[virt-tools-list] [PATCH V2] addhardware: Deal with the conflict host device

[Lin Ma]

>>> Cole Robinson <crobinso at redhat.com> 08/16/14 1:25 AM >>>
......
> I'd like to see this arranged in a bit different way. We already have similar
> code to match a hostdev to its nodedev in details.py:lookup_nodedev. Please
> move that code to a function like nodedev.py NodeDevice.compare_to_nodedev,
> then you can even separate the logic and implement the function in PCIDevice,
> USBDevice, etc. Bonus points for unit tests in tests/nodedev.py. That'll be
> one patch

I'll use NodeDevice.compare_to_nodedev instead of VirtualHostDevice.is_conflict_hostdev function.
In compare_to_nodedev function, I'll iterate all of hostdevs of all of vms, call lookup_nodedev for every 
hostdev to get the matched nodedev, Compare this nodedev with the passed nodedev which representing
the pci/usb device that users picked up while adding hardware.(put the comparson logic in PCIDevice 
and USBDevice)

 
> You'll still need details.py:lookup_nodedev to iterate over all the
> nodedevices and compare against the passed hostdev, but the comparison logic
> will be removed.

Do you mean I should use the code of lookup_nodedev function in compare_to_nodedev function?
If so, why ? If you want to make the comparson happens between nodedev and nodedev, Just calling
lookup_nodedev function to get the matched nodedev and comparing this nodedev with passed nodedev, just like what 
I mentioned above.

If you want to make the comparson happens between nodedev and hostdev, Just comparing the nodedev attributes with
the hostdev attributes, In this situation, We don't need the code of lookup_nodedev or calling lookup_nodedev, 
We don't need to iterate over all the nodedevices, We need to iterate all the hostdevices in all the vms, and compare against the 
given/passed nodedev.

What do you think?

Thanks,
Lin


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/virt-tools-list/attachments/20140819/017e2b0a/attachment.htm>