Re: 2.6.14-rt21: slow-running clock

[Jonathan Woithe]

> > > > When running Ingo's 2.6.14-rt21 (and in fact rt kernels back to at least
> > > > 2.6.13-rc days), the clock on my i915-based laptop runs slow.  The degree
> > > > of slowness appears directly related to how busy the machine is.  If
> > > > it is just sitting around doing very little the time is kept rather
> > > > well.  However, as soon as the load increases the RTC and system time
> > > > diverge significantly.  For example, running jackd for 2 minutes results
> > > > in the system time loosing as much as 20 seconds compared to the CMOS RTC.
> > > > Processes doing HDD I/O also seem to affect the system time similarly.
> > > > 
> > > > Selectively disabling different timer-related kernel options does not make
> > > > any difference.  However, the clock seems fine under vanilla 2.6.14,
> > > > suggesting an issue somewhere in the rt patches.
> > > 
> > > Could you please send me your dmesg and the output of:
> > > 
> > > 	cat  /sys/devices/system/clocksource/clocksource0/*
> > 
> > First the contents of the above /sys/ files:
> > 
> >   /sys/devices/system/clocksource/clocksource0/current_clocksource:
> >     c3tsc
> > 
> >   /sys/devices/system/clocksource/clocksource0/available_clocksource
> >     acpi_pm jiffies c3tsc pit
> 
> Odd. I'm not sure why the acpi_pm wasn't chosen by default if it was
> available and the TSC fell back to the c3tsc. It might be something in
> the -RT tree that's changed that bit. Could you try the following and
> see if it doesn't resolve the timekeeping problems you're seeing?
> 
> echo "acpi_pm" >  /sys/devices/system/clocksource/clocksource0/current_clocksource

Will test it tonight.

> Still it sounds like something isn't right w/ either the c3tsc code or
> the cpufreq notification code. Would you be willing to test further
> patches?

Yes, no problem.

> Also could you try booting w/ idle=poll and see if that changes the
> behavior?

Again, it will be done tonight and I'll advise tomorrow.

Regards
  jonathan
-
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/