mnt: Account for MS_RDONLY in fs_fully_visible
authorEric W. Biederman <ebiederm@xmission.com>
Fri, 10 Jun 2016 17:21:40 +0000 (12:21 -0500)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 27 Jul 2016 16:47:28 +0000 (09:47 -0700)
commit6256d2f541fc893083ebebb1369ff41803f6b129
treec2527abdf8e977cef477c0b2feb2434ee7b62fcd
parent57eb6e3dbdd3389a4325b8d5a42203591f64153d
mnt: Account for MS_RDONLY in fs_fully_visible

commit 695e9df010e40f407f4830dc11d53dce957710ba upstream.

In rare cases it is possible for s_flags & MS_RDONLY to be set but
MNT_READONLY to be clear.  This starting combination can cause
fs_fully_visible to fail to ensure that the new mount is readonly.
Therefore force MNT_LOCK_READONLY in the new mount if MS_RDONLY
is set on the source filesystem of the mount.

In general both MS_RDONLY and MNT_READONLY are set at the same for
mounts so I don't expect any programs to care.  Nor do I expect
MS_RDONLY to be set on proc or sysfs in the initial user namespace,
which further decreases the likelyhood of problems.

Which means this change should only affect system configurations by
paranoid sysadmins who should welcome the additional protection
as it keeps people from wriggling out of their policies.

Fixes: 8c6cf9cc829f ("mnt: Modify fs_fully_visible to deal with locked ro nodev and atime")
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/namespace.c