On 08/11/2016 01:59 PM, Ján Tomko wrote:
On Wed, Aug 10, 2016 at 04:27:58PM +0200, Boris Fiuczynski wrote:
> On 08/08/2016 06:35 PM, Ján Tomko wrote:
>>
https://bugzilla.redhat.com/show_bug.cgi?id=1227354
>>
>> Translate the optional <virtio revision> attribute to
>> disable-legacy=on/off and disable-modern=on/off options
>> for the following devices:
>>
>> <memballoon> virtio-balloon-pci
>> <disk> virtio-blk-pci
>> <controller> virtio-scsi-pci
>> virtio-serial-pci
>> <filesystem> virtio-9p-pci
>> <interface> virtio-net-pci
>> <rng> virtio-rng-pci
>> <video> virtio-gpu-pci
>> <input> virtio-input-host-pci
>> virtio-keyboard-pci
>> virtio-mouse-pci
>> virtio-tablet-pci
>> ---
>> src/qemu/qemu_command.c | 56
>> +++++++++++++++++++
>> .../qemuxml2argv-virtio-revision.args | 62
>> ++++++++++++++++++++++
>> tests/qemuxml2argvtest.c | 11 ++++
>> 3 files changed, 129 insertions(+)
>
> Although you mention in all above devices the type of bus in the code
> that differentiation does not really exist. It originates from the way
> the capability QEMU_CAPS_DEVICE_VIRTIO_REVISION is sensed and set, e.g.
> virQEMUCapsObjectPropsVirtioBalloon[] is also used for
> virtio-balloon-ccw devices.
> I would suggest to create a QEMU_CAPS_DEVICE_VIRTIO_PCI_REVISION and
> sense the capability for pci only and also generate the command line
> parameters for pci only since the virtio pci revision handling does not
> work at all for virtio ccw.
I have sent a new version that only probes the capability on PCI
devices:
https://www.redhat.com/archives/libvir-list/2016-August/msg00592.html
Thanks for catching that.
Jan
Thanks, and another thanks for adding pci-only in the command line
generation with this patch:
https://www.redhat.com/archives/libvir-list/2016-August/msg00593.html
--
Mit freundlichen Grüßen/Kind regards
Boris Fiuczynski
IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martina Köderitz
Geschäftsführung: Dirk Wittkopp
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294