Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 04 Jan 2012 17:32:20 +0100
From: Samuele Giovanni Tonon <samu@...uxasylum.net>
To: john-dev@...ts.openwall.com
Subject: Re: getting the CUDA & OpenCL stuff into -jumbo

On 01/04/12 14:29, Solar Designer wrote:
> On Wed, Jan 04, 2012 at 02:13:19PM +0100, Samuele Giovanni Tonon wrote:
>> this is a patch to be applied on latest jumbo (jumbo-5) .
>>
>> As you can see there are little to nothing changes inside of Makefile.
>>
>> I was able to cleanly compile it, on both nvidia and Ati machines .
> 
> Thanks!  Unfortunately, I think this patch breaks non-OpenCL builds and
> thus is unsuitable to be merged into -jumbo as-is.  Specifically, I
> think the *_fmt_plug.c files introduced by the patch will be picked up
> in all builds, and I see no logic to disable them or their contents when
> OpenCL is not available.  I think you need to rename them to *_fmt.c
> (not "plug") and have the formats registered in john.c explicitly only
> in OpenCL builds (similarly to what we do for formats that require
> OpenSSL 0.9.8+).

this could do the trick you're looking for :-)

i deleted the 32-opencl target since i haven't a machine to test it
therefore at the moment opencl is only on 64bit, if someone is
interested at 32bit target we can make some tests.

Regards
Samuele

Download attachment "john-1.7.9-jumbo-5-opencl-2.diff.bz2" of type "application/x-bzip" (11385 bytes)

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.