On Wed, Jul 25, 2012 at 09:01:10AM +0900, Simon Horman wrote:
> On Tue, Jul 24, 2012 at 01:31:41PM -0700, Ben Pfaff wrote:
> > On Mon, Jul 23, 2012 at 03:16:35PM +0900, Simon Horman wrote:
> > > Signed-off-by: Simon Horman <ho...@verge.net.au>
> > > struct ofpbuf *
> > > -ofputil_encode_barrier_request(void)
> > > +ofputil_encode_barrier_request(uint8_t ofp_version)
> > > {
> > > - return ofpraw_alloc(OFPRAW_OFPT10_BARRIER_REQUEST, OFP10_VERSION, 0);
> > > + enum ofpraw type;
> > > +
> > > + switch (ofp_version) {
> > > + case OFP12_VERSION:
> > > + case OFP11_VERSION:
> > > + type = OFPRAW_OFPT11_BARRIER_REQUEST;
> > > + break;
> > > +
> > > + case OFP10_VERSION:
> > > + type = OFPRAW_OFPT10_BARRIER_REQUEST;
> > > + break;
> > > +
> > > + default:
> > > + NOT_REACHED();
> > > + }
> > > +
> > > + return ofpraw_alloc(type, ofp_version, 0);
> > > }
> >
> > This will need to change when we add new OpenFlow versions, but I
> > don't expect that new OpenFlow versions will actually change anything
> > in the barrier request. So I'd be inclined to do something like:
> > raw = (version == OFP10_VERSION
> > ? OFPRAW_OFPT10_BARRIER_REQUEST
> > : OFPRAW_OFPT11_BARRIER_REQUEST);
> > instead of a switch that definitely needs to be updated.
>
> I do have a slight preference for the case style, but I'll change
> update your patch as per your suggestion.
So, here's another option that I've been thinking about anyway:
instead of using macros for OFP10_VERSION etc., define an "enum
ofp_version" with those as enumeration constants. Then as long as we
passed in the ofp_version as an enum ofp_version, we'd automatically
get compiler warnings whenever we add a new version.
If you are willing to write up that change then I'd be happy to use a
switch statement here.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
