Comment 6 for bug 1940635

Can somebody confirm my assumptions on the architecture of the test ?

First, it sets up a veth with router_eth42 at one end and test_eth42 at the other (the itf names might be different ?). dnsmasq is bound against router_eth42 which is configured manually via iproute with both IPv4 and IPv6 addresses.

networkd is in charge of configuring test_eth42. In this specific case, it is expected that it sends a DHCPv6 request via test_eth42, and that we get an answer from dnsmasq with an lease to an address between 2600::10 and 2600::20.

When running tcpdump on test_eth42, I can only see some ICMPv6 router solicitations (from both ends of the tunnel), with no router advertisement as answer. This bit is expected AFAIK. However, I do not see any other traffic, crucially, no DHCPv6 traffic at all.