From 7f5f81406e2b36785f1e25fe5209edd9dd3610d7 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Fri, 17 Jul 2020 16:37:25 -0700 Subject: rhashtable: drop duplicated word in Drop the doubled word "be" in a comment. Signed-off-by: Randy Dunlap Cc: Thomas Graf Cc: Herbert Xu Cc: netdev@vger.kernel.org Signed-off-by: David S. Miller --- include/linux/rhashtable.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/rhashtable.h b/include/linux/rhashtable.h index 70ebef866cc8..d3432ee65de7 100644 --- a/include/linux/rhashtable.h +++ b/include/linux/rhashtable.h @@ -33,7 +33,7 @@ * of two or more hash tables when the rhashtable is being resized. * The end of the chain is marked with a special nulls marks which has * the least significant bit set but otherwise stores the address of - * the hash bucket. This allows us to be be sure we've found the end + * the hash bucket. This allows us to be sure we've found the end * of the right list. * The value stored in the hash bucket has BIT(0) used as a lock bit. * This bit must be atomically set before any changes are made to -- cgit v1.2.3 From 76be93fc0702322179bb0ea87295d820ee46ad14 Mon Sep 17 00:00:00 2001 From: Yuchung Cheng Date: Thu, 23 Jul 2020 12:00:06 -0700 Subject: tcp: allow at most one TLP probe per flight 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 Signed-off-by: Neal Cardwell Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- include/linux/tcp.h | 6 ++++-- net/ipv4/tcp_input.c | 11 ++++++----- net/ipv4/tcp_output.c | 13 ++++++++----- 3 files changed, 18 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 9aac824c523c..a1bbaa1c1a3a 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -220,7 +220,9 @@ struct tcp_sock { } rack; u16 advmss; /* Advertised MSS */ u8 compressed_ack; - u8 dup_ack_counter; + u8 dup_ack_counter:2, + tlp_retrans:1, /* TLP is a retransmission */ + unused:5; 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 */ @@ -243,7 +245,7 @@ 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 */ diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 9615e72656d1..518f04355fbf 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3488,10 +3488,8 @@ static void tcp_replace_ts_recent(struct tcp_sock *tp, u32 seq) } } -/* 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) { @@ -3500,7 +3498,10 @@ static void tcp_process_tlp_ack(struct sock *sk, u32 ack, int flag) 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)) { diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 5f5b2f0b0e60..0bc05d68cd74 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2624,6 +2624,11 @@ 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; @@ -2641,10 +2646,6 @@ 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; @@ -2666,10 +2667,12 @@ 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; -- cgit v1.2.3 From c2b69f24ebd166a13cdc9909b50f33228895998b Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Fri, 24 Jul 2020 10:50:22 +1000 Subject: flow_offload: Move rhashtable inclusion to the source file I noticed that touching linux/rhashtable.h causes lib/vsprintf.c to be rebuilt. This dependency came through a bogus inclusion in the file net/flow_offload.h. This patch moves it to the right place. This patch also removes a lingering rhashtable inclusion in cls_api created by the same commit. Fixes: 4e481908c51b ("flow_offload: move tc indirect block to...") Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- include/net/flow_offload.h | 1 - net/core/flow_offload.c | 1 + net/sched/cls_api.c | 1 - 3 files changed, 1 insertion(+), 2 deletions(-) (limited to 'include') diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h index 6315324b9dc2..3eaf25f68b79 100644 --- a/include/net/flow_offload.h +++ b/include/net/flow_offload.h @@ -5,7 +5,6 @@ #include #include #include -#include struct flow_match { struct flow_dissector *dissector; diff --git a/net/core/flow_offload.c b/net/core/flow_offload.c index b739cfab796e..2076219b8ba5 100644 --- a/net/core/flow_offload.c +++ b/net/core/flow_offload.c @@ -4,6 +4,7 @@ #include #include #include +#include struct flow_rule *flow_rule_alloc(unsigned int num_actions) { diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index e62beec0d844..4619cb3cb0a8 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -20,7 +20,6 @@ #include #include #include -#include #include #include #include -- cgit v1.2.3