Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 30 Jun 2012 09:03:41 +0200
From: Frank Dittrich <>
Subject: Re: For some dynamic formats on linux-x86-mmx build cracking
 depends on password candidate sequence

On 06/29/2012 11:52 PM, jfoug wrote:
> But this does appear to be something along the lines of improper key cleaning. 

Please note that with --mkpc=127 the bug disappears as well.

> I just wonder why you were not seeing any failures when the pw file was not in reversed order?

I think I know.

Among other such lines, line 5118 contains a password of 110 consecutive
[0-9] characters.
5118 = 40 * 128 - 2

If I change that word to a much shorter one (I used 0123456789fdtest),
the clang build (which failed to crack 18 passwords before), cracks
these 6 additional passwords, increasing the total count from 1482 to 1488:

Sword��sh__3     (u778-dynamic_2)
asdfasfga__7     (u658-dynamic_2)
characters__5    (u532-dynamic_2)
PIII__4          (u406-dynamic_2)
out�eft          (u169-dynamic_2)
swordfish        (u47-dynamic_2)

$ LC_ALL=C grep -n "^Sword...*sh__3$" pw.dic|grep -v fi

4994 = 39 * 128 + 2

Interesting that this word is *before* the one that I changed!
May be we have a bug in a totally different location?

$ grep -n "^asdfasfga__7$" pw.dic

5122 = 40 * 128 + 2

$ grep -n "^characters__5$" pw.dic

5250 = 41 * 128 + 2

$ grep -n "^PIII__4$" pw.dic

5378 = 42 * 128 + 2

So there is a pattern.

May be we need to carefully construct the pw.dic, so that very long
passwords appear close to the boundaries of (MAX_KEYS_PER_CRYPT) blocks
of passwords.
And it is a good thing for testing that dupes suppression is not
switched on.
Running with --dupes-suppression, I would not have hit this bug.


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.