tcp: remove tp->lost_out guard to make joining diff nicer
authorIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Sun, 21 Sep 2008 04:23:49 +0000 (21:23 -0700)
committerDavid S. Miller <davem@davemloft.net>
Sun, 21 Sep 2008 04:23:49 +0000 (21:23 -0700)
The validity of the retransmit_high must then be ensured
if no L'ed skb exits!

This makes a minor change to behavior, we now have to
iterate the head to find out that the loop terminates.

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

index 2f24ecc370672fdd5211723c0f9973815e55ab9e..9f44be633ef645ac47dd72e1cf8abf61fd91ace9 100644 (file)
@@ -2034,53 +2034,54 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
        struct sk_buff *skb;
        int mib_idx;
 
+       if (!tp->lost_out)
+               tp->retransmit_high = tp->snd_una;
+
        if (tp->retransmit_skb_hint)
                skb = tp->retransmit_skb_hint;
        else
                skb = tcp_write_queue_head(sk);
 
        /* First pass: retransmit lost packets. */
-       if (tp->lost_out) {
-               tcp_for_write_queue_from(skb, sk) {
-                       __u8 sacked = TCP_SKB_CB(skb)->sacked;
+       tcp_for_write_queue_from(skb, sk) {
+               __u8 sacked = TCP_SKB_CB(skb)->sacked;
 
-                       if (skb == tcp_send_head(sk))
-                               break;
-                       /* we could do better than to assign each time */
-                       tp->retransmit_skb_hint = skb;
-
-                       /* Assume this retransmit will generate
-                        * only one packet for congestion window
-                        * calculation purposes.  This works because
-                        * tcp_retransmit_skb() will chop up the
-                        * packet to be MSS sized and all the
-                        * packet counting works out.
-                        */
-                       if (tcp_packets_in_flight(tp) >= tp->snd_cwnd)
-                               return;
-                       if (!before(TCP_SKB_CB(skb)->seq, tp->retransmit_high))
-                               break;
-                       if (sacked & (TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))
-                               continue;
+               if (skb == tcp_send_head(sk))
+                       break;
+               /* we could do better than to assign each time */
+               tp->retransmit_skb_hint = skb;
+
+               /* Assume this retransmit will generate
+                * only one packet for congestion window
+                * calculation purposes.  This works because
+                * tcp_retransmit_skb() will chop up the
+                * packet to be MSS sized and all the
+                * packet counting works out.
+                */
+               if (tcp_packets_in_flight(tp) >= tp->snd_cwnd)
+                       return;
+               if (!before(TCP_SKB_CB(skb)->seq, tp->retransmit_high))
+                       break;
+               if (sacked & (TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))
+                       continue;
 
-                       if (!(sacked & TCPCB_LOST))
-                               continue;
+               if (!(sacked & TCPCB_LOST))
+                       continue;
 
-                       if (tcp_retransmit_skb(sk, skb)) {
-                               tp->retransmit_skb_hint = NULL;
-                               return;
-                       }
-                       if (icsk->icsk_ca_state != TCP_CA_Loss)
-                               mib_idx = LINUX_MIB_TCPFASTRETRANS;
-                       else
-                               mib_idx = LINUX_MIB_TCPSLOWSTARTRETRANS;
-                       NET_INC_STATS_BH(sock_net(sk), mib_idx);
-
-                       if (skb == tcp_write_queue_head(sk))
-                               inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
-                                                         inet_csk(sk)->icsk_rto,
-                                                         TCP_RTO_MAX);
+               if (tcp_retransmit_skb(sk, skb)) {
+                       tp->retransmit_skb_hint = NULL;
+                       return;
                }
+               if (icsk->icsk_ca_state != TCP_CA_Loss)
+                       mib_idx = LINUX_MIB_TCPFASTRETRANS;
+               else
+                       mib_idx = LINUX_MIB_TCPSLOWSTARTRETRANS;
+               NET_INC_STATS_BH(sock_net(sk), mib_idx);
+
+               if (skb == tcp_write_queue_head(sk))
+                       inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
+                                                 inet_csk(sk)->icsk_rto,
+                                                 TCP_RTO_MAX);
        }
 
        /* OK, demanded retransmission is finished. */