diff mbox

[2/2] virtio: correct the memory barrier in virtqueue_kick_prepare()

Message ID 20120120081708.50747.60241.stgit@amd-6168-8-1.englab.nay.redhat.com
State Not Applicable, archived
Delegated to: David Miller
Headers show

Commit Message

Jason Wang Jan. 20, 2012, 8:17 a.m. UTC
Use virtio_mb() to make sure the available index to be exposed before
checking the the avail event. Otherwise we may get stale value of
avail event in guest and never kick the host after.

Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 drivers/virtio/virtio_ring.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Michael S. Tsirkin Jan. 22, 2012, 11:47 a.m. UTC | #1
On Fri, Jan 20, 2012 at 04:17:08PM +0800, Jason Wang wrote:
> Use virtio_mb() to make sure the available index to be exposed before
> checking the the avail event. Otherwise we may get stale value of
> avail event in guest and never kick the host after.
> 
> Signed-off-by: Jason Wang <jasowang@redhat.com>

Good catch.
Note: this fixes a bug introduced by ee7cd8981e15bcb365fc762afe3fc47b8242f630.

Acked-by: Michael S. Tsirkin <mst@redhat.com>

> ---
>  drivers/virtio/virtio_ring.c |    6 +++---
>  1 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index 78428a8..07d9360 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -308,9 +308,9 @@ bool virtqueue_kick_prepare(struct virtqueue *_vq)
>  	bool needs_kick;
>  
>  	START_USE(vq);
> -	/* Descriptors and available array need to be set before we expose the
> -	 * new available array entries. */
> -	virtio_wmb(vq);
> +	/* We need expose available array entries before checking avail

Nit:
We need expose -> Need to expose

> +	 * event. */
> +	virtio_mb(vq);
>  
>  	old = vq->vring.avail->idx - vq->num_added;
>  	new = vq->vring.avail->idx;
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rusty Russell Jan. 22, 2012, 11:34 p.m. UTC | #2
On Sun, 22 Jan 2012 13:47:57 +0200, "Michael S. Tsirkin" <mst@redhat.com> wrote:
> On Fri, Jan 20, 2012 at 04:17:08PM +0800, Jason Wang wrote:
> > Use virtio_mb() to make sure the available index to be exposed before
> > checking the the avail event. Otherwise we may get stale value of
> > avail event in guest and never kick the host after.
> > 
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> 
> Good catch.
> Note: this fixes a bug introduced by ee7cd8981e15bcb365fc762afe3fc47b8242f630.

Indeed, great catch.  It replaces the missing barrier, and after some
careful thought I've convinced myself that it is necessary for both the
avail index and flag cases.

> >  	START_USE(vq);
> > -	/* Descriptors and available array need to be set before we expose the
> > -	 * new available array entries. */
> > -	virtio_wmb(vq);
> > +	/* We need expose available array entries before checking avail
> 
> Nit:
> We need expose -> Need to expose
> 
> > +	 * event. */
> > +	virtio_mb(vq);

Fixed typo and applied.

Thanks,
Rusty.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index 78428a8..07d9360 100644
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -308,9 +308,9 @@  bool virtqueue_kick_prepare(struct virtqueue *_vq)
 	bool needs_kick;
 
 	START_USE(vq);
-	/* Descriptors and available array need to be set before we expose the
-	 * new available array entries. */
-	virtio_wmb(vq);
+	/* We need expose available array entries before checking avail
+	 * event. */
+	virtio_mb(vq);
 
 	old = vq->vring.avail->idx - vq->num_added;
 	new = vq->vring.avail->idx;