Re: [PATCH] i386 tsc: remove xtime_lock'ing around cpufreq notifier

[Andi Kleen <ak@novell.com>]

On Thursday 12 April 2007 19:55:59 Andrew Morton wrote:
> On Thu, 12 Apr 2007 10:43:03 -0700
> Jeremy Fitzhardinge <jeremy@goop.org> wrote:
> 
> > Andrew Morton wrote:
> > > hm.  People (ab)use sched_clock() for all sorts of things nowadays.  I wouldn't
> > > do anything to degrade it just on behalf of printk-timestamping.
> > >   
> > 
> > printk was the only non-scheduler-ish use I could find.  Are there others?
> > 
> 
> blktrace.  I've seen a couple of trace/debug-style things which use
> sched_clock for timestamping event collection. I think lttng does exotic
> things with TSCs, performing private skew correction, although that might
> have changed now.

They should always just store the cpu too and educate the users that
only (cpu, timestamp) pairs make sense to compare.

That said at least my new sched_clock should not normally show 
large non differences between CPUs, so it can be usually ignored, but they can 
happen. I believe some of the already existing sched_clocks() (like the
one used on Altix) have the same property. 

But on VMI/Xen as currently implemented the differences will be large.

-Andi