Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 29 Sep 2008 16:39:31 +0200
From: Bucsay Balázs <>
Subject: Re: quite fast MD5 hashing implementation

Hmm, i doesnt understand you, or you doesnt understand me :)
The size of the hash is 128bit (md5), we have the hash, so we have the
a64,b64,c64,d64. If you want to get the a63 and the rest, you have to
know too (i think) the F function (we know it from the md5
specification), the ac value (we know it too) and the x. Probably we
dont know the password, so we dont know x.

So how can we do a reverse technic? And if we can do a reverse technic,
why can we only revserse only the round 4?

I hope you will understand what i want to say :)

Balázs Bucsay

Simon Marechal wrote:
> Bucsay Balázs a écrit :
>> Hello!
>> I dont see the solution here, maybe because i'm "blind" :)
>> a63 -= F(b64,c64,d64) + x + ac;
>> You can tell here the a63, if you know the "x", but the x is the part of
>> the right password, so if you know x, you will know something about the
>> password. Am i right? I hope im not and i can learn a new trick :)
> Oh I see your problem. Actually you pretend that you do know the
> password, except for the first four bytes. Then you somehow "know" x,
> until you have to reverse K[0], which is at the start of the last round.
> That way you save 25% while the last part of the passwords you are
> testing stays the same.

To unsubscribe, e-mail and reply
to the automated confirmation request that will be sent to you.

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.