Oracle (10G and lower) Hash Support?

[aliby]

Just read over the Adding Multiforcer Hash Types portion of the Developer Guide on your Wiki (http://cryptohaze.com/wiki/index.php/De ... Hash_Types) in the hope/idea that I might try to code this myself. Unfortunately that portion of the Developer Guide seems to be a little neglected

Then I looked at the Doxygen documentation and got myself all kinds of confused. If you've got some info/pointers, I'd be more than happy to take a stab at coding it. But if you've already got this started, I won't bother.

Thanks!

[Bitweasil]

Just read over the Adding Multiforcer Hash Types portion of the Developer Guide on your Wiki (http://cryptohaze.com/wiki/index.php/De ... Hash_Types) in the hope/idea that I might try to code this myself. Unfortunately that portion of the Developer Guide seems to be a little neglected

Then I looked at the Doxygen documentation and got myself all kinds of confused. If you've got some info/pointers, I'd be more than happy to take a stab at coding it. But if you've already got this started, I won't bother.

Thanks!

Documentation in general is neglected. It's something I'm working on as I have time.

The trick is that, right now, I'm porting everything to the MFN Framework. So the old stuff isn't still valid.

If you could write me up some concise host-CPU C code that took a password & made an Oracle hash out of it, using ONLY primitive DES operations (you have access to a single DES operation that will take a block & a key, and generate the output - it will not do any CBC chaining or such), it would be quick for me to port that to CUDA.

If you've got the logic to handle different length passwords/usernames, and can perform one operation, I can plug it in pretty quick. I'm planning to add the salted hash type to the MFN Framework this weekend if I have time (I'm in the middle of moving and changing jobs, so things are truly chaotic).

[aliby]

Well, I'm not that great at programming. Assuming the code/info I attached earlier didn't help, would C++ algorithim code from a different open source Oracle brute forcer (woraauthbf 0.22 - http://www.soonerorlater.hu/index.khtml?article_id=513) help?

If so, it's attached. If not, I'll try to brush up on my programming skills soon...

[Bitweasil]

Looks good. Between that & what Scoobz posted, I should be able to work it out. CUDA or OpenCL first?

[aliby]

Well, I'm currently running this on a box with an NVIDIA card... So I'd say CUDA first.

Although I may have a new box to play with here soon/later today that has 4x hd6990s in it... So... Both?

[Bitweasil]

I quite honestly don't know that my code works on a 6990.

I've never tried it on the dual GPU systems for OpenCL/ATI. It's quite possible that it won't work properly if at all.

[aliby]

Well, let's go with CUDA first then.

I'll be more than happy to test your code out on that box once it gets up and running!