From 15ecbe94a45ef88491ca459b26efdd02f91edb6d Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 27 Jun 2018 08:47:21 -0700 Subject: tcp: add one more quick ack after after ECN events Larry Brakmo proposal ( https://patchwork.ozlabs.org/patch/935233/ tcp: force cwnd at least 2 in tcp_cwnd_reduction) made us rethink about our recent patch removing ~16 quick acks after ECN events. tcp_enter_quickack_mode(sk, 1) makes sure one immediate ack is sent, but in the case the sender cwnd was lowered to 1, we do not want to have a delayed ack for the next packet we will receive. Fixes: 522040ea5fdd ("tcp: do not aggressively quick ack after ECN events") Signed-off-by: Eric Dumazet Reported-by: Neal Cardwell Cc: Lawrence Brakmo Acked-by: Neal Cardwell Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 355d3dffd021..045d930d01a9 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -265,7 +265,7 @@ static void __tcp_ecn_check_ce(struct sock *sk, const struct sk_buff *skb) * it is probably a retransmit. */ if (tp->ecn_flags & TCP_ECN_SEEN) - tcp_enter_quickack_mode(sk, 1); + tcp_enter_quickack_mode(sk, 2); break; case INET_ECN_CE: if (tcp_ca_needs_ecn(sk)) @@ -273,7 +273,7 @@ static void __tcp_ecn_check_ce(struct sock *sk, const struct sk_buff *skb) if (!(tp->ecn_flags & TCP_ECN_DEMAND_CWR)) { /* Better not delay acks, sender can have a very low cwnd */ - tcp_enter_quickack_mode(sk, 1); + tcp_enter_quickack_mode(sk, 2); tp->ecn_flags |= TCP_ECN_DEMAND_CWR; } tp->ecn_flags |= TCP_ECN_SEEN; -- cgit v1.2.3 From 1236f22fbae15df3736ab4a984c64c0c6ee6254c Mon Sep 17 00:00:00 2001 From: Ilpo Järvinen Date: Fri, 29 Jun 2018 13:07:53 +0300 Subject: tcp: prevent bogus FRTO undos with non-SACK flows MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If SACK is not enabled and the first cumulative ACK after the RTO retransmission covers more than the retransmitted skb, a spurious FRTO undo will trigger (assuming FRTO is enabled for that RTO). The reason is that any non-retransmitted segment acknowledged will set FLAG_ORIG_SACK_ACKED in tcp_clean_rtx_queue even if there is no indication that it would have been delivered for real (the scoreboard is not kept with TCPCB_SACKED_ACKED bits in the non-SACK case so the check for that bit won't help like it does with SACK). Having FLAG_ORIG_SACK_ACKED set results in the spurious FRTO undo in tcp_process_loss. We need to use more strict condition for non-SACK case and check that none of the cumulatively ACKed segments were retransmitted to prove that progress is due to original transmissions. Only then keep FLAG_ORIG_SACK_ACKED set, allowing FRTO undo to proceed in non-SACK case. (FLAG_ORIG_SACK_ACKED is planned to be renamed to FLAG_ORIG_PROGRESS to better indicate its purpose but to keep this change minimal, it will be done in another patch). Besides burstiness and congestion control violations, this problem can result in RTO loop: When the loss recovery is prematurely undoed, only new data will be transmitted (if available) and the next retransmission can occur only after a new RTO which in case of multiple losses (that are not for consecutive packets) requires one RTO per loss to recover. Signed-off-by: Ilpo Järvinen Tested-by: Neal Cardwell Acked-by: Neal Cardwell Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'net/ipv4/tcp_input.c') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 045d930d01a9..8e5522c6833a 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3181,6 +3181,15 @@ static int tcp_clean_rtx_queue(struct sock *sk, u32 prior_fack, if (tcp_is_reno(tp)) { tcp_remove_reno_sacks(sk, pkts_acked); + + /* If any of the cumulatively ACKed segments was + * retransmitted, non-SACK case cannot confirm that + * progress was due to original transmission due to + * lack of TCPCB_SACKED_ACKED bits even if some of + * the packets may have been never retransmitted. + */ + if (flag & FLAG_RETRANS_DATA_ACKED) + flag &= ~FLAG_ORIG_SACK_ACKED; } else { int delta; -- cgit v1.2.3