[TCP]: Appropriate Byte Count support

This is an updated version of the RFC3465 ABC patch originally
for Linux 2.6.11-rc4 by Yee-Ting Li. ABC is a way of counting
bytes ack'd rather than packets when updating congestion control.

The orignal ABC described in the RFC applied to a Reno style
algorithm. For advanced congestion control there is little
change after leaving slow start.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Documentation/networking/ip-sysctl.txt

@@ -78,6 +78,11 @@ inet_peer_gc_maxtime - INTEGER
 
 TCP variables: 
 
+tcp_abc - INTEGER
+	Controls Appropriate Byte Count defined in RFC3465. If set to
+	0 then does congestion avoid once per ack. 1 is conservative
+	value, and 2 is more agressive.
+
 tcp_syn_retries - INTEGER
 	Number of times initial SYNs for an active TCP connection attempt
 	will be retransmitted. Should not be higher than 255. Default value

include/linux/sysctl.h

@@ -390,6 +390,7 @@ enum
 	NET_TCP_BIC_BETA=108,
 	NET_IPV4_ICMP_ERRORS_USE_INBOUND_IFADDR=109,
 	NET_TCP_CONG_CONTROL=110,
+	NET_TCP_ABC=111,
 };
 
 enum {

include/linux/tcp.h

@@ -326,6 +326,7 @@ struct tcp_sock {
 	__u32	snd_up;		/* Urgent pointer		*/
 
 	__u32	total_retrans;	/* Total retransmits for entire connection */
+	__u32	bytes_acked;	/* Appropriate Byte Counting - RFC3465 */
 
 	unsigned int		keepalive_time;	  /* time before keep alive takes place */
 	unsigned int		keepalive_intvl;  /* time interval between keep alive probes */

include/net/tcp.h

@@ -218,6 +218,7 @@ extern int sysctl_tcp_low_latency;
 extern int sysctl_tcp_nometrics_save;
 extern int sysctl_tcp_moderate_rcvbuf;
 extern int sysctl_tcp_tso_win_divisor;
+extern int sysctl_tcp_abc;
 
 extern atomic_t tcp_memory_allocated;
 extern atomic_t tcp_sockets_allocated;
@@ -770,6 +771,23 @@ static inline __u32 tcp_current_ssthresh(const struct sock *sk)
  */
 static inline void tcp_slow_start(struct tcp_sock *tp)
 {
+	if (sysctl_tcp_abc) {
+		/* RFC3465: Slow Start
+		 * TCP sender SHOULD increase cwnd by the number of
+		 * previously unacknowledged bytes ACKed by each incoming
+		 * acknowledgment, provided the increase is not more than L
+		 */
+		if (tp->bytes_acked < tp->mss_cache)
+			return;
+
+		/* We MAY increase by 2 if discovered delayed ack */
+		if (sysctl_tcp_abc > 1 && tp->bytes_acked > 2*tp->mss_cache) {
+			if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+				tp->snd_cwnd++;
+		}
+	}
+	tp->bytes_acked = 0;
+
 	if (tp->snd_cwnd < tp->snd_cwnd_clamp)
 		tp->snd_cwnd++;
 }
@@ -804,6 +822,7 @@ static inline void tcp_enter_cwr(struct sock *sk)
 	struct tcp_sock *tp = tcp_sk(sk);
 
 	tp->prior_ssthresh = 0;
+	tp->bytes_acked = 0;
 	if (inet_csk(sk)->icsk_ca_state < TCP_CA_CWR) {
 		__tcp_enter_cwr(sk);
 		tcp_set_ca_state(sk, TCP_CA_CWR);

net/ipv4/sysctl_net_ipv4.c

@@ -645,6 +645,14 @@ ctl_table ipv4_table[] = {
 		.proc_handler	= &proc_tcp_congestion_control,
 		.strategy	= &sysctl_tcp_congestion_control,
 	},
+	{
+		.ctl_name	= NET_TCP_ABC,
+		.procname	= "tcp_abc",
+		.data		= &sysctl_tcp_abc,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
+	},
 
 	{ .ctl_name = 0 }
 };

net/ipv4/tcp.c

@@ -1669,6 +1669,7 @@ int tcp_disconnect(struct sock *sk, int flags)
 	tp->packets_out = 0;
 	tp->snd_ssthresh = 0x7fffffff;
 	tp->snd_cwnd_cnt = 0;
+	tp->bytes_acked = 0;
 	tcp_set_ca_state(sk, TCP_CA_Open);
 	tcp_clear_retrans(tp);
 	inet_csk_delack_init(sk);

net/ipv4/tcp_cong.c

@@ -192,17 +192,26 @@ void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 rtt, u32 in_flight,
 	/* In "safe" area, increase. */
         if (tp->snd_cwnd <= tp->snd_ssthresh)
 		tcp_slow_start(tp);
-	else {
-		/* In dangerous area, increase slowly.
-		 * In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd
-		 */
-		if (tp->snd_cwnd_cnt >= tp->snd_cwnd) {
-			if (tp->snd_cwnd < tp->snd_cwnd_clamp)
-				tp->snd_cwnd++;
-			tp->snd_cwnd_cnt = 0;
-		} else
-			tp->snd_cwnd_cnt++;
-	}
+
+ 	/* In dangerous area, increase slowly. */
+	else if (sysctl_tcp_abc) {
+ 		/* RFC3465: Apppriate Byte Count
+ 		 * increase once for each full cwnd acked
+ 		 */
+ 		if (tp->bytes_acked >= tp->snd_cwnd*tp->mss_cache) {
+ 			tp->bytes_acked -= tp->snd_cwnd*tp->mss_cache;
+ 			if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ 				tp->snd_cwnd++;
+ 		}
+ 	} else {
+ 		/* In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd */
+ 		if (tp->snd_cwnd_cnt >= tp->snd_cwnd) {
+ 			if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ 				tp->snd_cwnd++;
+ 			tp->snd_cwnd_cnt = 0;
+ 		} else
+ 			tp->snd_cwnd_cnt++;
+ 	}
 }
 EXPORT_SYMBOL_GPL(tcp_reno_cong_avoid);
 

net/ipv4/tcp_input.c

@@ -89,6 +89,7 @@ int sysctl_tcp_frto;
 int sysctl_tcp_nometrics_save;
 
 int sysctl_tcp_moderate_rcvbuf = 1;
+int sysctl_tcp_abc = 1;
 
 #define FLAG_DATA		0x01 /* Incoming frame contained data.		*/
 #define FLAG_WIN_UPDATE		0x02 /* Incoming ACK was a window update.	*/
@@ -1247,6 +1248,7 @@ void tcp_enter_loss(struct sock *sk, int how)
 	tp->snd_cwnd_cnt   = 0;
 	tp->snd_cwnd_stamp = tcp_time_stamp;
 
+	tp->bytes_acked = 0;
 	tcp_clear_retrans(tp);
 
 	/* Push undo marker, if it was plain RTO and nothing
@@ -1904,6 +1906,7 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una,
 			TCP_ECN_queue_cwr(tp);
 		}
 
+		tp->bytes_acked = 0;
 		tp->snd_cwnd_cnt = 0;
 		tcp_set_ca_state(sk, TCP_CA_Recovery);
 	}
@@ -2310,6 +2313,9 @@ static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
 	if (before(ack, prior_snd_una))
 		goto old_ack;
 
+	if (sysctl_tcp_abc && icsk->icsk_ca_state < TCP_CA_CWR)
+		tp->bytes_acked += ack - prior_snd_una;
+
 	if (!(flag&FLAG_SLOWPATH) && after(ack, prior_snd_una)) {
 		/* Window is constant, pure forward advance.
 		 * No more checks are required.
@@ -4370,6 +4376,7 @@ discard:
 
 EXPORT_SYMBOL(sysctl_tcp_ecn);
 EXPORT_SYMBOL(sysctl_tcp_reordering);
+EXPORT_SYMBOL(sysctl_tcp_abc);
 EXPORT_SYMBOL(tcp_parse_options);
 EXPORT_SYMBOL(tcp_rcv_established);
 EXPORT_SYMBOL(tcp_rcv_state_process);

net/ipv4/tcp_minisocks.c

@@ -380,6 +380,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
 		 */
 		newtp->snd_cwnd = 2;
 		newtp->snd_cwnd_cnt = 0;
+		newtp->bytes_acked = 0;
 
 		newtp->frto_counter = 0;
 		newtp->frto_highmark = 0;