Re: [PATCH v11 7/9] cpuset: Expose cpus.effective and mems.effective on cgroup v2 root

[Waiman Long]

On 07/06/2018 04:32 PM, Waiman Long wrote:
> On 07/03/2018 11:58 AM, Tejun Heo wrote:
>> Hello, Waiman.
>>
>> On Tue, Jul 03, 2018 at 08:41:31AM +0800, Waiman Long wrote:
>>>> So, effective changing when enabling partition on a child feels wrong
>>>> to me.  It's supposed to contain what's actually allowed to the cgroup
>>>> from its parent and that shouldn't change regardless of how those
>>>> resources are used.  It's still given to the cgroup from its parent.
>>> Another way to work around this issue is to expose the reserved_cpus in
>>> the parent for holding CPUs that can taken by a chid partition. That
>>> will require adding one more cpuset file for those cgroups that are
>>> partition roots.
>> Yeah, that should work.
>>
> Thinking about it a bit more, that approach will make creating a
> partition a multi-step process:
>
> 1) Reserve the CPUs in reserved_cpus.
> 2) enable sched.partition
> 3) Write the CPUs list into cpus.
>
> There are also more exception cases that need to be handled. The current
> approach, on the other hands, is much simpler and easier to understand
> and use.
>
>>> I don't mind restricting that to the first level children for now. That
>>> does restrict where we can put the container root if we want a separate
>>> partition for a container. Let's hear if others have any objection about
>>> that.
>> As currently implemented, partioning locks away the cpus which should
>> be a system level decision, not container level, so it makes sense to
>> me that it is only available to system root.
> So my preference is to allow partition only on the first level children
> of the root for the time being. I think it should cover most of the use
> cases. I will update the patchset to reflect that.
>
> Cheers,
> Longman
>
Below is the incremental patch that allow partitioning only on the first
level children of the root. Please let me know your thourght on that.

Thanks,
Longman

-------------[ Cut here ]-------------------------