Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 1 Aug 2013 23:03:05 +0200
From: Frank Dittrich <frank_dittrich@...mail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Bleeding: 4 out of 346 formats failing (linux-x86-64-native
 OMP build on Intel i7)

On 08/01/2013 10:22 PM, Dhiru Kholia wrote:
> On 08/01/13 at 09:55pm, Frank Dittrich wrote:
>> These formats fail:
>>
>> eCryptfs [65536x SHA-512]... (8xOMP) FAILED (get_hash[0](1))
>>
>> Sometimes I also get a core dump (instead of just a failing self test)
>> when eCryptfs is benchmarked.
> 
> I have pushed a fix for the eCryptfs format to bleeding-jumbo. Please
> pull and give it a try again.

Thanks, this seems to have fixed it.

I tested the OMP version several times (--test=0, --test; with and
without --format=...), and eCryptfs never failed, while some of the 3
dynamic formats still fail in every test run.

BTW:
I had a look at your changes.
You use ECRYPTFS_DEFAULT_NUM_HASH_ITERATIONS. The name seems to indicate
that the number of iterations can be configured.
If this is indeed true (no idea, since I don't know eCryptfs), shouldn't
the hash include the number of iterations, even if all 3 test hashes in
ecryptfs_fmt_plug.c use the default number of iterations?

Frank

Powered by blists - more mailing lists

Your e-mail address:

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