Return to BSD News archive
Newsgroups: comp.os.386bsd.questions Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!howland.reston.ans.net!pipex!uunet!news.smith.edu!sophia.smith.edu!jfieber From: jfieber@sophia.smith.edu (J Fieber) Subject: Re: timezone in kernel configuration (question) Message-ID: <1993Oct14.134728.24994@sophia.smith.edu> Sender: root@sophia.smith.edu (Operator) Nntp-Posting-Host: sophia.smith.edu Organization: Smith College, Northampton, MA, USA References: <1993Oct13.020048.18590@sophia.smith.edu> <2709@hcshh.hcs.de> Date: Thu, 14 Oct 1993 13:47:28 GMT Lines: 35 In article <2709@hcshh.hcs.de>, Hellmuth Michaelis <hm@hcshh.hcs.de> wrote: >When i recall it correctly, the number before "dst" is the number of hours >west of Greenwich (here it is one hour east = -1) and the number after the >"dst" is the Daylight Saving Time 'Method' to be applied when switching >(in this case) between MET and METDST. The "4" is for MET, i found this in >the System Administration Manual in the original BSD 4.3 manual set. I've been searching through the kernel code trying to find places where the dst value is actually used and have not found any. And, as I said, some of the docs indicate that the timezone information is no longer maintained in the kernel. However, I did find a pretty crude hack in /sys/i386/isa/clock.c relating to daylight savings time. As it stands, it assumes that the user manually adjusts the bios clock for daylight savings time. This is probably a reasonable assumption if the machine is also used for some other OS. However, the code that starts up the clock only makes a rough guess as to whether daylight savings time is in effect or not. As such your clock may jump back and forth around the change and if you live somewhere with different daylight savings time rules it will simply jump an hour unexpectedly twice a year *approximately* when the US changes time. Blech. I just #ifdef'd the offending bit of code out and set my bios clock to read GMT and all works perfectly. In the same file there are two places where the leap year calculation is wrong, though the bug won't show up till the year 2100. Again, for reference this comes from the FreeBSD GAMMA source. -john -- == John Fieber ================================== Young Science Library == == jfieber@sophia.smith.edu ======= Smith College, Northampton MA 01063 ==