Scalable random patchkit revisited

[Andi Kleen]

Hi Ted,

So far noone has come with a better solution, and the /dev/*random scaling
problem on large systems is still there and affecting workloads.

So I would like to resubmit the /dev/*random patchkit I posted some time
ago. Back then it was successfully reviewed, but then some alternative
proposals came up that nobody was interested in working on.

If new proposals are implemented they could be done on top of my patchkit
(even removing the extra pools if they are not needed anymore).

I forward ported the patches to the current master (4.5-rc*) and retested
it and it works just fine.

Please consider applying.

Thanks,
-Andi