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
57 .Oo Cm set Ar N Oc Cm table Ar name Cm modify Ar modify-options
59 .Oo Cm set Ar N Oc Cm table Ar name Cm swap Ar name
61 .Oo Cm set Ar N Oc Cm table Ar name Cm add Ar table-key Op Ar value
63 .Oo Cm set Ar N Oc Cm table Ar name Cm add Op Ar table-key Ar value ...
65 .Oo Cm set Ar N Oc Cm table Ar name Cm atomic add Op Ar table-key Ar value ...
67 .Oo Cm set Ar N Oc Cm table Ar name Cm delete Op Ar table-key ...
69 .Oo Cm set Ar N Oc Cm table Ar name Cm lookup Ar addr
71 .Oo Cm set Ar N Oc Cm table Ar name Cm lock
73 .Oo Cm set Ar N Oc Cm table Ar name Cm unlock
75 .Oo Cm set Ar N Oc Cm table
79 .Oo Cm set Ar N Oc Cm table
83 .Oo Cm set Ar N Oc Cm table
87 .Oo Cm set Ar N Oc Cm table
90 .Ss DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)
92 .Brq Cm pipe | queue | sched
98 .Brq Cm pipe | queue | sched
99 .Brq Cm delete | list | show
113 .Ss STATEFUL IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
115 .Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm create Ar create-options
117 .Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm config Ar config-options
119 .Oo Cm set Ar N Oc Cm nat64lsn
124 .Oo Cm set Ar N Oc Cm nat64lsn
128 .Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm stats Op Cm reset
129 .Ss STATELESS IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
131 .Oo Cm set Ar N Oc Cm nat64stl Ar name Cm create Ar create-options
133 .Oo Cm set Ar N Oc Cm nat64stl Ar name Cm config Ar config-options
135 .Oo Cm set Ar N Oc Cm nat64stl
139 .Oo Cm set Ar N Oc Cm nat64stl
143 .Oo Cm set Ar N Oc Cm nat64stl Ar name Cm stats Op Cm reset
144 .Ss XLAT464 CLAT IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
146 .Oo Cm set Ar N Oc Cm nat64clat Ar name Cm create Ar create-options
148 .Oo Cm set Ar N Oc Cm nat64clat Ar name Cm config Ar config-options
150 .Oo Cm set Ar N Oc Cm nat64clat
154 .Oo Cm set Ar N Oc Cm nat64clat
158 .Oo Cm set Ar N Oc Cm nat64clat Ar name Cm stats Op Cm reset
159 .Ss IPv6-to-IPv6 NETWORK PREFIX TRANSLATION
161 .Oo Cm set Ar N Oc Cm nptv6 Ar name Cm create Ar create-options
163 .Oo Cm set Ar N Oc Cm nptv6
167 .Oo Cm set Ar N Oc Cm nptv6
171 .Oo Cm set Ar N Oc Cm nptv6 Ar name Cm stats Op Cm reset
172 .Ss INTERNAL DIAGNOSTICS
179 .Ss LIST OF RULES AND PREPROCESSING
192 utility is the user interface for controlling the
196 traffic shaper/packet scheduler, and the
197 in-kernel NAT services.
199 A firewall configuration, or
203 numbered from 1 to 65535.
204 Packets are passed to the firewall
205 from a number of different places in the protocol stack
206 (depending on the source and destination of the packet,
207 it is possible for the firewall to be
208 invoked multiple times on the same packet).
209 The packet passed to the firewall is compared
210 against each of the rules in the
213 (multiple rules with the same number are permitted, in which case
214 they are processed in order of insertion).
215 When a match is found, the action corresponding to the
216 matching rule is performed.
218 Depending on the action and certain system settings, packets
219 can be reinjected into the firewall at some rule after the
220 matching one for further processing.
222 A ruleset always includes a
224 rule (numbered 65535) which cannot be modified or deleted,
225 and matches all packets.
226 The action associated with the
232 depending on how the kernel is configured.
234 If the ruleset includes one or more rules with the
241 the firewall will have a
243 behaviour, i.e., upon a match it will create
245 i.e., rules that match packets with the same 5-tuple
246 (protocol, source and destination addresses and ports)
247 as the packet which caused their creation.
248 Dynamic rules, which have a limited lifetime, are checked
249 at the first occurrence of a
254 rule, and are typically used to open the firewall on-demand to
255 legitimate traffic only.
262 for all packets (not only these matched by the rule) but
269 .Sx STATEFUL FIREWALL
272 Sections below for more information on the stateful behaviour of
275 All rules (including dynamic ones) have a few associated counters:
276 a packet count, a byte count, a log count and a timestamp
277 indicating the time of the last match.
278 Counters can be displayed or reset with
282 Each rule belongs to one of 32 different
286 commands to atomically manipulate sets, such as enable,
287 disable, swap sets, move all rules in a set to another
288 one, delete all rules in a set.
289 These can be useful to
290 install temporary configurations, or to test them.
293 for more information on
296 Rules can be added with the
298 command; deleted individually or in groups with the
300 command, and globally (except those in set 31) with the
302 command; displayed, optionally with the content of the
308 Finally, counters can be reset with the
314 The following general options are available when invoking
316 .Bl -tag -width indent
318 Show counter values when listing rules.
321 command implies this option.
323 Only show the action and the comment, not the body of a rule.
327 When entering or showing rules, print them in compact form,
328 i.e., omitting the "ip from any to any" string
329 when this does not carry any additional information.
331 When listing, show dynamic rules in addition to static ones.
333 When listing, show only dynamic states.
334 When deleting, delete only dynamic states.
336 Run without prompting for confirmation for commands that can cause problems if misused,
339 If there is no tty associated with the process, this is implied.
342 command with this flag ignores possible errors,
343 i.e., nonexistent rule number.
344 And for batched commands execution continues with the next command.
346 When listing a table (see the
348 section below for more information on lookup tables), format values
350 By default, values are shown as integers.
352 Only check syntax of the command strings, without actually passing
355 Try to resolve addresses and service names in output.
357 Be quiet when executing the
367 This is useful when updating rulesets by executing multiple
371 .Ql sh\ /etc/rc.firewall ) ,
372 or by processing a file with many
374 rules across a remote login session.
375 It also stops a table add or delete
376 from failing if the entry already exists or is not present.
378 The reason why this option may be important is that
379 for some of these actions,
381 may print a message; if the action results in blocking the
382 traffic to the remote client,
383 the remote login session will be closed
384 and the rest of the ruleset will not be processed.
385 Access to the console would then be required to recover.
387 When listing rules, show the
389 each rule belongs to.
390 If this flag is not specified, disabled rules will not be
393 When listing pipes, sort according to one of the four
394 counters (total or current packets or bytes).
396 When listing, show last match timestamp converted with
399 When listing, show last match timestamp as seconds from the epoch.
400 This form can be more convenient for postprocessing by scripts.
402 .Ss LIST OF RULES AND PREPROCESSING
403 To ease configuration, rules can be put into a file which is
406 as shown in the last synopsis line.
410 The file will be read line by line and applied as arguments to the
414 Optionally, a preprocessor can be specified using
418 is to be piped through.
419 Useful preprocessors include
425 does not start with a slash
427 as its first character, the usual
429 name search is performed.
430 Care should be taken with this in environments where not all
431 file systems are mounted (yet) by the time
433 is being run (e.g.\& when they are mounted over NFS).
436 has been specified, any additional arguments are passed on to the preprocessor
438 This allows for flexible configuration files (like conditionalizing
439 them on the local hostname) and the use of macros to centralize
440 frequently required arguments like IP addresses.
441 .Ss TRAFFIC SHAPER CONFIGURATION
447 commands are used to configure the traffic shaper and packet scheduler.
449 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
450 Section below for details.
452 If the world and the kernel get out of sync the
454 ABI may break, preventing you from being able to add any rules.
455 This can adversely affect the booting process.
460 to temporarily disable the firewall to regain access to the network,
461 allowing you to fix the problem.
463 A packet is checked against the active ruleset in multiple places
464 in the protocol stack, under control of several sysctl variables.
465 These places and variables are shown below, and it is important to
466 have this picture in mind in order to design a correct ruleset.
467 .Bd -literal -offset indent
470 +----------->-----------+
472 [ip(6)_input] [ip(6)_output] net.inet(6).ip(6).fw.enable=1
475 [ether_demux] [ether_output_frame] net.link.ether.ipfw=1
477 +-->--[bdg_forward]-->--+ net.link.bridge.ipfw=1
483 times the same packet goes through the firewall can
484 vary between 0 and 4 depending on packet source and
485 destination, and system configuration.
487 Note that as packets flow through the stack, headers can be
488 stripped or added to it, and so they may or may not be available
490 E.g., incoming packets will include the MAC header when
494 but the same packets will have the MAC header stripped off when
501 Also note that each packet is always checked against the complete ruleset,
502 irrespective of the place where the check occurs, or the source of the packet.
503 If a rule contains some match patterns or actions which are not valid
504 for the place of invocation (e.g.\& trying to match a MAC header within
508 the match pattern will not match, but a
510 operator in front of such patterns
514 match on those packets.
515 It is thus the responsibility of
516 the programmer, if necessary, to write a suitable ruleset to
517 differentiate among the possible places.
519 rules can be useful here, as an example:
520 .Bd -literal -offset indent
521 # packets from ether_demux or bdg_forward
522 ipfw add 10 skipto 1000 all from any to any layer2 in
523 # packets from ip_input
524 ipfw add 10 skipto 2000 all from any to any not layer2 in
525 # packets from ip_output
526 ipfw add 10 skipto 3000 all from any to any not layer2 out
527 # packets from ether_output_frame
528 ipfw add 10 skipto 4000 all from any to any layer2 out
531 (yes, at the moment there is no way to differentiate between
532 ether_demux and bdg_forward).
534 Also note that only actions
543 frames and all other actions act as if they were
546 Full set of actions is supported for IP packets without
551 action does not divert
555 In general, each keyword or argument must be provided as
556 a separate command line argument, with no leading or trailing
558 Keywords are case-sensitive, whereas arguments may
559 or may not be case-sensitive depending on their nature
560 (e.g.\& uid's are, hostnames are not).
562 Some arguments (e.g., port or address lists) are comma-separated
564 In this case, spaces after commas ',' are allowed to make
565 the line more readable.
566 You can also put the entire
567 command (including flags) into a single argument.
568 E.g., the following forms are equivalent:
569 .Bd -literal -offset indent
570 ipfw -q add deny src-ip 10.0.0.0/24,127.0.0.1/8
571 ipfw -q add deny src-ip 10.0.0.0/24, 127.0.0.1/8
572 ipfw "-q add deny src-ip 10.0.0.0/24, 127.0.0.1/8"
575 The format of firewall rules is the following:
576 .Bd -ragged -offset indent
579 .Op Cm set Ar set_number
580 .Op Cm prob Ar match_probability
582 .Op Cm log Op Cm logamount Ar number
592 where the body of the rule specifies which information is used
593 for filtering packets, among the following:
595 .Bl -tag -width "Source and dest. addresses and ports" -offset XXX -compact
596 .It Layer-2 header fields
598 .It IPv4 and IPv6 Protocol
599 SCTP, TCP, UDP, ICMP, etc.
600 .It Source and dest. addresses and ports
604 .It Transmit and receive interface
606 .It Misc. IP header fields
607 Version, type of service, datagram length, identification,
611 .It IPv6 Extension headers
612 Fragmentation, Hop-by-Hop options,
613 Routing Headers, Source routing rthdr0, Mobile IPv6 rthdr2, IPSec options.
615 .It Misc. TCP header fields
616 TCP flags (SYN, FIN, ACK, RST, etc.),
617 sequence number, acknowledgment number,
625 When the packet can be associated with a local socket.
627 Whether a packet came from a divert socket (e.g.,
629 .It Fib annotation state
630 Whether a packet has been tagged for using a specific FIB (routing table)
631 in future forwarding decisions.
634 Note that some of the above information, e.g.\& source MAC or IP addresses and
635 TCP/UDP ports, can be easily spoofed, so filtering on those fields
636 alone might not guarantee the desired results.
637 .Bl -tag -width indent
639 Each rule is associated with a
641 in the range 1..65535, with the latter reserved for the
644 Rules are checked sequentially by rule number.
645 Multiple rules can have the same number, in which case they are
646 checked (and listed) according to the order in which they have
648 If a rule is entered without specifying a number, the kernel will
649 assign one in such a way that the rule becomes the last one
653 Automatic rule numbers are assigned by incrementing the last
654 non-default rule number by the value of the sysctl variable
655 .Ar net.inet.ip.fw.autoinc_step
656 which defaults to 100.
657 If this is not possible (e.g.\& because we would go beyond the
658 maximum allowed rule number), the number of the last
659 non-default value is used instead.
660 .It Cm set Ar set_number
661 Each rule is associated with a
664 Sets can be individually disabled and enabled, so this parameter
665 is of fundamental importance for atomic ruleset manipulation.
666 It can be also used to simplify deletion of groups of rules.
667 If a rule is entered without specifying a set number,
670 Set 31 is special in that it cannot be disabled,
671 and rules in set 31 are not deleted by the
673 command (but you can delete them with the
674 .Nm ipfw delete set 31
676 Set 31 is also used for the
679 .It Cm prob Ar match_probability
680 A match is only declared with the specified probability
681 (floating point number between 0 and 1).
682 This can be useful for a number of applications such as
683 random packet drop or
686 to simulate the effect of multiple paths leading to out-of-order
689 Note: this condition is checked before any other condition, including
696 .It Cm log Op Cm logamount Ar number
697 Packets matching a rule with the
699 keyword will be made available for logging in two ways:
700 if the sysctl variable
701 .Va net.inet.ip.fw.verbose
702 is set to 0 (default), one can use
707 This pseudo interface can be created manually after a system
708 boot by using the following command:
709 .Bd -literal -offset indent
710 # ifconfig ipfw0 create
713 Or, automatically at boot time by adding the following
717 .Bd -literal -offset indent
721 There is zero overhead when no
723 is attached to the pseudo interface.
726 .Va net.inet.ip.fw.verbose
727 is set to 1, packets will be logged to
731 facility up to a maximum of
736 is specified, the limit is taken from the sysctl variable
737 .Va net.inet.ip.fw.verbose_limit .
738 In both cases, a value of 0 means unlimited logging.
740 Once the limit is reached, logging can be re-enabled by
741 clearing the logging counter or the packet counter for that entry, see the
745 Note: logging is done after all other packet matching conditions
746 have been successfully verified, and before performing the final
747 action (accept, deny, etc.) on the packet.
749 When a packet matches a rule with the
751 keyword, the numeric tag for the given
753 in the range 1..65534 will be attached to the packet.
754 The tag acts as an internal marker (it is not sent out over
755 the wire) that can be used to identify these packets later on.
756 This can be used, for example, to provide trust between interfaces
757 and to start doing policy-based filtering.
758 A packet can have multiple tags at the same time.
759 Tags are "sticky", meaning once a tag is applied to a packet by a
760 matching rule it exists until explicit removal.
761 Tags are kept with the packet everywhere within the kernel, but are
762 lost when packet leaves the kernel, for example, on transmitting
763 packet out to the network or sending packet to a
767 To check for previously applied tags, use the
770 To delete previously applied tag, use the
774 Note: since tags are kept with the packet everywhere in kernelspace,
775 they can be set and unset anywhere in the kernel network subsystem
778 facility), not only by means of the
784 For example, there can be a specialized
786 node doing traffic analyzing and tagging for later inspecting
788 .It Cm untag Ar number
789 When a packet matches a rule with the
791 keyword, the tag with the number
793 is searched among the tags attached to this packet and,
794 if found, removed from it.
795 Other tags bound to packet, if present, are left untouched.
797 When a packet matches a rule with the
799 keyword, the ALTQ identifier for the given
804 Note that this ALTQ tag is only meaningful for packets going "out" of IPFW,
805 and not being rejected or going to divert sockets.
806 Note that if there is insufficient memory at the time the packet is
807 processed, it will not be tagged, so it is wise to make your ALTQ
808 "default" queue policy account for this.
811 rules match a single packet, only the first one adds the ALTQ classification
813 In doing so, traffic may be shaped by using
814 .Cm count Cm altq Ar queue
815 rules for classification early in the ruleset, then later applying
816 the filtering decision.
821 rules may come later and provide the actual filtering decisions in
822 addition to the fallback ALTQ tag.
826 to set up the queues before IPFW will be able to look them up by name,
827 and if the ALTQ disciplines are rearranged, the rules in containing the
828 queue identifiers in the kernel will likely have gone stale and need
830 Stale queue identifiers will probably result in misclassification.
832 All system ALTQ processing can be turned on or off via
837 .Cm disable Ar altq .
839 .Va net.inet.ip.fw.one_pass
840 is irrelevant to ALTQ traffic shaping, as the actual rule action is followed
841 always after adding an ALTQ tag.
844 A rule can be associated with one of the following actions, which
845 will be executed when the packet matches the body of the rule.
846 .Bl -tag -width indent
847 .It Cm allow | accept | pass | permit
848 Allow packets that match rule.
849 The search terminates.
850 .It Cm check-state Op Ar :flowname | Cm :any
851 Checks the packet against the dynamic ruleset.
852 If a match is found, execute the action associated with
853 the rule which generated this dynamic rule, otherwise
854 move to the next rule.
857 rules do not have a body.
860 rule is found, the dynamic ruleset is checked at the first
867 is symbolic name assigned to dynamic rule by
872 can be used to ignore states flowname when matching.
875 keyword is special name used for compatibility with old rulesets.
877 Update counters for all packets that match rule.
878 The search continues with the next rule.
880 Discard packets that match this rule.
881 The search terminates.
882 .It Cm divert Ar port
883 Divert packets that match this rule to the
887 The search terminates.
888 .It Cm fwd | forward Ar ipaddr | tablearg Ns Op , Ns Ar port
889 Change the next-hop on matching packets to
891 which can be an IP address or a host name.
892 The next hop can also be supplied by the last table
893 looked up for the packet by using the
895 keyword instead of an explicit address.
896 The search terminates if this rule matches.
900 is a local address, then matching packets will be forwarded to
902 (or the port number in the packet if one is not specified in the rule)
903 on the local machine.
907 is not a local address, then the port number
908 (if specified) is ignored, and the packet will be
909 forwarded to the remote address, using the route as found in
910 the local routing table for that IP.
914 rule will not match layer-2 packets (those received
915 on ether_input, ether_output, or bridged).
919 action does not change the contents of the packet at all.
920 In particular, the destination address remains unmodified, so
921 packets forwarded to another system will usually be rejected by that system
922 unless there is a matching rule on that system to capture them.
923 For packets forwarded locally,
924 the local address of the socket will be
925 set to the original destination address of the packet.
928 entry look rather weird but is intended for
929 use with transparent proxy servers.
930 .It Cm nat Ar nat_nr | global | tablearg
933 (for network address translation, address redirect, etc.):
935 .Sx NETWORK ADDRESS TRANSLATION (NAT)
936 Section for further information.
937 .It Cm nat64lsn Ar name
938 Pass packet to a stateful NAT64 instance (for IPv6/IPv4 network address and
939 protocol translation): see the
940 .Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
941 Section for further information.
942 .It Cm nat64stl Ar name
943 Pass packet to a stateless NAT64 instance (for IPv6/IPv4 network address and
944 protocol translation): see the
945 .Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
946 Section for further information.
947 .It Cm nat64clat Ar name
948 Pass packet to a CLAT NAT64 instance (for client-side IPv6/IPv4 network address and
949 protocol translation): see the
950 .Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
951 Section for further information.
953 Pass packet to a NPTv6 instance (for IPv6-to-IPv6 network prefix translation):
955 .Sx IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
956 Section for further information.
957 .It Cm pipe Ar pipe_nr
961 (for bandwidth limitation, delay, etc.).
963 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
964 Section for further information.
965 The search terminates; however, on exit from the pipe and if
969 .Va net.inet.ip.fw.one_pass
970 is not set, the packet is passed again to the firewall code
971 starting from the next rule.
972 .It Cm queue Ar queue_nr
976 (for bandwidth limitation using WF2Q+).
982 Discard packets that match this rule, and if the
983 packet is a TCP packet, try to send a TCP reset (RST) notice.
984 The search terminates.
986 Discard packets that match this rule, and if the
987 packet is a TCP packet, try to send a TCP reset (RST) notice.
988 The search terminates.
989 .It Cm skipto Ar number | tablearg
990 Skip all subsequent rules numbered less than
992 The search continues with the first rule numbered
995 It is possible to use the
997 keyword with a skipto for a
1000 Skipto may work either in O(log(N)) or in O(1) depending
1001 on amount of memory and/or sysctl variables.
1003 .Sx SYSCTL VARIABLES
1004 section for more details.
1005 .It Cm call Ar number | tablearg
1006 The current rule number is saved in the internal stack and
1007 ruleset processing continues with the first rule numbered
1010 If later a rule with the
1012 action is encountered, the processing returns to the first rule
1015 rule plus one or higher
1016 (the same behaviour as with packets returning from
1021 This could be used to make somewhat like an assembly language
1023 calls to rules with common checks for different interfaces, etc.
1025 Rule with any number could be called, not just forward jumps as with
1027 So, to prevent endless loops in case of mistakes, both
1031 actions don't do any jumps and simply go to the next rule if memory
1032 cannot be allocated or stack overflowed/underflowed.
1034 Internally stack for rule numbers is implemented using
1036 facility and currently has size of 16 entries.
1037 As mbuf tags are lost when packet leaves the kernel,
1039 should not be used in subroutines to avoid endless loops
1040 and other undesired effects.
1042 Takes rule number saved to internal stack by the last
1044 action and returns ruleset processing to the first rule
1045 with number greater than number of corresponding
1048 See description of the
1050 action for more details.
1056 and thus are unconditional, but
1058 command-line utility currently requires every action except
1061 While it is sometimes useful to return only on some packets,
1062 usually you want to print just
1065 A workaround for this is to use new syntax and
1068 .Bd -literal -offset indent
1069 # Add a rule without actual body
1070 ipfw add 2999 return via any
1072 # List rules without "from any to any" part
1076 This cosmetic annoyance may be fixed in future releases.
1078 Send a copy of packets matching this rule to the
1080 socket bound to port
1082 The search continues with the next rule.
1083 .It Cm unreach Ar code
1084 Discard packets that match this rule, and try to send an ICMP
1085 unreachable notice with code
1089 is a number from 0 to 255, or one of these aliases:
1090 .Cm net , host , protocol , port ,
1091 .Cm needfrag , srcfail , net-unknown , host-unknown ,
1092 .Cm isolated , net-prohib , host-prohib , tosnet ,
1093 .Cm toshost , filter-prohib , host-precedence
1095 .Cm precedence-cutoff .
1096 The search terminates.
1097 .It Cm unreach6 Ar code
1098 Discard packets that match this rule, and try to send an ICMPv6
1099 unreachable notice with code
1103 is a number from 0, 1, 3 or 4, or one of these aliases:
1104 .Cm no-route, admin-prohib, address
1107 The search terminates.
1108 .It Cm netgraph Ar cookie
1109 Divert packet into netgraph with given
1111 The search terminates.
1112 If packet is later returned from netgraph it is either
1113 accepted or continues with the next rule, depending on
1114 .Va net.inet.ip.fw.one_pass
1116 .It Cm ngtee Ar cookie
1117 A copy of packet is diverted into netgraph, original
1118 packet continues with the next rule.
1121 for more information on
1126 .It Cm setfib Ar fibnum | tablearg
1127 The packet is tagged so as to use the FIB (routing table)
1129 in any subsequent forwarding decisions.
1130 In the current implementation, this is limited to the values 0 through 15, see
1132 Processing continues at the next rule.
1133 It is possible to use the
1135 keyword with setfib.
1136 If the tablearg value is not within the compiled range of fibs,
1137 the packet's fib is set to 0.
1138 .It Cm setdscp Ar DSCP | number | tablearg
1139 Set specified DiffServ codepoint for an IPv4/IPv6 packet.
1140 Processing continues at the next rule.
1141 Supported values are:
1187 Additionally, DSCP value can be specified by number (0..63).
1188 It is also possible to use the
1190 keyword with setdscp.
1191 If the tablearg value is not within the 0..63 range, lower 6 bits of supplied
1193 .It Cm tcp-setmss Ar mss
1194 Set the Maximum Segment Size (MSS) in the TCP segment to value
1198 should be loaded or kernel should have
1199 .Cm options IPFIREWALL_PMOD
1200 to be able use this action.
1201 This command does not change a packet if original MSS value is lower than
1203 Both TCP over IPv4 and over IPv6 are supported.
1204 Regardless of matched a packet or not by the
1206 rule, the search continues with the next rule.
1208 Queue and reassemble IPv4 fragments.
1209 If the packet is not fragmented, counters are updated and
1210 processing continues with the next rule.
1211 If the packet is the last logical fragment, the packet is reassembled and, if
1212 .Va net.inet.ip.fw.one_pass
1213 is set to 0, processing continues with the next rule.
1214 Otherwise, the packet is allowed to pass and the search terminates.
1215 If the packet is a fragment in the middle of a logical group of fragments,
1217 processing stops immediately.
1219 Fragment handling can be tuned via
1220 .Va net.inet.ip.maxfragpackets
1222 .Va net.inet.ip.maxfragsperpacket
1223 which limit, respectively, the maximum number of processable
1224 fragments (default: 800) and
1225 the maximum number of fragments per packet (default: 16).
1227 NOTA BENE: since fragments do not contain port numbers,
1228 they should be avoided with the
1231 Alternatively, direction-based (like
1235 ) and source-based (like
1237 ) match patterns can be used to select fragments.
1239 Usually a simple rule like:
1240 .Bd -literal -offset indent
1241 # reassemble incoming fragments
1242 ipfw add reass all from any to any in
1245 is all you need at the beginning of your ruleset.
1247 Discard packets that match this rule, and if the packet is an SCTP packet,
1248 try to send an SCTP packet containing an ABORT chunk.
1249 The search terminates.
1251 Discard packets that match this rule, and if the packet is an SCTP packet,
1252 try to send an SCTP packet containing an ABORT chunk.
1253 The search terminates.
1256 The body of a rule contains zero or more patterns (such as
1257 specific source and destination addresses or ports,
1258 protocol options, incoming or outgoing interfaces, etc.)
1259 that the packet must match in order to be recognised.
1260 In general, the patterns are connected by (implicit)
1262 operators -- i.e., all must match in order for the
1264 Individual patterns can be prefixed by the
1266 operator to reverse the result of the match, as in
1268 .Dl "ipfw add 100 allow ip from not 1.2.3.4 to any"
1270 Additionally, sets of alternative match patterns
1272 can be constructed by putting the patterns in
1273 lists enclosed between parentheses ( ) or braces { }, and
1276 operator as follows:
1278 .Dl "ipfw add 100 allow ip from { x or not y or z } to any"
1280 Only one level of parentheses is allowed.
1281 Beware that most shells have special meanings for parentheses
1282 or braces, so it is advisable to put a backslash \\ in front of them
1283 to prevent such interpretations.
1285 The body of a rule must in general include a source and destination
1289 can be used in various places to specify that the content of
1290 a required field is irrelevant.
1292 The rule body has the following format:
1293 .Bd -ragged -offset indent
1294 .Op Ar proto Cm from Ar src Cm to Ar dst
1298 The first part (proto from src to dst) is for backward
1299 compatibility with earlier versions of
1303 any match pattern (including MAC headers, IP protocols,
1304 addresses and ports) can be specified in the
1308 Rule fields have the following meaning:
1309 .Bl -tag -width indent
1310 .It Ar proto : protocol | Cm { Ar protocol Cm or ... }
1311 .It Ar protocol : Oo Cm not Oc Ar protocol-name | protocol-number
1312 An IP protocol specified by number or name
1313 (for a complete list see
1314 .Pa /etc/protocols ) ,
1315 or one of the following keywords:
1316 .Bl -tag -width indent
1318 Matches IPv4 packets.
1320 Matches IPv6 packets.
1329 option will be treated as inner protocol.
1337 .Cm { Ar protocol Cm or ... }
1340 is provided for convenience only but its use is deprecated.
1341 .It Ar src No and Ar dst : Bro Cm addr | Cm { Ar addr Cm or ... } Brc Op Oo Cm not Oc Ar ports
1342 An address (or a list, see below)
1343 optionally followed by
1349 with multiple addresses) is provided for convenience only and
1350 its use is discouraged.
1351 .It Ar addr : Oo Cm not Oc Bro
1352 .Cm any | me | me6 |
1353 .Cm table Ns Pq Ar name Ns Op , Ns Ar value
1354 .Ar | addr-list | addr-set
1356 .Bl -tag -width indent
1358 Matches any IP address.
1360 Matches any IP address configured on an interface in the system.
1362 Matches any IPv6 address configured on an interface in the system.
1363 The address list is evaluated at the time the packet is
1365 .It Cm table Ns Pq Ar name Ns Op , Ns Ar value
1366 Matches any IPv4 or IPv6 address for which an entry exists in the lookup table
1368 If an optional 32-bit unsigned
1370 is also specified, an entry will match only if it has this value.
1373 section below for more information on lookup tables.
1375 .It Ar addr-list : ip-addr Ns Op Ns , Ns Ar addr-list
1377 A host or subnet address specified in one of the following ways:
1378 .Bl -tag -width indent
1379 .It Ar numeric-ip | hostname
1380 Matches a single IPv4 address, specified as dotted-quad or a hostname.
1381 Hostnames are resolved at the time the rule is added to the firewall list.
1382 .It Ar addr Ns / Ns Ar masklen
1383 Matches all addresses with base
1385 (specified as an IP address, a network number, or a hostname)
1389 As an example, 1.2.3.4/25 or 1.2.3.0/25 will match
1390 all IP numbers from 1.2.3.0 to 1.2.3.127 .
1391 .It Ar addr Ns : Ns Ar mask
1392 Matches all addresses with base
1394 (specified as an IP address, a network number, or a hostname)
1397 specified as a dotted quad.
1398 As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match
1400 This form is advised only for non-contiguous
1402 It is better to resort to the
1403 .Ar addr Ns / Ns Ar masklen
1404 format for contiguous masks, which is more compact and less
1407 .It Ar addr-set : addr Ns Oo Ns / Ns Ar masklen Oc Ns Cm { Ns Ar list Ns Cm }
1408 .It Ar list : Bro Ar num | num-num Brc Ns Op Ns , Ns Ar list
1409 Matches all addresses with base address
1411 (specified as an IP address, a network number, or a hostname)
1412 and whose last byte is in the list between braces { } .
1413 Note that there must be no spaces between braces and
1414 numbers (spaces after commas are allowed).
1415 Elements of the list can be specified as single entries
1419 field is used to limit the size of the set of addresses,
1420 and can have any value between 24 and 32.
1422 it will be assumed as 24.
1424 This format is particularly useful to handle sparse address sets
1425 within a single rule.
1426 Because the matching occurs using a
1427 bitmask, it takes constant time and dramatically reduces
1428 the complexity of rulesets.
1430 As an example, an address specified as 1.2.3.4/24{128,35-55,89}
1431 or 1.2.3.0/24{128,35-55,89}
1432 will match the following IP addresses:
1434 1.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .
1435 .It Ar addr6-list : ip6-addr Ns Op Ns , Ns Ar addr6-list
1437 A host or subnet specified one of the following ways:
1438 .Bl -tag -width indent
1439 .It Ar numeric-ip | hostname
1440 Matches a single IPv6 address as allowed by
1443 Hostnames are resolved at the time the rule is added to the firewall
1445 .It Ar addr Ns / Ns Ar masklen
1446 Matches all IPv6 addresses with base
1448 (specified as allowed by
1454 .It Ar addr Ns / Ns Ar mask
1455 Matches all IPv6 addresses with base
1457 (specified as allowed by
1462 specified as allowed by
1464 As an example, fe::640:0:0/ffff::ffff:ffff:0:0 will match
1466 This form is advised only for non-contiguous
1468 It is better to resort to the
1469 .Ar addr Ns / Ns Ar masklen
1470 format for contiguous masks, which is more compact and less
1474 No support for sets of IPv6 addresses is provided because IPv6 addresses
1475 are typically random past the initial prefix.
1476 .It Ar ports : Bro Ar port | port Ns \&- Ns Ar port Ns Brc Ns Op , Ns Ar ports
1477 For protocols which support port numbers (such as SCTP, TCP and UDP), optional
1479 may be specified as one or more ports or port ranges, separated
1480 by commas but no spaces, and an optional
1485 notation specifies a range of ports (including boundaries).
1489 may be used instead of numeric port values.
1490 The length of the port list is limited to 30 ports or ranges,
1491 though one can specify larger ranges by using an
1495 section of the rule.
1499 can be used to escape the dash
1501 character in a service name (from a shell, the backslash must be
1502 typed twice to avoid the shell itself interpreting it as an escape
1505 .Dl "ipfw add count tcp from any ftp\e\e-data-ftp to any"
1507 Fragmented packets which have a non-zero offset (i.e., not the first
1508 fragment) will never match a rule which has one or more port
1512 option for details on matching fragmented packets.
1514 .Ss RULE OPTIONS (MATCH PATTERNS)
1515 Additional match patterns can be used within
1517 Zero or more of these so-called
1519 can be present in a rule, optionally prefixed by the
1521 operand, and possibly grouped into
1524 The following match patterns can be used (listed in alphabetical order):
1525 .Bl -tag -width indent
1526 .It Cm // this is a comment .
1527 Inserts the specified text as a comment in the rule.
1528 Everything following // is considered as a comment and stored in the rule.
1529 You can have comment-only rules, which are listed as having a
1531 action followed by the comment.
1535 .It Cm defer-immediate-action | defer-action
1536 A rule with this option will not perform normal action
1538 This option is intended to be used with
1542 as the dynamic rule, created but ignored on match, will work
1547 .Cm defer-immediate-action
1548 create a dynamic rule and continue with the next rule without actually
1549 performing the action part of this rule.
1550 When the rule is later activated via the state table, the action is
1553 Matches only packets generated by a divert socket.
1554 .It Cm diverted-loopback
1555 Matches only packets coming from a divert socket back into the IP stack
1557 .It Cm diverted-output
1558 Matches only packets going from a divert socket back outward to the IP
1559 stack output for delivery.
1560 .It Cm dst-ip Ar ip-address
1561 Matches IPv4 packets whose destination IP is one of the address(es)
1562 specified as argument.
1563 .It Bro Cm dst-ip6 | dst-ipv6 Brc Ar ip6-address
1564 Matches IPv6 packets whose destination IP is one of the address(es)
1565 specified as argument.
1566 .It Cm dst-port Ar ports
1567 Matches IP packets whose destination port is one of the port(s)
1568 specified as argument.
1570 Matches TCP packets that have the RST or ACK bits set.
1571 .It Cm ext6hdr Ar header
1572 Matches IPv6 packets containing the extended header given by
1574 Supported headers are:
1580 any type of Routing Header
1582 Source routing Routing Header Type 0
1584 Mobile IPv6 Routing Header Type 2
1588 IPSec authentication headers
1590 and IPsec encapsulated security payload headers
1592 .It Cm fib Ar fibnum
1593 Matches a packet that has been tagged to use
1594 the given FIB (routing table) number.
1595 .It Cm flow Ar table Ns Pq Ar name Ns Op , Ns Ar value
1596 Search for the flow entry in lookup table
1598 If not found, the match fails.
1599 Otherwise, the match succeeds and
1601 is set to the value extracted from the table.
1603 This option can be useful to quickly dispatch traffic based on
1604 certain packet fields.
1607 section below for more information on lookup tables.
1608 .It Cm flow-id Ar labels
1609 Matches IPv6 packets containing any of the flow labels given in
1612 is a comma separated list of numeric flow labels.
1614 Matches IPv4 packets whose
1616 field contains the comma separated list of IPv4 fragmentation
1617 options specified in
1619 The recognized options are:
1621 .Pq Dv don't fragment ,
1623 .Pq Dv more fragments ,
1625 .Pq Dv reserved fragment bit
1627 .Pq Dv non-zero fragment offset .
1628 The absence of a particular options may be denoted
1632 Empty list of options defaults to matching on non-zero fragment offset.
1633 Such rule would match all not the first fragment datagrams,
1635 This is a backward compatibility with older rulesets.
1637 Matches all TCP or UDP packets sent by or received for a
1641 may be specified by name or number.
1643 Matches all TCP or UDP packets sent by or received for the
1644 jail whose ID or name is
1646 .It Cm icmptypes Ar types
1647 Matches ICMP packets whose ICMP type is in the list
1649 The list may be specified as any combination of
1650 individual types (numeric) separated by commas.
1651 .Em Ranges are not allowed .
1652 The supported ICMP types are:
1656 destination unreachable
1664 router advertisement
1668 time-to-live exceeded
1680 address mask request
1682 and address mask reply
1684 .It Cm icmp6types Ar types
1685 Matches ICMP6 packets whose ICMP6 type is in the list of
1687 The list may be specified as any combination of
1688 individual types (numeric) separated by commas.
1689 .Em Ranges are not allowed .
1691 Matches incoming or outgoing packets, respectively.
1695 are mutually exclusive (in fact,
1699 .It Cm ipid Ar id-list
1700 Matches IPv4 packets whose
1702 field has value included in
1704 which is either a single value or a list of values or ranges
1705 specified in the same way as
1707 .It Cm iplen Ar len-list
1708 Matches IP packets whose total length, including header and data, is
1711 which is either a single value or a list of values or ranges
1712 specified in the same way as
1714 .It Cm ipoptions Ar spec
1715 Matches packets whose IPv4 header contains the comma separated list of
1716 options specified in
1718 The supported IP options are:
1721 (strict source route),
1723 (loose source route),
1725 (record packet route) and
1728 The absence of a particular option may be denoted
1731 .It Cm ipprecedence Ar precedence
1732 Matches IPv4 packets whose precedence field is equal to
1735 Matches packets that have IPSEC history associated with them
1736 (i.e., the packet comes encapsulated in IPSEC, the kernel
1737 has IPSEC support, and can correctly decapsulate it).
1739 Note that specifying
1741 is different from specifying
1743 as the latter will only look at the specific IP protocol field,
1744 irrespective of IPSEC kernel support and the validity of the IPSEC data.
1746 Further note that this flag is silently ignored in kernels without
1748 It does not affect rule processing when given and the
1749 rules are handled as if with no
1752 .It Cm iptos Ar spec
1753 Matches IPv4 packets whose
1755 field contains the comma separated list of
1756 service types specified in
1758 The supported IP types of service are:
1761 .Pq Dv IPTOS_LOWDELAY ,
1763 .Pq Dv IPTOS_THROUGHPUT ,
1765 .Pq Dv IPTOS_RELIABILITY ,
1767 .Pq Dv IPTOS_MINCOST ,
1769 .Pq Dv IPTOS_ECN_CE .
1770 The absence of a particular type may be denoted
1773 .It Cm dscp spec Ns Op , Ns Ar spec
1774 Matches IPv4/IPv6 packets whose
1776 field value is contained in
1779 Multiple values can be specified via
1780 the comma separated list.
1781 Value can be one of keywords used in
1783 action or exact number.
1784 .It Cm ipttl Ar ttl-list
1785 Matches IPv4 packets whose time to live is included in
1787 which is either a single value or a list of values or ranges
1788 specified in the same way as
1790 .It Cm ipversion Ar ver
1791 Matches IP packets whose IP version field is
1793 .It Cm keep-state Op Ar :flowname
1794 Upon a match, the firewall will create a dynamic rule, whose
1795 default behaviour is to match bidirectional traffic between
1796 source and destination IP/port using the same protocol.
1797 The rule has a limited lifetime (controlled by a set of
1799 variables), and the lifetime is refreshed every time a matching
1803 is used to assign additional to addresses, ports and protocol parameter
1805 It can be used for more accurate matching by
1810 keyword is special name used for compatibility with old rulesets.
1812 Matches only layer2 packets, i.e., those passed to
1817 .Fn ether_output_frame .
1818 .It Cm limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N Op Ar :flowname
1819 The firewall will only allow
1821 connections with the same
1822 set of parameters as specified in the rule.
1824 of source and destination addresses and ports can be
1826 .It Cm lookup Bro Cm dst-ip | dst-port | src-ip | src-port | uid | jail Brc Ar name
1827 Search an entry in lookup table
1829 that matches the field specified as argument.
1830 If not found, the match fails.
1831 Otherwise, the match succeeds and
1833 is set to the value extracted from the table.
1835 This option can be useful to quickly dispatch traffic based on
1836 certain packet fields.
1839 section below for more information on lookup tables.
1840 .It Cm { MAC | mac } Ar dst-mac src-mac
1841 Match packets with a given
1845 addresses, specified as the
1847 keyword (matching any MAC address), or six groups of hex digits
1848 separated by colons,
1849 and optionally followed by a mask indicating the significant bits.
1850 The mask may be specified using either of the following methods:
1851 .Bl -enum -width indent
1855 followed by the number of significant bits.
1856 For example, an address with 33 significant bits could be specified as:
1858 .Dl "MAC 10:20:30:40:50:60/33 any"
1862 followed by a bitmask specified as six groups of hex digits separated
1864 For example, an address in which the last 16 bits are significant could
1867 .Dl "MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any"
1869 Note that the ampersand character has a special meaning in many shells
1870 and should generally be escaped.
1872 Note that the order of MAC addresses (destination first,
1874 the same as on the wire, but the opposite of the one used for
1876 .It Cm mac-type Ar mac-type
1877 Matches packets whose Ethernet Type field
1878 corresponds to one of those specified as argument.
1880 is specified in the same way as
1882 (i.e., one or more comma-separated single values or ranges).
1883 You can use symbolic names for known values such as
1884 .Em vlan , ipv4, ipv6 .
1885 Values can be entered as decimal or hexadecimal (if prefixed by 0x),
1886 and they are always printed as hexadecimal (unless the
1888 option is used, in which case symbolic resolution will be attempted).
1889 .It Cm proto Ar protocol
1890 Matches packets with the corresponding IP protocol.
1892 Upon a match, the firewall will create a dynamic rule as if
1895 However, this option doesn't imply an implicit
1899 .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
1900 Matches packets received, transmitted or going through,
1901 respectively, the interface specified by exact name
1905 by IP address, or through some interface.
1908 may be used to match interface by its kernel ifindex.
1911 section below for more information on lookup tables.
1915 keyword causes the interface to always be checked.
1922 then only the receive or transmit interface (respectively)
1924 By specifying both, it is possible to match packets based on
1925 both receive and transmit interface, e.g.:
1927 .Dl "ipfw add deny ip from any to any out recv ed0 xmit ed1"
1931 interface can be tested on either incoming or outgoing packets,
1934 interface can only be tested on outgoing packets.
1939 is invalid) whenever
1943 A packet might not have a receive or transmit interface: packets
1944 originating from the local host have no receive interface,
1945 while packets destined for the local host have no transmit
1947 .It Cm set-limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N
1950 but does not have an implicit
1954 Matches TCP packets that have the SYN bit set but no ACK bit.
1955 This is the short form of
1956 .Dq Li tcpflags\ syn,!ack .
1958 Matches packets that are associated to a local socket and
1959 for which the SO_USER_COOKIE socket option has been set
1960 to a non-zero value.
1961 As a side effect, the value of the
1962 option is made available as
1964 value, which in turn can be used as
1969 .It Cm src-ip Ar ip-address
1970 Matches IPv4 packets whose source IP is one of the address(es)
1971 specified as an argument.
1972 .It Cm src-ip6 Ar ip6-address
1973 Matches IPv6 packets whose source IP is one of the address(es)
1974 specified as an argument.
1975 .It Cm src-port Ar ports
1976 Matches IP packets whose source port is one of the port(s)
1977 specified as argument.
1978 .It Cm tagged Ar tag-list
1979 Matches packets whose tags are included in
1981 which is either a single value or a list of values or ranges
1982 specified in the same way as
1984 Tags can be applied to the packet using
1986 rule action parameter (see it's description for details on tags).
1987 .It Cm tcpack Ar ack
1989 Match if the TCP header acknowledgment number field is set to
1991 .It Cm tcpdatalen Ar tcpdatalen-list
1992 Matches TCP packets whose length of TCP data is
1993 .Ar tcpdatalen-list ,
1994 which is either a single value or a list of values or ranges
1995 specified in the same way as
1997 .It Cm tcpflags Ar spec
1999 Match if the TCP header contains the comma separated list of
2002 The supported TCP flags are:
2011 The absence of a particular flag may be denoted
2014 A rule which contains a
2016 specification can never match a fragmented packet which has
2020 option for details on matching fragmented packets.
2021 .It Cm tcpmss Ar tcpmss-list
2022 Matches TCP packets whose MSS (maximum segment size) value is set to
2024 which is either a single value or a list of values or ranges
2025 specified in the same way as
2027 .It Cm tcpseq Ar seq
2029 Match if the TCP header sequence number field is set to
2031 .It Cm tcpwin Ar tcpwin-list
2032 Matches TCP packets whose header window field is set to
2034 which is either a single value or a list of values or ranges
2035 specified in the same way as
2037 .It Cm tcpoptions Ar spec
2039 Match if the TCP header contains the comma separated list of
2040 options specified in
2042 The supported TCP options are:
2045 (maximum segment size),
2047 (tcp window advertisement),
2051 (rfc1323 timestamp) and
2053 (rfc1644 t/tcp connection count).
2054 The absence of a particular option may be denoted
2058 Match all TCP or UDP packets sent by or received for a
2062 may be matched by name or identification number.
2064 For incoming packets,
2065 a routing table lookup is done on the packet's source address.
2066 If the interface on which the packet entered the system matches the
2067 outgoing interface for the route,
2069 If the interfaces do not match up,
2070 the packet does not match.
2071 All outgoing packets or packets with no incoming interface match.
2073 The name and functionality of the option is intentionally similar to
2074 the Cisco IOS command:
2076 .Dl ip verify unicast reverse-path
2078 This option can be used to make anti-spoofing rules to reject all
2079 packets with source addresses not from this interface.
2083 For incoming packets,
2084 a routing table lookup is done on the packet's source address.
2085 If a route to the source address exists, but not the default route
2086 or a blackhole/reject route, the packet matches.
2087 Otherwise, the packet does not match.
2088 All outgoing packets match.
2090 The name and functionality of the option is intentionally similar to
2091 the Cisco IOS command:
2093 .Dl ip verify unicast source reachable-via any
2095 This option can be used to make anti-spoofing rules to reject all
2096 packets whose source address is unreachable.
2098 For incoming packets, the packet's source address is checked if it
2099 belongs to a directly connected network.
2100 If the network is directly connected, then the interface the packet
2101 came on in is compared to the interface the network is connected to.
2102 When incoming interface and directly connected interface are not the
2103 same, the packet does not match.
2104 Otherwise, the packet does match.
2105 All outgoing packets match.
2107 This option can be used to make anti-spoofing rules to reject all
2108 packets that pretend to be from a directly connected network but do
2109 not come in through that interface.
2110 This option is similar to but more restricted than
2112 because it engages only on packets with source addresses of directly
2113 connected networks instead of all source addresses.
2116 Lookup tables are useful to handle large sparse sets of
2117 addresses or other search keys (e.g., ports, jail IDs, interface names).
2118 In the rest of this section we will use the term ``key''.
2119 Table name needs to match the following spec:
2121 Tables with the same name can be created in different
2123 However, rule links to the tables in
2126 This behavior can be controlled by
2127 .Va net.inet.ip.fw.tables_sets
2131 section for more information.
2132 There may be up to 65535 different lookup tables.
2134 The following table types are supported:
2135 .Bl -tag -width indent
2136 .It Ar table-type : Ar addr | iface | number | flow
2137 .It Ar table-key : Ar addr Ns Oo / Ns Ar masklen Oc | iface-name | number | flow-spec
2138 .It Ar flow-spec : Ar flow-field Ns Op , Ns Ar flow-spec
2139 .It Ar flow-field : src-ip | proto | src-port | dst-ip | dst-port
2141 Matches IPv4 or IPv6 address.
2142 Each entry is represented by an
2143 .Ar addr Ns Op / Ns Ar masklen
2144 and will match all addresses with base
2146 (specified as an IPv4/IPv6 address, or a hostname) and mask width of
2151 is not specified, it defaults to 32 for IPv4 and 128 for IPv6.
2152 When looking up an IP address in a table, the most specific
2155 Matches interface names.
2156 Each entry is represented by string treated as interface name.
2157 Wildcards are not supported.
2159 Matches protocol ports, uids/gids or jail IDs.
2160 Each entry is represented by 32-bit unsigned integer.
2161 Ranges are not supported.
2163 Matches packet fields specified by
2165 type suboptions with table entries.
2168 Tables require explicit creation via
2172 The following creation options are supported:
2173 .Bl -tag -width indent
2174 .It Ar create-options : Ar create-option | create-options
2175 .It Ar create-option : Cm type Ar table-type | Cm valtype Ar value-mask | Cm algo Ar algo-desc |
2176 .Cm limit Ar number | Cm locked | Cm missing | Cm or-flush
2182 Table algorithm to use (see below).
2184 Maximum number of items that may be inserted into table.
2186 Restrict any table modifications.
2188 Do not fail if table already exists and has exactly same options as new one.
2190 Flush existing table with same name instead of returning error.
2193 so existing table must be compatible with new one.
2196 Some of these options may be modified later via
2199 The following options can be changed:
2200 .Bl -tag -width indent
2201 .It Ar modify-options : Ar modify-option | modify-options
2202 .It Ar modify-option : Cm limit Ar number
2204 Alter maximum number of items that may be inserted into table.
2207 Additionally, table can be locked or unlocked using
2215 can be swapped with each other using
2218 Swap may fail if tables limits are set and data exchange
2219 would result in limits hit.
2220 Operation is performed atomically.
2222 One or more entries can be added to a table at once using
2225 Addition of all items are performed atomically.
2226 By default, error in addition of one entry does not influence
2227 addition of other entries.
2228 However, non-zero error code is returned in that case.
2231 keyword may be specified before
2233 to indicate all-or-none add request.
2235 One or more entries can be removed from a table at once using
2238 By default, error in removal of one entry does not influence
2239 removing of other entries.
2240 However, non-zero error code is returned in that case.
2242 It may be possible to check what entry will be found on particular
2248 This functionality is optional and may be unsupported in some algorithms.
2250 The following operations can be performed on
2255 .Bl -tag -width indent
2259 Removes all entries.
2261 Shows generic table information.
2263 Shows generic table information and algo-specific data.
2266 The following lookup algorithms are supported:
2267 .Bl -tag -width indent
2268 .It Ar algo-desc : algo-name | "algo-name algo-data"
2269 .It Ar algo-name : Ar addr: radix | addr: hash | iface: array | number: array | flow: hash
2271 Separate Radix trees for IPv4 and IPv6, the same way as the routing table (see
2277 Separate auto-growing hashes for IPv4 and IPv6.
2278 Accepts entries with the same mask length specified initially via
2279 .Cm "addr:hash masks=/v4,/v6"
2280 algorithm creation options.
2281 Assume /32 and /128 masks by default.
2282 Search removes host bits (according to mask) from supplied address and checks
2283 resulting key in appropriate hash.
2284 Mostly optimized for /64 and byte-ranged IPv6 masks.
2286 Array storing sorted indexes for entries which are presented in the system.
2287 Optimized for very fast lookup.
2289 Array storing sorted u32 numbers.
2291 Auto-growing hash storing flow entries.
2292 Search calculates hash on required packet fields and searches for matching
2293 entries in selected bucket.
2298 feature provides the ability to use a value, looked up in the table, as
2299 the argument for a rule action, action parameter or rule option.
2300 This can significantly reduce number of rules in some configurations.
2301 If two tables are used in a rule, the result of the second (destination)
2304 Each record may hold one or more values according to
2306 This mask is set on table creation via
2309 The following value types are supported:
2310 .Bl -tag -width indent
2311 .It Ar value-mask : Ar value-type Ns Op , Ns Ar value-mask
2312 .It Ar value-type : Ar skipto | pipe | fib | nat | dscp | tag | divert |
2313 .Ar netgraph | limit | ipv4
2315 rule number to jump to.
2319 fib number to match/set.
2321 nat number to jump to.
2323 dscp value to match/set.
2325 tag number to match/set.
2327 port number to divert traffic to.
2329 hook number to move packet to.
2331 maximum number of connections.
2333 IPv4 nexthop to fwd packets to.
2335 IPv6 nexthop to fwd packets to.
2340 argument can be used with the following actions:
2341 .Cm nat, pipe, queue, divert, tee, netgraph, ngtee, fwd, skipto, setfib ,
2349 action, the user should be aware that the code will walk the ruleset
2350 up to a rule equal to, or past, the given number.
2354 Section for example usage of tables and the tablearg keyword.
2356 Each rule or table belongs to one of 32 different
2359 Set 31 is reserved for the default rule.
2361 By default, rules or tables are put in set 0, unless you use the
2363 attribute when adding a new rule or table.
2364 Sets can be individually and atomically enabled or disabled,
2365 so this mechanism permits an easy way to store multiple configurations
2366 of the firewall and quickly (and atomically) switch between them.
2368 By default, tables from set 0 are referenced when adding rule with
2369 table opcodes regardless of rule set.
2370 This behavior can be changed by setting
2371 .Va net.inet.ip.fw.tables_sets
2373 Rule's set will then be used for table references.
2375 The command to enable/disable sets is
2376 .Bd -ragged -offset indent
2378 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
2385 sections can be specified.
2386 Command execution is atomic on all the sets specified in the command.
2387 By default, all sets are enabled.
2389 When you disable a set, its rules behave as if they do not exist
2390 in the firewall configuration, with only one exception:
2391 .Bd -ragged -offset indent
2392 dynamic rules created from a rule before it had been disabled
2393 will still be active until they expire.
2395 dynamic rules you have to explicitly delete the parent rule
2396 which generated them.
2399 The set number of rules can be changed with the command
2400 .Bd -ragged -offset indent
2403 .Brq Cm rule Ar rule-number | old-set
2407 Also, you can atomically swap two rulesets with the command
2408 .Bd -ragged -offset indent
2410 .Cm set swap Ar first-set second-set
2415 Section on some possible uses of sets of rules.
2416 .Sh STATEFUL FIREWALL
2417 Stateful operation is a way for the firewall to dynamically
2418 create rules for specific flows when packets that
2419 match a given pattern are detected.
2420 Support for stateful
2421 operation comes through the
2422 .Cm check-state , keep-state , record-state , limit
2428 Dynamic rules are created when a packet matches a
2434 rule, causing the creation of a
2436 rule which will match all and only packets with
2440 .Em src-ip/src-port dst-ip/dst-port
2445 are used here only to denote the initial match addresses, but they
2446 are completely equivalent afterwards).
2452 This name is used in matching together with addresses, ports and protocol.
2453 Dynamic rules will be checked at the first
2454 .Cm check-state, keep-state
2457 occurrence, and the action performed upon a match will be the same
2458 as in the parent rule.
2460 Note that no additional attributes other than protocol and IP addresses
2461 and ports and :flowname are checked on dynamic rules.
2463 The typical use of dynamic rules is to keep a closed firewall configuration,
2464 but let the first TCP SYN packet from the inside network install a
2465 dynamic rule for the flow so that packets belonging to that session
2466 will be allowed through the firewall:
2468 .Dl "ipfw add check-state :OUTBOUND"
2469 .Dl "ipfw add allow tcp from my-subnet to any setup keep-state :OUTBOUND"
2470 .Dl "ipfw add deny tcp from any to any"
2472 A similar approach can be used for UDP, where an UDP packet coming
2473 from the inside will install a dynamic rule to let the response through
2476 .Dl "ipfw add check-state :OUTBOUND"
2477 .Dl "ipfw add allow udp from my-subnet to any keep-state :OUTBOUND"
2478 .Dl "ipfw add deny udp from any to any"
2480 Dynamic rules expire after some time, which depends on the status
2481 of the flow and the setting of some
2485 .Sx SYSCTL VARIABLES
2487 For TCP sessions, dynamic rules can be instructed to periodically
2488 send keepalive packets to refresh the state of the rule when it is
2493 for more examples on how to use dynamic rules.
2494 .Sh TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
2496 is also the user interface for the
2498 traffic shaper, packet scheduler and network emulator, a subsystem that
2499 can artificially queue, delay or drop packets
2500 emulating the behaviour of certain network links
2501 or queueing systems.
2504 operates by first using the firewall to select packets
2505 using any match pattern that can be used in
2508 Matching packets are then passed to either of two
2509 different objects, which implement the traffic regulation:
2510 .Bl -hang -offset XXXX
2516 with given bandwidth and propagation delay,
2517 driven by a FIFO scheduler and a single queue with programmable
2518 queue size and packet loss rate.
2519 Packets are appended to the queue as they come out from
2521 and then transferred in FIFO order to the link at the desired rate.
2525 is an abstraction used to implement packet scheduling
2526 using one of several packet scheduling algorithms.
2529 are first grouped into flows according to a mask on the 5-tuple.
2530 Flows are then passed to the scheduler associated to the
2532 and each flow uses scheduling parameters (weight and others)
2533 as configured in the
2536 A scheduler in turn is connected to an emulated link,
2537 and arbitrates the link's bandwidth among backlogged flows according to
2538 weights and to the features of the scheduling algorithm in use.
2543 can be used to set hard limits to the bandwidth that a flow can use, whereas
2545 can be used to determine how different flows share the available bandwidth.
2547 A graphical representation of the binding of queues,
2548 flows, schedulers and links is below.
2549 .Bd -literal -offset indent
2550 (flow_mask|sched_mask) sched_mask
2551 +---------+ weight Wx +-------------+
2552 | |->-[flow]-->--| |-+
2553 -->--| QUEUE x | ... | | |
2554 | |->-[flow]-->--| SCHEDuler N | |
2556 ... | +--[LINK N]-->--
2557 +---------+ weight Wy | | +--[LINK N]-->--
2558 | |->-[flow]-->--| | |
2559 -->--| QUEUE y | ... | | |
2560 | |->-[flow]-->--| | |
2561 +---------+ +-------------+ |
2564 It is important to understand the role of the SCHED_MASK
2565 and FLOW_MASK, which are configured through the commands
2566 .Dl "ipfw sched N config mask SCHED_MASK ..."
2568 .Dl "ipfw queue X config mask FLOW_MASK ..." .
2570 The SCHED_MASK is used to assign flows to one or more
2571 scheduler instances, one for each
2572 value of the packet's 5-tuple after applying SCHED_MASK.
2573 As an example, using ``src-ip 0xffffff00'' creates one instance
2574 for each /24 destination subnet.
2576 The FLOW_MASK, together with the SCHED_MASK, is used to split
2578 As an example, using
2579 ``src-ip 0x000000ff''
2580 together with the previous SCHED_MASK makes a flow for
2581 each individual source address.
2582 In turn, flows for each /24
2583 subnet will be sent to the same scheduler instance.
2585 The above diagram holds even for the
2587 case, with the only restriction that a
2589 only supports a SCHED_MASK, and forces the use of a FIFO
2590 scheduler (these are for backward compatibility reasons;
2591 in fact, internally, a
2593 pipe is implemented exactly as above).
2595 There are two modes of
2603 mode tries to emulate a real link: the
2605 scheduler ensures that the packet will not leave the pipe faster than it
2606 would on the real link with a given bandwidth.
2609 mode allows certain packets to bypass the
2611 scheduler (if packet flow does not exceed pipe's bandwidth).
2612 This is the reason why the
2614 mode requires less CPU cycles per packet (on average) and packet latency
2615 can be significantly lower in comparison to a real link with the same
2621 mode can be enabled by setting the
2622 .Va net.inet.ip.dummynet.io_fast
2624 variable to a non-zero value.
2625 .Ss PIPE, QUEUE AND SCHEDULER CONFIGURATION
2631 configuration commands are the following:
2632 .Bd -ragged -offset indent
2633 .Cm pipe Ar number Cm config Ar pipe-configuration
2635 .Cm queue Ar number Cm config Ar queue-configuration
2637 .Cm sched Ar number Cm config Ar sched-configuration
2640 The following parameters can be configured for a pipe:
2642 .Bl -tag -width indent -compact
2643 .It Cm bw Ar bandwidth | device
2644 Bandwidth, measured in
2647 .Brq Cm bit/s | Byte/s .
2650 A value of 0 (default) means unlimited bandwidth.
2651 The unit must immediately follow the number, as in
2653 .Dl "dnctl pipe 1 config bw 300Kbit/s"
2655 If a device name is specified instead of a numeric value, as in
2657 .Dl "dnctl pipe 1 config bw tun0"
2659 then the transmit clock is supplied by the specified device.
2660 At the moment only the
2662 device supports this
2663 functionality, for use in conjunction with
2666 .It Cm delay Ar ms-delay
2667 Propagation delay, measured in milliseconds.
2668 The value is rounded to the next multiple of the clock tick
2669 (typically 10ms, but it is a good practice to run kernels
2671 .Dq "options HZ=1000"
2673 the granularity to 1ms or less).
2674 The default value is 0, meaning no delay.
2676 .It Cm burst Ar size
2677 If the data to be sent exceeds the pipe's bandwidth limit
2678 (and the pipe was previously idle), up to
2680 bytes of data are allowed to bypass the
2682 scheduler, and will be sent as fast as the physical link allows.
2683 Any additional data will be transmitted at the rate specified
2687 The burst size depends on how long the pipe has been idle;
2688 the effective burst size is calculated as follows:
2695 .It Cm profile Ar filename
2696 A file specifying the additional overhead incurred in the transmission
2697 of a packet on the link.
2699 Some link types introduce extra delays in the transmission
2700 of a packet, e.g., because of MAC level framing, contention on
2701 the use of the channel, MAC level retransmissions and so on.
2702 From our point of view, the channel is effectively unavailable
2703 for this extra time, which is constant or variable depending
2705 Additionally, packets may be dropped after this
2706 time (e.g., on a wireless link after too many retransmissions).
2707 We can model the additional delay with an empirical curve
2708 that represents its distribution.
2709 .Bd -literal -offset indent
2710 cumulative probability
2720 +-------*------------------->
2723 The empirical curve may have both vertical and horizontal lines.
2724 Vertical lines represent constant delay for a range of
2726 Horizontal lines correspond to a discontinuity in the delay
2727 distribution: the pipe will use the largest delay for a
2730 The file format is the following, with whitespace acting as
2731 a separator and '#' indicating the beginning a comment:
2732 .Bl -tag -width indent
2733 .It Cm name Ar identifier
2734 optional name (listed by "dnctl pipe show")
2735 to identify the delay distribution;
2737 the bandwidth used for the pipe.
2738 If not specified here, it must be present
2739 explicitly as a configuration parameter for the pipe;
2740 .It Cm loss-level Ar L
2741 the probability above which packets are lost.
2742 (0.0 <= L <= 1.0, default 1.0 i.e., no loss);
2744 the number of samples used in the internal
2745 representation of the curve (2..1024; default 100);
2746 .It Cm "delay prob" | "prob delay"
2747 One of these two lines is mandatory and defines
2748 the format of the following lines with data points.
2750 2 or more lines representing points in the curve,
2751 with either delay or probability first, according
2752 to the chosen format.
2753 The unit for delay is milliseconds.
2754 Data points do not need to be sorted.
2755 Also, the number of actual lines can be different
2756 from the value of the "samples" parameter:
2758 utility will sort and interpolate
2759 the curve as needed.
2762 Example of a profile file:
2763 .Bd -literal -offset indent
2768 0 200 # minimum overhead is 200ms
2774 #configuration file end
2778 The following parameters can be configured for a queue:
2780 .Bl -tag -width indent -compact
2781 .It Cm pipe Ar pipe_nr
2782 Connects a queue to the specified pipe.
2783 Multiple queues (with the same or different weights) can be connected to
2784 the same pipe, which specifies the aggregate rate for the set of queues.
2786 .It Cm weight Ar weight
2787 Specifies the weight to be used for flows matching this queue.
2788 The weight must be in the range 1..100, and defaults to 1.
2791 The following case-insensitive parameters can be configured for a
2794 .Bl -tag -width indent -compact
2795 .It Cm type Ar {fifo | wf2q+ | rr | qfq | fq_codel | fq_pie}
2796 specifies the scheduling algorithm to use.
2797 .Bl -tag -width indent -compact
2799 is just a FIFO scheduler (which means that all packets
2800 are stored in the same queue as they arrive to the scheduler).
2801 FIFO has O(1) per-packet time complexity, with very low
2802 constants (estimate 60-80ns on a 2GHz desktop machine)
2803 but gives no service guarantees.
2805 implements the WF2Q+ algorithm, which is a Weighted Fair Queueing
2806 algorithm which permits flows to share bandwidth according to
2808 Note that weights are not priorities; even a flow
2809 with a minuscule weight will never starve.
2810 WF2Q+ has O(log N) per-packet processing cost, where N is the number
2811 of flows, and is the default algorithm used by previous versions
2814 implements the Deficit Round Robin algorithm, which has O(1) processing
2815 costs (roughly, 100-150ns per packet)
2816 and permits bandwidth allocation according to weights, but
2817 with poor service guarantees.
2819 implements the QFQ algorithm, which is a very fast variant of
2820 WF2Q+, with similar service guarantees and O(1) processing
2821 costs (roughly, 200-250ns per packet).
2823 implements the FQ-CoDel (FlowQueue-CoDel) scheduler/AQM algorithm, which
2824 uses a modified Deficit Round Robin scheduler to manage two lists of sub-queues
2825 (old sub-queues and new sub-queues) for providing brief periods of priority to
2826 lightweight or short burst flows.
2827 By default, the total number of sub-queues is 1024.
2828 FQ-CoDel's internal, dynamically
2829 created sub-queues are controlled by separate instances of CoDel AQM.
2831 implements the FQ-PIE (FlowQueue-PIE) scheduler/AQM algorithm, which similar to
2833 but uses per sub-queue PIE AQM instance to control the queue delay.
2837 inherits AQM parameters and options from
2841 inherits AQM parameters and options from
2844 Additionally, both of
2848 have shared scheduler parameters which are:
2849 .Bl -tag -width indent
2852 specifies the quantum (credit) of the scheduler.
2854 is the number of bytes a queue can serve before being moved to the tail
2856 The default is 1514 bytes, and the maximum acceptable value
2860 specifies the hard size limit (in unit of packets) of all queues managed by an
2861 instance of the scheduler.
2862 The default value of
2864 is 10240 packets, and the maximum acceptable value is 20480 packets.
2867 specifies the total number of flow queues (sub-queues) that fq_*
2868 creates and manages.
2869 By default, 1024 sub-queues are created when an instance
2870 of the fq_{codel/pie} scheduler is created.
2871 The maximum acceptable value is
2875 Note that any token after
2879 is considered a parameter for fq_{codel/pie}.
2880 So, ensure all scheduler
2881 configuration options not related to fq_{codel/pie} are written before
2886 In addition to the type, all parameters allowed for a pipe can also
2887 be specified for a scheduler.
2889 Finally, the following parameters can be configured for both
2892 .Bl -tag -width XXXX -compact
2893 .It Cm buckets Ar hash-table-size
2894 Specifies the size of the hash table used for storing the
2896 Default value is 64 controlled by the
2899 .Va net.inet.ip.dummynet.hash_size ,
2900 allowed range is 16 to 65536.
2902 .It Cm mask Ar mask-specifier
2903 Packets sent to a given pipe or queue by an
2905 rule can be further classified into multiple flows, each of which is then
2909 A flow identifier is constructed by masking the IP addresses,
2910 ports and protocol types as specified with the
2912 options in the configuration of the pipe or queue.
2913 For each different flow identifier, a new pipe or queue is created
2914 with the same parameters as the original object, and matching packets
2919 are used, each flow will get the same bandwidth as defined by the pipe,
2922 are used, each flow will share the parent's pipe bandwidth evenly
2923 with other flows generated by the same queue (note that other queues
2924 with different weights might be connected to the same pipe).
2926 Available mask specifiers are a combination of one or more of the following:
2928 .Cm dst-ip Ar mask ,
2929 .Cm dst-ip6 Ar mask ,
2930 .Cm src-ip Ar mask ,
2931 .Cm src-ip6 Ar mask ,
2932 .Cm dst-port Ar mask ,
2933 .Cm src-port Ar mask ,
2934 .Cm flow-id Ar mask ,
2939 where the latter means all bits in all fields are significant.
2942 When a packet is dropped by a
2944 queue or pipe, the error
2945 is normally reported to the caller routine in the kernel, in the
2946 same way as it happens when a device queue fills up.
2948 option reports the packet as successfully delivered, which can be
2949 needed for some experimental setups where you want to simulate
2950 loss or congestion at a remote router.
2952 .It Cm plr Ar packet-loss-rate
2955 .Ar packet-loss-rate
2956 is a floating-point number between 0 and 1, with 0 meaning no
2957 loss, 1 meaning 100% loss.
2958 The loss rate is internally represented on 31 bits.
2960 .It Cm queue Brq Ar slots | size Ns Cm Kbytes
2965 Default value is 50 slots, which
2966 is the typical queue size for Ethernet devices.
2967 Note that for slow speed links you should keep the queue
2968 size short or your traffic might be affected by a significant
2970 E.g., 50 max-sized Ethernet packets (1500 bytes) mean 600Kbit
2971 or 20s of queue on a 30Kbit/s pipe.
2972 Even worse effects can result if you get packets from an
2973 interface with a much larger MTU, e.g.\& the loopback interface
2974 with its 16KB packets.
2978 .Em net.inet.ip.dummynet.pipe_byte_limit
2980 .Em net.inet.ip.dummynet.pipe_slot_limit
2981 control the maximum lengths that can be specified.
2983 .It Cm red | gred Ar w_q Ns / Ns Ar min_th Ns / Ns Ar max_th Ns / Ns Ar max_p
2985 Make use of the RED (Random Early Detection) queue management algorithm.
2990 point numbers between 0 and 1 (inclusive), while
2994 are integer numbers specifying thresholds for queue management
2995 (thresholds are computed in bytes if the queue has been defined
2996 in bytes, in slots otherwise).
2997 The two parameters can also be of the same value if needed.
3000 also supports the gentle RED variant (gred) and ECN (Explicit Congestion
3001 Notification) as optional.
3004 variables can be used to control the RED behaviour:
3005 .Bl -tag -width indent
3006 .It Va net.inet.ip.dummynet.red_lookup_depth
3007 specifies the accuracy in computing the average queue
3008 when the link is idle (defaults to 256, must be greater than zero)
3009 .It Va net.inet.ip.dummynet.red_avg_pkt_size
3010 specifies the expected average packet size (defaults to 512, must be
3012 .It Va net.inet.ip.dummynet.red_max_pkt_size
3013 specifies the expected maximum packet size, only used when queue
3014 thresholds are in bytes (defaults to 1500, must be greater than zero).
3017 .It Cm codel Oo Cm target Ar time Oc Oo Cm interval Ar time Oc Oo Cm ecn |
3019 Make use of the CoDel (Controlled-Delay) queue management algorithm.
3021 is interpreted as milliseconds by default but seconds (s), milliseconds (ms) or
3022 microseconds (us) can be specified instead.
3023 CoDel drops or marks (ECN) packets
3024 depending on packet sojourn time in the queue.
3027 (5ms by default) is the minimum acceptable persistent queue delay that CoDel
3029 CoDel does not drop packets directly after packets sojourn time becomes
3036 (100ms default) before dropping.
3039 should be set to maximum RTT for all expected connections.
3041 enables (disabled by default) packet marking (instead of dropping) for
3042 ECN-enabled TCP flows when queue delay becomes high.
3044 Note that any token after
3046 is considered a parameter for CoDel.
3047 So, ensure all pipe/queue
3048 configuration options are written before
3055 .Va net.inet.ip.dummynet.codel.target
3057 .Va net.inet.ip.dummynet.codel.interval
3058 can be used to set CoDel default parameters.
3060 .It Cm pie Oo Cm target Ar time Oc Oo Cm tupdate Ar time Oc Oo
3061 .Cm alpha Ar n Oc Oo Cm beta Ar n Oc Oo Cm max_burst Ar time Oc Oo
3062 .Cm max_ecnth Ar n Oc Oo Cm ecn | Cm noecn Oc Oo Cm capdrop |
3063 .Cm nocapdrop Oc Oo Cm drand | Cm nodrand Oc Oo Cm onoff
3064 .Oc Oo Cm dre | Cm ts Oc
3065 Make use of the PIE (Proportional Integral controller Enhanced) queue management
3067 PIE drops or marks packets depending on a calculated drop probability during
3068 en-queue process, with the aim of achieving high throughput while keeping queue
3070 At regular time intervals of
3073 (15ms by default) a background process (re)calculates the probability based on queue delay
3077 (15ms by default) and queue delay trends.
3078 PIE approximates current queue
3079 delay by using a departure rate estimation method, or (optionally) by using a
3080 packet timestamp method similar to CoDel.
3082 is interpreted as milliseconds by default but seconds (s), milliseconds (ms) or
3083 microseconds (us) can be specified instead.
3084 The other PIE parameters and options are as follows:
3085 .Bl -tag -width indent
3088 is a floating point number between 0 and 7 which specifies the weight of queue
3089 delay deviations that is used in drop probability calculation.
3090 0.125 is the default.
3093 is a floating point number between 0 and 7 which specifies is the weight of queue
3094 delay trend that is used in drop probability calculation.
3095 1.25 is the default.
3096 .It Cm max_burst Ar time
3097 The maximum period of time that PIE does not drop/mark packets.
3099 default and 10s is the maximum value.
3100 .It Cm max_ecnth Ar n
3101 Even when ECN is enabled, PIE drops packets instead of marking them when drop
3102 probability becomes higher than ECN probability threshold
3104 , the default is 0.1 (i.e 10%) and 1 is the maximum value.
3106 enable or disable ECN marking for ECN-enabled TCP flows.
3107 Disabled by default.
3108 .It Cm capdrop | nocapdrop
3109 enable or disable cap drop adjustment.
3110 Cap drop adjustment is enabled by default.
3111 .It Cm drand | nodrand
3112 enable or disable drop probability de-randomisation.
3113 De-randomisation eliminates
3114 the problem of dropping packets too close or too far.
3115 De-randomisation is enabled by default.
3117 enable turning PIE on and off depending on queue load.
3118 If this option is enabled,
3119 PIE turns on when over 1/3 of queue becomes full.
3120 This option is disabled by
3123 Calculate queue delay using departure rate estimation
3131 Note that any token after
3133 is considered a parameter for PIE.
3134 So ensure all pipe/queue
3135 the configuration options are written before
3139 variables can be used to control the
3143 .Sx SYSCTL VARIABLES
3144 section for more details.
3147 When used with IPv6 data,
3149 currently has several limitations.
3150 Information necessary to route link-local packets to an
3151 interface is not available after processing by
3153 so those packets are dropped in the output path.
3154 Care should be taken to ensure that link-local packets are not passed to
3157 Here are some important points to consider when designing your
3161 Remember that you filter both packets going
3165 Most connections need packets going in both directions.
3167 Remember to test very carefully.
3168 It is a good idea to be near the console when doing this.
3169 If you cannot be near the console,
3170 use an auto-recovery script such as the one in
3171 .Pa /usr/share/examples/ipfw/change_rules.sh .
3173 Do not forget the loopback interface.
3178 There are circumstances where fragmented datagrams are unconditionally
3180 TCP packets are dropped if they do not contain at least 20 bytes of
3181 TCP header, UDP packets are dropped if they do not contain a full 8
3182 byte UDP header, and ICMP packets are dropped if they do not contain
3183 4 bytes of ICMP header, enough to specify the ICMP type, code, and
3185 These packets are simply logged as
3187 since there may not be enough good data in the packet to produce a
3188 meaningful log entry.
3190 Another type of packet is unconditionally dropped, a TCP packet with a
3191 fragment offset of one.
3192 This is a valid packet, but it only has one use, to try
3193 to circumvent firewalls.
3194 When logging is enabled, these packets are
3195 reported as being dropped by rule -1.
3197 If you are logged in over a network, loading the
3201 is probably not as straightforward as you would think.
3202 The following command line is recommended:
3203 .Bd -literal -offset indent
3205 ipfw add 32000 allow ip from any to any
3208 Along the same lines, doing an
3209 .Bd -literal -offset indent
3213 in similar surroundings is also a bad idea.
3217 filter list may not be modified if the system security level
3218 is set to 3 or higher
3221 for information on system security levels).
3223 .Sh PACKET DIVERSION
3226 socket bound to the specified port will receive all packets
3227 diverted to that port.
3228 If no socket is bound to the destination port, or if the divert module is
3229 not loaded, or if the kernel was not compiled with divert socket support,
3230 the packets are dropped.
3231 .Sh NETWORK ADDRESS TRANSLATION (NAT)
3233 support in-kernel NAT using the kernel version of
3237 should be loaded or kernel should have
3238 .Cm options IPFIREWALL_NAT
3241 The nat configuration command is the following:
3242 .Bd -ragged -offset indent
3247 .Ar nat-configuration
3251 The following parameters can be configured:
3252 .Bl -tag -width indent
3253 .It Cm ip Ar ip_address
3254 Define an ip address to use for aliasing.
3256 Use ip address of NIC for aliasing, dynamically changing
3257 it if NIC's ip address changes.
3259 Enable logging on this nat instance.
3261 Deny any incoming connection from outside world.
3263 Try to leave the alias port numbers unchanged from
3264 the actual local port numbers.
3266 Traffic on the local network not originating from a RFC 1918
3267 unregistered address spaces will be ignored.
3269 Like unreg_only, but includes the RFC 6598 (Carrier Grade NAT)
3272 Reset table of the packet aliasing engine on address change.
3274 Reverse the way libalias handles aliasing.
3276 Obey transparent proxy rules only, packet aliasing is not performed.
3278 Skip instance in case of global state lookup (see below).
3279 .It Cm port_range Ar lower-upper
3280 Set the aliasing ports between the ranges given. Upper port has to be greater
3284 Some special values can be supplied instead of
3286 in nat rule actions:
3287 .Bl -tag -width indent
3289 Looks up translation state in all configured nat instances.
3290 If an entry is found, packet is aliased according to that entry.
3291 If no entry was found in any of the instances, packet is passed unchanged,
3292 and no new entry will be created.
3294 .Sx MULTIPLE INSTANCES
3297 for more information.
3299 Uses argument supplied in lookup table.
3302 section below for more information on lookup tables.
3305 To let the packet continue after being (de)aliased, set the sysctl variable
3306 .Va net.inet.ip.fw.one_pass
3308 For more information about aliasing modes, refer to
3312 for some examples of nat usage.
3313 .Ss REDIRECT AND LSNAT SUPPORT IN IPFW
3314 Redirect and LSNAT support follow closely the syntax used in
3318 for some examples on how to do redirect and lsnat.
3319 .Ss SCTP NAT SUPPORT
3320 SCTP nat can be configured in a similar manner to TCP through the
3323 The main difference is that
3325 does not do port translation.
3326 Since the local and global side ports will be the same,
3327 there is no need to specify both.
3328 Ports are redirected as follows:
3329 .Bd -ragged -offset indent
3335 .Cm redirect_port sctp
3336 .Ar ip_address [,addr_list] {[port | port-port] [,ports]}
3342 configuration can be done in real-time through the
3345 All may be changed dynamically, though the hash_table size will only
3350 .Sx SYSCTL VARIABLES
3352 .Sh IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
3353 .Ss Stateful translation
3355 supports in-kernel IPv6/IPv4 network address and protocol translation.
3356 Stateful NAT64 translation allows IPv6-only clients to contact IPv4 servers
3357 using unicast TCP, UDP or ICMP protocols.
3358 One or more IPv4 addresses assigned to a stateful NAT64 translator are shared
3359 among several IPv6-only clients.
3360 When stateful NAT64 is used in conjunction with DNS64, no changes are usually
3361 required in the IPv6 client or the IPv4 server.
3364 should be loaded or kernel should have
3365 .Cm options IPFIREWALL_NAT64
3366 to be able use stateful NAT64 translator.
3368 Stateful NAT64 uses a bunch of memory for several types of objects.
3369 When IPv6 client initiates connection, NAT64 translator creates a host entry
3370 in the states table.
3371 Each host entry uses preallocated IPv4 alias entry.
3372 Each alias entry has a number of ports group entries allocated on demand.
3373 Ports group entries contains connection state entries.
3374 There are several options to control limits and lifetime for these objects.
3376 NAT64 translator follows RFC7915 when does ICMPv6/ICMP translation,
3377 unsupported message types will be silently dropped.
3378 IPv6 needs several ICMPv6 message types to be explicitly allowed for correct
3380 Make sure that ND6 neighbor solicitation (ICMPv6 type 135) and neighbor
3381 advertisement (ICMPv6 type 136) messages will not be handled by translation
3384 After translation NAT64 translator by default sends packets through
3385 corresponding netisr queue.
3386 Thus translator host should be configured as IPv4 and IPv6 router.
3387 Also this means, that a packet is handled by firewall twice.
3388 First time an original packet is handled and consumed by translator,
3389 and then it is handled again as translated packet.
3390 This behavior can be changed by sysctl variable
3391 .Va net.inet.ip.fw.nat64_direct_output .
3392 Also translated packet can be tagged using
3394 rule action, and then matched by
3396 opcode to avoid loops and extra overhead.
3398 The stateful NAT64 configuration command is the following:
3399 .Bd -ragged -offset indent
3408 The following parameters can be configured:
3409 .Bl -tag -width indent
3410 .It Cm prefix4 Ar ipv4_prefix/plen
3411 The IPv4 prefix with mask defines the pool of IPv4 addresses used as
3412 source address after translation.
3413 Stateful NAT64 module translates IPv6 source address of client to one
3414 IPv4 address from this pool.
3415 Note that incoming IPv4 packets that don't have corresponding state entry
3416 in the states table will be dropped by translator.
3417 Make sure that translation rules handle packets, destined to configured prefix.
3418 .It Cm prefix6 Ar ipv6_prefix/length
3419 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3420 to represent IPv4 addresses.
3421 This IPv6 prefix should be configured in DNS64.
3422 The translator implementation follows RFC6052, that restricts the length of
3423 prefixes to one of following: 32, 40, 48, 56, 64, or 96.
3424 The Well-Known IPv6 Prefix 64:ff9b:: must be 96 bits long.
3427 prefix can be used to handle several IPv6 prefixes with one NAT64 instance.
3428 The NAT64 instance will determine a destination IPv4 address from prefix
3430 .It Cm states_chunks Ar number
3431 The number of states chunks in single ports group.
3432 Each ports group by default can keep 64 state entries in single chunk.
3433 The above value affects the maximum number of states that can be associated with single IPv4 alias address and port.
3434 The value must be power of 2, and up to 128.
3435 .It Cm host_del_age Ar seconds
3436 The number of seconds until the host entry for a IPv6 client will be deleted
3437 and all its resources will be released due to inactivity.
3440 .It Cm pg_del_age Ar seconds
3441 The number of seconds until a ports group with unused state entries will
3445 .It Cm tcp_syn_age Ar seconds
3446 The number of seconds while a state entry for TCP connection with only SYN
3448 If TCP connection establishing will not be finished,
3449 state entry will be deleted.
3452 .It Cm tcp_est_age Ar seconds
3453 The number of seconds while a state entry for established TCP connection
3457 .It Cm tcp_close_age Ar seconds
3458 The number of seconds while a state entry for closed TCP connection
3460 Keeping state entries for closed connections is needed, because IPv4 servers
3461 typically keep closed connections in a TIME_WAIT state for a several minutes.
3462 Since translator's IPv4 addresses are shared among all IPv6 clients,
3463 new connections from the same addresses and ports may be rejected by server,
3464 because these connections are still in a TIME_WAIT state.
3465 Keeping them in translator's state table protects from such rejects.
3468 .It Cm udp_age Ar seconds
3469 The number of seconds while translator keeps state entry in a waiting for
3470 reply to the sent UDP datagram.
3473 .It Cm icmp_age Ar seconds
3474 The number of seconds while translator keeps state entry in a waiting for
3475 reply to the sent ICMP message.
3479 Turn on logging of all handled packets via BPF through
3483 is a pseudo interface and can be created after a boot manually with
3486 Note that it has different purpose than
3489 Translators sends to BPF an additional information with each packet.
3492 you are able to see each handled packet before and after translation.
3494 Turn off logging of all handled packets via BPF.
3495 .It Cm allow_private
3496 Turn on processing private IPv4 addresses.
3497 By default IPv6 packets with destinations mapped to private address ranges
3498 defined by RFC1918 are not processed.
3499 .It Cm -allow_private
3500 Turn off private address handling in
3505 To inspect a states table of stateful NAT64 the following command can be used:
3506 .Bd -ragged -offset indent
3514 Stateless NAT64 translator doesn't use a states table for translation
3515 and converts IPv4 addresses to IPv6 and vice versa solely based on the
3516 mappings taken from configured lookup tables.
3517 Since a states table doesn't used by stateless translator,
3518 it can be configured to pass IPv4 clients to IPv6-only servers.
3520 The stateless NAT64 configuration command is the following:
3521 .Bd -ragged -offset indent
3530 The following parameters can be configured:
3531 .Bl -tag -width indent
3532 .It Cm prefix6 Ar ipv6_prefix/length
3533 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3534 to represent IPv4 addresses.
3535 This IPv6 prefix should be configured in DNS64.
3536 .It Cm table4 Ar table46
3539 contains mapping how IPv4 addresses should be translated to IPv6 addresses.
3540 .It Cm table6 Ar table64
3543 contains mapping how IPv6 addresses should be translated to IPv4 addresses.
3545 Turn on logging of all handled packets via BPF through
3549 Turn off logging of all handled packets via BPF.
3550 .It Cm allow_private
3551 Turn on processing private IPv4 addresses.
3552 By default IPv6 packets with destinations mapped to private address ranges
3553 defined by RFC1918 are not processed.
3554 .It Cm -allow_private
3555 Turn off private address handling in
3560 Note that the behavior of stateless translator with respect to not matched
3561 packets differs from stateful translator.
3562 If corresponding addresses was not found in the lookup tables, the packet
3563 will not be dropped and the search continues.
3564 .Ss XLAT464 CLAT translation
3565 XLAT464 CLAT NAT64 translator implements client-side stateless translation as
3566 defined in RFC6877 and is very similar to statless NAT64 translator
3568 Instead of lookup tables it uses one-to-one mapping between IPv4 and IPv6
3569 addresses using configured prefixes.
3570 This mode can be used as a replacement of DNS64 service for applications
3571 that are not using it (e.g. VoIP) allowing them to access IPv4-only Internet
3572 over IPv6-only networks with help of remote NAT64 translator.
3574 The CLAT NAT64 configuration command is the following:
3575 .Bd -ragged -offset indent
3584 The following parameters can be configured:
3585 .Bl -tag -width indent
3586 .It Cm clat_prefix Ar ipv6_prefix/length
3587 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3588 to represent source IPv4 addresses.
3589 .It Cm plat_prefix Ar ipv6_prefix/length
3590 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3591 to represent destination IPv4 addresses.
3592 This IPv6 prefix should be configured on a remote NAT64 translator.
3594 Turn on logging of all handled packets via BPF through
3598 Turn off logging of all handled packets via BPF.
3599 .It Cm allow_private
3600 Turn on processing private IPv4 addresses.
3603 instance will not process IPv4 packets with destination address from private
3604 ranges as defined in RFC1918.
3605 .It Cm -allow_private
3606 Turn off private address handling in
3611 Note that the behavior of CLAT translator with respect to not matched
3612 packets differs from stateful translator.
3613 If corresponding addresses were not matched against prefixes configured,
3614 the packet will not be dropped and the search continues.
3615 .Sh IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
3617 supports in-kernel IPv6-to-IPv6 network prefix translation as described
3621 should be loaded or kernel should has
3622 .Cm options IPFIREWALL_NPTV6
3623 to be able use NPTv6 translator.
3625 The NPTv6 configuration command is the following:
3626 .Bd -ragged -offset indent
3635 The following parameters can be configured:
3636 .Bl -tag -width indent
3637 .It Cm int_prefix Ar ipv6_prefix
3638 IPv6 prefix used in internal network.
3639 NPTv6 module translates source address when it matches this prefix.
3640 .It Cm ext_prefix Ar ipv6_prefix
3641 IPv6 prefix used in external network.
3642 NPTv6 module translates destination address when it matches this prefix.
3643 .It Cm ext_if Ar nic
3644 The NPTv6 module will use first global IPv6 address from interface
3647 It can be useful when IPv6 prefix of external network is dynamically obtained.
3651 options are mutually exclusive.
3652 .It Cm prefixlen Ar length
3653 The length of specified IPv6 prefixes.
3654 It must be in range from 8 to 64.
3657 Note that the prefix translation rules are silently ignored when IPv6 packet
3658 forwarding is disabled.
3659 To enable the packet forwarding, set the sysctl variable
3660 .Va net.inet6.ip6.forwarding
3663 To let the packet continue after being translated, set the sysctl variable
3664 .Va net.inet.ip.fw.one_pass
3667 Tunables can be set in
3673 before ipfw module gets loaded.
3674 .Bl -tag -width indent
3675 .It Va net.inet.ip.fw.default_to_accept : No 0
3676 Defines ipfw last rule behavior.
3677 This value overrides
3678 .Cd "options IPFW_DEFAULT_TO_(ACCEPT|DENY)"
3679 from kernel configuration file.
3680 .It Va net.inet.ip.fw.tables_max : No 128
3681 Defines number of tables available in ipfw.
3682 Number cannot exceed 65534.
3684 .Sh SYSCTL VARIABLES
3687 variables controls the behaviour of the firewall and
3689 .Pq Nm dummynet , bridge , sctp nat .
3690 These are shown below together with their default value
3691 (but always check with the
3693 command what value is actually in use) and meaning:
3694 .Bl -tag -width indent
3695 .It Va net.inet.ip.alias.sctp.accept_global_ootb_addip : No 0
3698 responds to receipt of global OOTB ASCONF-AddIP:
3699 .Bl -tag -width indent
3701 No response (unless a partially matching association exists -
3702 ports and vtags match but global address does not)
3705 will accept and process all OOTB global AddIP messages.
3708 Option 1 should never be selected as this forms a security risk.
3710 establish multiple fake associations by sending AddIP messages.
3711 .It Va net.inet.ip.alias.sctp.chunk_proc_limit : No 5
3712 Defines the maximum number of chunks in an SCTP packet that will be
3714 packet that matches an existing association.
3715 This value is enforced to be greater or equal than
3716 .Cm net.inet.ip.alias.sctp.initialising_chunk_proc_limit .
3718 a DoS risk yet setting too low a value may result in
3719 important control chunks in
3720 the packet not being located and parsed.
3721 .It Va net.inet.ip.alias.sctp.error_on_ootb : No 1
3724 responds to any Out-of-the-Blue (OOTB) packets with ErrorM packets.
3725 An OOTB packet is a packet that arrives with no existing association
3728 and is not an INIT or ASCONF-AddIP packet:
3729 .Bl -tag -width indent
3731 ErrorM is never sent in response to OOTB packets.
3733 ErrorM is only sent to OOTB packets received on the local side.
3735 ErrorM is sent to the local side and on the global side ONLY if there is a
3736 partial match (ports and vtags match but the source global IP does not).
3737 This value is only useful if the
3739 is tracking global IP addresses.
3741 ErrorM is sent in response to all OOTB packets on both
3742 the local and global side
3746 At the moment the default is 0, since the ErrorM packet is not yet
3747 supported by most SCTP stacks.
3748 When it is supported, and if not tracking
3749 global addresses, we recommend setting this value to 1 to allow
3750 multi-homed local hosts to function with the
3752 To track global addresses, we recommend setting this value to 2 to
3753 allow global hosts to be informed when they need to (re)send an
3755 Value 3 should never be chosen (except for debugging) as the
3757 will respond to all OOTB global packets (a DoS risk).
3758 .It Va net.inet.ip.alias.sctp.hashtable_size : No 2003
3759 Size of hash tables used for
3761 lookups (100 < prime_number > 1000001).
3764 size for any future created
3766 instance and therefore must be set prior to creating a
3769 The table sizes may be changed to suit specific needs.
3770 If there will be few
3771 concurrent associations, and memory is scarce, you may make these smaller.
3772 If there will be many thousands (or millions) of concurrent associations, you
3773 should make these larger.
3774 A prime number is best for the table size.
3776 update function will adjust your input value to the next highest prime number.
3777 .It Va net.inet.ip.alias.sctp.holddown_time : No 0
3778 Hold association in table for this many seconds after receiving a
3780 This allows endpoints to correct shutdown gracefully if a
3781 shutdown_complete is lost and retransmissions are required.
3782 .It Va net.inet.ip.alias.sctp.init_timer : No 15
3783 Timeout value while waiting for (INIT-ACK|AddIP-ACK).
3784 This value cannot be 0.
3785 .It Va net.inet.ip.alias.sctp.initialising_chunk_proc_limit : No 2
3786 Defines the maximum number of chunks in an SCTP packet that will be parsed when
3787 no existing association exists that matches that packet.
3789 will only be an INIT or ASCONF-AddIP packet.
3790 A higher value may become a DoS
3791 risk as malformed packets can consume processing resources.
3792 .It Va net.inet.ip.alias.sctp.param_proc_limit : No 25
3793 Defines the maximum number of parameters within a chunk that will be
3796 As for other similar sysctl variables, larger values pose a DoS risk.
3797 .It Va net.inet.ip.alias.sctp.log_level : No 0
3798 Level of detail in the system log messages (0 \- minimal, 1 \- event,
3799 2 \- info, 3 \- detail, 4 \- debug, 5 \- max debug).
3801 option in high loss environments.
3802 .It Va net.inet.ip.alias.sctp.shutdown_time : No 15
3803 Timeout value while waiting for SHUTDOWN-COMPLETE.
3804 This value cannot be 0.
3805 .It Va net.inet.ip.alias.sctp.track_global_addresses : No 0
3806 Enables/disables global IP address tracking within the
3809 upper limit on the number of addresses tracked for each association:
3810 .Bl -tag -width indent
3812 Global tracking is disabled
3814 Enables tracking, the maximum number of addresses tracked for each
3815 association is limited to this value
3818 This variable is fully dynamic, the new value will be adopted for all newly
3819 arriving associations, existing associations are treated
3820 as they were previously.
3821 Global tracking will decrease the number of collisions within the
3824 of increased processing load, memory usage, complexity, and possible
3827 problems in complex networks with multiple
3829 We recommend not tracking
3830 global IP addresses, this will still result in a fully functional
3832 .It Va net.inet.ip.alias.sctp.up_timer : No 300
3833 Timeout value to keep an association up with no traffic.
3834 This value cannot be 0.
3835 .It Va net.inet.ip.dummynet.codel.interval : No 100000
3838 AQM interval in microseconds.
3839 The value must be in the range 1..5000000.
3840 .It Va net.inet.ip.dummynet.codel.target : No 5000
3843 AQM target delay time in microseconds (the minimum acceptable persistent queue
3845 The value must be in the range 1..5000000.
3846 .It Va net.inet.ip.dummynet.expire : No 1
3847 Lazily delete dynamic pipes/queue once they have no pending traffic.
3848 You can disable this by setting the variable to 0, in which case
3849 the pipes/queues will only be deleted when the threshold is reached.
3850 .It Va net.inet.ip.dummynet.fqcodel.flows : No 1024
3851 Defines the default total number of flow queues (sub-queues) that
3853 creates and manages.
3854 The value must be in the range 1..65536.
3855 .It Va net.inet.ip.dummynet.fqcodel.interval : No 100000
3858 scheduler/AQM interval in microseconds.
3859 The value must be in the range 1..5000000.
3860 .It Va net.inet.ip.dummynet.fqcodel.limit : No 10240
3861 The default hard size limit (in unit of packet) of all queues managed by an
3865 The value must be in the range 1..20480.
3866 .It Va net.inet.ip.dummynet.fqcodel.quantum : No 1514
3867 The default quantum (credit) of the
3870 The value must be in the range 1..9000.
3871 .It Va net.inet.ip.dummynet.fqcodel.target : No 5000
3874 scheduler/AQM target delay time in microseconds (the minimum acceptable
3875 persistent queue delay).
3876 The value must be in the range 1..5000000.
3877 .It Va net.inet.ip.dummynet.fqpie.alpha : No 125
3880 parameter (scaled by 1000) for
3883 The value must be in the range 1..7000.
3884 .It Va net.inet.ip.dummynet.fqpie.beta : No 1250
3887 parameter (scaled by 1000) for
3890 The value must be in the range 1..7000.
3891 .It Va net.inet.ip.dummynet.fqpie.flows : No 1024
3892 Defines the default total number of flow queues (sub-queues) that
3894 creates and manages.
3895 The value must be in the range 1..65536.
3896 .It Va net.inet.ip.dummynet.fqpie.limit : No 10240
3897 The default hard size limit (in unit of packet) of all queues managed by an
3901 The value must be in the range 1..20480.
3902 .It Va net.inet.ip.dummynet.fqpie.max_burst : No 150000
3903 The default maximum period of microseconds that
3905 scheduler/AQM does not drop/mark packets.
3906 The value must be in the range 1..10000000.
3907 .It Va net.inet.ip.dummynet.fqpie.max_ecnth : No 99
3908 The default maximum ECN probability threshold (scaled by 1000) for
3911 The value must be in the range 1..7000.
3912 .It Va net.inet.ip.dummynet.fqpie.quantum : No 1514
3913 The default quantum (credit) of the
3916 The value must be in the range 1..9000.
3917 .It Va net.inet.ip.dummynet.fqpie.target : No 15000
3922 in unit of microsecond.
3923 The value must be in the range 1..5000000.
3924 .It Va net.inet.ip.dummynet.fqpie.tupdate : No 15000
3929 in unit of microsecond.
3930 The value must be in the range 1..5000000.
3931 .It Va net.inet.ip.dummynet.hash_size : No 64
3932 Default size of the hash table used for dynamic pipes/queues.
3933 This value is used when no
3935 option is specified when configuring a pipe/queue.
3936 .It Va net.inet.ip.dummynet.io_fast : No 0
3937 If set to a non-zero value,
3942 operation (see above) is enabled.
3943 .It Va net.inet.ip.dummynet.io_pkt
3944 Number of packets passed to
3946 .It Va net.inet.ip.dummynet.io_pkt_drop
3947 Number of packets dropped by
3949 .It Va net.inet.ip.dummynet.io_pkt_fast
3950 Number of packets bypassed by the
3953 .It Va net.inet.ip.dummynet.max_chain_len : No 16
3954 Target value for the maximum number of pipes/queues in a hash bucket.
3956 .Cm max_chain_len*hash_size
3957 is used to determine the threshold over which empty pipes/queues
3958 will be expired even when
3959 .Cm net.inet.ip.dummynet.expire=0 .
3960 .It Va net.inet.ip.dummynet.red_lookup_depth : No 256
3961 .It Va net.inet.ip.dummynet.red_avg_pkt_size : No 512
3962 .It Va net.inet.ip.dummynet.red_max_pkt_size : No 1500
3963 Parameters used in the computations of the drop probability
3964 for the RED algorithm.
3965 .It Va net.inet.ip.dummynet.pie.alpha : No 125
3968 parameter (scaled by 1000) for
3971 The value must be in the range 1..7000.
3972 .It Va net.inet.ip.dummynet.pie.beta : No 1250
3975 parameter (scaled by 1000) for
3978 The value must be in the range 1..7000.
3979 .It Va net.inet.ip.dummynet.pie.max_burst : No 150000
3980 The default maximum period of microseconds that
3982 AQM does not drop/mark packets.
3983 The value must be in the range 1..10000000.
3984 .It Va net.inet.ip.dummynet.pie.max_ecnth : No 99
3985 The default maximum ECN probability threshold (scaled by 1000) for
3988 The value must be in the range 1..7000.
3989 .It Va net.inet.ip.dummynet.pie.target : No 15000
3994 AQM in unit of microsecond.
3995 The value must be in the range 1..5000000.
3996 .It Va net.inet.ip.dummynet.pie.tupdate : No 15000
4001 AQM in unit of microsecond.
4002 The value must be in the range 1..5000000.
4003 .It Va net.inet.ip.dummynet.pipe_byte_limit : No 1048576
4004 .It Va net.inet.ip.dummynet.pipe_slot_limit : No 100
4005 The maximum queue size that can be specified in bytes or packets.
4006 These limits prevent accidental exhaustion of resources such as mbufs.
4007 If you raise these limits,
4008 you should make sure the system is configured so that sufficient resources
4010 .It Va net.inet.ip.fw.autoinc_step : No 100
4011 Delta between rule numbers when auto-generating them.
4012 The value must be in the range 1..1000.
4013 .It Va net.inet.ip.fw.curr_dyn_buckets : Va net.inet.ip.fw.dyn_buckets
4014 The current number of buckets in the hash table for dynamic rules
4016 .It Va net.inet.ip.fw.debug : No 1
4017 Controls debugging messages produced by
4019 .It Va net.inet.ip.fw.default_rule : No 65535
4020 The default rule number (read-only).
4022 .Nm , the default rule is the last one, so its number
4023 can also serve as the highest number allowed for a rule.
4024 .It Va net.inet.ip.fw.dyn_buckets : No 256
4025 The number of buckets in the hash table for dynamic rules.
4026 Must be a power of 2, up to 65536.
4027 It only takes effect when all dynamic rules have expired, so you
4028 are advised to use a
4030 command to make sure that the hash table is resized.
4031 .It Va net.inet.ip.fw.dyn_count : No 3
4032 Current number of dynamic rules
4034 .It Va net.inet.ip.fw.dyn_keepalive : No 1
4035 Enables generation of keepalive packets for
4037 rules on TCP sessions.
4038 A keepalive is generated to both
4039 sides of the connection every 5 seconds for the last 20
4040 seconds of the lifetime of the rule.
4041 .It Va net.inet.ip.fw.dyn_max : No 8192
4042 Maximum number of dynamic rules.
4043 When you hit this limit, no more dynamic rules can be
4044 installed until old ones expire.
4045 .It Va net.inet.ip.fw.dyn_ack_lifetime : No 300
4046 .It Va net.inet.ip.fw.dyn_syn_lifetime : No 20
4047 .It Va net.inet.ip.fw.dyn_fin_lifetime : No 1
4048 .It Va net.inet.ip.fw.dyn_rst_lifetime : No 1
4049 .It Va net.inet.ip.fw.dyn_udp_lifetime : No 5
4050 .It Va net.inet.ip.fw.dyn_short_lifetime : No 30
4051 These variables control the lifetime, in seconds, of dynamic
4053 Upon the initial SYN exchange the lifetime is kept short,
4054 then increased after both SYN have been seen, then decreased
4055 again during the final FIN exchange or when a RST is received.
4057 .Em dyn_fin_lifetime
4059 .Em dyn_rst_lifetime
4060 must be strictly lower than 5 seconds, the period of
4061 repetition of keepalives.
4062 The firewall enforces that.
4063 .It Va net.inet.ip.fw.dyn_keep_states : No 0
4064 Keep dynamic states on rule/set deletion.
4065 States are relinked to default rule (65535).
4066 This can be handly for ruleset reload.
4067 Turned off by default.
4068 .It Va net.inet.ip.fw.enable : No 1
4069 Enables the firewall.
4070 Setting this variable to 0 lets you run your machine without
4071 firewall even if compiled in.
4072 .It Va net.inet6.ip6.fw.enable : No 1
4073 provides the same functionality as above for the IPv6 case.
4074 .It Va net.inet.ip.fw.one_pass : No 1
4075 When set, the packet exiting from the
4079 node is not passed though the firewall again.
4080 Otherwise, after an action, the packet is
4081 reinjected into the firewall at the next rule.
4082 .It Va net.inet.ip.fw.tables_max : No 128
4083 Maximum number of tables.
4084 .It Va net.inet.ip.fw.verbose : No 1
4085 Enables verbose messages.
4086 .It Va net.inet.ip.fw.verbose_limit : No 0
4087 Limits the number of messages produced by a verbose firewall.
4088 .It Va net.inet6.ip6.fw.deny_unknown_exthdrs : No 1
4089 If enabled packets with unknown IPv6 Extension Headers will be denied.
4090 .It Va net.link.ether.ipfw : No 0
4091 Controls whether layer-2 packets are passed to
4094 .It Va net.link.bridge.ipfw : No 0
4095 Controls whether bridged packets are passed to
4098 .It Va net.inet.ip.fw.nat64_debug : No 0
4099 Controls debugging messages produced by
4102 .It Va net.inet.ip.fw.nat64_direct_output : No 0
4103 Controls the output method used by
4106 .Bl -tag -width indent
4108 A packet is handled by
4111 First time an original packet is handled by
4116 Then translated packet is queued via netisr to input processing again.
4118 A packet is handled by
4120 only once, and after translation it will be pushed directly to outgoing
4124 .Sh INTERNAL DIAGNOSTICS
4125 There are some commands that may be useful to understand current state
4126 of certain subsystems inside kernel module.
4127 These commands provide debugging output which may change without notice.
4129 Currently the following commands are available as
4132 .Bl -tag -width indent
4134 Lists all interface which are currently tracked by
4136 with their in-kernel status.
4138 List all table lookup algorithms currently available.
4141 There are far too many possible uses of
4143 so this Section will only give a small set of examples.
4144 .Ss BASIC PACKET FILTERING
4145 This command adds an entry which denies all tcp packets from
4146 .Em cracker.evil.org
4147 to the telnet port of
4149 from being forwarded by the host:
4151 .Dl "ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet"
4153 This one disallows any connection from the entire cracker's
4156 .Dl "ipfw add deny ip from 123.45.67.0/24 to my.host.org"
4158 A first and efficient way to limit access (not using dynamic rules)
4159 is the use of the following rules:
4161 .Dl "ipfw add allow tcp from any to any established"
4162 .Dl "ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup"
4163 .Dl "ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup"
4165 .Dl "ipfw add deny tcp from any to any"
4167 The first rule will be a quick match for normal TCP packets,
4168 but it will not match the initial SYN packet, which will be
4171 rules only for selected source/destination pairs.
4172 All other SYN packets will be rejected by the final
4176 If you administer one or more subnets, you can take advantage
4177 of the address sets and or-blocks and write extremely
4178 compact rulesets which selectively enable services to blocks
4179 of clients, as below:
4181 .Dl "goodguys=\*q{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }\*q"
4182 .Dl "badguys=\*q10.1.2.0/24{8,38,60}\*q"
4184 .Dl "ipfw add allow ip from ${goodguys} to any"
4185 .Dl "ipfw add deny ip from ${badguys} to any"
4186 .Dl "... normal policies ..."
4190 option could be used to do automated anti-spoofing by adding the
4191 following to the top of a ruleset:
4193 .Dl "ipfw add deny ip from any to any not verrevpath in"
4195 This rule drops all incoming packets that appear to be coming to the
4196 system on the wrong interface.
4197 For example, a packet with a source
4198 address belonging to a host on a protected internal network would be
4199 dropped if it tried to enter the system from an external interface.
4203 option could be used to do similar but more restricted anti-spoofing
4204 by adding the following to the top of a ruleset:
4206 .Dl "ipfw add deny ip from any to any not antispoof in"
4208 This rule drops all incoming packets that appear to be coming from another
4209 directly connected system but on the wrong interface.
4210 For example, a packet with a source address of
4211 .Li 192.168.0.0/24 ,
4220 option could be used to (re)mark user traffic,
4221 by adding the following to the appropriate place in ruleset:
4223 .Dl "ipfw add setdscp be ip from any to any dscp af11,af21"
4224 .Ss SELECTIVE MIRRORING
4225 If your network has network traffic analyzer
4226 connected to your host directly via dedicated interface
4227 or remotely via RSPAN vlan, you can selectively mirror
4228 some Ethernet layer2 frames to the analyzer.
4230 First, make sure your firewall is already configured and runs.
4231 Then, enable layer2 processing if not already enabled:
4233 .Dl "sysctl net.link.ether.ipfw=1"
4235 Next, load needed additional kernel modules:
4237 .Dl "kldload ng_ether ng_ipfw"
4239 Optionally, make system load these modules automatically
4242 .Dl sysrc kld_list+="ng_ether ng_ipfw"
4246 kernel module to transmit mirrored copies of layer2 frames
4247 out via vlan900 interface:
4249 .Dl "ngctl connect ipfw: vlan900: 1 lower"
4251 Think of "1" here as of "mirroring instance index" and vlan900 is its
4253 You can have arbitrary number of instances.
4258 At last, actually start mirroring of selected frames using "instance 1".
4259 For frames incoming from em0 interface:
4261 .Dl "ipfw add ngtee 1 ip from any to 192.168.0.1 layer2 in recv em0"
4263 For frames outgoing to em0 interface:
4265 .Dl "ipfw add ngtee 1 ip from any to 192.168.0.1 layer2 out xmit em0"
4267 For both incoming and outgoing frames while flowing through em0:
4269 .Dl "ipfw add ngtee 1 ip from any to 192.168.0.1 layer2 via em0"
4271 Make sure you do not perform mirroring for already duplicated frames
4272 or kernel may hang as there is no safety net.
4274 In order to protect a site from flood attacks involving fake
4275 TCP packets, it is safer to use dynamic rules:
4277 .Dl "ipfw add check-state"
4278 .Dl "ipfw add deny tcp from any to any established"
4279 .Dl "ipfw add allow tcp from my-net to any setup keep-state"
4281 This will let the firewall install dynamic rules only for
4282 those connection which start with a regular SYN packet coming
4283 from the inside of our network.
4284 Dynamic rules are checked when encountering the first
4293 rule should usually be placed near the beginning of the
4294 ruleset to minimize the amount of work scanning the ruleset.
4295 Your mileage may vary.
4297 For more complex scenarios with dynamic rules
4301 can be used to precisely control creation and checking of dynamic rules.
4302 Example of usage of these options are provided in
4303 .Sx NETWORK ADDRESS TRANSLATION (NAT)
4306 To limit the number of connections a user can open
4307 you can use the following type of rules:
4309 .Dl "ipfw add allow tcp from my-net/24 to any setup limit src-addr 10"
4310 .Dl "ipfw add allow tcp from any to me setup limit src-addr 4"
4312 The former (assuming it runs on a gateway) will allow each host
4313 on a /24 network to open at most 10 TCP connections.
4314 The latter can be placed on a server to make sure that a single
4315 client does not use more than 4 simultaneous connections.
4318 stateful rules can be subject to denial-of-service attacks
4319 by a SYN-flood which opens a huge number of dynamic rules.
4320 The effects of such attacks can be partially limited by
4323 variables which control the operation of the firewall.
4325 Here is a good usage of the
4327 command to see accounting records and timestamp information:
4331 or in short form without timestamps:
4335 which is equivalent to:
4339 Next rule diverts all incoming packets from 192.168.2.0/24
4340 to divert port 5000:
4342 .Dl ipfw divert 5000 ip from 192.168.2.0/24 to any in
4344 The following rules show some of the applications of
4348 for simulations and the like.
4350 This rule drops random incoming packets with a probability
4353 .Dl "ipfw add prob 0.05 deny ip from any to any in"
4355 A similar effect can be achieved making use of
4359 .Dl "dnctl add pipe 10 ip from any to any"
4360 .Dl "dnctl pipe 10 config plr 0.05"
4362 We can use pipes to artificially limit bandwidth, e.g.\& on a
4363 machine acting as a router, if we want to limit traffic from
4364 local clients on 192.168.2.0/24 we do:
4366 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
4367 .Dl "dnctl pipe 1 config bw 300Kbit/s queue 50KBytes"
4369 note that we use the
4371 modifier so that the rule is not used twice.
4372 Remember in fact that
4374 rules are checked both on incoming and outgoing packets.
4376 Should we want to simulate a bidirectional link with bandwidth
4377 limitations, the correct way is the following:
4379 .Dl "ipfw add pipe 1 ip from any to any out"
4380 .Dl "ipfw add pipe 2 ip from any to any in"
4381 .Dl "dnctl pipe 1 config bw 64Kbit/s queue 10Kbytes"
4382 .Dl "dnctl pipe 2 config bw 64Kbit/s queue 10Kbytes"
4384 The above can be very useful, e.g.\& if you want to see how
4385 your fancy Web page will look for a residential user who
4386 is connected only through a slow link.
4387 You should not use only one pipe for both directions, unless
4388 you want to simulate a half-duplex medium (e.g.\& AppleTalk,
4390 It is not necessary that both pipes have the same configuration,
4391 so we can also simulate asymmetric links.
4393 Should we want to verify network performance with the RED queue
4394 management algorithm:
4396 .Dl "ipfw add pipe 1 ip from any to any"
4397 .Dl "dnctl pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1"
4399 Another typical application of the traffic shaper is to
4400 introduce some delay in the communication.
4401 This can significantly affect applications which do a lot of Remote
4402 Procedure Calls, and where the round-trip-time of the
4403 connection often becomes a limiting factor much more than
4406 .Dl "ipfw add pipe 1 ip from any to any out"
4407 .Dl "ipfw add pipe 2 ip from any to any in"
4408 .Dl "dnctl pipe 1 config delay 250ms bw 1Mbit/s"
4409 .Dl "dnctl pipe 2 config delay 250ms bw 1Mbit/s"
4411 Per-flow queueing can be useful for a variety of purposes.
4412 A very simple one is counting traffic:
4414 .Dl "ipfw add pipe 1 tcp from any to any"
4415 .Dl "ipfw add pipe 1 udp from any to any"
4416 .Dl "ipfw add pipe 1 ip from any to any"
4417 .Dl "dnctl pipe 1 config mask all"
4419 The above set of rules will create queues (and collect
4420 statistics) for all traffic.
4421 Because the pipes have no limitations, the only effect is
4422 collecting statistics.
4423 Note that we need 3 rules, not just the last one, because
4426 tries to match IP packets it will not consider ports, so we
4427 would not see connections on separate ports as different
4430 A more sophisticated example is limiting the outbound traffic
4431 on a net with per-host limits, rather than per-network limits:
4433 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
4434 .Dl "ipfw add pipe 2 ip from any to 192.168.2.0/24 in"
4435 .Dl "dnctl pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
4436 .Dl "dnctl pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
4438 In the following example, we need to create several traffic bandwidth
4439 classes and we need different hosts/networks to fall into different classes.
4440 We create one pipe for each class and configure them accordingly.
4441 Then we create a single table and fill it with IP subnets and addresses.
4442 For each subnet/host we set the argument equal to the number of the pipe
4444 Then we classify traffic using a single rule:
4446 .Dl "dnctl pipe 1 config bw 1000Kbyte/s"
4447 .Dl "dnctl pipe 4 config bw 4000Kbyte/s"
4449 .Dl "ipfw table T1 create type addr"
4450 .Dl "ipfw table T1 add 192.168.2.0/24 1"
4451 .Dl "ipfw table T1 add 192.168.0.0/27 4"
4452 .Dl "ipfw table T1 add 192.168.0.2 1"
4454 .Dl "ipfw add pipe tablearg ip from 'table(T1)' to any"
4458 action, the table entries may include hostnames and IP addresses.
4460 .Dl "ipfw table T2 create type addr valtype ipv4"
4461 .Dl "ipfw table T2 add 192.168.2.0/24 10.23.2.1"
4462 .Dl "ipfw table T2 add 192.168.0.0/27 router1.dmz"
4464 .Dl "ipfw add 100 fwd tablearg ip from any to 'table(T2)'"
4466 In the following example per-interface firewall is created:
4468 .Dl "ipfw table IN create type iface valtype skipto,fib"
4469 .Dl "ipfw table IN add vlan20 12000,12"
4470 .Dl "ipfw table IN add vlan30 13000,13"
4471 .Dl "ipfw table OUT create type iface valtype skipto"
4472 .Dl "ipfw table OUT add vlan20 22000"
4473 .Dl "ipfw table OUT add vlan30 23000"
4475 .Dl "ipfw add 100 setfib tablearg ip from any to any recv 'table(IN)' in"
4476 .Dl "ipfw add 200 skipto tablearg ip from any to any recv 'table(IN)' in"
4477 .Dl "ipfw add 300 skipto tablearg ip from any to any xmit 'table(OUT)' out"
4479 The following example illustrate usage of flow tables:
4481 .Dl "ipfw table fl create type flow:src-ip,proto,dst-ip,dst-port"
4482 .Dl "ipfw table fl add 2a02:6b8:77::88,tcp,2a02:6b8:77::99,80 11"
4483 .Dl "ipfw table fl add 10.0.0.1,udp,10.0.0.2,53 12"
4485 .Dl "ipfw add 100 allow ip from any to any flow 'table(fl,11)' recv ix0"
4487 To add a set of rules atomically, e.g.\& set 18:
4489 .Dl "ipfw set disable 18"
4490 .Dl "ipfw add NN set 18 ... # repeat as needed"
4491 .Dl "ipfw set enable 18"
4493 To delete a set of rules atomically the command is simply:
4495 .Dl "ipfw delete set 18"
4497 To test a ruleset and disable it and regain control if something goes wrong:
4499 .Dl "ipfw set disable 18"
4500 .Dl "ipfw add NN set 18 ... # repeat as needed"
4501 .Dl "ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18"
4503 Here if everything goes well, you press control-C before the "sleep"
4504 terminates, and your ruleset will be left active.
4505 Otherwise, e.g.\& if
4506 you cannot access your box, the ruleset will be disabled after
4507 the sleep terminates thus restoring the previous situation.
4509 To show rules of the specific set:
4511 .Dl "ipfw set 18 show"
4513 To show rules of the disabled set:
4515 .Dl "ipfw -S set 18 show"
4517 To clear a specific rule counters of the specific set:
4519 .Dl "ipfw set 18 zero NN"
4521 To delete a specific rule of the specific set:
4523 .Dl "ipfw set 18 delete NN"
4524 .Ss NAT, REDIRECT AND LSNAT
4525 First redirect all the traffic to nat instance 123:
4527 .Dl "ipfw add nat 123 all from any to any"
4529 Then to configure nat instance 123 to alias all the outgoing traffic with ip
4530 192.168.0.123, blocking all incoming connections, trying to keep
4531 same ports on both sides, clearing aliasing table on address change
4532 and keeping a log of traffic/link statistics:
4534 .Dl "ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports"
4536 Or to change address of instance 123, aliasing table will be cleared (see
4539 .Dl "ipfw nat 123 config ip 10.0.0.1"
4541 To see configuration of nat instance 123:
4543 .Dl "ipfw nat 123 show config"
4545 To show logs of all instances:
4547 .Dl "ipfw nat show log"
4549 To see configurations of all instances:
4551 .Dl "ipfw nat show config"
4553 Or a redirect rule with mixed modes could looks like:
4554 .Bd -literal -offset 2n
4555 ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66
4556 redirect_port tcp 192.168.0.1:80 500
4557 redirect_proto udp 192.168.1.43 192.168.1.1
4558 redirect_addr 192.168.0.10,192.168.0.11
4560 redirect_port tcp 192.168.0.1:80,192.168.0.10:22
4564 or it could be split in:
4565 .Bd -literal -offset 2n
4566 ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66
4567 ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500
4568 ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1
4569 ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12
4571 ipfw nat 5 config redirect_port tcp
4572 192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500
4575 Sometimes you may want to mix NAT and dynamic rules.
4576 It could be achieved with
4581 Problem is, you need to create dynamic rule before NAT and check it
4582 after NAT actions (or vice versa) to have consistent addresses and ports.
4585 option will trigger activation of existing dynamic state, and action of such
4586 rule will be performed as soon as rule is matched.
4589 rule packet need to be passed to NAT, not allowed as soon is possible.
4591 There is example of set of rules to achieve this.
4592 Bear in mind that this is example only and it is not very useful by itself.
4594 On way out, after all checks place this rules:
4596 .Dl "ipfw add allow record-state skip-action"
4597 .Dl "ipfw add nat 1"
4599 And on way in there should be something like this:
4601 .Dl "ipfw add nat 1"
4602 .Dl "ipfw add check-state"
4604 Please note, that first rule on way out doesn't allow packet and doesn't
4605 execute existing dynamic rules.
4606 All it does, create new dynamic rule with
4608 action, if it is not created yet.
4609 Later, this dynamic rule is used on way in by
4612 .Ss CONFIGURING CODEL, PIE, FQ-CODEL and FQ-PIE AQM
4616 AQM can be configured for
4626 AQM using default configuration for traffic from 192.168.0.0/24 and 1Mbits/s
4629 .Dl "dnctl pipe 1 config bw 1mbits/s codel"
4630 .Dl "ipfw add 100 pipe 1 ip from 192.168.0.0/24 to any"
4636 AQM using different configurations parameters for traffic from
4637 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4639 .Dl "dnctl pipe 1 config bw 1mbits/s"
4640 .Dl "dnctl queue 1 config pipe 1 codel target 8ms interval 160ms ecn"
4641 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4647 AQM using default configuration for traffic from 192.168.0.0/24 and 1Mbits/s
4650 .Dl "dnctl pipe 1 config bw 1mbits/s pie"
4651 .Dl "ipfw add 100 pipe 1 ip from 192.168.0.0/24 to any"
4657 AQM using different configuration parameters for traffic from
4658 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4660 .Dl "dnctl pipe 1 config bw 1mbits/s"
4661 .Dl "dnctl queue 1 config pipe 1 pie target 20ms tupdate 30ms ecn"
4662 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4667 AQM can be configured for
4673 scheduler using different configurations parameters for traffic from
4674 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4676 .Dl "dnctl pipe 1 config bw 1mbits/s"
4677 .Dl "dnctl sched 1 config pipe 1 type fq_codel"
4678 .Dl "dnctl queue 1 config sched 1"
4679 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4683 default configuration for a
4685 such as disable ECN and change the
4689 .Dl "dnctl sched 1 config pipe 1 type fq_codel target 10ms noecn"
4695 scheduler using different configurations parameters for traffic from
4696 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4698 .Dl "dnctl pipe 1 config bw 1mbits/s"
4699 .Dl "dnctl sched 1 config pipe 1 type fq_pie"
4700 .Dl "dnctl queue 1 config sched 1"
4701 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4703 The configurations of
4706 can be changed in a similar way as for
4730 utility first appeared in
4735 Stateful extensions were introduced in
4738 was introduced in Summer 2002.
4740 .An Ugen J. S. Antsilevich ,
4741 .An Poul-Henning Kamp ,
4745 .An Rasool Al-Saadi .
4748 API based upon code written by
4752 Dummynet has been introduced by Luigi Rizzo in 1997-1998.
4754 Some early work (1999-2000) on the
4756 traffic shaper supported by Akamba Corp.
4758 The ipfw core (ipfw2) has been completely redesigned and
4759 reimplemented by Luigi Rizzo in summer 2002.
4762 options have been added by various developers over the years.
4765 In-kernel NAT support written by
4766 .An Paolo Pisati Aq Mt piso@FreeBSD.org
4767 as part of a Summer of Code 2005 project.
4771 support has been developed by
4772 .An The Centre for Advanced Internet Architectures (CAIA) Aq http://www.caia.swin.edu.au .
4773 The primary developers and maintainers are David Hayes and Jason But.
4774 For further information visit:
4775 .Aq http://www.caia.swin.edu.au/urp/SONATA
4777 Delay profiles have been developed by Alessandro Cerri and
4778 Luigi Rizzo, supported by the
4779 European Commission within Projects Onelab and Onelab2.
4781 CoDel, PIE, FQ-CoDel and FQ-PIE AQM for Dummynet have been implemented by
4782 .An The Centre for Advanced Internet Architectures (CAIA)
4783 in 2016, supported by The Comcast Innovation Fund.
4784 The primary developer is
4787 The syntax has grown over the years and sometimes it might be confusing.
4788 Unfortunately, backward compatibility prevents cleaning up mistakes
4789 made in the definition of the syntax.
4793 Misconfiguring the firewall can put your computer in an unusable state,
4794 possibly shutting down network services and requiring console access to
4795 regain control of it.
4797 Incoming packet fragments diverted by
4799 are reassembled before delivery to the socket.
4800 The action used on those packet is the one from the
4801 rule which matches the first fragment of the packet.
4803 Packets diverted to userland, and then reinserted by a userland process
4804 may lose various packet attributes.
4805 The packet source interface name
4806 will be preserved if it is shorter than 8 bytes and the userland process
4807 saves and reuses the sockaddr_in
4810 otherwise, it may be lost.
4811 If a packet is reinserted in this manner, later rules may be incorrectly
4812 applied, making the order of
4814 rules in the rule sequence very important.
4816 Dummynet drops all packets with IPv6 link-local addresses.
4822 may not behave as expected.
4823 In particular, incoming SYN packets may
4824 have no uid or gid associated with them since they do not yet belong
4825 to a TCP connection, and the uid/gid associated with a packet may not
4826 be as expected if the associated process calls
4828 or similar system calls.
4830 Rule syntax is subject to the command line environment and some patterns
4831 may need to be escaped with the backslash character
4832 or quoted appropriately.
4834 Due to the architecture of
4836 ipfw nat is not compatible with the TCP segmentation offloading (TSO).
4837 Thus, to reliably nat your network traffic, please disable TSO
4841 ICMP error messages are not implicitly matched by dynamic rules
4842 for the respective conversations.
4843 To avoid failures of network error detection and path MTU discovery,
4844 ICMP error messages may need to be allowed explicitly through static
4851 actions may lead to confusing behaviour if ruleset has mistakes,
4852 and/or interaction with other subsystems (netgraph, dummynet, etc.) is used.
4853 One possible case for this is packet leaving
4855 in subroutine on the input pass, while later on output encountering unpaired
4858 As the call stack is kept intact after input pass, packet will suddenly
4859 return to the rule number used on input pass, not on output one.
4860 Order of processing should be checked carefully to avoid such mistakes.
projects / FreeBSD / FreeBSD.git / blob