[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Sat, 07 Apr 2012 23:29:06 +0200
From: arcctgx <arcctgx@...zta.onet.pl>
To: john-users@...ts.openwall.com
Subject: inconsistent number of cracked hashes in dynamic_6
Dear all,
I ran into a problem with the dynamic_6 format in john-1.7.9-jumbo5. I
was trying to crack a list of hashes using the password.lst dictionary
provided with john-1.7.9.
john-1.7.8-jumbo7 cracked 64 hashes from the list:
> john -for="md5_gen(6)" -w=password.lst hashes.txt
Loaded 443 password hashes with 443 different salts ( md5_gen(6):
md5(md5($p).$s) [SSE2 16x4x2 (intr)])
<snip>
guesses: 64
But john-1.7.9-jumbo5 did not:
> john -for=dynamic_6 -w=password.lst hashes.txt
Loaded 443 password hashes with 443 different salts (dynamic_6:
md5(md5($p).$s) [SSE2i 10x4x3])
<snip>
guesses: 48
What is interesting, after I removed the john.pot file and tried again,
I got a different number of hashes cracked:
> rm john.pot && john -for=dynamic_6 -w=password.lst hashes.txt
Loaded 443 password hashes with 443 different salts (dynamic_6:
md5(md5($p).$s) [SSE2i 10x4x3])
<snip>
guesses: 18
I did that many times, and the number of cracked hashes varied from
session to session. On the other hand, john-1.7.8-jumbo7 always cracks
exactly 64 hashes.
I tested john-1.7.9-jumbo5 on 4 machines with various CPUs, make targets
and compiler versions:
Intel Q9550 Core2 Quad, linux-x86-64, gcc-4.2.4
Intel E6750 Core2 Duo, linux-x86-64, gcc-4.6.2
Intel Pentium 4, linux-x86-sse2, gcc-4.5.0
AMD Athlon XP 2000+, linux-x86-mmx, gcc-4.5.3
I got the same behaviour everywhere. Is this a bug, or did I do
something wrong?
In case anyone else would like to test this, the list of hashes is here:
http://pastebin.com/NpM5PS56
Best regards,
arcctgx
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.
