Re: [PATCH] cpuacct: VIRT_CPU_ACCOUNTING don't prevent percpucputime count

[Bharata B Rao]

On Tue, Apr 28, 2009 at 05:10:47PM +0900, KOSAKI Motohiro wrote:
> > On Tue, Apr 28, 2009 at 04:38:51PM +0900, KOSAKI Motohiro wrote:
> > > > On Tue, Apr 28, 2009 at 03:53:32PM +0900, KOSAKI Motohiro wrote:
> > > 
> > > > BTW, did you observe any real problem with the percpu counter spinlock ?
> > > 
> > > No.
> > > I review percpu_counter() caller recently and it seems stragen usage.
> > > 
> > 
> > I should have phrased the question better ...
> > 
> > So have you found any performance degradation with any benchmarks/workloads
> > on archs which define VIRT_CPU_ACCOUNTING due to percpu_counter spinlock
> > being taken on every tick ? If the answer is no, don't you think we could
> > wait before making the kind of change you are proposing ?
> 
> maybe, I don't understand your point.

My point is, let us not make this change if it is not a real problem that
has been observed on archs which define VIRT_CPU_ACCOUNTING.

Regards,
Bharata.
--
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/