Attention is currently required from: Arthur Heymans, Christian Walter, Johnny Lin, Jonathan Zhang, Lean Sheng Tan, Nico Huber, Patrick Rudolph, Tim Chu.
Shuo Liu has posted comments on this change by Shuo Liu. ( https://review.coreboot.org/c/coreboot/+/83538?usp=email )
Change subject: soc/intel/xeon_sp: Reserve MMIO high range ......................................................................
Patch Set 2:
(4 comments)
Patchset:
PS2:
Side note: This whole situation is becoming too messy for my taste. Especially […]
The cause is around FSP's two layer IIO resource allocation structure (IIO stack -> PCIe domains).
I need to bring the inputs internally for some discussion, while this CL could help in SPR and GNR timeframe to save further efforts here.
Another aspect is that, if other SoCs has similar needs of using MMIO high, maybe some common changes could be justified then?
Commit Message:
https://review.coreboot.org/c/coreboot/+/83538/comment/18c53a6c_2ad7ddf1?usp... : PS1, Line 7: soc/intel/xeon_sp: Reserve FSP MMIO high window
GNR and later SoC also fit. […]
Sorry for being busy with some affairs recent days and I'm glad that some dedicated time today could be reserved for discussion :-)
For 'non PCI domain MMIOs', they are used for misc devices belonged to an IIO stack but not belonged to any PCIe domains under that stack. For Gen6, VTd BAR is counted to this class, similarly is the CXL host bridge control register BAR.
Not like Gen1 (https://review.coreboot.org/c/coreboot/+/83136), in Gen6, all these misc BARs do not needs to be explicitly reserved anymore.
This CL (https://review.coreboot.org/c/coreboot/+/83538) is initially raised for SPR, but also will help GNR. I updated the commit message and not sure if that could be clearer.
https://review.coreboot.org/c/coreboot/+/83538/comment/80757e69_48f78c7c?usp... : PS1, Line 19: is especially important : on systems with 2 or more sockets, where each socket has multiple : domains.
Well, you could also tell it to use WB default instead. But I get it, Linux […]
Sure, updated the commit message.
File src/soc/intel/xeon_sp/uncore.c:
https://review.coreboot.org/c/coreboot/+/83538/comment/4c177fbc_9cc62e2b?usp... : PS2, Line 371: * MMIO high window has to be added in set_resources instead of read_resources : * due to that adding in read_resources will cause the whole window reserved : * and cannot be used for resource allocation.
Done