[libre-riscv-dev] daily kan-ban update 08jul2020

[Jacob Lifshay]

today:
finished adding support for mul* madd* and additional div* variants to
power-instruction-analyzer
the python bindings pass all tests

discovered I was accidentally running the code without using any
native inline-assembly because I forgot to enable the native_instrs
feature. changed it so the gen-output.sh script defaults to using
native instructions and can be passed a flag to disable native
instructions to allow it to run on non-PowerISA computers.

turns out POWER9 is weird, but not in the same way as qemu
for all the gory details, see:
https://salsa.debian.org/Kazan-team/power-instruction-analyzer/-/blob/master/src/instr_models.rs

Jacob