[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: heimdal and OpenSSL
On Fri, Jan 05, 2001 at 01:51:55AM +0100, Richard Levitte - VMS Whacker wrote:
> As an OpenSSL developer, I'd like to know what's the actual benefit of
> Heimdals des_new_random_key() over OpenSSLs des_random_key(). I
> understand it's based on a different PRNG, is that the only real
> difference, or does the scrambling with des_ecb_encrypt() have a
> meaning I can't detect yet?
Gee, after looking at the various sources more closely, I am beginning
to think that the history might be something like this:
KTH libdes: des_random_key a lame PRNG
SSLeay: des_random_key same PRNG as above
KTH libdes: des_new_random_key a replacement PRNG
OpenSSL 0.9.5: des_random_key basically a wrapper for RAND_bytes?
Which would imply that OpenSSL des_random_key and Heimdal
des_new_random_key are indeed interchangeable. It would be nice if
someone who knows could confirm whether or not I'm off the deep end
here.
> The actual main difference that I can detect is that the PRNG in
> Heimdals rnd_keys.c can take seeding from any of /dev/{,s,u}random
> (it's quite possible that I'll borrow some ideas for OpenSSL
> there...).
Hmm, OpenSSL only uses /dev/urandom by default? Isn't that dangerous?
I guess that's OK if you don't have to seed very often. I'll have to
try building it with /dev/random and see how it runs.
> In any case, since des_random_key() is provided in Heimdal for
> backward compatibility, can one assume that it and
> des_new_random_key() are actually interchangeable? The comments in
> Heimdals des.h seem to suggest that...
>
> In that case, it might be possible for us to provide
> des_new_random_key() as an entry point in OpenSSL. We'll see...
That would help, although one might as easily do this in the Heimdal
sources, or even just rename des_new_random_key to des_random_key in
the Heimdal sources. Yes, I like this latter approach.
--
Jacques Vidrine / n@nectar.com / jvidrine@verio.net / nectar@FreeBSD.org