Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [day] [month] [year] [list]
Date: Tue, 30 Aug 2016 20:30:12 +0000 (UTC)
From: Sebastian Heyn <>
To:  <>
Subject: Re: Which is the correct hash?

Seems like this is correct, and the idea is brilliant! Even though it did not reduce the size of the hashfile itself, the speedup is drastical.
I have to get into openmpi next. Maybe a small openmpi-enabled live cd (USB) or something.

jfoug <> schrieb am Di, 30.8.2016:

 Betreff: Re: [john-users] Which is the correct hash?
 Datum: Dienstag, 30. August, 2016 22:23 Uhr
 8/30/2016 1:36 PM, magnum wrote:
 According to Sebastian's mail this morning, the only
 usable file for 
 > attacking is a fairly
 large one (16 MB) and we may need to CRC all of 
 > it for each guess. That's why speed is
 hit and this is what the "file 
 magic" stuff could possibly work around.
 This was exactly why I created
 the magic logic. Since the blob was a 
 file (IF the builder of the blob was not trying to confuse
 us by 
 renaming a jgp picture into a .zip),
 then the file magic would only need 
 fully decode the first 4 bytes, and then ONLY check if those
 ended up being
 PK\x3\x4   The magic was not noticeably faster
 for tiny 
 files, but the larger the file
 was, the more noticeable, and as the size 
 of file got huge, without magic, the cracking
 speed simply almost crawls 
 to a stop.
 But  (and this is a big but),
 we are fully assuming that we know what 
 file magic looks like.  That is not always correct
 (especially if 
 the user purposely renamed
 something as more camouflage).
 Was this a pkzip file, or a winzip AES
 encrypted file?  I though only 
 the pkzip
 had the magic logic in it, and that the winzip method did
 care how big the file blob was.
 @magnum: is my rememberance wrong here.

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.