Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 10 Mar 2015 16:11:49 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: bcrypt-opencl

On 2015-03-10 15:49, Sayantan Datta wrote:
> On Tue, Mar 10, 2015 at 6:18 PM, magnum <john.magnum@...hmail.com> wrote:
> 
>> - in December 2014, Jim ripped common code from bcrypt and
>> bcrypt-opencl, and placed in BF_common.h & BF_common_plug.c. At that
>> point and up until today, the opencl would "accept" $2y$ but it's
>> unclear to me if it did so 100% correctly.
> 
> Why ? Did the CPU and the openCL version show different behavior ?

Not to my knowledge, and it may be that all was fine. It's just that I
do not think Jim actually added any support to the OpenCL format (in
case that was needed), I think he just had them share valid/binary/salt.

Actually, I'm considering re-applying Jim's patch. But possibly re-write
it in a way that doesn't confuse Git as much as it did. So maybe you
shouldn't bother with this any more than helping me understand the
*practical* differences between the four tags in the OpenCL format.

valid() is simple. We should just accept all four variants.

In salt(), the CPU format currently has this:

        if (ciphertext[2] == 'a' || ciphertext[2] == 'b')
                salt.subtype = 'y';
        else
                salt.subtype = ciphertext[2];

        return &salt;

Will that be correct?

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.