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
46 .Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
49 .Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
53 .Cm table Ar number Cm add Ar addr Ns Oo / Ns Ar masklen Oc Op Ar value
55 .Cm table Ar number Cm delete Ar addr Ns Op / Ns Ar masklen
65 .Ss DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)
67 .Brq Cm pipe | queue | sched
73 .Brq Cm pipe | queue | sched
74 .Brq Cm delete | list | show
97 utility is the user interface for controlling the
101 traffic shaper/packet scheduler, and the
102 in-kernel NAT services.
104 A firewall configuration, or
108 numbered from 1 to 65535.
109 Packets are passed to the firewall
110 from a number of different places in the protocol stack
111 (depending on the source and destination of the packet,
112 it is possible for the firewall to be
113 invoked multiple times on the same packet).
114 The packet passed to the firewall is compared
115 against each of the rules in the
118 (multiple rules with the same number are permitted, in which case
119 they are processed in order of insertion).
120 When a match is found, the action corresponding to the
121 matching rule is performed.
123 Depending on the action and certain system settings, packets
124 can be reinjected into the firewall at some rule after the
125 matching one for further processing.
127 A ruleset always includes a
129 rule (numbered 65535) which cannot be modified or deleted,
130 and matches all packets.
131 The action associated with the
137 depending on how the kernel is configured.
139 If the ruleset includes one or more rules with the
144 the firewall will have a
146 behaviour, i.e., upon a match it will create
148 i.e. rules that match packets with the same 5-tuple
149 (protocol, source and destination addresses and ports)
150 as the packet which caused their creation.
151 Dynamic rules, which have a limited lifetime, are checked
152 at the first occurrence of a
157 rule, and are typically used to open the firewall on-demand to
158 legitimate traffic only.
160 .Sx STATEFUL FIREWALL
163 Sections below for more information on the stateful behaviour of
166 All rules (including dynamic ones) have a few associated counters:
167 a packet count, a byte count, a log count and a timestamp
168 indicating the time of the last match.
169 Counters can be displayed or reset with
173 Each rule belongs to one of 32 different
177 commands to atomically manipulate sets, such as enable,
178 disable, swap sets, move all rules in a set to another
179 one, delete all rules in a set.
180 These can be useful to
181 install temporary configurations, or to test them.
184 for more information on
188 Rules can be added with the
190 command; deleted individually or in groups with the
192 command, and globally (except those in set 31) with the
194 command; displayed, optionally with the content of the
200 Finally, counters can be reset with the
207 The following general options are available when invoking
209 .Bl -tag -width indent
211 Show counter values when listing rules.
214 command implies this option.
216 Only show the action and the comment, not the body of a rule.
220 When entering or showing rules, print them in compact form,
221 i.e., omitting the "ip from any to any" string
222 when this does not carry any additional information.
224 When listing, show dynamic rules in addition to static ones.
228 is specified, also show expired dynamic rules.
230 Do not ask for confirmation for commands that can cause problems
233 If there is no tty associated with the process, this is implied.
235 When listing a table (see the
237 section below for more information on lookup tables), format values
238 as IP addresses. By default, values are shown as integers.
240 Only check syntax of the command strings, without actually passing
243 Try to resolve addresses and service names in output.
245 Be quiet when executing the
255 This is useful when updating rulesets by executing multiple
259 .Ql sh\ /etc/rc.firewall ) ,
260 or by processing a file with many
262 rules across a remote login session.
263 It also stops a table add or delete
264 from failing if the entry already exists or is not present.
266 The reason why this option may be important is that
267 for some of these actions,
269 may print a message; if the action results in blocking the
270 traffic to the remote client,
271 the remote login session will be closed
272 and the rest of the ruleset will not be processed.
273 Access to the console would then be required to recover.
275 When listing rules, show the
277 each rule belongs to.
278 If this flag is not specified, disabled rules will not be
281 When listing pipes, sort according to one of the four
282 counters (total or current packets or bytes).
284 When listing, show last match timestamp converted with ctime().
286 When listing, show last match timestamp as seconds from the epoch.
287 This form can be more convenient for postprocessing by scripts.
290 .Ss LIST OF RULES AND PREPROCESSING
291 To ease configuration, rules can be put into a file which is
294 as shown in the last synopsis line.
298 The file will be read line by line and applied as arguments to the
302 Optionally, a preprocessor can be specified using
306 is to be piped through.
307 Useful preprocessors include
313 does not start with a slash
315 as its first character, the usual
317 name search is performed.
318 Care should be taken with this in environments where not all
319 file systems are mounted (yet) by the time
321 is being run (e.g.\& when they are mounted over NFS).
324 has been specified, any additional arguments are passed on to the preprocessor
326 This allows for flexible configuration files (like conditionalizing
327 them on the local hostname) and the use of macros to centralize
328 frequently required arguments like IP addresses.
330 .Ss TRAFFIC SHAPER CONFIGURATION
336 commands are used to configure the traffic shaper and packet scheduler.
338 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
339 Section below for details.
341 If the world and the kernel get out of sync the
343 ABI may break, preventing you from being able to add any rules.
345 adversely effect the booting process.
350 to temporarily disable the firewall to regain access to the network,
351 allowing you to fix the problem.
353 A packet is checked against the active ruleset in multiple places
354 in the protocol stack, under control of several sysctl variables.
355 These places and variables are shown below, and it is important to
356 have this picture in mind in order to design a correct ruleset.
357 .Bd -literal -offset indent
360 +----------->-----------+
362 [ip(6)_input] [ip(6)_output] net.inet(6).ip(6).fw.enable=1
365 [ether_demux] [ether_output_frame] net.link.ether.ipfw=1
367 +-->--[bdg_forward]-->--+ net.link.bridge.ipfw=1
373 times the same packet goes through the firewall can
374 vary between 0 and 4 depending on packet source and
375 destination, and system configuration.
377 Note that as packets flow through the stack, headers can be
378 stripped or added to it, and so they may or may not be available
380 E.g., incoming packets will include the MAC header when
384 but the same packets will have the MAC header stripped off when
391 Also note that each packet is always checked against the complete ruleset,
392 irrespective of the place where the check occurs, or the source of the packet.
393 If a rule contains some match patterns or actions which are not valid
394 for the place of invocation (e.g.\& trying to match a MAC header within
398 the match pattern will not match, but a
400 operator in front of such patterns
404 match on those packets.
405 It is thus the responsibility of
406 the programmer, if necessary, to write a suitable ruleset to
407 differentiate among the possible places.
409 rules can be useful here, as an example:
410 .Bd -literal -offset indent
411 # packets from ether_demux or bdg_forward
412 ipfw add 10 skipto 1000 all from any to any layer2 in
413 # packets from ip_input
414 ipfw add 10 skipto 2000 all from any to any not layer2 in
415 # packets from ip_output
416 ipfw add 10 skipto 3000 all from any to any not layer2 out
417 # packets from ether_output_frame
418 ipfw add 10 skipto 4000 all from any to any layer2 out
421 (yes, at the moment there is no way to differentiate between
422 ether_demux and bdg_forward).
424 In general, each keyword or argument must be provided as
425 a separate command line argument, with no leading or trailing
427 Keywords are case-sensitive, whereas arguments may
428 or may not be case-sensitive depending on their nature
429 (e.g.\& uid's are, hostnames are not).
431 Some arguments (e.g. port or address lists) are comma-separated
433 In this case, spaces after commas ',' are allowed to make
434 the line more readable.
435 You can also put the entire
436 command (including flags) into a single argument.
437 E.g., the following forms are equivalent:
438 .Bd -literal -offset indent
439 ipfw -q add deny src-ip 10.0.0.0/24,127.0.0.1/8
440 ipfw -q add deny src-ip 10.0.0.0/24, 127.0.0.1/8
441 ipfw "-q add deny src-ip 10.0.0.0/24, 127.0.0.1/8"
444 The format of firewall rules is the following:
445 .Bd -ragged -offset indent
448 .Op Cm set Ar set_number
449 .Op Cm prob Ar match_probability
451 .Op Cm log Op Cm logamount Ar number
461 where the body of the rule specifies which information is used
462 for filtering packets, among the following:
464 .Bl -tag -width "Source and dest. addresses and ports" -offset XXX -compact
465 .It Layer-2 header fields
467 .It IPv4 and IPv6 Protocol
469 .It Source and dest. addresses and ports
473 .It Transmit and receive interface
475 .It Misc. IP header fields
476 Version, type of service, datagram length, identification,
477 fragment flag (non-zero IP offset),
480 .It IPv6 Extension headers
481 Fragmentation, Hop-by-Hop options,
482 Routing Headers, Source routing rthdr0, Mobile IPv6 rthdr2, IPSec options.
484 .It Misc. TCP header fields
485 TCP flags (SYN, FIN, ACK, RST, etc.),
486 sequence number, acknowledgment number,
494 When the packet can be associated with a local socket.
496 Whether a packet came from a divert socket (e.g.,
498 .It Fib annotation state
499 Whether a packet has been tagged for using a specific FIB (routing table)
500 in future forwarding decisions.
503 Note that some of the above information, e.g.\& source MAC or IP addresses and
504 TCP/UDP ports, can be easily spoofed, so filtering on those fields
505 alone might not guarantee the desired results.
506 .Bl -tag -width indent
508 Each rule is associated with a
510 in the range 1..65535, with the latter reserved for the
513 Rules are checked sequentially by rule number.
514 Multiple rules can have the same number, in which case they are
515 checked (and listed) according to the order in which they have
517 If a rule is entered without specifying a number, the kernel will
518 assign one in such a way that the rule becomes the last one
522 Automatic rule numbers are assigned by incrementing the last
523 non-default rule number by the value of the sysctl variable
524 .Ar net.inet.ip.fw.autoinc_step
525 which defaults to 100.
526 If this is not possible (e.g.\& because we would go beyond the
527 maximum allowed rule number), the number of the last
528 non-default value is used instead.
529 .It Cm set Ar set_number
530 Each rule is associated with a
533 Sets can be individually disabled and enabled, so this parameter
534 is of fundamental importance for atomic ruleset manipulation.
535 It can be also used to simplify deletion of groups of rules.
536 If a rule is entered without specifying a set number,
539 Set 31 is special in that it cannot be disabled,
540 and rules in set 31 are not deleted by the
542 command (but you can delete them with the
543 .Nm ipfw delete set 31
545 Set 31 is also used for the
548 .It Cm prob Ar match_probability
549 A match is only declared with the specified probability
550 (floating point number between 0 and 1).
551 This can be useful for a number of applications such as
552 random packet drop or
555 to simulate the effect of multiple paths leading to out-of-order
558 Note: this condition is checked before any other condition, including
559 ones such as keep-state or check-state which might have side effects.
560 .It Cm log Op Cm logamount Ar number
561 When a packet matches a rule with the
563 keyword, a message will be
569 The logging only occurs if the sysctl variable
570 .Va net.inet.ip.fw.verbose
572 (which is the default when the kernel is compiled with
573 .Dv IPFIREWALL_VERBOSE )
574 and the number of packets logged so far for that
575 particular rule does not exceed the
580 is specified, the limit is taken from the sysctl variable
581 .Va net.inet.ip.fw.verbose_limit .
582 In both cases, a value of 0 removes the logging limit.
584 Once the limit is reached, logging can be re-enabled by
585 clearing the logging counter or the packet counter for that entry, see the
589 Note: logging is done after all other packet matching conditions
590 have been successfully verified, and before performing the final
591 action (accept, deny, etc.) on the packet.
593 When a packet matches a rule with the
595 keyword, the numeric tag for the given
597 in the range 1..65534 will be attached to the packet.
598 The tag acts as an internal marker (it is not sent out over
599 the wire) that can be used to identify these packets later on.
600 This can be used, for example, to provide trust between interfaces
601 and to start doing policy-based filtering.
602 A packet can have multiple tags at the same time.
603 Tags are "sticky", meaning once a tag is applied to a packet by a
604 matching rule it exists until explicit removal.
605 Tags are kept with the packet everywhere within the kernel, but are
606 lost when packet leaves the kernel, for example, on transmitting
607 packet out to the network or sending packet to a
611 To check for previously applied tags, use the
614 To delete previously applied tag, use the
618 Note: since tags are kept with the packet everywhere in kernelspace,
619 they can be set and unset anywhere in the kernel network subsystem
622 facility), not only by means of the
628 For example, there can be a specialized
630 node doing traffic analyzing and tagging for later inspecting
632 .It Cm untag Ar number
633 When a packet matches a rule with the
635 keyword, the tag with the number
637 is searched among the tags attached to this packet and,
638 if found, removed from it.
639 Other tags bound to packet, if present, are left untouched.
641 When a packet matches a rule with the
643 keyword, the ALTQ identifier for the given
648 Note that this ALTQ tag is only meaningful for packets going "out" of IPFW,
649 and not being rejected or going to divert sockets.
650 Note that if there is insufficient memory at the time the packet is
651 processed, it will not be tagged, so it is wise to make your ALTQ
652 "default" queue policy account for this.
655 rules match a single packet, only the first one adds the ALTQ classification
657 In doing so, traffic may be shaped by using
658 .Cm count Cm altq Ar queue
659 rules for classification early in the ruleset, then later applying
660 the filtering decision.
665 rules may come later and provide the actual filtering decisions in
666 addition to the fallback ALTQ tag.
670 to set up the queues before IPFW will be able to look them up by name,
671 and if the ALTQ disciplines are rearranged, the rules in containing the
672 queue identifiers in the kernel will likely have gone stale and need
674 Stale queue identifiers will probably result in misclassification.
676 All system ALTQ processing can be turned on or off via
681 .Cm disable Ar altq .
683 .Va net.inet.ip.fw.one_pass
684 is irrelevant to ALTQ traffic shaping, as the actual rule action is followed
685 always after adding an ALTQ tag.
688 A rule can be associated with one of the following actions, which
689 will be executed when the packet matches the body of the rule.
690 .Bl -tag -width indent
691 .It Cm allow | accept | pass | permit
692 Allow packets that match rule.
693 The search terminates.
695 Checks the packet against the dynamic ruleset.
696 If a match is found, execute the action associated with
697 the rule which generated this dynamic rule, otherwise
698 move to the next rule.
701 rules do not have a body.
704 rule is found, the dynamic ruleset is checked at the first
710 Update counters for all packets that match rule.
711 The search continues with the next rule.
713 Discard packets that match this rule.
714 The search terminates.
715 .It Cm divert Ar port
716 Divert packets that match this rule to the
720 The search terminates.
721 .It Cm fwd | forward Ar ipaddr | tablearg Ns Op , Ns Ar port
722 Change the next-hop on matching packets to
724 which can be an IP address or a host name.
725 The next hop can also be supplied by the last table
726 looked up for the packet by using the
728 keyword instead of an explicit address.
729 The search terminates if this rule matches.
733 is a local address, then matching packets will be forwarded to
735 (or the port number in the packet if one is not specified in the rule)
736 on the local machine.
740 is not a local address, then the port number
741 (if specified) is ignored, and the packet will be
742 forwarded to the remote address, using the route as found in
743 the local routing table for that IP.
747 rule will not match layer-2 packets (those received
748 on ether_input, ether_output, or bridged).
752 action does not change the contents of the packet at all.
753 In particular, the destination address remains unmodified, so
754 packets forwarded to another system will usually be rejected by that system
755 unless there is a matching rule on that system to capture them.
756 For packets forwarded locally,
757 the local address of the socket will be
758 set to the original destination address of the packet.
761 entry look rather weird but is intended for
762 use with transparent proxy servers.
766 a custom kernel needs to be compiled with the option
767 .Cd "options IPFIREWALL_FORWARD" .
771 (for network address translation, address redirect, etc.):
773 .Sx NETWORK ADDRESS TRANSLATION (NAT)
774 Section for further information.
775 .It Cm pipe Ar pipe_nr
779 (for bandwidth limitation, delay, etc.).
781 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
782 Section for further information.
783 The search terminates; however, on exit from the pipe and if
787 .Va net.inet.ip.fw.one_pass
788 is not set, the packet is passed again to the firewall code
789 starting from the next rule.
790 .It Cm queue Ar queue_nr
794 (for bandwidth limitation using WF2Q+).
800 Discard packets that match this rule, and if the
801 packet is a TCP packet, try to send a TCP reset (RST) notice.
802 The search terminates.
804 Discard packets that match this rule, and if the
805 packet is a TCP packet, try to send a TCP reset (RST) notice.
806 The search terminates.
807 .It Cm skipto Ar number | tablearg
808 Skip all subsequent rules numbered less than
810 The search continues with the first rule numbered
813 It is possible to use the
815 keyword with a skipto for a
817 skipto, but care should be used, as no destination caching
818 is possible in this case so the rules are always walked to find it,
821 .It Cm call Ar number | tablearg
822 The current rule number is saved in the internal stack and
823 ruleset processing continues with the first rule numbered
826 If later a rule with the
828 action is encountered, the processing returns to the first rule
831 rule plus one or higher
832 (the same behaviour as with packets returning from
837 This could be used to make somewhat like an assembly language
839 calls to rules with common checks for different interfaces, etc.
841 Rule with any number could be called, not just forward jumps as with
843 So, to prevent endless loops in case of mistakes, both
847 actions don't do any jumps and simply go to the next rule if memory
848 can't be allocated or stack overflowed/undeflowed.
850 Internally stack for rule numbers is implemented using
852 facility and currently has size of 16 entries.
853 As mbuf tags are lost when packet leaves the kernel,
855 should not be used in subroutines to avoid endless loops
856 and other undesired effects.
858 Takes rule number saved to internal stack by the last
860 action and returns ruleset processing to the first rule
861 with number greater than number of corresponding
863 rule. See description of the
865 action for more details.
871 and thus are unconditional, but
873 command-line utility currently requires every action except
876 While it is sometimes useful to return only on some packets,
877 usually you want to print just
880 A workaround for this is to use new syntax and
884 .Bd -literal -offset indent
885 # Add a rule without actual body
886 ipfw add 2999 return via any
888 # List rules without "from any to any" part
892 This cosmetic annoyance may be fixed in future releases.
894 Send a copy of packets matching this rule to the
898 The search continues with the next rule.
899 .It Cm unreach Ar code
900 Discard packets that match this rule, and try to send an ICMP
901 unreachable notice with code
905 is a number from 0 to 255, or one of these aliases:
906 .Cm net , host , protocol , port ,
907 .Cm needfrag , srcfail , net-unknown , host-unknown ,
908 .Cm isolated , net-prohib , host-prohib , tosnet ,
909 .Cm toshost , filter-prohib , host-precedence
911 .Cm precedence-cutoff .
912 The search terminates.
913 .It Cm unreach6 Ar code
914 Discard packets that match this rule, and try to send an ICMPv6
915 unreachable notice with code
919 is a number from 0, 1, 3 or 4, or one of these aliases:
920 .Cm no-route, admin-prohib, address
923 The search terminates.
924 .It Cm netgraph Ar cookie
925 Divert packet into netgraph with given
927 The search terminates.
928 If packet is later returned from netgraph it is either
929 accepted or continues with the next rule, depending on
930 .Va net.inet.ip.fw.one_pass
932 .It Cm ngtee Ar cookie
933 A copy of packet is diverted into netgraph, original
934 packet continues with the next rule.
937 for more information on
942 .It Cm setfib Ar fibnum | tablearg
943 The packet is tagged so as to use the FIB (routing table)
945 in any subsequent forwarding decisions.
946 Initially this is limited to the values 0 through 15, see
948 Processing continues at the next rule.
949 It is possible to use the
951 keyword with a setfib. If tablearg value is not within compiled FIB range packet fib is set to 0.
953 Queue and reassemble ip fragments.
954 If the packet is not fragmented, counters are updated and processing continues with the next rule.
955 If the packet is the last logical fragment, the packet is reassembled and, if
956 .Va net.inet.ip.fw.one_pass
957 is set to 0, processing continues with the next rule, else packet is allowed to pass and search terminates.
958 If the packet is a fragment in the middle, it is consumed and processing stops immediately.
960 Fragments handling can be tuned via
961 .Va net.inet.ip.maxfragpackets
963 .Va net.inet.ip.maxfragsperpacket
964 which limit, respectively, the maximum number of processable fragments (default: 800) and
965 the maximum number of fragments per packet (default: 16).
967 NOTA BENE: since fragments do not contain port numbers, they should be avoided with the
970 Alternatively, direction-based (like
974 ) and source-based (like
976 ) match patterns can be used to select fragments.
978 Usually a simple rule like:
979 .Bd -literal -offset indent
980 # reassemble incoming fragments
981 ipfw add reass all from any to any in
984 is all you need at the beginning of your ruleset.
987 The body of a rule contains zero or more patterns (such as
988 specific source and destination addresses or ports,
989 protocol options, incoming or outgoing interfaces, etc.)
990 that the packet must match in order to be recognised.
991 In general, the patterns are connected by (implicit)
993 operators -- i.e., all must match in order for the
995 Individual patterns can be prefixed by the
997 operator to reverse the result of the match, as in
999 .Dl "ipfw add 100 allow ip from not 1.2.3.4 to any"
1001 Additionally, sets of alternative match patterns
1003 can be constructed by putting the patterns in
1004 lists enclosed between parentheses ( ) or braces { }, and
1007 operator as follows:
1009 .Dl "ipfw add 100 allow ip from { x or not y or z } to any"
1011 Only one level of parentheses is allowed.
1012 Beware that most shells have special meanings for parentheses
1013 or braces, so it is advisable to put a backslash \\ in front of them
1014 to prevent such interpretations.
1016 The body of a rule must in general include a source and destination
1020 can be used in various places to specify that the content of
1021 a required field is irrelevant.
1023 The rule body has the following format:
1024 .Bd -ragged -offset indent
1025 .Op Ar proto Cm from Ar src Cm to Ar dst
1029 The first part (proto from src to dst) is for backward
1030 compatibility with earlier versions of
1034 any match pattern (including MAC headers, IP protocols,
1035 addresses and ports) can be specified in the
1039 Rule fields have the following meaning:
1040 .Bl -tag -width indent
1041 .It Ar proto : protocol | Cm { Ar protocol Cm or ... }
1042 .It Ar protocol : Oo Cm not Oc Ar protocol-name | protocol-number
1043 An IP protocol specified by number or name
1044 (for a complete list see
1045 .Pa /etc/protocols ) ,
1046 or one of the following keywords:
1047 .Bl -tag -width indent
1049 Matches IPv4 packets.
1051 Matches IPv6 packets.
1060 option will be treated as inner protocol.
1068 .Cm { Ar protocol Cm or ... }
1071 is provided for convenience only but its use is deprecated.
1072 .It Ar src No and Ar dst : Bro Cm addr | Cm { Ar addr Cm or ... } Brc Op Oo Cm not Oc Ar ports
1073 An address (or a list, see below)
1074 optionally followed by
1080 with multiple addresses) is provided for convenience only and
1081 its use is discouraged.
1082 .It Ar addr : Oo Cm not Oc Bro
1083 .Bl -tag -width indent
1084 .Cm any | me | me6 |
1085 .Cm table Ns Pq Ar number Ns Op , Ns Ar value
1086 .Ar | addr-list | addr-set
1089 matches any IP address.
1091 matches any IP address configured on an interface in the system.
1093 matches any IPv6 address configured on an interface in the system.
1094 The address list is evaluated at the time the packet is
1096 .It Cm table Ns Pq Ar number Ns Op , Ns Ar value
1097 Matches any IPv4 address for which an entry exists in the lookup table
1099 If an optional 32-bit unsigned
1101 is also specified, an entry will match only if it has this value.
1104 section below for more information on lookup tables.
1106 .It Ar addr-list : ip-addr Ns Op Ns , Ns Ar addr-list
1108 A host or subnet address specified in one of the following ways:
1109 .Bl -tag -width indent
1110 .It Ar numeric-ip | hostname
1111 Matches a single IPv4 address, specified as dotted-quad or a hostname.
1112 Hostnames are resolved at the time the rule is added to the firewall list.
1113 .It Ar addr Ns / Ns Ar masklen
1114 Matches all addresses with base
1116 (specified as an IP address, a network number, or a hostname)
1120 As an example, 1.2.3.4/25 or 1.2.3.0/25 will match
1121 all IP numbers from 1.2.3.0 to 1.2.3.127 .
1122 .It Ar addr Ns : Ns Ar mask
1123 Matches all addresses with base
1125 (specified as an IP address, a network number, or a hostname)
1128 specified as a dotted quad.
1129 As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match
1131 This form is advised only for non-contiguous
1133 It is better to resort to the
1134 .Ar addr Ns / Ns Ar masklen
1135 format for contiguous masks, which is more compact and less
1138 .It Ar addr-set : addr Ns Oo Ns / Ns Ar masklen Oc Ns Cm { Ns Ar list Ns Cm }
1139 .It Ar list : Bro Ar num | num-num Brc Ns Op Ns , Ns Ar list
1140 Matches all addresses with base address
1142 (specified as an IP address, a network number, or a hostname)
1143 and whose last byte is in the list between braces { } .
1144 Note that there must be no spaces between braces and
1145 numbers (spaces after commas are allowed).
1146 Elements of the list can be specified as single entries
1150 field is used to limit the size of the set of addresses,
1151 and can have any value between 24 and 32.
1153 it will be assumed as 24.
1155 This format is particularly useful to handle sparse address sets
1156 within a single rule.
1157 Because the matching occurs using a
1158 bitmask, it takes constant time and dramatically reduces
1159 the complexity of rulesets.
1161 As an example, an address specified as 1.2.3.4/24{128,35-55,89}
1162 or 1.2.3.0/24{128,35-55,89}
1163 will match the following IP addresses:
1165 1.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .
1166 .It Ar addr6-list : ip6-addr Ns Op Ns , Ns Ar addr6-list
1168 A host or subnet specified one of the following ways:
1170 .Bl -tag -width indent
1171 .It Ar numeric-ip | hostname
1172 Matches a single IPv6 address as allowed by
1175 Hostnames are resolved at the time the rule is added to the firewall
1177 .It Ar addr Ns / Ns Ar masklen
1178 Matches all IPv6 addresses with base
1180 (specified as allowed by
1188 No support for sets of IPv6 addresses is provided because IPv6 addresses
1189 are typically random past the initial prefix.
1190 .It Ar ports : Bro Ar port | port Ns \&- Ns Ar port Ns Brc Ns Op , Ns Ar ports
1191 For protocols which support port numbers (such as TCP and UDP), optional
1193 may be specified as one or more ports or port ranges, separated
1194 by commas but no spaces, and an optional
1199 notation specifies a range of ports (including boundaries).
1203 may be used instead of numeric port values.
1204 The length of the port list is limited to 30 ports or ranges,
1205 though one can specify larger ranges by using an
1209 section of the rule.
1213 can be used to escape the dash
1215 character in a service name (from a shell, the backslash must be
1216 typed twice to avoid the shell itself interpreting it as an escape
1219 .Dl "ipfw add count tcp from any ftp\e\e-data-ftp to any"
1221 Fragmented packets which have a non-zero offset (i.e., not the first
1222 fragment) will never match a rule which has one or more port
1226 option for details on matching fragmented packets.
1228 .Ss RULE OPTIONS (MATCH PATTERNS)
1229 Additional match patterns can be used within
1231 Zero or more of these so-called
1233 can be present in a rule, optionally prefixed by the
1235 operand, and possibly grouped into
1238 The following match patterns can be used (listed in alphabetical order):
1239 .Bl -tag -width indent
1240 .It Cm // this is a comment.
1241 Inserts the specified text as a comment in the rule.
1242 Everything following // is considered as a comment and stored in the rule.
1243 You can have comment-only rules, which are listed as having a
1245 action followed by the comment.
1250 Matches only packets generated by a divert socket.
1251 .It Cm diverted-loopback
1252 Matches only packets coming from a divert socket back into the IP stack
1254 .It Cm diverted-output
1255 Matches only packets going from a divert socket back outward to the IP
1256 stack output for delivery.
1257 .It Cm dst-ip Ar ip-address
1258 Matches IPv4 packets whose destination IP is one of the address(es)
1259 specified as argument.
1260 .It Bro Cm dst-ip6 | dst-ipv6 Brc Ar ip6-address
1261 Matches IPv6 packets whose destination IP is one of the address(es)
1262 specified as argument.
1263 .It Cm dst-port Ar ports
1264 Matches IP packets whose destination port is one of the port(s)
1265 specified as argument.
1267 Matches TCP packets that have the RST or ACK bits set.
1268 .It Cm ext6hdr Ar header
1269 Matches IPv6 packets containing the extended header given by
1271 Supported headers are:
1277 any type of Routing Header
1279 Source routing Routing Header Type 0
1281 Mobile IPv6 Routing Header Type 2
1285 IPSec authentication headers
1287 and IPsec encapsulated security payload headers
1289 .It Cm fib Ar fibnum
1290 Matches a packet that has been tagged to use
1291 the given FIB (routing table) number.
1292 .It Cm flow-id Ar labels
1293 Matches IPv6 packets containing any of the flow labels given in
1296 is a comma separated list of numeric flow labels.
1298 Matches packets that are fragments and not the first
1299 fragment of an IP datagram.
1300 Note that these packets will not have
1301 the next protocol header (e.g.\& TCP, UDP) so options that look into
1302 these headers cannot match.
1304 Matches all TCP or UDP packets sent by or received for a
1308 may be specified by name or number.
1309 .It Cm jail Ar prisonID
1310 Matches all TCP or UDP packets sent by or received for the
1311 jail whos prison ID is
1313 .It Cm icmptypes Ar types
1314 Matches ICMP packets whose ICMP type is in the list
1316 The list may be specified as any combination of
1317 individual types (numeric) separated by commas.
1318 .Em Ranges are not allowed .
1319 The supported ICMP types are:
1323 destination unreachable
1331 router advertisement
1335 time-to-live exceeded
1347 address mask request
1349 and address mask reply
1351 .It Cm icmp6types Ar types
1352 Matches ICMP6 packets whose ICMP6 type is in the list of
1354 The list may be specified as any combination of
1355 individual types (numeric) separated by commas.
1356 .Em Ranges are not allowed .
1358 Matches incoming or outgoing packets, respectively.
1362 are mutually exclusive (in fact,
1366 .It Cm ipid Ar id-list
1367 Matches IPv4 packets whose
1369 field has value included in
1371 which is either a single value or a list of values or ranges
1372 specified in the same way as
1374 .It Cm iplen Ar len-list
1375 Matches IP packets whose total length, including header and data, is
1378 which is either a single value or a list of values or ranges
1379 specified in the same way as
1381 .It Cm ipoptions Ar spec
1382 Matches packets whose IPv4 header contains the comma separated list of
1383 options specified in
1385 The supported IP options are:
1388 (strict source route),
1390 (loose source route),
1392 (record packet route) and
1395 The absence of a particular option may be denoted
1398 .It Cm ipprecedence Ar precedence
1399 Matches IPv4 packets whose precedence field is equal to
1402 Matches packets that have IPSEC history associated with them
1403 (i.e., the packet comes encapsulated in IPSEC, the kernel
1404 has IPSEC support and IPSEC_FILTERTUNNEL option, and can correctly
1407 Note that specifying
1409 is different from specifying
1411 as the latter will only look at the specific IP protocol field,
1412 irrespective of IPSEC kernel support and the validity of the IPSEC data.
1414 Further note that this flag is silently ignored in kernels without
1416 It does not affect rule processing when given and the
1417 rules are handled as if with no
1420 .It Cm iptos Ar spec
1421 Matches IPv4 packets whose
1423 field contains the comma separated list of
1424 service types specified in
1426 The supported IP types of service are:
1429 .Pq Dv IPTOS_LOWDELAY ,
1431 .Pq Dv IPTOS_THROUGHPUT ,
1433 .Pq Dv IPTOS_RELIABILITY ,
1435 .Pq Dv IPTOS_MINCOST ,
1437 .Pq Dv IPTOS_ECN_CE .
1438 The absence of a particular type may be denoted
1441 .It Cm ipttl Ar ttl-list
1442 Matches IPv4 packets whose time to live is included in
1444 which is either a single value or a list of values or ranges
1445 specified in the same way as
1447 .It Cm ipversion Ar ver
1448 Matches IP packets whose IP version field is
1451 Upon a match, the firewall will create a dynamic rule, whose
1452 default behaviour is to match bidirectional traffic between
1453 source and destination IP/port using the same protocol.
1454 The rule has a limited lifetime (controlled by a set of
1456 variables), and the lifetime is refreshed every time a matching
1459 Matches only layer2 packets, i.e., those passed to
1461 from ether_demux() and ether_output_frame().
1462 .It Cm limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N
1463 The firewall will only allow
1465 connections with the same
1466 set of parameters as specified in the rule.
1468 of source and destination addresses and ports can be
1471 only IPv4 flows are supported.
1472 .It Cm lookup Bro Cm dst-ip | dst-port | src-ip | src-port | uid | jail Brc Ar N
1473 Search an entry in lookup table
1475 that matches the field specified as argument.
1476 If not found, the match fails.
1477 Otherwise, the match succeeds and
1479 is set to the value extracted from the table.
1481 This option can be useful to quickly dispatch traffic based on
1482 certain packet fields.
1485 section below for more information on lookup tables.
1486 .It Cm { MAC | mac } Ar dst-mac src-mac
1487 Match packets with a given
1491 addresses, specified as the
1493 keyword (matching any MAC address), or six groups of hex digits
1494 separated by colons,
1495 and optionally followed by a mask indicating the significant bits.
1496 The mask may be specified using either of the following methods:
1497 .Bl -enum -width indent
1501 followed by the number of significant bits.
1502 For example, an address with 33 significant bits could be specified as:
1504 .Dl "MAC 10:20:30:40:50:60/33 any"
1509 followed by a bitmask specified as six groups of hex digits separated
1511 For example, an address in which the last 16 bits are significant could
1514 .Dl "MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any"
1516 Note that the ampersand character has a special meaning in many shells
1517 and should generally be escaped.
1520 Note that the order of MAC addresses (destination first,
1522 the same as on the wire, but the opposite of the one used for
1524 .It Cm mac-type Ar mac-type
1525 Matches packets whose Ethernet Type field
1526 corresponds to one of those specified as argument.
1528 is specified in the same way as
1530 (i.e., one or more comma-separated single values or ranges).
1531 You can use symbolic names for known values such as
1532 .Em vlan , ipv4, ipv6 .
1533 Values can be entered as decimal or hexadecimal (if prefixed by 0x),
1534 and they are always printed as hexadecimal (unless the
1536 option is used, in which case symbolic resolution will be attempted).
1537 .It Cm proto Ar protocol
1538 Matches packets with the corresponding IP protocol.
1539 .It Cm recv | xmit | via Brq Ar ifX | Ar if Ns Cm * | Ar table Ns Pq Ar number Ns Op , Ns Ar value | Ar ipno | Ar any
1540 Matches packets received, transmitted or going through,
1541 respectively, the interface specified by exact name
1542 .Ns No ( Ar ifX Ns No ),
1544 .Ns No ( Ar if Ns Ar * Ns No ),
1545 by IP address, or through some interface.
1549 keyword causes the interface to always be checked.
1556 then only the receive or transmit interface (respectively)
1558 By specifying both, it is possible to match packets based on
1559 both receive and transmit interface, e.g.:
1561 .Dl "ipfw add deny ip from any to any out recv ed0 xmit ed1"
1565 interface can be tested on either incoming or outgoing packets,
1568 interface can only be tested on outgoing packets.
1573 is invalid) whenever
1577 A packet might not have a receive or transmit interface: packets
1578 originating from the local host have no receive interface,
1579 while packets destined for the local host have no transmit
1582 Matches TCP packets that have the SYN bit set but no ACK bit.
1583 This is the short form of
1584 .Dq Li tcpflags\ syn,!ack .
1585 .It Cm src-ip Ar ip-address
1586 Matches IPv4 packets whose source IP is one of the address(es)
1587 specified as an argument.
1588 .It Cm src-ip6 Ar ip6-address
1589 Matches IPv6 packets whose source IP is one of the address(es)
1590 specified as an argument.
1591 .It Cm src-port Ar ports
1592 Matches IP packets whose source port is one of the port(s)
1593 specified as argument.
1594 .It Cm tagged Ar tag-list
1595 Matches packets whose tags are included in
1597 which is either a single value or a list of values or ranges
1598 specified in the same way as
1600 Tags can be applied to the packet using
1602 rule action parameter (see it's description for details on tags).
1603 .It Cm tcpack Ar ack
1605 Match if the TCP header acknowledgment number field is set to
1607 .It Cm tcpdatalen Ar tcpdatalen-list
1608 Matches TCP packets whose length of TCP data is
1609 .Ar tcpdatalen-list ,
1610 which is either a single value or a list of values or ranges
1611 specified in the same way as
1613 .It Cm tcpflags Ar spec
1615 Match if the TCP header contains the comma separated list of
1618 The supported TCP flags are:
1627 The absence of a particular flag may be denoted
1630 A rule which contains a
1632 specification can never match a fragmented packet which has
1636 option for details on matching fragmented packets.
1637 .It Cm tcpseq Ar seq
1639 Match if the TCP header sequence number field is set to
1641 .It Cm tcpwin Ar win
1643 Match if the TCP header window field is set to
1645 .It Cm tcpoptions Ar spec
1647 Match if the TCP header contains the comma separated list of
1648 options specified in
1650 The supported TCP options are:
1653 (maximum segment size),
1655 (tcp window advertisement),
1659 (rfc1323 timestamp) and
1661 (rfc1644 t/tcp connection count).
1662 The absence of a particular option may be denoted
1666 Match all TCP or UDP packets sent by or received for a
1670 may be matched by name or identification number.
1672 For incoming packets,
1673 a routing table lookup is done on the packet's source address.
1674 If the interface on which the packet entered the system matches the
1675 outgoing interface for the route,
1677 If the interfaces do not match up,
1678 the packet does not match.
1679 All outgoing packets or packets with no incoming interface match.
1681 The name and functionality of the option is intentionally similar to
1682 the Cisco IOS command:
1684 .Dl ip verify unicast reverse-path
1686 This option can be used to make anti-spoofing rules to reject all
1687 packets with source addresses not from this interface.
1691 For incoming packets,
1692 a routing table lookup is done on the packet's source address.
1693 If a route to the source address exists, but not the default route
1694 or a blackhole/reject route, the packet matches.
1695 Otherwise, the packet does not match.
1696 All outgoing packets match.
1698 The name and functionality of the option is intentionally similar to
1699 the Cisco IOS command:
1701 .Dl ip verify unicast source reachable-via any
1703 This option can be used to make anti-spoofing rules to reject all
1704 packets whose source address is unreachable.
1706 For incoming packets, the packet's source address is checked if it
1707 belongs to a directly connected network.
1708 If the network is directly connected, then the interface the packet
1709 came on in is compared to the interface the network is connected to.
1710 When incoming interface and directly connected interface are not the
1711 same, the packet does not match.
1712 Otherwise, the packet does match.
1713 All outgoing packets match.
1715 This option can be used to make anti-spoofing rules to reject all
1716 packets that pretend to be from a directly connected network but do
1717 not come in through that interface.
1718 This option is similar to but more restricted than
1720 because it engages only on packets with source addresses of directly
1721 connected networks instead of all source addresses.
1724 Lookup tables are useful to handle large sparse sets of
1725 addresses or other search keys (e.g. ports, jail IDs, interface names).
1726 In the rest of this section we will use the term ``address''.
1727 There may be up to 4096 different lookup tables, numbered 0 to 4095.
1729 Each entry is represented by an
1730 .Ar addr Ns Op / Ns Ar masklen
1731 and will match all addresses with base
1733 (specified as an IPv4/IPv6 address, a hostname or an unsigned integer)
1739 is not specified, it defaults to 32 for IPv4 and 128 for IPv6.
1740 When looking up an IP address in a table, the most specific
1742 Associated with each entry is a 32-bit unsigned
1744 which can optionally be checked by a rule matching code.
1745 When adding an entry, if
1747 is not specified, it defaults to 0.
1749 An entry can be added to a table
1751 or removed from a table
1753 A table can be examined
1758 Internally, each table is stored in a Radix tree, the same way as
1759 the routing table (see
1762 Lookup tables currently support only ports, jail IDs, IPv4/IPv6 addresses
1763 and interface names. Wildcards is not supported for interface names.
1767 feature provides the ability to use a value, looked up in the table, as
1768 the argument for a rule action, action parameter or rule option.
1769 This can significantly reduce number of rules in some configurations.
1770 If two tables are used in a rule, the result of the second (destination)
1774 argument can be used with the following actions:
1775 .Cm nat, pipe , queue, divert, tee, netgraph, ngtee, fwd, skipto, setfib,
1783 it is possible to supply table entries with values
1784 that are in the form of IP addresses or hostnames.
1787 Section for example usage of tables and the tablearg keyword.
1791 action, the user should be aware that the code will walk the ruleset
1792 up to a rule equal to, or past, the given number, and should therefore try keep the
1793 ruleset compact between the skipto and the target rules.
1795 Each rule belongs to one of 32 different
1798 Set 31 is reserved for the default rule.
1800 By default, rules are put in set 0, unless you use the
1802 attribute when entering a new rule.
1803 Sets can be individually and atomically enabled or disabled,
1804 so this mechanism permits an easy way to store multiple configurations
1805 of the firewall and quickly (and atomically) switch between them.
1806 The command to enable/disable sets is
1807 .Bd -ragged -offset indent
1809 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
1816 sections can be specified.
1817 Command execution is atomic on all the sets specified in the command.
1818 By default, all sets are enabled.
1820 When you disable a set, its rules behave as if they do not exist
1821 in the firewall configuration, with only one exception:
1822 .Bd -ragged -offset indent
1823 dynamic rules created from a rule before it had been disabled
1824 will still be active until they expire.
1826 dynamic rules you have to explicitly delete the parent rule
1827 which generated them.
1830 The set number of rules can be changed with the command
1831 .Bd -ragged -offset indent
1834 .Brq Cm rule Ar rule-number | old-set
1838 Also, you can atomically swap two rulesets with the command
1839 .Bd -ragged -offset indent
1841 .Cm set swap Ar first-set second-set
1846 Section on some possible uses of sets of rules.
1847 .Sh STATEFUL FIREWALL
1848 Stateful operation is a way for the firewall to dynamically
1849 create rules for specific flows when packets that
1850 match a given pattern are detected.
1851 Support for stateful
1852 operation comes through the
1853 .Cm check-state , keep-state
1859 Dynamic rules are created when a packet matches a
1863 rule, causing the creation of a
1865 rule which will match all and only packets with
1869 .Em src-ip/src-port dst-ip/dst-port
1874 are used here only to denote the initial match addresses, but they
1875 are completely equivalent afterwards).
1876 Dynamic rules will be checked at the first
1877 .Cm check-state, keep-state
1880 occurrence, and the action performed upon a match will be the same
1881 as in the parent rule.
1883 Note that no additional attributes other than protocol and IP addresses
1884 and ports are checked on dynamic rules.
1886 The typical use of dynamic rules is to keep a closed firewall configuration,
1887 but let the first TCP SYN packet from the inside network install a
1888 dynamic rule for the flow so that packets belonging to that session
1889 will be allowed through the firewall:
1891 .Dl "ipfw add check-state"
1892 .Dl "ipfw add allow tcp from my-subnet to any setup keep-state"
1893 .Dl "ipfw add deny tcp from any to any"
1895 A similar approach can be used for UDP, where an UDP packet coming
1896 from the inside will install a dynamic rule to let the response through
1899 .Dl "ipfw add check-state"
1900 .Dl "ipfw add allow udp from my-subnet to any keep-state"
1901 .Dl "ipfw add deny udp from any to any"
1903 Dynamic rules expire after some time, which depends on the status
1904 of the flow and the setting of some
1908 .Sx SYSCTL VARIABLES
1910 For TCP sessions, dynamic rules can be instructed to periodically
1911 send keepalive packets to refresh the state of the rule when it is
1916 for more examples on how to use dynamic rules.
1917 .Sh TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
1919 is also the user interface for the
1921 traffic shaper, packet scheduler and network emulator, a subsystem that
1922 can artificially queue, delay or drop packets
1923 emulating the behaviour of certain network links
1924 or queueing systems.
1927 operates by first using the firewall to select packets
1928 using any match pattern that can be used in
1931 Matching packets are then passed to either of two
1932 different objects, which implement the traffic regulation:
1933 .Bl -hang -offset XXXX
1939 with given bandwidth and propagation delay,
1940 driven by a FIFO scheduler and a single queue with programmable
1941 queue size and packet loss rate.
1942 Packets are appended to the queue as they come out from
1944 and then transferred in FIFO order to the link at the desired rate.
1948 is an abstraction used to implement packet scheduling
1949 using one of several packet scheduling algorithms.
1952 are first grouped into flows according to a mask on the 5-tuple.
1953 Flows are then passed to the scheduler associated to the
1955 and each flow uses scheduling parameters (weight and others)
1956 as configured in the
1959 A scheduler in turn is connected to an emulated link,
1960 and arbitrates the link's bandwidth among backlogged flows according to
1961 weights and to the features of the scheduling algorithm in use.
1966 can be used to set hard limits to the bandwidth that a flow can use, whereas
1968 can be used to determine how different flows share the available bandwidth.
1970 A graphical representation of the binding of queues,
1971 flows, schedulers and links is below.
1972 .Bd -literal -offset indent
1973 (flow_mask|sched_mask) sched_mask
1974 +---------+ weight Wx +-------------+
1975 | |->-[flow]-->--| |-+
1976 -->--| QUEUE x | ... | | |
1977 | |->-[flow]-->--| SCHEDuler N | |
1979 ... | +--[LINK N]-->--
1980 +---------+ weight Wy | | +--[LINK N]-->--
1981 | |->-[flow]-->--| | |
1982 -->--| QUEUE y | ... | | |
1983 | |->-[flow]-->--| | |
1984 +---------+ +-------------+ |
1987 It is important to understand the role of the SCHED_MASK
1988 and FLOW_MASK, which are configured through the commands
1989 .Dl "ipfw sched N config mask SCHED_MASK ..."
1991 .Dl "ipfw queue X config mask FLOW_MASK ..." .
1993 The SCHED_MASK is used to assign flows to one or more
1994 scheduler instances, one for each
1995 value of the packet's 5-fuple after applying SCHED_MASK.
1996 As an example, using ``src-ip 0xffffff00'' creates one instance
1997 for each /24 destination subnet.
1999 The FLOW_MASK, together with the SCHED_MASK, is used to split
2000 packets into flows. As an example, using
2001 ``src-ip 0x000000ff''
2002 together with the previous SCHED_MASK makes a flow for
2003 each individual source address. In turn, flows for each /24
2004 subnet will be sent to the same scheduler instance.
2006 The above diagram holds even for the
2008 case, with the only restriction that a
2010 only supports a SCHED_MASK, and forces the use of a FIFO
2011 scheduler (these are for backward compatibility reasons;
2012 in fact, internally, a
2014 pipe is implemented exactly as above).
2016 There are two modes of
2024 mode tries to emulate a real link: the
2026 scheduler ensures that the packet will not leave the pipe faster than it
2027 would on the real link with a given bandwidth.
2030 mode allows certain packets to bypass the
2032 scheduler (if packet flow does not exceed pipe's bandwidth).
2033 This is the reason why the
2035 mode requires less CPU cycles per packet (on average) and packet latency
2036 can be significantly lower in comparison to a real link with the same
2042 mode can be enabled by setting the
2043 .Va net.inet.ip.dummynet.io_fast
2045 variable to a non-zero value.
2047 .Ss PIPE, QUEUE AND SCHEDULER CONFIGURATION
2053 configuration commands are the following:
2054 .Bd -ragged -offset indent
2055 .Cm pipe Ar number Cm config Ar pipe-configuration
2057 .Cm queue Ar number Cm config Ar queue-configuration
2059 .Cm sched Ar number Cm config Ar sched-configuration
2062 The following parameters can be configured for a pipe:
2064 .Bl -tag -width indent -compact
2065 .It Cm bw Ar bandwidth | device
2066 Bandwidth, measured in
2069 .Brq Cm bit/s | Byte/s .
2072 A value of 0 (default) means unlimited bandwidth.
2073 The unit must immediately follow the number, as in
2075 .Dl "ipfw pipe 1 config bw 300Kbit/s"
2077 If a device name is specified instead of a numeric value, as in
2079 .Dl "ipfw pipe 1 config bw tun0"
2081 then the transmit clock is supplied by the specified device.
2082 At the moment only the
2084 device supports this
2085 functionality, for use in conjunction with
2088 .It Cm delay Ar ms-delay
2089 Propagation delay, measured in milliseconds.
2090 The value is rounded to the next multiple of the clock tick
2091 (typically 10ms, but it is a good practice to run kernels
2093 .Dq "options HZ=1000"
2095 the granularity to 1ms or less).
2096 The default value is 0, meaning no delay.
2098 .It Cm burst Ar size
2099 If the data to be sent exceeds the pipe's bandwidth limit
2100 (and the pipe was previously idle), up to
2102 bytes of data are allowed to bypass the
2104 scheduler, and will be sent as fast as the physical link allows.
2105 Any additional data will be transmitted at the rate specified
2109 The burst size depends on how long the pipe has been idle;
2110 the effective burst size is calculated as follows:
2117 .It Cm profile Ar filename
2118 A file specifying the additional overhead incurred in the transmission
2119 of a packet on the link.
2121 Some link types introduce extra delays in the transmission
2122 of a packet, e.g. because of MAC level framing, contention on
2123 the use of the channel, MAC level retransmissions and so on.
2124 From our point of view, the channel is effectively unavailable
2125 for this extra time, which is constant or variable depending
2126 on the link type. Additionally, packets may be dropped after this
2127 time (e.g. on a wireless link after too many retransmissions).
2128 We can model the additional delay with an empirical curve
2129 that represents its distribution.
2130 .Bd -literal -offset indent
2131 cumulative probability
2141 +-------*------------------->
2144 The empirical curve may have both vertical and horizontal lines.
2145 Vertical lines represent constant delay for a range of
2147 Horizontal lines correspond to a discontinuity in the delay
2148 distribution: the pipe will use the largest delay for a
2151 The file format is the following, with whitespace acting as
2152 a separator and '#' indicating the beginning a comment:
2153 .Bl -tag -width indent
2154 .It Cm name Ar identifier
2155 optional name (listed by "ipfw pipe show")
2156 to identify the delay distribution;
2158 the bandwidth used for the pipe.
2159 If not specified here, it must be present
2160 explicitly as a configuration parameter for the pipe;
2161 .It Cm loss-level Ar L
2162 the probability above which packets are lost.
2163 (0.0 <= L <= 1.0, default 1.0 i.e. no loss);
2165 the number of samples used in the internal
2166 representation of the curve (2..1024; default 100);
2167 .It Cm "delay prob" | "prob delay"
2168 One of these two lines is mandatory and defines
2169 the format of the following lines with data points.
2171 2 or more lines representing points in the curve,
2172 with either delay or probability first, according
2173 to the chosen format.
2174 The unit for delay is milliseconds.
2175 Data points do not need to be sorted.
2176 Also, tne number of actual lines can be different
2177 from the value of the "samples" parameter:
2179 utility will sort and interpolate
2180 the curve as needed.
2183 Example of a profile file:
2184 .Bd -literal -offset indent
2189 0 200 # minimum overhead is 200ms
2195 #configuration file end
2199 The following parameters can be configured for a queue:
2201 .Bl -tag -width indent -compact
2202 .It Cm pipe Ar pipe_nr
2203 Connects a queue to the specified pipe.
2204 Multiple queues (with the same or different weights) can be connected to
2205 the same pipe, which specifies the aggregate rate for the set of queues.
2207 .It Cm weight Ar weight
2208 Specifies the weight to be used for flows matching this queue.
2209 The weight must be in the range 1..100, and defaults to 1.
2212 The following case-insensitive parameters can be configured for a
2215 .Bl -tag -width indent -compact
2216 .It Cm type Ar {fifo | wf2q+ | rr | qfq}
2217 specifies the scheduling algorithm to use.
2218 .Bl -tag -width indent -compact
2220 is just a FIFO scheduler (which means that all packets
2221 are stored in the same queue as they arrive to the scheduler).
2222 FIFO has O(1) per-packet time complexity, with very low
2223 constants (estimate 60-80ns on a 2GHz desktop machine)
2224 but gives no service guarantees.
2226 implements the WF2Q+ algorithm, which is a Weighted Fair Queueing
2227 algorithm which permits flows to share bandwidth according to
2228 their weights. Note that weights are not priorities; even a flow
2229 with a minuscule weight will never starve.
2230 WF2Q+ has O(log N) per-packet processing cost, where N is the number
2231 of flows, and is the default algorithm used by previous versions
2234 implements the Deficit Round Robin algorithm, which has O(1) processing
2235 costs (roughly, 100-150ns per packet)
2236 and permits bandwidth allocation according to weights, but
2237 with poor service guarantees.
2239 implements the QFQ algorithm, which is a very fast variant of
2240 WF2Q+, with similar service guarantees and O(1) processing
2241 costs (roughly, 200-250ns per packet).
2245 In addition to the type, all parameters allowed for a pipe can also
2246 be specified for a scheduler.
2248 Finally, the following parameters can be configured for both
2251 .Bl -tag -width XXXX -compact
2253 .It Cm buckets Ar hash-table-size
2254 Specifies the size of the hash table used for storing the
2256 Default value is 64 controlled by the
2259 .Va net.inet.ip.dummynet.hash_size ,
2260 allowed range is 16 to 65536.
2262 .It Cm mask Ar mask-specifier
2263 Packets sent to a given pipe or queue by an
2265 rule can be further classified into multiple flows, each of which is then
2269 A flow identifier is constructed by masking the IP addresses,
2270 ports and protocol types as specified with the
2272 options in the configuration of the pipe or queue.
2273 For each different flow identifier, a new pipe or queue is created
2274 with the same parameters as the original object, and matching packets
2279 are used, each flow will get the same bandwidth as defined by the pipe,
2282 are used, each flow will share the parent's pipe bandwidth evenly
2283 with other flows generated by the same queue (note that other queues
2284 with different weights might be connected to the same pipe).
2286 Available mask specifiers are a combination of one or more of the following:
2288 .Cm dst-ip Ar mask ,
2289 .Cm dst-ip6 Ar mask ,
2290 .Cm src-ip Ar mask ,
2291 .Cm src-ip6 Ar mask ,
2292 .Cm dst-port Ar mask ,
2293 .Cm src-port Ar mask ,
2294 .Cm flow-id Ar mask ,
2299 where the latter means all bits in all fields are significant.
2302 When a packet is dropped by a
2304 queue or pipe, the error
2305 is normally reported to the caller routine in the kernel, in the
2306 same way as it happens when a device queue fills up.
2308 option reports the packet as successfully delivered, which can be
2309 needed for some experimental setups where you want to simulate
2310 loss or congestion at a remote router.
2312 .It Cm plr Ar packet-loss-rate
2315 .Ar packet-loss-rate
2316 is a floating-point number between 0 and 1, with 0 meaning no
2317 loss, 1 meaning 100% loss.
2318 The loss rate is internally represented on 31 bits.
2320 .It Cm queue Brq Ar slots | size Ns Cm Kbytes
2325 Default value is 50 slots, which
2326 is the typical queue size for Ethernet devices.
2327 Note that for slow speed links you should keep the queue
2328 size short or your traffic might be affected by a significant
2330 E.g., 50 max-sized ethernet packets (1500 bytes) mean 600Kbit
2331 or 20s of queue on a 30Kbit/s pipe.
2332 Even worse effects can result if you get packets from an
2333 interface with a much larger MTU, e.g.\& the loopback interface
2334 with its 16KB packets.
2338 .Em net.inet.ip.dummynet.pipe_byte_limit
2340 .Em net.inet.ip.dummynet.pipe_slot_limit
2341 control the maximum lengths that can be specified.
2343 .It Cm red | gred Ar w_q Ns / Ns Ar min_th Ns / Ns Ar max_th Ns / Ns Ar max_p
2344 Make use of the RED (Random Early Detection) queue management algorithm.
2349 point numbers between 0 and 1 (0 not included), while
2353 are integer numbers specifying thresholds for queue management
2354 (thresholds are computed in bytes if the queue has been defined
2355 in bytes, in slots otherwise).
2358 also supports the gentle RED variant (gred).
2361 variables can be used to control the RED behaviour:
2362 .Bl -tag -width indent
2363 .It Va net.inet.ip.dummynet.red_lookup_depth
2364 specifies the accuracy in computing the average queue
2365 when the link is idle (defaults to 256, must be greater than zero)
2366 .It Va net.inet.ip.dummynet.red_avg_pkt_size
2367 specifies the expected average packet size (defaults to 512, must be
2369 .It Va net.inet.ip.dummynet.red_max_pkt_size
2370 specifies the expected maximum packet size, only used when queue
2371 thresholds are in bytes (defaults to 1500, must be greater than zero).
2375 When used with IPv6 data,
2377 currently has several limitations.
2378 Information necessary to route link-local packets to an
2379 interface is not available after processing by
2381 so those packets are dropped in the output path.
2382 Care should be taken to insure that link-local packets are not passed to
2385 Here are some important points to consider when designing your
2389 Remember that you filter both packets going
2393 Most connections need packets going in both directions.
2395 Remember to test very carefully.
2396 It is a good idea to be near the console when doing this.
2397 If you cannot be near the console,
2398 use an auto-recovery script such as the one in
2399 .Pa /usr/share/examples/ipfw/change_rules.sh .
2401 Do not forget the loopback interface.
2406 There are circumstances where fragmented datagrams are unconditionally
2408 TCP packets are dropped if they do not contain at least 20 bytes of
2409 TCP header, UDP packets are dropped if they do not contain a full 8
2410 byte UDP header, and ICMP packets are dropped if they do not contain
2411 4 bytes of ICMP header, enough to specify the ICMP type, code, and
2413 These packets are simply logged as
2415 since there may not be enough good data in the packet to produce a
2416 meaningful log entry.
2418 Another type of packet is unconditionally dropped, a TCP packet with a
2419 fragment offset of one.
2420 This is a valid packet, but it only has one use, to try
2421 to circumvent firewalls.
2422 When logging is enabled, these packets are
2423 reported as being dropped by rule -1.
2425 If you are logged in over a network, loading the
2429 is probably not as straightforward as you would think.
2430 The following command line is recommended:
2431 .Bd -literal -offset indent
2433 ipfw add 32000 allow ip from any to any
2436 Along the same lines, doing an
2437 .Bd -literal -offset indent
2441 in similar surroundings is also a bad idea.
2445 filter list may not be modified if the system security level
2446 is set to 3 or higher
2449 for information on system security levels).
2451 .Sh PACKET DIVERSION
2454 socket bound to the specified port will receive all packets
2455 diverted to that port.
2456 If no socket is bound to the destination port, or if the divert module is
2457 not loaded, or if the kernel was not compiled with divert socket support,
2458 the packets are dropped.
2459 .Sh NETWORK ADDRESS TRANSLATION (NAT)
2462 support in-kernel NAT using the kernel version of
2465 The nat configuration command is the following:
2466 .Bd -ragged -offset indent
2471 .Ar nat-configuration
2475 The following parameters can be configured:
2476 .Bl -tag -width indent
2477 .It Cm ip Ar ip_address
2478 Define an ip address to use for aliasing.
2480 Use ip address of NIC for aliasing, dynamically changing
2481 it if NIC's ip address changes.
2483 Enable logging on this nat instance.
2485 Deny any incoming connection from outside world.
2487 Try to leave the alias port numbers unchanged from
2488 the actual local port numbers.
2490 Traffic on the local network not originating from an
2491 unregistered address spaces will be ignored.
2493 Reset table of the packet aliasing engine on address change.
2495 Reverse the way libalias handles aliasing.
2497 Obey transparent proxy rules only, packet aliasing is not performed.
2499 Skip instance in case of global state lookup (see below).
2502 Some specials value can be supplied instead of
2504 .Bl -tag -width indent
2506 Looks up translation state in all configured nat instances.
2507 If an entry is found, packet is aliased according to that entry.
2508 If no entry was found in any of the instances, packet is passed unchanged,
2509 and no new entry will be created.
2511 .Sx MULTIPLE INSTANCES
2514 for more information.
2516 Uses argument supplied in lookup table. See
2518 section below for more information on lookup tables.
2521 To let the packet continue after being (de)aliased, set the sysctl variable
2522 .Va net.inet.ip.fw.one_pass
2524 For more information about aliasing modes, refer to
2528 for some examples about nat usage.
2529 .Ss REDIRECT AND LSNAT SUPPORT IN IPFW
2530 Redirect and LSNAT support follow closely the syntax used in
2534 for some examples on how to do redirect and lsnat.
2535 .Ss SCTP NAT SUPPORT
2536 SCTP nat can be configured in a similar manner to TCP through the
2539 The main difference is that
2541 does not do port translation.
2542 Since the local and global side ports will be the same,
2543 there is no need to specify both.
2544 Ports are redirected as follows:
2545 .Bd -ragged -offset indent
2551 .Cm redirect_port sctp
2552 .Ar ip_address [,addr_list] {[port | port-port] [,ports]}
2558 configuration can be done in real-time through the
2561 All may be changed dynamically, though the hash_table size will only
2566 .Sx SYSCTL VARIABLES
2569 Tunables can be set in
2575 before ipfw module gets loaded.
2576 .Bl -tag -width indent
2577 .It Va net.inet.ip.fw.default_to_accept: No 0
2578 Defines ipfw last rule behavior. This value overrides
2579 .Cd "options IPFW_DEFAULT_TO_(ACCEPT|DENY)"
2580 from kernel configuration file.
2581 .It Va net.inet.ip.fw.tables_max: No 128
2582 Defines number of tables available in ipfw. Number cannot exceed 65534.
2584 .Sh SYSCTL VARIABLES
2587 variables controls the behaviour of the firewall and
2589 .Pq Nm dummynet , bridge , sctp nat .
2590 These are shown below together with their default value
2591 (but always check with the
2593 command what value is actually in use) and meaning:
2594 .Bl -tag -width indent
2595 .It Va net.inet.ip.alias.sctp.accept_global_ootb_addip: No 0
2598 responds to receipt of global OOTB ASCONF-AddIP:
2599 .Bl -tag -width indent
2601 No response (unless a partially matching association exists -
2602 ports and vtags match but global address does not)
2605 will accept and process all OOTB global AddIP messages.
2608 Option 1 should never be selected as this forms a security risk.
2610 establish multiple fake associations by sending AddIP messages.
2611 .It Va net.inet.ip.alias.sctp.chunk_proc_limit: No 5
2612 Defines the maximum number of chunks in an SCTP packet that will be parsed for a
2613 packet that matches an existing association.
2614 This value is enforced to be greater or equal than
2615 .Cm net.inet.ip.alias.sctp.initialising_chunk_proc_limit .
2617 a DoS risk yet setting too low a value may result in important control chunks in
2618 the packet not being located and parsed.
2619 .It Va net.inet.ip.alias.sctp.error_on_ootb: No 1
2622 responds to any Out-of-the-Blue (OOTB) packets with ErrorM packets.
2623 An OOTB packet is a packet that arrives with no existing association
2626 and is not an INIT or ASCONF-AddIP packet:
2627 .Bl -tag -width indent
2629 ErrorM is never sent in response to OOTB packets.
2631 ErrorM is only sent to OOTB packets received on the local side.
2633 ErrorM is sent to the local side and on the global side ONLY if there is a
2634 partial match (ports and vtags match but the source global IP does not).
2635 This value is only useful if the
2637 is tracking global IP addresses.
2639 ErrorM is sent in response to all OOTB packets on both the local and global side
2643 At the moment the default is 0, since the ErrorM packet is not yet
2644 supported by most SCTP stacks.
2645 When it is supported, and if not tracking
2646 global addresses, we recommend setting this value to 1 to allow
2647 multi-homed local hosts to function with the
2649 To track global addresses, we recommend setting this value to 2 to
2650 allow global hosts to be informed when they need to (re)send an
2652 Value 3 should never be chosen (except for debugging) as the
2654 will respond to all OOTB global packets (a DoS risk).
2655 .It Va net.inet.ip.alias.sctp.hashtable_size: No 2003
2656 Size of hash tables used for
2658 lookups (100 < prime_number > 1000001).
2661 size for any future created
2663 instance and therefore must be set prior to creating a
2666 The table sizes may be changed to suit specific needs.
2667 If there will be few
2668 concurrent associations, and memory is scarce, you may make these smaller.
2669 If there will be many thousands (or millions) of concurrent associations, you
2670 should make these larger.
2671 A prime number is best for the table size.
2673 update function will adjust your input value to the next highest prime number.
2674 .It Va net.inet.ip.alias.sctp.holddown_time: No 0
2675 Hold association in table for this many seconds after receiving a
2677 This allows endpoints to correct shutdown gracefully if a
2678 shutdown_complete is lost and retransmissions are required.
2679 .It Va net.inet.ip.alias.sctp.init_timer: No 15
2680 Timeout value while waiting for (INIT-ACK|AddIP-ACK).
2681 This value cannot be 0.
2682 .It Va net.inet.ip.alias.sctp.initialising_chunk_proc_limit: No 2
2683 Defines the maximum number of chunks in an SCTP packet that will be parsed when
2684 no existing association exists that matches that packet.
2686 will only be an INIT or ASCONF-AddIP packet.
2687 A higher value may become a DoS
2688 risk as malformed packets can consume processing resources.
2689 .It Va net.inet.ip.alias.sctp.param_proc_limit: No 25
2690 Defines the maximum number of parameters within a chunk that will be parsed in a
2692 As for other similar sysctl variables, larger values pose a DoS risk.
2693 .It Va net.inet.ip.alias.sctp.log_level: No 0
2694 Level of detail in the system log messages (0 \- minimal, 1 \- event,
2695 2 \- info, 3 \- detail, 4 \- debug, 5 \- max debug). May be a good
2696 option in high loss environments.
2697 .It Va net.inet.ip.alias.sctp.shutdown_time: No 15
2698 Timeout value while waiting for SHUTDOWN-COMPLETE.
2699 This value cannot be 0.
2700 .It Va net.inet.ip.alias.sctp.track_global_addresses: No 0
2701 Enables/disables global IP address tracking within the
2704 upper limit on the number of addresses tracked for each association:
2705 .Bl -tag -width indent
2707 Global tracking is disabled
2709 Enables tracking, the maximum number of addresses tracked for each
2710 association is limited to this value
2713 This variable is fully dynamic, the new value will be adopted for all newly
2714 arriving associations, existing associations are treated as they were previously.
2715 Global tracking will decrease the number of collisions within the
2718 of increased processing load, memory usage, complexity, and possible
2721 problems in complex networks with multiple
2723 We recommend not tracking
2724 global IP addresses, this will still result in a fully functional
2726 .It Va net.inet.ip.alias.sctp.up_timer: No 300
2727 Timeout value to keep an association up with no traffic.
2728 This value cannot be 0.
2729 .It Va net.inet.ip.dummynet.expire : No 1
2730 Lazily delete dynamic pipes/queue once they have no pending traffic.
2731 You can disable this by setting the variable to 0, in which case
2732 the pipes/queues will only be deleted when the threshold is reached.
2733 .It Va net.inet.ip.dummynet.hash_size : No 64
2734 Default size of the hash table used for dynamic pipes/queues.
2735 This value is used when no
2737 option is specified when configuring a pipe/queue.
2738 .It Va net.inet.ip.dummynet.io_fast : No 0
2739 If set to a non-zero value,
2744 operation (see above) is enabled.
2745 .It Va net.inet.ip.dummynet.io_pkt
2746 Number of packets passed to
2748 .It Va net.inet.ip.dummynet.io_pkt_drop
2749 Number of packets dropped by
2751 .It Va net.inet.ip.dummynet.io_pkt_fast
2752 Number of packets bypassed by the
2755 .It Va net.inet.ip.dummynet.max_chain_len : No 16
2756 Target value for the maximum number of pipes/queues in a hash bucket.
2758 .Cm max_chain_len*hash_size
2759 is used to determine the threshold over which empty pipes/queues
2760 will be expired even when
2761 .Cm net.inet.ip.dummynet.expire=0 .
2762 .It Va net.inet.ip.dummynet.red_lookup_depth : No 256
2763 .It Va net.inet.ip.dummynet.red_avg_pkt_size : No 512
2764 .It Va net.inet.ip.dummynet.red_max_pkt_size : No 1500
2765 Parameters used in the computations of the drop probability
2766 for the RED algorithm.
2767 .It Va net.inet.ip.dummynet.pipe_byte_limit : No 1048576
2768 .It Va net.inet.ip.dummynet.pipe_slot_limit : No 100
2769 The maximum queue size that can be specified in bytes or packets.
2770 These limits prevent accidental exhaustion of resources such as mbufs.
2771 If you raise these limits,
2772 you should make sure the system is configured so that sufficient resources
2774 .It Va net.inet.ip.fw.autoinc_step : No 100
2775 Delta between rule numbers when auto-generating them.
2776 The value must be in the range 1..1000.
2777 .It Va net.inet.ip.fw.curr_dyn_buckets : Va net.inet.ip.fw.dyn_buckets
2778 The current number of buckets in the hash table for dynamic rules
2780 .It Va net.inet.ip.fw.debug : No 1
2781 Controls debugging messages produced by
2783 .It Va net.inet.ip.fw.default_rule : No 65535
2784 The default rule number (read-only).
2786 .Nm , the default rule is the last one, so its number
2787 can also serve as the highest number allowed for a rule.
2788 .It Va net.inet.ip.fw.dyn_buckets : No 256
2789 The number of buckets in the hash table for dynamic rules.
2790 Must be a power of 2, up to 65536.
2791 It only takes effect when all dynamic rules have expired, so you
2792 are advised to use a
2794 command to make sure that the hash table is resized.
2795 .It Va net.inet.ip.fw.dyn_count : No 3
2796 Current number of dynamic rules
2798 .It Va net.inet.ip.fw.dyn_keepalive : No 1
2799 Enables generation of keepalive packets for
2801 rules on TCP sessions.
2802 A keepalive is generated to both
2803 sides of the connection every 5 seconds for the last 20
2804 seconds of the lifetime of the rule.
2805 .It Va net.inet.ip.fw.dyn_max : No 8192
2806 Maximum number of dynamic rules.
2807 When you hit this limit, no more dynamic rules can be
2808 installed until old ones expire.
2809 .It Va net.inet.ip.fw.dyn_ack_lifetime : No 300
2810 .It Va net.inet.ip.fw.dyn_syn_lifetime : No 20
2811 .It Va net.inet.ip.fw.dyn_fin_lifetime : No 1
2812 .It Va net.inet.ip.fw.dyn_rst_lifetime : No 1
2813 .It Va net.inet.ip.fw.dyn_udp_lifetime : No 5
2814 .It Va net.inet.ip.fw.dyn_short_lifetime : No 30
2815 These variables control the lifetime, in seconds, of dynamic
2817 Upon the initial SYN exchange the lifetime is kept short,
2818 then increased after both SYN have been seen, then decreased
2819 again during the final FIN exchange or when a RST is received.
2821 .Em dyn_fin_lifetime
2823 .Em dyn_rst_lifetime
2824 must be strictly lower than 5 seconds, the period of
2825 repetition of keepalives.
2826 The firewall enforces that.
2827 .It Va net.inet.ip.fw.enable : No 1
2828 Enables the firewall.
2829 Setting this variable to 0 lets you run your machine without
2830 firewall even if compiled in.
2831 .It Va net.inet6.ip6.fw.enable : No 1
2832 provides the same functionality as above for the IPv6 case.
2833 .It Va net.inet.ip.fw.one_pass : No 1
2834 When set, the packet exiting from the
2838 node is not passed though the firewall again.
2839 Otherwise, after an action, the packet is
2840 reinjected into the firewall at the next rule.
2841 .It Va net.inet.ip.fw.tables_max : No 128
2842 Maximum number of tables (read-only).
2843 .It Va net.inet.ip.fw.verbose : No 1
2844 Enables verbose messages.
2845 .It Va net.inet.ip.fw.verbose_limit : No 0
2846 Limits the number of messages produced by a verbose firewall.
2847 .It Va net.inet6.ip6.fw.deny_unknown_exthdrs : No 1
2848 If enabled packets with unknown IPv6 Extension Headers will be denied.
2849 .It Va net.link.ether.ipfw : No 0
2850 Controls whether layer-2 packets are passed to
2853 .It Va net.link.bridge.ipfw : No 0
2854 Controls whether bridged packets are passed to
2860 There are far too many possible uses of
2862 so this Section will only give a small set of examples.
2864 .Ss BASIC PACKET FILTERING
2865 This command adds an entry which denies all tcp packets from
2866 .Em cracker.evil.org
2867 to the telnet port of
2869 from being forwarded by the host:
2871 .Dl "ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet"
2873 This one disallows any connection from the entire cracker's
2876 .Dl "ipfw add deny ip from 123.45.67.0/24 to my.host.org"
2878 A first and efficient way to limit access (not using dynamic rules)
2879 is the use of the following rules:
2881 .Dl "ipfw add allow tcp from any to any established"
2882 .Dl "ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup"
2883 .Dl "ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup"
2885 .Dl "ipfw add deny tcp from any to any"
2887 The first rule will be a quick match for normal TCP packets,
2888 but it will not match the initial SYN packet, which will be
2891 rules only for selected source/destination pairs.
2892 All other SYN packets will be rejected by the final
2896 If you administer one or more subnets, you can take advantage
2897 of the address sets and or-blocks and write extremely
2898 compact rulesets which selectively enable services to blocks
2899 of clients, as below:
2901 .Dl "goodguys=\*q{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }\*q"
2902 .Dl "badguys=\*q10.1.2.0/24{8,38,60}\*q"
2904 .Dl "ipfw add allow ip from ${goodguys} to any"
2905 .Dl "ipfw add deny ip from ${badguys} to any"
2906 .Dl "... normal policies ..."
2910 option could be used to do automated anti-spoofing by adding the
2911 following to the top of a ruleset:
2913 .Dl "ipfw add deny ip from any to any not verrevpath in"
2915 This rule drops all incoming packets that appear to be coming to the
2916 system on the wrong interface.
2917 For example, a packet with a source
2918 address belonging to a host on a protected internal network would be
2919 dropped if it tried to enter the system from an external interface.
2923 option could be used to do similar but more restricted anti-spoofing
2924 by adding the following to the top of a ruleset:
2926 .Dl "ipfw add deny ip from any to any not antispoof in"
2928 This rule drops all incoming packets that appear to be coming from another
2929 directly connected system but on the wrong interface.
2930 For example, a packet with a source address of
2931 .Li 192.168.0.0/24 ,
2938 In order to protect a site from flood attacks involving fake
2939 TCP packets, it is safer to use dynamic rules:
2941 .Dl "ipfw add check-state"
2942 .Dl "ipfw add deny tcp from any to any established"
2943 .Dl "ipfw add allow tcp from my-net to any setup keep-state"
2945 This will let the firewall install dynamic rules only for
2946 those connection which start with a regular SYN packet coming
2947 from the inside of our network.
2948 Dynamic rules are checked when encountering the first
2955 rule should usually be placed near the beginning of the
2956 ruleset to minimize the amount of work scanning the ruleset.
2957 Your mileage may vary.
2959 To limit the number of connections a user can open
2960 you can use the following type of rules:
2962 .Dl "ipfw add allow tcp from my-net/24 to any setup limit src-addr 10"
2963 .Dl "ipfw add allow tcp from any to me setup limit src-addr 4"
2965 The former (assuming it runs on a gateway) will allow each host
2966 on a /24 network to open at most 10 TCP connections.
2967 The latter can be placed on a server to make sure that a single
2968 client does not use more than 4 simultaneous connections.
2971 stateful rules can be subject to denial-of-service attacks
2972 by a SYN-flood which opens a huge number of dynamic rules.
2973 The effects of such attacks can be partially limited by
2976 variables which control the operation of the firewall.
2978 Here is a good usage of the
2980 command to see accounting records and timestamp information:
2984 or in short form without timestamps:
2988 which is equivalent to:
2992 Next rule diverts all incoming packets from 192.168.2.0/24
2993 to divert port 5000:
2995 .Dl ipfw divert 5000 ip from 192.168.2.0/24 to any in
2998 The following rules show some of the applications of
3002 for simulations and the like.
3004 This rule drops random incoming packets with a probability
3007 .Dl "ipfw add prob 0.05 deny ip from any to any in"
3009 A similar effect can be achieved making use of
3013 .Dl "ipfw add pipe 10 ip from any to any"
3014 .Dl "ipfw pipe 10 config plr 0.05"
3016 We can use pipes to artificially limit bandwidth, e.g.\& on a
3017 machine acting as a router, if we want to limit traffic from
3018 local clients on 192.168.2.0/24 we do:
3020 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
3021 .Dl "ipfw pipe 1 config bw 300Kbit/s queue 50KBytes"
3023 note that we use the
3025 modifier so that the rule is not used twice.
3026 Remember in fact that
3028 rules are checked both on incoming and outgoing packets.
3030 Should we want to simulate a bidirectional link with bandwidth
3031 limitations, the correct way is the following:
3033 .Dl "ipfw add pipe 1 ip from any to any out"
3034 .Dl "ipfw add pipe 2 ip from any to any in"
3035 .Dl "ipfw pipe 1 config bw 64Kbit/s queue 10Kbytes"
3036 .Dl "ipfw pipe 2 config bw 64Kbit/s queue 10Kbytes"
3038 The above can be very useful, e.g.\& if you want to see how
3039 your fancy Web page will look for a residential user who
3040 is connected only through a slow link.
3041 You should not use only one pipe for both directions, unless
3042 you want to simulate a half-duplex medium (e.g.\& AppleTalk,
3044 It is not necessary that both pipes have the same configuration,
3045 so we can also simulate asymmetric links.
3047 Should we want to verify network performance with the RED queue
3048 management algorithm:
3050 .Dl "ipfw add pipe 1 ip from any to any"
3051 .Dl "ipfw pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1"
3053 Another typical application of the traffic shaper is to
3054 introduce some delay in the communication.
3055 This can significantly affect applications which do a lot of Remote
3056 Procedure Calls, and where the round-trip-time of the
3057 connection often becomes a limiting factor much more than
3060 .Dl "ipfw add pipe 1 ip from any to any out"
3061 .Dl "ipfw add pipe 2 ip from any to any in"
3062 .Dl "ipfw pipe 1 config delay 250ms bw 1Mbit/s"
3063 .Dl "ipfw pipe 2 config delay 250ms bw 1Mbit/s"
3065 Per-flow queueing can be useful for a variety of purposes.
3066 A very simple one is counting traffic:
3068 .Dl "ipfw add pipe 1 tcp from any to any"
3069 .Dl "ipfw add pipe 1 udp from any to any"
3070 .Dl "ipfw add pipe 1 ip from any to any"
3071 .Dl "ipfw pipe 1 config mask all"
3073 The above set of rules will create queues (and collect
3074 statistics) for all traffic.
3075 Because the pipes have no limitations, the only effect is
3076 collecting statistics.
3077 Note that we need 3 rules, not just the last one, because
3080 tries to match IP packets it will not consider ports, so we
3081 would not see connections on separate ports as different
3084 A more sophisticated example is limiting the outbound traffic
3085 on a net with per-host limits, rather than per-network limits:
3087 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
3088 .Dl "ipfw add pipe 2 ip from any to 192.168.2.0/24 in"
3089 .Dl "ipfw pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
3090 .Dl "ipfw pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
3092 In the following example, we need to create several traffic bandwidth
3093 classes and we need different hosts/networks to fall into different classes.
3094 We create one pipe for each class and configure them accordingly.
3095 Then we create a single table and fill it with IP subnets and addresses.
3096 For each subnet/host we set the argument equal to the number of the pipe
3098 Then we classify traffic using a single rule:
3100 .Dl "ipfw pipe 1 config bw 1000Kbyte/s"
3101 .Dl "ipfw pipe 4 config bw 4000Kbyte/s"
3103 .Dl "ipfw table 1 add 192.168.2.0/24 1"
3104 .Dl "ipfw table 1 add 192.168.0.0/27 4"
3105 .Dl "ipfw table 1 add 192.168.0.2 1"
3107 .Dl "ipfw add pipe tablearg ip from table(1) to any"
3111 action, the table entries may include hostnames and IP addresses.
3113 .Dl "ipfw table 1 add 192.168.2.0/24 10.23.2.1"
3114 .Dl "ipfw table 1 add 192.168.0.0/27 router1.dmz"
3116 .Dl "ipfw add 100 fwd tablearg ip from any to table(1)"
3118 In the following example per-interface firewall is created:
3120 .Dl "ipfw table 10 add vlan20 12000"
3121 .Dl "ipfw table 10 add vlan30 13000"
3122 .Dl "ipfw table 20 add vlan20 22000"
3123 .Dl "ipfw table 20 add vlan30 23000"
3125 .Dl "ipfw add 100 ipfw skipto tablearg ip from any to any recv 'table(10)' in"
3126 .Dl "ipfw add 200 ipfw skipto tablearg ip from any to any xmit 'table(10)' out"
3128 To add a set of rules atomically, e.g.\& set 18:
3130 .Dl "ipfw set disable 18"
3131 .Dl "ipfw add NN set 18 ... # repeat as needed"
3132 .Dl "ipfw set enable 18"
3134 To delete a set of rules atomically the command is simply:
3136 .Dl "ipfw delete set 18"
3138 To test a ruleset and disable it and regain control if something goes wrong:
3140 .Dl "ipfw set disable 18"
3141 .Dl "ipfw add NN set 18 ... # repeat as needed"
3142 .Dl "ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18"
3144 Here if everything goes well, you press control-C before the "sleep"
3145 terminates, and your ruleset will be left active.
3146 Otherwise, e.g.\& if
3147 you cannot access your box, the ruleset will be disabled after
3148 the sleep terminates thus restoring the previous situation.
3150 To show rules of the specific set:
3152 .Dl "ipfw set 18 show"
3154 To show rules of the disabled set:
3156 .Dl "ipfw -S set 18 show"
3158 To clear a specific rule counters of the specific set:
3160 .Dl "ipfw set 18 zero NN"
3162 To delete a specific rule of the specific set:
3164 .Dl "ipfw set 18 delete NN"
3165 .Ss NAT, REDIRECT AND LSNAT
3166 First redirect all the traffic to nat instance 123:
3168 .Dl "ipfw add nat 123 all from any to any"
3170 Then to configure nat instance 123 to alias all the outgoing traffic with ip
3171 192.168.0.123, blocking all incoming connections, trying to keep
3172 same ports on both sides, clearing aliasing table on address change
3173 and keeping a log of traffic/link statistics:
3175 .Dl "ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports"
3177 Or to change address of instance 123, aliasing table will be cleared (see
3180 .Dl "ipfw nat 123 config ip 10.0.0.1"
3182 To see configuration of nat instance 123:
3184 .Dl "ipfw nat 123 show config"
3186 To show logs of all the instances in range 111-999:
3188 .Dl "ipfw nat 111-999 show"
3190 To see configurations of all instances:
3192 .Dl "ipfw nat show config"
3194 Or a redirect rule with mixed modes could looks like:
3196 .Dl "ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66"
3197 .Dl " redirect_port tcp 192.168.0.1:80 500"
3198 .Dl " redirect_proto udp 192.168.1.43 192.168.1.1"
3199 .Dl " redirect_addr 192.168.0.10,192.168.0.11"
3200 .Dl " 10.0.0.100 # LSNAT"
3201 .Dl " redirect_port tcp 192.168.0.1:80,192.168.0.10:22"
3204 or it could be split in:
3206 .Dl "ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66"
3207 .Dl "ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500"
3208 .Dl "ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1"
3209 .Dl "ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12"
3211 .Dl "ipfw nat 5 config redirect_port tcp"
3212 .Dl " 192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500"
3234 utility first appeared in
3239 Stateful extensions were introduced in
3242 was introduced in Summer 2002.
3244 .An Ugen J. S. Antsilevich ,
3245 .An Poul-Henning Kamp ,
3251 API based upon code written by
3255 Dummynet has been introduced by Luigi Rizzo in 1997-1998.
3257 Some early work (1999-2000) on the
3259 traffic shaper supported by Akamba Corp.
3261 The ipfw core (ipfw2) has been completely redesigned and
3262 reimplemented by Luigi Rizzo in summer 2002. Further
3264 options have been added by various developer over the years.
3267 In-kernel NAT support written by
3268 .An Paolo Pisati Aq piso@FreeBSD.org
3269 as part of a Summer of Code 2005 project.
3273 support has been developed by
3274 .An The Centre for Advanced Internet Architectures (CAIA) Aq http://www.caia.swin.edu.au .
3275 The primary developers and maintainers are David Hayes and Jason But.
3276 For further information visit:
3277 .Aq http://www.caia.swin.edu.au/urp/SONATA
3279 Delay profiles have been developed by Alessandro Cerri and
3280 Luigi Rizzo, supported by the
3281 European Commission within Projects Onelab and Onelab2.
3283 The syntax has grown over the years and sometimes it might be confusing.
3284 Unfortunately, backward compatibility prevents cleaning up mistakes
3285 made in the definition of the syntax.
3289 Misconfiguring the firewall can put your computer in an unusable state,
3290 possibly shutting down network services and requiring console access to
3291 regain control of it.
3293 Incoming packet fragments diverted by
3295 are reassembled before delivery to the socket.
3296 The action used on those packet is the one from the
3297 rule which matches the first fragment of the packet.
3299 Packets diverted to userland, and then reinserted by a userland process
3300 may lose various packet attributes.
3301 The packet source interface name
3302 will be preserved if it is shorter than 8 bytes and the userland process
3303 saves and reuses the sockaddr_in
3306 otherwise, it may be lost.
3307 If a packet is reinserted in this manner, later rules may be incorrectly
3308 applied, making the order of
3310 rules in the rule sequence very important.
3312 Dummynet drops all packets with IPv6 link-local addresses.
3318 may not behave as expected.
3319 In particular, incoming SYN packets may
3320 have no uid or gid associated with them since they do not yet belong
3321 to a TCP connection, and the uid/gid associated with a packet may not
3322 be as expected if the associated process calls
3324 or similar system calls.
3326 Rule syntax is subject to the command line environment and some patterns
3327 may need to be escaped with the backslash character
3328 or quoted appropriately.
3330 Due to the architecture of
3332 ipfw nat is not compatible with the TCP segmentation offloading (TSO).
3333 Thus, to reliably nat your network traffic, please disable TSO
3337 ICMP error messages are not implicitly matched by dynamic rules
3338 for the respective conversations.
3339 To avoid failures of network error detection and path MTU discovery,
3340 ICMP error messages may need to be allowed explicitly through static
3347 actions may lead to confusing behaviour if ruleset has mistakes,
3348 and/or interaction with other subsystems (netgraph, dummynet, etc.) is used.
3349 One possible case for this is packet leaving
3351 in subroutine on the input pass, while later on output encountering unpaired
3354 As the call stack is kept intact after input pass, packet will suddenly
3355 return to the rule number used on input pass, not on output one.
3356 Order of processing should be checked carefully to avoid such mistakes.