Re: [PATCH 2/2] bdi: Avoid oops on device removal

[Jan Kara <jack@suse.cz>]

On Thu 27-02-14 15:07:48, Tejun Heo wrote:
> Hello,
> 
> On Tue, Feb 25, 2014 at 11:29:14PM +0100, Jan Kara wrote:
> > +static void bdi_wakeup_thread(struct backing_dev_info *bdi)
> > +{
> > +	spin_lock_bh(&bdi->wb_lock);
> > +	if (test_bit(BDI_registered, &bdi->state))
> > +		mod_delayed_work(bdi_wq, &bdi->wb.dwork, 0);
> > +	spin_unlock_bh(&bdi->wb_lock);
> > +}
> 
> I wonder whether this can be smarter without requiring wb_lock each
> timer but this probably is the simplest for -stable backports.
  We could be clever, check whether the work is already queued for
execution and bail out without taking wb_lock if yes (that would also
save us some unnecessary juggling in try_to_grab_pending() for the situation
were the work is already queued). But I'm not sure how to cleanly implement
this...

> >  static void bdi_queue_work(struct backing_dev_info *bdi,
> >  			   struct wb_writeback_work *work)
> >  {
> >  	trace_writeback_queue(bdi, work);
> >  
> >  	spin_lock_bh(&bdi->wb_lock);
> > +	if (!test_bit(BDI_registered, &bdi->state)) {
> > +		if (work->done)
> > +			complete(work->done);
> > +		goto out_unlock;
> > +	}
> >  	list_add_tail(&work->list, &bdi->work_list);
> > -	spin_unlock_bh(&bdi->wb_lock);
> > -
> >  	mod_delayed_work(bdi_wq, &bdi->wb.dwork, 0);
> > +out_unlock:
> > +	spin_unlock_bh(&bdi->wb_lock);
> >  }
> >  
> > +
> > +
> 
> Why three blank lines?
  A mistake. Will fix.

> Other than that,
> 
> Reviewed-by: Tejun Heo <tj@kernel.org>
  Thanks!

							Honza
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR