Re: [PATCH 2/4] workqueue: Split apply attrs code from its locking

[Frederic Weisbecker <fweisbec@gmail.com>]

On Thu, Apr 24, 2014 at 10:48:32AM -0400, Tejun Heo wrote:
> On Thu, Apr 24, 2014 at 04:37:34PM +0200, Frederic Weisbecker wrote:
> > +static int apply_workqueue_attrs_locked(struct workqueue_struct *wq,
> > +					const struct workqueue_attrs *attrs)
> >  {
> >  	struct workqueue_attrs *new_attrs, *tmp_attrs;
> >  	struct pool_workqueue **pwq_tbl, *dfl_pwq;
> > @@ -3976,15 +3960,6 @@ int apply_workqueue_attrs(struct workqueue_struct *wq,
> >  	copy_workqueue_attrs(tmp_attrs, new_attrs);
> >  
> >  	/*
> > -	 * CPUs should stay stable across pwq creations and installations.
> > -	 * Pin CPUs, determine the target cpumask for each node and create
> > -	 * pwqs accordingly.
> > -	 */
> > -	get_online_cpus();
> > -
> > -	mutex_lock(&wq_pool_mutex);
> 
> lockdep_assert_held()

Not sure... Only a small part of the function actually needs to be locked. Namely
those doing the pwq allocations, which already have the lockdep_assert_held().