Re: [PATCH v14 00/12] Enable cpuset controller in default hierarchy

[Waiman Long]

On 11/07/2018 04:32 PM, Tejun Heo wrote:
> Hello,
>
> On Tue, Nov 06, 2018 at 09:06:01AM -0500, Waiman Long wrote:
>> On 11/06/2018 06:55 AM, Peter Zijlstra wrote:
>>> On Tue, Nov 06, 2018 at 12:53:35PM +0100, Peter Zijlstra wrote:
>>>> On Mon, Nov 05, 2018 at 08:36:56AM -0800, Tejun Heo wrote:
>>>>> Hello,
>>>>>
>>>>> So, this looks good to me.  Peter, I'm gonna roll the series into
>>>>> cgroup/for-4.21-cpuset.  Please holler if you have any objections /
>>>>> comments.
>>>> Mostly nits, except I think the interface for cpuset.sched.partition
>>>> needs to find another term for "not a parition", "normal" really doesn't
>>>> work for me.
>>> cpuset.sched.partition: "true", "false", "error"
>>>
>>> makes more sense to me.
>> The terms "true", "false", "error" look good to me. I can change the
>> patchset to use those if Tejun has no objection. I will fix the other
>> nits you have as well.
> How about the following?
>
> Interface file is named "cpuset.cpus.partition" as it's a partition of
> cpus.  The writeable keys are "root" and "member" - a partition root
> and a partition member.  When "root" is requested but can't be
> satisfied it shows "root invalid".  This seems pretty consistent with
> cgroup.type and other cpuset files.
>
> Thanks.
>
Yes, that looks good. I will revise my patchset to include those changes
unless Peter has other better idea.

Thanks,
Longman