poor default seeding of RNG

Tom Holroyd tomh at po.crl.go.jp
Wed Mar 14 17:18:29 EST 2001

On Wed, 14 Mar 2001, Damien Miller wrote:

> On Wed, 14 Mar 2001, Tom Holroyd wrote:
> > Correct me if I'm wrong, but init_rng() in entropy.c doesn't call
> > seed_rng(), and in fact seed_rng() isn't called from _anywhere_ (in
> It is called from arc4random_stir which is used fairly pervasively
> and also implicitly from arc4random.

Ah, yes, I posted too soon -- I forgot about the compatibility library,
thanks.  There is a call to seed_rng() in arc4random_stir(); however,
arc4random_stir() isn't called in sshconnect2.c (or I just haven't found
it yet :-). Should there be a call to arc4random_stir() in sshconnect2.c?

I guess it's quite possible that arc4random() would get called pretty soon
under most circumstances, but sshconnect1.c calls arc4random_stir(), so
it would be logical for ssh2connect.c to do so as well, or maybe just in

(Does arc4random_stir() really seed the BN_rand() generator on systems
that don't use the compatibility lib?)

Dr. Tom Holroyd
"I am, as I said, inspired by the biological phenomena in which
chemical forces are used in repetitious fashion to produce all
kinds of weird effects (one of which is the author)."
	-- Richard Feynman, _There's Plenty of Room at the Bottom_

More information about the openssh-unix-dev mailing list