GIF(4) BSD Kernel Interfaces Manual GIF(4)NAME
gif — generic tunnel interface
The gif interface is a generic tunnelling device for IPv4 and IPv6. It
can tunnel IPv traffic over IPv. Therefore, there can be four
possible configurations. The behavior of gif is mainly based on RFC2893
IPv6-over-IPv4 configured tunnel. On NetBSD, gif can also tunnel ISO
traffic over IPv using EON encapsulation. Note that gif does not
perform GRE encapsulation; use gre(4) for GRE encapsulation.
Each gif interface is created at runtime using interface cloning. This
is most easily done with the “ifconfig create” command or using the
ifconfig_⟨interface⟩ variable in rc.conf(5).
To use gif, the administrator needs to configure the protocol and
addresses used for the outer header. This can be done by using
ifconfig(8) tunnel, or SIOCSIFPHYADDR ioctl. The administrator also
needs to configure the protocol and addresses for the inner header, with
ifconfig(8). Note that IPv6 link-local addresses (those that start with
fe80::) will be automatically configured whenever possible. You may need
to remove IPv6 link-local addresses manually using ifconfig(8), if you
want to disable the use of IPv6 as the inner header (for example, if you
need a pure IPv4-over-IPv6 tunnel). Finally, you must modify the routing
table to route the packets through the gif interface.
The gif device can be configured to be ECN friendly. This can be config‐
ured by IFF_LINK1.
ECN friendly behavior
The gif device can be configured to be ECN friendly, as described in
draft-ietf-ipsec-ecn-02.txt. This is turned off by default, and can be
turned on by the IFF_LINK1 interface flag.
Without IFF_LINK1, gif will show normal behavior, as described in
RFC2893. This can be summarized as follows:
Ingress Set outer TOS bit to 0.
Egress Drop outer TOS bit.
With IFF_LINK1, gif will copy ECN bits (0x02 and 0x01 on IPv4 TOS byte or
IPv6 traffic class byte) on egress and ingress, as follows:
Ingress Copy TOS bits except for ECN CE (masked with 0xfe) from
inner to outer. Set ECN CE bit to 0.
Egress Use inner TOS bits with some change. If outer ECN CE bit
is 1, enable ECN CE bit on the inner.
Note that the ECN friendly behavior violates RFC2893. This should be
used in mutual agreement with the peer.
A malicious party may try to circumvent security filters by using tun‐
nelled packets. For better protection, gif performs both martian and
ingress filtering against the outer source address on egress. Note that
martian/ingress filters are in no way complete. You may want to secure
your node by using packet filters. Ingress filtering can break tunnel
operation in an asymmetrically routed network. It can be turned off by
Processing each packet requires two route lookups: first on the packet
itself, and second on the tunnel destination. This second route can be
cached, increasing tunnel performance. However, in a dynamically routed
network, the tunnel will stick to the cached route, ignoring routing ta‐
ble updates. Route caching can be enabled with the IFF_LINK0 flag.
By default, gif tunnels may not be nested. This behavior may be modified
at runtime by setting the sysctl(8) variable net.link.gif.max_nesting to
the desired level of nesting. Additionally, gif tunnels are restricted
to one per pair of end points. Parallel tunnels may be enabled by set‐
ting the sysctl(8) variable net.link.gif.parallel_tunnels to 1.
SEE ALSOgre(4), inet(4), inet6(4), ifconfig(8)
R. Gilligan and E. Nordmark, "Transition Mechanisms for IPv6 Hosts and
Routers", RFC2893, August 2000, ftp://ftp.isi.edu/in-notes/rfc2893.txt.
Sally Floyd, David L. Black, and K. K. Ramakrishnan, IPsec Interactions
with ECN, December 1999, draft-ietf-ipsec-ecn-02.txt.
The gif device first appeared in the WIDE hydrangea IPv6 kit.
There are many tunnelling protocol specifications, all defined differ‐
ently from each other. The gif device may not interoperate with peers
which are based on different specifications, and are picky about outer
header fields. For example, you cannot usually use gif to talk with
IPsec devices that use IPsec tunnel mode.
The current code does not check if the ingress address (outer source
address) configured in the gif interface makes sense. Make sure to spec‐
ify an address which belongs to your node. Otherwise, your node will not
be able to receive packets from the peer, and it will generate packets
with a spoofed source address.
If the outer protocol is IPv4, gif does not try to perform path MTU dis‐
covery for the encapsulated packet (DF bit is set to 0).
If the outer protocol is IPv6, path MTU discovery for encapsulated pack‐
ets may affect communication over the interface. The first bigger-than-
pmtu packet may be lost. To avoid the problem, you may want to set the
interface MTU for gif to 1240 or smaller, when the outer header is IPv6
and the inner header is IPv4.
The gif device does not translate ICMP messages for the outer header into
the inner header.
In the past, gif had a multi-destination behavior, configurable via
IFF_LINK0 flag. The behavior is obsolete and is no longer supported.
On FreeBSD 6.1, 6.2, 6.3, 7.0, 7.1, and 7.2 the gif sends and receives
incorrect EtherIP packets with reversed version field when if_bridge(4)
is used together. As a workaround on this interoperability issue, the
following two ifconfig(8) flags can be used:
accept_rev_ethip_ver accepts both correct EtherIP packets and ones
with reversed version field, if enabled. If
disabled, the gif accepts the correct packets
only. This flag is enabled by default.
send_rev_ethip_ver sends EtherIP packets with reversed version
field intentionally, if enabled. If dis‐
abled, the gif sends the correct packets
only. This flag is disabled by default.
If interoperability with the older FreeBSD machines is needed, both of
these two flags must be enabled.
BSD June 8, 2009 BSD