1 .\" $OpenBSD: pf.conf.5,v 1.402 2008/06/11 07:21:00 jmc Exp $
3 .\" Copyright (c) 2002, Daniel Hartmeier
4 .\" All rights reserved.
6 .\" Redistribution and use in source and binary forms, with or without
7 .\" modification, are permitted provided that the following conditions
10 .\" - Redistributions of source code must retain the above copyright
11 .\" notice, this list of conditions and the following disclaimer.
12 .\" - Redistributions in binary form must reproduce the above
13 .\" copyright notice, this list of conditions and the following
14 .\" disclaimer in the documentation and/or other materials provided
15 .\" with the distribution.
17 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18 .\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19 .\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
20 .\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
21 .\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
22 .\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
23 .\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
24 .\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
25 .\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
27 .\" ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28 .\" POSSIBILITY OF SUCH DAMAGE.
35 .Nd packet filter configuration file
39 packet filter modifies, drops or passes packets according to rules or
40 definitions specified in
43 There are seven types of statements in
47 User-defined variables may be defined and used later, simplifying
48 the configuration file.
49 Macros must be defined before they are referenced in
52 Tables provide a mechanism for increasing the performance and flexibility of
53 rules with large numbers of source or destination addresses.
55 Options tune the behaviour of the packet filtering engine.
56 .It Cm Traffic Normalization Li (e.g.\& Em scrub )
57 Traffic normalization protects internal machines against inconsistencies
58 in Internet protocols and implementations.
60 Queueing provides rule-based bandwidth control.
61 .It Cm Translation Li (Various forms of NAT)
62 Translation rules specify how addresses are to be mapped or redirected to
64 .It Cm Packet Filtering
65 Packet filtering provides rule-based blocking or passing of packets.
72 the types of statements should be grouped and appear in
74 in the order shown above, as this matches the operation of the underlying
75 packet filtering engine.
78 enforces this order (see
82 Comments can be put anywhere in the file using a hash mark
84 and extend to the end of the current line.
86 Additional configuration files can be included with the
89 .Bd -literal -offset indent
90 include "/etc/pf/sub.filter.conf"
93 Macros can be defined that will later be expanded in context.
94 Macro names must start with a letter, and may contain letters, digits
96 Macro names may not be reserved words (for example
100 Macros are not expanded inside quotes.
103 .Bd -literal -offset indent
105 all_ifs = \&"{\&" $ext_if lo0 \&"}\&"
106 pass out on $ext_if from any to any
107 pass in on $ext_if proto tcp from any to any port 25
110 Tables are named structures which can hold a collection of addresses and
112 Lookups against tables in
114 are relatively fast, making a single rule with tables much more efficient,
116 processor usage and memory consumption, than a large number of rules which
117 differ only in IP address (either created explicitly or automatically by rule
120 Tables can be used as the source or destination of filter rules,
124 translation rules such as
128 (see below for details on the various rule types).
129 Tables can also be used for the redirect address of
133 rules and in the routing options of filter rules, but only for
137 Tables can be defined with any of the following
140 As with macros, reserved words may not be used as table names.
141 .Bl -tag -width "manually"
143 Persistent tables can be manually created with the
149 before or after the ruleset has been loaded.
151 Table definitions can be placed directly in this file, and loaded at the
152 same time as other rules are loaded, atomically.
153 Table definitions inside
157 statement, and are especially useful to define non-persistent tables.
158 The contents of a pre-existing table defined without a list of addresses
159 to initialize it is not altered when
162 A table initialized with the empty list,
164 will be cleared on load.
167 Tables may be defined with the following two attributes:
168 .Bl -tag -width persist
172 flag forces the kernel to keep the table even when no rules refer to it.
173 If the flag is not set, the kernel will automatically remove the table
174 when the last rule referring to it is flushed.
178 flag prevents the user from altering the contents of the table once it
182 can be used to add or remove addresses from the table at any time, even
189 flag enables per-address packet and byte counters which can be displayed with
194 .Bd -literal -offset indent
195 table \*(Ltprivate\*(Gt const { 10/8, 172.16/12, 192.168/16 }
196 table \*(Ltbadhosts\*(Gt persist
197 block on fxp0 from { \*(Ltprivate\*(Gt, \*(Ltbadhosts\*(Gt } to any
200 creates a table called private, to hold RFC 1918 private network
201 blocks, and a table called badhosts, which is initially empty.
202 A filter rule is set up to block all traffic coming from addresses listed in
204 The private table cannot have its contents changed and the badhosts table
205 will exist even when no active filter rules reference it.
206 Addresses may later be added to the badhosts table, so that traffic from
207 these hosts can be blocked by using
208 .Bd -literal -offset indent
209 # pfctl -t badhosts -Tadd 204.92.77.111
212 A table can also be initialized with an address list specified in one or more
213 external files, using the following syntax:
214 .Bd -literal -offset indent
215 table \*(Ltspam\*(Gt persist file \&"/etc/spammers\&" file \&"/etc/openrelays\&"
216 block on fxp0 from \*(Ltspam\*(Gt to any
223 list IP addresses, one per line.
224 Any lines beginning with a # are treated as comments and ignored.
225 In addition to being specified by IP address, hosts may also be
226 specified by their hostname.
227 When the resolver is called to add a hostname to a table,
229 resulting IPv4 and IPv6 addresses are placed into the table.
230 IP addresses can also be entered in a table by specifying a valid interface
233 keyword, in which case all addresses assigned to the interface will be
237 may be tuned for various situations using the
243 .Bl -tag -width "src.track" -compact
245 Interval between purging expired states and fragments.
247 Seconds before an unassembled fragment is expired.
249 Length of time to retain a source tracking entry after the last state
253 When a packet matches a stateful connection, the seconds to live for the
254 connection will be updated to that of the
256 which corresponds to the connection state.
257 Each packet which matches this state will reset the TTL.
258 Tuning these values may improve the performance of the
259 firewall at the risk of dropping valid idle connections.
261 .Bl -tag -width xxxx -compact
263 The state after the first packet.
265 The state before the destination host ever sends a packet.
266 .It Ar tcp.established
267 The fully established state.
269 The state after the first FIN has been sent.
271 The state after both FINs have been exchanged and the connection is closed.
272 Some hosts (notably web servers on Solaris) send TCP packets even after closing
278 can prevent blocking of such packets.
280 The state after one endpoint sends an RST.
283 ICMP and UDP are handled in a fashion similar to TCP, but with a much more
284 limited set of states:
286 .Bl -tag -width xxxx -compact
288 The state after the first packet.
290 The state if the source host sends more than one packet but the destination
291 host has never sent one back.
293 The state if both hosts have sent packets.
295 The state after the first packet.
297 The state after an ICMP error came back in response to an ICMP packet.
300 Other protocols are handled similarly to UDP:
302 .Bl -tag -width xxxx -compact
305 .It Ar other.multiple
308 Timeout values can be reduced adaptively as the number of state table
311 .Bl -tag -width xxxx -compact
312 .It Ar adaptive.start
313 When the number of state entries exceeds this value, adaptive scaling
315 All timeout values are scaled linearly with factor
316 (adaptive.end - number of states) / (adaptive.end - adaptive.start).
318 When reaching this number of state entries, all timeout values become
319 zero, effectively purging all state entries immediately.
320 This value is used to define the scale factor, it should not actually
321 be reached (set a lower state limit, see below).
324 Adaptive timeouts are enabled by default, with an adaptive.start value
325 equal to 60% of the state limit, and an adaptive.end value equal to
326 120% of the state limit.
327 They can be disabled by setting both adaptive.start and adaptive.end to 0.
329 The adaptive timeout values can be defined both globally and for each rule.
330 When used on a per-rule basis, the values relate to the number of
331 states created by the rule, otherwise to the total number of
335 .Bd -literal -offset indent
336 set timeout tcp.first 120
337 set timeout tcp.established 86400
338 set timeout { adaptive.start 6000, adaptive.end 12000 }
339 set limit states 10000
342 With 9000 state table entries, the timeout values are scaled to 50%
343 (tcp.first 60, tcp.established 43200).
345 .It Ar set loginterface
346 Enable collection of packet and byte count statistics for the given interface.
347 These statistics can be viewed using
348 .Bd -literal -offset indent
354 collects statistics on the interface named dc0:
355 .Bd -literal -offset indent
359 One can disable the loginterface using:
360 .Bd -literal -offset indent
361 set loginterface none
365 Sets hard limits on the memory zones used by the packet filter.
368 for an explanation of memory zones.
371 .Bd -literal -offset indent
372 set limit states 20000
375 sets the maximum number of entries in the memory pool used by state table
376 entries (generated by
378 rules which do not specify
382 .Bd -literal -offset indent
383 set limit frags 20000
386 sets the maximum number of entries in the memory pool used for fragment
387 reassembly (generated by
391 .Bd -literal -offset indent
392 set limit src-nodes 2000
395 sets the maximum number of entries in the memory pool used for tracking
396 source IP addresses (generated by the
402 sets limits on the memory pools used by tables.
403 The first limits the number of tables that can exist to 1000.
404 The second limits the overall number of addresses that can be stored
407 Various limits can be combined on a single line:
408 .Bd -literal -offset indent
409 set limit { states 20000, frags 20000, src-nodes 2000 }
412 .It Ar set ruleset-optimization
413 .Bl -tag -width xxxxxxxx -compact
415 Disable the ruleset optimizer.
417 Enable basic ruleset optimization.
418 This is the default behaviour.
419 Basic ruleset optimization does four things to improve the
420 performance of ruleset evaluations:
424 remove duplicate rules
426 remove rules that are a subset of another rule
428 combine multiple rules into a table when advantageous
430 re-order the rules to improve evaluation performance
434 Uses the currently loaded ruleset as a feedback profile to tailor the
435 ordering of quick rules to actual network traffic.
438 It is important to note that the ruleset optimizer will modify the ruleset
439 to improve performance.
440 A side effect of the ruleset modification is that per-rule accounting
441 statistics will have different meanings than before.
442 If per-rule accounting is important for billing purposes or whatnot,
443 either the ruleset optimizer should not be used or a label field should
444 be added to all of the accounting rules to act as optimization barriers.
446 Optimization can also be set as a command-line argument to
448 overriding the settings in
450 .It Ar set optimization
451 Optimize state timeouts for one of the following network environments:
453 .Bl -tag -width xxxx -compact
455 A normal network environment.
456 Suitable for almost all networks.
458 A high-latency environment (such as a satellite connection).
463 Aggressively expire connections.
464 This can greatly reduce the memory usage of the firewall at the cost of
465 dropping idle connections early.
467 Extremely conservative settings.
468 Avoid dropping legitimate connections at the
469 expense of greater memory utilization (possibly much greater on a busy
470 network) and slightly increased processor utilization.
474 .Bd -literal -offset indent
475 set optimization aggressive
478 .It Ar set keep-policy keep_rule
481 option sets the default state retention policy for all
485 .Sx STATEFUL TRACKING OPTIONS
491 .Ar no Ns / Ns Ar keep Ns / Ns Ar modulate Ns / Ns Ar synproxy state
494 rule will override the default.
496 .Bd -literal -offset indent
497 set keep-policy keep state (pickups)
500 .It Ar set block-policy
503 option sets the default behaviour for the packet
507 .Bl -tag -width xxxxxxxx -compact
509 Packet is silently dropped.
511 A TCP RST is returned for blocked TCP packets,
512 an ICMP UNREACHABLE is returned for blocked UDP packets,
513 and all other packets are silently dropped.
517 .Bd -literal -offset indent
518 set block-policy return
520 .It Ar set state-policy
523 option sets the default behaviour for states:
525 .Bl -tag -width if-bound -compact
527 States are bound to interface.
529 States can match packets on any interfaces (the default).
533 .Bd -literal -offset indent
534 set state-policy if-bound
539 identifies this firewall's state table entries to other firewalls
543 By default the hostid is set to a pseudo-random value, however it may be
544 desirable to manually configure it, for example to more easily identify the
545 source of state table entries.
546 .Bd -literal -offset indent
550 The hostid may be specified in either decimal or hexadecimal.
551 .It Ar set require-order
554 enforces an ordering of the statement types in the ruleset to:
560 Setting this option to
562 disables this enforcement.
563 There may be non-trivial and non-obvious implications to an out of
565 Consider carefully before disabling the order enforcement.
566 .It Ar set fingerprints
567 Load fingerprints of known operating systems from the given filename.
568 By default fingerprints of known operating systems are automatically
573 but can be overridden via this option.
574 Setting this option may leave a small period of time where the fingerprints
575 referenced by the currently active ruleset are inconsistent until the new
576 ruleset finishes loading.
580 .Dl set fingerprints \&"/etc/pf.os.devel\&"
582 .It Ar set skip on Aq Ar ifspec
583 List interfaces for which packets should not be filtered.
584 Packets passing in or out on such interfaces are passed as if pf was
585 disabled, i.e. pf does not process them in any way.
586 This can be useful on loopback and other virtual interfaces, when
587 packet filtering is not desired and can have unexpected effects.
595 to one of the following:
597 .Bl -tag -width xxxxxxxxxxxx -compact
599 Don't generate debug messages.
601 Generate debug messages only for serious errors.
603 Generate debug messages for various errors.
605 Generate debug messages for common conditions.
608 .Sh TRAFFIC NORMALIZATION
609 Traffic normalization is used to sanitize packet content in such
610 a way that there are no ambiguities in packet interpretation on
612 The normalizer does IP fragment reassembly to prevent attacks
613 that confuse intrusion detection systems by sending overlapping
615 Packet normalization is invoked with the
620 has the following options:
625 bit from a matching IP packet.
626 Some operating systems are known to generate fragmented packets with the
629 This is particularly true with NFS.
631 will drop such fragmented
637 Unfortunately some operating systems also generate their
639 packets with a zero IP identification field.
642 bit on packets with a zero IP ID may cause deleterious results if an
643 upstream router later fragments the packet.
646 modifier (see below) is recommended in combination with the
648 modifier to ensure unique IP identifiers.
649 .It Ar min-ttl Aq Ar number
650 Enforces a minimum TTL for matching IP packets.
651 .It Ar max-mss Aq Ar number
652 Enforces a maximum MSS for matching TCP packets.
653 .It Xo Ar set-tos Aq Ar string
654 .No \*(Ba Aq Ar number
658 for matching IP packets.
665 or as either hex or decimal.
667 Replaces the IP identification field with random values to compensate
668 for predictable values generated by many hosts.
669 This option only applies to packets that are not fragmented
670 after the optional fragment reassembly.
671 .It Ar fragment reassemble
674 rules, fragments can be reassembled by normalization.
675 In this case, fragments are buffered until they form a complete
676 packet, and only the completed packet is passed on to the filter.
677 The advantage is that filter rules have to deal only with complete
678 packets, and can ignore fragments.
679 The drawback of caching fragments is the additional memory cost.
680 But the full reassembly method is the only method that currently works
682 This is the default behavior of a
684 rule if no fragmentation modifier is supplied.
686 The default fragment reassembly method is expensive, hence the option
690 will track the fragments and cache a small range descriptor.
691 Duplicate fragments are dropped and overlaps are cropped.
692 Thus data will only occur once on the wire with ambiguities resolving to
693 the first occurrence.
695 .Ar fragment reassemble
696 modifier, fragments are not buffered, they are passed as soon as they
700 reassembly mechanism does not yet work with NAT.
702 .It Ar fragment drop-ovl
703 This option is similar to the
705 modifier except that all overlapping or duplicate fragments will be
706 dropped, and all further corresponding fragments will be
708 .It Ar reassemble tcp
709 Statefully normalizes TCP connections.
710 .Ar scrub reassemble tcp
711 rules may not have the direction (in/out) specified.
713 performs the following normalizations:
715 .Bl -tag -width timeout -compact
717 Neither side of the connection is allowed to reduce their IP TTL.
718 An attacker may send a packet such that it reaches the firewall, affects
719 the firewall state, and expires before reaching the destination host.
721 will raise the TTL of all packets back up to the highest value seen on
723 .It timestamp modulation
724 Modern TCP stacks will send a timestamp on every TCP packet and echo
725 the other endpoint's timestamp back to them.
726 Many operating systems will merely start the timestamp at zero when
727 first booted, and increment it several times a second.
728 The uptime of the host can be deduced by reading the timestamp and multiplying
730 Also observing several different timestamps can be used to count hosts
732 And spoofing TCP packets into a connection requires knowing or guessing
734 Timestamps merely need to be monotonically increasing and not derived off a
739 to modulate the TCP timestamps with a random number.
740 .It extended PAWS checks
741 There is a problem with TCP on long fat pipes, in that a packet might get
742 delayed for longer than it takes the connection to wrap its 32-bit sequence
744 In such an occurrence, the old packet would be indistinguishable from a
745 new packet and would be accepted as such.
746 The solution to this is called PAWS: Protection Against Wrapped Sequence
748 It protects against it by making sure the timestamp on each packet does
751 also makes sure the timestamp on the packet does not go forward more
755 artificially extends the security of TCP sequence numbers by 10 to 18
756 bits when the host uses appropriately randomized timestamps, since a
757 blind attacker would have to guess the timestamp as well.
762 .Bd -literal -offset indent
763 scrub in on $ext_if all fragment reassemble
768 option prefixed to a scrub rule causes matching packets to remain unscrubbed,
769 much in the same way as
771 works in the packet filter (see below).
772 This mechanism should be used when it is necessary to exclude specific packets
773 from broader scrub rules.
775 Packets can be assigned to queues for the purpose of bandwidth
777 At least two declarations are required to configure queues, and later
778 any packet filtering rule can reference the defined queues by name.
779 During the filtering component of
783 name is where any packets from
785 rules will be queued, while for
787 rules it specifies where any resulting ICMP or TCP RST
788 packets should be queued.
791 defines the algorithm used to decide which packets get delayed, dropped, or
792 sent out immediately.
796 .Bl -tag -width ".Ar fairq"
798 Class Based Queueing.
800 attached to an interface build a tree, thus each
802 can have further child
804 Each queue can have a
810 mainly controls the time packets take to get sent out, while
812 has primarily effects on throughput.
814 achieves both partitioning and sharing of link bandwidth
815 by hierarchically structured classes.
816 Each class has its own
818 and is assigned its share of
820 A child class can borrow bandwidth from its parent class
821 as long as excess bandwidth is available
828 are flat attached to the interface, thus,
830 cannot have further child
836 assigned, ranging from 0 to 15.
843 Hierarchical Fair Service Curve.
845 attached to an interface build a tree, thus each
847 can have further child
849 Each queue can have a
855 mainly controls the time packets take to get sent out, while
857 primarily affects throughput.
859 supports both link-sharing and guaranteed real-time services.
860 It employs a service curve based QoS model,
861 and its unique feature is an ability to decouple
869 are flat attached to the interface, thus,
871 cannot have further child
873 Each queue must be given a unique
875 and one must be marked
876 as the default queue.
877 Each queue implements a number of
879 (default 256) which sorts the
880 traffic based on a hash key generated by the
885 Each bucket contains a list of packets controlled by
889 to function properly,
891 must be enabled on most of the rule sets that route packets to the queue.
892 Any rules for which keep state is not enabled are added to the end of the
894 If you do not wish keep state to do TCP sequence space checks use
895 .Ar "keep state (no-pickups)"
897 .Ar "keep state (hash-only)" .
899 Packet selection operates as follows:
900 The queues are scanned from highest priority to lowest priority.
901 If a queue has pending packets and is under its bandwidth minimum the
902 scan stops and a packet is selected from that queue.
903 If all queues have reached their bandwidth minimum a scale factor based
904 on each queue's bandwidth minimum versus that queue's current bandwidth
905 usage is calculated and the queue with the lowest scale factor is selected.
906 This effectively uses the minimum bandwidth specification as a relative
907 weighting for apportioning any remaining bandwidth on the link.
909 The priority mechanic is only applicable in cases where the aggregate
910 minimum bandwidth guarantees exceed the link bandwidth, and also has
911 a small effect on queue selection when prioritizing between equal scale
916 round robins between its
918 extracting one packet from each bucket.
919 This essentially prevents large backlogs of packets from high volume
920 connections from destroying the interactive response of other connections.
926 is guaranteed minimum and more will be used if no higher priority traffic is
928 Creating a queue with one bucket as a catch-all for
930 rules not characterized by
933 Such a queue serves as a basic priority queue with a bandwidth specification.
935 Also note that when specifying rules it is always a good idea to specify
936 a secondary queue for any tcp rules.
937 The secondary queue is selected for pure ACKs without payloads and should
938 generally be dedicated to that purpose with a minimum bandwidth specification
939 sufficient to max-out the bandwidth for your incoming traffic.
942 The interfaces on which queueing should be activated are declared using
947 has the following keywords:
950 Queueing is enabled on the named interface.
952 Specifies which queueing scheduler to use.
953 Currently supported values
956 for Class Based Queueing,
958 for Priority Queueing,
960 for the Hierarchical Fair Service Curve scheduler, and
962 for the Fair Queueing.
963 .It Ar bandwidth <bw>
964 The maximum bitrate for all queues on an
965 interface may be specified using the
968 The value can be specified as an absolute value or as a
969 percentage of the interface bandwidth.
970 When using an absolute value, the suffixes
976 are used to represent bits, kilobits, megabits, and
977 gigabits per second, respectively.
978 The value must not exceed the interface bandwidth.
981 is not specified, the interface bandwidth is used
982 (but take note that some interfaces do not know their bandwidth,
983 or can adapt their bandwidth rates).
988 specifies a guaranteed minimum but the fairq is allowed to exceed it.
989 .It Ar qlimit <limit>
990 The maximum number of packets held in the queue.
992 .It Ar tbrsize Aq Ar size
993 Adjusts the size, in bytes, of the token bucket regulator.
994 If not specified, heuristics based on the
995 interface bandwidth are used to determine the size.
996 .It Ar queue Aq Ar list
997 Defines a list of subqueues to create on an interface.
1000 In the following example, the interface dc0
1001 should queue up to 5 Mbit/s in four second-level queues using
1002 Class Based Queueing.
1003 Those four queues will be shown in a later example.
1004 .Bd -literal -offset indent
1005 altq on dc0 cbq bandwidth 5Mb queue { std, http, mail, ssh }
1008 Once interfaces are activated for queueing using the
1010 directive, a sequence of
1012 directives may be defined.
1013 The name associated with a
1015 must match a queue defined in the
1017 directive (e.g.\& mail), or, except for the
1025 The following keywords can be used:
1026 .Bl -tag -width xxxx
1027 .It Ar on Aq Ar interface
1028 Specifies the interface the queue operates on.
1029 If not given, it operates on all matching interfaces.
1030 .It Ar bandwidth Aq Ar bw
1031 Specifies the maximum bitrate to be processed by the queue.
1032 This value must not exceed the value of the parent
1034 and can be specified as an absolute value or a percentage of the parent
1036 If not specified, defaults to 100% of the parent queue's bandwidth.
1039 scheduler does not support bandwidth specification.
1042 scheduler uses the bandwidth specification as a guaranteed minimum and
1044 .It Ar priority Aq Ar level
1045 Between queues a priority level can be set.
1051 the range is 0 to 7 and for
1053 the range is 0 to 15.
1054 The default for all is 1.
1056 queues with a higher priority are always served first.
1058 queues with a higher priority are served first unless they exceed their
1059 bandwidth specification.
1063 queues with a higher priority are preferred in the case of overload.
1064 .It Ar qlimit Aq Ar limit
1065 The maximum number of packets held in the queue.
1069 this specified the maximum number of packets held per bucket.
1074 can get additional parameters with
1076 .Pf ( Aq Ar parameters ) .
1078 Parameters are as follows:
1081 Packets not matched by another queue are assigned to this one.
1082 Exactly one default queue is required.
1084 Enable RED (Random Early Detection) on this queue.
1085 RED drops packets with a probability proportional to the average
1088 Enables RIO on this queue.
1089 RIO is RED with IN/OUT, thus running
1090 RED two times more than RIO would achieve the same effect.
1092 Enables ECN (Explicit Congestion Notification) on this queue.
1099 supports the following additional options:
1101 .It Ar buckets <number>
1102 Specify the number of buckets, from 1 to 2048 in powers of 2.
1103 A bucket size of 1 causes a
1105 to essentially degenerate into a priority queue.
1106 .It Ar linkshare <sc>
1107 The bandwidth share of a backlogged queue.
1108 This option is parsed but not yet supported.
1109 .It Ar hogs <bandwidth>
1110 This option allows low bandwidth connections to burst up to the specified
1111 bandwidth by not advancing the round robin when taking packets out of
1113 When using this option a small value no greater than 1/20 available interface
1114 bandwidth is recommended.
1120 supports an additional option:
1123 The queue can borrow bandwidth from the parent.
1129 supports some additional options:
1131 .It Ar realtime Aq Ar sc
1132 The minimum required bandwidth for the queue.
1133 .It Ar upperlimit Aq Ar sc
1134 The maximum allowed bandwidth for the queue.
1135 .It Ar linkshare Aq Ar sc
1136 The bandwidth share of a backlogged queue.
1143 The format for service curve specifications is
1148 controls the bandwidth assigned to the queue.
1152 are optional and can be used to control the initial bandwidth assignment.
1155 milliseconds the queue gets the bandwidth given as
1157 afterwards the value given in
1164 child queues can be specified as in an
1166 declaration, thus building a tree of queues using a part of
1167 their parent's bandwidth.
1169 Packets can be assigned to queues based on filter rules by using the
1174 is specified; when a second one is specified it will instead be used for
1175 packets which have a
1179 and for TCP ACKs with no data payload.
1181 To continue the previous example, the examples below would specify the
1183 queues, plus a few child queues.
1186 sessions get priority over bulk transfers like
1190 The queues may then be referenced by filtering rules (see
1191 .Sx PACKET FILTERING
1194 queue std bandwidth 10% cbq(default)
1195 queue http bandwidth 60% priority 2 cbq(borrow red) \e
1196 { employees, developers }
1197 queue developers bandwidth 75% cbq(borrow)
1198 queue employees bandwidth 15%
1199 queue mail bandwidth 10% priority 0 cbq(borrow ecn)
1200 queue ssh bandwidth 20% cbq(borrow) { ssh_interactive, ssh_bulk }
1201 queue ssh_interactive bandwidth 50% priority 7 cbq(borrow)
1202 queue ssh_bulk bandwidth 50% priority 0 cbq(borrow)
1204 block return out on dc0 inet all queue std
1205 pass out on dc0 inet proto tcp from $developerhosts to any port 80 \e
1207 pass out on dc0 inet proto tcp from $employeehosts to any port 80 \e
1209 pass out on dc0 inet proto tcp from any to any port 22 \e
1210 queue(ssh_bulk, ssh_interactive)
1211 pass out on dc0 inet proto tcp from any to any port 25 \e
1215 Translation rules modify either the source or destination address of the
1216 packets associated with a stateful connection.
1217 A stateful connection is automatically created to track packets matching
1218 such a rule as long as they are not blocked by the filtering section of
1220 The translation engine modifies the specified address and/or port in the
1221 packet, recalculates IP, TCP and UDP checksums as necessary, and passes it to
1222 the packet filter for evaluation.
1224 Since translation occurs before filtering the filter
1225 engine will see packets as they look after any
1226 addresses and ports have been translated.
1227 Filter rules will therefore have to filter based on the translated
1228 address and port number.
1229 Packets that match a translation rule are only automatically passed if
1232 modifier is given, otherwise they are
1239 The state entry created permits
1241 to keep track of the original address for traffic associated with that state
1242 and correctly direct return traffic for that connection.
1244 Various types of translation are possible with pf:
1245 .Bl -tag -width xxxx
1249 rule specifies a bidirectional mapping between an external IP netblock
1250 and an internal IP netblock.
1254 rule specifies that IP addresses are to be changed as the packet
1255 traverses the given interface.
1256 This technique allows one or more IP addresses
1257 on the translating host to support network traffic for a larger range of
1258 machines on an "inside" network.
1259 Although in theory any IP address can be used on the inside, it is strongly
1260 recommended that one of the address ranges defined by RFC 1918 be used.
1261 These netblocks are:
1263 10.0.0.0 - 10.255.255.255 (all of net 10, i.e., 10/8)
1264 172.16.0.0 - 172.31.255.255 (i.e., 172.16/12)
1265 192.168.0.0 - 192.168.255.255 (i.e., 192.168/16)
1268 The packet is redirected to another destination and possibly a
1271 rules can optionally specify port ranges instead of single ports.
1272 rdr ... port 2000:2999 -\*(Gt ... port 4000
1273 redirects ports 2000 to 2999 (inclusive) to port 4000.
1274 rdr ... port 2000:2999 -\*(Gt ... port 4000:*
1275 redirects port 2000 to 4000, 2001 to 4001, ..., 2999 to 4999.
1278 In addition to modifying the address, some translation rules may modify
1279 source or destination ports for
1283 connections; implicitly in the case of
1285 rules and explicitly in the case of
1288 Port numbers are never translated with a
1292 Evaluation order of the translation rules is dependent on the type
1293 of the translation rules and of the direction of a packet.
1295 rules are always evaluated first.
1298 rules are evaluated on an inbound packet or the
1300 rules on an outbound packet.
1301 Rules of the same type are evaluated in the same order in which they
1302 appear in the ruleset.
1303 The first matching rule decides what action is taken.
1307 option prefixed to a translation rule causes packets to remain untranslated,
1308 much in the same way as
1310 works in the packet filter (see below).
1311 If no rule matches the packet it is passed to the filter engine unmodified.
1313 Translation rules apply only to packets that pass through
1314 the specified interface, and if no interface is specified,
1315 translation is applied to packets on all interfaces.
1316 For instance, redirecting port 80 on an external interface to an internal
1317 web server will only work for connections originating from the outside.
1318 Connections to the address of the external interface from local hosts will
1319 not be redirected, since such packets do not actually pass through the
1321 Redirections cannot reflect packets back through the interface they arrive
1322 on, they can only be redirected to hosts connected to different interfaces
1323 or to the firewall itself.
1325 Note that redirecting external incoming connections to the loopback
1327 .Bd -literal -offset indent
1328 rdr on ne3 inet proto tcp to port smtp -\*(Gt 127.0.0.1 port spamd
1331 will effectively allow an external host to connect to daemons
1332 bound solely to the loopback address, circumventing the traditional
1333 blocking of such connections on a real interface.
1334 Unless this effect is desired, any of the local non-loopback addresses
1335 should be used as redirection target instead, which allows external
1336 connections only to daemons bound to this address or not bound to
1340 .Sx TRANSLATION EXAMPLES
1342 .Sh PACKET FILTERING
1348 packets based on attributes of their layer 3 (see
1358 In addition, packets may also be
1359 assigned to queues for the purpose of bandwidth control.
1361 For each packet processed by the packet filter, the filter rules are
1362 evaluated in sequential order, from first to last.
1363 The last matching rule decides what action is taken.
1364 If no rule matches the packet, the default action is to pass
1367 The following actions can be used in the filter:
1368 .Bl -tag -width xxxx
1370 The packet is blocked.
1371 There are a number of ways in which a
1373 rule can behave when blocking a packet.
1374 The default behaviour is to
1376 packets silently, however this can be overridden or made
1377 explicit either globally, by setting the
1379 option, or on a per-rule basis with one of the following options:
1381 .Bl -tag -width xxxx -compact
1383 The packet is silently dropped.
1385 This applies only to
1387 packets, and issues a TCP RST which closes the
1391 This causes ICMP messages to be returned for packets which match the rule.
1392 By default this is an ICMP UNREACHABLE message, however this
1393 can be overridden by specifying a message as a code or number.
1395 This causes a TCP RST to be returned for
1397 packets and an ICMP UNREACHABLE for UDP and other packets.
1400 Options returning ICMP packets currently have no effect if
1404 as the code to support this feature has not yet been implemented.
1406 The simplest mechanism to block everything by default and only pass
1407 packets that match explicit rules is specify a first filter rule of:
1408 .Bd -literal -offset indent
1412 The packet is passed;
1413 state is created unless the
1415 option is specified.
1420 filters packets statefully; the first time a packet matches a
1422 rule, a state entry is created; for subsequent packets the filter checks
1423 whether the packet matches any state.
1424 If it does, the packet is passed without evaluation of any rules.
1425 After the connection is closed or times out, the state entry is automatically
1428 This has several advantages.
1429 For TCP connections, comparing a packet to a state involves checking
1430 its sequence numbers, as well as TCP timestamps if a
1431 .Ar scrub reassemble tcp
1432 rule applies to the connection.
1433 If these values are outside the narrow windows of expected
1434 values, the packet is dropped.
1435 This prevents spoofing attacks, such as when an attacker sends packets with
1436 a fake source address/port but does not know the connection's sequence
1440 knows how to match ICMP replies to states.
1442 .Bd -literal -offset indent
1443 pass out inet proto icmp all icmp-type echoreq
1446 allows echo requests (such as those created by
1448 out statefully, and matches incoming echo replies correctly to states.
1450 Also, looking up states is usually faster than evaluating rules.
1451 If there are 50 rules, all of them are evaluated sequentially in O(n).
1452 Even with 50000 states, only 16 comparisons are needed to match a
1453 state, since states are stored in a binary search tree that allows
1454 searches in O(log2 n).
1456 Furthermore, correct handling of ICMP error messages is critical to
1457 many protocols, particularly TCP.
1459 matches ICMP error messages to the correct connection, checks them against
1460 connection parameters, and passes them if appropriate.
1461 For example if an ICMP source quench message referring to a stateful TCP
1462 connection arrives, it will be matched to the state and get passed.
1464 Finally, state tracking is required for
1465 .Ar nat , binat No and Ar rdr
1466 rules, in order to track address and port translations and reverse the
1467 translation on returning packets.
1470 will also create state for other protocols which are effectively stateless by
1472 UDP packets are matched to states using only host addresses and ports,
1473 and other protocols are matched to states using only the host addresses.
1475 If stateless filtering of individual packets is desired,
1478 keyword can be used to specify that state will not be created
1479 if this is the last matching rule.
1480 A number of parameters can also be set to affect how
1482 handles state tracking.
1484 .Sx STATEFUL TRACKING OPTIONS
1485 below for further details.
1487 The rule parameters specify the packets to which a rule applies.
1488 A packet always comes in on, or goes out through, one interface.
1489 Most parameters are optional.
1490 If a parameter is specified, the rule only applies to packets with
1491 matching attributes.
1492 Certain parameters can be expressed as lists, in which case
1494 generates all needed rule combinations.
1495 .Bl -tag -width xxxx
1496 .It Ar in No or Ar out
1497 This rule applies to incoming or outgoing packets.
1502 are specified, the rule will match packets in both directions.
1504 In addition to the action specified, a log message is generated.
1505 Only the packet that establishes the state is logged,
1508 option is specified.
1509 The logged packets are sent to a
1511 interface, by default
1513 This interface is monitored by the
1515 logging daemon, which dumps the logged packets to the file
1521 Used to force logging of all packets for a connection.
1522 This is not necessary when
1524 is explicitly specified.
1527 packets are logged to
1532 user ID of the user that owns the socket and the PID of the process that
1533 has the socket open where the packet is sourced from or destined to
1534 (depending on which socket is local).
1535 This is in addition to the normal information logged.
1536 .It Ar log (to Aq Ar interface )
1537 Send logs to the specified
1539 interface instead of
1542 If a packet matches a rule which has the
1544 option set, this rule
1545 is considered the last matching rule, and evaluation of subsequent rules
1547 .It Ar on Aq Ar interface
1548 This rule applies only to packets coming in on, or going out through, this
1549 particular interface.
1551 This rule applies only to packets of this address family.
1552 Supported values are
1556 .It Ar proto Aq Ar protocol
1557 This rule applies only to packets of this protocol.
1558 Common protocols are
1564 For a list of all the protocol name to number mappings used by
1567 .Pa /etc/protocols .
1569 .Ar from Aq Ar source
1570 .Ar port Aq Ar source
1575 This rule applies only to packets with the specified source and destination
1576 addresses and ports.
1578 Addresses can be specified in CIDR notation (matching netblocks), as
1579 symbolic host names or interface names, or as any of the following keywords:
1581 .Bl -tag -width xxxxxxxxxxxxxx -compact
1584 .It Ar route Aq Ar label
1585 Any address whose associated route has label
1592 Any address which is not currently routable.
1594 Any source address that fails a unicast reverse path forwarding (URPF)
1595 check, i.e. packets coming in on an interface other than that which holds
1596 the route back to the packet's source address.
1598 Any address that matches the given table.
1601 Ranges of addresses are specified by using the
1605 .Dq 10.1.1.10 - 10.1.1.12
1606 means all addresses from 10.1.1.10 to 10.1.1.12,
1607 hence addresses 10.1.1.10, 10.1.1.11, and 10.1.1.12.
1609 Interface names can have modifiers appended:
1611 .Bl -tag -width xxxxxxxxxxxx -compact
1613 Translates to the network(s) attached to the interface.
1615 Translates to the interface's broadcast address(es).
1617 Translates to the point to point interface's peer address(es).
1619 Do not include interface aliases.
1622 Host names may also have the
1624 option appended to restrict the name resolution to the first of each
1625 v4 and v6 address found.
1627 Host name resolution and interface to address translation are done at
1629 When the address of an interface (or host name) changes (under DHCP or PPP,
1630 for instance), the ruleset must be reloaded for the change to be reflected
1632 Surrounding the interface name (and optional modifiers) in parentheses
1633 changes this behaviour.
1634 When the interface name is surrounded by parentheses, the rule is
1635 automatically updated whenever the interface changes its address.
1636 The ruleset does not need to be reloaded.
1637 This is especially useful with
1640 Ports can be specified either by number or by name.
1641 For example, port 80 can be specified as
1643 For a list of all port name to number mappings used by
1648 Ports and ranges of ports are specified by using these operators:
1649 .Bd -literal -offset indent
1653 \*(Le (less than or equal)
1654 \*(Gt (greater than)
1655 \*(Ge (greater than or equal)
1656 : (range including boundaries)
1657 \*(Gt\*(Lt (range excluding boundaries)
1658 \*(Lt\*(Gt (except range)
1665 are binary operators (they take two arguments).
1668 .It Ar port 2000:2004
1670 .Sq all ports \*(Ge 2000 and \*(Le 2004 ,
1671 hence ports 2000, 2001, 2002, 2003 and 2004.
1672 .It Ar port 2000 \*(Gt\*(Lt 2004
1674 .Sq all ports \*(Gt 2000 and \*(Lt 2004 ,
1675 hence ports 2001, 2002 and 2003.
1676 .It Ar port 2000 \*(Lt\*(Gt 2004
1678 .Sq all ports \*(Lt 2000 or \*(Gt 2004 ,
1679 hence ports 1-1999 and 2005-65535.
1682 The operating system of the source host can be specified in the case of TCP
1687 .Sx OPERATING SYSTEM FINGERPRINTING
1688 section for more information.
1690 The host, port and OS specifications are optional, as in the following examples:
1691 .Bd -literal -offset indent
1693 pass in from any to any
1694 pass in proto tcp from any port \*(Le 1024 to any
1695 pass in proto tcp from any to any port 25
1696 pass in proto tcp from 10.0.0.0/8 port \*(Gt 1024 \e
1697 to ! 10.1.2.3 port != ssh
1698 pass in proto tcp from any os "OpenBSD"
1699 pass in proto tcp from route "DTAG"
1702 This is equivalent to "from any to any".
1703 .It Ar group Aq Ar group
1706 this rule only applies to packets of sockets owned by the specified group.
1707 .It Ar user Aq Ar user
1708 This rule only applies to packets of sockets owned by the specified user.
1709 For outgoing connections initiated from the firewall, this is the user
1710 that opened the connection.
1711 For incoming connections to the firewall itself, this is the user that
1712 listens on the destination port.
1713 For forwarded connections, where the firewall is not a connection endpoint,
1714 the user and group are
1717 All packets, both outgoing and incoming, of one connection are associated
1718 with the same user and group.
1719 Only TCP and UDP packets can be associated with users; for other protocols
1720 these parameters are ignored.
1722 User and group refer to the effective (as opposed to the real) IDs, in
1723 case the socket is created by a setuid/setgid process.
1724 User and group IDs are stored when a socket is created;
1725 when a process creates a listening socket as root (for instance, by
1726 binding to a privileged port) and subsequently changes to another
1727 user ID (to drop privileges), the credentials will remain root.
1729 User and group IDs can be specified as either numbers or names.
1730 The syntax is similar to the one for ports.
1733 matches packets of forwarded connections.
1735 can only be used with the operators
1739 Other constructs like
1740 .Cm user \*(Ge unknown
1742 Forwarded packets with unknown user and group ID match only rules
1743 that explicitly compare against
1751 does not match forwarded packets.
1752 The following example allows only selected users to open outgoing
1754 .Bd -literal -offset indent
1755 block out proto { tcp, udp } all
1756 pass out proto { tcp, udp } all user { \*(Lt 1000, dhartmei }
1758 .It Xo Ar flags Aq Ar a
1760 .No \*(Ba / Ns Aq Ar b
1763 This rule only applies to TCP packets that have the flags
1767 Flags not specified in
1770 For stateful connections, the default is
1772 To indicate that flags should not be checked at all, specify
1774 The flags are: (F)IN, (S)YN, (R)ST, (P)USH, (A)CK, (U)RG, (E)CE, and C(W)R.
1778 The other flags are ignored.
1780 This is the default setting for stateful connections.
1781 Out of SYN and ACK, exactly SYN may be set.
1782 SYN, SYN+PSH and SYN+RST match, but SYN+ACK, ACK and ACK+RST do not.
1783 This is more restrictive than the previous example.
1785 If the first set is not specified, it defaults to none.
1786 All of SYN, FIN, RST and ACK must be unset.
1791 is applied by default (unless
1793 is specified), only the initial SYN packet of a TCP handshake will create
1794 a state for a TCP connection.
1795 It is possible to be less restrictive, and allow state creation from
1798 packets, by specifying
1802 to synchronize to existing connections, for instance
1803 if one flushes the state table.
1804 However, states created from such intermediate packets may be missing
1805 connection details such as the TCP window scaling factor.
1806 States which modify the packet flow, such as those affected by
1807 .Ar nat , binat No or Ar rdr
1809 .Ar modulate No or Ar synproxy state
1810 options, or scrubbed with
1812 will also not be recoverable from intermediate packets.
1813 Such connections will stall and time out.
1814 .It Xo Ar icmp-type Aq Ar type
1817 .It Xo Ar icmp6-type Aq Ar type
1820 This rule only applies to ICMP or ICMPv6 packets with the specified type
1822 Text names for ICMP types and codes are listed in
1826 This parameter is only valid for rules that cover protocols ICMP or
1828 The protocol and the ICMP type indicator
1835 .It Xo Ar tos Aq Ar string
1836 .No \*(Ba Aq Ar number
1838 This rule applies to packets with the specified
1847 or as either hex or decimal.
1849 For example, the following rules are identical:
1850 .Bd -literal -offset indent
1851 pass all tos lowdelay
1856 By default, IPv4 packets with IP options or IPv6 packets with routing
1857 extension headers are blocked.
1862 rule, packets that pass the filter based on that rule (last matching)
1863 do so even if they contain IP options or routing extension headers.
1864 For packets that match state, the rule that initially created the
1868 rule that is used when a packet does not match any rules does not
1870 .It Ar label Aq Ar string
1871 Adds a label (name) to the rule, which can be used to identify the rule.
1874 shows per-rule statistics for rules that have labels.
1876 The following macros can be used in labels:
1878 .Bl -tag -width $srcaddr -compact -offset indent
1882 The source IP address.
1884 The destination IP address.
1886 The source port specification.
1888 The destination port specification.
1896 .Bd -literal -offset indent
1897 ips = \&"{ 1.2.3.4, 1.2.3.5 }\&"
1898 pass in proto tcp from any to $ips \e
1899 port \*(Gt 1023 label \&"$dstaddr:$dstport\&"
1903 .Bd -literal -offset indent
1904 pass in inet proto tcp from any to 1.2.3.4 \e
1905 port \*(Gt 1023 label \&"1.2.3.4:\*(Gt1023\&"
1906 pass in inet proto tcp from any to 1.2.3.5 \e
1907 port \*(Gt 1023 label \&"1.2.3.5:\*(Gt1023\&"
1910 The macro expansion for the
1912 directive occurs only at configuration file parse time, not during runtime.
1913 .It Xo Ar queue Aq Ar queue
1914 .No \*(Ba ( Aq Ar queue ,
1917 Packets matching this rule will be assigned to the specified queue.
1918 If two queues are given, packets which have a
1922 and TCP ACKs with no data payload will be assigned to the second one.
1928 .Bd -literal -offset indent
1929 pass in proto tcp to port 25 queue mail
1930 pass in proto tcp to port 22 queue(ssh_bulk, ssh_prio)
1932 .It Ar tag Aq Ar string
1933 Packets matching this rule will be tagged with the
1935 The tag acts as an internal marker that can be used to
1936 identify these packets later on.
1937 This can be used, for example, to provide trust between
1938 interfaces and to determine if packets have been
1939 processed by translation rules.
1942 meaning that the packet will be tagged even if the rule
1943 is not the last matching rule.
1944 Further matching rules can replace the tag with a
1945 new one but will not remove a previously applied tag.
1946 A packet is only ever assigned one tag at a time.
1947 Packet tagging can be done during
1952 rules in addition to filter rules.
1953 Tags take the same macros as labels (see above).
1954 .It Ar tagged Aq Ar string
1955 Used with filter, translation or scrub rules
1956 to specify that packets must already
1957 be tagged with the given tag in order to match the rule.
1958 Inverse tag matching can also be done
1964 .It Ar rtable Aq Ar number
1965 Used to select an alternate routing table for the routing lookup.
1966 Only effective before the route lookup happened, i.e. when filtering inbound.
1967 .It Xo Ar divert-to Aq Ar host
1970 Used to redirect packets to a local socket bound to
1974 The packets will not be modified, so
1976 on the socket will return the original destination address of the packet.
1978 Used to receive replies for sockets that are bound to addresses
1979 which are not local to the machine.
1982 for information on how to bind these sockets.
1983 .It Ar probability Aq Ar number
1984 A probability attribute can be attached to a rule, with a value set between
1985 0 and 1, bounds not included.
1986 In that case, the rule will be honoured using the given probability value
1988 For example, the following rule will drop 20% of incoming ICMP packets:
1989 .Bd -literal -offset indent
1990 block in proto icmp probability 20%
1994 If a packet matches a rule with a route option set, the packet filter will
1995 route the packet according to the type of route option.
1996 When such a rule creates state, the route option is also applied to all
1997 packets matching the same connection.
1998 .Bl -tag -width xxxx
2002 option does a normal route lookup to find the next hop for the packet.
2006 option routes the packet to the specified interface with an optional address
2010 rule creates state, only packets that pass in the same direction as the
2011 filter rule specifies will be routed in this way.
2012 Packets passing in the opposite direction (replies) are not affected
2013 and are routed normally.
2017 option is similar to
2019 but routes packets that pass in the opposite direction (replies) to the
2020 specified interface.
2021 Opposite direction is only defined in the context of a state entry, and
2023 is useful only in rules that create state.
2024 It can be used on systems with multiple external connections to
2025 route all outgoing packets of a connection through the interface
2026 the incoming connection arrived through (symmetric routing enforcement).
2030 option creates a duplicate of the packet and routes it like
2032 The original packet gets routed as it normally would.
2039 rules, (as well as for the
2044 rule options) for which there is a single redirection address which has a
2045 subnet mask smaller than 32 for IPv4 or 128 for IPv6 (more than one IP
2046 address), a variety of different methods for assigning this address can be
2048 .Bl -tag -width xxxx
2052 option applies the network portion of the redirection address to the address
2053 to be modified (source with
2060 option selects an address at random within the defined block of addresses.
2064 option uses a hash of the source address to determine the redirection address,
2065 ensuring that the redirection address is always the same for a given source.
2066 An optional key can be specified after this keyword either in hex or as a
2069 randomly generates a key for source-hash every time the
2070 ruleset is reloaded.
2074 option loops through the redirection address(es).
2076 When more than one redirection address is specified,
2078 is the only permitted pool type.
2086 from modifying the source port on TCP and UDP packets.
2091 option can be specified to help ensure that multiple connections from the
2092 same source are mapped to the same redirection address.
2093 This option can be used with the
2098 Note that by default these associations are destroyed as soon as there are
2099 no longer states which refer to them; in order to make the mappings last
2100 beyond the lifetime of the states, increase the global options with
2101 .Ar set timeout src.track .
2103 .Sx STATEFUL TRACKING OPTIONS
2104 for more ways to control the source tracking.
2105 .Sh STATE MODULATION
2106 Much of the security derived from TCP is attributable to how well the
2107 initial sequence numbers (ISNs) are chosen.
2108 Some popular stack implementations choose
2110 poor ISNs and thus are normally susceptible to ISN prediction exploits.
2113 rule to a TCP connection,
2115 will create a high quality random sequence number for each connection
2120 directive implicitly keeps state on the rule and is
2121 only applicable to TCP connections.
2124 .Bd -literal -offset indent
2126 pass out proto tcp from any to any modulate state
2127 pass in proto tcp from any to any port 25 flags S/SFRA modulate state
2130 Note that modulated connections will not recover when the state table
2131 is lost (firewall reboot, flushing the state table, etc...).
2133 will not be able to infer a connection again after the state table flushes
2134 the connection's modulator.
2135 When the state is lost, the connection may be left dangling until the
2136 respective endpoints time out the connection.
2137 It is possible on a fast local network for the endpoints to start an ACK
2138 storm while trying to resynchronize after the loss of the modulator.
2141 settings (or a more strict equivalent) should be used on
2143 rules to prevent ACK storms.
2145 Note that alternative methods are available
2146 to prevent loss of the state table
2147 and allow for firewall failover.
2152 for further information.
2156 passes packets that are part of a
2158 handshake between the endpoints.
2161 option can be used to cause
2163 itself to complete the handshake with the active endpoint, perform a handshake
2164 with the passive endpoint, and then forward packets between the endpoints.
2166 No packets are sent to the passive endpoint before the active endpoint has
2167 completed the handshake, hence so-called SYN floods with spoofed source
2168 addresses will not reach the passive endpoint, as the sender can't complete the
2171 The proxy is transparent to both endpoints, they each see a single
2172 connection from/to the other endpoint.
2174 chooses random initial sequence numbers for both handshakes.
2175 Once the handshakes are completed, the sequence number modulators
2176 (see previous section) are used to translate further packets of the
2180 .Ar modulate state .
2190 .Bd -literal -offset indent
2191 pass in proto tcp from any to any port www synproxy state
2193 .Sh STATEFUL TRACKING OPTIONS
2194 A number of options related to stateful tracking can be applied on a
2200 support these options, and
2202 must be specified explicitly to apply options to a rule.
2204 .Bl -tag -width xxxx -compact
2205 .It Ar max Aq Ar number
2206 Limits the number of concurrent states the rule may create.
2207 When this limit is reached, further packets that would create
2208 state will not match this rule until existing states time out.
2210 Prevent state changes for states created by this rule from appearing on the
2213 .It Xo Aq Ar timeout
2216 Changes the timeout values used for states created by this rule.
2217 For a list of all valid timeout names, see
2221 Uses a sloppy TCP connection tracker that does not check sequence
2222 numbers at all, which makes insertion and ICMP teardown attacks way
2224 This is intended to be used in situations where one does not see all
2225 packets of a connection, e.g. in asymmetric routing situations.
2226 Cannot be used with modulate or synproxy state.
2229 Multiple options can be specified, separated by commas:
2230 .Bd -literal -offset indent
2231 pass in proto tcp from any to any \e
2232 port www keep state \e
2233 (max 100, source-track rule, max-src-nodes 75, \e
2234 max-src-states 3, tcp.established 60, tcp.closing 5)
2239 keyword is specified, the number of states per source IP is tracked.
2241 .Bl -tag -width xxxx -compact
2242 .It Ar source-track rule
2243 The maximum number of states created by this rule is limited by the rule's
2248 Only state entries created by this particular rule count toward the rule's
2250 .It Ar source-track global
2251 The number of states created by all rules that use this option is limited.
2252 Each rule can specify different
2256 options, however state entries created by any participating rule count towards
2257 each individual rule's limits.
2260 The following limits can be set:
2262 .Bl -tag -width xxxx -compact
2263 .It Ar max-src-nodes Aq Ar number
2264 Limits the maximum number of source addresses which can simultaneously
2265 have state table entries.
2266 .It Ar max-src-states Aq Ar number
2267 Limits the maximum number of simultaneous state entries that a single
2268 source address can create with this rule.
2270 Specify that mid-stream pickups are to be allowed.
2271 The default is to NOT allow mid-stream pickups and implies flags
2272 S/SA for TCP connections.
2273 If pickups are enabled, flags S/SA are not implied
2274 for TCP connections and state can be created for any packet.
2276 The implied flags parameters need not be specified in either case
2277 unless you explicitly wish to override them, which also allows
2278 you to roll-up several protocols into a single rule.
2280 Certain validations are disabled when mid-stream pickups occur.
2281 For example, the window scaling options are not known for
2282 TCP pickups and sequence space comparisons must be disabled.
2284 This does not effect state representing fully quantified
2285 connections (for which the SYN/SYN-ACK passed through the routing
2287 Those connections continue to be fully validated.
2289 Specify that mid-stream pickups are to be allowed, but unconditionally
2290 disables sequence space checks even if full state is available.
2292 Specify that mid-stream pickups are not to be allowed.
2294 default and this keyword does not normally need to be specified.
2295 However, if you are concerned about rule set portability then
2296 specifying this keyword will at least result in an error from
2298 if it doesn't understand the feature.
2299 TCP flags of S/SA are implied
2300 and do not need to explicitly specified.
2303 For stateful TCP connections, limits on established connections (connections
2304 which have completed the TCP 3-way handshake) can also be enforced
2307 .Bl -tag -width xxxx -compact
2308 .It Ar max-src-conn Aq Ar number
2309 Limits the maximum number of simultaneous TCP connections which have
2310 completed the 3-way handshake that a single host can make.
2311 .It Xo Ar max-src-conn-rate Aq Ar number
2314 Limit the rate of new connections over a time interval.
2315 The connection rate is an approximation calculated as a moving average.
2318 Because the 3-way handshake ensures that the source address is not being
2319 spoofed, more aggressive action can be taken based on these limits.
2321 .Ar overload Aq Ar table
2322 state option, source IP addresses which hit either of the limits on
2323 established connections will be added to the named table.
2324 This table can be used in the ruleset to block further activity from
2325 the offending host, redirect it to a tarpit process, or restrict its
2330 keyword kills all states created by the matching rule which originate
2331 from the host which exceeds these limits.
2334 modifier to the flush command kills all states originating from the
2335 offending host, regardless of which rule created the state.
2337 For example, the following rules will protect the webserver against
2338 hosts making more than 100 connections in 10 seconds.
2339 Any host which connects faster than this rate will have its address added
2342 table and have all states originating from it flushed.
2343 Any new packets arriving from this host will be dropped unconditionally
2345 .Bd -literal -offset indent
2346 block quick from \*(Ltbad_hosts\*(Gt
2347 pass in on $ext_if proto tcp to $webserver port www keep state \e
2348 (max-src-conn-rate 100/10, overload \*(Ltbad_hosts\*(Gt flush global)
2350 .Sh OPERATING SYSTEM FINGERPRINTING
2351 Passive OS Fingerprinting is a mechanism to inspect nuances of a TCP
2352 connection's initial SYN packet and guess at the host's operating system.
2353 Unfortunately these nuances are easily spoofed by an attacker so the
2354 fingerprint is not useful in making security decisions.
2355 But the fingerprint is typically accurate enough to make policy decisions
2358 The fingerprints may be specified by operating system class, by
2359 version, or by subtype/patchlevel.
2360 The class of an operating system is typically the vendor or genre
2366 The version of the oldest available
2368 release on the main FTP site
2369 would be 2.6 and the fingerprint would be written
2371 .Dl \&"OpenBSD 2.6\&"
2373 The subtype of an operating system is typically used to describe the
2374 patchlevel if that patch led to changes in the TCP stack behavior.
2377 the only subtype is for a fingerprint that was
2380 scrub option and would be specified as
2382 .Dl \&"OpenBSD 3.3 no-df\&"
2384 Fingerprints for most popular operating systems are provided by
2388 is running, a complete list of known operating system fingerprints may
2389 be listed by running:
2393 Filter rules can enforce policy at any level of operating system specification
2394 assuming a fingerprint is present.
2395 Policy could limit traffic to approved operating systems or even ban traffic
2396 from hosts that aren't at the latest service pack.
2400 class can also be used as the fingerprint which will match packets for
2401 which no operating system fingerprint is known.
2404 .Bd -literal -offset indent
2405 pass out proto tcp from any os OpenBSD
2406 block out proto tcp from any os Doors
2407 block out proto tcp from any os "Doors PT"
2408 block out proto tcp from any os "Doors PT SP3"
2409 block out from any os "unknown"
2410 pass on lo0 proto tcp from any os "OpenBSD 3.3 lo0"
2413 Operating system fingerprinting is limited only to the TCP SYN packet.
2414 This means that it will not work on other protocols and will not match
2415 a currently established connection.
2417 Caveat: operating system fingerprints are occasionally wrong.
2418 There are three problems: an attacker can trivially craft his packets to
2419 appear as any operating system he chooses;
2420 an operating system patch could change the stack behavior and no fingerprints
2421 will match it until the database is updated;
2422 and multiple operating systems may have the same fingerprint.
2423 .Sh BLOCKING SPOOFED TRAFFIC
2424 "Spoofing" is the faking of IP addresses, typically for malicious
2428 directive expands to a set of filter rules which will block all
2429 traffic with a source IP from the network(s) directly connected
2430 to the specified interface(s) from entering the system through
2431 any other interface.
2433 For example, the line
2434 .Bd -literal -offset indent
2439 .Bd -literal -offset indent
2440 block drop in on ! lo0 inet from 127.0.0.1/8 to any
2441 block drop in on ! lo0 inet6 from ::1 to any
2444 For non-loopback interfaces, there are additional rules to block incoming
2445 packets with a source IP address identical to the interface's IP(s).
2446 For example, assuming the interface wi0 had an IP address of 10.0.0.1 and a
2447 netmask of 255.255.255.0,
2449 .Bd -literal -offset indent
2450 antispoof for wi0 inet
2454 .Bd -literal -offset indent
2455 block drop in on ! wi0 inet from 10.0.0.0/24 to any
2456 block drop in inet from 10.0.0.1 to any
2459 Caveat: Rules created by the
2461 directive interfere with packets sent over loopback interfaces
2463 One should pass these explicitly.
2464 .Sh FRAGMENT HANDLING
2465 The size of IP datagrams (packets) can be significantly larger than the
2466 maximum transmission unit (MTU) of the network.
2467 In cases when it is necessary or more efficient to send such large packets,
2468 the large packet will be fragmented into many smaller packets that will each
2470 Unfortunately for a firewalling device, only the first logical fragment will
2471 contain the necessary header information for the subprotocol that allows
2473 to filter on things such as TCP ports or to perform NAT.
2477 rules as described in
2478 .Sx TRAFFIC NORMALIZATION
2479 above, there are three options for handling fragments in the packet filter.
2481 One alternative is to filter individual fragments with filter rules.
2484 rule applies to a fragment, it is passed to the filter.
2485 Filter rules with matching IP header parameters decide whether the
2486 fragment is passed or blocked, in the same way as complete packets
2488 Without reassembly, fragments can only be filtered based on IP header
2489 fields (source/destination address, protocol), since subprotocol header
2490 fields are not available (TCP/UDP port numbers, ICMP code/type).
2493 option can be used to restrict filter rules to apply only to
2494 fragments, but not complete packets.
2495 Filter rules without the
2497 option still apply to fragments, if they only specify IP header fields.
2498 For instance, the rule
2499 .Bd -literal -offset indent
2500 pass in proto tcp from any to any port 80
2503 never applies to a fragment, even if the fragment is part of a TCP
2504 packet with destination port 80, because without reassembly this information
2505 is not available for each fragment.
2506 This also means that fragments cannot create new or match existing
2507 state table entries, which makes stateful filtering and address
2508 translation (NAT, redirection) for fragments impossible.
2510 It's also possible to reassemble only certain fragments by specifying
2511 source or destination addresses or protocols as parameters in
2515 In most cases, the benefits of reassembly outweigh the additional
2516 memory cost, and it's recommended to use
2519 all fragments via the
2520 .Ar fragment reassemble
2523 The memory allocated for fragment caching can be limited using
2525 Once this limit is reached, fragments that would have to be cached
2526 are dropped until other entries time out.
2527 The timeout value can also be adjusted.
2529 Currently, only IPv4 fragments are supported and IPv6 fragments
2530 are blocked unconditionally.
2532 Besides the main ruleset,
2534 can load rulesets into
2539 is a container that can hold rules, address tables, and other anchors.
2543 has a name which specifies the path where
2545 can be used to access the anchor to perform operations on it, such as
2546 attaching child anchors to it or loading rules into it.
2547 Anchors may be nested, with components separated by
2549 characters, similar to how file system hierarchies are laid out.
2550 The main ruleset is actually the default anchor, so filter and
2551 translation rules, for example, may also be contained in any anchor.
2553 An anchor can reference another
2556 using the following kinds
2558 .Bl -tag -width xxxx
2559 .It Ar nat-anchor Aq Ar name
2562 rules in the specified
2564 .It Ar rdr-anchor Aq Ar name
2567 rules in the specified
2569 .It Ar binat-anchor Aq Ar name
2572 rules in the specified
2574 .It Ar anchor Aq Ar name
2575 Evaluates the filter rules in the specified
2577 .It Xo Ar load anchor
2581 Loads the rules from the specified file into the
2586 When evaluation of the main ruleset reaches an
2590 will proceed to evaluate all rules specified in that anchor.
2592 Matching filter and translation rules marked with the
2594 option are final and abort the evaluation of the rules in other
2595 anchors and the main ruleset.
2598 itself is marked with the
2601 ruleset evaluation will terminate when the anchor is exited if the packet is
2602 matched by any rule within the anchor.
2605 rules are evaluated relative to the anchor in which they are contained.
2608 rules specified in the main ruleset will reference anchor
2609 attachment points underneath the main ruleset, and
2611 rules specified in a file loaded from a
2613 rule will be attached under that anchor point.
2615 Rules may be contained in
2617 attachment points which do not contain any rules when the main ruleset
2618 is loaded, and later such anchors can be manipulated through
2620 without reloading the main ruleset or other anchors.
2622 .Bd -literal -offset indent
2624 block on $ext_if all
2626 pass out on $ext_if all
2627 pass in on $ext_if proto tcp from any \e
2628 to $ext_if port smtp
2631 blocks all packets on the external interface by default, then evaluates
2634 named "spam", and finally passes all outgoing connections and
2635 incoming connections to port 25.
2636 .Bd -literal -offset indent
2637 # echo \&"block in quick from 1.2.3.4 to any\&" \&| \e
2641 This loads a single rule into the
2643 which blocks all packets from a specific address.
2645 The anchor can also be populated by adding a
2650 .Bd -literal -offset indent
2652 load anchor spam from "/etc/pf-spam.conf"
2659 it will also load all the rules from the file
2660 .Pa /etc/pf-spam.conf
2665 rules can specify packet filtering parameters using the same syntax as
2667 When parameters are used, the
2669 rule is only evaluated for matching packets.
2670 This allows conditional evaluation of anchors, like:
2671 .Bd -literal -offset indent
2672 block on $ext_if all
2673 anchor spam proto tcp from any to any port smtp
2674 pass out on $ext_if all
2675 pass in on $ext_if proto tcp from any to $ext_if port smtp
2680 spam are only evaluated for
2682 packets with destination port 25.
2684 .Bd -literal -offset indent
2685 # echo \&"block in quick from 1.2.3.4 to any" \&| \e
2689 will only block connections from 1.2.3.4 to port 25.
2691 Anchors may end with the asterisk
2693 character, which signifies that all anchors attached at that point
2694 should be evaluated in the alphabetical ordering of their anchor name.
2696 .Bd -literal -offset indent
2700 will evaluate each rule in each anchor attached to the
2703 Note that it will only evaluate anchors that are directly attached to the
2705 anchor, and will not descend to evaluate anchors recursively.
2707 Since anchors are evaluated relative to the anchor in which they are
2708 contained, there is a mechanism for accessing the parent and ancestor
2709 anchors of a given anchor.
2710 Similar to file system path name resolution, if the sequence
2712 appears as an anchor path component, the parent anchor of the current
2713 anchor in the path evaluation at that point will become the new current
2715 As an example, consider the following:
2716 .Bd -literal -offset indent
2717 # echo ' anchor "spam/allowed" ' | pfctl -f -
2718 # echo -e ' anchor "../banned" \en pass' | \e
2719 pfctl -a spam/allowed -f -
2722 Evaluation of the main ruleset will lead into the
2724 anchor, which will evaluate the rules in the
2726 anchor, if any, before finally evaluating the
2732 can also be loaded inline in the ruleset within a brace ('{' '}') delimited
2734 Brace delimited blocks may contain rules or other brace-delimited blocks.
2735 When anchors are loaded this way the anchor name becomes optional.
2736 .Bd -literal -offset indent
2737 anchor "external" on egress {
2740 pass proto tcp from any to port { 25, 80, 443 }
2742 pass in proto tcp to any port 22
2746 Since the parser specification for anchor names is a string, any
2747 reference to an anchor name containing
2749 characters will require double quote
2751 characters around the anchor name.
2752 .Sh TRANSLATION EXAMPLES
2753 This example maps incoming requests on port 80 to port 8080, on
2754 which a daemon is running (because, for example, it is not run as root,
2755 and therefore lacks permission to bind to port 80).
2757 # use a macro for the interface name, so it can be changed easily
2760 # map daemon on 8080 to appear to be on 80
2761 rdr on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 port 8080
2766 modifier is given, packets matching the translation rule are passed without
2767 inspecting the filter rules:
2769 rdr pass on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 \e
2773 In the example below, vlan12 is configured as 192.168.168.1;
2774 the machine translates all packets coming from 192.168.168.0/24 to 204.92.77.111
2775 when they are going out any interface except vlan12.
2776 This has the net effect of making traffic from the 192.168.168.0/24
2777 network appear as though it is the Internet routable address
2778 204.92.77.111 to nodes behind any interface on the router except
2779 for the nodes on vlan12.
2780 (Thus, 192.168.168.1 can talk to the 192.168.168.0/24 nodes.)
2782 nat on ! vlan12 from 192.168.168.0/24 to any -\*(Gt 204.92.77.111
2785 In the example below, the machine sits between a fake internal 144.19.74.*
2786 network, and a routable external IP of 204.92.77.100.
2789 rule excludes protocol AH from being translated.
2792 no nat on $ext_if proto ah from 144.19.74.0/24 to any
2793 nat on $ext_if from 144.19.74.0/24 to any -\*(Gt 204.92.77.100
2796 In the example below, packets bound for one specific server, as well as those
2797 generated by the sysadmins are not proxied; all other connections are.
2800 no rdr on $int_if proto { tcp, udp } from any to $server port 80
2801 no rdr on $int_if proto { tcp, udp } from $sysadmins to any port 80
2802 rdr on $int_if proto { tcp, udp } from any to any port 80 -\*(Gt 127.0.0.1 \e
2806 This longer example uses both a NAT and a redirection.
2807 The external interface has the address 157.161.48.183.
2808 On localhost, we are running
2810 waiting for FTP sessions to be redirected to it.
2811 The three mandatory anchors for
2813 are omitted from this example; see the
2818 # Translate outgoing packets' source addresses (any protocol).
2819 # In this case, any address but the gateway's external address is mapped.
2820 nat on $ext_if inet from ! ($ext_if) to any -\*(Gt ($ext_if)
2823 # Map outgoing packets' source port to an assigned proxy port instead of
2824 # an arbitrary port.
2825 # In this case, proxy outgoing isakmp with port 500 on the gateway.
2826 nat on $ext_if inet proto udp from any port = isakmp to any -\*(Gt ($ext_if) \e
2830 # Translate outgoing packets' source address (any protocol).
2831 # Translate incoming packets' destination address to an internal machine
2833 binat on $ext_if from 10.1.2.150 to any -\*(Gt $ext_if
2836 # Translate incoming packets' destination addresses.
2837 # As an example, redirect a TCP and UDP port to an internal machine.
2838 rdr on $ext_if inet proto tcp from any to ($ext_if) port 8080 \e
2839 -\*(Gt 10.1.2.151 port 22
2840 rdr on $ext_if inet proto udp from any to ($ext_if) port 8080 \e
2841 -\*(Gt 10.1.2.151 port 53
2844 # Translate outgoing ftp control connections to send them to localhost
2845 # for proxying with ftp-proxy(8) running on port 8021.
2846 rdr on $int_if proto tcp from any to any port 21 -\*(Gt 127.0.0.1 port 8021
2849 In this example, a NAT gateway is set up to translate internal addresses
2850 using a pool of public addresses (192.0.2.16/28) and to redirect
2851 incoming web server connections to a group of web servers on the internal
2855 # Translate outgoing packets' source addresses using an address pool.
2856 # A given source address is always translated to the same pool address by
2857 # using the source-hash keyword.
2858 nat on $ext_if inet from any to any -\*(Gt 192.0.2.16/28 source-hash
2861 # Translate incoming web server connections to a group of web servers on
2862 # the internal network.
2863 rdr on $ext_if proto tcp from any to any port 80 \e
2864 -\*(Gt { 10.1.2.155, 10.1.2.160, 10.1.2.161 } round-robin
2868 # The external interface is kue0
2869 # (157.161.48.183, the only routable address)
2870 # and the private network is 10.0.0.0/8, for which we are doing NAT.
2872 # use a macro for the interface name, so it can be changed easily
2875 # normalize all incoming traffic
2876 scrub in on $ext_if all fragment reassemble
2878 # block and log everything by default
2879 block return log on $ext_if all
2881 # block anything coming from source we have no back routes for
2882 block in from no-route to any
2884 # block packets whose ingress interface does not match the one in
2885 # the route back to their source address
2886 block in from urpf-failed to any
2888 # block and log outgoing packets that do not have our address as source,
2889 # they are either spoofed or something is misconfigured (NAT disabled,
2890 # for instance), we want to be nice and do not send out garbage.
2891 block out log quick on $ext_if from ! 157.161.48.183 to any
2893 # silently drop broadcasts (cable modem noise)
2894 block in quick on $ext_if from any to 255.255.255.255
2896 # block and log incoming packets from reserved address space and invalid
2897 # addresses, they are either spoofed or misconfigured, we cannot reply to
2898 # them anyway (hence, no return-rst).
2899 block in log quick on $ext_if from { 10.0.0.0/8, 172.16.0.0/12, \e
2900 192.168.0.0/16, 255.255.255.255/32 } to any
2904 # pass out/in certain ICMP queries and keep state (ping)
2905 # state matching is done on host addresses and ICMP id (not type/code),
2906 # so replies (like 0/0 for 8/0) will match queries
2907 # ICMP error messages (which always refer to a TCP/UDP packet) are
2908 # handled by the TCP/UDP states
2909 pass on $ext_if inet proto icmp all icmp-type 8 code 0
2913 # pass out all UDP connections and keep state
2914 pass out on $ext_if proto udp all
2916 # pass in certain UDP connections and keep state (DNS)
2917 pass in on $ext_if proto udp from any to any port domain
2921 # pass out all TCP connections and modulate state
2922 pass out on $ext_if proto tcp all modulate state
2924 # pass in certain TCP connections and keep state (SSH, SMTP, DNS, IDENT)
2925 pass in on $ext_if proto tcp from any to any port { ssh, smtp, domain, \e
2928 # Do not allow Windows 9x SMTP connections since they are typically
2929 # a viral worm. Alternately we could limit these OSes to 1 connection each.
2930 block in on $ext_if proto tcp from any os {"Windows 95", "Windows 98"} \e
2934 # pass in/out all IPv6 traffic: note that we have to enable this in two
2935 # different ways, on both our physical interface and our tunnel
2936 pass quick on gif0 inet6
2937 pass quick on $ext_if proto ipv6
2939 # Using the pickup options to keep/modulate/synproxy state
2941 # no-pickups (default) Do not allow connections to be picked up in the
2942 # middle. Implies flags S/SA (the 'no-pickups' option need
2943 # not be specified, it is the default).
2945 # pickups Allow connections to be picked up in the middle, even if
2946 # no window scaling information is known. Such connections
2947 # will disable sequence space checks. Implies no flag
2950 # hash-only Do not fail packets on sequence space checks. Implies no
2951 # flag restrictions.
2953 pass in on $ext_if proto tcp ... keep state (no-pickups)
2954 pass in on $ext_if proto tcp ... keep state (pickups)
2955 pass in on $ext_if proto tcp ... keep state (hash-only)
2959 # three interfaces: $int_if, $ext_if, and $wifi_if (wireless). NAT is
2960 # being done on $ext_if for all outgoing packets. tag packets in on
2961 # $int_if and pass those tagged packets out on $ext_if. all other
2962 # outgoing packets (i.e., packets from the wireless network) are only
2963 # permitted to access port 80.
2965 pass in on $int_if from any to any tag INTNET
2966 pass in on $wifi_if from any to any
2968 block out on $ext_if from any to any
2969 pass out quick on $ext_if tagged INTNET
2970 pass out on $ext_if proto tcp from any to any port 80
2972 # tag incoming packets as they are redirected to spamd(8). use the tag
2973 # to pass those packets through the packet filter.
2975 rdr on $ext_if inet proto tcp from \*(Ltspammers\*(Gt to port smtp \e
2976 tag SPAMD -\*(Gt 127.0.0.1 port spamd
2979 pass in on $ext_if inet proto tcp tagged SPAMD
2986 line = option | pf-rule | nat-rule | binat-rule | rdr-rule |
2987 antispoof-rule | altq-rule | queue-rule | trans-anchors |
2988 anchor-rule | anchor-close | load-anchor | table-rule | include
2990 option = "set" ( [ "timeout" ( timeout | "{" timeout-list "}" ) ] |
2991 [ "ruleset-optimization" [ "none" | "basic" | "profile" ] ] |
2992 [ "optimization" [ "default" | "normal" |
2993 "high-latency" | "satellite" |
2994 "aggressive" | "conservative" ] ]
2995 [ "limit" ( limit-item | "{" limit-list "}" ) ] |
2996 [ "loginterface" ( interface-name | "none" ) ] |
2997 [ "block-policy" ( "drop" | "return" ) ] |
2998 [ "keep-policy" keep ] |
2999 [ "state-policy" ( "if-bound" | "floating" ) ]
3000 [ "require-order" ( "yes" | "no" ) ]
3001 [ "fingerprints" filename ] |
3002 [ "skip on" ( interface-name | "{" interface-list "}" ) ] |
3003 [ "debug" ( "none" | "urgent" | "misc" | "loud" ) ] )
3005 pf-rule = action [ "in" | "out" ]
3006 [ "log" [ "(" logopts ")"] ] [ "quick" ]
3007 [ "on" ifspec ] [ "fastroute" | route ] [ af ] [ protospec ]
3008 hosts [ filteropt-list ]
3010 logopts = logopt [ "," logopts ]
3011 logopt = "all" | "user" | "to" interface-name
3013 filteropt-list = filteropt-list filteropt | filteropt
3014 filteropt = user | group | flags | icmp-type | icmp6-type | tos |
3015 keep | "fragment" | "no-df" | "min-ttl" number |
3016 "max-mss" number | "random-id" | "reassemble tcp" |
3017 fragmentation | "allow-opts" |
3018 "label" string | "tag" string | [ ! ] "tagged" string |
3019 "queue" ( string | "(" string [ [ "," ] string ] ")" ) |
3020 "probability" number"%"
3022 keep = "no" "state" |
3023 ( "keep" | "modulate" | "synproxy" ) "state"
3024 [ "(" state-opts ")" ]
3026 nat-rule = [ "no" ] "nat" [ "pass" [ "log" [ "(" logopts ")" ] ] ]
3027 [ "on" ifspec ] [ af ]
3028 [ protospec ] hosts [ "tag" string ] [ "tagged" string ]
3029 [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" )
3030 [ portspec ] [ pooltype ] [ "static-port" ] ]
3032 binat-rule = [ "no" ] "binat" [ "pass" [ "log" [ "(" logopts ")" ] ] ]
3033 [ "on" interface-name ] [ af ]
3034 [ "proto" ( proto-name | proto-number ) ]
3035 "from" address [ "/" mask-bits ] "to" ipspec
3036 [ "tag" string ] [ "tagged" string ]
3037 [ "-\*(Gt" address [ "/" mask-bits ] ]
3039 rdr-rule = [ "no" ] "rdr" [ "pass" [ "log" [ "(" logopts ")" ] ] ]
3040 [ "on" ifspec ] [ af ]
3041 [ protospec ] hosts [ "tag" string ] [ "tagged" string ]
3042 [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" )
3043 [ portspec ] [ pooltype ] ]
3045 antispoof-rule = "antispoof" [ "log" ] [ "quick" ]
3046 "for" ( interface-name | "{" interface-list "}" )
3047 [ af ] [ "label" string ]
3049 table-rule = "table" "\*(Lt" string "\*(Gt" [ tableopts-list ]
3050 tableopts-list = tableopts-list tableopts | tableopts
3051 tableopts = "persist" | "const" | "counters" | "file" string |
3052 "{" [ tableaddr-list ] "}"
3053 tableaddr-list = tableaddr-list [ "," ] tableaddr-spec | tableaddr-spec
3054 tableaddr-spec = [ "!" ] tableaddr [ "/" mask-bits ]
3055 tableaddr = hostname | ipv4-dotted-quad | ipv6-coloned-hex |
3056 interface-name | "self"
3058 altq-rule = "altq on" interface-name queueopts-list
3060 queue-rule = "queue" string [ "on" interface-name ] queueopts-list
3063 anchor-rule = "anchor" string [ "in" | "out" ] [ "on" ifspec ]
3064 [ af ] [ "proto" ] [ protospec ] [ hosts ]
3066 trans-anchors = ( "nat-anchor" | "rdr-anchor" | "binat-anchor" ) string
3067 [ "on" ifspec ] [ af ] [ "proto" ] [ protospec ] [ hosts ]
3069 load-anchor = "load anchor" string "from" filename
3071 queueopts-list = queueopts-list queueopts | queueopts
3072 queueopts = "bandwidth" bandwidth-spec |
3073 "qlimit" number | "tbrsize" number |
3074 "priority" number | schedulers
3075 schedulers = cbq-def | hfsc-def | priq-def | fairq-def
3076 bandwidth-spec = "number" ( "b" | "Kb" | "Mb" | "Gb" | "%" )
3078 action = "pass" | "block" [ return ] | [ "no" ] "scrub"
3079 return = "drop" | "return" | "return-rst" [ "( ttl" number ")" ] |
3080 "return-icmp" [ "(" icmpcode [ [ "," ] icmp6code ] ")" ] |
3081 "return-icmp6" [ "(" icmp6code ")" ]
3082 icmpcode = icmp-code-name | icmp-code-number
3083 icmp6code = icmp6-code-name | icmp6-code-number
3085 ifspec = ( [ "!" ] interface-name ) | "{" interface-list "}"
3086 interface-list = [ "!" ] interface-name [ [ "," ] interface-list ]
3087 route = ( "route-to" | "reply-to" | "dup-to" )
3088 ( routehost | "{" routehost-list "}" )
3090 af = "inet" | "inet6"
3092 protospec = "proto" ( proto-name | proto-number |
3093 "{" proto-list "}" )
3094 proto-list = ( proto-name | proto-number ) [ [ "," ] proto-list ]
3097 "from" ( "any" | "no-route" | "urpf-failed" | "self" | host |
3098 "{" host-list "}" | "route" string ) [ port ] [ os ]
3099 "to" ( "any" | "no-route" | "self" | host |
3100 "{" host-list "}" | "route" string ) [ port ]
3102 ipspec = "any" | host | "{" host-list "}"
3103 host = [ "!" ] ( address [ "/" mask-bits ] | "\*(Lt" string "\*(Gt" )
3104 redirhost = address [ "/" mask-bits ]
3105 routehost = "(" interface-name [ address [ "/" mask-bits ] ] ")"
3106 address = interface-name | "(" interface-name ")" | hostname |
3107 ipv4-dotted-quad | ipv6-coloned-hex
3108 host-list = host [ [ "," ] host-list ]
3109 redirhost-list = redirhost [ [ "," ] redirhost-list ]
3110 routehost-list = routehost [ [ "," ] routehost-list ]
3112 port = "port" ( unary-op | binary-op | "{" op-list "}" )
3113 portspec = "port" ( number | name ) [ ":" ( "*" | number | name ) ]
3114 os = "os" ( os-name | "{" os-list "}" )
3115 user = "user" ( unary-op | binary-op | "{" op-list "}" )
3116 group = "group" ( unary-op | binary-op | "{" op-list "}" )
3118 unary-op = [ "=" | "!=" | "\*(Lt" | "\*(Le" | "\*(Gt" | "\*(Ge" ]
3120 binary-op = number ( "\*(Lt\*(Gt" | "\*(Gt\*(Lt" | ":" ) number
3121 op-list = ( unary-op | binary-op ) [ [ "," ] op-list ]
3123 os-name = operating-system-name
3124 os-list = os-name [ [ "," ] os-list ]
3126 flags = "flags" ( [ flag-set ] "/" flag-set | "any" )
3127 flag-set = [ "F" ] [ "S" ] [ "R" ] [ "P" ] [ "A" ] [ "U" ] [ "E" ]
3130 icmp-type = "icmp-type" ( icmp-type-code | "{" icmp-list "}" )
3131 icmp6-type = "icmp6-type" ( icmp-type-code | "{" icmp-list "}" )
3132 icmp-type-code = ( icmp-type-name | icmp-type-number )
3133 [ "code" ( icmp-code-name | icmp-code-number ) ]
3134 icmp-list = icmp-type-code [ [ "," ] icmp-list ]
3136 tos = ( "lowdelay" | "throughput" | "reliability" |
3139 state-opts = state-opt [ [ "," ] state-opts ]
3140 state-opt = "max" number | "no-sync" | timeout |
3141 "source-track" [ "rule" | "global" ] |
3142 "max-src-nodes" number | "max-src-states" number |
3143 "max-src-conn" number |
3144 "max-src-conn-rate" number "/" number |
3145 "overload" "\*(Lt" string "\*(Gt" [ "flush" ] |
3146 "if-bound" | "floating" |
3147 "pickups" | "no-pickups" | "hash-only"
3149 fragmentation = [ "fragment reassemble" | "fragment crop" |
3150 "fragment drop-ovl" ]
3152 timeout-list = timeout [ [ "," ] timeout-list ]
3153 timeout = ( "tcp.first" | "tcp.opening" | "tcp.established" |
3154 "tcp.closing" | "tcp.finwait" | "tcp.closed" |
3155 "udp.first" | "udp.single" | "udp.multiple" |
3156 "icmp.first" | "icmp.error" |
3157 "other.first" | "other.single" | "other.multiple" |
3158 "frag" | "interval" | "src.track" |
3159 "adaptive.start" | "adaptive.end" ) number
3161 limit-list = limit-item [ [ "," ] limit-list ]
3162 limit-item = ( "states" | "frags" | "src-nodes" ) number
3164 pooltype = ( "bitmask" | "random" |
3165 "source-hash" [ hex-key | string-key ] |
3166 "round-robin" ) [ sticky-address ]
3168 subqueue = string | "{" queue-list "}"
3169 queue-list = string [ [ "," ] string ]
3171 cbq-def = "cbq" [ "(" cbq-opts ")" ]
3172 priq-def = "priq" [ "(" priq-opts ")" ]
3173 hfsc-def = "hfsc" [ "(" hfsc-opts ")" ]
3174 fairq-def = "fairq" [ "(" fairq-opts ")" ]
3176 cbq-opts = cbq-opt [ [ "," ] cbq-opts ]
3177 priq-opts = priq-opt [ [ "," ] priq-opts ]
3178 hfsc-opts = hfsc-opt [ [ "," ] hfsc-opts ]
3179 fairq-opts = fairq-opt [ [ "," ] fairq-opts ]
3181 cbq-opt = "default" | "borrow" | "red" | "ecn" | "rio"
3182 priq-opt = "default" | "red" | "ecn" | "rio"
3183 hfsc-opt = "default" | "red" | "ecn" | "rio" |
3184 linkshare-sc | realtime-sc | upperlimit-sc
3185 fairq-opt = "default" | "red" | "ecn" | "rio" |
3186 "buckets" number | "hogs" number | linkshare-sc
3188 linkshare-sc = "linkshare" sc-spec
3189 realtime-sc = "realtime" sc-spec
3190 upperlimit-sc = "upperlimit" sc-spec
3191 sc-spec = ( bandwidth-spec |
3192 "(" bandwidth-spec number bandwidth-spec ")" )
3193 include = "include" filename
3196 .Bl -tag -width ".Pa /usr/share/examples/pf" -compact
3200 Default location of the ruleset file.
3202 Default location of OS fingerprints.
3203 .It Pa /etc/protocols
3204 Protocol name database.
3205 .It Pa /etc/services
3206 Service name database.
3207 .It Pa /usr/share/examples/pf
3232 file format first appeared in