Re: [PATCH] cgroup/cpuset: Support multiple source/destination cpusets using pids pattern

[Ridong Chen]

On 6/6/2026 1:15 AM, Waiman Long wrote:
> On 6/5/26 3:35 AM, Ridong Chen wrote:
>>
>> On 6/4/2026 2:47 AM, Waiman Long wrote:
>>> On 6/3/26 6:26 AM, Ridong Chen wrote:
>>>> The current cpuset_can_attach() and cpuset_attach() functions assume
>>>> task
>>>> migration is from one source cpuset to one destination cpuset. This
>>>> can be
>>>> wrong in several scenarios:
>>>>    - Moving a multi-threaded process with threads in different cpusets
>>>>    - Disabling the cpuset controller (many children to one parent)
>>>>    - Enabling the cpuset controller (one parent to many children)
>>>>
>>>> Fix this by adopting the pids subsystem's per-task accounting pattern.
>>>> In cpuset_can_attach(), use task_cs(task) to get the correct source
>>>> cpuset
>>>> for each task (like pids_can_attach uses task_css), adjust
>>>> nr_deadline_tasks
>>>> and reserve DL bandwidth per-task, and increment attach_in_progress
>>>> per-task
>>>> on the destination cpuset. In cpuset_attach(), handle destination
>>>> cpuset
>>>> changes within the task iteration loop.
>>>>
>>>> A shared helper cpuset_undo_attach() reverses the per-task
>>>> operations for
>>>> both partial rollback in cpuset_can_attach() and full reversal in
>>>> cpuset_cancel_attach().
>>>>
>>>> When multiple source cpusets are detected in can_attach(), set
>>>> attach_many_sources so that cpuset_attach() forces cpus_updated and
>>>> mems_updated to true, ensuring all tasks get properly updated
>>>> regardless
>>>> of which source cpuset cpuset_attach_old_cs points to.
>>>>
>>>> This eliminates the need for nr_migrate_dl_tasks, sum_migrate_dl_bw,
>>>> and
>>>> dl_bw_cpu fields in struct cpuset.
>>>>
>>>> Fixes: 4ec22e9c5a90 ("cpuset: Enable cpuset controller in default
>>>> hierarchy")
>>>> Signed-off-by: Ridong Chen <ridong.chen@xxxxxxxxx>
>>> It is not a problem doing per-task DL BW allocation and eliminating the
>>> *dl_bw* fields. However, updating nr_deadline_tasks before it is
>>> committed can be problematic.
>>>
>> Good to hear that.
>>
>>> nr_deadline_tasks is used in dl_rebuild_rd_accounting() which is called
>>> by partition_sched_domains_locked(). After the release of cpuset_mutex
>>> at the end of cpuset_can_attach() and before cpuset_attach() or
>>> cpuset_cancel_attach() is called, it is possible
>>> that partition_sched_domains_locked() can be called
>>> and dl_rebuild_rd_accounting() is not getting the right DL BW accounting
>>> information. So unless there is a way to confirm that this situation
>>> cannot happen, we can't change nr_deadline_tasks before the attach is
>>> commited.
>>>
>> We can keep the nr_migrate_dl_tasks field and update nr_deadline_tasks
>> once migration is complete. I think this will be much simpler than
>> fixing the issue using lists.
>>
> But we still need to track the set of source and destination cpusets to
> commit or cancel the change. Doing it task-by-task will add code in the
> cpuset_attach() and cpuset_cancel_attach() to check if a task is a DL
> task and act accordingly. So we are just trading task-by-task code with
> code to handle the lists.
> 

I resend a patch [1] that keeps nr_migrate_dl_tasks but eliminates
sum_migrate_dl_bw, dl_bw_cpu, attach_node, and attach_llist_head from
the cpuset structure by doing per-task dl_bw_alloc directly in
cpuset_can_attach().

I just offer a way to discuss whether we can solve this issue in a
simpler way.

[1]
https://lore.kernel.org/cgroups/20260602023203.248077-1-longman@xxxxxxxxxx/T/#mb2c6a3ae44f34f571db5dffa888212eaeaaea17a
-- 
Best regards,
Ridong