tcp: early retransmit

This patch implements RFC 5827 early retransmit (ER) for TCP.
It reduces DUPACK threshold (dupthresh) if outstanding packets are
less than 4 to recover losses by fast recovery instead of timeout.

While the algorithm is simple, small but frequent network reordering
makes this feature dangerous: the connection repeatedly enter
false recovery and degrade performance. Therefore we implement
a mitigation suggested in the appendix of the RFC that delays
entering fast recovery by a small interval, i.e., RTT/4. Currently
ER is conservative and is disabled for the rest of the connection
after the first reordering event. A large scale web server
experiment on the performance impact of ER is summarized in
section 6 of the paper "Proportional Rate Reduction for TCP”,
IMC 2011. http://conferences.sigcomm.org/imc/2011/docs/p155.pdf

Note that Linux has a similar feature called THIN_DUPACK. The
differences are THIN_DUPACK do not mitigate reorderings and is only
used after slow start. Currently ER is disabled if THIN_DUPACK is
enabled. I would be happy to merge THIN_DUPACK feature with ER if
people think it's a good idea.

ER is enabled by sysctl_tcp_early_retrans:
  0: Disables ER

  1: Reduce dupthresh to packets_out - 1 when outstanding packets < 4.

  2: (Default) reduce dupthresh like mode 1. In addition, delay
     entering fast recovery by RTT/4.

Note: mode 2 is implemented in the third part of this patch series.

Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 33417f8..ef32956 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -27,6 +27,7 @@
 #include <net/tcp_memcontrol.h>
 
 static int zero;
+static int two = 2;
 static int tcp_retr1_max = 255;
 static int ip_local_port_range_min[] = { 1, 1 };
 static int ip_local_port_range_max[] = { 65535, 65535 };
@@ -677,6 +678,15 @@
 		.proc_handler   = proc_dointvec
 	},
 	{
+		.procname	= "tcp_early_retrans",
+		.data		= &sysctl_tcp_early_retrans,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &two,
+	},
+	{
 		.procname	= "udp_mem",
 		.data		= &sysctl_udp_mem,
 		.maxlen		= sizeof(sysctl_udp_mem),
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 9670af3..6802c89 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -395,6 +395,7 @@
 	tp->mss_cache = TCP_MSS_DEFAULT;
 
 	tp->reordering = sysctl_tcp_reordering;
+	tcp_enable_early_retrans(tp);
 	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
 
 	sk->sk_state = TCP_CLOSE;
@@ -2495,6 +2496,8 @@
 			err = -EINVAL;
 		else
 			tp->thin_dupack = val;
+			if (tp->thin_dupack)
+				tcp_disable_early_retrans(tp);
 		break;
 
 	case TCP_REPAIR:
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index be8e09d..e042cab 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -99,6 +99,7 @@
 
 int sysctl_tcp_moderate_rcvbuf __read_mostly = 1;
 int sysctl_tcp_abc __read_mostly;
+int sysctl_tcp_early_retrans __read_mostly = 2;
 
 #define FLAG_DATA		0x01 /* Incoming frame contained data.		*/
 #define FLAG_WIN_UPDATE		0x02 /* Incoming ACK was a window update.	*/
@@ -906,6 +907,7 @@
 	if (dst_metric(dst, RTAX_REORDERING) &&
 	    tp->reordering != dst_metric(dst, RTAX_REORDERING)) {
 		tcp_disable_fack(tp);
+		tcp_disable_early_retrans(tp);
 		tp->reordering = dst_metric(dst, RTAX_REORDERING);
 	}
 
@@ -988,6 +990,9 @@
 #endif
 		tcp_disable_fack(tp);
 	}
+
+	if (metric > 0)
+		tcp_disable_early_retrans(tp);
 }
 
 /* This must be called before lost_out is incremented */
@@ -2492,6 +2497,16 @@
 	    tcp_is_sack(tp) && !tcp_send_head(sk))
 		return 1;
 
+	/* Trick#6: TCP early retransmit, per RFC5827.  To avoid spurious
+	 * retransmissions due to small network reorderings, we implement
+	 * Mitigation A.3 in the RFC and delay the retransmission for a short
+	 * interval if appropriate.
+	 */
+	if (tp->do_early_retrans && !tp->retrans_out && tp->sacked_out &&
+	    (tp->packets_out == (tp->sacked_out + 1) && tp->packets_out < 4) &&
+	    !tcp_may_send_now(sk))
+		return 1;
+
 	return 0;
 }
 
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 3cabafb..6f6a918 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -482,6 +482,7 @@
 		newtp->sacked_out = 0;
 		newtp->fackets_out = 0;
 		newtp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
+		tcp_enable_early_retrans(newtp);
 
 		/* So many TCP implementations out there (incorrectly) count the
 		 * initial SYN frame in their delayed-ACK and congestion control