Re: [RFC PATCH v3 0/7] Use ACPI PCI hot-plug for Q35

[Julia Suvorova]

On Thu, Oct 1, 2020 at 3:02 PM Ani Sinha <ani@anisinha.ca> wrote:
>
>
>
> On Thu, Oct 1, 2020 at 17:10 Michael S. Tsirkin <mst@redhat.com> wrote:
>>
>> On Thu, Oct 01, 2020 at 10:55:35AM +0200, Julia Suvorova wrote:
>> > On Thu, Sep 24, 2020 at 11:20 AM Michael S. Tsirkin <mst@redhat.com> wrote:
>> > >
>> > > On Thu, Sep 24, 2020 at 09:00:06AM +0200, Julia Suvorova wrote:
>> > > > The patch set consists of two parts:
>> > > > patches 1-4: introduce new feature
>> > > >              'acpi-pci-hotplug-with-bridge-support' on Q35
>> > > > patches 5-7: make the feature default along with changes in ACPI tables
>> > > >
>> > > > This way maintainers can decide which way to choose without breaking
>> > > > the patch set.
>> > > >
>> > > > With the feature disabled Q35 falls back to the native hot-plug.
>> > > >
>> > > > Pros
>> > > >     * no racy behavior during boot (see 110c477c2ed)
>> > > >     * eject is possible - according to PCIe spec, attention button
>> > > >       press should lead to power off, and then the adapter should be
>> > > >       removed manually. As there is no power down state exists in QEMU,
>> > > >       we cannot distinguish between an eject and a power down
>> > > >       request.
>> > > >     * no delay during deleting - after the actual power off software
>> > > >       must wait at least 1 second before indicating about it. This case
>> > > >       is quite important for users, it even has its own bug:
>> > > >           https://bugzilla.redhat.com/show_bug.cgi?id=1594168
>> > > >     * no timer-based behavior - in addition to the previous example,
>> > > >       the attention button has a 5-second waiting period, during which
>> > > >       the operation can be canceled with a second press. While this
>> > > >       looks fine for manual button control, automation will result in
>> > > >       the need to queue or drop events, and the software receiving
>> > > >       events in all sort of unspecified combinations of attention/power
>> > > >       indicator states, which is racy and uppredictable.
>> > > >     * fixes:
>> > > >         * https://bugzilla.redhat.com/show_bug.cgi?id=1752465
>> > > >         * https://bugzilla.redhat.com/show_bug.cgi?id=1690256
>> > > >
>> > > > Cons:
>> > > >     * lose per-port control over hot-plug (can be resolved)
>> > > >     * no access to possible features presented in slot capabilities
>> > > >       (this is only surprise removal AFAIK)
>> > >
>> > > something I don't quite get is whether with this one can still add
>> > > new pcie bridges and then hotplug into these with native
>> > > hotplug.
>> >
>> > Right now I disable native if there is acpihp anywhere, but even if
>> > you enable it for hotplugged devices, native hot-plug will not work.
>>
>> So that's a minor regression in functionality, right?
>> Why is that the case? Because you disable it in ACPI?
>> What if we don't?
>
>
> Stupid question. If both native and acpi is enabled which one does OS chose? 
> Does this choice vary between OSes and different flavours of the same OS like 
> Windows?

It will always choose native.

>>
>>
>> > > the challenge to answering this with certainty is that right now qemu
>> > > does not allow hotplugging complex devices such as bridges at all,
>> > > we only have a hack for multifunction devices.
>> > > Maybe adding a bridge as function 1 on command line, then hotplugging 
>> > > another device as
>> > > function 0 will work to test this.
>> > >
>> > >
>> > >
>> > > > v3:
>> > > >     * drop change of _OSC to allow SHPC on hotplugged bridges
>> > > >     * use 'acpi-root-pci-hotplug'
>> > > >     * add migration states [Igor]
>> > > >     * minor style changes
>> > > >
>> > > > v2:
>> > > >     * new ioport range for acpiphp [Gerd]
>> > > >     * drop find_pci_host() [Igor]
>> > > >     * explain magic numbers in _OSC [Igor]
>> > > >     * drop build_q35_pci_hotplug() wrapper [Igor]
>> > > >
>> > > > Julia Suvorova (7):
>> > > >   hw/acpi/pcihp: Enhance acpi_pcihp_disable_root_bus() to support Q35
>> > > >   hw/i386/acpi-build: Add ACPI PCI hot-plug methods to Q35
>> > > >   hw/pci/pcie: Do not initialize slot capability if acpihp is used
>> > > >   hw/acpi/ich9: Enable ACPI PCI hot-plug
>> > > >   bios-tables-test: Allow changes in DSDT ACPI tables
>> > > >   hw/acpi/ich9: Set ACPI PCI hot-plug as default
>> > > >   bios-tables-test: Update golden binaries
>> > > >
>> > > >  hw/i386/acpi-build.h              |   7 ++++
>> > > >  include/hw/acpi/ich9.h            |   5 +++
>> > > >  include/hw/acpi/pcihp.h           |   3 +-
>> > > >  hw/acpi/ich9.c                    |  67 ++++++++++++++++++++++++++++++
>> > > >  hw/acpi/pcihp.c                   |  16 ++++---
>> > > >  hw/acpi/piix4.c                   |   4 +-
>> > > >  hw/i386/acpi-build.c              |  31 ++++++++------
>> > > >  hw/i386/pc.c                      |   1 +
>> > > >  hw/pci/pcie.c                     |  16 +++++++
>> > > >  tests/data/acpi/q35/DSDT          | Bin 7678 -> 7950 bytes
>> > > >  tests/data/acpi/q35/DSDT.acpihmat | Bin 9002 -> 9274 bytes
>> > > >  tests/data/acpi/q35/DSDT.bridge   | Bin 7695 -> 9865 bytes
>> > > >  tests/data/acpi/q35/DSDT.cphp     | Bin 8141 -> 8413 bytes
>> > > >  tests/data/acpi/q35/DSDT.dimmpxm  | Bin 9331 -> 9603 bytes
>> > > >  tests/data/acpi/q35/DSDT.ipmibt   | Bin 7753 -> 8025 bytes
>> > > >  tests/data/acpi/q35/DSDT.memhp    | Bin 9037 -> 9309 bytes
>> > > >  tests/data/acpi/q35/DSDT.mmio64   | Bin 8808 -> 9080 bytes
>> > > >  tests/data/acpi/q35/DSDT.numamem  | Bin 7684 -> 7956 bytes
>> > > >  tests/data/acpi/q35/DSDT.tis      | Bin 8283 -> 8555 bytes
>> > > >  19 files changed, 129 insertions(+), 21 deletions(-)
>> > > >
>> > > > --
>> > > > 2.25.4
>> > >
>>