On Mon, Oct 29, 2012 at 09:58:58AM -0700, Ben Pfaff wrote: > On Sat, Oct 27, 2012 at 03:05:57PM +0900, Simon Horman wrote: > > There are (or at least will be) cases where this check can produce false > > positives. For example, a flow which matches a non-MPLS packet and then > > applies an MPLS push action followed by an action to load the MPLS label. > > So, I remember the discussion, but I don't remember coming to exactly > this conclusion. > > The problem seems to be, at this point, MPLS specific, because for > most prerequisites, whether it is satisfied cannot be affected by > actions. For example, no action can transform a TCP packet into a > non-TCP packet, and vice versa. > > My thought is, then, for now we should define a new function to test > whether a prerequisite could ever be satisfied given a particular > flow. For all prerequisites other than MPLS, this would return the > same thing as mf_are_prereqs_ok(); for MPLS, it could just return > true.
I think that I am comfortable with that approach, I will implement it and see how things go. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
