drm/amdgpu: add some extra VM error handling

If updating the PDs fails we now invalidate all entries to try again later.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

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

@@ -1103,6 +1103,32 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 	return r;
 }
 
+/*
+ * amdgpu_vm_invalidate_level - mark all PD levels as invalid
+ *
+ * @parent: parent PD
+ *
+ * Mark all PD level as invalid after an error.
+ */
+static void amdgpu_vm_invalidate_level(struct amdgpu_vm_pt *parent)
+{
+	unsigned pt_idx;
+
+	/*
+	 * Recurse into the subdirectories. This recursion is harmless because
+	 * we only have a maximum of 5 layers.
+	 */
+	for (pt_idx = 0; pt_idx <= parent->last_entry_used; ++pt_idx) {
+		struct amdgpu_vm_pt *entry = &parent->entries[pt_idx];
+
+		if (!entry->bo)
+			continue;
+
+		entry->addr = ~0ULL;
+		amdgpu_vm_invalidate_level(entry);
+	}
+}
+
 /*
  * amdgpu_vm_update_directories - make sure that all directories are valid
  *
@@ -1115,7 +1141,13 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 int amdgpu_vm_update_directories(struct amdgpu_device *adev,
 				 struct amdgpu_vm *vm)
 {
-	return amdgpu_vm_update_level(adev, vm, &vm->root, 0);
+	int r;
+
+	r = amdgpu_vm_update_level(adev, vm, &vm->root, 0);
+	if (r)
+		amdgpu_vm_invalidate_level(&vm->root);
+
+	return r;
 }
 
 /**