[Cryptography] OpenSSL CSPRNG work

[Patrick Chkoreff]

John Denker via cryptography wrote on 07/06/2017 07:19 PM:

>> assuming it was seeded with at least 128 unpredictable bits when first initiated
> 
> That assumes away the root of the problem.

It emphasizes the nature of the solution.

Unfortunately the solution sounds difficult, as Theodore points out.  On
the subject of continuity, he says:

> If we can build a way of getting that seed passed into the kernel from
> the bootloader, that will probably go a long way towards solving much
> of the problem. 

On the subject of initiation, he says:

> You still have to solve the problem of how do you reliable and secure
> the seed when the system is booted for the first time right after it
> has been unpacked from the box.

That's an especially pressing problem for those who wish to generate SSH
keys within milliseconds of booting a brand new machine fresh out of the
packing foam -- though I've never quite understood the need to do that.
("Do Not Generate Keys Before The Machine Has Booted Up!" -- Bear)

On my own machine, I suppose that upon booting up every time I could
roll 52 6-sided dice and echo the result into /dev/random, but that
doesn't sound very user-friendly.  The problem is knowing when the
equivalent of that process has been achieved from internal randomness,
and blocking like a stubborn mule until that has happened.


-- Patrick