Date: Mon, 12 Sep 2005 18:45:52 +0200 From: "Poul-Henning Kamp" <phk@phk.freebsd.dk> To: Jung-uk Kim <jkim@FreeBSD.org> Cc: cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/kern vfs_subr.c Message-ID: <11401.1126543552@phk.freebsd.dk> In-Reply-To: Your message of "Mon, 12 Sep 2005 12:40:49 EDT." <200509121240.51775.jkim@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <200509121240.51775.jkim@FreeBSD.org>, Jung-uk Kim writes: >Please make sure to use `time_uptime' if you need monotonic timestamp. >Especially many network stacks seem to make sequence IDs and >timestamps to expire from `time_second' instead of time_uptime. >time_second is not guaranteed to be monotonic! time_uptime is. > >http://docs.freebsd.org/cgi/mid.cgi?8153.1126340565 >http://docs.freebsd.org/cgi/mid.cgi?20050909214808.GA6021 I would like to add that this is *also* true for userland where uptime is called clock_gettime(CLOCK_MONOTONIC,...). Please consider carefully, if you need a timestamp for a certain time of day or just a certain inverval of time measured from now. In the latter case, it is _always_ uptime/CLOCK_MONOTONIC you want. Failure to do this means that your program/kernel code will do the wrong thing when the clock is stepped (manually or by NTP). -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?11401.1126543552>