summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/Kconfig1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/Makefile8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/aldebaran.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu.h2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h5
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c20
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c26
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_device.c285
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c19
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_display.c80
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_display.h5
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c5
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c46
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c17
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c101
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_job.c17
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c19
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c24
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_object.c212
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_object.h12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c65
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c345
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h15
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c77
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h21
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c21
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c81
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h31
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c113
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h21
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c209
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h26
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c11
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/cik.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/cik_ih.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/cik_sdma.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/cz_ih.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/dce_v10_0.c31
-rw-r--r--drivers/gpu/drm/amd/amdgpu/dce_v11_0.c31
-rw-r--r--drivers/gpu/drm/amd/amdgpu/dce_v6_0.c31
-rw-r--r--drivers/gpu/drm/amd/amdgpu/dce_v8_0.c30
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c47
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h56
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm124
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c47
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c75
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c16
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c46
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c78
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c57
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c45
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c9
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c9
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c59
-rw-r--r--drivers/gpu/drm/amd/amdgpu/iceland_ih.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/ih_v6_0.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/ih_v6_1.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/ih_v7_0.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/imu_v11_0.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/imu_v12_0.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c104
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c20
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c708
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h29
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mes_v11_0.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mes_v12_0.c189
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/navi10_ih.c11
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c81
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nv.c26
-rw-r--r--drivers/gpu/drm/amd/amdgpu/psp_v13_0.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c138
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c304
-rw-r--r--drivers/gpu/drm/amd/amdgpu/si.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/si_dma.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/si_ih.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/soc15.c55
-rw-r--r--drivers/gpu/drm/amd/amdgpu/soc21.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/soc24.c27
-rw-r--r--drivers/gpu/drm/amd/amdgpu/ta_ras_if.h3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h24
-rw-r--r--drivers/gpu/drm/amd/amdgpu/tonga_ih.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/umc_v12_0.c267
-rw-r--r--drivers/gpu/drm/amd/amdgpu/umc_v12_0.h17
-rw-r--r--drivers/gpu/drm/amd/amdgpu/umc_v8_14.c160
-rw-r--r--drivers/gpu/drm/amd/amdgpu/umc_v8_14.h51
-rw-r--r--drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c17
-rw-r--r--drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c17
-rw-r--r--drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vce_v2_0.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vce_v3_0.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vce_v4_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c25
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c24
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c33
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c34
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c40
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c121
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c34
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c79
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h5
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c1118
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h29
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vega10_ih.c11
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vega20_ih.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vi.c49
161 files changed, 5697 insertions, 1397 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/Kconfig b/drivers/gpu/drm/amd/amdgpu/Kconfig
index 41fa3377d9cf..1a11cab741ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/Kconfig
+++ b/drivers/gpu/drm/amd/amdgpu/Kconfig
@@ -26,6 +26,7 @@ config DRM_AMDGPU
select DRM_BUDDY
select DRM_SUBALLOC_HELPER
select DRM_EXEC
+ select DRM_PANEL_BACKLIGHT_QUIRKS
# amdgpu depends on ACPI_VIDEO when ACPI is enabled, for select to work
# ACPI_VIDEO's dependencies must also be selected.
select INPUT if ACPI
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index c7b18c52825d..5b21674b07fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -1,5 +1,5 @@
#
-# Copyright 2017 Advanced Micro Devices, Inc.
+# Copyright 2017-2024 Advanced Micro Devices, Inc. All rights reserved.
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
@@ -105,7 +105,7 @@ amdgpu-y += \
# add UMC block
amdgpu-y += \
- umc_v6_0.o umc_v6_1.o umc_v6_7.o umc_v8_7.o umc_v8_10.o umc_v12_0.o
+ umc_v6_0.o umc_v6_1.o umc_v6_7.o umc_v8_7.o umc_v8_10.o umc_v12_0.o umc_v8_14.o
# add IH block
amdgpu-y += \
@@ -200,6 +200,7 @@ amdgpu-y += \
vcn_v4_0_3.o \
vcn_v4_0_5.o \
vcn_v5_0_0.o \
+ vcn_v5_0_1.o \
amdgpu_jpeg.o \
jpeg_v1_0.o \
jpeg_v2_0.o \
@@ -208,7 +209,8 @@ amdgpu-y += \
jpeg_v4_0.o \
jpeg_v4_0_3.o \
jpeg_v4_0_5.o \
- jpeg_v5_0_0.o
+ jpeg_v5_0_0.o \
+ jpeg_v5_0_1.o
# add VPE block
amdgpu-y += \
diff --git a/drivers/gpu/drm/amd/amdgpu/aldebaran.c b/drivers/gpu/drm/amd/amdgpu/aldebaran.c
index f44de9d4b6a1..e13fbd974141 100644
--- a/drivers/gpu/drm/amd/amdgpu/aldebaran.c
+++ b/drivers/gpu/drm/amd/amdgpu/aldebaran.c
@@ -334,6 +334,8 @@ aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
AMDGPU_INIT_LEVEL_RESET_RECOVERY);
dev_info(tmp_adev->dev,
"GPU reset succeeded, trying to resume\n");
+ /*TBD: Ideally should clear only GFX, SDMA blocks*/
+ amdgpu_ras_clear_err_state(tmp_adev);
r = aldebaran_mode2_restore_ip(tmp_adev);
if (r)
goto end;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 4653a8d2823a..69895fccb474 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -880,6 +880,7 @@ struct amdgpu_device {
bool need_swiotlb;
bool accel_working;
struct notifier_block acpi_nb;
+ struct notifier_block pm_nb;
struct amdgpu_i2c_chan *i2c_bus[AMDGPU_MAX_I2C_BUS];
struct debugfs_blob_wrapper debugfs_vbios_blob;
struct debugfs_blob_wrapper debugfs_discovery_blob;
@@ -1174,7 +1175,6 @@ struct amdgpu_device {
struct work_struct reset_work;
- bool job_hang;
bool dc_enabled;
/* Mask of active clusters */
uint32_t aid_mask;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
index 5ef6b745f222..f3289d289913 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
@@ -71,6 +71,11 @@ struct ras_query_context;
#define ACA_ERROR_CE_MASK BIT_MASK(ACA_ERROR_TYPE_CE)
#define ACA_ERROR_DEFERRED_MASK BIT_MASK(ACA_ERROR_TYPE_DEFERRED)
+#define mmSMNAID_AID0_MCA_SMU 0x03b30400 /* SMN AID AID0 */
+#define mmSMNAID_XCD0_MCA_SMU 0x36430400 /* SMN AID XCD0 */
+#define mmSMNAID_XCD1_MCA_SMU 0x38430400 /* SMN AID XCD1 */
+#define mmSMNXCD_XCD0_MCA_SMU 0x40430400 /* SMN XCD XCD0 */
+
enum aca_reg_idx {
ACA_REG_IDX_CTL = 0,
ACA_REG_IDX_STATUS = 1,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
index ec5e0dcf8613..deb0785350e8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
@@ -140,7 +140,7 @@ static int acp_poweroff(struct generic_pm_domain *genpd)
* 2. power off the acp tiles
* 3. check and enter ulv state
*/
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true, 0);
return 0;
}
@@ -157,7 +157,7 @@ static int acp_poweron(struct generic_pm_domain *genpd)
* 2. turn on acp clock
* 3. power on acp tiles
*/
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false, 0);
return 0;
}
@@ -236,7 +236,7 @@ static int acp_hw_init(struct amdgpu_ip_block *ip_block)
ip_block->version->major, ip_block->version->minor);
/* -ENODEV means board uses AZ rather than ACP */
if (r == -ENODEV) {
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true, 0);
return 0;
} else if (r) {
return r;
@@ -508,7 +508,7 @@ static int acp_hw_fini(struct amdgpu_ip_block *ip_block)
/* return early if no ACP */
if (!adev->acp.acp_genpd) {
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false, 0);
return 0;
}
@@ -565,7 +565,7 @@ static int acp_suspend(struct amdgpu_ip_block *ip_block)
/* power up on suspend */
if (!adev->acp.acp_cell)
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false, 0);
return 0;
}
@@ -575,7 +575,7 @@ static int acp_resume(struct amdgpu_ip_block *ip_block)
/* power down again on resume */
if (!adev->acp.acp_cell)
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true, 0);
return 0;
}
@@ -584,19 +584,19 @@ static bool acp_is_idle(void *handle)
return true;
}
-static int acp_set_clockgating_state(void *handle,
+static int acp_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int acp_set_powergating_state(void *handle,
+static int acp_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, enable);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, enable, 0);
return 0;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index c4e733c2e75e..2c1b38c5cfc6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -368,7 +368,7 @@ void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void **mem_obj)
{
struct amdgpu_bo **bo = (struct amdgpu_bo **) mem_obj;
- amdgpu_bo_reserve(*bo, true);
+ (void)amdgpu_bo_reserve(*bo, true);
amdgpu_bo_kunmap(*bo);
amdgpu_bo_unpin(*bo);
amdgpu_bo_unreserve(*bo);
@@ -725,7 +725,9 @@ void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle)
/* Disable GFXOFF and PG. Temporary workaround
* to fix some compute applications issue on GFX9.
*/
- adev->ip_blocks[AMD_IP_BLOCK_TYPE_GFX].version->funcs->set_powergating_state((void *)adev, state);
+ struct amdgpu_ip_block *gfx_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+ if (gfx_block != NULL)
+ gfx_block->version->funcs->set_powergating_state((void *)gfx_block, state);
}
amdgpu_dpm_switch_power_profile(adev,
PP_SMC_POWER_PROFILE_COMPUTE,
@@ -835,7 +837,7 @@ int amdgpu_amdkfd_unmap_hiq(struct amdgpu_device *adev, u32 doorbell_off,
if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
return -EINVAL;
- if (!kiq_ring->sched.ready || adev->job_hang)
+ if (!kiq_ring->sched.ready || amdgpu_in_reset(adev))
return 0;
ring_funcs = kzalloc(sizeof(*ring_funcs), GFP_KERNEL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 4b80ad860639..8af67f18500a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -433,6 +433,9 @@ void kgd2kfd_unlock_kfd(void);
int kgd2kfd_start_sched(struct kfd_dev *kfd, uint32_t node_id);
int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id);
bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id);
+bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry,
+ bool retry_fault);
+
#else
static inline int kgd2kfd_init(void)
{
@@ -518,5 +521,12 @@ static inline bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id)
{
return false;
}
+
+static inline bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry,
+ bool retry_fault)
+{
+ return false;
+}
+
#endif
#endif /* AMDGPU_AMDKFD_H_INCLUDED */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
index cc66ebb7bae1..441568163e20 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
@@ -1131,6 +1131,9 @@ uint64_t kgd_gfx_v9_hqd_get_pq_addr(struct amdgpu_device *adev,
uint32_t low, high;
uint64_t queue_addr = 0;
+ if (!amdgpu_gpu_recovery)
+ return 0;
+
kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
amdgpu_gfx_rlc_enter_safe_mode(adev, inst);
@@ -1179,6 +1182,9 @@ uint64_t kgd_gfx_v9_hqd_reset(struct amdgpu_device *adev,
uint32_t low, high, pipe_reset_data = 0;
uint64_t queue_addr = 0;
+ if (!amdgpu_gpu_recovery)
+ return 0;
+
kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
amdgpu_gfx_rlc_enter_safe_mode(adev, inst);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index f30548f4c3b3..1e998f972c30 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -730,7 +730,7 @@ kfd_mem_dmaunmap_userptr(struct kgd_mem *mem,
return;
amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
- ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+ (void)ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
sg_free_table(ttm->sg);
@@ -779,7 +779,7 @@ kfd_mem_dmaunmap_sg_bo(struct kgd_mem *mem,
}
amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
- ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+ (void)ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
dir = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
@@ -989,7 +989,7 @@ unwind:
if (!attachment[i])
continue;
if (attachment[i]->bo_va) {
- amdgpu_bo_reserve(bo[i], true);
+ (void)amdgpu_bo_reserve(bo[i], true);
if (--attachment[i]->bo_va->ref_count == 0)
amdgpu_vm_bo_del(adev, attachment[i]->bo_va);
amdgpu_bo_unreserve(bo[i]);
@@ -1259,11 +1259,11 @@ static int unmap_bo_from_gpuvm(struct kgd_mem *mem,
return -EBUSY;
}
- amdgpu_vm_bo_unmap(adev, bo_va, entry->va);
+ (void)amdgpu_vm_bo_unmap(adev, bo_va, entry->va);
- amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
+ (void)amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
- amdgpu_sync_fence(sync, bo_va->last_pt_update);
+ (void)amdgpu_sync_fence(sync, bo_va->last_pt_update);
return 0;
}
@@ -2352,7 +2352,7 @@ void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_mem *mem)
{
struct amdgpu_bo *bo = mem->bo;
- amdgpu_bo_reserve(bo, true);
+ (void)amdgpu_bo_reserve(bo, true);
amdgpu_bo_kunmap(bo);
amdgpu_bo_unpin(bo);
amdgpu_bo_unreserve(bo);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
index 45affc02548c..423fd2eebe1e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
@@ -47,35 +47,37 @@
/* Check if current bios is an ATOM BIOS.
* Return true if it is ATOM BIOS. Otherwise, return false.
*/
-static bool check_atom_bios(uint8_t *bios, size_t size)
+static bool check_atom_bios(struct amdgpu_device *adev, size_t size)
{
uint16_t tmp, bios_header_start;
+ uint8_t *bios = adev->bios;
if (!bios || size < 0x49) {
- DRM_INFO("vbios mem is null or mem size is wrong\n");
+ dev_dbg(adev->dev, "VBIOS mem is null or mem size is wrong\n");
return false;
}
if (!AMD_IS_VALID_VBIOS(bios)) {
- DRM_INFO("BIOS signature incorrect %x %x\n", bios[0], bios[1]);
+ dev_dbg(adev->dev, "VBIOS signature incorrect %x %x\n", bios[0],
+ bios[1]);
return false;
}
bios_header_start = bios[0x48] | (bios[0x49] << 8);
if (!bios_header_start) {
- DRM_INFO("Can't locate bios header\n");
+ dev_dbg(adev->dev, "Can't locate VBIOS header\n");
return false;
}
tmp = bios_header_start + 4;
if (size < tmp) {
- DRM_INFO("BIOS header is broken\n");
+ dev_dbg(adev->dev, "VBIOS header is broken\n");
return false;
}
if (!memcmp(bios + tmp, "ATOM", 4) ||
!memcmp(bios + tmp, "MOTA", 4)) {
- DRM_DEBUG("ATOMBIOS detected\n");
+ dev_dbg(adev->dev, "ATOMBIOS detected\n");
return true;
}
@@ -118,7 +120,7 @@ static bool amdgpu_read_bios_from_vram(struct amdgpu_device *adev)
memcpy_fromio(adev->bios, bios, size);
iounmap(bios);
- if (!check_atom_bios(adev->bios, size)) {
+ if (!check_atom_bios(adev, size)) {
kfree(adev->bios);
return false;
}
@@ -146,7 +148,7 @@ bool amdgpu_read_bios(struct amdgpu_device *adev)
memcpy_fromio(adev->bios, bios, size);
pci_unmap_rom(adev->pdev, bios);
- if (!check_atom_bios(adev->bios, size)) {
+ if (!check_atom_bios(adev, size)) {
kfree(adev->bios);
return false;
}
@@ -186,7 +188,7 @@ static bool amdgpu_read_bios_from_rom(struct amdgpu_device *adev)
/* read complete BIOS */
amdgpu_asic_read_bios_from_rom(adev, adev->bios, len);
- if (!check_atom_bios(adev->bios, len)) {
+ if (!check_atom_bios(adev, len)) {
kfree(adev->bios);
return false;
}
@@ -216,7 +218,7 @@ static bool amdgpu_read_platform_bios(struct amdgpu_device *adev)
memcpy_fromio(adev->bios, bios, romlen);
iounmap(bios);
- if (!check_atom_bios(adev->bios, romlen))
+ if (!check_atom_bios(adev, romlen))
goto free_bios;
adev->bios_size = romlen;
@@ -324,7 +326,7 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
break;
}
- if (!check_atom_bios(adev->bios, size)) {
+ if (!check_atom_bios(adev, size)) {
kfree(adev->bios);
return false;
}
@@ -389,7 +391,7 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
vhdr->ImageLength,
GFP_KERNEL);
- if (!check_atom_bios(adev->bios, vhdr->ImageLength)) {
+ if (!check_atom_bios(adev, vhdr->ImageLength)) {
kfree(adev->bios);
return false;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
index 16153d275d7a..68bce6a6d09d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
@@ -414,7 +414,9 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
return -EINVAL;
}
- err = amdgpu_ucode_request(adev, &adev->pm.fw, "%s", fw_name);
+ err = amdgpu_ucode_request(adev, &adev->pm.fw,
+ AMDGPU_UCODE_REQUIRED,
+ "%s", fw_name);
if (err) {
DRM_ERROR("Failed to load firmware \"%s\"", fw_name);
amdgpu_ucode_release(&adev->pm.fw);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 5df21529b3b1..5cc5f59e3018 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1105,7 +1105,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
* We can't use gang submit on with reserved VMIDs when the VM changes
* can't be invalidated by more than one engine at the same time.
*/
- if (p->gang_size > 1 && !p->adev->vm_manager.concurrent_flush) {
+ if (p->gang_size > 1 && !adev->vm_manager.concurrent_flush) {
for (i = 0; i < p->gang_size; ++i) {
struct drm_sched_entity *entity = p->entities[i];
struct drm_gpu_scheduler *sched = entity->rq->sched;
@@ -1189,7 +1189,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
if (!bo)
continue;
- amdgpu_vm_bo_invalidate(adev, bo, false);
+ amdgpu_vm_bo_invalidate(bo, false);
}
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index a68338cb7b4a..49ca8c814455 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -2095,6 +2095,7 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
if (amdgpu_umsch_mm & amdgpu_umsch_mm_fwlog)
amdgpu_debugfs_umsch_fwlog_init(adev, &adev->umsch_mm);
+ amdgpu_debugfs_vcn_sched_mask_init(adev);
amdgpu_debugfs_jpeg_sched_mask_init(adev);
amdgpu_debugfs_gfx_sched_mask_init(adev);
amdgpu_debugfs_compute_sched_mask_init(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index cd4fac120834..018dfccd771b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -199,14 +199,16 @@ void amdgpu_set_init_level(struct amdgpu_device *adev,
}
static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev);
+static int amdgpu_device_pm_notifier(struct notifier_block *nb, unsigned long mode,
+ void *data);
/**
* DOC: pcie_replay_count
*
* The amdgpu driver provides a sysfs API for reporting the total number
- * of PCIe replays (NAKs)
+ * of PCIe replays (NAKs).
* The file pcie_replay_count is used for this and returns the total
- * number of replays as a sum of the NAKs generated and NAKs received
+ * number of replays as a sum of the NAKs generated and NAKs received.
*/
static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev,
@@ -432,8 +434,8 @@ bool amdgpu_device_supports_boco(struct drm_device *dev)
* @dev: drm_device pointer
*
* Return:
- * 1 if the device supporte BACO;
- * 3 if the device support MACO (only works if BACO is supported)
+ * 1 if the device supports BACO;
+ * 3 if the device supports MACO (only works if BACO is supported)
* otherwise return 0.
*/
int amdgpu_device_supports_baco(struct drm_device *dev)
@@ -580,7 +582,7 @@ void amdgpu_device_mm_access(struct amdgpu_device *adev, loff_t pos,
}
/**
- * amdgpu_device_aper_access - access vram by vram aperature
+ * amdgpu_device_aper_access - access vram by vram aperture
*
* @adev: amdgpu_device pointer
* @pos: offset of the buffer in vram
@@ -671,7 +673,7 @@ bool amdgpu_device_skip_hw_access(struct amdgpu_device *adev)
* here is that the GPU reset is not running on another thread in parallel.
*
* For this we trylock the read side of the reset semaphore, if that succeeds
- * we know that the reset is not running in paralell.
+ * we know that the reset is not running in parallel.
*
* If the trylock fails we assert that we are either already holding the read
* side of the lock or are the reset thread itself and hold the write side of
@@ -1402,6 +1404,7 @@ static int amdgpu_device_asic_init(struct amdgpu_device *adev)
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0) ||
amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0)) {
amdgpu_psp_wait_for_bootloader(adev);
ret = amdgpu_atomfirmware_asic_init(adev, true);
@@ -1635,6 +1638,13 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev))
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 &&
+ adev->pdev->device == 0x731f &&
+ adev->pdev->subsystem_vendor == PCI_VENDOR_ID_DELL)
+ return 0;
+
/* PCI_EXT_CAP_ID_VNDR extended capability is located at 0x100 */
if (!pci_find_ext_capability(adev->pdev, PCI_EXT_CAP_ID_VNDR))
DRM_WARN("System can't access extended configuration space, please check!!\n");
@@ -1736,7 +1746,7 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev)
uint32_t fw_ver;
err = request_firmware(&adev->pm.fw, "amdgpu/fiji_smc.bin", adev->dev);
- /* force vPost if error occured */
+ /* force vPost if error occurred */
if (err)
return true;
@@ -2168,7 +2178,7 @@ int amdgpu_device_ip_set_clockgating_state(void *dev,
if (!adev->ip_blocks[i].version->funcs->set_clockgating_state)
continue;
r = adev->ip_blocks[i].version->funcs->set_clockgating_state(
- (void *)adev, state);
+ &adev->ip_blocks[i], state);
if (r)
DRM_ERROR("set_clockgating_state of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
@@ -2202,7 +2212,7 @@ int amdgpu_device_ip_set_powergating_state(void *dev,
if (!adev->ip_blocks[i].version->funcs->set_powergating_state)
continue;
r = adev->ip_blocks[i].version->funcs->set_powergating_state(
- (void *)adev, state);
+ &adev->ip_blocks[i], state);
if (r)
DRM_ERROR("set_powergating_state of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
@@ -2362,8 +2372,8 @@ int amdgpu_device_ip_block_add(struct amdgpu_device *adev,
break;
}
- DRM_INFO("add ip block number %d <%s>\n", adev->num_ip_blocks,
- ip_block_version->funcs->name);
+ dev_info(adev->dev, "detected ip block number %d <%s>\n",
+ adev->num_ip_blocks, ip_block_version->funcs->name);
adev->ip_blocks[adev->num_ip_blocks].adev = adev;
@@ -2381,7 +2391,7 @@ int amdgpu_device_ip_block_add(struct amdgpu_device *adev,
* the module parameter virtual_display. This feature provides a virtual
* display hardware on headless boards or in virtualized environments.
* This function parses and validates the configuration string specified by
- * the user and configues the virtual display configuration (number of
+ * the user and configures the virtual display configuration (number of
* virtual connectors, crtcs, etc.) specified.
*/
static void amdgpu_device_enable_virtual_display(struct amdgpu_device *adev)
@@ -2444,7 +2454,7 @@ void amdgpu_device_set_sriov_virtual_display(struct amdgpu_device *adev)
* @adev: amdgpu_device pointer
*
* Parses the asic configuration parameters specified in the gpu info
- * firmware and makes them availale to the driver for use in configuring
+ * firmware and makes them available to the driver for use in configuring
* the asic.
* Returns 0 on success, -EINVAL on failure.
*/
@@ -2485,6 +2495,7 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->firmware.gpu_info_fw,
+ AMDGPU_UCODE_OPTIONAL,
"amdgpu/%s_gpu_info.bin", chip_name);
if (err) {
dev_err(adev->dev,
@@ -2504,7 +2515,7 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
le32_to_cpu(hdr->header.ucode_array_offset_bytes));
/*
- * Should be droped when DAL no longer needs it.
+ * Should be dropped when DAL no longer needs it.
*/
if (adev->asic_type == CHIP_NAVI12)
goto parse_soc_bounding_box;
@@ -3064,7 +3075,7 @@ init_failed:
*
* Writes a reset magic value to the gart pointer in VRAM. The driver calls
* this function before a GPU reset. If the value is retained after a
- * GPU reset, VRAM has not been lost. Some GPU resets may destry VRAM contents.
+ * GPU reset, VRAM has not been lost. Some GPU resets may destroy VRAM contents.
*/
static void amdgpu_device_fill_reset_magic(struct amdgpu_device *adev)
{
@@ -3140,7 +3151,7 @@ int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_JPEG &&
adev->ip_blocks[i].version->funcs->set_clockgating_state) {
/* enable clockgating to save power */
- r = adev->ip_blocks[i].version->funcs->set_clockgating_state((void *)adev,
+ r = adev->ip_blocks[i].version->funcs->set_clockgating_state(&adev->ip_blocks[i],
state);
if (r) {
DRM_ERROR("set_clockgating_state(gate) of IP block <%s> failed %d\n",
@@ -3177,7 +3188,7 @@ int amdgpu_device_set_pg_state(struct amdgpu_device *adev,
adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_JPEG &&
adev->ip_blocks[i].version->funcs->set_powergating_state) {
/* enable powergating to save power */
- r = adev->ip_blocks[i].version->funcs->set_powergating_state((void *)adev,
+ r = adev->ip_blocks[i].version->funcs->set_powergating_state(&adev->ip_blocks[i],
state);
if (r) {
DRM_ERROR("set_powergating_state(gate) of IP block <%s> failed %d\n",
@@ -3379,7 +3390,7 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
- /* Workaroud for ASICs need to disable SMC first */
+ /* Workaround for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
@@ -3481,7 +3492,7 @@ static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
WARN_ON_ONCE(adev->gfx.gfx_off_state);
WARN_ON_ONCE(adev->gfx.gfx_off_req_count);
- if (!amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, true))
+ if (!amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, true, 0))
adev->gfx.gfx_off_state = true;
}
@@ -4309,7 +4320,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
/*
* Reset domain needs to be present early, before XGMI hive discovered
- * (if any) and intitialized to use reset sem and in_gpu reset flag
+ * (if any) and initialized to use reset sem and in_gpu reset flag
* early on during init and before calling to RREG32.
*/
adev->reset_domain = amdgpu_reset_create_reset_domain(SINGLE_DEVICE, "amdgpu-reset-dev");
@@ -4599,6 +4610,11 @@ fence_driver_init:
amdgpu_device_check_iommu_direct_map(adev);
+ adev->pm_nb.notifier_call = amdgpu_device_pm_notifier;
+ r = register_pm_notifier(&adev->pm_nb);
+ if (r)
+ goto failed;
+
return 0;
release_ras_con:
@@ -4663,6 +4679,8 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
drain_workqueue(adev->mman.bdev.wq);
adev->shutdown = true;
+ unregister_pm_notifier(&adev->pm_nb);
+
/* make sure IB test finished before entering exclusive mode
* to avoid preemption on IB test
*/
@@ -4781,8 +4799,8 @@ static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
{
int ret;
- /* No need to evict vram on APUs for suspend to ram or s2idle */
- if ((adev->in_s3 || adev->in_s0ix) && (adev->flags & AMD_IS_APU))
+ /* No need to evict vram on APUs unless going to S4 */
+ if (!adev->in_s4 && (adev->flags & AMD_IS_APU))
return 0;
ret = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM);
@@ -4795,6 +4813,41 @@ static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
* Suspend & resume.
*/
/**
+ * amdgpu_device_pm_notifier - Notification block for Suspend/Hibernate events
+ * @nb: notifier block
+ * @mode: suspend mode
+ * @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.
+ */
+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;
+ }
+
+ return NOTIFY_DONE;
+}
+
+/**
* amdgpu_device_prepare - prepare for device suspend
*
* @dev: drm dev pointer
@@ -4833,7 +4886,7 @@ int amdgpu_device_prepare(struct drm_device *dev)
return 0;
unprepare:
- adev->in_s0ix = adev->in_s3 = false;
+ adev->in_s0ix = adev->in_s3 = adev->in_s4 = false;
return r;
}
@@ -5184,7 +5237,7 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
if (r)
return r;
- amdgpu_ras_set_fed(adev, false);
+ amdgpu_ras_clear_err_state(adev);
amdgpu_irq_gpu_reset_resume_helper(adev);
/* some sw clean up VF needs to do before recover */
@@ -5241,16 +5294,18 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
}
/**
- * amdgpu_device_has_job_running - check if there is any job in mirror list
+ * amdgpu_device_has_job_running - check if there is any unfinished job
*
* @adev: amdgpu_device pointer
*
- * check if there is any job in mirror list
+ * check if there is any job running on the device when guest driver receives
+ * FLR notification from host driver. If there are still jobs running, then
+ * the guest driver will not respond the FLR reset. Instead, let the job hit
+ * the timeout and guest driver then issue the reset request.
*/
bool amdgpu_device_has_job_running(struct amdgpu_device *adev)
{
int i;
- struct drm_sched_job *job;
for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
struct amdgpu_ring *ring = adev->rings[i];
@@ -5258,11 +5313,7 @@ bool amdgpu_device_has_job_running(struct amdgpu_device *adev)
if (!amdgpu_ring_sched_ready(ring))
continue;
- spin_lock(&ring->sched.job_list_lock);
- job = list_first_entry_or_null(&ring->sched.pending_list,
- struct drm_sched_job, list);
- spin_unlock(&ring->sched.job_list_lock);
- if (job)
+ if (amdgpu_fence_count_emitted(ring))
return true;
}
return false;
@@ -5487,7 +5538,7 @@ int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context)
amdgpu_set_init_level(tmp_adev, init_level);
if (full_reset) {
/* post card */
- amdgpu_ras_set_fed(tmp_adev, false);
+ amdgpu_ras_clear_err_state(tmp_adev);
r = amdgpu_device_asic_init(tmp_adev);
if (r) {
dev_warn(tmp_adev->dev, "asic atom init failed!");
@@ -5821,6 +5872,18 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
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);
@@ -5903,7 +5966,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
amdgpu_amdkfd_pre_reset(tmp_adev, reset_context);
/*
- * Mark these ASICs to be reseted as untracked first
+ * Mark these ASICs to be reset as untracked first
* And add them back after reset completed
*/
amdgpu_unregister_gpu_instance(tmp_adev);
@@ -6102,19 +6165,56 @@ static void amdgpu_device_partner_bandwidth(struct amdgpu_device *adev,
}
/**
+ * amdgpu_device_gpu_bandwidth - find the bandwidth of the GPU
+ *
+ * @adev: amdgpu_device pointer
+ * @speed: pointer to the speed of the link
+ * @width: pointer to the width of the link
+ *
+ * Evaluate the hierarchy to find the speed and bandwidth capabilities of the
+ * AMD dGPU which may be a virtual upstream bridge.
+ */
+static void amdgpu_device_gpu_bandwidth(struct amdgpu_device *adev,
+ enum pci_bus_speed *speed,
+ enum pcie_link_width *width)
+{
+ struct pci_dev *parent = adev->pdev;
+
+ if (!speed || !width)
+ return;
+
+ parent = pci_upstream_bridge(parent);
+ if (parent && parent->vendor == PCI_VENDOR_ID_ATI) {
+ /* use the upstream/downstream switches internal to dGPU */
+ *speed = pcie_get_speed_cap(parent);
+ *width = pcie_get_width_cap(parent);
+ while ((parent = pci_upstream_bridge(parent))) {
+ if (parent->vendor == PCI_VENDOR_ID_ATI) {
+ /* use the upstream/downstream switches internal to dGPU */
+ *speed = pcie_get_speed_cap(parent);
+ *width = pcie_get_width_cap(parent);
+ }
+ }
+ } else {
+ /* use the device itself */
+ *speed = pcie_get_speed_cap(adev->pdev);
+ *width = pcie_get_width_cap(adev->pdev);
+ }
+}
+
+/**
* amdgpu_device_get_pcie_info - fence pcie info about the PCIE slot
*
* @adev: amdgpu_device pointer
*
- * Fetchs and stores in the driver the PCIE capabilities (gen speed
+ * Fetches and stores in the driver the PCIE capabilities (gen speed
* and lanes) of the slot the device is in. Handles APUs and
* virtualized environments where PCIE config space may not be available.
*/
static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
{
- struct pci_dev *pdev;
enum pci_bus_speed speed_cap, platform_speed_cap;
- enum pcie_link_width platform_link_width;
+ enum pcie_link_width platform_link_width, link_width;
if (amdgpu_pcie_gen_cap)
adev->pm.pcie_gen_mask = amdgpu_pcie_gen_cap;
@@ -6136,11 +6236,10 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
amdgpu_device_partner_bandwidth(adev, &platform_speed_cap,
&platform_link_width);
+ amdgpu_device_gpu_bandwidth(adev, &speed_cap, &link_width);
if (adev->pm.pcie_gen_mask == 0) {
/* asic caps */
- pdev = adev->pdev;
- speed_cap = pcie_get_speed_cap(pdev);
if (speed_cap == PCI_SPEED_UNKNOWN) {
adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
@@ -6196,51 +6295,103 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
}
}
if (adev->pm.pcie_mlw_mask == 0) {
+ /* asic caps */
+ if (link_width == PCIE_LNK_WIDTH_UNKNOWN) {
+ adev->pm.pcie_mlw_mask |= AMDGPU_DEFAULT_ASIC_PCIE_MLW_MASK;
+ } else {
+ switch (link_width) {
+ case PCIE_LNK_X32:
+ adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X32 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X16 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+ break;
+ case PCIE_LNK_X16:
+ adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X16 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+ break;
+ case PCIE_LNK_X12:
+ adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+ break;
+ case PCIE_LNK_X8:
+ adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+ break;
+ case PCIE_LNK_X4:
+ adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+ break;
+ case PCIE_LNK_X2:
+ adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+ break;
+ case PCIE_LNK_X1:
+ adev->pm.pcie_mlw_mask |= CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1;
+ break;
+ default:
+ break;
+ }
+ }
+ /* platform caps */
if (platform_link_width == PCIE_LNK_WIDTH_UNKNOWN) {
adev->pm.pcie_mlw_mask |= AMDGPU_DEFAULT_PCIE_MLW_MASK;
} else {
switch (platform_link_width) {
case PCIE_LNK_X32:
- adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+ adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
break;
case PCIE_LNK_X16:
- adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+ adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
break;
case PCIE_LNK_X12:
- adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+ adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
break;
case PCIE_LNK_X8:
- adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+ adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
break;
case PCIE_LNK_X4:
- adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+ adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
break;
case PCIE_LNK_X2:
- adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
- CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+ adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+ CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
break;
case PCIE_LNK_X1:
- adev->pm.pcie_mlw_mask = CAIL_PCIE_LINK_WIDTH_SUPPORT_X1;
+ adev->pm.pcie_mlw_mask |= CAIL_PCIE_LINK_WIDTH_SUPPORT_X1;
break;
default:
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 1040204ac8b9..949d74eff294 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -1,5 +1,5 @@
/*
- * Copyright 2018 Advanced Micro Devices, Inc.
+ * Copyright 2018-2024 Advanced Micro Devices, Inc. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -104,7 +104,9 @@
#include "smuio_v13_0_6.h"
#include "smuio_v14_0_2.h"
#include "vcn_v5_0_0.h"
+#include "vcn_v5_0_1.h"
#include "jpeg_v5_0_0.h"
+#include "jpeg_v5_0_1.h"
#include "amdgpu_vpe.h"
#if defined(CONFIG_DRM_AMD_ISP)
@@ -1340,7 +1342,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
*/
if (adev->vcn.num_vcn_inst <
AMDGPU_MAX_VCN_INSTANCES) {
- adev->vcn.vcn_config[adev->vcn.num_vcn_inst] =
+ adev->vcn.inst[adev->vcn.num_vcn_inst].vcn_config =
ip->revision & 0xc0;
adev->vcn.num_vcn_inst++;
adev->vcn.inst_mask |=
@@ -1705,7 +1707,7 @@ static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev)
* so this won't overflow.
*/
for (v = 0; v < adev->vcn.num_vcn_inst; v++) {
- adev->vcn.vcn_codec_disable_mask[v] =
+ adev->vcn.inst[v].vcn_codec_disable_mask =
le32_to_cpu(vcn_info->v1.instance_info[v].fuse_data.all_bits);
}
break;
@@ -1836,6 +1838,7 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev)
case IP_VERSION(9, 4, 2):
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
amdgpu_device_ip_block_add(adev, &vega10_common_ip_block);
break;
case IP_VERSION(10, 1, 10):
@@ -1890,6 +1893,7 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev)
case IP_VERSION(9, 4, 2):
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block);
break;
case IP_VERSION(10, 1, 10):
@@ -2013,6 +2017,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev)
case IP_VERSION(13, 0, 8):
case IP_VERSION(13, 0, 10):
case IP_VERSION(13, 0, 11):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
case IP_VERSION(14, 0, 0):
case IP_VERSION(14, 0, 1):
@@ -2184,6 +2189,7 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev)
break;
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
amdgpu_device_ip_block_add(adev, &gfx_v9_4_3_ip_block);
break;
case IP_VERSION(10, 1, 10):
@@ -2238,6 +2244,7 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev)
break;
case IP_VERSION(4, 4, 2):
case IP_VERSION(4, 4, 5):
+ case IP_VERSION(4, 4, 4):
amdgpu_device_ip_block_add(adev, &sdma_v4_4_2_ip_block);
break;
case IP_VERSION(5, 0, 0):
@@ -2361,6 +2368,10 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &vcn_v5_0_0_ip_block);
amdgpu_device_ip_block_add(adev, &jpeg_v5_0_0_ip_block);
break;
+ case IP_VERSION(5, 0, 1):
+ amdgpu_device_ip_block_add(adev, &vcn_v5_0_1_ip_block);
+ amdgpu_device_ip_block_add(adev, &jpeg_v5_0_1_ip_block);
+ break;
default:
dev_err(adev->dev,
"Failed to add vcn/jpeg ip block(UVD_HWIP:0x%x)\n",
@@ -2405,6 +2416,7 @@ static void amdgpu_discovery_init_soc_config(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
aqua_vanjaram_init_soc_config(adev);
break;
default:
@@ -2652,6 +2664,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
case IP_VERSION(9, 4, 2):
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
adev->family = AMDGPU_FAMILY_AI;
break;
case IP_VERSION(9, 1, 0):
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
index b119d27271c1..35c778426a7c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
@@ -33,6 +33,7 @@
#include "soc15_common.h"
#include "gc/gc_11_0_0_offset.h"
#include "gc/gc_11_0_0_sh_mask.h"
+#include "bif/bif_4_1_d.h"
#include <asm/div64.h>
#include <linux/pci.h>
@@ -1788,3 +1789,82 @@ int amdgpu_display_resume_helper(struct amdgpu_device *adev)
return 0;
}
+/* panic_bo is set in amdgpu_dm_plane_get_scanout_buffer() and only used in amdgpu_dm_set_pixel()
+ * they are called from the panic handler, and protected by the drm_panic spinlock.
+ */
+static struct amdgpu_bo *panic_abo;
+
+/* Use the indirect MMIO to write each pixel to the GPU VRAM,
+ * This is a simplified version of amdgpu_device_mm_access()
+ */
+static void amdgpu_display_set_pixel(struct drm_scanout_buffer *sb,
+ unsigned int x,
+ unsigned int y,
+ u32 color)
+{
+ struct amdgpu_res_cursor cursor;
+ unsigned long offset;
+ struct amdgpu_bo *abo = panic_abo;
+ struct amdgpu_device *adev = amdgpu_ttm_adev(abo->tbo.bdev);
+ uint32_t tmp;
+
+ offset = x * 4 + y * sb->pitch[0];
+ amdgpu_res_first(abo->tbo.resource, offset, 4, &cursor);
+
+ tmp = cursor.start >> 31;
+ WREG32_NO_KIQ(mmMM_INDEX, ((uint32_t) cursor.start) | 0x80000000);
+ if (tmp != 0xffffffff)
+ WREG32_NO_KIQ(mmMM_INDEX_HI, tmp);
+ WREG32_NO_KIQ(mmMM_DATA, color);
+}
+
+int amdgpu_display_get_scanout_buffer(struct drm_plane *plane,
+ struct drm_scanout_buffer *sb)
+{
+ struct amdgpu_bo *abo;
+ struct drm_framebuffer *fb = plane->state->fb;
+
+ if (!fb)
+ return -EINVAL;
+
+ DRM_DEBUG_KMS("Framebuffer %dx%d %p4cc\n", fb->width, fb->height, &fb->format->format);
+
+ abo = gem_to_amdgpu_bo(fb->obj[0]);
+ if (!abo)
+ return -EINVAL;
+
+ sb->width = fb->width;
+ sb->height = fb->height;
+ /* Use the generic linear format, because tiling will be disabled in panic_flush() */
+ sb->format = drm_format_info(fb->format->format);
+ if (!sb->format)
+ return -EINVAL;
+
+ sb->pitch[0] = fb->pitches[0];
+
+ if (abo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS) {
+ if (abo->tbo.resource->mem_type != TTM_PL_VRAM) {
+ drm_warn(plane->dev, "amdgpu panic, framebuffer not in VRAM\n");
+ return -EINVAL;
+ }
+ /* Only handle 32bits format, to simplify mmio access */
+ if (fb->format->cpp[0] != 4) {
+ drm_warn(plane->dev, "amdgpu panic, pixel format is not 32bits\n");
+ return -EINVAL;
+ }
+ sb->set_pixel = amdgpu_display_set_pixel;
+ panic_abo = abo;
+ return 0;
+ }
+ if (!abo->kmap.virtual &&
+ ttm_bo_kmap(&abo->tbo, 0, PFN_UP(abo->tbo.base.size), &abo->kmap)) {
+ drm_warn(plane->dev, "amdgpu bo map failed, panic won't be displayed\n");
+ return -ENOMEM;
+ }
+ if (abo->kmap.bo_kmap_type & TTM_BO_MAP_IOMEM_MASK)
+ iosys_map_set_vaddr_iomem(&sb->map[0], abo->kmap.virtual);
+ else
+ iosys_map_set_vaddr(&sb->map[0], abo->kmap.virtual);
+
+ return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
index 9d19940f73c8..dfa0d642ac16 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
@@ -23,6 +23,8 @@
#ifndef __AMDGPU_DISPLAY_H__
#define __AMDGPU_DISPLAY_H__
+#include <drm/drm_panic.h>
+
#define amdgpu_display_vblank_get_counter(adev, crtc) (adev)->mode_info.funcs->vblank_get_counter((adev), (crtc))
#define amdgpu_display_backlight_set_level(adev, e, l) (adev)->mode_info.funcs->backlight_set_level((e), (l))
#define amdgpu_display_backlight_get_level(adev, e) (adev)->mode_info.funcs->backlight_get_level((e))
@@ -49,4 +51,7 @@ amdgpu_lookup_format_info(u32 format, uint64_t modifier);
int amdgpu_display_suspend_helper(struct amdgpu_device *adev);
int amdgpu_display_resume_helper(struct amdgpu_device *adev);
+int amdgpu_display_get_scanout_buffer(struct drm_plane *plane,
+ struct drm_scanout_buffer *sb);
+
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 8e81a83d37d8..9f627caedc3f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -36,6 +36,7 @@
#include "amdgpu_gem.h"
#include "amdgpu_dma_buf.h"
#include "amdgpu_xgmi.h"
+#include "amdgpu_vm.h"
#include <drm/amdgpu_drm.h>
#include <drm/ttm/ttm_tt.h>
#include <linux/dma-buf.h>
@@ -60,6 +61,8 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
if (pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
attach->peer2peer = false;
+ amdgpu_vm_bo_update_shared(bo);
+
return 0;
}
@@ -345,7 +348,7 @@ amdgpu_dma_buf_move_notify(struct dma_buf_attachment *attach)
/* FIXME: This should be after the "if", but needs a fix to make sure
* DMABuf imports are initialized in the right VM list.
*/
- amdgpu_vm_bo_invalidate(adev, bo, false);
+ amdgpu_vm_bo_invalidate(bo, false);
if (!bo->tbo.resource || bo->tbo.resource->mem_type == TTM_PL_SYSTEM)
return;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 38686203bea6..c0ddbe7d6f0b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -23,7 +23,7 @@
*/
#include <drm/amdgpu_drm.h>
-#include <drm/drm_client_setup.h>
+#include <drm/clients/drm_client_setup.h>
#include <drm/drm_drv.h>
#include <drm/drm_fbdev_ttm.h>
#include <drm/drm_gem.h>
@@ -119,9 +119,11 @@
* - 3.57.0 - Compute tunneling on GFX10+
* - 3.58.0 - Add GFX12 DCC support
* - 3.59.0 - Cleared VRAM
+ * - 3.60.0 - Add AMDGPU_TILING_GFX12_DCC_WRITE_COMPRESS_DISABLE (Vulkan requirement)
+ * - 3.61.0 - Contains fix for RV/PCO compute queues
*/
#define KMS_DRIVER_MAJOR 3
-#define KMS_DRIVER_MINOR 59
+#define KMS_DRIVER_MINOR 61
#define KMS_DRIVER_PATCHLEVEL 0
/*
@@ -280,7 +282,7 @@ module_param_named(gartsize, amdgpu_gart_size, uint, 0600);
/**
* DOC: gttsize (int)
* Restrict the size of GTT domain (for userspace use) in MiB for testing.
- * The default is -1 (Use 1/2 RAM, minimum value is 3GB).
+ * The default is -1 (Use value specified by TTM).
*/
MODULE_PARM_DESC(gttsize, "Size of the GTT userspace domain in megabytes (-1 = auto)");
module_param_named(gttsize, amdgpu_gtt_size, int, 0600);
@@ -399,7 +401,7 @@ module_param_named(runpm, amdgpu_runtime_pm, int, 0444);
* the kernel log for the list of IPs on the asic. The default is 0xffffffff (enable all blocks on a device).
*/
MODULE_PARM_DESC(ip_block_mask, "IP Block Mask (all blocks enabled (default))");
-module_param_named(ip_block_mask, amdgpu_ip_block_mask, uint, 0444);
+module_param_named_unsafe(ip_block_mask, amdgpu_ip_block_mask, uint, 0444);
/**
* DOC: bapm (int)
@@ -457,7 +459,7 @@ module_param_named(vm_update_mode, amdgpu_vm_update_mode, int, 0444);
* Enable experimental hw support (1 = enable). The default is 0 (disabled).
*/
MODULE_PARM_DESC(exp_hw_support, "experimental hw support (1 = enable, 0 = disable (default))");
-module_param_named(exp_hw_support, amdgpu_exp_hw_support, int, 0444);
+module_param_named_unsafe(exp_hw_support, amdgpu_exp_hw_support, int, 0444);
/**
* DOC: dc (int)
@@ -568,14 +570,14 @@ module_param_named(compute_multipipe, amdgpu_compute_multipipe, int, 0444);
* Set to enable GPU recovery mechanism (1 = enable, 0 = disable). The default is -1 (auto, disabled except SRIOV).
*/
MODULE_PARM_DESC(gpu_recovery, "Enable GPU recovery mechanism, (1 = enable, 0 = disable, -1 = auto)");
-module_param_named(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
+module_param_named_unsafe(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
/**
* DOC: emu_mode (int)
* Set value 1 to enable emulation mode. This is only needed when running on an emulator. The default is 0 (disabled).
*/
MODULE_PARM_DESC(emu_mode, "Emulation mode, (1 = enable, 0 = disable)");
-module_param_named(emu_mode, amdgpu_emu_mode, int, 0444);
+module_param_named_unsafe(emu_mode, amdgpu_emu_mode, int, 0444);
/**
* DOC: ras_enable (int)
@@ -730,7 +732,7 @@ module_param_named(noretry, amdgpu_noretry, int, 0644);
*/
MODULE_PARM_DESC(force_asic_type,
"A non negative value used to specify the asic type for all supported GPUs");
-module_param_named(force_asic_type, amdgpu_force_asic_type, int, 0444);
+module_param_named_unsafe(force_asic_type, amdgpu_force_asic_type, int, 0444);
/**
* DOC: use_xgmi_p2p (int)
@@ -749,7 +751,7 @@ module_param_named(use_xgmi_p2p, amdgpu_use_xgmi_p2p, int, 0444);
* assigns queues to HQDs.
*/
int sched_policy = KFD_SCHED_POLICY_HWS;
-module_param(sched_policy, int, 0444);
+module_param_unsafe(sched_policy, int, 0444);
MODULE_PARM_DESC(sched_policy,
"Scheduling policy (0 = HWS (Default), 1 = HWS without over-subscription, 2 = Non-HWS (Used for debugging only)");
@@ -799,7 +801,7 @@ MODULE_PARM_DESC(send_sigterm,
* Setting 1 enables halt on hang.
*/
int halt_if_hws_hang;
-module_param(halt_if_hws_hang, int, 0644);
+module_param_unsafe(halt_if_hws_hang, int, 0644);
MODULE_PARM_DESC(halt_if_hws_hang, "Halt if HWS hang is detected (0 = off (default), 1 = on)");
/**
@@ -808,7 +810,7 @@ MODULE_PARM_DESC(halt_if_hws_hang, "Halt if HWS hang is detected (0 = off (defau
* check says. Default value: false (rely on MEC2 firmware version check).
*/
bool hws_gws_support;
-module_param(hws_gws_support, bool, 0444);
+module_param_unsafe(hws_gws_support, bool, 0444);
MODULE_PARM_DESC(hws_gws_support, "Assume MEC2 FW supports GWS barriers (false = rely on FW version check (Default), true = force supported)");
/**
@@ -841,7 +843,7 @@ MODULE_PARM_DESC(no_system_mem_limit, "disable system memory limit (false = defa
*/
int amdgpu_no_queue_eviction_on_vm_fault;
MODULE_PARM_DESC(no_queue_eviction_on_vm_fault, "No queue eviction on VM fault (0 = queue eviction, 1 = no queue eviction)");
-module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
+module_param_named_unsafe(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
#endif
/**
@@ -849,7 +851,7 @@ module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm
*/
int amdgpu_mtype_local;
MODULE_PARM_DESC(mtype_local, "MTYPE for local memory (0 = MTYPE_RW (default), 1 = MTYPE_NC, 2 = MTYPE_CC)");
-module_param_named(mtype_local, amdgpu_mtype_local, int, 0444);
+module_param_named_unsafe(mtype_local, amdgpu_mtype_local, int, 0444);
/**
* DOC: pcie_p2p (bool)
@@ -953,7 +955,7 @@ module_param_named(freesync_video, amdgpu_freesync_vid_mode, uint, 0444);
* GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco)
*/
MODULE_PARM_DESC(reset_method, "GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco/bamaco)");
-module_param_named(reset_method, amdgpu_reset_method, int, 0644);
+module_param_named_unsafe(reset_method, amdgpu_reset_method, int, 0644);
/**
* DOC: bad_page_threshold (int) Bad page threshold is specifies the
@@ -1049,7 +1051,7 @@ module_param_named(seamless, amdgpu_seamless, int, 0444);
* - 0x4: Disable GPU soft recovery, always do a full reset
*/
MODULE_PARM_DESC(debug_mask, "debug options for amdgpu, disabled by default");
-module_param_named(debug_mask, amdgpu_debug_mask, uint, 0444);
+module_param_named_unsafe(debug_mask, amdgpu_debug_mask, uint, 0444);
/**
* DOC: agp (int)
@@ -2552,9 +2554,7 @@ static int amdgpu_pmops_freeze(struct device *dev)
struct amdgpu_device *adev = drm_to_adev(drm_dev);
int r;
- adev->in_s4 = true;
r = amdgpu_device_suspend(drm_dev, true);
- adev->in_s4 = false;
if (r)
return r;
@@ -2566,8 +2566,13 @@ 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;
- return amdgpu_device_resume(drm_dev, true);
+ r = amdgpu_device_resume(drm_dev, true);
+ adev->in_s4 = false;
+
+ return r;
}
static int amdgpu_pmops_poweroff(struct device *dev)
@@ -2580,6 +2585,9 @@ 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);
}
@@ -2916,7 +2924,6 @@ static const struct drm_driver amdgpu_kms_driver = {
.name = DRIVER_NAME,
.desc = DRIVER_DESC,
- .date = DRIVER_DATE,
.major = KMS_DRIVER_MAJOR,
.minor = KMS_DRIVER_MINOR,
.patchlevel = KMS_DRIVER_PATCHLEVEL,
@@ -2940,7 +2947,6 @@ const struct drm_driver amdgpu_partition_driver = {
.name = DRIVER_NAME,
.desc = DRIVER_DESC,
- .date = DRIVER_DATE,
.major = KMS_DRIVER_MAJOR,
.minor = KMS_DRIVER_MINOR,
.patchlevel = KMS_DRIVER_PATCHLEVEL,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
index 5bc2cb661af7..2d86cc6f7f4d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
@@ -40,7 +40,6 @@
#define DRIVER_NAME "amdgpu"
#define DRIVER_DESC "AMD GPU"
-#define DRIVER_DATE "20150101"
extern const struct drm_driver amdgpu_partition_driver;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c
index df2cf5c33925..91d638098889 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c
@@ -60,7 +60,7 @@ void amdgpu_show_fdinfo(struct drm_printer *p, struct drm_file *file)
struct amdgpu_fpriv *fpriv = file->driver_priv;
struct amdgpu_vm *vm = &fpriv->vm;
- struct amdgpu_mem_stats stats[__AMDGPU_PL_LAST + 1] = { };
+ struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM];
ktime_t usage[AMDGPU_HW_IP_NUM];
const char *pl_name[] = {
[TTM_PL_VRAM] = "vram",
@@ -72,15 +72,8 @@ void amdgpu_show_fdinfo(struct drm_printer *p, struct drm_file *file)
[AMDGPU_PL_DOORBELL] = "doorbell",
};
unsigned int hw_ip, i;
- int ret;
-
- ret = amdgpu_bo_reserve(vm->root.bo, false);
- if (ret)
- return;
-
- amdgpu_vm_get_memory(vm, stats, ARRAY_SIZE(stats));
- amdgpu_bo_unreserve(vm->root.bo);
+ amdgpu_vm_get_memory(vm, stats);
amdgpu_ctx_mgr_usage(&fpriv->ctx_mgr, usage);
/*
@@ -114,9 +107,11 @@ void amdgpu_show_fdinfo(struct drm_printer *p, struct drm_file *file)
drm_printf(p, "amd-evicted-vram:\t%llu KiB\n",
stats[TTM_PL_VRAM].evicted/1024UL);
drm_printf(p, "amd-requested-vram:\t%llu KiB\n",
- stats[TTM_PL_VRAM].requested/1024UL);
+ (stats[TTM_PL_VRAM].drm.shared +
+ stats[TTM_PL_VRAM].drm.private) / 1024UL);
drm_printf(p, "amd-requested-gtt:\t%llu KiB\n",
- stats[TTM_PL_TT].requested/1024UL);
+ (stats[TTM_PL_TT].drm.shared +
+ stats[TTM_PL_TT].drm.private) / 1024UL);
for (hw_ip = 0; hw_ip < AMDGPU_HW_IP_NUM; ++hw_ip) {
if (!usage[hw_ip])
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
index ceb5163480f4..09c9194d5bd5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
@@ -384,7 +384,7 @@ int amdgpu_fru_sysfs_init(struct amdgpu_device *adev)
void amdgpu_fru_sysfs_fini(struct amdgpu_device *adev)
{
- if (!is_fru_eeprom_supported(adev, NULL) || !adev->fru_info)
+ if (!adev->fru_info)
return;
sysfs_remove_files(&adev->dev->kobj, amdgpu_fru_attributes);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h
index bc58dca18035..98f3196599ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h
@@ -32,7 +32,7 @@ struct amdgpu_fru_info {
char product_name[AMDGPU_PRODUCT_NAME_LEN];
char serial[20];
char manufacturer_name[32];
- char fru_id[32];
+ char fru_id[50];
};
int amdgpu_fru_get_product_info(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 1a5df8b94661..69429df09477 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -42,6 +42,7 @@
#include "amdgpu_dma_buf.h"
#include "amdgpu_hmm.h"
#include "amdgpu_xgmi.h"
+#include "amdgpu_vm.h"
static vm_fault_t amdgpu_gem_fault(struct vm_fault *vmf)
{
@@ -87,10 +88,8 @@ static void amdgpu_gem_object_free(struct drm_gem_object *gobj)
{
struct amdgpu_bo *aobj = gem_to_amdgpu_bo(gobj);
- if (aobj) {
- amdgpu_hmm_unregister(aobj);
- ttm_bo_put(&aobj->tbo);
- }
+ amdgpu_hmm_unregister(aobj);
+ ttm_bo_put(&aobj->tbo);
}
int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
@@ -179,6 +178,7 @@ static int amdgpu_gem_object_open(struct drm_gem_object *obj,
if (r)
return r;
+ amdgpu_vm_bo_update_shared(abo);
bo_va = amdgpu_vm_bo_find(vm, abo);
if (!bo_va)
bo_va = amdgpu_vm_bo_add(adev, vm, abo);
@@ -252,6 +252,7 @@ static void amdgpu_gem_object_close(struct drm_gem_object *obj,
goto out_unlock;
amdgpu_vm_bo_del(adev, bo_va);
+ amdgpu_vm_bo_update_shared(bo);
if (!amdgpu_vm_ready(vm))
goto out_unlock;
@@ -839,7 +840,6 @@ error:
int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
struct drm_file *filp)
{
- struct amdgpu_device *adev = drm_to_adev(dev);
struct drm_amdgpu_gem_op *args = data;
struct drm_gem_object *gobj;
struct amdgpu_vm_bo_base *base;
@@ -899,7 +899,7 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
robj->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT;
if (robj->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
- amdgpu_vm_bo_invalidate(adev, robj, true);
+ amdgpu_vm_bo_invalidate(robj, true);
amdgpu_bo_unreserve(robj);
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 1d155463d044..c1f35ded684e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -515,7 +515,7 @@ int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id)
if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
return -EINVAL;
- if (!kiq_ring->sched.ready || adev->job_hang || amdgpu_in_reset(adev))
+ if (!kiq_ring->sched.ready || amdgpu_in_reset(adev))
return 0;
spin_lock(&kiq->ring_lock);
@@ -567,7 +567,7 @@ int amdgpu_gfx_disable_kgq(struct amdgpu_device *adev, int xcc_id)
if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
return -EINVAL;
- if (!adev->gfx.kiq[0].ring.sched.ready || adev->job_hang)
+ if (!adev->gfx.kiq[0].ring.sched.ready || amdgpu_in_reset(adev))
return 0;
if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) {
@@ -806,7 +806,7 @@ void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
/* If going to s2idle, no need to wait */
if (adev->in_s0ix) {
if (!amdgpu_dpm_set_powergating_by_smu(adev,
- AMD_IP_BLOCK_TYPE_GFX, true))
+ AMD_IP_BLOCK_TYPE_GFX, true, 0))
adev->gfx.gfx_off_state = true;
} else {
schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
@@ -818,7 +818,7 @@ void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
cancel_delayed_work_sync(&adev->gfx.gfx_off_delay_work);
if (adev->gfx.gfx_off_state &&
- !amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, false)) {
+ !amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, false, 0)) {
adev->gfx.gfx_off_state = false;
if (adev->gfx.funcs->init_spm_golden) {
@@ -1484,6 +1484,24 @@ static int amdgpu_gfx_run_cleaner_shader(struct amdgpu_device *adev, int xcp_id)
return 0;
}
+/**
+ * amdgpu_gfx_set_run_cleaner_shader - Execute the AMDGPU GFX Cleaner Shader
+ * @dev: The device structure
+ * @attr: The device attribute structure
+ * @buf: The buffer containing the input data
+ * @count: The size of the input data
+ *
+ * Provides the sysfs interface to manually run a cleaner shader, which is
+ * used to clear the GPU state between different tasks. Writing a value to the
+ * 'run_cleaner_shader' sysfs file triggers the cleaner shader execution.
+ * The value written corresponds to the partition index on multi-partition
+ * devices. On single-partition devices, the value should be '0'.
+ *
+ * The cleaner shader clears the Local Data Store (LDS) and General Purpose
+ * Registers (GPRs) to ensure data isolation between GPU workloads.
+ *
+ * Return: The number of bytes written to the sysfs file.
+ */
static ssize_t amdgpu_gfx_set_run_cleaner_shader(struct device *dev,
struct device_attribute *attr,
const char *buf,
@@ -1532,6 +1550,19 @@ static ssize_t amdgpu_gfx_set_run_cleaner_shader(struct device *dev,
return count;
}
+/**
+ * amdgpu_gfx_get_enforce_isolation - Query AMDGPU GFX Enforce Isolation Settings
+ * @dev: The device structure
+ * @attr: The device attribute structure
+ * @buf: The buffer to store the output data
+ *
+ * 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.
+ *
+ * Return: The number of bytes read from the sysfs file.
+ */
static ssize_t amdgpu_gfx_get_enforce_isolation(struct device *dev,
struct device_attribute *attr,
char *buf)
@@ -1555,6 +1586,20 @@ static ssize_t amdgpu_gfx_get_enforce_isolation(struct device *dev,
return size;
}
+/**
+ * amdgpu_gfx_set_enforce_isolation - Control AMDGPU GFX Enforce Isolation
+ * @dev: The device structure
+ * @attr: The device attribute structure
+ * @buf: The buffer containing the input data
+ * @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.
+ *
+ * Return: The number of bytes written to the sysfs file.
+ */
static ssize_t amdgpu_gfx_set_enforce_isolation(struct device *dev,
struct device_attribute *attr,
const char *buf, size_t count)
@@ -1593,22 +1638,19 @@ static ssize_t amdgpu_gfx_set_enforce_isolation(struct device *dev,
}
mutex_lock(&adev->enforce_isolation_mutex);
-
for (i = 0; i < num_partitions; i++) {
- if (adev->enforce_isolation[i] && !partition_values[i]) {
+ if (adev->enforce_isolation[i] && !partition_values[i])
/* Going from enabled to disabled */
amdgpu_vmid_free_reserved(adev, AMDGPU_GFXHUB(i));
- amdgpu_mes_set_enforce_isolation(adev, i, false);
- } else if (!adev->enforce_isolation[i] && partition_values[i]) {
+ else if (!adev->enforce_isolation[i] && partition_values[i])
/* Going from disabled to enabled */
amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(i));
- amdgpu_mes_set_enforce_isolation(adev, i, true);
- }
adev->enforce_isolation[i] = partition_values[i];
}
-
mutex_unlock(&adev->enforce_isolation_mutex);
+ amdgpu_mes_update_enforce_isolation(adev);
+
return count;
}
@@ -1940,6 +1982,17 @@ void amdgpu_gfx_enforce_isolation_handler(struct work_struct *work)
mutex_unlock(&adev->enforce_isolation_mutex);
}
+/**
+ * amdgpu_gfx_enforce_isolation_wait_for_kfd - Manage KFD wait period for process isolation
+ * @adev: amdgpu_device pointer
+ * @idx: Index of the GPU partition
+ *
+ * When kernel submissions come in, the jobs are given a time slice and once
+ * that time slice is up, if there are KFD user queues active, kernel
+ * submissions are blocked until KFD has had its time slice. Once the KFD time
+ * slice is up, KFD user queues are preempted and kernel submissions are
+ * unblocked and allowed to run again.
+ */
static void
amdgpu_gfx_enforce_isolation_wait_for_kfd(struct amdgpu_device *adev,
u32 idx)
@@ -1985,6 +2038,15 @@ amdgpu_gfx_enforce_isolation_wait_for_kfd(struct amdgpu_device *adev,
msleep(GFX_SLICE_PERIOD_MS);
}
+/**
+ * amdgpu_gfx_enforce_isolation_ring_begin_use - Begin use of a ring with enforced isolation
+ * @ring: Pointer to the amdgpu_ring structure
+ *
+ * Ring begin_use helper implementation for gfx which serializes access to the
+ * gfx IP between kernel submission IOCTLs and KFD user queues when isolation
+ * enforcement is enabled. The kernel submission IOCTLs and KFD user queues
+ * each get a time slice when both are active.
+ */
void amdgpu_gfx_enforce_isolation_ring_begin_use(struct amdgpu_ring *ring)
{
struct amdgpu_device *adev = ring->adev;
@@ -2016,6 +2078,15 @@ void amdgpu_gfx_enforce_isolation_ring_begin_use(struct amdgpu_ring *ring)
amdgpu_gfx_kfd_sch_ctrl(adev, idx, false);
}
+/**
+ * amdgpu_gfx_enforce_isolation_ring_end_use - End use of a ring with enforced isolation
+ * @ring: Pointer to the amdgpu_ring structure
+ *
+ * Ring end_use helper implementation for gfx which serializes access to the
+ * gfx IP between kernel submission IOCTLs and KFD user queues when isolation
+ * enforcement is enabled. The kernel submission IOCTLs and KFD user queues
+ * each get a time slice when both are active.
+ */
void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring)
{
struct amdgpu_device *adev = ring->adev;
@@ -2058,7 +2129,7 @@ static int amdgpu_debugfs_gfx_sched_mask_set(void *data, u64 val)
if (!adev)
return -ENODEV;
- mask = (1 << adev->gfx.num_gfx_rings) - 1;
+ mask = (1ULL << adev->gfx.num_gfx_rings) - 1;
if ((val & mask) == 0)
return -EINVAL;
@@ -2086,7 +2157,7 @@ static int amdgpu_debugfs_gfx_sched_mask_get(void *data, u64 *val)
for (i = 0; i < adev->gfx.num_gfx_rings; ++i) {
ring = &adev->gfx.gfx_ring[i];
if (ring->sched.ready)
- mask |= 1 << i;
+ mask |= 1ULL << i;
}
*val = mask;
@@ -2128,7 +2199,7 @@ static int amdgpu_debugfs_compute_sched_mask_set(void *data, u64 val)
if (!adev)
return -ENODEV;
- mask = (1 << adev->gfx.num_compute_rings) - 1;
+ mask = (1ULL << adev->gfx.num_compute_rings) - 1;
if ((val & mask) == 0)
return -EINVAL;
@@ -2157,7 +2228,7 @@ static int amdgpu_debugfs_compute_sched_mask_get(void *data, u64 *val)
for (i = 0; i < adev->gfx.num_compute_rings; ++i) {
ring = &adev->gfx.compute_ring[i];
if (ring->sched.ready)
- mask |= 1 << i;
+ mask |= 1ULL << i;
}
*val = mask;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 071f187f5e28..2ea98ec60220 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -89,16 +89,14 @@ int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
/**
* amdgpu_ib_free - free an IB (Indirect Buffer)
*
- * @adev: amdgpu_device pointer
* @ib: IB object to free
* @f: the fence SA bo need wait on for the ib alloation
*
* Free an IB (all asics).
*/
-void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
- struct dma_fence *f)
+void amdgpu_ib_free(struct amdgpu_ib *ib, struct dma_fence *f)
{
- amdgpu_sa_bo_free(adev, &ib->sa_bo, f);
+ amdgpu_sa_bo_free(&ib->sa_bo, f);
}
/**
@@ -299,7 +297,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs,
amdgpu_ring_patch_cond_exec(ring, cond_exec);
ring->current_ctx = fence_ctx;
- if (vm && ring->funcs->emit_switch_buffer)
+ if (job && ring->funcs->emit_switch_buffer)
amdgpu_ring_emit_switch_buffer(ring);
if (ring->funcs->emit_wave_limit &&
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
index f3b0aaf3ebc6..901f8b12c672 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
@@ -298,3 +298,9 @@ uint64_t amdgpu_ih_decode_iv_ts_helper(struct amdgpu_ih_ring *ih, u32 rptr,
dw2 = le32_to_cpu(ih->ring[ring_index + 2]);
return dw1 | ((u64)(dw2 & 0xffff) << 32);
}
+
+const char *amdgpu_ih_ring_name(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
+{
+ return ih == &adev->irq.ih ? "ih" : ih == &adev->irq.ih_soft ? "sw ih" :
+ ih == &adev->irq.ih1 ? "ih1" : ih == &adev->irq.ih2 ? "ih2" : "unknown";
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
index 508f02eb0cf8..7d4395a5d8ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
@@ -110,4 +110,5 @@ void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
struct amdgpu_iv_entry *entry);
uint64_t amdgpu_ih_decode_iv_ts_helper(struct amdgpu_ih_ring *ih, u32 rptr,
signed int offset);
+const char *amdgpu_ih_ring_name(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c
index 263ce1811cc8..732744488b03 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c
@@ -77,7 +77,8 @@ static int isp_load_fw_by_psp(struct amdgpu_device *adev)
sizeof(ucode_prefix));
/* read isp fw */
- r = amdgpu_ucode_request(adev, &adev->isp.fw, "amdgpu/%s.bin", ucode_prefix);
+ r = amdgpu_ucode_request(adev, &adev->isp.fw, AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s.bin", ucode_prefix);
if (r) {
amdgpu_ucode_release(&adev->isp.fw);
return r;
@@ -128,13 +129,13 @@ static bool isp_is_idle(void *handle)
return true;
}
-static int isp_set_clockgating_state(void *handle,
+static int isp_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int isp_set_powergating_state(void *handle,
+static int isp_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index a21c510c408e..100f04475943 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -102,8 +102,6 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job)
return DRM_GPU_SCHED_STAT_ENODEV;
}
- adev->job_hang = true;
-
/*
* Do the coredump immediately after a job timeout to get a very
* close dump/snapshot/representation of GPU's current error status
@@ -181,7 +179,6 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job)
}
exit:
- adev->job_hang = false;
drm_dev_exit(idx);
return DRM_GPU_SCHED_STAT_NOMINAL;
}
@@ -197,11 +194,6 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, struct amdgpu_vm *vm,
if (!*job)
return -ENOMEM;
- /*
- * Initialize the scheduler to at least some ring so that we always
- * have a pointer to adev.
- */
- (*job)->base.sched = &adev->rings[0]->sched;
(*job)->vm = vm;
amdgpu_sync_create(&(*job)->explicit_sync);
@@ -267,7 +259,7 @@ void amdgpu_job_free_resources(struct amdgpu_job *job)
f = NULL;
for (i = 0; i < job->num_ibs; ++i)
- amdgpu_ib_free(NULL, &job->ibs[i], f);
+ amdgpu_ib_free(&job->ibs[i], f);
}
static void amdgpu_job_free_cb(struct drm_sched_job *s_job)
@@ -366,6 +358,13 @@ amdgpu_job_prepare_job(struct drm_sched_job *sched_job,
dev_err(ring->adev->dev, "Error getting VM ID (%d)\n", r);
goto error;
}
+ /*
+ * The VM structure might be released after the VMID is
+ * assigned, we had multiple problems with people trying to use
+ * the VM pointer so better set it to NULL.
+ */
+ if (!fence)
+ job->vm = NULL;
}
return fence;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
index 3eb4a4653fce..d9cb343a8708 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
@@ -27,7 +27,8 @@
#include "amdgpu_ras.h"
#define AMDGPU_MAX_JPEG_INSTANCES 4
-#define AMDGPU_MAX_JPEG_RINGS 8
+#define AMDGPU_MAX_JPEG_RINGS 10
+#define AMDGPU_MAX_JPEG_RINGS_4_0_3 8
#define AMDGPU_JPEG_HARVEST_JPEG0 (1 << 0)
#define AMDGPU_JPEG_HARVEST_JPEG1 (1 << 1)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 016a6f6c4267..98528ee94c15 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -846,7 +846,7 @@ out:
case AMDGPU_INFO_DEV_INFO: {
struct drm_amdgpu_info_device *dev_info;
uint64_t vm_size;
- uint32_t pcie_gen_mask;
+ uint32_t pcie_gen_mask, pcie_width_mask;
dev_info = kzalloc(sizeof(*dev_info), GFP_KERNEL);
if (!dev_info)
@@ -934,15 +934,18 @@ out:
dev_info->tcc_disabled_mask = adev->gfx.config.tcc_disabled_mask;
/* Combine the chip gen mask with the platform (CPU/mobo) mask. */
- pcie_gen_mask = adev->pm.pcie_gen_mask & (adev->pm.pcie_gen_mask >> 16);
+ pcie_gen_mask = adev->pm.pcie_gen_mask &
+ (adev->pm.pcie_gen_mask >> CAIL_PCIE_LINK_SPEED_SUPPORT_SHIFT);
+ pcie_width_mask = adev->pm.pcie_mlw_mask &
+ (adev->pm.pcie_mlw_mask >> CAIL_PCIE_LINK_WIDTH_SUPPORT_SHIFT);
dev_info->pcie_gen = fls(pcie_gen_mask);
dev_info->pcie_num_lanes =
- adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 ? 32 :
- adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 ? 16 :
- adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 ? 12 :
- adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 ? 8 :
- adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 ? 4 :
- adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 ? 2 : 1;
+ pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X32 ? 32 :
+ pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X16 ? 16 :
+ pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 ? 12 :
+ pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 ? 8 :
+ pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 ? 4 :
+ pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 ? 2 : 1;
dev_info->tcp_cache_size = adev->gfx.config.gc_tcp_l1_size;
dev_info->num_sqc_per_wgp = adev->gfx.config.gc_num_sqc_per_wgp;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
index 59ec20b07a6a..709c11cbeabd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
@@ -1610,10 +1610,12 @@ int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe)
pipe == AMDGPU_MES_SCHED_PIPE ? "" : "1");
}
- r = amdgpu_ucode_request(adev, &adev->mes.fw[pipe], "%s", fw_name);
+ r = amdgpu_ucode_request(adev, &adev->mes.fw[pipe], AMDGPU_UCODE_REQUIRED,
+ "%s", fw_name);
if (r && need_retry && pipe == AMDGPU_MES_SCHED_PIPE) {
dev_info(adev->dev, "try to fall back to %s_mes.bin\n", ucode_prefix);
r = amdgpu_ucode_request(adev, &adev->mes.fw[pipe],
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mes.bin", ucode_prefix);
}
@@ -1679,7 +1681,8 @@ bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev)
}
/* Fix me -- node_id is used to identify the correct MES instances in the future */
-int amdgpu_mes_set_enforce_isolation(struct amdgpu_device *adev, uint32_t node_id, bool enable)
+static int amdgpu_mes_set_enforce_isolation(struct amdgpu_device *adev,
+ uint32_t node_id, bool enable)
{
struct mes_misc_op_input op_input = {0};
int r;
@@ -1701,6 +1704,23 @@ error:
return r;
}
+int amdgpu_mes_update_enforce_isolation(struct amdgpu_device *adev)
+{
+ int i, r = 0;
+
+ 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])
+ r |= amdgpu_mes_set_enforce_isolation(adev, i, true);
+ else
+ r |= amdgpu_mes_set_enforce_isolation(adev, i, false);
+ }
+ mutex_unlock(&adev->enforce_isolation_mutex);
+ }
+ return r;
+}
+
#if defined(CONFIG_DEBUG_FS)
static int amdgpu_debugfs_mes_event_log_show(struct seq_file *m, void *unused)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
index c6f93cbd6739..e98ea7ede1ba 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
@@ -40,7 +40,7 @@
#define AMDGPU_MES_VERSION_MASK 0x00000fff
#define AMDGPU_MES_API_VERSION_MASK 0x00fff000
#define AMDGPU_MES_FEAT_VERSION_MASK 0xff000000
-#define AMDGPU_MES_MSCRATCH_SIZE 0x8000
+#define AMDGPU_MES_MSCRATCH_SIZE 0x40000
enum amdgpu_mes_priority_level {
AMDGPU_MES_PRIORITY_LEVEL_LOW = 0,
@@ -534,6 +534,6 @@ static inline void amdgpu_mes_unlock(struct amdgpu_mes *mes)
bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev);
-int amdgpu_mes_set_enforce_isolation(struct amdgpu_device *adev, uint32_t node_id, bool enable);
+int amdgpu_mes_update_enforce_isolation(struct amdgpu_device *adev);
#endif /* __AMDGPU_MES_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 6852d50caa89..96f4b8904e9a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -41,6 +41,7 @@
#include "amdgpu_amdkfd.h"
#include "amdgpu_vram_mgr.h"
#include "amdgpu_vm.h"
+#include "amdgpu_dma_buf.h"
/**
* DOC: amdgpu_object
@@ -324,6 +325,9 @@ error_free:
*
* Allocates and pins a BO for kernel internal use.
*
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to create and access the kernel BO.
+ *
* Note: For bo_ptr new BO is only created if bo_ptr points to NULL.
*
* Returns:
@@ -347,6 +351,76 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
return 0;
}
+EXPORT_SYMBOL(amdgpu_bo_create_kernel);
+
+/**
+ * amdgpu_bo_create_isp_user - create user BO for isp
+ *
+ * @adev: amdgpu device object
+ * @dma_buf: DMABUF handle for isp buffer
+ * @domain: where to place it
+ * @bo: used to initialize BOs in structures
+ * @gpu_addr: GPU addr of the pinned BO
+ *
+ * Imports isp DMABUF to allocate and pin a user BO for isp internal use. It does
+ * GART alloc to generate gpu_addr for BO to make it accessible through the
+ * GART aperture for ISP HW.
+ *
+ * This function is exported to allow the V4L2 isp device external to drm device
+ * to create and access the isp user BO.
+ *
+ * Returns:
+ * 0 on success, negative error code otherwise.
+ */
+int amdgpu_bo_create_isp_user(struct amdgpu_device *adev,
+ struct dma_buf *dma_buf, u32 domain, struct amdgpu_bo **bo,
+ u64 *gpu_addr)
+
+{
+ struct drm_gem_object *gem_obj;
+ int r;
+
+ gem_obj = amdgpu_gem_prime_import(&adev->ddev, dma_buf);
+ *bo = gem_to_amdgpu_bo(gem_obj);
+ if (!(*bo)) {
+ dev_err(adev->dev, "failed to get valid isp user bo\n");
+ return -EINVAL;
+ }
+
+ r = amdgpu_bo_reserve(*bo, false);
+ if (r) {
+ dev_err(adev->dev, "(%d) failed to reserve isp user bo\n", r);
+ return r;
+ }
+
+ r = amdgpu_bo_pin(*bo, domain);
+ if (r) {
+ dev_err(adev->dev, "(%d) isp user bo pin failed\n", r);
+ goto error_unreserve;
+ }
+
+ r = amdgpu_ttm_alloc_gart(&(*bo)->tbo);
+ if (r) {
+ dev_err(adev->dev, "%p bind failed\n", *bo);
+ goto error_unpin;
+ }
+
+ if (!WARN_ON(!gpu_addr))
+ *gpu_addr = amdgpu_bo_gpu_offset(*bo);
+
+ amdgpu_bo_unreserve(*bo);
+
+ return 0;
+
+error_unpin:
+ amdgpu_bo_unpin(*bo);
+error_unreserve:
+ amdgpu_bo_unreserve(*bo);
+ amdgpu_bo_unref(bo);
+
+ return r;
+}
+EXPORT_SYMBOL(amdgpu_bo_create_isp_user);
/**
* amdgpu_bo_create_kernel_at - create BO for kernel use at specific location
@@ -423,6 +497,9 @@ error:
* @cpu_addr: pointer to where the BO's CPU memory space address was stored
*
* unmaps and unpin a BO for kernel internal use.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to free the kernel BO.
*/
void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
void **cpu_addr)
@@ -447,6 +524,30 @@ void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
if (cpu_addr)
*cpu_addr = NULL;
}
+EXPORT_SYMBOL(amdgpu_bo_free_kernel);
+
+/**
+ * amdgpu_bo_free_isp_user - free BO for isp use
+ *
+ * @bo: amdgpu isp user BO to free
+ *
+ * unpin and unref BO for isp internal use.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to free the isp user BO.
+ */
+void amdgpu_bo_free_isp_user(struct amdgpu_bo *bo)
+{
+ if (bo == NULL)
+ return;
+
+ if (amdgpu_bo_reserve(bo, true) == 0) {
+ amdgpu_bo_unpin(bo);
+ amdgpu_bo_unreserve(bo);
+ }
+ amdgpu_bo_unref(&bo);
+}
+EXPORT_SYMBOL(amdgpu_bo_free_isp_user);
/* Validate bo size is bit bigger than the request domain */
static bool amdgpu_bo_validate_size(struct amdgpu_device *adev,
@@ -1150,7 +1251,6 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
bool evict,
struct ttm_resource *new_mem)
{
- struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
struct ttm_resource *old_mem = bo->resource;
struct amdgpu_bo *abo;
@@ -1158,7 +1258,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
return;
abo = ttm_to_amdgpu_bo(bo);
- amdgpu_vm_bo_invalidate(adev, abo, evict);
+ amdgpu_vm_bo_move(abo, new_mem, evict);
amdgpu_bo_kunmap(abo);
@@ -1171,75 +1271,6 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
old_mem ? old_mem->mem_type : -1);
}
-void amdgpu_bo_get_memory(struct amdgpu_bo *bo,
- struct amdgpu_mem_stats *stats,
- unsigned int sz)
-{
- const unsigned int domain_to_pl[] = {
- [ilog2(AMDGPU_GEM_DOMAIN_CPU)] = TTM_PL_SYSTEM,
- [ilog2(AMDGPU_GEM_DOMAIN_GTT)] = TTM_PL_TT,
- [ilog2(AMDGPU_GEM_DOMAIN_VRAM)] = TTM_PL_VRAM,
- [ilog2(AMDGPU_GEM_DOMAIN_GDS)] = AMDGPU_PL_GDS,
- [ilog2(AMDGPU_GEM_DOMAIN_GWS)] = AMDGPU_PL_GWS,
- [ilog2(AMDGPU_GEM_DOMAIN_OA)] = AMDGPU_PL_OA,
- [ilog2(AMDGPU_GEM_DOMAIN_DOORBELL)] = AMDGPU_PL_DOORBELL,
- };
- struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
- struct ttm_resource *res = bo->tbo.resource;
- struct drm_gem_object *obj = &bo->tbo.base;
- uint64_t size = amdgpu_bo_size(bo);
- unsigned int type;
-
- if (!res) {
- /*
- * If no backing store use one of the preferred domain for basic
- * stats. We take the MSB since that should give a reasonable
- * view.
- */
- BUILD_BUG_ON(TTM_PL_VRAM < TTM_PL_TT ||
- TTM_PL_VRAM < TTM_PL_SYSTEM);
- type = fls(bo->preferred_domains & AMDGPU_GEM_DOMAIN_MASK);
- if (!type)
- return;
- type--;
- if (drm_WARN_ON_ONCE(&adev->ddev,
- type >= ARRAY_SIZE(domain_to_pl)))
- return;
- type = domain_to_pl[type];
- } else {
- type = res->mem_type;
- }
-
- if (drm_WARN_ON_ONCE(&adev->ddev, type >= sz))
- return;
-
- /* DRM stats common fields: */
-
- if (drm_gem_object_is_shared_for_memory_stats(obj))
- stats[type].drm.shared += size;
- else
- stats[type].drm.private += size;
-
- if (res) {
- stats[type].drm.resident += size;
-
- if (!dma_resv_test_signaled(obj->resv, DMA_RESV_USAGE_BOOKKEEP))
- stats[type].drm.active += size;
- else if (bo->flags & AMDGPU_GEM_CREATE_DISCARDABLE)
- stats[type].drm.purgeable += size;
- }
-
- /* amdgpu specific stats: */
-
- if (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) {
- stats[TTM_PL_VRAM].requested += size;
- if (type != TTM_PL_VRAM)
- stats[TTM_PL_VRAM].evicted += size;
- } else if (bo->preferred_domains & AMDGPU_GEM_DOMAIN_GTT) {
- stats[TTM_PL_TT].requested += size;
- }
-}
-
/**
* amdgpu_bo_release_notify - notification about a BO being released
* @bo: pointer to a buffer object
@@ -1455,6 +1486,45 @@ u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo)
}
/**
+ * amdgpu_bo_mem_stats_placement - bo placement for memory accounting
+ * @bo: the buffer object we should look at
+ *
+ * BO can have multiple preferred placements, to avoid double counting we want
+ * to file it under a single placement for memory stats.
+ * Luckily, if we take the highest set bit in preferred_domains the result is
+ * quite sensible.
+ *
+ * Returns:
+ * Which of the placements should the BO be accounted under.
+ */
+uint32_t amdgpu_bo_mem_stats_placement(struct amdgpu_bo *bo)
+{
+ uint32_t domain = bo->preferred_domains & AMDGPU_GEM_DOMAIN_MASK;
+
+ if (!domain)
+ return TTM_PL_SYSTEM;
+
+ switch (rounddown_pow_of_two(domain)) {
+ case AMDGPU_GEM_DOMAIN_CPU:
+ return TTM_PL_SYSTEM;
+ case AMDGPU_GEM_DOMAIN_GTT:
+ return TTM_PL_TT;
+ case AMDGPU_GEM_DOMAIN_VRAM:
+ return TTM_PL_VRAM;
+ case AMDGPU_GEM_DOMAIN_GDS:
+ return AMDGPU_PL_GDS;
+ case AMDGPU_GEM_DOMAIN_GWS:
+ return AMDGPU_PL_GWS;
+ case AMDGPU_GEM_DOMAIN_OA:
+ return AMDGPU_PL_OA;
+ case AMDGPU_GEM_DOMAIN_DOORBELL:
+ return AMDGPU_PL_DOORBELL;
+ default:
+ return TTM_PL_SYSTEM;
+ }
+}
+
+/**
* amdgpu_bo_get_preferred_domain - get preferred domain
* @adev: amdgpu device object
* @domain: allowed :ref:`memory domains <amdgpu_memory_domains>`
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index be6769852ece..375448627f7b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -260,6 +260,10 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
unsigned long size, int align,
u32 domain, struct amdgpu_bo **bo_ptr,
u64 *gpu_addr, void **cpu_addr);
+int amdgpu_bo_create_isp_user(struct amdgpu_device *adev,
+ struct dma_buf *dbuf, u32 domain,
+ struct amdgpu_bo **bo,
+ u64 *gpu_addr);
int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev,
uint64_t offset, uint64_t size,
struct amdgpu_bo **bo_ptr, void **cpu_addr);
@@ -271,6 +275,7 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev,
struct amdgpu_bo_vm **ubo_ptr);
void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
void **cpu_addr);
+void amdgpu_bo_free_isp_user(struct amdgpu_bo *bo);
int amdgpu_bo_kmap(struct amdgpu_bo *bo, void **ptr);
void *amdgpu_bo_kptr(struct amdgpu_bo *bo);
void amdgpu_bo_kunmap(struct amdgpu_bo *bo);
@@ -300,9 +305,7 @@ int amdgpu_bo_sync_wait_resv(struct amdgpu_device *adev, struct dma_resv *resv,
int amdgpu_bo_sync_wait(struct amdgpu_bo *bo, void *owner, bool intr);
u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo);
u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo);
-void amdgpu_bo_get_memory(struct amdgpu_bo *bo,
- struct amdgpu_mem_stats *stats,
- unsigned int size);
+uint32_t amdgpu_bo_mem_stats_placement(struct amdgpu_bo *bo);
uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev,
uint32_t domain);
@@ -337,8 +340,7 @@ int amdgpu_sa_bo_manager_start(struct amdgpu_device *adev,
int amdgpu_sa_bo_new(struct amdgpu_sa_manager *sa_manager,
struct drm_suballoc **sa_bo,
unsigned int size);
-void amdgpu_sa_bo_free(struct amdgpu_device *adev,
- struct drm_suballoc **sa_bo,
+void amdgpu_sa_bo_free(struct drm_suballoc **sa_bo,
struct dma_fence *fence);
#if defined(CONFIG_DEBUG_FS)
void amdgpu_sa_bo_dump_debug_info(struct amdgpu_sa_manager *sa_manager,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 448f9e742983..e5fc80ed06ea 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -208,6 +208,7 @@ static int psp_early_init(struct amdgpu_ip_block *ip_block)
psp->boot_time_tmr = false;
fallthrough;
case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
psp_v13_0_set_psp_funcs(psp);
psp->autoload_supported = false;
@@ -359,6 +360,7 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev,
int i;
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))
return false;
@@ -870,6 +872,7 @@ static bool psp_skip_tmr(struct psp_context *psp)
case IP_VERSION(13, 0, 2):
case IP_VERSION(13, 0, 6):
case IP_VERSION(13, 0, 10):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
return true;
default:
@@ -2264,7 +2267,8 @@ int psp_securedisplay_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
return -EINVAL;
if (ta_cmd_id != TA_SECUREDISPLAY_COMMAND__QUERY_TA &&
- ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC)
+ ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC &&
+ ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC_V2)
return -EINVAL;
ret = psp_ta_invoke(psp, ta_cmd_id, &psp->securedisplay_context.context);
@@ -2385,6 +2389,15 @@ static int psp_hw_start(struct psp_context *psp)
}
}
+ if ((is_psp_fw_valid(psp->spdm_drv)) &&
+ (psp->funcs->bootloader_load_spdm_drv != NULL)) {
+ ret = psp_bootloader_load_spdm_drv(psp);
+ if (ret) {
+ dev_err(adev->dev, "PSP load spdm_drv failed!\n");
+ return ret;
+ }
+ }
+
if ((is_psp_fw_valid(psp->sos)) &&
(psp->funcs->bootloader_load_sos != NULL)) {
ret = psp_bootloader_load_sos(psp);
@@ -3007,10 +3020,7 @@ static int psp_hw_init(struct amdgpu_ip_block *ip_block)
struct amdgpu_device *adev = ip_block->adev;
mutex_lock(&adev->firmware.mutex);
- /*
- * This sequence is just used on hw_init only once, no need on
- * resume.
- */
+
ret = amdgpu_ucode_init_bo(adev);
if (ret)
goto failed;
@@ -3135,6 +3145,10 @@ static int psp_resume(struct amdgpu_ip_block *ip_block)
mutex_lock(&adev->firmware.mutex);
+ ret = amdgpu_ucode_init_bo(adev);
+ if (ret)
+ goto failed;
+
ret = psp_hw_start(psp);
if (ret)
goto failed;
@@ -3289,7 +3303,8 @@ int psp_init_asd_microcode(struct psp_context *psp, const char *chip_name)
const struct psp_firmware_header_v1_0 *asd_hdr;
int err = 0;
- err = amdgpu_ucode_request(adev, &adev->psp.asd_fw, "amdgpu/%s_asd.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->psp.asd_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_asd.bin", chip_name);
if (err)
goto out;
@@ -3311,7 +3326,8 @@ int psp_init_toc_microcode(struct psp_context *psp, const char *chip_name)
const struct psp_firmware_header_v1_0 *toc_hdr;
int err = 0;
- err = amdgpu_ucode_request(adev, &adev->psp.toc_fw, "amdgpu/%s_toc.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->psp.toc_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_toc.bin", chip_name);
if (err)
goto out;
@@ -3407,6 +3423,12 @@ static int parse_sos_bin_descriptor(struct psp_context *psp,
psp->ipkeymgr_drv.size_bytes = le32_to_cpu(desc->size_bytes);
psp->ipkeymgr_drv.start_addr = ucode_start_addr;
break;
+ case PSP_FW_TYPE_PSP_SPDM_DRV:
+ psp->spdm_drv.fw_version = le32_to_cpu(desc->fw_version);
+ psp->spdm_drv.feature_version = le32_to_cpu(desc->fw_version);
+ psp->spdm_drv.size_bytes = le32_to_cpu(desc->size_bytes);
+ psp->spdm_drv.start_addr = ucode_start_addr;
+ break;
default:
dev_warn(psp->adev->dev, "Unsupported PSP FW type: %d\n", desc->fw_type);
break;
@@ -3474,7 +3496,8 @@ int psp_init_sos_microcode(struct psp_context *psp, const char *chip_name)
uint8_t *ucode_array_start_addr;
int err = 0;
- err = amdgpu_ucode_request(adev, &adev->psp.sos_fw, "amdgpu/%s_sos.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->psp.sos_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_sos.bin", chip_name);
if (err)
goto out;
@@ -3750,7 +3773,8 @@ int psp_init_ta_microcode(struct psp_context *psp, const char *chip_name)
struct amdgpu_device *adev = psp->adev;
int err;
- err = amdgpu_ucode_request(adev, &adev->psp.ta_fw, "amdgpu/%s_ta.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->psp.ta_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_ta.bin", chip_name);
if (err)
return err;
@@ -3785,14 +3809,16 @@ int psp_init_cap_microcode(struct psp_context *psp, const char *chip_name)
return -EINVAL;
}
- err = amdgpu_ucode_request(adev, &adev->psp.cap_fw, "amdgpu/%s_cap.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->psp.cap_fw, AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s_cap.bin", chip_name);
if (err) {
if (err == -ENODEV) {
dev_warn(adev->dev, "cap microcode does not exist, skip\n");
err = 0;
- goto out;
+ } else {
+ dev_err(adev->dev, "fail to initialize cap microcode\n");
}
- dev_err(adev->dev, "fail to initialize cap microcode\n");
+ goto out;
}
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CAP];
@@ -3849,13 +3875,13 @@ int psp_config_sq_perfmon(struct psp_context *psp,
return ret;
}
-static int psp_set_clockgating_state(void *handle,
+static int psp_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int psp_set_powergating_state(void *handle,
+static int psp_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
@@ -3867,10 +3893,12 @@ static ssize_t psp_usbc_pd_fw_sysfs_read(struct device *dev,
{
struct drm_device *ddev = dev_get_drvdata(dev);
struct amdgpu_device *adev = drm_to_adev(ddev);
+ struct amdgpu_ip_block *ip_block;
uint32_t fw_ver;
int ret;
- if (!adev->ip_blocks[AMD_IP_BLOCK_TYPE_PSP].status.late_initialized) {
+ ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP);
+ if (!ip_block || !ip_block->status.late_initialized) {
dev_info(adev->dev, "PSP block is not ready yet\n.");
return -EBUSY;
}
@@ -3899,8 +3927,10 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
struct amdgpu_bo *fw_buf_bo = NULL;
uint64_t fw_pri_mc_addr;
void *fw_pri_cpu_addr;
+ struct amdgpu_ip_block *ip_block;
- if (!adev->ip_blocks[AMD_IP_BLOCK_TYPE_PSP].status.late_initialized) {
+ ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP);
+ if (!ip_block || !ip_block->status.late_initialized) {
dev_err(adev->dev, "PSP block is not ready yet.");
return -EBUSY;
}
@@ -3908,7 +3938,8 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
if (!drm_dev_enter(ddev, &idx))
return -ENODEV;
- ret = amdgpu_ucode_request(adev, &usbc_pd_fw, "amdgpu/%s", buf);
+ ret = amdgpu_ucode_request(adev, &usbc_pd_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s", buf);
if (ret)
goto fail;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index 567cb1f924ca..8d5acc415d38 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -80,6 +80,7 @@ enum psp_bootloader_cmd {
PSP_BL__DRAM_LONG_TRAIN = 0x100000,
PSP_BL__DRAM_SHORT_TRAIN = 0x200000,
PSP_BL__LOAD_TOS_SPL_TABLE = 0x10000000,
+ PSP_BL__LOAD_SPDMDRV = 0x20000000,
};
enum psp_ring_type {
@@ -120,6 +121,7 @@ struct psp_funcs {
int (*bootloader_load_dbg_drv)(struct psp_context *psp);
int (*bootloader_load_ras_drv)(struct psp_context *psp);
int (*bootloader_load_ipkeymgr_drv)(struct psp_context *psp);
+ int (*bootloader_load_spdm_drv)(struct psp_context *psp);
int (*bootloader_load_sos)(struct psp_context *psp);
int (*ring_create)(struct psp_context *psp,
enum psp_ring_type ring_type);
@@ -343,6 +345,7 @@ struct psp_context {
struct psp_bin_desc dbg_drv;
struct psp_bin_desc ras_drv;
struct psp_bin_desc ipkeymgr_drv;
+ struct psp_bin_desc spdm_drv;
/* tmr buffer */
struct amdgpu_bo *tmr_bo;
@@ -434,6 +437,9 @@ struct amdgpu_psp_funcs {
#define psp_bootloader_load_ipkeymgr_drv(psp) \
((psp)->funcs->bootloader_load_ipkeymgr_drv ? \
(psp)->funcs->bootloader_load_ipkeymgr_drv((psp)) : 0)
+#define psp_bootloader_load_spdm_drv(psp) \
+ ((psp)->funcs->bootloader_load_spdm_drv ? \
+ (psp)->funcs->bootloader_load_spdm_drv((psp)) : 0)
#define psp_bootloader_load_sos(psp) \
((psp)->funcs->bootloader_load_sos ? (psp)->funcs->bootloader_load_sos((psp)) : 0)
#define psp_smu_reload_quirk(psp) \
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 4c9fa24dd972..f0924aa3f4e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -36,6 +36,7 @@
#include "amdgpu_xgmi.h"
#include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
#include "nbio_v4_3.h"
+#include "nbif_v6_3_1.h"
#include "nbio_v7_9.h"
#include "atom.h"
#include "amdgpu_reset.h"
@@ -192,7 +193,7 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre
if (amdgpu_bad_page_threshold != 0) {
amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
- err_data.err_addr_cnt);
+ err_data.err_addr_cnt, false);
amdgpu_ras_save_bad_pages(adev, NULL);
}
@@ -2015,6 +2016,7 @@ static bool amdgpu_ras_aca_is_supported(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
ret = true;
break;
@@ -2156,6 +2158,16 @@ 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
+ * interrupt. The error gets propagated to other devices as well, so
+ * make sure to ack the interrupt regardless of FED status.
+ */
+ if (!amdgpu_ras_get_fed_status(adev) &&
+ amdgpu_ras_is_err_state(adev, AMDGPU_RAS_BLOCK__ANY))
+ return;
if (adev->nbio.ras &&
adev->nbio.ras->handle_ras_controller_intr_no_bifring)
@@ -2185,6 +2197,7 @@ static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager *
if (ret)
return;
+ amdgpu_ras_set_err_poison(adev, block_obj->ras_comm.block);
/* both query_poison_status and handle_poison_consumption are optional,
* but at least one of them should be implemented if we need poison
* consumption handler
@@ -2717,40 +2730,203 @@ static int amdgpu_ras_realloc_eh_data_space(struct amdgpu_device *adev,
return 0;
}
+static int amdgpu_ras_mca2pa_by_idx(struct amdgpu_device *adev,
+ struct eeprom_table_record *bps,
+ struct ras_err_data *err_data)
+{
+ struct ta_ras_query_address_input addr_in;
+ uint32_t socket = 0;
+ int ret = 0;
+
+ if (adev->smuio.funcs && adev->smuio.funcs->get_socket_id)
+ socket = adev->smuio.funcs->get_socket_id(adev);
+
+ /* reinit err_data */
+ err_data->err_addr_cnt = 0;
+ err_data->err_addr_len = adev->umc.retire_unit;
+
+ memset(&addr_in, 0, sizeof(addr_in));
+ addr_in.ma.err_addr = bps->address;
+ addr_in.ma.socket_id = socket;
+ addr_in.ma.ch_inst = bps->mem_channel;
+ /* tell RAS TA the node instance is not used */
+ addr_in.ma.node_inst = TA_RAS_INV_NODE;
+
+ if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+ ret = adev->umc.ras->convert_ras_err_addr(adev, err_data,
+ &addr_in, NULL, false);
+
+ return ret;
+}
+
+static int amdgpu_ras_mca2pa(struct amdgpu_device *adev,
+ struct eeprom_table_record *bps,
+ struct ras_err_data *err_data)
+{
+ struct ta_ras_query_address_input addr_in;
+ uint32_t die_id, socket = 0;
+
+ if (adev->smuio.funcs && adev->smuio.funcs->get_socket_id)
+ socket = adev->smuio.funcs->get_socket_id(adev);
+
+ /* although die id is gotten from PA in nps1 mode, the id is
+ * fitable for any nps mode
+ */
+ if (adev->umc.ras && adev->umc.ras->get_die_id_from_pa)
+ die_id = adev->umc.ras->get_die_id_from_pa(adev, bps->address,
+ bps->retired_page << AMDGPU_GPU_PAGE_SHIFT);
+ else
+ return -EINVAL;
+
+ /* reinit err_data */
+ err_data->err_addr_cnt = 0;
+ err_data->err_addr_len = adev->umc.retire_unit;
+
+ memset(&addr_in, 0, sizeof(addr_in));
+ addr_in.ma.err_addr = bps->address;
+ addr_in.ma.ch_inst = bps->mem_channel;
+ addr_in.ma.umc_inst = bps->mcumc_id;
+ addr_in.ma.node_inst = die_id;
+ addr_in.ma.socket_id = socket;
+
+ if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+ return adev->umc.ras->convert_ras_err_addr(adev, err_data,
+ &addr_in, NULL, false);
+ else
+ return -EINVAL;
+}
+
/* it deal with vram only. */
int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
- struct eeprom_table_record *bps, int pages)
+ struct eeprom_table_record *bps, int pages, bool from_rom)
{
struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
struct ras_err_handler_data *data;
+ struct ras_err_data err_data;
+ struct eeprom_table_record *err_rec;
+ struct amdgpu_ras_eeprom_control *control =
+ &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, j, loop_cnt = 1;
+ bool find_pages_per_pa = false;
if (!con || !con->eh_data || !bps || pages <= 0)
return 0;
+ if (from_rom) {
+ err_data.err_addr =
+ kcalloc(adev->umc.retire_unit,
+ sizeof(struct eeprom_table_record), GFP_KERNEL);
+ if (!err_data.err_addr) {
+ dev_warn(adev->dev, "Failed to alloc UMC error address record in mca2pa conversion!\n");
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ err_rec = err_data.err_addr;
+ loop_cnt = adev->umc.retire_unit;
+ if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+ nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+ }
+
mutex_lock(&con->recovery_lock);
data = con->eh_data;
- if (!data)
- goto out;
+ if (!data) {
+ /* Returning 0 as the absence of eh_data is acceptable */
+ goto free;
+ }
for (i = 0; i < pages; i++) {
- if (amdgpu_ras_check_bad_page_unlock(con,
- bps[i].retired_page << AMDGPU_GPU_PAGE_SHIFT))
- continue;
+ if (from_rom &&
+ control->rec_type == AMDGPU_RAS_EEPROM_REC_MCA) {
+ if (!find_pages_per_pa) {
+ if (amdgpu_ras_mca2pa_by_idx(adev, &bps[i], &err_data)) {
+ if (!i && nps == AMDGPU_NPS1_PARTITION_MODE) {
+ /* may use old RAS TA, use PA to find pages in
+ * one row
+ */
+ if (amdgpu_umc_pages_in_a_row(adev, &err_data,
+ bps[i].retired_page <<
+ AMDGPU_GPU_PAGE_SHIFT)) {
+ ret = -EINVAL;
+ goto free;
+ } else {
+ find_pages_per_pa = true;
+ }
+ } else {
+ /* unsupported cases */
+ ret = -EOPNOTSUPP;
+ goto free;
+ }
+ }
+ } else {
+ if (amdgpu_umc_pages_in_a_row(adev, &err_data,
+ bps[i].retired_page << AMDGPU_GPU_PAGE_SHIFT)) {
+ ret = -EINVAL;
+ goto free;
+ }
+ }
+ } else {
+ if (from_rom && !find_pages_per_pa) {
+ if (bps[i].retired_page & UMC_CHANNEL_IDX_V2) {
+ /* bad page in any NPS mode in eeprom */
+ if (amdgpu_ras_mca2pa_by_idx(adev, &bps[i], &err_data)) {
+ ret = -EINVAL;
+ goto free;
+ }
+ } else {
+ /* legacy bad page in eeprom, generated only in
+ * NPS1 mode
+ */
+ if (amdgpu_ras_mca2pa(adev, &bps[i], &err_data)) {
+ /* old RAS TA or ASICs which don't support to
+ * convert addrss via mca address
+ */
+ if (!i && nps == AMDGPU_NPS1_PARTITION_MODE) {
+ find_pages_per_pa = true;
+ err_rec = &bps[i];
+ loop_cnt = 1;
+ } else {
+ /* non-nps1 mode, old RAS TA
+ * can't support it
+ */
+ ret = -EOPNOTSUPP;
+ goto free;
+ }
+ }
+ }
- if (!data->space_left &&
- amdgpu_ras_realloc_eh_data_space(adev, data, 256)) {
- ret = -ENOMEM;
- goto out;
+ if (!find_pages_per_pa)
+ i += (adev->umc.retire_unit - 1);
+ } else {
+ err_rec = &bps[i];
+ }
}
- amdgpu_ras_reserve_page(adev, bps[i].retired_page);
+ for (j = 0; j < loop_cnt; j++) {
+ if (amdgpu_ras_check_bad_page_unlock(con,
+ err_rec[j].retired_page << AMDGPU_GPU_PAGE_SHIFT))
+ continue;
+
+ if (!data->space_left &&
+ amdgpu_ras_realloc_eh_data_space(adev, data, 256)) {
+ ret = -ENOMEM;
+ goto free;
+ }
- memcpy(&data->bps[data->count], &bps[i], sizeof(*data->bps));
- data->count++;
- data->space_left--;
+ amdgpu_ras_reserve_page(adev, err_rec[j].retired_page);
+
+ memcpy(&data->bps[data->count], &(err_rec[j]),
+ sizeof(struct eeprom_table_record));
+ data->count++;
+ data->space_left--;
+ }
}
+
+free:
+ if (from_rom)
+ kfree(err_data.err_addr);
out:
mutex_unlock(&con->recovery_lock);
@@ -2768,7 +2944,7 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
struct ras_err_handler_data *data;
struct amdgpu_ras_eeprom_control *control;
- int save_count;
+ int save_count, unit_num, bad_page_num, i;
if (!con || !con->eh_data) {
if (new_cnt)
@@ -2780,19 +2956,32 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
mutex_lock(&con->recovery_lock);
control = &con->eeprom_control;
data = con->eh_data;
- save_count = data->count - control->ras_num_recs;
+ bad_page_num = control->ras_num_bad_pages;
+ save_count = data->count - bad_page_num;
mutex_unlock(&con->recovery_lock);
+ unit_num = save_count / adev->umc.retire_unit;
if (new_cnt)
- *new_cnt = save_count / adev->umc.retire_unit;
+ *new_cnt = unit_num;
/* only new entries are saved */
if (save_count > 0) {
- if (amdgpu_ras_eeprom_append(control,
- &data->bps[control->ras_num_recs],
- save_count)) {
- dev_err(adev->dev, "Failed to save EEPROM table data!");
- return -EIO;
+ if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA) {
+ if (amdgpu_ras_eeprom_append(control,
+ &data->bps[control->ras_num_recs],
+ save_count)) {
+ dev_err(adev->dev, "Failed to save EEPROM table data!");
+ return -EIO;
+ }
+ } else {
+ for (i = 0; i < unit_num; i++) {
+ if (amdgpu_ras_eeprom_append(control,
+ &data->bps[bad_page_num + i * adev->umc.retire_unit],
+ 1)) {
+ dev_err(adev->dev, "Failed to save EEPROM table data!");
+ return -EIO;
+ }
+ }
}
dev_info(adev->dev, "Saved %d pages to EEPROM table.\n", save_count);
@@ -2821,11 +3010,32 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)
return -ENOMEM;
ret = amdgpu_ras_eeprom_read(control, bps, control->ras_num_recs);
- if (ret)
+ if (ret) {
dev_err(adev->dev, "Failed to load EEPROM table records!");
- else
- ret = amdgpu_ras_add_bad_pages(adev, bps, control->ras_num_recs);
+ } else {
+ if (control->ras_num_recs > 1 &&
+ adev->umc.ras && adev->umc.ras->convert_ras_err_addr) {
+ if ((bps[0].address == bps[1].address) &&
+ (bps[0].mem_channel == bps[1].mem_channel))
+ control->rec_type = AMDGPU_RAS_EEPROM_REC_PA;
+ else
+ control->rec_type = AMDGPU_RAS_EEPROM_REC_MCA;
+ }
+
+ ret = amdgpu_ras_eeprom_check(control);
+ if (ret)
+ goto out;
+
+ /* HW not usable */
+ if (amdgpu_ras_is_rma(adev)) {
+ ret = -EHWPOISON;
+ goto out;
+ }
+ ret = amdgpu_ras_add_bad_pages(adev, bps, control->ras_num_recs, true);
+ }
+
+out:
kfree(bps);
return ret;
}
@@ -3205,31 +3415,36 @@ static int amdgpu_ras_page_retirement_thread(void *param)
int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev)
{
struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+ struct amdgpu_ras_eeprom_control *control;
int ret;
if (!con || amdgpu_sriov_vf(adev))
return 0;
- ret = amdgpu_ras_eeprom_init(&con->eeprom_control);
-
+ control = &con->eeprom_control;
+ ret = amdgpu_ras_eeprom_init(control);
if (ret)
return ret;
- /* HW not usable */
- if (amdgpu_ras_is_rma(adev))
- return -EHWPOISON;
+ if (!adev->umc.ras || !adev->umc.ras->convert_ras_err_addr)
+ control->rec_type = AMDGPU_RAS_EEPROM_REC_PA;
+
+ /* default status is MCA storage */
+ if (control->ras_num_recs <= 1 &&
+ adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+ control->rec_type = AMDGPU_RAS_EEPROM_REC_MCA;
- if (con->eeprom_control.ras_num_recs) {
+ if (control->ras_num_recs) {
ret = amdgpu_ras_load_bad_pages(adev);
if (ret)
return ret;
amdgpu_dpm_send_hbm_bad_pages_num(
- adev, con->eeprom_control.ras_num_recs);
+ adev, control->ras_num_bad_pages);
if (con->update_channel_flag == true) {
amdgpu_dpm_send_hbm_bad_channel_flag(
- adev, con->eeprom_control.bad_channel_bitmap);
+ adev, control->bad_channel_bitmap);
con->update_channel_flag = false;
}
}
@@ -3366,6 +3581,7 @@ static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
case IP_VERSION(13, 0, 2):
case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
return true;
default:
@@ -3378,7 +3594,9 @@ static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev)
case IP_VERSION(13, 0, 0):
case IP_VERSION(13, 0, 6):
case IP_VERSION(13, 0, 10):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
+ case IP_VERSION(14, 0, 3):
return true;
default:
return false;
@@ -3629,6 +3847,7 @@ static void amdgpu_ras_init_reserved_vram_size(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
case IP_VERSION(13, 0, 2):
case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
con->reserved_pages_in_bytes = AMDGPU_RAS_RESERVED_VRAM_SIZE;
break;
@@ -3704,7 +3923,19 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
* check DF RAS */
adev->nbio.ras = &nbio_v4_3_ras;
break;
+ case IP_VERSION(6, 3, 1):
+ if (adev->ras_hw_enabled & (1 << AMDGPU_RAS_BLOCK__DF))
+ /* unlike other generation of nbio ras,
+ * nbif v6_3_1 only support fatal error interrupt
+ * to inform software that DF is freezed due to
+ * system fatal error event. driver should not
+ * enable nbio ras in such case. Instead,
+ * check DF RAS
+ */
+ adev->nbio.ras = &nbif_v6_3_1_ras;
+ break;
case IP_VERSION(7, 9, 0):
+ case IP_VERSION(7, 9, 1):
if (!adev->gmc.is_app_apu)
adev->nbio.ras = &nbio_v7_9_ras;
break;
@@ -4083,7 +4314,7 @@ bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev)
if (!ras)
return false;
- return atomic_read(&ras->fed);
+ return test_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
}
void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status)
@@ -4091,8 +4322,48 @@ void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status)
struct amdgpu_ras *ras;
ras = amdgpu_ras_get_context(adev);
+ if (ras) {
+ if (status)
+ set_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+ else
+ clear_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+ }
+}
+
+void amdgpu_ras_clear_err_state(struct amdgpu_device *adev)
+{
+ struct amdgpu_ras *ras;
+
+ ras = amdgpu_ras_get_context(adev);
+ if (ras)
+ ras->ras_err_state = 0;
+}
+
+void amdgpu_ras_set_err_poison(struct amdgpu_device *adev,
+ enum amdgpu_ras_block block)
+{
+ struct amdgpu_ras *ras;
+
+ ras = amdgpu_ras_get_context(adev);
if (ras)
- atomic_set(&ras->fed, !!status);
+ set_bit(block, &ras->ras_err_state);
+}
+
+bool amdgpu_ras_is_err_state(struct amdgpu_device *adev, int block)
+{
+ struct amdgpu_ras *ras;
+
+ ras = amdgpu_ras_get_context(adev);
+ if (ras) {
+ if (block == AMDGPU_RAS_BLOCK__ANY)
+ return (ras->ras_err_state != 0);
+ else
+ return test_bit(block, &ras->ras_err_state) ||
+ test_bit(AMDGPU_RAS_BLOCK__LAST,
+ &ras->ras_err_state);
+ }
+
+ return false;
}
static struct ras_event_manager *__get_ras_event_mgr(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index 6db772ecfee4..82db986c36a0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -99,7 +99,8 @@ enum amdgpu_ras_block {
AMDGPU_RAS_BLOCK__IH,
AMDGPU_RAS_BLOCK__MPIO,
- AMDGPU_RAS_BLOCK__LAST
+ AMDGPU_RAS_BLOCK__LAST,
+ AMDGPU_RAS_BLOCK__ANY = -1
};
enum amdgpu_ras_mca_block {
@@ -482,6 +483,8 @@ struct ras_ecc_err {
uint64_t ipid;
uint64_t addr;
uint64_t pa_pfn;
+ /* save global channel index across all UMC instances */
+ uint32_t channel_idx;
struct ras_err_pages err_pages;
};
@@ -558,8 +561,8 @@ struct amdgpu_ras {
struct ras_ecc_log_info umc_ecc_log;
struct delayed_work page_retirement_dwork;
- /* Fatal error detected flag */
- atomic_t fed;
+ /* ras errors detected */
+ unsigned long ras_err_state;
/* RAS event manager */
struct ras_event_manager __event_mgr;
@@ -750,7 +753,7 @@ int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
/* error handling functions */
int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
- struct eeprom_table_record *bps, int pages);
+ struct eeprom_table_record *bps, int pages, bool from_rom);
int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
unsigned long *new_cnt);
@@ -952,6 +955,10 @@ ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *a
void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status);
bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev);
+void amdgpu_ras_set_err_poison(struct amdgpu_device *adev,
+ enum amdgpu_ras_block block);
+void amdgpu_ras_clear_err_state(struct amdgpu_device *adev);
+bool amdgpu_ras_is_err_state(struct amdgpu_device *adev, int block);
u64 amdgpu_ras_acquire_event_id(struct amdgpu_device *adev, enum ras_event_type type);
int amdgpu_ras_mark_ras_event_caller(struct amdgpu_device *adev, enum ras_event_type type,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index f28f6b4ba765..52c16bfeccaa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -470,9 +470,10 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
res = __write_table_ras_info(control);
control->ras_num_recs = 0;
+ control->ras_num_bad_pages = 0;
control->ras_fri = 0;
- amdgpu_dpm_send_hbm_bad_pages_num(adev, control->ras_num_recs);
+ amdgpu_dpm_send_hbm_bad_pages_num(adev, control->ras_num_bad_pages);
control->bad_channel_bitmap = 0;
amdgpu_dpm_send_hbm_bad_channel_flag(adev, control->bad_channel_bitmap);
@@ -559,7 +560,7 @@ bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev)
if (con->eeprom_control.tbl_hdr.header == RAS_TABLE_HDR_BAD) {
if (amdgpu_bad_page_threshold == -1) {
dev_warn(adev->dev, "RAS records:%d exceed threshold:%d",
- con->eeprom_control.ras_num_recs, con->bad_page_cnt_threshold);
+ con->eeprom_control.ras_num_bad_pages, con->bad_page_cnt_threshold);
dev_warn(adev->dev,
"But GPU can be operated due to bad_page_threshold = -1.\n");
return false;
@@ -621,6 +622,7 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
const u32 num)
{
struct amdgpu_ras *con = amdgpu_ras_get_context(to_amdgpu_device(control));
+ struct amdgpu_device *adev = to_amdgpu_device(control);
u32 a, b, i;
u8 *buf, *pp;
int res;
@@ -723,6 +725,12 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
control->ras_num_recs = 1 + (control->ras_max_record_count + b
- control->ras_fri)
% control->ras_max_record_count;
+
+ if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA)
+ control->ras_num_bad_pages = control->ras_num_recs;
+ else
+ control->ras_num_bad_pages =
+ control->ras_num_recs * adev->umc.retire_unit;
Out:
kfree(buf);
return res;
@@ -740,10 +748,10 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
/* Modify the header if it exceeds.
*/
if (amdgpu_bad_page_threshold != 0 &&
- control->ras_num_recs >= ras->bad_page_cnt_threshold) {
+ control->ras_num_bad_pages >= ras->bad_page_cnt_threshold) {
dev_warn(adev->dev,
"Saved bad pages %d reaches threshold value %d\n",
- control->ras_num_recs, ras->bad_page_cnt_threshold);
+ control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
control->tbl_hdr.header = RAS_TABLE_HDR_BAD;
if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) {
control->tbl_rai.rma_status = GPU_RETIRED__ECC_REACH_THRESHOLD;
@@ -798,9 +806,9 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
*/
if (amdgpu_bad_page_threshold != 0 &&
control->tbl_hdr.version == RAS_TABLE_VER_V2_1 &&
- control->ras_num_recs < ras->bad_page_cnt_threshold)
+ control->ras_num_bad_pages < ras->bad_page_cnt_threshold)
control->tbl_rai.health_percent = ((ras->bad_page_cnt_threshold -
- control->ras_num_recs) * 100) /
+ control->ras_num_bad_pages) * 100) /
ras->bad_page_cnt_threshold;
/* Recalc the checksum.
@@ -841,7 +849,7 @@ int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
const u32 num)
{
struct amdgpu_device *adev = to_amdgpu_device(control);
- int res;
+ int res, i;
if (!__is_ras_eeprom_supported(adev))
return 0;
@@ -855,6 +863,10 @@ int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
return -EINVAL;
}
+ /* set the new channel index flag */
+ for (i = 0; i < num; i++)
+ record[i].retired_page |= UMC_CHANNEL_IDX_V2;
+
mutex_lock(&control->ras_tbl_mutex);
res = amdgpu_ras_eeprom_append_table(control, record, num);
@@ -864,6 +876,11 @@ int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
amdgpu_ras_debugfs_set_ret_size(control);
mutex_unlock(&control->ras_tbl_mutex);
+
+ /* clear channel index flag, the flag is only saved on eeprom */
+ for (i = 0; i < num; i++)
+ record[i].retired_page &= ~UMC_CHANNEL_IDX_V2;
+
return res;
}
@@ -1373,9 +1390,35 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
}
control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset);
+ return 0;
+}
+
+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;
+
+ if (!__is_ras_eeprom_supported(adev))
+ return 0;
+
+ /* Verify i2c adapter is initialized */
+ if (!adev->pm.ras_eeprom_i2c_bus || !adev->pm.ras_eeprom_i2c_bus->algo)
+ return -ENOENT;
+
+ if (!__get_eeprom_i2c_addr(adev, control))
+ return -EINVAL;
+
+ if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA)
+ control->ras_num_bad_pages = control->ras_num_recs;
+ else
+ control->ras_num_bad_pages =
+ control->ras_num_recs * adev->umc.retire_unit;
+
if (hdr->header == RAS_TABLE_HDR_VAL) {
DRM_DEBUG_DRIVER("Found existing EEPROM table with %d records",
- control->ras_num_recs);
+ control->ras_num_bad_pages);
if (hdr->version == RAS_TABLE_VER_V2_1) {
res = __read_table_ras_info(control);
@@ -1390,9 +1433,9 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
/* Warn if we are at 90% of the threshold or above
*/
- if (10 * control->ras_num_recs >= 9 * ras->bad_page_cnt_threshold)
+ if (10 * control->ras_num_bad_pages >= 9 * ras->bad_page_cnt_threshold)
dev_warn(adev->dev, "RAS records:%u exceeds 90%% of threshold:%d",
- control->ras_num_recs,
+ control->ras_num_bad_pages,
ras->bad_page_cnt_threshold);
} else if (hdr->header == RAS_TABLE_HDR_BAD &&
amdgpu_bad_page_threshold != 0) {
@@ -1403,10 +1446,12 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
}
res = __verify_ras_table_checksum(control);
- if (res)
- DRM_ERROR("RAS Table incorrect checksum or error:%d\n",
+ if (res) {
+ dev_err(adev->dev, "RAS Table incorrect checksum or error:%d\n",
res);
- if (ras->bad_page_cnt_threshold > control->ras_num_recs) {
+ return -EINVAL;
+ }
+ if (ras->bad_page_cnt_threshold > control->ras_num_bad_pages) {
/* This means that, the threshold was increased since
* the last time the system was booted, and now,
* ras->bad_page_cnt_threshold - control->num_recs > 0,
@@ -1416,13 +1461,13 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
dev_info(adev->dev,
"records:%d threshold:%d, resetting "
"RAS table header signature",
- control->ras_num_recs,
+ control->ras_num_bad_pages,
ras->bad_page_cnt_threshold);
res = amdgpu_ras_eeprom_correct_header_tag(control,
RAS_TABLE_HDR_VAL);
} else {
dev_err(adev->dev, "RAS records:%d exceed threshold:%d",
- control->ras_num_recs, ras->bad_page_cnt_threshold);
+ control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
if (amdgpu_bad_page_threshold == -1) {
dev_warn(adev->dev, "GPU will be initialized due to bad_page_threshold = -1.");
res = 0;
@@ -1431,7 +1476,7 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
dev_err(adev->dev,
"RAS records:%d exceed threshold:%d, "
"GPU will not be initialized. Replace this GPU or increase the threshold",
- control->ras_num_recs, ras->bad_page_cnt_threshold);
+ control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
}
}
} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
index b9ebda577797..81d55cb7b397 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
@@ -43,6 +43,19 @@ enum amdgpu_ras_eeprom_err_type {
AMDGPU_RAS_EEPROM_ERR_COUNT,
};
+/*
+ * one UMC MCA address could map to multiply physical address (PA),
+ * such as 1:16, we use eeprom_table_record.address to store MCA
+ * address and use eeprom_table_record.retired_page to save PA.
+ *
+ * AMDGPU_RAS_EEPROM_REC_PA: one record store one PA
+ * AMDGPU_RAS_EEPROM_REC_MCA: one record store one MCA address
+ */
+enum amdgpu_ras_eeprom_rec_type {
+ AMDGPU_RAS_EEPROM_REC_PA,
+ AMDGPU_RAS_EEPROM_REC_MCA,
+};
+
struct amdgpu_ras_eeprom_table_header {
uint32_t header;
uint32_t version;
@@ -82,6 +95,11 @@ struct amdgpu_ras_eeprom_control {
*/
u32 ras_num_recs;
+ /* the bad page number is ras_num_recs or
+ * ras_num_recs * umc.retire_unit
+ */
+ u32 ras_num_bad_pages;
+
/* First record index to read, 0-based.
* Range is [0, num_recs-1]. This is
* an absolute index, starting right after
@@ -102,6 +120,7 @@ struct amdgpu_ras_eeprom_control {
/* Record channel info which occurred bad pages
*/
u32 bad_channel_bitmap;
+ enum amdgpu_ras_eeprom_rec_type rec_type;
};
/*
@@ -145,6 +164,8 @@ uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *co
void amdgpu_ras_debugfs_set_ret_size(struct amdgpu_ras_eeprom_control *control);
+int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control);
+
extern const struct file_operations amdgpu_ras_debugfs_eeprom_size_ops;
extern const struct file_operations amdgpu_ras_debugfs_eeprom_table_ops;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c
index a0acb65f4b40..dabfbdf6f1ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c
@@ -183,6 +183,7 @@ int amdgpu_reset_init(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
case IP_VERSION(13, 0, 2):
case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
ret = aldebaran_reset_init(adev);
break;
@@ -206,6 +207,7 @@ int amdgpu_reset_fini(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
case IP_VERSION(13, 0, 2):
case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
ret = aldebaran_reset_fini(adev);
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
index 36fc9578c53c..dee5a1b4e572 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
@@ -462,8 +462,7 @@ int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
unsigned size,
enum amdgpu_ib_pool_type pool,
struct amdgpu_ib *ib);
-void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
- struct dma_fence *f);
+void amdgpu_ib_free(struct amdgpu_ib *ib, struct dma_fence *f);
int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
struct amdgpu_ib *ibs, struct amdgpu_job *job,
struct dma_fence **f);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
index 10df731998b2..39070b2a4c04 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
@@ -93,8 +93,7 @@ int amdgpu_sa_bo_new(struct amdgpu_sa_manager *sa_manager,
return 0;
}
-void amdgpu_sa_bo_free(struct amdgpu_device *adev, struct drm_suballoc **sa_bo,
- struct dma_fence *fence)
+void amdgpu_sa_bo_free(struct drm_suballoc **sa_bo, struct dma_fence *fence)
{
if (sa_bo == NULL || *sa_bo == NULL) {
return;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
index 113f0d242618..174badca27e7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
@@ -219,9 +219,11 @@ int amdgpu_sdma_init_microcode(struct amdgpu_device *adev,
amdgpu_ucode_ip_version_decode(adev, SDMA0_HWIP, ucode_prefix, sizeof(ucode_prefix));
if (instance == 0)
err = amdgpu_ucode_request(adev, &adev->sdma.instance[instance].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s.bin", ucode_prefix);
else
err = amdgpu_ucode_request(adev, &adev->sdma.instance[instance].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s%d.bin", ucode_prefix, instance);
if (err)
goto out;
@@ -261,6 +263,8 @@ int amdgpu_sdma_init_microcode(struct amdgpu_device *adev,
if ((amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
IP_VERSION(4, 4, 2) ||
amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+ IP_VERSION(4, 4, 4) ||
+ amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
IP_VERSION(4, 4, 5)) &&
adev->firmware.load_type ==
AMDGPU_FW_LOAD_PSP &&
@@ -358,13 +362,13 @@ static int amdgpu_debugfs_sdma_sched_mask_set(void *data, u64 val)
if (!adev)
return -ENODEV;
- mask = (1 << adev->sdma.num_instances) - 1;
+ mask = BIT_ULL(adev->sdma.num_instances) - 1;
if ((val & mask) == 0)
return -EINVAL;
for (i = 0; i < adev->sdma.num_instances; ++i) {
ring = &adev->sdma.instance[i].ring;
- if (val & (1 << i))
+ if (val & BIT_ULL(i))
ring->sched.ready = true;
else
ring->sched.ready = false;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
index 2db58b5812a8..5f60736051d1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
@@ -107,6 +107,7 @@ struct amdgpu_sdma {
struct amdgpu_irq_src doorbell_invalid_irq;
struct amdgpu_irq_src pool_timeout_irq;
struct amdgpu_irq_src srbm_write_irq;
+ struct amdgpu_irq_src ctxt_empty_irq;
int num_instances;
uint32_t sdma_mask;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index c8180cad0abd..262bd010a283 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -309,7 +309,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
mutex_lock(&adev->mman.gtt_window_lock);
while (src_mm.remaining) {
uint64_t from, to, cur_size, tiling_flags;
- uint32_t num_type, data_format, max_com;
+ uint32_t num_type, data_format, max_com, write_compress_disable;
struct dma_fence *next;
/* Never copy more than 256MiB at once to avoid a timeout */
@@ -340,9 +340,13 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
max_com = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_MAX_COMPRESSED_BLOCK);
num_type = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_NUMBER_TYPE);
data_format = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_DATA_FORMAT);
+ write_compress_disable =
+ AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_WRITE_COMPRESS_DISABLE);
copy_flags |= (AMDGPU_COPY_FLAGS_SET(MAX_COMPRESSED, max_com) |
AMDGPU_COPY_FLAGS_SET(NUMBER_TYPE, num_type) |
- AMDGPU_COPY_FLAGS_SET(DATA_FORMAT, data_format));
+ AMDGPU_COPY_FLAGS_SET(DATA_FORMAT, data_format) |
+ AMDGPU_COPY_FLAGS_SET(WRITE_COMPRESS_DISABLE,
+ write_compress_disable));
}
r = amdgpu_copy_buffer(ring, from, to, cur_size, resv,
@@ -1762,7 +1766,8 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
if (!adev->bios &&
(amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)))
reserve_size = max(reserve_size, (uint32_t)280 << 20);
else if (!reserve_size)
reserve_size = DISCOVERY_TMR_OFFSET;
@@ -2065,6 +2070,7 @@ void amdgpu_ttm_fini(struct amdgpu_device *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);
+ ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_DOORBELL);
ttm_device_fini(&adev->mman.bdev);
adev->mman.initialized = false;
DRM_INFO("amdgpu: ttm finalized\n");
@@ -2275,7 +2281,7 @@ int amdgpu_ttm_clear_buffer(struct amdgpu_bo *bo,
struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
struct amdgpu_res_cursor cursor;
u64 addr;
- int r;
+ int r = 0;
if (!adev->mman.buffer_funcs_enabled)
return -EINVAL;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 2852a6064c9a..208b7d1d8a27 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -26,15 +26,15 @@
#include <linux/dma-direction.h>
#include <drm/gpu_scheduler.h>
+#include <drm/ttm/ttm_placement.h>
#include "amdgpu_vram_mgr.h"
-#include "amdgpu.h"
#define AMDGPU_PL_GDS (TTM_PL_PRIV + 0)
#define AMDGPU_PL_GWS (TTM_PL_PRIV + 1)
#define AMDGPU_PL_OA (TTM_PL_PRIV + 2)
#define AMDGPU_PL_PREEMPT (TTM_PL_PRIV + 3)
#define AMDGPU_PL_DOORBELL (TTM_PL_PRIV + 4)
-#define __AMDGPU_PL_LAST (TTM_PL_PRIV + 4)
+#define __AMDGPU_PL_NUM (TTM_PL_PRIV + 5)
#define AMDGPU_GTT_MAX_TRANSFER_SIZE 512
#define AMDGPU_GTT_NUM_TRANSFER_WINDOWS 2
@@ -119,6 +119,8 @@ struct amdgpu_copy_mem {
#define AMDGPU_COPY_FLAGS_NUMBER_TYPE_MASK 0x07
#define AMDGPU_COPY_FLAGS_DATA_FORMAT_SHIFT 8
#define AMDGPU_COPY_FLAGS_DATA_FORMAT_MASK 0x3f
+#define AMDGPU_COPY_FLAGS_WRITE_COMPRESS_DISABLE_SHIFT 14
+#define AMDGPU_COPY_FLAGS_WRITE_COMPRESS_DISABLE_MASK 0x1
#define AMDGPU_COPY_FLAGS_SET(field, value) \
(((__u32)(value) & AMDGPU_COPY_FLAGS_##field##_MASK) << AMDGPU_COPY_FLAGS_##field##_SHIFT)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
index 4c7b53648a50..cf700824b960 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
@@ -1434,6 +1434,7 @@ void amdgpu_ucode_ip_version_decode(struct amdgpu_device *adev, int block_type,
*
* @adev: amdgpu device
* @fw: pointer to load firmware to
+ * @required: whether the firmware is required
* @fmt: firmware name format string
* @...: variable arguments
*
@@ -1442,7 +1443,7 @@ void amdgpu_ucode_ip_version_decode(struct amdgpu_device *adev, int block_type,
* the error code to -ENODEV, so that early_init functions will fail to load.
*/
int amdgpu_ucode_request(struct amdgpu_device *adev, const struct firmware **fw,
- const char *fmt, ...)
+ enum amdgpu_ucode_required required, const char *fmt, ...)
{
char fname[AMDGPU_UCODE_NAME_MAX];
va_list ap;
@@ -1456,16 +1457,24 @@ int amdgpu_ucode_request(struct amdgpu_device *adev, const struct firmware **fw,
return -EOVERFLOW;
}
- r = request_firmware(fw, fname, adev->dev);
+ if (required == AMDGPU_UCODE_REQUIRED)
+ r = request_firmware(fw, fname, adev->dev);
+ else {
+ r = firmware_request_nowarn(fw, fname, adev->dev);
+ if (r)
+ drm_info(&adev->ddev, "Optional firmware \"%s\" was not found\n", fname);
+ }
if (r)
return -ENODEV;
r = amdgpu_ucode_validate(*fw);
- if (r) {
+ if (r)
+ /*
+ * The amdgpu_ucode_request() should be paired with amdgpu_ucode_release()
+ * regardless of success/failure, and the amdgpu_ucode_release() takes care of
+ * firmware release and need to avoid redundant release FW operation here.
+ */
dev_dbg(adev->dev, "\"%s\" failed to validate\n", fname);
- release_firmware(*fw);
- *fw = NULL;
- }
return r;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
index 4150ec0aa10d..4eedd92f000b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
@@ -126,6 +126,7 @@ enum psp_fw_type {
PSP_FW_TYPE_PSP_DBG_DRV,
PSP_FW_TYPE_PSP_RAS_DRV,
PSP_FW_TYPE_PSP_IPKEYMGR_DRV,
+ PSP_FW_TYPE_PSP_SPDM_DRV,
PSP_FW_TYPE_MAX_INDEX,
};
@@ -551,6 +552,11 @@ enum amdgpu_firmware_load_type {
AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO,
};
+enum amdgpu_ucode_required {
+ AMDGPU_UCODE_OPTIONAL,
+ AMDGPU_UCODE_REQUIRED,
+};
+
/* conform to smu_ucode_xfer_cz.h */
#define AMDGPU_SDMA0_UCODE_LOADED 0x00000001
#define AMDGPU_SDMA1_UCODE_LOADED 0x00000002
@@ -604,9 +610,9 @@ void amdgpu_ucode_print_rlc_hdr(const struct common_firmware_header *hdr);
void amdgpu_ucode_print_sdma_hdr(const struct common_firmware_header *hdr);
void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr);
void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr);
-__printf(3, 4)
+__printf(4, 5)
int amdgpu_ucode_request(struct amdgpu_device *adev, const struct firmware **fw,
- const char *fmt, ...);
+ enum amdgpu_ucode_required required, const char *fmt, ...);
void amdgpu_ucode_release(const struct firmware **fw);
bool amdgpu_ucode_hdr_version(union amdgpu_firmware_header *hdr,
uint16_t hdr_major, uint16_t hdr_minor);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index 896f3609b0ee..eafe20d8fe0b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -78,7 +78,7 @@ int amdgpu_umc_page_retirement_mca(struct amdgpu_device *adev,
if (amdgpu_bad_page_threshold != 0) {
amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
- err_data.err_addr_cnt);
+ err_data.err_addr_cnt, false);
amdgpu_ras_save_bad_pages(adev, NULL);
}
@@ -166,10 +166,11 @@ void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
if ((amdgpu_bad_page_threshold != 0) &&
err_data->err_addr_cnt) {
amdgpu_ras_add_bad_pages(adev, err_data->err_addr,
- err_data->err_addr_cnt);
+ err_data->err_addr_cnt, false);
amdgpu_ras_save_bad_pages(adev, &err_count);
- amdgpu_dpm_send_hbm_bad_pages_num(adev, con->eeprom_control.ras_num_recs);
+ amdgpu_dpm_send_hbm_bad_pages_num(adev,
+ con->eeprom_control.ras_num_bad_pages);
if (con->update_channel_flag == true) {
amdgpu_dpm_send_hbm_bad_channel_flag(adev, con->eeprom_control.bad_channel_bitmap);
@@ -444,3 +445,77 @@ int amdgpu_umc_logs_ecc_err(struct amdgpu_device *adev,
return ret;
}
+
+int amdgpu_umc_pages_in_a_row(struct amdgpu_device *adev,
+ struct ras_err_data *err_data, uint64_t pa_addr)
+{
+ struct ta_ras_query_address_output addr_out;
+
+ /* reinit err_data */
+ err_data->err_addr_cnt = 0;
+ err_data->err_addr_len = adev->umc.retire_unit;
+
+ addr_out.pa.pa = pa_addr;
+ if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+ return adev->umc.ras->convert_ras_err_addr(adev, err_data, NULL,
+ &addr_out, false);
+ else
+ return -EINVAL;
+}
+
+int amdgpu_umc_lookup_bad_pages_in_a_row(struct amdgpu_device *adev,
+ uint64_t pa_addr, uint64_t *pfns, int len)
+{
+ int i, ret;
+ struct ras_err_data err_data;
+
+ err_data.err_addr = kcalloc(adev->umc.retire_unit,
+ sizeof(struct eeprom_table_record), GFP_KERNEL);
+ if (!err_data.err_addr) {
+ dev_warn(adev->dev, "Failed to alloc memory in bad page lookup!\n");
+ return 0;
+ }
+
+ ret = amdgpu_umc_pages_in_a_row(adev, &err_data, pa_addr);
+ if (ret)
+ goto out;
+
+ for (i = 0; i < adev->umc.retire_unit; i++) {
+ if (i >= len)
+ goto out;
+
+ pfns[i] = err_data.err_addr[i].retired_page;
+ }
+ ret = i;
+
+out:
+ kfree(err_data.err_addr);
+ return ret;
+}
+
+int amdgpu_umc_mca_to_addr(struct amdgpu_device *adev,
+ uint64_t err_addr, uint32_t ch, uint32_t umc,
+ uint32_t node, uint32_t socket,
+ struct ta_ras_query_address_output *addr_out, bool dump_addr)
+{
+ struct ta_ras_query_address_input addr_in;
+ int ret;
+
+ memset(&addr_in, 0, sizeof(addr_in));
+ addr_in.ma.err_addr = err_addr;
+ addr_in.ma.ch_inst = ch;
+ addr_in.ma.umc_inst = umc;
+ addr_in.ma.node_inst = node;
+ addr_in.ma.socket_id = socket;
+
+ if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr) {
+ ret = adev->umc.ras->convert_ras_err_addr(adev, NULL, &addr_in,
+ addr_out, dump_addr);
+ if (ret)
+ return ret;
+ } else {
+ return 0;
+ }
+
+ return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
index ce4179db2a6d..a4a7e61817aa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
@@ -54,6 +54,22 @@
/* Page retirement tag */
#define UMC_ECC_NEW_DETECTED_TAG 0x1
+/*
+ * a flag to indicate v2 of channel index stored in eeprom
+ *
+ * v1 (legacy way): store channel index within a umc instance in eeprom
+ * range in UMC v12: 0 ~ 7
+ * v2: store global channel index in eeprom
+ * range in UMC v12: 0 ~ 127
+ *
+ * NOTE: it's better to store it in eeprom_table_record.mem_channel,
+ * but there is only 8 bits in mem_channel, and the channel number may
+ * increase in the future, we decide to save it in
+ * eeprom_table_record.retired_page. retired_page is useless in v2,
+ * we depend on eeprom_table_record.address instead of retired_page in v2.
+ * Only 48 bits are saved on eeprom, use bit 47 here.
+ */
+#define UMC_CHANNEL_IDX_V2 BIT_ULL(47)
typedef int (*umc_func)(struct amdgpu_device *adev, uint32_t node_inst,
uint32_t umc_inst, uint32_t ch_inst, void *data);
@@ -70,6 +86,13 @@ struct amdgpu_umc_ras {
enum amdgpu_mca_error_type type, void *ras_error_status);
int (*update_ecc_status)(struct amdgpu_device *adev,
uint64_t status, uint64_t ipid, uint64_t addr);
+ int (*convert_ras_err_addr)(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 (*get_die_id_from_pa)(struct amdgpu_device *adev,
+ uint64_t mca_addr, uint64_t retired_page);
};
struct amdgpu_umc_funcs {
@@ -134,4 +157,12 @@ int amdgpu_umc_logs_ecc_err(struct amdgpu_device *adev,
void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
void *ras_error_status);
+int amdgpu_umc_pages_in_a_row(struct amdgpu_device *adev,
+ struct ras_err_data *err_data, uint64_t pa_addr);
+int amdgpu_umc_lookup_bad_pages_in_a_row(struct amdgpu_device *adev,
+ uint64_t pa_addr, uint64_t *pfns, int len);
+int amdgpu_umc_mca_to_addr(struct amdgpu_device *adev,
+ uint64_t err_addr, uint32_t ch, uint32_t umc,
+ uint32_t node, uint32_t socket,
+ struct ta_ras_query_address_output *addr_out, bool dump_addr);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
index bd2d3863c3ed..dde15c6a96e1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
@@ -587,7 +587,8 @@ int amdgpu_umsch_mm_init_microcode(struct amdgpu_umsch_mm *umsch)
break;
}
- r = amdgpu_ucode_request(adev, &adev->umsch_mm.fw, "%s", fw_name);
+ r = amdgpu_ucode_request(adev, &adev->umsch_mm.fw, AMDGPU_UCODE_REQUIRED,
+ "%s", fw_name);
if (r) {
release_firmware(adev->umsch_mm.fw);
adev->umsch_mm.fw = NULL;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 65bb26215e86..74758b5ffc6c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -260,7 +260,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
return -EINVAL;
}
- r = amdgpu_ucode_request(adev, &adev->uvd.fw, "%s", fw_name);
+ r = amdgpu_ucode_request(adev, &adev->uvd.fw, AMDGPU_UCODE_REQUIRED, "%s", fw_name);
if (r) {
dev_err(adev->dev, "amdgpu_uvd: Can't validate firmware \"%s\"\n",
fw_name);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
index 599d3ca4e0ef..b9060bcd4806 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
@@ -158,7 +158,7 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
return -EINVAL;
}
- r = amdgpu_ucode_request(adev, &adev->vce.fw, "%s", fw_name);
+ r = amdgpu_ucode_request(adev, &adev->vce.fw, AMDGPU_UCODE_REQUIRED, "%s", fw_name);
if (r) {
dev_err(adev->dev, "amdgpu_vce: Can't validate firmware \"%s\"\n",
fw_name);
@@ -503,7 +503,7 @@ static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
ib->ptr[i] = 0x0;
r = amdgpu_job_submit_direct(job, ring, &f);
- amdgpu_ib_free(ring->adev, &ib_msg, f);
+ amdgpu_ib_free(&ib_msg, f);
if (r)
goto err;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 3e94c3ba1ba2..83faf6e6788a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -1,5 +1,5 @@
/*
- * Copyright 2016 Advanced Micro Devices, Inc.
+ * Copyright 2016-2024 Advanced Micro Devices, Inc.
* All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
@@ -62,6 +62,7 @@
#define FIRMWARE_VCN4_0_6 "amdgpu/vcn_4_0_6.bin"
#define FIRMWARE_VCN4_0_6_1 "amdgpu/vcn_4_0_6_1.bin"
#define FIRMWARE_VCN5_0_0 "amdgpu/vcn_5_0_0.bin"
+#define FIRMWARE_VCN5_0_1 "amdgpu/vcn_5_0_1.bin"
MODULE_FIRMWARE(FIRMWARE_RAVEN);
MODULE_FIRMWARE(FIRMWARE_PICASSO);
@@ -88,6 +89,7 @@ MODULE_FIRMWARE(FIRMWARE_VCN4_0_5);
MODULE_FIRMWARE(FIRMWARE_VCN4_0_6);
MODULE_FIRMWARE(FIRMWARE_VCN4_0_6_1);
MODULE_FIRMWARE(FIRMWARE_VCN5_0_0);
+MODULE_FIRMWARE(FIRMWARE_VCN5_0_1);
static void amdgpu_vcn_idle_work_handler(struct work_struct *work);
@@ -99,11 +101,15 @@ int amdgpu_vcn_early_init(struct amdgpu_device *adev)
amdgpu_ucode_ip_version_decode(adev, UVD_HWIP, ucode_prefix, sizeof(ucode_prefix));
for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
if (i == 1 && amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 6))
- r = amdgpu_ucode_request(adev, &adev->vcn.fw[i], "amdgpu/%s_%d.bin", ucode_prefix, i);
+ r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_%d.bin", ucode_prefix, i);
else
- r = amdgpu_ucode_request(adev, &adev->vcn.fw[i], "amdgpu/%s.bin", ucode_prefix);
+ r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s.bin", ucode_prefix);
if (r) {
- amdgpu_ucode_release(&adev->vcn.fw[i]);
+ amdgpu_ucode_release(&adev->vcn.inst[i].fw);
return r;
}
}
@@ -151,7 +157,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
adev->vcn.using_unified_queue =
amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0);
- hdr = (const struct common_firmware_header *)adev->vcn.fw[0]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[0].fw->data;
adev->vcn.fw_version = le32_to_cpu(hdr->ucode_version);
/* Bit 20-23, it is encode major and non-zero for new naming convention.
@@ -270,7 +276,7 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)
for (i = 0; i < adev->vcn.num_enc_rings; ++i)
amdgpu_ring_fini(&adev->vcn.inst[j].ring_enc[i]);
- amdgpu_ucode_release(&adev->vcn.fw[j]);
+ amdgpu_ucode_release(&adev->vcn.inst[j].fw);
}
mutex_destroy(&adev->vcn.vcn1_jpeg1_workaround);
@@ -282,7 +288,7 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)
bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, enum vcn_ring_type type, uint32_t vcn_instance)
{
bool ret = false;
- int vcn_config = adev->vcn.vcn_config[vcn_instance];
+ int vcn_config = adev->vcn.inst[vcn_instance].vcn_config;
if ((type == VCN_ENCODE_RING) && (vcn_config & VCN_BLOCK_ENCODE_DISABLE_MASK))
ret = true;
@@ -362,12 +368,12 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev)
const struct common_firmware_header *hdr;
unsigned int offset;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[i]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data;
if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
if (drm_dev_enter(adev_to_drm(adev), &idx)) {
memcpy_toio(adev->vcn.inst[i].cpu_addr,
- adev->vcn.fw[i]->data + offset,
+ adev->vcn.inst[i].fw->data + offset,
le32_to_cpu(hdr->ucode_size_bytes));
drm_dev_exit(idx);
}
@@ -580,7 +586,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring,
if (r)
goto err_free;
- amdgpu_ib_free(adev, ib_msg, f);
+ amdgpu_ib_free(ib_msg, f);
if (fence)
*fence = dma_fence_get(f);
@@ -591,7 +597,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring,
err_free:
amdgpu_job_free(job);
err:
- amdgpu_ib_free(adev, ib_msg, f);
+ amdgpu_ib_free(ib_msg, f);
return r;
}
@@ -773,7 +779,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring,
if (r)
goto err_free;
- amdgpu_ib_free(adev, ib_msg, f);
+ amdgpu_ib_free(ib_msg, f);
if (fence)
*fence = dma_fence_get(f);
@@ -784,7 +790,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring,
err_free:
amdgpu_job_free(job);
err:
- amdgpu_ib_free(adev, ib_msg, f);
+ amdgpu_ib_free(ib_msg, f);
return r;
}
@@ -1014,7 +1020,7 @@ int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = 0;
error:
- amdgpu_ib_free(adev, &ib, fence);
+ amdgpu_ib_free(&ib, fence);
dma_fence_put(fence);
return r;
@@ -1025,7 +1031,8 @@ int amdgpu_vcn_unified_ring_test_ib(struct amdgpu_ring *ring, long timeout)
struct amdgpu_device *adev = ring->adev;
long r;
- if (amdgpu_ip_version(adev, UVD_HWIP, 0) != IP_VERSION(4, 0, 3)) {
+ if ((amdgpu_ip_version(adev, UVD_HWIP, 0) != IP_VERSION(4, 0, 3)) &&
+ (amdgpu_ip_version(adev, UVD_HWIP, 0) != IP_VERSION(5, 0, 1))) {
r = amdgpu_vcn_enc_ring_test_ib(ring, timeout);
if (r)
goto error;
@@ -1063,7 +1070,7 @@ void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev)
if (adev->vcn.harvest_config & (1 << i))
continue;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[i]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data;
/* currently only support 2 FW instances */
if (i >= 2) {
dev_info(adev->dev, "More then 2 VCN FW instances!\n");
@@ -1071,12 +1078,14 @@ void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev)
}
idx = AMDGPU_UCODE_ID_VCN + i;
adev->firmware.ucode[idx].ucode_id = idx;
- adev->firmware.ucode[idx].fw = adev->vcn.fw[i];
+ adev->firmware.ucode[idx].fw = adev->vcn.inst[i].fw;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE);
if (amdgpu_ip_version(adev, UVD_HWIP, 0) ==
- IP_VERSION(4, 0, 3))
+ IP_VERSION(4, 0, 3) ||
+ amdgpu_ip_version(adev, UVD_HWIP, 0) ==
+ IP_VERSION(5, 0, 1))
break;
}
}
@@ -1320,3 +1329,71 @@ void amdgpu_vcn_sysfs_reset_mask_fini(struct amdgpu_device *adev)
device_remove_file(adev->dev, &dev_attr_vcn_reset_mask);
}
}
+
+/*
+ * debugfs to enable/disable vcn job submission to specific core or
+ * instance. It is created only if the queue type is unified.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_debugfs_vcn_sched_mask_set(void *data, u64 val)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)data;
+ u32 i;
+ u64 mask;
+ struct amdgpu_ring *ring;
+
+ if (!adev)
+ return -ENODEV;
+
+ mask = (1ULL << adev->vcn.num_vcn_inst) - 1;
+ if ((val & mask) == 0)
+ return -EINVAL;
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ ring = &adev->vcn.inst[i].ring_enc[0];
+ if (val & (1ULL << i))
+ ring->sched.ready = true;
+ else
+ ring->sched.ready = false;
+ }
+ /* publish sched.ready flag update effective immediately across smp */
+ smp_rmb();
+ return 0;
+}
+
+static int amdgpu_debugfs_vcn_sched_mask_get(void *data, u64 *val)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)data;
+ u32 i;
+ u64 mask = 0;
+ struct amdgpu_ring *ring;
+
+ if (!adev)
+ return -ENODEV;
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ ring = &adev->vcn.inst[i].ring_enc[0];
+ if (ring->sched.ready)
+ mask |= 1ULL << i;
+ }
+ *val = mask;
+ return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_vcn_sched_mask_fops,
+ amdgpu_debugfs_vcn_sched_mask_get,
+ amdgpu_debugfs_vcn_sched_mask_set, "%llx\n");
+#endif
+
+void amdgpu_debugfs_vcn_sched_mask_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+ struct drm_minor *minor = adev_to_drm(adev)->primary;
+ struct dentry *root = minor->debugfs_root;
+ char name[32];
+
+ if (adev->vcn.num_vcn_inst <= 1 || !adev->vcn.using_unified_queue)
+ return;
+ sprintf(name, "amdgpu_vcn_sched_mask");
+ debugfs_create_file(name, 0600, root, adev,
+ &amdgpu_debugfs_vcn_sched_mask_fops);
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 1e32311c1dff..adaf4388ad28 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -1,5 +1,5 @@
/*
- * Copyright 2016 Advanced Micro Devices, Inc.
+ * Copyright 2016-2024 Advanced Micro Devices, Inc. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -163,20 +163,30 @@
#define SOC24_DPG_MODE_OFFSET(ip, inst_idx, reg) \
({ \
uint32_t internal_reg_offset, addr; \
- bool video_range, aon_range; \
+ bool video_range, video1_range, aon_range, aon1_range; \
\
addr = (adev->reg_offset[ip##_HWIP][inst_idx][reg##_BASE_IDX] + reg); \
addr <<= 2; \
video_range = ((((0xFFFFF & addr) >= (VCN_VID_SOC_ADDRESS)) && \
((0xFFFFF & addr) < ((VCN_VID_SOC_ADDRESS + 0x2600))))); \
+ video1_range = ((((0xFFFFF & addr) >= (VCN1_VID_SOC_ADDRESS)) && \
+ ((0xFFFFF & addr) < ((VCN1_VID_SOC_ADDRESS + 0x2600))))); \
aon_range = ((((0xFFFFF & addr) >= (VCN_AON_SOC_ADDRESS)) && \
((0xFFFFF & addr) < ((VCN_AON_SOC_ADDRESS + 0x600))))); \
+ aon1_range = ((((0xFFFFF & addr) >= (VCN1_AON_SOC_ADDRESS)) && \
+ ((0xFFFFF & addr) < ((VCN1_AON_SOC_ADDRESS + 0x600))))); \
if (video_range) \
internal_reg_offset = ((0xFFFFF & addr) - (VCN_VID_SOC_ADDRESS) + \
(VCN_VID_IP_ADDRESS)); \
else if (aon_range) \
internal_reg_offset = ((0xFFFFF & addr) - (VCN_AON_SOC_ADDRESS) + \
(VCN_AON_IP_ADDRESS)); \
+ else if (video1_range) \
+ internal_reg_offset = ((0xFFFFF & addr) - (VCN1_VID_SOC_ADDRESS) + \
+ (VCN_VID_IP_ADDRESS)); \
+ else if (aon1_range) \
+ internal_reg_offset = ((0xFFFFF & addr) - (VCN1_AON_SOC_ADDRESS) + \
+ (VCN_AON_IP_ADDRESS)); \
else \
internal_reg_offset = (0xFFFFF & addr); \
\
@@ -297,6 +307,9 @@ struct amdgpu_vcn_inst {
atomic_t dpg_enc_submission_cnt;
struct amdgpu_vcn_fw_shared fw_shared;
uint8_t aid_id;
+ const struct firmware *fw; /* VCN firmware */
+ uint8_t vcn_config;
+ uint32_t vcn_codec_disable_mask;
};
struct amdgpu_vcn_ras {
@@ -306,15 +319,12 @@ struct amdgpu_vcn_ras {
struct amdgpu_vcn {
unsigned fw_version;
struct delayed_work idle_work;
- const struct firmware *fw[AMDGPU_MAX_VCN_INSTANCES]; /* VCN firmware */
unsigned num_enc_rings;
enum amd_powergating_state cur_state;
bool indirect_sram;
uint8_t num_vcn_inst;
struct amdgpu_vcn_inst inst[AMDGPU_MAX_VCN_INSTANCES];
- uint8_t vcn_config[AMDGPU_MAX_VCN_INSTANCES];
- uint32_t vcn_codec_disable_mask[AMDGPU_MAX_VCN_INSTANCES];
struct amdgpu_vcn_reg internal;
struct mutex vcn_pg_lock;
struct mutex vcn1_jpeg1_workaround;
@@ -523,5 +533,6 @@ int amdgpu_vcn_psp_update_sram(struct amdgpu_device *adev, int inst_idx,
int amdgpu_vcn_save_vcpu_bo(struct amdgpu_device *adev);
int amdgpu_vcn_sysfs_reset_mask_init(struct amdgpu_device *adev);
void amdgpu_vcn_sysfs_reset_mask_fini(struct amdgpu_device *adev);
+void amdgpu_debugfs_vcn_sched_mask_init(struct amdgpu_device *adev);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index c704e9803e11..0af469ec6fcc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -1263,12 +1263,10 @@ static int amdgpu_virt_cache_host_error_counts(struct amdgpu_device *adev,
if (used_size > (AMD_SRIOV_RAS_TELEMETRY_SIZE_KB << 10))
return 0;
- tmp = kmalloc(used_size, GFP_KERNEL);
+ tmp = kmemdup(&host_telemetry->body.error_count, used_size, GFP_KERNEL);
if (!tmp)
return -ENOMEM;
- memcpy(tmp, &host_telemetry->body.error_count, used_size);
-
if (checksum != amd_sriov_msg_checksum(tmp, used_size, 0, 0))
goto out;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
index 8bf28d336807..7507d9443028 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
@@ -188,8 +188,8 @@ static int amdgpu_vkms_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
amdgpu_crtc->connector = NULL;
amdgpu_crtc->vsync_timer_enabled = AMDGPU_IRQ_STATE_DISABLE;
- hrtimer_init(&amdgpu_crtc->vblank_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
- amdgpu_crtc->vblank_timer.function = &amdgpu_vkms_vblank_simulate;
+ hrtimer_setup(&amdgpu_crtc->vblank_timer, &amdgpu_vkms_vblank_simulate, CLOCK_MONOTONIC,
+ HRTIMER_MODE_REL);
return ret;
}
@@ -632,13 +632,13 @@ static bool amdgpu_vkms_is_idle(void *handle)
return true;
}
-static int amdgpu_vkms_set_clockgating_state(void *handle,
+static int amdgpu_vkms_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int amdgpu_vkms_set_powergating_state(void *handle,
+static int amdgpu_vkms_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index c9c48b782ec1..5c07777d3239 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -36,6 +36,7 @@
#include <drm/ttm/ttm_tt.h>
#include <drm/drm_exec.h>
#include "amdgpu.h"
+#include "amdgpu_vm.h"
#include "amdgpu_trace.h"
#include "amdgpu_amdkfd.h"
#include "amdgpu_gmc.h"
@@ -311,6 +312,111 @@ static void amdgpu_vm_bo_reset_state_machine(struct amdgpu_vm *vm)
}
/**
+ * amdgpu_vm_update_shared - helper to update shared memory stat
+ * @base: base structure for tracking BO usage in a VM
+ *
+ * Takes the vm status_lock and updates the shared memory stat. If the basic
+ * stat changed (e.g. buffer was moved) amdgpu_vm_update_stats need to be called
+ * as well.
+ */
+static void amdgpu_vm_update_shared(struct amdgpu_vm_bo_base *base)
+{
+ struct amdgpu_vm *vm = base->vm;
+ struct amdgpu_bo *bo = base->bo;
+ uint64_t size = amdgpu_bo_size(bo);
+ uint32_t bo_memtype = amdgpu_bo_mem_stats_placement(bo);
+ bool shared;
+
+ spin_lock(&vm->status_lock);
+ shared = drm_gem_object_is_shared_for_memory_stats(&bo->tbo.base);
+ if (base->shared != shared) {
+ base->shared = shared;
+ if (shared) {
+ vm->stats[bo_memtype].drm.shared += size;
+ vm->stats[bo_memtype].drm.private -= size;
+ } else {
+ vm->stats[bo_memtype].drm.shared -= size;
+ vm->stats[bo_memtype].drm.private += size;
+ }
+ }
+ spin_unlock(&vm->status_lock);
+}
+
+/**
+ * amdgpu_vm_bo_update_shared - callback when bo gets shared/unshared
+ * @bo: amdgpu buffer object
+ *
+ * Update the per VM stats for all the vm if needed from private to shared or
+ * vice versa.
+ */
+void amdgpu_vm_bo_update_shared(struct amdgpu_bo *bo)
+{
+ struct amdgpu_vm_bo_base *base;
+
+ for (base = bo->vm_bo; base; base = base->next)
+ amdgpu_vm_update_shared(base);
+}
+
+/**
+ * amdgpu_vm_update_stats_locked - helper to update normal memory stat
+ * @base: base structure for tracking BO usage in a VM
+ * @res: the ttm_resource to use for the purpose of accounting, may or may not
+ * be bo->tbo.resource
+ * @sign: if we should add (+1) or subtract (-1) from the stat
+ *
+ * Caller need to have the vm status_lock held. Useful for when multiple update
+ * need to happen at the same time.
+ */
+static void amdgpu_vm_update_stats_locked(struct amdgpu_vm_bo_base *base,
+ struct ttm_resource *res, int sign)
+{
+ struct amdgpu_vm *vm = base->vm;
+ struct amdgpu_bo *bo = base->bo;
+ int64_t size = sign * amdgpu_bo_size(bo);
+ uint32_t bo_memtype = amdgpu_bo_mem_stats_placement(bo);
+
+ /* For drm-total- and drm-shared-, BO are accounted by their preferred
+ * placement, see also amdgpu_bo_mem_stats_placement.
+ */
+ if (base->shared)
+ vm->stats[bo_memtype].drm.shared += size;
+ else
+ vm->stats[bo_memtype].drm.private += size;
+
+ if (res && res->mem_type < __AMDGPU_PL_NUM) {
+ uint32_t res_memtype = res->mem_type;
+
+ vm->stats[res_memtype].drm.resident += size;
+ /* BO only count as purgeable if it is resident,
+ * since otherwise there's nothing to purge.
+ */
+ if (bo->flags & AMDGPU_GEM_CREATE_DISCARDABLE)
+ vm->stats[res_memtype].drm.purgeable += size;
+ if (!(bo->preferred_domains & amdgpu_mem_type_to_domain(res_memtype)))
+ vm->stats[bo_memtype].evicted += size;
+ }
+}
+
+/**
+ * amdgpu_vm_update_stats - helper to update normal memory stat
+ * @base: base structure for tracking BO usage in a VM
+ * @res: the ttm_resource to use for the purpose of accounting, may or may not
+ * be bo->tbo.resource
+ * @sign: if we should add (+1) or subtract (-1) from the stat
+ *
+ * Updates the basic memory stat when bo is added/deleted/moved.
+ */
+void amdgpu_vm_update_stats(struct amdgpu_vm_bo_base *base,
+ struct ttm_resource *res, int sign)
+{
+ struct amdgpu_vm *vm = base->vm;
+
+ spin_lock(&vm->status_lock);
+ amdgpu_vm_update_stats_locked(base, res, sign);
+ spin_unlock(&vm->status_lock);
+}
+
+/**
* amdgpu_vm_bo_base_init - Adds bo to the list of bos associated with the vm
*
* @base: base structure for tracking BO usage in a VM
@@ -333,6 +439,11 @@ void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base,
base->next = bo->vm_bo;
bo->vm_bo = base;
+ spin_lock(&vm->status_lock);
+ base->shared = drm_gem_object_is_shared_for_memory_stats(&bo->tbo.base);
+ amdgpu_vm_update_stats_locked(base, bo->tbo.resource, +1);
+ spin_unlock(&vm->status_lock);
+
if (!amdgpu_vm_is_bo_always_valid(vm, bo))
return;
@@ -1083,53 +1194,11 @@ error_free:
return r;
}
-static void amdgpu_vm_bo_get_memory(struct amdgpu_bo_va *bo_va,
- struct amdgpu_mem_stats *stats,
- unsigned int size)
-{
- struct amdgpu_vm *vm = bo_va->base.vm;
- struct amdgpu_bo *bo = bo_va->base.bo;
-
- if (!bo)
- return;
-
- /*
- * For now ignore BOs which are currently locked and potentially
- * changing their location.
- */
- if (!amdgpu_vm_is_bo_always_valid(vm, bo) &&
- !dma_resv_trylock(bo->tbo.base.resv))
- return;
-
- amdgpu_bo_get_memory(bo, stats, size);
- if (!amdgpu_vm_is_bo_always_valid(vm, bo))
- dma_resv_unlock(bo->tbo.base.resv);
-}
-
void amdgpu_vm_get_memory(struct amdgpu_vm *vm,
- struct amdgpu_mem_stats *stats,
- unsigned int size)
+ struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM])
{
- struct amdgpu_bo_va *bo_va, *tmp;
-
spin_lock(&vm->status_lock);
- list_for_each_entry_safe(bo_va, tmp, &vm->idle, base.vm_status)
- amdgpu_vm_bo_get_memory(bo_va, stats, size);
-
- list_for_each_entry_safe(bo_va, tmp, &vm->evicted, base.vm_status)
- amdgpu_vm_bo_get_memory(bo_va, stats, size);
-
- list_for_each_entry_safe(bo_va, tmp, &vm->relocated, base.vm_status)
- amdgpu_vm_bo_get_memory(bo_va, stats, size);
-
- list_for_each_entry_safe(bo_va, tmp, &vm->moved, base.vm_status)
- amdgpu_vm_bo_get_memory(bo_va, stats, size);
-
- list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, base.vm_status)
- amdgpu_vm_bo_get_memory(bo_va, stats, size);
-
- list_for_each_entry_safe(bo_va, tmp, &vm->done, base.vm_status)
- amdgpu_vm_bo_get_memory(bo_va, stats, size);
+ memcpy(stats, vm->stats, sizeof(*stats) * __AMDGPU_PL_NUM);
spin_unlock(&vm->status_lock);
}
@@ -2075,6 +2144,7 @@ void amdgpu_vm_bo_del(struct amdgpu_device *adev,
if (*base != &bo_va->base)
continue;
+ amdgpu_vm_update_stats(*base, bo->tbo.resource, -1);
*base = bo_va->base.next;
break;
}
@@ -2143,14 +2213,12 @@ bool amdgpu_vm_evictable(struct amdgpu_bo *bo)
/**
* amdgpu_vm_bo_invalidate - mark the bo as invalid
*
- * @adev: amdgpu_device pointer
* @bo: amdgpu buffer object
* @evicted: is the BO evicted
*
* Mark @bo as invalid.
*/
-void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
- struct amdgpu_bo *bo, bool evicted)
+void amdgpu_vm_bo_invalidate(struct amdgpu_bo *bo, bool evicted)
{
struct amdgpu_vm_bo_base *bo_base;
@@ -2176,6 +2244,32 @@ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
}
/**
+ * amdgpu_vm_bo_move - handle BO move
+ *
+ * @bo: amdgpu buffer object
+ * @new_mem: the new placement of the BO move
+ * @evicted: is the BO evicted
+ *
+ * Update the memory stats for the new placement and mark @bo as invalid.
+ */
+void amdgpu_vm_bo_move(struct amdgpu_bo *bo, struct ttm_resource *new_mem,
+ bool evicted)
+{
+ struct amdgpu_vm_bo_base *bo_base;
+
+ for (bo_base = bo->vm_bo; bo_base; bo_base = bo_base->next) {
+ struct amdgpu_vm *vm = bo_base->vm;
+
+ spin_lock(&vm->status_lock);
+ amdgpu_vm_update_stats_locked(bo_base, bo->tbo.resource, -1);
+ amdgpu_vm_update_stats_locked(bo_base, new_mem, +1);
+ spin_unlock(&vm->status_lock);
+ }
+
+ amdgpu_vm_bo_invalidate(bo, evicted);
+}
+
+/**
* amdgpu_vm_get_block_size - calculate VM page table size as power of two
*
* @vm_size: VM size
@@ -2594,6 +2688,16 @@ void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
vm->is_compute_context = false;
}
+static int amdgpu_vm_stats_is_zero(struct amdgpu_vm *vm)
+{
+ for (int i = 0; i < __AMDGPU_PL_NUM; ++i) {
+ if (!(drm_memory_stats_is_zero(&vm->stats[i].drm) &&
+ vm->stats[i].evicted == 0))
+ return false;
+ }
+ return true;
+}
+
/**
* amdgpu_vm_fini - tear down a vm instance
*
@@ -2617,7 +2721,6 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
root = amdgpu_bo_ref(vm->root.bo);
amdgpu_bo_reserve(root, true);
- amdgpu_vm_put_task_info(vm->task_info);
amdgpu_vm_set_pasid(adev, vm, 0);
dma_fence_wait(vm->last_unlocked, false);
dma_fence_put(vm->last_unlocked);
@@ -2666,6 +2769,16 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
}
ttm_lru_bulk_move_fini(&adev->mman.bdev, &vm->lru_bulk_move);
+
+ if (!amdgpu_vm_stats_is_zero(vm)) {
+ struct amdgpu_task_info *ti = vm->task_info;
+
+ dev_warn(adev->dev,
+ "VM memory stats for proc %s(%d) task %s(%d) is non-zero when fini\n",
+ ti->process_name, ti->pid, ti->task_name, ti->tgid);
+ }
+
+ amdgpu_vm_put_task_info(vm->task_info);
}
/**
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 5d119ac26c4f..a3e128e373bc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -35,6 +35,7 @@
#include "amdgpu_sync.h"
#include "amdgpu_ring.h"
#include "amdgpu_ids.h"
+#include "amdgpu_ttm.h"
struct drm_exec;
@@ -202,9 +203,13 @@ struct amdgpu_vm_bo_base {
/* protected by bo being reserved */
struct amdgpu_vm_bo_base *next;
- /* protected by spinlock */
+ /* protected by vm status_lock */
struct list_head vm_status;
+ /* if the bo is counted as shared in mem stats
+ * protected by vm status_lock */
+ bool shared;
+
/* protected by the BO being reserved */
bool moved;
};
@@ -324,10 +329,7 @@ struct amdgpu_vm_fault_info {
struct amdgpu_mem_stats {
struct drm_memory_stats drm;
- /* buffers that requested this placement */
- uint64_t requested;
- /* buffers that requested this placement
- * but are currently evicted */
+ /* buffers that requested this placement but are currently evicted */
uint64_t evicted;
};
@@ -345,6 +347,9 @@ struct amdgpu_vm {
/* Lock to protect vm_bo add/del/move on all lists of vm */
spinlock_t status_lock;
+ /* Memory statistics for this vm, protected by status_lock */
+ struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM];
+
/* Per-VM and PT BOs who needs a validation */
struct list_head evicted;
@@ -524,8 +529,12 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
struct amdgpu_bo_va *bo_va,
bool clear);
bool amdgpu_vm_evictable(struct amdgpu_bo *bo);
-void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
- struct amdgpu_bo *bo, bool evicted);
+void amdgpu_vm_bo_invalidate(struct amdgpu_bo *bo, bool evicted);
+void amdgpu_vm_update_stats(struct amdgpu_vm_bo_base *base,
+ struct ttm_resource *new_res, int sign);
+void amdgpu_vm_bo_update_shared(struct amdgpu_bo *bo);
+void amdgpu_vm_bo_move(struct amdgpu_bo *bo, struct ttm_resource *new_mem,
+ bool evicted);
uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr);
struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
struct amdgpu_bo *bo);
@@ -576,8 +585,7 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm);
void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
struct amdgpu_vm *vm);
void amdgpu_vm_get_memory(struct amdgpu_vm *vm,
- struct amdgpu_mem_stats *stats,
- unsigned int size);
+ struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM]);
int amdgpu_vm_pt_clear(struct amdgpu_device *adev, struct amdgpu_vm *vm,
struct amdgpu_bo_vm *vmbo, bool immediate);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
index f78a0434a48f..b0bf21682115 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
@@ -537,6 +537,7 @@ static void amdgpu_vm_pt_free(struct amdgpu_vm_bo_base *entry)
if (!entry->bo)
return;
+ amdgpu_vm_update_stats(entry, entry->bo->tbo.resource, -1);
entry->bo->vm_bo = NULL;
ttm_bo_set_bulk_move(&entry->bo->tbo, NULL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
index 110b120d7375..121ee17b522b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
@@ -236,7 +236,8 @@ int amdgpu_vpe_init_microcode(struct amdgpu_vpe *vpe)
int ret;
amdgpu_ucode_ip_version_decode(adev, VPE_HWIP, fw_prefix, sizeof(fw_prefix));
- ret = amdgpu_ucode_request(adev, &adev->vpe.fw, "amdgpu/%s.bin", fw_prefix);
+ ret = amdgpu_ucode_request(adev, &adev->vpe.fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s.bin", fw_prefix);
if (ret)
goto out;
@@ -646,16 +647,16 @@ static int vpe_ring_preempt_ib(struct amdgpu_ring *ring)
return r;
}
-static int vpe_set_clockgating_state(void *handle,
+static int vpe_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int vpe_set_powergating_state(void *handle,
+static int vpe_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
struct amdgpu_vpe *vpe = &adev->vpe;
if (!adev->pm.dpm_enabled)
@@ -833,7 +834,7 @@ static int vpe_ring_test_ib(struct amdgpu_ring *ring, long timeout)
ret = (le32_to_cpu(adev->wb.wb[index]) == test_pattern) ? 0 : -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
index e209b5e101df..23b6f7a4aa4a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
@@ -427,7 +427,7 @@ void amdgpu_xcp_release_sched(struct amdgpu_device *adev,
return;
sched = entity->entity.rq->sched;
- if (sched->ready) {
+ if (drm_sched_wqueue_ready(sched)) {
ring = to_amdgpu_ring(entity->entity.rq->sched);
atomic_dec(&adev->xcp_mgr->xcp[ring->xcp_id].ref_cnt);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c
index e2cb1f080e88..08d6787893b3 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik.c
@@ -2161,13 +2161,13 @@ static int cik_common_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int cik_common_set_clockgating_state(void *handle,
+static int cik_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int cik_common_set_powergating_state(void *handle,
+static int cik_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c
index 1da17755ad53..444563486769 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c
@@ -402,13 +402,13 @@ static int cik_ih_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int cik_ih_set_clockgating_state(void *handle,
+static int cik_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int cik_ih_set_powergating_state(void *handle,
+static int cik_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
index ede1a028d48d..d9bd8f3f17e2 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
@@ -133,9 +133,11 @@ static int cik_sdma_init_microcode(struct amdgpu_device *adev)
for (i = 0; i < adev->sdma.num_instances; i++) {
if (i == 0)
err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sdma.bin", chip_name);
else
err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sdma1.bin", chip_name);
if (err)
goto out;
@@ -696,7 +698,7 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
@@ -1189,11 +1191,11 @@ static int cik_sdma_process_illegal_inst_irq(struct amdgpu_device *adev,
return 0;
}
-static int cik_sdma_set_clockgating_state(void *handle,
+static int cik_sdma_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
bool gate = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_CG_STATE_GATE)
gate = true;
@@ -1204,7 +1206,7 @@ static int cik_sdma_set_clockgating_state(void *handle,
return 0;
}
-static int cik_sdma_set_powergating_state(void *handle,
+static int cik_sdma_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c
index d72973bd570d..82586b76aeda 100644
--- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c
@@ -398,14 +398,14 @@ static int cz_ih_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int cz_ih_set_clockgating_state(void *handle,
+static int cz_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
// TODO
return 0;
}
-static int cz_ih_set_powergating_state(void *handle,
+static int cz_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
// TODO
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
index 5098c50d54c8..c5e3d2251b18 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
@@ -2687,6 +2687,32 @@ static const struct drm_crtc_helper_funcs dce_v10_0_crtc_helper_funcs = {
.get_scanout_position = amdgpu_crtc_get_scanout_position,
};
+static void dce_v10_0_panic_flush(struct drm_plane *plane)
+{
+ struct drm_framebuffer *fb;
+ struct amdgpu_crtc *amdgpu_crtc;
+ struct amdgpu_device *adev;
+ uint32_t fb_format;
+
+ if (!plane->fb)
+ return;
+
+ fb = plane->fb;
+ amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+ adev = drm_to_adev(fb->dev);
+
+ /* Disable DC tiling */
+ fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+ fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK;
+ WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+
+}
+
+static const struct drm_plane_helper_funcs dce_v10_0_drm_primary_plane_helper_funcs = {
+ .get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+ .panic_flush = dce_v10_0_panic_flush,
+};
+
static int dce_v10_0_crtc_init(struct amdgpu_device *adev, int index)
{
struct amdgpu_crtc *amdgpu_crtc;
@@ -2734,6 +2760,7 @@ static int dce_v10_0_crtc_init(struct amdgpu_device *adev, int index)
amdgpu_crtc->encoder = NULL;
amdgpu_crtc->connector = NULL;
drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v10_0_crtc_helper_funcs);
+ drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v10_0_drm_primary_plane_helper_funcs);
return 0;
}
@@ -3302,13 +3329,13 @@ static int dce_v10_0_hpd_irq(struct amdgpu_device *adev,
return 0;
}
-static int dce_v10_0_set_clockgating_state(void *handle,
+static int dce_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int dce_v10_0_set_powergating_state(void *handle,
+static int dce_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
index c5680ff4ab9f..ea42a4472bf6 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
@@ -2800,6 +2800,32 @@ static const struct drm_crtc_helper_funcs dce_v11_0_crtc_helper_funcs = {
.get_scanout_position = amdgpu_crtc_get_scanout_position,
};
+static void dce_v11_0_panic_flush(struct drm_plane *plane)
+{
+ struct drm_framebuffer *fb;
+ struct amdgpu_crtc *amdgpu_crtc;
+ struct amdgpu_device *adev;
+ uint32_t fb_format;
+
+ if (!plane->fb)
+ return;
+
+ fb = plane->fb;
+ amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+ adev = drm_to_adev(fb->dev);
+
+ /* Disable DC tiling */
+ fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+ fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK;
+ WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+
+}
+
+static const struct drm_plane_helper_funcs dce_v11_0_drm_primary_plane_helper_funcs = {
+ .get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+ .panic_flush = dce_v11_0_panic_flush,
+};
+
static int dce_v11_0_crtc_init(struct amdgpu_device *adev, int index)
{
struct amdgpu_crtc *amdgpu_crtc;
@@ -2847,6 +2873,7 @@ static int dce_v11_0_crtc_init(struct amdgpu_device *adev, int index)
amdgpu_crtc->encoder = NULL;
amdgpu_crtc->connector = NULL;
drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v11_0_crtc_helper_funcs);
+ drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v11_0_drm_primary_plane_helper_funcs);
return 0;
}
@@ -3434,13 +3461,13 @@ static int dce_v11_0_hpd_irq(struct amdgpu_device *adev,
return 0;
}
-static int dce_v11_0_set_clockgating_state(void *handle,
+static int dce_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int dce_v11_0_set_powergating_state(void *handle,
+static int dce_v11_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
index eb7de9122d99..915804a6a1d7 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
@@ -2602,6 +2602,32 @@ static const struct drm_crtc_helper_funcs dce_v6_0_crtc_helper_funcs = {
.get_scanout_position = amdgpu_crtc_get_scanout_position,
};
+static void dce_v6_0_panic_flush(struct drm_plane *plane)
+{
+ struct drm_framebuffer *fb;
+ struct amdgpu_crtc *amdgpu_crtc;
+ struct amdgpu_device *adev;
+ uint32_t fb_format;
+
+ if (!plane->fb)
+ return;
+
+ fb = plane->fb;
+ amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+ adev = drm_to_adev(fb->dev);
+
+ /* Disable DC tiling */
+ fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+ fb_format &= ~GRPH_ARRAY_MODE(0x7);
+ WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+
+}
+
+static const struct drm_plane_helper_funcs dce_v6_0_drm_primary_plane_helper_funcs = {
+ .get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+ .panic_flush = dce_v6_0_panic_flush,
+};
+
static int dce_v6_0_crtc_init(struct amdgpu_device *adev, int index)
{
struct amdgpu_crtc *amdgpu_crtc;
@@ -2629,6 +2655,7 @@ static int dce_v6_0_crtc_init(struct amdgpu_device *adev, int index)
amdgpu_crtc->encoder = NULL;
amdgpu_crtc->connector = NULL;
drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v6_0_crtc_helper_funcs);
+ drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v6_0_drm_primary_plane_helper_funcs);
return 0;
}
@@ -3124,13 +3151,13 @@ static int dce_v6_0_hpd_irq(struct amdgpu_device *adev,
}
-static int dce_v6_0_set_clockgating_state(void *handle,
+static int dce_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int dce_v6_0_set_powergating_state(void *handle,
+static int dce_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
index 04b79ff87f75..f2edc0fece5b 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
@@ -2613,6 +2613,31 @@ static const struct drm_crtc_helper_funcs dce_v8_0_crtc_helper_funcs = {
.get_scanout_position = amdgpu_crtc_get_scanout_position,
};
+static void dce_v8_0_panic_flush(struct drm_plane *plane)
+{
+ struct drm_framebuffer *fb;
+ struct amdgpu_crtc *amdgpu_crtc;
+ struct amdgpu_device *adev;
+ uint32_t fb_format;
+
+ if (!plane->fb)
+ return;
+
+ fb = plane->fb;
+ amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+ adev = drm_to_adev(fb->dev);
+
+ /* Disable DC tiling */
+ fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+ fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK;
+ WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+}
+
+static const struct drm_plane_helper_funcs dce_v8_0_drm_primary_plane_helper_funcs = {
+ .get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+ .panic_flush = dce_v8_0_panic_flush,
+};
+
static int dce_v8_0_crtc_init(struct amdgpu_device *adev, int index)
{
struct amdgpu_crtc *amdgpu_crtc;
@@ -2640,6 +2665,7 @@ static int dce_v8_0_crtc_init(struct amdgpu_device *adev, int index)
amdgpu_crtc->encoder = NULL;
amdgpu_crtc->connector = NULL;
drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v8_0_crtc_helper_funcs);
+ drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v8_0_drm_primary_plane_helper_funcs);
return 0;
}
@@ -3212,13 +3238,13 @@ static int dce_v8_0_hpd_irq(struct amdgpu_device *adev,
}
-static int dce_v8_0_set_clockgating_state(void *handle,
+static int dce_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int dce_v8_0_set_powergating_state(void *handle,
+static int dce_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 24dce803a829..5ba263fe5512 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -45,6 +45,7 @@
#include "clearstate_gfx10.h"
#include "v10_structs.h"
#include "gfx_v10_0.h"
+#include "gfx_v10_0_cleaner_shader.h"
#include "nbio_v2_3.h"
/*
@@ -3673,7 +3674,7 @@ static void gfx_v10_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
static void gfx_v10_0_update_spm_vmid_internal(struct amdgpu_device *adev,
unsigned int vmid);
-static int gfx_v10_0_set_powergating_state(void *handle,
+static int gfx_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static void gfx10_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue_mask)
{
@@ -4036,7 +4037,7 @@ static int gfx_v10_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
else
r = -EINVAL;
err2:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
amdgpu_device_wb_free(adev, index);
@@ -4138,18 +4139,21 @@ static int gfx_v10_0_init_microcode(struct amdgpu_device *adev)
amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp%s.bin", ucode_prefix, wks);
if (err)
goto out;
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_PFP);
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me%s.bin", ucode_prefix, wks);
if (err)
goto out;
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_ME);
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_ce%s.bin", ucode_prefix, wks);
if (err)
goto out;
@@ -4173,6 +4177,7 @@ static int gfx_v10_0_init_microcode(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec%s.bin", ucode_prefix, wks);
if (err)
goto out;
@@ -4180,6 +4185,7 @@ static int gfx_v10_0_init_microcode(struct amdgpu_device *adev)
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1_JT);
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec2%s.bin", ucode_prefix, wks);
if (!err) {
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2);
@@ -4733,6 +4739,23 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block)
break;
}
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+ case IP_VERSION(10, 3, 0):
+ case IP_VERSION(10, 3, 2):
+ case IP_VERSION(10, 3, 4):
+ case IP_VERSION(10, 3, 5):
+ 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 >= 64 &&
+ adev->gfx.pfp_fw_version >= 100 &&
+ adev->gfx.mec_fw_version >= 122) {
+ 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;
@@ -5952,7 +5975,7 @@ static int gfx_v10_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
else
WREG32_SOC15(GC, 0, mmCP_ME_CNTL, tmp);
- if (adev->job_hang && !enable)
+ if (amdgpu_in_reset(adev) && !enable)
return 0;
for (i = 0; i < adev->usec_timeout; i++) {
@@ -6599,17 +6622,13 @@ static void gfx_v10_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp);
- tmp |= 0x80;
- WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp);
+ WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp | 0x80);
break;
default:
tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp | 0x80);
break;
}
}
@@ -7457,7 +7476,7 @@ static int gfx_v10_0_hw_fini(struct amdgpu_ip_block *ip_block)
* otherwise the gfxoff disallowing will be failed to set.
*/
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 3, 1))
- gfx_v10_0_set_powergating_state(ip_block->adev, AMD_PG_STATE_UNGATE);
+ gfx_v10_0_set_powergating_state(ip_block, AMD_PG_STATE_UNGATE);
if (!adev->no_hw_access) {
if (amdgpu_async_gfx_ring) {
@@ -8345,10 +8364,10 @@ static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs_sriov = {
.is_rlcg_access_range = gfx_v10_0_is_rlcg_access_range,
};
-static int gfx_v10_0_set_powergating_state(void *handle,
+static int gfx_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
@@ -8383,10 +8402,10 @@ static int gfx_v10_0_set_powergating_state(void *handle,
return 0;
}
-static int gfx_v10_0_set_clockgating_state(void *handle,
+static int gfx_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
new file mode 100644
index 000000000000..663c2572d440
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
@@ -0,0 +1,56 @@
+/* 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.
+ */
+
+/* Define the cleaner shader gfx_10_3_0 */
+static const u32 gfx_10_3_0_cleaner_shader_hex[] = {
+ 0xb0804004, 0xbf8a0000,
+ 0xbe8203b8, 0xbefc0380,
+ 0x7e008480, 0x7e028480,
+ 0x7e048480, 0x7e068480,
+ 0x7e088480, 0x7e0a8480,
+ 0x7e0c8480, 0x7e0e8480,
+ 0xbefc0302, 0x80828802,
+ 0xbf84fff5, 0xbe8203ff,
+ 0x80000000, 0x87020002,
+ 0xbf840012, 0xbefe03c1,
+ 0xbeff03c1, 0xd7650001,
+ 0x0001007f, 0xd7660001,
+ 0x0002027e, 0x16020288,
+ 0xbe8203bf, 0xbefc03c1,
+ 0xd9382000, 0x00020201,
+ 0xd9386040, 0x00040401,
+ 0xd70f6a01, 0x000202ff,
+ 0x00000400, 0x80828102,
+ 0xbf84fff7, 0xbefc03ff,
+ 0x00000068, 0xbe803080,
+ 0xbe813080, 0xbe823080,
+ 0xbe833080, 0x80fc847c,
+ 0xbf84fffa, 0xbeea0480,
+ 0xbeec0480, 0xbeee0480,
+ 0xbef00480, 0xbef20480,
+ 0xbef40480, 0xbef60480,
+ 0xbef80480, 0xbefa0480,
+ 0xbf810000, 0xbf9f0000,
+ 0xbf9f0000, 0xbf9f0000,
+ 0xbf9f0000, 0xbf9f0000,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm
new file mode 100644
index 000000000000..0e1c246166c0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm
@@ -0,0 +1,124 @@
+/* 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.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is first 64 Dwords or 256 bytes of 192 Dwords cleaner shader.
+//To turn this shader program on for complitaion change this to main and lower shader main to main_1
+
+// GFX10.3 : Clear SGPRs, VGPRs and LDS
+// Launch 32 waves per CU (16 per SIMD) as a workgroup (threadgroup) to fill every wave slot
+// Waves are "wave32" and have 64 VGPRs each, which uses all 1024 VGPRs per SIMD
+// Waves are launched in "CU" mode, and the workgroup shares 64KB of LDS (half of the WGP's LDS)
+// It takes 2 workgroups to use all of LDS: one on each CU of the WGP
+// Each wave clears SGPRs 0 - 107
+// Each wave clears VGPRs 0 - 63
+// The first wave of the workgroup clears its 64KB of LDS
+// The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+// before any wave in the workgroup could end. Without this, it is possible not all SGPRs get cleared.
+
+
+shader main
+ asic(GFX10)
+ type(CS)
+ wave_size(32)
+// Note: original source code from SQ team
+
+//
+// Create 32 waves in a threadgroup (CS waves)
+// Each allocates 64 VGPRs
+// The workgroup allocates all of LDS (64kbytes)
+//
+// Takes about 2500 clocks to run.
+// (theorhetical fastest = 1024clks vgpr + 640lds = 1660 clks)
+//
+ S_BARRIER
+ s_mov_b32 s2, 0x00000038 // Loop 64/8=8 times (loop unrolled for performance)
+ s_mov_b32 m0, 0
+ //
+ // CLEAR VGPRs
+ //
+label_0005:
+ v_movreld_b32 v0, 0
+ v_movreld_b32 v1, 0
+ v_movreld_b32 v2, 0
+ v_movreld_b32 v3, 0
+ v_movreld_b32 v4, 0
+ v_movreld_b32 v5, 0
+ v_movreld_b32 v6, 0
+ v_movreld_b32 v7, 0
+ s_mov_b32 m0, s2
+ s_sub_u32 s2, s2, 8
+ s_cbranch_scc0 label_0005
+ //
+ s_mov_b32 s2, 0x80000000 // Bit31 is first_wave
+ s_and_b32 s2, s2, s0 // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+ s_cbranch_scc0 label_0023 // Clean LDS if its first wave of ThreadGroup/WorkGroup
+ // CLEAR LDS
+ //
+ s_mov_b32 exec_lo, 0xffffffff
+ s_mov_b32 exec_hi, 0xffffffff
+ v_mbcnt_lo_u32_b32 v1, exec_hi, 0 // Set V1 to thread-ID (0..63)
+ v_mbcnt_hi_u32_b32 v1, exec_lo, v1 // Set V1 to thread-ID (0..63)
+ v_mul_u32_u24 v1, 0x00000008, v1 // * 8, so each thread is a double-dword address (8byte)
+ s_mov_b32 s2, 0x00000003f // 64 loop iterations
+ s_mov_b32 m0, 0xffffffff
+ // Clear all of LDS space
+ // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+ ds_write2_b64 v1, v[2:3], v[2:3] offset1:32
+ ds_write2_b64 v1, v[4:5], v[4:5] offset0:64 offset1:96
+ v_add_co_u32 v1, vcc, 0x00000400, v1
+ s_sub_u32 s2, s2, 1
+ s_cbranch_scc0 label_001F
+
+ //
+ // CLEAR SGPRs
+ //
+label_0023:
+ s_mov_b32 m0, 0x00000068 // Loop 108/4=27 times (loop unrolled for performance)
+label_sgpr_loop:
+ s_movreld_b32 s0, 0
+ s_movreld_b32 s1, 0
+ s_movreld_b32 s2, 0
+ s_movreld_b32 s3, 0
+ s_sub_u32 m0, m0, 4
+ s_cbranch_scc0 label_sgpr_loop
+
+ //clear vcc
+ s_mov_b32 flat_scratch_lo, 0 //clear flat scratch lo SGPR
+ s_mov_b32 flat_scratch_hi, 0 //clear flat scratch hi SGPR
+ s_mov_b64 vcc, 0 //clear vcc
+ s_mov_b64 ttmp0, 0 //Clear ttmp0 and ttmp1
+ s_mov_b64 ttmp2, 0 //Clear ttmp2 and ttmp3
+ s_mov_b64 ttmp4, 0 //Clear ttmp4 and ttmp5
+ s_mov_b64 ttmp6, 0 //Clear ttmp6 and ttmp7
+ s_mov_b64 ttmp8, 0 //Clear ttmp8 and ttmp9
+ s_mov_b64 ttmp10, 0 //Clear ttmp10 and ttmp11
+ s_mov_b64 ttmp12, 0 //Clear ttmp12 and ttmp13
+ s_mov_b64 ttmp14, 0 //Clear ttmp14 and ttmp15
+
+ s_endpgm
+
+end
+
+
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index 2ae058a224f4..56c06b72a70a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -615,7 +615,7 @@ static int gfx_v11_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err2:
if (!ring->is_mes_queue)
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
if (!ring->is_mes_queue)
@@ -639,6 +639,7 @@ static int gfx_v11_0_init_toc_microcode(struct amdgpu_device *adev, const char *
int err = 0;
err = amdgpu_ucode_request(adev, &adev->psp.toc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_toc.bin", ucode_prefix);
if (err)
goto out;
@@ -688,6 +689,7 @@ static int gfx_v11_0_init_microcode(struct amdgpu_device *adev)
amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", ucode_prefix);
if (err)
goto out;
@@ -705,6 +707,7 @@ static int gfx_v11_0_init_microcode(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", ucode_prefix);
if (err)
goto out;
@@ -720,9 +723,11 @@ static int gfx_v11_0_init_microcode(struct amdgpu_device *adev)
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 0) &&
adev->pdev->revision == 0xCE)
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/gc_11_0_0_rlc_1.bin");
else
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", ucode_prefix);
if (err)
goto out;
@@ -735,6 +740,7 @@ static int gfx_v11_0_init_microcode(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec.bin", ucode_prefix);
if (err)
goto out;
@@ -1885,6 +1891,7 @@ static u32 gfx_v11_0_get_rb_active_bitmap(struct amdgpu_device *adev)
static void gfx_v11_0_setup_rb(struct amdgpu_device *adev)
{
+ u32 rb_bitmap_per_sa;
u32 rb_bitmap_width_per_sa;
u32 max_sa;
u32 active_sa_bitmap;
@@ -1902,9 +1909,11 @@ static void gfx_v11_0_setup_rb(struct amdgpu_device *adev)
adev->gfx.config.max_sh_per_se;
rb_bitmap_width_per_sa = adev->gfx.config.max_backends_per_se /
adev->gfx.config.max_sh_per_se;
+ rb_bitmap_per_sa = amdgpu_gfx_create_bitmask(rb_bitmap_width_per_sa);
+
for (i = 0; i < max_sa; i++) {
if (active_sa_bitmap & (1 << i))
- active_rb_bitmap |= (0x3 << (i * rb_bitmap_width_per_sa));
+ active_rb_bitmap |= (rb_bitmap_per_sa << (i * rb_bitmap_width_per_sa));
}
active_rb_bitmap &= global_active_rb_bitmap;
@@ -3918,9 +3927,7 @@ static void gfx_v11_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
}
static void gfx_v11_0_cp_set_doorbell_range(struct amdgpu_device *adev)
@@ -5458,10 +5465,10 @@ static void gfx_v11_cntl_pg(struct amdgpu_device *adev, bool enable)
amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
-static int gfx_v11_0_set_powergating_state(void *handle,
+static int gfx_v11_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
@@ -5494,10 +5501,10 @@ static int gfx_v11_0_set_powergating_state(void *handle,
return 0;
}
-static int gfx_v11_0_set_clockgating_state(void *handle,
+static int gfx_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -6646,30 +6653,14 @@ static int gfx_v11_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid)
static int gfx_v11_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid)
{
struct amdgpu_device *adev = ring->adev;
- int i, r = 0;
+ int r = 0;
if (amdgpu_sriov_vf(adev))
return -EINVAL;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
- mutex_lock(&adev->srbm_mutex);
- soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
- WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2);
- WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1);
-
- /* make sure dequeue is complete*/
- for (i = 0; i < adev->usec_timeout; i++) {
- if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
- break;
- udelay(1);
- }
- if (i >= adev->usec_timeout)
- r = -ETIMEDOUT;
- soc21_grbm_select(adev, 0, 0, 0, 0);
- mutex_unlock(&adev->srbm_mutex);
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+ r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, true);
if (r) {
- dev_err(adev->dev, "fail to wait on hqd deactivate\n");
+ dev_err(adev->dev, "reset via MMIO failed %d\n", r);
return r;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
index da327ab48a57..48ff00427882 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -513,7 +513,7 @@ static int gfx_v12_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err2:
if (!ring->is_mes_queue)
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
if (!ring->is_mes_queue)
@@ -537,6 +537,7 @@ static int gfx_v12_0_init_toc_microcode(struct amdgpu_device *adev, const char *
int err = 0;
err = amdgpu_ucode_request(adev, &adev->psp.toc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_toc.bin", ucode_prefix);
if (err)
goto out;
@@ -566,6 +567,7 @@ static int gfx_v12_0_init_microcode(struct amdgpu_device *adev)
amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", ucode_prefix);
if (err)
goto out;
@@ -573,6 +575,7 @@ static int gfx_v12_0_init_microcode(struct amdgpu_device *adev)
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK);
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", ucode_prefix);
if (err)
goto out;
@@ -581,6 +584,7 @@ static int gfx_v12_0_init_microcode(struct amdgpu_device *adev)
if (!amdgpu_sriov_vf(adev)) {
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", ucode_prefix);
if (err)
goto out;
@@ -593,6 +597,7 @@ static int gfx_v12_0_init_microcode(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec.bin", ucode_prefix);
if (err)
goto out;
@@ -1347,6 +1352,14 @@ 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.me_fw_version >= 2480 &&
+ adev->gfx.pfp_fw_version >= 2530 &&
+ adev->gfx.mec_fw_version >= 2680 &&
+ adev->mes.fw_version[0] >= 100)
+ adev->gfx.enable_cleaner_shader = true;
+ break;
default:
adev->gfx.enable_cleaner_shader = false;
break;
@@ -1437,11 +1450,19 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
}
}
- /* TODO: Add queue reset mask when FW fully supports it */
adev->gfx.gfx_supported_reset =
amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
adev->gfx.compute_supported_reset =
amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+ 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 >= 2660) &&
+ (adev->gfx.mec_fw_version >= 2920)) {
+ adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+ adev->gfx.gfx_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+ }
+ }
if (!adev->enable_mes_kiq) {
r = amdgpu_gfx_kiq_init(adev, GFX12_MEC_HPD_SIZE, 0);
@@ -1610,6 +1631,7 @@ static u32 gfx_v12_0_get_rb_active_bitmap(struct amdgpu_device *adev)
static void gfx_v12_0_setup_rb(struct amdgpu_device *adev)
{
+ u32 rb_bitmap_per_sa;
u32 rb_bitmap_width_per_sa;
u32 max_sa;
u32 active_sa_bitmap;
@@ -1627,12 +1649,14 @@ static void gfx_v12_0_setup_rb(struct amdgpu_device *adev)
adev->gfx.config.max_sh_per_se;
rb_bitmap_width_per_sa = adev->gfx.config.max_backends_per_se /
adev->gfx.config.max_sh_per_se;
+ rb_bitmap_per_sa = amdgpu_gfx_create_bitmask(rb_bitmap_width_per_sa);
+
for (i = 0; i < max_sa; i++) {
if (active_sa_bitmap & (1 << i))
- active_rb_bitmap |= (0x3 << (i * rb_bitmap_width_per_sa));
+ active_rb_bitmap |= (rb_bitmap_per_sa << (i * rb_bitmap_width_per_sa));
}
- active_rb_bitmap |= global_active_rb_bitmap;
+ active_rb_bitmap &= global_active_rb_bitmap;
adev->gfx.config.backend_enable_mask = active_rb_bitmap;
adev->gfx.config.num_rbs = hweight32(active_rb_bitmap);
}
@@ -2413,7 +2437,7 @@ static int gfx_v12_0_cp_gfx_load_me_microcode_rs64(struct amdgpu_device *adev)
(void **)&adev->gfx.me.me_fw_data_ptr);
if (r) {
dev_err(adev->dev, "(%d) failed to create me data bo\n", r);
- gfx_v12_0_pfp_fini(adev);
+ gfx_v12_0_me_fini(adev);
return r;
}
@@ -2832,9 +2856,7 @@ static void gfx_v12_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
}
static void gfx_v12_0_cp_set_doorbell_range(struct amdgpu_device *adev)
@@ -3864,10 +3886,10 @@ static void gfx_v12_cntl_pg(struct amdgpu_device *adev, bool enable)
}
#endif
-static int gfx_v12_0_set_powergating_state(void *handle,
+static int gfx_v12_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
@@ -3999,17 +4021,6 @@ static void gfx_v12_0_update_coarse_grain_clock_gating(struct amdgpu_device *ade
if (def != data)
WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D, data);
-
- data = RREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL);
- data &= ~SDMA0_RLC_CGCG_CTRL__CGCG_INT_ENABLE_MASK;
- WREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL, data);
-
- /* Some ASICs only have one SDMA instance, not need to configure SDMA1 */
- if (adev->sdma.num_instances > 1) {
- data = RREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL);
- data &= ~SDMA1_RLC_CGCG_CTRL__CGCG_INT_ENABLE_MASK;
- WREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL, data);
- }
}
}
@@ -4115,10 +4126,10 @@ static int gfx_v12_0_update_gfx_clock_gating(struct amdgpu_device *adev,
return 0;
}
-static int gfx_v12_0_set_clockgating_state(void *handle,
+static int gfx_v12_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -5233,24 +5244,16 @@ static int gfx_v12_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid)
static int gfx_v12_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid)
{
struct amdgpu_device *adev = ring->adev;
- int r, i;
+ int r;
if (amdgpu_sriov_vf(adev))
return -EINVAL;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
- mutex_lock(&adev->srbm_mutex);
- soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
- WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2);
- WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1);
- for (i = 0; i < adev->usec_timeout; i++) {
- if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
- break;
- udelay(1);
+ 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;
}
- soc24_grbm_select(adev, 0, 0, 0, 0);
- mutex_unlock(&adev->srbm_mutex);
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
r = amdgpu_bo_reserve(ring->mqd_obj, false);
if (unlikely(r != 0)) {
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h
index bcc9c72ccbde..f7184b2dc4e8 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h
@@ -26,4 +26,6 @@
extern const struct amdgpu_ip_block_version gfx_v12_0_ip_block;
+int gfx_v12_0_request_gfx_index_mutex(struct amdgpu_device *adev,
+ bool req);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
index 41f50bf380c4..f26e2cdec07a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
@@ -337,6 +337,7 @@ static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", chip_name);
if (err)
goto out;
@@ -345,6 +346,7 @@ static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", chip_name);
if (err)
goto out;
@@ -353,6 +355,7 @@ static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_ce.bin", chip_name);
if (err)
goto out;
@@ -361,6 +364,7 @@ static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", chip_name);
if (err)
goto out;
@@ -1906,7 +1910,7 @@ static int gfx_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
error:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
return r;
}
@@ -3373,11 +3377,11 @@ static int gfx_v6_0_priv_inst_irq(struct amdgpu_device *adev,
return 0;
}
-static int gfx_v6_0_set_clockgating_state(void *handle,
+static int gfx_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
bool gate = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_CG_STATE_GATE)
gate = true;
@@ -3395,11 +3399,11 @@ static int gfx_v6_0_set_clockgating_state(void *handle,
return 0;
}
-static int gfx_v6_0_set_powergating_state(void *handle,
+static int gfx_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
bool gate = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_PG_STATE_GATE)
gate = true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index 824d5913103b..84745b2453ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -934,33 +934,39 @@ static int gfx_v7_0_init_microcode(struct amdgpu_device *adev)
}
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", chip_name);
if (err)
goto out;
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", chip_name);
if (err)
goto out;
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_ce.bin", chip_name);
if (err)
goto out;
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec.bin", chip_name);
if (err)
goto out;
if (adev->asic_type == CHIP_KAVERI) {
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec2.bin", chip_name);
if (err)
goto out;
}
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", chip_name);
out:
if (err) {
@@ -2324,7 +2330,7 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
error:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
return r;
}
@@ -4846,11 +4852,11 @@ static int gfx_v7_0_priv_inst_irq(struct amdgpu_device *adev,
return 0;
}
-static int gfx_v7_0_set_clockgating_state(void *handle,
+static int gfx_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
bool gate = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_CG_STATE_GATE)
gate = true;
@@ -4869,11 +4875,11 @@ static int gfx_v7_0_set_clockgating_state(void *handle,
return 0;
}
-static int gfx_v7_0_set_powergating_state(void *handle,
+static int gfx_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
bool gate = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_PG_STATE_GATE)
gate = true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index b7006c41e270..6a025438f9d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -914,7 +914,7 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err2:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
amdgpu_device_wb_free(adev, index);
@@ -982,13 +982,16 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_OPTIONAL,
"amdgpu/%s_pfp_2.bin", chip_name);
if (err == -ENODEV) {
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", chip_name);
}
} else {
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", chip_name);
}
if (err)
@@ -999,13 +1002,16 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_OPTIONAL,
"amdgpu/%s_me_2.bin", chip_name);
if (err == -ENODEV) {
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", chip_name);
}
} else {
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", chip_name);
}
if (err)
@@ -1017,13 +1023,16 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_OPTIONAL,
"amdgpu/%s_ce_2.bin", chip_name);
if (err == -ENODEV) {
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_ce.bin", chip_name);
}
} else {
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_ce.bin", chip_name);
}
if (err)
@@ -1044,6 +1053,7 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
adev->virt.chained_ib_support = false;
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", chip_name);
if (err)
goto out;
@@ -1093,13 +1103,16 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_OPTIONAL,
"amdgpu/%s_mec_2.bin", chip_name);
if (err == -ENODEV) {
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec.bin", chip_name);
}
} else {
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec.bin", chip_name);
}
if (err)
@@ -1112,13 +1125,16 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
(adev->asic_type != CHIP_TOPAZ)) {
if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_OPTIONAL,
"amdgpu/%s_mec2_2.bin", chip_name);
if (err == -ENODEV) {
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec2.bin", chip_name);
}
} else {
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec2.bin", chip_name);
}
if (!err) {
@@ -1640,7 +1656,7 @@ static int gfx_v8_0_do_edc_gpr_workarounds(struct amdgpu_device *adev)
RREG32(sec_ded_counter_registers[i]);
fail:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
return r;
@@ -4304,9 +4320,7 @@ static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32(mmRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32(mmRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32(mmRLC_CP_SCHEDULERS, tmp);
+ WREG32(mmRLC_CP_SCHEDULERS, tmp | 0x80);
}
static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
@@ -5321,7 +5335,7 @@ static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *ade
(adev->asic_type == CHIP_POLARIS12) ||
(adev->asic_type == CHIP_VEGAM))
/* Send msg to SMU via Powerplay */
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, enable);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, enable, 0);
WREG32_FIELD(RLC_PG_CNTL, STATIC_PER_CU_PG_ENABLE, enable ? 1 : 0);
}
@@ -5367,10 +5381,10 @@ static void cz_update_gfx_cg_power_gating(struct amdgpu_device *adev,
}
}
-static int gfx_v8_0_set_powergating_state(void *handle,
+static int gfx_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
@@ -5625,8 +5639,6 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
{
uint32_t temp, data;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
/* It is disabled by HW by default */
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) {
@@ -5720,8 +5732,6 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
/* 7- wait for RLC_SERDES_CU_MASTER & RLC_SERDES_NONCU_MASTER idle */
gfx_v8_0_wait_for_rlc_serdes(adev);
}
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
@@ -5731,8 +5741,6 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
temp = data = RREG32(mmRLC_CGCG_CGLS_CTRL);
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
temp1 = data1 = RREG32(mmRLC_CGTT_MGCG_OVERRIDE);
data1 &= ~RLC_CGTT_MGCG_OVERRIDE__CGCG_MASK;
@@ -5813,12 +5821,12 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
}
gfx_v8_0_wait_for_rlc_serdes(adev);
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
bool enable)
{
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
if (enable) {
/* CGCG/CGLS should be enabled after MGCG/MGLS/TS(CG/LS)
* === MGCG + MGLS + TS(CG/LS) ===
@@ -5832,6 +5840,8 @@ static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
gfx_v8_0_update_coarse_grain_clock_gating(adev, enable);
gfx_v8_0_update_medium_grain_clock_gating(adev, enable);
}
+
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
return 0;
}
@@ -5982,10 +5992,10 @@ static int gfx_v8_0_polaris_update_gfx_clock_gating(struct amdgpu_device *adev,
return 0;
}
-static int gfx_v8_0_set_clockgating_state(void *handle,
+static int gfx_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 0b6f09f2cc9b..0dce4421418c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -1243,7 +1243,7 @@ static int gfx_v9_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err2:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
amdgpu_device_wb_free(adev, index);
@@ -1429,18 +1429,21 @@ static int gfx_v9_0_init_cp_gfx_microcode(struct amdgpu_device *adev,
int err;
err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_pfp.bin", chip_name);
if (err)
goto out;
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_PFP);
err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_me.bin", chip_name);
if (err)
goto out;
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_ME);
err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_ce.bin", chip_name);
if (err)
goto out;
@@ -1476,6 +1479,7 @@ static int gfx_v9_0_init_rlc_microcode(struct amdgpu_device *adev,
(((adev->pdev->revision >= 0xC8) && (adev->pdev->revision <= 0xCF)) ||
((adev->pdev->revision >= 0xD8) && (adev->pdev->revision <= 0xDF))))
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc_am4.bin", chip_name);
else if (!strcmp(chip_name, "raven") && (amdgpu_pm_load_smu_firmware(adev, &smu_version) == 0) &&
(smu_version >= 0x41e2b))
@@ -1483,9 +1487,11 @@ static int gfx_v9_0_init_rlc_microcode(struct amdgpu_device *adev,
*SMC is loaded by SBIOS on APU and it's able to get the SMU version directly.
*/
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_kicker_rlc.bin", chip_name);
else
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", chip_name);
if (err)
goto out;
@@ -1518,9 +1524,11 @@ static int gfx_v9_0_init_cp_compute_microcode(struct amdgpu_device *adev,
if (amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_ALDEBARAN))
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
- "amdgpu/%s_sjt_mec.bin", chip_name);
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_sjt_mec.bin", chip_name);
else
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec.bin", chip_name);
if (err)
goto out;
@@ -1531,9 +1539,11 @@ static int gfx_v9_0_init_cp_compute_microcode(struct amdgpu_device *adev,
if (gfx_v9_0_load_mec2_fw_bin_support(adev)) {
if (amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_ALDEBARAN))
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sjt_mec2.bin", chip_name);
else
err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_mec2.bin", chip_name);
if (!err) {
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2);
@@ -3488,9 +3498,7 @@ static void gfx_v9_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp | 0x80);
}
static void gfx_v9_0_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd)
@@ -4780,7 +4788,7 @@ static int gfx_v9_0_do_edc_gpr_workarounds(struct amdgpu_device *adev)
}
fail:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
return r;
@@ -4956,8 +4964,6 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
{
uint32_t data, def;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
/* It is disabled by HW by default */
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
/* 1 - RLC_CGTT_MGCG_OVERRIDE */
@@ -5022,8 +5028,6 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
WREG32_SOC15(GC, 0, mmCP_MEM_SLP_CNTL, data);
}
}
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
@@ -5034,8 +5038,6 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
if (!adev->gfx.num_gfx_rings)
return;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
/* Enable 3D CGCG/CGLS */
if (enable) {
/* write cmd to clear cgcg/cgls ov */
@@ -5077,8 +5079,6 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
if (def != data)
WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D, data);
}
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
@@ -5086,8 +5086,6 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
{
uint32_t def, data;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE);
/* unset CGCG override */
@@ -5129,13 +5127,12 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
if (def != data)
WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL, data);
}
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static int gfx_v9_0_update_gfx_clock_gating(struct amdgpu_device *adev,
bool enable)
{
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
if (enable) {
/* CGCG/CGLS should be enabled after MGCG/MGLS
* === MGCG + MGLS ===
@@ -5155,6 +5152,7 @@ static int gfx_v9_0_update_gfx_clock_gating(struct amdgpu_device *adev,
/* === MGCG + MGLS === */
gfx_v9_0_update_medium_grain_clock_gating(adev, enable);
}
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
return 0;
}
@@ -5232,10 +5230,10 @@ static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = {
.is_rlcg_access_range = gfx_v9_0_is_rlcg_access_range,
};
-static int gfx_v9_0_set_powergating_state(void *handle,
+static int gfx_v9_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
@@ -5277,10 +5275,10 @@ static int gfx_v9_0_set_powergating_state(void *handle,
return 0;
}
-static int gfx_v9_0_set_clockgating_state(void *handle,
+static int gfx_v9_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -7439,6 +7437,38 @@ static void gfx_v9_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
amdgpu_ring_write(ring, 0); /* RESERVED field, programmed to zero */
}
+static void gfx_v9_0_ring_begin_use_compute(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+ struct amdgpu_ip_block *gfx_block =
+ amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+
+ amdgpu_gfx_enforce_isolation_ring_begin_use(ring);
+
+ /* Raven and PCO APUs seem to have stability issues
+ * with compute and gfxoff and gfx pg. Disable gfx pg during
+ * submission and allow again afterwards.
+ */
+ if (gfx_block && amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0))
+ gfx_v9_0_set_powergating_state(gfx_block, AMD_PG_STATE_UNGATE);
+}
+
+static void gfx_v9_0_ring_end_use_compute(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+ struct amdgpu_ip_block *gfx_block =
+ amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+
+ /* Raven and PCO APUs seem to have stability issues
+ * with compute and gfxoff and gfx pg. Disable gfx pg during
+ * submission and allow again afterwards.
+ */
+ if (gfx_block && amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0))
+ gfx_v9_0_set_powergating_state(gfx_block, AMD_PG_STATE_GATE);
+
+ amdgpu_gfx_enforce_isolation_ring_end_use(ring);
+}
+
static const struct amd_ip_funcs gfx_v9_0_ip_funcs = {
.name = "gfx_v9_0",
.early_init = gfx_v9_0_early_init,
@@ -7615,8 +7645,8 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
.emit_wave_limit = gfx_v9_0_emit_wave_limit,
.reset = gfx_v9_0_reset_kcq,
.emit_cleaner_shader = gfx_v9_0_ring_emit_cleaner_shader,
- .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use,
- .end_use = amdgpu_gfx_enforce_isolation_ring_end_use,
+ .begin_use = gfx_v9_0_ring_begin_use_compute,
+ .end_use = gfx_v9_0_ring_end_use_compute,
};
static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
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 3f4fd2f08163..d81449f9d822 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c
@@ -412,7 +412,7 @@ static int gfx_v9_4_2_run_shader(struct amdgpu_device *adev,
r = amdgpu_ib_schedule(ring, 1, ib, NULL, fence_ptr);
if (r) {
dev_err(adev->dev, "ib submit failed (%d).\n", r);
- amdgpu_ib_free(adev, ib, NULL);
+ amdgpu_ib_free(ib, NULL);
}
return r;
}
@@ -611,16 +611,16 @@ static int gfx_v9_4_2_do_sgprs_init(struct amdgpu_device *adev)
}
disp2_failed:
- amdgpu_ib_free(adev, &disp_ibs[2], NULL);
+ amdgpu_ib_free(&disp_ibs[2], NULL);
dma_fence_put(fences[2]);
disp1_failed:
- amdgpu_ib_free(adev, &disp_ibs[1], NULL);
+ amdgpu_ib_free(&disp_ibs[1], NULL);
dma_fence_put(fences[1]);
disp0_failed:
- amdgpu_ib_free(adev, &disp_ibs[0], NULL);
+ amdgpu_ib_free(&disp_ibs[0], NULL);
dma_fence_put(fences[0]);
pro_end:
- amdgpu_ib_free(adev, &wb_ib, NULL);
+ amdgpu_ib_free(&wb_ib, NULL);
if (r)
dev_info(adev->dev, "Init SGPRS Failed\n");
@@ -687,10 +687,10 @@ static int gfx_v9_4_2_do_vgprs_init(struct amdgpu_device *adev)
}
disp_failed:
- amdgpu_ib_free(adev, &disp_ib, NULL);
+ amdgpu_ib_free(&disp_ib, NULL);
dma_fence_put(fence);
pro_end:
- amdgpu_ib_free(adev, &wb_ib, NULL);
+ amdgpu_ib_free(&wb_ib, NULL);
if (r)
dev_info(adev->dev, "Init VGPRS Failed\n");
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 54459254bd37..2ba185875baa 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -43,8 +43,10 @@
MODULE_FIRMWARE("amdgpu/gc_9_4_3_mec.bin");
MODULE_FIRMWARE("amdgpu/gc_9_4_4_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_5_0_mec.bin");
MODULE_FIRMWARE("amdgpu/gc_9_4_3_rlc.bin");
MODULE_FIRMWARE("amdgpu/gc_9_4_4_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_5_0_rlc.bin");
MODULE_FIRMWARE("amdgpu/gc_9_4_3_sjt_mec.bin");
MODULE_FIRMWARE("amdgpu/gc_9_4_4_sjt_mec.bin");
@@ -54,10 +56,6 @@ MODULE_FIRMWARE("amdgpu/gc_9_4_4_sjt_mec.bin");
#define GOLDEN_GB_ADDR_CONFIG 0x2a114042
#define CP_HQD_PERSISTENT_STATE_DEFAULT 0xbe05301
-#define mmSMNAID_XCD0_MCA_SMU 0x36430400 /* SMN AID XCD0 */
-#define mmSMNAID_XCD1_MCA_SMU 0x38430400 /* SMN AID XCD1 */
-#define mmSMNXCD_XCD0_MCA_SMU 0x40430400 /* SMN XCD XCD0 */
-
#define XCC_REG_RANGE_0_LOW 0x2000 /* XCC gfxdec0 lower Bound */
#define XCC_REG_RANGE_0_HIGH 0x3400 /* XCC gfxdec0 upper Bound */
#define XCC_REG_RANGE_1_LOW 0xA000 /* XCC gfxdec1 lower Bound */
@@ -351,13 +349,17 @@ static void gfx_v9_4_3_init_golden_registers(struct amdgpu_device *adev)
WREG32_SOC15(GC, dev_inst, regGB_ADDR_CONFIG,
GOLDEN_GB_ADDR_CONFIG);
- /* Golden settings applied by driver for ASIC with rev_id 0 */
- if (adev->rev_id == 0) {
- WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL1,
- REDUCE_FIFO_DEPTH_BY_2, 2);
+ if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) {
+ WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL2, SPARE, 0x1);
} else {
- WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL2,
- SPARE, 0x1);
+ /* Golden settings applied by driver for ASIC with rev_id 0 */
+ if (adev->rev_id == 0) {
+ WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL1,
+ REDUCE_FIFO_DEPTH_BY_2, 2);
+ } else {
+ WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL2,
+ SPARE, 0x1);
+ }
}
}
}
@@ -501,7 +503,7 @@ static int gfx_v9_4_3_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err2:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
amdgpu_device_wb_free(adev, index);
@@ -545,6 +547,7 @@ static int gfx_v9_4_3_init_rlc_microcode(struct amdgpu_device *adev,
err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_rlc.bin", chip_name);
if (err)
goto out;
@@ -576,12 +579,19 @@ static int gfx_v9_4_3_init_cp_compute_microcode(struct amdgpu_device *adev,
{
int err;
- if (amdgpu_sriov_vf(adev))
+ if (amdgpu_sriov_vf(adev)) {
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
- "amdgpu/%s_sjt_mec.bin", chip_name);
- else
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_sjt_mec.bin", chip_name);
+
+ if (err)
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mec.bin", chip_name);
+ } else
err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
- "amdgpu/%s_mec.bin", chip_name);
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mec.bin", chip_name);
if (err)
goto out;
amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1);
@@ -935,6 +945,7 @@ static int gfx_v9_4_3_gpu_early_init(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
adev->gfx.config.max_hw_contexts = 8;
adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -1785,9 +1796,7 @@ static void gfx_v9_4_3_xcc_kiq_setting(struct amdgpu_ring *ring, int xcc_id)
tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp | 0x80);
}
static void gfx_v9_4_3_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd)
@@ -2770,16 +2779,16 @@ static const struct amdgpu_rlc_funcs gfx_v9_4_3_rlc_funcs = {
.is_rlcg_access_range = gfx_v9_4_3_is_rlcg_access_range,
};
-static int gfx_v9_4_3_set_powergating_state(void *handle,
+static int gfx_v9_4_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
}
-static int gfx_v9_4_3_set_clockgating_state(void *handle,
+static int gfx_v9_4_3_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int i, num_xcc;
if (amdgpu_sriov_vf(adev))
@@ -4659,7 +4668,6 @@ static void gfx_v9_4_3_ip_dump(struct amdgpu_ip_block *ip_block)
num_xcc = NUM_XCC(adev->gfx.xcc_mask);
- amdgpu_gfx_off_ctrl(adev, false);
for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
xcc_offset = xcc_id * reg_count;
for (i = 0; i < reg_count; i++)
@@ -4667,7 +4675,6 @@ static void gfx_v9_4_3_ip_dump(struct amdgpu_ip_block *ip_block)
RREG32(SOC15_REG_ENTRY_OFFSET_INST(gc_reg_list_9_4_3[i],
GET_INST(GC, xcc_id)));
}
- amdgpu_gfx_off_ctrl(adev, true);
/* dump compute queue registers for all instances */
if (!adev->gfx.ip_dump_compute_queues)
@@ -4676,7 +4683,6 @@ static void gfx_v9_4_3_ip_dump(struct amdgpu_ip_block *ip_block)
num_inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
adev->gfx.mec.num_queue_per_pipe;
reg_count = ARRAY_SIZE(gc_cp_reg_list_9_4_3);
- amdgpu_gfx_off_ctrl(adev, false);
mutex_lock(&adev->srbm_mutex);
for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
xcc_offset = xcc_id * reg_count * num_inst;
@@ -4703,7 +4709,6 @@ static void gfx_v9_4_3_ip_dump(struct amdgpu_ip_block *ip_block)
}
soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
- amdgpu_gfx_off_ctrl(adev, true);
}
static void gfx_v9_4_3_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
@@ -4866,6 +4871,7 @@ static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
/* 9.4.3 removed all the GDS internal memory,
* only support GWS opcode in kernel, like barrier
* semaphore.etc */
@@ -4879,6 +4885,7 @@ static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev)
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
/* deprecated for 9.4.3, no usage at all */
adev->gds.gds_compute_max_wave_id = 0;
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c
index ed8e130c7d19..5470cef7e9bd 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c
@@ -368,7 +368,9 @@ static void gfxhub_v1_2_xcc_setup_vmid_config(struct amdgpu_device *adev,
amdgpu_ip_version(adev, GC_HWIP, 0) ==
IP_VERSION(9, 4, 3) ||
amdgpu_ip_version(adev, GC_HWIP, 0) ==
- IP_VERSION(9, 4, 4));
+ IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) ==
+ IP_VERSION(9, 5, 0));
WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT1_CNTL,
i * hub->ctx_distance, tmp);
WREG32_SOC15_OFFSET(GC, GET_INST(GC, j),
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 697599c46240..9bedca9a79c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -1088,11 +1088,11 @@ static int gmc_v10_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int gmc_v10_0_set_clockgating_state(void *handle,
+static int gmc_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
int r;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
/*
* The issue mmhub can't disconnect from DF with MMHUB clock gating being disabled
@@ -1131,7 +1131,7 @@ static void gmc_v10_0_get_clockgating_state(void *handle, u64 *flags)
athub_v2_0_get_clockgating(adev, flags);
}
-static int gmc_v10_0_set_powergating_state(void *handle,
+static int gmc_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index f893ab4c14df..72751ab4c766 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -996,11 +996,11 @@ static int gmc_v11_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int gmc_v11_0_set_clockgating_state(void *handle,
+static int gmc_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
int r;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
r = adev->mmhub.funcs->set_clockgating(adev, state);
if (r)
@@ -1018,7 +1018,7 @@ static void gmc_v11_0_get_clockgating_state(void *handle, u64 *flags)
athub_v3_0_get_clockgating(adev, flags);
}
-static int gmc_v11_0_set_powergating_state(void *handle,
+static int gmc_v11_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
index d22b027fd0bb..c3c144a4f45e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
@@ -40,7 +40,7 @@
#include "gfxhub_v12_0.h"
#include "mmhub_v4_1_0.h"
#include "athub_v4_1_0.h"
-
+#include "umc_v8_14.h"
static int gmc_v12_0_ecc_interrupt_state(struct amdgpu_device *adev,
struct amdgpu_irq_src *src,
@@ -501,9 +501,6 @@ static void gmc_v12_0_get_vm_pte(struct amdgpu_device *adev,
uint64_t *flags)
{
struct amdgpu_bo *bo = mapping->bo_va->base.bo;
- struct amdgpu_device *bo_adev;
- bool coherent, is_system;
-
*flags &= ~AMDGPU_PTE_EXECUTABLE;
*flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE;
@@ -519,25 +516,11 @@ static void gmc_v12_0_get_vm_pte(struct amdgpu_device *adev,
*flags &= ~AMDGPU_PTE_VALID;
}
- if (!bo)
- return;
-
- if (bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
- AMDGPU_GEM_CREATE_UNCACHED))
- *flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_UC);
-
- bo_adev = amdgpu_ttm_adev(bo->tbo.bdev);
- coherent = bo->flags & AMDGPU_GEM_CREATE_COHERENT;
- is_system = (bo->tbo.resource->mem_type == TTM_PL_TT) ||
- (bo->tbo.resource->mem_type == AMDGPU_PL_PREEMPT);
-
if (bo && bo->flags & AMDGPU_GEM_CREATE_GFX12_DCC)
*flags |= AMDGPU_PTE_DCC;
- /* WA for HW bug */
- if (is_system || ((bo_adev != adev) && coherent))
- *flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_NC);
-
+ if (bo && bo->flags & AMDGPU_GEM_CREATE_UNCACHED)
+ *flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_UC);
}
static unsigned gmc_v12_0_get_vbios_fb_size(struct amdgpu_device *adev)
@@ -581,6 +564,18 @@ static void gmc_v12_0_set_gmc_funcs(struct amdgpu_device *adev)
static void gmc_v12_0_set_umc_funcs(struct amdgpu_device *adev)
{
+ switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
+ case IP_VERSION(8, 14, 0):
+ adev->umc.channel_inst_num = UMC_V8_14_CHANNEL_INSTANCE_NUM;
+ adev->umc.umc_inst_num = UMC_V8_14_UMC_INSTANCE_NUM(adev);
+ adev->umc.node_inst_num = 0;
+ adev->umc.max_ras_err_cnt_per_query = UMC_V8_14_TOTAL_CHANNEL_NUM(adev);
+ adev->umc.channel_offs = UMC_V8_14_PER_CHANNEL_OFFSET;
+ adev->umc.ras = &umc_v8_14_ras;
+ break;
+ default:
+ break;
+ }
}
@@ -829,6 +824,10 @@ static int gmc_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
amdgpu_vm_manager_init(adev);
+ r = amdgpu_gmc_ras_sw_init(adev);
+ if (r)
+ return r;
+
return 0;
}
@@ -980,11 +979,11 @@ static int gmc_v12_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int gmc_v12_0_set_clockgating_state(void *handle,
+static int gmc_v12_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
int r;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
r = adev->mmhub.funcs->set_clockgating(adev, state);
if (r)
@@ -1002,7 +1001,7 @@ static void gmc_v12_0_get_clockgating_state(void *handle, u64 *flags)
athub_v4_1_0_get_clockgating(adev, flags);
}
-static int gmc_v12_0_set_powergating_state(void *handle,
+static int gmc_v12_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
index ca000b3d1afc..2245dda92021 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
@@ -131,7 +131,8 @@ static int gmc_v6_0_init_microcode(struct amdgpu_device *adev)
if (((RREG32(mmMC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58)
chip_name = "si58";
- err = amdgpu_ucode_request(adev, &adev->gmc.fw, "amdgpu/%s_mc.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->gmc.fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mc.bin", chip_name);
if (err) {
dev_err(adev->dev,
"si_mc: Failed to load firmware \"%s_mc.bin\"\n",
@@ -1094,13 +1095,13 @@ static int gmc_v6_0_process_interrupt(struct amdgpu_device *adev,
return 0;
}
-static int gmc_v6_0_set_clockgating_state(void *handle,
+static int gmc_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int gmc_v6_0_set_powergating_state(void *handle,
+static int gmc_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
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 b6016f11956e..9aac4b1101e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -157,7 +157,8 @@ static int gmc_v7_0_init_microcode(struct amdgpu_device *adev)
return -EINVAL;
}
- err = amdgpu_ucode_request(adev, &adev->gmc.fw, "amdgpu/%s_mc.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->gmc.fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mc.bin", chip_name);
if (err) {
pr_err("cik_mc: Failed to load firmware \"%s_mc.bin\"\n", chip_name);
amdgpu_ucode_release(&adev->gmc.fw);
@@ -1317,11 +1318,11 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev,
return 0;
}
-static int gmc_v7_0_set_clockgating_state(void *handle,
+static int gmc_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
bool gate = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_CG_STATE_GATE)
gate = true;
@@ -1337,7 +1338,7 @@ static int gmc_v7_0_set_clockgating_state(void *handle,
return 0;
}
-static int gmc_v7_0_set_powergating_state(void *handle,
+static int gmc_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 12d5967ecd45..d06585207c33 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -259,7 +259,8 @@ static int gmc_v8_0_init_microcode(struct amdgpu_device *adev)
return -EINVAL;
}
- err = amdgpu_ucode_request(adev, &adev->gmc.fw, "amdgpu/%s_mc.bin", chip_name);
+ err = amdgpu_ucode_request(adev, &adev->gmc.fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mc.bin", chip_name);
if (err) {
pr_err("mc: Failed to load firmware \"%s_mc.bin\"\n", chip_name);
amdgpu_ucode_release(&adev->gmc.fw);
@@ -1658,10 +1659,10 @@ static void fiji_update_mc_light_sleep(struct amdgpu_device *adev,
}
}
-static int gmc_v8_0_set_clockgating_state(void *handle,
+static int gmc_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1679,7 +1680,7 @@ static int gmc_v8_0_set_clockgating_state(void *handle,
return 0;
}
-static int gmc_v8_0_set_powergating_state(void *handle,
+static int gmc_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 50c5da3020cb..291549765c38 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -623,6 +623,9 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
}
}
+ if (kgd2kfd_vmfault_fast_path(adev, entry, retry_fault))
+ return 1;
+
if (!printk_ratelimit())
return 0;
@@ -645,7 +648,8 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
soc15_ih_clientid_name[entry->client_id]);
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
dev_err(adev->dev, " cookie node_id %d fault from die %s%d%s\n",
node_id, node_id % 4 == 3 ? "RSV" : "AID", node_id / 4,
node_id % 4 == 1 ? ".XCD0" : node_id % 4 == 2 ? ".XCD1" : "");
@@ -795,7 +799,8 @@ static bool gmc_v9_0_use_invalidate_semaphore(struct amdgpu_device *adev,
{
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
return false;
return ((vmhub == AMDGPU_MMHUB0(0) ||
@@ -1138,12 +1143,13 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
bool uncached = bo->flags & AMDGPU_GEM_CREATE_UNCACHED;
struct amdgpu_vm *vm = mapping->bo_va->base.vm;
unsigned int mtype_local, mtype;
+ uint32_t gc_ip_version = amdgpu_ip_version(adev, GC_HWIP, 0);
bool snoop = false;
bool is_local;
dma_resv_assert_held(bo->tbo.base.resv);
- switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+ switch (gc_ip_version) {
case IP_VERSION(9, 4, 1):
case IP_VERSION(9, 4, 2):
if (is_vram) {
@@ -1157,10 +1163,7 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
/* FIXME: is this still needed? Or does
* amdgpu_ttm_tt_pde_flags already handle this?
*/
- if ((amdgpu_ip_version(adev, GC_HWIP, 0) ==
- IP_VERSION(9, 4, 2) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) ==
- IP_VERSION(9, 4, 3)) &&
+ if (gc_ip_version == IP_VERSION(9, 4, 2) &&
adev->gmc.xgmi.connected_to_cpu)
snoop = true;
} else {
@@ -1184,6 +1187,7 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
break;
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
/* Only local VRAM BOs or system memory on non-NUMA APUs
* can be assumed to be local in their entirety. Choose
* MTYPE_NC as safe fallback for all system memory BOs on
@@ -1208,7 +1212,7 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
if (uncached) {
mtype = MTYPE_UC;
} else if (ext_coherent) {
- if (adev->rev_id)
+ if (gc_ip_version == IP_VERSION(9, 5, 0) || adev->rev_id)
mtype = is_local ? MTYPE_CC : MTYPE_UC;
else
mtype = MTYPE_UC;
@@ -1218,10 +1222,10 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
/* dGPU */
if (is_local)
mtype = mtype_local;
- else if (is_vram)
- mtype = MTYPE_NC;
- else
+ else if (gc_ip_version < IP_VERSION(9, 5, 0) && !is_vram)
mtype = MTYPE_UC;
+ else
+ mtype = MTYPE_NC;
}
break;
@@ -1275,7 +1279,8 @@ static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev,
* memory can use more efficient MTYPEs.
*/
if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 3) &&
- amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 4) &&
+ amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 5, 0))
return;
/* Only direct-mapped memory allows us to determine the NUMA node from
@@ -1540,6 +1545,7 @@ static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev)
adev->mmhub.ras = &mmhub_v1_7_ras;
break;
case IP_VERSION(1, 8, 0):
+ case IP_VERSION(1, 8, 1):
adev->mmhub.ras = &mmhub_v1_8_ras;
break;
default:
@@ -1551,7 +1557,8 @@ static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev)
static void gmc_v9_0_set_gfxhub_funcs(struct amdgpu_device *adev)
{
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
adev->gfxhub.funcs = &gfxhub_v1_2_funcs;
else
adev->gfxhub.funcs = &gfxhub_v1_0_funcs;
@@ -1619,7 +1626,8 @@ static int gmc_v9_0_early_init(struct amdgpu_ip_block *ip_block)
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 0) ||
amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) ||
amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
adev->gmc.xgmi.supported = true;
if (amdgpu_ip_version(adev, XGMI_HWIP, 0) == IP_VERSION(6, 1, 0)) {
@@ -1792,6 +1800,7 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
case IP_VERSION(9, 4, 2):
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
default:
adev->gmc.gart_size = 512ULL << 20;
break;
@@ -2070,7 +2079,8 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
spin_lock_init(&adev->gmc.invalidate_lock);
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) {
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) {
gmc_v9_4_3_init_vram_info(adev);
} else if (!adev->bios) {
if (adev->flags & AMD_IS_APU) {
@@ -2154,6 +2164,7 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
break;
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
bitmap_set(adev->vmhubs_mask, AMDGPU_GFXHUB(0),
NUM_XCC(adev->gfx.xcc_mask));
@@ -2220,7 +2231,8 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
amdgpu_gmc_get_vbios_allocations(adev);
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) {
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) {
r = gmc_v9_0_init_mem_ranges(adev);
if (r)
return r;
@@ -2250,7 +2262,8 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
(amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) ||
amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) ?
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) ?
3 :
8;
@@ -2263,7 +2276,8 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
return r;
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
amdgpu_gmc_sysfs_init(adev);
return 0;
@@ -2274,7 +2288,8 @@ static int gmc_v9_0_sw_fini(struct amdgpu_ip_block *ip_block)
struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
- amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+ amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
amdgpu_gmc_sysfs_fini(adev);
amdgpu_gmc_ras_fini(adev);
@@ -2544,10 +2559,10 @@ static int gmc_v9_0_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int gmc_v9_0_set_clockgating_state(void *handle,
+static int gmc_v9_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
adev->mmhub.funcs->set_clockgating(adev, state);
@@ -2565,7 +2580,7 @@ static void gmc_v9_0_get_clockgating_state(void *handle, u64 *flags)
athub_v1_0_get_clockgating(adev, flags);
}
-static int gmc_v9_0_set_powergating_state(void *handle,
+static int gmc_v9_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
index 7f45e93c0397..8ac3d3282268 100644
--- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
@@ -392,13 +392,13 @@ static int iceland_ih_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int iceland_ih_set_clockgating_state(void *handle,
+static int iceland_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int iceland_ih_set_powergating_state(void *handle,
+static int iceland_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
index 38f953fd65d9..f8a485164437 100644
--- a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
@@ -693,10 +693,10 @@ static void ih_v6_0_update_clockgating_state(struct amdgpu_device *adev,
}
}
-static int ih_v6_0_set_clockgating_state(void *handle,
+static int ih_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
ih_v6_0_update_clockgating_state(adev,
state == AMD_CG_STATE_GATE);
@@ -756,10 +756,10 @@ static void ih_v6_0_update_ih_mem_power_gating(struct amdgpu_device *adev,
WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
}
-static int ih_v6_0_set_powergating_state(void *handle,
+static int ih_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (adev->pg_flags & AMD_PG_SUPPORT_IH_SRAM_PG)
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c
index 61381e0c3795..dd0042efceec 100644
--- a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c
@@ -674,10 +674,10 @@ static void ih_v6_1_update_clockgating_state(struct amdgpu_device *adev,
return;
}
-static int ih_v6_1_set_clockgating_state(void *handle,
+static int ih_v6_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
ih_v6_1_update_clockgating_state(adev,
state == AMD_CG_STATE_GATE);
@@ -737,10 +737,10 @@ static void ih_v6_1_update_ih_mem_power_gating(struct amdgpu_device *adev,
WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
}
-static int ih_v6_1_set_powergating_state(void *handle,
+static int ih_v6_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (adev->pg_flags & AMD_PG_SUPPORT_IH_SRAM_PG)
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c
index d2428cf5d385..8f9b15c171f3 100644
--- a/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c
@@ -664,10 +664,10 @@ static void ih_v7_0_update_clockgating_state(struct amdgpu_device *adev,
return;
}
-static int ih_v7_0_set_clockgating_state(void *handle,
+static int ih_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
ih_v7_0_update_clockgating_state(adev,
state == AMD_CG_STATE_GATE);
@@ -727,10 +727,10 @@ static void ih_v7_0_update_ih_mem_power_gating(struct amdgpu_device *adev,
WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
}
-static int ih_v7_0_set_powergating_state(void *handle,
+static int ih_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (adev->pg_flags & AMD_PG_SUPPORT_IH_SRAM_PG)
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
index d4f72e47ae9e..aeca5c08ea2f 100644
--- a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
@@ -50,7 +50,8 @@ static int imu_v11_0_init_microcode(struct amdgpu_device *adev)
DRM_DEBUG("\n");
amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
- err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, "amdgpu/%s_imu.bin", ucode_prefix);
+ err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_imu.bin", ucode_prefix);
if (err)
goto out;
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c b/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c
index 1341f0292031..df898dbb746e 100644
--- a/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c
@@ -47,7 +47,8 @@ static int imu_v12_0_init_microcode(struct amdgpu_device *adev)
DRM_DEBUG("\n");
amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
- err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, "amdgpu/%s_imu.bin", ucode_prefix);
+ err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_imu.bin", ucode_prefix);
if (err)
goto out;
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
index 6e29b69894a5..7c9251c03815 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
@@ -35,7 +35,7 @@
static void jpeg_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v2_0_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v2_0_set_powergating_state(void *handle,
+static int jpeg_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
/**
@@ -154,7 +154,7 @@ static int jpeg_v2_0_hw_fini(struct amdgpu_ip_block *ip_block)
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS))
- jpeg_v2_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ jpeg_v2_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
return 0;
}
@@ -675,14 +675,14 @@ static int jpeg_v2_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return ret;
}
-static int jpeg_v2_0_set_clockgating_state(void *handle,
+static int jpeg_v2_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
if (enable) {
- if (!jpeg_v2_0_is_idle(handle))
+ if (!jpeg_v2_0_is_idle(adev))
return -EBUSY;
jpeg_v2_0_enable_clock_gating(adev);
} else {
@@ -692,10 +692,10 @@ static int jpeg_v2_0_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v2_0_set_powergating_state(void *handle,
+static int jpeg_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (state == adev->jpeg.cur_state)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
index 9ac421486f05..11f6af2646e7 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
@@ -38,7 +38,7 @@
static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v2_5_set_powergating_state(void *handle,
+static int jpeg_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static void jpeg_v2_5_set_ras_funcs(struct amdgpu_device *adev);
@@ -219,7 +219,7 @@ static int jpeg_v2_5_hw_fini(struct amdgpu_ip_block *ip_block)
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS))
- jpeg_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ jpeg_v2_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG))
amdgpu_irq_put(adev, &adev->jpeg.inst[i].ras_poison_irq, 0);
@@ -518,10 +518,10 @@ static int jpeg_v2_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int jpeg_v2_5_set_clockgating_state(void *handle,
+static int jpeg_v2_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
int i;
@@ -530,7 +530,7 @@ static int jpeg_v2_5_set_clockgating_state(void *handle,
continue;
if (enable) {
- if (!jpeg_v2_5_is_idle(handle))
+ if (!jpeg_v2_5_is_idle(adev))
return -EBUSY;
jpeg_v2_5_enable_clock_gating(adev, i);
} else {
@@ -541,10 +541,10 @@ static int jpeg_v2_5_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v2_5_set_powergating_state(void *handle,
+static int jpeg_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (state == adev->jpeg.cur_state)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
index e0df6800502c..4eca65ea9053 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
@@ -36,7 +36,7 @@
static void jpeg_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v3_0_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v3_0_set_powergating_state(void *handle,
+static int jpeg_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
/**
@@ -168,7 +168,7 @@ static int jpeg_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS))
- jpeg_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ jpeg_v3_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
return 0;
}
@@ -466,14 +466,14 @@ static int jpeg_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
}
-static int jpeg_v3_0_set_clockgating_state(void *handle,
+static int jpeg_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = state == AMD_CG_STATE_GATE;
if (enable) {
- if (!jpeg_v3_0_is_idle(handle))
+ if (!jpeg_v3_0_is_idle(adev))
return -EBUSY;
jpeg_v3_0_enable_clock_gating(adev);
} else {
@@ -483,10 +483,10 @@ static int jpeg_v3_0_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v3_0_set_powergating_state(void *handle,
+static int jpeg_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if(state == adev->jpeg.cur_state)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
index eca1963c33b6..0aef1f64afd0 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
@@ -39,7 +39,7 @@
static int jpeg_v4_0_start_sriov(struct amdgpu_device *adev);
static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v4_0_set_powergating_state(void *handle,
+static int jpeg_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev);
@@ -206,7 +206,7 @@ static int jpeg_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
if (!amdgpu_sriov_vf(adev)) {
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS))
- jpeg_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ jpeg_v4_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG))
amdgpu_irq_put(adev, &adev->jpeg.inst->ras_poison_irq, 0);
@@ -635,14 +635,14 @@ static int jpeg_v4_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
}
-static int jpeg_v4_0_set_clockgating_state(void *handle,
+static int jpeg_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = state == AMD_CG_STATE_GATE;
if (enable) {
- if (!jpeg_v4_0_is_idle(handle))
+ if (!jpeg_v4_0_is_idle(adev))
return -EBUSY;
jpeg_v4_0_enable_clock_gating(adev);
} else {
@@ -652,10 +652,10 @@ static int jpeg_v4_0_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v4_0_set_powergating_state(void *handle,
+static int jpeg_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (amdgpu_sriov_vf(adev)) {
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
index 67b51bcbacd1..88f9771c1686 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
@@ -43,7 +43,7 @@ enum jpeg_engin_status {
static void jpeg_v4_0_3_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v4_0_3_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v4_0_3_set_powergating_state(void *handle,
+static int jpeg_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static void jpeg_v4_0_3_set_ras_funcs(struct amdgpu_device *adev);
static void jpeg_v4_0_3_dec_ring_set_wptr(struct amdgpu_ring *ring);
@@ -76,7 +76,7 @@ static int jpeg_v4_0_3_early_init(struct amdgpu_ip_block *ip_block)
{
struct amdgpu_device *adev = ip_block->adev;
- adev->jpeg.num_jpeg_rings = AMDGPU_MAX_JPEG_RINGS;
+ adev->jpeg.num_jpeg_rings = AMDGPU_MAX_JPEG_RINGS_4_0_3;
jpeg_v4_0_3_set_dec_ring_funcs(adev);
jpeg_v4_0_3_set_irq_funcs(adev);
@@ -321,7 +321,7 @@ static int jpeg_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block)
if (r)
return r;
- for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ 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];
ring->wptr = 0;
@@ -379,7 +379,7 @@ static int jpeg_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block)
if (!amdgpu_sriov_vf(adev)) {
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE)
- ret = jpeg_v4_0_3_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ ret = jpeg_v4_0_3_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
return ret;
@@ -949,16 +949,16 @@ static int jpeg_v4_0_3_wait_for_idle(struct amdgpu_ip_block *ip_block)
return ret;
}
-static int jpeg_v4_0_3_set_clockgating_state(void *handle,
+static int jpeg_v4_0_3_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = state == AMD_CG_STATE_GATE;
int i;
for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
if (enable) {
- if (!jpeg_v4_0_3_is_idle(handle))
+ if (!jpeg_v4_0_3_is_idle(adev))
return -EBUSY;
jpeg_v4_0_3_enable_clock_gating(adev, i);
} else {
@@ -968,10 +968,10 @@ static int jpeg_v4_0_3_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v4_0_3_set_powergating_state(void *handle,
+static int jpeg_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (amdgpu_sriov_vf(adev)) {
@@ -1231,9 +1231,95 @@ static const struct amdgpu_ras_block_hw_ops jpeg_v4_0_3_ras_hw_ops = {
.reset_ras_error_count = jpeg_v4_0_3_reset_ras_error_count,
};
+static int jpeg_v4_0_3_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+ enum aca_smu_type type, void *data)
+{
+ struct aca_bank_info info;
+ u64 misc0;
+ int ret;
+
+ ret = aca_bank_info_decode(bank, &info);
+ if (ret)
+ return ret;
+
+ misc0 = bank->regs[ACA_REG_IDX_MISC0];
+ switch (type) {
+ case ACA_SMU_TYPE_UE:
+ ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+ 1ULL);
+ break;
+ case ACA_SMU_TYPE_CE:
+ ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE,
+ ACA_REG__MISC0__ERRCNT(misc0));
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return ret;
+}
+
+/* reference to smu driver if header file */
+static int jpeg_v4_0_3_err_codes[] = {
+ 16, 17, 18, 19, 20, 21, 22, 23, /* JPEG[0-7][S|D] */
+ 24, 25, 26, 27, 28, 29, 30, 31
+};
+
+static bool jpeg_v4_0_3_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+ enum aca_smu_type type, void *data)
+{
+ u32 instlo;
+
+ instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+ instlo &= GENMASK(31, 1);
+
+ if (instlo != mmSMNAID_AID0_MCA_SMU)
+ return false;
+
+ if (aca_bank_check_error_codes(handle->adev, bank,
+ jpeg_v4_0_3_err_codes,
+ ARRAY_SIZE(jpeg_v4_0_3_err_codes)))
+ return false;
+
+ return true;
+}
+
+static const struct aca_bank_ops jpeg_v4_0_3_aca_bank_ops = {
+ .aca_bank_parser = jpeg_v4_0_3_aca_bank_parser,
+ .aca_bank_is_valid = jpeg_v4_0_3_aca_bank_is_valid,
+};
+
+static const struct aca_info jpeg_v4_0_3_aca_info = {
+ .hwip = ACA_HWIP_TYPE_SMU,
+ .mask = ACA_ERROR_UE_MASK,
+ .bank_ops = &jpeg_v4_0_3_aca_bank_ops,
+};
+
+static int jpeg_v4_0_3_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+ int r;
+
+ r = amdgpu_ras_block_late_init(adev, ras_block);
+ if (r)
+ return r;
+
+ r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__JPEG,
+ &jpeg_v4_0_3_aca_info, NULL);
+ if (r)
+ goto late_fini;
+
+ return 0;
+
+late_fini:
+ amdgpu_ras_block_late_fini(adev, ras_block);
+
+ return r;
+}
+
static struct amdgpu_jpeg_ras jpeg_v4_0_3_ras = {
.ras_block = {
.hw_ops = &jpeg_v4_0_3_ras_hw_ops,
+ .ras_late_init = jpeg_v4_0_3_ras_late_init,
},
};
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
index 1d9e3b101c3a..6b3656984957 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
@@ -48,7 +48,7 @@
static void jpeg_v4_0_5_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v4_0_5_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v4_0_5_set_powergating_state(void *handle,
+static int jpeg_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static void jpeg_v4_0_5_dec_ring_set_wptr(struct amdgpu_ring *ring);
@@ -236,7 +236,7 @@ static int jpeg_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block)
if (!amdgpu_sriov_vf(adev)) {
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, i, regUVD_JRBC_STATUS))
- jpeg_v4_0_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ jpeg_v4_0_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
}
return 0;
@@ -660,10 +660,10 @@ static int jpeg_v4_0_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int jpeg_v4_0_5_set_clockgating_state(void *handle,
+static int jpeg_v4_0_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
int i;
@@ -672,7 +672,7 @@ static int jpeg_v4_0_5_set_clockgating_state(void *handle,
continue;
if (enable) {
- if (!jpeg_v4_0_5_is_idle(handle))
+ if (!jpeg_v4_0_5_is_idle(adev))
return -EBUSY;
jpeg_v4_0_5_enable_clock_gating(adev, i);
@@ -684,10 +684,10 @@ static int jpeg_v4_0_5_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v4_0_5_set_powergating_state(void *handle,
+static int jpeg_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (amdgpu_sriov_vf(adev)) {
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
index 58fb1e5fa89c..d5cf0f2799d4 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
@@ -31,12 +31,12 @@
#include "vcn/vcn_5_0_0_offset.h"
#include "vcn/vcn_5_0_0_sh_mask.h"
-#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_5_0.h"
#include "jpeg_v5_0_0.h"
static void jpeg_v5_0_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void jpeg_v5_0_0_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v5_0_0_set_powergating_state(void *handle,
+static int jpeg_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
/**
@@ -74,7 +74,7 @@ static int jpeg_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block)
/* JPEG TRAP */
r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
- VCN_4_0__SRCID__JPEG_DECODE, &adev->jpeg.inst->irq);
+ VCN_5_0__SRCID__JPEG_DECODE, &adev->jpeg.inst->irq);
if (r)
return r;
@@ -172,7 +172,7 @@ static int jpeg_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block)
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS))
- jpeg_v5_0_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ jpeg_v5_0_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
return 0;
}
@@ -560,14 +560,14 @@ static int jpeg_v5_0_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
}
-static int jpeg_v5_0_0_set_clockgating_state(void *handle,
+static int jpeg_v5_0_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
if (enable) {
- if (!jpeg_v5_0_0_is_idle(handle))
+ if (!jpeg_v5_0_0_is_idle(adev))
return -EBUSY;
jpeg_v5_0_0_enable_clock_gating(adev);
} else {
@@ -577,10 +577,10 @@ static int jpeg_v5_0_0_set_clockgating_state(void *handle,
return 0;
}
-static int jpeg_v5_0_0_set_powergating_state(void *handle,
+static int jpeg_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (state == adev->jpeg.cur_state)
@@ -612,7 +612,7 @@ static int jpeg_v5_0_0_process_interrupt(struct amdgpu_device *adev,
DRM_DEBUG("IH: JPEG TRAP\n");
switch (entry->src_id) {
- case VCN_4_0__SRCID__JPEG_DECODE:
+ case VCN_5_0__SRCID__JPEG_DECODE:
amdgpu_fence_process(adev->jpeg.inst->ring_dec);
break;
default:
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
new file mode 100644
index 000000000000..40d4c32a8c2a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
@@ -0,0 +1,708 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2014-2024 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * 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_jpeg.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "jpeg_v4_0_3.h"
+#include "jpeg_v5_0_1.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 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,
+ enum amd_powergating_state state);
+static void jpeg_v5_0_1_dec_ring_set_wptr(struct amdgpu_ring *ring);
+
+static int amdgpu_ih_srcid_jpeg[] = {
+ VCN_5_0__SRCID__JPEG_DECODE,
+ VCN_5_0__SRCID__JPEG1_DECODE,
+ VCN_5_0__SRCID__JPEG2_DECODE,
+ VCN_5_0__SRCID__JPEG3_DECODE,
+ VCN_5_0__SRCID__JPEG4_DECODE,
+ VCN_5_0__SRCID__JPEG5_DECODE,
+ VCN_5_0__SRCID__JPEG6_DECODE,
+ VCN_5_0__SRCID__JPEG7_DECODE,
+ VCN_5_0__SRCID__JPEG8_DECODE,
+ VCN_5_0__SRCID__JPEG9_DECODE,
+};
+
+static int jpeg_v5_0_1_core_reg_offset(u32 pipe)
+{
+ if (pipe <= AMDGPU_MAX_JPEG_RINGS_4_0_3)
+ return ((0x40 * pipe) - 0xc80);
+ else
+ return ((0x40 * pipe) - 0x440);
+}
+
+/**
+ * jpeg_v5_0_1_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int jpeg_v5_0_1_early_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+
+ if (!adev->jpeg.num_jpeg_inst || adev->jpeg.num_jpeg_inst > AMDGPU_MAX_JPEG_INSTANCES)
+ return -ENOENT;
+
+ adev->jpeg.num_jpeg_rings = AMDGPU_MAX_JPEG_RINGS;
+ jpeg_v5_0_1_set_dec_ring_funcs(adev);
+ jpeg_v5_0_1_set_irq_funcs(adev);
+
+ return 0;
+}
+
+/**
+ * jpeg_v5_0_1_sw_init - sw init for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int jpeg_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ring *ring;
+ int i, j, r, jpeg_inst;
+
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ /* JPEG TRAP */
+ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+ amdgpu_ih_srcid_jpeg[j], &adev->jpeg.inst->irq);
+ if (r)
+ return r;
+ }
+
+ r = amdgpu_jpeg_sw_init(adev);
+ if (r)
+ return r;
+
+ r = amdgpu_jpeg_resume(adev);
+ if (r)
+ return r;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ jpeg_inst = GET_INST(JPEG, i);
+
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ ring = &adev->jpeg.inst[i].ring_dec[j];
+ ring->use_doorbell = false;
+ 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;
+ }
+ 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,
+ AMDGPU_RING_PRIO_DEFAULT, NULL);
+ if (r)
+ return r;
+
+ adev->jpeg.internal.jpeg_pitch[j] =
+ regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET;
+ adev->jpeg.inst[i].external.jpeg_pitch[j] =
+ SOC15_REG_OFFSET1(JPEG, jpeg_inst, regUVD_JRBC_SCRATCH0,
+ (j ? jpeg_v5_0_1_core_reg_offset(j) : 0));
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * jpeg_v5_0_1_sw_fini - sw fini for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * JPEG suspend and free up sw allocation
+ */
+static int jpeg_v5_0_1_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int r;
+
+ r = amdgpu_jpeg_suspend(adev);
+ if (r)
+ return r;
+
+ r = amdgpu_jpeg_sw_fini(adev);
+
+ return r;
+}
+
+/**
+ * jpeg_v5_0_1_hw_init - start and test JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int jpeg_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ring *ring;
+ int i, j, r, jpeg_inst;
+
+ if (amdgpu_sriov_vf(adev)) {
+ /* jpeg_v5_0_1_start_sriov(adev); */
+ 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];
+ ring->wptr = 0;
+ ring->wptr_old = 0;
+ jpeg_v5_0_1_dec_ring_set_wptr(ring);
+ ring->sched.ready = true;
+ }
+ }
+ return 0;
+ }
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ jpeg_inst = GET_INST(JPEG, i);
+ ring = adev->jpeg.inst[i].ring_dec;
+ if (ring->use_doorbell)
+ adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+ (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 11 * jpeg_inst,
+ adev->jpeg.inst[i].aid_id);
+
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ 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->doorbell_index <<
+ VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+ VCN_JPEG_DB_CTRL__EN_MASK);
+ r = amdgpu_ring_test_helper(ring);
+ if (r)
+ return r;
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * jpeg_v5_0_1_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the JPEG block, mark ring as not ready any more
+ */
+static int jpeg_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int ret = 0;
+
+ 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);
+
+ return ret;
+}
+
+/**
+ * jpeg_v5_0_1_suspend - suspend JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend JPEG block
+ */
+static int jpeg_v5_0_1_suspend(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int r;
+
+ r = jpeg_v5_0_1_hw_fini(ip_block);
+ if (r)
+ return r;
+
+ r = amdgpu_jpeg_suspend(adev);
+
+ return r;
+}
+
+/**
+ * jpeg_v5_0_1_resume - resume JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init JPEG block
+ */
+static int jpeg_v5_0_1_resume(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int r;
+
+ r = amdgpu_jpeg_resume(adev);
+ if (r)
+ return r;
+
+ r = jpeg_v5_0_1_hw_init(ip_block);
+
+ return r;
+}
+
+static int jpeg_v5_0_1_disable_antihang(struct amdgpu_device *adev, int inst_idx)
+{
+ int jpeg_inst;
+
+ jpeg_inst = GET_INST(JPEG, inst_idx);
+ /* disable anti hang mechanism */
+ WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS), 0,
+ ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+ /* keep the JPEG in static PG mode */
+ WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS), 0,
+ ~UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK);
+
+ return 0;
+}
+
+static int jpeg_v5_0_1_enable_antihang(struct amdgpu_device *adev, int inst_idx)
+{
+ int jpeg_inst;
+
+ jpeg_inst = GET_INST(JPEG, inst_idx);
+ /* enable anti hang mechanism */
+ WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS),
+ UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+ ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+ return 0;
+}
+
+/**
+ * jpeg_v5_0_1_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v5_0_1_start(struct amdgpu_device *adev)
+{
+ struct amdgpu_ring *ring;
+ int i, j, jpeg_inst, r;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ jpeg_inst = GET_INST(JPEG, i);
+
+ /* disable antihang */
+ r = jpeg_v5_0_1_disable_antihang(adev, i);
+ if (r)
+ return r;
+
+ /* MJPEG global tiling registers */
+ WREG32_SOC15(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG,
+ adev->gfx.config.gb_addr_config);
+
+ /* enable JMI channel */
+ WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), 0,
+ ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ int reg_offset = (j ? jpeg_v5_0_1_core_reg_offset(j) : 0);
+ u32 reg, data, mask;
+
+ ring = &adev->jpeg.inst[i].ring_dec[j];
+
+ /* enable System Interrupt for JRBC */
+ reg = SOC15_REG_OFFSET(JPEG, jpeg_inst, regJPEG_SYS_INT_EN);
+ if (j < AMDGPU_MAX_JPEG_RINGS_4_0_3) {
+ data = JPEG_SYS_INT_EN__DJRBC0_MASK << j;
+ mask = ~(JPEG_SYS_INT_EN__DJRBC0_MASK << j);
+ WREG32_P(reg, data, mask);
+ } else {
+ data = JPEG_SYS_INT_EN__DJRBC0_MASK << (j+12);
+ mask = ~(JPEG_SYS_INT_EN__DJRBC0_MASK << (j+12));
+ WREG32_P(reg, data, mask);
+ }
+
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_LMI_JRBC_RB_VMID,
+ reg_offset, 0);
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_JRBC_RB_CNTL,
+ reg_offset,
+ (0x00000001L | 0x00000002L));
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+ reg_offset, lower_32_bits(ring->gpu_addr));
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+ reg_offset, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_JRBC_RB_RPTR,
+ reg_offset, 0);
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_JRBC_RB_WPTR,
+ reg_offset, 0);
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_JRBC_RB_CNTL,
+ reg_offset, 0x00000002L);
+ WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+ regUVD_JRBC_RB_SIZE,
+ reg_offset, ring->ring_size / 4);
+ ring->wptr = RREG32_SOC15_OFFSET(JPEG, jpeg_inst, regUVD_JRBC_RB_WPTR,
+ reg_offset);
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * jpeg_v5_0_1_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v5_0_1_stop(struct amdgpu_device *adev)
+{
+ int i, jpeg_inst, r;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ jpeg_inst = GET_INST(JPEG, i);
+ /* reset JMI */
+ WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL),
+ UVD_JMI_CNTL__SOFT_RESET_MASK,
+ ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+ /* enable antihang */
+ r = jpeg_v5_0_1_enable_antihang(adev, i);
+ if (r)
+ return r;
+ }
+
+ return 0;
+}
+
+/**
+ * jpeg_v5_0_1_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t jpeg_v5_0_1_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC_RB_RPTR,
+ ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0);
+}
+
+/**
+ * jpeg_v5_0_1_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t jpeg_v5_0_1_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring->use_doorbell)
+ return adev->wb.wb[ring->wptr_offs];
+
+ return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC_RB_WPTR,
+ ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0);
+}
+
+/**
+ * jpeg_v5_0_1_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void jpeg_v5_0_1_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring->use_doorbell) {
+ adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+ WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+ } else {
+ WREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me),
+ regUVD_JRBC_RB_WPTR,
+ (ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0),
+ lower_32_bits(ring->wptr));
+ }
+}
+
+static bool jpeg_v5_0_1_is_idle(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ bool ret = false;
+ int i, j;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ int reg_offset = (j ? jpeg_v5_0_1_core_reg_offset(j) : 0);
+
+ ret &= ((RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, i),
+ regUVD_JRBC_STATUS, reg_offset) &
+ UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
+ UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+ }
+ }
+
+ return ret;
+}
+
+static int jpeg_v5_0_1_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int ret = 0;
+ int i, j;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ int reg_offset = (j ? jpeg_v5_0_1_core_reg_offset(j) : 0);
+
+ ret &= SOC15_WAIT_ON_RREG_OFFSET(JPEG, GET_INST(JPEG, i),
+ regUVD_JRBC_STATUS, reg_offset,
+ UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
+ UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+ }
+ }
+ return ret;
+}
+
+static int jpeg_v5_0_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_clockgating_state state)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
+
+ int i;
+
+ if (!enable)
+ return 0;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ if (!jpeg_v5_0_1_is_idle(adev))
+ return -EBUSY;
+ }
+
+ return 0;
+}
+
+static int jpeg_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int ret;
+
+ if (state == adev->jpeg.cur_state)
+ return 0;
+
+ if (state == AMD_PG_STATE_GATE)
+ ret = jpeg_v5_0_1_stop(adev);
+ else
+ ret = jpeg_v5_0_1_start(adev);
+
+ if (!ret)
+ adev->jpeg.cur_state = state;
+
+ return ret;
+}
+
+static int jpeg_v5_0_1_set_interrupt_state(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *source,
+ unsigned int type,
+ enum amdgpu_interrupt_state state)
+{
+ return 0;
+}
+
+static int jpeg_v5_0_1_process_interrupt(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *source,
+ struct amdgpu_iv_entry *entry)
+{
+ u32 i, inst;
+
+ i = node_id_to_phys_map[entry->node_id];
+ DRM_DEV_DEBUG(adev->dev, "IH: JPEG TRAP\n");
+
+ for (inst = 0; inst < adev->jpeg.num_jpeg_inst; ++inst)
+ if (adev->jpeg.inst[inst].aid_id == i)
+ break;
+
+ if (inst >= adev->jpeg.num_jpeg_inst) {
+ dev_WARN_ONCE(adev->dev, 1,
+ "Interrupt received for unknown JPEG instance %d",
+ entry->node_id);
+ return 0;
+ }
+
+ switch (entry->src_id) {
+ case VCN_5_0__SRCID__JPEG_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[0]);
+ break;
+ case VCN_5_0__SRCID__JPEG1_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[1]);
+ break;
+ case VCN_5_0__SRCID__JPEG2_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[2]);
+ break;
+ case VCN_5_0__SRCID__JPEG3_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[3]);
+ break;
+ case VCN_5_0__SRCID__JPEG4_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[4]);
+ break;
+ case VCN_5_0__SRCID__JPEG5_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[5]);
+ break;
+ case VCN_5_0__SRCID__JPEG6_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[6]);
+ break;
+ case VCN_5_0__SRCID__JPEG7_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[7]);
+ break;
+ case VCN_5_0__SRCID__JPEG8_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[8]);
+ break;
+ case VCN_5_0__SRCID__JPEG9_DECODE:
+ amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[9]);
+ break;
+ default:
+ DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+ entry->src_id, entry->src_data[0]);
+ break;
+ }
+
+ return 0;
+}
+
+static const struct amd_ip_funcs jpeg_v5_0_1_ip_funcs = {
+ .name = "jpeg_v5_0_1",
+ .early_init = jpeg_v5_0_1_early_init,
+ .late_init = NULL,
+ .sw_init = jpeg_v5_0_1_sw_init,
+ .sw_fini = jpeg_v5_0_1_sw_fini,
+ .hw_init = jpeg_v5_0_1_hw_init,
+ .hw_fini = jpeg_v5_0_1_hw_fini,
+ .suspend = jpeg_v5_0_1_suspend,
+ .resume = jpeg_v5_0_1_resume,
+ .is_idle = jpeg_v5_0_1_is_idle,
+ .wait_for_idle = jpeg_v5_0_1_wait_for_idle,
+ .check_soft_reset = NULL,
+ .pre_soft_reset = NULL,
+ .soft_reset = NULL,
+ .post_soft_reset = NULL,
+ .set_clockgating_state = jpeg_v5_0_1_set_clockgating_state,
+ .set_powergating_state = jpeg_v5_0_1_set_powergating_state,
+ .dump_ip_state = NULL,
+ .print_ip_state = NULL,
+};
+
+static const struct amdgpu_ring_funcs jpeg_v5_0_1_dec_ring_vm_funcs = {
+ .type = AMDGPU_RING_TYPE_VCN_JPEG,
+ .align_mask = 0xf,
+ .get_rptr = jpeg_v5_0_1_dec_ring_get_rptr,
+ .get_wptr = jpeg_v5_0_1_dec_ring_get_wptr,
+ .set_wptr = jpeg_v5_0_1_dec_ring_set_wptr,
+ .emit_frame_size =
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+ 8 + /* jpeg_v5_0_1_dec_ring_emit_vm_flush */
+ 22 + 22 + /* jpeg_v5_0_1_dec_ring_emit_fence x2 vm fence */
+ 8 + 16,
+ .emit_ib_size = 22, /* jpeg_v5_0_1_dec_ring_emit_ib */
+ .emit_ib = jpeg_v4_0_3_dec_ring_emit_ib,
+ .emit_fence = jpeg_v4_0_3_dec_ring_emit_fence,
+ .emit_vm_flush = jpeg_v4_0_3_dec_ring_emit_vm_flush,
+ .test_ring = amdgpu_jpeg_dec_ring_test_ring,
+ .test_ib = amdgpu_jpeg_dec_ring_test_ib,
+ .insert_nop = jpeg_v4_0_3_dec_ring_nop,
+ .insert_start = jpeg_v4_0_3_dec_ring_insert_start,
+ .insert_end = jpeg_v4_0_3_dec_ring_insert_end,
+ .pad_ib = amdgpu_ring_generic_pad_ib,
+ .begin_use = amdgpu_jpeg_ring_begin_use,
+ .end_use = amdgpu_jpeg_ring_end_use,
+ .emit_wreg = jpeg_v4_0_3_dec_ring_emit_wreg,
+ .emit_reg_wait = jpeg_v4_0_3_dec_ring_emit_reg_wait,
+ .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
+static void jpeg_v5_0_1_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+ int i, j, jpeg_inst;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+ for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+ adev->jpeg.inst[i].ring_dec[j].funcs = &jpeg_v5_0_1_dec_ring_vm_funcs;
+ adev->jpeg.inst[i].ring_dec[j].me = i;
+ adev->jpeg.inst[i].ring_dec[j].pipe = j;
+ }
+ jpeg_inst = GET_INST(JPEG, i);
+ adev->jpeg.inst[i].aid_id =
+ jpeg_inst / adev->jpeg.num_inst_per_aid;
+ }
+}
+
+static const struct amdgpu_irq_src_funcs jpeg_v5_0_1_irq_funcs = {
+ .set = jpeg_v5_0_1_set_interrupt_state,
+ .process = jpeg_v5_0_1_process_interrupt,
+};
+
+static void jpeg_v5_0_1_set_irq_funcs(struct amdgpu_device *adev)
+{
+ int i;
+
+ for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i)
+ adev->jpeg.inst->irq.num_types += adev->jpeg.num_jpeg_rings;
+
+ adev->jpeg.inst->irq.funcs = &jpeg_v5_0_1_irq_funcs;
+}
+
+const struct amdgpu_ip_block_version jpeg_v5_0_1_ip_block = {
+ .type = AMD_IP_BLOCK_TYPE_JPEG,
+ .major = 5,
+ .minor = 0,
+ .rev = 1,
+ .funcs = &jpeg_v5_0_1_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h
new file mode 100644
index 000000000000..8ce146c00bb6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h
@@ -0,0 +1,29 @@
+/*
+ * 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.
+ *
+ */
+
+#ifndef __JPEG_V5_0_1_H__
+#define __JPEG_V5_0_1_H__
+
+extern const struct amdgpu_ip_block_version jpeg_v5_0_1_ip_block;
+
+#endif /* __JPEG_V5_0_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index 9c905b9e9376..f9a4d08eef92 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -1505,9 +1505,7 @@ static void mes_v11_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
}
static void mes_v11_0_kiq_clear(struct amdgpu_device *adev)
@@ -1635,6 +1633,10 @@ static int mes_v11_0_hw_init(struct amdgpu_ip_block *ip_block)
goto failure;
}
+ r = amdgpu_mes_update_enforce_isolation(adev);
+ if (r)
+ goto failure;
+
out:
/*
* Disable KIQ ring usage from the driver once MES is enabled.
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
index 9ecc5d61e49b..0fd0fa6ed518 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
@@ -24,6 +24,7 @@
#include <linux/firmware.h>
#include <linux/module.h>
#include "amdgpu.h"
+#include "gfx_v12_0.h"
#include "soc15_common.h"
#include "soc21.h"
#include "gc/gc_12_0_0_offset.h"
@@ -350,6 +351,132 @@ static int mes_v12_0_remove_hw_queue(struct amdgpu_mes *mes,
offsetof(union MESAPI__REMOVE_QUEUE, api_status));
}
+int gfx_v12_0_request_gfx_index_mutex(struct amdgpu_device *adev,
+ bool req)
+{
+ u32 i, tmp, val;
+
+ for (i = 0; i < adev->usec_timeout; i++) {
+ /* Request with MeId=2, PipeId=0 */
+ tmp = REG_SET_FIELD(0, CP_GFX_INDEX_MUTEX, REQUEST, req);
+ tmp = REG_SET_FIELD(tmp, CP_GFX_INDEX_MUTEX, CLIENTID, 4);
+ WREG32_SOC15(GC, 0, regCP_GFX_INDEX_MUTEX, tmp);
+
+ val = RREG32_SOC15(GC, 0, regCP_GFX_INDEX_MUTEX);
+ if (req) {
+ if (val == tmp)
+ break;
+ } else {
+ tmp = REG_SET_FIELD(tmp, CP_GFX_INDEX_MUTEX,
+ REQUEST, 1);
+
+ /* unlocked or locked by firmware */
+ if (val != tmp)
+ break;
+ }
+ udelay(1);
+ }
+
+ if (i >= adev->usec_timeout)
+ return -EINVAL;
+
+ return 0;
+}
+
+static int mes_v12_0_reset_queue_mmio(struct amdgpu_mes *mes, uint32_t queue_type,
+ uint32_t me_id, uint32_t pipe_id,
+ uint32_t queue_id, uint32_t vmid)
+{
+ struct amdgpu_device *adev = mes->adev;
+ uint32_t value, reg;
+ int i, r = 0;
+
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+ if (queue_type == AMDGPU_RING_TYPE_GFX) {
+ dev_info(adev->dev, "reset gfx queue (%d:%d:%d: vmid:%d)\n",
+ me_id, pipe_id, queue_id, vmid);
+
+ mutex_lock(&adev->gfx.reset_sem_mutex);
+ gfx_v12_0_request_gfx_index_mutex(adev, true);
+ /* all se allow writes */
+ WREG32_SOC15(GC, 0, regGRBM_GFX_INDEX,
+ (uint32_t)(0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT));
+ value = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+ if (pipe_id == 0)
+ value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE0_QUEUES, 1 << queue_id);
+ else
+ value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE1_QUEUES, 1 << queue_id);
+ WREG32_SOC15(GC, 0, regCP_VMID_RESET, value);
+ gfx_v12_0_request_gfx_index_mutex(adev, false);
+ mutex_unlock(&adev->gfx.reset_sem_mutex);
+
+ mutex_lock(&adev->srbm_mutex);
+ soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+ /* wait till dequeue take effects */
+ for (i = 0; i < adev->usec_timeout; i++) {
+ if (!(RREG32_SOC15(GC, 0, regCP_GFX_HQD_ACTIVE) & 1))
+ break;
+ udelay(1);
+ }
+ if (i >= adev->usec_timeout) {
+ dev_err(adev->dev, "failed to wait on gfx hqd deactivate\n");
+ r = -ETIMEDOUT;
+ }
+
+ soc21_grbm_select(adev, 0, 0, 0, 0);
+ mutex_unlock(&adev->srbm_mutex);
+ } else if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
+ dev_info(adev->dev, "reset compute queue (%d:%d:%d)\n",
+ me_id, pipe_id, queue_id);
+ mutex_lock(&adev->srbm_mutex);
+ soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+ WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2);
+ WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1);
+
+ /* wait till dequeue take effects */
+ for (i = 0; i < adev->usec_timeout; i++) {
+ if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+ break;
+ udelay(1);
+ }
+ if (i >= adev->usec_timeout) {
+ dev_err(adev->dev, "failed to wait on hqd deactivate\n");
+ r = -ETIMEDOUT;
+ }
+ soc21_grbm_select(adev, 0, 0, 0, 0);
+ mutex_unlock(&adev->srbm_mutex);
+ } else if (queue_type == AMDGPU_RING_TYPE_SDMA) {
+ dev_info(adev->dev, "reset sdma queue (%d:%d:%d)\n",
+ me_id, pipe_id, queue_id);
+ switch (me_id) {
+ case 1:
+ reg = SOC15_REG_OFFSET(GC, 0, regSDMA1_QUEUE_RESET_REQ);
+ break;
+ case 0:
+ default:
+ reg = SOC15_REG_OFFSET(GC, 0, regSDMA0_QUEUE_RESET_REQ);
+ break;
+ }
+
+ value = 1 << queue_id;
+ WREG32(reg, value);
+ /* wait for queue reset done */
+ for (i = 0; i < adev->usec_timeout; i++) {
+ if (!(RREG32(reg) & value))
+ break;
+ udelay(1);
+ }
+ if (i >= adev->usec_timeout) {
+ dev_err(adev->dev, "failed to wait on sdma queue reset done\n");
+ r = -ETIMEDOUT;
+ }
+ }
+
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+ return r;
+}
+
static int mes_v12_0_reset_hw_queue(struct amdgpu_mes *mes,
struct mes_reset_queue_input *input)
{
@@ -629,7 +756,8 @@ static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes, int pipe)
if (amdgpu_mes_log_enable) {
mes_set_hw_res_pkt.enable_mes_event_int_logging = 1;
- mes_set_hw_res_pkt.event_intr_history_gpu_mc_ptr = mes->event_log_gpu_addr + pipe * AMDGPU_MES_LOG_BUFFER_SIZE;
+ mes_set_hw_res_pkt.event_intr_history_gpu_mc_ptr = mes->event_log_gpu_addr +
+ pipe * (AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE);
}
if (enforce_isolation)
@@ -721,6 +849,11 @@ static int mes_v12_0_reset_legacy_queue(struct amdgpu_mes *mes,
union MESAPI__RESET mes_reset_queue_pkt;
int pipe;
+ if (input->use_mmio)
+ return mes_v12_0_reset_queue_mmio(mes, input->queue_type,
+ input->me_id, input->pipe_id,
+ input->queue_id, input->vmid);
+
memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt));
mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
@@ -851,29 +984,50 @@ static void mes_v12_0_enable(struct amdgpu_device *adev, bool enable)
uint32_t pipe, data = 0;
if (enable) {
- data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
- data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
- data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET, 1);
- WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
-
mutex_lock(&adev->srbm_mutex);
for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
soc21_grbm_select(adev, 3, pipe, 0, 0);
+ if (amdgpu_mes_log_enable) {
+ u32 log_size = AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE;
+ /* In case uni mes is not enabled, only program for pipe 0 */
+ if (adev->mes.event_log_size >= (pipe + 1) * log_size) {
+ WREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_LO,
+ lower_32_bits(adev->mes.event_log_gpu_addr +
+ pipe * log_size + AMDGPU_MES_LOG_BUFFER_SIZE));
+ WREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_HI,
+ upper_32_bits(adev->mes.event_log_gpu_addr +
+ pipe * log_size + AMDGPU_MES_LOG_BUFFER_SIZE));
+ dev_info(adev->dev, "Setup CP MES MSCRATCH address : 0x%x. 0x%x\n",
+ RREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_HI),
+ RREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_LO));
+ }
+ }
+
+ data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
+ if (pipe == 0)
+ data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
+ else
+ data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET, 1);
+ WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
ucode_addr = adev->mes.uc_start_addr[pipe] >> 2;
WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START,
lower_32_bits(ucode_addr));
WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START_HI,
upper_32_bits(ucode_addr));
+
+ /* unhalt MES and activate one pipe each loop */
+ data = REG_SET_FIELD(0, CP_MES_CNTL, MES_PIPE0_ACTIVE, 1);
+ if (pipe)
+ data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE, 1);
+ dev_info(adev->dev, "program CP_MES_CNTL : 0x%x\n", data);
+
+ WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+
}
soc21_grbm_select(adev, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
- /* unhalt MES and activate pipe0 */
- data = REG_SET_FIELD(0, CP_MES_CNTL, MES_PIPE0_ACTIVE, 1);
- data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE, 1);
- WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
-
if (amdgpu_emu_mode)
msleep(100);
else if (adev->enable_uni_mes)
@@ -1347,8 +1501,9 @@ static int mes_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
adev->mes.kiq_hw_fini = &mes_v12_0_kiq_hw_fini;
adev->mes.enable_legacy_queue_map = true;
- adev->mes.event_log_size = adev->enable_uni_mes ? (AMDGPU_MAX_MES_PIPES * AMDGPU_MES_LOG_BUFFER_SIZE) : AMDGPU_MES_LOG_BUFFER_SIZE;
-
+ adev->mes.event_log_size = adev->enable_uni_mes ?
+ (AMDGPU_MAX_MES_PIPES * (AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE)) :
+ (AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE);
r = amdgpu_mes_init(adev);
if (r)
return r;
@@ -1455,9 +1610,7 @@ static void mes_v12_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
}
static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev)
@@ -1590,6 +1743,10 @@ static int mes_v12_0_hw_init(struct amdgpu_ip_block *ip_block)
goto failure;
}
+ r = amdgpu_mes_update_enforce_isolation(adev);
+ if (r)
+ goto failure;
+
out:
/*
* Disable KIQ ring usage from the driver once MES is enabled.
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
index e9a6f33ca710..243eabda0607 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
@@ -356,7 +356,7 @@ static void mmhub_v1_0_update_power_gating(struct amdgpu_device *adev,
if (adev->pg_flags & AMD_PG_SUPPORT_MMHUB)
amdgpu_dpm_set_powergating_by_smu(adev,
AMD_IP_BLOCK_TYPE_GMC,
- enable);
+ enable, 0);
}
static int mmhub_v1_0_gart_enable(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
index b01bb759d0f4..e646e5cef0a2 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
@@ -33,7 +33,6 @@
#define regVM_L2_CNTL3_DEFAULT 0x80100007
#define regVM_L2_CNTL4_DEFAULT 0x000000c1
-#define mmSMNAID_AID0_MCA_SMU 0x03b30400
static u64 mmhub_v1_8_get_fb_location(struct amdgpu_device *adev)
{
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
index 0820ed62e2e8..62cdfe10e6f4 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
@@ -434,9 +434,8 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev,
* this should allow us to catch up.
*/
tmp = (wptr + 32) & ih->ptr_mask;
- dev_warn(adev->dev, "IH ring buffer overflow "
- "(0x%08X, 0x%08X, 0x%08X)\n",
- wptr, ih->rptr, tmp);
+ dev_warn(adev->dev, "%s ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+ amdgpu_ih_ring_name(adev, ih), wptr, ih->rptr, tmp);
ih->rptr = tmp;
tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
@@ -667,17 +666,17 @@ static void navi10_ih_update_clockgating_state(struct amdgpu_device *adev,
}
}
-static int navi10_ih_set_clockgating_state(void *handle,
+static int navi10_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
navi10_ih_update_clockgating_state(adev,
state == AMD_CG_STATE_GATE);
return 0;
}
-static int navi10_ih_set_powergating_state(void *handle,
+static int navi10_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c
index 39919e0892c1..c92875ceb31f 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c
@@ -28,6 +28,7 @@
#include "nbif/nbif_6_3_1_sh_mask.h"
#include "pcie/pcie_6_1_0_offset.h"
#include "pcie/pcie_6_1_0_sh_mask.h"
+#include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
#include <uapi/linux/kfd_ioctl.h>
static void nbif_v6_3_1_remap_hdp_registers(struct amdgpu_device *adev)
@@ -518,3 +519,83 @@ const struct amdgpu_nbio_funcs nbif_v6_3_1_sriov_funcs = {
.get_rom_offset = nbif_v6_3_1_get_rom_offset,
.set_reg_remap = nbif_v6_3_1_set_reg_remap,
};
+
+static int nbif_v6_3_1_set_ras_err_event_athub_irq_state(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *src,
+ unsigned type,
+ enum amdgpu_interrupt_state state)
+{
+ /* The ras_controller_irq enablement should be done in psp bl when it
+ * tries to enable ras feature. Driver only need to set the correct interrupt
+ * vector for bare-metal and sriov use case respectively
+ */
+ uint32_t bif_doorbell_int_cntl;
+
+ bif_doorbell_int_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+ bif_doorbell_int_cntl = REG_SET_FIELD(bif_doorbell_int_cntl,
+ BIF_BX0_BIF_DOORBELL_INT_CNTL,
+ RAS_ATHUB_ERR_EVENT_INTERRUPT_DISABLE,
+ (state == AMDGPU_IRQ_STATE_ENABLE) ? 0 : 1);
+ WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_int_cntl);
+
+ return 0;
+}
+
+static int nbif_v6_3_1_process_err_event_athub_irq(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *source,
+ struct amdgpu_iv_entry *entry)
+{
+ /* By design, the ih cookie for err_event_athub_irq should be written
+ * to bif ring. since bif ring is not enabled, just leave process callback
+ * as a dummy one.
+ */
+ return 0;
+}
+
+static const struct amdgpu_irq_src_funcs nbif_v6_3_1_ras_err_event_athub_irq_funcs = {
+ .set = nbif_v6_3_1_set_ras_err_event_athub_irq_state,
+ .process = nbif_v6_3_1_process_err_event_athub_irq,
+};
+
+static void nbif_v6_3_1_handle_ras_err_event_athub_intr_no_bifring(struct amdgpu_device *adev)
+{
+ uint32_t bif_doorbell_int_cntl;
+
+ bif_doorbell_int_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+ if (REG_GET_FIELD(bif_doorbell_int_cntl,
+ BIF_BX0_BIF_DOORBELL_INT_CNTL,
+ RAS_ATHUB_ERR_EVENT_INTERRUPT_STATUS)) {
+ /* driver has to clear the interrupt status when bif ring is disabled */
+ bif_doorbell_int_cntl = REG_SET_FIELD(bif_doorbell_int_cntl,
+ BIF_BX0_BIF_DOORBELL_INT_CNTL,
+ RAS_ATHUB_ERR_EVENT_INTERRUPT_CLEAR, 1);
+ WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_int_cntl);
+ amdgpu_ras_global_ras_isr(adev);
+ }
+}
+
+static int nbif_v6_3_1_init_ras_err_event_athub_interrupt(struct amdgpu_device *adev)
+{
+ int r;
+
+ /* init the irq funcs */
+ adev->nbio.ras_err_event_athub_irq.funcs =
+ &nbif_v6_3_1_ras_err_event_athub_irq_funcs;
+ adev->nbio.ras_err_event_athub_irq.num_types = 1;
+
+ /* register ras err event athub interrupt
+ * nbif v6_3_1 uses the same irq source as nbio v7_4
+ */
+ r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_BIF,
+ NBIF_7_4__SRCID__ERREVENT_ATHUB_INTERRUPT,
+ &adev->nbio.ras_err_event_athub_irq);
+
+ return r;
+}
+
+struct amdgpu_nbio_ras nbif_v6_3_1_ras = {
+ .handle_ras_err_event_athub_intr_no_bifring =
+ nbif_v6_3_1_handle_ras_err_event_athub_intr_no_bifring,
+ .init_ras_err_event_athub_interrupt =
+ nbif_v6_3_1_init_ras_err_event_athub_interrupt,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h
index b7f2e0d88905..9ac4831d39e1 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h
+++ b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h
@@ -29,5 +29,6 @@
extern const struct nbio_hdp_flush_reg nbif_v6_3_1_hdp_flush_reg;
extern const struct amdgpu_nbio_funcs nbif_v6_3_1_funcs;
extern const struct amdgpu_nbio_funcs nbif_v6_3_1_sriov_funcs;
+extern struct amdgpu_nbio_ras nbif_v6_3_1_ras;
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 3bad565ded73..95c609317a8d 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -78,12 +78,12 @@ static const struct amdgpu_video_codecs nv_video_codecs_encode = {
/* Navi1x */
static const struct amdgpu_video_codec_info nv_video_codecs_decode_array[] = {
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 8192, 8192, 0)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
};
@@ -104,10 +104,10 @@ static const struct amdgpu_video_codecs sc_video_codecs_encode = {
};
static const struct amdgpu_video_codec_info sc_video_codecs_decode_array_vcn0[] = {
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
@@ -115,10 +115,10 @@ static const struct amdgpu_video_codec_info sc_video_codecs_decode_array_vcn0[]
};
static const struct amdgpu_video_codec_info sc_video_codecs_decode_array_vcn1[] = {
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
@@ -1039,10 +1039,10 @@ static bool nv_common_is_idle(void *handle)
return true;
}
-static int nv_common_set_clockgating_state(void *handle,
+static int nv_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1070,7 +1070,7 @@ static int nv_common_set_clockgating_state(void *handle,
return 0;
}
-static int nv_common_set_powergating_state(void *handle,
+static int nv_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* TODO */
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
index c4b775aaee9f..cc621064610f 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
@@ -51,6 +51,8 @@ MODULE_FIRMWARE("amdgpu/psp_13_0_11_toc.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_11_ta.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_6_sos.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_6_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_12_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_12_ta.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_14_sos.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_14_ta.bin");
MODULE_FIRMWARE("amdgpu/psp_14_0_0_toc.bin");
@@ -122,6 +124,7 @@ static int psp_v13_0_init_microcode(struct psp_context *psp)
case IP_VERSION(13, 0, 6):
case IP_VERSION(13, 0, 7):
case IP_VERSION(13, 0, 10):
+ case IP_VERSION(13, 0, 12):
case IP_VERSION(13, 0, 14):
err = psp_init_sos_microcode(psp, ucode_prefix);
if (err)
@@ -177,6 +180,7 @@ static int psp_v13_0_wait_for_bootloader(struct psp_context *psp)
retry_cnt =
((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))) ?
PSP_VMBX_POLLING_LIMIT :
10;
@@ -203,6 +207,7 @@ static int psp_v13_0_wait_for_bootloader_steady_state(struct psp_context *psp)
int ret;
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)) {
ret = psp_v13_0_wait_for_vmbx_ready(psp);
if (ret)
@@ -288,6 +293,11 @@ static int psp_v13_0_bootloader_load_ras_drv(struct psp_context *psp)
return psp_v13_0_bootloader_load_component(psp, &psp->ras_drv, PSP_BL__LOAD_RASDRV);
}
+static int psp_v13_0_bootloader_load_spdm_drv(struct psp_context *psp)
+{
+ return psp_v13_0_bootloader_load_component(psp, &psp->spdm_drv, PSP_BL__LOAD_SPDMDRV);
+}
+
static inline void psp_v13_0_init_sos_version(struct psp_context *psp)
{
struct amdgpu_device *adev = psp->adev;
@@ -798,6 +808,7 @@ static bool psp_v13_0_get_ras_capability(struct psp_context *psp)
return false;
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)) &&
(!(adev->flags & AMD_IS_APU))) {
reg_data = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_127);
@@ -857,6 +868,7 @@ static const struct psp_funcs psp_v13_0_funcs = {
.bootloader_load_intf_drv = psp_v13_0_bootloader_load_intf_drv,
.bootloader_load_dbg_drv = psp_v13_0_bootloader_load_dbg_drv,
.bootloader_load_ras_drv = psp_v13_0_bootloader_load_ras_drv,
+ .bootloader_load_spdm_drv = psp_v13_0_bootloader_load_spdm_drv,
.bootloader_load_sos = psp_v13_0_bootloader_load_sos,
.ring_create = psp_v13_0_ring_create,
.ring_stop = psp_v13_0_ring_stop,
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
index 7948d74f8722..135c5099bfb8 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
@@ -145,9 +145,11 @@ static int sdma_v2_4_init_microcode(struct amdgpu_device *adev)
for (i = 0; i < adev->sdma.num_instances; i++) {
if (i == 0)
err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sdma.bin", chip_name);
else
err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sdma1.bin", chip_name);
if (err)
goto out;
@@ -631,7 +633,7 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
@@ -1080,14 +1082,14 @@ static int sdma_v2_4_process_illegal_inst_irq(struct amdgpu_device *adev,
return 0;
}
-static int sdma_v2_4_set_clockgating_state(void *handle,
+static int sdma_v2_4_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
/* XXX handled via the smc on VI */
return 0;
}
-static int sdma_v2_4_set_powergating_state(void *handle,
+static int sdma_v2_4_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
index 9a3d729545a7..c611328671ed 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
@@ -305,9 +305,11 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev)
for (i = 0; i < adev->sdma.num_instances; i++) {
if (i == 0)
err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sdma.bin", chip_name);
else
err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+ AMDGPU_UCODE_REQUIRED,
"amdgpu/%s_sdma1.bin", chip_name);
if (err)
goto out;
@@ -904,7 +906,7 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
else
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
@@ -1483,10 +1485,10 @@ static void sdma_v3_0_update_sdma_medium_grain_light_sleep(
}
}
-static int sdma_v3_0_set_clockgating_state(void *handle,
+static int sdma_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1506,7 +1508,7 @@ static int sdma_v3_0_set_clockgating_state(void *handle,
return 0;
}
-static int sdma_v3_0_set_powergating_state(void *handle,
+static int sdma_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index c1f98f6cf20d..b48d9c0b2e1c 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -1565,7 +1565,7 @@ static int sdma_v4_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
@@ -1956,7 +1956,7 @@ static int sdma_v4_0_hw_init(struct amdgpu_ip_block *ip_block)
struct amdgpu_device *adev = ip_block->adev;
if (adev->flags & AMD_IS_APU)
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false, 0);
if (!amdgpu_sriov_vf(adev))
sdma_v4_0_init_golden_registers(adev);
@@ -1983,7 +1983,7 @@ static int sdma_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
sdma_v4_0_enable(adev, false);
if (adev->flags & AMD_IS_APU)
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, true);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, true, 0);
return 0;
}
@@ -2297,10 +2297,10 @@ static void sdma_v4_0_update_medium_grain_light_sleep(
}
}
-static int sdma_v4_0_set_clockgating_state(void *handle,
+static int sdma_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -2312,10 +2312,10 @@ static int sdma_v4_0_set_clockgating_state(void *handle,
return 0;
}
-static int sdma_v4_0_set_powergating_state(void *handle,
+static int sdma_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
case IP_VERSION(4, 1, 0):
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 a38553f38fdc..5e0066cd6c51 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
@@ -189,6 +189,7 @@ static int sdma_v4_4_2_init_microcode(struct amdgpu_device *adev)
for (i = 0; i < adev->sdma.num_instances; i++) {
if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+ amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 4) ||
amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 5)) {
ret = amdgpu_sdma_init_microcode(adev, 0, true);
break;
@@ -667,11 +668,12 @@ static uint32_t sdma_v4_4_2_rb_cntl(struct amdgpu_ring *ring, uint32_t rb_cntl)
*
* @adev: amdgpu_device pointer
* @i: instance to resume
+ * @restore: used to restore wptr when restart
*
* Set up the gfx DMA ring buffers and enable them.
* Returns 0 for success, error for failure.
*/
-static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i)
+static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i, bool restore)
{
struct amdgpu_ring *ring = &adev->sdma.instance[i].ring;
u32 rb_cntl, ib_cntl, wptr_poll_cntl;
@@ -698,16 +700,24 @@ static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i)
WREG32_SDMA(i, regSDMA_GFX_RB_BASE, ring->gpu_addr >> 8);
WREG32_SDMA(i, regSDMA_GFX_RB_BASE_HI, ring->gpu_addr >> 40);
- ring->wptr = 0;
+ if (!restore)
+ ring->wptr = 0;
/* before programing wptr to a less value, need set minor_ptr_update first */
WREG32_SDMA(i, regSDMA_GFX_MINOR_PTR_UPDATE, 1);
/* Initialize the ring buffer's read and write pointers */
- WREG32_SDMA(i, regSDMA_GFX_RB_RPTR, 0);
- WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_HI, 0);
- WREG32_SDMA(i, regSDMA_GFX_RB_WPTR, 0);
- WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_HI, 0);
+ if (restore) {
+ WREG32_SDMA(i, regSDMA_GFX_RB_RPTR, lower_32_bits(ring->wptr << 2));
+ WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_HI, upper_32_bits(ring->wptr << 2));
+ WREG32_SDMA(i, regSDMA_GFX_RB_WPTR, lower_32_bits(ring->wptr << 2));
+ WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_HI, upper_32_bits(ring->wptr << 2));
+ } else {
+ WREG32_SDMA(i, regSDMA_GFX_RB_RPTR, 0);
+ WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_HI, 0);
+ WREG32_SDMA(i, regSDMA_GFX_RB_WPTR, 0);
+ WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_HI, 0);
+ }
doorbell = RREG32_SDMA(i, regSDMA_GFX_DOORBELL);
doorbell_offset = RREG32_SDMA(i, regSDMA_GFX_DOORBELL_OFFSET);
@@ -755,11 +765,12 @@ static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i)
*
* @adev: amdgpu_device pointer
* @i: instance to resume
+ * @restore: boolean to say restore needed or not
*
* Set up the page DMA ring buffers and enable them.
* Returns 0 for success, error for failure.
*/
-static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i)
+static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i, bool restore)
{
struct amdgpu_ring *ring = &adev->sdma.instance[i].page;
u32 rb_cntl, ib_cntl, wptr_poll_cntl;
@@ -775,10 +786,17 @@ static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i)
WREG32_SDMA(i, regSDMA_PAGE_RB_CNTL, rb_cntl);
/* Initialize the ring buffer's read and write pointers */
- WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR, 0);
- WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_HI, 0);
- WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR, 0);
- WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_HI, 0);
+ if (restore) {
+ WREG32_SDMA(i, regSDMA_GFX_RB_RPTR, lower_32_bits(ring->wptr << 2));
+ WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_HI, upper_32_bits(ring->wptr << 2));
+ WREG32_SDMA(i, regSDMA_GFX_RB_WPTR, lower_32_bits(ring->wptr << 2));
+ WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_HI, upper_32_bits(ring->wptr << 2));
+ } else {
+ WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR, 0);
+ WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_HI, 0);
+ WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR, 0);
+ WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_HI, 0);
+ }
/* set the wb address whether it's enabled or not */
WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_ADDR_HI,
@@ -792,7 +810,8 @@ static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i)
WREG32_SDMA(i, regSDMA_PAGE_RB_BASE, ring->gpu_addr >> 8);
WREG32_SDMA(i, regSDMA_PAGE_RB_BASE_HI, ring->gpu_addr >> 40);
- ring->wptr = 0;
+ if (!restore)
+ ring->wptr = 0;
/* before programing wptr to a less value, need set minor_ptr_update first */
WREG32_SDMA(i, regSDMA_PAGE_MINOR_PTR_UPDATE, 1);
@@ -911,12 +930,13 @@ static int sdma_v4_4_2_inst_load_microcode(struct amdgpu_device *adev,
*
* @adev: amdgpu_device pointer
* @inst_mask: mask of dma engine instances to be enabled
+ * @restore: boolean to say restore needed or not
*
* Set up the DMA engines and enable them.
* Returns 0 for success, error for failure.
*/
static int sdma_v4_4_2_inst_start(struct amdgpu_device *adev,
- uint32_t inst_mask)
+ uint32_t inst_mask, bool restore)
{
struct amdgpu_ring *ring;
uint32_t tmp_mask;
@@ -927,7 +947,7 @@ static int sdma_v4_4_2_inst_start(struct amdgpu_device *adev,
sdma_v4_4_2_inst_enable(adev, false, inst_mask);
} else {
/* bypass sdma microcode loading on Gopher */
- if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP &&
+ if (!restore && adev->firmware.load_type != AMDGPU_FW_LOAD_PSP &&
adev->sdma.instance[0].fw) {
r = sdma_v4_4_2_inst_load_microcode(adev, inst_mask);
if (r)
@@ -946,17 +966,19 @@ static int sdma_v4_4_2_inst_start(struct amdgpu_device *adev,
uint32_t temp;
WREG32_SDMA(i, regSDMA_SEM_WAIT_FAIL_TIMER_CNTL, 0);
- sdma_v4_4_2_gfx_resume(adev, i);
+ sdma_v4_4_2_gfx_resume(adev, i, restore);
if (adev->sdma.has_page_queue)
- sdma_v4_4_2_page_resume(adev, i);
+ sdma_v4_4_2_page_resume(adev, i, restore);
/* set utc l1 enable flag always to 1 */
temp = RREG32_SDMA(i, regSDMA_CNTL);
temp = REG_SET_FIELD(temp, SDMA_CNTL, UTC_L1_ENABLE, 1);
- /* enable context empty interrupt during initialization */
- temp = REG_SET_FIELD(temp, SDMA_CNTL, CTXEMPTY_INT_ENABLE, 1);
- WREG32_SDMA(i, regSDMA_CNTL, temp);
+ if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) < IP_VERSION(4, 4, 5)) {
+ /* enable context empty interrupt during initialization */
+ temp = REG_SET_FIELD(temp, SDMA_CNTL, CTXEMPTY_INT_ENABLE, 1);
+ WREG32_SDMA(i, regSDMA_CNTL, temp);
+ }
if (!amdgpu_sriov_vf(adev)) {
if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
/* unhalt engine */
@@ -1110,7 +1132,7 @@ static int sdma_v4_4_2_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
@@ -1384,6 +1406,12 @@ static int sdma_v4_4_2_sw_init(struct amdgpu_ip_block *ip_block)
&adev->sdma.srbm_write_irq);
if (r)
return r;
+
+ r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+ SDMA0_4_0__SRCID__SDMA_CTXEMPTY,
+ &adev->sdma.ctxt_empty_irq);
+ if (r)
+ return r;
}
for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -1466,6 +1494,7 @@ static int sdma_v4_4_2_sw_fini(struct amdgpu_ip_block *ip_block)
amdgpu_sdma_sysfs_reset_mask_fini(adev);
if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+ amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 4) ||
amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 5))
amdgpu_sdma_destroy_inst_ctx(adev, true);
else
@@ -1486,7 +1515,7 @@ static int sdma_v4_4_2_hw_init(struct amdgpu_ip_block *ip_block)
if (!amdgpu_sriov_vf(adev))
sdma_v4_4_2_inst_init_golden_registers(adev, inst_mask);
- r = sdma_v4_4_2_inst_start(adev, inst_mask);
+ r = sdma_v4_4_2_inst_start(adev, inst_mask, false);
return r;
}
@@ -1514,7 +1543,7 @@ static int sdma_v4_4_2_hw_fini(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int sdma_v4_4_2_set_clockgating_state(void *handle,
+static int sdma_v4_4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state);
static int sdma_v4_4_2_suspend(struct amdgpu_ip_block *ip_block)
@@ -1522,7 +1551,7 @@ static int sdma_v4_4_2_suspend(struct amdgpu_ip_block *ip_block)
struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_in_reset(adev))
- sdma_v4_4_2_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+ sdma_v4_4_2_set_clockgating_state(ip_block, AMD_CG_STATE_UNGATE);
return sdma_v4_4_2_hw_fini(ip_block);
}
@@ -1573,6 +1602,42 @@ static int sdma_v4_4_2_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
+static int sdma_v4_4_2_reset_queue(struct amdgpu_ring *ring, unsigned int vmid)
+{
+ struct amdgpu_device *adev = ring->adev;
+ int i, r;
+ u32 inst_mask;
+
+ if (amdgpu_sriov_vf(adev))
+ return -EINVAL;
+
+ /* stop queue */
+ inst_mask = 1 << ring->me;
+ sdma_v4_4_2_inst_gfx_stop(adev, inst_mask);
+ if (adev->sdma.has_page_queue)
+ sdma_v4_4_2_inst_page_stop(adev, inst_mask);
+
+ r = amdgpu_dpm_reset_sdma(adev, 1 << GET_INST(SDMA0, ring->me));
+ if (r)
+ return r;
+
+ udelay(50);
+
+ for (i = 0; i < adev->usec_timeout; i++) {
+ if (!REG_GET_FIELD(RREG32_SDMA(ring->me, regSDMA_F32_CNTL), SDMA_F32_CNTL, HALT))
+ break;
+ udelay(1);
+ }
+
+ if (i == adev->usec_timeout) {
+ dev_err(adev->dev, "timed out waiting for SDMA%d unhalt after reset\n",
+ ring->me);
+ return -ETIMEDOUT;
+ }
+
+ return sdma_v4_4_2_inst_start(adev, inst_mask, true);
+}
+
static int sdma_v4_4_2_set_trap_irq_state(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
unsigned type,
@@ -1755,6 +1820,16 @@ static int sdma_v4_4_2_process_srbm_write_irq(struct amdgpu_device *adev,
return 0;
}
+static int sdma_v4_4_2_process_ctxt_empty_irq(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *source,
+ struct amdgpu_iv_entry *entry)
+{
+ /* There is nothing useful to be done here, only kept for debug */
+ dev_dbg_ratelimited(adev->dev, "SDMA context empty interrupt");
+ sdma_v4_4_2_print_iv_entry(adev, entry);
+ return 0;
+}
+
static void sdma_v4_4_2_inst_update_medium_grain_light_sleep(
struct amdgpu_device *adev, bool enable, uint32_t inst_mask)
{
@@ -1821,10 +1896,10 @@ static void sdma_v4_4_2_inst_update_medium_grain_clock_gating(
}
}
-static int sdma_v4_4_2_set_clockgating_state(void *handle,
+static int sdma_v4_4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
uint32_t inst_mask;
if (amdgpu_sriov_vf(adev))
@@ -1839,7 +1914,7 @@ static int sdma_v4_4_2_set_clockgating_state(void *handle,
return 0;
}
-static int sdma_v4_4_2_set_powergating_state(void *handle,
+static int sdma_v4_4_2_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
@@ -1895,7 +1970,6 @@ static void sdma_v4_4_2_dump_ip_state(struct amdgpu_ip_block *ip_block)
if (!adev->sdma.ip_dump)
return;
- amdgpu_gfx_off_ctrl(adev, false);
for (i = 0; i < adev->sdma.num_instances; i++) {
instance_offset = i * reg_count;
for (j = 0; j < reg_count; j++)
@@ -1903,7 +1977,6 @@ static void sdma_v4_4_2_dump_ip_state(struct amdgpu_ip_block *ip_block)
RREG32(sdma_v4_4_2_get_reg_offset(adev, i,
sdma_reg_list_4_4_2[j].reg_offset));
}
- amdgpu_gfx_off_ctrl(adev, true);
}
const struct amd_ip_funcs sdma_v4_4_2_ip_funcs = {
@@ -1955,6 +2028,7 @@ static const struct amdgpu_ring_funcs sdma_v4_4_2_ring_funcs = {
.emit_wreg = sdma_v4_4_2_ring_emit_wreg,
.emit_reg_wait = sdma_v4_4_2_ring_emit_reg_wait,
.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+ .reset = sdma_v4_4_2_reset_queue,
};
static const struct amdgpu_ring_funcs sdma_v4_4_2_page_ring_funcs = {
@@ -2038,6 +2112,10 @@ static const struct amdgpu_irq_src_funcs sdma_v4_4_2_srbm_write_irq_funcs = {
.process = sdma_v4_4_2_process_srbm_write_irq,
};
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_ctxt_empty_irq_funcs = {
+ .process = sdma_v4_4_2_process_ctxt_empty_irq,
+};
+
static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev)
{
adev->sdma.trap_irq.num_types = adev->sdma.num_instances;
@@ -2046,6 +2124,7 @@ static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev)
adev->sdma.doorbell_invalid_irq.num_types = adev->sdma.num_instances;
adev->sdma.pool_timeout_irq.num_types = adev->sdma.num_instances;
adev->sdma.srbm_write_irq.num_types = adev->sdma.num_instances;
+ adev->sdma.ctxt_empty_irq.num_types = adev->sdma.num_instances;
adev->sdma.trap_irq.funcs = &sdma_v4_4_2_trap_irq_funcs;
adev->sdma.illegal_inst_irq.funcs = &sdma_v4_4_2_illegal_inst_irq_funcs;
@@ -2054,6 +2133,7 @@ static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev)
adev->sdma.doorbell_invalid_irq.funcs = &sdma_v4_4_2_doorbell_invalid_irq_funcs;
adev->sdma.pool_timeout_irq.funcs = &sdma_v4_4_2_pool_timeout_irq_funcs;
adev->sdma.srbm_write_irq.funcs = &sdma_v4_4_2_srbm_write_irq_funcs;
+ adev->sdma.ctxt_empty_irq.funcs = &sdma_v4_4_2_ctxt_empty_irq_funcs;
}
/**
@@ -2167,7 +2247,7 @@ static int sdma_v4_4_2_xcp_resume(void *handle, uint32_t inst_mask)
if (!amdgpu_sriov_vf(adev))
sdma_v4_4_2_inst_init_golden_registers(adev, inst_mask);
- r = sdma_v4_4_2_inst_start(adev, inst_mask);
+ r = sdma_v4_4_2_inst_start(adev, inst_mask, false);
return r;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
index fa9b40934957..b764550834a0 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
@@ -1194,7 +1194,7 @@ static int sdma_v5_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
if (!ring->is_mes_queue)
@@ -1853,10 +1853,10 @@ static void sdma_v5_0_update_medium_grain_light_sleep(struct amdgpu_device *adev
}
}
-static int sdma_v5_0_set_clockgating_state(void *handle,
+static int sdma_v5_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1877,7 +1877,7 @@ static int sdma_v5_0_set_clockgating_state(void *handle,
return 0;
}
-static int sdma_v5_0_set_powergating_state(void *handle,
+static int sdma_v5_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index ba5160399ab2..b1818e87889a 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -1050,7 +1050,7 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
if (!ring->is_mes_queue)
@@ -1812,10 +1812,10 @@ static void sdma_v5_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
}
}
-static int sdma_v5_2_set_clockgating_state(void *handle,
+static int sdma_v5_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1841,7 +1841,7 @@ static int sdma_v5_2_set_clockgating_state(void *handle,
return 0;
}
-static int sdma_v5_2_set_powergating_state(void *handle,
+static int sdma_v5_2_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
index d46128b0ec92..1a023b45f0be 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
@@ -1063,7 +1063,7 @@ static int sdma_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
if (!ring->is_mes_queue)
@@ -1601,13 +1601,13 @@ static int sdma_v6_0_process_illegal_inst_irq(struct amdgpu_device *adev,
return 0;
}
-static int sdma_v6_0_set_clockgating_state(void *handle,
+static int sdma_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int sdma_v6_0_set_powergating_state(void *handle,
+static int sdma_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
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 d2ce6b6a7ff6..7e10e94624e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
@@ -490,162 +490,185 @@ static void sdma_v7_0_enable(struct amdgpu_device *adev, bool enable)
}
/**
- * sdma_v7_0_gfx_resume - setup and start the async dma engines
+ * sdma_v7_0_gfx_resume_instance - start/restart a certain sdma engine
*
* @adev: amdgpu_device pointer
+ * @i: instance
+ * @restore: used to restore wptr when restart
*
- * Set up the gfx DMA ring buffers and enable them.
- * Returns 0 for success, error for failure.
+ * Set up the gfx DMA ring buffers and enable them. On restart, we will restore wptr and rptr.
+ * Return 0 for success.
*/
-static int sdma_v7_0_gfx_resume(struct amdgpu_device *adev)
+static int sdma_v7_0_gfx_resume_instance(struct amdgpu_device *adev, int i, bool restore)
{
struct amdgpu_ring *ring;
u32 rb_cntl, ib_cntl;
u32 rb_bufsz;
u32 doorbell;
u32 doorbell_offset;
- u32 tmp;
+ u32 temp;
u64 wptr_gpu_addr;
- int i, r;
-
- for (i = 0; i < adev->sdma.num_instances; i++) {
- ring = &adev->sdma.instance[i].ring;
+ int r;
- //if (!amdgpu_sriov_vf(adev))
- // WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
+ ring = &adev->sdma.instance[i].ring;
- /* Set ring buffer size in dwords */
- rb_bufsz = order_base_2(ring->ring_size / 4);
- rb_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL));
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SIZE, rb_bufsz);
+ /* Set ring buffer size in dwords */
+ rb_bufsz = order_base_2(ring->ring_size / 4);
+ rb_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL));
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SIZE, rb_bufsz);
#ifdef __BIG_ENDIAN
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SWAP_ENABLE, 1);
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL,
- RPTR_WRITEBACK_SWAP_ENABLE, 1);
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SWAP_ENABLE, 1);
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL,
+ RPTR_WRITEBACK_SWAP_ENABLE, 1);
#endif
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_PRIV, 1);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
-
- /* Initialize the ring buffer's read and write pointers */
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_PRIV, 1);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+
+ /* Initialize the ring buffer's read and write pointers */
+ if (restore) {
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR), lower_32_bits(ring->wptr << 2));
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_HI), upper_32_bits(ring->wptr << 2));
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr << 2));
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr << 2));
+ } else {
WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR), 0);
WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_HI), 0);
WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), 0);
WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), 0);
+ }
+ /* setup the wptr shadow polling */
+ wptr_gpu_addr = ring->wptr_gpu_addr;
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_LO),
+ lower_32_bits(wptr_gpu_addr));
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_HI),
+ upper_32_bits(wptr_gpu_addr));
+
+ /* set the wb address whether it's enabled or not */
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_HI),
+ upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_LO),
+ lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
+
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+ if (amdgpu_sriov_vf(adev))
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 1);
+ else
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 0);
- /* setup the wptr shadow polling */
- wptr_gpu_addr = ring->wptr_gpu_addr;
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_LO),
- lower_32_bits(wptr_gpu_addr));
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_HI),
- upper_32_bits(wptr_gpu_addr));
-
- /* set the wb address whether it's enabled or not */
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_HI),
- upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_LO),
- lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
-
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
- if (amdgpu_sriov_vf(adev))
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 1);
- else
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 0);
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, MCU_WPTR_POLL_ENABLE, 1);
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, MCU_WPTR_POLL_ENABLE, 1);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE), ring->gpu_addr >> 8);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE_HI), ring->gpu_addr >> 40);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE), ring->gpu_addr >> 8);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE_HI), ring->gpu_addr >> 40);
+ if (!restore)
ring->wptr = 0;
- /* before programing wptr to a less value, need set minor_ptr_update first */
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 1);
+ /* before programing wptr to a less value, need set minor_ptr_update first */
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 1);
- if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr) << 2);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
- }
+ if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr) << 2);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
+ }
- doorbell = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL));
- doorbell_offset = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET));
+ doorbell = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL));
+ doorbell_offset = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET));
- if (ring->use_doorbell) {
- doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 1);
- doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_QUEUE0_DOORBELL_OFFSET,
- OFFSET, ring->doorbell_index);
- } else {
- doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 0);
- }
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL), doorbell);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET), doorbell_offset);
-
- if (i == 0)
- adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
- ring->doorbell_index,
- adev->doorbell_index.sdma_doorbell_range * adev->sdma.num_instances);
-
- if (amdgpu_sriov_vf(adev))
- sdma_v7_0_ring_set_wptr(ring);
-
- /* set minor_ptr_update to 0 after wptr programed */
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 0);
-
- /* Set up sdma hang watchdog */
- tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL));
- /* 100ms per unit */
- tmp = REG_SET_FIELD(tmp, SDMA0_WATCHDOG_CNTL, QUEUE_HANG_COUNT,
- max(adev->usec_timeout/100000, 1));
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL), tmp);
-
- /* Set up RESP_MODE to non-copy addresses */
- tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL));
- tmp = REG_SET_FIELD(tmp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
- tmp = REG_SET_FIELD(tmp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL), tmp);
-
- /* program default cache read and write policy */
- tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE));
- /* clean read policy and write policy bits */
- tmp &= 0xFF0FFF;
- tmp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) |
- (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE), tmp);
-
- if (!amdgpu_sriov_vf(adev)) {
- /* unhalt engine */
- tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL));
- tmp = REG_SET_FIELD(tmp, SDMA0_MCU_CNTL, HALT, 0);
- tmp = REG_SET_FIELD(tmp, SDMA0_MCU_CNTL, RESET, 0);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL), tmp);
- }
+ if (ring->use_doorbell) {
+ doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 1);
+ doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_QUEUE0_DOORBELL_OFFSET,
+ OFFSET, ring->doorbell_index);
+ } else {
+ doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 0);
+ }
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL), doorbell);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET), doorbell_offset);
- /* enable DMA RB */
- rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_ENABLE, 1);
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+ if (i == 0)
+ adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
+ ring->doorbell_index,
+ adev->doorbell_index.sdma_doorbell_range * adev->sdma.num_instances);
- ib_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL));
- ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_ENABLE, 1);
+ if (amdgpu_sriov_vf(adev))
+ sdma_v7_0_ring_set_wptr(ring);
+
+ /* set minor_ptr_update to 0 after wptr programed */
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 0);
+
+ /* Set up sdma hang watchdog */
+ temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL));
+ /* 100ms per unit */
+ temp = REG_SET_FIELD(temp, SDMA0_WATCHDOG_CNTL, QUEUE_HANG_COUNT,
+ max(adev->usec_timeout/100000, 1));
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL), temp);
+
+ /* Set up RESP_MODE to non-copy addresses */
+ temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL));
+ temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
+ temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL), temp);
+
+ /* program default cache read and write policy */
+ temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE));
+ /* clean read policy and write policy bits */
+ temp &= 0xFF0FFF;
+ temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) |
+ (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE), temp);
+
+ if (!amdgpu_sriov_vf(adev)) {
+ /* unhalt engine */
+ temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL));
+ temp = REG_SET_FIELD(temp, SDMA0_MCU_CNTL, HALT, 0);
+ temp = REG_SET_FIELD(temp, SDMA0_MCU_CNTL, RESET, 0);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL), temp);
+ }
+
+ /* enable DMA RB */
+ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_ENABLE, 1);
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+
+ ib_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL));
+ ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_ENABLE, 1);
#ifdef __BIG_ENDIAN
- ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_SWAP_ENABLE, 1);
+ ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_SWAP_ENABLE, 1);
#endif
- /* enable DMA IBs */
- WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl);
+ /* enable DMA IBs */
+ WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl);
+ ring->sched.ready = true;
- ring->sched.ready = true;
+ if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
+ sdma_v7_0_ctx_switch_enable(adev, true);
+ sdma_v7_0_enable(adev, true);
+ }
- if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
- sdma_v7_0_ctx_switch_enable(adev, true);
- sdma_v7_0_enable(adev, true);
- }
+ r = amdgpu_ring_test_helper(ring);
+ if (r)
+ ring->sched.ready = false;
- r = amdgpu_ring_test_helper(ring);
- if (r) {
- ring->sched.ready = false;
- return r;
- }
+ return r;
+}
+
+/**
+ * sdma_v7_0_gfx_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v7_0_gfx_resume(struct amdgpu_device *adev)
+{
+ int i, r;
+ for (i = 0; i < adev->sdma.num_instances; i++) {
+ r = sdma_v7_0_gfx_resume_instance(adev, i, false);
+ if (r)
+ return r;
}
return 0;
+
}
/**
@@ -806,6 +829,31 @@ static bool sdma_v7_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
return false;
}
+static int sdma_v7_0_reset_queue(struct amdgpu_ring *ring, unsigned int vmid)
+{
+ struct amdgpu_device *adev = ring->adev;
+ int i, r;
+
+ if (amdgpu_sriov_vf(adev))
+ return -EINVAL;
+
+ for (i = 0; i < adev->sdma.num_instances; i++) {
+ if (ring == &adev->sdma.instance[i].ring)
+ break;
+ }
+
+ if (i == adev->sdma.num_instances) {
+ DRM_ERROR("sdma instance not found\n");
+ return -EINVAL;
+ }
+
+ r = amdgpu_mes_reset_legacy_queue(adev, ring, vmid, true);
+ if (r)
+ return r;
+
+ return sdma_v7_0_gfx_resume_instance(adev, i, true);
+}
+
/**
* sdma_v7_0_start - setup and start the async dma engines
*
@@ -1060,7 +1108,7 @@ static int sdma_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
if (!ring->is_mes_queue)
@@ -1316,6 +1364,13 @@ static int sdma_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
return r;
}
+ adev->sdma.supported_reset =
+ amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring);
+ adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+ r = amdgpu_sdma_sysfs_reset_mask_init(adev);
+ if (r)
+ return r;
/* Allocate memory for SDMA IP Dump buffer */
ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
if (ptr)
@@ -1334,6 +1389,7 @@ static int sdma_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
for (i = 0; i < adev->sdma.num_instances; i++)
amdgpu_ring_fini(&adev->sdma.instance[i].ring);
+ amdgpu_sdma_sysfs_reset_mask_fini(adev);
amdgpu_sdma_destroy_inst_ctx(adev, true);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT)
@@ -1524,13 +1580,13 @@ static int sdma_v7_0_process_illegal_inst_irq(struct amdgpu_device *adev,
return 0;
}
-static int sdma_v7_0_set_clockgating_state(void *handle,
+static int sdma_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int sdma_v7_0_set_powergating_state(void *handle,
+static int sdma_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
@@ -1636,6 +1692,7 @@ static const struct amdgpu_ring_funcs sdma_v7_0_ring_funcs = {
.emit_reg_write_reg_wait = sdma_v7_0_ring_emit_reg_write_reg_wait,
.init_cond_exec = sdma_v7_0_ring_init_cond_exec,
.preempt_ib = sdma_v7_0_ring_preempt_ib,
+ .reset = sdma_v7_0_reset_queue,
};
static void sdma_v7_0_set_ring_funcs(struct amdgpu_device *adev)
@@ -1684,11 +1741,12 @@ static void sdma_v7_0_emit_copy_buffer(struct amdgpu_ib *ib,
uint32_t byte_count,
uint32_t copy_flags)
{
- uint32_t num_type, data_format, max_com;
+ uint32_t num_type, data_format, max_com, write_cm;
max_com = AMDGPU_COPY_FLAGS_GET(copy_flags, MAX_COMPRESSED);
data_format = AMDGPU_COPY_FLAGS_GET(copy_flags, DATA_FORMAT);
num_type = AMDGPU_COPY_FLAGS_GET(copy_flags, NUMBER_TYPE);
+ write_cm = AMDGPU_COPY_FLAGS_GET(copy_flags, WRITE_COMPRESS_DISABLE) ? 2 : 1;
ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COPY) |
SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
@@ -1705,7 +1763,7 @@ static void sdma_v7_0_emit_copy_buffer(struct amdgpu_ib *ib,
if ((copy_flags & (AMDGPU_COPY_FLAGS_READ_DECOMPRESSED | AMDGPU_COPY_FLAGS_WRITE_COMPRESSED)))
ib->ptr[ib->length_dw++] = SDMA_DCC_DATA_FORMAT(data_format) | SDMA_DCC_NUM_TYPE(num_type) |
((copy_flags & AMDGPU_COPY_FLAGS_READ_DECOMPRESSED) ? SDMA_DCC_READ_CM(2) : 0) |
- ((copy_flags & AMDGPU_COPY_FLAGS_WRITE_COMPRESSED) ? SDMA_DCC_WRITE_CM(1) : 0) |
+ ((copy_flags & AMDGPU_COPY_FLAGS_WRITE_COMPRESSED) ? SDMA_DCC_WRITE_CM(write_cm) : 0) |
SDMA_DCC_MAX_COM(max_com) | SDMA_DCC_MAX_UCOM(1);
else
ib->ptr[ib->length_dw++] = 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c
index 00f63d3fbea7..77ef7da2e4fe 100644
--- a/drivers/gpu/drm/amd/amdgpu/si.c
+++ b/drivers/gpu/drm/amd/amdgpu/si.c
@@ -2649,13 +2649,13 @@ static bool si_common_is_idle(void *handle)
return true;
}
-static int si_common_set_clockgating_state(void *handle,
+static int si_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int si_common_set_powergating_state(void *handle,
+static int si_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c
index 47647a6083e8..dbd78d5345a4 100644
--- a/drivers/gpu/drm/amd/amdgpu/si_dma.c
+++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c
@@ -286,7 +286,7 @@ static int si_dma_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err1:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err0:
amdgpu_device_wb_free(adev, index);
@@ -629,13 +629,13 @@ static int si_dma_process_trap_irq(struct amdgpu_device *adev,
return 0;
}
-static int si_dma_set_clockgating_state(void *handle,
+static int si_dma_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
u32 orig, data, offset;
int i;
bool enable;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
enable = (state == AMD_CG_STATE_GATE);
@@ -672,12 +672,12 @@ static int si_dma_set_clockgating_state(void *handle,
return 0;
}
-static int si_dma_set_powergating_state(void *handle,
+static int si_dma_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
u32 tmp;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
WREG32(DMA_PGFSM_WRITE, 0x00002000);
WREG32(DMA_PGFSM_CONFIG, 0x100010ff);
diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c
index 2ec1ebe4db11..a32b6243c1f8 100644
--- a/drivers/gpu/drm/amd/amdgpu/si_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c
@@ -263,13 +263,13 @@ static int si_ih_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int si_ih_set_clockgating_state(void *handle,
+static int si_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int si_ih_set_powergating_state(void *handle,
+static int si_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index ede072758dab..e98fb3fa36a8 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -103,12 +103,11 @@ static const struct amdgpu_video_codecs vega_video_codecs_encode =
/* Vega */
static const struct amdgpu_video_codec_info vega_video_codecs_decode_array[] =
{
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 4096, 186)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
};
static const struct amdgpu_video_codecs vega_video_codecs_decode =
@@ -120,12 +119,12 @@ static const struct amdgpu_video_codecs vega_video_codecs_decode =
/* Raven */
static const struct amdgpu_video_codec_info rv_video_codecs_decode_array[] =
{
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 4096, 186)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 8192, 8192, 0)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 4096, 4096, 0)},
};
@@ -138,10 +137,10 @@ static const struct amdgpu_video_codecs rv_video_codecs_decode =
/* Renoir, Arcturus */
static const struct amdgpu_video_codec_info rn_video_codecs_decode_array[] =
{
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
- {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
@@ -171,6 +170,24 @@ static const struct amdgpu_video_codecs vcn_4_0_3_video_codecs_encode = {
.codec_array = NULL,
};
+static const struct amdgpu_video_codecs vcn_5_0_1_video_codecs_encode_vcn0 = {
+ .codec_count = 0,
+ .codec_array = NULL,
+};
+
+static const struct amdgpu_video_codec_info vcn_5_0_1_video_codecs_decode_array_vcn0[] = {
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_5_0_1_video_codecs_decode_vcn0 = {
+ .codec_count = ARRAY_SIZE(vcn_5_0_1_video_codecs_decode_array_vcn0),
+ .codec_array = vcn_5_0_1_video_codecs_decode_array_vcn0,
+};
+
static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode,
const struct amdgpu_video_codecs **codecs)
{
@@ -209,6 +226,12 @@ static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode,
else
*codecs = &vcn_4_0_3_video_codecs_decode;
return 0;
+ case IP_VERSION(5, 0, 1):
+ if (encode)
+ *codecs = &vcn_5_0_1_video_codecs_encode_vcn0;
+ else
+ *codecs = &vcn_5_0_1_video_codecs_decode_vcn0;
+ return 0;
default:
return -EINVAL;
}
@@ -327,6 +350,7 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev)
if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(12, 0, 0) ||
amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(12, 0, 1) ||
amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+ amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 12) ||
amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 14))
return 10000;
if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(10, 0, 0) ||
@@ -556,6 +580,7 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
break;
case IP_VERSION(13, 0, 6):
case IP_VERSION(13, 0, 14):
+ case IP_VERSION(13, 0, 12):
/* Use gpu_recovery param to target a reset method.
* Enable triggering of GPU reset only if specified
* by module parameter.
@@ -1177,6 +1202,7 @@ static int soc15_common_early_init(struct amdgpu_ip_block *ip_block)
break;
case IP_VERSION(9, 4, 3):
case IP_VERSION(9, 4, 4):
+ case IP_VERSION(9, 5, 0):
adev->asic_funcs = &aqua_vanjaram_asic_funcs;
adev->cg_flags =
AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG |
@@ -1385,10 +1411,10 @@ static void soc15_update_drm_light_sleep(struct amdgpu_device *adev, bool enable
WREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_LIGHT_SLEEP_CTRL), data);
}
-static int soc15_common_set_clockgating_state(void *handle,
+static int soc15_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1453,6 +1479,7 @@ static void soc15_common_get_clockgating_state(void *handle, u64 *flags)
if ((amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 2)) &&
(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))) {
/* AMD_CG_SUPPORT_DRM_MGCG */
data = RREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_CGTT_CTRL0));
@@ -1473,7 +1500,7 @@ static void soc15_common_get_clockgating_state(void *handle, u64 *flags)
adev->df.funcs->get_clockgating_state(adev, flags);
}
-static int soc15_common_set_powergating_state(void *handle,
+static int soc15_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* todo */
diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
index d6999835918f..62ad67d0b598 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc21.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -928,10 +928,10 @@ static bool soc21_common_is_idle(void *handle)
return true;
}
-static int soc21_common_set_clockgating_state(void *handle,
+static int soc21_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
case IP_VERSION(4, 3, 0):
@@ -954,10 +954,10 @@ static int soc21_common_set_clockgating_state(void *handle,
return 0;
}
-static int soc21_common_set_powergating_state(void *handle,
+static int soc21_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (amdgpu_ip_version(adev, LSDMA_HWIP, 0)) {
case IP_VERSION(6, 0, 0):
diff --git a/drivers/gpu/drm/amd/amdgpu/soc24.c b/drivers/gpu/drm/amd/amdgpu/soc24.c
index be96de92b2f5..6b8e078ee7c7 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc24.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc24.c
@@ -444,8 +444,18 @@ static int soc24_common_late_init(struct amdgpu_ip_block *ip_block)
{
struct amdgpu_device *adev = ip_block->adev;
- if (amdgpu_sriov_vf(adev))
+ if (amdgpu_sriov_vf(adev)) {
xgpu_nv_mailbox_get_irq(adev);
+ } else {
+ if (adev->nbio.ras &&
+ adev->nbio.ras_err_event_athub_irq.funcs)
+ /* don't need to fail gpu late init
+ * if enabling athub_err_event interrupt failed
+ * nbif v6_3_1 only support fatal error hanlding
+ * just enable the interrupt directly
+ */
+ amdgpu_irq_get(adev, &adev->nbio.ras_err_event_athub_irq, 0);
+ }
/* Enable selfring doorbell aperture late because doorbell BAR
* aperture will change if resize BAR successfully in gmc sw_init.
@@ -501,8 +511,13 @@ static int soc24_common_hw_fini(struct amdgpu_ip_block *ip_block)
adev->nbio.funcs->enable_doorbell_aperture(adev, false);
adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, false);
- if (amdgpu_sriov_vf(adev))
+ if (amdgpu_sriov_vf(adev)) {
xgpu_nv_mailbox_put_irq(adev);
+ } else {
+ if (adev->nbio.ras &&
+ adev->nbio.ras_err_event_athub_irq.funcs)
+ amdgpu_irq_put(adev, &adev->nbio.ras_err_event_athub_irq, 0);
+ }
return 0;
}
@@ -522,10 +537,10 @@ static bool soc24_common_is_idle(void *handle)
return true;
}
-static int soc24_common_set_clockgating_state(void *handle,
+static int soc24_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
case IP_VERSION(6, 3, 1):
@@ -542,10 +557,10 @@ static int soc24_common_set_clockgating_state(void *handle,
return 0;
}
-static int soc24_common_set_powergating_state(void *handle,
+static int soc24_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (amdgpu_ip_version(adev, LSDMA_HWIP, 0)) {
case IP_VERSION(7, 0, 0):
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
index 21b71a427b1f..64891f099366 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
@@ -30,6 +30,9 @@
#define RSP_ID_MASK (1U << 31)
#define RSP_ID(cmdId) (((uint32_t)(cmdId)) | RSP_ID_MASK)
+/* invalid node instance value */
+#define TA_RAS_INV_NODE 0xffff
+
/* RAS related enumerations */
/**********************************************************/
enum ras_command {
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h b/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
index 00d8bdb8254f..9ec2e03d41c7 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
@@ -31,10 +31,12 @@
* Secure Display Command ID
*/
enum ta_securedisplay_command {
- /* Query whether TA is responding used only for validation purpose */
+ /* Query whether TA is responding. It is used only for validation purpose */
TA_SECUREDISPLAY_COMMAND__QUERY_TA = 1,
/* Send region of Interest and CRC value to I2C */
TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC = 2,
+ /* V2 to send multiple regions of Interest and CRC value to I2C */
+ TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC_V2 = 3,
/* Maximum Command ID */
TA_SECUREDISPLAY_COMMAND__MAX_ID = 0x7FFFFFFF,
};
@@ -83,6 +85,8 @@ enum ta_securedisplay_ta_query_cmd_ret {
enum ta_securedisplay_buffer_size {
/* 15 bytes = 8 byte (ROI) + 6 byte(CRC) + 1 byte(phy_id) */
TA_SECUREDISPLAY_I2C_BUFFER_SIZE = 15,
+ /* 16 bytes = 8 byte (ROI) + 6 byte(CRC) + 1 byte(phy_id) + 1 byte(roi_idx) */
+ TA_SECUREDISPLAY_V2_I2C_BUFFER_SIZE = 16,
};
/** Input/output structures for Secure Display commands */
@@ -95,7 +99,15 @@ enum ta_securedisplay_buffer_size {
* Physical ID to determine which DIO scratch register should be used to get ROI
*/
struct ta_securedisplay_send_roi_crc_input {
- uint32_t phy_id; /* Physical ID */
+ /* Physical ID */
+ uint32_t phy_id;
+};
+
+struct ta_securedisplay_send_roi_crc_v2_input {
+ /* Physical ID */
+ uint32_t phy_id;
+ /* Region of interest index */
+ uint8_t roi_idx;
};
/** @union ta_securedisplay_cmd_input
@@ -104,6 +116,8 @@ struct ta_securedisplay_send_roi_crc_input {
union ta_securedisplay_cmd_input {
/* send ROI and CRC input buffer format */
struct ta_securedisplay_send_roi_crc_input send_roi_crc;
+ /* send ROI and CRC input buffer format, v2 adds a ROI index */
+ struct ta_securedisplay_send_roi_crc_v2_input send_roi_crc_v2;
uint32_t reserved[4];
};
@@ -128,6 +142,10 @@ struct ta_securedisplay_send_roi_crc_output {
uint8_t reserved;
};
+struct ta_securedisplay_send_roi_crc_v2_output {
+ uint8_t i2c_buf[TA_SECUREDISPLAY_V2_I2C_BUFFER_SIZE]; /* I2C buffer */
+};
+
/** @union ta_securedisplay_cmd_output
* Output buffer
*/
@@ -136,6 +154,8 @@ union ta_securedisplay_cmd_output {
struct ta_securedisplay_query_ta_output query_ta;
/* Send ROI CRC output buffer format used only for validation purpose */
struct ta_securedisplay_send_roi_crc_output send_roi_crc;
+ /* Send ROI CRC output buffer format used only for validation purpose */
+ struct ta_securedisplay_send_roi_crc_v2_output send_roi_crc_v2;
uint32_t reserved[4];
};
diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
index 5a04a6770138..0968e551f7b5 100644
--- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
@@ -448,13 +448,13 @@ static int tonga_ih_soft_reset(struct amdgpu_ip_block *ip_block)
return 0;
}
-static int tonga_ih_set_clockgating_state(void *handle,
+static int tonga_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int tonga_ih_set_powergating_state(void *handle,
+static int tonga_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
index 1a8ea834efa6..a7b9c358a2d4 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
@@ -173,156 +173,96 @@ 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_convert_error_address(struct amdgpu_device *adev,
+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_input *addr_in,
+ struct ta_ras_query_address_output *addr_out,
+ bool dump_addr)
{
- uint32_t col, row, row_xor, bank, channel_index;
- uint64_t soc_pa, retired_page, column, err_addr;
- struct ta_ras_query_address_output addr_out;
+ uint32_t col, col_lower, row, row_lower, bank;
+ uint32_t channel_index = 0, umc_inst = 0;
+ uint32_t i, loop_bits[UMC_V12_0_RETIRE_LOOP_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)
+ paddr_out = &addr_out_tmp;
+ else
+ paddr_out = addr_out;
- err_addr = addr_in->ma.err_addr;
- addr_in->addr_type = TA_RAS_MCA_TO_PA;
- if (psp_ras_query_address(&adev->psp, addr_in, &addr_out)) {
- dev_warn(adev->dev, "Failed to query RAS physical address for 0x%llx",
- err_addr);
+ err_addr = bank = 0;
+ if (addr_in) {
+ err_addr = addr_in->ma.err_addr;
+ addr_in->addr_type = TA_RAS_MCA_TO_PA;
+ ret = psp_ras_query_address(&adev->psp, addr_in, paddr_out);
+ if (ret) {
+ dev_warn(adev->dev, "Failed to query RAS physical address for 0x%llx",
+ err_addr);
- return;
- }
+ goto out;
+ }
- soc_pa = addr_out.pa.pa;
- bank = addr_out.pa.bank;
- channel_index = addr_out.pa.channel_idx;
-
- col = (err_addr >> 1) & 0x1fULL;
- row = (err_addr >> 10) & 0x3fffULL;
- row_xor = row ^ (0x1ULL << 13);
- /* clear [C3 C2] in soc physical address */
- soc_pa &= ~(0x3ULL << UMC_V12_0_PA_C2_BIT);
- /* clear [C4] in soc physical address */
- soc_pa &= ~(0x1ULL << UMC_V12_0_PA_C4_BIT);
-
- /* loop for all possibilities of [C4 C3 C2] */
- for (column = 0; column < UMC_V12_0_NA_MAP_PA_NUM; column++) {
- retired_page = soc_pa | ((column & 0x3) << UMC_V12_0_PA_C2_BIT);
- retired_page |= (((column & 0x4) >> 2) << UMC_V12_0_PA_C4_BIT);
- /* include column bit 0 and 1 */
- col &= 0x3;
- col |= (column << 2);
- dev_info(adev->dev,
- "Error Address(PA):0x%-10llx Row:0x%-4x Col:0x%-2x Bank:0x%x Channel:0x%x\n",
- retired_page, row, col, bank, channel_index);
- amdgpu_umc_fill_error_record(err_data, err_addr,
- retired_page, channel_index, addr_in->ma.umc_inst);
-
- /* shift R13 bit */
- retired_page ^= (0x1ULL << UMC_V12_0_PA_R13_BIT);
- dev_info(adev->dev,
- "Error Address(PA):0x%-10llx Row:0x%-4x Col:0x%-2x Bank:0x%x Channel:0x%x\n",
- retired_page, row_xor, col, bank, channel_index);
- amdgpu_umc_fill_error_record(err_data, err_addr,
- retired_page, channel_index, addr_in->ma.umc_inst);
+ bank = paddr_out->pa.bank;
+ /* no need to care about umc inst if addr_in is NULL */
+ umc_inst = addr_in->ma.umc_inst;
}
-}
-static void umc_v12_0_dump_addr_info(struct amdgpu_device *adev,
- struct ta_ras_query_address_output *addr_out,
- uint64_t err_addr)
-{
- uint32_t col, row, row_xor, bank, channel_index;
- uint64_t soc_pa, retired_page, column;
-
- soc_pa = addr_out->pa.pa;
- bank = addr_out->pa.bank;
- channel_index = addr_out->pa.channel_idx;
-
- col = (err_addr >> 1) & 0x1fULL;
- row = (err_addr >> 10) & 0x3fffULL;
- row_xor = row ^ (0x1ULL << 13);
- /* clear [C3 C2] in soc physical address */
- soc_pa &= ~(0x3ULL << UMC_V12_0_PA_C2_BIT);
- /* clear [C4] in soc physical address */
- soc_pa &= ~(0x1ULL << UMC_V12_0_PA_C4_BIT);
-
- /* loop for all possibilities of [C4 C3 C2] */
- for (column = 0; column < UMC_V12_0_NA_MAP_PA_NUM; column++) {
- retired_page = soc_pa | ((column & 0x3) << UMC_V12_0_PA_C2_BIT);
- retired_page |= (((column & 0x4) >> 2) << UMC_V12_0_PA_C4_BIT);
- /* include column bit 0 and 1 */
- col &= 0x3;
- col |= (column << 2);
- dev_info(adev->dev,
- "Error Address(PA):0x%-10llx Row:0x%-4x Col:0x%-2x Bank:0x%x Channel:0x%x\n",
- retired_page, row, col, bank, channel_index);
-
- /* shift R13 bit */
- retired_page ^= (0x1ULL << UMC_V12_0_PA_R13_BIT);
- dev_info(adev->dev,
- "Error Address(PA):0x%-10llx Row:0x%-4x Col:0x%-2x Bank:0x%x Channel:0x%x\n",
- retired_page, row_xor, col, bank, channel_index);
- }
-}
+ 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;
-static int umc_v12_0_lookup_bad_pages_in_a_row(struct amdgpu_device *adev,
- uint64_t pa_addr, uint64_t *pfns, int len)
-{
- uint64_t soc_pa, retired_page, column;
- uint32_t pos = 0;
-
- soc_pa = pa_addr;
- /* clear [C3 C2] in soc physical address */
- soc_pa &= ~(0x3ULL << UMC_V12_0_PA_C2_BIT);
- /* clear [C4] in soc physical address */
- soc_pa &= ~(0x1ULL << UMC_V12_0_PA_C4_BIT);
-
- /* loop for all possibilities of [C4 C3 C2] */
- for (column = 0; column < UMC_V12_0_NA_MAP_PA_NUM; column++) {
- retired_page = soc_pa | ((column & 0x3) << UMC_V12_0_PA_C2_BIT);
- retired_page |= (((column & 0x4) >> 2) << UMC_V12_0_PA_C4_BIT);
-
- if (pos >= len)
- return 0;
- pfns[pos++] = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
-
- /* shift R13 bit */
- retired_page ^= (0x1ULL << UMC_V12_0_PA_R13_BIT);
-
- if (pos >= len)
- return 0;
- pfns[pos++] = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
+ 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;
}
- return pos;
-}
-
-static int umc_v12_0_convert_mca_to_addr(struct amdgpu_device *adev,
- uint64_t err_addr, uint32_t ch, uint32_t umc,
- uint32_t node, uint32_t socket,
- uint64_t *addr, bool dump_addr)
-{
- struct ta_ras_query_address_input addr_in;
- struct ta_ras_query_address_output addr_out;
-
- memset(&addr_in, 0, sizeof(addr_in));
- addr_in.ma.err_addr = err_addr;
- addr_in.ma.ch_inst = ch;
- addr_in.ma.umc_inst = umc;
- addr_in.ma.node_inst = node;
- addr_in.ma.socket_id = socket;
- addr_in.addr_type = TA_RAS_MCA_TO_PA;
- if (psp_ras_query_address(&adev->psp, &addr_in, &addr_out)) {
- dev_warn(adev->dev, "Failed to query RAS physical address for 0x%llx",
- err_addr);
- return -EINVAL;
+ 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]);
+
+ 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 */
+ row_lower = (err_addr >> UMC_V12_0_MA_R0_BIT) & 0x1fffULL;
+
+ 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++) {
+ 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]);
+
+ col = ((column & 0x7) << 2) | col_lower;
+ /* add row bit 13 */
+ row = ((column >> 3) << 13) | row_lower;
+
+ if (dump_addr)
+ dev_info(adev->dev,
+ "Error Address(PA):0x%-10llx Row:0x%-4x Col:0x%-2x Bank:0x%x Channel:0x%x\n",
+ soc_pa, row, col, bank, channel_index);
+
+ if (err_data)
+ amdgpu_umc_fill_error_record(err_data, err_addr,
+ soc_pa, channel_index, umc_inst);
}
- if (dump_addr)
- umc_v12_0_dump_addr_info(adev, &addr_out, err_addr);
-
- *addr = addr_out.pa.pa;
-
- return 0;
+out:
+ return ret;
}
static int umc_v12_0_query_error_address(struct amdgpu_device *adev,
@@ -374,7 +314,7 @@ static int umc_v12_0_query_error_address(struct amdgpu_device *adev,
addr_in.ma.umc_inst = umc_inst;
addr_in.ma.node_inst = node_inst;
- umc_v12_0_convert_error_address(adev, err_data, &addr_in);
+ umc_v12_0_convert_error_address(adev, err_data, &addr_in, NULL, true);
}
/* clear umc status */
@@ -526,6 +466,9 @@ static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev,
uint64_t page_pfn[UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL];
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;
int count, ret, i;
hwid = REG_GET_FIELD(ipid, MCMP1_IPIDT0, HardwareID);
@@ -552,10 +495,10 @@ static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev,
MCA_IPID_2_UMC_CH(ipid),
err_addr);
- ret = umc_v12_0_convert_mca_to_addr(adev,
+ ret = amdgpu_umc_mca_to_addr(adev,
err_addr, MCA_IPID_2_UMC_CH(ipid),
MCA_IPID_2_UMC_INST(ipid), MCA_IPID_2_DIE_ID(ipid),
- MCA_IPID_2_SOCKET_ID(ipid), &pa_addr, true);
+ MCA_IPID_2_SOCKET_ID(ipid), &addr_out, true);
if (ret)
return ret;
@@ -563,14 +506,21 @@ static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev,
if (!ecc_err)
return -ENOMEM;
+ pa_addr = addr_out.pa.pa;
ecc_err->status = status;
ecc_err->ipid = ipid;
ecc_err->addr = addr;
- ecc_err->pa_pfn = UMC_V12_ADDR_MASK_BAD_COLS(pa_addr) >> AMDGPU_GPU_PAGE_SHIFT;
+ 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(UMC_V12_0_PA_C4_BIT) >> AMDGPU_GPU_PAGE_SHIFT;
+ ecc_err->pa_pfn = BIT_ULL(shift_bit) >> AMDGPU_GPU_PAGE_SHIFT;
ret = amdgpu_umc_logs_ecc_err(adev, &con->umc_ecc_log.de_page_tree, ecc_err);
if (ret) {
@@ -586,7 +536,7 @@ static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev,
con->umc_ecc_log.de_queried_count++;
memset(page_pfn, 0, sizeof(page_pfn));
- count = umc_v12_0_lookup_bad_pages_in_a_row(adev,
+ count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
pa_addr,
page_pfn, ARRAY_SIZE(page_pfn));
if (count <= 0) {
@@ -629,7 +579,7 @@ static int umc_v12_0_fill_error_record(struct amdgpu_device *adev,
return -EINVAL;
memset(page_pfn, 0, sizeof(page_pfn));
- count = umc_v12_0_lookup_bad_pages_in_a_row(adev,
+ count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
ecc_err->pa_pfn << AMDGPU_GPU_PAGE_SHIFT,
page_pfn, ARRAY_SIZE(page_pfn));
@@ -637,7 +587,7 @@ static int umc_v12_0_fill_error_record(struct amdgpu_device *adev,
ret = amdgpu_umc_fill_error_record(err_data,
ecc_err->addr,
page_pfn[i] << AMDGPU_GPU_PAGE_SHIFT,
- MCA_IPID_2_UMC_CH(ecc_err->ipid),
+ ecc_err->channel_idx,
MCA_IPID_2_UMC_INST(ecc_err->ipid));
if (ret)
break;
@@ -676,6 +626,31 @@ static void umc_v12_0_query_ras_ecc_err_addr(struct amdgpu_device *adev,
mutex_unlock(&con->umc_ecc_log.lock);
}
+static uint32_t umc_v12_0_get_die_id(struct amdgpu_device *adev,
+ uint64_t mca_addr, uint64_t retired_page)
+{
+ uint32_t die = 0;
+
+ /* we only calculate die id for nps1 mode right now */
+ die += ((((retired_page >> 12) & 0x1ULL)^
+ ((retired_page >> 20) & 0x1ULL) ^
+ ((retired_page >> 27) & 0x1ULL) ^
+ ((retired_page >> 34) & 0x1ULL) ^
+ ((retired_page >> 41) & 0x1ULL)) << 0);
+
+ /* the original PA_C4 and PA_R13 may be cleared in retired_page, so
+ * get them from mca_addr.
+ */
+ die += ((((retired_page >> 13) & 0x1ULL) ^
+ ((mca_addr >> 5) & 0x1ULL) ^
+ ((retired_page >> 28) & 0x1ULL) ^
+ ((mca_addr >> 23) & 0x1ULL) ^
+ ((retired_page >> 42) & 0x1ULL)) << 1);
+ die &= 3;
+
+ return die;
+}
+
struct amdgpu_umc_ras umc_v12_0_ras = {
.ras_block = {
.hw_ops = &umc_v12_0_ras_hw_ops,
@@ -686,5 +661,7 @@ struct amdgpu_umc_ras umc_v12_0_ras = {
.ecc_info_query_ras_error_address = umc_v12_0_query_ras_ecc_err_addr,
.check_ecc_err_status = umc_v12_0_check_ecc_err_status,
.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,
};
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
index be5598d76c1d..9298018d938f 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
@@ -55,12 +55,24 @@
#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
+#define UMC_V12_0_PA_C3_BIT 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_R11_BIT 33
#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
+/* row bits in MCA address */
+#define UMC_V12_0_MA_R0_BIT 10
#define MCA_UMC_HWID_V12_0 0x96
#define MCA_UMC_MCATYPE_V12_0 0x0
@@ -81,11 +93,6 @@
(((REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdLo) & 0x1) << 2) | \
(REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdHi) & 0x03))
-#define UMC_V12_ADDR_MASK_BAD_COLS(addr) \
- ((addr) & ~((0x3ULL << UMC_V12_0_PA_C2_BIT) | \
- (0x1ULL << UMC_V12_0_PA_C4_BIT) | \
- (0x1ULL << UMC_V12_0_PA_R13_BIT)))
-
bool umc_v12_0_is_deferred_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
bool umc_v12_0_is_uncorrectable_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
bool umc_v12_0_is_correctable_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_14.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.c
new file mode 100644
index 000000000000..eaca10a3c4a9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.c
@@ -0,0 +1,160 @@
+/*
+ * 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 "umc_v8_14.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_umc.h"
+#include "amdgpu.h"
+#include "umc/umc_8_14_0_offset.h"
+#include "umc/umc_8_14_0_sh_mask.h"
+
+static inline uint32_t get_umc_v8_14_reg_offset(struct amdgpu_device *adev,
+ uint32_t umc_inst,
+ uint32_t ch_inst)
+{
+ return adev->umc.channel_offs * ch_inst + UMC_V8_14_INST_DIST * umc_inst;
+}
+
+static int umc_v8_14_clear_error_count_per_channel(struct amdgpu_device *adev,
+ uint32_t node_inst, uint32_t umc_inst,
+ uint32_t ch_inst, void *data)
+{
+ uint32_t ecc_err_cnt_addr;
+ uint32_t umc_reg_offset =
+ get_umc_v8_14_reg_offset(adev, umc_inst, ch_inst);
+
+ ecc_err_cnt_addr =
+ SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+ /* clear error count */
+ WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4,
+ UMC_V8_14_CE_CNT_INIT);
+
+ return 0;
+}
+
+static void umc_v8_14_clear_error_count(struct amdgpu_device *adev)
+{
+ amdgpu_umc_loop_channels(adev,
+ umc_v8_14_clear_error_count_per_channel, NULL);
+}
+
+static void umc_v8_14_query_correctable_error_count(struct amdgpu_device *adev,
+ uint32_t umc_reg_offset,
+ unsigned long *error_count)
+{
+ uint32_t ecc_err_cnt, ecc_err_cnt_addr;
+
+ /* UMC 8_14 registers */
+ ecc_err_cnt_addr =
+ SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+ ecc_err_cnt = RREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4);
+ *error_count +=
+ (REG_GET_FIELD(ecc_err_cnt, UMCCH0_GeccErrCnt, GeccErrCnt) -
+ UMC_V8_14_CE_CNT_INIT);
+}
+
+static void umc_v8_14_query_uncorrectable_error_count(struct amdgpu_device *adev,
+ uint32_t umc_reg_offset,
+ unsigned long *error_count)
+{
+ uint32_t ecc_err_cnt, ecc_err_cnt_addr;
+ /* UMC 8_14 registers */
+ ecc_err_cnt_addr =
+ SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+ ecc_err_cnt = RREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4);
+ *error_count +=
+ (REG_GET_FIELD(ecc_err_cnt, UMCCH0_GeccErrCnt, GeccUnCorrErrCnt) -
+ UMC_V8_14_CE_CNT_INIT);
+}
+
+static int umc_v8_14_query_error_count_per_channel(struct amdgpu_device *adev,
+ uint32_t node_inst, uint32_t umc_inst,
+ uint32_t ch_inst, void *data)
+{
+ struct ras_err_data *err_data = (struct ras_err_data *)data;
+ uint32_t umc_reg_offset =
+ get_umc_v8_14_reg_offset(adev, umc_inst, ch_inst);
+
+ umc_v8_14_query_correctable_error_count(adev,
+ umc_reg_offset,
+ &(err_data->ce_count));
+ umc_v8_14_query_uncorrectable_error_count(adev,
+ umc_reg_offset,
+ &(err_data->ue_count));
+
+ return 0;
+}
+
+static void umc_v8_14_query_ras_error_count(struct amdgpu_device *adev,
+ void *ras_error_status)
+{
+ amdgpu_umc_loop_channels(adev,
+ umc_v8_14_query_error_count_per_channel, ras_error_status);
+
+ umc_v8_14_clear_error_count(adev);
+}
+
+static int umc_v8_14_err_cnt_init_per_channel(struct amdgpu_device *adev,
+ uint32_t node_inst, uint32_t umc_inst,
+ uint32_t ch_inst, void *data)
+{
+ uint32_t ecc_err_cnt_sel, ecc_err_cnt_sel_addr;
+ uint32_t ecc_err_cnt_addr;
+ uint32_t umc_reg_offset =
+ get_umc_v8_14_reg_offset(adev, umc_inst, ch_inst);
+
+ ecc_err_cnt_sel_addr =
+ SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCntSel);
+ ecc_err_cnt_addr =
+ SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+ ecc_err_cnt_sel = RREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4);
+
+ /* set ce error interrupt type to APIC based interrupt */
+ ecc_err_cnt_sel = REG_SET_FIELD(ecc_err_cnt_sel, UMCCH0_GeccErrCntSel,
+ GeccErrInt, 0x1);
+ WREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4, ecc_err_cnt_sel);
+ /* set error count to initial value */
+ WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4, UMC_V8_14_CE_CNT_INIT);
+
+ return 0;
+}
+
+static void umc_v8_14_err_cnt_init(struct amdgpu_device *adev)
+{
+ amdgpu_umc_loop_channels(adev,
+ umc_v8_14_err_cnt_init_per_channel, NULL);
+}
+
+const struct amdgpu_ras_block_hw_ops umc_v8_14_ras_hw_ops = {
+ .query_ras_error_count = umc_v8_14_query_ras_error_count,
+};
+
+struct amdgpu_umc_ras umc_v8_14_ras = {
+ .ras_block = {
+ .hw_ops = &umc_v8_14_ras_hw_ops,
+ },
+ .err_cnt_init = umc_v8_14_err_cnt_init,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_14.h b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.h
new file mode 100644
index 000000000000..20a258f0017a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.h
@@ -0,0 +1,51 @@
+/*
+ * 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.
+ *
+ */
+#ifndef __UMC_V8_14_H__
+#define __UMC_V8_14_H__
+
+#include "soc15_common.h"
+#include "amdgpu.h"
+
+/* number of umc channel instance with memory map register access */
+#define UMC_V8_14_CHANNEL_INSTANCE_NUM 2
+/* number of umc instance with memory map register access */
+#define UMC_V8_14_UMC_INSTANCE_NUM(adev) ((adev)->umc.node_inst_num)
+
+/* Total channel instances for all available umc nodes */
+#define UMC_V8_14_TOTAL_CHANNEL_NUM(adev) \
+ (UMC_V8_14_CHANNEL_INSTANCE_NUM * (adev)->gmc.num_umc)
+
+/* UMC register per channel offset */
+#define UMC_V8_14_PER_CHANNEL_OFFSET 0x400
+
+#define UMC_V8_14_INST_DIST 0x40000
+
+/* EccErrCnt max value */
+#define UMC_V8_14_CE_CNT_MAX 0xffff
+/* umc ce interrupt threshold */
+#define UMC_V8_14_CE_INT_THRESHOLD 0xffff
+/* umc ce count initial value */
+#define UMC_V8_14_CE_CNT_INIT (UMC_V8_14_CE_CNT_MAX - UMC_V8_14_CE_INT_THRESHOLD)
+
+extern struct amdgpu_umc_ras umc_v8_14_ras;
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
index bdbca25d80c4..5830e799c0a3 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
@@ -790,13 +790,13 @@ static int uvd_v3_1_soft_reset(struct amdgpu_ip_block *ip_block)
return uvd_v3_1_start(adev);
}
-static int uvd_v3_1_set_clockgating_state(void *handle,
+static int uvd_v3_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int uvd_v3_1_set_powergating_state(void *handle,
+static int uvd_v3_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
index a836dc9cfcad..f93079e09215 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
@@ -44,7 +44,7 @@ static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev);
static void uvd_v4_2_set_irq_funcs(struct amdgpu_device *adev);
static int uvd_v4_2_start(struct amdgpu_device *adev);
static void uvd_v4_2_stop(struct amdgpu_device *adev);
-static int uvd_v4_2_set_clockgating_state(void *handle,
+static int uvd_v4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state);
static void uvd_v4_2_set_dcm(struct amdgpu_device *adev,
bool sw_mode);
@@ -708,13 +708,13 @@ static int uvd_v4_2_process_interrupt(struct amdgpu_device *adev,
return 0;
}
-static int uvd_v4_2_set_clockgating_state(void *handle,
+static int uvd_v4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
return 0;
}
-static int uvd_v4_2_set_powergating_state(void *handle,
+static int uvd_v4_2_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the UVD block.
@@ -724,7 +724,7 @@ static int uvd_v4_2_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_PG_STATE_GATE) {
uvd_v4_2_stop(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
index ab55fae3569e..050a0f309390 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
@@ -42,7 +42,7 @@ static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev);
static void uvd_v5_0_set_irq_funcs(struct amdgpu_device *adev);
static int uvd_v5_0_start(struct amdgpu_device *adev);
static void uvd_v5_0_stop(struct amdgpu_device *adev);
-static int uvd_v5_0_set_clockgating_state(void *handle,
+static int uvd_v5_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state);
static void uvd_v5_0_enable_mgcg(struct amdgpu_device *adev,
bool enable);
@@ -155,7 +155,7 @@ static int uvd_v5_0_hw_init(struct amdgpu_ip_block *ip_block)
int r;
amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
- uvd_v5_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+ uvd_v5_0_set_clockgating_state(ip_block, AMD_CG_STATE_UNGATE);
uvd_v5_0_enable_mgcg(adev, true);
r = amdgpu_ring_test_helper(ring);
@@ -790,16 +790,11 @@ static void uvd_v5_0_enable_mgcg(struct amdgpu_device *adev,
}
}
-static int uvd_v5_0_set_clockgating_state(void *handle,
+static int uvd_v5_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
- struct amdgpu_ip_block *ip_block;
-
- ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_UVD);
- if (!ip_block)
- return -EINVAL;
if (enable) {
/* wait for STATUS to clear */
@@ -817,7 +812,7 @@ static int uvd_v5_0_set_clockgating_state(void *handle,
return 0;
}
-static int uvd_v5_0_set_powergating_state(void *handle,
+static int uvd_v5_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the UVD block.
@@ -827,7 +822,7 @@ static int uvd_v5_0_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret = 0;
if (state == AMD_PG_STATE_GATE) {
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
index 39f8c3d3a135..d9d036ee51fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
@@ -48,7 +48,7 @@ static void uvd_v6_0_set_irq_funcs(struct amdgpu_device *adev);
static int uvd_v6_0_start(struct amdgpu_device *adev);
static void uvd_v6_0_stop(struct amdgpu_device *adev);
static void uvd_v6_0_set_sw_clock_gating(struct amdgpu_device *adev);
-static int uvd_v6_0_set_clockgating_state(void *handle,
+static int uvd_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state);
static void uvd_v6_0_enable_mgcg(struct amdgpu_device *adev,
bool enable);
@@ -467,7 +467,7 @@ static int uvd_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
int i, r;
amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
- uvd_v6_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+ uvd_v6_0_set_clockgating_state(ip_block, AMD_CG_STATE_UNGATE);
uvd_v6_0_enable_mgcg(adev, true);
r = amdgpu_ring_test_helper(ring);
@@ -1450,17 +1450,12 @@ static void uvd_v6_0_enable_mgcg(struct amdgpu_device *adev,
}
}
-static int uvd_v6_0_set_clockgating_state(void *handle,
+static int uvd_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
- struct amdgpu_ip_block *ip_block;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
- ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_UVD);
- if (!ip_block)
- return -EINVAL;
-
if (enable) {
/* wait for STATUS to clear */
if (uvd_v6_0_wait_for_idle(ip_block))
@@ -1476,7 +1471,7 @@ static int uvd_v6_0_set_clockgating_state(void *handle,
return 0;
}
-static int uvd_v6_0_set_powergating_state(void *handle,
+static int uvd_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the UVD block.
@@ -1486,7 +1481,7 @@ static int uvd_v6_0_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret = 0;
WREG32(mmUVD_POWER_STATUS, UVD_POWER_STATUS__UVD_PG_EN_MASK);
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index 3c8ab8698af8..9d237b5937fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -1511,7 +1511,7 @@ static int uvd_v7_0_process_interrupt(struct amdgpu_device *adev,
return 0;
}
-static int uvd_v7_0_set_clockgating_state(void *handle,
+static int uvd_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
/* needed for driver unload*/
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
index c1ed91b39415..09fd6ef99b3d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
@@ -284,7 +284,7 @@ static int vce_v2_0_stop(struct amdgpu_device *adev)
return 0;
}
- ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_VCN);
+ ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_VCE);
if (!ip_block)
return -EINVAL;
@@ -578,13 +578,13 @@ static int vce_v2_0_process_interrupt(struct amdgpu_device *adev,
return 0;
}
-static int vce_v2_0_set_clockgating_state(void *handle,
+static int vce_v2_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
bool gate = false;
bool sw_cg = false;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_CG_STATE_GATE) {
gate = true;
@@ -596,7 +596,7 @@ static int vce_v2_0_set_clockgating_state(void *handle,
return 0;
}
-static int vce_v2_0_set_powergating_state(void *handle,
+static int vce_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the VCE block.
@@ -606,7 +606,7 @@ static int vce_v2_0_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_PG_STATE_GATE)
return vce_v2_0_stop(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
index 6bb318a06f19..f8bddcd19b68 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
@@ -65,7 +65,7 @@ static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx);
static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev);
static void vce_v3_0_set_irq_funcs(struct amdgpu_device *adev);
static int vce_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block);
-static int vce_v3_0_set_clockgating_state(void *handle,
+static int vce_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state);
/**
* vce_v3_0_ring_get_rptr - get read pointer
@@ -497,7 +497,7 @@ static int vce_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
return r;
vce_v3_0_stop(adev);
- return vce_v3_0_set_clockgating_state(adev, AMD_CG_STATE_GATE);
+ return vce_v3_0_set_clockgating_state(ip_block, AMD_CG_STATE_GATE);
}
static int vce_v3_0_suspend(struct amdgpu_ip_block *ip_block)
@@ -760,10 +760,10 @@ static int vce_v3_0_process_interrupt(struct amdgpu_device *adev,
return 0;
}
-static int vce_v3_0_set_clockgating_state(void *handle,
+static int vce_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
int i;
@@ -801,7 +801,7 @@ static int vce_v3_0_set_clockgating_state(void *handle,
return 0;
}
-static int vce_v3_0_set_powergating_state(void *handle,
+static int vce_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the VCE block.
@@ -811,7 +811,7 @@ static int vce_v3_0_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret = 0;
if (state == AMD_PG_STATE_GATE) {
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
index 79ee555768a5..335bda64ff5b 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
@@ -684,14 +684,14 @@ static void vce_v4_0_mc_resume(struct amdgpu_device *adev)
~VCE_SYS_INT_EN__VCE_SYS_INT_TRAP_INTERRUPT_EN_MASK);
}
-static int vce_v4_0_set_clockgating_state(void *handle,
+static int vce_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
/* needed for driver unload*/
return 0;
}
-static int vce_v4_0_set_powergating_state(void *handle,
+static int vce_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the VCE block.
@@ -701,7 +701,7 @@ static int vce_v4_0_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == AMD_PG_STATE_GATE)
return vce_v4_0_stop(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index 10e99c926fb8..5ea96c983517 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -85,7 +85,8 @@ static int vcn_v1_0_stop(struct amdgpu_device *adev);
static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev);
static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v1_0_set_powergating_state(void *handle, enum amd_powergating_state state);
+static int vcn_v1_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state);
static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -281,7 +282,7 @@ static int vcn_v1_0_hw_fini(struct amdgpu_ip_block *ip_block)
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(VCN, 0, mmUVD_STATUS))) {
- vcn_v1_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ vcn_v1_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
return 0;
@@ -303,7 +304,7 @@ static int vcn_v1_0_suspend(struct amdgpu_ip_block *ip_block)
idle_work_unexecuted = cancel_delayed_work_sync(&adev->vcn.idle_work);
if (idle_work_unexecuted) {
if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ amdgpu_dpm_enable_vcn(adev, false, 0);
}
r = vcn_v1_0_hw_fini(ip_block);
@@ -344,7 +345,7 @@ static int vcn_v1_0_resume(struct amdgpu_ip_block *ip_block)
*/
static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_device *adev)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[0]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
@@ -411,7 +412,7 @@ static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_device *adev)
static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_device *adev)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[0]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
@@ -1394,15 +1395,15 @@ static int vcn_v1_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return ret;
}
-static int vcn_v1_0_set_clockgating_state(void *handle,
+static int vcn_v1_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
if (enable) {
/* wait for STATUS to clear */
- if (!vcn_v1_0_is_idle(handle))
+ if (!vcn_v1_0_is_idle(adev))
return -EBUSY;
vcn_v1_0_enable_clock_gating(adev);
} else {
@@ -1799,7 +1800,7 @@ static void vcn_v1_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t coun
}
}
-static int vcn_v1_0_set_powergating_state(void *handle,
+static int vcn_v1_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the VCN block.
@@ -1810,7 +1811,7 @@ static int vcn_v1_0_set_powergating_state(void *handle,
* the smc and the hw blocks
*/
int ret;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (state == adev->vcn.cur_state)
return 0;
@@ -1856,7 +1857,7 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work)
if (fences == 0) {
amdgpu_gfx_off_ctrl(adev, true);
if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ amdgpu_dpm_enable_vcn(adev, false, 0);
else
amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
AMD_PG_STATE_GATE);
@@ -1886,7 +1887,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks)
if (set_clocks) {
amdgpu_gfx_off_ctrl(adev, false);
if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ amdgpu_dpm_enable_vcn(adev, true, 0);
else
amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
AMD_PG_STATE_UNGATE);
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
index e0322cbca3ec..e42cfc731ad8 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
@@ -92,7 +92,7 @@ static const struct amdgpu_hwip_reg_entry vcn_reg_list_2_0[] = {
static void vcn_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void vcn_v2_0_set_enc_ring_funcs(struct amdgpu_device *adev);
static void vcn_v2_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v2_0_set_powergating_state(void *handle,
+static int vcn_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v2_0_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -318,7 +318,7 @@ static int vcn_v2_0_hw_fini(struct amdgpu_ip_block *ip_block)
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(VCN, 0, mmUVD_STATUS)))
- vcn_v2_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ vcn_v2_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
return 0;
}
@@ -372,7 +372,7 @@ static int vcn_v2_0_resume(struct amdgpu_ip_block *ip_block)
*/
static void vcn_v2_0_mc_resume(struct amdgpu_device *adev)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[0]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
uint32_t offset;
if (amdgpu_sriov_vf(adev))
@@ -428,7 +428,7 @@ static void vcn_v2_0_mc_resume(struct amdgpu_device *adev)
static void vcn_v2_0_mc_resume_dpg_mode(struct amdgpu_device *adev, bool indirect)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[0]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
@@ -978,7 +978,7 @@ static int vcn_v2_0_start(struct amdgpu_device *adev)
int i, j, r;
if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ amdgpu_dpm_enable_vcn(adev, true, 0);
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
return vcn_v2_0_start_dpg_mode(adev, adev->vcn.indirect_sram);
@@ -1235,7 +1235,7 @@ static int vcn_v2_0_stop(struct amdgpu_device *adev)
power_off:
if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ amdgpu_dpm_enable_vcn(adev, false, 0);
return 0;
}
@@ -1335,10 +1335,10 @@ static int vcn_v2_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return ret;
}
-static int vcn_v2_0_set_clockgating_state(void *handle,
+static int vcn_v2_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
@@ -1346,7 +1346,7 @@ static int vcn_v2_0_set_clockgating_state(void *handle,
if (enable) {
/* wait for STATUS to clear */
- if (!vcn_v2_0_is_idle(handle))
+ if (!vcn_v2_0_is_idle(adev))
return -EBUSY;
vcn_v2_0_enable_clock_gating(adev);
} else {
@@ -1796,7 +1796,7 @@ int vcn_v2_0_dec_ring_test_ring(struct amdgpu_ring *ring)
}
-static int vcn_v2_0_set_powergating_state(void *handle,
+static int vcn_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
/* This doesn't actually powergate the VCN block.
@@ -1807,7 +1807,7 @@ static int vcn_v2_0_set_powergating_state(void *handle,
* the smc and the hw blocks
*/
int ret;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev)) {
adev->vcn.cur_state = AMD_PG_STATE_UNGATE;
@@ -1920,7 +1920,7 @@ static int vcn_v2_0_start_sriov(struct amdgpu_device *adev)
init_table += header->vcn_table_offset;
- size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[0]->size + 4);
+ size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
MMSCH_V2_0_INSERT_DIRECT_RD_MOD_WT(
SOC15_REG_OFFSET(UVD, i, mmUVD_STATUS),
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
index 6aa08281d094..b518202955ca 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
@@ -95,7 +95,7 @@ static const struct amdgpu_hwip_reg_entry vcn_reg_list_2_5[] = {
static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev);
static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev);
static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v2_5_set_powergating_state(void *handle,
+static int vcn_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v2_5_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -399,7 +399,7 @@ static int vcn_v2_5_hw_fini(struct amdgpu_ip_block *ip_block)
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(VCN, i, mmUVD_STATUS)))
- vcn_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ vcn_v2_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN))
amdgpu_irq_put(adev, &adev->vcn.inst[i].ras_poison_irq, 0);
@@ -465,7 +465,7 @@ static void vcn_v2_5_mc_resume(struct amdgpu_device *adev)
if (adev->vcn.harvest_config & (1 << i))
continue;
- size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[i]->size + 4);
+ size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
/* cache window 0: fw */
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
@@ -514,7 +514,7 @@ static void vcn_v2_5_mc_resume(struct amdgpu_device *adev)
static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[inst_idx]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst_idx].fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
@@ -1012,8 +1012,10 @@ static int vcn_v2_5_start(struct amdgpu_device *adev)
uint32_t rb_bufsz, tmp;
int i, j, k, r;
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, true, i);
+ }
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
@@ -1285,7 +1287,7 @@ static int vcn_v2_5_sriov_start(struct amdgpu_device *adev)
SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS),
~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
- size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[i]->size + 4);
+ size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
/* mc resume*/
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
MMSCH_V1_0_INSERT_DIRECT_WT(
@@ -1485,8 +1487,10 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev)
~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
}
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, false, i);
+ }
return 0;
}
@@ -1778,6 +1782,7 @@ static bool vcn_v2_5_is_idle(void *handle)
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
continue;
+
ret &= (RREG32_SOC15(VCN, i, mmUVD_STATUS) == UVD_STATUS__IDLE);
}
@@ -1801,17 +1806,17 @@ static int vcn_v2_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
return ret;
}
-static int vcn_v2_5_set_clockgating_state(void *handle,
+static int vcn_v2_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
return 0;
if (enable) {
- if (!vcn_v2_5_is_idle(handle))
+ if (!vcn_v2_5_is_idle(adev))
return -EBUSY;
vcn_v2_5_enable_clock_gating(adev);
} else {
@@ -1821,10 +1826,10 @@ static int vcn_v2_5_set_clockgating_state(void *handle,
return 0;
}
-static int vcn_v2_5_set_powergating_state(void *handle,
+static int vcn_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (amdgpu_sriov_vf(adev))
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 6732ad7f16f5..63ddd4cca910 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -105,7 +105,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev);
static void vcn_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev);
static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v3_0_set_powergating_state(void *handle,
+static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -430,9 +430,9 @@ static int vcn_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
if (!amdgpu_sriov_vf(adev)) {
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
- (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
- RREG32_SOC15(VCN, i, mmUVD_STATUS))) {
- vcn_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
+ RREG32_SOC15(VCN, i, mmUVD_STATUS))) {
+ vcn_v3_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
}
}
@@ -490,7 +490,7 @@ static int vcn_v3_0_resume(struct amdgpu_ip_block *ip_block)
*/
static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[inst]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst].fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
@@ -540,7 +540,7 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
{
- uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[inst_idx]->size + 4);
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst_idx].fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
@@ -1141,8 +1141,10 @@ static int vcn_v3_0_start(struct amdgpu_device *adev)
uint32_t rb_bufsz, tmp;
int i, j, k, r;
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, true, i);
+ }
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
@@ -1373,7 +1375,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev)
mmUVD_STATUS),
~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
- cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[i]->size + 4);
+ cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
MMSCH_V3_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
@@ -1632,8 +1634,10 @@ static int vcn_v3_0_stop(struct amdgpu_device *adev)
vcn_v3_0_enable_static_power_gating(adev, i);
}
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, false, i);
+ }
return 0;
}
@@ -2132,10 +2136,10 @@ static int vcn_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
return ret;
}
-static int vcn_v3_0_set_clockgating_state(void *handle,
+static int vcn_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = state == AMD_CG_STATE_GATE;
int i;
@@ -2155,10 +2159,10 @@ static int vcn_v3_0_set_clockgating_state(void *handle,
return 0;
}
-static int vcn_v3_0_set_powergating_state(void *handle,
+static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
/* for SRIOV, guest should not control VCN Power-gating
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
index fcc8511e91ee..00551d6f0370 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
@@ -96,7 +96,7 @@ static int amdgpu_ih_clientid_vcns[] = {
static int vcn_v4_0_start_sriov(struct amdgpu_device *adev);
static void vcn_v4_0_set_unified_ring_funcs(struct amdgpu_device *adev);
static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v4_0_set_powergating_state(void *handle,
+static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v4_0_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -366,9 +366,9 @@ static int vcn_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
continue;
if (!amdgpu_sriov_vf(adev)) {
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
- (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
- RREG32_SOC15(VCN, i, regUVD_STATUS))) {
- vcn_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
+ RREG32_SOC15(VCN, i, regUVD_STATUS))) {
+ vcn_v4_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
}
if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN))
@@ -431,7 +431,7 @@ static void vcn_v4_0_mc_resume(struct amdgpu_device *adev, int inst)
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -491,7 +491,7 @@ static void vcn_v4_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
{
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst_idx]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -1097,8 +1097,10 @@ static int vcn_v4_0_start(struct amdgpu_device *adev)
uint32_t tmp;
int i, j, k, r;
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, true, i);
+ }
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
@@ -1341,7 +1343,7 @@ static int vcn_v4_0_start_sriov(struct amdgpu_device *adev)
regUVD_STATUS),
~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
- cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw[i]->size + 4);
+ cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
@@ -1623,8 +1625,10 @@ static int vcn_v4_0_stop(struct amdgpu_device *adev)
vcn_v4_0_enable_static_power_gating(adev, i);
}
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, false, i);
+ }
return 0;
}
@@ -2007,14 +2011,15 @@ static int vcn_v4_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
/**
* vcn_v4_0_set_clockgating_state - set VCN block clockgating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: clock gating state
*
* Set VCN block clockgating state
*/
-static int vcn_v4_0_set_clockgating_state(void *handle, enum amd_clockgating_state state)
+static int vcn_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = state == AMD_CG_STATE_GATE;
int i;
@@ -2037,14 +2042,15 @@ static int vcn_v4_0_set_clockgating_state(void *handle, enum amd_clockgating_sta
/**
* vcn_v4_0_set_powergating_state - set VCN block powergating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: power gating state
*
* Set VCN block powergating state
*/
-static int vcn_v4_0_set_powergating_state(void *handle, enum amd_powergating_state state)
+static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
/* for SRIOV, guest should not control VCN Power-gating
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 3f69b9b2bcd0..ecdc027f8220 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
@@ -87,7 +87,7 @@ static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0_3[] = {
static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev);
static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev);
static void vcn_v4_0_3_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v4_0_3_set_powergating_state(void *handle,
+static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v4_0_3_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -349,7 +349,7 @@ static int vcn_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block)
cancel_delayed_work_sync(&adev->vcn.idle_work);
if (adev->vcn.cur_state != AMD_PG_STATE_GATE)
- vcn_v4_0_3_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ vcn_v4_0_3_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
return 0;
}
@@ -407,7 +407,7 @@ static void vcn_v4_0_3_mc_resume(struct amdgpu_device *adev, int inst_idx)
uint32_t offset, size, vcn_inst;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst_idx]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
vcn_inst = GET_INST(VCN, inst_idx);
@@ -482,7 +482,7 @@ static void vcn_v4_0_3_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_i
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst_idx]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -957,6 +957,8 @@ static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev)
for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
vcn_inst = GET_INST(VCN, i);
+ vcn_v4_0_3_fw_shared_init(adev, vcn_inst);
+
memset(&header, 0, sizeof(struct mmsch_v4_0_3_init_header));
header.version = MMSCH_VERSION;
header.total_size = sizeof(struct mmsch_v4_0_3_init_header) >> 2;
@@ -969,7 +971,7 @@ static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev)
MMSCH_V4_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.fw[i]->size + 4);
+ cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
@@ -1121,8 +1123,10 @@ static int vcn_v4_0_3_start(struct amdgpu_device *adev)
int i, j, k, r, vcn_inst;
uint32_t tmp;
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, true, i);
+ }
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
@@ -1395,8 +1399,10 @@ static int vcn_v4_0_3_stop(struct amdgpu_device *adev)
vcn_v4_0_3_enable_clock_gating(adev, i);
}
Done:
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, false, i);
+ }
return 0;
}
@@ -1616,15 +1622,15 @@ static int vcn_v4_0_3_wait_for_idle(struct amdgpu_ip_block *ip_block)
/* vcn_v4_0_3_set_clockgating_state - set VCN block clockgating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: clock gating state
*
* Set VCN block clockgating state
*/
-static int vcn_v4_0_3_set_clockgating_state(void *handle,
+static int vcn_v4_0_3_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = state == AMD_CG_STATE_GATE;
int i;
@@ -1644,15 +1650,15 @@ static int vcn_v4_0_3_set_clockgating_state(void *handle,
/**
* vcn_v4_0_3_set_powergating_state - set VCN block powergating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: power gating state
*
* Set VCN block powergating state
*/
-static int vcn_v4_0_3_set_powergating_state(void *handle,
+static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
/* for SRIOV, guest should not control VCN Power-gating
@@ -1911,9 +1917,94 @@ static const struct amdgpu_ras_block_hw_ops vcn_v4_0_3_ras_hw_ops = {
.reset_ras_error_count = vcn_v4_0_3_reset_ras_error_count,
};
+static int vcn_v4_0_3_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+ enum aca_smu_type type, void *data)
+{
+ struct aca_bank_info info;
+ u64 misc0;
+ int ret;
+
+ ret = aca_bank_info_decode(bank, &info);
+ if (ret)
+ return ret;
+
+ misc0 = bank->regs[ACA_REG_IDX_MISC0];
+ switch (type) {
+ case ACA_SMU_TYPE_UE:
+ ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+ 1ULL);
+ break;
+ case ACA_SMU_TYPE_CE:
+ ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE,
+ ACA_REG__MISC0__ERRCNT(misc0));
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return ret;
+}
+
+/* reference to smu driver if header file */
+static int vcn_v4_0_3_err_codes[] = {
+ 14, 15, /* VCN */
+};
+
+static bool vcn_v4_0_3_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+ enum aca_smu_type type, void *data)
+{
+ u32 instlo;
+
+ instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+ instlo &= GENMASK(31, 1);
+
+ if (instlo != mmSMNAID_AID0_MCA_SMU)
+ return false;
+
+ if (aca_bank_check_error_codes(handle->adev, bank,
+ vcn_v4_0_3_err_codes,
+ ARRAY_SIZE(vcn_v4_0_3_err_codes)))
+ return false;
+
+ return true;
+}
+
+static const struct aca_bank_ops vcn_v4_0_3_aca_bank_ops = {
+ .aca_bank_parser = vcn_v4_0_3_aca_bank_parser,
+ .aca_bank_is_valid = vcn_v4_0_3_aca_bank_is_valid,
+};
+
+static const struct aca_info vcn_v4_0_3_aca_info = {
+ .hwip = ACA_HWIP_TYPE_SMU,
+ .mask = ACA_ERROR_UE_MASK,
+ .bank_ops = &vcn_v4_0_3_aca_bank_ops,
+};
+
+static int vcn_v4_0_3_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+ int r;
+
+ r = amdgpu_ras_block_late_init(adev, ras_block);
+ if (r)
+ return r;
+
+ r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__VCN,
+ &vcn_v4_0_3_aca_info, NULL);
+ if (r)
+ goto late_fini;
+
+ return 0;
+
+late_fini:
+ amdgpu_ras_block_late_fini(adev, ras_block);
+
+ return r;
+}
+
static struct amdgpu_vcn_ras vcn_v4_0_3_ras = {
.ras_block = {
.hw_ops = &vcn_v4_0_3_ras_hw_ops,
+ .ras_late_init = vcn_v4_0_3_ras_late_init,
},
};
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 71961fb3f7ff..23d3c16c9d9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
@@ -95,7 +95,7 @@ static int amdgpu_ih_clientid_vcns[] = {
static void vcn_v4_0_5_set_unified_ring_funcs(struct amdgpu_device *adev);
static void vcn_v4_0_5_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v4_0_5_set_powergating_state(void *handle,
+static int vcn_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v4_0_5_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -309,7 +309,7 @@ static int vcn_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block)
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(VCN, i, regUVD_STATUS))) {
- vcn_v4_0_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ vcn_v4_0_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
}
}
@@ -370,7 +370,7 @@ static void vcn_v4_0_5_mc_resume(struct amdgpu_device *adev, int inst)
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -431,7 +431,7 @@ static void vcn_v4_0_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_i
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst_idx]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -1000,8 +1000,10 @@ static int vcn_v4_0_5_start(struct amdgpu_device *adev)
uint32_t tmp;
int i, j, k, r;
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, true, i);
+ }
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
@@ -1277,8 +1279,10 @@ static int vcn_v4_0_5_stop(struct amdgpu_device *adev)
vcn_v4_0_5_enable_static_power_gating(adev, i);
}
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, false, i);
+ }
return 0;
}
@@ -1492,14 +1496,15 @@ static int vcn_v4_0_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
/**
* vcn_v4_0_5_set_clockgating_state - set VCN block clockgating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: clock gating state
*
* Set VCN block clockgating state
*/
-static int vcn_v4_0_5_set_clockgating_state(void *handle, enum amd_clockgating_state state)
+static int vcn_v4_0_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
int i;
@@ -1522,14 +1527,15 @@ static int vcn_v4_0_5_set_clockgating_state(void *handle, enum amd_clockgating_s
/**
* vcn_v4_0_5_set_powergating_state - set VCN block powergating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: power gating state
*
* Set VCN block powergating state
*/
-static int vcn_v4_0_5_set_powergating_state(void *handle, enum amd_powergating_state state)
+static int vcn_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (state == adev->vcn.cur_state)
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 bd3d2bbdc16b..b6d78381ebfb 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
@@ -32,7 +32,7 @@
#include "vcn/vcn_5_0_0_offset.h"
#include "vcn/vcn_5_0_0_sh_mask.h"
-#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_5_0.h"
#include "vcn_v5_0_0.h"
#include <drm/drm_drv.h>
@@ -78,7 +78,7 @@ static int amdgpu_ih_clientid_vcns[] = {
static void vcn_v5_0_0_set_unified_ring_funcs(struct amdgpu_device *adev);
static void vcn_v5_0_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v5_0_0_set_powergating_state(void *handle,
+static int vcn_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state);
static int vcn_v5_0_0_pause_dpg_mode(struct amdgpu_device *adev,
int inst_idx, struct dpg_pause_state *new_state);
@@ -105,6 +105,21 @@ static int vcn_v5_0_0_early_init(struct amdgpu_ip_block *ip_block)
return amdgpu_vcn_early_init(adev);
}
+void vcn_v5_0_0_alloc_ip_dump(struct amdgpu_device *adev)
+{
+ uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_5_0);
+ uint32_t *ptr;
+
+ /* Allocate memory for VCN IP Dump buffer */
+ ptr = kcalloc(adev->vcn.num_vcn_inst * reg_count, sizeof(uint32_t), GFP_KERNEL);
+ if (!ptr) {
+ DRM_ERROR("Failed to allocate memory for VCN IP Dump\n");
+ adev->vcn.ip_dump = NULL;
+ } else {
+ adev->vcn.ip_dump = ptr;
+ }
+}
+
/**
* vcn_v5_0_0_sw_init - sw init for VCN block
*
@@ -117,8 +132,6 @@ static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block)
struct amdgpu_ring *ring;
struct amdgpu_device *adev = ip_block->adev;
int i, r;
- uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_5_0);
- uint32_t *ptr;
r = amdgpu_vcn_sw_init(adev);
if (r)
@@ -140,13 +153,13 @@ static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block)
/* VCN UNIFIED TRAP */
r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
- VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[i].irq);
+ VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[i].irq);
if (r)
return r;
/* VCN POISON TRAP */
r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
- VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq);
+ VCN_5_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq);
if (r)
return r;
@@ -177,14 +190,7 @@ static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block)
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
adev->vcn.pause_dpg_mode = vcn_v5_0_0_pause_dpg_mode;
- /* Allocate memory for VCN IP Dump buffer */
- ptr = kcalloc(adev->vcn.num_vcn_inst * reg_count, sizeof(uint32_t), GFP_KERNEL);
- if (!ptr) {
- DRM_ERROR("Failed to allocate memory for VCN IP Dump\n");
- adev->vcn.ip_dump = NULL;
- } else {
- adev->vcn.ip_dump = ptr;
- }
+ vcn_v5_0_0_alloc_ip_dump(adev);
r = amdgpu_vcn_sysfs_reset_mask_init(adev);
if (r)
@@ -283,7 +289,7 @@ static int vcn_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block)
if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(VCN, i, regUVD_STATUS))) {
- vcn_v5_0_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+ vcn_v5_0_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
}
}
}
@@ -344,7 +350,7 @@ static void vcn_v5_0_0_mc_resume(struct amdgpu_device *adev, int inst)
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -405,7 +411,7 @@ static void vcn_v5_0_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_i
uint32_t offset, size;
const struct common_firmware_header *hdr;
- hdr = (const struct common_firmware_header *)adev->vcn.fw[inst_idx]->data;
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
/* cache window 0: fw */
@@ -771,8 +777,10 @@ static int vcn_v5_0_0_start(struct amdgpu_device *adev)
uint32_t tmp;
int i, j, k, r;
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, true);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, true, i);
+ }
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
@@ -1018,8 +1026,10 @@ static int vcn_v5_0_0_stop(struct amdgpu_device *adev)
vcn_v5_0_0_enable_static_power_gating(adev, i);
}
- if (adev->pm.dpm_enabled)
- amdgpu_dpm_enable_uvd(adev, false);
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_vcn(adev, false, i);
+ }
return 0;
}
@@ -1229,14 +1239,15 @@ static int vcn_v5_0_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
/**
* vcn_v5_0_0_set_clockgating_state - set VCN block clockgating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: clock gating state
*
* Set VCN block clockgating state
*/
-static int vcn_v5_0_0_set_clockgating_state(void *handle, enum amd_clockgating_state state)
+static int vcn_v5_0_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
int i;
@@ -1259,14 +1270,15 @@ static int vcn_v5_0_0_set_clockgating_state(void *handle, enum amd_clockgating_s
/**
* vcn_v5_0_0_set_powergating_state - set VCN block powergating state
*
- * @handle: amdgpu_device pointer
+ * @ip_block: amdgpu_ip_block pointer
* @state: power gating state
*
* Set VCN block powergating state
*/
-static int vcn_v5_0_0_set_powergating_state(void *handle, enum amd_powergating_state state)
+static int vcn_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int ret;
if (state == adev->vcn.cur_state)
@@ -1312,10 +1324,10 @@ static int vcn_v5_0_0_process_interrupt(struct amdgpu_device *adev, struct amdgp
DRM_DEBUG("IH: VCN TRAP\n");
switch (entry->src_id) {
- case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+ case VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]);
break;
- case VCN_4_0__SRCID_UVD_POISON:
+ case VCN_5_0__SRCID_UVD_POISON:
amdgpu_vcn_process_poison_irq(adev, source, entry);
break;
default:
@@ -1351,7 +1363,8 @@ static void vcn_v5_0_0_set_irq_funcs(struct amdgpu_device *adev)
}
}
-static void vcn_v5_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+void vcn_v5_0_0_print_ip_state(struct amdgpu_ip_block *ip_block,
+ struct drm_printer *p)
{
struct amdgpu_device *adev = ip_block->adev;
int i, j;
@@ -1383,7 +1396,7 @@ static void vcn_v5_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm
}
}
-static void vcn_v5_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
+void vcn_v5_0_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
{
struct amdgpu_device *adev = ip_block->adev;
int i, j;
@@ -1424,8 +1437,8 @@ static const struct amd_ip_funcs vcn_v5_0_0_ip_funcs = {
.wait_for_idle = vcn_v5_0_0_wait_for_idle,
.set_clockgating_state = vcn_v5_0_0_set_clockgating_state,
.set_powergating_state = vcn_v5_0_0_set_powergating_state,
- .dump_ip_state = vcn_v5_0_dump_ip_state,
- .print_ip_state = vcn_v5_0_print_ip_state,
+ .dump_ip_state = vcn_v5_0_0_dump_ip_state,
+ .print_ip_state = vcn_v5_0_0_print_ip_state,
};
const struct amdgpu_ip_block_version vcn_v5_0_0_ip_block = {
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h
index 51bbccd4360f..b8927652bc50 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h
@@ -32,6 +32,11 @@
#define VCN_VID_IP_ADDRESS 0x0
#define VCN_AON_IP_ADDRESS 0x30000
+void vcn_v5_0_0_alloc_ip_dump(struct amdgpu_device *adev);
+void vcn_v5_0_0_print_ip_state(struct amdgpu_ip_block *ip_block,
+ struct drm_printer *p);
+void vcn_v5_0_0_dump_ip_state(struct amdgpu_ip_block *ip_block);
+
extern const struct amdgpu_ip_block_version vcn_v5_0_0_ip_block;
#endif /* __VCN_V5_0_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
new file mode 100644
index 000000000000..8b463c977d08
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
@@ -0,0 +1,1118 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * 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/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_hw_ip.h"
+#include "vcn_v2_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"
+#include "vcn_v5_0_0.h"
+#include "vcn_v5_0_1.h"
+
+#include <drm/drm_drv.h>
+
+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_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state);
+static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring);
+
+/**
+ * vcn_v5_0_1_early_init - set function pointers and load microcode
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ * Load microcode from filesystem
+ */
+static int vcn_v5_0_1_early_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+
+ /* re-use enc ring as unified ring */
+ adev->vcn.num_enc_rings = 1;
+
+ vcn_v5_0_1_set_unified_ring_funcs(adev);
+ vcn_v5_0_1_set_irq_funcs(adev);
+
+ return amdgpu_vcn_early_init(adev);
+}
+
+/**
+ * vcn_v5_0_1_sw_init - sw init for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ring *ring;
+ int i, r, vcn_inst;
+
+ r = amdgpu_vcn_sw_init(adev);
+ if (r)
+ return r;
+
+ amdgpu_vcn_setup_ucode(adev);
+
+ r = amdgpu_vcn_resume(adev);
+ if (r)
+ return r;
+
+ /* VCN UNIFIED TRAP */
+ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+ VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq);
+ if (r)
+ return r;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+ volatile struct amdgpu_vcn5_fw_shared *fw_shared;
+
+ vcn_inst = GET_INST(VCN, i);
+
+ 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;
+
+ ring->vm_hub = AMDGPU_MMHUB0(adev->vcn.inst[i].aid_id);
+ sprintf(ring->name, "vcn_unified_%d", adev->vcn.inst[i].aid_id);
+
+ r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
+ AMDGPU_RING_PRIO_DEFAULT, &adev->vcn.inst[i].sched_score);
+ if (r)
+ return r;
+
+ fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+ fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+ fw_shared->sq.is_enabled = true;
+
+ if (amdgpu_vcnfw_log)
+ amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+ }
+
+ /* 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]);
+
+ vcn_v5_0_0_alloc_ip_dump(adev);
+
+ return amdgpu_vcn_sysfs_reset_mask_init(adev);
+}
+
+/**
+ * vcn_v5_0_1_sw_fini - sw fini for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v5_0_1_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int i, r, idx;
+
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+ volatile struct amdgpu_vcn4_fw_shared *fw_shared;
+
+ fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+ fw_shared->present_flag_0 = 0;
+ fw_shared->sq.is_enabled = 0;
+ }
+
+ drm_dev_exit(idx);
+ }
+
+ r = amdgpu_vcn_suspend(adev);
+ if (r)
+ return r;
+
+ r = amdgpu_vcn_sw_fini(adev);
+
+ amdgpu_vcn_sysfs_reset_mask_fini(adev);
+
+ kfree(adev->vcn.ip_dump);
+
+ return r;
+}
+
+/**
+ * vcn_v5_0_1_hw_init - start and test VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ring *ring;
+ int i, r, vcn_inst;
+
+ 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);
+
+ r = amdgpu_ring_test_helper(ring);
+ if (r)
+ return r;
+ }
+
+ return 0;
+}
+
+/**
+ * vcn_v5_0_1_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+
+ cancel_delayed_work_sync(&adev->vcn.idle_work);
+
+ return 0;
+}
+
+/**
+ * vcn_v5_0_1_suspend - suspend VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v5_0_1_suspend(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int r;
+
+ r = vcn_v5_0_1_hw_fini(ip_block);
+ if (r)
+ return r;
+
+ r = amdgpu_vcn_suspend(adev);
+
+ return r;
+}
+
+/**
+ * vcn_v5_0_1_resume - resume VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v5_0_1_resume(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int r;
+
+ r = amdgpu_vcn_resume(adev);
+ if (r)
+ return r;
+
+ r = vcn_v5_0_1_hw_init(ip_block);
+
+ return r;
+}
+
+/**
+ * vcn_v5_0_1_mc_resume - memory controller programming
+ *
+ * @adev: amdgpu_device pointer
+ * @inst: instance number
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v5_0_1_mc_resume(struct amdgpu_device *adev, int inst)
+{
+ uint32_t offset, size, vcn_inst;
+ const struct common_firmware_header *hdr;
+
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
+ size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+ vcn_inst = GET_INST(VCN, inst);
+ /* cache window 0: fw */
+ if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+ (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_lo));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+ (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_hi));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0, 0);
+ offset = 0;
+ } else {
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+ lower_32_bits(adev->vcn.inst[inst].gpu_addr));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+ upper_32_bits(adev->vcn.inst[inst].gpu_addr));
+ offset = size;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0,
+ AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+ }
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE0, size);
+
+ /* cache window 1: stack */
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+ lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+ upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET1, 0);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+ /* cache window 2: context */
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+ lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+ upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET2, 0);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+ /* non-cache window */
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+ lower_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+ upper_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_NONCACHE_OFFSET0, 0);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_NONCACHE_SIZE0,
+ AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+}
+
+/**
+ * vcn_v5_0_1_mc_resume_dpg_mode - memory controller programming for dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ * @indirect: indirectly write sram
+ *
+ * Let the VCN memory controller know it's offsets with dpg mode
+ */
+static void vcn_v5_0_1_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+{
+ uint32_t offset, size;
+ const struct common_firmware_header *hdr;
+
+ hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+ size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+ /* cache window 0: fw */
+ if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+ if (!indirect) {
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+ (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN +
+ inst_idx].tmr_mc_addr_lo), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+ (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN +
+ inst_idx].tmr_mc_addr_hi), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+ } else {
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+ }
+ offset = 0;
+ } else {
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+ lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+ upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+ offset = size;
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_OFFSET0),
+ AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+ }
+
+ if (!indirect)
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+ else
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+ /* cache window 1: stack */
+ if (!indirect) {
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+ lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+ upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+ } else {
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+ }
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+ /* cache window 2: context */
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+ lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+ AMDGPU_VCN_STACK_SIZE), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+ upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+ AMDGPU_VCN_STACK_SIZE), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+ /* non-cache window */
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+ lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+ upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_NONCACHE_SIZE0),
+ AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)), 0, indirect);
+
+ /* 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_v5_0_1_disable_clock_gating - disable VCN clock gating
+ *
+ * @adev: amdgpu_device pointer
+ * @inst: instance number
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v5_0_1_disable_clock_gating(struct amdgpu_device *adev, int inst)
+{
+}
+
+/**
+ * vcn_v5_0_1_enable_clock_gating - enable VCN clock gating
+ *
+ * @adev: amdgpu_device pointer
+ * @inst: instance number
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v5_0_1_enable_clock_gating(struct amdgpu_device *adev, int inst)
+{
+}
+
+/**
+ * vcn_v5_0_1_start_dpg_mode - VCN start with dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ * @indirect: indirectly write sram
+ *
+ * Start VCN block with dpg mode
+ */
+static int vcn_v5_0_1_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+{
+ volatile struct amdgpu_vcn4_fw_shared *fw_shared =
+ adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+ struct amdgpu_ring *ring;
+ int vcn_inst;
+ uint32_t tmp;
+
+ vcn_inst = GET_INST(VCN, inst_idx);
+
+ /* disable register anti-hang mechanism */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 1,
+ ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+ /* enable dynamic power gating mode */
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS);
+ tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS, tmp);
+
+ if (indirect) {
+ adev->vcn.inst[inst_idx].dpg_sram_curr_addr =
+ (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+ /* Use dummy register 0xDEADBEEF passing AID selection to PSP FW */
+ WREG32_SOC24_DPG_MODE(inst_idx, 0xDEADBEEF,
+ adev->vcn.inst[inst_idx].aid_id, 0, true);
+ }
+
+ /* enable VCPU clock */
+ tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+ tmp |= UVD_VCPU_CNTL__CLK_EN_MASK | UVD_VCPU_CNTL__BLK_RST_MASK;
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+ /* disable master interrupt */
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_MASTINT_EN), 0, 0, indirect);
+
+ /* setup regUVD_LMI_CTRL */
+ tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+ UVD_LMI_CTRL__REQ_MODE_MASK |
+ UVD_LMI_CTRL__CRC_RESET_MASK |
+ UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+ UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+ UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+ (8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+ 0x00100000L);
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_CTRL), tmp, 0, indirect);
+
+ vcn_v5_0_1_mc_resume_dpg_mode(adev, inst_idx, indirect);
+
+ tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+ tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+ /* enable LMI MC and UMC channels */
+ tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT;
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_LMI_CTRL2), tmp, 0, indirect);
+
+ /* enable master interrupt */
+ WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+ VCN, 0, regUVD_MASTINT_EN),
+ UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+ if (indirect)
+ amdgpu_vcn_psp_update_sram(adev, inst_idx, AMDGPU_UCODE_ID_VCN0_RAM);
+
+ ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, lower_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, ring->ring_size / sizeof(uint32_t));
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+ tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+ WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+ fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0);
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, tmp);
+ ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+ tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+ fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+ 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_v5_0_1_start - VCN start
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Start VCN block
+ */
+static int vcn_v5_0_1_start(struct amdgpu_device *adev)
+{
+ volatile struct amdgpu_vcn4_fw_shared *fw_shared;
+ struct amdgpu_ring *ring;
+ uint32_t tmp;
+ int i, j, k, r, vcn_inst;
+
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_uvd(adev, true);
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+
+ if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+ r = vcn_v5_0_1_start_dpg_mode(adev, i, adev->vcn.indirect_sram);
+ continue;
+ }
+
+ vcn_inst = GET_INST(VCN, i);
+
+ /* set VCN status busy */
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp);
+
+ /* enable VCPU clock */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+ UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+ /* disable master interrupt */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0,
+ ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+ /* enable LMI MC and UMC channels */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0,
+ ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+ tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+ tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+ /* setup regUVD_LMI_CTRL */
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, tmp |
+ UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+ UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+ UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+ UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+ vcn_v5_0_1_mc_resume(adev, i);
+
+ /* VCN global tiling registers */
+ WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG,
+ adev->gfx.config.gb_addr_config);
+
+ /* unblock VCPU register access */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0,
+ ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+ /* release VCPU reset to boot */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+ ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+ for (j = 0; j < 10; ++j) {
+ uint32_t status;
+
+ for (k = 0; k < 100; ++k) {
+ status = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+ if (status & 2)
+ break;
+ mdelay(100);
+ if (amdgpu_emu_mode == 1)
+ msleep(20);
+ }
+
+ if (amdgpu_emu_mode == 1) {
+ r = -1;
+ if (status & 2) {
+ r = 0;
+ break;
+ }
+ } else {
+ r = 0;
+ if (status & 2)
+ break;
+
+ dev_err(adev->dev,
+ "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i);
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+ UVD_VCPU_CNTL__BLK_RST_MASK,
+ ~UVD_VCPU_CNTL__BLK_RST_MASK);
+ mdelay(10);
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+ ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+ mdelay(10);
+ r = -1;
+ }
+ }
+
+ if (r) {
+ dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i);
+ return r;
+ }
+
+ /* enable master interrupt */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN),
+ UVD_MASTINT_EN__VCPU_EN_MASK,
+ ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+ /* clear the busy bit of VCN_STATUS */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0,
+ ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+ ring = &adev->vcn.inst[i].ring_enc[0];
+
+ 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);
+
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, ring->gpu_addr);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, ring->ring_size / 4);
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+ tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+ WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+ fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0);
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR);
+ WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, tmp);
+ ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+
+ tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+ tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+ fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+ }
+
+ return 0;
+}
+
+/**
+ * vcn_v5_0_1_stop_dpg_mode - VCN stop with dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ *
+ * Stop VCN block with dpg mode
+ */
+static void vcn_v5_0_1_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+{
+ uint32_t tmp;
+ int vcn_inst;
+
+ vcn_inst = GET_INST(VCN, inst_idx);
+
+ /* Wait for power status to be 1 */
+ SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1,
+ UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+ /* wait for read ptr to be equal to write ptr */
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+ SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_RB_RPTR, tmp, 0xFFFFFFFF);
+
+ /* disable dynamic power gating mode */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 0,
+ ~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+}
+
+/**
+ * vcn_v5_0_1_stop - VCN stop
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Stop VCN block
+ */
+static int vcn_v5_0_1_stop(struct amdgpu_device *adev)
+{
+ volatile struct amdgpu_vcn4_fw_shared *fw_shared;
+ uint32_t tmp;
+ int i, r = 0, vcn_inst;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ vcn_inst = GET_INST(VCN, i);
+
+ fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+ fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF;
+
+ if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+ vcn_v5_0_1_stop_dpg_mode(adev, i);
+ continue;
+ }
+
+ /* wait for vcn idle */
+ r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+ if (r)
+ return r;
+
+ tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+ UVD_LMI_STATUS__READ_CLEAN_MASK |
+ UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+ UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+ r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp);
+ if (r)
+ return r;
+
+ /* disable LMI UMC channel */
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2);
+ tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp);
+ tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
+ UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+ r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp);
+ if (r)
+ return r;
+
+ /* block VCPU register access */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL),
+ UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+ ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+ /* reset VCPU */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+ UVD_VCPU_CNTL__BLK_RST_MASK,
+ ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+ /* disable VCPU clock */
+ WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+ ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+ /* apply soft reset */
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+ tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+ tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+ tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+ /* clear status */
+ WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0);
+ }
+
+ if (adev->pm.dpm_enabled)
+ amdgpu_dpm_enable_uvd(adev, false);
+
+ return 0;
+}
+
+/**
+ * vcn_v5_0_1_unified_ring_get_rptr - get unified read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified read pointer
+ */
+static uint64_t vcn_v5_0_1_unified_ring_get_rptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+ DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+ return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_RPTR);
+}
+
+/**
+ * vcn_v5_0_1_unified_ring_get_wptr - get unified write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified write pointer
+ */
+static uint64_t vcn_v5_0_1_unified_ring_get_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+ DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+ if (ring->use_doorbell)
+ return *ring->wptr_cpu_addr;
+ else
+ return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_WPTR);
+}
+
+/**
+ * vcn_v5_0_1_unified_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+ DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+ if (ring->use_doorbell) {
+ *ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+ WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+ } else {
+ WREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_WPTR,
+ lower_32_bits(ring->wptr));
+ }
+}
+
+static const struct amdgpu_ring_funcs vcn_v5_0_1_unified_ring_vm_funcs = {
+ .type = AMDGPU_RING_TYPE_VCN_ENC,
+ .align_mask = 0x3f,
+ .nop = VCN_ENC_CMD_NO_OP,
+ .get_rptr = vcn_v5_0_1_unified_ring_get_rptr,
+ .get_wptr = vcn_v5_0_1_unified_ring_get_wptr,
+ .set_wptr = vcn_v5_0_1_unified_ring_set_wptr,
+ .emit_frame_size =
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+ 4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+ 5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+ 1, /* vcn_v2_0_enc_ring_insert_end */
+ .emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+ .emit_ib = vcn_v2_0_enc_ring_emit_ib,
+ .emit_fence = vcn_v2_0_enc_ring_emit_fence,
+ .emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+ .test_ring = amdgpu_vcn_enc_ring_test_ring,
+ .test_ib = amdgpu_vcn_unified_ring_test_ib,
+ .insert_nop = amdgpu_ring_insert_nop,
+ .insert_end = vcn_v2_0_enc_ring_insert_end,
+ .pad_ib = amdgpu_ring_generic_pad_ib,
+ .begin_use = amdgpu_vcn_ring_begin_use,
+ .end_use = amdgpu_vcn_ring_end_use,
+ .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,
+};
+
+/**
+ * vcn_v5_0_1_set_unified_ring_funcs - set unified ring functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set unified ring functions
+ */
+static void vcn_v5_0_1_set_unified_ring_funcs(struct amdgpu_device *adev)
+{
+ int i, vcn_inst;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v5_0_1_unified_ring_vm_funcs;
+ adev->vcn.inst[i].ring_enc[0].me = i;
+ vcn_inst = GET_INST(VCN, i);
+ adev->vcn.inst[i].aid_id = vcn_inst / adev->vcn.num_inst_per_aid;
+ }
+}
+
+/**
+ * vcn_v5_0_1_is_idle - check VCN block is idle
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * Check whether VCN block is idle
+ */
+static bool vcn_v5_0_1_is_idle(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ int i, ret = 1;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+ ret &= (RREG32_SOC15(VCN, GET_INST(VCN, i), regUVD_STATUS) == UVD_STATUS__IDLE);
+
+ return ret;
+}
+
+/**
+ * vcn_v5_0_1_wait_for_idle - wait for VCN block idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Wait for VCN block idle
+ */
+static int vcn_v5_0_1_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int i, ret = 0;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ ret = SOC15_WAIT_ON_RREG(VCN, GET_INST(VCN, i), regUVD_STATUS, UVD_STATUS__IDLE,
+ UVD_STATUS__IDLE);
+ if (ret)
+ return ret;
+ }
+
+ return ret;
+}
+
+/**
+ * vcn_v5_0_1_set_clockgating_state - set VCN block clockgating state
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ * @state: clock gating state
+ *
+ * Set VCN block clockgating state
+ */
+static int vcn_v5_0_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_clockgating_state state)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ bool enable = state == AMD_CG_STATE_GATE;
+ int i;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ if (enable) {
+ if (RREG32_SOC15(VCN, GET_INST(VCN, i), regUVD_STATUS) != UVD_STATUS__IDLE)
+ return -EBUSY;
+ vcn_v5_0_1_enable_clock_gating(adev, i);
+ } else {
+ vcn_v5_0_1_disable_clock_gating(adev, i);
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * vcn_v5_0_1_set_powergating_state - set VCN block powergating state
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ * @state: power gating state
+ *
+ * Set VCN block powergating state
+ */
+static int vcn_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
+ enum amd_powergating_state state)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ int ret;
+
+ if (state == adev->vcn.cur_state)
+ return 0;
+
+ if (state == AMD_PG_STATE_GATE)
+ ret = vcn_v5_0_1_stop(adev);
+ else
+ ret = vcn_v5_0_1_start(adev);
+
+ if (!ret)
+ adev->vcn.cur_state = state;
+
+ return ret;
+}
+
+/**
+ * vcn_v5_0_1_process_interrupt - process VCN block interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @entry: interrupt entry from clients and sources
+ *
+ * Process VCN block interrupt
+ */
+static int vcn_v5_0_1_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source,
+ struct amdgpu_iv_entry *entry)
+{
+ uint32_t i, inst;
+
+ i = node_id_to_phys_map[entry->node_id];
+
+ DRM_DEV_DEBUG(adev->dev, "IH: VCN TRAP\n");
+
+ for (inst = 0; inst < adev->vcn.num_vcn_inst; ++inst)
+ if (adev->vcn.inst[inst].aid_id == i)
+ break;
+ if (inst >= adev->vcn.num_vcn_inst) {
+ dev_WARN_ONCE(adev->dev, 1,
+ "Interrupt received for unknown VCN instance %d",
+ entry->node_id);
+ return 0;
+ }
+
+ switch (entry->src_id) {
+ case VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+ amdgpu_fence_process(&adev->vcn.inst[inst].ring_enc[0]);
+ break;
+ default:
+ DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+ entry->src_id, entry->src_data[0]);
+ break;
+ }
+
+ return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v5_0_1_irq_funcs = {
+ .process = vcn_v5_0_1_process_interrupt,
+};
+
+/**
+ * vcn_v5_0_1_set_irq_funcs - set VCN block interrupt irq functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set VCN block interrupt irq functions
+ */
+static void vcn_v5_0_1_set_irq_funcs(struct amdgpu_device *adev)
+{
+ int i;
+
+ for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+ adev->vcn.inst->irq.num_types++;
+ adev->vcn.inst->irq.funcs = &vcn_v5_0_1_irq_funcs;
+}
+
+static const struct amd_ip_funcs vcn_v5_0_1_ip_funcs = {
+ .name = "vcn_v5_0_1",
+ .early_init = vcn_v5_0_1_early_init,
+ .late_init = NULL,
+ .sw_init = vcn_v5_0_1_sw_init,
+ .sw_fini = vcn_v5_0_1_sw_fini,
+ .hw_init = vcn_v5_0_1_hw_init,
+ .hw_fini = vcn_v5_0_1_hw_fini,
+ .suspend = vcn_v5_0_1_suspend,
+ .resume = vcn_v5_0_1_resume,
+ .is_idle = vcn_v5_0_1_is_idle,
+ .wait_for_idle = vcn_v5_0_1_wait_for_idle,
+ .check_soft_reset = NULL,
+ .pre_soft_reset = NULL,
+ .soft_reset = NULL,
+ .post_soft_reset = NULL,
+ .set_clockgating_state = vcn_v5_0_1_set_clockgating_state,
+ .set_powergating_state = vcn_v5_0_1_set_powergating_state,
+ .dump_ip_state = vcn_v5_0_0_dump_ip_state,
+ .print_ip_state = vcn_v5_0_0_print_ip_state,
+};
+
+const struct amdgpu_ip_block_version vcn_v5_0_1_ip_block = {
+ .type = AMD_IP_BLOCK_TYPE_VCN,
+ .major = 5,
+ .minor = 0,
+ .rev = 1,
+ .funcs = &vcn_v5_0_1_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h
new file mode 100644
index 000000000000..82ac709f44bf
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * 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 __VCN_v5_0_1_H__
+#define __VCN_v5_0_1_H__
+
+extern const struct amdgpu_ip_block_version vcn_v5_0_1_ip_block;
+
+#endif /* __VCN_v5_0_1_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
index 0fedadd0a6a4..98fc6941159e 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
@@ -364,9 +364,8 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev,
* this should allow us to catchup.
*/
tmp = (wptr + 32) & ih->ptr_mask;
- dev_warn(adev->dev, "IH ring buffer overflow "
- "(0x%08X, 0x%08X, 0x%08X)\n",
- wptr, ih->rptr, tmp);
+ dev_warn_ratelimited(adev->dev, "%s ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+ amdgpu_ih_ring_name(adev, ih), wptr, ih->rptr, tmp);
ih->rptr = tmp;
tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
@@ -605,10 +604,10 @@ static void vega10_ih_update_clockgating_state(struct amdgpu_device *adev,
}
}
-static int vega10_ih_set_clockgating_state(void *handle,
+static int vega10_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
vega10_ih_update_clockgating_state(adev,
state == AMD_CG_STATE_GATE);
@@ -616,7 +615,7 @@ static int vega10_ih_set_clockgating_state(void *handle,
}
-static int vega10_ih_set_powergating_state(void *handle,
+static int vega10_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
index 1c9aff742e43..e9e3b2ed4b7b 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
@@ -366,6 +366,7 @@ static int vega20_ih_irq_init(struct amdgpu_device *adev)
/* Enable IH Retry CAM */
if (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 0) ||
amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+ amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 4) ||
amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 5))
WREG32_FIELD15(OSSSYS, 0, IH_RETRY_INT_CAM_CNTL_ALDEBARAN,
ENABLE, 1);
@@ -443,9 +444,8 @@ static u32 vega20_ih_get_wptr(struct amdgpu_device *adev,
* this should allow us to catchup.
*/
tmp = (wptr + 32) & ih->ptr_mask;
- dev_warn(adev->dev, "IH ring buffer overflow "
- "(0x%08X, 0x%08X, 0x%08X)\n",
- wptr, ih->rptr, tmp);
+ dev_warn_ratelimited(adev->dev, "%s ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+ amdgpu_ih_ring_name(adev, ih), wptr, ih->rptr, tmp);
ih->rptr = tmp;
tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
@@ -697,10 +697,10 @@ static void vega20_ih_update_clockgating_state(struct amdgpu_device *adev,
}
}
-static int vega20_ih_set_clockgating_state(void *handle,
+static int vega20_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
vega20_ih_update_clockgating_state(adev,
state == AMD_CG_STATE_GATE);
@@ -708,7 +708,7 @@ static int vega20_ih_set_clockgating_state(void *handle,
}
-static int vega20_ih_set_powergating_state(void *handle,
+static int vega20_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c
index a83505815d39..0c9c4d8b7b71 100644
--- a/drivers/gpu/drm/amd/amdgpu/vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/vi.c
@@ -167,16 +167,16 @@ static const struct amdgpu_video_codec_info tonga_video_codecs_decode_array[] =
{
{
.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
+ .max_width = 1920,
+ .max_height = 1088,
+ .max_pixels_per_frame = 1920 * 1088,
.max_level = 3,
},
{
.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
+ .max_width = 1920,
+ .max_height = 1088,
+ .max_pixels_per_frame = 1920 * 1088,
.max_level = 5,
},
{
@@ -188,9 +188,9 @@ static const struct amdgpu_video_codec_info tonga_video_codecs_decode_array[] =
},
{
.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
+ .max_width = 1920,
+ .max_height = 1088,
+ .max_pixels_per_frame = 1920 * 1088,
.max_level = 4,
},
};
@@ -206,16 +206,16 @@ static const struct amdgpu_video_codec_info cz_video_codecs_decode_array[] =
{
{
.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
+ .max_width = 1920,
+ .max_height = 1088,
+ .max_pixels_per_frame = 1920 * 1088,
.max_level = 3,
},
{
.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
+ .max_width = 1920,
+ .max_height = 1088,
+ .max_pixels_per_frame = 1920 * 1088,
.max_level = 5,
},
{
@@ -227,9 +227,9 @@ static const struct amdgpu_video_codec_info cz_video_codecs_decode_array[] =
},
{
.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
+ .max_width = 1920,
+ .max_height = 1088,
+ .max_pixels_per_frame = 1920 * 1088,
.max_level = 4,
},
{
@@ -239,13 +239,6 @@ static const struct amdgpu_video_codec_info cz_video_codecs_decode_array[] =
.max_pixels_per_frame = 4096 * 4096,
.max_level = 186,
},
- {
- .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG,
- .max_width = 4096,
- .max_height = 4096,
- .max_pixels_per_frame = 4096 * 4096,
- .max_level = 0,
- },
};
static const struct amdgpu_video_codecs cz_video_codecs_decode =
@@ -1945,10 +1938,10 @@ static int vi_common_set_clockgating_state_by_smu(void *handle,
return 0;
}
-static int vi_common_set_clockgating_state(void *handle,
+static int vi_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -1988,7 +1981,7 @@ static int vi_common_set_clockgating_state(void *handle,
return 0;
}
-static int vi_common_set_powergating_state(void *handle,
+static int vi_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
return 0;