wg-ip is a tool to assign ip addresses derived from the public keys to wireguard interfaces.
I personally just run wg-ip apply dev wg0 after adding the peers and
bringing the interface up. In my automation scripts, I obtain the private
ip address of a given node by using wg-ip gen <node-pubkey>.
wg-ip is written to be portable across shells. It should run fine on openwrt (if not, it's a bug).
Usage: wg-ip [-4|-6|--subnet <subnet>][--asnet <length>] [gen <pubkey>|[dev <device> ]apply|[dev <device> ]dryrun]
Subnet used:
-4
sets the subnet to 10.0.0.0/8
-6
sets the subnet to fd1a:6126:2887::/48
--subnet <subnet>
sets an arbitrary (ipv4 or ipv6) subnet
When the subnet is specified several times, only the last
occurrence applies. By default -6 is enabled.
Device used:
dev <device>:
specify the device to which the commands apply.
When the device option is used several times, only the last
occurrence applies. If no device is specified, commands are
applied to the first entry returned by 'wg show interfaces'
Commands:
--help
show this message
--test
run the embedded test suite
gen <pubkey>:
generate an ip from the pubkey, in the specified subnet.
when the --asnet option is used, the generated ip is cut to
length to generate a subnet prefix instead of a whole ip.
dryrun:
generate an ip from the pubkey of the specified interface (see
"device used"), as well as for all the peers of that same
interface, in the specified subnet (see "subnet used"). Create
commands to add this ip to the interface, as well as commands
to modify allowed_ips for each peer. The commands are simply
displayed, not applied.
apply:
actually run the commands displayed by 'dryrun'
When several commands are specified, only the last one is
executed. Executing '' without argument is the same as running
' dryrun'
Example:
dev wg0 apply
See https://lists.zx2c4.com/pipermail/wireguard/2018-April/002593.html