Date: Fri, 2 Jan 2009 04:29:49 +0100 From: Erik Trulsson <ertr1013@student.uu.se> To: Darren Pilgrim <freebsd@bitfreak.org> Cc: freebsd-hackers@freebsd.org Subject: Re: Is the FreeBSD clock UTC or TAI? Message-ID: <20090102032949.GA33780@owl.midgard.homeip.net> In-Reply-To: <495D7421.1090208@bitfreak.org> References: <495D7421.1090208@bitfreak.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jan 01, 2009 at 05:55:45PM -0800, Darren Pilgrim wrote: > This came up during discussion of leap seconds and why UTC and TAI are > different. My question is, does FreeBSD's internal clock use UTC or TAI > for timekeeping? That is, is wallclock calculated from an exact count > of the number of seconds since epoch (TAI), then adjusted with a leap > seconds table to match UTC, or does it internally use UTC and have code > to deal with the ambiguous seconds that occur at each leap second? I think the answer is no. Instead I believe FreeBSD follows the POSIX rules which mandates using UTC, while completely ignoring the concept of leap seconds. There was a long thread over on freebsd-current@ in early January 2006 titled "FreeBSD handles leapsecond correctly" that discussed this at length. (The general consensus seems to have been that leap seconds are evil and a PITA and essentially impossible to handle 'correctly' since various standards differ on how they should be handled. What to do about them is less clear however.) -- <Insert your favourite quote here.> Erik Trulsson ertr1013@student.uu.se
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090102032949.GA33780>