diff options
author | Yuchung Cheng <ycheng@google.com> | 2013-09-03 17:14:35 -0400 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2013-09-04 14:41:55 -0400 |
commit | 52f20e655d9f6f7f937a1cdacf219d9df3ab6166 (patch) | |
tree | 234a887f331af07ebb2ff5a1ae80823f8e1456b5 /net/ipv4 | |
parent | 430eda6d6d568eded71dfd1be5a16c0c1379e201 (diff) |
tcp: better comments for RTO initiallization
Commit 1b7fdd2ab585("tcp: do not use cached RTT for RTT estimation")
removes important comments on how RTO is initialized and updated.
Hopefully this patch puts those information back.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv4')
-rw-r--r-- | net/ipv4/tcp_metrics.c | 26 |
1 files changed, 20 insertions, 6 deletions
diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c index 273ed735cca2..4a22f3e715df 100644 --- a/net/ipv4/tcp_metrics.c +++ b/net/ipv4/tcp_metrics.c | |||
@@ -481,13 +481,27 @@ void tcp_init_metrics(struct sock *sk) | |||
481 | crtt = tcp_metric_get_jiffies(tm, TCP_METRIC_RTT); | 481 | crtt = tcp_metric_get_jiffies(tm, TCP_METRIC_RTT); |
482 | rcu_read_unlock(); | 482 | rcu_read_unlock(); |
483 | reset: | 483 | reset: |
484 | /* The initial RTT measurement from the SYN/SYN-ACK is not ideal | ||
485 | * to seed the RTO for later data packets because SYN packets are | ||
486 | * small. Use the per-dst cached values to seed the RTO but keep | ||
487 | * the RTT estimator variables intact (e.g., srtt, mdev, rttvar). | ||
488 | * Later the RTO will be updated immediately upon obtaining the first | ||
489 | * data RTT sample (tcp_rtt_estimator()). Hence the cached RTT only | ||
490 | * influences the first RTO but not later RTT estimation. | ||
491 | * | ||
492 | * But if RTT is not available from the SYN (due to retransmits or | ||
493 | * syn cookies) or the cache, force a conservative 3secs timeout. | ||
494 | * | ||
495 | * A bit of theory. RTT is time passed after "normal" sized packet | ||
496 | * is sent until it is ACKed. In normal circumstances sending small | ||
497 | * packets force peer to delay ACKs and calculation is correct too. | ||
498 | * The algorithm is adaptive and, provided we follow specs, it | ||
499 | * NEVER underestimate RTT. BUT! If peer tries to make some clever | ||
500 | * tricks sort of "quick acks" for time long enough to decrease RTT | ||
501 | * to low value, and then abruptly stops to do it and starts to delay | ||
502 | * ACKs, wait for troubles. | ||
503 | */ | ||
484 | if (crtt > tp->srtt) { | 504 | if (crtt > tp->srtt) { |
485 | /* Initial RTT (tp->srtt) from SYN usually don't measure | ||
486 | * serialization delay on low BW links well so RTO may be | ||
487 | * under-estimated. Stay conservative and seed RTO with | ||
488 | * the RTTs from past data exchanges, using the same seeding | ||
489 | * formula in tcp_rtt_estimator(). | ||
490 | */ | ||
491 | inet_csk(sk)->icsk_rto = crtt + max(crtt >> 2, tcp_rto_min(sk)); | 505 | inet_csk(sk)->icsk_rto = crtt + max(crtt >> 2, tcp_rto_min(sk)); |
492 | } else if (tp->srtt == 0) { | 506 | } else if (tp->srtt == 0) { |
493 | /* RFC6298: 5.7 We've failed to get a valid RTT sample from | 507 | /* RFC6298: 5.7 We've failed to get a valid RTT sample from |