[langsec-discuss] Breaking X.509 cert validation with frankencerts

travis+ml-langsec at subspacefield.org travis+ml-langsec at subspacefield.org
Wed Apr 9 20:32:04 UTC 2014


On Sun, Apr 06, 2014 at 08:26:41PM +0200, Meredith L. Patterson wrote:
> Oh, that's clever. I'd been headed down the symbolic-execution path for
> X.509 parse tree differentials, but GA fuzzing bred from real-world certs
> and synthetic cert parts is both hilarious and clearly effective. I wonder
> if we should invite them for an afternoon presentation, if they'll be
> around in time.

On the off chance you didn't already know this, mutation and
recombination of valid inputs are pretty much the standard "dumb"
fuzzing technique these days.  5 lines of python for the former:

http://fuzzinginfo.files.wordpress.com/2012/05/cmiller-csw-2010.pdf

There's a reasonable book on the subject that has sat unopened on my
desk but has pointers to lots of tools:

http://www.fuzzing.org/

The main difficulty seems to be in setting up the test harness and
configuring a (private?) cloud of (presumably Linux) boxes to run it.

Microsoft seems enthusiastic:

http://research.microsoft.com/pubs/121494/paper.pdf

Codenomicon has Testing as a Service:

http://www.codenomicon.com/news/press-releases/2012-01-24.shtml

And they found a prior TLS bug that way:

http://www.codenomicon.com/news/press-releases/2012-05-11.shtml
-- 
http://www.subspacefield.org/~travis/
Remediating... like a BOSS.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 834 bytes
Desc: not available
URL: <https://mail.langsec.org/pipermail/langsec-discuss/attachments/20140409/760ac221/attachment.pgp>


More information about the langsec-discuss mailing list