random(4) plugin infrastructure for mulitple RNG in a modular fashion

David O'Brien obrien at FreeBSD.org
Wed Aug 7 18:32:43 UTC 2013


Replying to an earlier comments made:

On Fri, Aug 02, 2013 at 09:31:30AM +0200, Pawel Jakub Dawidek wrote:
> Not to mention that hardware RNGs are often hard to trust. I think Intel
> refused to publish RDRAND specs?

No argument there.  I know some folks feel RDRAND could be tainted for
NSA's sake.


> With PRNG in place you get consistent performance, it is seeded from
> various different sources (hardware RNG being one of them), so it is

This isn't correct.  If you have hardware RNG today, you use it instead
of software RNG.  The HW-RNG is not an entropy source into Yarrow.

Another thing we're fixing if we can get past this infrastructure change.

-- 
-- David    (obrien at NUXI.org)


More information about the freebsd-arch mailing list