Merge branch 'tcp-in-slow-start'

Yuchung Cheng says:

====================
tcp: fixes some congestion control corner cases

This patch series fixes corner cases of TCP congestion control.
First issue is to avoid continuing slow start when cwnd reaches ssthresh.
Second issue is incorrectly processing order of congestion state and
cwnd update when entering fast recovery or undoing cwnd.
====================

Signed-off-by: David S. Miller <[e-mail]>

include/net/tcp.h

···
 
 #define TCP_INFINITE_SSTHRESH	0x7fffffff
 
+static inline bool tcp_in_slow_start(const struct tcp_sock *tp)
+{
+	return tp->snd_cwnd < tp->snd_ssthresh;
+}
+
 static inline bool tcp_in_initial_slowstart(const struct tcp_sock *tp)
 {
 	return tp->snd_ssthresh >= TCP_INFINITE_SSTHRESH;
···
 	const struct tcp_sock *tp = tcp_sk(sk);
 
 	/* If in slow start, ensure cwnd grows to twice what was ACKed. */
-	if (tp->snd_cwnd <= tp->snd_ssthresh)
+	if (tcp_in_slow_start(tp))
 		return tp->snd_cwnd < 2 * tp->max_packets_out;
 
 	return tp->is_cwnd_limited;

net/ipv4/tcp_bic.c

···
 	if (!tcp_is_cwnd_limited(sk))
 		return;
 
-	if (tp->snd_cwnd <= tp->snd_ssthresh)
+	if (tcp_in_slow_start(tp))
 		tcp_slow_start(tp, acked);
 	else {
 		bictcp_update(ca, tp->snd_cwnd);

net/ipv4/tcp_cdg.c

···
 	u32 prior_snd_cwnd;
 	u32 incr;
 
-	if (tp->snd_cwnd < tp->snd_ssthresh && hystart_detect)
+	if (tcp_in_slow_start(tp) && hystart_detect)
 		tcp_cdg_hystart_update(sk);
 
 	if (after(ack, ca->rtt_seq) && ca->rtt.v64) {

net/ipv4/tcp_cong.c

···
  */
 u32 tcp_slow_start(struct tcp_sock *tp, u32 acked)
 {
-	u32 cwnd = tp->snd_cwnd + acked;
+	u32 cwnd = min(tp->snd_cwnd + acked, tp->snd_ssthresh);
 
-	if (cwnd > tp->snd_ssthresh)
-		cwnd = tp->snd_ssthresh + 1;
 	acked -= cwnd - tp->snd_cwnd;
 	tp->snd_cwnd = min(cwnd, tp->snd_cwnd_clamp);
 
···
 		return;
 
 	/* In "safe" area, increase. */
-	if (tp->snd_cwnd <= tp->snd_ssthresh) {
+	if (tcp_in_slow_start(tp)) {
 		acked = tcp_slow_start(tp, acked);
 		if (!acked)
 			return;

net/ipv4/tcp_cubic.c

···
 	if (!tcp_is_cwnd_limited(sk))
 		return;
 
-	if (tp->snd_cwnd <= tp->snd_ssthresh) {
+	if (tcp_in_slow_start(tp)) {
 		if (hystart && after(ack, ca->end_seq))
 			bictcp_hystart_reset(sk);
 		acked = tcp_slow_start(tp, acked);
···
 		ca->delay_min = delay;
 
 	/* hystart triggers when cwnd is larger than some threshold */
-	if (hystart && tp->snd_cwnd <= tp->snd_ssthresh &&
+	if (hystart && tcp_in_slow_start(tp) &&
 	    tp->snd_cwnd >= hystart_low_window)
 		hystart_update(sk, delay);
 }

net/ipv4/tcp_highspeed.c

···
 	if (!tcp_is_cwnd_limited(sk))
 		return;
 
-	if (tp->snd_cwnd <= tp->snd_ssthresh)
+	if (tcp_in_slow_start(tp))
 		tcp_slow_start(tp, acked);
 	else {
 		/* Update AIMD parameters.

net/ipv4/tcp_htcp.c

···
 	if (!tcp_is_cwnd_limited(sk))
 		return;
 
-	if (tp->snd_cwnd <= tp->snd_ssthresh)
+	if (tcp_in_slow_start(tp))
 		tcp_slow_start(tp, acked);
 	else {
 		/* In dangerous area, increase slowly.

net/ipv4/tcp_hybla.c

···
 
 	rho_fractions = ca->rho_3ls - (ca->rho << 3);
 
-	if (tp->snd_cwnd < tp->snd_ssthresh) {
+	if (tcp_in_slow_start(tp)) {
 		/*
 		 * slow start
 		 *      INC = 2^RHO - 1

net/ipv4/tcp_illinois.c

···
 		return;
 
 	/* In slow start */
-	if (tp->snd_cwnd <= tp->snd_ssthresh)
+	if (tcp_in_slow_start(tp))
 		tcp_slow_start(tp, acked);
 
 	else {

net/ipv4/tcp_input.c

···
 				    &sack_state);
 	acked -= tp->packets_out;
 
-	/* Advance cwnd if state allows */
-	if (tcp_may_raise_cwnd(sk, flag))
-		tcp_cong_avoid(sk, ack, acked);
-
 	if (tcp_ack_is_dubious(sk, flag)) {
 		is_dupack = !(flag & (FLAG_SND_UNA_ADVANCED | FLAG_NOT_DUP));
 		tcp_fastretrans_alert(sk, acked, prior_unsacked,
···
 	}
 	if (tp->tlp_high_seq)
 		tcp_process_tlp_ack(sk, ack, flag);
+
+	/* Advance cwnd if state allows */
+	if (tcp_may_raise_cwnd(sk, flag))
+		tcp_cong_avoid(sk, ack, acked);
 
 	if ((flag & FLAG_FORWARD_PROGRESS) || !(flag & FLAG_NOT_DUP)) {
 		struct dst_entry *dst = __sk_dst_get(sk);

net/ipv4/tcp_metrics.c

···
 				tcp_metric_set(tm, TCP_METRIC_CWND,
 					       tp->snd_cwnd);
 		}
-	} else if (tp->snd_cwnd > tp->snd_ssthresh &&
+	} else if (!tcp_in_slow_start(tp) &&
 		   icsk->icsk_ca_state == TCP_CA_Open) {
 		/* Cong. avoidance phase, cwnd is reliable. */
 		if (!tcp_metric_locked(tm, TCP_METRIC_SSTHRESH))

net/ipv4/tcp_scalable.c

···
 	if (!tcp_is_cwnd_limited(sk))
 		return;
 
-	if (tp->snd_cwnd <= tp->snd_ssthresh)
+	if (tcp_in_slow_start(tp))
 		tcp_slow_start(tp, acked);
 	else
 		tcp_cong_avoid_ai(tp, min(tp->snd_cwnd, TCP_SCALABLE_AI_CNT),

net/ipv4/tcp_vegas.c

···
 			 */
 			diff = tp->snd_cwnd * (rtt-vegas->baseRTT) / vegas->baseRTT;
 
-			if (diff > gamma && tp->snd_cwnd <= tp->snd_ssthresh) {
+			if (diff > gamma && tcp_in_slow_start(tp)) {
 				/* Going too fast. Time to slow down
 				 * and switch to congestion avoidance.
 				 */
···
 				tp->snd_cwnd = min(tp->snd_cwnd, (u32)target_cwnd+1);
 				tp->snd_ssthresh = tcp_vegas_ssthresh(tp);
 
-			} else if (tp->snd_cwnd <= tp->snd_ssthresh) {
+			} else if (tcp_in_slow_start(tp)) {
 				/* Slow start.  */
 				tcp_slow_start(tp, acked);
 			} else {
···
 		vegas->minRTT = 0x7fffffff;
 	}
 	/* Use normal slow start */
-	else if (tp->snd_cwnd <= tp->snd_ssthresh)
+	else if (tcp_in_slow_start(tp))
 		tcp_slow_start(tp, acked);
 }
 

net/ipv4/tcp_veno.c

···
 
 		veno->diff = (tp->snd_cwnd << V_PARAM_SHIFT) - target_cwnd;
 
-		if (tp->snd_cwnd <= tp->snd_ssthresh) {
+		if (tcp_in_slow_start(tp)) {
 			/* Slow start.  */
 			tcp_slow_start(tp, acked);
 		} else {