xref: /freebsd/sbin/ipfw/ipfw.8 (revision a25896ca1270e25b657ceaa8d47d5699515f5c25)
1.\"
2.\" $FreeBSD$
3.\"
4.Dd October 21, 2018
5.Dt IPFW 8
6.Os
7.Sh NAME
8.Nm ipfw
9.Nd User interface for firewall, traffic shaper, packet scheduler,
10in-kernel NAT.
11.Sh SYNOPSIS
12.Ss FIREWALL CONFIGURATION
13.Nm
14.Op Fl cq
15.Cm add
16.Ar rule
17.Nm
18.Op Fl acdefnNStT
19.Op Cm set Ar N
20.Brq Cm list | show
21.Op Ar rule | first-last ...
22.Nm
23.Op Fl f | q
24.Op Cm set Ar N
25.Cm flush
26.Nm
27.Op Fl q
28.Op Cm set Ar N
29.Brq Cm delete | zero | resetlog
30.Op Ar number ...
31.Pp
32.Nm
33.Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
34.Nm
35.Cm set move
36.Op Cm rule
37.Ar number Cm to Ar number
38.Nm
39.Cm set swap Ar number number
40.Nm
41.Cm set show
42.Ss SYSCTL SHORTCUTS
43.Nm
44.Cm enable
45.Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
46.Nm
47.Cm disable
48.Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
49.Ss LOOKUP TABLES
50.Nm
51.Oo Cm set Ar N Oc Cm table Ar name Cm create Ar create-options
52.Nm
53.Oo Cm set Ar N Oc Cm table
54.Brq Ar name | all
55.Cm destroy
56.Nm
57.Oo Cm set Ar N Oc Cm table Ar name Cm modify Ar modify-options
58.Nm
59.Oo Cm set Ar N Oc Cm table Ar name Cm swap Ar name
60.Nm
61.Oo Cm set Ar N Oc Cm table Ar name Cm add Ar table-key Op Ar value
62.Nm
63.Oo Cm set Ar N Oc Cm table Ar name Cm add Op Ar table-key Ar value ...
64.Nm
65.Oo Cm set Ar N Oc Cm table Ar name Cm atomic add Op Ar table-key Ar value ...
66.Nm
67.Oo Cm set Ar N Oc Cm table Ar name Cm delete Op Ar table-key ...
68.Nm
69.Oo Cm set Ar N Oc Cm table Ar name Cm lookup Ar addr
70.Nm
71.Oo Cm set Ar N Oc Cm table Ar name Cm lock
72.Nm
73.Oo Cm set Ar N Oc Cm table Ar name Cm unlock
74.Nm
75.Oo Cm set Ar N Oc Cm table
76.Brq Ar name | all
77.Cm list
78.Nm
79.Oo Cm set Ar N Oc Cm table
80.Brq Ar name | all
81.Cm info
82.Nm
83.Oo Cm set Ar N Oc Cm table
84.Brq Ar name | all
85.Cm detail
86.Nm
87.Oo Cm set Ar N Oc Cm table
88.Brq Ar name | all
89.Cm flush
90.Ss DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)
91.Nm
92.Brq Cm pipe | queue | sched
93.Ar number
94.Cm config
95.Ar config-options
96.Nm
97.Op Fl s Op Ar field
98.Brq Cm pipe | queue | sched
99.Brq Cm delete | list | show
100.Op Ar number ...
101.Ss IN-KERNEL NAT
102.Nm
103.Op Fl q
104.Cm nat
105.Ar number
106.Cm config
107.Ar config-options
108.Pp
109.Nm
110.Op Fl cfnNqS
111.Oo
112.Fl p Ar preproc
113.Oo
114.Ar preproc-flags
115.Oc
116.Oc
117.Ar pathname
118.Ss STATEFUL IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
119.Nm
120.Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm create Ar create-options
121.Nm
122.Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm config Ar config-options
123.Nm
124.Oo Cm set Ar N Oc Cm nat64lsn
125.Brq Ar name | all
126.Brq Cm list | show
127.Op Cm states
128.Nm
129.Oo Cm set Ar N Oc Cm nat64lsn
130.Brq Ar name | all
131.Cm destroy
132.Nm
133.Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm stats Op Cm reset
134.Ss STATELESS IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
135.Nm
136.Oo Cm set Ar N Oc Cm nat64stl Ar name Cm create Ar create-options
137.Nm
138.Oo Cm set Ar N Oc Cm nat64stl Ar name Cm config Ar config-options
139.Nm
140.Oo Cm set Ar N Oc Cm nat64stl
141.Brq Ar name | all
142.Brq Cm list | show
143.Nm
144.Oo Cm set Ar N Oc Cm nat64stl
145.Brq Ar name | all
146.Cm destroy
147.Nm
148.Oo Cm set Ar N Oc Cm nat64stl Ar name Cm stats Op Cm reset
149.Ss IPv6-to-IPv6 NETWORK PREFIX TRANSLATION
150.Nm
151.Oo Cm set Ar N Oc Cm nptv6 Ar name Cm create Ar create-options
152.Nm
153.Oo Cm set Ar N Oc Cm nptv6
154.Brq Ar name | all
155.Brq Cm list | show
156.Nm
157.Oo Cm set Ar N Oc Cm nptv6
158.Brq Ar name | all
159.Cm destroy
160.Nm
161.Oo Cm set Ar N Oc Cm nptv6 Ar name Cm stats Op Cm reset
162.Ss INTERNAL DIAGNOSTICS
163.Nm
164.Cm internal iflist
165.Nm
166.Cm internal talist
167.Nm
168.Cm internal vlist
169.Sh DESCRIPTION
170The
171.Nm
172utility is the user interface for controlling the
173.Xr ipfw 4
174firewall, the
175.Xr dummynet 4
176traffic shaper/packet scheduler, and the
177in-kernel NAT services.
178.Pp
179A firewall configuration, or
180.Em ruleset ,
181is made of a list of
182.Em rules
183numbered from 1 to 65535.
184Packets are passed to the firewall
185from a number of different places in the protocol stack
186(depending on the source and destination of the packet,
187it is possible for the firewall to be
188invoked multiple times on the same packet).
189The packet passed to the firewall is compared
190against each of the rules in the
191.Em ruleset ,
192in rule-number order
193(multiple rules with the same number are permitted, in which case
194they are processed in order of insertion).
195When a match is found, the action corresponding to the
196matching rule is performed.
197.Pp
198Depending on the action and certain system settings, packets
199can be reinjected into the firewall at some rule after the
200matching one for further processing.
201.Pp
202A ruleset always includes a
203.Em default
204rule (numbered 65535) which cannot be modified or deleted,
205and matches all packets.
206The action associated with the
207.Em default
208rule can be either
209.Cm deny
210or
211.Cm allow
212depending on how the kernel is configured.
213.Pp
214If the ruleset includes one or more rules with the
215.Cm keep-state ,
216.Cm record-state ,
217.Cm limit
218or
219.Cm set-limit
220option,
221the firewall will have a
222.Em stateful
223behaviour, i.e., upon a match it will create
224.Em dynamic rules ,
225i.e., rules that match packets with the same 5-tuple
226(protocol, source and destination addresses and ports)
227as the packet which caused their creation.
228Dynamic rules, which have a limited lifetime, are checked
229at the first occurrence of a
230.Cm check-state ,
231.Cm keep-state
232or
233.Cm limit
234rule, and are typically used to open the firewall on-demand to
235legitimate traffic only.
236Please, note, that
237.Cm keep-state
238amd
239.Cm limit
240imply implicit
241.Cm check-state
242for all packets (not only these matched by the rule) but
243.Cm record-state
244and
245.Cm set-limit
246have no implicit
247.Cm check-state .
248See the
249.Sx STATEFUL FIREWALL
250and
251.Sx EXAMPLES
252Sections below for more information on the stateful behaviour of
253.Nm .
254.Pp
255All rules (including dynamic ones) have a few associated counters:
256a packet count, a byte count, a log count and a timestamp
257indicating the time of the last match.
258Counters can be displayed or reset with
259.Nm
260commands.
261.Pp
262Each rule belongs to one of 32 different
263.Em sets
264, and there are
265.Nm
266commands to atomically manipulate sets, such as enable,
267disable, swap sets, move all rules in a set to another
268one, delete all rules in a set.
269These can be useful to
270install temporary configurations, or to test them.
271See Section
272.Sx SETS OF RULES
273for more information on
274.Em sets .
275.Pp
276Rules can be added with the
277.Cm add
278command; deleted individually or in groups with the
279.Cm delete
280command, and globally (except those in set 31) with the
281.Cm flush
282command; displayed, optionally with the content of the
283counters, using the
284.Cm show
285and
286.Cm list
287commands.
288Finally, counters can be reset with the
289.Cm zero
290and
291.Cm resetlog
292commands.
293.Pp
294.Ss COMMAND OPTIONS
295The following general options are available when invoking
296.Nm :
297.Bl -tag -width indent
298.It Fl a
299Show counter values when listing rules.
300The
301.Cm show
302command implies this option.
303.It Fl b
304Only show the action and the comment, not the body of a rule.
305Implies
306.Fl c .
307.It Fl c
308When entering or showing rules, print them in compact form,
309i.e., omitting the "ip from any to any" string
310when this does not carry any additional information.
311.It Fl d
312When listing, show dynamic rules in addition to static ones.
313.It Fl e
314When listing and
315.Fl d
316is specified, also show expired dynamic rules.
317.It Fl f
318Run without prompting for confirmation for commands that can cause problems if misused,
319i.e.,
320.Cm flush .
321If there is no tty associated with the process, this is implied.
322The
323.Cm delete
324command with this flag ignores possible errors,
325i.e., nonexistent rule number.
326And for batched commands execution continues with the next command.
327.It Fl i
328When listing a table (see the
329.Sx LOOKUP TABLES
330section below for more information on lookup tables), format values
331as IP addresses.
332By default, values are shown as integers.
333.It Fl n
334Only check syntax of the command strings, without actually passing
335them to the kernel.
336.It Fl N
337Try to resolve addresses and service names in output.
338.It Fl q
339Be quiet when executing the
340.Cm add ,
341.Cm nat ,
342.Cm zero ,
343.Cm resetlog
344or
345.Cm flush
346commands;
347(implies
348.Fl f ) .
349This is useful when updating rulesets by executing multiple
350.Nm
351commands in a script
352(e.g.,
353.Ql sh\ /etc/rc.firewall ) ,
354or by processing a file with many
355.Nm
356rules across a remote login session.
357It also stops a table add or delete
358from failing if the entry already exists or is not present.
359.Pp
360The reason why this option may be important is that
361for some of these actions,
362.Nm
363may print a message; if the action results in blocking the
364traffic to the remote client,
365the remote login session will be closed
366and the rest of the ruleset will not be processed.
367Access to the console would then be required to recover.
368.It Fl S
369When listing rules, show the
370.Em set
371each rule belongs to.
372If this flag is not specified, disabled rules will not be
373listed.
374.It Fl s Op Ar field
375When listing pipes, sort according to one of the four
376counters (total or current packets or bytes).
377.It Fl t
378When listing, show last match timestamp converted with ctime().
379.It Fl T
380When listing, show last match timestamp as seconds from the epoch.
381This form can be more convenient for postprocessing by scripts.
382.El
383.Ss LIST OF RULES AND PREPROCESSING
384To ease configuration, rules can be put into a file which is
385processed using
386.Nm
387as shown in the last synopsis line.
388An absolute
389.Ar pathname
390must be used.
391The file will be read line by line and applied as arguments to the
392.Nm
393utility.
394.Pp
395Optionally, a preprocessor can be specified using
396.Fl p Ar preproc
397where
398.Ar pathname
399is to be piped through.
400Useful preprocessors include
401.Xr cpp 1
402and
403.Xr m4 1 .
404If
405.Ar preproc
406does not start with a slash
407.Pq Ql /
408as its first character, the usual
409.Ev PATH
410name search is performed.
411Care should be taken with this in environments where not all
412file systems are mounted (yet) by the time
413.Nm
414is being run (e.g.\& when they are mounted over NFS).
415Once
416.Fl p
417has been specified, any additional arguments are passed on to the preprocessor
418for interpretation.
419This allows for flexible configuration files (like conditionalizing
420them on the local hostname) and the use of macros to centralize
421frequently required arguments like IP addresses.
422.Ss TRAFFIC SHAPER CONFIGURATION
423The
424.Nm
425.Cm pipe , queue
426and
427.Cm sched
428commands are used to configure the traffic shaper and packet scheduler.
429See the
430.Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
431Section below for details.
432.Pp
433If the world and the kernel get out of sync the
434.Nm
435ABI may break, preventing you from being able to add any rules.
436This can adversely affect the booting process.
437You can use
438.Nm
439.Cm disable
440.Cm firewall
441to temporarily disable the firewall to regain access to the network,
442allowing you to fix the problem.
443.Sh PACKET FLOW
444A packet is checked against the active ruleset in multiple places
445in the protocol stack, under control of several sysctl variables.
446These places and variables are shown below, and it is important to
447have this picture in mind in order to design a correct ruleset.
448.Bd -literal -offset indent
449       ^    to upper layers    V
450       |                       |
451       +----------->-----------+
452       ^                       V
453 [ip(6)_input]           [ip(6)_output]     net.inet(6).ip(6).fw.enable=1
454       |                       |
455       ^                       V
456 [ether_demux]        [ether_output_frame]  net.link.ether.ipfw=1
457       |                       |
458       +-->--[bdg_forward]-->--+            net.link.bridge.ipfw=1
459       ^                       V
460       |      to devices       |
461.Ed
462.Pp
463The number of
464times the same packet goes through the firewall can
465vary between 0 and 4 depending on packet source and
466destination, and system configuration.
467.Pp
468Note that as packets flow through the stack, headers can be
469stripped or added to it, and so they may or may not be available
470for inspection.
471E.g., incoming packets will include the MAC header when
472.Nm
473is invoked from
474.Cm ether_demux() ,
475but the same packets will have the MAC header stripped off when
476.Nm
477is invoked from
478.Cm ip_input()
479or
480.Cm ip6_input() .
481.Pp
482Also note that each packet is always checked against the complete ruleset,
483irrespective of the place where the check occurs, or the source of the packet.
484If a rule contains some match patterns or actions which are not valid
485for the place of invocation (e.g.\& trying to match a MAC header within
486.Cm ip_input
487or
488.Cm ip6_input ),
489the match pattern will not match, but a
490.Cm not
491operator in front of such patterns
492.Em will
493cause the pattern to
494.Em always
495match on those packets.
496It is thus the responsibility of
497the programmer, if necessary, to write a suitable ruleset to
498differentiate among the possible places.
499.Cm skipto
500rules can be useful here, as an example:
501.Bd -literal -offset indent
502# packets from ether_demux or bdg_forward
503ipfw add 10 skipto 1000 all from any to any layer2 in
504# packets from ip_input
505ipfw add 10 skipto 2000 all from any to any not layer2 in
506# packets from ip_output
507ipfw add 10 skipto 3000 all from any to any not layer2 out
508# packets from ether_output_frame
509ipfw add 10 skipto 4000 all from any to any layer2 out
510.Ed
511.Pp
512(yes, at the moment there is no way to differentiate between
513ether_demux and bdg_forward).
514.Pp
515Also note that only actions
516.Cm allow,
517.Cm deny,
518.Cm netgraph,
519.Cm ngtee
520and related to
521.Cm dummynet
522are processed for
523.Cm layer2
524frames and all other actions act as if they were
525.Cm allow
526for such frames.
527Full set of actions is supported for IP packets without
528.Cm layer2
529headers only.
530For example,
531.Cm divert
532action does not divert
533.Cm layer2
534frames.
535.Sh SYNTAX
536In general, each keyword or argument must be provided as
537a separate command line argument, with no leading or trailing
538spaces.
539Keywords are case-sensitive, whereas arguments may
540or may not be case-sensitive depending on their nature
541(e.g.\& uid's are, hostnames are not).
542.Pp
543Some arguments (e.g., port or address lists) are comma-separated
544lists of values.
545In this case, spaces after commas ',' are allowed to make
546the line more readable.
547You can also put the entire
548command (including flags) into a single argument.
549E.g., the following forms are equivalent:
550.Bd -literal -offset indent
551ipfw -q add deny src-ip 10.0.0.0/24,127.0.0.1/8
552ipfw -q add deny src-ip 10.0.0.0/24, 127.0.0.1/8
553ipfw "-q add deny src-ip 10.0.0.0/24, 127.0.0.1/8"
554.Ed
555.Sh RULE FORMAT
556The format of firewall rules is the following:
557.Bd -ragged -offset indent
558.Bk -words
559.Op Ar rule_number
560.Op Cm set Ar set_number
561.Op Cm prob Ar match_probability
562.Ar action
563.Op Cm log Op Cm logamount Ar number
564.Op Cm altq Ar queue
565.Oo
566.Bro Cm tag | untag
567.Brc Ar number
568.Oc
569.Ar body
570.Ek
571.Ed
572.Pp
573where the body of the rule specifies which information is used
574for filtering packets, among the following:
575.Pp
576.Bl -tag -width "Source and dest. addresses and ports" -offset XXX -compact
577.It Layer-2 header fields
578When available
579.It IPv4 and IPv6 Protocol
580SCTP, TCP, UDP, ICMP, etc.
581.It Source and dest. addresses and ports
582.It Direction
583See Section
584.Sx PACKET FLOW
585.It Transmit and receive interface
586By name or address
587.It Misc. IP header fields
588Version, type of service, datagram length, identification,
589fragment flag (non-zero IP offset),
590Time To Live
591.It IP options
592.It IPv6 Extension headers
593Fragmentation, Hop-by-Hop options,
594Routing Headers, Source routing rthdr0, Mobile IPv6 rthdr2, IPSec options.
595.It IPv6 Flow-ID
596.It Misc. TCP header fields
597TCP flags (SYN, FIN, ACK, RST, etc.),
598sequence number, acknowledgment number,
599window
600.It TCP options
601.It ICMP types
602for ICMP packets
603.It ICMP6 types
604for ICMP6 packets
605.It User/group ID
606When the packet can be associated with a local socket.
607.It Divert status
608Whether a packet came from a divert socket (e.g.,
609.Xr natd 8 ) .
610.It Fib annotation state
611Whether a packet has been tagged for using a specific FIB (routing table)
612in future forwarding decisions.
613.El
614.Pp
615Note that some of the above information, e.g.\& source MAC or IP addresses and
616TCP/UDP ports, can be easily spoofed, so filtering on those fields
617alone might not guarantee the desired results.
618.Bl -tag -width indent
619.It Ar rule_number
620Each rule is associated with a
621.Ar rule_number
622in the range 1..65535, with the latter reserved for the
623.Em default
624rule.
625Rules are checked sequentially by rule number.
626Multiple rules can have the same number, in which case they are
627checked (and listed) according to the order in which they have
628been added.
629If a rule is entered without specifying a number, the kernel will
630assign one in such a way that the rule becomes the last one
631before the
632.Em default
633rule.
634Automatic rule numbers are assigned by incrementing the last
635non-default rule number by the value of the sysctl variable
636.Ar net.inet.ip.fw.autoinc_step
637which defaults to 100.
638If this is not possible (e.g.\& because we would go beyond the
639maximum allowed rule number), the number of the last
640non-default value is used instead.
641.It Cm set Ar set_number
642Each rule is associated with a
643.Ar set_number
644in the range 0..31.
645Sets can be individually disabled and enabled, so this parameter
646is of fundamental importance for atomic ruleset manipulation.
647It can be also used to simplify deletion of groups of rules.
648If a rule is entered without specifying a set number,
649set 0 will be used.
650.br
651Set 31 is special in that it cannot be disabled,
652and rules in set 31 are not deleted by the
653.Nm ipfw flush
654command (but you can delete them with the
655.Nm ipfw delete set 31
656command).
657Set 31 is also used for the
658.Em default
659rule.
660.It Cm prob Ar match_probability
661A match is only declared with the specified probability
662(floating point number between 0 and 1).
663This can be useful for a number of applications such as
664random packet drop or
665(in conjunction with
666.Nm dummynet )
667to simulate the effect of multiple paths leading to out-of-order
668packet delivery.
669.Pp
670Note: this condition is checked before any other condition, including
671ones such as
672.Cm keep-state
673or
674.Cm check-state
675which might have
676side effects.
677.It Cm log Op Cm logamount Ar number
678Packets matching a rule with the
679.Cm log
680keyword will be made available for logging in two ways:
681if the sysctl variable
682.Va net.inet.ip.fw.verbose
683is set to 0 (default), one can use
684.Xr bpf 4
685attached to the
686.Li ipfw0
687pseudo interface.
688This pseudo interface can be created after a boot
689manually by using the following command:
690.Bd -literal -offset indent
691# ifconfig ipfw0 create
692.Ed
693.Pp
694Or, automatically at boot time by adding the following
695line to the
696.Xr rc.conf 5
697file:
698.Bd -literal -offset indent
699firewall_logif="YES"
700.Ed
701.Pp
702There is no overhead if no
703.Xr bpf 4
704is attached to the pseudo interface.
705.Pp
706If
707.Va net.inet.ip.fw.verbose
708is set to 1, packets will be logged to
709.Xr syslogd 8
710with a
711.Dv LOG_SECURITY
712facility up to a maximum of
713.Cm logamount
714packets.
715If no
716.Cm logamount
717is specified, the limit is taken from the sysctl variable
718.Va net.inet.ip.fw.verbose_limit .
719In both cases, a value of 0 means unlimited logging.
720.Pp
721Once the limit is reached, logging can be re-enabled by
722clearing the logging counter or the packet counter for that entry, see the
723.Cm resetlog
724command.
725.Pp
726Note: logging is done after all other packet matching conditions
727have been successfully verified, and before performing the final
728action (accept, deny, etc.) on the packet.
729.It Cm tag Ar number
730When a packet matches a rule with the
731.Cm tag
732keyword, the numeric tag for the given
733.Ar number
734in the range 1..65534 will be attached to the packet.
735The tag acts as an internal marker (it is not sent out over
736the wire) that can be used to identify these packets later on.
737This can be used, for example, to provide trust between interfaces
738and to start doing policy-based filtering.
739A packet can have multiple tags at the same time.
740Tags are "sticky", meaning once a tag is applied to a packet by a
741matching rule it exists until explicit removal.
742Tags are kept with the packet everywhere within the kernel, but are
743lost when packet leaves the kernel, for example, on transmitting
744packet out to the network or sending packet to a
745.Xr divert 4
746socket.
747.Pp
748To check for previously applied tags, use the
749.Cm tagged
750rule option.
751To delete previously applied tag, use the
752.Cm untag
753keyword.
754.Pp
755Note: since tags are kept with the packet everywhere in kernelspace,
756they can be set and unset anywhere in the kernel network subsystem
757(using the
758.Xr mbuf_tags 9
759facility), not only by means of the
760.Xr ipfw 4
761.Cm tag
762and
763.Cm untag
764keywords.
765For example, there can be a specialized
766.Xr netgraph 4
767node doing traffic analyzing and tagging for later inspecting
768in firewall.
769.It Cm untag Ar number
770When a packet matches a rule with the
771.Cm untag
772keyword, the tag with the number
773.Ar number
774is searched among the tags attached to this packet and,
775if found, removed from it.
776Other tags bound to packet, if present, are left untouched.
777.It Cm altq Ar queue
778When a packet matches a rule with the
779.Cm altq
780keyword, the ALTQ identifier for the given
781.Ar queue
782(see
783.Xr altq 4 )
784will be attached.
785Note that this ALTQ tag is only meaningful for packets going "out" of IPFW,
786and not being rejected or going to divert sockets.
787Note that if there is insufficient memory at the time the packet is
788processed, it will not be tagged, so it is wise to make your ALTQ
789"default" queue policy account for this.
790If multiple
791.Cm altq
792rules match a single packet, only the first one adds the ALTQ classification
793tag.
794In doing so, traffic may be shaped by using
795.Cm count Cm altq Ar queue
796rules for classification early in the ruleset, then later applying
797the filtering decision.
798For example,
799.Cm check-state
800and
801.Cm keep-state
802rules may come later and provide the actual filtering decisions in
803addition to the fallback ALTQ tag.
804.Pp
805You must run
806.Xr pfctl 8
807to set up the queues before IPFW will be able to look them up by name,
808and if the ALTQ disciplines are rearranged, the rules in containing the
809queue identifiers in the kernel will likely have gone stale and need
810to be reloaded.
811Stale queue identifiers will probably result in misclassification.
812.Pp
813All system ALTQ processing can be turned on or off via
814.Nm
815.Cm enable Ar altq
816and
817.Nm
818.Cm disable Ar altq .
819The usage of
820.Va net.inet.ip.fw.one_pass
821is irrelevant to ALTQ traffic shaping, as the actual rule action is followed
822always after adding an ALTQ tag.
823.El
824.Ss RULE ACTIONS
825A rule can be associated with one of the following actions, which
826will be executed when the packet matches the body of the rule.
827.Bl -tag -width indent
828.It Cm allow | accept | pass | permit
829Allow packets that match rule.
830The search terminates.
831.It Cm check-state Op Ar :flowname | Cm :any
832Checks the packet against the dynamic ruleset.
833If a match is found, execute the action associated with
834the rule which generated this dynamic rule, otherwise
835move to the next rule.
836.br
837.Cm Check-state
838rules do not have a body.
839If no
840.Cm check-state
841rule is found, the dynamic ruleset is checked at the first
842.Cm keep-state
843or
844.Cm limit
845rule.
846The
847.Ar :flowname
848is symbolic name assigned to dynamic rule by
849.Cm keep-state
850opcode.
851The special flowname
852.Cm :any
853can be used to ignore states flowname when matching.
854The
855.Cm :default
856keyword is special name used for compatibility with old rulesets.
857.It Cm count
858Update counters for all packets that match rule.
859The search continues with the next rule.
860.It Cm deny | drop
861Discard packets that match this rule.
862The search terminates.
863.It Cm divert Ar port
864Divert packets that match this rule to the
865.Xr divert 4
866socket bound to port
867.Ar port .
868The search terminates.
869.It Cm fwd | forward Ar ipaddr | tablearg Ns Op , Ns Ar port
870Change the next-hop on matching packets to
871.Ar ipaddr ,
872which can be an IP address or a host name.
873The next hop can also be supplied by the last table
874looked up for the packet by using the
875.Cm tablearg
876keyword instead of an explicit address.
877The search terminates if this rule matches.
878.Pp
879If
880.Ar ipaddr
881is a local address, then matching packets will be forwarded to
882.Ar port
883(or the port number in the packet if one is not specified in the rule)
884on the local machine.
885.br
886If
887.Ar ipaddr
888is not a local address, then the port number
889(if specified) is ignored, and the packet will be
890forwarded to the remote address, using the route as found in
891the local routing table for that IP.
892.br
893A
894.Ar fwd
895rule will not match layer-2 packets (those received
896on ether_input, ether_output, or bridged).
897.br
898The
899.Cm fwd
900action does not change the contents of the packet at all.
901In particular, the destination address remains unmodified, so
902packets forwarded to another system will usually be rejected by that system
903unless there is a matching rule on that system to capture them.
904For packets forwarded locally,
905the local address of the socket will be
906set to the original destination address of the packet.
907This makes the
908.Xr netstat 1
909entry look rather weird but is intended for
910use with transparent proxy servers.
911.It Cm nat Ar nat_nr | tablearg
912Pass packet to a
913nat instance
914(for network address translation, address redirect, etc.):
915see the
916.Sx NETWORK ADDRESS TRANSLATION (NAT)
917Section for further information.
918.It Cm nat64lsn Ar name
919Pass packet to a stateful NAT64 instance (for IPv6/IPv4 network address and
920protocol translation): see the
921.Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
922Section for further information.
923.It Cm nat64stl Ar name
924Pass packet to a stateless NAT64 instance (for IPv6/IPv4 network address and
925protocol translation): see the
926.Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
927Section for further information.
928.It Cm nptv6 Ar name
929Pass packet to a NPTv6 instance (for IPv6-to-IPv6 network prefix translation):
930see the
931.Sx IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
932Section for further information.
933.It Cm pipe Ar pipe_nr
934Pass packet to a
935.Nm dummynet
936.Dq pipe
937(for bandwidth limitation, delay, etc.).
938See the
939.Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
940Section for further information.
941The search terminates; however, on exit from the pipe and if
942the
943.Xr sysctl 8
944variable
945.Va net.inet.ip.fw.one_pass
946is not set, the packet is passed again to the firewall code
947starting from the next rule.
948.It Cm queue Ar queue_nr
949Pass packet to a
950.Nm dummynet
951.Dq queue
952(for bandwidth limitation using WF2Q+).
953.It Cm reject
954(Deprecated).
955Synonym for
956.Cm unreach host .
957.It Cm reset
958Discard packets that match this rule, and if the
959packet is a TCP packet, try to send a TCP reset (RST) notice.
960The search terminates.
961.It Cm reset6
962Discard packets that match this rule, and if the
963packet is a TCP packet, try to send a TCP reset (RST) notice.
964The search terminates.
965.It Cm skipto Ar number | tablearg
966Skip all subsequent rules numbered less than
967.Ar number .
968The search continues with the first rule numbered
969.Ar number
970or higher.
971It is possible to use the
972.Cm tablearg
973keyword with a skipto for a
974.Em computed
975skipto. Skipto may work either in O(log(N)) or in O(1) depending
976on amount of memory and/or sysctl variables.
977See the
978.Sx SYSCTL VARIABLES
979section for more details.
980.It Cm call Ar number | tablearg
981The current rule number is saved in the internal stack and
982ruleset processing continues with the first rule numbered
983.Ar number
984or higher.
985If later a rule with the
986.Cm return
987action is encountered, the processing returns to the first rule
988with number of this
989.Cm call
990rule plus one or higher
991(the same behaviour as with packets returning from
992.Xr divert 4
993socket after a
994.Cm divert
995action).
996This could be used to make somewhat like an assembly language
997.Dq subroutine
998calls to rules with common checks for different interfaces, etc.
999.Pp
1000Rule with any number could be called, not just forward jumps as with
1001.Cm skipto .
1002So, to prevent endless loops in case of mistakes, both
1003.Cm call
1004and
1005.Cm return
1006actions don't do any jumps and simply go to the next rule if memory
1007cannot be allocated or stack overflowed/underflowed.
1008.Pp
1009Internally stack for rule numbers is implemented using
1010.Xr mbuf_tags 9
1011facility and currently has size of 16 entries.
1012As mbuf tags are lost when packet leaves the kernel,
1013.Cm divert
1014should not be used in subroutines to avoid endless loops
1015and other undesired effects.
1016.It Cm return
1017Takes rule number saved to internal stack by the last
1018.Cm call
1019action and returns ruleset processing to the first rule
1020with number greater than number of corresponding
1021.Cm call
1022rule.
1023See description of the
1024.Cm call
1025action for more details.
1026.Pp
1027Note that
1028.Cm return
1029rules usually end a
1030.Dq subroutine
1031and thus are unconditional, but
1032.Nm
1033command-line utility currently requires every action except
1034.Cm check-state
1035to have body.
1036While it is sometimes useful to return only on some packets,
1037usually you want to print just
1038.Dq return
1039for readability.
1040A workaround for this is to use new syntax and
1041.Fl c
1042switch:
1043.Bd -literal -offset indent
1044# Add a rule without actual body
1045ipfw add 2999 return via any
1046
1047# List rules without "from any to any" part
1048ipfw -c list
1049.Ed
1050.Pp
1051This cosmetic annoyance may be fixed in future releases.
1052.It Cm tee Ar port
1053Send a copy of packets matching this rule to the
1054.Xr divert 4
1055socket bound to port
1056.Ar port .
1057The search continues with the next rule.
1058.It Cm unreach Ar code
1059Discard packets that match this rule, and try to send an ICMP
1060unreachable notice with code
1061.Ar code ,
1062where
1063.Ar code
1064is a number from 0 to 255, or one of these aliases:
1065.Cm net , host , protocol , port ,
1066.Cm needfrag , srcfail , net-unknown , host-unknown ,
1067.Cm isolated , net-prohib , host-prohib , tosnet ,
1068.Cm toshost , filter-prohib , host-precedence
1069or
1070.Cm precedence-cutoff .
1071The search terminates.
1072.It Cm unreach6 Ar code
1073Discard packets that match this rule, and try to send an ICMPv6
1074unreachable notice with code
1075.Ar code ,
1076where
1077.Ar code
1078is a number from 0, 1, 3 or 4, or one of these aliases:
1079.Cm no-route, admin-prohib, address
1080or
1081.Cm port .
1082The search terminates.
1083.It Cm netgraph Ar cookie
1084Divert packet into netgraph with given
1085.Ar cookie .
1086The search terminates.
1087If packet is later returned from netgraph it is either
1088accepted or continues with the next rule, depending on
1089.Va net.inet.ip.fw.one_pass
1090sysctl variable.
1091.It Cm ngtee Ar cookie
1092A copy of packet is diverted into netgraph, original
1093packet continues with the next rule.
1094See
1095.Xr ng_ipfw 4
1096for more information on
1097.Cm netgraph
1098and
1099.Cm ngtee
1100actions.
1101.It Cm setfib Ar fibnum | tablearg
1102The packet is tagged so as to use the FIB (routing table)
1103.Ar fibnum
1104in any subsequent forwarding decisions.
1105In the current implementation, this is limited to the values 0 through 15, see
1106.Xr setfib 2 .
1107Processing continues at the next rule.
1108It is possible to use the
1109.Cm tablearg
1110keyword with setfib.
1111If the tablearg value is not within the compiled range of fibs,
1112the packet's fib is set to 0.
1113.It Cm setdscp Ar DSCP | number | tablearg
1114Set specified DiffServ codepoint for an IPv4/IPv6 packet.
1115Processing continues at the next rule.
1116Supported values are:
1117.Pp
1118.Cm cs0
1119.Pq Dv 000000 ,
1120.Cm cs1
1121.Pq Dv 001000 ,
1122.Cm cs2
1123.Pq Dv 010000 ,
1124.Cm cs3
1125.Pq Dv 011000 ,
1126.Cm cs4
1127.Pq Dv 100000 ,
1128.Cm cs5
1129.Pq Dv 101000 ,
1130.Cm cs6
1131.Pq Dv 110000 ,
1132.Cm cs7
1133.Pq Dv 111000 ,
1134.Cm af11
1135.Pq Dv 001010 ,
1136.Cm af12
1137.Pq Dv 001100 ,
1138.Cm af13
1139.Pq Dv 001110 ,
1140.Cm af21
1141.Pq Dv 010010 ,
1142.Cm af22
1143.Pq Dv 010100 ,
1144.Cm af23
1145.Pq Dv 010110 ,
1146.Cm af31
1147.Pq Dv 011010 ,
1148.Cm af32
1149.Pq Dv 011100 ,
1150.Cm af33
1151.Pq Dv 011110 ,
1152.Cm af41
1153.Pq Dv 100010 ,
1154.Cm af42
1155.Pq Dv 100100 ,
1156.Cm af43
1157.Pq Dv 100110 ,
1158.Cm ef
1159.Pq Dv 101110 ,
1160.Cm be
1161.Pq Dv 000000 .
1162Additionally, DSCP value can be specified by number (0..64).
1163It is also possible to use the
1164.Cm tablearg
1165keyword with setdscp.
1166If the tablearg value is not within the 0..64 range, lower 6 bits of supplied
1167value are used.
1168.It Cm tcp-setmss Ar mss
1169Set the Maximum Segment Size (MSS) in the TCP segment to value
1170.Ar mss .
1171The kernel module
1172.Cm ipfw_pmod
1173should be loaded or kernel should have
1174.Cm options IPFIREWALL_PMOD
1175to be able use this action.
1176This command does not change a packet if original MSS value is lower than
1177specified value.
1178Both TCP over IPv4 and over IPv6 are supported.
1179Regardless of matched a packet or not by the
1180.Cm tcp-setmss
1181rule, the search continues with the next rule.
1182.It Cm reass
1183Queue and reassemble IPv4 fragments.
1184If the packet is not fragmented, counters are updated and
1185processing continues with the next rule.
1186If the packet is the last logical fragment, the packet is reassembled and, if
1187.Va net.inet.ip.fw.one_pass
1188is set to 0, processing continues with the next rule.
1189Otherwise, the packet is allowed to pass and the search terminates.
1190If the packet is a fragment in the middle of a logical group of fragments,
1191it is consumed and
1192processing stops immediately.
1193.Pp
1194Fragment handling can be tuned via
1195.Va net.inet.ip.maxfragpackets
1196and
1197.Va net.inet.ip.maxfragsperpacket
1198which limit, respectively, the maximum number of processable
1199fragments (default: 800) and
1200the maximum number of fragments per packet (default: 16).
1201.Pp
1202NOTA BENE: since fragments do not contain port numbers,
1203they should be avoided with the
1204.Nm reass
1205rule.
1206Alternatively, direction-based (like
1207.Nm in
1208/
1209.Nm out
1210) and source-based (like
1211.Nm via
1212) match patterns can be used to select fragments.
1213.Pp
1214Usually a simple rule like:
1215.Bd -literal -offset indent
1216# reassemble incoming fragments
1217ipfw add reass all from any to any in
1218.Ed
1219.Pp
1220is all you need at the beginning of your ruleset.
1221.It Cm abort
1222Discard packets that match this rule, and if the packet is an SCTP packet,
1223try to send an SCTP packet containing an ABORT chunk.
1224The search terminates.
1225.It Cm abort6
1226Discard packets that match this rule, and if the packet is an SCTP packet,
1227try to send an SCTP packet containing an ABORT chunk.
1228The search terminates.
1229.El
1230.Ss RULE BODY
1231The body of a rule contains zero or more patterns (such as
1232specific source and destination addresses or ports,
1233protocol options, incoming or outgoing interfaces, etc.)
1234that the packet must match in order to be recognised.
1235In general, the patterns are connected by (implicit)
1236.Cm and
1237operators -- i.e., all must match in order for the
1238rule to match.
1239Individual patterns can be prefixed by the
1240.Cm not
1241operator to reverse the result of the match, as in
1242.Pp
1243.Dl "ipfw add 100 allow ip from not 1.2.3.4 to any"
1244.Pp
1245Additionally, sets of alternative match patterns
1246.Pq Em or-blocks
1247can be constructed by putting the patterns in
1248lists enclosed between parentheses ( ) or braces { }, and
1249using the
1250.Cm or
1251operator as follows:
1252.Pp
1253.Dl "ipfw add 100 allow ip from { x or not y or z } to any"
1254.Pp
1255Only one level of parentheses is allowed.
1256Beware that most shells have special meanings for parentheses
1257or braces, so it is advisable to put a backslash \\ in front of them
1258to prevent such interpretations.
1259.Pp
1260The body of a rule must in general include a source and destination
1261address specifier.
1262The keyword
1263.Ar any
1264can be used in various places to specify that the content of
1265a required field is irrelevant.
1266.Pp
1267The rule body has the following format:
1268.Bd -ragged -offset indent
1269.Op Ar proto Cm from Ar src Cm to Ar dst
1270.Op Ar options
1271.Ed
1272.Pp
1273The first part (proto from src to dst) is for backward
1274compatibility with earlier versions of
1275.Fx .
1276In modern
1277.Fx
1278any match pattern (including MAC headers, IP protocols,
1279addresses and ports) can be specified in the
1280.Ar options
1281section.
1282.Pp
1283Rule fields have the following meaning:
1284.Bl -tag -width indent
1285.It Ar proto : protocol | Cm { Ar protocol Cm or ... }
1286.It Ar protocol : Oo Cm not Oc Ar protocol-name | protocol-number
1287An IP protocol specified by number or name
1288(for a complete list see
1289.Pa /etc/protocols ) ,
1290or one of the following keywords:
1291.Bl -tag -width indent
1292.It Cm ip4 | ipv4
1293Matches IPv4 packets.
1294.It Cm ip6 | ipv6
1295Matches IPv6 packets.
1296.It Cm ip | all
1297Matches any packet.
1298.El
1299.Pp
1300The
1301.Cm ipv6
1302in
1303.Cm proto
1304option will be treated as inner protocol.
1305And, the
1306.Cm ipv4
1307is not available in
1308.Cm proto
1309option.
1310.Pp
1311The
1312.Cm { Ar protocol Cm or ... }
1313format (an
1314.Em or-block )
1315is provided for convenience only but its use is deprecated.
1316.It Ar src No and Ar dst : Bro Cm addr | Cm { Ar addr Cm or ... } Brc Op Oo Cm not Oc Ar ports
1317An address (or a list, see below)
1318optionally followed by
1319.Ar ports
1320specifiers.
1321.Pp
1322The second format
1323.Em ( or-block
1324with multiple addresses) is provided for convenience only and
1325its use is discouraged.
1326.It Ar addr : Oo Cm not Oc Bro
1327.Cm any | me | me6 |
1328.Cm table Ns Pq Ar name Ns Op , Ns Ar value
1329.Ar | addr-list | addr-set
1330.Brc
1331.Bl -tag -width indent
1332.It Cm any
1333matches any IP address.
1334.It Cm me
1335matches any IP address configured on an interface in the system.
1336.It Cm me6
1337matches any IPv6 address configured on an interface in the system.
1338The address list is evaluated at the time the packet is
1339analysed.
1340.It Cm table Ns Pq Ar name Ns Op , Ns Ar value
1341Matches any IPv4 or IPv6 address for which an entry exists in the lookup table
1342.Ar number .
1343If an optional 32-bit unsigned
1344.Ar value
1345is also specified, an entry will match only if it has this value.
1346See the
1347.Sx LOOKUP TABLES
1348section below for more information on lookup tables.
1349.El
1350.It Ar addr-list : ip-addr Ns Op Ns , Ns Ar addr-list
1351.It Ar ip-addr :
1352A host or subnet address specified in one of the following ways:
1353.Bl -tag -width indent
1354.It Ar numeric-ip | hostname
1355Matches a single IPv4 address, specified as dotted-quad or a hostname.
1356Hostnames are resolved at the time the rule is added to the firewall list.
1357.It Ar addr Ns / Ns Ar masklen
1358Matches all addresses with base
1359.Ar addr
1360(specified as an IP address, a network number, or a hostname)
1361and mask width of
1362.Cm masklen
1363bits.
1364As an example, 1.2.3.4/25 or 1.2.3.0/25 will match
1365all IP numbers from 1.2.3.0 to 1.2.3.127 .
1366.It Ar addr Ns : Ns Ar mask
1367Matches all addresses with base
1368.Ar addr
1369(specified as an IP address, a network number, or a hostname)
1370and the mask of
1371.Ar mask ,
1372specified as a dotted quad.
1373As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match
13741.*.3.*.
1375This form is advised only for non-contiguous
1376masks.
1377It is better to resort to the
1378.Ar addr Ns / Ns Ar masklen
1379format for contiguous masks, which is more compact and less
1380error-prone.
1381.El
1382.It Ar addr-set : addr Ns Oo Ns / Ns Ar masklen Oc Ns Cm { Ns Ar list Ns Cm }
1383.It Ar list : Bro Ar num | num-num Brc Ns Op Ns , Ns Ar list
1384Matches all addresses with base address
1385.Ar addr
1386(specified as an IP address, a network number, or a hostname)
1387and whose last byte is in the list between braces { } .
1388Note that there must be no spaces between braces and
1389numbers (spaces after commas are allowed).
1390Elements of the list can be specified as single entries
1391or ranges.
1392The
1393.Ar masklen
1394field is used to limit the size of the set of addresses,
1395and can have any value between 24 and 32.
1396If not specified,
1397it will be assumed as 24.
1398.br
1399This format is particularly useful to handle sparse address sets
1400within a single rule.
1401Because the matching occurs using a
1402bitmask, it takes constant time and dramatically reduces
1403the complexity of rulesets.
1404.br
1405As an example, an address specified as 1.2.3.4/24{128,35-55,89}
1406or 1.2.3.0/24{128,35-55,89}
1407will match the following IP addresses:
1408.br
14091.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .
1410.It Ar addr6-list : ip6-addr Ns Op Ns , Ns Ar addr6-list
1411.It Ar ip6-addr :
1412A host or subnet specified one of the following ways:
1413.Bl -tag -width indent
1414.It Ar numeric-ip | hostname
1415Matches a single IPv6 address as allowed by
1416.Xr inet_pton 3
1417or a hostname.
1418Hostnames are resolved at the time the rule is added to the firewall
1419list.
1420.It Ar addr Ns / Ns Ar masklen
1421Matches all IPv6 addresses with base
1422.Ar addr
1423(specified as allowed by
1424.Xr inet_pton
1425or a hostname)
1426and mask width of
1427.Cm masklen
1428bits.
1429.It Ar addr Ns / Ns Ar mask
1430Matches all IPv6 addresses with base
1431.Ar addr
1432(specified as allowed by
1433.Xr inet_pton
1434or a hostname)
1435and the mask of
1436.Ar mask ,
1437specified as allowed by
1438.Xr inet_pton.
1439As an example, fe::640:0:0/ffff::ffff:ffff:0:0 will match
1440fe:*:*:*:0:640:*:*.
1441This form is advised only for non-contiguous
1442masks.
1443It is better to resort to the
1444.Ar addr Ns / Ns Ar masklen
1445format for contiguous masks, which is more compact and less
1446error-prone.
1447.El
1448.Pp
1449No support for sets of IPv6 addresses is provided because IPv6 addresses
1450are typically random past the initial prefix.
1451.It Ar ports : Bro Ar port | port Ns \&- Ns Ar port Ns Brc Ns Op , Ns Ar ports
1452For protocols which support port numbers (such as SCTP, TCP and UDP), optional
1453.Cm ports
1454may be specified as one or more ports or port ranges, separated
1455by commas but no spaces, and an optional
1456.Cm not
1457operator.
1458The
1459.Ql \&-
1460notation specifies a range of ports (including boundaries).
1461.Pp
1462Service names (from
1463.Pa /etc/services )
1464may be used instead of numeric port values.
1465The length of the port list is limited to 30 ports or ranges,
1466though one can specify larger ranges by using an
1467.Em or-block
1468in the
1469.Cm options
1470section of the rule.
1471.Pp
1472A backslash
1473.Pq Ql \e
1474can be used to escape the dash
1475.Pq Ql -
1476character in a service name (from a shell, the backslash must be
1477typed twice to avoid the shell itself interpreting it as an escape
1478character).
1479.Pp
1480.Dl "ipfw add count tcp from any ftp\e\e-data-ftp to any"
1481.Pp
1482Fragmented packets which have a non-zero offset (i.e., not the first
1483fragment) will never match a rule which has one or more port
1484specifications.
1485See the
1486.Cm frag
1487option for details on matching fragmented packets.
1488.El
1489.Ss RULE OPTIONS (MATCH PATTERNS)
1490Additional match patterns can be used within
1491rules.
1492Zero or more of these so-called
1493.Em options
1494can be present in a rule, optionally prefixed by the
1495.Cm not
1496operand, and possibly grouped into
1497.Em or-blocks .
1498.Pp
1499The following match patterns can be used (listed in alphabetical order):
1500.Bl -tag -width indent
1501.It Cm // this is a comment.
1502Inserts the specified text as a comment in the rule.
1503Everything following // is considered as a comment and stored in the rule.
1504You can have comment-only rules, which are listed as having a
1505.Cm count
1506action followed by the comment.
1507.It Cm bridged
1508Alias for
1509.Cm layer2 .
1510.It Cm defer-immediate-action | defer-action
1511A rule with this option will not perform normal action
1512upon a match. This option is intended to be used with
1513.Cm record-state
1514or
1515.Cm keep-state
1516as the dynamic rule, created but ignored on match, will work
1517as intended.
1518Rules with both
1519.Cm record-state
1520and
1521.Cm defer-immediate-action
1522create a dynamic rule and continue with the next rule without actually
1523performing the action part of this rule. When the rule is later activated
1524via the state table, the action is performed as usual.
1525.It Cm diverted
1526Matches only packets generated by a divert socket.
1527.It Cm diverted-loopback
1528Matches only packets coming from a divert socket back into the IP stack
1529input for delivery.
1530.It Cm diverted-output
1531Matches only packets going from a divert socket back outward to the IP
1532stack output for delivery.
1533.It Cm dst-ip Ar ip-address
1534Matches IPv4 packets whose destination IP is one of the address(es)
1535specified as argument.
1536.It Bro Cm dst-ip6 | dst-ipv6 Brc Ar ip6-address
1537Matches IPv6 packets whose destination IP is one of the address(es)
1538specified as argument.
1539.It Cm dst-port Ar ports
1540Matches IP packets whose destination port is one of the port(s)
1541specified as argument.
1542.It Cm established
1543Matches TCP packets that have the RST or ACK bits set.
1544.It Cm ext6hdr Ar header
1545Matches IPv6 packets containing the extended header given by
1546.Ar header .
1547Supported headers are:
1548.Pp
1549Fragment,
1550.Pq Cm frag ,
1551Hop-to-hop options
1552.Pq Cm hopopt ,
1553any type of Routing Header
1554.Pq Cm route ,
1555Source routing Routing Header Type 0
1556.Pq Cm rthdr0 ,
1557Mobile IPv6 Routing Header Type 2
1558.Pq Cm rthdr2 ,
1559Destination options
1560.Pq Cm dstopt ,
1561IPSec authentication headers
1562.Pq Cm ah ,
1563and IPsec encapsulated security payload headers
1564.Pq Cm esp .
1565.It Cm fib Ar fibnum
1566Matches a packet that has been tagged to use
1567the given FIB (routing table) number.
1568.It Cm flow Ar table Ns Pq Ar name Ns Op , Ns Ar value
1569Search for the flow entry in lookup table
1570.Ar name .
1571If not found, the match fails.
1572Otherwise, the match succeeds and
1573.Cm tablearg
1574is set to the value extracted from the table.
1575.Pp
1576This option can be useful to quickly dispatch traffic based on
1577certain packet fields.
1578See the
1579.Sx LOOKUP TABLES
1580section below for more information on lookup tables.
1581.It Cm flow-id Ar labels
1582Matches IPv6 packets containing any of the flow labels given in
1583.Ar labels .
1584.Ar labels
1585is a comma separated list of numeric flow labels.
1586.It Cm frag
1587Matches packets that are fragments and not the first
1588fragment of an IP datagram.
1589Note that these packets will not have
1590the next protocol header (e.g.\& TCP, UDP) so options that look into
1591these headers cannot match.
1592.It Cm gid Ar group
1593Matches all TCP or UDP packets sent by or received for a
1594.Ar group .
1595A
1596.Ar group
1597may be specified by name or number.
1598.It Cm jail Ar jail
1599Matches all TCP or UDP packets sent by or received for the
1600jail whose ID or name is
1601.Ar jail .
1602.It Cm icmptypes Ar types
1603Matches ICMP packets whose ICMP type is in the list
1604.Ar types .
1605The list may be specified as any combination of
1606individual types (numeric) separated by commas.
1607.Em Ranges are not allowed .
1608The supported ICMP types are:
1609.Pp
1610echo reply
1611.Pq Cm 0 ,
1612destination unreachable
1613.Pq Cm 3 ,
1614source quench
1615.Pq Cm 4 ,
1616redirect
1617.Pq Cm 5 ,
1618echo request
1619.Pq Cm 8 ,
1620router advertisement
1621.Pq Cm 9 ,
1622router solicitation
1623.Pq Cm 10 ,
1624time-to-live exceeded
1625.Pq Cm 11 ,
1626IP header bad
1627.Pq Cm 12 ,
1628timestamp request
1629.Pq Cm 13 ,
1630timestamp reply
1631.Pq Cm 14 ,
1632information request
1633.Pq Cm 15 ,
1634information reply
1635.Pq Cm 16 ,
1636address mask request
1637.Pq Cm 17
1638and address mask reply
1639.Pq Cm 18 .
1640.It Cm icmp6types Ar types
1641Matches ICMP6 packets whose ICMP6 type is in the list of
1642.Ar types .
1643The list may be specified as any combination of
1644individual types (numeric) separated by commas.
1645.Em Ranges are not allowed .
1646.It Cm in | out
1647Matches incoming or outgoing packets, respectively.
1648.Cm in
1649and
1650.Cm out
1651are mutually exclusive (in fact,
1652.Cm out
1653is implemented as
1654.Cm not in Ns No ).
1655.It Cm ipid Ar id-list
1656Matches IPv4 packets whose
1657.Cm ip_id
1658field has value included in
1659.Ar id-list ,
1660which is either a single value or a list of values or ranges
1661specified in the same way as
1662.Ar ports .
1663.It Cm iplen Ar len-list
1664Matches IP packets whose total length, including header and data, is
1665in the set
1666.Ar len-list ,
1667which is either a single value or a list of values or ranges
1668specified in the same way as
1669.Ar ports .
1670.It Cm ipoptions Ar spec
1671Matches packets whose IPv4 header contains the comma separated list of
1672options specified in
1673.Ar spec .
1674The supported IP options are:
1675.Pp
1676.Cm ssrr
1677(strict source route),
1678.Cm lsrr
1679(loose source route),
1680.Cm rr
1681(record packet route) and
1682.Cm ts
1683(timestamp).
1684The absence of a particular option may be denoted
1685with a
1686.Ql \&! .
1687.It Cm ipprecedence Ar precedence
1688Matches IPv4 packets whose precedence field is equal to
1689.Ar precedence .
1690.It Cm ipsec
1691Matches packets that have IPSEC history associated with them
1692(i.e., the packet comes encapsulated in IPSEC, the kernel
1693has IPSEC support, and can correctly decapsulate it).
1694.Pp
1695Note that specifying
1696.Cm ipsec
1697is different from specifying
1698.Cm proto Ar ipsec
1699as the latter will only look at the specific IP protocol field,
1700irrespective of IPSEC kernel support and the validity of the IPSEC data.
1701.Pp
1702Further note that this flag is silently ignored in kernels without
1703IPSEC support.
1704It does not affect rule processing when given and the
1705rules are handled as if with no
1706.Cm ipsec
1707flag.
1708.It Cm iptos Ar spec
1709Matches IPv4 packets whose
1710.Cm tos
1711field contains the comma separated list of
1712service types specified in
1713.Ar spec .
1714The supported IP types of service are:
1715.Pp
1716.Cm lowdelay
1717.Pq Dv IPTOS_LOWDELAY ,
1718.Cm throughput
1719.Pq Dv IPTOS_THROUGHPUT ,
1720.Cm reliability
1721.Pq Dv IPTOS_RELIABILITY ,
1722.Cm mincost
1723.Pq Dv IPTOS_MINCOST ,
1724.Cm congestion
1725.Pq Dv IPTOS_ECN_CE .
1726The absence of a particular type may be denoted
1727with a
1728.Ql \&! .
1729.It Cm dscp spec Ns Op , Ns Ar spec
1730Matches IPv4/IPv6 packets whose
1731.Cm DS
1732field value is contained in
1733.Ar spec
1734mask.
1735Multiple values can be specified via
1736the comma separated list.
1737Value can be one of keywords used in
1738.Cm setdscp
1739action or exact number.
1740.It Cm ipttl Ar ttl-list
1741Matches IPv4 packets whose time to live is included in
1742.Ar ttl-list ,
1743which is either a single value or a list of values or ranges
1744specified in the same way as
1745.Ar ports .
1746.It Cm ipversion Ar ver
1747Matches IP packets whose IP version field is
1748.Ar ver .
1749.It Cm keep-state Op Ar :flowname
1750Upon a match, the firewall will create a dynamic rule, whose
1751default behaviour is to match bidirectional traffic between
1752source and destination IP/port using the same protocol.
1753The rule has a limited lifetime (controlled by a set of
1754.Xr sysctl 8
1755variables), and the lifetime is refreshed every time a matching
1756packet is found.
1757The
1758.Ar :flowname
1759is used to assign additional to addresses, ports and protocol parameter
1760to dynamic rule. It can be used for more accurate matching by
1761.Cm check-state
1762rule.
1763The
1764.Cm :default
1765keyword is special name used for compatibility with old rulesets.
1766.It Cm layer2
1767Matches only layer2 packets, i.e., those passed to
1768.Nm
1769from ether_demux() and ether_output_frame().
1770.It Cm limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N Op Ar :flowname
1771The firewall will only allow
1772.Ar N
1773connections with the same
1774set of parameters as specified in the rule.
1775One or more
1776of source and destination addresses and ports can be
1777specified.
1778.It Cm lookup Bro Cm dst-ip | dst-port | src-ip | src-port | uid | jail Brc Ar name
1779Search an entry in lookup table
1780.Ar name
1781that matches the field specified as argument.
1782If not found, the match fails.
1783Otherwise, the match succeeds and
1784.Cm tablearg
1785is set to the value extracted from the table.
1786.Pp
1787This option can be useful to quickly dispatch traffic based on
1788certain packet fields.
1789See the
1790.Sx LOOKUP TABLES
1791section below for more information on lookup tables.
1792.It Cm { MAC | mac } Ar dst-mac src-mac
1793Match packets with a given
1794.Ar dst-mac
1795and
1796.Ar src-mac
1797addresses, specified as the
1798.Cm any
1799keyword (matching any MAC address), or six groups of hex digits
1800separated by colons,
1801and optionally followed by a mask indicating the significant bits.
1802The mask may be specified using either of the following methods:
1803.Bl -enum -width indent
1804.It
1805A slash
1806.Pq /
1807followed by the number of significant bits.
1808For example, an address with 33 significant bits could be specified as:
1809.Pp
1810.Dl "MAC 10:20:30:40:50:60/33 any"
1811.It
1812An ampersand
1813.Pq &
1814followed by a bitmask specified as six groups of hex digits separated
1815by colons.
1816For example, an address in which the last 16 bits are significant could
1817be specified as:
1818.Pp
1819.Dl "MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any"
1820.Pp
1821Note that the ampersand character has a special meaning in many shells
1822and should generally be escaped.
1823.El
1824Note that the order of MAC addresses (destination first,
1825source second) is
1826the same as on the wire, but the opposite of the one used for
1827IP addresses.
1828.It Cm mac-type Ar mac-type
1829Matches packets whose Ethernet Type field
1830corresponds to one of those specified as argument.
1831.Ar mac-type
1832is specified in the same way as
1833.Cm port numbers
1834(i.e., one or more comma-separated single values or ranges).
1835You can use symbolic names for known values such as
1836.Em vlan , ipv4, ipv6 .
1837Values can be entered as decimal or hexadecimal (if prefixed by 0x),
1838and they are always printed as hexadecimal (unless the
1839.Cm -N
1840option is used, in which case symbolic resolution will be attempted).
1841.It Cm proto Ar protocol
1842Matches packets with the corresponding IP protocol.
1843.It Cm record-state
1844Upon a match, the firewall will create a dynamic rule as if
1845.Cm keep-state
1846was specified.
1847However, this option doesn't imply an implicit
1848.Cm check-state
1849in contrast to
1850.Cm keep-state .
1851.It Cm recv | xmit | via Brq Ar ifX | Ar if Ns Cm * | Ar table Ns Po Ar name Ns Oo , Ns Ar value Oc Pc | Ar ipno | Ar any
1852Matches packets received, transmitted or going through,
1853respectively, the interface specified by exact name
1854.Po Ar ifX Pc ,
1855by device name
1856.Po Ar if* Pc ,
1857by IP address, or through some interface.
1858Table
1859.Ar name
1860may be used to match interface by its kernel ifindex.
1861See the
1862.Sx LOOKUP TABLES
1863section below for more information on lookup tables.
1864.Pp
1865The
1866.Cm via
1867keyword causes the interface to always be checked.
1868If
1869.Cm recv
1870or
1871.Cm xmit
1872is used instead of
1873.Cm via ,
1874then only the receive or transmit interface (respectively)
1875is checked.
1876By specifying both, it is possible to match packets based on
1877both receive and transmit interface, e.g.:
1878.Pp
1879.Dl "ipfw add deny ip from any to any out recv ed0 xmit ed1"
1880.Pp
1881The
1882.Cm recv
1883interface can be tested on either incoming or outgoing packets,
1884while the
1885.Cm xmit
1886interface can only be tested on outgoing packets.
1887So
1888.Cm out
1889is required (and
1890.Cm in
1891is invalid) whenever
1892.Cm xmit
1893is used.
1894.Pp
1895A packet might not have a receive or transmit interface: packets
1896originating from the local host have no receive interface,
1897while packets destined for the local host have no transmit
1898interface.
1899.It Cm set-limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N
1900Works like
1901.Cm limit
1902but does not have an implicit
1903.Cm check-state
1904attached to it.
1905.It Cm setup
1906Matches TCP packets that have the SYN bit set but no ACK bit.
1907This is the short form of
1908.Dq Li tcpflags\ syn,!ack .
1909.It Cm sockarg
1910Matches packets that are associated to a local socket and
1911for which the SO_USER_COOKIE socket option has been set
1912to a non-zero value.
1913As a side effect, the value of the
1914option is made available as
1915.Cm tablearg
1916value, which in turn can be used as
1917.Cm skipto
1918or
1919.Cm pipe
1920number.
1921.It Cm src-ip Ar ip-address
1922Matches IPv4 packets whose source IP is one of the address(es)
1923specified as an argument.
1924.It Cm src-ip6 Ar ip6-address
1925Matches IPv6 packets whose source IP is one of the address(es)
1926specified as an argument.
1927.It Cm src-port Ar ports
1928Matches IP packets whose source port is one of the port(s)
1929specified as argument.
1930.It Cm tagged Ar tag-list
1931Matches packets whose tags are included in
1932.Ar tag-list ,
1933which is either a single value or a list of values or ranges
1934specified in the same way as
1935.Ar ports .
1936Tags can be applied to the packet using
1937.Cm tag
1938rule action parameter (see it's description for details on tags).
1939.It Cm tcpack Ar ack
1940TCP packets only.
1941Match if the TCP header acknowledgment number field is set to
1942.Ar ack .
1943.It Cm tcpdatalen Ar tcpdatalen-list
1944Matches TCP packets whose length of TCP data is
1945.Ar tcpdatalen-list ,
1946which is either a single value or a list of values or ranges
1947specified in the same way as
1948.Ar ports .
1949.It Cm tcpflags Ar spec
1950TCP packets only.
1951Match if the TCP header contains the comma separated list of
1952flags specified in
1953.Ar spec .
1954The supported TCP flags are:
1955.Pp
1956.Cm fin ,
1957.Cm syn ,
1958.Cm rst ,
1959.Cm psh ,
1960.Cm ack
1961and
1962.Cm urg .
1963The absence of a particular flag may be denoted
1964with a
1965.Ql \&! .
1966A rule which contains a
1967.Cm tcpflags
1968specification can never match a fragmented packet which has
1969a non-zero offset.
1970See the
1971.Cm frag
1972option for details on matching fragmented packets.
1973.It Cm tcpseq Ar seq
1974TCP packets only.
1975Match if the TCP header sequence number field is set to
1976.Ar seq .
1977.It Cm tcpwin Ar tcpwin-list
1978Matches TCP packets whose  header window field is set to
1979.Ar tcpwin-list ,
1980which is either a single value or a list of values or ranges
1981specified in the same way as
1982.Ar ports .
1983.It Cm tcpoptions Ar spec
1984TCP packets only.
1985Match if the TCP header contains the comma separated list of
1986options specified in
1987.Ar spec .
1988The supported TCP options are:
1989.Pp
1990.Cm mss
1991(maximum segment size),
1992.Cm window
1993(tcp window advertisement),
1994.Cm sack
1995(selective ack),
1996.Cm ts
1997(rfc1323 timestamp) and
1998.Cm cc
1999(rfc1644 t/tcp connection count).
2000The absence of a particular option may be denoted
2001with a
2002.Ql \&! .
2003.It Cm uid Ar user
2004Match all TCP or UDP packets sent by or received for a
2005.Ar user .
2006A
2007.Ar user
2008may be matched by name or identification number.
2009.It Cm verrevpath
2010For incoming packets,
2011a routing table lookup is done on the packet's source address.
2012If the interface on which the packet entered the system matches the
2013outgoing interface for the route,
2014the packet matches.
2015If the interfaces do not match up,
2016the packet does not match.
2017All outgoing packets or packets with no incoming interface match.
2018.Pp
2019The name and functionality of the option is intentionally similar to
2020the Cisco IOS command:
2021.Pp
2022.Dl ip verify unicast reverse-path
2023.Pp
2024This option can be used to make anti-spoofing rules to reject all
2025packets with source addresses not from this interface.
2026See also the option
2027.Cm antispoof .
2028.It Cm versrcreach
2029For incoming packets,
2030a routing table lookup is done on the packet's source address.
2031If a route to the source address exists, but not the default route
2032or a blackhole/reject route, the packet matches.
2033Otherwise, the packet does not match.
2034All outgoing packets match.
2035.Pp
2036The name and functionality of the option is intentionally similar to
2037the Cisco IOS command:
2038.Pp
2039.Dl ip verify unicast source reachable-via any
2040.Pp
2041This option can be used to make anti-spoofing rules to reject all
2042packets whose source address is unreachable.
2043.It Cm antispoof
2044For incoming packets, the packet's source address is checked if it
2045belongs to a directly connected network.
2046If the network is directly connected, then the interface the packet
2047came on in is compared to the interface the network is connected to.
2048When incoming interface and directly connected interface are not the
2049same, the packet does not match.
2050Otherwise, the packet does match.
2051All outgoing packets match.
2052.Pp
2053This option can be used to make anti-spoofing rules to reject all
2054packets that pretend to be from a directly connected network but do
2055not come in through that interface.
2056This option is similar to but more restricted than
2057.Cm verrevpath
2058because it engages only on packets with source addresses of directly
2059connected networks instead of all source addresses.
2060.El
2061.Sh LOOKUP TABLES
2062Lookup tables are useful to handle large sparse sets of
2063addresses or other search keys (e.g., ports, jail IDs, interface names).
2064In the rest of this section we will use the term ``key''.
2065Table name needs to match the following spec:
2066.Ar table-name .
2067Tables with the same name can be created in different
2068.Ar sets .
2069However, rule links to the tables in
2070.Ar set 0
2071by default.
2072This behavior can be controlled by
2073.Va net.inet.ip.fw.tables_sets
2074variable.
2075See the
2076.Sx SETS OF RULES
2077section for more information.
2078There may be up to 65535 different lookup tables.
2079.Pp
2080The following table types are supported:
2081.Bl -tag -width indent
2082.It Ar table-type : Ar addr | iface | number | flow
2083.It Ar table-key : Ar addr Ns Oo / Ns Ar masklen Oc | iface-name | number | flow-spec
2084.It Ar flow-spec : Ar flow-field Ns Op , Ns Ar flow-spec
2085.It Ar flow-field : src-ip | proto | src-port | dst-ip | dst-port
2086.It Cm addr
2087matches IPv4 or IPv6 address.
2088Each entry is represented by an
2089.Ar addr Ns Op / Ns Ar masklen
2090and will match all addresses with base
2091.Ar addr
2092(specified as an IPv4/IPv6 address, or a hostname) and mask width of
2093.Ar masklen
2094bits.
2095If
2096.Ar masklen
2097is not specified, it defaults to 32 for IPv4 and 128 for IPv6.
2098When looking up an IP address in a table, the most specific
2099entry will match.
2100.It Cm iface
2101matches interface names.
2102Each entry is represented by string treated as interface name.
2103Wildcards are not supported.
2104.It Cm number
2105maches protocol ports, uids/gids or jail IDs.
2106Each entry is represented by 32-bit unsigned integer.
2107Ranges are not supported.
2108.It Cm flow
2109Matches packet fields specified by
2110.Ar flow
2111type suboptions with table entries.
2112.El
2113.Pp
2114Tables require explicit creation via
2115.Cm create
2116before use.
2117.Pp
2118The following creation options are supported:
2119.Bl -tag -width indent
2120.It Ar create-options : Ar create-option | create-options
2121.It Ar create-option : Cm type Ar table-type | Cm valtype Ar value-mask | Cm algo Ar algo-desc |
2122.Cm limit Ar number | Cm locked
2123.It Cm type
2124Table key type.
2125.It Cm valtype
2126Table value mask.
2127.It Cm algo
2128Table algorithm to use (see below).
2129.It Cm limit
2130Maximum number of items that may be inserted into table.
2131.It Cm locked
2132Restrict any table modifications.
2133.El
2134.Pp
2135Some of these options may be modified later via
2136.Cm modify
2137keyword.
2138The following options can be changed:
2139.Bl -tag -width indent
2140.It Ar modify-options : Ar modify-option | modify-options
2141.It Ar modify-option : Cm limit Ar number
2142.It Cm limit
2143Alter maximum number of items that may be inserted into table.
2144.El
2145.Pp
2146Additionally, table can be locked or unlocked using
2147.Cm lock
2148or
2149.Cm unlock
2150commands.
2151.Pp
2152Tables of the same
2153.Ar type
2154can be swapped with each other using
2155.Cm swap Ar name
2156command.
2157Swap may fail if tables limits are set and data exchange
2158would result in limits hit.
2159Operation is performed atomically.
2160.Pp
2161One or more entries can be added to a table at once using
2162.Cm add
2163command.
2164Addition of all items are performed atomically.
2165By default, error in addition of one entry does not influence
2166addition of other entries. However, non-zero error code is returned
2167in that case.
2168Special
2169.Cm atomic
2170keyword may be specified before
2171.Cm add
2172to indicate all-or-none add request.
2173.Pp
2174One or more entries can be removed from a table at once using
2175.Cm delete
2176command.
2177By default, error in removal of one entry does not influence
2178removing of other entries. However, non-zero error code is returned
2179in that case.
2180.Pp
2181It may be possible to check what entry will be found on particular
2182.Ar table-key
2183using
2184.Cm lookup
2185.Ar table-key
2186command.
2187This functionality is optional and may be unsupported in some algorithms.
2188.Pp
2189The following operations can be performed on
2190.Ar one
2191or
2192.Cm all
2193tables:
2194.Bl -tag -width indent
2195.It Cm list
2196List all entries.
2197.It Cm flush
2198Removes all entries.
2199.It Cm info
2200Shows generic table information.
2201.It Cm detail
2202Shows generic table information and algo-specific data.
2203.El
2204.Pp
2205The following lookup algorithms are supported:
2206.Bl -tag -width indent
2207.It Ar algo-desc : algo-name | "algo-name algo-data"
2208.It Ar algo-name: Ar addr:radix | addr:hash | iface:array | number:array | flow:hash
2209.It Cm addr:radix
2210Separate Radix trees for IPv4 and IPv6, the same way as the routing table (see
2211.Xr route 4 ) .
2212Default choice for
2213.Ar addr
2214type.
2215.It Cm addr:hash
2216Separate auto-growing hashes for IPv4 and IPv6.
2217Accepts entries with the same mask length specified initially via
2218.Cm "addr:hash masks=/v4,/v6"
2219algorithm creation options.
2220Assume /32 and /128 masks by default.
2221Search removes host bits (according to mask) from supplied address and checks
2222resulting key in appropriate hash.
2223Mostly optimized for /64 and byte-ranged IPv6 masks.
2224.It Cm iface:array
2225Array storing sorted indexes for entries which are presented in the system.
2226Optimized for very fast lookup.
2227.It Cm number:array
2228Array storing sorted u32 numbers.
2229.It Cm flow:hash
2230Auto-growing hash storing flow entries.
2231Search calculates hash on required packet fields and searches for matching
2232entries in selected bucket.
2233.El
2234.Pp
2235The
2236.Cm tablearg
2237feature provides the ability to use a value, looked up in the table, as
2238the argument for a rule action, action parameter or rule option.
2239This can significantly reduce number of rules in some configurations.
2240If two tables are used in a rule, the result of the second (destination)
2241is used.
2242.Pp
2243Each record may hold one or more values according to
2244.Ar value-mask .
2245This mask is set on table creation via
2246.Cm valtype
2247option.
2248The following value types are supported:
2249.Bl -tag -width indent
2250.It Ar value-mask : Ar value-type Ns Op , Ns Ar value-mask
2251.It Ar value-type : Ar skipto | pipe | fib | nat | dscp | tag | divert |
2252.Ar netgraph | limit | ipv4
2253.It Cm skipto
2254rule number to jump to.
2255.It Cm pipe
2256Pipe number to use.
2257.It Cm fib
2258fib number to match/set.
2259.It Cm nat
2260nat number to jump to.
2261.It Cm dscp
2262dscp value to match/set.
2263.It Cm tag
2264tag number to match/set.
2265.It Cm divert
2266port number to divert traffic to.
2267.It Cm netgraph
2268hook number to move packet to.
2269.It Cm limit
2270maximum number of connections.
2271.It Cm ipv4
2272IPv4 nexthop to fwd packets to.
2273.It Cm ipv6
2274IPv6 nexthop to fwd packets to.
2275.El
2276.Pp
2277The
2278.Cm tablearg
2279argument can be used with the following actions:
2280.Cm nat, pipe , queue, divert, tee, netgraph, ngtee, fwd, skipto, setfib,
2281action parameters:
2282.Cm tag, untag,
2283rule options:
2284.Cm limit, tagged.
2285.Pp
2286When used with the
2287.Cm skipto
2288action, the user should be aware that the code will walk the ruleset
2289up to a rule equal to, or past, the given number.
2290.Pp
2291See the
2292.Sx EXAMPLES
2293Section for example usage of tables and the tablearg keyword.
2294.Sh SETS OF RULES
2295Each rule or table belongs to one of 32 different
2296.Em sets
2297, numbered 0 to 31.
2298Set 31 is reserved for the default rule.
2299.Pp
2300By default, rules or tables are put in set 0, unless you use the
2301.Cm set N
2302attribute when adding a new rule or table.
2303Sets can be individually and atomically enabled or disabled,
2304so this mechanism permits an easy way to store multiple configurations
2305of the firewall and quickly (and atomically) switch between them.
2306.Pp
2307By default, tables from set 0 are referenced when adding rule with
2308table opcodes regardless of rule set.
2309This behavior can be changed by setting
2310.Va net.inet.ip.fw.tables_sets
2311variable to 1.
2312Rule's set will then be used for table references.
2313.Pp
2314The command to enable/disable sets is
2315.Bd -ragged -offset indent
2316.Nm
2317.Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
2318.Ed
2319.Pp
2320where multiple
2321.Cm enable
2322or
2323.Cm disable
2324sections can be specified.
2325Command execution is atomic on all the sets specified in the command.
2326By default, all sets are enabled.
2327.Pp
2328When you disable a set, its rules behave as if they do not exist
2329in the firewall configuration, with only one exception:
2330.Bd -ragged -offset indent
2331dynamic rules created from a rule before it had been disabled
2332will still be active until they expire.
2333In order to delete
2334dynamic rules you have to explicitly delete the parent rule
2335which generated them.
2336.Ed
2337.Pp
2338The set number of rules can be changed with the command
2339.Bd -ragged -offset indent
2340.Nm
2341.Cm set move
2342.Brq Cm rule Ar rule-number | old-set
2343.Cm to Ar new-set
2344.Ed
2345.Pp
2346Also, you can atomically swap two rulesets with the command
2347.Bd -ragged -offset indent
2348.Nm
2349.Cm set swap Ar first-set second-set
2350.Ed
2351.Pp
2352See the
2353.Sx EXAMPLES
2354Section on some possible uses of sets of rules.
2355.Sh STATEFUL FIREWALL
2356Stateful operation is a way for the firewall to dynamically
2357create rules for specific flows when packets that
2358match a given pattern are detected.
2359Support for stateful
2360operation comes through the
2361.Cm check-state , keep-state , record-state , limit
2362and
2363.Cm set-limit
2364options of
2365.Nm rules .
2366.Pp
2367Dynamic rules are created when a packet matches a
2368.Cm keep-state ,
2369.Cm record-state ,
2370.Cm limit
2371or
2372.Cm set-limit
2373rule, causing the creation of a
2374.Em dynamic
2375rule which will match all and only packets with
2376a given
2377.Em protocol
2378between a
2379.Em src-ip/src-port dst-ip/dst-port
2380pair of addresses
2381.Em ( src
2382and
2383.Em dst
2384are used here only to denote the initial match addresses, but they
2385are completely equivalent afterwards).
2386Rules created by
2387.Cm keep-state
2388option also have a
2389.Ar :flowname
2390taken from it.
2391This name is used in matching together with addresses, ports and protocol.
2392Dynamic rules will be checked at the first
2393.Cm check-state, keep-state
2394or
2395.Cm limit
2396occurrence, and the action performed upon a match will be the same
2397as in the parent rule.
2398.Pp
2399Note that no additional attributes other than protocol and IP addresses
2400and ports and :flowname are checked on dynamic rules.
2401.Pp
2402The typical use of dynamic rules is to keep a closed firewall configuration,
2403but let the first TCP SYN packet from the inside network install a
2404dynamic rule for the flow so that packets belonging to that session
2405will be allowed through the firewall:
2406.Pp
2407.Dl "ipfw add check-state :OUTBOUND"
2408.Dl "ipfw add allow tcp from my-subnet to any setup keep-state :OUTBOUND"
2409.Dl "ipfw add deny tcp from any to any"
2410.Pp
2411A similar approach can be used for UDP, where an UDP packet coming
2412from the inside will install a dynamic rule to let the response through
2413the firewall:
2414.Pp
2415.Dl "ipfw add check-state :OUTBOUND"
2416.Dl "ipfw add allow udp from my-subnet to any keep-state :OUTBOUND"
2417.Dl "ipfw add deny udp from any to any"
2418.Pp
2419Dynamic rules expire after some time, which depends on the status
2420of the flow and the setting of some
2421.Cm sysctl
2422variables.
2423See Section
2424.Sx SYSCTL VARIABLES
2425for more details.
2426For TCP sessions, dynamic rules can be instructed to periodically
2427send keepalive packets to refresh the state of the rule when it is
2428about to expire.
2429.Pp
2430See Section
2431.Sx EXAMPLES
2432for more examples on how to use dynamic rules.
2433.Sh TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
2434.Nm
2435is also the user interface for the
2436.Nm dummynet
2437traffic shaper, packet scheduler and network emulator, a subsystem that
2438can artificially queue, delay or drop packets
2439emulating the behaviour of certain network links
2440or queueing systems.
2441.Pp
2442.Nm dummynet
2443operates by first using the firewall to select packets
2444using any match pattern that can be used in
2445.Nm
2446rules.
2447Matching packets are then passed to either of two
2448different objects, which implement the traffic regulation:
2449.Bl -hang -offset XXXX
2450.It Em pipe
2451A
2452.Em pipe
2453emulates a
2454.Em link
2455with given bandwidth and propagation delay,
2456driven by a FIFO scheduler and a single queue with programmable
2457queue size and packet loss rate.
2458Packets are appended to the queue as they come out from
2459.Nm ipfw ,
2460and then transferred in FIFO order to the link at the desired rate.
2461.It Em queue
2462A
2463.Em queue
2464is an abstraction used to implement packet scheduling
2465using one of several packet scheduling algorithms.
2466Packets sent to a
2467.Em queue
2468are first grouped into flows according to a mask on the 5-tuple.
2469Flows are then passed to the scheduler associated to the
2470.Em queue ,
2471and each flow uses scheduling parameters (weight and others)
2472as configured in the
2473.Em queue
2474itself.
2475A scheduler in turn is connected to an emulated link,
2476and arbitrates the link's bandwidth among backlogged flows according to
2477weights and to the features of the scheduling algorithm in use.
2478.El
2479.Pp
2480In practice,
2481.Em pipes
2482can be used to set hard limits to the bandwidth that a flow can use, whereas
2483.Em queues
2484can be used to determine how different flows share the available bandwidth.
2485.Pp
2486A graphical representation of the binding of queues,
2487flows, schedulers and links is below.
2488.Bd -literal -offset indent
2489                 (flow_mask|sched_mask)  sched_mask
2490         +---------+   weight Wx  +-------------+
2491         |         |->-[flow]-->--|             |-+
2492    -->--| QUEUE x |   ...        |             | |
2493         |         |->-[flow]-->--| SCHEDuler N | |
2494         +---------+              |             | |
2495             ...                  |             +--[LINK N]-->--
2496         +---------+   weight Wy  |             | +--[LINK N]-->--
2497         |         |->-[flow]-->--|             | |
2498    -->--| QUEUE y |   ...        |             | |
2499         |         |->-[flow]-->--|             | |
2500         +---------+              +-------------+ |
2501                                    +-------------+
2502.Ed
2503It is important to understand the role of the SCHED_MASK
2504and FLOW_MASK, which are configured through the commands
2505.Dl "ipfw sched N config mask SCHED_MASK ..."
2506and
2507.Dl "ipfw queue X config mask FLOW_MASK ..." .
2508.Pp
2509The SCHED_MASK is used to assign flows to one or more
2510scheduler instances, one for each
2511value of the packet's 5-tuple after applying SCHED_MASK.
2512As an example, using ``src-ip 0xffffff00'' creates one instance
2513for each /24 destination subnet.
2514.Pp
2515The FLOW_MASK, together with the SCHED_MASK, is used to split
2516packets into flows.
2517As an example, using
2518``src-ip 0x000000ff''
2519together with the previous SCHED_MASK makes a flow for
2520each individual source address.
2521In turn, flows for each /24
2522subnet will be sent to the same scheduler instance.
2523.Pp
2524The above diagram holds even for the
2525.Em pipe
2526case, with the only restriction that a
2527.Em pipe
2528only supports a SCHED_MASK, and forces the use of a FIFO
2529scheduler (these are for backward compatibility reasons;
2530in fact, internally, a
2531.Nm dummynet's
2532pipe is implemented exactly as above).
2533.Pp
2534There are two modes of
2535.Nm dummynet
2536operation:
2537.Dq normal
2538and
2539.Dq fast .
2540The
2541.Dq normal
2542mode tries to emulate a real link: the
2543.Nm dummynet
2544scheduler ensures that the packet will not leave the pipe faster than it
2545would on the real link with a given bandwidth.
2546The
2547.Dq fast
2548mode allows certain packets to bypass the
2549.Nm dummynet
2550scheduler (if packet flow does not exceed pipe's bandwidth).
2551This is the reason why the
2552.Dq fast
2553mode requires less CPU cycles per packet (on average) and packet latency
2554can be significantly lower in comparison to a real link with the same
2555bandwidth.
2556The default mode is
2557.Dq normal .
2558The
2559.Dq fast
2560mode can be enabled by setting the
2561.Va net.inet.ip.dummynet.io_fast
2562.Xr sysctl 8
2563variable to a non-zero value.
2564.Pp
2565.Ss PIPE, QUEUE AND SCHEDULER CONFIGURATION
2566The
2567.Em pipe ,
2568.Em queue
2569and
2570.Em scheduler
2571configuration commands are the following:
2572.Bd -ragged -offset indent
2573.Cm pipe Ar number Cm config Ar pipe-configuration
2574.Pp
2575.Cm queue Ar number Cm config Ar queue-configuration
2576.Pp
2577.Cm sched Ar number Cm config Ar sched-configuration
2578.Ed
2579.Pp
2580The following parameters can be configured for a pipe:
2581.Pp
2582.Bl -tag -width indent -compact
2583.It Cm bw Ar bandwidth | device
2584Bandwidth, measured in
2585.Sm off
2586.Op Cm K | M | G
2587.Brq Cm bit/s | Byte/s .
2588.Sm on
2589.Pp
2590A value of 0 (default) means unlimited bandwidth.
2591The unit must immediately follow the number, as in
2592.Pp
2593.Dl "ipfw pipe 1 config bw 300Kbit/s"
2594.Pp
2595If a device name is specified instead of a numeric value, as in
2596.Pp
2597.Dl "ipfw pipe 1 config bw tun0"
2598.Pp
2599then the transmit clock is supplied by the specified device.
2600At the moment only the
2601.Xr tun 4
2602device supports this
2603functionality, for use in conjunction with
2604.Xr ppp 8 .
2605.Pp
2606.It Cm delay Ar ms-delay
2607Propagation delay, measured in milliseconds.
2608The value is rounded to the next multiple of the clock tick
2609(typically 10ms, but it is a good practice to run kernels
2610with
2611.Dq "options HZ=1000"
2612to reduce
2613the granularity to 1ms or less).
2614The default value is 0, meaning no delay.
2615.Pp
2616.It Cm burst Ar size
2617If the data to be sent exceeds the pipe's bandwidth limit
2618(and the pipe was previously idle), up to
2619.Ar size
2620bytes of data are allowed to bypass the
2621.Nm dummynet
2622scheduler, and will be sent as fast as the physical link allows.
2623Any additional data will be transmitted at the rate specified
2624by the
2625.Nm pipe
2626bandwidth.
2627The burst size depends on how long the pipe has been idle;
2628the effective burst size is calculated as follows:
2629MAX(
2630.Ar size
2631,
2632.Nm bw
2633* pipe_idle_time).
2634.Pp
2635.It Cm profile Ar filename
2636A file specifying the additional overhead incurred in the transmission
2637of a packet on the link.
2638.Pp
2639Some link types introduce extra delays in the transmission
2640of a packet, e.g., because of MAC level framing, contention on
2641the use of the channel, MAC level retransmissions and so on.
2642From our point of view, the channel is effectively unavailable
2643for this extra time, which is constant or variable depending
2644on the link type.
2645Additionally, packets may be dropped after this
2646time (e.g., on a wireless link after too many retransmissions).
2647We can model the additional delay with an empirical curve
2648that represents its distribution.
2649.Bd -literal -offset indent
2650      cumulative probability
2651      1.0 ^
2652          |
2653      L   +-- loss-level          x
2654          |                 ******
2655          |                *
2656          |           *****
2657          |          *
2658          |        **
2659          |       *
2660          +-------*------------------->
2661                      delay
2662.Ed
2663The empirical curve may have both vertical and horizontal lines.
2664Vertical lines represent constant delay for a range of
2665probabilities.
2666Horizontal lines correspond to a discontinuity in the delay
2667distribution: the pipe will use the largest delay for a
2668given probability.
2669.Pp
2670The file format is the following, with whitespace acting as
2671a separator and '#' indicating the beginning a comment:
2672.Bl -tag -width indent
2673.It Cm name Ar identifier
2674optional name (listed by "ipfw pipe show")
2675to identify the delay distribution;
2676.It Cm bw Ar value
2677the bandwidth used for the pipe.
2678If not specified here, it must be present
2679explicitly as a configuration parameter for the pipe;
2680.It Cm loss-level Ar L
2681the probability above which packets are lost.
2682(0.0 <= L <= 1.0, default 1.0 i.e., no loss);
2683.It Cm samples Ar N
2684the number of samples used in the internal
2685representation of the curve (2..1024; default 100);
2686.It Cm "delay prob" | "prob delay"
2687One of these two lines is mandatory and defines
2688the format of the following lines with data points.
2689.It Ar XXX Ar YYY
26902 or more lines representing points in the curve,
2691with either delay or probability first, according
2692to the chosen format.
2693The unit for delay is milliseconds.
2694Data points do not need to be sorted.
2695Also, the number of actual lines can be different
2696from the value of the "samples" parameter:
2697.Nm
2698utility will sort and interpolate
2699the curve as needed.
2700.El
2701.Pp
2702Example of a profile file:
2703.Bd -literal -offset indent
2704name    bla_bla_bla
2705samples 100
2706loss-level    0.86
2707prob    delay
27080       200	# minimum overhead is 200ms
27090.5     200
27100.5     300
27110.8     1000
27120.9     1300
27131       1300
2714#configuration file end
2715.Ed
2716.El
2717.Pp
2718The following parameters can be configured for a queue:
2719.Pp
2720.Bl -tag -width indent -compact
2721.It Cm pipe Ar pipe_nr
2722Connects a queue to the specified pipe.
2723Multiple queues (with the same or different weights) can be connected to
2724the same pipe, which specifies the aggregate rate for the set of queues.
2725.Pp
2726.It Cm weight Ar weight
2727Specifies the weight to be used for flows matching this queue.
2728The weight must be in the range 1..100, and defaults to 1.
2729.El
2730.Pp
2731The following case-insensitive parameters can be configured for a
2732scheduler:
2733.Pp
2734.Bl -tag -width indent -compact
2735.It Cm type Ar {fifo | wf2q+ | rr | qfq | fq_codel | fq_pie}
2736specifies the scheduling algorithm to use.
2737.Bl -tag -width indent -compact
2738.It Cm fifo
2739is just a FIFO scheduler (which means that all packets
2740are stored in the same queue as they arrive to the scheduler).
2741FIFO has O(1) per-packet time complexity, with very low
2742constants (estimate 60-80ns on a 2GHz desktop machine)
2743but gives no service guarantees.
2744.It Cm wf2q+
2745implements the WF2Q+ algorithm, which is a Weighted Fair Queueing
2746algorithm which permits flows to share bandwidth according to
2747their weights.
2748Note that weights are not priorities; even a flow
2749with a minuscule weight will never starve.
2750WF2Q+ has O(log N) per-packet processing cost, where N is the number
2751of flows, and is the default algorithm used by previous versions
2752dummynet's queues.
2753.It Cm rr
2754implements the Deficit Round Robin algorithm, which has O(1) processing
2755costs (roughly, 100-150ns per packet)
2756and permits bandwidth allocation according to weights, but
2757with poor service guarantees.
2758.It Cm qfq
2759implements the QFQ algorithm, which is a very fast variant of
2760WF2Q+, with similar service guarantees and O(1) processing
2761costs (roughly, 200-250ns per packet).
2762.It Cm fq_codel
2763implements the FQ-CoDel (FlowQueue-CoDel) scheduler/AQM algorithm, which
2764uses a modified Deficit Round Robin scheduler to manage two lists of sub-queues
2765(old sub-queues and new sub-queues) for providing brief periods of priority to
2766lightweight or short burst flows.
2767By default, the total number of sub-queues is 1024.
2768FQ-CoDel's internal, dynamically
2769created sub-queues are controlled by separate instances of CoDel AQM.
2770.It Cm fq_pie
2771implements the FQ-PIE (FlowQueue-PIE) scheduler/AQM algorithm, which similar to
2772.Cm fq_codel
2773but uses per sub-queue PIE AQM instance to control the queue delay.
2774.El
2775.Pp
2776.Cm fq_codel
2777inherits AQM parameters and options from
2778.Cm codel
2779(see below), and
2780.Cm fq_pie
2781inherits AQM parameters and options from
2782.Cm pie
2783(see below).
2784Additionally, both of
2785.Cm fq_codel
2786and
2787.Cm fq_pie
2788have shared scheduler parameters which are:
2789.Bl -tag -width indent
2790.It Cm quantum
2791.Ar m
2792specifies the quantum (credit) of the scheduler.
2793.Ar m
2794is the number of bytes a queue can serve before being moved to the tail
2795of old queues list.
2796The default is 1514 bytes, and the maximum accepable value
2797is 9000 bytes.
2798.It Cm limit
2799.Ar m
2800specifies the hard size limit (in unit of packets) of all queues managed by an
2801instance of the scheduler.
2802The default value of
2803.Ar m
2804is 10240 packets, and the maximum accepable value is 20480 packets.
2805.It Cm flows
2806.Ar m
2807specifies the total number of flow queues (sub-queues) that fq_*
2808creates and manages.
2809By default, 1024 sub-queues are created when an instance
2810of the fq_{codel/pie} scheduler is created.
2811The maximum accepable value is
281265536.
2813.El
2814.Pp
2815Note that any token after
2816.Cm fq_codel
2817or
2818.Cm fq_pie
2819is considered a parameter for fq_{codel/pie}.
2820So, ensure all scheduler
2821configuration options not related to fq_{codel/pie} are written before
2822.Cm fq_codel/fq_pie
2823tokens.
2824.El
2825.Pp
2826In addition to the type, all parameters allowed for a pipe can also
2827be specified for a scheduler.
2828.Pp
2829Finally, the following parameters can be configured for both
2830pipes and queues:
2831.Pp
2832.Bl -tag -width XXXX -compact
2833.It Cm buckets Ar hash-table-size
2834Specifies the size of the hash table used for storing the
2835various queues.
2836Default value is 64 controlled by the
2837.Xr sysctl 8
2838variable
2839.Va net.inet.ip.dummynet.hash_size ,
2840allowed range is 16 to 65536.
2841.Pp
2842.It Cm mask Ar mask-specifier
2843Packets sent to a given pipe or queue by an
2844.Nm
2845rule can be further classified into multiple flows, each of which is then
2846sent to a different
2847.Em dynamic
2848pipe or queue.
2849A flow identifier is constructed by masking the IP addresses,
2850ports and protocol types as specified with the
2851.Cm mask
2852options in the configuration of the pipe or queue.
2853For each different flow identifier, a new pipe or queue is created
2854with the same parameters as the original object, and matching packets
2855are sent to it.
2856.Pp
2857Thus, when
2858.Em dynamic pipes
2859are used, each flow will get the same bandwidth as defined by the pipe,
2860whereas when
2861.Em dynamic queues
2862are used, each flow will share the parent's pipe bandwidth evenly
2863with other flows generated by the same queue (note that other queues
2864with different weights might be connected to the same pipe).
2865.br
2866Available mask specifiers are a combination of one or more of the following:
2867.Pp
2868.Cm dst-ip Ar mask ,
2869.Cm dst-ip6 Ar mask ,
2870.Cm src-ip Ar mask ,
2871.Cm src-ip6 Ar mask ,
2872.Cm dst-port Ar mask ,
2873.Cm src-port Ar mask ,
2874.Cm flow-id Ar mask ,
2875.Cm proto Ar mask
2876or
2877.Cm all ,
2878.Pp
2879where the latter means all bits in all fields are significant.
2880.Pp
2881.It Cm noerror
2882When a packet is dropped by a
2883.Nm dummynet
2884queue or pipe, the error
2885is normally reported to the caller routine in the kernel, in the
2886same way as it happens when a device queue fills up.
2887Setting this
2888option reports the packet as successfully delivered, which can be
2889needed for some experimental setups where you want to simulate
2890loss or congestion at a remote router.
2891.Pp
2892.It Cm plr Ar packet-loss-rate
2893Packet loss rate.
2894Argument
2895.Ar packet-loss-rate
2896is a floating-point number between 0 and 1, with 0 meaning no
2897loss, 1 meaning 100% loss.
2898The loss rate is internally represented on 31 bits.
2899.Pp
2900.It Cm queue Brq Ar slots | size Ns Cm Kbytes
2901Queue size, in
2902.Ar slots
2903or
2904.Cm KBytes .
2905Default value is 50 slots, which
2906is the typical queue size for Ethernet devices.
2907Note that for slow speed links you should keep the queue
2908size short or your traffic might be affected by a significant
2909queueing delay.
2910E.g., 50 max-sized ethernet packets (1500 bytes) mean 600Kbit
2911or 20s of queue on a 30Kbit/s pipe.
2912Even worse effects can result if you get packets from an
2913interface with a much larger MTU, e.g.\& the loopback interface
2914with its 16KB packets.
2915The
2916.Xr sysctl 8
2917variables
2918.Em net.inet.ip.dummynet.pipe_byte_limit
2919and
2920.Em net.inet.ip.dummynet.pipe_slot_limit
2921control the maximum lengths that can be specified.
2922.Pp
2923.It Cm red | gred Ar w_q Ns / Ns Ar min_th Ns / Ns Ar max_th Ns / Ns Ar max_p
2924[ecn]
2925Make use of the RED (Random Early Detection) queue management algorithm.
2926.Ar w_q
2927and
2928.Ar max_p
2929are floating
2930point numbers between 0 and 1 (inclusive), while
2931.Ar min_th
2932and
2933.Ar max_th
2934are integer numbers specifying thresholds for queue management
2935(thresholds are computed in bytes if the queue has been defined
2936in bytes, in slots otherwise).
2937The two parameters can also be of the same value if needed. The
2938.Nm dummynet
2939also supports the gentle RED variant (gred) and ECN (Explicit Congestion
2940Notification) as optional. Three
2941.Xr sysctl 8
2942variables can be used to control the RED behaviour:
2943.Bl -tag -width indent
2944.It Va net.inet.ip.dummynet.red_lookup_depth
2945specifies the accuracy in computing the average queue
2946when the link is idle (defaults to 256, must be greater than zero)
2947.It Va net.inet.ip.dummynet.red_avg_pkt_size
2948specifies the expected average packet size (defaults to 512, must be
2949greater than zero)
2950.It Va net.inet.ip.dummynet.red_max_pkt_size
2951specifies the expected maximum packet size, only used when queue
2952thresholds are in bytes (defaults to 1500, must be greater than zero).
2953.El
2954.Pp
2955.It Cm codel Oo Cm target Ar time Oc Oo Cm interval Ar time Oc Oo Cm ecn |
2956.Cm noecn Oc
2957Make use of the CoDel (Controlled-Delay) queue management algorithm.
2958.Ar time
2959is interpreted as milliseconds by default but seconds (s), milliseconds (ms) or
2960microseconds (us) can be specified instead.
2961CoDel drops or marks (ECN) packets
2962depending on packet sojourn time in the queue.
2963.Cm target
2964.Ar time
2965(5ms by default) is the minimum acceptable persistent queue delay that CoDel
2966allows.
2967CoDel does not drop packets directly after packets sojourn time becomes
2968higher than
2969.Cm target
2970.Ar time
2971but waits for
2972.Cm interval
2973.Ar time
2974(100ms default) before dropping.
2975.Cm interval
2976.Ar time
2977should be set to maximum RTT for all expected connections.
2978.Cm ecn
2979enables (disabled by default) packet marking (instead of dropping) for
2980ECN-enabled TCP flows when queue delay becomes high.
2981.Pp
2982Note that any token after
2983.Cm codel
2984is considered a parameter for CoDel.
2985So, ensure all pipe/queue
2986configuration options are written before
2987.Cm codel
2988token.
2989.Pp
2990The
2991.Xr sysctl 8
2992variables
2993.Va net.inet.ip.dummynet.codel.target
2994and
2995.Va net.inet.ip.dummynet.codel.interval
2996can be used to set CoDel default parameters.
2997.Pp
2998.It Cm pie Oo Cm target Ar time Oc Oo Cm tupdate Ar time Oc Oo
2999.Cm alpha Ar n Oc Oo Cm beta Ar n Oc Oo Cm max_burst Ar time Oc Oo
3000.Cm max_ecnth Ar n Oc Oo Cm ecn | Cm noecn Oc Oo Cm capdrop |
3001.Cm nocapdrop Oc Oo Cm drand | Cm nodrand Oc Oo Cm onoff
3002.Oc Oo Cm dre | Cm ts Oc
3003Make use of the PIE (Proportional Integral controller Enhanced) queue management
3004algorithm.
3005PIE drops or marks packets depending on a calculated drop probability during
3006en-queue process, with the aim of achieving high throughput while keeping queue
3007delay low.
3008At regular time intervals of
3009.Cm tupdate
3010.Ar time
3011(15ms by default) a background process (re)calculates the probability based on queue delay
3012deviations from
3013.Cm target
3014.Ar time
3015(15ms by default) and queue delay trends.
3016PIE approximates current queue
3017delay by using a departure rate estimation method, or (optionally) by using a
3018packet timestamp method similar to CoDel.
3019.Ar time
3020is interpreted as milliseconds by default but seconds (s), milliseconds (ms) or
3021microseconds (us) can be specified instead.
3022The other PIE parameters and options are as follows:
3023.Bl -tag -width indent
3024.It Cm alpha Ar n
3025.Ar n
3026is a floating point number between 0 and 7 which specifies the weight of queue
3027delay deviations that is used in drop probability calculation.
30280.125 is the default.
3029.It Cm beta Ar n
3030.Ar n
3031is a floating point number between 0 and 7 which specifies is the weight of queue
3032delay trend that is used in drop probability calculation.
30331.25 is the default.
3034.It Cm max_burst Ar time
3035The maximum period of time that PIE does not drop/mark packets.
3036150ms is the
3037default and 10s is the maximum value.
3038.It Cm max_ecnth Ar n
3039Even when ECN is enabled, PIE drops packets instead of marking them when drop
3040probability becomes higher than ECN probability threshold
3041.Cm max_ecnth Ar n
3042, the default is 0.1 (i.e 10%) and 1 is the maximum value.
3043.It Cm ecn | noecn
3044enable or disable ECN marking for ECN-enabled TCP flows.
3045Disabled by default.
3046.It Cm capdrop | nocapdrop
3047enable or disable cap drop adjustment.
3048Cap drop adjustment is enabled by default.
3049.It Cm drand | nodrand
3050enable or disable drop probability de-randomisation.
3051De-randomisation eliminates
3052the problem of dropping packets too close or too far.
3053De-randomisation is enabled by default.
3054.It Cm onoff
3055enable turning PIE on and off depending on queue load.
3056If this option is enabled,
3057PIE turnes on when over 1/3 of queue becomes full.
3058This option is disabled by
3059default.
3060.It Cm dre | ts
3061Calculate queue delay using departure rate estimation
3062.Cm dre
3063or timestamps
3064.Cm ts .
3065.Cm dre
3066is used by default.
3067.El
3068.Pp
3069Note that any token after
3070.Cm pie
3071is considered a parameter for PIE.
3072So ensure all pipe/queue
3073the configuration options are written before
3074.Cm pie
3075token.
3076.Xr sysctl 8
3077variables can be used to control the
3078.Cm pie
3079default parameters.
3080See the
3081.Sx SYSCTL VARIABLES
3082section for more details.
3083.El
3084.Pp
3085When used with IPv6 data,
3086.Nm dummynet
3087currently has several limitations.
3088Information necessary to route link-local packets to an
3089interface is not available after processing by
3090.Nm dummynet
3091so those packets are dropped in the output path.
3092Care should be taken to ensure that link-local packets are not passed to
3093.Nm dummynet .
3094.Sh CHECKLIST
3095Here are some important points to consider when designing your
3096rules:
3097.Bl -bullet
3098.It
3099Remember that you filter both packets going
3100.Cm in
3101and
3102.Cm out .
3103Most connections need packets going in both directions.
3104.It
3105Remember to test very carefully.
3106It is a good idea to be near the console when doing this.
3107If you cannot be near the console,
3108use an auto-recovery script such as the one in
3109.Pa /usr/share/examples/ipfw/change_rules.sh .
3110.It
3111Do not forget the loopback interface.
3112.El
3113.Sh FINE POINTS
3114.Bl -bullet
3115.It
3116There are circumstances where fragmented datagrams are unconditionally
3117dropped.
3118TCP packets are dropped if they do not contain at least 20 bytes of
3119TCP header, UDP packets are dropped if they do not contain a full 8
3120byte UDP header, and ICMP packets are dropped if they do not contain
31214 bytes of ICMP header, enough to specify the ICMP type, code, and
3122checksum.
3123These packets are simply logged as
3124.Dq pullup failed
3125since there may not be enough good data in the packet to produce a
3126meaningful log entry.
3127.It
3128Another type of packet is unconditionally dropped, a TCP packet with a
3129fragment offset of one.
3130This is a valid packet, but it only has one use, to try
3131to circumvent firewalls.
3132When logging is enabled, these packets are
3133reported as being dropped by rule -1.
3134.It
3135If you are logged in over a network, loading the
3136.Xr kld 4
3137version of
3138.Nm
3139is probably not as straightforward as you would think.
3140The following command line is recommended:
3141.Bd -literal -offset indent
3142kldload ipfw && \e
3143ipfw add 32000 allow ip from any to any
3144.Ed
3145.Pp
3146Along the same lines, doing an
3147.Bd -literal -offset indent
3148ipfw flush
3149.Ed
3150.Pp
3151in similar surroundings is also a bad idea.
3152.It
3153The
3154.Nm
3155filter list may not be modified if the system security level
3156is set to 3 or higher
3157(see
3158.Xr init 8
3159for information on system security levels).
3160.El
3161.Sh PACKET DIVERSION
3162A
3163.Xr divert 4
3164socket bound to the specified port will receive all packets
3165diverted to that port.
3166If no socket is bound to the destination port, or if the divert module is
3167not loaded, or if the kernel was not compiled with divert socket support,
3168the packets are dropped.
3169.Sh NETWORK ADDRESS TRANSLATION (NAT)
3170.Nm
3171support in-kernel NAT using the kernel version of
3172.Xr libalias 3 .
3173The kernel module
3174.Cm ipfw_nat
3175should be loaded or kernel should have
3176.Cm options IPFIREWALL_NAT
3177to be able use NAT.
3178.Pp
3179The nat configuration command is the following:
3180.Bd -ragged -offset indent
3181.Bk -words
3182.Cm nat
3183.Ar nat_number
3184.Cm config
3185.Ar nat-configuration
3186.Ek
3187.Ed
3188.Pp
3189The following parameters can be configured:
3190.Bl -tag -width indent
3191.It Cm ip Ar ip_address
3192Define an ip address to use for aliasing.
3193.It Cm if Ar nic
3194Use ip address of NIC for aliasing, dynamically changing
3195it if NIC's ip address changes.
3196.It Cm log
3197Enable logging on this nat instance.
3198.It Cm deny_in
3199Deny any incoming connection from outside world.
3200.It Cm same_ports
3201Try to leave the alias port numbers unchanged from
3202the actual local port numbers.
3203.It Cm unreg_only
3204Traffic on the local network not originating from an
3205unregistered address spaces will be ignored.
3206.It Cm reset
3207Reset table of the packet aliasing engine on address change.
3208.It Cm reverse
3209Reverse the way libalias handles aliasing.
3210.It Cm proxy_only
3211Obey transparent proxy rules only, packet aliasing is not performed.
3212.It Cm skip_global
3213Skip instance in case of global state lookup (see below).
3214.El
3215.Pp
3216Some specials value can be supplied instead of
3217.Va nat_number:
3218.Bl -tag -width indent
3219.It Cm global
3220Looks up translation state in all configured nat instances.
3221If an entry is found, packet is aliased according to that entry.
3222If no entry was found in any of the instances, packet is passed unchanged,
3223and no new entry will be created.
3224See section
3225.Sx MULTIPLE INSTANCES
3226in
3227.Xr natd 8
3228for more information.
3229.It Cm tablearg
3230Uses argument supplied in lookup table.
3231See
3232.Sx LOOKUP TABLES
3233section below for more information on lookup tables.
3234.El
3235.Pp
3236To let the packet continue after being (de)aliased, set the sysctl variable
3237.Va net.inet.ip.fw.one_pass
3238to 0.
3239For more information about aliasing modes, refer to
3240.Xr libalias 3 .
3241See Section
3242.Sx EXAMPLES
3243for some examples about nat usage.
3244.Ss REDIRECT AND LSNAT SUPPORT IN IPFW
3245Redirect and LSNAT support follow closely the syntax used in
3246.Xr natd 8 .
3247See Section
3248.Sx EXAMPLES
3249for some examples on how to do redirect and lsnat.
3250.Ss SCTP NAT SUPPORT
3251SCTP nat can be configured in a similar manner to TCP through the
3252.Nm
3253command line tool.
3254The main difference is that
3255.Nm sctp nat
3256does not do port translation.
3257Since the local and global side ports will be the same,
3258there is no need to specify both.
3259Ports are redirected as follows:
3260.Bd -ragged -offset indent
3261.Bk -words
3262.Cm nat
3263.Ar nat_number
3264.Cm config if
3265.Ar nic
3266.Cm redirect_port sctp
3267.Ar ip_address [,addr_list] {[port | port-port] [,ports]}
3268.Ek
3269.Ed
3270.Pp
3271Most
3272.Nm sctp nat
3273configuration can be done in real-time through the
3274.Xr sysctl 8
3275interface.
3276All may be changed dynamically, though the hash_table size will only
3277change for new
3278.Nm nat
3279instances.
3280See
3281.Sx SYSCTL VARIABLES
3282for more info.
3283.Sh IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
3284.Nm
3285supports in-kernel IPv6/IPv4 network address and protocol translation.
3286Stateful NAT64 translation allows IPv6-only clients to contact IPv4 servers
3287using unicast TCP, UDP or ICMP protocols.
3288One or more IPv4 addresses assigned to a stateful NAT64 translator are shared
3289among several IPv6-only clients.
3290When stateful NAT64 is used in conjunction with DNS64, no changes are usually
3291required in the IPv6 client or the IPv4 server.
3292The kernel module
3293.Cm ipfw_nat64
3294should be loaded or kernel should have
3295.Cm options IPFIREWALL_NAT64
3296to be able use stateful NAT64 translator.
3297.Pp
3298Stateful NAT64 uses a bunch of memory for several types of objects.
3299When IPv6 client initiates connection, NAT64 translator creates a host entry
3300in the states table.
3301Each host entry has a number of ports group entries allocated on demand.
3302Ports group entries contains connection state entries.
3303There are several options to control limits and lifetime for these objects.
3304.Pp
3305NAT64 translator follows RFC7915 when does ICMPv6/ICMP translation,
3306unsupported message types will be silently dropped.
3307IPv6 needs several ICMPv6 message types to be explicitly allowed for correct
3308operation.
3309Make sure that ND6 neighbor solicitation (ICMPv6 type 135) and neighbor
3310advertisement (ICMPv6 type 136) messages will not be handled by translation
3311rules.
3312.Pp
3313After translation NAT64 translator by default sends packets through
3314corresponding netisr queue.
3315Thus translator host should be configured as IPv4 and IPv6 router.
3316Also this means, that a packet is handled by firewall twice.
3317First time an original packet is handled and consumed by translator,
3318and then it is handled again as translated packet.
3319This behavior can be changed by sysctl variable
3320.Va net.inet.ip.fw.nat64_direct_output .
3321.Pp
3322The stateful NAT64 configuration command is the following:
3323.Bd -ragged -offset indent
3324.Bk -words
3325.Cm nat64lsn
3326.Ar name
3327.Cm create
3328.Ar create-options
3329.Ek
3330.Ed
3331.Pp
3332The following parameters can be configured:
3333.Bl -tag -width indent
3334.It Cm prefix4 Ar ipv4_prefix/plen
3335The IPv4 prefix with mask defines the pool of IPv4 addresses used as
3336source address after translation.
3337Stateful NAT64 module translates IPv6 source address of client to one
3338IPv4 address from this pool.
3339Note that incoming IPv4 packets that don't have corresponding state entry
3340in the states table will be dropped by translator.
3341Make sure that translation rules handle packets, destined to configured prefix.
3342.It Cm prefix6 Ar ipv6_prefix/length
3343The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3344to represent IPv4 addresses. This IPv6 prefix should be configured in DNS64.
3345The translator implementation follows RFC6052, that restricts the length of
3346prefixes to one of following: 32, 40, 48, 56, 64, or 96.
3347The Well-Known IPv6 Prefix 64:ff9b:: must be 96 bits long.
3348.It Cm max_ports Ar number
3349Maximum number of ports reserved for upper level protocols to one IPv6 client.
3350All reserved ports are divided into chunks between supported protocols.
3351The number of connections from one IPv6 client is limited by this option.
3352Note that closed TCP connections still remain in the list of connections until
3353.Cm tcp_close_age
3354interval will not expire.
3355Default value is
3356.Ar 2048 .
3357.It Cm host_del_age Ar seconds
3358The number of seconds until the host entry for a IPv6 client will be deleted
3359and all its resources will be released due to inactivity.
3360Default value is
3361.Ar 3600 .
3362.It Cm pg_del_age Ar seconds
3363The number of seconds until a ports group with unused state entries will
3364be released.
3365Default value is
3366.Ar 900 .
3367.It Cm tcp_syn_age Ar seconds
3368The number of seconds while a state entry for TCP connection with only SYN
3369sent will be kept.
3370If TCP connection establishing will not be finished,
3371state entry will be deleted.
3372Default value is
3373.Ar 10 .
3374.It Cm tcp_est_age Ar seconds
3375The number of seconds while a state entry for established TCP connection
3376will be kept.
3377Default value is
3378.Ar 7200 .
3379.It Cm tcp_close_age Ar seconds
3380The number of seconds while a state entry for closed TCP connection
3381will be kept.
3382Keeping state entries for closed connections is needed, because IPv4 servers
3383typically keep closed connections in a TIME_WAIT state for a several minutes.
3384Since translator's IPv4 addresses are shared among all IPv6 clients,
3385new connections from the same addresses and ports may be rejected by server,
3386because these connections are still in a TIME_WAIT state.
3387Keeping them in translator's state table protects from such rejects.
3388Default value is
3389.Ar 180 .
3390.It Cm udp_age Ar seconds
3391The number of seconds while translator keeps state entry in a waiting for
3392reply to the sent UDP datagram.
3393Default value is
3394.Ar 120 .
3395.It Cm icmp_age Ar seconds
3396The number of seconds while translator keeps state entry in a waiting for
3397reply to the sent ICMP message.
3398Default value is
3399.Ar 60 .
3400.It Cm log
3401Turn on logging of all handled packets via BPF through
3402.Ar ipfwlog0
3403interface.
3404.Ar ipfwlog0
3405is a pseudo interface and can be created after a boot manually with
3406.Cm ifconfig
3407command.
3408Note that it has different purpose than
3409.Ar ipfw0
3410interface.
3411Translators sends to BPF an additional information with each packet.
3412With
3413.Cm tcpdump
3414you are able to see each handled packet before and after translation.
3415.It Cm -log
3416Turn off logging of all handled packets via BPF.
3417.El
3418.Pp
3419To inspect a states table of stateful NAT64 the following command can be used:
3420.Bd -ragged -offset indent
3421.Bk -words
3422.Cm nat64lsn
3423.Ar name
3424.Cm show Cm states
3425.Ek
3426.Ed
3427.Pp
3428.Pp
3429Stateless NAT64 translator doesn't use a states table for translation
3430and converts IPv4 addresses to IPv6 and vice versa solely based on the
3431mappings taken from configured lookup tables.
3432Since a states table doesn't used by stateless translator,
3433it can be configured to pass IPv4 clients to IPv6-only servers.
3434.Pp
3435The stateless NAT64 configuration command is the following:
3436.Bd -ragged -offset indent
3437.Bk -words
3438.Cm nat64stl
3439.Ar name
3440.Cm create
3441.Ar create-options
3442.Ek
3443.Ed
3444.Pp
3445The following parameters can be configured:
3446.Bl -tag -width indent
3447.It Cm prefix6 Ar ipv6_prefix/length
3448The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3449to represent IPv4 addresses. This IPv6 prefix should be configured in DNS64.
3450.It Cm table4 Ar table46
3451The lookup table
3452.Ar table46
3453contains mapping how IPv4 addresses should be translated to IPv6 addresses.
3454.It Cm table6 Ar table64
3455The lookup table
3456.Ar table64
3457contains mapping how IPv6 addresses should be translated to IPv4 addresses.
3458.It Cm log
3459Turn on logging of all handled packets via BPF through
3460.Ar ipfwlog0
3461interface.
3462.It Cm -log
3463Turn off logging of all handled packets via BPF.
3464.El
3465.Pp
3466Note that the behavior of stateless translator with respect to not matched
3467packets differs from stateful translator.
3468If corresponding addresses was not found in the lookup tables, the packet
3469will not be dropped and the search continues.
3470.Sh IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
3471.Nm
3472supports in-kernel IPv6-to-IPv6 network prefix translation as described
3473in RFC6296.
3474The kernel module
3475.Cm ipfw_nptv6
3476should be loaded or kernel should has
3477.Cm options IPFIREWALL_NPTV6
3478to be able use NPTv6 translator.
3479.Pp
3480The NPTv6 configuration command is the following:
3481.Bd -ragged -offset indent
3482.Bk -words
3483.Cm nptv6
3484.Ar name
3485.Cm create
3486.Ar create-options
3487.Ek
3488.Ed
3489.Pp
3490The following parameters can be configured:
3491.Bl -tag -width indent
3492.It Cm int_prefix Ar ipv6_prefix
3493IPv6 prefix used in internal network.
3494NPTv6 module translates source address when it matches this prefix.
3495.It Cm ext_prefix Ar ipv6_prefix
3496IPv6 prefix used in external network.
3497NPTv6 module translates destination address when it matches this prefix.
3498.It Cm prefixlen Ar length
3499The length of specified IPv6 prefixes. It must be in range from 8 to 64.
3500.El
3501.Pp
3502Note that the prefix translation rules are silently ignored when IPv6 packet
3503forwarding is disabled.
3504To enable the packet forwarding, set the sysctl variable
3505.Va net.inet6.ip6.forwarding
3506to 1.
3507.Pp
3508To let the packet continue after being translated, set the sysctl variable
3509.Va net.inet.ip.fw.one_pass
3510to 0.
3511.Sh LOADER TUNABLES
3512Tunables can be set in
3513.Xr loader 8
3514prompt,
3515.Xr loader.conf 5
3516or
3517.Xr kenv 1
3518before ipfw module gets loaded.
3519.Bl -tag -width indent
3520.It Va net.inet.ip.fw.default_to_accept: No 0
3521Defines ipfw last rule behavior.
3522This value overrides
3523.Cd "options IPFW_DEFAULT_TO_(ACCEPT|DENY)"
3524from kernel configuration file.
3525.It Va net.inet.ip.fw.tables_max: No 128
3526Defines number of tables available in ipfw.
3527Number cannot exceed 65534.
3528.El
3529.Sh SYSCTL VARIABLES
3530A set of
3531.Xr sysctl 8
3532variables controls the behaviour of the firewall and
3533associated modules
3534.Pq Nm dummynet , bridge , sctp nat .
3535These are shown below together with their default value
3536(but always check with the
3537.Xr sysctl 8
3538command what value is actually in use) and meaning:
3539.Bl -tag -width indent
3540.It Va net.inet.ip.alias.sctp.accept_global_ootb_addip: No 0
3541Defines how the
3542.Nm nat
3543responds to receipt of global OOTB ASCONF-AddIP:
3544.Bl -tag -width indent
3545.It Cm 0
3546No response (unless a partially matching association exists -
3547ports and vtags match but global address does not)
3548.It Cm 1
3549.Nm nat
3550will accept and process all OOTB global AddIP messages.
3551.El
3552.Pp
3553Option 1 should never be selected as this forms a security risk.
3554An attacker can
3555establish multiple fake associations by sending AddIP messages.
3556.It Va net.inet.ip.alias.sctp.chunk_proc_limit: No 5
3557Defines the maximum number of chunks in an SCTP packet that will be
3558parsed for a
3559packet that matches an existing association.
3560This value is enforced to be greater or equal than
3561.Cm net.inet.ip.alias.sctp.initialising_chunk_proc_limit .
3562A high value is
3563a DoS risk yet setting too low a value may result in
3564important control chunks in
3565the packet not being located and parsed.
3566.It Va net.inet.ip.alias.sctp.error_on_ootb: No 1
3567Defines when the
3568.Nm nat
3569responds to any Out-of-the-Blue (OOTB) packets with ErrorM packets.
3570An OOTB packet is a packet that arrives with no existing association
3571registered in the
3572.Nm nat
3573and is not an INIT or ASCONF-AddIP packet:
3574.Bl -tag -width indent
3575.It Cm 0
3576ErrorM is never sent in response to OOTB packets.
3577.It Cm 1
3578ErrorM is only sent to OOTB packets received on the local side.
3579.It Cm 2
3580ErrorM is sent to the local side and on the global side ONLY if there is a
3581partial match (ports and vtags match but the source global IP does not).
3582This value is only useful if the
3583.Nm nat
3584is tracking global IP addresses.
3585.It Cm 3
3586ErrorM is sent in response to all OOTB packets on both
3587the local and global side
3588(DoS risk).
3589.El
3590.Pp
3591At the moment the default is 0, since the ErrorM packet is not yet
3592supported by most SCTP stacks.
3593When it is supported, and if not tracking
3594global addresses, we recommend setting this value to 1 to allow
3595multi-homed local hosts to function with the
3596.Nm nat .
3597To track global addresses, we recommend setting this value to 2 to
3598allow global hosts to be informed when they need to (re)send an
3599ASCONF-AddIP.
3600Value 3 should never be chosen (except for debugging) as the
3601.Nm nat
3602will respond to all OOTB global packets (a DoS risk).
3603.It Va net.inet.ip.alias.sctp.hashtable_size: No 2003
3604Size of hash tables used for
3605.Nm nat
3606lookups (100 < prime_number > 1000001).
3607This value sets the
3608.Nm hash table
3609size for any future created
3610.Nm nat
3611instance and therefore must be set prior to creating a
3612.Nm nat
3613instance.
3614The table sizes may be changed to suit specific needs.
3615If there will be few
3616concurrent associations, and memory is scarce, you may make these smaller.
3617If there will be many thousands (or millions) of concurrent associations, you
3618should make these larger.
3619A prime number is best for the table size.
3620The sysctl
3621update function will adjust your input value to the next highest prime number.
3622.It Va net.inet.ip.alias.sctp.holddown_time:  No 0
3623Hold association in table for this many seconds after receiving a
3624SHUTDOWN-COMPLETE.
3625This allows endpoints to correct shutdown gracefully if a
3626shutdown_complete is lost and retransmissions are required.
3627.It Va net.inet.ip.alias.sctp.init_timer: No 15
3628Timeout value while waiting for (INIT-ACK|AddIP-ACK).
3629This value cannot be 0.
3630.It Va net.inet.ip.alias.sctp.initialising_chunk_proc_limit: No 2
3631Defines the maximum number of chunks in an SCTP packet that will be parsed when
3632no existing association exists that matches that packet.
3633Ideally this packet
3634will only be an INIT or ASCONF-AddIP packet.
3635A higher value may become a DoS
3636risk as malformed packets can consume processing resources.
3637.It Va net.inet.ip.alias.sctp.param_proc_limit: No 25
3638Defines the maximum number of parameters within a chunk that will be
3639parsed in a
3640packet.
3641As for other similar sysctl variables, larger values pose a DoS risk.
3642.It Va net.inet.ip.alias.sctp.log_level: No 0
3643Level of detail in the system log messages (0 \- minimal, 1 \- event,
36442 \- info, 3 \- detail, 4 \- debug, 5 \- max debug).
3645May be a good
3646option in high loss environments.
3647.It Va net.inet.ip.alias.sctp.shutdown_time: No 15
3648Timeout value while waiting for SHUTDOWN-COMPLETE.
3649This value cannot be 0.
3650.It Va net.inet.ip.alias.sctp.track_global_addresses: No 0
3651Enables/disables global IP address tracking within the
3652.Nm nat
3653and places an
3654upper limit on the number of addresses tracked for each association:
3655.Bl -tag -width indent
3656.It Cm 0
3657Global tracking is disabled
3658.It Cm >1
3659Enables tracking, the maximum number of addresses tracked for each
3660association is limited to this value
3661.El
3662.Pp
3663This variable is fully dynamic, the new value will be adopted for all newly
3664arriving associations, existing associations are treated
3665as they were previously.
3666Global tracking will decrease the number of collisions within the
3667.Nm nat
3668at a cost
3669of increased processing load, memory usage, complexity, and possible
3670.Nm nat
3671state
3672problems in complex networks with multiple
3673.Nm nats .
3674We recommend not tracking
3675global IP addresses, this will still result in a fully functional
3676.Nm nat .
3677.It Va net.inet.ip.alias.sctp.up_timer: No 300
3678Timeout value to keep an association up with no traffic.
3679This value cannot be 0.
3680.It Va net.inet.ip.dummynet.codel.interval : No 100000
3681Default
3682.Cm codel
3683AQM interval in microseconds.
3684The value must be in the range 1..5000000.
3685.It Va net.inet.ip.dummynet.codel.target : No 5000
3686Default
3687.Cm codel
3688AQM target delay time in microseconds (the minimum acceptable persistent queue
3689delay).
3690The value must be in the range 1..5000000.
3691.It Va net.inet.ip.dummynet.expire : No 1
3692Lazily delete dynamic pipes/queue once they have no pending traffic.
3693You can disable this by setting the variable to 0, in which case
3694the pipes/queues will only be deleted when the threshold is reached.
3695.It Va net.inet.ip.dummynet.fqcodel.flows : No 1024
3696Defines the default total number of flow queues (sub-queues) that
3697.Cm fq_codel
3698creates and manages.
3699The value must be in the range 1..65536.
3700.It Va net.inet.ip.dummynet.fqcodel.interval : No 100000
3701Default
3702.Cm fq_codel
3703scheduler/AQM interval in microseconds.
3704The value must be in the range 1..5000000.
3705.It Va net.inet.ip.dummynet.fqcodel.limit : No 10240
3706The default hard size limit (in unit of packet) of all queues managed by an
3707instance of the
3708.Cm fq_codel
3709scheduler.
3710The value must be in the range 1..20480.
3711.It Va net.inet.ip.dummynet.fqcodel.quantum : No 1514
3712The default quantum (credit) of the
3713.Cm fq_codel
3714in unit of byte.
3715The value must be in the range 1..9000.
3716.It Va net.inet.ip.dummynet.fqcodel.target : No 5000
3717Default
3718.Cm fq_codel
3719scheduler/AQM target delay time in microseconds (the minimum acceptable
3720persistent queue delay).
3721The value must be in the range 1..5000000.
3722.It Va net.inet.ip.dummynet.fqpie.alpha : No 125
3723The default
3724.Ar alpha
3725parameter (scaled by 1000) for
3726.Cm fq_pie
3727scheduler/AQM.
3728The value must be in the range 1..7000.
3729.It Va net.inet.ip.dummynet.fqpie.beta : No 1250
3730The default
3731.Ar beta
3732parameter (scaled by 1000) for
3733.Cm fq_pie
3734scheduler/AQM.
3735The value must be in the range 1..7000.
3736.It Va net.inet.ip.dummynet.fqpie.flows : No 1024
3737Defines the default total number of flow queues (sub-queues) that
3738.Cm fq_pie
3739creates and manages.
3740The value must be in the range 1..65536.
3741.It Va net.inet.ip.dummynet.fqpie.limit : No 10240
3742The default hard size limit (in unit of packet) of all queues managed by an
3743instance of the
3744.Cm fq_pie
3745scheduler.
3746The value must be in the range 1..20480.
3747.It Va net.inet.ip.dummynet.fqpie.max_burst : No 150000
3748The default maximum period of microseconds that
3749.Cm fq_pie
3750scheduler/AQM does not drop/mark packets.
3751The value must be in the range 1..10000000.
3752.It Va net.inet.ip.dummynet.fqpie.max_ecnth : No 99
3753The default maximum ECN probability threshold (scaled by 1000) for
3754.Cm fq_pie
3755scheduler/AQM.
3756The value must be in the range 1..7000.
3757.It Va net.inet.ip.dummynet.fqpie.quantum : No 1514
3758The default quantum (credit) of the
3759.Cm fq_pie
3760in unit of byte.
3761The value must be in the range 1..9000.
3762.It Va net.inet.ip.dummynet.fqpie.target : No 15000
3763The default
3764.Cm target
3765delay of the
3766.Cm fq_pie
3767in unit of microsecond.
3768The value must be in the range 1..5000000.
3769.It Va net.inet.ip.dummynet.fqpie.tupdate : No 15000
3770The default
3771.Cm tupdate
3772of the
3773.Cm fq_pie
3774in unit of microsecond.
3775The value must be in the range 1..5000000.
3776.It Va net.inet.ip.dummynet.hash_size : No 64
3777Default size of the hash table used for dynamic pipes/queues.
3778This value is used when no
3779.Cm buckets
3780option is specified when configuring a pipe/queue.
3781.It Va net.inet.ip.dummynet.io_fast : No 0
3782If set to a non-zero value,
3783the
3784.Dq fast
3785mode of
3786.Nm dummynet
3787operation (see above) is enabled.
3788.It Va net.inet.ip.dummynet.io_pkt
3789Number of packets passed to
3790.Nm dummynet .
3791.It Va net.inet.ip.dummynet.io_pkt_drop
3792Number of packets dropped by
3793.Nm dummynet .
3794.It Va net.inet.ip.dummynet.io_pkt_fast
3795Number of packets bypassed by the
3796.Nm dummynet
3797scheduler.
3798.It Va net.inet.ip.dummynet.max_chain_len : No 16
3799Target value for the maximum number of pipes/queues in a hash bucket.
3800The product
3801.Cm max_chain_len*hash_size
3802is used to determine the threshold over which empty pipes/queues
3803will be expired even when
3804.Cm net.inet.ip.dummynet.expire=0 .
3805.It Va net.inet.ip.dummynet.red_lookup_depth : No 256
3806.It Va net.inet.ip.dummynet.red_avg_pkt_size : No 512
3807.It Va net.inet.ip.dummynet.red_max_pkt_size : No 1500
3808Parameters used in the computations of the drop probability
3809for the RED algorithm.
3810.It Va net.inet.ip.dummynet.pie.alpha : No 125
3811The default
3812.Ar alpha
3813parameter (scaled by 1000) for
3814.Cm pie
3815AQM.
3816The value must be in the range 1..7000.
3817.It Va net.inet.ip.dummynet.pie.beta : No 1250
3818The default
3819.Ar beta
3820parameter (scaled by 1000) for
3821.Cm pie
3822AQM.
3823The value must be in the range 1..7000.
3824.It Va net.inet.ip.dummynet.pie.max_burst : No 150000
3825The default maximum period of microseconds that
3826.Cm pie
3827AQM does not drop/mark packets.
3828The value must be in the range 1..10000000.
3829.It Va net.inet.ip.dummynet.pie.max_ecnth : No 99
3830The default maximum ECN probability threshold (scaled by 1000) for
3831.Cm pie
3832AQM.
3833The value must be in the range 1..7000.
3834.It Va net.inet.ip.dummynet.pie.target : No 15000
3835The default
3836.Cm target
3837delay of
3838.Cm pie
3839AQM in unit of microsecond.
3840The value must be in the range 1..5000000.
3841.It Va net.inet.ip.dummynet.pie.tupdate : No 15000
3842The default
3843.Cm tupdate
3844of
3845.Cm pie
3846AQM in unit of microsecond.
3847The value must be in the range 1..5000000.
3848.It Va net.inet.ip.dummynet.pipe_byte_limit : No 1048576
3849.It Va net.inet.ip.dummynet.pipe_slot_limit : No 100
3850The maximum queue size that can be specified in bytes or packets.
3851These limits prevent accidental exhaustion of resources such as mbufs.
3852If you raise these limits,
3853you should make sure the system is configured so that sufficient resources
3854are available.
3855.It Va net.inet.ip.fw.autoinc_step : No 100
3856Delta between rule numbers when auto-generating them.
3857The value must be in the range 1..1000.
3858.It Va net.inet.ip.fw.curr_dyn_buckets : Va net.inet.ip.fw.dyn_buckets
3859The current number of buckets in the hash table for dynamic rules
3860(readonly).
3861.It Va net.inet.ip.fw.debug : No 1
3862Controls debugging messages produced by
3863.Nm .
3864.It Va net.inet.ip.fw.default_rule : No 65535
3865The default rule number (read-only).
3866By the design of
3867.Nm , the default rule is the last one, so its number
3868can also serve as the highest number allowed for a rule.
3869.It Va net.inet.ip.fw.dyn_buckets : No 256
3870The number of buckets in the hash table for dynamic rules.
3871Must be a power of 2, up to 65536.
3872It only takes effect when all dynamic rules have expired, so you
3873are advised to use a
3874.Cm flush
3875command to make sure that the hash table is resized.
3876.It Va net.inet.ip.fw.dyn_count : No 3
3877Current number of dynamic rules
3878(read-only).
3879.It Va net.inet.ip.fw.dyn_keepalive : No 1
3880Enables generation of keepalive packets for
3881.Cm keep-state
3882rules on TCP sessions.
3883A keepalive is generated to both
3884sides of the connection every 5 seconds for the last 20
3885seconds of the lifetime of the rule.
3886.It Va net.inet.ip.fw.dyn_max : No 8192
3887Maximum number of dynamic rules.
3888When you hit this limit, no more dynamic rules can be
3889installed until old ones expire.
3890.It Va net.inet.ip.fw.dyn_ack_lifetime : No 300
3891.It Va net.inet.ip.fw.dyn_syn_lifetime : No 20
3892.It Va net.inet.ip.fw.dyn_fin_lifetime : No 1
3893.It Va net.inet.ip.fw.dyn_rst_lifetime : No 1
3894.It Va net.inet.ip.fw.dyn_udp_lifetime : No 5
3895.It Va net.inet.ip.fw.dyn_short_lifetime : No 30
3896These variables control the lifetime, in seconds, of dynamic
3897rules.
3898Upon the initial SYN exchange the lifetime is kept short,
3899then increased after both SYN have been seen, then decreased
3900again during the final FIN exchange or when a RST is received.
3901Both
3902.Em dyn_fin_lifetime
3903and
3904.Em dyn_rst_lifetime
3905must be strictly lower than 5 seconds, the period of
3906repetition of keepalives.
3907The firewall enforces that.
3908.It Va net.inet.ip.fw.dyn_keep_states: No 0
3909Keep dynamic states on rule/set deletion.
3910States are relinked to default rule (65535).
3911This can be handly for ruleset reload.
3912Turned off by default.
3913.It Va net.inet.ip.fw.enable : No 1
3914Enables the firewall.
3915Setting this variable to 0 lets you run your machine without
3916firewall even if compiled in.
3917.It Va net.inet6.ip6.fw.enable : No 1
3918provides the same functionality as above for the IPv6 case.
3919.It Va net.inet.ip.fw.one_pass : No 1
3920When set, the packet exiting from the
3921.Nm dummynet
3922pipe or from
3923.Xr ng_ipfw 4
3924node is not passed though the firewall again.
3925Otherwise, after an action, the packet is
3926reinjected into the firewall at the next rule.
3927.It Va net.inet.ip.fw.tables_max : No 128
3928Maximum number of tables.
3929.It Va net.inet.ip.fw.verbose : No 1
3930Enables verbose messages.
3931.It Va net.inet.ip.fw.verbose_limit : No 0
3932Limits the number of messages produced by a verbose firewall.
3933.It Va net.inet6.ip6.fw.deny_unknown_exthdrs : No 1
3934If enabled packets with unknown IPv6 Extension Headers will be denied.
3935.It Va net.link.ether.ipfw : No 0
3936Controls whether layer-2 packets are passed to
3937.Nm .
3938Default is no.
3939.It Va net.link.bridge.ipfw : No 0
3940Controls whether bridged packets are passed to
3941.Nm .
3942Default is no.
3943.It Va net.inet.ip.fw.nat64_allow_private : No 0
3944Defines how
3945.Nm nat64
3946handles private IPv4 addresses:
3947.Bl -tag -width indent
3948.It Cm 0
3949Packets with private IPv4 will not be handled by translator
3950.It Cm 1
3951Translator will accept and process packets with private IPv4 addresses.
3952.El
3953.It Va net.inet.ip.fw.nat64_debug : No 0
3954Controls debugging messages produced by
3955.Nm ipfw_nat64
3956module.
3957.It Va net.inet.ip.fw.nat64_direct_output : No 0
3958Controls the output method used by
3959.Nm ipfw_nat64
3960module:
3961.Bl -tag -width indent
3962.It Cm 0
3963A packet is handled by
3964.Nm ipfw
3965twice.
3966First time an original packet is handled by
3967.Nm ipfw
3968and consumed by
3969.Nm ipfw_nat64
3970translator.
3971Then translated packet is queued via netisr to input processing again.
3972.It Cm 1
3973A packet is handled by
3974.Nm ipfw
3975only once, and after translation it will be pushed directly to outgoing
3976interface.
3977.El
3978.El
3979.Sh INTERNAL DIAGNOSTICS
3980There are some commands that may be useful to understand current state
3981of certain subsystems inside kernel module.
3982These commands provide debugging output which may change without notice.
3983.Pp
3984Currently the following commands are available as
3985.Cm internal
3986sub-options:
3987.Bl -tag -width indent
3988.It Cm iflist
3989Lists all interface which are currently tracked by
3990.Nm
3991with their in-kernel status.
3992.It Cm talist
3993List all table lookup algorithms currently available.
3994.El
3995.Sh EXAMPLES
3996There are far too many possible uses of
3997.Nm
3998so this Section will only give a small set of examples.
3999.Pp
4000.Ss BASIC PACKET FILTERING
4001This command adds an entry which denies all tcp packets from
4002.Em cracker.evil.org
4003to the telnet port of
4004.Em wolf.tambov.su
4005from being forwarded by the host:
4006.Pp
4007.Dl "ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet"
4008.Pp
4009This one disallows any connection from the entire cracker's
4010network to my host:
4011.Pp
4012.Dl "ipfw add deny ip from 123.45.67.0/24 to my.host.org"
4013.Pp
4014A first and efficient way to limit access (not using dynamic rules)
4015is the use of the following rules:
4016.Pp
4017.Dl "ipfw add allow tcp from any to any established"
4018.Dl "ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup"
4019.Dl "ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup"
4020.Dl "..."
4021.Dl "ipfw add deny tcp from any to any"
4022.Pp
4023The first rule will be a quick match for normal TCP packets,
4024but it will not match the initial SYN packet, which will be
4025matched by the
4026.Cm setup
4027rules only for selected source/destination pairs.
4028All other SYN packets will be rejected by the final
4029.Cm deny
4030rule.
4031.Pp
4032If you administer one or more subnets, you can take advantage
4033of the address sets and or-blocks and write extremely
4034compact rulesets which selectively enable services to blocks
4035of clients, as below:
4036.Pp
4037.Dl "goodguys=\*q{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }\*q"
4038.Dl "badguys=\*q10.1.2.0/24{8,38,60}\*q"
4039.Dl ""
4040.Dl "ipfw add allow ip from ${goodguys} to any"
4041.Dl "ipfw add deny ip from ${badguys} to any"
4042.Dl "... normal policies ..."
4043.Pp
4044The
4045.Cm verrevpath
4046option could be used to do automated anti-spoofing by adding the
4047following to the top of a ruleset:
4048.Pp
4049.Dl "ipfw add deny ip from any to any not verrevpath in"
4050.Pp
4051This rule drops all incoming packets that appear to be coming to the
4052system on the wrong interface.
4053For example, a packet with a source
4054address belonging to a host on a protected internal network would be
4055dropped if it tried to enter the system from an external interface.
4056.Pp
4057The
4058.Cm antispoof
4059option could be used to do similar but more restricted anti-spoofing
4060by adding the following to the top of a ruleset:
4061.Pp
4062.Dl "ipfw add deny ip from any to any not antispoof in"
4063.Pp
4064This rule drops all incoming packets that appear to be coming from another
4065directly connected system but on the wrong interface.
4066For example, a packet with a source address of
4067.Li 192.168.0.0/24 ,
4068configured on
4069.Li fxp0 ,
4070but coming in on
4071.Li fxp1
4072would be dropped.
4073.Pp
4074The
4075.Cm setdscp
4076option could be used to (re)mark user traffic,
4077by adding the following to the appropriate place in ruleset:
4078.Pp
4079.Dl "ipfw add setdscp be ip from any to any dscp af11,af21"
4080.Ss DYNAMIC RULES
4081In order to protect a site from flood attacks involving fake
4082TCP packets, it is safer to use dynamic rules:
4083.Pp
4084.Dl "ipfw add check-state"
4085.Dl "ipfw add deny tcp from any to any established"
4086.Dl "ipfw add allow tcp from my-net to any setup keep-state"
4087.Pp
4088This will let the firewall install dynamic rules only for
4089those connection which start with a regular SYN packet coming
4090from the inside of our network.
4091Dynamic rules are checked when encountering the first
4092occurrence of a
4093.Cm check-state ,
4094.Cm keep-state
4095or
4096.Cm limit
4097rule.
4098A
4099.Cm check-state
4100rule should usually be placed near the beginning of the
4101ruleset to minimize the amount of work scanning the ruleset.
4102Your mileage may vary.
4103.Pp
4104For more complex scenarios with dynamic rules
4105.Cm record-state
4106and
4107.Cm defer-action
4108can be used to precisely control creation and checking of dynamic rules.
4109Example of usage of these options are provided in
4110.Sx NETWORK ADDRESS TRANSLATION (NAT)
4111Section.
4112.Pp
4113To limit the number of connections a user can open
4114you can use the following type of rules:
4115.Pp
4116.Dl "ipfw add allow tcp from my-net/24 to any setup limit src-addr 10"
4117.Dl "ipfw add allow tcp from any to me setup limit src-addr 4"
4118.Pp
4119The former (assuming it runs on a gateway) will allow each host
4120on a /24 network to open at most 10 TCP connections.
4121The latter can be placed on a server to make sure that a single
4122client does not use more than 4 simultaneous connections.
4123.Pp
4124.Em BEWARE :
4125stateful rules can be subject to denial-of-service attacks
4126by a SYN-flood which opens a huge number of dynamic rules.
4127The effects of such attacks can be partially limited by
4128acting on a set of
4129.Xr sysctl 8
4130variables which control the operation of the firewall.
4131.Pp
4132Here is a good usage of the
4133.Cm list
4134command to see accounting records and timestamp information:
4135.Pp
4136.Dl ipfw -at list
4137.Pp
4138or in short form without timestamps:
4139.Pp
4140.Dl ipfw -a list
4141.Pp
4142which is equivalent to:
4143.Pp
4144.Dl ipfw show
4145.Pp
4146Next rule diverts all incoming packets from 192.168.2.0/24
4147to divert port 5000:
4148.Pp
4149.Dl ipfw divert 5000 ip from 192.168.2.0/24 to any in
4150.Ss TRAFFIC SHAPING
4151The following rules show some of the applications of
4152.Nm
4153and
4154.Nm dummynet
4155for simulations and the like.
4156.Pp
4157This rule drops random incoming packets with a probability
4158of 5%:
4159.Pp
4160.Dl "ipfw add prob 0.05 deny ip from any to any in"
4161.Pp
4162A similar effect can be achieved making use of
4163.Nm dummynet
4164pipes:
4165.Pp
4166.Dl "ipfw add pipe 10 ip from any to any"
4167.Dl "ipfw pipe 10 config plr 0.05"
4168.Pp
4169We can use pipes to artificially limit bandwidth, e.g.\& on a
4170machine acting as a router, if we want to limit traffic from
4171local clients on 192.168.2.0/24 we do:
4172.Pp
4173.Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
4174.Dl "ipfw pipe 1 config bw 300Kbit/s queue 50KBytes"
4175.Pp
4176note that we use the
4177.Cm out
4178modifier so that the rule is not used twice.
4179Remember in fact that
4180.Nm
4181rules are checked both on incoming and outgoing packets.
4182.Pp
4183Should we want to simulate a bidirectional link with bandwidth
4184limitations, the correct way is the following:
4185.Pp
4186.Dl "ipfw add pipe 1 ip from any to any out"
4187.Dl "ipfw add pipe 2 ip from any to any in"
4188.Dl "ipfw pipe 1 config bw 64Kbit/s queue 10Kbytes"
4189.Dl "ipfw pipe 2 config bw 64Kbit/s queue 10Kbytes"
4190.Pp
4191The above can be very useful, e.g.\& if you want to see how
4192your fancy Web page will look for a residential user who
4193is connected only through a slow link.
4194You should not use only one pipe for both directions, unless
4195you want to simulate a half-duplex medium (e.g.\& AppleTalk,
4196Ethernet, IRDA).
4197It is not necessary that both pipes have the same configuration,
4198so we can also simulate asymmetric links.
4199.Pp
4200Should we want to verify network performance with the RED queue
4201management algorithm:
4202.Pp
4203.Dl "ipfw add pipe 1 ip from any to any"
4204.Dl "ipfw pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1"
4205.Pp
4206Another typical application of the traffic shaper is to
4207introduce some delay in the communication.
4208This can significantly affect applications which do a lot of Remote
4209Procedure Calls, and where the round-trip-time of the
4210connection often becomes a limiting factor much more than
4211bandwidth:
4212.Pp
4213.Dl "ipfw add pipe 1 ip from any to any out"
4214.Dl "ipfw add pipe 2 ip from any to any in"
4215.Dl "ipfw pipe 1 config delay 250ms bw 1Mbit/s"
4216.Dl "ipfw pipe 2 config delay 250ms bw 1Mbit/s"
4217.Pp
4218Per-flow queueing can be useful for a variety of purposes.
4219A very simple one is counting traffic:
4220.Pp
4221.Dl "ipfw add pipe 1 tcp from any to any"
4222.Dl "ipfw add pipe 1 udp from any to any"
4223.Dl "ipfw add pipe 1 ip from any to any"
4224.Dl "ipfw pipe 1 config mask all"
4225.Pp
4226The above set of rules will create queues (and collect
4227statistics) for all traffic.
4228Because the pipes have no limitations, the only effect is
4229collecting statistics.
4230Note that we need 3 rules, not just the last one, because
4231when
4232.Nm
4233tries to match IP packets it will not consider ports, so we
4234would not see connections on separate ports as different
4235ones.
4236.Pp
4237A more sophisticated example is limiting the outbound traffic
4238on a net with per-host limits, rather than per-network limits:
4239.Pp
4240.Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
4241.Dl "ipfw add pipe 2 ip from any to 192.168.2.0/24 in"
4242.Dl "ipfw pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
4243.Dl "ipfw pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
4244.Ss LOOKUP TABLES
4245In the following example, we need to create several traffic bandwidth
4246classes and we need different hosts/networks to fall into different classes.
4247We create one pipe for each class and configure them accordingly.
4248Then we create a single table and fill it with IP subnets and addresses.
4249For each subnet/host we set the argument equal to the number of the pipe
4250that it should use.
4251Then we classify traffic using a single rule:
4252.Pp
4253.Dl "ipfw pipe 1 config bw 1000Kbyte/s"
4254.Dl "ipfw pipe 4 config bw 4000Kbyte/s"
4255.Dl "..."
4256.Dl "ipfw table T1 create type addr"
4257.Dl "ipfw table T1 add 192.168.2.0/24 1"
4258.Dl "ipfw table T1 add 192.168.0.0/27 4"
4259.Dl "ipfw table T1 add 192.168.0.2 1"
4260.Dl "..."
4261.Dl "ipfw add pipe tablearg ip from 'table(T1)' to any"
4262.Pp
4263Using the
4264.Cm fwd
4265action, the table entries may include hostnames and IP addresses.
4266.Pp
4267.Dl "ipfw table T2 create type addr ftype ip"
4268.Dl "ipfw table T2 add 192.168.2.0/24 10.23.2.1"
4269.Dl "ipfw table T21 add 192.168.0.0/27 router1.dmz"
4270.Dl "..."
4271.Dl "ipfw add 100 fwd tablearg ip from any to table(1)"
4272.Pp
4273In the following example per-interface firewall is created:
4274.Pp
4275.Dl "ipfw table IN create type iface valtype skipto,fib"
4276.Dl "ipfw table IN add vlan20 12000,12"
4277.Dl "ipfw table IN add vlan30 13000,13"
4278.Dl "ipfw table OUT create type iface valtype skipto"
4279.Dl "ipfw table OUT add vlan20 22000"
4280.Dl "ipfw table OUT add vlan30 23000"
4281.Dl ".."
4282.Dl "ipfw add 100 setfib tablearg ip from any to any recv 'table(IN)' in"
4283.Dl "ipfw add 200 skipto tablearg ip from any to any recv 'table(IN)' in"
4284.Dl "ipfw add 300 skipto tablearg ip from any to any xmit 'table(OUT)' out"
4285.Pp
4286The following example illustrate usage of flow tables:
4287.Pp
4288.Dl "ipfw table fl create type flow:flow:src-ip,proto,dst-ip,dst-port"
4289.Dl "ipfw table fl add 2a02:6b8:77::88,tcp,2a02:6b8:77::99,80 11"
4290.Dl "ipfw table fl add 10.0.0.1,udp,10.0.0.2,53 12"
4291.Dl ".."
4292.Dl "ipfw add 100 allow ip from any to any flow 'table(fl,11)' recv ix0"
4293.Ss SETS OF RULES
4294To add a set of rules atomically, e.g.\& set 18:
4295.Pp
4296.Dl "ipfw set disable 18"
4297.Dl "ipfw add NN set 18 ...         # repeat as needed"
4298.Dl "ipfw set enable 18"
4299.Pp
4300To delete a set of rules atomically the command is simply:
4301.Pp
4302.Dl "ipfw delete set 18"
4303.Pp
4304To test a ruleset and disable it and regain control if something goes wrong:
4305.Pp
4306.Dl "ipfw set disable 18"
4307.Dl "ipfw add NN set 18 ...         # repeat as needed"
4308.Dl "ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18"
4309.Pp
4310Here if everything goes well, you press control-C before the "sleep"
4311terminates, and your ruleset will be left active.
4312Otherwise, e.g.\& if
4313you cannot access your box, the ruleset will be disabled after
4314the sleep terminates thus restoring the previous situation.
4315.Pp
4316To show rules of the specific set:
4317.Pp
4318.Dl "ipfw set 18 show"
4319.Pp
4320To show rules of the disabled set:
4321.Pp
4322.Dl "ipfw -S set 18 show"
4323.Pp
4324To clear a specific rule counters of the specific set:
4325.Pp
4326.Dl "ipfw set 18 zero NN"
4327.Pp
4328To delete a specific rule of the specific set:
4329.Pp
4330.Dl "ipfw set 18 delete NN"
4331.Ss NAT, REDIRECT AND LSNAT
4332First redirect all the traffic to nat instance 123:
4333.Pp
4334.Dl "ipfw add nat 123 all from any to any"
4335.Pp
4336Then to configure nat instance 123 to alias all the outgoing traffic with ip
4337192.168.0.123, blocking all incoming connections, trying to keep
4338same ports on both sides, clearing aliasing table on address change
4339and keeping a log of traffic/link statistics:
4340.Pp
4341.Dl "ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports"
4342.Pp
4343Or to change address of instance 123, aliasing table will be cleared (see
4344reset option):
4345.Pp
4346.Dl "ipfw nat 123 config ip 10.0.0.1"
4347.Pp
4348To see configuration of nat instance 123:
4349.Pp
4350.Dl "ipfw nat 123 show config"
4351.Pp
4352To show logs of all the instances in range 111-999:
4353.Pp
4354.Dl "ipfw nat 111-999 show"
4355.Pp
4356To see configurations of all instances:
4357.Pp
4358.Dl "ipfw nat show config"
4359.Pp
4360Or a redirect rule with mixed modes could looks like:
4361.Pp
4362.Dl "ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66"
4363.Dl "			 redirect_port tcp 192.168.0.1:80 500"
4364.Dl "			 redirect_proto udp 192.168.1.43 192.168.1.1"
4365.Dl "			 redirect_addr 192.168.0.10,192.168.0.11"
4366.Dl "			 	    10.0.0.100	# LSNAT"
4367.Dl "			 redirect_port tcp 192.168.0.1:80,192.168.0.10:22"
4368.Dl "			 	    500		# LSNAT"
4369.Pp
4370or it could be split in:
4371.Pp
4372.Dl "ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66"
4373.Dl "ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500"
4374.Dl "ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1"
4375.Dl "ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12"
4376.Dl "				         10.0.0.100"
4377.Dl "ipfw nat 5 config redirect_port tcp"
4378.Dl "			192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500"
4379.Pp
4380Sometimes you may want to mix NAT and dynamic rules. It could be achived with
4381.Cm record-state
4382and
4383.Cm defer-action
4384options. Problem is, you need to create dynamic rule before NAT and check it
4385after NAT actions (or vice versa) to have consistent addresses and ports.
4386Rule with
4387.Cm keep-state
4388option will trigger activation of existing dynamic state, and action of such
4389rule will be performed as soon as rule is matched. In case of NAT and
4390.Cm allow
4391rule packet need to be passed to NAT, not allowed as soon is possible.
4392.Pp
4393There is example of set of rules to achive this. Bear in mind that this
4394is exmaple only and it is not very usefult by itself.
4395.Pp
4396On way out, after all checks place this rules:
4397.Pp
4398.Dl "ipfw add allow record-state skip-action"
4399.Dl "ipfw add nat 1"
4400.Pp
4401And on way in there should be something like this:
4402.Pp
4403.Dl "ipfw add nat 1"
4404.Dl "ipfw add check-state"
4405.Pp
4406Please note, that first rule on way out doesn't allow packet and doesn't
4407execute existing dynamic rules. All it does, create new dynamic rule with
4408.Cm allow
4409action, if it is not created yet. Later, this dynamic rule is used on way
4410in by
4411.Cm check-state
4412rule.
4413.Ss CONFIGURING CODEL, PIE, FQ-CODEL and FQ-PIE AQM
4414.Cm codel
4415and
4416.Cm pie
4417AQM can be configured for
4418.Nm dummynet
4419.Cm pipe
4420or
4421.Cm queue.
4422.Pp
4423To configure a
4424.Cm pipe
4425with
4426.Cm codel
4427AQM using default configuration for traffic from 192.168.0.0/24 and 1Mbits/s
4428rate limit, we do:
4429.Pp
4430.Dl "ipfw pipe 1 config bw 1mbits/s codel"
4431.Dl "ipfw add 100 pipe 1 ip from 192.168.0.0/24 to any"
4432.Pp
4433To configure a
4434.Cm queue
4435with
4436.Cm codel
4437AQM using different configurations parameters for traffic from
4438192.168.0.0/24 and 1Mbits/s rate limit, we do:
4439.Pp
4440.Dl "ipfw pipe 1 config bw 1mbits/s"
4441.Dl "ipfw queue 1 config pipe 1 codel target 8ms interval 160ms ecn"
4442.Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4443.Pp
4444To configure a
4445.Cm pipe
4446with
4447.Cm pie
4448AQM using default configuration for traffic from 192.168.0.0/24 and 1Mbits/s
4449rate limit, we do:
4450.Pp
4451.Dl "ipfw pipe 1 config bw 1mbits/s pie"
4452.Dl "ipfw add 100 pipe 1 ip from 192.168.0.0/24 to any"
4453.Pp
4454To configure a
4455.Cm queue
4456with
4457.Cm pie
4458AQM using different configuration parameters for traffic from
4459192.168.0.0/24 and 1Mbits/s rate limit, we do:
4460.Pp
4461.Dl "ipfw pipe 1 config bw 1mbits/s"
4462.Dl "ipfw queue 1 config pipe 1 pie target 20ms tupdate 30ms ecn"
4463.Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4464.Pp
4465.Cm fq_codel
4466and
4467.Cm fq_pie
4468AQM can be configured for
4469.Nm dummynet
4470schedulers.
4471.Pp
4472To configure
4473.Cm fq_codel
4474scheduler using different configurations parameters for traffic from
4475192.168.0.0/24 and 1Mbits/s rate limit, we do:
4476.Pp
4477.Dl "ipfw pipe 1 config bw 1mbits/s"
4478.Dl "ipfw sched 1 config pipe 1 type fq_codel"
4479.Dl "ipfw queue 1 config sched 1"
4480.Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4481.Pp
4482To change
4483.Cm fq_codel
4484default configuration for a
4485.Cm sched
4486such as disable ECN and change the
4487.Ar target
4488to 10ms, we do:
4489.Pp
4490.Dl "ipfw sched 1 config pipe 1 type fq_codel target 10ms noecn"
4491.Pp
4492Similar to
4493.Cm fq_codel,
4494to configure
4495.Cm fq_pie
4496scheduler using different configurations parameters for traffic from
4497192.168.0.0/24 and 1Mbits/s rate limit, we do:
4498.Pp
4499.Dl "ipfw pipe 1 config bw 1mbits/s"
4500.Dl "ipfw sched 1 config pipe 1 type fq_pie"
4501.Dl "ipfw queue 1 config sched 1"
4502.Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4503.Pp
4504The configurations of
4505.Cm fq_pie
4506.Cm sched
4507can be changed in a similar way as for
4508.Cm fq_codel
4509.Sh SEE ALSO
4510.Xr cpp 1 ,
4511.Xr m4 1 ,
4512.Xr altq 4 ,
4513.Xr divert 4 ,
4514.Xr dummynet 4 ,
4515.Xr if_bridge 4 ,
4516.Xr ip 4 ,
4517.Xr ipfirewall 4 ,
4518.Xr ng_ipfw 4 ,
4519.Xr protocols 5 ,
4520.Xr services 5 ,
4521.Xr init 8 ,
4522.Xr kldload 8 ,
4523.Xr reboot 8 ,
4524.Xr sysctl 8 ,
4525.Xr syslogd 8
4526.Sh HISTORY
4527The
4528.Nm
4529utility first appeared in
4530.Fx 2.0 .
4531.Nm dummynet
4532was introduced in
4533.Fx 2.2.8 .
4534Stateful extensions were introduced in
4535.Fx 4.0 .
4536.Nm ipfw2
4537was introduced in Summer 2002.
4538.Sh AUTHORS
4539.An Ugen J. S. Antsilevich ,
4540.An Poul-Henning Kamp ,
4541.An Alex Nash ,
4542.An Archie Cobbs ,
4543.An Luigi Rizzo ,
4544.An Rasool Al-Saadi .
4545.Pp
4546.An -nosplit
4547API based upon code written by
4548.An Daniel Boulet
4549for BSDI.
4550.Pp
4551Dummynet has been introduced by Luigi Rizzo in 1997-1998.
4552.Pp
4553Some early work (1999-2000) on the
4554.Nm dummynet
4555traffic shaper supported by Akamba Corp.
4556.Pp
4557The ipfw core (ipfw2) has been completely redesigned and
4558reimplemented by Luigi Rizzo in summer 2002.
4559Further
4560actions and
4561options have been added by various developer over the years.
4562.Pp
4563.An -nosplit
4564In-kernel NAT support written by
4565.An Paolo Pisati Aq Mt piso@FreeBSD.org
4566as part of a Summer of Code 2005 project.
4567.Pp
4568SCTP
4569.Nm nat
4570support has been developed by
4571.An The Centre for Advanced Internet Architectures (CAIA) Aq http://www.caia.swin.edu.au .
4572The primary developers and maintainers are David Hayes and Jason But.
4573For further information visit:
4574.Aq http://www.caia.swin.edu.au/urp/SONATA
4575.Pp
4576Delay profiles have been developed by Alessandro Cerri and
4577Luigi Rizzo, supported by the
4578European Commission within Projects Onelab and Onelab2.
4579.Pp
4580CoDel, PIE, FQ-CoDel and FQ-PIE AQM for Dummynet have been implemented by
4581.An The Centre for Advanced Internet Architectures (CAIA)
4582in 2016, supported by The Comcast Innovation Fund.
4583The primary developer is
4584Rasool Al-Saadi.
4585.Sh BUGS
4586The syntax has grown over the years and sometimes it might be confusing.
4587Unfortunately, backward compatibility prevents cleaning up mistakes
4588made in the definition of the syntax.
4589.Pp
4590.Em !!! WARNING !!!
4591.Pp
4592Misconfiguring the firewall can put your computer in an unusable state,
4593possibly shutting down network services and requiring console access to
4594regain control of it.
4595.Pp
4596Incoming packet fragments diverted by
4597.Cm divert
4598are reassembled before delivery to the socket.
4599The action used on those packet is the one from the
4600rule which matches the first fragment of the packet.
4601.Pp
4602Packets diverted to userland, and then reinserted by a userland process
4603may lose various packet attributes.
4604The packet source interface name
4605will be preserved if it is shorter than 8 bytes and the userland process
4606saves and reuses the sockaddr_in
4607(as does
4608.Xr natd 8 ) ;
4609otherwise, it may be lost.
4610If a packet is reinserted in this manner, later rules may be incorrectly
4611applied, making the order of
4612.Cm divert
4613rules in the rule sequence very important.
4614.Pp
4615Dummynet drops all packets with IPv6 link-local addresses.
4616.Pp
4617Rules using
4618.Cm uid
4619or
4620.Cm gid
4621may not behave as expected.
4622In particular, incoming SYN packets may
4623have no uid or gid associated with them since they do not yet belong
4624to a TCP connection, and the uid/gid associated with a packet may not
4625be as expected if the associated process calls
4626.Xr setuid 2
4627or similar system calls.
4628.Pp
4629Rule syntax is subject to the command line environment and some patterns
4630may need to be escaped with the backslash character
4631or quoted appropriately.
4632.Pp
4633Due to the architecture of
4634.Xr libalias 3 ,
4635ipfw nat is not compatible with the TCP segmentation offloading (TSO).
4636Thus, to reliably nat your network traffic, please disable TSO
4637on your NICs using
4638.Xr ifconfig 8 .
4639.Pp
4640ICMP error messages are not implicitly matched by dynamic rules
4641for the respective conversations.
4642To avoid failures of network error detection and path MTU discovery,
4643ICMP error messages may need to be allowed explicitly through static
4644rules.
4645.Pp
4646Rules using
4647.Cm call
4648and
4649.Cm return
4650actions may lead to confusing behaviour if ruleset has mistakes,
4651and/or interaction with other subsystems (netgraph, dummynet, etc.) is used.
4652One possible case for this is packet leaving
4653.Nm
4654in subroutine on the input pass, while later on output encountering unpaired
4655.Cm return
4656first.
4657As the call stack is kept intact after input pass, packet will suddenly
4658return to the rule number used on input pass, not on output one.
4659Order of processing should be checked carefully to avoid such mistakes.
4660