Re: [PATCH v2 2/2] docs: cgroup-v2: Describe effective CPUs of partition roots
From: Guopeng Zhang
Date: Tue Sep 22 2026 - 05:17:54 EST
在 2026/9/22 14:36, Ridong Chen 写道:
>
>
> On 9/22/2026 11:21 AM, Guopeng Zhang wrote:
>> From: Guopeng Zhang <zhangguopeng@xxxxxxxxxx>
>>
>> The cpuset.cpus.effective description says that the effective CPUs are
>> granted by the parent and are normally a subset of cpuset.cpus. This
>> does not always hold for a valid non-root partition root. Its effective
>> CPUs are the active CPUs in cpuset.cpus.exclusive.effective after CPUs
>> assigned to valid child partition roots are removed.
>>
>> If cpuset.cpus.exclusive differs from cpuset.cpus, the effective CPUs
>> may therefore be outside cpuset.cpus. Remote partitions also do not
>> necessarily get their CPUs from their immediate parent.
>>
>> Suggested-by: Ridong Chen <ridong.chen@xxxxxxxxx>
>> Signed-off-by: Guopeng Zhang <zhangguopeng@xxxxxxxxxx>
>> ---
>> Documentation/admin-guide/cgroup-v2.rst | 56 ++++++++++++++++++++-----
>> 1 file changed, 46 insertions(+), 10 deletions(-)
>>
>> diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst
>> index 18a01ce4c334..d21bbae3312a 100644
>> --- a/Documentation/admin-guide/cgroup-v2.rst
>> +++ b/Documentation/admin-guide/cgroup-v2.rst
>> @@ -2505,16 +2505,52 @@ Cpuset Interface Files
>> A read-only multiple values file which exists on all
>> cpuset-enabled cgroups.
>> - It lists the onlined CPUs that are actually granted to this
>> - cgroup by its parent. These CPUs are allowed to be used by
>> - tasks within the current cgroup.
>> -
>> - If "cpuset.cpus" is empty, the "cpuset.cpus.effective" file shows
>> - all the CPUs from the parent cgroup that can be available to
>> - be used by this cgroup. Otherwise, it should be a subset of
>> - "cpuset.cpus" unless none of the CPUs listed in "cpuset.cpus"
>> - can be granted. In this case, it will be treated just like an
>> - empty "cpuset.cpus".
>> + It lists the active CPUs available to tasks in this cgroup.
>> +
>> + For a cgroup that is not a valid partition root, an empty
>> + "cpuset.cpus" makes "cpuset.cpus.effective" show all CPUs
>> + available from the parent cgroup. Otherwise, it is a subset
>> + of "cpuset.cpus" unless none of the requested CPUs can be
>> + granted. In that case, it is treated like an empty
>> + "cpuset.cpus".
>> +
>> + For example::
>> +
>> + P (cpuset.cpus.effective=0-3)
>> + |
>> + +-- C0 (member): cpuset.cpus=1-2
>> + | cpuset.cpus.effective=1-2
>> + +-- C1 (invalid root): cpuset.cpus=4-5
>> + | cpuset.cpus.effective=0-3
>> + \-- C2 (member): cpuset.cpus=<empty>
>> + cpuset.cpus.effective=0-3
>> +
>
> That is clear. Thanks.
>
>> + C0 gets CPUs 1-2 because they are available from P.
>> + None of C1's requested CPUs 4-5 are available from P, so C1 uses P's
>> + effective CPUs 0-3 instead.
>> + C2 has an empty "cpuset.cpus" and therefore also uses P's effective
>> + CPUs 0-3.
>> +
>> + For a valid non-root partition root, "cpuset.cpus.effective" contains
>> + the active CPUs from "cpuset.cpus.exclusive.effective" except
>> + those granted to valid child partition roots. If
>> + "cpuset.cpus.exclusive" is set to a value different from
>> + "cpuset.cpus", the effective CPUs need not be a subset of
>> + "cpuset.cpus".
>> +
>> + For example, if CPUs 2-5 are active, available, and not assigned
>> + to a valid child partition root::
>> +
>> + top (root)
>> + |
>> + \-- P (valid partition root)
>> + cpuset.cpus=0-3
>> + cpuset.cpus.exclusive=2-5
>> + cpuset.cpus.exclusive.effective=2-5
>> + cpuset.cpus.effective=2-5
>> +
>> + P has 2-5 in "cpuset.cpus.effective" even though its
>> + "cpuset.cpus" is 0-3.
>> Its value will be affected by CPU hotplug events.
>>
>
> I would suggest offering an example where "cpuset.cpus.effective" differs from "cpuset.cpus.exclusive.effective", meaning that "cpuset.cpus.exclusive.effective" contains offline CPUs. This would make it much clearer that "cpuset.cpus.effective" does not contain offline CPUs.
>
> With that, feel free to add:
>
> Reviewed-by: Ridong Chen <ridong.chen@xxxxxxxxx>
>
Thanks for the suggestion. I'll update the example in v3 to cover the offline CPU case and add your Reviewed-by.
Thanks,
Guopeng