1.\" Copyright (c) 2001, Matthew Dillon. Terms and conditions are those of 2.\" the BSD Copyright as specified in the file "/usr/src/COPYRIGHT" in 3.\" the source tree. 4.\" 5.\" $FreeBSD$ 6.\" 7.Dd May 26, 2001 8.Dt FIREWALL 7 9.Os 10.Sh NAME 11.Nm firewall 12.Nd simple firewalls under FreeBSD 13.Sh FIREWALL BASICS 14A Firewall is most commonly used to protect an internal network 15from an outside network by preventing the outside network from 16making arbitrary connections into the internal network. Firewalls 17are also used to prevent outside entities from spoofing internal 18IP addresses and to isolate services such as NFS or SMBFS (Windows 19file sharing) within LAN segments. 20.Pp 21The 22.Fx 23firewalling system also has the capability to limit bandwidth using 24.Xr dummynet 4 . 25This feature can be useful when you need to guarantee a certain 26amount of bandwidth for a critical purpose. For example, if you 27are doing video conferencing over the Internet via your 28office T1 (1.5 MBits), you may wish to bandwidth-limit all other 29T1 traffic to 1 MBit in order to reserve at least 0.5 MBits 30for your video conferencing connections. Similarly if you are 31running a popular web or ftp site from a colocation facility 32you might want to limit bandwidth to prevent excessive band 33width charges from your provider. 34.Pp 35Finally, 36.Fx 37firewalls may be used to divert packets or change the next-hop 38address for packets to help route them to the correct destination. 39Packet diversion is most often used to support NAT (network 40address translation), which allows an internal network using 41a private IP space to make connections to the outside for browsing 42or other purposes. 43.Pp 44Constructing a firewall may appear to be trivial, but most people 45get them wrong. The most common mistake is to create an exclusive 46firewall rather than an inclusive firewall. An exclusive firewall 47allows all packets through except for those matching a set of rules. 48An inclusive firewall allows only packets matching the ruleset 49through. Inclusive firewalls are much, much safer than exclusive 50firewalls but a tad more difficult to build properly. The 51second most common mistake is to blackhole everything except the 52particular port you want to let through. TCP/IP needs to be able 53to get certain types of ICMP errors to function properly - for 54example, to implement MTU discovery. Also, a number of common 55system daemons make reverse connections to the 56.Sy auth 57service in an attempt to authenticate the user making a connection. 58Auth is rather dangerous but the proper implementation is to return 59a TCP reset for the connection attempt rather than simply blackholing 60the packet. We cover these and other quirks involved with constructing 61a firewall in the sample firewall section below. 62.Sh IPFW KERNEL CONFIGURATION 63You do not need to create a custom kernel to use the IP firewalling features. 64If you enable firewalling in your 65.Em /etc/rc.conf 66(see below), the ipfw kernel module will be loaded automatically 67when necessary. 68However, 69if you are paranoid you can compile IPFW directly into the 70.Fx 71kernel by using the 72.Sy IPFIREWALL 73option set. If compiled in the kernel, ipfw denies all 74packets by default, which means that, if you do not load in 75a permissive ruleset via 76.Em /etc/rc.conf , 77rebooting into your new kernel will take the network offline. 78This can prevent you from being able to access your system if you 79are not sitting at the console. It is also quite common to 80update a kernel to a new release and reboot before updating 81the binaries. This can result in an incompatibility between 82the 83.Xr ipfw 8 84program and the kernel which prevents it from running in the 85boot sequence, also resulting in an inaccessible machine. 86Because of these problems the 87.Sy IPFIREWALL_DEFAULT_TO_ACCEPT 88kernel option is also available which changes the default firewall 89to pass through all packets. Note, however, that using this option 90may open a small window of opportunity during booting where your 91firewall passes all packets. Still, it's a good option to use 92while getting up to speed with 93.Fx 94firewalling. Get rid of it once you understand how it all works 95to close the loophole, though. There is a third option called 96.Sy IPDIVERT 97which allows you to use the firewall to divert packets to a user program 98and is necessary if you wish to use 99.Xr natd 8 100to give private internal networks access to the outside world. 101If you want to be able to limit the bandwidth used by certain types of 102traffic, the 103.Sy DUMMYNET 104option must be used to enable 105.Em ipfw pipe 106rules. 107.Sh SAMPLE IPFW-BASED FIREWALL 108Here is an example ipfw-based firewall taken from a machine with three 109interface cards. fxp0 is connected to the 'exposed' LAN. Machines 110on this LAN are dual-homed with both internal 10. IP addresses and 111Internet-routed IP addresses. In our example, 192.100.5.x represents 112the Internet-routed IP block while 10.x.x.x represents the internal 113networks. While it isn't relevant to the example, 10.0.1.x is 114assigned as the internal address block for the LAN on fxp0, 10.0.2.x 115for the LAN on fxp1, and 10.0.3.x for the LAN on fxp2. 116.Pp 117In this example we want to isolate all three LANs from the Internet 118as well as isolate them from each other, and we want to give all 119internal addresses access to the Internet through a NAT gateway running 120on this machine. To make the NAT gateway work, the firewall machine 121is given two Internet-exposed addresses on fxp0 in addition to an 122internal 10. address on fxp0: one exposed address (not shown) 123represents the machine's official address, and the second exposed 124address (192.100.5.5 in our example) represents the NAT gateway 125rendezvous IP. We make the example more complex by giving the machines 126on the exposed LAN internal 10.0.0.x addresses as well as exposed 127addresses. The idea here is that you can bind internal services 128to internal addresses even on exposed machines and still protect 129those services from the Internet. The only services you run on 130exposed IP addresses would be the ones you wish to expose to the 131Internet. 132.Pp 133It is important to note that the 10.0.0.x network in our example 134is not protected by our firewall. You must make sure that your 135Internet router protects this network from outside spoofing. 136Also, in our example, we pretty much give the exposed hosts free 137reign on our internal network when operating services through 138internal IP addresses (10.0.0.x). This is somewhat of security 139risk... what if an exposed host is compromised? To remove the 140risk and force everything coming in via LAN0 to go through 141the firewall, remove rules 01010 and 01011. 142.Pp 143Finally, note that the use of internal addresses represents a 144big piece of our firewall protection mechanism. With proper 145spoofing safeguards in place, nothing outside can directly 146access an internal (LAN1 or LAN2) host. 147.Bd -literal 148# /etc/rc.conf 149# 150firewall_enable="YES" 151firewall_type="/etc/ipfw.conf" 152 153# temporary port binding range let 154# through the firewall. 155# 156# NOTE: heavily loaded services running through the firewall may require 157# a larger port range for local-size binding. 4000-10000 or 4000-30000 158# might be a better choice. 159ip_portrange_first=4000 160ip_portrange_last=5000 161\&... 162.Ed 163.Pp 164.Bd -literal 165# /etc/ipfw.conf 166# 167# FIREWALL: the firewall machine / nat gateway 168# LAN0 10.0.0.X and 192.100.5.X (dual homed) 169# LAN1 10.0.1.X 170# LAN2 10.0.2.X 171# sw: ethernet switch (unmanaged) 172# 173# 192.100.5.x represents IP addresses exposed to the Internet 174# (i.e. Internet routeable). 10.x.x.x represent internal IPs 175# (not exposed) 176# 177# [LAN1] 178# ^ 179# | 180# FIREWALL -->[LAN2] 181# | 182# [LAN0] 183# | 184# +--> exposed host A 185# +--> exposed host B 186# +--> exposed host C 187# | 188# INTERNET (secondary firewall) 189# ROUTER 190# | 191# [Internet] 192# 193# NOT SHOWN: The INTERNET ROUTER must contain rules to disallow 194# all packets with source IP addresses in the 10. block in order 195# to protect the dual-homed 10.0.0.x block. Exposed hosts are 196# not otherwise protected in this example - they should only bind 197# exposed services to exposed IPs but can safely bind internal 198# services to internal IPs. 199# 200# The NAT gateway works by taking packets sent from internal 201# IP addresses to external IP addresses and routing them to natd, which 202# is listening on port 8668. This is handled by rule 00300. Data coming 203# back to natd from the outside world must also be routed to natd using 204# rule 00301. To make the example interesting, we note that we do 205# NOT have to run internal requests to exposed hosts through natd 206# (rule 00290) because those exposed hosts know about our 207# 10. network. This can reduce the load on natd. Also note that we 208# of course do not have to route internal<->internal traffic through 209# natd since those hosts know how to route our 10. internal network. 210# The natd command we run from /etc/rc.local is shown below. See 211# also the in-kernel version of natd, ipnat. 212# 213# natd -s -u -a 208.161.114.67 214# 215# 216add 00290 skipto 1000 ip from 10.0.0.0/8 to 192.100.5.0/24 217add 00300 divert 8668 ip from 10.0.0.0/8 to not 10.0.0.0/8 218add 00301 divert 8668 ip from not 10.0.0.0/8 to 192.100.5.5 219 220# Short cut the rules to avoid running high bandwidths through 221# the entire rule set. Allow established tcp connections through, 222# and shortcut all outgoing packets under the assumption that 223# we need only firewall incoming packets. 224# 225# Allowing established tcp connections through creates a small 226# hole but may be necessary to avoid overloading your firewall. 227# If you are worried, you can move the rule to after the spoof 228# checks. 229# 230add 01000 allow tcp from any to any established 231add 01001 allow all from any to any out via fxp0 232add 01001 allow all from any to any out via fxp1 233add 01001 allow all from any to any out via fxp2 234 235# Spoof protection. This depends on how well you trust your 236# internal networks. Packets received via fxp1 MUST come from 237# 10.0.1.x. Packets received via fxp2 MUST come from 10.0.2.x. 238# Packets received via fxp0 cannot come from the LAN1 or LAN2 239# blocks. We can't protect 10.0.0.x here, the Internet router 240# must do that for us. 241# 242add 01500 deny all from not 10.0.1.0/24 in via fxp1 243add 01500 deny all from not 10.0.2.0/24 in via fxp2 244add 01501 deny all from 10.0.1.0/24 in via fxp0 245add 01501 deny all from 10.0.2.0/24 in via fxp0 246 247# In this example rule set there are no restrictions between 248# internal hosts, even those on the exposed LAN (as long as 249# they use an internal IP address). This represents a 250# potential security hole (what if an exposed host is 251# compromised?). If you want full restrictions to apply 252# between the three LANs, firewalling them off from each 253# other for added security, remove these two rules. 254# 255# If you want to isolate LAN1 and LAN2, but still want 256# to give exposed hosts free reign with each other, get 257# rid of rule 01010 and keep rule 01011. 258# 259# (commented out, uncomment for less restrictive firewall) 260#add 01010 allow all from 10.0.0.0/8 to 10.0.0.0/8 261#add 01011 allow all from 192.100.5.0/24 to 192.100.5.0/24 262# 263 264# SPECIFIC SERVICES ALLOWED FROM SPECIFIC LANS 265# 266# If using a more restrictive firewall, allow specific LANs 267# access to specific services running on the firewall itself. 268# In this case we assume LAN1 needs access to filesharing running 269# on the firewall. If using a less restrictive firewall 270# (allowing rule 01010), you don't need these rules. 271# 272add 01012 allow tcp from 10.0.1.0/8 to 10.0.1.1 139 273add 01012 allow udp from 10.0.1.0/8 to 10.0.1.1 137,138 274 275# GENERAL SERVICES ALLOWED TO CROSS INTERNAL AND EXPOSED LANS 276# 277# We allow specific UDP services through: DNS lookups, ntalk, and ntp. 278# Note that internal services are protected by virtue of having 279# spoof-proof internal IP addresses (10. net), so these rules 280# really only apply to services bound to exposed IPs. We have 281# to allow UDP fragments or larger fragmented UDP packets will 282# not survive the firewall. 283# 284# If we want to expose high-numbered temporary service ports 285# for things like DNS lookup responses we can use a port range, 286# in this example 4000-65535, and we set to /etc/rc.conf variables 287# on all exposed machines to make sure they bind temporary ports 288# to the exposed port range (see rc.conf example above) 289# 290add 02000 allow udp from any to any 4000-65535,domain,ntalk,ntp 291add 02500 allow udp from any to any frag 292 293# Allow similar services for TCP. Again, these only apply to 294# services bound to exposed addresses. NOTE: we allow 'auth' 295# through but do not actually run an identd server on any exposed 296# port. This allows the machine being authed to respond with a 297# TCP RESET. Throwing the packet away would result in delays 298# when connecting to remote services that do reverse ident lookups. 299# 300# Note that we do not allow tcp fragments through, and that we do 301# not allow fragments in general (except for UDP fragments). We 302# expect the TCP mtu discovery protocol to work properly so there 303# should be no TCP fragments. 304# 305add 03000 allow tcp from any to any http,https 306add 03000 allow tcp from any to any 4000-65535,ssh,smtp,domain,ntalk 307add 03000 allow tcp from any to any auth,pop3,ftp,ftp-data 308 309# It is important to allow certain ICMP types through, here is a list 310# of general ICMP types. Note that it is important to let ICMP type 3 311# through. 312# 313# 0 Echo Reply 314# 3 Destination Unreachable (used by TCP MTU discovery, aka 315# packet-too-big) 316# 4 Source Quench (typically not allowed) 317# 5 Redirect (typically not allowed - can be dangerous!) 318# 8 Echo 319# 11 Time Exceeded 320# 12 Parameter Problem 321# 13 Timestamp 322# 14 Timestamp Reply 323# 324# Sometimes people need to allow ICMP REDIRECT packets, which is 325# type 5, but if you allow it make sure that your Internet router 326# disallows it. 327 328add 04000 allow icmp from any to any icmptypes 0,3,8,11,12,13,14 329 330# log any remaining fragments that get through. Might be useful, 331# otherwise don't bother. Have a final deny rule as a safety to 332# guarantee that your firewall is inclusive no matter how the kernel 333# is configured. 334# 335add 05000 deny log ip from any to any frag 336add 06000 deny all from any to any 337.Ed 338.Sh PORT BINDING INTERNAL AND EXTERNAL SERVICES 339We've mentioned multi-homing hosts and binding services to internal or 340external addresses but we haven't really explained it. When you have a 341host with multiple IP addresses assigned to it, you can bind services run 342on that host to specific IPs or interfaces rather than all IPs. Take 343the firewall machine for example: With three interfaces 344and two exposed IP addresses 345on one of those interfaces, the firewall machine is known by 5 different 346IP addresses (10.0.0.1, 10.0.1.1, 10.0.2.1, 192.100.5.5, and say 347192.100.5.1). If the firewall is providing file sharing services to the 348windows LAN segment (say it is LAN1), you can use samba's 'bind interfaces' 349directive to specifically bind it to just the LAN1 IP address. That 350way the file sharing services will not be made available to other LAN 351segments. The same goes for NFS. If LAN2 has your UNIX engineering 352workstations, you can tell nfsd to bind specifically to 10.0.2.1. You 353can specify how to bind virtually every service on the machine and you 354can use a light 355.Xr jail 8 356to indirectly bind services that do not otherwise give you the option. 357.Sh SEE ALSO 358.Xr ipnat 1 , 359.Xr dummynet 4 , 360.Xr ipnat 5 , 361.Xr rc.conf 5 , 362.Xr smb.conf 5 [ /usr/ports/net/samba ] , 363.Xr samba 7 [ /usr/ports/net/samba ] , 364.Xr config 8 , 365.Xr ipfw 8 , 366.Xr jail 8 , 367.Xr natd 8 , 368.Xr nfsd 8 369.Sh ADDITIONAL READING 370.Xr ipf 5 , 371.Xr ipf 8 , 372.Xr ipfstat 8 373.Sh HISTORY 374The 375.Nm 376manual page was originally written by 377.An Matthew Dillon 378and first appeared 379in 380.Fx 4.3 , 381May 2001. 382