Bug 1392932 (CVE-2016-9386, xsa191)

Summary: CVE-2016-9386 xsa191 xen: x86 null segments not always treated as unusable (XSA-191)
Product: [Other] Security Response Reporter: Adam Mariš <amaris>
Component: vulnerabilityAssignee: Red Hat Product Security <security-response-team>
Status: CLOSED WONTFIX QA Contact:
Severity: medium Docs Contact:
Priority: medium    
Version: unspecifiedCC: security-response-team
Target Milestone: ---Keywords: Security
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2018-03-07 10:20:45 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 1397383    
Bug Blocks: 1392955    
Attachments:
Description Flags
xen-unstable, Xen 4.7.x
none
Xen 4.6.x, Xen 4.5.x, Xen 4.4.x none

Description Adam Mariš 2016-11-08 13:56:02 UTC
ISSUE DESCRIPTION
=================

The Xen x86 emulator erroneously failed to consider the unusability of
segments when performing memory accesses.

The intended behaviour is as follows: The user data segment (%ds, %es,
%fs and %gs) selectors may be NULL in 32-bit to prevent access.  In
64-bit, NULL has a special meaning for user segments, and there is no
way of preventing access.  However, in both 32-bit and 64-bit, a NULL
LDT system segment is intended to prevent access.

On Intel hardware, loading a NULL selector zeros the base as well as most
attributes, but sets the limit field to its largest possible value.  On AMD
hardware, loading a NULL selector zeros the attributes, leaving the stale base
and limit intact.

Xen may erroneously permit the access using unexpected base/limit values.

Ability to exploit this vulnerability on Intel is easy, but on AMD depends in
a complicated way on how the guest kernel manages LDTs.

IMPACT
======

An unprivileged guest user program may be able to elevate its privilege
to that of the guest operating system.

VULNERABLE SYSTEMS
==================

The vulnerability is only exposed to HVM guests.

ARM systems are NOT vulnerable.

All versions of Xen are affected.

However, we believe that the vulnerability cannot be exploited on Xen
4.7 by completely unprivileged guest processes, unless the VM has been
explicitly configured with a non-default cpu vendor string (in xm/xl,
this would be done with a `cpuid=' domain config option).

MITIGATION
==========

Running only PV guests will avoid this issue.

External References:

http://xenbits.xen.org/xsa/advisory-191.html

Acknowledgements:

Name: the Xen project
Upstream: Andrew Cooper (Citrix)

Comment 1 Adam Mariš 2016-11-08 14:37:12 UTC
Created attachment 1218530 [details]
xen-unstable, Xen 4.7.x

Comment 2 Adam Mariš 2016-11-08 14:37:39 UTC
Created attachment 1218531 [details]
Xen 4.6.x, Xen 4.5.x, Xen 4.4.x

Comment 3 Martin Prpič 2016-11-22 12:26:23 UTC
Created xen tracking bugs for this issue:

Affects: fedora-all [bug 1397383]