policyd-spf.conf - Man Page
policyd-spf python configuration parameters
Version
3.1.0
Usage
Usage:
policyd-spf [/etc/policyd-spf/policyd-spf.conf]
Other Documentation
This documentation assumes you have read Postfix's README_FILES/ SMTPD_POLICY_README and are generally familiar with Sender Policy Framework (SPF). See RFC 7208 for details.
man 1 policyd-spf provides general operation documentation for this package.
man 5 policyd-spf.peruser provides documentation on setting up and using different configuration options on a per user (mail reciepient) basis.
Synopsis
spf-engine provides both a Postfix policy server front end and a milter front end that can be used with any mail server which supports the milter protocol. Most options apply to both policy server and milter usage. Options specific to the milter are listed at the end. The package operates with a default installed configuration file and set of default configuration options that are used if the configuration file cannot be found. These options can be changed by changing the installed configuration files or through giving a path to an alternate configuration file.
Description
Additionally, whitelisting certain IP addresses from SPF checks is supported. This man page and the sample configuration file show the format to use. These options can be adjusted on a per user (mail recipient) basis. Details on per user settings can be found in policyd-spf.peruser(5). Per user settings are applicable only to the policy server.
Options
Logging
"debugLevel" controls the amount of information logged by the policy server.
The default, 1, logs no debugging messages, just basic policy results and errors generated through the policy server. This value can be increased up to 5 (values higher than 5 will not cause an error, but will not log any additional information).
debug level 2 adds a log message if no client address (IP address from which the connection was made), Mail From address, or HELO/EHLO name is received by the policy server, and logs SPF results for each Mail From and HELO check.
debug level 3 generates a log message each time the policy server starts and each time it exits, as well as logging a copy of the exact header returned to Postfix to be prepended into the message. Each time the policy server starts, debug level 3 also logs the configuration information used by the policy server.
debug level 4 logs the complete data set received by Postfix via the policy interface and when the end of the entry is read.
debug level 5 is used to debug config file processing and, for this purpose, can only be set in code and not via the config file. It also provides additional internal status details generally of interest only to developers.
If debug level is 0, then the policy server logs errors only.
When set to -1, nothing is logged once the configuration file has been read in, not even error messages. Use with caution.
Default:
debugLevel = 1
Test Operation
The policy server can operate in a test only mode. This allows you to see the potential impact of SPF checking in your mail logs without rejecting mail. Headers are prepended in messages, but message delivery is not affected. This mode is not enabled by default. To enable it, set TestOnly = 0. If you want test operation with no header added, also set Header_Type = None.
Note: This option was previously named defaultSeedOnly. This is still accepted, but logs an error.
Default:
TestOnly = 1
Helo/Ehlo Checking
HELO check rejection policy options are:
Fail - (Default) Reject only on HELO Fail. HELO/EHLO is known first in the SMTP dialogue and there is no practical reason to waste resources on Mail From checks if the HELO check will already cause the message to be rejected. This should not cause interoperability problems when used for HELO.
SPF_Not_Pass - Reject if result not Pass, None, or Temperror (alternatively put, reject if the SPF result is Fail, Softfail, Neutral, PermError). Unlike Mail From checking, there are no standard e-mail use cases where a HELO check should not Pass if there is an SPF record for the HELO name (transparent forwarding, for example, is not an issue). HELO/EHLO is known first in the SMTP dialogue and there is no practical reason to waste resources on Mail From checks if the HELO check will already cause the message to be rejected. This is not consistent with the RFC 7208 requirement to treat none and neutral the same, but should not cause interoperability problems when used for HELO.
Softfail - Reject on HELO Softfail or Fail. HELO/EHLO is known first in the SMTP dialogue and there is no practical reason to waste resources on Mail From checks if the HELO check will already cause the message to be rejected. This should not cause interoperability problems when used for HELO.
Null - Only reject HELO Fail for Null sender (SPF Classic). This is the approach used by the pre-RFC 4408 reference implementation and many of the pre- RFC specifications. Use of at least this option (SPF_Not_Pass or Fail) are preferred) is highly recommended.
False - Never reject on HELO, append header only. This is useful for post-SMTP spam filters such as SpamAssassin.
No_Check - Never check HELO. This is only recommended if you are calling the policy server twice (once for HELO checks and once for Mail From) with two different configuration files. This approach is useful to get both the HELO and Mail From headers prepended to a message.
Note: If No_Mail is set to True, it overrides rejections in any case where the SPF record is not "v=spf1 -all".
Default:
HELO_reject = Fail
Helo/Ehlo Pass Restriction
HELO Pass Restriction allows integration with other Postfix access controls by provding a user supplied name of a postfix access restriction to be applied to a message when the HELO checking result is Pass. The indicated restriction must be an action as defined for a Postfix SMTP server access table access(5) and explained in the Postfix RESTRICTION CLASS README. The README.per_user_whitelisting file provided with this distribution provides examples. Note: A helo pass restriction will be the returned result even if the mail from result would cause the message to be rejected.
Example:
HELO_pass_restriction = helo_passed_spf
Default:
None
Mail From CHECKING
Mail From rejection policy options are:
SPF_Not_Pass - Reject if result not Pass/None/Tempfail. This option is not RFC 7208 compliant since the mail with an SPF Neutral result is treated differently than mail with no SPF record and Softfail results are not supposed to cause mail rejection. Global use of this option is not recommended. Use per-domain if needed (per-domain usage described below).
Softfail - Reject on Mail From Softfail or Fail. Global use of this option is not recommended. Use per-domain if needed (per-domain usage described below).
Fail (default) - Reject on Mail From Fail.
False - Never reject on Mail From, append header only. This is useful for post-SMTP spam filters such as SpamAssassin.
No_Check - Never check Mail From/Return Path. This is only recommended if you are calling the policy server twice (once for HELO checks and once for Mail From) with two different configuration files. This approach is useful to get both the HELO and Mail From headers prepended to a message. It could also be used to do HELO checking only (because HELO checking has a lower false positive risk than Mail From checking), but this approach may not be fully RFC 7208 compliant since the Mail From identity is mandatory if HELO checking does not reach a definitive result.
Note: If No_Mail is set to True, it overrides rejections in any case where the SPF record is not "v=spf1 -all". Default:
Mail_From_reject = Fail
Mail From PASS RESTRICTION
Mail From Pass Restriction allows integration with other Postfix access contlols by provding a user supplied name of a postfix access restriction to be applied to a message when the Mail From checking result is Pass. The indicated restriction must be an action as defined for a Postfix SMTP server access table access(5) and explained in the Postfix RESTRICTION CLASS README. The README.per_user_whitelisting file provided with this distribution provides examples. Note: A mail from pass restriction will be the returned result even if the helo result would cause the message to be rejected.
Example:
Mail_From_pass_restriction = mfrom_passed_spf
Default:
None
Limit Rejections To Domains That Send No Mail
No_Mail - Only reject when SPF indicates the host/domain sends no mail. This option will only cause mail to be rejected if the HELO/Mail From record is "v=spf1 -all". This option is useful for rejecting mail in situations where the tolerance for rejecting wanted mail is very low. It operates on both HELO and Mail From identities if set.
Default:
No_Mail = False
Domain Specific Receiver Policy
Using this option, a list of domains can be defined for special processing when messages do not Pass SPF. This can be useful for commonly spoofed domains that are not yet publishing SPF records with -all. Specifically, if mail from a domain in this list has a Neutral/Softfail result, it will be rejected (as if it had a Fail result). If needed, it is better to do it on a per-domain basis rather than globally.
Example:
Reject_Not_Pass_Domains = aol.com,hotmail.com
Default:
None
Permanent Error Processing
Policy for rejecting due to SPF PermError options are:
True - Reject the message if the SPF result (for HELO or Mail From) is PermError. This has a higher short-term false positive risk, but does result in senders getting feedback that they have a problem with their SPF record.
False - Treat PermError the same as no SPF record at all. This is consistet with the pre-RFC usage (the pre-RFC name for this error was "Unknown").
This is a global option that affects both HELO and Mail From scopes when checks for that scope are enabled. The only per scope setting that can over-ride this is Mail_From/HELO_reject = False/
Default:
PermError_reject = False
Temporary Error Processing
Policy for deferring messages due to SPF TempError options are:
True - Defer the message if the SPF result (for HELO or Mail From) is TempError. This is the traditional usage and has proven useful in reducing acceptance of unwanted messages. Sometimes spam senders do not retry. Sometimes by the time a message is retried the sending IP has made it onto a DNS RBL and can then be rejected. This is not the default because it is possible for some DNS errors that are classified as "Temporary" per RFC 7208 to be permanent in the sense that they require operator intervention to correct.
This is a global option that affects both HELO and Mail From scopes when checks for that scope are enabled. The only per scope setting that can over-ride this is Mail_From/HELO_reject = False/
False - Treat TempError the same as no SPF record at all. This is the default to minimize false positive risk.
Default:
TempError_Defer = False
Prospective SPF Check
Prospective SPF checking - Check to see if mail sent from the defined IP address would pass. This is useful for outbound MTAs to avoid sending mail that would Fail SPF checks when received. Disable HELO checking when using this option. It's only potentially useful for Mail From checking. SPF Received headers are not added when this option is used.
Prospective = 192.168.0.4
Default:
None
Local SPF Bypass List
Do not check SPF for localhost addresses - add to skip addresses to skip SPF for internal networks if desired. Defaults are standard IPv4 and IPv6 localhost addresses. This can also be used, to allow mail from local clients submitting mail to an MTA also acting as a Mail Submission Agent (MSA) to be skipped. An x-header is prepended indicating SPF checks were skipped due to a local address. This is a trace header only. Note the lack of spaces in the list.
Default:
skip_addresses = 127.0.0.0/8,::ffff:127.0.0.0/104,::1
SPF IP Whitelist
A comma separated CIDR Notation list of IP addresses to skip SPF checks for. Use this list to whitelist trusted relays (such as a secondary MX and trusted forwarders). An x-header is prepended indicating the IP was whitelisted against SPF checks. This is a trace header only. Note the lack of spaces in the list.
Example:
Whitelist = 192.168.0.0/31,192.168.1.0/30
Default:
None
SPF Helo Whitelist
A comma separated HELO/EHLO host names to skip SPF checks for. Use this list to whitelist trusted relays (such as a secondary MX and trusted forwarders) or to work around a host with a buggy SPF record. An x-header is prepended indicating the host was whitelisted against SPF checks. This is a trace header only. Note the lack of spaces in the list.
This option includes a check to ensure the connect IP address is referenced in an A or AAAA record by the HELO/EHLO domain that is whitelisted. This is to avoid inadvertent bypass of SPF checks if HELO/EHLO names are forged. If a HELO/EHLO domain is unable to pass such a forward IP address match check, then use an SPF IP Whitelist for the host's IP address instead.
Example:
HELO_Whitelist = relay.example.com,sender.example.org
Default:
None
SPF Domain Whitelist
Domain_Whitelist: List of domains whose sending IPs should be whitelisted from SPF checks. Use this to list trusted forwarders by domain name. Client IP addresses are tested against SPF records published by the listed domains. This is useful for large forwarders with complex outbound infrastructures and SPF records. This option is less scalable than the SPF IP Whitelist. An x-header is prepended indicating the IP was whitelisted against SPF checks. This is a trace header only. This option does nothing if the domain does not have an SPF record. In this case use the SPF IP Whitelist described above or Domain_Whitelist_PTR (below). Note the lack of spaces in the list.
Example:
Domain_Whitelist = pobox.com,trustedforwarder.org
Default:
None
PTR Domain Whitelist
Domain_Whitelist_PTR: List of domains (and subdomains) whose sending IPs should be whitelisted from SPF checks based on PTR match of the domain. Use this to list trusted forwarders by domain name if they do not publish SPF records. Client IP addresses PTR names are tested to see if they match the listed domains. This is useful for large forwarders with complex outbound infrastructures, but no SPF records and predictable host naming. Matching is done using the same rules as the SPF PTR mechanism as described in RFC 7208. List the parent domain and all subdomains will match. This option is less scalable than the SPF IP Whitelist. An x-header is prepended indicating the IP was whitelisted against SPF checks. This is a trace header only. This option does nothing if the host does not have a PTR record record. In this case use the SPF IP Whitelist described above. Note the lack of spaces in the list.
Example:
Domain_Whitelist_PTR = yahoo.com,yahoogroups.com
Default:
None
SPF Enhanced Status Codes
By default, Postfix will use the 4/5.7.1 enhanced status code for defer/reject actions from the policy server (originally defined in RFC 1893, RFC 3463 is the current reference). New email authentication (including SPF) specific codes were defined in RFC 7372. The policy server now overrides the Postfix enhanced status codes to use the RFC 7372 values. This can be disabled by setting this option to "No" in the event of interoperability issues.
Example:
SPF_Enhanced_Status_Codes = No
Default:
Yes
Results Header
The standard method for documenting SPF results in a message (for consumption by downstream processes) is the Received-SPF header defined in RFC 7208. This is the default header to use. Results can also be documented in the Authentication-Results header, which is also covered in RFC 7208. The default is Received-SPF (SPF), but inclusion of Authentication-Results (AR) headers as an alternative to Received-SPF can be specified.
If there is a requirement to prepend both Received-SPF and Authentication- Results headers, then it must be done by processing the message with more than one instance of the policy server using different configuration files with different Header_Type settings.
For no header at all, use Header_Type = None.
Examples:
Header_Type = SPF or Header_Type = AR
Default:
SPF
Hide RCPT to in Results Header
Both Received-SPF and Authentication-Results (AR) header fields include the receiving (RCPT TO) host, if available. In this application, it will always be the host of the first recipient sent by the sending MTA, even if that recipient is a BCC recipient. This is unavoidable as neither the Postfix policy interface nor the milter interface provides any indication if the recipient is BCC or not (this information is not available in until after SMTP DATA in the body of the message). This presented a possible avenue for a privacy breach, but in version 3.0 it was modified to be only the host name and not the full address.
Hide_Receiver can set to Yes interest of maximizing privacy. If per user processing is not in use, this setting will replace the actual host name with <UNKNOWN> both in header fields and SMTP responses. The latter may make it more difficult for senders to troubleshoot issues with their SPF deployments. As an implementation detail, currently specifying any value other than No will result in the recipient being hidden, but that may change in the future. If per user processing is in use, this setting has no effect.
For the milter interface, checks are always done at the Mail From stage in the SMTP protocol, so receiver data is never available. When using the milter, this setting has no effect.
Examples:
Hide_Receiver = Yes or Hide_Receiver = No
Default:
No
Authentications Results Authentication Identifier
Every Authentication-Results header field has an authentication identifier field ('Authserv_Id'). This is similar in syntax to a fully-qualified domain name. See policyd-spf.conf.5 and RFC 7001 paragraph 2.4 for details. Default is HOSTNAME. The results of socket.gethostname will be used unless an alternate value is specified. An Authserv-Id must be provided if Header_Type 'AR' is used.
The authentication identifier field provides a unique identifier that refers to the authenticating service within a given administrative domain. The identifier MUST be unique to that domain. This identifier is intended to be machine-readable and not necessarily meaningful to users.
Example:
Authserv_Id = mx.example.com
Default:
HOSTNAME
DNS Timeout Limit
RFC 7208 recommends an elapsed time limit for SPF checks of at least 20 seconds. Lookup_Time allows the maximum time (seconds) to be adjusted. 20 seconds is the default. This limit is applied separately to Mail From and HELO/EHLO checks, so if both are performed, the lookups may take up to twice Lookup_Time (plus any additional time required for whitelisting related DNS lookups).
It is important that the combination of Lookup_Time(s) and applicable Whitelist_Lookup_Time(s) be less than the smtpd_policy_service_timeout defined for the service (default 100 seconds). Since HELO and Mail From time limits are independent, smtpd_policy_service_timeout needs to be at least double the time allowed for the SPF policy server per entity type.
Example:
Lookup_Time = 20
Default
20 (seconds)
Whitelist DNS Timeout Limit
Some of the available whitelisting mechanisms, i.e. Domain_Whitelist, Domain_Whitelist_PTR, and HELO_Whitelist, require specific non-SPF DNS lookups to determine if a connection should be white listed from SPF checks. The maximum amount of time (in seconds) allocated for each of these checks, when used (none are enabled by default), is controlled by the Whitelist_Lookup_Time parameter. It defaults to 10 seconds and is applied independently to each whitelisting method in use (e.g. if both a Domain_Whitelist_PTR and HELO_Whitelist are defined, together they may take up to 20 seconds). This is in addition to the time allowed for SPF Lookup_Time.
It is important that the combination of Lookup_Time(s) and applicable Whitelist_Lookup_Time(s) be less than the smtpd_policy_service_timeout defined for the service (default 100 seconds). Since HELO and Mail From time limits are independent, smtpd_policy_service_timeout needs to be at least double the time allowed for the SPF policy server per entity type.
Example:
Whitelist_Lookup_Time = 10
Default
10 (seconds)
DNS Void Lookup Limit
RFC 7208 adds a new processing limit called "void lookup limit" (See section 4.6.4). Void lookups are DNS queries within an SPF record for which DNS queries return either a positive answer (RCODE 0) with an answer count of 0, or a "Name Error" (RCODE 3) answer. This should not need to be changed. Although new in an RFC in RFC 7208, this limit has been widely deployed in the Mail::SPF perl library without issue. Default is 2, but it can be adjusted.
Example:
Void_Limit = 2
Default
2
Mock SPF Check To Add Milter Compatibility Header Field
In some versions of postfix, for bizarre Sendmail compatibility reasons, the first header field added by a policy server is not visible to milters. To make this easy to work around, set the Mock value to true and a fixed header field will be inserted so the actual SPF check will be the second field and visible to milters such as DMARC milter.
To use this feature requires additional postfix configuration to execute the second, mock, instance of the policy server:
Add a second service to /etc/postfix/master.cf:
policyd-spf-mock unix - n n - 0 spawn
user=nobody argv=/usr/bin/policyd-spf '/etc/pypolicyd-spf/mock
Configure the Postfix additional policy service in /etc/postfix/main.cf to run before the regular service:
smtpd_recipient_restrictions =
...
reject_unauth_destination
check_policy_service unix:private/policyd-spf-mock
check_policy_service unix:private/policyd-spf
...
Create the specified configuration file in the location indicated with Mock set to True.
Example:
Mock = True
Default
False
Log Queue ID
Include Postfix Queue ID in policy server log messages. This option only applies to log level 0 or 1 messages that are unique to the policy server interface when configured for post-queue processing. Not used for milter or common code logging. Included to assist with correlating logging with specific Postfix message processing.
Example:
QueueID = True
Default
True
Reason Message/URL
If a message is rejected or deferred because of the SPF policy, a reason is given for logging and debugging purposes. The String configured supports the following format specifiers:
rejectdefer - either the string 'rejected' or 'deferred'
spf - SPF result code
url - Parameterized URL to http://www.openspf.net/ explaining SPF
The default url can be overridden by setting Reason_URL. It defaults to the defunct openspf.org URL due to a lack of a standardized alternative. This is intended for use with a web service which provides additional information to a user. The following components are supported and required:
{0} - SPF result type (pass, fail, etc.) {1} - Result rationale provided by the SPF library {2} - Connect IP address {3} - Recipient address, if available
All four components must be included in your custom URL. As shown in the default and the example, '=' may be used in the URL. This is the only configuration setting where it is allowed.
Example:
Reason_Message = Message {rejectdefer} due to: {spf}. Reason_URL = http://www.openspf.net/Why?s={0};id={1};ip={2};r={3}
Default:
Reason_Message = Message {rejectdefer} due to: {spf}. Please see {url} Reason_URL = http://www.openspf.net/Why?s={0};id={1};ip={2};r={3}
Milter Only Options
- Socket (string)
Specifies the socket that should be established by the filter to receive connections from postfix(1) in order to provide service. socketspec is in one of two forms: local:path, which creates a UNIX domain socket at the specified path, or inet:port[@host] or inet6:port[@host] which creates a TCP socket on the specified port and in the specified protocol family. If the host is not given as either a hostname or an IP address, the socket will be listening on all interfaces. A literal IP address must be enclosed in square brackets. This option is mandatory in the configuration file.
- PidFile (string)
Specifies the path to a file that should be created at process start containing the process ID. If not specified, no such file will be created.
- UserID (string)
Attempts to become the specified userid before starting operations. The value is of the form userid[:group]. The process will be assigned all of the groups and primary group ID of the named userid unless an alternate group is specified.
- UMask (integer)
Requests a specific permissions mask to be used for file creation. This only really applies to creation of the socket when Socket specifies a UNIX domain socket, and to the PidFile (if any); temporary files are created by the mkstemp(3) function that enforces a specific file mode on creation regardless of the process umask. See umask(2) for more information.
- InternalHosts (dataset)
Identifies a set internal hosts whose mail should be signed rather than verified. The set should contain on each line a hostname, domain name (e.g. ".example.com"), IP address, an IPv6 address (including an IPv4 mapped address), or a CIDR-style IP specification (e.g. "192.168.1.0/24"). An entry beginning with a bang ("!") character means "not", allowing exclusions of specific hosts that are otherwise members of larger sets. Host and domain names are matched first, then the IP or IPv6 address depending on the connection type. More precise entries are preferred over less precise ones, i.e. "192.168.1.1" will match before "!192.168.1.0/24". The text form of IPv6 addresses will be forced to lowercase when queried (RFC5952), so the contents of this data set should also use lowercase. The IP address portion of an entry may optionally contain square brackets; both forms (with and without) will be checked. If not specified, the default of "127.0.0.1" is applied. Naturally, providing a value here overrides the default, so if mail from 127.0.0.1 should be signed, the list provided here should include that address explicitly.
- MacroList (dataset)
Defines a set of MTA-provided macros that should be checked to see if the sender has been determined to be a local user and therefore whether or not the message checked. If a value is specified matching a macro name in the data set, the value of the macro must match a value specified (matching is case-sensitive), otherwise the macro must be defined but may contain any value. The set is empty by default, meaning macros are not considered when making the verify decision. The general format of the value is value1[|value2[|...]]; if one or more value is defined then the macro must be set to one of the listed values, otherwise the macro must be set but can contain any value.
In order for the macro and its value to be available to the filter for checking, the MTA must send it during the protocol exchange. This is either accomplished via manual configuration of the MTA to send the desired macros or, for MTA/filter combinations that support the feature, the filter can request those macros that are of interest. The latter is a feature negotiated at the time the filter receives a connection from the MTA and its availability depends upon the version of milter used to compile the filter and the version of the MTA making the connection.
See Also
man 1 policyd-spf, man 5 policyd-spf.peruser, python-spf, <http://www.openspf.net>, RFC 7208, RFC 7001, RFC 7372
Authors
This version of policyd-spf was written by Copyright © 2007-2016, Scott Kitterman <scott@kitterman.com>. It is derived from Tumgreyspf, written by Sean Reifschneider, tummy.com, ltd <jafo@tummy.com>. Portions of the documentation were written by Meng Weng Wong <mengwong@pobox.com>.
This man-page was created by Scott Kitterman <scott@kitterman.com> and is licensed under the same terms as the program.