On Fri 03-04-20 17:15:44, Jules Irenge wrote:
>
>
> On Wed, 1 Apr 2020, Jan Kara wrote:
>
> > On Tue 31-03-20 21:46:38, Jules Irenge wrote:
> > > Sparse reports a warning at fsnotify_finish_user_wait()
> > >
> > > warning: context imbalance in fsnotify_finish_user_wait()
> > > - wrong count at exit
> > >
> > > The root cause is the missing annotation at fsnotify_finish_user_wait()
> > > Add the missing __acquires(&fsnotify_mark_srcu) annotation.
> > >
> > > Signed-off-by: Jules Irenge <jbi.octave@xxxxxxxxx>
> >
> > OK, but then fsnotify_prepare_user_wait() needs __releases annotation as
> > well if we're going to be serious about sparse warnings in this code?
> >
> > Honza
> >
> > > ---
> > > fs/notify/mark.c | 1 +
> > > 1 file changed, 1 insertion(+)
> > >
> > > diff --git a/fs/notify/mark.c b/fs/notify/mark.c
> > > index 1d96216dffd1..44fea637bb02 100644
> > > --- a/fs/notify/mark.c
> > > +++ b/fs/notify/mark.c
> > > @@ -350,6 +350,7 @@ bool fsnotify_prepare_user_wait(struct fsnotify_iter_info *iter_info)
> > > }
> > >
> > > void fsnotify_finish_user_wait(struct fsnotify_iter_info *iter_info)
> > > + __acquires(&fsnotify_mark_srcu)
> > > {
> > > int type;
> > >
> > > --
> > > 2.24.1
> > >
> > --
> > Jan Kara <jack@xxxxxxxx>
> > SUSE Labs, CR
> >
>
> Thanks for the reply. I think adding an annotation at
> fsnotify_prepare_user_wait() will not theoretically remove the warning.
> That's the only reason why I skipped it .
Well, I think the goal isn't really to remove warnings but to make
annotations correct... So even if sparse was not clever enough to spot that
missing annotation, you should add it if you've decided to fix sparse
annotations for fsnotify code.
Honza
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
