[PATCH] virtio: remove overzealous BUG_ON.

[PATCH] virtio: remove overzealous BUG_ON.

[Rusty Russell]

The 'disable_cb' callback is designed as an optimization to tell the host
we don't need callbacks now.  As it is not reliable, the debug check is
overzealous: it can happen on two CPUs at the same time.  Document this.

Even if it were reliable, the virtio_net driver doesn't disable
callbacks on transmit so the START_USE/END_USE debugging reentrance
protection can be easily tripped even on UP.

Thanks to Balaji Rao for the bug report and testing.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
CC: Balaji Rao <balajirrao@gmail.com>

diff -r 0b3e64be7696 drivers/virtio/virtio_ring.c
--- a/drivers/virtio/virtio_ring.c	Mon Apr 07 07:40:08 2008 +1000
+++ b/drivers/virtio/virtio_ring.c	Mon Apr 07 13:37:03 2008 +1000
@@ -214,10 +215,7 @@ static void vring_disable_cb(struct virt
 {
 	struct vring_virtqueue *vq = to_vvq(_vq);
 
-	START_USE(vq);
-	BUG_ON(vq->vring.avail->flags & VRING_AVAIL_F_NO_INTERRUPT);
 	vq->vring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT;
-	END_USE(vq);
 }
 
 static bool vring_enable_cb(struct virtqueue *_vq)
diff -r 0b3e64be7696 include/linux/virtio.h
--- a/include/linux/virtio.h	Mon Apr 07 07:40:08 2008 +1000
+++ b/include/linux/virtio.h	Mon Apr 07 13:37:03 2008 +1000
@@ -41,6 +41,8 @@ struct virtqueue
  *	Returns NULL or the "data" token handed to add_buf.
  * @disable_cb: disable callbacks
  *	vq: the struct virtqueue we're talking about.
+ *	Note that this is not necessarily synchronous, hence unreliable and only
+ *	useful as an optimization.
  * @enable_cb: restart callbacks after disable_cb.
  *	vq: the struct virtqueue we're talking about.
  *	This re-enables callbacks; it returns "false" if there are pending
@@ -48,7 +50,8 @@ struct virtqueue
  *	checking for more work, and enabling callbacks.
  *
  * Locking rules are straightforward: the driver is responsible for
- * locking.  No two operations may be invoked simultaneously.
+ * locking.  No two operations may be invoked simultaneously, with the exception
+ * of @disable_cb.
  *
  * All operations can be called in any context.
  */

Re: [PATCH] virtio: remove overzealous BUG_ON.

[Rusty Russell]

On Monday 07 April 2008 23:50:24 Anthony Liguori wrote:
> Rusty Russell wrote:
> > The 'disable_cb' callback is designed as an optimization to tell the host
> > we don't need callbacks now.  As it is not reliable, the debug check is
> > overzealous: it can happen on two CPUs at the same time.  Document this.
> >
> > Even if it were reliable, the virtio_net driver doesn't disable
> > callbacks on transmit so the START_USE/END_USE debugging reentrance
> > protection can be easily tripped even on UP.
> >
> > Thanks to Balaji Rao for the bug report and testing.
>
> I think the printk() in start_xmit should be removed too (or at least
> rate limited).

Yep, please send patch.  I'm about to jump on a plane, so consider this a 
pre-canned ack.

Thanks,
Rusty.

Re: [PATCH] virtio: remove overzealous BUG_ON.

[Anthony Liguori]

Rusty Russell wrote:
> The 'disable_cb' callback is designed as an optimization to tell the host
> we don't need callbacks now.  As it is not reliable, the debug check is
> overzealous: it can happen on two CPUs at the same time.  Document this.
> 
> Even if it were reliable, the virtio_net driver doesn't disable
> callbacks on transmit so the START_USE/END_USE debugging reentrance
> protection can be easily tripped even on UP.
> 
> Thanks to Balaji Rao for the bug report and testing.

I think the printk() in start_xmit should be removed too (or at least 
rate limited).

Regards,

Anthony Liguori