Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 8 Jun 2012 08:09:17 -0500
From: "jfoug" <jfoug@....net>
To: <john-dev@...ts.openwall.com>
Subject: RE: rawSHA1_LI internals

>Here is a patch that does what I mean. I just don't understand how it
>can possibly work without this patch. BTW the new self-tests fails
>unless this is done to cmp_all().

Your change to non-sse2 is still a problem. I have added your 000's to the
self test (leaving the original hashes there also (i.e. the dupe condition),
and have re-done cmp_one.  Cmp_one is now an cmp_exact for both SSE2 and
non. So cmp_exact simply returns a 1

This still does not change the behavior of getting the get_source to
'correct' the hashes.  To do that, can ONLY be done in a call right after a
crypt_all.  I will have to find the proper completed binary from within the
format.  The caller does not 'give' us the computed hashes, it only gives us
the original binary, which has the 00000's.  But I 'do' think this is
achievable.

This is a patch against bleeding, but it still should apply properly to
jumbo-5 and or to magnum-jumbo (a few lines offset).

Jim.

Download attachment "raw-sha1_LI_fix.diff" of type "application/octet-stream" (1836 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.