[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 23 Jun 2011 10:25:55 +0200
From: S3nf <thes3nf@...glemail.com>
To: john-dev@...ts.openwall.com
Subject: Re: mscash2 tag + iteration count
Hello,
sorry for my really late response and thanks to JimF who already fixed the
bug concerning the salt length (buffer reuse issue). I've also found the bug
this morning but obviously I was too late again. =)
I've updated the unoptimized sample code for DCC2 hashes in the Openwall
Wiki.
The performance increase of the DCC2 code is also good news and I'm really
impressed by the improvements of John within the last few months. And in my
opinion the proposed changes to the mscash2 format (tag + iteration cout)
are a very good idea.
Keep up the good work.
Best regards,
S3nf
2011/6/22 magnum <rawsmooth@...dband.net>
> JimF has nailed the salt-length problem *without* using the OpenSSL
> workaround and we are now finalizing some performance optimisations (current
> boost is 17-18% on my gear!). I think it would be wise to include the below
> change before releasing it:
>
> The current mscash2 format (dcc2) use the same tagging (eg. in john.pot) as
> the older mscash format (dcc1). I think this is a bad thing. Furthermore, I
> believe the iteration count can be altered but it's currently not stored
> with the hash but hard-coded (with a #define) in mscash2_fmt.c.
>
> - I propose we keep the old format for mscash (dcc1):
>
> M$test1#**64cd29e36a8431a2b111378564a106**31
>
>
> - I propose we use this new tag + iteration count for mscash2 (dcc2):
>
> $DCC2$10240#test1#**607bbe89611e37446e736f7856515b**f8
>
>
> We will probably be able to attack different iteration counts in the same
> run.
>
> Anyone agreeing or disagreeing on this?
>
> magnum
>
--
All work and no play makes S3nf a dull boy.
Content of type "text/html" skipped
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.
