Re: [PATCH] fsnotify: Cleanup spinlock assertions

[Jan Kara <jack@suse.cz>]

On Thu 22-09-16 13:45:28, Jeff Layton wrote:
> On Thu, 2016-09-22 at 11:43 +0200, Jan Kara wrote:
> > Use assert_spin_locked() macro instead of hand-made BUG_ON statements.
> > 
> > Suggested-by: Heiner Kallweit <hkallweit1@gmail.com>
> > Signed-off-by: Jan Kara <jack@suse.cz>
> > ---
> >  fs/notify/fanotify/fanotify_user.c | 3 +--
> >  fs/notify/notification.c           | 9 +++------
> >  2 files changed, 4 insertions(+), 8 deletions(-)
> > 
> > Andrew, can you please add this cleanup to the fanotify patches you carry?
> > Thanks!
> > 
> > diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c
> > index 189fab3ac4e6..7ebfca6a1427 100644
> > --- a/fs/notify/fanotify/fanotify_user.c
> > +++ b/fs/notify/fanotify/fanotify_user.c
> > @@ -54,8 +54,7 @@ struct kmem_cache *fanotify_perm_event_cachep __read_mostly;
> >  static struct fsnotify_event *get_one_event(struct fsnotify_group *group,
> >  					    size_t count)
> >  {
> > -	BUG_ON(IS_ENABLED(CONFIG_SMP) &&
> > -	       !spin_is_locked(&group->notification_lock));
> > +	assert_spin_locked(&group->notification_lock);
> >  
> >  	pr_debug("%s: group=%p count=%zd\n", __func__, group, count);
> >  
> > diff --git a/fs/notify/notification.c b/fs/notify/notification.c
> > index 1a8010e7a2a0..66f85c651c52 100644
> > --- a/fs/notify/notification.c
> > +++ b/fs/notify/notification.c
> > @@ -63,8 +63,7 @@ EXPORT_SYMBOL_GPL(fsnotify_get_cookie);
> >  /* return true if the notify queue is empty, false otherwise */
> >  bool fsnotify_notify_queue_is_empty(struct fsnotify_group *group)
> >  {
> > -	BUG_ON(IS_ENABLED(CONFIG_SMP) &&
> > -	       !spin_is_locked(&group->notification_lock));
> > +	assert_spin_locked(&group->notification_lock);
> >  	return list_empty(&group->notification_list) ? true : false;
> >  }
> >  
> > @@ -149,8 +148,7 @@ struct fsnotify_event *fsnotify_remove_first_event(struct fsnotify_group *group)
> >  {
> >  	struct fsnotify_event *event;
> >  
> > -	BUG_ON(IS_ENABLED(CONFIG_SMP) &&
> > -	       !spin_is_locked(&group->notification_lock));
> > +	assert_spin_locked(&group->notification_lock);
> >  
> >  	pr_debug("%s: group=%p\n", __func__, group);
> >  
> > @@ -172,8 +170,7 @@ struct fsnotify_event *fsnotify_remove_first_event(struct fsnotify_group *group)
> >   */
> >  struct fsnotify_event *fsnotify_peek_first_event(struct fsnotify_group *group)
> >  {
> > -	BUG_ON(IS_ENABLED(CONFIG_SMP) &&
> > -	       !spin_is_locked(&group->notification_lock));
> > +	assert_spin_locked(&group->notification_lock);
> >  
> >  	return list_first_entry(&group->notification_list,
> >  				struct fsnotify_event, list);
> 
> Much cleaner.
> 
> That said, I have a personal preference for lockdep_assert_held() in
> these situations, which not only tells you whether the lock is locked,
> but (I believe) whether it was locked by the current task as well.

Yes, it does.

> Theoretically you could have a different task take this spinlock, and
> then call into here without holding it and not get the assertion since
> it was locked at the time. Of course, that does require lockdep...

Yeah, I personally don't have a strong preference. Both have advantages and
disadvantages - as you said, lockdep_assert_held() is reliable when lockdep
is enabled but there's much less testing happening with lockdep enabled and
also lockdep changes the timing enough that some cases just need not
trigger...
 
> In any case, this is still an improvement:
> 
> Reviewed-by: Jeff Layton <jlayton@redhat.com>

Thanks!

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