On Sun, 2014-01-05 at 21:38 +0100, Jan Tulak wrote:
I'm also developing it (
https://github.com/BroukPytlik/RdRand).
ReadMe still needs to be filled with reasonable texts, I have to copy it from
man pages (why write the same things twice). :-)
About the security concerns... I have done some statistical testing of it
(PractRand, TestU01) and even after many terabytes on four machines it didn't
found anything suspicious.
So I would not used it directly for something important (closed things are
closed things, and with NSA paying to RSA for backdoors...), but for casual
usage or as one of more entropy sources (or as a seed for a CSPRNG) it can
work pretty well.
My package is including the C library and also a simple application usable by
users directly (i.e. usable in shell scripts) if they do not want to pull data
from /dev/[u]random.
Hello and welcome. Some questions based on your description of rdrand.
How would you expect someone to use this library? I mean if /dev/urandom
is more portable why use the rdrand tool or library? Also how does your
project relates to rng-tools that do use the hardware generator to
feed /dev/random?
regards,
Nikos