9 .Nd User interface for firewall, traffic shaper, packet scheduler,
12 .Ss FIREWALL CONFIGURATION
21 .Op Ar rule | first-last ...
29 .Brq Cm delete | zero | resetlog
33 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
37 .Ar number Cm to Ar number
39 .Cm set swap Ar number number
45 .Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
48 .Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
51 .Cm table Ar number Cm add Ar addr Ns Oo / Ns Ar masklen Oc Op Ar value
53 .Cm table Ar number Cm delete Ar addr Ns Op / Ns Ar masklen
62 .Ss DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)
64 .Brq Cm pipe | queue | sched
70 .Brq Cm pipe | queue | sched
71 .Brq Cm delete | list | show
93 utility is the user interface for controlling the
97 traffic shaper/packet scheduler, and the
98 in-kernel NAT services.
100 A firewall configuration, or
104 numbered from 1 to 65535.
105 Packets are passed to the firewall
106 from a number of different places in the protocol stack
107 (depending on the source and destination of the packet,
108 it is possible for the firewall to be
109 invoked multiple times on the same packet).
110 The packet passed to the firewall is compared
111 against each of the rules in the
114 (multiple rules with the same number are permitted, in which case
115 they are processed in order of insertion).
116 When a match is found, the action corresponding to the
117 matching rule is performed.
119 Depending on the action and certain system settings, packets
120 can be reinjected into the firewall at some rule after the
121 matching one for further processing.
123 A ruleset always includes a
125 rule (numbered 65535) which cannot be modified or deleted,
126 and matches all packets.
127 The action associated with the
133 depending on how the kernel is configured.
135 If the ruleset includes one or more rules with the
140 the firewall will have a
142 behaviour, i.e., upon a match it will create
144 i.e., rules that match packets with the same 5-tuple
145 (protocol, source and destination addresses and ports)
146 as the packet which caused their creation.
147 Dynamic rules, which have a limited lifetime, are checked
148 at the first occurrence of a
153 rule, and are typically used to open the firewall on-demand to
154 legitimate traffic only.
156 .Sx STATEFUL FIREWALL
159 Sections below for more information on the stateful behaviour of
162 All rules (including dynamic ones) have a few associated counters:
163 a packet count, a byte count, a log count and a timestamp
164 indicating the time of the last match.
165 Counters can be displayed or reset with
169 Each rule belongs to one of 32 different
173 commands to atomically manipulate sets, such as enable,
174 disable, swap sets, move all rules in a set to another
175 one, delete all rules in a set.
176 These can be useful to
177 install temporary configurations, or to test them.
180 for more information on
183 Rules can be added with the
185 command; deleted individually or in groups with the
187 command, and globally (except those in set 31) with the
189 command; displayed, optionally with the content of the
195 Finally, counters can be reset with the
202 The following general options are available when invoking
204 .Bl -tag -width indent
206 Show counter values when listing rules.
209 command implies this option.
211 Only show the action and the comment, not the body of a rule.
215 When entering or showing rules, print them in compact form,
216 i.e., omitting the "ip from any to any" string
217 when this does not carry any additional information.
219 When listing, show dynamic rules in addition to static ones.
223 is specified, also show expired dynamic rules.
225 Do not ask for confirmation for commands that can cause problems
228 If there is no tty associated with the process, this is implied.
230 When listing a table (see the
232 section below for more information on lookup tables), format values
234 By default, values are shown as integers.
236 Only check syntax of the command strings, without actually passing
239 Try to resolve addresses and service names in output.
241 Be quiet when executing the
251 This is useful when updating rulesets by executing multiple
255 .Ql sh\ /etc/rc.firewall ) ,
256 or by processing a file with many
258 rules across a remote login session.
259 It also stops a table add or delete
260 from failing if the entry already exists or is not present.
262 The reason why this option may be important is that
263 for some of these actions,
265 may print a message; if the action results in blocking the
266 traffic to the remote client,
267 the remote login session will be closed
268 and the rest of the ruleset will not be processed.
269 Access to the console would then be required to recover.
271 When listing rules, show the
273 each rule belongs to.
274 If this flag is not specified, disabled rules will not be
277 When listing pipes, sort according to one of the four
278 counters (total or current packets or bytes).
280 When listing, show last match timestamp converted with ctime().
282 When listing, show last match timestamp as seconds from the epoch.
283 This form can be more convenient for postprocessing by scripts.
285 .Ss LIST OF RULES AND PREPROCESSING
286 To ease configuration, rules can be put into a file which is
289 as shown in the last synopsis line.
293 The file will be read line by line and applied as arguments to the
297 Optionally, a preprocessor can be specified using
301 is to be piped through.
302 Useful preprocessors include
308 does not start with a slash
310 as its first character, the usual
312 name search is performed.
313 Care should be taken with this in environments where not all
314 file systems are mounted (yet) by the time
316 is being run (e.g.\& when they are mounted over NFS).
319 has been specified, any additional arguments are passed on to the preprocessor
321 This allows for flexible configuration files (like conditionalizing
322 them on the local hostname) and the use of macros to centralize
323 frequently required arguments like IP addresses.
324 .Ss TRAFFIC SHAPER CONFIGURATION
330 commands are used to configure the traffic shaper and packet scheduler.
332 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
333 Section below for details.
335 If the world and the kernel get out of sync the
337 ABI may break, preventing you from being able to add any rules.
338 This can adversely affect the booting process.
343 to temporarily disable the firewall to regain access to the network,
344 allowing you to fix the problem.
346 A packet is checked against the active ruleset in multiple places
347 in the protocol stack, under control of several sysctl variables.
348 These places and variables are shown below, and it is important to
349 have this picture in mind in order to design a correct ruleset.
350 .Bd -literal -offset indent
353 +----------->-----------+
355 [ip(6)_input] [ip(6)_output] net.inet(6).ip(6).fw.enable=1
358 [ether_demux] [ether_output_frame] net.link.ether.ipfw=1
360 +-->--[bdg_forward]-->--+ net.link.bridge.ipfw=1
366 times the same packet goes through the firewall can
367 vary between 0 and 4 depending on packet source and
368 destination, and system configuration.
370 Note that as packets flow through the stack, headers can be
371 stripped or added to it, and so they may or may not be available
373 E.g., incoming packets will include the MAC header when
377 but the same packets will have the MAC header stripped off when
384 Also note that each packet is always checked against the complete ruleset,
385 irrespective of the place where the check occurs, or the source of the packet.
386 If a rule contains some match patterns or actions which are not valid
387 for the place of invocation (e.g.\& trying to match a MAC header within
391 the match pattern will not match, but a
393 operator in front of such patterns
397 match on those packets.
398 It is thus the responsibility of
399 the programmer, if necessary, to write a suitable ruleset to
400 differentiate among the possible places.
402 rules can be useful here, as an example:
403 .Bd -literal -offset indent
404 # packets from ether_demux or bdg_forward
405 ipfw add 10 skipto 1000 all from any to any layer2 in
406 # packets from ip_input
407 ipfw add 10 skipto 2000 all from any to any not layer2 in
408 # packets from ip_output
409 ipfw add 10 skipto 3000 all from any to any not layer2 out
410 # packets from ether_output_frame
411 ipfw add 10 skipto 4000 all from any to any layer2 out
414 (yes, at the moment there is no way to differentiate between
415 ether_demux and bdg_forward).
417 In general, each keyword or argument must be provided as
418 a separate command line argument, with no leading or trailing
420 Keywords are case-sensitive, whereas arguments may
421 or may not be case-sensitive depending on their nature
422 (e.g.\& uid's are, hostnames are not).
424 Some arguments (e.g., port or address lists) are comma-separated
426 In this case, spaces after commas ',' are allowed to make
427 the line more readable.
428 You can also put the entire
429 command (including flags) into a single argument.
430 E.g., the following forms are equivalent:
431 .Bd -literal -offset indent
432 ipfw -q add deny src-ip 10.0.0.0/24,127.0.0.1/8
433 ipfw -q add deny src-ip 10.0.0.0/24, 127.0.0.1/8
434 ipfw "-q add deny src-ip 10.0.0.0/24, 127.0.0.1/8"
437 The format of firewall rules is the following:
438 .Bd -ragged -offset indent
441 .Op Cm set Ar set_number
442 .Op Cm prob Ar match_probability
444 .Op Cm log Op Cm logamount Ar number
454 where the body of the rule specifies which information is used
455 for filtering packets, among the following:
457 .Bl -tag -width "Source and dest. addresses and ports" -offset XXX -compact
458 .It Layer-2 header fields
460 .It IPv4 and IPv6 Protocol
462 .It Source and dest. addresses and ports
466 .It Transmit and receive interface
468 .It Misc. IP header fields
469 Version, type of service, datagram length, identification,
470 fragment flag (non-zero IP offset),
473 .It IPv6 Extension headers
474 Fragmentation, Hop-by-Hop options,
475 Routing Headers, Source routing rthdr0, Mobile IPv6 rthdr2, IPSec options.
477 .It Misc. TCP header fields
478 TCP flags (SYN, FIN, ACK, RST, etc.),
479 sequence number, acknowledgment number,
487 When the packet can be associated with a local socket.
489 Whether a packet came from a divert socket (e.g.,
491 .It Fib annotation state
492 Whether a packet has been tagged for using a specific FIB (routing table)
493 in future forwarding decisions.
496 Note that some of the above information, e.g.\& source MAC or IP addresses and
497 TCP/UDP ports, can be easily spoofed, so filtering on those fields
498 alone might not guarantee the desired results.
499 .Bl -tag -width indent
501 Each rule is associated with a
503 in the range 1..65535, with the latter reserved for the
506 Rules are checked sequentially by rule number.
507 Multiple rules can have the same number, in which case they are
508 checked (and listed) according to the order in which they have
510 If a rule is entered without specifying a number, the kernel will
511 assign one in such a way that the rule becomes the last one
515 Automatic rule numbers are assigned by incrementing the last
516 non-default rule number by the value of the sysctl variable
517 .Ar net.inet.ip.fw.autoinc_step
518 which defaults to 100.
519 If this is not possible (e.g.\& because we would go beyond the
520 maximum allowed rule number), the number of the last
521 non-default value is used instead.
522 .It Cm set Ar set_number
523 Each rule is associated with a
526 Sets can be individually disabled and enabled, so this parameter
527 is of fundamental importance for atomic ruleset manipulation.
528 It can be also used to simplify deletion of groups of rules.
529 If a rule is entered without specifying a set number,
532 Set 31 is special in that it cannot be disabled,
533 and rules in set 31 are not deleted by the
535 command (but you can delete them with the
536 .Nm ipfw delete set 31
538 Set 31 is also used for the
541 .It Cm prob Ar match_probability
542 A match is only declared with the specified probability
543 (floating point number between 0 and 1).
544 This can be useful for a number of applications such as
545 random packet drop or
548 to simulate the effect of multiple paths leading to out-of-order
551 Note: this condition is checked before any other condition, including
552 ones such as keep-state or check-state which might have side effects.
553 .It Cm log Op Cm logamount Ar number
554 Packets matching a rule with the
556 keyword will be made available for logging in two ways:
557 if the sysctl variable
558 .Va net.inet.ip.fw.verbose
559 is set to 0 (default), one can use
564 This pseudo interface can be created after a boot
565 manually by using the following command:
566 .Bd -literal -offset indent
567 # ifconfig ipfw0 create
570 Or, automatically at boot time by adding the following
574 .Bd -literal -offset indent
578 There is no overhead if no
580 is attached to the pseudo interface.
583 .Va net.inet.ip.fw.verbose
584 is set to 1, packets will be logged to
588 facility up to a maximum of
593 is specified, the limit is taken from the sysctl variable
594 .Va net.inet.ip.fw.verbose_limit .
595 In both cases, a value of 0 means unlimited logging.
597 Once the limit is reached, logging can be re-enabled by
598 clearing the logging counter or the packet counter for that entry, see the
602 Note: logging is done after all other packet matching conditions
603 have been successfully verified, and before performing the final
604 action (accept, deny, etc.) on the packet.
606 When a packet matches a rule with the
608 keyword, the numeric tag for the given
610 in the range 1..65534 will be attached to the packet.
611 The tag acts as an internal marker (it is not sent out over
612 the wire) that can be used to identify these packets later on.
613 This can be used, for example, to provide trust between interfaces
614 and to start doing policy-based filtering.
615 A packet can have multiple tags at the same time.
616 Tags are "sticky", meaning once a tag is applied to a packet by a
617 matching rule it exists until explicit removal.
618 Tags are kept with the packet everywhere within the kernel, but are
619 lost when packet leaves the kernel, for example, on transmitting
620 packet out to the network or sending packet to a
624 To check for previously applied tags, use the
627 To delete previously applied tag, use the
631 Note: since tags are kept with the packet everywhere in kernelspace,
632 they can be set and unset anywhere in the kernel network subsystem
635 facility), not only by means of the
641 For example, there can be a specialized
643 node doing traffic analyzing and tagging for later inspecting
645 .It Cm untag Ar number
646 When a packet matches a rule with the
648 keyword, the tag with the number
650 is searched among the tags attached to this packet and,
651 if found, removed from it.
652 Other tags bound to packet, if present, are left untouched.
654 When a packet matches a rule with the
656 keyword, the ALTQ identifier for the given
661 Note that this ALTQ tag is only meaningful for packets going "out" of IPFW,
662 and not being rejected or going to divert sockets.
663 Note that if there is insufficient memory at the time the packet is
664 processed, it will not be tagged, so it is wise to make your ALTQ
665 "default" queue policy account for this.
668 rules match a single packet, only the first one adds the ALTQ classification
670 In doing so, traffic may be shaped by using
671 .Cm count Cm altq Ar queue
672 rules for classification early in the ruleset, then later applying
673 the filtering decision.
678 rules may come later and provide the actual filtering decisions in
679 addition to the fallback ALTQ tag.
683 to set up the queues before IPFW will be able to look them up by name,
684 and if the ALTQ disciplines are rearranged, the rules in containing the
685 queue identifiers in the kernel will likely have gone stale and need
687 Stale queue identifiers will probably result in misclassification.
689 All system ALTQ processing can be turned on or off via
694 .Cm disable Ar altq .
696 .Va net.inet.ip.fw.one_pass
697 is irrelevant to ALTQ traffic shaping, as the actual rule action is followed
698 always after adding an ALTQ tag.
701 A rule can be associated with one of the following actions, which
702 will be executed when the packet matches the body of the rule.
703 .Bl -tag -width indent
704 .It Cm allow | accept | pass | permit
705 Allow packets that match rule.
706 The search terminates.
708 Checks the packet against the dynamic ruleset.
709 If a match is found, execute the action associated with
710 the rule which generated this dynamic rule, otherwise
711 move to the next rule.
714 rules do not have a body.
717 rule is found, the dynamic ruleset is checked at the first
723 Update counters for all packets that match rule.
724 The search continues with the next rule.
726 Discard packets that match this rule.
727 The search terminates.
728 .It Cm divert Ar port
729 Divert packets that match this rule to the
733 The search terminates.
734 .It Cm fwd | forward Ar ipaddr | tablearg Ns Op , Ns Ar port
735 Change the next-hop on matching packets to
737 which can be an IP address or a host name.
738 For IPv4, the next hop can also be supplied by the last table
739 looked up for the packet by using the
741 keyword instead of an explicit address.
742 The search terminates if this rule matches.
746 is a local address, then matching packets will be forwarded to
748 (or the port number in the packet if one is not specified in the rule)
749 on the local machine.
753 is not a local address, then the port number
754 (if specified) is ignored, and the packet will be
755 forwarded to the remote address, using the route as found in
756 the local routing table for that IP.
760 rule will not match layer-2 packets (those received
761 on ether_input, ether_output, or bridged).
765 action does not change the contents of the packet at all.
766 In particular, the destination address remains unmodified, so
767 packets forwarded to another system will usually be rejected by that system
768 unless there is a matching rule on that system to capture them.
769 For packets forwarded locally,
770 the local address of the socket will be
771 set to the original destination address of the packet.
774 entry look rather weird but is intended for
775 use with transparent proxy servers.
776 .It Cm nat Ar nat_nr | tablearg
779 (for network address translation, address redirect, etc.):
781 .Sx NETWORK ADDRESS TRANSLATION (NAT)
782 Section for further information.
783 .It Cm pipe Ar pipe_nr
787 (for bandwidth limitation, delay, etc.).
789 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
790 Section for further information.
791 The search terminates; however, on exit from the pipe and if
795 .Va net.inet.ip.fw.one_pass
796 is not set, the packet is passed again to the firewall code
797 starting from the next rule.
798 .It Cm queue Ar queue_nr
802 (for bandwidth limitation using WF2Q+).
808 Discard packets that match this rule, and if the
809 packet is a TCP packet, try to send a TCP reset (RST) notice.
810 The search terminates.
812 Discard packets that match this rule, and if the
813 packet is a TCP packet, try to send a TCP reset (RST) notice.
814 The search terminates.
815 .It Cm skipto Ar number | tablearg
816 Skip all subsequent rules numbered less than
818 The search continues with the first rule numbered
821 It is possible to use the
823 keyword with a skipto for a
825 skipto, but care should be used, as no destination caching
826 is possible in this case so the rules are always walked to find it,
829 .It Cm call Ar number | tablearg
830 The current rule number is saved in the internal stack and
831 ruleset processing continues with the first rule numbered
834 If later a rule with the
836 action is encountered, the processing returns to the first rule
839 rule plus one or higher
840 (the same behaviour as with packets returning from
845 This could be used to make somewhat like an assembly language
847 calls to rules with common checks for different interfaces, etc.
849 Rule with any number could be called, not just forward jumps as with
851 So, to prevent endless loops in case of mistakes, both
855 actions don't do any jumps and simply go to the next rule if memory
856 cannot be allocated or stack overflowed/underflowed.
858 Internally stack for rule numbers is implemented using
860 facility and currently has size of 16 entries.
861 As mbuf tags are lost when packet leaves the kernel,
863 should not be used in subroutines to avoid endless loops
864 and other undesired effects.
866 Takes rule number saved to internal stack by the last
868 action and returns ruleset processing to the first rule
869 with number greater than number of corresponding
872 See description of the
874 action for more details.
880 and thus are unconditional, but
882 command-line utility currently requires every action except
885 While it is sometimes useful to return only on some packets,
886 usually you want to print just
889 A workaround for this is to use new syntax and
892 .Bd -literal -offset indent
893 # Add a rule without actual body
894 ipfw add 2999 return via any
896 # List rules without "from any to any" part
900 This cosmetic annoyance may be fixed in future releases.
902 Send a copy of packets matching this rule to the
906 The search continues with the next rule.
907 .It Cm unreach Ar code
908 Discard packets that match this rule, and try to send an ICMP
909 unreachable notice with code
913 is a number from 0 to 255, or one of these aliases:
914 .Cm net , host , protocol , port ,
915 .Cm needfrag , srcfail , net-unknown , host-unknown ,
916 .Cm isolated , net-prohib , host-prohib , tosnet ,
917 .Cm toshost , filter-prohib , host-precedence
919 .Cm precedence-cutoff .
920 The search terminates.
921 .It Cm unreach6 Ar code
922 Discard packets that match this rule, and try to send an ICMPv6
923 unreachable notice with code
927 is a number from 0, 1, 3 or 4, or one of these aliases:
928 .Cm no-route, admin-prohib, address
931 The search terminates.
932 .It Cm netgraph Ar cookie
933 Divert packet into netgraph with given
935 The search terminates.
936 If packet is later returned from netgraph it is either
937 accepted or continues with the next rule, depending on
938 .Va net.inet.ip.fw.one_pass
940 .It Cm ngtee Ar cookie
941 A copy of packet is diverted into netgraph, original
942 packet continues with the next rule.
945 for more information on
950 .It Cm setfib Ar fibnum | tablearg
951 The packet is tagged so as to use the FIB (routing table)
953 in any subsequent forwarding decisions.
954 In the current implementation, this is limited to the values 0 through 15, see
956 Processing continues at the next rule.
957 It is possible to use the
960 If the tablearg value is not within the compiled range of fibs,
961 the packet's fib is set to 0.
962 .It Cm setdscp Ar DSCP | number | tablearg
963 Set specified DiffServ codepoint for an IPv4/IPv6 packet.
964 Processing continues at the next rule.
965 Supported values are:
1011 Additionally, DSCP value can be specified by number (0..64).
1012 It is also possible to use the
1014 keyword with setdscp.
1015 If the tablearg value is not within the 0..64 range, lower 6 bits of supplied
1018 Queue and reassemble IP fragments.
1019 If the packet is not fragmented, counters are updated and
1020 processing continues with the next rule.
1021 If the packet is the last logical fragment, the packet is reassembled and, if
1022 .Va net.inet.ip.fw.one_pass
1023 is set to 0, processing continues with the next rule.
1024 Otherwise, the packet is allowed to pass and the search terminates.
1025 If the packet is a fragment in the middle of a logical group of fragments,
1027 processing stops immediately.
1029 Fragment handling can be tuned via
1030 .Va net.inet.ip.maxfragpackets
1032 .Va net.inet.ip.maxfragsperpacket
1033 which limit, respectively, the maximum number of processable
1034 fragments (default: 800) and
1035 the maximum number of fragments per packet (default: 16).
1037 NOTA BENE: since fragments do not contain port numbers,
1038 they should be avoided with the
1041 Alternatively, direction-based (like
1045 ) and source-based (like
1047 ) match patterns can be used to select fragments.
1049 Usually a simple rule like:
1050 .Bd -literal -offset indent
1051 # reassemble incoming fragments
1052 ipfw add reass all from any to any in
1055 is all you need at the beginning of your ruleset.
1058 The body of a rule contains zero or more patterns (such as
1059 specific source and destination addresses or ports,
1060 protocol options, incoming or outgoing interfaces, etc.)
1061 that the packet must match in order to be recognised.
1062 In general, the patterns are connected by (implicit)
1064 operators -- i.e., all must match in order for the
1066 Individual patterns can be prefixed by the
1068 operator to reverse the result of the match, as in
1070 .Dl "ipfw add 100 allow ip from not 1.2.3.4 to any"
1072 Additionally, sets of alternative match patterns
1074 can be constructed by putting the patterns in
1075 lists enclosed between parentheses ( ) or braces { }, and
1078 operator as follows:
1080 .Dl "ipfw add 100 allow ip from { x or not y or z } to any"
1082 Only one level of parentheses is allowed.
1083 Beware that most shells have special meanings for parentheses
1084 or braces, so it is advisable to put a backslash \\ in front of them
1085 to prevent such interpretations.
1087 The body of a rule must in general include a source and destination
1091 can be used in various places to specify that the content of
1092 a required field is irrelevant.
1094 The rule body has the following format:
1095 .Bd -ragged -offset indent
1096 .Op Ar proto Cm from Ar src Cm to Ar dst
1100 The first part (proto from src to dst) is for backward
1101 compatibility with earlier versions of
1105 any match pattern (including MAC headers, IP protocols,
1106 addresses and ports) can be specified in the
1110 Rule fields have the following meaning:
1111 .Bl -tag -width indent
1112 .It Ar proto : protocol | Cm { Ar protocol Cm or ... }
1113 .It Ar protocol : Oo Cm not Oc Ar protocol-name | protocol-number
1114 An IP protocol specified by number or name
1115 (for a complete list see
1116 .Pa /etc/protocols ) ,
1117 or one of the following keywords:
1118 .Bl -tag -width indent
1120 Matches IPv4 packets.
1122 Matches IPv6 packets.
1131 option will be treated as inner protocol.
1139 .Cm { Ar protocol Cm or ... }
1142 is provided for convenience only but its use is deprecated.
1143 .It Ar src No and Ar dst : Bro Cm addr | Cm { Ar addr Cm or ... } Brc Op Oo Cm not Oc Ar ports
1144 An address (or a list, see below)
1145 optionally followed by
1151 with multiple addresses) is provided for convenience only and
1152 its use is discouraged.
1153 .It Ar addr : Oo Cm not Oc Bro
1154 .Cm any | me | me6 |
1155 .Cm table Ns Pq Ar number Ns Op , Ns Ar value
1156 .Ar | addr-list | addr-set
1158 .Bl -tag -width indent
1160 matches any IP address.
1162 matches any IP address configured on an interface in the system.
1164 matches any IPv6 address configured on an interface in the system.
1165 The address list is evaluated at the time the packet is
1167 .It Cm table Ns Pq Ar number Ns Op , Ns Ar value
1168 Matches any IPv4 address for which an entry exists in the lookup table
1170 If an optional 32-bit unsigned
1172 is also specified, an entry will match only if it has this value.
1175 section below for more information on lookup tables.
1177 .It Ar addr-list : ip-addr Ns Op Ns , Ns Ar addr-list
1179 A host or subnet address specified in one of the following ways:
1180 .Bl -tag -width indent
1181 .It Ar numeric-ip | hostname
1182 Matches a single IPv4 address, specified as dotted-quad or a hostname.
1183 Hostnames are resolved at the time the rule is added to the firewall list.
1184 .It Ar addr Ns / Ns Ar masklen
1185 Matches all addresses with base
1187 (specified as an IP address, a network number, or a hostname)
1191 As an example, 1.2.3.4/25 or 1.2.3.0/25 will match
1192 all IP numbers from 1.2.3.0 to 1.2.3.127 .
1193 .It Ar addr Ns : Ns Ar mask
1194 Matches all addresses with base
1196 (specified as an IP address, a network number, or a hostname)
1199 specified as a dotted quad.
1200 As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match
1202 This form is advised only for non-contiguous
1204 It is better to resort to the
1205 .Ar addr Ns / Ns Ar masklen
1206 format for contiguous masks, which is more compact and less
1209 .It Ar addr-set : addr Ns Oo Ns / Ns Ar masklen Oc Ns Cm { Ns Ar list Ns Cm }
1210 .It Ar list : Bro Ar num | num-num Brc Ns Op Ns , Ns Ar list
1211 Matches all addresses with base address
1213 (specified as an IP address, a network number, or a hostname)
1214 and whose last byte is in the list between braces { } .
1215 Note that there must be no spaces between braces and
1216 numbers (spaces after commas are allowed).
1217 Elements of the list can be specified as single entries
1221 field is used to limit the size of the set of addresses,
1222 and can have any value between 24 and 32.
1224 it will be assumed as 24.
1226 This format is particularly useful to handle sparse address sets
1227 within a single rule.
1228 Because the matching occurs using a
1229 bitmask, it takes constant time and dramatically reduces
1230 the complexity of rulesets.
1232 As an example, an address specified as 1.2.3.4/24{128,35-55,89}
1233 or 1.2.3.0/24{128,35-55,89}
1234 will match the following IP addresses:
1236 1.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .
1237 .It Ar addr6-list : ip6-addr Ns Op Ns , Ns Ar addr6-list
1239 A host or subnet specified one of the following ways:
1240 .Bl -tag -width indent
1241 .It Ar numeric-ip | hostname
1242 Matches a single IPv6 address as allowed by
1245 Hostnames are resolved at the time the rule is added to the firewall
1247 .It Ar addr Ns / Ns Ar masklen
1248 Matches all IPv6 addresses with base
1250 (specified as allowed by
1258 No support for sets of IPv6 addresses is provided because IPv6 addresses
1259 are typically random past the initial prefix.
1260 .It Ar ports : Bro Ar port | port Ns \&- Ns Ar port Ns Brc Ns Op , Ns Ar ports
1261 For protocols which support port numbers (such as TCP and UDP), optional
1263 may be specified as one or more ports or port ranges, separated
1264 by commas but no spaces, and an optional
1269 notation specifies a range of ports (including boundaries).
1273 may be used instead of numeric port values.
1274 The length of the port list is limited to 30 ports or ranges,
1275 though one can specify larger ranges by using an
1279 section of the rule.
1283 can be used to escape the dash
1285 character in a service name (from a shell, the backslash must be
1286 typed twice to avoid the shell itself interpreting it as an escape
1289 .Dl "ipfw add count tcp from any ftp\e\e-data-ftp to any"
1291 Fragmented packets which have a non-zero offset (i.e., not the first
1292 fragment) will never match a rule which has one or more port
1296 option for details on matching fragmented packets.
1298 .Ss RULE OPTIONS (MATCH PATTERNS)
1299 Additional match patterns can be used within
1301 Zero or more of these so-called
1303 can be present in a rule, optionally prefixed by the
1305 operand, and possibly grouped into
1308 The following match patterns can be used (listed in alphabetical order):
1309 .Bl -tag -width indent
1310 .It Cm // this is a comment.
1311 Inserts the specified text as a comment in the rule.
1312 Everything following // is considered as a comment and stored in the rule.
1313 You can have comment-only rules, which are listed as having a
1315 action followed by the comment.
1320 Matches only packets generated by a divert socket.
1321 .It Cm diverted-loopback
1322 Matches only packets coming from a divert socket back into the IP stack
1324 .It Cm diverted-output
1325 Matches only packets going from a divert socket back outward to the IP
1326 stack output for delivery.
1327 .It Cm dst-ip Ar ip-address
1328 Matches IPv4 packets whose destination IP is one of the address(es)
1329 specified as argument.
1330 .It Bro Cm dst-ip6 | dst-ipv6 Brc Ar ip6-address
1331 Matches IPv6 packets whose destination IP is one of the address(es)
1332 specified as argument.
1333 .It Cm dst-port Ar ports
1334 Matches IP packets whose destination port is one of the port(s)
1335 specified as argument.
1337 Matches TCP packets that have the RST or ACK bits set.
1338 .It Cm ext6hdr Ar header
1339 Matches IPv6 packets containing the extended header given by
1341 Supported headers are:
1347 any type of Routing Header
1349 Source routing Routing Header Type 0
1351 Mobile IPv6 Routing Header Type 2
1355 IPSec authentication headers
1357 and IPsec encapsulated security payload headers
1359 .It Cm fib Ar fibnum
1360 Matches a packet that has been tagged to use
1361 the given FIB (routing table) number.
1362 .It Cm flow-id Ar labels
1363 Matches IPv6 packets containing any of the flow labels given in
1366 is a comma separated list of numeric flow labels.
1368 Matches packets that are fragments and not the first
1369 fragment of an IP datagram.
1370 Note that these packets will not have
1371 the next protocol header (e.g.\& TCP, UDP) so options that look into
1372 these headers cannot match.
1374 Matches all TCP or UDP packets sent by or received for a
1378 may be specified by name or number.
1379 .It Cm jail Ar prisonID
1380 Matches all TCP or UDP packets sent by or received for the
1381 jail whos prison ID is
1383 .It Cm icmptypes Ar types
1384 Matches ICMP packets whose ICMP type is in the list
1386 The list may be specified as any combination of
1387 individual types (numeric) separated by commas.
1388 .Em Ranges are not allowed .
1389 The supported ICMP types are:
1393 destination unreachable
1401 router advertisement
1405 time-to-live exceeded
1417 address mask request
1419 and address mask reply
1421 .It Cm icmp6types Ar types
1422 Matches ICMP6 packets whose ICMP6 type is in the list of
1424 The list may be specified as any combination of
1425 individual types (numeric) separated by commas.
1426 .Em Ranges are not allowed .
1428 Matches incoming or outgoing packets, respectively.
1432 are mutually exclusive (in fact,
1436 .It Cm ipid Ar id-list
1437 Matches IPv4 packets whose
1439 field has value included in
1441 which is either a single value or a list of values or ranges
1442 specified in the same way as
1444 .It Cm iplen Ar len-list
1445 Matches IP packets whose total length, including header and data, is
1448 which is either a single value or a list of values or ranges
1449 specified in the same way as
1451 .It Cm ipoptions Ar spec
1452 Matches packets whose IPv4 header contains the comma separated list of
1453 options specified in
1455 The supported IP options are:
1458 (strict source route),
1460 (loose source route),
1462 (record packet route) and
1465 The absence of a particular option may be denoted
1468 .It Cm ipprecedence Ar precedence
1469 Matches IPv4 packets whose precedence field is equal to
1472 Matches packets that have IPSEC history associated with them
1473 (i.e., the packet comes encapsulated in IPSEC, the kernel
1474 has IPSEC support and IPSEC_FILTERTUNNEL option, and can correctly
1477 Note that specifying
1479 is different from specifying
1481 as the latter will only look at the specific IP protocol field,
1482 irrespective of IPSEC kernel support and the validity of the IPSEC data.
1484 Further note that this flag is silently ignored in kernels without
1486 It does not affect rule processing when given and the
1487 rules are handled as if with no
1490 .It Cm iptos Ar spec
1491 Matches IPv4 packets whose
1493 field contains the comma separated list of
1494 service types specified in
1496 The supported IP types of service are:
1499 .Pq Dv IPTOS_LOWDELAY ,
1501 .Pq Dv IPTOS_THROUGHPUT ,
1503 .Pq Dv IPTOS_RELIABILITY ,
1505 .Pq Dv IPTOS_MINCOST ,
1507 .Pq Dv IPTOS_ECN_CE .
1508 The absence of a particular type may be denoted
1511 .It Cm dscp spec Ns Op , Ns Ar spec
1512 Matches IPv4/IPv6 packets whose
1514 field value is contained in
1517 Multiple values can be specified via
1518 the comma separated list.
1519 Value can be one of keywords used in
1521 action or exact number.
1522 .It Cm ipttl Ar ttl-list
1523 Matches IPv4 packets whose time to live is included in
1525 which is either a single value or a list of values or ranges
1526 specified in the same way as
1528 .It Cm ipversion Ar ver
1529 Matches IP packets whose IP version field is
1532 Upon a match, the firewall will create a dynamic rule, whose
1533 default behaviour is to match bidirectional traffic between
1534 source and destination IP/port using the same protocol.
1535 The rule has a limited lifetime (controlled by a set of
1537 variables), and the lifetime is refreshed every time a matching
1540 Matches only layer2 packets, i.e., those passed to
1542 from ether_demux() and ether_output_frame().
1543 .It Cm limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N
1544 The firewall will only allow
1546 connections with the same
1547 set of parameters as specified in the rule.
1549 of source and destination addresses and ports can be
1552 only IPv4 flows are supported.
1553 .It Cm lookup Bro Cm dst-ip | dst-port | src-ip | src-port | uid | jail Brc Ar N
1554 Search an entry in lookup table
1556 that matches the field specified as argument.
1557 If not found, the match fails.
1558 Otherwise, the match succeeds and
1560 is set to the value extracted from the table.
1562 This option can be useful to quickly dispatch traffic based on
1563 certain packet fields.
1566 section below for more information on lookup tables.
1567 .It Cm { MAC | mac } Ar dst-mac src-mac
1568 Match packets with a given
1572 addresses, specified as the
1574 keyword (matching any MAC address), or six groups of hex digits
1575 separated by colons,
1576 and optionally followed by a mask indicating the significant bits.
1577 The mask may be specified using either of the following methods:
1578 .Bl -enum -width indent
1582 followed by the number of significant bits.
1583 For example, an address with 33 significant bits could be specified as:
1585 .Dl "MAC 10:20:30:40:50:60/33 any"
1590 followed by a bitmask specified as six groups of hex digits separated
1592 For example, an address in which the last 16 bits are significant could
1595 .Dl "MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any"
1597 Note that the ampersand character has a special meaning in many shells
1598 and should generally be escaped.
1601 Note that the order of MAC addresses (destination first,
1603 the same as on the wire, but the opposite of the one used for
1605 .It Cm mac-type Ar mac-type
1606 Matches packets whose Ethernet Type field
1607 corresponds to one of those specified as argument.
1609 is specified in the same way as
1611 (i.e., one or more comma-separated single values or ranges).
1612 You can use symbolic names for known values such as
1613 .Em vlan , ipv4, ipv6 .
1614 Values can be entered as decimal or hexadecimal (if prefixed by 0x),
1615 and they are always printed as hexadecimal (unless the
1617 option is used, in which case symbolic resolution will be attempted).
1618 .It Cm proto Ar protocol
1619 Matches packets with the corresponding IP protocol.
1620 .It Cm recv | xmit | via Brq Ar ifX | Ar if Ns Cm * | Ar table Ns Pq Ar number Ns Op , Ns Ar value | Ar ipno | Ar any
1621 Matches packets received, transmitted or going through,
1622 respectively, the interface specified by exact name
1626 by IP address, or through some interface.
1630 keyword causes the interface to always be checked.
1637 then only the receive or transmit interface (respectively)
1639 By specifying both, it is possible to match packets based on
1640 both receive and transmit interface, e.g.:
1642 .Dl "ipfw add deny ip from any to any out recv ed0 xmit ed1"
1646 interface can be tested on either incoming or outgoing packets,
1649 interface can only be tested on outgoing packets.
1654 is invalid) whenever
1658 A packet might not have a receive or transmit interface: packets
1659 originating from the local host have no receive interface,
1660 while packets destined for the local host have no transmit
1663 Matches TCP packets that have the SYN bit set but no ACK bit.
1664 This is the short form of
1665 .Dq Li tcpflags\ syn,!ack .
1667 Matches packets that are associated to a local socket and
1668 for which the SO_USER_COOKIE socket option has been set
1669 to a non-zero value.
1670 As a side effect, the value of the
1671 option is made available as
1673 value, which in turn can be used as
1678 .It Cm src-ip Ar ip-address
1679 Matches IPv4 packets whose source IP is one of the address(es)
1680 specified as an argument.
1681 .It Cm src-ip6 Ar ip6-address
1682 Matches IPv6 packets whose source IP is one of the address(es)
1683 specified as an argument.
1684 .It Cm src-port Ar ports
1685 Matches IP packets whose source port is one of the port(s)
1686 specified as argument.
1687 .It Cm tagged Ar tag-list
1688 Matches packets whose tags are included in
1690 which is either a single value or a list of values or ranges
1691 specified in the same way as
1693 Tags can be applied to the packet using
1695 rule action parameter (see it's description for details on tags).
1696 .It Cm tcpack Ar ack
1698 Match if the TCP header acknowledgment number field is set to
1700 .It Cm tcpdatalen Ar tcpdatalen-list
1701 Matches TCP packets whose length of TCP data is
1702 .Ar tcpdatalen-list ,
1703 which is either a single value or a list of values or ranges
1704 specified in the same way as
1706 .It Cm tcpflags Ar spec
1708 Match if the TCP header contains the comma separated list of
1711 The supported TCP flags are:
1720 The absence of a particular flag may be denoted
1723 A rule which contains a
1725 specification can never match a fragmented packet which has
1729 option for details on matching fragmented packets.
1730 .It Cm tcpseq Ar seq
1732 Match if the TCP header sequence number field is set to
1734 .It Cm tcpwin Ar tcpwin-list
1735 Matches TCP packets whose header window field is set to
1737 which is either a single value or a list of values or ranges
1738 specified in the same way as
1740 .It Cm tcpoptions Ar spec
1742 Match if the TCP header contains the comma separated list of
1743 options specified in
1745 The supported TCP options are:
1748 (maximum segment size),
1750 (tcp window advertisement),
1754 (rfc1323 timestamp) and
1756 (rfc1644 t/tcp connection count).
1757 The absence of a particular option may be denoted
1761 Match all TCP or UDP packets sent by or received for a
1765 may be matched by name or identification number.
1767 For incoming packets,
1768 a routing table lookup is done on the packet's source address.
1769 If the interface on which the packet entered the system matches the
1770 outgoing interface for the route,
1772 If the interfaces do not match up,
1773 the packet does not match.
1774 All outgoing packets or packets with no incoming interface match.
1776 The name and functionality of the option is intentionally similar to
1777 the Cisco IOS command:
1779 .Dl ip verify unicast reverse-path
1781 This option can be used to make anti-spoofing rules to reject all
1782 packets with source addresses not from this interface.
1786 For incoming packets,
1787 a routing table lookup is done on the packet's source address.
1788 If a route to the source address exists, but not the default route
1789 or a blackhole/reject route, the packet matches.
1790 Otherwise, the packet does not match.
1791 All outgoing packets match.
1793 The name and functionality of the option is intentionally similar to
1794 the Cisco IOS command:
1796 .Dl ip verify unicast source reachable-via any
1798 This option can be used to make anti-spoofing rules to reject all
1799 packets whose source address is unreachable.
1801 For incoming packets, the packet's source address is checked if it
1802 belongs to a directly connected network.
1803 If the network is directly connected, then the interface the packet
1804 came on in is compared to the interface the network is connected to.
1805 When incoming interface and directly connected interface are not the
1806 same, the packet does not match.
1807 Otherwise, the packet does match.
1808 All outgoing packets match.
1810 This option can be used to make anti-spoofing rules to reject all
1811 packets that pretend to be from a directly connected network but do
1812 not come in through that interface.
1813 This option is similar to but more restricted than
1815 because it engages only on packets with source addresses of directly
1816 connected networks instead of all source addresses.
1819 Lookup tables are useful to handle large sparse sets of
1820 addresses or other search keys (e.g., ports, jail IDs, interface names).
1821 In the rest of this section we will use the term ``address''.
1822 There may be up to 65535 different lookup tables, numbered 0 to 65534.
1824 Each entry is represented by an
1825 .Ar addr Ns Op / Ns Ar masklen
1826 and will match all addresses with base
1828 (specified as an IPv4/IPv6 address, a hostname or an unsigned integer)
1834 is not specified, it defaults to 32 for IPv4 and 128 for IPv6.
1835 When looking up an IP address in a table, the most specific
1837 Associated with each entry is a 32-bit unsigned
1839 which can optionally be checked by a rule matching code.
1840 When adding an entry, if
1842 is not specified, it defaults to 0.
1844 An entry can be added to a table
1846 or removed from a table
1848 A table can be examined
1853 Internally, each table is stored in a Radix tree, the same way as
1854 the routing table (see
1857 Lookup tables currently support only ports, jail IDs, IPv4/IPv6 addresses
1858 and interface names.
1859 Wildcards is not supported for interface names.
1863 feature provides the ability to use a value, looked up in the table, as
1864 the argument for a rule action, action parameter or rule option.
1865 This can significantly reduce number of rules in some configurations.
1866 If two tables are used in a rule, the result of the second (destination)
1870 argument can be used with the following actions:
1871 .Cm nat, pipe , queue, divert, tee, netgraph, ngtee, fwd, skipto, setfib,
1879 it is possible to supply table entries with values
1880 that are in the form of IP addresses or hostnames.
1883 Section for example usage of tables and the tablearg keyword.
1887 action, the user should be aware that the code will walk the ruleset
1888 up to a rule equal to, or past, the given number,
1889 and should therefore try keep the
1890 ruleset compact between the skipto and the target rules.
1892 Each rule belongs to one of 32 different
1895 Set 31 is reserved for the default rule.
1897 By default, rules are put in set 0, unless you use the
1899 attribute when entering a new rule.
1900 Sets can be individually and atomically enabled or disabled,
1901 so this mechanism permits an easy way to store multiple configurations
1902 of the firewall and quickly (and atomically) switch between them.
1903 The command to enable/disable sets is
1904 .Bd -ragged -offset indent
1906 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
1913 sections can be specified.
1914 Command execution is atomic on all the sets specified in the command.
1915 By default, all sets are enabled.
1917 When you disable a set, its rules behave as if they do not exist
1918 in the firewall configuration, with only one exception:
1919 .Bd -ragged -offset indent
1920 dynamic rules created from a rule before it had been disabled
1921 will still be active until they expire.
1923 dynamic rules you have to explicitly delete the parent rule
1924 which generated them.
1927 The set number of rules can be changed with the command
1928 .Bd -ragged -offset indent
1931 .Brq Cm rule Ar rule-number | old-set
1935 Also, you can atomically swap two rulesets with the command
1936 .Bd -ragged -offset indent
1938 .Cm set swap Ar first-set second-set
1943 Section on some possible uses of sets of rules.
1944 .Sh STATEFUL FIREWALL
1945 Stateful operation is a way for the firewall to dynamically
1946 create rules for specific flows when packets that
1947 match a given pattern are detected.
1948 Support for stateful
1949 operation comes through the
1950 .Cm check-state , keep-state
1956 Dynamic rules are created when a packet matches a
1960 rule, causing the creation of a
1962 rule which will match all and only packets with
1966 .Em src-ip/src-port dst-ip/dst-port
1971 are used here only to denote the initial match addresses, but they
1972 are completely equivalent afterwards).
1973 Dynamic rules will be checked at the first
1974 .Cm check-state, keep-state
1977 occurrence, and the action performed upon a match will be the same
1978 as in the parent rule.
1980 Note that no additional attributes other than protocol and IP addresses
1981 and ports are checked on dynamic rules.
1983 The typical use of dynamic rules is to keep a closed firewall configuration,
1984 but let the first TCP SYN packet from the inside network install a
1985 dynamic rule for the flow so that packets belonging to that session
1986 will be allowed through the firewall:
1988 .Dl "ipfw add check-state"
1989 .Dl "ipfw add allow tcp from my-subnet to any setup keep-state"
1990 .Dl "ipfw add deny tcp from any to any"
1992 A similar approach can be used for UDP, where an UDP packet coming
1993 from the inside will install a dynamic rule to let the response through
1996 .Dl "ipfw add check-state"
1997 .Dl "ipfw add allow udp from my-subnet to any keep-state"
1998 .Dl "ipfw add deny udp from any to any"
2000 Dynamic rules expire after some time, which depends on the status
2001 of the flow and the setting of some
2005 .Sx SYSCTL VARIABLES
2007 For TCP sessions, dynamic rules can be instructed to periodically
2008 send keepalive packets to refresh the state of the rule when it is
2013 for more examples on how to use dynamic rules.
2014 .Sh TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
2016 is also the user interface for the
2018 traffic shaper, packet scheduler and network emulator, a subsystem that
2019 can artificially queue, delay or drop packets
2020 emulating the behaviour of certain network links
2021 or queueing systems.
2024 operates by first using the firewall to select packets
2025 using any match pattern that can be used in
2028 Matching packets are then passed to either of two
2029 different objects, which implement the traffic regulation:
2030 .Bl -hang -offset XXXX
2036 with given bandwidth and propagation delay,
2037 driven by a FIFO scheduler and a single queue with programmable
2038 queue size and packet loss rate.
2039 Packets are appended to the queue as they come out from
2041 and then transferred in FIFO order to the link at the desired rate.
2045 is an abstraction used to implement packet scheduling
2046 using one of several packet scheduling algorithms.
2049 are first grouped into flows according to a mask on the 5-tuple.
2050 Flows are then passed to the scheduler associated to the
2052 and each flow uses scheduling parameters (weight and others)
2053 as configured in the
2056 A scheduler in turn is connected to an emulated link,
2057 and arbitrates the link's bandwidth among backlogged flows according to
2058 weights and to the features of the scheduling algorithm in use.
2063 can be used to set hard limits to the bandwidth that a flow can use, whereas
2065 can be used to determine how different flows share the available bandwidth.
2067 A graphical representation of the binding of queues,
2068 flows, schedulers and links is below.
2069 .Bd -literal -offset indent
2070 (flow_mask|sched_mask) sched_mask
2071 +---------+ weight Wx +-------------+
2072 | |->-[flow]-->--| |-+
2073 -->--| QUEUE x | ... | | |
2074 | |->-[flow]-->--| SCHEDuler N | |
2076 ... | +--[LINK N]-->--
2077 +---------+ weight Wy | | +--[LINK N]-->--
2078 | |->-[flow]-->--| | |
2079 -->--| QUEUE y | ... | | |
2080 | |->-[flow]-->--| | |
2081 +---------+ +-------------+ |
2084 It is important to understand the role of the SCHED_MASK
2085 and FLOW_MASK, which are configured through the commands
2086 .Dl "ipfw sched N config mask SCHED_MASK ..."
2088 .Dl "ipfw queue X config mask FLOW_MASK ..." .
2090 The SCHED_MASK is used to assign flows to one or more
2091 scheduler instances, one for each
2092 value of the packet's 5-tuple after applying SCHED_MASK.
2093 As an example, using ``src-ip 0xffffff00'' creates one instance
2094 for each /24 destination subnet.
2096 The FLOW_MASK, together with the SCHED_MASK, is used to split
2098 As an example, using
2099 ``src-ip 0x000000ff''
2100 together with the previous SCHED_MASK makes a flow for
2101 each individual source address.
2102 In turn, flows for each /24
2103 subnet will be sent to the same scheduler instance.
2105 The above diagram holds even for the
2107 case, with the only restriction that a
2109 only supports a SCHED_MASK, and forces the use of a FIFO
2110 scheduler (these are for backward compatibility reasons;
2111 in fact, internally, a
2113 pipe is implemented exactly as above).
2115 There are two modes of
2123 mode tries to emulate a real link: the
2125 scheduler ensures that the packet will not leave the pipe faster than it
2126 would on the real link with a given bandwidth.
2129 mode allows certain packets to bypass the
2131 scheduler (if packet flow does not exceed pipe's bandwidth).
2132 This is the reason why the
2134 mode requires less CPU cycles per packet (on average) and packet latency
2135 can be significantly lower in comparison to a real link with the same
2141 mode can be enabled by setting the
2142 .Va net.inet.ip.dummynet.io_fast
2144 variable to a non-zero value.
2146 .Ss PIPE, QUEUE AND SCHEDULER CONFIGURATION
2152 configuration commands are the following:
2153 .Bd -ragged -offset indent
2154 .Cm pipe Ar number Cm config Ar pipe-configuration
2156 .Cm queue Ar number Cm config Ar queue-configuration
2158 .Cm sched Ar number Cm config Ar sched-configuration
2161 The following parameters can be configured for a pipe:
2163 .Bl -tag -width indent -compact
2164 .It Cm bw Ar bandwidth | device
2165 Bandwidth, measured in
2168 .Brq Cm bit/s | Byte/s .
2171 A value of 0 (default) means unlimited bandwidth.
2172 The unit must immediately follow the number, as in
2174 .Dl "ipfw pipe 1 config bw 300Kbit/s"
2176 If a device name is specified instead of a numeric value, as in
2178 .Dl "ipfw pipe 1 config bw tun0"
2180 then the transmit clock is supplied by the specified device.
2181 At the moment only the
2183 device supports this
2184 functionality, for use in conjunction with
2187 .It Cm delay Ar ms-delay
2188 Propagation delay, measured in milliseconds.
2189 The value is rounded to the next multiple of the clock tick
2190 (typically 10ms, but it is a good practice to run kernels
2192 .Dq "options HZ=1000"
2194 the granularity to 1ms or less).
2195 The default value is 0, meaning no delay.
2197 .It Cm burst Ar size
2198 If the data to be sent exceeds the pipe's bandwidth limit
2199 (and the pipe was previously idle), up to
2201 bytes of data are allowed to bypass the
2203 scheduler, and will be sent as fast as the physical link allows.
2204 Any additional data will be transmitted at the rate specified
2208 The burst size depends on how long the pipe has been idle;
2209 the effective burst size is calculated as follows:
2216 .It Cm profile Ar filename
2217 A file specifying the additional overhead incurred in the transmission
2218 of a packet on the link.
2220 Some link types introduce extra delays in the transmission
2221 of a packet, e.g., because of MAC level framing, contention on
2222 the use of the channel, MAC level retransmissions and so on.
2223 From our point of view, the channel is effectively unavailable
2224 for this extra time, which is constant or variable depending
2226 Additionally, packets may be dropped after this
2227 time (e.g., on a wireless link after too many retransmissions).
2228 We can model the additional delay with an empirical curve
2229 that represents its distribution.
2230 .Bd -literal -offset indent
2231 cumulative probability
2241 +-------*------------------->
2244 The empirical curve may have both vertical and horizontal lines.
2245 Vertical lines represent constant delay for a range of
2247 Horizontal lines correspond to a discontinuity in the delay
2248 distribution: the pipe will use the largest delay for a
2251 The file format is the following, with whitespace acting as
2252 a separator and '#' indicating the beginning a comment:
2253 .Bl -tag -width indent
2254 .It Cm name Ar identifier
2255 optional name (listed by "ipfw pipe show")
2256 to identify the delay distribution;
2258 the bandwidth used for the pipe.
2259 If not specified here, it must be present
2260 explicitly as a configuration parameter for the pipe;
2261 .It Cm loss-level Ar L
2262 the probability above which packets are lost.
2263 (0.0 <= L <= 1.0, default 1.0 i.e., no loss);
2265 the number of samples used in the internal
2266 representation of the curve (2..1024; default 100);
2267 .It Cm "delay prob" | "prob delay"
2268 One of these two lines is mandatory and defines
2269 the format of the following lines with data points.
2271 2 or more lines representing points in the curve,
2272 with either delay or probability first, according
2273 to the chosen format.
2274 The unit for delay is milliseconds.
2275 Data points do not need to be sorted.
2276 Also, the number of actual lines can be different
2277 from the value of the "samples" parameter:
2279 utility will sort and interpolate
2280 the curve as needed.
2283 Example of a profile file:
2284 .Bd -literal -offset indent
2289 0 200 # minimum overhead is 200ms
2295 #configuration file end
2299 The following parameters can be configured for a queue:
2301 .Bl -tag -width indent -compact
2302 .It Cm pipe Ar pipe_nr
2303 Connects a queue to the specified pipe.
2304 Multiple queues (with the same or different weights) can be connected to
2305 the same pipe, which specifies the aggregate rate for the set of queues.
2307 .It Cm weight Ar weight
2308 Specifies the weight to be used for flows matching this queue.
2309 The weight must be in the range 1..100, and defaults to 1.
2312 The following case-insensitive parameters can be configured for a
2315 .Bl -tag -width indent -compact
2316 .It Cm type Ar {fifo | wf2q+ | rr | qfq}
2317 specifies the scheduling algorithm to use.
2318 .Bl -tag -width indent -compact
2320 is just a FIFO scheduler (which means that all packets
2321 are stored in the same queue as they arrive to the scheduler).
2322 FIFO has O(1) per-packet time complexity, with very low
2323 constants (estimate 60-80ns on a 2GHz desktop machine)
2324 but gives no service guarantees.
2326 implements the WF2Q+ algorithm, which is a Weighted Fair Queueing
2327 algorithm which permits flows to share bandwidth according to
2329 Note that weights are not priorities; even a flow
2330 with a minuscule weight will never starve.
2331 WF2Q+ has O(log N) per-packet processing cost, where N is the number
2332 of flows, and is the default algorithm used by previous versions
2335 implements the Deficit Round Robin algorithm, which has O(1) processing
2336 costs (roughly, 100-150ns per packet)
2337 and permits bandwidth allocation according to weights, but
2338 with poor service guarantees.
2340 implements the QFQ algorithm, which is a very fast variant of
2341 WF2Q+, with similar service guarantees and O(1) processing
2342 costs (roughly, 200-250ns per packet).
2346 In addition to the type, all parameters allowed for a pipe can also
2347 be specified for a scheduler.
2349 Finally, the following parameters can be configured for both
2352 .Bl -tag -width XXXX -compact
2353 .It Cm buckets Ar hash-table-size
2354 Specifies the size of the hash table used for storing the
2356 Default value is 64 controlled by the
2359 .Va net.inet.ip.dummynet.hash_size ,
2360 allowed range is 16 to 65536.
2362 .It Cm mask Ar mask-specifier
2363 Packets sent to a given pipe or queue by an
2365 rule can be further classified into multiple flows, each of which is then
2369 A flow identifier is constructed by masking the IP addresses,
2370 ports and protocol types as specified with the
2372 options in the configuration of the pipe or queue.
2373 For each different flow identifier, a new pipe or queue is created
2374 with the same parameters as the original object, and matching packets
2379 are used, each flow will get the same bandwidth as defined by the pipe,
2382 are used, each flow will share the parent's pipe bandwidth evenly
2383 with other flows generated by the same queue (note that other queues
2384 with different weights might be connected to the same pipe).
2386 Available mask specifiers are a combination of one or more of the following:
2388 .Cm dst-ip Ar mask ,
2389 .Cm dst-ip6 Ar mask ,
2390 .Cm src-ip Ar mask ,
2391 .Cm src-ip6 Ar mask ,
2392 .Cm dst-port Ar mask ,
2393 .Cm src-port Ar mask ,
2394 .Cm flow-id Ar mask ,
2399 where the latter means all bits in all fields are significant.
2402 When a packet is dropped by a
2404 queue or pipe, the error
2405 is normally reported to the caller routine in the kernel, in the
2406 same way as it happens when a device queue fills up.
2408 option reports the packet as successfully delivered, which can be
2409 needed for some experimental setups where you want to simulate
2410 loss or congestion at a remote router.
2412 .It Cm plr Ar packet-loss-rate
2415 .Ar packet-loss-rate
2416 is a floating-point number between 0 and 1, with 0 meaning no
2417 loss, 1 meaning 100% loss.
2418 The loss rate is internally represented on 31 bits.
2420 .It Cm queue Brq Ar slots | size Ns Cm Kbytes
2425 Default value is 50 slots, which
2426 is the typical queue size for Ethernet devices.
2427 Note that for slow speed links you should keep the queue
2428 size short or your traffic might be affected by a significant
2430 E.g., 50 max-sized ethernet packets (1500 bytes) mean 600Kbit
2431 or 20s of queue on a 30Kbit/s pipe.
2432 Even worse effects can result if you get packets from an
2433 interface with a much larger MTU, e.g.\& the loopback interface
2434 with its 16KB packets.
2438 .Em net.inet.ip.dummynet.pipe_byte_limit
2440 .Em net.inet.ip.dummynet.pipe_slot_limit
2441 control the maximum lengths that can be specified.
2443 .It Cm red | gred Ar w_q Ns / Ns Ar min_th Ns / Ns Ar max_th Ns / Ns Ar max_p
2445 Make use of the RED (Random Early Detection) queue management algorithm.
2450 point numbers between 0 and 1 (inclusive), while
2454 are integer numbers specifying thresholds for queue management
2455 (thresholds are computed in bytes if the queue has been defined
2456 in bytes, in slots otherwise).
2457 The two parameters can also be of the same value if needed. The
2459 also supports the gentle RED variant (gred) and ECN (Explicit Congestion
2460 Notification) as optional. Three
2462 variables can be used to control the RED behaviour:
2463 .Bl -tag -width indent
2464 .It Va net.inet.ip.dummynet.red_lookup_depth
2465 specifies the accuracy in computing the average queue
2466 when the link is idle (defaults to 256, must be greater than zero)
2467 .It Va net.inet.ip.dummynet.red_avg_pkt_size
2468 specifies the expected average packet size (defaults to 512, must be
2470 .It Va net.inet.ip.dummynet.red_max_pkt_size
2471 specifies the expected maximum packet size, only used when queue
2472 thresholds are in bytes (defaults to 1500, must be greater than zero).
2476 When used with IPv6 data,
2478 currently has several limitations.
2479 Information necessary to route link-local packets to an
2480 interface is not available after processing by
2482 so those packets are dropped in the output path.
2483 Care should be taken to ensure that link-local packets are not passed to
2486 Here are some important points to consider when designing your
2490 Remember that you filter both packets going
2494 Most connections need packets going in both directions.
2496 Remember to test very carefully.
2497 It is a good idea to be near the console when doing this.
2498 If you cannot be near the console,
2499 use an auto-recovery script such as the one in
2500 .Pa /usr/share/examples/ipfw/change_rules.sh .
2502 Do not forget the loopback interface.
2507 There are circumstances where fragmented datagrams are unconditionally
2509 TCP packets are dropped if they do not contain at least 20 bytes of
2510 TCP header, UDP packets are dropped if they do not contain a full 8
2511 byte UDP header, and ICMP packets are dropped if they do not contain
2512 4 bytes of ICMP header, enough to specify the ICMP type, code, and
2514 These packets are simply logged as
2516 since there may not be enough good data in the packet to produce a
2517 meaningful log entry.
2519 Another type of packet is unconditionally dropped, a TCP packet with a
2520 fragment offset of one.
2521 This is a valid packet, but it only has one use, to try
2522 to circumvent firewalls.
2523 When logging is enabled, these packets are
2524 reported as being dropped by rule -1.
2526 If you are logged in over a network, loading the
2530 is probably not as straightforward as you would think.
2531 The following command line is recommended:
2532 .Bd -literal -offset indent
2534 ipfw add 32000 allow ip from any to any
2537 Along the same lines, doing an
2538 .Bd -literal -offset indent
2542 in similar surroundings is also a bad idea.
2546 filter list may not be modified if the system security level
2547 is set to 3 or higher
2550 for information on system security levels).
2552 .Sh PACKET DIVERSION
2555 socket bound to the specified port will receive all packets
2556 diverted to that port.
2557 If no socket is bound to the destination port, or if the divert module is
2558 not loaded, or if the kernel was not compiled with divert socket support,
2559 the packets are dropped.
2560 .Sh NETWORK ADDRESS TRANSLATION (NAT)
2562 support in-kernel NAT using the kernel version of
2565 The nat configuration command is the following:
2566 .Bd -ragged -offset indent
2571 .Ar nat-configuration
2575 The following parameters can be configured:
2576 .Bl -tag -width indent
2577 .It Cm ip Ar ip_address
2578 Define an ip address to use for aliasing.
2580 Use ip address of NIC for aliasing, dynamically changing
2581 it if NIC's ip address changes.
2583 Enable logging on this nat instance.
2585 Deny any incoming connection from outside world.
2587 Try to leave the alias port numbers unchanged from
2588 the actual local port numbers.
2590 Traffic on the local network not originating from an
2591 unregistered address spaces will be ignored.
2593 Reset table of the packet aliasing engine on address change.
2595 Reverse the way libalias handles aliasing.
2597 Obey transparent proxy rules only, packet aliasing is not performed.
2599 Skip instance in case of global state lookup (see below).
2602 Some specials value can be supplied instead of
2604 .Bl -tag -width indent
2606 Looks up translation state in all configured nat instances.
2607 If an entry is found, packet is aliased according to that entry.
2608 If no entry was found in any of the instances, packet is passed unchanged,
2609 and no new entry will be created.
2611 .Sx MULTIPLE INSTANCES
2614 for more information.
2616 Uses argument supplied in lookup table.
2619 section below for more information on lookup tables.
2622 To let the packet continue after being (de)aliased, set the sysctl variable
2623 .Va net.inet.ip.fw.one_pass
2625 For more information about aliasing modes, refer to
2629 for some examples about nat usage.
2630 .Ss REDIRECT AND LSNAT SUPPORT IN IPFW
2631 Redirect and LSNAT support follow closely the syntax used in
2635 for some examples on how to do redirect and lsnat.
2636 .Ss SCTP NAT SUPPORT
2637 SCTP nat can be configured in a similar manner to TCP through the
2640 The main difference is that
2642 does not do port translation.
2643 Since the local and global side ports will be the same,
2644 there is no need to specify both.
2645 Ports are redirected as follows:
2646 .Bd -ragged -offset indent
2652 .Cm redirect_port sctp
2653 .Ar ip_address [,addr_list] {[port | port-port] [,ports]}
2659 configuration can be done in real-time through the
2662 All may be changed dynamically, though the hash_table size will only
2667 .Sx SYSCTL VARIABLES
2670 Tunables can be set in
2676 before ipfw module gets loaded.
2677 .Bl -tag -width indent
2678 .It Va net.inet.ip.fw.default_to_accept: No 0
2679 Defines ipfw last rule behavior.
2680 This value overrides
2681 .Cd "options IPFW_DEFAULT_TO_(ACCEPT|DENY)"
2682 from kernel configuration file.
2683 .It Va net.inet.ip.fw.tables_max: No 128
2684 Defines number of tables available in ipfw.
2685 Number cannot exceed 65534.
2687 .Sh SYSCTL VARIABLES
2690 variables controls the behaviour of the firewall and
2692 .Pq Nm dummynet , bridge , sctp nat .
2693 These are shown below together with their default value
2694 (but always check with the
2696 command what value is actually in use) and meaning:
2697 .Bl -tag -width indent
2698 .It Va net.inet.ip.alias.sctp.accept_global_ootb_addip: No 0
2701 responds to receipt of global OOTB ASCONF-AddIP:
2702 .Bl -tag -width indent
2704 No response (unless a partially matching association exists -
2705 ports and vtags match but global address does not)
2708 will accept and process all OOTB global AddIP messages.
2711 Option 1 should never be selected as this forms a security risk.
2713 establish multiple fake associations by sending AddIP messages.
2714 .It Va net.inet.ip.alias.sctp.chunk_proc_limit: No 5
2715 Defines the maximum number of chunks in an SCTP packet that will be
2717 packet that matches an existing association.
2718 This value is enforced to be greater or equal than
2719 .Cm net.inet.ip.alias.sctp.initialising_chunk_proc_limit .
2721 a DoS risk yet setting too low a value may result in
2722 important control chunks in
2723 the packet not being located and parsed.
2724 .It Va net.inet.ip.alias.sctp.error_on_ootb: No 1
2727 responds to any Out-of-the-Blue (OOTB) packets with ErrorM packets.
2728 An OOTB packet is a packet that arrives with no existing association
2731 and is not an INIT or ASCONF-AddIP packet:
2732 .Bl -tag -width indent
2734 ErrorM is never sent in response to OOTB packets.
2736 ErrorM is only sent to OOTB packets received on the local side.
2738 ErrorM is sent to the local side and on the global side ONLY if there is a
2739 partial match (ports and vtags match but the source global IP does not).
2740 This value is only useful if the
2742 is tracking global IP addresses.
2744 ErrorM is sent in response to all OOTB packets on both
2745 the local and global side
2749 At the moment the default is 0, since the ErrorM packet is not yet
2750 supported by most SCTP stacks.
2751 When it is supported, and if not tracking
2752 global addresses, we recommend setting this value to 1 to allow
2753 multi-homed local hosts to function with the
2755 To track global addresses, we recommend setting this value to 2 to
2756 allow global hosts to be informed when they need to (re)send an
2758 Value 3 should never be chosen (except for debugging) as the
2760 will respond to all OOTB global packets (a DoS risk).
2761 .It Va net.inet.ip.alias.sctp.hashtable_size: No 2003
2762 Size of hash tables used for
2764 lookups (100 < prime_number > 1000001).
2767 size for any future created
2769 instance and therefore must be set prior to creating a
2772 The table sizes may be changed to suit specific needs.
2773 If there will be few
2774 concurrent associations, and memory is scarce, you may make these smaller.
2775 If there will be many thousands (or millions) of concurrent associations, you
2776 should make these larger.
2777 A prime number is best for the table size.
2779 update function will adjust your input value to the next highest prime number.
2780 .It Va net.inet.ip.alias.sctp.holddown_time: No 0
2781 Hold association in table for this many seconds after receiving a
2783 This allows endpoints to correct shutdown gracefully if a
2784 shutdown_complete is lost and retransmissions are required.
2785 .It Va net.inet.ip.alias.sctp.init_timer: No 15
2786 Timeout value while waiting for (INIT-ACK|AddIP-ACK).
2787 This value cannot be 0.
2788 .It Va net.inet.ip.alias.sctp.initialising_chunk_proc_limit: No 2
2789 Defines the maximum number of chunks in an SCTP packet that will be parsed when
2790 no existing association exists that matches that packet.
2792 will only be an INIT or ASCONF-AddIP packet.
2793 A higher value may become a DoS
2794 risk as malformed packets can consume processing resources.
2795 .It Va net.inet.ip.alias.sctp.param_proc_limit: No 25
2796 Defines the maximum number of parameters within a chunk that will be
2799 As for other similar sysctl variables, larger values pose a DoS risk.
2800 .It Va net.inet.ip.alias.sctp.log_level: No 0
2801 Level of detail in the system log messages (0 \- minimal, 1 \- event,
2802 2 \- info, 3 \- detail, 4 \- debug, 5 \- max debug).
2804 option in high loss environments.
2805 .It Va net.inet.ip.alias.sctp.shutdown_time: No 15
2806 Timeout value while waiting for SHUTDOWN-COMPLETE.
2807 This value cannot be 0.
2808 .It Va net.inet.ip.alias.sctp.track_global_addresses: No 0
2809 Enables/disables global IP address tracking within the
2812 upper limit on the number of addresses tracked for each association:
2813 .Bl -tag -width indent
2815 Global tracking is disabled
2817 Enables tracking, the maximum number of addresses tracked for each
2818 association is limited to this value
2821 This variable is fully dynamic, the new value will be adopted for all newly
2822 arriving associations, existing associations are treated
2823 as they were previously.
2824 Global tracking will decrease the number of collisions within the
2827 of increased processing load, memory usage, complexity, and possible
2830 problems in complex networks with multiple
2832 We recommend not tracking
2833 global IP addresses, this will still result in a fully functional
2835 .It Va net.inet.ip.alias.sctp.up_timer: No 300
2836 Timeout value to keep an association up with no traffic.
2837 This value cannot be 0.
2838 .It Va net.inet.ip.dummynet.expire : No 1
2839 Lazily delete dynamic pipes/queue once they have no pending traffic.
2840 You can disable this by setting the variable to 0, in which case
2841 the pipes/queues will only be deleted when the threshold is reached.
2842 .It Va net.inet.ip.dummynet.hash_size : No 64
2843 Default size of the hash table used for dynamic pipes/queues.
2844 This value is used when no
2846 option is specified when configuring a pipe/queue.
2847 .It Va net.inet.ip.dummynet.io_fast : No 0
2848 If set to a non-zero value,
2853 operation (see above) is enabled.
2854 .It Va net.inet.ip.dummynet.io_pkt
2855 Number of packets passed to
2857 .It Va net.inet.ip.dummynet.io_pkt_drop
2858 Number of packets dropped by
2860 .It Va net.inet.ip.dummynet.io_pkt_fast
2861 Number of packets bypassed by the
2864 .It Va net.inet.ip.dummynet.max_chain_len : No 16
2865 Target value for the maximum number of pipes/queues in a hash bucket.
2867 .Cm max_chain_len*hash_size
2868 is used to determine the threshold over which empty pipes/queues
2869 will be expired even when
2870 .Cm net.inet.ip.dummynet.expire=0 .
2871 .It Va net.inet.ip.dummynet.red_lookup_depth : No 256
2872 .It Va net.inet.ip.dummynet.red_avg_pkt_size : No 512
2873 .It Va net.inet.ip.dummynet.red_max_pkt_size : No 1500
2874 Parameters used in the computations of the drop probability
2875 for the RED algorithm.
2876 .It Va net.inet.ip.dummynet.pipe_byte_limit : No 1048576
2877 .It Va net.inet.ip.dummynet.pipe_slot_limit : No 100
2878 The maximum queue size that can be specified in bytes or packets.
2879 These limits prevent accidental exhaustion of resources such as mbufs.
2880 If you raise these limits,
2881 you should make sure the system is configured so that sufficient resources
2883 .It Va net.inet.ip.fw.autoinc_step : No 100
2884 Delta between rule numbers when auto-generating them.
2885 The value must be in the range 1..1000.
2886 .It Va net.inet.ip.fw.curr_dyn_buckets : Va net.inet.ip.fw.dyn_buckets
2887 The current number of buckets in the hash table for dynamic rules
2889 .It Va net.inet.ip.fw.debug : No 1
2890 Controls debugging messages produced by
2892 .It Va net.inet.ip.fw.default_rule : No 65535
2893 The default rule number (read-only).
2895 .Nm , the default rule is the last one, so its number
2896 can also serve as the highest number allowed for a rule.
2897 .It Va net.inet.ip.fw.dyn_buckets : No 256
2898 The number of buckets in the hash table for dynamic rules.
2899 Must be a power of 2, up to 65536.
2900 It only takes effect when all dynamic rules have expired, so you
2901 are advised to use a
2903 command to make sure that the hash table is resized.
2904 .It Va net.inet.ip.fw.dyn_count : No 3
2905 Current number of dynamic rules
2907 .It Va net.inet.ip.fw.dyn_keepalive : No 1
2908 Enables generation of keepalive packets for
2910 rules on TCP sessions.
2911 A keepalive is generated to both
2912 sides of the connection every 5 seconds for the last 20
2913 seconds of the lifetime of the rule.
2914 .It Va net.inet.ip.fw.dyn_max : No 8192
2915 Maximum number of dynamic rules.
2916 When you hit this limit, no more dynamic rules can be
2917 installed until old ones expire.
2918 .It Va net.inet.ip.fw.dyn_ack_lifetime : No 300
2919 .It Va net.inet.ip.fw.dyn_syn_lifetime : No 20
2920 .It Va net.inet.ip.fw.dyn_fin_lifetime : No 1
2921 .It Va net.inet.ip.fw.dyn_rst_lifetime : No 1
2922 .It Va net.inet.ip.fw.dyn_udp_lifetime : No 5
2923 .It Va net.inet.ip.fw.dyn_short_lifetime : No 30
2924 These variables control the lifetime, in seconds, of dynamic
2926 Upon the initial SYN exchange the lifetime is kept short,
2927 then increased after both SYN have been seen, then decreased
2928 again during the final FIN exchange or when a RST is received.
2930 .Em dyn_fin_lifetime
2932 .Em dyn_rst_lifetime
2933 must be strictly lower than 5 seconds, the period of
2934 repetition of keepalives.
2935 The firewall enforces that.
2936 .It Va net.inet.ip.fw.dyn_keep_states: No 0
2937 Keep dynamic states on rule/set deletion.
2938 States are relinked to default rule (65535).
2939 This can be handly for ruleset reload.
2940 Turned off by default.
2941 .It Va net.inet.ip.fw.enable : No 1
2942 Enables the firewall.
2943 Setting this variable to 0 lets you run your machine without
2944 firewall even if compiled in.
2945 .It Va net.inet6.ip6.fw.enable : No 1
2946 provides the same functionality as above for the IPv6 case.
2947 .It Va net.inet.ip.fw.one_pass : No 1
2948 When set, the packet exiting from the
2952 node is not passed though the firewall again.
2953 Otherwise, after an action, the packet is
2954 reinjected into the firewall at the next rule.
2955 .It Va net.inet.ip.fw.tables_max : No 128
2956 Maximum number of tables.
2957 .It Va net.inet.ip.fw.verbose : No 1
2958 Enables verbose messages.
2959 .It Va net.inet.ip.fw.verbose_limit : No 0
2960 Limits the number of messages produced by a verbose firewall.
2961 .It Va net.inet6.ip6.fw.deny_unknown_exthdrs : No 1
2962 If enabled packets with unknown IPv6 Extension Headers will be denied.
2963 .It Va net.link.ether.ipfw : No 0
2964 Controls whether layer-2 packets are passed to
2967 .It Va net.link.bridge.ipfw : No 0
2968 Controls whether bridged packets are passed to
2973 There are far too many possible uses of
2975 so this Section will only give a small set of examples.
2977 .Ss BASIC PACKET FILTERING
2978 This command adds an entry which denies all tcp packets from
2979 .Em cracker.evil.org
2980 to the telnet port of
2982 from being forwarded by the host:
2984 .Dl "ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet"
2986 This one disallows any connection from the entire cracker's
2989 .Dl "ipfw add deny ip from 123.45.67.0/24 to my.host.org"
2991 A first and efficient way to limit access (not using dynamic rules)
2992 is the use of the following rules:
2994 .Dl "ipfw add allow tcp from any to any established"
2995 .Dl "ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup"
2996 .Dl "ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup"
2998 .Dl "ipfw add deny tcp from any to any"
3000 The first rule will be a quick match for normal TCP packets,
3001 but it will not match the initial SYN packet, which will be
3004 rules only for selected source/destination pairs.
3005 All other SYN packets will be rejected by the final
3009 If you administer one or more subnets, you can take advantage
3010 of the address sets and or-blocks and write extremely
3011 compact rulesets which selectively enable services to blocks
3012 of clients, as below:
3014 .Dl "goodguys=\*q{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }\*q"
3015 .Dl "badguys=\*q10.1.2.0/24{8,38,60}\*q"
3017 .Dl "ipfw add allow ip from ${goodguys} to any"
3018 .Dl "ipfw add deny ip from ${badguys} to any"
3019 .Dl "... normal policies ..."
3023 option could be used to do automated anti-spoofing by adding the
3024 following to the top of a ruleset:
3026 .Dl "ipfw add deny ip from any to any not verrevpath in"
3028 This rule drops all incoming packets that appear to be coming to the
3029 system on the wrong interface.
3030 For example, a packet with a source
3031 address belonging to a host on a protected internal network would be
3032 dropped if it tried to enter the system from an external interface.
3036 option could be used to do similar but more restricted anti-spoofing
3037 by adding the following to the top of a ruleset:
3039 .Dl "ipfw add deny ip from any to any not antispoof in"
3041 This rule drops all incoming packets that appear to be coming from another
3042 directly connected system but on the wrong interface.
3043 For example, a packet with a source address of
3044 .Li 192.168.0.0/24 ,
3053 option could be used to (re)mark user traffic,
3054 by adding the following to the appropriate place in ruleset:
3056 .Dl "ipfw add setdscp be ip from any to any dscp af11,af21"
3058 In order to protect a site from flood attacks involving fake
3059 TCP packets, it is safer to use dynamic rules:
3061 .Dl "ipfw add check-state"
3062 .Dl "ipfw add deny tcp from any to any established"
3063 .Dl "ipfw add allow tcp from my-net to any setup keep-state"
3065 This will let the firewall install dynamic rules only for
3066 those connection which start with a regular SYN packet coming
3067 from the inside of our network.
3068 Dynamic rules are checked when encountering the first
3077 rule should usually be placed near the beginning of the
3078 ruleset to minimize the amount of work scanning the ruleset.
3079 Your mileage may vary.
3081 To limit the number of connections a user can open
3082 you can use the following type of rules:
3084 .Dl "ipfw add allow tcp from my-net/24 to any setup limit src-addr 10"
3085 .Dl "ipfw add allow tcp from any to me setup limit src-addr 4"
3087 The former (assuming it runs on a gateway) will allow each host
3088 on a /24 network to open at most 10 TCP connections.
3089 The latter can be placed on a server to make sure that a single
3090 client does not use more than 4 simultaneous connections.
3093 stateful rules can be subject to denial-of-service attacks
3094 by a SYN-flood which opens a huge number of dynamic rules.
3095 The effects of such attacks can be partially limited by
3098 variables which control the operation of the firewall.
3100 Here is a good usage of the
3102 command to see accounting records and timestamp information:
3106 or in short form without timestamps:
3110 which is equivalent to:
3114 Next rule diverts all incoming packets from 192.168.2.0/24
3115 to divert port 5000:
3117 .Dl ipfw divert 5000 ip from 192.168.2.0/24 to any in
3119 The following rules show some of the applications of
3123 for simulations and the like.
3125 This rule drops random incoming packets with a probability
3128 .Dl "ipfw add prob 0.05 deny ip from any to any in"
3130 A similar effect can be achieved making use of
3134 .Dl "ipfw add pipe 10 ip from any to any"
3135 .Dl "ipfw pipe 10 config plr 0.05"
3137 We can use pipes to artificially limit bandwidth, e.g.\& on a
3138 machine acting as a router, if we want to limit traffic from
3139 local clients on 192.168.2.0/24 we do:
3141 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
3142 .Dl "ipfw pipe 1 config bw 300Kbit/s queue 50KBytes"
3144 note that we use the
3146 modifier so that the rule is not used twice.
3147 Remember in fact that
3149 rules are checked both on incoming and outgoing packets.
3151 Should we want to simulate a bidirectional link with bandwidth
3152 limitations, the correct way is the following:
3154 .Dl "ipfw add pipe 1 ip from any to any out"
3155 .Dl "ipfw add pipe 2 ip from any to any in"
3156 .Dl "ipfw pipe 1 config bw 64Kbit/s queue 10Kbytes"
3157 .Dl "ipfw pipe 2 config bw 64Kbit/s queue 10Kbytes"
3159 The above can be very useful, e.g.\& if you want to see how
3160 your fancy Web page will look for a residential user who
3161 is connected only through a slow link.
3162 You should not use only one pipe for both directions, unless
3163 you want to simulate a half-duplex medium (e.g.\& AppleTalk,
3165 It is not necessary that both pipes have the same configuration,
3166 so we can also simulate asymmetric links.
3168 Should we want to verify network performance with the RED queue
3169 management algorithm:
3171 .Dl "ipfw add pipe 1 ip from any to any"
3172 .Dl "ipfw pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1"
3174 Another typical application of the traffic shaper is to
3175 introduce some delay in the communication.
3176 This can significantly affect applications which do a lot of Remote
3177 Procedure Calls, and where the round-trip-time of the
3178 connection often becomes a limiting factor much more than
3181 .Dl "ipfw add pipe 1 ip from any to any out"
3182 .Dl "ipfw add pipe 2 ip from any to any in"
3183 .Dl "ipfw pipe 1 config delay 250ms bw 1Mbit/s"
3184 .Dl "ipfw pipe 2 config delay 250ms bw 1Mbit/s"
3186 Per-flow queueing can be useful for a variety of purposes.
3187 A very simple one is counting traffic:
3189 .Dl "ipfw add pipe 1 tcp from any to any"
3190 .Dl "ipfw add pipe 1 udp from any to any"
3191 .Dl "ipfw add pipe 1 ip from any to any"
3192 .Dl "ipfw pipe 1 config mask all"
3194 The above set of rules will create queues (and collect
3195 statistics) for all traffic.
3196 Because the pipes have no limitations, the only effect is
3197 collecting statistics.
3198 Note that we need 3 rules, not just the last one, because
3201 tries to match IP packets it will not consider ports, so we
3202 would not see connections on separate ports as different
3205 A more sophisticated example is limiting the outbound traffic
3206 on a net with per-host limits, rather than per-network limits:
3208 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
3209 .Dl "ipfw add pipe 2 ip from any to 192.168.2.0/24 in"
3210 .Dl "ipfw pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
3211 .Dl "ipfw pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
3213 In the following example, we need to create several traffic bandwidth
3214 classes and we need different hosts/networks to fall into different classes.
3215 We create one pipe for each class and configure them accordingly.
3216 Then we create a single table and fill it with IP subnets and addresses.
3217 For each subnet/host we set the argument equal to the number of the pipe
3219 Then we classify traffic using a single rule:
3221 .Dl "ipfw pipe 1 config bw 1000Kbyte/s"
3222 .Dl "ipfw pipe 4 config bw 4000Kbyte/s"
3224 .Dl "ipfw table 1 add 192.168.2.0/24 1"
3225 .Dl "ipfw table 1 add 192.168.0.0/27 4"
3226 .Dl "ipfw table 1 add 192.168.0.2 1"
3228 .Dl "ipfw add pipe tablearg ip from table(1) to any"
3232 action, the table entries may include hostnames and IP addresses.
3234 .Dl "ipfw table 1 add 192.168.2.0/24 10.23.2.1"
3235 .Dl "ipfw table 1 add 192.168.0.0/27 router1.dmz"
3237 .Dl "ipfw add 100 fwd tablearg ip from any to table(1)"
3239 In the following example per-interface firewall is created:
3241 .Dl "ipfw table 10 add vlan20 12000"
3242 .Dl "ipfw table 10 add vlan30 13000"
3243 .Dl "ipfw table 20 add vlan20 22000"
3244 .Dl "ipfw table 20 add vlan30 23000"
3246 .Dl "ipfw add 100 ipfw skipto tablearg ip from any to any recv 'table(10)' in"
3247 .Dl "ipfw add 200 ipfw skipto tablearg ip from any to any xmit 'table(10)' out"
3249 To add a set of rules atomically, e.g.\& set 18:
3251 .Dl "ipfw set disable 18"
3252 .Dl "ipfw add NN set 18 ... # repeat as needed"
3253 .Dl "ipfw set enable 18"
3255 To delete a set of rules atomically the command is simply:
3257 .Dl "ipfw delete set 18"
3259 To test a ruleset and disable it and regain control if something goes wrong:
3261 .Dl "ipfw set disable 18"
3262 .Dl "ipfw add NN set 18 ... # repeat as needed"
3263 .Dl "ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18"
3265 Here if everything goes well, you press control-C before the "sleep"
3266 terminates, and your ruleset will be left active.
3267 Otherwise, e.g.\& if
3268 you cannot access your box, the ruleset will be disabled after
3269 the sleep terminates thus restoring the previous situation.
3271 To show rules of the specific set:
3273 .Dl "ipfw set 18 show"
3275 To show rules of the disabled set:
3277 .Dl "ipfw -S set 18 show"
3279 To clear a specific rule counters of the specific set:
3281 .Dl "ipfw set 18 zero NN"
3283 To delete a specific rule of the specific set:
3285 .Dl "ipfw set 18 delete NN"
3286 .Ss NAT, REDIRECT AND LSNAT
3287 First redirect all the traffic to nat instance 123:
3289 .Dl "ipfw add nat 123 all from any to any"
3291 Then to configure nat instance 123 to alias all the outgoing traffic with ip
3292 192.168.0.123, blocking all incoming connections, trying to keep
3293 same ports on both sides, clearing aliasing table on address change
3294 and keeping a log of traffic/link statistics:
3296 .Dl "ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports"
3298 Or to change address of instance 123, aliasing table will be cleared (see
3301 .Dl "ipfw nat 123 config ip 10.0.0.1"
3303 To see configuration of nat instance 123:
3305 .Dl "ipfw nat 123 show config"
3307 To show logs of all the instances in range 111-999:
3309 .Dl "ipfw nat 111-999 show"
3311 To see configurations of all instances:
3313 .Dl "ipfw nat show config"
3315 Or a redirect rule with mixed modes could looks like:
3317 .Dl "ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66"
3318 .Dl " redirect_port tcp 192.168.0.1:80 500"
3319 .Dl " redirect_proto udp 192.168.1.43 192.168.1.1"
3320 .Dl " redirect_addr 192.168.0.10,192.168.0.11"
3321 .Dl " 10.0.0.100 # LSNAT"
3322 .Dl " redirect_port tcp 192.168.0.1:80,192.168.0.10:22"
3325 or it could be split in:
3327 .Dl "ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66"
3328 .Dl "ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500"
3329 .Dl "ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1"
3330 .Dl "ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12"
3332 .Dl "ipfw nat 5 config redirect_port tcp"
3333 .Dl " 192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500"
3354 utility first appeared in
3359 Stateful extensions were introduced in
3362 was introduced in Summer 2002.
3364 .An Ugen J. S. Antsilevich ,
3365 .An Poul-Henning Kamp ,
3371 API based upon code written by
3375 Dummynet has been introduced by Luigi Rizzo in 1997-1998.
3377 Some early work (1999-2000) on the
3379 traffic shaper supported by Akamba Corp.
3381 The ipfw core (ipfw2) has been completely redesigned and
3382 reimplemented by Luigi Rizzo in summer 2002.
3385 options have been added by various developer over the years.
3388 In-kernel NAT support written by
3389 .An Paolo Pisati Aq piso@FreeBSD.org
3390 as part of a Summer of Code 2005 project.
3394 support has been developed by
3395 .An The Centre for Advanced Internet Architectures (CAIA) Aq http://www.caia.swin.edu.au .
3396 The primary developers and maintainers are David Hayes and Jason But.
3397 For further information visit:
3398 .Aq http://www.caia.swin.edu.au/urp/SONATA
3400 Delay profiles have been developed by Alessandro Cerri and
3401 Luigi Rizzo, supported by the
3402 European Commission within Projects Onelab and Onelab2.
3404 The syntax has grown over the years and sometimes it might be confusing.
3405 Unfortunately, backward compatibility prevents cleaning up mistakes
3406 made in the definition of the syntax.
3410 Misconfiguring the firewall can put your computer in an unusable state,
3411 possibly shutting down network services and requiring console access to
3412 regain control of it.
3414 Incoming packet fragments diverted by
3416 are reassembled before delivery to the socket.
3417 The action used on those packet is the one from the
3418 rule which matches the first fragment of the packet.
3420 Packets diverted to userland, and then reinserted by a userland process
3421 may lose various packet attributes.
3422 The packet source interface name
3423 will be preserved if it is shorter than 8 bytes and the userland process
3424 saves and reuses the sockaddr_in
3427 otherwise, it may be lost.
3428 If a packet is reinserted in this manner, later rules may be incorrectly
3429 applied, making the order of
3431 rules in the rule sequence very important.
3433 Dummynet drops all packets with IPv6 link-local addresses.
3439 may not behave as expected.
3440 In particular, incoming SYN packets may
3441 have no uid or gid associated with them since they do not yet belong
3442 to a TCP connection, and the uid/gid associated with a packet may not
3443 be as expected if the associated process calls
3445 or similar system calls.
3447 Rule syntax is subject to the command line environment and some patterns
3448 may need to be escaped with the backslash character
3449 or quoted appropriately.
3451 Due to the architecture of
3453 ipfw nat is not compatible with the TCP segmentation offloading (TSO).
3454 Thus, to reliably nat your network traffic, please disable TSO
3458 ICMP error messages are not implicitly matched by dynamic rules
3459 for the respective conversations.
3460 To avoid failures of network error detection and path MTU discovery,
3461 ICMP error messages may need to be allowed explicitly through static
3468 actions may lead to confusing behaviour if ruleset has mistakes,
3469 and/or interaction with other subsystems (netgraph, dummynet, etc.) is used.
3470 One possible case for this is packet leaving
3472 in subroutine on the input pass, while later on output encountering unpaired
3475 As the call stack is kept intact after input pass, packet will suddenly
3476 return to the rule number used on input pass, not on output one.
3477 Order of processing should be checked carefully to avoid such mistakes.