tcp: fix possible socket refcount problem for ipv6
Julian Anastasov [Wed, 5 Sep 2012 10:53:18 +0000 (10:53 +0000)]
commit 144d56e91044181ec0ef67aeca91e9a8b5718348
("tcp: fix possible socket refcount problem") is missing
the IPv6 part. As tcp_release_cb is shared by both protocols
we should hold sock reference for the TCP_MTU_REDUCED_DEFERRED
bit.

Signed-off-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv6/tcp_ipv6.c

index a3e60cc..acd32e3 100644 (file)
@@ -403,8 +403,9 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
                tp->mtu_info = ntohl(info);
                if (!sock_owned_by_user(sk))
                        tcp_v6_mtu_reduced(sk);
-               else
-                       set_bit(TCP_MTU_REDUCED_DEFERRED, &tp->tsq_flags);
+               else if (!test_and_set_bit(TCP_MTU_REDUCED_DEFERRED,
+                                          &tp->tsq_flags))
+                       sock_hold(sk);
                goto out;
        }