Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 3 Mar 2011 07:58:28 -0600
From: "JimF" <>
To: <>
Subject: Re: MD5 Generic improvements

----- Original Message ----- 
From: "magnum" <>
>> On 03/02/2011 07:56 AM, jfoug wrote:
>> There is now a -v3 'early' on the wiki.  This issue should be fixed.
> Confirmed, using md5_gen(0), thin raw-md5


>(it only works with the extra fix)

The 'extra fix' I listed for the raw-md5 'thin' which was done in the SSE 
intrisic was not
part of my -v3 patch, so yes, it would have to be done on its own.  I do not
think this was put into the sse-intrinsic patch after I listed it.

That patch (from a prior email) is this, and it IS needed:

static void rawmd5_init(void)
  md5_gen_RESET_LINK(&fmt_rawMD5go, Convert(Conv_Buf, 
rawmd5_tests[0].ciphertext), "raw-md5");
  fmt_rawMD5go.methods.binary = rawmd5_binary;
- fmt_rawMD5go.methods.salt = fmt_default_salt;
+ fmt_rawMD5go.methods.salt = fmt_MD5gen.methods.salt;

So, the order of patching is:

My -v3 patch
intrinsics patch  (I think 2 of them, against jumbo-12)  NOTE, pretty much 
only useful if doing a 64 bit build.
the hand patch above, to the raw-md5 init() function.  (if you patched with 
Any other patches wanted (hmail? or others).

That will get you intrisics, and md5-gen which uses them.


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.