[Monotone-devel] on usage of /dev/random

[Lapo Luchini]

Just to avoid losing track of the issue, discussed shortly on IRC:
(slightly edited for shortness and clarity, see the links for the
original sentences)

http://colabti.org/irclogger/irclogger_log/monotone?date=2009-11-03
tommyd  I'm not an expert in this area, but maybe somebody else
        could reply to this? http://savannah.nongnu.org/bugs/?27929
thm     the question is, do we need a rng for anonynous pull at all?

http://colabti.org/irclogger/irclogger_log/monotone?date=2009-11-04
lapo    I'm not a fan of a blocking /dev/random per se (as long as it
        is a CSPRNG like Yarrow or -better- Fortuna, it doesn't need
        to block, new entroy is only necessary to have forward secrecy),
        but I do agree that we ought to use /dev/random only
        to generate keys at most… or even not at all (?)

http://colabti.org/irclogger/irclogger_log/monotone?date=2009-11-05
zwol    which rng issue?
thm     we use the rng without need
zwol    well, we shouldn't
thm     thus emptying the entropy pool
zwol    um, I think the network layer needs it for nonces?
        arguably it should use /dev/urandom for that but I don't
        think botan exposes the difference
        and /dev/urandom drains the entropy pool too

-- 
Lapo Luchini - http://lapo.it/

“Anyone who considers protocol unimportant has never dealt with a cat.”
(Robert A. Heinlein, , "The Cat Who Walks Through Walls", 1985)