Hallo,
On Tue, Nov 10, 2015, at 02:11, Eric Dumazet wrote:
> On Thu, 2015-05-21 at 17:00 +0200, Hannes Frederic Sowa wrote:
>
> > +
> > +static ssize_t skb_unix_socket_splice(struct sock *sk,
> > + struct pipe_inode_info *pipe,
> > + struct splice_pipe_desc *spd)
> > +{
> > + int ret;
> > + struct unix_sock *u = unix_sk(sk);
> > +
> > + mutex_unlock(&u->readlock);
> > + ret = splice_to_pipe(pipe, spd);
> > + mutex_lock(&u->readlock);
> > +
> > + return ret;
> > +}
> > +
>
> Hi Hannes
>
> Since we release u->readlock, what prevents another thread to read() the
> same af_unix socket and consume the skb while we splice it ?
>
> TCP stack has special code to take care of this possibility.
Hm, I do see the problem, yes. I left a window open for races happening.
Without having a closer look what spontaneously comes to my mind:
increment skb->users before splice_to_pipe, re-peek the socket queue
after the splice, recheck if we have the same skb pointer in the front,
and then consume_skb the skb properly and indicating a short read.
Let me try tomorrow if that works, but I fear we can read data multiple
times with this fix, so I don't know if this is a proper fix. I have to
check the TCP code closer. Thanks for the hint!
Thanks for the report,
Hannes
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
