From: Al Viro Date: Mon, 27 Apr 2020 10:26:22 -0400 Subject: propagate_one(): mnt_set_mountpoint() needs mount_lock Origin: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit?id=fa87bf609aa173b5dce91d23cd3dcebd9e846124 commit b0d3869ce9eeacbb1bbd541909beeef4126426d5 upstream. ... to protect the modification of mp->m_count done by it. Most of the places that modify that thing also have namespace_lock held, but not all of them can do so, so we really need mount_lock here. Kudos to Piotr Krysiuk , who'd spotted a related bug in pivot_root(2) (fixed unnoticed in 5.3); search for other similar turds has caught out this one. Cc: stable@kernel.org Signed-off-by: Al Viro Signed-off-by: Piotr Krysiuk Signed-off-by: Greg Kroah-Hartman --- fs/pnode.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/fs/pnode.c b/fs/pnode.c index 53d411a371ce..7910ae91f17e 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -266,14 +266,13 @@ static int propagate_one(struct mount *m) if (IS_ERR(child)) return PTR_ERR(child); child->mnt.mnt_flags &= ~MNT_LOCKED; + read_seqlock_excl(&mount_lock); mnt_set_mountpoint(m, mp, child); + if (m->mnt_master != dest_master) + SET_MNT_MARK(m->mnt_master); + read_sequnlock_excl(&mount_lock); last_dest = m; last_source = child; - if (m->mnt_master != dest_master) { - read_seqlock_excl(&mount_lock); - SET_MNT_MARK(m->mnt_master); - read_sequnlock_excl(&mount_lock); - } hlist_add_head(&child->mnt_hash, list); return count_mounts(m->mnt_ns, child); }