On Thu 22-12-16 18:13:11, Paul Moore wrote: > On Thu, Dec 22, 2016 at 4:15 AM, Jan Kara <jack(a)suse.cz&gt; wrote: > > There are only two calls sites of fsnotify_duplicate_mark(). Those are > > in kernel/audit_tree.c and both are bogus. Vfsmount pointer is unused > > for audit tree, inode pointer and group gets set in > > fsnotify_add_mark_locked() later anyway, mask and free_mark are already > > set in alloc_chunk(). In fact, calling fsnotify_duplicate_mark() is > > actively harmful because following fsnotify_add_mark_locked() will leak > > group reference by overwriting the group pointer. So just remove the two > > calls to fsnotify_duplicate_mark() and the function. > > > > Signed-off-by: Jan Kara <jack(a)suse.cz&gt; > > --- > > fs/notify/mark.c | 12 ------------ > > include/linux/fsnotify_backend.h | 2 -- > > kernel/audit_tree.c | 6 ++---- > > 3 files changed, 2 insertions(+), 18 deletions(-) > > At first glance this looks reasonable, although you keep mentioning > "fsnotify_add_mark_locked" above when untag_chunk() is calling > "fsnotify_add_mark"; I just wanted to make sure you hadn't intended to > take the mutex in the audit code instead of relying on the locking in > fsnotify_add_mark(). No, I didn't want to take mutex in the audit code. It is just that fsnotify_add_mark() is a thin wrapper around fsnotify_add_mark_locked() so I was speaking about that function.

> > > diff --git a/fs/notify/mark.c b/fs/notify/mark.c > > index d3fea0bd89e2..6043306e8e21 100644 > > --- a/fs/notify/mark.c > > +++ b/fs/notify/mark.c > > @@ -510,18 +510,6 @@ void fsnotify_detach_group_marks(struct fsnotify_group *group) > > } > > } > > > > -void fsnotify_duplicate_mark(struct fsnotify_mark *new, struct fsnotify_mark *old) > > -{ > > - assert_spin_locked(&old->lock); > > - new->inode = old->inode; > > - new->mnt = old->mnt; > > - if (old->group) > > - fsnotify_get_group(old->group); > > - new->group = old->group; > > - new->mask = old->mask; > > - new->free_mark = old->free_mark; > > -} > > - > > /* > > * Nothing fancy, just initialize lists and locks and counters. > > */ > > diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h > > index 0cf34d6cc253..487246546ebe 100644 > > --- a/include/linux/fsnotify_backend.h > > +++ b/include/linux/fsnotify_backend.h > > @@ -323,8 +323,6 @@ extern void fsnotify_init_mark(struct fsnotify_mark *mark, void (*free_mark)(str > > extern struct fsnotify_mark *fsnotify_find_inode_mark(struct fsnotify_group *group, struct inode *inode); > > /* find (and take a reference) to a mark associated with group and vfsmount */ > > extern struct fsnotify_mark *fsnotify_find_vfsmount_mark(struct fsnotify_group *group, struct vfsmount *mnt); > > -/* copy the values from old into new */ > > -extern void fsnotify_duplicate_mark(struct fsnotify_mark *new, struct fsnotify_mark *old); > > /* set the ignored_mask of a mark */ > > extern void fsnotify_set_mark_ignored_mask_locked(struct fsnotify_mark *mark, __u32 mask); > > /* set the mask of a mark (might pin the object into memory */ > > diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c > > index 8b1dde96a0fa..f3130eb0a4bd 100644 > > --- a/kernel/audit_tree.c > > +++ b/kernel/audit_tree.c > > @@ -258,8 +258,7 @@ static void untag_chunk(struct node *p) > > if (!new) > > goto Fallback; > > > > - fsnotify_duplicate_mark(&new->mark, entry); > > - if (fsnotify_add_mark(&new->mark, new->mark.group, new->mark.inode, NULL, 1)) { > > + if (fsnotify_add_mark(&new->mark, entry->group, entry->inode, NULL, 1)) { > > fsnotify_put_mark(&new->mark); > > goto Fallback; > > } > > @@ -395,8 +394,7 @@ static int tag_chunk(struct inode *inode, struct audit_tree *tree) > > return -ENOENT; > > } > > > > - fsnotify_duplicate_mark(chunk_entry, old_entry); > > - if (fsnotify_add_mark(chunk_entry, chunk_entry->group, chunk_entry->inode, NULL, 1)) { > > + if (fsnotify_add_mark(chunk_entry, old_entry->group, old_entry->inode, NULL, 1)) { > > spin_unlock(&old_entry->lock); > > fsnotify_put_mark(chunk_entry); > > fsnotify_put_mark(old_entry); > > -- > > 2.10.2 > > > > -- > paul moore > www.paul-moore.com -- Jan Kara <jack(a)suse.com&gt; SUSE Labs, CR