1 <?xml version='
1.0'
?> <!--*-nxml-*-->
2 <!DOCTYPE refentry PUBLIC
"-//OASIS//DTD DocBook XML V4.2//EN"
3 "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
6 SPDX-License-Identifier: LGPL-2.1+
8 This file is part of systemd.
10 Copyright 2013 Tom Gundersen
13 <refentry id=
"systemd.netdev" conditional='ENABLE_NETWORKD'
>
16 <title>systemd.network
</title>
17 <productname>systemd
</productname>
21 <contrib>Developer
</contrib>
22 <firstname>Tom
</firstname>
23 <surname>Gundersen
</surname>
24 <email>teg@jklm.no
</email>
30 <refentrytitle>systemd.netdev
</refentrytitle>
31 <manvolnum>5</manvolnum>
35 <refname>systemd.netdev
</refname>
36 <refpurpose>Virtual Network Device configuration
</refpurpose>
40 <para><filename><replaceable>netdev
</replaceable>.netdev
</filename></para>
44 <title>Description
</title>
46 <para>Network setup is performed by
47 <citerefentry><refentrytitle>systemd-networkd
</refentrytitle><manvolnum>8</manvolnum></citerefentry>.
50 <para>The main Virtual Network Device file must have the extension
<filename>.netdev
</filename>;
51 other extensions are ignored. Virtual network devices are created as soon as networkd is
52 started. If a netdev with the specified name already exists, networkd will use that as-is rather
53 than create its own. Note that the settings of the pre-existing netdev will not be changed by
56 <para>The
<filename>.netdev
</filename> files are read from the files located in the system
57 network directory
<filename>/usr/lib/systemd/network
</filename>, the volatile runtime network
58 directory
<filename>/run/systemd/network
</filename> and the local administration network
59 directory
<filename>/etc/systemd/network
</filename>. All configuration files are collectively
60 sorted and processed in lexical order, regardless of the directories in which they live.
61 However, files with identical filenames replace each other. Files in
<filename>/etc
</filename>
62 have the highest priority, files in
<filename>/run
</filename> take precedence over files with
63 the same name in
<filename>/usr/lib
</filename>. This can be used to override a system-supplied
64 configuration file with a local file if needed. As a special case, an empty file (file size
0)
65 or symlink with the same name pointing to
<filename>/dev/null
</filename> disables the
66 configuration file entirely (it is
"masked").
</para>
68 <para>Along with the netdev file
<filename>foo.netdev
</filename>, a
"drop-in" directory
69 <filename>foo.netdev.d/
</filename> may exist. All files with the suffix
<literal>.conf
</literal>
70 from this directory will be parsed after the file itself is parsed. This is useful to alter or
71 add configuration settings, without having to modify the main configuration file. Each drop-in
72 file must have appropriate section headers.
</para>
74 <para>In addition to
<filename>/etc/systemd/network
</filename>, drop-in
<literal>.d
</literal>
75 directories can be placed in
<filename>/usr/lib/systemd/network
</filename> or
76 <filename>/run/systemd/network
</filename> directories. Drop-in files in
77 <filename>/etc
</filename> take precedence over those in
<filename>/run
</filename> which in turn
78 take precedence over those in
<filename>/usr/lib
</filename>. Drop-in files under any of these
79 directories take precedence over the main netdev file wherever located. (Of course, since
80 <filename>/run
</filename> is temporary and
<filename>/usr/lib
</filename> is for vendors, it is
81 unlikely drop-ins should be used in either of those places.)
</para>
85 <title>Supported netdev kinds
</title>
87 <para>The following kinds of virtual network devices may be
88 configured in
<filename>.netdev
</filename> files:
</para>
91 <title>Supported kinds of virtual network devices
</title>
94 <colspec colname='kind'
/>
95 <colspec colname='explanation'
/>
98 <entry>Description
</entry>
101 <row><entry><varname>bond
</varname></entry>
102 <entry>A bond device is an aggregation of all its slave devices. See
<ulink url=
"https://www.kernel.org/doc/Documentation/networking/bonding.txt">Linux Ethernet Bonding Driver HOWTO
</ulink> for details.Local configuration
</entry></row>
104 <row><entry><varname>bridge
</varname></entry>
105 <entry>A bridge device is a software switch, and each of its slave devices and the bridge itself are ports of the switch.
</entry></row>
107 <row><entry><varname>dummy
</varname></entry>
108 <entry>A dummy device drops all packets sent to it.
</entry></row>
110 <row><entry><varname>gre
</varname></entry>
111 <entry>A Level
3 GRE tunnel over IPv4. See
<ulink url=
"https://tools.ietf.org/html/rfc2784">RFC
2784</ulink> for details.
</entry></row>
113 <row><entry><varname>gretap
</varname></entry>
114 <entry>A Level
2 GRE tunnel over IPv4.
</entry></row>
116 <row><entry><varname>ip6gre
</varname></entry>
117 <entry>A Level
3 GRE tunnel over IPv6.
</entry></row>
119 <row><entry><varname>ip6tnl
</varname></entry>
120 <entry>An IPv4 or IPv6 tunnel over IPv6
</entry></row>
122 <row><entry><varname>ip6gretap
</varname></entry>
123 <entry>A Level
2 GRE tunnel over IPv6.
</entry></row>
125 <row><entry><varname>ipip
</varname></entry>
126 <entry>An IPv4 over IPv4 tunnel.
</entry></row>
128 <row><entry><varname>ipvlan
</varname></entry>
129 <entry>An ipvlan device is a stacked device which receives packets from its underlying device based on IP address filtering.
</entry></row>
131 <row><entry><varname>macvlan
</varname></entry>
132 <entry>A macvlan device is a stacked device which receives packets from its underlying device based on MAC address filtering.
</entry></row>
134 <row><entry><varname>macvtap
</varname></entry>
135 <entry>A macvtap device is a stacked device which receives packets from its underlying device based on MAC address filtering.
</entry></row>
137 <row><entry><varname>sit
</varname></entry>
138 <entry>An IPv6 over IPv4 tunnel.
</entry></row>
140 <row><entry><varname>tap
</varname></entry>
141 <entry>A persistent Level
2 tunnel between a network device and a device node.
</entry></row>
143 <row><entry><varname>tun
</varname></entry>
144 <entry>A persistent Level
3 tunnel between a network device and a device node.
</entry></row>
146 <row><entry><varname>veth
</varname></entry>
147 <entry>An Ethernet tunnel between a pair of network devices.
</entry></row>
149 <row><entry><varname>vlan
</varname></entry>
150 <entry>A VLAN is a stacked device which receives packets from its underlying device based on VLAN tagging. See
<ulink url=
"http://www.ieee802.org/1/pages/802.1Q.html">IEEE
802.1Q
</ulink> for details.
</entry></row>
152 <row><entry><varname>vti
</varname></entry>
153 <entry>An IPv4 over IPSec tunnel.
</entry></row>
155 <row><entry><varname>vti6
</varname></entry>
156 <entry>An IPv6 over IPSec tunnel.
</entry></row>
158 <row><entry><varname>vxlan
</varname></entry>
159 <entry>A virtual extensible LAN (vxlan), for connecting Cloud computing deployments.
</entry></row>
161 <row><entry><varname>geneve
</varname></entry>
162 <entry>A GEneric NEtwork Virtualization Encapsulation (GENEVE) netdev driver.
</entry></row>
164 <row><entry><varname>vrf
</varname></entry>
165 <entry>A Virtual Routing and Forwarding (
<ulink url=
"https://www.kernel.org/doc/Documentation/networking/vrf.txt">VRF
</ulink>) interface to create separate routing and forwarding domains.
</entry></row>
167 <row><entry><varname>vcan
</varname></entry>
168 <entry>The virtual CAN driver (vcan). Similar to the network loopback devices, vcan offers a virtual local CAN interface.
</entry></row>
170 <row><entry><varname>vxcan
</varname></entry>
171 <entry>The virtual CAN tunnel driver (vxcan). Similar to the virtual ethernet driver veth, vxcan implements a local CAN traffic tunnel between two virtual CAN network devices. When creating a vxcan, two vxcan devices are created as pair. When one end receives the packet it appears on its pair and vice versa. The vxcan can be used for cross namespace communication.
174 <row><entry><varname>wireguard
</varname></entry>
175 <entry>WireGuard Secure Network Tunnel.
</entry></row>
184 <title>[Match] Section Options
</title>
186 <para>A virtual network device is only created if the
187 <literal>[Match]
</literal> section matches the current
188 environment, or if the section is empty. The following keys are
191 <variablelist class='network-directives'
>
193 <term><varname>Host=
</varname></term>
195 <para>Matches against the hostname or machine ID of the
196 host. See
<literal>ConditionHost=
</literal> in
197 <citerefentry><refentrytitle>systemd.unit
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
203 <term><varname>Virtualization=
</varname></term>
205 <para>Checks whether the system is executed in a virtualized
206 environment and optionally test whether it is a specific
208 <literal>ConditionVirtualization=
</literal> in
209 <citerefentry><refentrytitle>systemd.unit
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
215 <term><varname>KernelCommandLine=
</varname></term>
217 <para>Checks whether a specific kernel command line option
218 is set (or if prefixed with the exclamation mark unset). See
219 <literal>ConditionKernelCommandLine=
</literal> in
220 <citerefentry><refentrytitle>systemd.unit
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
226 <term><varname>KernelVersion=
</varname></term>
228 <para>Checks whether the kernel version (as reported by
<command>uname -r
</command>) matches a certain
229 expression (or if prefixed with the exclamation mark does not match it). See
230 <literal>ConditionKernelVersion=
</literal> in
231 <citerefentry><refentrytitle>systemd.unit
</refentrytitle><manvolnum>5</manvolnum></citerefentry> for details.
236 <term><varname>Architecture=
</varname></term>
238 <para>Checks whether the system is running on a specific
239 architecture. See
<literal>ConditionArchitecture=
</literal> in
240 <citerefentry><refentrytitle>systemd.unit
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
250 <title>[NetDev] Section Options
</title>
252 <para>The
<literal>[NetDev]
</literal> section accepts the
253 following keys:
</para>
255 <variablelist class='network-directives'
>
257 <term><varname>Description=
</varname></term>
259 <para>A free-form description of the netdev.
</para>
263 <term><varname>Name=
</varname></term>
265 <para>The interface name used when creating the netdev.
266 This option is compulsory.
</para>
270 <term><varname>Kind=
</varname></term>
272 <para>The netdev kind. This option is compulsory. See the
273 <literal>Supported netdev kinds
</literal> section for the
278 <term><varname>MTUBytes=
</varname></term>
280 <para>The maximum transmission unit in bytes to set for
281 the device. The usual suffixes K, M, G, are supported and
282 are understood to the base of
1024. This key is not
283 currently supported for
<literal>tun
</literal> or
284 <literal>tap
</literal> devices.
289 <term><varname>MACAddress=
</varname></term>
291 <para>The MAC address to use for the device. If none is
292 given, one is generated based on the interface name and
294 <citerefentry><refentrytitle>machine-id
</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
295 This key is not currently supported for
296 <literal>tun
</literal> or
<literal>tap
</literal> devices.
304 <title>[Bridge] Section Options
</title>
306 <para>The
<literal>[Bridge]
</literal> section only applies for
307 netdevs of kind
<literal>bridge
</literal>, and accepts the
308 following keys:
</para>
310 <variablelist class='network-directives'
>
312 <term><varname>HelloTimeSec=
</varname></term>
314 <para>HelloTimeSec specifies the number of seconds between two hello packets
315 sent out by the root bridge and the designated bridges. Hello packets are
316 used to communicate information about the topology throughout the entire
317 bridged local area network.
</para>
321 <term><varname>MaxAgeSec=
</varname></term>
323 <para>MaxAgeSec specifies the number of seconds of maximum message age.
324 If the last seen (received) hello packet is more than this number of
325 seconds old, the bridge in question will start the takeover procedure
326 in attempt to become the Root Bridge itself.
</para>
330 <term><varname>ForwardDelaySec=
</varname></term>
332 <para>ForwardDelaySec specifies the number of seconds spent in each
333 of the Listening and Learning states before the Forwarding state is entered.
</para>
337 <term><varname>AgeingTimeSec=
</varname></term>
339 <para>This specifies the number of seconds a MAC Address will be kept in
340 the forwarding database after having a packet received from this MAC Address.
</para>
344 <term><varname>Priority=
</varname></term>
346 <para>The priority of the bridge. An integer between
0 and
65535. A lower value
347 means higher priority. The bridge having the lowest priority will be elected as root bridge.
</para>
351 <term><varname>GroupForwardMask=
</varname></term>
353 <para>A
16-bit bitmask represented as an integer which allows forwarding of link
354 local frames with
802.1D reserved addresses (
01:
80:C2:
00:
00:
0X). A logical AND
355 is performed between the specified bitmask and the exponentiation of
2^X, the
356 lower nibble of the last octet of the MAC address. For example, a value of
8
357 would allow forwarding of frames addressed to
01:
80:C2:
00:
00:
03 (
802.1X PAE).
</para>
361 <term><varname>DefaultPVID=
</varname></term>
363 <para>This specifies the default port VLAN ID of a newly attached bridge port.
364 Set this to an integer in the range
1–
4094 or
<literal>none
</literal> to disable the PVID.
</para>
368 <term><varname>MulticastQuerier=
</varname></term>
370 <para>A boolean. This setting controls the IFLA_BR_MCAST_QUERIER option in the kernel.
371 If enabled, the kernel will send general ICMP queries from a zero source address.
372 This feature should allow faster convergence on startup, but it causes some
373 multicast-aware switches to misbehave and disrupt forwarding of multicast packets.
374 When unset, the kernel's default setting applies.
379 <term><varname>MulticastSnooping=
</varname></term>
381 <para>A boolean. This setting controls the IFLA_BR_MCAST_SNOOPING option in the kernel.
382 If enabled, IGMP snooping monitors the Internet Group Management Protocol (IGMP) traffic
383 between hosts and multicast routers. When unset, the kernel's default setting applies.
388 <term><varname>VLANFiltering=
</varname></term>
390 <para>A boolean. This setting controls the IFLA_BR_VLAN_FILTERING option in the kernel.
391 If enabled, the bridge will be started in VLAN-filtering mode. When unset, the kernel's
392 default setting applies.
397 <term><varname>STP=
</varname></term>
399 <para>A boolean. This enables the bridge's Spanning Tree Protocol (STP). When unset,
400 the kernel's default setting applies.
408 <title>[VLAN] Section Options
</title>
410 <para>The
<literal>[VLAN]
</literal> section only applies for
411 netdevs of kind
<literal>vlan
</literal>, and accepts the
412 following key:
</para>
414 <variablelist class='network-directives'
>
416 <term><varname>Id=
</varname></term>
418 <para>The VLAN ID to use. An integer in the range
0–
4094.
419 This option is compulsory.
</para>
423 <term><varname>GVRP=
</varname></term>
425 <para>The Generic VLAN Registration Protocol (GVRP) is a protocol that
426 allows automatic learning of VLANs on a network. A boolean. When unset,
427 the kernel's default setting applies.
</para>
431 <term><varname>MVRP=
</varname></term>
433 <para>Multiple VLAN Registration Protocol (MVRP) formerly known as GARP VLAN
434 Registration Protocol (GVRP) is a standards-based Layer
2 network protocol,
435 for automatic configuration of VLAN information on switches. It was defined
436 in the
802.1ak amendment to
802.1Q-
2005. A boolean. When unset, the kernel's
437 default setting applies.
</para>
441 <term><varname>LooseBinding=
</varname></term>
443 <para>The VLAN loose binding mode, in which only the operational state is passed
444 from the parent to the associated VLANs, but the VLAN device state is not changed.
445 A boolean. When unset, the kernel's default setting applies.
</para>
449 <term><varname>ReorderHeader=
</varname></term>
451 <para>The VLAN reorder header is set VLAN interfaces behave like physical interfaces.
452 A boolean. When unset, the kernel's default setting applies.
</para>
459 <title>[MACVLAN] Section Options
</title>
461 <para>The
<literal>[MACVLAN]
</literal> section only applies for
462 netdevs of kind
<literal>macvlan
</literal>, and accepts the
463 following key:
</para>
465 <variablelist class='network-directives'
>
467 <term><varname>Mode=
</varname></term>
469 <para>The MACVLAN mode to use. The supported options are
470 <literal>private
</literal>,
471 <literal>vepa
</literal>,
472 <literal>bridge
</literal>, and
473 <literal>passthru
</literal>.
482 <title>[MACVTAP] Section Options
</title>
484 <para>The
<literal>[MACVTAP]
</literal> section applies for
485 netdevs of kind
<literal>macvtap
</literal> and accepts the
486 same key as
<literal>[MACVLAN]
</literal>.
</para>
491 <title>[IPVLAN] Section Options
</title>
493 <para>The
<literal>[IPVLAN]
</literal> section only applies for
494 netdevs of kind
<literal>ipvlan
</literal>, and accepts the
495 following key:
</para>
497 <variablelist class='network-directives'
>
499 <term><varname>Mode=
</varname></term>
501 <para>The IPVLAN mode to use. The supported options are
502 <literal>L2
</literal>,
<literal>L3
</literal> and
<literal>L3S
</literal>.
507 <term><varname>Flags=
</varname></term>
509 <para>The IPVLAN flags to use. The supported options are
510 <literal>bridge
</literal>,
<literal>private
</literal> and
<literal>vepa
</literal>.
519 <title>[VXLAN] Section Options
</title>
520 <para>The
<literal>[VXLAN]
</literal> section only applies for
521 netdevs of kind
<literal>vxlan
</literal>, and accepts the
522 following keys:
</para>
524 <variablelist class='network-directives'
>
526 <term><varname>Id=
</varname></term>
528 <para>The VXLAN ID to use.
</para>
532 <term><varname>Remote=
</varname></term>
534 <para>Configures destination IP address.
</para>
538 <term><varname>Local=
</varname></term>
540 <para>Configures local IP address.
</para>
544 <term><varname>TOS=
</varname></term>
546 <para>The Type Of Service byte value for a vxlan interface.
</para>
550 <term><varname>TTL=
</varname></term>
552 <para>A fixed Time To Live N on Virtual eXtensible Local
553 Area Network packets. N is a number in the range
1–
255.
0
554 is a special value meaning that packets inherit the TTL
559 <term><varname>MacLearning=
</varname></term>
561 <para>A boolean. When true, enables dynamic MAC learning
562 to discover remote MAC addresses.
</para>
566 <term><varname>FDBAgeingSec=
</varname></term>
568 <para>The lifetime of Forwarding Database entry learnt by
569 the kernel, in seconds.
</para>
573 <term><varname>MaximumFDBEntries=
</varname></term>
575 <para>Configures maximum number of FDB entries.
</para>
579 <term><varname>ReduceARPProxy=
</varname></term>
581 <para>A boolean. When true, bridge-connected VXLAN tunnel
582 endpoint answers ARP requests from the local bridge on behalf
583 of remote Distributed Overlay Virtual Ethernet
584 <ulink url=
"https://en.wikipedia.org/wiki/Distributed_Overlay_Virtual_Ethernet">
585 (DVOE)
</ulink> clients. Defaults to false.
</para>
589 <term><varname>L2MissNotification=
</varname></term>
591 <para>A boolean. When true, enables netlink LLADDR miss
592 notifications.
</para>
596 <term><varname>L3MissNotification=
</varname></term>
598 <para>A boolean. When true, enables netlink IP address miss
599 notifications.
</para>
603 <term><varname>RouteShortCircuit=
</varname></term>
605 <para>A boolean. When true, route short circuiting is turned
610 <term><varname>UDPChecksum=
</varname></term>
612 <para>A boolean. When true, transmitting UDP checksums when doing VXLAN/IPv4 is turned on.
</para>
616 <term><varname>UDP6ZeroChecksumTx=
</varname></term>
618 <para>A boolean. When true, sending zero checksums in VXLAN/IPv6 is turned on.
</para>
622 <term><varname>UDP6ZeroChecksumRx=
</varname></term>
624 <para>A boolean. When true, receiving zero checksums in VXLAN/IPv6 is turned on.
</para>
628 <term><varname>RemoteChecksumTx=
</varname></term>
630 <para>A boolean. When true, remote transmit checksum offload of VXLAN is turned on.
</para>
634 <term><varname>RemoteChecksumRx=
</varname></term>
636 <para>A boolean. When true, remote receive checksum offload in VXLAN is turned on.
</para>
640 <term><varname>GroupPolicyExtension=
</varname></term>
642 <para>A boolean. When true, it enables Group Policy VXLAN extension security label mechanism
643 across network peers based on VXLAN. For details about the Group Policy VXLAN, see the
644 <ulink url=
"https://tools.ietf.org/html/draft-smith-vxlan-group-policy">
645 VXLAN Group Policy
</ulink> document. Defaults to false.
</para>
649 <term><varname>DestinationPort=
</varname></term>
651 <para>Configures the default destination UDP port on a per-device basis.
652 If destination port is not specified then Linux kernel default will be used.
653 Set destination port
4789 to get the IANA assigned value. If not set or if the
654 destination port is assigned the empty string the default port of
4789 is used.
</para>
658 <term><varname>PortRange=
</varname></term>
660 <para>Configures VXLAN port range. VXLAN bases source
661 UDP port based on flow to help the receiver to be able
662 to load balance based on outer header flow. It
663 restricts the port range to the normal UDP local
664 ports, and allows overriding via configuration.
</para>
668 <term><varname>FlowLabel=
</varname></term>
670 <para>Specifies the flow label to use in outgoing packets.
671 The valid range is
0-
1048575.
678 <title>[GENEVE] Section Options
</title>
679 <para>The
<literal>[GENEVE]
</literal> section only applies for
680 netdevs of kind
<literal>geneve
</literal>, and accepts the
681 following keys:
</para>
683 <variablelist class='network-directives'
>
685 <term><varname>Id=
</varname></term>
687 <para>Specifies the Virtual Network Identifier (VNI) to use. Ranges [
0-
16777215].
</para>
691 <term><varname>Remote=
</varname></term>
693 <para>Specifies the unicast destination IP address to use in outgoing packets.
</para>
697 <term><varname>TOS=
</varname></term>
699 <para>Specifies the TOS value to use in outgoing packets. Ranges [
1-
255].
</para>
703 <term><varname>TTL=
</varname></term>
705 <para>Specifies the TTL value to use in outgoing packets. Ranges [
1-
255].
</para>
709 <term><varname>UDPChecksum=
</varname></term>
711 <para>A boolean. When true, specifies if UDP checksum is calculated for transmitted packets over IPv4.
</para>
715 <term><varname>UDP6ZeroChecksumTx=
</varname></term>
717 <para>A boolean. When true, skip UDP checksum calculation for transmitted packets over IPv6.
</para>
721 <term><varname>UDP6ZeroChecksumRx=
</varname></term>
723 <para>A boolean. When true, allows incoming UDP packets over IPv6 with zero checksum field.
</para>
727 <term><varname>DestinationPort=
</varname></term>
729 <para>Specifies destination port. Defaults to
6081. If not set or assigned the empty string, the default
730 port of
6081 is used.
</para>
734 <term><varname>FlowLabel=
</varname></term>
736 <para>Specifies the flow label to use in outgoing packets.
</para>
742 <title>[Tunnel] Section Options
</title>
744 <para>The
<literal>[Tunnel]
</literal> section only applies for
746 <literal>ipip
</literal>,
747 <literal>sit
</literal>,
748 <literal>gre
</literal>,
749 <literal>gretap
</literal>,
750 <literal>ip6gre
</literal>,
751 <literal>ip6gretap
</literal>,
752 <literal>vti
</literal>,
753 <literal>vti6
</literal>, and
754 <literal>ip6tnl
</literal> and accepts
755 the following keys:
</para>
757 <variablelist class='network-directives'
>
759 <term><varname>Local=
</varname></term>
761 <para>A static local address for tunneled packets. It must
762 be an address on another interface of this host.
</para>
766 <term><varname>Remote=
</varname></term>
768 <para>The remote endpoint of the tunnel.
</para>
772 <term><varname>TOS=
</varname></term>
774 <para>The Type Of Service byte value for a tunnel interface.
775 For details about the TOS, see the
776 <ulink url=
"http://tools.ietf.org/html/rfc1349"> Type of
777 Service in the Internet Protocol Suite
</ulink> document.
782 <term><varname>TTL=
</varname></term>
784 <para>A fixed Time To Live N on tunneled packets. N is a
785 number in the range
1–
255.
0 is a special value meaning that
786 packets inherit the TTL value. The default value for IPv4
787 tunnels is: inherit. The default value for IPv6 tunnels is
792 <term><varname>DiscoverPathMTU=
</varname></term>
794 <para>A boolean. When true, enables Path MTU Discovery on
799 <term><varname>IPv6FlowLabel=
</varname></term>
801 <para>Configures the
20-bit flow label (see
<ulink url=
"https://tools.ietf.org/html/rfc6437">
802 RFC
6437</ulink>) field in the IPv6 header (see
<ulink url=
"https://tools.ietf.org/html/rfc2460">
803 RFC
2460</ulink>), which is used by a node to label packets of a flow.
804 It is only used for IPv6 tunnels.
805 A flow label of zero is used to indicate packets that have
807 It can be configured to a value in the range
0–
0xFFFFF, or be
808 set to
<literal>inherit
</literal>, in which case the original flowlabel is used.
</para>
812 <term><varname>CopyDSCP=
</varname></term>
814 <para>A boolean. When true, the Differentiated Service Code
815 Point (DSCP) field will be copied to the inner header from
816 outer header during the decapsulation of an IPv6 tunnel
817 packet. DSCP is a field in an IP packet that enables different
818 levels of service to be assigned to network traffic.
819 Defaults to
<literal>no
</literal>.
824 <term><varname>EncapsulationLimit=
</varname></term>
826 <para>The Tunnel Encapsulation Limit option specifies how many additional
827 levels of encapsulation are permitted to be prepended to the packet.
828 For example, a Tunnel Encapsulation Limit option containing a limit
829 value of zero means that a packet carrying that option may not enter
830 another tunnel before exiting the current tunnel.
831 (see
<ulink url=
"https://tools.ietf.org/html/rfc2473#section-4.1.1"> RFC
2473</ulink>).
832 The valid range is
0–
255 and
<literal>none
</literal>. Defaults to
4.
837 <term><varname>Key=
</varname></term>
839 <para>The
<varname>Key=
</varname> parameter specifies the same key to use in
840 both directions (
<varname>InputKey=
</varname> and
<varname>OutputKey=
</varname>).
841 The
<varname>Key=
</varname> is either a number or an IPv4 address-like dotted quad.
842 It is used as mark-configured SAD/SPD entry as part of the lookup key (both in data
843 and control path) in ip xfrm (framework used to implement IPsec protocol).
844 See
<ulink url=
"http://man7.org/linux/man-pages/man8/ip-xfrm.8.html">
845 ip-xfrm — transform configuration
</ulink> for details. It is only used for VTI/VTI6
850 <term><varname>InputKey=
</varname></term>
852 <para>The
<varname>InputKey=
</varname> parameter specifies the key to use for input.
853 The format is same as
<varname>Key=
</varname>. It is only used for VTI/VTI6 tunnels.
</para>
857 <term><varname>OutputKey=
</varname></term>
859 <para>The
<varname>OutputKey=
</varname> parameter specifies the key to use for output.
860 The format is same as
<varname>Key=
</varname>. It is only used for VTI/VTI6 tunnels.
</para>
864 <term><varname>Mode=
</varname></term>
866 <para>An
<literal>ip6tnl
</literal> tunnel can be in one of three
868 <literal>ip6ip6
</literal> for IPv6 over IPv6,
869 <literal>ipip6
</literal> for IPv4 over IPv6 or
870 <literal>any
</literal> for either.
875 <term><varname>Independent=
</varname></term>
877 <para>A boolean. When true tunnel does not require .network file. Created as
"tunnel@NONE".
878 Defaults to
<literal>false
</literal>.
883 <term><varname>AllowLocalRemote=
</varname></term>
885 <para>A boolean. When true allows tunnel traffic on
<varname>ip6tnl
</varname> devices where the remote endpoint is a local host address.
893 <title>[Peer] Section Options
</title>
895 <para>The
<literal>[Peer]
</literal> section only applies for
896 netdevs of kind
<literal>veth
</literal> and accepts the
897 following keys:
</para>
899 <variablelist class='network-directives'
>
901 <term><varname>Name=
</varname></term>
903 <para>The interface name used when creating the netdev.
904 This option is compulsory.
</para>
908 <term><varname>MACAddress=
</varname></term>
910 <para>The peer MACAddress, if not set, it is generated in
911 the same way as the MAC address of the main
918 <title>[VXCAN] Section Options
</title>
919 <para>The
<literal>[VXCAN]
</literal> section only applies for
920 netdevs of kind
<literal>vxcan
</literal> and accepts the
921 following key:
</para>
923 <variablelist class='network-directives'
>
925 <term><varname>Peer=
</varname></term>
927 <para>The peer interface name used when creating the netdev.
928 This option is compulsory.
</para>
934 <title>[Tun] Section Options
</title>
936 <para>The
<literal>[Tun]
</literal> section only applies for
937 netdevs of kind
<literal>tun
</literal>, and accepts the following
940 <variablelist class='network-directives'
>
942 <term><varname>OneQueue=
</varname></term>
943 <listitem><para>Takes a boolean argument. Configures whether
944 all packets are queued at the device (enabled), or a fixed
945 number of packets are queued at the device and the rest at the
946 <literal>qdisc
</literal>. Defaults to
947 <literal>no
</literal>.
</para>
951 <term><varname>MultiQueue=
</varname></term>
952 <listitem><para>Takes a boolean argument. Configures whether
953 to use multiple file descriptors (queues) to parallelize
954 packets sending and receiving. Defaults to
955 <literal>no
</literal>.
</para>
959 <term><varname>PacketInfo=
</varname></term>
960 <listitem><para>Takes a boolean argument. Configures whether
961 packets should be prepended with four extra bytes (two flag
962 bytes and two protocol bytes). If disabled, it indicates that
963 the packets will be pure IP packets. Defaults to
964 <literal>no
</literal>.
</para>
968 <term><varname>VNetHeader=
</varname></term>
969 <listitem><para>Takes a boolean argument. Configures
970 IFF_VNET_HDR flag for a tap device. It allows sending
971 and receiving larger Generic Segmentation Offload (GSO)
972 packets. This may increase throughput significantly.
974 <literal>no
</literal>.
</para>
978 <term><varname>User=
</varname></term>
979 <listitem><para>User to grant access to the
980 <filename>/dev/net/tun
</filename> device.
</para>
984 <term><varname>Group=
</varname></term>
985 <listitem><para>Group to grant access to the
986 <filename>/dev/net/tun
</filename> device.
</para>
995 <title>[Tap] Section Options
</title>
997 <para>The
<literal>[Tap]
</literal> section only applies for
998 netdevs of kind
<literal>tap
</literal>, and accepts the same keys
999 as the
<literal>[Tun]
</literal> section.
</para>
1003 <title>[WireGuard] Section Options
</title>
1005 <para>The
<literal>[WireGuard]
</literal> section accepts the following
1008 <variablelist class='network-directives'
>
1010 <term><varname>PrivateKey=
</varname></term>
1012 <para>The Base64 encoded private key for the interface. It can be
1013 generated using the
<command>wg genkey
</command> command
1014 (see
<citerefentry project=
"wireguard"><refentrytitle>wg
</refentrytitle><manvolnum>8</manvolnum></citerefentry>).
1015 This option is mandatory to use WireGuard.
1016 Note that because this information is secret, you may want to set
1017 the permissions of the .netdev file to be owned by
<literal>root:systemd-networkd
</literal>
1018 with a
<literal>0640</literal> file mode.
</para>
1022 <term><varname>ListenPort=
</varname></term>
1024 <para>Sets UDP port for listening. Takes either value between
1 and
65535
1025 or
<literal>auto
</literal>. If
<literal>auto
</literal> is specified,
1026 the port is automatically generated based on interface name.
1027 Defaults to
<literal>auto
</literal>.
</para>
1031 <term><varname>FwMark=
</varname></term>
1033 <para>Sets a firewall mark on outgoing WireGuard packets from this interface.
</para>
1040 <title>[WireGuardPeer] Section Options
</title>
1042 <para>The
<literal>[WireGuardPeer]
</literal> section accepts the following
1045 <variablelist class='network-directives'
>
1047 <term><varname>PublicKey=
</varname></term>
1049 <para>Sets a Base64 encoded public key calculated by
<command>wg pubkey
</command>
1050 (see
<citerefentry project=
"wireguard"><refentrytitle>wg
</refentrytitle><manvolnum>8</manvolnum></citerefentry>)
1051 from a private key, and usually transmitted out of band to the
1052 author of the configuration file. This option is mandatory for this
1057 <term><varname>PresharedKey=
</varname></term>
1059 <para>Optional preshared key for the interface. It can be generated
1060 by the
<command>wg genpsk
</command> command. This option adds an
1061 additional layer of symmetric-key cryptography to be mixed into the
1062 already existing public-key cryptography, for post-quantum
1064 Note that because this information is secret, you may want to set
1065 the permissions of the .netdev file to be owned by
<literal>root:systemd-networkd
</literal>
1066 with a
<literal>0640</literal> file mode.
</para>
1070 <term><varname>AllowedIPs=
</varname></term>
1072 <para>Sets a comma-separated list of IP (v4 or v6) addresses with CIDR masks
1073 from which this peer is allowed to send incoming traffic and to
1074 which outgoing traffic for this peer is directed. The catch-all
1075 0.0.0.0/
0 may be specified for matching all IPv4 addresses, and
1076 ::/
0 may be specified for matching all IPv6 addresses.
</para>
1080 <term><varname>Endpoint=
</varname></term>
1082 <para>Sets an endpoint IP address or hostname, followed by a colon, and then
1083 a port number. This endpoint will be updated automatically once to
1084 the most recent source IP address and port of correctly
1085 authenticated packets from the peer at configuration time.
</para>
1089 <term><varname>PersistentKeepalive=
</varname></term>
1091 <para>Sets a seconds interval, between
1 and
65535 inclusive, of how often
1092 to send an authenticated empty packet to the peer for the purpose
1093 of keeping a stateful firewall or NAT mapping valid persistently.
1094 For example, if the interface very rarely sends traffic, but it
1095 might at anytime receive traffic from a peer, and it is behind NAT,
1096 the interface might benefit from having a persistent keepalive
1097 interval of
25 seconds. If set to
0 or
"off", this option is
1098 disabled. By default or when unspecified, this option is off.
1099 Most users will not need this.
</para>
1106 <title>[Bond] Section Options
</title>
1108 <para>The
<literal>[Bond]
</literal> section accepts the following
1111 <variablelist class='network-directives'
>
1113 <term><varname>Mode=
</varname></term>
1115 <para>Specifies one of the bonding policies. The default is
1116 <literal>balance-rr
</literal> (round robin). Possible values are
1117 <literal>balance-rr
</literal>,
1118 <literal>active-backup
</literal>,
1119 <literal>balance-xor
</literal>,
1120 <literal>broadcast
</literal>,
1121 <literal>802.3ad
</literal>,
1122 <literal>balance-tlb
</literal>, and
1123 <literal>balance-alb
</literal>.
1129 <term><varname>TransmitHashPolicy=
</varname></term>
1131 <para>Selects the transmit hash policy to use for slave
1132 selection in balance-xor,
802.3ad, and tlb modes. Possible
1134 <literal>layer2
</literal>,
1135 <literal>layer3+
4</literal>,
1136 <literal>layer2+
3</literal>,
1137 <literal>encap2+
3</literal>, and
1138 <literal>encap3+
4</literal>.
1144 <term><varname>LACPTransmitRate=
</varname></term>
1146 <para>Specifies the rate with which link partner transmits
1147 Link Aggregation Control Protocol Data Unit packets in
1148 802.3ad mode. Possible values are
<literal>slow
</literal>,
1149 which requests partner to transmit LACPDUs every
30 seconds,
1150 and
<literal>fast
</literal>, which requests partner to
1151 transmit LACPDUs every second. The default value is
1152 <literal>slow
</literal>.
</para>
1157 <term><varname>MIIMonitorSec=
</varname></term>
1159 <para>Specifies the frequency that Media Independent
1160 Interface link monitoring will occur. A value of zero
1161 disables MII link monitoring. This value is rounded down to
1162 the nearest millisecond. The default value is
0.
</para>
1167 <term><varname>UpDelaySec=
</varname></term>
1169 <para>Specifies the delay before a link is enabled after a
1170 link up status has been detected. This value is rounded down
1171 to a multiple of MIIMonitorSec. The default value is
1177 <term><varname>DownDelaySec=
</varname></term>
1179 <para>Specifies the delay before a link is disabled after a
1180 link down status has been detected. This value is rounded
1181 down to a multiple of MIIMonitorSec. The default value is
1187 <term><varname>LearnPacketIntervalSec=
</varname></term>
1189 <para>Specifies the number of seconds between instances where the bonding
1190 driver sends learning packets to each slave peer switch.
1191 The valid range is
1–
0x7fffffff; the default value is
1. This option
1192 has an effect only for the balance-tlb and balance-alb modes.
</para>
1197 <term><varname>AdSelect=
</varname></term>
1199 <para>Specifies the
802.3ad aggregation selection logic to use. Possible values are
1200 <literal>stable
</literal>,
1201 <literal>bandwidth
</literal> and
1202 <literal>count
</literal>.
1208 <term><varname>FailOverMACPolicy=
</varname></term>
1210 <para>Specifies whether the active-backup mode should set all slaves to
1211 the same MAC address at the time of enslavement or, when enabled, to perform special handling of the
1212 bond's MAC address in accordance with the selected policy. The default policy is none.
1214 <literal>none
</literal>,
1215 <literal>active
</literal> and
1216 <literal>follow
</literal>.
1222 <term><varname>ARPValidate=
</varname></term>
1224 <para>Specifies whether or not ARP probes and replies should be
1225 validated in any mode that supports ARP monitoring, or whether
1226 non-ARP traffic should be filtered (disregarded) for link
1227 monitoring purposes. Possible values are
1228 <literal>none
</literal>,
1229 <literal>active
</literal>,
1230 <literal>backup
</literal> and
1231 <literal>all
</literal>.
1237 <term><varname>ARPIntervalSec=
</varname></term>
1239 <para>Specifies the ARP link monitoring frequency in milliseconds.
1240 A value of
0 disables ARP monitoring. The default value is
0.
1246 <term><varname>ARPIPTargets=
</varname></term>
1248 <para>Specifies the IP addresses to use as ARP monitoring peers when
1249 ARPIntervalSec is greater than
0. These are the targets of the ARP request
1250 sent to determine the health of the link to the targets.
1251 Specify these values in IPv4 dotted decimal format. At least one IP
1252 address must be given for ARP monitoring to function. The
1253 maximum number of targets that can be specified is
16. The
1254 default value is no IP addresses.
1260 <term><varname>ARPAllTargets=
</varname></term>
1262 <para>Specifies the quantity of ARPIPTargets that must be reachable
1263 in order for the ARP monitor to consider a slave as being up.
1264 This option affects only active-backup mode for slaves with
1265 ARPValidate enabled. Possible values are
1266 <literal>any
</literal> and
1267 <literal>all
</literal>.
1273 <term><varname>PrimaryReselectPolicy=
</varname></term>
1275 <para>Specifies the reselection policy for the primary slave. This
1276 affects how the primary slave is chosen to become the active slave
1277 when failure of the active slave or recovery of the primary slave
1278 occurs. This option is designed to prevent flip-flopping between
1279 the primary slave and other slaves. Possible values are
1280 <literal>always
</literal>,
1281 <literal>better
</literal> and
1282 <literal>failure
</literal>.
1288 <term><varname>ResendIGMP=
</varname></term>
1290 <para>Specifies the number of IGMP membership reports to be issued after
1291 a failover event. One membership report is issued immediately after
1292 the failover, subsequent packets are sent in each
200ms interval.
1293 The valid range is
0–
255. Defaults to
1. A value of
0
1294 prevents the IGMP membership report from being issued in response
1295 to the failover event.
1301 <term><varname>PacketsPerSlave=
</varname></term>
1303 <para>Specify the number of packets to transmit through a slave before
1304 moving to the next one. When set to
0, then a slave is chosen at
1305 random. The valid range is
0–
65535. Defaults to
1. This option
1306 only has effect when in balance-rr mode.
1312 <term><varname>GratuitousARP=
</varname></term>
1314 <para>Specify the number of peer notifications (gratuitous ARPs and
1315 unsolicited IPv6 Neighbor Advertisements) to be issued after a
1316 failover event. As soon as the link is up on the new slave,
1317 a peer notification is sent on the bonding device and each
1318 VLAN sub-device. This is repeated at each link monitor interval
1319 (ARPIntervalSec or MIIMonitorSec, whichever is active) if the number is
1320 greater than
1. The valid range is
0–
255. The default value is
1.
1321 These options affect only the active-backup mode.
1327 <term><varname>AllSlavesActive=
</varname></term>
1329 <para>A boolean. Specifies that duplicate frames (received on inactive ports)
1330 should be dropped when false, or delivered when true. Normally, bonding will drop
1331 duplicate frames (received on inactive ports), which is desirable for
1332 most users. But there are some times it is nice to allow duplicate
1333 frames to be delivered. The default value is false (drop duplicate frames
1334 received on inactive ports).
1340 <term><varname>MinLinks=
</varname></term>
1342 <para>Specifies the minimum number of links that must be active before
1343 asserting carrier. The default value is
0.
1349 <para>For more detail information see
1350 <ulink url=
"https://www.kernel.org/doc/Documentation/networking/bonding.txt">
1351 Linux Ethernet Bonding Driver HOWTO
</ulink></para>
1356 <title>Example
</title>
1358 <title>/etc/systemd/network/
25-bridge.netdev
</title>
1360 <programlisting>[NetDev]
1362 Kind=bridge
</programlisting>
1366 <title>/etc/systemd/network/
25-vlan1.netdev
</title>
1368 <programlisting>[Match]
1376 Id=
1</programlisting>
1379 <title>/etc/systemd/network/
25-ipip.netdev
</title>
1380 <programlisting>[NetDev]
1386 Local=
192.168.223.238
1387 Remote=
192.169.224.239
1388 TTL=
64</programlisting>
1391 <title>/etc/systemd/network/
25-tap.netdev
</title>
1392 <programlisting>[NetDev]
1398 PacketInfo=true
</programlisting> </example>
1401 <title>/etc/systemd/network/
25-sit.netdev
</title>
1402 <programlisting>[NetDev]
1409 Remote=
10.65.223.239</programlisting>
1413 <title>/etc/systemd/network/
25-gre.netdev
</title>
1414 <programlisting>[NetDev]
1421 Remote=
10.65.223.239</programlisting>
1425 <title>/etc/systemd/network/
25-vti.netdev
</title>
1427 <programlisting>[NetDev]
1434 Remote=
10.65.223.239</programlisting>
1438 <title>/etc/systemd/network/
25-veth.netdev
</title>
1439 <programlisting>[NetDev]
1444 Name=veth-peer
</programlisting>
1448 <title>/etc/systemd/network/
25-bond.netdev
</title>
1449 <programlisting>[NetDev]
1455 TransmitHashPolicy=layer3+
4
1457 LACPTransmitRate=fast
1462 <title>/etc/systemd/network/
25-dummy.netdev
</title>
1463 <programlisting>[NetDev]
1466 MACAddress=
12:
34:
56:
78:
9a:bc
</programlisting>
1469 <title>/etc/systemd/network/
25-vrf.netdev
</title>
1470 <para>Create a VRF interface with table
42.
</para>
1471 <programlisting>[NetDev]
1476 Table=
42</programlisting>
1480 <title>/etc/systemd/network/
25-macvtap.netdev
</title>
1481 <para>Create a MacVTap device.
</para>
1482 <programlisting>[NetDev]
1488 <title>/etc/systemd/network/
25-wireguard.netdev
</title>
1489 <programlisting>[NetDev]
1494 PrivateKey=EEGlnEPYJV//kbvvIqxKkQwOiS+UENyPncC4bF46ong=
1498 PublicKey=RDf+LSpeEre7YEIKaxg+wbpsNV7du+ktR99uBEtIiCA=
1499 AllowedIPs=fd31:bf08:
57cb::/
48,
192.168.26.0/
24
1500 Endpoint=wireguard.example.com:
51820</programlisting>
1504 <title>See Also
</title>
1506 <citerefentry><refentrytitle>systemd
</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
1507 <citerefentry><refentrytitle>systemd-networkd
</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
1508 <citerefentry><refentrytitle>systemd.link
</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
1509 <citerefentry><refentrytitle>systemd.network
</refentrytitle><manvolnum>5</manvolnum></citerefentry>