Kyösti Mälkki has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/42073 )
Change subject: [WIP] sb,soc/amd: Allow dynamic ACPIMMIO base address ......................................................................
Patch Set 2:
(2 comments)
https://review.coreboot.org/c/coreboot/+/42073/2/src/soc/amd/common/block/in... File src/soc/amd/common/block/include/amdblocks/acpimmio_map.h:
https://review.coreboot.org/c/coreboot/+/42073/2/src/soc/amd/common/block/in... PS2, Line 47: #else
Are you thinking we would be directly assigning to the pointer objects for each mapping when running […]
I have no docs and did not search for the code in psp-verstage application how it is done there. But yes, I was thinking a one shot function early in the application to assign all the pointers.
I know some syscall like interface is there to set up the mapping and my interpretation was banks would have a constant address through the lifetime of the application.
I was also wondering how atomicity of read-modify-write ops to those banks is guaranteed.
https://review.coreboot.org/c/coreboot/+/42073/2/src/soc/amd/common/block/in... PS2, Line 83: #define ACPIMMIO_BASE(base, x) (void *)(base + ACPIMMIO_ ## x ## _BANK)
Why add the 'base' here? It also doesn't appear this macro is used any more? I think adding a commen […]
I had assumed the entire ACPIMMIO block of 8 KB would be accessible for psp-verstage application as a one mapping. It's one of the comments in patchset #1.