Attention is currently required from: Andrey Petrov, Arthur Heymans, Frans Hendriks, Paul Menzel, Yu-Ping Wu.
Julius Werner has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/63419?usp=email )
Change subject: Don't build a separate romstage by default on most x86 targets ......................................................................
Patch Set 55:
(1 comment)
File src/Kconfig:
https://review.coreboot.org/c/coreboot/+/63419/comment/755f49f3_50a3a1e9 : PS11, Line 1389: default n if ARCH_X86
No, I think my point was rather about what to base the heuristic upon for where it makes sense to […]
(Sorry for the delay, this must have slipped by me somehow.)
I'm still trying to figure out _why_ this should be `default n if ARCH_X86`. I mean, if you're saying that for all current x86 boards it happens to be the best choice that may be true, but it still doesn't seem like a good idea to just tie it to that because future x86 boards may be governed by different factors. I think it would be better to tie it to something that is logically connected to the performance difference from separating the ramstage, like the compression/XIP thing (and probably also add a comment to document the reasoning for that).
If you're saying it also makes sense for AMD then maybe that should be a separate consideration. We don't have a `CONFIG_BOOTBLOCK_IS_AUTOMATICALLY_DECOMPRESSED_BY_PLATFORM`, and this seems to be a pretty unique case so it may not be worth to add it just for this... so maybe just `default n if !NO_XIP_EARLY_STAGES || RESET_VECTOR_IN_RAM`?