On Thu, Nov 6, 2014 at 2:49 PM, Ben Pfaff <b...@nicira.com> wrote: > On Thu, Nov 06, 2014 at 02:37:01PM -0800, Jesse Gross wrote: >> I'm curious how you ended up laying this out. The OpenFlow spec says >> that the extra space should be used as an vendor ID in the form of an >> OUI. How did you reconcile this? > > You can probably get away with a lot if you make the most significant > byte of the extra 32 bits nonzero, because OpenFlow only specifies the > format of the other 24 bits if that byte is zero and leaves the rest of > the space reserved for future standardization. > > But that's kind of a nasty thing to do.
This crossed my mind as well. > Have you considered allowing database configuration to specify what kind > of attributes are expected and how to parse them? The configuration > could be P4-based, for example. This is possible but it seems kind of ugly and complicated so I hope to not have to go down this path. I think there is a simpler variant, which is to just have a mapping table of Geneve IDs to a pool of OXMs. That's all that is really needed since the action to be taken is just matching against the payload so there isn't much parsing. But I still would like it avoid this if possible. _______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss
