Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Sun, 10 Feb 2013 00:22:52 +0100
From: magnum <>
Subject: Re: cracking partial/inexact hashes (was: Support for PBKDF2 (SHA512))

On 9 Feb, 2013, at 23:57 , Solar Designer <> wrote:
> I am moving this to john-dev.
> On Sat, Feb 09, 2013 at 11:45:29PM +0100, magnum wrote:
>> I could not resist implementing a proper cmp_all() just to verify that the 1024 bit OSX hashes were indeed longer PBKDF2 output. They are. This is pretty silly considering that you don't have to calculate more than the first 512-bit block for knowing you have cracked the right password. We do it now anyway just for good measure, and because we can do it without slowing things down. With a totally na??ve approach, those hashes would take twice as long to calculate.
> Maybe we should print a warning when the first 512 bits match, but
> further data does not?  This would indicate data corruption/typo in the
> extra data.

We already do! It will print this to stderr:

pbkdf2-hmac-sha512: Note: False positive from cmp_one() for
'crackable4us'. This is a bug or a miswritten input line.

Maybe the wording is not optimal for end users though. I could change it to something about "partial match".


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.