You are not logged in.
Hi
I have a router running Arch Linux that acts as NAT for IPv4 and as a bridge for IPv6 because my ISP has native IPv6. I'm using systemd-networkd and I use the underlying interfaces that are added to the bridge for IPv4. The bridge is three interfaces including the wireless interface. Everything works except for IPv6 connectivity to the internet with the wireless clients. Hostapd adds the wireless interface to the IPv6 bridge (I'm not adding it with systemd-networkd). I am able to ssh from another IPv6 host somewhere else on the internet to the IPv6 address of one of the ethernet connected hosts behind my router. I have noticed that hostapd changes the sysctl settings for the wireless interface but it seems to me that it shouldn't matter when it's in the bridge and NATed IPv4 with the underlying interfaces works just fine. Is there something I'm missing or this a bug in hostapd?
I'm not looking for workarounds and not interested in changing my setup to something else.
# networkctl -a status
● 1: lo
Link File: /usr/lib/systemd/network/99-default.link
Network File: n/a
Type: loopback
State: carrier (unmanaged)
Address: 127.0.0.1
::1
● 3: enp0s10
Link File: /usr/lib/systemd/network/99-default.link
Network File: /etc/systemd/network/enp0s10.network
Type: ether
State: routable (configuring)
Path: pci-0000:00:0a.0
Driver: via-rhine
Vendor: VIA Technologies, Inc.
Model: VT6105M [Rhine-III]
HW Address: XX:XX:XX:XX:XX:XX
Address: 192.168.1.1
● 4: enp0s11
Link File: /usr/lib/systemd/network/99-default.link
Network File: /etc/systemd/network/enp0s11.network
Type: ether
State: routable (configuring)
Path: pci-0000:00:0b.0
Driver: via-rhine
Vendor: VIA Technologies, Inc.
Model: VT6105M [Rhine-III]
HW Address: XX:XX:XX:XX:XX:XX
Address: XXX.XXX.XXX.XXX
Gateway: XXX.XXX.XXX.254
DNS: XXX.XXX.XXX.XXX
XXX.XXX.XXX.XXX
● 5: wlp0s15f5u1
Link File: /usr/lib/systemd/network/99-default.link
Network File: /etc/systemd/network/wlp0s15f5u1.network
Type: wlan
State: carrier (configuring)
Path: pci-0000:00:0f.5-usb-0:1:1.0
Driver: rt73usb
Vendor: Ralink Technology, Corp.
Model: RT2501/RT2573 Wireless Adapter
HW Address: 00:1f:1f:08:f2:d2 (Edimax Technology Co. Ltd.)
Address: 192.168.2.1
● 6: br0
Link File: /usr/lib/systemd/network/99-default.link
Network File: /etc/systemd/network/ipv6.network
Type: ether
State: routable (configured)
Driver: bridge
HW Address: XX:XX:XX:XX:XX:XX
Address: XXXX:XXX:XXXX:XXXX:XXXX:XXX:XXXX:XXXX
fe80::1062:fff:fe71:2061
DNS: fe80::1062:fff:fe71:2061
XXXX:XXX::1
XXXX:XXX::2
# brctl show
bridge name bridge id STP enabled interfaces
br0 XXXX.XXXXXXXXXXXX no enp0s10
enp0s11
wlp0s15f5u1
Last edited by cjsthompson (2017-06-30 13:23:07)
Offline
I suppose your wireless clients are not getting global IPv6 addresses (like br0 does, "XXXX:XXX:XXXX:XXXX:XXXX:XXX:XXXX:XXXX") though?
My ISP also provides IPv6 capability and it is done in a way like this:
It provides me with an optical fibre modem with 4 LAN ports. Say I connect two devices to two of the ports, both of them get an external IPv4 address, but only one gets a global IPv6 address. However, if one of the device is a IPv6-capable router, it also gets another prefix, which is different from that of its IPv6 address and is used to derive global IPv6 addresses for its clients.
If your ISP implement IPv6 in a similar manner, I think your machine is now like the IPv6 router in the aforementioned scenario. I suppose you will need to set up something like radvd to make use of the prefix it gets (well, assuming it does).
Also, why are there addresses/routes/dns configured for enp0s10/enp0s11/wlp0s15f5u1 when they are bridge slaves?
Last edited by tom.ty89 (2017-07-02 02:58:57)
Offline