r570_disp_get_supported, + .get_connect_state = r570_disp_get_connect_state, + .get_active = r570_disp_get_active, + .bl_ctrl = r570_bl_ctrl, + .dp = { + .get_caps = r570_dp_get_caps, + .set_indexed_link_rates = r570_dp_set_indexed_link_rates, + }, + .chan = { + .set_pushbuf = r570_disp_chan_set_pushbuf, + .dmac_alloc = r570_dmac_alloc, + }, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/fbsr.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/fbsr.c new file mode 100644 index 000000000000..2945d5b4e570 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/fbsr.c @@ -0,0 +1,149 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <subdev/instmem/priv.h> +#include <subdev/bar.h> +#include <subdev/gsp.h> +#include <subdev/mmu/vmm.h> + +#include "nvrm/fbsr.h" +#include "nvrm/fifo.h" + +static int +r570_fbsr_suspend_channels(struct nvkm_gsp *gsp, bool suspend) +{ + NV2080_CTRL_CMD_INTERNAL_FIFO_TOGGLE_ACTIVE_CHANNEL_SCHEDULING_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_FIFO_TOGGLE_ACTIVE_CHANNEL_SCHEDULING, + sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->bDisableActiveChannels = suspend; + + return nvkm_gsp_rm_ctrl_wr(&gsp->internal.device.subdevice, ctrl); +} + +static void +r570_fbsr_resume(struct nvkm_gsp *gsp) +{ + struct nvkm_device *device = gsp->subdev.device; + struct nvkm_instmem *imem = device->imem; + struct nvkm_instobj *iobj; + struct nvkm_vmm *vmm; + + /* Restore BAR2 page tables via BAR0 window, and re-enable BAR2. */ + list_for_each_entry(iobj, &imem->boot, head) { + if (iobj->suspend) + nvkm_instobj_load(iobj); + } + + device->bar->bar2 = true; + + vmm = nvkm_bar_bar2_vmm(device); + vmm->func->flush(vmm, 0); + + /* Restore remaining BAR2 allocations (including BAR1 page tables) via BAR2. */ + list_for_each_entry(iobj, &imem->list, head) { + if (iobj->suspend) + nvkm_instobj_load(iobj); + } + + vmm = nvkm_bar_bar1_vmm(device); + vmm->func->flush(vmm, 0); + + /* Resume channel scheduling. */ + r570_fbsr_suspend_channels(device->gsp, false); + + /* Finish cleaning up. */ + r535_fbsr_resume(gsp); +} + +static int +r570_fbsr_init(struct nvkm_gsp *gsp, struct sg_table *sgt, u64 size) +{ + NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS *ctrl; + struct nvkm_gsp_object memlist; + int ret; + + ret = r535_fbsr_memlist(&gsp->internal.device, 0xcaf00003, NVKM_MEM_TARGET_HOST, + 0, size, sgt, &memlist); + if (ret) + return ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_FBSR_INIT, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->hClient = gsp->internal.client.object.handle; + ctrl->hSysMem = memlist.handle; + ctrl->sysmemAddrOfSuspendResumeData = gsp->sr.meta.addr; + ctrl->bEnteringGcoffState = 1; + + ret = nvkm_gsp_rm_ctrl_wr(&gsp->internal.device.subdevice, ctrl); + if (ret) + return ret; + + nvkm_gsp_rm_free(&memlist); + return 0; +} + +static int +r570_fbsr_suspend(struct nvkm_gsp *gsp) +{ + struct nvkm_subdev *subdev = &gsp->subdev; + struct nvkm_device *device = subdev->device; + struct nvkm_instmem *imem = device->imem; + struct nvkm_instobj *iobj; + u64 size; + int ret; + + /* Stop channel scheduling. */ + r570_fbsr_suspend_channels(gsp, true); + + /* Save BAR2 allocations to system memory. */ + list_for_each_entry(iobj, &imem->list, head) { + if (iobj->preserve) { + ret = nvkm_instobj_save(iobj); + if (ret) + return ret; + } + } + + list_for_each_entry(iobj, &imem->boot, head) { + ret = nvkm_instobj_save(iobj); + if (ret) + return ret; + } + + /* Disable BAR2 access. */ + device->bar->bar2 = false; + + /* Allocate system memory to hold RM's VRAM allocations across suspend. */ + size = gsp->fb.heap.size; + size += gsp->fb.rsvd_size; + size += gsp->fb.bios.vga_workspace.size; + nvkm_debug(subdev, "fbsr: size: 0x%llx bytes\n", size); + + ret = nvkm_gsp_sg(device, size, &gsp->sr.fbsr); + if (ret) + return ret; + + /* Initialise FBSR on RM. */ + ret = r570_fbsr_init(gsp, &gsp->sr.fbsr, size); + if (ret) { + nvkm_gsp_sg_free(device, &gsp->sr.fbsr); + return ret; + } + + return 0; +} + +const struct nvkm_rm_api_fbsr +r570_fbsr = { + .suspend = r570_fbsr_suspend, + .resume = r570_fbsr_resume, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/fifo.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/fifo.c new file mode 100644 index 000000000000..79132805cfcf --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/fifo.c @@ -0,0 +1,217 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/rm.h> + +#include <subdev/mmu.h> +#include <engine/fifo/priv.h> +#include <engine/fifo/chan.h> +#include <engine/fifo/runl.h> + +#include "nvhw/drf.h" + +#include "nvrm/fifo.h" +#include "nvrm/engine.h" + +#define CHID_PER_USERD 8 + +static int +r570_chan_alloc(struct nvkm_gsp_device *device, u32 handle, u32 nv2080_engine_type, u8 runq, + bool priv, int chid, u64 inst_addr, u64 userd_addr, u64 mthdbuf_addr, + struct nvkm_vmm *vmm, u64 gpfifo_offset, u32 gpfifo_length, + struct nvkm_gsp_object *chan) +{ + struct nvkm_gsp *gsp = device->object.client->gsp; + struct nvkm_fifo *fifo = gsp->subdev.device->fifo; + const int userd_p = chid / CHID_PER_USERD; + const int userd_i = chid % CHID_PER_USERD; + NV_CHANNELGPFIFO_ALLOCATION_PARAMETERS *args; + + args = nvkm_gsp_rm_alloc_get(&device->object, handle, + fifo->func->chan.user.oclass, sizeof(*args), chan); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); + + args->gpFifoOffset = gpfifo_offset; + args->gpFifoEntries = gpfifo_length / 8; + + args->flags = NVDEF(NVOS04, FLAGS, CHANNEL_TYPE, PHYSICAL); + args->flags |= NVDEF(NVOS04, FLAGS, VPR, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_SKIP_MAP_REFCOUNTING, FALSE); + args->flags |= NVVAL(NVOS04, FLAGS, GROUP_CHANNEL_RUNQUEUE, runq); + if (!priv) + args->flags |= NVDEF(NVOS04, FLAGS, PRIVILEGED_CHANNEL, FALSE); + else + args->flags |= NVDEF(NVOS04, FLAGS, PRIVILEGED_CHANNEL, TRUE); + args->flags |= NVDEF(NVOS04, FLAGS, DELAY_CHANNEL_SCHEDULING, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_DENY_PHYSICAL_MODE_CE, FALSE); + + args->flags |= NVVAL(NVOS04, FLAGS, CHANNEL_USERD_INDEX_VALUE, userd_i); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_USERD_INDEX_FIXED, FALSE); + args->flags |= NVVAL(NVOS04, FLAGS, CHANNEL_USERD_INDEX_PAGE_VALUE, userd_p); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_USERD_INDEX_PAGE_FIXED, TRUE); + + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_DENY_AUTH_LEVEL_PRIV, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_SKIP_SCRUBBER, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_CLIENT_MAP_FIFO, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, SET_EVICT_LAST_CE_PREFETCH_CHANNEL, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_VGPU_PLUGIN_CONTEXT, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, CHANNEL_PBDMA_ACQUIRE_TIMEOUT, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, GROUP_CHANNEL_THREAD, DEFAULT); + args->flags |= NVDEF(NVOS04, FLAGS, MAP_CHANNEL, FALSE); + args->flags |= NVDEF(NVOS04, FLAGS, SKIP_CTXBUFFER_ALLOC, FALSE); + + args->hVASpace = vmm->rm.object.handle; + args->engineType = nv2080_engine_type; + + args->instanceMem.base = inst_addr; + args->instanceMem.size = fifo->func->chan.func->inst->size; + args->instanceMem.addressSpace = 2; + args->instanceMem.cacheAttrib = 1; + + args->userdMem.base = userd_addr; + args->userdMem.size = fifo->func->chan.func->userd->size; + args->userdMem.addressSpace = 2; + args->userdMem.cacheAttrib = 1; + + args->ramfcMem.base = inst_addr; + args->ramfcMem.size = 0x200; + args->ramfcMem.addressSpace = 2; + args->ramfcMem.cacheAttrib = 1; + + args->mthdbufMem.base = mthdbuf_addr; + args->mthdbufMem.size = fifo->rm.mthdbuf_size; + args->mthdbufMem.addressSpace = 1; + args->mthdbufMem.cacheAttrib = 0; + + if (!priv) + args->internalFlags = NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, PRIVILEGE, USER); + else + args->internalFlags = NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, PRIVILEGE, ADMIN); + args->internalFlags |= NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, ERROR_NOTIFIER_TYPE, NONE); + args->internalFlags |= NVDEF(NV_KERNELCHANNEL, ALLOC_INTERNALFLAGS, ECC_ERROR_NOTIFIER_TYPE, NONE); + + return nvkm_gsp_rm_alloc_wr(chan, args); +} + +static int +r570_fifo_rc_triggered(void *priv, u32 fn, void *repv, u32 repc) +{ + rpc_rc_triggered_v17_02 *msg = repv; + struct nvkm_gsp *gsp = priv; + + if (WARN_ON(repc < sizeof(*msg))) + return -EINVAL; + + nvkm_error(&gsp->subdev, "rc engn:%08x chid:%d gfid:%d level:%d type:%d scope:%d part:%d " + "fault_addr:%08x%08x fault_type:%08x\n", + msg->nv2080EngineType, msg->chid, msg->gfid, msg->exceptLevel, msg->exceptType, + msg->scope, msg->partitionAttributionId, + msg->mmuFaultAddrHi, msg->mmuFaultAddrLo, msg->mmuFaultType); + + r535_fifo_rc_chid(gsp->subdev.device->fifo, msg->chid); + return 0; +} + +static int +r570_fifo_ectx_size(struct nvkm_fifo *fifo) +{ + NV2080_CTRL_GPU_GET_CONSTRUCTED_FALCON_INFO_PARAMS *ctrl; + struct nvkm_gsp *gsp = fifo->engine.subdev.device->gsp; + struct nvkm_runl *runl; + struct nvkm_engn *engn; + + ctrl = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_GPU_GET_CONSTRUCTED_FALCON_INFO, + sizeof(*ctrl)); + if (WARN_ON(IS_ERR(ctrl))) + return PTR_ERR(ctrl); + + for (int i = 0; i < ctrl->numConstructedFalcons; i++) { + nvkm_runl_foreach(runl, fifo) { + nvkm_runl_foreach_engn(engn, runl) { + if (engn->rm.desc == ctrl->constructedFalconsTable[i].engDesc) { + engn->rm.size = + ctrl->constructedFalconsTable[i].ctxBufferSize; + break; + } + } + } + } + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); + return 0; +} + +static int +r570_fifo_xlat_rm_engine_type(u32 rm, enum nvkm_subdev_type *ptype, int *p2080) +{ +#define RM_ENGINE_TYPE(RM,NVKM,INST) \ + RM_ENGINE_TYPE_##RM: \ + *ptype = NVKM_ENGINE_##NVKM; \ + *p2080 = NV2080_ENGINE_TYPE_##RM; \ + return INST + + switch (rm) { + case RM_ENGINE_TYPE( GR0, GR, 0); + case RM_ENGINE_TYPE( COPY0, CE, 0); + case RM_ENGINE_TYPE( COPY1, CE, 1); + case RM_ENGINE_TYPE( COPY2, CE, 2); + case RM_ENGINE_TYPE( COPY3, CE, 3); + case RM_ENGINE_TYPE( COPY4, CE, 4); + case RM_ENGINE_TYPE( COPY5, CE, 5); + case RM_ENGINE_TYPE( COPY6, CE, 6); + case RM_ENGINE_TYPE( COPY7, CE, 7); + case RM_ENGINE_TYPE( COPY8, CE, 8); + case RM_ENGINE_TYPE( COPY9, CE, 9); + case RM_ENGINE_TYPE( COPY10, CE, 10); + case RM_ENGINE_TYPE( COPY11, CE, 11); + case RM_ENGINE_TYPE( COPY12, CE, 12); + case RM_ENGINE_TYPE( COPY13, CE, 13); + case RM_ENGINE_TYPE( COPY14, CE, 14); + case RM_ENGINE_TYPE( COPY15, CE, 15); + case RM_ENGINE_TYPE( COPY16, CE, 16); + case RM_ENGINE_TYPE( COPY17, CE, 17); + case RM_ENGINE_TYPE( COPY18, CE, 18); + case RM_ENGINE_TYPE( COPY19, CE, 19); + case RM_ENGINE_TYPE( NVDEC0, NVDEC, 0); + case RM_ENGINE_TYPE( NVDEC1, NVDEC, 1); + case RM_ENGINE_TYPE( NVDEC2, NVDEC, 2); + case RM_ENGINE_TYPE( NVDEC3, NVDEC, 3); + case RM_ENGINE_TYPE( NVDEC4, NVDEC, 4); + case RM_ENGINE_TYPE( NVDEC5, NVDEC, 5); + case RM_ENGINE_TYPE( NVDEC6, NVDEC, 6); + case RM_ENGINE_TYPE( NVDEC7, NVDEC, 7); + case RM_ENGINE_TYPE( NVENC0, NVENC, 0); + case RM_ENGINE_TYPE( NVENC1, NVENC, 1); + case RM_ENGINE_TYPE( NVENC2, NVENC, 2); + case RM_ENGINE_TYPE( NVENC3, NVENC, 3); + case RM_ENGINE_TYPE(NVJPEG0, NVJPG, 0); + case RM_ENGINE_TYPE(NVJPEG1, NVJPG, 1); + case RM_ENGINE_TYPE(NVJPEG2, NVJPG, 2); + case RM_ENGINE_TYPE(NVJPEG3, NVJPG, 3); + case RM_ENGINE_TYPE(NVJPEG4, NVJPG, 4); + case RM_ENGINE_TYPE(NVJPEG5, NVJPG, 5); + case RM_ENGINE_TYPE(NVJPEG6, NVJPG, 6); + case RM_ENGINE_TYPE(NVJPEG7, NVJPG, 7); + case RM_ENGINE_TYPE( SW, SW, 0); + case RM_ENGINE_TYPE( SEC2, SEC2, 0); + case RM_ENGINE_TYPE( OFA0, OFA, 0); + case RM_ENGINE_TYPE( OFA1, OFA, 1); + default: + return -EINVAL; + } +#undef RM_ENGINE_TYPE +} + +const struct nvkm_rm_api_fifo +r570_fifo = { + .xlat_rm_engine_type = r570_fifo_xlat_rm_engine_type, + .ectx_size = r570_fifo_ectx_size, + .rsvd_chids = 1, + .rc_triggered = r570_fifo_rc_triggered, + .chan = { + .alloc = r570_chan_alloc, + }, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/gr.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/gr.c new file mode 100644 index 000000000000..b6cced9b8aa1 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/gr.c @@ -0,0 +1,191 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/gr.h> + +#include <subdev/mmu.h> +#include <engine/fifo.h> +#include <engine/fifo/chid.h> +#include <engine/gr/priv.h> + +#include "nvrm/gr.h" +#include "nvrm/engine.h" + +int +r570_gr_tpc_mask(struct nvkm_gsp *gsp, int gpc, u32 *pmask) +{ + NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS *ctrl; + int ret; + + ctrl = nvkm_gsp_rm_ctrl_get(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_GPU_GET_FERMI_TPC_INFO, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->gpcId = gpc; + + ret = nvkm_gsp_rm_ctrl_push(&gsp->internal.device.subdevice, &ctrl, sizeof(*ctrl)); + if (ret) + return ret; + + *pmask = ctrl->tpcMask; + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); + return 0; +} + +int +r570_gr_gpc_mask(struct nvkm_gsp *gsp, u32 *pmask) +{ + NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_GPU_GET_FERMI_GPC_INFO, sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + *pmask = ctrl->gpcMask; + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, ctrl); + return 0; +} + +static int +r570_gr_scrubber_ctrl(struct r535_gr *gr, bool teardown) +{ + NV2080_CTRL_INTERNAL_GR_INIT_BUG4208224_WAR_PARAMS *ctrl; + + ctrl = nvkm_gsp_rm_ctrl_get(&gr->scrubber.vmm->rm.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_KGR_INIT_BUG4208224_WAR, + sizeof(*ctrl)); + if (IS_ERR(ctrl)) + return PTR_ERR(ctrl); + + ctrl->bTeardown = teardown; + + return nvkm_gsp_rm_ctrl_wr(&gr->scrubber.vmm->rm.device.subdevice, ctrl); +} + +static void +r570_gr_scrubber_fini(struct r535_gr *gr) +{ + /* Teardown scrubber channel on RM. */ + if (gr->scrubber.enabled) { + WARN_ON(r570_gr_scrubber_ctrl(gr, true)); + gr->scrubber.enabled = false; + } + + /* Free scrubber channel. */ + nvkm_gsp_rm_free(&gr->scrubber.threed); + nvkm_gsp_rm_free(&gr->scrubber.chan); + + for (int i = 0; i < gr->ctxbuf_nr; i++) { + nvkm_vmm_put(gr->scrubber.vmm, &gr->scrubber.ctxbuf.vma[i]); + nvkm_memory_unref(&gr->scrubber.ctxbuf.mem[i]); + } + + nvkm_vmm_unref(&gr->scrubber.vmm); + nvkm_memory_unref(&gr->scrubber.inst); +} + +static int +r570_gr_scrubber_init(struct r535_gr *gr) +{ + struct nvkm_subdev *subdev = &gr->base.engine.subdev; + struct nvkm_device *device = subdev->device; + struct nvkm_gsp *gsp = device->gsp; + struct nvkm_rm *rm = gsp->rm; + int ret; + + /* Scrubber channel only required on TU10x. */ + switch (device->chipset) { + case 0x162: + case 0x164: + case 0x166: + break; + default: + return 0; + } + + if (gr->scrubber.chid < 0) { + gr->scrubber.chid = nvkm_chid_get(device->fifo->chid, NULL); + if (gr->scrubber.chid < 0) + return gr->scrubber.chid; + } + + /* Allocate scrubber channel. */ + ret = nvkm_memory_new(device, NVKM_MEM_TARGET_INST, + 0x2000 + device->fifo->rm.mthdbuf_size, 0, true, + &gr->scrubber.inst); + if (ret) + goto done; + + ret = nvkm_vmm_new(device, 0x1000, 0, NULL, 0, NULL, "grScrubberVmm", + &gr->scrubber.vmm); + if (ret) + goto done; + + ret = r535_mmu_vaspace_new(gr->scrubber.vmm, KGRAPHICS_SCRUBBER_HANDLE_VAS, false); + if (ret) + goto done; + + ret = rm->api->fifo->chan.alloc(&gr->scrubber.vmm->rm.device, KGRAPHICS_SCRUBBER_HANDLE_CHANNEL, + NV2080_ENGINE_TYPE_GR0, 0, false, gr->scrubber.chid, + nvkm_memory_addr(gr->scrubber.inst), + nvkm_memory_addr(gr->scrubber.inst) + 0x1000, + nvkm_memory_addr(gr->scrubber.inst) + 0x2000, + gr->scrubber.vmm, 0, 0x1000, &gr->scrubber.chan); + if (ret) + goto done; + + ret = r535_gr_promote_ctx(gr, false, gr->scrubber.vmm, gr->scrubber.ctxbuf.mem, + gr->scrubber.ctxbuf.vma, &gr->scrubber.chan); + if (ret) + goto done; + + ret = nvkm_gsp_rm_alloc(&gr->scrubber.chan, KGRAPHICS_SCRUBBER_HANDLE_3DOBJ, + rm->gpu->gr.class.threed, 0, &gr->scrubber.threed); + if (ret) + goto done; + + /* Initialise scrubber channel on RM. */ + ret = r570_gr_scrubber_ctrl(gr, false); + if (ret) + goto done; + + gr->scrubber.enabled = true; + +done: + if (ret) + r570_gr_scrubber_fini(gr); + + return ret; +} + +static int +r570_gr_get_ctxbufs_info(struct r535_gr *gr) +{ + NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS *info; + struct nvkm_subdev *subdev = &gr->base.engine.subdev; + struct nvkm_gsp *gsp = subdev->device->gsp; + + info = nvkm_gsp_rm_ctrl_rd(&gsp->internal.device.subdevice, + NV2080_CTRL_CMD_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO, + sizeof(*info)); + if (WARN_ON(IS_ERR(info))) + return PTR_ERR(info); + + for (int i = 0; i < ARRAY_SIZE(info->engineContextBuffersInfo[0].engine); i++) + r535_gr_get_ctxbuf_info(gr, i, &info->engineContextBuffersInfo[0].engine[i]); + + nvkm_gsp_rm_ctrl_done(&gsp->internal.device.subdevice, info); + return 0; +} + +const struct nvkm_rm_api_gr +r570_gr = { + .get_ctxbufs_info = r570_gr_get_ctxbufs_info, + .scrubber.init = r570_gr_scrubber_init, + .scrubber.fini = r570_gr_scrubber_fini, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/gsp.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/gsp.c new file mode 100644 index 000000000000..9d2fa4e66d59 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/gsp.c @@ -0,0 +1,216 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/rm.h> +#include <rm/rpc.h> + +#include <asm-generic/video.h> + +#include "nvrm/gsp.h" +#include "nvrm/rpcfn.h" +#include "nvrm/msgfn.h" + +#include <core/pci.h> +#include <subdev/pci/priv.h> + +static u32 +r570_gsp_sr_data_size(struct nvkm_gsp *gsp) +{ + GspFwWprMeta *meta = gsp->wpr_meta.data; + + return (meta->frtsOffset + meta->frtsSize) - + (meta->nonWprHeapOffset + meta->nonWprHeapSize); +} + +static void +r570_gsp_drop_post_nocat_record(struct nvkm_gsp *gsp) +{ + if (gsp->subdev.debug < NV_DBG_DEBUG) { + r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_GSP_POST_NOCAT_RECORD, NULL, NULL); + r535_gsp_msg_ntfy_add(gsp, NV_VGPU_MSG_EVENT_GSP_LOCKDOWN_NOTICE, NULL, NULL); + } +} + +static bool +r570_gsp_xlat_mc_engine_idx(u32 mc_engine_idx, enum nvkm_subdev_type *ptype, int *pinst) +{ + switch (mc_engine_idx) { + case MC_ENGINE_IDX_GSP: + *ptype = NVKM_SUBDEV_GSP; + *pinst = 0; + return true; + case MC_ENGINE_IDX_DISP: + *ptype = NVKM_ENGINE_DISP; + *pinst = 0; + return true; + case MC_ENGINE_IDX_CE0 ... MC_ENGINE_IDX_CE19: + *ptype = NVKM_ENGINE_CE; + *pinst = mc_engine_idx - MC_ENGINE_IDX_CE0; + return true; + case MC_ENGINE_IDX_GR0: + *ptype = NVKM_ENGINE_GR; + *pinst = 0; + return true; + case MC_ENGINE_IDX_NVDEC0 ... MC_ENGINE_IDX_NVDEC7: + *ptype = NVKM_ENGINE_NVDEC; + *pinst = mc_engine_idx - MC_ENGINE_IDX_NVDEC0; + return true; + case MC_ENGINE_IDX_NVENC ... MC_ENGINE_IDX_NVENC3: + *ptype = NVKM_ENGINE_NVENC; + *pinst = mc_engine_idx - MC_ENGINE_IDX_NVENC; + return true; + case MC_ENGINE_IDX_NVJPEG0 ... MC_ENGINE_IDX_NVJPEG7: + *ptype = NVKM_ENGINE_NVJPG; + *pinst = mc_engine_idx - MC_ENGINE_IDX_NVJPEG0; + return true; + case MC_ENGINE_IDX_OFA0 ... MC_ENGINE_IDX_OFA1: + *ptype = NVKM_ENGINE_OFA; + *pinst = mc_engine_idx - MC_ENGINE_IDX_OFA0; + return true; + default: + return false; + } +} + +static int +r570_gsp_get_static_info(struct nvkm_gsp *gsp) +{ + GspStaticConfigInfo *rpc; + u32 gpc_mask; + u32 tpc_mask; + int ret; + + rpc = nvkm_gsp_rpc_rd(gsp, NV_VGPU_MSG_FUNCTION_GET_GSP_STATIC_INFO, sizeof(*rpc)); + if (IS_ERR(rpc)) + return PTR_ERR(rpc); + + gsp->internal.client.object.client = &gsp->internal.client; + gsp->internal.client.object.parent = NULL; + gsp->internal.client.object.handle = rpc->hInternalClient; + gsp->internal.client.gsp = gsp; + INIT_LIST_HEAD(&gsp->internal.client.events); + + gsp->internal.device.object.client = &gsp->internal.client; + gsp->internal.device.object.parent = &gsp->internal.client.object; + gsp->internal.device.object.handle = rpc->hInternalDevice; + + gsp->internal.device.subdevice.client = &gsp->internal.client; + gsp->internal.device.subdevice.parent = &gsp->internal.device.object; + gsp->internal.device.subdevice.handle = rpc->hInternalSubdevice; + + gsp->bar.rm_bar1_pdb = rpc->bar1PdeBase; + gsp->bar.rm_bar2_pdb = rpc->bar2PdeBase; + + r535_gsp_get_static_info_fb(gsp, &rpc->fbRegionInfoParams); + + if (gsp->rm->wpr->offset_set_by_acr) { + GspFwWprMeta *meta = gsp->wpr_meta.data; + + meta->nonWprHeapOffset = rpc->fwWprLayoutOffset.nonWprHeapOffset; + meta->frtsOffset = rpc->fwWprLayoutOffset.frtsOffset; + } + + nvkm_gsp_rpc_done(gsp, rpc); + + ret = r570_gr_gpc_mask(gsp, &gpc_mask); + if (ret) + return ret; + + for (int gpc = 0; gpc < 32; gpc++) { + if (gpc_mask & BIT(gpc)) { + ret = r570_gr_tpc_mask(gsp, gpc, &tpc_mask); + if (ret) + return ret; + + gsp->gr.tpcs += hweight32(tpc_mask); + gsp->gr.gpcs++; + } + } + + return 0; +} + +static void +r570_gsp_acpi_info(struct nvkm_gsp *gsp, ACPI_METHOD_DATA *acpi) +{ +#if defined(CONFIG_ACPI) && defined(CONFIG_X86) + acpi_handle handle = ACPI_HANDLE(gsp->subdev.device->dev); + + if (!handle) + return; + + acpi->bValid = 1; + + r535_gsp_acpi_dod(handle, &acpi->dodMethodData); + r535_gsp_acpi_jt(handle, &acpi->jtMethodData); + r535_gsp_acpi_caps(handle, &acpi->capsMethodData); +#endif +} + +static int +r570_gsp_set_system_info(struct nvkm_gsp *gsp) +{ + struct nvkm_device *device = gsp->subdev.device; + struct pci_dev *pdev = container_of(device, struct nvkm_device_pci, device)->pdev; + GspSystemInfo *info; + + if (WARN_ON(device->type == NVKM_DEVICE_TEGRA)) + return -ENOSYS; + + info = nvkm_gsp_rpc_get(gsp, NV_VGPU_MSG_FUNCTION_GSP_SET_SYSTEM_INFO, sizeof(*info)); + if (IS_ERR(info)) + return PTR_ERR(info); + + info->gpuPhysAddr = device->func->resource_addr(device, NVKM_BAR0_PRI); + info->gpuPhysFbAddr = device->func->resource_addr(device, NVKM_BAR1_FB); + info->gpuPhysInstAddr = device->func->resource_addr(device, NVKM_BAR2_INST); + info->nvDomainBusDeviceFunc = pci_dev_id(pdev); + info->maxUserVa = TASK_SIZE; + info->pciConfigMirrorBase = device->pci->func->cfg.addr; + info->pciConfigMirrorSize = device->pci->func->cfg.size; + info->PCIDeviceID = (pdev->device << 16) | pdev->vendor; + info->PCISubDeviceID = (pdev->subsystem_device << 16) | pdev->subsystem_vendor; + info->PCIRevisionID = pdev->revision; + r570_gsp_acpi_info(gsp, &info->acpiMethodData); + info->bIsPrimary = video_is_primary_device(device->dev); + info->bPreserveVideoMemoryAllocations = false; + + return nvkm_gsp_rpc_wr(gsp, info, NVKM_GSP_RPC_REPLY_NOWAIT); +} + +static void +r570_gsp_set_rmargs(struct nvkm_gsp *gsp, bool resume) +{ + GSP_ARGUMENTS_CACHED *args; + + args = gsp->rmargs.data; + args->messageQueueInitArguments.sharedMemPhysAddr = gsp->shm.mem.addr; + args->messageQueueInitArguments.pageTableEntryCount = gsp->shm.ptes.nr; + args->messageQueueInitArguments.cmdQueueOffset = + (u8 *)gsp->shm.cmdq.ptr - (u8 *)gsp->shm.mem.data; + args->messageQueueInitArguments.statQueueOffset = + (u8 *)gsp->shm.msgq.ptr - (u8 *)gsp->shm.mem.data; + + if (!resume) { + args->srInitArguments.oldLevel = 0; + args->srInitArguments.flags = 0; + args->srInitArguments.bInPMTransition = 0; + } else { + args->srInitArguments.oldLevel = NV2080_CTRL_GPU_SET_POWER_STATE_GPU_LEVEL_3; + args->srInitArguments.flags = 0; + args->srInitArguments.bInPMTransition = 1; + } + + args->bDmemStack = 1; +} + +const struct nvkm_rm_api_gsp +r570_gsp = { + .set_rmargs = r570_gsp_set_rmargs, + .set_system_info = r570_gsp_set_system_info, + .get_static_info = r570_gsp_get_static_info, + .xlat_mc_engine_idx = r570_gsp_xlat_mc_engine_idx, + .drop_post_nocat_record = r570_gsp_drop_post_nocat_record, + .sr_data_size = r570_gsp_sr_data_size, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/client.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/client.h new file mode 100644 index 000000000000..e8714e0abc37 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/client.h @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2024, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_CLIENT_H__ +#define __NVRM_CLIENT_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define NV01_ROOT (0x0U) /* finn: Evaluated from "NV0000_ALLOC_PARAMETERS_MESSAGE_ID" */ + +#define NV_PROC_NAME_MAX_LENGTH 100U + +typedef struct NV0000_ALLOC_PARAMETERS { + NvHandle hClient; /* CORERM-2934: hClient must remain the first member until all allocations use these params */ + NvU32 processID; + char processName[NV_PROC_NAME_MAX_LENGTH]; + NV_DECLARE_ALIGNED(NvP64 pOsPidInfo, 8); +} NV0000_ALLOC_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/disp.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/disp.h new file mode 100644 index 000000000000..06e972835d77 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/disp.h @@ -0,0 +1,355 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_DISP_H__ +#define __NVRM_DISP_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_GET_STATIC_INFO (0x20800a01) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS { + NvU32 feHwSysCap; + NvU32 windowPresentMask; + NvBool bFbRemapperEnabled; + NvU32 numHeads; + NvU32 i2cPort; + NvU32 internalDispActiveMask; + NvU32 embeddedDisplayPortMask; + NvBool bExternalMuxSupported; + NvBool bInternalMuxSupported; + NvU32 numDispChannels; +} NV2080_CTRL_INTERNAL_DISPLAY_GET_STATIC_INFO_PARAMS; + +#define NV0073_CTRL_CMD_SYSTEM_GET_SUPPORTED (0x730107U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayMask; + NvU32 displayMaskDDC; +} NV0073_CTRL_SYSTEM_GET_SUPPORTED_PARAMS; + +#define NV0073_CTRL_MAX_CONNECTORS 4U + +#define NV0073_CTRL_CMD_SPECIFIC_GET_CONNECTOR_DATA (0x730250U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 flags; + NvU32 DDCPartners; + NvU32 count; + struct { + NvU32 index; + NvU32 type; + NvU32 location; + } data[NV0073_CTRL_MAX_CONNECTORS]; + NvU32 platform; +} NV0073_CTRL_SPECIFIC_GET_CONNECTOR_DATA_PARAMS; + +#define NV0073_CTRL_CMD_DP_GET_CAPS (0x731369U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_CMD_DSC_CAP_PARAMS { + NvBool bDscSupported; + NvU32 encoderColorFormatMask; + NvU32 lineBufferSizeKB; + NvU32 rateBufferSizeKB; + NvU32 bitsPerPixelPrecision; + NvU32 maxNumHztSlices; + NvU32 lineBufferBitDepth; +} NV0073_CTRL_CMD_DSC_CAP_PARAMS; + +typedef struct NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS { + NvU32 subDeviceInstance; + NvU32 sorIndex; + NvU32 maxLinkRate; + NvU32 dpVersionsSupported; + NvU32 UHBRSupportedByGpu; + NvU32 minPClkForCompressed; + NvBool bIsMultistreamSupported; + NvBool bIsSCEnabled; + NvBool bHasIncreasedWatermarkLimits; + NvBool bIsPC2Disabled; + NvBool isSingleHeadMSTSupported; + NvBool bFECSupported; + NvBool bIsTrainPhyRepeater; + NvBool bOverrideLinkBw; + NvBool bUseRgFlushSequence; + NvBool bSupportDPDownSpread; + NV0073_CTRL_CMD_DSC_CAP_PARAMS DSC; +} NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS; + +#define NV0073_CTRL_CMD_DP_GET_CAPS (0x731369U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID" */ +#define NV0073_CTRL_CMD_DP_GET_CAPS_PARAMS_MESSAGE_ID (0x69U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2 0:0 +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_2_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4 1:1 +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP1_4_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP2_0 2:2 +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP2_0_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DP_VERSIONS_SUPPORTED_DP2_0_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE 2:0 +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_NONE (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_1_62 (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_2_70 (0x00000002U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_5_40 (0x00000003U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_MAX_LINK_RATE_8_10 (0x00000004U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR10_0 0:0 +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR10_0_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR10_0_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR13_5 1:1 +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR13_5_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR13_5_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR20_0 2:2 +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR20_0_NO (0x00000000U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_UHBR_SUPPORTED_UHBR20_0_YES (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_RGB (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_444 (0x00000002U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_NATIVE_422 (0x00000004U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_ENCODER_COLOR_FORMAT_Y_CB_CR_NATIVE_420 (0x00000008U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_16 (0x00000001U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_8 (0x00000002U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_4 (0x00000003U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1_2 (0x00000004U) +#define NV0073_CTRL_CMD_DP_GET_CAPS_DSC_BITS_PER_PIXEL_PRECISION_1 (0x00000005U) + +#define NV0073_CTRL_CMD_SYSTEM_GET_CONNECT_STATE (0x730108U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS { + NvU32 subDeviceInstance; + NvU32 flags; + NvU32 displayMask; + NvU32 retryTimeMs; +} NV0073_CTRL_SYSTEM_GET_CONNECT_STATE_PARAMS; + +#define NV0073_CTRL_CMD_DFP_GET_INFO (0x731140U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DFP_INTERFACE_ID << 8) | NV0073_CTRL_DFP_GET_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DFP_GET_INFO_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 flags; + NvU32 UHBRSupportedByDfp; +} NV0073_CTRL_DFP_GET_INFO_PARAMS; + +#define NV0073_CTRL_DFP_FLAGS_SIGNAL 2:0 +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_TMDS (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_LVDS (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_SDI (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_DISPLAYPORT (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_DSI (0x00000004U) +#define NV0073_CTRL_DFP_FLAGS_SIGNAL_WRBK (0x00000005U) +#define NV0073_CTRL_DFP_FLAGS_LANE 5:3 +#define NV0073_CTRL_DFP_FLAGS_LANE_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_LANE_SINGLE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_LANE_DUAL (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_LANE_QUAD (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_LANE_OCT (0x00000004U) +#define NV0073_CTRL_DFP_FLAGS_LIMIT 6:6 +#define NV0073_CTRL_DFP_FLAGS_LIMIT_DISABLE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_LIMIT_60HZ_RR (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER 7:7 +#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER_NORMAL (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_SLI_SCALER_DISABLE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE 8:8 +#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE 9:9 +#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_LIMITED_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE 10:10 +#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_RANGE_AUTO_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE 11:11 +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR422_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE 12:12 +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_FORMAT_YCBCR444_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_TYPE_C_TO_DP_CONNECTOR 13:13 +#define NV0073_CTRL_DFP_FLAGS_TYPE_C_TO_DP_CONNECTOR_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_TYPE_C_TO_DP_CONNECTOR_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED 14:14 +#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_HDMI_ALLOWED_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT 15:15 +#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_EMBEDDED_DISPLAYPORT_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT 16:16 +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_CONSTRAINT_PREFER_RBR (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW 19:17 +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_1_62GBPS (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_2_70GBPS (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_5_40GBPS (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_DP_LINK_BW_8_10GBPS (0x00000004U) +#define NV0073_CTRL_DFP_FLAGS_LINK 21:20 +#define NV0073_CTRL_DFP_FLAGS_LINK_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_LINK_SINGLE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_LINK_DUAL (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID 22:22 +#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DP_FORCE_RM_EDID_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID 24:23 +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_NONE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_A (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_B (0x00000002U) +#define NV0073_CTRL_DFP_FLAGS_DSI_DEVICE_ID_DSI_GANGED (0x00000003U) +#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED 25:25 +#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DP_POST_CURSOR2_DISABLED_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS_DP_PHY_REPEATER_COUNT 29:26 +#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE 30:30 +#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS_DYNAMIC_MUX_CAPABLE_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_10_0GBPS 0:0 +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_10_0GBPS_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_10_0GBPS_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_13_5GBPS 1:1 +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_13_5GBPS_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_13_5GBPS_TRUE (0x00000001U) +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_20_0GBPS 2:2 +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_20_0GBPS_FALSE (0x00000000U) +#define NV0073_CTRL_DFP_FLAGS2_DP_UHBR_SUPPORTED_20_0GBPS_TRUE (0x00000001U) + +#define NV0073_CTRL_CMD_SYSTEM_GET_ACTIVE (0x73010cU) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SYSTEM_INTERFACE_ID << 8) | NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS { + NvU32 subDeviceInstance; + NvU32 head; + NvU32 flags; + NvU32 displayId; +} NV0073_CTRL_SYSTEM_GET_ACTIVE_PARAMS; + +#define NV0073_CTRL_CMD_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS (0x730292U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_SET_BACKLIGHT_BRIGHTNESS_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_CMD_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS (0x730291U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_SPECIFIC_INTERFACE_ID << 8) | NV0073_CTRL_SPECIFIC_GET_BACKLIGHT_BRIGHTNESS_PARAMS_MESSAGE_ID" */ + +typedef struct NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 brightness; + NvBool bUncalibrated; + NvU8 brightnessType; +} NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_PARAMS; + +#define NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES (0x731377U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS_MESSAGE_ID" */ + +#define NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES 8U + +typedef struct NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS { + // In + NvU32 subDeviceInstance; + NvU32 displayId; + NvU16 linkRateTbl[NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES]; + + // Out + NvU16 linkBwTbl[NV0073_CTRL_DP_MAX_INDEXED_LINK_RATES]; + NvU8 linkBwCount; +} NV0073_CTRL_CMD_DP_CONFIG_INDEXED_LINK_RATES_PARAMS; + +#define NV0073_CTRL_CMD_DP_CTRL (0x731343U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_CTRL_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DP_CTRL_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 cmd; + NvU32 data; + NvU32 err; + NvU32 retryTimeMs; + NvU32 eightLaneDpcdBaseAddr; +} NV0073_CTRL_DP_CTRL_PARAMS; + +typedef struct NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS { + NvU32 subDeviceInstance; + NvU32 head; + NvU32 sorIndex; + NvU32 dpLink; + + NvBool bEnableOverride; + NvBool bMST; + NvU32 singleHeadMultistreamMode; + NvU32 hBlankSym; + NvU32 vBlankSym; + NvU32 colorFormat; + NvBool bEnableTwoHeadOneOr; + + struct { + NvU32 slotStart; + NvU32 slotEnd; + NvU32 PBN; + NvU32 Timeslice; + NvBool sendACT; // deprecated -Use NV0073_CTRL_CMD_DP_SEND_ACT + NvU32 singleHeadMSTPipeline; + NvBool bEnableAudioOverRightPanel; + } MST; + + struct { + NvBool bEnhancedFraming; + NvU32 tuSize; + NvU32 waterMark; + NvBool bEnableAudioOverRightPanel; + } SST; +} NV0073_CTRL_CMD_DP_CONFIG_STREAM_PARAMS; + +#define NV0073_CTRL_CMD_DP_SET_AUDIO_MUTESTREAM (0x731359U) /* finn: Evaluated from "(FINN_NV04_DISPLAY_COMMON_DP_INTERFACE_ID << 8) | NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS_MESSAGE_ID" */ +typedef struct NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS { + NvU32 subDeviceInstance; + NvU32 displayId; + NvU32 mute; +} NV0073_CTRL_DP_SET_AUDIO_MUTESTREAM_PARAMS; + +#define NV2080_CTRL_CMD_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER (0x20800a58) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS { + NvU32 addressSpace; + NV_DECLARE_ALIGNED(NvU64 physicalAddr, 8); + NV_DECLARE_ALIGNED(NvU64 limit, 8); + NvU32 cacheSnoop; + NvU32 hclass; + NvU32 channelInstance; + NvBool valid; + NvU32 pbTargetAperture; + NvU32 channelPBSize; + NvU32 subDeviceId; +} NV2080_CTRL_INTERNAL_DISPLAY_CHANNEL_PUSHBUFFER_PARAMS; + +#define ADDR_SYSMEM (1) + +#define ADDR_FBMEM 2 // Frame buffer memory space + +typedef enum +{ + PB_SIZE_4KB = 0, + PB_SIZE_8KB, + PB_SIZE_16KB, + PB_SIZE_32KB, + PB_SIZE_64KB +} ChannelPBSize; + +typedef struct +{ + NvV32 channelInstance; // One of the n channel instances of a given channel type. + // Note that core channel has only one instance + // while all others have two (one per head). + NvHandle hObjectBuffer; // ctx dma handle for DMA push buffer + NvHandle hObjectNotify; // ctx dma handle for an area (of type NvNotification defined in sdk/nvidia/inc/nvtypes.h) where RM can write errors/notifications + NvU32 offset; // Initial offset for put/get, usually zero. + NvP64 pControl NV_ALIGN_BYTES(8); // pControl gives virt addr of UDISP GET/PUT regs + + NvU32 flags; + ChannelPBSize channelPBSize; // Size of Push Buffer requested by client (allowed values in enum) +#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB 1:1 +#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB_YES 0x00000000 +#define NV50VAIO_CHANNELDMA_ALLOCATION_FLAGS_CONNECT_PB_AT_GRAB_NO 0x00000001 + + NvU32 subDeviceId; // One-hot encoded subDeviceId (i.e. SDM) that will be used to address the channel in the pushbuffer stream (via SSDM method) +} NV50VAIO_CHANNELDMA_ALLOCATION_PARAMETERS; + +#define NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_TYPE_PERCENT100 1 +#define NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_TYPE_PERCENT1000 2 +#define NV0073_CTRL_SPECIFIC_BACKLIGHT_BRIGHTNESS_TYPE_NITS 3 + +typedef enum +{ + IOVA, + PHYS_NVM, + PHYS_PCI, + PHYS_PCI_COHERENT +} PBTARGETAPERTURE; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/engine.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/engine.h new file mode 100644 index 000000000000..7997050a4f29 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/engine.h @@ -0,0 +1,318 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_ENGINE_H__ +#define __NVRM_ENGINE_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define MC_ENGINE_IDX_NULL 0 // This must be 0 +#define MC_ENGINE_IDX_TMR 1 +#define MC_ENGINE_IDX_DISP 2 +#define MC_ENGINE_IDX_FB 3 +#define MC_ENGINE_IDX_FIFO 4 +#define MC_ENGINE_IDX_VIDEO 5 +#define MC_ENGINE_IDX_MD 6 +#define MC_ENGINE_IDX_BUS 7 +#define MC_ENGINE_IDX_PMGR 8 +#define MC_ENGINE_IDX_VP2 9 +#define MC_ENGINE_IDX_CIPHER 10 +#define MC_ENGINE_IDX_BIF 11 +#define MC_ENGINE_IDX_PPP 12 +#define MC_ENGINE_IDX_PRIVRING 13 +#define MC_ENGINE_IDX_PMU 14 +#define MC_ENGINE_IDX_CE0 15 +#define MC_ENGINE_IDX_CE1 16 +#define MC_ENGINE_IDX_CE2 17 +#define MC_ENGINE_IDX_CE3 18 +#define MC_ENGINE_IDX_CE4 19 +#define MC_ENGINE_IDX_CE5 20 +#define MC_ENGINE_IDX_CE6 21 +#define MC_ENGINE_IDX_CE7 22 +#define MC_ENGINE_IDX_CE8 23 +#define MC_ENGINE_IDX_CE9 24 +#define MC_ENGINE_IDX_CE10 25 +#define MC_ENGINE_IDX_CE11 26 +#define MC_ENGINE_IDX_CE12 27 +#define MC_ENGINE_IDX_CE13 28 +#define MC_ENGINE_IDX_CE14 29 +#define MC_ENGINE_IDX_CE15 30 +#define MC_ENGINE_IDX_CE16 31 +#define MC_ENGINE_IDX_CE17 32 +#define MC_ENGINE_IDX_CE18 33 +#define MC_ENGINE_IDX_CE19 34 +#define MC_ENGINE_IDX_CE_MAX MC_ENGINE_IDX_CE19 +#define MC_ENGINE_IDX_VIC 35 +#define MC_ENGINE_IDX_ISOHUB 36 +#define MC_ENGINE_IDX_VGPU 37 +#define MC_ENGINE_IDX_NVENC 38 +#define MC_ENGINE_IDX_NVENC1 39 +#define MC_ENGINE_IDX_NVENC2 40 +#define MC_ENGINE_IDX_NVENC3 41 +#define MC_ENGINE_IDX_C2C 42 +#define MC_ENGINE_IDX_LTC 43 +#define MC_ENGINE_IDX_FBHUB 44 +#define MC_ENGINE_IDX_HDACODEC 45 +#define MC_ENGINE_IDX_GMMU 46 +#define MC_ENGINE_IDX_SEC2 47 +#define MC_ENGINE_IDX_FSP 48 +#define MC_ENGINE_IDX_NVLINK 49 +#define MC_ENGINE_IDX_GSP 50 +#define MC_ENGINE_IDX_NVJPG 51 +#define MC_ENGINE_IDX_NVJPEG MC_ENGINE_IDX_NVJPG +#define MC_ENGINE_IDX_NVJPEG0 MC_ENGINE_IDX_NVJPEG +#define MC_ENGINE_IDX_NVJPEG1 52 +#define MC_ENGINE_IDX_NVJPEG2 53 +#define MC_ENGINE_IDX_NVJPEG3 54 +#define MC_ENGINE_IDX_NVJPEG4 55 +#define MC_ENGINE_IDX_NVJPEG5 56 +#define MC_ENGINE_IDX_NVJPEG6 57 +#define MC_ENGINE_IDX_NVJPEG7 58 +#define MC_ENGINE_IDX_REPLAYABLE_FAULT 59 +#define MC_ENGINE_IDX_ACCESS_CNTR 60 +#define MC_ENGINE_IDX_NON_REPLAYABLE_FAULT 61 +#define MC_ENGINE_IDX_REPLAYABLE_FAULT_ERROR 62 +#define MC_ENGINE_IDX_NON_REPLAYABLE_FAULT_ERROR 63 +#define MC_ENGINE_IDX_INFO_FAULT 64 +#define MC_ENGINE_IDX_BSP 65 +#define MC_ENGINE_IDX_NVDEC MC_ENGINE_IDX_BSP +#define MC_ENGINE_IDX_NVDEC0 MC_ENGINE_IDX_NVDEC +#define MC_ENGINE_IDX_NVDEC1 66 +#define MC_ENGINE_IDX_NVDEC2 67 +#define MC_ENGINE_IDX_NVDEC3 68 +#define MC_ENGINE_IDX_NVDEC4 69 +#define MC_ENGINE_IDX_NVDEC5 70 +#define MC_ENGINE_IDX_NVDEC6 71 +#define MC_ENGINE_IDX_NVDEC7 72 +#define MC_ENGINE_IDX_CPU_DOORBELL 73 +#define MC_ENGINE_IDX_PRIV_DOORBELL 74 +#define MC_ENGINE_IDX_MMU_ECC_ERROR 75 +#define MC_ENGINE_IDX_BLG 76 +#define MC_ENGINE_IDX_PERFMON 77 +#define MC_ENGINE_IDX_BUF_RESET 78 +#define MC_ENGINE_IDX_XBAR 79 +#define MC_ENGINE_IDX_ZPW 80 +#define MC_ENGINE_IDX_OFA0 81 +#define MC_ENGINE_IDX_OFA1 82 +#define MC_ENGINE_IDX_TEGRA 83 +#define MC_ENGINE_IDX_GR 84 +#define MC_ENGINE_IDX_GR0 MC_ENGINE_IDX_GR +#define MC_ENGINE_IDX_GR1 85 +#define MC_ENGINE_IDX_GR2 86 +#define MC_ENGINE_IDX_GR3 87 +#define MC_ENGINE_IDX_GR4 88 +#define MC_ENGINE_IDX_GR5 89 +#define MC_ENGINE_IDX_GR6 90 +#define MC_ENGINE_IDX_GR7 91 +#define MC_ENGINE_IDX_ESCHED 92 +#define MC_ENGINE_IDX_ESCHED__SIZE 64 +#define MC_ENGINE_IDX_GR_FECS_LOG 156 +#define MC_ENGINE_IDX_GR0_FECS_LOG MC_ENGINE_IDX_GR_FECS_LOG +#define MC_ENGINE_IDX_GR1_FECS_LOG 157 +#define MC_ENGINE_IDX_GR2_FECS_LOG 158 +#define MC_ENGINE_IDX_GR3_FECS_LOG 159 +#define MC_ENGINE_IDX_GR4_FECS_LOG 160 +#define MC_ENGINE_IDX_GR5_FECS_LOG 161 +#define MC_ENGINE_IDX_GR6_FECS_LOG 162 +#define MC_ENGINE_IDX_GR7_FECS_LOG 163 +#define MC_ENGINE_IDX_TMR_SWRL 164 +#define MC_ENGINE_IDX_DISP_GSP 165 +#define MC_ENGINE_IDX_REPLAYABLE_FAULT_CPU 166 +#define MC_ENGINE_IDX_NON_REPLAYABLE_FAULT_CPU 167 +#define MC_ENGINE_IDX_PXUC 168 +#define MC_ENGINE_IDX_SYSLTC 169 +#define MC_ENGINE_IDX_LRCC 170 +#define MC_ENGINE_IDX_GSPLITE 171 +#define MC_ENGINE_IDX_GSPLITE0 MC_ENGINE_IDX_GSPLITE +#define MC_ENGINE_IDX_GSPLITE1 172 +#define MC_ENGINE_IDX_GSPLITE2 173 +#define MC_ENGINE_IDX_GSPLITE3 174 +#define MC_ENGINE_IDX_GSPLITE_MAX MC_ENGINE_IDX_GSPLITE3 +#define MC_ENGINE_IDX_DPAUX 175 +#define MC_ENGINE_IDX_DISP_LOW 176 +#define MC_ENGINE_IDX_MAX 177 + +typedef enum +{ + RM_ENGINE_TYPE_NULL = (0x00000000), + RM_ENGINE_TYPE_GR0 = (0x00000001), + RM_ENGINE_TYPE_GR1 = (0x00000002), + RM_ENGINE_TYPE_GR2 = (0x00000003), + RM_ENGINE_TYPE_GR3 = (0x00000004), + RM_ENGINE_TYPE_GR4 = (0x00000005), + RM_ENGINE_TYPE_GR5 = (0x00000006), + RM_ENGINE_TYPE_GR6 = (0x00000007), + RM_ENGINE_TYPE_GR7 = (0x00000008), + RM_ENGINE_TYPE_COPY0 = (0x00000009), + RM_ENGINE_TYPE_COPY1 = (0x0000000a), + RM_ENGINE_TYPE_COPY2 = (0x0000000b), + RM_ENGINE_TYPE_COPY3 = (0x0000000c), + RM_ENGINE_TYPE_COPY4 = (0x0000000d), + RM_ENGINE_TYPE_COPY5 = (0x0000000e), + RM_ENGINE_TYPE_COPY6 = (0x0000000f), + RM_ENGINE_TYPE_COPY7 = (0x00000010), + RM_ENGINE_TYPE_COPY8 = (0x00000011), + RM_ENGINE_TYPE_COPY9 = (0x00000012), + RM_ENGINE_TYPE_COPY10 = (0x00000013), + RM_ENGINE_TYPE_COPY11 = (0x00000014), + RM_ENGINE_TYPE_COPY12 = (0x00000015), + RM_ENGINE_TYPE_COPY13 = (0x00000016), + RM_ENGINE_TYPE_COPY14 = (0x00000017), + RM_ENGINE_TYPE_COPY15 = (0x00000018), + RM_ENGINE_TYPE_COPY16 = (0x00000019), + RM_ENGINE_TYPE_COPY17 = (0x0000001a), + RM_ENGINE_TYPE_COPY18 = (0x0000001b), + RM_ENGINE_TYPE_COPY19 = (0x0000001c), + RM_ENGINE_TYPE_NVDEC0 = (0x0000001d), + RM_ENGINE_TYPE_NVDEC1 = (0x0000001e), + RM_ENGINE_TYPE_NVDEC2 = (0x0000001f), + RM_ENGINE_TYPE_NVDEC3 = (0x00000020), + RM_ENGINE_TYPE_NVDEC4 = (0x00000021), + RM_ENGINE_TYPE_NVDEC5 = (0x00000022), + RM_ENGINE_TYPE_NVDEC6 = (0x00000023), + RM_ENGINE_TYPE_NVDEC7 = (0x00000024), + RM_ENGINE_TYPE_NVENC0 = (0x00000025), + RM_ENGINE_TYPE_NVENC1 = (0x00000026), + RM_ENGINE_TYPE_NVENC2 = (0x00000027), + // Bug 4175886 - Use this new value for all chips once GB20X is released + RM_ENGINE_TYPE_NVENC3 = (0x00000028), + RM_ENGINE_TYPE_VP = (0x00000029), + RM_ENGINE_TYPE_ME = (0x0000002a), + RM_ENGINE_TYPE_PPP = (0x0000002b), + RM_ENGINE_TYPE_MPEG = (0x0000002c), + RM_ENGINE_TYPE_SW = (0x0000002d), + RM_ENGINE_TYPE_TSEC = (0x0000002e), + RM_ENGINE_TYPE_VIC = (0x0000002f), + RM_ENGINE_TYPE_MP = (0x00000030), + RM_ENGINE_TYPE_SEC2 = (0x00000031), + RM_ENGINE_TYPE_HOST = (0x00000032), + RM_ENGINE_TYPE_DPU = (0x00000033), + RM_ENGINE_TYPE_PMU = (0x00000034), + RM_ENGINE_TYPE_FBFLCN = (0x00000035), + RM_ENGINE_TYPE_NVJPEG0 = (0x00000036), + RM_ENGINE_TYPE_NVJPEG1 = (0x00000037), + RM_ENGINE_TYPE_NVJPEG2 = (0x00000038), + RM_ENGINE_TYPE_NVJPEG3 = (0x00000039), + RM_ENGINE_TYPE_NVJPEG4 = (0x0000003a), + RM_ENGINE_TYPE_NVJPEG5 = (0x0000003b), + RM_ENGINE_TYPE_NVJPEG6 = (0x0000003c), + RM_ENGINE_TYPE_NVJPEG7 = (0x0000003d), + RM_ENGINE_TYPE_OFA0 = (0x0000003e), + RM_ENGINE_TYPE_OFA1 = (0x0000003f), + RM_ENGINE_TYPE_RESERVED40 = (0x00000040), + RM_ENGINE_TYPE_RESERVED41 = (0x00000041), + RM_ENGINE_TYPE_RESERVED42 = (0x00000042), + RM_ENGINE_TYPE_RESERVED43 = (0x00000043), + RM_ENGINE_TYPE_RESERVED44 = (0x00000044), + RM_ENGINE_TYPE_RESERVED45 = (0x00000045), + RM_ENGINE_TYPE_RESERVED46 = (0x00000046), + RM_ENGINE_TYPE_RESERVED47 = (0x00000047), + RM_ENGINE_TYPE_RESERVED48 = (0x00000048), + RM_ENGINE_TYPE_RESERVED49 = (0x00000049), + RM_ENGINE_TYPE_RESERVED4a = (0x0000004a), + RM_ENGINE_TYPE_RESERVED4b = (0x0000004b), + RM_ENGINE_TYPE_RESERVED4c = (0x0000004c), + RM_ENGINE_TYPE_RESERVED4d = (0x0000004d), + RM_ENGINE_TYPE_RESERVED4e = (0x0000004e), + RM_ENGINE_TYPE_RESERVED4f = (0x0000004f), + RM_ENGINE_TYPE_RESERVED50 = (0x00000050), + RM_ENGINE_TYPE_RESERVED51 = (0x00000051), + RM_ENGINE_TYPE_RESERVED52 = (0x00000052), + RM_ENGINE_TYPE_RESERVED53 = (0x00000053), + RM_ENGINE_TYPE_LAST = (0x00000054), +} RM_ENGINE_TYPE; + +#define NV2080_ENGINE_TYPE_NULL (0x00000000) +#define NV2080_ENGINE_TYPE_GRAPHICS (0x00000001) +#define NV2080_ENGINE_TYPE_GR0 NV2080_ENGINE_TYPE_GRAPHICS +#define NV2080_ENGINE_TYPE_GR1 (0x00000002) +#define NV2080_ENGINE_TYPE_GR2 (0x00000003) +#define NV2080_ENGINE_TYPE_GR3 (0x00000004) +#define NV2080_ENGINE_TYPE_GR4 (0x00000005) +#define NV2080_ENGINE_TYPE_GR5 (0x00000006) +#define NV2080_ENGINE_TYPE_GR6 (0x00000007) +#define NV2080_ENGINE_TYPE_GR7 (0x00000008) +#define NV2080_ENGINE_TYPE_COPY0 (0x00000009) +#define NV2080_ENGINE_TYPE_COPY1 (0x0000000a) +#define NV2080_ENGINE_TYPE_COPY2 (0x0000000b) +#define NV2080_ENGINE_TYPE_COPY3 (0x0000000c) +#define NV2080_ENGINE_TYPE_COPY4 (0x0000000d) +#define NV2080_ENGINE_TYPE_COPY5 (0x0000000e) +#define NV2080_ENGINE_TYPE_COPY6 (0x0000000f) +#define NV2080_ENGINE_TYPE_COPY7 (0x00000010) +#define NV2080_ENGINE_TYPE_COPY8 (0x00000011) +#define NV2080_ENGINE_TYPE_COPY9 (0x00000012) +#define NV2080_ENGINE_TYPE_BSP (0x00000013) +#define NV2080_ENGINE_TYPE_NVDEC0 NV2080_ENGINE_TYPE_BSP +#define NV2080_ENGINE_TYPE_NVDEC1 (0x00000014) +#define NV2080_ENGINE_TYPE_NVDEC2 (0x00000015) +#define NV2080_ENGINE_TYPE_NVDEC3 (0x00000016) +#define NV2080_ENGINE_TYPE_NVDEC4 (0x00000017) +#define NV2080_ENGINE_TYPE_NVDEC5 (0x00000018) +#define NV2080_ENGINE_TYPE_NVDEC6 (0x00000019) +#define NV2080_ENGINE_TYPE_NVDEC7 (0x0000001a) +#define NV2080_ENGINE_TYPE_MSENC (0x0000001b) +#define NV2080_ENGINE_TYPE_NVENC0 NV2080_ENGINE_TYPE_MSENC /* Mutually exclusive alias */ +#define NV2080_ENGINE_TYPE_NVENC1 (0x0000001c) +#define NV2080_ENGINE_TYPE_NVENC2 (0x0000001d) +#define NV2080_ENGINE_TYPE_VP (0x0000001e) +#define NV2080_ENGINE_TYPE_ME (0x0000001f) +#define NV2080_ENGINE_TYPE_PPP (0x00000020) +#define NV2080_ENGINE_TYPE_MPEG (0x00000021) +#define NV2080_ENGINE_TYPE_SW (0x00000022) +#define NV2080_ENGINE_TYPE_CIPHER (0x00000023) +#define NV2080_ENGINE_TYPE_TSEC NV2080_ENGINE_TYPE_CIPHER +#define NV2080_ENGINE_TYPE_VIC (0x00000024) +#define NV2080_ENGINE_TYPE_MP (0x00000025) +#define NV2080_ENGINE_TYPE_SEC2 (0x00000026) +#define NV2080_ENGINE_TYPE_HOST (0x00000027) +#define NV2080_ENGINE_TYPE_DPU (0x00000028) +#define NV2080_ENGINE_TYPE_PMU (0x00000029) +#define NV2080_ENGINE_TYPE_FBFLCN (0x0000002a) +#define NV2080_ENGINE_TYPE_NVJPG (0x0000002b) +#define NV2080_ENGINE_TYPE_NVJPEG0 NV2080_ENGINE_TYPE_NVJPG +#define NV2080_ENGINE_TYPE_NVJPEG1 (0x0000002c) +#define NV2080_ENGINE_TYPE_NVJPEG2 (0x0000002d) +#define NV2080_ENGINE_TYPE_NVJPEG3 (0x0000002e) +#define NV2080_ENGINE_TYPE_NVJPEG4 (0x0000002f) +#define NV2080_ENGINE_TYPE_NVJPEG5 (0x00000030) +#define NV2080_ENGINE_TYPE_NVJPEG6 (0x00000031) +#define NV2080_ENGINE_TYPE_NVJPEG7 (0x00000032) +#define NV2080_ENGINE_TYPE_OFA (0x00000033) +#define NV2080_ENGINE_TYPE_OFA0 NV2080_ENGINE_TYPE_OFA +#define NV2080_ENGINE_TYPE_COPY10 (0x00000034) +#define NV2080_ENGINE_TYPE_COPY11 (0x00000035) +#define NV2080_ENGINE_TYPE_COPY12 (0x00000036) +#define NV2080_ENGINE_TYPE_COPY13 (0x00000037) +#define NV2080_ENGINE_TYPE_COPY14 (0x00000038) +#define NV2080_ENGINE_TYPE_COPY15 (0x00000039) +#define NV2080_ENGINE_TYPE_COPY16 (0x0000003a) +#define NV2080_ENGINE_TYPE_COPY17 (0x0000003b) +#define NV2080_ENGINE_TYPE_COPY18 (0x0000003c) +#define NV2080_ENGINE_TYPE_COPY19 (0x0000003d) +#define NV2080_ENGINE_TYPE_OFA1 (0x0000003e) +#define NV2080_ENGINE_TYPE_NVENC3 (0x0000003f) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY0 (0x00000040) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY1 (0x00000041) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY2 (0x00000042) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY3 (0x00000043) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY4 (0x00000044) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY5 (0x00000045) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY6 (0x00000046) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY7 (0x00000047) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY8 (0x00000048) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY9 (0x00000049) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY10 (0x0000004a) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY11 (0x0000004b) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY12 (0x0000004c) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY13 (0x0000004d) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY14 (0x0000004e) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY15 (0x0000004f) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY16 (0x00000050) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY17 (0x00000051) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY18 (0x00000052) +#define NV2080_ENGINE_TYPE_COMP_DECOMP_COPY19 (0x00000053) +#define NV2080_ENGINE_TYPE_LAST (0x00000054) +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/fbsr.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/fbsr.h new file mode 100644 index 000000000000..8af432375f7a --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/fbsr.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_FBSR_H__ +#define __NVRM_FBSR_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define NV2080_CTRL_CMD_INTERNAL_FBSR_INIT (0x20800ac2) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS { + NvHandle hClient; + NvHandle hSysMem; + NvBool bEnteringGcoffState; + NV_DECLARE_ALIGNED(NvU64 sysmemAddrOfSuspendResumeData, 8); +} NV2080_CTRL_INTERNAL_FBSR_INIT_PARAMS; + +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/alloc/alloc_channel.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/fifo.h index 7157c7757698..2b002ca64e0f 100644 --- a/drivers/gpu/drm/nouveau/include/nvrm/535.113.01/common/sdk/nvidia/inc/alloc/alloc_channel.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/fifo.h @@ -1,31 +1,14 @@ -#ifndef __src_common_sdk_nvidia_inc_alloc_alloc_channel_h__ -#define __src_common_sdk_nvidia_inc_alloc_alloc_channel_h__ -#include <nvrm/535.113.01/common/sdk/nvidia/inc/nvlimits.h> - -/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/535.113.01 */ - -/* - * SPDX-FileCopyrightText: Copyright (c) 2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved. - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_FIFO_H__ +#define __NVRM_FIFO_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define NV_MAX_SUBDEVICES 8 typedef struct NV_MEMORY_DESC_PARAMS { NV_DECLARE_ALIGNED(NvU64 base, 8); @@ -34,137 +17,197 @@ typedef struct NV_MEMORY_DESC_PARAMS { NvU32 cacheAttrib; } NV_MEMORY_DESC_PARAMS; +#define CC_CHAN_ALLOC_IV_SIZE_DWORD 3U + +#define CC_CHAN_ALLOC_NONCE_SIZE_DWORD 8U + +typedef struct NV_CHANNEL_ALLOC_PARAMS { + + NvHandle hObjectError; // error context DMA + NvHandle hObjectBuffer; // no longer used + NV_DECLARE_ALIGNED(NvU64 gpFifoOffset, 8); // offset to beginning of GP FIFO + NvU32 gpFifoEntries; // number of GP FIFO entries + + NvU32 flags; + + + NvHandle hContextShare; // context share handle + NvHandle hVASpace; // VASpace for the channel + + // handle to UserD memory object for channel, ignored if hUserdMemory[0]=0 + NvHandle hUserdMemory[NV_MAX_SUBDEVICES]; + + // offset to beginning of UserD within hUserdMemory[x] + NV_DECLARE_ALIGNED(NvU64 userdOffset[NV_MAX_SUBDEVICES], 8); + + // engine type(NV2080_ENGINE_TYPE_*) with which this channel is associated + NvU32 engineType; + // Channel identifier that is unique for the duration of a RM session + NvU32 cid; + // One-hot encoded bitmask to match SET_SUBDEVICE_MASK methods + NvU32 subDeviceId; + NvHandle hObjectEccError; // ECC error context DMA + + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS instanceMem, 8); + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS userdMem, 8); + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS ramfcMem, 8); + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS mthdbufMem, 8); + + NvHandle hPhysChannelGroup; // reserved + NvU32 internalFlags; // reserved + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS errorNotifierMem, 8); // reserved + NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS eccErrorNotifierMem, 8); // reserved + NvU32 ProcessID; // reserved + NvU32 SubProcessID; // reserved + + // IV used for CPU-side encryption / GPU-side decryption. + NvU32 encryptIv[CC_CHAN_ALLOC_IV_SIZE_DWORD]; // reserved + // IV used for CPU-side decryption / GPU-side encryption. + NvU32 decryptIv[CC_CHAN_ALLOC_IV_SIZE_DWORD]; // reserved + // Nonce used CPU-side signing / GPU-side signature verification. + NvU32 hmacNonce[CC_CHAN_ALLOC_NONCE_SIZE_DWORD]; // reserved + NvU32 tpcConfigID; // TPC Configuration Id as supported by DTD-PG Feature +} NV_CHANNEL_ALLOC_PARAMS; + +typedef NV_CHANNEL_ALLOC_PARAMS NV_CHANNELGPFIFO_ALLOCATION_PARAMETERS; + #define NVOS04_FLAGS_CHANNEL_TYPE 1:0 #define NVOS04_FLAGS_CHANNEL_TYPE_PHYSICAL 0x00000000 #define NVOS04_FLAGS_CHANNEL_TYPE_VIRTUAL 0x00000001 // OBSOLETE #define NVOS04_FLAGS_CHANNEL_TYPE_PHYSICAL_FOR_VIRTUAL 0x00000002 // OBSOLETE - #define NVOS04_FLAGS_VPR 2:2 #define NVOS04_FLAGS_VPR_FALSE 0x00000000 #define NVOS04_FLAGS_VPR_TRUE 0x00000001 - #define NVOS04_FLAGS_CC_SECURE 2:2 #define NVOS04_FLAGS_CC_SECURE_FALSE 0x00000000 #define NVOS04_FLAGS_CC_SECURE_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_SKIP_MAP_REFCOUNTING 3:3 #define NVOS04_FLAGS_CHANNEL_SKIP_MAP_REFCOUNTING_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_SKIP_MAP_REFCOUNTING_TRUE 0x00000001 - #define NVOS04_FLAGS_GROUP_CHANNEL_RUNQUEUE 4:4 #define NVOS04_FLAGS_GROUP_CHANNEL_RUNQUEUE_DEFAULT 0x00000000 #define NVOS04_FLAGS_GROUP_CHANNEL_RUNQUEUE_ONE 0x00000001 - #define NVOS04_FLAGS_PRIVILEGED_CHANNEL 5:5 #define NVOS04_FLAGS_PRIVILEGED_CHANNEL_FALSE 0x00000000 #define NVOS04_FLAGS_PRIVILEGED_CHANNEL_TRUE 0x00000001 - #define NVOS04_FLAGS_DELAY_CHANNEL_SCHEDULING 6:6 #define NVOS04_FLAGS_DELAY_CHANNEL_SCHEDULING_FALSE 0x00000000 #define NVOS04_FLAGS_DELAY_CHANNEL_SCHEDULING_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_DENY_PHYSICAL_MODE_CE 7:7 #define NVOS04_FLAGS_CHANNEL_DENY_PHYSICAL_MODE_CE_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_DENY_PHYSICAL_MODE_CE_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_VALUE 10:8 - #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_FIXED 11:11 #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_FIXED_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_FIXED_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_VALUE 20:12 - #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_FIXED 21:21 #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_FIXED_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_USERD_INDEX_PAGE_FIXED_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_DENY_AUTH_LEVEL_PRIV 22:22 #define NVOS04_FLAGS_CHANNEL_DENY_AUTH_LEVEL_PRIV_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_DENY_AUTH_LEVEL_PRIV_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_SKIP_SCRUBBER 23:23 #define NVOS04_FLAGS_CHANNEL_SKIP_SCRUBBER_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_SKIP_SCRUBBER_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_CLIENT_MAP_FIFO 24:24 #define NVOS04_FLAGS_CHANNEL_CLIENT_MAP_FIFO_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_CLIENT_MAP_FIFO_TRUE 0x00000001 - #define NVOS04_FLAGS_SET_EVICT_LAST_CE_PREFETCH_CHANNEL 25:25 #define NVOS04_FLAGS_SET_EVICT_LAST_CE_PREFETCH_CHANNEL_FALSE 0x00000000 #define NVOS04_FLAGS_SET_EVICT_LAST_CE_PREFETCH_CHANNEL_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_VGPU_PLUGIN_CONTEXT 26:26 #define NVOS04_FLAGS_CHANNEL_VGPU_PLUGIN_CONTEXT_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_VGPU_PLUGIN_CONTEXT_TRUE 0x00000001 - #define NVOS04_FLAGS_CHANNEL_PBDMA_ACQUIRE_TIMEOUT 27:27 #define NVOS04_FLAGS_CHANNEL_PBDMA_ACQUIRE_TIMEOUT_FALSE 0x00000000 #define NVOS04_FLAGS_CHANNEL_PBDMA_ACQUIRE_TIMEOUT_TRUE 0x00000001 - #define NVOS04_FLAGS_GROUP_CHANNEL_THREAD 29:28 #define NVOS04_FLAGS_GROUP_CHANNEL_THREAD_DEFAULT 0x00000000 #define NVOS04_FLAGS_GROUP_CHANNEL_THREAD_ONE 0x00000001 #define NVOS04_FLAGS_GROUP_CHANNEL_THREAD_TWO 0x00000002 - #define NVOS04_FLAGS_MAP_CHANNEL 30:30 #define NVOS04_FLAGS_MAP_CHANNEL_FALSE 0x00000000 #define NVOS04_FLAGS_MAP_CHANNEL_TRUE 0x00000001 - #define NVOS04_FLAGS_SKIP_CTXBUFFER_ALLOC 31:31 #define NVOS04_FLAGS_SKIP_CTXBUFFER_ALLOC_FALSE 0x00000000 #define NVOS04_FLAGS_SKIP_CTXBUFFER_ALLOC_TRUE 0x00000001 -#define CC_CHAN_ALLOC_IV_SIZE_DWORD 3U -#define CC_CHAN_ALLOC_NONCE_SIZE_DWORD 8U - -typedef struct NV_CHANNEL_ALLOC_PARAMS { - - NvHandle hObjectError; // error context DMA - NvHandle hObjectBuffer; // no longer used - NV_DECLARE_ALIGNED(NvU64 gpFifoOffset, 8); // offset to beginning of GP FIFO - NvU32 gpFifoEntries; // number of GP FIFO entries - - NvU32 flags; - - - NvHandle hContextShare; // context share handle - NvHandle hVASpace; // VASpace for the channel - - // handle to UserD memory object for channel, ignored if hUserdMemory[0]=0 - NvHandle hUserdMemory[NV_MAX_SUBDEVICES]; - - // offset to beginning of UserD within hUserdMemory[x] - NV_DECLARE_ALIGNED(NvU64 userdOffset[NV_MAX_SUBDEVICES], 8); - - // engine type(NV2080_ENGINE_TYPE_*) with which this channel is associated - NvU32 engineType; - // Channel identifier that is unique for the duration of a RM session - NvU32 cid; - // One-hot encoded bitmask to match SET_SUBDEVICE_MASK methods - NvU32 subDeviceId; - NvHandle hObjectEccError; // ECC error context DMA - - NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS instanceMem, 8); - NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS userdMem, 8); - NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS ramfcMem, 8); - NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS mthdbufMem, 8); - - NvHandle hPhysChannelGroup; // reserved - NvU32 internalFlags; // reserved - NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS errorNotifierMem, 8); // reserved - NV_DECLARE_ALIGNED(NV_MEMORY_DESC_PARAMS eccErrorNotifierMem, 8); // reserved - NvU32 ProcessID; // reserved - NvU32 SubProcessID; // reserved - // IV used for CPU-side encryption / GPU-side decryption. - NvU32 encryptIv[CC_CHAN_ALLOC_IV_SIZE_DWORD]; // reserved - // IV used for CPU-side decryption / GPU-side encryption. - NvU32 decryptIv[CC_CHAN_ALLOC_IV_SIZE_DWORD]; // reserved - // Nonce used CPU-side signing / GPU-side signature verification. - NvU32 hmacNonce[CC_CHAN_ALLOC_NONCE_SIZE_DWORD]; // reserved -} NV_CHANNEL_ALLOC_PARAMS; - -typedef NV_CHANNEL_ALLOC_PARAMS NV_CHANNELGPFIFO_ALLOCATION_PARAMETERS; - +typedef enum { + /*! + * Initial state as passed in NV_CHANNEL_ALLOC_PARAMS by + * kernel CPU-RM clients. + */ + ERROR_NOTIFIER_TYPE_UNKNOWN = 0, + /*! @brief Error notifier is explicitly not set. + * + * The corresponding hErrorContext or hEccErrorContext must be + * NV01_NULL_OBJECT. + */ + ERROR_NOTIFIER_TYPE_NONE, + /*! @brief Error notifier is a ContextDma */ + ERROR_NOTIFIER_TYPE_CTXDMA, + /*! @brief Error notifier is a NvNotification array in sysmem/vidmem */ + ERROR_NOTIFIER_TYPE_MEMORY +} ErrorNotifierType; + +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE 1:0 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_USER 0x0 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_ADMIN 0x1 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_PRIVILEGE_KERNEL 0x2 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE 3:2 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_UNKNOWN ERROR_NOTIFIER_TYPE_UNKNOWN +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_NONE ERROR_NOTIFIER_TYPE_NONE +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_CTXDMA ERROR_NOTIFIER_TYPE_CTXDMA +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ERROR_NOTIFIER_TYPE_MEMORY ERROR_NOTIFIER_TYPE_MEMORY +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE 5:4 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_UNKNOWN ERROR_NOTIFIER_TYPE_UNKNOWN +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_NONE ERROR_NOTIFIER_TYPE_NONE +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_CTXDMA ERROR_NOTIFIER_TYPE_CTXDMA +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_ECC_ERROR_NOTIFIER_TYPE_MEMORY ERROR_NOTIFIER_TYPE_MEMORY +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_GSP_OWNED 6:6 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_GSP_OWNED_NO 0x0 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_GSP_OWNED_YES 0x1 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_UVM_OWNED 7:7 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_UVM_OWNED_NO 0x0 +#define NV_KERNELCHANNEL_ALLOC_INTERNALFLAGS_UVM_OWNED_YES 0x1 + +typedef struct rpc_rc_triggered_v17_02 +{ + NvU32 nv2080EngineType; + NvU32 chid; + NvU32 gfid; + NvU32 exceptLevel; + NvU32 exceptType; + NvU32 scope; + NvU16 partitionAttributionId; + NvU32 mmuFaultAddrLo; + NvU32 mmuFaultAddrHi; + NvU32 mmuFaultType; + NvBool bCallbackNeeded; + NvU32 rcJournalBufferSize; + NvU8 rcJournalBuffer[]; +} rpc_rc_triggered_v17_02; + +#define NV2080_CTRL_GPU_MAX_CONSTRUCTED_FALCONS 0x40 + +typedef struct NV2080_CTRL_GPU_CONSTRUCTED_FALCON_INFO { + NvU32 engDesc; + NvU32 ctxAttr; + NvU32 ctxBufferSize; + NvU32 addrSpaceList; + NvU32 registerBase; +} NV2080_CTRL_GPU_CONSTRUCTED_FALCON_INFO; + +#define NV2080_CTRL_CMD_GPU_GET_CONSTRUCTED_FALCON_INFO (0x208001b0) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_GPU_INTERFACE_ID << 8) | NV2080_CTRL_GPU_GET_CONSTRUCTED_FALCON_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_GPU_GET_CONSTRUCTED_FALCON_INFO_PARAMS { + NvU32 numConstructedFalcons; + NV2080_CTRL_GPU_CONSTRUCTED_FALCON_INFO constructedFalconsTable[NV2080_CTRL_GPU_MAX_CONSTRUCTED_FALCONS]; +} NV2080_CTRL_GPU_GET_CONSTRUCTED_FALCON_INFO_PARAMS; + +typedef struct NV2080_CTRL_CMD_INTERNAL_FIFO_TOGGLE_ACTIVE_CHANNEL_SCHEDULING_PARAMS { + NvBool bDisableActiveChannels; +} NV2080_CTRL_CMD_INTERNAL_FIFO_TOGGLE_ACTIVE_CHANNEL_SCHEDULING_PARAMS; + +#define NV2080_CTRL_CMD_INTERNAL_FIFO_TOGGLE_ACTIVE_CHANNEL_SCHEDULING (0x20800ac3) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_CMD_INTERNAL_FIFO_TOGGLE_ACTIVE_CHANNEL_SCHEDULING_PARAMS_MESSAGE_ID" */ #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/gr.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/gr.h new file mode 100644 index 000000000000..feed1dabd9d2 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/gr.h @@ -0,0 +1,79 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_GR_H__ +#define __NVRM_GR_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define NV2080_CTRL_CMD_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO (0x20800a32) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_STATIC_KGR_GET_CONTEXT_BUFFERS_INFO_PARAMS_MESSAGE_ID" */ + +#define NV2080_CTRL_INTERNAL_GR_MAX_ENGINES 8 + +#define NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT 0x1a + +typedef struct NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO { + NvU32 size; + NvU32 alignment; +} NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO; + +typedef struct NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO { + NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_BUFFER_INFO engine[NV2080_CTRL_INTERNAL_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT]; +} NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO; + +typedef struct NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS { + NV2080_CTRL_INTERNAL_STATIC_GR_CONTEXT_BUFFERS_INFO engineContextBuffersInfo[NV2080_CTRL_INTERNAL_GR_MAX_ENGINES]; +} NV2080_CTRL_INTERNAL_STATIC_GR_GET_CONTEXT_BUFFERS_INFO_PARAMS; + +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID 4:0 +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS (0x00000000) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_VLD (0x00000001) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_VIDEO (0x00000002) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_MPEG (0x00000003) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_CAPTURE (0x00000004) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_DISPLAY (0x00000005) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_ENCRYPTION (0x00000006) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_POSTPROCESS (0x00000007) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_ZCULL (0x00000008) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PM (0x00000009) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COMPUTE_PREEMPT (0x0000000a) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PREEMPT (0x0000000b) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_SPILL (0x0000000c) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PAGEPOOL (0x0000000d) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_BETACB (0x0000000e) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_RTV (0x0000000f) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PATCH (0x00000010) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_BUNDLE_CB (0x00000011) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PAGEPOOL_GLOBAL (0x00000012) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_ATTRIBUTE_CB (0x00000013) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_RTV_CB_GLOBAL (0x00000014) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_GFXP_POOL (0x00000015) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_GFXP_CTRL_BLK (0x00000016) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_FECS_EVENT (0x00000017) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_PRIV_ACCESS_MAP (0x00000018) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_GRAPHICS_SETUP (0x00000019) +#define NV0080_CTRL_FIFO_GET_ENGINE_CONTEXT_PROPERTIES_ENGINE_ID_COUNT (0x0000001a) + +#define NV2080_CTRL_CMD_GPU_GET_FERMI_GPC_INFO (0x20800137U) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_GPU_INTERFACE_ID << 8) | NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS { + NvU32 gpcMask; +} NV2080_CTRL_GPU_GET_FERMI_GPC_INFO_PARAMS; + +#define NV2080_CTRL_CMD_GPU_GET_FERMI_TPC_INFO (0x20800138U) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_GPU_INTERFACE_ID << 8) | NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS_MESSAGE_ID" */ +typedef struct NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS { + NvU32 gpcId; + NvU32 tpcMask; +} NV2080_CTRL_GPU_GET_FERMI_TPC_INFO_PARAMS; + +#define KGRAPHICS_SCRUBBER_HANDLE_VAS 0xdada0042 +#define KGRAPHICS_SCRUBBER_HANDLE_CHANNEL (KGRAPHICS_SCRUBBER_HANDLE_VAS + 3) +#define KGRAPHICS_SCRUBBER_HANDLE_3DOBJ (KGRAPHICS_SCRUBBER_HANDLE_VAS + 4) + +typedef struct NV2080_CTRL_INTERNAL_GR_INIT_BUG4208224_WAR_PARAMS { + NvBool bTeardown; +} NV2080_CTRL_INTERNAL_GR_INIT_BUG4208224_WAR_PARAMS; + +#define NV2080_CTRL_CMD_INTERNAL_KGR_INIT_BUG4208224_WAR (0x20800a46) /* finn: Evaluated from "(FINN_NV20_SUBDEVICE_0_INTERNAL_INTERFACE_ID << 8) | NV2080_CTRL_INTERNAL_KGR_INIT_BUG4208224_WAR_PARAMS_MESSAGE_ID" */ +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/gsp.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/gsp.h new file mode 100644 index 000000000000..b6075021e74f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/gsp.h @@ -0,0 +1,634 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_GSP_H__ +#define __NVRM_GSP_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#define NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MAX_ENTRIES 16U + +#define NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MEM_TYPES 17U + +typedef NvBool NV2080_CTRL_CMD_FB_GET_FB_REGION_SURFACE_MEM_TYPE_FLAG[NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MEM_TYPES]; + +typedef struct NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO { + NV_DECLARE_ALIGNED(NvU64 base, 8); + NV_DECLARE_ALIGNED(NvU64 limit, 8); + NV_DECLARE_ALIGNED(NvU64 reserved, 8); + NvU32 performance; + NvBool supportCompressed; + NvBool supportISO; + NvBool bProtected; + NV2080_CTRL_CMD_FB_GET_FB_REGION_SURFACE_MEM_TYPE_FLAG blackList; +} NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO; + +typedef struct NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS { + NvU32 numFBRegions; + NV_DECLARE_ALIGNED(NV2080_CTRL_CMD_FB_GET_FB_REGION_FB_REGION_INFO fbRegion[NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_MAX_ENTRIES], 8); +} NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS; + +#define NV0080_CTRL_GR_CAPS_TBL_SIZE 23 + +#define NV2080_GPU_MAX_GID_LENGTH (0x000000100ULL) + +typedef struct NV2080_CTRL_GPU_GET_GID_INFO_PARAMS { + NvU32 index; + NvU32 flags; + NvU32 length; + NvU8 data[NV2080_GPU_MAX_GID_LENGTH]; +} NV2080_CTRL_GPU_GET_GID_INFO_PARAMS; + +typedef struct NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS { + NvU32 BoardID; + char chipSKU[9]; + char chipSKUMod[5]; + NvU32 skuConfigVersion; + char project[5]; + char projectSKU[5]; + char CDP[6]; + char projectSKUMod[2]; + NvU32 businessCycle; +} NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS; + +#define MAX_GPC_COUNT 32 + +typedef struct NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS { + NvU32 totalVFs; + NvU32 firstVfOffset; + NvU32 vfFeatureMask; + NV_DECLARE_ALIGNED(NvU64 FirstVFBar0Address, 8); + NV_DECLARE_ALIGNED(NvU64 FirstVFBar1Address, 8); + NV_DECLARE_ALIGNED(NvU64 FirstVFBar2Address, 8); + NV_DECLARE_ALIGNED(NvU64 bar0Size, 8); + NV_DECLARE_ALIGNED(NvU64 bar1Size, 8); + NV_DECLARE_ALIGNED(NvU64 bar2Size, 8); + NvBool b64bitBar0; + NvBool b64bitBar1; + NvBool b64bitBar2; + NvBool bSriovEnabled; + NvBool bSriovHeavyEnabled; + NvBool bEmulateVFBar0TlbInvalidationRegister; + NvBool bClientRmAllocatedCtxBuffer; + NvBool bNonPowerOf2ChannelCountSupported; + NvBool bVfResizableBAR1Supported; +} NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS; + +#include "engine.h" + +#define NVGPU_ENGINE_CAPS_MASK_BITS 32 + +#define NVGPU_ENGINE_CAPS_MASK_ARRAY_MAX ((RM_ENGINE_TYPE_LAST-1)/NVGPU_ENGINE_CAPS_MASK_BITS + 1) + +#define NV2080_GPU_MAX_NAME_STRING_LENGTH (0x0000040U) + +typedef struct VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS +{ + NvU32 numHeads; + NvU32 maxNumHeads; +} VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS; + +typedef struct VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS +{ + NvU32 headIndex; + NvU32 maxHResolution; + NvU32 maxVResolution; +} VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS; + +#define MAX_GROUP_COUNT 2 + +typedef struct +{ + NvU32 ecidLow; + NvU32 ecidHigh; + NvU32 ecidExtended; +} EcidManufacturingInfo; + +typedef struct +{ + NvU64 nonWprHeapOffset; + NvU64 frtsOffset; +} FW_WPR_LAYOUT_OFFSET; + +typedef struct GspStaticConfigInfo_t +{ + NvU8 grCapsBits[NV0080_CTRL_GR_CAPS_TBL_SIZE]; + NV2080_CTRL_GPU_GET_GID_INFO_PARAMS gidInfo; + NV2080_CTRL_BIOS_GET_SKU_INFO_PARAMS SKUInfo; + NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS fbRegionInfoParams; + + NV0080_CTRL_GPU_GET_SRIOV_CAPS_PARAMS sriovCaps; + NvU32 sriovMaxGfid; + + NvU32 engineCaps[NVGPU_ENGINE_CAPS_MASK_ARRAY_MAX]; + + NvBool poisonFuseEnabled; + + NvU64 fb_length; + NvU64 fbio_mask; + NvU32 fb_bus_width; + NvU32 fb_ram_type; + NvU64 fbp_mask; + NvU32 l2_cache_size; + + NvU8 gpuNameString[NV2080_GPU_MAX_NAME_STRING_LENGTH]; + NvU8 gpuShortNameString[NV2080_GPU_MAX_NAME_STRING_LENGTH]; + NvU16 gpuNameString_Unicode[NV2080_GPU_MAX_NAME_STRING_LENGTH]; + NvBool bGpuInternalSku; + NvBool bIsQuadroGeneric; + NvBool bIsQuadroAd; + NvBool bIsNvidiaNvs; + NvBool bIsVgx; + NvBool bGeforceSmb; + NvBool bIsTitan; + NvBool bIsTesla; + NvBool bIsMobile; + NvBool bIsGc6Rtd3Allowed; + NvBool bIsGc8Rtd3Allowed; + NvBool bIsGcOffRtd3Allowed; + NvBool bIsGcoffLegacyAllowed; + NvBool bIsMigSupported; + + /* "Total Board Power" refers to power requirement of GPU, + * while in GC6 state. Majority of this power will be used + * to keep V-RAM active to preserve its content. + * Some energy maybe consumed by Always-on components on GPU chip. + * This power will be provided by 3.3v voltage rail. + */ + NvU16 RTD3GC6TotalBoardPower; + + /* PERST# (i.e. PCI Express Reset) is a sideband signal + * generated by the PCIe Host to indicate the PCIe devices, + * that the power-rails and the reference-clock are stable. + * The endpoint device typically uses this signal as a global reset. + */ + NvU16 RTD3GC6PerstDelay; + + NvU64 bar1PdeBase; + NvU64 bar2PdeBase; + + NvBool bVbiosValid; + NvU32 vbiosSubVendor; + NvU32 vbiosSubDevice; + + NvBool bPageRetirementSupported; + + NvBool bSplitVasBetweenServerClientRm; + + NvBool bClRootportNeedsNosnoopWAR; + + VIRTUAL_DISPLAY_GET_NUM_HEADS_PARAMS displaylessMaxHeads; + VIRTUAL_DISPLAY_GET_MAX_RESOLUTION_PARAMS displaylessMaxResolution; + NvU64 displaylessMaxPixels; + + // Client handle for internal RMAPI control. + NvHandle hInternalClient; + + // Device handle for internal RMAPI control. + NvHandle hInternalDevice; + + // Subdevice handle for internal RMAPI control. + NvHandle hInternalSubdevice; + + NvBool bSelfHostedMode; + NvBool bAtsSupported; + + NvBool bIsGpuUefi; + NvBool bIsEfiInit; + + EcidManufacturingInfo ecidInfo[MAX_GROUP_COUNT]; + + FW_WPR_LAYOUT_OFFSET fwWprLayoutOffset; +} GspStaticConfigInfo; + +typedef struct +{ + NvU16 deviceID; // deviceID + NvU16 vendorID; // vendorID + NvU16 subdeviceID; // subsystem deviceID + NvU16 subvendorID; // subsystem vendorID + NvU8 revisionID; // revision ID +} BUSINFO; + +#define NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS (16U) + +typedef struct DOD_METHOD_DATA +{ + NV_STATUS status; + NvU32 acpiIdListLen; + NvU32 acpiIdList[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; +} DOD_METHOD_DATA; + +typedef struct JT_METHOD_DATA +{ + NV_STATUS status; + NvU32 jtCaps; + NvU16 jtRevId; + NvBool bSBIOSCaps; +} JT_METHOD_DATA; + +typedef struct MUX_METHOD_DATA_ELEMENT +{ + NvU32 acpiId; + NvU32 mode; + NV_STATUS status; +} MUX_METHOD_DATA_ELEMENT; + +#define NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS (16U) + +typedef struct MUX_METHOD_DATA +{ + NvU32 tableLen; + MUX_METHOD_DATA_ELEMENT acpiIdMuxModeTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; + MUX_METHOD_DATA_ELEMENT acpiIdMuxPartTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; + MUX_METHOD_DATA_ELEMENT acpiIdMuxStateTable[NV0073_CTRL_SYSTEM_ACPI_ID_MAP_MAX_DISPLAYS]; +} MUX_METHOD_DATA; + +typedef struct CAPS_METHOD_DATA +{ + NV_STATUS status; + NvU32 optimusCaps; +} CAPS_METHOD_DATA; + +typedef struct ACPI_METHOD_DATA +{ + NvBool bValid; + DOD_METHOD_DATA dodMethodData; + JT_METHOD_DATA jtMethodData; + MUX_METHOD_DATA muxMethodData; + CAPS_METHOD_DATA capsMethodData; +} ACPI_METHOD_DATA; + +typedef struct GSP_VF_INFO +{ + NvU32 totalVFs; + NvU32 firstVFOffset; + NvU64 FirstVFBar0Address; + NvU64 FirstVFBar1Address; + NvU64 FirstVFBar2Address; + NvBool b64bitBar0; + NvBool b64bitBar1; + NvBool b64bitBar2; +} GSP_VF_INFO; + +typedef struct +{ + // Link capabilities + NvU32 linkCap; +} GSP_PCIE_CONFIG_REG; + +typedef struct GspSystemInfo +{ + NvU64 gpuPhysAddr; + NvU64 gpuPhysFbAddr; + NvU64 gpuPhysInstAddr; + NvU64 gpuPhysIoAddr; + NvU64 nvDomainBusDeviceFunc; + NvU64 simAccessBufPhysAddr; + NvU64 notifyOpSharedSurfacePhysAddr; + NvU64 pcieAtomicsOpMask; + NvU64 consoleMemSize; + NvU64 maxUserVa; + NvU32 pciConfigMirrorBase; + NvU32 pciConfigMirrorSize; + NvU32 PCIDeviceID; + NvU32 PCISubDeviceID; + NvU32 PCIRevisionID; + NvU32 pcieAtomicsCplDeviceCapMask; + NvU8 oorArch; + NvU64 clPdbProperties; + NvU32 Chipset; + NvBool bGpuBehindBridge; + NvBool bFlrSupported; + NvBool b64bBar0Supported; + NvBool bMnocAvailable; + NvU32 chipsetL1ssEnable; + NvBool bUpstreamL0sUnsupported; + NvBool bUpstreamL1Unsupported; + NvBool bUpstreamL1PorSupported; + NvBool bUpstreamL1PorMobileOnly; + NvBool bSystemHasMux; + NvU8 upstreamAddressValid; + BUSINFO FHBBusInfo; + BUSINFO chipsetIDInfo; + ACPI_METHOD_DATA acpiMethodData; + NvU32 hypervisorType; + NvBool bIsPassthru; + NvU64 sysTimerOffsetNs; + GSP_VF_INFO gspVFInfo; + NvBool bIsPrimary; + NvBool isGridBuild; + GSP_PCIE_CONFIG_REG pcieConfigReg; + NvU32 gridBuildCsp; + NvBool bPreserveVideoMemoryAllocations; + NvBool bTdrEventSupported; + NvBool bFeatureStretchVblankCapable; + NvBool bEnableDynamicGranularityPageArrays; + NvBool bClockBoostSupported; + NvBool bRouteDispIntrsToCPU; + NvU64 hostPageSize; +} GspSystemInfo; + +typedef struct rpc_os_error_log_v17_00 +{ + NvU32 exceptType; + NvU32 runlistId; + NvU32 chid; + char errString[0x100]; + NvU32 preemptiveRemovalPreviousXid; +} rpc_os_error_log_v17_00; + +typedef struct +{ + // Magic + // BL to use for verification (i.e. Booter locked it in WPR2) + NvU64 magic; // = 0xdc3aae21371a60b3; + + // Revision number of Booter-BL-Sequencer handoff interface + // Bumped up when we change this interface so it is not backward compatible. + // Bumped up when we revoke GSP-RM ucode + NvU64 revision; // = 1; + + // ---- Members regarding data in SYSMEM ---------------------------- + // Consumed by Booter for DMA + + NvU64 sysmemAddrOfRadix3Elf; + NvU64 sizeOfRadix3Elf; + + NvU64 sysmemAddrOfBootloader; + NvU64 sizeOfBootloader; + + // Offsets inside bootloader image needed by Booter + NvU64 bootloaderCodeOffset; + NvU64 bootloaderDataOffset; + NvU64 bootloaderManifestOffset; + + union + { + // Used only at initial boot + struct + { + NvU64 sysmemAddrOfSignature; + NvU64 sizeOfSignature; + }; + + // + // Used at suspend/resume to read GspFwHeapFreeList + // Offset relative to GspFwWprMeta FBMEM PA (gspFwWprStart) + // + struct + { + NvU32 gspFwHeapFreeListWprOffset; + NvU32 unused0; + NvU64 unused1; + }; + }; + + // ---- Members describing FB layout -------------------------------- + NvU64 gspFwRsvdStart; + + NvU64 nonWprHeapOffset; + NvU64 nonWprHeapSize; + + NvU64 gspFwWprStart; + + // GSP-RM to use to setup heap. + NvU64 gspFwHeapOffset; + NvU64 gspFwHeapSize; + + // BL to use to find ELF for jump + NvU64 gspFwOffset; + // Size is sizeOfRadix3Elf above. + + NvU64 bootBinOffset; + // Size is sizeOfBootloader above. + + NvU64 frtsOffset; + NvU64 frtsSize; + + NvU64 gspFwWprEnd; + + // GSP-RM to use for fbRegionInfo? + NvU64 fbSize; + + // ---- Other members ----------------------------------------------- + + // GSP-RM to use for fbRegionInfo? + NvU64 vgaWorkspaceOffset; + NvU64 vgaWorkspaceSize; + + // Boot count. Used to determine whether to load the firmware image. + NvU64 bootCount; + + // This union is organized the way it is to start at an 8-byte boundary and achieve natural + // packing of the internal struct fields. + union + { + struct + { + // TODO: the partitionRpc* fields below do not really belong in this + // structure. The values are patched in by the partition bootstrapper + // when GSP-RM is booted in a partition, and this structure was a + // convenient place for the bootstrapper to access them. These should + // be moved to a different comm. mechanism between the bootstrapper + // and the GSP-RM tasks. + + // Shared partition RPC memory (physical address) + NvU64 partitionRpcAddr; + + // Offsets relative to partitionRpcAddr + NvU16 partitionRpcRequestOffset; + NvU16 partitionRpcReplyOffset; + + // Code section and dataSection offset and size. + NvU32 elfCodeOffset; + NvU32 elfDataOffset; + NvU32 elfCodeSize; + NvU32 elfDataSize; + + // Used during GSP-RM resume to check for revocation + NvU32 lsUcodeVersion; + }; + + struct + { + // Pad for the partitionRpc* fields, plus 4 bytes + NvU32 partitionRpcPadding[4]; + + // CrashCat (contiguous) buffer size/location - occupies same bytes as the + // elf(Code|Data)(Offset|Size) fields above. + // TODO: move to GSP_FMC_INIT_PARAMS + NvU64 sysmemAddrOfCrashReportQueue; + NvU32 sizeOfCrashReportQueue; + + // Pad for the lsUcodeVersion field + NvU32 lsUcodeVersionPadding[1]; + }; + }; + + // Number of VF partitions allocating sub-heaps from the WPR heap + // Used during boot to ensure the heap is adequately sized + NvU8 gspFwHeapVfPartitionCount; + + // Flags to help decide GSP-FW flow. + NvU8 flags; + + // Pad structure to exactly 256 bytes. Can replace padding with additional + // fields without incrementing revision. Padding initialized to 0. + NvU8 padding[2]; + + // + // Starts at gspFwWprEnd+frtsSize b/c FRTS is positioned + // to end where this allocation starts (when RM requests FSP to create + // FRTS). + // + NvU32 pmuReservedSize; + + // BL to use for verification (i.e. Booter says OK to boot) + NvU64 verified; // 0x0 -> unverified, 0xa0a0a0a0a0a0a0a0 -> verified +} GspFwWprMeta; + +#define GSP_FW_WPR_META_MAGIC 0xdc3aae21371a60b3ULL + +#define GSP_FW_WPR_META_REVISION 1 + +typedef struct { + NvU64 sharedMemPhysAddr; + NvU32 pageTableEntryCount; + NvLength cmdQueueOffset; + NvLength statQueueOffset; +} MESSAGE_QUEUE_INIT_ARGUMENTS; + +typedef struct { + NvU32 oldLevel; + NvU32 flags; + NvBool bInPMTransition; +} GSP_SR_INIT_ARGUMENTS; + +typedef struct +{ + MESSAGE_QUEUE_INIT_ARGUMENTS messageQueueInitArguments; + GSP_SR_INIT_ARGUMENTS srInitArguments; + NvU32 gpuInstance; + NvBool bDmemStack; + + struct + { + NvU64 pa; + NvU64 size; + } profilerArgs; +} GSP_ARGUMENTS_CACHED; + +#define NV2080_CTRL_GPU_SET_POWER_STATE_GPU_LEVEL_3 (0x00000003U) + +typedef struct +{ + // Magic for verification by secure ucode + NvU64 magic; // = GSP_FW_SR_META_MAGIC; + + // + // Revision number + // Bumped up when we change this interface so it is not backward compatible. + // + NvU64 revision; // = GSP_FW_SR_META_MAGIC_REVISION; + + // Members regarding data in SYSMEM + NvU64 sysmemAddrOfSuspendResumeData; + NvU64 sizeOfSuspendResumeData; + + // + // Internal members for use by secure ucode + // Must be exactly GSP_FW_SR_META_INTERNAL_SIZE bytes. + // + NvU32 internal[32]; + + // Same as flags of GspFwWprMeta + NvU32 flags; + + // Subrevision number used by secure ucode + NvU32 subrevision; + + // + // Pad structure to exactly 256 bytes (1 DMA chunk). + // Padding initialized to zero. + // + NvU32 padding[22]; +} GspFwSRMeta; + +#define GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS2 (0 << 20) // No FB heap usage + +#define GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3_BAREMETAL (22 << 20) + +#define GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X (8 << 20) // Turing thru Ada + +#define GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS2_MIN_MB (64u) + +#define BULLSEYE_ROOT_HEAP_ALLOC_RM_DATA_SECTION_SIZE_DELTA (12u) + +#define BULLSEYE_ROOT_HEAP_ALLOC_BAREMETAL_LIBOS_HEAP_SIZE_DELTA (70u) + +#define GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB \ + (88u + (BULLSEYE_ROOT_HEAP_ALLOC_RM_DATA_SECTION_SIZE_DELTA) + \ + (BULLSEYE_ROOT_HEAP_ALLOC_BAREMETAL_LIBOS_HEAP_SIZE_DELTA)) + +typedef struct GSP_FMC_INIT_PARAMS +{ + // CC initialization "registry keys" + NvU32 regkeys; +} GSP_FMC_INIT_PARAMS; + +typedef enum { + GSP_DMA_TARGET_LOCAL_FB, + GSP_DMA_TARGET_COHERENT_SYSTEM, + GSP_DMA_TARGET_NONCOHERENT_SYSTEM, + GSP_DMA_TARGET_COUNT +} GSP_DMA_TARGET; + +typedef struct GSP_ACR_BOOT_GSP_RM_PARAMS +{ + // Physical memory aperture through which gspRmDescPa is accessed + GSP_DMA_TARGET target; + // Size in bytes of the GSP-RM descriptor structure + NvU32 gspRmDescSize; + // Physical offset in the target aperture of the GSP-RM descriptor structure + NvU64 gspRmDescOffset; + // Physical offset in FB to set the start of the WPR containing GSP-RM + NvU64 wprCarveoutOffset; + // Size in bytes of the WPR containing GSP-RM + NvU32 wprCarveoutSize; + // Whether to boot GSP-RM or GSP-Proxy through ACR + NvBool bIsGspRmBoot; +} GSP_ACR_BOOT_GSP_RM_PARAMS; + +typedef struct GSP_RM_PARAMS +{ + // Physical memory aperture through which bootArgsOffset is accessed + GSP_DMA_TARGET target; + // Physical offset in the memory aperture that will be passed to GSP-RM + NvU64 bootArgsOffset; +} GSP_RM_PARAMS; + +typedef struct GSP_SPDM_PARAMS +{ + // Physical Memory Aperture through which all addresses are accessed + GSP_DMA_TARGET target; + + // Physical offset in the memory aperture where SPDM payload is stored + NvU64 payloadBufferOffset; + + // Size of the above payload buffer + NvU32 payloadBufferSize; +} GSP_SPDM_PARAMS; + +typedef struct GSP_FMC_BOOT_PARAMS +{ + GSP_FMC_INIT_PARAMS initParams; + GSP_ACR_BOOT_GSP_RM_PARAMS bootGspRmParams; + GSP_RM_PARAMS gspRmParams; + GSP_SPDM_PARAMS gspSpdmParams; +} GSP_FMC_BOOT_PARAMS; + +#define GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100 (14 << 20) // Hopper+ +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/msgfn.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/msgfn.h new file mode 100644 index 000000000000..e06643f57695 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/msgfn.h @@ -0,0 +1,57 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_MSGFN_H__ +#define __NVRM_MSGFN_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#ifndef E +# define E(RPC, VAL) NV_VGPU_MSG_EVENT_##RPC = VAL, +# define DEFINING_E_IN_RPC_GLOBAL_ENUMS_H +enum { +#endif + E(FIRST_EVENT, 0x1000) + E(GSP_INIT_DONE, 0x1001) + E(GSP_RUN_CPU_SEQUENCER, 0x1002) + E(POST_EVENT, 0x1003) + E(RC_TRIGGERED, 0x1004) + E(MMU_FAULT_QUEUED, 0x1005) + E(OS_ERROR_LOG, 0x1006) + E(RG_LINE_INTR, 0x1007) + E(GPUACCT_PERFMON_UTIL_SAMPLES, 0x1008) + E(SIM_READ, 0x1009) + E(SIM_WRITE, 0x100a) + E(SEMAPHORE_SCHEDULE_CALLBACK, 0x100b) + E(UCODE_LIBOS_PRINT, 0x100c) + E(VGPU_GSP_PLUGIN_TRIGGERED, 0x100d) + E(PERF_GPU_BOOST_SYNC_LIMITS_CALLBACK, 0x100e) + E(PERF_BRIDGELESS_INFO_UPDATE, 0x100f) + E(VGPU_CONFIG, 0x1010) + E(DISPLAY_MODESET, 0x1011) + E(EXTDEV_INTR_SERVICE, 0x1012) + E(NVLINK_INBAND_RECEIVED_DATA_256, 0x1013) + E(NVLINK_INBAND_RECEIVED_DATA_512, 0x1014) + E(NVLINK_INBAND_RECEIVED_DATA_1024, 0x1015) + E(NVLINK_INBAND_RECEIVED_DATA_2048, 0x1016) + E(NVLINK_INBAND_RECEIVED_DATA_4096, 0x1017) + E(TIMED_SEMAPHORE_RELEASE, 0x1018) + E(NVLINK_IS_GPU_DEGRADED, 0x1019) + E(PFM_REQ_HNDLR_STATE_SYNC_CALLBACK, 0x101a) + E(NVLINK_FAULT_UP, 0x101b) + E(GSP_LOCKDOWN_NOTICE, 0x101c) + E(MIG_CI_CONFIG_UPDATE, 0x101d) + E(UPDATE_GSP_TRACE, 0x101e) + E(NVLINK_FATAL_ERROR_RECOVERY, 0x101f) + E(GSP_POST_NOCAT_RECORD, 0x1020) + E(FECS_ERROR, 0x1021) + E(RECOVERY_ACTION, 0x1022) + E(NUM_EVENTS, 0x1023) +#ifdef DEFINING_E_IN_RPC_GLOBAL_ENUMS_H +}; +# undef E +# undef DEFINING_E_IN_RPC_GLOBAL_ENUMS_H +#endif +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/ofa.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/ofa.h new file mode 100644 index 000000000000..fcaef7f553a6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/ofa.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_OFA_H__ +#define __NVRM_OFA_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +typedef struct +{ + NvU32 size; + NvU32 prohibitMultipleInstances; // Prohibit multiple allocations of OFA? + NvU32 engineInstance; +} NV_OFA_ALLOCATION_PARAMETERS; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/rpcfn.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/rpcfn.h new file mode 100644 index 000000000000..2d67b598c58b --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/nvrm/rpcfn.h @@ -0,0 +1,249 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. */ + +#ifndef __NVRM_RPCFN_H__ +#define __NVRM_RPCFN_H__ +#include <nvrm/nvtypes.h> + +/* Excerpt of RM headers from https://github.com/NVIDIA/open-gpu-kernel-modules/tree/570.144 */ + +#ifndef X +# define X(UNIT, RPC, VAL) NV_VGPU_MSG_FUNCTION_##RPC = VAL, +# define DEFINING_X_IN_RPC_GLOBAL_ENUMS_H +enum { +#endif + X(RM, NOP, 0) + X(RM, SET_GUEST_SYSTEM_INFO, 1) + X(RM, ALLOC_ROOT, 2) + X(RM, ALLOC_DEVICE, 3) // deprecated + X(RM, ALLOC_MEMORY, 4) + X(RM, ALLOC_CTX_DMA, 5) + X(RM, ALLOC_CHANNEL_DMA, 6) + X(RM, MAP_MEMORY, 7) + X(RM, BIND_CTX_DMA, 8) // deprecated + X(RM, ALLOC_OBJECT, 9) + X(RM, FREE, 10) + X(RM, LOG, 11) + X(RM, ALLOC_VIDMEM, 12) + X(RM, UNMAP_MEMORY, 13) + X(RM, MAP_MEMORY_DMA, 14) + X(RM, UNMAP_MEMORY_DMA, 15) + X(RM, GET_EDID, 16) // deprecated + X(RM, ALLOC_DISP_CHANNEL, 17) + X(RM, ALLOC_DISP_OBJECT, 18) + X(RM, ALLOC_SUBDEVICE, 19) + X(RM, ALLOC_DYNAMIC_MEMORY, 20) + X(RM, DUP_OBJECT, 21) + X(RM, IDLE_CHANNELS, 22) + X(RM, ALLOC_EVENT, 23) + X(RM, SEND_EVENT, 24) // deprecated + X(RM, REMAPPER_CONTROL, 25) // deprecated + X(RM, DMA_CONTROL, 26) // deprecated + X(RM, DMA_FILL_PTE_MEM, 27) + X(RM, MANAGE_HW_RESOURCE, 28) + X(RM, BIND_ARBITRARY_CTX_DMA, 29) // deprecated + X(RM, CREATE_FB_SEGMENT, 30) + X(RM, DESTROY_FB_SEGMENT, 31) + X(RM, ALLOC_SHARE_DEVICE, 32) + X(RM, DEFERRED_API_CONTROL, 33) + X(RM, REMOVE_DEFERRED_API, 34) + X(RM, SIM_ESCAPE_READ, 35) + X(RM, SIM_ESCAPE_WRITE, 36) + X(RM, SIM_MANAGE_DISPLAY_CONTEXT_DMA, 37) + X(RM, FREE_VIDMEM_VIRT, 38) + X(RM, PERF_GET_PSTATE_INFO, 39) // deprecated + X(RM, PERF_GET_PERFMON_SAMPLE, 40) + X(RM, PERF_GET_VIRTUAL_PSTATE_INFO, 41) // deprecated + X(RM, PERF_GET_LEVEL_INFO, 42) + X(RM, MAP_SEMA_MEMORY, 43) + X(RM, UNMAP_SEMA_MEMORY, 44) + X(RM, SET_SURFACE_PROPERTIES, 45) + X(RM, CLEANUP_SURFACE, 46) + X(RM, UNLOADING_GUEST_DRIVER, 47) + X(RM, TDR_SET_TIMEOUT_STATE, 48) + X(RM, SWITCH_TO_VGA, 49) + X(RM, GPU_EXEC_REG_OPS, 50) + X(RM, GET_STATIC_INFO, 51) + X(RM, ALLOC_VIRTMEM, 52) + X(RM, UPDATE_PDE_2, 53) + X(RM, SET_PAGE_DIRECTORY, 54) + X(RM, GET_STATIC_PSTATE_INFO, 55) + X(RM, TRANSLATE_GUEST_GPU_PTES, 56) + X(RM, RESERVED_57, 57) + X(RM, RESET_CURRENT_GR_CONTEXT, 58) + X(RM, SET_SEMA_MEM_VALIDATION_STATE, 59) + X(RM, GET_ENGINE_UTILIZATION, 60) + X(RM, UPDATE_GPU_PDES, 61) + X(RM, GET_ENCODER_CAPACITY, 62) + X(RM, VGPU_PF_REG_READ32, 63) // deprecated + X(RM, SET_GUEST_SYSTEM_INFO_EXT, 64) + X(GSP, GET_GSP_STATIC_INFO, 65) + X(RM, RMFS_INIT, 66) // deprecated + X(RM, RMFS_CLOSE_QUEUE, 67) // deprecated + X(RM, RMFS_CLEANUP, 68) // deprecated + X(RM, RMFS_TEST, 69) // deprecated + X(RM, UPDATE_BAR_PDE, 70) + X(RM, CONTINUATION_RECORD, 71) + X(RM, GSP_SET_SYSTEM_INFO, 72) + X(RM, SET_REGISTRY, 73) + X(GSP, GSP_INIT_POST_OBJGPU, 74) // deprecated + X(RM, SUBDEV_EVENT_SET_NOTIFICATION, 75) // deprecated + X(GSP, GSP_RM_CONTROL, 76) + X(RM, GET_STATIC_INFO2, 77) + X(RM, DUMP_PROTOBUF_COMPONENT, 78) + X(RM, UNSET_PAGE_DIRECTORY, 79) + X(RM, GET_CONSOLIDATED_STATIC_INFO, 80) // deprecated + X(RM, GMMU_REGISTER_FAULT_BUFFER, 81) // deprecated + X(RM, GMMU_UNREGISTER_FAULT_BUFFER, 82) // deprecated + X(RM, GMMU_REGISTER_CLIENT_SHADOW_FAULT_BUFFER, 83) // deprecated + X(RM, GMMU_UNREGISTER_CLIENT_SHADOW_FAULT_BUFFER, 84) // deprecated + X(RM, CTRL_SET_VGPU_FB_USAGE, 85) + X(RM, CTRL_NVFBC_SW_SESSION_UPDATE_INFO, 86) + X(RM, CTRL_NVENC_SW_SESSION_UPDATE_INFO, 87) + X(RM, CTRL_RESET_CHANNEL, 88) + X(RM, CTRL_RESET_ISOLATED_CHANNEL, 89) + X(RM, CTRL_GPU_HANDLE_VF_PRI_FAULT, 90) + X(RM, CTRL_CLK_GET_EXTENDED_INFO, 91) + X(RM, CTRL_PERF_BOOST, 92) + X(RM, CTRL_PERF_VPSTATES_GET_CONTROL, 93) + X(RM, CTRL_GET_ZBC_CLEAR_TABLE, 94) + X(RM, CTRL_SET_ZBC_COLOR_CLEAR, 95) + X(RM, CTRL_SET_ZBC_DEPTH_CLEAR, 96) + X(RM, CTRL_GPFIFO_SCHEDULE, 97) + X(RM, CTRL_SET_TIMESLICE, 98) + X(RM, CTRL_PREEMPT, 99) + X(RM, CTRL_FIFO_DISABLE_CHANNELS, 100) + X(RM, CTRL_SET_TSG_INTERLEAVE_LEVEL, 101) + X(RM, CTRL_SET_CHANNEL_INTERLEAVE_LEVEL, 102) + X(GSP, GSP_RM_ALLOC, 103) + X(RM, CTRL_GET_P2P_CAPS_V2, 104) + X(RM, CTRL_CIPHER_AES_ENCRYPT, 105) + X(RM, CTRL_CIPHER_SESSION_KEY, 106) + X(RM, CTRL_CIPHER_SESSION_KEY_STATUS, 107) + X(RM, CTRL_DBG_CLEAR_ALL_SM_ERROR_STATES, 108) + X(RM, CTRL_DBG_READ_ALL_SM_ERROR_STATES, 109) + X(RM, CTRL_DBG_SET_EXCEPTION_MASK, 110) + X(RM, CTRL_GPU_PROMOTE_CTX, 111) + X(RM, CTRL_GR_CTXSW_PREEMPTION_BIND, 112) + X(RM, CTRL_GR_SET_CTXSW_PREEMPTION_MODE, 113) + X(RM, CTRL_GR_CTXSW_ZCULL_BIND, 114) + X(RM, CTRL_GPU_INITIALIZE_CTX, 115) + X(RM, CTRL_VASPACE_COPY_SERVER_RESERVED_PDES, 116) + X(RM, CTRL_FIFO_CLEAR_FAULTED_BIT, 117) + X(RM, CTRL_GET_LATEST_ECC_ADDRESSES, 118) + X(RM, CTRL_MC_SERVICE_INTERRUPTS, 119) + X(RM, CTRL_DMA_SET_DEFAULT_VASPACE, 120) + X(RM, CTRL_GET_CE_PCE_MASK, 121) + X(RM, CTRL_GET_ZBC_CLEAR_TABLE_ENTRY, 122) + X(RM, CTRL_GET_NVLINK_PEER_ID_MASK, 123) // deprecated + X(RM, CTRL_GET_NVLINK_STATUS, 124) + X(RM, CTRL_GET_P2P_CAPS, 125) + X(RM, CTRL_GET_P2P_CAPS_MATRIX, 126) + X(RM, RESERVED_0, 127) + X(RM, CTRL_RESERVE_PM_AREA_SMPC, 128) + X(RM, CTRL_RESERVE_HWPM_LEGACY, 129) + X(RM, CTRL_B0CC_EXEC_REG_OPS, 130) + X(RM, CTRL_BIND_PM_RESOURCES, 131) + X(RM, CTRL_DBG_SUSPEND_CONTEXT, 132) + X(RM, CTRL_DBG_RESUME_CONTEXT, 133) + X(RM, CTRL_DBG_EXEC_REG_OPS, 134) + X(RM, CTRL_DBG_SET_MODE_MMU_DEBUG, 135) + X(RM, CTRL_DBG_READ_SINGLE_SM_ERROR_STATE, 136) + X(RM, CTRL_DBG_CLEAR_SINGLE_SM_ERROR_STATE, 137) + X(RM, CTRL_DBG_SET_MODE_ERRBAR_DEBUG, 138) + X(RM, CTRL_DBG_SET_NEXT_STOP_TRIGGER_TYPE, 139) + X(RM, CTRL_ALLOC_PMA_STREAM, 140) + X(RM, CTRL_PMA_STREAM_UPDATE_GET_PUT, 141) + X(RM, CTRL_FB_GET_INFO_V2, 142) + X(RM, CTRL_FIFO_SET_CHANNEL_PROPERTIES, 143) + X(RM, CTRL_GR_GET_CTX_BUFFER_INFO, 144) + X(RM, CTRL_KGR_GET_CTX_BUFFER_PTES, 145) + X(RM, CTRL_GPU_EVICT_CTX, 146) + X(RM, CTRL_FB_GET_FS_INFO, 147) + X(RM, CTRL_GRMGR_GET_GR_FS_INFO, 148) + X(RM, CTRL_STOP_CHANNEL, 149) + X(RM, CTRL_GR_PC_SAMPLING_MODE, 150) + X(RM, CTRL_PERF_RATED_TDP_GET_STATUS, 151) + X(RM, CTRL_PERF_RATED_TDP_SET_CONTROL, 152) + X(RM, CTRL_FREE_PMA_STREAM, 153) + X(RM, CTRL_TIMER_SET_GR_TICK_FREQ, 154) + X(RM, CTRL_FIFO_SETUP_VF_ZOMBIE_SUBCTX_PDB, 155) + X(RM, GET_CONSOLIDATED_GR_STATIC_INFO, 156) + X(RM, CTRL_DBG_SET_SINGLE_SM_SINGLE_STEP, 157) + X(RM, CTRL_GR_GET_TPC_PARTITION_MODE, 158) + X(RM, CTRL_GR_SET_TPC_PARTITION_MODE, 159) + X(UVM, UVM_PAGING_CHANNEL_ALLOCATE, 160) + X(UVM, UVM_PAGING_CHANNEL_DESTROY, 161) + X(UVM, UVM_PAGING_CHANNEL_MAP, 162) + X(UVM, UVM_PAGING_CHANNEL_UNMAP, 163) + X(UVM, UVM_PAGING_CHANNEL_PUSH_STREAM, 164) + X(UVM, UVM_PAGING_CHANNEL_SET_HANDLES, 165) + X(UVM, UVM_METHOD_STREAM_GUEST_PAGES_OPERATION, 166) + X(RM, CTRL_INTERNAL_QUIESCE_PMA_CHANNEL, 167) + X(RM, DCE_RM_INIT, 168) + X(RM, REGISTER_VIRTUAL_EVENT_BUFFER, 169) + X(RM, CTRL_EVENT_BUFFER_UPDATE_GET, 170) + X(RM, GET_PLCABLE_ADDRESS_KIND, 171) + X(RM, CTRL_PERF_LIMITS_SET_STATUS_V2, 172) + X(RM, CTRL_INTERNAL_SRIOV_PROMOTE_PMA_STREAM, 173) + X(RM, CTRL_GET_MMU_DEBUG_MODE, 174) + X(RM, CTRL_INTERNAL_PROMOTE_FAULT_METHOD_BUFFERS, 175) + X(RM, CTRL_FLCN_GET_CTX_BUFFER_SIZE, 176) + X(RM, CTRL_FLCN_GET_CTX_BUFFER_INFO, 177) + X(RM, DISABLE_CHANNELS, 178) + X(RM, CTRL_FABRIC_MEMORY_DESCRIBE, 179) + X(RM, CTRL_FABRIC_MEM_STATS, 180) + X(RM, SAVE_HIBERNATION_DATA, 181) + X(RM, RESTORE_HIBERNATION_DATA, 182) + X(RM, CTRL_INTERNAL_MEMSYS_SET_ZBC_REFERENCED, 183) + X(RM, CTRL_EXEC_PARTITIONS_CREATE, 184) + X(RM, CTRL_EXEC_PARTITIONS_DELETE, 185) + X(RM, CTRL_GPFIFO_GET_WORK_SUBMIT_TOKEN, 186) + X(RM, CTRL_GPFIFO_SET_WORK_SUBMIT_TOKEN_NOTIF_INDEX, 187) + X(RM, PMA_SCRUBBER_SHARED_BUFFER_GUEST_PAGES_OPERATION, 188) + X(RM, CTRL_MASTER_GET_VIRTUAL_FUNCTION_ERROR_CONT_INTR_MASK, 189) + X(RM, SET_SYSMEM_DIRTY_PAGE_TRACKING_BUFFER, 190) + X(RM, CTRL_SUBDEVICE_GET_P2P_CAPS, 191) + X(RM, CTRL_BUS_SET_P2P_MAPPING, 192) + X(RM, CTRL_BUS_UNSET_P2P_MAPPING, 193) + X(RM, CTRL_FLA_SETUP_INSTANCE_MEM_BLOCK, 194) + X(RM, CTRL_GPU_MIGRATABLE_OPS, 195) + X(RM, CTRL_GET_TOTAL_HS_CREDITS, 196) + X(RM, CTRL_GET_HS_CREDITS, 197) + X(RM, CTRL_SET_HS_CREDITS, 198) + X(RM, CTRL_PM_AREA_PC_SAMPLER, 199) + X(RM, INVALIDATE_TLB, 200) + X(RM, CTRL_GPU_QUERY_ECC_STATUS, 201) // deprecated + X(RM, ECC_NOTIFIER_WRITE_ACK, 202) + X(RM, CTRL_DBG_GET_MODE_MMU_DEBUG, 203) + X(RM, RM_API_CONTROL, 204) + X(RM, CTRL_CMD_INTERNAL_GPU_START_FABRIC_PROBE, 205) + X(RM, CTRL_NVLINK_GET_INBAND_RECEIVED_DATA, 206) + X(RM, GET_STATIC_DATA, 207) + X(RM, RESERVED_208, 208) + X(RM, CTRL_GPU_GET_INFO_V2, 209) + X(RM, GET_BRAND_CAPS, 210) + X(RM, CTRL_CMD_NVLINK_INBAND_SEND_DATA, 211) + X(RM, UPDATE_GPM_GUEST_BUFFER_INFO, 212) + X(RM, CTRL_CMD_INTERNAL_CONTROL_GSP_TRACE, 213) + X(RM, CTRL_SET_ZBC_STENCIL_CLEAR, 214) + X(RM, CTRL_SUBDEVICE_GET_VGPU_HEAP_STATS, 215) + X(RM, CTRL_SUBDEVICE_GET_LIBOS_HEAP_STATS, 216) + X(RM, CTRL_DBG_SET_MODE_MMU_GCC_DEBUG, 217) + X(RM, CTRL_DBG_GET_MODE_MMU_GCC_DEBUG, 218) + X(RM, CTRL_RESERVE_HES, 219) + X(RM, CTRL_RELEASE_HES, 220) + X(RM, CTRL_RESERVE_CCU_PROF, 221) + X(RM, CTRL_RELEASE_CCU_PROF, 222) + X(RM, RESERVED, 223) + X(RM, CTRL_CMD_GET_CHIPLET_HS_CREDIT_POOL, 224) + X(RM, CTRL_CMD_GET_HS_CREDITS_MAPPING, 225) + X(RM, CTRL_EXEC_PARTITIONS_EXPORT, 226) + X(RM, NUM_FUNCTIONS, 227) +#ifdef DEFINING_X_IN_RPC_GLOBAL_ENUMS_H +}; +# undef X +# undef DEFINING_X_IN_RPC_GLOBAL_ENUMS_H +#endif +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/ofa.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/ofa.c new file mode 100644 index 000000000000..6fb3083edde3 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/ofa.c @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/engine.h> + +#include "nvrm/ofa.h" + +static int +r570_ofa_alloc(struct nvkm_gsp_object *parent, u32 handle, u32 oclass, int inst, + struct nvkm_gsp_object *ofa) +{ + NV_OFA_ALLOCATION_PARAMETERS *args; + + args = nvkm_gsp_rm_alloc_get(parent, handle, oclass, sizeof(*args), ofa); + if (WARN_ON(IS_ERR(args))) + return PTR_ERR(args); + + args->size = sizeof(*args); + args->engineInstance = inst; + + return nvkm_gsp_rm_alloc_wr(ofa, args); +} + +const struct nvkm_rm_api_engine +r570_ofa = { + .alloc = r570_ofa_alloc, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/rm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/rm.c new file mode 100644 index 000000000000..498658d0c60c --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r570/rm.c @@ -0,0 +1,99 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <rm/rm.h> + +#include "nvrm/gsp.h" + +static const struct nvkm_rm_wpr +r570_wpr_libos2 = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS2, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS2_MIN_MB, +}; + +static const struct nvkm_rm_wpr +r570_wpr_libos3 = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3_BAREMETAL, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB, +}; + +static const struct nvkm_rm_wpr +r570_wpr_libos3_gh100 = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3_BAREMETAL, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB, + .heap_size_non_wpr = 0x200000, + .offset_set_by_acr = true, +}; + +static const struct nvkm_rm_wpr +r570_wpr_libos3_gb10x = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3_BAREMETAL, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB, + .heap_size_non_wpr = 0x200000, + .rsvd_size_pmu = ALIGN(0x0800000 + 0x1000000 + 0x0001000, 0x20000), + .offset_set_by_acr = true, +}; + +static const struct nvkm_rm_wpr +r570_wpr_libos3_gb20x = { + .os_carveout_size = GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3_BAREMETAL, + .base_size = GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100, + .heap_size_min = GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB, + .heap_size_non_wpr = 0x220000, + .rsvd_size_pmu = ALIGN(0x0800000 + 0x1000000 + 0x0001000, 0x20000), + .offset_set_by_acr = true, +}; + +static const struct nvkm_rm_api +r570_api = { + .gsp = &r570_gsp, + .rpc = &r535_rpc, + .ctrl = &r535_ctrl, + .alloc = &r535_alloc, + .client = &r570_client, + .device = &r535_device, + .fbsr = &r570_fbsr, + .disp = &r570_disp, + .fifo = &r570_fifo, + .ce = &r535_ce, + .gr = &r570_gr, + .nvdec = &r535_nvdec, + .nvenc = &r535_nvenc, + .nvjpg = &r535_nvjpg, + .ofa = &r570_ofa, +}; + +const struct nvkm_rm_impl +r570_rm_tu102 = { + .wpr = &r570_wpr_libos2, + .api = &r570_api, +}; + +const struct nvkm_rm_impl +r570_rm_ga102 = { + .wpr = &r570_wpr_libos3, + .api = &r570_api, +}; + +const struct nvkm_rm_impl +r570_rm_gh100 = { + .wpr = &r570_wpr_libos3_gh100, + .api = &r570_api, +}; + +const struct nvkm_rm_impl +r570_rm_gb10x = { + .wpr = &r570_wpr_libos3_gb10x, + .api = &r570_api, +}; + +const struct nvkm_rm_impl +r570_rm_gb20x = { + .wpr = &r570_wpr_libos3_gb20x, + .api = &r570_api, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/rm.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/rm.h new file mode 100644 index 000000000000..393ea775941f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/rm.h @@ -0,0 +1,191 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include <subdev/gsp.h> +#ifndef __NVKM_RM_H__ +#define __NVKM_RM_H__ +#include "handles.h" +struct nvkm_outp; +struct r535_gr; + +struct nvkm_rm_impl { + const struct nvkm_rm_wpr *wpr; + const struct nvkm_rm_api *api; +}; + +struct nvkm_rm { + struct nvkm_device *device; + const struct nvkm_rm_gpu *gpu; + const struct nvkm_rm_wpr *wpr; + const struct nvkm_rm_api *api; +}; + +struct nvkm_rm_wpr { + u32 os_carveout_size; + u32 base_size; + u64 heap_size_min; + u32 heap_size_non_wpr; + u32 rsvd_size_pmu; + bool offset_set_by_acr; +}; + +struct nvkm_rm_api { + const struct nvkm_rm_api_gsp { + void (*set_rmargs)(struct nvkm_gsp *, bool resume); + int (*set_system_info)(struct nvkm_gsp *); + int (*get_static_info)(struct nvkm_gsp *); + bool (*xlat_mc_engine_idx)(u32 mc_engine_idx, enum nvkm_subdev_type *, int *inst); + void (*drop_send_user_shared_data)(struct nvkm_gsp *); + void (*drop_post_nocat_record)(struct nvkm_gsp *); + u32 (*sr_data_size)(struct nvkm_gsp *); + } *gsp; + + const struct nvkm_rm_api_rpc { + void *(*get)(struct nvkm_gsp *, u32 fn, u32 argc); + void *(*push)(struct nvkm_gsp *gsp, void *argv, + enum nvkm_gsp_rpc_reply_policy policy, u32 repc); + void (*done)(struct nvkm_gsp *gsp, void *repv); + } *rpc; + + const struct nvkm_rm_api_ctrl { + void *(*get)(struct nvkm_gsp_object *, u32 cmd, u32 params_size); + int (*push)(struct nvkm_gsp_object *, void **params, u32 repc); + void (*done)(struct nvkm_gsp_object *, void *params); + } *ctrl; + + const struct nvkm_rm_api_alloc { + void *(*get)(struct nvkm_gsp_object *, u32 oclass, u32 params_size); + void *(*push)(struct nvkm_gsp_object *, void *params); + void (*done)(struct nvkm_gsp_object *, void *params); + + int (*free)(struct nvkm_gsp_object *); + } *alloc; + + const struct nvkm_rm_api_client { + int (*ctor)(struct nvkm_gsp_client *, u32 handle); + } *client; + + const struct nvkm_rm_api_device { + int (*ctor)(struct nvkm_gsp_client *, struct nvkm_gsp_device *); + void (*dtor)(struct nvkm_gsp_device *); + + struct { + int (*ctor)(struct nvkm_gsp_device *, u32 handle, u32 id, + nvkm_gsp_event_func, struct nvkm_gsp_event *); + void (*dtor)(struct nvkm_gsp_event *); + } event; + } *device; + + const struct nvkm_rm_api_fbsr { + int (*suspend)(struct nvkm_gsp *); + void (*resume)(struct nvkm_gsp *); + } *fbsr; + + const struct nvkm_rm_api_disp { + int (*get_static_info)(struct nvkm_disp *); + int (*get_supported)(struct nvkm_disp *, unsigned long *display_mask); + int (*get_connect_state)(struct nvkm_disp *, unsigned display_id); + int (*get_active)(struct nvkm_disp *, unsigned head, u32 *display_id); + + int (*bl_ctrl)(struct nvkm_disp *, unsigned display_id, bool set, int *val); + + struct { + int (*get_caps)(struct nvkm_disp *, int *link_bw, bool *mst, bool *wm); + int (*set_indexed_link_rates)(struct nvkm_outp *); + } dp; + + struct { + int (*set_pushbuf)(struct nvkm_disp *, s32 oclass, int inst, + struct nvkm_memory *); + int (*dmac_alloc)(struct nvkm_disp *, u32 oclass, int inst, u32 put_offset, + struct nvkm_gsp_object *); + } chan; + } *disp; + + const struct nvkm_rm_api_fifo { + int (*xlat_rm_engine_type)(u32 rm_engine_type, + enum nvkm_subdev_type *, int *nv2080_type); + int (*ectx_size)(struct nvkm_fifo *); + unsigned rsvd_chids; + int (*rc_triggered)(void *priv, u32 fn, void *repv, u32 repc); + struct { + int (*alloc)(struct nvkm_gsp_device *, u32 handle, + u32 nv2080_engine_type, u8 runq, bool priv, int chid, + u64 inst_addr, u64 userd_addr, u64 mthdbuf_addr, + struct nvkm_vmm *, u64 gpfifo_offset, u32 gpfifo_length, + struct nvkm_gsp_object *); + } chan; + } *fifo; + + const struct nvkm_rm_api_engine { + int (*alloc)(struct nvkm_gsp_object *chan, u32 handle, u32 class, int inst, + struct nvkm_gsp_object *); + } *ce, *nvdec, *nvenc, *nvjpg, *ofa; + + const struct nvkm_rm_api_gr { + int (*get_ctxbufs_info)(struct r535_gr *); + struct { + int (*init)(struct r535_gr *); + void (*fini)(struct r535_gr *); + } scrubber; + } *gr; +}; + +extern const struct nvkm_rm_impl r535_rm_tu102; +extern const struct nvkm_rm_impl r535_rm_ga102; +extern const struct nvkm_rm_api_gsp r535_gsp; +typedef struct DOD_METHOD_DATA DOD_METHOD_DATA; +typedef struct JT_METHOD_DATA JT_METHOD_DATA; +typedef struct CAPS_METHOD_DATA CAPS_METHOD_DATA; +void r535_gsp_acpi_dod(acpi_handle, DOD_METHOD_DATA *); +void r535_gsp_acpi_jt(acpi_handle, JT_METHOD_DATA *); +void r535_gsp_acpi_caps(acpi_handle, CAPS_METHOD_DATA *); +struct NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS; +void r535_gsp_get_static_info_fb(struct nvkm_gsp *, + const struct NV2080_CTRL_CMD_FB_GET_FB_REGION_INFO_PARAMS *); +extern const struct nvkm_rm_api_rpc r535_rpc; +extern const struct nvkm_rm_api_ctrl r535_ctrl; +extern const struct nvkm_rm_api_alloc r535_alloc; +extern const struct nvkm_rm_api_client r535_client; +void r535_gsp_client_dtor(struct nvkm_gsp_client *); +extern const struct nvkm_rm_api_device r535_device; +int r535_mmu_vaspace_new(struct nvkm_vmm *, u32 handle, bool external); +void r535_mmu_vaspace_del(struct nvkm_vmm *); +extern const struct nvkm_rm_api_fbsr r535_fbsr; +void r535_fbsr_resume(struct nvkm_gsp *); +int r535_fbsr_memlist(struct nvkm_gsp_device *, u32 handle, enum nvkm_memory_target, + u64 phys, u64 size, struct sg_table *, struct nvkm_gsp_object *); +extern const struct nvkm_rm_api_disp r535_disp; +extern const struct nvkm_rm_api_fifo r535_fifo; +void r535_fifo_rc_chid(struct nvkm_fifo *, int chid); +extern const struct nvkm_rm_api_engine r535_ce; +extern const struct nvkm_rm_api_gr r535_gr; +void *r535_gr_dtor(struct nvkm_gr *); +int r535_gr_oneinit(struct nvkm_gr *); +u64 r535_gr_units(struct nvkm_gr *); +int r535_gr_chan_new(struct nvkm_gr *, struct nvkm_chan *, const struct nvkm_oclass *, + struct nvkm_object **); +int r535_gr_promote_ctx(struct r535_gr *, bool golden, struct nvkm_vmm *, + struct nvkm_memory **pctxbuf_mem, struct nvkm_vma **pctxbuf_vma, + struct nvkm_gsp_object *chan); +extern const struct nvkm_rm_api_engine r535_nvdec; +extern const struct nvkm_rm_api_engine r535_nvenc; +extern const struct nvkm_rm_api_engine r535_nvjpg; +extern const struct nvkm_rm_api_engine r535_ofa; + +extern const struct nvkm_rm_impl r570_rm_tu102; +extern const struct nvkm_rm_impl r570_rm_ga102; +extern const struct nvkm_rm_impl r570_rm_gh100; +extern const struct nvkm_rm_impl r570_rm_gb10x; +extern const struct nvkm_rm_impl r570_rm_gb20x; +extern const struct nvkm_rm_api_gsp r570_gsp; +extern const struct nvkm_rm_api_client r570_client; +extern const struct nvkm_rm_api_fbsr r570_fbsr; +extern const struct nvkm_rm_api_disp r570_disp; +extern const struct nvkm_rm_api_fifo r570_fifo; +extern const struct nvkm_rm_api_gr r570_gr; +int r570_gr_gpc_mask(struct nvkm_gsp *, u32 *mask); +int r570_gr_tpc_mask(struct nvkm_gsp *, int gpc, u32 *mask); +extern const struct nvkm_rm_api_engine r570_ofa; +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/rpc.h b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/rpc.h new file mode 100644 index 000000000000..4431e33b3304 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/rpc.h @@ -0,0 +1,18 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#ifndef __NVKM_RM_RPC_H__ +#define __NVKM_RM_RPC_H__ +#include "rm.h" + +#define to_payload_hdr(p, header) \ + container_of((void *)p, typeof(*header), params) + +int r535_gsp_rpc_poll(struct nvkm_gsp *, u32 fn); + +struct nvfw_gsp_rpc *r535_gsp_msg_recv(struct nvkm_gsp *, int fn, u32 gsp_rpc_len); +int r535_gsp_msg_ntfy_add(struct nvkm_gsp *, u32 fn, nvkm_gsp_msg_ntfy_func, void *priv); + +int r535_rpc_status_to_errno(uint32_t rpc_status); +#endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/tu1xx.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/tu1xx.c new file mode 100644 index 000000000000..423502f870db --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/tu1xx.c @@ -0,0 +1,38 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "gpu.h" + +#include <engine/fifo/priv.h> + +#include <nvif/class.h> + +const struct nvkm_rm_gpu +tu1xx_gpu = { + .disp.class = { + .root = TU102_DISP, + .caps = GV100_DISP_CAPS, + .core = TU102_DISP_CORE_CHANNEL_DMA, + .wndw = TU102_DISP_WINDOW_CHANNEL_DMA, + .wimm = TU102_DISP_WINDOW_IMM_CHANNEL_DMA, + .curs = TU102_DISP_CURSOR, + }, + + .usermode.class = TURING_USERMODE_A, + + .fifo.chan = { + .class = TURING_CHANNEL_GPFIFO_A, + .doorbell_handle = tu102_chan_doorbell_handle, + }, + + .ce.class = TURING_DMA_COPY_A, + .gr.class = { + .i2m = KEPLER_INLINE_TO_MEMORY_B, + .twod = FERMI_TWOD_A, + .threed = TURING_A, + .compute = TURING_COMPUTE_A, + }, + .nvdec.class = NVC4B0_VIDEO_DECODER, + .nvenc.class = NVC4B7_VIDEO_ENCODER, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c index 59c5f2b9172a..58e233bc53b1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c @@ -22,11 +22,45 @@ #include "priv.h" #include <subdev/fb.h> +#include <engine/sec2.h> + +#include <rm/r535/nvrm/gsp.h> #include <nvfw/flcn.h> #include <nvfw/fw.h> #include <nvfw/hs.h> +static int +tu102_gsp_booter_unload(struct nvkm_gsp *gsp, u32 mbox0, u32 mbox1) +{ + struct nvkm_subdev *subdev = &gsp->subdev; + struct nvkm_device *device = subdev->device; + u32 wpr2_hi; + int ret; + + wpr2_hi = nvkm_rd32(device, 0x1fa828); + if (!wpr2_hi) { + nvkm_debug(subdev, "WPR2 not set - skipping booter unload\n"); + return 0; + } + + ret = nvkm_falcon_fw_boot(&gsp->booter.unload, &gsp->subdev, true, &mbox0, &mbox1, 0, 0); + if (WARN_ON(ret)) + return ret; + + wpr2_hi = nvkm_rd32(device, 0x1fa828); + if (WARN_ON(wpr2_hi)) + return -EIO; + + return 0; +} + +static int +tu102_gsp_booter_load(struct nvkm_gsp *gsp, u32 mbox0, u32 mbox1) +{ + return nvkm_falcon_fw_boot(&gsp->booter.load, &gsp->subdev, true, &mbox0, &mbox1, 0, 0); +} + int tu102_gsp_booter_ctor(struct nvkm_gsp *gsp, const char *name, const struct firmware *blob, struct nvkm_falcon *falcon, struct nvkm_falcon_fw *fw) @@ -114,6 +148,118 @@ tu102_gsp_reset(struct nvkm_gsp *gsp) return gsp->falcon.func->reset_eng(&gsp->falcon); } +int +tu102_gsp_fini(struct nvkm_gsp *gsp, bool suspend) +{ + u32 mbox0 = 0xff, mbox1 = 0xff; + int ret; + + ret = r535_gsp_fini(gsp, suspend); + if (ret && suspend) + return ret; + + nvkm_falcon_reset(&gsp->falcon); + + ret = nvkm_gsp_fwsec_sb(gsp); + WARN_ON(ret); + + if (suspend) { + mbox0 = lower_32_bits(gsp->sr.meta.addr); + mbox1 = upper_32_bits(gsp->sr.meta.addr); + } + + ret = tu102_gsp_booter_unload(gsp, mbox0, mbox1); + WARN_ON(ret); + return 0; +} + +int +tu102_gsp_init(struct nvkm_gsp *gsp) +{ + u32 mbox0, mbox1; + int ret; + + if (!gsp->sr.meta.data) { + mbox0 = lower_32_bits(gsp->wpr_meta.addr); + mbox1 = upper_32_bits(gsp->wpr_meta.addr); + } else { + gsp->rm->api->gsp->set_rmargs(gsp, true); + + mbox0 = lower_32_bits(gsp->sr.meta.addr); + mbox1 = upper_32_bits(gsp->sr.meta.addr); + } + + /* Execute booter to handle (eventually...) booting GSP-RM. */ + ret = tu102_gsp_booter_load(gsp, mbox0, mbox1); + if (WARN_ON(ret)) + return ret; + + return r535_gsp_init(gsp); +} + +static int +tu102_gsp_wpr_meta_init(struct nvkm_gsp *gsp) +{ + GspFwWprMeta *meta; + int ret; + + ret = nvkm_gsp_mem_ctor(gsp, sizeof(*meta), &gsp->wpr_meta); + if (ret) + return ret; + + meta = gsp->wpr_meta.data; + + meta->magic = GSP_FW_WPR_META_MAGIC; + meta->revision = GSP_FW_WPR_META_REVISION; + + meta->sysmemAddrOfRadix3Elf = gsp->radix3.lvl0.addr; + meta->sizeOfRadix3Elf = gsp->fb.wpr2.elf.size; + + meta->sysmemAddrOfBootloader = gsp->boot.fw.addr; + meta->sizeOfBootloader = gsp->boot.fw.size; + meta->bootloaderCodeOffset = gsp->boot.code_offset; + meta->bootloaderDataOffset = gsp->boot.data_offset; + meta->bootloaderManifestOffset = gsp->boot.manifest_offset; + + meta->sysmemAddrOfSignature = gsp->sig.addr; + meta->sizeOfSignature = gsp->sig.size; + + meta->gspFwRsvdStart = gsp->fb.heap.addr; + meta->nonWprHeapOffset = gsp->fb.heap.addr; + meta->nonWprHeapSize = gsp->fb.heap.size; + meta->gspFwWprStart = gsp->fb.wpr2.addr; + meta->gspFwHeapOffset = gsp->fb.wpr2.heap.addr; + meta->gspFwHeapSize = gsp->fb.wpr2.heap.size; + meta->gspFwOffset = gsp->fb.wpr2.elf.addr; + meta->bootBinOffset = gsp->fb.wpr2.boot.addr; + meta->frtsOffset = gsp->fb.wpr2.frts.addr; + meta->frtsSize = gsp->fb.wpr2.frts.size; + meta->gspFwWprEnd = ALIGN_DOWN(gsp->fb.bios.vga_workspace.addr, 0x20000); + meta->fbSize = gsp->fb.size; + meta->vgaWorkspaceOffset = gsp->fb.bios.vga_workspace.addr; + meta->vgaWorkspaceSize = gsp->fb.bios.vga_workspace.size; + meta->bootCount = 0; + meta->partitionRpcAddr = 0; + meta->partitionRpcRequestOffset = 0; + meta->partitionRpcReplyOffset = 0; + meta->verified = 0; + return 0; +} + +u64 +tu102_gsp_wpr_heap_size(struct nvkm_gsp *gsp) +{ + u32 fb_size_gb = DIV_ROUND_UP_ULL(gsp->fb.size, 1 << 30); + u64 heap_size; + + heap_size = gsp->rm->wpr->os_carveout_size + + gsp->rm->wpr->base_size + + ALIGN(GSP_FW_HEAP_PARAM_SIZE_PER_GB_FB * fb_size_gb, 1 << 20) + + ALIGN(GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE, 1 << 20); + + return max(heap_size, gsp->rm->wpr->heap_size_min); +} + static u64 tu102_gsp_vga_workspace_addr(struct nvkm_gsp *gsp, u64 fb_size) { @@ -136,14 +282,67 @@ tu102_gsp_vga_workspace_addr(struct nvkm_gsp *gsp, u64 fb_size) int tu102_gsp_oneinit(struct nvkm_gsp *gsp) { - gsp->fb.size = nvkm_fb_vidmem_size(gsp->subdev.device); + struct nvkm_device *device = gsp->subdev.device; + int ret; + + gsp->fb.size = nvkm_fb_vidmem_size(device); gsp->fb.bios.vga_workspace.addr = tu102_gsp_vga_workspace_addr(gsp, gsp->fb.size); gsp->fb.bios.vga_workspace.size = gsp->fb.size - gsp->fb.bios.vga_workspace.addr; gsp->fb.bios.addr = gsp->fb.bios.vga_workspace.addr; gsp->fb.bios.size = gsp->fb.bios.vga_workspace.size; - return r535_gsp_oneinit(gsp); + ret = gsp->func->booter.ctor(gsp, "booter-load", gsp->fws.booter.load, + &device->sec2->falcon, &gsp->booter.load); + if (ret) + return ret; + + ret = gsp->func->booter.ctor(gsp, "booter-unload", gsp->fws.booter.unload, + &device->sec2->falcon, &gsp->booter.unload); + if (ret) + return ret; + + ret = r535_gsp_oneinit(gsp); + if (ret) + return ret; + + /* Calculate FB layout. */ + gsp->fb.wpr2.frts.size = 0x100000; + gsp->fb.wpr2.frts.addr = ALIGN_DOWN(gsp->fb.bios.addr, 0x20000) - gsp->fb.wpr2.frts.size; + + gsp->fb.wpr2.boot.size = gsp->boot.fw.size; + gsp->fb.wpr2.boot.addr = ALIGN_DOWN(gsp->fb.wpr2.frts.addr - gsp->fb.wpr2.boot.size, 0x1000); + + gsp->fb.wpr2.elf.size = gsp->fw.len; + gsp->fb.wpr2.elf.addr = ALIGN_DOWN(gsp->fb.wpr2.boot.addr - gsp->fb.wpr2.elf.size, 0x10000); + + gsp->fb.wpr2.heap.size = tu102_gsp_wpr_heap_size(gsp); + + gsp->fb.wpr2.heap.addr = ALIGN_DOWN(gsp->fb.wpr2.elf.addr - gsp->fb.wpr2.heap.size, 0x100000); + gsp->fb.wpr2.heap.size = ALIGN_DOWN(gsp->fb.wpr2.elf.addr - gsp->fb.wpr2.heap.addr, 0x100000); + + gsp->fb.wpr2.addr = ALIGN_DOWN(gsp->fb.wpr2.heap.addr - sizeof(GspFwWprMeta), 0x100000); + gsp->fb.wpr2.size = gsp->fb.wpr2.frts.addr + gsp->fb.wpr2.frts.size - gsp->fb.wpr2.addr; + + gsp->fb.heap.size = 0x100000; + gsp->fb.heap.addr = gsp->fb.wpr2.addr - gsp->fb.heap.size; + + ret = tu102_gsp_wpr_meta_init(gsp); + if (ret) + return ret; + + ret = nvkm_gsp_fwsec_frts(gsp); + if (WARN_ON(ret)) + return ret; + + /* Reset GSP into RISC-V mode. */ + ret = gsp->func->reset(gsp); + if (ret) + return ret; + + nvkm_falcon_wr32(&gsp->falcon, 0x040, lower_32_bits(gsp->libos.addr)); + nvkm_falcon_wr32(&gsp->falcon, 0x044, upper_32_bits(gsp->libos.addr)); + return 0; } const struct nvkm_falcon_func @@ -163,29 +362,73 @@ tu102_gsp_flcn = { }; static const struct nvkm_gsp_func -tu102_gsp_r535_113_01 = { +tu102_gsp = { .flcn = &tu102_gsp_flcn, .fwsec = &tu102_gsp_fwsec, .sig_section = ".fwsignature_tu10x", - .wpr_heap.base_size = 8 << 20, - .wpr_heap.min_size = 64 << 20, - .booter.ctor = tu102_gsp_booter_ctor, .dtor = r535_gsp_dtor, .oneinit = tu102_gsp_oneinit, - .init = r535_gsp_init, - .fini = r535_gsp_fini, + .init = tu102_gsp_init, + .fini = tu102_gsp_fini, .reset = tu102_gsp_reset, - .rm = &r535_gsp_rm, + .rm.gpu = &tu1xx_gpu, }; +int +tu102_gsp_load_rm(struct nvkm_gsp *gsp, const struct nvkm_gsp_fwif *fwif) +{ + struct nvkm_subdev *subdev = &gsp->subdev; + bool enable_gsp = fwif->enable; + int ret; + +#if IS_ENABLED(CONFIG_DRM_NOUVEAU_GSP_DEFAULT) + enable_gsp = true; +#endif + if (!nvkm_boolopt(subdev->device->cfgopt, "NvGspRm", enable_gsp)) + return -EINVAL; + + ret = nvkm_gsp_load_fw(gsp, "gsp", fwif->ver, &gsp->fws.rm); + if (ret) + return ret; + + ret = nvkm_gsp_load_fw(gsp, "bootloader", fwif->ver, &gsp->fws.bl); + if (ret) + return ret; + + return 0; +} + +int +tu102_gsp_load(struct nvkm_gsp *gsp, int ver, const struct nvkm_gsp_fwif *fwif) +{ + int ret; + + ret = tu102_gsp_load_rm(gsp, fwif); + if (ret) + goto done; + + ret = nvkm_gsp_load_fw(gsp, "booter_load", fwif->ver, &gsp->fws.booter.load); + if (ret) + goto done; + + ret = nvkm_gsp_load_fw(gsp, "booter_unload", fwif->ver, &gsp->fws.booter.unload); + +done: + if (ret) + nvkm_gsp_dtor_fws(gsp); + + return ret; +} + static struct nvkm_gsp_fwif tu102_gsps[] = { - { 0, r535_gsp_load, &tu102_gsp_r535_113_01, "535.113.01" }, + { 1, tu102_gsp_load, &tu102_gsp, &r570_rm_tu102, "570.144" }, + { 0, tu102_gsp_load, &tu102_gsp, &r535_rm_tu102, "535.113.01" }, { -1, gv100_gsp_nofw, &gv100_gsp }, {} }; @@ -196,3 +439,11 @@ tu102_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, { return nvkm_gsp_new_(tu102_gsps, device, type, inst, pgsp); } + +NVKM_GSP_FIRMWARE_BOOTER(tu102, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(tu104, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(tu106, 535.113.01); + +NVKM_GSP_FIRMWARE_BOOTER(tu102, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(tu104, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(tu106, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu116.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu116.c index 04fbd9ed28b1..97eb046c25d0 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu116.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu116.c @@ -22,29 +22,27 @@ #include "priv.h" static const struct nvkm_gsp_func -tu116_gsp_r535_113_01 = { +tu116_gsp = { .flcn = &tu102_gsp_flcn, .fwsec = &tu102_gsp_fwsec, .sig_section = ".fwsignature_tu11x", - .wpr_heap.base_size = 8 << 20, - .wpr_heap.min_size = 64 << 20, - .booter.ctor = tu102_gsp_booter_ctor, .dtor = r535_gsp_dtor, .oneinit = tu102_gsp_oneinit, - .init = r535_gsp_init, - .fini = r535_gsp_fini, + .init = tu102_gsp_init, + .fini = tu102_gsp_fini, .reset = tu102_gsp_reset, - .rm = &r535_gsp_rm, + .rm.gpu = &tu1xx_gpu, }; static struct nvkm_gsp_fwif tu116_gsps[] = { - { 0, r535_gsp_load, &tu116_gsp_r535_113_01, "535.113.01" }, + { 1, tu102_gsp_load, &tu116_gsp, &r570_rm_tu102, "570.144" }, + { 0, tu102_gsp_load, &tu116_gsp, &r535_rm_tu102, "535.113.01" }, { -1, gv100_gsp_nofw, &gv100_gsp }, {} }; @@ -55,3 +53,9 @@ tu116_gsp_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, { return nvkm_gsp_new_(tu116_gsps, device, type, inst, pgsp); } + +NVKM_GSP_FIRMWARE_BOOTER(tu116, 535.113.01); +NVKM_GSP_FIRMWARE_BOOTER(tu117, 535.113.01); + +NVKM_GSP_FIRMWARE_BOOTER(tu116, 570.144); +NVKM_GSP_FIRMWARE_BOOTER(tu117, 570.144); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/Kbuild index 553d540f2736..fa7a2862dd1f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/Kbuild @@ -4,5 +4,4 @@ nvkm-y += nvkm/subdev/instmem/nv04.o nvkm-y += nvkm/subdev/instmem/nv40.o nvkm-y += nvkm/subdev/instmem/nv50.o nvkm-y += nvkm/subdev/instmem/gk20a.o - -nvkm-y += nvkm/subdev/instmem/r535.o +nvkm-y += nvkm/subdev/instmem/gh100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c index a2cd3330efc6..2f55bab8e132 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/base.c @@ -182,9 +182,11 @@ nvkm_instmem_fini(struct nvkm_subdev *subdev, bool suspend) int ret; if (suspend) { - ret = imem->func->suspend(imem); - if (ret) - return ret; + if (imem->func->suspend) { + ret = imem->func->suspend(imem); + if (ret) + return ret; + } imem->suspend = true; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gh100.c new file mode 100644 index 000000000000..8d8dd5f8a6c7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gh100.c @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/ref/gh100/pri_nv_xal_ep.h> + +static void +gh100_instmem_set_bar0_window_addr(struct nvkm_device *device, u64 addr) +{ + nvkm_wr32(device, NV_XAL_EP_BAR0_WINDOW, addr >> NV_XAL_EP_BAR0_WINDOW_BASE_SHIFT); +} + +static const struct nvkm_instmem_func +gh100_instmem = { + .fini = nv50_instmem_fini, + .memory_new = nv50_instobj_new, + .memory_wrap = nv50_instobj_wrap, + .set_bar0_window_addr = gh100_instmem_set_bar0_window_addr, +}; + +int +gh100_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_instmem **pimem) +{ + return r535_instmem_new(&gh100_instmem, device, type, inst, pimem); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c index 6b462f960922..2544b9f0ec85 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c @@ -239,7 +239,6 @@ nv40_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int ins struct nvkm_instmem **pimem) { struct nv40_instmem *imem; - int bar; if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL))) return -ENOMEM; @@ -247,13 +246,8 @@ nv40_instmem_new(struct nvkm_device *device, enum nvkm_subdev_type type, int ins *pimem = &imem->base; /* map bar */ - if (device->func->resource_size(device, 2)) - bar = 2; - else - bar = 3; - - imem->iomem = ioremap_wc(device->func->resource_addr(device, bar), - device->func->resource_size(device, bar)); + imem->iomem = ioremap_wc(device->func->resource_addr(device, NVKM_BAR2_INST), + device->func->resource_size(device, NVKM_BAR2_INST)); if (!imem->iomem) { nvkm_error(&imem->base.subdev, "unable to map PRAMIN BAR\n"); return -EFAULT; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c index dd5b5a17ece0..4ca6fb30743d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c @@ -65,7 +65,7 @@ nv50_instobj_wr32_slow(struct nvkm_memory *memory, u64 offset, u32 data) spin_lock_irqsave(&imem->base.lock, flags); if (unlikely(imem->addr != base)) { - nvkm_wr32(device, 0x001700, base >> 16); + imem->base.func->set_bar0_window_addr(device, base); imem->addr = base; } nvkm_wr32(device, 0x700000 + addr, data); @@ -85,7 +85,7 @@ nv50_instobj_rd32_slow(struct nvkm_memory *memory, u64 offset) spin_lock_irqsave(&imem->base.lock, flags); if (unlikely(imem->addr != base)) { - nvkm_wr32(device, 0x001700, base >> 16); + imem->base.func->set_bar0_window_addr(device, base); imem->addr = base; } data = nvkm_rd32(device, 0x700000 + addr); @@ -172,7 +172,7 @@ nv50_instobj_kmap(struct nv50_instobj *iobj, struct nvkm_vmm *vmm) /* Make the mapping visible to the host. */ iobj->bar = bar; - iobj->map = ioremap_wc(device->func->resource_addr(device, 3) + + iobj->map = ioremap_wc(device->func->resource_addr(device, NVKM_BAR2_INST) + (u32)iobj->bar->addr, size); if (!iobj->map) { nvkm_warn(subdev, "PRAMIN ioremap failed\n"); @@ -353,7 +353,7 @@ nv50_instobj_func = { .map = nv50_instobj_map, }; -static int +int nv50_instobj_wrap(struct nvkm_instmem *base, struct nvkm_memory *memory, struct nvkm_memory **pmemory) { @@ -373,7 +373,7 @@ nv50_instobj_wrap(struct nvkm_instmem *base, return 0; } -static int +int nv50_instobj_new(struct nvkm_instmem *imem, u32 size, u32 align, bool zero, struct nvkm_memory **pmemory) { @@ -395,6 +395,12 @@ nv50_instobj_new(struct nvkm_instmem *imem, u32 size, u32 align, bool zero, *****************************************************************************/ static void +nv50_instmem_set_bar0_window_addr(struct nvkm_device *device, u64 addr) +{ + nvkm_wr32(device, 0x001700, addr >> 16); +} + +void nv50_instmem_fini(struct nvkm_instmem *base) { nv50_instmem(base)->addr = ~0ULL; @@ -415,6 +421,7 @@ nv50_instmem = { .memory_new = nv50_instobj_new, .memory_wrap = nv50_instobj_wrap, .zero = false, + .set_bar0_window_addr = nv50_instmem_set_bar0_window_addr, }; int diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h index 4c14c96fb60a..87bbdd786eaa 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/priv.h @@ -16,10 +16,16 @@ struct nvkm_instmem_func { bool zero, struct nvkm_memory **); int (*memory_wrap)(struct nvkm_instmem *, struct nvkm_memory *, struct nvkm_memory **); bool zero; + void (*set_bar0_window_addr)(struct nvkm_device *, u64 addr); }; int nv50_instmem_new_(const struct nvkm_instmem_func *, struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem **); +void nv50_instmem_fini(struct nvkm_instmem *); +int nv50_instobj_new(struct nvkm_instmem *, u32 size, u32 align, bool zero, + struct nvkm_memory **); +int nv50_instobj_wrap(struct nvkm_instmem *, struct nvkm_memory *vram, + struct nvkm_memory **bar2); void nvkm_instmem_ctor(const struct nvkm_instmem_func *, struct nvkm_device *, enum nvkm_subdev_type, int, struct nvkm_instmem *); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild index 7ba35ea59c06..ea4848931540 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild @@ -15,8 +15,7 @@ nvkm-y += nvkm/subdev/mmu/gp100.o nvkm-y += nvkm/subdev/mmu/gp10b.o nvkm-y += nvkm/subdev/mmu/gv100.o nvkm-y += nvkm/subdev/mmu/tu102.o - -nvkm-y += nvkm/subdev/mmu/r535.o +nvkm-y += nvkm/subdev/mmu/gh100.o nvkm-y += nvkm/subdev/mmu/mem.o nvkm-y += nvkm/subdev/mmu/memnv04.o @@ -38,6 +37,7 @@ nvkm-y += nvkm/subdev/mmu/vmmgp100.o nvkm-y += nvkm/subdev/mmu/vmmgp10b.o nvkm-y += nvkm/subdev/mmu/vmmgv100.o nvkm-y += nvkm/subdev/mmu/vmmtu102.o +nvkm-y += nvkm/subdev/mmu/vmmgh100.o nvkm-y += nvkm/subdev/mmu/umem.o nvkm-y += nvkm/subdev/mmu/ummu.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gh100.c new file mode 100644 index 000000000000..2918fb32cc91 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/gh100.c @@ -0,0 +1,25 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "mem.h" +#include "vmm.h" + +#include <nvif/class.h> + +static const struct nvkm_mmu_func +gh100_mmu = { + .dma_bits = 52, + .mmu = {{ -1, -1, NVIF_CLASS_MMU_GF100}}, + .mem = {{ -1, 0, NVIF_CLASS_MEM_GF100}, gf100_mem_new, gf100_mem_map }, + .vmm = {{ -1, 0, NVIF_CLASS_VMM_GP100}, gh100_vmm_new }, + .kind = tu102_mmu_kind, + .kind_sys = true, +}; + +int +gh100_mmu_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_mmu **pmmu) +{ + return r535_mmu_new(&gh100_mmu, device, type, inst, pmmu); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memgf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memgf100.c index d9c9bee45222..160a5749a29f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memgf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memgf100.c @@ -60,7 +60,7 @@ gf100_mem_map(struct nvkm_mmu *mmu, struct nvkm_memory *memory, void *argv, if (ret) return ret; - *paddr = device->func->resource_addr(device, 1) + (*pvma)->addr; + *paddr = device->func->resource_addr(device, NVKM_BAR1_FB) + (*pvma)->addr; *psize = (*pvma)->size; return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv04.c index 79a3b0cc9f5b..1e3db52de6cb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv04.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv04.c @@ -41,7 +41,7 @@ nv04_mem_map(struct nvkm_mmu *mmu, struct nvkm_memory *memory, void *argv, if ((ret = nvif_unvers(ret, &argv, &argc, args->vn))) return ret; - *paddr = device->func->resource_addr(device, 1) + addr; + *paddr = device->func->resource_addr(device, NVKM_BAR1_FB) + addr; *psize = nvkm_memory_size(memory); *pvma = ERR_PTR(-ENODEV); return 0; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv50.c index 46759b89fc1f..33b2321e9d87 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/memnv50.c @@ -57,7 +57,7 @@ nv50_mem_map(struct nvkm_mmu *mmu, struct nvkm_memory *memory, void *argv, if (ret) return ret; - *paddr = device->func->resource_addr(device, 1) + (*pvma)->addr; + *paddr = device->func->resource_addr(device, NVKM_BAR1_FB) + (*pvma)->addr; *psize = (*pvma)->size; return nvkm_memory_map(memory, 0, bar, *pvma, &uvmm, sizeof(uvmm)); } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h index e9ca6537778c..90efef8f0b54 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/priv.h @@ -53,6 +53,8 @@ const u8 *gf100_mmu_kind(struct nvkm_mmu *, int *count, u8 *invalid); const u8 *gm200_mmu_kind(struct nvkm_mmu *, int *, u8 *); +const u8 *tu102_mmu_kind(struct nvkm_mmu *, int *, u8 *); + struct nvkm_mmu_pt { union { struct nvkm_mmu_ptc *ptc; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c index df662ce4a4b0..7acff3642e20 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu102.c @@ -28,7 +28,7 @@ #include <nvif/class.h> -static const u8 * +const u8 * tu102_mmu_kind(struct nvkm_mmu *mmu, int *count, u8 *invalid) { static const u8 diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c index 9c97800fe037..f95c58b67633 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c @@ -19,7 +19,7 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. */ -#define NVKM_VMM_LEVELS_MAX 5 +#define NVKM_VMM_LEVELS_MAX 6 #include "vmm.h" #include <subdev/fb.h> @@ -1030,12 +1030,8 @@ nvkm_vmm_dtor(struct nvkm_vmm *vmm) struct nvkm_vma *vma; struct rb_node *node; - if (vmm->rm.client.gsp) { - nvkm_gsp_rm_free(&vmm->rm.object); - nvkm_gsp_device_dtor(&vmm->rm.device); - nvkm_gsp_client_dtor(&vmm->rm.client); - nvkm_vmm_put(vmm, &vmm->rm.rsvd); - } + if (vmm->rm.client.gsp) + r535_mmu_vaspace_del(vmm); if (0) nvkm_vmm_dump(vmm); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h index f9bc30cdb2b3..4586a425dbe4 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h @@ -143,6 +143,8 @@ struct nvkm_vmm_func { int (*aper)(enum nvkm_memory_target); int (*valid)(struct nvkm_vmm *, void *argv, u32 argc, struct nvkm_vmm_map *); + int (*valid2)(struct nvkm_vmm *, bool ro, bool priv, u8 kind, u8 comp, + struct nvkm_vmm_map *); void (*flush)(struct nvkm_vmm *, int depth); int (*mthd)(struct nvkm_vmm *, struct nvkm_client *, @@ -254,6 +256,8 @@ void gp100_vmm_invalidate_pdb(struct nvkm_vmm *, u64 addr); int gv100_vmm_join(struct nvkm_vmm *, struct nvkm_memory *); +void tu102_vmm_flush(struct nvkm_vmm *, int depth); + int nv04_vmm_new(struct nvkm_mmu *, bool, u64, u64, void *, u32, struct lock_class_key *, const char *, struct nvkm_vmm **); int nv41_vmm_new(struct nvkm_mmu *, bool, u64, u64, void *, u32, @@ -296,6 +300,9 @@ int gv100_vmm_new(struct nvkm_mmu *, bool, u64, u64, void *, u32, int tu102_vmm_new(struct nvkm_mmu *, bool, u64, u64, void *, u32, struct lock_class_key *, const char *, struct nvkm_vmm **); +int gh100_vmm_new(struct nvkm_mmu *, bool, u64, u64, void *, u32, + struct lock_class_key *, const char *, + struct nvkm_vmm **); #define VMM_PRINT(l,v,p,f,a...) do { \ struct nvkm_vmm *_vmm = (v); \ diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgh100.c new file mode 100644 index 000000000000..5614df3432da --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgh100.c @@ -0,0 +1,306 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "vmm.h" + +#include <subdev/fb.h> + +#include <nvhw/drf.h> +#include <nvhw/ref/gh100/dev_mmu.h> + +static inline void +gh100_vmm_pgt_pte(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, u32 ptei, u32 ptes, + struct nvkm_vmm_map *map, u64 addr) +{ + u64 data = addr | map->type; + + while (ptes--) { + VMM_WO064(pt, vmm, ptei++ * NV_MMU_VER3_PTE__SIZE, data); + data += map->next; + } +} + +static void +gh100_vmm_pgt_sgl(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, u32 ptei, u32 ptes, + struct nvkm_vmm_map *map) +{ + VMM_MAP_ITER_SGL(vmm, pt, ptei, ptes, map, gh100_vmm_pgt_pte); +} + +static void +gh100_vmm_pgt_dma(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, u32 ptei, u32 ptes, + struct nvkm_vmm_map *map) +{ + if (map->page->shift == PAGE_SHIFT) { + VMM_SPAM(vmm, "DMAA %08x %08x PTE(s)", ptei, ptes); + + nvkm_kmap(pt->memory); + while (ptes--) { + const u64 data = *map->dma++ | map->type; + + VMM_WO064(pt, vmm, ptei++ * NV_MMU_VER3_PTE__SIZE, data); + } + nvkm_done(pt->memory); + return; + } + + VMM_MAP_ITER_DMA(vmm, pt, ptei, ptes, map, gh100_vmm_pgt_pte); +} + +static void +gh100_vmm_pgt_mem(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, u32 ptei, u32 ptes, + struct nvkm_vmm_map *map) +{ + VMM_MAP_ITER_MEM(vmm, pt, ptei, ptes, map, gh100_vmm_pgt_pte); +} + +static void +gh100_vmm_pgt_sparse(struct nvkm_vmm *vmm, + struct nvkm_mmu_pt *pt, u32 ptei, u32 ptes) +{ + const u64 data = NVDEF(NV_MMU, VER3_PTE, PCF, SPARSE); + + VMM_FO064(pt, vmm, ptei * NV_MMU_VER3_PTE__SIZE, data, ptes); +} + +static const struct nvkm_vmm_desc_func +gh100_vmm_desc_spt = { + .unmap = gf100_vmm_pgt_unmap, + .sparse = gh100_vmm_pgt_sparse, + .mem = gh100_vmm_pgt_mem, + .dma = gh100_vmm_pgt_dma, + .sgl = gh100_vmm_pgt_sgl, +}; + +static void +gh100_vmm_lpt_invalid(struct nvkm_vmm *vmm, + struct nvkm_mmu_pt *pt, u32 ptei, u32 ptes) +{ + const u64 data = NVDEF(NV_MMU, VER3_PTE, PCF, NO_VALID_4KB_PAGE); + + VMM_FO064(pt, vmm, ptei * NV_MMU_VER3_PTE__SIZE, data, ptes); +} + +static const struct nvkm_vmm_desc_func +gh100_vmm_desc_lpt = { + .invalid = gh100_vmm_lpt_invalid, + .unmap = gf100_vmm_pgt_unmap, + .sparse = gh100_vmm_pgt_sparse, + .mem = gh100_vmm_pgt_mem, +}; + +static inline void +gh100_vmm_pd0_pte(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, + u32 ptei, u32 ptes, struct nvkm_vmm_map *map, u64 addr) +{ + u64 data = addr | map->type; + + while (ptes--) { + VMM_WO128(pt, vmm, ptei++ * NV_MMU_VER3_DUAL_PDE__SIZE, data, 0ULL); + data += map->next; + } +} + +static void +gh100_vmm_pd0_mem(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, + u32 ptei, u32 ptes, struct nvkm_vmm_map *map) +{ + VMM_MAP_ITER_MEM(vmm, pt, ptei, ptes, map, gh100_vmm_pd0_pte); +} + +static inline bool +gh100_vmm_pde(struct nvkm_mmu_pt *pt, u64 *data) +{ + switch (nvkm_memory_target(pt->memory)) { + case NVKM_MEM_TARGET_VRAM: + *data |= NVDEF(NV_MMU, VER3_PDE, APERTURE, VIDEO_MEMORY); + *data |= NVDEF(NV_MMU, VER3_PDE, PCF, VALID_CACHED_ATS_NOT_ALLOWED); + break; + case NVKM_MEM_TARGET_HOST: + *data |= NVDEF(NV_MMU, VER3_PDE, APERTURE, SYSTEM_COHERENT_MEMORY); + *data |= NVDEF(NV_MMU, VER3_PDE, PCF, VALID_UNCACHED_ATS_ALLOWED); + break; + case NVKM_MEM_TARGET_NCOH: + *data |= NVDEF(NV_MMU, VER3_PDE, APERTURE, SYSTEM_NON_COHERENT_MEMORY); + *data |= NVDEF(NV_MMU, VER3_PDE, PCF, VALID_CACHED_ATS_ALLOWED); + break; + default: + WARN_ON(1); + return false; + } + + *data |= pt->addr; + return true; +} + +static void +gh100_vmm_pd0_pde(struct nvkm_vmm *vmm, struct nvkm_vmm_pt *pgd, u32 pdei) +{ + struct nvkm_vmm_pt *pgt = pgd->pde[pdei]; + struct nvkm_mmu_pt *pd = pgd->pt[0]; + u64 data[2] = {}; + + if (pgt->pt[0] && !gh100_vmm_pde(pgt->pt[0], &data[0])) + return; + if (pgt->pt[1] && !gh100_vmm_pde(pgt->pt[1], &data[1])) + return; + + nvkm_kmap(pd->memory); + VMM_WO128(pd, vmm, pdei * NV_MMU_VER3_DUAL_PDE__SIZE, data[0], data[1]); + nvkm_done(pd->memory); +} + +static void +gh100_vmm_pd0_sparse(struct nvkm_vmm *vmm, + struct nvkm_mmu_pt *pt, u32 pdei, u32 pdes) +{ + const u64 data = NVDEF(NV_MMU, VER3_DUAL_PDE, PCF_BIG, SPARSE_ATS_ALLOWED); + + VMM_FO128(pt, vmm, pdei * NV_MMU_VER3_DUAL_PDE__SIZE, data, 0ULL, pdes); +} + +static void +gh100_vmm_pd0_unmap(struct nvkm_vmm *vmm, + struct nvkm_mmu_pt *pt, u32 pdei, u32 pdes) +{ + VMM_FO128(pt, vmm, pdei * NV_MMU_VER3_DUAL_PDE__SIZE, 0ULL, 0ULL, pdes); +} + +static const struct nvkm_vmm_desc_func +gh100_vmm_desc_pd0 = { + .unmap = gh100_vmm_pd0_unmap, + .sparse = gh100_vmm_pd0_sparse, + .pde = gh100_vmm_pd0_pde, + .mem = gh100_vmm_pd0_mem, +}; + +static void +gh100_vmm_pd1_pde(struct nvkm_vmm *vmm, struct nvkm_vmm_pt *pgd, u32 pdei) +{ + struct nvkm_vmm_pt *pgt = pgd->pde[pdei]; + struct nvkm_mmu_pt *pd = pgd->pt[0]; + u64 data = 0; + + if (!gh100_vmm_pde(pgt->pt[0], &data)) + return; + + nvkm_kmap(pd->memory); + VMM_WO064(pd, vmm, pdei * NV_MMU_VER3_PDE__SIZE, data); + nvkm_done(pd->memory); +} + +static const struct nvkm_vmm_desc_func +gh100_vmm_desc_pd1 = { + .unmap = gf100_vmm_pgt_unmap, + .sparse = gh100_vmm_pgt_sparse, + .pde = gh100_vmm_pd1_pde, +}; + +static const struct nvkm_vmm_desc +gh100_vmm_desc_16[] = { + { LPT, 5, 8, 0x0100, &gh100_vmm_desc_lpt }, + { PGD, 8, 16, 0x1000, &gh100_vmm_desc_pd0 }, + { PGD, 9, 8, 0x1000, &gh100_vmm_desc_pd1 }, + { PGD, 9, 8, 0x1000, &gh100_vmm_desc_pd1 }, + { PGD, 9, 8, 0x1000, &gh100_vmm_desc_pd1 }, + { PGD, 1, 8, 0x1000, &gh100_vmm_desc_pd1 }, + {} +}; + +static const struct nvkm_vmm_desc +gh100_vmm_desc_12[] = { + { SPT, 9, 8, 0x1000, &gh100_vmm_desc_spt }, + { PGD, 8, 16, 0x1000, &gh100_vmm_desc_pd0 }, + { PGD, 9, 8, 0x1000, &gh100_vmm_desc_pd1 }, + { PGD, 9, 8, 0x1000, &gh100_vmm_desc_pd1 }, + { PGD, 9, 8, 0x1000, &gh100_vmm_desc_pd1 }, + { PGD, 1, 8, 0x1000, &gh100_vmm_desc_pd1 }, + {} +}; + +static int +gh100_vmm_valid(struct nvkm_vmm *vmm, bool ro, bool priv, u8 kind, u8 comp, + struct nvkm_vmm_map *map) +{ + const enum nvkm_memory_target target = nvkm_memory_target(map->memory); + const bool vol = target == NVKM_MEM_TARGET_HOST; + const struct nvkm_vmm_page *page = map->page; + u8 kind_inv, pcf; + int kindn, aper; + const u8 *kindm; + + map->next = 1ULL << page->shift; + map->type = 0; + + aper = vmm->func->aper(target); + if (WARN_ON(aper < 0)) + return aper; + + kindm = vmm->mmu->func->kind(vmm->mmu, &kindn, &kind_inv); + if (kind >= kindn || kindm[kind] == kind_inv) { + VMM_DEBUG(vmm, "kind %02x", kind); + return -EINVAL; + } + + if (priv) { + if (ro) { + if (vol) + pcf = NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_ATOMIC_UNCACHED_ACD; + else + pcf = NV_MMU_VER3_PTE_PCF_PRIVILEGE_RO_ATOMIC_CACHED_ACD; + } else { + if (vol) + pcf = NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_ATOMIC_UNCACHED_ACD; + else + pcf = NV_MMU_VER3_PTE_PCF_PRIVILEGE_RW_ATOMIC_CACHED_ACD; + } + } else { + if (ro) { + if (vol) + pcf = NV_MMU_VER3_PTE_PCF_REGULAR_RO_ATOMIC_UNCACHED_ACD; + else + pcf = NV_MMU_VER3_PTE_PCF_REGULAR_RO_ATOMIC_CACHED_ACD; + } else { + if (vol) + pcf = NV_MMU_VER3_PTE_PCF_REGULAR_RW_ATOMIC_UNCACHED_ACD; + else + pcf = NV_MMU_VER3_PTE_PCF_REGULAR_RW_ATOMIC_CACHED_ACD; + } + } + + map->type |= NVDEF(NV_MMU, VER3_PTE, VALID, TRUE); + map->type |= NVVAL(NV_MMU, VER3_PTE, APERTURE, aper); + map->type |= NVVAL(NV_MMU, VER3_PTE, PCF, pcf); + map->type |= NVVAL(NV_MMU, VER3_PTE, KIND, kind); + return 0; +} + +static const struct nvkm_vmm_func +gh100_vmm = { + .join = gv100_vmm_join, + .part = gf100_vmm_part, + .aper = gf100_vmm_aper, + .valid = gp100_vmm_valid, + .valid2 = gh100_vmm_valid, + .flush = tu102_vmm_flush, + .page = { + { 56, &gh100_vmm_desc_16[5], NVKM_VMM_PAGE_Sxxx }, + { 47, &gh100_vmm_desc_16[4], NVKM_VMM_PAGE_Sxxx }, + { 38, &gh100_vmm_desc_16[3], NVKM_VMM_PAGE_Sxxx }, + { 29, &gh100_vmm_desc_16[2], NVKM_VMM_PAGE_SVxC }, + { 21, &gh100_vmm_desc_16[1], NVKM_VMM_PAGE_SVxC }, + { 16, &gh100_vmm_desc_16[0], NVKM_VMM_PAGE_SVxC }, + { 12, &gh100_vmm_desc_12[0], NVKM_VMM_PAGE_SVHx }, + {} + } +}; + +int +gh100_vmm_new(struct nvkm_mmu *mmu, bool managed, u64 addr, u64 size, + void *argv, u32 argc, struct lock_class_key *key, + const char *name, struct nvkm_vmm **pvmm) +{ + return gp100_vmm_new_(&gh100_vmm, mmu, managed, addr, size, + argv, argc, key, name, pvmm); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c index bddac77f48f0..851fd847a2a9 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c @@ -436,6 +436,9 @@ gp100_vmm_valid(struct nvkm_vmm *vmm, void *argv, u32 argc, return ret; } + if (vmm->func->valid2) + return vmm->func->valid2(vmm, ro, priv, kind, 0, map); + aper = vmm->func->aper(target); if (WARN_ON(aper < 0)) return aper; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c index 8379e72d77ab..4b30eab40bba 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c @@ -23,7 +23,7 @@ #include <subdev/timer.h> -static void +void tu102_vmm_flush(struct nvkm_vmm *vmm, int depth) { struct nvkm_device *device = vmm->mmu->subdev.device; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild index 174bdf995271..a14ea0f7b1c8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild @@ -13,3 +13,4 @@ nvkm-y += nvkm/subdev/pci/gf100.o nvkm-y += nvkm/subdev/pci/gf106.o nvkm-y += nvkm/subdev/pci/gk104.o nvkm-y += nvkm/subdev/pci/gp100.o +nvkm-y += nvkm/subdev/pci/gh100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c index 5a0de45d36ce..6867934256a7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c @@ -39,26 +39,26 @@ nvkm_pci_msi_rearm(struct nvkm_device *device) u32 nvkm_pci_rd32(struct nvkm_pci *pci, u16 addr) { - return pci->func->rd32(pci, addr); + return nvkm_rd32(pci->subdev.device, pci->func->cfg.addr + addr); } void nvkm_pci_wr08(struct nvkm_pci *pci, u16 addr, u8 data) { - pci->func->wr08(pci, addr, data); + nvkm_wr08(pci->subdev.device, pci->func->cfg.addr + addr, data); } void nvkm_pci_wr32(struct nvkm_pci *pci, u16 addr, u32 data) { - pci->func->wr32(pci, addr, data); + nvkm_wr32(pci->subdev.device, pci->func->cfg.addr + addr, data); } u32 nvkm_pci_mask(struct nvkm_pci *pci, u16 addr, u32 mask, u32 value) { - u32 data = pci->func->rd32(pci, addr); - pci->func->wr32(pci, addr, (data & ~mask) | value); + u32 data = nvkm_pci_rd32(pci, addr); + nvkm_pci_wr32(pci, addr, (data & ~mask) | value); return data; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c index 5b29aacedef3..5308f6539a3f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g84.c @@ -132,10 +132,9 @@ g84_pcie_init(struct nvkm_pci *pci) static const struct nvkm_pci_func g84_pci_func = { + .cfg = { .addr = 0x088000, .size = 0x1000 }, + .init = g84_pci_init, - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, .msi_rearm = nv46_pci_msi_rearm, .pcie.init = g84_pcie_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c index a9e0674009c6..8ae7aa02e675 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g92.c @@ -33,10 +33,9 @@ g92_pcie_version_supported(struct nvkm_pci *pci) static const struct nvkm_pci_func g92_pci_func = { + .cfg = { .addr = 0x088000, .size = 0x1000 }, + .init = g84_pci_init, - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, .msi_rearm = nv46_pci_msi_rearm, .pcie.init = g84_pcie_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c index 7bacd0693283..df745d0690ca 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/g94.c @@ -25,10 +25,9 @@ static const struct nvkm_pci_func g94_pci_func = { + .cfg = { .addr = 0x088000, .size = 0x1000 }, + .init = g84_pci_init, - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, .msi_rearm = nv40_pci_msi_rearm, .pcie.init = g84_pcie_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c index 099906092fe1..6ce941df87b7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf100.c @@ -78,10 +78,9 @@ gf100_pcie_set_link(struct nvkm_pci *pci, enum nvkm_pcie_speed speed, u8 width) static const struct nvkm_pci_func gf100_pci_func = { + .cfg = { .addr = 0x088000, .size = 0x1000 }, + .init = g84_pci_init, - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, .msi_rearm = gf100_pci_msi_rearm, .pcie.init = gf100_pcie_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c index bcde609ba866..712ca7e0959a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gf106.c @@ -25,10 +25,9 @@ static const struct nvkm_pci_func gf106_pci_func = { + .cfg = { .addr = 0x088000, .size = 0x1000 }, + .init = g84_pci_init, - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, .msi_rearm = nv40_pci_msi_rearm, .pcie.init = gf100_pcie_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gh100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gh100.c new file mode 100644 index 000000000000..42da92d7a5fe --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gh100.c @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: MIT + * + * Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + */ +#include "priv.h" + +#include <nvhw/drf.h> +#include <nvhw/ref/gh100/dev_xtl_ep_pri.h> + +static void +gh100_pci_msi_rearm(struct nvkm_pci *pci) +{ + /* Handled by top-level intr ACK. */ +} + +static const struct nvkm_pci_func +gh100_pci = { + .cfg = { + .addr = DRF_LO(NV_EP_PCFGM), + .size = DRF_HI(NV_EP_PCFGM) - DRF_LO(NV_EP_PCFGM) + 1, + }, + .msi_rearm = gh100_pci_msi_rearm, +}; + +int +gh100_pci_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst, + struct nvkm_pci **ppci) +{ + return nvkm_pci_new_(&gh100_pci, device, type, inst, ppci); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c index 6be87ecffc89..ec6d0a7de995 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gk104.c @@ -204,10 +204,9 @@ gk104_pcie_set_link(struct nvkm_pci *pci, enum nvkm_pcie_speed speed, u8 width) static const struct nvkm_pci_func gk104_pci_func = { + .cfg = { .addr = 0x088000, .size = 0x1000 }, + .init = g84_pci_init, - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, .msi_rearm = nv40_pci_msi_rearm, .pcie.init = gk104_pcie_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c index a5fafda0014d..4204316a544f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c @@ -31,9 +31,7 @@ gp100_pci_msi_rearm(struct nvkm_pci *pci) static const struct nvkm_pci_func gp100_pci_func = { - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, + .cfg = { .addr = 0x088000, .size = 0x1000 }, .msi_rearm = gp100_pci_msi_rearm, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c index 9ab64194b185..b8a3f6850fa7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv04.c @@ -23,32 +23,9 @@ */ #include "priv.h" -static u32 -nv04_pci_rd32(struct nvkm_pci *pci, u16 addr) -{ - struct nvkm_device *device = pci->subdev.device; - return nvkm_rd32(device, 0x001800 + addr); -} - -static void -nv04_pci_wr08(struct nvkm_pci *pci, u16 addr, u8 data) -{ - struct nvkm_device *device = pci->subdev.device; - nvkm_wr08(device, 0x001800 + addr, data); -} - -static void -nv04_pci_wr32(struct nvkm_pci *pci, u16 addr, u32 data) -{ - struct nvkm_device *device = pci->subdev.device; - nvkm_wr32(device, 0x001800 + addr, data); -} - static const struct nvkm_pci_func nv04_pci_func = { - .rd32 = nv04_pci_rd32, - .wr08 = nv04_pci_wr08, - .wr32 = nv04_pci_wr32, + .cfg = { .addr = 0x001800, .size = 0x1000 }, }; int diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c index 6a3c31cf0200..1971dbbdeb2b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv40.c @@ -23,27 +23,6 @@ */ #include "priv.h" -u32 -nv40_pci_rd32(struct nvkm_pci *pci, u16 addr) -{ - struct nvkm_device *device = pci->subdev.device; - return nvkm_rd32(device, 0x088000 + addr); -} - -void -nv40_pci_wr08(struct nvkm_pci *pci, u16 addr, u8 data) -{ - struct nvkm_device *device = pci->subdev.device; - nvkm_wr08(device, 0x088000 + addr, data); -} - -void -nv40_pci_wr32(struct nvkm_pci *pci, u16 addr, u32 data) -{ - struct nvkm_device *device = pci->subdev.device; - nvkm_wr32(device, 0x088000 + addr, data); -} - void nv40_pci_msi_rearm(struct nvkm_pci *pci) { @@ -52,9 +31,7 @@ nv40_pci_msi_rearm(struct nvkm_pci *pci) static const struct nvkm_pci_func nv40_pci_func = { - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, + .cfg = { .addr = 0x088000, .size = 0x1000 }, .msi_rearm = nv40_pci_msi_rearm, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c index 9cad17f178ec..0093eabac9ae 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv46.c @@ -38,9 +38,7 @@ nv46_pci_msi_rearm(struct nvkm_pci *pci) static const struct nvkm_pci_func nv46_pci_func = { - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, + .cfg = { .addr = 0x088000, .size = 0x1000 }, .msi_rearm = nv46_pci_msi_rearm, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c index 741e34bf307c..b445081bb80e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/nv4c.c @@ -25,9 +25,7 @@ static const struct nvkm_pci_func nv4c_pci_func = { - .rd32 = nv40_pci_rd32, - .wr08 = nv40_pci_wr08, - .wr32 = nv40_pci_wr32, + .cfg = { .addr = 0x088000, .size = 0x1000 }, }; int diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h index 9b7583532962..988eeee1471c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/priv.h @@ -8,10 +8,12 @@ int nvkm_pci_new_(const struct nvkm_pci_func *, struct nvkm_device *, enum nvkm_ struct nvkm_pci **); struct nvkm_pci_func { + struct { + u32 addr; + u16 size; + } cfg; + void (*init)(struct nvkm_pci *); - u32 (*rd32)(struct nvkm_pci *, u16 addr); - void (*wr08)(struct nvkm_pci *, u16 addr, u8 data); - void (*wr32)(struct nvkm_pci *, u16 addr, u32 data); void (*msi_rearm)(struct nvkm_pci *); struct { @@ -27,9 +29,6 @@ struct nvkm_pci_func { } pcie; }; -u32 nv40_pci_rd32(struct nvkm_pci *, u16); -void nv40_pci_wr08(struct nvkm_pci *, u16, u8); -void nv40_pci_wr32(struct nvkm_pci *, u16, u32); void nv40_pci_msi_rearm(struct nvkm_pci *); void nv46_pci_msi_rearm(struct nvkm_pci *); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/r535.c b/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/r535.c index dce337306cab..9446049642e1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/r535.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/r535.c @@ -21,6 +21,8 @@ */ #include "priv.h" +#include <rm/gpu.h> + static void r535_vfn_dtor(struct nvkm_vfn *vfn) { @@ -32,6 +34,7 @@ r535_vfn_new(const struct nvkm_vfn_func *hw, struct nvkm_device *device, enum nvkm_subdev_type type, int inst, u32 addr, struct nvkm_vfn **pvfn) { + const struct nvkm_rm_gpu *gpu = device->gsp->rm->gpu; struct nvkm_vfn_func *rm; int ret; @@ -39,8 +42,12 @@ r535_vfn_new(const struct nvkm_vfn_func *hw, return -ENOMEM; rm->dtor = r535_vfn_dtor; - rm->intr = hw->intr; - rm->user = hw->user; + rm->intr = &tu102_vfn_intr, + rm->user.addr = 0x030000; + rm->user.size = 0x010000; + rm->user.base.minver = -1; + rm->user.base.maxver = -1; + rm->user.base.oclass = gpu->usermode.class; ret = nvkm_vfn_new_(rm, device, type, inst, addr, pvfn); if (ret) diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/uvfn.c b/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/uvfn.c index c5460a14c541..4e64d8843373 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/uvfn.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/vfn/uvfn.c @@ -36,7 +36,7 @@ nvkm_uvfn_map(struct nvkm_object *object, void *argv, u32 argc, struct nvkm_vfn *vfn = nvkm_uvfn(object)->vfn; struct nvkm_device *device = vfn->subdev.device; - *addr = device->func->resource_addr(device, 0) + vfn->addr.user; + *addr = device->func->resource_addr(device, NVKM_BAR0_PRI) + vfn->addr.user; *size = vfn->func->user.size; *type = NVKM_OBJECT_MAP_IO; return 0; diff --git a/drivers/gpu/drm/nova/Kconfig b/drivers/gpu/drm/nova/Kconfig new file mode 100644 index 000000000000..cca6a3fea879 --- /dev/null +++ b/drivers/gpu/drm/nova/Kconfig @@ -0,0 +1,14 @@ +config DRM_NOVA + tristate "Nova DRM driver" + depends on DRM=y + depends on PCI + depends on RUST + select AUXILIARY_BUS + default n + help + Choose this if you want to build the Nova DRM driver for Nvidia + GSP-based GPUs. + + This driver is work in progress and may not be functional. + + If M is selected, the module will be called nova. diff --git a/drivers/gpu/drm/nova/Makefile b/drivers/gpu/drm/nova/Makefile new file mode 100644 index 000000000000..42019bff3173 --- /dev/null +++ b/drivers/gpu/drm/nova/Makefile @@ -0,0 +1,3 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-$(CONFIG_DRM_NOVA) += nova.o diff --git a/drivers/gpu/drm/nova/driver.rs b/drivers/gpu/drm/nova/driver.rs new file mode 100644 index 000000000000..b28b2e05cc15 --- /dev/null +++ b/drivers/gpu/drm/nova/driver.rs @@ -0,0 +1,69 @@ +// SPDX-License-Identifier: GPL-2.0 + +use kernel::{auxiliary, c_str, device::Core, drm, drm::gem, drm::ioctl, prelude::*, types::ARef}; + +use crate::file::File; +use crate::gem::NovaObject; + +pub(crate) struct NovaDriver { + #[expect(unused)] + drm: ARef<drm::Device<Self>>, +} + +/// Convienence type alias for the DRM device type for this driver +pub(crate) type NovaDevice = drm::Device<NovaDriver>; + +#[pin_data] +pub(crate) struct NovaData { + pub(crate) adev: ARef<auxiliary::Device>, +} + +const INFO: drm::DriverInfo = drm::DriverInfo { + major: 0, + minor: 0, + patchlevel: 0, + name: c_str!("nova"), + desc: c_str!("Nvidia Graphics"), +}; + +const NOVA_CORE_MODULE_NAME: &CStr = c_str!("NovaCore"); +const AUXILIARY_NAME: &CStr = c_str!("nova-drm"); + +kernel::auxiliary_device_table!( + AUX_TABLE, + MODULE_AUX_TABLE, + <NovaDriver as auxiliary::Driver>::IdInfo, + [( + auxiliary::DeviceId::new(NOVA_CORE_MODULE_NAME, AUXILIARY_NAME), + () + )] +); + +impl auxiliary::Driver for NovaDriver { + type IdInfo = (); + const ID_TABLE: auxiliary::IdTable<Self::IdInfo> = &AUX_TABLE; + + fn probe(adev: &auxiliary::Device<Core>, _info: &Self::IdInfo) -> Result<Pin<KBox<Self>>> { + let data = try_pin_init!(NovaData { adev: adev.into() }); + + let drm = drm::Device::<Self>::new(adev.as_ref(), data)?; + drm::Registration::new_foreign_owned(&drm, adev.as_ref(), 0)?; + + Ok(KBox::new(Self { drm }, GFP_KERNEL)?.into()) + } +} + +#[vtable] +impl drm::Driver for NovaDriver { + type Data = NovaData; + type File = File; + type Object = gem::Object<NovaObject>; + + const INFO: drm::DriverInfo = INFO; + + kernel::declare_drm_ioctls! { + (NOVA_GETPARAM, drm_nova_getparam, ioctl::RENDER_ALLOW, File::get_param), + (NOVA_GEM_CREATE, drm_nova_gem_create, ioctl::AUTH | ioctl::RENDER_ALLOW, File::gem_create), + (NOVA_GEM_INFO, drm_nova_gem_info, ioctl::AUTH | ioctl::RENDER_ALLOW, File::gem_info), + } +} diff --git a/drivers/gpu/drm/nova/file.rs b/drivers/gpu/drm/nova/file.rs new file mode 100644 index 000000000000..7e59a34b830d --- /dev/null +++ b/drivers/gpu/drm/nova/file.rs @@ -0,0 +1,74 @@ +// SPDX-License-Identifier: GPL-2.0 + +use crate::driver::{NovaDevice, NovaDriver}; +use crate::gem::NovaObject; +use crate::uapi::{GemCreate, GemInfo, Getparam}; +use kernel::{ + alloc::flags::*, + drm::{self, gem::BaseObject}, + pci, + prelude::*, + types::Opaque, + uapi, +}; + +pub(crate) struct File; + +impl drm::file::DriverFile for File { + type Driver = NovaDriver; + + fn open(_dev: &NovaDevice) -> Result<Pin<KBox<Self>>> { + Ok(KBox::new(Self, GFP_KERNEL)?.into()) + } +} + +impl File { + /// IOCTL: get_param: Query GPU / driver metadata. + pub(crate) fn get_param( + dev: &NovaDevice, + getparam: &Opaque<uapi::drm_nova_getparam>, + _file: &drm::File<File>, + ) -> Result<u32> { + let adev = &dev.adev; + let parent = adev.parent().ok_or(ENOENT)?; + let pdev: &pci::Device = parent.try_into()?; + let getparam: &Getparam = getparam.into(); + + let value = match getparam.param() as u32 { + uapi::NOVA_GETPARAM_VRAM_BAR_SIZE => pdev.resource_len(1)?, + _ => return Err(EINVAL), + }; + + getparam.set_value(value); + + Ok(0) + } + + /// IOCTL: gem_create: Create a new DRM GEM object. + pub(crate) fn gem_create( + dev: &NovaDevice, + req: &Opaque<uapi::drm_nova_gem_create>, + file: &drm::File<File>, + ) -> Result<u32> { + let req: &GemCreate = req.into(); + let obj = NovaObject::new(dev, req.size().try_into()?)?; + + req.set_handle(obj.create_handle(file)?); + + Ok(0) + } + + /// IOCTL: gem_info: Query GEM metadata. + pub(crate) fn gem_info( + _dev: &NovaDevice, + req: &Opaque<uapi::drm_nova_gem_info>, + file: &drm::File<File>, + ) -> Result<u32> { + let req: &GemInfo = req.into(); + let bo = NovaObject::lookup_handle(file, req.handle())?; + + req.set_size(bo.size().try_into()?); + + Ok(0) + } +} diff --git a/drivers/gpu/drm/nova/gem.rs b/drivers/gpu/drm/nova/gem.rs new file mode 100644 index 000000000000..33b62d21400c --- /dev/null +++ b/drivers/gpu/drm/nova/gem.rs @@ -0,0 +1,49 @@ +// SPDX-License-Identifier: GPL-2.0 + +use kernel::{ + drm, + drm::{gem, gem::BaseObject}, + prelude::*, + types::ARef, +}; + +use crate::{ + driver::{NovaDevice, NovaDriver}, + file::File, +}; + +/// GEM Object inner driver data +#[pin_data] +pub(crate) struct NovaObject {} + +impl gem::BaseDriverObject<gem::Object<NovaObject>> for NovaObject { + fn new(_dev: &NovaDevice, _size: usize) -> impl PinInit<Self, Error> { + try_pin_init!(NovaObject {}) + } +} + +impl gem::DriverObject for NovaObject { + type Driver = NovaDriver; +} + +impl NovaObject { + /// Create a new DRM GEM object. + pub(crate) fn new(dev: &NovaDevice, size: usize) -> Result<ARef<gem::Object<Self>>> { + let aligned_size = size.next_multiple_of(1 << 12); + + if size == 0 || size > aligned_size { + return Err(EINVAL); + } + + gem::Object::new(dev, aligned_size) + } + + /// Look up a GEM object handle for a `File` and return an `ObjectRef` for it. + #[inline] + pub(crate) fn lookup_handle( + file: &drm::File<File>, + handle: u32, + ) -> Result<ARef<gem::Object<Self>>> { + gem::Object::lookup_handle(file, handle) + } +} diff --git a/drivers/gpu/drm/nova/nova.rs b/drivers/gpu/drm/nova/nova.rs new file mode 100644 index 000000000000..902876aa14d1 --- /dev/null +++ b/drivers/gpu/drm/nova/nova.rs @@ -0,0 +1,18 @@ +// SPDX-License-Identifier: GPL-2.0 + +//! Nova DRM Driver + +mod driver; +mod file; +mod gem; +mod uapi; + +use crate::driver::NovaDriver; + +kernel::module_auxiliary_driver! { + type: NovaDriver, + name: "Nova", + author: "Danilo Krummrich", + description: "Nova GPU driver", + license: "GPL v2", +} diff --git a/drivers/gpu/drm/nova/uapi.rs b/drivers/gpu/drm/nova/uapi.rs new file mode 100644 index 000000000000..eb228a58d423 --- /dev/null +++ b/drivers/gpu/drm/nova/uapi.rs @@ -0,0 +1,61 @@ +// SPDX-License-Identifier: GPL-2.0 + +use kernel::uapi; + +// TODO Work out some common infrastructure to avoid boilerplate code for uAPI abstractions. + +macro_rules! define_uapi_abstraction { + ($name:ident <= $inner:ty) => { + #[repr(transparent)] + pub struct $name(::kernel::types::Opaque<$inner>); + + impl ::core::convert::From<&::kernel::types::Opaque<$inner>> for &$name { + fn from(value: &::kernel::types::Opaque<$inner>) -> Self { + // SAFETY: `Self` is a transparent wrapper of `$inner`. + unsafe { ::core::mem::transmute(value) } + } + } + }; +} + +define_uapi_abstraction!(Getparam <= uapi::drm_nova_getparam); + +impl Getparam { + pub fn param(&self) -> u64 { + // SAFETY: `self.get()` is a valid pointer to a `struct drm_nova_getparam`. + unsafe { (*self.0.get()).param } + } + + pub fn set_value(&self, v: u64) { + // SAFETY: `self.get()` is a valid pointer to a `struct drm_nova_getparam`. + unsafe { (*self.0.get()).value = v }; + } +} + +define_uapi_abstraction!(GemCreate <= uapi::drm_nova_gem_create); + +impl GemCreate { + pub fn size(&self) -> u64 { + // SAFETY: `self.get()` is a valid pointer to a `struct drm_nova_gem_create`. + unsafe { (*self.0.get()).size } + } + + pub fn set_handle(&self, handle: u32) { + // SAFETY: `self.get()` is a valid pointer to a `struct drm_nova_gem_create`. + unsafe { (*self.0.get()).handle = handle }; + } +} + +define_uapi_abstraction!(GemInfo <= uapi::drm_nova_gem_info); + +impl GemInfo { + pub fn handle(&self) -> u32 { + // SAFETY: `self.get()` is a valid pointer to a `struct drm_nova_gem_info`. + unsafe { (*self.0.get()).handle } + } + + pub fn set_size(&self, size: u64) { + // SAFETY: `self.get()` is a valid pointer to a `struct drm_nova_gem_info`. + unsafe { (*self.0.get()).size = size }; + } +} diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h index 2db40789235c..a7caac5b8ac8 100644 --- a/drivers/gpu/drm/radeon/atombios.h +++ b/drivers/gpu/drm/radeon/atombios.h @@ -3825,8 +3825,7 @@ typedef struct _ATOM_DPCD_INFO // note2: From RV770, the memory is more than 32bit addressable, so we will change // ucTableFormatRevision=1,ucTableContentRevision=4, the structure remains // exactly same as 1.1 and 1.2 (1.3 is never in use), but ulStartAddrUsedByFirmware -// (in offset to start of memory address) is KB aligned instead of byte aligend. -/***********************************************************************************/ +// (in offset to start of memory address) is KB aligned instead of byte aligned. // Note3: /* If we change usReserved to "usFBUsedbyDrvInKB", then to VBIOS this usFBUsedbyDrvInKB is a predefined, unchanged constant across VGA or non VGA adapter, for CAIL, The size of FB access area is known, only thing missing is the Offset of FB Access area, so we can have: diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index 11a492f21157..51a3e0fc2f56 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -8548,7 +8548,7 @@ int cik_suspend(struct radeon_device *rdev) */ int cik_init(struct radeon_device *rdev) { - struct radeon_ring *ring; + struct radeon_ring *ring, *ring_cp1, *ring_cp2; int r; /* Read BIOS */ @@ -8623,19 +8623,22 @@ int cik_init(struct radeon_device *rdev) ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); - ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; - ring->ring_obj = NULL; - r600_ring_init(rdev, ring, 1024 * 1024); - r = radeon_doorbell_get(rdev, &ring->doorbell_index); + ring_cp1 = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; + ring_cp2 = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; + ring_cp1->ring_obj = NULL; + ring_cp2->ring_obj = NULL; + ring_cp1->doorbell_index = RADEON_MAX_DOORBELLS; + ring_cp2->doorbell_index = RADEON_MAX_DOORBELLS; + + r600_ring_init(rdev, ring_cp1, 1024 * 1024); + r = radeon_doorbell_get(rdev, &ring_cp1->doorbell_index); if (r) return r; - ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; - ring->ring_obj = NULL; - r600_ring_init(rdev, ring, 1024 * 1024); - r = radeon_doorbell_get(rdev, &ring->doorbell_index); + r600_ring_init(rdev, ring_cp2, 1024 * 1024); + r = radeon_doorbell_get(rdev, &ring_cp2->doorbell_index); if (r) - return r; + goto out; ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; ring->ring_obj = NULL; @@ -8653,12 +8656,16 @@ int cik_init(struct radeon_device *rdev) r = r600_pcie_gart_init(rdev); if (r) - return r; + goto out; rdev->accel_working = true; r = cik_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); + radeon_doorbell_free(rdev, ring_cp1->doorbell_index); + radeon_doorbell_free(rdev, ring_cp2->doorbell_index); + ring_cp1->doorbell_index = RADEON_MAX_DOORBELLS; + ring_cp2->doorbell_index = RADEON_MAX_DOORBELLS; cik_cp_fini(rdev); cik_sdma_fini(rdev); cik_irq_fini(rdev); @@ -8678,10 +8685,16 @@ int cik_init(struct radeon_device *rdev) */ if (!rdev->mc_fw && !(rdev->flags & RADEON_IS_IGP)) { DRM_ERROR("radeon: MC ucode required for NI+.\n"); - return -EINVAL; + r = -EINVAL; + goto out; } return 0; + +out: + radeon_doorbell_free(rdev, ring_cp1->doorbell_index); + radeon_doorbell_free(rdev, ring_cp2->doorbell_index); + return r; } /** @@ -8695,6 +8708,7 @@ int cik_init(struct radeon_device *rdev) */ void cik_fini(struct radeon_device *rdev) { + struct radeon_ring *ring; radeon_pm_fini(rdev); cik_cp_fini(rdev); cik_sdma_fini(rdev); @@ -8708,6 +8722,10 @@ void cik_fini(struct radeon_device *rdev) radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); uvd_v1_0_fini(rdev); + ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; + radeon_doorbell_free(rdev, ring->doorbell_index); + ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; + radeon_doorbell_free(rdev, ring->doorbell_index); radeon_uvd_fini(rdev); radeon_vce_fini(rdev); cik_pcie_gart_fini(rdev); diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c index 661f374f5f27..9758f3a9df75 100644 --- a/drivers/gpu/drm/radeon/r600_hdmi.c +++ b/drivers/gpu/drm/radeon/r600_hdmi.c @@ -290,28 +290,6 @@ int r600_hdmi_buffer_status_changed(struct drm_encoder *encoder) return result; } -/* - * write the audio workaround status to the hardware - */ -void r600_hdmi_audio_workaround(struct drm_encoder *encoder) -{ - struct drm_device *dev = encoder->dev; - struct radeon_device *rdev = dev->dev_private; - struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); - struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; - uint32_t offset = dig->afmt->offset; - bool hdmi_audio_workaround = false; /* FIXME */ - u32 value; - - if (!hdmi_audio_workaround || - r600_hdmi_is_audio_buffer_filled(encoder)) - value = 0; /* disable workaround */ - else - value = HDMI0_AUDIO_TEST_EN; /* enable workaround */ - WREG32_P(HDMI0_AUDIO_PACKET_CONTROL + offset, - value, ~HDMI0_AUDIO_TEST_EN); -} - void r600_hdmi_audio_set_dto(struct radeon_device *rdev, struct radeon_crtc *crtc, unsigned int clock) { diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 8605c074d9f7..63c47585afbc 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -394,9 +394,6 @@ long radeon_fence_wait_timeout(struct radeon_fence *fence, bool interruptible, l int radeon_fence_wait(struct radeon_fence *fence, bool interruptible); int radeon_fence_wait_next(struct radeon_device *rdev, int ring); int radeon_fence_wait_empty(struct radeon_device *rdev, int ring); -int radeon_fence_wait_any(struct radeon_device *rdev, - struct radeon_fence **fences, - bool intr); struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence); void radeon_fence_unref(struct radeon_fence **fence); unsigned radeon_fence_count_emitted(struct radeon_device *rdev, int ring); diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index 8f5e07834fcc..9e697f10f9ca 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -401,7 +401,6 @@ void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock); void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder, void *buffer, size_t size); void r600_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t clock); -void r600_hdmi_audio_workaround(struct drm_encoder *encoder); int r600_hdmi_buffer_status_changed(struct drm_encoder *encoder); void r600_hdmi_update_audio_settings(struct drm_encoder *encoder); u32 r600_get_xclk(struct radeon_device *rdev); diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 64b26bfeafc9..b8e6202f1d5b 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -409,7 +409,6 @@ static int cmp_size_smaller_first(void *priv, const struct list_head *a, * radeon_cs_parser_fini() - clean parser states * @parser: parser structure holding parsing context. * @error: error number - * @backoff: indicator to backoff the reservation * * If error is set than unvalidate buffer, otherwise just free memory * used by parsing context. diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c index 8ff4f18b51a9..5b5b54e876d4 100644 --- a/drivers/gpu/drm/radeon/radeon_fence.c +++ b/drivers/gpu/drm/radeon/radeon_fence.c @@ -575,48 +575,6 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr) } /** - * radeon_fence_wait_any - wait for a fence to signal on any ring - * - * @rdev: radeon device pointer - * @fences: radeon fence object(s) - * @intr: use interruptable sleep - * - * Wait for any requested fence to signal (all asics). Fence - * array is indexed by ring id. @intr selects whether to use - * interruptable (true) or non-interruptable (false) sleep when - * waiting for the fences. Used by the suballocator. - * Returns 0 if any fence has passed, error for all other cases. - */ -int radeon_fence_wait_any(struct radeon_device *rdev, - struct radeon_fence **fences, - bool intr) -{ - uint64_t seq[RADEON_NUM_RINGS]; - unsigned int i, num_rings = 0; - long r; - - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - seq[i] = 0; - - if (!fences[i]) - continue; - - seq[i] = fences[i]->seq; - ++num_rings; - } - - /* nothing to wait for ? */ - if (num_rings == 0) - return -ENOENT; - - r = radeon_fence_wait_seq_timeout(rdev, seq, intr, MAX_SCHEDULE_TIMEOUT); - if (r < 0) - return r; - - return 0; -} - -/** * radeon_fence_wait_next - wait for the next fence to signal * * @rdev: radeon device pointer diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h index 65a911ddd509..f9267b026f8d 100644 --- a/drivers/gpu/drm/radeon/sid.h +++ b/drivers/gpu/drm/radeon/sid.h @@ -1109,7 +1109,7 @@ #define MIN_POWER_SHIFT 0 #define MAX_POWER(x) ((x) << 16) #define MAX_POWER_MASK (0x3fff << 16) -#define MAX_POWER_SHIFT 0 +#define MAX_POWER_SHIFT 16 #define SQ_POWER_THROTTLE2 0x8e5c #define MAX_POWER_DELTA(x) ((x) << 0) #define MAX_POWER_DELTA_MASK (0x3fff << 0) diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile index c5d6681645ed..e4bf484d4121 100644 --- a/drivers/gpu/drm/xe/Makefile +++ b/drivers/gpu/drm/xe/Makefile @@ -187,7 +187,6 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \ display/xe_display.o \ display/xe_display_misc.o \ display/xe_display_rpm.o \ - display/xe_display_rps.o \ display/xe_display_wa.o \ display/xe_dsb_buffer.o \ display/xe_fb_pin.o \ @@ -198,7 +197,6 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \ # SOC code shared with i915 xe-$(CONFIG_DRM_XE_DISPLAY) += \ i915-soc/intel_dram.o \ - i915-soc/intel_pch.o \ i915-soc/intel_rom.o # Display code shared with i915 @@ -273,6 +271,7 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \ i915-display/intel_panel.o \ i915-display/intel_pfit.o \ i915-display/intel_pmdemand.o \ + i915-display/intel_pch.o \ i915-display/intel_pps.o \ i915-display/intel_psr.o \ i915-display/intel_qp_tables.o \ diff --git a/drivers/gpu/drm/xe/compat-i915-headers/gt/intel_rps.h b/drivers/gpu/drm/xe/compat-i915-headers/gt/intel_rps.h deleted file mode 100644 index 21fec9cc837c..000000000000 --- a/drivers/gpu/drm/xe/compat-i915-headers/gt/intel_rps.h +++ /dev/null @@ -1,11 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -/* - * Copyright © 2023 Intel Corporation - */ - -#ifndef __INTEL_RPS_H__ -#define __INTEL_RPS_H__ - -#define gen5_rps_irq_handler(x) ({}) - -#endif /* __INTEL_RPS_H__ */ diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h index f89bd5e3520d..9b7572e06f34 100644 --- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h +++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h @@ -21,28 +21,12 @@ static inline struct drm_i915_private *to_i915(const struct drm_device *dev) return container_of(dev, struct drm_i915_private, drm); } +/* compat platform checks only for soc/ usage */ #define IS_PLATFORM(xe, x) ((xe)->info.platform == x) -#define INTEL_INFO(dev_priv) (&((dev_priv)->info)) -#define IS_I830(dev_priv) (dev_priv && 0) -#define IS_I845G(dev_priv) (dev_priv && 0) -#define IS_I85X(dev_priv) (dev_priv && 0) -#define IS_I865G(dev_priv) (dev_priv && 0) #define IS_I915G(dev_priv) (dev_priv && 0) #define IS_I915GM(dev_priv) (dev_priv && 0) -#define IS_I945G(dev_priv) (dev_priv && 0) -#define IS_I945GM(dev_priv) (dev_priv && 0) -#define IS_I965G(dev_priv) (dev_priv && 0) -#define IS_I965GM(dev_priv) (dev_priv && 0) -#define IS_G45(dev_priv) (dev_priv && 0) -#define IS_GM45(dev_priv) (dev_priv && 0) -#define IS_G4X(dev_priv) (dev_priv && 0) #define IS_PINEVIEW(dev_priv) (dev_priv && 0) -#define IS_G33(dev_priv) (dev_priv && 0) -#define IS_IRONLAKE(dev_priv) (dev_priv && 0) -#define IS_IRONLAKE_M(dev_priv) (dev_priv && 0) -#define IS_SANDYBRIDGE(dev_priv) (dev_priv && 0) #define IS_IVYBRIDGE(dev_priv) (dev_priv && 0) -#define IS_IVB_GT1(dev_priv) (dev_priv && 0) #define IS_VALLEYVIEW(dev_priv) (dev_priv && 0) #define IS_CHERRYVIEW(dev_priv) (dev_priv && 0) #define IS_HASWELL(dev_priv) (dev_priv && 0) @@ -70,39 +54,10 @@ static inline struct drm_i915_private *to_i915(const struct drm_device *dev) #define IS_HASWELL_ULT(dev_priv) (dev_priv && 0) #define IS_BROADWELL_ULT(dev_priv) (dev_priv && 0) -#define IS_BROADWELL_ULX(dev_priv) (dev_priv && 0) #define IS_MOBILE(xe) (xe && 0) -#define IS_TIGERLAKE_UY(xe) (xe && 0) -#define IS_COMETLAKE_ULX(xe) (xe && 0) -#define IS_COFFEELAKE_ULX(xe) (xe && 0) -#define IS_KABYLAKE_ULX(xe) (xe && 0) -#define IS_SKYLAKE_ULX(xe) (xe && 0) -#define IS_HASWELL_ULX(xe) (xe && 0) -#define IS_COMETLAKE_ULT(xe) (xe && 0) -#define IS_COFFEELAKE_ULT(xe) (xe && 0) -#define IS_KABYLAKE_ULT(xe) (xe && 0) -#define IS_SKYLAKE_ULT(xe) (xe && 0) - -#define IS_DG2_G10(xe) ((xe)->info.subplatform == XE_SUBPLATFORM_DG2_G10) -#define IS_DG2_G11(xe) ((xe)->info.subplatform == XE_SUBPLATFORM_DG2_G11) -#define IS_DG2_G12(xe) ((xe)->info.subplatform == XE_SUBPLATFORM_DG2_G12) -#define IS_RAPTORLAKE_U(xe) ((xe)->info.subplatform == XE_SUBPLATFORM_ALDERLAKE_P_RPLU) -#define IS_ICL_WITH_PORT_F(xe) (xe && 0) #define HAS_FLAT_CCS(xe) (xe_device_has_flat_ccs(xe)) - #define HAS_128_BYTE_Y_TILING(xe) (xe || 1) -#ifdef CONFIG_ARM64 -/* - * arm64 indirectly includes linux/rtc.h, - * which defines a irq_lock, so include it - * here before #define-ing it - */ -#include <linux/rtc.h> -#endif - -#define irq_lock irq.lock - #endif diff --git a/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h deleted file mode 100644 index 9c46556d33a4..000000000000 --- a/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h +++ /dev/null @@ -1,6 +0,0 @@ -/* SPDX-License-Identifier: MIT */ -/* - * Copyright © 2023 Intel Corporation - */ - -#include "../../../i915/soc/intel_pch.h" diff --git a/drivers/gpu/drm/xe/display/intel_fbdev_fb.c b/drivers/gpu/drm/xe/display/intel_fbdev_fb.c index 267f31697343..e8191562d122 100644 --- a/drivers/gpu/drm/xe/display/intel_fbdev_fb.c +++ b/drivers/gpu/drm/xe/display/intel_fbdev_fb.c @@ -79,11 +79,11 @@ err: return ERR_CAST(fb); } -int intel_fbdev_fb_fill_info(struct drm_i915_private *i915, struct fb_info *info, - struct drm_gem_object *_obj, struct i915_vma *vma) +int intel_fbdev_fb_fill_info(struct intel_display *display, struct fb_info *info, + struct drm_gem_object *_obj, struct i915_vma *vma) { struct xe_bo *obj = gem_to_xe_bo(_obj); - struct pci_dev *pdev = to_pci_dev(i915->drm.dev); + struct pci_dev *pdev = to_pci_dev(display->drm->dev); if (!(obj->flags & XE_BO_FLAG_SYSTEM)) { if (obj->flags & XE_BO_FLAG_STOLEN) diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c index 20c3bcd953b7..68f064f33d4b 100644 --- a/drivers/gpu/drm/xe/display/xe_display.c +++ b/drivers/gpu/drm/xe/display/xe_display.c @@ -133,9 +133,6 @@ int xe_display_init_early(struct xe_device *xe) /* Fake uncore lock */ spin_lock_init(&xe->uncore.lock); - /* This must be called before any calls to HAS_PCH_* */ - intel_detect_pch(xe); - intel_display_driver_early_probe(display); /* Early display init.. */ diff --git a/drivers/gpu/drm/xe/display/xe_display_rps.c b/drivers/gpu/drm/xe/display/xe_display_rps.c deleted file mode 100644 index fa616f9688a5..000000000000 --- a/drivers/gpu/drm/xe/display/xe_display_rps.c +++ /dev/null @@ -1,17 +0,0 @@ -// SPDX-License-Identifier: MIT -/* - * Copyright © 2023 Intel Corporation - */ - -#include "intel_display_rps.h" - -void intel_display_rps_boost_after_vblank(struct drm_crtc *crtc, - struct dma_fence *fence) -{ -} - -void intel_display_rps_mark_interactive(struct intel_display *display, - struct intel_atomic_state *state, - bool interactive) -{ -} diff --git a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c index 7c02323e9531..b35a6f201d4a 100644 --- a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c +++ b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c @@ -9,7 +9,6 @@ #include "abi/gsc_command_header_abi.h" #include "intel_hdcp_gsc.h" -#include "intel_hdcp_gsc_message.h" #include "xe_bo.h" #include "xe_device.h" #include "xe_device_types.h" @@ -22,7 +21,8 @@ #define HECI_MEADDRESS_HDCP 18 -struct intel_hdcp_gsc_message { +struct intel_hdcp_gsc_context { + struct xe_device *xe; struct xe_bo *hdcp_bo; u64 hdcp_cmd_in; u64 hdcp_cmd_out; @@ -30,14 +30,9 @@ struct intel_hdcp_gsc_message { #define HDCP_GSC_HEADER_SIZE sizeof(struct intel_gsc_mtl_header) -bool intel_hdcp_gsc_cs_required(struct intel_display *display) +bool intel_hdcp_gsc_check_status(struct drm_device *drm) { - return DISPLAY_VER(display) >= 14; -} - -bool intel_hdcp_gsc_check_status(struct intel_display *display) -{ - struct xe_device *xe = to_xe_device(display->drm); + struct xe_device *xe = to_xe_device(drm); struct xe_tile *tile = xe_device_get_root_tile(xe); struct xe_gt *gt = tile->media_gt; struct xe_gsc *gsc = >->uc.gsc; @@ -69,10 +64,9 @@ out: } /*This function helps allocate memory for the command that we will send to gsc cs */ -static int intel_hdcp_gsc_initialize_message(struct intel_display *display, - struct intel_hdcp_gsc_message *hdcp_message) +static int intel_hdcp_gsc_initialize_message(struct xe_device *xe, + struct intel_hdcp_gsc_context *gsc_context) { - struct xe_device *xe = to_xe_device(display->drm); struct xe_bo *bo = NULL; u64 cmd_in, cmd_out; int ret = 0; @@ -84,7 +78,7 @@ static int intel_hdcp_gsc_initialize_message(struct intel_display *display, XE_BO_FLAG_GGTT); if (IS_ERR(bo)) { - drm_err(display->drm, "Failed to allocate bo for HDCP streaming command!\n"); + drm_err(&xe->drm, "Failed to allocate bo for HDCP streaming command!\n"); ret = PTR_ERR(bo); goto out; } @@ -93,104 +87,60 @@ static int intel_hdcp_gsc_initialize_message(struct intel_display *display, cmd_out = cmd_in + PAGE_SIZE; xe_map_memset(xe, &bo->vmap, 0, 0, bo->size); - hdcp_message->hdcp_bo = bo; - hdcp_message->hdcp_cmd_in = cmd_in; - hdcp_message->hdcp_cmd_out = cmd_out; + gsc_context->hdcp_bo = bo; + gsc_context->hdcp_cmd_in = cmd_in; + gsc_context->hdcp_cmd_out = cmd_out; + gsc_context->xe = xe; + out: return ret; } -static int intel_hdcp_gsc_hdcp2_init(struct intel_display *display) +struct intel_hdcp_gsc_context *intel_hdcp_gsc_context_alloc(struct drm_device *drm) { - struct intel_hdcp_gsc_message *hdcp_message; + struct xe_device *xe = to_xe_device(drm); + struct intel_hdcp_gsc_context *gsc_context; int ret; - hdcp_message = kzalloc(sizeof(*hdcp_message), GFP_KERNEL); - - if (!hdcp_message) - return -ENOMEM; + gsc_context = kzalloc(sizeof(*gsc_context), GFP_KERNEL); + if (!gsc_context) + return ERR_PTR(-ENOMEM); /* * NOTE: No need to lock the comp mutex here as it is already * going to be taken before this function called */ - ret = intel_hdcp_gsc_initialize_message(display, hdcp_message); + ret = intel_hdcp_gsc_initialize_message(xe, gsc_context); if (ret) { - drm_err(display->drm, "Could not initialize hdcp_message\n"); - kfree(hdcp_message); - return ret; + drm_err(&xe->drm, "Could not initialize gsc_context\n"); + kfree(gsc_context); + gsc_context = ERR_PTR(ret); } - display->hdcp.hdcp_message = hdcp_message; - return ret; -} - -static const struct i915_hdcp_ops gsc_hdcp_ops = { - .initiate_hdcp2_session = intel_hdcp_gsc_initiate_session, - .verify_receiver_cert_prepare_km = - intel_hdcp_gsc_verify_receiver_cert_prepare_km, - .verify_hprime = intel_hdcp_gsc_verify_hprime, - .store_pairing_info = intel_hdcp_gsc_store_pairing_info, - .initiate_locality_check = intel_hdcp_gsc_initiate_locality_check, - .verify_lprime = intel_hdcp_gsc_verify_lprime, - .get_session_key = intel_hdcp_gsc_get_session_key, - .repeater_check_flow_prepare_ack = - intel_hdcp_gsc_repeater_check_flow_prepare_ack, - .verify_mprime = intel_hdcp_gsc_verify_mprime, - .enable_hdcp_authentication = intel_hdcp_gsc_enable_authentication, - .close_hdcp_session = intel_hdcp_gsc_close_session, -}; - -int intel_hdcp_gsc_init(struct intel_display *display) -{ - struct i915_hdcp_arbiter *data; - int ret; - - data = kzalloc(sizeof(*data), GFP_KERNEL); - if (!data) - return -ENOMEM; - - mutex_lock(&display->hdcp.hdcp_mutex); - display->hdcp.arbiter = data; - display->hdcp.arbiter->hdcp_dev = display->drm->dev; - display->hdcp.arbiter->ops = &gsc_hdcp_ops; - ret = intel_hdcp_gsc_hdcp2_init(display); - if (ret) - kfree(data); - - mutex_unlock(&display->hdcp.hdcp_mutex); - - return ret; + return gsc_context; } -void intel_hdcp_gsc_fini(struct intel_display *display) +void intel_hdcp_gsc_context_free(struct intel_hdcp_gsc_context *gsc_context) { - struct intel_hdcp_gsc_message *hdcp_message = - display->hdcp.hdcp_message; - struct i915_hdcp_arbiter *arb = display->hdcp.arbiter; - - if (hdcp_message) { - xe_bo_unpin_map_no_vm(hdcp_message->hdcp_bo); - kfree(hdcp_message); - display->hdcp.hdcp_message = NULL; - } + if (!gsc_context) + return; - kfree(arb); - display->hdcp.arbiter = NULL; + xe_bo_unpin_map_no_vm(gsc_context->hdcp_bo); + kfree(gsc_context); } static int xe_gsc_send_sync(struct xe_device *xe, - struct intel_hdcp_gsc_message *hdcp_message, + struct intel_hdcp_gsc_context *gsc_context, u32 msg_size_in, u32 msg_size_out, u32 addr_out_off) { - struct xe_gt *gt = hdcp_message->hdcp_bo->tile->media_gt; - struct iosys_map *map = &hdcp_message->hdcp_bo->vmap; + struct xe_gt *gt = gsc_context->hdcp_bo->tile->media_gt; + struct iosys_map *map = &gsc_context->hdcp_bo->vmap; struct xe_gsc *gsc = >->uc.gsc; int ret; - ret = xe_gsc_pkt_submit_kernel(gsc, hdcp_message->hdcp_cmd_in, msg_size_in, - hdcp_message->hdcp_cmd_out, msg_size_out); + ret = xe_gsc_pkt_submit_kernel(gsc, gsc_context->hdcp_cmd_in, msg_size_in, + gsc_context->hdcp_cmd_out, msg_size_out); if (ret) { drm_err(&xe->drm, "failed to send gsc HDCP msg (%d)\n", ret); return ret; @@ -205,12 +155,12 @@ static int xe_gsc_send_sync(struct xe_device *xe, return ret; } -ssize_t intel_hdcp_gsc_msg_send(struct xe_device *xe, u8 *msg_in, - size_t msg_in_len, u8 *msg_out, - size_t msg_out_len) +ssize_t intel_hdcp_gsc_msg_send(struct intel_hdcp_gsc_context *gsc_context, + void *msg_in, size_t msg_in_len, + void *msg_out, size_t msg_out_len) { + struct xe_device *xe = gsc_context->xe; const size_t max_msg_size = PAGE_SIZE - HDCP_GSC_HEADER_SIZE; - struct intel_hdcp_gsc_message *hdcp_message; u64 host_session_id; u32 msg_size_in, msg_size_out; u32 addr_out_off, addr_in_wr_off = 0; @@ -223,15 +173,14 @@ ssize_t intel_hdcp_gsc_msg_send(struct xe_device *xe, u8 *msg_in, msg_size_in = msg_in_len + HDCP_GSC_HEADER_SIZE; msg_size_out = msg_out_len + HDCP_GSC_HEADER_SIZE; - hdcp_message = xe->display.hdcp.hdcp_message; addr_out_off = PAGE_SIZE; host_session_id = xe_gsc_create_host_session_id(); xe_pm_runtime_get_noresume(xe); - addr_in_wr_off = xe_gsc_emit_header(xe, &hdcp_message->hdcp_bo->vmap, + addr_in_wr_off = xe_gsc_emit_header(xe, &gsc_context->hdcp_bo->vmap, addr_in_wr_off, HECI_MEADDRESS_HDCP, host_session_id, msg_in_len); - xe_map_memcpy_to(xe, &hdcp_message->hdcp_bo->vmap, addr_in_wr_off, + xe_map_memcpy_to(xe, &gsc_context->hdcp_bo->vmap, addr_in_wr_off, msg_in, msg_in_len); /* * Keep sending request in case the pending bit is set no need to add @@ -240,7 +189,7 @@ ssize_t intel_hdcp_gsc_msg_send(struct xe_device *xe, u8 *msg_in, * 20 times each message 50 ms apart */ do { - ret = xe_gsc_send_sync(xe, hdcp_message, msg_size_in, msg_size_out, + ret = xe_gsc_send_sync(xe, gsc_context, msg_size_in, msg_size_out, addr_out_off); /* Only try again if gsc says so */ @@ -254,7 +203,7 @@ ssize_t intel_hdcp_gsc_msg_send(struct xe_device *xe, u8 *msg_in, if (ret) goto out; - xe_map_memcpy_from(xe, msg_out, &hdcp_message->hdcp_bo->vmap, + xe_map_memcpy_from(xe, msg_out, &gsc_context->hdcp_bo->vmap, addr_out_off + HDCP_GSC_HEADER_SIZE, msg_out_len); diff --git a/drivers/gpu/drm/xe/tests/xe_mocs.c b/drivers/gpu/drm/xe/tests/xe_mocs.c index ef1e5256c56a..0e502feaca81 100644 --- a/drivers/gpu/drm/xe/tests/xe_mocs.c +++ b/drivers/gpu/drm/xe/tests/xe_mocs.c @@ -46,8 +46,11 @@ static void read_l3cc_table(struct xe_gt *gt, unsigned int fw_ref, i; u32 reg_val; - fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT); - KUNIT_ASSERT_NE_MSG(test, fw_ref, 0, "Forcewake Failed.\n"); + fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL); + if (!xe_force_wake_ref_has_domain(fw_ref, XE_FORCEWAKE_ALL)) { + xe_force_wake_put(gt_to_fw(gt), fw_ref); + KUNIT_ASSERT_TRUE_MSG(test, true, "Forcewake Failed.\n"); + } for (i = 0; i < info->num_mocs_regs; i++) { if (!(i & 1)) { diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index 3a84a9d92c48..d99d91fe8aa9 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -2569,7 +2569,7 @@ static int gem_create_user_ext_set_property(struct xe_device *xe, int err; u32 idx; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; @@ -2606,7 +2606,7 @@ static int gem_create_user_extensions(struct xe_device *xe, struct xe_bo *bo, if (XE_IOCTL_DBG(xe, ext_number >= MAX_USER_EXTENSIONS)) return -E2BIG; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; diff --git a/drivers/gpu/drm/xe/xe_devcoredump.c b/drivers/gpu/drm/xe/xe_devcoredump.c index a9e618abf8ac..7a8af2311318 100644 --- a/drivers/gpu/drm/xe/xe_devcoredump.c +++ b/drivers/gpu/drm/xe/xe_devcoredump.c @@ -177,6 +177,8 @@ static ssize_t xe_devcoredump_read(char *buffer, loff_t offset, struct xe_devcoredump *coredump = data; struct xe_devcoredump_snapshot *ss; ssize_t byte_copied; + u32 chunk_offset; + ssize_t new_chunk_position; if (!coredump) return -ENODEV; @@ -201,10 +203,14 @@ static ssize_t xe_devcoredump_read(char *buffer, loff_t offset, return 0; } + new_chunk_position = div_u64_rem(offset, + XE_DEVCOREDUMP_CHUNK_MAX, + &chunk_offset); + if (offset >= ss->read.chunk_position + XE_DEVCOREDUMP_CHUNK_MAX || offset < ss->read.chunk_position) { - ss->read.chunk_position = - ALIGN_DOWN(offset, XE_DEVCOREDUMP_CHUNK_MAX); + ss->read.chunk_position = new_chunk_position * + XE_DEVCOREDUMP_CHUNK_MAX; __xe_devcoredump_read(ss->read.buffer, XE_DEVCOREDUMP_CHUNK_MAX, @@ -213,8 +219,7 @@ static ssize_t xe_devcoredump_read(char *buffer, loff_t offset, byte_copied = count < ss->read.size - offset ? count : ss->read.size - offset; - memcpy(buffer, ss->read.buffer + - (offset % XE_DEVCOREDUMP_CHUNK_MAX), byte_copied); + memcpy(buffer, ss->read.buffer + chunk_offset, byte_copied); mutex_unlock(&coredump->lock); diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index 75e753e0a682..c02c4c4e9412 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -26,6 +26,7 @@ #include "xe_bo_evict.h" #include "xe_debugfs.h" #include "xe_devcoredump.h" +#include "xe_device_sysfs.h" #include "xe_dma_buf.h" #include "xe_drm_client.h" #include "xe_drv.h" @@ -915,6 +916,10 @@ int xe_device_probe(struct xe_device *xe) if (err) goto err_unregister_display; + err = xe_device_sysfs_init(xe); + if (err) + goto err_unregister_display; + xe_debugfs_register(xe); err = xe_hwmon_register(xe); diff --git a/drivers/gpu/drm/xe/xe_device_sysfs.c b/drivers/gpu/drm/xe/xe_device_sysfs.c index 7efbd4c52791..2e657692e5b5 100644 --- a/drivers/gpu/drm/xe/xe_device_sysfs.c +++ b/drivers/gpu/drm/xe/xe_device_sysfs.c @@ -3,14 +3,16 @@ * Copyright © 2023 Intel Corporation */ +#include <linux/device.h> #include <linux/kobject.h> #include <linux/pci.h> #include <linux/sysfs.h> -#include <drm/drm_managed.h> - #include "xe_device.h" #include "xe_device_sysfs.h" +#include "xe_mmio.h" +#include "xe_pcode_api.h" +#include "xe_pcode.h" #include "xe_pm.h" /** @@ -63,11 +65,94 @@ vram_d3cold_threshold_store(struct device *dev, struct device_attribute *attr, static DEVICE_ATTR_RW(vram_d3cold_threshold); +/** + * DOC: PCIe Gen5 Limitations + * + * Default link speed of discrete GPUs is determined by configuration parameters + * stored in their flash memory, which are subject to override through user + * initiated firmware updates. It has been observed that devices configured with + * PCIe Gen5 as their default link speed can come across link quality issues due + * to host or motherboard limitations and may have to auto-downgrade their link + * to PCIe Gen4 speed when faced with unstable link at Gen5, which makes + * firmware updates rather risky on such setups. It is required to ensure that + * the device is capable of auto-downgrading its link to PCIe Gen4 speed before + * pushing the firmware image with PCIe Gen5 as default configuration. This can + * be done by reading ``auto_link_downgrade_capable`` sysfs entry, which will + * denote if the device is capable of auto-downgrading its link to PCIe Gen4 + * speed with boolean output value of ``0`` or ``1``, meaning `incapable` or + * `capable` respectively. + * + * .. code-block:: shell + * + * $ cat /sys/bus/pci/devices/<bdf>/auto_link_downgrade_capable + * + * Pushing the firmware image with PCIe Gen5 as default configuration on a auto + * link downgrade incapable device and facing link instability due to host or + * motherboard limitations can result in driver failing to bind to the device, + * making further firmware updates impossible with RMA being the only last + * resort. + * + * Link downgrade status of auto link downgrade capable devices is available + * through ``auto_link_downgrade_status`` sysfs entry with boolean output value + * of ``0`` or ``1``, where ``0`` means no auto-downgrading was required during + * link training (which is the optimal scenario) and ``1`` means the device has + * auto-downgraded its link to PCIe Gen4 speed due to unstable Gen5 link. + * + * .. code-block:: shell + * + * $ cat /sys/bus/pci/devices/<bdf>/auto_link_downgrade_status + */ + +static ssize_t +auto_link_downgrade_capable_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct xe_device *xe = pdev_to_xe_device(pdev); + u32 cap, val; + + xe_pm_runtime_get(xe); + val = xe_mmio_read32(xe_root_tile_mmio(xe), BMG_PCIE_CAP); + xe_pm_runtime_put(xe); + + cap = REG_FIELD_GET(LINK_DOWNGRADE, val); + return sysfs_emit(buf, "%u\n", cap == DOWNGRADE_CAPABLE ? true : false); +} +static DEVICE_ATTR_ADMIN_RO(auto_link_downgrade_capable); + +static ssize_t +auto_link_downgrade_status_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct xe_device *xe = pdev_to_xe_device(pdev); + /* default the auto_link_downgrade status to 0 */ + u32 val = 0; + int ret; + + xe_pm_runtime_get(xe); + ret = xe_pcode_read(xe_device_get_root_tile(xe), + PCODE_MBOX(DGFX_PCODE_STATUS, DGFX_GET_INIT_STATUS, 0), + &val, NULL); + xe_pm_runtime_put(xe); + + return ret ?: sysfs_emit(buf, "%u\n", REG_FIELD_GET(DGFX_LINK_DOWNGRADE_STATUS, val)); +} +static DEVICE_ATTR_ADMIN_RO(auto_link_downgrade_status); + +static const struct attribute *auto_link_downgrade_attrs[] = { + &dev_attr_auto_link_downgrade_capable.attr, + &dev_attr_auto_link_downgrade_status.attr, + NULL +}; + static void xe_device_sysfs_fini(void *arg) { struct xe_device *xe = arg; - sysfs_remove_file(&xe->drm.dev->kobj, &dev_attr_vram_d3cold_threshold.attr); + if (xe->d3cold.capable) + sysfs_remove_file(&xe->drm.dev->kobj, &dev_attr_vram_d3cold_threshold.attr); + + if (xe->info.platform == XE_BATTLEMAGE) + sysfs_remove_files(&xe->drm.dev->kobj, auto_link_downgrade_attrs); } int xe_device_sysfs_init(struct xe_device *xe) @@ -75,9 +160,17 @@ int xe_device_sysfs_init(struct xe_device *xe) struct device *dev = xe->drm.dev; int ret; - ret = sysfs_create_file(&dev->kobj, &dev_attr_vram_d3cold_threshold.attr); - if (ret) - return ret; + if (xe->d3cold.capable) { + ret = sysfs_create_file(&dev->kobj, &dev_attr_vram_d3cold_threshold.attr); + if (ret) + return ret; + } + + if (xe->info.platform == XE_BATTLEMAGE) { + ret = sysfs_create_files(&dev->kobj, auto_link_downgrade_attrs); + if (ret) + return ret; + } return devm_add_action_or_reset(dev, xe_device_sysfs_fini, xe); } diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h index 495bc00ebed4..06c65dace026 100644 --- a/drivers/gpu/drm/xe/xe_device_types.h +++ b/drivers/gpu/drm/xe/xe_device_types.h @@ -31,7 +31,6 @@ #endif #if IS_ENABLED(CONFIG_DRM_XE_DISPLAY) -#include "soc/intel_pch.h" #include "intel_display_core.h" #include "intel_display_device.h" #endif @@ -589,7 +588,6 @@ struct xe_device { * migrating to the right sub-structs */ struct intel_display display; - enum intel_pch pch_type; struct dram_info { bool wm_lv_0_adjust_needed; diff --git a/drivers/gpu/drm/xe/xe_eu_stall.c b/drivers/gpu/drm/xe/xe_eu_stall.c index e2bb156c71fb..96732613b4b7 100644 --- a/drivers/gpu/drm/xe/xe_eu_stall.c +++ b/drivers/gpu/drm/xe/xe_eu_stall.c @@ -283,7 +283,7 @@ static int xe_eu_stall_user_ext_set_property(struct xe_device *xe, u64 extension int err; u32 idx; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; @@ -313,7 +313,7 @@ static int xe_eu_stall_user_extensions(struct xe_device *xe, u64 extension, if (XE_IOCTL_DBG(xe, ext_number >= MAX_USER_EXTENSIONS)) return -E2BIG; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index b75adfc99fb7..44364c042ad7 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -176,8 +176,8 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) } if (xe_exec_queue_is_parallel(q)) { - err = __copy_from_user(addresses, addresses_user, sizeof(u64) * - q->width); + err = copy_from_user(addresses, addresses_user, sizeof(u64) * + q->width); if (err) { err = -EFAULT; goto err_syncs; diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index 606922d9dd73..21d4ced31dd9 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -479,7 +479,7 @@ static int exec_queue_user_ext_set_property(struct xe_device *xe, int err; u32 idx; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; @@ -518,7 +518,7 @@ static int exec_queue_user_extensions(struct xe_device *xe, struct xe_exec_queue if (XE_IOCTL_DBG(xe, ext_number >= MAX_USER_EXTENSIONS)) return -E2BIG; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; @@ -618,9 +618,8 @@ int xe_exec_queue_create_ioctl(struct drm_device *dev, void *data, if (XE_IOCTL_DBG(xe, !len || len > XE_HW_ENGINE_MAX_INSTANCE)) return -EINVAL; - err = __copy_from_user(eci, user_eci, - sizeof(struct drm_xe_engine_class_instance) * - len); + err = copy_from_user(eci, user_eci, + sizeof(struct drm_xe_engine_class_instance) * len); if (XE_IOCTL_DBG(xe, err)) return -EFAULT; diff --git a/drivers/gpu/drm/xe/xe_gsc.c b/drivers/gpu/drm/xe/xe_gsc.c index fd41113f8572..0bcf97063ff6 100644 --- a/drivers/gpu/drm/xe/xe_gsc.c +++ b/drivers/gpu/drm/xe/xe_gsc.c @@ -555,6 +555,28 @@ void xe_gsc_wait_for_worker_completion(struct xe_gsc *gsc) flush_work(&gsc->work); } +void xe_gsc_stop_prepare(struct xe_gsc *gsc) +{ + struct xe_gt *gt = gsc_to_gt(gsc); + int ret; + + if (!xe_uc_fw_is_loadable(&gsc->fw) || xe_uc_fw_is_in_error_state(&gsc->fw)) + return; + + xe_force_wake_assert_held(gt_to_fw(gt), XE_FW_GSC); + + /* + * If the GSC FW load or the proxy init are interrupted, the only way + * to recover it is to do an FLR and reload the GSC from scratch. + * Therefore, let's wait for the init to complete before stopping + * operations. The proxy init is the last step, so we can just wait on + * that + */ + ret = xe_gsc_wait_for_proxy_init_done(gsc); + if (ret) + xe_gt_err(gt, "failed to wait for GSC init completion before uc stop\n"); +} + /* * wa_14015076503: if the GSC FW is loaded, we need to alert it before doing a * GSC engine reset by writing a notification bit in the GS1 register and then diff --git a/drivers/gpu/drm/xe/xe_gsc.h b/drivers/gpu/drm/xe/xe_gsc.h index d99f66c38075..b8b8e0810ad9 100644 --- a/drivers/gpu/drm/xe/xe_gsc.h +++ b/drivers/gpu/drm/xe/xe_gsc.h @@ -16,6 +16,7 @@ struct xe_hw_engine; int xe_gsc_init(struct xe_gsc *gsc); int xe_gsc_init_post_hwconfig(struct xe_gsc *gsc); void xe_gsc_wait_for_worker_completion(struct xe_gsc *gsc); +void xe_gsc_stop_prepare(struct xe_gsc *gsc); void xe_gsc_load_start(struct xe_gsc *gsc); void xe_gsc_hwe_irq_handler(struct xe_hw_engine *hwe, u16 intr_vec); diff --git a/drivers/gpu/drm/xe/xe_gsc_proxy.c b/drivers/gpu/drm/xe/xe_gsc_proxy.c index 8cf70b228ff3..d0519cd6704a 100644 --- a/drivers/gpu/drm/xe/xe_gsc_proxy.c +++ b/drivers/gpu/drm/xe/xe_gsc_proxy.c @@ -71,6 +71,17 @@ bool xe_gsc_proxy_init_done(struct xe_gsc *gsc) HECI1_FWSTS1_PROXY_STATE_NORMAL; } +int xe_gsc_wait_for_proxy_init_done(struct xe_gsc *gsc) +{ + struct xe_gt *gt = gsc_to_gt(gsc); + + /* Proxy init can take up to 500ms, so wait double that for safety */ + return xe_mmio_wait32(>->mmio, HECI_FWSTS1(MTL_GSC_HECI1_BASE), + HECI1_FWSTS1_CURRENT_STATE, + HECI1_FWSTS1_PROXY_STATE_NORMAL, + USEC_PER_SEC, NULL, false); +} + static void __gsc_proxy_irq_rmw(struct xe_gsc *gsc, u32 clr, u32 set) { struct xe_gt *gt = gsc_to_gt(gsc); diff --git a/drivers/gpu/drm/xe/xe_gsc_proxy.h b/drivers/gpu/drm/xe/xe_gsc_proxy.h index fdef56995cd4..765602221dbc 100644 --- a/drivers/gpu/drm/xe/xe_gsc_proxy.h +++ b/drivers/gpu/drm/xe/xe_gsc_proxy.h @@ -12,6 +12,7 @@ struct xe_gsc; int xe_gsc_proxy_init(struct xe_gsc *gsc); bool xe_gsc_proxy_init_done(struct xe_gsc *gsc); +int xe_gsc_wait_for_proxy_init_done(struct xe_gsc *gsc); int xe_gsc_proxy_start(struct xe_gsc *gsc); int xe_gsc_proxy_request_handler(struct xe_gsc *gsc); diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c index 8068b4bc0a09..0e5d243c9451 100644 --- a/drivers/gpu/drm/xe/xe_gt.c +++ b/drivers/gpu/drm/xe/xe_gt.c @@ -899,7 +899,7 @@ void xe_gt_suspend_prepare(struct xe_gt *gt) fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL); - xe_uc_stop_prepare(>->uc); + xe_uc_suspend_prepare(>->uc); xe_force_wake_put(gt_to_fw(gt), fw_ref); } diff --git a/drivers/gpu/drm/xe/xe_gt_debugfs.c b/drivers/gpu/drm/xe/xe_gt_debugfs.c index a88076e9cc7d..119a55bb7580 100644 --- a/drivers/gpu/drm/xe/xe_gt_debugfs.c +++ b/drivers/gpu/drm/xe/xe_gt_debugfs.c @@ -92,22 +92,23 @@ static int hw_engines(struct xe_gt *gt, struct drm_printer *p) struct xe_hw_engine *hwe; enum xe_hw_engine_id id; unsigned int fw_ref; + int ret = 0; xe_pm_runtime_get(xe); fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL); if (!xe_force_wake_ref_has_domain(fw_ref, XE_FORCEWAKE_ALL)) { - xe_pm_runtime_put(xe); - xe_force_wake_put(gt_to_fw(gt), fw_ref); - return -ETIMEDOUT; + ret = -ETIMEDOUT; + goto fw_put; } for_each_hw_engine(hwe, gt, id) xe_hw_engine_print(hwe, p); +fw_put: xe_force_wake_put(gt_to_fw(gt), fw_ref); xe_pm_runtime_put(xe); - return 0; + return ret; } static int powergate_info(struct xe_gt *gt, struct drm_printer *p) diff --git a/drivers/gpu/drm/xe/xe_guc.c b/drivers/gpu/drm/xe/xe_guc.c index c5aace59b62c..bac5471a1a78 100644 --- a/drivers/gpu/drm/xe/xe_guc.c +++ b/drivers/gpu/drm/xe/xe_guc.c @@ -1510,30 +1510,32 @@ void xe_guc_print_info(struct xe_guc *guc, struct drm_printer *p) xe_uc_fw_print(&guc->fw, p); - fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT); - if (!fw_ref) - return; + if (!IS_SRIOV_VF(gt_to_xe(gt))) { + fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT); + if (!fw_ref) + return; + + status = xe_mmio_read32(>->mmio, GUC_STATUS); + + drm_printf(p, "\nGuC status 0x%08x:\n", status); + drm_printf(p, "\tBootrom status = 0x%x\n", + REG_FIELD_GET(GS_BOOTROM_MASK, status)); + drm_printf(p, "\tuKernel status = 0x%x\n", + REG_FIELD_GET(GS_UKERNEL_MASK, status)); + drm_printf(p, "\tMIA Core status = 0x%x\n", + REG_FIELD_GET(GS_MIA_MASK, status)); + drm_printf(p, "\tLog level = %d\n", + xe_guc_log_get_level(&guc->log)); + + drm_puts(p, "\nScratch registers:\n"); + for (i = 0; i < SOFT_SCRATCH_COUNT; i++) { + drm_printf(p, "\t%2d: \t0x%x\n", + i, xe_mmio_read32(>->mmio, SOFT_SCRATCH(i))); + } - status = xe_mmio_read32(>->mmio, GUC_STATUS); - - drm_printf(p, "\nGuC status 0x%08x:\n", status); - drm_printf(p, "\tBootrom status = 0x%x\n", - REG_FIELD_GET(GS_BOOTROM_MASK, status)); - drm_printf(p, "\tuKernel status = 0x%x\n", - REG_FIELD_GET(GS_UKERNEL_MASK, status)); - drm_printf(p, "\tMIA Core status = 0x%x\n", - REG_FIELD_GET(GS_MIA_MASK, status)); - drm_printf(p, "\tLog level = %d\n", - xe_guc_log_get_level(&guc->log)); - - drm_puts(p, "\nScratch registers:\n"); - for (i = 0; i < SOFT_SCRATCH_COUNT; i++) { - drm_printf(p, "\t%2d: \t0x%x\n", - i, xe_mmio_read32(>->mmio, SOFT_SCRATCH(i))); + xe_force_wake_put(gt_to_fw(gt), fw_ref); } - xe_force_wake_put(gt_to_fw(gt), fw_ref); - drm_puts(p, "\n"); xe_guc_ct_print(&guc->ct, p, false); diff --git a/drivers/gpu/drm/xe/xe_guc_debugfs.c b/drivers/gpu/drm/xe/xe_guc_debugfs.c index f33013f8a0f3..0b102ab46c4d 100644 --- a/drivers/gpu/drm/xe/xe_guc_debugfs.c +++ b/drivers/gpu/drm/xe/xe_guc_debugfs.c @@ -113,23 +113,34 @@ static const struct drm_info_list vf_safe_debugfs_list[] = { { "guc_ctb", .show = guc_debugfs_show, .data = guc_ctb }, }; +/* For GuC debugfs files that require the SLPC support */ +static const struct drm_info_list slpc_debugfs_list[] = { + { "guc_pc", .show = guc_debugfs_show, .data = guc_pc }, +}; + /* everything else should be added here */ static const struct drm_info_list pf_only_debugfs_list[] = { { "guc_log", .show = guc_debugfs_show, .data = guc_log }, { "guc_log_dmesg", .show = guc_debugfs_show, .data = guc_log_dmesg }, - { "guc_pc", .show = guc_debugfs_show, .data = guc_pc }, }; void xe_guc_debugfs_register(struct xe_guc *guc, struct dentry *parent) { - struct drm_minor *minor = guc_to_xe(guc)->drm.primary; + struct xe_device *xe = guc_to_xe(guc); + struct drm_minor *minor = xe->drm.primary; drm_debugfs_create_files(vf_safe_debugfs_list, ARRAY_SIZE(vf_safe_debugfs_list), parent, minor); - if (!IS_SRIOV_VF(guc_to_xe(guc))) + if (!IS_SRIOV_VF(xe)) { drm_debugfs_create_files(pf_only_debugfs_list, ARRAY_SIZE(pf_only_debugfs_list), parent, minor); + + if (!xe->info.skip_guc_pc) + drm_debugfs_create_files(slpc_debugfs_list, + ARRAY_SIZE(slpc_debugfs_list), + parent, minor); + } } diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c index 813c3c0bb250..369be36f7dc5 100644 --- a/drivers/gpu/drm/xe/xe_guc_submit.c +++ b/drivers/gpu/drm/xe/xe_guc_submit.c @@ -1179,9 +1179,12 @@ trigger_reset: process_name = q->vm->xef->process_name; pid = q->vm->xef->pid; } - xe_gt_notice(guc_to_gt(guc), "Timedout job: seqno=%u, lrc_seqno=%u, guc_id=%d, flags=0x%lx in %s [%d]", - xe_sched_job_seqno(job), xe_sched_job_lrc_seqno(job), - q->guc->id, q->flags, process_name, pid); + + if (!exec_queue_killed(q)) + xe_gt_notice(guc_to_gt(guc), + "Timedout job: seqno=%u, lrc_seqno=%u, guc_id=%d, flags=0x%lx in %s [%d]", + xe_sched_job_seqno(job), xe_sched_job_lrc_seqno(job), + q->guc->id, q->flags, process_name, pid); trace_xe_sched_job_timedout(job); diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index 346f357b3d1f..fb842fa0552e 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -1301,7 +1301,7 @@ static int xe_oa_user_ext_set_property(struct xe_oa *oa, enum xe_oa_user_extn_fr int err; u32 idx; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(oa->xe, err)) return -EFAULT; @@ -1338,7 +1338,7 @@ static int xe_oa_user_extensions(struct xe_oa *oa, enum xe_oa_user_extn_from fro if (XE_IOCTL_DBG(oa->xe, ext_number >= MAX_USER_EXTENSIONS)) return -E2BIG; - err = __copy_from_user(&ext, address, sizeof(ext)); + err = copy_from_user(&ext, address, sizeof(ext)); if (XE_IOCTL_DBG(oa->xe, err)) return -EFAULT; @@ -2281,7 +2281,7 @@ int xe_oa_add_config_ioctl(struct drm_device *dev, u64 data, struct drm_file *fi return -EACCES; } - err = __copy_from_user(¶m, u64_to_user_ptr(data), sizeof(param)); + err = copy_from_user(¶m, u64_to_user_ptr(data), sizeof(param)); if (XE_IOCTL_DBG(oa->xe, err)) return -EFAULT; diff --git a/drivers/gpu/drm/xe/xe_pcode_api.h b/drivers/gpu/drm/xe/xe_pcode_api.h index e622ae17f08d..127d4d26c4cf 100644 --- a/drivers/gpu/drm/xe/xe_pcode_api.h +++ b/drivers/gpu/drm/xe/xe_pcode_api.h @@ -34,6 +34,7 @@ #define DGFX_PCODE_STATUS 0x7E #define DGFX_GET_INIT_STATUS 0x0 #define DGFX_INIT_STATUS_COMPLETE 0x1 +#define DGFX_LINK_DOWNGRADE_STATUS REG_BIT(31) #define PCODE_POWER_SETUP 0x7C #define POWER_SETUP_SUBCOMMAND_READ_I1 0x4 @@ -66,6 +67,10 @@ /* Auxiliary info bits */ #define AUXINFO_HISTORY_OFFSET REG_GENMASK(31, 29) +#define BMG_PCIE_CAP XE_REG(0x138340) +#define LINK_DOWNGRADE REG_GENMASK(1, 0) +#define DOWNGRADE_CAPABLE 2 + struct pcode_err_decode { int errno; const char *str; diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c index 38514cef817e..693866def183 100644 --- a/drivers/gpu/drm/xe/xe_pm.c +++ b/drivers/gpu/drm/xe/xe_pm.c @@ -16,7 +16,6 @@ #include "xe_bo.h" #include "xe_bo_evict.h" #include "xe_device.h" -#include "xe_device_sysfs.h" #include "xe_ggtt.h" #include "xe_gt.h" #include "xe_guc.h" @@ -273,6 +272,7 @@ int xe_pm_init_early(struct xe_device *xe) if (err) return err; + xe->d3cold.capable = xe_pm_pci_d3cold_capable(xe); return 0; } ALLOW_ERROR_INJECTION(xe_pm_init_early, ERRNO); /* See xe_pci_probe() */ @@ -344,13 +344,7 @@ int xe_pm_init(struct xe_device *xe) if (!xe_device_uc_enabled(xe)) return 0; - xe->d3cold.capable = xe_pm_pci_d3cold_capable(xe); - if (xe->d3cold.capable) { - err = xe_device_sysfs_init(xe); - if (err) - goto err_unregister; - vram_threshold = vram_threshold_value(xe); err = xe_pm_set_vram_threshold(xe, vram_threshold); if (err) diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c index 890f6b2f40e9..d25f02c8d7fc 100644 --- a/drivers/gpu/drm/xe/xe_svm.c +++ b/drivers/gpu/drm/xe/xe_svm.c @@ -726,7 +726,6 @@ static int xe_svm_alloc_vram(struct xe_vm *vm, struct xe_tile *tile, } #endif - /** * xe_svm_handle_pagefault() - SVM handle page fault * @vm: The VM. @@ -966,3 +965,15 @@ int xe_devm_add(struct xe_tile *tile, struct xe_vram_region *vr) return 0; } #endif + +/** + * xe_svm_flush() - SVM flush + * @vm: The VM. + * + * Flush all SVM actions. + */ +void xe_svm_flush(struct xe_vm *vm) +{ + if (xe_vm_in_fault_mode(vm)) + flush_work(&vm->svm.garbage_collector.work); +} diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h index 3d441eb1f7ea..2881af1e60b2 100644 --- a/drivers/gpu/drm/xe/xe_svm.h +++ b/drivers/gpu/drm/xe/xe_svm.h @@ -96,6 +96,8 @@ static inline bool xe_svm_range_has_dma_mapping(struct xe_svm_range *range) #define xe_svm_notifier_unlock(vm__) \ drm_gpusvm_notifier_unlock(&(vm__)->svm.gpusvm) +void xe_svm_flush(struct xe_vm *vm); + #else #include <linux/interval_tree.h> @@ -180,5 +182,9 @@ static inline void xe_svm_notifier_lock(struct xe_vm *vm) static inline void xe_svm_notifier_unlock(struct xe_vm *vm) { } + +static inline void xe_svm_flush(struct xe_vm *vm) +{ +} #endif #endif diff --git a/drivers/gpu/drm/xe/xe_uc.c b/drivers/gpu/drm/xe/xe_uc.c index c14bd2282044..3a8751a8b92d 100644 --- a/drivers/gpu/drm/xe/xe_uc.c +++ b/drivers/gpu/drm/xe/xe_uc.c @@ -244,7 +244,7 @@ void xe_uc_gucrc_disable(struct xe_uc *uc) void xe_uc_stop_prepare(struct xe_uc *uc) { - xe_gsc_wait_for_worker_completion(&uc->gsc); + xe_gsc_stop_prepare(&uc->gsc); xe_guc_stop_prepare(&uc->guc); } @@ -278,6 +278,12 @@ again: goto again; } +void xe_uc_suspend_prepare(struct xe_uc *uc) +{ + xe_gsc_wait_for_worker_completion(&uc->gsc); + xe_guc_stop_prepare(&uc->guc); +} + int xe_uc_suspend(struct xe_uc *uc) { /* GuC submission not enabled, nothing to do */ diff --git a/drivers/gpu/drm/xe/xe_uc.h b/drivers/gpu/drm/xe/xe_uc.h index 3813c1ede450..c23e6f5e2514 100644 --- a/drivers/gpu/drm/xe/xe_uc.h +++ b/drivers/gpu/drm/xe/xe_uc.h @@ -18,6 +18,7 @@ int xe_uc_reset_prepare(struct xe_uc *uc); void xe_uc_stop_prepare(struct xe_uc *uc); void xe_uc_stop(struct xe_uc *uc); int xe_uc_start(struct xe_uc *uc); +void xe_uc_suspend_prepare(struct xe_uc *uc); int xe_uc_suspend(struct xe_uc *uc); int xe_uc_sanitize_reset(struct xe_uc *uc); void xe_uc_declare_wedged(struct xe_uc *uc); diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 80e56e232685..79323c78130f 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -3101,9 +3101,9 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe, struct xe_vm *vm, if (!*bind_ops) return args->num_binds > 1 ? -ENOBUFS : -ENOMEM; - err = __copy_from_user(*bind_ops, bind_user, - sizeof(struct drm_xe_vm_bind_op) * - args->num_binds); + err = copy_from_user(*bind_ops, bind_user, + sizeof(struct drm_xe_vm_bind_op) * + args->num_binds); if (XE_IOCTL_DBG(xe, err)) { err = -EFAULT; goto free_bind_ops; @@ -3331,8 +3331,7 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) } /* Ensure all UNMAPs visible */ - if (xe_vm_in_fault_mode(vm)) - flush_work(&vm->svm.garbage_collector.work); + xe_svm_flush(vm); err = down_write_killable(&vm->lock); if (err) |