On Fri, Apr 2, 2010 at 3:59 AM, Peter Zijlstra <peterz@infradead.org> wrote:
quoted text
> On Thu, 2010-04-01 at 22:18 +0200, Thomas Gleixner wrote:
>> Well, the uninterruptible count imbalance is preventing us to calc the
>> load avg per cpu, which is something the power saving folks are
>> interested in.
>>
>> If that is fixed we probably have a good chance to collect the per cpu
>> stuff and build a combined one - have not looked into the gory details
>> of the math yet.
>
> The problem with all this per-cpu loadavg is that it would require a
> steady tick on each cpu to age this loadavg, which is quite in
> contradiction with power savings as they rather like the nohz feature.
>
> No idea yet on what to do about that.

What if we stored the time at which a cpu goes idle in its rq. Then,
when the time comes around to calculate the load avg one cpu should be
able to scan the run queues of all the cpus and use their load avg if
not idle, or use their load avg + idle timestamp math if idle.

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