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 .Oo Cm set Ar N Oc Cm table Ar name Cm create Ar create-options
53 .Oo Cm set Ar N Oc Cm table Ar name Cm destroy
55 .Oo Cm set Ar N Oc Cm table Ar name Cm modify Ar modify-options
57 .Oo Cm set Ar N Oc Cm table Ar name Cm swap Ar name
59 .Oo Cm set Ar N Oc Cm table Ar name Cm add Ar table-key Op Ar value
61 .Oo Cm set Ar N Oc Cm table Ar name Cm add Op Ar table-key Ar value ...
63 .Oo Cm set Ar N Oc Cm table Ar name Cm atomic add Op Ar table-key Ar value ...
65 .Oo Cm set Ar N Oc Cm table Ar name Cm delete Op Ar table-key ...
67 .Oo Cm set Ar N Oc Cm table Ar name Cm lookup Ar addr
69 .Oo Cm set Ar N Oc Cm table Ar name Cm lock
71 .Oo Cm set Ar N Oc Cm table Ar name Cm unlock
73 .Oo Cm set Ar N Oc Cm table
77 .Oo Cm set Ar N Oc Cm table
81 .Oo Cm set Ar N Oc Cm table
85 .Oo Cm set Ar N Oc Cm table
88 .Ss DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)
90 .Brq Cm pipe | queue | sched
96 .Brq Cm pipe | queue | sched
97 .Brq Cm delete | list | show
116 .Ss IPv6-to-IPv6 NETWORK PREFIX TRANSLATION
118 .Oo Cm set Ar N Oc Cm nptv6 Ar name Cm create Ar create-options
120 .Oo Cm set Ar N Oc Cm nptv6
124 .Oo Cm set Ar N Oc Cm nptv6
128 .Oo Cm set Ar N Oc Cm nptv6 Ar name Cm stats Op Cm reset
129 .Ss INTERNAL DIAGNOSTICS
139 utility is the user interface for controlling the
143 traffic shaper/packet scheduler, and the
144 in-kernel NAT services.
146 A firewall configuration, or
150 numbered from 1 to 65535.
151 Packets are passed to the firewall
152 from a number of different places in the protocol stack
153 (depending on the source and destination of the packet,
154 it is possible for the firewall to be
155 invoked multiple times on the same packet).
156 The packet passed to the firewall is compared
157 against each of the rules in the
160 (multiple rules with the same number are permitted, in which case
161 they are processed in order of insertion).
162 When a match is found, the action corresponding to the
163 matching rule is performed.
165 Depending on the action and certain system settings, packets
166 can be reinjected into the firewall at some rule after the
167 matching one for further processing.
169 A ruleset always includes a
171 rule (numbered 65535) which cannot be modified or deleted,
172 and matches all packets.
173 The action associated with the
179 depending on how the kernel is configured.
181 If the ruleset includes one or more rules with the
186 the firewall will have a
188 behaviour, i.e., upon a match it will create
190 i.e., rules that match packets with the same 5-tuple
191 (protocol, source and destination addresses and ports)
192 as the packet which caused their creation.
193 Dynamic rules, which have a limited lifetime, are checked
194 at the first occurrence of a
199 rule, and are typically used to open the firewall on-demand to
200 legitimate traffic only.
202 .Sx STATEFUL FIREWALL
205 Sections below for more information on the stateful behaviour of
208 All rules (including dynamic ones) have a few associated counters:
209 a packet count, a byte count, a log count and a timestamp
210 indicating the time of the last match.
211 Counters can be displayed or reset with
215 Each rule belongs to one of 32 different
219 commands to atomically manipulate sets, such as enable,
220 disable, swap sets, move all rules in a set to another
221 one, delete all rules in a set.
222 These can be useful to
223 install temporary configurations, or to test them.
226 for more information on
229 Rules can be added with the
231 command; deleted individually or in groups with the
233 command, and globally (except those in set 31) with the
235 command; displayed, optionally with the content of the
241 Finally, counters can be reset with the
248 The following general options are available when invoking
250 .Bl -tag -width indent
252 Show counter values when listing rules.
255 command implies this option.
257 Only show the action and the comment, not the body of a rule.
261 When entering or showing rules, print them in compact form,
262 i.e., omitting the "ip from any to any" string
263 when this does not carry any additional information.
265 When listing, show dynamic rules in addition to static ones.
269 is specified, also show expired dynamic rules.
271 Do not ask for confirmation for commands that can cause problems
274 If there is no tty associated with the process, this is implied.
276 When listing a table (see the
278 section below for more information on lookup tables), format values
280 By default, values are shown as integers.
282 Only check syntax of the command strings, without actually passing
285 Try to resolve addresses and service names in output.
287 Be quiet when executing the
297 This is useful when updating rulesets by executing multiple
301 .Ql sh\ /etc/rc.firewall ) ,
302 or by processing a file with many
304 rules across a remote login session.
305 It also stops a table add or delete
306 from failing if the entry already exists or is not present.
308 The reason why this option may be important is that
309 for some of these actions,
311 may print a message; if the action results in blocking the
312 traffic to the remote client,
313 the remote login session will be closed
314 and the rest of the ruleset will not be processed.
315 Access to the console would then be required to recover.
317 When listing rules, show the
319 each rule belongs to.
320 If this flag is not specified, disabled rules will not be
323 When listing pipes, sort according to one of the four
324 counters (total or current packets or bytes).
326 When listing, show last match timestamp converted with ctime().
328 When listing, show last match timestamp as seconds from the epoch.
329 This form can be more convenient for postprocessing by scripts.
331 .Ss LIST OF RULES AND PREPROCESSING
332 To ease configuration, rules can be put into a file which is
335 as shown in the last synopsis line.
339 The file will be read line by line and applied as arguments to the
343 Optionally, a preprocessor can be specified using
347 is to be piped through.
348 Useful preprocessors include
354 does not start with a slash
356 as its first character, the usual
358 name search is performed.
359 Care should be taken with this in environments where not all
360 file systems are mounted (yet) by the time
362 is being run (e.g.\& when they are mounted over NFS).
365 has been specified, any additional arguments are passed on to the preprocessor
367 This allows for flexible configuration files (like conditionalizing
368 them on the local hostname) and the use of macros to centralize
369 frequently required arguments like IP addresses.
370 .Ss TRAFFIC SHAPER CONFIGURATION
376 commands are used to configure the traffic shaper and packet scheduler.
378 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
379 Section below for details.
381 If the world and the kernel get out of sync the
383 ABI may break, preventing you from being able to add any rules.
384 This can adversely affect the booting process.
389 to temporarily disable the firewall to regain access to the network,
390 allowing you to fix the problem.
392 A packet is checked against the active ruleset in multiple places
393 in the protocol stack, under control of several sysctl variables.
394 These places and variables are shown below, and it is important to
395 have this picture in mind in order to design a correct ruleset.
396 .Bd -literal -offset indent
399 +----------->-----------+
401 [ip(6)_input] [ip(6)_output] net.inet(6).ip(6).fw.enable=1
404 [ether_demux] [ether_output_frame] net.link.ether.ipfw=1
406 +-->--[bdg_forward]-->--+ net.link.bridge.ipfw=1
412 times the same packet goes through the firewall can
413 vary between 0 and 4 depending on packet source and
414 destination, and system configuration.
416 Note that as packets flow through the stack, headers can be
417 stripped or added to it, and so they may or may not be available
419 E.g., incoming packets will include the MAC header when
423 but the same packets will have the MAC header stripped off when
430 Also note that each packet is always checked against the complete ruleset,
431 irrespective of the place where the check occurs, or the source of the packet.
432 If a rule contains some match patterns or actions which are not valid
433 for the place of invocation (e.g.\& trying to match a MAC header within
437 the match pattern will not match, but a
439 operator in front of such patterns
443 match on those packets.
444 It is thus the responsibility of
445 the programmer, if necessary, to write a suitable ruleset to
446 differentiate among the possible places.
448 rules can be useful here, as an example:
449 .Bd -literal -offset indent
450 # packets from ether_demux or bdg_forward
451 ipfw add 10 skipto 1000 all from any to any layer2 in
452 # packets from ip_input
453 ipfw add 10 skipto 2000 all from any to any not layer2 in
454 # packets from ip_output
455 ipfw add 10 skipto 3000 all from any to any not layer2 out
456 # packets from ether_output_frame
457 ipfw add 10 skipto 4000 all from any to any layer2 out
460 (yes, at the moment there is no way to differentiate between
461 ether_demux and bdg_forward).
463 In general, each keyword or argument must be provided as
464 a separate command line argument, with no leading or trailing
466 Keywords are case-sensitive, whereas arguments may
467 or may not be case-sensitive depending on their nature
468 (e.g.\& uid's are, hostnames are not).
470 Some arguments (e.g., port or address lists) are comma-separated
472 In this case, spaces after commas ',' are allowed to make
473 the line more readable.
474 You can also put the entire
475 command (including flags) into a single argument.
476 E.g., the following forms are equivalent:
477 .Bd -literal -offset indent
478 ipfw -q add deny src-ip 10.0.0.0/24,127.0.0.1/8
479 ipfw -q add deny src-ip 10.0.0.0/24, 127.0.0.1/8
480 ipfw "-q add deny src-ip 10.0.0.0/24, 127.0.0.1/8"
483 The format of firewall rules is the following:
484 .Bd -ragged -offset indent
487 .Op Cm set Ar set_number
488 .Op Cm prob Ar match_probability
490 .Op Cm log Op Cm logamount Ar number
500 where the body of the rule specifies which information is used
501 for filtering packets, among the following:
503 .Bl -tag -width "Source and dest. addresses and ports" -offset XXX -compact
504 .It Layer-2 header fields
506 .It IPv4 and IPv6 Protocol
508 .It Source and dest. addresses and ports
512 .It Transmit and receive interface
514 .It Misc. IP header fields
515 Version, type of service, datagram length, identification,
516 fragment flag (non-zero IP offset),
519 .It IPv6 Extension headers
520 Fragmentation, Hop-by-Hop options,
521 Routing Headers, Source routing rthdr0, Mobile IPv6 rthdr2, IPSec options.
523 .It Misc. TCP header fields
524 TCP flags (SYN, FIN, ACK, RST, etc.),
525 sequence number, acknowledgment number,
533 When the packet can be associated with a local socket.
535 Whether a packet came from a divert socket (e.g.,
537 .It Fib annotation state
538 Whether a packet has been tagged for using a specific FIB (routing table)
539 in future forwarding decisions.
542 Note that some of the above information, e.g.\& source MAC or IP addresses and
543 TCP/UDP ports, can be easily spoofed, so filtering on those fields
544 alone might not guarantee the desired results.
545 .Bl -tag -width indent
547 Each rule is associated with a
549 in the range 1..65535, with the latter reserved for the
552 Rules are checked sequentially by rule number.
553 Multiple rules can have the same number, in which case they are
554 checked (and listed) according to the order in which they have
556 If a rule is entered without specifying a number, the kernel will
557 assign one in such a way that the rule becomes the last one
561 Automatic rule numbers are assigned by incrementing the last
562 non-default rule number by the value of the sysctl variable
563 .Ar net.inet.ip.fw.autoinc_step
564 which defaults to 100.
565 If this is not possible (e.g.\& because we would go beyond the
566 maximum allowed rule number), the number of the last
567 non-default value is used instead.
568 .It Cm set Ar set_number
569 Each rule is associated with a
572 Sets can be individually disabled and enabled, so this parameter
573 is of fundamental importance for atomic ruleset manipulation.
574 It can be also used to simplify deletion of groups of rules.
575 If a rule is entered without specifying a set number,
578 Set 31 is special in that it cannot be disabled,
579 and rules in set 31 are not deleted by the
581 command (but you can delete them with the
582 .Nm ipfw delete set 31
584 Set 31 is also used for the
587 .It Cm prob Ar match_probability
588 A match is only declared with the specified probability
589 (floating point number between 0 and 1).
590 This can be useful for a number of applications such as
591 random packet drop or
594 to simulate the effect of multiple paths leading to out-of-order
597 Note: this condition is checked before any other condition, including
598 ones such as keep-state or check-state which might have side effects.
599 .It Cm log Op Cm logamount Ar number
600 Packets matching a rule with the
602 keyword will be made available for logging in two ways:
603 if the sysctl variable
604 .Va net.inet.ip.fw.verbose
605 is set to 0 (default), one can use
610 This pseudo interface can be created after a boot
611 manually by using the following command:
612 .Bd -literal -offset indent
613 # ifconfig ipfw0 create
616 Or, automatically at boot time by adding the following
620 .Bd -literal -offset indent
624 There is no overhead if no
626 is attached to the pseudo interface.
629 .Va net.inet.ip.fw.verbose
630 is set to 1, packets will be logged to
634 facility up to a maximum of
639 is specified, the limit is taken from the sysctl variable
640 .Va net.inet.ip.fw.verbose_limit .
641 In both cases, a value of 0 means unlimited logging.
643 Once the limit is reached, logging can be re-enabled by
644 clearing the logging counter or the packet counter for that entry, see the
648 Note: logging is done after all other packet matching conditions
649 have been successfully verified, and before performing the final
650 action (accept, deny, etc.) on the packet.
652 When a packet matches a rule with the
654 keyword, the numeric tag for the given
656 in the range 1..65534 will be attached to the packet.
657 The tag acts as an internal marker (it is not sent out over
658 the wire) that can be used to identify these packets later on.
659 This can be used, for example, to provide trust between interfaces
660 and to start doing policy-based filtering.
661 A packet can have multiple tags at the same time.
662 Tags are "sticky", meaning once a tag is applied to a packet by a
663 matching rule it exists until explicit removal.
664 Tags are kept with the packet everywhere within the kernel, but are
665 lost when packet leaves the kernel, for example, on transmitting
666 packet out to the network or sending packet to a
670 To check for previously applied tags, use the
673 To delete previously applied tag, use the
677 Note: since tags are kept with the packet everywhere in kernelspace,
678 they can be set and unset anywhere in the kernel network subsystem
681 facility), not only by means of the
687 For example, there can be a specialized
689 node doing traffic analyzing and tagging for later inspecting
691 .It Cm untag Ar number
692 When a packet matches a rule with the
694 keyword, the tag with the number
696 is searched among the tags attached to this packet and,
697 if found, removed from it.
698 Other tags bound to packet, if present, are left untouched.
700 When a packet matches a rule with the
702 keyword, the ALTQ identifier for the given
707 Note that this ALTQ tag is only meaningful for packets going "out" of IPFW,
708 and not being rejected or going to divert sockets.
709 Note that if there is insufficient memory at the time the packet is
710 processed, it will not be tagged, so it is wise to make your ALTQ
711 "default" queue policy account for this.
714 rules match a single packet, only the first one adds the ALTQ classification
716 In doing so, traffic may be shaped by using
717 .Cm count Cm altq Ar queue
718 rules for classification early in the ruleset, then later applying
719 the filtering decision.
724 rules may come later and provide the actual filtering decisions in
725 addition to the fallback ALTQ tag.
729 to set up the queues before IPFW will be able to look them up by name,
730 and if the ALTQ disciplines are rearranged, the rules in containing the
731 queue identifiers in the kernel will likely have gone stale and need
733 Stale queue identifiers will probably result in misclassification.
735 All system ALTQ processing can be turned on or off via
740 .Cm disable Ar altq .
742 .Va net.inet.ip.fw.one_pass
743 is irrelevant to ALTQ traffic shaping, as the actual rule action is followed
744 always after adding an ALTQ tag.
747 A rule can be associated with one of the following actions, which
748 will be executed when the packet matches the body of the rule.
749 .Bl -tag -width indent
750 .It Cm allow | accept | pass | permit
751 Allow packets that match rule.
752 The search terminates.
753 .It Cm check-state Op Ar :flowname | Cm :any
754 Checks the packet against the dynamic ruleset.
755 If a match is found, execute the action associated with
756 the rule which generated this dynamic rule, otherwise
757 move to the next rule.
760 rules do not have a body.
763 rule is found, the dynamic ruleset is checked at the first
770 is symbolic name assigned to dynamic rule by
775 can be used to ignore states flowname when matching.
778 keyword is special name used for compatibility with old rulesets.
780 Update counters for all packets that match rule.
781 The search continues with the next rule.
783 Discard packets that match this rule.
784 The search terminates.
785 .It Cm divert Ar port
786 Divert packets that match this rule to the
790 The search terminates.
791 .It Cm fwd | forward Ar ipaddr | tablearg Ns Op , Ns Ar port
792 Change the next-hop on matching packets to
794 which can be an IP address or a host name.
795 For IPv4, the next hop can also be supplied by the last table
796 looked up for the packet by using the
798 keyword instead of an explicit address.
799 The search terminates if this rule matches.
803 is a local address, then matching packets will be forwarded to
805 (or the port number in the packet if one is not specified in the rule)
806 on the local machine.
810 is not a local address, then the port number
811 (if specified) is ignored, and the packet will be
812 forwarded to the remote address, using the route as found in
813 the local routing table for that IP.
817 rule will not match layer-2 packets (those received
818 on ether_input, ether_output, or bridged).
822 action does not change the contents of the packet at all.
823 In particular, the destination address remains unmodified, so
824 packets forwarded to another system will usually be rejected by that system
825 unless there is a matching rule on that system to capture them.
826 For packets forwarded locally,
827 the local address of the socket will be
828 set to the original destination address of the packet.
831 entry look rather weird but is intended for
832 use with transparent proxy servers.
833 .It Cm nat Ar nat_nr | tablearg
836 (for network address translation, address redirect, etc.):
838 .Sx NETWORK ADDRESS TRANSLATION (NAT)
839 Section for further information.
841 Pass packet to a NPTv6 instance (for IPv6-to-IPv6 network prefix translation):
843 .Sx IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
844 Section for further information.
845 .It Cm pipe Ar pipe_nr
849 (for bandwidth limitation, delay, etc.).
851 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
852 Section for further information.
853 The search terminates; however, on exit from the pipe and if
857 .Va net.inet.ip.fw.one_pass
858 is not set, the packet is passed again to the firewall code
859 starting from the next rule.
860 .It Cm queue Ar queue_nr
864 (for bandwidth limitation using WF2Q+).
870 Discard packets that match this rule, and if the
871 packet is a TCP packet, try to send a TCP reset (RST) notice.
872 The search terminates.
874 Discard packets that match this rule, and if the
875 packet is a TCP packet, try to send a TCP reset (RST) notice.
876 The search terminates.
877 .It Cm skipto Ar number | tablearg
878 Skip all subsequent rules numbered less than
880 The search continues with the first rule numbered
883 It is possible to use the
885 keyword with a skipto for a
887 skipto. Skipto may work either in O(log(N)) or in O(1) depending
888 on amount of memory and/or sysctl variables.
891 section for more details.
892 .It Cm call Ar number | tablearg
893 The current rule number is saved in the internal stack and
894 ruleset processing continues with the first rule numbered
897 If later a rule with the
899 action is encountered, the processing returns to the first rule
902 rule plus one or higher
903 (the same behaviour as with packets returning from
908 This could be used to make somewhat like an assembly language
910 calls to rules with common checks for different interfaces, etc.
912 Rule with any number could be called, not just forward jumps as with
914 So, to prevent endless loops in case of mistakes, both
918 actions don't do any jumps and simply go to the next rule if memory
919 cannot be allocated or stack overflowed/underflowed.
921 Internally stack for rule numbers is implemented using
923 facility and currently has size of 16 entries.
924 As mbuf tags are lost when packet leaves the kernel,
926 should not be used in subroutines to avoid endless loops
927 and other undesired effects.
929 Takes rule number saved to internal stack by the last
931 action and returns ruleset processing to the first rule
932 with number greater than number of corresponding
935 See description of the
937 action for more details.
943 and thus are unconditional, but
945 command-line utility currently requires every action except
948 While it is sometimes useful to return only on some packets,
949 usually you want to print just
952 A workaround for this is to use new syntax and
955 .Bd -literal -offset indent
956 # Add a rule without actual body
957 ipfw add 2999 return via any
959 # List rules without "from any to any" part
963 This cosmetic annoyance may be fixed in future releases.
965 Send a copy of packets matching this rule to the
969 The search continues with the next rule.
970 .It Cm unreach Ar code
971 Discard packets that match this rule, and try to send an ICMP
972 unreachable notice with code
976 is a number from 0 to 255, or one of these aliases:
977 .Cm net , host , protocol , port ,
978 .Cm needfrag , srcfail , net-unknown , host-unknown ,
979 .Cm isolated , net-prohib , host-prohib , tosnet ,
980 .Cm toshost , filter-prohib , host-precedence
982 .Cm precedence-cutoff .
983 The search terminates.
984 .It Cm unreach6 Ar code
985 Discard packets that match this rule, and try to send an ICMPv6
986 unreachable notice with code
990 is a number from 0, 1, 3 or 4, or one of these aliases:
991 .Cm no-route, admin-prohib, address
994 The search terminates.
995 .It Cm netgraph Ar cookie
996 Divert packet into netgraph with given
998 The search terminates.
999 If packet is later returned from netgraph it is either
1000 accepted or continues with the next rule, depending on
1001 .Va net.inet.ip.fw.one_pass
1003 .It Cm ngtee Ar cookie
1004 A copy of packet is diverted into netgraph, original
1005 packet continues with the next rule.
1008 for more information on
1013 .It Cm setfib Ar fibnum | tablearg
1014 The packet is tagged so as to use the FIB (routing table)
1016 in any subsequent forwarding decisions.
1017 In the current implementation, this is limited to the values 0 through 15, see
1019 Processing continues at the next rule.
1020 It is possible to use the
1022 keyword with setfib.
1023 If the tablearg value is not within the compiled range of fibs,
1024 the packet's fib is set to 0.
1025 .It Cm setdscp Ar DSCP | number | tablearg
1026 Set specified DiffServ codepoint for an IPv4/IPv6 packet.
1027 Processing continues at the next rule.
1028 Supported values are:
1074 Additionally, DSCP value can be specified by number (0..64).
1075 It is also possible to use the
1077 keyword with setdscp.
1078 If the tablearg value is not within the 0..64 range, lower 6 bits of supplied
1081 Queue and reassemble IP fragments.
1082 If the packet is not fragmented, counters are updated and
1083 processing continues with the next rule.
1084 If the packet is the last logical fragment, the packet is reassembled and, if
1085 .Va net.inet.ip.fw.one_pass
1086 is set to 0, processing continues with the next rule.
1087 Otherwise, the packet is allowed to pass and the search terminates.
1088 If the packet is a fragment in the middle of a logical group of fragments,
1090 processing stops immediately.
1092 Fragment handling can be tuned via
1093 .Va net.inet.ip.maxfragpackets
1095 .Va net.inet.ip.maxfragsperpacket
1096 which limit, respectively, the maximum number of processable
1097 fragments (default: 800) and
1098 the maximum number of fragments per packet (default: 16).
1100 NOTA BENE: since fragments do not contain port numbers,
1101 they should be avoided with the
1104 Alternatively, direction-based (like
1108 ) and source-based (like
1110 ) match patterns can be used to select fragments.
1112 Usually a simple rule like:
1113 .Bd -literal -offset indent
1114 # reassemble incoming fragments
1115 ipfw add reass all from any to any in
1118 is all you need at the beginning of your ruleset.
1121 The body of a rule contains zero or more patterns (such as
1122 specific source and destination addresses or ports,
1123 protocol options, incoming or outgoing interfaces, etc.)
1124 that the packet must match in order to be recognised.
1125 In general, the patterns are connected by (implicit)
1127 operators -- i.e., all must match in order for the
1129 Individual patterns can be prefixed by the
1131 operator to reverse the result of the match, as in
1133 .Dl "ipfw add 100 allow ip from not 1.2.3.4 to any"
1135 Additionally, sets of alternative match patterns
1137 can be constructed by putting the patterns in
1138 lists enclosed between parentheses ( ) or braces { }, and
1141 operator as follows:
1143 .Dl "ipfw add 100 allow ip from { x or not y or z } to any"
1145 Only one level of parentheses is allowed.
1146 Beware that most shells have special meanings for parentheses
1147 or braces, so it is advisable to put a backslash \\ in front of them
1148 to prevent such interpretations.
1150 The body of a rule must in general include a source and destination
1154 can be used in various places to specify that the content of
1155 a required field is irrelevant.
1157 The rule body has the following format:
1158 .Bd -ragged -offset indent
1159 .Op Ar proto Cm from Ar src Cm to Ar dst
1163 The first part (proto from src to dst) is for backward
1164 compatibility with earlier versions of
1168 any match pattern (including MAC headers, IP protocols,
1169 addresses and ports) can be specified in the
1173 Rule fields have the following meaning:
1174 .Bl -tag -width indent
1175 .It Ar proto : protocol | Cm { Ar protocol Cm or ... }
1176 .It Ar protocol : Oo Cm not Oc Ar protocol-name | protocol-number
1177 An IP protocol specified by number or name
1178 (for a complete list see
1179 .Pa /etc/protocols ) ,
1180 or one of the following keywords:
1181 .Bl -tag -width indent
1183 Matches IPv4 packets.
1185 Matches IPv6 packets.
1194 option will be treated as inner protocol.
1202 .Cm { Ar protocol Cm or ... }
1205 is provided for convenience only but its use is deprecated.
1206 .It Ar src No and Ar dst : Bro Cm addr | Cm { Ar addr Cm or ... } Brc Op Oo Cm not Oc Ar ports
1207 An address (or a list, see below)
1208 optionally followed by
1214 with multiple addresses) is provided for convenience only and
1215 its use is discouraged.
1216 .It Ar addr : Oo Cm not Oc Bro
1217 .Cm any | me | me6 |
1218 .Cm table Ns Pq Ar name Ns Op , Ns Ar value
1219 .Ar | addr-list | addr-set
1221 .Bl -tag -width indent
1223 matches any IP address.
1225 matches any IP address configured on an interface in the system.
1227 matches any IPv6 address configured on an interface in the system.
1228 The address list is evaluated at the time the packet is
1230 .It Cm table Ns Pq Ar name Ns Op , Ns Ar value
1231 Matches any IPv4 or IPv6 address for which an entry exists in the lookup table
1233 If an optional 32-bit unsigned
1235 is also specified, an entry will match only if it has this value.
1238 section below for more information on lookup tables.
1240 .It Ar addr-list : ip-addr Ns Op Ns , Ns Ar addr-list
1242 A host or subnet address specified in one of the following ways:
1243 .Bl -tag -width indent
1244 .It Ar numeric-ip | hostname
1245 Matches a single IPv4 address, specified as dotted-quad or a hostname.
1246 Hostnames are resolved at the time the rule is added to the firewall list.
1247 .It Ar addr Ns / Ns Ar masklen
1248 Matches all addresses with base
1250 (specified as an IP address, a network number, or a hostname)
1254 As an example, 1.2.3.4/25 or 1.2.3.0/25 will match
1255 all IP numbers from 1.2.3.0 to 1.2.3.127 .
1256 .It Ar addr Ns : Ns Ar mask
1257 Matches all addresses with base
1259 (specified as an IP address, a network number, or a hostname)
1262 specified as a dotted quad.
1263 As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match
1265 This form is advised only for non-contiguous
1267 It is better to resort to the
1268 .Ar addr Ns / Ns Ar masklen
1269 format for contiguous masks, which is more compact and less
1272 .It Ar addr-set : addr Ns Oo Ns / Ns Ar masklen Oc Ns Cm { Ns Ar list Ns Cm }
1273 .It Ar list : Bro Ar num | num-num Brc Ns Op Ns , Ns Ar list
1274 Matches all addresses with base address
1276 (specified as an IP address, a network number, or a hostname)
1277 and whose last byte is in the list between braces { } .
1278 Note that there must be no spaces between braces and
1279 numbers (spaces after commas are allowed).
1280 Elements of the list can be specified as single entries
1284 field is used to limit the size of the set of addresses,
1285 and can have any value between 24 and 32.
1287 it will be assumed as 24.
1289 This format is particularly useful to handle sparse address sets
1290 within a single rule.
1291 Because the matching occurs using a
1292 bitmask, it takes constant time and dramatically reduces
1293 the complexity of rulesets.
1295 As an example, an address specified as 1.2.3.4/24{128,35-55,89}
1296 or 1.2.3.0/24{128,35-55,89}
1297 will match the following IP addresses:
1299 1.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .
1300 .It Ar addr6-list : ip6-addr Ns Op Ns , Ns Ar addr6-list
1302 A host or subnet specified one of the following ways:
1303 .Bl -tag -width indent
1304 .It Ar numeric-ip | hostname
1305 Matches a single IPv6 address as allowed by
1308 Hostnames are resolved at the time the rule is added to the firewall
1310 .It Ar addr Ns / Ns Ar masklen
1311 Matches all IPv6 addresses with base
1313 (specified as allowed by
1319 .It Ar addr Ns / Ns Ar mask
1320 Matches all IPv6 addresses with base
1322 (specified as allowed by
1327 specified as allowed by
1329 As an example, fe::640:0:0/ffff::ffff:ffff:0:0 will match
1331 This form is advised only for non-contiguous
1333 It is better to resort to the
1334 .Ar addr Ns / Ns Ar masklen
1335 format for contiguous masks, which is more compact and less
1339 No support for sets of IPv6 addresses is provided because IPv6 addresses
1340 are typically random past the initial prefix.
1341 .It Ar ports : Bro Ar port | port Ns \&- Ns Ar port Ns Brc Ns Op , Ns Ar ports
1342 For protocols which support port numbers (such as TCP and UDP), optional
1344 may be specified as one or more ports or port ranges, separated
1345 by commas but no spaces, and an optional
1350 notation specifies a range of ports (including boundaries).
1354 may be used instead of numeric port values.
1355 The length of the port list is limited to 30 ports or ranges,
1356 though one can specify larger ranges by using an
1360 section of the rule.
1364 can be used to escape the dash
1366 character in a service name (from a shell, the backslash must be
1367 typed twice to avoid the shell itself interpreting it as an escape
1370 .Dl "ipfw add count tcp from any ftp\e\e-data-ftp to any"
1372 Fragmented packets which have a non-zero offset (i.e., not the first
1373 fragment) will never match a rule which has one or more port
1377 option for details on matching fragmented packets.
1379 .Ss RULE OPTIONS (MATCH PATTERNS)
1380 Additional match patterns can be used within
1382 Zero or more of these so-called
1384 can be present in a rule, optionally prefixed by the
1386 operand, and possibly grouped into
1389 The following match patterns can be used (listed in alphabetical order):
1390 .Bl -tag -width indent
1391 .It Cm // this is a comment.
1392 Inserts the specified text as a comment in the rule.
1393 Everything following // is considered as a comment and stored in the rule.
1394 You can have comment-only rules, which are listed as having a
1396 action followed by the comment.
1401 Matches only packets generated by a divert socket.
1402 .It Cm diverted-loopback
1403 Matches only packets coming from a divert socket back into the IP stack
1405 .It Cm diverted-output
1406 Matches only packets going from a divert socket back outward to the IP
1407 stack output for delivery.
1408 .It Cm dst-ip Ar ip-address
1409 Matches IPv4 packets whose destination IP is one of the address(es)
1410 specified as argument.
1411 .It Bro Cm dst-ip6 | dst-ipv6 Brc Ar ip6-address
1412 Matches IPv6 packets whose destination IP is one of the address(es)
1413 specified as argument.
1414 .It Cm dst-port Ar ports
1415 Matches IP packets whose destination port is one of the port(s)
1416 specified as argument.
1418 Matches TCP packets that have the RST or ACK bits set.
1419 .It Cm ext6hdr Ar header
1420 Matches IPv6 packets containing the extended header given by
1422 Supported headers are:
1428 any type of Routing Header
1430 Source routing Routing Header Type 0
1432 Mobile IPv6 Routing Header Type 2
1436 IPSec authentication headers
1438 and IPsec encapsulated security payload headers
1440 .It Cm fib Ar fibnum
1441 Matches a packet that has been tagged to use
1442 the given FIB (routing table) number.
1443 .It Cm flow Ar table Ns Pq Ar name Ns Op , Ns Ar value
1444 Search for the flow entry in lookup table
1446 If not found, the match fails.
1447 Otherwise, the match succeeds and
1449 is set to the value extracted from the table.
1451 This option can be useful to quickly dispatch traffic based on
1452 certain packet fields.
1455 section below for more information on lookup tables.
1456 .It Cm flow-id Ar labels
1457 Matches IPv6 packets containing any of the flow labels given in
1460 is a comma separated list of numeric flow labels.
1462 Matches packets that are fragments and not the first
1463 fragment of an IP datagram.
1464 Note that these packets will not have
1465 the next protocol header (e.g.\& TCP, UDP) so options that look into
1466 these headers cannot match.
1468 Matches all TCP or UDP packets sent by or received for a
1472 may be specified by name or number.
1473 .It Cm jail Ar prisonID
1474 Matches all TCP or UDP packets sent by or received for the
1475 jail whos prison ID is
1477 .It Cm icmptypes Ar types
1478 Matches ICMP packets whose ICMP type is in the list
1480 The list may be specified as any combination of
1481 individual types (numeric) separated by commas.
1482 .Em Ranges are not allowed .
1483 The supported ICMP types are:
1487 destination unreachable
1495 router advertisement
1499 time-to-live exceeded
1511 address mask request
1513 and address mask reply
1515 .It Cm icmp6types Ar types
1516 Matches ICMP6 packets whose ICMP6 type is in the list of
1518 The list may be specified as any combination of
1519 individual types (numeric) separated by commas.
1520 .Em Ranges are not allowed .
1522 Matches incoming or outgoing packets, respectively.
1526 are mutually exclusive (in fact,
1530 .It Cm ipid Ar id-list
1531 Matches IPv4 packets whose
1533 field has value included in
1535 which is either a single value or a list of values or ranges
1536 specified in the same way as
1538 .It Cm iplen Ar len-list
1539 Matches IP packets whose total length, including header and data, is
1542 which is either a single value or a list of values or ranges
1543 specified in the same way as
1545 .It Cm ipoptions Ar spec
1546 Matches packets whose IPv4 header contains the comma separated list of
1547 options specified in
1549 The supported IP options are:
1552 (strict source route),
1554 (loose source route),
1556 (record packet route) and
1559 The absence of a particular option may be denoted
1562 .It Cm ipprecedence Ar precedence
1563 Matches IPv4 packets whose precedence field is equal to
1566 Matches packets that have IPSEC history associated with them
1567 (i.e., the packet comes encapsulated in IPSEC, the kernel
1568 has IPSEC support, and can correctly decapsulate it).
1570 Note that specifying
1572 is different from specifying
1574 as the latter will only look at the specific IP protocol field,
1575 irrespective of IPSEC kernel support and the validity of the IPSEC data.
1577 Further note that this flag is silently ignored in kernels without
1579 It does not affect rule processing when given and the
1580 rules are handled as if with no
1583 .It Cm iptos Ar spec
1584 Matches IPv4 packets whose
1586 field contains the comma separated list of
1587 service types specified in
1589 The supported IP types of service are:
1592 .Pq Dv IPTOS_LOWDELAY ,
1594 .Pq Dv IPTOS_THROUGHPUT ,
1596 .Pq Dv IPTOS_RELIABILITY ,
1598 .Pq Dv IPTOS_MINCOST ,
1600 .Pq Dv IPTOS_ECN_CE .
1601 The absence of a particular type may be denoted
1604 .It Cm dscp spec Ns Op , Ns Ar spec
1605 Matches IPv4/IPv6 packets whose
1607 field value is contained in
1610 Multiple values can be specified via
1611 the comma separated list.
1612 Value can be one of keywords used in
1614 action or exact number.
1615 .It Cm ipttl Ar ttl-list
1616 Matches IPv4 packets whose time to live is included in
1618 which is either a single value or a list of values or ranges
1619 specified in the same way as
1621 .It Cm ipversion Ar ver
1622 Matches IP packets whose IP version field is
1624 .It Cm keep-state Op Ar :flowname
1625 Upon a match, the firewall will create a dynamic rule, whose
1626 default behaviour is to match bidirectional traffic between
1627 source and destination IP/port using the same protocol.
1628 The rule has a limited lifetime (controlled by a set of
1630 variables), and the lifetime is refreshed every time a matching
1634 is used to assign additional to addresses, ports and protocol parameter
1635 to dynamic rule. It can be used for more accurate matching by
1640 keyword is special name used for compatibility with old rulesets.
1642 Matches only layer2 packets, i.e., those passed to
1644 from ether_demux() and ether_output_frame().
1645 .It Cm limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N Op Ar :flowname
1646 The firewall will only allow
1648 connections with the same
1649 set of parameters as specified in the rule.
1651 of source and destination addresses and ports can be
1653 .It Cm lookup Bro Cm dst-ip | dst-port | src-ip | src-port | uid | jail Brc Ar name
1654 Search an entry in lookup table
1656 that matches the field specified as argument.
1657 If not found, the match fails.
1658 Otherwise, the match succeeds and
1660 is set to the value extracted from the table.
1662 This option can be useful to quickly dispatch traffic based on
1663 certain packet fields.
1666 section below for more information on lookup tables.
1667 .It Cm { MAC | mac } Ar dst-mac src-mac
1668 Match packets with a given
1672 addresses, specified as the
1674 keyword (matching any MAC address), or six groups of hex digits
1675 separated by colons,
1676 and optionally followed by a mask indicating the significant bits.
1677 The mask may be specified using either of the following methods:
1678 .Bl -enum -width indent
1682 followed by the number of significant bits.
1683 For example, an address with 33 significant bits could be specified as:
1685 .Dl "MAC 10:20:30:40:50:60/33 any"
1689 followed by a bitmask specified as six groups of hex digits separated
1691 For example, an address in which the last 16 bits are significant could
1694 .Dl "MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any"
1696 Note that the ampersand character has a special meaning in many shells
1697 and should generally be escaped.
1699 Note that the order of MAC addresses (destination first,
1701 the same as on the wire, but the opposite of the one used for
1703 .It Cm mac-type Ar mac-type
1704 Matches packets whose Ethernet Type field
1705 corresponds to one of those specified as argument.
1707 is specified in the same way as
1709 (i.e., one or more comma-separated single values or ranges).
1710 You can use symbolic names for known values such as
1711 .Em vlan , ipv4, ipv6 .
1712 Values can be entered as decimal or hexadecimal (if prefixed by 0x),
1713 and they are always printed as hexadecimal (unless the
1715 option is used, in which case symbolic resolution will be attempted).
1716 .It Cm proto Ar protocol
1717 Matches packets with the corresponding IP protocol.
1718 .It Cm recv | xmit | via Brq Ar ifX | Ar if Ns Cm * | Ar table Ns Po Ar name Ns Oo , Ns Ar value Oc Pc | Ar ipno | Ar any
1719 Matches packets received, transmitted or going through,
1720 respectively, the interface specified by exact name
1724 by IP address, or through some interface.
1727 may be used to match interface by its kernel ifindex.
1730 section below for more information on lookup tables.
1734 keyword causes the interface to always be checked.
1741 then only the receive or transmit interface (respectively)
1743 By specifying both, it is possible to match packets based on
1744 both receive and transmit interface, e.g.:
1746 .Dl "ipfw add deny ip from any to any out recv ed0 xmit ed1"
1750 interface can be tested on either incoming or outgoing packets,
1753 interface can only be tested on outgoing packets.
1758 is invalid) whenever
1762 A packet might not have a receive or transmit interface: packets
1763 originating from the local host have no receive interface,
1764 while packets destined for the local host have no transmit
1767 Matches TCP packets that have the SYN bit set but no ACK bit.
1768 This is the short form of
1769 .Dq Li tcpflags\ syn,!ack .
1771 Matches packets that are associated to a local socket and
1772 for which the SO_USER_COOKIE socket option has been set
1773 to a non-zero value.
1774 As a side effect, the value of the
1775 option is made available as
1777 value, which in turn can be used as
1782 .It Cm src-ip Ar ip-address
1783 Matches IPv4 packets whose source IP is one of the address(es)
1784 specified as an argument.
1785 .It Cm src-ip6 Ar ip6-address
1786 Matches IPv6 packets whose source IP is one of the address(es)
1787 specified as an argument.
1788 .It Cm src-port Ar ports
1789 Matches IP packets whose source port is one of the port(s)
1790 specified as argument.
1791 .It Cm tagged Ar tag-list
1792 Matches packets whose tags are included in
1794 which is either a single value or a list of values or ranges
1795 specified in the same way as
1797 Tags can be applied to the packet using
1799 rule action parameter (see it's description for details on tags).
1800 .It Cm tcpack Ar ack
1802 Match if the TCP header acknowledgment number field is set to
1804 .It Cm tcpdatalen Ar tcpdatalen-list
1805 Matches TCP packets whose length of TCP data is
1806 .Ar tcpdatalen-list ,
1807 which is either a single value or a list of values or ranges
1808 specified in the same way as
1810 .It Cm tcpflags Ar spec
1812 Match if the TCP header contains the comma separated list of
1815 The supported TCP flags are:
1824 The absence of a particular flag may be denoted
1827 A rule which contains a
1829 specification can never match a fragmented packet which has
1833 option for details on matching fragmented packets.
1834 .It Cm tcpseq Ar seq
1836 Match if the TCP header sequence number field is set to
1838 .It Cm tcpwin Ar tcpwin-list
1839 Matches TCP packets whose header window field is set to
1841 which is either a single value or a list of values or ranges
1842 specified in the same way as
1844 .It Cm tcpoptions Ar spec
1846 Match if the TCP header contains the comma separated list of
1847 options specified in
1849 The supported TCP options are:
1852 (maximum segment size),
1854 (tcp window advertisement),
1858 (rfc1323 timestamp) and
1860 (rfc1644 t/tcp connection count).
1861 The absence of a particular option may be denoted
1865 Match all TCP or UDP packets sent by or received for a
1869 may be matched by name or identification number.
1871 For incoming packets,
1872 a routing table lookup is done on the packet's source address.
1873 If the interface on which the packet entered the system matches the
1874 outgoing interface for the route,
1876 If the interfaces do not match up,
1877 the packet does not match.
1878 All outgoing packets or packets with no incoming interface match.
1880 The name and functionality of the option is intentionally similar to
1881 the Cisco IOS command:
1883 .Dl ip verify unicast reverse-path
1885 This option can be used to make anti-spoofing rules to reject all
1886 packets with source addresses not from this interface.
1890 For incoming packets,
1891 a routing table lookup is done on the packet's source address.
1892 If a route to the source address exists, but not the default route
1893 or a blackhole/reject route, the packet matches.
1894 Otherwise, the packet does not match.
1895 All outgoing packets match.
1897 The name and functionality of the option is intentionally similar to
1898 the Cisco IOS command:
1900 .Dl ip verify unicast source reachable-via any
1902 This option can be used to make anti-spoofing rules to reject all
1903 packets whose source address is unreachable.
1905 For incoming packets, the packet's source address is checked if it
1906 belongs to a directly connected network.
1907 If the network is directly connected, then the interface the packet
1908 came on in is compared to the interface the network is connected to.
1909 When incoming interface and directly connected interface are not the
1910 same, the packet does not match.
1911 Otherwise, the packet does match.
1912 All outgoing packets match.
1914 This option can be used to make anti-spoofing rules to reject all
1915 packets that pretend to be from a directly connected network but do
1916 not come in through that interface.
1917 This option is similar to but more restricted than
1919 because it engages only on packets with source addresses of directly
1920 connected networks instead of all source addresses.
1923 Lookup tables are useful to handle large sparse sets of
1924 addresses or other search keys (e.g., ports, jail IDs, interface names).
1925 In the rest of this section we will use the term ``key''.
1926 Table name needs to match the following spec:
1928 Tables with the same name can be created in different
1930 However, rule links to the tables in
1933 This behavior can be controlled by
1934 .Va net.inet.ip.fw.tables_sets
1938 section for more information.
1939 There may be up to 65535 different lookup tables.
1941 The following table types are supported:
1942 .Bl -tag -width indent
1943 .It Ar table-type : Ar addr | iface | number | flow
1944 .It Ar table-key : Ar addr Ns Oo / Ns Ar masklen Oc | iface-name | number | flow-spec
1945 .It Ar flow-spec : Ar flow-field Ns Op , Ns Ar flow-spec
1946 .It Ar flow-field : src-ip | proto | src-port | dst-ip | dst-port
1948 matches IPv4 or IPv6 address.
1949 Each entry is represented by an
1950 .Ar addr Ns Op / Ns Ar masklen
1951 and will match all addresses with base
1953 (specified as an IPv4/IPv6 address, or a hostname) and mask width of
1958 is not specified, it defaults to 32 for IPv4 and 128 for IPv6.
1959 When looking up an IP address in a table, the most specific
1962 matches interface names.
1963 Each entry is represented by string treated as interface name.
1964 Wildcards are not supported.
1966 maches protocol ports, uids/gids or jail IDs.
1967 Each entry is represented by 32-bit unsigned integer.
1968 Ranges are not supported.
1970 Matches packet fields specified by
1972 type suboptions with table entries.
1975 Tables require explicit creation via
1979 The following creation options are supported:
1980 .Bl -tag -width indent
1981 .It Ar create-options : Ar create-option | create-options
1982 .It Ar create-option : Cm type Ar table-type | Cm valtype Ar value-mask | Cm algo Ar algo-desc |
1983 .Cm limit Ar number | Cm locked
1989 Table algorithm to use (see below).
1991 Maximum number of items that may be inserted into table.
1993 Restrict any table modifications.
1996 Some of these options may be modified later via
1999 The following options can be changed:
2000 .Bl -tag -width indent
2001 .It Ar modify-options : Ar modify-option | modify-options
2002 .It Ar modify-option : Cm limit Ar number
2004 Alter maximum number of items that may be inserted into table.
2007 Additionally, table can be locked or unlocked using
2015 can be swapped with each other using
2018 Swap may fail if tables limits are set and data exchange
2019 would result in limits hit.
2020 Operation is performed atomically.
2022 One or more entries can be added to a table at once using
2025 Addition of all items are performed atomically.
2026 By default, error in addition of one entry does not influence
2027 addition of other entries. However, non-zero error code is returned
2031 keyword may be specified before
2033 to indicate all-or-none add request.
2035 One or more entries can be removed from a table at once using
2038 By default, error in removal of one entry does not influence
2039 removing of other entries. However, non-zero error code is returned
2042 It may be possible to check what entry will be found on particular
2048 This functionality is optional and may be unsupported in some algorithms.
2050 The following operations can be performed on
2055 .Bl -tag -width indent
2059 Removes all entries.
2061 Shows generic table information.
2063 Shows generic table information and algo-specific data.
2066 The following lookup algorithms are supported:
2067 .Bl -tag -width indent
2068 .It Ar algo-desc : algo-name | "algo-name algo-data"
2069 .It Ar algo-name: Ar addr:radix | addr:hash | iface:array | number:array | flow:hash
2071 Separate Radix trees for IPv4 and IPv6, the same way as the routing table (see
2077 Separate auto-growing hashes for IPv4 and IPv6.
2078 Accepts entries with the same mask length specified initially via
2079 .Cm "addr:hash masks=/v4,/v6"
2080 algorithm creation options.
2081 Assume /32 and /128 masks by default.
2082 Search removes host bits (according to mask) from supplied address and checks
2083 resulting key in appropriate hash.
2084 Mostly optimized for /64 and byte-ranged IPv6 masks.
2086 Array storing sorted indexes for entries which are presented in the system.
2087 Optimized for very fast lookup.
2089 Array storing sorted u32 numbers.
2091 Auto-growing hash storing flow entries.
2092 Search calculates hash on required packet fields and searches for matching
2093 entries in selected bucket.
2098 feature provides the ability to use a value, looked up in the table, as
2099 the argument for a rule action, action parameter or rule option.
2100 This can significantly reduce number of rules in some configurations.
2101 If two tables are used in a rule, the result of the second (destination)
2104 Each record may hold one or more values according to
2106 This mask is set on table creation via
2109 The following value types are supported:
2110 .Bl -tag -width indent
2111 .It Ar value-mask : Ar value-type Ns Op , Ns Ar value-mask
2112 .It Ar value-type : Ar skipto | pipe | fib | nat | dscp | tag | divert |
2113 .Ar netgraph | limit | ipv4
2115 rule number to jump to.
2119 fib number to match/set.
2121 nat number to jump to.
2123 dscp value to match/set.
2125 tag number to match/set.
2127 port number to divert traffic to.
2129 hook number to move packet to.
2131 maximum number of connections.
2133 IPv4 nexthop to fwd packets to.
2135 IPv6 nexthop to fwd packets to.
2140 argument can be used with the following actions:
2141 .Cm nat, pipe , queue, divert, tee, netgraph, ngtee, fwd, skipto, setfib,
2149 action, the user should be aware that the code will walk the ruleset
2150 up to a rule equal to, or past, the given number.
2154 Section for example usage of tables and the tablearg keyword.
2156 Each rule or table belongs to one of 32 different
2159 Set 31 is reserved for the default rule.
2161 By default, rules or tables are put in set 0, unless you use the
2163 attribute when adding a new rule or table.
2164 Sets can be individually and atomically enabled or disabled,
2165 so this mechanism permits an easy way to store multiple configurations
2166 of the firewall and quickly (and atomically) switch between them.
2168 By default, tables from set 0 are referenced when adding rule with
2169 table opcodes regardless of rule set.
2170 This behavior can be changed by setting
2171 .Va net.inet.ip.fw.tables_set
2173 Rule's set will then be used for table references.
2175 The command to enable/disable sets is
2176 .Bd -ragged -offset indent
2178 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
2185 sections can be specified.
2186 Command execution is atomic on all the sets specified in the command.
2187 By default, all sets are enabled.
2189 When you disable a set, its rules behave as if they do not exist
2190 in the firewall configuration, with only one exception:
2191 .Bd -ragged -offset indent
2192 dynamic rules created from a rule before it had been disabled
2193 will still be active until they expire.
2195 dynamic rules you have to explicitly delete the parent rule
2196 which generated them.
2199 The set number of rules can be changed with the command
2200 .Bd -ragged -offset indent
2203 .Brq Cm rule Ar rule-number | old-set
2207 Also, you can atomically swap two rulesets with the command
2208 .Bd -ragged -offset indent
2210 .Cm set swap Ar first-set second-set
2215 Section on some possible uses of sets of rules.
2216 .Sh STATEFUL FIREWALL
2217 Stateful operation is a way for the firewall to dynamically
2218 create rules for specific flows when packets that
2219 match a given pattern are detected.
2220 Support for stateful
2221 operation comes through the
2222 .Cm check-state , keep-state
2228 Dynamic rules are created when a packet matches a
2232 rule, causing the creation of a
2234 rule which will match all and only packets with
2238 .Em src-ip/src-port dst-ip/dst-port
2243 are used here only to denote the initial match addresses, but they
2244 are completely equivalent afterwards).
2250 This name is used in matching together with addresses, ports and protocol.
2251 Dynamic rules will be checked at the first
2252 .Cm check-state, keep-state
2255 occurrence, and the action performed upon a match will be the same
2256 as in the parent rule.
2258 Note that no additional attributes other than protocol and IP addresses
2259 and ports and :flowname are checked on dynamic rules.
2261 The typical use of dynamic rules is to keep a closed firewall configuration,
2262 but let the first TCP SYN packet from the inside network install a
2263 dynamic rule for the flow so that packets belonging to that session
2264 will be allowed through the firewall:
2266 .Dl "ipfw add check-state :OUTBOUND"
2267 .Dl "ipfw add allow tcp from my-subnet to any setup keep-state :OUTBOUND"
2268 .Dl "ipfw add deny tcp from any to any"
2270 A similar approach can be used for UDP, where an UDP packet coming
2271 from the inside will install a dynamic rule to let the response through
2274 .Dl "ipfw add check-state :OUTBOUND"
2275 .Dl "ipfw add allow udp from my-subnet to any keep-state :OUTBOUND"
2276 .Dl "ipfw add deny udp from any to any"
2278 Dynamic rules expire after some time, which depends on the status
2279 of the flow and the setting of some
2283 .Sx SYSCTL VARIABLES
2285 For TCP sessions, dynamic rules can be instructed to periodically
2286 send keepalive packets to refresh the state of the rule when it is
2291 for more examples on how to use dynamic rules.
2292 .Sh TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
2294 is also the user interface for the
2296 traffic shaper, packet scheduler and network emulator, a subsystem that
2297 can artificially queue, delay or drop packets
2298 emulating the behaviour of certain network links
2299 or queueing systems.
2302 operates by first using the firewall to select packets
2303 using any match pattern that can be used in
2306 Matching packets are then passed to either of two
2307 different objects, which implement the traffic regulation:
2308 .Bl -hang -offset XXXX
2314 with given bandwidth and propagation delay,
2315 driven by a FIFO scheduler and a single queue with programmable
2316 queue size and packet loss rate.
2317 Packets are appended to the queue as they come out from
2319 and then transferred in FIFO order to the link at the desired rate.
2323 is an abstraction used to implement packet scheduling
2324 using one of several packet scheduling algorithms.
2327 are first grouped into flows according to a mask on the 5-tuple.
2328 Flows are then passed to the scheduler associated to the
2330 and each flow uses scheduling parameters (weight and others)
2331 as configured in the
2334 A scheduler in turn is connected to an emulated link,
2335 and arbitrates the link's bandwidth among backlogged flows according to
2336 weights and to the features of the scheduling algorithm in use.
2341 can be used to set hard limits to the bandwidth that a flow can use, whereas
2343 can be used to determine how different flows share the available bandwidth.
2345 A graphical representation of the binding of queues,
2346 flows, schedulers and links is below.
2347 .Bd -literal -offset indent
2348 (flow_mask|sched_mask) sched_mask
2349 +---------+ weight Wx +-------------+
2350 | |->-[flow]-->--| |-+
2351 -->--| QUEUE x | ... | | |
2352 | |->-[flow]-->--| SCHEDuler N | |
2354 ... | +--[LINK N]-->--
2355 +---------+ weight Wy | | +--[LINK N]-->--
2356 | |->-[flow]-->--| | |
2357 -->--| QUEUE y | ... | | |
2358 | |->-[flow]-->--| | |
2359 +---------+ +-------------+ |
2362 It is important to understand the role of the SCHED_MASK
2363 and FLOW_MASK, which are configured through the commands
2364 .Dl "ipfw sched N config mask SCHED_MASK ..."
2366 .Dl "ipfw queue X config mask FLOW_MASK ..." .
2368 The SCHED_MASK is used to assign flows to one or more
2369 scheduler instances, one for each
2370 value of the packet's 5-tuple after applying SCHED_MASK.
2371 As an example, using ``src-ip 0xffffff00'' creates one instance
2372 for each /24 destination subnet.
2374 The FLOW_MASK, together with the SCHED_MASK, is used to split
2376 As an example, using
2377 ``src-ip 0x000000ff''
2378 together with the previous SCHED_MASK makes a flow for
2379 each individual source address.
2380 In turn, flows for each /24
2381 subnet will be sent to the same scheduler instance.
2383 The above diagram holds even for the
2385 case, with the only restriction that a
2387 only supports a SCHED_MASK, and forces the use of a FIFO
2388 scheduler (these are for backward compatibility reasons;
2389 in fact, internally, a
2391 pipe is implemented exactly as above).
2393 There are two modes of
2401 mode tries to emulate a real link: the
2403 scheduler ensures that the packet will not leave the pipe faster than it
2404 would on the real link with a given bandwidth.
2407 mode allows certain packets to bypass the
2409 scheduler (if packet flow does not exceed pipe's bandwidth).
2410 This is the reason why the
2412 mode requires less CPU cycles per packet (on average) and packet latency
2413 can be significantly lower in comparison to a real link with the same
2419 mode can be enabled by setting the
2420 .Va net.inet.ip.dummynet.io_fast
2422 variable to a non-zero value.
2424 .Ss PIPE, QUEUE AND SCHEDULER CONFIGURATION
2430 configuration commands are the following:
2431 .Bd -ragged -offset indent
2432 .Cm pipe Ar number Cm config Ar pipe-configuration
2434 .Cm queue Ar number Cm config Ar queue-configuration
2436 .Cm sched Ar number Cm config Ar sched-configuration
2439 The following parameters can be configured for a pipe:
2441 .Bl -tag -width indent -compact
2442 .It Cm bw Ar bandwidth | device
2443 Bandwidth, measured in
2446 .Brq Cm bit/s | Byte/s .
2449 A value of 0 (default) means unlimited bandwidth.
2450 The unit must immediately follow the number, as in
2452 .Dl "ipfw pipe 1 config bw 300Kbit/s"
2454 If a device name is specified instead of a numeric value, as in
2456 .Dl "ipfw pipe 1 config bw tun0"
2458 then the transmit clock is supplied by the specified device.
2459 At the moment only the
2461 device supports this
2462 functionality, for use in conjunction with
2465 .It Cm delay Ar ms-delay
2466 Propagation delay, measured in milliseconds.
2467 The value is rounded to the next multiple of the clock tick
2468 (typically 10ms, but it is a good practice to run kernels
2470 .Dq "options HZ=1000"
2472 the granularity to 1ms or less).
2473 The default value is 0, meaning no delay.
2475 .It Cm burst Ar size
2476 If the data to be sent exceeds the pipe's bandwidth limit
2477 (and the pipe was previously idle), up to
2479 bytes of data are allowed to bypass the
2481 scheduler, and will be sent as fast as the physical link allows.
2482 Any additional data will be transmitted at the rate specified
2486 The burst size depends on how long the pipe has been idle;
2487 the effective burst size is calculated as follows:
2494 .It Cm profile Ar filename
2495 A file specifying the additional overhead incurred in the transmission
2496 of a packet on the link.
2498 Some link types introduce extra delays in the transmission
2499 of a packet, e.g., because of MAC level framing, contention on
2500 the use of the channel, MAC level retransmissions and so on.
2501 From our point of view, the channel is effectively unavailable
2502 for this extra time, which is constant or variable depending
2504 Additionally, packets may be dropped after this
2505 time (e.g., on a wireless link after too many retransmissions).
2506 We can model the additional delay with an empirical curve
2507 that represents its distribution.
2508 .Bd -literal -offset indent
2509 cumulative probability
2519 +-------*------------------->
2522 The empirical curve may have both vertical and horizontal lines.
2523 Vertical lines represent constant delay for a range of
2525 Horizontal lines correspond to a discontinuity in the delay
2526 distribution: the pipe will use the largest delay for a
2529 The file format is the following, with whitespace acting as
2530 a separator and '#' indicating the beginning a comment:
2531 .Bl -tag -width indent
2532 .It Cm name Ar identifier
2533 optional name (listed by "ipfw pipe show")
2534 to identify the delay distribution;
2536 the bandwidth used for the pipe.
2537 If not specified here, it must be present
2538 explicitly as a configuration parameter for the pipe;
2539 .It Cm loss-level Ar L
2540 the probability above which packets are lost.
2541 (0.0 <= L <= 1.0, default 1.0 i.e., no loss);
2543 the number of samples used in the internal
2544 representation of the curve (2..1024; default 100);
2545 .It Cm "delay prob" | "prob delay"
2546 One of these two lines is mandatory and defines
2547 the format of the following lines with data points.
2549 2 or more lines representing points in the curve,
2550 with either delay or probability first, according
2551 to the chosen format.
2552 The unit for delay is milliseconds.
2553 Data points do not need to be sorted.
2554 Also, the number of actual lines can be different
2555 from the value of the "samples" parameter:
2557 utility will sort and interpolate
2558 the curve as needed.
2561 Example of a profile file:
2562 .Bd -literal -offset indent
2567 0 200 # minimum overhead is 200ms
2573 #configuration file end
2577 The following parameters can be configured for a queue:
2579 .Bl -tag -width indent -compact
2580 .It Cm pipe Ar pipe_nr
2581 Connects a queue to the specified pipe.
2582 Multiple queues (with the same or different weights) can be connected to
2583 the same pipe, which specifies the aggregate rate for the set of queues.
2585 .It Cm weight Ar weight
2586 Specifies the weight to be used for flows matching this queue.
2587 The weight must be in the range 1..100, and defaults to 1.
2590 The following case-insensitive parameters can be configured for a
2593 .Bl -tag -width indent -compact
2594 .It Cm type Ar {fifo | wf2q+ | rr | qfq}
2595 specifies the scheduling algorithm to use.
2596 .Bl -tag -width indent -compact
2598 is just a FIFO scheduler (which means that all packets
2599 are stored in the same queue as they arrive to the scheduler).
2600 FIFO has O(1) per-packet time complexity, with very low
2601 constants (estimate 60-80ns on a 2GHz desktop machine)
2602 but gives no service guarantees.
2604 implements the WF2Q+ algorithm, which is a Weighted Fair Queueing
2605 algorithm which permits flows to share bandwidth according to
2607 Note that weights are not priorities; even a flow
2608 with a minuscule weight will never starve.
2609 WF2Q+ has O(log N) per-packet processing cost, where N is the number
2610 of flows, and is the default algorithm used by previous versions
2613 implements the Deficit Round Robin algorithm, which has O(1) processing
2614 costs (roughly, 100-150ns per packet)
2615 and permits bandwidth allocation according to weights, but
2616 with poor service guarantees.
2618 implements the QFQ algorithm, which is a very fast variant of
2619 WF2Q+, with similar service guarantees and O(1) processing
2620 costs (roughly, 200-250ns per packet).
2624 In addition to the type, all parameters allowed for a pipe can also
2625 be specified for a scheduler.
2627 Finally, the following parameters can be configured for both
2630 .Bl -tag -width XXXX -compact
2631 .It Cm buckets Ar hash-table-size
2632 Specifies the size of the hash table used for storing the
2634 Default value is 64 controlled by the
2637 .Va net.inet.ip.dummynet.hash_size ,
2638 allowed range is 16 to 65536.
2640 .It Cm mask Ar mask-specifier
2641 Packets sent to a given pipe or queue by an
2643 rule can be further classified into multiple flows, each of which is then
2647 A flow identifier is constructed by masking the IP addresses,
2648 ports and protocol types as specified with the
2650 options in the configuration of the pipe or queue.
2651 For each different flow identifier, a new pipe or queue is created
2652 with the same parameters as the original object, and matching packets
2657 are used, each flow will get the same bandwidth as defined by the pipe,
2660 are used, each flow will share the parent's pipe bandwidth evenly
2661 with other flows generated by the same queue (note that other queues
2662 with different weights might be connected to the same pipe).
2664 Available mask specifiers are a combination of one or more of the following:
2666 .Cm dst-ip Ar mask ,
2667 .Cm dst-ip6 Ar mask ,
2668 .Cm src-ip Ar mask ,
2669 .Cm src-ip6 Ar mask ,
2670 .Cm dst-port Ar mask ,
2671 .Cm src-port Ar mask ,
2672 .Cm flow-id Ar mask ,
2677 where the latter means all bits in all fields are significant.
2680 When a packet is dropped by a
2682 queue or pipe, the error
2683 is normally reported to the caller routine in the kernel, in the
2684 same way as it happens when a device queue fills up.
2686 option reports the packet as successfully delivered, which can be
2687 needed for some experimental setups where you want to simulate
2688 loss or congestion at a remote router.
2690 .It Cm plr Ar packet-loss-rate
2693 .Ar packet-loss-rate
2694 is a floating-point number between 0 and 1, with 0 meaning no
2695 loss, 1 meaning 100% loss.
2696 The loss rate is internally represented on 31 bits.
2698 .It Cm queue Brq Ar slots | size Ns Cm Kbytes
2703 Default value is 50 slots, which
2704 is the typical queue size for Ethernet devices.
2705 Note that for slow speed links you should keep the queue
2706 size short or your traffic might be affected by a significant
2708 E.g., 50 max-sized ethernet packets (1500 bytes) mean 600Kbit
2709 or 20s of queue on a 30Kbit/s pipe.
2710 Even worse effects can result if you get packets from an
2711 interface with a much larger MTU, e.g.\& the loopback interface
2712 with its 16KB packets.
2716 .Em net.inet.ip.dummynet.pipe_byte_limit
2718 .Em net.inet.ip.dummynet.pipe_slot_limit
2719 control the maximum lengths that can be specified.
2721 .It Cm red | gred Ar w_q Ns / Ns Ar min_th Ns / Ns Ar max_th Ns / Ns Ar max_p
2723 Make use of the RED (Random Early Detection) queue management algorithm.
2728 point numbers between 0 and 1 (inclusive), while
2732 are integer numbers specifying thresholds for queue management
2733 (thresholds are computed in bytes if the queue has been defined
2734 in bytes, in slots otherwise).
2735 The two parameters can also be of the same value if needed. The
2737 also supports the gentle RED variant (gred) and ECN (Explicit Congestion
2738 Notification) as optional. Three
2740 variables can be used to control the RED behaviour:
2741 .Bl -tag -width indent
2742 .It Va net.inet.ip.dummynet.red_lookup_depth
2743 specifies the accuracy in computing the average queue
2744 when the link is idle (defaults to 256, must be greater than zero)
2745 .It Va net.inet.ip.dummynet.red_avg_pkt_size
2746 specifies the expected average packet size (defaults to 512, must be
2748 .It Va net.inet.ip.dummynet.red_max_pkt_size
2749 specifies the expected maximum packet size, only used when queue
2750 thresholds are in bytes (defaults to 1500, must be greater than zero).
2754 When used with IPv6 data,
2756 currently has several limitations.
2757 Information necessary to route link-local packets to an
2758 interface is not available after processing by
2760 so those packets are dropped in the output path.
2761 Care should be taken to ensure that link-local packets are not passed to
2764 Here are some important points to consider when designing your
2768 Remember that you filter both packets going
2772 Most connections need packets going in both directions.
2774 Remember to test very carefully.
2775 It is a good idea to be near the console when doing this.
2776 If you cannot be near the console,
2777 use an auto-recovery script such as the one in
2778 .Pa /usr/share/examples/ipfw/change_rules.sh .
2780 Do not forget the loopback interface.
2785 There are circumstances where fragmented datagrams are unconditionally
2787 TCP packets are dropped if they do not contain at least 20 bytes of
2788 TCP header, UDP packets are dropped if they do not contain a full 8
2789 byte UDP header, and ICMP packets are dropped if they do not contain
2790 4 bytes of ICMP header, enough to specify the ICMP type, code, and
2792 These packets are simply logged as
2794 since there may not be enough good data in the packet to produce a
2795 meaningful log entry.
2797 Another type of packet is unconditionally dropped, a TCP packet with a
2798 fragment offset of one.
2799 This is a valid packet, but it only has one use, to try
2800 to circumvent firewalls.
2801 When logging is enabled, these packets are
2802 reported as being dropped by rule -1.
2804 If you are logged in over a network, loading the
2808 is probably not as straightforward as you would think.
2809 The following command line is recommended:
2810 .Bd -literal -offset indent
2812 ipfw add 32000 allow ip from any to any
2815 Along the same lines, doing an
2816 .Bd -literal -offset indent
2820 in similar surroundings is also a bad idea.
2824 filter list may not be modified if the system security level
2825 is set to 3 or higher
2828 for information on system security levels).
2830 .Sh PACKET DIVERSION
2833 socket bound to the specified port will receive all packets
2834 diverted to that port.
2835 If no socket is bound to the destination port, or if the divert module is
2836 not loaded, or if the kernel was not compiled with divert socket support,
2837 the packets are dropped.
2838 .Sh NETWORK ADDRESS TRANSLATION (NAT)
2840 support in-kernel NAT using the kernel version of
2843 The nat configuration command is the following:
2844 .Bd -ragged -offset indent
2849 .Ar nat-configuration
2853 The following parameters can be configured:
2854 .Bl -tag -width indent
2855 .It Cm ip Ar ip_address
2856 Define an ip address to use for aliasing.
2858 Use ip address of NIC for aliasing, dynamically changing
2859 it if NIC's ip address changes.
2861 Enable logging on this nat instance.
2863 Deny any incoming connection from outside world.
2865 Try to leave the alias port numbers unchanged from
2866 the actual local port numbers.
2868 Traffic on the local network not originating from an
2869 unregistered address spaces will be ignored.
2871 Reset table of the packet aliasing engine on address change.
2873 Reverse the way libalias handles aliasing.
2875 Obey transparent proxy rules only, packet aliasing is not performed.
2877 Skip instance in case of global state lookup (see below).
2880 Some specials value can be supplied instead of
2882 .Bl -tag -width indent
2884 Looks up translation state in all configured nat instances.
2885 If an entry is found, packet is aliased according to that entry.
2886 If no entry was found in any of the instances, packet is passed unchanged,
2887 and no new entry will be created.
2889 .Sx MULTIPLE INSTANCES
2892 for more information.
2894 Uses argument supplied in lookup table.
2897 section below for more information on lookup tables.
2900 To let the packet continue after being (de)aliased, set the sysctl variable
2901 .Va net.inet.ip.fw.one_pass
2903 For more information about aliasing modes, refer to
2907 for some examples about nat usage.
2908 .Ss REDIRECT AND LSNAT SUPPORT IN IPFW
2909 Redirect and LSNAT support follow closely the syntax used in
2913 for some examples on how to do redirect and lsnat.
2914 .Ss SCTP NAT SUPPORT
2915 SCTP nat can be configured in a similar manner to TCP through the
2918 The main difference is that
2920 does not do port translation.
2921 Since the local and global side ports will be the same,
2922 there is no need to specify both.
2923 Ports are redirected as follows:
2924 .Bd -ragged -offset indent
2930 .Cm redirect_port sctp
2931 .Ar ip_address [,addr_list] {[port | port-port] [,ports]}
2937 configuration can be done in real-time through the
2940 All may be changed dynamically, though the hash_table size will only
2945 .Sx SYSCTL VARIABLES
2947 .Sh IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
2949 support in-kernel IPv6-to-IPv6 network prefix translation as described
2953 should be loaded or kernel should has
2954 .Cm options IPFIREWALL_NPTV6
2955 to be able use NPTv6 translator.
2957 The NPTv6 configuration command is the following:
2958 .Bd -ragged -offset indent
2967 The following parameters can be configured:
2968 .Bl -tag -width indent
2969 .It Cm int_prefix Ar ipv6_prefix
2970 IPv6 prefix used in internal network.
2971 NPTv6 module translates source address when it matches this prefix.
2972 .It Cm ext_prefix Ar ipv6_prefix
2973 IPv6 prefix used in external network.
2974 NPTv6 module translates destination address when it matches this prefix.
2975 .It Cm prefixlen Ar length
2976 The length of specified IPv6 prefixes. It must be in range from 8 to 64.
2979 Note that the prefix translation rules are silently ignored when IPv6 packet
2980 forwarding is disabled.
2981 To enable the packet forwarding, set the sysctl variable
2982 .Va net.inet6.ip6.forwarding
2985 To let the packet continue after being translated, set the sysctl variable
2986 .Va net.inet.ip.fw.one_pass
2989 Tunables can be set in
2995 before ipfw module gets loaded.
2996 .Bl -tag -width indent
2997 .It Va net.inet.ip.fw.default_to_accept: No 0
2998 Defines ipfw last rule behavior.
2999 This value overrides
3000 .Cd "options IPFW_DEFAULT_TO_(ACCEPT|DENY)"
3001 from kernel configuration file.
3002 .It Va net.inet.ip.fw.tables_max: No 128
3003 Defines number of tables available in ipfw.
3004 Number cannot exceed 65534.
3006 .Sh SYSCTL VARIABLES
3009 variables controls the behaviour of the firewall and
3011 .Pq Nm dummynet , bridge , sctp nat .
3012 These are shown below together with their default value
3013 (but always check with the
3015 command what value is actually in use) and meaning:
3016 .Bl -tag -width indent
3017 .It Va net.inet.ip.alias.sctp.accept_global_ootb_addip: No 0
3020 responds to receipt of global OOTB ASCONF-AddIP:
3021 .Bl -tag -width indent
3023 No response (unless a partially matching association exists -
3024 ports and vtags match but global address does not)
3027 will accept and process all OOTB global AddIP messages.
3030 Option 1 should never be selected as this forms a security risk.
3032 establish multiple fake associations by sending AddIP messages.
3033 .It Va net.inet.ip.alias.sctp.chunk_proc_limit: No 5
3034 Defines the maximum number of chunks in an SCTP packet that will be
3036 packet that matches an existing association.
3037 This value is enforced to be greater or equal than
3038 .Cm net.inet.ip.alias.sctp.initialising_chunk_proc_limit .
3040 a DoS risk yet setting too low a value may result in
3041 important control chunks in
3042 the packet not being located and parsed.
3043 .It Va net.inet.ip.alias.sctp.error_on_ootb: No 1
3046 responds to any Out-of-the-Blue (OOTB) packets with ErrorM packets.
3047 An OOTB packet is a packet that arrives with no existing association
3050 and is not an INIT or ASCONF-AddIP packet:
3051 .Bl -tag -width indent
3053 ErrorM is never sent in response to OOTB packets.
3055 ErrorM is only sent to OOTB packets received on the local side.
3057 ErrorM is sent to the local side and on the global side ONLY if there is a
3058 partial match (ports and vtags match but the source global IP does not).
3059 This value is only useful if the
3061 is tracking global IP addresses.
3063 ErrorM is sent in response to all OOTB packets on both
3064 the local and global side
3068 At the moment the default is 0, since the ErrorM packet is not yet
3069 supported by most SCTP stacks.
3070 When it is supported, and if not tracking
3071 global addresses, we recommend setting this value to 1 to allow
3072 multi-homed local hosts to function with the
3074 To track global addresses, we recommend setting this value to 2 to
3075 allow global hosts to be informed when they need to (re)send an
3077 Value 3 should never be chosen (except for debugging) as the
3079 will respond to all OOTB global packets (a DoS risk).
3080 .It Va net.inet.ip.alias.sctp.hashtable_size: No 2003
3081 Size of hash tables used for
3083 lookups (100 < prime_number > 1000001).
3086 size for any future created
3088 instance and therefore must be set prior to creating a
3091 The table sizes may be changed to suit specific needs.
3092 If there will be few
3093 concurrent associations, and memory is scarce, you may make these smaller.
3094 If there will be many thousands (or millions) of concurrent associations, you
3095 should make these larger.
3096 A prime number is best for the table size.
3098 update function will adjust your input value to the next highest prime number.
3099 .It Va net.inet.ip.alias.sctp.holddown_time: No 0
3100 Hold association in table for this many seconds after receiving a
3102 This allows endpoints to correct shutdown gracefully if a
3103 shutdown_complete is lost and retransmissions are required.
3104 .It Va net.inet.ip.alias.sctp.init_timer: No 15
3105 Timeout value while waiting for (INIT-ACK|AddIP-ACK).
3106 This value cannot be 0.
3107 .It Va net.inet.ip.alias.sctp.initialising_chunk_proc_limit: No 2
3108 Defines the maximum number of chunks in an SCTP packet that will be parsed when
3109 no existing association exists that matches that packet.
3111 will only be an INIT or ASCONF-AddIP packet.
3112 A higher value may become a DoS
3113 risk as malformed packets can consume processing resources.
3114 .It Va net.inet.ip.alias.sctp.param_proc_limit: No 25
3115 Defines the maximum number of parameters within a chunk that will be
3118 As for other similar sysctl variables, larger values pose a DoS risk.
3119 .It Va net.inet.ip.alias.sctp.log_level: No 0
3120 Level of detail in the system log messages (0 \- minimal, 1 \- event,
3121 2 \- info, 3 \- detail, 4 \- debug, 5 \- max debug).
3123 option in high loss environments.
3124 .It Va net.inet.ip.alias.sctp.shutdown_time: No 15
3125 Timeout value while waiting for SHUTDOWN-COMPLETE.
3126 This value cannot be 0.
3127 .It Va net.inet.ip.alias.sctp.track_global_addresses: No 0
3128 Enables/disables global IP address tracking within the
3131 upper limit on the number of addresses tracked for each association:
3132 .Bl -tag -width indent
3134 Global tracking is disabled
3136 Enables tracking, the maximum number of addresses tracked for each
3137 association is limited to this value
3140 This variable is fully dynamic, the new value will be adopted for all newly
3141 arriving associations, existing associations are treated
3142 as they were previously.
3143 Global tracking will decrease the number of collisions within the
3146 of increased processing load, memory usage, complexity, and possible
3149 problems in complex networks with multiple
3151 We recommend not tracking
3152 global IP addresses, this will still result in a fully functional
3154 .It Va net.inet.ip.alias.sctp.up_timer: No 300
3155 Timeout value to keep an association up with no traffic.
3156 This value cannot be 0.
3157 .It Va net.inet.ip.dummynet.expire : No 1
3158 Lazily delete dynamic pipes/queue once they have no pending traffic.
3159 You can disable this by setting the variable to 0, in which case
3160 the pipes/queues will only be deleted when the threshold is reached.
3161 .It Va net.inet.ip.dummynet.hash_size : No 64
3162 Default size of the hash table used for dynamic pipes/queues.
3163 This value is used when no
3165 option is specified when configuring a pipe/queue.
3166 .It Va net.inet.ip.dummynet.io_fast : No 0
3167 If set to a non-zero value,
3172 operation (see above) is enabled.
3173 .It Va net.inet.ip.dummynet.io_pkt
3174 Number of packets passed to
3176 .It Va net.inet.ip.dummynet.io_pkt_drop
3177 Number of packets dropped by
3179 .It Va net.inet.ip.dummynet.io_pkt_fast
3180 Number of packets bypassed by the
3183 .It Va net.inet.ip.dummynet.max_chain_len : No 16
3184 Target value for the maximum number of pipes/queues in a hash bucket.
3186 .Cm max_chain_len*hash_size
3187 is used to determine the threshold over which empty pipes/queues
3188 will be expired even when
3189 .Cm net.inet.ip.dummynet.expire=0 .
3190 .It Va net.inet.ip.dummynet.red_lookup_depth : No 256
3191 .It Va net.inet.ip.dummynet.red_avg_pkt_size : No 512
3192 .It Va net.inet.ip.dummynet.red_max_pkt_size : No 1500
3193 Parameters used in the computations of the drop probability
3194 for the RED algorithm.
3195 .It Va net.inet.ip.dummynet.pipe_byte_limit : No 1048576
3196 .It Va net.inet.ip.dummynet.pipe_slot_limit : No 100
3197 The maximum queue size that can be specified in bytes or packets.
3198 These limits prevent accidental exhaustion of resources such as mbufs.
3199 If you raise these limits,
3200 you should make sure the system is configured so that sufficient resources
3202 .It Va net.inet.ip.fw.autoinc_step : No 100
3203 Delta between rule numbers when auto-generating them.
3204 The value must be in the range 1..1000.
3205 .It Va net.inet.ip.fw.curr_dyn_buckets : Va net.inet.ip.fw.dyn_buckets
3206 The current number of buckets in the hash table for dynamic rules
3208 .It Va net.inet.ip.fw.debug : No 1
3209 Controls debugging messages produced by
3211 .It Va net.inet.ip.fw.default_rule : No 65535
3212 The default rule number (read-only).
3214 .Nm , the default rule is the last one, so its number
3215 can also serve as the highest number allowed for a rule.
3216 .It Va net.inet.ip.fw.dyn_buckets : No 256
3217 The number of buckets in the hash table for dynamic rules.
3218 Must be a power of 2, up to 65536.
3219 It only takes effect when all dynamic rules have expired, so you
3220 are advised to use a
3222 command to make sure that the hash table is resized.
3223 .It Va net.inet.ip.fw.dyn_count : No 3
3224 Current number of dynamic rules
3226 .It Va net.inet.ip.fw.dyn_keepalive : No 1
3227 Enables generation of keepalive packets for
3229 rules on TCP sessions.
3230 A keepalive is generated to both
3231 sides of the connection every 5 seconds for the last 20
3232 seconds of the lifetime of the rule.
3233 .It Va net.inet.ip.fw.dyn_max : No 8192
3234 Maximum number of dynamic rules.
3235 When you hit this limit, no more dynamic rules can be
3236 installed until old ones expire.
3237 .It Va net.inet.ip.fw.dyn_ack_lifetime : No 300
3238 .It Va net.inet.ip.fw.dyn_syn_lifetime : No 20
3239 .It Va net.inet.ip.fw.dyn_fin_lifetime : No 1
3240 .It Va net.inet.ip.fw.dyn_rst_lifetime : No 1
3241 .It Va net.inet.ip.fw.dyn_udp_lifetime : No 5
3242 .It Va net.inet.ip.fw.dyn_short_lifetime : No 30
3243 These variables control the lifetime, in seconds, of dynamic
3245 Upon the initial SYN exchange the lifetime is kept short,
3246 then increased after both SYN have been seen, then decreased
3247 again during the final FIN exchange or when a RST is received.
3249 .Em dyn_fin_lifetime
3251 .Em dyn_rst_lifetime
3252 must be strictly lower than 5 seconds, the period of
3253 repetition of keepalives.
3254 The firewall enforces that.
3255 .It Va net.inet.ip.fw.dyn_keep_states: No 0
3256 Keep dynamic states on rule/set deletion.
3257 States are relinked to default rule (65535).
3258 This can be handly for ruleset reload.
3259 Turned off by default.
3260 .It Va net.inet.ip.fw.enable : No 1
3261 Enables the firewall.
3262 Setting this variable to 0 lets you run your machine without
3263 firewall even if compiled in.
3264 .It Va net.inet6.ip6.fw.enable : No 1
3265 provides the same functionality as above for the IPv6 case.
3266 .It Va net.inet.ip.fw.one_pass : No 1
3267 When set, the packet exiting from the
3271 node is not passed though the firewall again.
3272 Otherwise, after an action, the packet is
3273 reinjected into the firewall at the next rule.
3274 .It Va net.inet.ip.fw.tables_max : No 128
3275 Maximum number of tables.
3276 .It Va net.inet.ip.fw.verbose : No 1
3277 Enables verbose messages.
3278 .It Va net.inet.ip.fw.verbose_limit : No 0
3279 Limits the number of messages produced by a verbose firewall.
3280 .It Va net.inet6.ip6.fw.deny_unknown_exthdrs : No 1
3281 If enabled packets with unknown IPv6 Extension Headers will be denied.
3282 .It Va net.link.ether.ipfw : No 0
3283 Controls whether layer-2 packets are passed to
3286 .It Va net.link.bridge.ipfw : No 0
3287 Controls whether bridged packets are passed to
3291 .Sh INTERNAL DIAGNOSTICS
3292 There are some commands that may be useful to understand current state
3293 of certain subsystems inside kernel module.
3294 These commands provide debugging output which may change without notice.
3296 Currently the following commands are available as
3299 .Bl -tag -width indent
3301 Lists all interface which are currently tracked by
3303 with their in-kernel status.
3305 List all table lookup algorithms currently available.
3308 There are far too many possible uses of
3310 so this Section will only give a small set of examples.
3312 .Ss BASIC PACKET FILTERING
3313 This command adds an entry which denies all tcp packets from
3314 .Em cracker.evil.org
3315 to the telnet port of
3317 from being forwarded by the host:
3319 .Dl "ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet"
3321 This one disallows any connection from the entire cracker's
3324 .Dl "ipfw add deny ip from 123.45.67.0/24 to my.host.org"
3326 A first and efficient way to limit access (not using dynamic rules)
3327 is the use of the following rules:
3329 .Dl "ipfw add allow tcp from any to any established"
3330 .Dl "ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup"
3331 .Dl "ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup"
3333 .Dl "ipfw add deny tcp from any to any"
3335 The first rule will be a quick match for normal TCP packets,
3336 but it will not match the initial SYN packet, which will be
3339 rules only for selected source/destination pairs.
3340 All other SYN packets will be rejected by the final
3344 If you administer one or more subnets, you can take advantage
3345 of the address sets and or-blocks and write extremely
3346 compact rulesets which selectively enable services to blocks
3347 of clients, as below:
3349 .Dl "goodguys=\*q{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }\*q"
3350 .Dl "badguys=\*q10.1.2.0/24{8,38,60}\*q"
3352 .Dl "ipfw add allow ip from ${goodguys} to any"
3353 .Dl "ipfw add deny ip from ${badguys} to any"
3354 .Dl "... normal policies ..."
3358 option could be used to do automated anti-spoofing by adding the
3359 following to the top of a ruleset:
3361 .Dl "ipfw add deny ip from any to any not verrevpath in"
3363 This rule drops all incoming packets that appear to be coming to the
3364 system on the wrong interface.
3365 For example, a packet with a source
3366 address belonging to a host on a protected internal network would be
3367 dropped if it tried to enter the system from an external interface.
3371 option could be used to do similar but more restricted anti-spoofing
3372 by adding the following to the top of a ruleset:
3374 .Dl "ipfw add deny ip from any to any not antispoof in"
3376 This rule drops all incoming packets that appear to be coming from another
3377 directly connected system but on the wrong interface.
3378 For example, a packet with a source address of
3379 .Li 192.168.0.0/24 ,
3388 option could be used to (re)mark user traffic,
3389 by adding the following to the appropriate place in ruleset:
3391 .Dl "ipfw add setdscp be ip from any to any dscp af11,af21"
3393 In order to protect a site from flood attacks involving fake
3394 TCP packets, it is safer to use dynamic rules:
3396 .Dl "ipfw add check-state"
3397 .Dl "ipfw add deny tcp from any to any established"
3398 .Dl "ipfw add allow tcp from my-net to any setup keep-state"
3400 This will let the firewall install dynamic rules only for
3401 those connection which start with a regular SYN packet coming
3402 from the inside of our network.
3403 Dynamic rules are checked when encountering the first
3412 rule should usually be placed near the beginning of the
3413 ruleset to minimize the amount of work scanning the ruleset.
3414 Your mileage may vary.
3416 To limit the number of connections a user can open
3417 you can use the following type of rules:
3419 .Dl "ipfw add allow tcp from my-net/24 to any setup limit src-addr 10"
3420 .Dl "ipfw add allow tcp from any to me setup limit src-addr 4"
3422 The former (assuming it runs on a gateway) will allow each host
3423 on a /24 network to open at most 10 TCP connections.
3424 The latter can be placed on a server to make sure that a single
3425 client does not use more than 4 simultaneous connections.
3428 stateful rules can be subject to denial-of-service attacks
3429 by a SYN-flood which opens a huge number of dynamic rules.
3430 The effects of such attacks can be partially limited by
3433 variables which control the operation of the firewall.
3435 Here is a good usage of the
3437 command to see accounting records and timestamp information:
3441 or in short form without timestamps:
3445 which is equivalent to:
3449 Next rule diverts all incoming packets from 192.168.2.0/24
3450 to divert port 5000:
3452 .Dl ipfw divert 5000 ip from 192.168.2.0/24 to any in
3454 The following rules show some of the applications of
3458 for simulations and the like.
3460 This rule drops random incoming packets with a probability
3463 .Dl "ipfw add prob 0.05 deny ip from any to any in"
3465 A similar effect can be achieved making use of
3469 .Dl "ipfw add pipe 10 ip from any to any"
3470 .Dl "ipfw pipe 10 config plr 0.05"
3472 We can use pipes to artificially limit bandwidth, e.g.\& on a
3473 machine acting as a router, if we want to limit traffic from
3474 local clients on 192.168.2.0/24 we do:
3476 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
3477 .Dl "ipfw pipe 1 config bw 300Kbit/s queue 50KBytes"
3479 note that we use the
3481 modifier so that the rule is not used twice.
3482 Remember in fact that
3484 rules are checked both on incoming and outgoing packets.
3486 Should we want to simulate a bidirectional link with bandwidth
3487 limitations, the correct way is the following:
3489 .Dl "ipfw add pipe 1 ip from any to any out"
3490 .Dl "ipfw add pipe 2 ip from any to any in"
3491 .Dl "ipfw pipe 1 config bw 64Kbit/s queue 10Kbytes"
3492 .Dl "ipfw pipe 2 config bw 64Kbit/s queue 10Kbytes"
3494 The above can be very useful, e.g.\& if you want to see how
3495 your fancy Web page will look for a residential user who
3496 is connected only through a slow link.
3497 You should not use only one pipe for both directions, unless
3498 you want to simulate a half-duplex medium (e.g.\& AppleTalk,
3500 It is not necessary that both pipes have the same configuration,
3501 so we can also simulate asymmetric links.
3503 Should we want to verify network performance with the RED queue
3504 management algorithm:
3506 .Dl "ipfw add pipe 1 ip from any to any"
3507 .Dl "ipfw pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1"
3509 Another typical application of the traffic shaper is to
3510 introduce some delay in the communication.
3511 This can significantly affect applications which do a lot of Remote
3512 Procedure Calls, and where the round-trip-time of the
3513 connection often becomes a limiting factor much more than
3516 .Dl "ipfw add pipe 1 ip from any to any out"
3517 .Dl "ipfw add pipe 2 ip from any to any in"
3518 .Dl "ipfw pipe 1 config delay 250ms bw 1Mbit/s"
3519 .Dl "ipfw pipe 2 config delay 250ms bw 1Mbit/s"
3521 Per-flow queueing can be useful for a variety of purposes.
3522 A very simple one is counting traffic:
3524 .Dl "ipfw add pipe 1 tcp from any to any"
3525 .Dl "ipfw add pipe 1 udp from any to any"
3526 .Dl "ipfw add pipe 1 ip from any to any"
3527 .Dl "ipfw pipe 1 config mask all"
3529 The above set of rules will create queues (and collect
3530 statistics) for all traffic.
3531 Because the pipes have no limitations, the only effect is
3532 collecting statistics.
3533 Note that we need 3 rules, not just the last one, because
3536 tries to match IP packets it will not consider ports, so we
3537 would not see connections on separate ports as different
3540 A more sophisticated example is limiting the outbound traffic
3541 on a net with per-host limits, rather than per-network limits:
3543 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
3544 .Dl "ipfw add pipe 2 ip from any to 192.168.2.0/24 in"
3545 .Dl "ipfw pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
3546 .Dl "ipfw pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
3548 In the following example, we need to create several traffic bandwidth
3549 classes and we need different hosts/networks to fall into different classes.
3550 We create one pipe for each class and configure them accordingly.
3551 Then we create a single table and fill it with IP subnets and addresses.
3552 For each subnet/host we set the argument equal to the number of the pipe
3554 Then we classify traffic using a single rule:
3556 .Dl "ipfw pipe 1 config bw 1000Kbyte/s"
3557 .Dl "ipfw pipe 4 config bw 4000Kbyte/s"
3559 .Dl "ipfw table T1 create type addr"
3560 .Dl "ipfw table T1 add 192.168.2.0/24 1"
3561 .Dl "ipfw table T1 add 192.168.0.0/27 4"
3562 .Dl "ipfw table T1 add 192.168.0.2 1"
3564 .Dl "ipfw add pipe tablearg ip from 'table(T1)' to any"
3568 action, the table entries may include hostnames and IP addresses.
3570 .Dl "ipfw table T2 create type addr ftype ip"
3571 .Dl "ipfw table T2 add 192.168.2.0/24 10.23.2.1"
3572 .Dl "ipfw table T21 add 192.168.0.0/27 router1.dmz"
3574 .Dl "ipfw add 100 fwd tablearg ip from any to table(1)"
3576 In the following example per-interface firewall is created:
3578 .Dl "ipfw table IN create type iface valtype skipto,fib"
3579 .Dl "ipfw table IN add vlan20 12000,12"
3580 .Dl "ipfw table IN add vlan30 13000,13"
3581 .Dl "ipfw table OUT create type iface valtype skipto"
3582 .Dl "ipfw table OUT add vlan20 22000"
3583 .Dl "ipfw table OUT add vlan30 23000"
3585 .Dl "ipfw add 100 ipfw setfib tablearg ip from any to any recv 'table(IN)' in"
3586 .Dl "ipfw add 200 ipfw skipto tablearg ip from any to any recv 'table(IN)' in"
3587 .Dl "ipfw add 300 ipfw skipto tablearg ip from any to any xmit 'table(OUT)' out"
3589 The following example illustrate usage of flow tables:
3591 .Dl "ipfw table fl create type flow:flow:src-ip,proto,dst-ip,dst-port"
3592 .Dl "ipfw table fl add 2a02:6b8:77::88,tcp,2a02:6b8:77::99,80 11"
3593 .Dl "ipfw table fl add 10.0.0.1,udp,10.0.0.2,53 12"
3595 .Dl "ipfw add 100 allow ip from any to any flow 'table(fl,11)' recv ix0"
3597 To add a set of rules atomically, e.g.\& set 18:
3599 .Dl "ipfw set disable 18"
3600 .Dl "ipfw add NN set 18 ... # repeat as needed"
3601 .Dl "ipfw set enable 18"
3603 To delete a set of rules atomically the command is simply:
3605 .Dl "ipfw delete set 18"
3607 To test a ruleset and disable it and regain control if something goes wrong:
3609 .Dl "ipfw set disable 18"
3610 .Dl "ipfw add NN set 18 ... # repeat as needed"
3611 .Dl "ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18"
3613 Here if everything goes well, you press control-C before the "sleep"
3614 terminates, and your ruleset will be left active.
3615 Otherwise, e.g.\& if
3616 you cannot access your box, the ruleset will be disabled after
3617 the sleep terminates thus restoring the previous situation.
3619 To show rules of the specific set:
3621 .Dl "ipfw set 18 show"
3623 To show rules of the disabled set:
3625 .Dl "ipfw -S set 18 show"
3627 To clear a specific rule counters of the specific set:
3629 .Dl "ipfw set 18 zero NN"
3631 To delete a specific rule of the specific set:
3633 .Dl "ipfw set 18 delete NN"
3634 .Ss NAT, REDIRECT AND LSNAT
3635 First redirect all the traffic to nat instance 123:
3637 .Dl "ipfw add nat 123 all from any to any"
3639 Then to configure nat instance 123 to alias all the outgoing traffic with ip
3640 192.168.0.123, blocking all incoming connections, trying to keep
3641 same ports on both sides, clearing aliasing table on address change
3642 and keeping a log of traffic/link statistics:
3644 .Dl "ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports"
3646 Or to change address of instance 123, aliasing table will be cleared (see
3649 .Dl "ipfw nat 123 config ip 10.0.0.1"
3651 To see configuration of nat instance 123:
3653 .Dl "ipfw nat 123 show config"
3655 To show logs of all the instances in range 111-999:
3657 .Dl "ipfw nat 111-999 show"
3659 To see configurations of all instances:
3661 .Dl "ipfw nat show config"
3663 Or a redirect rule with mixed modes could looks like:
3665 .Dl "ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66"
3666 .Dl " redirect_port tcp 192.168.0.1:80 500"
3667 .Dl " redirect_proto udp 192.168.1.43 192.168.1.1"
3668 .Dl " redirect_addr 192.168.0.10,192.168.0.11"
3669 .Dl " 10.0.0.100 # LSNAT"
3670 .Dl " redirect_port tcp 192.168.0.1:80,192.168.0.10:22"
3673 or it could be split in:
3675 .Dl "ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66"
3676 .Dl "ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500"
3677 .Dl "ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1"
3678 .Dl "ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12"
3680 .Dl "ipfw nat 5 config redirect_port tcp"
3681 .Dl " 192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500"
3702 utility first appeared in
3707 Stateful extensions were introduced in
3710 was introduced in Summer 2002.
3712 .An Ugen J. S. Antsilevich ,
3713 .An Poul-Henning Kamp ,
3719 API based upon code written by
3723 Dummynet has been introduced by Luigi Rizzo in 1997-1998.
3725 Some early work (1999-2000) on the
3727 traffic shaper supported by Akamba Corp.
3729 The ipfw core (ipfw2) has been completely redesigned and
3730 reimplemented by Luigi Rizzo in summer 2002.
3733 options have been added by various developer over the years.
3736 In-kernel NAT support written by
3737 .An Paolo Pisati Aq Mt piso@FreeBSD.org
3738 as part of a Summer of Code 2005 project.
3742 support has been developed by
3743 .An The Centre for Advanced Internet Architectures (CAIA) Aq http://www.caia.swin.edu.au .
3744 The primary developers and maintainers are David Hayes and Jason But.
3745 For further information visit:
3746 .Aq http://www.caia.swin.edu.au/urp/SONATA
3748 Delay profiles have been developed by Alessandro Cerri and
3749 Luigi Rizzo, supported by the
3750 European Commission within Projects Onelab and Onelab2.
3752 The syntax has grown over the years and sometimes it might be confusing.
3753 Unfortunately, backward compatibility prevents cleaning up mistakes
3754 made in the definition of the syntax.
3758 Misconfiguring the firewall can put your computer in an unusable state,
3759 possibly shutting down network services and requiring console access to
3760 regain control of it.
3762 Incoming packet fragments diverted by
3764 are reassembled before delivery to the socket.
3765 The action used on those packet is the one from the
3766 rule which matches the first fragment of the packet.
3768 Packets diverted to userland, and then reinserted by a userland process
3769 may lose various packet attributes.
3770 The packet source interface name
3771 will be preserved if it is shorter than 8 bytes and the userland process
3772 saves and reuses the sockaddr_in
3775 otherwise, it may be lost.
3776 If a packet is reinserted in this manner, later rules may be incorrectly
3777 applied, making the order of
3779 rules in the rule sequence very important.
3781 Dummynet drops all packets with IPv6 link-local addresses.
3787 may not behave as expected.
3788 In particular, incoming SYN packets may
3789 have no uid or gid associated with them since they do not yet belong
3790 to a TCP connection, and the uid/gid associated with a packet may not
3791 be as expected if the associated process calls
3793 or similar system calls.
3795 Rule syntax is subject to the command line environment and some patterns
3796 may need to be escaped with the backslash character
3797 or quoted appropriately.
3799 Due to the architecture of
3801 ipfw nat is not compatible with the TCP segmentation offloading (TSO).
3802 Thus, to reliably nat your network traffic, please disable TSO
3806 ICMP error messages are not implicitly matched by dynamic rules
3807 for the respective conversations.
3808 To avoid failures of network error detection and path MTU discovery,
3809 ICMP error messages may need to be allowed explicitly through static
3816 actions may lead to confusing behaviour if ruleset has mistakes,
3817 and/or interaction with other subsystems (netgraph, dummynet, etc.) is used.
3818 One possible case for this is packet leaving
3820 in subroutine on the input pass, while later on output encountering unpaired
3823 As the call stack is kept intact after input pass, packet will suddenly
3824 return to the rule number used on input pass, not on output one.
3825 Order of processing should be checked carefully to avoid such mistakes.