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

[Warner Losh]

On Aug 24, 2013, at 2:35 AM, Mark R V Murray wrote:

> 
> On 24 Aug 2013, at 00:18, John-Mark Gurney <jmg at funkthat.com> wrote:
> 
>> Tim Kientzle wrote this message on Sun, Aug 18, 2013 at 12:27 -0700:
>>> But clearly some people really want to be able to
>>> force /dev/random to be the unconditioned output
>>> of a particular HW RNG.  I don't know if this is a
>>> good idea or not, but clearly there are people who
>>> want this.
>> 
>> Considering that the Ivy bridge's rdrand implementation already uses
>> AES to condition the raw entropy source, using Yarrow/etc to
>> additionally condition it seems excesive, hence why some people want
>> to use it directly…
> 
> Nehemiah as well, using a Davies-meyer hash in software. However, what
> may be excessive for one person may be another's requirement.

Especially in light of the recent NSA revelations...  I for one won't trust hardware random number generation...

I find it interesting that earlier in the thread there was a desire by a certain router company to fulfill the NSA's requirement that it use the random number stream from the intel chips directly, and that's what kicked off this overly long thread.

Warner