Route-based XFRMi

Libreswan has, new in 2019, support for route based vpn using Linux XFRM device. There will be an ipsec1 device where you can add routes or "routing policies".

Route-based VPNs using libreswan XFRM device

By default Libreswan set up IPseec/VPN tunnels using IPsec policy alone, without "ip routing". On Linux systems this is called a policy-based VPN or IPsec. In libreswan, these policies are specified with leftsubnet= and rightsubnet= and optionally also with leftprotoport= and rightprotport=. Libreswan now, Kernel 4.19+ Libreswan 3.30+, allow you to setup a route-based VPN. A route based VPN need both policy and route. Libreswan will add the polices and basic routing in simple cases. In more advanced cases you can add addition destination based routes or routing policies. A ypical use case is policy covers everything 0/0 and add specific routes to send traffic via the tunnels. This is basically a policy-based VPN with leftsubnet=0.0.0.0/0 and rightsubnet=0.0.0.0/0.

The routes or routing policy (ip rule) go via XFRM device. Which is created by librewan and attached to the IPsec policy. Now, whenever a packet is routed into this XFRM device/interface, it will be encrypted. Everything that goes via this device will either encrypted or dropped based on associated IPsec policy. The encrypted EPS packet is routed using main routing table. While this type of setup is much less secure, it is easier to manage because you just need to update the routing table instead of adding or modifying IPsec policies. While libreswan supported route based VPN with KLIPS using the ipsec0 interface, as of libreswan-3.30, adds support using the Linux XFRMi device.

An additional advantage of using libreswan with XFRM is that you have a real network device (unlike the nflogXX interface) that supports firewalling with iptables, running tcpdump, etc. It is really functions as the old KLIPS ipsec0 interface.

New ipsec.conf keywords for XFRM support

The following new keywords have been added to support XFRM:

 ipsec-interface=no|yes|<n> . <n> to create/use interface "ipsecN" default, when "yes" is 1. Allowed range 1-UINT32_MAX (4294967295U) 

Using XFRM interface

An example configuration is shown below. It is a regular VPN connection with the additional options to turn it into a route-based VPN.

conn routed-vpn
    left=192.1.2.23
    right=192.1.2.45
    authby=rsasigkey
    leftsubnet=0.0.0.0/0
    rightsubnet=0.0.0.0/0
    auto=start
    ipsec-interface=yes

This will create the ipsec1 interface. If you want to encrypt all traffic to 10.0.0.0/8 using this VPN, simply run:

ip route add 10.0.0.0/8 dev ipsec1

To see or monitor the traffic sent before encryption or received after decrypion traffic, run:

tcpdump -i ipsec1 -n

And you can add firewall rules if you want:

# do not allow IRC traffic on port 6666
iptables -I INPUT -j DROP -p tcp --dport 6666 -i ipsec1

To see the post-encrypt and pre-decrypt (assuming your external interface is eth0), run:

tcpdump -i eth0 -n esp or udp port 4500

To see more information about XFRM device:

ip -d link show dev ipsec1
ip -s link show dev ipsec1

Setting up a route-based VPN with Amazon

[ to be filled in ]

XFRM related advanced use cases

updown script

The standard updown script (_updown.netkey) tries to autodetect what scenario is being deployed and will reconfigure the XFRM interface and network options accordingly. As this feature is still new, we expect it to not be perfect yet. If your scenario is not covered, please contact the developers and explain your use case.

Device options

Each device has a standard set of options associated with it. These can be found in /proc/sys/net/ipv4/conf/ipsec/*. The default updown script sets some of these (rp_filter, forwarding, policy_disable) but specific use cases might require different settings. Setting disable_xfrm will cause the device to completely fail, so do not do that.

XFRM device details

The xfrm device is internally attached to another network device such as eth0. Note "ipsec1@eth0" in the following output.

ip -d link show dev  ipsec1
2: ipsec1@eth0: <NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
    link/none 06:c7:58:c4:b2:c6 brd ff:ff:ff:ff:ff:ff promiscuity 0 minmtu 68 maxmtu 1500
    xfrm if_id 0x1 addrgenmode eui64 numtxqueues 1 numrxqueues 1 gso_max_size 65536 gso_max_segs 65535

#to see the policy details, with iproute5.1 or later
root@road:/#
ip xfrm state
src 192.1.2.23 dst 192.1.3.209
	proto esp spi 0xa7c66a14 reqid 16393 mode tunnel
	replay-window 32 flag af-unspec
	output-mark 0x1
	aead rfc4106(gcm(aes)) 0xd371dde7df8be108215590f49a084d665417ad63aa1896d51c03173b85d5037d02384567 128
	encap type espinudp sport 4500 dport 4500 addr 0.0.0.0
	anti-replay context: seq 0x0, oseq 0x0, bitmap 0x00000000
	if_id 0x1
src 192.1.3.209 dst 192.1.2.23
	proto esp spi 0x873318d8 reqid 16393 mode tunnel
	replay-window 32 flag af-unspec
	output-mark 0x1
	aead rfc4106(gcm(aes)) 0xbda3403de033c690a4c6c54651493bd8590042a56997ae127965e1e1f01de405843c4e55 128
	encap type espinudp sport 4500 dport 4500 addr 0.0.0.0
	anti-replay context: seq 0x0, oseq 0x0, bitmap 0x00000000

ip xfrm policy
src 192.0.2.1/32 dst 0.0.0.0/0
	dir out priority 1040383 ptype main
	tmpl src 192.1.3.209 dst 192.1.2.23
		proto esp reqid 16393 mode tunnel
	if_id 0x1
src 0.0.0.0/0 dst 192.0.2.1/32
	dir fwd priority 1040383 ptype main
	tmpl src 192.1.2.23 dst 192.1.3.209
		proto esp reqid 16393 mode tunnel
	if_id 0x1
src 0.0.0.0/0 dst 192.0.2.1/32
	dir in priority 1040383 ptype main
	tmpl src 192.1.2.23 dst 192.1.3.209
		proto esp reqid 16393 mode tunnel
	if_id 0x1

Note new fields: "if_id 0x1", "output-mark 0x1", "xfrm if_id 0x1". The output mark is not configurable, it is the same as the XFRM interface ID. Also note this mark not the same as the one used by VTI.

routing table for a typical roadwarrior client setup

ip rule show
0:	from all lookup local
100:	from all fwmark 0x1 lookup 50
32766:	from all lookup main
32767:	from all lookup default
road #
 ip route show
0.0.0.0/1 dev ipsec1 scope link src 192.0.2.1
default via 192.1.3.254 dev eth0
128.0.0.0/1 dev ipsec1 scope link src 192.0.2.1
192.1.3.0/24 dev eth0 proto kernel scope link src 192.1.3.209
road #
 ip route show table 50
192.1.2.23 via 192.1.3.254 dev eth0

Sharing ipsec-ineterface/XFRMi interface

these interfaces can be shared between different connection. The xfrm policy will match based on source and destination pick the right policy. So the simple case ipsec-interface=yes can share many connections. Also more than one address can be added to the same interface.

possible limittions of xfrmi??

no IKEv1 support in libreswan

Currently 3.30 ipsec-interface need ikev2=yes. May be something for the future to add.

0/0 route and ospf

During testing one issue was noticed by github issue

ipsec-interface on /32-to-32 responder

seems to create routing issue. On the responder IKE_AUTH response is sent through the tunnel and initiator never establish the connection. The initiator has no tunnel yet. So it won't see the response.

Configuration example

OpenWRT Static tunnel

/etc/config/network

config interface 'ipsec1'
        option proto 'xfrm'
        option mtu '1438'
        option zone 'VPN'
        option tunlink 'lan'
        option ifid '1'

config interface 'ipsec1_static'
        option proto 'static'
        option ifname '@ipsec1'
        option auto '0'
        list ipaddr '192.0.1.45/24'

libreswan connection config

/etc/ipsec.conf

conn staticvpn
     leftid=west
     rightid=east
     left=192.1.2.45
     right=192.1.2.23
     leftsubnet=192.0.1.0/24
     rightsubnet=192.0.2./24
     ipsec-interface=yes
     authby=secret