Nico Huber has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/57627 )
Change subject: arch/x86,cpu/x86: Introduce new method for accessing cpu_info ......................................................................
Patch Set 7:
(1 comment)
Commit Message:
https://review.coreboot.org/c/coreboot/+/57627/comment/651a0eba_4d31054c PS6, Line 19: This means that when smm_do_relocation is executed, it is running in SMM
COOP_MULTITASKING doesn't change SMM code. It's only compiled into romstage/ramstage right now. […]
That's why I said "code compiled into SMM relocation". I don't care if it's "ramstage code" in our build system. I'm still not sure if I fully understand what is going on.
This combination of SMM relocation with code that calls cpu_info() is a new corner case, right? To me it seems like that's an issue of the caller and not of cpu_info(). Yet, your commit message states the opposite. IMO, we have to be very careful about commit messages that state too bluntly that something is broken. It will likely confuse people when they read the history later.
Please don't merge commits when the commit message is not clear yet. IMHO, review can't even start before it is (unless the changes don't need any explanation of course). If the background is not fully understood, reviewers can't suggest alternatives, for instance.