1.\" $NetBSD: pfctl.8,v 1.4 2004/11/14 11:26:48 yamt Exp $ 2.\" $OpenBSD: pfctl.8,v 1.115 2004/08/26 16:14:58 jaredy Exp $ 3.\" 4.\" Copyright (c) 2001 Kjell Wooding. All rights reserved. 5.\" 6.\" Redistribution and use in source and binary forms, with or without 7.\" modification, are permitted provided that the following conditions 8.\" are met: 9.\" 1. Redistributions of source code must retain the above copyright 10.\" notice, this list of conditions and the following disclaimer. 11.\" 2. Redistributions in binary form must reproduce the above copyright 12.\" notice, this list of conditions and the following disclaimer in the 13.\" documentation and/or other materials provided with the distribution. 14.\" 3. The name of the author may not be used to endorse or promote products 15.\" derived from this software without specific prior written permission. 16.\" 17.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR 18.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 19.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 20.\" IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, 21.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 22.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 23.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 24.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 25.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 26.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 27.\" 28.Dd November 20, 2002 29.Dt PFCTL 8 30.Os 31.Sh NAME 32.Nm pfctl 33.Nd "control the packet filter (PF) and network address translation (NAT) device" 34.Sh SYNOPSIS 35.Nm pfctl 36.Bk -words 37.Op Fl AdeghNnOoqRrvz 38.Op Fl a Ar anchor 39.Xo 40.Oo Fl D 41.Ar macro Ns = Ns Ar value Oc 42.Xc 43.Op Fl F Ar modifier 44.Op Fl f Ar file 45.Op Fl i Ar interface 46.Op Fl k Ar host 47.Op Fl p Ar device 48.Op Fl s Ar modifier 49.Oo Xo 50.Fl t Ar table 51.Fl T Ar command 52.Op Ar address ... Oc 53.Xc 54.Op Fl x Ar level 55.Ek 56.Sh DESCRIPTION 57The 58.Nm 59utility communicates with the packet filter device using the 60ioctl interface described in 61.Xr pf 4 . 62It allows ruleset and parameter configuration and retrieval of status 63information from the packet filter. 64.Pp 65Packet filtering restricts the types of packets that pass through 66network interfaces entering or leaving the host based on filter 67rules as described in 68.Xr pf.conf 5 . 69The packet filter can also replace addresses and ports of packets. 70Replacing source addresses and ports of outgoing packets is called 71NAT (Network Address Translation) and is used to connect an internal 72network (usually reserved address space) to an external one (the 73Internet) by making all connections to external hosts appear to 74come from the gateway. 75Replacing destination addresses and ports of incoming packets 76is used to redirect connections to different hosts and/or ports. 77A combination of both translations, bidirectional NAT, is also 78supported. 79Translation rules are described in 80.Xr pf.conf 5 . 81.Pp 82When the variable 83.Va pf 84is set to 85.Dv YES 86in 87.Xr rc.conf.local 8 , 88the rule file specified with the variable 89.Va pf_rules 90is loaded automatically by the 91.Xr rc 8 92scripts and the packet filter is enabled. 93.Pp 94The packet filter does not itself forward packets between interfaces. 95Forwarding can be enabled by setting the 96.Xr sysctl 8 97variables 98.Em net.inet.ip.forwarding 99and/or 100.Em net.inet6.ip6.forwarding 101to 1. 102Set them permanently in 103.Xr sysctl.conf 5 . 104.Pp 105The 106.Nm 107utility provides several commands. 108The options are as follows: 109.Bl -tag -width Ds 110.It Fl A 111Load only the queue rules present in the rule file. 112Other rules and options are ignored. 113.It Fl a Ar anchor 114Apply flags 115.Fl f , 116.Fl F , 117and 118.Fl s 119only to the rules in the specified 120.Ar anchor . 121In addition to the main ruleset, 122.Nm 123can load and manipulate additional rulesets by name, 124called anchors. 125The main ruleset is the default anchor. 126.Pp 127Anchors are referenced by name and may be nested, 128with the various components of the anchor path separated by 129.Sq / 130characters, similar to how file system hierarchies are laid out. 131The last component of the anchor path is where ruleset operations are 132performed. 133.Pp 134Evaluation of 135.Ar anchor 136rules from the main ruleset is described in 137.Xr pf.conf 5 . 138.Pp 139For example, the following will show all filter rules (see the 140.Fl s 141flag below) inside the anchor 142.Li authpf/smith , 143which would have been created for user smith by 144.Xr authpf 8 : 145.Bd -literal -offset indent 146# pfctl -a authpf/smith -s rules 147.Ed 148.Pp 149Private tables can also be put inside anchors, either by having table 150statements in the 151.Xr pf.conf 5 152file that is loaded in the anchor, or by using regular table commands, as in: 153.Bd -literal -offset indent 154# pfctl -a foo/bar -t mytable -T add 1.2.3.4 5.6.7.8 155.Ed 156.Pp 157When a rule referring to a table is loaded in an anchor, the rule will use the 158private table if one is defined, and then fallback to the table defined in the 159main ruleset, if there is one. 160This is similar to C rules for variable scope. 161It is possible to create distinct tables with the same name in the global 162ruleset and in an anchor, but this is often bad design and a warning will be 163issued in that case. 164.It Fl D Ar macro Ns = Ns Ar value 165Define 166.Ar macro 167to be set to 168.Ar value 169on the command line. 170Overrides the definition of 171.Ar macro 172in the ruleset. 173.It Fl d 174Disable the packet filter. 175.It Fl e 176Enable the packet filter. 177.It Fl F Ar modifier 178Flush the filter parameters specified by 179.Ar modifier 180(may be abbreviated): 181.Pp 182.Bl -tag -width xxxxxxxxxxxx -compact 183.It Fl F Cm nat 184Flush the NAT rules. 185.It Fl F Cm queue 186Flush the queue rules. 187.It Fl F Cm rules 188Flush the filter rules. 189.It Fl F Cm state 190Flush the state table (NAT and filter). 191.It Fl F Cm Sources 192Flush the source tracking table. 193.It Fl F Cm info 194Flush the filter information (statistics that are not bound to rules). 195.It Fl F Cm Tables 196Flush the tables. 197.It Fl F Cm osfp 198Flush the passive operating system fingerprints. 199.It Fl F Cm all 200Flush all of the above. 201.El 202.It Fl f Ar file 203Load the rules contained in 204.Ar file . 205This 206.Ar file 207may contain macros, tables, options, and normalization, queueing, 208translation, and filtering rules. 209With the exception of macros and tables, the statements must appear in that 210order. 211.It Fl g 212Include output helpful for debugging. 213.It Fl h 214Help. 215.It Fl i Ar interface 216Restrict the operation to the given 217.Ar interface . 218.It Fl k Ar host 219Kill all of the state entries originating from the specified 220.Ar host . 221A second 222.Fl k Ar host 223option may be specified, which will kill all the state entries 224from the first 225.Ar host 226to the second 227.Ar host . 228For example, to kill all of the state entries originating from 229.Li host : 230.Bd -literal -offset indent 231# pfctl -k host 232.Ed 233.Pp 234To kill all of the state entries from 235.Li host1 236to 237.Li host2 : 238.Bd -literal -offset indent 239# pfctl -k host1 -k host2 240.Ed 241.It Fl N 242Load only the NAT rules present in the rule file. 243Other rules and options are ignored. 244.It Fl n 245Do not actually load rules, just parse them. 246.It Fl O 247Load only the options present in the rule file. 248Other rules and options are ignored. 249.It Fl o 250Enable the ruleset optimizer. 251The ruleset optimizer attempts to improve rulesets by removing rule 252duplication and making better use of rule ordering. 253Specifically, it does four things: 254.Pp 255.Bl -enum -compact 256.It 257remove duplicate rules 258.It 259remove rules that are a subset of another rule 260.It 261combine multiple rules into a table when advantageous 262.It 263re-order the rules to improve evaluation performance 264.El 265.Pp 266A second 267.Fl o 268may be specified to use the currently loaded ruleset as a feedback profile 269to tailor the optimization of the 270.Ar quick 271rules to the actual network behavior. 272.Pp 273It is important to note that the ruleset optimizer will modify the ruleset 274to improve performance. 275A side effect of the ruleset modification is that per-rule accounting 276statistics will have different meanings than before. 277If per-rule accounting is important for billing purposes or whatnot, either 278the ruleset optimizer should not be used or a 279.Ar label 280field should be added to all of the accounting rules to act as optimization 281barriers. 282.It Fl p Ar device 283Use the device file 284.Ar device 285instead of the default 286.Pa /dev/pf . 287.It Fl q 288Only print errors and warnings. 289.It Fl R 290Load only the filter rules present in the rule file. 291Other rules and options are ignored. 292.It Fl r 293Perform reverse DNS lookups on states when displaying them. 294.It Fl s Ar modifier 295Show the filter parameters specified by 296.Ar modifier 297(may be abbreviated): 298.Pp 299.Bl -tag -width xxxxxxxxxxxxx -compact 300.It Fl s Cm nat 301Show the currently loaded NAT rules. 302.It Fl s Cm queue 303Show the currently loaded queue rules. 304When used together with 305.Fl v , 306per-queue statistics are also shown. 307When used together with 308.Fl v v , 309.Nm 310will loop and show updated queue statistics every five seconds, including 311measured bandwidth and packets per second. 312.It Fl s Cm rules 313Show the currently loaded filter rules. 314When used together with 315.Fl v , 316the per-rule statistics (number of evaluations, 317packets and bytes) are also shown. 318Note that the 319.Dq skip step 320optimization done automatically by the kernel 321will skip evaluation of rules where possible. 322Packets passed statefully are counted in the rule that created the state 323(even though the rule isn't evaluated more than once for the entire 324connection). 325.It Fl s Cm Anchors 326Show the currently loaded anchors directly attached to the main ruleset. 327If 328.Fl a Ar anchor 329is specified as well, the anchors loaded directly below the given 330.Ar anchor 331are shown instead. 332If 333.Fl v 334is specified, all anchors attached under the target anchor will be 335displayed recursively. 336.It Fl s Cm state 337Show the contents of the state table. 338.It Fl s Cm Sources 339Show the contents of the source tracking table. 340.It Fl s Cm info 341Show filter information (statistics and counters). 342When used together with 343.Fl v , 344source tracking statistics are also shown. 345.It Fl s Cm labels 346Show per-rule statistics (label, evaluations, packets, bytes) of 347filter rules with labels, useful for accounting. 348.It Fl s Cm timeouts 349Show the current global timeouts. 350.It Fl s Cm memory 351Show the current pool memory hard limits. 352.It Fl s Cm Tables 353Show the list of tables. 354.It Fl s Cm osfp 355Show the list of operating system fingerprints. 356.It Fl s Cm Interfaces 357Show the list of interfaces and interface drivers available to PF. 358When used together with a double 359.Fl v , 360interface statistics are also shown. 361.Fl i 362can be used to select an interface or a group of interfaces. 363.It Fl s Cm all 364Show all of the above, except for the lists of interfaces and operating 365system fingerprints. 366.El 367.It Fl T Ar command Op Ar address ... 368Specify the 369.Ar command 370(may be abbreviated) to apply to the table. 371Commands include: 372.Pp 373.Bl -tag -width xxxxxxxxxxxx -compact 374.It Fl T Cm kill 375Kill a table. 376.It Fl T Cm flush 377Flush all addresses of a table. 378.It Fl T Cm add 379Add one or more addresses in a table. 380Automatically create a nonexisting table. 381.It Fl T Cm delete 382Delete one or more addresses from a table. 383.It Fl T Cm replace 384Replace the addresses of the table. 385Automatically create a nonexisting table. 386.It Fl T Cm show 387Show the content (addresses) of a table. 388.It Fl T Cm test 389Test if the given addresses match a table. 390.It Fl T Cm zero 391Clear all the statistics of a table. 392.It Fl T Cm load 393Load only the table definitions from 394.Xr pf.conf 5 . 395This is used in conjunction with the 396.Fl f 397flag, as in: 398.Bd -literal -offset indent 399# pfctl -Tl -f pf.conf 400.Ed 401.El 402.Pp 403For the 404.Cm add , 405.Cm delete , 406.Cm replace , 407and 408.Cm test 409commands, the list of addresses can be specified either directly on the command 410line and/or in an unformatted text file, using the 411.Fl f 412flag. 413Comments starting with a 414.Sq # 415are allowed in the text file. 416With these commands, the 417.Fl v 418flag can also be used once or twice, in which case 419.Nm 420will print the 421detailed result of the operation for each individual address, prefixed by 422one of the following letters: 423.Pp 424.Bl -tag -width XXX -compact 425.It A 426The address/network has been added. 427.It C 428The address/network has been changed (negated). 429.It D 430The address/network has been deleted. 431.It M 432The address matches 433.Po 434.Cm test 435operation only 436.Pc . 437.It X 438The address/network is duplicated and therefore ignored. 439.It Y 440The address/network cannot be added/deleted due to conflicting 441.Sq \&! 442attributes. 443.It Z 444The address/network has been cleared (statistics). 445.El 446.Pp 447Each table maintains a set of counters that can be retrieved using the 448.Fl v 449flag of 450.Nm . 451For example, the following commands define a wide open firewall which will keep 452track of packets going to or coming from the 453.Ox 454FTP server. 455The following commands configure the firewall and send 10 pings to the FTP 456server: 457.Bd -literal -offset indent 458# printf "table \*[Lt]test\*[Gt] { ftp.NetBSD.org }\en \e 459 pass out to \*[Lt]test\*[Gt] keep state\en" | pfctl -f- 460# ping -qc10 ftp.NetBSD.org 461.Ed 462.Pp 463We can now use the table 464.Cm show 465command to output, for each address and packet direction, the number of packets 466and bytes that are being passed or blocked by rules referencing the table. 467The time at which the current accounting started is also shown with the 468.Dq Cleared 469line. 470.Bd -literal -offset indent 471# pfctl -t test -vTshow 472 129.128.5.191 473 Cleared: Thu Feb 13 18:55:18 2003 474 In/Block: [ Packets: 0 Bytes: 0 ] 475 In/Pass: [ Packets: 10 Bytes: 840 ] 476 Out/Block: [ Packets: 0 Bytes: 0 ] 477 Out/Pass: [ Packets: 10 Bytes: 840 ] 478.Ed 479.Pp 480Similarly, it is possible to view global information about the tables 481by using the 482.Fl v 483modifier twice and the 484.Fl s 485.Cm Tables 486command. 487This will display the number of addresses on each table, 488the number of rules which reference the table, and the global 489packet statistics for the whole table: 490.Bd -literal -offset indent 491# pfctl -vvsTables 492--a-r- test 493 Addresses: 1 494 Cleared: Thu Feb 13 18:55:18 2003 495 References: [ Anchors: 0 Rules: 1 ] 496 Evaluations: [ NoMatch: 3496 Match: 1 ] 497 In/Block: [ Packets: 0 Bytes: 0 ] 498 In/Pass: [ Packets: 10 Bytes: 840 ] 499 In/XPass: [ Packets: 0 Bytes: 0 ] 500 Out/Block: [ Packets: 0 Bytes: 0 ] 501 Out/Pass: [ Packets: 10 Bytes: 840 ] 502 Out/XPass: [ Packets: 0 Bytes: 0 ] 503.Ed 504.Pp 505As we can see here, only one packet \- the initial ping request \- matched the 506table, but all packets passing as the result of the state are correctly 507accounted for. 508Reloading the table(s) or ruleset will not affect packet accounting in any way. 509The two 510.Dq XPass 511counters are incremented instead of the 512.Dq Pass 513counters when a 514.Dq stateful 515packet is passed but doesn't match the table anymore. 516This will happen in our example if someone flushes the table while the 517.Xr ping 8 518command is running. 519.Pp 520When used with a single 521.Fl v , 522.Nm 523will only display the first line containing the table flags and name. 524The flags are defined as follows: 525.Pp 526.Bl -tag -width XXX -compact 527.It c 528For constant tables, which cannot be altered outside 529.Xr pf.conf 5 . 530.It p 531For persistent tables, which don't get automatically killed when no rules 532refer to them. 533.It a 534For tables which are part of the 535.Em active 536tableset. 537Tables without this flag do not really exist, cannot contain addresses, and are 538only listed if the 539.Fl g 540flag is given. 541.It i 542For tables which are part of the 543.Em inactive 544tableset. 545This flag can only be witnessed briefly during the loading of 546.Xr pf.conf 5 . 547.It r 548For tables which are referenced (used) by rules. 549.It h 550This flag is set when a table in the main ruleset is hidden by one or more 551tables of the same name from anchors attached below it. 552.El 553.It Fl t Ar table 554Specify the name of the table. 555.It Fl v 556Produce more verbose output. 557A second use of 558.Fl v 559will produce even more verbose output including ruleset warnings. 560See the previous section for its effect on table commands. 561.It Fl x Ar level 562Set the debug 563.Ar level 564(may be abbreviated) to one of the following: 565.Pp 566.Bl -tag -width xxxxxxxxxxxx -compact 567.It Fl x Cm none 568Don't generate debug messages. 569.It Fl x Cm urgent 570Generate debug messages only for serious errors. 571.It Fl x Cm misc 572Generate debug messages for various errors. 573.It Fl x Cm loud 574Generate debug messages for common conditions. 575.El 576.It Fl z 577Clear per-rule statistics. 578.El 579.Sh FILES 580.Bl -tag -width "/etc/pf.conf" -compact 581.It Pa /etc/pf.conf 582Packet filter rules file. 583.It Pa /etc/pf.os 584Passive operating system fingerprint database. 585.El 586.Sh SEE ALSO 587.Xr pf 4 , 588.Xr pf.conf 5 , 589.Xr pf.os 5 , 590.Xr sysctl.conf 5 , 591.Xr authpf 8 , 592.Xr ftp-proxy 8 , 593.Xr rc 8 , 594.Xr rc.conf 8 , 595.Xr sysctl 8 596.Sh HISTORY 597The 598.Nm 599program and the 600.Xr pf 4 601filter mechanism first appeared in 602.Ox 3.0 . 603