phc-discussions - Re: [PHC] yescrypt to gain Catena features (Re: [PHC] gambit wiki strength)

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: <CAOLP8p51UG25g+sqrU4GYGR_w77993vfyuqE9fF6THpQWXMXiQ@mail.gmail.com>
Date: Mon, 14 Apr 2014 19:19:08 -0400
From: Bill Cox <waywardgeek@...il.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] yescrypt to gain Catena features (Re: [PHC] gambit wiki strength)

On Mon, Apr 14, 2014 at 4:55 PM, Solar Designer <solar@...nwall.com> wrote:

> On Mon, Apr 14, 2014 at 03:09:19PM -0400, Bill Cox wrote:
> > Yescript: The is surprisingly little in the Yescript code that seems
> > inspired by Catena.  I think Alexander is a fan of client-independent
> > update, but I don't see support for it in the code.  Perhaps Solar
> Designer
> > feels worse about using other people's good ideas than me.  He does have
> > server-relief, but it's an extended version meant to be nearly compatible
> > with SCRAM.
>
> I can add cost upgrades (client-independent updates in Catena terms)
> without breaking compatibility with yescrypt's current hashes.  I intend
> to add this maybe in the tweaks period.
>
> I also intend to add cache-timing resistant mode(s).
>
> The reason these things aren't in there yet is that I focused on core
> functionality first, whereas I view these as nice extras.  I simply
> didn't have time for these extras without impacting my work on the core.
> I also feel that yescrypt is complicated as it is, so having even more
> functionality in there might be counter-productive for initial reviews.
>
> I have no problem using other people's good ideas, with due credit and
> as long as the ideas aren't patented - e.g., yescrypt builds upon scrypt,
> and a future version will likely optionally use Catena's bit reversal.
>
> Alexander
>

Sounds good to me.  I was actually thinking that such functionality could
be provided in the API wrapper for the PHC submissions.  Functionality we
could add there includes:

- "strongly secure" PRK derivation (which we'd pass as the password to PHS)
- Client-indpenented updates
- Server relief
- Clearing the password before calling PHS
- Dealing with salt encode/decode

I almost wish someone had submitted such a framework with a pluggable
password hashing algorithm, maybe something like Catena without the Catena
algorithm inside...

Bill

Content of type "text/html" skipped