Re: Hardware RNG

"Andy Green" <andy@xxxxxxxxxxx> wrote in message news:19zTl.291881$HA3.51321@xxxxxxxxxxxxxxxx
On Wed, 27 May 2009 23:09:12 -0700, Joseph Ashwood wrote:

"Boon" <root@localhost> wrote in message
By way of LWN, I found the following page, which discusses how to build
what looks like a (true) random number generator.

Would anyone care to comment?

I have some concerns.

FWIW I put the 12GByte continuous sample from Whirlygig, which I still
have, through ENT and got this:

I'm going to stop you right there. Predefined tests on the output will only reveal enormous errors, not subtle flaws. This is the first lesson of cryptographically strong RNGs, not knowing this does not say good things about the design.