In message <26650.1022702207@eterna.com.au>matthew green writes
   [...]
   
   >
   >
   >would someone please explain to me why our _kernel_ random() needs
   >to worry about POSIX or being reproducible (in general) ?
   
   Verification that the in-kernel security subsystem or cryptosystem
   (or other tool using random(9)) really does what it's supposed to do.
   By reverse-engineering the behaviour and comparing to behaviour
   on the (known, reproducible)  random(9) stream.
   
   And for debugging, when it doesn't match.


these are special case scenarios, not "in general."
i don't see why it matters "in general."