[coreboot-gerrit] [S] Change in coreboot[master]: cpu/x86/mp_init.c: Fix the logic to add AP cpus

[Bill XIE (Code Review)]

Attention is currently required from: Bill XIE, Arthur Heymans.

Hello Tim Wawrzynczak, Angel Pons, Arthur Heymans,

I'd like you to reexamine a change. Please visit

https://review.coreboot.org/c/coreboot/+/69215

to look at the new patch set (#2).

Change subject: cpu/x86/mp_init.c: Fix the logic to add AP cpus
......................................................................

cpu/x86/mp_init.c: Fix the logic to add AP cpus

commit 48c825ebd1beadb8ac4fd3ae687a36ff4705ab7a assumes that
(g_)cpu_bus->children->sibling represents the first AP cpu, but this
is not true when device tree (like mb/asus/p8x7x-series/devicetree.cb)
contains additional static lapic nodes. In such case, Linux kernel can
find one cpu unable to wake up, and an additional working cpu without
corresponding ACPI firmware node:

smpboot: Allowing 5 CPUs, 0 hotplug CPUs
    smpboot: CPU0: Intel(R) Xeon(R) CPU E3-1225 V2 @ 3.20GHz (family:
         0x6, model: 0x3a, stepping: 0x9)
    smpboot: do_boot_cpu failed(-1) to wakeup CPU#2

Now, in the end of init_bsp(), the last apic node is held on
g_last_apic, and all node representing an AP cpu should be inserted
and set up after it. The consistency lost with commit 48c825e should
be restored now.

Signed-off-by: Bill XIE 
Change-Id: I1075bc52091e619ef79cbeab0a0f6b57cff1d708
---
M src/cpu/x86/mp_init.c
1 file changed, 38 insertions(+), 4 deletions(-)

git pull ssh://review.coreboot.org:29418/coreboot refs/changes/15/69215/2