On Fri, May 18, 2018 at 11:27 PM, Tejun Heo <tj@xxxxxxxxxx> wrote:
> Hello, Lai.
>
> On Thu, May 17, 2018 at 09:50:00PM +0800, Lai Jiangshan wrote:
>> > +void wq_worker_comm(char *buf, size_t size, struct task_struct *task)
>> > +{
>> > + struct worker *worker;
>> > + struct worker_pool *pool;
>> > +
>> > + mutex_lock(&wq_pool_attach_mutex);
>> > +
>>
>> Need to recheck
>> task->flags | PF_WQ_WORKER
>> again, because @task is often not the current task,
>> the worker returned by kthread_data(task)
>> could have been freed when the worker
>> was destroyed.
>
> Hmm... but we're pinning the task and kthread_data doesn't get freed
> till the task is released. I don't see how the worker could go away.
>
pinning the task doesn't stop the kthread_data from
getting freed, especially when worker_thread()
free kthread_data by itself.
Even for rescuer_thread, it is hard to say if
the kthread_data is being protected.
> ...
>> BTW, is it possible to keep attach_mutex
>> unchanged? rcu_sched might be enough.
>
> It's simpler to use a global mutex, no?
>
> Thanks.
>
> --
> tejun
--
To unsubscribe from this list: send the line "unsubscribe linux-api" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
