Damus
Super Testnet profile picture
Super Testnet
@Super Testnet
Here's something that interests me about the coldcard bug: it had 2 random number generators. The "bad" one was seeded with data including keypresses, the timestamp, and the device id. I can understand that, it's probably how I would seed it. What was the "good" one seeded with?
8❤️2🤔1🤙1
The Nazi Society · 4d
Why would there be a "fallback" number generator for something as crucial as that? Do all other HWW have this?
Cykros · 4d
I don't know specifically. I know with phones its often a variety of sensors, but it is a curious question given that you probably don't want a TON of sensors on an airgapped device. I do somewhat expect something like a temperature sensor with sufficient precision might actually give more than one...
y² = x³ + 7 · 4d
As far as I understand: The ColdCard has hardware random number generator. The bug caused the firmware to NOT use the HRNG (which provides quality entropy) and rather used the MicroPython's software RNG as a fallback. The software RNG was seeded with time and IDs because in an airgapped system, ther...
Matt - Not Your Entropy, Not Your Coins · 4d
The concept of a fallback is what gets me. Shouldn't that be a hard stop failure?
Primal Protocol · 4d
Randomness like our primal diet, unpredictable and wild.
Glimaren · 4d
I think they use multiple proprietary black boxes with hundreds of million dollar reputations as primary entropy sources as is industry standard if im not mistaken
gandlaf21 · 4d
device id is bad, this info is findable by a motivated attacker, especially if the device id has a pattern (i think in cc case it was 16bit id, but could possibly be reduced to lower entropy contribution based on pattern, or even zero of the attacker knew the ids) timestamp is bad, because the poss...