Re: [patch, 2.6.10-rc2] sched: fix ->nr_uninterruptible handling bugs

[Peter Williams <pwil3058@bigpond.net.au>]

Ingo Molnar wrote:
> PREEMPT_RT on SMP systems triggered weird (very high) load average
> values rather easily, which turned out to be a mainline kernel
> ->nr_uninterruptible handling bug in try_to_wake_up().
> 
> the following code:
> 
>         if (old_state == TASK_UNINTERRUPTIBLE) {
>                 old_rq->nr_uninterruptible--;
> 
> potentially executes with old_rq potentially being != rq, and hence
> updating ->nr_uninterruptible without the lock held. Given a
> sufficiently concurrent preemption workload the count can get out of
> whack and updates might get lost, permanently skewing the global count. 
> Nothing except the load-average uses nr_uninterruptible() so this
> condition can go unnoticed quite easily.
> 
> the fix is to update ->nr_uninterruptible always on the runqueue where
> the task currently is. (this is also a tiny performance plus for
> try_to_wake_up() as a stackslot gets freed up.)

Couldn't this part of the problem have been solved by using an atomic_t 
for nr_uninterruptible as for nr_iowait?  It would also remove the need 
for migrate_nr_uninterruptible().

Peter
-- 
Peter Williams                                   pwil3058@bigpond.net.au

"Learning, n. The kind of ignorance distinguishing the studious."
  -- Ambrose Bierce