Re: [TLS] Certificate validation can of worms

[Kurt Roeckx <[e-mail]>]

On Fri, Apr 04, 2014 at 06:48:54PM -0700, Yan Zhu wrote:
> On 04/04/2014 06:35 PM, Watson Ladd wrote:
> > Dear all,
> > https://www.cs.utexas.edu/~shmat/shmat_oak14.pdf contains tests of
> > many TLS implementations. Interestingly all tested implementations
> > contain errors, and all but OpenSSL erroneous accepts. Cryptlib was
> > not tested, because it doesn't validate certificates.
> > 
> > At the core of these issues is the complexity of certificate
> > validation. Things for this committee to consider:
> > 
> > 1: How will DANE make this worse?
> > 2: Is this really the best we can do? What features of X509 led to
> > these problems?
> > 3: This focused on server authentication. How does client authentication fare?
> 
> In case anyone wants to investigate these questions, the code used in
> that study to generate mutated certificates is at
> https://github.com/sumanj/frankencert. (You will need a corpus of
> certificates to start from; the mirrors of EFF's SSL Observatory scan
> results seem to all be unavailable at the moment, but that should be
> fixed soon.)

I have a copy of that corpus.  But you might also want to look at
https://scans.io/


Kurt