xref: /openbsd-src/share/man/man5/pf.conf.5 (revision fc405d53b73a2d73393cb97f684863d17b583e38)
1.\"	$OpenBSD: pf.conf.5,v 1.600 2022/11/18 18:11:10 kn Exp $
2.\"
3.\" Copyright (c) 2002, Daniel Hartmeier
4.\" Copyright (c) 2003 - 2013 Henning Brauer <henning@openbsd.org>
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 $Mdocdate: November 18 2022 $
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 .
43.Pp
44This is an overview of the sections in this manual page:
45.Bl -inset
46.It Sx PACKET FILTERING
47including network address translation (NAT).
48.It Sx OPTIONS
49globally tune the behaviour of the packet filtering engine.
50.It Sx QUEUEING
51provides rule-based bandwidth and traffic control.
52.It Sx TABLES
53provide a method for dealing with large numbers of addresses.
54.It Sx ANCHORS
55are containers for rules and tables.
56.It Sx STATEFUL FILTERING
57tracks packets by state.
58.It Sx TRAFFIC NORMALISATION
59includes scrub, fragment handling, and blocking spoofed traffic.
60.It Sx OPERATING SYSTEM FINGERPRINTING
61is a method for detecting a host's operating system.
62.It Sx EXAMPLES
63provides some example rulesets.
64.It Sx GRAMMAR
65provides a complete BNF grammar reference.
66.El
67.Pp
68The current line can be extended over multiple lines using a backslash
69.Pq Sq \e .
70Comments can be put anywhere in the file using a hash mark
71.Pq Sq # ,
72and extend to the end of the current line.
73Care should be taken when commenting out multi-line text:
74the comment is effective until the end of the entire block.
75.Pp
76Argument names not beginning with a letter, digit, or underscore
77must be quoted.
78.Pp
79Additional configuration files can be included with the
80.Ic include
81keyword, for example:
82.Bd -literal -offset indent
83include "/etc/pf/sub.filter.conf"
84.Ed
85.Pp
86Macros can be defined that will later be expanded in context.
87Macro names must start with a letter, digit, or underscore,
88and may contain any of those characters.
89Macro names may not be reserved words (for example
90.Ic pass ,
91.Cm in ,
92.Cm out ) .
93Macros are not expanded inside quotes.
94.Pp
95For example:
96.Bd -literal -offset indent
97ext_if = "kue0"
98all_ifs = "{" $ext_if lo0 "}"
99pass out on $ext_if from any to any
100pass in  on $ext_if proto tcp from any to any port 25
101.Ed
102.Sh PACKET FILTERING
103.Xr pf 4
104has the ability to
105.Ic block ,
106.Ic pass ,
107and
108.Ic match
109packets based on attributes of their layer 3
110and layer 4 headers.
111Filter rules determine which of these actions are taken;
112filter parameters specify the packets to which a rule applies.
113.Pp
114Each time a packet processed by the packet filter comes in on or
115goes out through an interface, the filter rules are evaluated in
116sequential order, from first to last.
117For
118.Ic block
119and
120.Ic pass ,
121the last matching rule decides what action is taken;
122if no rule matches the packet, the default action is to pass
123the packet without creating a state.
124For
125.Ic match ,
126rules are evaluated every time they match;
127the pass/block state of a packet remains unchanged.
128.Pp
129Most parameters are optional.
130If a parameter is specified, the rule only applies to packets with
131matching attributes.
132The matching for some parameters can be inverted with the
133.Cm !\&
134operator.
135Certain parameters can be expressed as lists, in which case
136.Xr pfctl 8
137generates all needed rule combinations.
138.Pp
139By default
140.Xr pf 4
141filters packets statefully:
142the first time a packet matches a
143.Ic pass
144rule, a state entry is created.
145The packet filter examines each packet to see if it matches an existing state.
146If it does, the packet is passed without evaluation of any rules.
147After the connection is closed or times out, the state entry is automatically
148removed.
149.Pp
150The following actions can be used in the filter:
151.Bl -tag -width Ds
152.It Ic block
153The packet is blocked.
154There are a number of ways in which a
155.Ic block
156rule can behave when blocking a packet.
157The default behaviour is to
158.Cm drop
159packets silently, however this can be overridden or made
160explicit either globally, by setting the
161.Cm block-policy
162option, or on a per-rule basis with one of the following options:
163.Pp
164.Bl -tag -width return-icmp6 -compact
165.It Cm drop
166The packet is silently dropped.
167.It Cm return
168This causes a TCP RST to be returned for TCP packets
169and an ICMP UNREACHABLE for other types of packets.
170.It Cm return-icmp
171.It Cm return-icmp6
172This causes ICMP messages to be returned for packets which match the rule.
173By default this is an ICMP UNREACHABLE message, however this
174can be overridden by specifying a message as a code or number.
175.It Cm return-rst
176This applies only to TCP packets,
177and issues a TCP RST which closes the connection.
178An optional parameter,
179.Cm ttl ,
180may be given with a TTL value.
181.El
182.Pp
183Options returning ICMP packets currently have no effect if
184.Xr pf 4
185operates on a
186.Xr bridge 4 ,
187as the code to support this feature has not yet been implemented.
188.Pp
189The simplest mechanism to block everything by default and only pass
190packets that match explicit rules is specify a first filter rule of:
191.Pp
192.Dl block all
193.It Ic match
194The packet is matched.
195This mechanism is used to provide fine grained filtering
196without altering the block/pass state of a packet.
197.Ic match
198rules differ from
199.Ic block
200and
201.Ic pass
202rules in that parameters are set every time a packet matches the
203rule, not only on the last matching rule.
204For the following parameters,
205this means that the parameter effectively becomes
206.Dq sticky
207until explicitly overridden:
208.Cm nat-to ,
209.Cm binat-to ,
210.Cm rdr-to ,
211.Cm queue ,
212.Cm rtable ,
213and
214.Cm scrub .
215.Pp
216.Cm log
217is different still,
218in that the action happens every time a rule matches
219i.e. a single packet can get logged more than once.
220.It Ic pass
221The packet is passed;
222state is created unless the
223.Cm no state
224option is specified.
225.El
226.Pp
227The following parameters can be used in the filter:
228.Bl -tag -width Ds
229.It Cm in No or Cm out
230A packet always comes in on, or goes out through, one interface.
231.Cm in
232and
233.Cm out
234apply to incoming and outgoing packets;
235if neither are specified,
236the rule will match packets in both directions.
237.It Cm log Pq Cm all | matches | to Ar interface | Cm user
238In addition to any action specified,
239log the packet.
240Only the packet that establishes the state is logged,
241unless the
242.Cm no state
243option is specified.
244The logged packets are sent to a
245.Xr pflog 4
246interface, by default
247.Pa pflog0 ;
248pflog0 is monitored by the
249.Xr pflogd 8
250logging daemon which logs to the file
251.Pa /var/log/pflog
252in pcap binary format.
253.Pp
254The keywords
255.Cm all , matches , to ,
256and
257.Cm user
258are optional and can be combined using commas,
259but must be enclosed in parentheses if given.
260.Pp
261Use
262.Cm all
263to force logging of all packets for a connection.
264This is not necessary when
265.Cm no state
266is explicitly specified.
267.Pp
268If
269.Cm matches
270is specified,
271it logs the packet on all subsequent matching rules.
272It is often combined with
273.Cm to Ar interface
274to avoid adding noise to the default log file.
275.Pp
276The keyword
277.Cm user
278logs the UID and PID of the
279socket on the local host used to send or receive a packet,
280in addition to the normal information.
281.Pp
282To specify a logging interface other than
283.Pa pflog0 ,
284use the syntax
285.Cm to Ar interface .
286.It Cm quick
287If a packet matches a rule which has the
288.Cm quick
289option set, this rule
290is considered the last matching rule, and evaluation of subsequent rules
291is skipped.
292.It Cm on Ar interface | Cm any
293This rule applies only to packets coming in on, or going out through, this
294particular interface or interface group.
295For more information on interface groups,
296see the
297.Ic group
298keyword in
299.Xr ifconfig 8 .
300.Cm any
301will match any existing interface except loopback ones.
302.It Cm on rdomain Ar number
303This rule applies only to packets coming in on, or going out through, this
304particular routing domain.
305.It Cm inet | inet6
306This rule applies only to packets of this address family.
307.It Cm proto Ar protocol
308This rule applies only to packets of this protocol.
309Common protocols are ICMP, ICMP6, TCP, and UDP.
310For a list of all the protocol name to number mappings used by
311.Xr pfctl 8 ,
312see the file
313.Pa /etc/protocols .
314.It Xo
315.Cm from Ar source
316.Cm port Ar source
317.Cm os Ar source
318.Cm to Ar dest
319.Cm port Ar dest
320.Xc
321This rule applies only to packets with the specified source and destination
322addresses and ports.
323.Pp
324Addresses can be specified in CIDR notation (matching netblocks), as
325symbolic host names, interface names or interface group names, or as any
326of the following keywords:
327.Pp
328.Bl -tag -width urpf-failed -compact
329.It Cm any
330Any address.
331.It Cm no-route
332Any address which is not currently routable.
333.It Cm route Ar label
334Any address matching the given
335.Xr route 8
336.Ar label .
337.It Cm self
338Expands to all addresses assigned to all interfaces.
339.It Pf < Ar table Ns >
340Any address matching the given
341.Ar table .
342.It Cm urpf-failed
343Any source address that fails a unicast reverse path forwarding (URPF)
344check, i.e. packets coming in on an interface other than that which holds
345the route back to the packet's source address.
346.El
347.Pp
348Ranges of addresses are specified using the
349.Sq -
350operator.
351For instance:
352.Dq 10.1.1.10 - 10.1.1.12
353means all addresses from 10.1.1.10 to 10.1.1.12,
354hence addresses 10.1.1.10, 10.1.1.11, and 10.1.1.12.
355.Pp
356Interface names, interface group names, and
357.Cm self
358can have modifiers appended:
359.Pp
360.Bl -tag -width :broadcast -compact
361.It Cm :0
362Do not include interface aliases.
363.It Cm :broadcast
364Translates to the interface's broadcast address(es).
365.It Cm :network
366Translates to the network(s) attached to the interface.
367.It Cm :peer
368Translates to the point-to-point interface's peer address(es).
369.El
370.Pp
371Host names may also have the
372.Cm :0
373modifier appended to restrict the name resolution to the first of each
374v4 and v6 address found.
375.Pp
376Host name resolution and interface to address translation are done at
377ruleset load-time.
378When the address of an interface (or host name) changes (under DHCP or PPP,
379for instance), the ruleset must be reloaded for the change to be reflected
380in the kernel.
381Surrounding the interface name (and optional modifiers) in parentheses
382changes this behaviour.
383When the interface name is surrounded by parentheses, the rule is
384automatically updated whenever the interface changes its address.
385The ruleset does not need to be reloaded.
386This is especially useful with NAT.
387.Pp
388Ports can be specified either by number or by name.
389For example, port 80 can be specified as
390.Cm www .
391For a list of all port name to number mappings used by
392.Xr pfctl 8 ,
393see the file
394.Pa /etc/services .
395.Pp
396Ports and ranges of ports are specified using these operators:
397.Bd -literal -offset indent
398=	(equal)
399!=	(unequal)
400<	(less than)
401<=	(less than or equal)
402>	(greater than)
403>=	(greater than or equal)
404:	(range including boundaries)
405><	(range excluding boundaries)
406<>	(except range)
407.Ed
408.Pp
409.Sq >< ,
410.Sq <>
411and
412.Sq \&:
413are binary operators (they take two arguments).
414For instance:
415.Bl -tag -width Ds
416.It Li port 2000:2004
417means
418.Sq all ports \(>= 2000 and \(<= 2004 ,
419hence ports 2000, 2001, 2002, 2003, and 2004.
420.It Li port 2000 >< 2004
421means
422.Sq all ports > 2000 and < 2004 ,
423hence ports 2001, 2002, and 2003.
424.It Li port 2000 <> 2004
425means
426.Sq all ports < 2000 or > 2004 ,
427hence ports 1\(en1999 and 2005\(en65535.
428.El
429.Pp
430The operating system of the source host can be specified in the case of TCP
431rules with the
432.Cm os
433modifier.
434See the
435.Sx OPERATING SYSTEM FINGERPRINTING
436section for more information.
437.Pp
438The
439.Cm host ,
440.Cm port ,
441and
442.Cm os
443specifications are optional, as in the following examples:
444.Bd -literal -offset indent
445pass in all
446pass in from any to any
447pass in proto tcp from any port < 1024 to any
448pass in proto tcp from any to any port 25
449pass in proto tcp from 10.0.0.0/8 port >= 1024 \e
450      to ! 10.1.2.3 port != ssh
451pass in proto tcp from any os "OpenBSD"
452pass in proto tcp from route "DTAG"
453.Ed
454.El
455.Pp
456The following additional parameters can be used in the filter:
457.Pp
458.Bl -tag -width Ds -compact
459.It Cm all
460This is equivalent to
461.Ql from any to any .
462.Pp
463.It Cm allow-opts
464By default, packets with IPv4 options or IPv6 hop-by-hop or destination
465options header are blocked.
466When
467.Cm allow-opts
468is specified for a
469.Ic pass
470rule, packets that pass the filter based on that rule (last matching)
471do so even if they contain options.
472For packets that match state, the rule that initially created the
473state is used.
474The implicit pass rule, that is used when a packet does not match
475any rules, does not allow IP options or option headers.
476Note that IPv6 packets with type 0 routing headers are always dropped.
477.Pp
478.It Cm divert-packet port Ar port
479Used to send matching packets to
480.Xr divert 4
481sockets bound to port
482.Ar port .
483If the default option of fragment reassembly is enabled, scrubbing with
484.Cm reassemble tcp
485is also enabled for
486.Cm divert-packet
487rules.
488.Pp
489.It Cm divert-reply
490Used to receive replies for sockets that are bound to addresses
491which are not local to the machine.
492See
493.Xr setsockopt 2
494for information on how to bind these sockets.
495.Pp
496.It Cm divert-to Ar host Cm port Ar port
497Used to redirect packets to a local socket bound to
498.Ar host
499and
500.Ar port .
501The packets will not be modified, preserving the original destination
502address for the application to access.
503.Dv SOCK_STREAM
504connections can access the original destination address using
505.Xr getsockname 2 .
506.Dv SOCK_DGRAM
507sockets can be configured with the
508.Xr ip 4
509.Dv IP_RECVDSTADDR
510and
511.Dv IP_RECVDSTPORT
512socket options when receiving IPv4 packets, or the
513.Xr ip6 4
514.Dv IPV6_RECVPKTINFO
515and
516.Dv IPV6_RECVDSTPORT
517socket options when receiving IPv6 packets.
518.Pp
519.It Cm flags Ar a Ns / Ns Ar b | Cm any
520This rule only applies to TCP packets that have the flags
521.Ar a
522set out of set
523.Ar b .
524Flags not specified in
525.Ar b
526are ignored.
527For stateful connections, the default is
528.Cm flags S/SA .
529To indicate that flags should not be checked at all, specify
530.Cm flags any .
531The flags are: (F)IN, (S)YN, (R)ST, (P)USH, (A)CK, (U)RG, (E)CE, and C(W)R.
532.Bl -tag -width "flags /SFRA"
533.It Cm flags S/S
534Flag SYN is set.
535The other flags are ignored.
536.It Cm flags S/SA
537This is the default setting for stateful connections.
538Out of SYN and ACK, exactly SYN may be set.
539SYN, SYN+PSH, and SYN+RST match, but SYN+ACK, ACK, and ACK+RST do not.
540This is more restrictive than the previous example.
541.It Cm flags /SFRA
542If the first set is not specified, it defaults to none.
543All of SYN, FIN, RST, and ACK must be unset.
544.El
545.Pp
546Because
547.Cm flags S/SA
548is applied by default (unless
549.Cm no state
550is specified), only the initial SYN packet of a TCP handshake will create
551a state for a TCP connection.
552It is possible to be less restrictive, and allow state creation from
553intermediate
554.Pq non-SYN
555packets, by specifying
556.Cm flags any .
557This will cause
558.Xr pf 4
559to synchronize to existing connections, for instance
560if one flushes the state table.
561However, states created from such intermediate packets may be missing
562connection details such as the TCP window scaling factor.
563States which modify the packet flow, such as those affected by
564.Cm af-to ,
565.Cm modulate state ,
566.Cm nat-to ,
567.Cm rdr-to ,
568or
569.Cm synproxy state
570options, or scrubbed with
571.Cm reassemble tcp ,
572will also not be recoverable from intermediate packets.
573Such connections will stall and time out.
574.Pp
575.It Cm group Ar group
576Similar to
577.Cm user ,
578this rule only applies to packets of sockets owned by the specified
579.Ar group .
580.Pp
581.It Cm icmp-type Ar type Op Cm code Ar code
582.It Cm icmp6-type Ar type Op Cm code Ar code
583This rule only applies to ICMP or ICMP6 packets with the specified type
584and code.
585Text names for ICMP types and codes are listed in
586.Xr icmp 4
587and
588.Xr icmp6 4 .
589The protocol and the ICMP type indicator
590.Po
591.Cm icmp-type
592or
593.Cm icmp6-type
594.Pc
595must match.
596.Pp
597ICMP responses are not permitted unless they either match an
598existing request, or unless
599.Cm no state
600or
601.Cm keep state (sloppy)
602is specified.
603.Pp
604.It Cm label Ar string
605Adds a label to the rule, which can be used to identify the rule.
606For instance,
607.Ql pfctl -s labels
608shows per-rule statistics for rules that have labels.
609.Pp
610The following macros can be used in labels:
611.Pp
612.Bl -tag -width "$srcaddrXXX" -compact -offset indent
613.It Va $dstaddr
614The destination IP address.
615.It Va $dstport
616The destination port specification.
617.It Va $if
618The interface.
619.It Va $nr
620The rule number.
621.It Va $proto
622The protocol name.
623.It Va $srcaddr
624The source IP address.
625.It Va $srcport
626The source port specification.
627.El
628.Pp
629For example:
630.Bd -literal -offset indent -compact
631ips = "{ 1.2.3.4, 1.2.3.5 }"
632pass in proto tcp from any to $ips \e
633      port > 1023 label "$dstaddr:$dstport"
634.Ed
635.Pp
636Expands to:
637.Bd -literal -offset indent -compact
638pass in inet proto tcp from any to 1.2.3.4 \e
639      port > 1023 label "1.2.3.4:>1023"
640pass in inet proto tcp from any to 1.2.3.5 \e
641      port > 1023 label "1.2.3.5:>1023"
642.Ed
643.Pp
644The macro expansion for the
645.Cm label
646directive occurs only at configuration file parse time, not during runtime.
647.Pp
648.It Cm max-pkt-rate Ar number Ns / Ns Ar seconds
649Measure the rate of packets matching the rule and states created by it.
650When the specified rate is exceeded, the rule stops matching.
651Only packets in the direction in which the state was created are considered,
652so that typically requests are counted and replies are not.
653For example,
654to pass up to 100 ICMP packets per 10 seconds:
655.Bd -literal -offset indent
656block in proto icmp
657pass in proto icmp max-pkt-rate 100/10
658.Ed
659.Pp
660When the rate is exceeded, all ICMP is blocked until the rate falls below
661100 per 10 seconds again.
662.Pp
663.It Cm once
664Create a one shot rule.
665The first matching packet marks the rule as expired.
666Expired rules are skipped and hidden, unless
667.Xr pfctl 8
668is used in debug or verbose mode.
669.Pp
670.It Cm probability Ar number Ns %
671A probability attribute can be attached to a rule,
672with a value set between 0 and 100%,
673in which case the rule is honoured using the given probability value.
674For example, the following rule will drop 20% of incoming ICMP packets:
675.Pp
676.Dl block in proto icmp probability 20%
677.Pp
678.It Cm prio Ar number
679Only match packets which have the given queueing priority assigned.
680.Pp
681.It Oo Cm \&! Oc Ns Cm received-on Ar interface
682Only match packets which were received on the specified
683.Cm interface
684(or interface group).
685.Cm any
686will match any existing interface except loopback ones.
687.Pp
688.It Cm rtable Ar number
689Used to select an alternate routing table for the routing lookup.
690Only effective before the route lookup happened, i.e. when filtering inbound.
691.Pp
692.It Cm set delay Ar milliseconds
693Packets matching this rule will be delayed at the outbound interface by the
694given number of milliseconds.
695.Pp
696.It Cm set prio Ar priority | Pq Ar priority , priority
697Packets matching this rule will be assigned a specific queueing priority.
698Priorities are assigned as integers 0 through 7,
699with a default priority of 3.
700If the packet is transmitted on a
701.Xr vlan 4
702interface, the queueing priority will also be written as the priority
703code point in the 802.1Q VLAN header.
704If two priorities are given, TCP ACKs with no data payload and packets
705which have a TOS of
706.Cm lowdelay
707will be assigned to the second one.
708Packets with a higher priority number are processed first,
709and packets with the same priority are processed
710in the order in which they are received.
711.Pp
712For example:
713.Bd -literal -offset indent
714pass in proto tcp to port 25 set prio 2
715pass in proto tcp to port 22 set prio (2, 5)
716.Ed
717.Pp
718The interface priority queues accessed by the
719.Cm set prio
720keyword are always enabled and do not require any additional
721configuration, unlike the queues described below and in the
722.Sx QUEUEING
723section.
724.Pp
725.It Cm set queue Ar queue | Pq Ar queue , queue
726Packets matching this rule will be assigned to the specified
727.Ar queue .
728If two queues are given, packets which have a TOS of
729.Cm lowdelay
730and TCP ACKs with no data payload will be assigned to the second one.
731See
732.Sx QUEUEING
733for setup details.
734.Pp
735For example:
736.Bd -literal -offset indent
737pass in proto tcp to port 25 set queue mail
738pass in proto tcp to port 22 set queue(ssh_bulk, ssh_prio)
739.Ed
740.Pp
741.It Cm set tos Ar string | number
742Enforces a TOS for matching packets.
743.Ar string
744may be one of
745.Cm critical ,
746.Cm inetcontrol ,
747.Cm lowdelay ,
748.Cm netcontrol ,
749.Cm throughput ,
750.Cm reliability ,
751or one of the DiffServ Code Points:
752.Cm ef ,
753.Cm af11 No ... Cm af43 ,
754.Cm cs0 No ... Cm cs7 ;
755.Ar number
756may be either a hex or decimal number.
757.Pp
758.It Cm tag Ar string
759Packets matching this rule will be tagged with the specified
760.Ar string .
761The tag acts as an internal marker that can be used to
762identify these packets later on.
763This can be used, for example, to provide trust between
764interfaces and to determine if packets have been
765processed by translation rules.
766Tags are
767.Dq sticky ,
768meaning that the packet will be tagged even if the rule
769is not the last matching rule.
770Further matching rules can replace the tag with a
771new one but will not remove a previously applied tag.
772A packet is only ever assigned one tag at a time.
773Tags take the same macros as labels (see above).
774.Pp
775.It Oo Cm \&! Oc Ns Cm tagged Ar string
776Used with filter or translation rules
777to specify that packets must already
778be tagged with the given
779.Ar string
780in order to match the rule.
781.Pp
782.It Cm tos Ar string | number
783This rule applies to packets with the specified TOS bits set.
784.Ar string
785may be one of
786.Cm critical ,
787.Cm inetcontrol ,
788.Cm lowdelay ,
789.Cm netcontrol ,
790.Cm throughput ,
791.Cm reliability ,
792or one of the DiffServ Code Points:
793.Cm ef ,
794.Cm af11 No ... Cm af43 ,
795.Cm cs0 No ... Cm cs7 ;
796.Ar number
797may be either a hex or decimal number.
798.Pp
799For example, the following rules are identical:
800.Bd -literal -offset indent
801pass all tos lowdelay
802pass all tos 0x10
803pass all tos 16
804.Ed
805.Pp
806.It Cm user Ar user
807This rule only applies to packets of sockets owned by the specified
808.Ar user .
809For outgoing connections initiated from the firewall, this is the user
810that opened the connection.
811For incoming connections to the firewall itself, this is the user that
812listens on the destination port.
813.Pp
814When listening sockets are bound to the wildcard address,
815.Xr pf 4
816cannot determine if a connection is destined for the firewall itself.
817To avoid false matches on just the destination port, combine a
818.Cm user
819rule with source or destination address
820.Cm self .
821.Pp
822All packets, both outgoing and incoming, of one connection are associated
823with the same user and group.
824Only TCP and UDP packets can be associated with users.
825.Pp
826The
827.Ar user
828and
829.Ar group
830arguments refer to the effective (as opposed to the real) IDs, in
831case the socket is created by a setuid/setgid process.
832User and group IDs are stored when a socket is created;
833when a process creates a listening socket as root (for instance, by
834binding to a privileged port) and subsequently changes to another
835user ID (to drop privileges), the credentials will remain root.
836.Pp
837User and group IDs can be specified as either numbers or names.
838The syntax is similar to the one for ports.
839The following example allows only selected users to open outgoing
840connections:
841.Bd -literal -offset indent
842block out proto tcp all
843pass  out proto tcp from self user { < 1000, dhartmei }
844.Ed
845.Pp
846The example below permits users with uid between 1000 and 1500
847to open connections:
848.Bd -literal -offset indent
849block out proto tcp all
850pass  out proto tcp from self user { 999 >< 1501 }
851.Ed
852.Pp
853The
854.Sq \&:
855operator, which works for port number matching, does not work for
856.Cm user
857and
858.Cm group
859match.
860.El
861.Ss Translation
862Translation options modify either the source or destination address and
863port of the packets associated with a stateful connection.
864.Xr pf 4
865modifies the specified address and/or port in the packet and recalculates
866IP, TCP, and UDP checksums as necessary.
867.Pp
868If specified on a
869.Ic match
870rule, subsequent rules will see packets as they look
871after any addresses and ports have been translated.
872These rules will therefore have to filter based on the translated
873address and port number.
874.Pp
875The state entry created permits
876.Xr pf 4
877to keep track of the original address for traffic associated with that state
878and correctly direct return traffic for that connection.
879.Pp
880Different types of translation are possible with pf:
881.Bl -tag -width binat-to
882.It Cm af-to
883Translation between different address families (NAT64) is handled
884using
885.Cm af-to
886rules.
887Because address family translation overrides the routing table, it's
888only possible to use
889.Cm af-to
890on inbound rules, and a source address for the resulting translation
891must always be specified.
892.Pp
893The optional second argument is the host or subnet the original
894addresses are translated into for the destination.
895The lowest bits of the original destination address form the host
896part of the new destination address according to the specified subnet.
897It is possible to embed a complete IPv4 address into an IPv6 address
898using a network prefix of /96 or smaller.
899.Pp
900When a destination address is not specified, it is assumed that the host
901part is 32-bit long.
902For IPv6 to IPv4 translation this would mean using only the lower 32
903bits of the original IPv6 destination address.
904For IPv4 to IPv6 translation the destination subnet defaults to the
905subnet of the new IPv6 source address with a prefix length of /96.
906See RFC 6052 Section 2.2 for details on how the prefix determines the
907destination address encoding.
908.Pp
909For example, the following rules are identical:
910.Bd -literal -offset indent
911pass in inet af-to inet6 from 2001:db8::1 to 2001:db8::/96
912pass in inet af-to inet6 from 2001:db8::1
913.Ed
914.Pp
915In the above example the matching IPv4 packets will be modified to
916have a source address of 2001:db8::1 and a destination address will
917get prefixed with 2001:db8::/96, e.g. 198.51.100.100 will be
918translated to 2001:db8::c633:6464.
919.Pp
920In the reverse case the following rules are identical:
921.Bd -literal -offset indent
922pass in inet6 from any to 64:ff9b::/96 af-to inet \e
923	from 198.51.100.1 to 0.0.0.0/0
924pass in inet6 from any to 64:ff9b::/96 af-to inet \e
925	from 198.51.100.1
926.Ed
927.Pp
928The destination IPv4 address is assumed to be embedded inside the
929original IPv6 destination address, e.g. 64:ff9b::c633:6464 will be
930translated to 198.51.100.100.
931.Pp
932The current implementation will only extract IPv4 addresses from the
933IPv6 addresses with a prefix length of /96 and greater.
934.It Cm binat-to
935A
936.Cm binat-to
937rule specifies a bidirectional mapping between an external IP
938netblock and an internal IP netblock.
939It expands to an outbound
940.Cm nat-to
941rule and an inbound
942.Cm rdr-to
943rule.
944.It Cm nat-to
945A
946.Cm nat-to
947option specifies that IP addresses are to be changed as the packet
948traverses the given interface.
949This technique allows one or more IP addresses
950on the translating host to support network traffic for a larger range of
951machines on an
952.Dq inside
953network.
954Although in theory any IP address can be used on the inside, it is strongly
955recommended that one of the address ranges defined by RFC 1918 be used.
956Those netblocks are:
957.Bd -literal -offset indent
95810.0.0.0 \(en 10.255.255.255 (all of net 10, i.e. 10/8)
959172.16.0.0 \(en 172.31.255.255 (i.e. 172.16/12)
960192.168.0.0 \(en 192.168.255.255 (i.e. 192.168/16)
961.Ed
962.Pp
963.Cm nat-to
964is usually applied outbound.
965If applied inbound, nat-to to a local IP address is not supported.
966.It Cm rdr-to
967The packet is redirected to another destination and possibly a
968different port.
969.Cm rdr-to
970can optionally specify port ranges instead of single ports.
971For instance:
972.Bl -tag -width Ds
973.It match in ... port 2000:2999 rdr-to ... port 4000
974redirects ports 2000 to 2999 (inclusive) to port 4000.
975.It match in ... port 2000:2999 rdr-to ... port 4000:*
976redirects port 2000 to 4000, port 2001 to 4001, ..., port 2999 to 4999.
977.El
978.Pp
979.Cm rdr-to
980is usually applied inbound.
981If applied outbound, rdr-to to a local IP address is not supported.
982.El
983.Pp
984In addition to modifying the address, some translation rules may modify
985source or destination ports for TCP or UDP connections;
986implicitly in the case of
987.Cm nat-to
988options and explicitly in the case of
989.Cm rdr-to
990ones.
991Port numbers are never translated with a
992.Cm binat-to
993rule.
994.Pp
995Translation options apply only to packets that pass through the specified
996interface, and if no interface is specified, translation is applied
997to packets on all interfaces.
998For instance, redirecting port 80 on an external interface to an internal
999web server will only work for connections originating from the outside.
1000Connections to the address of the external interface from local hosts will
1001not be redirected, since such packets do not actually pass through the
1002external interface.
1003Redirections cannot reflect packets back through the interface they arrive
1004on, they can only be redirected to hosts connected to different interfaces
1005or to the firewall itself.
1006.Pp
1007However packets may be redirected to hosts connected to the interface the
1008packet arrived on by using redirection with NAT.
1009For example:
1010.Bd -literal -offset indent
1011pass in on $int_if proto tcp from $int_net to $ext_if port 80 \e
1012	rdr-to $server
1013pass out on $int_if proto tcp to $server port 80 \e
1014	received-on $int_if nat-to $int_if
1015.Ed
1016.Pp
1017Note that redirecting external incoming connections to the loopback address
1018will effectively allow an external host to connect to daemons
1019bound solely to the loopback address, circumventing the traditional
1020blocking of such connections on a real interface.
1021For example:
1022.Bd -literal -offset indent
1023pass in on egress proto tcp from any to any port smtp \e
1024	rdr-to 127.0.0.1 port spamd
1025.Ed
1026.Pp
1027Unless this effect is desired, any of the local non-loopback addresses
1028should be used instead as the redirection target, which allows external
1029connections only to daemons bound to this address or not bound to
1030any address.
1031.Pp
1032For
1033.Cm af-to ,
1034.Cm nat-to
1035and
1036.Cm rdr-to
1037options for which there is a single redirection address which has a
1038subnet mask smaller than 32 for IPv4 or 128 for IPv6 (more than one IP
1039address), a variety of different methods for assigning this address can be
1040used:
1041.Bl -tag -width xxxx
1042.It Cm bitmask
1043The
1044.Cm bitmask
1045option applies the network portion of the redirection address to the address
1046to be modified (source with
1047.Cm nat-to ,
1048destination with
1049.Cm rdr-to ) .
1050.It Cm least-states Op Cm sticky-address
1051The
1052.Cm least-states
1053option selects the address with the least active states from
1054a given address pool and considers given weights
1055associated with address(es).
1056Weights can be specified between 1 and 65535.
1057Addresses with higher weights are selected more often.
1058.Pp
1059.Cm sticky-address
1060can be specified to ensure that multiple connections from the
1061same source are mapped to the same redirection address.
1062Associations are destroyed as soon as there are
1063no longer states which refer to them;
1064in order to make the mappings last
1065beyond the lifetime of the states,
1066increase the global options with
1067.Ic set Cm timeout src.track .
1068.It Cm random Op Cm sticky-address
1069The
1070.Cm random
1071option selects an address at random within the defined block of addresses.
1072.Cm sticky-address
1073is as described above.
1074.It Cm round-robin Op Cm sticky-address
1075The
1076.Cm round-robin
1077option loops through the redirection address(es) and considers given weights
1078associated with address(es).
1079Weights can be specified between 1 and 65535.
1080Addresses with higher weights are selected more often.
1081.Cm sticky-address
1082is as described above.
1083.It Cm source-hash Oo Ar key Oc Op Cm sticky-address
1084The
1085.Cm source-hash
1086option uses a hash of the source address to determine the redirection address,
1087ensuring that the redirection address is always the same for a given source.
1088An optional
1089.Ar key
1090can be specified after this keyword either in hex or as a string;
1091by default
1092.Xr pfctl 8
1093randomly generates a key for source-hash every time the
1094ruleset is reloaded.
1095.Cm sticky-address
1096is as described above.
1097.It Cm static-port
1098With
1099.Cm nat-to
1100rules, the
1101.Cm static-port
1102option prevents
1103.Xr pf 4
1104from modifying the source port on TCP and UDP packets.
1105.El
1106.Pp
1107When more than one redirection address or a table is specified,
1108.Cm bitmask
1109is not permitted as a pool type.
1110.Ss Routing
1111If a packet matches a rule with one of the following route options set,
1112the packet filter will route the packet according to the type of route option.
1113When such a rule creates state, the route option is also applied to all
1114packets matching the same connection.
1115.Bl -tag -width route-to
1116.It Cm dup-to
1117The
1118.Cm dup-to
1119option creates a duplicate of the packet and routes it like
1120.Cm route-to .
1121The original packet gets routed as it normally would.
1122.It Cm reply-to
1123The
1124.Cm reply-to
1125option is similar to
1126.Cm route-to ,
1127but routes packets that pass in the opposite direction (replies) to the
1128specified address.
1129Opposite direction is only defined in the context of a state entry, and
1130.Cm reply-to
1131is useful only in rules that create state.
1132It can be used on systems with multiple paths to the internet to ensure
1133that replies to an incoming network connection to a particular address
1134are sent using the path associated with that address (symmetric routing
1135enforcement).
1136.It Cm route-to
1137The
1138.Cm route-to
1139option routes the packet to the specified destination address instead
1140of the destination address in the packet header.
1141When a
1142.Cm route-to
1143rule creates state, only packets that pass in the same direction as the
1144filter rule specifies will be routed in this way.
1145Packets passing in the opposite direction (replies) are not affected
1146and are routed normally.
1147.El
1148.Pp
1149For the
1150.Cm dup-to ,
1151.Cm reply-to ,
1152and
1153.Cm route-to
1154route options
1155for which there is a single redirection address which has a
1156subnet mask smaller than 32 for IPv4 or 128 for IPv6 (more than one IP
1157address),
1158the methods
1159.Cm least-states ,
1160.Cm random ,
1161.Cm round-robin ,
1162and
1163.Cm source-hash ,
1164as described above,
1165can be used.
1166.Sh OPTIONS
1167.Xr pf 4
1168may be tuned for various situations using the
1169.Ic set
1170command.
1171.Bl -tag -width Ds
1172.It Ic set Cm block-policy drop | return
1173The
1174.Cm block-policy
1175option sets the default behaviour for the packet
1176.Ic block
1177action:
1178.Pp
1179.Bl -tag -width return -compact
1180.It Cm drop
1181Packet is silently dropped.
1182.It Cm return
1183A TCP RST is returned for blocked TCP packets,
1184an ICMP UNREACHABLE is returned for blocked UDP packets,
1185and all other packets are silently dropped.
1186.El
1187.Pp
1188The default value is
1189.Cm drop .
1190.It Ic set Cm debug Ar level
1191Set the debug
1192.Ar level ,
1193which limits the severity of log messages printed by
1194.Xr pf 4 .
1195This should be a keyword from the following ordered list
1196(highest to lowest):
1197.Cm emerg ,
1198.Cm alert ,
1199.Cm crit ,
1200.Cm err ,
1201.Cm warning ,
1202.Cm notice ,
1203.Cm info ,
1204and
1205.Cm debug .
1206These keywords correspond to the similar (LOG_) values specified to the
1207.Xr syslog 3
1208library routine.
1209The default value is
1210.Cm err .
1211.It Cm set Cm fingerprints Ar filename
1212Load fingerprints of known operating systems from the given
1213.Ar filename .
1214By default fingerprints of known operating systems are automatically
1215loaded from
1216.Xr pf.os 5 ,
1217but can be overridden via this option.
1218Setting this option may leave a small period of time where the fingerprints
1219referenced by the currently active ruleset are inconsistent until the new
1220ruleset finishes loading.
1221The default location for fingerprints is
1222.Pa /etc/pf.os .
1223.It Ic set Cm hostid Ar number
1224The 32-bit hostid
1225.Ar number
1226identifies this firewall's state table entries to other firewalls
1227in a
1228.Xr pfsync 4
1229failover cluster.
1230By default the hostid is set to a pseudo-random value, however it may be
1231desirable to manually configure it, for example to more easily identify the
1232source of state table entries.
1233The hostid may be specified in either decimal or hexadecimal.
1234.It Ic set Cm limit Ar limit-item number
1235Sets hard limits on the memory pools used by the packet filter.
1236See
1237.Xr pool 9
1238for an explanation of memory pools.
1239.Pp
1240Limits can be set on the following:
1241.Bl -tag -width pktdelay_pkts
1242.It Cm states
1243Set the maximum number of entries in the memory pool used by state table
1244entries (those generated by
1245.Ic pass
1246rules which do not specify
1247.Cm no state ) .
1248The default is 100000.
1249.It Cm src-nodes
1250Set the maximum number of entries in the memory pool used for tracking
1251source IP addresses (generated by the
1252.Cm sticky-address
1253and
1254.Cm src.track
1255options).
1256The default is 10000.
1257.It Cm frags
1258Set the maximum number of entries in the memory pool used for fragment
1259reassembly.
1260The maximum may not exceed, and should be well below,
1261the maximum number of mbuf clusters
1262.Pq sysctl kern.maxclusters
1263in the system.
1264The default is NMBCLUSTERS/32.
1265.Dv NMBCLUSTERS
1266defines the total number of packets which can exist in-system at any one time.
1267Refer to
1268.In machine/param.h
1269for the platform-specific value.
1270.It Cm tables
1271Set the number of tables that can exist.
1272The default is 1000.
1273.It Cm table-entries
1274Set the number of addresses that can be stored in tables.
1275The default is 200000, or 100000 on machines with
1276less than 100MB of physical memory.
1277.It Cm pktdelay_pkts
1278Set the maximum number of packets that can be held in the delay queue.
1279The default is 10000.
1280.It Cm anchors
1281Set the number of anchors that can exist.
1282The default is 512.
1283.El
1284.Pp
1285Multiple limits can be combined on a single line:
1286.Bd -literal -offset indent
1287set limit { states 20000, frags 2000, src-nodes 2000 }
1288.Ed
1289.It Ic set Cm loginterface Ar interface | Cm none
1290Enable collection of packet and byte count statistics for the given
1291interface or interface group.
1292These statistics can be viewed using:
1293.Pp
1294.Dl # pfctl -s info
1295.Pp
1296In this example
1297.Xr pf 4
1298collects statistics on the interface named dc0:
1299.Pp
1300.Dl set loginterface dc0
1301.Pp
1302One can disable the loginterface using:
1303.Pp
1304.Dl set loginterface none
1305.Pp
1306The default value is
1307.Cm none .
1308.It Ic set Cm optimization Ar environment
1309Optimize state timeouts for one of the following network environments:
1310.Pp
1311.Bl -tag -width Ds -compact
1312.It Cm aggressive
1313Aggressively expire connections.
1314This can greatly reduce the memory usage of the firewall at the cost of
1315dropping idle connections early.
1316.It Cm conservative
1317Extremely conservative settings.
1318Avoid dropping legitimate connections at the
1319expense of greater memory utilization (possibly much greater on a busy
1320network) and slightly increased processor utilization.
1321.It Cm high-latency
1322A high-latency environment (such as a satellite connection).
1323.It Cm normal
1324A normal network environment.
1325Suitable for almost all networks.
1326.It Cm satellite
1327Alias for
1328.Cm high-latency .
1329.El
1330.Pp
1331The default value is
1332.Cm normal .
1333.It Ic set Cm reassemble yes | no Op Cm no-df
1334The
1335.Cm reassemble
1336option is used to enable or disable the reassembly of fragmented packets,
1337and can be set to
1338.Cm yes
1339(the default) or
1340.Cm no .
1341If
1342.Cm no-df
1343is also specified, fragments with the
1344.Dq dont-fragment
1345bit set are reassembled too,
1346instead of being dropped;
1347the reassembled packet will have the
1348.Dq dont-fragment
1349bit cleared.
1350The default value is
1351.Cm yes .
1352.It Ic set Cm ruleset-optimization Ar level
1353.Bl -tag -width profile -compact
1354.It Cm basic
1355Enable basic ruleset optimization.
1356This is the default behaviour.
1357Basic ruleset optimization does four things to improve the
1358performance of ruleset evaluations:
1359.Pp
1360.Bl -enum -compact
1361.It
1362remove duplicate rules
1363.It
1364remove rules that are a subset of another rule
1365.It
1366combine multiple rules into a table when advantageous
1367.It
1368reorder the rules to improve evaluation performance
1369.El
1370.Pp
1371.It Cm none
1372Disable the ruleset optimizer.
1373.It Cm profile
1374Uses the currently loaded ruleset as a feedback profile to tailor the
1375ordering of
1376.Cm quick
1377rules to actual network traffic.
1378.El
1379.Pp
1380It is important to note that the ruleset optimizer will modify the ruleset
1381to improve performance.
1382A side effect of the ruleset modification is that per-rule accounting
1383statistics will have different meanings than before.
1384If per-rule accounting is important for billing purposes or whatnot,
1385either the ruleset optimizer should not be used or a label field should
1386be added to all of the accounting rules to act as optimization barriers.
1387.Pp
1388Optimization can also be set as a command-line argument to
1389.Xr pfctl 8 ,
1390overriding the settings in
1391.Nm .
1392.It Ic set Cm skip on Ar ifspec
1393List interfaces for which packets should not be filtered.
1394Packets passing in or out on such interfaces are passed as if pf was
1395disabled, i.e. pf does not process them in any way.
1396This can be useful on loopback and other virtual interfaces, when
1397packet filtering is not desired and can have unexpected effects.
1398PF filters traffic on all interfaces by default.
1399.It Ic set Cm state-defaults Ar state-option , ...
1400The
1401.Cm state-defaults
1402option sets the state options for states created from rules
1403without an explicit
1404.Cm keep state .
1405For example:
1406.Pp
1407.Dl set state-defaults pflow, no-sync
1408.It Ic set Cm state-policy if-bound | floating
1409The
1410.Cm state-policy
1411option sets the default behaviour for states:
1412.Pp
1413.Bl -tag -width if-bound -compact
1414.It Cm if-bound
1415States are bound to an interface.
1416.It Cm floating
1417States can match packets on any interfaces (the default).
1418.El
1419.It Ic set Cm syncookies never | always | adaptive
1420When
1421.Cm syncookies
1422are active, pf will answer each and every incoming TCP SYN with a
1423syncookie SYNACK, without allocating any resources.
1424Upon reception of the client's ACK in response to the syncookie
1425SYNACK, pf will evaluate the ruleset and create state if the ruleset
1426permits it, complete the three way handshake with the target host,
1427and continue the connection with synproxy in place.
1428This allows pf to be resilient against large synflood attacks,
1429which could otherwise exhaust the state table.
1430Due to the blind answers to each and every SYN,
1431syncookies share the caveats of synproxy:
1432seemingly accepting connections that will be dropped later on.
1433.Pp
1434.Bl -tag -width adaptive -compact
1435.It Cm never
1436pf will never send syncookie SYNACKs (the default).
1437.It Cm always
1438pf will always send syncookie SYNACKs.
1439.It Cm adaptive
1440pf will enable syncookie mode when a given percentage of the state table
1441is used up by half-open TCP connections, such as those that saw the initial
1442SYN but didn't finish the three way handshake.
1443The thresholds for entering and leaving syncookie mode can be specified using:
1444.Bd -literal -offset indent
1445set syncookies adaptive (start 25%, end 12%)
1446.Ed
1447.El
1448.It Ic set Cm timeout Ar variable value
1449.Bl -tag -width "src.track" -compact
1450.It Cm frag
1451Seconds before an unassembled fragment is expired (60 by default).
1452.It Cm interval
1453Interval between purging expired states and fragments (10 seconds by default).
1454.It Cm src.track
1455Length of time to retain a source tracking entry after the last state
1456expires (0 by default, which means there is no global limit.
1457The value is defined by the rule which creates the state.).
1458.El
1459.Pp
1460When a packet matches a stateful connection, the seconds to live for the
1461connection will be updated to that of the
1462protocol and modifier
1463which corresponds to the connection state.
1464Each packet which matches this state will reset the TTL.
1465Tuning these values may improve the performance of the
1466firewall at the risk of dropping valid idle connections.
1467.Pp
1468.Bl -tag -width Ds -compact
1469.It Cm tcp.closed Pq 90 seconds by default
1470The state after one endpoint sends an RST.
1471.It Cm tcp.closing Pq 900 seconds by default
1472The state after the first FIN has been sent.
1473.It Cm tcp.established Pq 24 hours by default
1474The fully established state.
1475.It Cm tcp.finwait Pq 45 seconds by default
1476The state after both FINs have been exchanged and the connection is closed.
1477Some hosts (notably web servers on Solaris) send TCP packets even after closing
1478the connection.
1479Increasing
1480.Cm tcp.finwait
1481(and possibly
1482.Cm tcp.closing )
1483can prevent blocking of such packets.
1484.It Cm tcp.first Pq 120 seconds by default
1485The state after the first packet.
1486.It Cm tcp.opening Pq 30 seconds by default
1487The state after the second packet but before both endpoints have
1488acknowledged the connection.
1489.El
1490.Pp
1491ICMP and UDP are handled in a fashion similar to TCP, but with a much more
1492limited set of states:
1493.Pp
1494.Bl -tag -width Ds -compact
1495.It Cm icmp.error Pq 10 seconds by default
1496The state after an ICMP error came back in response to an ICMP packet.
1497.It Cm icmp.first Pq 20 seconds by default
1498The state after the first packet.
1499.It Cm udp.first Pq 60 seconds by default
1500The state after the first packet.
1501.It Cm udp.multiple Pq 60 seconds by default
1502The state if both hosts have sent packets.
1503.It Cm udp.single Pq 30 seconds by default
1504The state if the source host sends more than one packet but the destination
1505host has never sent one back.
1506.El
1507.Pp
1508Other protocols are handled similarly to UDP:
1509.Pp
1510.Bl -tag -width xxxx -compact
1511.It Cm other.first Pq 60 seconds by default
1512.It Cm other.multiple Pq 60 seconds by default
1513.It Cm other.single Pq 30 seconds by default
1514.El
1515.Pp
1516Timeout values can be reduced adaptively as the number of state table
1517entries grows.
1518.Pp
1519.Bl -tag -width Ds -compact
1520.It Cm adaptive.start Pq 60000 states by default
1521When the number of state entries exceeds this value, adaptive scaling
1522begins.
1523All timeout values are scaled linearly with factor
1524(adaptive.end \- number of states) / (adaptive.end \- adaptive.start).
1525.It Cm adaptive.end Pq 120000 states by default
1526When reaching this number of state entries, all timeout values become
1527zero, effectively purging all state entries immediately.
1528This value is used to define the scale factor; it should not actually
1529be reached (set a lower state limit, see below).
1530.El
1531.Pp
1532Adaptive timeouts are enabled by default, with an adaptive.start value
1533equal to 60% of the state limit, and an adaptive.end value equal to
1534120% of the state limit.
1535They can be disabled by setting both adaptive.start and adaptive.end to 0.
1536.Pp
1537The adaptive timeout values can be defined both globally and for each rule.
1538When used on a per-rule basis, the values relate to the number of
1539states created by the rule, otherwise to the total number of
1540states.
1541.Pp
1542For example:
1543.Bd -literal -offset indent
1544set timeout tcp.first 120
1545set timeout tcp.established 86400
1546set timeout { adaptive.start 60000, adaptive.end 120000 }
1547set limit states 100000
1548.Ed
1549.Pp
1550With 90000 state table entries, the timeout values are scaled to 50%
1551(tcp.first 60, tcp.established 43200).
1552.El
1553.Pp
1554.Dq pfctl -F Reset
1555restores default values for the following options: debug, all limit options,
1556loginterface, reassemble, skip, syncookies, all timeouts.
1557.Sh QUEUEING
1558Packets can be assigned to queues for the purpose of bandwidth
1559control.
1560At least one declaration is required to configure queues, and later
1561any packet filtering rule can reference the defined queues by name.
1562When filtering, the last referenced
1563.Ar queue
1564name is where any passed packets will be queued, while for
1565blocked packets it specifies where any resulting ICMP or TCP RST
1566packets should be queued.
1567If the referenced queue does not exist on the outgoing interface,
1568the default queue for that interface is used.
1569Queues attached to an interface build a tree,
1570thus each queue can have further child queues.
1571Only leaf queues, i.e. queues without children, can be used to assign
1572packets to.
1573The root queue must specifically reference an interface, all other queues
1574pick up the interfaces they should be created on from their parent queues.
1575.Pp
1576In the following example, a queue named std is created on the interface em0,
1577with 3 child queues ssh, mail and http:
1578.Bd -literal -offset indent
1579queue std on em0 bandwidth 100M
1580queue ssh parent std bandwidth 10M
1581queue mail parent std bandwidth 10M
1582queue http parent std bandwidth 80M default
1583.Ed
1584.Pp
1585The specified bandwidth is the target bandwidth, every queue can receive
1586more bandwidth as long as the parent still has some available.
1587The maximum bandwidth that should be assigned to a given queue can be limited
1588using the
1589.Cm max
1590keyword.
1591If a limitation isn't imposed on the root queue, borrowing can result in
1592saturating the bandwidth of the outgoing interface.
1593Similarly, a minimum (reserved) bandwidth can be specified:
1594.Pp
1595.Dl queue ssh parent std bandwidth 10M min 5M max 25M
1596.Pp
1597For each of these 3 bandwidth specifications an additional burst bandwidth and
1598time can be specified:
1599.Pp
1600.Dl queue ssh parent std bandwidth 10M burst 90M for 100ms
1601.Pp
1602All
1603.Cm bandwidth
1604values are specified as bits per second or using the suffixes
1605.Cm K ,
1606.Cm M ,
1607and
1608.Cm G
1609to represent kilobits, megabits, and gigabits per second, respectively.
1610The value must not exceed the interface bandwidth.
1611.Pp
1612If multiple connections are assigned the same queue, they're not guaranteed
1613to share the queue bandwidth fairly.
1614An alternative flow queue manager can be used to achieve fair sharing by
1615indicating how many simultaneous states are expected with a
1616.Cm flows
1617option, unless a minimum bandwidth has been specified as well.
1618.Pp
1619When packets are classified by the stateful inspection engine, a flow
1620identifier is assigned to all packets belonging to the state,
1621thus limiting the number of individual flows that can be recognized
1622by the resolution of a flow identifier.
1623The current implementation is able to classify traffic into 32767 distinct
1624flows.
1625However, efficient fair sharing is observed even with a much smaller number
1626of flows.
1627For example on a 10Mbit/s DSL or a cable modem uplink, the following simple
1628configuration can be used:
1629.Bd -literal -offset 4n
1630queue outq on em0 bandwidth 9M max 9M flows 1024 qlimit 1024 \e
1631      default
1632.Ed
1633.Pp
1634It's important to specify the upper bound within 90-95% of the expected
1635bandwidth and raise the default queue limit.
1636.Pp
1637If a
1638.Cm flows
1639option appears without a
1640.Cm bandwidth
1641specification, the flow queue manager is selected as the queueing discipline
1642for the corresponding interface acting as a default queue for all outgoing
1643packets.
1644In such a scenario, a queueing hierarchy is not supported.
1645.Pp
1646In addition to the bandwidth and flow specifications, queues support the
1647following options:
1648.Bl -tag -width xxxx
1649.It Cm default
1650Packets not matched by another queue are assigned to this queue.
1651Exactly one default queue per interface is required.
1652.It Cm on Ar interface
1653Specifies the interface the queue operates on.
1654If not given, it operates on all matching interfaces.
1655.It Cm parent Ar name
1656Defines which parent queue the queue should be attached to.
1657Mandatory for all queues except root queues.
1658The parent queue must exist.
1659.It Cm quantum Ar size
1660Specifies the quantum of service for the flow queue manager.
1661The lower the quantum size the more advantage is given to streams of smaller
1662packets at the expense of bulk transfers.
1663The default value is set to the configured Maximum Transmission Unit (MTU)
1664of the specified interface.
1665.It Cm qlimit Ar limit
1666The maximum number of packets held in the queue.
1667The default is 50.
1668.El
1669.Pp
1670Packets can be assigned to queues based on filter rules by using the
1671.Cm queue
1672keyword.
1673Normally only one
1674.Ar queue
1675is specified; when a second one is specified it will instead be used for
1676packets which have a TOS of
1677.Cm lowdelay
1678and for TCP ACKs with no data payload.
1679.Pp
1680To continue the previous example, the examples below would specify the
1681four referenced
1682queues, plus a few child queues.
1683Interactive
1684.Xr ssh 1
1685sessions get a queue with a minimum bandwidth;
1686.Xr scp 1
1687and
1688.Xr sftp 1
1689bulk transfers go to a separate queue.
1690The queues are then referenced by filtering rules.
1691.Bd -literal -offset 4n
1692queue rootq on em0 bandwidth 100M max 100M
1693queue http parent rootq bandwidth 60M burst 90M for 100ms
1694queue  developers parent http bandwidth 45M
1695queue  employees parent http bandwidth 15M
1696queue mail parent rootq bandwidth 10M
1697queue ssh parent rootq bandwidth 20M
1698queue  ssh_interactive parent ssh bandwidth 10M min 5M
1699queue  ssh_bulk parent ssh bandwidth 10M
1700queue std parent rootq bandwidth 20M default
1701
1702block return out on em0 inet all set queue std
1703pass out on em0 inet proto tcp from $developerhosts to any port 80 \e
1704      set queue developers
1705pass out on em0 inet proto tcp from $employeehosts to any port 80 \e
1706      set queue employees
1707pass out on em0 inet proto tcp from any to any port 22 \e
1708      set queue(ssh_bulk, ssh_interactive)
1709pass out on em0 inet proto tcp from any to any port 25 \e
1710      set queue mail
1711.Ed
1712.Sh TABLES
1713Tables are named structures which can hold a collection of addresses and
1714networks.
1715Lookups against tables in
1716.Xr pf 4
1717are relatively fast, making a single rule with tables much more efficient,
1718in terms of
1719processor usage and memory consumption, than a large number of rules which
1720differ only in IP address (either created explicitly or automatically by rule
1721expansion).
1722.Pp
1723Tables can be used as the source or destination of filter
1724or translation rules.
1725They can also be used for the redirect address of
1726.Cm nat-to
1727and
1728.Cm rdr-to
1729and in the routing options of filter rules, but not for
1730.Cm bitmask
1731pools.
1732.Pp
1733Tables can be defined with any of the following
1734.Xr pfctl 8
1735mechanisms.
1736As with macros, reserved words may not be used as table names.
1737.Bl -tag -width "manually"
1738.It manually
1739Persistent tables can be manually created with the
1740.Cm add
1741or
1742.Cm replace
1743option of
1744.Xr pfctl 8 ,
1745before or after the ruleset has been loaded.
1746.It Nm
1747Table definitions can be placed directly in this file and loaded at the
1748same time as other rules are loaded, atomically.
1749Table definitions inside
1750.Nm
1751use the
1752.Ic table
1753statement, and are especially useful to define non-persistent tables.
1754The contents of a pre-existing table defined without a list of addresses
1755to initialize it is not altered when
1756.Nm
1757is loaded.
1758A table initialized with the empty list,
1759.Li { } ,
1760will be cleared on load.
1761.El
1762.Pp
1763Tables may be defined with the following attributes:
1764.Bl -tag -width counters
1765.It Cm const
1766The
1767.Cm const
1768flag prevents the user from altering the contents of the table once it
1769has been created.
1770Without that flag,
1771.Xr pfctl 8
1772can be used to add or remove addresses from the table at any time, even
1773when running with
1774.Xr securelevel 7
1775= 2.
1776.It Cm counters
1777The
1778.Cm counters
1779flag enables per-address packet and byte counters, which can be displayed with
1780.Xr pfctl 8 .
1781.It Cm persist
1782The
1783.Cm persist
1784flag forces the kernel to keep the table even when no rules refer to it.
1785If the flag is not set, the kernel will automatically remove the table
1786when the last rule referring to it is flushed.
1787.El
1788.Pp
1789This example creates a table called
1790.Dq private ,
1791to hold RFC 1918 private network blocks,
1792and a table called
1793.Dq badhosts ,
1794which is initially empty.
1795A filter rule is set up to block all traffic coming from addresses listed in
1796either table:
1797.Bd -literal -offset indent
1798table <private> const { 10/8, 172.16/12, 192.168/16 }
1799table <badhosts> persist
1800block on fxp0 from { <private>, <badhosts> } to any
1801.Ed
1802.Pp
1803The private table cannot have its contents changed and the badhosts table
1804will exist even when no active filter rules reference it.
1805Addresses may later be added to the badhosts table, so that traffic from
1806these hosts can be blocked by using the following:
1807.Pp
1808.Dl # pfctl -t badhosts -Tadd 204.92.77.111
1809.Pp
1810A table can also be initialized with an address list specified in one or more
1811external files, using the following syntax:
1812.Bd -literal -offset indent
1813table <spam> persist file "/etc/spammers" file "/etc/openrelays"
1814block on fxp0 from <spam> to any
1815.Ed
1816.Pp
1817The files
1818.Pa /etc/spammers
1819and
1820.Pa /etc/openrelays
1821list IP addresses, one per line.
1822Any lines beginning with a
1823.Sq #
1824are treated as comments and ignored.
1825In addition to being specified by IP address, hosts may also be
1826specified by their hostname.
1827When the resolver is called to add a hostname to a table,
1828.Em all
1829resulting IPv4 and IPv6 addresses are placed into the table.
1830IP addresses can also be entered in a table by specifying a valid interface
1831name, a valid interface group, or the
1832.Cm self
1833keyword, in which case all addresses assigned to the interface(s) will be
1834added to the table.
1835.Sh ANCHORS
1836Besides the main ruleset,
1837.Nm
1838can specify anchor attachment points.
1839An anchor is a container that can hold rules,
1840address tables, and other anchors.
1841When evaluation of the main ruleset reaches an
1842.Ic anchor
1843rule,
1844.Xr pf 4
1845will proceed to evaluate all rules specified in that anchor.
1846.Pp
1847The following example blocks all packets on the external interface by default,
1848then evaluates all rules in the anchor named "spam",
1849and finally passes all outgoing connections and
1850incoming connections to port 25:
1851.Bd -literal -offset indent
1852ext_if = "kue0"
1853block on $ext_if all
1854anchor spam
1855pass out on $ext_if all
1856pass in on $ext_if proto tcp from any to $ext_if port smtp
1857.Ed
1858.Pp
1859Anchors can be manipulated through
1860.Xr pfctl 8
1861without reloading the main ruleset or other anchors.
1862This loads a single rule into the anchor,
1863which blocks all packets from a specific address:
1864.Bd -literal -offset indent
1865# echo "block in quick from 1.2.3.4 to any" | pfctl -a spam -f -
1866.Ed
1867.Pp
1868The anchor can also be populated by adding a
1869.Ic load anchor
1870rule after the anchor rule.
1871When
1872.Xr pfctl 8
1873loads
1874.Nm ,
1875it will also load all the rules from the file
1876.Pa /etc/pf-spam.conf
1877into the anchor.
1878.Bd -literal -offset indent
1879anchor spam
1880load anchor spam from "/etc/pf-spam.conf"
1881.Ed
1882.Pp
1883An anchor rule can also contain a filter ruleset
1884in a brace-delimited block.
1885In that case, no separate loading of rules into the anchor
1886is required.
1887Brace delimited blocks may contain rules or other brace-delimited blocks.
1888When an anchor is populated this way, the anchor name becomes optional.
1889Since the parser specification for anchor names is a string,
1890double quote characters
1891.Pq Sq \&"
1892should be placed around the anchor name.
1893.Bd -literal -offset indent
1894anchor "external" on egress {
1895	block
1896	anchor out {
1897		pass proto tcp from any to port { 25, 80, 443 }
1898	}
1899	pass in proto tcp to any port 22
1900}
1901.Ed
1902.Pp
1903Anchor rules can also specify packet filtering parameters
1904using the same syntax as filter rules.
1905When parameters are used,
1906the anchor rule is only evaluated for matching packets.
1907This allows conditional evaluation of anchors, like:
1908.Bd -literal -offset indent
1909block on $ext_if all
1910anchor spam proto tcp from any to any port smtp
1911pass out on $ext_if all
1912pass in on $ext_if proto tcp from any to $ext_if port smtp
1913.Ed
1914.Pp
1915The rules inside anchor "spam" are only evaluated
1916for TCP packets with destination port 25.
1917Hence, the following
1918will only block connections from 1.2.3.4 to port 25:
1919.Bd -literal -offset indent
1920# echo "block in quick from 1.2.3.4 to any" | pfctl -a spam -f -
1921.Ed
1922.Pp
1923Matching filter and translation rules marked with the
1924.Cm quick
1925option are final and abort the evaluation of the rules in other
1926anchors and the main ruleset.
1927If the anchor itself is marked with the
1928.Cm quick
1929option,
1930ruleset evaluation will terminate when the anchor is exited if the packet is
1931matched by any rule within the anchor.
1932.Pp
1933An anchor references other anchor attachment points
1934using the following syntax:
1935.Bl -tag -width xxxx
1936.It Ic anchor Ar name
1937Evaluates the filter rules in the specified anchor.
1938.El
1939.Pp
1940An anchor has a name which specifies the path where
1941.Xr pfctl 8
1942can be used to access the anchor to perform operations on it, such as
1943attaching child anchors to it or loading rules into it.
1944Anchors may be nested, with components separated by
1945.Sq /
1946characters, similar to how file system hierarchies are laid out.
1947The main ruleset is actually the default anchor, so filter and
1948translation rules, for example, may also be contained in any anchor.
1949.Pp
1950Anchor rules are evaluated relative to the anchor in which they are contained.
1951For example,
1952all anchor rules specified in the main ruleset will reference
1953anchor attachment points underneath the main ruleset,
1954and anchor rules specified in a file loaded from a
1955.Ic load anchor
1956rule will be attached under that anchor point.
1957.Pp
1958Anchors may end with the asterisk
1959.Pq Sq *
1960character, which signifies that all anchors attached at that point
1961should be evaluated in the alphabetical ordering of their anchor name.
1962For example,
1963the following
1964will evaluate each rule in each anchor attached to the "spam" anchor:
1965.Bd -literal -offset indent
1966anchor "spam/*"
1967.Ed
1968.Pp
1969Note that it will only evaluate anchors that are directly attached to the
1970"spam" anchor, and will not descend to evaluate anchors recursively.
1971.Pp
1972Since anchors are evaluated relative to the anchor in which they are
1973contained, there is a mechanism for accessing the parent and ancestor
1974anchors of a given anchor.
1975Similar to file system path name resolution, if the sequence
1976.Sq ..
1977appears as an anchor path component, the parent anchor of the current
1978anchor in the path evaluation at that point will become the new current
1979anchor.
1980As an example, consider the following:
1981.Bd -literal -offset indent
1982# printf 'anchor "spam/allowed"\en' | pfctl -f -
1983# printf 'anchor "../banned"\enpass\en' | pfctl -a spam/allowed -f -
1984.Ed
1985.Pp
1986Evaluation of the main ruleset will lead into the
1987spam/allowed anchor, which will evaluate the rules in the
1988spam/banned anchor, if any, before finally evaluating the
1989.Ic pass
1990rule.
1991.Sh STATEFUL FILTERING
1992.Xr pf 4
1993filters packets statefully,
1994which has several advantages.
1995For TCP connections, comparing a packet to a state involves checking
1996its sequence numbers, as well as TCP timestamps if a rule using the
1997.Cm reassemble tcp
1998parameter applies to the connection.
1999If these values are outside the narrow windows of expected
2000values, the packet is dropped.
2001This prevents spoofing attacks, such as when an attacker sends packets with
2002a fake source address/port but does not know the connection's sequence
2003numbers.
2004Similarly,
2005.Xr pf 4
2006knows how to match ICMP replies to states.
2007For example,
2008to allow echo requests (such as those created by
2009.Xr ping 8 )
2010out statefully and match incoming echo replies correctly to states:
2011.Pp
2012.Dl pass out inet proto icmp all icmp-type echoreq
2013.Pp
2014Also, looking up states is usually faster than evaluating rules.
2015If there are 50 rules, all of them are evaluated sequentially in O(n).
2016Even with 50000 states, only 16 comparisons are needed to match a
2017state, since states are stored in a binary search tree that allows
2018searches in O(log2\~n).
2019.Pp
2020Furthermore, correct handling of ICMP error messages is critical to
2021many protocols, particularly TCP.
2022.Xr pf 4
2023matches ICMP error messages to the correct connection, checks them against
2024connection parameters, and passes them if appropriate.
2025For example if an ICMP source quench message referring to a stateful TCP
2026connection arrives, it will be matched to the state and get passed.
2027.Pp
2028Finally, state tracking is required for
2029.Cm nat-to
2030and
2031.Cm rdr-to
2032options, in order to track address and port translations and reverse the
2033translation on returning packets.
2034.Pp
2035.Xr pf 4
2036will also create state for other protocols which are effectively stateless by
2037nature.
2038UDP packets are matched to states using only host addresses and ports,
2039and other protocols are matched to states using only the host addresses.
2040.Pp
2041If stateless filtering of individual packets is desired,
2042the
2043.Cm no state
2044keyword can be used to specify that state will not be created
2045if this is the last matching rule.
2046Note that packets which match neither block nor pass rules,
2047and thus are passed by default,
2048are effectively passed as if
2049.Cm no state
2050had been specified.
2051.Pp
2052A number of parameters can also be set to affect how
2053.Xr pf 4
2054handles state tracking,
2055as detailed below.
2056.Ss State Modulation
2057Much of the security derived from TCP is attributable to how well the
2058initial sequence numbers (ISNs) are chosen.
2059Some popular stack implementations choose
2060.Em very
2061poor ISNs and thus are normally susceptible to ISN prediction exploits.
2062By applying a
2063.Cm modulate state
2064rule to a TCP connection,
2065.Xr pf 4
2066will create a high quality random sequence number for each connection
2067endpoint.
2068.Pp
2069The
2070.Cm modulate state
2071directive implicitly keeps state on the rule and is
2072only applicable to TCP connections.
2073.Pp
2074For instance:
2075.Bd -literal -offset indent
2076block all
2077pass out proto tcp from any to any modulate state
2078pass in  proto tcp from any to any port 25 flags S/SFRA \e
2079      modulate state
2080.Ed
2081.Pp
2082Note that modulated connections will not recover when the state table
2083is lost (firewall reboot, flushing the state table, etc.).
2084.Xr pf 4
2085will not be able to infer a connection again after the state table flushes
2086the connection's modulator.
2087When the state is lost, the connection may be left dangling until the
2088respective endpoints time out the connection.
2089It is possible on a fast local network for the endpoints to start an ACK
2090storm while trying to resynchronize after the loss of the modulator.
2091The default
2092.Cm flags
2093settings (or a more strict equivalent) should be used on
2094.Cm modulate state
2095rules to prevent ACK storms.
2096.Pp
2097Note that alternative methods are available
2098to prevent loss of the state table
2099and allow for firewall failover.
2100See
2101.Xr carp 4
2102and
2103.Xr pfsync 4
2104for further information.
2105.Ss SYN Proxy
2106By default,
2107.Xr pf 4
2108passes packets that are part of a
2109TCP handshake between the endpoints.
2110The
2111.Cm synproxy state
2112option can be used to cause
2113.Xr pf 4
2114itself to complete the handshake with the active endpoint, perform a handshake
2115with the passive endpoint, and then forward packets between the endpoints.
2116.Pp
2117No packets are sent to the passive endpoint before the active endpoint has
2118completed the handshake, hence so-called SYN floods with spoofed source
2119addresses will not reach the passive endpoint, as the sender can't complete the
2120handshake.
2121.Pp
2122The proxy is transparent to both endpoints; they each see a single
2123connection from/to the other endpoint.
2124.Xr pf 4
2125chooses random initial sequence numbers for both handshakes.
2126Once the handshakes are completed, the sequence number modulators
2127(see previous section) are used to translate further packets of the
2128connection.
2129.Cm synproxy state
2130includes
2131.Cm modulate state .
2132.Pp
2133Rules with
2134.Cm synproxy state
2135will not work if
2136.Xr pf 4
2137operates on a
2138.Xr bridge 4 .
2139Also they act on incoming SYN packets only.
2140.Pp
2141Example:
2142.Bd -literal -offset indent
2143pass in proto tcp from any to any port www synproxy state
2144.Ed
2145.Ss Stateful Tracking Options
2146A number of options related to stateful tracking can be applied on a
2147per-rule basis.
2148One of
2149.Cm keep state ,
2150.Cm modulate state ,
2151or
2152.Cm synproxy state
2153must be specified explicitly to apply these options to a rule.
2154.Pp
2155.Bl -tag -width xxxx -compact
2156.It Cm floating
2157States can match packets on any interfaces
2158(the opposite of
2159.Cm if-bound ) .
2160This is the default.
2161.It Cm if-bound
2162States are bound to an interface
2163(the opposite of
2164.Cm floating ) .
2165.It Cm max Ar number
2166Limits the number of concurrent states the rule may create.
2167When this limit is reached, further packets that would create
2168state are dropped until existing states time out.
2169.It Cm no-sync
2170Prevent state changes for states created by this rule from appearing on the
2171.Xr pfsync 4
2172interface.
2173.It Cm pflow
2174States created by this rule are exported on the
2175.Xr pflow 4
2176interface.
2177.It Cm sloppy
2178For TCP, uses a sloppy connection tracker that does not check sequence
2179numbers at all, which makes insertion and ICMP teardown attacks way
2180easier.
2181This is intended to be used in situations where one does not see all
2182packets of a connection, e.g. in asymmetric routing situations.
2183It cannot be used with
2184.Cm modulate state
2185or
2186.Cm synproxy state .
2187For ICMP, this option allows states to be created from replies,
2188not just requests.
2189.It Ar timeout seconds
2190Changes the
2191.Ar timeout
2192values used for states created by this rule.
2193For a list of all valid
2194.Ar timeout
2195names, see
2196.Sx OPTIONS
2197above.
2198.El
2199.Pp
2200Multiple options can be specified, separated by commas:
2201.Bd -literal -offset indent
2202pass in proto tcp from any to any \e
2203      port www keep state \e
2204      (max 100, source-track rule, max-src-nodes 75, \e
2205      max-src-states 3, tcp.established 60, tcp.closing 5)
2206.Ed
2207.Pp
2208When the
2209.Cm source-track
2210keyword is specified, the number of states per source IP is tracked.
2211.Pp
2212.Bl -tag -width xxxx -compact
2213.It Cm source-track global
2214The number of states created by all rules that use this option is limited.
2215Each rule can specify different
2216.Cm max-src-nodes
2217and
2218.Cm max-src-states
2219options, however state entries created by any participating rule count towards
2220each individual rule's limits.
2221.It Cm source-track rule
2222The maximum number of states created by this rule is limited by the rule's
2223.Cm max-src-nodes
2224and
2225.Cm max-src-states
2226options.
2227Only state entries created by this particular rule count toward the rule's
2228limits.
2229.El
2230.Pp
2231The following limits can be set:
2232.Pp
2233.Bl -tag -width xxxx -compact
2234.It Cm max-src-nodes Ar number
2235Limits the maximum number of source addresses which can simultaneously
2236have state table entries.
2237.It Cm max-src-states Ar number
2238Limits the maximum number of simultaneous state entries that a single
2239source address can create with this rule.
2240.El
2241.Pp
2242For stateful TCP connections, limits on established connections (connections
2243which have completed the TCP 3-way handshake) can also be enforced
2244per source IP.
2245.Pp
2246.Bl -tag -width xxxx -compact
2247.It Cm max-src-conn Ar number
2248Limits the maximum number of simultaneous TCP connections which have
2249completed the 3-way handshake that a single host can make.
2250.It Cm max-src-conn-rate Ar number Ns / Ns Ar seconds
2251Limit the rate of new connections over a time interval.
2252The connection rate is an approximation calculated as a moving average.
2253.El
2254.Pp
2255When one of these limits is reached, further packets that would create
2256state are dropped until existing states time out.
2257.Pp
2258Because the 3-way handshake ensures that the source address is not being
2259spoofed, more aggressive action can be taken based on these limits.
2260With the
2261.Cm overload Pf < Ar table Ns >
2262state option, source IP addresses which hit either of the limits on
2263established connections will be added to the named
2264.Ar table .
2265This table can be used in the ruleset to block further activity from
2266the offending host, redirect it to a tarpit process, or restrict its
2267bandwidth.
2268.Pp
2269The optional
2270.Cm flush
2271keyword kills all states created by the matching rule which originate
2272from the host which exceeds these limits.
2273The
2274.Cm global
2275modifier to the
2276.Cm flush
2277command kills all states originating from the
2278offending host, regardless of which rule created the state.
2279.Pp
2280For example, the following rules will protect the webserver against
2281hosts making more than 100 connections in 10 seconds.
2282Any host which connects faster than this rate will have its address added
2283to the <bad_hosts> table and have all states originating from it flushed.
2284Any new packets arriving from this host will be dropped unconditionally
2285by the block rule.
2286.Bd -literal -offset indent
2287block quick from <bad_hosts>
2288pass in on $ext_if proto tcp to $webserver port www keep state \e
2289      (max-src-conn-rate 100/10, overload <bad_hosts> flush global)
2290.Ed
2291.Sh TRAFFIC NORMALISATION
2292Traffic normalisation is a broad umbrella term
2293for aspects of the packet filter which deal with
2294verifying packets, packet fragments, spoof traffic,
2295and other irregularities.
2296.Ss Scrub
2297Scrub involves sanitising packet content in such a way
2298that there are no ambiguities in packet interpretation on the receiving side.
2299It is invoked with the
2300.Cm scrub
2301option, added to regular rules.
2302.Pp
2303Parameters are specified enclosed in parentheses.
2304At least one of the following parameters must be specified:
2305.Bl -tag -width xxxx
2306.It Cm max-mss Ar number
2307Reduces the maximum segment size (MSS)
2308on TCP SYN packets to be no greater than
2309.Ar number .
2310This is sometimes required in scenarios where the two endpoints
2311of a TCP connection are not able to carry similar sized packets
2312and the resulting mismatch can lead to packet fragmentation or loss.
2313Note that setting the MSS this way can have undesirable effects,
2314such as interfering with the OS detection features of
2315.Xr pf 4 .
2316.It Cm min-ttl Ar number
2317Enforces a minimum TTL for matching IP packets.
2318.It Cm no-df
2319Clears the
2320.Dq dont-fragment
2321bit from a matching IPv4 packet.
2322Some operating systems have NFS implementations
2323which are known to generate fragmented packets with the
2324.Dq dont-fragment
2325bit set.
2326.Xr pf 4
2327will drop such fragmented
2328.Dq dont-fragment
2329packets unless
2330.Cm no-df
2331is specified.
2332.Pp
2333Unfortunately some operating systems also generate their
2334.Dq dont-fragment
2335packets with a zero IP identification field.
2336Clearing the
2337.Dq dont-fragment
2338bit on packets with a zero IP ID may cause deleterious results if an
2339upstream router later fragments the packet.
2340Using
2341.Cm random-id
2342is recommended in combination with
2343.Cm no-df
2344to ensure unique IP identifiers.
2345.It Cm random-id
2346Replaces the IPv4 identification field with random values to compensate
2347for predictable values generated by many hosts.
2348This option only applies to packets that are not fragmented
2349after the optional fragment reassembly.
2350.It Cm reassemble tcp
2351Statefully normalises TCP connections.
2352.Cm reassemble tcp
2353performs the following normalisations:
2354.Bl -ohang
2355.It TTL
2356Neither side of the connection is allowed to reduce their IP TTL.
2357An attacker may send a packet such that it reaches the firewall, affects
2358the firewall state, and expires before reaching the destination host.
2359.Cm reassemble tcp
2360will raise the TTL of all packets back up to the highest value seen on
2361the connection.
2362.It Timestamp Modulation
2363Modern TCP stacks will send a timestamp on every TCP packet and echo
2364the other endpoint's timestamp back to them.
2365Many operating systems will merely start the timestamp at zero when
2366first booted, and increment it several times a second.
2367The uptime of the host can be deduced by reading the timestamp and multiplying
2368by a constant.
2369Also observing several different timestamps can be used to count hosts
2370behind a NAT device.
2371And spoofing TCP packets into a connection requires knowing or guessing
2372valid timestamps.
2373Timestamps merely need to be monotonically increasing and not derived off a
2374guessable base time.
2375.Cm reassemble tcp
2376will cause
2377.Cm scrub
2378to modulate the TCP timestamps with a random number.
2379.It Extended PAWS Checks
2380There is a problem with TCP on long fat pipes, in that a packet might get
2381delayed for longer than it takes the connection to wrap its 32-bit sequence
2382space.
2383In such an occurrence, the old packet would be indistinguishable from a
2384new packet and would be accepted as such.
2385The solution to this is called PAWS: Protection Against Wrapped Sequence
2386numbers.
2387It protects against it by making sure the timestamp on each packet does
2388not go backwards.
2389.Cm reassemble tcp
2390also makes sure the timestamp on the packet does not go forward more
2391than the RFC allows.
2392By doing this,
2393.Xr pf 4
2394artificially extends the security of TCP sequence numbers by 10 to 18
2395bits when the host uses appropriately randomized timestamps, since a
2396blind attacker would have to guess the timestamp as well.
2397.El
2398.El
2399.Pp
2400For example:
2401.Pp
2402.Dl match in all scrub (no-df random-id max-mss 1440)
2403.Ss Fragment Handling
2404The size of IP datagrams (packets) can be significantly larger than the
2405maximum transmission unit (MTU) of the network.
2406In cases when it is necessary or more efficient to send such large packets,
2407the large packet will be fragmented into many smaller packets that will each
2408fit onto the wire.
2409Unfortunately for a firewalling device, only the first logical fragment will
2410contain the necessary header information for the subprotocol that allows
2411.Xr pf 4
2412to filter on things such as TCP ports or to perform NAT.
2413.Pp
2414One alternative is to filter individual fragments with filter rules.
2415If packet reassembly is turned off, it is passed to the filter.
2416Filter rules with matching IP header parameters decide whether the
2417fragment is passed or blocked, in the same way as complete packets
2418are filtered.
2419Without reassembly, fragments can only be filtered based on IP header
2420fields (source/destination address, protocol), since subprotocol header
2421fields are not available (TCP/UDP port numbers, ICMP code/type).
2422The
2423.Cm fragment
2424option can be used to restrict filter rules to apply only to
2425fragments, but not complete packets.
2426Filter rules without the
2427.Cm fragment
2428option still apply to fragments, if they only specify IP header fields.
2429For instance:
2430.Bd -literal -offset indent
2431pass in proto tcp from any to any port 80
2432.Ed
2433.Pp
2434The rule above never applies to a fragment,
2435even if the fragment is part of a TCP packet with destination port 80,
2436because without reassembly this information
2437is not available for each fragment.
2438This also means that fragments cannot create new or match existing
2439state table entries, which makes stateful filtering and address
2440translation (NAT, redirection) for fragments impossible.
2441.Pp
2442In most cases, the benefits of reassembly outweigh the additional
2443memory cost,
2444so reassembly is on by default.
2445.Pp
2446The memory allocated for fragment caching can be limited using
2447.Xr pfctl 8 .
2448Once this limit is reached, fragments that would have to be cached
2449are dropped until other entries time out.
2450The timeout value can also be adjusted.
2451.Pp
2452When forwarding reassembled IPv6 packets, pf refragments them with
2453the original maximum fragment size.
2454This allows the sender to determine the optimal fragment size by
2455path MTU discovery.
2456.Ss Blocking Spoofed Traffic
2457Spoofing is the faking of IP addresses,
2458typically for malicious purposes.
2459The
2460.Ic antispoof
2461directive expands to a set of filter rules which will block all
2462traffic with a source IP from the network(s) directly connected
2463to the specified interface(s) from entering the system through
2464any other interface.
2465.Pp
2466For example:
2467.Dl antispoof for lo0
2468.Pp
2469Expands to:
2470.Bd -literal -offset indent -compact
2471block drop in on ! lo0 inet from 127.0.0.1/8 to any
2472block drop in on ! lo0 inet6 from ::1 to any
2473.Ed
2474.Pp
2475For non-loopback interfaces, there are additional rules to block incoming
2476packets with a source IP address identical to the interface's IP(s).
2477For example, assuming the interface wi0 had an IP address of 10.0.0.1 and a
2478netmask of 255.255.255.0:
2479.Pp
2480.Dl antispoof for wi0 inet
2481.Pp
2482Expands to:
2483.Bd -literal -offset indent -compact
2484block drop in on ! wi0 inet from 10.0.0.0/24 to any
2485block drop in inet from 10.0.0.1 to any
2486.Ed
2487.Pp
2488Caveat: Rules created by the
2489.Ic antispoof
2490directive interfere with packets sent over loopback interfaces
2491to local addresses.
2492One should pass these explicitly.
2493.Sh OPERATING SYSTEM FINGERPRINTING
2494Passive OS fingerprinting is a mechanism to inspect nuances of a TCP
2495connection's initial SYN packet and guess at the host's operating system.
2496Unfortunately these nuances are easily spoofed by an attacker so the
2497fingerprint is not useful in making security decisions.
2498But the fingerprint is typically accurate enough to make policy decisions
2499upon.
2500.Pp
2501The fingerprints may be specified by operating system class, by
2502version, or by subtype/patchlevel.
2503The class of an operating system is typically the vendor or genre
2504and would be
2505.Ox
2506for the
2507.Xr pf 4
2508firewall itself.
2509The version of the oldest available
2510.Ox
2511release on the main FTP site
2512would be 2.6 and the fingerprint would be written as:
2513.Pp
2514.Dl \&"OpenBSD 2.6\&"
2515.Pp
2516The subtype of an operating system is typically used to describe the
2517patchlevel if that patch led to changes in the TCP stack behavior.
2518In the case of
2519.Ox ,
2520the only subtype is for a fingerprint that was
2521normalised by the
2522.Cm no-df
2523scrub option and would be specified as:
2524.Pp
2525.Dl \&"OpenBSD 3.3 no-df\&"
2526.Pp
2527Fingerprints for most popular operating systems are provided by
2528.Xr pf.os 5 .
2529Once
2530.Xr pf 4
2531is running, a complete list of known operating system fingerprints may
2532be listed by running:
2533.Pp
2534.Dl # pfctl -so
2535.Pp
2536Filter rules can enforce policy at any level of operating system specification
2537assuming a fingerprint is present.
2538Policy could limit traffic to approved operating systems or even ban traffic
2539from hosts that aren't at the latest service pack.
2540.Pp
2541The
2542.Cm unknown
2543class can also be used as the fingerprint which will match packets for
2544which no operating system fingerprint is known.
2545.Pp
2546Examples:
2547.Bd -literal -offset indent
2548pass  out proto tcp from any os OpenBSD
2549block out proto tcp from any os Doors
2550block out proto tcp from any os "Doors PT"
2551block out proto tcp from any os "Doors PT SP3"
2552block out from any os "unknown"
2553pass on lo0 proto tcp from any os "OpenBSD 3.3 lo0"
2554.Ed
2555.Pp
2556Operating system fingerprinting is limited only to the TCP SYN packet.
2557This means that it will not work on other protocols and will not match
2558a currently established connection.
2559.Pp
2560Caveat: operating system fingerprints are occasionally wrong.
2561There are three problems: an attacker can trivially craft packets to
2562appear as any operating system;
2563an operating system patch could change the stack behavior and no fingerprints
2564will match it until the database is updated;
2565and multiple operating systems may have the same fingerprint.
2566.Sh EXAMPLES
2567In this example,
2568the external interface is
2569.Pa kue0 .
2570We use a macro for the interface name, so it can be changed easily.
2571All incoming traffic is "normalised",
2572and everything is blocked and logged by default.
2573.Bd -literal -offset 4n
2574ext_if = "kue0"
2575match in all scrub (no-df max-mss 1440)
2576block return log on $ext_if all
2577.Ed
2578.Pp
2579Here we specifically block packets we don't want:
2580anything coming from source we have no back routes for;
2581packets whose ingress interface does not match the one in
2582the route back to their source address;
2583anything that does not have our address (157.161.48.183) as source;
2584broadcasts (cable modem noise);
2585and anything from reserved address space or invalid addresses.
2586.Bd -literal -offset 4n
2587block in from no-route to any
2588block in from urpf-failed to any
2589block out log quick on $ext_if from ! 157.161.48.183 to any
2590block in quick on $ext_if from any to 255.255.255.255
2591block in log quick on $ext_if from { 10.0.0.0/8, 172.16.0.0/12, \e
2592    192.168.0.0/16, 255.255.255.255/32 } to any
2593.Ed
2594.Pp
2595For ICMP,
2596pass out/in ping queries.
2597State matching is done on host addresses and ICMP ID (not type/code),
2598so replies (like 0/0 for 8/0) will match queries.
2599ICMP error messages (which always refer to a TCP/UDP packet)
2600are handled by the TCP/UDP states.
2601.Bd -literal -offset 4n
2602pass on $ext_if inet proto icmp all icmp-type 8 code 0
2603.Ed
2604.Pp
2605For UDP,
2606pass out all UDP connections.
2607DNS connections are passed in.
2608.Bd -literal -offset 4n
2609pass out on $ext_if proto udp all
2610pass in on $ext_if proto udp from any to any port domain
2611.Ed
2612.Pp
2613For TCP,
2614pass out all TCP connections and modulate state.
2615SSH, SMTP, DNS, and IDENT connections are passed in.
2616We do not allow Windows 9x SMTP connections since they are typically
2617a viral worm.
2618.Bd -literal -offset 4n
2619pass out on $ext_if proto tcp all modulate state
2620pass in on $ext_if proto tcp from any to any \e
2621    port { ssh, smtp, domain, auth }
2622block in on $ext_if proto tcp from any \e
2623    os { "Windows 95", "Windows 98" } to any port smtp
2624.Ed
2625.Pp
2626Here we pass in/out all IPv6 traffic:
2627note that we have to enable this in two different ways,
2628on both our physical interface and our tunnel.
2629.Bd -literal -offset 4n
2630pass quick on gif0 inet6
2631pass quick on $ext_if proto ipv6
2632.Ed
2633.Pp
2634This example illustrates packet tagging.
2635There are three interfaces: $int_if, $ext_if, and $wifi_if (wireless).
2636NAT is being done on $ext_if for all outgoing packets.
2637Packets in on $int_if are tagged and passed out on $ext_if.
2638All other outgoing packets
2639(i.e. packets from the wireless network)
2640are only permitted to access port 80.
2641.Bd -literal -offset 4n
2642pass in on $int_if from any to any tag INTNET
2643pass in on $wifi_if from any to any
2644
2645block out on $ext_if from any to any
2646pass out quick on $ext_if tagged INTNET
2647pass out on $ext_if proto tcp from any to any port 80
2648.Ed
2649.Pp
2650In this example,
2651we tag incoming packets as they are redirected to
2652.Xr spamd 8 .
2653The tag is used to pass those packets through the packet filter.
2654.Bd -literal -offset 4n
2655match in on $ext_if inet proto tcp from <spammers> to port smtp \e
2656     tag SPAMD rdr-to 127.0.0.1 port spamd
2657
2658block in on $ext_if
2659pass in on $ext_if inet proto tcp tagged SPAMD
2660.Ed
2661.Pp
2662This example maps incoming requests on port 80 to port 8080, on
2663which a daemon is running (because, for example, it is not run as root,
2664and therefore lacks permission to bind to port 80).
2665.Bd -literal -offset 4n
2666match in on $ext_if proto tcp from any to any port 80 \e
2667      rdr-to 127.0.0.1 port 8080
2668.Ed
2669.Pp
2670If a
2671.Ic pass
2672rule is used with the
2673.Cm quick
2674modifier, packets matching the translation rule are passed without
2675inspecting subsequent filter rules.
2676.Bd -literal -offset 4n
2677pass in quick on $ext_if proto tcp from any to any port 80 \e
2678      rdr-to 127.0.0.1 port 8080
2679.Ed
2680.Pp
2681In the example below, vlan12 is configured as 192.168.168.1;
2682the machine translates all packets coming from 192.168.168.0/24 to 204.92.77.111
2683when they are going out any interface except vlan12.
2684This has the net effect of making traffic from the 192.168.168.0/24
2685network appear as though it is the Internet routable address
2686204.92.77.111 to nodes behind any interface on the router except
2687for the nodes on vlan12.
2688Thus, 192.168.168.1 can talk to the 192.168.168.0/24 nodes.
2689.Bd -literal -offset 4n
2690match out on ! vlan12 from 192.168.168.0/24 to any nat-to 204.92.77.111
2691.Ed
2692.Pp
2693In the example below, the machine sits between a fake internal
2694144.19.74.* network, and a routable external IP of 204.92.77.100.
2695The last rule excludes protocol AH from being translated.
2696.Bd -literal -offset 4n
2697pass out on $ext_if from 144.19.74.0/24 nat-to 204.92.77.100
2698pass out on $ext_if proto ah from 144.19.74.0/24
2699.Ed
2700.Pp
2701In the example below, packets bound for one specific server, as well as those
2702generated by the sysadmins are not proxied; all other connections are.
2703.Bd -literal -offset 4n
2704pass in on $int_if proto { tcp, udp } from any to any port 80 \e
2705      rdr-to 127.0.0.1 port 80
2706pass in on $int_if proto { tcp, udp } from any to $server port 80
2707pass in on $int_if proto { tcp, udp } from $sysadmins to any port 80
2708.Ed
2709.Pp
2710This example maps outgoing packets' source port
2711to an assigned proxy port instead of an arbitrary port.
2712In this case, proxy outgoing isakmp with port 500 on the gateway.
2713.Bd -literal -offset 4n
2714match out on $ext_if inet proto udp from any port isakmp to any \e
2715    nat-to ($ext_if) port 500
2716.Ed
2717.Pp
2718One more example uses
2719.Cm rdr-to
2720to redirect a TCP and UDP port to an internal machine.
2721.Bd -literal -offset 4n
2722match in on $ext_if inet proto tcp from any to ($ext_if) port 8080 \e
2723      rdr-to 10.1.2.151 port 22
2724match in on $ext_if inet proto udp from any to ($ext_if) port 8080 \e
2725      rdr-to 10.1.2.151 port 53
2726.Ed
2727.Pp
2728In this example, a NAT gateway is set up to translate internal addresses
2729using a pool of public addresses (192.0.2.16/28).
2730A given source address is always translated to the same pool address by
2731using the
2732.Cm source-hash
2733keyword.
2734The gateway also translates incoming web server connections
2735to a group of web servers on the internal network.
2736.Bd -literal -offset 4n
2737match out on $ext_if inet from any to any nat-to 192.0.2.16/28 \e
2738    source-hash
2739match in  on $ext_if proto tcp from any to any port 80 \e
2740    rdr-to { 10.1.2.155 weight 2, 10.1.2.160 weight 1, \e
2741             10.1.2.161 weight 8 } round-robin
2742.Ed
2743.Pp
2744The bidirectional address translation example uses a single
2745.Cm binat-to
2746rule that expands to a
2747.Cm nat-to
2748and an
2749.Cm rdr-to
2750rule.
2751.Bd -literal -offset 4n
2752pass on $ext_if from 10.1.2.120 to any binat-to 192.0.2.17
2753.Ed
2754.Pp
2755The previous example is identical to the following set of rules:
2756.Bd -literal -offset 4n
2757pass out on $ext_if inet from 10.1.2.120 to any \e
2758      nat-to 192.0.2.17 static-port
2759pass in on $ext_if inet from any to 192.0.2.17 rdr-to 10.1.2.120
2760.Ed
2761.Pp
2762In the example below, a router handling both address families
2763translates an internal IPv4 subnet to IPv6 using the well-known
276464:ff9b::/96 prefix:
2765.Bd -literal -offset 4n
2766pass in on $v4_if inet af-to inet6 from ($v6_if) to 64:ff9b::/96
2767.Ed
2768.Pp
2769Paired with the example above, the example below can be used on
2770another router handling both address families to translate back
2771to IPv4:
2772.Bd -literal -offset 4n
2773pass in on $v6_if inet6 to 64:ff9b::/96 af-to inet from ($v4_if)
2774.Ed
2775.Sh GRAMMAR
2776Syntax for
2777.Nm
2778in BNF:
2779.Bd -literal
2780line           = ( option | pf-rule |
2781                 antispoof-rule | queue-rule | anchor-rule |
2782                 anchor-close | load-anchor | table-rule | include )
2783
2784option         = "set" ( [ "timeout" ( timeout | "{" timeout-list "}" ) ] |
2785                 [ "ruleset-optimization" [ "none" | "basic" |
2786                 "profile" ] ] |
2787                 [ "optimization" [ "default" | "normal" | "high-latency" |
2788                 "satellite" | "aggressive" | "conservative" ] ]
2789                 [ "limit" ( limit-item | "{" limit-list "}" ) ] |
2790                 [ "loginterface" ( interface-name | "none" ) ] |
2791                 [ "block-policy" ( "drop" | "return" ) ] |
2792                 [ "state-policy" ( "if-bound" | "floating" ) ]
2793                 [ "state-defaults" state-opts ]
2794                 [ "fingerprints" filename ] |
2795                 [ "skip on" ifspec ] |
2796                 [ "debug" ( "emerg" | "alert" | "crit" | "err" |
2797                 "warning" | "notice" | "info" | "debug" ) ] |
2798                 [ "reassemble" ( "yes" | "no" ) [ "no-df" ] ] )
2799
2800pf-rule        = action [ ( "in" | "out" ) ]
2801                 [ "log" [ "(" logopts ")"] ] [ "quick" ]
2802                 [ "on" ( ifspec | "rdomain" number ) ] [ af ]
2803                 [ protospec ] [ hosts ] [ filteropts ]
2804
2805logopts        = logopt [ [ "," ] logopts ]
2806logopt         = "all" | "matches" | "user" | "to" interface-name
2807
2808filteropts     = filteropt [ [ "," ] filteropts ]
2809filteropt      = user | group | flags | icmp-type | icmp6-type |
2810                 "tos" tos |
2811                 ( "no" | "keep" | "modulate" | "synproxy" ) "state"
2812                 [ "(" state-opts ")" ] | "scrub" "(" scrubopts ")" |
2813                 "fragment" | "allow-opts" | "once" |
2814                 "divert-packet" "port" port | "divert-reply" |
2815                 "divert-to" host "port" port |
2816                 "label" string | "tag" string | [ "!" ] "tagged" string |
2817                 "max-pkt-rate" number "/" seconds |
2818                 "set delay" number |
2819                 "set prio" ( number | "(" number [ [ "," ] number ] ")" ) |
2820                 "set queue" ( string | "(" string [ [ "," ] string ] ")" ) |
2821                 "rtable" number | "probability" number"%" | "prio" number |
2822                 "af-to" af "from" ( redirhost | "{" redirhost-list "}" )
2823                 [ "to" ( redirhost | "{" redirhost-list "}" ) ] |
2824                 "binat-to" ( redirhost | "{" redirhost-list "}" )
2825                 [ portspec ] [ pooltype ] |
2826                 "rdr-to" ( redirhost | "{" redirhost-list "}" )
2827                 [ portspec ] [ pooltype ] |
2828                 "nat-to" ( redirhost | "{" redirhost-list "}" )
2829                 [ portspec ] [ pooltype ] [ "static-port" ] |
2830                 [ route ] | [ "set tos" tos ] |
2831                 [ [ "!" ] "received-on" ( interface-name | interface-group ) ]
2832
2833scrubopts      = scrubopt [ [ "," ] scrubopts ]
2834scrubopt       = "no-df" | "min-ttl" number | "max-mss" number |
2835                 "reassemble tcp" | "random-id"
2836
2837antispoof-rule = "antispoof" [ "log" ] [ "quick" ]
2838                 "for" ifspec [ af ] [ "label" string ]
2839
2840table-rule     = "table" "<" string ">" [ tableopts ]
2841tableopts      = tableopt [ tableopts ]
2842tableopt       = "persist" | "const" | "counters" |
2843                 "file" string | "{" [ tableaddrs ] "}"
2844tableaddrs     = tableaddr-spec [ [ "," ] tableaddrs ]
2845tableaddr-spec = [ "!" ] tableaddr [ "/" mask-bits ]
2846tableaddr      = hostname | ifspec | "self" |
2847                 ipv4-dotted-quad | ipv6-coloned-hex
2848
2849queue-rule     = "queue" string [ "on" interface-name ] queueopts-list
2850
2851anchor-rule    = "anchor" [ string ] [ ( "in" | "out" ) ] [ "on" ifspec ]
2852                 [ af ] [ protospec ] [ hosts ] [ filteropt-list ] [ "{" ]
2853
2854anchor-close   = "}"
2855
2856load-anchor    = "load anchor" string "from" filename
2857
2858queueopts-list = queueopts-list queueopts | queueopts
2859queueopts      = ([ "bandwidth" bandwidth ] | [ "min" bandwidth ] |
2860                 [ "max" bandwidth ] | [ "parent" string ] |
2861                 [ "default" ]) |
2862                 ([ "flows" number ] | [ "quantum" number ]) |
2863                 [ "qlimit" number ]
2864
2865bandwidth      = bandwidth-spec [ "burst" bandwidth-spec "for" number "ms" ]
2866bandwidth-spec = number ( "" | "K" | "M" | "G" )
2867
2868action         = "pass" | "match" | "block" [ return ]
2869return         = "drop" | "return" |
2870                 "return-rst" [ "(" "ttl" number ")" ] |
2871                 "return-icmp" [ "(" icmpcode [ [ "," ] icmp6code ] ")" ] |
2872                 "return-icmp6" [ "(" icmp6code ")" ]
2873icmpcode       = ( icmp-code-name | icmp-code-number )
2874icmp6code      = ( icmp6-code-name | icmp6-code-number )
2875
2876ifspec         = ( [ "!" ] ( interface-name | interface-group ) ) |
2877                 "{" interface-list "}"
2878interface-list = [ "!" ] ( interface-name | interface-group )
2879                 [ [ "," ] interface-list ]
2880route          = ( "route-to" | "reply-to" | "dup-to" )
2881                 ( redirhost | "{" redirhost-list "}" )
2882af             = "inet" | "inet6"
2883
2884protospec      = "proto" ( proto-name | proto-number |
2885                 "{" proto-list "}" )
2886proto-list     = ( proto-name | proto-number ) [ [ "," ] proto-list ]
2887
2888hosts          = "all" |
2889                 "from" ( "any" | "no-route" | "urpf-failed" | "self" |
2890                 host | "{" host-list "}" | "route" string ) [ port ]
2891                 [ os ]
2892                 "to"   ( "any" | "no-route" | "self" | host |
2893                 "{" host-list "}" | "route" string ) [ port ]
2894
2895ipspec         = "any" | host | "{" host-list "}"
2896host           = [ "!" ] ( address [ "weight" number ] |
2897                 address [ "/" mask-bits ] [ "weight" number ] |
2898                 "<" string ">" )
2899redirhost      = address [ "/" mask-bits ]
2900address        = ( interface-name | interface-group |
2901                 "(" ( interface-name | interface-group ) ")" |
2902                 hostname | ipv4-dotted-quad | ipv6-coloned-hex )
2903host-list      = host [ [ "," ] host-list ]
2904redirhost-list = redirhost [ [ "," ] redirhost-list ]
2905
2906port           = "port" ( unary-op | binary-op | "{" op-list "}" )
2907portspec       = "port" ( number | name ) [ ":" ( "*" | number | name ) ]
2908os             = "os"  ( os-name | "{" os-list "}" )
2909user           = "user" ( unary-op | binary-op | "{" op-list "}" )
2910group          = "group" ( unary-op | binary-op | "{" op-list "}" )
2911
2912unary-op       = [ "=" | "!=" | "<" | "<=" | ">" | ">=" ]
2913                 ( name | number )
2914binary-op      = number ( "<>" | "><" | ":" ) number
2915op-list        = ( unary-op | binary-op ) [ [ "," ] op-list ]
2916
2917os-name        = operating-system-name
2918os-list        = os-name [ [ "," ] os-list ]
2919
2920flags          = "flags" ( [ flag-set ] "/"  flag-set | "any" )
2921flag-set       = [ "F" ] [ "S" ] [ "R" ] [ "P" ] [ "A" ] [ "U" ] [ "E" ]
2922                 [ "W" ]
2923
2924icmp-type      = "icmp-type" ( icmp-type-code | "{" icmp-list "}" )
2925icmp6-type     = "icmp6-type" ( icmp-type-code | "{" icmp-list "}" )
2926icmp-type-code = ( icmp-type-name | icmp-type-number )
2927                 [ "code" ( icmp-code-name | icmp-code-number ) ]
2928icmp-list      = icmp-type-code [ [ "," ] icmp-list ]
2929
2930tos            = ( "lowdelay" | "throughput" | "reliability" |
2931                 [ "0x" ] number )
2932
2933state-opts     = state-opt [ [ "," ] state-opts ]
2934state-opt      = ( "max" number | "no-sync" | timeout | "sloppy" |
2935                 "pflow" | "source-track" [ ( "rule" | "global" ) ] |
2936                 "max-src-nodes" number | "max-src-states" number |
2937                 "max-src-conn" number |
2938                 "max-src-conn-rate" number "/" number |
2939                 "overload" "<" string ">" [ "flush" [ "global" ] ] |
2940                 "if-bound" | "floating" )
2941
2942timeout-list   = timeout [ [ "," ] timeout-list ]
2943timeout        = ( "tcp.first" | "tcp.opening" | "tcp.established" |
2944                 "tcp.closing" | "tcp.finwait" | "tcp.closed" |
2945                 "udp.first" | "udp.single" | "udp.multiple" |
2946                 "icmp.first" | "icmp.error" |
2947                 "other.first" | "other.single" | "other.multiple" |
2948                 "frag" | "interval" | "src.track" |
2949                 "adaptive.start" | "adaptive.end" ) number
2950
2951limit-list     = limit-item [ [ "," ] limit-list ]
2952limit-item     = ( "states" | "frags" | "src-nodes" | "tables" |
2953                 "table-entries" ) number
2954
2955pooltype       = ( "bitmask" | "least-states" |
2956                 "random" | "round-robin" |
2957                 "source-hash" [ ( hex-key | string-key ) ] )
2958                 [ "sticky-address" ]
2959
2960include        = "include" filename
2961.Ed
2962.Sh FILES
2963.Bl -tag -width /etc/examples/pf.conf -compact
2964.It Pa /etc/hosts
2965Host name database.
2966.It Pa /etc/pf.conf
2967Default location of the ruleset file.
2968.It Pa /etc/examples/pf.conf
2969Example ruleset file.
2970.It Pa /etc/pf.os
2971Default location of OS fingerprints.
2972.It Pa /etc/protocols
2973Protocol name database.
2974.It Pa /etc/services
2975Service name database.
2976.El
2977.Sh SEE ALSO
2978.Xr pf 4 ,
2979.Xr pflow 4 ,
2980.Xr pfsync 4 ,
2981.Xr pf.os 5 ,
2982.Xr pfctl 8 ,
2983.Xr pflogd 8
2984.Sh HISTORY
2985The
2986.Nm
2987file format first appeared in
2988.Ox 3.0 .
2989