ipv6: introduce tcp_v6_iif()
authorEric Dumazet <edumazet@google.com>
Fri, 17 Oct 2014 16:17:20 +0000 (09:17 -0700)
committerDavid S. Miller <davem@davemloft.net>
Sat, 18 Oct 2014 03:48:07 +0000 (23:48 -0400)
commit870c3151382c980590d4d609babf3b0243e7db93
tree5d472dc0982bfda41f56c08ff98c46d9b0f4afd8
parent70b33fb0ddec827cbbd14cdc664fc27b2ef4a6b6
ipv6: introduce tcp_v6_iif()

Commit 971f10eca186 ("tcp: better TCP_SKB_CB layout to reduce cache line
misses") added a regression for SO_BINDTODEVICE on IPv6.

This is because we still use inet6_iif() which expects that IP6 control
block is still at the beginning of skb->cb[]

This patch adds tcp_v6_iif() helper and uses it where necessary.

Because __inet6_lookup_skb() is used by TCP and DCCP, we add an iif
parameter to it.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Fixes: 971f10eca186 ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
Acked-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/net/inet6_hashtables.h
include/net/tcp.h
net/dccp/ipv6.c
net/ipv6/syncookies.c
net/ipv6/tcp_ipv6.c