mnt: Clarify and correct the disconnect logic in umount_tree

rmdir mntpoint will result in an infinite loop when there is
a mount locked on the mountpoint in another mount namespace.

This is because the logic to test to see if a mount should
be disconnected in umount_tree is buggy.

Move the logic to decide if a mount should remain connected to
it's mountpoint into it's own function disconnect_mount so that
clarity of expression instead of terseness of expression becomes
a virtue.

When the conditions where it is invalid to leave a mount connected
are first ruled out, the logic for deciding if a mount should
be disconnected becomes much clearer and simpler.

Fixes: e0c9c0afd2fc958ffa34b697972721d81df8a56f mnt: Update detach_mounts to leave mounts connected
Fixes: ce07d891a0891d3c0d0c2d73d577490486b809e1 mnt: Honor MNT_LOCKED when detaching mounts
Cc: stable@vger.kernel.org
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>

1 files changed, 31 insertions, 4 deletions

diff --git a/fs/namespace.c b/fs/namespace.c
index c7cb8a526c05..7829eab0b706 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1361,6 +1361,36 @@ enum umount_tree_flags {
 	UMOUNT_PROPAGATE = 2,
 	UMOUNT_CONNECTED = 4,
 };
+
+static bool disconnect_mount(struct mount *mnt, enum umount_tree_flags how)
+{
+	/* Leaving mounts connected is only valid for lazy umounts */
+	if (how & UMOUNT_SYNC)
+		return true;
+
+	/* A mount without a parent has nothing to be connected to */
+	if (!mnt_has_parent(mnt))
+		return true;
+
+	/* Because the reference counting rules change when mounts are
+	 * unmounted and connected, umounted mounts may not be
+	 * connected to mounted mounts.
+	 */
+	if (!(mnt->mnt_parent->mnt.mnt_flags & MNT_UMOUNT))
+		return true;
+
+	/* Has it been requested that the mount remain connected? */
+	if (how & UMOUNT_CONNECTED)
+		return false;
+
+	/* Is the mount locked such that it needs to remain connected? */
+	if (IS_MNT_LOCKED(mnt))
+		return false;
+
+	/* By default disconnect the mount */
+	return true;
+}
+
 /*
  * mount_lock must be held
  * namespace_sem must be held for write
@@ -1398,10 +1428,7 @@ static void umount_tree(struct mount *mnt, enum umount_tree_flags how)
 		if (how & UMOUNT_SYNC)
 			p->mnt.mnt_flags |= MNT_SYNC_UMOUNT;
 
-		disconnect = !(((how & UMOUNT_CONNECTED) &&
-				mnt_has_parent(p) &&
-				(p->mnt_parent->mnt.mnt_flags & MNT_UMOUNT)) ||
-			       IS_MNT_LOCKED_AND_LAZY(p));
+		disconnect = disconnect_mount(p, how);
 
 		pin_insert_group(&p->mnt_umount, &p->mnt_parent->mnt,
 				 disconnect ? &unmounted : NULL);