Re: [PATCH v24 00/12] /dev/random - a new approach with full SP800-90B compliance

Am Dienstag, 12. November 2019, 14:23:10 CET schrieb Florian Weimer:

Hi Florian,

* Stephan Müller:

quoted

* support deactivation of TRNG (i.e. blocking behavior of /dev/random)

  at compile time. If deactivated, /dev/random behaves like
  getrandom(2).

I don't quite understand this comment.  Doesn't getrandom with the
GRND_RANDOM always behave like /dev/random?  Presumably, without the
TRNG tap, the GRND_RANDOM flag for getrandom is ignored, and reading
from /dev/random behaves like reading from /dev/urandom.

Absolutely. Apologies for the imprecision here. I will correct that.

The idea is that the constant blocking behavior of /dev/random and GRND_RANDOM 
is replaced with the blocking behavior of getrandom(2) without the GRND_RANDOM 
flag (i.e. the interface waits until the LRNG thinks it is completely seeded 
before it provides ulimited data).

Anyway, reading the accompanying PDF, this looks rather impressive:
the userspace bootstrapping problem is gone (the issue where waiting
for more entropy prevents the collection of more entropy), *and* we
can still make the standards people happy.

(Replying from my other account due to mail issues, sorry.)

Ciao
Stephan