[libvirt] [PATCH 03/15] RNG updates, new xml parser/formatter code to support forward mode=hostdev

[Laine Stump]

On 08/15/2012 07:25 AM, Shradha Shah wrote:
> On 08/14/2012 06:27 AM, Laine Stump wrote:
>> On 08/10/2012 12:23 PM, Shradha Shah wrote:
>>> diff --git a/tests/networkxml2xmlin/hostdev-pf.xml b/tests/networkxml2xmlin/hostdev-pf.xml
>>> new file mode 100644
>>> index 0000000..e07db69
>>> --- /dev/null
>>> +++ b/tests/networkxml2xmlin/hostdev-pf.xml
>>> @@ -0,0 +1,11 @@
>>> +<network>
>>> +  <name>hostdev</name>
>>> +  <uuid>81ff0d90-c91e-6742-64da-4a736edb9a9b</uuid>
>>> +  <forward mode="hostdev" managed="yes">
>>> +    <pf dev='eth2'/>
>>> +    <address type='pci' domain='0' bus='3' slot='0' function='1'/>
>> Is it legal to have a pf using a netdev name along with a list of pci
>> devices? I thought it was either/or?
> This is a test to test the inactive flag that can be used with net-dumpxml.
> An active network will have a pf along with the interface pool but when the 
> inactive flag is used only the inactive part of the network xml should be 
> formatted. Hence the address part of the xml is thrown out.

But that means if somebody mistakenly believes that they can specify
(for example), one entire PF, plus several VFs of another, that config
will be accepted and the VFs from the second PF will be silently
discarded rather than logging an error. I think you may need to log an
error rather than ignoring it.