[libvirt] Network device abstraction aka virtual switch - V3

[Daniel P. Berrange]

On Thu, Jun 16, 2011 at 04:40:19PM +0200, Gerhard Stenzel wrote:
> On Sun, 2011-06-12 at 20:29 -0400, Laine Stump wrote:
> > <!-- A macvtap passthrough connection (one guest interface per dev)
> > -->
> > <network>
> > <name>red-network</name>
> > <forward layer='link' mode='passthrough' dev='eth10'/>
> > <interface dev='eth10'/>
> > <interface dev='eth11'/>
> > <interface dev='eth12'/>
> > <interface dev='eth13'/>
> > <interface dev='eth14'/>
> > <interface dev='eth15'/>
> > <interface dev='eth16'/>
> > <interface dev='eth17'/>
> > </forward>
> > </network> 
> 
> If this example describes a scenario with a SR-IOV card, where eth10 is
> the physical function and eth11-eth17 are the virtual functions and
> libvirt can attach a VM to any of the VFs, then I would not list eth10
> in the interface pool for passthrough devices.

All interfaces listed here should be considered equal for attaching VMs
to. I don't think the network code has to even care about whether a NIC
in the XML is a virtual or a physical function. The application will
discover NICs and whether they are virtual/physical functions via the
node device APIs in libvirt. It will then decide which of the NICs to
use when creating the network XML.

Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|