. Update the Copyright year (only) when changing content.
. /////////////////////////////////////////////////////////////////////////////
-.set previousversion "4.93"
+.set previousversion "4.94"
.include ./local_params
.set ACL "access control lists (ACLs)"
&%local_part_suffix%& is in use, the prefix or suffix is removed from the local
part before this check. If you want to do precondition tests on local parts
that include affixes, you can do so by using a &%condition%& option (see below)
-.new
that uses the variables &$local_part$&, &$local_part_prefix$&,
&$local_part_prefix_v$&, &$local_part_suffix$&
and &$local_part_suffix_v$& as necessary.
-.wen
.next
.vindex "&$local_user_uid$&"
.vindex "&$local_user_gid$&"
by Exim in conjunction with the &%-MC%& option. It signifies that the
remote host supports the ESMTP &_DSN_& extension.
-.new
.vitem &%-MCd%&
.oindex "&%-MCd%&"
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-d%& option
to pass on an information string on the purpose of the process.
-.wen
.vitem &%-MCG%&&~<&'queue&~name'&>
.oindex "&%-MCG%&"
signals the final completion of the sequence of processes that are passing
messages through the same SMTP connection.
+.new
+.vitem &%-MCq%&&~<&'recipient&~address'&>&~<&'size'&>
+.oindex "&%-MCq%&"
+This option is not intended for use by external callers. It is used internally
+by Exim to implement quota checking for local users.
+.wen
+
.vitem &%-MCS%&
.oindex "&%-MCS%&"
This option is not intended for use by external callers. It is used internally
without &%-bd%&, this is the only way of causing Exim to write a pid file,
because in those cases, the normal pid file is not used.
-.new
.vitem &%-oPX%&
.oindex "&%-oPX%&"
.cindex "pid (process id)" "of daemon"
The daemon uses it when terminating due to a SIGTEM, possibly in
combination with &%-oP%&&~<&'path'&>.
It causes the pid file to be removed.
-.wen
.vitem &%-or%&&~<&'time'&>
.oindex "&%-or%&"
.cindex "queue" "routing"
.cindex "routing" "whole queue before delivery"
.cindex "first pass routing"
+.cindex "queue runner" "two phase"
An option starting with &%-qq%& requests a two-stage queue run. In the first
stage, the queue is scanned as if the &%queue_smtp_domains%& option matched
every domain. Addresses are routed, local deliveries happen, but no remote
transports are run.
-.new
Performance will be best if the &%queue_run_in_order%& option is false.
-.wen
+If that is so and the &%queue_fast_ramp%& option is true then
+in the first phase of the run,
+once a threshold number of messages are routed for a given host,
+a delivery process is forked in parallel with the rest of the scan.
.cindex "hints database" "remembering routing"
The hints database that remembers which messages are waiting for specific hosts
This &(appendfile)& transport is used for local delivery to user mailboxes in
traditional BSD mailbox format.
-.new
We prefer to avoid using &$local_part$& directly to define the mailbox filename,
as it is provided by a potential bad actor.
Instead we use &$local_part_data$&,
the result of looking up &$local_part$& in the user database
(done by using &%check_local_user%& in the the router).
-.wen
By default &(appendfile)& runs under the uid and gid of the
local user, which requires the sticky bit to be set on the &_/var/mail_&
lookup. Lookups of this type are conditional expansion items. Different results
can be defined for the cases of lookup success and failure. See chapter
&<<CHAPexpand>>&, where string expansions are described in detail.
-The key for the lookup is specified as part of the string expansion.
+The key for the lookup is &*specified*& as part of the string expansion.
.next
Lists of domains, hosts, and email addresses can contain lookup requests as a
way of avoiding excessively long linear lists. In this case, the data that is
returned by the lookup is often (but not always) discarded; whether the lookup
succeeds or fails is what really counts. These kinds of list are described in
chapter &<<CHAPdomhosaddlists>>&.
-The key for the lookup is given by the context in which the list is expanded.
+The key for the lookup is &*implicit*&,
+given by the context in which the list is expanded.
.endlist
String expansions, lists, and lookups interact with each other in such a way
In the second example, the lookup is a single item in a domain list. It causes
Exim to use a lookup to see if the domain that is being processed can be found
-in the file. The file could contains lines like this:
+in the file.
+The file could contains lines like this:
.code
domain1:
domain2:
and a comma-separated list of options.
Each option is a &"name=value"& pair.
Whether an option is meaningful depands on the lookup type.
+
+All lookups support the option &"cache=no_rd"&.
+If this is given then the cache that Exim manages for lookup results
+is not checked before diong the lookup.
+The result of the lookup is still written to the cache.
.wen
The rest of this chapter describes the different lookup types that are
The &'single-key'& type requires the specification of a file in which to look,
and a single key to search for. The key must be a non-empty string for the
lookup to succeed. The lookup type determines how the file is searched.
-.new
.cindex "tainted data" "single-key lookups"
The file string may not be tainted
(either underlying implementation or cached value)
returns data, the result is replaced with a non-tainted
version of the lookup key.
-.wen
+.cindex "tainted data" "de-tainting"
.next
.cindex "query-style lookup" "definition of"
The &'query-style'& type accepts a generalized database query. No particular
.cindex "lookup" "dsearch"
.cindex "dsearch lookup type"
&(dsearch)&: The given file must be an
-.new
absolute
-.wen
directory path; this is searched for an entry
whose name is the key by calling the &[lstat()]& function.
The key may not
contain any forward slash characters.
If &[lstat()]& succeeds then so does the lookup.
-.new
.cindex "tainted data" "dsearch result"
The result is regarded as untainted.
.endd
The default matching is for any entry type, including directories
and symlinks.
-.wen
An example of how this
lookup can be used to support virtual domains is given in section
or array; for the latter two a string-representation of the JSON
is returned.
For elements of type string, the returned value is de-quoted.
+
+
+.new
+.next
+.cindex LMDB
+.cindex lookup lmdb
+.cindex database lmdb
+&(lmdb)&: The given file is an LMDB database.
+LMDB is a memory-mapped key-value store,
+with API modeled loosely on that of BerkeleyDB.
+See &url(https://symas.com/products/lightning-memory-mapped-database/)
+for the feature set and operation modes.
+
+Exim provides read-only access via the LMDB C library.
+The library can be obtained from &url(https://github.com/LMDB/lmdb)
+or your operating system package repository.
+To enable LMDB support in Exim set LOOKUP_LMDB=yes in &_Local/Makefile_&.
+
+You will need to separately create the LMDB database file,
+possibly using the &"mdb_load"& utility.
+.wen
+
+
.next
.cindex "linear search"
.cindex "lookup" "lsearch"
.cindex "sqlite lookup type"
.cindex "lookup" "sqlite"
&(sqlite)&: The format of the query is
-new
-an optional filename
-.wen
-followed by an SQL statement
-that is passed to an SQLite database. See section &<<SECTsqlite>>&.
+an SQL statement that is passed to an SQLite database. See section &<<SECTsqlite>>&.
.next
&(testdb)&: This is a lookup type that is used for testing Exim. It is
escapes whitespace and backslash characters with a backslash.
.section "Specifying the server in the query" "SECTspeserque"
-.new
For MySQL, PostgreSQL and Redis lookups (but not currently for Oracle and InterBase),
it is possible to specify a list of servers with an individual query. This is
done by appending a comma-separated option to the query type:
.display
-.endd
&`,servers=`&&'server1:server2:server3:...'&
-.wen
+.endd
Each item in the list may take one of two forms:
.olist
If it contains no slashes it is assumed to be just a host name. The appropriate
${lookup pgsql,servers=master/db/name/pw {UPDATE ...} }
.endd
-.new
-An older syntax places the servers speciification before the qury,
+An older syntax places the servers specification before the query,
semicolon separated:
.code
${lookup mysql{servers=master; UPDATE ...} }
The new version avoids potential issues with tainted
arguments in the query, for explicit expansion.
&*Note*&: server specifications in list-style lookups are still problematic.
-.wen
.section "Special MySQL features" "SECID73"
.new
.oindex &%sqlite_dbfile%&
-The preferred way of specifying the file is by using the
-&%sqlite_dbfile%& option, set to
-an absolute path.
+There are two ways of
+specifying the file.
+The first is is by using the &%sqlite_dbfile%& main option.
+The second, which allows separate files for each query,
+is to use an option appended, comma-separated, to the &"sqlite"&
+lookup type word. The option is the word &"file"&, then an equals,
+then the filename.
+The filename in this case cannot contain whitespace or open-brace charachters.
.wen
+
A deprecated method is available, prefixing the query with the filename
separated by white space.
-This means that the path name cannot contain white space.
+This means that
.cindex "tainted data" "sqlite file"
-It also means that the query cannot use any tainted values, as that taints
+the query cannot use any tainted values, as that taints
the entire query including the filename - resulting in a refusal to open
the file.
-.new
+In all the above cases the filename must be an absolute path.
+
Here is a lookup expansion example:
.code
sqlite_dbfile = /some/thing/sqlitedb
domainlist relay_to_domains = sqlite;\
select * from relays where ip='$sender_host_address';
.endd
-.wen
The only character affected by the &%quote_sqlite%& operator is a single
quote, which it doubles.
-.new
.section "Results of list checking" SECTlistresults
The primary result of doing a list check is a truth value.
In some contexts additional information is stored
The detail of the additional information depends on the
type of match and is given below as the &*value*& information.
-.wen
where &'x.y'& does not match. It's best to avoid negation altogether in
referenced lists if you can.
-.new
.cindex "hiding named list values"
.cindex "named lists" "hiding value of"
Some named list definitions may contain sensitive data, for example, passwords for
.code
hide domainlist filter_for_domains = ldap;PASS=secret ldap::/// ...
.endd
-.wen
Named lists may have a performance advantage. When Exim is routing an
possible to use the same configuration file on several different hosts that
differ only in their names.
-.new
The value for a match will be the primary host name.
-.wen
.next
In today's Internet, the use of domain literals is controversial;
see the &%allow_domain_literals%& main option.
-.new
The value for a match will be the string &`@[]`&.
-.wen
.next
domains = <? @mx_any/ignore=<;127.0.0.1;::1 ? \
an.other.domain ? ...
.endd
-.new
The value for a match will be the list element string (starting &`@mx_`&).
-.wen
.next
list item such as &`*key.ex`& matches &'donkey.ex'& as well as
&'cipher.key.ex'&.
-.new
The value for a match will be the list element string (starting with the asterisk).
Additionally, &$0$& will be set to the matched string
and &$1$& to the variable portion which the asterisk matched.
-.wen
.next
.cindex "regular expressions" "in domain list"
it is not to be expanded (unless you really do want to build a regular
expression by expansion, of course).
-.new
The value for a match will be the list element string (starting with the circumflex).
Additionally, &$0$& will be set to the string matching the regular expression,
and &$1$& (onwards) to any submatches identified by parentheses.
-.wen
The value will be untainted.
.next
-.new
If the pattern starts with the name of a lookup type
of either kind (single-key or query-style) it may be
followed by a comma and options,
The options are lookup-type specific and consist of a comma-separated list.
Each item starts with a tag and and equals "=".
-.wen
.next
.cindex "domain list" "matching literal domain name"
.cindex "tainted data" "de-tainting"
Note that this is commonly untainted
(depending on the way the list was created).
+Specifically, explicit text in the configuration file in not tainted.
This is a useful way of obtaining an untainted equivalent to
the domain, for later operations.
+
+However if the list (including one-element lists)
+is created by expanding a variable containing tainted data,
+it is tainted and so will the match value be.
.endlist
conditions do not expand some of the brace-delimited branches, for security
reasons,
.cindex "tainted data" expansion
+.cindex "tainted data" definition
.cindex expansion "tainted data"
and expansion of data deriving from the sender (&"tainted data"&)
is not permitted.
+.new
+Common ways of obtaining untainted equivalents of variables with
+tainted values
+.cindex "tainted data" "de-tainting"
+come down to using the tainted value as a lookup key in a trusted database.
+This database could be the filesystem structure,
+or the password file,
+or accessed via a DBMS.
+Specific methods are indexed under &"de-tainting"&.
+.wen
+
.section "Literal text in expanded strings" "SECTlittext"
those headers that contain lists of addresses, a comma is also inserted at the
junctions between headers. This does not happen for the &%rheader%& expansion.
-.new
.cindex "tainted data"
When the headers are from an incoming message,
the result of expanding any of these variables is tainted.
-.wen
.vitem &*${hmac{*&<&'hashname'&>&*}{*&<&'secret'&>&*}{*&<&'string'&>&*}}*&
You can use &`fail`& instead of {<&'string3'&>} as in a string extract.
-.vitem "&*${lookup{*&<&'key'&>&*}&~*&<&'search&~type'&>&*&~&&&
- {*&<&'file'&>&*}&~{*&<&'string1'&>&*}&~{*&<&'string2'&>&*}}*&"
-This is the first of one of two different types of lookup item, which are both
-described in the next item.
+.new
+.vitem &*${listquote{*&<&'separator'&>&*}{*&<&'string'&>&*}}*&
+.cindex quoting "for list"
+.cindex list quoting
+This item doubles any occurrence of the separator character
+in the given string.
+An empty string is replaced with a single space.
+This converts the string into a safe form for use as a list element,
+in a list using the given separator.
+.wen
+
-.vitem "&*${lookup&~*&<&'search&~type'&>&*&~{*&<&'query'&>&*}&~&&&
+.vitem "&*${lookup&~{*&<&'key'&>&*}&~*&<&'search&~type'&>&*&~&&&
+ {*&<&'file'&>&*}&~{*&<&'string1'&>&*}&~{*&<&'string2'&>&*}}*&" &&&
+ "&*${lookup&~*&<&'search&~type'&>&*&~{*&<&'query'&>&*}&~&&&
{*&<&'string1'&>&*}&~{*&<&'string2'&>&*}}*&"
.cindex "expansion" "lookup in"
.cindex "file" "lookups"
${readsocket{/socket/name}{request string}{3s:shutdown=no}}
.endd
-.new
The following option names are recognised:
.ilist
&*cache*&
Values are &"yes"& or &"no"& (the default).
If it is enabled, a shutdown as descripbed above is never done.
.endlist
-.wen
A fourth argument allows you to change any newlines that are in the data
will sort an MX lookup into priority order.
+
+.new
+.vitem &*${srs_encode&~{*&<&'secret'&>&*}{*&<&'return&~path'&>&*}{*&<&'original&~domain'&>&*}}*&
+SRS encoding. See SECT &<<SECTSRS>>& for details.
+.wen
+
+
+
.vitem &*${substr{*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&
.cindex "&%substr%& expansion item"
.cindex "substring extraction"
case-independent.
Case and collation order are defined per the system C locale.
+
+.new
+.vitem &*inbound_srs&~{*&<&'local&~part'&>&*}{*&<&'secret'&>&*}*&
+SRS decode. See SECT &<<SECTSRS>>& for details.
+.wen
+
+
.vitem &*inlist&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
&*inlisti&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "string" "comparison"
There can be problems if any of the strings are permitted to contain colon
characters. In the usual way, these have to be doubled to avoid being taken as
-separators. If the data is being inserted from a variable, the &%sg%& expansion
-item can be used to double any existing colons. For example, the configuration
+separators.
+The &%listquote%& expansion item can be used for this.
+For example, the configuration
of a LOGIN authenticator might contain this setting:
.code
-server_condition = ${if pam{$auth1:${sg{$auth2}{:}{::}}}}
-.endd
-For a PLAIN authenticator you could use:
-.code
-server_condition = ${if pam{$auth2:${sg{$auth3}{:}{::}}}}
+server_condition = ${if pam{$auth1:${listquote{:}{$auth2}}}}
.endd
In some operating systems, PAM authentication can be done only from a process
running as root. Since Exim is running as the Exim user when receiving
the complete argument of the ETRN command (see section &<<SECTETRN>>&).
.endlist
-.new
.cindex "tainted data"
If the origin of the data is an incoming message,
-the result of expanding this variable is tainted.
-When un untainted version is needed, one should be obtained from
+the result of expanding this variable is tainted and may not
+be further expanded or used as a filename.
+When an untainted version is needed, one should be obtained from
looking up the value in a local (therefore trusted) database.
Often &$domain_data$& is usable in this role.
-.wen
.vitem &$domain_data$&
.vindex "&$domain_data$&"
-When the &%domains%& option on a router matches a domain by
-means of a lookup, the data read by the lookup is available during the running
-of the router as &$domain_data$&. In addition, if the driver routes the
+When the &%domains%& condition on a router
+.new
+or an ACL
+matches a domain
+against a list, the match value is copied to &$domain_data$&.
+This is an enhancement over previous versions of Exim, when it only
+applied to the data read by a lookup.
+For details on match values see section &<<SECTlistresults>>& et. al.
+.wen
+
+If the router routes the
address to a transport, the value is available in that transport. If the
transport is handling multiple addresses, the value from the first address is
used.
-&$domain_data$& is also set when the &%domains%& condition in an ACL matches a
-domain by means of a lookup. The data read by the lookup is available during
-the rest of the ACL statement. In all other situations, this variable expands
-to nothing.
+&$domain_data$& set in an ACL is available during
+the rest of the ACL statement.
.vitem &$exim_gid$&
.vindex "&$exim_gid$&"
because a message may have many recipients and the system filter is called just
once.
-.new
.cindex "tainted data"
If the origin of the data is an incoming message,
-the result of expanding this variable is tainted.
+the result of expanding this variable is tainted and
+may not be further expanded or used as a filename.
&*Warning*&: the content of this variable is usually provided by a potential
attacker.
Often &$local_part_data$& is usable in this role.
If needed, use a router &%address_data%& or &%set%& option for
the retrieved data.
-.wen
When a message is being delivered to a file, pipe, or autoreply transport as a
result of aliasing or forwarding, &$local_part$& is set to the local part of
.vitem &$local_part_data$&
.vindex "&$local_part_data$&"
-When the &%local_parts%& option on a router matches a local part by means of a
-lookup, the data read by the lookup is available during the running of the
-router as &$local_part_data$&. In addition, if the driver routes the address
-to a transport, the value is available in that transport. If the transport is
-handling multiple addresses, the value from the first address is used.
-
+When the &%local_parts%& condition on a router or ACL
+matches a local part list
.new
-The &%check_local_user%& router option also sets this variable.
+the match value is copied to &$local_part_data$&.
+This is an enhancement over previous versions of Exim, when it only
+applied to the data read by a lookup.
+For details on match values see section &<<SECTlistresults>>& et. al.
.wen
-&$local_part_data$& is also set when the &%local_parts%& condition in an ACL
-matches a local part by means of a lookup. The data read by the lookup is
-available during the rest of the ACL statement. In all other situations, this
-variable expands to nothing.
+The &%check_local_user%& router option also sets this variable.
.vindex &$local_part_prefix$& &&&
&$local_part_prefix_v$& &&&
value of &$local_part$& during routing and subsequent delivery. The values of
any prefix or suffix are in &$local_part_prefix$& and
&$local_part_suffix$&, respectively.
-.new
.cindex "tainted data"
If the specification did not include a wildcard then
the affix variable value is not tainted.
the affix matched by the wildcard is in
&$local_part_prefix_v$& or &$local_part_suffix_v$& as appropriate,
and both the whole and varying values are tainted.
-.wen
.vitem &$local_scan_data$&
.vindex "&$local_scan_data$&"
dns_dnssec_ok = 1
.endd
-.new
In addition, on Linux with glibc 2.31 or newer the resolver library will
default to stripping out a successful validation status.
This will break a previously working Exim installation.
.code
options trust-ad
.endd
-.wen
Exim does not perform DNSSEC validation itself, instead leaving that to a
validating resolver (e.g. unbound, or bind with suitable configuration).
If certificate verification fails it may refer to a failing chain element
which is not the leaf.
+
+.new
+.vitem &$tls_in_resumption$& &&&
+ &$tls_out_resumption$&
+.vindex &$tls_in_resumption$&
+.vindex &$tls_out_resumption$&
+.cindex TLS resumption
+Observability for TLS session resumption. See &<<SECTresumption>>& for details.
+.wen
+
+
.vitem &$tls_in_sni$&
.vindex "&$tls_in_sni$&"
.vindex "&$tls_sni$&"
.cindex "TLS" "Server Name Indication"
+.cindex "TLS" SNI
+.cindex SNI "observability on server"
When a TLS session is being established, if the client sends the Server
Name Indication extension, the value will be placed in this variable.
If the variable appears in &%tls_certificate%& then this option and
.vitem &$tls_out_sni$&
.vindex "&$tls_out_sni$&"
.cindex "TLS" "Server Name Indication"
+.cindex "TLS" SNI
+.cindex SNI "observability in client"
During outbound
SMTP deliveries, this variable reflects the value of the &%tls_sni%& option on
the transport.
.row &%hold_domains%& "hold delivery for these domains"
.row &%local_interfaces%& "for routing checks"
.row &%queue_domains%& "no immediate delivery for these"
+.row &%queue_fast_ramp%& "parallel delivery with 2-phase queue run"
.row &%queue_only%& "no immediate delivery at all"
.row &%queue_only_file%& "no immediate delivery if file exists"
.row &%queue_only_load%& "no immediate delivery if load is high"
This option defines a template file containing paragraphs of text to be used
for constructing bounce messages. Details of the file's contents are given in
chapter &<<CHAPemsgcust>>&.
-.new
.cindex bounce_message_file "tainted data"
The option is expanded to give the file path, which must be
absolute and untainted.
-.wen
See also &%warn_message_file%&.
to handle IPv6 literal addresses.
-.new
.option dkim_verify_hashes main "string list" "sha256 : sha512"
.cindex DKIM "selecting signature algorithms"
This option gives a list of hash types which are acceptable in signatures,
-.wen
and an order of processing.
Signatures with algorithms not in the list will be ignored.
Signatures with algorithms not in the list will be ignored.
-.new
.option dkim_verify_min_keysizes main "string list" "rsa=1024 ed25519=250"
This option gives a list of key sizes which are acceptable in signatures.
The list is keyed by the algorithm type for the key; the values are in bits.
Signatures with keys smaller than given by this option will fail verification.
The default enforces the RFC 8301 minimum key size for RSA signatures.
-.wen
.option dkim_verify_minimal main boolean false
If set to true, verification of signatures will terminate after the
If the resolver library does not support DNSSEC then this option has no effect.
-.new
On Linux with glibc 2.31 or newer this is insufficient, the resolver library
will default to stripping out a successful validation status.
This will break a previously working Exim installation.
.code
options trust-ad
.endd
-.wen
.option dns_ipv4_lookup main "domain list&!!" unset
A NOTIFY=SUCCESS option requests success-DSN messages.
A NOTIFY= option with no argument requests that no delay or failure DSNs
are sent.
-.new
&*Note*&: Supplying success-DSN messages has been criticised
on privacy grounds; it can leak details of internal forwarding.
-.wen
.option dsn_from main "string&!!" "see below"
.cindex "&'From:'& header line" "in bounces"
transport driver.
-.new
.option notifier_socket main string "$spool_directory/exim_daemon_notify"
This option gives the name for a unix-domain socket on which the daemon
listens for work and information-requests.
If the Exim command line uses a &%-oX%& option and does not use &%-oP%&
then a notifier socket is not created.
-.wen
.option openssl_options main "string list" "+no_sslv2 +no_sslv3 +single_dh_use +no_ticket +no_renegotiation"
next queue run. See also &%hold_domains%& and &%queue_smtp_domains%&.
+.new
+.option queue_fast_ramp main boolean false
+.cindex "queue runner" "two phase"
+.cindex "queue" "double scanning"
+If set to true, two-phase queue runs, initiated using &%-qq%& on the
+command line, may start parallel delivery processes during their first
+phase. This will be done when a threshold number of messages have been
+routed for a single host.
+.wen
+
+
.option queue_list_requires_admin main boolean true
.cindex "restricting access to features"
.oindex "&%-bp%&"
smtp_etrn_command = /etc/etrn_command $domain \
$sender_host_address
.endd
-.new
If the option is not set, the argument for the ETRN command must
be a &'#'& followed by an address string.
In this case an &'exim -R <string>'& command is used;
if the ETRN ACL has set up a named-queue then &'-MCG <queue>'& is appended.
-.wen
A new process is created to run the command, but Exim does not wait for it to
complete. Consequently, its status cannot be checked. If the command cannot be
This option is available when Exim is compiled with SPF support.
See section &<<SECSPF>>& for more details.
-.new
.option spf_smtp_comment_template main string&!! "Please%_see%_http://www.open-spf.org/Why"
This option is available when Exim is compiled with SPF support. It
allows the customisation of the SMTP comment that the SPF library
A note on using Exim variables: As
currently the SPF library is initialized before the SMTP EHLO phase,
the variables useful for expansion are quite limited.
-.wen
.option split_spool_directory main boolean false
.option syslog_timestamp main boolean true
.cindex "syslog" "timestamps"
+.cindex timestamps syslog
If &%syslog_timestamp%& is set false, the timestamps on Exim's log lines are
omitted when these lines are sent to syslog. See chapter &<<CHAPlog>>& for
details of Exim's logging.
file is used, the &$tls_in_ourcert$& variable is unreliable.
The macro "_TLS_BAD_MULTICERT_IN_OURCERT" will be defined for those versions.
+.cindex SNI "selecting server certificate based on"
If the option contains &$tls_out_sni$& and Exim is built against OpenSSL, then
if the OpenSSL build supports TLS extensions and the TLS client sends the
Server Name Indication extension, then this option and others documented in
&<<SECTreqciphssl>>& and &<<SECTreqciphgnu>>&.
+.new
+.option tls_resumption_hosts main "host list&!!" unset
+.cindex TLS resumption
+This option controls which connections to offer the TLS resumption feature.
+See &<<SECTresumption>>& for details.
+.wen
+
+
.option tls_try_verify_hosts main "host list&!!" unset
.cindex "TLS" "client certificate verification"
.cindex "certificate" "verification of client"
been in the queue for a specified amount of time, as specified by
&%delay_warning%&. Details of the file's contents are given in chapter
&<<CHAPemsgcust>>&.
-.new
.cindex warn_message_file "tainted data"
The option is expanded to give the file path, which must be
absolute and untainted.
-.wen
See also &%bounce_message_file%&.
that is associated with any addresses that are accepted by the router.
However, the option has no effect when an address is just being verified.
Each list item is separately expanded, at transport time.
-.new
If an item ends in *, it will match any header with the given prefix.
-.wen
The way in which
the text is used to remove header lines at transport time is described in
section &<<SECTheadersaddrem>>&. Header lines are not actually removed until
This behaviour can be overridden by setting &%rcpt_include_affixes%& true on
the relevant transport.
-.new
.vindex &$local_part_prefix_v$&
If wildcarding (above) was used then the part of the prefix matching the
wildcard is available in &$local_part_prefix_v$&.
-.wen
When an address is being verified, &%local_part_prefix%& affects only the
behaviour of the router. If the callout feature of verification is in use, this
yield empty addresses, for example, items containing only RFC 2822 address
comments.
-.new
.cindex "tainted data" "in filenames"
.cindex redirect "tainted data"
Tainted data may not be used for a filename.
In the examples above, &$local_part$& is used for looking up data held locally
on the system, and not used directly (the second example derives &$home$& via
the passsword file or database, using &$local_part$&).
-.wen
.code
list1: :include:/opt/lists/list1
.endd
-.new
.cindex "tainted data" "in filenames"
.cindex redirect "tainted data"
Tainted data may not be used for a filename.
-.wen
.next
.cindex "address redirection" "to black hole"
.cindex "delivery" "discard"
If the result of the expansion is an empty string, or if the expansion
is forced to fail, no action is taken. Other expansion failures are treated as
errors and cause the delivery to be deferred.
-.new
If an item ends in *, it will match any header with the given prefix.
-.wen
Matching headers are omitted from the message as it is transported, as described
in section &<<SECTheadersaddrem>>&. Header removal can also be specified by
used to determine the file or directory name for the delivery. Normally, the
contents of &$address_file$& are used in some way in the string expansion.
.endlist
-.new
.cindex "tainted data" "in filenames"
.cindex appendfile "tainted data"
Tainted data may not be used for a file or directory name.
This means that, for instance, &$local_part$& cannot be used directly
as a component of a path. It can however be used as the key for a lookup
which returns a path (or component).
-.wen
.cindex "Sieve filter" "configuring &(appendfile)&"
details of the local delivery environment and chapter &<<CHAPbatching>>&
for a discussion of local delivery batching.
-.new
.cindex "tainted data" "in pipe command"
.cindex pipe "tainted data"
Tainted data may not be used for the command name.
-.wen
.section "Concurrent delivery" "SECID140"
.cindex "RFC 3030" "CHUNKING"
This option provides a list of servers to which, provided they announce
CHUNKING support, Exim will attempt to use BDAT commands rather than DATA.
-.new
Unless DKIM signing is being done,
-.wen
BDAT will not be used in conjunction with a transport filter.
.option hosts_try_dane smtp "host list&!!" *
permits this.
+.new
+.option message_linelength_limit smtp integer 998
+.cindex "line length" limit
+This option sets the maximum line length, in bytes, that the transport
+will send. Any messages with lines exceeding the given value
+will fail and a failure-DSN ("bounce") message will if possible be returned
+to the sender.
+The default value is that defined by the SMTP standards.
+
+It is generally wise to also check in the data ACL so that messages
+received via SMTP can be refused without producing a bounce.
+.wen
+
+
.option multi_domain smtp boolean&!! true
.vindex "&$domain$&"
When this option is set, the &(smtp)& transport can handle a number of
ciphers is a preference order.
+.new
+.option tls_resumption_hosts smtp "host list&!!" unset
+.cindex TLS resumption
+This option controls which connections to use the TLS resumption feature.
+See &<<SECTresumption>>& for details.
+.wen
+
+
.option tls_sni smtp string&!! unset
.cindex "TLS" "Server Name Indication"
+.cindex "TLS" SNI
+.cindex SNI "setting in client"
.vindex "&$tls_sni$&"
If this option is set then it sets the $tls_out_sni variable and causes any
TLS session to pass this value as the Server Name Indication extension to
option is passed. When authentication succeeds, the identity of the user
who authenticated is placed in &$auth1$&.
-.new
The Dovecot configuration to match the above wil look
something like:
.code
auth_mechanisms = plain login ntlm
.endd
-.wen
.ecindex IIDdcotauth1
.ecindex IIDdcotauth2
made that any particular new authentication mechanism will be supported
without code changes in Exim.
-.new
The library is expected to add support in an upcoming
realease for the SCRAM-SHA-256 method.
The macro _HAVE_AUTH_GSASL_SCRAM_SHA_256 will be defined
.option client_username gsasl string&!! unset
This option is exapanded before use, and should result in
the account name to be used.
-.wen
-.new
.option client_spassword gsasl string&!! unset
If a SCRAM mechanism is being used and this option is set
it is used in preference to &%client_password%&.
with the PBKDF2-prepared password, hex-encoded.
Note that this value will depend on the salt and iteration-count
supplied by the server.
-.wen
non-issue, as a man-in-the-middle attack will cause the correct client and
server to see different identifiers and authentication will fail.
-.new
This is
only usable by mechanisms which support "channel binding"; at time of
writing, that's the SCRAM family.
When using this feature the "-PLUS" variants of the method names need to be used.
-.wen
This defaults off to ensure smooth upgrade across Exim releases, in case
this option causes some clients to start failing. Some future release
.option server_scram_iter gsasl string&!! 4096
This option provides data for the SCRAM family of mechanisms.
-.new
The &$auth1$&, &$auth2$& and &$auth3$& variables are available
when this option is expanded.
either the iteration count or the salt).
A minimum value of 4096 is required by the standards
for all current SCRAM mechanism variants.
-.wen
.option server_scram_salt gsasl string&!! unset
This option provides data for the SCRAM family of mechanisms.
-.new
The &$auth1$&, &$auth2$& and &$auth3$& variables are available
when this option is expanded.
The value should be a base64-encoded string,
of random data typically 4-to-16 bytes long.
If unset or empty after expansion the library will provides a value for the
protocol conversation.
-.wen
-.new
.option server_key gsasl string&!! unset
.option server_skey gsasl string&!! unset
These options can be used for the SCRAM family of mechanisms
The libgsasl library release includes a utility &'gsasl'& which can be used
to generate these values.
-.wen
.option server_service gsasl string &`smtp`&
or need not succeed respectively.
The &%tls_verify_cert_hostnames%& option lists hosts for which additional
-checks are made: that the host name (the one in the DNS A record)
-is valid for the certificate.
+name checks are made on the server certificate.
+.new
+The match against this list is, as per other Exim usage, the
+IP for the host. That is most closely associated with the
+name on the DNS A (or AAAA) record for the host.
+However, the name that needs to be in the certificate
+is the one at the head of any CNAME chain leading to the A record.
+.wen
The option defaults to always checking.
The &(smtp)& transport has two OCSP-related options:
.section "Use of TLS Server Name Indication" "SECTtlssni"
.cindex "TLS" "Server Name Indication"
+.cindex "TLS" SNI
+.cindex SNI
.vindex "&$tls_in_sni$&"
.oindex "&%tls_in_sni%&"
With TLS1.0 or above, there is an extension mechanism by which extra
.ecindex IIDencsmtp2
+.new
+.section "TLS Resumption" "SECTresumption"
+.cindex TLS resumption
+TLS Session Resumption for TLS 1.2 and TLS 1.3 connections can be used (defined
+in RFC 5077 for 1.2). The support for this requires GnuTLS 3.6.3 or OpenSSL 1.1.1
+(or later).
+
+Session resumption (this is the "stateless" variant) involves the server sending
+a "session ticket" to the client on one connection, which can be stored by the
+client and used for a later session. The ticket contains sufficient state for
+the server to reconstruct the TLS session, avoiding some expensive crypto
+calculation and (on TLS1.2) one full packet roundtrip time.
+
+.ilist
+Operational cost/benefit:
+
+ The extra data being transmitted costs a minor amount, and the client has
+ extra costs in storing and retrieving the data.
+
+ In the Exim/Gnutls implementation the extra cost on an initial connection
+ which is TLS1.2 over a loopback path is about 6ms on 2017-laptop class hardware.
+ The saved cost on a subsequent connection is about 4ms; three or more
+ connections become a net win. On longer network paths, two or more
+ connections will have an average lower startup time thanks to the one
+ saved packet roundtrip. TLS1.3 will save the crypto cpu costs but not any
+ packet roundtrips.
+
+.cindex "hints database" tls
+ Since a new hints DB is used on the TLS client,
+ the hints DB maintenance should be updated to additionally handle "tls".
+
+.next
+Security aspects:
+
+ The session ticket is encrypted, but is obviously an additional security
+ vulnarability surface. An attacker able to decrypt it would have access
+ all connections using the resumed session.
+ The session ticket encryption key is not committed to storage by the server
+ and is rotated regularly (OpenSSL: 1hr, and one previous key is used for
+ overlap; GnuTLS 6hr but does not specify any overlap).
+ Tickets have limited lifetime (2hr, and new ones issued after 1hr under
+ OpenSSL. GnuTLS 2hr, appears to not do overlap).
+
+ There is a question-mark over the security of the Diffie-Helman parameters
+ used for session negotiation.
+
+.next
+Observability:
+
+ The &%log_selector%& "tls_resumption" appends an asterisk to the tls_cipher "X="
+ element.
+
+ The variables &$tls_in_resumption$& and &$tls_out_resumption$&
+ have bits 0-4 indicating respectively
+ support built, client requested ticket, client offered session,
+ server issued ticket, resume used. A suitable decode list is provided
+ in the builtin macro _RESUME_DECODE for in &%listextract%& expansions.
+
+.next
+Control:
+
+The &%tls_resumption_hosts%& main option specifies a hostlist for which
+exim, operating as a server, will offer resumption to clients.
+Current best practice is to not offer the feature to MUA connection.
+Commonly this can be done like this:
+.code
+tls_resumption_hosts = ${if inlist {$received_port}{587:465} {:}{*}}
+.endd
+If the peer host matches the list after expansion then resumption
+is offered and/or accepted.
+
+The &%tls_resumption_hosts%& smtp transport option performs the
+equivalent function for operation as a client.
+If the peer host matches the list after expansion then resumption
+is attempted (if a stored session is available) or the information
+stored (if supplied by the peer).
+
+
+.next
+Issues:
+
+ In a resumed session:
+.ilist
+ The variables &$tls_{in,out}_cipher$& will have values different
+ to the original (under GnuTLS).
+.next
+ The variables &$tls_{in,out}_ocsp$& will be "not requested" or "no response",
+ and the &%hosts_require_ocsp%& smtp trasnport option will fail.
+. XXX need to do something with that hosts_require_ocsp
+.endlist
+
+.endlist
+.wen
+
.section DANE "SECDANE"
.cindex DANE
controlled by &%acl_smtp_connect%& or &%acl_smtp_helo%&. See also
&%pipelining_advertise_hosts%&.
-.new
.vitem &*control&~=&~queue/*&<&'options'&>* &&&
&*control&~=&~queue_only*&
.oindex "&%queue%&"
The control only applies to the current message, not to any subsequent ones that
may be received in the same SMTP connection.
-.wen
.vitem &*control&~=&~submission/*&<&'options'&>
.cindex "message" "submission"
.cindex "header lines" "verifying header names only ASCII"
.cindex "verifying" "header names only ASCII"
This condition is relevant only in an ACL that is run after a message has been
-received, that is, in an ACL specified by &%acl_smtp_data%& or
-&%acl_not_smtp%&. It checks all header names (not the content) to make sure
+received.
+This usually means an ACL specified by &%acl_smtp_data%& or &%acl_not_smtp%&.
+It checks all header names (not the content) to make sure
there are no non-ASCII characters, also excluding control characters. The
allowable characters are decimal ASCII values 33 through 126.
(eg. is generated from the received message)
they must be protected from the options parsing by doubling:
.code
-verify = sender=${sg{${address:$h_sender:}}{/}{//}}
+verify = sender=${listquote{/}{${address:$h_sender:}}}
.endd
.endlist
immediately after a successful redirection. By default, if a redirection
generates just one address, that address is also verified. See further
discussion in section &<<SECTredirwhilveri>>&.
+.new
+.next
+If the &%quota%& option is specified for recipient verify,
+successful routing to an appendfile transport is followed by a call into
+the transport to evaluate the quota status for the recipient.
+No actual delivery is done, but verification will succeed if the quota
+is sufficient for the message (if the sender gave a message size) or
+not already exceeded (otherwise).
+.wen
.endlist
.cindex "verifying address" "differentiating failures"
&%recipient%&: The RCPT command in a callout was rejected.
.next
&%postmaster%&: The postmaster check in a callout was rejected.
+.new
+.next
+&%quota%&: The quota check for a local recipient did non pass.
.endlist
The main use of these variables is expected to be to distinguish between
+.new
+.section "Quota caching" "SECTquotacache"
+.cindex "hints database" "quota cache"
+.cindex "quota" "cache, description of"
+.cindex "caching" "quota"
+Exim caches the results of quota verification
+in order to reduce the amount of resources used.
+The &"callout"& hints database is used.
+
+The default cache periods are five minutes for a positive (good) result
+and one hour for a negative result.
+To change the periods the &%quota%& option can be followed by an equals sign
+and a number of optional paramemters, separated by commas.
+For example:
+.code
+verify = recipient/quota=cachepos=1h,cacheneg=1d
+.endd
+Possible parameters are:
+.vlist
+.vitem &*cachepos&~=&~*&<&'time&~interval'&>
+.cindex "quota cache" "positive entry expiry, specifying"
+Set the lifetime for a positive cache entry.
+A value of zero seconds is legitimate.
+
+.vitem &*cacheneg&~=&~*&<&'time&~interval'&>
+.cindex "quota cache" "negative entry expiry, specifying"
+As above, for a negative entry.
+
+.vitem &*no_cache*&
+Set both positive and negative lifetimes to zero.
+.wen
+
.section "Sender address verification reporting" "SECTsenaddver"
.cindex "verifying" "suppressing error details"
See section &<<SECTaddressverification>>& for a general discussion of
LOCAL_SCAN_SOURCE=Local/local_scan.c
.endd
for example. The function must be called &[local_scan()]&;
-.new
the source file(s) for it should first #define LOCAL_SCAN
and then #include "local_scan.h".
-.wen
It is called by
Exim after it has received a message, when the success return code is about to
be sent. This is after all the ACLs have been run. The return code from your
arguments. It flushes the output stream, and returns a non-zero value if there
is an error.
-.new
.vitem &*void&~*store_get(int,BOOL)*&
This function accesses Exim's internal store (memory) manager. It gets a new
chunk of memory whose size is given by the first argument.
FALSE if it is locally-sourced.
Exim bombs out if it ever
runs out of memory. See the next section for a discussion of memory handling.
-.wen
.vitem &*void&~*store_get_perm(int,BOOL)*&
This function is like &'store_get()'&, but it always gets memory from the
.section "Per-address filtering" "SECTperaddfil"
-.vindex "&$domain$&"
-.vindex "&$local_part$&"
+.vindex "&$domain_data$&"
+.vindex "&$local_part_data$&"
In contrast to the system filter, which is run just once per message for each
delivery attempt, it is also possible to set up a system-wide filtering
operation that runs once for each recipient address. In this case, variables
-such as &$local_part$& and &$domain$& can be used, and indeed, the choice of
-filter file could be made dependent on them. This is an example of a router
-which implements such a filter:
+such as &$local_part_data$& and &$domain_data$& can be used,
+and indeed, the choice of filter file could be made dependent on them.
+This is an example of a router which implements such a filter:
.code
central_filter:
check_local_user
data = ${lookup{$local_part}lsearch{/etc/mail/virtual/$domain_data}}
no_more
.endd
-.new
The &%domains%& option specifies that the router is to be skipped, unless there
is a file in the &_/etc/mail/virtual_& directory whose name is the same as the
domain that is being processed.
The &(dsearch)& lookup used results in an untainted version of &$domain$&
being placed into the &$domain_data$& variable.
-.wen
When the router runs, it looks up the local
part in the file to find a new address (or list of addresses). The &%no_more%&
.code
my_mailboxes:
driver = appendfile
- file = /var/mail/$domain/$local_part_data
+ file = /var/mail/$domain_data/$local_part_data
user = mail
.endd
This uses a directory of mailboxes for each domain. The &%user%& setting is
cases by testing the variable &$local_part_suffix$&. For example:
.code
if $local_part_suffix contains -special then
-save /home/$local_part/Mail/special
+save /home/$local_part_data/Mail/special
endif
.endd
If the filter file does not exist, or does not deal with such addresses, they
&`*tls_certificate_verified `& certificate verification status
&`*tls_cipher `& TLS cipher suite on <= and => lines
&` tls_peerdn `& TLS peer DN on <= and => lines
+&` tls_resumption `& append * to cipher field
&` tls_sni `& TLS SNI on <= lines
&` unknown_in_list `& DNS lookup failed in list match
connection, and a certificate is supplied by the remote host, the peer DN is
added to the log line, preceded by DN=.
.next
+.cindex "log" "TLS resumption"
+.cindex "TLS" "logging session resumption"
+.new
+&%tls_resumption%&: When a message is sent or received over an encrypted
+connection and the TLS session resumed one used on a previous TCP connection,
+an asterisk is appended to the X= cipher field in the log line.
+.wen
+.next
.cindex "log" "TLS SNI"
.cindex "TLS" "logging SNI"
+.cindex SNI logging
&%tls_sni%&: When a message is received over an encrypted connection, and
the remote host provided the Server Name Indication extension, the SNI is
added to the log line, preceded by SNI=.
.next
&'ratelimit'&: the data for implementing the ratelimit ACL condition
.next
+.new
+&'tls'&: TLS session resumption data
+.wen
+.next
&'misc'&: other hints data
.endlist
. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////
-.chapter "DKIM, SPF and DMARC" "CHAPdkim" &&&
- "DKIM, SPF and DMARC Support"
+.chapter "DKIM, SPF, SRS and DMARC" "CHAPdkim" &&&
+ "DKIM, SPF, SRS and DMARC Support"
.section "DKIM (DomainKeys Identified Mail)" SECDKIM
.cindex "DKIM"
whether or not each header is present in the message.
The default list is available for the expansion in the macro
&"_DKIM_SIGN_HEADERS"&
-.new
and an oversigning variant is in &"_DKIM_OVERSIGN_HEADERS"&.
-.wen
If a name is repeated, multiple headers by that name (or the absence thereof)
will be signed. The textually later headers in the headers part of the
.vitem &%$dkim_key_length%&
Number of bits in the key.
-.new
Valid only once the key is loaded, which is at the time the header signature
is verified, which is after the body hash is.
-.wen
Note that RFC 8301 says:
.code
.vindex &%spf_smtp_comment_template%&
This contains a string that can be used in a SMTP response
to the calling party. Useful for "fail".
-.new
The string is generated by the SPF library from the template configured in the main config
option &%spf_smtp_comment_template%&.
-.wen
.endlist
+.section "SRS (Sender Rewriting Scheme)" SECTSRS
+.cindex SRS "sender rewriting scheme"
+
+.new
+SRS can be used to modify sender addresses when forwarding so that
+SPF verification does not object to them.
+It operates by encoding the original envelope sender in a new
+sender local part and using a domain run by the forwarding site
+as the new domain for the sender. Any DSN message should be returned
+to this new sender at the forwarding site, which can extract the
+original sender from the coded local part and forward the DSN to
+the originator.
+
+This is a way of avoiding the breakage that SPF does to forwarding.
+The constructed local-part will be longer than the original,
+leading to possible problems with very long addresses.
+The changing of the sender address also hinders the tracing of mail
+problems.
+
+Exim can be built to include native SRS support. To do this
+SUPPORT_SRS=yes must be defined in &_Local/Makefile_&.
+If this has been done, the macros _HAVE_SRS and _HAVE_NATIVE_SRS
+will be defined.
+The support is limited to SRS0-encoding; SRS1 is not supported.
+
+.cindex SRS excoding
+To encode an address use this expansion item:
+.vlist
+.vitem &*${srs_encode&~{*&<&'secret'&>&*}{*&<&'return&~path'&>&*}{*&<&'original&~domain'&>&*}}*&
+.cindex "&%srs_encode%& expansion item"
+.cindex SRS "expansion item"
+The first argument should be a secret known and used by all systems
+handling the recipient domain for the original message.
+There is no need to periodically change this key; a timestamp is also
+encoded.
+The second argument should be given as the envelope sender address before this
+encoding operation.
+The third argument should be the recipient domain of the message when
+it arrived at this system.
+.endlist
+
+.cindex SRS decoding
+To decode an address use this expansion condition:
+.vlist
+.vitem &*inbound_srs&~{*&<&'local&~part'&>&*}{*&<&'secret'&>&*}*&
+The first argument should be the recipient local prt as is was received.
+The second argument is the site secret.
+
+If the messages is not for an SRS-encoded recipient the condition will
+return false. If it is, the condition will return true and the variable
+&$srs_recipient$& will be set to the decoded (original) value.
+.endlist
+
+Example usage:
+.code
+ #macro
+ SRS_SECRET = <pick something unique for your site for this. Use on all MXs.>
+
+ #routers
+
+ outbound:
+ driver = dnslookup
+ # if outbound, and forwarding has been done, use an alternate transport
+ domains = ! +my_domains
+ transport = ${if eq {$local_part@$domain} \
+ {$original_local_part@$original_domain} \
+ {remote_smtp} {remote_forwarded_smtp}}
+
+ inbound_srs:
+ driver = redirect
+ senders = :
+ domains = +my_domains
+ # detect inbound bounces which are SRS'd, and decode them
+ condition = ${if inbound_srs {$local_part} {SRS_SECRET}}
+ data = $srs_recipient
+
+ inbound_srs_failure:
+ driver = redirect
+ senders = :
+ domains = +my_domains
+ # detect inbound bounces which look SRS'd but are invalid
+ condition = ${if inbound_srs {$local_part} {}}
+ allow_fail
+ data = :fail: Invalid SRS recipient address
+
+ #... further routers here
+
+
+ # transport; should look like the non-forward outbound
+ # one, plus the max_rcpt and return_path options
+ remote_forwarded_smtp:
+ driver = smtp
+ # modify the envelope from, for mails that we forward
+ max_rcpt = 1
+ return_path = ${srs_encode {SRS_SECRET} {$return_path} {$original_domain}}
+.endd
+
+
+.wen
+
+
+
.section DMARC SECDMARC
.cindex DMARC verification
the most current version can be downloaded
from a link at &url(https://publicsuffix.org/list/public_suffix_list.dat).
See also the util/renew-opendmarc-tlds.sh script.
-.new
The default for the option is unset.
If not set, DMARC processing is disabled.
-.wen
The &%dmarc_history_file%& option, if set
The name is placed in the variable &$event_name$& and the event action
expansion must check this, as it will be called for every possible event type.
-.new
The current list of events is:
-.wen
.display
&`dane:fail after transport `& per connection
&`msg:complete after main `& per message
git://git.exim.org / users / heiko / exim.git / blobdiff