Re: [RFC/PATCH] debug workqueue deadlocks with lockdep

[Ingo Molnar <mingo@elte.hu>]

* Johannes Berg <johannes@sipsolutions.net> wrote:

> > > @@ -257,7 +260,9 @@ static void run_workqueue(struct cpu_wor
> > >
> > >  		BUG_ON(get_wq_data(work) != cwq);
> > >  		work_clear_pending(work);
> > > +		lock_acquire(&cwq->wq->lockdep_map, 0, 0, 0, 2, _THIS_IP_);
> > >  		f(work);
> > > +		lock_release(&cwq->wq->lockdep_map, 0, _THIS_IP_);
> >                                                    ^^^
> > Isn't it better to call lock_release() with nested == 1 ?
> 
> Not sure, Ingo?

well, in this case the lock/unlock should nest perfectly (i.e. it should 
always be balanced perfectly), so indeed calling with nested==1 leads to 
stricter checking.

non-nested unlocks occur when people do stuff like:

	spin_lock(&lock1);
	spin_lock(&lock2);
	spin_unlock(&lock1);
	spin_unlock(&lock2);

the first unlock is not 'nested perfectly'. Now for the workqueue 
dep_map this shouldnt be a legal combination, right?

	Ingo