&` _DRIVER_ROUTER_* `& router drivers
&` _DRIVER_TRANSPORT_* `& transport drivers
&` _DRIVER_AUTHENTICATOR_* `& authenticator drivers
+&` _EXP_COND_* `& expansion conditions
+&` _EXP_ITEM_* `& expansion items
+&` _EXP_OP_* `& expansion operators
+&` _EXP_VAR_* `& expansion variables
&` _LOG_* `& log_selector values
&` _OPT_MAIN_* `& main config options
&` _OPT_ROUTERS_* `& generic router options
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 to be expanded.
.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 &*implicit*&,
-given by the context in which the list is expanded.
+Depending on the lookup type (see below)
+the key for the lookup may need to be &*specified*& as above
+or may be &*implicit*&,
+given by the context in which the list is being checked.
.endlist
String expansions, lists, and lookups interact with each other in such a way
domains = ${lookup{$sender_host_address}lsearch{/some/file}}
domains = lsearch;/some/file
.endd
+.ilist
The first uses a string expansion, the result of which must be a domain list.
The key for an expansion-style lookup must be given explicitly.
No strings have been specified for a successful or a failing lookup; the
When the lookup succeeds, the result of the expansion is a list of domains (and
possibly other types of item that are allowed in domain lists).
.cindex "tainted data" "de-tainting"
-.cindex "de-tainting" "using a lookup expansion""
+.cindex "de-tainting" "using a lookup expansion"
The result of the expansion is not tainted.
+.next
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 key for a list-style lookup is implicit, from the lookup context, if
the lookup is a single-key type (see below).
-For query-style lookup types the key must be given explicitly.
+For query-style lookup types the query must be given explicitly.
+.endlist
It is possible, though no doubt confusing, to use both kinds of lookup at once.
Consider a file containing lines like this:
(either underlying implementation or cached value)
returns data, the result is replaced with a non-tainted
version of the lookup key.
+
.next
.cindex "query-style lookup" "definition of"
The &'query-style'& type accepts a generalized database query. No particular
.cindex "single-key lookup" "list of types"
The following single-key lookup types are implemented:
-.ilist
+.subsection cdb
.cindex "cdb" "description of"
.cindex "lookup" "cdb"
.cindex "binary zero" "in lookup key"
-&(cdb)&: The given file is searched as a Constant DataBase file, using the key
+The given file is searched as a Constant DataBase file, using the key
string without a terminating binary zero. The cdb format is designed for
indexed files that are read frequently and never updated, except by total
re-creation. As such, it is particularly suitable for large files containing
because the code for reading cdb files is included directly in Exim itself.
However, no means of building or testing cdb files is provided with Exim, so
you need to obtain a cdb distribution in order to do this.
-.next
+
+.subsection dbm
.cindex "DBM" "lookup type"
.cindex "lookup" "dbm"
.cindex "binary zero" "in lookup key"
-&(dbm)&: Calls to DBM library functions are used to extract data from the given
+Calls to DBM library functions are used to extract data from the given
DBM file by looking up the record with the given key. A terminating binary
zero is included in the key that is passed to the DBM library. See section
&<<SECTdb>>& for a discussion of DBM libraries.
the DB_UNKNOWN option. This enables it to handle any of the types of database
that the library supports, and can be useful for accessing DBM files created by
other applications. (For earlier DB versions, DB_HASH is always used.)
-.next
+
+.subsection dbmjz
.cindex "lookup" "dbmjz"
.cindex "lookup" "dbm &-- embedded NULs"
.cindex "sasldb2"
.cindex "dbmjz lookup type"
-&(dbmjz)&: This is the same as &(dbm)&, except that the lookup key is
+This is the same as &(dbm)&, except that the lookup key is
interpreted as an Exim list; the elements of the list are joined together with
ASCII NUL characters to form the lookup key. An example usage would be to
authenticate incoming SMTP calls using the passwords from Cyrus SASL's
&_/etc/sasldb2_& file with the &(gsasl)& authenticator or Exim's own
&(cram_md5)& authenticator.
-.next
+
+.subsection dbmnz
.cindex "lookup" "dbmnz"
.cindex "lookup" "dbm &-- terminating zero"
.cindex "binary zero" "in lookup key"
.cindex "Courier"
.cindex "&_/etc/userdbshadow.dat_&"
.cindex "dbmnz lookup type"
-&(dbmnz)&: This is the same as &(dbm)&, except that a terminating binary zero
+This is the same as &(dbm)&, except that a terminating binary zero
is not included in the key that is passed to the DBM library. You may need this
if you want to look up data in files that are created by or shared with some
other application that does not use terminating zeros. For example, you need to
calls using the passwords from Courier's &_/etc/userdbshadow.dat_& file. Exim's
utility program for creating DBM files (&'exim_dbmbuild'&) includes the zeros
by default, but has an option to omit them (see section &<<SECTdbmbuild>>&).
-.next
+
+.subsection dsearch
.cindex "lookup" "dsearch"
.cindex "dsearch lookup type"
-&(dsearch)&: The given file must be an
-absolute
-directory path; this is searched for an entry
+The given file must be an absolute 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.
+The key may not contain any forward slash characters.
If &[lstat()]& succeeds then so does the lookup.
.cindex "tainted data" "dsearch result"
The result is regarded as untainted.
An example of how this
lookup can be used to support virtual domains is given in section
&<<SECTvirtualdomains>>&.
-.next
+
+.subsection iplsearch
.cindex "lookup" "iplsearch"
.cindex "iplsearch lookup type"
-&(iplsearch)&: The given file is a text file containing keys and data. A key is
+The given file is a text file containing keys and data. A key is
terminated by a colon or white space or the end of the line. The keys in the
file must be IP addresses, or IP addresses with CIDR masks. Keys that involve
IPv6 addresses must be enclosed in quotes to prevent the first internal colon
rather than omitting the key portion.
Note however that the key portion will have been de-quoted.
-.next
+.subsection json
.cindex lookup json
.cindex json "lookup type"
.cindex JSON expansions
-&(json)&: The given file is a text file with a JSON structure.
+The given file is a text file with a JSON structure.
An element of the structure is extracted, defined by the search key.
The key is a list of subelement selectors
(colon-separated by default but changeable in the usual way)
For elements of type string, the returned value is de-quoted.
-.next
+.subsection lmdb
.cindex LMDB
.cindex lookup lmdb
.cindex database lmdb
-&(lmdb)&: The given file is an LMDB database.
+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/)
possibly using the &"mdb_load"& utility.
-.next
+.subsection lsearch
.cindex "linear search"
.cindex "lookup" "lsearch"
.cindex "lsearch lookup type"
.cindex "case sensitivity" "in lsearch lookup"
-&(lsearch)&: The given file is a text file that is searched linearly for a
+The given file is a text file that is searched linearly for a
line beginning with the search key, terminated by a colon or white space or the
end of the line. The search is case-insensitive; that is, upper and lower case
letters are treated as the same. The first occurrence of the key that is found
quoted keys (exactly as for unquoted keys). There is no special handling of
quotes for the data part of an &(lsearch)& line.
-.next
+.subsection nis
.cindex "NIS lookup type"
.cindex "lookup" "NIS"
.cindex "binary zero" "in lookup key"
-&(nis)&: The given file is the name of a NIS map, and a NIS lookup is done with
+The given file is the name of a NIS map, and a NIS lookup is done with
the given key, without a terminating binary zero. There is a variant called
&(nis0)& which does include the terminating binary zero in the key. This is
reportedly needed for Sun-style alias files. Exim does not recognize NIS
aliases; the full map names must be used.
-.next
+.subsection (n)wildlsearch
.cindex "wildlsearch lookup type"
.cindex "lookup" "wildlsearch"
.cindex "nwildlsearch lookup type"
file that are regular expressions can be made case-sensitive by the use of
&`(-i)`& within the pattern. The following forms of wildcard are recognized:
-. ==== As this is a nested list, any displays it contains must be indented
-. ==== as otherwise they are too far to the left.
-
.olist
The string may begin with an asterisk to mean &"ends with"&. For example:
.code
- *.a.b.c data for anything.a.b.c
- *fish data for anythingfish
+*.a.b.c data for anything.a.b.c
+*fish data for anythingfish
.endd
.next
The string may begin with a circumflex to indicate a regular expression. For
example, for &(wildlsearch)&:
.code
- ^\N\d+\.a\.b\N data for <digits>.a.b
+^\N\d+\.a\.b\N data for <digits>.a.b
.endd
Note the use of &`\N`& to disable expansion of the contents of the regular
expression. If you are using &(nwildlsearch)&, where the keys are not
string-expanded, the equivalent entry is:
.code
- ^\d+\.a\.b data for <digits>.a.b
+^\d+\.a\.b data for <digits>.a.b
.endd
The case-insensitive flag is set at the start of compiling the regular
expression, but it can be turned off by using &`(-i)`& at an appropriate point.
For example, to make the entire pattern case-sensitive:
.code
- ^(?-i)\d+\.a\.b data for <digits>.a.b
+^(?-i)\d+\.a\.b data for <digits>.a.b
.endd
If the regular expression contains white space or colon characters, you must
lookup name terminated by a semicolon, and followed by lookup data. For
example:
.code
- cdb;/some/file data for keys that match the file
+cdb;/some/file data for keys that match the file
.endd
The data that is obtained from the nested lookup is discarded.
.endlist olist
&((n)wildlsearch)& can &'not'& be turned into a DBM or cdb file, because those
lookup types support only literal keys.
-.next
+.subsection spf
.cindex "spf lookup type"
.cindex "lookup" "spf"
-&(spf)&: If Exim is built with SPF support, manual lookups can be done
+If Exim is built with SPF support, manual lookups can be done
(as opposed to the standard ACL condition method).
For details see section &<<SECSPF>>&.
-.endlist ilist
.section "Query-style lookup types" "SECTquerystylelookups"
The supported query-style lookup types are listed below. Further details about
many of them are given in later sections.
-.ilist
+.subsection dnsdb
.cindex "DNS" "as a lookup type"
.cindex "lookup" "DNS"
-&(dnsdb)&: This does a DNS search for one or more records whose domain names
+This does a DNS search for one or more records whose domain names
are given in the supplied query. The resulting data is the contents of the
records. See section &<<SECTdnsdb>>&.
-.next
+
+.subsection ibase
.cindex "InterBase lookup type"
.cindex "lookup" "InterBase"
-&(ibase)&: This does a lookup in an InterBase database.
-.next
+This does a lookup in an InterBase database.
+
+.subsection ldap
.cindex "LDAP" "lookup type"
.cindex "lookup" "LDAP"
-&(ldap)&: This does an LDAP lookup using a query in the form of a URL, and
+This does an LDAP lookup using a query in the form of a URL, and
returns attributes from a single entry. There is a variant called &(ldapm)&
that permits values from multiple entries to be returned. A third variant
called &(ldapdn)& returns the Distinguished Name of a single entry instead of
any attribute values. See section &<<SECTldap>>&.
-.next
+
+.subsection mysql
.cindex "MySQL" "lookup type"
.cindex "lookup" "MySQL"
-&(mysql)&: The format of the query is an SQL statement that is passed to a
+The format of the query is an SQL statement that is passed to a
MySQL database. See section &<<SECTsql>>&.
-.next
+
+.subsection nisplus
.cindex "NIS+ lookup type"
.cindex "lookup" "NIS+"
-&(nisplus)&: This does a NIS+ lookup using a query that can specify the name of
+This does a NIS+ lookup using a query that can specify the name of
the field to be returned. See section &<<SECTnisplus>>&.
-.next
+
+.subsection oracle
.cindex "Oracle" "lookup type"
.cindex "lookup" "Oracle"
-&(oracle)&: The format of the query is an SQL statement that is passed to an
+The format of the query is an SQL statement that is passed to an
Oracle database. See section &<<SECTsql>>&.
-.next
+
+.subsection passwd
.cindex "lookup" "passwd"
.cindex "passwd lookup type"
.cindex "&_/etc/passwd_&"
-&(passwd)& is a query-style lookup with queries that are just user names. The
+This is a query-style lookup with queries that are just user names. The
lookup calls &[getpwnam()]& to interrogate the system password data, and on
success, the result string is the same as you would get from an &(lsearch)&
lookup on a traditional &_/etc/passwd file_&, though with &`*`& for the
.code
*:42:42:King Rat:/home/kr:/bin/bash
.endd
-.next
+
+.subsection pgsql
.cindex "PostgreSQL lookup type"
.cindex "lookup" "PostgreSQL"
-&(pgsql)&: The format of the query is an SQL statement that is passed to a
+The format of the query is an SQL statement that is passed to a
PostgreSQL database. See section &<<SECTsql>>&.
-.next
+.subsection redis
.cindex "Redis lookup type"
.cindex lookup Redis
-&(redis)&: The format of the query is either a simple get or simple set,
+The format of the query is either a simple get or simple set,
passed to a Redis database. See section &<<SECTsql>>&.
-.next
+.subsection sqlite
.cindex "sqlite lookup type"
.cindex "lookup" "sqlite"
-&(sqlite)&: The format of the query is
+The format of the query is
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
+.subsection testdb
+This is a lookup type that is used for testing Exim. It is
not likely to be useful in normal operation.
-.next
+
+.subsection whoson
.cindex "whoson lookup type"
.cindex "lookup" "whoson"
-. --- still http:-only, 2018-09-07
-&(whoson)&: &'Whoson'& (&url(http://whoson.sourceforge.net)) is a protocol that
+&'Whoson'& (&url(http://whoson.sourceforge.net)) is a protocol that
allows a server to check whether a particular (dynamically allocated) IP
address is currently allocated to a known (trusted) user and, optionally, to
obtain the identity of the said user. For SMTP servers, &'Whoson'& was popular
the authenticated user, which is stored in the variable &$value$&. However, in
this example, the data in &$value$& is not used; the result of the lookup is
one of the fixed strings &"yes"& or &"no"&.
-.endlist
on the number of entries returned, and no time limit on queries.
When a DN is quoted in the USER= setting for LDAP authentication, Exim
-removes any URL quoting that it may contain before passing it LDAP. Apparently
+removes any URL quoting that it may contain before passing it to the LDAP library.
+Apparently
some libraries do this for themselves, but some do not. Removing the URL
quoting has two advantages:
.code
add_header = :at_start:${authresults {$primary_hostname}}
.endd
-This is safe even if no authentication results are available.
+This is safe even if no authentication results are available
+.new
+and would generally be placed in the DATA ACL.
+.wen
.vitem "&*${certextract{*&<&'field'&>&*}{*&<&'certificate'&>&*}&&&
default, but the separator can be changed in the usual way (&<<SECTlistsepchange>>&).
For each item
in this list, its value is place in &$item$&, and then the condition is
-evaluated. If the condition is true, &$item$& is added to the output as an
+evaluated.
+.new
+Any modification of &$value$& by this evaluation is discarded.
+.wen
+If the condition is true, &$item$& is added to the output as an
item in a new list; if the condition is false, the item is discarded. The
separator used for the output list is the same as the one used for the
input, but a separator setting is not included in the output. For example:
${filter{a:b:c}{!eq{$item}{b}}}
.endd
yields &`a:c`&. At the end of the expansion, the value of &$item$& is restored
-to what it was before. See also the &%map%& and &%reduce%& expansion items.
+to what it was before.
+See also the &%map%& and &%reduce%& expansion items.
.vitem &*${hash{*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&
restored to what they were before. See also the &%filter%& and &%map%&
expansion items.
+. A bit of a special-case logic error in writing an expansion;
+. probably not worth including in the mainline of documentation.
+. If only we had footnotes (the html output variant is the problem).
+.
+. .new
+. &*Note*&: if an &'expansion condition'& is used in <&'string3'&>
+. and that condition modifies &$value$&,
+. then the string expansions dependent on the condition cannot use
+. the &$value$& of the reduce iteration.
+. .wen
+
.vitem &*$rheader_*&<&'header&~name'&>&*:*&&~or&~&*$rh_*&<&'header&~name'&>&*:*&
This item inserts &"raw"& header lines. It is described with the &%header%&
expansion item in section &<<SECTexpansionitems>>& above.
.cindex "first delivery"
.cindex "expansion" "first delivery test"
.cindex "&%first_delivery%& expansion condition"
+.cindex retry condition
This condition, which has no data, is true during a message's first delivery
attempt. It is false during any subsequent delivery attempts.
SRS decode. See SECT &<<SECTSRS>>& for details.
-.vitem &*inlist&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
- &*inlisti&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
+.vitem &*inlist&~{*&<&'subject'&>&*}{*&<&'list'&>&*}*& &&&
+ &*inlisti&~{*&<&'subject'&>&*}{*&<&'list'&>&*}*&
.cindex "string" "comparison"
.cindex "list" "iterative conditions"
Both strings are expanded; the second string is treated as a list of simple
nowadays the ACL specified by &%acl_smtp_connect%& can also reject incoming
connections immediately.
+.new
+If the connection is on a TLS-on-connect port then the TCP connection is
+just dropped. Otherwise, an SMTP error is sent first.
+.wen
+
The ability to give an immediate rejection (either by this option or using an
ACL) is provided for use in unusual cases. Many hosts will just try again,
sometimes without much delay. Normally, it is better to use an ACL to reject
.cindex "banner for SMTP"
.cindex "welcome banner for SMTP"
.cindex "customizing" "SMTP banner"
-This string, which is expanded every time it is used, is output as the initial
+If a connect ACL does not supply a message,
+this string (which is expanded every time it is used) is output as the initial
positive response to an SMTP connection. The default setting is:
.code
smtp_banner = $smtp_active_hostname ESMTP Exim \
$version_number $tod_full
.endd
-Failure to expand the string causes a panic error. If you want to create a
+.new
+Failure to expand the string causes a panic error;
+a forced fail just closes the connection.
+.wen
+If you want to create a
multiline response to the initial SMTP connection, use &"\n"& in the string at
appropriate points, but not at the end. Note that the 220 code is not included
in this string. Exim adds it automatically (several times in the case of a
resent to other recipients.
&*Note:*& If used on a transport handling multiple recipients
-(the smtp transport unless &%rcpt_max%& is 1, the appendfile, pipe or lmtp
+(the smtp transport unless &%max_rcpt%& is 1, the appendfile, pipe or lmtp
transport if &%batch_max%& is greater than 1)
then information about Bcc recipients will be leaked.
Doing so is generally not advised.
&<<CHAPSMTPAUTH>>& for details of authentication.
-.option hosts_request_ocsp smtp "host list&!!" *
+.option hosts_request_ocsp smtp "host list&!!" "see below"
.cindex "TLS" "requiring for certain servers"
Exim will request a Certificate Status on a
TLS session for any host that matches this list.
&%tls_verify_certificates%& should also be set for the transport.
+.new
+The default is &"**"& if DANE is not in use for the connection,
+or if DANE-TA us used.
+It is empty if DANE-EE is used.
+.wen
+
.option hosts_require_alpn smtp "host list&!!" unset
.cindex ALPN "require negotiation in client"
.cindex TLS ALPN
string &`IGNOREQUOTA`& is added to RCPT commands, provided that the LMTP server
has advertised support for IGNOREQUOTA in its response to the LHLO command.
-.option max_rcpt smtp integer 100
+.option max_rcpt smtp integer&!! 100
.cindex "RCPT" "maximum number of outgoing"
-This option limits the number of RCPT commands that are sent in a single
-SMTP message transaction. Each set of addresses is treated independently, and
+This option,
+.new
+after expansion,
+.wen
+limits the number of RCPT commands that are sent in a single
+SMTP message transaction.
+A value setting of zero disables the limit.
+
+.new
+If a constant is given,
+.wen
+each set of addresses is treated independently, and
so can cause parallel connections to the same host if &%remote_max_parallel%&
-permits this. A value setting of zero disables the limit.
+permits this.
.option message_linelength_limit smtp integer 998
operation is as if this option selected all hosts.
&*Warning*&: Including a host in &%tls_verify_hosts%& does not require
that connections use TLS.
-Fallback to in-clear communication will be done unless restricted by
+Fallback to in-clear communication will be done unless restricted by
the &%hosts_require_tls%& option.
.option utf8_downconvert smtp integer&!! -1
(again depending on the &%tls_cipher%& log selector).
-.subsection "Requesting and verifying client certificates" SECID183
+.subsection "Requesting and verifying client certificates"
.cindex "certificate" "verification of client"
.cindex "TLS" "client certificate verification"
If you want an Exim server to request a certificate when negotiating a TLS
certificate is supplied, &$tls_in_peerdn$& is empty.
-.section "Revoked certificates" "SECID184"
-.cindex "TLS" "revoked certificates"
-.cindex "revocation list"
-.cindex "certificate" "revocation list"
-.cindex "OCSP" "stapling"
-Certificate issuing authorities issue Certificate Revocation Lists (CRLs) when
-certificates are revoked. If you have such a list, you can pass it to an Exim
-server using the global option called &%tls_crl%& and to an Exim client using
-an identically named option for the &(smtp)& transport. In each case, the value
-of the option is expanded and must then be the name of a file that contains a
-CRL in PEM format.
-The downside is that clients have to periodically re-download a potentially huge
-file from every certificate authority they know of.
-
-The way with most moving parts at query time is Online Certificate
-Status Protocol (OCSP), where the client verifies the certificate
-against an OCSP server run by the CA. This lets the CA track all
-usage of the certs. It requires running software with access to the
-private key of the CA, to sign the responses to the OCSP queries. OCSP
-is based on HTTP and can be proxied accordingly.
-
-The only widespread OCSP server implementation (known to this writer)
-comes as part of OpenSSL and aborts on an invalid request, such as
-connecting to the port and then disconnecting. This requires
-re-entering the passphrase each time some random client does this.
-
-The third way is OCSP Stapling; in this, the server using a certificate
-issued by the CA periodically requests an OCSP proof of validity from
-the OCSP server, then serves it up inline as part of the TLS
-negotiation. This approach adds no extra round trips, does not let the
-CA track users, scales well with number of certs issued by the CA and is
-resilient to temporary OCSP server failures, as long as the server
-starts retrying to fetch an OCSP proof some time before its current
-proof expires. The downside is that it requires server support.
-
-Unless Exim is built with the support disabled,
-or with GnuTLS earlier than version 3.3.16 / 3.4.8
-support for OCSP stapling is included.
-
-There is a global option called &%tls_ocsp_file%&.
-The file specified therein is expected to be in DER format, and contain
-an OCSP proof. Exim will serve it as part of the TLS handshake. This
-option will be re-expanded for SNI, if the &%tls_certificate%& option
-contains &`tls_in_sni`&, as per other TLS options.
-
-Exim does not at this time implement any support for fetching a new OCSP
-proof. The burden is on the administrator to handle this, outside of
-Exim. The file specified should be replaced atomically, so that the
-contents are always valid. Exim will expand the &%tls_ocsp_file%& option
-on each connection, so a new file will be handled transparently on the
-next connection.
-
-When built with OpenSSL Exim will check for a valid next update timestamp
-in the OCSP proof; if not present, or if the proof has expired, it will be
-ignored.
-
-For the client to be able to verify the stapled OCSP the server must
-also supply, in its stapled information, any intermediate
-certificates for the chain leading to the OCSP proof from the signer
-of the server certificate. There may be zero or one such. These
-intermediate certificates should be added to the server OCSP stapling
-file named by &%tls_ocsp_file%&.
-
-Note that the proof only covers the terminal server certificate,
-not any of the chain from CA to it.
-
-There is no current way to staple a proof for a client certificate.
-
-.code
- A helper script "ocsp_fetch.pl" for fetching a proof from a CA
- OCSP server is supplied. The server URL may be included in the
- server certificate, if the CA is helpful.
-
- One failure mode seen was the OCSP Signer cert expiring before the end
- of validity of the OCSP proof. The checking done by Exim/OpenSSL
- noted this as invalid overall, but the re-fetch script did not.
-.endd
-
-
-.section "Caching of static server configuration items" "SECTserverTLScache"
+.subsection "Caching of static server configuration items" "SSECTserverTLScache"
.cindex certificate caching
.cindex privatekey caching
.cindex crl caching
0.5.10. (Its presence predates the current API which Exim uses, so if Exim
built, then you have SNI support).
+.subsection ALPN
.cindex TLS ALPN
.cindex ALPN "general information"
.cindex TLS "Application Layer Protocol Names"
The client for the connection proposes a set of protocol names, and
the server responds with a selected one.
It is not, as of 2021, commonly used for SMTP connections.
-However, to guard against misirected or malicious use of web clients
+However, to guard against misdirected or malicious use of web clients
(which often do use ALPN) against MTA ports, Exim by default check that
there is no incompatible ALPN specified by a client for a TLS connection.
If there is, the connection is rejected.
&%tls_alpn%& and &%hosts_require_alpn%&.
There are no variables providing observability.
Some feature-specific logging may appear on denied connections, but this
-depends on the behavious of the peer
+depends on the behaviour of the peer
(not all peers can send a feature-specific TLS Alert).
This feature is available when Exim is built with
user certificates, see the &'General implementation overview'& chapter of the
Open-source PKI book, available online at
&url(https://sourceforge.net/projects/ospkibook/).
+
+
+.subsection "Revoked certificates"
+.cindex "TLS" "revoked certificates"
+.cindex "revocation list"
+.cindex "certificate" "revocation list"
+.cindex "OCSP" "stapling"
+There are three ways for a certificate to be made unusable
+before its expiry.
+
+.ilist
+Certificate issuing authorities issue Certificate Revocation Lists (CRLs) when
+certificates are revoked. If you have such a list, you can pass it to an Exim
+server using the global option called &%tls_crl%& and to an Exim client using
+an identically named option for the &(smtp)& transport. In each case, the value
+of the option is expanded and must then be the name of a file that contains a
+CRL in PEM format.
+The downside is that clients have to periodically re-download a potentially huge
+file from every certificate authority they know of.
+
+.next
+The way with most moving parts at query time is Online Certificate
+Status Protocol (OCSP), where the client verifies the certificate
+against an OCSP server run by the CA. This lets the CA track all
+usage of the certs. It requires running software with access to the
+private key of the CA, to sign the responses to the OCSP queries. OCSP
+is based on HTTP and can be proxied accordingly.
+
+The only widespread OCSP server implementation (known to this writer)
+comes as part of OpenSSL and aborts on an invalid request, such as
+connecting to the port and then disconnecting. This requires
+re-entering the passphrase each time some random client does this.
+
+.next
+The third way is OCSP Stapling; in this, the server using a certificate
+issued by the CA periodically requests an OCSP proof of validity from
+the OCSP server, then serves it up inline as part of the TLS
+negotiation. This approach adds no extra round trips, does not let the
+CA track users, scales well with number of certs issued by the CA and is
+resilient to temporary OCSP server failures, as long as the server
+starts retrying to fetch an OCSP proof some time before its current
+proof expires. The downside is that it requires server support.
+
+Unless Exim is built with the support disabled,
+or with GnuTLS earlier than version 3.3.16 / 3.4.8
+support for OCSP stapling is included.
+
+There is a global option called &%tls_ocsp_file%&.
+The file specified therein is expected to be in DER format, and contain
+an OCSP proof. Exim will serve it as part of the TLS handshake. This
+option will be re-expanded for SNI, if the &%tls_certificate%& option
+contains &`tls_in_sni`&, as per other TLS options.
+
+Exim does not at this time implement any support for fetching a new OCSP
+proof. The burden is on the administrator to handle this, outside of
+Exim. The file specified should be replaced atomically, so that the
+contents are always valid. Exim will expand the &%tls_ocsp_file%& option
+on each connection, so a new file will be handled transparently on the
+next connection.
+
+When built with OpenSSL Exim will check for a valid next update timestamp
+in the OCSP proof; if not present, or if the proof has expired, it will be
+ignored.
+
+For the client to be able to verify the stapled OCSP the server must
+also supply, in its stapled information, any intermediate
+certificates for the chain leading to the OCSP proof from the signer
+of the server certificate. There may be zero or one such. These
+intermediate certificates should be added to the server OCSP stapling
+file named by &%tls_ocsp_file%&.
+
+Note that the proof only covers the terminal server certificate,
+not any of the chain from CA to it.
+
+There is no current way to staple a proof for a client certificate.
+
+.code
+ A helper script "ocsp_fetch.pl" for fetching a proof from a CA
+ OCSP server is supplied. The server URL may be included in the
+ server certificate, if the CA is helpful.
+
+ One failure mode seen was the OCSP Signer cert expiring before the end
+ of validity of the OCSP proof. The checking done by Exim/OpenSSL
+ noted this as invalid overall, but the re-fetch script did not.
+.endd
+.endlist
+
+
.ecindex IIDencsmtp1
.ecindex IIDencsmtp2
for every possible target server. It also scales (slightly) better than having to maintain on an SMTP
client a copy of the standard CAs bundle. It also means not having to pay a CA for certificates.
-DANE requires a server operator to do three things: 1) run DNSSEC. This provides assurance to clients
+DANE requires a server operator to do three things:
+.olist
+Run DNSSEC. This provides assurance to clients
that DNS lookups they do for the server have not been tampered with. The domain MX record applying
to this server, its A record, its TLSA record and any associated CNAME records must all be covered by
DNSSEC.
-2) add TLSA DNS records. These say what the server certificate for a TLS connection should be.
-3) offer a server certificate, or certificate chain, in TLS connections which is is anchored by one of the TLSA records.
+.next
+Add TLSA DNS records. These say what the server certificate for a TLS connection should be.
+.next
+Offer a server certificate, or certificate chain, in TLS connections which is is anchored by one of the TLSA records.
+.endlist
There are no changes to Exim specific to server-side operation of DANE.
Support for client-side operation of DANE can be included at compile time by defining SUPPORT_DANE=yes
in &_Local/Makefile_&.
If it has been included, the macro "_HAVE_DANE" will be defined.
+.subsection "DNS records"
A TLSA record consist of 4 fields, the "Certificate Usage", the
"Selector", the "Matching type", and the "Certificate Association Data".
For a detailed description of the TLSA record see
This means no MD5 and no SHA-1. SHA2-256 is the minimum for reliable
interoperability (and probably the maximum too, in 2018).
+.subsection "Interaction with OCSP"
The use of OCSP-stapling should be considered, allowing for fast revocation of certificates (which would otherwise
be limited by the DNS TTL on the TLSA records). However, this is likely to only be usable with DANE-TA. NOTE: the
default of requesting OCSP for all hosts is modified iff DANE is in use, to:
those who use &%hosts_require_ocsp%&, should consider the interaction with DANE in their OCSP settings.
+.subsection "Client configuration"
For client-side DANE there are three new smtp transport options, &%hosts_try_dane%&, &%hosts_require_dane%&
and &%dane_require_tls_ciphers%&.
The &"require"& variant will result in failure if the target host is not
The router and transport option &%dnssec_request_domains%& must not be
set to &"never"&, and &%dnssec_require_domains%& is ignored.
+.subsection Observability
If verification was successful using DANE then the "CV" item in the delivery log line will show as "CV=dane".
There is a new variable &$tls_out_dane$& which will have "yes" if
The &$event_data$& will be one of the Result Types defined in
Section 4.3 of that document.
+.subsection General
Under GnuTLS, DANE is only supported from version 3.0.0 onwards.
DANE is specified in published RFCs and decouples certificate authority trust
selection from a "race to the bottom" of "you must trust everything for mail
-to get through". There is an alternative technology called MTA-STS, which
+to get through".
+There is an alternative technology called MTA-STS, which
instead publishes MX trust anchor information on an HTTPS website. At the
time this text was last updated, MTA-STS was still a draft, not yet an RFC.
Exim has no support for MTA-STS as a client, but Exim mail server operators
the message override the banner message that is otherwise specified by the
&%smtp_banner%& option.
-For tls-on-connect connections, the ACL is run after the TLS connection
-is accepted (however, &%host_reject_connection%& is tested before).
+.new
+For tls-on-connect connections, the ACL is run before the TLS connection
+is accepted; if the ACL does not accept then the TCP connection is dropped without
+any TLS startup attempt and without any SMTP response being transmitted.
+.wen
.subsection "The EHLO/HELO ACL" SECID192
immediate writes to file are done as normal.
trigger=<&'reason'&> This option selects cause for the pretrigger buffer
- see above) to be copied to file. A reason of $*now*
+ see above) to be copied to file. A reason of &*now*&
take effect immediately; one of &*paniclog*& triggers
on a write to the panic log.
.endd
The &%seen%& ACL condition can be used to test whether a
situation has been previously met.
It uses a hints database to record a timestamp against a key.
-host. The syntax of the condition is:
+The syntax of the condition is:
.display
&`seen =`& <&'optional flag'&><&'time interval'&> &`/`& <&'options'&>
.endd
.subsection "Ratelimit options for what is being measured" ratoptmea
.cindex "rate limiting" "per_* options"
-The &%per_conn%& option limits the client's connection rate. It is not
+.vlist
+.vitem per_conn
+.cindex "rate limiting" per_conn
+This option limits the client's connection rate. It is not
normally used in the &%acl_not_smtp%&, &%acl_not_smtp_mime%&, or
&%acl_not_smtp_start%& ACLs.
-The &%per_mail%& option limits the client's rate of sending messages. This is
+.vitem per_mail
+.cindex "rate limiting" per_conn
+This option limits the client's rate of sending messages. This is
the default if none of the &%per_*%& options is specified. It can be used in
&%acl_smtp_mail%&, &%acl_smtp_rcpt%&, &%acl_smtp_predata%&, &%acl_smtp_mime%&,
&%acl_smtp_data%&, or &%acl_not_smtp%&.
-The &%per_byte%& option limits the sender's email bandwidth. It can be used in
+.vitem per_byte
+.cindex "rate limiting" per_conn
+This option limits the sender's email bandwidth. It can be used in
the same ACLs as the &%per_mail%& option, though it is best to use this option
in the &%acl_smtp_mime%&, &%acl_smtp_data%& or &%acl_not_smtp%& ACLs; if it is
used in an earlier ACL, Exim relies on the SIZE parameter given by the client
follow the limit &'m'& in the configuration with K, M, or G to specify limits
in kilobytes, megabytes, or gigabytes, respectively.
-The &%per_rcpt%& option causes Exim to limit the rate at which recipients are
+.vitem per_rcpt
+.cindex "rate limiting" per_rcpt
+This option causes Exim to limit the rate at which recipients are
accepted. It can be used in the &%acl_smtp_rcpt%&, &%acl_smtp_predata%&,
&%acl_smtp_mime%&, or &%acl_smtp_data%& ACLs. In
&%acl_smtp_rcpt%& the rate is updated one recipient at a time; in the other
in either case the rate limiting engine will see a message with many
recipients as a large high-speed burst.
-The &%per_addr%& option is like the &%per_rcpt%& option, except it counts the
+.vitem per_addr
+.cindex "rate limiting" per_addr
+This option is like the &%per_rcpt%& option, except it counts the
number of different recipients that the client has sent messages to in the
last time period. That is, if the client repeatedly sends messages to the same
recipient, its measured rate is not increased. This option can only be used in
&%acl_smtp_rcpt%&.
-The &%per_cmd%& option causes Exim to recompute the rate every time the
+.vitem per_cmd
+.cindex "rate limiting" per_cmd
+This option causes Exim to recompute the rate every time the
condition is processed. This can be used to limit the rate of any SMTP
command. If it is used in multiple ACLs it can limit the aggregate rate of
multiple different commands.
-The &%count=%& option can be used to alter how much Exim adds to the client's
-measured rate. For example, the &%per_byte%& option is equivalent to
-&`per_mail/count=$message_size`&. If there is no &%count=%& option, Exim
+.vitem count
+.cindex "rate limiting" count
+This option can be used to alter how much Exim adds to the client's
+measured rate.
+A value is required, after an equals sign.
+For example, the &%per_byte%& option is equivalent to
+&`per_mail/count=$message_size`&.
+If there is no &%count=%& option, Exim
increases the measured rate by one (except for the &%per_rcpt%& option in ACLs
-other than &%acl_smtp_rcpt%&). The count does not have to be an integer.
+other than &%acl_smtp_rcpt%&).
+The count does not have to be an integer.
-The &%unique=%& option is described in section &<<ratoptuniq>>& below.
+.vitem unique
+.cindex "rate limiting" unique
+This option is described in section &<<ratoptuniq>>& below.
+.endlist
.subsection "Ratelimit update modes" ratoptupd
is done. If a message fails to be completely delivered at the first attempt,
the system filter is run again at the start of every retry.
If you want your filter to do something only once per message, you can make use
+.cindex retry condition
of the &%first_delivery%& condition in an &%if%& command in the filter to
prevent it happening on retries.
.irow &`etrn`& * "ETRN commands"
.irow &`host_lookup_failed`& * "as it says"
.irow &`ident_timeout`& "timeout for ident connection"
-.irow &`incoming_interface`& "local interface on <= and => lines"
+.irow &`incoming_interface`& "local interface & port on <= and => lines"
.irow &`incoming_port`& "remote port on <= lines"
.irow &`lost_incoming_connection`& * "as it says (includes timeouts)"
.irow &`millisec`& "millisecond timestamps and RT,QT,DT,D times"
.subsection ACL SSECDMARCACL
.cindex DMARC "ACL condition"
-DMARC checks cam be run on incoming SMTP messages by using the
+DMARC checks can be run on incoming SMTP messages by using the
&"dmarc_status"& ACL condition in the DATA ACL. You are required to
call the &"spf"& condition first in the ACLs, then the &"dmarc_status"&
condition. Putting this condition in the ACLs is required in order