Attention is currently required from: Lance Zhao, Raul Rangel, Martin Roth, Tim Wawrzynczak, Julius Werner. Patrick Georgi 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:
The order right now has been hand crafted and analyzed extensively. […]
How much is going on in ELOG or APOB? Does the prefetch use case require a buffer for writes to flash?
As in, there's a region in RAM where data to write to flash is cached in a simple format. After reads completed, the prefetch routine flips a flag somewhere. The next attempt to write to flash doesn't add to the buffer but writes out the buffer to flash before writing its own data. From there, the only extra cost is looking at the flag.
It adds a fair amount of complexity, but OTOH with something like this we could really just initialize a list of files to fetch and their target addresses early on, rather than hand crafting things to ensure that things are ready just in time without blocking anything.
(If the DMA engine supports it, it might even be feasible to queue SPI writes as DMA engine commands to process after the reads are completed, so that they're also completed in parallel to the main task.)