tcp: convert retransmit_cnt_hint to seqno

Main benefit in this is that we can then freely point
the retransmit_skb_hint to anywhere we want to because
there's no longer need to know what would be the count
changes involve, and since this is really used only as a
terminator, unnecessary work is one time walk at most,
and if some retransmissions are necessary after that
point later on, the walk is not full waste of time
anyway.

Since retransmit_high must be kept valid, all lost
markers must ensure that.

Now I also have learned how those "holes" in the
rexmittable skbs can appear, mtu probe does them. So
I removed the misleading comment as well.

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

4 files changed, 30 insertions, 33 deletions

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 2e2557388e36..d7637c4b2840 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -358,7 +358,7 @@ struct tcp_sock {
                                          */
 
         int     lost_cnt_hint;
-        int     retransmit_cnt_hint;
+        u32     retransmit_high;        /* L-bits may be on up to this seqno */
 
         u32     lost_retrans_low;       /* Sent seq after any rxmit (lowest) */
 
diff --git a/include/net/tcp.h b/include/net/tcp.h
index b71676326950..d0e90c50722b 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -472,6 +472,8 @@ extern void tcp_send_delayed_ack(struct sock *sk);
 
 /* tcp_input.c */
 extern void tcp_cwnd_application_limited(struct sock *sk);
+extern void tcp_skb_mark_lost_uncond_verify(struct tcp_sock *tp,
+                                            struct sk_buff *skb);
 
 /* tcp_timer.c */
 extern void tcp_init_xmit_timers(struct sock *);
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 12512336dbd8..d271cc825005 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -979,17 +979,17 @@ static void tcp_update_reordering(struct sock *sk, const int metric,
         }
 }
 
-/* RFC: This is from the original, I doubt that this is necessary at all:
- * clear xmit_retrans hint if seq of this skb is beyond hint. How could we
- * retransmitted past LOST markings in the first place? I'm not fully sure
- * about undo and end of connection cases, which can cause R without L?
- */
+/* This must be called before lost_out is incremented */
 static void tcp_verify_retransmit_hint(struct tcp_sock *tp, struct sk_buff *skb)
 {
-        if ((tp->retransmit_skb_hint != NULL) &&
+        if ((tp->retransmit_skb_hint == NULL) ||
             before(TCP_SKB_CB(skb)->seq,
                    TCP_SKB_CB(tp->retransmit_skb_hint)->seq))
-                tp->retransmit_skb_hint = NULL;
+                tp->retransmit_skb_hint = skb;
+
+        if (!tp->lost_out ||
+            after(TCP_SKB_CB(skb)->end_seq, tp->retransmit_high))
+                tp->retransmit_high = TCP_SKB_CB(skb)->end_seq;
 }
 
 static void tcp_skb_mark_lost(struct tcp_sock *tp, struct sk_buff *skb)
@@ -1002,6 +1002,16 @@ static void tcp_skb_mark_lost(struct tcp_sock *tp, struct sk_buff *skb)
         }
 }
 
+void tcp_skb_mark_lost_uncond_verify(struct tcp_sock *tp, struct sk_buff *skb)
+{
+        tcp_verify_retransmit_hint(tp, skb);
+
+        if (!(TCP_SKB_CB(skb)->sacked & (TCPCB_LOST|TCPCB_SACKED_ACKED))) {
+                tp->lost_out += tcp_skb_pcount(skb);
+                TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
+        }
+}
+
 /* This procedure tags the retransmission queue when SACKs arrive.
  *
  * We have three tag bits: SACKED(S), RETRANS(R) and LOST(L).
@@ -1178,13 +1188,7 @@ static void tcp_mark_lost_retrans(struct sock *sk)
                         TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
                         tp->retrans_out -= tcp_skb_pcount(skb);
 
-                        /* clear lost hint */
-                        tp->retransmit_skb_hint = NULL;
-
-                        if (!(TCP_SKB_CB(skb)->sacked & (TCPCB_LOST|TCPCB_SACKED_ACKED))) {
-                                tp->lost_out += tcp_skb_pcount(skb);
-                                TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
-                        }
+                        tcp_skb_mark_lost_uncond_verify(tp, skb);
                         NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPLOSTRETRANSMIT);
                 } else {
                         if (before(ack_seq, new_low_seq))
@@ -1890,6 +1894,7 @@ static void tcp_enter_frto_loss(struct sock *sk, int allowed_segments, int flag)
                 if (!(TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) {
                         TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                         tp->lost_out += tcp_skb_pcount(skb);
+                        tp->retransmit_high = TCP_SKB_CB(skb)->end_seq;
                 }
         }
         tcp_verify_left_out(tp);
@@ -1974,6 +1979,7 @@ void tcp_enter_loss(struct sock *sk, int how)
                         TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_ACKED;
                         TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
                         tp->lost_out += tcp_skb_pcount(skb);
+                        tp->retransmit_high = TCP_SKB_CB(skb)->end_seq;
                 }
         }
         tcp_verify_left_out(tp);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 11490958a096..cfae61b40c44 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1838,7 +1838,7 @@ void tcp_simple_retransmit(struct sock *sk)
         struct tcp_sock *tp = tcp_sk(sk);
         struct sk_buff *skb;
         unsigned int mss = tcp_current_mss(sk, 0);
-        int lost = 0;
+        u32 prior_lost = tp->lost_out;
 
         tcp_for_write_queue(skb, sk) {
                 if (skb == tcp_send_head(sk))
@@ -1849,17 +1849,13 @@ void tcp_simple_retransmit(struct sock *sk)
                                 TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
                                 tp->retrans_out -= tcp_skb_pcount(skb);
                         }
-                        if (!(TCP_SKB_CB(skb)->sacked & TCPCB_LOST)) {
-                                TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
-                                tp->lost_out += tcp_skb_pcount(skb);
-                                lost = 1;
-                        }
+                        tcp_skb_mark_lost_uncond_verify(tp, skb);
                 }
         }
 
         tcp_clear_all_retrans_hints(tp);
 
-        if (!lost)
+        if (prior_lost == tp->lost_out)
                 return;
 
         if (tcp_is_reno(tp))
@@ -2009,15 +2005,11 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
         const struct inet_connection_sock *icsk = inet_csk(sk);
         struct tcp_sock *tp = tcp_sk(sk);
         struct sk_buff *skb;
-        int packet_cnt;
 
-        if (tp->retransmit_skb_hint) {
+        if (tp->retransmit_skb_hint)
                 skb = tp->retransmit_skb_hint;
-                packet_cnt = tp->retransmit_cnt_hint;
-        } else {
+        else
                 skb = tcp_write_queue_head(sk);
-                packet_cnt = 0;
-        }
 
         /* First pass: retransmit lost packets. */
         if (tp->lost_out) {
@@ -2028,7 +2020,6 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
                                 break;
                         /* we could do better than to assign each time */
                         tp->retransmit_skb_hint = skb;
-                        tp->retransmit_cnt_hint = packet_cnt;
 
                         /* Assume this retransmit will generate
                          * only one packet for congestion window
@@ -2039,6 +2030,8 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
                          */
                         if (tcp_packets_in_flight(tp) >= tp->snd_cwnd)
                                 return;
+                        if (!before(TCP_SKB_CB(skb)->seq, tp->retransmit_high))
+                                break;
 
                         if (sacked & TCPCB_LOST) {
                                 if (!(sacked & (TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))) {
@@ -2059,10 +2052,6 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
                                                                           inet_csk(sk)->icsk_rto,
                                                                           TCP_RTO_MAX);
                                 }
-
-                                packet_cnt += tcp_skb_pcount(skb);
-                                if (packet_cnt >= tp->lost_out)
-                                        break;
                         }
                 }
         }