ppc32: Weird process scheduling behaviour with 2.6.24-rc

[Benjamin Herrenschmidt]

On Wed, 2008-01-23 at 14:14 +0100, Michel Dänzer wrote:
> > Does ppc32 (or your instance thereof) have a high resolution
> > sched_clock()?
> 
> I'm not sure (FWIW, we did get support for NO_HZ and HIGH_RES_TIMERS
> in
> 2.6.24-rc as well, but playing with these config options and even
> reverting the code didn't seem to have any effect), can someone from
> the
> linuxppc-dev list answer this?

We do have a hires sched_clock() based on the processor internal
timebase and scaled to ns. Maybe we screwed something up there ? The
implementation is in arch/powerpc/kernel/timer.c

/*
 * Scheduler clock - returns current time in nanosec units.
 *
 * Note: mulhdu(a, b) (multiply high double unsigned) returns
 * the high 64 bits of a * b, i.e. (a * b) >> 64, where a and b
 * are 64-bit unsigned numbers.
 */
unsigned long long sched_clock(void)
{
	if (__USE_RTC())
		return get_rtc();
	return mulhdu(get_tb() - boot_tb, tb_to_ns_scale) << tb_to_ns_shift;
}

(You can mostly ignore the RTC() case which is the native ns clock of
the old 601 processor.

Ben.