Re: [PATCH v3] vduse: Add suspend

[Dan Carpenter]

On Thu, Jun 11, 2026 at 05:03:24AM -0400, Michael S. Tsirkin wrote:
> On Thu, Jun 11, 2026 at 10:18:51AM +0300, Dan Carpenter wrote:
> > Hi Eugenio,
> > 
> > kernel test robot noticed the following build warnings:
> > 
> > https://git-scm.com/docs/git-format-patch#_base_tree_information]
> > 
> > url:    https://github.com/intel-lab-lkp/linux/commits/Eugenio-P-rez/vduse-Add-suspend/20260610-164534
> > base:   next-20260609
> > patch link:    https://lore.kernel.org/r/20260610083452.477759-1-eperezma%40redhat.com
> > patch subject: [PATCH v3] vduse: Add suspend
> > config: arm64-randconfig-r072-20260610 (https://download.01.org/0day-ci/archive/20260611/202606111115.tKKe1qCE-lkp@xxxxxxxxx/config)
> > compiler: aarch64-linux-gcc (GCC) 8.5.0
> > smatch: v0.5.0-9185-gbcc58b9c
> > 
> > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > the same patch/commit), kindly add following tags
> > | Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > | Reported-by: Dan Carpenter <error27@xxxxxxxxx>
> > | Closes: https://lore.kernel.org/r/202606111115.tKKe1qCE-lkp@xxxxxxxxx/
> > 
> > smatch warnings:
> > drivers/vdpa/vdpa_user/vduse_dev.c:577 vduse_vq_kick() warn: inconsistent returns '&vq->kick_lock'.
> > drivers/vdpa/vdpa_user/vduse_dev.c:1302 vduse_dev_queue_irq_work() warn: inconsistent returns '&dev->rwsem'.
> > 
> > vim +577 drivers/vdpa/vdpa_user/vduse_dev.c
> > 
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  562  static void vduse_vq_kick(struct vduse_virtqueue *vq)
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  563  {
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  564  	spin_lock(&vq->kick_lock);
> >                                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  565  	if (!vq->ready)
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  566  		goto unlock;
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  567  
> > 9c4307e82fa1dc Eugenio Pérez     2026-06-10  568  	guard(rwsem_read)(&vq->dev->rwsem);
> > 9c4307e82fa1dc Eugenio Pérez     2026-06-10  569  	if (vq->dev->suspended)
> > 9c4307e82fa1dc Eugenio Pérez     2026-06-10  570  		return;
> > 
> > unlock before returning?
> > 
> > 9c4307e82fa1dc Eugenio Pérez     2026-06-10  571  
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  572  	if (vq->kickfd)
> > 3652117f854819 Christian Brauner 2023-11-22  573  		eventfd_signal(vq->kickfd);
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  574  	else
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  575  		vq->kicked = true;
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  576  unlock:
> > c8a6153b6c59d9 Xie Yongji        2021-08-31 @577  	spin_unlock(&vq->kick_lock);
> > c8a6153b6c59d9 Xie Yongji        2021-08-31  578  }
> 
> 
> I think this is fixed by:
> 
> commit e4a249d15eb2d4b28213bebb1eefaf2e6d99de0b (HEAD -> vhost, linux-next-vhost/linux-next, kernel.org/vhost, kernel.org/test)
> Author: Nathan Chancellor <nathan@xxxxxxxxxx>
> Date:   Wed Jun 10 12:16:49 2026 -0700
> 
>     vduse: Fix error around jumping over a __cleanup() variable
>     
> right?

These things haven't hit linux-next yet.  I found the email.
https://lore.kernel.org/all/20260610-vduse_vq_kick-fix-guard-usage-v1-1-0ce02c08006e@xxxxxxxxxx/

That only fixes the bug in vduse_vq_kick(), not the bug in
vduse_dev_queue_irq_work().  I don't see a fix for that
yet on lore but I may have missed it.

regards,
dan carpenter