Re: [BUG] fs: inotify_handle_event() reading un-init memory

From: Jan Kara
Date: Wed Jun 17 2015 - 04:13:39 EST


On Tue 16-06-15 22:22:34, Jesper Dangaard Brouer wrote:
>
> On Tue, 16 Jun 2015 15:52:09 +0200 Jan Kara <jack@xxxxxxx> wrote:
>
> > On Tue 16-06-15 13:33:18, Jesper Dangaard Brouer wrote:
> > > Caught by kmemcheck.
> > >
> > > Don't know the fix... just pointed at the bug.
> > >
> > > Introduced in commit 7053aee26a3 ("fsnotify: do not share
> > > events between notification groups").
> > > ---
> > > fs/notify/inotify/inotify_fsnotify.c | 3 ++-
> > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/fs/notify/inotify/inotify_fsnotify.c b/fs/notify/inotify/inotify_fsnotify.c
> > > index 2cd900c2c737..370d66dc4ddb 100644
> > > --- a/fs/notify/inotify/inotify_fsnotify.c
> > > +++ b/fs/notify/inotify/inotify_fsnotify.c
> > > @@ -96,11 +96,12 @@ int inotify_handle_event(struct fsnotify_group *group,
> > > i_mark = container_of(inode_mark, struct inotify_inode_mark,
> > > fsn_mark);
> > >
> > > + // new object alloc here
> > > event = kmalloc(alloc_len, GFP_KERNEL);
> > > if (unlikely(!event))
> > > return -ENOMEM;
> > >
> > > - fsn_event = &event->fse;
> > > + fsn_event = &event->fse; // This looks wrong!?! read from un-init mem?
> >
> > Where is here any read? This is just a pointer arithmetics where we add
> > offset of 'fse' entry to 'event' address.
>
> I was kmemcheck that complained, perhaps it is a false-positive?
May be. What was the kmemcheck warning you saw? And can you also attach
disassembly of inotify_handle_event() from your kernel? Thanks!

Honza
--
Jan Kara <jack@xxxxxxx>
SUSE Labs, CR
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/