On Sat, May 5, 2018 at 4:26 PM, Kyösti Mälkki kyosti.malkki@gmail.com wrote:
On Sun, May 6, 2018 at 12:17 AM, Aaron Durbin adurbin@google.com wrote:
I find it particularly hard to be civil on your first question, so trying with sarcasm instead. After 5000 or so development hours and direct support from AMD, is the boot sequence for soc/stoneyridge prototypes equally bad, that is, AP CPUs execute through bootblock and verstage?
They currently jump to where they are told at the moment. See bootblock_c_entry() in src/soc/amd/stoneyridge/bootblock/bootblock.c. There is no running through multiple stages within coreboot.
I think CAR_GLOBAL equally fails with soc/amd/stoneyridge if you had SoC with multiple compute units where MTRRs are not shared, thus setup becomes asymmetric between some cores. Marshall's experiments support this happens.
Agreed. That's a concern w/ running APs through all the stages and not having a symmetric view of the address space.
I think you are right that guarding things w/ boot_cpu() would work. Is CONFIG_SQUELCH_EARLY_SMP set for all those types of boards? Or do we have a weird mix? Or should we have a Kconfig that says 'CAR' is not symmetric across cpus thus can't be used?
I would rather grant AP CPUs access to CAR_GLOBALs by manipulating the MTRR setup.
Kyösti
-
Aaron Durbin