drm/amdgpu: have bos for PDs/PTS cpu accessible when kfd uses cpu to update vm

When kfd uses cpu to update vm iterates all current PDs/PTs bos, adds
AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED flag and kmap them to kernel virtual
address space before kfd updates the vm that was created by gfx.
Signed-off-by: Xiaogang Chen <Xiaogang.Chen@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_vm.c

@@ -2279,16 +2279,13 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 			goto unreserve_bo;
 
 		vm->update_funcs = &amdgpu_vm_cpu_funcs;
+		r = amdgpu_vm_pt_map_tables(adev, vm);
+		if (r)
+			goto unreserve_bo;
+
 	} else {
 		vm->update_funcs = &amdgpu_vm_sdma_funcs;
 	}
-	/*
-	 * Make sure root PD gets mapped. As vm_update_mode could be changed
-	 * when turning a GFX VM into a compute VM.
-	 */
-	r = vm->update_funcs->map_table(to_amdgpu_bo_vm(vm->root.bo));
-	if (r)
-		goto unreserve_bo;
 
 	dma_fence_put(vm->last_update);
 	vm->last_update = dma_fence_get_stub();

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

@@ -497,6 +497,8 @@ void amdgpu_vm_pt_free_work(struct work_struct *work);
 void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m);
 #endif
 
+int amdgpu_vm_pt_map_tables(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+
 /**
  * amdgpu_vm_tlb_seq - return tlb flush sequence number
  * @vm: the amdgpu_vm structure to query

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

@@ -31,6 +31,7 @@
  */
 static int amdgpu_vm_cpu_map_table(struct amdgpu_bo_vm *table)
 {
+	table->bo.flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	return amdgpu_bo_kmap(&table->bo, NULL);
 }
 

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

@@ -1075,3 +1075,31 @@ int amdgpu_vm_ptes_update(struct amdgpu_vm_update_params *params,
 
 	return 0;
 }
+
+/**
+ * amdgpu_vm_pt_map_tables - have bo of root PD cpu accessible
+ * @adev: amdgpu device structure
+ * @vm: amdgpu vm structure
+ *
+ * make root page directory and everything below it cpu accessible.
+ */
+int amdgpu_vm_pt_map_tables(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+{
+	struct amdgpu_vm_pt_cursor cursor;
+	struct amdgpu_vm_bo_base *entry;
+
+	for_each_amdgpu_vm_pt_dfs_safe(adev, vm, NULL, cursor, entry) {
+
+		struct amdgpu_bo_vm *bo;
+		int r;
+
+		if (entry->bo) {
+			bo = to_amdgpu_bo_vm(entry->bo);
+			r = vm->update_funcs->map_table(bo);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}