Messages in this thread Patch in this message | | | Date | Wed, 01 Feb 2012 13:00:36 -0800 | From | Greg KH <> | Subject | [72/89] af_unix: fix EPOLLET regression for stream sockets |
| |
3.2-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Dumazet <eric.dumazet@gmail.com>
[ Upstream commit 6f01fd6e6f6809061b56e78f1e8d143099716d70 ]
Commit 0884d7aa24 (AF_UNIX: Fix poll blocking problem when reading from a stream socket) added a regression for epoll() in Edge Triggered mode (EPOLLET)
Appropriate fix is to use skb_peek()/skb_unlink() instead of skb_dequeue(), and only call skb_unlink() when skb is fully consumed.
This remove the need to requeue a partial skb into sk_receive_queue head and the extra sk->sk_data_ready() calls that added the regression.
This is safe because once skb is given to sk_receive_queue, it is not modified by a writer, and readers are serialized by u->readlock mutex.
This also reduce number of spinlock acquisition for small reads or MSG_PEEK users so should improve overall performance.
Reported-by: Nick Mathewson <nickm@freehaven.net> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: Alexey Moiseytsev <himeraster@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> --- net/unix/af_unix.c | 19 ++++--------------- 1 file changed, 4 insertions(+), 15 deletions(-)
--- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -1915,7 +1915,7 @@ static int unix_stream_recvmsg(struct ki struct sk_buff *skb; unix_state_lock(sk); - skb = skb_dequeue(&sk->sk_receive_queue); + skb = skb_peek(&sk->sk_receive_queue); if (skb == NULL) { unix_sk(sk)->recursion_level = 0; if (copied >= target) @@ -1955,11 +1955,8 @@ static int unix_stream_recvmsg(struct ki if (check_creds) { /* Never glue messages from different writers */ if ((UNIXCB(skb).pid != siocb->scm->pid) || - (UNIXCB(skb).cred != siocb->scm->cred)) { - skb_queue_head(&sk->sk_receive_queue, skb); - sk->sk_data_ready(sk, skb->len); + (UNIXCB(skb).cred != siocb->scm->cred)) break; - } } else { /* Copy credentials */ scm_set_cred(siocb->scm, UNIXCB(skb).pid, UNIXCB(skb).cred); @@ -1974,8 +1971,6 @@ static int unix_stream_recvmsg(struct ki chunk = min_t(unsigned int, skb->len, size); if (memcpy_toiovec(msg->msg_iov, skb->data, chunk)) { - skb_queue_head(&sk->sk_receive_queue, skb); - sk->sk_data_ready(sk, skb->len); if (copied == 0) copied = -EFAULT; break; @@ -1990,13 +1985,10 @@ static int unix_stream_recvmsg(struct ki if (UNIXCB(skb).fp) unix_detach_fds(siocb->scm, skb); - /* put the skb back if we didn't use it up.. */ - if (skb->len) { - skb_queue_head(&sk->sk_receive_queue, skb); - sk->sk_data_ready(sk, skb->len); + if (skb->len) break; - } + skb_unlink(skb, &sk->sk_receive_queue); consume_skb(skb); if (siocb->scm->fp) @@ -2007,9 +1999,6 @@ static int unix_stream_recvmsg(struct ki if (UNIXCB(skb).fp) siocb->scm->fp = scm_fp_dup(UNIXCB(skb).fp); - /* put message back and return */ - skb_queue_head(&sk->sk_receive_queue, skb); - sk->sk_data_ready(sk, skb->len); break; } } while (size);
| |