lkml.org 
[lkml]   [2013]   [Mar]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
SubjectRe: BUG: IPv4: Attempt to release TCP socket in state 1
From
Date
On Sun, 2013-03-17 at 09:52 -0700, Eric Dumazet wrote:

>
> We perform a sock_hold() somewhere while the socket is already dead.

Oh well, thats not exactly that.

When processing ICMP messages we perform a full inet[6] lookup and can
find a LISTEN socket.

Please try the following fix :

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index b7ab868..3d250c6 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -407,7 +407,8 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
if (code > NR_ICMP_UNREACH)
goto out;

- if (code == ICMP_FRAG_NEEDED) { /* PMTU discovery (RFC1191) */
+ if (sk->sk_state != TCP_LISTEN &&
+ code == ICMP_FRAG_NEEDED) { /* PMTU discovery (RFC1191) */
tp->mtu_info = info;
if (!sock_owned_by_user(sk)) {
tcp_v4_mtu_reduced(sk);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 9b64600..c12b3ff 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -388,7 +388,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
dst->ops->redirect(dst, sk, skb);
}

- if (type == ICMPV6_PKT_TOOBIG) {
+ if (type == ICMPV6_PKT_TOOBIG && sk->sk_state != TCP_LISTEN) {
tp->mtu_info = ntohl(info);
if (!sock_owned_by_user(sk))
tcp_v6_mtu_reduced(sk);




\
 
 \ /
  Last update: 2013-03-17 20:21    [W:0.053 / U:0.068 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site