Intermediate · IPv4 over an IPv6-only core (RFC 8950) · section 5 of 8

Turning on RFC 8950

✓ Lab-tested 17 October 2026 on VyOS 2026.10.16-0024-rolling, FRR 10.2.1

VyOS needs one line per IPv6 neighbor that carries IPv4: capability extended-nexthop. Both ends must have it; the capability is only used if both sides advertise it. Add it on R1 for C1 and R2, and on R2 for R1.

After the sessions re-establish:

show bgp neighbors shows Extended nexthop: advertised and received.
Both IPv4 routes are valid and best, with IPv6 link-local next hops. The detail view shows the peer's global IPv6 address and, in brackets, the link-local address actually used for forwarding.
R1's IPv4 routing table holds two BGP routes and nothing else. There's no connected IPv4 route, because R1 has no IPv4 subnet.
In the Linux kernel underneath, the routes are via inet6 fe80::…: an IPv4 route with an IPv6 gateway. The kernel has supported this (the RTA_VIA attribute) since Linux 5.2.

R1 also passes both routes on (PfxSnt 2), so each customer now learns the other's prefix.

R1: extended next hop on both IPv6 sessions

set protocols bgp neighbor 2001:db8:0:1::2 capability extended-nexthop
set protocols bgp neighbor 2001:db8:0:2::2 capability extended-nexthop

Lines 1–2: One line per IPv6 neighbor that carries IPv4.

R2: and on its side of the R1 session

set protocols bgp neighbor 2001:db8:0:2::1 capability extended-nexthop

Line 1: Both ends must advertise it.

IPv4 routes with IPv6 next hops (real output)

vyos@R1:~$ show ip bgp summary

IPv4 Unicast Summary:
BGP router identifier 0.0.0.1, local AS number 64500 VRF default vrf-id 0
BGP table version 2
RIB entries 2, using 320 bytes of memory
Peers 2, using 60 KiB of memory

Neighbor        V         AS   MsgRcvd   MsgSent   TblVer  InQ OutQ  Up/Down State/PfxRcd   PfxSnt Desc
2001:db8:0:1::2 4      64512         9         9        2    0    0 00:00:17            1        2 C1
2001:db8:0:2::2 4      64501        15        13        2    0    0 00:00:03            1        2 R2

Total number of neighbors 2
vyos@R1:~$ show bgp neighbors 2001:db8:0:2::2 | match "Extended nexthop|Address Family IPv4 Unicast:"
    Extended nexthop: advertised and received
    Address Family IPv4 Unicast: advertised and received
vyos@R1:~$ show ip bgp
BGP table version is 2, local router ID is 0.0.0.1, vrf id 0
Default local pref 100, local AS 64500
Status codes:  s suppressed, d damped, h history, u unsorted, * valid, > best, = multipath,
               i internal, r RIB-failure, S Stale, R Removed
Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self
Origin codes:  i - IGP, e - EGP, ? - incomplete
RPKI validation codes: V valid, I invalid, N Not found

     Network          Next Hop            Metric LocPrf Weight Path
 *>  198.51.100.0/24  fe80::78c8:f1ff:fe05:427
                                               0             0 64512 i
 *>  203.0.113.0/24   fe80::5054:ff:fe02:1
                                                             0 64501 64502 i

Displayed 2 routes and 2 total paths
vyos@R1:~$ show bgp ipv4 203.0.113.0/24
BGP routing table entry for 203.0.113.0/24, version 2
Paths: (1 available, best #1, table default)
  Advertised to peers:
  2001:db8:0:1::2 2001:db8:0:2::2
  64501 64502
    2001:db8:0:2::2 from 2001:db8:0:2::2 (192.0.2.1)
    (fe80::5054:ff:fe02:1) (used)
      Origin IGP, valid, external, best (First path received)
      Last update: Sat Oct 17 01:26:13 2026
vyos@R1:~$ show ip route 203.0.113.0/24
Routing entry for 203.0.113.0/24
  Known via "bgp", distance 20, metric 0, best
  Last update 00:00:05 ago
  Flags: Selected
  Status: Installed
  * fe80::5054:ff:fe02:1, via eth1, weight 1

vyos@R1:~$ show ip route
Codes: K - kernel route, C - connected, L - local, S - static,
       R - RIP, O - OSPF, I - IS-IS, B - BGP, E - EIGRP, N - NHRP,
       T - Table, v - VNC, V - VNC-Direct, A - Babel, F - PBR,
       f - OpenFabric, t - Table-Direct,
       > - selected route, * - FIB route, q - queued, r - rejected, b - backup
       t - trapped, o - offload failure

IPv4 unicast VRF default:
B>* 198.51.100.0/24 [20/0] via fe80::78c8:f1ff:fe05:427, eth0, weight 1, 00:00:19
B>* 203.0.113.0/24 [20/0] via fe80::5054:ff:fe02:1, eth1, weight 1, 00:00:05

Lines 10–11: Routes now flow both ways (PfxSnt 2).
Line 15: The capability is negotiated.
Lines 27–30: IPv4 prefixes, IPv6 link-local next hops, valid and best.
Line 39: Learnt from R2's global address…
Line 40: …forwarded via its link-local address.
Lines 60–61: R1's whole IPv4 routing table: two BGP routes, no connected routes.

In the Linux kernel (real output)

vyos@R1:~$ ip -4 route
198.51.100.0/24 nhid 12 via inet6 fe80::78c8:f1ff:fe05:427 dev eth0 proto bgp metric 20
203.0.113.0/24 nhid 14 via inet6 fe80::5054:ff:fe02:1 dev eth1 proto bgp metric 20
vyos@R1:~$ ip -4 addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever

Lines 2–3: IPv4 routes with IPv6 gateways: via inet6.
Lines 4–7: R1's only IPv4 address is 127.0.0.1.

What VyOS rendered into FRR (real output)

vyos@R1:~$ vtysh -c 'show running-config bgpd'
...
router bgp 64500
 bgp router-id 0.0.0.1
 no bgp ebgp-requires-policy
 no bgp reject-as-sets
 no bgp default ipv4-unicast
 no bgp network import-check
 neighbor 2001:db8:0:1::2 remote-as 64512
 neighbor 2001:db8:0:1::2 description C1
 no neighbor 2001:db8:0:1::2 enforce-first-as
 neighbor 2001:db8:0:1::2 capability extended-nexthop
 neighbor 2001:db8:0:2::2 remote-as 64501
 neighbor 2001:db8:0:2::2 description R2
 no neighbor 2001:db8:0:2::2 enforce-first-as
 neighbor 2001:db8:0:2::2 capability extended-nexthop
 !
 address-family ipv4 unicast
  neighbor 2001:db8:0:1::2 activate
  neighbor 2001:db8:0:2::2 activate
 exit-address-family
exit

Line 4: The manual router ID.
Lines 12–16: capability extended-nexthop on each IPv6 neighbor.
Lines 18–21: IPv4 unicast activated for both.