phc-discussions - Re: [PHC] Specification of a modular crypt format

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

Message-ID: <20151007223813.GA18621@openwall.com>
Date: Thu, 8 Oct 2015 01:38:13 +0300
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Specification of a modular crypt format

On Sat, Sep 19, 2015 at 07:02:51PM +0200, Thomas Pornin wrote:
> The code includes some generic Base64 encoder/decoder (about 150 lines),
> then code specific to Argon2i (about 200 lines), and some test vectors.
> The Base64 code is constant-time with regards to the data contents (no
> data-dependent table lookups or branches); this is probably useless in
> most situations, but easy enough to achieve.

It is in fact (at least theoretically) desirable to have salt decoding
cache-timing-safe when (or just in case) the hashing scheme itself is
not cache-timing-safe or/and the hash comparison is not timing-safe.

This also means that en/decoding the salt, instead of using it verbatim,
results in extra risk.  Yet we seem to have chosen to do that to provide
a more generic interface to the underlying hashing scheme.  (I did post
this sort of rationale in here.  I should have mentioned the above
drawback in that same message.)

While we probably can implement things safely ourselves, third-party
implementations are a (bigger) concern.

Alexander