Arthur Heymans has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/36290 )
Change subject: arch/x86/cbmem.c: Drop API to save cbmem_top across stages ......................................................................
Patch Set 2:
Patch Set 2:
Patch Set 2:
Ok. I'll document the need for nvram store of cbmem_top. The question was a bit rhetorical as I think the AMD_S3LATE_PARAMS is just a poorly designed interface as it forces the caller to know in advance where some data will be, which is not the case for coreboot.
AMD/SAGE never got S3 right with AGESAv5. A bunch of things were requested during amd/stoneyridge reviews and only a fraction got fixed at first.
Nothing to do about that now... Do you think it's a good idea to make the need for backup/restore more explicit by explicitly calling them on both the post and resume paths instead of in the cbmem_top implementation, while still using the cbmem_top implementation I intend to use here to implement cbmem_top? IMO it improves the clarity of the bootflow a little, but I have no strong feelings about it.