Proxmark3 community

iceman

Administrator

Registered: 2013-04-25

Posts: 9,507

Website

Yet anther crapto thread

As the subject states, its yet anther crapto thread.   I'm dedicating this one to the original author @blapost which apparently reads this forum still.

To start with, I got an email for @blapost where he said was a bit tired of bad patches onto his crapto1 code. Hence why the craptoev1 code never gonna be on github but that is a different story.    I got curious and started looking into the crapto1 imp again.

The PM3 source has 4 copies of crapto1 located on the following places.
1) armsource
2) client/nonce2key
3) tools/mfkey
4) tools/nonce2key

@blapost latest release of Crapto1 source is v3.3.   By the look of it, we don't have that version in PM3.

So what has been changed?  Well, the armsrc version has some minor changes not to worry about.
Both the /tools doesn't have changes but the most modified one is the 2) under the client folder.

This one has been adapted for attacks against chinese tag with no ACKs.  It kind of looks like a mess and was added around 2013.
The most important isssue is the changes @pwpivi did 2015 to make the "hf mf mifare" not have the WDT resets when facing a tag with corrected prng, nack response.    These changes makes all the changes done to the nonce2key.c 2013 obsolete.  The code paths is also a mess in there so it did call some of these function changes unnessesary. 

I'm all for dividing the special darkside attact (when par is zero) but it would be nice if that code so as good as the rest.  Maybe @blapost, @piwi could divert some attention on it?

And another side question to @blapost,   why did you make a hardnestedbinreader into txt nonce file? Why not just use the bin file format?

piwi

Contributor

Registered: 2013-06-04

Posts: 704

Re: Yet anther crapto thread

The modifications in client/nonce2key/crapto1.c are not for the no-ACK no-par case. The (my) modifications in crapto1.c replace the previous quicksort-merge by a more performant bucket sort. The no-ACK no-par modifications are in client/nonce2key/nonce2key.c. I agree that they need not be there but could be in the calling function. I remember a respective Pull Request by the author to revert that. I am currently sorting those pieces.

iceman

Administrator

Registered: 2013-04-25

Posts: 9,507

Website

Re: Yet anther crapto thread

This is an old thread.  I've addressed this in my fork. The bucket sort didnt decrease execution time,  but as an algo it was interesting to watch and learn.  [@piwi, your c-skills is godlike]