Re: Lightweight support for instruction RNGs

[Thor Lancelot Simon <tls%panix.com@localhost>]

On Mon, Dec 21, 2015 at 08:52:16PM -0600, David Young wrote:
> 
> I think Alistair is concerned that the implementation of "NIST SP800-90
> CTR_DRBG" could be incorrect, or else that it could be embedded in
> a system in which the correct behavior is not, for whatever reason,
> manifest in the userland output.  Thus the statistical properties of the

Let me be crystal clear: I sent the mailing list a patch which in no
way changes anything that even directly calls the CTR_DRBG.  The patch I
sent (and on which I asked for comment *which I foolishly, it seems, assumed
would come from people who had actually read the patch!*) does not get
within a mile of the goddamned CTR_DRBG.

Of course if you want to know if the CTR_DRBG works right, it is entirely
sensible to run statistical tests on its output!

However, it takes a pretty determined reader, I would say, to look at
the beginning of this discussion and actually believe that's what we
were talking about; and one who is either sorely mislead, sorely
misinformed, *or is rudely commenting in a basically off-topic way
without actually reading the code* to think that the patch I actually
posted (again, for crying out loud, has anyone commenting even read it?)
would plausibly make any difference in the CTR_DRBG's output at all.

Thor