Re: [patch 12/15] sched: maintain throttled rqs as a list

[Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>]

Hi Paul,

(2011/03/23 12:03), Paul Turner wrote:
> @@ -1434,7 +1435,11 @@ static void throttle_cfs_rq(struct cfs_r
>  	if (!se)
>  		rq->nr_running += task_delta;
>  
> +	raw_spin_lock(&cfs_b->lock);
>  	cfs_rq->throttled = 1;
> +	list_add_tail_rcu(&cfs_rq->throttled_list, &cfs_b->throttled_cfs_rq);
> +	raw_spin_unlock(&cfs_b->lock);
> +
>  	cfs_rq->throttled_timestamp = rq_of(cfs_rq)->clock;
>  }
>  

Though I'm looking forward to see your updates soon, just FYI,
I found that somehow throttle_cfs_rq() was called with already
throttled cfs_rq.

So it breaks list throttled_cfs_rq by repeating list_add_tail_rcu()
and results in lockups like that Bharata and Xiao already reported.

There should be better fix, but following first aid to put_prev_entity()
worked for me:

-       if (!within_bandwidth(cfs_rq))
+       if (!within_bandwidth(cfs_rq) && !cfs_rq->throttled)
                throttle_cfs_rq(cfs_rq);
        else
                return_cfs_rq_quota(cfs_rq);

I believe you can do better next time.


Thanks,
H.Seto