Re: kernel knowledge of localtime (user-level implememntation)

Albert D. Cahalan (acahalan@cs.uml.edu)
Tue, 1 Dec 1998 04:54:48 -0500 (EST)


a sun writes:

>>> so, where do people think a timezone updater should be stuck? do
>>> people not like using the update daemon?
>
>> My machine has an rdate call in /etc/crontab to synchronise its time
>> to an atomic clock, and that call is set to run every three hours...
>
> well, some random program you stick in crontab probably won't do. for
> example, i use ntpd to keep all my clocks up-to-date. should we stick
> it in there as well?

No, it can stand alone as a plain cron job. (and run it at boot)

> this issue has less to do with making sure your actual time is
> accurate than making sure the kernel is in sync with the
> timezone. sticking things in crontab only works if the
> smbfs/fat/ncpfs/hfs users all put it in there (and it gets run
> pretty often due to the chance that the timezone may be switched
> at any time).

The offset change must be done from the kernel, or perhaps from
a real-time process. I'd feel safer with a tiny bit of kernel code.

Calculation of the change must be done in userspace, but you can do
that a month in advance. The kernel only needs two things:

1. time in UTC when the offset will change next
2. new offset

So, once a month your cron job sets up the next change (if any).

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/