On 22.01.19 16:03, Christian Borntraeger wrote: > > > On 22.01.2019 13:52, David Hildenbrand wrote: >> On 22.01.19 13:44, Cornelia Huck wrote: >>> On Tue, 22 Jan 2019 10:41:43 +0100 >>> David Hildenbrand <[email protected]> wrote: >>> >>>> We decided to always create the PCI host bridge, even if 'zpci' is not >>>> enabled (due to migration compatibility). This however right now allows >>>> to add zPCI/PCI devices to a VM although the guest will never actually see >>>> them, confusing people that are using a simple CPU model that has no >>>> 'zpci' enabled - "Why isn't this working" (David Hildenbrand) >>>> >>>> Let's check for 'zpci' and at least print a warning that this will not >>>> work as expected. We could also bail out, however that might break >>>> existing QEMU commandlines. >>>> >>>> Signed-off-by: David Hildenbrand <[email protected]> >>>> --- >>>> hw/s390x/s390-pci-bus.c | 5 +++++ >>>> 1 file changed, 5 insertions(+) >>>> >>>> diff --git a/hw/s390x/s390-pci-bus.c b/hw/s390x/s390-pci-bus.c >>>> index b86a8bdcd4..e7d4f49611 100644 >>>> --- a/hw/s390x/s390-pci-bus.c >>>> +++ b/hw/s390x/s390-pci-bus.c >>>> @@ -863,6 +863,11 @@ static void s390_pcihost_pre_plug(HotplugHandler >>>> *hotplug_dev, DeviceState *dev, >>>> { >>>> S390pciState *s = S390_PCI_HOST_BRIDGE(hotplug_dev); >>>> >>>> + if (!s390_has_feat(S390_FEAT_ZPCI)) { >>>> + warn_report("Adding PCI or zPCI devices without the 'zpci' CPU >>>> feature." >>>> + " The guest will not be able to see/use these >>>> devices."); >>>> + } >>>> + >>>> if (object_dynamic_cast(OBJECT(dev), TYPE_PCI_DEVICE)) { >>>> PCIDevice *pdev = PCI_DEVICE(dev); >>>> >>> >>> That's hotplug only, isn't it? IIRC coldplugging already fails? >>> >> >> No, applies also to coldplugging. > > Back then we made this a conscious decision, because removing the bridge > triggered a > lot of issues regarding migration. And the current behaviour actually is a > good > match to the real hardware, there are PCI devices in the system that can not > be used > by guests. I understand that this is kind of surprising, so I am fine with > the warn_report > but I do not want to have a hard error right now. >
So you agree to this patch, unmodified, correct? Thanks! -- Thanks, David / dhildenb
