Attention is currently required from: Lance Zhao, Raul Rangel, Patrick Georgi, Martin Roth, Tim Wawrzynczak. Julius Werner has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/59504 )
Change subject: acpi,Makefile: Add preload_acpi_dsdt ......................................................................
Patch Set 2:
(1 comment)
Patchset:
PS2:
IMO, it needs to be at a lower layer. APOB prefetch (start_apob_cache_read) only uses fmap_locate_area_as_rdev, so making it call into cbfs gets a little strange... Come to think of it, ELOG also uses a raw region.
Right, I didn't say CBFS, I just said it should be its own API rather than being integrated into rdev_writeat() under the hood. More like an outside wrapper around the rdev functions.
Now that it's decoupled we can in theory resize the thread stack size.
That sounds dangerous, though. coreboot programmers make certain assumptions about how much stack usage is okay vs too much. It's not always obvious (and may change after the fact) whether a function you're writing may be called from a special worker thread or not. Detect stack overflows after the fact at runtime is also quite brittle and dangerous to rely on. I'd much rather we don't create new stack overflow hazards and instead optimize usage of the stacks we have by being frugal with the amount of threads we create.
-
Julius Werner (Code Review)