Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 14 Feb 2013 21:33:46 -0800 (PST)
From: deepika dutta <deepikadutta_19@...oo.com>
To: "john-dev@...ts.openwall.com" <john-dev@...ts.openwall.com>
Subject: Re: mschapv2-bitsliced conversion

Ok I will do that.


 
Cheers,
Deepika


________________________________
 From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com 
Sent: Friday, February 15, 2013 1:05 AM
Subject: Re: [john-dev] mschapv2-bitsliced conversion
 
On 14 Feb, 2013, at 19:44 , magnum <john.magnum@...hmail.com> wrote:
> I have merged your code with the latest mschapv2-naive version (it now has Freeradius support too) and dropped the old version in favor of this bitslice format. I have tested on x86 32/64 SSE and generic, as well as Sparc32 big-endian. And I ran it through out test suite. Everything seems fine. You can see the changes in unstable-jumbo branch (soon to be Jumbo 8) and bleeding-jumbo branch (future) here:
> https://github.com/magnumripper/JohnTheRipper.git
> 
> Note that your code is now named --format=mschapv2-naive and file is MSCHAPv2_bs_fmt_plug.c. The other (default) format now exploits a weakness so it doesn't use DES at all in the inner loop (massive speedup). But that format loads very slow so the BS format is better in some situations. It is also an excellent example for future DES stuff. We should do the same to NETNTLM_old_fmt_plug.c for a starter - it is almost identical to MSCHAPv2.
> 
> There are some whitespace changes, so diff it with -w to see the changes. Please base any future contributions on latest Git. Thanks!
> 
> Solar, I committed this to unstable too. We can revert it if problems surface. Old code is moved to unused/ and modifed so it will happily co-exist with the two others if moved back for testing.

I just realized the new code does not work for -mmx builds. I do not think many people need such a target but I guess we still have to support it. Deepika, could you try to implement one of the follwing?

a) Implement MMX version of DES_bs_crypt_plain() in x86-mmx.S

or

b) Add some #ifdef-ninja so MSCHAPv2 use generic code for MMX builds - without affecting DES and LM formats.

Please base this on the unstable-jumbo branch of Github as above.

magnum
Content of type "text/html" skipped

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.