Nico Huber has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/36216 )
Change subject: soc/intel: common,skl,cnl,icl: drop reserved mmio memory size calculation ......................................................................
Patch Set 8:
(1 comment)
https://review.coreboot.org/c/coreboot/+/36216/5/src/soc/intel/common/block/... File src/soc/intel/common/block/systemagent/systemagent.c:
https://review.coreboot.org/c/coreboot/+/36216/5/src/soc/intel/common/block/... PS5, Line 187: - reserved_mmio_size;
Where did we land on the reserved ram vs reserved mmio differences? It's no difference for many of t […]
There is no difference. `reserved_mmio_size` always fills the gap between cbmem_top() and DPR exactly. I have confirmed that theory by reading all the `memmap.c` over again and again. And Michael confirmed it by testing.
The code in the memmaps isn't hard to follow btw. we always have
+-----------------------------+ TOLUD | traditional (including DPR) | +-----------------------------+ DPR base | reserved_mmio_size | +-----------------------------+ cbmem_top() | CBMEM |
I don't see anything between the reserved part and cbmem_top() in any combinations of options and platforms.
When the calculation here was introduced it made a difference. Platforms had soc_reserved_mmio_size() implemented step by step. But that wasn't necessary. The result is the same, everything from cbmem_top() to TSEG (inclusive) is marked as reserved RAM, everything above TSEG to TOLUD is marked as MMIO.
I really doubt there is anything to miss, so this change shouldn't alter anything. It just gets rid of the unnecessary boilerplate.