[Libre-soc-dev] daily kan-ban update 10jan2022

[Luke Kenneth Casson Leighton]

yesterday i fixed a bug in PortInterface and its usage where
addresses had been truncated to 48-bits.  this left linux cut
off which uses addresses with the top 2 bits set (0xc000....NNNN)
i re-ran the (3 hour) simulation and it *still* failed but this time
with a 0x300 exception.

today therefore i added a means to capture gtkwave traces from
a specific point, spotting a combination of PC and instruction will
trigger VCD dumps from that point onwards.

the simulation is now re-running yet again and will take another 2
hours to get to the exact same failure point.

clearly this is an O(N^2) progression that urgently needs some
dump-of-state and restore in order to not have this kind of latency.

l.

---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68