Re: [PATCH net,stable] vhost: fix skb leak in handle_rx()

[Wei Xu <wexu@redhat.com>]

On Wed, Nov 29, 2017 at 01:06:28PM +0800, Jason Wang wrote:
> 
> 
> On 2017年11月29日 09:53, Jason Wang wrote:
> > 
> > 
> > On 2017年11月29日 01:17, wexu@redhat.com wrote:
> > > From: Wei Xu <wexu@redhat.com>
> > > 
> > > Matthew found a roughly 40% tcp throughput regression with commit
> > > c67df11f(vhost_net: try batch dequing from skb array) as discussed
> > > in the following thread:
> > > https://www.mail-archive.com/netdev@vger.kernel.org/msg187936.html
> > > 
> > > Eventually we figured out that it was a skb leak in handle_rx()
> > > when sending packets to the VM. This usually happens when a guest
> > > can not drain out vq as fast as vhost fills in, afterwards it sets
> > > off the traffic jam and leaks skb(s) which occurs as no headcount
> > > to send on the vq from vhost side.
> > > 
> > > This can be avoided by making sure we have got enough headcount
> > > before actually consuming a skb from the batched rx array while
> > > transmitting, which is simply done by deferring it a moment later
> > > in this patch.
> > > 
> > > Signed-off-by: Wei Xu <wexu@redhat.com>
> > > ---
> > >   drivers/vhost/net.c | 4 ++--
> > >   1 file changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
> > > index 8d626d7..e76535e 100644
> > > --- a/drivers/vhost/net.c
> > > +++ b/drivers/vhost/net.c
> > > @@ -778,8 +778,6 @@ static void handle_rx(struct vhost_net *net)
> > >           /* On error, stop handling until the next kick. */
> > >           if (unlikely(headcount < 0))
> > >               goto out;
> > > -        if (nvq->rx_array)
> > > -            msg.msg_control = vhost_net_buf_consume(&nvq->rxq);
> > >           /* On overrun, truncate and discard */
> > >           if (unlikely(headcount > UIO_MAXIOV)) {
> > 
> > You need do msg.msg_control = vhost_net_buf_consume() here too,
> > otherwise we may still get it leaked.
> > 
> > Thanks
> 
> Not a leak actually, but the packet won't be consumed and we will hit
> UIO_MAXIOV forever in this case.

I see, thanks, will make a v2.

> 
> Thanks