phc-discussions - Re: [PHC] Re: Mechanical 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] [thread-next>] [day] [month] [year] [list]

Date: Mon, 7 Apr 2014 06:00:21 -0400
From: Bill Cox <waywardgeek@...il.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Re: Mechanical tests

On Mon, Apr 7, 2014 at 2:58 AM, Christian Forler
<christian.forler@...-weimar.de> wrote:
> On 05.04.2014 17:53, Poul-Henning Kamp wrote:
>  [...]
>>
>> I'm not defending POMELO in any way, I havn't even looked at it yet.
>>
>> I'm just pointing out that while dieharder is a damn good too for
>> what it is designed for, it is not designed for what we're trying
>> to do here.
>
>
> A common use-case for a password hashing schemes is password based key
> derivation. Cryptographers assume that keys to be random. This
> assumption is crucial for the security of almost all cryptographic
> schemes. Therefore, (for an adversary) the output of a good password
> hashing scheme should be indistinguishable from random. Hence, dieharder
> is IMHO a valid sanity check tool.
>
>
> Best regards,
> Christian

By the way, in order to test Catena, I had to make a one-line change:

diff --git a/Catena/code/src/catena.c b/Catena/code/src/catena.c
index 8c84f82..5316045 100644
--- a/Catena/code/src/catena.c
+++ b/Catena/code/src/catena.c
@@ -283,7 +283,7 @@ int PHS(void *out, size_t outlen,  const void *in,
size_t inlen,
        unsigned int m_cost) {

   return __Catena((const uint8_t *) in, inlen, salt, saltlen, (const uint8_t *)
-                 "", 0, t_cost, MIN_GARLIC, m_cost, outlen, REGULAR,
+                 "", 0, t_cost, m_cost, m_cost, outlen, REGULAR,
                  PASSWORD_HASHING_MODE, out);
 }

Without this, all m_cost values under 18 result in an error, and with
an m_cost of 18, it takes too long to generate a GiB of data.  Would
it be OK to check in this change along with the other changes I've
made to PHC entries to get them to compile with a common main.c?

Bill