On Thu, Jan 13, 2000 at 10:11:45PM +0000, wayne@moneyworld.co.uk wrote:
> Hi there,
>
> Before I start, I am running kernel 2.2.12 on a RedHat 6.0 box.
>
> I've just noticed some weird messages in the /var/log/messages on one of
> my machines... A section of the log is attached... The mesages I am
> worried about are the ones like
> Jan 13 22:04:33 he4 kernel: set_rtc_mmss: can't update from 53 to 4
>
> Is this a problem and if so, is it known ?
>
> ---------------------
>
> Jan 13 22:04:33 he4 kernel: set_rtc_mmss: can't update from 53 to 4
> Jan 13 22:05:34 he4 kernel: set_rtc_mmss: can't update from 53 to 5
> Jan 13 22:06:35 he4 kernel: set_rtc_mmss: can't update from 53 to 6
> Jan 13 22:07:36 he4 kernel: set_rtc_mmss: can't update from 53 to 7
> -------------------------
This means that you somehow triggered the infamous "11-minute RTC update"
where the kernel fiddles with RTC every 11 minutes, because it thinks
the RTC contents is worthless compared to its own superior data.
Such things happen if you use software that calls adjtimex(), for example xntpd.
It furthermore means that the kernel reports that its attempt at
fiddling failed. Since the kernel does not want to know too much
(it assumes that RTC time and kernel time should differ by an
integral number of half hours, never mind the people in Nepal)
it will not touch the hours part of the RTC and if an update would
require changing the hours part it comes with this printk and waits
for another opportunity.
Finally, it means that either your RTC is stopped or is not being read
correctly. You might experiment a bit with a recent hwclock.
If the kernel fiddling can cause the RTC to stop, that would be
one more reason to consider removing this code.
Since the RTC is not used anywhere, this kernel fiddling is fairly
useless, and if it fails, the failure is fairly harmless.
The only thing is that you may have a strange time and/or date
immediately after the next reboot.
Andries
[Note also that there may be a bad interaction between the kernel
11-minute mode and hwclock --adjust. Hwclock tries to find the drift
of the RTC, and if it looks at the RTC before and after a kernel adjustment
it will conclude that the RTC has an enormous drift. Thus, you may find
that /etc/adjtime contains ridiculous numbers.]
-
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/
This archive was generated by hypermail 2b29 : Sat Jan 15 2000 - 21:00:26 EST