Undetectable Rootkits Through Virtualization?

techmuse writes "eWeek has an article about a prototype rootkit that is implemented using a virtual machine hypervisor running on top of AMD's Pacifica virtualization implementation. The idea is that the target OS, or software running on it, would not be able to detect the rootkit, because the OS would be running virtualized on top of the rootkit. The prototype is supposed to be demonstrated at the Syscan conference and the Black Hat Briefings over the next month."

Before people start the Windows flamefest

fta:
Rutkowska stressed that the Blue Pill technology does not rely on any bug of the underlying operating system. "I have implemented a working prototype for Vista x64, but I see no reasons why it should not be possible to port it to other operating systems, like Linux or BSD which can be run on x64 platform," she added.

Let's make this a bit easier to understand.

I'm sure someone will correct me if I'm wrong but ...

This is not really different from running WinXP, then installing VMWare Workstation, then installing Win2K in a virtual machine.

The "host" OS is what gets infected. That would be WinXP. Of course nothing running in the "guest OS (Win2K) would be able to detect it. But ... so what? And that would directly contradict their claim:

Rutkowska stressed that the Blue Pill technology does not rely on any bug of the underlying operating system.

There are only three (3) ways for the "underlying operating system" to be infected.

#1. Worm
#2. Virus
#3. Trojan

If we aren't talking "nude pictures of celebrities", then it's either a worm or a virus and both of those are bugs in the OS.

If it's a trojan, then WTF are you doing installing unknown apps on the host OS?

Now, the only way this would be interesting would be if the worm / virus / trojan installed the virtualization software, moved the existing OS to a virtual machine and faked the names of all the interfaces (NIC, IDE controller, etc). If you can do that, VMWare really wants to talk to you.

Re:Let's make this a bit easier to understand.

Now, the only way this would be interesting would be if the worm / virus / trojan installed the virtualization software, moved the existing OS to a virtual machine

That's exactly what it does, according to this paper [umich.edu] that somebody else posted in the comments. I don't know that it fakes all the hardware names and such (unlikely), but I doubt that the typical user would recognize that the hardware in their control panel was any different than before.

Think about what it means if they're right.

I don't think they're right. Look at page 3 where they have their diagram showing the VMM in direct contact with the hardware.

Here's a simple test to see if they're right.

Put in a NIC that your host OS does not have drivers for. Your host OS will not be able to connect to the network. Now, if the virtual machine in their example can access the network, then they're correct.

There's no end of hype for "threats" that never seem to materialize (or are vastly over-stated). If they can do what their diagrams indicate, then this would revolutionize the computer industry. I really mean that.

For example, you would NEVER again have any problem with wireless networking under Linux. Or sound. Or any peripheral. Or hardware accelerated video. No more nVidia drivers needed! The VMM handles it for you!

So, no, I don't believe that what they claim is actually what they can deliver.

MOD PARENT UP, he's right.

...and he provides a critical rejoinder to grandparent who misunderstood what BluePill does (or rather what it claims it does).

Grandparent seems to think that BluePill merely is a mal-VMM that sits between any guest OS and the host OS. So the guest OS won't know that he's being thwarted. What these folks are claiming is two-fold:

• They'll do what SubVirt did -- move the VMM which is usually operating as a process on a host OS below that host OS. So, not only are all the guest OSs not going to know a/b the the mal-VMM, but also the host OS itself effectively becomes another guest OS.
• Unlike SubVirt which required that the mal-VMM exploit a vulnerability in the *host OS* in order to do this swallowing-up of the host OS, these folks' claim is that there are generic mechanisms to inject code into the Vista kernel. And these generic mechanisms are sufficient for this subversion.
• Moreover, they're saying that this is the case, despite security mechanisms in Vista that prevent kernel-mode code from running if that code is not signed (by a trusted party).

Anyway these are some pretty tall claims (particularly, re: the ability to inject arbitrary code into the Vista kernel). I initially thought the same thing as the grandparent: that they were saying that you could create a mal-VMM so that any VM running on that mal-VMM would not be able to detect the badness of the VMM (which is pretty trivial, actually).

Re:Before people start the Windows flamefest

Rutkowska stressed that the Blue Pill technology does not rely on any bug of the underlying operating system.

It's doesn't rely on any bug of the guest operating system, and isn't detectable from the guest operating system. But if something is mitigating access between multiple guest operating systems to hardware, then that thing is itself some sort of minimal operating system, and it is there that the problem lies. As far as the guest operating systems are concerned, this is really more like what would previously have been a hardware hack, in fact it's almost like your healthy computer is running behind a compromised firewall that's sending out the spam or whatever.

Getting to the point, people act as if virtualization simplifies things, But really it's an additional layer of abstraction and complication, another mass of code and/or hardware to go wrong. Now there will have to be software tools to manange this new underlying minimal OS, and maybe virus/rootkit software. I think the applicability will be limited.

said this before

Implementing malware with virtual machines [umich.edu]

ok, but...

Who runs anything *real* on a virtual server?

Re:ok, but...

Sadly, and in a large part due to the way commercial IT is funded, this can actually look good on paper - to the technology accountant: "as many servers as we like, that can be created and destroyed at will? Yes please". We also need virtual finance teams, virtual staff, virtual customers - hell don't bother running a real business at all - just model the entire thing, and play it like a RTS sim - with your score linked directly to the corporate stock price!

Technology finance will cretae some bizarre technical solutions, if sombody in the organisation doesn't put the brakes on - another good example is "hmm terminal server runs all the same apps that native desltops do for the remote workers - let's just issue everyone a Windows TS "device" and host everyone's sessions inside a big servers in the data centre - it's cheaper, and there's no difference right? This is where someone gets to try and explain latency, and how it's different from "bandwidth", to an accountant :D "yeah but we just paid for a 1 Ooodlegig/s line - it'll be super quick!"

It's not new either - mainframes have operated like this for years. IBM would have you create your entire data centre inside z/VM - including the routers, switches and firewalls. It's great for development and testing - need more Linux/Apache/WAS/Oracle servers? sure just wish 3 more into existence, re-test your fancy shmancy clustering and treacle bending widget, and then bin them off again with another wave of the virtual wand.

We have clusters of Websphere AS inside one LPAR - not for speed I hasten to add - that would be silly, but to create resilience, seperate the Java VMs and add flexibitlty for software releases.

Is it *really* undetectable?

Can't you just take the hard drive out, mount it from another computer, and see all the malicious DLLs the rootkit was trying to hide from you?

the side effects are detactable

Current virtualization doesn't virtualize anything but basic VGA graphics. That's certainly noticable.

Boss asks: are you playing games at work?!

Me: Just checking for rootkits boss!

Motherboards already block this...

Some, albeit high end, motherboards support a visual warning message that alerts the user to a program, or the OS trying to modify the boot sector on the hard disk. If you had this enabled it would stop this rootkit dead in its tracks. It's just a shame that more bioses / motherboards don't offer this support by default.

If you have this on your motherboard I highly recommend you turn it on, it isn't too often that you reinstall the OS and pressing F9 isn't that much of an inconvenience even if you did it once a day.

PS - All of the "My favorite OS is secure" posts below this are wrong if the Operating System supports some type of driver, or root program (running in the kernels memory space).

Re:Motherboards already block this...

Hmm.. I have quite a pile of system boards here, dating from old 486 systems upto p4 and athlon xp, with ami, award, phoenix and biosses, and all of them have the boot sector virus protection option (tho sometimes just called virus protection).

This offers at best a partial protection. While the MBR is important, the actual boot is done from the partition boot record, mot the master boot record, and this badly named feature is not going to help against that. Why badly named? because it does monitor (attempted) changes to the bootrecord and doesn't know anything about viruses.

Next. even if you could protect against that, things just get a bit more OS and possibly OS version dependent because you have to move to the file that gets loaded by the partition bootrecord.

Oh, quite a few 'boot managers' change the mbr on every boot.

So while it offers some protection, that protection is extremely limited, and can be quite inconvenient.

This just reinforces the good old principle

If your system suffered a successful intrusion, you wipe.

Of course, there were LKM rootkits (pretty hard to detect) for a good while now, this is just taking it to an all new level.

I wish the spread of better hidden rootkits on Windows, because only that will further sane security policies and wipe the stupid idea of virus scanners out (when it's doing IDS not IPS). There ain't such thing as 'intrusion removal'. It's like putting on a condom after sex. Oh wait, it's slashdot. Let me rephrase. It is like trying to recover data from /dev/null.

A win-win situation for everyone

From TFA:

Rutkowska says of the Blue Pill concept, "I am very excited about the chance to work with Sony on how this technology can be used to protect their next generation of music CDs, DVDs, and high-definition Bluray discs. I believe it will be a win-win situation for everyone involved. Well, everyone important, anyway."

Not much less detectable

I don't think this changes the situation much. Viruses have always tried to hide. This just requires different methods to detect them. Ultimately some viruses can only be reliably detected by booting off of readonly media. The same now as before. I think OS providers should provide a boot disk for routine scanning as a matter of standard procedure.

Maybe it's time for some new paradigm

Perhaps there could be an OS that wouldn't allow malware to be injected through root-trust, signed applications, memory compartmentalization with read, write, execute permissions and 4 privilege levels (instead of 2). Of course, that wouldn't be Windows or Linux or BSD or any other generic OS.

Real Beneficiaries of Hardware Virtualization...

So, just as you would expect, the future of having CPUs with hardware support for virtualization will be wonderful for preserving absolutely perfect security and cloaking for rootkits and their owners. In fact, thinking of why a certain class of non-blackhat beneficiaries would very much like such a possibility, this could be why both Intel and AMD are planning to ensure that all future CPUs, including even those in ordinary non-server desktop PCs, will have compulsory (permanently enabled) hardware support for virtualization. You know the routine - think of the children etc etc.

Is this a "root kit"?

Technically it's not rooting an OS but actually is almost it's own OS (hypervisor actually) that is running the OS in a virtual machine. Couldn't you get the same effect by hacking BIOS?

While the subject is scary

Is this really a surprise? Given the layered design of software, if you have something that can sit between the hardware and the software (and monitor what passes between, and control said information), they why would it not have complete control? The question is how could this easily be placed on someone's machine? The next question is why can a level of virtualization be introduced between the operating system and the hardware during execution?

"your operating system swallows the Blue Pill and it awakes inside the Matrix controlled by the ultra thin Blue Pill hypervisor. This all happens on-the-fly (i.e. without restarting the system)"

Brilliant

The Matrix has you.

livecd

Microsoft already working on one to address teh issue. I dont know if you have to pay additional money as MS hinted that ms antivirus will be a subscription service.

Running everything off a livecd is a good idea since most infected pc's are as slow as a 486 and could take hours to days to scan. In this case it would be ineffective.

I wonder if bios virii are next/? Its the only way to go above even booting off a pc and some malware and spyware makers are working on this. That way it can't be removed at all. Claria should be taken out to a field and shot.

Towards a runtime for Voight-Kampff machines

"Is this testing whether I'm a virtual machine or a lesbian, Mr. Dowd?"

Microsoft

I remember an article a couple of months ago where Microsoft employees had done something similiar, that is using virtualization to create a low level rootkit.

Because, y'know, the only way to protect yourself against attacks like these are with Trusted Platform Modules.

20 bucks Microsoft sponsored this research in some way.

So what?

I don't think that anyone is suprised by this. After all, it is common sense that the virtual OS's security is at the pleasure of the host, in much the same way that the security of a user-mode process is at the pleasure of the operating system. If there is anything between you and the actual naked hardware, then there is always the possiblity that that layer is doing something with your data that you don't lie.

The Reverse: Using Host to Protect Virtual Servers

This is regarding Linux rather than Windows but:

Host machine with Vserver kernel running Tripwire or Aide
    with configuration adjustments to detect changes in client "machines"

Host machine well protected
  client machines doing ftp or web services or email or.....

Although Vserver is particular to Linux: Other schemes doing
reasonably strong virtualization can also do the job in Linux,
Solaris (Zones), BSD (Containers), Windows, etc.

It should greatly decrease the ability of something as clever as
BluePill to do damage if it was infecting a well-partitioned virtual
machine rather than a regular machine.

Vserver: http://linux-vserver.org/ [linux-vserver.org]
AIDE: http://www.securityfocus.com/infocus/1424 [securityfocus.com]

TPM

This is actually the good side to the Trusted Platform Module - you can set up your machine to refuse to boot, warn you, whatever if anything in the boot process changes. As it's implemented in BIOS with hashing of the boot process before it even loads it from disk, there's no real way around this short of having physical access to the machine and turning off the TPM.

The bad side of the TPM is when you lose control of it - then the machine isn't yours any more but the xxAA's.

Virtualisation used for rootkit-safe environments

Can't the same trick be used to make a rootkit-safe environment? Launch a watchdog application and let that watchdog application launch the real OS in a virtualized environment, as soon as a rootkit wants to fiddle the watchdog application takes notice and there would be no way for the rootkit to either detect or by pass the watchdog. Or even more drastic, launch each (or most) process in a virtualized environment, would probally be a little slow, but should provide a extremly secure OS.

Whoa. Déjà vu.

"A Slashdot article just went by, and then another one that looks just like it!"

"It's a glitch in the rootkit! It happens when it changes something!"

"No, I said a SLASHDOT article."

"Ah, you're probably fine then."

Once untrusted software is run on your machine...

... it's not your machine anymore.

So let me get this straight...

A virtual machine can't tell anything about the state of the host it runs on other than what's exposed to it? Isn't this kinda like saying that if you use an oscilloscope to monitor bit flips on the bus, the OS can't detect it? How is this news?

The virus must use some memory

The virus must use some memory. This probably makes it detectable. It would probably appear that your computer has less memory than is actually installed.

What could the virus do? I doubt it could swap to disk to cover that. I guess it could try using compression on a small part of the guest OS.

Bah, humbug!

Exactly the same thing was done using the ancient "cookie monster" program on Multics, long before Unix was even a gleam in T&R's eye.

The perpetrator created a user-ring instance of a user (a virtual-machine-like process), loaded in the cookie mosnter, then loaded the command interpreter and handed the result to an unsuspecting user, my boss.

He searchrd high and low, never suspecting the program that kept saying "Want cookie!" was down below the shell.

--dave

Undetectable to software maybe

I don't know about you, but when I start up VMWare in windows (or parrellels in OS X) I definatley notice a performance hit. I find it difficult to believe that somebody wouldn't notice the fact that larg(ish depending on how exactly this is implemented) chunks of RAM and disk space are suddenly in use by some rogue program.

DRM?

Can we use this to bypass the DRM included in Vista?

LiveCDs?

Ok, I see how that's a great technical feat, but how is this any different? If I suspect a machine at work I wipe it. If it's a friend asking for help, and there are tons of settings I can't easily copy, I use a LiveCD and hope for the best. The illusion of real-time scanners and programs like AdAware passed a long time ago. Unless the virus flashes the BIOS (kudoz to the writer), the LiveCD should still be able to track it down, right? Provided the signature is known of course. I don't see how this changes anything?

Nothing new, really.

The fundamental question of systems administration: once you have had a root compromise, what can you do to the machine to get it back up and running, in a known good configuration, with all chances of future compromise as a result of the initial compromise removed?

Answer: either compare the system (booted from known good media) to a known good set of files, or reinstall from known good media.

There's no other answer. Any tools you run on the compromised system are by definition suspect; they might be good, or they might be compromised. You have no way of knowing; anything they tell you is suspect. Even if you have tool binaries that you know are good, you don't know that the data they're gathering reflects reality or has been altered to give you a wrong impression.

So the fact that this software is undetectable doesn't really change anything; you're still finding out about the compromise through unusual activity, so that's 'status quo'. The only thing that's different is the layer that's compromised.

The interesting question is how the software gets in place in the first instance to compromise the system. The answer is that it was run as root (or administrator, or supervisor, or whatever the super-user is called). How did it get root privileges? Two possible answers: (1) a flaw in the OS (defined as the kernel, and any processes running with root privileges); or (2) the end user ran it somehow as root.

In the first case, it's the standard security problem. The OS is flawed; anything can get root. That's a bug. In the second case, it's end user stupidity. Nothing you run as an end user should require root privileges. (If the OS is designed in such a way that you do, again, that's a flaw in the OS. If the application expects it when it doesn't really need it, that's a bug in the application, and the vendor should be shot.)

So there's another layer the rootkit can hide in. Be still, my beating heart! This is, and remains, nothing fundamentally new. [acm.org]

What's that you say?

The next version of WGA will be undetectable? Thanks, Microsoft! ;)

Subvirt

This is old news. SubVirt [umich.edu] by Peter Chen's group at UMich is the original system which proposed this idea. FYI.

Useful Applications?

I remember somewhere reading that the founder of Ubuntu, MArk Shuttleworth, wanted to get linux to the point of being as easy as Firefox to install. What better way than to install linux, and still virtualize Windows on top of it? Not very practical at this point, I admit, but this could be quite an impressive way to switch over, not to mention quite easy for the end-user. You could even set it up so that you could easily switch between the two.

Great (start) Paper

SubVirt is the training wheels version, check this:

http://www.securityfocus.com/columnists/402 [securityfocus.com]

Want to Talk?

http://www.securityfocus.com/comments/columns/402/ 33600#33600 [securityfocus.com]

Let me guess...

The rootkit's code name is "The Matrix"?

As I understand, a likely initial vector is...

...the virtual hardware drivers. The driver software which is part of the vm management software is by definition standardized and thus a prime target for attack. It also operates below most antivirus software so if an exploit can be found in the virtual hardware, it concievable could leap from the virtual machine into the host operating system. At least that's the way I understand the process.

First it was the chocolate in the peanut butter...

Now it's the invading code in the 'software' that's supposed to be hardware. Not even the BIOS is safe anymore.

Shhhhhhhhh.......

Please don't let Sony hear...

Kind of obvious

This requires them to move the host OS to a VM. Getting someone to install a trojan is getting easier but something this disruptive would be obvious to even the dumbest lamer.

Actually quite detectable

A key point about virtualization (even hardware virtualization) that people miss is that it does not guarentee that programs run as they normally if those programs are timing sensitive. This isn't a new revelation. If you go back to the Popek/Goldberg paper from the 70s, they make it quite clear.

So how do you exploit this to detect that you're in a VM? If you're an operating system, the easiest approach is to disable interrupts periodically and wait out a few time slices. You would then compare wallclock time and see if you're wait took longer than you expected it should. If it did, you're being pre-empted. With interrupts off, that's a sure sign that you're in a VM.

The above is a general solution to the problem. It's funny the author used SVM (a.k.a. Pacifica). SVM has a feature called dynamic attestation. This essentially introduces an unemulatable instruction that one can use precisely for the purpose of determining whether you're in a VM or not.

An Obvious Solution

So the scenario described in TFA is like this:

Hardware [ Operating System ]
C:\rootkit.exe
Hardware [ Rootkit [ Operating System ] ]

But if we go back a step...

Hardware [ Hypervisor [ Operating System ] ]
VirtualC:\rootkit.exe
Hardware [ Hypervisor [ Rootkit [ Operating System ] ] ]

Then the Hypervisor can, in theory, detect the rootkit. So the obvious defense against such shenanigans is to install your own hypervisor before the bad guys do it first.

All of this is meaningless, of course, because if a piece of malware has sufficient privileges to install a hypervisor, you're already boned. Fix the privilege hole and you're as safe as you ever were.

Re:The only defense

that is not true.
All thr AV companies have labds where they make new exploits. Then design a way to detect that TYPE of exlpoit.

Besides, have software to protect your systems helps with the know problems bouncing around out there even if not the zero day ones. Fortuasntly there aren't a lot of zero day issues.

Re:Is the solution DRM?

No, the solution is to not give the malware the path to even be able to do this by using a capability-type system.

Re:The only defense

Are you staring blankly at me?

No, I'm staring at you like you're an effing loon who doesn't know what the hell he's talking about.

Re:The only defense

I've been telling people this for a while, mainly to blank stares; you cannot detect if you have a virus/keylogger/spyware on your system .... They only detect the known malware, but nobody knows about the undetected hacks.

Not true.

Many detection tools will look for specific signatures of known exploits. Thus this part of the detection will not detect anything else. We're in agreement up to this point.

However...

There are other means of detection. One can look to see if certain system calls have been hooked in some way, files placed in certain places, alternate calls to read the same file return different results, system behaviour typical of an exploit, so forth. Code sequences with known execution times can be run and if the results are too far off, you know something is up. Network traffic can be examined on the machine and passively tapped just off the machine, and the difference can be enlightening. Even if your malware author is a certified genius and masks every single possible activity (ha!), then how on earth are they going to hide the CPU power required to implement it? And so on and on...

I'd be surprised if there were many modern anti-malware utilities that didn't implement a few of the more basic generic checks. Your assertion is not true.

Heck, even in this case, bugs in the implementation of the virtualisation can be used to detect if we're running or not. Code sequences exist that can detect whether you're in a virtual machine by the subtle differences between a true machine and a virtual one. Look at VMware I/O addresses and drive IDs, for example. Any difference between the _huge_ interface between virtual machine and the real machine can potentially be tested for and used for detection.

Re:The only defense

>Accepting this then, the only truly safe way to compute today is to keep your
> boot/OS/application drive from being writable. Baring this, the next best
>step is to re-image your drive from non-writable media daily.

You'd certainly get a blank stare from me.

That's not very practical. Depending on your OS and partioning scheme, you would be losing logs, patches, and preferences with each re-image.

A better approach is to start with a clean system, run something like tripwire, and keep an eye out for unusual changes.