xref: /freebsd/sbin/routed/routed.8 (revision 1669d8afc64812c8d2d1d147ae1fd42ff441e1b1)
1.\"   $Revision: 2.26 $
2.\"
3.\" Copyright (c) 1983, 1991, 1993
4.\"	The Regents of the University of California.  All rights reserved.
5.\"
6.\" Redistribution and use in source and binary forms, with or without
7.\" modification, are permitted provided that the following conditions
8.\" are met:
9.\" 1. Redistributions of source code must retain the above copyright
10.\"    notice, this list of conditions and the following disclaimer.
11.\" 2. Redistributions in binary form must reproduce the above copyright
12.\"    notice, this list of conditions and the following disclaimer in the
13.\"    documentation and/or other materials provided with the distribution.
14.\" 4. Neither the name of the University nor the names of its contributors
15.\"    may be used to endorse or promote products derived from this software
16.\"    without specific prior written permission.
17.\"
18.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
19.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
22.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28.\" SUCH DAMAGE.
29.\"
30.\"     @(#)routed.8	8.2 (Berkeley) 12/11/93
31.\" $FreeBSD$
32.\"
33.Dd June 1, 1996
34.Dt ROUTED 8
35.Os
36.Sh NAME
37.Nm routed ,
38.Nm rdisc
39.Nd network RIP and router discovery routing daemon
40.Sh SYNOPSIS
41.Nm
42.Op Fl sqdghmpAtv
43.Op Fl T Ar tracefile
44.Oo
45.Fl F
46.Ar net Ns Op /mask Ns Op ,metric
47.Oc
48.Op Fl P Ar parms
49.Sh DESCRIPTION
50The
51.Nm
52utility is a daemon invoked at boot time to manage the network
53routing tables.
54It uses Routing Information Protocol, RIPv1 (RFC\ 1058),
55RIPv2 (RFC\ 1723),
56and Internet Router Discovery Protocol (RFC 1256)
57to maintain the kernel routing table.
58The RIPv1 protocol is based on the reference
59.Bx 4.3
60daemon.
61.Pp
62It listens on the
63.Xr udp 4
64socket for the
65.Xr route 8
66service (see
67.Xr services 5 )
68for Routing Information Protocol packets.
69It also sends and receives multicast Router Discovery ICMP messages.
70If the host is a router,
71.Nm
72periodically supplies copies
73of its routing tables to any directly connected hosts and networks.
74It also advertises or solicits default routes using Router Discovery
75ICMP messages.
76.Pp
77When started (or when a network interface is later turned on),
78.Nm
79uses an AF_ROUTE address family facility to find those
80directly connected interfaces configured into the
81system and marked "up".
82It adds necessary routes for the interfaces
83to the kernel routing table.
84Soon after being first started, and provided there is at least one
85interface on which RIP has not been disabled,
86.Nm
87deletes all pre-existing
88non-static routes in kernel table.
89Static routes in the kernel table are preserved and
90included in RIP responses if they have a valid RIP metric
91(see
92.Xr route 8 ) .
93.Pp
94If more than one interface is present (not counting the loopback interface),
95it is assumed that the host should forward packets among the
96connected networks.
97After transmitting a RIP
98.Em request
99and
100Router Discovery Advertisements or Solicitations on a new interface,
101the daemon enters a loop, listening for
102RIP request and response and Router Discovery packets from other hosts.
103.Pp
104When a
105.Em request
106packet is received,
107.Nm
108formulates a reply based on the information maintained in its
109internal tables.
110The
111.Em response
112packet generated contains a list of known routes, each marked
113with a "hop count" metric (a count of 16 or greater is
114considered "infinite").
115The advertised metric for a route reflects the metrics associated
116with interfaces
117(see
118.Xr ifconfig 8 )
119though which it is received and sent,
120so setting the metric on an interface
121is an effective way to steer traffic.
122See also
123.Cm adj_inmetric
124and
125.Cm adj_outmetric
126parameters below.
127.Pp
128Responses do not include routes with a first hop on the requesting
129network to implement in part
130.Em split-horizon .
131Requests from query programs
132such as
133.Xr rtquery 8
134are answered with the complete table.
135.Pp
136The routing table maintained by the daemon
137includes space for several gateways for each destination
138to speed recovery from a failing router.
139RIP
140.Em response
141packets received are used to update the routing tables provided they are
142from one of the several currently recognized gateways or
143advertise a better metric than at least one of the existing
144gateways.
145.Pp
146When an update is applied,
147.Nm
148records the change in its own tables and updates the kernel routing table
149if the best route to the destination changes.
150The change in the kernel routing table is reflected in the next batch of
151.Em response
152packets sent.
153If the next response is not scheduled for a while, a
154.Em flash update
155response containing only recently changed routes is sent.
156.Pp
157In addition to processing incoming packets,
158.Nm
159also periodically checks the routing table entries.
160If an entry has not been updated for 3 minutes, the entry's metric
161is set to infinity and marked for deletion.
162Deletions are delayed until the route has been advertised with
163an infinite metric to ensure the invalidation
164is propagated throughout the local internet.
165This is a form of
166.Em poison reverse .
167.Pp
168Routes in the kernel table that are added or changed as a result
169of ICMP Redirect messages are deleted after a while to minimize
170.Em black-holes .
171When a TCP connection suffers a timeout,
172the kernel tells
173.Nm ,
174which deletes all redirected routes
175through the gateway involved, advances the age of all RIP routes through
176the gateway to allow an alternate to be chosen, and advances of the
177age of any relevant Router Discovery Protocol default routes.
178.Pp
179Hosts acting as internetwork routers gratuitously supply their
180routing tables every 30 seconds to all directly connected hosts
181and networks.
182These RIP responses are sent to the broadcast address on nets that support
183broadcasting,
184to the destination address on point-to-point links, and to the router's
185own address on other networks.
186If RIPv2 is enabled, multicast packets are sent on interfaces that
187support multicasting.
188.Pp
189If no response is received on a remote interface, if there are errors
190while sending responses,
191or if there are more errors than input or output (see
192.Xr netstat 1 ) ,
193then the cable or some other part of the interface is assumed to be
194disconnected or broken, and routes are adjusted appropriately.
195.Pp
196The
197.Em Internet Router Discovery Protocol
198is handled similarly.
199When the daemon is supplying RIP routes, it also listens for
200Router Discovery Solicitations and sends Advertisements.
201When it is quiet and listening to other RIP routers, it
202sends Solicitations and listens for Advertisements.
203If it receives
204a good Advertisement and it is not multi-homed,
205it stops listening for broadcast or multicast RIP responses.
206It tracks several advertising routers to speed recovery when the
207currently chosen router dies.
208If all discovered routers disappear,
209the daemon resumes listening to RIP responses.
210It continues listening to RIP while using Router Discovery
211if multi-homed to ensure all interfaces are used.
212.Pp
213The Router Discovery standard requires that advertisements
214have a default "lifetime" of 30 minutes.
215That means should
216something happen, a client can be without a good route for
21730 minutes.
218It is a good idea to reduce the default to 45
219seconds using
220.Fl P Cm rdisc_interval=45
221on the command line or
222.Cm rdisc_interval=45
223in the
224.Pa /etc/gateways
225file.
226.Pp
227While using Router Discovery (which happens by default when
228the system has a single network interface and a Router Discover Advertisement
229is received), there is a single default route and a variable number of
230redirected host routes in the kernel table.
231On a host with more than one network interface,
232this default route will be via only one of the interfaces.
233Thus, multi-homed hosts running with
234.Fl q
235might need
236.Cm no_rdisc
237described below.
238.Pp
239See the
240.Cm pm_rdisc
241facility described below to support "legacy" systems
242that can handle neither RIPv2 nor Router Discovery.
243.Pp
244By default, neither Router Discovery advertisements nor solicitations
245are sent over point to point links (e.g.\& PPP).
246The netmask associated with point-to-point links (such as SLIP
247or PPP, with the IFF_POINTOPOINT flag) is used by
248.Nm
249to infer the netmask used by the remote system when RIPv1 is used.
250.Pp
251The following options are available:
252.Bl -tag -width indent
253.It Fl s
254force
255.Nm
256to supply routing information.
257This is the default if multiple network interfaces are present on which
258RIP or Router Discovery have not been disabled, and if the kernel switch
259ipforwarding=1.
260.It Fl q
261is the opposite of the
262.Fl s
263option.
264This is the default when only one interface is present.
265With this explicit option, the daemon is always in "quiet-mode" for RIP
266and does not supply routing information to other computers.
267.It Fl d
268do not run in the background.
269This option is meant for interactive use.
270.It Fl g
271used on internetwork routers to offer a route
272to the "default" destination.
273It is equivalent to
274.Fl F
275.Cm 0/0,1
276and is present mostly for historical reasons.
277A better choice is
278.Fl P Cm pm_rdisc
279on the command line or
280.Cm pm_rdisc
281in the
282.Pa /etc/gateways
283file,
284since a larger metric
285will be used, reducing the spread of the potentially dangerous
286default route.
287This is typically used on a gateway to the Internet,
288or on a gateway that uses another routing protocol whose routes
289are not reported to other local routers.
290Notice that because a metric of 1 is used, this feature is
291dangerous.
292It is more commonly accidentally used to create chaos with a
293routing loop than to solve problems.
294.It Fl h
295cause host or point-to-point routes to not be advertised,
296provided there is a network route going the same direction.
297That is a limited kind of aggregation.
298This option is useful on gateways to Ethernets that have other gateway
299machines connected with point-to-point links such as SLIP.
300.It Fl m
301cause the machine to advertise a host or point-to-point route to
302its primary interface.
303It is useful on multi-homed machines such as NFS servers.
304This option should not be used except when the cost of
305the host routes it generates is justified by the popularity of
306the server.
307It is effective only when the machine is supplying
308routing information, because there is more than one interface.
309The
310.Fl m
311option overrides the
312.Fl q
313option to the limited extent of advertising the host route.
314.It Fl A
315do not ignore RIPv2 authentication if we do not care about RIPv2
316authentication.
317This option is required for conformance with RFC 1723.
318However, it makes no sense and breaks using RIP as a discovery protocol
319to ignore all RIPv2 packets that carry authentication when this machine
320does not care about authentication.
321.It Fl t
322increase the debugging level, which causes more information to be logged
323on the tracefile specified with
324.Fl T
325or standard out.
326The debugging level can be increased or decreased
327with the
328.Em SIGUSR1
329or
330.Em SIGUSR2
331signals or with the
332.Xr rtquery 8
333command.
334.It Fl T Ar tracefile
335increases the debugging level to at least 1 and
336causes debugging information to be appended to the trace file.
337Note that because of security concerns, it is wisest to not run
338.Nm
339routinely with tracing directed to a file.
340.It Fl v
341display and logs the version of daemon.
342.It Fl F Ar net[/mask][,metric]
343minimize routes in transmissions via interfaces with addresses that match
344.Em net/mask ,
345and synthesizes a default route to this machine with the
346.Em metric .
347The intent is to reduce RIP traffic on slow, point-to-point links
348such as PPP links by replacing many large UDP packets of RIP information
349with a single, small packet containing a "fake" default route.
350If
351.Em metric
352is absent, a value of 14 is assumed to limit
353the spread of the "fake" default route.
354This is a dangerous feature that when used carelessly can cause routing
355loops.
356Notice also that more than one interface can match the specified network
357number and mask.
358See also
359.Fl g .
360.It Fl P Ar parms
361is equivalent to adding the parameter
362line
363.Em parms
364to the
365.Pa /etc/gateways
366file.
367.El
368.Pp
369Any other argument supplied is interpreted as the name
370of a file in which the actions of
371.Nm
372should be logged.
373It is better to use
374.Fl T
375instead of
376appending the name of the trace file to the command.
377.Pp
378The
379.Nm
380utility also supports the notion of
381"distant"
382.Em passive
383or
384.Em active
385gateways.
386When
387.Nm
388is started, it reads the file
389.Pa /etc/gateways
390to find such distant gateways which may not be located using
391only information from a routing socket, to discover if some
392of the local gateways are
393.Em passive ,
394and to obtain other parameters.
395Gateways specified in this manner should be marked passive
396if they are not expected to exchange routing information,
397while gateways marked active
398should be willing to exchange RIP packets.
399Routes through
400.Em passive
401gateways are installed in the
402kernel's routing tables once upon startup and are not included in
403transmitted RIP responses.
404.Pp
405Distant active gateways are treated like network interfaces.
406RIP responses are sent
407to the distant
408.Em active
409gateway.
410If no responses are received, the associated route is deleted from
411the kernel table and RIP responses advertised via other interfaces.
412If the distant gateway resumes sending RIP responses, the associated
413route is restored.
414.Pp
415Such gateways can be useful on media that do not support broadcasts
416or multicasts but otherwise act like classic shared media like
417Ethernets such as some ATM networks.
418One can list all RIP routers reachable on the HIPPI or ATM network in
419.Pa /etc/gateways
420with a series of
421"host" lines.
422Note that it is usually desirable to use RIPv2 in such situations
423to avoid generating lists of inferred host routes.
424.Pp
425Gateways marked
426.Em external
427are also passive, but are not placed in the kernel
428routing table nor are they included in routing updates.
429The function of external entries is to indicate
430that another routing process
431will install such a route if necessary,
432and that other routes to that destination should not be installed
433by
434.Nm .
435Such entries are only required when both routers may learn of routes
436to the same destination.
437.Pp
438The
439.Pa /etc/gateways
440file is comprised of a series of lines, each in
441one of the following two formats or consist of parameters described later.
442Blank lines and lines starting with '#' are comments.
443.Pp
444.Bd -ragged
445.Cm net
446.Ar Nname[/mask]
447.Cm gateway
448.Ar Gname
449.Cm metric
450.Ar value
451.Pf < Cm passive No \&|
452.Cm active No \&|
453.Cm extern Ns >
454.Ed
455.Bd -ragged
456.Cm host
457.Ar Hname
458.Cm gateway
459.Ar Gname
460.Cm metric
461.Ar value
462.Pf < Cm passive No \&|
463.Cm active No \&|
464.Cm extern Ns >
465.Ed
466.Pp
467.Ar Nname
468or
469.Ar Hname
470is the name of the destination network or host.
471It may be a symbolic network name or an Internet address
472specified in "dot" notation (see
473.Xr inet 3 ) .
474(If it is a name, then it must either be defined in
475.Pa /etc/networks
476or
477.Pa /etc/hosts ,
478or
479.Xr named 8 ,
480must have been started before
481.Nm . )
482.Pp
483.Ar Mask
484is an optional number between 1 and 32 indicating the netmask associated
485with
486.Ar Nname .
487.Pp
488.Ar Gname
489is the name or address of the gateway to which RIP responses should
490be forwarded.
491.Pp
492.Ar Value
493is the hop count to the destination host or network.
494.Pp
495.Cm Host Ar hname
496is equivalent to
497.Cm net Ar nname/32 .
498.Pp
499One of the keywords
500.Cm passive ,
501.Cm active
502or
503.Cm external
504must be present to indicate whether the gateway should be treated as
505.Cm passive
506or
507.Cm active
508(as described above),
509or whether the gateway is
510.Cm external
511to the scope of the RIP protocol.
512.Pp
513As can be seen when debugging is turned on with
514.Fl t ,
515such lines create pseudo-interfaces.
516To set parameters for remote or external interfaces,
517a line starting with
518.Cm if=alias(Hname) ,
519.Cm if=remote(Hname) ,
520etc.\& should be used.
521.Ss Parameters
522Lines that start with neither "net" nor "host" must consist of one
523or more of the following parameter settings, separated by commas or
524blanks:
525.Bl -tag -width indent
526.It Cm if Ns = Ns Ar ifname
527indicates that the other parameters on the line apply to the interface
528name
529.Ar ifname .
530.It Cm subnet Ns = Ns Ar nname Ns Oo / Ns Ar mask Oc Ns Op , Ns Ar metric
531advertises a route to network
532.Ar nname
533with mask
534.Ar mask
535and the supplied metric (default 1).
536This is useful for filling "holes" in CIDR allocations.
537This parameter must appear by itself on a line.
538The network number must specify a full, 32-bit value, as in 192.0.2.0
539instead of 192.0.2.
540.Pp
541Do not use this feature unless necessary.
542It is dangerous.
543.It Cm ripv1_mask Ns = Ns Ar nname Ns / Ns Ar mask1 , Ns Ar mask2
544specifies that netmask of the network of which
545.Ar nname Ns / Ns Ar mask1
546is
547a subnet should be
548.Ar mask2 .
549For example,
550.Dq Li ripv1_mask=192.0.2.16/28,27
551marks 192.0.2.16/28
552as a subnet of 192.0.2.0/27 instead of 192.0.2.0/24.
553It is better to turn on RIPv2 instead of using this facility, for example
554with
555.Cm ripv2_out .
556.It Cm passwd Ns = Ns Ar XXX[|KeyID[start|stop]]
557specifies a RIPv2 cleartext password that will be included on
558all RIPv2 responses sent, and checked on all RIPv2 responses received.
559Any blanks, tab characters, commas, or '#', '|', or NULL characters in the
560password must be escaped with a backslash (\\).
561The common escape sequences \\n, \\r, \\t, \\b, and \\xxx have their
562usual meanings.
563The
564.Cm KeyID
565must be unique but is ignored for cleartext passwords.
566If present,
567.Cm start
568and
569.Cm stop
570are timestamps in the form year/month/day@hour:minute.
571They specify when the password is valid.
572The valid password with the most future is used on output packets, unless
573all passwords have expired, in which case the password that expired most
574recently is used, or unless no passwords are valid yet, in which case
575no password is output.
576Incoming packets can carry any password that is valid, will
577be valid within the next 24 hours, or that was valid within the preceding
57824 hours.
579To protect the secrets, the passwd settings are valid only in the
580.Em /etc/gateways
581file and only when that file is readable only by UID 0.
582.It Cm md5_passwd Ns \&= Ns Ar XXX|KeyID[start|stop]
583specifies a RIPv2 MD5 password.
584Except that a
585.Cm KeyID
586is required, this keyword is similar to
587.Cm passwd .
588.It Cm no_ag
589turns off aggregation of subnets in RIPv1 and RIPv2 responses.
590.It Cm no_super_ag
591turns off aggregation of networks into supernets in RIPv2 responses.
592.It Cm passive
593marks the interface to not be advertised in updates sent via other
594interfaces, and turns off all RIP and router discovery through the interface.
595.It Cm no_rip
596disables all RIP processing on the specified interface.
597If no interfaces are allowed to process RIP packets,
598.Nm
599acts purely as a router discovery daemon.
600.Pp
601Note that turning off RIP without explicitly turning on router
602discovery advertisements with
603.Cm rdisc_adv
604or
605.Fl s
606causes
607.Nm
608to act as a client router discovery daemon, not advertising.
609.It Cm no_rip_mcast
610causes RIPv2 packets to be broadcast instead of multicast.
611.It Cm no_rip_out
612causes no RIP updates to be sent.
613.It Cm no_ripv1_in
614causes RIPv1 received responses to be ignored.
615.It Cm no_ripv2_in
616causes RIPv2 received responses to be ignored.
617.It Cm ripv2_out
618turns on RIPv2 output and causes RIPv2 advertisements to be
619multicast when possible.
620.It Cm ripv2
621is equivalent to
622.Cm no_ripv1_in
623and
624.Cm no_ripv1_out .
625This enables RIPv2.
626.It Cm no_rdisc
627disables the Internet Router Discovery Protocol.
628.It Cm no_solicit
629disables the transmission of Router Discovery Solicitations.
630.It Cm send_solicit
631specifies that Router Discovery solicitations should be sent,
632even on point-to-point links,
633which by default only listen to Router Discovery messages.
634.It Cm no_rdisc_adv
635disables the transmission of Router Discovery Advertisements.
636.It Cm rdisc_adv
637specifies that Router Discovery Advertisements should be sent,
638even on point-to-point links,
639which by default only listen to Router Discovery messages.
640.It Cm bcast_rdisc
641specifies that Router Discovery packets should be broadcast instead of
642multicast.
643.It Cm rdisc_pref Ns \&= Ns Ar N
644sets the preference in Router Discovery Advertisements to the optionally
645signed integer
646.Ar N .
647The default preference is 0.
648Default routes with smaller or more negative preferences are preferred by
649clients.
650.It Cm rdisc_interval Ns \&= Ns Ar N
651sets the nominal interval with which Router Discovery Advertisements
652are transmitted to N seconds and their lifetime to 3*N.
653.It Cm fake_default Ns \&= Ns Ar metric
654has an identical effect to
655.Fl F Ar net[/mask][=metric]
656with the network and mask coming from the specified interface.
657.It Cm pm_rdisc
658is similar to
659.Cm fake_default .
660When RIPv2 routes are multicast, so that RIPv1 listeners cannot
661receive them, this feature causes a RIPv1 default route to be
662broadcast to RIPv1 listeners.
663Unless modified with
664.Cm fake_default ,
665the default route is broadcast with a metric of 14.
666That serves as a "poor man's router discovery" protocol.
667.It Cm adj_inmetric Ns \&= Ns Ar delta
668adjusts the hop count or metric of received RIP routes by
669.Ar delta .
670The metric of every received RIP route is increased by the sum
671of two values associated with the interface.
672One is the adj_inmetric value and the other is the interface
673metric set with
674.Xr ifconfig 8 .
675.It Cm adj_outmetric Ns \&= Ns Ar delta
676adjusts the hop count or metric of advertised RIP routes by
677.Ar delta .
678The metric of every received RIP route is increased by the metric
679associated with the interface by which it was received, or by 1 if
680the interface does not have a non-zero metric.
681The metric of the received route is then increased by the
682adj_outmetric associated with the interface.
683Every advertised route is increased by a total of four
684values,
685the metric set for the interface by which it was received with
686.Xr ifconfig 8 ,
687the
688.Cm adj_inmetric Ar delta
689of the receiving interface,
690the metric set for the interface by which it is transmitted with
691.Xr ifconfig 8 ,
692and the
693.Cm adj_outmetric Ar delta
694of the transmitting interface.
695.It Cm trust_gateway Ns \&= Ns Ar rname[|net1/mask1|net2/mask2|...]
696causes RIP packets from router
697.Ar rname
698and other routers named in other
699.Cm trust_gateway
700keywords to be accepted, and packets from other routers to be ignored.
701If networks are specified, then routes to other networks will be ignored
702from that router.
703.It Cm redirect_ok
704allows the kernel to listen ICMP Redirect messages when the system is acting
705as a router and forwarding packets.
706Otherwise, ICMP Redirect messages are overridden and deleted when the
707system is acting as a router.
708.El
709.Sh FILES
710.Bl -tag -width /etc/gateways -compact
711.It Pa /etc/gateways
712for distant gateways
713.El
714.Sh SEE ALSO
715.Xr icmp 4 ,
716.Xr udp 4 ,
717.Xr rtquery 8
718.Rs
719.%T Internet Transport Protocols
720.%R XSIS 028112
721.%Q Xerox System Integration Standard
722.Re
723.Sh HISTORY
724The
725.Nm
726utility appeared in
727.Bx 4.2 .
728.\"  LocalWords:  loopback ICMP rtquery ifconfig multicasting Solicitations RIPv
729.\"  LocalWords:  netstat rdisc
730.Sh BUGS
731It does not always detect unidirectional failures in network interfaces,
732for example, when the output side fails.
733