[Cryptography] Two round secure hashing

[Bill Cox]

And now I've found a way to create input collisions... oh, well!  In case
anyone spent time reading it, for a 1024 sized super-block, pick a random
bit location, and flip it in m0, m1, m512, and m513.  If the probability of
R(R(S1^m0)^m1) == R(R(S1^m0^(1<<bitpos))) is significant, and similarly for
the other pairs, then the probability of a collision is significant.  Since
I didn't specify how strong R has to be, it could easily pass along a
flipped input bit to the output as it's only change.

This probably wont work out, though I'm having fun playing with it.

Bill
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20140424/140a6a4f/attachment.html>