Slashdot Mirror
Airgap-Jumping Malware May Use Ultrasonic Networking To Communicate
Hugh Pickens DOT Com writes "Dan Goodwin writes at Ars Technica about a rootkit that seems straight out of a science-fiction thriller. According to security consultant Dragos Ruiu one day his MacBook Air, on which he had just installed a fresh copy of OS X, spontaneously updated the firmware that helps it boot. Stranger still, when Ruiu then tried to boot the machine off a CD ROM, it refused and he also found that the machine could delete data and undo configuration changes with no prompting. Next a computer running the Open BSD operating system also began to modify its settings and delete its data without explanation or prompting and further investigation showed that multiple variants of Windows and Linux were also affected. But the story gets stranger still. Ruiu began observing encrypted data packets being sent to and from an infected laptop that had no obvious network connection with—but was in close proximity to—another badBIOS-infected computer. The packets were transmitted even when the laptop had its Wi-Fi and Bluetooth cards removed. Ruiu also disconnected the machine's power cord so it ran only on battery to rule out the possibility it was receiving signals over the electrical connection. Even then, forensic tools showed the packets continued to flow over the airgapped machine. Then, when Ruiu removed internal speaker and microphone connected to the airgapped machine, the packets suddenly stopped. With the speakers and mic intact, Ruiu said, the isolated computer seemed to be using the high-frequency connection to maintain the integrity of the badBIOS infection as he worked to dismantle software components the malware relied on. It's too early to say with confidence that what Ruiu has been observing is a USB-transmitted rootkit that can burrow into a computer's lowest levels and use it as a jumping off point to infect a variety of operating systems with malware that can't be detected. It's even harder to know for sure that infected systems are using high-frequency sounds to communicate with isolated machines. But after almost two weeks of online discussion, no one has been able to rule out these troubling scenarios, either. 'It looks like the state of the art in intrusion stuff is a lot more advanced than we assumed it was,' says Ruiu. 'The take-away from this is a lot of our forensic procedures are weak when faced with challenges like this. A lot of companies have to take a lot more care when they use forensic data if they're faced with sophisticated attackers.'"
http://tech.slashdot.org/story/13/10/31/1955239/ars-cross-platform-malware-communicates-with-sound
Is it really SO hard to get rid of dupes that are less than 24 hours old? You seriously call yourself editor if you don't even manage to get those basic things straight?
Bust out an oscilloscope and a logic analyzer and start looking at these signals. It shouldn't be hard to get a waveform capture of the audio running over the speaker and the handshake between a USB device and the host.
A certain alphabet agency that's been in trouble for tapping all kinds of folks lately? Or are they too clueless to put together a monster like this?
1. You'd have to write a boot loader that a) loads your bare-metal-level sound and microphone driver, networking driver, sonic network protocol, and payload.
2. You'd have to write the forementioned a) bare-metal-level sound and mic drivers. Network drivers that might as well be bare-metal, implement a sonic network protocol, and then get them to successfully transmit your payload.
3. You have to TEST this combo on many different machines.
We're either looking at someone who has a LOT of free time and hardware on his hands, or a 1st or 2nd world military-level dev team with LOTS of cash to spend, IMO.
The next Slashdot story will be ready soon, but subscribers can beat the rush and slashdot the links early!
You've discounted the most obvious option - an attention whore who isn't adverse to making shit up.
What is being 'proposed' is NOT anything infecting through the speaker/microphone, but a pre-existing inection (that was probably USB based)
then communication through these methods - a VERY VERY different thing.
The hype and BS layers need to be peeled off this.
There is no possible infection vector via microphone/speaker, or via power cord as semi-implied (unless you had a powerline modem..), it is simply a
way to get data out of the airgapped but INFECTED machine to others that may not be airgapped.
The 'solution' here is simple, remove the infection! there is more to security than just network airgapping!
Time to go back to security 101.
April Fools Day is five months away. Come back and repost this then.
Where, exactly, were these "packets" flowing when the networking cards were removed?
Are they UDP or TCP?
How long does it take you to download a movie over your speaker?
Sheesh, evil *and* a jerk. -- Jade
I read the original article, but I don't see any part where someone recorded what was going out the speaker and looked at it. If someone is sending data over audio, it will show on a scope. Clearly that's not going to do much unless the receiving side has some kind of modem code listening for it.
Then there are claims like "It seemed to send TLS encrypted commands in the HostOptions field of DHCP packets." Attacking via DHCP packets is plausible; DHCP clients get told a lot of things they're supposed to do, and some of the older vendor-specific extensions are very insecure. But TLS? TLS isn't used within the DHCP protocol itself. There's a way to store DHCP configuration info in an LDAP server and have a DHCP server access it via LDAP.
If someone is seeing strange DHCP packets, and reloading the BIOS won't help, it's possible that what's going on involves an attack via the network controller. The fancier network controller parts now have CPUs and EEPROM. This may be an attack which puts code in the network controller which in turn patches the BIOS.
The people studying this need to list exactly what network ICs the machines involved are using. Some network devices are too dumb to be used as an attack vector, but some have whole protocol stacks, WiFi support, remote administration support, etc. It would not be surprising if those were attackable.
I've expected attacks via network controllers for years. That's been used to attack servers. There's a known attack on PCI controllers which can survive rebooting and reloading the BIOS.
If the machine has wireless networking hardware and the attack exploits the network controller, it may be able to do wireless networking even if the user thinks they have the hardware disabled. Time to open up the machine, clip onto the JTAG port on the network controller, and read out the device memory with a JTAG debugger. Compare the dumps with other machines.
I think it's transmitted by LSD. My computer stopped doing that kind of stuff as soon as I stopped taking it.
Sheesh, evil *and* a jerk. -- Jade
No, you're still wrong.
Here's how it works:
Because you couldn't here my clear my through [sic] when I typed the word adult in reference to the /. community.
See how easy that is?
Required reading for internet skeptics
But people just beat their chest and ridiculed the people posting, locking and shuffling threads or in some cases on commercial antivirus forums, deleting threads and moving them to hidden sections or trashed them altogether.
I believe this is a huge conspiracy which has been going on for years. People in malware forums have been shouting from the rooftops about this but no one wanted to listen.
What you overlooked and should have read:
1. Nobody Seems To Notice and Nobody Seems To Care - Government & Stealth Malware
http://anonymous.livelyblog.com/2012/10/05/nobody-seems-to-notice-and-nobody-seems-to-care-government-stealth-malware/
2. Spy agency ASIO are hacking into personal computers
http://anonymous.livelyblog.com/2013/01/13/spy-agency-asio-are-hacking-into-personal-computers/
3. Will security firms detect police spyware?
http://anonymous.livelyblog.com/2013/09/17/will-security-firms-detect-police-spyware/
And several PDF files on blackhat pages, forums, and conferences.
These attacks against non-networked computers runs deep - some changes are so subtle and appear to blend into normal black box Windows activities people overlook them. Read article #1 which includes the sad state of malware detection on *nix.
When you Google enough for firmware, PCI, AGP, BIOS, sound card malware, SDR, FRS, and why some distros autoload the ax25, rose, and netrom modules by default (including TAILS, check it for yourself with lsmod), it is quite unusual. Why would a distribution like TAILS need hamradio modules? They're in there, too, in addition to the ax25, rose, netrom modules. Batman mesh networking is included in TAILS too.
People repeat the same mantra: the only safe computer is a non-networked computer. This is a lie. The truth is, an entirely shielded TEMPEST room with no network connections and shielding down to every piece of the computer is the best test environment, but who is going to take such precautions? Is the shielded computer in the shielded room bound for other locations outside of this safe room?
Wikileaks have released Spy Files, listing many companies developing malware to root your box beyond detection often aimed at Governments and Military sources. These secret communications are no secret, and some have been detected via FRS, but that's only one source out of many.
"Because you couldn't here my clear my through when I typed the word adult in reference to the /. community. "
I had to read that about 15 times before it started to make sense. I think you were trying to be sarcastic. Is that possible? English doesn't seem to be your first, or even second language, but to indicate sarcasm one uses quotes.
The latin "sic" means THIS, you use it when you are copying something verbatim but you know it is wrong.
"Sorry that one went over your head"
You might want to check your arrogant attitude and tone it down a bit. You aren't as "adult" as you think you are and could benefit from LISTENING to others and maybe LEARN something instead of looking like a complete JACKASS.
As the Ars article points out, the individual pieces needed to do all this have already been proven over the years.
Here's why it makes even more sense to me.
A military minded person cannot allow threats to exist anywhere. If anyone anywhere has a weapon that they don't, they must immediately take steps to duplicate it, and defend against it.
Now take that mindset, combine it with a large team of military hackers. Now every single exploit ever publicly disclosed becomes a checkbox on a list somewhere. As a recent Snowden leak story showed, 0-day vulnerabilities have been purchased by the government. We can be sure they run the largest honeypot networks in existence and immediately dissect every new worm, root kit and exploit that touches them.
Every theoretical exploit must be tested for feasibility, turned into a proof-of-concept and then packaged as a tool.
And all that $$ and hacker power is under the command of someone who wants turnkey solutions and "kill switches" for everything.
So it's definitely possible that such tools exist. But why would he be a target? I dunno, maybe someone wants advance notice on what the presenters at upcoming security conferences might be talking about so they can Barnaby Jack them?
Sometimes people will claim something they strongly believe already exists in order to motivate people to look for it and find their proof. Sometimes they get lucky and proof is found, other times they get exposed for it. I hope he's wrong, I really want him to be wrong, but part of me believes it's real because it's definitely possible. After all, if it's just a few years out, then "they" have had it for a decade or more.
Cwm, fjord-bank glyphs vext quiz
These machines do two things:
1. They try to infect other machines. They seem to use several methods for this. One is infecting USB sticks and other media. They have been observed abusing an old windows exploit that uses true type fonts as the vector for that.
2. They are trying to communicate with other infected machines. They use some rather inventive carriers for that it seems. One of these appears to be sound. How it works isn't published yet. Another seems to be to use out-of-band communication by putting data inside host-option packets in DHCP. It's obvious that the malware uses such side channels to avoid detection. The OOB communication is done purely to keep in touch with "the swarm" and is not used to infect other machines.
The real nastiness appears to be that this malware is able to infect multiple operating systems that are usually passed by malware manufacturers and also happens to be able to nest itself on the eeprom of infected machines. Both are more or less "a first" and the combination hasn't been seen in the wild either.
Right now, there's a lot of discovery being done and a lot of speculation taking place as to who made it, what it can do, how it gets itself in eeprom and prevents itself from being overwritten during reflashing of the bios. It's not known if the virus will attempt to infect virtual machines, or will only infect machines that will let it nest in it's bios. Also, anything malicious apart from infecting and communicating hasn't been observed. For all we know, it may be a true worm that does nothing but replicate and is an out of control experiment.
So far, no infections appear to have been seen on virtual machines, or machines that don't have an intel chipset. I haven't seen any linux infected machines mentioned, but don't hold your breath on that, if *BSD and OSX have been infected, Linux may very well be infected too. Windows is infected for certain, but what versions are exactly vulnerable isn't clear to me at this time.
Thus far, the only thing that can be advised to prevent infection is the usual; don't trust content/media from sources that could be spreading infections, knowingly or not and keep your system up to date. If applicable, set your bios read-only with hardware switches or jumpers and if at all possible, put passwords on bioses and put software blocks on updates as well. To this date it's not known if and what software blocks will prevent the malware, but it's best to give it as few attack surfaces as possible.
I was promised a flying car. Where is my flying car?
"You have to TEST this combo on many different machines."
I'm calling hoax as fuck on this whole thing, but for just your microphone and speakers, the majority of laptops are using RealTek. Bare metal for that shouldn't be too hard to handle, as the driverset remains the same across all AC97 models and HD models. Two compliant bare-metal drivers shouldn't be too hard to fit in. Now, transmitting over ultrasonic is a whole different beast, and to do this through a supposedly truly airgapped room via noise should be impossible, as real airgaps will easily kill those frequencies.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
Hey buddy its real. The bandwidth of this type of communication is low but the hardware will do it. The startup I work for is focused on transmitting data through high frequency audio and we're not the only ones.
Case studies include Yamaha info sound, Sonic Notify, and LISNR.
The only reason I'd doubt this story is because the bandwidth is less than 300 bits per second in most implementations I've seen.
I haven't yet seen mention of someone setting up microphones sensitive to ultrasonic frequencies to check to see what, if any, odd sounds are being made by the computers. A lot of extraordinary claims are being made and I just don't see the requisite extraordinary evidence.
> Sure they can. Maybe not very efficiently, and not far above the range of human hearing, but they are analog devices, so there is no sharp cutoff at some limit.
To explain a little more: The requirement for mic/speaker on a Mac is to generate/record audio in the audible frequency range in high quality. To have high quality on the high end of that spectrum, you'll have to use a mic/speaker that will still work at yet higher frequencies (read: ultrasonic), with decreasing quality the higher you go.
So in the ultrasonic range you do have a working mic/speaker with mediocre quality. Add: ... and you have a working network link.
- filters to compensate for different output volume at different frequencies (sorry -- missing the technical terms here)
- detection for frequencies that should better be avoided because the signal/noise ratio is too bad
- error detection/correction on the digital side
- retransmission of lost packets
For an engineer with embedded programming experience, this shouldn't be that big of a deal. The challenge isn't only in coding it up, it is also in looking up and comprehending possibly vast documentation needed to pull it off. The code, presumably, runs in system management mode on x86 machines.
A successful API design takes a mixture of software design and pedagogy.
An air gap merely means that no network or other data cables cross it. It doesn't mean keeping things physically away!
A successful API design takes a mixture of software design and pedagogy.
Why do you think network security engineers always have headphones on? They're not listening to music, they're packet-sniffing.
GrpA
Enjoy science fiction? "Turing Evolved" - AI, Mecha, Androids and rail-gun battles. What more could you want?
Pretty sure the Mac can be set to record and playback af 48k samples per second.That gives you at least 4kHz of bandwidth above the limits of human hearing right there. With modern encodings, that's probably good for around 20kbps.
I think the claim is that it's going to keep infected through the mic, that is, new rootkit pieces being put in through it.
so the badbios would have mic input drivers built in, which would still allow throughput to regular audio functionality.
the author should have provided examples of the communications. I mean, isn't this missing the usual proofs, like dumps of the said bios supermalware, dumps of the audio communications etc..
world was created 5 seconds before this post as it is.
Are you paying attention? A speaker is an analog device. It doesn't have a "cutoff", it has a frequency response curve. Speakers typically used in laptops are quite small, so tend to perform better at higher frequencies than lower ones. Typically I'd guess they're +/- 3dB between 200Hz and 15kHz, with more attenuation outside of that range. Better ones (as might be fitted to a Mac) might manage to stay within +/- 3dB between 100Hz and 20kHz.
Quickly! To the Batdetector!
I just called it "tanks", modeled it from games I'd played in arcades. There's a Windows tanks game from a decade ago that's very similar, except it's in color (the computer I wrote it for was black and white only). The Windows tanks game weighed in at over 4 megabytes, mine was probably less than 400 bytes including timing loops to slow it down enough to be playable. This was 1983 on a really primitive TS-1000, 1 mHz Z-80 CPU and 4k of memory.
As to favorite tanks games, I haven't really played many in the last ten years, but there was a first person shooter tanks game in the arcade at Disney World in the early '80s that was awesome (I worked at Disney then, spent a lot of time in that arcade). It steered with two sticks like a real tank.
Free Martian Whores!
It's using Microcode in the CPU that is received over 3G cellular.
Remember SandyBridge advertised this capability for supposedly stopping theft....
But it's really just a backdoor so they always have a network connection to your box. They can run compiler trust attacks or just read arbitrary data from memory after scanning application fingerprints.
I've been saying for awhile now that this is the next attack vector but the last few times I've mentioned it, you trolls downmodded me to infinity.
So please listen again. It's not the sound card.... they use that to detect when people are close to avoid transmitting if I were to guess. His tinkering proved they should stop before being detected.
Yeah, I thought of this, too. Here's some background info on the tech involved. It seems to fit, the article doesn't specifically say only certain newer intel processors are at risk, but it doesn't give any counterexamples that would rule it out, either. This is an obscure deliberately OOB data transmission channel that seems like it could well be the hidden vector, only... Surely a security specialist would be aware of this as a possible mechanism? Also, why would disconnecting the mic/speakers stop a transmission if it's really using 3G? Could be wrong, but I've reluctantly concluded that this line of investigation is probably a red herring in regards to the case at hand, although it's certainly alarming enough in its own right.
makes a fine covert channel to get data to or from a compromised router, and NSA has shown interest in mass-pwning routers.
/. -- the Free Republic of technology.