From: Yuchung Cheng ycheng@google.com
[ Upstream commit 76be93fc0702322179bb0ea87295d820ee46ad14 ]
Previously TLP may send multiple probes of new data in one flight. This happens when the sender is cwnd limited. After the initial TLP containing new data is sent, the sender receives another ACK that acks partial inflight. It may re-arm another TLP timer to send more, if no further ACK returns before the next TLP timeout (PTO) expires. The sender may send in theory a large amount of TLP until send queue is depleted. This only happens if the sender sees such irregular uncommon ACK pattern. But it is generally undesirable behavior during congestion especially.
The original TLP design restrict only one TLP probe per inflight as published in "Reducing Web Latency: the Virtue of Gentle Aggression", SIGCOMM 2013. This patch changes TLP to send at most one probe per inflight.
Note that if the sender is app-limited, TLP retransmits old data and did not have this issue.
Signed-off-by: Yuchung Cheng ycheng@google.com Signed-off-by: Neal Cardwell ncardwell@google.com Signed-off-by: Eric Dumazet edumazet@google.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- include/linux/tcp.h | 4 +++- net/ipv4/tcp_input.c | 11 ++++++----- net/ipv4/tcp_output.c | 13 ++++++++----- 3 files changed, 17 insertions(+), 11 deletions(-)
--- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -216,6 +216,8 @@ struct tcp_sock { } rack; u16 advmss; /* Advertised MSS */ u8 compressed_ack; + u8 tlp_retrans:1, /* TLP is a retransmission */ + unused_1:7; u32 chrono_start; /* Start time in jiffies of a TCP chrono */ u32 chrono_stat[3]; /* Time in jiffies for chrono_stat stats */ u8 chrono_type:2, /* current chronograph type */ @@ -238,7 +240,7 @@ struct tcp_sock { save_syn:1, /* Save headers of SYN packet */ is_cwnd_limited:1,/* forward progress limited by snd_cwnd? */ syn_smc:1; /* SYN includes SMC */ - u32 tlp_high_seq; /* snd_nxt at the time of TLP retransmit. */ + u32 tlp_high_seq; /* snd_nxt at the time of TLP */
u32 tcp_tx_delay; /* delay (in usec) added to TX packets */ u64 tcp_wstamp_ns; /* departure time for next sent data packet */ --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3505,10 +3505,8 @@ static void tcp_replace_ts_recent(struct } }
-/* This routine deals with acks during a TLP episode. - * We mark the end of a TLP episode on receiving TLP dupack or when - * ack is after tlp_high_seq. - * Ref: loss detection algorithm in draft-dukkipati-tcpm-tcp-loss-probe. +/* This routine deals with acks during a TLP episode and ends an episode by + * resetting tlp_high_seq. Ref: TLP algorithm in draft-ietf-tcpm-rack */ static void tcp_process_tlp_ack(struct sock *sk, u32 ack, int flag) { @@ -3517,7 +3515,10 @@ static void tcp_process_tlp_ack(struct s if (before(ack, tp->tlp_high_seq)) return;
- if (flag & FLAG_DSACKING_ACK) { + if (!tp->tlp_retrans) { + /* TLP of new data has been acknowledged */ + tp->tlp_high_seq = 0; + } else if (flag & FLAG_DSACKING_ACK) { /* This DSACK means original and TLP probe arrived; no loss */ tp->tlp_high_seq = 0; } else if (after(ack, tp->tlp_high_seq)) { --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2564,6 +2564,11 @@ void tcp_send_loss_probe(struct sock *sk int pcount; int mss = tcp_current_mss(sk);
+ /* At most one outstanding TLP */ + if (tp->tlp_high_seq) + goto rearm_timer; + + tp->tlp_retrans = 0; skb = tcp_send_head(sk); if (skb && tcp_snd_wnd_test(tp, skb, mss)) { pcount = tp->packets_out; @@ -2581,10 +2586,6 @@ void tcp_send_loss_probe(struct sock *sk return; }
- /* At most one outstanding TLP retransmission. */ - if (tp->tlp_high_seq) - goto rearm_timer; - if (skb_still_in_host_queue(sk, skb)) goto rearm_timer;
@@ -2606,10 +2607,12 @@ void tcp_send_loss_probe(struct sock *sk if (__tcp_retransmit_skb(sk, skb, 1)) goto rearm_timer;
+ tp->tlp_retrans = 1; + +probe_sent: /* Record snd_nxt for loss detection. */ tp->tlp_high_seq = tp->snd_nxt;
-probe_sent: NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPLOSSPROBES); /* Reset s.t. tcp_rearm_rto will restart timer from now */ inet_csk(sk)->icsk_pending = 0;