Re: [PATCH v3 2/6] account guest time per-cgroup as well.

[Paul Turner]

On Wed, May 30, 2012 at 3:36 AM, Glauber Costa <glommer@xxxxxxxxxxxxx> wrote:
> On 05/30/2012 02:32 PM, Peter Zijlstra wrote:
>>
>> On Wed, 2012-05-30 at 13:48 +0400, Glauber Costa wrote:
>>>
>>> >  We already track multiple tick statistics per-cgroup, using
>>> >  the task_group_account_field facility. This patch accounts
>>> >  guest_time in that manner as well.
>>
>> So this leaves IOWAIT and IDLE the only fields not using
>> task_group_account_field(), right?
>
>
> Yes, because they are essentially global, and their meaning is ill-defined
> from within a cgroup.
>
> If you look further out in the patchset, I intend to export idle from cpu,
> instead of cpuacct, because something that can be used as idle value is
> already computed anyway from the schedstats, so I'm just using it.
>
> iowait will be left blank for now. Me and Paul agreed last time we talked
> that it is not uber important to have iowait values per-cgroup.

Stronger:  it lacks a definition you can sanely measure without atomic
counters everywhere (similarly for group-idle).

> --
> To unsubscribe from this list: send the line "unsubscribe cgroups" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
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/