public inbox for qemu-devel@nongnu.org
 help / color / mirror / Atom feed
From: Zhao Liu <zhao1.liu@intel.com>
To: Peter Maydell <peter.maydell@linaro.org>,
	Igor Mammedov <imammedo@redhat.com>
Cc: "Paolo Bonzini" <pbonzini@redhat.com>,
	qemu-devel@nongnu.org, "Fabiano Rosas" <farosas@suse.de>,
	"Peter Xu" <peterx@redhat.com>,
	"Philippe Mathieu-Daudé" <philmd@linaro.org>,
	"Aurelien Jarno" <aurelien@aurel32.net>,
	"Zhao Liu" <zhao1.liu@intel.com>
Subject: Re: [PULL 04/41] pc: Start with modern CPU hotplug interface by default
Date: Tue, 17 Mar 2026 12:23:46 +0800	[thread overview]
Message-ID: <abjXUtJNrcmSFUP4@intel.com> (raw)
In-Reply-To: <CAFEAcA-O4g5ZrRdK1s5kBjMfRbWhPnD0hLZwhpVHVHVbE7LKdQ@mail.gmail.com>

+Philippe & Aurelien

(And seeking Igor's input)

> The problem is that the vmstate_cpu_hotplug VMStateDescription
> that is being used has a NULL "fields" (it is the "stub" one).
> This used to be OK because the "piix4_pm/cpuhp" vmstate had
> a "needed" function that would return false for Malta, so we
> never tried to use the stub code.
> 
> What's the intention here ? Shouldn't we still have a "needed"
> function that returns "false" for the cases where we would be
> using the stub version of vmstate_cpu_hotplug ? Otherwise we
> would be changing the migration data. We can permit a compat
> break for MIPS boards, but the commit message suggests that
> this wasn't intended to be a compat break.

Thanks, this is my bad - I didn't realize it would break Malta. The
"needed" method looks necessary, but previously making the "stub" one
depend on cpu_hotplug_legacy=true also sounds a bit not proper.

Maybe in the needed() method we should also check whether the board/
machine supports CPU hotplug: check mc->has_hotpluggable_cpus. IIUC,
Malta doesn't support this.

> In this thread Fabiano suggests a fix which is probably OK
> if we want to take the "break migration compat" route:
> https://lore.kernel.org/qemu-devel/CAFEAcA9BzSp9F8yv4eZv5eeUM4JXSy9T-QnRr_UbJZpNmhtHyw@mail.gmail.com/T/#m6f4a672ddfa8a0a34dfdf9b57bb04e529ab5c9a4

I think Fabiano's patch is a clean solution. Or I think there's another
option: we can bring needed() back but check mc->has_hotpluggable_cpus
instead, like acpi-ged/cpuhp did ().

Igor, what do you think?

---
diff --git a/hw/acpi/ich9.c b/hw/acpi/ich9.c
index bbb1bd60a206..5c7dfb2c69db 100644
--- a/hw/acpi/ich9.c
+++ b/hw/acpi/ich9.c
@@ -184,10 +184,18 @@ static const VMStateDescription vmstate_tco_io_state = {
     }
 };

+static bool cpuhp_needed(void *opaque)
+{
+    MachineClass *mc = MACHINE_GET_CLASS(qdev_get_machine());
+
+    return mc->has_hotpluggable_cpus;
+}
+
 static const VMStateDescription vmstate_cpuhp_state = {
     .name = "ich9_pm/cpuhp",
     .version_id = 1,
     .minimum_version_id = 1,
+    .needed = cpuhp_needed,
     .fields = (const VMStateField[]) {
         VMSTATE_CPU_HOTPLUG(cpuhp_state, ICH9LPCPMRegs),
         VMSTATE_END_OF_LIST()
diff --git a/hw/acpi/piix4.c b/hw/acpi/piix4.c
index 43860d122780..9b7f50c7afac 100644
--- a/hw/acpi/piix4.c
+++ b/hw/acpi/piix4.c
@@ -195,10 +195,18 @@ static const VMStateDescription vmstate_memhp_state = {
     }
 };

+static bool cpuhp_needed(void *opaque)
+{
+    MachineClass *mc = MACHINE_GET_CLASS(qdev_get_machine());
+
+    return mc->has_hotpluggable_cpus;
+}
+
 static const VMStateDescription vmstate_cpuhp_state = {
     .name = "piix4_pm/cpuhp",
     .version_id = 1,
     .minimum_version_id = 1,
+    .needed = cpuhp_needed,
     .fields = (const VMStateField[]) {
         VMSTATE_CPU_HOTPLUG(cpuhp_state, PIIX4PMState),
         VMSTATE_END_OF_LIST()





  reply	other threads:[~2026-03-17  3:58 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-02-12 14:42 [PULL 00/41] QOM, target/i386 patches for 2026-02-12 Paolo Bonzini
2026-02-12 14:42 ` [PULL 01/41] target/i386: convert SEV-ES termination requests to guest panic events Paolo Bonzini
2026-02-12 14:42 ` [PULL 02/41] hw/i386/pc: Remove deprecated pc-q35-2.6 and pc-i440fx-2.6 machines Paolo Bonzini
2026-02-12 14:42 ` [PULL 03/41] tests/acpi: Allow DSDT table change for x86 machines Paolo Bonzini
2026-02-12 14:42 ` [PULL 04/41] pc: Start with modern CPU hotplug interface by default Paolo Bonzini
2026-03-16 16:01   ` Peter Maydell
2026-03-17  4:23     ` Zhao Liu [this message]
2026-03-17 14:21       ` Peter Maydell
2026-02-12 14:42 ` [PULL 05/41] hw/i386/pc: Remove PCMachineClass::legacy_cpu_hotplug field Paolo Bonzini
2026-02-12 14:42 ` [PULL 06/41] acpi: Remove legacy cpu hotplug utilities Paolo Bonzini
2026-02-12 14:42 ` [PULL 07/41] docs/specs/acpi_cpu_hotplug: Remove legacy cpu hotplug descriptions Paolo Bonzini
2026-02-12 14:42 ` [PULL 08/41] tests/acpi: Update DSDT tables for pc & q35 machines Paolo Bonzini
2026-02-12 14:42 ` [PULL 09/41] hw/nvram/fw_cfg: Rename fw_cfg_init_mem() with '_nodma' suffix Paolo Bonzini
2026-02-12 14:42 ` [PULL 10/41] hw/mips/loongson3_virt: Prefer using fw_cfg_init_mem_nodma() Paolo Bonzini
2026-02-12 14:42 ` [PULL 11/41] hw/nvram/fw_cfg: Factor fw_cfg_init_mem_internal() out Paolo Bonzini
2026-02-12 14:42 ` [PULL 12/41] hw/nvram/fw_cfg: Rename fw_cfg_init_mem_wide() -> fw_cfg_init_mem_dma() Paolo Bonzini
2026-02-12 14:42 ` [PULL 13/41] hw/i386/x86: Remove X86MachineClass::fwcfg_dma_enabled field Paolo Bonzini
2026-02-12 14:42 ` [PULL 14/41] hw/i386/pc: Remove multiboot.bin Paolo Bonzini
2026-02-12 14:42 ` [PULL 15/41] hw/i386: Assume fw_cfg DMA is always enabled Paolo Bonzini
2026-02-12 14:42 ` [PULL 16/41] hw/i386: Remove linuxboot.bin Paolo Bonzini
2026-03-24 12:12   ` Daniel P. Berrangé
2026-03-24 12:14     ` Paolo Bonzini
2026-02-12 14:42 ` [PULL 17/41] hw/i386/pc: Remove pc_compat_2_6[] array Paolo Bonzini
2026-02-12 14:42 ` [PULL 18/41] hw/intc/apic: Remove APICCommonState::legacy_instance_id field Paolo Bonzini
2026-02-12 14:42 ` [PULL 19/41] hw/core/machine: Remove hw_compat_2_6[] array Paolo Bonzini
2026-02-12 14:42 ` [PULL 20/41] hw/virtio/virtio-mmio: Remove VirtIOMMIOProxy::format_transport_address field Paolo Bonzini
2026-02-12 14:42 ` [PULL 21/41] hw/i386/pc: Remove deprecated pc-q35-2.7 and pc-i440fx-2.7 machines Paolo Bonzini
2026-02-12 14:42 ` [PULL 22/41] hw/i386/pc: Remove pc_compat_2_7[] array Paolo Bonzini
2026-02-12 14:42 ` [PULL 23/41] target/i386/cpu: Remove CPUX86State::full_cpuid_auto_level field Paolo Bonzini
2026-02-12 14:42 ` [PULL 24/41] hw/audio/pcspk: Remove PCSpkState::migrate field Paolo Bonzini
2026-02-12 14:42 ` [PULL 25/41] hw/core/machine: Remove hw_compat_2_7[] array Paolo Bonzini
2026-02-12 14:42 ` [PULL 26/41] hw/i386/intel_iommu: Remove IntelIOMMUState::buggy_eim field Paolo Bonzini
2026-02-12 14:42 ` [PULL 27/41] hw/virtio/virtio-pci: Remove VirtIOPCIProxy::ignore_backend_features field Paolo Bonzini
2026-02-12 14:42 ` [PULL 28/41] hw/char/virtio-serial: Do not expose the 'emergency-write' property Paolo Bonzini
2026-02-16 17:21   ` Alexander Bulekov
2026-02-12 14:42 ` [PULL 29/41] i386/cpu: Fix incorrect initializer in Diamond Rapids definition Paolo Bonzini
2026-02-12 14:42 ` [PULL 30/41] target/i386/kvm: set KVM_PMU_CAP_DISABLE if "-pmu" is configured Paolo Bonzini
2026-02-12 14:42 ` [PULL 31/41] target/i386/kvm: extract unrelated code out of kvm_x86_build_cpuid() Paolo Bonzini
2026-02-12 14:42 ` [PULL 32/41] target/i386/kvm: rename architectural PMU variables Paolo Bonzini
2026-02-12 14:42 ` [PULL 33/41] target/i386/kvm: reset AMD PMU registers during VM reset Paolo Bonzini
2026-02-12 14:42 ` [PULL 34/41] target/i386/kvm: support perfmon-v2 for reset Paolo Bonzini
2026-02-12 14:42 ` [PULL 35/41] qdev: Free property array on release Paolo Bonzini
2026-02-12 14:42 ` [PULL 36/41] qdev: make release_string() idempotent Paolo Bonzini
2026-02-12 14:42 ` [PULL 37/41] qdev: make release_drive() idempotent Paolo Bonzini
2026-02-12 14:42 ` [PULL 38/41] qdev: make release_tpm() idempotent Paolo Bonzini
2026-02-12 14:42 ` [PULL 39/41] tests: add /qdev/free-properties test Paolo Bonzini
2026-02-12 14:42 ` [PULL 40/41] accel/mshv: Remove remap overlapping mappings code Paolo Bonzini
2026-02-12 14:42 ` [PULL 41/41] target/i386/mshv: remove unused optimization of gva=>gpa translation Paolo Bonzini
2026-02-12 16:26 ` [PULL 00/41] QOM, target/i386 patches for 2026-02-12 Peter Maydell

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=abjXUtJNrcmSFUP4@intel.com \
    --to=zhao1.liu@intel.com \
    --cc=aurelien@aurel32.net \
    --cc=farosas@suse.de \
    --cc=imammedo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=peterx@redhat.com \
    --cc=philmd@linaro.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox