The push_nsh action creates a non-Ethernet packet in the OF pipeline, which is not at all supported by OVS prior to Simon's patch. It is not good enough to promise that an SFC controller will always send a push_eth action right next to fix this.
A new action in OVS must be sane in the sense that when applied to a packet that fulfills the action's pre-requisites the outcome is a valid packet that can be processed further. Simon's patch provides a base for further processing of non-Ethernet packets, such as sending to a "L3" (or rather non-Ethernet) tunnel like VXLAN-GPE or GRE, or pushing an Ethernet header with push_eth. BR, Jan > From: Jesse Gross [mailto:je...@kernel.org] > Sent: Wednesday, 13 July, 2016 19:53 > > Please see previous comments in this thread, such as this one: > http://openvswitch.org/pipermail/dev/2016-July/074980.html > > On Wed, Jul 13, 2016 at 10:06 AM, Brady Allen Johnson > <brady.allen.john...@ericsson.com> wrote: > > Is the current implementation really dependent on Simon's patch? > > > > I understood that the current implementation is for ethernet+NSH and > > VXLAN+ethernet+NSH which doesnt require Simon's patch. Simon's patch > > VXLAN+ethernet+would > > be needed for VXLAN-GPE+NSH, which is not in this implementation. > > Maybe the authors can verify this. > > _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
