"Dr. Stephen Henson" <[email protected]> writes:

> On Wed, Aug 03, 2011, Henrik Grindal Bakken wrote:
>
>> 
>> Hi.  I'm working on FIPS-validating a product using OpenSSL (but with
>> a crypto module spanning wider, so we can't easily use the OpenSSL
>> crypto module).  During code review, some questions about the RNG
>> tests have come up.  Most specifically, from what I can read, SP
>> 800-90 requires that (in 11.3.4) the reseed function shall perform a
>> known-answer test before reseeding, and from what I can read in the
>> code, this doesn't happen.
>> 
>> Is there a reason for this?
>
> The OpenSSL implementation supports prediction resistance so 11.3.4
> #2 should apply in this case.
>
> I notice however that it doesn't directly test entropy failure on an
> explicit reseed: though it does on a reseed caused by a PR
> request. I'll update it to include an explicit test too.

Hi Steve, and thanks for your quick reply.

Since OpenSSL supports prediction resistance, wouldn't that mean that
11.3.4 #1 applies?  However, even if it's #1 or #2, shouldn't a KAT be
performed at some point before the reseed actually takes place?  I
must confess I don't really understand how SP 800-90 intends the KAT
to be carried out...



-- 
Henrik Grindal Bakken <[email protected]>
PGP ID: 8D436E52
Fingerprint: 131D 9590 F0CF 47EF 7963  02AF 9236 D25A 8D43 6E52
______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [email protected]
Automated List Manager                           [email protected]

Reply via email to