On Mon, Sep 10, 2012 at 09:59:28AM +0900, Simon Horman wrote: > On Sat, Sep 08, 2012 at 12:15:53PM +0900, Simon Horman wrote: > > On Fri, Sep 07, 2012 at 10:33:00AM -0700, Ben Pfaff wrote: > > > On Fri, Aug 31, 2012 at 05:42:48PM +0900, Simon Horman wrote: > > > > On Thu, Aug 30, 2012 at 10:02:51AM +0900, Simon Horman wrote: > > > > > On Wed, Aug 29, 2012 at 10:07:55AM -0700, Ben Pfaff wrote: > > > > > > On Tue, Aug 21, 2012 at 01:55:36PM +0900, Simon Horman wrote: > > > > > > > Signed-off-by: Simon Horman <ho...@verge.net.au> > > > > > > > > > > > > I'm not really happy with this approach. It seems fairly awkward. > > > > > > I > > > > > > would much rather define some "struct ofputil_*" that has a > > > > > > superset of > > > > > > the 1.[012] table stats and define converter functions. > > > > > > > > > > Sure, I'll see what I can do. > > > > > I do agree that the approach I have taken is somewhat awkward. > > > > > > > > is this closer to what you would like? > > > > > > Much closer. I think that if you move the conversion code from > > > ofproto.c to single-purpose functions in lib/ofp-util.c then it's what > > > I had in mind. > > > > Sure, I'll make that change and see how it looks. > > > > > (I did mean to actually invent a new struct for that, > > > but it seems like the OF1.2 version is good enough for now.) > > > > (me too) > > Hi Ben, > > The patch below attempts to implement your suggestion. > > I have created ofputil function to both convert to Open Flow version > specific table stats structures, and create Open Flow version specific > table stats reply messages. This much seems clean. Going beyond that > seemed to get a bit messy.
Thanks. It's moving in the right direction. I'm working on massaging it a little more toward further abstraction. (If you have time for anything else in the series, please don't let that hold you up.) Thanks, Ben. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
