DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com C7D8A806CA
Date: Fri, 11 Aug 2017 00:21:16 +0300
From: "Michael S. Tsirkin" <mst@redhat.com>
To: "Richard W.M. Jones" <rjones@redhat.com>
Cc: jejb@linux.vnet.ibm.com, martin.petersen@oracle.com,
        jasowang@redhat.com, linux-scsi@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        virtualization@lists.linux-foundation.org, hch@lst.de,
        pbonzini@redhat.com
Subject: Re: [PATCH 1/2] virtio: Reduce BUG if total_sg > virtqueue size to
 WARN.
Message-ID: <20170811001447-mutt-send-email-mst@kernel.org>
References: <20170810164035.19963-1-rjones@redhat.com>
 <20170810164035.19963-2-rjones@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170810164035.19963-2-rjones@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1512
Lines: 48

On Thu, Aug 10, 2017 at 05:40:34PM +0100, Richard W.M. Jones wrote:
> If using indirect descriptors, you can make the total_sg as large as
> you want.

That would be a spec violation though, even if it happens to
work on current QEMU.

The spec says:
	A driver MUST NOT create a descriptor chain longer than the Queue Size of the device.

What prompted this patch?  Do we ever encounter this situation?

>  If not, BUG is too serious because the function later
> returns -ENOSPC.
> 
> Thanks Paolo Bonzini, Christoph Hellwig.
> 
> Signed-off-by: Richard W.M. Jones <rjones@redhat.com>
> ---
>  drivers/virtio/virtio_ring.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index 5e1b548828e6..27cbc1eab868 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -296,7 +296,6 @@ static inline int virtqueue_add(struct virtqueue *_vq,
>  	}
>  #endif
>  
> -	BUG_ON(total_sg > vq->vring.num);
>  	BUG_ON(total_sg == 0);
>  
>  	head = vq->free_head;
> @@ -305,8 +304,10 @@ static inline int virtqueue_add(struct virtqueue *_vq,
>  	 * buffers, then go indirect. FIXME: tune this threshold */
>  	if (vq->indirect && total_sg > 1 && vq->vq.num_free)
>  		desc = alloc_indirect(_vq, total_sg, gfp);
> -	else
> +	else {
>  		desc = NULL;
> +		WARN_ON_ONCE(total_sg > vq->vring.num && !vq->indirect);
> +	}
>  
>  	if (desc) {
>  		/* Use a single buffer which doesn't continue */
> -- 
> 2.13.1