Re: [BUG] Regression on behavior of EPOLLET | EPOLLIN for AF_UNIXsockets in 3.2

[Eric Dumazet]

Le vendredi 27 janvier 2012 Ã 22:17 +0400, Glauber Costa a Ãcrit :
> On 01/27/2012 09:53 PM, Eric Dumazet wrote:
> > Le vendredi 27 janvier 2012 Ã 12:05 -0500, Nick Mathewson a Ãcrit :
> >> [1.] One line summary of the problem:
> >>
> >
> > Hi
> >
> > Probably coming from commit 0884d7aa24e15e72b3c07f7da910a13bb7df3592
> > (AF_UNIX: Fix poll blocking problem when reading from a stream socket)
> >
> > When we requeue skb because not completely eaten, we call again
> >
> > sk->sk_data_ready(sk, skb->len);
> >
> For the record, I just confirmed this to be the case.

A fix would be to change unix_poll() and not call sk_data_ready() when
skb is requeued.

if (!skb_queue_empty(&sk->sk_receive_queue))
	mask |= POLLIN | POLLRDNORM;

Might be tricky if we want to keep unix_poll() lockless, but quite
possible.

Or... not dequeue skb from sk_received_queue unless fully consumed.


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/