Re: [PATCH V2] workqueue: fix possible race condition when rescuer VS pwq-release

[Lai Jiangshan <laijs@cn.fujitsu.com>]

On 04/17/2014 12:50 AM, Tejun Heo wrote:
> Hello, Lai.
> 
> On Thu, Apr 17, 2014 at 12:21:21AM +0800, Lai Jiangshan wrote:
>> OK. It is better to use get_pwq(). I will also change the above comments to:
>>
>>   The base ref and the possible ref from rerscuer(stopped) are never
>> dropped on per-cpu pwqs.
>>   Directly free the pwqs and wq.
> 
> Hmmm, why does that matter?  The base ref is the business of
> create/destroy path and as long as base ref is never put other refs
> just don't matter.
> 
>> The reason I quickly dropped V1 and wrote the V2 is that I saw this comment.
>> "The base ref" is precise after I used get_pwq() in V1.
>>
>> Or to avoid to change to this comments.
>> I can also move the following code down to the bottom of the rescuer_thread().
>>
>> if (kthread_should_stop()) {
>> __set_current_state(TASK_RUNNING);
>> rescuer->task->flags &= ~PF_WQ_WORKER;
>> return 0;
>> }
>>
>> (I reply this email on browser, never mind the syntax).
>> Maybe the second choice are better.
> 
> Hmmm... yeah, regardlesss of the above, it's a bit nasty that rescuer
> may exit with non-empty mayday list.
> 

+		 * The wq->maydays list maybe still have some pwqs linked,
+		 * but it is safe to free them all together since the rescuer
+		 * is stopped.

I just comment it without fix. It has no harm before, but it harms after I added a get_pwq().
Sorry.
Lai