[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20101112.133620.241452814.davem@davemloft.net>
Date: Fri, 12 Nov 2010 13:36:20 -0800 (PST)
From: David Miller <davem@...emloft.net>
To: eric.dumazet@...il.com
Cc: kuznet@....inr.ac.ru, kaber@...sh.net, equinox@...c24.net,
eparis@...hat.com, hzhong@...il.com, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, pekkas@...core.fi, jmorris@...ei.org,
yoshfuji@...ux-ipv6.org, paul.moore@...com,
damian@....rwth-aachen.de
Subject: Re: a problem tcp_v4_err()
From: Eric Dumazet <eric.dumazet@...il.com>
Date: Fri, 12 Nov 2010 22:18:38 +0100
> Acked-by: Eric Dumazet <eric.dumazet@...il.com>
Thanks for reviewing, here is the final bit I committed and
I'll queue this up for whatever -stable trees need this.
--------------------
tcp: Don't change unlocked socket state in tcp_v4_err().
Alexey Kuznetsov noticed a regression introduced by
commit f1ecd5d9e7366609d640ff4040304ea197fbc618
("Revert Backoff [v3]: Revert RTO on ICMP destination unreachable")
The RTO and timer modification code added to tcp_v4_err()
doesn't check sock_owned_by_user(), which if true means we
don't have exclusive access to the socket and therefore cannot
modify it's critical state.
Just skip this new code block if sock_owned_by_user() is true
and eliminate the now superfluous sock_owned_by_user() code
block contained within.
Reported-by: Alexey Kuznetsov <kuznet@....inr.ac.ru>
Signed-off-by: David S. Miller <davem@...emloft.net>
CC: Damian Lukowski <damian@....rwth-aachen.de>
Acked-by: Eric Dumazet <eric.dumazet@...il.com>
---
net/ipv4/tcp_ipv4.c | 8 +++-----
1 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 8f8527d..69ccbc1 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -415,6 +415,9 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
!icsk->icsk_backoff)
break;
+ if (sock_owned_by_user(sk))
+ break;
+
icsk->icsk_backoff--;
inet_csk(sk)->icsk_rto = __tcp_set_rto(tp) <<
icsk->icsk_backoff;
@@ -429,11 +432,6 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
if (remaining) {
inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
remaining, TCP_RTO_MAX);
- } else if (sock_owned_by_user(sk)) {
- /* RTO revert clocked out retransmission,
- * but socket is locked. Will defer. */
- inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
- HZ/20, TCP_RTO_MAX);
} else {
/* RTO revert clocked out retransmission.
* Will retransmit now */
--
1.7.3.2
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists