Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 18 Apr 2012 16:22:36 +0200
From: Frank Dittrich <frank_dittrich@...mail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Weekly report 1

On 04/18/2012 02:38 PM, myrice wrote:
> Yes, I just noticed this. I took a look at crack.c.
> In crk_password_loop(), we invoke crypt_all for crypt a bunch of passwords.
> And next we invoke cmp_all for all hashes with same salt. But I still not
> sure about how to use get_hash*().

Whether cmp_all or one of the get_hash* functions (and which one of
those) gets called depends on the number of different hashes per salt
you have.
If there is just one hash per salt or just a few hashes per salt,
cmp_all will be called, otherwise one of the get_hash* functions (which
one again depends on the number of hashes per salt).

For decent hash algorithms and correct implementations, you'll usually
have just one hash per salt, since there are so many possible salts, and
the salt should be generated randomly.

There are, however, poorly designed hash algorithms which allow only a
very limited number of different salts.
And there are also broken implementations which do not pick a truly
random salt when computing a hash for a changed password.

That's why, you cannot rely on cmp_all being called, even if in most
cases (decent hash algorithm and correct implementation) you'll just
find one hash per salt in a file of password hashes.


BTW: Somehow you always manage to insert your replies to quoted text in
a way that looks like the first line of your reply is part of the quoted
text. This is somewhat confusing.
Could you try to put an empty line between text you quote and your
reply? This should it make easier to read your mails and recognise which
part is quoted text, and which part is your reply.


Thanks,

Frank

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ