FIREWALL CONFIGURATION

ipfw	[ -cq] add rule

ipfw	[ -acdefnNStT][ set N]{ list | show}[ rule | first-last ...]

ipfw	[ -f | -q][ set N] flush

ipfw	[ -q][ set N]{ delete | zero | resetlog}[ number ...]

ipfw	set [ disable number ...][ enable number ...]

ipfw	set move [ rule] number to number

ipfw	set swap number number

ipfw

set show

SYSCTL SHORTCUTS

ipfw	enable { firewall | altq | one_pass | debug | verbose | dyn_keepalive}

ipfw	disable { firewall | altq | one_pass | debug | verbose | dyn_keepalive}

LOOKUP TABLES

ipfw	table number add addr[ / masklen][ value]

ipfw	table number delete addr[ / masklen]

ipfw	table { number | all} flush

ipfw	table { number | all} list

DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)

ipfw	{ pipe | queue | sched} number config config-options

ipfw	[ -s [ field]]{ pipe | queue | sched}{ delete | list | show}[ number ...]

IN-KERNEL NAT

ipfw	[ -q] nat number config config-options

ipfw	[ -cfnNqS][ -p preproc [ preproc-flags]] pathname

COMMAND OPTIONS

The following general options are available when invoking ipfw:

-a

Show counter values when listing rules. The show command implies this option.

-b

Only show the action and the comment, not the body of a rule. Implies -c.

-c

When entering or showing rules, print them in compact form, i.e., omitting the "ip from any to any" string when this does not carry any additional information.

-d

When listing, show dynamic rules in addition to static ones.

-e

When listing and -d is specified, also show expired dynamic rules.

-f

Do not ask for confirmation for commands that can cause problems if misused, i.e., flush. If there is no tty associated with the process, this is implied.

-i

When listing a table (see the LOOKUP TABLES section below for more information on lookup tables), format values as IP addresses. By default, values are shown as integers.

-n

Only check syntax of the command strings, without actually passing them to the kernel.

-N

Try to resolve addresses and service names in output.

-q

Be quiet when executing the add, nat, zero, resetlog or flush commands; (implies -f). This is useful when updating rulesets by executing multiple ipfw commands in a script (e.g., ‘ sh /etc/rc.firewall’), or by processing a file with many ipfw rules across a remote login session. It also stops a table add or delete from failing if the entry already exists or is not present.

The reason why this option may be important is that for some of these actions, ipfw may print a message; if the action results in blocking the traffic to the remote client, the remote login session will be closed and the rest of the ruleset will not be processed. Access to the console would then be required to recover.

-S

When listing rules, show the set each rule belongs to. If this flag is not specified, disabled rules will not be listed.

-s [ field]

When listing pipes, sort according to one of the four counters (total or current packets or bytes).

-t

When listing, show last match timestamp converted with ctime().

-T

When listing, show last match timestamp as seconds from the epoch. This form can be more convenient for postprocessing by scripts.

LIST OF RULES AND PREPROCESSING

To ease configuration, rules can be put into a file which is processed using ipfw as shown in the last synopsis line. An absolute pathname must be used. The file will be read line by line and applied as arguments to the ipfw utility.

Optionally, a preprocessor can be specified using -p preproc where pathname is to be piped through. Useful preprocessors include cpp(1) and m4(1). If preproc does not start with a slash (‘ /’) as its first character, the usual PATH name search is performed. Care should be taken with this in environments where not all file systems are mounted (yet) by the time ipfw is being run (e.g. when they are mounted over NFS). Once -p has been specified, any additional arguments are passed on to the preprocessor for interpretation. This allows for flexible configuration files (like conditionalizing them on the local hostname) and the use of macros to centralize frequently required arguments like IP addresses.

TRAFFIC SHAPER CONFIGURATION

The ipfw pipe, queue and sched commands are used to configure the traffic shaper and packet scheduler. See the TRAFFIC SHAPER (DUMMYNET) CONFIGURATION Section below for details.

If the world and the kernel get out of sync the ipfw ABI may break, preventing you from being able to add any rules. This can adversely effect the booting process. You can use ipfw disable firewall to temporarily disable the firewall to regain access to the network, allowing you to fix the problem.

RULE ACTIONS

A rule can be associated with one of the following actions, which will be executed when the packet matches the body of the rule.

allow | accept | pass | permit

Allow packets that match rule. The search terminates.

check-state

Checks the packet against the dynamic ruleset. If a match is found, execute the action associated with the rule which generated this dynamic rule, otherwise move to the next rule.

 

Check-state rules do not have a body. If no check-state rule is found, the dynamic ruleset is checked at the first keep-state or limit rule.

count

Update counters for all packets that match rule. The search continues with the next rule.

deny | drop

Discard packets that match this rule. The search terminates.

divert port

Divert packets that match this rule to the divert(4) socket bound to port port. The search terminates.

fwd | forward ipaddr | tablearg[ , port]

Change the next-hop on matching packets to ipaddr, which can be an IP address or a host name. For IPv4, the next hop can also be supplied by the last table looked up for the packet by using the tablearg keyword instead of an explicit address. The search terminates if this rule matches.

If ipaddr is a local address, then matching packets will be forwarded to port (or the port number in the packet if one is not specified in the rule) on the local machine.

 

If ipaddr is not a local address, then the port number (if specified) is ignored, and the packet will be forwarded to the remote address, using the route as found in the local routing table for that IP.

 

A fwd rule will not match layer-2 packets (those received on ether_input, ether_output, or bridged).

 

The fwd action does not change the contents of the packet at all. In particular, the destination address remains unmodified, so packets forwarded to another system will usually be rejected by that system unless there is a matching rule on that system to capture them. For packets forwarded locally, the local address of the socket will be set to the original destination address of the packet. This makes the netstat(1) entry look rather weird but is intended for use with transparent proxy servers.

nat nat_nr | tablearg

Pass packet to a nat instance (for network address translation, address redirect, etc.): see the NETWORK ADDRESS TRANSLATION (NAT) Section for further information.

pipe pipe_nr

Pass packet to a dummynet “pipe” (for bandwidth limitation, delay, etc.). See the TRAFFIC SHAPER (DUMMYNET) CONFIGURATION Section for further information. The search terminates; however, on exit from the pipe and if the sysctl(8) variable net.inet.ip.fw.one_pass is not set, the packet is passed again to the firewall code starting from the next rule.

queue queue_nr

Pass packet to a dummynet “queue” (for bandwidth limitation using WF2Q+).

reject

(Deprecated). Synonym for unreach host.

reset

Discard packets that match this rule, and if the packet is a TCP packet, try to send a TCP reset (RST) notice. The search terminates.

reset6

Discard packets that match this rule, and if the packet is a TCP packet, try to send a TCP reset (RST) notice. The search terminates.

skipto number | tablearg

Skip all subsequent rules numbered less than number. The search continues with the first rule numbered number or higher. It is possible to use the tablearg keyword with a skipto for a computed skipto, but care should be used, as no destination caching is possible in this case so the rules are always walked to find it, starting from the skipto.

call number | tablearg

The current rule number is saved in the internal stack and ruleset processing continues with the first rule numbered number or higher. If later a rule with the return action is encountered, the processing returns to the first rule with number of this call rule plus one or higher (the same behaviour as with packets returning from divert(4) socket after a divert action). This could be used to make somewhat like an assembly language “subroutine” calls to rules with common checks for different interfaces, etc.

Rule with any number could be called, not just forward jumps as with skipto. So, to prevent endless loops in case of mistakes, both call and return actions don't do any jumps and simply go to the next rule if memory cannot be allocated or stack overflowed/underflowed.

Internally stack for rule numbers is implemented using mbuf_tags(9) facility and currently has size of 16 entries. As mbuf tags are lost when packet leaves the kernel, divert should not be used in subroutines to avoid endless loops and other undesired effects.

return

Takes rule number saved to internal stack by the last call action and returns ruleset processing to the first rule with number greater than number of corresponding call rule. See description of the call action for more details.

Note that return rules usually end a “subroutine” and thus are unconditional, but ipfw command-line utility currently requires every action except check-state to have body. While it is sometimes useful to return only on some packets, usually you want to print just “return” for readability. A workaround for this is to use new syntax and -c switch:

# Add a rule without actual body 
ipfw add 2999 return via any 
 
# List rules without "from any to any" part 
ipfw -c list

This cosmetic annoyance may be fixed in future releases.

tee port

Send a copy of packets matching this rule to the divert(4) socket bound to port port. The search continues with the next rule.

unreach code

Discard packets that match this rule, and try to send an ICMP unreachable notice with code code, where code is a number from 0 to 255, or one of these aliases: net, host, protocol, port, needfrag, srcfail, net-unknown, host-unknown, isolated, net-prohib, host-prohib, tosnet, toshost, filter-prohib, host-precedence or precedence-cutoff. The search terminates.

unreach6 code

Discard packets that match this rule, and try to send an ICMPv6 unreachable notice with code code, where code is a number from 0, 1, 3 or 4, or one of these aliases: no-route, admin-prohib, address or port. The search terminates.

netgraph cookie

Divert packet into netgraph with given cookie. The search terminates. If packet is later returned from netgraph it is either accepted or continues with the next rule, depending on net.inet.ip.fw.one_pass sysctl variable.

ngtee cookie

A copy of packet is diverted into netgraph, original packet continues with the next rule. See ng_ipfw(4) for more information on netgraph and ngtee actions.

setfib fibnum | tablearg

The packet is tagged so as to use the FIB (routing table) fibnum in any subsequent forwarding decisions. In the current implementation, this is limited to the values 0 through 15, see setfib(2). Processing continues at the next rule. It is possible to use the tablearg keyword with setfib. If the tablearg value is not within the compiled range of fibs, the packet's fib is set to 0.

setdscp DSCP | number | tablearg

Set specified DiffServ codepoint for an IPv4/IPv6 packet. Processing continues at the next rule. Supported values are:

CS0 ( 000000), CS1 ( 001000), CS2 ( 010000), CS3 ( 011000), CS4 ( 100000), CS5 ( 101000), CS6 ( 110000), CS7 ( 111000), AF11 ( 001010), AF12 ( 001100), AF13 ( 001110), AF21 ( 010010), AF22 ( 010100), AF23 ( 010110), AF31 ( 011010), AF32 ( 011100), AF33 ( 011110), AF41 ( 100010), AF42 ( 100100), AF43 ( 100110), EF ( 101110), BE ( 000000). Additionally, DSCP value can be specified by number (0..64). It is also possible to use the tablearg keyword with setdscp. If the tablearg value is not within the 0..64 range, lower 6 bits of supplied value are used.

reass

Queue and reassemble IP fragments. If the packet is not fragmented, counters are updated and processing continues with the next rule. If the packet is the last logical fragment, the packet is reassembled and, if net.inet.ip.fw.one_pass is set to 0, processing continues with the next rule. Otherwise, the packet is allowed to pass and the search terminates. If the packet is a fragment in the middle of a logical group of fragments, it is consumed and processing stops immediately.

Fragment handling can be tuned via net.inet.ip.maxfragpackets and net.inet.ip.maxfragsperpacket which limit, respectively, the maximum number of processable fragments (default: 800) and the maximum number of fragments per packet (default: 16).

NOTA BENE: since fragments do not contain port numbers, they should be avoided with the reass rule. Alternatively, direction-based (like in / out ) and source-based (like via ) match patterns can be used to select fragments.

Usually a simple rule like:

# reassemble incoming fragments 
ipfw add reass all from any to any in

is all you need at the beginning of your ruleset.

RULE BODY

The body of a rule contains zero or more patterns (such as specific source and destination addresses or ports, protocol options, incoming or outgoing interfaces, etc.) that the packet must match in order to be recognised. In general, the patterns are connected by (implicit) and operators -- i.e., all must match in order for the rule to match. Individual patterns can be prefixed by the not operator to reverse the result of the match, as in

ipfw add 100 allow ip from not 1.2.3.4 to any

Additionally, sets of alternative match patterns ( or-blocks) can be constructed by putting the patterns in lists enclosed between parentheses ( ) or braces { }, and using the or operator as follows:

ipfw add 100 allow ip from { x or not y or z } to any

Only one level of parentheses is allowed. Beware that most shells have special meanings for parentheses or braces, so it is advisable to put a backslash \ in front of them to prevent such interpretations.

The body of a rule must in general include a source and destination address specifier. The keyword any can be used in various places to specify that the content of a required field is irrelevant.

The rule body has the following format:

The first part (proto from src to dst) is for backward compatibility with earlier versions of FreeBSD. In modern FreeBSD any match pattern (including MAC headers, IP protocols, addresses and ports) can be specified in the options section.

Rule fields have the following meaning:

proto: protocol | { protocol or ... }

protocol: [ not] protocol-name | protocol-number

An IP protocol specified by number or name (for a complete list see /etc/protocols), or one of the following keywords:

ip4 | ipv4

Matches IPv4 packets.

ip6 | ipv6

Matches IPv6 packets.

ip | all

Matches any packet.

The ipv6 in proto option will be treated as inner protocol. And, the ipv4 is not available in proto option.

The { protocol or ... } format (an or-block) is provided for convenience only but its use is deprecated.

src and dst: { addr | { addr or ... }}[ [ not] ports]

An address (or a list, see below) optionally followed by ports specifiers.

The second format ( or-block with multiple addresses) is provided for convenience only and its use is discouraged.

addr: [ not]{ any | me | me6 | table( number[ , value]) | addr-list | addr-set}

any

matches any IP address.

me

matches any IP address configured on an interface in the system.

me6

matches any IPv6 address configured on an interface in the system. The address list is evaluated at the time the packet is analysed.

table( number[ , value])

Matches any IPv4 address for which an entry exists in the lookup table number. If an optional 32-bit unsigned value is also specified, an entry will match only if it has this value. See the LOOKUP TABLES section below for more information on lookup tables.

addr-list: ip-addr[ , addr-list]

ip-addr:

A host or subnet address specified in one of the following ways:

numeric-ip | hostname

Matches a single IPv4 address, specified as dotted-quad or a hostname. Hostnames are resolved at the time the rule is added to the firewall list.

addr/ masklen

Matches all addresses with base addr (specified as an IP address, a network number, or a hostname) and mask width of masklen bits. As an example, 1.2.3.4/25 or 1.2.3.0/25 will match all IP numbers from 1.2.3.0 to 1.2.3.127 .

addr: mask

Matches all addresses with base addr (specified as an IP address, a network number, or a hostname) and the mask of mask, specified as a dotted quad. As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match 1.*.3.*. This form is advised only for non-contiguous masks. It is better to resort to the addr/ masklen format for contiguous masks, which is more compact and less error-prone.

addr-set: addr[ / masklen] { list }

list: { num | num-num}[ , list]

Matches all addresses with base address addr (specified as an IP address, a network number, or a hostname) and whose last byte is in the list between braces { } . Note that there must be no spaces between braces and numbers (spaces after commas are allowed). Elements of the list can be specified as single entries or ranges. The masklen field is used to limit the size of the set of addresses, and can have any value between 24 and 32. If not specified, it will be assumed as 24.

 

This format is particularly useful to handle sparse address sets within a single rule. Because the matching occurs using a bitmask, it takes constant time and dramatically reduces the complexity of rulesets.

 

As an example, an address specified as 1.2.3.4/24{128,35-55,89} or 1.2.3.0/24{128,35-55,89} will match the following IP addresses:

 

1.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .

addr6-list: ip6-addr[ , addr6-list]

ip6-addr:

A host or subnet specified one of the following ways:

numeric-ip | hostname

Matches a single IPv6 address as allowed by inet_pton(3) or a hostname. Hostnames are resolved at the time the rule is added to the firewall list.

addr/ masklen

Matches all IPv6 addresses with base addr (specified as allowed by inet_pton or a hostname) and mask width of masklen bits.

No support for sets of IPv6 addresses is provided because IPv6 addresses are typically random past the initial prefix.

ports: { port | port- port}[ , ports]

For protocols which support port numbers (such as TCP and UDP), optional ports may be specified as one or more ports or port ranges, separated by commas but no spaces, and an optional not operator. The ‘ -’ notation specifies a range of ports (including boundaries).

Service names (from /etc/services) may be used instead of numeric port values. The length of the port list is limited to 30 ports or ranges, though one can specify larger ranges by using an or-block in the options section of the rule.

A backslash (‘ \’) can be used to escape the dash (‘ -’) character in a service name (from a shell, the backslash must be typed twice to avoid the shell itself interpreting it as an escape character).

ipfw add count tcp from any ftp\\-data-ftp to any

Fragmented packets which have a non-zero offset (i.e., not the first fragment) will never match a rule which has one or more port specifications. See the frag option for details on matching fragmented packets.

RULE OPTIONS (MATCH PATTERNS)

Additional match patterns can be used within rules. Zero or more of these so-called options can be present in a rule, optionally prefixed by the not operand, and possibly grouped into or-blocks.

The following match patterns can be used (listed in alphabetical order):

// this is a comment.

Inserts the specified text as a comment in the rule. Everything following // is considered as a comment and stored in the rule. You can have comment-only rules, which are listed as having a count action followed by the comment.

bridged

Alias for layer2.

diverted

Matches only packets generated by a divert socket.

diverted-loopback

Matches only packets coming from a divert socket back into the IP stack input for delivery.

diverted-output

Matches only packets going from a divert socket back outward to the IP stack output for delivery.

dst-ip ip-address

Matches IPv4 packets whose destination IP is one of the address(es) specified as argument.

{ dst-ip6 | dst-ipv6} ip6-address

Matches IPv6 packets whose destination IP is one of the address(es) specified as argument.

dst-port ports

Matches IP packets whose destination port is one of the port(s) specified as argument.

established

Matches TCP packets that have the RST or ACK bits set.

ext6hdr header

Matches IPv6 packets containing the extended header given by header. Supported headers are:

Fragment, ( frag), Hop-to-hop options ( hopopt), any type of Routing Header ( route), Source routing Routing Header Type 0 ( rthdr0), Mobile IPv6 Routing Header Type 2 ( rthdr2), Destination options ( dstopt), IPSec authentication headers ( ah), and IPsec encapsulated security payload headers ( esp).

fib fibnum

Matches a packet that has been tagged to use the given FIB (routing table) number.

flow-id labels

Matches IPv6 packets containing any of the flow labels given in labels. labels is a comma separated list of numeric flow labels.

frag

Matches packets that are fragments and not the first fragment of an IP datagram. Note that these packets will not have the next protocol header (e.g. TCP, UDP) so options that look into these headers cannot match.

gid group

Matches all TCP or UDP packets sent by or received for a group. A group may be specified by name or number.

jail prisonID

Matches all TCP or UDP packets sent by or received for the jail whos prison ID is prisonID.

icmptypes types

Matches ICMP packets whose ICMP type is in the list types. The list may be specified as any combination of individual types (numeric) separated by commas. Ranges are not allowed. The supported ICMP types are:

echo reply ( 0), destination unreachable ( 3), source quench ( 4), redirect ( 5), echo request ( 8), router advertisement ( 9), router solicitation ( 10), time-to-live exceeded ( 11), IP header bad ( 12), timestamp request ( 13), timestamp reply ( 14), information request ( 15), information reply ( 16), address mask request ( 17) and address mask reply ( 18).

icmp6types types

Matches ICMP6 packets whose ICMP6 type is in the list of types. The list may be specified as any combination of individual types (numeric) separated by commas. Ranges are not allowed.

in | out

Matches incoming or outgoing packets, respectively. in and out are mutually exclusive (in fact, out is implemented as not in).

ipid id-list

Matches IPv4 packets whose ip_id field has value included in id-list, which is either a single value or a list of values or ranges specified in the same way as ports.

iplen len-list

Matches IP packets whose total length, including header and data, is in the set len-list, which is either a single value or a list of values or ranges specified in the same way as ports.

ipoptions spec

Matches packets whose IPv4 header contains the comma separated list of options specified in spec. The supported IP options are:

ssrr (strict source route), lsrr (loose source route), rr (record packet route) and ts (timestamp). The absence of a particular option may be denoted with a ‘ !’.

ipprecedence precedence

Matches IPv4 packets whose precedence field is equal to precedence.

ipsec

Matches packets that have IPSEC history associated with them (i.e., the packet comes encapsulated in IPSEC, the kernel has IPSEC support and IPSEC_FILTERTUNNEL option, and can correctly decapsulate it).

Note that specifying ipsec is different from specifying proto ipsec as the latter will only look at the specific IP protocol field, irrespective of IPSEC kernel support and the validity of the IPSEC data.

Further note that this flag is silently ignored in kernels without IPSEC support. It does not affect rule processing when given and the rules are handled as if with no ipsec flag.

iptos spec

Matches IPv4 packets whose tos field contains the comma separated list of service types specified in spec. The supported IP types of service are:

lowdelay ( IPTOS_LOWDELAY), throughput ( IPTOS_THROUGHPUT), reliability ( IPTOS_RELIABILITY), mincost ( IPTOS_MINCOST), congestion ( IPTOS_ECN_CE). The absence of a particular type may be denoted with a ‘ !’.

dscp spec[ , spec]

Matches IPv4/IPv6 packets whose DS field value is contained in spec mask. Multiple values can be specified via the comma separated list. Value can be one of keywords used in setdscp action or exact number.

ipttl ttl-list

Matches IPv4 packets whose time to live is included in ttl-list, which is either a single value or a list of values or ranges specified in the same way as ports.

ipversion ver

Matches IP packets whose IP version field is ver.

keep-state

Upon a match, the firewall will create a dynamic rule, whose default behaviour is to match bidirectional traffic between source and destination IP/port using the same protocol. The rule has a limited lifetime (controlled by a set of sysctl(8) variables), and the lifetime is refreshed every time a matching packet is found.

layer2

Matches only layer2 packets, i.e., those passed to ipfw from ether_demux() and ether_output_frame().

limit { src-addr | src-port | dst-addr | dst-port} N

The firewall will only allow N connections with the same set of parameters as specified in the rule. One or more of source and destination addresses and ports can be specified. Currently, only IPv4 flows are supported.

lookup { dst-ip | dst-port | src-ip | src-port | uid | jail} N

Search an entry in lookup table N that matches the field specified as argument. If not found, the match fails. Otherwise, the match succeeds and tablearg is set to the value extracted from the table.

This option can be useful to quickly dispatch traffic based on certain packet fields. See the LOOKUP TABLES section below for more information on lookup tables.

{ MAC | mac } dst-mac src-mac

Match packets with a given dst-mac and src-mac addresses, specified as the any keyword (matching any MAC address), or six groups of hex digits separated by colons, and optionally followed by a mask indicating the significant bits. The mask may be specified using either of the following methods:

1. A slash (/) followed by the number of significant bits. For example, an address with 33 significant bits could be specified as:

   MAC 10:20:30:40:50:60/33 any

2. An ampersand (&) followed by a bitmask specified as six groups of hex digits separated by colons. For example, an address in which the last 16 bits are significant could be specified as:

   MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any

   Note that the ampersand character has a special meaning in many shells and should generally be escaped.

Note that the order of MAC addresses (destination first, source second) is the same as on the wire, but the opposite of the one used for IP addresses.

mac-type mac-type

Matches packets whose Ethernet Type field corresponds to one of those specified as argument. mac-type is specified in the same way as port numbers (i.e., one or more comma-separated single values or ranges). You can use symbolic names for known values such as vlan, ipv4, ipv6. Values can be entered as decimal or hexadecimal (if prefixed by 0x), and they are always printed as hexadecimal (unless the -N option is used, in which case symbolic resolution will be attempted).

proto protocol

Matches packets with the corresponding IP protocol.

recv | xmit | via { ifX | if * | table( number[ , value | ipno | any])}

Matches packets received, transmitted or going through, respectively, the interface specified by exact name ( ifX), by device name ( if*), by IP address, or through some interface.

The via keyword causes the interface to always be checked. If recv or xmit is used instead of via, then only the receive or transmit interface (respectively) is checked. By specifying both, it is possible to match packets based on both receive and transmit interface, e.g.:

ipfw add deny ip from any to any out recv ed0 xmit ed1

The recv interface can be tested on either incoming or outgoing packets, while the xmit interface can only be tested on outgoing packets. So out is required (and in is invalid) whenever xmit is used.

A packet might not have a receive or transmit interface: packets originating from the local host have no receive interface, while packets destined for the local host have no transmit interface.

setup

Matches TCP packets that have the SYN bit set but no ACK bit. This is the short form of “ tcpflags syn,!ack”.

sockarg

Matches packets that are associated to a local socket and for which the SO_USER_COOKIE socket option has been set to a non-zero value. As a side effect, the value of the option is made available as tablearg value, which in turn can be used as skipto or pipe number.

src-ip ip-address

Matches IPv4 packets whose source IP is one of the address(es) specified as an argument.

src-ip6 ip6-address

Matches IPv6 packets whose source IP is one of the address(es) specified as an argument.

src-port ports

Matches IP packets whose source port is one of the port(s) specified as argument.

tagged tag-list

Matches packets whose tags are included in tag-list, which is either a single value or a list of values or ranges specified in the same way as ports. Tags can be applied to the packet using tag rule action parameter (see it's description for details on tags).

tcpack ack

TCP packets only. Match if the TCP header acknowledgment number field is set to ack.

tcpdatalen tcpdatalen-list

Matches TCP packets whose length of TCP data is tcpdatalen-list, which is either a single value or a list of values or ranges specified in the same way as ports.

tcpflags spec

TCP packets only. Match if the TCP header contains the comma separated list of flags specified in spec. The supported TCP flags are:

fin, syn, rst, psh, ack and urg. The absence of a particular flag may be denoted with a ‘ !’. A rule which contains a tcpflags specification can never match a fragmented packet which has a non-zero offset. See the frag option for details on matching fragmented packets.

tcpseq seq

TCP packets only. Match if the TCP header sequence number field is set to seq.

tcpwin tcpwin-list

Matches TCP packets whose header window field is set to tcpwin-list, which is either a single value or a list of values or ranges specified in the same way as ports.

tcpoptions spec

TCP packets only. Match if the TCP header contains the comma separated list of options specified in spec. The supported TCP options are:

mss (maximum segment size), window (tcp window advertisement), sack (selective ack), ts (rfc1323 timestamp) and cc (rfc1644 t/tcp connection count). The absence of a particular option may be denoted with a ‘ !’.

uid user

Match all TCP or UDP packets sent by or received for a user. A user may be matched by name or identification number.

verrevpath

For incoming packets, a routing table lookup is done on the packet's source address. If the interface on which the packet entered the system matches the outgoing interface for the route, the packet matches. If the interfaces do not match up, the packet does not match. All outgoing packets or packets with no incoming interface match.

The name and functionality of the option is intentionally similar to the Cisco IOS command:

ip verify unicast reverse-path

This option can be used to make anti-spoofing rules to reject all packets with source addresses not from this interface. See also the option antispoof.

versrcreach

For incoming packets, a routing table lookup is done on the packet's source address. If a route to the source address exists, but not the default route or a blackhole/reject route, the packet matches. Otherwise, the packet does not match. All outgoing packets match.

The name and functionality of the option is intentionally similar to the Cisco IOS command:

ip verify unicast source reachable-via any

This option can be used to make anti-spoofing rules to reject all packets whose source address is unreachable.

antispoof

For incoming packets, the packet's source address is checked if it belongs to a directly connected network. If the network is directly connected, then the interface the packet came on in is compared to the interface the network is connected to. When incoming interface and directly connected interface are not the same, the packet does not match. Otherwise, the packet does match. All outgoing packets match.

This option can be used to make anti-spoofing rules to reject all packets that pretend to be from a directly connected network but do not come in through that interface. This option is similar to but more restricted than verrevpath because it engages only on packets with source addresses of directly connected networks instead of all source addresses.

                 (flow_mask|sched_mask)  sched_mask 
         +---------+   weight Wx  +-------------+ 
         |         |->-[flow]-->--|             |-+ 
    -->--| QUEUE x |   ...        |             | | 
         |         |->-[flow]-->--| SCHEDuler N | | 
         +---------+              |             | | 
             ...                  |             +--[LINK N]-->-- 
         +---------+   weight Wy  |             | +--[LINK N]-->-- 
         |         |->-[flow]-->--|             | | 
    -->--| QUEUE y |   ...        |             | | 
         |         |->-[flow]-->--|             | | 
         +---------+              +-------------+ | 
                                    +-------------+

PIPE, QUEUE AND SCHEDULER CONFIGURATION

The pipe, queue and scheduler configuration commands are the following:

The following parameters can be configured for a pipe:

bw bandwidth | device

Bandwidth, measured in [ K| M]{ bit/s| Byte/s}.

A value of 0 (default) means unlimited bandwidth. The unit must immediately follow the number, as in

ipfw pipe 1 config bw 300Kbit/s

If a device name is specified instead of a numeric value, as in

ipfw pipe 1 config bw tun0

then the transmit clock is supplied by the specified device. At the moment only the tun(4) device supports this functionality, for use in conjunction with ppp(8).

delay ms-delay

Propagation delay, measured in milliseconds. The value is rounded to the next multiple of the clock tick (typically 10ms, but it is a good practice to run kernels with “options HZ=1000” to reduce the granularity to 1ms or less). The default value is 0, meaning no delay.

burst size

If the data to be sent exceeds the pipe's bandwidth limit (and the pipe was previously idle), up to size bytes of data are allowed to bypass the dummynet scheduler, and will be sent as fast as the physical link allows. Any additional data will be transmitted at the rate specified by the pipe bandwidth. The burst size depends on how long the pipe has been idle; the effective burst size is calculated as follows: MAX( size , bw * pipe_idle_time).

profile filename

A file specifying the additional overhead incurred in the transmission of a packet on the link.

Some link types introduce extra delays in the transmission of a packet, e.g., because of MAC level framing, contention on the use of the channel, MAC level retransmissions and so on. From our point of view, the channel is effectively unavailable for this extra time, which is constant or variable depending on the link type. Additionally, packets may be dropped after this time (e.g., on a wireless link after too many retransmissions). We can model the additional delay with an empirical curve that represents its distribution.

      cumulative probability 
      1.0 ^ 
          | 
      L   +-- loss-level          x 
          |                 ****** 
          |                * 
          |           ***** 
          |          * 
          |        ** 
          |       * 
          +-------*-------------------> 
                      delay

The empirical curve may have both vertical and horizontal lines. Vertical lines represent constant delay for a range of probabilities. Horizontal lines correspond to a discontinuity in the delay distribution: the pipe will use the largest delay for a given probability.

The file format is the following, with whitespace acting as a separator and '#' indicating the beginning a comment:

name identifier

optional name (listed by "ipfw pipe show") to identify the delay distribution;

bw value

the bandwidth used for the pipe. If not specified here, it must be present explicitly as a configuration parameter for the pipe;

loss-level L

the probability above which packets are lost. (0.0 <= L <= 1.0, default 1.0 i.e., no loss);

samples N

the number of samples used in the internal representation of the curve (2..1024; default 100);

delay prob | prob delay

One of these two lines is mandatory and defines the format of the following lines with data points.

XXX YYY

2 or more lines representing points in the curve, with either delay or probability first, according to the chosen format. The unit for delay is milliseconds. Data points do not need to be sorted. Also, the number of actual lines can be different from the value of the "samples" parameter: ipfw utility will sort and interpolate the curve as needed.

Example of a profile file:

name    bla_bla_bla 
samples 100 
loss-level    0.86 
prob    delay 
0       200 # minimum overhead is 200ms 
0.5     200 
0.5     300 
0.8     1000 
0.9     1300 
1       1300 
#configuration file end

The following parameters can be configured for a queue:

pipe pipe_nr

Connects a queue to the specified pipe. Multiple queues (with the same or different weights) can be connected to the same pipe, which specifies the aggregate rate for the set of queues.

weight weight

Specifies the weight to be used for flows matching this queue. The weight must be in the range 1..100, and defaults to 1.

The following case-insensitive parameters can be configured for a scheduler:

type {fifo | wf2q+ | rr | qfq}

specifies the scheduling algorithm to use.

fifo

is just a FIFO scheduler (which means that all packets are stored in the same queue as they arrive to the scheduler). FIFO has O(1) per-packet time complexity, with very low constants (estimate 60-80ns on a 2GHz desktop machine) but gives no service guarantees.

wf2q+

implements the WF2Q+ algorithm, which is a Weighted Fair Queueing algorithm which permits flows to share bandwidth according to their weights. Note that weights are not priorities; even a flow with a minuscule weight will never starve. WF2Q+ has O(log N) per-packet processing cost, where N is the number of flows, and is the default algorithm used by previous versions dummynet's queues.

rr

implements the Deficit Round Robin algorithm, which has O(1) processing costs (roughly, 100-150ns per packet) and permits bandwidth allocation according to weights, but with poor service guarantees.

qfq

implements the QFQ algorithm, which is a very fast variant of WF2Q+, with similar service guarantees and O(1) processing costs (roughly, 200-250ns per packet).

In addition to the type, all parameters allowed for a pipe can also be specified for a scheduler.

Finally, the following parameters can be configured for both pipes and queues:

buckets hash-table-size

Specifies the size of the hash table used for storing the various queues. Default value is 64 controlled by the sysctl(8) variable net.inet.ip.dummynet.hash_size, allowed range is 16 to 65536.

mask mask-specifier

Packets sent to a given pipe or queue by an ipfw rule can be further classified into multiple flows, each of which is then sent to a different dynamic pipe or queue. A flow identifier is constructed by masking the IP addresses, ports and protocol types as specified with the mask options in the configuration of the pipe or queue. For each different flow identifier, a new pipe or queue is created with the same parameters as the original object, and matching packets are sent to it.

Thus, when dynamic pipes are used, each flow will get the same bandwidth as defined by the pipe, whereas when dynamic queues are used, each flow will share the parent's pipe bandwidth evenly with other flows generated by the same queue (note that other queues with different weights might be connected to the same pipe).

 

Available mask specifiers are a combination of one or more of the following:

dst-ip mask, dst-ip6 mask, src-ip mask, src-ip6 mask, dst-port mask, src-port mask, flow-id mask, proto mask or all,

where the latter means all bits in all fields are significant.

noerror

When a packet is dropped by a dummynet queue or pipe, the error is normally reported to the caller routine in the kernel, in the same way as it happens when a device queue fills up. Setting this option reports the packet as successfully delivered, which can be needed for some experimental setups where you want to simulate loss or congestion at a remote router.

plr packet-loss-rate

Packet loss rate. Argument packet-loss-rate is a floating-point number between 0 and 1, with 0 meaning no loss, 1 meaning 100% loss. The loss rate is internally represented on 31 bits.

queue { slots | size Kbytes}

Queue size, in slots or KBytes. Default value is 50 slots, which is the typical queue size for Ethernet devices. Note that for slow speed links you should keep the queue size short or your traffic might be affected by a significant queueing delay. E.g., 50 max-sized ethernet packets (1500 bytes) mean 600Kbit or 20s of queue on a 30Kbit/s pipe. Even worse effects can result if you get packets from an interface with a much larger MTU, e.g. the loopback interface with its 16KB packets. The sysctl(8) variables net.inet.ip.dummynet.pipe_byte_limit and net.inet.ip.dummynet.pipe_slot_limit control the maximum lengths that can be specified.

red | gred w_q/ min_th/ max_th/ max_p

Make use of the RED (Random Early Detection) queue management algorithm. w_q and max_p are floating point numbers between 0 and 1 (0 not included), while min_th and max_th are integer numbers specifying thresholds for queue management (thresholds are computed in bytes if the queue has been defined in bytes, in slots otherwise). The dummynet also supports the gentle RED variant (gred). Three sysctl(8) variables can be used to control the RED behaviour:

net.inet.ip.dummynet.red_lookup_depth

specifies the accuracy in computing the average queue when the link is idle (defaults to 256, must be greater than zero)

net.inet.ip.dummynet.red_avg_pkt_size

specifies the expected average packet size (defaults to 512, must be greater than zero)

net.inet.ip.dummynet.red_max_pkt_size

specifies the expected maximum packet size, only used when queue thresholds are in bytes (defaults to 1500, must be greater than zero).

When used with IPv6 data, dummynet currently has several limitations. Information necessary to route link-local packets to an interface is not available after processing by dummynet so those packets are dropped in the output path. Care should be taken to ensure that link-local packets are not passed to dummynet.

REDIRECT AND LSNAT SUPPORT IN IPFW

Redirect and LSNAT support follow closely the syntax used in natd(8). See Section EXAMPLES for some examples on how to do redirect and lsnat.

SCTP NAT SUPPORT

SCTP nat can be configured in a similar manner to TCP through the ipfw command line tool. The main difference is that sctp nat does not do port translation. Since the local and global side ports will be the same, there is no need to specify both. Ports are redirected as follows:

Most sctp nat configuration can be done in real-time through the sysctl(8) interface. All may be changed dynamically, though the hash_table size will only change for new nat instances. See SYSCTL VARIABLES for more info.

BASIC PACKET FILTERING

This command adds an entry which denies all tcp packets from cracker.evil.org to the telnet port of wolf.tambov.su from being forwarded by the host:

ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet

This one disallows any connection from the entire cracker's network to my host:

ipfw add deny ip from 123.45.67.0/24 to my.host.org

A first and efficient way to limit access (not using dynamic rules) is the use of the following rules:

ipfw add allow tcp from any to any established

ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup

ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup

...

ipfw add deny tcp from any to any

The first rule will be a quick match for normal TCP packets, but it will not match the initial SYN packet, which will be matched by the setup rules only for selected source/destination pairs. All other SYN packets will be rejected by the final deny rule.

If you administer one or more subnets, you can take advantage of the address sets and or-blocks and write extremely compact rulesets which selectively enable services to blocks of clients, as below:

goodguys="{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }"

badguys="10.1.2.0/24{8,38,60}"

ipfw add allow ip from ${goodguys} to any

ipfw add deny ip from ${badguys} to any

... normal policies ...

The verrevpath option could be used to do automated anti-spoofing by adding the following to the top of a ruleset:

ipfw add deny ip from any to any not verrevpath in

This rule drops all incoming packets that appear to be coming to the system on the wrong interface. For example, a packet with a source address belonging to a host on a protected internal network would be dropped if it tried to enter the system from an external interface.

The antispoof option could be used to do similar but more restricted anti-spoofing by adding the following to the top of a ruleset:

ipfw add deny ip from any to any not antispoof in

This rule drops all incoming packets that appear to be coming from another directly connected system but on the wrong interface. For example, a packet with a source address of 192.168.0.0/24, configured on fxp0, but coming in on fxp1 would be dropped.

The setdscp option could be used to (re)mark user traffic, by adding the following to the appropriate place in ruleset:

ipfw add setdscp be ip from any to any dscp af11,af21

DYNAMIC RULES

In order to protect a site from flood attacks involving fake TCP packets, it is safer to use dynamic rules:

ipfw add check-state

ipfw add deny tcp from any to any established

ipfw add allow tcp from my-net to any setup keep-state

This will let the firewall install dynamic rules only for those connection which start with a regular SYN packet coming from the inside of our network. Dynamic rules are checked when encountering the first occurrence of a check-state, keep-state or limit rule. A check-state rule should usually be placed near the beginning of the ruleset to minimize the amount of work scanning the ruleset. Your mileage may vary.

To limit the number of connections a user can open you can use the following type of rules:

ipfw add allow tcp from my-net/24 to any setup limit src-addr 10

ipfw add allow tcp from any to me setup limit src-addr 4

The former (assuming it runs on a gateway) will allow each host on a /24 network to open at most 10 TCP connections. The latter can be placed on a server to make sure that a single client does not use more than 4 simultaneous connections.

BEWARE: stateful rules can be subject to denial-of-service attacks by a SYN-flood which opens a huge number of dynamic rules. The effects of such attacks can be partially limited by acting on a set of sysctl(8) variables which control the operation of the firewall.

Here is a good usage of the list command to see accounting records and timestamp information:

ipfw -at list

or in short form without timestamps:

ipfw -a list

which is equivalent to:

ipfw show

Next rule diverts all incoming packets from 192.168.2.0/24 to divert port 5000:

ipfw divert 5000 ip from 192.168.2.0/24 to any in

TRAFFIC SHAPING

The following rules show some of the applications of ipfw and dummynet for simulations and the like.

This rule drops random incoming packets with a probability of 5%:

ipfw add prob 0.05 deny ip from any to any in

A similar effect can be achieved making use of dummynet pipes:

ipfw add pipe 10 ip from any to any

ipfw pipe 10 config plr 0.05

We can use pipes to artificially limit bandwidth, e.g. on a machine acting as a router, if we want to limit traffic from local clients on 192.168.2.0/24 we do:

ipfw add pipe 1 ip from 192.168.2.0/24 to any out

ipfw pipe 1 config bw 300Kbit/s queue 50KBytes

note that we use the out modifier so that the rule is not used twice. Remember in fact that ipfw rules are checked both on incoming and outgoing packets.

Should we want to simulate a bidirectional link with bandwidth limitations, the correct way is the following:

ipfw add pipe 1 ip from any to any out

ipfw add pipe 2 ip from any to any in

ipfw pipe 1 config bw 64Kbit/s queue 10Kbytes

ipfw pipe 2 config bw 64Kbit/s queue 10Kbytes

The above can be very useful, e.g. if you want to see how your fancy Web page will look for a residential user who is connected only through a slow link. You should not use only one pipe for both directions, unless you want to simulate a half-duplex medium (e.g. AppleTalk, Ethernet, IRDA). It is not necessary that both pipes have the same configuration, so we can also simulate asymmetric links.

Should we want to verify network performance with the RED queue management algorithm:

ipfw add pipe 1 ip from any to any

ipfw pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1

Another typical application of the traffic shaper is to introduce some delay in the communication. This can significantly affect applications which do a lot of Remote Procedure Calls, and where the round-trip-time of the connection often becomes a limiting factor much more than bandwidth:

ipfw add pipe 1 ip from any to any out

ipfw add pipe 2 ip from any to any in

ipfw pipe 1 config delay 250ms bw 1Mbit/s

ipfw pipe 2 config delay 250ms bw 1Mbit/s

Per-flow queueing can be useful for a variety of purposes. A very simple one is counting traffic:

ipfw add pipe 1 tcp from any to any

ipfw add pipe 1 udp from any to any

ipfw add pipe 1 ip from any to any

ipfw pipe 1 config mask all

The above set of rules will create queues (and collect statistics) for all traffic. Because the pipes have no limitations, the only effect is collecting statistics. Note that we need 3 rules, not just the last one, because when ipfw tries to match IP packets it will not consider ports, so we would not see connections on separate ports as different ones.

A more sophisticated example is limiting the outbound traffic on a net with per-host limits, rather than per-network limits:

ipfw add pipe 1 ip from 192.168.2.0/24 to any out

ipfw add pipe 2 ip from any to 192.168.2.0/24 in

ipfw pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes

ipfw pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes

LOOKUP TABLES

In the following example, we need to create several traffic bandwidth classes and we need different hosts/networks to fall into different classes. We create one pipe for each class and configure them accordingly. Then we create a single table and fill it with IP subnets and addresses. For each subnet/host we set the argument equal to the number of the pipe that it should use. Then we classify traffic using a single rule:

ipfw pipe 1 config bw 1000Kbyte/s

ipfw pipe 4 config bw 4000Kbyte/s

...

ipfw table 1 add 192.168.2.0/24 1

ipfw table 1 add 192.168.0.0/27 4

ipfw table 1 add 192.168.0.2 1

...

ipfw add pipe tablearg ip from table(1) to any

Using the fwd action, the table entries may include hostnames and IP addresses.

ipfw table 1 add 192.168.2.0/24 10.23.2.1

ipfw table 1 add 192.168.0.0/27 router1.dmz

...

ipfw add 100 fwd tablearg ip from any to table(1)

In the following example per-interface firewall is created:

ipfw table 10 add vlan20 12000

ipfw table 10 add vlan30 13000

ipfw table 20 add vlan20 22000

ipfw table 20 add vlan30 23000

..

ipfw add 100 ipfw skipto tablearg ip from any to any recv 'table(10)' in

ipfw add 200 ipfw skipto tablearg ip from any to any xmit 'table(10)' out

SETS OF RULES

To add a set of rules atomically, e.g. set 18:

ipfw set disable 18

ipfw add NN set 18 ... # repeat as needed

ipfw set enable 18

To delete a set of rules atomically the command is simply:

ipfw delete set 18

To test a ruleset and disable it and regain control if something goes wrong:

ipfw set disable 18

ipfw add NN set 18 ... # repeat as needed

ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18

Here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates thus restoring the previous situation.

To show rules of the specific set:

ipfw set 18 show

To show rules of the disabled set:

ipfw -S set 18 show

To clear a specific rule counters of the specific set:

ipfw set 18 zero NN

To delete a specific rule of the specific set:

ipfw set 18 delete NN

NAT, REDIRECT AND LSNAT

First redirect all the traffic to nat instance 123:

ipfw add nat 123 all from any to any

Then to configure nat instance 123 to alias all the outgoing traffic with ip 192.168.0.123, blocking all incoming connections, trying to keep same ports on both sides, clearing aliasing table on address change and keeping a log of traffic/link statistics:

ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports

Or to change address of instance 123, aliasing table will be cleared (see reset option):

ipfw nat 123 config ip 10.0.0.1

To see configuration of nat instance 123:

ipfw nat 123 show config

To show logs of all the instances in range 111-999:

ipfw nat 111-999 show

To see configurations of all instances:

ipfw nat show config

Or a redirect rule with mixed modes could looks like:

ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66

redirect_port tcp 192.168.0.1:80 500

redirect_proto udp 192.168.1.43 192.168.1.1

redirect_addr 192.168.0.10,192.168.0.11

10.0.0.100 # LSNAT

redirect_port tcp 192.168.0.1:80,192.168.0.10:22

500 # LSNAT

or it could be split in:

ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66

ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500

ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1

ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12

10.0.0.100

ipfw nat 5 config redirect_port tcp

192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500