net: Remove 'unlikely' qualifier in skb_steal_sock()
authorVijay Subramanian <subramanian.vijay@gmail.com>
Sun, 24 Jun 2012 13:03:07 +0000 (13:03 +0000)
committerDavid S. Miller <davem@davemloft.net>
Mon, 25 Jun 2012 23:08:36 +0000 (16:08 -0700)
With early demux enabled by default for TCP flows, there is high chance that
skb->sk will be non-null. 'unlikely()' was removed from __inet_lookup_skb() but
maybe it can be removed from skb_steal_sock() as well.

Note: skb_steal_sock() is also called by __inet6_lookup_skb() and
__udp4_lib_lookup_skb() but they are protected by their own 'unlikely' calls.

Signed-off-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/net/sock.h

index 87b424ae750a96c08913a7ab183e68c52975afc6..21086036e348dce616d6ff4933077c4aca8035fa 100644 (file)
@@ -2154,7 +2154,7 @@ static inline void sk_change_net(struct sock *sk, struct net *net)
 
 static inline struct sock *skb_steal_sock(struct sk_buff *skb)
 {
-       if (unlikely(skb->sk)) {
+       if (skb->sk) {
                struct sock *sk = skb->sk;
 
                skb->destructor = NULL;