oss-sec mailing list archives

Re: Re: CVE request Linux kernel: ns: user namespaces panic


From: ebiederm () xmission com (Eric W. Biederman)
Date: Fri, 05 Jun 2015 12:31:23 -0500

P J P <ppandit () redhat com> writes:

+-- On Thu, 4 Jun 2015, cve-assign () mitre org wrote --+
| >> From: ebiederm () xmission com (Eric W. Biederman)
| >> 
| >> The core issue is that a unprivileged user could call umount(MNT_DETACH)
| >> and in the right circumstances gain access to every file on essentially
| >> any filesystem in the mount namespace.
| >> 
| >> e0c9c0afd2fc958ffa34b697972721d81df8a56f mnt: Update detach_mounts to leave mounts connected
| >> is the real bug fix that fixes a fairly scary issue.
| >>
| >> So I believe cd4a40174b71acd021877341684d8bb1dc8ea4ae prevents a
| >> difficult to trigger crash if you have
| >> e0c9c0afd2fc958ffa34b697972721d81df8a56f applied.
| >>
|
| Use CVE-2015-4176 for the issue fixed in
| e0c9c0afd2fc958ffa34b697972721d81df8a56f . This code change is present
| in 4.0.2.
| 
| Use CVE-2015-4177 for the issue fixed in
| cd4a40174b71acd021877341684d8bb1dc8ea4ae. This code change is not
| present in 4.0.2.

  Ie IIUC, existence of CVE-2015-4177 depends on CVE-2015-4176 being fixed, 
which in turn depends on commit

  -> https://git.kernel.org/linus/ce07d891a0891d3c0d0c2d73d577490486b809e1

being present. (...Eric?)

There are no dependencies between
e0c9c0afd2fc958ffa34b697972721d81df8a56f and
cd4a40174b71acd021877341684d8bb1dc8ea4ae

Neither commit (cec9c... or cd4a4017...) missing is a problem unless
ce07d891a0891d3c0d0c2d73d577490486b809e1 is present.

Commits (cec9c... and cd4a4017...) are harmless on their own.

I hope that clarifies things.

Eric


Current thread: