net: fix sk_forward_alloc corruptions
authorEric Dumazet <eric.dumazet@gmail.com>
Sat, 29 May 2010 07:20:48 +0000 (00:20 -0700)
committerDavid S. Miller <davem@davemloft.net>
Sat, 29 May 2010 07:20:48 +0000 (00:20 -0700)
commit2903037400a26e7c0cc93ab75a7d62abfacdf485
treec29d8b3199d8fa52856e5edcd374411f962d7c11
parent7dfde179c38056b91d51e60f3d50902387f27c84
net: fix sk_forward_alloc corruptions

As David found out, sock_queue_err_skb() should be called with socket
lock hold, or we risk sk_forward_alloc corruption, since we use non
atomic operations to update this field.

This patch adds bh_lock_sock()/bh_unlock_sock() pair to three spots.
(BH already disabled)

1) skb_tstamp_tx()
2) Before calling ip_icmp_error(), in __udp4_lib_err()
3) Before calling ipv6_icmp_error(), in __udp6_lib_err()

Reported-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/core/skbuff.c
net/ipv4/udp.c
net/ipv6/udp.c