Troy Monaghen wrote:
>1) I don't see any Diffie-Hellman code in the fips part of the source
>tree except for the dh_test() function in fips_test_suite.c. Will DH be
>available to use in an application that will be running in FIPS mode
>without violating the security policy?
DH will not be available in FIPS mode (the reference in fips_test_suite.c
is to prove to NIST that it is disabled in FIPS mode). Leaving it out
was, frankly, a stupid oversight. No testing would have been needed, we
would only have had to put the relevant source files in the ./fips/ tree
with an entry in the fingerprint.sha1 file. But by the time we realized
this oversight we were out of time, money, and patience.
This FIPS-140 validation is *very* time consuming, with a lot of false
starts and rework. Ben and I have been working on this for over 18
months, with both of us donating a _lot_ more time than we ever intended
to. We just ran out of steam.
Now that this effort is close to completion we are seeing some interest
in expanding the scope of the validation. With the difficult and
expensive part already done those requests will be relatively easy to
satisfy with contributions of money and/or labor. And I would expect to
include DH in any follow-on validation.
-Steve M.
