fix: do not advertise Pod IPv4 CIDR in a Cluster composed of IPv6-only Nodes

[abbshr]

in case of pods subnet configured with both IPv4 and IPv6 CIDR, but nodes only have IPv6 addresses.

kube-router will report error and won't exit:

2025-11-13T12:27:06.433971867+08:00 stderr F E1113 04:27:06.433945  789979 network_routes_controller.go:393] Error advertising route: invalid nexthop address: <nil>

even though providing the key command line options as follow:

              "--run-router=true",
              "--run-firewall=false",
              "--run-service-proxy=true",
              "--enable-ipv4=true",
              "--enable-ipv6=true",
              "--router-id=generate",
              "--enable-overlay=true",

as a result, the BGP rib will be empty, and the results are quite confusing.

so i modify the code, when nodes in a IPv6-only environment, kube-router won't advertise the IPv4 Pod CIDR.

[aauren]

Thanks for this submission @abbshr!

I'm sorry to say that this was a slightly obvious oversight. I guess even with IPv6 support added I was still treating IPv4 addresses as first-class citizens.