Re: Torvalds Slams Theoretical Security

Lawrence D'Oliveiro <ldo@nz.invalid> writes:

Richard Kettlewell wrote:

Lawrence D'Oliveiro <ldo@nz.invalid> writes:

Weak cryptography is easy to fix. The hard part to fix is weak random
numbers.

 
Other way round. A bad RNG can be swapped out for a better one with
little or no impact on anything else.

>
Unfortunately, you can never be sure your RNG is good.

You can’t be 100% certain of anything. There’s a 2^-256 chance someone
could guess an AES-256 key in a single try, for instance. But you can be
certain enough for practical purposes.

Accidents (DSA-1571, Sony) and compromises (Juniper, Crypto AG) happen;
so does cryptanalysis (Enigma). But the majority of vulnerabilities are
not cryptographic in nature.

--
https://www.greenend.org.uk/rjk/