[libre-riscv-dev] [Bug 186] Create decoder for SOC: Power ISA and RISC-V

[bugzilla-daemon at libre-riscv.org]

http://bugs.libre-riscv.org/show_bug.cgi?id=186

--- Comment #52 from Michael Nolan <mtnolan2640 at gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #50)
> (In reply to Michael Nolan from comment #49)
> > (In reply to Luke Kenneth Casson Leighton from comment #47)
> > 
> > > 
> > > the idea being, we just have to write a tree-like array of specifications,
> > > pass that into *one* PowerDecoder, and it will (recursively) pass the next
> > > part of the spec-array to more PowerDecoder instances.
> > > 
> > > otherwise we need to set that up "by hand"... with a top-level hand-coded
> > > Switch / Case statement, which, once done, i guarantee you will go,
> > > "hmm, this looks pretty much exactly like a PowerDecoder instance" :)
> > > 
> > > you want to have a go at that?
> > 
> > I have something like that semi-working in b2bb2b, 
> 
> nice, will take a look.
> 
> > but it doesn't include
> > extra.csv or minor_19_00000. 
> 
> ok.
> 
> 
> > I get what extra.csv is for, though I'm not entirely sure how to integrate
> > it with the decoder.
> 
> yeah it is effectively at the same "level" as the major opcodes, and needs
> to be done.. mmmm... before them?
> 
> in effect it is extra case statements that need to go into major.csv.
> 
> the only "problem" is, the widths are different.  therefore, we need to have
> *two* switch statements.
> 
> i *think* this is doable by making all args of PowerDecode into lists,
> disabling the "default" switch, and initialising the "default" at the
> beginning.
> 
> > What is minor_19_00000.csv for though?
> 
> just as extra.csv is to be added in at the same level as major.csv,
> minor_19_00000 is likewise needing to be added to minor_19 ad a 2nd set of
> switches, but with *different* widths.
> 
> hence the idea of making PowerDecode args into lists then putting a for-loop
> around the switch generation.
> 
> what do you think?

I see now. So instead of passing in the bit selection and csv name, we'd pass
in a list of bit selections and csvs, and it would cascade to the next csv if
the first one doesn't match?

-- 
You are receiving this mail because:
You are on the CC list for the bug.