Nico Huber has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/35763 )
Change subject: soc/intel/skylake: rework and fix platform detection ......................................................................
Patch Set 8:
(1 comment)
I would prefer to treat the platform detection and changes to the RP swapping in separate commits. The former can be generally useful, e.g. as simple functions is_spt() is_unp() (the LP can actually be derived from Kconfig, could still implement a func- tion for that, that acts on the config setting).
https://review.coreboot.org/c/coreboot/+/35763/7//COMMIT_MSG Commit Message:
https://review.coreboot.org/c/coreboot/+/35763/7//COMMIT_MSG@17 PS7, Line 17: this problem in general, depends on platform detection.
Done
Ok, I'm starting to understand what you say. An example would have made this more clear. It fails to call pcie_update_device_tree() if all ports of the first group are disabled (otherwise we would not read 0xffff but the swapped IDs)... Which is really not what you said, but a potential problem.
So the real problem is that we make decisions for all groups based on the state of one group...