I've had a 'gut feel' recently that the system time on my system
is running slightly fast.  I've not checked this, but have noticed
something odd in some of the statistics.

If you look at the interrupt counts with vmstat -i you might get:

pic0 pin 0                                   259304       99 intr
cpu0 softclock                               260001      100 intr

The first is the hardware timer interrupt and the second the
one that links hardclock() to softclock().  Checking with ddb
shows that both hardclock_ticks and softclock_ticks both
equal 'cpu0 softclock' [1].

So it looks as though hardclock() is being called too many times.
This is on a 'boring' single cpu system without ioapics.

Any thoughts?

	David

[1] nfs ensures that softclock() has to be called for every tick.

-- 
David Laight: david@l8s.co.uk