Re: [PATCH] sched: Support current clocksource handling in fallback

[Thomas Gleixner <tglx@linutronix.de>]

On Tue, 26 May 2009, john stultz wrote:

> On Tue, 2009-05-26 at 16:38 +0200, Peter Zijlstra wrote:
> > Added the generic clock and timer folks to CC.
> > 
> > On Tue, 2009-05-26 at 16:31 +0200, Linus Walleij wrote:
> > > 2009/5/26 Paul Mundt <lethal@linux-sh.org>:
> > > 
> > > >  */
> > > >  unsigned long long __attribute__((weak)) sched_clock(void)
> > > >  {
> > > > +       /*
> > > > +        * Use the current clocksource when it becomes available later in
> > > > +        * the boot process, and ensure that it has a high enough rating
> > > > +        * to make it suitable for general use.
> > > > +        */
> > > > +       if (clock && clock->rating >= 100)
> > > > +               return cyc2ns(clock, clocksource_read(clock));
> 
> I'm not super familiar with the recent sched_clock changes, but how will
> this work if the clocksource wraps (ACPI PM wraps every 2-5 seconds).

You don't want to use ACPI PM for sched_clock, never ever.
 
> Also there's no locking here, so the clocksource could change under you.
> 
> Further, checking for rating being greater then 100 really doesn't mean
> anything. Probably need to check if the clocksource is continuous
> instead.

I'd like to have an explicit flag for this, so we can avoid that stuff
like pmtimer and other slow access clock sources are used.
 
> Overall, I'd probably suggest thinking this through a bit more. At some
> point doing this right will cause sched_clock() to be basically the same
> as ktime_get(). So why not just use that instead of remaking it?

ktime_get() involves xtime lock and the scheduler does not care about
a slightly wrong value. sched_clock does not have the accuracy
requirements of time keeping,

If we have an explicit flag we can replace lots of arch/embedded
sched_clock implementations with a generic one which is a Good Thing.

There is a world beside the broken x86 timers :)

Thanks,

	tglx