1 files changed, 29 insertions, 5 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
index df63dc3bca18..dea1a64be44d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
@@ -502,6 +502,7 @@ exit:
 int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			int level, bool immediate, struct amdgpu_bo_vm **vmbo)
 {
+	struct amdgpu_fpriv *fpriv = container_of(vm, struct amdgpu_fpriv, vm);
 	struct amdgpu_bo_param bp;
 	struct amdgpu_bo *bo;
 	struct dma_resv *resv;
@@ -512,7 +513,12 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 
 	bp.size = amdgpu_vm_pt_size(adev, level);
 	bp.byte_align = AMDGPU_GPU_PAGE_SIZE;
-	bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
+
+	if (!adev->gmc.is_app_apu)
+		bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
+	else
+		bp.domain = AMDGPU_GEM_DOMAIN_GTT;
+
 	bp.domain = amdgpu_bo_get_preferred_domain(adev, bp.domain);
 	bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
 		AMDGPU_GEM_CREATE_CPU_GTT_USWC;
@@ -529,6 +535,8 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 
 	bp.type = ttm_bo_type_kernel;
 	bp.no_wait_gpu = immediate;
+	bp.xcp_id_plus1 = fpriv->xcp_id == ~0 ? 0 : fpriv->xcp_id + 1;
+
 	if (vm->root.bo)
 		bp.resv = vm->root.bo->tbo.base.resv;
 
@@ -553,6 +561,7 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	bp.type = ttm_bo_type_kernel;
 	bp.resv = bo->tbo.base.resv;
 	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+	bp.xcp_id_plus1 = fpriv->xcp_id == ~0 ? 0 : fpriv->xcp_id + 1;
 
 	r = amdgpu_bo_create(adev, &bp, &(*vmbo)->shadow);
 
@@ -564,7 +573,6 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		return r;
 	}
 
-	(*vmbo)->shadow->parent = amdgpu_bo_ref(bo);
 	amdgpu_bo_add_to_shadow_list(*vmbo);
 
 	return 0;
@@ -781,13 +789,14 @@ static void amdgpu_vm_pte_update_flags(struct amdgpu_vm_update_params *params,
 				       uint64_t pe, uint64_t addr,
 				       unsigned int count, uint32_t incr,
 				       uint64_t flags)
-
 {
+	struct amdgpu_device *adev = params->adev;
+
 	if (level != AMDGPU_VM_PTB) {
 		flags |= AMDGPU_PDE_PTE;
-		amdgpu_gmc_get_vm_pde(params->adev, level, &addr, &flags);
+		amdgpu_gmc_get_vm_pde(adev, level, &addr, &flags);
 
-	} else if (params->adev->asic_type >= CHIP_VEGA10 &&
+	} else if (adev->asic_type >= CHIP_VEGA10 &&
 		   !(flags & AMDGPU_PTE_VALID) &&
 		   !(flags & AMDGPU_PTE_PRT)) {
 
@@ -795,6 +804,21 @@ static void amdgpu_vm_pte_update_flags(struct amdgpu_vm_update_params *params,
 		flags |= AMDGPU_PTE_EXECUTABLE;
 	}
 
+	/* APUs mapping system memory may need different MTYPEs on different
+	 * NUMA nodes. Only do this for contiguous ranges that can be assumed
+	 * to be on the same NUMA node.
+	 */
+	if ((flags & AMDGPU_PTE_SYSTEM) && (adev->flags & AMD_IS_APU) &&
+	    adev->gmc.gmc_funcs->override_vm_pte_flags &&
+	    num_possible_nodes() > 1) {
+		if (!params->pages_addr)
+			amdgpu_gmc_override_vm_pte_flags(adev, params->vm,
+							 addr, &flags);
+		else
+			dev_dbg(adev->dev,
+				"override_vm_pte_flags skipped: non-contiguous\n");
+	}
+
 	params->vm->update_funcs->update(params, pt, pe, addr, count, incr,
 					 flags);
 }