Author Topic: Breaking the Ledger Security Model  (Read 1267 times)

0 Members and 1 Guest are viewing this topic.

Offline bitwelderTopic starter

  • Frequent Contributor
  • **
  • Posts: 967
  • Country: fi
Breaking the Ledger Security Model
« on: March 20, 2018, 03:42:52 pm »
https://saleemrashid.com/2018/03/20/breaking-ledger-security-model/

Not much impressed by the general behaviour of the Ledger team
 

Offline thm_w

  • Super Contributor
  • ***
  • Posts: 6384
  • Country: ca
  • Non-expert
Re: Breaking the Ledger Security Model
« Reply #1 on: March 21, 2018, 07:26:18 pm »
Good write up.
The guy is apparently 15 years old as well, if they were smart they would offer him a job.
Profile -> Modify profile -> Look and Layout ->  Don't show users' signatures
 

Offline bitwelderTopic starter

  • Frequent Contributor
  • **
  • Posts: 967
  • Country: fi
Re: Breaking the Ledger Security Model
« Reply #2 on: March 21, 2018, 10:45:29 pm »
By the way, the young guy took already the other h/w wallet for a spin  :box:
https://saleemrashid.com/2017/08/17/extracting-trezor-secrets-sram/
 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13748
  • Country: gb
    • Mike's Electric Stuff
Re: Breaking the Ledger Security Model
« Reply #3 on: March 22, 2018, 12:36:10 am »
So if the magic security chip is a peripheral, and the MCU controls the USB and OLED, why couldn't malicious firmware just ignore the security chip altogether, and implement fake security functionality internally?
Am I missing something ?
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline thm_w

  • Super Contributor
  • ***
  • Posts: 6384
  • Country: ca
  • Non-expert
Re: Breaking the Ledger Security Model
« Reply #4 on: March 22, 2018, 09:15:20 pm »
So if the magic security chip is a peripheral, and the MCU controls the USB and OLED, why couldn't malicious firmware just ignore the security chip altogether, and implement fake security functionality internally?
Am I missing something ?

If you do this after the user has setup their account, etc. then you would not have access to the private keys on the secure chip (SE).

If you do this before initialization, and feed the user a false seed or similar, that would be an option. But then I assume it would not authenticate via their web application. There must be some Ledger private keys in the SE as well that allow firmware updates and SE verification.

https://support.ledgerwallet.com/hc/en-us/articles/115005321449-How-to-verify-the-security-integrity-of-my-Nano-S-

Quote
4) The root of trust for the current batch is the following secp256k1 public key : 0490f5c9d15a0134bb019d2afd0bf297149738459706e7ac5be4abc350a1f818057224fce12ec9a65de18ec34d6e8c24db927835ea1692b14c32e9836a75dad609 - as checked here Genuine.py
Profile -> Modify profile -> Look and Layout ->  Don't show users' signatures
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6721
  • Country: nl
Re: Breaking the Ledger Security Model
« Reply #5 on: March 23, 2018, 12:23:23 am »
This is why I trust open source tokens you can just program yourself more (of course physical access is an instant loss). It would be nice to have open source and tamper proof chips, but if I have to chose I prefer open source when doing business with this kind of company with so little skin in the game.

It's different for my bank, then again if some hack happens against the hardware I got from my bank they'll generally just refund any money lost ... just like they did for skimming. I have more cause to trust them.
« Last Edit: March 23, 2018, 12:25:40 am by Marco »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf