Re: [PATCH] sched/fair: Fix RCU stall upon ENOMEM atsched_create_group().

[Peter Zijlstra <peterz@infradead.org>]

On Tue, Jan 06, 2015 at 11:45:29PM +0900, Tetsuo Handa wrote:
> Peter Zijlstra wrote:
> > On Thu, Dec 25, 2014 at 10:10:45PM +0900, Tetsuo Handa wrote:
> > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > > index ef2b104..586ee15 100644
> > > --- a/kernel/sched/fair.c
> > > +++ b/kernel/sched/fair.c
> > > @@ -7756,8 +7756,12 @@ static void task_move_group_fair(struct task_struct *p, int queued)
> > >  void free_fair_sched_group(struct task_group *tg)
> > >  {
> > >  	int i;
> > > +	struct cfs_bandwidth *cfs_b;
> > >  
> > > -	destroy_cfs_bandwidth(tg_cfs_bandwidth(tg));
> > > +	/* Check whether init_cfs_bandwidth() was called. */
> > > +	cfs_b = tg_cfs_bandwidth(tg);
> > > +	if (cfs_b->throttled_cfs_rq.next)
> > > +		destroy_cfs_bandwidth(cfs_b);
> > >  
> > >  	for_each_possible_cpu(i) {
> > >  		if (tg->cfs_rq)
> > 
> > This doesn't actually compile for !CONFIG_CFS_BANDWIDTH, how about:
> > 
> > ---
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -4005,6 +4005,10 @@ void __start_cfs_bandwidth(struct cfs_ba
> >  
> >  static void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
> >  {
> > +	/* init_cfs_bandwidth() was not called */
> > +	if (!cfs_b->throttled_cfs_rq.next)
> > +		return;
> > +
> >  	hrtimer_cancel(&cfs_b->period_timer);
> >  	hrtimer_cancel(&cfs_b->slack_timer);
> >  }
> > 
> I posted "#ifdef CONFIG_CFS_BANDWIDTH" version 7 minutes ago.
> But I'm fine with your patch. Thank you.

Yeah saw that, its what I did first too, then figured the ifdef was ugly
;-)