1.\" $FreeBSD$ 2.\" $OpenBSD: pf.conf.5,v 1.406 2009/01/31 19:37:12 sobrado Exp $ 3.\" 4.\" Copyright (c) 2002, Daniel Hartmeier 5.\" All rights reserved. 6.\" 7.\" Redistribution and use in source and binary forms, with or without 8.\" modification, are permitted provided that the following conditions 9.\" are met: 10.\" 11.\" - Redistributions of source code must retain the above copyright 12.\" notice, this list of conditions and the following disclaimer. 13.\" - Redistributions in binary form must reproduce the above 14.\" copyright notice, this list of conditions and the following 15.\" disclaimer in the documentation and/or other materials provided 16.\" with the distribution. 17.\" 18.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 19.\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 20.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 21.\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 22.\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 23.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, 24.\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 25.\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 26.\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN 28.\" ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29.\" POSSIBILITY OF SUCH DAMAGE. 30.\" 31.Dd September 25, 2021 32.Dt PF.CONF 5 33.Os 34.Sh NAME 35.Nm pf.conf 36.Nd packet filter configuration file 37.Sh DESCRIPTION 38The 39.Xr pf 4 40packet filter modifies, drops or passes packets according to rules or 41definitions specified in 42.Nm pf.conf . 43.Sh STATEMENT ORDER 44There are seven types of statements in 45.Nm pf.conf : 46.Bl -tag -width xxxx 47.It Cm Macros 48User-defined variables may be defined and used later, simplifying 49the configuration file. 50Macros must be defined before they are referenced in 51.Nm pf.conf . 52.It Cm Tables 53Tables provide a mechanism for increasing the performance and flexibility of 54rules with large numbers of source or destination addresses. 55.It Cm Options 56Options tune the behaviour of the packet filtering engine. 57.It Cm Traffic Normalization Li (e.g. Em scrub ) 58Traffic normalization protects internal machines against inconsistencies 59in Internet protocols and implementations. 60.It Cm Queueing 61Queueing provides rule-based bandwidth control. 62.It Cm Translation Li (Various forms of NAT) 63Translation rules specify how addresses are to be mapped or redirected to 64other addresses. 65.It Cm Packet Filtering 66Packet filtering provides rule-based blocking or passing of packets. 67.El 68.Pp 69With the exception of 70.Cm macros 71and 72.Cm tables , 73the types of statements should be grouped and appear in 74.Nm pf.conf 75in the order shown above, as this matches the operation of the underlying 76packet filtering engine. 77By default 78.Xr pfctl 8 79enforces this order (see 80.Ar set require-order 81below). 82.Pp 83Comments can be put anywhere in the file using a hash mark 84.Pq Sq # , 85and extend to the end of the current line. 86.Pp 87Additional configuration files can be included with the 88.Ic include 89keyword, for example: 90.Bd -literal -offset indent 91include "/etc/pf/sub.filter.conf" 92.Ed 93.Sh MACROS 94Macros can be defined that will later be expanded in context. 95Macro names must start with a letter, and may contain letters, digits 96and underscores. 97Macro names may not be reserved words (for example 98.Ar pass , 99.Ar in , 100.Ar out ) . 101Macros are not expanded inside quotes. 102.Pp 103For example, 104.Bd -literal -offset indent 105ext_if = \&"kue0\&" 106all_ifs = \&"{\&" $ext_if lo0 \&"}\&" 107pass out on $ext_if from any to any 108pass in on $ext_if proto tcp from any to any port 25 109.Ed 110.Sh TABLES 111Tables are named structures which can hold a collection of addresses and 112networks. 113Lookups against tables in 114.Xr pf 4 115are relatively fast, making a single rule with tables much more efficient, 116in terms of 117processor usage and memory consumption, than a large number of rules which 118differ only in IP address (either created explicitly or automatically by rule 119expansion). 120.Pp 121Tables can be used as the source or destination of filter rules, 122.Ar scrub 123rules 124or 125translation rules such as 126.Ar nat 127or 128.Ar rdr 129(see below for details on the various rule types). 130Tables can also be used for the redirect address of 131.Ar nat 132and 133.Ar rdr 134rules and in the routing options of filter rules, but only for 135.Ar round-robin 136pools. 137.Pp 138Tables can be defined with any of the following 139.Xr pfctl 8 140mechanisms. 141As with macros, reserved words may not be used as table names. 142.Bl -tag -width "manually" 143.It Ar manually 144Persistent tables can be manually created with the 145.Ar add 146or 147.Ar replace 148option of 149.Xr pfctl 8 , 150before or after the ruleset has been loaded. 151.It Pa pf.conf 152Table definitions can be placed directly in this file, and loaded at the 153same time as other rules are loaded, atomically. 154Table definitions inside 155.Nm pf.conf 156use the 157.Ar table 158statement, and are especially useful to define non-persistent tables. 159The contents of a pre-existing table defined without a list of addresses 160to initialize it is not altered when 161.Nm pf.conf 162is loaded. 163A table initialized with the empty list, 164.Li { } , 165will be cleared on load. 166.El 167.Pp 168Tables may be defined with the following attributes: 169.Bl -tag -width persist 170.It Ar persist 171The 172.Ar persist 173flag forces the kernel to keep the table even when no rules refer to it. 174If the flag is not set, the kernel will automatically remove the table 175when the last rule referring to it is flushed. 176.It Ar const 177The 178.Ar const 179flag prevents the user from altering the contents of the table once it 180has been created. 181Without that flag, 182.Xr pfctl 8 183can be used to add or remove addresses from the table at any time, even 184when running with 185.Xr securelevel 7 186= 2. 187.It Ar counters 188The 189.Ar counters 190flag enables per-address packet and byte counters which can be displayed with 191.Xr pfctl 8 . 192Note that this feature carries significant memory overhead for large tables. 193.El 194.Pp 195For example, 196.Bd -literal -offset indent 197table \*(Ltprivate\*(Gt const { 10/8, 172.16/12, 192.168/16 } 198table \*(Ltbadhosts\*(Gt persist 199block on fxp0 from { \*(Ltprivate\*(Gt, \*(Ltbadhosts\*(Gt } to any 200.Ed 201.Pp 202creates a table called private, to hold RFC 1918 private network 203blocks, and a table called badhosts, which is initially empty. 204A filter rule is set up to block all traffic coming from addresses listed in 205either table. 206The private table cannot have its contents changed and the badhosts table 207will exist even when no active filter rules reference it. 208Addresses may later be added to the badhosts table, so that traffic from 209these hosts can be blocked by using 210.Bd -literal -offset indent 211# pfctl -t badhosts -Tadd 204.92.77.111 212.Ed 213.Pp 214A table can also be initialized with an address list specified in one or more 215external files, using the following syntax: 216.Bd -literal -offset indent 217table \*(Ltspam\*(Gt persist file \&"/etc/spammers\&" file \&"/etc/openrelays\&" 218block on fxp0 from \*(Ltspam\*(Gt to any 219.Ed 220.Pp 221The files 222.Pa /etc/spammers 223and 224.Pa /etc/openrelays 225list IP addresses, one per line. 226Any lines beginning with a # are treated as comments and ignored. 227In addition to being specified by IP address, hosts may also be 228specified by their hostname. 229When the resolver is called to add a hostname to a table, 230.Em all 231resulting IPv4 and IPv6 addresses are placed into the table. 232IP addresses can also be entered in a table by specifying a valid interface 233name, a valid interface group or the 234.Em self 235keyword, in which case all addresses assigned to the interface(s) will be 236added to the table. 237.Sh OPTIONS 238.Xr pf 4 239may be tuned for various situations using the 240.Ar set 241command. 242.Bl -tag -width xxxx 243.It Ar set timeout 244.Pp 245.Bl -tag -width "src.track" -compact 246.It Ar interval 247Interval between purging expired states and fragments. 248.It Ar frag 249Seconds before an unassembled fragment is expired. 250.It Ar src.track 251Length of time to retain a source tracking entry after the last state 252expires. 253.El 254.Pp 255When a packet matches a stateful connection, the seconds to live for the 256connection will be updated to that of the 257.Ar proto.modifier 258which corresponds to the connection state. 259Each packet which matches this state will reset the TTL. 260Tuning these values may improve the performance of the 261firewall at the risk of dropping valid idle connections. 262.Pp 263.Bl -tag -width xxxx -compact 264.It Ar tcp.first 265The state after the first packet. 266.It Ar tcp.opening 267The state before the destination host ever sends a packet. 268.It Ar tcp.established 269The fully established state. 270.It Ar tcp.closing 271The state after the first FIN has been sent. 272.It Ar tcp.finwait 273The state after both FINs have been exchanged and the connection is closed. 274Some hosts (notably web servers on Solaris) send TCP packets even after closing 275the connection. 276Increasing 277.Ar tcp.finwait 278(and possibly 279.Ar tcp.closing ) 280can prevent blocking of such packets. 281.It Ar tcp.closed 282The state after one endpoint sends an RST. 283.El 284.Pp 285ICMP and UDP are handled in a fashion similar to TCP, but with a much more 286limited set of states: 287.Pp 288.Bl -tag -width xxxx -compact 289.It Ar udp.first 290The state after the first packet. 291.It Ar udp.single 292The state if the source host sends more than one packet but the destination 293host has never sent one back. 294.It Ar udp.multiple 295The state if both hosts have sent packets. 296.It Ar icmp.first 297The state after the first packet. 298.It Ar icmp.error 299The state after an ICMP error came back in response to an ICMP packet. 300.El 301.Pp 302Other protocols are handled similarly to UDP: 303.Pp 304.Bl -tag -width xxxx -compact 305.It Ar other.first 306.It Ar other.single 307.It Ar other.multiple 308.El 309.Pp 310Timeout values can be reduced adaptively as the number of state table 311entries grows. 312.Pp 313.Bl -tag -width xxxx -compact 314.It Ar adaptive.start 315When the number of state entries exceeds this value, adaptive scaling 316begins. 317All timeout values are scaled linearly with factor 318(adaptive.end - number of states) / (adaptive.end - adaptive.start). 319.It Ar adaptive.end 320When reaching this number of state entries, all timeout values become 321zero, effectively purging all state entries immediately. 322This value is used to define the scale factor, it should not actually 323be reached (set a lower state limit, see below). 324.El 325.Pp 326Adaptive timeouts are enabled by default, with an adaptive.start value 327equal to 60% of the state limit, and an adaptive.end value equal to 328120% of the state limit. 329They can be disabled by setting both adaptive.start and adaptive.end to 0. 330.Pp 331The adaptive timeout values can be defined both globally and for each rule. 332When used on a per-rule basis, the values relate to the number of 333states created by the rule, otherwise to the total number of 334states. 335.Pp 336For example: 337.Bd -literal -offset indent 338set timeout tcp.first 120 339set timeout tcp.established 86400 340set timeout { adaptive.start 6000, adaptive.end 12000 } 341set limit states 10000 342.Ed 343.Pp 344With 9000 state table entries, the timeout values are scaled to 50% 345(tcp.first 60, tcp.established 43200). 346.It Ar set loginterface 347Enable collection of packet and byte count statistics for the given 348interface or interface group. 349These statistics can be viewed using 350.Bd -literal -offset indent 351# pfctl -s info 352.Ed 353.Pp 354In this example 355.Xr pf 4 356collects statistics on the interface named dc0: 357.Bd -literal -offset indent 358set loginterface dc0 359.Ed 360.Pp 361One can disable the loginterface using: 362.Bd -literal -offset indent 363set loginterface none 364.Ed 365.It Ar set limit 366Sets hard limits on the memory pools used by the packet filter. 367See 368.Xr zone 9 369for an explanation of memory pools. 370.Pp 371For example, 372.Bd -literal -offset indent 373set limit states 20000 374.Ed 375.Pp 376sets the maximum number of entries in the memory pool used by state table 377entries (generated by 378.Ar pass 379rules which do not specify 380.Ar no state ) 381to 20000. 382Using 383.Bd -literal -offset indent 384set limit frags 20000 385.Ed 386.Pp 387sets the maximum number of entries in the memory pool used for fragment 388reassembly (generated by 389.Ar scrub 390rules) to 20000. 391Using 392.Bd -literal -offset indent 393set limit src-nodes 2000 394.Ed 395.Pp 396sets the maximum number of entries in the memory pool used for tracking 397source IP addresses (generated by the 398.Ar sticky-address 399and 400.Ar src.track 401options) to 2000. 402Using 403.Bd -literal -offset indent 404set limit tables 1000 405set limit table-entries 100000 406.Ed 407.Pp 408sets limits on the memory pools used by tables. 409The first limits the number of tables that can exist to 1000. 410The second limits the overall number of addresses that can be stored 411in tables to 100000. 412.Pp 413Various limits can be combined on a single line: 414.Bd -literal -offset indent 415set limit { states 20000, frags 20000, src-nodes 2000 } 416.Ed 417.It Ar set ruleset-optimization 418.Bl -tag -width xxxxxxxx -compact 419.It Ar none 420Disable the ruleset optimizer. 421.It Ar basic 422Enable basic ruleset optimization. 423This is the default behaviour. 424Basic ruleset optimization does four things to improve the 425performance of ruleset evaluations: 426.Pp 427.Bl -enum -compact 428.It 429remove duplicate rules 430.It 431remove rules that are a subset of another rule 432.It 433combine multiple rules into a table when advantageous 434.It 435re-order the rules to improve evaluation performance 436.El 437.Pp 438.It Ar profile 439Uses the currently loaded ruleset as a feedback profile to tailor the 440ordering of quick rules to actual network traffic. 441.El 442.Pp 443It is important to note that the ruleset optimizer will modify the ruleset 444to improve performance. 445A side effect of the ruleset modification is that per-rule accounting 446statistics will have different meanings than before. 447If per-rule accounting is important for billing purposes or whatnot, 448either the ruleset optimizer should not be used or a label field should 449be added to all of the accounting rules to act as optimization barriers. 450.Pp 451Optimization can also be set as a command-line argument to 452.Xr pfctl 8 , 453overriding the settings in 454.Nm . 455.It Ar set optimization 456Optimize state timeouts for one of the following network environments: 457.Pp 458.Bl -tag -width xxxx -compact 459.It Ar normal 460A normal network environment. 461Suitable for almost all networks. 462.It Ar high-latency 463A high-latency environment (such as a satellite connection). 464.It Ar satellite 465Alias for 466.Ar high-latency . 467.It Ar aggressive 468Aggressively expire connections. 469This can greatly reduce the memory usage of the firewall at the cost of 470dropping idle connections early. 471.It Ar conservative 472Extremely conservative settings. 473Avoid dropping legitimate connections at the 474expense of greater memory utilization (possibly much greater on a busy 475network) and slightly increased processor utilization. 476.El 477.Pp 478For example: 479.Bd -literal -offset indent 480set optimization aggressive 481.Ed 482.It Ar set block-policy 483The 484.Ar block-policy 485option sets the default behaviour for the packet 486.Ar block 487action: 488.Pp 489.Bl -tag -width xxxxxxxx -compact 490.It Ar drop 491Packet is silently dropped. 492.It Ar return 493A TCP RST is returned for blocked TCP packets, 494an ICMP UNREACHABLE is returned for blocked UDP packets, 495and all other packets are silently dropped. 496.El 497.Pp 498For example: 499.Bd -literal -offset indent 500set block-policy return 501.Ed 502.It Ar set fail-policy 503The 504.Ar fail-policy 505option sets the behaviour of rules which should pass a packet but were 506unable to do so. 507This might happen when a nat or route-to rule uses an empty table as list 508of targets or if a rule fails to create state or source node. 509The following 510.Ar block 511actions are possible: 512.Pp 513.Bl -tag -width xxxxxxxx -compact 514.It Ar drop 515Incoming packet is silently dropped. 516.It Ar return 517Incoming packet is dropped and TCP RST is returned for TCP packets, 518an ICMP UNREACHABLE is returned for UDP packets, 519and no response is sent for other packets. 520.El 521.Pp 522For example: 523.Bd -literal -offset indent 524set fail-policy return 525.Ed 526.It Ar set state-policy 527The 528.Ar state-policy 529option sets the default behaviour for states: 530.Pp 531.Bl -tag -width group-bound -compact 532.It Ar if-bound 533States are bound to interface. 534.It Ar floating 535States can match packets on any interfaces (the default). 536.El 537.Pp 538For example: 539.Bd -literal -offset indent 540set state-policy if-bound 541.Ed 542.It Ar set syncookies never | always | adaptive 543When 544.Cm syncookies 545are active, pf will answer each incoming TCP SYN with a syncookie SYNACK, 546without allocating any resources. 547Upon reception of the client's ACK in response to the syncookie 548SYNACK, pf will evaluate the ruleset and create state if the ruleset 549permits it, complete the three way handshake with the target host and 550continue the connection with synproxy in place. 551This allows pf to be resilient against large synflood attacks which would 552run the state table against its limits otherwise. 553Due to the blind answers to every incoming SYN syncookies share the caveats of 554synproxy, namely seemingly accepting connections that will be dropped later on. 555.Pp 556.Bl -tag -width adaptive -compact 557.It Cm never 558pf will never send syncookie SYNACKs (the default). 559.It Cm always 560pf will always send syncookie SYNACKs. 561.It Cm adaptive 562pf will enable syncookie mode when a given percentage of the state table 563is used up by half-open TCP connections, as in, those that saw the initial 564SYN but didn't finish the three way handshake. 565The thresholds for entering and leaving syncookie mode can be specified using 566.Bd -literal -offset indent 567set syncookies adaptive (start 25%, end 12%) 568.Ed 569.El 570.It Ar set state-defaults 571The 572.Ar state-defaults 573option sets the state options for states created from rules 574without an explicit 575.Ar keep state . 576For example: 577.Bd -literal -offset indent 578set state-defaults no-sync 579.Ed 580.It Ar set hostid 581The 32-bit 582.Ar hostid 583identifies this firewall's state table entries to other firewalls 584in a 585.Xr pfsync 4 586failover cluster. 587By default the hostid is set to a pseudo-random value, however it may be 588desirable to manually configure it, for example to more easily identify the 589source of state table entries. 590.Bd -literal -offset indent 591set hostid 1 592.Ed 593.Pp 594The hostid may be specified in either decimal or hexadecimal. 595.It Ar set require-order 596By default 597.Xr pfctl 8 598enforces an ordering of the statement types in the ruleset to: 599.Em options , 600.Em normalization , 601.Em queueing , 602.Em translation , 603.Em filtering . 604Setting this option to 605.Ar no 606disables this enforcement. 607There may be non-trivial and non-obvious implications to an out of 608order ruleset. 609Consider carefully before disabling the order enforcement. 610.It Ar set fingerprints 611Load fingerprints of known operating systems from the given filename. 612By default fingerprints of known operating systems are automatically 613loaded from 614.Xr pf.os 5 615in 616.Pa /etc 617but can be overridden via this option. 618Setting this option may leave a small period of time where the fingerprints 619referenced by the currently active ruleset are inconsistent until the new 620ruleset finishes loading. 621.Pp 622For example: 623.Pp 624.Dl set fingerprints \&"/etc/pf.os.devel\&" 625.It Ar set skip on Aq Ar ifspec 626List interfaces for which packets should not be filtered. 627Packets passing in or out on such interfaces are passed as if pf was 628disabled, i.e. pf does not process them in any way. 629This can be useful on loopback and other virtual interfaces, when 630packet filtering is not desired and can have unexpected effects. 631For example: 632.Pp 633.Dl set skip on lo0 634.It Ar set debug 635Set the debug 636.Ar level 637to one of the following: 638.Pp 639.Bl -tag -width xxxxxxxxxxxx -compact 640.It Ar none 641Don't generate debug messages. 642.It Ar urgent 643Generate debug messages only for serious errors. 644.It Ar misc 645Generate debug messages for various errors. 646.It Ar loud 647Generate debug messages for common conditions. 648.El 649.It Ar set keepcounters 650Preserve rule counters across rule updates. 651Usually rule counters are reset to zero on every update of the ruleset. 652With 653.Ar keepcounters 654set pf will attempt to find matching rules between old and new rulesets 655and preserve the rule counters. 656.El 657.Sh TRAFFIC NORMALIZATION 658Traffic normalization is used to sanitize packet content in such 659a way that there are no ambiguities in packet interpretation on 660the receiving side. 661The normalizer does IP fragment reassembly to prevent attacks 662that confuse intrusion detection systems by sending overlapping 663IP fragments. 664Packet normalization is invoked with the 665.Ar scrub 666directive. 667.Pp 668.Ar scrub 669has the following options: 670.Bl -tag -width xxxx 671.It Ar no-df 672Clears the 673.Ar dont-fragment 674bit from a matching IP packet. 675Some operating systems are known to generate fragmented packets with the 676.Ar dont-fragment 677bit set. 678This is particularly true with NFS. 679.Ar Scrub 680will drop such fragmented 681.Ar dont-fragment 682packets unless 683.Ar no-df 684is specified. 685.Pp 686Unfortunately some operating systems also generate their 687.Ar dont-fragment 688packets with a zero IP identification field. 689Clearing the 690.Ar dont-fragment 691bit on packets with a zero IP ID may cause deleterious results if an 692upstream router later fragments the packet. 693Using the 694.Ar random-id 695modifier (see below) is recommended in combination with the 696.Ar no-df 697modifier to ensure unique IP identifiers. 698.It Ar min-ttl Aq Ar number 699Enforces a minimum TTL for matching IP packets. 700.It Ar max-mss Aq Ar number 701Enforces a maximum MSS for matching TCP packets. 702.It Xo Ar set-tos Aq Ar string 703.No \*(Ba Aq Ar number 704.Xc 705Enforces a 706.Em TOS 707for matching IP packets. 708.Em TOS 709may be 710given as one of 711.Ar critical , 712.Ar inetcontrol , 713.Ar lowdelay , 714.Ar netcontrol , 715.Ar throughput , 716.Ar reliability , 717or one of the DiffServ Code Points: 718.Ar ef , 719.Ar va , 720.Ar af11 No ... Ar af43 , 721.Ar cs0 No ... Ar cs7 ; 722or as either hex or decimal. 723.It Ar random-id 724Replaces the IP identification field with random values to compensate 725for predictable values generated by many hosts. 726This option only applies to packets that are not fragmented 727after the optional fragment reassembly. 728.It Ar fragment reassemble 729Using 730.Ar scrub 731rules, fragments can be reassembled by normalization. 732In this case, fragments are buffered until they form a complete 733packet, and only the completed packet is passed on to the filter. 734The advantage is that filter rules have to deal only with complete 735packets, and can ignore fragments. 736The drawback of caching fragments is the additional memory cost. 737.It Ar reassemble tcp 738Statefully normalizes TCP connections. 739.Ar scrub reassemble tcp 740rules may not have the direction (in/out) specified. 741.Ar reassemble tcp 742performs the following normalizations: 743.Pp 744.Bl -tag -width timeout -compact 745.It ttl 746Neither side of the connection is allowed to reduce their IP TTL. 747An attacker may send a packet such that it reaches the firewall, affects 748the firewall state, and expires before reaching the destination host. 749.Ar reassemble tcp 750will raise the TTL of all packets back up to the highest value seen on 751the connection. 752.It timestamp modulation 753Modern TCP stacks will send a timestamp on every TCP packet and echo 754the other endpoint's timestamp back to them. 755Many operating systems will merely start the timestamp at zero when 756first booted, and increment it several times a second. 757The uptime of the host can be deduced by reading the timestamp and multiplying 758by a constant. 759Also observing several different timestamps can be used to count hosts 760behind a NAT device. 761And spoofing TCP packets into a connection requires knowing or guessing 762valid timestamps. 763Timestamps merely need to be monotonically increasing and not derived off a 764guessable base time. 765.Ar reassemble tcp 766will cause 767.Ar scrub 768to modulate the TCP timestamps with a random number. 769.It extended PAWS checks 770There is a problem with TCP on long fat pipes, in that a packet might get 771delayed for longer than it takes the connection to wrap its 32-bit sequence 772space. 773In such an occurrence, the old packet would be indistinguishable from a 774new packet and would be accepted as such. 775The solution to this is called PAWS: Protection Against Wrapped Sequence 776numbers. 777It protects against it by making sure the timestamp on each packet does 778not go backwards. 779.Ar reassemble tcp 780also makes sure the timestamp on the packet does not go forward more 781than the RFC allows. 782By doing this, 783.Xr pf 4 784artificially extends the security of TCP sequence numbers by 10 to 18 785bits when the host uses appropriately randomized timestamps, since a 786blind attacker would have to guess the timestamp as well. 787.El 788.El 789.Pp 790For example, 791.Bd -literal -offset indent 792scrub in on $ext_if all fragment reassemble 793.Ed 794.Pp 795The 796.Ar no 797option prefixed to a scrub rule causes matching packets to remain unscrubbed, 798much in the same way as 799.Ar drop quick 800works in the packet filter (see below). 801This mechanism should be used when it is necessary to exclude specific packets 802from broader scrub rules. 803.Sh QUEUEING with ALTQ 804The ALTQ system is currently not available in the GENERIC kernel nor as 805loadable modules. 806In order to use the herein after called queueing options one has to use a 807custom built kernel. 808Please refer to 809.Xr altq 4 810to learn about the related kernel options. 811.Pp 812Packets can be assigned to queues for the purpose of bandwidth 813control. 814At least two declarations are required to configure queues, and later 815any packet filtering rule can reference the defined queues by name. 816During the filtering component of 817.Nm pf.conf , 818the last referenced 819.Ar queue 820name is where any packets from 821.Ar pass 822rules will be queued, while for 823.Ar block 824rules it specifies where any resulting ICMP or TCP RST 825packets should be queued. 826The 827.Ar scheduler 828defines the algorithm used to decide which packets get delayed, dropped, or 829sent out immediately. 830There are three 831.Ar schedulers 832currently supported. 833.Bl -tag -width xxxx 834.It Ar cbq 835Class Based Queueing. 836.Ar Queues 837attached to an interface build a tree, thus each 838.Ar queue 839can have further child 840.Ar queues . 841Each queue can have a 842.Ar priority 843and a 844.Ar bandwidth 845assigned. 846.Ar Priority 847mainly controls the time packets take to get sent out, while 848.Ar bandwidth 849has primarily effects on throughput. 850.Ar cbq 851achieves both partitioning and sharing of link bandwidth 852by hierarchically structured classes. 853Each class has its own 854.Ar queue 855and is assigned its share of 856.Ar bandwidth . 857A child class can borrow bandwidth from its parent class 858as long as excess bandwidth is available 859(see the option 860.Ar borrow , 861below). 862.It Ar priq 863Priority Queueing. 864.Ar Queues 865are flat attached to the interface, thus, 866.Ar queues 867cannot have further child 868.Ar queues . 869Each 870.Ar queue 871has a unique 872.Ar priority 873assigned, ranging from 0 to 15. 874Packets in the 875.Ar queue 876with the highest 877.Ar priority 878are processed first. 879.It Ar hfsc 880Hierarchical Fair Service Curve. 881.Ar Queues 882attached to an interface build a tree, thus each 883.Ar queue 884can have further child 885.Ar queues . 886Each queue can have a 887.Ar priority 888and a 889.Ar bandwidth 890assigned. 891.Ar Priority 892mainly controls the time packets take to get sent out, while 893.Ar bandwidth 894primarily affects throughput. 895.Ar hfsc 896supports both link-sharing and guaranteed real-time services. 897It employs a service curve based QoS model, 898and its unique feature is an ability to decouple 899.Ar delay 900and 901.Ar bandwidth 902allocation. 903.El 904.Pp 905The interfaces on which queueing should be activated are declared using 906the 907.Ar altq on 908declaration. 909.Ar altq on 910has the following keywords: 911.Bl -tag -width xxxx 912.It Aq Ar interface 913Queueing is enabled on the named interface. 914.It Aq Ar scheduler 915Specifies which queueing scheduler to use. 916Currently supported values 917are 918.Ar cbq 919for Class Based Queueing, 920.Ar priq 921for Priority Queueing and 922.Ar hfsc 923for the Hierarchical Fair Service Curve scheduler. 924.It Ar bandwidth Aq Ar bw 925The maximum bitrate for all queues on an 926interface may be specified using the 927.Ar bandwidth 928keyword. 929The value can be specified as an absolute value or as a 930percentage of the interface bandwidth. 931When using an absolute value, the suffixes 932.Ar b , 933.Ar Kb , 934.Ar Mb , 935and 936.Ar Gb 937are used to represent bits, kilobits, megabits, and 938gigabits per second, respectively. 939The value must not exceed the interface bandwidth. 940If 941.Ar bandwidth 942is not specified, the interface bandwidth is used 943(but take note that some interfaces do not know their bandwidth, 944or can adapt their bandwidth rates). 945.It Ar qlimit Aq Ar limit 946The maximum number of packets held in the queue. 947The default is 50. 948.It Ar tbrsize Aq Ar size 949Adjusts the size, in bytes, of the token bucket regulator. 950If not specified, heuristics based on the 951interface bandwidth are used to determine the size. 952.It Ar queue Aq Ar list 953Defines a list of subqueues to create on an interface. 954.El 955.Pp 956In the following example, the interface dc0 957should queue up to 5Mbps in four second-level queues using 958Class Based Queueing. 959Those four queues will be shown in a later example. 960.Bd -literal -offset indent 961altq on dc0 cbq bandwidth 5Mb queue { std, http, mail, ssh } 962.Ed 963.Pp 964Once interfaces are activated for queueing using the 965.Ar altq 966directive, a sequence of 967.Ar queue 968directives may be defined. 969The name associated with a 970.Ar queue 971must match a queue defined in the 972.Ar altq 973directive (e.g. mail), or, except for the 974.Ar priq 975.Ar scheduler , 976in a parent 977.Ar queue 978declaration. 979The following keywords can be used: 980.Bl -tag -width xxxx 981.It Ar on Aq Ar interface 982Specifies the interface the queue operates on. 983If not given, it operates on all matching interfaces. 984.It Ar bandwidth Aq Ar bw 985Specifies the maximum bitrate to be processed by the queue. 986This value must not exceed the value of the parent 987.Ar queue 988and can be specified as an absolute value or a percentage of the parent 989queue's bandwidth. 990If not specified, defaults to 100% of the parent queue's bandwidth. 991The 992.Ar priq 993scheduler does not support bandwidth specification. 994.It Ar priority Aq Ar level 995Between queues a priority level can be set. 996For 997.Ar cbq 998and 999.Ar hfsc , 1000the range is 0 to 7 and for 1001.Ar priq , 1002the range is 0 to 15. 1003The default for all is 1. 1004.Ar Priq 1005queues with a higher priority are always served first. 1006.Ar Cbq 1007and 1008.Ar Hfsc 1009queues with a higher priority are preferred in the case of overload. 1010.It Ar qlimit Aq Ar limit 1011The maximum number of packets held in the queue. 1012The default is 50. 1013.El 1014.Pp 1015The 1016.Ar scheduler 1017can get additional parameters with 1018.Xo Aq Ar scheduler 1019.Pf ( Aq Ar parameters ) . 1020.Xc 1021Parameters are as follows: 1022.Bl -tag -width Fl 1023.It Ar default 1024Packets not matched by another queue are assigned to this one. 1025Exactly one default queue is required. 1026.It Ar red 1027Enable RED (Random Early Detection) on this queue. 1028RED drops packets with a probability proportional to the average 1029queue length. 1030.It Ar rio 1031Enables RIO on this queue. 1032RIO is RED with IN/OUT, thus running 1033RED two times more than RIO would achieve the same effect. 1034RIO is currently not supported in the GENERIC kernel. 1035.It Ar ecn 1036Enables ECN (Explicit Congestion Notification) on this queue. 1037ECN implies RED. 1038.El 1039.Pp 1040The 1041.Ar cbq 1042.Ar scheduler 1043supports an additional option: 1044.Bl -tag -width Fl 1045.It Ar borrow 1046The queue can borrow bandwidth from the parent. 1047.El 1048.Pp 1049The 1050.Ar hfsc 1051.Ar scheduler 1052supports some additional options: 1053.Bl -tag -width Fl 1054.It Ar realtime Aq Ar sc 1055The minimum required bandwidth for the queue. 1056.It Ar upperlimit Aq Ar sc 1057The maximum allowed bandwidth for the queue. 1058.It Ar linkshare Aq Ar sc 1059The bandwidth share of a backlogged queue. 1060.El 1061.Pp 1062.Aq Ar sc 1063is an acronym for 1064.Ar service curve . 1065.Pp 1066The format for service curve specifications is 1067.Ar ( m1 , d , m2 ) . 1068.Ar m2 1069controls the bandwidth assigned to the queue. 1070.Ar m1 1071and 1072.Ar d 1073are optional and can be used to control the initial bandwidth assignment. 1074For the first 1075.Ar d 1076milliseconds the queue gets the bandwidth given as 1077.Ar m1 , 1078afterwards the value given in 1079.Ar m2 . 1080.Pp 1081Furthermore, with 1082.Ar cbq 1083and 1084.Ar hfsc , 1085child queues can be specified as in an 1086.Ar altq 1087declaration, thus building a tree of queues using a part of 1088their parent's bandwidth. 1089.Pp 1090Packets can be assigned to queues based on filter rules by using the 1091.Ar queue 1092keyword. 1093Normally only one 1094.Ar queue 1095is specified; when a second one is specified it will instead be used for 1096packets which have a 1097.Em TOS 1098of 1099.Em lowdelay 1100and for TCP ACKs with no data payload. 1101.Pp 1102To continue the previous example, the examples below would specify the 1103four referenced 1104queues, plus a few child queues. 1105Interactive 1106.Xr ssh 1 1107sessions get priority over bulk transfers like 1108.Xr scp 1 1109and 1110.Xr sftp 1 . 1111The queues may then be referenced by filtering rules (see 1112.Sx PACKET FILTERING 1113below). 1114.Bd -literal 1115queue std bandwidth 10% cbq(default) 1116queue http bandwidth 60% priority 2 cbq(borrow red) \e 1117 { employees, developers } 1118queue developers bandwidth 75% cbq(borrow) 1119queue employees bandwidth 15% 1120queue mail bandwidth 10% priority 0 cbq(borrow ecn) 1121queue ssh bandwidth 20% cbq(borrow) { ssh_interactive, ssh_bulk } 1122queue ssh_interactive bandwidth 50% priority 7 cbq(borrow) 1123queue ssh_bulk bandwidth 50% priority 0 cbq(borrow) 1124 1125block return out on dc0 inet all queue std 1126pass out on dc0 inet proto tcp from $developerhosts to any port 80 \e 1127 queue developers 1128pass out on dc0 inet proto tcp from $employeehosts to any port 80 \e 1129 queue employees 1130pass out on dc0 inet proto tcp from any to any port 22 \e 1131 queue(ssh_bulk, ssh_interactive) 1132pass out on dc0 inet proto tcp from any to any port 25 \e 1133 queue mail 1134.Ed 1135.Sh QUEUEING with dummynet 1136Queueing can also be done with 1137.Xr dummynet 4 . 1138Queues and pipes can be created with 1139.Xr dnctl 8 . 1140.Pp 1141Packets can be assigned to queues and pipes using 1142.Ar dnqueue 1143and 1144.Ar dnpipe 1145respectively. 1146.Pp 1147Both 1148.Ar dnqueue 1149and 1150.Ar dnpipe 1151take either a single pipe or queue number or two numbers as arguments. 1152The first pipe or queue number will be used to shape the traffic in the rule 1153direction, the second will be used to shape the traffic in the reverse 1154direction. 1155If the rule does not specify a direction the first packet to create state will 1156be shaped according to the first number, and the response traffic according to 1157the second. 1158.Pp 1159If the 1160.Xr dummynet 4 1161module is not loaded any traffic sent into a queue or pipe will be dropped. 1162.Sh TRANSLATION 1163Translation rules modify either the source or destination address of the 1164packets associated with a stateful connection. 1165A stateful connection is automatically created to track packets matching 1166such a rule as long as they are not blocked by the filtering section of 1167.Nm pf.conf . 1168The translation engine modifies the specified address and/or port in the 1169packet, recalculates IP, TCP and UDP checksums as necessary, and passes it to 1170the packet filter for evaluation. 1171.Pp 1172Since translation occurs before filtering the filter 1173engine will see packets as they look after any 1174addresses and ports have been translated. 1175Filter rules will therefore have to filter based on the translated 1176address and port number. 1177Packets that match a translation rule are only automatically passed if 1178the 1179.Ar pass 1180modifier is given, otherwise they are 1181still subject to 1182.Ar block 1183and 1184.Ar pass 1185rules. 1186.Pp 1187The state entry created permits 1188.Xr pf 4 1189to keep track of the original address for traffic associated with that state 1190and correctly direct return traffic for that connection. 1191.Pp 1192Various types of translation are possible with pf: 1193.Bl -tag -width xxxx 1194.It Ar binat 1195A 1196.Ar binat 1197rule specifies a bidirectional mapping between an external IP netblock 1198and an internal IP netblock. 1199.It Ar nat 1200A 1201.Ar nat 1202rule specifies that IP addresses are to be changed as the packet 1203traverses the given interface. 1204This technique allows one or more IP addresses 1205on the translating host to support network traffic for a larger range of 1206machines on an "inside" network. 1207Although in theory any IP address can be used on the inside, it is strongly 1208recommended that one of the address ranges defined by RFC 1918 be used. 1209These netblocks are: 1210.Bd -literal 121110.0.0.0 - 10.255.255.255 (all of net 10, i.e., 10/8) 1212172.16.0.0 - 172.31.255.255 (i.e., 172.16/12) 1213192.168.0.0 - 192.168.255.255 (i.e., 192.168/16) 1214.Ed 1215.It Pa rdr 1216The packet is redirected to another destination and possibly a 1217different port. 1218.Ar rdr 1219rules can optionally specify port ranges instead of single ports. 1220rdr ... port 2000:2999 -\*(Gt ... port 4000 1221redirects ports 2000 to 2999 (inclusive) to port 4000. 1222rdr ... port 2000:2999 -\*(Gt ... port 4000:* 1223redirects port 2000 to 4000, 2001 to 4001, ..., 2999 to 4999. 1224.El 1225.Pp 1226In addition to modifying the address, some translation rules may modify 1227source or destination ports for 1228.Xr tcp 4 1229or 1230.Xr udp 4 1231connections; implicitly in the case of 1232.Ar nat 1233rules and explicitly in the case of 1234.Ar rdr 1235rules. 1236Port numbers are never translated with a 1237.Ar binat 1238rule. 1239.Pp 1240Evaluation order of the translation rules is dependent on the type 1241of the translation rules and of the direction of a packet. 1242.Ar binat 1243rules are always evaluated first. 1244Then either the 1245.Ar rdr 1246rules are evaluated on an inbound packet or the 1247.Ar nat 1248rules on an outbound packet. 1249Rules of the same type are evaluated in the same order in which they 1250appear in the ruleset. 1251The first matching rule decides what action is taken. 1252.Pp 1253The 1254.Ar no 1255option prefixed to a translation rule causes packets to remain untranslated, 1256much in the same way as 1257.Ar drop quick 1258works in the packet filter (see below). 1259If no rule matches the packet it is passed to the filter engine unmodified. 1260.Pp 1261Translation rules apply only to packets that pass through 1262the specified interface, and if no interface is specified, 1263translation is applied to packets on all interfaces. 1264For instance, redirecting port 80 on an external interface to an internal 1265web server will only work for connections originating from the outside. 1266Connections to the address of the external interface from local hosts will 1267not be redirected, since such packets do not actually pass through the 1268external interface. 1269Redirections cannot reflect packets back through the interface they arrive 1270on, they can only be redirected to hosts connected to different interfaces 1271or to the firewall itself. 1272.Pp 1273Note that redirecting external incoming connections to the loopback 1274address, as in 1275.Bd -literal -offset indent 1276rdr on ne3 inet proto tcp to port smtp -\*(Gt 127.0.0.1 port spamd 1277.Ed 1278.Pp 1279will effectively allow an external host to connect to daemons 1280bound solely to the loopback address, circumventing the traditional 1281blocking of such connections on a real interface. 1282Unless this effect is desired, any of the local non-loopback addresses 1283should be used as redirection target instead, which allows external 1284connections only to daemons bound to this address or not bound to 1285any address. 1286.Pp 1287See 1288.Sx TRANSLATION EXAMPLES 1289below. 1290.Sh PACKET FILTERING 1291.Xr pf 4 1292has the ability to 1293.Ar block 1294, 1295.Ar pass 1296and 1297.Ar match 1298packets based on attributes of their layer 3 (see 1299.Xr ip 4 1300and 1301.Xr ip6 4 ) 1302and layer 4 (see 1303.Xr icmp 4 , 1304.Xr icmp6 4 , 1305.Xr tcp 4 , 1306.Xr udp 4 ) 1307headers. 1308In addition, packets may also be 1309assigned to queues for the purpose of bandwidth control. 1310.Pp 1311For each packet processed by the packet filter, the filter rules are 1312evaluated in sequential order, from first to last. 1313For 1314.Ar block 1315and 1316.Ar pass 1317, the last matching rule decides what action is taken. 1318For 1319.Ar match 1320, rules are evaulated every time they match; the pass/block state of a packet 1321remains unchanged. 1322If no rule matches the packet, the default action is to pass 1323the packet. 1324.Pp 1325The following actions can be used in the filter: 1326.Bl -tag -width xxxx 1327.It Ar block 1328The packet is blocked. 1329There are a number of ways in which a 1330.Ar block 1331rule can behave when blocking a packet. 1332The default behaviour is to 1333.Ar drop 1334packets silently, however this can be overridden or made 1335explicit either globally, by setting the 1336.Ar block-policy 1337option, or on a per-rule basis with one of the following options: 1338.Pp 1339.Bl -tag -width xxxx -compact 1340.It Ar drop 1341The packet is silently dropped. 1342.It Ar return-rst 1343This applies only to 1344.Xr tcp 4 1345packets, and issues a TCP RST which closes the 1346connection. 1347.It Ar return-icmp 1348.It Ar return-icmp6 1349This causes ICMP messages to be returned for packets which match the rule. 1350By default this is an ICMP UNREACHABLE message, however this 1351can be overridden by specifying a message as a code or number. 1352.It Ar return 1353This causes a TCP RST to be returned for 1354.Xr tcp 4 1355packets and an ICMP UNREACHABLE for UDP and other packets. 1356.El 1357.Pp 1358Options returning ICMP packets currently have no effect if 1359.Xr pf 4 1360operates on a 1361.Xr if_bridge 4 , 1362as the code to support this feature has not yet been implemented. 1363.Pp 1364The simplest mechanism to block everything by default and only pass 1365packets that match explicit rules is specify a first filter rule of: 1366.Bd -literal -offset indent 1367block all 1368.Ed 1369.It Ar match 1370The packet is matched. 1371This mechanism is used to provide fine grained filtering without altering the 1372block/pass state of a packet. 1373.Ar match 1374rules differ from 1375.Ar block 1376and 1377.Ar pass 1378rules in that parameters are set every time a packet matches the rule, not only 1379on the last matching rule. 1380For the following parameters, this means that the parameter effectively becomes 1381"sticky" until explicitly overridden: 1382.Ar queue , 1383.Ar dnpipe , 1384.Ar dnqueue 1385. 1386.It Ar pass 1387The packet is passed; 1388state is created unless the 1389.Ar no state 1390option is specified. 1391.El 1392.Pp 1393By default 1394.Xr pf 4 1395filters packets statefully; the first time a packet matches a 1396.Ar pass 1397rule, a state entry is created; for subsequent packets the filter checks 1398whether the packet matches any state. 1399If it does, the packet is passed without evaluation of any rules. 1400After the connection is closed or times out, the state entry is automatically 1401removed. 1402.Pp 1403This has several advantages. 1404For TCP connections, comparing a packet to a state involves checking 1405its sequence numbers, as well as TCP timestamps if a 1406.Ar scrub reassemble tcp 1407rule applies to the connection. 1408If these values are outside the narrow windows of expected 1409values, the packet is dropped. 1410This prevents spoofing attacks, such as when an attacker sends packets with 1411a fake source address/port but does not know the connection's sequence 1412numbers. 1413Similarly, 1414.Xr pf 4 1415knows how to match ICMP replies to states. 1416For example, 1417.Bd -literal -offset indent 1418pass out inet proto icmp all icmp-type echoreq 1419.Ed 1420.Pp 1421allows echo requests (such as those created by 1422.Xr ping 8 ) 1423out statefully, and matches incoming echo replies correctly to states. 1424.Pp 1425Also, looking up states is usually faster than evaluating rules. 1426If there are 50 rules, all of them are evaluated sequentially in O(n). 1427Even with 50000 states, only 16 comparisons are needed to match a 1428state, since states are stored in a binary search tree that allows 1429searches in O(log2 n). 1430.Pp 1431Furthermore, correct handling of ICMP error messages is critical to 1432many protocols, particularly TCP. 1433.Xr pf 4 1434matches ICMP error messages to the correct connection, checks them against 1435connection parameters, and passes them if appropriate. 1436For example if an ICMP source quench message referring to a stateful TCP 1437connection arrives, it will be matched to the state and get passed. 1438.Pp 1439Finally, state tracking is required for 1440.Ar nat , binat No and Ar rdr 1441rules, in order to track address and port translations and reverse the 1442translation on returning packets. 1443.Pp 1444.Xr pf 4 1445will also create state for other protocols which are effectively stateless by 1446nature. 1447UDP packets are matched to states using only host addresses and ports, 1448and other protocols are matched to states using only the host addresses. 1449.Pp 1450If stateless filtering of individual packets is desired, 1451the 1452.Ar no state 1453keyword can be used to specify that state will not be created 1454if this is the last matching rule. 1455A number of parameters can also be set to affect how 1456.Xr pf 4 1457handles state tracking. 1458See 1459.Sx STATEFUL TRACKING OPTIONS 1460below for further details. 1461.Sh PARAMETERS 1462The rule parameters specify the packets to which a rule applies. 1463A packet always comes in on, or goes out through, one interface. 1464Most parameters are optional. 1465If a parameter is specified, the rule only applies to packets with 1466matching attributes. 1467Certain parameters can be expressed as lists, in which case 1468.Xr pfctl 8 1469generates all needed rule combinations. 1470.Bl -tag -width xxxx 1471.It Ar in No or Ar out 1472This rule applies to incoming or outgoing packets. 1473If neither 1474.Ar in 1475nor 1476.Ar out 1477are specified, the rule will match packets in both directions. 1478.It Ar log 1479In addition to the action specified, a log message is generated. 1480Only the packet that establishes the state is logged, 1481unless the 1482.Ar no state 1483option is specified. 1484The logged packets are sent to a 1485.Xr pflog 4 1486interface, by default 1487.Ar pflog0 . 1488This interface is monitored by the 1489.Xr pflogd 8 1490logging daemon, which dumps the logged packets to the file 1491.Pa /var/log/pflog 1492in 1493.Xr pcap 3 1494binary format. 1495.It Ar log (all) 1496Used to force logging of all packets for a connection. 1497This is not necessary when 1498.Ar no state 1499is explicitly specified. 1500As with 1501.Ar log , 1502packets are logged to 1503.Xr pflog 4 . 1504.It Ar log (user) 1505Logs the 1506.Ux 1507user ID of the user that owns the socket and the PID of the process that 1508has the socket open where the packet is sourced from or destined to 1509(depending on which socket is local). 1510This is in addition to the normal information logged. 1511.Pp 1512Only the first packet 1513logged via 1514.Ar log (all, user) 1515will have the user credentials logged when using stateful matching. 1516.It Ar log (to Aq Ar interface ) 1517Send logs to the specified 1518.Xr pflog 4 1519interface instead of 1520.Ar pflog0 . 1521.It Ar quick 1522If a packet matches a rule which has the 1523.Ar quick 1524option set, this rule 1525is considered the last matching rule, and evaluation of subsequent rules 1526is skipped. 1527.It Ar on Aq Ar interface 1528This rule applies only to packets coming in on, or going out through, this 1529particular interface or interface group. 1530For more information on interface groups, 1531see the 1532.Ic group 1533keyword in 1534.Xr ifconfig 8 . 1535.It Aq Ar af 1536This rule applies only to packets of this address family. 1537Supported values are 1538.Ar inet 1539and 1540.Ar inet6 . 1541.It Ar proto Aq Ar protocol 1542This rule applies only to packets of this protocol. 1543Common protocols are 1544.Xr icmp 4 , 1545.Xr icmp6 4 , 1546.Xr tcp 4 , 1547and 1548.Xr udp 4 . 1549For a list of all the protocol name to number mappings used by 1550.Xr pfctl 8 , 1551see the file 1552.Pa /etc/protocols . 1553.It Xo 1554.Ar from Aq Ar source 1555.Ar port Aq Ar source 1556.Ar os Aq Ar source 1557.Ar to Aq Ar dest 1558.Ar port Aq Ar dest 1559.Xc 1560This rule applies only to packets with the specified source and destination 1561addresses and ports. 1562.Pp 1563Addresses can be specified in CIDR notation (matching netblocks), as 1564symbolic host names, interface names or interface group names, or as any 1565of the following keywords: 1566.Pp 1567.Bl -tag -width xxxxxxxxxxxxxx -compact 1568.It Ar any 1569Any address. 1570.It Ar no-route 1571Any address which is not currently routable. 1572.It Ar urpf-failed 1573Any source address that fails a unicast reverse path forwarding (URPF) 1574check, i.e. packets coming in on an interface other than that which holds 1575the route back to the packet's source address. 1576.It Aq Ar table 1577Any address that matches the given table. 1578.El 1579.Pp 1580Ranges of addresses are specified by using the 1581.Sq - 1582operator. 1583For instance: 1584.Dq 10.1.1.10 - 10.1.1.12 1585means all addresses from 10.1.1.10 to 10.1.1.12, 1586hence addresses 10.1.1.10, 10.1.1.11, and 10.1.1.12. 1587.Pp 1588Interface names and interface group names can have modifiers appended: 1589.Pp 1590.Bl -tag -width xxxxxxxxxxxx -compact 1591.It Ar :network 1592Translates to the network(s) attached to the interface. 1593.It Ar :broadcast 1594Translates to the interface's broadcast address(es). 1595.It Ar :peer 1596Translates to the point-to-point interface's peer address(es). 1597.It Ar :0 1598Do not include interface aliases. 1599.El 1600.Pp 1601Host names may also have the 1602.Ar :0 1603option appended to restrict the name resolution to the first of each 1604v4 and non-link-local v6 address found. 1605.Pp 1606Host name resolution and interface to address translation are done at 1607ruleset load-time. 1608When the address of an interface (or host name) changes (under DHCP or PPP, 1609for instance), the ruleset must be reloaded for the change to be reflected 1610in the kernel. 1611Surrounding the interface name (and optional modifiers) in parentheses 1612changes this behaviour. 1613When the interface name is surrounded by parentheses, the rule is 1614automatically updated whenever the interface changes its address. 1615The ruleset does not need to be reloaded. 1616This is especially useful with 1617.Ar nat . 1618.Pp 1619Ports can be specified either by number or by name. 1620For example, port 80 can be specified as 1621.Em www . 1622For a list of all port name to number mappings used by 1623.Xr pfctl 8 , 1624see the file 1625.Pa /etc/services . 1626.Pp 1627Ports and ranges of ports are specified by using these operators: 1628.Bd -literal -offset indent 1629= (equal) 1630!= (unequal) 1631\*(Lt (less than) 1632\*(Le (less than or equal) 1633\*(Gt (greater than) 1634\*(Ge (greater than or equal) 1635: (range including boundaries) 1636\*(Gt\*(Lt (range excluding boundaries) 1637\*(Lt\*(Gt (except range) 1638.Ed 1639.Pp 1640.Sq \*(Gt\*(Lt , 1641.Sq \*(Lt\*(Gt 1642and 1643.Sq \&: 1644are binary operators (they take two arguments). 1645For instance: 1646.Bl -tag -width Fl 1647.It Ar port 2000:2004 1648means 1649.Sq all ports \*(Ge 2000 and \*(Le 2004 , 1650hence ports 2000, 2001, 2002, 2003 and 2004. 1651.It Ar port 2000 \*(Gt\*(Lt 2004 1652means 1653.Sq all ports \*(Gt 2000 and \*(Lt 2004 , 1654hence ports 2001, 2002 and 2003. 1655.It Ar port 2000 \*(Lt\*(Gt 2004 1656means 1657.Sq all ports \*(Lt 2000 or \*(Gt 2004 , 1658hence ports 1-1999 and 2005-65535. 1659.El 1660.Pp 1661The operating system of the source host can be specified in the case of TCP 1662rules with the 1663.Ar OS 1664modifier. 1665See the 1666.Sx OPERATING SYSTEM FINGERPRINTING 1667section for more information. 1668.Pp 1669The host, port and OS specifications are optional, as in the following examples: 1670.Bd -literal -offset indent 1671pass in all 1672pass in from any to any 1673pass in proto tcp from any port \*(Le 1024 to any 1674pass in proto tcp from any to any port 25 1675pass in proto tcp from 10.0.0.0/8 port \*(Gt 1024 \e 1676 to ! 10.1.2.3 port != ssh 1677pass in proto tcp from any os "OpenBSD" 1678.Ed 1679.It Ar all 1680This is equivalent to "from any to any". 1681.It Ar group Aq Ar group 1682Similar to 1683.Ar user , 1684this rule only applies to packets of sockets owned by the specified group. 1685.It Ar user Aq Ar user 1686This rule only applies to packets of sockets owned by the specified user. 1687For outgoing connections initiated from the firewall, this is the user 1688that opened the connection. 1689For incoming connections to the firewall itself, this is the user that 1690listens on the destination port. 1691For forwarded connections, where the firewall is not a connection endpoint, 1692the user and group are 1693.Em unknown . 1694.Pp 1695All packets, both outgoing and incoming, of one connection are associated 1696with the same user and group. 1697Only TCP and UDP packets can be associated with users; for other protocols 1698these parameters are ignored. 1699.Pp 1700User and group refer to the effective (as opposed to the real) IDs, in 1701case the socket is created by a setuid/setgid process. 1702User and group IDs are stored when a socket is created; 1703when a process creates a listening socket as root (for instance, by 1704binding to a privileged port) and subsequently changes to another 1705user ID (to drop privileges), the credentials will remain root. 1706.Pp 1707User and group IDs can be specified as either numbers or names. 1708The syntax is similar to the one for ports. 1709The value 1710.Em unknown 1711matches packets of forwarded connections. 1712.Em unknown 1713can only be used with the operators 1714.Cm = 1715and 1716.Cm != . 1717Other constructs like 1718.Cm user \*(Ge unknown 1719are invalid. 1720Forwarded packets with unknown user and group ID match only rules 1721that explicitly compare against 1722.Em unknown 1723with the operators 1724.Cm = 1725or 1726.Cm != . 1727For instance 1728.Cm user \*(Ge 0 1729does not match forwarded packets. 1730The following example allows only selected users to open outgoing 1731connections: 1732.Bd -literal -offset indent 1733block out proto { tcp, udp } all 1734pass out proto { tcp, udp } all user { \*(Lt 1000, dhartmei } 1735.Ed 1736.It Xo Ar flags Aq Ar a 1737.Pf / Ns Aq Ar b 1738.No \*(Ba / Ns Aq Ar b 1739.No \*(Ba any 1740.Xc 1741This rule only applies to TCP packets that have the flags 1742.Aq Ar a 1743set out of set 1744.Aq Ar b . 1745Flags not specified in 1746.Aq Ar b 1747are ignored. 1748For stateful connections, the default is 1749.Ar flags S/SA . 1750To indicate that flags should not be checked at all, specify 1751.Ar flags any . 1752The flags are: (F)IN, (S)YN, (R)ST, (P)USH, (A)CK, (U)RG, (E)CE, and C(W)R. 1753.Bl -tag -width Fl 1754.It Ar flags S/S 1755Flag SYN is set. 1756The other flags are ignored. 1757.It Ar flags S/SA 1758This is the default setting for stateful connections. 1759Out of SYN and ACK, exactly SYN may be set. 1760SYN, SYN+PSH and SYN+RST match, but SYN+ACK, ACK and ACK+RST do not. 1761This is more restrictive than the previous example. 1762.It Ar flags /SFRA 1763If the first set is not specified, it defaults to none. 1764All of SYN, FIN, RST and ACK must be unset. 1765.El 1766.Pp 1767Because 1768.Ar flags S/SA 1769is applied by default (unless 1770.Ar no state 1771is specified), only the initial SYN packet of a TCP handshake will create 1772a state for a TCP connection. 1773It is possible to be less restrictive, and allow state creation from 1774intermediate 1775.Pq non-SYN 1776packets, by specifying 1777.Ar flags any . 1778This will cause 1779.Xr pf 4 1780to synchronize to existing connections, for instance 1781if one flushes the state table. 1782However, states created from such intermediate packets may be missing 1783connection details such as the TCP window scaling factor. 1784States which modify the packet flow, such as those affected by 1785.Ar nat , binat No or Ar rdr 1786rules, 1787.Ar modulate No or Ar synproxy state 1788options, or scrubbed with 1789.Ar reassemble tcp 1790will also not be recoverable from intermediate packets. 1791Such connections will stall and time out. 1792.It Xo Ar icmp-type Aq Ar type 1793.Ar code Aq Ar code 1794.Xc 1795.It Xo Ar icmp6-type Aq Ar type 1796.Ar code Aq Ar code 1797.Xc 1798This rule only applies to ICMP or ICMPv6 packets with the specified type 1799and code. 1800Text names for ICMP types and codes are listed in 1801.Xr icmp 4 1802and 1803.Xr icmp6 4 . 1804This parameter is only valid for rules that cover protocols ICMP or 1805ICMP6. 1806The protocol and the ICMP type indicator 1807.Po 1808.Ar icmp-type 1809or 1810.Ar icmp6-type 1811.Pc 1812must match. 1813.It Xo Ar tos Aq Ar string 1814.No \*(Ba Aq Ar number 1815.Xc 1816This rule applies to packets with the specified 1817.Em TOS 1818bits set. 1819.Em TOS 1820may be 1821given as one of 1822.Ar critical , 1823.Ar inetcontrol , 1824.Ar lowdelay , 1825.Ar netcontrol , 1826.Ar throughput , 1827.Ar reliability , 1828or one of the DiffServ Code Points: 1829.Ar ef , 1830.Ar va , 1831.Ar af11 No ... Ar af43 , 1832.Ar cs0 No ... Ar cs7 ; 1833or as either hex or decimal. 1834.Pp 1835For example, the following rules are identical: 1836.Bd -literal -offset indent 1837pass all tos lowdelay 1838pass all tos 0x10 1839pass all tos 16 1840.Ed 1841.It Ar allow-opts 1842By default, IPv4 packets with IP options or IPv6 packets with routing 1843extension headers are blocked. 1844When 1845.Ar allow-opts 1846is specified for a 1847.Ar pass 1848rule, packets that pass the filter based on that rule (last matching) 1849do so even if they contain IP options or routing extension headers. 1850For packets that match state, the rule that initially created the 1851state is used. 1852The implicit 1853.Ar pass 1854rule that is used when a packet does not match any rules does not 1855allow IP options. 1856.It Ar label Aq Ar string 1857Adds a label (name) to the rule, which can be used to identify the rule. 1858For instance, 1859pfctl -s labels 1860shows per-rule statistics for rules that have labels. 1861.Pp 1862The following macros can be used in labels: 1863.Pp 1864.Bl -tag -width $srcaddr -compact -offset indent 1865.It Ar $if 1866The interface. 1867.It Ar $srcaddr 1868The source IP address. 1869.It Ar $dstaddr 1870The destination IP address. 1871.It Ar $srcport 1872The source port specification. 1873.It Ar $dstport 1874The destination port specification. 1875.It Ar $proto 1876The protocol name. 1877.It Ar $nr 1878The rule number. 1879.El 1880.Pp 1881For example: 1882.Bd -literal -offset indent 1883ips = \&"{ 1.2.3.4, 1.2.3.5 }\&" 1884pass in proto tcp from any to $ips \e 1885 port \*(Gt 1023 label \&"$dstaddr:$dstport\&" 1886.Ed 1887.Pp 1888expands to 1889.Bd -literal -offset indent 1890pass in inet proto tcp from any to 1.2.3.4 \e 1891 port \*(Gt 1023 label \&"1.2.3.4:\*(Gt1023\&" 1892pass in inet proto tcp from any to 1.2.3.5 \e 1893 port \*(Gt 1023 label \&"1.2.3.5:\*(Gt1023\&" 1894.Ed 1895.Pp 1896The macro expansion for the 1897.Ar label 1898directive occurs only at configuration file parse time, not during runtime. 1899.It Ar ridentifier Aq Ar number 1900Add an identifier (number) to the rule, which can be used to correlate the rule 1901to pflog entries, even after ruleset updates. 1902.It Xo Ar queue Aq Ar queue 1903.No \*(Ba ( Aq Ar queue , 1904.Aq Ar queue ) 1905.Xc 1906Packets matching this rule will be assigned to the specified queue. 1907If two queues are given, packets which have a 1908.Em TOS 1909of 1910.Em lowdelay 1911and TCP ACKs with no data payload will be assigned to the second one. 1912See 1913.Sx QUEUEING 1914for setup details. 1915.Pp 1916For example: 1917.Bd -literal -offset indent 1918pass in proto tcp to port 25 queue mail 1919pass in proto tcp to port 22 queue(ssh_bulk, ssh_prio) 1920.Ed 1921.Pp 1922.It Cm set prio Ar priority | Pq Ar priority , priority 1923Packets matching this rule will be assigned a specific queueing priority. 1924Priorities are assigned as integers 0 through 7. 1925If the packet is transmitted on a 1926.Xr vlan 4 1927interface, the queueing priority will be written as the priority 1928code point in the 802.1Q VLAN header. 1929If two priorities are given, packets which have a TOS of 1930.Cm lowdelay 1931and TCP ACKs with no data payload will be assigned to the second one. 1932.Pp 1933For example: 1934.Bd -literal -offset indent 1935pass in proto tcp to port 25 set prio 2 1936pass in proto tcp to port 22 set prio (2, 5) 1937.Ed 1938.Pp 1939.It Ar tag Aq Ar string 1940Packets matching this rule will be tagged with the 1941specified string. 1942The tag acts as an internal marker that can be used to 1943identify these packets later on. 1944This can be used, for example, to provide trust between 1945interfaces and to determine if packets have been 1946processed by translation rules. 1947Tags are 1948.Qq sticky , 1949meaning that the packet will be tagged even if the rule 1950is not the last matching rule. 1951Further matching rules can replace the tag with a 1952new one but will not remove a previously applied tag. 1953A packet is only ever assigned one tag at a time. 1954Packet tagging can be done during 1955.Ar nat , 1956.Ar rdr , 1957or 1958.Ar binat 1959rules in addition to filter rules. 1960Tags take the same macros as labels (see above). 1961.It Ar tagged Aq Ar string 1962Used with filter, translation or scrub rules 1963to specify that packets must already 1964be tagged with the given tag in order to match the rule. 1965Inverse tag matching can also be done 1966by specifying the 1967.Cm !\& 1968operator before the 1969.Ar tagged 1970keyword. 1971.It Ar rtable Aq Ar number 1972Used to select an alternate routing table for the routing lookup. 1973Only effective before the route lookup happened, i.e. when filtering inbound. 1974.It Xo Ar divert-to Aq Ar host 1975.Ar port Aq Ar port 1976.Xc 1977Used to redirect packets to a local socket bound to 1978.Ar host 1979and 1980.Ar port . 1981The packets will not be modified, so 1982.Xr getsockname 2 1983on the socket will return the original destination address of the packet. 1984.It Ar divert-reply 1985Used to receive replies for sockets that are bound to addresses 1986which are not local to the machine. 1987See 1988.Xr setsockopt 2 1989for information on how to bind these sockets. 1990.It Ar probability Aq Ar number 1991A probability attribute can be attached to a rule, with a value set between 19920 and 1, bounds not included. 1993In that case, the rule will be honoured using the given probability value 1994only. 1995For example, the following rule will drop 20% of incoming ICMP packets: 1996.Bd -literal -offset indent 1997block in proto icmp probability 20% 1998.Ed 1999.It Ar prio Aq Ar number 2000Only match packets which have the given queueing priority assigned. 2001.Pp 2002.El 2003.Sh ROUTING 2004If a packet matches a rule with a route option set, the packet filter will 2005route the packet according to the type of route option. 2006When such a rule creates state, the route option is also applied to all 2007packets matching the same connection. 2008.Bl -tag -width xxxx 2009.It Ar route-to 2010The 2011.Ar route-to 2012option routes the packet to the specified interface with an optional address 2013for the next hop. 2014When a 2015.Ar route-to 2016rule creates state, only packets that pass in the same direction as the 2017filter rule specifies will be routed in this way. 2018Packets passing in the opposite direction (replies) are not affected 2019and are routed normally. 2020.It Ar reply-to 2021The 2022.Ar reply-to 2023option is similar to 2024.Ar route-to , 2025but routes packets that pass in the opposite direction (replies) to the 2026specified interface. 2027Opposite direction is only defined in the context of a state entry, and 2028.Ar reply-to 2029is useful only in rules that create state. 2030It can be used on systems with multiple external connections to 2031route all outgoing packets of a connection through the interface 2032the incoming connection arrived through (symmetric routing enforcement). 2033.It Ar dup-to 2034The 2035.Ar dup-to 2036option creates a duplicate of the packet and routes it like 2037.Ar route-to . 2038The original packet gets routed as it normally would. 2039.El 2040.Sh POOL OPTIONS 2041For 2042.Ar nat 2043and 2044.Ar rdr 2045rules, (as well as for the 2046.Ar route-to , 2047.Ar reply-to 2048and 2049.Ar dup-to 2050rule options) for which there is a single redirection address which has a 2051subnet mask smaller than 32 for IPv4 or 128 for IPv6 (more than one IP 2052address), a variety of different methods for assigning this address can be 2053used: 2054.Bl -tag -width xxxx 2055.It Ar bitmask 2056The 2057.Ar bitmask 2058option applies the network portion of the redirection address to the address 2059to be modified (source with 2060.Ar nat , 2061destination with 2062.Ar rdr ) . 2063.It Ar random 2064The 2065.Ar random 2066option selects an address at random within the defined block of addresses. 2067.It Ar source-hash 2068The 2069.Ar source-hash 2070option uses a hash of the source address to determine the redirection address, 2071ensuring that the redirection address is always the same for a given source. 2072An optional key can be specified after this keyword either in hex or as a 2073string; by default 2074.Xr pfctl 8 2075randomly generates a key for source-hash every time the 2076ruleset is reloaded. 2077.It Ar round-robin 2078The 2079.Ar round-robin 2080option loops through the redirection address(es). 2081.Pp 2082When more than one redirection address is specified, 2083.Ar round-robin 2084is the only permitted pool type. 2085.It Ar static-port 2086With 2087.Ar nat 2088rules, the 2089.Ar static-port 2090option prevents 2091.Xr pf 4 2092from modifying the source port on TCP and UDP packets. 2093.It Xo Ar map-e-portset Aq Ar psid-offset 2094.No / Aq Ar psid-len 2095.No / Aq Ar psid 2096.Xc 2097With 2098.Ar nat 2099rules, the 2100.Ar map-e-portset 2101option enables the source port translation of MAP-E (RFC 7597) Customer Edge. 2102In order to make the host act as a MAP-E Customer Edge, setting up a tunneling 2103interface and pass rules for encapsulated packets are required in addition 2104to the map-e-portset nat rule. 2105.Pp 2106For example: 2107.Bd -literal -offset indent 2108nat on $gif_mape_if from $int_if:network to any \e 2109 -> $ipv4_mape_src map-e-portset 6/8/0x34 2110.Ed 2111.Pp 2112sets PSID offset 6, PSID length 8, PSID 0x34. 2113.Ed 2114.El 2115.Pp 2116Additionally, the 2117.Ar sticky-address 2118option can be specified to help ensure that multiple connections from the 2119same source are mapped to the same redirection address. 2120This option can be used with the 2121.Ar random 2122and 2123.Ar round-robin 2124pool options. 2125Note that by default these associations are destroyed as soon as there are 2126no longer states which refer to them; in order to make the mappings last 2127beyond the lifetime of the states, increase the global options with 2128.Ar set timeout src.track . 2129See 2130.Sx STATEFUL TRACKING OPTIONS 2131for more ways to control the source tracking. 2132.Sh STATE MODULATION 2133Much of the security derived from TCP is attributable to how well the 2134initial sequence numbers (ISNs) are chosen. 2135Some popular stack implementations choose 2136.Em very 2137poor ISNs and thus are normally susceptible to ISN prediction exploits. 2138By applying a 2139.Ar modulate state 2140rule to a TCP connection, 2141.Xr pf 4 2142will create a high quality random sequence number for each connection 2143endpoint. 2144.Pp 2145The 2146.Ar modulate state 2147directive implicitly keeps state on the rule and is 2148only applicable to TCP connections. 2149.Pp 2150For instance: 2151.Bd -literal -offset indent 2152block all 2153pass out proto tcp from any to any modulate state 2154pass in proto tcp from any to any port 25 flags S/SFRA modulate state 2155.Ed 2156.Pp 2157Note that modulated connections will not recover when the state table 2158is lost (firewall reboot, flushing the state table, etc...). 2159.Xr pf 4 2160will not be able to infer a connection again after the state table flushes 2161the connection's modulator. 2162When the state is lost, the connection may be left dangling until the 2163respective endpoints time out the connection. 2164It is possible on a fast local network for the endpoints to start an ACK 2165storm while trying to resynchronize after the loss of the modulator. 2166The default 2167.Ar flags 2168settings (or a more strict equivalent) should be used on 2169.Ar modulate state 2170rules to prevent ACK storms. 2171.Pp 2172Note that alternative methods are available 2173to prevent loss of the state table 2174and allow for firewall failover. 2175See 2176.Xr carp 4 2177and 2178.Xr pfsync 4 2179for further information. 2180.Sh SYN PROXY 2181By default, 2182.Xr pf 4 2183passes packets that are part of a 2184.Xr tcp 4 2185handshake between the endpoints. 2186The 2187.Ar synproxy state 2188option can be used to cause 2189.Xr pf 4 2190itself to complete the handshake with the active endpoint, perform a handshake 2191with the passive endpoint, and then forward packets between the endpoints. 2192.Pp 2193No packets are sent to the passive endpoint before the active endpoint has 2194completed the handshake, hence so-called SYN floods with spoofed source 2195addresses will not reach the passive endpoint, as the sender can't complete the 2196handshake. 2197.Pp 2198The proxy is transparent to both endpoints, they each see a single 2199connection from/to the other endpoint. 2200.Xr pf 4 2201chooses random initial sequence numbers for both handshakes. 2202Once the handshakes are completed, the sequence number modulators 2203(see previous section) are used to translate further packets of the 2204connection. 2205.Ar synproxy state 2206includes 2207.Ar modulate state . 2208.Pp 2209Rules with 2210.Ar synproxy 2211will not work if 2212.Xr pf 4 2213operates on a 2214.Xr bridge 4 . 2215.Pp 2216Example: 2217.Bd -literal -offset indent 2218pass in proto tcp from any to any port www synproxy state 2219.Ed 2220.Sh STATEFUL TRACKING OPTIONS 2221A number of options related to stateful tracking can be applied on a 2222per-rule basis. 2223.Ar keep state , 2224.Ar modulate state 2225and 2226.Ar synproxy state 2227support these options, and 2228.Ar keep state 2229must be specified explicitly to apply options to a rule. 2230.Pp 2231.Bl -tag -width xxxx -compact 2232.It Ar max Aq Ar number 2233Limits the number of concurrent states the rule may create. 2234When this limit is reached, further packets that would create 2235state will not match this rule until existing states time out. 2236.It Ar no-sync 2237Prevent state changes for states created by this rule from appearing on the 2238.Xr pfsync 4 2239interface. 2240.It Xo Aq Ar timeout 2241.Aq Ar seconds 2242.Xc 2243Changes the timeout values used for states created by this rule. 2244For a list of all valid timeout names, see 2245.Sx OPTIONS 2246above. 2247.It Ar sloppy 2248Uses a sloppy TCP connection tracker that does not check sequence 2249numbers at all, which makes insertion and ICMP teardown attacks way 2250easier. 2251This is intended to be used in situations where one does not see all 2252packets of a connection, e.g. in asymmetric routing situations. 2253Cannot be used with modulate or synproxy state. 2254.El 2255.Pp 2256Multiple options can be specified, separated by commas: 2257.Bd -literal -offset indent 2258pass in proto tcp from any to any \e 2259 port www keep state \e 2260 (max 100, source-track rule, max-src-nodes 75, \e 2261 max-src-states 3, tcp.established 60, tcp.closing 5) 2262.Ed 2263.Pp 2264When the 2265.Ar source-track 2266keyword is specified, the number of states per source IP is tracked. 2267.Pp 2268.Bl -tag -width xxxx -compact 2269.It Ar source-track rule 2270The maximum number of states created by this rule is limited by the rule's 2271.Ar max-src-nodes 2272and 2273.Ar max-src-states 2274options. 2275Only state entries created by this particular rule count toward the rule's 2276limits. 2277.It Ar source-track global 2278The number of states created by all rules that use this option is limited. 2279Each rule can specify different 2280.Ar max-src-nodes 2281and 2282.Ar max-src-states 2283options, however state entries created by any participating rule count towards 2284each individual rule's limits. 2285.El 2286.Pp 2287The following limits can be set: 2288.Pp 2289.Bl -tag -width xxxx -compact 2290.It Ar max-src-nodes Aq Ar number 2291Limits the maximum number of source addresses which can simultaneously 2292have state table entries. 2293.It Ar max-src-states Aq Ar number 2294Limits the maximum number of simultaneous state entries that a single 2295source address can create with this rule. 2296.El 2297.Pp 2298For stateful TCP connections, limits on established connections (connections 2299which have completed the TCP 3-way handshake) can also be enforced 2300per source IP. 2301.Pp 2302.Bl -tag -width xxxx -compact 2303.It Ar max-src-conn Aq Ar number 2304Limits the maximum number of simultaneous TCP connections which have 2305completed the 3-way handshake that a single host can make. 2306.It Xo Ar max-src-conn-rate Aq Ar number 2307.No / Aq Ar seconds 2308.Xc 2309Limit the rate of new connections over a time interval. 2310The connection rate is an approximation calculated as a moving average. 2311.El 2312.Pp 2313Because the 3-way handshake ensures that the source address is not being 2314spoofed, more aggressive action can be taken based on these limits. 2315With the 2316.Ar overload Aq Ar table 2317state option, source IP addresses which hit either of the limits on 2318established connections will be added to the named table. 2319This table can be used in the ruleset to block further activity from 2320the offending host, redirect it to a tarpit process, or restrict its 2321bandwidth. 2322.Pp 2323The optional 2324.Ar flush 2325keyword kills all states created by the matching rule which originate 2326from the host which exceeds these limits. 2327The 2328.Ar global 2329modifier to the flush command kills all states originating from the 2330offending host, regardless of which rule created the state. 2331.Pp 2332For example, the following rules will protect the webserver against 2333hosts making more than 100 connections in 10 seconds. 2334Any host which connects faster than this rate will have its address added 2335to the 2336.Aq bad_hosts 2337table and have all states originating from it flushed. 2338Any new packets arriving from this host will be dropped unconditionally 2339by the block rule. 2340.Bd -literal -offset indent 2341block quick from \*(Ltbad_hosts\*(Gt 2342pass in on $ext_if proto tcp to $webserver port www keep state \e 2343 (max-src-conn-rate 100/10, overload \*(Ltbad_hosts\*(Gt flush global) 2344.Ed 2345.Sh OPERATING SYSTEM FINGERPRINTING 2346Passive OS Fingerprinting is a mechanism to inspect nuances of a TCP 2347connection's initial SYN packet and guess at the host's operating system. 2348Unfortunately these nuances are easily spoofed by an attacker so the 2349fingerprint is not useful in making security decisions. 2350But the fingerprint is typically accurate enough to make policy decisions 2351upon. 2352.Pp 2353The fingerprints may be specified by operating system class, by 2354version, or by subtype/patchlevel. 2355The class of an operating system is typically the vendor or genre 2356and would be 2357.Ox 2358for the 2359.Xr pf 4 2360firewall itself. 2361The version of the oldest available 2362.Ox 2363release on the main FTP site 2364would be 2.6 and the fingerprint would be written 2365.Pp 2366.Dl \&"OpenBSD 2.6\&" 2367.Pp 2368The subtype of an operating system is typically used to describe the 2369patchlevel if that patch led to changes in the TCP stack behavior. 2370In the case of 2371.Ox , 2372the only subtype is for a fingerprint that was 2373normalized by the 2374.Ar no-df 2375scrub option and would be specified as 2376.Pp 2377.Dl \&"OpenBSD 3.3 no-df\&" 2378.Pp 2379Fingerprints for most popular operating systems are provided by 2380.Xr pf.os 5 . 2381Once 2382.Xr pf 4 2383is running, a complete list of known operating system fingerprints may 2384be listed by running: 2385.Pp 2386.Dl # pfctl -so 2387.Pp 2388Filter rules can enforce policy at any level of operating system specification 2389assuming a fingerprint is present. 2390Policy could limit traffic to approved operating systems or even ban traffic 2391from hosts that aren't at the latest service pack. 2392.Pp 2393The 2394.Ar unknown 2395class can also be used as the fingerprint which will match packets for 2396which no operating system fingerprint is known. 2397.Pp 2398Examples: 2399.Bd -literal -offset indent 2400pass out proto tcp from any os OpenBSD 2401block out proto tcp from any os Doors 2402block out proto tcp from any os "Doors PT" 2403block out proto tcp from any os "Doors PT SP3" 2404block out from any os "unknown" 2405pass on lo0 proto tcp from any os "OpenBSD 3.3 lo0" 2406.Ed 2407.Pp 2408Operating system fingerprinting is limited only to the TCP SYN packet. 2409This means that it will not work on other protocols and will not match 2410a currently established connection. 2411.Pp 2412Caveat: operating system fingerprints are occasionally wrong. 2413There are three problems: an attacker can trivially craft his packets to 2414appear as any operating system he chooses; 2415an operating system patch could change the stack behavior and no fingerprints 2416will match it until the database is updated; 2417and multiple operating systems may have the same fingerprint. 2418.Sh BLOCKING SPOOFED TRAFFIC 2419"Spoofing" is the faking of IP addresses, typically for malicious 2420purposes. 2421The 2422.Ar antispoof 2423directive expands to a set of filter rules which will block all 2424traffic with a source IP from the network(s) directly connected 2425to the specified interface(s) from entering the system through 2426any other interface. 2427.Pp 2428For example, the line 2429.Bd -literal -offset indent 2430antispoof for lo0 2431.Ed 2432.Pp 2433expands to 2434.Bd -literal -offset indent 2435block drop in on ! lo0 inet from 127.0.0.1/8 to any 2436block drop in on ! lo0 inet6 from ::1 to any 2437.Ed 2438.Pp 2439For non-loopback interfaces, there are additional rules to block incoming 2440packets with a source IP address identical to the interface's IP(s). 2441For example, assuming the interface wi0 had an IP address of 10.0.0.1 and a 2442netmask of 255.255.255.0, 2443the line 2444.Bd -literal -offset indent 2445antispoof for wi0 inet 2446.Ed 2447.Pp 2448expands to 2449.Bd -literal -offset indent 2450block drop in on ! wi0 inet from 10.0.0.0/24 to any 2451block drop in inet from 10.0.0.1 to any 2452.Ed 2453.Pp 2454Caveat: Rules created by the 2455.Ar antispoof 2456directive interfere with packets sent over loopback interfaces 2457to local addresses. 2458One should pass these explicitly. 2459.Sh FRAGMENT HANDLING 2460The size of IP datagrams (packets) can be significantly larger than the 2461maximum transmission unit (MTU) of the network. 2462In cases when it is necessary or more efficient to send such large packets, 2463the large packet will be fragmented into many smaller packets that will each 2464fit onto the wire. 2465Unfortunately for a firewalling device, only the first logical fragment will 2466contain the necessary header information for the subprotocol that allows 2467.Xr pf 4 2468to filter on things such as TCP ports or to perform NAT. 2469.Pp 2470Besides the use of 2471.Ar scrub 2472rules as described in 2473.Sx TRAFFIC NORMALIZATION 2474above, there are three options for handling fragments in the packet filter. 2475.Pp 2476One alternative is to filter individual fragments with filter rules. 2477If no 2478.Ar scrub 2479rule applies to a fragment, it is passed to the filter. 2480Filter rules with matching IP header parameters decide whether the 2481fragment is passed or blocked, in the same way as complete packets 2482are filtered. 2483Without reassembly, fragments can only be filtered based on IP header 2484fields (source/destination address, protocol), since subprotocol header 2485fields are not available (TCP/UDP port numbers, ICMP code/type). 2486The 2487.Ar fragment 2488option can be used to restrict filter rules to apply only to 2489fragments, but not complete packets. 2490Filter rules without the 2491.Ar fragment 2492option still apply to fragments, if they only specify IP header fields. 2493For instance, the rule 2494.Bd -literal -offset indent 2495pass in proto tcp from any to any port 80 2496.Ed 2497.Pp 2498never applies to a fragment, even if the fragment is part of a TCP 2499packet with destination port 80, because without reassembly this information 2500is not available for each fragment. 2501This also means that fragments cannot create new or match existing 2502state table entries, which makes stateful filtering and address 2503translation (NAT, redirection) for fragments impossible. 2504.Pp 2505It's also possible to reassemble only certain fragments by specifying 2506source or destination addresses or protocols as parameters in 2507.Ar scrub 2508rules. 2509.Pp 2510In most cases, the benefits of reassembly outweigh the additional 2511memory cost, and it's recommended to use 2512.Ar scrub 2513rules to reassemble 2514all fragments via the 2515.Ar fragment reassemble 2516modifier. 2517.Pp 2518The memory allocated for fragment caching can be limited using 2519.Xr pfctl 8 . 2520Once this limit is reached, fragments that would have to be cached 2521are dropped until other entries time out. 2522The timeout value can also be adjusted. 2523.Pp 2524When forwarding reassembled IPv6 packets, pf refragments them with 2525the original maximum fragment size. 2526This allows the sender to determine the optimal fragment size by 2527path MTU discovery. 2528.Sh ANCHORS 2529Besides the main ruleset, 2530.Xr pfctl 8 2531can load rulesets into 2532.Ar anchor 2533attachment points. 2534An 2535.Ar anchor 2536is a container that can hold rules, address tables, and other anchors. 2537.Pp 2538An 2539.Ar anchor 2540has a name which specifies the path where 2541.Xr pfctl 8 2542can be used to access the anchor to perform operations on it, such as 2543attaching child anchors to it or loading rules into it. 2544Anchors may be nested, with components separated by 2545.Sq / 2546characters, similar to how file system hierarchies are laid out. 2547The main ruleset is actually the default anchor, so filter and 2548translation rules, for example, may also be contained in any anchor. 2549.Pp 2550An anchor can reference another 2551.Ar anchor 2552attachment point 2553using the following kinds 2554of rules: 2555.Bl -tag -width xxxx 2556.It Ar nat-anchor Aq Ar name 2557Evaluates the 2558.Ar nat 2559rules in the specified 2560.Ar anchor . 2561.It Ar rdr-anchor Aq Ar name 2562Evaluates the 2563.Ar rdr 2564rules in the specified 2565.Ar anchor . 2566.It Ar binat-anchor Aq Ar name 2567Evaluates the 2568.Ar binat 2569rules in the specified 2570.Ar anchor . 2571.It Ar anchor Aq Ar name 2572Evaluates the filter rules in the specified 2573.Ar anchor . 2574.It Xo Ar load anchor 2575.Aq Ar name 2576.Ar from Aq Ar file 2577.Xc 2578Loads the rules from the specified file into the 2579anchor 2580.Ar name . 2581.El 2582.Pp 2583When evaluation of the main ruleset reaches an 2584.Ar anchor 2585rule, 2586.Xr pf 4 2587will proceed to evaluate all rules specified in that anchor. 2588.Pp 2589Matching filter and translation rules marked with the 2590.Ar quick 2591option are final and abort the evaluation of the rules in other 2592anchors and the main ruleset. 2593If the 2594.Ar anchor 2595itself is marked with the 2596.Ar quick 2597option, 2598ruleset evaluation will terminate when the anchor is exited if the packet is 2599matched by any rule within the anchor. 2600.Pp 2601.Ar anchor 2602rules are evaluated relative to the anchor in which they are contained. 2603For example, all 2604.Ar anchor 2605rules specified in the main ruleset will reference anchor 2606attachment points underneath the main ruleset, and 2607.Ar anchor 2608rules specified in a file loaded from a 2609.Ar load anchor 2610rule will be attached under that anchor point. 2611.Pp 2612Rules may be contained in 2613.Ar anchor 2614attachment points which do not contain any rules when the main ruleset 2615is loaded, and later such anchors can be manipulated through 2616.Xr pfctl 8 2617without reloading the main ruleset or other anchors. 2618For example, 2619.Bd -literal -offset indent 2620ext_if = \&"kue0\&" 2621block on $ext_if all 2622anchor spam 2623pass out on $ext_if all 2624pass in on $ext_if proto tcp from any \e 2625 to $ext_if port smtp 2626.Ed 2627.Pp 2628blocks all packets on the external interface by default, then evaluates 2629all rules in the 2630.Ar anchor 2631named "spam", and finally passes all outgoing connections and 2632incoming connections to port 25. 2633.Bd -literal -offset indent 2634# echo \&"block in quick from 1.2.3.4 to any\&" \&| \e 2635 pfctl -a spam -f - 2636.Ed 2637.Pp 2638This loads a single rule into the 2639.Ar anchor , 2640which blocks all packets from a specific address. 2641.Pp 2642The anchor can also be populated by adding a 2643.Ar load anchor 2644rule after the 2645.Ar anchor 2646rule: 2647.Bd -literal -offset indent 2648anchor spam 2649load anchor spam from "/etc/pf-spam.conf" 2650.Ed 2651.Pp 2652When 2653.Xr pfctl 8 2654loads 2655.Nm pf.conf , 2656it will also load all the rules from the file 2657.Pa /etc/pf-spam.conf 2658into the anchor. 2659.Pp 2660Optionally, 2661.Ar anchor 2662rules can specify packet filtering parameters using the same syntax as 2663filter rules. 2664When parameters are used, the 2665.Ar anchor 2666rule is only evaluated for matching packets. 2667This allows conditional evaluation of anchors, like: 2668.Bd -literal -offset indent 2669block on $ext_if all 2670anchor spam proto tcp from any to any port smtp 2671pass out on $ext_if all 2672pass in on $ext_if proto tcp from any to $ext_if port smtp 2673.Ed 2674.Pp 2675The rules inside 2676.Ar anchor 2677spam are only evaluated for 2678.Ar tcp 2679packets with destination port 25. 2680Hence, 2681.Bd -literal -offset indent 2682# echo \&"block in quick from 1.2.3.4 to any" \&| \e 2683 pfctl -a spam -f - 2684.Ed 2685.Pp 2686will only block connections from 1.2.3.4 to port 25. 2687.Pp 2688Anchors may end with the asterisk 2689.Pq Sq * 2690character, which signifies that all anchors attached at that point 2691should be evaluated in the alphabetical ordering of their anchor name. 2692For example, 2693.Bd -literal -offset indent 2694anchor "spam/*" 2695.Ed 2696.Pp 2697will evaluate each rule in each anchor attached to the 2698.Li spam 2699anchor. 2700Note that it will only evaluate anchors that are directly attached to the 2701.Li spam 2702anchor, and will not descend to evaluate anchors recursively. 2703.Pp 2704Since anchors are evaluated relative to the anchor in which they are 2705contained, there is a mechanism for accessing the parent and ancestor 2706anchors of a given anchor. 2707Similar to file system path name resolution, if the sequence 2708.Dq .. 2709appears as an anchor path component, the parent anchor of the current 2710anchor in the path evaluation at that point will become the new current 2711anchor. 2712As an example, consider the following: 2713.Bd -literal -offset indent 2714# echo ' anchor "spam/allowed" ' | pfctl -f - 2715# echo -e ' anchor "../banned" \en pass' | \e 2716 pfctl -a spam/allowed -f - 2717.Ed 2718.Pp 2719Evaluation of the main ruleset will lead into the 2720.Li spam/allowed 2721anchor, which will evaluate the rules in the 2722.Li spam/banned 2723anchor, if any, before finally evaluating the 2724.Ar pass 2725rule. 2726.Pp 2727Filter rule 2728.Ar anchors 2729can also be loaded inline in the ruleset within a brace ('{' '}') delimited 2730block. 2731Brace delimited blocks may contain rules or other brace-delimited blocks. 2732When anchors are loaded this way the anchor name becomes optional. 2733.Bd -literal -offset indent 2734anchor "external" on $ext_if { 2735 block 2736 anchor out { 2737 pass proto tcp from any to port { 25, 80, 443 } 2738 } 2739 pass in proto tcp to any port 22 2740} 2741.Ed 2742.Pp 2743Since the parser specification for anchor names is a string, any 2744reference to an anchor name containing 2745.Sq / 2746characters will require double quote 2747.Pq Sq \&" 2748characters around the anchor name. 2749.Sh TRANSLATION EXAMPLES 2750This example maps incoming requests on port 80 to port 8080, on 2751which a daemon is running (because, for example, it is not run as root, 2752and therefore lacks permission to bind to port 80). 2753.Bd -literal 2754# use a macro for the interface name, so it can be changed easily 2755ext_if = \&"ne3\&" 2756 2757# map daemon on 8080 to appear to be on 80 2758rdr on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 port 8080 2759.Ed 2760.Pp 2761If the 2762.Ar pass 2763modifier is given, packets matching the translation rule are passed without 2764inspecting the filter rules: 2765.Bd -literal 2766rdr pass on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 \e 2767 port 8080 2768.Ed 2769.Pp 2770In the example below, vlan12 is configured as 192.168.168.1; 2771the machine translates all packets coming from 192.168.168.0/24 to 204.92.77.111 2772when they are going out any interface except vlan12. 2773This has the net effect of making traffic from the 192.168.168.0/24 2774network appear as though it is the Internet routable address 2775204.92.77.111 to nodes behind any interface on the router except 2776for the nodes on vlan12. 2777(Thus, 192.168.168.1 can talk to the 192.168.168.0/24 nodes.) 2778.Bd -literal 2779nat on ! vlan12 from 192.168.168.0/24 to any -\*(Gt 204.92.77.111 2780.Ed 2781.Pp 2782In the example below, the machine sits between a fake internal 144.19.74.* 2783network, and a routable external IP of 204.92.77.100. 2784The 2785.Ar no nat 2786rule excludes protocol AH from being translated. 2787.Bd -literal 2788# NO NAT 2789no nat on $ext_if proto ah from 144.19.74.0/24 to any 2790nat on $ext_if from 144.19.74.0/24 to any -\*(Gt 204.92.77.100 2791.Ed 2792.Pp 2793In the example below, packets bound for one specific server, as well as those 2794generated by the sysadmins are not proxied; all other connections are. 2795.Bd -literal 2796# NO RDR 2797no rdr on $int_if proto { tcp, udp } from any to $server port 80 2798no rdr on $int_if proto { tcp, udp } from $sysadmins to any port 80 2799rdr on $int_if proto { tcp, udp } from any to any port 80 -\*(Gt 127.0.0.1 \e 2800 port 80 2801.Ed 2802.Pp 2803This longer example uses both a NAT and a redirection. 2804The external interface has the address 157.161.48.183. 2805On localhost, we are running 2806.Xr ftp-proxy 8 , 2807waiting for FTP sessions to be redirected to it. 2808The three mandatory anchors for 2809.Xr ftp-proxy 8 2810are omitted from this example; see the 2811.Xr ftp-proxy 8 2812manpage. 2813.Bd -literal 2814# NAT 2815# Translate outgoing packets' source addresses (any protocol). 2816# In this case, any address but the gateway's external address is mapped. 2817nat on $ext_if inet from ! ($ext_if) to any -\*(Gt ($ext_if) 2818 2819# NAT PROXYING 2820# Map outgoing packets' source port to an assigned proxy port instead of 2821# an arbitrary port. 2822# In this case, proxy outgoing isakmp with port 500 on the gateway. 2823nat on $ext_if inet proto udp from any port = isakmp to any -\*(Gt ($ext_if) \e 2824 port 500 2825 2826# BINAT 2827# Translate outgoing packets' source address (any protocol). 2828# Translate incoming packets' destination address to an internal machine 2829# (bidirectional). 2830binat on $ext_if from 10.1.2.150 to any -\*(Gt $ext_if 2831 2832# RDR 2833# Translate incoming packets' destination addresses. 2834# As an example, redirect a TCP and UDP port to an internal machine. 2835rdr on $ext_if inet proto tcp from any to ($ext_if) port 8080 \e 2836 -\*(Gt 10.1.2.151 port 22 2837rdr on $ext_if inet proto udp from any to ($ext_if) port 8080 \e 2838 -\*(Gt 10.1.2.151 port 53 2839 2840# RDR 2841# Translate outgoing ftp control connections to send them to localhost 2842# for proxying with ftp-proxy(8) running on port 8021. 2843rdr on $int_if proto tcp from any to any port 21 -\*(Gt 127.0.0.1 port 8021 2844.Ed 2845.Pp 2846In this example, a NAT gateway is set up to translate internal addresses 2847using a pool of public addresses (192.0.2.16/28) and to redirect 2848incoming web server connections to a group of web servers on the internal 2849network. 2850.Bd -literal 2851# NAT LOAD BALANCE 2852# Translate outgoing packets' source addresses using an address pool. 2853# A given source address is always translated to the same pool address by 2854# using the source-hash keyword. 2855nat on $ext_if inet from any to any -\*(Gt 192.0.2.16/28 source-hash 2856 2857# RDR ROUND ROBIN 2858# Translate incoming web server connections to a group of web servers on 2859# the internal network. 2860rdr on $ext_if proto tcp from any to any port 80 \e 2861 -\*(Gt { 10.1.2.155, 10.1.2.160, 10.1.2.161 } round-robin 2862.Ed 2863.Sh FILTER EXAMPLES 2864.Bd -literal 2865# The external interface is kue0 2866# (157.161.48.183, the only routable address) 2867# and the private network is 10.0.0.0/8, for which we are doing NAT. 2868 2869# use a macro for the interface name, so it can be changed easily 2870ext_if = \&"kue0\&" 2871 2872# normalize all incoming traffic 2873scrub in on $ext_if all fragment reassemble 2874 2875# block and log everything by default 2876block return log on $ext_if all 2877 2878# block anything coming from source we have no back routes for 2879block in from no-route to any 2880 2881# block packets whose ingress interface does not match the one in 2882# the route back to their source address 2883block in from urpf-failed to any 2884 2885# block and log outgoing packets that do not have our address as source, 2886# they are either spoofed or something is misconfigured (NAT disabled, 2887# for instance), we want to be nice and do not send out garbage. 2888block out log quick on $ext_if from ! 157.161.48.183 to any 2889 2890# silently drop broadcasts (cable modem noise) 2891block in quick on $ext_if from any to 255.255.255.255 2892 2893# block and log incoming packets from reserved address space and invalid 2894# addresses, they are either spoofed or misconfigured, we cannot reply to 2895# them anyway (hence, no return-rst). 2896block in log quick on $ext_if from { 10.0.0.0/8, 172.16.0.0/12, \e 2897 192.168.0.0/16, 255.255.255.255/32 } to any 2898 2899# ICMP 2900 2901# pass out/in certain ICMP queries and keep state (ping) 2902# state matching is done on host addresses and ICMP id (not type/code), 2903# so replies (like 0/0 for 8/0) will match queries 2904# ICMP error messages (which always refer to a TCP/UDP packet) are 2905# handled by the TCP/UDP states 2906pass on $ext_if inet proto icmp all icmp-type 8 code 0 2907 2908# UDP 2909 2910# pass out all UDP connections and keep state 2911pass out on $ext_if proto udp all 2912 2913# pass in certain UDP connections and keep state (DNS) 2914pass in on $ext_if proto udp from any to any port domain 2915 2916# TCP 2917 2918# pass out all TCP connections and modulate state 2919pass out on $ext_if proto tcp all modulate state 2920 2921# pass in certain TCP connections and keep state (SSH, SMTP, DNS, IDENT) 2922pass in on $ext_if proto tcp from any to any port { ssh, smtp, domain, \e 2923 auth } 2924 2925# Do not allow Windows 9x SMTP connections since they are typically 2926# a viral worm. Alternately we could limit these OSes to 1 connection each. 2927block in on $ext_if proto tcp from any os {"Windows 95", "Windows 98"} \e 2928 to any port smtp 2929 2930# IPv6 2931# pass in/out all IPv6 traffic: note that we have to enable this in two 2932# different ways, on both our physical interface and our tunnel 2933pass quick on gif0 inet6 2934pass quick on $ext_if proto ipv6 2935 2936# Packet Tagging 2937 2938# three interfaces: $int_if, $ext_if, and $wifi_if (wireless). NAT is 2939# being done on $ext_if for all outgoing packets. tag packets in on 2940# $int_if and pass those tagged packets out on $ext_if. all other 2941# outgoing packets (i.e., packets from the wireless network) are only 2942# permitted to access port 80. 2943 2944pass in on $int_if from any to any tag INTNET 2945pass in on $wifi_if from any to any 2946 2947block out on $ext_if from any to any 2948pass out quick on $ext_if tagged INTNET 2949pass out on $ext_if proto tcp from any to any port 80 2950 2951# tag incoming packets as they are redirected to spamd(8). use the tag 2952# to pass those packets through the packet filter. 2953 2954rdr on $ext_if inet proto tcp from \*(Ltspammers\*(Gt to port smtp \e 2955 tag SPAMD -\*(Gt 127.0.0.1 port spamd 2956 2957block in on $ext_if 2958pass in on $ext_if inet proto tcp tagged SPAMD 2959.Ed 2960.Sh GRAMMAR 2961Syntax for 2962.Nm 2963in BNF: 2964.Bd -literal 2965line = ( option | pf-rule | nat-rule | binat-rule | rdr-rule | 2966 antispoof-rule | altq-rule | queue-rule | trans-anchors | 2967 anchor-rule | anchor-close | load-anchor | table-rule | 2968 include ) 2969 2970option = "set" ( [ "timeout" ( timeout | "{" timeout-list "}" ) ] | 2971 [ "ruleset-optimization" [ "none" | "basic" | "profile" ]] | 2972 [ "optimization" [ "default" | "normal" | 2973 "high-latency" | "satellite" | 2974 "aggressive" | "conservative" ] ] 2975 [ "limit" ( limit-item | "{" limit-list "}" ) ] | 2976 [ "loginterface" ( interface-name | "none" ) ] | 2977 [ "block-policy" ( "drop" | "return" ) ] | 2978 [ "state-policy" ( "if-bound" | "floating" ) ] 2979 [ "state-defaults" state-opts ] 2980 [ "require-order" ( "yes" | "no" ) ] 2981 [ "fingerprints" filename ] | 2982 [ "skip on" ifspec ] | 2983 [ "debug" ( "none" | "urgent" | "misc" | "loud" ) ] 2984 [ "keepcounters" ] ) 2985 2986pf-rule = action [ ( "in" | "out" ) ] 2987 [ "log" [ "(" logopts ")"] ] [ "quick" ] 2988 [ "on" ifspec ] [ route ] [ af ] [ protospec ] 2989 hosts [ filteropt-list ] 2990 2991logopts = logopt [ "," logopts ] 2992logopt = "all" | "user" | "to" interface-name 2993 2994filteropt-list = filteropt-list filteropt | filteropt 2995filteropt = user | group | flags | icmp-type | icmp6-type | "tos" tos | 2996 ( "no" | "keep" | "modulate" | "synproxy" ) "state" 2997 [ "(" state-opts ")" ] | 2998 "fragment" | "no-df" | "min-ttl" number | "set-tos" tos | 2999 "max-mss" number | "random-id" | "reassemble tcp" | 3000 fragmentation | "allow-opts" | 3001 "label" string | "tag" string | [ ! ] "tagged" string | 3002 "set prio" ( number | "(" number [ [ "," ] number ] ")" ) | 3003 "queue" ( string | "(" string [ [ "," ] string ] ")" ) | 3004 "rtable" number | "probability" number"%" | "prio" number | 3005 "dnpipe" ( number | "(" number "," number ")" ) | 3006 "dnqueue" ( number | "(" number "," number ")" ) | 3007 "ridentifier" number 3008 3009nat-rule = [ "no" ] "nat" [ "pass" [ "log" [ "(" logopts ")" ] ] ] 3010 [ "on" ifspec ] [ af ] 3011 [ protospec ] hosts [ "tag" string ] [ "tagged" string ] 3012 [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" ) 3013 [ portspec ] [ pooltype ] [ "static-port" ] 3014 [ "map-e-portset" number "/" number "/" number ] ] 3015 3016binat-rule = [ "no" ] "binat" [ "pass" [ "log" [ "(" logopts ")" ] ] ] 3017 [ "on" interface-name ] [ af ] 3018 [ "proto" ( proto-name | proto-number ) ] 3019 "from" address [ "/" mask-bits ] "to" ipspec 3020 [ "tag" string ] [ "tagged" string ] 3021 [ "-\*(Gt" address [ "/" mask-bits ] ] 3022 3023rdr-rule = [ "no" ] "rdr" [ "pass" [ "log" [ "(" logopts ")" ] ] ] 3024 [ "on" ifspec ] [ af ] 3025 [ protospec ] hosts [ "tag" string ] [ "tagged" string ] 3026 [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" ) 3027 [ portspec ] [ pooltype ] ] 3028 3029antispoof-rule = "antispoof" [ "log" ] [ "quick" ] 3030 "for" ifspec [ af ] [ "label" string ] 3031 [ "ridentifier" number ] 3032 3033table-rule = "table" "\*(Lt" string "\*(Gt" [ tableopts-list ] 3034tableopts-list = tableopts-list tableopts | tableopts 3035tableopts = "persist" | "const" | "counters" | "file" string | 3036 "{" [ tableaddr-list ] "}" 3037tableaddr-list = tableaddr-list [ "," ] tableaddr-spec | tableaddr-spec 3038tableaddr-spec = [ "!" ] tableaddr [ "/" mask-bits ] 3039tableaddr = hostname | ifspec | "self" | 3040 ipv4-dotted-quad | ipv6-coloned-hex 3041 3042altq-rule = "altq on" interface-name queueopts-list 3043 "queue" subqueue 3044queue-rule = "queue" string [ "on" interface-name ] queueopts-list 3045 subqueue 3046 3047anchor-rule = "anchor" [ string ] [ ( "in" | "out" ) ] [ "on" ifspec ] 3048 [ af ] [ protospec ] [ hosts ] [ filteropt-list ] [ "{" ] 3049 3050anchor-close = "}" 3051 3052trans-anchors = ( "nat-anchor" | "rdr-anchor" | "binat-anchor" ) string 3053 [ "on" ifspec ] [ af ] [ "proto" ] [ protospec ] [ hosts ] 3054 3055load-anchor = "load anchor" string "from" filename 3056 3057queueopts-list = queueopts-list queueopts | queueopts 3058queueopts = [ "bandwidth" bandwidth-spec ] | 3059 [ "qlimit" number ] | [ "tbrsize" number ] | 3060 [ "priority" number ] | [ schedulers ] 3061schedulers = ( cbq-def | priq-def | hfsc-def ) 3062bandwidth-spec = "number" ( "b" | "Kb" | "Mb" | "Gb" | "%" ) 3063 3064action = "pass" | "block" [ return ] | [ "no" ] "scrub" 3065return = "drop" | "return" | "return-rst" [ "( ttl" number ")" ] | 3066 "return-icmp" [ "(" icmpcode [ [ "," ] icmp6code ] ")" ] | 3067 "return-icmp6" [ "(" icmp6code ")" ] 3068icmpcode = ( icmp-code-name | icmp-code-number ) 3069icmp6code = ( icmp6-code-name | icmp6-code-number ) 3070 3071ifspec = ( [ "!" ] ( interface-name | interface-group ) ) | 3072 "{" interface-list "}" 3073interface-list = [ "!" ] ( interface-name | interface-group ) 3074 [ [ "," ] interface-list ] 3075route = ( "route-to" | "reply-to" | "dup-to" ) 3076 ( routehost | "{" routehost-list "}" ) 3077 [ pooltype ] 3078af = "inet" | "inet6" 3079 3080protospec = "proto" ( proto-name | proto-number | 3081 "{" proto-list "}" ) 3082proto-list = ( proto-name | proto-number ) [ [ "," ] proto-list ] 3083 3084hosts = "all" | 3085 "from" ( "any" | "no-route" | "urpf-failed" | "self" | host | 3086 "{" host-list "}" ) [ port ] [ os ] 3087 "to" ( "any" | "no-route" | "self" | host | 3088 "{" host-list "}" ) [ port ] 3089 3090ipspec = "any" | host | "{" host-list "}" 3091host = [ "!" ] ( address [ "/" mask-bits ] | "\*(Lt" string "\*(Gt" ) 3092redirhost = address [ "/" mask-bits ] 3093routehost = "(" interface-name [ address [ "/" mask-bits ] ] ")" 3094address = ( interface-name | interface-group | 3095 "(" ( interface-name | interface-group ) ")" | 3096 hostname | ipv4-dotted-quad | ipv6-coloned-hex ) 3097host-list = host [ [ "," ] host-list ] 3098redirhost-list = redirhost [ [ "," ] redirhost-list ] 3099routehost-list = routehost [ [ "," ] routehost-list ] 3100 3101port = "port" ( unary-op | binary-op | "{" op-list "}" ) 3102portspec = "port" ( number | name ) [ ":" ( "*" | number | name ) ] 3103os = "os" ( os-name | "{" os-list "}" ) 3104user = "user" ( unary-op | binary-op | "{" op-list "}" ) 3105group = "group" ( unary-op | binary-op | "{" op-list "}" ) 3106 3107unary-op = [ "=" | "!=" | "\*(Lt" | "\*(Le" | "\*(Gt" | "\*(Ge" ] 3108 ( name | number ) 3109binary-op = number ( "\*(Lt\*(Gt" | "\*(Gt\*(Lt" | ":" ) number 3110op-list = ( unary-op | binary-op ) [ [ "," ] op-list ] 3111 3112os-name = operating-system-name 3113os-list = os-name [ [ "," ] os-list ] 3114 3115flags = "flags" ( [ flag-set ] "/" flag-set | "any" ) 3116flag-set = [ "F" ] [ "S" ] [ "R" ] [ "P" ] [ "A" ] [ "U" ] [ "E" ] 3117 [ "W" ] 3118 3119icmp-type = "icmp-type" ( icmp-type-code | "{" icmp-list "}" ) 3120icmp6-type = "icmp6-type" ( icmp-type-code | "{" icmp-list "}" ) 3121icmp-type-code = ( icmp-type-name | icmp-type-number ) 3122 [ "code" ( icmp-code-name | icmp-code-number ) ] 3123icmp-list = icmp-type-code [ [ "," ] icmp-list ] 3124 3125tos = ( "lowdelay" | "throughput" | "reliability" | 3126 [ "0x" ] number ) 3127 3128state-opts = state-opt [ [ "," ] state-opts ] 3129state-opt = ( "max" number | "no-sync" | timeout | "sloppy" | 3130 "source-track" [ ( "rule" | "global" ) ] | 3131 "max-src-nodes" number | "max-src-states" number | 3132 "max-src-conn" number | 3133 "max-src-conn-rate" number "/" number | 3134 "overload" "\*(Lt" string "\*(Gt" [ "flush" ] | 3135 "if-bound" | "floating" ) 3136 3137fragmentation = [ "fragment reassemble" ] 3138 3139timeout-list = timeout [ [ "," ] timeout-list ] 3140timeout = ( "tcp.first" | "tcp.opening" | "tcp.established" | 3141 "tcp.closing" | "tcp.finwait" | "tcp.closed" | 3142 "udp.first" | "udp.single" | "udp.multiple" | 3143 "icmp.first" | "icmp.error" | 3144 "other.first" | "other.single" | "other.multiple" | 3145 "frag" | "interval" | "src.track" | 3146 "adaptive.start" | "adaptive.end" ) number 3147 3148limit-list = limit-item [ [ "," ] limit-list ] 3149limit-item = ( "states" | "frags" | "src-nodes" ) number 3150 3151pooltype = ( "bitmask" | "random" | 3152 "source-hash" [ ( hex-key | string-key ) ] | 3153 "round-robin" ) [ sticky-address ] 3154 3155subqueue = string | "{" queue-list "}" 3156queue-list = string [ [ "," ] string ] 3157cbq-def = "cbq" [ "(" cbq-opt [ [ "," ] cbq-opt ] ")" ] 3158priq-def = "priq" [ "(" priq-opt [ [ "," ] priq-opt ] ")" ] 3159hfsc-def = "hfsc" [ "(" hfsc-opt [ [ "," ] hfsc-opt ] ")" ] 3160cbq-opt = ( "default" | "borrow" | "red" | "ecn" | "rio" ) 3161priq-opt = ( "default" | "red" | "ecn" | "rio" ) 3162hfsc-opt = ( "default" | "red" | "ecn" | "rio" | 3163 linkshare-sc | realtime-sc | upperlimit-sc ) 3164linkshare-sc = "linkshare" sc-spec 3165realtime-sc = "realtime" sc-spec 3166upperlimit-sc = "upperlimit" sc-spec 3167sc-spec = ( bandwidth-spec | 3168 "(" bandwidth-spec number bandwidth-spec ")" ) 3169include = "include" filename 3170.Ed 3171.Sh FILES 3172.Bl -tag -width "/etc/protocols" -compact 3173.It Pa /etc/hosts 3174Host name database. 3175.It Pa /etc/pf.conf 3176Default location of the ruleset file. 3177The file has to be created manually as it is not installed with a 3178standard installation. 3179.It Pa /etc/pf.os 3180Default location of OS fingerprints. 3181.It Pa /etc/protocols 3182Protocol name database. 3183.It Pa /etc/services 3184Service name database. 3185.El 3186.Sh SEE ALSO 3187.Xr altq 4 , 3188.Xr carp 4 , 3189.Xr icmp 4 , 3190.Xr icmp6 4 , 3191.Xr ip 4 , 3192.Xr ip6 4 , 3193.Xr pf 4 , 3194.Xr pfsync 4 , 3195.Xr tcp 4 , 3196.Xr udp 4 , 3197.Xr hosts 5 , 3198.Xr pf.os 5 , 3199.Xr protocols 5 , 3200.Xr services 5 , 3201.Xr ftp-proxy 8 , 3202.Xr pfctl 8 , 3203.Xr pflogd 8 3204.Sh HISTORY 3205The 3206.Nm 3207file format first appeared in 3208.Ox 3.0 . 3209