Merge tag 'fs.idmapped.mount_setattr.v5.13-rc3' of gitolite.kernel.org:pub/scm/linux/kernel/git/brauner/linux

Pull mount_setattr fix from Christian Brauner:
 "This makes an underlying idmapping assumption more explicit.

  We currently don't have any filesystems that support idmapped mounts
  which are mountable inside a user namespace, i.e. where s_user_ns !=
  init_user_ns. That was a deliberate decision for now as userns root
  can just mount the filesystem themselves.

  Express this restriction explicitly and enforce it until there's a
  real use-case for this. This way we can notice it and will have a
  chance to adapt and audit our translation helpers and fstests
  appropriately if we need to support such filesystems"

* tag 'fs.idmapped.mount_setattr.v5.13-rc3' of gitolite.kernel.org:pub/scm/linux/kernel/git/brauner/linux:
  fs/mount_setattr: tighten permission checks

Author: torvalds

fs/namespace.c

@@ -3855,8 +3855,12 @@ static int can_idmap_mount(const struct mount_kattr *kattr, struct mount *mnt)
 	if (!(m->mnt_sb->s_type->fs_flags & FS_ALLOW_IDMAP))
 		return -EINVAL;
 
+	/* Don't yet support filesystem mountable in user namespaces. */
+	if (m->mnt_sb->s_user_ns != &init_user_ns)
+		return -EINVAL;
+
 	/* We're not controlling the superblock. */
-	if (!ns_capable(m->mnt_sb->s_user_ns, CAP_SYS_ADMIN))
+	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
 	/* Mount has already been visible in the filesystem hierarchy. */