Pages being used to run x86 guests in shadow mode are reference counted
to track their uses. Unfortunately the overflow check when trying to
obtain a new reference used a mask one bit wider than the reference
count actually is, rendering the entire check ineffective.
A malicious or buggy guest may cause a hypervisor crash, resulting in
a Denial of Service (DoS) affecting the entire host, or cause hypervisor
memory corruption. We cannot rule out a guest being able to escalate
Xen versions 4.1 and later are affected. Xen versions 4.0 and earlier
are not affected.
x86 systems are vulnerable. ARM systems are not vulnerable.
Only guests run in shadow mode can exploit the vulnerability.
PV guests typically only run in shadow mode during live migration, as
well as for features like VM snapshot.
Note that save / restore does *not* use shadow mode, and so does not
expose this vulnerability. Some downstreams also include a "non-live
migration" feature, which also does not use shadow mode (and thus does
not expose this vulnerability).
HVM guests run in shadow mode on hardware without HAP support, or when
HAP is disabled (globally or in the VM configuration file). Live
migration does not affect an HVM guest's use of shadow mode.
For HVM guest explicitly configured to use shadow paging (e.g. via the
`hap=0' xl domain configuration file parameter), changing to HAP (e.g.
by setting `hap=1') will avoid exposing the vulnerability to those
guests. HAP is the default (in upstream Xen), where the hardware
supports it; so this mitigation is only applicable if HAP has been
disabled by configuration.
For PV guests, avoiding their live migration avoids the vulnerability.
Name: the Xen project
Upstream: Jan Beulich (SUSE)
Created xen tracking bugs for this issue:
Affects: fedora-all [bug 1525018]