Re: [PATCH] sched: Support current clocksource handling in fallback sched_clock().

[Paul Mundt <lethal@linux-sh.org>]

On Tue, May 26, 2009 at 04:50:09PM +0200, Peter Zijlstra wrote:
> On Tue, 2009-05-26 at 23:43 +0900, Paul Mundt wrote:
> > > Else you might want an additional criteria, like
> > > cyc2ns(1) (much less than) jiffies_to_usecs(1)*1000
> > > (however you do that the best way)
> > > so you don't pick something
> > > that isn't substantially faster than the jiffy counter atleast?
> > > 
> > This rather defeats the purpose of sched_clock() being fast. If we want
> > to add a flag that means this in to the clocksource instead of consulting
> > the rating, that is fine with me too. I know which clocksources I prefer
> > to use for a sched_clock() and they are all better than jiffies. The
> > semantics of how we tell sched_clock() that are not so important. Rating
> > seemed like a good choice from the documentation in struct clocksource at
> > least.
> 
> Am I confused or are we talking about fast HZ vs fast cycles?
> 
> sched_clock() should be fast cycles, that is, we don't want to read a
> clock that takes about 1000 cycles.
> 
> sched_clock() is about providing a high resolution clock that is fast
> (low cycle count) to acquire, and need not be strictly monotonic on smp.

I don't think there's any confusion here. My point is that I didn't want
to add too much logic in to sched_clock() given that it is supposed to be
fast. So if the rating test by itself is not sufficient, then we need
another way to flag a clocksource as being usable for sched_clock().