From owner-svn-src-all@freebsd.org Thu Feb 21 09:22:05 2019 Return-Path: Delivered-To: svn-src-all@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id E2D9514FF62E; Thu, 21 Feb 2019 09:22:04 +0000 (UTC) (envelope-from tuexen@freebsd.org) Received: from drew.franken.de (drew.ipv6.franken.de [IPv6:2001:638:a02:a001:20e:cff:fe4a:feaa]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "*.franken.de", Issuer "COMODO RSA Domain Validation Secure Server CA" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 68D8D69C89; Thu, 21 Feb 2019 09:22:04 +0000 (UTC) (envelope-from tuexen@freebsd.org) Received: from [192.168.1.9] (p57BB46EC.dip0.t-ipconnect.de [87.187.70.236]) (Authenticated sender: macmic) by mail-n.franken.de (Postfix) with ESMTPSA id CA965721E281A; Thu, 21 Feb 2019 10:21:57 +0100 (CET) Content-Type: text/plain; charset=us-ascii Mime-Version: 1.0 (Mac OS X Mail 12.2 \(3445.102.3\)) Subject: Re: svn commit: r344368 - head/sys/netinet From: Michael Tuexen In-Reply-To: <7b4f29d5-e909-6397-7c24-b5979a3963c1@FreeBSD.org> Date: Thu, 21 Feb 2019 10:21:56 +0100 Cc: src-committers , svn-src-all@freebsd.org, svn-src-head@freebsd.org, freebsd-transport@freebsd.org Content-Transfer-Encoding: quoted-printable Message-Id: <1141FFDF-F764-4CBB-B245-8E92815FCE8F@freebsd.org> References: <201902201803.x1KI3haW078567@repo.freebsd.org> <7b4f29d5-e909-6397-7c24-b5979a3963c1@FreeBSD.org> To: "koobs@freebsd.org" X-Mailer: Apple Mail (2.3445.102.3) X-Spam-Status: No, score=-2.9 required=5.0 tests=ALL_TRUSTED,BAYES_00 autolearn=disabled version=3.4.1 X-Spam-Checker-Version: SpamAssassin 3.4.1 (2015-04-28) on mail-n.franken.de X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 21 Feb 2019 09:22:05 -0000 > On 21. Feb 2019, at 09:32, Kubilay Kocak wrote: >=20 > On 21/02/2019 5:03 am, Michael Tuexen wrote: >> Author: tuexen >> Date: Wed Feb 20 18:03:43 2019 >> New Revision: 344368 >> URL: https://svnweb.freebsd.org/changeset/base/344368 >> Log: >> Reduce the TCP initial retransmission timeout from 3 seconds to >> 1 second as allowed by RFC 6298. >> Reviewed by: kbowling@, Richard Scheffenegger >> Sponsored by: Netflix, Inc. >> Differential Revision: https://reviews.freebsd.org/D18941 >> Modified: >> head/sys/netinet/tcp_syncache.c >> head/sys/netinet/tcp_timer.h >> Modified: head/sys/netinet/tcp_syncache.c >> = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D >> --- head/sys/netinet/tcp_syncache.c Wed Feb 20 17:56:38 2019 = (r344367) >> +++ head/sys/netinet/tcp_syncache.c Wed Feb 20 18:03:43 2019 = (r344368) >> @@ -159,7 +159,7 @@ static int syncookie_cmp(struct = in_conninfo *inc, str >> * tcp_backoff[1] + >> * tcp_backoff[2] + >> * tcp_backoff[3]) + 3 * tcp_rexmit_slop, >> - * 3000 ms * (1 + 2 + 4 + 8) + 3 * 200 ms =3D 45600 ms, >> + * 1000 ms * (1 + 2 + 4 + 8) + 3 * 200 ms =3D 15600 ms, >> * the odds are that the user has given up attempting to connect by = then. >> */ >> #define SYNCACHE_MAXREXMTS 3 >> Modified: head/sys/netinet/tcp_timer.h >> = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D >> --- head/sys/netinet/tcp_timer.h Wed Feb 20 17:56:38 2019 = (r344367) >> +++ head/sys/netinet/tcp_timer.h Wed Feb 20 18:03:43 2019 = (r344368) >> @@ -77,7 +77,7 @@ >> #define TCPTV_MSL ( 30*hz) /* max seg = lifetime (hah!) */ >> #define TCPTV_SRTTBASE 0 /* base = roundtrip time; >> if 0, no idea yet */ >> -#define TCPTV_RTOBASE ( 3*hz) /* assumed RTO = if no info */ >> +#define TCPTV_RTOBASE ( 1*hz) /* assumed RTO = if no info */ >> #define TCPTV_PERSMIN ( 5*hz) /* minimum = persist interval */ >> #define TCPTV_PERSMAX ( 60*hz) /* maximum = persist interval */ >> _______________________________________________ >=20 > Any way this can be a run-time tunable? That is definitely possible. However, changing the default should be = separate from making it sysctl-able. I'll bring this up in the bi-weekly transport call and see what others = think. Best regards Michael >=20