[Cryptech Tech] [Cryptech-Commits] [core/platform/novena] 21/21: Sick hacks to compensate for sparse MUX within TRNG core.

[Paul Selkirk]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 09/29/2015 11:33 AM, Joachim Strömbergson wrote:
> 
> I think we want to keep the internal mux of the TRNG.
> 
> We could strip away all API functionalities from the different sub 
> modules (avalanche_entropy, rosc_entropy, mixer (and sha512) and
> csprng (and chacha)) and make a single trng.v that don't expose the
> name, versions of the sub modules and compresses the space.
> 
> The reason why we (imho) should keep the structure is that it
> makes switching, adding and modifying the structure easier. Adding
> a third entropy provider, changing the mixer. But otoh, removing
> the mux and headers would save space and is not really hard to do.
> Mabye we should bite the bullet.

As I understand it, the reason for giving each TRNG sub-module its own
register address space was to be able to tap the raw entropy/mixer
state. Can we fit the relevant registers into a 0x100 block under
trng? If so, I would say go for it.

				paul
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJWDGqCAAoJELAqzRn9CEWT0JsH/RmWvzhrVof2/tah4j/kJ8ea
+30dQrvZfnmSkPR8w11SiNeD783qj5yOD+eF+Y5+LGH12HHX1cF+Ij28PFd95O+I
becS2nLdiA2xETUXi12TVbWj273VEGlH8ovw1V1x5KeDKHiR+YXifP8hpf/ZI8ru
HKhCxRddQOW7pAyk2mYabuSXm2f04lSCPk7/5rsKRsBf5FbUuvrMZDTn6tF0tsdk
mLTUlipYr32TUBsyEI+pyabGFTwW4MEyYr1Nd7B5WnvsSupjvLmZKSGdzEyrEI1W
T9Sc4Q7Ug9wT6uOVAnZtjlixPHh6wK3GqTPPwKIIxef35JzavRP4V6c2H9R1m80=
=wxqp
-----END PGP SIGNATURE-----