On Fri, Feb 13, 2015 at 01:51:54PM -0800, Ben Pfaff wrote: > On Fri, Jan 30, 2015 at 11:41:50AM +0900, Simon Horman wrote: > > NMX selection method > > Signed-off-by: Simon Horman <simon.hor...@netronome.com> > > > > --- > > * v2 > > Use array of TLVs rather than OF1.1 match for fields field of > > NMX selection method property > > I'm a little surprised that no selection method names are defined in > this or in the documentation. Is it expected that initially only the > empty string is used?
My thinking was to start with a framework no methods defined. And then to add methods with their implementation as use-cases arise. The last patch of this series implements a "hash" method although it is primarily intended as an example. If you think it is useful I would be happy to document some strings and what their behaviour might be. This would act as a reservation of sorts, defining what a well defined implementation of a given string should be. I would also be happy to provide implementations if that is useful. Though naturally I'd prefer some bound on the number of algorithms to be implemented up front. All that said, I do think it is reasonable to have a framework and add users on a case-by-case basis. > I'm happy enough with the structure otherwise. It might be > appropriate to squash this with other patches (I haven't read them yet). _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
