Re: [PATCH V2 RFC 9/9] virtio_scsi: verify if queue is broken after virtqueue_get_buf()

[Rusty Russell <rusty@xxxxxxxxxxxxxxx>]

Heinz Graalfs <graalfs@xxxxxxxxxxxxxxxxxx> writes:
> If virtqueue_get_buf() returned with a NULL pointer avoid a possibly
> endless loop by checking for a broken virtqueue.
>
> Signed-off-by: Heinz Graalfs <graalfs@xxxxxxxxxxxxxxxxxx>
> ---
>  drivers/scsi/virtio_scsi.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
> index 74b88ef..45bcdb5 100644
> --- a/drivers/scsi/virtio_scsi.c
> +++ b/drivers/scsi/virtio_scsi.c
> @@ -224,7 +224,8 @@ static void virtscsi_vq_done(struct virtio_scsi *vscsi,
>  		virtqueue_disable_cb(vq);
>  		while ((buf = virtqueue_get_buf(vq, &len)) != NULL)
>  			fn(vscsi, buf);
> -	} while (!virtqueue_enable_cb(vq));
> +	} while (unlikely(!virtqueue_is_broken(vq)) &&
> +		 !virtqueue_enable_cb(vq));
>  	spin_unlock_irqrestore(&virtscsi_vq->vq_lock, flags);

unlikely(!virtqueue_is_broken(vq))?

I didn't apply this one.

I'd prefer:

	spin_lock_irqsave(&virtscsi_vq->vq_lock, flags);
	do {
		virtqueue_disable_cb(vq);
		while ((buf = virtqueue_get_buf(vq, &len)) != NULL)
			fn(vscsi, buf);

                if (unlikely(virtqueue_is_broken(vq)))
                        break;
	} while (!virtqueue_enable_cb(vq));
	spin_unlock_irqrestore(&virtscsi_vq->vq_lock, flags);

Thanks,
Rusty.
_______________________________________________
Virtualization mailing list
Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linuxfoundation.org/mailman/listinfo/virtualization