Re: [tip:sched/core] sched: Untangle cpu-load and timekeeping code

[Peter Zijlstra]

On Thu, 2011-01-27 at 15:13 +0800, Yong Zhang wrote:
> On Wed, Jan 26, 2011 at 8:14 PM, tip-bot for Peter Zijlstra
> <a.p.zijlstra@xxxxxxxxx> wrote:
> > Commit-ID:  18a57c3b43781ea83c3bb21c3add84ba890d5fb3
> > Gitweb:     http://git.kernel.org/tip/18a57c3b43781ea83c3bb21c3add84ba890d5fb3
> > Author:     Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
> > AuthorDate: Tue, 25 Jan 2011 11:30:35 +0100
> > Committer:  Ingo Molnar <mingo@xxxxxxx>
> > CommitDate: Wed, 26 Jan 2011 12:33:23 +0100
> >
> > sched: Untangle cpu-load and timekeeping code
> >
> > Remove the calc_global_load() call from the timekeeping code and make
> > it local to the scheduler.
> >
> > Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
> > LKML-Reference: <new-submission>
> > Signed-off-by: Ingo Molnar <mingo@xxxxxxx>
> 
> Now another question comes into my head:
> 
> Previous we stop load calculation only when we goes to full idle,
> say every CPU goes to idle. tick_do_timer_cpu is jumping to
> the no-idle one so we can always get the real load.
> 
> But now, if the CPU on which the global_load_timer resident
> goes to idle and there is still no-idle cpu in system, the load
> calculation stops. So it's not reflecting the real load after
> that.

Hmm, good point, Thomas is there any way we can make this timer stay on
an active cpu for as long as there is one?

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/