Re: [PATCH] sched: Fix rq->nr_iowait ordering

[Mel Gorman]

On Tue, Nov 17, 2020 at 10:38:29AM +0100, Peter Zijlstra wrote:
> Subject: sched: Fix rq->nr_iowait ordering
> From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Date: Thu, 24 Sep 2020 13:50:42 +0200
> 
>   schedule()				ttwu()
>     deactivate_task();			  if (p->on_rq && ...) // false
> 					    atomic_dec(&task_rq(p)->nr_iowait);
>     if (prev->in_iowait)
>       atomic_inc(&rq->nr_iowait);
> 
> Allows nr_iowait to be decremented before it gets incremented,
> resulting in more dodgy IO-wait numbers than usual.
> 
> Note that because we can now do ttwu_queue_wakelist() before
> p->on_cpu==0, we lose the natural ordering and have to further delay
> the decrement.
> 
> Fixes: Fixes: c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")
> Reported-by: Tejun Heo <tj@xxxxxxxxxx>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>

s/Fixes: Fixes:/Fixes:/

Ok, very minor hazard that the same logic gets duplicated that someone
might try "fix" but git blame should help. Otherwise, it makes sense as
I've received more than one "bug" that complained that a number was larger
than they expected even if no other problem was present so

Acked-by: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>

-- 
Mel Gorman
SUSE Labs