3 ms·

Linus's response is pretty angry, as usual: “Where do I start a petition to raise the IQ and kernel knowledge of people? Guys, go read drivers/char/random.c. T

by beedogs 13y ago

Linus's response is pretty angry, as usual: “Where do I start a petition to raise the IQ and kernel knowledge of people? Guys, go read drivers/char/random.c. Then, learn about cryptography. Finally, come back here and admit to the world that you were wrong. Short answer: we actually know what we are doing. You don't. Long answer: we use rdrand as _one_ of many inputs into the random pool, and we use it as a way to _improve_ that random pool. So even if rdrand were to be back-doored by the NSA, our use of rdrand actually improves the quality of the random numbers you get from /dev/random. Really short answer: you're ignorant.”

jdiez17 13y ago
To be honest, he is right. Not only what he said is correct, but cryptography is one of those strange fields where saying "I know more than you about this and I am right" is an acceptable thing to do.
- dingaling 13y ago
  I'll be the fool here since I'm not mathematical: how is it sustainable to mix any potentially flawed source into the entropy pool? For example if my RNG just outputs a series of 5s, surely that would weaken the pool even if the other nine sources were lovely lovely entropy? What is the 'threshold' for tolerable degradation? The patch that was finally accepted for the Linux kernel was to mix-in RdRand but not to increment the entropy counter. To me as a layman that suggests that if the other inputs are weak at any time ( e.g. soon after boot ) then RdRand will be dominant.
  - jn7 13y ago
    You don't mention that you have read random.c or anything about the SHA mixing that goes on. There's a fairly self explanatory comment that is worth reading. If you ignore the delivery, Linus' advice was pretty clear.
- beedogs 13y ago
  I agree; at this point you pretty much know what you're in for when you engage Linus in a discussion in which he's likely to have forgotten more about the subject than you'll ever know.

by beedogs 13y ago

jdiez17 13y ago

To be honest, he is right. Not only what he said is correct, but cryptography is one of those strange fields where saying "I know more than you about this and I am right" is an acceptable thing to do.

dingaling 13y ago
I'll be the fool here since I'm not mathematical: how is it sustainable to mix any potentially flawed source into the entropy pool? For example if my RNG just outputs a series of 5s, surely that would weaken the pool even if the other nine sources were lovely lovely entropy? What is the 'threshold' for tolerable degradation? The patch that was finally accepted for the Linux kernel was to mix-in RdRand but not to increment the entropy counter. To me as a layman that suggests that if the other inputs are weak at any time ( e.g. soon after boot ) then RdRand will be dominant.
- jn7 13y ago
  You don't mention that you have read random.c or anything about the SHA mixing that goes on. There's a fairly self explanatory comment that is worth reading. If you ignore the delivery, Linus' advice was pretty clear.
beedogs 13y ago
I agree; at this point you pretty much know what you're in for when you engage Linus in a discussion in which he's likely to have forgotten more about the subject than you'll ever know.