Techwizz/kernel-fxtec-pro1x
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

tcp: retry more conservatively on local congestion

Browse source 
Previously when the sender fails to retransmit a data packet on
timeout due to congestion in the local host (e.g. throttling in
qdisc), it'll retry within an RTO up to 500ms.

In low-RTT networks such as data-centers, RTO is often far
below the default minimum 200ms (and the cap 500ms). Then local
host congestion could trigger a retry storm pouring gas to the
fire. Worse yet, the retry counter (icsk_retransmits) is not
properly updated so the aggressive retry may exceed the system
limit (15 rounds) until the packet finally slips through.

On such rare events, it's wise to retry more conservatively (500ms)
and update the stats properly to reflect these incidents and follow
the system limit. Note that this is consistent with the behavior
when a keep-alive probe is dropped due to local congestion.

Change-Id: I39ca48f733faefcda4e9cc091fb3c860fc9e0708
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Git-Commit: 590d2026d62418bb27de9ca87526e9131c1f48af
Git-repo: https://android.googlesource.com/kernel/common/
Signed-off-by: Kaustubh Pandey <kapandey@codeaurora.org>
This commit is contained in:
Yuchung Cheng 2019-01-16 15:05:34 -08:00 committed by Kaustubh Pandey
parent 5eb4f3dce3
commit 8591132ecd
1 changed files with 3 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
net/ipv4/tcp_timer.c
View file

@ -531,14 +531,13 @@ void tcp_retransmit_timer(struct sock *sk)
tcp_enter_loss(sk);
icsk->icsk_retransmits++;
if (tcp_retransmit_skb(sk, tcp_rtx_queue_head(sk), 1) > 0) {
/* Retransmission failed because of local congestion,
* do not backoff.
* Let senders fight for local resources conservatively.
*/
if (!icsk->icsk_retransmits)
icsk->icsk_retransmits = 1;
inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
min(icsk->icsk_rto, TCP_RESOURCE_PROBE_INTERVAL),
TCP_RESOURCE_PROBE_INTERVAL,
TCP_RTO_MAX);
goto out;
}
@ -559,7 +558,6 @@ void tcp_retransmit_timer(struct sock *sk)
* the 120 second clamps though!
*/
icsk->icsk_backoff++;
icsk->icsk_retransmits++;
out_reset_timer:
/* If stream is thin, use linear timeouts. Since 'icsk_backoff' is