On 03/27/15 08:49, Thomas Graf wrote:

So at that point you get fairly close to cls_flow in terms
of functionality. Not sure why you want to keep them separate.

IMO, it is different enough to be separate.
Flow could do more of course (as could u32) but the lookup
approaches are different. And yes, ONF will probably make a press
release if you called it the openflow classifier (fingers crossed).

cheers,
jamal