Prev: [tip:irq/core] genirq: Clear CPU mask in affinity_hint when none is provided
Next: 2.6.33.2 kmalloc-8 slab leaks ~512 objects per second
From: Juan Quintela on 12 May 2010 05:30
"Michael S. Tsirkin" <mst(a)redhat.com> wrote:
> According to memory-barriers.txt, an smp memory barrier
> should always be paired with another smp memory barrier,
> and I quote "a lack of appropriate pairing is almost certainly an
> error".
>
> In case of vhost, failure to flush out used index
> update before looking at the interrupt disable flag
> could result in missed interrupts, resulting in
> networking hang under stress.
>
> This might happen when flags read bypasses used index write.
> So we see interrupts disabled and do not interrupt, at the
> same time guest writes flags value to enable interrupt,
> reads an old used index value, thinks that
> used ring is empty and waits for interrupt.
>
> Note: the barrier we pair with here is in
> drivers/virtio/virtio_ring.c, function
> vring_enable_cb.
>
> Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com>
> ---
>
> Dave, I think this is needed in 2.6.34, I'll send a pull
> request after doing some more testing.
>
> Rusty, Juan, could you take a look as well please?
> Thanks!

I would have prefered to put it:

void vhost_add_used_and_signal(struct vhost_dev *dev,
struct vhost_virtqueue *vq,
unsigned int head, int len)
{
vhost_add_used(vq, head, len);
>>>> smp_mb();
vhost_signal(dev, vq);
}

Because it looks strange to have a barrier as the 1st instruction of a
function. And this way it is clearer (at least to me) what we are
protecting.

But on the other hand, we would have to put a comment explainingthat all
users of vhost_signal() have to put that smp_mb() so .....

Perhaps just improving the commet stating that the corresponding barrier
is there?

> Note: the barrier we pair with here is in
> drivers/virtio/virtio_ring.c, function
> vring_enable_cb.

Good catch.

Later, Juan.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: [tip:irq/core] genirq: Clear CPU mask in affinity_hint when none is provided
Next: 2.6.33.2 kmalloc-8 slab leaks ~512 objects per second