mac/crc.py: optimize CRC validation

[rowanG077]

This is a WIP!! Do not merge yet. I need to validate it first.

The idea here is that we can pre-calculate different residues for a zero padded stream. We can just throw the full width into the full width engine, while making sure we set invalid bytes to 0 and then compare with the appropriate check value. This saves us from having to instantiate multiple crc engines on the receive path, which is nice since they are relatively expensive.

I did not do hardware tests yet. I will post resource usage differences later once I have validated it works

[rowanG077]

By default this is not actually more efficient on yosys. This is due to YosysHQ/yosys#4293 closing because if it does not provide the benefit by default it might as well not be done.