[Cryptography] Dieharder & symmetric cryptosystems

[Michel Arboi]

Le ven. 17 janv. 2020 à 18:51, james hughes <hughejp at me.com> a écrit :
> Yes, counter mode repeats at the block size 2^128 and never before.

I see.
Anyway, the chance that it hits a short cycle that is picked up by
Dieharder seems pretty low. That's odd.

Note that I don't have WEAK results every time. Just too often to make
me comfortable
There is a new 3.29.4beta Dieharder version on the web site. I'll
check if the behaviour is different.

> Maybe the best recommendation is to use the OS supplied recommendations for cryptographically secure random numbers.

Actually, I was confused. I thought this was more or less related to
what I saw with urandom.
But urandom is based upon Chacha20; its property are different.

--
PGP key ID : 0x85A1C6A1 - 0x05054F8485A1C6A1
Fingerprint: 1DC3 8857 B930 0B6B 9420  5D56 0505 4F84 85A1 C6A1