[TCP]: NAGLE_PUSH seems to be a wrong way around
authorIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Wed, 5 Dec 2007 10:25:32 +0000 (02:25 -0800)
committerDavid S. Miller <davem@sunset.davemloft.net>
Wed, 5 Dec 2007 13:37:31 +0000 (05:37 -0800)
The comment in tcp_nagle_test suggests that. This bug is very
very old, even 2.4.0 seems to have it.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp_output.c

index e5130a7fe181865f70476f402007d941b37af707..f4c1eef89af0b3ac1e63ed91ecf3725a761d6949 100644 (file)
@@ -1162,8 +1162,7 @@ int tcp_may_send_now(struct sock *sk)
        return (skb &&
                tcp_snd_test(sk, skb, tcp_current_mss(sk, 1),
                             (tcp_skb_is_last(sk, skb) ?
-                             TCP_NAGLE_PUSH :
-                             tp->nonagle)));
+                             tp->nonagle : TCP_NAGLE_PUSH)));
 }
 
 /* Trim TSO SKB to LEN bytes, put the remaining data into a new packet