xref: /freebsd/share/man/man5/pf.conf.5 (revision f3c5273d315a64826d2149ac453ff8c4583ddbe8)
1.\"    $FreeBSD$
2.\"	$OpenBSD: pf.conf.5,v 1.406 2009/01/31 19:37:12 sobrado Exp $
3.\"
4.\" Copyright (c) 2002, Daniel Hartmeier
5.\" All rights reserved.
6.\"
7.\" Redistribution and use in source and binary forms, with or without
8.\" modification, are permitted provided that the following conditions
9.\" are met:
10.\"
11.\"    - Redistributions of source code must retain the above copyright
12.\"      notice, this list of conditions and the following disclaimer.
13.\"    - Redistributions in binary form must reproduce the above
14.\"      copyright notice, this list of conditions and the following
15.\"      disclaimer in the documentation and/or other materials provided
16.\"      with the distribution.
17.\"
18.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19.\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
21.\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
22.\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
23.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
24.\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
25.\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
26.\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
28.\" ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29.\" POSSIBILITY OF SUCH DAMAGE.
30.\"
31.Dd January 5, 2019
32.Dt PF.CONF 5
33.Os
34.Sh NAME
35.Nm pf.conf
36.Nd packet filter configuration file
37.Sh DESCRIPTION
38The
39.Xr pf 4
40packet filter modifies, drops or passes packets according to rules or
41definitions specified in
42.Nm pf.conf .
43.Sh STATEMENT ORDER
44There are seven types of statements in
45.Nm pf.conf :
46.Bl -tag -width xxxx
47.It Cm Macros
48User-defined variables may be defined and used later, simplifying
49the configuration file.
50Macros must be defined before they are referenced in
51.Nm pf.conf .
52.It Cm Tables
53Tables provide a mechanism for increasing the performance and flexibility of
54rules with large numbers of source or destination addresses.
55.It Cm Options
56Options tune the behaviour of the packet filtering engine.
57.It Cm Traffic Normalization Li (e.g. Em scrub )
58Traffic normalization protects internal machines against inconsistencies
59in Internet protocols and implementations.
60.It Cm Queueing
61Queueing provides rule-based bandwidth control.
62.It Cm Translation Li (Various forms of NAT)
63Translation rules specify how addresses are to be mapped or redirected to
64other addresses.
65.It Cm Packet Filtering
66Packet filtering provides rule-based blocking or passing of packets.
67.El
68.Pp
69With the exception of
70.Cm macros
71and
72.Cm tables ,
73the types of statements should be grouped and appear in
74.Nm pf.conf
75in the order shown above, as this matches the operation of the underlying
76packet filtering engine.
77By default
78.Xr pfctl 8
79enforces this order (see
80.Ar set require-order
81below).
82.Pp
83Comments can be put anywhere in the file using a hash mark
84.Pq Sq # ,
85and extend to the end of the current line.
86.Pp
87Additional configuration files can be included with the
88.Ic include
89keyword, for example:
90.Bd -literal -offset indent
91include "/etc/pf/sub.filter.conf"
92.Ed
93.Sh MACROS
94Macros can be defined that will later be expanded in context.
95Macro names must start with a letter, and may contain letters, digits
96and underscores.
97Macro names may not be reserved words (for example
98.Ar pass ,
99.Ar in ,
100.Ar out ) .
101Macros are not expanded inside quotes.
102.Pp
103For example,
104.Bd -literal -offset indent
105ext_if = \&"kue0\&"
106all_ifs = \&"{\&" $ext_if lo0 \&"}\&"
107pass out on $ext_if from any to any
108pass in  on $ext_if proto tcp from any to any port 25
109.Ed
110.Sh TABLES
111Tables are named structures which can hold a collection of addresses and
112networks.
113Lookups against tables in
114.Xr pf 4
115are relatively fast, making a single rule with tables much more efficient,
116in terms of
117processor usage and memory consumption, than a large number of rules which
118differ only in IP address (either created explicitly or automatically by rule
119expansion).
120.Pp
121Tables can be used as the source or destination of filter rules,
122.Ar scrub
123rules
124or
125translation rules such as
126.Ar nat
127or
128.Ar rdr
129(see below for details on the various rule types).
130Tables can also be used for the redirect address of
131.Ar nat
132and
133.Ar rdr
134rules and in the routing options of filter rules, but only for
135.Ar round-robin
136pools.
137.Pp
138Tables can be defined with any of the following
139.Xr pfctl 8
140mechanisms.
141As with macros, reserved words may not be used as table names.
142.Bl -tag -width "manually"
143.It Ar manually
144Persistent tables can be manually created with the
145.Ar add
146or
147.Ar replace
148option of
149.Xr pfctl 8 ,
150before or after the ruleset has been loaded.
151.It Pa pf.conf
152Table definitions can be placed directly in this file, and loaded at the
153same time as other rules are loaded, atomically.
154Table definitions inside
155.Nm pf.conf
156use the
157.Ar table
158statement, and are especially useful to define non-persistent tables.
159The contents of a pre-existing table defined without a list of addresses
160to initialize it is not altered when
161.Nm pf.conf
162is loaded.
163A table initialized with the empty list,
164.Li { } ,
165will be cleared on load.
166.El
167.Pp
168Tables may be defined with the following attributes:
169.Bl -tag -width persist
170.It Ar persist
171The
172.Ar persist
173flag forces the kernel to keep the table even when no rules refer to it.
174If the flag is not set, the kernel will automatically remove the table
175when the last rule referring to it is flushed.
176.It Ar const
177The
178.Ar const
179flag prevents the user from altering the contents of the table once it
180has been created.
181Without that flag,
182.Xr pfctl 8
183can be used to add or remove addresses from the table at any time, even
184when running with
185.Xr securelevel 7
186= 2.
187.It Ar counters
188The
189.Ar counters
190flag enables per-address packet and byte counters which can be displayed with
191.Xr pfctl 8 .
192.El
193.Pp
194For example,
195.Bd -literal -offset indent
196table \*(Ltprivate\*(Gt const { 10/8, 172.16/12, 192.168/16 }
197table \*(Ltbadhosts\*(Gt persist
198block on fxp0 from { \*(Ltprivate\*(Gt, \*(Ltbadhosts\*(Gt } to any
199.Ed
200.Pp
201creates a table called private, to hold RFC 1918 private network
202blocks, and a table called badhosts, which is initially empty.
203A filter rule is set up to block all traffic coming from addresses listed in
204either table.
205The private table cannot have its contents changed and the badhosts table
206will exist even when no active filter rules reference it.
207Addresses may later be added to the badhosts table, so that traffic from
208these hosts can be blocked by using
209.Bd -literal -offset indent
210# pfctl -t badhosts -Tadd 204.92.77.111
211.Ed
212.Pp
213A table can also be initialized with an address list specified in one or more
214external files, using the following syntax:
215.Bd -literal -offset indent
216table \*(Ltspam\*(Gt persist file \&"/etc/spammers\&" file \&"/etc/openrelays\&"
217block on fxp0 from \*(Ltspam\*(Gt to any
218.Ed
219.Pp
220The files
221.Pa /etc/spammers
222and
223.Pa /etc/openrelays
224list IP addresses, one per line.
225Any lines beginning with a # are treated as comments and ignored.
226In addition to being specified by IP address, hosts may also be
227specified by their hostname.
228When the resolver is called to add a hostname to a table,
229.Em all
230resulting IPv4 and IPv6 addresses are placed into the table.
231IP addresses can also be entered in a table by specifying a valid interface
232name, a valid interface group or the
233.Em self
234keyword, in which case all addresses assigned to the interface(s) will be
235added to the table.
236.Sh OPTIONS
237.Xr pf 4
238may be tuned for various situations using the
239.Ar set
240command.
241.Bl -tag -width xxxx
242.It Ar set timeout
243.Pp
244.Bl -tag -width "src.track" -compact
245.It Ar interval
246Interval between purging expired states and fragments.
247.It Ar frag
248Seconds before an unassembled fragment is expired.
249.It Ar src.track
250Length of time to retain a source tracking entry after the last state
251expires.
252.El
253.Pp
254When a packet matches a stateful connection, the seconds to live for the
255connection will be updated to that of the
256.Ar proto.modifier
257which corresponds to the connection state.
258Each packet which matches this state will reset the TTL.
259Tuning these values may improve the performance of the
260firewall at the risk of dropping valid idle connections.
261.Pp
262.Bl -tag -width xxxx -compact
263.It Ar tcp.first
264The state after the first packet.
265.It Ar tcp.opening
266The state before the destination host ever sends a packet.
267.It Ar tcp.established
268The fully established state.
269.It Ar tcp.closing
270The state after the first FIN has been sent.
271.It Ar tcp.finwait
272The state after both FINs have been exchanged and the connection is closed.
273Some hosts (notably web servers on Solaris) send TCP packets even after closing
274the connection.
275Increasing
276.Ar tcp.finwait
277(and possibly
278.Ar tcp.closing )
279can prevent blocking of such packets.
280.It Ar tcp.closed
281The state after one endpoint sends an RST.
282.El
283.Pp
284ICMP and UDP are handled in a fashion similar to TCP, but with a much more
285limited set of states:
286.Pp
287.Bl -tag -width xxxx -compact
288.It Ar udp.first
289The state after the first packet.
290.It Ar udp.single
291The state if the source host sends more than one packet but the destination
292host has never sent one back.
293.It Ar udp.multiple
294The state if both hosts have sent packets.
295.It Ar icmp.first
296The state after the first packet.
297.It Ar icmp.error
298The state after an ICMP error came back in response to an ICMP packet.
299.El
300.Pp
301Other protocols are handled similarly to UDP:
302.Pp
303.Bl -tag -width xxxx -compact
304.It Ar other.first
305.It Ar other.single
306.It Ar other.multiple
307.El
308.Pp
309Timeout values can be reduced adaptively as the number of state table
310entries grows.
311.Pp
312.Bl -tag -width xxxx -compact
313.It Ar adaptive.start
314When the number of state entries exceeds this value, adaptive scaling
315begins.
316All timeout values are scaled linearly with factor
317(adaptive.end - number of states) / (adaptive.end - adaptive.start).
318.It Ar adaptive.end
319When reaching this number of state entries, all timeout values become
320zero, effectively purging all state entries immediately.
321This value is used to define the scale factor, it should not actually
322be reached (set a lower state limit, see below).
323.El
324.Pp
325Adaptive timeouts are enabled by default, with an adaptive.start value
326equal to 60% of the state limit, and an adaptive.end value equal to
327120% of the state limit.
328They can be disabled by setting both adaptive.start and adaptive.end to 0.
329.Pp
330The adaptive timeout values can be defined both globally and for each rule.
331When used on a per-rule basis, the values relate to the number of
332states created by the rule, otherwise to the total number of
333states.
334.Pp
335For example:
336.Bd -literal -offset indent
337set timeout tcp.first 120
338set timeout tcp.established 86400
339set timeout { adaptive.start 6000, adaptive.end 12000 }
340set limit states 10000
341.Ed
342.Pp
343With 9000 state table entries, the timeout values are scaled to 50%
344(tcp.first 60, tcp.established 43200).
345.It Ar set loginterface
346Enable collection of packet and byte count statistics for the given
347interface or interface group.
348These statistics can be viewed using
349.Bd -literal -offset indent
350# pfctl -s info
351.Ed
352.Pp
353In this example
354.Xr pf 4
355collects statistics on the interface named dc0:
356.Bd -literal -offset indent
357set loginterface dc0
358.Ed
359.Pp
360One can disable the loginterface using:
361.Bd -literal -offset indent
362set loginterface none
363.Ed
364.It Ar set limit
365Sets hard limits on the memory pools used by the packet filter.
366See
367.Xr zone 9
368for an explanation of memory pools.
369.Pp
370For example,
371.Bd -literal -offset indent
372set limit states 20000
373.Ed
374.Pp
375sets the maximum number of entries in the memory pool used by state table
376entries (generated by
377.Ar pass
378rules which do not specify
379.Ar no state )
380to 20000.
381Using
382.Bd -literal -offset indent
383set limit frags 20000
384.Ed
385.Pp
386sets the maximum number of entries in the memory pool used for fragment
387reassembly (generated by
388.Ar scrub
389rules) to 20000.
390Using
391.Bd -literal -offset indent
392set limit src-nodes 2000
393.Ed
394.Pp
395sets the maximum number of entries in the memory pool used for tracking
396source IP addresses (generated by the
397.Ar sticky-address
398and
399.Ar src.track
400options) to 2000.
401Using
402.Bd -literal -offset indent
403set limit tables 1000
404set limit table-entries 100000
405.Ed
406.Pp
407sets limits on the memory pools used by tables.
408The first limits the number of tables that can exist to 1000.
409The second limits the overall number of addresses that can be stored
410in tables to 100000.
411.Pp
412Various limits can be combined on a single line:
413.Bd -literal -offset indent
414set limit { states 20000, frags 20000, src-nodes 2000 }
415.Ed
416.It Ar set ruleset-optimization
417.Bl -tag -width xxxxxxxx -compact
418.It Ar none
419Disable the ruleset optimizer.
420.It Ar basic
421Enable basic ruleset optimization.
422This is the default behaviour.
423Basic ruleset optimization does four things to improve the
424performance of ruleset evaluations:
425.Pp
426.Bl -enum -compact
427.It
428remove duplicate rules
429.It
430remove rules that are a subset of another rule
431.It
432combine multiple rules into a table when advantageous
433.It
434re-order the rules to improve evaluation performance
435.El
436.Pp
437.It Ar profile
438Uses the currently loaded ruleset as a feedback profile to tailor the
439ordering of quick rules to actual network traffic.
440.El
441.Pp
442It is important to note that the ruleset optimizer will modify the ruleset
443to improve performance.
444A side effect of the ruleset modification is that per-rule accounting
445statistics will have different meanings than before.
446If per-rule accounting is important for billing purposes or whatnot,
447either the ruleset optimizer should not be used or a label field should
448be added to all of the accounting rules to act as optimization barriers.
449.Pp
450Optimization can also be set as a command-line argument to
451.Xr pfctl 8 ,
452overriding the settings in
453.Nm .
454.It Ar set optimization
455Optimize state timeouts for one of the following network environments:
456.Pp
457.Bl -tag -width xxxx -compact
458.It Ar normal
459A normal network environment.
460Suitable for almost all networks.
461.It Ar high-latency
462A high-latency environment (such as a satellite connection).
463.It Ar satellite
464Alias for
465.Ar high-latency .
466.It Ar aggressive
467Aggressively expire connections.
468This can greatly reduce the memory usage of the firewall at the cost of
469dropping idle connections early.
470.It Ar conservative
471Extremely conservative settings.
472Avoid dropping legitimate connections at the
473expense of greater memory utilization (possibly much greater on a busy
474network) and slightly increased processor utilization.
475.El
476.Pp
477For example:
478.Bd -literal -offset indent
479set optimization aggressive
480.Ed
481.It Ar set block-policy
482The
483.Ar block-policy
484option sets the default behaviour for the packet
485.Ar block
486action:
487.Pp
488.Bl -tag -width xxxxxxxx -compact
489.It Ar drop
490Packet is silently dropped.
491.It Ar return
492A TCP RST is returned for blocked TCP packets,
493an ICMP UNREACHABLE is returned for blocked UDP packets,
494and all other packets are silently dropped.
495.El
496.Pp
497For example:
498.Bd -literal -offset indent
499set block-policy return
500.Ed
501
502.It Ar set fail-policy
503The
504.Ar fail-policy
505option sets the behaviour of rules which should pass a packet but were unable to
506do so. This might happen when a nat or route-to rule uses an empty table as list
507of targets or if a rule fails to create state or source node.
508The following
509.Ar block
510actions are possible:
511.Pp
512.Bl -tag -width xxxxxxxx -compact
513.It Ar drop
514Incoming packet is silently dropped.
515.It Ar return
516Incoming packet is dropped and TCP RST is returned for TCP packets,
517an ICMP UNREACHABLE is returned for UDP packets,
518and no response is sent for other packets.
519.El
520.Pp
521For example:
522.Bd -literal -offset indent
523set fail-policy return
524.Ed
525
526.It Ar set state-policy
527The
528.Ar state-policy
529option sets the default behaviour for states:
530.Pp
531.Bl -tag -width group-bound -compact
532.It Ar if-bound
533States are bound to interface.
534.It Ar floating
535States can match packets on any interfaces (the default).
536.El
537.Pp
538For example:
539.Bd -literal -offset indent
540set state-policy if-bound
541.Ed
542.It Ar set state-defaults
543The
544.Ar state-defaults
545option sets the state options for states created from rules
546without an explicit
547.Ar keep state .
548For example:
549.Bd -literal -offset indent
550set state-defaults no-sync
551.Ed
552.It Ar set hostid
553The 32-bit
554.Ar hostid
555identifies this firewall's state table entries to other firewalls
556in a
557.Xr pfsync 4
558failover cluster.
559By default the hostid is set to a pseudo-random value, however it may be
560desirable to manually configure it, for example to more easily identify the
561source of state table entries.
562.Bd -literal -offset indent
563set hostid 1
564.Ed
565.Pp
566The hostid may be specified in either decimal or hexadecimal.
567.It Ar set require-order
568By default
569.Xr pfctl 8
570enforces an ordering of the statement types in the ruleset to:
571.Em options ,
572.Em normalization ,
573.Em queueing ,
574.Em translation ,
575.Em filtering .
576Setting this option to
577.Ar no
578disables this enforcement.
579There may be non-trivial and non-obvious implications to an out of
580order ruleset.
581Consider carefully before disabling the order enforcement.
582.It Ar set fingerprints
583Load fingerprints of known operating systems from the given filename.
584By default fingerprints of known operating systems are automatically
585loaded from
586.Xr pf.os 5
587in
588.Pa /etc
589but can be overridden via this option.
590Setting this option may leave a small period of time where the fingerprints
591referenced by the currently active ruleset are inconsistent until the new
592ruleset finishes loading.
593.Pp
594For example:
595.Pp
596.Dl set fingerprints \&"/etc/pf.os.devel\&"
597.It Ar set skip on Aq Ar ifspec
598List interfaces for which packets should not be filtered.
599Packets passing in or out on such interfaces are passed as if pf was
600disabled, i.e. pf does not process them in any way.
601This can be useful on loopback and other virtual interfaces, when
602packet filtering is not desired and can have unexpected effects.
603For example:
604.Pp
605.Dl set skip on lo0
606.It Ar set debug
607Set the debug
608.Ar level
609to one of the following:
610.Pp
611.Bl -tag -width xxxxxxxxxxxx -compact
612.It Ar none
613Don't generate debug messages.
614.It Ar urgent
615Generate debug messages only for serious errors.
616.It Ar misc
617Generate debug messages for various errors.
618.It Ar loud
619Generate debug messages for common conditions.
620.El
621.El
622.Sh TRAFFIC NORMALIZATION
623Traffic normalization is used to sanitize packet content in such
624a way that there are no ambiguities in packet interpretation on
625the receiving side.
626The normalizer does IP fragment reassembly to prevent attacks
627that confuse intrusion detection systems by sending overlapping
628IP fragments.
629Packet normalization is invoked with the
630.Ar scrub
631directive.
632.Pp
633.Ar scrub
634has the following options:
635.Bl -tag -width xxxx
636.It Ar no-df
637Clears the
638.Ar dont-fragment
639bit from a matching IP packet.
640Some operating systems are known to generate fragmented packets with the
641.Ar dont-fragment
642bit set.
643This is particularly true with NFS.
644.Ar Scrub
645will drop such fragmented
646.Ar dont-fragment
647packets unless
648.Ar no-df
649is specified.
650.Pp
651Unfortunately some operating systems also generate their
652.Ar dont-fragment
653packets with a zero IP identification field.
654Clearing the
655.Ar dont-fragment
656bit on packets with a zero IP ID may cause deleterious results if an
657upstream router later fragments the packet.
658Using the
659.Ar random-id
660modifier (see below) is recommended in combination with the
661.Ar no-df
662modifier to ensure unique IP identifiers.
663.It Ar min-ttl Aq Ar number
664Enforces a minimum TTL for matching IP packets.
665.It Ar max-mss Aq Ar number
666Enforces a maximum MSS for matching TCP packets.
667.It Xo Ar set-tos Aq Ar string
668.No \*(Ba Aq Ar number
669.Xc
670Enforces a
671.Em TOS
672for matching IP packets.
673.Em TOS
674may be
675given as one of
676.Ar critical ,
677.Ar inetcontrol ,
678.Ar lowdelay ,
679.Ar netcontrol ,
680.Ar throughput ,
681.Ar reliability ,
682or one of the DiffServ Code Points:
683.Ar ef ,
684.Ar af11 No ... Ar af43 ,
685.Ar cs0 No ... Ar cs7 ;
686or as either hex or decimal.
687.It Ar random-id
688Replaces the IP identification field with random values to compensate
689for predictable values generated by many hosts.
690This option only applies to packets that are not fragmented
691after the optional fragment reassembly.
692.It Ar fragment reassemble
693Using
694.Ar scrub
695rules, fragments can be reassembled by normalization.
696In this case, fragments are buffered until they form a complete
697packet, and only the completed packet is passed on to the filter.
698The advantage is that filter rules have to deal only with complete
699packets, and can ignore fragments.
700The drawback of caching fragments is the additional memory cost.
701.It Ar reassemble tcp
702Statefully normalizes TCP connections.
703.Ar scrub reassemble tcp
704rules may not have the direction (in/out) specified.
705.Ar reassemble tcp
706performs the following normalizations:
707.Pp
708.Bl -tag -width timeout -compact
709.It ttl
710Neither side of the connection is allowed to reduce their IP TTL.
711An attacker may send a packet such that it reaches the firewall, affects
712the firewall state, and expires before reaching the destination host.
713.Ar reassemble tcp
714will raise the TTL of all packets back up to the highest value seen on
715the connection.
716.It timestamp modulation
717Modern TCP stacks will send a timestamp on every TCP packet and echo
718the other endpoint's timestamp back to them.
719Many operating systems will merely start the timestamp at zero when
720first booted, and increment it several times a second.
721The uptime of the host can be deduced by reading the timestamp and multiplying
722by a constant.
723Also observing several different timestamps can be used to count hosts
724behind a NAT device.
725And spoofing TCP packets into a connection requires knowing or guessing
726valid timestamps.
727Timestamps merely need to be monotonically increasing and not derived off a
728guessable base time.
729.Ar reassemble tcp
730will cause
731.Ar scrub
732to modulate the TCP timestamps with a random number.
733.It extended PAWS checks
734There is a problem with TCP on long fat pipes, in that a packet might get
735delayed for longer than it takes the connection to wrap its 32-bit sequence
736space.
737In such an occurrence, the old packet would be indistinguishable from a
738new packet and would be accepted as such.
739The solution to this is called PAWS: Protection Against Wrapped Sequence
740numbers.
741It protects against it by making sure the timestamp on each packet does
742not go backwards.
743.Ar reassemble tcp
744also makes sure the timestamp on the packet does not go forward more
745than the RFC allows.
746By doing this,
747.Xr pf 4
748artificially extends the security of TCP sequence numbers by 10 to 18
749bits when the host uses appropriately randomized timestamps, since a
750blind attacker would have to guess the timestamp as well.
751.El
752.El
753.Pp
754For example,
755.Bd -literal -offset indent
756scrub in on $ext_if all fragment reassemble
757.Ed
758.Pp
759The
760.Ar no
761option prefixed to a scrub rule causes matching packets to remain unscrubbed,
762much in the same way as
763.Ar drop quick
764works in the packet filter (see below).
765This mechanism should be used when it is necessary to exclude specific packets
766from broader scrub rules.
767.Sh QUEUEING
768The ALTQ system is currently not available in the GENERIC kernel nor as
769loadable modules.
770In order to use the herein after called queueing options one has to use a
771custom built kernel.
772Please refer to
773.Xr altq 4
774to learn about the related kernel options.
775.Pp
776Packets can be assigned to queues for the purpose of bandwidth
777control.
778At least two declarations are required to configure queues, and later
779any packet filtering rule can reference the defined queues by name.
780During the filtering component of
781.Nm pf.conf ,
782the last referenced
783.Ar queue
784name is where any packets from
785.Ar pass
786rules will be queued, while for
787.Ar block
788rules it specifies where any resulting ICMP or TCP RST
789packets should be queued.
790The
791.Ar scheduler
792defines the algorithm used to decide which packets get delayed, dropped, or
793sent out immediately.
794There are three
795.Ar schedulers
796currently supported.
797.Bl -tag -width xxxx
798.It Ar cbq
799Class Based Queueing.
800.Ar Queues
801attached to an interface build a tree, thus each
802.Ar queue
803can have further child
804.Ar queues .
805Each queue can have a
806.Ar priority
807and a
808.Ar bandwidth
809assigned.
810.Ar Priority
811mainly controls the time packets take to get sent out, while
812.Ar bandwidth
813has primarily effects on throughput.
814.Ar cbq
815achieves both partitioning and sharing of link bandwidth
816by hierarchically structured classes.
817Each class has its own
818.Ar queue
819and is assigned its share of
820.Ar bandwidth .
821A child class can borrow bandwidth from its parent class
822as long as excess bandwidth is available
823(see the option
824.Ar borrow ,
825below).
826.It Ar priq
827Priority Queueing.
828.Ar Queues
829are flat attached to the interface, thus,
830.Ar queues
831cannot have further child
832.Ar queues .
833Each
834.Ar queue
835has a unique
836.Ar priority
837assigned, ranging from 0 to 15.
838Packets in the
839.Ar queue
840with the highest
841.Ar priority
842are processed first.
843.It Ar hfsc
844Hierarchical Fair Service Curve.
845.Ar Queues
846attached to an interface build a tree, thus each
847.Ar queue
848can have further child
849.Ar queues .
850Each queue can have a
851.Ar priority
852and a
853.Ar bandwidth
854assigned.
855.Ar Priority
856mainly controls the time packets take to get sent out, while
857.Ar bandwidth
858primarily affects throughput.
859.Ar hfsc
860supports both link-sharing and guaranteed real-time services.
861It employs a service curve based QoS model,
862and its unique feature is an ability to decouple
863.Ar delay
864and
865.Ar bandwidth
866allocation.
867.El
868.Pp
869The interfaces on which queueing should be activated are declared using
870the
871.Ar altq on
872declaration.
873.Ar altq on
874has the following keywords:
875.Bl -tag -width xxxx
876.It Aq Ar interface
877Queueing is enabled on the named interface.
878.It Aq Ar scheduler
879Specifies which queueing scheduler to use.
880Currently supported values
881are
882.Ar cbq
883for Class Based Queueing,
884.Ar priq
885for Priority Queueing and
886.Ar hfsc
887for the Hierarchical Fair Service Curve scheduler.
888.It Ar bandwidth Aq Ar bw
889The maximum bitrate for all queues on an
890interface may be specified using the
891.Ar bandwidth
892keyword.
893The value can be specified as an absolute value or as a
894percentage of the interface bandwidth.
895When using an absolute value, the suffixes
896.Ar b ,
897.Ar Kb ,
898.Ar Mb ,
899and
900.Ar Gb
901are used to represent bits, kilobits, megabits, and
902gigabits per second, respectively.
903The value must not exceed the interface bandwidth.
904If
905.Ar bandwidth
906is not specified, the interface bandwidth is used
907(but take note that some interfaces do not know their bandwidth,
908or can adapt their bandwidth rates).
909.It Ar qlimit Aq Ar limit
910The maximum number of packets held in the queue.
911The default is 50.
912.It Ar tbrsize Aq Ar size
913Adjusts the size, in bytes, of the token bucket regulator.
914If not specified, heuristics based on the
915interface bandwidth are used to determine the size.
916.It Ar queue Aq Ar list
917Defines a list of subqueues to create on an interface.
918.El
919.Pp
920In the following example, the interface dc0
921should queue up to 5Mbps in four second-level queues using
922Class Based Queueing.
923Those four queues will be shown in a later example.
924.Bd -literal -offset indent
925altq on dc0 cbq bandwidth 5Mb queue { std, http, mail, ssh }
926.Ed
927.Pp
928Once interfaces are activated for queueing using the
929.Ar altq
930directive, a sequence of
931.Ar queue
932directives may be defined.
933The name associated with a
934.Ar queue
935must match a queue defined in the
936.Ar altq
937directive (e.g. mail), or, except for the
938.Ar priq
939.Ar scheduler ,
940in a parent
941.Ar queue
942declaration.
943The following keywords can be used:
944.Bl -tag -width xxxx
945.It Ar on Aq Ar interface
946Specifies the interface the queue operates on.
947If not given, it operates on all matching interfaces.
948.It Ar bandwidth Aq Ar bw
949Specifies the maximum bitrate to be processed by the queue.
950This value must not exceed the value of the parent
951.Ar queue
952and can be specified as an absolute value or a percentage of the parent
953queue's bandwidth.
954If not specified, defaults to 100% of the parent queue's bandwidth.
955The
956.Ar priq
957scheduler does not support bandwidth specification.
958.It Ar priority Aq Ar level
959Between queues a priority level can be set.
960For
961.Ar cbq
962and
963.Ar hfsc ,
964the range is 0 to 7 and for
965.Ar priq ,
966the range is 0 to 15.
967The default for all is 1.
968.Ar Priq
969queues with a higher priority are always served first.
970.Ar Cbq
971and
972.Ar Hfsc
973queues with a higher priority are preferred in the case of overload.
974.It Ar qlimit Aq Ar limit
975The maximum number of packets held in the queue.
976The default is 50.
977.El
978.Pp
979The
980.Ar scheduler
981can get additional parameters with
982.Xo Aq Ar scheduler
983.Pf ( Aq Ar parameters ) .
984.Xc
985Parameters are as follows:
986.Bl -tag -width Fl
987.It Ar default
988Packets not matched by another queue are assigned to this one.
989Exactly one default queue is required.
990.It Ar red
991Enable RED (Random Early Detection) on this queue.
992RED drops packets with a probability proportional to the average
993queue length.
994.It Ar rio
995Enables RIO on this queue.
996RIO is RED with IN/OUT, thus running
997RED two times more than RIO would achieve the same effect.
998RIO is currently not supported in the GENERIC kernel.
999.It Ar ecn
1000Enables ECN (Explicit Congestion Notification) on this queue.
1001ECN implies RED.
1002.El
1003.Pp
1004The
1005.Ar cbq
1006.Ar scheduler
1007supports an additional option:
1008.Bl -tag -width Fl
1009.It Ar borrow
1010The queue can borrow bandwidth from the parent.
1011.El
1012.Pp
1013The
1014.Ar hfsc
1015.Ar scheduler
1016supports some additional options:
1017.Bl -tag -width Fl
1018.It Ar realtime Aq Ar sc
1019The minimum required bandwidth for the queue.
1020.It Ar upperlimit Aq Ar sc
1021The maximum allowed bandwidth for the queue.
1022.It Ar linkshare Aq Ar sc
1023The bandwidth share of a backlogged queue.
1024.El
1025.Pp
1026.Aq Ar sc
1027is an acronym for
1028.Ar service curve .
1029.Pp
1030The format for service curve specifications is
1031.Ar ( m1 , d , m2 ) .
1032.Ar m2
1033controls the bandwidth assigned to the queue.
1034.Ar m1
1035and
1036.Ar d
1037are optional and can be used to control the initial bandwidth assignment.
1038For the first
1039.Ar d
1040milliseconds the queue gets the bandwidth given as
1041.Ar m1 ,
1042afterwards the value given in
1043.Ar m2 .
1044.Pp
1045Furthermore, with
1046.Ar cbq
1047and
1048.Ar hfsc ,
1049child queues can be specified as in an
1050.Ar altq
1051declaration, thus building a tree of queues using a part of
1052their parent's bandwidth.
1053.Pp
1054Packets can be assigned to queues based on filter rules by using the
1055.Ar queue
1056keyword.
1057Normally only one
1058.Ar queue
1059is specified; when a second one is specified it will instead be used for
1060packets which have a
1061.Em TOS
1062of
1063.Em lowdelay
1064and for TCP ACKs with no data payload.
1065.Pp
1066To continue the previous example, the examples below would specify the
1067four referenced
1068queues, plus a few child queues.
1069Interactive
1070.Xr ssh 1
1071sessions get priority over bulk transfers like
1072.Xr scp 1
1073and
1074.Xr sftp 1 .
1075The queues may then be referenced by filtering rules (see
1076.Sx PACKET FILTERING
1077below).
1078.Bd -literal
1079queue std bandwidth 10% cbq(default)
1080queue http bandwidth 60% priority 2 cbq(borrow red) \e
1081      { employees, developers }
1082queue  developers bandwidth 75% cbq(borrow)
1083queue  employees bandwidth 15%
1084queue mail bandwidth 10% priority 0 cbq(borrow ecn)
1085queue ssh bandwidth 20% cbq(borrow) { ssh_interactive, ssh_bulk }
1086queue  ssh_interactive bandwidth 50% priority 7 cbq(borrow)
1087queue  ssh_bulk bandwidth 50% priority 0 cbq(borrow)
1088
1089block return out on dc0 inet all queue std
1090pass out on dc0 inet proto tcp from $developerhosts to any port 80 \e
1091      queue developers
1092pass out on dc0 inet proto tcp from $employeehosts to any port 80 \e
1093      queue employees
1094pass out on dc0 inet proto tcp from any to any port 22 \e
1095      queue(ssh_bulk, ssh_interactive)
1096pass out on dc0 inet proto tcp from any to any port 25 \e
1097      queue mail
1098.Ed
1099.Sh TRANSLATION
1100Translation rules modify either the source or destination address of the
1101packets associated with a stateful connection.
1102A stateful connection is automatically created to track packets matching
1103such a rule as long as they are not blocked by the filtering section of
1104.Nm pf.conf .
1105The translation engine modifies the specified address and/or port in the
1106packet, recalculates IP, TCP and UDP checksums as necessary, and passes it to
1107the packet filter for evaluation.
1108.Pp
1109Since translation occurs before filtering the filter
1110engine will see packets as they look after any
1111addresses and ports have been translated.
1112Filter rules will therefore have to filter based on the translated
1113address and port number.
1114Packets that match a translation rule are only automatically passed if
1115the
1116.Ar pass
1117modifier is given, otherwise they are
1118still subject to
1119.Ar block
1120and
1121.Ar pass
1122rules.
1123.Pp
1124The state entry created permits
1125.Xr pf 4
1126to keep track of the original address for traffic associated with that state
1127and correctly direct return traffic for that connection.
1128.Pp
1129Various types of translation are possible with pf:
1130.Bl -tag -width xxxx
1131.It Ar binat
1132A
1133.Ar binat
1134rule specifies a bidirectional mapping between an external IP netblock
1135and an internal IP netblock.
1136.It Ar nat
1137A
1138.Ar nat
1139rule specifies that IP addresses are to be changed as the packet
1140traverses the given interface.
1141This technique allows one or more IP addresses
1142on the translating host to support network traffic for a larger range of
1143machines on an "inside" network.
1144Although in theory any IP address can be used on the inside, it is strongly
1145recommended that one of the address ranges defined by RFC 1918 be used.
1146These netblocks are:
1147.Bd -literal
114810.0.0.0 - 10.255.255.255 (all of net 10, i.e., 10/8)
1149172.16.0.0 - 172.31.255.255 (i.e., 172.16/12)
1150192.168.0.0 - 192.168.255.255 (i.e., 192.168/16)
1151.Ed
1152.It Pa rdr
1153The packet is redirected to another destination and possibly a
1154different port.
1155.Ar rdr
1156rules can optionally specify port ranges instead of single ports.
1157rdr ... port 2000:2999 -\*(Gt ... port 4000
1158redirects ports 2000 to 2999 (inclusive) to port 4000.
1159rdr ... port 2000:2999 -\*(Gt ... port 4000:*
1160redirects port 2000 to 4000, 2001 to 4001, ..., 2999 to 4999.
1161.El
1162.Pp
1163In addition to modifying the address, some translation rules may modify
1164source or destination ports for
1165.Xr tcp 4
1166or
1167.Xr udp 4
1168connections; implicitly in the case of
1169.Ar nat
1170rules and explicitly in the case of
1171.Ar rdr
1172rules.
1173Port numbers are never translated with a
1174.Ar binat
1175rule.
1176.Pp
1177Evaluation order of the translation rules is dependent on the type
1178of the translation rules and of the direction of a packet.
1179.Ar binat
1180rules are always evaluated first.
1181Then either the
1182.Ar rdr
1183rules are evaluated on an inbound packet or the
1184.Ar nat
1185rules on an outbound packet.
1186Rules of the same type are evaluated in the same order in which they
1187appear in the ruleset.
1188The first matching rule decides what action is taken.
1189.Pp
1190The
1191.Ar no
1192option prefixed to a translation rule causes packets to remain untranslated,
1193much in the same way as
1194.Ar drop quick
1195works in the packet filter (see below).
1196If no rule matches the packet it is passed to the filter engine unmodified.
1197.Pp
1198Translation rules apply only to packets that pass through
1199the specified interface, and if no interface is specified,
1200translation is applied to packets on all interfaces.
1201For instance, redirecting port 80 on an external interface to an internal
1202web server will only work for connections originating from the outside.
1203Connections to the address of the external interface from local hosts will
1204not be redirected, since such packets do not actually pass through the
1205external interface.
1206Redirections cannot reflect packets back through the interface they arrive
1207on, they can only be redirected to hosts connected to different interfaces
1208or to the firewall itself.
1209.Pp
1210Note that redirecting external incoming connections to the loopback
1211address, as in
1212.Bd -literal -offset indent
1213rdr on ne3 inet proto tcp to port smtp -\*(Gt 127.0.0.1 port spamd
1214.Ed
1215.Pp
1216will effectively allow an external host to connect to daemons
1217bound solely to the loopback address, circumventing the traditional
1218blocking of such connections on a real interface.
1219Unless this effect is desired, any of the local non-loopback addresses
1220should be used as redirection target instead, which allows external
1221connections only to daemons bound to this address or not bound to
1222any address.
1223.Pp
1224See
1225.Sx TRANSLATION EXAMPLES
1226below.
1227.Sh PACKET FILTERING
1228.Xr pf 4
1229has the ability to
1230.Ar block
1231and
1232.Ar pass
1233packets based on attributes of their layer 3 (see
1234.Xr ip 4
1235and
1236.Xr ip6 4 )
1237and layer 4 (see
1238.Xr icmp 4 ,
1239.Xr icmp6 4 ,
1240.Xr tcp 4 ,
1241.Xr udp 4 )
1242headers.
1243In addition, packets may also be
1244assigned to queues for the purpose of bandwidth control.
1245.Pp
1246For each packet processed by the packet filter, the filter rules are
1247evaluated in sequential order, from first to last.
1248The last matching rule decides what action is taken.
1249If no rule matches the packet, the default action is to pass
1250the packet.
1251.Pp
1252The following actions can be used in the filter:
1253.Bl -tag -width xxxx
1254.It Ar block
1255The packet is blocked.
1256There are a number of ways in which a
1257.Ar block
1258rule can behave when blocking a packet.
1259The default behaviour is to
1260.Ar drop
1261packets silently, however this can be overridden or made
1262explicit either globally, by setting the
1263.Ar block-policy
1264option, or on a per-rule basis with one of the following options:
1265.Pp
1266.Bl -tag -width xxxx -compact
1267.It Ar drop
1268The packet is silently dropped.
1269.It Ar return-rst
1270This applies only to
1271.Xr tcp 4
1272packets, and issues a TCP RST which closes the
1273connection.
1274.It Ar return-icmp
1275.It Ar return-icmp6
1276This causes ICMP messages to be returned for packets which match the rule.
1277By default this is an ICMP UNREACHABLE message, however this
1278can be overridden by specifying a message as a code or number.
1279.It Ar return
1280This causes a TCP RST to be returned for
1281.Xr tcp 4
1282packets and an ICMP UNREACHABLE for UDP and other packets.
1283.El
1284.Pp
1285Options returning ICMP packets currently have no effect if
1286.Xr pf 4
1287operates on a
1288.Xr if_bridge 4 ,
1289as the code to support this feature has not yet been implemented.
1290.Pp
1291The simplest mechanism to block everything by default and only pass
1292packets that match explicit rules is specify a first filter rule of:
1293.Bd -literal -offset indent
1294block all
1295.Ed
1296.It Ar pass
1297The packet is passed;
1298state is created unless the
1299.Ar no state
1300option is specified.
1301.El
1302.Pp
1303By default
1304.Xr pf 4
1305filters packets statefully; the first time a packet matches a
1306.Ar pass
1307rule, a state entry is created; for subsequent packets the filter checks
1308whether the packet matches any state.
1309If it does, the packet is passed without evaluation of any rules.
1310After the connection is closed or times out, the state entry is automatically
1311removed.
1312.Pp
1313This has several advantages.
1314For TCP connections, comparing a packet to a state involves checking
1315its sequence numbers, as well as TCP timestamps if a
1316.Ar scrub reassemble tcp
1317rule applies to the connection.
1318If these values are outside the narrow windows of expected
1319values, the packet is dropped.
1320This prevents spoofing attacks, such as when an attacker sends packets with
1321a fake source address/port but does not know the connection's sequence
1322numbers.
1323Similarly,
1324.Xr pf 4
1325knows how to match ICMP replies to states.
1326For example,
1327.Bd -literal -offset indent
1328pass out inet proto icmp all icmp-type echoreq
1329.Ed
1330.Pp
1331allows echo requests (such as those created by
1332.Xr ping 8 )
1333out statefully, and matches incoming echo replies correctly to states.
1334.Pp
1335Also, looking up states is usually faster than evaluating rules.
1336If there are 50 rules, all of them are evaluated sequentially in O(n).
1337Even with 50000 states, only 16 comparisons are needed to match a
1338state, since states are stored in a binary search tree that allows
1339searches in O(log2 n).
1340.Pp
1341Furthermore, correct handling of ICMP error messages is critical to
1342many protocols, particularly TCP.
1343.Xr pf 4
1344matches ICMP error messages to the correct connection, checks them against
1345connection parameters, and passes them if appropriate.
1346For example if an ICMP source quench message referring to a stateful TCP
1347connection arrives, it will be matched to the state and get passed.
1348.Pp
1349Finally, state tracking is required for
1350.Ar nat , binat No and Ar rdr
1351rules, in order to track address and port translations and reverse the
1352translation on returning packets.
1353.Pp
1354.Xr pf 4
1355will also create state for other protocols which are effectively stateless by
1356nature.
1357UDP packets are matched to states using only host addresses and ports,
1358and other protocols are matched to states using only the host addresses.
1359.Pp
1360If stateless filtering of individual packets is desired,
1361the
1362.Ar no state
1363keyword can be used to specify that state will not be created
1364if this is the last matching rule.
1365A number of parameters can also be set to affect how
1366.Xr pf 4
1367handles state tracking.
1368See
1369.Sx STATEFUL TRACKING OPTIONS
1370below for further details.
1371.Sh PARAMETERS
1372The rule parameters specify the packets to which a rule applies.
1373A packet always comes in on, or goes out through, one interface.
1374Most parameters are optional.
1375If a parameter is specified, the rule only applies to packets with
1376matching attributes.
1377Certain parameters can be expressed as lists, in which case
1378.Xr pfctl 8
1379generates all needed rule combinations.
1380.Bl -tag -width xxxx
1381.It Ar in No or Ar out
1382This rule applies to incoming or outgoing packets.
1383If neither
1384.Ar in
1385nor
1386.Ar out
1387are specified, the rule will match packets in both directions.
1388.It Ar log
1389In addition to the action specified, a log message is generated.
1390Only the packet that establishes the state is logged,
1391unless the
1392.Ar no state
1393option is specified.
1394The logged packets are sent to a
1395.Xr pflog 4
1396interface, by default
1397.Ar pflog0 .
1398This interface is monitored by the
1399.Xr pflogd 8
1400logging daemon, which dumps the logged packets to the file
1401.Pa /var/log/pflog
1402in
1403.Xr pcap 3
1404binary format.
1405.It Ar log (all)
1406Used to force logging of all packets for a connection.
1407This is not necessary when
1408.Ar no state
1409is explicitly specified.
1410As with
1411.Ar log ,
1412packets are logged to
1413.Xr pflog 4 .
1414.It Ar log (user)
1415Logs the
1416.Ux
1417user ID of the user that owns the socket and the PID of the process that
1418has the socket open where the packet is sourced from or destined to
1419(depending on which socket is local).
1420This is in addition to the normal information logged.
1421.Pp
1422Only the first packet
1423logged via
1424.Ar log (all, user)
1425will have the user credentials logged when using stateful matching.
1426.It Ar log (to Aq Ar interface )
1427Send logs to the specified
1428.Xr pflog 4
1429interface instead of
1430.Ar pflog0 .
1431.It Ar quick
1432If a packet matches a rule which has the
1433.Ar quick
1434option set, this rule
1435is considered the last matching rule, and evaluation of subsequent rules
1436is skipped.
1437.It Ar on Aq Ar interface
1438This rule applies only to packets coming in on, or going out through, this
1439particular interface or interface group.
1440For more information on interface groups,
1441see the
1442.Ic group
1443keyword in
1444.Xr ifconfig 8 .
1445.It Aq Ar af
1446This rule applies only to packets of this address family.
1447Supported values are
1448.Ar inet
1449and
1450.Ar inet6 .
1451.It Ar proto Aq Ar protocol
1452This rule applies only to packets of this protocol.
1453Common protocols are
1454.Xr icmp 4 ,
1455.Xr icmp6 4 ,
1456.Xr tcp 4 ,
1457and
1458.Xr udp 4 .
1459For a list of all the protocol name to number mappings used by
1460.Xr pfctl 8 ,
1461see the file
1462.Pa /etc/protocols .
1463.It Xo
1464.Ar from Aq Ar source
1465.Ar port Aq Ar source
1466.Ar os Aq Ar source
1467.Ar to Aq Ar dest
1468.Ar port Aq Ar dest
1469.Xc
1470This rule applies only to packets with the specified source and destination
1471addresses and ports.
1472.Pp
1473Addresses can be specified in CIDR notation (matching netblocks), as
1474symbolic host names, interface names or interface group names, or as any
1475of the following keywords:
1476.Pp
1477.Bl -tag -width xxxxxxxxxxxxxx -compact
1478.It Ar any
1479Any address.
1480.It Ar no-route
1481Any address which is not currently routable.
1482.It Ar urpf-failed
1483Any source address that fails a unicast reverse path forwarding (URPF)
1484check, i.e. packets coming in on an interface other than that which holds
1485the route back to the packet's source address.
1486.It Aq Ar table
1487Any address that matches the given table.
1488.El
1489.Pp
1490Ranges of addresses are specified by using the
1491.Sq -
1492operator.
1493For instance:
1494.Dq 10.1.1.10 - 10.1.1.12
1495means all addresses from 10.1.1.10 to 10.1.1.12,
1496hence addresses 10.1.1.10, 10.1.1.11, and 10.1.1.12.
1497.Pp
1498Interface names and interface group names can have modifiers appended:
1499.Pp
1500.Bl -tag -width xxxxxxxxxxxx -compact
1501.It Ar :network
1502Translates to the network(s) attached to the interface.
1503.It Ar :broadcast
1504Translates to the interface's broadcast address(es).
1505.It Ar :peer
1506Translates to the point-to-point interface's peer address(es).
1507.It Ar :0
1508Do not include interface aliases.
1509.El
1510.Pp
1511Host names may also have the
1512.Ar :0
1513option appended to restrict the name resolution to the first of each
1514v4 and non-link-local v6 address found.
1515.Pp
1516Host name resolution and interface to address translation are done at
1517ruleset load-time.
1518When the address of an interface (or host name) changes (under DHCP or PPP,
1519for instance), the ruleset must be reloaded for the change to be reflected
1520in the kernel.
1521Surrounding the interface name (and optional modifiers) in parentheses
1522changes this behaviour.
1523When the interface name is surrounded by parentheses, the rule is
1524automatically updated whenever the interface changes its address.
1525The ruleset does not need to be reloaded.
1526This is especially useful with
1527.Ar nat .
1528.Pp
1529Ports can be specified either by number or by name.
1530For example, port 80 can be specified as
1531.Em www .
1532For a list of all port name to number mappings used by
1533.Xr pfctl 8 ,
1534see the file
1535.Pa /etc/services .
1536.Pp
1537Ports and ranges of ports are specified by using these operators:
1538.Bd -literal -offset indent
1539=	(equal)
1540!=	(unequal)
1541\*(Lt	(less than)
1542\*(Le	(less than or equal)
1543\*(Gt	(greater than)
1544\*(Ge	(greater than or equal)
1545:	(range including boundaries)
1546\*(Gt\*(Lt	(range excluding boundaries)
1547\*(Lt\*(Gt	(except range)
1548.Ed
1549.Pp
1550.Sq \*(Gt\*(Lt ,
1551.Sq \*(Lt\*(Gt
1552and
1553.Sq \&:
1554are binary operators (they take two arguments).
1555For instance:
1556.Bl -tag -width Fl
1557.It Ar port 2000:2004
1558means
1559.Sq all ports \*(Ge 2000 and \*(Le 2004 ,
1560hence ports 2000, 2001, 2002, 2003 and 2004.
1561.It Ar port 2000 \*(Gt\*(Lt 2004
1562means
1563.Sq all ports \*(Gt 2000 and \*(Lt 2004 ,
1564hence ports 2001, 2002 and 2003.
1565.It Ar port 2000 \*(Lt\*(Gt 2004
1566means
1567.Sq all ports \*(Lt 2000 or \*(Gt 2004 ,
1568hence ports 1-1999 and 2005-65535.
1569.El
1570.Pp
1571The operating system of the source host can be specified in the case of TCP
1572rules with the
1573.Ar OS
1574modifier.
1575See the
1576.Sx OPERATING SYSTEM FINGERPRINTING
1577section for more information.
1578.Pp
1579The host, port and OS specifications are optional, as in the following examples:
1580.Bd -literal -offset indent
1581pass in all
1582pass in from any to any
1583pass in proto tcp from any port \*(Le 1024 to any
1584pass in proto tcp from any to any port 25
1585pass in proto tcp from 10.0.0.0/8 port \*(Gt 1024 \e
1586      to ! 10.1.2.3 port != ssh
1587pass in proto tcp from any os "OpenBSD"
1588.Ed
1589.It Ar all
1590This is equivalent to "from any to any".
1591.It Ar group Aq Ar group
1592Similar to
1593.Ar user ,
1594this rule only applies to packets of sockets owned by the specified group.
1595.It Ar user Aq Ar user
1596This rule only applies to packets of sockets owned by the specified user.
1597For outgoing connections initiated from the firewall, this is the user
1598that opened the connection.
1599For incoming connections to the firewall itself, this is the user that
1600listens on the destination port.
1601For forwarded connections, where the firewall is not a connection endpoint,
1602the user and group are
1603.Em unknown .
1604.Pp
1605All packets, both outgoing and incoming, of one connection are associated
1606with the same user and group.
1607Only TCP and UDP packets can be associated with users; for other protocols
1608these parameters are ignored.
1609.Pp
1610User and group refer to the effective (as opposed to the real) IDs, in
1611case the socket is created by a setuid/setgid process.
1612User and group IDs are stored when a socket is created;
1613when a process creates a listening socket as root (for instance, by
1614binding to a privileged port) and subsequently changes to another
1615user ID (to drop privileges), the credentials will remain root.
1616.Pp
1617User and group IDs can be specified as either numbers or names.
1618The syntax is similar to the one for ports.
1619The value
1620.Em unknown
1621matches packets of forwarded connections.
1622.Em unknown
1623can only be used with the operators
1624.Cm =
1625and
1626.Cm != .
1627Other constructs like
1628.Cm user \*(Ge unknown
1629are invalid.
1630Forwarded packets with unknown user and group ID match only rules
1631that explicitly compare against
1632.Em unknown
1633with the operators
1634.Cm =
1635or
1636.Cm != .
1637For instance
1638.Cm user \*(Ge 0
1639does not match forwarded packets.
1640The following example allows only selected users to open outgoing
1641connections:
1642.Bd -literal -offset indent
1643block out proto { tcp, udp } all
1644pass  out proto { tcp, udp } all user { \*(Lt 1000, dhartmei }
1645.Ed
1646.It Xo Ar flags Aq Ar a
1647.Pf / Ns Aq Ar b
1648.No \*(Ba / Ns Aq Ar b
1649.No \*(Ba any
1650.Xc
1651This rule only applies to TCP packets that have the flags
1652.Aq Ar a
1653set out of set
1654.Aq Ar b .
1655Flags not specified in
1656.Aq Ar b
1657are ignored.
1658For stateful connections, the default is
1659.Ar flags S/SA .
1660To indicate that flags should not be checked at all, specify
1661.Ar flags any .
1662The flags are: (F)IN, (S)YN, (R)ST, (P)USH, (A)CK, (U)RG, (E)CE, and C(W)R.
1663.Bl -tag -width Fl
1664.It Ar flags S/S
1665Flag SYN is set.
1666The other flags are ignored.
1667.It Ar flags S/SA
1668This is the default setting for stateful connections.
1669Out of SYN and ACK, exactly SYN may be set.
1670SYN, SYN+PSH and SYN+RST match, but SYN+ACK, ACK and ACK+RST do not.
1671This is more restrictive than the previous example.
1672.It Ar flags /SFRA
1673If the first set is not specified, it defaults to none.
1674All of SYN, FIN, RST and ACK must be unset.
1675.El
1676.Pp
1677Because
1678.Ar flags S/SA
1679is applied by default (unless
1680.Ar no state
1681is specified), only the initial SYN packet of a TCP handshake will create
1682a state for a TCP connection.
1683It is possible to be less restrictive, and allow state creation from
1684intermediate
1685.Pq non-SYN
1686packets, by specifying
1687.Ar flags any .
1688This will cause
1689.Xr pf 4
1690to synchronize to existing connections, for instance
1691if one flushes the state table.
1692However, states created from such intermediate packets may be missing
1693connection details such as the TCP window scaling factor.
1694States which modify the packet flow, such as those affected by
1695.Ar nat , binat No or Ar rdr
1696rules,
1697.Ar modulate No or Ar synproxy state
1698options, or scrubbed with
1699.Ar reassemble tcp
1700will also not be recoverable from intermediate packets.
1701Such connections will stall and time out.
1702.It Xo Ar icmp-type Aq Ar type
1703.Ar code Aq Ar code
1704.Xc
1705.It Xo Ar icmp6-type Aq Ar type
1706.Ar code Aq Ar code
1707.Xc
1708This rule only applies to ICMP or ICMPv6 packets with the specified type
1709and code.
1710Text names for ICMP types and codes are listed in
1711.Xr icmp 4
1712and
1713.Xr icmp6 4 .
1714This parameter is only valid for rules that cover protocols ICMP or
1715ICMP6.
1716The protocol and the ICMP type indicator
1717.Po
1718.Ar icmp-type
1719or
1720.Ar icmp6-type
1721.Pc
1722must match.
1723.It Xo Ar tos Aq Ar string
1724.No \*(Ba Aq Ar number
1725.Xc
1726This rule applies to packets with the specified
1727.Em TOS
1728bits set.
1729.Em TOS
1730may be
1731given as one of
1732.Ar critical ,
1733.Ar inetcontrol ,
1734.Ar lowdelay ,
1735.Ar netcontrol ,
1736.Ar throughput ,
1737.Ar reliability ,
1738or one of the DiffServ Code Points:
1739.Ar ef ,
1740.Ar af11 No ... Ar af43 ,
1741.Ar cs0 No ... Ar cs7 ;
1742or as either hex or decimal.
1743.Pp
1744For example, the following rules are identical:
1745.Bd -literal -offset indent
1746pass all tos lowdelay
1747pass all tos 0x10
1748pass all tos 16
1749.Ed
1750.It Ar allow-opts
1751By default, IPv4 packets with IP options or IPv6 packets with routing
1752extension headers are blocked.
1753When
1754.Ar allow-opts
1755is specified for a
1756.Ar pass
1757rule, packets that pass the filter based on that rule (last matching)
1758do so even if they contain IP options or routing extension headers.
1759For packets that match state, the rule that initially created the
1760state is used.
1761The implicit
1762.Ar pass
1763rule that is used when a packet does not match any rules does not
1764allow IP options.
1765.It Ar label Aq Ar string
1766Adds a label (name) to the rule, which can be used to identify the rule.
1767For instance,
1768pfctl -s labels
1769shows per-rule statistics for rules that have labels.
1770.Pp
1771The following macros can be used in labels:
1772.Pp
1773.Bl -tag -width $srcaddr -compact -offset indent
1774.It Ar $if
1775The interface.
1776.It Ar $srcaddr
1777The source IP address.
1778.It Ar $dstaddr
1779The destination IP address.
1780.It Ar $srcport
1781The source port specification.
1782.It Ar $dstport
1783The destination port specification.
1784.It Ar $proto
1785The protocol name.
1786.It Ar $nr
1787The rule number.
1788.El
1789.Pp
1790For example:
1791.Bd -literal -offset indent
1792ips = \&"{ 1.2.3.4, 1.2.3.5 }\&"
1793pass in proto tcp from any to $ips \e
1794      port \*(Gt 1023 label \&"$dstaddr:$dstport\&"
1795.Ed
1796.Pp
1797expands to
1798.Bd -literal -offset indent
1799pass in inet proto tcp from any to 1.2.3.4 \e
1800      port \*(Gt 1023 label \&"1.2.3.4:\*(Gt1023\&"
1801pass in inet proto tcp from any to 1.2.3.5 \e
1802      port \*(Gt 1023 label \&"1.2.3.5:\*(Gt1023\&"
1803.Ed
1804.Pp
1805The macro expansion for the
1806.Ar label
1807directive occurs only at configuration file parse time, not during runtime.
1808.It Xo Ar queue Aq Ar queue
1809.No \*(Ba ( Aq Ar queue ,
1810.Aq Ar queue )
1811.Xc
1812Packets matching this rule will be assigned to the specified queue.
1813If two queues are given, packets which have a
1814.Em TOS
1815of
1816.Em lowdelay
1817and TCP ACKs with no data payload will be assigned to the second one.
1818See
1819.Sx QUEUEING
1820for setup details.
1821.Pp
1822For example:
1823.Bd -literal -offset indent
1824pass in proto tcp to port 25 queue mail
1825pass in proto tcp to port 22 queue(ssh_bulk, ssh_prio)
1826.Ed
1827.Pp
1828.It Cm set prio Ar priority | Pq Ar priority , priority
1829Packets matching this rule will be assigned a specific queueing priority.
1830Priorities are assigned as integers 0 through 7.
1831If the packet is transmitted on a
1832.Xr vlan 4
1833interface, the queueing priority will be written as the priority
1834code point in the 802.1Q VLAN header.
1835If two priorities are given, packets which have a TOS of
1836.Cm lowdelay
1837and TCP ACKs with no data payload will be assigned to the second one.
1838.Pp
1839For example:
1840.Bd -literal -offset indent
1841pass in proto tcp to port 25 set prio 2
1842pass in proto tcp to port 22 set prio (2, 5)
1843.Ed
1844.Pp
1845.It Ar tag Aq Ar string
1846Packets matching this rule will be tagged with the
1847specified string.
1848The tag acts as an internal marker that can be used to
1849identify these packets later on.
1850This can be used, for example, to provide trust between
1851interfaces and to determine if packets have been
1852processed by translation rules.
1853Tags are
1854.Qq sticky ,
1855meaning that the packet will be tagged even if the rule
1856is not the last matching rule.
1857Further matching rules can replace the tag with a
1858new one but will not remove a previously applied tag.
1859A packet is only ever assigned one tag at a time.
1860Packet tagging can be done during
1861.Ar nat ,
1862.Ar rdr ,
1863or
1864.Ar binat
1865rules in addition to filter rules.
1866Tags take the same macros as labels (see above).
1867.It Ar tagged Aq Ar string
1868Used with filter, translation or scrub rules
1869to specify that packets must already
1870be tagged with the given tag in order to match the rule.
1871Inverse tag matching can also be done
1872by specifying the
1873.Cm !\&
1874operator before the
1875.Ar tagged
1876keyword.
1877.It Ar rtable Aq Ar number
1878Used to select an alternate routing table for the routing lookup.
1879Only effective before the route lookup happened, i.e. when filtering inbound.
1880.It Xo Ar divert-to Aq Ar host
1881.Ar port Aq Ar port
1882.Xc
1883Used to redirect packets to a local socket bound to
1884.Ar host
1885and
1886.Ar port .
1887The packets will not be modified, so
1888.Xr getsockname 2
1889on the socket will return the original destination address of the packet.
1890.It Ar divert-reply
1891Used to receive replies for sockets that are bound to addresses
1892which are not local to the machine.
1893See
1894.Xr setsockopt 2
1895for information on how to bind these sockets.
1896.It Ar probability Aq Ar number
1897A probability attribute can be attached to a rule, with a value set between
18980 and 1, bounds not included.
1899In that case, the rule will be honoured using the given probability value
1900only.
1901For example, the following rule will drop 20% of incoming ICMP packets:
1902.Bd -literal -offset indent
1903block in proto icmp probability 20%
1904.Ed
1905.It Ar prio Aq Ar number
1906Only match packets which have the given queueing priority assigned.
1907.Pp
1908.El
1909.Sh ROUTING
1910If a packet matches a rule with a route option set, the packet filter will
1911route the packet according to the type of route option.
1912When such a rule creates state, the route option is also applied to all
1913packets matching the same connection.
1914.Bl -tag -width xxxx
1915.It Ar route-to
1916The
1917.Ar route-to
1918option routes the packet to the specified interface with an optional address
1919for the next hop.
1920When a
1921.Ar route-to
1922rule creates state, only packets that pass in the same direction as the
1923filter rule specifies will be routed in this way.
1924Packets passing in the opposite direction (replies) are not affected
1925and are routed normally.
1926.It Ar reply-to
1927The
1928.Ar reply-to
1929option is similar to
1930.Ar route-to ,
1931but routes packets that pass in the opposite direction (replies) to the
1932specified interface.
1933Opposite direction is only defined in the context of a state entry, and
1934.Ar reply-to
1935is useful only in rules that create state.
1936It can be used on systems with multiple external connections to
1937route all outgoing packets of a connection through the interface
1938the incoming connection arrived through (symmetric routing enforcement).
1939.It Ar dup-to
1940The
1941.Ar dup-to
1942option creates a duplicate of the packet and routes it like
1943.Ar route-to .
1944The original packet gets routed as it normally would.
1945.El
1946.Sh POOL OPTIONS
1947For
1948.Ar nat
1949and
1950.Ar rdr
1951rules, (as well as for the
1952.Ar route-to ,
1953.Ar reply-to
1954and
1955.Ar dup-to
1956rule options) for which there is a single redirection address which has a
1957subnet mask smaller than 32 for IPv4 or 128 for IPv6 (more than one IP
1958address), a variety of different methods for assigning this address can be
1959used:
1960.Bl -tag -width xxxx
1961.It Ar bitmask
1962The
1963.Ar bitmask
1964option applies the network portion of the redirection address to the address
1965to be modified (source with
1966.Ar nat ,
1967destination with
1968.Ar rdr ) .
1969.It Ar random
1970The
1971.Ar random
1972option selects an address at random within the defined block of addresses.
1973.It Ar source-hash
1974The
1975.Ar source-hash
1976option uses a hash of the source address to determine the redirection address,
1977ensuring that the redirection address is always the same for a given source.
1978An optional key can be specified after this keyword either in hex or as a
1979string; by default
1980.Xr pfctl 8
1981randomly generates a key for source-hash every time the
1982ruleset is reloaded.
1983.It Ar round-robin
1984The
1985.Ar round-robin
1986option loops through the redirection address(es).
1987.Pp
1988When more than one redirection address is specified,
1989.Ar round-robin
1990is the only permitted pool type.
1991.It Ar static-port
1992With
1993.Ar nat
1994rules, the
1995.Ar static-port
1996option prevents
1997.Xr pf 4
1998from modifying the source port on TCP and UDP packets.
1999.El
2000.Pp
2001Additionally, the
2002.Ar sticky-address
2003option can be specified to help ensure that multiple connections from the
2004same source are mapped to the same redirection address.
2005This option can be used with the
2006.Ar random
2007and
2008.Ar round-robin
2009pool options.
2010Note that by default these associations are destroyed as soon as there are
2011no longer states which refer to them; in order to make the mappings last
2012beyond the lifetime of the states, increase the global options with
2013.Ar set timeout src.track .
2014See
2015.Sx STATEFUL TRACKING OPTIONS
2016for more ways to control the source tracking.
2017.Sh STATE MODULATION
2018Much of the security derived from TCP is attributable to how well the
2019initial sequence numbers (ISNs) are chosen.
2020Some popular stack implementations choose
2021.Em very
2022poor ISNs and thus are normally susceptible to ISN prediction exploits.
2023By applying a
2024.Ar modulate state
2025rule to a TCP connection,
2026.Xr pf 4
2027will create a high quality random sequence number for each connection
2028endpoint.
2029.Pp
2030The
2031.Ar modulate state
2032directive implicitly keeps state on the rule and is
2033only applicable to TCP connections.
2034.Pp
2035For instance:
2036.Bd -literal -offset indent
2037block all
2038pass out proto tcp from any to any modulate state
2039pass in  proto tcp from any to any port 25 flags S/SFRA modulate state
2040.Ed
2041.Pp
2042Note that modulated connections will not recover when the state table
2043is lost (firewall reboot, flushing the state table, etc...).
2044.Xr pf 4
2045will not be able to infer a connection again after the state table flushes
2046the connection's modulator.
2047When the state is lost, the connection may be left dangling until the
2048respective endpoints time out the connection.
2049It is possible on a fast local network for the endpoints to start an ACK
2050storm while trying to resynchronize after the loss of the modulator.
2051The default
2052.Ar flags
2053settings (or a more strict equivalent) should be used on
2054.Ar modulate state
2055rules to prevent ACK storms.
2056.Pp
2057Note that alternative methods are available
2058to prevent loss of the state table
2059and allow for firewall failover.
2060See
2061.Xr carp 4
2062and
2063.Xr pfsync 4
2064for further information.
2065.Sh SYN PROXY
2066By default,
2067.Xr pf 4
2068passes packets that are part of a
2069.Xr tcp 4
2070handshake between the endpoints.
2071The
2072.Ar synproxy state
2073option can be used to cause
2074.Xr pf 4
2075itself to complete the handshake with the active endpoint, perform a handshake
2076with the passive endpoint, and then forward packets between the endpoints.
2077.Pp
2078No packets are sent to the passive endpoint before the active endpoint has
2079completed the handshake, hence so-called SYN floods with spoofed source
2080addresses will not reach the passive endpoint, as the sender can't complete the
2081handshake.
2082.Pp
2083The proxy is transparent to both endpoints, they each see a single
2084connection from/to the other endpoint.
2085.Xr pf 4
2086chooses random initial sequence numbers for both handshakes.
2087Once the handshakes are completed, the sequence number modulators
2088(see previous section) are used to translate further packets of the
2089connection.
2090.Ar synproxy state
2091includes
2092.Ar modulate state .
2093.Pp
2094Rules with
2095.Ar synproxy
2096will not work if
2097.Xr pf 4
2098operates on a
2099.Xr bridge 4 .
2100.Pp
2101Example:
2102.Bd -literal -offset indent
2103pass in proto tcp from any to any port www synproxy state
2104.Ed
2105.Sh STATEFUL TRACKING OPTIONS
2106A number of options related to stateful tracking can be applied on a
2107per-rule basis.
2108.Ar keep state ,
2109.Ar modulate state
2110and
2111.Ar synproxy state
2112support these options, and
2113.Ar keep state
2114must be specified explicitly to apply options to a rule.
2115.Pp
2116.Bl -tag -width xxxx -compact
2117.It Ar max Aq Ar number
2118Limits the number of concurrent states the rule may create.
2119When this limit is reached, further packets that would create
2120state will not match this rule until existing states time out.
2121.It Ar no-sync
2122Prevent state changes for states created by this rule from appearing on the
2123.Xr pfsync 4
2124interface.
2125.It Xo Aq Ar timeout
2126.Aq Ar seconds
2127.Xc
2128Changes the timeout values used for states created by this rule.
2129For a list of all valid timeout names, see
2130.Sx OPTIONS
2131above.
2132.It Ar sloppy
2133Uses a sloppy TCP connection tracker that does not check sequence
2134numbers at all, which makes insertion and ICMP teardown attacks way
2135easier.
2136This is intended to be used in situations where one does not see all
2137packets of a connection, e.g. in asymmetric routing situations.
2138Cannot be used with modulate or synproxy state.
2139.El
2140.Pp
2141Multiple options can be specified, separated by commas:
2142.Bd -literal -offset indent
2143pass in proto tcp from any to any \e
2144      port www keep state \e
2145      (max 100, source-track rule, max-src-nodes 75, \e
2146      max-src-states 3, tcp.established 60, tcp.closing 5)
2147.Ed
2148.Pp
2149When the
2150.Ar source-track
2151keyword is specified, the number of states per source IP is tracked.
2152.Pp
2153.Bl -tag -width xxxx -compact
2154.It Ar source-track rule
2155The maximum number of states created by this rule is limited by the rule's
2156.Ar max-src-nodes
2157and
2158.Ar max-src-states
2159options.
2160Only state entries created by this particular rule count toward the rule's
2161limits.
2162.It Ar source-track global
2163The number of states created by all rules that use this option is limited.
2164Each rule can specify different
2165.Ar max-src-nodes
2166and
2167.Ar max-src-states
2168options, however state entries created by any participating rule count towards
2169each individual rule's limits.
2170.El
2171.Pp
2172The following limits can be set:
2173.Pp
2174.Bl -tag -width xxxx -compact
2175.It Ar max-src-nodes Aq Ar number
2176Limits the maximum number of source addresses which can simultaneously
2177have state table entries.
2178.It Ar max-src-states Aq Ar number
2179Limits the maximum number of simultaneous state entries that a single
2180source address can create with this rule.
2181.El
2182.Pp
2183For stateful TCP connections, limits on established connections (connections
2184which have completed the TCP 3-way handshake) can also be enforced
2185per source IP.
2186.Pp
2187.Bl -tag -width xxxx -compact
2188.It Ar max-src-conn Aq Ar number
2189Limits the maximum number of simultaneous TCP connections which have
2190completed the 3-way handshake that a single host can make.
2191.It Xo Ar max-src-conn-rate Aq Ar number
2192.No / Aq Ar seconds
2193.Xc
2194Limit the rate of new connections over a time interval.
2195The connection rate is an approximation calculated as a moving average.
2196.El
2197.Pp
2198Because the 3-way handshake ensures that the source address is not being
2199spoofed, more aggressive action can be taken based on these limits.
2200With the
2201.Ar overload Aq Ar table
2202state option, source IP addresses which hit either of the limits on
2203established connections will be added to the named table.
2204This table can be used in the ruleset to block further activity from
2205the offending host, redirect it to a tarpit process, or restrict its
2206bandwidth.
2207.Pp
2208The optional
2209.Ar flush
2210keyword kills all states created by the matching rule which originate
2211from the host which exceeds these limits.
2212The
2213.Ar global
2214modifier to the flush command kills all states originating from the
2215offending host, regardless of which rule created the state.
2216.Pp
2217For example, the following rules will protect the webserver against
2218hosts making more than 100 connections in 10 seconds.
2219Any host which connects faster than this rate will have its address added
2220to the
2221.Aq bad_hosts
2222table and have all states originating from it flushed.
2223Any new packets arriving from this host will be dropped unconditionally
2224by the block rule.
2225.Bd -literal -offset indent
2226block quick from \*(Ltbad_hosts\*(Gt
2227pass in on $ext_if proto tcp to $webserver port www keep state \e
2228	(max-src-conn-rate 100/10, overload \*(Ltbad_hosts\*(Gt flush global)
2229.Ed
2230.Sh OPERATING SYSTEM FINGERPRINTING
2231Passive OS Fingerprinting is a mechanism to inspect nuances of a TCP
2232connection's initial SYN packet and guess at the host's operating system.
2233Unfortunately these nuances are easily spoofed by an attacker so the
2234fingerprint is not useful in making security decisions.
2235But the fingerprint is typically accurate enough to make policy decisions
2236upon.
2237.Pp
2238The fingerprints may be specified by operating system class, by
2239version, or by subtype/patchlevel.
2240The class of an operating system is typically the vendor or genre
2241and would be
2242.Ox
2243for the
2244.Xr pf 4
2245firewall itself.
2246The version of the oldest available
2247.Ox
2248release on the main FTP site
2249would be 2.6 and the fingerprint would be written
2250.Pp
2251.Dl \&"OpenBSD 2.6\&"
2252.Pp
2253The subtype of an operating system is typically used to describe the
2254patchlevel if that patch led to changes in the TCP stack behavior.
2255In the case of
2256.Ox ,
2257the only subtype is for a fingerprint that was
2258normalized by the
2259.Ar no-df
2260scrub option and would be specified as
2261.Pp
2262.Dl \&"OpenBSD 3.3 no-df\&"
2263.Pp
2264Fingerprints for most popular operating systems are provided by
2265.Xr pf.os 5 .
2266Once
2267.Xr pf 4
2268is running, a complete list of known operating system fingerprints may
2269be listed by running:
2270.Pp
2271.Dl # pfctl -so
2272.Pp
2273Filter rules can enforce policy at any level of operating system specification
2274assuming a fingerprint is present.
2275Policy could limit traffic to approved operating systems or even ban traffic
2276from hosts that aren't at the latest service pack.
2277.Pp
2278The
2279.Ar unknown
2280class can also be used as the fingerprint which will match packets for
2281which no operating system fingerprint is known.
2282.Pp
2283Examples:
2284.Bd -literal -offset indent
2285pass  out proto tcp from any os OpenBSD
2286block out proto tcp from any os Doors
2287block out proto tcp from any os "Doors PT"
2288block out proto tcp from any os "Doors PT SP3"
2289block out from any os "unknown"
2290pass on lo0 proto tcp from any os "OpenBSD 3.3 lo0"
2291.Ed
2292.Pp
2293Operating system fingerprinting is limited only to the TCP SYN packet.
2294This means that it will not work on other protocols and will not match
2295a currently established connection.
2296.Pp
2297Caveat: operating system fingerprints are occasionally wrong.
2298There are three problems: an attacker can trivially craft his packets to
2299appear as any operating system he chooses;
2300an operating system patch could change the stack behavior and no fingerprints
2301will match it until the database is updated;
2302and multiple operating systems may have the same fingerprint.
2303.Sh BLOCKING SPOOFED TRAFFIC
2304"Spoofing" is the faking of IP addresses, typically for malicious
2305purposes.
2306The
2307.Ar antispoof
2308directive expands to a set of filter rules which will block all
2309traffic with a source IP from the network(s) directly connected
2310to the specified interface(s) from entering the system through
2311any other interface.
2312.Pp
2313For example, the line
2314.Bd -literal -offset indent
2315antispoof for lo0
2316.Ed
2317.Pp
2318expands to
2319.Bd -literal -offset indent
2320block drop in on ! lo0 inet from 127.0.0.1/8 to any
2321block drop in on ! lo0 inet6 from ::1 to any
2322.Ed
2323.Pp
2324For non-loopback interfaces, there are additional rules to block incoming
2325packets with a source IP address identical to the interface's IP(s).
2326For example, assuming the interface wi0 had an IP address of 10.0.0.1 and a
2327netmask of 255.255.255.0,
2328the line
2329.Bd -literal -offset indent
2330antispoof for wi0 inet
2331.Ed
2332.Pp
2333expands to
2334.Bd -literal -offset indent
2335block drop in on ! wi0 inet from 10.0.0.0/24 to any
2336block drop in inet from 10.0.0.1 to any
2337.Ed
2338.Pp
2339Caveat: Rules created by the
2340.Ar antispoof
2341directive interfere with packets sent over loopback interfaces
2342to local addresses.
2343One should pass these explicitly.
2344.Sh FRAGMENT HANDLING
2345The size of IP datagrams (packets) can be significantly larger than the
2346maximum transmission unit (MTU) of the network.
2347In cases when it is necessary or more efficient to send such large packets,
2348the large packet will be fragmented into many smaller packets that will each
2349fit onto the wire.
2350Unfortunately for a firewalling device, only the first logical fragment will
2351contain the necessary header information for the subprotocol that allows
2352.Xr pf 4
2353to filter on things such as TCP ports or to perform NAT.
2354.Pp
2355Besides the use of
2356.Ar scrub
2357rules as described in
2358.Sx TRAFFIC NORMALIZATION
2359above, there are three options for handling fragments in the packet filter.
2360.Pp
2361One alternative is to filter individual fragments with filter rules.
2362If no
2363.Ar scrub
2364rule applies to a fragment, it is passed to the filter.
2365Filter rules with matching IP header parameters decide whether the
2366fragment is passed or blocked, in the same way as complete packets
2367are filtered.
2368Without reassembly, fragments can only be filtered based on IP header
2369fields (source/destination address, protocol), since subprotocol header
2370fields are not available (TCP/UDP port numbers, ICMP code/type).
2371The
2372.Ar fragment
2373option can be used to restrict filter rules to apply only to
2374fragments, but not complete packets.
2375Filter rules without the
2376.Ar fragment
2377option still apply to fragments, if they only specify IP header fields.
2378For instance, the rule
2379.Bd -literal -offset indent
2380pass in proto tcp from any to any port 80
2381.Ed
2382.Pp
2383never applies to a fragment, even if the fragment is part of a TCP
2384packet with destination port 80, because without reassembly this information
2385is not available for each fragment.
2386This also means that fragments cannot create new or match existing
2387state table entries, which makes stateful filtering and address
2388translation (NAT, redirection) for fragments impossible.
2389.Pp
2390It's also possible to reassemble only certain fragments by specifying
2391source or destination addresses or protocols as parameters in
2392.Ar scrub
2393rules.
2394.Pp
2395In most cases, the benefits of reassembly outweigh the additional
2396memory cost, and it's recommended to use
2397.Ar scrub
2398rules to reassemble
2399all fragments via the
2400.Ar fragment reassemble
2401modifier.
2402.Pp
2403The memory allocated for fragment caching can be limited using
2404.Xr pfctl 8 .
2405Once this limit is reached, fragments that would have to be cached
2406are dropped until other entries time out.
2407The timeout value can also be adjusted.
2408.Pp
2409When forwarding reassembled IPv6 packets, pf refragments them with
2410the original maximum fragment size.
2411This allows the sender to determine the optimal fragment size by
2412path MTU discovery.
2413.Sh ANCHORS
2414Besides the main ruleset,
2415.Xr pfctl 8
2416can load rulesets into
2417.Ar anchor
2418attachment points.
2419An
2420.Ar anchor
2421is a container that can hold rules, address tables, and other anchors.
2422.Pp
2423An
2424.Ar anchor
2425has a name which specifies the path where
2426.Xr pfctl 8
2427can be used to access the anchor to perform operations on it, such as
2428attaching child anchors to it or loading rules into it.
2429Anchors may be nested, with components separated by
2430.Sq /
2431characters, similar to how file system hierarchies are laid out.
2432The main ruleset is actually the default anchor, so filter and
2433translation rules, for example, may also be contained in any anchor.
2434.Pp
2435An anchor can reference another
2436.Ar anchor
2437attachment point
2438using the following kinds
2439of rules:
2440.Bl -tag -width xxxx
2441.It Ar nat-anchor Aq Ar name
2442Evaluates the
2443.Ar nat
2444rules in the specified
2445.Ar anchor .
2446.It Ar rdr-anchor Aq Ar name
2447Evaluates the
2448.Ar rdr
2449rules in the specified
2450.Ar anchor .
2451.It Ar binat-anchor Aq Ar name
2452Evaluates the
2453.Ar binat
2454rules in the specified
2455.Ar anchor .
2456.It Ar anchor Aq Ar name
2457Evaluates the filter rules in the specified
2458.Ar anchor .
2459.It Xo Ar load anchor
2460.Aq Ar name
2461.Ar from Aq Ar file
2462.Xc
2463Loads the rules from the specified file into the
2464anchor
2465.Ar name .
2466.El
2467.Pp
2468When evaluation of the main ruleset reaches an
2469.Ar anchor
2470rule,
2471.Xr pf 4
2472will proceed to evaluate all rules specified in that anchor.
2473.Pp
2474Matching filter and translation rules marked with the
2475.Ar quick
2476option are final and abort the evaluation of the rules in other
2477anchors and the main ruleset.
2478If the
2479.Ar anchor
2480itself is marked with the
2481.Ar quick
2482option,
2483ruleset evaluation will terminate when the anchor is exited if the packet is
2484matched by any rule within the anchor.
2485.Pp
2486.Ar anchor
2487rules are evaluated relative to the anchor in which they are contained.
2488For example, all
2489.Ar anchor
2490rules specified in the main ruleset will reference anchor
2491attachment points underneath the main ruleset, and
2492.Ar anchor
2493rules specified in a file loaded from a
2494.Ar load anchor
2495rule will be attached under that anchor point.
2496.Pp
2497Rules may be contained in
2498.Ar anchor
2499attachment points which do not contain any rules when the main ruleset
2500is loaded, and later such anchors can be manipulated through
2501.Xr pfctl 8
2502without reloading the main ruleset or other anchors.
2503For example,
2504.Bd -literal -offset indent
2505ext_if = \&"kue0\&"
2506block on $ext_if all
2507anchor spam
2508pass out on $ext_if all
2509pass in on $ext_if proto tcp from any \e
2510      to $ext_if port smtp
2511.Ed
2512.Pp
2513blocks all packets on the external interface by default, then evaluates
2514all rules in the
2515.Ar anchor
2516named "spam", and finally passes all outgoing connections and
2517incoming connections to port 25.
2518.Bd -literal -offset indent
2519# echo \&"block in quick from 1.2.3.4 to any\&" \&| \e
2520      pfctl -a spam -f -
2521.Ed
2522.Pp
2523This loads a single rule into the
2524.Ar anchor ,
2525which blocks all packets from a specific address.
2526.Pp
2527The anchor can also be populated by adding a
2528.Ar load anchor
2529rule after the
2530.Ar anchor
2531rule:
2532.Bd -literal -offset indent
2533anchor spam
2534load anchor spam from "/etc/pf-spam.conf"
2535.Ed
2536.Pp
2537When
2538.Xr pfctl 8
2539loads
2540.Nm pf.conf ,
2541it will also load all the rules from the file
2542.Pa /etc/pf-spam.conf
2543into the anchor.
2544.Pp
2545Optionally,
2546.Ar anchor
2547rules can specify packet filtering parameters using the same syntax as
2548filter rules.
2549When parameters are used, the
2550.Ar anchor
2551rule is only evaluated for matching packets.
2552This allows conditional evaluation of anchors, like:
2553.Bd -literal -offset indent
2554block on $ext_if all
2555anchor spam proto tcp from any to any port smtp
2556pass out on $ext_if all
2557pass in on $ext_if proto tcp from any to $ext_if port smtp
2558.Ed
2559.Pp
2560The rules inside
2561.Ar anchor
2562spam are only evaluated for
2563.Ar tcp
2564packets with destination port 25.
2565Hence,
2566.Bd -literal -offset indent
2567# echo \&"block in quick from 1.2.3.4 to any" \&| \e
2568      pfctl -a spam -f -
2569.Ed
2570.Pp
2571will only block connections from 1.2.3.4 to port 25.
2572.Pp
2573Anchors may end with the asterisk
2574.Pq Sq *
2575character, which signifies that all anchors attached at that point
2576should be evaluated in the alphabetical ordering of their anchor name.
2577For example,
2578.Bd -literal -offset indent
2579anchor "spam/*"
2580.Ed
2581.Pp
2582will evaluate each rule in each anchor attached to the
2583.Li spam
2584anchor.
2585Note that it will only evaluate anchors that are directly attached to the
2586.Li spam
2587anchor, and will not descend to evaluate anchors recursively.
2588.Pp
2589Since anchors are evaluated relative to the anchor in which they are
2590contained, there is a mechanism for accessing the parent and ancestor
2591anchors of a given anchor.
2592Similar to file system path name resolution, if the sequence
2593.Dq ..
2594appears as an anchor path component, the parent anchor of the current
2595anchor in the path evaluation at that point will become the new current
2596anchor.
2597As an example, consider the following:
2598.Bd -literal -offset indent
2599# echo ' anchor "spam/allowed" ' | pfctl -f -
2600# echo -e ' anchor "../banned" \en pass' | \e
2601      pfctl -a spam/allowed -f -
2602.Ed
2603.Pp
2604Evaluation of the main ruleset will lead into the
2605.Li spam/allowed
2606anchor, which will evaluate the rules in the
2607.Li spam/banned
2608anchor, if any, before finally evaluating the
2609.Ar pass
2610rule.
2611.Pp
2612Filter rule
2613.Ar anchors
2614can also be loaded inline in the ruleset within a brace ('{' '}') delimited
2615block.
2616Brace delimited blocks may contain rules or other brace-delimited blocks.
2617When anchors are loaded this way the anchor name becomes optional.
2618.Bd -literal -offset indent
2619anchor "external" on egress {
2620	block
2621	anchor out {
2622		pass proto tcp from any to port { 25, 80, 443 }
2623	}
2624	pass in proto tcp to any port 22
2625}
2626.Ed
2627.Pp
2628Since the parser specification for anchor names is a string, any
2629reference to an anchor name containing
2630.Sq /
2631characters will require double quote
2632.Pq Sq \&"
2633characters around the anchor name.
2634.Sh TRANSLATION EXAMPLES
2635This example maps incoming requests on port 80 to port 8080, on
2636which a daemon is running (because, for example, it is not run as root,
2637and therefore lacks permission to bind to port 80).
2638.Bd -literal
2639# use a macro for the interface name, so it can be changed easily
2640ext_if = \&"ne3\&"
2641
2642# map daemon on 8080 to appear to be on 80
2643rdr on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 port 8080
2644.Ed
2645.Pp
2646If the
2647.Ar pass
2648modifier is given, packets matching the translation rule are passed without
2649inspecting the filter rules:
2650.Bd -literal
2651rdr pass on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 \e
2652      port 8080
2653.Ed
2654.Pp
2655In the example below, vlan12 is configured as 192.168.168.1;
2656the machine translates all packets coming from 192.168.168.0/24 to 204.92.77.111
2657when they are going out any interface except vlan12.
2658This has the net effect of making traffic from the 192.168.168.0/24
2659network appear as though it is the Internet routable address
2660204.92.77.111 to nodes behind any interface on the router except
2661for the nodes on vlan12.
2662(Thus, 192.168.168.1 can talk to the 192.168.168.0/24 nodes.)
2663.Bd -literal
2664nat on ! vlan12 from 192.168.168.0/24 to any -\*(Gt 204.92.77.111
2665.Ed
2666.Pp
2667In the example below, the machine sits between a fake internal 144.19.74.*
2668network, and a routable external IP of 204.92.77.100.
2669The
2670.Ar no nat
2671rule excludes protocol AH from being translated.
2672.Bd -literal
2673# NO NAT
2674no nat on $ext_if proto ah from 144.19.74.0/24 to any
2675nat on $ext_if from 144.19.74.0/24 to any -\*(Gt 204.92.77.100
2676.Ed
2677.Pp
2678In the example below, packets bound for one specific server, as well as those
2679generated by the sysadmins are not proxied; all other connections are.
2680.Bd -literal
2681# NO RDR
2682no rdr on $int_if proto { tcp, udp } from any to $server port 80
2683no rdr on $int_if proto { tcp, udp } from $sysadmins to any port 80
2684rdr on $int_if proto { tcp, udp } from any to any port 80 -\*(Gt 127.0.0.1 \e
2685      port 80
2686.Ed
2687.Pp
2688This longer example uses both a NAT and a redirection.
2689The external interface has the address 157.161.48.183.
2690On localhost, we are running
2691.Xr ftp-proxy 8 ,
2692waiting for FTP sessions to be redirected to it.
2693The three mandatory anchors for
2694.Xr ftp-proxy 8
2695are omitted from this example; see the
2696.Xr ftp-proxy 8
2697manpage.
2698.Bd -literal
2699# NAT
2700# Translate outgoing packets' source addresses (any protocol).
2701# In this case, any address but the gateway's external address is mapped.
2702nat on $ext_if inet from ! ($ext_if) to any -\*(Gt ($ext_if)
2703
2704# NAT PROXYING
2705# Map outgoing packets' source port to an assigned proxy port instead of
2706# an arbitrary port.
2707# In this case, proxy outgoing isakmp with port 500 on the gateway.
2708nat on $ext_if inet proto udp from any port = isakmp to any -\*(Gt ($ext_if) \e
2709      port 500
2710
2711# BINAT
2712# Translate outgoing packets' source address (any protocol).
2713# Translate incoming packets' destination address to an internal machine
2714# (bidirectional).
2715binat on $ext_if from 10.1.2.150 to any -\*(Gt $ext_if
2716
2717# RDR
2718# Translate incoming packets' destination addresses.
2719# As an example, redirect a TCP and UDP port to an internal machine.
2720rdr on $ext_if inet proto tcp from any to ($ext_if) port 8080 \e
2721      -\*(Gt 10.1.2.151 port 22
2722rdr on $ext_if inet proto udp from any to ($ext_if) port 8080 \e
2723      -\*(Gt 10.1.2.151 port 53
2724
2725# RDR
2726# Translate outgoing ftp control connections to send them to localhost
2727# for proxying with ftp-proxy(8) running on port 8021.
2728rdr on $int_if proto tcp from any to any port 21 -\*(Gt 127.0.0.1 port 8021
2729.Ed
2730.Pp
2731In this example, a NAT gateway is set up to translate internal addresses
2732using a pool of public addresses (192.0.2.16/28) and to redirect
2733incoming web server connections to a group of web servers on the internal
2734network.
2735.Bd -literal
2736# NAT LOAD BALANCE
2737# Translate outgoing packets' source addresses using an address pool.
2738# A given source address is always translated to the same pool address by
2739# using the source-hash keyword.
2740nat on $ext_if inet from any to any -\*(Gt 192.0.2.16/28 source-hash
2741
2742# RDR ROUND ROBIN
2743# Translate incoming web server connections to a group of web servers on
2744# the internal network.
2745rdr on $ext_if proto tcp from any to any port 80 \e
2746      -\*(Gt { 10.1.2.155, 10.1.2.160, 10.1.2.161 } round-robin
2747.Ed
2748.Sh FILTER EXAMPLES
2749.Bd -literal
2750# The external interface is kue0
2751# (157.161.48.183, the only routable address)
2752# and the private network is 10.0.0.0/8, for which we are doing NAT.
2753
2754# use a macro for the interface name, so it can be changed easily
2755ext_if = \&"kue0\&"
2756
2757# normalize all incoming traffic
2758scrub in on $ext_if all fragment reassemble
2759
2760# block and log everything by default
2761block return log on $ext_if all
2762
2763# block anything coming from source we have no back routes for
2764block in from no-route to any
2765
2766# block packets whose ingress interface does not match the one in
2767# the route back to their source address
2768block in from urpf-failed to any
2769
2770# block and log outgoing packets that do not have our address as source,
2771# they are either spoofed or something is misconfigured (NAT disabled,
2772# for instance), we want to be nice and do not send out garbage.
2773block out log quick on $ext_if from ! 157.161.48.183 to any
2774
2775# silently drop broadcasts (cable modem noise)
2776block in quick on $ext_if from any to 255.255.255.255
2777
2778# block and log incoming packets from reserved address space and invalid
2779# addresses, they are either spoofed or misconfigured, we cannot reply to
2780# them anyway (hence, no return-rst).
2781block in log quick on $ext_if from { 10.0.0.0/8, 172.16.0.0/12, \e
2782      192.168.0.0/16, 255.255.255.255/32 } to any
2783
2784# ICMP
2785
2786# pass out/in certain ICMP queries and keep state (ping)
2787# state matching is done on host addresses and ICMP id (not type/code),
2788# so replies (like 0/0 for 8/0) will match queries
2789# ICMP error messages (which always refer to a TCP/UDP packet) are
2790# handled by the TCP/UDP states
2791pass on $ext_if inet proto icmp all icmp-type 8 code 0
2792
2793# UDP
2794
2795# pass out all UDP connections and keep state
2796pass out on $ext_if proto udp all
2797
2798# pass in certain UDP connections and keep state (DNS)
2799pass in on $ext_if proto udp from any to any port domain
2800
2801# TCP
2802
2803# pass out all TCP connections and modulate state
2804pass out on $ext_if proto tcp all modulate state
2805
2806# pass in certain TCP connections and keep state (SSH, SMTP, DNS, IDENT)
2807pass in on $ext_if proto tcp from any to any port { ssh, smtp, domain, \e
2808      auth }
2809
2810# Do not allow Windows 9x SMTP connections since they are typically
2811# a viral worm. Alternately we could limit these OSes to 1 connection each.
2812block in on $ext_if proto tcp from any os {"Windows 95", "Windows 98"} \e
2813      to any port smtp
2814
2815# IPv6
2816# pass in/out all IPv6 traffic: note that we have to enable this in two
2817# different ways, on both our physical interface and our tunnel
2818pass quick on gif0 inet6
2819pass quick on $ext_if proto ipv6
2820
2821# Packet Tagging
2822
2823# three interfaces: $int_if, $ext_if, and $wifi_if (wireless). NAT is
2824# being done on $ext_if for all outgoing packets. tag packets in on
2825# $int_if and pass those tagged packets out on $ext_if.  all other
2826# outgoing packets (i.e., packets from the wireless network) are only
2827# permitted to access port 80.
2828
2829pass in on $int_if from any to any tag INTNET
2830pass in on $wifi_if from any to any
2831
2832block out on $ext_if from any to any
2833pass out quick on $ext_if tagged INTNET
2834pass out on $ext_if proto tcp from any to any port 80
2835
2836# tag incoming packets as they are redirected to spamd(8). use the tag
2837# to pass those packets through the packet filter.
2838
2839rdr on $ext_if inet proto tcp from \*(Ltspammers\*(Gt to port smtp \e
2840	tag SPAMD -\*(Gt 127.0.0.1 port spamd
2841
2842block in on $ext_if
2843pass in on $ext_if inet proto tcp tagged SPAMD
2844.Ed
2845.Sh GRAMMAR
2846Syntax for
2847.Nm
2848in BNF:
2849.Bd -literal
2850line           = ( option | pf-rule | nat-rule | binat-rule | rdr-rule |
2851                 antispoof-rule | altq-rule | queue-rule | trans-anchors |
2852                 anchor-rule | anchor-close | load-anchor | table-rule |
2853                 include )
2854
2855option         = "set" ( [ "timeout" ( timeout | "{" timeout-list "}" ) ] |
2856                 [ "ruleset-optimization" [ "none" | "basic" | "profile" ]] |
2857                 [ "optimization" [ "default" | "normal" |
2858                 "high-latency" | "satellite" |
2859                 "aggressive" | "conservative" ] ]
2860                 [ "limit" ( limit-item | "{" limit-list "}" ) ] |
2861                 [ "loginterface" ( interface-name | "none" ) ] |
2862                 [ "block-policy" ( "drop" | "return" ) ] |
2863                 [ "state-policy" ( "if-bound" | "floating" ) ]
2864                 [ "state-defaults" state-opts ]
2865                 [ "require-order" ( "yes" | "no" ) ]
2866                 [ "fingerprints" filename ] |
2867                 [ "skip on" ifspec ] |
2868                 [ "debug" ( "none" | "urgent" | "misc" | "loud" ) ] )
2869
2870pf-rule        = action [ ( "in" | "out" ) ]
2871                 [ "log" [ "(" logopts ")"] ] [ "quick" ]
2872                 [ "on" ifspec ] [ route ] [ af ] [ protospec ]
2873                 hosts [ filteropt-list ]
2874
2875logopts        = logopt [ "," logopts ]
2876logopt         = "all" | "user" | "to" interface-name
2877
2878filteropt-list = filteropt-list filteropt | filteropt
2879filteropt      = user | group | flags | icmp-type | icmp6-type | "tos" tos |
2880                 ( "no" | "keep" | "modulate" | "synproxy" ) "state"
2881                 [ "(" state-opts ")" ] |
2882                 "fragment" | "no-df" | "min-ttl" number | "set-tos" tos |
2883                 "max-mss" number | "random-id" | "reassemble tcp" |
2884                 fragmentation | "allow-opts" |
2885                 "label" string | "tag" string | [ ! ] "tagged" string |
2886                 "set prio" ( number | "(" number [ [ "," ] number ] ")" ) |
2887                 "queue" ( string | "(" string [ [ "," ] string ] ")" ) |
2888                 "rtable" number | "probability" number"%" | "prio" number
2889
2890nat-rule       = [ "no" ] "nat" [ "pass" [ "log" [ "(" logopts ")" ] ] ]
2891                 [ "on" ifspec ] [ af ]
2892                 [ protospec ] hosts [ "tag" string ] [ "tagged" string ]
2893                 [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" )
2894                 [ portspec ] [ pooltype ] [ "static-port" ] ]
2895
2896binat-rule     = [ "no" ] "binat" [ "pass" [ "log" [ "(" logopts ")" ] ] ]
2897                 [ "on" interface-name ] [ af ]
2898                 [ "proto" ( proto-name | proto-number ) ]
2899                 "from" address [ "/" mask-bits ] "to" ipspec
2900                 [ "tag" string ] [ "tagged" string ]
2901                 [ "-\*(Gt" address [ "/" mask-bits ] ]
2902
2903rdr-rule       = [ "no" ] "rdr" [ "pass" [ "log" [ "(" logopts ")" ] ] ]
2904                 [ "on" ifspec ] [ af ]
2905                 [ protospec ] hosts [ "tag" string ] [ "tagged" string ]
2906                 [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" )
2907                 [ portspec ] [ pooltype ] ]
2908
2909antispoof-rule = "antispoof" [ "log" ] [ "quick" ]
2910                 "for" ifspec [ af ] [ "label" string ]
2911
2912table-rule     = "table" "\*(Lt" string "\*(Gt" [ tableopts-list ]
2913tableopts-list = tableopts-list tableopts | tableopts
2914tableopts      = "persist" | "const" | "counters" | "file" string |
2915                 "{" [ tableaddr-list ] "}"
2916tableaddr-list = tableaddr-list [ "," ] tableaddr-spec | tableaddr-spec
2917tableaddr-spec = [ "!" ] tableaddr [ "/" mask-bits ]
2918tableaddr      = hostname | ifspec | "self" |
2919                 ipv4-dotted-quad | ipv6-coloned-hex
2920
2921altq-rule      = "altq on" interface-name queueopts-list
2922                 "queue" subqueue
2923queue-rule     = "queue" string [ "on" interface-name ] queueopts-list
2924                 subqueue
2925
2926anchor-rule    = "anchor" [ string ] [ ( "in" | "out" ) ] [ "on" ifspec ]
2927                 [ af ] [ protospec ] [ hosts ] [ filteropt-list ] [ "{" ]
2928
2929anchor-close   = "}"
2930
2931trans-anchors  = ( "nat-anchor" | "rdr-anchor" | "binat-anchor" ) string
2932                 [ "on" ifspec ] [ af ] [ "proto" ] [ protospec ] [ hosts ]
2933
2934load-anchor    = "load anchor" string "from" filename
2935
2936queueopts-list = queueopts-list queueopts | queueopts
2937queueopts      = [ "bandwidth" bandwidth-spec ] |
2938                 [ "qlimit" number ] | [ "tbrsize" number ] |
2939                 [ "priority" number ] | [ schedulers ]
2940schedulers     = ( cbq-def | priq-def | hfsc-def )
2941bandwidth-spec = "number" ( "b" | "Kb" | "Mb" | "Gb" | "%" )
2942
2943action         = "pass" | "block" [ return ] | [ "no" ] "scrub"
2944return         = "drop" | "return" | "return-rst" [ "( ttl" number ")" ] |
2945                 "return-icmp" [ "(" icmpcode [ [ "," ] icmp6code ] ")" ] |
2946                 "return-icmp6" [ "(" icmp6code ")" ]
2947icmpcode       = ( icmp-code-name | icmp-code-number )
2948icmp6code      = ( icmp6-code-name | icmp6-code-number )
2949
2950ifspec         = ( [ "!" ] ( interface-name | interface-group ) ) |
2951                 "{" interface-list "}"
2952interface-list = [ "!" ] ( interface-name | interface-group )
2953                 [ [ "," ] interface-list ]
2954route          = ( "route-to" | "reply-to" | "dup-to" )
2955                 ( routehost | "{" routehost-list "}" )
2956                 [ pooltype ]
2957af             = "inet" | "inet6"
2958
2959protospec      = "proto" ( proto-name | proto-number |
2960                 "{" proto-list "}" )
2961proto-list     = ( proto-name | proto-number ) [ [ "," ] proto-list ]
2962
2963hosts          = "all" |
2964                 "from" ( "any" | "no-route" | "urpf-failed" | "self" | host |
2965                 "{" host-list "}" ) [ port ] [ os ]
2966                 "to"   ( "any" | "no-route" | "self" | host |
2967                 "{" host-list "}" ) [ port ]
2968
2969ipspec         = "any" | host | "{" host-list "}"
2970host           = [ "!" ] ( address [ "/" mask-bits ] | "\*(Lt" string "\*(Gt" )
2971redirhost      = address [ "/" mask-bits ]
2972routehost      = "(" interface-name [ address [ "/" mask-bits ] ] ")"
2973address        = ( interface-name | interface-group |
2974                 "(" ( interface-name | interface-group ) ")" |
2975                 hostname | ipv4-dotted-quad | ipv6-coloned-hex )
2976host-list      = host [ [ "," ] host-list ]
2977redirhost-list = redirhost [ [ "," ] redirhost-list ]
2978routehost-list = routehost [ [ "," ] routehost-list ]
2979
2980port           = "port" ( unary-op | binary-op | "{" op-list "}" )
2981portspec       = "port" ( number | name ) [ ":" ( "*" | number | name ) ]
2982os             = "os"  ( os-name | "{" os-list "}" )
2983user           = "user" ( unary-op | binary-op | "{" op-list "}" )
2984group          = "group" ( unary-op | binary-op | "{" op-list "}" )
2985
2986unary-op       = [ "=" | "!=" | "\*(Lt" | "\*(Le" | "\*(Gt" | "\*(Ge" ]
2987                 ( name | number )
2988binary-op      = number ( "\*(Lt\*(Gt" | "\*(Gt\*(Lt" | ":" ) number
2989op-list        = ( unary-op | binary-op ) [ [ "," ] op-list ]
2990
2991os-name        = operating-system-name
2992os-list        = os-name [ [ "," ] os-list ]
2993
2994flags          = "flags" ( [ flag-set ] "/"  flag-set | "any" )
2995flag-set       = [ "F" ] [ "S" ] [ "R" ] [ "P" ] [ "A" ] [ "U" ] [ "E" ]
2996                 [ "W" ]
2997
2998icmp-type      = "icmp-type" ( icmp-type-code | "{" icmp-list "}" )
2999icmp6-type     = "icmp6-type" ( icmp-type-code | "{" icmp-list "}" )
3000icmp-type-code = ( icmp-type-name | icmp-type-number )
3001                 [ "code" ( icmp-code-name | icmp-code-number ) ]
3002icmp-list      = icmp-type-code [ [ "," ] icmp-list ]
3003
3004tos            = ( "lowdelay" | "throughput" | "reliability" |
3005                 [ "0x" ] number )
3006
3007state-opts     = state-opt [ [ "," ] state-opts ]
3008state-opt      = ( "max" number | "no-sync" | timeout | "sloppy" |
3009                 "source-track" [ ( "rule" | "global" ) ] |
3010                 "max-src-nodes" number | "max-src-states" number |
3011                 "max-src-conn" number |
3012                 "max-src-conn-rate" number "/" number |
3013                 "overload" "\*(Lt" string "\*(Gt" [ "flush" ] |
3014                 "if-bound" | "floating" )
3015
3016fragmentation  = [ "fragment reassemble" ]
3017
3018timeout-list   = timeout [ [ "," ] timeout-list ]
3019timeout        = ( "tcp.first" | "tcp.opening" | "tcp.established" |
3020                 "tcp.closing" | "tcp.finwait" | "tcp.closed" |
3021                 "udp.first" | "udp.single" | "udp.multiple" |
3022                 "icmp.first" | "icmp.error" |
3023                 "other.first" | "other.single" | "other.multiple" |
3024                 "frag" | "interval" | "src.track" |
3025                 "adaptive.start" | "adaptive.end" ) number
3026
3027limit-list     = limit-item [ [ "," ] limit-list ]
3028limit-item     = ( "states" | "frags" | "src-nodes" ) number
3029
3030pooltype       = ( "bitmask" | "random" |
3031                 "source-hash" [ ( hex-key | string-key ) ] |
3032                 "round-robin" ) [ sticky-address ]
3033
3034subqueue       = string | "{" queue-list "}"
3035queue-list     = string [ [ "," ] string ]
3036cbq-def        = "cbq" [ "(" cbq-opt [ [ "," ] cbq-opt ] ")" ]
3037priq-def       = "priq" [ "(" priq-opt [ [ "," ] priq-opt ] ")" ]
3038hfsc-def       = "hfsc" [ "(" hfsc-opt [ [ "," ] hfsc-opt ] ")" ]
3039cbq-opt        = ( "default" | "borrow" | "red" | "ecn" | "rio" )
3040priq-opt       = ( "default" | "red" | "ecn" | "rio" )
3041hfsc-opt       = ( "default" | "red" | "ecn" | "rio" |
3042                 linkshare-sc | realtime-sc | upperlimit-sc )
3043linkshare-sc   = "linkshare" sc-spec
3044realtime-sc    = "realtime" sc-spec
3045upperlimit-sc  = "upperlimit" sc-spec
3046sc-spec        = ( bandwidth-spec |
3047                 "(" bandwidth-spec number bandwidth-spec ")" )
3048include        = "include" filename
3049.Ed
3050.Sh FILES
3051.Bl -tag -width "/etc/protocols" -compact
3052.It Pa /etc/hosts
3053Host name database.
3054.It Pa /etc/pf.conf
3055Default location of the ruleset file.
3056.It Pa /etc/pf.os
3057Default location of OS fingerprints.
3058.It Pa /etc/protocols
3059Protocol name database.
3060.It Pa /etc/services
3061Service name database.
3062.El
3063.Sh SEE ALSO
3064.Xr altq 4 ,
3065.Xr carp 4 ,
3066.Xr icmp 4 ,
3067.Xr icmp6 4 ,
3068.Xr ip 4 ,
3069.Xr ip6 4 ,
3070.Xr pf 4 ,
3071.Xr pfsync 4 ,
3072.Xr tcp 4 ,
3073.Xr udp 4 ,
3074.Xr hosts 5 ,
3075.Xr pf.os 5 ,
3076.Xr protocols 5 ,
3077.Xr services 5 ,
3078.Xr ftp-proxy 8 ,
3079.Xr pfctl 8 ,
3080.Xr pflogd 8
3081.Sh HISTORY
3082The
3083.Nm
3084file format first appeared in
3085.Ox 3.0 .
3086