[FEAT] IPv6 subnet support

[simonmysun]

Is this a new feature request?

• I have searched the existing issues

Wanted change

I look forward to supports for IPv6 subnet. Currently /root/etc/s6-overlay/s6-rc.d/init-wireguard-confs/run is based on the assumption that users only uses IPv4 for wireguard traffics. However, both docker and wireguard are IPv6 ready. I have manually adjusted the config files to allocate IPv6 addresses to the wg peers. It will be wiped away when there is any change in docker environment variables and I will have to manually revert the changes. I would appreciate if IPv6 configuration was supported and I would be happy to adjust my set up to follow the change.

Reason for change

Docker and wireguard are IPv6 ready. IPv4 will be phased out in the future. This is a change that has to made sooner or later.

Proposed code change

No response

[github-actions]

Thanks for opening your first issue here! Be sure to follow the relevant issue templates, or risk having this issue marked as invalid.

[LinuxServer-CI]

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

[simonmysun]

Please don't. Automatic reply does not require feedback from OP.

[drizuid]

i don't follow the expectation. You can see the default route supports ipv6 and ipv4, we dont want to use "private ip space" for ipv6 and we don't want to nat, so we can't provide a dump subnet. You should be assigning a properly routed subnet from your /56, /60, or /64. For example, I carve out a /112 for my docker network and allocate an ip from that to various containers, including wireguard. I then configure my conf to use that block.

This does assume you've properly configured your custom bridge to support ipv6 and that you've properly created routes for the subnet.

Can you be a bit clearer on what you are proposing we do here?

[interna1error]

+1 on this one. The author is probably proposing support for an environment variable similar to INTERNAL_SUBNET for IPv4, so each client and the server itself get an automatically assigned IPv6 address from the specified subnet for inter-device communication within the wireguard network. Maybe rename INTERNAL_SUBNET to INTERNAL_IPV4_SUBNET and add a new INTERNAL_IPV6_SUBNET variable?

[simonmysun]

Yes. INTERNAL_SUBNET is assumed to be an IPv4 Address:

docker-wireguard/root/etc/s6-overlay/s6-rc.d/init-wireguard-confs/run

Line 160 in 3bebdf6

INTERFACE=$(echo "$INTERNAL_SUBNET" | awk 'BEGIN{FS=OFS="."} NF--')

Additionally when I manually allocated IPv6 addresses to the peers, they will be rewritten in generate_confs. It's difficult for me to update the peer confs because I have to either apply changes manually, or let generate_confs overwrite and then revert the changes on peer addreses.

[drizuid]

let us discuss internally, thank you for the clarification

[drizuid]

I'm a bit delayed getting back to reply here, but we are not opposed to adding support for this, however, i am the only ipv6 user on the team and do not have time to put into this, so the PR will need to come from the users wanting this support. The PR would also need to be thoroughly tested before it could be merged.