Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [day] [month] [year] [list]
Date: Tue, 18 Sep 2012 21:36:05 -0400 (EDT)
From: jfoug@....net
To: john-dev@...ts.openwall.com
Subject: RE: hmac-md5/sha1 in magnum-jumbo (was: 1.7.9-jumbo-7)

This is EXACTLY what I see.  A few hashed missed, and then a few hashes 
missed with the .pot file is re-run.

I will state this.  It is ONLY a few hashes VERY early on.  Like in the 
first SSE block, or first 2 blocks of SSE.  It is the same ones every 
time.

I wonder if there was a memset or something missed in init() that later 
gets cleaned up, but in that first block run, cause a few items here and 
there to miss. NOTE, I have not dug into this (no time), but that is 
just an educated 'guess'.


> After fixing the SHA_BUF_SIZ bug in the -fixes branch (which made 
> about half of the 1500 fail) and merging that to magnum-jumbo, I now 
> probably see the same as you in the latter branch:
>
>
> form=hmac-sha1                    guesses: 1497 time: 0:00:00:00 : 
> Expected count(s) (1500)  [!!!FAILED!!!]
> .pot CHK:hmac-sha1                guesses: 1493 time: 0:00:00:00 : 
> Expected count(s) (1500)  [!!!FAILED!!!]
>
> form=hmacMD5                      guesses: 1498 time: 0:00:00:00 : 
> Expected count(s) (1500)  [!!!FAILED!!!]
> .pot CHK:hmacMD5                  guesses: 1493 time: 0:00:00:00 : 
> Expected count(s) (1500)  [!!!FAILED!!!]
>
>
> This happened when I introduced clear_keys() in the hmac SSE2 formats. 
> We could just revert that but I'm very curious why that fails. It 
> might be worth trying to nail it.
>
> magnum

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ