Routing question

David Demland plug-discuss@lists.plug.phoenix.az.us
Mon, 20 Jan 2003 15:45:40 -0700


I just got my server up and running after having some problems upgrading the
kernel. Now my iptables scripts will not route traffic. I have been using
this script for a while without any problems before. What would cause the
problem? The script is:

#!/bin/bash
# ----------------------------------------------------------------------|
#	      This is it...MonMotha's Firewall 2.3.7!			|
#	     All your hackers are belong to Linux/Netfilter!		|
# ----------------------------------------------------------------------|
# 2.3 RELEASE NOTES: This is the 2.2 series with some extra stuff,	|
# including MAC address matching, stateful matching, port forwarding,	|
# per-proto accept behavior, and some other stuff that I might think	|
# about adding later.							|
# ----------------------------------------------------------------------|
# COMMENTS from MonMotha:						|
#									|
# I will be entering "feature freeze" when 2.3.8 goes final. Please	|
# make sure to have any patches or feature requests in by then.		|
# I expect 2.3.7 to be closing in on deserving the "stable" marking.	|
#						--MonMotha		|
#									|
# Please note the change of my e-mail address.  The new address is:	|
# obi-wan@starwarsfan.com.  The old address (bvmopen@usa.net) will be	|
# discontinued as of July 31, 2001.					|
#						--MonMotha		|
#									|
# When e-mailing to report a bug, please check first that it has not	|
# already been fixed in the next prerelease (which can be found at the	|
# distribution site).							|
#						--MonMotha		|
#									|
# Before e-mailing me, please check the distribution site (which can be	|
# found at http://freshmeat.net/projects/mothafirewall as it changes	|
# sometimes) for a new version.						|
#						--MonMotha		|
#									|
# Please...PLEASE give me feedback on your experiences with this script	|
# I would really like to know what everyone wants, what works, and	|
# about the inevitable bugs present in anything.			|
#									|
# 		Direct all feedback to: obi-wan@starwarsfan.com		|
# 						--MonMotha		|
#									|
# When e-mailing with problems, please include firewall script version,	|
# iptables version, kernel version, and GNU BASH version.  If you think	|
# your problem might be related to kernel configuration, please attach	|
# the .config file for your kernel.					|
#						--MonMotha		|
#									|
# ----------------------------------------------------------------------|
# SYSTEM REQUIREMENTS: You must have either compiled the appropriate	|
# iptables support into your 2.4 kernel or have loaded all the		|
# applicable modules BEFORE you run this script.  This script will not	|
# load modules for you.							|
#									|
# 	You will need (at least) the following kernel options to use	|
# this firewall: CONFIG_NETFILTER, CONFIG_IP_NF_IPTABLES,		|
# CONFIG_IP_NF_FILTER, CONFIG_IP_NF_MATCH_STATE and			|
# CONFIG_IP_NF_TARGET_REJECT.						|
# 	To use the masquerading you will also need (at least):		|
# CONFIG_IP_NF_CONNTRACK, CONFIG_IP_NF_NAT, CONFIG_IP_NF_NAT_NEEDED	|
# and CONFIG_IP_NF_TARGET_MASQUERADE.					|
# 	Additional options may be needed to use other features.		|
# 									|
# You need iptables.  Get it at "http://netfilter.filewatcher.org".	|
# Some of the features will need patches only present in the CVS	|
#									|
# This script was written (and partially tested) with iptables CVS	|
# and kernel 2.4.x (non testing) in mind.				|
# 									|
# Also, this is a BASH shell script...any 2.x version of GNU BASH	|
# should work.								|
# ----------------------------------------------------------------------|
#									|
# ALL USERS, READ THE FOLLOWING:					|
#									|
# This is distributed under the modified BSD liscense:			|
#									|
# Redistribution and use in source and binary forms, with or without 	|
# modification, are permitted provided that the following conditions 	|
# are met:								|
#									|
#    1.Redistributions of source code must retain the above copyright	|
#      notice, this list of conditions and the following disclaimer.	|
#    2.Redistributions in binary form must reproduce the above		|
#      copyright notice, this list of conditions and the following	|
#      disclaimer in the documentation and/or other materials provided	|
#      with the distribution.						|
#    3.The name of the author may not be used to endorse or promote	|
#      products derived from this software without specific prior	|
#      written permission.						|
#									|
# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR	|
# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED	|
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE	|
# ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY	|
# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL	|
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE	|
# GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS		|
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER	|
# IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR	|
# OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN	|
# IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE				|
#									|
# While this may be used freely for commercial use, I do REQUEST that	|
# any commercial users please tell me via e-mail at			|
# obi-wan@starwarsfan.com that they are using it, why they chose it,	|
# how well it works, etc.						|
#									|
# ----------------------------------------------------------------------|
# IMPORTANT: This firewall is currently in beta! It may be too		|
# restrictive or insecure.						|
# ----------------------------------------------------------------------|
# CHANGELOG: (Since 2.3.0-pre1a only)					|
#	version 2.3.7: No changes from pre8				|
#	version 2.3.7-pre8: Change email address on liscense		|
#			    Revert to pre6 behavior of dropping ICMP	|
#				echo-request (take global DROP= policy)	|
#			    Allow everything from interface lo		|
#			    Correct pre7 changelog			|
#			    Special rules for DHCP servers		|
#	version 2.3.7-pre7: Fix version number in changelog entry below	|
#			    Fix 127.0.0.1 INPUT handling.		|
#			    Only enable IP forwarding if it's needed	|
#				(INTERNAL_LAN defined)			|
#			    Tweak flood parameters			|
#			    Hostwise allows now override explicit,	|
#				denies but not blackholes		|
#			    ICMP echo-request will no longer take the	|
#				specified drop policy when it doesn't	|
#				comply with limits, straight DROP will	|
#				be used instead				|
#			    Fix REJECT handling in TREJECT and LTREJECT	|
#			    Add transparent proxy support (Joshua Link)	|
#	version 2.3.7-pre6: Fix status reporting on SSR SysCtl loop	|
#			    Fix the SSR SysCtl loop			|
#			    Remove stateful match from forward chain	|
#	version 2.3.7-pre5: Make the default policy actually be DROP	|
#				instead of just saying it is		|
#			    Add stateful matching to forward chain to	|
#				prevent people from routing onto your	|
#				internal network (please tell me if 	|
#				breaks anything). Thanks to Martin	|
#				Mosny for noticing this			|
#			    Block Source Routed Packets to help with	|
#				the above problem			|
#			    Add option for TCP SynCookies on or off	|
#			    Fix BLACKHOLE directive (was being applied	|
#				to INPUT/OUTPUT after the jump for	|
#				INETIN/INETOUT so didn't apply for	|
#				the internet).  Thanks to Gerry Doris	|
#				for noticing this			|
#			    Add DHCP client to default UDP port allows	|
#			    Note email address change			|
#			    Changed emphesis in comments		|
#			    Forwarding of port ranges (Vinny and Eddie)	|
#	version 2.3.7-pre4: Line 414, missing subnet match caused all	|
#				packets from anywhere to be allowed.	|
#				Fixed.					|
#	version 2.3.7-pre3: Fix missing fi (fatal syntax error)		|
#			    Fix logging in TCPACCEPT chain		|
#	version 2.3.7-pre2: Add route verification (thanks to Jeremy	|
#				Frank)					|
#			    Add blackhole directive			|
#			    Updated configuration sanity checks		|
#			    Ripped out SSH Stuff as it isn't needed	|
#			    True default DROP on INPUT			|
#			    Don't run the INTERNAL_LAN loop if no nets	|
#			    Upped the default SYN limit as large	|
#				numbers of small FTP transfers would	|
#				overload it quickly			|
#			    Form cleanups				|
#	version 2.3.7-pre1: Maybe the FTP will work now (fixes for the	|
#				RELATED state)				|
#			    Now works with both LAN and DMZ iface null	|
#			    Moved static NAT to stable options		|
#			    Change parser to /bin/bash not /bin/sh	|
#	version 2.3.6: Add TTL mangling					|
#		       Added some more EFNet servers to the list	|
#		       Fix in the DMZOUT chain				|
#		       Fix FTP stuff					|
#	version 2.3.5: Fixes to make port forwarding work again		|
#	version 2.3.4: USE_MASQ has been changed to MASQ_LAN in port fw	|
#		       Fix syntax error in TCP port forwards		|
#		       General cleanup					|
#		       Fixes in port forwarding				|
#		       It's LTREJECT, not TLREJECT			|
#		       More TOS mangling				|
#	version 2.3.3: Fatal syntax error in IP forward detect fix	|
#		       Don't bail on no IP forward for no LAN		|
#	version 2.3.3-pre1: Reject with tcp-reset for TCP option	|
#			    Removed the huge list of censorship		|
#			    Moved the port forwards to stable options	|
#			    Moved the TOS mangling to stable options	|
#			    Check before enabling IP Forwarding and	|
#				IP SynCookies				|
#			    Don't run censorship loop if no rules	|
#			    Request low latency TOS on UDP packets for	|
#				games on ports 4000-7000 (Diablo II)	|
#			    Fix bad syntax in the port forwarding loops	|
#			    Reversed DMZIN and DMZOUT fixed		|
#			    Various syntax fixes			|
#			    Stateful inspection on forward chain	|
#			    Other stateful matching changes		|
#	version 2.3.2: Fixed bad syntax in DMZ_IFACE loop		|
#	version 2.3.2-pre2: Put a real liscense on it (BSD liscense)	|
#			    Changed format of ALLOW_HOSTWISE and	|
#				DENY_HOSTWISE to be less confusing	|
#				(the ":" was changed to ">")		|
#			    Added LOG_FLOOD option to tweak log limit	|
#			    Added SYN_FLOOD option to tweak SYN limit	|
#			    Added PING_FLOOD option to tweak PING limit	|
#	version 2.3.2-pre1: Stateful matching on active FTP and SSH	|
#				rules (thanks to Len Padilla)		|
#			    Fixed a minor bug in chain creation order	|
#				(thanks to Peter Lindman)		|
#			    TOS Optimizations (thanks to vesa alatalo)	|
#			    Begin DMZ Support				|
#			    Proofread comments and correct		|
#			    Use BASH builtins instead of sed 		|
#				(thanks to Craig Ludington)		|
#			    Fixed "USE_SNAT" bug in port forwarding	|
#				(has been changed to "SNAT_LAN")	|
#				(thanks to Frédéric Marchand)		|
#			    Tuned down default TCP allows (remove POP3)	|
#	version 2.3.1: Option for 1:1 or subnet:1 static NAT		|
#		       Internet censorship options			|
#	version 2.3.1-pre2: Added option to deny specific ports from	|
#				specific hosts				|
#			    Added limiting to logging chains to prevent	|
#				log DoSing				|
#			    Spiffed up comments				|
#			    Changed the "AUTH_ALLOW" and "DNS" options	|
#				to be more generic and flexible		|
#	version 2.3.1-pre1: Updated comments for new kernel version	|
#			    Removed double drop setting			|
#			    Updated for iptables-1.2			|
#			    Began a kernel option list			|
#	version 2.3.0: No changes from pre1g				|
#	version 2.3.0-pre1g: Tuned down default TCP allows		|
#			     Restructure to SSH loop			|
#			     Status Reporting Fixes (newlines, etc.)	|
#			     Fix log prefix length on accept loops	|
#	version 2.3.0-pre1f: Moved the ICMP echo-request limit to where	|
#				it should have been			|
#			     Allows the rest of the ICMP like it should	|
#			     Remove the interface matching from ICMP	|
#				echo-request (not needed)		|
#	version 2.3.0-pre1e: Fixed an issue in the invalid matching	|
#	version 2.3.0-pre1d: Spiffed up comments			|
#			     Port Forwarding				|
#			     Moved the deny setting to normal options	|
#	version 2.3.0-pre1c: Minor fixes that don't (currently) affect	|
#				functionality				|
#	version 2.3.0-pre1b: Security fix documented in 2.1.13		|
#			     Slight logic change in TCP_ALLOW loop	|
#			     Don't print allow messages if nothign is	|
#				allowed by that loop			|
#			     Changed IPTables download URL		|
#       version 2.3.0-pre1a: Initial branch from 2.1.12                 |
#                            Add stuff in release notes except port fw  |
# ----------------------------------------------------------------------|
# USE AT YOUR OWN RISK! THIS IS WITHOUT ANY WARRANTY STATED OR IMPLIED!	|
# That means NONE whatsoever, not even fitfullness for purpose!		|
# ----------------------------------------------------------------------|
# You NEED to set this!							|
# Configuration follows:						|
#									|
# Notes about configuration: 						|
# Some things take more than one option; separate with spaces.		|
# You probably don't want all the ports I have under here open, portscan|
#	yourself to find what you want open.				|
# If you want to used host-based identd allowing, do NOT put 113 in	|
#	TCP_ALLOW and DO set ALLOW_TCP_HOSTWISE (using 113 as the port).|
#	Of course, you can also put 113 in TCP_ALLOW to allow anyone.	|
# The same applies to DNS zone transfers (only use port 53 and UDP).	|
# MAC_MASQ is ONLY used for the purposes of masquerading and it will	|
#	override the MASQ_LAN setting for masquerading.	However, you	|
#	must still define MASQ_LAN properly.				|
# INTERNAL_LAN must always be properly defined.				|
# You can use hostnames anywhere, but you'll need to have access to the	|
# DNS server when the script runs and you might not get the expected	|
#	results since the DNS lookup is only done once.			|
# You can mix and match hosts with public IPs and masqueraded hosts in	|
# 	INTERNAL_LAN as long as you define the ones to use NAT later.	|
# DMZ support can currently be considered (at best) PREALPHA.		|
# It should work without a LAN, leave INTERNAL_LAN blank.		|


# Main configuration, modify to suit your setup.  Help can be found at:
#    http://www.mplug.org/phpwiki/index.php?MonMothaReferenceGuide
IPTABLES="/sbin/iptables"		# set to your iptables location, must be set
TCP_ALLOW="22"					# TCP ports to allow
UDP_ALLOW="68 6112 6119 4000"			# UDP ports to allow
INET_IFACE="eth1"				# the interface your internet's on (one only), must be
set
LAN_IFACE="eth0"				# the interface your LAN's on (one only)
INTERNAL_LAN="192.168.0.0/24 192.168.1.0/24"	# The internal LAN (including
DMZs but not censored hosts)
MASQ_LAN="192.168.0.0/24 192.168.1.0/24"	# the internal network(s) to be
masqueraded (this is overridden by MAC_MASQ)
SNAT_LAN=""					# Internal networks/hosts to use static NAT (format is
<internal ip or network>:<external ip>) (this is overridden by MAC_SNAT)
DROP="TREJECT"					# What to do with packets we don't want: DROP, REJECT,
TREJECT (Reject with tcp-reset for TCP), LDROP (log and drop), LREJECT (log
and reject), LTREJECT (log and reject with tcp-reset)
DENY_ALL=""					# Internet hosts to explicitly deny from accessing your
system at all
DENY_HOSTWISE_TCP=""				# Specific hosts to deny access to specific TCP
ports; format is "IP>PORT"
DENY_HOSTWISE_UDP=""				# Specific hosts to deny access to specific UDP
ports; format is "IP>PORT"
BLACKHOLE=""					# People you don't want to have anything to do with
(equivlent of my old TK_DROP).  This is a bidirectional drop.
BLACKHOLE_DROP="DROP"				# What to do for the blackholes (same options as
DROP directive above)
ALLOW_HOSTWISE_TCP=""				# Specific hosts allowed access to specific TCP
ports; format is "IP>PORT"
ALLOW_HOSTWISE_UDP=""				# Specific hosts allowed access to specific UDP
ports; format is "IP>PORT"
TCP_FW=""					# TCP port forwards, form is "SPORT:DPORT>IP"
UDP_FW=""					# UDP port forwards, form is "SPORT:DPORT>IP"
MANGLE_TOS_OPTIMIZE="TRUE"			# TOS "optimizations" on or off (TRUE/FALSE
toggle)
ENABLE="Y"					# Set to 'Y' when it's configured; this is for your own
saftey


# Flood Params.  You will still recieve the packets and the bandwidth will
be used, but this will cause floods to be ignored (useful against SYNFLOODS
especially)
LOG_FLOOD="2/s"					# Limit on logging (for LTREJECT, LREJECT and LDROP, the
packet will always take the policy regardless of logging)
SYN_FLOOD="20/s"				# GLOBAL limit on SYN packets (servers will probably
need even higher sustained rates as this isn't on a per IP basis)
PING_FLOOD="1/s"				# GLOBAL limit on ICMP echo-requests to reply to

# Outbound filters (they work, but are of limited functionality), probably
better to use a proxy here
ALLOW_OUT_TCP=""				# Internal hosts allowed to be forwarded out on TCP
(internet censorship!) (do not put this/these host/s in INTERNAL_LAN, but do
define their method of access [snat, masq] if not a public ip)

# Below here is experimental (please report your successes/failures)
MAC_MASQ=""					# MAC addresses permitted to use masquerading, leave blank
to not use
MAC_SNAT=""					# MAC addresses permitted to use static NAT, leave blank to
not use (format is <MAC Address>:<external ip>)
TTL_SAFE=""					# How many hops packets need to make once they get on your
LAN (null disables the mangling) (requires patch from patch-o-matic)
USE_SYNCOOKIES="TRUE"				# TCP SynCookies on or off (TRUE/FALSE toggle)
PROXY=""					# Redirect for Squid or other transparent proxy. Syntax to
specify the proxy is "host:port".
DHCP_SERVER="FALSE"				# Set to true if you run a DHCP server. DHCP clients
do not need this. This allows broadcasts to the server from potential
clients on the LAN to succeede. MUST DEFINE LAN_IFACE IF YOU USE THIS!

# Only touch these if you're daring (PREALPHA stuff, as in basically
non-functional)
DMZ_IFACE=""					# Interface your DMZ is on (leave blank if you don't have
one) MUST DEFINE LAN_IFACE IF YOU USE THIS!

# ----------------------------------------------------------------------|
# These control basic script behavior, there should be no need to	|
#	any of these settings for normal use.				|
# ----------------------------------------------------------------------|
FILTER_CHAINS="INETIN INETOUT DMZIN DMZOUT TCPACCEPT UDPACCEPT LDROP LREJECT
TREJECT LTREJECT"
LOOP_IFACE="lo"

# ----------------------------------------------------------------------|
# You shouldn't need to modify anything below here			|
# Main Script Starts							|
# ----------------------------------------------------------------------|

# Let's load it!
echo "Loading iptables firewall:"

# Configuration Sanity Checks
echo -n "Checking configuration..."
if ! [ -x $IPTABLES ] ; then
	echo
	echo "ERROR IN CONFIGURATION: IPTABLES doesn't exist or isn't executable!"
	exit 1
fi
if [ "$DMZ_IFACE" = "$LAN_IFACE" ] && [ "$LAN_IFACE" != "" ]; then
	echo
	echo "ERROR IN CONFIGURATION: DMZ_IFACE and LAN_IFACE can't be the same!"
	exit 1
fi
if [ "$DROP" = "" ] ; then
	echo
	echo "There needs to be a DROP policy (try TREJECT)!"
	exit 1
fi
if [ "$DROP" = "ACCEPT" ] ; then
	echo
	echo "The DROP policy is set to ACCEPT; there is no point in loading the
firewall as there wouldn't be one."
	exit 2
fi
if [ "$BLACKHOLE" != "" ] && [ "$BLACKHOLE_DROP" = "" ] ; then
	echo
	echo "You can't use blackholes and not have a policy for them!"
	exit 1
fi
if ! [ "$ENABLE" = "Y" ] ; then
	echo
	echo "You need to edit your configuration and set ENABLE to Y!"
	exit 99
fi
echo "passed"


# ===============================================
# -------Set some Kernel stuff via SysCTL--------
# ===============================================

# Turn on IP forwarding

if [ "$INTERNAL_LAN" != "" ] ; then
	echo -n "Checking IP Forwarding..."
	if [ -e /proc/sys/net/ipv4/ip_forward ] ; then
		echo 1 > /proc/sys/net/ipv4/ip_forward
		echo "enabled."
	else
		echo "support not found! This will cause problems if you need to do any
routing."
	fi
fi

# Enable TCP Syncookies
if [ "$USE_SYNCOOKIES" = "TRUE" ] ; then
	echo -n "Checking IP SynCookies..."
	if [ -e /proc/sys/net/ipv4/tcp_syncookies ] ; then
		echo 1 > /proc/sys/net/ipv4/tcp_syncookies
		echo "enabled."
	else
		echo "support not found, but that's OK."
	fi
fi

# Enable Route Verification to prevent martians and other such crud that
# seems to be commonplace on the internet today
echo -n "Checking Route Verification..."
if [ "$INET_IFACE" != "" ] ; then
	if [ -e /proc/sys/net/ipv4/conf/$INET_IFACE/rp_filter ] ; then
		echo 1 > /proc/sys/net/ipv4/conf/$INET_IFACE/rp_filter
		echo -n "activated:${INET_IFACE} "
	else
		echo "not found:${INET_IFACE} "
	fi
fi

if [ "$LAN_IFACE" != "" ] ; then
	if [ -e /proc/sys/net/ipv4/conf/$LAN_IFACE/rp_filter ] ; then
		echo 1 > /proc/sys/net/ipv4/conf/$LAN_IFACE/rp_filter
		echo -n "activated:${LAN_IFACE} "
	else
		echo "not found:${LAN_IFACE} "
	fi
fi

if [ "$DMZ_IFACE" != "" ] ; then
	if [ -e /proc/sys/net/ipv4/conf/$DMZ_IFACE/rp_filter ] ; then
		echo 1 > /proc/sys/net/ipv4/conf/$DMZ_IFACE/rp_filter
		echo -n "activated:${DMZ_IFACE} "
	else
		echo "not found:${DMZ_IFACE} "
	fi
fi
echo

# Tell the Kernel to Ignore Source Routed Packets
echo -n "Refusing SSR Packets via SysCtl..."
if [ "$INET_IFACE" != "" ] ; then
	if [ -e /proc/sys/net/ipv4/conf/$INET_IFACE/accept_source_route ] ; then
		echo "0" > /proc/sys/net/ipv4/conf/$INET_IFACE/accept_source_route
		echo -n "activated:${INET_IFACE} "
	else
		echo "not found:${INET_IFACE} "
	fi
fi

if [ "$LAN_IFACE" != "" ] ; then
	if [ -e /proc/sys/net/ipv4/conf/$LAN_IFACE/accept_source_route ] ; then
		echo "0" > /proc/sys/net/ipv4/conf/$LAN_IFACE/accept_source_route
		echo -n "activated:${LAN_IFACE} "
	else
		echo "not found:${LAN_IFACE} "
	fi
fi

if [ "$DMZ_IFACE" != "" ] ; then
	if [ -e /proc/sys/net/ipv4/conf/$DMZ_IFACE/accept_source_route ] ; then
		echo "0" > /proc/sys/net/ipv4/conf/$DMZ_IFACE/accept_source_route
		echo -n "activated:${DMZ_IFACE} "
	else
		echo "not found:${DMZ_IFACE} "
	fi
fi
echo

# ===============================================
# --------Actual NetFilter Stuff Follows---------
# ===============================================

# Flush everything
# If you need compatability, you can comment some or all of these out,
# but remember, if you re-run it, it'll just add the new rules in, it
# won't remove the old ones for you then, this is how it removes them.
echo -n "Flush: "
${IPTABLES} -t filter -F INPUT
echo -n "INPUT "
${IPTABLES} -t filter -F OUTPUT
echo -n "OUTPUT1 "
${IPTABLES} -t filter -F FORWARD
echo -n "FORWARD "
${IPTABLES} -t nat -F PREROUTING
echo -n "PREROUTING1 "
${IPTABLES} -t nat -F OUTPUT
echo -n "OUTPUT2 "
${IPTABLES} -t nat -F POSTROUTING
echo -n "POSTROUTING "
${IPTABLES} -t mangle -F PREROUTING
echo -n "PREROUTING2 "
${IPTABLES} -t mangle -F OUTPUT
echo -n "OUTPUT3"
echo

# Create new chains
# Output to /dev/null in case they don't exist from a previous invocation
echo -n "Creating chains: "
for chain in ${FILTER_CHAINS} ; do
	${IPTABLES} -t filter -F ${chain} > /dev/null 2>&1
	${IPTABLES} -t filter -X ${chain} > /dev/null 2>&1
	${IPTABLES} -t filter -N ${chain}
	echo -n "${chain} "
done
echo

# Default Policies
# INPUT policy is drop as of 2.3.7-pre5
# Policy can't be reject becuase of kernel limitations
echo -n "Default Policies: "
${IPTABLES} -t filter -P INPUT DROP
echo -n "INPUT:DROP "
${IPTABLES} -t filter -P OUTPUT ACCEPT
echo -n "OUTPUT:ACCEPT "
${IPTABLES} -t filter -P FORWARD DROP
echo -n "FORWARD:DROP "
echo

# Local traffic to internet or crossing subnets
# This should cover what we need if we don't use masquerading
# Unfortunately, MAC address matching isn't bidirectional (for
#   obvious reasons), so IP based matching is done here
echo -n "Local Traffic Rules: "
if [ "$INTERNAL_LAN" != "" ] ; then
	for subnet in ${INTERNAL_LAN} ; do
		${IPTABLES} -t filter -A FORWARD -s ${subnet} -j ACCEPT
		${IPTABLES} -t filter -A FORWARD -d ${subnet} -j ACCEPT
		${IPTABLES} -t filter -A INPUT -s ${subnet} -j ACCEPT
		echo -n "${subnet}:ACCEPT "
	done
	echo
fi

# 127.0.0.0/8 used to need an entry in INTERNAL_LAN, but routing of that
isn't needed
# so an allow is placed on INPUT so that the computer can talk to itself :)
${IPTABLES} -t filter -A INPUT -i ${LOOP_IFACE} -j ACCEPT
echo -n "loopback:ACCEPT "

# DHCP server magic
# Allow broadcasts from LAN to UDP port 67 (DHCP server)
if [ "$DHCP_SERVER" = "TRUE" ] ; then
	${IPTABLES} -t filter -A INPUT -i ${LAN_IFACE} -p udp --dport 67 -j ACCEPT
fi
echo #newline from local traffic rules



if [ "$PROXY" != "" ] ; then
	echo -n "Setting up Transparent Proxy to ${PROXY}: "
	for subnet in ${INTERNAL_LAN} ; do
			echo "$PROXY" | {
                                IFS=':' read host port
                                if [ "$host" = "localhost" ] || [ "$host" =
"127.0.0.1" ] ; then
                                        ${IPTABLES} -t nat -A PREROUTING -s
${subnet} -p tcp --dport 80 -j REDIRECT --to-port ${port}
                                        echo -n "${subnet}:PROXY "
                                else
                                        ${IPTABLES} -t nat -A PREROUTING -s
${subnet} -p tcp --dport 80 -j DNAT --to ${host}:${port}
                                        echo -n "${subnet}:PROXY "
                                fi
                        }
        done
        echo
fi

if [ "$ALLOW_OUT_TCP" != "" ] ; then
	echo -n "Internet censorship TCP allows: "
	for rule in ${ALLOW_OUT_TCP} ; do
		echo "$rule" | {
			IFS=':' read intip destip
		        ${IPTABLES} -t filter -A FORWARD -s ${intip} -d ${destip} -o
${INET_IFACE} -j ACCEPT
			${IPTABLES} -t filter -A FORWARD -d ${intip} -s ${destip} -i
${INET_IFACE} -j ACCEPT
			echo -n "${intip}:${destip} "
		}
	done
	echo
fi

# Set up basic NAT if the user wants it
if [ "$MASQ_LAN" != "" ] ; then
	echo -n "Setting up masquerading: "
	if [ "$MAC_MASQ" = "" ] ; then
		for subnet in ${MASQ_LAN} ; do
			${IPTABLES} -t nat -A POSTROUTING -s ${subnet} -o ${INET_IFACE} -j
MASQUERADE
			echo -n "${subnet}:MASQUERADE "
		done
	else
		for address in ${MAC_MASQ} ; do
			${IPTABLES} -t nat -A POSTROUTING -m mac --mac-source ${address} -o
${INET_IFACE} -j MASQUERADE
			echo -n "${address}:MASQUERADE "
		done
	fi
	echo
fi
if [ "$SNAT_LAN" != "" ] ; then #Static NAT used
	echo -n "Setting up static NAT: "
        if [ "$MAC_SNAT" = "" ] ; then
                for rule in ${SNAT_LAN} ; do
			echo "$rule" | {
				IFS=':' read host destip
	                        ${IPTABLES} -t nat -A POSTROUTING -s ${host} -o
${INET_IFACE} -j SNAT --to-source ${destip}
	                        echo -n "${subnet}:SNAT "
			}
                done
        else
                for rule in ${MAC_SNAT} ; do
			echo "$rule" | {
				IFS=':' read address destip
	                        ${IPTABLES} -t nat -A POSTROUTING -m
mac --mac-source ${address} -o ${INET_IFACE} -j SNAT --to-source ${destip}
	                        echo -n "${address}:SNAT "
			}
                done
        fi
        echo
fi

#TCP Port-Forwards
if [ "$TCP_FW" != "" ] ; then
	echo -n "TCP Port Forwards: "
	for rule in ${TCP_FW} ; do
		echo "$rule" | {
			IFS=':>' read srcport destport host
				echo "$srcport" | {
					IFS='-' read fsp lsp
					if [ "$lsp" != "" ] ; then
						echo "$destport" | {
							IFS='-' read fdp ldp
							${IPTABLES} -t nat -A PREROUTING -p tcp -i ${INET_IFACE} --dport
${fsp}:${lsp} -j DNAT --to-destination ${host}:${destport}
							${IPTABLES} -t filter -A FORWARD -p tcp -d ${host} --dport
${fdp}:${ldp} -j ACCEPT
						}
					else
						${IPTABLES} -t nat -A PREROUTING -p tcp -i ${INET_IFACE} --dport
${srcport} -j DNAT --to-destination ${host}:${destport}
						${IPTABLES} -t filter -A FORWARD -p tcp -d ${host} --dport
${destport} -j ACCEPT
					fi
					echo -n "${rule} "
				}
		}
	done
	echo
fi

#UDP Port Forwards
if [ "$UDP_FW" != "" ] ; then
	echo -n "UDP Port Forwards: "
	for rule in ${UDP_FW} ; do
		echo "$rule" | {
			IFS=':>' read srcport destport host
				echo "$srcport" | {
					IFS='-' read fsp lsp
					if [ "$lsp" != "" ] ; then
						echo "$destport" | {
							IFS='-' read fdp ldp
	        					${IPTABLES} -t nat -A PREROUTING -p udp -i
${INET_IFACE} --dport ${fsp}:${lsp} -j DNAT --to-destination
${host}:${destport}
							${IPTABLES} -t filter -A FORWARD -p udp -d ${host} --dport
${fdp}:${ldp} -j ACCEPT
						}
					else
						${IPTABLES} -t nat -A PREROUTING -p udp -i ${INET_IFACE} --dport
${srcport} -j DNAT --to-destination ${host}:${destport}
						${IPTABLES} -t filter -A FORWARD -p udp -d ${host} --dport
${destport} -j ACCEPT
					fi
					echo -n "${rule} "
	        		}
		}
        done
        echo
fi

# ===============================================
# -------Chain setup before jumping to them------
# ===============================================


# Set up INET chains
echo -n "Setting up INET chains: "
${IPTABLES} -t filter -A INPUT -i ${INET_IFACE} -j INETIN
echo -n "INETIN "
${IPTABLES} -t filter -A OUTPUT -o ${INET_IFACE} -j INETOUT
echo -n "INETOUT "
echo

# For now we'll subject the DMZ to the same rules as the internet when going
onto the trusted LAN
# And we'll let it go anywhere on the internet
if [ "$DMZ_IFACE" != "" ] ; then
	echo -n "Setting up DMZ Chains: "
	${IPTABLES} -A OUTPUT -o ${DMZ_IFACE} -j DMZOUT
	echo -n "DMZOUT "
	${IPTABLES} -A INPUT -i ${DMZ_IFACE} -j DMZIN
	echo -n "DMZIN "
	echo
	echo -n "DMZ for LAN Forwarding to INETIN..."
	${IPTABLES} -A DMZOUT -o ${LAN_IFACE} -j INETIN
	echo "done"
	echo -n "DMZ for Internet Forwarding to INETOUT..."
	${IPTABLES} -A DMZOUT -o ${INET_IFACE} -j INETOUT
	echo -n "done"
fi


#These logging chains are valid to specify in DROP= above
#Set up LDROP
echo -n "Setting up drop chains chains: "
${IPTABLES} -t filter -A LDROP -p tcp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "TCP Dropped "
${IPTABLES} -t filter -A LDROP -p udp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "UDP Dropped "
${IPTABLES} -t filter -A LDROP -p icmp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "ICMP Dropped "
${IPTABLES} -t filter -A LDROP -f -m limit --limit ${LOG_FLOOD} -j
LOG --log-level warning --log-prefix "FRAGMENT Dropped "
${IPTABLES} -t filter -A LDROP -j DROP
echo -n "LDROP "

#And LREJECT too
${IPTABLES} -t filter -A LREJECT -p tcp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "TCP Rejected "
${IPTABLES} -t filter -A LREJECT -p udp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "UDP Rejected "
${IPTABLES} -t filter -A LREJECT -p icmp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "ICMP Dropped "
${IPTABLES} -t filter -A LREJECT -f -m limit --limit ${LOG_FLOOD} -j
LOG --log-level warning --log-prefix "FRAGMENT Rejected "
${IPTABLES} -t filter -A LREJECT -j REJECT
echo -n "LREJECT "

#Don't forget TREJECT
${IPTABLES} -t filter -A TREJECT -p tcp -j REJECT --reject-with tcp-reset
${IPTABLES} -t filter -A TREJECT -p udp -j REJECT --reject-with
icmp-port-unreachable
${IPTABLES} -t filter -A TREJECT -p icmp -j DROP
${IPTABLES} -t filter -A TREJECT -j REJECT
echo -n "TREJECT "

#And LTREJECT
${IPTABLES} -t filter -A LTREJECT -p tcp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "TCP Rejected "
${IPTABLES} -t filter -A LTREJECT -p udp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "UDP Rejected "
${IPTABLES} -t filter -A LTREJECT -p icmp -m limit --limit ${LOG_FLOOD} -j
LOG --log-level info --log-prefix "ICMP Dropped "
${IPTABLES} -t filter -A LTREJECT -f -m limit --limit ${LOG_FLOOD} -j
LOG --log-level warning --log-prefix "FRAGMENT Rejected "
${IPTABLES} -t filter -A LTREJECT -p tcp -j REJECT --reject-with tcp-reset
${IPTABLES} -t filter -A LTREJECT -p udp -j REJECT --reject-with
icmp-port-unreachable
${IPTABLES} -t filter -A LTREJECT -p icmp -j DROP
${IPTABLES} -t filter -A LTREJECT -j REJECT
echo -n "LTREJECT "

#newline
echo


# Set up the per-proto ACCEPT chains
echo -n "Setting up per-proto ACCEPT: "

# TCPACCEPT
# SYN Flood "Protection"
${IPTABLES} -t filter -A TCPACCEPT -p tcp --syn -m limit --limit
${SYN_FLOOD} -j ACCEPT
${IPTABLES} -t filter -A TCPACCEPT -p tcp --syn -m limit --limit
${LOG_FLOOD} -j LOG --log-prefix "Possible SynFlood "
${IPTABLES} -t filter -A TCPACCEPT -p tcp --syn -j ${DROP}
${IPTABLES} -t filter -A TCPACCEPT -p tcp ! --syn -j ACCEPT
# Log anything that hasn't matched yet and ${DROP} it since it isn't TCP and
shouldn't be here
${IPTABLES} -t filter -A TCPACCEPT -m limit --limit ${LOG_FLOOD} -j
LOG --log-prefix "Mismatch in TCPACCEPT "
${IPTABLES} -t filter -A TCPACCEPT -j ${DROP}
echo -n "TCPACCEPT "

#UDPACCEPT
${IPTABLES} -t filter -A UDPACCEPT -p udp -j ACCEPT
# Log anything not UDP and ${DROP} it since it's not supposed to be here
${IPTABLES} -t filter -A UDPACCEPT -m limit --limit ${LOG_FLOOD} -j
LOG --log-prefix "Mismatch on UDPACCEPT "
${IPTABLES} -t filter -A UDPACCEPT -j ${DROP}
echo -n "UDPACCEPT "

#Done
echo


# ===============================================
# --------------Apply configuration--------------
# ===============================================

#Blackholes will not be overridden by hostwise allows
if [ "$BLACKHOLE" != "" ] ; then
	echo -n "Blackholes: "
	for host in ${BLACKHOLE} ; do
		${IPTABLES} -t filter -A INETIN -s ${host} -j ${BLACKHOLE_DROP}
		${IPTABLES} -t filter -A INETOUT -d ${host} -j ${BLACKHOLE_DROP}
		echo -n "${host} "
	done
	echo
fi

#Hostwise allows (to bypass explicit denies)
if [ "$ALLOW_HOSTWISE_TCP" != "" ] ; then
	echo -n "Hostwise TCP Allows: "
	for rule in ${ALLOW_HOSTWISE_TCP} ; do
		echo "$rule" | {
			IFS='>' read  host port
			${IPTABLES} -t filter -A INETIN -p tcp -s ${host} --dport ${port} -j
ACCEPT
			echo -n "${rule} "
		}
	done
	echo
fi

if [ "$ALLOW_HOSTWISE_UDP" != "" ] ; then
	echo -n "Hostwise UDP Allows: "
	for rule in ${ALLOW_HOSTWISE_UDP} ; do
		echo "$rule" | {
			IFS='>' read  host port
			${IPTABLES} -t filter -A INETIN -p udp -s ${host} --dport ${port} -j
ACCEPT
			echo -n "${rule} "
		}
	done
	echo
fi

# =================================================
# ----------------Explicit Denies------------------
# =================================================



if [ "$DENY_ALL" != "" ] ; then
	echo -n "Denying hosts: "
	for host in ${DENY_ALL} ; do
		${IPTABLES} -t filter -A INETIN -s ${host} -j ${DROP}
		echo -n "${host}:${DROP}"
	done
	echo
fi

if [ "$DENY_HOSTWISE_TCP" != "" ] ; then
	echo -n "Hostwise TCP Denies: "
	for rule in ${DENY_HOSTWISE_TCP} ; do
		echo "$rule" | {
			IFS='>' read host port
			${IPTABLES} -t filter -A INETIN -p tcp -s ${host} --dport ${port} -j
${DROP}
			echo -n "${rule} "
		}
	done
	echo
fi

if [ "$DENY_HOSTWISE_UDP" != "" ] ; then
	echo -n "Hostwise UDP Denies: "
	for rule in ${DENY_HOSTWISE_UDP} ; do
		echo "$rule" | {
			IFS='>' read host port
			${IPTABLES} -t filter -A INETIN -p udp -s ${host} --dport ${port} -j
${DROP}
			echo -n "${rule} "
		}
	done
	echo
fi

#Invalid packets are always annoying
echo -n "${DROP}ing invalid packets..."
${IPTABLES} -t filter -A INETIN -m state --state INVALID -j ${DROP}
echo "done"


# ================================================================
# ------------Allow stuff we have chosen to allow in--------------
# ================================================================

# Flood "security"
# You'll still respond to these if they comply with the limits (set in
config)
# There is a more elegant way to set this using sysctl, however this has the
# 	advantage that the kernel ICMP stack never has to process it, lessening
#	the chance of a very serious flood overloading your kernel.
# This is just a packet limit, you still get the packets on the interface
and
#    still may experience lag if the flood is heavy enough
echo -n "Flood limiting: "
# Ping Floods (ICMP echo-request)
${IPTABLES} -t filter -A INETIN -p icmp --icmp-type echo-request -m
limit --limit ${PING_FLOOD} -j ACCEPT
${IPTABLES} -t filter -A INETIN -p icmp --icmp-type echo-request -j ${DROP}
echo -n "ICMP-PING "
echo

# Probably shouldn't let em all back in, someone give me a good list of the
#	WORTHWHILE ICMP messages.
echo -n "Allowing the rest of the ICMP messages in..."
${IPTABLES} -t filter -A INETIN -p icmp --icmp-type ! echo-request -j ACCEPT
echo "done"

if [ "$TCP_ALLOW" != "" ] ; then
	echo -n "TCP Input Allow: "
	for port in ${TCP_ALLOW} ; do
 		${IPTABLES} -t filter -A INETIN -p tcp --dport ${port} -j TCPACCEPT
		echo -n "${port} "
	done
	echo
fi

if [ "$UDP_ALLOW" != "" ] ; then
	echo -n "UDP Input Allow: "
	for port in ${UDP_ALLOW} ; do
		${IPTABLES} -t filter -A INETIN -p udp --dport ${port} -j UDPACCEPT
		echo -n "${port} "
	done
	echo
fi

echo -n "Allowing established outbound connections back in..."
${IPTABLES} -t filter -A INETIN -m state --state ESTABLISHED -j ACCEPT
echo "done"

# RELATED on high ports only for security
echo -n "Allowing related inbound connections..."
${IPTABLES} -t filter -A INETIN -p tcp --dport 1024:65535 -m state --state
RELATED -j TCPACCEPT
${IPTABLES} -t filter -A INETIN -p udp --dport 1024:65535 -m state --state
RELATED -j UDPACCEPT
echo "done"


# =================================================
# ----------------Packet Mangling------------------
# =================================================


# TTL mangling
# This is probably just for the paranoid, but hey, isn't that what
#	all security guys are? :)
if [ "$TTL_SAFE" != "" ] ; then
	${IPTABLES} -t mangle -A PREROUTING -i ${INET_IFACE} -j TTL --ttl-set
${TTL_SAFE}
fi

# Type of Service mangle optimizations (the ACTIVE FTP one will only work
for uploads)
if [ "$MANGLE_TOS_OPTIMIZE" = "TRUE" ] ; then
	echo -n "Optimizing traffic: "
	${IPTABLES} -t mangle -A OUTPUT -p tcp --dport 23 -j TOS --set-tos
Minimize-Delay
	echo -n "telnet "
	${IPTABLES} -t mangle -A OUTPUT -p tcp --dport 22 -j TOS --set-tos
Minimize-Delay
	echo -n "ssh "
	${IPTABLES} -t mangle -A OUTPUT -p tcp --dport 20 -j TOS --set-tos
Minimize-Cost
	echo -n "ftp-data "
	${IPTABLES} -t mangle -A OUTPUT -p tcp --dport 21 -j TOS --set-tos
Minimize-Delay
	echo -n "ftp-control "
	${IPTABLES} -t mangle -A OUTPUT -p udp --dport 4000:7000 -j TOS --set-tos
Minimize-Delay
	echo -n "diablo2 "
	echo
fi

#What to do on those INET chains when we hit the end
echo -n "Setting up INET policies: "
#Drop if we cant find a valid inbound rule.
${IPTABLES} -t filter -A INETIN -j ${DROP}
echo -n "INETIN:${DROP} "
#We can send what we want to the internet
${IPTABLES} -t filter -A INETOUT -j ACCEPT
echo -n "INETOUT:ACCEPT "
echo

#All done!
echo "Done loading the firewall!"

Anyone have an idea?

David