Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Tue, 24 Feb 2015 7:17:32 -0500
From:  <jfoug@....net>
To: john-dev@...ts.openwall.com
Subject: Re: SIMD SHA-crypt, etc.


---- Solar Designer <solar@...nwall.com> wrote: 
> On Tue, Feb 24, 2015 at 09:34:06AM +0300, Solar Designer wrote:
> > I guess your next step is to update Drupal7 and some other SHA-2 using
> > formats to use this new code as well?
> 
> I notice that many are already updated, but many are not.
> 
> HMAC-SHA384 fails self-test with multiple threads:

Yes, this is a known issue in the new code. I have stared at it a long time, and can not see the problem.  The only 'real' change between sha384 and sha512, is that I have to clean the tail end of the crypt buffer prior to performing the 2nd crypt (cut it from 64 to 48 bytes of data).  But I do the same thing in the sha224 variant (but different sized of course).  The 224 works like a champ.  Magnum has also stared at it, and can not yet see the problem.  We will get it at some time.  Note, this is why it is called bleeding edge.  I did find a block of code where SHA512 was being computed (I did hmac-sha512 first for SIMD ,then cut/pasted and tweaked for sha384), but it was in code that is not used, since PLAINTEXT length is 125 which is less than the 128 [io]pad size for 64 bit SHA2.  I have not been able to fix the OMP issue however.  I will likely not have a chance to get on it until this weekend.

Powered by blists - more mailing lists

Your e-mail address:

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