shorewall6.conf (5) - Linux Manuals

shorewall6.conf: Shorewall6 global configuration file


shorewall6.conf - Shorewall6 global configuration file




This file sets options that apply to Shorewall6 as a whole.

The file consists of Shell comments (lines beginning with '#'), blank lines and assignment statements (variable=value). If the value contains shell meta characters or white-space, then it must be enclosed in quotes. Example: LOG_LEVEL="NFLOG(1,0,1)".


Many options have as their value a log-level. Log levels are a method of describing to syslog (8) the importance of a message and a number of parameters in this file have log levels as their value.

These levels are defined by syslog and are used to determine the destination of the messages through entries in /etc/syslog.conf (5). The syslog documentation refers to these as "priorities"; Netfilter calls them "levels" and Shorewall6 also uses that term.

Valid levels are:

       7       debug
       6       info
       5       notice
       4       warning
       3       err
       2       crit
       1       alert
       0       emerg

For most Shorewall6 logging, a level of 6 (info) is appropriate. Shorewall6 log messages are generated by NetFilter and are logged using facility 'kern' and the level that you specify. If you are unsure of the level to choose, 6 (info) is a safe bet. You may specify levels by name or by number.

If you have built your kernel with NFLOG target support, you may also specify a log level of NFLOG (must be all caps). Rather than log its messages to syslogd, Shorewall6 will direct netfilter to log the messages via the NFLOG target which will send them to a process called 'ulogd'. ulogd is available with most Linux distributions (although it probably isn't installed by default). Ulogd is also available from m[blue][] and can be configured to log all Shorewall6 message to their own log file


If you want to specify parameters to ULOG or NFLOG (e.g., NFLOG(1,0,1)), then you must quote the setting.



Beginning with Shorewall 5.0.0, the log level may be followed by a colon (":") and a log tag. The log tag normally follows the packet disposition in Shorewall-generated Netfilter log messages, separated from the disposition by a colon (e.g, "DROP:mytag"). See LOGTAGONLY below for additional information.

The following options may be set in shorewall6.conf.






DROP_DEFAULT describes the rules to be applied before a connection request is dropped by a DROP policy; REJECT_DEFAULT describes the rules to be applied if a connection request is rejected by a REJECT policy. The other three are similar for ACCEPT, QUEUE and NFQUEUE policies.

The value applied to these may be:

a) The name of an action. The
      name may optionally be followed by a comma-separated list of
      parameters enclosed in parentheses if the specified action accepts
      parameters (e.g., 'Drop(audit)').
c) None or none
The default values are:
If you set the value of either option to "None" then no default action will be used and the default action or macro must be specified in m[blue]shorewall6-policym[][1](5).

You can pass parameters to the specified action or macro (e.g., myaction(audit,DROP)).

Beginning with Shorewall 4.5.10, the action name can be followed optionally by a colon and a log level. The level will be applied to each rule in the action or macro body that does not already have a log level.


Added in Shorewall 4.4.7. If set to Yes, Shorewall6 accounting is enabled (see m[blue]shorewall6-accountingm[][2](5)). If not specified or set to the empty value, ACCOUNTING=Yes is assumed.


Added in Shorewall 4.4.20. This setting determines which Netfilter table the accounting rules are added in. By default, ACCOUNTING_TABLE=filter is assumed. See also m[blue]shorewall6-accountingm[][2](5).


The value of this variable affects Shorewall's stopped state. The behavior differs depending on whether m[blue]shorewall-routestoppedm[][3](5) or m[blue]shorewall-stoppedrulesm[][4](5) is used:


When ADMINISABSENTMINDED=No, only traffic to/from those addresses listed in routestopped is accepted when Shorewall is stopped. When ADMINISABSENTMINDED=Yes, in addition to traffic to/from addresses in routestopped, connections that were active when Shorewall stopped continue to work and all new connections from the firewall system itself are allowed.

Note that the routestopped file is not supported in Shorewall 5.0 and later versions.


All existing connections continue to work. To sever all existing connections when the firewall is stopped, install the conntrack utility and place the command conntrack -F in the stopped user exit (/etc/shorewall6/stopped).

If ADMINISABSENTMINDED=No, only new connections matching entries in stoppedrules are accepted when Shorewall is stopped. Response packets and related connections are automatically accepted.

If ADMINISABSENTMINDED=Yes, in addition to connections matching entries in stoppedrules, all new connections from the firewall system itself are allowed when the firewall is stopped. Response packets and related connections are automatically accepted.

If this variable is not set or is given the empty value then ADMINISABSENTMINDED=No is assumed.


Added in Shorewall 4.5.11. Normally, if an unknown shell variable is encountered in a configuration file (except in ?IF and ?ELSIF directives), the compiler raises a fatal error. If IGNOREUNKNOWNVARIABLES is set to Yes, then such variables simply expand to an empty string. Default is No.


Formerly named AUTO_COMMENT. If set, if there is not a current comment when a macro is invoked, the behavior is as if the first line of the macro file was "COMMENT <macro name>". The AUTO_COMMENT option has a default value of 'Yes'.


Added in Shorewall 4.5.7. When set to Yes (the default), the generated ruleset will automatically associate helpers with applications that require them (FTP, IRC, etc.). When configuring your firewall on systems running kernel 3.5 or later, it is recommended that you:


2. Modify the HELPERS setting (see below) to list the helpers that you need.

3. Either:

1. Modify m[blue]shorewall6-conntrackm[][5] (5) to only apply helpers where they are required; or

2. Specify the appropriate helper in the HELPER column in m[blue]shorewall6-rulesm[][6] (5).

The macros for those applications requiring a helper automatically specify the appropriate HELPER where required.


If set, the behavior of the start, reload and restart commands is changed; if no files in the CONFIG_PATH (see below) have been changed since the last successful start, reload or restart command, then the compilation step is skipped and the compiled script that executed the last start, reload or restart command is used. The default is AUTOMAKE=No.

The setting of the AUTOMAKE option is ignored if the start, reload or restart command includes a directory name (e.g., shorewall6 restart /etc/


Added in Shorewall-4.6.0. When set to Yes, causes entries in m[blue]shorewall6-tcfilters(5)m[][7] to generate a basic filter rather than a u32 filter. This setting requires the Basic Ematch capability in your kernel and iptables.

One of the advantages of basic filters is that ipset matches are supported in newer iproute2 and kernel versions. Because Shorewall6 cannot reliably detect this capability, use of basic filters is controlled by this option.

The default value is No which causes u32 filters to be generated.



Added in Shorewall 4.5.13 to replace the BLACKLISTNEWONLY option. Specifies the connection tracking states that are to be subject to blacklist screening. If BLACKLIST is not specified then the states subject to blacklisting are NEW,ESTABLISHED,INVALID,UNTRACKED.

Note: The ESTABLISHED state may not be specified if FASTACCEPT is specified.


This parameter determines the disposition of packets from blacklisted hosts. It may have the value DROP if the packets are to be dropped or REJECT if the packets are to be replied with an ICMP port unreachable reply or a TCP RST (tcp only). If you do not assign a value or if you assign an empty value then DROP is assumed. The setting determines the disposition of packets sent to the blacklog target of m[blue]shorewall6-blrulesm[][8](5).


Formerly named BLACKLIST_LOGLEVEL. This parameter determines if packets from blacklisted hosts are logged and it determines the syslog level that they are to be logged at. Its value is a syslog level (Example: BLACKLIST_LOG_LEVEL=debug). If you do not assign a value or if you assign an empty value then packets from blacklisted hosts are not logged. The setting determines the log level of packets sent to the blacklog target of m[blue]shorewall6-blrulesm[][8](5).


Added in Shorewall 4.5.16. Prior to the availability of BEGIN PERL....END PERL in configuration files, the only way to execute a chain-specific script was to create a script file with the same name as the chain and place it in a directory on the CONFIG_PATH. That facility has the drawback that the compiler will attempt to run a non-script file just because it has the same name as a chain. To disable this facility, set CHAIN_SCRIPTS=No. If not specified or specified as the empty value, CHAIN_SCRIPTS=Yes is assumed.


This parameter enables the TCP Clamp MSS to PMTU feature of Netfilter and is usually required when your internet connection is through PPPoE or PPTP. If set to Yes or yes, the feature is enabled. If left blank or set to No or no, the feature is not enabled.

Important: This option requires CONFIG_IP_NF_TARGET_TCPMSS in your kernel.

You may also set CLAMPMSS to a numeric value (e.g., CLAMPMSS=1400). This will set the MSS field in TCP SYN packets going through the firewall to the value that you specify.


If this option is set to No then Shorewall6 won't clear the current traffic control rules during [re]start or reload. This setting is intended for use by people that prefer to configure traffic shaping when the network interfaces come up rather than when the firewall is started. If that is what you want to do, set TC_ENABLED=Yes and CLEAR_TC=No and do not supply an /etc/shorewall6/tcstart file. That way, your traffic shaping rules can still use the "fwmark" classifier based on packet marking defined in m[blue]shorewall6-tcrulesm[][9](5). If not specified, CLEAR_TC=No is assumed.

If you also run Shorewall and if you have TC_ENABLED=Internal in your m[blue]shorewall-confm[][10](5), then you will want CLEAR_TC=No in this file.


Added in Shorewall6 4.4.12. When you set this option to Yes, you are asserting that the configuration is complete so that your set of zones encompasses any hosts that can send or receive traffic to/from/through the firewall. This causes Shorewall6 to omit the rules that catch packets in which the source or destination IP address is outside of any of your zones. Default is No. It is recommended that this option only be set to Yes if:

• You have defined an interface whose effective physical setting is '+'.

• That interface is assigned to a zone.

• You have no CONTINUE policies or rules.


Specifies where configuration files other than shorewall6.conf may be found. CONFIG_PATH is specifies as a list of directory names separated by colons (":"). When looking for a configuration file:

• If the command is "try" or a "<configuration directory>" was specified in the command (e.g., shorewall6 check ./gateway) then the directory given in the command is searched first.

• Next, each directory in the CONFIG_PATH setting is searched in sequence.

If CONFIG_PATH is not given or if it is set to the empty value then the contents of /usr/share/shorewall6/configpath are used. As released from, that file sets the CONFIG_PATH to /etc/shorewall6:/usr/share/shorewall6:/usr/share/shorewall but your particular distribution may set it differently. See the output of shorewall6 show config for the default on your system.


Added in Shorewall 4.5.12. When set to 'Yes' (the default), DNS names are validated in the compiler and then passed on to the generated script where they are resolved by ip6tables-restore. This is an advantage if you use AUTOMAKE=Yes and the IP address associated with the DNS name is subject to change. When DEFER_DNS_RESOLUTION=No, DNS names are converted into IP addresses by the compiler. This has the advantage that when AUTOMAKE=Yes the start, reload and restart commands will succeed even if no DNS server is reachable (assuming that the configuration hasn't changed since the compiled script was last generated).


If set to Yes (the default value), entries in the /etc/shorewall6/rtrules file cause an 'ip rule del' command to be generated in addition to an 'ip rule add' command. Setting this option to No, causes the 'ip rule del' command to be omitted.


Causes Shorewall6 to not load the listed kernel modules.


Added in Shorewall 4.4.7. When set to No or no, chain-based dynamic blacklisting using shorewall6 drop, shorewall6 reject, shorewall6 logdrop and shorewall6 logreject is disabled. Default is Yes. Beginning with Shorewall 5.0.8, ipset-based dynamic blacklisting using shorewall6 blacklist is also supported. The name of the set (setname) and the level (log_level), if any, at which blacklisted traffic is to be logged may also be specified. The default set name is SW_DBL6 and the default log level is none (no logging). if ipset-only is given, then chain-based dynamic blacklisting is disabled just as if DYNAMIC_BLACKLISTING=No had been specified.

Possible options are:


Normally, only packets whose source address matches an entry in the ipset are dropped. If src-dst is included, then packets whose destination address matches an entry in the ipset are also dropped.


The disconnect option was added in Shorewall 5.0.13 and requires that the conntrack utility be installed on the firewall system. When an address is blacklisted using the blacklist command, all connections originating from that address are disconnected. if the src-dst option was also specified, then all connections to that address are also disconnected.


Added in Shorewall 5.0.13. Normally, Shorewall creates the dynamic blacklisting ipset with timeout 0 which means that entries are permanent. If you want entries in the set that are not accessed for a period of time to be deleted from the set, you may specify that period using this option. Note that the blacklist command can override the ipset's timeout setting.

Once the dynamic blacklisting ipset has been created, changing this option setting requires a complete restart of the firewall; shorewall6 restart if RESTART=restart, otherwise shorewall6 stop && shorewall6 start

When ipset-based dynamic blacklisting is enabled, the contents of the blacklist will be preserved over stop/reboot/start sequences if SAVE_IPSETS=Yes or if setname is included in the list of sets to be saved in SAVE_IPSETS.


Normally, when the SOURCE or DEST columns in shorewall-policy(5) contains 'all', a single policy chain is created and the policy is enforced in that chain. For example, if the policy entry is

#                   LEVEL
net     all  DROP   info

then the chain name is 'net-all' ("net2all" if ZONE2ZONE=2) which is also the chain named in Shorewall log messages generated as a result of the policy. If EXPAND_POLICIES=Yes, then Shorewall will create a separate chain for each pair of zones covered by the policy. This makes the resulting log messages easier to interpret since the chain in the messages will have a name of the form 'a2b' where 'a' is the SOURCE zone and 'b' is the DEST zone.


Added in Shorewall 4.4.17. When set to Yes when compiling for use by Shorewall6 Lite (shorewall6 load, shorewall6 reload or shorewall6 export commands), the compiler will copy the modules or helpers file from the administrative system into the script. When set to No or not specified, the compiler will not copy the modules or helpers file from /usr/share/shorewall6 but will copy those found in another location on the CONFIG_PATH.

When compiling for direct use by Shorewall6, causes the contents of the local module or helpers file to be copied into the compiled script. When set to No or not set, the compiled script reads the file itself.


Normally, Shorewall6 defers accepting ESTABLISHED/RELATED packets until these packets reach the chain in which the original connection was accepted. So for packets going from the 'loc' zone to the 'net' zone, ESTABLISHED/RELATED packets are ACCEPTED in the 'loc-net' or 'loc2net' chain, depending on the setting of ZONE2ZONE (see below).

If you set FASTACCEPT=Yes, then ESTABLISHED/RELATED packets are accepted early in the INPUT, FORWARD and OUTPUT chains. If you set FASTACCEPT=Yes then you may not include rules in the ESTABLISHED or RELATED sections of m[blue]shorewall6-rulesm[][6](5).


This option was added in Shorewall 5.0.13 and may be used on an administrative system in directories containing the configurations of remote firewalls. The contents of the variable are the default value for the system parameter to the remote-start, remote-reload and remote-restart commands.


Added in Shorewall 4.4.11. Traditionally, Shorewall has cleared the packet mark in the first rule in the mangle FORWARD chain. This behavior is maintained with the default setting of this option (FORWARD_CLEAR_MARK=Yes). If FORWARD_CLEAR_MARK is set to 'No', packet marks set in the mangle PREROUTING chain are retained in the FORWARD chains.


Added in Shorewall 4.5.4. Specifies the pathname of the directory containing the GeoIP Match database. See m[blue][][11]. If not specified, the default value is /usr/share/xt_geoip/LE which is the default location of the little-endian database.


Added in Shorewall 4.5.7. This option specifies a comma-separated list naming the Netfilter application helpers that are to be enabled. If not specified, the default is to enable all helpers.

Possible values for helper are:






none - This special value was added in Shorewall 4.5.16 and indicates that no helpers are to be enabled. It also prevents the compiler for probing for helper support; such probing generates messages on the system log of the form "xt_CT: No such helper XXX" where XXX is the helper name. When used, none must be the only helper specified.






When HELPERS is specified on a system running Kernel 3.5.0 or later, automatic association of helpers to connections is disabled.


When this option is set to Yes, it causes subzones to be treated differently with respect to policies.

Subzones are defined by following their name with ":" and a list of parent zones (in m[blue]shorewall6-zonesm[][12](5)). Normally, you want to have a set of special rules for the subzone and if a connection doesn't match any of those subzone-specific rules then you want the parent zone rules and policies to be applied; see m[blue]shorewall6-nestingm[][13](5). With IMPLICIT_CONTINUE=Yes, that happens automatically.

If IMPLICIT_CONTINUE=No or if IMPLICIT_CONTINUE is not set, then subzones are not subject to this special treatment. With IMPLICIT_CONTINUE=Yes, an implicit CONTINUE policy may be overridden by including an explicit policy (one that does not specify "all" in either the SOURCE or the DEST columns).


Added in Shorewall 4.6.0. Traditionally in m[blue]shorewall6-rulesm[][6](5), a semicolon separates column-oriented specifications on the left from m[blue]alternative specificaitonsm[][14] on the right.. When INLINE_MATCHES=Yes is specified, the specifications on the right are interpreted as if INLINE had been specified in the ACTION column. This also applies to m[blue]shorewall6-masq(5)m[][15] and m[blue]shorewall6-mangle(5m[][16]) which also support INLINE. If not specified or if specified as the empty value, the value 'No' is assumed for backward compatibility.

Beginning with Shorewall 5.0.0, it is no longer necessary to set INLINE_MATCHES=Yes in order to be able to specify your own iptables text in a rule and INLINE_MATCHES=Yes is deprecated. Beginning with 5.0.0, you may simply preface your text with a pair of semicolons (";;"). If alternate input is also specified in the rule, it should appear before the semicolons and may be separated from normal column input by a single semicolon or enclosed in curly braces ("{....}").


Added in Shorewall 4.5.13. Shorewall has traditionally passed INVALID packets through the NEW section of m[blue]shorewall-rulesm[][6] (5). When a packet in INVALID state fails to match any rule in the INVALID section, the packet is disposed of based on this setting. The default value is CONTINUE for compatibility with earlier versions.


Added in Shorewall 4.5.13. Packets in the INVALID state that do not match any rule in the INVALID section of m[blue]shorewall6-rulesm[][6] (5) are logged at this level. The default value is empty which means no logging is performed.


If specified, gives the pathname of the 'ip' executable. If not specified, 'ip' is assumed and the utility will be located using the current PATH setting.


This rather useless parameter determines whether Shorewall6 enables or disables IPV6 Packet Forwarding on all interfaces (/proc/sys/net/ipv6/config/all/forwarding). Possible values are:

On or on

packet forwarding will be enabled.

Off or off

packet forwarding will be disabled.

Keep or keep

Shorewall6 will neither enable nor disable packet forwarding

If this variable is not set or is given an empty value (IP_FORWARD="") then IP_FORWARD=On is assumed.


This parameter names the ip6tables executable to be used by Shorewall6. If not specified or if specified as a null value, then the ip6tables executable located using the PATH option is used.

Regardless of how the ip6tables utility is located (specified via IP6TABLES= or located via PATH), Shorewall6 uses the ip6tables-restore and ip6tables-save utilities from that same directory.


If specified, gives the pathname of the 'ipset' executable. If not specified, 'ipset' is assumed and the utility will be located using the current PATH setting.


Added in Shorewall 4.5.2. Default is Yes. When set, causes the rules compiler to issue a warning when:

• The compiler is being run by root and an ipset specified in the configuration does not exists. Only one warning is issued for each missing ipset.

• When [src] is specified in a destination column and when [dst] is specified in a source column.


When set to Yes, this option prevents scripts generated by Shorewall6 from altering the /etc/iproute2/rt_tables database when there are entries in /etc/shorewall6/providers. If you set this option to Yes while Shorewall6 (Shorewall6-lite) is running, you should remove the file /var/lib/shorewall6/rt_tables (/var/lib/shorewall6-lite/rt_tables) before your next stop, refresh, restore, reload or restart command.

The default is KEEP_RT_TABLES=No.


Added in Shorewall 4.4.7. When set to Yes, restricts the set of modules loaded by shorewall to those listed in /var/lib/shorewall6/helpers and those that are actually used. When not set, or set to the empty value, LOAD_HELPERS_ONLY=No is assumed.


Specifies the name of the Shorewall6 lock file, used to prevent simultaneous state-changing commands. If not specified, ${VARDIR}/shorewall6/lock is assumed (${VARDIR} is normally /var/lib but can be changed when Shorewall-core is installed -- see the output of shorewall6 show vardir).


Added in Shorewall 4.6.4. LOG_BACKEND determines the logging backend to be used for the iptrace command (see m[blue]shorewall6(8)m[][17]).

backend is one of:


Use standard kernel logging.


Use netlink logging to ulogd version 2 or later.


This option controls the amount of information logged to the file specified in the STARTUP_LOG option.

Values are:

-1 - Logging is disabled
0 - Silent. Only error messages are logged.
1 - Major progress messages logged.
2 - All progress messages logged
If not specified, then -1 is assumed.


This option is intended for use as a debugging aid. When set to a log level, this option causes Shorewall6 to generate a logging rule as the first rule in each builtin chain.

• The table name is used as the chain name in the log prefix.

• The chain name is used as the target in the log prefix.

For example, using the default LOGFORMAT, the log prefix for logging from the nat table's PREROUTING chain is:


To help insure that all packets in the NEW state are logged, rate limiting (LOGLIMIT) should be disabled when using LOGALLNEW. Use LOGALLNEW at your own risk; it may cause high CPU and disk utilization and you may not be able to control your firewall after you enable this option.

Do not use this option if the resulting log messages will be sent to another system.


This parameter tells the /sbin/shorewall6 program where to look for Shorewall6 messages when processing the dump, logwatch, show log, and hits commands. If not assigned or if assigned an empty value, /var/log/messages is assumed. Beginning with Shorewall, you may specify systemd to use journelctl -r to read the log.


The value of this variable generate the --log-prefix setting for Shorewall6 logging rules. It contains a "printf" formatting template which accepts three arguments (the chain name, logging rule number (optional) and the disposition). To use LOGFORMAT with fireparse, set it as:

    LOGFORMAT="fp=%s:%d a=%s "

If the LOGFORMAT value contains the substring "%d" then the logging rule number is calculated and formatted in that position; if that substring is not included then the rule number is not included. If not supplied or supplied as empty (LOGFORMAT="") then "Shorewall6:%s:%s:" is assumed.

The setting of LOGFORMAT has an effect of the permitted length of zone names. See m[blue]shorewall6-zonesm[][12] (5).


Added in Shorewall 4.4.12. Limits the logging rate, either overall, or by source or destination IP address.

If the value starts with 's:' then logging is limited per source IP. If the value starts with 'd:', then logging is limited per destination IP. Otherwise, the overall logging rate is limited.

If burst is not specified, then a value of 5 is assumed.

The keywords second and minute are accepted beginning with Shorewall 4.6.13.


Using the default LOGFORMAT, chain names may not exceed 11 characters or truncation of the log prefix may occur. Longer chain names may be used with log tags if you set LOGTAGONLY=Yes. With LOGTAGONLY=Yes, if a log tag is specified then the tag is included in the log prefix in place of the chain name.

Beginning with Shorewall 4.5.12, when LOGTAGONLY=Yes, you have more control over the generated log prefix. Beginning with that release, the tag is interpreted as a chain name and a disposition separated by a comma. So this rule:

#ACTION                                SOURCE         DEST
LOG:info:foo,bar                 net                   fw

would generate the following log prefix when using the default LOGFORMAT setting:


#ACTION                               SOURCE            DEST
LOG:info:,bar                        net                    fw

would generate



Determines the disposition of connections requests that fail MAC Verification and must have the value ACCEPT (accept the connection request anyway), REJECT (reject the connection request) or DROP (ignore the connection request). If not set or if set to the empty value (e.g., MACLIST_DISPOSITION="") then MACLIST_DISPOSITION=REJECT is assumed.

A_DROP and A_REJECT are audited versions of DROP and REJECT respectively and were added in Shorewall 4.4.20. They require AUDIT_TARGET in the kernel and ip6tables.


Determines the syslog level for logging connection requests that fail MAC Verification. The value must be a valid syslogd log level. If you don't want to log these connection requests, set to the empty value (e.g., MACLIST_LOG_LEVEL="").


Normally, MAC verification occurs in the filter table (INPUT and FORWARD) chains. When forwarding a packet from an interface with MAC verification to a bridge interface, that doesn't work.

This problem can be worked around by setting MACLIST_TABLE=mangle which will cause Mac verification to occur out of the PREROUTING chain. Because REJECT isn't available in that environment, you may not specify MACLIST_DISPOSITION=REJECT with MACLIST_TABLE=mangle.


The performance of configurations with a large numbers of entries in m[blue]shorewall6-maclistm[][18](5) can be improved by setting the MACLIST_TTL variable in m[blue]shorewall6.confm[][19](5).

If your iptables and kernel support the "Recent Match" (see the output of "shorewall check" near the top), you can cache the results of a 'maclist' file lookup and thus reduce the overhead associated with MAC Verification.

When a new connection arrives from a 'maclist' interface, the packet passes through then list of entries for that interface in m[blue]shorewall6-maclistm[][18](5). If there is a match then the source IP address is added to the 'Recent' set for that interface. Subsequent connection attempts from that IP address occurring within $MACLIST_TTL seconds will be accepted without having to scan all of the entries. After $MACLIST_TTL from the first accepted connection request from an IP address, the next connection request from that IP address will be checked against the entire list.

If MACLIST_TTL is not specified or is specified as empty (e.g, MACLIST_TTL="" or is specified as zero then 'maclist' lookups will not be cached).


Determines whether Shorewall will generate rules in the Netfilter mangle table. Setting MANGLE_ENABLED=No disables all Shorewall features that require the mangle table. The default is MANGLE_ENABLED=Yes.


If your kernel has a FORWARD chain in the mangle table, you may set MARK_IN_FORWARD_CHAIN=Yes to cause the marking specified in the tcrules file to occur in that chain rather than in the PREROUTING chain. This permits you to mark inbound traffic based on its destination address when DNAT is in use. To determine if your kernel has a FORWARD chain in the mangle table, use the /shorewall6 show mangle command; if a FORWARD chain is displayed then your kernel will support this option. If this option is not specified or if it is given the empty value (e.g., MARK_IN_FORWARD_CHAIN="") then MARK_IN_FORWARD_CHAIN=No is assumed.


Added in Shorewall 4.4.26. Number of bits on the right of the 32-bit packet mark to be masked when clearing the traffic shaping mark. Must be >= TC_BITS and <= PROVIDER_OFFSET (if PROVIDER_OFFSET > 0). Prior to Shorewall 5.0.0, default value and the default values of the other mark layout options is determined as follows:

Table 1. Default Packet Mark Layout


From 5.0.0 onward, the default value of MASK_BITS is 8, the default value of PROVIDER_BITS, TC_BITS, MASK_BITS and PROVIDER_OFFSET is 8.

MODULE_SUFFIX=["extension ..."]

The value of this option determines the possible file extensions of kernel modules. The default value is "ko ko.gz ko.xz o o.gz o.xz gz xz".


This parameter specifies the directory/directories where your kernel netfilter modules may be found. If you leave the variable empty, Shorewall will supply the value "/lib/modules/$uname/kernel/net/ipv${g_family}/netfilter:/lib/modules/$uname/kernel/net/netfilter:/lib/modules/$uname/kernel/net/sched:/lib/modules/$uname/extra:/lib/modules/$uname/extra/ipset" where uname holds the output of 'uname -r' and g_family holds '6'.

The option plus sign ('+') was added in Shorewall 5.0.3 and causes the listed pathnames to be appended to the default list above.


The value of this variable determines the number of seconds that programs will wait for exclusive access to the Shorewall6 lock file. After the number of seconds corresponding to the value of this variable, programs will assume that the last program to hold the lock died without releasing the lock.

If not set or set to the empty value, a value of 60 (60 seconds) is assumed.

An appropriate value for this parameter would be twice the length of time that it takes your firewall system to process a shorewall6 restart command.


Added in Shorewall 4.5.7. Specifies the pathname of the nfacct utility. If not specified, Shorewall will use the PATH setting to find the program.


The specified value enables certain optimizations. Each optimization category is associated with a power of two. To enable multiple optimization categories, simply add their corresponding numbers together.

Beginning with Shorewall 4.5.20, you may specify OPTIMIZE=All to enable all optimization categories, and you may also specify OPTIMIZE=None to disable optimization.

• Optimization category 1 - Traditionally, Shorewall has created rules for m[blue]the complete matrix of host groups defined by the zones, interfaces and hosts filesm[][20]. Any traffic that didn't correspond to an element of that matrix was rejected in one of the built-in chains. When the matrix is sparse, this results in lots of largely useless rules.

These extra rules can be eliminated by setting the 1 bit in OPTIMIZE.

The 1 bit setting also controls the suppression of redundant wildcard rules (those specifying "all" in the SOURCE or DEST column). A wildcard rule is considered to be redundant when it has the same ACTION and Log Level as the applicable policy.

Optimization level 1 is ignored when optimization level 4 is also selected, since level 4 performs similar optimizations in a more robust way.

• Optimization category 2 - Added in Shorewall 4.4.7. When set, suppresses superfluous ACCEPT rules in a policy chain that implements an ACCEPT policy. Any ACCEPT rules that immediately precede the final blanket ACCEPT rule in the chain are now omitted.

• Optimization category 4 - Added in Shorewall 4.4.7. When set, causes short chains (those with less than 2 rules) to be optimized away. The following chains are excluded from optimization:

• accounting chains (unless OPTIMIZE_ACCOUNTING=Yes)

• action chains (user-defined)

• 'blacklst' chain

• dynamic


• If a built-in chain has a single rule that branches to a second chain, then the rules from the second chain are moved to the built-in chain and the target chain is omitted.

• Chains with no references are deleted.

• Accounting chains are subject to optimization if the OPTIMIZE_ACCOUNTING option is set to 'Yes'.

• If a chain ends with an unconditional branch to a second chain (other than to 'reject'), then the branch is deleted from the first chain and the rules from the second chain are appended to it.

An additional optimization was added in Shorewall 4.5.4. If the last rule in a chain is an unqualified jump to a simple target, then all immediately preceding rules with the same simple target are omitted.

For example, consider this chain:

        -A fw-net -p udp --dport 67:68 -j ACCEPT
        -A fw-net -p udp --sport 1194 -j ACCEPT
        -A fw-net -p 41 -j ACCEPT
        -A fw-net -j ACCEPT

Since all of the rules are jumps to the simple target ACCEPT, this chain is totally optimized away and jumps to the chain are replace with jumps to ACCEPT.

• Optimization category 8 - Added in Shorewall 4.4.9. When set, causes chains with identical rules to be collapsed into a single chain.

• Optimization category 16 - Added in Shorewall 4.4.26. When set, causes sequences of compatible rules to be combined into a single rule. Rules are considered compatible if they differ only in their destination ports and comments.

A sequence of compatible rules is often generated when macros are invoked in sequence.

The ability to combine adjacent rules is limited by two factors:

• Destination port lists may only be combined up to a maximum of 15 ports, where a port-pair counts as two ports.

• Rules may only be combined until the length of their concatenated comment reaches 255 characters.

When either of these limits would be exceeded, the current combined rule is emitted and the compiler attempts to combine rules beginning with the one that would have exceeded the limit. Adjacent combined comments are separated by ', '. Empty comments at the front of a group of combined comments are replaced by 'Others and'. Empty comments at the end of a group of combined comments are replaced by 'and others'.

Beginning in Shorewall 4.5.10, this option also suppresses duplicate adjacent rules and duplicate non-adjacent rules that don't include mark, connmark, dscp, ecn, set, tos or u32 matches.

Example 1:

Rules with comments "FOO", <empty> and "BAR" would result in the combined comment "FOO and others, BAR".

Example 2:

Rules with comments <empty>, "FOO" and "BAR" would result in the combined comment "Others and FOO, BAR". Note: Optimize level 16 requires "Extended Multi-port Match" in your iptables and kernel.

The default value is zero which disables all optimizations.


Added in Shorewall 4.4.7. If set to Yes, Shorewall accounting changes are subject to optimization (OPTIMIZE=4,5,6 or 7). If not specified or set to the empty value, OPTIMIZE_ACCOUNTING=No is assumed.


Added in Shorewall 5.0.6. Specifies a path name of a pager program like less or more. When PAGER is given, the output of verbose status commands and the dump command are piped through the named program when the output file is a terminal.

Beginning with Shorewall 5.0.12, the default value of this option is the DEFAULT_PAGER setting in shorewallrc.


Determines the order in which Shorewall6 searches directories for executable files.


Added in Shorewall 4.4.11 RC1. Specifies the path name of the Perl executable. Default is /usr/bin/perl. If the pathname specified by this option does not exist or the named file is not executable, then Shorewall6 falls back to /usr/bin/perl


Added in Shorewall 4.4.26. The number of bits in the 32-bit packet mark to be used for provider numbers. May be zero. See MASK_BITS above for default value.


Added in Shorewall 4.4.26. The offset from the right (low-order end) of the provider number field in the 32-bit packet mark. If non-zero, must be >= TC_BITS (Shorewall automatically adjusts PROVIDER_OFFSET's value). PROVIDER_OFFSET + PROVIDER_BITS + ZONE_BITS must be < 32. See MASK_BITS above for default value.



Earlier generations of Shorewall6 Lite required that remote root login via ssh be enabled in order to use the load and reload commands. Beginning with release 3.9.5, you may define an alternative means for accessing the remote firewall system. In that release, two new options were added to shorewall6.conf:.RS 4 RSH_COMMAND
The default values for these are as follows:.RS 4 RSH_COMMAND: ssh ${root}@${system} ${command}
RCP_COMMAND: scp ${files}
Shell variables that will be set when the commands are invoked are as follows:.RS 4 root - root user. Normally
        root but may be overridden using the '-r'
system - The name/IP address
        of the remote firewall system.
command - For RSH_COMMAND,
        the command to be executed on the firewall system.
files - For RCP_COMMAND, a
        space-separated list of files to be copied to the remote
        firewall system.
destination - The directory
        on the remote system that the files are to be copied


Added in Shorewall 4.4.27. Shorewall has traditionally ACCEPTed RELATED packets that don't match any rule in the RELATED section of m[blue]shorewall6-rulesm[][6] (5). Concern about the safety of this practice resulted in the addition of this option. When a packet in RELATED state fails to match any rule in the RELATED section, the packet is disposed of based on this setting. The default value is ACCEPT for compatibility with earlier versions.


Added in Shorewall 4.4.27. Packets in the related state that do not match any rule in the RELATED section of m[blue]shorewall6-rulesm[][6] (5) are logged at this level. The default value is empty which means no logging is performed.


Added in Shorewall 4.5.21. When a REJECT target is specified, Shorewall normally handles the response as follows:

• If the destination address of the packet is a broadcast or multicast address, the packet is dropped.

• if the protocol is ICMP (58) then the packet is dropped.

• if the protocol is TCP (6) then the packet is rejected with an RST.

• if the protocol is UDP (17) then the packet is rejected with an 'port-unreachable' ICMP (ICMP6).

• if the protocol is ICMP (1) then the packet is rejected with a 'addr-unreachable' ICMP.

• otherwise, the packet is rejected with a 'adm-prohibited' ICMP.

You can modify this behavior by implementing your own action that handles REJECT and specifying it's name in this option. The nolog and inline options will automatically be assumed for the specified action.

The following action implements the standard behavior:

?format 2
Broadcast(DROP) -       -       -
DROP            -       -       2
INLINE          -       -       6       ;; -j REJECT --reject-with tcp-reset
INLINE          -       -       17      ;; -j REJECT
?if __IPV4
INLINE          -       -       1       ;; -j REJECT --reject-with icmp-host-unreachable
INLINE          -       -       -       ;; -j REJECT --reject-with icmp-host-prohibited
INLINE          -       -       58      ;; -j REJECT --reject-with icmp6-addr-unreachable
INLINE          -       -       -       ;; -j REJECT --reject-with icmp6-adm-prohibited
INLINE          -       -       -       ;; -j REJECT


Added in Shorewall 4.4.10. The default is No. If set to Yes, at least one optional interface must be up in order for the firewall to be in the started state. Intended to be used with the m[blue]Shorewall Init Packagem[][21].


Added in Shorewall 5.0.1 to replace LEGACY_RESTART which was added in Shorewall 5.0.0. In that release, the reload command was redefined to do what restart had done in earlier releases and restart became a true restart (equivalent to stop followed by start). When RESTART=reload, the restart command performs the same operation as the reload command making it compatible with earlier releases. If not specified, RESTART=reload is assumed.


Added in Shorewall 4.5.9. When set to Yes (the default), provider marks are restored unconditionally at the top of the mangle OUTPUT and PREROUTING chains, even if the saved mark is zero. When this option is set to No, the mark is restored only if it is non-zero. If you have problems with IPSEC ESP packets not being routed correctly on output, try setting this option to No.


Specifies the simple name of a file in /var/lib/shorewall6 to be used as the default restore script in the shorewall6 save, shorewall6 restore, shorewall6 forget and shorewall6 -f start commands.


Added in Shorewall 4.5.7. Determines the disposition of packets entering from interfaces with the rpfilter option (see m[blue]shorewall6-interfacesm[][22](5)). Packets disposed of by this option are those whose response packets would not be sent through the same interface receiving the packet.


Added in shorewall 4.5.7. Determines the logging of packets disposed via the RPFILTER_DISPOSITION. The default value is info.


Re-enabled in Shorewall 4.4.6. If SAVE_IPSETS=Yes, then the current contents of your ipsets will be saved by the shorewall stop and shorewall save commands and restored by the shorewall start and shorewall restore commands.

Beginning with Shorewall 4.6.4, you can restrict the set of ipsets saved by specifying a setlist (a comma-separated list of ipv6 ipset names).


This option is used to specify the shell program to be used to interpret the compiled script. If not specified or specified as a null value, /bin/sh is assumed. Using a light-weight shell such as ash or dash can significantly improve performance.


Added in Shorewall 4.4.20. The default setting is DROP which causes smurf packets (see the nosmurfs option in m[blue]shorewall6-interfacesm[][22](5)) to be dropped. A_DROP causes the packets to be audited prior to being dropped and requires AUDIT_TARGET support in the kernel and ip6tables.


Specifies the logging level for smurf packets (see the nosmurfs option in m[blue]shorewall6-interfacesm[][22](5)). If set to the empty value ( SMURF_LOG_LEVEL="" ) then smurfs are not logged.


Added in Shorewall 4.4.20. Determines the disposition of packets matching the sfilter option (see m[blue]shorewall6-interfacesm[][22](5)) and of hairpin packets on interfaces without the routeback option.[23] interfaces without the routeback option.


Added on Shorewall 4.4.20. Determines the logging of packets matching the sfilter option (see m[blue]shorewall6-interfacesm[][22](5)) and of hairpin packets on interfaces without the routeback option.[24] interfaces without the routeback option. The default is info. If you don't wish for these packets to be logged, use SFILTER_LOG_LEVEL=none.


Determines if Shorewall6 is allowed to start. As released from, this option is set to No. When set to Yes or yes, Shorewall6 may be started. Used as a guard against Shorewall6 being accidentally started before it has been configured.


If specified, determines where Shorewall will log the details of each start, reload, restart, refresh, try, and safe-* command. Logging verbosity is determined by the setting of LOG_VERBOSITY above.


This parameter should be set to the name of a file that the firewall should create if it starts successfully and remove when it stops. Creating and removing this file allows Shorewall6 to work with your distribution's initscripts. For RedHat, this should be set to /var/lock/subsys/shorewall6. For Debian, the value is /var/lock/shorewall6 and in LEAF it is /var/run/shorewall.


If specified, gives the pathname of the 'tc' executable. If not specified, 'tc' is assumed and the utility will be located using the current PATH setting.


Added in Shorewall 4.4.26. The number of bits at the low end of the 32-bit packet mark to be used for traffic shaping marking. May be zero. See MASK_BITS above for default value.


If you say Yes or yes here, Shorewall6 will use a script that you supply to configure traffic shaping. The script must be named 'tcstart' and must be placed in a directory on your CONFIG_PATH.

If you say No or no then traffic shaping is not enabled.

If you set TC_ENABLED=Internal or internal or leave the option empty then Shorewall6 will use its builtin traffic shaper (tc4shorewall6 written by Arne Bernin.

If you set TC_ENABLED=Simple (Shorewall 4.4.6 and later), simple traffic shaping using m[blue]shorewall-tcinterfacesm[][25](5) and m[blue]shorewall-tcprim[][26](5) is enabled.

Beginning with Shorewall 4.4.15, if you set TC_ENABLED=Shared or shared, then you should create symbolic links from your Shorewall6 configuration directory (normally /etc/shorewall6/) to your Shorewall tcdevices and tcclasses files. This allows the compiler to have access to your Shorewall traffic shaping configuration so that it can validate CLASSIFY rules in m[blue]shorewall6-tcrulesm[][9] (5).

If you also run Shorewall and if you have TC_ENABLED=Internal in your m[blue]shorewall-confm[][10](5), then you will want TC_ENABLED=No or TC_ENABLED=Shared in this file.


Normally, Shorewall6 tries to protect users from themselves by preventing PREROUTING and OUTPUT tcrules from being applied to packets that have been marked by the 'track' option in m[blue]shorewall6-providersm[][27](5).

If you know what you are doing, you can set TC_EXPERT=Yes and Shorewall6 will not include these cautionary checks.


Added in Shorewall 4.4.6. Determines the mapping of a packet's TOS field to priority bands. See m[blue]shorewall6-tcprim[][28](5). The map consists of 16 space-separated digits with values 1, 2 or 3. A value of 1 corresponds to Linux priority 0, 2 to Linux priority 1, and 3 to Linux Priority 2. The first entry gives the priority of TOS value 0, the second of TOS value 1, and so on. See tc-prio(8) for additional information.

The default setting is TC_PRIOMAP="2 3 3 3 2 3 1 1 2 2 2 2 2 2 2 2".


Determines the disposition of TCP packets that fail the checks enabled by the tcpflags interface option (see m[blue]shorewall6-interfacesm[][22](5)) and must have a value of ACCEPT (accept the packet), REJECT (send an RST response) or DROP (ignore the packet). If not set or if set to the empty value (e.g., TCP_FLAGS_DISPOSITION="") then TCP_FLAGS_DISPOSITION=DROP is assumed.


Determines the syslog level for logging packets that fail the checks enabled by the tcpflags interface option. The value must be a valid syslogd log level. If you don't want to log these packets, set to the empty value (e.g., TCP_FLAGS_LOG_LEVEL="").


Added in Shorewall 4.4.3. When set to Yes, causes the track option to be assumed on all providers defined in m[blue]shorewall6-providersm[][27](5). May be overridden on an individual provider through use of the notrack option. The default value is 'No'.

Beginning in Shorewall 4.4.6, setting this option to 'Yes' also simplifies PREROUTING rules in m[blue]shorewall6-tcrulesm[][9](5). Previously, when TC_EXPERT=No, packets arriving through 'tracked' provider interfaces were unconditionally passed to the PREROUTING tcrules. This was done so that tcrules could reset the packet mark to zero, thus allowing the packet to be routed using the 'main' routing table. Using the main table allowed dynamic routes (such as those added for VPNs) to be effective. The m[blue]shorewall6-rtrulesm[][29](5) file was created to provide a better alternative to clearing the packet mark. As a consequence, passing these packets to PREROUTING complicates things without providing any real benefit. Beginning with Shorewall 4.4.6, when TRACK_PROVIDERS=Yes and TC_EXPERT=No, packets arriving through 'tracked' interfaces will not be passed to the PREROUTING rules. Since TRACK_PROVIDERS was just introduced in 4.4.3, this change should be transparent to most, if not all, users.


Added in Shorewall 4.5.20. If set to Yes, causes the compiler to add a comment to iptables rules to indicate the file name and line number of the configuration entry that generated the rule. If set to No (the default), then no such comments are added.

Setting this option to Yes requires the Comments capability in iptables and kernel.

Beginning with Shorewall 5.0.5, the option may also be set to File. That setting causes similar comments to be added to the .ip6tables-restore-input file, which is normally created in /var/lib/shorewall6.


Added in Shorewall 4.5.13. Shorewall has traditionally passed UNTRACKED packets through the NEW section of m[blue]shorewall6-rulesm[][6] (5). When a packet in UNTRACKED state fails to match any rule in the UNTRACKED section, the packet is disposed of based on this setting. The default value is CONTINUE for compatibility with earlier versions.


Added in Shorewall 4.5.13. Packets in the UNTRACKED state that do not match any rule in the UNTRACKED section of m[blue]shorewall6-rulesm[][6] (5) are logged at this level. The default value is empty which means no logging is performed.


Added in Shorewall6 4.4.25. When set to 'Yes', this option causes the Shorewall6 multi-ISP feature to create a set of routing rules which are resilient to changes in the main routing table. Such changes can occur for a number of reasons, VPNs going up and down being an example. The idea is to send packets through the main table prior to applying any of the Shorewall6-generated routing rules. So changes to the main table will affect the routing of packets by default.


1. Both the DUPLICATE and the COPY columns in m[blue]shorewall6-providersm[][27](5) file must remain empty (or contain "-").

2. The default route is added to the the 'default' table rather than to the main table.

3. balance is assumed unless loose is specified.

4. Packets are sent through the main routing table by a rule with priority 999. In m[blue]shorewall6-routing_rulesm[][30](5), the range 1-998 may be used for inserting rules that bypass the main table.

5. All provider gateways must be specified explicitly in the GATEWAY column. detect may not be specified.

6. You should disable all default route management outside of Shorewall6. If a default route is added to the main table while Shorewall is started, then all policy routing will stop working (except for those routing rules in the priority range 1-998).

Prior to Shorewall 4.6.0, if USE_DEFAULT_RT was not set or if it was set to the empty string then USE_DEFAULT_RT=No was assumed. Beginning with Shorewall 4.6.0, the default is USE_DEFAULT_RT=Yes and use of USE_DEFAULT_RT=No is deprecated.

The enable, disable and reenable commands do not work correctly when USE_DEFAULT_RT=No.


Added in Shorewall 4.4.27. Normally, when Shorewall creates a Netfilter chain that relates to an interface, it uses the interface's logical name as the base of the chain name. For example, if the logical name for an interface is OAKLAND, then the input chain for traffic arriving on that interface would be 'OAKLAND_in'. If this option is set to Yes, then the physical name of the interface will be used the base of the chain name.


Added in Shorewall 4.5.15. When set to 'Yes', Shorewall will use routing table (provider) names in the generated script rather than table numbers. When set to 'No' (the default), routing table numbers will be used.

If you set USE_RT_NAMES=Yes and KEEP_RT_TABLES=Yes, then you must insure that all of your providers have entries in /etc/iproute2/rt_tables as well as the following entries:
255 local
254 main
253 default
250 balance
0 unspec
Without these entries, the firewall will fail to start.


Added in Shorewall 5.0.9. When Yes (the default), messages produced by the ?INFO and ?WARNING directives include the filename and linenumber of the directive. When set to No, that additional information is omitted. The setting may be overridden on a directive by directive basis by following ?INFO or ?WARNING with '!' (no intervening white space).


Shorewall6 has traditionally been very noisy (produced lots of output). You may set the default level of verbosity using the VERBOSITY OPTION.

Values are:

0 - Silent. You may make it more verbose using the -v
1 - Major progress messages displayed
2 - All progress messages displayed (pre Shorewall6-3.2.0
If not specified, then 2 is assumed.


Added in Shorewall 4.5.12. When set to Yes (the default), the compiler issues a warning when it finds a capabilities file that doesn't specify all of the capabilities supported by the compiler. When WARNOLDCAPVERSION is set to No, no warning is issued.


Deprecated in Shorewall 4.4.26 in favor of TC_BITS and MASK_BITS.

When set to No (the default), traffic shaping marks are 8 bytes wide (possible values are 1-255). When WIDE_TC_MARKS=Yes, traffic shaping marks are 14 bytes wide (values 1-16383). The setting of WIDE_TC_MARKS also has an effect on the HIGH_ROUTE_MARKS option (see above).


Added in Shorewall 4.6.11. Over time, there have been a number of changes in Shorewall that work around defects in other products such as iptables and ipset. When WORKAROUNDS=Yes, these workarounds are enabled; when WORKAROUNDS=No, they are disabled. If not specified or if specified as empty, WORKAROUNDS=Yes is assumed.

Do not set WORKAROUNDS=Yes if you need to be able to use Shorewall-generated scripts (such as created by the save command) built by Shorewall 4.4.7 or older.


Added in Shorewall 5.0.12, this is a workaround for an issue where packet marks are not zeroed by the kernel. It should be set to No (the default) unless you find that incoming packets are being mis-routed for no apparent reasons.

Do not set this option to Yes if you have IPSEC software running on the firewall system.


Added in Shorewall 4.4.26. When non-zero, enables automatic packet marking by source zone and determines the number of bits in the 32-bit packet mark to be used for the zone mark. Default value is 0.


Added in Shorewall 4.4.4. This option determines how Shorewall constructs chain names involving zone names and/or 'all'. Beginning with Shorewall 4.6.0, the default is '-' (e.g., fw-net); prior to that release, the default was '2' (e.g., fw2net).




alternative specificaitons
the complete matrix of host groups defined by the zones, interfaces and hosts files
Shorewall Init Package
Hairpin packets are packets that are routed out of the same interface that they arrived on.
Hairpin packets are packets that are routed out of the same interface that they arrived on.