HMAC appears to be working - we can set the key registers and stream data to the fifo and it produces the correct value in the digest registers.
It took me longer than intended to get the timing right :(
Also
Small simplification to sha256, that I can remove if annoying for current proof work
Changed the order of realignment and fifo in comparison to OpenTitan as it was simpler to encode in Cava2
There are some implicit assumptions such as not writing to fifo after CMD_PROCESS or when fifo is full (this should be obeyed by a correct operating host)
HMAC appears to be working - we can set the key registers and stream data to the fifo and it produces the correct value in the digest registers.
It took me longer than intended to get the timing right :(
Also