Messages in this thread | | | Date | Tue, 25 Nov 2008 12:09:35 +0100 | From | Andi Kleen <> | Subject | Re: [2.6.28-rc6] oprofile: "opcontrol --start" output two warnings |
| |
> if so, I think get_stagger() is a bit strange. > it depend on caller cpu. then if PREEMPT=Y, it return radom result.
Even without PREEMPT it is random because there is no guarantee the init code is executing on CPU 0
> > I'm not sure about oprofile design. > but if you are right, I think p4_fill_in_addresses shoudn't use smp_processor_id().
Correct.
> > Am I missing any point?
No you're right. Always returning 0 in get_stagger() should be ok I think, at least it wouldn't make anything worse.
Or perhaps figure out if the per cpu addresses are really needed, if yes then this would need much more changes. But I hope that would not be needed.
But someone should better test it, the P4 perfmon handling is certainly hairy and I don't claim to understand all its intricate details.
-Andi
-- ak@linux.intel.com
| |