Re: [PATCH 1/2] sched: Rework migrate_tasks()

[Kirill Tkhai <tkhai@yandex.ru>]

11.06.2014, 17:15, "Srikar Dronamraju" <srikar@linux.vnet.ibm.com>:
>>>  * Kirill Tkhai <ktkhai@parallels.com> [2014-06-11 13:52:10]:
>>>>   Currently migrate_tasks() skips throttled tasks,
>>>>   because they are not pickable by pick_next_task().
>>>  Before migrate_tasks() is called, we do call set_rq_offline(), in
>>>  migration_call().
>>>
>>>  Shouldnt this take care of unthrottling the tasks and making sure that
>>>  they can be picked by pick_next_task().
>>  If we do this separate for every class, we'll have to do this 3 times.
>>  Furthermore, deadline class does not have a list of throttled tasks.
>>  So we'll have to the same as I did: to lock tasklist_lock and to iterate
>>  throw all of the tasks in the system just to found deadline tasks.
>
> I think you misread my comment.
>
> Currently migrate_task() gets called from migration_call() and in the
> migration_call() before migrate_tasks(), set_rq_offline() should put
> tasks back using unthrottle_cfs_rq().
>
> So my question is: Why are these tasks not getting unthrottled
> through we are calling set_rq_offline? To me set_rq_offline is
> calling the actual sched class routines to do the needful.
>
> I can understand about deadline tasks, because we don't have a deadline
> But thats the only tasks that we need to fix.

Hm, I tested that on fair class tasks. They used to disappear from
/proc/sched_debug and used to hang. I'll check all once again.

I'm agree with you, if set_rq_offline() already presents, we should use it.

/me went to clarify why it does not work in my test.

>>>>   These tasks stay on dead cpu even after they
>>>>   becomes unthrottled. They are not schedulable
>>>>   till user manually changes their affinity or till
>>>>   cpu becomes alive again.
>>>  If we are still seeing tasks not being picked by pick_next_task(), then
>>>  can it probably mean that rq->rd was NULL?
>>  Unthrottle functions dl_task_timer() and unthrottle_cfs_rq() put tasks and
>>  queues back. They do not look at rq->rd.
>
> What I meant was only if rq->rd isn't set, then we don't call
> set_rq_offline, which seems very reasonable.
> --
> Thanks and Regards
> Srikar Dronamraju