From 47fc644f801e4414753a9b7e87ed41f991cd68c3 Mon Sep 17 00:00:00 2001
From: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Date: Wed, 19 Apr 2023 16:12:45 +0530
Subject: drm/amd/amdgpu: Fix style errors in amdgpu_drv.c & amdgpu_device.c
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Fix following checkpatch style errors in amdgpu_drv.c &
amdgpu_device.c
ERROR: exactly one space required after that #ifdef
ERROR: spaces required around that '+=' (ctx:WxV)
ERROR: space required before the open brace '{'
ERROR: spaces required around that '||' (ctx:VxE)
ERROR: space prohibited before that close parenthesis ')'
ERROR: space required before the open parenthesis '('
ERROR: space required before the open brace '{'
ERROR: code indent should use tabs where possible
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index b1ca1ab6d6ad..b400d598b75a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -1660,7 +1660,7 @@ static const u16 amdgpu_unsupported_pciidlist[] = {
};
static const struct pci_device_id pciidlist[] = {
-#ifdef CONFIG_DRM_AMDGPU_SI
+#ifdef CONFIG_DRM_AMDGPU_SI
{0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
{0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
{0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
--
cgit v1.2.3
From 8a93c691248e7ff2b3944107a1ead2671b6854f2 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 10 Mar 2023 13:17:11 -0500
Subject: drm/amdgpu: bump driver version number for CP GFX shadow
So UMDs can determine whether the kernel supports this.
Mesa MR: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21986
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index b400d598b75a..b987022e11b4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -110,9 +110,10 @@
* 3.52.0 - Add AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD, add device_info fields:
* tcp_cache_size, num_sqc_per_wgp, sqc_data_cache_size, sqc_inst_cache_size,
* gl1c_cache_size, gl2c_cache_size, mall_size, enabled_rb_pipes_mask_hi
+ * 3.53.0 - Support for GFX11 CP GFX shadowing
*/
#define KMS_DRIVER_MAJOR 3
-#define KMS_DRIVER_MINOR 52
+#define KMS_DRIVER_MINOR 53
#define KMS_DRIVER_PATCHLEVEL 0
unsigned int amdgpu_vram_limit = UINT_MAX;
--
cgit v1.2.3
From 489763af891d5dc35c0b64e18af284d6591286cf Mon Sep 17 00:00:00 2001
From: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Date: Mon, 24 Apr 2023 19:25:32 +0200
Subject: drm/amdgpu: add new flag to AMDGPU_CTX_QUERY2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
OpenGL EXT_robustness extension expects the driver to stop reporting
GUILTY_CONTEXT_RESET when the reset has completed and the GPU is ready
to accept submission again.
This commit adds a AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS flag,
that let the UMD know that the reset is still not finished.
Mesa MR: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22290
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
include/uapi/drm/amdgpu_drm.h | 2 ++
3 files changed, 7 insertions(+), 1 deletion(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index d2139ac12159..e1f642a3dc2f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -576,6 +576,9 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
if (atomic_read(&ctx->guilty))
out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY;
+ if (amdgpu_in_reset(adev))
+ out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS;
+
if (adev->ras_enabled && con) {
/* Return the cached values in O(1),
* and schedule delayed work to cache
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index b987022e11b4..45544ebe576e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -111,9 +111,10 @@
* tcp_cache_size, num_sqc_per_wgp, sqc_data_cache_size, sqc_inst_cache_size,
* gl1c_cache_size, gl2c_cache_size, mall_size, enabled_rb_pipes_mask_hi
* 3.53.0 - Support for GFX11 CP GFX shadowing
+ * 3.54.0 - Add AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS support
*/
#define KMS_DRIVER_MAJOR 3
-#define KMS_DRIVER_MINOR 53
+#define KMS_DRIVER_MINOR 54
#define KMS_DRIVER_PATCHLEVEL 0
unsigned int amdgpu_vram_limit = UINT_MAX;
diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h
index cc78528c3b4b..79b14828d542 100644
--- a/include/uapi/drm/amdgpu_drm.h
+++ b/include/uapi/drm/amdgpu_drm.h
@@ -245,6 +245,8 @@ union drm_amdgpu_bo_list {
/* indicate some errors are detected by RAS */
#define AMDGPU_CTX_QUERY2_FLAGS_RAS_CE (1<<3)
#define AMDGPU_CTX_QUERY2_FLAGS_RAS_UE (1<<4)
+/* indicate that the reset hasn't completed yet */
+#define AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS (1<<5)
/* Context priority level */
#define AMDGPU_CTX_PRIORITY_UNSET -2048
--
cgit v1.2.3
From 0fa49d108386c201b5c2cce68066a9b8f66883a5 Mon Sep 17 00:00:00 2001
From: Shiwu Zhang <shiwu.zhang@amd.com>
Date: Fri, 17 Dec 2021 11:27:53 +0800
Subject: drm/amdgpu: override partition mode through module parameter
Add a module parameter to override the partition mode.
Signed-off-by: Shiwu Zhang <shiwu.zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 13 +++++++++++++
drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c | 25 +++++++++++++++++++++++--
3 files changed, 38 insertions(+), 2 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 0f163d266812..a277bdc86057 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -242,6 +242,8 @@ extern int amdgpu_num_kcq;
extern int amdgpu_vcnfw_log;
extern int amdgpu_sg_display;
+extern uint amdgpu_user_partt_mode;
+
#define AMDGPU_VM_MAX_NUM_CTX 4096
#define AMDGPU_SG_THRESHOLD (256*1024*1024)
#define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS 3000
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 45544ebe576e..9e9da2ac5c82 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -193,6 +193,7 @@ int amdgpu_smartshift_bias;
int amdgpu_use_xgmi_p2p = 1;
int amdgpu_vcnfw_log;
int amdgpu_sg_display = -1; /* auto */
+uint amdgpu_user_partt_mode;
static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work);
@@ -950,6 +951,18 @@ MODULE_PARM_DESC(smu_pptable_id,
"specify pptable id to be used (-1 = auto(default) value, 0 = use pptable from vbios, > 0 = soft pptable id)");
module_param_named(smu_pptable_id, amdgpu_smu_pptable_id, int, 0444);
+/**
+ * DOC: partition_mode (int)
+ * Used to override the default SPX mode.
+ */
+MODULE_PARM_DESC(user_partt_mode,
+ "specify partition mode to be used (0 = AMDGPU_SPX_PARTITION_MODE(default value), \
+ 1 = AMDGPU_DPX_PARTITION_MODE, \
+ 2 = AMDGPU_TPX_PARTITION_MODE, \
+ 3 = AMDGPU_QPX_PARTITION_MODE, \
+ 4 = AMDGPU_CPX_PARTITION_MODE)");
+module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
+
/* These devices are not supported by amdgpu.
* They are supported by the mach64, r128, radeon drivers
*/
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 1d15db9423c9..2676a185c232 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -2162,8 +2162,29 @@ static int gfx_v9_4_3_early_init(void *handle)
/* hardcode in emulation phase */
adev->gfx.num_xcd = 1;
- adev->gfx.num_xcc_per_xcp = 1;
- adev->gfx.partition_mode = AMDGPU_SPX_PARTITION_MODE;
+
+ adev->gfx.partition_mode = amdgpu_user_partt_mode;
+ /* calculate the num_xcc_in_xcp for the partition mode*/
+ switch (amdgpu_user_partt_mode) {
+ case AMDGPU_SPX_PARTITION_MODE:
+ adev->gfx.num_xcc_per_xcp = adev->gfx.num_xcd;
+ break;
+ case AMDGPU_DPX_PARTITION_MODE:
+ adev->gfx.num_xcc_per_xcp = adev->gfx.num_xcd / 2;
+ break;
+ case AMDGPU_TPX_PARTITION_MODE:
+ adev->gfx.num_xcc_per_xcp = adev->gfx.num_xcd / 3;
+ break;
+ case AMDGPU_QPX_PARTITION_MODE:
+ adev->gfx.num_xcc_per_xcp = adev->gfx.num_xcd / 4;
+ break;
+ case AMDGPU_CPX_PARTITION_MODE:
+ adev->gfx.num_xcc_per_xcp = 1;
+ break;
+ default:
+ adev->gfx.num_xcc_per_xcp = adev->gfx.num_xcd;
+ break;
+ }
adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
AMDGPU_MAX_COMPUTE_RINGS);
--
cgit v1.2.3
From 4d5275ab0b18d17697392aafd93e206e6b9de647 Mon Sep 17 00:00:00 2001
From: Lijo Lazar <lijo.lazar@amd.com>
Date: Fri, 27 Jan 2023 18:18:17 +0530
Subject: drm/amdgpu: Add parsing of acpi xcc objects
Add parsing of ACPI xcc objects and fill in relevant info from them by
invoking the DSM methods.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-and-tested-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +
drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c | 294 +++++++++++++++++++++++++++++++
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 1 +
3 files changed, 297 insertions(+)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 39743d44b567..880bf9d67284 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1404,11 +1404,13 @@ int amdgpu_acpi_pcie_notify_device_ready(struct amdgpu_device *adev);
void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_caps *caps);
bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev);
void amdgpu_acpi_detect(void);
+void amdgpu_acpi_release(void);
#else
static inline int amdgpu_acpi_init(struct amdgpu_device *adev) { return 0; }
static inline void amdgpu_acpi_fini(struct amdgpu_device *adev) { }
static inline bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev) { return false; }
static inline void amdgpu_acpi_detect(void) { }
+static inline void amdgpu_acpi_release(void) { }
static inline bool amdgpu_acpi_is_power_shift_control_supported(void) { return false; }
static inline int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev,
u8 dev_state, bool drv_state) { return 0; }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index aeeec211861c..a3a2ef43abfc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -38,6 +38,43 @@
#include "amd_acpi.h"
#include "atom.h"
+/* Declare GUID for AMD _DSM method for XCCs */
+static const guid_t amd_xcc_dsm_guid = GUID_INIT(0x8267f5d5, 0xa556, 0x44f2,
+ 0xb8, 0xb4, 0x45, 0x56, 0x2e,
+ 0x8c, 0x5b, 0xec);
+
+#define AMD_XCC_HID_START 3000
+#define AMD_XCC_DSM_GET_NUM_FUNCS 0
+#define AMD_XCC_DSM_GET_SUPP_MODE 1
+#define AMD_XCC_DSM_GET_XCP_MODE 2
+#define AMD_XCC_DSM_GET_VF_XCC_MAPPING 4
+#define AMD_XCC_DSM_GET_TMR_INFO 5
+#define AMD_XCC_DSM_NUM_FUNCS 5
+
+#define AMD_XCC_MAX_HID 24
+
+/* Encapsulates the XCD acpi object information */
+struct amdgpu_acpi_xcc_info {
+ struct list_head list;
+ int mem_node;
+ uint8_t xcp_node;
+ uint8_t phy_id;
+ acpi_handle handle;
+};
+
+struct amdgpu_acpi_dev_info {
+ struct list_head list;
+ struct list_head xcc_list;
+ uint16_t bdf;
+ uint16_t supp_xcp_mode;
+ uint16_t xcp_mode;
+ uint16_t mem_mode;
+ uint64_t tmr_base;
+ uint64_t tmr_size;
+};
+
+struct list_head amdgpu_acpi_dev_list;
+
struct amdgpu_atif_notification_cfg {
bool enabled;
int command_code;
@@ -801,6 +838,240 @@ int amdgpu_acpi_smart_shift_update(struct drm_device *dev, enum amdgpu_ss ss_sta
return r;
}
+/**
+ * amdgpu_acpi_get_node_id - obtain the NUMA node id for corresponding amdgpu
+ * acpi device handle
+ *
+ * @handle: acpi handle
+ * @nid: NUMA Node id returned by the platform firmware
+ *
+ * Queries the ACPI interface to fetch the corresponding NUMA Node ID for a
+ * given amdgpu acpi device.
+ *
+ * Returns ACPI STATUS OK with Node ID on success or the corresponding failure reason
+ */
+acpi_status amdgpu_acpi_get_node_id(acpi_handle handle, int *nid)
+{
+#ifdef CONFIG_ACPI_NUMA
+ u64 pxm;
+ acpi_status status;
+
+ status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
+
+ if (ACPI_FAILURE(status))
+ return status;
+
+ *nid = pxm_to_node(pxm);
+
+ return_ACPI_STATUS(AE_OK);
+#else
+ return_ACPI_STATUS(AE_NOT_EXIST);
+#endif
+}
+
+struct amdgpu_acpi_dev_info *amdgpu_acpi_get_dev(u16 bdf)
+{
+ struct amdgpu_acpi_dev_info *acpi_dev;
+
+ if (list_empty(&amdgpu_acpi_dev_list))
+ return NULL;
+
+ list_for_each_entry(acpi_dev, &amdgpu_acpi_dev_list, list)
+ if (acpi_dev->bdf == bdf)
+ return acpi_dev;
+
+ return NULL;
+}
+
+static int amdgpu_acpi_dev_init(struct amdgpu_acpi_dev_info **dev_info,
+ struct amdgpu_acpi_xcc_info *xcc_info, u16 bdf)
+{
+ struct amdgpu_acpi_dev_info *tmp;
+ union acpi_object *obj;
+ int ret = -ENOENT;
+
+ *dev_info = NULL;
+ tmp = kzalloc(sizeof(struct amdgpu_acpi_dev_info), GFP_KERNEL);
+ if (!tmp)
+ return -ENOMEM;
+
+ INIT_LIST_HEAD(&tmp->xcc_list);
+ INIT_LIST_HEAD(&tmp->list);
+ tmp->bdf = bdf;
+
+ obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+ AMD_XCC_DSM_GET_SUPP_MODE, NULL,
+ ACPI_TYPE_INTEGER);
+
+ if (!obj) {
+ acpi_handle_debug(xcc_info->handle,
+ "_DSM function %d evaluation failed",
+ AMD_XCC_DSM_GET_SUPP_MODE);
+ ret = -ENOENT;
+ goto out;
+ }
+
+ tmp->supp_xcp_mode = obj->integer.value & 0xFFFF;
+ ACPI_FREE(obj);
+
+ obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+ AMD_XCC_DSM_GET_XCP_MODE, NULL,
+ ACPI_TYPE_INTEGER);
+
+ if (!obj) {
+ acpi_handle_debug(xcc_info->handle,
+ "_DSM function %d evaluation failed",
+ AMD_XCC_DSM_GET_XCP_MODE);
+ ret = -ENOENT;
+ goto out;
+ }
+
+ tmp->xcp_mode = obj->integer.value & 0xFFFF;
+ tmp->mem_mode = (obj->integer.value >> 32) & 0xFFFF;
+ ACPI_FREE(obj);
+
+ /* Evaluate DSMs and fill XCC information */
+ obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+ AMD_XCC_DSM_GET_TMR_INFO, NULL,
+ ACPI_TYPE_PACKAGE);
+
+ if (!obj || obj->package.count < 2) {
+ acpi_handle_debug(xcc_info->handle,
+ "_DSM function %d evaluation failed",
+ AMD_XCC_DSM_GET_TMR_INFO);
+ ret = -ENOENT;
+ goto out;
+ }
+
+ tmp->tmr_base = obj->package.elements[0].integer.value;
+ tmp->tmr_size = obj->package.elements[1].integer.value;
+ ACPI_FREE(obj);
+
+ DRM_DEBUG_DRIVER(
+ "New dev(%x): Supported xcp mode: %x curr xcp_mode : %x mem mode : %x, tmr base: %llx tmr size: %llx ",
+ tmp->bdf, tmp->supp_xcp_mode, tmp->xcp_mode, tmp->mem_mode,
+ tmp->tmr_base, tmp->tmr_size);
+ list_add_tail(&tmp->list, &amdgpu_acpi_dev_list);
+ *dev_info = tmp;
+
+ return 0;
+
+out:
+ if (obj)
+ ACPI_FREE(obj);
+ kfree(tmp);
+
+ return ret;
+}
+
+static int amdgpu_acpi_get_xcc_info(struct amdgpu_acpi_xcc_info *xcc_info,
+ u16 *bdf)
+{
+ union acpi_object *obj;
+ acpi_status status;
+ int ret = -ENOENT;
+
+ obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+ AMD_XCC_DSM_GET_NUM_FUNCS, NULL,
+ ACPI_TYPE_INTEGER);
+
+ if (!obj || obj->integer.value != AMD_XCC_DSM_NUM_FUNCS)
+ goto out;
+ ACPI_FREE(obj);
+
+ /* Evaluate DSMs and fill XCC information */
+ obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+ AMD_XCC_DSM_GET_VF_XCC_MAPPING, NULL,
+ ACPI_TYPE_INTEGER);
+
+ if (!obj) {
+ acpi_handle_debug(xcc_info->handle,
+ "_DSM function %d evaluation failed",
+ AMD_XCC_DSM_GET_VF_XCC_MAPPING);
+ ret = -EINVAL;
+ goto out;
+ }
+
+ /* PF xcc id [39:32] */
+ xcc_info->phy_id = (obj->integer.value >> 32) & 0xFF;
+ /* xcp node of this xcc [47:40] */
+ xcc_info->xcp_node = (obj->integer.value >> 40) & 0xFF;
+ /* PF bus/dev/fn of this xcc [63:48] */
+ *bdf = (obj->integer.value >> 48) & 0xFFFF;
+ ACPI_FREE(obj);
+ obj = NULL;
+
+ status = amdgpu_acpi_get_node_id(xcc_info->handle, &xcc_info->mem_node);
+
+ /* TODO: check if this check is required */
+ if (ACPI_SUCCESS(status))
+ ret = 0;
+out:
+ if (obj)
+ ACPI_FREE(obj);
+
+ return ret;
+}
+
+static int amdgpu_acpi_enumerate_xcc(void)
+{
+ struct amdgpu_acpi_dev_info *dev_info = NULL;
+ struct amdgpu_acpi_xcc_info *xcc_info;
+ struct acpi_device *acpi_dev;
+ char hid[ACPI_ID_LEN];
+ int ret, id;
+ u16 bdf;
+
+ INIT_LIST_HEAD(&amdgpu_acpi_dev_list);
+
+ for (id = 0; id < AMD_XCC_MAX_HID; id++) {
+ sprintf(hid, "%s%d", "AMD", AMD_XCC_HID_START + id);
+ acpi_dev = acpi_dev_get_first_match_dev(hid, NULL, -1);
+ /* These ACPI objects are expected to be in sequential order. If
+ * one is not found, no need to check the rest.
+ */
+ if (!acpi_dev) {
+ DRM_DEBUG_DRIVER("No matching acpi device found for %s",
+ hid);
+ break;
+ }
+
+ xcc_info = kzalloc(sizeof(struct amdgpu_acpi_xcc_info),
+ GFP_KERNEL);
+ if (!xcc_info) {
+ DRM_ERROR("Failed to allocate memory for xcc info\n");
+ return -ENOMEM;
+ }
+
+ INIT_LIST_HEAD(&xcc_info->list);
+ xcc_info->handle = acpi_device_handle(acpi_dev);
+ acpi_dev_put(acpi_dev);
+
+ ret = amdgpu_acpi_get_xcc_info(xcc_info, &bdf);
+ if (ret) {
+ kfree(xcc_info);
+ continue;
+ }
+
+ dev_info = amdgpu_acpi_get_dev(bdf);
+
+ if (!dev_info)
+ ret = amdgpu_acpi_dev_init(&dev_info, xcc_info, bdf);
+
+ if (ret == -ENOMEM)
+ return ret;
+
+ if (!dev_info) {
+ kfree(xcc_info);
+ continue;
+ }
+
+ list_add_tail(&xcc_info->list, &dev_info->xcc_list);
+ }
+
+ return 0;
+}
+
/**
* amdgpu_acpi_event - handle notify events
*
@@ -1054,6 +1325,29 @@ void amdgpu_acpi_detect(void)
} else {
atif->backlight_caps.caps_valid = false;
}
+
+ amdgpu_acpi_enumerate_xcc();
+}
+
+void amdgpu_acpi_release(void)
+{
+ struct amdgpu_acpi_dev_info *dev_info, *dev_tmp;
+ struct amdgpu_acpi_xcc_info *xcc_info, *xcc_tmp;
+
+ if (list_empty(&amdgpu_acpi_dev_list))
+ return;
+
+ list_for_each_entry_safe(dev_info, dev_tmp, &amdgpu_acpi_dev_list,
+ list) {
+ list_for_each_entry_safe(xcc_info, xcc_tmp, &dev_info->xcc_list,
+ list) {
+ list_del(&xcc_info->list);
+ kfree(xcc_info);
+ }
+
+ list_del(&dev_info->list);
+ kfree(dev_info);
+ }
}
#if IS_ENABLED(CONFIG_SUSPEND)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 9e9da2ac5c82..f319318a8813 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2898,6 +2898,7 @@ static void __exit amdgpu_exit(void)
amdgpu_amdkfd_fini();
pci_unregister_driver(&amdgpu_kms_pci_driver);
amdgpu_unregister_atpx_handler();
+ amdgpu_acpi_release();
amdgpu_sync_fini();
amdgpu_fence_slab_fini();
mmu_notifier_synchronize();
--
cgit v1.2.3
From 570de94b9c5d93e1c5bc4e357946efb93c662da9 Mon Sep 17 00:00:00 2001
From: Lijo Lazar <lijo.lazar@amd.com>
Date: Mon, 13 Feb 2023 19:26:18 +0530
Subject: drm/amdgpu: Add auto mode for compute partition
When auto mode is specified, driver will choose the right compute
partition mode.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Le Ma <le.ma@amd.com>
Reviewed-by: Philip Yang <philip.yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 8 ++++---
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h | 2 ++
.../gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c | 28 +++++++++++++++++++++-
4 files changed, 35 insertions(+), 5 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index f2bafab15ceb..cb9373f8c25a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -242,7 +242,7 @@ extern int amdgpu_num_kcq;
extern int amdgpu_vcnfw_log;
extern int amdgpu_sg_display;
-extern uint amdgpu_user_partt_mode;
+extern int amdgpu_user_partt_mode;
#define AMDGPU_VM_MAX_NUM_CTX 4096
#define AMDGPU_SG_THRESHOLD (256*1024*1024)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index f319318a8813..da4e50aef95a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -193,7 +193,7 @@ int amdgpu_smartshift_bias;
int amdgpu_use_xgmi_p2p = 1;
int amdgpu_vcnfw_log;
int amdgpu_sg_display = -1; /* auto */
-uint amdgpu_user_partt_mode;
+int amdgpu_user_partt_mode = AMDGPU_AUTO_COMPUTE_PARTITION_MODE;
static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work);
@@ -955,8 +955,10 @@ module_param_named(smu_pptable_id, amdgpu_smu_pptable_id, int, 0444);
* DOC: partition_mode (int)
* Used to override the default SPX mode.
*/
-MODULE_PARM_DESC(user_partt_mode,
- "specify partition mode to be used (0 = AMDGPU_SPX_PARTITION_MODE(default value), \
+MODULE_PARM_DESC(
+ user_partt_mode,
+ "specify partition mode to be used (-2 = AMDGPU_AUTO_COMPUTE_PARTITION_MODE(default value) \
+ 0 = AMDGPU_SPX_PARTITION_MODE, \
1 = AMDGPU_DPX_PARTITION_MODE, \
2 = AMDGPU_TPX_PARTITION_MODE, \
3 = AMDGPU_QPX_PARTITION_MODE, \
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index 728977f8afe7..e9c93f6e12b8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -62,6 +62,8 @@ enum amdgpu_gfx_partition {
AMDGPU_QPX_PARTITION_MODE = 3,
AMDGPU_CPX_PARTITION_MODE = 4,
AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE = -1,
+ /* Automatically choose the right mode */
+ AMDGPU_AUTO_COMPUTE_PARTITION_MODE = -2,
};
#define NUM_XCC(x) hweight16(x)
diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c
index 7469de3fd6fe..a165b51e9e58 100644
--- a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c
+++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c
@@ -235,6 +235,30 @@ int __aqua_vanjaram_get_xcp_ip_info(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id,
return 0;
}
+static enum amdgpu_gfx_partition
+__aqua_vanjaram_get_auto_mode(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+ struct amdgpu_device *adev = xcp_mgr->adev;
+ int num_xcc;
+
+ num_xcc = NUM_XCC(xcp_mgr->adev->gfx.xcc_mask);
+
+ if (adev->gmc.num_mem_partitions == 1)
+ return AMDGPU_SPX_PARTITION_MODE;
+
+ if (adev->gmc.num_mem_partitions == num_xcc)
+ return AMDGPU_CPX_PARTITION_MODE;
+
+ if (adev->gmc.num_mem_partitions == num_xcc / 2)
+ return (adev->flags & AMD_IS_APU) ? AMDGPU_TPX_PARTITION_MODE :
+ AMDGPU_QPX_PARTITION_MODE;
+
+ if (adev->gmc.num_mem_partitions == 2 && !(adev->flags & AMD_IS_APU))
+ return AMDGPU_DPX_PARTITION_MODE;
+
+ return AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE;
+}
+
static bool __aqua_vanjaram_is_valid_mode(struct amdgpu_xcp_mgr *xcp_mgr,
enum amdgpu_gfx_partition mode)
{
@@ -304,7 +328,9 @@ static int aqua_vanjaram_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr,
adev = xcp_mgr->adev;
num_xcc = NUM_XCC(adev->gfx.xcc_mask);
- if (!__aqua_vanjaram_is_valid_mode(xcp_mgr, mode))
+ if (mode == AMDGPU_AUTO_COMPUTE_PARTITION_MODE)
+ mode = __aqua_vanjaram_get_auto_mode(xcp_mgr);
+ else if (!__aqua_vanjaram_is_valid_mode(xcp_mgr, mode))
return -EINVAL;
if (adev->kfd.init_complete)
--
cgit v1.2.3
From 895797d9193b38e759bc01268a8e3887e521f682 Mon Sep 17 00:00:00 2001
From: Graham Sider <Graham.Sider@amd.com>
Date: Mon, 6 Feb 2023 14:04:42 -0500
Subject: drm/amdgpu/bu: Add use_mtype_cc_wa module param
By default, set use_mtype_cc_wa to 1 to set PTE coherence flag MTYPE_CC
instead of MTYPE_RW by default. This is required for the time being to
mitigate a bug causing XCCs to hit stale data due to TCC marking fully
dirty lines as exclusive.
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Reviewed-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 7 +++++++
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 10 +++++++---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 7 +++++--
4 files changed, 20 insertions(+), 5 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index cb9373f8c25a..cd2a29a7e26d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -212,6 +212,7 @@ extern int amdgpu_noretry;
extern int amdgpu_force_asic_type;
extern int amdgpu_smartshift_bias;
extern int amdgpu_use_xgmi_p2p;
+extern bool amdgpu_use_mtype_cc_wa;
#ifdef CONFIG_HSA_AMD
extern int sched_policy;
extern bool debug_evictions;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index da4e50aef95a..8bc37826a99f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -822,6 +822,13 @@ MODULE_PARM_DESC(no_queue_eviction_on_vm_fault, "No queue eviction on VM fault (
module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
#endif
+/**
+ * DOC: use_mtype_cc_wa (bool)
+ */
+bool amdgpu_use_mtype_cc_wa = true;
+MODULE_PARM_DESC(use_mtype_cc_wa, "Use MTYPE_CC workaround (0 = use MTYPE_RW where applicable, 1 = use MTYPE_CC where applicable (default))");
+module_param_named(use_mtype_cc_wa, amdgpu_use_mtype_cc_wa, bool, 0444);
+
/**
* DOC: pcie_p2p (bool)
* Enable PCIe P2P (requires large-BAR). Default value: true (on)
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 2eb67b53e497..8623b93c05ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1187,6 +1187,7 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
bool coherent = bo->flags & AMDGPU_GEM_CREATE_COHERENT;
bool uncached = bo->flags & AMDGPU_GEM_CREATE_UNCACHED;
unsigned int mtype;
+ unsigned int mtype_default;
bool snoop = false;
switch (adev->ip_versions[GC_HWIP][0]) {
@@ -1230,7 +1231,10 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
/* FIXME: Needs more work for handling multiple memory
* partitions (> NPS1 mode) e.g. NPS4 for both APU and dGPU
* modes.
+ * FIXME: Temporarily using MTYPE_CC instead of MTYPE_RW where applicable.
+ * To force use of MTYPE_RW, set use_mtype_cc_wa=0
*/
+ mtype_default = amdgpu_use_mtype_cc_wa ? MTYPE_CC : MTYPE_RW;
snoop = true;
if (uncached) {
mtype = MTYPE_UC;
@@ -1245,14 +1249,14 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
* socket should be treated as remote access so MTYPE_RW
* cannot be used always.
*/
- mtype = MTYPE_RW;
+ mtype = mtype_default;
} else if (adev->flags & AMD_IS_APU) {
/* APU on carve out mode */
- mtype = MTYPE_RW;
+ mtype = mtype_default;
} else {
/* dGPU */
if (is_vram && bo_adev == adev)
- mtype = MTYPE_RW;
+ mtype = mtype_default;
else if (is_vram)
mtype = MTYPE_NC;
else
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 2b2129dd1e4a..477ef9294203 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -1198,9 +1198,12 @@ svm_range_get_pte_flags(struct kfd_node *node,
if (uncached) {
mapping_flags |= AMDGPU_VM_MTYPE_UC;
} else if (domain == SVM_RANGE_VRAM_DOMAIN) {
- /* local HBM region close to partition */
+ /* local HBM region close to partition
+ * FIXME: Temporarily using MTYPE_CC instead of MTYPE_RW where applicable.
+ * To force use of MTYPE_RW, set use_mtype_cc_wa=0
+ */
if (bo_node == node)
- mapping_flags |= AMDGPU_VM_MTYPE_RW;
+ mapping_flags |= amdgpu_use_mtype_cc_wa ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
/* local HBM region far from partition or remote XGMI GPU */
else if (svm_nodes_in_same_hive(bo_node, node))
mapping_flags |= AMDGPU_VM_MTYPE_NC;
--
cgit v1.2.3
From 76eb9c95a409ea820b2e7c968c220e7a38f27d76 Mon Sep 17 00:00:00 2001
From: David Francis <David.Francis@amd.com>
Date: Mon, 27 Feb 2023 10:33:11 -0500
Subject: drm/amdgpu/bu: add mtype_local as a module parameter
Selects the MTYPE to be used for local memory,
(0 = MTYPE_CC (default), 1 = MTYPE_NC, 2 = MTYPE_RW)
v2: squash in build fix (Alex)
Reviewed-by: Graham Sider <Graham.Sider@amd.com>
Signed-off-by: David Francis <David.Francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 8 ++++----
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 19 ++++++++++++++++---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 3 +--
4 files changed, 22 insertions(+), 10 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index cd2a29a7e26d..c2feaf2fd070 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -212,7 +212,7 @@ extern int amdgpu_noretry;
extern int amdgpu_force_asic_type;
extern int amdgpu_smartshift_bias;
extern int amdgpu_use_xgmi_p2p;
-extern bool amdgpu_use_mtype_cc_wa;
+extern int amdgpu_mtype_local;
#ifdef CONFIG_HSA_AMD
extern int sched_policy;
extern bool debug_evictions;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 8bc37826a99f..706ba4af062f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -823,11 +823,11 @@ module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm
#endif
/**
- * DOC: use_mtype_cc_wa (bool)
+ * DOC: mtype_local (int)
*/
-bool amdgpu_use_mtype_cc_wa = true;
-MODULE_PARM_DESC(use_mtype_cc_wa, "Use MTYPE_CC workaround (0 = use MTYPE_RW where applicable, 1 = use MTYPE_CC where applicable (default))");
-module_param_named(use_mtype_cc_wa, amdgpu_use_mtype_cc_wa, bool, 0444);
+int amdgpu_mtype_local;
+MODULE_PARM_DESC(mtype_local, "MTYPE for local memory (0 = MTYPE_CC (default), 1 = MTYPE_NC, 2 = MTYPE_RW)");
+module_param_named(mtype_local, amdgpu_mtype_local, int, 0444);
/**
* DOC: pcie_p2p (bool)
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index c64a69f75da2..5a1414300271 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1235,7 +1235,16 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
* NUMA systems. Their MTYPE can be overridden per-page in
* gmc_v9_0_override_vm_pte_flags.
*/
- mtype_local = amdgpu_use_mtype_cc_wa ? MTYPE_CC : MTYPE_RW;
+ mtype_local = MTYPE_CC;
+ if (amdgpu_mtype_local == 1) {
+ DRM_INFO_ONCE("Using MTYPE_NC for local memory\n");
+ mtype_local = MTYPE_NC;
+ } else if (amdgpu_mtype_local == 2) {
+ DRM_INFO_ONCE("Using MTYPE_RW for local memory\n");
+ mtype_local = MTYPE_RW;
+ } else {
+ DRM_INFO_ONCE("Using MTYPE_CC for local memory\n");
+ }
is_local = (!is_vram && (adev->flags & AMD_IS_APU) &&
num_possible_nodes() <= 1) ||
(is_vram && adev == bo_adev /* TODO: memory partitions &&
@@ -1349,9 +1358,13 @@ static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev,
dev_dbg(adev->dev, "vm->mem_id=%d, local_node=%d, nid=%d\n",
/*vm->mem_id*/0, local_node, nid);
if (nid == local_node) {
- unsigned int mtype_local =
- amdgpu_use_mtype_cc_wa ? MTYPE_CC : MTYPE_RW;
uint64_t old_flags = *flags;
+ unsigned int mtype_local = MTYPE_CC;
+
+ if (amdgpu_mtype_local == 1)
+ mtype_local = MTYPE_NC;
+ else if (amdgpu_mtype_local == 2)
+ mtype_local = MTYPE_RW;
*flags = (*flags & ~AMDGPU_PTE_MTYPE_VG10_MASK) |
AMDGPU_PTE_MTYPE_VG10(mtype_local);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 4eec75b28917..df0ed5677609 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -1192,8 +1192,7 @@ svm_range_get_pte_flags(struct kfd_node *node,
}
break;
case IP_VERSION(9, 4, 3):
- mtype_local = amdgpu_use_mtype_cc_wa ? AMDGPU_VM_MTYPE_CC :
- AMDGPU_VM_MTYPE_RW;
+ mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC : (amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_RW : AMDGPU_VM_MTYPE_CC);
snoop = true;
if (uncached) {
mapping_flags |= AMDGPU_VM_MTYPE_UC;
--
cgit v1.2.3
From b9cbd51000ad3541351ca832b00600870ac08e5c Mon Sep 17 00:00:00 2001
From: Graham Sider <Graham.Sider@amd.com>
Date: Mon, 6 Mar 2023 17:56:44 -0500
Subject: drm/amdgpu/bu: update mtype_local parameter settings
Update mtype_local module parameter to use MTYPE_RW by default.
0: MTYPE_RW (default)
1: MTYPE_NC
2: MTYPE_CC
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 12 ++++++------
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 3 ++-
3 files changed, 9 insertions(+), 8 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 706ba4af062f..aa466a9eb956 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -826,7 +826,7 @@ module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm
* DOC: mtype_local (int)
*/
int amdgpu_mtype_local;
-MODULE_PARM_DESC(mtype_local, "MTYPE for local memory (0 = MTYPE_CC (default), 1 = MTYPE_NC, 2 = MTYPE_RW)");
+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);
/**
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 5a1414300271..32eb4f4f5492 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1235,15 +1235,15 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
* NUMA systems. Their MTYPE can be overridden per-page in
* gmc_v9_0_override_vm_pte_flags.
*/
- mtype_local = MTYPE_CC;
+ mtype_local = MTYPE_RW;
if (amdgpu_mtype_local == 1) {
DRM_INFO_ONCE("Using MTYPE_NC for local memory\n");
mtype_local = MTYPE_NC;
} else if (amdgpu_mtype_local == 2) {
- DRM_INFO_ONCE("Using MTYPE_RW for local memory\n");
- mtype_local = MTYPE_RW;
- } else {
DRM_INFO_ONCE("Using MTYPE_CC for local memory\n");
+ mtype_local = MTYPE_CC;
+ } else {
+ DRM_INFO_ONCE("Using MTYPE_RW for local memory\n");
}
is_local = (!is_vram && (adev->flags & AMD_IS_APU) &&
num_possible_nodes() <= 1) ||
@@ -1359,12 +1359,12 @@ static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev,
/*vm->mem_id*/0, local_node, nid);
if (nid == local_node) {
uint64_t old_flags = *flags;
- unsigned int mtype_local = MTYPE_CC;
+ unsigned int mtype_local = MTYPE_RW;
if (amdgpu_mtype_local == 1)
mtype_local = MTYPE_NC;
else if (amdgpu_mtype_local == 2)
- mtype_local = MTYPE_RW;
+ mtype_local = MTYPE_CC;
*flags = (*flags & ~AMDGPU_PTE_MTYPE_VG10_MASK) |
AMDGPU_PTE_MTYPE_VG10(mtype_local);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index df0ed5677609..e6348d4133fd 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -1192,7 +1192,8 @@ svm_range_get_pte_flags(struct kfd_node *node,
}
break;
case IP_VERSION(9, 4, 3):
- mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC : (amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_RW : AMDGPU_VM_MTYPE_CC);
+ mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC :
+ (amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW);
snoop = true;
if (uncached) {
mapping_flags |= AMDGPU_VM_MTYPE_UC;
--
cgit v1.2.3
From 2c1c7ba457d4ecf475c0e220ac5359971355c6eb Mon Sep 17 00:00:00 2001
From: James Zhu <James.Zhu@amd.com>
Date: Mon, 15 Aug 2022 16:55:02 -0400
Subject: drm/amdgpu: support partition drm devices
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Support partition drm devices on GC_HWIP IP_VERSION(9, 4, 3).
This is a temporary solution and will be superceded.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-and-tested-by: Philip Yang<Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 32 ++++++++++++++++
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h | 2 +
drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c | 59 +++++++++++++++++++++++++++++-
drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h | 5 +++
6 files changed, 99 insertions(+), 1 deletion(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index c2feaf2fd070..eb2fb968e3af 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -107,6 +107,7 @@
#include "amdgpu_fdinfo.h"
#include "amdgpu_mca.h"
#include "amdgpu_ras.h"
+#include "amdgpu_xcp.h"
#define MAX_GPU_INSTANCE 64
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 0c6f983fb2ad..4d9c535bcb0c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -6065,6 +6065,7 @@ void amdgpu_device_halt(struct amdgpu_device *adev)
struct pci_dev *pdev = adev->pdev;
struct drm_device *ddev = adev_to_drm(adev);
+ amdgpu_xcp_dev_unplug(adev);
drm_dev_unplug(ddev);
amdgpu_irq_disable_all(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index aa466a9eb956..7300df2a342c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2185,6 +2185,10 @@ retry_init:
goto err_pci;
}
+ ret = amdgpu_xcp_dev_register(adev, ent);
+ if (ret)
+ goto err_pci;
+
/*
* 1. don't init fbdev on hw without DCE
* 2. don't init fbdev if there are no connectors
@@ -2257,6 +2261,7 @@ amdgpu_pci_remove(struct pci_dev *pdev)
struct drm_device *dev = pci_get_drvdata(pdev);
struct amdgpu_device *adev = drm_to_adev(dev);
+ amdgpu_xcp_dev_unplug(adev);
drm_dev_unplug(dev);
if (adev->pm.rpm_mode != AMDGPU_RUNPM_NONE) {
@@ -2840,6 +2845,33 @@ static const struct drm_driver amdgpu_kms_driver = {
.patchlevel = KMS_DRIVER_PATCHLEVEL,
};
+const struct drm_driver amdgpu_partition_driver = {
+ .driver_features =
+ DRIVER_GEM | DRIVER_RENDER | DRIVER_SYNCOBJ |
+ DRIVER_SYNCOBJ_TIMELINE,
+ .open = amdgpu_driver_open_kms,
+ .postclose = amdgpu_driver_postclose_kms,
+ .lastclose = amdgpu_driver_lastclose_kms,
+ .ioctls = amdgpu_ioctls_kms,
+ .num_ioctls = ARRAY_SIZE(amdgpu_ioctls_kms),
+ .dumb_create = amdgpu_mode_dumb_create,
+ .dumb_map_offset = amdgpu_mode_dumb_mmap,
+ .fops = &amdgpu_driver_kms_fops,
+ .release = &amdgpu_driver_release_kms,
+
+ .prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+ .prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+ .gem_prime_import = amdgpu_gem_prime_import,
+ .gem_prime_mmap = drm_gem_prime_mmap,
+
+ .name = DRIVER_NAME,
+ .desc = DRIVER_DESC,
+ .date = DRIVER_DATE,
+ .major = KMS_DRIVER_MAJOR,
+ .minor = KMS_DRIVER_MINOR,
+ .patchlevel = KMS_DRIVER_PATCHLEVEL,
+};
+
static struct pci_error_handlers amdgpu_pci_err_handler = {
.error_detected = amdgpu_pci_error_detected,
.mmio_enabled = amdgpu_pci_mmio_enabled,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
index 8178323e4bef..5bc2cb661af7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
@@ -42,6 +42,8 @@
#define DRIVER_DESC "AMD GPU"
#define DRIVER_DATE "20150101"
+extern const struct drm_driver amdgpu_partition_driver;
+
long amdgpu_drm_ioctl(struct file *filp,
unsigned int cmd, unsigned long arg);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
index bca226cc4e0b..8b28b18e4291 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
@@ -22,6 +22,9 @@
*/
#include "amdgpu.h"
#include "amdgpu_xcp.h"
+#include "amdgpu_drv.h"
+
+#include <drm/drm_drv.h>
static int __amdgpu_xcp_run(struct amdgpu_xcp_mgr *xcp_mgr,
struct amdgpu_xcp_ip *xcp_ip, int xcp_state)
@@ -217,6 +220,31 @@ int amdgpu_xcp_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags)
return mode;
}
+static int amdgpu_xcp_dev_alloc(struct amdgpu_device *adev)
+{
+ struct drm_device *p_ddev;
+ struct pci_dev *pdev;
+ struct drm_device *ddev;
+ int i;
+
+ pdev = adev->pdev;
+ ddev = adev_to_drm(adev);
+
+ for (i = 0; i < MAX_XCP; i++) {
+ p_ddev = drm_dev_alloc(&amdgpu_partition_driver,
+ &pci_upstream_bridge(pdev)->dev);
+ if (IS_ERR(p_ddev))
+ return PTR_ERR(p_ddev);
+
+ /* Redirect all IOCTLs to the primary device */
+ p_ddev->render->dev = ddev;
+ p_ddev->vma_offset_manager = ddev->vma_offset_manager;
+ adev->xcp_mgr->xcp[i].ddev = p_ddev;
+ }
+
+ return 0;
+}
+
int amdgpu_xcp_mgr_init(struct amdgpu_device *adev, int init_mode,
int init_num_xcps,
struct amdgpu_xcp_mgr_funcs *xcp_funcs)
@@ -242,7 +270,7 @@ int amdgpu_xcp_mgr_init(struct amdgpu_device *adev, int init_mode,
adev->xcp_mgr = xcp_mgr;
- return 0;
+ return amdgpu_xcp_dev_alloc(adev);
}
int amdgpu_xcp_get_partition(struct amdgpu_xcp_mgr *xcp_mgr,
@@ -278,3 +306,32 @@ int amdgpu_xcp_get_inst_details(struct amdgpu_xcp *xcp,
return 0;
}
+
+int amdgpu_xcp_dev_register(struct amdgpu_device *adev,
+ const struct pci_device_id *ent)
+{
+ int i, ret;
+
+ if (!adev->xcp_mgr)
+ return 0;
+
+ for (i = 0; i < MAX_XCP; i++) {
+ ret = drm_dev_register(adev->xcp_mgr->xcp[i].ddev, ent->driver_data);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev)
+{
+ int i;
+
+ if (!adev->xcp_mgr)
+ return;
+
+ for (i = 0; i < MAX_XCP; i++)
+ drm_dev_unplug(adev->xcp_mgr->xcp[i].ddev);
+}
+
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h
index e1319b887bf3..dad0b98d1ae7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h
@@ -70,6 +70,7 @@ struct amdgpu_xcp {
uint8_t id;
uint8_t mem_id;
bool valid;
+ struct drm_device *ddev;
};
struct amdgpu_xcp_mgr {
@@ -115,6 +116,10 @@ int amdgpu_xcp_get_inst_details(struct amdgpu_xcp *xcp,
enum AMDGPU_XCP_IP_BLOCK ip,
uint32_t *inst_mask);
+int amdgpu_xcp_dev_register(struct amdgpu_device *adev,
+ const struct pci_device_id *ent);
+void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev);
+
static inline int amdgpu_xcp_get_num_xcp(struct amdgpu_xcp_mgr *xcp_mgr)
{
if (!xcp_mgr)
--
cgit v1.2.3
From 5d6cd20075c823565e7550f8de70d7615ec3c8b7 Mon Sep 17 00:00:00 2001
From: Shiwu Zhang <shiwu.zhang@amd.com>
Date: Mon, 22 May 2023 15:58:10 +0800
Subject: drm/amdgpu: add the accelerator pcie class
v2: add the base class id for accelerator (lijo)
v3: add the new pci class in amdgpu tree (hawking)
Signed-off-by: Shiwu Zhang <shiwu.zhang@amd.com>
Acked-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 5 +++++
drivers/gpu/drm/amd/include/amd_shared.h | 1 +
2 files changed, 6 insertions(+)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 7300df2a342c..422c36ed8f36 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2041,6 +2041,11 @@ static const struct pci_device_id pciidlist[] = {
.class_mask = 0xffffff,
.driver_data = CHIP_IP_DISCOVERY },
+ { PCI_DEVICE(0x1002, PCI_ANY_ID),
+ .class = AMD_ACCELERATOR_PROCESSING << 8,
+ .class_mask = 0xffffff,
+ .driver_data = CHIP_IP_DISCOVERY },
+
{0, 0, 0}
};
diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h
index f175e65b853a..57d95e2cc54b 100644
--- a/drivers/gpu/drm/amd/include/amd_shared.h
+++ b/drivers/gpu/drm/amd/include/amd_shared.h
@@ -27,6 +27,7 @@
#define AMD_MAX_USEC_TIMEOUT 1000000 /* 1000 ms */
+#define AMD_ACCELERATOR_PROCESSING 0x1200 /* hardcoded pci class */
/*
* Chip flags
--
cgit v1.2.3
From 9938333a46c9e20539c85ca7df42a739541b0493 Mon Sep 17 00:00:00 2001
From: James Zhu <James.Zhu@amd.com>
Date: Tue, 25 Apr 2023 17:02:48 -0400
Subject: drm/amdgpu: use amdxcp platform device as spatial partition
Use amdxcp platform device as spatial partition device.
-v2: remove unused variable
Signed-off-by: James Zhu <James.Zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 2 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c | 12 +++++-------
2 files changed, 7 insertions(+), 7 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 422c36ed8f36..7489b2b1a0d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -50,6 +50,7 @@
#include "amdgpu_ras.h"
#include "amdgpu_xgmi.h"
#include "amdgpu_reset.h"
+#include "../amdxcp/amdgpu_xcp_drv.h"
/*
* KMS wrapper.
@@ -2948,6 +2949,7 @@ static void __exit amdgpu_exit(void)
amdgpu_sync_fini();
amdgpu_fence_slab_fini();
mmu_notifier_synchronize();
+ amdgpu_xcp_drv_release();
}
module_init(amdgpu_init);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
index 86087faab689..d733fa6e7477 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
@@ -25,6 +25,7 @@
#include "amdgpu_drv.h"
#include <drm/drm_drv.h>
+#include "../amdxcp/amdgpu_xcp_drv.h"
static int __amdgpu_xcp_run(struct amdgpu_xcp_mgr *xcp_mgr,
struct amdgpu_xcp_ip *xcp_ip, int xcp_state)
@@ -226,18 +227,15 @@ int amdgpu_xcp_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags)
static int amdgpu_xcp_dev_alloc(struct amdgpu_device *adev)
{
struct drm_device *p_ddev;
- struct pci_dev *pdev;
struct drm_device *ddev;
- int i;
+ int i, ret;
- pdev = adev->pdev;
ddev = adev_to_drm(adev);
for (i = 0; i < MAX_XCP; i++) {
- p_ddev = drm_dev_alloc(&amdgpu_partition_driver,
- &pci_upstream_bridge(pdev)->dev);
- if (IS_ERR(p_ddev))
- return PTR_ERR(p_ddev);
+ ret = amdgpu_xcp_drm_dev_alloc(&p_ddev);
+ if (ret)
+ return ret;
/* Redirect all IOCTLs to the primary device */
adev->xcp_mgr->xcp[i].rdev = p_ddev->render->dev;
--
cgit v1.2.3
From c1ac2ea802f5adfd1d128fc01375af9c5f113932 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Tue, 6 Jun 2023 11:14:04 -0400
Subject: drm/amdgpu: add missing radeon secondary PCI ID
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
0x5b70 is a missing RV370 secondary id. Add it so
we don't try and probe it with amdgpu.
Cc: michel@daenzer.net
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Tested-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 1 +
1 file changed, 1 insertion(+)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 7489b2b1a0d0..b8a1e4571cd9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -1640,6 +1640,7 @@ static const u16 amdgpu_unsupported_pciidlist[] = {
0x5874,
0x5940,
0x5941,
+ 0x5b70,
0x5b72,
0x5b73,
0x5b74,
--
cgit v1.2.3
From 9d65b1b4bcf3918164e17365eec169875eef8ee3 Mon Sep 17 00:00:00 2001
From: Shiwu Zhang <shiwu.zhang@amd.com>
Date: Tue, 23 May 2023 12:02:32 +0800
Subject: drm/amdgpu: add the accelerator PCIe class
Add the accelerator PCIe class and match the
class in amdgpu for 0x1002 devices of that class.
From PCI spec:
"PCI Code and ID Assignment, r1.9, sec 1, 1.19"
Signed-off-by: Shiwu Zhang <shiwu.zhang@amd.com>
Acked-by: Lijo Lazar <lijo.lazar@amd.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com> # pci_ids.h
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 +-
drivers/gpu/drm/amd/include/amd_shared.h | 1 -
include/linux/pci_ids.h | 3 +++
4 files changed, 5 insertions(+), 3 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index b8a1e4571cd9..8e58d187b173 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2044,7 +2044,7 @@ static const struct pci_device_id pciidlist[] = {
.driver_data = CHIP_IP_DISCOVERY },
{ PCI_DEVICE(0x1002, PCI_ANY_ID),
- .class = AMD_ACCELERATOR_PROCESSING << 8,
+ .class = PCI_CLASS_ACCELERATOR_PROCESSING << 8,
.class_mask = 0xffffff,
.driver_data = CHIP_IP_DISCOVERY },
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 0f1ca0136f50..25b4d7f0bd35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -57,7 +57,7 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev)
/* enable virtual display */
if (adev->asic_type != CHIP_ALDEBARAN &&
adev->asic_type != CHIP_ARCTURUS &&
- ((adev->pdev->class >> 8) != AMD_ACCELERATOR_PROCESSING)) {
+ ((adev->pdev->class >> 8) != PCI_CLASS_ACCELERATOR_PROCESSING)) {
if (adev->mode_info.num_crtc == 0)
adev->mode_info.num_crtc = 1;
adev->enable_virtual_display = true;
diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h
index 57d95e2cc54b..f175e65b853a 100644
--- a/drivers/gpu/drm/amd/include/amd_shared.h
+++ b/drivers/gpu/drm/amd/include/amd_shared.h
@@ -27,7 +27,6 @@
#define AMD_MAX_USEC_TIMEOUT 1000000 /* 1000 ms */
-#define AMD_ACCELERATOR_PROCESSING 0x1200 /* hardcoded pci class */
/*
* Chip flags
diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h
index 45c3d62e616d..0fbfbda3dc26 100644
--- a/include/linux/pci_ids.h
+++ b/include/linux/pci_ids.h
@@ -151,6 +151,9 @@
#define PCI_CLASS_SP_DPIO 0x1100
#define PCI_CLASS_SP_OTHER 0x1180
+#define PCI_BASE_CLASS_ACCELERATOR 0x12
+#define PCI_CLASS_ACCELERATOR_PROCESSING 0x1200
+
#define PCI_CLASS_OTHERS 0xff
/* Vendors and devices. Sort key: vendor first, device next. */
--
cgit v1.2.3
From 80e709ee6ecc9eba8bd8d188218472822e1b38bd Mon Sep 17 00:00:00 2001
From: Chong Li <chongli2@amd.com>
Date: Wed, 7 Jun 2023 15:56:12 +0800
Subject: drm/amdgpu: add option params to enforce process isolation between
graphics and compute
enforce process isolation between graphics and compute via using the same reserved vmid.
v2: remove params "struct amdgpu_vm *vm" from
amdgpu_vmid_alloc_reserved and amdgpu_vmid_free_reserved.
Signed-off-by: Chong Li <chongli2@amd.com>
Reviewed-by: Christian Koenig <Christian.Koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 10 +++++++++-
drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c | 17 +++++++----------
drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h | 6 ++----
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 25 +++++++++++++++++--------
5 files changed, 36 insertions(+), 23 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 74104d46a7be..a84bd4a0c421 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -214,6 +214,7 @@ extern int amdgpu_force_asic_type;
extern int amdgpu_smartshift_bias;
extern int amdgpu_use_xgmi_p2p;
extern int amdgpu_mtype_local;
+extern bool enforce_isolation;
#ifdef CONFIG_HSA_AMD
extern int sched_policy;
extern bool debug_evictions;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 8e58d187b173..999d008b6b48 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -153,7 +153,7 @@ uint amdgpu_pg_mask = 0xffffffff;
uint amdgpu_sdma_phase_quantum = 32;
char *amdgpu_disable_cu;
char *amdgpu_virtual_display;
-
+bool enforce_isolation;
/*
* OverDrive(bit 14) disabled by default
* GFX DCS(bit 19) disabled by default
@@ -973,6 +973,14 @@ MODULE_PARM_DESC(
4 = AMDGPU_CPX_PARTITION_MODE)");
module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
+
+/**
+ * DOC: enforce_isolation (bool)
+ * enforce process isolation between graphics and compute via using the same reserved vmid.
+ */
+module_param(enforce_isolation, bool, 0444);
+MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on");
+
/* These devices are not supported by amdgpu.
* They are supported by the mach64, r128, radeon drivers
*/
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
index c991ca0b7a1c..ff1ea99292fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
@@ -409,7 +409,7 @@ int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
if (r || !idle)
goto error;
- if (vm->reserved_vmid[vmhub]) {
+ if (vm->reserved_vmid[vmhub] || (enforce_isolation && (vmhub == AMDGPU_GFXHUB(0)))) {
r = amdgpu_vmid_grab_reserved(vm, ring, job, &id, fence);
if (r || !id)
goto error;
@@ -460,14 +460,11 @@ error:
}
int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
- struct amdgpu_vm *vm,
unsigned vmhub)
{
struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
mutex_lock(&id_mgr->lock);
- if (vm->reserved_vmid[vmhub])
- goto unlock;
++id_mgr->reserved_use_count;
if (!id_mgr->reserved) {
@@ -479,27 +476,23 @@ int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
list_del_init(&id->list);
id_mgr->reserved = id;
}
- vm->reserved_vmid[vmhub] = true;
-unlock:
mutex_unlock(&id_mgr->lock);
return 0;
}
void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
- struct amdgpu_vm *vm,
unsigned vmhub)
{
struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
mutex_lock(&id_mgr->lock);
- if (vm->reserved_vmid[vmhub] &&
- !--id_mgr->reserved_use_count) {
+ if (!--id_mgr->reserved_use_count) {
/* give the reserved ID back to normal round robin */
list_add(&id_mgr->reserved->list, &id_mgr->ids_lru);
id_mgr->reserved = NULL;
}
- vm->reserved_vmid[vmhub] = false;
+
mutex_unlock(&id_mgr->lock);
}
@@ -578,6 +571,10 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev)
list_add_tail(&id_mgr->ids[j].list, &id_mgr->ids_lru);
}
}
+ /* alloc a default reserved vmid to enforce isolation */
+ if (enforce_isolation)
+ amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(0));
+
}
/**
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
index d1cc09b45da4..fa8c42c83d5d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
@@ -79,11 +79,9 @@ void amdgpu_pasid_free_delayed(struct dma_resv *resv,
bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
struct amdgpu_vmid *id);
int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
- struct amdgpu_vm *vm,
- unsigned vmhub);
+ unsigned vmhub);
void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
- struct amdgpu_vm *vm,
- unsigned vmhub);
+ unsigned vmhub);
int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
struct amdgpu_job *job, struct dma_fence **fence);
void amdgpu_vmid_reset(struct amdgpu_device *adev, unsigned vmhub,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 76d57bc7ac62..dc80c9c8fd14 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -2284,8 +2284,14 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
}
dma_fence_put(vm->last_update);
- for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
- amdgpu_vmid_free_reserved(adev, vm, i);
+
+ for (i = 0; i < AMDGPU_MAX_VMHUBS; i++) {
+ if (vm->reserved_vmid[i]) {
+ amdgpu_vmid_free_reserved(adev, i);
+ vm->reserved_vmid[i] = false;
+ }
+ }
+
}
/**
@@ -2368,7 +2374,6 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
union drm_amdgpu_vm *args = data;
struct amdgpu_device *adev = drm_to_adev(dev);
struct amdgpu_fpriv *fpriv = filp->driver_priv;
- int r;
/* No valid flags defined yet */
if (args->in.flags)
@@ -2377,13 +2382,17 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
switch (args->in.op) {
case AMDGPU_VM_OP_RESERVE_VMID:
/* We only have requirement to reserve vmid from gfxhub */
- r = amdgpu_vmid_alloc_reserved(adev, &fpriv->vm,
- AMDGPU_GFXHUB(0));
- if (r)
- return r;
+ if (!fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)]) {
+ amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(0));
+ fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)] = true;
+ }
+
break;
case AMDGPU_VM_OP_UNRESERVE_VMID:
- amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB(0));
+ if (fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)]) {
+ amdgpu_vmid_free_reserved(adev, AMDGPU_GFXHUB(0));
+ fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)] = false;
+ }
break;
default:
return -EINVAL;
--
cgit v1.2.3