fs: drop peer group ids under namespace lock

commit cb2239c198ad9fbd5aced22cf93e45562da781eb upstream.

When cleaning up peer group ids in the failure path we need to make sure
to hold on to the namespace lock. Otherwise another thread might just
turn the mount from a shared into a non-shared mount concurrently.

Link: https://lore.kernel.org/lkml/00000000000088694505f8132d77@google.com
Fixes: 2a1867219c7b ("fs: add mount_setattr()")
Reported-by: syzbot+8ac3859139c685c4f597@syzkaller.appspotmail.com
Cc: stable@vger.kernel.org # 5.12+
Message-Id: <20230330-vfs-mount_setattr-propagation-fix-v1-1-37548d91533b@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

fs/namespace.c

diff --git a/fs/namespace.c b/fs/namespace.c
index c9b2983..1a9df6a 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -4121,9 +4121,9 @@ static int do_mount_setattr(struct path *path, struct mount_kattr *kattr)
 	unlock_mount_hash();
 
 	if (kattr->propagation) {
-		namespace_unlock();
 		if (err)
 			cleanup_group_ids(mnt, NULL);
+		namespace_unlock();
 	}
 
 	return err;