From e73ebb0881ea5534ce606c1d71b4ac44db5c6930 Mon Sep 17 00:00:00 2001 From: Neal Cardwell Date: Wed, 28 Jan 2015 20:01:35 -0500 Subject: tcp: stretch ACK fixes prep LRO, GRO, delayed ACKs, and middleboxes can cause "stretch ACKs" that cover more than the RFC-specified maximum of 2 packets. These stretch ACKs can cause serious performance shortfalls in common congestion control algorithms that were designed and tuned years ago with receiver hosts that were not using LRO or GRO, and were instead politely ACKing every other packet. This patch series fixes Reno and CUBIC to handle stretch ACKs. This patch prepares for the upcoming stretch ACK bug fix patches. It adds an "acked" parameter to tcp_cong_avoid_ai() to allow for future fixes to tcp_cong_avoid_ai() to correctly handle stretch ACKs, and changes all congestion control algorithms to pass in 1 for the ACKed count. It also changes tcp_slow_start() to return the number of packet ACK "credits" that were not processed in slow start mode, and can be processed by the congestion control module in additive increase mode. In future patches we will fix tcp_cong_avoid_ai() to handle stretch ACKs, and fix Reno and CUBIC handling of stretch ACKs in slow start and additive increase mode. Reported-by: Eyal Perry Signed-off-by: Neal Cardwell Signed-off-by: Yuchung Cheng Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/tcp_cong.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'net/ipv4/tcp_cong.c') diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c index 27ead0dd16bc..6826017c12d1 100644 --- a/net/ipv4/tcp_cong.c +++ b/net/ipv4/tcp_cong.c @@ -291,25 +291,28 @@ int tcp_set_congestion_control(struct sock *sk, const char *name) * ABC caps N to 2. Slow start exits when cwnd grows over ssthresh and * returns the leftover acks to adjust cwnd in congestion avoidance mode. */ -void tcp_slow_start(struct tcp_sock *tp, u32 acked) +u32 tcp_slow_start(struct tcp_sock *tp, u32 acked) { u32 cwnd = tp->snd_cwnd + acked; 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 acked; } EXPORT_SYMBOL_GPL(tcp_slow_start); /* In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd (or alternative w) */ -void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w) +void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w, u32 acked) { if (tp->snd_cwnd_cnt >= w) { if (tp->snd_cwnd < tp->snd_cwnd_clamp) tp->snd_cwnd++; tp->snd_cwnd_cnt = 0; } else { - tp->snd_cwnd_cnt++; + tp->snd_cwnd_cnt += acked; } } EXPORT_SYMBOL_GPL(tcp_cong_avoid_ai); @@ -333,7 +336,7 @@ void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 acked) tcp_slow_start(tp, acked); /* In dangerous area, increase slowly. */ else - tcp_cong_avoid_ai(tp, tp->snd_cwnd); + tcp_cong_avoid_ai(tp, tp->snd_cwnd, 1); } EXPORT_SYMBOL_GPL(tcp_reno_cong_avoid); -- cgit v1.2.3-58-ga151 From 814d488c61260521b1b3cc97063700a5a6667c8f Mon Sep 17 00:00:00 2001 From: Neal Cardwell Date: Wed, 28 Jan 2015 20:01:36 -0500 Subject: tcp: fix the timid additive increase on stretch ACKs tcp_cong_avoid_ai() was too timid (snd_cwnd increased too slowly) on "stretch ACKs" -- cases where the receiver ACKed more than 1 packet in a single ACK. For example, suppose w is 10 and we get a stretch ACK for 20 packets, so acked is 20. We ought to increase snd_cwnd by 2 (since acked/w = 20/10 = 2), but instead we were only increasing cwnd by 1. This patch fixes that behavior. Reported-by: Eyal Perry Signed-off-by: Neal Cardwell Signed-off-by: Yuchung Cheng Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/tcp_cong.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) (limited to 'net/ipv4/tcp_cong.c') diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c index 6826017c12d1..faaee5338bea 100644 --- a/net/ipv4/tcp_cong.c +++ b/net/ipv4/tcp_cong.c @@ -304,16 +304,19 @@ u32 tcp_slow_start(struct tcp_sock *tp, u32 acked) } EXPORT_SYMBOL_GPL(tcp_slow_start); -/* In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd (or alternative w) */ +/* In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd (or alternative w), + * for every packet that was ACKed. + */ void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w, u32 acked) { + tp->snd_cwnd_cnt += acked; if (tp->snd_cwnd_cnt >= w) { - if (tp->snd_cwnd < tp->snd_cwnd_clamp) - tp->snd_cwnd++; - tp->snd_cwnd_cnt = 0; - } else { - tp->snd_cwnd_cnt += acked; + u32 delta = tp->snd_cwnd_cnt / w; + + tp->snd_cwnd_cnt -= delta * w; + tp->snd_cwnd += delta; } + tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_cwnd_clamp); } EXPORT_SYMBOL_GPL(tcp_cong_avoid_ai); -- cgit v1.2.3-58-ga151 From c22bdca94782f05b9337d8548bde51b2f38ef17f Mon Sep 17 00:00:00 2001 From: Neal Cardwell Date: Wed, 28 Jan 2015 20:01:37 -0500 Subject: tcp: fix stretch ACK bugs in Reno Change Reno to properly handle stretch ACKs in additive increase mode by passing in the count of ACKed packets to tcp_cong_avoid_ai(). In addition, if snd_cwnd crosses snd_ssthresh during slow start processing, and we then exit slow start mode, we need to carry over any remaining "credit" for packets ACKed and apply that to additive increase by passing this remaining "acked" count to tcp_cong_avoid_ai(). Reported-by: Eyal Perry Signed-off-by: Neal Cardwell Signed-off-by: Yuchung Cheng Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/tcp_cong.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'net/ipv4/tcp_cong.c') diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c index faaee5338bea..8670e68e2ce6 100644 --- a/net/ipv4/tcp_cong.c +++ b/net/ipv4/tcp_cong.c @@ -335,11 +335,13 @@ void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 acked) return; /* In "safe" area, increase. */ - if (tp->snd_cwnd <= tp->snd_ssthresh) - tcp_slow_start(tp, acked); + if (tp->snd_cwnd <= tp->snd_ssthresh) { + acked = tcp_slow_start(tp, acked); + if (!acked) + return; + } /* In dangerous area, increase slowly. */ - else - tcp_cong_avoid_ai(tp, tp->snd_cwnd, 1); + tcp_cong_avoid_ai(tp, tp->snd_cwnd, acked); } EXPORT_SYMBOL_GPL(tcp_reno_cong_avoid); -- cgit v1.2.3-58-ga151