man iptables by iptables-save v1.3.5

可不是原创文章,没办法,找不到合适的标题项!

Generated by iptables-save v1.3.5 on Fri Mar 31 18:52:44 2006
man iptables |col -b >iptables.txt

得到的说明方法,呵呵,可以以TXT的格式看,没有了乱码,觉得不错,相当不错,贴出来共享下,好好学学啦!
感谢我同事,小沈对此获取方法的提供!!
这年头,要看英文的说明啦!!


IPTABLES(8)          IPTABLES(8)

NAME
       iptables - administration tool for IPv4 packet filtering and NAT
SYNOPSIS
       iptables [-t table] -[AD] chain rule-specification [options]
       iptables [-t table] -I chain [rulenum] rule-specification [options]
       iptables [-t table] -R chain rulenum rule-specification [options]
       iptables [-t table] -D chain rulenum [options]
       iptables [-t table] -[LFZ] [chain] [options]
       iptables [-t table] -N chain
       iptables [-t table] -X [chain]
       iptables [-t table] -P chain target [options]
       iptables [-t table] -E old-chain-name new-chain-name
DESCRIPTION
       Iptables  is  used  to  set  up, maintain, and inspect the tables of IP
       packet filter rules in the Linux kernel.  Several different tables  may
       be  defined.   Each  table contains a number of built-in chains and may
       also contain user-defined chains.
       Each chain is a list of rules which can match a set of  packets.   Each
       rule specifies what to do with a packet that matches.  This is called a
       'target', which may be a jump to a user-defined chain in the  same  ta-
       ble.

TARGETS
       A  firewall rule specifies criteria for a packet, and a target. If the
       packet does not match, the next rule in the chain is the  examined;  if
       it does match, then the next rule is specified by the value of the tar-
       get, which can be the name of a user-defined chain or one of  the  spe-
       cial values ACCEPT, Drop, QUEUE, or RETURN.
       ACCEPT  means to let the packet through.  Drop means to drop the packet
       on the floor.  QUEUE means to pass the packet to userspace.   (How  the
       packet can be received by a userspace process differs by the particular
       queue handler.  2.4.x and  2.6.x  kernels  up  to  2.6.13  include  the
       ip_queue  queue handler.  Kernels 2.6.14 and later additionally include
       the nfnetlink_queue queue handler.  Packets with a target of QUEUE will
       be  sent  to queue number '0' in this case. Please also see the NFQUEUE
       target as described  later  in  this  man  page.)   RETURN  means  stop
       traversing  this  chain and  resume  at  the next rule in the previous
       (calling) chain.  If the end of a built-in chain is reached or  a  rule
       in a built-in chain with target RETURN is matched, the target specified
       by the chain policy determines the fate of the packet.
TABLES
       There are currently three independent tables (which tables are  present
       at  any time depends on the kernel configuration options and which mod-
       ules are present).
       -t, --table table
       This option specifies the packet matching table which  the  com-
       mand  should operate on. If the kernel is configured with auto-
       matic module loading, an attempt will be made to load the appro-
       priate module for that table if it is not already there.
       The tables are as follows:
       filter:
    This is  the default table (if no -t option is passed).  It
    contains the built-in chains INPUT (for packets destined  to
    local  sockets),  FORWARD  (for packets being routed through
    the box), and OUTPUT (for locally-generated packets).
       nat:
    This table is consulted when a packet  that  creates a  new
    connection  is encountered.  It consists of three built-ins:
    PREROUTING (for altering packets as soon as they  come  in),
    OUTPUT  (for altering locally-generated packets before rout-
    ing), and POSTROUTING (for  altering packets  as  they  are
    about to go out).
       mangle:
    This table is used for specialized packet alteration.  Until
    kernel 2.4.17 it had two built-in  chains:  PREROUTING  (for
    altering  incoming  packets  before routing) and OUTPUT (for
    altering locally-generated packets before  routing).  Since
    kernel  2.4.18,  three  other  built-in chains are also sup-
    ported: INPUT (for packets coming into the box itself), FOR-
    WARD (for  altering packets being routed through the box),
    and POSTROUTING (for altering packets as they are  about  to
    go out).
       raw:
    This table  is  used mainly for configuring exemptions from
    connection tracking in combination with the NOTRACK  target.
    It registers at the netfilter hooks with higher priority and
    is thus called before ip_conntrack, or any other IP  tables.
    It  provides the following built-in chains: PREROUTING (for
    packets arriving via any  network  interface)  OUTPUT  (for
    packets generated by local processes)
OPTIONS
       The options that are recognized by iptables can be divided into several
       different groups.
   COMMANDS
       These options specify the specific action to perform.  Only one of them
       can  be specified on the command line unless otherwise specified below.
       For all the long versions of the command and option names, you need  to
       use  only  enough  letters to ensure that iptables can differentiate it
       from all other options.
       -A, --append chain rule-specification
       Append one or more rules to the end of the selected chain.  When
       the  source  and/or  destination names resolve to more than one
       address, a rule will be added for each possible address combina-
       tion.
       -D, --delete chain rule-specification
       -D, --delete chain rulenum
       Delete one or more rules from the selected chain.  There are two
       versions of this command: the rule can be specified as a number
       in  the  chain  (starting  at 1 for the first rule) or a rule to
       match.
       -I, --insert chain [rulenum] rule-specification
       Insert one or more rules in the selected chain as the given rule
       number. So,  if  the  rule  number is 1, the rule or rules are
       inserted at the head of the chain.  This is also the default  if
       no rule number is specified.
       -R, --replace chain rulenum rule-specification
       Replace a rule in the selected chain.  If the source and/or des-
       tination names resolve to multiple addresses, the  command  will
       fail.  Rules are numbered starting at 1.
       -L, --list [chain]
       List  all rules in the selected chain.  If no chain is selected,
       all chains are listed.  As  every  other iptables  command,  it
       applies  to  the specified table (filter is the default), so NAT
       rules get listed by
        iptables -t nat -n -L
       Please note that it is often used with the -n option,  in  order
       to  avoid  long reverse DNS lookups.  It is legal to specify the
       -Z (zero) option as well, in which case  the  chain(s)  will  be
       atomically  listed  and zeroed.  The exact output is affected by
       the other arguments given. The exact rules are suppressed  until
       you use
        iptables -L -v
       -F, --flush [chain]
       Flush the selected chain (all the chains in the table if none is
       given).  This is equivalent to deleting all  the rules  one  by
       one.
       -Z, --zero [chain]
       Zero the packet and byte counters in all chains. It is legal to
       specify the -L, --list (list) option as well, to see  the  coun-
       ters immediately before they are cleared. (See above.)
       -N, --new-chain chain
       Create  a  new user-defined chain by the given name.  There must
       be no target of that name already.
       -X, --delete-chain [chain]
       Delete the optional user-defined chain specified.  There must be
       no  references  to  the chain.  If there are, you must delete or
       replace the referring rules before the  chain  can  be  deleted.
       The  chain  must be  empty,  i.e. not contain any rules.  If no
       argument is given, it will attempt to delete  every  non-builtin
       chain in the table.
       -P, --policy chain target
       Set  the policy for the chain to the given target.  See the sec-
       tion TARGETS for the legal targets.   Only  built-in  (non-user-
       defined) chains can  have  policies,  and neither built-in nor
       user-defined chains can be policy targets.
       -E, --rename-chain old-chain new-chain
       Rename the user specified chain to the user supplied name.  This
       is cosmetic, and has no effect on the structure of the table.
       -h     Help.   Give a (currently very brief) description of the command
       syntax.
   PARAMETERS
       The following parameters make up a rule specification (as used  in  the
       add, delete, insert, replace and append commands).
       -p, --protocol [!] protocol
       The  protocol of the rule or of the packet to check.  The speci-
       fied protocol can be one of tcp, udp, icmp, or all, or it can be
       a  numeric  value, representing one of these protocols or a dif-
       ferent  one.   A protocol  name from  /etc/protocols  is  also
       allowed.  A  "!" argument before the protocol inverts the test.
       The number zero is equivalent to all.  Protocol all  will  match
       with  all  protocols and is taken as default when this option is
       omitted.
       -s, --source [!] address[/mask]
       Source specification.  Address can be either a network  name,  a
       hostname (please  note  that specifying any name to be resolved
       with a remote query such as DNS is a really bad idea), a network
       IP address (with /mask), or a plain IP address.  The mask can be
       either a network mask or a plain number, specifying  the number
       of 1's at the left side of the network mask.  Thus, a mask of 24
       is equivalent to 255.255.255.0.  A  "!"  argument  before  the
       address specification inverts the sense of the address. The flag
       --src is an alias for this option.
       -d, --destination [!] address[/mask]
       Destination  specification.   See  the  description  of  the  -s
       (source) flag  for  a  detailed description of the syntax.  The
       flag --dst is an alias for this option.
       -j, --jump target
       This specifies the target of the rule; i.e., what to do  if  the
       packet  matches  it.   The  target  can  be a user-defined chain
       (other than the one this rule is in), one of the special builtin
       targets  which  decide the fate of the packet immediately, or an
       extension (see EXTENSIONS below).  If this option is omitted  in
       a rule (and -g is not used), then matching the rule will have no
       effect on the packet's fate, but the counters on the  rule  will
       be incremented.
       -g, --goto chain
       This  specifies  that  the  processing should continue in a user
       specified chain. Unlike the --jump option return will  not  con-
       tinue  processing  in  this  chain but instead in the chain that
       called us via --jump.
       -i, --in-interface [!] name
       Name of an interface via which a packet was received  (only  for
       packets  entering  the  INPUT,  FORWARD  and PREROUTING chains).
       When the "!" argument is used before  the  interface  name,  the
       sense  is  inverted.   If the interface name ends in a "+", then
       any interface which begins with this name will match.   If  this
       option is omitted, any interface name will match.
       -o, --out-interface [!] name
       Name of an interface via which a packet is going to be sent (for
       packets entering the FORWARD, OUTPUT  and  POSTROUTING  chains).
       When  the  "!"  argument is used before the interface name, the
       sense is inverted.  If the interface name ends in  a  "+",  then
       any  interface  which begins with this name will match.  If this
       option is omitted, any interface name will match.
       [!]  -f, --fragment
       This means that the rule only refers to second and further frag-
       ments  of fragmented packets.  Since there is no way to tell the
       source or destination ports of such a  packet  (or  ICMP type),
       such a packet will not match any rules which specify them.  When
       the "!" argument precedes the "-f"  flag,  the  rule  will  only
       match head fragments, or unfragmented packets.
       -c, --set-counters PKTS BYTES
       This enables the administrator to initialize the packet and byte
       counters of a rule (during Insert, APPEND, REPLACE  operations).
   OTHER OPTIONS
       The following additional options can be specified:
       -v, --verbose
       Verbose  output.  This  option  makes the list command show the
       interface name, the rule options (if any), and  the  TOS masks.
       The  packet  and byte counters are also listed, with the suffix
       'K', 'M' or 'G' for 1000, 1,000,000 and 1,000,000,000  multipli-
       ers  respectively  (but  see  the  -x flag to change this).  For
       appending, insertion,  deletion  and  replacement,  this causes
       detailed information on the rule or rules to be printed.
       -n, --numeric
       Numeric  output.  IP addresses and port numbers will be printed
       in numeric format.  By default, the program will try to  display
       them  as host names, network names, or services (whenever appli-
       cable).
       -x, --exact
       Expand numbers.  Display the exact value of the packet and  byte
       counters,  instead  of only the rounded number in K's (multiples
       of 1000) M's (multiples of 1000K) or G's (multiples  of  1000M).
       This option is only relevant for the -L command.
       --line-numbers
       When  listing  rules,  add line numbers to the beginning of each
       rule, corresponding to that rule's position in the chain.
       --modprobe=command
       When adding or inserting rules into a chain, use command to load
       any necessary modules (targets, match extensions, etc).
MATCH EXTENSIONS
       iptables can use extended packet matching modules.  These are loaded in
       two ways: implicitly, when -p or --protocol is specified, or  with  the
       -m  or  --match options,  followed  by the matching module name; after
       these, various extra command line options become  available,  depending
       on  the specific module.  You can specify multiple extended match mod-
       ules in one line, and you can use the -h or --help  options  after  the
       module has been specified to receive help specific to that module.
       The  following  are included in the base package, and most of these can
       be preceded by a !  to invert the sense of the match.
   account
       Account traffic for all hosts in defined network/netmask.
       Features:
       - long (one counter per protocol TCP/UDP/IMCP/Other) and short  statis-
       tics
       - one iptables rule for all hosts in network/netmask
       - loading/saving counters (by reading/writting to procfs entries)

       --aaddr network/netmask
       defines network/netmask for which make statistics.
       --aname name
       defines  name  of  list  where statistics will be kept. If no is
       specified DEFAULT will be used.
       --ashort
       table will colect only short  statistics (only  total  counters
       without splitting it into protocols.
       Example usage:
       account traffic for/to 192.168.0.0/24 network into table mynetwork:
       #   iptables   -A   FORWARD   -m   account  --aname  mynetwork  --aaddr
       192.168.0.0/24
       account traffic for/to WWW serwer for 192.168.0.0/24 network into table
       mywwwserver:
       # iptables -A INPUT -p tcp --dport 80
  -m account --aname mywwwserver --aaddr 192.168.0.0/24 --ashort
       # iptables -A OUTPUT -p tcp --sport 80
  -m account --aname mywwwserver --aaddr 192.168.0.0/24 --ashort
       read counters:
       # cat    /proc/net/ipt_account/mynetwork   #    cat
       /proc/net/ipt_account/mywwwserver
       set counters:
       #    echo    "ip    = 192.168.0.1 packets_src =     0"     >
       /proc/net/ipt_account/mywwserver
       Webpage:
   [url]http://www.barbara.eu.org/~quaker/ipt_account/[/url]
   addrtype
       This module matches packets based on their address type.  Address types
       are used within the kernel networking stack  and  categorize  addresses
       into various groups.  The exact definition of that group depends on the
       specific layer three protocol.
       The following address types are possible:
       UNSPEC an unspecified address (i.e. 0.0.0.0) UNICAST an unicast address
       LOCAL  a local address BROADCAST a broadcast address ANYCAST an
       anycast packet MULTICAST a multicast address BLACKHOLE a black-
       hole  address UNREACHABLE an unreachable address PROHIBIT a pro-
       hibited address THROW FIXME NAT FIXME XRESOLVE FIXME
       --src-type type
       Matches if the source address is of given type
       --dst-type type
       Matches if the destination address is of given type
   ah
       This module matches the SPIs in Authentication header of IPsec packets.
       --ahspi [!] spi[:spi]
   childlevel
       This  is  an  experimental module.  It matches on whether the packet is
       part of a master connection or one of its children  (or grandchildren,
       etc).   For  instance,  most packets are level 0.  FTP data transfer is
       level 1.
       --childlevel [!] level
   comment
       Allows you to add comments (up to 256 characters) to any rule.
       --comment comment
       Example:
       iptables -A INPUT -s 192.168.0.0/16 -m comment --comment "A pri-
       vatized IP block"
   condition
       This matches if a specific /proc filename is '0' or '1'.
       --condition [!] filename
       Match  on  boolean value stored in /proc/net/ipt_condition/file-
       name file
   connbytes
       Match by how many bytes or packets a connection (or  one  of  the  two
       flows  constituting the connection) have tranferred so far, or by aver-
       age bytes per packet.
       The counters are 64bit and are thus not expected to overflow ;)
       The primary use is to detect long-lived downloads and mark them to  be
       scheduled using a lower priority band in traffic control.
       The  transfered bytes  per  connection can  also  be  viewed  through
       /proc/net/ip_conntrack and accessed via ctnetlink
       [!] --connbytes from:[to]
       match packets  from  a  connection  whose  packets/bytes/average
       packet size is more than FROM and less than TO bytes/packets. if
       TO is omitted only FROM check is done.  "!"  is  used  to  match
       packets not falling in the range.
       --connbytes-dir [original|reply|both]
       which packets to consider
       --connbytes-mode [packets|bytes|avgpkt]
       whether  to  check the amount of packets, number of bytes trans-
       ferred or the average size (in bytes) of all packets received so
       far.  Note  that when "both" is used together with "avgpkt", and
       data is going (mainly) only in one direction (for example HTTP),
       the  average  packet  size will be about half of the actual data
       packets.
       Example:
       iptables .. -m connbytes --connbytes  10000:100000  --connbytes-
       dir both --connbytes-mode bytes ...
   connlimit
       Allows  you  to restrict  the  number of parallel TCP connections to a
       server per client IP address (or address block).
       [!] --connlimit-above n
       match if the number of existing tcp connections is (not) above n
       --connlimit-mask bits
       group hosts using mask
       Examples:
       # allow 2 telnet connections per client host
       iptables -p tcp --syn --dport 23 -m connlimit --connlimit-above
       2 -j REJECT
       # you can also match the other way around:
       iptables -p tcp --syn --dport 23 -m  connlimit  !  --connlimit-
       above 2 -j ACCEPT
       #  limit  the nr of parallel http requests to 16 per class C sized net-
       work (24 bit netmask)
       iptables -p tcp --syn --dport 80 -m connlimit --connlimit-above
       16 --connlimit-mask 24 -j REJECT
   connmark
       This module matches the netfilter mark field associated with a  connec-
       tion (which can be set using the CONNMARK target below).
       --mark value[/mask]
       Matches  packets in connections with the given mark value (if a
       mask is specified, this is logically ANDed with the mark before
       the comparison).
   connrate
       This module matches the current transfer rate in a connection.
       --connrate [!] [from]:[to]
       Match  against the current connection transfer rate being within
       'from' and 'to' bytes per second. When the "!" argument is  used
       before the range, the sense of the match is inverted.
   conntrack
       This  module,  when combined with connection tracking, allows access to
       more connection tracking information than  the  "state" match.  (this
       module is present only if iptables was compiled under a kernel support-
       ing this feature)
       --ctstate state
       Where state is a comma separated list of the  connection states
       to  match.   Possible states are INVALID meaning that the packet
       is associated with no known connection, ESTABLISHED meaning that
       the  packet is associated with a connection which has seen pack-
       ets in both directions, NEW meaning that the packet has  started
       a  new  connection,  or  otherwise  associated with a connection
       which has not seen packets in both directions, and RELATED mean-
       ing that the packet is starting a new connection, but is associ-
       ated with an existing connection, such as an FTP data  transfer,
       or  an ICMP error.  SNAT A virtual state, matching if the origi-
       nal source address differs from the reply destination.   DNAT  A
       virtual state, matching if the original destination differs from
       the reply source.
       --ctproto proto
       Protocol to match (by number or name)
       --ctorigsrc [!] address[/mask]
       Match against original source address
       --ctorigdst [!] address[/mask]
       Match against original destination address
       --ctreplsrc [!] address[/mask]
       Match against reply source address
       --ctrepldst [!] address[/mask]
       Match against reply destination address
       --ctstatus [NONE|EXPECTED|SEEN_REPLY|ASSURED][,...]
       Match against internal conntrack states
       --ctexpire time[:time]
       Match remaining lifetime in seconds against given value or range
       of values (inclusive)
   dccp
       --source-port,--sport [!] port[:port]
       --destination-port,--dport [!] port[:port]
       --dccp-types [!] mask
       Match  when  the DCCP packet type is one of 'mask'. 'mask' is a
       comma-separated list of packet types.  Packet types are: REQUEST
       RESPONSE DATA  ACK  DATAACK  CLOSEREQ  CLOSE RESET SYNC SYNCACK
       INVALID.
       --dccp-option [!] number
       Match if DCP option set.
   dscp
       This module matches the 6 bit DSCP field within the TOS field in the IP
       header. DSCP has superseded TOS within the IETF.
       --dscp value
       Match against a numeric (decimal or hex) value [0-32].
       --dscp-class DiffServ Class
       Match  the  DiffServ class. This value may be any of the BE, EF,
       AFxx or CSx classes.   It  will  then  be  converted  into  it's
       according numeric value.
   dstlimit
       This  module  allows you to limit the packet per second (pps) rate on a
       per destination IP or per destination port base.   As  opposed  to  the
       'limit' match, every  destination  ip / destination port has it's own
       limit.
       THIS MODULE IS DEPRECATED AND HAS BEEN REPLACED BY ''hashlimit''
       --dstlimit avg
       Maximum average match rate (packets per second  unless  followed
       by /sec /minute /hour /day postfixes).
       --dstlimit-mode mode
       The limiting hashmode.  Is the specified limit per dstip, dstip-
       dstport tuple,  srcip-dstip  tuple,  or  per  srcipdstip-dstport
       tuple.
       --dstlimit-name name
       Name for /proc/net/ipt_dstlimit/* file entry
       [--dstlimit-burst burst]
       Number of packets to match in a burst.  Default: 5
       [--dstlimit-htable-size size]
       Number of buckets in the hashtable
       [--dstlimit-htable-max max]
       Maximum number of entries in the hashtable
       [--dstlimit-htable-gcinterval interval]
       Interval between  garbage  collection runs of the hashtable (in
       miliseconds).  Default is 1000 (1 second).
       [--dstlimit-htable-expire time
       After which time are idle entries  expired  from hashtable  (in
       miliseconds)?  Default is 10000 (10 seconds).
   ecn
       This  allows you to match the ECN bits of the IPv4 and TCP header.  ECN
       is the Explicit Congestion  Notification  mechanism  as  specified  in
       RFC3168
       --ecn-tcp-cwr
       This matches if the TCP ECN CWR (Congestion Window Received) bit
       is set.
       --ecn-tcp-ece
       This matches if the TCP ECN ECE (ECN Echo) bit is set.
       --ecn-ip-ect num
       This matches a particular IPv4 ECT (ECN-Capable Transport).  You
       have to specify a number between '0' and '3'.
   esp
       This module matches the SPIs in ESP header of IPsec packets.
       --espspi [!] spi[:spi]
   fuzzy
       This  module  matches  a  rate  limit based on a fuzzy logic controller
       [FLC]
       --lower-limit number
       Specifies the lower limit (in packets per second).
       --upper-limit number
       Specifies the upper limit (in packets per second).
   hashlimit
       This patch adds a new match called 'hashlimit'. The idea  is  to  have
       something  like 'limit',  but  either  per destination-ip or per (des-
       tip,destport) tuple.
       It gives you the ability to express
        '1000 packets per second for every host in 192.168.0.0/16'
        '100 packets per second for every service of 192.168.1.1'
       with a single iptables rule.
       --hashlimit rate
       A rate just like the limit match
       --hashlimit-burst num
       Burst value, just like limit match
       --hashlimit-mode destip | destip-destport
       Limit per IP or per port
       --hashlimit-name foo
       The name for the /proc/net/ipt_hashlimit/foo entry
       --hashlimit-htable-size num
       The number of buckets of the hash table
       --hashlimit-htable-max num
       Maximum entries in the hash
       --hashlimit-htable-expire num
       After how many miliseconds do hash entries expire
       --hashlimit-htable-gcinterval num
       How many miliseconds between garbage collection intervals
   helper
       This module matches packets related to a specific conntrack-helper.
       --helper string
       Matches packets related to the specified conntrack-helper.
       string can be "ftp" for packets  related to  a  ftp-session  on
       default  port.  For other ports append -portnr to the value, ie.
       "ftp-2121".
       Same rules apply for other conntrack-helpers.
   icmp
       This extension is loaded if '--protocol icmp' is  specified.   It  pro-
       vides the following option:
       --icmp-type [!] typename
       This  allows  specification  of  the  ICMP  type, which can be a
       numeric ICMP type, or one of the ICMP type names shown  by  the
       command
        iptables -p icmp -h
   iprange
       This matches on a given arbitrary range of IPv4 addresses
       [!]--src-range ip-ip
       Match source IP in the specified range.
       [!]--dst-range ip-ip
       Match destination IP in the specified range.
   ipv4options
       Match  on IPv4 header options like source routing, record route, times-
       tamp and router-alert.
       --ssrr To match packets with the flag strict source routing.
       --lsrr To match packets with the flag loose source routing.
       --no-srr
       To match packets with no flag for source routing.
       [!] --rr
       To match packets with the RR flag.
       [!] --ts
       To match packets with the TS flag.
       [!] --ra
       To match packets with the router-alert option.
       [!] --any-opt
       To match a packet with at least one IP option, or no  IP option
       at all if ! is chosen.
       Examples:
       $ iptables -A input -m ipv4options --rr -j Drop
       will drop packets with the record-route flag.
       $ iptables -A input -m ipv4options --ts -j Drop
       will drop packets with the timestamp flag.
   length
       This  module matches the length of a packet against a specific value or
       range of values.
       --length [!] length[:length]
   limit
       This module matches at a limited rate using a token bucket  filter.   A
       rule  using  this  extension  will  match  until  this limit is reached
       (unless the '!' flag is used).  It can be used in combination with  the
       LOG target to give limited logging, for example.
       --limit rate
       Maximum  average matching  rate: specified as a number, with an
       optional '/second', '/minute', '/hour', or  '/day'  suffix;  the
       default is 3/hour.
       --limit-burst number
       Maximum  initial number of  packets to match: this number gets
       recharged by one every time the limit  specified above  is  not
       reached, up to this number; the default is 5.
   mac
       --mac-source [!] address
       Match   source   MAC   address.  It   must   be  of  the  form
       XX:XX:XX:XX:XX:XX.  Note that this only makes sense for  packets
       coming from an Ethernet device and entering the PREROUTING, FOR-
       WARD or INPUT chains.
   mark
       This module matches the netfilter mark field associated with  a packet
       (which can be set using the MARK target below).
       --mark value[/mask]
       Matches packets with the given unsigned mark value (if a mask is
       specified, this is logically ANDed with the mask before the com-
       parison).
   mport
       This  module  matches  a  set of source or destination ports.  Up to 15
       ports can be specified. It can only be used in conjunction with -p tcp
       or -p udp.
       --source-ports port[,port[,port...]]
       Match  if  the  source port is one of the given ports.  The flag
       --sports is a convenient alias for this option.
       --destination-ports port[,port[,port...]]
       Match if the destination port is one of the  given  ports.   The
       flag --dports is a convenient alias for this option.
       --ports port[,port[,port...]]
       Match  if the both the source and destination ports are equal to
       each other and to one of the given ports.
   multiport
       This module matches a set of source or destination  ports.   Up to  15
       ports  can be specified.  A port range (port:port) counts as two ports.
       It can only be used in conjunction with -p tcp or -p udp.
       --source-ports [!] port[,port[,port:port...]]
       Match if the source port is one of the given  ports.   The  flag
       --sports is a convenient alias for this option.
       --destination-ports [!] port[,port[,port:port...]]
       Match  if  the  destination port is one of the given ports.  The
       flag --dports is a convenient alias for this option.
       --ports [!] port[,port[,port:port...]]
       Match if either the source or destination ports are equal to one
       of the given ports.
   nth
       This module matches every 'n'th packet
       --every value
       Match every 'value' packet
       [--counter num]
       Use internal counter number 'num'.  Default is '0'.
       [--start num]
       Initialize the counter at the number 'num' insetad of '0'.  Most
       between '0' and 'value'-1.
       [--packet num]
       Match on 'num' packet.  Most be between '0' and 'value'-1.
   osf
       The idea of passive OS fingerprint matching exists  for quite  a  long
       time,  but  was created as extension fo OpenBSD pf only some weeks ago.
       original  idea  was  lurked  in some  OpenBSD  mailing list   (thanks
       grange@open...) and  than  adopted for Linux netfilter in form of this
       code.
       original   fingerprint table was   created by   Michal   Zalewski
       <[email]lcamtuf@coredump.cx[/email]>.
       This module compares some data(WS, MSS, options and it's order, ttl, df
       and others) from first SYN packet (actually from packets with  SYN  bit
       set) with dynamically loaded OS fingerprints.
       --log 1/0
       If  present,  OSF  will log determined genres even if they don't
       match desired one.    0 - log all determined entries, 1  -  only
       first one.
       In syslog you find something like this:
       ipt_osf: Windows  [2000:SP3:Windows  XP Pro  SP1,  2000  SP3]:
       11.22.33.55:4024 -> 11.22.33.44:139
       ipt_osf:       Unknown:        16384:106:1:48:020405B401010402
       44.33.22.11:1239 -> 11.22.33.44:80
       --smart
       if  present, OSF will use some smartness to determine remote OS.
       OSF will use initial TTL only if source of connection is in  our
       local network.
       --netlink
       If  present,  OSF  will  log  all  events  also  through netlink
       NETLINK_NFLOG groupt 1.
       --genre [!] string
       Match a OS genre by passive fingerprinting
       Example:
       #iptables -I INPUT -j ACCEPT -p tcp  -m  osf  --genre  Linux  --log  1
       --smart
       NOTE: -p tcp is obviously required as it is a TCP match.
       Fingerprints  can  be  loaded  and  read through /proc/sys/net/ipv4/osf
       file.  One can flush all fingerprints with following command:
       echo -en FLUSH > /proc/sys/net/ipv4/osf
       Only one fingerprint per open/write/close.
       Fingerprints  can  be   downloaded   from    [url]http://www.openbsd.org/cgi-[/url]
       bin/cvsweb/src/etc/pf.os
   owner
       This  module  attempts  to  match various characteristics of the packet
       creator, for locally-generated packets. It is only valid in the OUTPUT
       chain,  and  even  this some packets (such as ICMP ping responses) may
       have no owner, and hence never match.
       --uid-owner userid
       Matches if the packet was created by a process  with  the  given
       effective user id.
       --gid-owner groupid
       Matches  if  the packet was created by a process with the given
       effective group id.
       --pid-owner processid
       Matches if the packet was created by a process  with  the  given
       process id.
       --sid-owner sessionid
       Matches if the packet was created by a process in the given ses-
       sion group.
       --cmd-owner name
       Matches if the packet was created by a process  with  the  given
       command name.  (this option is present only if iptables was com-
       piled under a kernel supporting this feature)
       NOTE: pid, sid and command matching are broken on SMP
   physdev
       This module matches  on the  bridge  port  input  and  output  devices
       enslaved  to  a bridge device. This module is a part of the infrastruc-
       ture that enables a transparent bridging IP firewall and is only useful
       for kernel versions above version 2.5.44.
       --physdev-in [!] name
       Name  of a bridge port via which a packet is received (only for
       packets entering the INPUT, FORWARD and PREROUTING  chains).  If
       the  interface  name  ends  in  a  "+", then any interface which
       begins with this name will match. If the packet didn't arrive
       through  a  bridge  device, this packet won't match this option,
       unless '!' is used.
       --physdev-out [!] name
       Name of a bridge port via which a packet is  going  to  be  sent
       (for  packets  entering  the  FORWARD,  OUTPUT  and  POSTROUTING
       chains). If the interface name ends in a "+", then  any inter-
       face  which  begins  with this name will match. Note that in the
       nat and mangle OUTPUT chains one cannot match on the bridge out-
       put  port,  however  one can in the filter OUTPUT chain. If the
       packet won't leave by a bridge device or it is yet unknown  what
       the  output  device  will  be,  then the packet won't match this
       option, unless
       [!] --physdev-is-in
       Matches if the packet has entered through a bridge interface.
       [!] --physdev-is-out
       Matches if the packet will leave through a bridge interface.
       [!] --physdev-is-bridged
       Matches if the packet is being  bridged  and  therefore  is  not
       being  routed.  This is only useful in the FORWARD and POSTROUT-
       ING chains.
   pkttype
       This module matches the link-layer packet type.
       --pkt-type [unicast|broadcast|multicast]
   policy
       This modules matches the policy used by IPsec for handling a packet.
       --dir in|out
       Used to select whether to match the policy used  for  decapsula-
       tion  or the policy that will be used for encapsulation.  in is
       valid in the PREROUTING, INPUT and FORWARD chains, out is  valid
       in the POSTROUTING, OUTPUT and FORWARD chains.
       --pol none|ipsec
       Matches if the packet is subject to IPsec processing.
       --strict
       Selects  whether to match the exact policy or match if any rule
       of the policy matches the given policy.
       --reqid id
       Matches the reqid of the policy rule. The reqid can be specified
       with setkey(8) using unique:id as level.
       --spi spi
       Matches the SPI of the SA.
       --proto ah|esp|ipcomp
       Matches the encapsulation protocol.
       --mode tunnel|transport
       Matches the encapsulation mode.
       --tunnel-src addr[/mask]
       Matches  the source end-point address of a tunnel mode SA.  Only
       valid with --mode tunnel.
       --tunnel-dst addr[/mask]
       Matches the destination end-point address of a tunnel  mode  SA.
       Only valid with --mode tunnel.
       --next Start  the next element in the policy specification. Can only be
       used with --strict
   psd
       Attempt to detect TCP and UDP port scans. This match was  derived  from
       Solar Designer's scanlogd.
       --psd-weight-threshold threshold
       Total weight of the latest TCP/UDP packets with different desti-
       nation ports coming from the same host to  be  treated  as  port
       scan sequence.
       --psd-delay-threshold delay
       Delay  (in  hundredths of second) for the packets with different
       destination ports coming from the same host  to  be  treated  as
       possible port scan subsequence.
       --psd-lo-ports-weight weight
       Weight  of the packet with privileged (<=1024) destination port.
       --psd-hi-ports-weight weight
       Weight of the packet with non-priviliged destination port.
   quota
       Implements network quotas by decrementing  a  byte  counter  with  each
       packet.
       --quota bytes
       The quota in bytes.
       KNOWN BUGS: this does not work on SMP systems.
   random
       This module randomly matches a certain percentage of all packets.
       --average percent
       Matches  the given percentage.  If omitted, a probability of 50%
       is set.
   realm
       This matches the routing realm. Routing realms are  used  in  complex
       routing setups involving dynamic routing protocols like BGP.
       --realm [!]value[/mask]
       Matches a given realm number (and optionally mask).
   recent
       Allows  you to dynamically create a list of IP addresses and then match
       against that list in a few different ways.
       For example, you can create a 'badguy' list out of people attempting to
       connect to  port 139 on your firewall and then Drop all future packets
       from them without considering them.
       --name name
       Specify the list to use for the commands. If no  name  is  given
       then 'DEFAULT' will be used.
       [!] --set
       This  will  add the source address of the packet to the list. If
       the source address is already in the list, this will update  the
       existing entry. This will always return success (or failure if
       '!' is passed in).
       [!] --rcheck
       Check if the source address of the packet is  currently  in  the
       list.
       [!] --update
       Like  --rcheck,  except it will update the "last seen" timestamp
       if it matches.
       [!] --remove
       Check if the source address of the packet is  currently  in  the
       list  and  if  so that address will be removed from the list and
       the rule will return true. If the address is not found, false is
       returned.
       [!] --seconds seconds
       This  option must be used in conjunction with one of --rcheck or
       --update. When used, this will narrow the match to  only happen
       when  the  address  is  in the list and was seen within the last
       given number of seconds.
       [!] --hitcount hits
       This option must be used in conjunction with one of --rcheck  or
       --update.  When  used, this will narrow the match to only happen
       when the address is in the list and packets  had been  received
       greater  than  or  equal to the given value. This option may be
       used along with --seconds  to  create  an  even  narrower  match
       requiring a certain number of hits within a specific time frame.
       --rttl This option must be used in conjunction with one of --rcheck  or
       --update.  When  used, this will narrow the match to only happen
       when the address is in the list  and  the  TTL  of  the  current
       packet matches that of the packet which hit the --set rule. This
       may be useful if you have  problems  with  people  faking  their
       source  address in order to DoS you via this module by disallow-
       ing others access to your site by sending bogus packets to  you.
       Examples:
       # iptables -A FORWARD -m recent --name badguy --rcheck --seconds
       60 -j Drop
       # iptables -A FORWARD -p tcp  -i eth0  --dport  139  -m recent
       --name badguy --set -j Drop
       Official  website  ( [url]http://snowman.net/projects/ipt_recent/[/url] )  also  has
       some examples of usage.
       /proc/net/ipt_recent/* are the current lists of addresses and  informa-
       tion about each entry of each list.
       Each  file in /proc/net/ipt_recent/ can be read from to see the current
       list or written two using the following commands to modify the list:
       echo xx.xx.xx.xx > /proc/net/ipt_recent/DEFAULT
       to Add to the DEFAULT list
       echo -xx.xx.xx.xx > /proc/net/ipt_recent/DEFAULT
       to Remove from the DEFAULT list
       echo clear > /proc/net/ipt_recent/DEFAULT
       to empty the DEFAULT list.
       The module itself accepts parameters, defaults shown:
       ip_list_tot=100
       Number of addresses remembered per table
       ip_pkt_list_tot=20
       Number of packets per address remembered
       ip_list_hash_size=0
       Hash table size. 0 means to calculate it based  on  ip_list_tot,
       default: 512
       ip_list_perms=0644
       Permissions for /proc/net/ipt_recent/* files
       debug=0
       Set to 1 to get lots of debugging info
   sctp
       --source-port,--sport [!] port[:port]
       --destination-port,--dport [!] port[:port]
       --chunk-types [!] all|any|only chunktype[:flags] [...]
       The  flag  letter  in  upper  case indicates that the flag is to
       match if set, in the lower case indicates to match if unset.
       Chunk types: DATA INIT  INIT_ACK SACK  HEARTBEAT  HEARTBEAT_ACK
       ABORT   SHUTDOWN  SHUTDOWN_ACK ERROR  COOKIE_ECHO  COOKIE_ACK
       ECN_ECNE ECN_CWR SHUTDOWN_COMPLETE ASCONF ASCONF_ACK
       chunk type     available flags
       DATA      U B E u b e
       ABORT      T t
       SHUTDOWN_COMPLETE     T t
       (lowercase means flag should be "off", uppercase means "on")
       Examples:
       iptables -A INPUT -p sctp --dport 80 -j Drop
       iptables -A INPUT -p sctp --chunk-types any DATA,INIT -j Drop
       iptables -A INPUT -p sctp --chunk-types any DATA:Be -j ACCEPT
   set
       This modules macthes IP sets which can be defined by ipset(8).
       --set setname flag[,flag...]
       where flags are src and/or dst and there can be no more than six
       of them. Hence the command
        iptables -A FORWARD -m set --set test src,dst
       will match packets, for which (depending on the type of the set)
       the source address or port number of the packet can be found  in
       the specified set. If there is a binding belonging to the mached
       set element or there is a default binding  for  the  given  set,
       then  the  rule  will  match  the  packet  only  if additionally
       (depending on the type of the set) the  destination  address  or
       port  number  of the packet can be found in the set according to
       the binding.
   state
       This module, when combined with connection tracking, allows  access  to
       the connection tracking state for this packet.
       --state state
       Where  state  is a comma separated list of the connection states
       to match.  Possible states are INVALID meaning that  the packet
       could  not  be identified for some reason which includes running
       out of memory and ICMP errors  which  don't  correspond  to  any
       known connection, ESTABLISHED meaning that the packet is associ-
       ated with a connection which has seen  packets  in  both direc-
       tions, NEW meaning that the packet has started a new connection,
       or otherwise associated with a connection  which has  not  seen
       packets  in both directions, and RELATED meaning that the packet
       is starting a new connection, but is associated with an existing
       connection, such as an FTP data transfer, or an ICMP error.
   string
       This  modules  matches  a  given  string by using some pattern matching
       strategy. It requires a linux kernel >= 2.6.14.
       --algo  bm|kmp
       Select the pattern matching strategy. (bm = Boyer-Moore, kmp  =
       Knuth-Pratt-Morris)
       --from offset
       Set the offset from which it starts looking for any matching. If
       not passed, default is 0.
       --to offset
       Set the offset from which it starts looking for any matching. If
       not passed, default is the packet size.
       --string pattern
       Matches  the  given  pattern.   --hex-string pattern Matches the
       given pattern in hex notation.
   tcp
       These extensions are loaded if '--protocol tcp' is specified.  It  pro-
       vides the following options:
       --source-port [!] port[:port]
       Source  port  or port range specification. This can either be a
       service name or a port number. An inclusive range  can  also  be
       specified,  using  the  format  port:port.  If the first port is
       omitted, "0" is assumed; if the  last  is  omitted,  "65535"  is
       assumed. If the second port greater then the first they will be
       swapped. The flag  --sport  is  a  convenient  alias  for  this
       option.
       --destination-port [!] port[:port]
       Destination  port or port range specification.  The flag --dport
       is a convenient alias for this option.
       --tcp-flags [!] mask comp
       Match when the TCP flags are as specified.  The  first  argument
       is  the  flags which we should examine, written as a comma-sepa-
       rated list, and the second argument is a comma-separated list of
       flags which must be set. Flags are: SYN ACK FIN RST URG PSH ALL
       NONE.  Hence the command
        iptables -A FORWARD -p tcp --tcp-flags SYN,ACK,FIN,RST SYN
       will only match packets with the SYN flag set, and the ACK,  FIN
       and RST flags unset.
       [!] --syn
       Only  match TCP packets with the SYN bit set and the ACK,RST and
       FIN bits cleared.  Such packets are used to request TCP  connec-
       tion initiation; for example, blocking such packets coming in an
       interface will prevent incoming TCP  connections,  but  outgoing
       TCP  connections will be unaffected.  It is equivalent to --tcp-
       flags  SYN,RST,ACK,FIN  SYN.   If  the  "!"  flag  precedes  the
       "--syn", the sense of the option is inverted.
       --tcp-option [!] number
       Match if TCP option set.
       --mss value[:value]
       Match  TCP  SYN  or SYN/ACK packets with the specified MSS value
       (or range), which control the maximum packet size for that  con-
       nection.
   tcpmss
       This  matches  the  TCP MSS  (maximum  segment size) field of the TCP
       header. You can only use this on TCP SYN or SYN/ACK packets, since the
       MSS  is only negotiated during the TCP handshake at connection startup
       time.
       [!] --mss value[:value]"
       Match a given TCP MSS value or range.
   time
       This matches if the packet arrival time/date is within a  given range.
       All options are facultative.
--timestart value
       Match  only  if  it is after 'value' (Inclusive, format: HH:MM ;
       default 00:00).
       --timestop  value
       Match only if it is before 'value' (Inclusive, format:  HH:MM  ;
       default 23:59).
       --days listofdays
       Match  only  if  today  is  one  of  the given  days.  (format:
       Mon,Tue,Wed,Thu,Fri,Sat,Sun ; default everyday)
       --datestart date
       Match  only  if  it  is after 'date'  (Inclusive,   format:
       YYYY[:MM[:DD[:hh[:mm[:ss]]]]]  ; h,m,s start from 0 ; default to
       1970)
       --datestop date
       Match  only  if  it  is  before 'date'  (Inclusive,   format:
       YYYY[:MM[:DD[:hh[:mm[:ss]]]]]  ; h,m,s start from 0 ; default to
       2037)
   tos
       This module matches the 8 bits of Type  of  Service  field  in  the  IP
       header (ie. including the precedence bits).
       --tos tos
       The argument is either a standard name, (use
        iptables -m tos -h
       to see the list), or a numeric value to match.
   ttl
       This module matches the time to live field in the IP header.
       --ttl-eq ttl
       Matches the given TTL value.
       --ttl-gt ttl
       Matches if TTL is greater than the given TTL value.
       --ttl-lt ttl
       Matches if TTL is less than the given TTL value.
   u32
       U32  allows  you  to extract quantities of up to 4 bytes from a packet,
       AND them with specified masks, shift them by specified amounts and test
       whether the results are in any of a set of specified ranges.  The spec-
       ification of what to extract is general enough  to  skip  over  headers
       with lengths stored in the packet, as in IP or TCP header lengths.
       Details and examples are in the kernel module source.
   udp
       These  extensions are loaded if '--protocol udp' is specified.  It pro-
       vides the following options:
       --source-port [!] port[:port]
       Source port or port range specification. See the description of
       the --source-port option of the TCP extension for details.
       --destination-port [!] port[:port]
       Destination  port or port range specification.  See the descrip-
       tion of the --destination-port option of the TCP extension  for
       details.
   unclean
       This  module takes no options, but attempts to match packets which seem
       malformed or unusual.  This is regarded as experimental.
TARGET EXTENSIONS
       iptables can use extended target modules: the following are included in
       the standard distribution.
   BALANCE
       This  allows  you to DNAT connections in a round-robin way over a given
       range of destination addresses.
       --to-destination ipaddr-ipaddr
       Address range to round-robin over.
   CLASSIFY
       This module allows you to set the skb->priority value (and  thus  clas-
       sify the packet into a specific CBQ class).
       --set-class MAJOR:MINOR
       Set the major and minor class value.
   CLUSTERIP
       This  module  allows  you  to  configure a simple cluster of nodes that
       share a certain IP and MAC address without an explicit load balancer in
       front  of  them.   Connections  are  statically distributed between the
       nodes in this cluster.
       --new  Create a new ClusterIP.  You always have to  set  this  on  the
       first rule for a given ClusterIP.
       --hashmode mode
       Specify  the hashing mode.  Has to be one of sourceip, sourceip-
       sourceport, sourceip-sourceport-destport
       --clustermac mac
       Specify the ClusterIP MAC address.  Has to be a link-layer  mul-
       ticast address
       --total-nodes num
       Number of total nodes within this cluster.
       --local-node num
       Local node number within this cluster.
       --hash-init rnd
       Specify the random seed used for hash initialization.
   CONNMARK
       This module sets the netfilter mark value associated with a connection
       --set-mark mark[/mask]
       Set connection mark. If a mask is specified then only those bits
       set in the mask is modified.
       --save-mark [--mask mask]
       Copy the netfilter packet mark value to the connection mark.  If
       a mask is specified then only those bits are copied.
       --restore-mark [--mask mask]
       Copy the connection mark value to the packet. If a mask is spec-
       ified then only those bits are copied. This is only valid in the
       mangle table.
   DNAT
       This  target is only valid in the nat table, in the PREROUTING and OUT-
       PUT chains, and user-defined chains which are only  called  from  those
       chains. It specifies that the destination address of the packet should
       be modified (and all future packets in this  connection will  also  be
       mangled),  and rules should cease being examined.  It takes one type of
       option:
       --to-destination ipaddr[-ipaddr][:port-port]
       which can specify a single new destination IP address, an inclu-
       sive  range of IP addresses, and optionally, a port range (which
       is only valid if the rule also specifies -p tcp or -p udp).   If
       no port range is specified, then the destination port will never
       be modified.
       In Kernels up to 2.6.10 you  can add  several  --to-destination
       options. For those kernels, if you specify more than one desti-
       nation address, either via an address range  or  multiple  --to-
       destination  options, a simple round-robin (one after another in
       cycle) load  balancing  takes  place  between  these  addresses.
       Later  Kernels  (>= 2.6.11-rc1) don't have the ability to NAT to
       multiple ranges anymore.

   DSCP
       This target allows to alter the value of the DSCP bits within  the  TOS
       header  of  the IPv4 packet.  As this manipulates a packet, it can only
       be used in the mangle table.
       --set-dscp value
       Set the DSCP field to a numerical value (can be decimal or hex)
       --set-dscp-class class
       Set the DSCP field to a DiffServ class.
   ECN
       This target allows to selectively work around known ECN blackholes.  It
       can only be used in the mangle table.
       --ecn-tcp-remove
       Remove all ECN bits from the TCP header. Of course, it can only
       be used in conjunction with -p tcp.
   IPMARK
       Allows you to mark a received packet basing on its IP address. This can
       replace many  mangle/mark  entries  with only one, if you use firewall
       based classifier.
       This target is to be used inside the mangle table, in  the  PREROUTING,
       POSTROUTING or FORWARD hooks.
       --addr src/dst
       Use source or destination IP address.
       --and-mask mask
       Perform bitwise 'and' on the IP address and this mask.
       --or-mask mask
       Perform bitwise 'or' on the IP address and this mask.
       The  order  of  IP  address  bytes  is reversed to meet "human order of
       bytes": 192.168.0.1 is 0xc0a80001. At first the 'and' operation is per-
       formed, then 'or'.
       Examples:
       We create a queue for each user, the queue number is adequate to the IP
       address of the user, e.g.: all packets going  to/from  192.168.5.2  are
       directed to 1:0502 queue, 192.168.5.12 -> 1:050c etc.
       We have one classifier rule:
       tc filter add dev eth3 parent 1:0 protocol ip fw
       Earlier we had many rules just like below:
       iptables -t mangle -A POSTROUTING -o eth3 -d 192.168.5.2 -j MARK
       --set-mark 0x10502
       iptables -t mangle -A POSTROUTING -o eth3 -d 192.168.5.3 -j MARK
       --set-mark 0x10503
       Using  IPMARK target we can replace all the mangle/mark rules with only
       one:
       iptables -t mangle -A POSTROUTING -o eth3 -j  IPMARK  --addr=dst
       --and-mask=0xffff --or-mask=0x10000
       On  the routers with hundreds of users there should be significant load
       decrease (e.g. twice).
   IPV4OPTSSTRIP
       Strip all the IP options from a packet.
       The target doesn't take any option, and therefore is extremly  easy  to
       use :
       # iptables -t mangle -A PREROUTING -j IPV4OPTSSTRIP
   LOG
       Turn  on  kernel  logging of matching packets.  When this option is set
       for a rule, the Linux kernel will print some information on all match-
       ing  packets  (like most IP header fields) via the kernel log (where it
       can be read with dmesg or syslogd(8)).  This is a "non-terminating tar-
       get",  i.e.  rule traversal continues at the next rule. So if you want
       to LOG the packets you refuse, use two separate rules  with  the  same
       matching criteria, first using target LOG then Drop (or REJECT).
       --log-level level
       Level of logging (numeric or see syslog.conf(5)).
       --log-prefix prefix
       Prefix  log messages with the specified prefix; up to 29 letters
       long, and useful for distinguishing messages in the logs.
       --log-tcp-sequence
       Log TCP sequence numbers. This is a security risk if the log  is
       readable by users.
       --log-tcp-options
       Log options from the TCP packet header.
       --log-ip-options
       Log options from the IP packet header.
       --log-uid
       Log the userid of the process which generated the packet.
   MARK
       This  is  used  to  set the  netfilter mark value associated with the
       packet. It is only valid in the mangle table.  It can for  example  be
       used in conjunction with iproute2.
       --set-mark mark
   MASQUERADE
       This  target  is only valid in the nat table, in the POSTROUTING chain.
       It should only be used with dynamically assigned  IP  (dialup)  connec-
       tions: if you have a static IP address, you should use the SNAT target.
       Masquerading is equivalent to specifying a mapping to the IP address of
       the  interface  the  packet  is going out, but also has the effect that
       connections are forgotten when the interface goes down.  This  is  the
       correct behavior  when the  next  dialup is unlikely to have the same
       interface address (and hence any established connections are lost  any-
       way).  It takes one option:
       --to-ports port[-port]
       This  specifies  a  range of source ports to use, overriding the
       default SNAT source port-selection heuristics (see above).  This
       is only valid if the rule also specifies -p tcp or -p udp.
   MIRROR
       This  is  an experimental demonstration target which inverts the source
       and destination fields in the IP header and retransmits the packet.  It
       is  only  valid in the INPUT, FORWARD and PREROUTING chains, and user-
       defined chains which are only called from those chains. Note that  the
       outgoing  packets  are NOT seen by any packet filtering chains, connec-
       tion tracking or NAT, to avoid loops and other problems.
   NETMAP
       This target allows you to statically map a whole network  of  addresses
       onto  another  network of addresses.  It can only be used from rules in
       the nat table.
       --to address[/mask]
       Network address to map to.  The resulting address will  be  con-
       structed in  the  following way: All 'one' bits in the mask are
       filled in from the new 'address'.  All bits that are zero in the
       mask are filled in from the original address.
   NFQUEUE
       This  target  is an extension of the QUEUE target. As opposed to QUEUE,
       it allows you to put a packet into any specific queue, identified  by
       its 16-bit queue number.
       --queue-num value
       This  specifies the QUEUE number to use. Valud queue numbers are
       0 to 65535. The default value is 0.
       It can only be used with Kernel versions  2.6.14  or  later,  since  it
       requires
       the nfnetlink_queue kernel support.
   NOTRACK
       This target disables connection tracking for all packets matching  that
       rule.
       It can only be used in the
       raw table.
   REDIRECT
       This  target is only valid in the nat table, in the PREROUTING and OUT-
       PUT chains, and user-defined chains which are only  called  from  those
       chains.  It redirects the packet to the machine itself by changing the
       destination IP  to  the primary  address  of  the  incoming  interface
       (locally-generated  packets  are  mapped to the 127.0.0.1 address).  It
       takes one option:
       --to-ports port[-port]
       This specifies a destination port or  range  of  ports  to  use:
       without  this,  the  destination port is never altered.  This is
       only valid if the rule also specifies -p tcp or -p udp.
   REJECT
       This is used to send back an error packet in response  to  the  matched
       packet: otherwise it is equivalent to Drop so it is a terminating TAR-
       GET, ending rule traversal.  This target is only valid  in  the INPUT,
       FORWARD and  OUTPUT  chains,  and  user-defined  chains which are only
       called from those chains.  The following option controls the nature  of
       the error packet returned:
       --reject-with type
       The type given can be
        icmp-net-unreachable
        icmp-host-unreachable
        icmp-port-unreachable
        icmp-proto-unreachable
        icmp-net-prohibited
        icmp-host-prohibited or
        icmp-admin-prohibited (*)
       which  return  the appropriate ICMP error message (port-unreach-
       able is the default).  The option tcp-reset can be used on rules
       which  only match the TCP protocol: this causes a TCP RST packet
       to be sent back. This  is  mainly  useful  for  blocking  ident
       (113/tcp)  probes  which frequently  occur when sending mail to
       broken mail hosts (which won't accept your mail otherwise).
       (*) Using icmp-admin-prohibited with kernels that  do  not  support  it
       will result in a plain Drop instead of REJECT
   ROUTE
       This is used to explicitly override the core  network  stack's  routing
       decision.  mangle table.
       --oif ifname
       Route the packet through 'ifname' network interface
       --iif ifname
       Change the packet's incoming interface to 'ifname'
       --gw IP_address
       Route the packet via this gateway
       --continue
       Behave like a non-terminating target and continue traversing the
       rules.  Not valid in combination with '--iif' or '--tee'
       --tee  Make a copy of the packet, and route that copy to the given des-
       tination.  For the original, uncopied packet, behave like a non-
       terminating target and continue traversing the rules.  Not valid
       in combination with '--iif' or '--continue'
   SAME
       Similar to SNAT/DNAT depending on chain: it takes a range of addresses
       ('--to 1.2.3.4-1.2.3.7') and gives a client the same  source-/destina-
       tion-address for each connection.
       --to <ipaddr>-<ipaddr>
       Addresses  to map source to. May be specified more than once for
       multiple ranges.
       --nodst
       Don't use the destination-ip in the calculations when  selecting
       the new source-ip
   SET
       This  modules  adds  and/or  deletes  entries from IP sets which can be
       defined by ipset(8).
       --add-set setname flag[,flag...]
       add the address(es)/port(s) of the packet to the sets
       --del-set setname flag[,flag...]
       delete the address(es)/port(s) of  the  packet  from  the  sets,
       where flags are src and/or dst and there can be no more than six
       of them.
       The bindings to follow must previously be defined in order to use
       multilevel adding/deleting by the SET target.
   SNAT
       This target is only valid in the nat table, in the  POSTROUTING chain.
       It  specifies  that the source address of the packet should be modified
       (and all future packets in this connection will also be mangled),  and
       rules should cease being examined.  It takes one type of option:
       --to-source  ipaddr[-ipaddr][:port-port]
       which  can  specify a single new source IP address, an inclusive
       range of IP addresses, and optionally, a port  range  (which  is
       only  valid if the rule also specifies -p tcp or -p udp).  If no
       port range is specified, then source ports  below  512  will  be
       mapped  to  other  ports below  512: those between 512 and 1023
       inclusive will be mapped to ports below 1024,  and  other  ports
       will  be mapped to 1024 or above. Where possible, no port alter-
       ation will occur.
       In Kernels  up  to  2.6.10,  you can  add  several  --to-source
       options. For those kernels, if you specify more than one source
       address, either via an address  range  or  multiple  --to-source
       options, a simple round-robin (one after another in cycle) takes
       place between these addresses.  Later  Kernels  (>=  2.6.11-rc1)
       don't have the ability to NAT to multiple ranges anymore.
   TARPIT
       Captures  and holds incoming TCP connections using no local per-connec-
       tion resources. Connections are accepted, but immediately  switched  to
       the persist state (0 byte window), in which the remote side stops send-
       ing data and asks to continue every 60-240 seconds.  Attempts to  close
       the  connection are  ignored,  forcing the remote side to time out the
       connection in 12-24 minutes.
       This  offers  similar   functionality   to   LaBrea   < [url]http://www.hack-[/url]
       busters.net/LaBrea/> but doesn't require dedicated hardware or IPs. Any
       TCP port that you would normally Drop or REJECT can  instead  become  a
       tarpit.
       To tarpit connections to TCP port 80 destined for the current machine:
       iptables -A INPUT -p tcp -m tcp --dport 80 -j TARPIT
       To significantly slow down Code Red/Nimda-style scans of unused address
       space, forward unused ip addresses to a Linux  box  not  acting  as  a
       router (e.g. "ip route 10.0.0.0 255.0.0.0 ip.of.linux.box" on a Cisco),
       enable IP forwarding on the Linux box, and add:
       iptables -A FORWARD -p tcp -j TARPIT
       iptables -A FORWARD -j Drop
       NOTE:  If you use the conntrack module while you are using TARPIT,  you
       should  also use the NOTRACK target, or the kernel will unneces-
       sarily allocate resources  for  each  TARPITted  connection.  To
       TARPIT incoming connections to the standard IRC port while using
       conntrack, you could:
       iptables -t raw -A PREROUTING -p tcp --dport 6667 -j NOTRACK
       iptables -A INPUT -p tcp --dport 6667 -j TARPIT
   TCPMSS
       This target allows to alter the MSS value of TCP SYN packets,  to  con-
       trol  the maximum size for that connection (usually limiting it to your
       outgoing interface's MTU minus 40).  Of course, it can only be used  in
       conjunction with -p tcp.  It is only valid in the mangle table.
       This  target  is  used to overcome criminally braindead ISPs or servers
       which block ICMP Fragmentation Needed packets.  The  symptoms  of  this
       problem are that everything works fine from your Linux firewall/router,
       but machines behind it can never exchange large packets:
1) Web browsers connect, then hang with no data received.
2) Small mail works fine, but large emails hang.
3) ssh works fine, but scp hangs after initial handshaking.
       Workaround: activate this option and add a rule to your firewall  con-
       figuration like:
iptables -t mangle -A FORWARD -p tcp --tcp-flags SYN,RST SYN \
      -j TCPMSS --clamp-mss-to-pmtu
       --set-mss value
       Explicitly set MSS option to specified value.
       --clamp-mss-to-pmtu
       Automatically clamp MSS value to (path_MTU - 40).
       These options are mutually exclusive.
   TOS
       This  is  used to set the 8-bit Type of Service field in the IP header.
       It is only valid in the mangle table.
       --set-tos tos
       You can use a numeric TOS values, or use
        iptables -j TOS -h
       to see the list of valid TOS names.
   TRACE
       This target has no options.  It just turns on packet  tracing  for  all
       packets that match this rule.
   TTL
       This is used to modify the IPv4 TTL header field.  The TTL field deter-
       mines how many hops (routers) a packet can traverse until it's time  to
       live is exceeded.
       Setting or  incrementing the TTL field can potentially be very danger-
       ous,
       so it should be avoided at any cost.
       Don't  ever set or increment the value on packets that leave your local
       network!
       mangle table.
       --ttl-set value
       Set the TTL value to 'value'.
       --ttl-dec value
       Decrement the TTL value 'value' times.
       --ttl-inc value
       Increment the TTL value 'value' times.
   ULOG
       This  target provides userspace logging of matching packets.  When this
       target is set for a rule, the Linux kernel will multicast  this packet
       through a netlink socket. One or more userspace processes may then sub-
       scribe to various multicast groups and receive the packets.  Like  LOG,
       this  is  a  "non-terminating target", i.e. rule traversal continues at
       the next rule.
       --ulog-nlgroup nlgroup
       This specifies the netlink group (1-32) to which the  packet  is
       sent.  Default value is 1.
       --ulog-prefix prefix
       Prefix  log messages with the specified prefix; up to 32 charac-
       ters long, and useful for distinguishing messages in the logs.
       --ulog-cprange size
       Number of bytes to be copied to userspace.  A value of 0 always
       copies the entire packet, regardless of its size.  Default is 0.
       --ulog-qthreshold size
       Number of packet to queue inside kernel. Setting this value to,
       e.g.  10 accumulates ten packets inside the kernel and transmits
       them as one netlink multipart message to userspace.  Default  is
       1 (for backwards compatibility).
   XOR
       Encrypt TCP and UDP traffic using a simple XOR encryption
       --key string
       Set key to "string"
       --block-size
       Set block size
DIAGNOSTICS
       Various error messages are printed to standard error.  The exit code is
       0 for correct functioning.  Errors which appear to be caused by invalid
       or  abused  command  line parameters cause an exit code of 2, and other
       errors cause an exit code of 1.
BUGS
       Bugs?  What's this? ;-) Well,  you  might  want  to  have  a  look  at
        [url]http://bugzilla.netfilter.org/[/url]
COMPATIBILITY WITH IPCHAINS
       This  iptables  is very similar to ipchains by Rusty Russell.  The main
       difference is that the chains INPUT and OUTPUT are only traversed  for
       packets coming into the local host and originating from the local host
       respectively.  Hence every packet only passes through one of the  three
       chains  (except loopback traffic, which involves both INPUT and OUTPUT
       chains); previously a forwarded packet would pass through all three.
       The other main difference is that -i refers to the input interface;  -o
       refers  to  the output interface,  and both are available for packets
       entering the FORWARD chain.
       iptables is a pure packet filter when using the default 'filter' table,
       with optional extension modules.  This should simplify much of the pre-
       vious confusion over the combination of IP masquerading and packet fil-
       tering  seen  previously.  So the following options are handled differ-
       ently:
-j MASQ
-M -S
-M -L
       There are several other changes in iptables.
SEE ALSO
       iptables-save(8), iptables-restore(8), ip6tables(8), ip6tables-save(8),
       ip6tables-restore(8), libipq(3).
       The packet-filtering-HOWTO details iptables usage for packet filtering,
       the NAT-HOWTO details NAT, the netfilter-extensions-HOWTO  details  the
       extensions  that  are not in the standard distribution, and the netfil-
       ter-hacking-HOWTO details the netfilter internals.
       See  [url]http://www.netfilter.org/.[/url]
AUTHORS
       Rusty Russell originally wrote iptables,  in  early  consultation  with
       Michael Neuling.
       Marc  Boucher  made  Rusty  abandon  ipnatctl by lobbying for a generic
       packet selection framework in iptables, then wrote  the mangle table,
       the owner match, the mark stuff, and ran around doing cool stuff every-
       where.
       James Morris wrote the TOS target, and tos match.
       Jozsef Kadlecsik wrote the REJECT target.
       Harald Welte wrote the ULOG and NFQUEUE target, the  new  libiptc,  as
       well as the TTL, DSCP, ECN matches and targets.
       The  Netfilter  Core  Team  is: Marc Boucher, Martin Josefsson, Jozsef
       Kadlecsik, Patrick McHardy, James Morris, Harald Welte and  Rusty  Rus-
       sell.
       Man page originally written by Herve Eychenne <[email]rv@wallfire.org[/email]>.

     Mar 09, 2002      IPTABLES(8)

 本文转自 rickyfang 51CTO博客，原文链接：http://blog.51cto.com/rickyfang/127408，如需转载请自行联系原作者