Summary
Add VyOS CLI support for VXLAN-GBP.

The intended packet flow is:

VyOS A: set packet mark in kernell
VyOS A: Encode the mark in the VXLAN-GBP header
VyOS B: Restore the packet mark in the Linux kernel from the VXLAN-GBP header

Use case
Carry group-policy context between VTEPs and apply receiving firewall policy to that context.

In ordinary non-external operation, Linux maps the packet mark to and from GBP.

The proposed change exposes the existing kernel facility through VyOS configuration.
GBP is documented in draft-smith-vxlan-group-policy, an expired individual Internet-Draft. It is implemented in iproute2 and supported in e.g. openVswitch.

Additional information

Why most probably not want to use GBP - It can be manipulated in transit, so you need to be very careful if you trust your underlay network or not or secure authenticity another way.

In our lab we mark a packet at ingress

set firewall ipv4 prerouting raw rule 100 set mark '20000'

Proposed config ( same as GBE), to pass the mark to be set in the vxlan extension header. We have tested it in our lab.

set interfaces vxlan vxlan100 gbp

Let me know what you think. We have it working in our lab, I can create PRs.