Linux IPv6 on Comcast
Linux and IPv6
IPv6 or Internet Protocol version 6 is the modern
network-layer protocol.
In many parts of the world including East Asia and
South Asia, IPv6 is the dominant protocol and IPv4 is used
only in small islands in an IPv6 sea.
About a year ago I had to replace my cable modem.
I checked first to verify that the one I planned to buy
at the local "big box" store supported IPv6.
However, I couldn't get IPv6 working, at least not at first.
Here is how to get IPv6 working with a
Motorola / Arris Surfboard SB6121
on a Comcast connection.
This involved some unexpected changes in the
ifcfg-* interface configuration file,
and modifications to one of the network control scripts.
Amazon
ASIN: B004XC6GJ0
Amazon
ASIN: B0100KRBA6
My main Linux system is connected to the cable modem, so its exterior interface gets a publicly reachable address. Network services are locked down with their configurations plus iptables (and now, ip6tables) rules.
I am doing this on Comcast, but I don't think there is
anything really Comcast-specific beyond the routing we
see in the traceroute6 output at the end.
I am running my own DNS server using BIND. I have noticed some erratic behavior from Comcast's DNS servers, but their routing is pretty reliable and so my DNS server can generally get me what I need.
I am using a Mageia distribution. Mandrake Linux was derived from Red Hat Linux version 5.1 back in 1998. There was a lawsuit from Hearst Corporation complaining about their trademarked character Mandrake the Magician in 2004, and then the acquisition of the Brazilian company Conectiva in early 2005, and so in April 2005 the Mandrake distribution became Mandriva Linux.
The last stable version was Mandriva Linux 2011, then the commercial Mandrake/Mandriva operation ended. Most of the laid-off developers joined the Mageia project.
Mageia remains somewhat Red-Hat-like, although not as similar as Oracle Linux or Scientific Linux, let alone CentOS.
Mageia uses RPM packages,
although it originally used the
URPMI
suite of commands —
urpmi,
urpme,
urpmf,
and
urpmq,
— rather than the
yum
high-level package management tool used on most
RPM-based distributions.
Now Mageia supports yum and
its replacement dnf.
The service control in general and the network
configuration details will be familiar to users of
most current Linux distributions.
Service control itself uses systemctl,
configuration is in /etc/sysconfig/,
network details and scripts are in
/etc/sysconfig/network-scripts/.
The Kernel Detects the Interfaces
The traditional network interface names eth0,
eth1, wlan0, and so on were replaced
starting with v197 systemd.
The names are now based on hardware details,
making them consistent from one boot session to the next.
Interface
Names
Interfaces on the motherboard may get names
eno1, eno2, and so on,
or maybe em1, em2, and so on.
Their names may encode their PCI bus addresses, such as
enp2s0 for an Ethernet device at
PCI bus address 02:00.0.
They may get something like p3p1
for a port on a PCI card, that would be for slot 3
and port 1.
My experience is that I cannot predict in advance what a given platform's interfaces will be named. But once we've seen it boot once, those names will always be the same.
The kernel detects the devices in whatever order its drivers
load and probe the buses, and assigns initial names
eth0, eth1, etc.
Then the interfaces are renamed.
The end result on my system is that p7p1
is the exterior Ethernet port, connected to the cable modem,
and p3p1 and em1 are on
interior LANs.
(em1 is on the motherboard,
p3p1 and p7p1
are on a dual-interface PCIe card)
A WLAN interface on a USB stick is wlp10s0u1,
encoding its USB device number plus the USB controller's PCI
bus address.
$ egrep 'p[0-9][ps][0-9]|e[mn][0-9]|eth[0-9]|wlan' /var/log/dmesg [ 4.791236] r8169 0000:06:00.0 eth0: RTL8168evl/8111evl at 0xffffc90000f96000, 00:13:3b:12:6f:a9, XID 0c900800 IRQ 44 [ 4.791238] r8169 0000:06:00.0 eth0: jumbo features [frames: 9200 bytes, tx checksumming: ko] [ 4.791727] r8169 0000:07:00.0 eth1: RTL8168evl/8111evl at 0xffffc90000f98000, 00:13:3b:12:6f:aa, XID 0c900800 IRQ 45 [ 4.791728] r8169 0000:07:00.0 eth1: jumbo features [frames: 9200 bytes, tx checksumming: ko] [ 4.792413] r8169 0000:0c:00.0 eth2: RTL8168f/8111f at 0xffffc90000f9c000, 08:62:66:2c:ab:1c, XID 08000800 IRQ 46 [ 4.792414] r8169 0000:0c:00.0 eth2: jumbo features [frames: 9200 bytes, tx checksumming: ko] [ 4.992382] r8169 0000:06:00.0 p3p1: renamed from eth0 [ 5.077048] r8169 0000:0c:00.0 em1: renamed from eth2 [ 5.085507] r8169 0000:07:00.0 p7p1: renamed from eth1 [ 6.268971] rt2800usb 2-1:1.0 wlp10s0u1: renamed from wlan0
Network Configuration and Scripts
General network settings go in
/etc/sysconfig/network —
hostname and whether to enable networking or not.
Individual network interfaces are configured in files
name ifcfg-* in the directory
/etc/sysconfig/network-scripts/.
In the following, p3p1 is on the interior LAN,
wlp10s0u1 is a USB-connected WLAN interface
using WPA/2 PSK with a key based on the passphrase
"PassPhraseHere",
and p7p1 is plugged into the cable modem
leading out to Comcast.
$ cat /etc/sysconfig/network HOSTNAME=thishost.example.com NETWORKING=yes $ cat /etc/sysconfig/network-scripts/ifcfg-p3p1 DEVICE=p3p1 ONBOOT=yes BOOTPROTO=static IPADDR=10.1.1.100 NETMASK=255.255.255.0 IPV6INIT=yes $ cat /etc/sysconfig/network-scripts/ifcfg-wlp10s0u1 DEVICE=wlp10s0u1 ONBOOT=yes BOOTPROTO=dhcp DHCP_CLIENT=dhclient METRIC=35 WIRELESS_MODE=Managed WIRELESS_ESSID=FBI_van4 WIRELESS_ENC_KEY="s:PassPhraseHere" WIRELESS_WPA_DRIVER=wext WIRELESS_WPA_REASSOCIATE=no IPV6INIT=yes $ cat /etc/sysconfig/network-scripts/ifcfg-p7p1 DEVICE=p7p1 ONBOOT=yes BOOTPROTO=dhcp DHCP_CLIENT=dhclient NEEDHOSTNAME=no DEFROUTE=yes IPV6INIT=yes
The interior interface p3p1 brought up IPv6
and gave itself a link-local auto-configured address
fe80::213:3bff:fe12:6fa9/64 (based on its Ethernet MAC
address 00:13:3b:12:6f:a9).
However, the exterior interface p7p1 did
not start IPv6, and attempts to manually assign an IPv6
address failed with a distinctive error:
# ip -6 addr add 2600:2600::1/64 dev p7p1 RTNETLINK answers: No buffer space available
That interface is ineligible for IPv6. The reasons are shown here, if you know what to look for:
$ cat /proc/net/if_inet6 00000000000000000000000000000001 01 80 10 80 lo fe80000000000000ca3a35fffecf3bb9 05 40 20 80 wlp10s0u1 fe8000000000000002133bfffe126fa9 02 40 20 80 p3p1 $ cat cat /sys/class/net/p7p1/mtu 576
The fields in if_inet6 are shown in
hexadecimal and list the:
• IPv6 address
• Netlink device number
• Prefix length
• Scope value
• Interface flags
Then the interface name.
But notice that p7p1 doesn't show up at all.
The reason is the MTU value.
IPv6 requires a minimum MTU of 1280 bytes
as per
RFC 2460.
(versus the IPv4 minimum of 576)
Enabling IPv6 on the Cable Modem Interface
Increase the IPv6-specific MTU setting.
You must use IPV6_MTU
and not MTU,
at least with the Mageia boot scripts.
$ cat /etc/sysconfig/network-scripts/ifcfg-p7p1 DEVICE=p7p1 ONBOOT=yes BOOTPROTO=dhcp DHCP_CLIENT=dhclient NEEDHOSTNAME=no DEFROUTE=yes IPV6INIT=yes IPV6_MTU=1500
We attempt to restart the network service. We don't get a global IPv6 address from the provider, but at least it brings up IPv6 and auto-configures a link-local address:
$ cat /proc/net/if_inet6 00000000000000000000000000000001 01 80 10 80 lo fe8000000000000002133bfffe126faa 03 40 20 80 p7p1 fe80000000000000ca3a35fffecf3bb9 05 40 20 80 wlp10s0u1 fe8000000000000002133bfffe126fa9 02 40 20 80 p3p1
Modifying the Boot Script
Now the service control scripts are trying to get an IPv6
address, but they fail.
You run:
systemctl, which calls
↪
/etc/sysconfig/network-scripts/ifup,
which calls
↪
/etc/sysconfig/network-scripts/ifup-ipv6 and
↪
/etc/sysconfig/network-scripts/ifup-eth
The last of those, ifup-eth, ends with
a block where it attempts to bring up IPv6 and then tests
the result.
The corresponding script in RHEL ends with an almost
identical block:
[ ... many lines not shown ... ]
# IPv6 initialisation?
/etc/sysconfig/network-scripts/ifup-ipv6 ${CONFIG}
if [[ "${DHCPV6C}" = [Yy1]* ]] && [ -x /sbin/dhclient ]; then
generate_config_file_name 6
generate_lease_file_name 6
echo
echo -n $"Determining IPv6 information for ${DEVICE}..."
if /sbin/dhclient -6 -1 ${DHCPV6C_OPTIONS} ${DHCLIENTCONF} -lf ${LEASEFILE} -pf /var/run/dhclient6-${DEVICE}.pid -H ${DHCP_HOSTNAME:-${HOSTNAME%%.*}} ${DEVICE} ; then
echo $" done."
else
echo $" failed."
if [ "${dhcpipv4}" = "good" -o -n "${IPADDR}" ]; then
net_log "Unable to obtain IPv6 DHCP address ${DEVICE}." warning
else
exit 1
fi
fi
fi
exec /etc/sysconfig/network-scripts/ifup-post ${CONFIG} ${2}
The ifup-eth script tries to pass a host name
with the -H flag, but the dhclient
program doesn't recognize that option.
It worked with dhclient version 4.1.1
but it no longer works with version 4.3.3,
see its documentation
for details.
Additionally, it seems to always fail when it's called too soon after the first instance ran to get the IPv4 configuration.
I need to modify the script.
Copy the original to ifup-eth-ORIGINAL
and make some changes.
The line marked *
was changed,
The lines marked +
were added, replacing what was there.
[ ... many lines not shown ... ] # IPv6 initialisation? /etc/sysconfig/network-scripts/ifup-ipv6 ${CONFIG} if [[ "${DHCPV6C}" = [Yy1]* ]] && [ -x /sbin/dhclient ]; then generate_config_file_name 6 generate_lease_file_name 6 echo echo -n $"Determining IPv6 information for ${DEVICE}..." attempts=1 * /sbin/dhclient -6 -1 ${DHCPV6C_OPTIONS} ${DHCLIENTCONF} -lf ${LEASEFILE} -pf /var/run/dhclient6-${DEVICE}.pid ${DEVICE} + success=$? + while [ $success != 0 ] + do + echo "" + attempts=$(( $attempts + 1 )) + sleep 1 + echo "attempt $attempts" + /sbin/dhclient -6 -1 ${DHCPV6C_OPTIONS} ${DHCLIENTCONF} -lf ${LEASEFILE} -pf /var/run/dhclient6-${DEVICE}.pid ${DEVICE} + success=$? + if [ $attempts -gt 10 ] + then + success=0 + fi + done fi exec /etc/sysconfig/network-scripts/ifup-post ${CONFIG} ${2}
It Works!
Let's see the configuration:
$ ip link show p7p1
3: p7p1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP mode DEFAULT group default qlen 1000
link/ether 00:13:3b:12:6f:aa brd ff:ff:ff:ff:ff:ff
$ ip addr show p7p1
3: p7p1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:13:3b:12:6f:aa brd ff:ff:ff:ff:ff:ff
inet 24.15.60.156/23 brd 255.255.255.255 scope global p7p1
valid_lft forever preferred_lft forever
inet6 2001:558:600d:16:12d:1ee2:8012:6869/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::213:3bff:fe12:6faa/64 scope link
valid_lft forever preferred_lft forever
$ cat /proc/net/if_inet6
00000000000000000000000000000001 01 80 10 80 lo
fe8000000000000002133bfffe126faa 03 40 20 80 p7p1
fe80000000000000ca3a35fffecf3bb9 05 40 20 80 wlp10s0u1
20010558600d0016012d1ee280126869 03 40 00 80 p7p1
fe8000000000000002133bfffe126fa9 02 40 20 80 p3p1
The fe80 address is auto-configured with
a link-local unicast address as before,
while the 2001:558::/31 IPv6 address block is assigned
to Comcast:
$ whois 2001:558:600d:16:12d:1ee2:8012:6869 [...] NetRange: 2001:558:: - 2001:559:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF CIDR: 2001:558::/31 NetName: COMCAST6NET NetHandle: NET6-2001-558-1 Parent: ARIN-001 (NET6-2001-400-0) NetType: Direct Allocation OriginAS: AS7922 Organization: Comcast Cable Communications, LLC (CCCS) [...]
Let's see the IPv4 and IPv6 routing tables:
$ ip route default via 24.15.60.1 dev p7p1 metric 5 10.1.1.0/24 dev p3p1 proto kernel scope link src 10.1.1.100 24.15.60.0/23 dev p7p1 proto kernel scope link src 24.15.60.156 metric 5 169.254.0.0/16 dev p7p1 scope link metric 5 169.254.0.0/16 dev wlp10s0u1 scope link metric 35 169.254.0.0/16 dev p3p1 scope link metric 1002 192.168.1.0/24 dev wlp10s0u1 proto kernel scope link src 192.168.1.101 metric 35 192.168.122.0/24 dev virbr0 proto kernel scope link src 192.168.122.1 linkdown $ ip -6 route unreachable ::/96 dev lo metric 1024 error -113 pref medium unreachable ::ffff:0.0.0.0/96 dev lo metric 1024 error -113 pref medium 2001:558:600d:16::/64 dev p7p1 proto kernel metric 256 pref medium unreachable 2002:a00::/24 dev lo metric 1024 error -113 pref medium unreachable 2002:7f00::/24 dev lo metric 1024 error -113 pref medium unreachable 2002:a9fe::/32 dev lo metric 1024 error -113 pref medium unreachable 2002:ac10::/28 dev lo metric 1024 error -113 pref medium unreachable 2002:c0a8::/32 dev lo metric 1024 error -113 pref medium unreachable 2002:e000::/19 dev lo metric 1024 error -113 pref medium unreachable 3ffe:ffff::/32 dev lo metric 1024 error -113 pref medium fe80::/64 dev wlp10s0u1 proto kernel metric 256 pref medium fe80::/64 dev p3p1 proto kernel metric 256 pref medium fe80::/64 dev p7p1 proto kernel metric 256 pref medium default via fe80::201:5cff:fe79:1446 dev p7p1 proto ra metric 1024 expires 1798sec pref medium
The 2001:558:600d:16::/64 subnet is directly connected
via p7p1.
The default route is the IPv6 next hop via Comcast's router.
The unreachable 2002::/16 routes are the
6to4 tunnel addresses
for localhost, the
RFC 1918
private IPv4 blocks (10/8, 172.16/12, and 192.168/16),
and the ZeroConf or Bonjour
169.254/16
block.
See RFC 5735
for the special-use IPv4 addresses.
The unreachable 3ffe::/16 address block was used for the
second 6bone experimental network.
Now let's see the ARP (IPv4) and NDP (IPv6) caches, with the Comcast router listed under its IPv4 and IPv6 addresses.
$ ip neigh show 24.15.60.1 dev p7p1 lladdr 00:01:5c:79:14:46 REACHABLE 10.1.1.234 dev p3p1 lladdr 00:30:c1:8b:d2:94 REACHABLE 10.1.1.231 dev p3p1 lladdr 00:1c:50:ac:72:1e DELAY 10.1.1.232 dev p3p1 lladdr b8:27:eb:69:be:bb REACHABLE 10.1.1.230 dev p3p1 lladdr 2c:27:d7:c5:d3:7b REACHABLE 192.168.1.100 dev wlp10s0u1 lladdr 00:37:6d:a9:b9:35 DELAY 10.1.1.252 dev p3p1 lladdr 00:1d:7e:2e:97:85 STALE 10.1.1.233 dev p3p1 lladdr b8:27:eb:01:d3:20 REACHABLE fe80::201:5cff:fe79:1446 dev p7p1 lladdr 00:01:5c:79:14:46 router REACHABLE
Finally, let's do some functional tests:
$ ping6 -n -c 5 www.google.com PING www.google.com(2607:f8b0:4009:80a::2004) 56 data bytes 64 bytes from 2607:f8b0:4009:80a::2004: icmp_seq=1 ttl=45 time=26.5 ms 64 bytes from 2607:f8b0:4009:80a::2004: icmp_seq=2 ttl=45 time=26.2 ms 64 bytes from 2607:f8b0:4009:80a::2004: icmp_seq=3 ttl=45 time=26.1 ms 64 bytes from 2607:f8b0:4009:80a::2004: icmp_seq=4 ttl=45 time=26.1 ms 64 bytes from 2607:f8b0:4009:80a::2004: icmp_seq=5 ttl=45 time=26.5 ms --- www.google.com ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4004ms rtt min/avg/max/mdev = 26.104/26.317/26.575/0.201 ms $ traceroute6 www.google.com traceroute to www.google.com (2607:f8b0:4009:80a::2004) from 2001:558:600d:16:12d:1ee2:8012:6869, 30 hops max, 24 byte packets 1 2001:558:600d:16::1 (2001:558:600d:16::1) 8.132 ms 8.575 ms 8.777 ms 2 te-5-4-ur01.lafayette.in.indiana.comcast.net (2001:558:322:ff89::1) 8.697 ms 8.278 ms 9.099 ms 3 te-9-2-ur02.lafayette.in.indiana.comcast.net (2001:558:300:40a::2) 8.603 ms 8.66 ms 10.193 ms 4 te-1-9-0-9-ar01.indianapolis.in.indiana.comcast.net (2001:558:300:40c::1) 18.756 ms 18.788 ms 15.965 ms 5 hu-1-13-0-0-ar01.area4.il.chicago.comcast.net (2001:558:300:25d::1) 27.876 ms 27.442 27.549 6 be-33491-cr02.350ecermak.il.ibone.comcast.net (2001:558:0:f6e9::1) 28.628 ms 27.001 ms 34.741 ms 7 hu-0-13-0-1-pe04.350ecermak.il.ibone.comcast.net (2001:558:0:f5df::2) 27.211 28.110 27.488 ms 8 as15169-7-c.350ecermak.il.ibone.comcast.net (2001:559::562) 28.239 ms 34.473 ms 26.295 ms 9 2001:4860::1:0:cb7d (2001:4860::1:0:cb7d) 27.782 ms 26.948 ms 36.19 ms 10 2001:4860:0:1::957 (2001:4860:0:1::957) 27.038 ms 26.581 ms 26.883 ms 11 ord30s21-in-x04.1e100.net (2607:f8b0:4009:80a::2004) 29.295 ms 27.844 ms 26.299 ms
That's 8-9 msec and 3 hops across West Lafayette and Lafayette, another 10 msec and one hop to Indianapolis, another 10 msec and 1 hop back past Lafayette to Chicago, 3 hops around the Chicago area, then into a Google data center.
Hops 9 and 10 are within the 2001:4860::/32 IPv6 address block assigned to Google but there aren't pointer records mapping those back to host names.
Cisco EZXS88W 8-port Ethernet switch, Cisco WRT54G wireless router and Ethernet switch, MFJ-1278 multi-mode data controller, and Harris SB6121 "SURFboard" cable modem.
Forwarding IPv6 to the World
The next step was to get IPv6 forwarding working
between my interior LAN and the outside world.
A shell script /etc/rc.d/rc.firewall
runs from /etc/rc.local at the end of
the booting process.
It uses iptables and ip6tables
to set up filtering rules.
[ ... many lines not shown ... ] echo " Finding interfaces..." EXT_INT="p7p1" EXT_IPV4="$( ip -4 addr show ${EXT_INT} | awk '/inet/ {print $2}' | sed 's@/.*@@' )" EXT_IPV6="$( ip -6 addr show ${EXT_INT} | awk '/inet6/ {print $2}' | sed 's@/.*@@' | grep -v '^fe80' )" [ ... many lines not shown ... ] echo " Turning on SNAT masqueraded forwarding..." # Turning on IPv6 forwarding will remove the IPv6 default route. # Record it first. DEFGW_IPV6=$(ip -6 route show | awk '/default/ {print $3}') # Now turn on forwarding and fix the IPv6 default route. echo 1 > /proc/sys/net/ipv4/ip_forward echo 1 > /proc/sys/net/ipv6/conf/all/forwarding if [ "$DEFGW_IPV6" != "" ] then ip route add default via $DEFGW_IPV6 dev p7p1 fi iptables -t nat -F ip6tables -t nat -F echo " Masq to ${EXT_INT}" iptables -t nat -A POSTROUTING -o ${EXT_INT} -j SNAT --to-source ${EXT_IPV4} ip6tables -t nat -A POSTROUTING -o ${EXT_INT} -j SNAT --to-source ${EXT_IPV6} [ ... many lines not shown ... ]
I need to run the
the IPv6 Router Advertisement Daemon.
It is configured in /etc/radvd.conf
and /etc/sysconfig/radvd
and controlled as service radvd.
My /etc/radvd.conf contains:
interface p3p1
{
AdvSendAdvert on;
# IPv6 routers will not forward (including with NAT) packets
# with link-local source addresses. So I advertise a
# fc00::/7 IPv6 block as per RFC 4193.
# https://tools.ietf.org/html/rfc4193
prefix fc00::/64
{
AdvOnLink on;
AdvAutonomous on;
AdvRouterAddr on;
};
};
I restart the Linux system and notice that a new IPv6 address is in use on the interior interface:
$ ip addr show p3p1
2: p3p1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:13:3b:12:6f:a9 brd ff:ff:ff:ff:ff:ff
inet 10.1.1.100/24 brd 10.1.1.255 scope global p3p1
valid_lft forever preferred_lft forever
inet6 fc00::213:3bff:fe12:6fa9/64 scope global mngtmpaddr dynamic
valid_lft 86157sec preferred_lft 14157sec
inet6 fe80::213:3bff:fe12:6fa9/64 scope link
valid_lft forever preferred_lft forever
When I restart an OpenBSD system on the interior LAN I now see and can do this:
$ cat /etc/hostname.re0 rtsol dhcp $ ifconfig re0 re0: flags=218843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,MPSAFE,AUTOCONF6> mtu 1500 lladdr 2c:27:d7:c5:d3:7b priority: 0 groups: egress media: Ethernet autoselect (100baseTX full-duplex) status: active inet 10.1.1.230 netmask 0xffffff00 broadcast 10.1.1.255 inet6 fe80::2e27:d7ff:fec5:d37b%re0 prefixlen 64 scopeid 0x1 inet6 fc00::2e27:d7ff:fec5:d37b prefixlen 64 autoconf pltime 14389 vltime 86389 inet6 fc00::38e3:cbcc:16ff:dd07 prefixlen 64 autoconf autoconfprivacy pltime 14269 vltime 86269 $ route show Routing tables Internet: Destination Gateway Flags Refs Use Mtu Prio Iface default 10.1.1.100 UGS 4 10 - 8 re0 10.1.1/24 10.1.1.230 UC 1 2 - 4 re0 10.1.1.100 00:13:3b:12:6f:a9 UHLc 2 132 - 4 re0 10.1.1.230 2c:27:d7:c5:d3:7b UHLl 0 41 - 1 re0 10.1.1.255 10.1.1.230 UHb 0 0 - 1 re0 loopback localhost UGRS 0 0 32768 8 lo0 localhost localhost UHl 2 3921 32768 1 lo0 BASE-ADDRESS.MCAST localhost URS 0 0 32768 8 lo0 Internet6: Destination Gateway Flags Refs Use Mtu Prio Iface ::/104 localhost UGRS 0 0 32768 8 lo0 ::/96 localhost UGRS 0 0 32768 8 lo0 default fe80::213:3bff:fe1 UG 0 3 - 56 re0 localhost localhost UHl 3 3 32768 1 lo0 ::127.0.0.0/104 localhost UGRS 0 0 32768 8 lo0 ::224.0.0.0/100 localhost UGRS 0 0 32768 8 lo0 ::255.0.0.0/104 localhost UGRS 0 0 32768 8 lo0 ::ffff:0.0.0.0/96 localhost UGRS 0 0 32768 8 lo0 2002::/24 localhost UGRS 0 0 32768 8 lo0 2002:7f00::/24 localhost UGRS 0 0 32768 8 lo0 2002:e000::/20 localhost UGRS 0 0 32768 8 lo0 2002:ff00::/24 localhost UGRS 0 0 32768 8 lo0 fc00::/64 fc00::2e27:d7ff:fe UC 0 0 - 4 re0 fc00::2e27:d7ff:fe 2c:27:d7:c5:d3:7b UHLl 0 0 - 1 re0 fc00::38e3:cbcc:16 2c:27:d7:c5:d3:7b UHLl 0 2 - 1 re0 fe80::/10 localhost UGRS 0 1 32768 8 lo0 fe80::%re0/64 fe80::2e27:d7ff:fe UC 1 5 - 4 re0 fe80::213:3bff:fe1 00:13:3b:12:6f:a9 UHLc 1 37 - 4 re0 fe80::2e27:d7ff:fe 2c:27:d7:c5:d3:7b UHLl 0 2 - 1 re0 fe80::1%lo0 fe80::1%lo0 UHl 0 0 32768 1 lo0 fec0::/10 localhost UGRS 0 0 32768 8 lo0 ff01::/16 localhost UGRS 0 1 32768 8 lo0 ff01::%re0/32 fe80::2e27:d7ff:fe UC 0 2 - 4 re0 ff01::%lo0/32 localhost UC 0 1 32768 4 lo0 ff02::/16 localhost UGRS 0 1 32768 8 lo0 ff02::%re0/32 fe80::2e27:d7ff:fe UC 0 2 - 4 re0 ff02::%lo0/32 localhost UC 0 1 32768 4 lo0 $ traceroute6 www.google.com traceroute6 to www.google.com (2607:f8b0:4002:c0c::93), 64 hops max, 60 byte packets 1 fc00::213:3bff:fe12:6fa9 (fc00::213:3bff:fe12:6fa9) 0.252 ms 0.189 ms 0.182 ms 2 2001:558:600d:16::1 (2001:558:600d:16::1) 16.989 ms 8.714 ms 8.407 ms 3 te-5-4-ur01.lafayette.in.indiana.comcast.net (2001:558:322:ff89::1) 40.658 ms 208.868 ms 207.967 ms 4 te-3-2-ur02.lafayette.in.indiana.comcast.net (2001:558:300:13f::2) 8.445 ms 8.243 ms 8.854 ms 5 te-0-7-0-4-ar01.indianapolis.in.indiana.comcast.net (2001:558:300:140::1) 17.06 ms 18.779 ms 16.809 ms 6 he-4-9-0-0-ar01.area4.il.chicago.comcast.net (2001:558:300:7::1) 27.611 ms 28.539 ms 27.51 ms [...]
And, from an interior Linux system:
$ ssh raspberrypi ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether b8:27:eb:69:be:bb brd ff:ff:ff:ff:ff:ff
inet 10.1.1.232/24 brd 10.1.1.255 scope global eth0
valid_lft forever preferred_lft forever
inet6 fc00::9fec:ef7f:c7d9:4ae3/64 scope global noprefixroute dynamic
valid_lft 86368sec preferred_lft 14368sec
inet6 fe80::ce3f:fcf6:59a6:7a08/64 scope link
valid_lft forever preferred_lft forever
$ ssh raspberrypi ip -6 route
fc00::/64 dev eth0 proto kernel metric 202
fe80::/64 dev eth0 proto kernel metric 256
default via fe80::213:3bff:fe12:6fa9 dev eth0 metric 202
$ ssh raspberrypi traceroute6 -m 6 www.google.com
traceroute to www.google.com (2607:f8b0:4002:c0c::93), 7 hops max, 80 byte packets
1 fc00::213:3bff:fe12:6fa9 (fc00::213:3bff:fe12:6fa9) 0.891 ms 0.309 ms 0.274 ms
2 2001:558:600d:16::1 (2001:558:600d:16::1) 8.422 ms 13.388 ms 14.183 ms
3 te-5-4-ur01.lafayette.in.indiana.comcast.net (2001:558:322:ff89::1) 14.858 ms 14.563 ms 14.266 ms
4 te-3-2-ur02.lafayette.in.indiana.comcast.net (2001:558:300:13f::2) 17.363 ms 18.034 ms 17.742 ms
5 te-0-7-0-4-ar01.indianapolis.in.indiana.comcast.net (2001:558:300:140::1) 19.396 ms 20.197 ms 19.876 ms
6 he-4-9-0-0-ar01.area4.il.chicago.comcast.net (2001:558:300:7::1) 28.450 ms 29.431 ms 33.517 ms