From: Al Viro <viro@zeniv.linux.org.uk>
Date: Fri, 23 May 2025 23:20:36 +0000 (-0400)
Subject: Don't propagate mounts into detached trees
X-Git-Tag: v6.16-rc1~106^2
X-Git-Url: https://git.kernel.dk/?a=commitdiff_plain;h=3b5260d12b1f;p=linux-block.git

Don't propagate mounts into detached trees

All versions up to 6.14 did not propagate mount events into detached
tree.  Shortly after 6.14 a merge of vfs-6.15-rc1.mount.namespace
(130e696aa68b) has changed that.

Unfortunately, that has caused userland regressions (reported in
https://lore.kernel.org/all/CAOYeF9WQhFDe+BGW=Dp5fK8oRy5AgZ6zokVyTj1Wp4EUiYgt4w@mail.gmail.com/)

Straight revert wouldn't be an option - in particular, the variant in 6.14
had a bug that got fixed in d1ddc6f1d9f0 ("fix IS_MNT_PROPAGATING uses")
and we don't want to bring the bug back.

This is a modification of manual revert posted by Christian, with changes
needed to avoid reintroducing the breakage in scenario described in
d1ddc6f1d9f0.

Cc: stable@vger.kernel.org
Reported-by: Allison Karlitskaya <lis@redhat.com>
Tested-by: Allison Karlitskaya <lis@redhat.com>
Acked-by: Christian Brauner <brauner@kernel.org>
Co-developed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---

diff --git a/fs/mount.h b/fs/mount.h
index 7aecf2a60472..ad7173037924 100644
--- a/fs/mount.h
+++ b/fs/mount.h
@@ -7,10 +7,6 @@
 
 extern struct list_head notify_list;
 
-typedef __u32 __bitwise mntns_flags_t;
-
-#define MNTNS_PROPAGATING	((__force mntns_flags_t)(1 << 0))
-
 struct mnt_namespace {
 	struct ns_common	ns;
 	struct mount *	root;
@@ -37,7 +33,6 @@ struct mnt_namespace {
 	struct rb_node		mnt_ns_tree_node; /* node in the mnt_ns_tree */
 	struct list_head	mnt_ns_list; /* entry in the sequential list of mounts namespace */
 	refcount_t		passive; /* number references not pinning @mounts */
-	mntns_flags_t		mntns_flags;
 } __randomize_layout;
 
 struct mnt_pcp {
diff --git a/fs/namespace.c b/fs/namespace.c
index 1b466c54a357..623cd110076d 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -3648,7 +3648,7 @@ static int do_move_mount(struct path *old_path,
 	if (!(attached ? check_mnt(old) : is_anon_ns(ns)))
 		goto out;
 
-	if (is_anon_ns(ns)) {
+	if (is_anon_ns(ns) && ns == p->mnt_ns) {
 		/*
 		 * Ending up with two files referring to the root of the
 		 * same anonymous mount namespace would cause an error
@@ -3656,16 +3656,7 @@ static int do_move_mount(struct path *old_path,
 		 * twice into the mount tree which would be rejected
 		 * later. But be explicit about it right here.
 		 */
-		if ((is_anon_ns(p->mnt_ns) && ns == p->mnt_ns))
-			goto out;
-
-		/*
-		 * If this is an anonymous mount tree ensure that mount
-		 * propagation can detect mounts that were just
-		 * propagated to the target mount tree so we don't
-		 * propagate onto them.
-		 */
-		ns->mntns_flags |= MNTNS_PROPAGATING;
+		goto out;
 	} else if (is_anon_ns(p->mnt_ns)) {
 		/*
 		 * Don't allow moving an attached mount tree to an
@@ -3722,8 +3713,6 @@ static int do_move_mount(struct path *old_path,
 	if (attached)
 		put_mountpoint(old_mp);
 out:
-	if (is_anon_ns(ns))
-		ns->mntns_flags &= ~MNTNS_PROPAGATING;
 	unlock_mount(mp);
 	if (!err) {
 		if (attached) {
diff --git a/fs/pnode.c b/fs/pnode.c
index fb77427df39e..ffd429b760d5 100644
--- a/fs/pnode.c
+++ b/fs/pnode.c
@@ -231,8 +231,8 @@ static int propagate_one(struct mount *m, struct mountpoint *dest_mp)
 	/* skip if mountpoint isn't visible in m */
 	if (!is_subdir(dest_mp->m_dentry, m->mnt.mnt_root))
 		return 0;
-	/* skip if m is in the anon_ns we are emptying */
-	if (m->mnt_ns->mntns_flags & MNTNS_PROPAGATING)
+	/* skip if m is in the anon_ns */
+	if (is_anon_ns(m->mnt_ns))
 		return 0;
 
 	if (peers(m, last_dest)) {