Merge tag 'amd-drm-next-5.17-2022-01-12' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-5.17-2022-01-12:

amdgpu:
- SR-IOV fixes
- Suspend/resume fixes
- Display fixes
- DMCUB fixes
- DP alt mode fixes
- RAS fixes
- UBSAN fix
- Navy Flounder VCN fix
- ttm resource manager cleanup
- default_groups change for kobj_type
- vkms fix
- Aldebaran fixes

amdkfd:
- SDMA ECC interrupt fix
- License clarification
- Pointer check fix
- DQM fixes for hawaii
- default_groups change for kobj_type
- Typo fixes

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220113030537.5758-1-alexander.deucher@amd.com

Author: danvet

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c

@@ -514,13 +514,6 @@ int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd,
 	return r;
 }
 
-uint64_t amdgpu_amdkfd_get_vram_usage(struct amdgpu_device *adev)
-{
-	struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
-
-	return amdgpu_vram_mgr_usage(vram_man);
-}
-
 uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst,
 					  struct amdgpu_device *src)
 {

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h

@@ -223,7 +223,6 @@ int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd,
 				  uint64_t *bo_size, void *metadata_buffer,
 				  size_t buffer_size, uint32_t *metadata_size,
 				  uint32_t *flags);
-uint64_t amdgpu_amdkfd_get_vram_usage(struct amdgpu_device *adev);
 uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst,
 					  struct amdgpu_device *src);
 int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst,

drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

@@ -298,7 +298,6 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 {
 	s64 time_us, increment_us;
 	u64 free_vram, total_vram, used_vram;
-	struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
 	/* Allow a maximum of 200 accumulated ms. This is basically per-IB
 	 * throttling.
 	 *
@@ -315,7 +314,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	}
 
 	total_vram = adev->gmc.real_vram_size - atomic64_read(&adev->vram_pin_size);
-	used_vram = amdgpu_vram_mgr_usage(vram_man);
+	used_vram = amdgpu_vram_mgr_usage(&adev->mman.vram_mgr);
 	free_vram = used_vram >= total_vram ? 0 : total_vram - used_vram;
 
 	spin_lock(&adev->mm_stats.lock);
@@ -362,7 +361,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	if (!amdgpu_gmc_vram_full_visible(&adev->gmc)) {
 		u64 total_vis_vram = adev->gmc.visible_vram_size;
 		u64 used_vis_vram =
-		  amdgpu_vram_mgr_vis_usage(vram_man);
+		  amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr);
 
 		if (used_vis_vram < total_vis_vram) {
 			u64 free_vis_vram = total_vis_vram - used_vis_vram;

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

@@ -552,7 +552,7 @@ void amdgpu_device_wreg(struct amdgpu_device *adev,
 }
 
 /**
- * amdgpu_mm_wreg_mmio_rlc -  write register either with mmio or with RLC path if in range
+ * amdgpu_mm_wreg_mmio_rlc -  write register either with direct/indirect mmio or with RLC path if in range
  *
  * this function is invoked only the debugfs register access
  */
@@ -567,6 +567,8 @@ void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
 	    adev->gfx.rlc.funcs->is_rlcg_access_range) {
 		if (adev->gfx.rlc.funcs->is_rlcg_access_range(adev, reg))
 			return adev->gfx.rlc.funcs->sriov_wreg(adev, reg, v, 0, 0);
+	} else if ((reg * 4) >= adev->rmmio_size) {
+		adev->pcie_wreg(adev, reg * 4, v);
 	} else {
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 	}
@@ -1448,7 +1450,7 @@ static int amdgpu_device_init_apu_flags(struct amdgpu_device *adev)
 			adev->apu_flags |= AMD_APU_IS_CYAN_SKILLFISH2;
 		break;
 	default:
-		return -EINVAL;
+		break;
 	}
 
 	return 0;
@@ -3496,9 +3498,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->psp.mutex);
 	mutex_init(&adev->notifier_lock);
 
-	r = amdgpu_device_init_apu_flags(adev);
-	if (r)
-		return r;
+	 amdgpu_device_init_apu_flags(adev);
 
 	r = amdgpu_device_check_arguments(adev);
 	if (r)
@@ -3833,6 +3833,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 
 static void amdgpu_device_unmap_mmio(struct amdgpu_device *adev)
 {
+
 	/* Clear all CPU mappings pointing to this device */
 	unmap_mapping_range(adev->ddev.anon_inode->i_mapping, 0, 0, 1);
 
@@ -3913,6 +3914,8 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 
 void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 {
+	int idx;
+
 	amdgpu_fence_driver_sw_fini(adev);
 	amdgpu_device_ip_fini(adev);
 	release_firmware(adev->firmware.gpu_info_fw);
@@ -3937,6 +3940,14 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA)
 		vga_client_unregister(adev->pdev);
 
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+
+		iounmap(adev->rmmio);
+		adev->rmmio = NULL;
+		amdgpu_device_doorbell_fini(adev);
+		drm_dev_exit(idx);
+	}
+
 	if (IS_ENABLED(CONFIG_PERF_EVENTS))
 		amdgpu_pmu_fini(adev);
 	if (adev->mman.discovery_bin)
@@ -3957,8 +3968,8 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
  */
 static void amdgpu_device_evict_resources(struct amdgpu_device *adev)
 {
-	/* No need to evict vram on APUs for suspend to ram */
-	if (adev->in_s3 && (adev->flags & AMD_IS_APU))
+	/* 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))
 		return;
 
 	if (amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM))
@@ -4005,16 +4016,11 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 	if (!adev->in_s0ix)
 		amdgpu_amdkfd_suspend(adev, adev->in_runpm);
 
-	/* First evict vram memory */
 	amdgpu_device_evict_resources(adev);
 
 	amdgpu_fence_driver_hw_fini(adev);
 
 	amdgpu_device_ip_suspend_phase2(adev);
-	/* This second call to evict device resources is to evict
-	 * the gart page table using the CPU.
-	 */
-	amdgpu_device_evict_resources(adev);
 
 	return 0;
 }
@@ -4359,8 +4365,6 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 		goto error;
 
 	amdgpu_virt_init_data_exchange(adev);
-	/* we need recover gart prior to run SMC/CP/SDMA resume */
-	amdgpu_gtt_mgr_recover(ttm_manager_type(&adev->mman.bdev, TTM_PL_TT));
 
 	r = amdgpu_device_fw_loading(adev);
 	if (r)
@@ -4680,10 +4684,6 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 					amdgpu_inc_vram_lost(tmp_adev);
 				}
 
-				r = amdgpu_gtt_mgr_recover(ttm_manager_type(&tmp_adev->mman.bdev, TTM_PL_TT));
-				if (r)
-					goto out;
-
 				r = amdgpu_device_fw_loading(tmp_adev);
 				if (r)
 					return r;

drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c

@@ -550,7 +550,8 @@ void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev)
 	}
 	/* some IP discovery tables on Navy Flounder don't have this set correctly */
 	if ((adev->ip_versions[UVD_HWIP][1] == IP_VERSION(3, 0, 1)) &&
-	    (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 2)))
+	    (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 2)) &&
+	    (adev->pdev->revision != 0xFF))
 		adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1;
 	if (vcn_harvest_count == adev->vcn.num_vcn_inst) {
 		adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK;

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -2166,9 +2166,9 @@ static int amdgpu_pmops_suspend(struct device *dev)
 
 	if (amdgpu_acpi_is_s0ix_active(adev))
 		adev->in_s0ix = true;
-	adev->in_s3 = true;
+	else
+		adev->in_s3 = true;
 	r = amdgpu_device_suspend(drm_dev, true);
-	adev->in_s3 = false;
 	if (r)
 		return r;
 	if (!adev->in_s0ix)
@@ -2189,6 +2189,8 @@ static int amdgpu_pmops_resume(struct device *dev)
 	r = amdgpu_device_resume(drm_dev, true);
 	if (amdgpu_acpi_is_s0ix_active(adev))
 		adev->in_s0ix = false;
+	else
+		adev->in_s3 = false;
 	return r;
 }
 

drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c

@@ -114,80 +114,12 @@ void amdgpu_gart_dummy_page_fini(struct amdgpu_device *adev)
  */
 int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev)
 {
-	int r;
-
-	if (adev->gart.bo == NULL) {
-		struct amdgpu_bo_param bp;
-
-		memset(&bp, 0, sizeof(bp));
-		bp.size = adev->gart.table_size;
-		bp.byte_align = PAGE_SIZE;
-		bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
-		bp.flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-			AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
-		bp.type = ttm_bo_type_kernel;
-		bp.resv = NULL;
-		bp.bo_ptr_size = sizeof(struct amdgpu_bo);
-
-		r = amdgpu_bo_create(adev, &bp, &adev->gart.bo);
-		if (r) {
-			return r;
-		}
-	}
-	return 0;
-}
-
-/**
- * amdgpu_gart_table_vram_pin - pin gart page table in vram
- *
- * @adev: amdgpu_device pointer
- *
- * Pin the GART page table in vram so it will not be moved
- * by the memory manager (pcie r4xx, r5xx+).  These asics require the
- * gart table to be in video memory.
- * Returns 0 for success, error for failure.
- */
-int amdgpu_gart_table_vram_pin(struct amdgpu_device *adev)
-{
-	int r;
-
-	r = amdgpu_bo_reserve(adev->gart.bo, false);
-	if (unlikely(r != 0))
-		return r;
-	r = amdgpu_bo_pin(adev->gart.bo, AMDGPU_GEM_DOMAIN_VRAM);
-	if (r) {
-		amdgpu_bo_unreserve(adev->gart.bo);
-		return r;
-	}
-	r = amdgpu_bo_kmap(adev->gart.bo, &adev->gart.ptr);
-	if (r)
-		amdgpu_bo_unpin(adev->gart.bo);
-	amdgpu_bo_unreserve(adev->gart.bo);
-	return r;
-}
-
-/**
- * amdgpu_gart_table_vram_unpin - unpin gart page table in vram
- *
- * @adev: amdgpu_device pointer
- *
- * Unpin the GART page table in vram (pcie r4xx, r5xx+).
- * These asics require the gart table to be in video memory.
- */
-void amdgpu_gart_table_vram_unpin(struct amdgpu_device *adev)
-{
-	int r;
+	if (adev->gart.bo != NULL)
+		return 0;
 
-	if (adev->gart.bo == NULL) {
-		return;
-	}
-	r = amdgpu_bo_reserve(adev->gart.bo, true);
-	if (likely(r == 0)) {
-		amdgpu_bo_kunmap(adev->gart.bo);
-		amdgpu_bo_unpin(adev->gart.bo);
-		amdgpu_bo_unreserve(adev->gart.bo);
-		adev->gart.ptr = NULL;
-	}
+	return amdgpu_bo_create_kernel(adev,  adev->gart.table_size, PAGE_SIZE,
+				       AMDGPU_GEM_DOMAIN_VRAM, &adev->gart.bo,
+				       NULL, (void *)&adev->gart.ptr);
 }
 
 /**
@@ -201,11 +133,7 @@ void amdgpu_gart_table_vram_unpin(struct amdgpu_device *adev)
  */
 void amdgpu_gart_table_vram_free(struct amdgpu_device *adev)
 {
-	if (adev->gart.bo == NULL) {
-		return;
-	}
-	amdgpu_bo_unref(&adev->gart.bo);
-	adev->gart.ptr = NULL;
+	amdgpu_bo_free_kernel(&adev->gart.bo, NULL, (void *)&adev->gart.ptr);
 }
 
 /*

drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c

@@ -264,9 +264,6 @@ static int amdgpu_gem_object_mmap(struct drm_gem_object *obj, struct vm_area_str
 	    !(vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC)))
 		vma->vm_flags &= ~VM_MAYWRITE;
 
-	if (bo->kfd_bo)
-		vma->vm_flags |= VM_DONTCOPY;
-
 	return drm_gem_ttm_mmap(obj, vma);
 }
 

drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c

@@ -77,10 +77,8 @@ static ssize_t amdgpu_mem_info_gtt_used_show(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	struct ttm_resource_manager *man;
 
-	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
-	return sysfs_emit(buf, "%llu\n", amdgpu_gtt_mgr_usage(man));
+	return sysfs_emit(buf, "%llu\n", amdgpu_gtt_mgr_usage(&adev->mman.gtt_mgr));
 }
 
 static DEVICE_ATTR(mem_info_gtt_total, S_IRUGO,
@@ -206,30 +204,27 @@ static void amdgpu_gtt_mgr_del(struct ttm_resource_manager *man,
 /**
  * amdgpu_gtt_mgr_usage - return usage of GTT domain
  *
- * @man: TTM memory type manager
+ * @mgr: amdgpu_gtt_mgr pointer
  *
  * Return how many bytes are used in the GTT domain
  */
-uint64_t amdgpu_gtt_mgr_usage(struct ttm_resource_manager *man)
+uint64_t amdgpu_gtt_mgr_usage(struct amdgpu_gtt_mgr *mgr)
 {
-	struct amdgpu_gtt_mgr *mgr = to_gtt_mgr(man);
-
 	return atomic64_read(&mgr->used) * PAGE_SIZE;
 }
 
 /**
  * amdgpu_gtt_mgr_recover - re-init gart
  *
- * @man: TTM memory type manager
+ * @mgr: amdgpu_gtt_mgr pointer
  *
  * Re-init the gart for each known BO in the GTT.
  */
-int amdgpu_gtt_mgr_recover(struct ttm_resource_manager *man)
+int amdgpu_gtt_mgr_recover(struct amdgpu_gtt_mgr *mgr)
 {
-	struct amdgpu_gtt_mgr *mgr = to_gtt_mgr(man);
-	struct amdgpu_device *adev;
 	struct amdgpu_gtt_node *node;
 	struct drm_mm_node *mm_node;
+	struct amdgpu_device *adev;
 	int r = 0;
 
 	adev = container_of(mgr, typeof(*adev), mman.gtt_mgr);