"Wired" article about key fob cloning

[E-Ryc]

2) Tesla has already done a lot to address it, via upgraded fobs on vehicles since June, and PIN to drive and software update allowing fixed fobs to be added to vulnerable older Model S vehicles. What have the other impacted manufacturers done? (Signal-blocking pouches. I am not joking.)
3) "Starts the engine?" Come on, Wired. You're better than that...

2/ Well, they added pain/extra steps to do to the owner (I don't see much difference from "use the pouch" solution). Once we can get the fobs replaced, it will be called "solution" (I'm not insisting on "for free")
3/ If I remember well, this sentence was in context "all keyless-entry cars" not "Tesla-specific"

 

[.jg.]

1) it's harder to pull off than the headline would have you believe (attacker needs to be within fob range of a target car, and then get within 3 feet of that car's fob. Then they can compute the future codes in seconds, on their computer with several-terabyte hard drive.

Where the other parameters are fixed, the range to the car or fob is a function of the antenna used by the hackers/thieves. The researchers used a small COTS antenna, which was not designed for use beyond about 10 - 20 cm. Blogs on RFID hacking decribe how to make antennas for longer ranges. All of the components of the mobile hacking kit in the video can be purchased for about $1000 or less.

2) Tesla has already done a lot to address it, via upgraded fobs on vehicles since June, and PIN to drive and software update allowing fixed fobs to be added to vulnerable older Model S vehicles. What have the other impacted manufacturers done? (Signal-blocking pouches. I am not joking.)

Agreed - Tesla has done more than the other manufacturers involved - and I like Tesla to maintain a lead in such matters.

 

[wdolson]

40 bits and no salt?
Why????
Is the car also not using a rolling code to generate the challenges?

We were using better stuff in the 90's (and you could optimize it to real time calculation).

From what I read there is a rolling code, but the hackers were able to work out the algorithm for the pattern by catching two key signals back to back and they can do that by mimicking the signals from the car and the key responds.

My understanding is this is a risk on the Model S only.

I heard the S and X use the same system, but the 3 doesn't have a fob.

And one more thing: You can watch the stolen car on your phone. And tell the cops where it is. I don't recall hearing that the fob turns off the phone app.

It hasn't happened in North America, but there have been some Teslas stolen in Europe that just vanished. The authorities believe the thieves disabled the sender so the car is incapable of sending out any signals to anyone. I suspect those cars ended up in Russia.

 

[Zaxxon]

Where the other parameters are fixed, the range to the car or fob is a function of the antenna used by the hackers/thieves. The researchers used a small COTS antenna, which was not designed for use beyond about 10 - 20 cm. Blogs on RFID hacking decribe how to make antennas for longer ranges. All of the components of the mobile hacking kit in the video can be purchased for about $1000 or less.

Definitely agreed. This is a problem, and it'd be nice if Tesla also makes the replacement fobs free or an at-cost option. I simply took issue with the sensationalistic headline, which gave the impression that someone could simply hack the car in seconds, then buried the details well within the article. They've got to have reasonably close physical address to the car, then very close physical access to the fob, then tangible physical access to the car.

And >= $1k worth of equipment that will not fit in, say, a handbag. It's very much not a hack that is likely to happen while you, say, walk away from your car in a parking lot. You'd have to be specifically targeted while stationary with your fob out.

 

[Zaxxon]

I heard the S and X use the same system, but the 3 doesn't have a fob.

The X fob had improved encryption from the start and is not susceptible to this hack.

Edit: I did read this somewhere, but the articles I Google now don't mention the X at all. Great journalism, guys! If someone has a link or refutation for this claim, hit us up.

 

[lklundin]

I found the video fairly scary and an incentive to use the new "Enter PIN to Drive" feature.
The vulnerabililty seems to be with the keyless entry systems of Teslas sold before June 2018 (and in other makes of car and motorcycles using simliar keyless entry systems).

HACKERS CAN STEAL A TESLA MODEL S IN SECONDS BY CLONING ITS KEY FOB

COSIC researchers hack Tesla Model S key fob

The Wired article does its best to be sensational, but the whole attack is the already well-known relay attack against 'passive entry' which is optionally available for the Model S:

"Aside from the PIN, Tesla also allows Model S owners to disable passive entry for its key fobs, meaning drivers would have to push a button on the fob to unlock the car. That would also stymie the KU Leuven attack".

So nothing new here - as far as I can read.

 

[mongo]

From what I read there is a rolling code, but the hackers were able to work out the algorithm for the pattern by catching two key signals back to back and they can do that
by mimicking the signals from the car and the key responds.

The article was a bit vague, but given there was a 6TB rainbow table, the fob was likely not adding a salt to the challenge or key. It sounds like it also used the same key for start as well as entry.

And >= $1k worth of equipment that will not fit in, say, a handbag. It's very much not a hack that is likely to happen while you, say, walk away from your car in a parking lot. You'd have to be specifically targeted while stationary with your fob out.

LF and RF Tx/Rx will fit in a purse for the data collection (check out Texas Instrument developmentkits). That can wifi to the computer with 6 TB drive (or go briefcase with larger battery pack). Would you notice a person with a purse (or satchel) walking near you for 10 seconds on a crowded sidewalk?

 

[mongo]

The Wired article does its best to be sensational, but the whole attack is the already well-known relay attack against 'passive entry' which is optionally available for the Model S:

"Aside from the PIN, Tesla also allows Model S owners to disable passive entry for its key fobs, meaning drivers would have to push a button on the fob to unlock the car. That would also stymie the KU Leuven attack".

So nothing new here - as far as I can read.

Relay attack is retransmission of the challenge/response in real time. This is back calculating the secret key such that you can duplicate the fob and no longer need the original.

 

[lklundin]

Relay attack is retransmission of the challenge/response in real time. This is back calculating the secret key such that you can duplicate the fob and no longer need the original.

I understand that reasoning. Feel free to try to explain what it means when Wired writes:
"Aside from the PIN, Tesla also allows Model S owners to disable passive entry for its key fobs, meaning drivers would have to push a button on the fob to unlock the car. That would also stymie the KU Leuven attack".

 

[Zaxxon]

In understand that reasoning. Feel free to try to explain what it means when Wired writes:
"Aside from the PIN, Tesla also allows Model S owners to disable passive entry for its key fobs, meaning drivers would have to push a button on the fob to unlock the car. That would also stymie the KU Leuven attack".

It would prevent the car or fob from responding, thus stymying this attack.

Also, a general relay attack lets the attacker get a one-time use code to get into the car. The difference with this attack is that they are effectively duplicating the fob, giving them permanent access.

Would you notice a person with a purse (or satchel) walking near you for 10 seconds on a crowded sidewalk?

Within 3 feet of the fob as per the article? Likely, yes. Again, this is a significant attack, but it's difficult to carry out.

 

[mongo]

I understand that reasoning. Feel free to try to explain what it means when Wired writes:
"Aside from the PIN, Tesla also allows Model S owners to disable passive entry for its key fobs, meaning drivers would have to push a button on the fob to unlock the car. That would also stymie the KU Leuven attack".

Passive entry is triggered by a message from the car (sent periodically looking for the fob) that causes the fob to send a message. With that feature off, the fob will not transmit without the user pressing a button. No button press, no messages, no messages, no data to use to back calculate the secret key.

Within 3 feet of the fob as per the article? Likely, yes. Again, this is a significant attack, but it's difficult to carry out.

Ah, you are in Colorado? That could be more apparent than in a more populous area.

 

[bob_p]

The 32.4 release feels like what Tesla did in late 2013 when a Model S high an object while driving on the highway. Tesla quickly distributed a software patch to disable the automatic vehicle lowering with adjustable air suspension at highway speeds, and then retrofitted extra armor plating to all vehicles to reduce the risk of battery damage from an object in the road bed - and they then re-enabled the air suspension.

Adding the PIN is a quick-and-dirty solution - and Tesla may want to go further for doing secondary authentication, such as detecting the smartphone that is usually present with a specific key fob or using voice recognition.

However, if there is a flaw in the key fobs - that should be fixed either as a warranty replacement or an option extra priced security upgrade.

Tesla should also send an e-mail or letter to every affected owner, because this appears to only affect certain vehicles.

 

[Zaxxon]

Ah, you are in Colorado? That could be more apparent than in a more populous area.

Yes. Would definitely be more do-able in, say, NYC or Tokyo.

 

[lklundin]

Passive entry is triggered by a message from the car (sent periodically looking for the fob) that causes the fob to send a message. With that feature off, the fob will not transmit without the user pressing a button. No button press, no messages, no messages, no data to use to back calculate the secret key.

Admittedly, I have only used a Tesla (Model S) for one day (in 2014) so I may be mistaken.

But isn't there also the scenario where the fob is pressed to lock the car (after which the owner presumably walks away)?

Couldn't that just as well be exploited to replicate the fob?

 

[g1000x]

Dang, my car was delivered on May 26, think I have the new fob? How can I tell?

 

[mongo]

Admittedly, I have only used a Tesla (Model S) for one day (in 2014) so I may be mistaken.

But isn't there also the scenario where the fob is pressed to lock the car (after which the owner presumably walks away)?

Couldn't that just as well be exploited to replicate the fob?

Possibly, but it should be harder. The RKE function includes a rolling code so that the message changes every time. That should require a lot more calculation to determine the rolling code and secret key. Whereas the Passive entry is only the secret key applied to the challenge from the car, so only the 40 bit value to figure out (if I am understanding their implementation properly).

 

[.jg.]

There's some more detail from the researchers themselves:
Fast, Furious and Insecure: Passive Keyless Entry and Start in Modern Supercars

The kit carried by the chap in the video:
The attacker device consists of a Raspberry Pi 3 Model B+, Proxmark3, Yard Stick One, and a USB battery pack. The Raspberry Pi connects to a smartphone’s WiFi hotspot allowing it to download files from a remote 6TB hard drive containing the TMTO tables.

About range from the fob:
Q. How close do you have to be to a key fob to clone it?
A. The maximum range is mostly affected by the transmission of the Low Frequency signals. Our experiments show that a key fob can successfully receive the signal transmitted by an unmodified Proxmark3 from a distance up to 1m.
This range can be increased by using purposely build antennas and transmission hardware. It has been shown in the past that this distance could be increased to eight meters [1].

 

[Tech_Guy]

Ive called Tesla multiple times now and no one has any idea whats going on with this. The phone support team has no internal memos or service bulletins on this but was able to look up the issue via google search. At my request they set me up for a call back to schedule a mobile service visit which I received today. The person I spoke to again had no idea what I was requesting or how to handle it. He also told me there are no new keys which is wrong according the article and what the Tesla representatives added to the article.

I've submitted an executive escalation requesting that they send out some kind of bulletin to the service team on how to handle this issue etc.

Ive also called the phone support and been re scheduled for a service center visit tomorrow which hopefully will resolve the issue but I honestly expect to go the same way as my phone calls...

 

[Reeler]

I thought new fobs were just $100 or so from Tesla? They should offer re-fobbing on service visits for some discount off the normal.

It would be worth a few bucks to get a couple new fobs and not have to enter a PIN into the screen every time.

 

[Tech_Guy]

just make sure your old fobs are removed / deactivated from the car once you have the new ones