XMHF is an eXtensible and Modular Hypervisor Framework
that strives to be a
comprehensible and flexible platform for performing
hypervisor research and development. The framework allows others to
build custom (security-sensitive) hypervisor-based solutions
(called "hypapps").
XMHF is designed to achieve three goals – modular extensibility,
automated verification, and high performance. XMHF includes a
core that provides functionality common to many hypervisor-based security
architectures and supports extensions that augment the core with
additional security or functional properties while preserving the
fundamental hypervisor security property of memory integrity
(i.e., ensuring that the hypervisor’s memory is not modified by
software running at a lower privilege level).
XMHF advocates a "rich" single-guest execution model where the
hypervisor framework supports only a single-guest and allows the
guest direct access to all performance-critical system devices and
device interrupts.
XMHF currently runs on recent multicore x86 hardware
virtualized platforms with support for dynamic root of trust
and nested (2-dimensional) paging. The framework is capable of
running unmodified legacy multiprocessor capable OSes such as
Windows and Linux.
The XMHF project includes the hypervisor framework and supporting
libraries along with several example hypapps:
XMHF: The eXtensible and Modular Hypervisor Framework
supporting custom hypervisor-based solutions (called "hypapps").
XMHF includes several example hypapps including
full-fledged hypapps such as TrustVisor and Lockdown:
TrustVisor: A special-purpose hypapp that provides
code integrity as well as data integrity and secrecy for userspace
Pieces of Application Logic (PALs).
Lockdown: A hypapp that provides the user with a red/green
system: an isolated and constrained environment for performing
online transactions, as well as a high-performance, general-purpose
environment for all other (non-security-sensitive) applications. An
external device verifies which environment is active and allows the
user to securely learn which environment is active and to switch
between them.
The XMHF project comprises code from multiple sources, under multiple
open source licenses. See COPYING.md for details.
There are a substantial number of known technical issues with this
codebase, many of them with implications for security. Please see the
ticket tracker for full
details. This absolutely remains EXPERIMENTAL software. Do not trust
important data to this software.
For bug reports, feature requests, etc., please use the sourceforge
tickets tool.
For other discussion and questions, please use the sourceforge
discussion tool. Note
that the discussion tool can also be used much like a traditional
mailing list, if you prefer. You will still need a sourceforge
account. You can subscribe to all messages or to individual message
threads through the web interface, after which you will receive
corresponding posts through email. You can also post by responding to
such notification messages, and start new threads by sending mail to
general@discussion.xmhf.p.re.sf.net. Posts via email must
originate from a sourceforge account's primary email address.
We are open to contributions. The easiest mechanism is probably to
fork
our git repository
through the web UI, make the changes on your fork, and then issue a
merge request
through the sourceforge web UI.
Maintainers:
Amit Vasudevan (XMHF, libbaremetal and Lockdown), Zongwei Zhou (TrustVisor and tee-sdk)
Other contributors: Jonathan McCune, James Newsome, Ning Qu, and Yanlin Li
Design, Implementation and Verification of an eXtensible and
Modular Hypervisor Framework. Amit Vasudevan, Sagar Chaki, Limin Jia,
Jonathan M. McCune, James Newsome, and Anupam Datta.
IEEE Symposium on Security and Privacy,
May 2013. pdf
Building Verifiable Trusted Path on Commodity x86 Computers.
Zongwei Zhou, Virgil Gligor, James Newsome, and Jonathan M. McCune.
IEEE Symposium on Security and Privacy (IEEE S&P), 2012.
pdf
"It's an app. It's a hypervisor. It's a hypapp.": Design and
Implementation of an eXtensible and Modular Hypervisor
Framework. Amit Vasudevan, Jonathan M. McCune, and James
Newsome. Technical Report CMU-CyLab-12-014, June 2012.
pdf
TrustVisor: Efficient TCB Reduction and Attestation. Jonathan
M. McCune, Yanlin Li, Ning Qu, Zongwei Zhou, Anupam Datta, Virgil
Gligor, and Adrian Perrig. IEEE Symposium on Security and Privacy,
May 2010. pdf
Lockdown: Towards a Safe and Practical Architecture for Security
Applications on Commodity Platforms. Amit Vasudevan and Bryan Parno
and Ning Qu and Virgil D. Gligor and Adrian Perrig. Proceedings of
the 5th International Conference on Trust and Trustworthy Computing
(TRUST), June 2012.
pdf
Lockdown: A Safe and Practical Environment for Security Applications
(CMU-CyLab-09-011) Amit Vasudevan and Bryan Parno and Ning Qu and
Virgil D. Gligor and Adrian Perrig. Technical Report
CMU-CyLab-09-011, June 2009.
pdf