# ############################################################################## # # File: IPTables::ChainMgr.pm # # Purpose: Perl interface to add and delete rules to an iptables chain. The # most common application of this module is to create a custom chain # and then add blocking rules to it. Rule additions are (mostly) # guaranteed to be unique. # # Author: Michael Rash (mbr@cipherdyne.org) # # Version: 0.9 # ############################################################################## # # $Id: ChainMgr.pm 990 2008-02-02 19:01:21Z mbr $ # package IPTables::ChainMgr; use 5.006; use POSIX ':sys_wait_h'; use Carp; use IPTables::Parse; use Net::IPv4Addr 'ipv4_network'; use strict; use warnings; use vars qw($VERSION); $VERSION = '0.9'; sub new() { my $class = shift; my %args = @_; my $self = { _iptables => $args{'iptables'} || '/sbin/iptables', _iptout => $args{'iptout'} || '/tmp/ipt.out', _ipterr => $args{'ipterr'} || '/tmp/ipt.err', _ipt_alarm => $args{'ipt_alarm'} || 30, _debug => $args{'debug'} || 0, _verbose => $args{'verbose'} || 0, _ipt_exec_style => $args{'ipt_exec_style'} || 'waitpid', _ipt_exec_sleep => $args{'ipt_exec_sleep'} || 0, _sigchld_handler => $args{'sigchld_handler'} || \&REAPER, }; croak "[*] $self->{'_iptables'} incorrect iptables path.\n" unless -e $self->{'_iptables'}; croak "[*] $self->{'_iptables'} not executable.\n" unless -x $self->{'_iptables'}; bless $self, $class; } sub chain_exists() { my $self = shift; my $table = shift || croak '[*] Must specify a table, e.g. "filter".'; my $chain = shift || croak '[*] Must specify a chain to create.'; my $iptables = $self->{'_iptables'}; ### see if the chain exists return $self->run_ipt_cmd("$iptables -t $table -v -n -L $chain"); } sub create_chain() { my $self = shift; my $table = shift || croak '[*] Must specify a table, e.g. "filter".'; my $chain = shift || croak '[*] Must specify a chain to create.'; my $iptables = $self->{'_iptables'}; ### see if the chain exists first my ($rv, $out_aref, $err_aref) = $self->chain_exists($table, $chain); ### the chain already exists return 1, $out_aref, $err_aref if $rv; ### create the chain return $self->run_ipt_cmd("$iptables -t $table -N $chain"); } sub flush_chain() { my $self = shift; my $table = shift || croak '[*] Must specify a table, e.g. "filter".'; my $chain = shift || croak '[*] Must specify a chain.'; my $iptables = $self->{'_iptables'}; ### flush the chain return $self->run_ipt_cmd("$iptables -t $table -F $chain"); } sub delete_chain() { my $self = shift; my $table = shift || croak '[*] Must specify a table, e.g. "filter".'; my $jump_from_chain = shift || croak '[*] Must specify a chain from which ', 'packets were jumped to this chain'; my $del_chain = shift || croak '[*] Must specify a chain to delete.'; my $iptables = $self->{'_iptables'}; ### see if the chain exists first my ($rv, $out_aref, $err_aref) = $self->chain_exists($table, $del_chain); ### return true if the chain doesn't exist (it is not an error condition) return 1, $out_aref, $err_aref unless $rv; ### flush the chain ($rv, $out_aref, $err_aref) = $self->flush_chain($table, $del_chain, $iptables); ### could not flush the chain return 0, $out_aref, $err_aref unless $rv; ### find and delete jump rules to this chain (we can't delete ### the chain until there are no references to it) my ($rulenum, $num_chain_rules) = $self->find_ip_rule('0.0.0.0/0', '0.0.0.0/0', $table, $jump_from_chain, $del_chain, {}); if ($rulenum) { $self->run_ipt_cmd( "$iptables -t $table -D $jump_from_chain $rulenum"); } ### note that we try to delete the chain now regardless ### of whether their were jump rules above (should probably ### parse for the "0 references" under the -nL output). return $self->run_ipt_cmd("$iptables -t $table -X $del_chain"); } sub append_ip_rule() { my $self = shift; my $src = shift || croak '[-] Must specify a src address/network.'; my $dst = shift || croak '[-] Must specify a dst address/network.'; my $table = shift || croak '[-] Must specify a table, e.g. "filter".'; my $chain = shift || croak '[-] Must specify a chain.'; my $target = shift || croak '[-] Must specify a iptables target, e.g. "DROP"'; ### optionally add port numbers and protocols, etc. my $extended_href = shift || {}; my $iptables = $self->{'_iptables'}; ### normalize src/dst if necessary; this is because iptables ### always reports the network address for subnets my $normalized_src = $self->normalize_net($src); my $normalized_dst = $self->normalize_net($dst); ### first check to see if this rule already exists my ($rule_position, $num_chain_rules) = $self->find_ip_rule($normalized_src, $normalized_dst, $table, $chain, $target, $extended_href); if ($rule_position) { my $msg = ''; if ($extended_href) { $msg = "Table: $table, chain: $chain, $normalized_src -> " . "$normalized_dst "; for my $key qw(protocol s_port d_port mac_source) { $msg .= "$key $extended_href->{$key} " if defined $extended_href->{$key}; } $msg .= 'rule already exists.'; } else { $msg = "Table: $table, chain: $chain, $normalized_src -> " . "$normalized_dst rule already exists."; } return 1, [$msg], []; } ### we need to add the rule my $ipt_cmd = ''; my $msg = ''; my $idx_err = ''; if ($extended_href) { $ipt_cmd = "$iptables -t $table -A $chain "; $ipt_cmd .= "-p $extended_href->{'protocol'} " if defined $extended_href->{'protocol'}; $ipt_cmd .= "-s $normalized_src "; $ipt_cmd .= "--sport $extended_href->{'s_port'} " if defined $extended_href->{'s_port'}; $ipt_cmd .= "-d $normalized_dst "; $ipt_cmd .= "--dport $extended_href->{'d_port'} " if defined $extended_href->{'d_port'}; $ipt_cmd .= "-m mac --mac-source $extended_href->{'mac_source'} " if defined $extended_href->{'mac_source'}; $ipt_cmd .= "-j $target"; $msg = "Table: $table, chain: $chain, added $normalized_src " . "-> $normalized_dst "; for my $key qw(protocol s_port d_port mac_source) { $msg .= "$key $extended_href->{$key} " if defined $extended_href->{$key}; } ### for NAT if (defined $extended_href->{'to_ip'} and defined $extended_href->{'to_port'}) { $ipt_cmd .= " --to $extended_href->{'to_ip'}:" . "$extended_href->{'to_port'}"; $msg .= "$extended_href->{'to_ip'}:$extended_href->{'to_port'}"; } $msg =~ s/\s*$//; } else { $ipt_cmd = "$iptables -t $table -A $chain " . "-s $normalized_src -d $normalized_dst -j $target"; $msg = "Table: $table, chain: $chain, added $normalized_src " . "-> $normalized_dst"; } my ($rv, $out_aref, $err_aref) = $self->run_ipt_cmd($ipt_cmd); if ($rv) { push @$out_aref, $msg if $msg; } push @$err_aref, $idx_err if $idx_err; return $rv, $out_aref, $err_aref; } sub add_ip_rule() { my $self = shift; my $src = shift || croak '[-] Must specify a src address/network.'; my $dst = shift || croak '[-] Must specify a dst address/network.'; my $rulenum = shift || croak '[-] Must specify an insert rule number.'; my $table = shift || croak '[-] Must specify a table, e.g. "filter".'; my $chain = shift || croak '[-] Must specify a chain.'; my $target = shift || croak '[-] Must specify a iptables target, e.g. "DROP"'; ### optionally add port numbers and protocols, etc. my $extended_href = shift || {}; my $iptables = $self->{'_iptables'}; ### normalize src/dst if necessary; this is because iptables ### always reports the network address for subnets my $normalized_src = $self->normalize_net($src); my $normalized_dst = $self->normalize_net($dst); ### first check to see if this rule already exists my ($rule_position, $num_chain_rules) = $self->find_ip_rule($normalized_src, $normalized_dst, $table, $chain, $target, $extended_href); if ($rule_position) { my $msg = ''; if ($extended_href) { $msg = "Table: $table, chain: $chain, $normalized_src -> " . "$normalized_dst "; for my $key qw(protocol s_port d_port mac_source) { $msg .= "$key $extended_href->{$key} " if defined $extended_href->{$key}; } $msg .= 'rule already exists.'; } else { $msg = "Table: $table, chain: $chain, $normalized_src -> " . "$normalized_dst rule already exists."; } return 1, [$msg], []; } ### we need to add the rule my $ipt_cmd = ''; my $msg = ''; my $idx_err = ''; ### check to see if the insertion index ($rulenum) is too big $rulenum = 1 if $rulenum <= 0; if ($rulenum > $num_chain_rules+1) { $idx_err = "Rule position $rulenum is past end of $chain " . "chain ($num_chain_rules rules), compensating." if $num_chain_rules > 0; $rulenum = $num_chain_rules + 1; } $rulenum = 1 if $rulenum == 0; if ($extended_href) { $ipt_cmd = "$iptables -t $table -I $chain $rulenum "; $ipt_cmd .= "-p $extended_href->{'protocol'} " if defined $extended_href->{'protocol'}; $ipt_cmd .= "-s $normalized_src "; $ipt_cmd .= "--sport $extended_href->{'s_port'} " if defined $extended_href->{'s_port'}; $ipt_cmd .= "-d $normalized_dst "; $ipt_cmd .= "--dport $extended_href->{'d_port'} " if defined $extended_href->{'d_port'}; $ipt_cmd .= "-m mac --mac-source $extended_href->{'mac_source'} " if defined $extended_href->{'mac_source'}; $ipt_cmd .= "-j $target"; $msg = "Table: $table, chain: $chain, added $normalized_src " . "-> $normalized_dst "; for my $key qw(protocol s_port d_port mac_source) { $msg .= "$key $extended_href->{$key} " if defined $extended_href->{$key}; } ### for NAT if (defined $extended_href->{'to_ip'} and defined $extended_href->{'to_port'}) { $ipt_cmd .= " --to $extended_href->{'to_ip'}:" . "$extended_href->{'to_port'}"; $msg .= "$extended_href->{'to_ip'}:$extended_href->{'to_port'}"; } $msg =~ s/\s*$//; } else { $ipt_cmd = "$iptables -t $table -I $chain $rulenum " . "-s $normalized_src -d $normalized_dst -j $target"; $msg = "Table: $table, chain: $chain, added $normalized_src " . "-> $normalized_dst"; } my ($rv, $out_aref, $err_aref) = $self->run_ipt_cmd($ipt_cmd); if ($rv) { push @$out_aref, $msg if $msg; } push @$err_aref, $idx_err if $idx_err; return $rv, $out_aref, $err_aref; } sub delete_ip_rule() { my $self = shift; my $src = shift || croak '[-] Must specify a src address/network.'; my $dst = shift || croak '[-] Must specify a dst address/network.'; my $table = shift || croak '[-] Must specify a table, e.g. "filter".'; my $chain = shift || croak '[-] Must specify a chain.'; my $target = shift || croak '[-] Must specify a iptables target, e.g. "DROP"'; ### optionally add port numbers and protocols, etc. my $extended_href = shift || {}; my $iptables = $self->{'_iptables'}; ### normalize src/dst if necessary; this is because iptables ### always reports network address for subnets my $normalized_src = $self->normalize_net($src); my $normalized_dst = $self->normalize_net($dst); ### first check to see if this rule already exists my ($rulenum, $num_chain_rules) = $self->find_ip_rule($normalized_src, $normalized_dst, $table, $chain, $target, $extended_href); if ($rulenum) { ### we need to delete the rule return $self->run_ipt_cmd("$iptables -t $table -D $chain $rulenum"); } my $extended_msg = ''; if ($extended_href) { for my $key qw(protocol s_port d_port mac_source) { $extended_msg .= "$key: $extended_href->{$key} " if defined $extended_href->{$key}; } ### for NAT if (defined $extended_href->{'to_ip'} and defined $extended_href->{'to_port'}) { $extended_msg .= "$extended_href->{'to_ip'}:" . "$extended_href->{'to_port'}"; } } $extended_msg =~ s/\s*$//; return 0, [], ["Table: $table, chain: $chain, rule $normalized_src " . "-> $normalized_dst $extended_msg does not exist."]; } sub find_ip_rule() { my $self = shift; my $debug = $self->{'_debug'}; my $verbose = $self->{'_verbose'}; my $src = shift || croak '[*] Must specify source address.'; my $dst = shift || croak '[*] Must specify destination address.'; my $table = shift || croak '[*] Must specify iptables table.'; my $chain = shift || croak '[*] Must specify iptables chain.'; my $target = shift || croak '[*] Must specify iptables target (this may be a chain).'; ### optionally add port numbers and protocols, etc. my $extended_href = shift || {}; my $iptables = $self->{'_iptables'}; my $ipt_parse = new IPTables::Parse( 'iptables' => $self->{'_iptables'}, 'iptout' => $self->{'_iptout'}, 'ipterr' => $self->{'_ipterr'}, 'debug' => $self->{'_debug'}, 'verbose' => $self->{'_verbose'}, 'ipt_alarm' => $self->{'_ipt_alarm'}, 'ipt_exec_style' => $self->{'_ipt_exec_style'}, 'ipt_exec_sleep' => $self->{'_ipt_exec_sleep'}, 'sigchld_handler' => $self->{'_sigchld_handler'}, ) or croak "[*] Could not acquire IPTables::Parse object"; my $fh = *STDERR; $fh = *STDOUT if $verbose; if ($debug or $verbose) { print $fh localtime() . " [+] IPTables::Parse::VERSION ", "$IPTables::Parse::VERSION\n" } my $chain_aref = $ipt_parse->chain_rules($table, $chain); my $rulenum = 1; for my $rule_href (@$chain_aref) { if ($rule_href->{'target'} eq $target and $rule_href->{'src'} eq $src and $rule_href->{'dst'} eq $dst) { if ($extended_href) { my $found = 1; for my $key qw( protocol s_port d_port to_ip to_port ) { if (defined $extended_href->{$key}) { unless ($extended_href->{$key} eq $rule_href->{$key}) { $found = 0 } } } return $rulenum, $#$chain_aref+1 if $found; } else { if ($rule_href->{'protocol'} eq 'all') { if ($target eq 'LOG' or $target eq 'ULOG') { ### built-in LOG and ULOG target rules always ### have extended information return $rulenum, $#$chain_aref+1; } elsif (not $rule_href->{'extended'}) { ### don't want any additional criteria (such as ### port numbers) in the rule. Note that we are ### also not checking interfaces return $rulenum, $#$chain_aref+1; } } } } $rulenum++; } return 0, $#$chain_aref+1; } sub normalize_net() { my $self = shift; my $net = shift || croak '[*] Must specify net.'; ### regex to match an IP address my $ip_re = '(?:\d{1,3}\.){3}\d{1,3}'; my $normalized_net = ''; if ($net =~ m|($ip_re)/($ip_re)|) { my ($net_addr, $cidr) = ipv4_network($1, $2); $normalized_net = "$net_addr/$cidr"; } elsif ($net =~ m|($ip_re)/(\d+)|) { my ($net_addr, $cidr) = ipv4_network($1, $2); $normalized_net = "$net_addr/$cidr"; } else { ### it is a hostname or an individual IP $normalized_net = $net; } return $normalized_net; } sub add_jump_rule() { my $self = shift; my $table = shift || croak '[-] Must specify a table, e.g. "filter".'; my $from_chain = shift || croak '[-] Must specify chain to jump from.'; my $rulenum = shift || croak '[-] Must specify jump rule chain position'; my $to_chain = shift || croak '[-] Must specify chain to jump to.'; my $iptables = $self->{'_iptables'}; my $idx_err = ''; if ($from_chain eq $to_chain) { return 0, ["Identical from_chain and to_chain ($from_chain) " . "not allowed."], []; } ### first check to see if the jump rule already exists my ($rule_position, $num_chain_rules) = $self->find_ip_rule('0.0.0.0/0', '0.0.0.0/0', $table, $from_chain, $to_chain, {}); ### check to see if the insertion index ($rulenum) is too big $rulenum = 1 if $rulenum <= 0; if ($rulenum > $num_chain_rules+1) { $idx_err = "Rule position $rulenum is past end of $from_chain " . "chain ($num_chain_rules rules), compensating." if $num_chain_rules > 0; $rulenum = $num_chain_rules + 1; } $rulenum = 1 if $rulenum == 0; if ($rule_position) { ### the rule already exists return 1, ["Table: $table, chain: $to_chain, jump rule already exists."], []; } ### we need to add the rule my ($rv, $out_aref, $err_aref) = $self->run_ipt_cmd( "$iptables -t $table -I $from_chain $rulenum -j $to_chain"); push @$err_aref, $idx_err if $idx_err; return $rv, $out_aref, $err_aref; } sub REAPER { my $stiff; while(($stiff = waitpid(-1,WNOHANG))>0){ # do something with $stiff if you want } local $SIG{'CHLD'} = \&REAPER; return; } sub run_ipt_cmd() { my $self = shift; my $cmd = shift || croak '[*] Must specify an iptables command to run.'; my $iptables = $self->{'_iptables'}; my $iptout = $self->{'_iptout'}; my $ipterr = $self->{'_ipterr'}; my $debug = $self->{'_debug'}; my $ipt_alarm = $self->{'_ipt_alarm'}; my $verbose = $self->{'_verbose'}; my $ipt_exec_style = $self->{'_ipt_exec_style'}; my $ipt_exec_sleep = $self->{'_ipt_exec_sleep'}; my $sigchld_handler = $self->{'_sigchld_handler'}; croak "[*] $cmd does not look like an iptables command." unless $cmd =~ m|^\s*iptables| or $cmd =~ m|^\S+/iptables|; my $rv = 1; my @stdout = (); my @stderr = (); my $fh = *STDERR; $fh = *STDOUT if $verbose; if ($debug or $verbose) { print $fh localtime() . " [+] IPTables::ChainMgr::", "run_ipt_cmd(${ipt_exec_style}()) $cmd\n"; if ($ipt_exec_sleep > 0) { print $fh localtime() . " [+] IPTables::ChainMgr::", "run_ipt_cmd() sleep seconds: $ipt_exec_sleep\n"; } } if ($ipt_exec_sleep > 0) { if ($debug or $verbose) { print $fh localtime() . " [+] IPTables::ChainMgr: ", "sleeping for $ipt_exec_sleep seconds before ", "executing iptables command.\n"; } sleep $ipt_exec_sleep; } if ($ipt_exec_style eq 'system') { system qq{$cmd > $iptout 2> $ipterr}; } elsif ($ipt_exec_style eq 'popen') { open CMD, "$cmd 2> $ipterr |" or croak "[*] Could not execute $cmd: $!"; @stdout = ; close CMD; open F, "> $iptout" or croak "[*] Could not open $iptout: $!"; print F for @stdout; close F; } else { my $ipt_pid; if ($debug or $verbose) { print $fh localtime() . " [+] IPTables::ChainMgr: " . "Setting SIGCHLD handler to: " . $sigchld_handler . "\n"; } local $SIG{'CHLD'} = $sigchld_handler; if ($ipt_pid = fork()) { eval { ### iptables should never take longer than 30 seconds to execute, ### unless there is some absolutely enormous policy or the kernel ### is exceedingly busy local $SIG{'ALRM'} = sub {die "[*] iptables command timeout.\n"}; alarm $ipt_alarm; waitpid($ipt_pid, 0); alarm 0; }; if ($@) { kill 9, $ipt_pid unless kill 15, $ipt_pid; } } else { croak "[*] Could not fork iptables: $!" unless defined $ipt_pid; ### exec the iptables command and preserve stdout and stderr exec qq{$cmd > $iptout 2> $ipterr}; } } if (not @stdout and -e $iptout) { open F, "< $iptout" or croak "[*] Could not open $iptout"; @stdout = ; close F; } if (-e $ipterr) { open F, "< $ipterr" or croak "[*] Could not open $ipterr"; @stderr = ; close F; $rv = 0 if @stderr; } if ($debug or $verbose) { print $fh localtime() . " iptables command stdout:\n"; for my $line (@stdout) { if ($line =~ /\n$/) { print $fh $line; } else { print $fh $line, "\n"; } } print $fh localtime() . " iptables command stderr:\n"; for my $line (@stderr) { if ($line =~ /\n$/) { print $fh $line; } else { print $fh $line, "\n"; } } } return $rv, \@stdout, \@stderr; } 1; __END__ =head1 NAME IPTables::ChainMgr - Perl extension for manipulating iptables policies =head1 SYNOPSIS use IPTables::ChainMgr; my %opts = ( 'iptables' => '/sbin/iptables', 'iptout' => '/tmp/iptables.out', 'ipterr' => '/tmp/iptables.err', 'debug' => 0, 'verbose' => 0 ### advanced options 'ipt_alarm' => 5, ### max seconds to wait for iptables execution. 'ipt_exec_style' => 'waitpid', ### can be 'waitpid', ### 'system', or 'popen'. 'ipt_exec_sleep' => 1, ### add in time delay between execution of ### iptables commands (default is 0). ); my $ipt_obj = new IPTables::ChainMgr(%opts) or die "[*] Could not acquire IPTables::ChainMgr object"; my $rv = 0; my $out_ar = []; my $errs_ar = []; # check to see if the 'CUSTOM' chain exists in the filter table ($rv, $out_ar, $errs_ar) = $ipt_obj->chain_exists('filter', 'CUSTOM'); if ($rv) { print "CUSTOM chain exists.\n"; ### flush all rules from the chain $ipt_obj->flush_chain('filter', 'CUSTOM'); ### now delete the chain (along with any jump rule in the ### INPUT chain) $ipt_obj->delete_chain('filter', 'INPUT', 'CUSTOM'); } # create new iptables chain in the 'filter' table $ipt_obj->create_chain('filter', 'CUSTOM'); # add rule to jump packets from the INPUT chain into CUSTOM at the # 4th rule position $ipt_obj->add_jump_rule('filter', 'INPUT', 4, 'CUSTOM'); # find rule that allows all traffic from 10.1.2.3 to 192.168.1.2 ($rv, $rule_num) = $ipt_obj->find_ip_rule('10.1.2.3', '192.168.1.2', 'filter', 'INPUT', 'ACCEPT', {}); # find rule that allows all TCP port 80 traffic from 10.1.2.3 to # 192.168.1.1 ($rv, $rule_num) = $ipt_obj->find_ip_rule('10.1.2.3', '192.168.1.2', 'filter', 'INPUT', 'ACCEPT', {'protocol' => 'tcp', 's_port' => 0, 'd_port' => 80}); # add rule at the 5th rule position to allow all traffic from # 10.1.2.3 to 192.168.1.2 via the INPUT chain in the filter table ($rv, $out_ar, $errs_ar) = $ipt_obj->add_ip_rule('10.1.2.3', '192.168.1.2', 5, 'filter', 'INPUT', 'ACCEPT', {}); # add rule at the 4th rule position to allow all traffic from # 10.1.2.3 to 192.168.1.2 over TCP port 80 via the CUSTOM chain # in the filter table ($rv, $out_ar, $errs_ar) = $ipt_obj->add_ip_rule('10.1.2.3', '192.168.1.2', 4, 'filter', 'CUSTOM', 'ACCEPT', {'protocol' => 'tcp', 's_port' => 0, 'd_port' => 80}); # append rule at the end of the CUSTOM chain in the filter table to # allow all traffic from 10.1.2.3 to 192.168.1.2 via port 80 ($rv, $out_ar, $errs_ar) = $ipt_obj->append_ip_rule('10.1.2.3', '192.168.1.2', 'filter', 'CUSTOM', 'ACCEPT', {'protocol' => 'tcp', 's_port' => 0, 'd_port' => 80}); # run an arbitrary iptables command and collect the output ($rv, $out_ar, $errs_ar) = $ipt_obj->run_ipt_cmd( '/sbin/iptables -v -n -L'); =head1 DESCRIPTION The C package provide an interface to manipulate iptables policies on Linux systems through the direct execution of iptables commands. Although making a perl extension of libiptc provided by the iptables project is possible (and has been done by the IPTables::libiptc module available from CPAN), it is also easy enough to just execute iptables commands directly in order to both parse and change the configuration of the policy. Further, this simplifies installation since the only external requirement is (in the spirit of scripting) to be able to point IPTables::ChainMgr at an installed iptables binary instead of having to compile against a library. =head1 FUNCTIONS The IPTables::ChainMgr extension provides an object interface to the following functions: =over 4 =item chain_exists($table, $chain) This function tests whether or not a chain (e.g. 'INPUT') exists within the specified table (e.g. 'filter'). This is most useful to test whether a custom chain has been added to the running iptables policy. The return values are (as with many IPTables::ChainMgr functions) an array of three things: a numeric value, and both the stdout and stderr of the iptables command in the form of array references. So, an example invocation of the chain_exists() function would be: ($rv, $out_ar, $errs_ar) = $ipt_obj->chain_exists('filter', 'CUSTOM'); If $rv is 1, then the CUSTOM chain exists in the filter table, and 0 otherwise. The $out_ar array reference contains the output of the command "/sbin/iptables -t filter -v -n -L CUSTOM", which will contain the rules in the CUSTOM chain (if it exists) or nothing (if not). The $errs_ar array reference contains the stderr of the iptables command. =item create_chain($table, $chain) This function creates a chain within the specified table. Again, three return values are given like so: ($rv, $out_ar, $errs_ar) = $ipt_obj->create_chain('filter', 'CUSTOM'); Behind the scenes, the create_chain() function in the example above runs the iptables command "/sbin/iptables -t filter -N CUSTOM". =item flush_chain($table, $chain) This function flushes all rules from chain in the specified table, and three values are returned: ($rv, $out_ar, $errs_ar) = $ipt_obj->flush_chain('filter', 'CUSTOM'); The flush_chain() function in the example above executes the iptables command "/sbin/iptables -t filter -F CUSTOM" =item delete_chain($table, $jump_from_chain, $chain) This function deletes a chain from the specified table along with any jump rule to which packets are jumped into this chain: ($rv, $out_ar, $errs_ar) = $ipt_obj->delete_chain('filter', 'INPUT', 'CUSTOM'); Internally a check is performed to see whether the chain exists within the table, and global jump rules are removed from the jump chain before deletion (a chain cannot be deleted until there are no references to it). In the example above, the CUSTOM chain is deleted after any jump rule to this chain from the INPUT chain is also deleted. =item find_ip_rule($src, $dst, $table, $chain, $target, %extended_info) This function parses the specified chain to see if there is a rule that matches the $src, $dst, $target, and (optionally) any %extended_info criteria. The return values are the rule number in the chain (or zero if it doesn't exist), and the total number of rules in the chain. Below are two examples; the first is to find an ACCEPT rule for 10.1.2.3 to communicate with 192.168.1.2 in the INPUT chain, and the second is the same except that the rule is restricted to TCP port 80: ($rulenum, $chain_rules) = $ipt_obj->find_ip_rule('10.1.2.3', '192.168.1.2', 'filter', 'INPUT', 'ACCEPT', {}); if ($rulenum) { print "matched rule $rulenum out of $chain_rules rules\n"; } ($rulenum, $chain_rules) = $ipt_obj->find_ip_rule('10.1.2.3', '192.168.1.2', 'filter', 'INPUT', 'ACCEPT', {'protocol' => 'tcp', 's_port' => 0, 'd_port' => 80}); if ($rulenum) { print "matched rule $rulenum out of $chain_rules rules\n"; } =item add_ip_rule($src, $dst, $rulenum, $table, $chain, $target, %extended_info) This function inserts a rule into the running iptables chain and table at the specified rule number. Return values are success or failure along with the iptables stdout and stderr. =item append_ip_rule($src, $dst, $table, $chain, $target, %extended_info) This function appends a rule at the end of the iptables chain in the specified table. Return values are success or failure along with the iptables stdout and stderr. =item delete_ip_rule($src, $dst, $table, $chain, $target, %extended_info) This function searches for and then deletes a matching rule within the specified chain. Return values are success or failure along with the iptables stdout and stderr. =item add_jump_rule($table, $from_chain, $rulenum, $to_chain) This function adds a jump rule (after making sure it doesn't already exist) into the specified chain. The $rulenum variable tells the function where within the calling chain the new jump rule should be placed. Here is an example to force all packets regardless of source or destination to be jumped to the CUSTOM chain from the INPUT chain at rule 4: ($rv, $out_ar, $errs_ar) = $ipt_obj->add_jump_rule('filter', 'INPUT', 4, 'CUSTOM'); =item run_ipt_cmd($cmd) This function is a generic work horse function for executing iptables commands, and is used internally by IPTables::ChainMgr functions. It can also be used by a script that imports the IPTables::ChainMgr extension to provide a consistent mechanism for executing iptables. Three return values are given: success (1) or failure (0) of the iptables command (yes, this backwards from the normal exit status of Linux/*NIX binaries), and array references to the iptables stdout and stderr. Here is an example to list all rules in the user-defined chain "CUSTOM": ($rv, $out_ar, $errs_ar) = $ipt_obj->run_ipt_cmd('/sbin/iptables -t filter -v -n -L CUSTOM'); if ($rv) { print "rules:\n"; print for @$out_ar; } =back =head1 SEE ALSO The IPTables::ChainMgr extension is closely associated with the IPTables::Parse extension, and both are heavily used by the psad, fwsnort, and fwknop projects to manipulate iptables policies based on various criteria (see the psad(8), fwsnort(8), and fwknop(8) man pages). As always, the iptables(8) man page provides the best information on command line execution and theory behind iptables. Although there is no mailing that is devoted specifically to the IPTables::ChainMgr extension, questions about the extension will be answered on the following lists: The psad mailing list: http://lists.sourceforge.net/lists/listinfo/psad-discuss The fwknop mailing list: http://lists.sourceforge.net/lists/listinfo/fwknop-discuss The fwsnort mailing list: http://lists.sourceforge.net/lists/listinfo/fwsnort-discuss The latest version of the IPTables::ChainMgr extension can be found at: http://www.cipherdyne.org/modules/ =head1 CREDITS Thanks to the following people: Franck Joncourt Grant Ferley Darien Kindlund =head1 AUTHOR The IPTables::ChainMgr extension was written by Michael Rash Fmbr@cipherdyne.orgE> to support the psad, fwknop, and fwsnort projects. Please send email to this address if there are any questions, comments, or bug reports. =head1 COPYRIGHT AND LICENSE Copyright (C) 2005-2008 by Michael Rash This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself, either Perl version 5.8.5 or, at your option, any later version of Perl 5 you may have available. =cut