Re: [PATCH 3/3] virtgpu: deallocate capsets on device deinit

[Stefano Garzarella]

On Mon, May 05, 2025 at 07:22:35PM +0300, Dmitry Osipenko wrote:
> On 5/5/25 18:58, Dmitry Osipenko wrote:
> > On 5/5/25 11:59, Manos Pitsidianakis wrote:
> > > diff --git a/drivers/gpu/drm/virtio/virtgpu_kms.c b/drivers/gpu/drm/virtio/virtgpu_kms.c
> > > index 7b3c4d314f8eee692e2842a7056d6dc64936fc2f..a8b751179332b9ec2fbba1392a6ee0e638a5192e 100644
> > > --- a/drivers/gpu/drm/virtio/virtgpu_kms.c
> > > +++ b/drivers/gpu/drm/virtio/virtgpu_kms.c
> > > @@ -286,6 +286,10 @@ void virtio_gpu_deinit(struct drm_device *dev)
> > >  	flush_work(&vgdev->cursorq.dequeue_work);
> > >  	flush_work(&vgdev->config_changed_work);
> > >  	virtio_reset_device(vgdev->vdev);
> > > +	spin_lock(&vgdev->display_info_lock);
> > > +	drmm_kfree(dev, vgdev->capsets);
> > > +	vgdev->capsets = NULL;
> > > +	spin_unlock(&vgdev->display_info_lock);
> >
> > Isn't this lock superfluous?
>
> Wait a minute, vgdev->capsets is allocated using drmm, hence it's
> auto-freed when DRM device is freed. This patch shouldn't be needed.

Yep, good point. I mean the patch is not wrong, but I think we can avoid 
it.

Thanks,
Stefano