Hackers Break Currency Validator To Pass Any Paper As Valid Euro

Trailrunner7 writes "If espionage is the world's second-oldest profession, counterfeiting may be in the running to be third on that list. People have been trying to forge currency for just about as long as currency has been circulating, and anti-counterfeiting methods have tried to keep pace with the state of the art. The anti-counterfeiting technology in use today of course relies on computers and software, and like all software, it has bugs, as researchers at IOActive discovered when they reverse-engineered the firmware in a popular Euro currency verifier and found that they could insert their own firmware and force the machine to verify any piece of paper as a valid Euro note. 'The impact is obvious. An attacker with temporary physical access to the device could install customized firmware and cause the device to accept counterfeit money. Taking into account the types of places where these devices are usually deployed (shops, mall, offices, etc.) this scenario is more than feasible.'"

Comments

Firmware update? Unlikely.

[mveloso]

I doubt that you'd be able to hang around a cash register with a serial cable and update some device's firmware without someone noticing. At that point why not just update the cash register's firmware and have it give you money directly?

Re:Firmware update? Unlikely.

[Qzukk]

"Hello, I'm from the maintenance department and I'm here to update your firmware to protect you from the exploit that was recently published on 2013-10-13."

Re:Firmware update? Unlikely.

[jandrese]

Unless this attack is a buffer overflow or something when you put in a particularly formatted note, I don't see the issue. "Oh, you can bypass the bill checker if you break the machine open, pull the ROM chips, and put in new ROM chips!"

According to TFA, the guy went and analyzed the firmware to discover how it worked, and then noted that you could bypass the check routines in it to always set the "good" pins high. About the only thing even mildly worrying is that there is apparently no crypto lock on the firmware, but a crypto lock on the firmware would be useless if you have physical access to the machine anyway, only slightly complicating the job of redesigning the internals, so that's not saying much. There's a reason these machines are secured with a lock and a sturdy metal case.

Re:Firmware update? Unlikely.

[mcrbids]

All you have to do is get a technician costume. You know, a big, black bag with lots of tools in it, perhaps a utility belt, a button-up, short-sleeve shirt with a generic company logo on it. Walk up to the unit with a slightly bored expression, casually pull out your cable, and get to work. Pay no attention to anybody around you.

Chances are, you just might get away with it.

SOURCE: I watched Burn Notice a few times.

Re:Firmware update? Unlikely.

[SuperCharlie]

When I was around 12 or so, my dad was in the army and worked on anti-aircraft systems. One Saturday he needed to get or do something at the shop so he drug me along for the ride. Both of us in our plain clothes. We walked up to the shop, 2 guards patrolling, he said hi, pulled out his keys and opened the door. I was in awe of what I saw inside.. 15 M163 Vulcan self-propelled anti-aircraft guns all in a line. We piddled with some things, he started one up and made sure to tell me repeatedly dont stand in front of this.. (the radar).. and after an hour or so we left.

Almost to the car, he said.. "you remember those two guards?" "Yes.." I said "I didnt know them from Adam. You can get away with anything if you look like you know what you are doing."

A lesson I have remembered all my life and used on more than one occasion.

Re:Firmware update? Unlikely.

[sootman]

> Which is a vulnerability of your employees
> allowing access to some stranger...

I work in an office with over 500 employees. Do you think I know everyone who works in security, telecom, and I.T.?

Re:Firmware update? Unlikely.

[Baloo Uriza]

And yet, it still surprises me how many places that have called my company for service see me arrive on site holding a clipboard, a cardboard carton, and a toolpouch full of screwdrivers, and automatically assume that I'm there to fix critical equipment touching customer data without so much as checking my paperwork, much less checking ID. People are stupid. This trick could very easily work in plain sight.

Re:Firmware update? Unlikely.

[mlts]

There are some fairly sane security measures a maker of a security device can do for fairly cheap to ensure that a tampered device isn't going to work without a lot of money and time put in:

1: If it is something static like a bill checker, take the time to heavily QA the device, including throw prototypes in the field for a while. Then, just have the firmware burned into a ROM (a true ROM, not an EEPROM, EPROM, PROM, flash, or an OS on a HDD... it goes into silicon and is not modifiable, period.) Of course, a bill checker might need updates when the currency gets a facelift, so a bill checker likely would need some type of upgrade mechanism.

2: If an update mechanism is needed, TPM chips are not expensive. In fact, some ARM CPUs have them built in. That solves 95% of the problem right there, because if the OS isn't signed, the OS won't be able to decrypt the last stage and boot.

3: As a subset of #2, the code that allows flashing of ROM images should be in a non-alterable, signed image. This way, if the main OS image has to change, it has to go through the "gatekeeper" image to be written to the boot medium, or it doesn't get on there.

4: Multiple images. This way, if a flash image is verified and copied to a temporary space and is being copied to the main storage, a power failure doesn't brick the device. The TPM boots, finds the signature of the first image fails, tries the backup, boots from that. The flash process updates both images, so only one would be inoperable during an update at a time.

5: To prevent flashing to a less secure previous version, the OS image that does the image update work can be set to look at version IDs, or optionally, if the ID is signed with a certain flag, can allow earlier versions to overwrite newer ones, or have beta images be able to be downgraded if needed.

6: The image flashing would have to be via a physical process, such as a USB connection. This way, devices can't be upgraded over the network, which shuts out a lot of potential exploits.

I'm sure I've missed a few items, but it doesn't take a lot of engineering to have an update mechanism in place that is tamper resistant.

Re:Firmware update? Unlikely.

[Wycliffe]

And how did that work out for him?

Don't be so smug. Crimes like these have a reverse survivorship bias. You usually
only hear about the ones that get caught or at least leave evidence behind.

Well duh

[PhilHibbs]

If you can physically access and modify a machine, you can change the way it behaves. Is this really news? Can they do it wirelessly? Over the internet?

Re:Well duh

[gstoddart]

f you can physically access and modify a machine, you can change the way it behaves. Is this really news?

This part of the article is what struck me:

After watching some videos from the vendor Inves on the machine's operations and reading through the machine's documentation, Santamarta came to the conclusions that some of the security claims the vendor makes were somewhat specious.

"Unfortunately, some of these claims are not completely true and others are simply false. It is possible to understand how Secureuro works; we can access the firmware and EEPROM without even needing hardware hacking. Also, there is no encryption system protecting the firmware"

So it sounds more like the company said "our stuff is secure, awesome, and hax0r proof", and someone essentially said "challenge accepted".

That he could do the initial reverse engineering without ever even having had the device (he downloaded just the free firmware) tells me that this device was pretty ripe for the picking.

Second-oldest profession FTFY

[Anonymous Coward]

Politics is the worlds second oldest profession, noted for it's uncanny likeness to the first.

Re:Second-oldest profession FTFY

[Chemisor]

You are absolutely right. Here are the top ten similarities between politics and programming:

• Design is always better than the implementation.
• Our number generators are random. Really.
• Polling is a lousy way to gather information.
• Codes always have bugs and loopholes. When they are found, lawyers are often involved.
• Old codes never die and never fade away.
• After failure, always blame the third party.
• Paying for support is expensive.
• DRM and vendor lock-in are the best means of increasing sales.
• Never listen to your your customers when they say they want fewer features. They must be lying.
• Power corrupts. That's why we have checksums and balancing.

Counterfeiting ?

[mbone]

If it accepts _any_ piece of paper, I don't see how that is counterfeiting - theft and fraud, sure, but if I make no effort to copy something, how is that still counterfeiting?