phc-discussions - Re: [PHC] friendly warning about randomness tests

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20140901222119.GB2249@openwall.com>
Date: Tue, 2 Sep 2014 02:21:19 +0400
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] friendly warning about randomness tests

On Mon, Sep 01, 2014 at 09:53:08PM +0000, Brandon Enright wrote:
> Yes.  Nmap (the port scanner) uses a LCG with some basic tweaks to it
> for generating "random" IP addresses to be scanned without producing
> any duplicates before cycling through all 2^32 IPs (Nmap's -iR feature).
> This passes every Dieharder randomness test.  It's probably not
> surprising but Dieharder doesn't have a check for treating the output
> as 32 bit numbers and then looking for expected duplicates.  If it did
> though Nmap's PRNG would obviously fail.

FWIW, extra/analyze.c included in the yescrypt submission should detect
the lack of duplicate 32-bit numbers.

Maybe we should run it on pre-final-hashing memory contents of PHC
candidates too.  (Naturally, I already did that for yescrypt.)

Failing randomness tests on pre-final-hashing memory contents is not
necessarily fatal (on actual PHS() outputs, it is), but it's useful
information for our analysis and decision-making.

> Randomness tests give me zero additional confidence in any candidate.

... but failing randomness tests give us additional concerns.

Alexander