Re: [PATCH 2/4] virtio_ring: switch to use BAD_RING()

[Jason Wang]

On Tue, Dec 27, 2022 at 7:36 AM Michael S. Tsirkin <mst@xxxxxxxxxx> wrote:
>
> On Mon, Dec 26, 2022 at 03:49:06PM +0800, Jason Wang wrote:
> > Switch to reuse BAD_RING() to allow common logic to be implemented in
> > BAD_RING().
> >
> > Signed-off-by: Jason Wang <jasowang@xxxxxxxxxx>
> > ---
> > Changes since V1:
> > - switch to use BAD_RING in virtio_break_device()
> > ---
> >  drivers/virtio/virtio_ring.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> > index 2e7689bb933b..5cfb2fa8abee 100644
> > --- a/drivers/virtio/virtio_ring.c
> > +++ b/drivers/virtio/virtio_ring.c
> > @@ -58,7 +58,8 @@
> >       do {                                                    \
> >               dev_err(&_vq->vq.vdev->dev,                     \
> >                       "%s:"fmt, (_vq)->vq.name, ##args);      \
> > -             (_vq)->broken = true;                           \
> > +             /* Pairs with READ_ONCE() in virtqueue_is_broken(). */ \
>
> I don't think WRITE_ONCE/READ_ONCE pair as such. Can you point
> me at documentation of such pairing?

Introduced by:

commit 60f0779862e4ab943810187752c462e85f5fa371
Author: Parav Pandit <parav@xxxxxxxxxx>
Date:   Wed Jul 21 17:26:45 2021 +0300

    virtio: Improve vq->broken access to avoid any compiler optimization

I think it might still apply here since virtqueue_is_broken() is still
put into a loop inside wait_event().

Thanks

>
> > +             WRITE_ONCE((_vq)->broken, true);                       \
> >       } while (0)
> >  #define START_USE(vq)
> >  #define END_USE(vq)
> > @@ -2237,7 +2238,7 @@ bool virtqueue_notify(struct virtqueue *_vq)
> >
> >       /* Prod other side to tell it about changes. */
> >       if (!vq->notify(_vq)) {
> > -             vq->broken = true;
> > +             BAD_RING(vq, "vq %d is broken\n", vq->vq.index);
> >               return false;
> >       }
> >       return true;
> > @@ -2786,8 +2787,7 @@ void virtio_break_device(struct virtio_device *dev)
> >       list_for_each_entry(_vq, &dev->vqs, list) {
> >               struct vring_virtqueue *vq = to_vvq(_vq);
> >
> > -             /* Pairs with READ_ONCE() in virtqueue_is_broken(). */
> > -             WRITE_ONCE(vq->broken, true);
> > +             BAD_RING(vq, "Device break vq %d", _vq->index);
> >       }
> >       spin_unlock(&dev->vqs_list_lock);
> >  }
> > --
> > 2.25.1
>