Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <c7e87fa11cdd60fda87f7ffd7448b50d@smtp.hushmail.com>
Date: Wed, 05 Nov 2014 17:42:45 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Discuss Mask-mode with GPUs and changes required for
 its support.

On 2014-11-05 16:02, Sayantan Datta wrote:
> Based on my earlier experience with mask-mode, it was necessary to write
> separate kernels for benchmark, mask-mode(password on GPU) and non-mask
> modes. However, as much as I would love to unify them under a common
> kernel, with all of them following the same code path, it is difficult to
> do so without making debatable changes. Some of the changes I propose are-
>
> 1. Apply a default mask='?w' for formats supporting GPU generation when
> non-mask modes are used.

You mean we'd technically run a "hybrid-mask kernel" but with a no-op 
mask? This sounds reasonable.

> 2.  Do away with self-test for formats with password generation on GPU.
> Instead, use the test-suit or write a separate self-test which is based on
> the same paradigm as the test-suite. Optionally, we'd want to use a
> non-default mask, that can crack the hashes loaded for checking.

IMO this is not acceptable due to the history of driver bugs coming and 
going. We will find out ways to self-test without using any special 
kernel. You can temporarily disable self-tests though, and I'll help 
putting the pieces together.

> 3. For benchmark, use a default benchmark mask, which corresponds to a
> generic usage scenario for the format. Also the number hashes cracked
> should be kept minimal so that it roughly resembles a generic cracking
> session.

Fair enough. But I would prefer we instead add support for combining the 
options: "-test -mask" would benchmark GPU generation (with the default 
mask from john.conf unless one is specified) while just "-test" would 
use the no-op '?w' mask. The benchmark should ideally not crack a single 
hash.

> Unfortunately, these are only the guidelines and I don't have much idea
> regarding where and how to incorporate those changes.

There will be needs for core changes but we'll sort it out together.


I think we will need at the very least a new format flag: FMT_GPU_MASK 
or something like that.

magnum

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.