FYI. Maybe my original guess is right...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
PCI controllers have an optional model attribute with possible values pci-root, pcie-root, pci-bridge, or dmi-to-pci-bridge. The root controllers (pci-root and pcie-root) have an optional pcihole64 element specifying how big (in kilobytes, or in the unit specified by pcihole64's unit attribute) the 64-bit PCI hole should be. Some guests (like Windows XP or Windows Server 2003) might crash when QEMU and Seabios are recent enough to support 64-bit PCI holes, unless this is disabled (set to 0). Since 1.1.2 (QEMU only)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The description above is from http://libvirt.org
/formatdomain.html#elementsControllers.

Thanks,
  Xu Wang
于 2014年04月22日 11:22, Xu Wang 写道:
I am sorry it may not be caused by the version of qemu-kvm. But I still have not found the real reason. If you know it please let me know.

Thanks,
  Xu Wang
于 2014年04月22日 10:59, Xu Wang 写道:
Dear John,
I just found an issue on RHEL-6.5. The reproduce steps,
1. Install a pure RHEL-6.5 system and just use rhn updates from RedHat.
2. Install and config libvirt-cim/cimtest just like before.
3. Run cimtest, you will find lots of testcases failed like this,
InvodeMethod(DefineSystem): CIM_ERR_FAILED: Failed to define domain:
internal error Unknown controller type 'pci' with return code 1

The root cause of error is, default version of qemu-kvm from RHEL-6.5 is 0.12.1.2-2,
too old for <controller type='pci' ...> (got that conclusion from link http://libvirt.org
/formatdomain.html#elementsControllers). In my opinion, we should take it into consideration,
or things like that will happen to the users who installed system like that because not everyone
will update qemu to the newer version. My suggestion is, shall we adjust cimtest a little?
Add a version checking into cimtest or just use another parameter replace it (type='pci')?

Thanks,
Xu Wang
于 2014年04月22日 00:21, John Ferlan 写道:

Do you feel outside of patch 1/10 that this series can be pushed once
the controller series is pushed?  With of course any "adjustments" to
the numbers based on the libvirt-cim commit numbers.

I can hold off on 1/10 and rework it later as there's just other things
going on and I don't have the same issue since I don't use aliases on my
localhost.

Thanks,

John


_______________________________________________
Libvirt-cim mailing list
Libvirt-cim@redhat.com
https://www.redhat.com/mailman/listinfo/libvirt-cim

_______________________________________________
Libvirt-cim mailing list
Libvirt-cim@redhat.com
https://www.redhat.com/mailman/listinfo/libvirt-cim