Re: i386/RTC: old problem, new solution?
Richard Gooch (rgooch@atnf.csiro.au)
Fri, 21 May 1999 11:40:40 +1000
Riley Williams writes:
> Hi Richard.
>
> >>> The only place I'm aware that the kernel needs to know the
> >>> timezone is this case with suspend/resume. So aside from this,
> >>> I don't see a need for the kernel to know the timezone.
>
> >> That's really the second place where it would be useful. The first
> >> is during the initial boot stages, from kernel loading to the point
> >> in the initialisation where `hwclock --hctosys` is run, since during
> >> this interval, messages sent to syslogd or klogd get stamped with
> >> the RTC time taken as UTC even if it isn't...
>
> > I don't see that. Provided you update the system clock before
> > starting syslogd, the times recorded will be correct. The kernel
> > ring buffer doesn't store timestamps. It's syslogd that adds the
> > timestamps.
>
> So the times appearing in /var/log/messages are the times when
> syslogd put its grubby paws on the message then, and not when it was
> issued?
Yep.
> Maybe I'm being stupid, but it seems to me that having the kernel
> tag the timestamp when the message was generated onto the message
> would help with tracking down races as it would provide a better
> idea of just when various problems occurred...
I don't think you really need that. The kernel messages are stored in
time order anyway. That's generally good enough to debug races.
Regards,
Richard....
-
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/