On 6/24/24 09:43, Vivek Kasireddy wrote: > +static void virtgpu_dma_buf_move_notify(struct dma_buf_attachment *attach) > +{ > + struct drm_gem_object *obj = attach->importer_priv; > + struct virtio_gpu_device *vgdev = obj->dev->dev_private; > + struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj); > + > + if (bo->created) { > + virtio_gpu_cmd_resource_detach_backing(vgdev, > + bo->hw_res_handle); > + if (bo->sgt) > + dma_buf_unmap_attachment(attach, bo->sgt, > + DMA_BIDIRECTIONAL); > + > + bo->sgt = NULL; > + } BO detachment should be fenced, AFAICT. See virtio_gpu_detach_object_fenced() in [1] for example. [1] https://lore.kernel.org/lkml/20240105184624.508603-1-dmitry.osipenko@xxxxxxxxxxxxx/T/#me62dca570fdae39c41bdf0867c9908ef62d5cc2e BTW, for next version please correct the list of emails recipients. I got only patch 0 in my inbox, for patches 1-5 only Gerd Hoffmann is CC'ed. -- Best regards, Dmitry