Re: [PATCH v2 8/9] sched: Don't reinvent the wheel but use existing llist API

From: Byungchul Park
Date: Mon Feb 13 2017 - 18:00:17 EST


On Mon, Feb 13, 2017 at 11:04:57AM +0100, Peter Zijlstra wrote:
> On Mon, Feb 13, 2017 at 04:21:08PM +0900, Byungchul Park wrote:
> > Although llist provides proper APIs, they are not used. Make them used.
> >
> > Signed-off-by: Byungchul Park <byungchul.park@xxxxxxx>
> > ---
> > kernel/sched/core.c | 13 ++-----------
> > 1 file changed, 2 insertions(+), 11 deletions(-)
> >
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index d01f9d0..417060b 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -1783,17 +1783,8 @@ void sched_ttwu_pending(void)
> > raw_spin_lock_irqsave(&rq->lock, flags);
> > rq_pin_lock(rq, &rf);
> >
> > - while (llist) {
> > - int wake_flags = 0;
> > -
> > - p = llist_entry(llist, struct task_struct, wake_entry);
> > - llist = llist_next(llist);
> > -
> > - if (p->sched_remote_wakeup)
> > - wake_flags = WF_MIGRATED;
> > -
> > - ttwu_do_activate(rq, p, wake_flags, &rf);
> > - }
> > + llist_for_each_entry(p, llist, wake_entry)
> > + ttwu_do_activate(rq, p, p->sched_remote_wakeup ? WF_MIGRATED : 0, &rf);
>
> I think this suffers the exact same problem the others did. After
> ttwu_do_activate() the llist entry can be reused, so doing list_next()
> after it is flaky.

Indeed. I thought it's safe since it's within rq locked and cannot be
reused. But I was wrong but it can be unlocked in ttwu_do_activate().
I will fix it. Thank you very much.