fs: add mount namespace to rbtree late

There's no point doing that under the namespace semaphore it just gives
the false impression that it protects the mount namespace rbtree and it
simply doesn't.

Link: https://lore.kernel.org/r/20241213-work-mount-rbtree-lockless-v3-2-6e3cdaf9b280@kernel.org
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>

diff --git a/fs/namespace.c b/fs/namespace.c
index a1832e9a4340bec853d8faec6d2715d9faacfc66..0ca2cda6bc16cfe0f15ec31146b56682b78ebf4d 100644 (file)
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -3981,7 +3981,6 @@ struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns,
                while (p->mnt.mnt_root != q->mnt.mnt_root)
                        p = next_mnt(skip_mnt_tree(p), old);
        }
-       mnt_ns_tree_add(new_ns);
        namespace_unlock();
 
        if (rootmnt)
@@ -3989,6 +3988,7 @@ struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns,
        if (pwdmnt)
                mntput(pwdmnt);
 
+       mnt_ns_tree_add(new_ns);
        return new_ns;
 }