Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 26 Jan 2012 17:29:41 -0600
From: "jfoug" <>
To: <>
Subject: RE: OpenCL vs Test Suite

Many of these 'may not' be problems.  What they often are, is the format has
a limitation of length of password that is 'able' to be tried.

That is fully acceptable.  If the format only handles words up to 18 bytes,
and there are 50 of them that are 19 to 32 in there, then expect them to
fail.  However, one HUGE issue that has been shown within the test suite, is
many formats had problems where they do have a max password, but do not
control the words that were being sent in, and some very long lines would
smash candidates after them.  Thus, if a user runs some targeted but dirty
files (such as large web sites, where all the text has been tokenized into
words, but some of them are long 'garbage' stuff, and this dictionary is
run, then often we would see these longer items causing side effects and
smashing testing of the candidate passwords either right before, or right
after (and even at times, causing john to core).

So, just because the ts shows 'FAILED', a user / developer (usually the ts
would first need to be used by the developer to make sure his code is clean
and working), would still need to diagnose, if the TS was going properly or

Now, in the way that the test cases are made, (using the script
file), we can tell it, to limit password actually encoded to X number of
bytes. In doing that, the script WILL read the input file, but will simply
skip over words that are too long. Thus, you can properly build an 18 char
hash input file (if the format handles 18 byte only), and use the exact same
dictionary / word file, and then the test should be 1500.  We may or may not
need to do some tweaks, and get proper hash input files for the OpenCL
builds.  I really do not know, since I have no way of building / running
them, and my time is already allocated in ways, I can not help out much with
john for quite some time.


>-----Original Message-----
>From: Lukas Odzioba []
>Sent: Thursday, January 26, 2012 3:31 PM
>Subject: Re: [john-dev] OpenCL vs Test Suite
>First of all thank you magnum for opencl patch for ts, and testing.
>I can confirm problems with opencl formats except cryptmd5-opencl.
>Below is my TS log.
>ukasz@...kstar$./ opencl
>- JtR-TestSuite (jtrts). Version 1.10, Dec 18, 2011.  By, Jim Fougeron
>- Testing:  John the Ripper password cracker, ver: 1.7.9-jumbo-5
>John Jumbo build detected.
>testing: john -test=0
>form=cryptmd5-opencl              guesses: 1500 time: 0:00:02:37
>.pot CHK:cryptmd5-opencl          guesses: 1500 time: 0:00:02:34
>form=phpass-opencl                guesses: 1470 time: 0:00:01:00 :
>Expected count(s) (1500)  [!!!FAILED!!!]
>.pot CHK:phpass-opencl            guesses: 1470 time: 0:00:01:00
>form=raw-md5-opencl               guesses: 1429 time: 0:00:00:04 :
>Expected count(s) (1500)  [!!!FAILED!!!]
>.pot CHK:raw-md5-opencl           guesses: 1429 time: 0:00:00:04
>form=raw-sha1-opencl              guesses: 1468 time: 0:00:00:00 :
>Expected count(s) (1500)  [!!!FAILED!!!]
>.pot CHK:raw-sha1-opencl          guesses: 1468 time: 0:00:00:00
>form=nt-opencl                    guesses: 1478 time: 0:00:00:01 :
>Expected count(s) (1500)  [!!!FAILED!!!]
>.pot CHK:nt-opencl                guesses: 1478 time: 0:00:00:01
>form=ssha-opencl                  guesses:    0 unk unk : Expected
>count(s) (1500)  [!!!FAILED!!!]
>sh: ./tst.pot: No such file or directory
>.pot CHK:ssha-opencl              guesses:    0 unk unk  [PASSED]
>Some tests had Errors. Performed 6 tests.5 errors
>Time used was 546 seconds
>Now I'll try to figure out what is going on.

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.