Re: Potential problem with 31e77c93e432dec7 ("sched/fair: Update blocked load when newly idle")

[Vincent Guittot]

Hi Niklas,

>> Thanks for the trace, I have been able to catch a problem with it.
>> Could you test the patch below to confirm that the problem is solved ?
>> The patch apply on-top of
>> c18bb396d3d261eb ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net")
>
> I can confirm that with the patch bellow I can no longer produce the
> problem. Thanks!

Thanks for testing
Do you mind if I add
Tested-by: Niklas SÃderlund <niklas.soderlund@xxxxxxxxxxxx>

Peter, Ingo,
Do you want me to re-send the patch with all tags or you will take
this version ?

Regards,
Vincent

>
>>
>> From: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
>> Date: Thu, 26 Apr 2018 12:19:32 +0200
>> Subject: [PATCH] sched/fair: fix the update of blocked load when newly idle
>> MIME-Version: 1.0
>> Content-Type: text/plain; charset=UTF-8
>> Content-Transfer-Encoding: 8bit
>>
>> With commit 31e77c93e432 ("sched/fair: Update blocked load when newly idle"),
>> we release the rq->lock when updating blocked load of idle CPUs. This open
>> a time window during which another CPU can add a task to this CPU's cfs_rq.
>> The check for newly added task of idle_balance() is not in the common path.
>> Move the out label to include this check.
>>
>> Fixes: 31e77c93e432 ("sched/fair: Update blocked load when newly idle")
>> Reported-by: Heiner Kallweit <hkallweit1@xxxxxxxxx>
>> Reported-by: Niklas SÃderlund <niklas.soderlund@xxxxxxxxxxxx>
>> Signed-off-by: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
>> ---
>>  kernel/sched/fair.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 0951d1c..15a9f5e 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -9847,6 +9847,7 @@ static int idle_balance(struct rq *this_rq, struct rq_flags *rf)
>>       if (curr_cost > this_rq->max_idle_balance_cost)
>>               this_rq->max_idle_balance_cost = curr_cost;
>>
>> +out:
>>       /*
>>        * While browsing the domains, we released the rq lock, a task could
>>        * have been enqueued in the meantime. Since we're not going idle,
>> @@ -9855,7 +9856,6 @@ static int idle_balance(struct rq *this_rq, struct rq_flags *rf)
>>       if (this_rq->cfs.h_nr_running && !pulled_task)
>>               pulled_task = 1;
>>
>> -out:
>>       /* Move the next balance forward */
>>       if (time_after(this_rq->next_balance, next_balance))
>>               this_rq->next_balance = next_balance;
>> --
>> 2.7.4
>>
>>
>>
>> [snip]
>>
>
> --
> Regards,
> Niklas SÃderlund