Re: [Qemu-devel] [PATCH v2] pc: Don't make die-id mandatory unless necessary

[Markus Armbruster]

Eduardo Habkost <address@hidden> writes:

> We have this issue reported when using libvirt to hotplug CPUs:
> https://bugzilla.redhat.com/show_bug.cgi?id=1741451
>
> Basically, libvirt is not copying die-id from
> query-hotpluggable-cpus, but die-id is now mandatory.
>
> We could blame libvirt and say it is not following the documented
> interface, because we have this buried in the QAPI schema
> documentation:
>
>> Note: currently there are 5 properties that could be present
>> but management should be prepared to pass through other
>> properties with device_add command to allow for future
>> interface extension. This also requires the filed names to be kept in
>> sync with the properties passed to -device/device_add.
>
> But I don't think this would be reasonable from us.  We can just
> make QEMU more flexible and let die-id to be omitted when there's
> no ambiguity.  This will allow us to keep compatibility with
> existing libvirt versions.
>
> Test case included to ensure we don't break this again.
>
> Fixes: commit 176d2cda0dee ("i386/cpu: Consolidate die-id validity in smp 
> context")
> Signed-off-by: Eduardo Habkost <address@hidden>
> ---
> Changes v1 -> v2:
> * v1 was "pc: Don't make CPU properties mandatory unless necessary"
> * Make only die-id optional (Igor Mammedov)
> ---
>  hw/i386/pc.c                             |  8 ++++++
>  tests/acceptance/pc_cpu_hotplug_props.py | 35 ++++++++++++++++++++++++
>  2 files changed, 43 insertions(+)
>  create mode 100644 tests/acceptance/pc_cpu_hotplug_props.py
>
> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> index 3ab4bcb3ca..9c3f6ae828 100644
> --- a/hw/i386/pc.c
> +++ b/hw/i386/pc.c
> @@ -2406,6 +2406,14 @@ static void pc_cpu_pre_plug(HotplugHandler 
> *hotplug_dev,
>          int max_socket = (ms->smp.max_cpus - 1) /
>                                  smp_threads / smp_cores / pcms->smp_dies;
>  
> +        /*
> +         * die-id was optional in QEMU 4.0 and older, so keep it optional
> +         * if there's only one die per socket.
> +         */

Is this comment worth its keep?

> +        if (cpu->die_id < 0 && pcms->smp_dies == 1) {
> +            cpu->die_id = 0;
> +        }
> +
>          if (cpu->socket_id < 0) {
>              error_setg(errp, "CPU socket-id is not set");
>              return;
> diff --git a/tests/acceptance/pc_cpu_hotplug_props.py 
> b/tests/acceptance/pc_cpu_hotplug_props.py
> new file mode 100644
> index 0000000000..08b7e632c6
> --- /dev/null
> +++ b/tests/acceptance/pc_cpu_hotplug_props.py
> @@ -0,0 +1,35 @@
> +#
> +# Ensure CPU die-id can be omitted on -device
> +#
> +#  Copyright (c) 2019 Red Hat Inc
> +#
> +# Author:
> +#  Eduardo Habkost <address@hidden>
> +#
> +# This library is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU Lesser General Public
> +# License as published by the Free Software Foundation; either
> +# version 2 of the License, or (at your option) any later version.
> +#
> +# This library is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +# Lesser General Public License for more details.
> +#
> +# You should have received a copy of the GNU Lesser General Public
> +# License along with this library; if not, see 
> <http://www.gnu.org/licenses/>.
> +#
> +
> +from avocado_qemu import Test
> +
> +class OmittedCPUProps(Test):
> +    """
> +    :avocado: tags=arch:x86_64
> +    """
> +    def test_no_die_id(self):
> +        self.vm.add_args('-nodefaults', '-S')
> +        self.vm.add_args('-smp', '1,sockets=2,cores=2,threads=2,maxcpus=8')
> +        self.vm.add_args('-cpu', 'qemu64')
> +        self.vm.add_args('-device', 
> 'qemu64-x86_64-cpu,socket-id=1,core-id=0,thread-id=0')
> +        self.vm.launch()
> +        self.assertEquals(len(self.vm.command('query-cpus')), 2)

This tests intentionally violates the "a CPU hot-plug with -device must
specify all the properties returned by query-hotpluggable-cpus" rule
laid down in qapi/machine.json, and the change to pc.c permits it.

Missing: a suitable update to the rule.

In addition, I'd like us to either work on making the rule stick in the
future (see my reply to Igor for an idea), or ditch the rule.  But
that's outside the scope of this regression fix.