On most architectures, during vCPU hot-plug and hot-unplug actions,
firmware or the VMM/QEMU can convey vCPU status to the OS by
toggling the ACPI method `_STA.Present` Bit. However, certain CPU
architecture specifications prohibit [1] modifications to CPU's
`presence` after the kernel has booted.
This limitation [2][3] arises because many per-CPU components, such as
interrupt controllers that are closely integrated with CPUs, or other
per-CPU features, may not support reconfiguration once the kernel is
initialized. These components often cannot be powered down as they might
belong to an `always-on` power domain. Consequently, specifications
require all CPUs to remain `_STA.Present` after system initialization.
To address this issue from an ACPI perspective, an `is_enabled` state has
been introduced in the `AcpiCpuStatus`. This state reflects whether a vCPU
has been hot-plugged or hot-unplugged in QEMU, thus indicating its
availability to the guest kernel. The guest can evaluate this status by
calling the ACPI `_STA` method and checking the `_STA.Enabled` bit after
QEMU signals the hot-plug event via GED.
References:
[1] Check comment 5 in the bugzilla entry
Link: https://bugzilla.tianocore.org/show_bug.cgi?id=4481#c5
[2] KVMForum 2023 Presentation: Challenges Revisited in Supporting Virt CPU
Hotplug on
architectures that don’t Support CPU Hotplug (like ARM64)
a. Kernel Link:
https://kvm-forum.qemu.org/2023/KVM-forum-cpu-hotplug_7OJ1YyJ.pdf
b. Qemu Link:
https://kvm-forum.qemu.org/2023/Challenges_Revisited_in_Supporting_Virt_CPU_Hotplug_-__ii0iNb3.pdf
[3] KVMForum 2020 Presentation: Challenges in Supporting Virtual CPU Hotplug on
SoC Based Systems (like ARM64)
Link: https://kvmforum2020.sched.com/event/eE4m
Signed-off-by: Salil Mehta <salil.me...@huawei.com>
Reviewed-by: Gustavo Romero <gustavo.rom...@linaro.org>
---
hw/acpi/cpu.c | 9 ++++++++-
include/hw/acpi/cpu.h | 2 +-
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/hw/acpi/cpu.c b/hw/acpi/cpu.c
index 5cb60ca8bc..238a0edbc1 100644
--- a/hw/acpi/cpu.c
+++ b/hw/acpi/cpu.c
@@ -225,7 +225,14 @@ void cpu_hotplug_hw_init(MemoryRegion *as, Object *owner,
state->dev_count = id_list->len;
state->devs = g_new0(typeof(*state->devs), state->dev_count);
for (i = 0; i < id_list->len; i++) {
- state->devs[i].cpu = CPU(id_list->cpus[i].cpu);
+ struct CPUState *cpu = CPU(id_list->cpus[i].cpu);
+
+ state->devs[i].cpu = cpu;
+ if (cpu && DEVICE(cpu)->realized) {
+ state->devs[i].is_enabled = true;
+ } else {
+ state->devs[i].is_enabled = false;
+ }
state->devs[i].arch_id = id_list->cpus[i].arch_id;
}
memory_region_init_io(&state->ctrl_reg, owner, &cpu_hotplug_ops, state,
diff --git a/include/hw/acpi/cpu.h b/include/hw/acpi/cpu.h
index 32654dc274..ff49caa501 100644
--- a/include/hw/acpi/cpu.h
+++ b/include/hw/acpi/cpu.h
@@ -26,6 +26,7 @@ typedef struct AcpiCpuStatus {
uint64_t arch_id;
bool is_inserting;
bool is_removing;
+ bool is_enabled;
bool fw_remove;
uint32_t ost_event;
uint32_t ost_status;
@@ -74,5 +75,4 @@ extern const VMStateDescription vmstate_cpu_hotplug;
#define VMSTATE_CPU_HOTPLUG(cpuhp, state) \
VMSTATE_STRUCT(cpuhp, state, 1, \
vmstate_cpu_hotplug, CPUHotplugState)
-
#endif
--
2.34.1
