[TCP]: Move LOSTRETRANS MIB outside !(L|S) check
authorIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Sat, 1 Dec 2007 22:47:56 +0000 (00:47 +0200)
committerDavid S. Miller <davem@davemloft.net>
Mon, 28 Jan 2008 22:55:39 +0000 (14:55 -0800)
Usually those skbs will have L set, not counting them as lost
retransmissions is misleading.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp_input.c

index 33d284e3f7f5680f9edde1dfd88039af3234b648..6ca77f8bceea2cc3b8ecf6d061a7c56391576d8c 100644 (file)
@@ -1160,8 +1160,8 @@ static int tcp_mark_lost_retrans(struct sock *sk)
                                tp->lost_out += tcp_skb_pcount(skb);
                                TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                                flag |= FLAG_DATA_SACKED;
-                               NET_INC_STATS_BH(LINUX_MIB_TCPLOSTRETRANSMIT);
                        }
+                       NET_INC_STATS_BH(LINUX_MIB_TCPLOSTRETRANSMIT);
                } else {
                        if (before(ack_seq, new_low_seq))
                                new_low_seq = ack_seq;