Re: [RFC PATCH 0/5] CPU Jitter RNG

["H. Peter Anvin" <hpa@zytor.com>]

On 02/04/2014 12:31 PM, Stephan Mueller wrote:
>>
>> The quantum noise sources there are in a system are generally two
>> independent clocks running against each other.  However, independent
>> clocks are rare; instead, most clocks are in fact slaved against each
>> other using PLLs and similar structures.  When mixing spread spectrum
>> clocks and non-spread-spectrum clocks that relationship can be very
>> complex, but at least for some designs it is still at its core
>> predictable.
> 
> But isn't there an additional clock? The clock used to drive the cache 
> and memory bus? When measuring memory accesses timings, larger 
> variations in the execution time are evident. This also applies when 
> hitting the caches (for L1, the variations are less than for L2 than for 
> L3). The variations in access timings would come from the CPU wait 
> states and their duration, would it not?
> 

Variations doesn't mean quantum unpredictable noise.  All the clocks you
are referring to are derived from the same BCLK and thus predictable.
What you have here is a PRNG with a large and obscure state space.

	-hpa