On Tue, 25 Jun 2024 12:57:11 +1000
"Nicholas Piggin" <npiggin@xxxxxxxxx> wrote:
[...]
> Hmm, you have a nice instr struct that you made earlier and now you're
> back to mask and shift... What about exposing that struct and add a
this is actually a nice idea, we could make a union of the structs for
the instruction formats we need, and then we can do what you propose
here
> function to create it so you could do grs[sblk_to_instr(sblk).r1]
probably something like _INSTR(sblk).${FORMAT}.r1
> here... Just a thought.
and maybe we could do this for the kernel as well
>
> Thanks,
> Nick
