* Igor Mammedov (imamm...@redhat.com) wrote:
> On Wed, 6 Apr 2022 17:11:09 +0100
> "Dr. David Alan Gilbert" <dgilb...@redhat.com> wrote:
>
> > * Igor Mammedov (imamm...@redhat.com) wrote:
> > > On Wed, 6 Apr 2022 10:38:51 +0100
> > > "Dr. David Alan Gilbert" <dgilb...@redhat.com> wrote:
> > >
> > > > * Igor Mammedov (imamm...@redhat.com) wrote:
> > > > > On Wed, 6 Apr 2022 09:35:31 +0100
> > > > > "Dr. David Alan Gilbert (git)" <dgilb...@redhat.com> wrote:
> > > > >
> > > > > > From: "Dr. David Alan Gilbert" <dgilb...@redhat.com>
> > > > > >
> > > > > > The 'acpi_index' field is a statically configured field, which for
> > > > > > some reason is migrated; this never makes much sense because it's
> > > > > > command line static.
> > > > >
> > > > > that's true only for the field that's part of PCIDEvice,
> > > > > however AcpiPciHpState::acpi_index is runtime state and _must_
> > > > > be migrated if set, otherwise guest might get wrong index
> > > > > if it's in process of querying it
> > > >
> > > > So this patch only changes the piix4.c version; I'm confused, is there
> > > > a AcpiPciHpState::acpi_index that's runtime setable in there?
> > >
> > > > > > VMSTATE_PCI_HOTPLUG(acpi_pci_hotplug, PIIX4PMState,
> > > ^^^ AcpiPciHpState
> > > > > > vmstate_test_use_acpi_hotplug_bridge,
> > > > > > - vmstate_acpi_pcihp_use_acpi_index),
> > >
> > > hw/acpi/pcihp.c:pci_write():
> > > s->acpi_index = object_property_get_uint(o, "acpi-index", NULL);
> > >
> > > s->acpi_index is runtime value that is supposed to be migrated if it's set
> > > to something other then 0
> > >
> > > I may have botched VMSTATE_PCI_HOTPLUG, intent was to migrate
> > > AcpiPciHpState::acpi_index if necessary. But I'm not sure how
> > > if I used correct approach for to migrate an optional value
> > > i.e. maybe instead of VMSTATE_UINT32_TEST(pcihp.acpi_index, state,
> > > test_acpi_index)
> > > I should've used subsection, because destination has no clue if
> > > acpi_index would be transmitted over wire or not?
> > >
> > > > > > However, on piix4 it's conditional, and the condition/test function
> > > > > > ends up having the wrong pointer passed to it (it gets a
> > > > > > PIIX4PMState
> > > > > > not the AcpiPciHpState it was expecting, because VMSTATE_PCI_HOTPLUG
> > > > > > is a macro and not another struct). This means the field is
> > > > > > randomly
> > > > > > loaded/saved based on a random pointer. In 6.x this random pointer
> > > > > > randomly seems to get 0 for everyone (!); in 7.0rc it's getting junk
> > > > > > and trying to load a field that the source didn't send. The
> > > > > > migration
> > > > > > stream gets out of line and hits the section footer.
> > > > >
> > > > > I'm a bit confused by description,
> > > > > do you have a reproducer for me to try?
> > > >
> > > > Yeh, see the linked gitlab case command line:
> > > > https://gitlab.com/qemu-project/qemu/-/issues/932
> > > >
> > > > ./x86_64-softmmu/qemu-system-x86_64 -M pc-q35-6.2 -m 512 -device
> > > > virtio-scsi-pci,id=scsihw0,bus=pcie.0,addr=0x5,acpi-index=3 -drive
> > > > if=none,my.qcow2,format=qcow2,id=drive-scsi0,node-name=scsi0 -device
> > > > 'scsi-hd,bus=scsihw0.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0,id=scsi0'
> > > > -nographic
> > > >
> > > > just migrating from a 6.2 to a head triggers this.
> > >
> > > Over here any migration from qemu-6.2 to HEAD at 3d31fe4d662f13c7
> > > fails even without acpi-index, as simple as this:
> > >
> > > qemu-system-x86_64-6.2 -M pc-i440fx-6.2 -m 512 -vnc :0 -monitor stdio
> > > (qemu) stop
> > > (qemu) migrate "exec:gzip -c > STATEFILE.gz"
> > >
> > > qemu-system-x86_64-7.0 -M pc-i440fx-6.2 -m 512 -vnc :0 -monitor stdio
> > > -incoming "exec: gzip -c -d STATEFILE.gz"
> > >
> > > (qemu) qemu-system-x86_64-7.0: Missing section footer for
> > > 0000:00:01.3/piix4_pm
> > > qemu-system-x86_64-7.0: load of migration failed: Invalid argument
> > >
> > >
> > > Like you pointed out in gitlab issue, vmstate_acpi_pcihp_use_acpi_index
> > > is broken. Following applied to HEAD should fix immediate issue on
> > > destination
> > > reading random value:
> > >
> > > diff --git a/hw/acpi/piix4.c b/hw/acpi/piix4.c
> > > index f0b5fac44a..c97db491c8 100644
> > > --- a/hw/acpi/piix4.c
> > > +++ b/hw/acpi/piix4.c
> > > @@ -269,6 +269,11 @@ static bool piix4_vmstate_need_smbus(void *opaque,
> > > int version_id)
> > > return pm_smbus_vmstate_needed();
> > > }
> > >
> > > +static bool vmstate_piix4_need_acpi_index(void *opaque, int version_id)
> > > +{
> > > + PIIX4PMState *s = PIIX4_PM(opaque);
> > > + return vmstate_acpi_pcihp_use_acpi_index(&(s->acpi_pci_hotplug),
> > > version_id);
> > > +}
> >
> > But if acpi_index was set on the source 6.2 host, it won't send the
> > index, but the 7.0 would expect it, and it would fail in the same way
> > wouldn't it?
>
> With piix4 fixed up 7.0 won't expect field as s->acpi_index initialized to 0
> so check will always return 0 and the field won't be expected.
> ( testing confirms it).
> If test on 6.2 host somehow manages to return 1, destination won't
> be able to accept it, because it has no idea about it (that is not fixable,
> I'm afraid).
>
> For Q35 we set check to NULL
> VMSTATE_PCI_HOTPLUG(acpi_pci_hotplug,
> ICH9LPCPMRegs,
> NULL, NULL),
>
> which if I read vmstate_load_state() correctly will always expect
> the field and will always store fields since field->version_id == 0
> for VMSTATE_UINT32_TEST.
>
> So we can't remove field without breaking Q35.
Yes.
> Net effect:
> * not send the field for PC machine (ever)
> * send field always for Q35 (always)
>
> So your patch is good with fixed commit message
> and a comment close to the field that it's not really used with piix4
Could you write a new commit message based on mine?
> And to make migration of acpi_index on PC machine working,
> we need add an extra subsection that should be able to
> handle conditional value.
Yes; I hadn't realised acpi_index was actually writeable.
Dave
>
> >
> > Dave
> >
> > > /* qemu-kvm 1.2 uses version 3 but advertised as 2
> > > * To support incoming qemu-kvm 1.2 migration, change version_id
> > > * and minimum_version_id to 2 below (which breaks migration from
> > > @@ -299,7 +304,7 @@ static const VMStateDescription vmstate_acpi = {
> > > struct AcpiPciHpPciStatus),
> > > VMSTATE_PCI_HOTPLUG(acpi_pci_hotplug, PIIX4PMState,
> > > vmstate_test_use_acpi_hotplug_bridge,
> > > - vmstate_acpi_pcihp_use_acpi_index),
> > > + vmstate_piix4_need_acpi_index),
> > > VMSTATE_END_OF_LIST()
> > > },
> > > .subsections = (const VMStateDescription*[]) {
> > >
> > >
> > > > Dave
> > > >
> > > > > > The bodge is on piix4 never to load the field:
> > > > > > a) Most 6.x builds never send it, so most of the time the
> > > > > > migration
> > > > > > will work.
> > > > > > b) We can backport this fix to 6.x to remove the boobytrap.
> > > > > > c) It should never have made a difference anyway since the
> > > > > > acpi-index
> > > > > > is command line configured and should be correct on the
> > > > > > destination
> > > > > > anyway
> > > > > > d) ich9 is still sending/receiving this (unconditionally all the
> > > > > > time)
> > > > > > but due to (c) should never notice. We could follow up to make
> > > > > > it
> > > > > > skip.
> > > > > >
> > > > > > It worries me just when (a) actually happens.
> > > > > >
> > > > > > Fixes: b32bd76 ("pci: introduce acpi-index property for PCI device")
> > > > > > Resolves: https://gitlab.com/qemu-project/qemu/-/issues/932
> > > > > >
> > > > > > Signed-off-by: Dr. David Alan Gilbert <dgilb...@redhat.com>
> > > > > > ---
> > > > > > hw/acpi/acpi-pci-hotplug-stub.c | 4 ----
> > > > > > hw/acpi/pcihp.c | 6 ------
> > > > > > hw/acpi/piix4.c | 11 ++++++++++-
> > > > > > include/hw/acpi/pcihp.h | 2 --
> > > > > > 4 files changed, 10 insertions(+), 13 deletions(-)
> > > > > >
> > > > > > diff --git a/hw/acpi/acpi-pci-hotplug-stub.c
> > > > > > b/hw/acpi/acpi-pci-hotplug-stub.c
> > > > > > index 734e4c5986..a43f6dafc9 100644
> > > > > > --- a/hw/acpi/acpi-pci-hotplug-stub.c
> > > > > > +++ b/hw/acpi/acpi-pci-hotplug-stub.c
> > > > > > @@ -41,7 +41,3 @@ void acpi_pcihp_reset(AcpiPciHpState *s, bool
> > > > > > acpihp_root_off)
> > > > > > return;
> > > > > > }
> > > > > >
> > > > > > -bool vmstate_acpi_pcihp_use_acpi_index(void *opaque, int
> > > > > > version_id)
> > > > > > -{
> > > > > > - return false;
> > > > > > -}
> > > > > > diff --git a/hw/acpi/pcihp.c b/hw/acpi/pcihp.c
> > > > > > index 6351bd3424..bf65bbea49 100644
> > > > > > --- a/hw/acpi/pcihp.c
> > > > > > +++ b/hw/acpi/pcihp.c
> > > > > > @@ -554,12 +554,6 @@ void acpi_pcihp_init(Object *owner,
> > > > > > AcpiPciHpState *s, PCIBus *root_bus,
> > > > > > OBJ_PROP_FLAG_READ);
> > > > > > }
> > > > > >
> > > > > > -bool vmstate_acpi_pcihp_use_acpi_index(void *opaque, int
> > > > > > version_id)
> > > > > > -{
> > > > > > - AcpiPciHpState *s = opaque;
> > > > > > - return s->acpi_index;
> > > > > > -}
> > > > > > -
> > > > > > const VMStateDescription vmstate_acpi_pcihp_pci_status = {
> > > > > > .name = "acpi_pcihp_pci_status",
> > > > > > .version_id = 1,
> > > > > > diff --git a/hw/acpi/piix4.c b/hw/acpi/piix4.c
> > > > > > index cc37fa3416..48aeedd5f0 100644
> > > > > > --- a/hw/acpi/piix4.c
> > > > > > +++ b/hw/acpi/piix4.c
> > > > > > @@ -267,6 +267,15 @@ static bool piix4_vmstate_need_smbus(void
> > > > > > *opaque, int version_id)
> > > > > > return pm_smbus_vmstate_needed();
> > > > > > }
> > > > > >
> > > > > > +/*
> > > > > > + * This is a fudge to turn off the acpi_index field, whose
> > > > > > + * test was always broken on piix4.
> > > > > > + */
> > > > > > +static bool vmstate_test_never(void *opaque, int version_id)
> > > > > > +{
> > > > > > + return false;
> > > > > > +}
> > > > > > +
> > > > > > /* qemu-kvm 1.2 uses version 3 but advertised as 2
> > > > > > * To support incoming qemu-kvm 1.2 migration, change version_id
> > > > > > * and minimum_version_id to 2 below (which breaks migration from
> > > > > > @@ -297,7 +306,7 @@ static const VMStateDescription vmstate_acpi = {
> > > > > > struct AcpiPciHpPciStatus),
> > > > > > VMSTATE_PCI_HOTPLUG(acpi_pci_hotplug, PIIX4PMState,
> > > > > > vmstate_test_use_acpi_hotplug_bridge,
> > > > > > - vmstate_acpi_pcihp_use_acpi_index),
> > > > > > + vmstate_test_never),
> > > > > > VMSTATE_END_OF_LIST()
> > > > > > },
> > > > > > .subsections = (const VMStateDescription*[]) {
> > > > > > diff --git a/include/hw/acpi/pcihp.h b/include/hw/acpi/pcihp.h
> > > > > > index af1a169fc3..7e268c2c9c 100644
> > > > > > --- a/include/hw/acpi/pcihp.h
> > > > > > +++ b/include/hw/acpi/pcihp.h
> > > > > > @@ -73,8 +73,6 @@ void acpi_pcihp_reset(AcpiPciHpState *s, bool
> > > > > > acpihp_root_off);
> > > > > >
> > > > > > extern const VMStateDescription vmstate_acpi_pcihp_pci_status;
> > > > > >
> > > > > > -bool vmstate_acpi_pcihp_use_acpi_index(void *opaque, int
> > > > > > version_id);
> > > > > > -
> > > > > > #define VMSTATE_PCI_HOTPLUG(pcihp, state, test_pcihp,
> > > > > > test_acpi_index) \
> > > > > > VMSTATE_UINT32_TEST(pcihp.hotplug_select, state, \
> > > > > > test_pcihp), \
> > > > >
> > >
>
--
Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
