Change in coreboot[master]: ec/acpi: Refactor to reduce code duplication, use named constants

[Abel Briggs (Code Review)]

Attention is currently required from: Lean Sheng Tan.
Abel Briggs has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/64012 )

Change subject: ec/acpi: Refactor to reduce code duplication, use named constants
......................................................................

Patch Set 2:

(1 comment)

Patchset:

PS2:
> has this been tested with any hardware?

Not yet. I am currently in the process of rebasing the WIP patchset
for the (currently unsupported) Lenovo mainboard I have in order to
test the general refactoring changes in this patchset, but for other
affected boards, I would need to consult the mailing list for willing
testers to ensure there are no regressions. I can move this back to
WIP until I have testing artifacts, if that's preferred.

The rationale for this patchset was that I needed to factor out the EC
of `mb/lenovo/s230u` into its own standalone folder to reuse it for
my mainboard, but after seeing the current state of various standalone
ECs, I was not comfortable duplicating code.

Semantically, this patchset (at least, this commit) should behave
identically to master in cases where the UUT has a known-working
EC which follows the standard EC hardware API. If the EC is
non-functional, the `send` and `recv` functions will now avoid
reading and writing to I/O space.

The vast majority of callers on these functions do not read the return
value and effectively do not care about errors, so this patch should change
little (they're already invoking the equivalent of undefined behavior if
the EC fails). For the callers which do care about errors, we should try
to ensure correct behavior.

I could not find a rationale as to why the return statements were
commented out in the original implementation of these functions.