Messages in this thread | | | Date | Sun, 04 Jan 2009 16:15:55 +0000 | From | Sitsofe Wheeler <> | Subject | Re: Bug: Status/Summary of slashdot leap-second crash on new years 2008-2009 |
| |
david@lang.hm wrote: > so are you saying that other 'correct' OS's have patches issued every > time a leap second is declared so that they have an in-kernel table of > them to use to calculate the correct time?
I think the number of other "correct" OSes that actually step the time on leap seconds is not that large (at least doing the announcement via NTP). According to http://www.ntp.org/ntpfaq/NTP-s-algo-real.htm#AEN2499 leap seconds are only changed via stepping if you have the right kernel discipline (notes on how to check whether a given OS has the kernel kernel discipline are mentioned on http://www.ntp.org/ntpfaq/NTP-s-algo-kernel.htm#AEN2220 ).
I have a feeling that OSX doesn't do it (there's a mailing list post from 2005 where someone was trying to add FreeBSD's ntp_adjtime to Darwin http://lists.apple.com/archives/Darwin-kernel/2005/Jan/msg00004.html ). Additionally folks I know using ntpd synchronized OSX machines said their machines were off by one second right after the new year.
Windows is also known not to do it without slewing: http://www.meinberg.de/english/info/leap-second.htm#os .
| |