Re: [patch v8 4/9] sched: fix slept time double counting in enqueue entity

[Alex Shi <seakeel@gmail.com>]

>>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>>> index c61a614..9640c66 100644
>>> --- a/kernel/sched/fair.c
>>> +++ b/kernel/sched/fair.c
>>> @@ -5856,7 +5856,8 @@ static void switched_from_fair(struct rq *rq,
>>> struct task_struct *p)
>>>         */
>>>         if (p->se.avg.decay_count) {
>>>                 struct cfs_rq *cfs_rq = cfs_rq_of(&p->se);
>>> -               __synchronize_entity_decay(&p->se);
>>> +               p->se.avg.last_runnable_update +=
>>> +                               __synchronize_entity_decay(&p->se);
>>
>> it is not needed, since the last_runnable_update will be reset if it
>> will be switched to fair.
> 
> When it is backed to fair class, the only thing scheduler would do
> is to check whether it could preempt current running task, so
> where to reset the last_runnable_update?

here, when the task enqueue to cfs_rq.
avg.decay_count == 0, so the last line will reset last_runnable_update.

Anyway, you may consider to add a comment in the function, since it is bit hard
to understand for you.

---
/* Add the load generated by se into cfs_rq's child load-average */
static inline void enqueue_entity_load_avg(struct cfs_rq *cfs_rq,
                                                  struct sched_entity *se,
                                                  int wakeup)
{
        /*
         * We track migrations using entity decay_count <= 0, on a wake-up
         * migration we use a negative decay count to track the remote decays
         * accumulated while sleeping.
         */
        if (unlikely(se->avg.decay_count <= 0)) {
                se->avg.last_runnable_update = rq_of(cfs_rq)->clock_task;

> 
> Thanks,
> Lei
> 
>>
>> --
>> Thanks
>>     Alex


-- 
Thanks
    Alex