IPv6 fix for NDP and DHCPv6

Some considerations about them:

NDP specific rules are needed because it is a Layer 3 protocol. Rules allow:
- exchanging neighbor-solicitation/neighbor-advertisement between pods and/or hosts
- receiving router-advertisement and redirect from routers
- sending router-solicitation to routers
Other ICMP scenarios, (e.g. Path MTU Discovery), involve traffic that is not mentioned by either MultinetworkPolicy or regular NetworkPolicy. It is usually allowed when there is no protocol specified in policies, but the user can't specify ICMP in protocol fields. Therefore, if we put some built-in logic around icmp (i.e allow all icmpv6 traffic), the user will not able to block it.
- IMO we should allow the traffic that is strictly needed for applications to work
- Note: ICMP scenarios like MTU discovery are affected by iptables also in IPv4. If we decide to put some built-in logic on them, we should apply it to both IPv6 and IPv4 families.
Rules for DHCPv6 protocol allow:
- Incoming UDP traffic to client port (546) with link-local address as the destination.
- Outgoing UDP traffic to server port (547) with All_DHCP_Relay_Agents_and_Servers (ff02::1:2)
- Note: outgoing traffic is needed because when receiving a RECONFIGURE packet, the client should initiate a Renew/Reply, Rebind/Reply handshake.
- Note 2: I was not able to test these rules against a DHCPv6 server, as dhcp ipam plugin currently does not support DHCPv6 protocol. Is there any other way to try it? Otherwise, we can move these changes to another PR.

Waiting for a feedback

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pkg/server/server.go	0	16	0.0%
<!--	Total:	22	38	57.89%	-->

k8snetworkplumbingwg / multi-networkpolicy-iptables