CMOS time jumps when ntpd ist running

[Juergen Beisert]

Hi list,

anybody here who saw such CMOS clock behaviour when ntpd is running?
I synced system and CMOS time and after that I ran a simple
  'watch -t -n 300 "date; hwclock -u" > /logfile'
(note: When the ntpd daemon is not running, the CMOS clock does not show this 
behaviour):

[...]
    Mon Aug 10 03:53:20 CEST 2009              <-- system time
    Mon Aug 10 03:53:20 2009  0.000000 seconds <-- CMOS time

    Mon Aug 10 03:58:20 CEST 2009
    Mon Aug 10 03:58:20 2009  0.000000 seconds

    Mon Aug 10 04:03:20 CEST 2009
    Sun Aug  9 22:03:20 2009  0.000000 seconds <----- BANG !!!!!!111!1!!

    Mon Aug 10 04:08:20 CEST 2009
    Sun Aug  9 22:08:20 2009  0.000000 seconds
[...]

It seems its always -6 hours the CMOS clock jumps. And sometimes -(6h + n * 
24h) with n from 1 to 3 I saw.

Is my CMOS clock broken? (BTW: x86 arch, 32 bit, Linux kernel 2.6.28.10 and 
there is no SMM code in the system, it runs coreboot)

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