[coreboot-gerrit] Change in coreboot[master]: Rename ECAM-specific MMCONF Kconfigs

[Nico Huber (Code Review)]

Attention is currently required from: Raul Rangel, Mariusz Szafrański, Jonathan Zhang, Stefan Reinauer, Kyösti Mälkki, Andrey Petrov, Patrick Rudolph, Anjaneya "Reddy" Chagam, Johnny Lin, Suresh Bellampalli, Morgan Jang, Michal Motyl, Alexander Couzens, Felix Held, Shelley Chen, Furquan Shaikh, Angel Pons, Lance Zhao, Jason Glenesk, Martin Roth, Damien Zammit, Lee Leahy, Marshall Dawson, Tim Wawrzynczak, Vanessa Eusebio, Huang Jin.
Nico Huber has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/57333 )

Change subject: Rename ECAM-specific MMCONF Kconfigs
......................................................................

Patch Set 8:

(2 comments)

Patchset:

PS5:
> +1. […]

My three questions above are still unanswered. And the parent commit
is flawed, I think it needs another Kconfig. So I'm still not convinced
that we need to rename anything at all.

File src/device/Kconfig:

https://review.coreboot.org/c/coreboot/+/57333/comment/bb2ffe5a_0cca2ce2 
PS8, Line 501: 
             : config PCI_NO_MMIO_OPS
             : 	bool
             : 	default n
             : 	help
             : 	  If this config is selected, the MMIO access for PCI config
             : 	  space is disabled.
             : 
The original option was just there to ease per-platform Kconfig selection.
Instead of `select MMCONF_SUPPORT` for 20+ times, it allowed us to have
a `default y` and only `select NO_MMCONF_SUPPORT` a few times. Hence, this
option should be renamed (if we rename at all) according to the other option,
just with the `NO_` prefix.
> MMIO ops and PCI_ECAM --> Use MMIO ops and config space is mapped in address space as defined by ECAM in PCIe spec.
> MMIO ops but !PCI_ECAM --> Use MMIO ops and config space is mapped in address space as defined by some platform specific mechanism.

Right. This shows there are actually two orthogonal questions: 1. Do we have
MMIO ops; 2. Do we have PCI_ECAM. But that also means that the parent commit
is flawed (it tries to act on a single Kconfig option) and we again discuss
things for naught here.