Re: [PATCH v3 1/2] mm: compaction: support triggering of proactive compaction by user

[Vlastimil Babka]

On 6/17/21 6:05 PM, Charan Teja Kalla wrote:
> The wait_event/freezable_timeout() documentation says that:
>  * Returns:
>  * 0 if the @condition evaluated to %false after the @timeout elapsed,
> 			or
>  * 1 if the @condition evaluated to %true after the @timeout elapsed,
>  * or the remaining jiffies (at least 1) if the @condition evaluated
>  * to %true before the @timeout elapsed.
> 
> which means the condition must be evaluated to true or timeout should be
> elapsed for the function wait_event_freezable_timeout() to return.
> 
> Please check the macro implementation of __wait_event, where it will be
> in for(;;) till the condition is evaluated to true or timeout happens.
> #define __wait_event_freezable_timeout(wq_head, condition, timeout)
> 
>         ___wait_event(wq_head, ___wait_cond_timeout(condition),
> 
>                       TASK_INTERRUPTIBLE, 0, timeout,
> 
>                       __ret = freezable_schedule_timeout(__ret))
> 
> Thus the plain wakeup of kcompactd don't do the proactive compact work.
> And so we should identify its wakeup for proactive work with a separate
> flag.

OK, you're right, I forgot that the macro has the for loop to guard against
spurious wakeups.