drm/amdgpu: install stub fence into potential unused fence pointers

When using cpu to update page tables, vm update fences are unused.
Install stub fence into these fence pointers instead of NULL
to avoid NULL dereference when calling dma_fence_wait() on them.

Suggested-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lang Yu <Lang.Yu@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

@@ -1358,6 +1358,7 @@ struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
 	amdgpu_vm_bo_base_init(&bo_va->base, vm, bo);
 
 	bo_va->ref_count = 1;
+	bo_va->last_pt_update = dma_fence_get_stub();
 	INIT_LIST_HEAD(&bo_va->valids);
 	INIT_LIST_HEAD(&bo_va->invalids);
 
@@ -2067,7 +2068,8 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		vm->update_funcs = &amdgpu_vm_cpu_funcs;
 	else
 		vm->update_funcs = &amdgpu_vm_sdma_funcs;
-	vm->last_update = NULL;
+
+	vm->last_update = dma_fence_get_stub();
 	vm->last_unlocked = dma_fence_get_stub();
 	vm->last_tlb_flush = dma_fence_get_stub();
 
@@ -2192,7 +2194,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		goto unreserve_bo;
 
 	dma_fence_put(vm->last_update);
-	vm->last_update = NULL;
+	vm->last_update = dma_fence_get_stub();
 	vm->is_compute_context = true;
 
 	/* Free the shadow bo for compute VM */