Hi guys, Is there a way to classify the input packets based on the QinQ field (i.e. SVLAN and CVLAN labels)?
It looks like by default all QinQ packets are dropped, so to avoid this a sub-interface has to be created for every QinQ label. The problem with this approach is that SVLAN and CVLAN labels are removed from the packet, therefore these fields can no longer be used by any classifier table. create sub TenGigabitEthernet86/0/1 svlan1cvlan100 dot1q 1 inner-dot1q 100 Also, we need to be able to create thousands of QinQ flows, so probably the sub-interface approach is not the best. Can the sub-interface approach be made to work for QinQ-based classifier? Is there a better way to handle QinQ flows that the sub-interface approach? Regards, Cristian
_______________________________________________ vpp-dev mailing list vpp-dev@lists.fd.io https://lists.fd.io/mailman/listinfo/vpp-dev