| Age | Commit message (Collapse) | Author |
|---|
|
- All cores use active-low reset now.
- TRNG cores are contiguous (but they still have their own mux, so occupy
a block of 16 cores).
- Use modexps6 instead of modexp.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
low priority, so just disable the optimization for now.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
cores. Tweak TRNG templates to support multiple instances, more for
consistency than than because we really expect multiple TRNGs.
|
|
|
|
different opinions about whether reset should be high or low, and the
core selector code is responsible for making this right. Hmm.
Address map may still be wrong, as addressing scheme seems to have
changed while the core_select branch was gathering dust.
|
|
|
|
|
|
The board_regs and comm_regs cores handle reset differently, but
there's also this sys_ena wire which appeared out of the ether one day
and is not yet in this movie. This version does NOT synthesize (nor
did the previous ones, but now we know it...).
|
|
|
|
|
|
|
|
additional methods for things like .h and .mk files.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
cores)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
testbench that are based on the real RSA key generated by Rob. Both tests works on the modexp in the Novena.
|
|
|
|
|
|
|
|
modulus.
|
