[ovs-dev] [PATCH 2/5] datapath: Add 'patch' vport.

Ben Pfaff blp at nicira.com
Tue May 18 20:36:42 UTC 2010


On Tue, May 18, 2010 at 01:35:05PM -0700, Jesse Gross wrote:
> On Tue, May 18, 2010 at 9:58 AM, Ben Pfaff <blp at nicira.com> wrote:
> 
> > On Mon, May 17, 2010 at 07:24:45PM -0700, Jesse Gross wrote:
> > > Adds a new 'patch' vport that acts like a virtual patch cable to
> > > connect together two datapaths.
> >
> > Looks good to me.
> >
> > Is there any value in a "patch" vport having a MAC address?  Since it's
> > not a network device there's no way to configure an IP address on it, so
> > can the MAC ever get used for anything?  (Actually I guess the same
> > question applies to GRE vports.  I see that Ian has the same question.)
> 
> The situation is a bit different from GRE since GRE uses the MAC for path
> MTU discovery.  The only thing that 'patch' uses the MAC address for is to
> make it look more like a real Ethernet device.  The MAC is reported over
> OpenFlow and it is conceivable that a controller could use it as a way to
> identify a port.  This approach will break on Xen but generally speaking, it
> makes me a bit nervous to have non-unique MAC addresses.

OK, no problem.




More information about the dev mailing list