X-Git-Url: https://git.exim.org/exim.git/blobdiff_plain/9723f9667322bf96db786fa49d53139a48fabc5e..8ac90765750f87c573300b9e953af3d8090cab8b:/doc/doc-docbook/spec.xfpt?ds=sidebyside diff --git a/doc/doc-docbook/spec.xfpt b/doc/doc-docbook/spec.xfpt index e5873f349..d0e3358b8 100644 --- a/doc/doc-docbook/spec.xfpt +++ b/doc/doc-docbook/spec.xfpt @@ -3966,8 +3966,17 @@ the messages are active, their status is not altered. This option can be used only by an admin user or by the user who originally caused the message to be placed on the queue. +. .new +. .vitem &%-MS%& +. .oindex "&%-MS%&" +. .cindex REQUIRETLS +. This option is used to request REQUIRETLS processing on the message. +. It is used internally by Exim in conjunction with -E when generating +. a bounce message. +. .wen + .vitem &%-Mset%&&~<&'message&~id'&> -.oindex "&%-Mset%& +.oindex "&%-Mset%&" .cindex "testing" "string expansion" .cindex "expansion" "testing" This option is useful only in conjunction with &%-be%& (that is, when testing @@ -9434,11 +9443,14 @@ letters appear. For example: &*$h_*&<&'header&~name'&>&*:*&" &&& "&*$bheader_*&<&'header&~name'&>&*:*&&~or&~&&& &*$bh_*&<&'header&~name'&>&*:*&" &&& + "&*$lheader_*&<&'header&~name'&>&*:*&&~or&~&&& + &*$lh_*&<&'header&~name'&>&*:*&" "&*$rheader_*&<&'header&~name'&>&*:*&&~or&~&&& &*$rh_*&<&'header&~name'&>&*:*&" .cindex "expansion" "header insertion" .vindex "&$header_$&" .vindex "&$bheader_$&" +.vindex "&$lheader_$&" .vindex "&$rheader_$&" .cindex "header lines" "in expansion strings" .cindex "header lines" "character sets" @@ -9451,7 +9463,7 @@ The newline that terminates a header line is not included in the expansion, but internal newlines (caused by splitting the header line over several physical lines) may be present. -The difference between &%rheader%&, &%bheader%&, and &%header%& is in the way +The difference between the four pairs of expansions is in the way the data in the header line is interpreted. .ilist @@ -9459,6 +9471,15 @@ the data in the header line is interpreted. &%rheader%& gives the original &"raw"& content of the header line, with no processing at all, and without the removal of leading and trailing white space. +.next +.cindex "list" "of header lines" +&%lheader%& gives a colon-separated list, one element per header when there +are multiple headers with a given name. +Any embedded colon characters within an element are doubled, so normal Exim +list-processing facilities can be used. +The terminating newline of each element is removed; in other respects +the content is &"raw"&. + .next .cindex "base64 encoding" "in header lines" &%bheader%& removes leading and trailing white space, and then decodes base64 @@ -9866,15 +9887,26 @@ extend what can be done. Firstly, you can vary the timeout. For example: .code ${readsocket{/socket/name}{request string}{3s}} .endd + The third argument is a list of options, of which the first element is the timeout and must be present if the argument is given. Further elements are options of form &'name=value'&. -One option type is currently recognised, defining whether (the default) +Two option types is currently recognised: shutdown and tls. +The first defines whether (the default) or not a shutdown is done on the connection after sending the request. Example, to not do so (preferred, eg. by some webservers): .code ${readsocket{/socket/name}{request string}{3s:shutdown=no}} .endd +.new +The second, tls, controls the use of TLS on the connection. Example: +.code +${readsocket{/socket/name}{request string}{3s:tls=yes}} +.endd +The default is to not use TLS. +If it is enabled, a shutdown as descripbed above is never done. +.wen + A fourth argument allows you to change any newlines that are in the data that is read, in the same way as for &%readfile%& (see above). This example turns them into spaces: @@ -10402,7 +10434,7 @@ abbreviation &%h%& can be used when &%hash%& is used as an operator. .cindex "expansion" "hex to base64" .cindex "&%hex2b64%& expansion item" This operator converts a hex string into one that is base64 encoded. This can -be useful for processing the output of the MD5 and SHA-1 hashing functions. +be useful for processing the output of the various hashing functions. @@ -13861,6 +13893,7 @@ listed in more than one group. .row &%av_scanner%& "specify virus scanner" .row &%check_rfc2047_length%& "check length of RFC 2047 &""encoded &&& words""&" +.row &%dns_cname_loops%& "follow CNAMEs returned by resolver" .row &%dns_csa_search_limit%& "control CSA parent search depth" .row &%dns_csa_use_reverse%& "en/disable CSA IP reverse search" .row &%header_maxsize%& "total size of message header" @@ -14763,6 +14796,19 @@ This option controls whether or not an IP address, given as a CSA domain, is reversed and looked up in the reverse DNS, as described in more detail in section &<>&. +.new +.option dns_cname_loops main integer 1 +.cindex DNS "CNAME following" +This option controls the following of CNAME chains, needed if the resolver does +not do it internally. +As of 2018 most should, and the default can be left. +If you have an ancient one, a value of 10 is likely needed. + +The default value of one CNAME-follow is needed +thanks to the observed return for an MX request, +given no MX presence but a CNAME to an A, of the CNAME. +.wen + .option dns_dnssec_ok main integer -1 .cindex "DNS" "resolver options" @@ -15082,7 +15128,7 @@ server. This reduces security slightly, but improves interworking with older implementations of TLS. -option gnutls_allow_auto_pkcs11 main boolean unset +.option gnutls_allow_auto_pkcs11 main boolean unset This option will let GnuTLS (2.12.0 or later) autoload PKCS11 modules with the p11-kit configuration files in &_/etc/pkcs11/modules/_&. @@ -17369,7 +17415,7 @@ The ordering of the two lists must match. .cindex SSMTP .cindex SMTPS This option specifies a list of incoming SSMTP (aka SMTPS) ports that should -operate the obsolete SSMTP (SMTPS) protocol, where a TLS session is immediately +operate the SSMTP (SMTPS) protocol, where a TLS session is immediately set up without waiting for the client to issue a STARTTLS command. For further details, see section &<>&. @@ -23950,14 +23996,15 @@ the message. As a result, the overall timeout for a message depends on the size of the message. Its value must not be zero. See also &%final_timeout%&. +.option dkim_canon smtp string&!! unset .option dkim_domain smtp string list&!! unset -.option dkim_selector smtp string&!! unset +.option dkim_hash smtp string&!! sha256 +.option dkim_identity smtp string&!! unset .option dkim_private_key smtp string&!! unset -.option dkim_canon smtp string&!! unset +.option dkim_selector smtp string&!! unset .option dkim_strict smtp string&!! unset .option dkim_sign_headers smtp string&!! "per RFC" -.option dkim_hash smtp string&!! sha256 -.option dkim_identity smtp string&!! unset +.option dkim_timestamps smtp string&!! unset DKIM signing options. For details see section &<>&. @@ -26742,7 +26789,7 @@ authenticator only. There is only one option: .option server_socket dovecot string unset -This option must specify the socket that is the interface to Dovecot +This option must specify the UNIX socket that is the interface to Dovecot authentication. The &%public_name%& option must specify an authentication mechanism that Dovecot is configured to support. You can have several authenticators for different mechanisms. For example: @@ -26792,15 +26839,17 @@ without code changes in Exim. .option server_channelbinding gsasl boolean false +Do not set this true without consulting a cryptographic engineer. + Some authentication mechanisms are able to use external context at both ends of the session to bind the authentication to that context, and fail the authentication process if that context differs. Specifically, some TLS ciphersuites can provide identifying information about the cryptographic context. -This means that certificate identity and verification becomes a non-issue, -as a man-in-the-middle attack will cause the correct client and server to -see different identifiers and authentication will fail. +This should have meant that certificate identity and verification becomes a +non-issue, as a man-in-the-middle attack will cause the correct client and +server to see different identifiers and authentication will fail. This is currently only supported when using the GnuTLS library. This is only usable by mechanisms which support "channel binding"; at time of @@ -26808,7 +26857,11 @@ writing, that's the SCRAM family. This defaults off to ensure smooth upgrade across Exim releases, in case this option causes some clients to start failing. Some future release -of Exim may switch the default to be true. +of Exim might have switched the default to be true. + +However, Channel Binding in TLS has proven to be broken in current versions. +Do not plan to rely upon this feature for security, ever, without consulting +with a subject matter expert (a cryptographic engineer). .option server_hostname gsasl string&!! "see below" @@ -27390,7 +27443,10 @@ the size of the generated prime, so it might still be too large. .oindex "&%tls_require_ciphers%&" "OpenSSL" There is a function in the OpenSSL library that can be passed a list of cipher suites before the cipher negotiation takes place. This specifies which ciphers -are acceptable. The list is colon separated and may contain names like +.new +are acceptable for TLS versions prior to 1.3. +.wen +The list is colon separated and may contain names like DES-CBC3-SHA. Exim passes the expanded value of &%tls_require_ciphers%& directly to this function call. Many systems will install the OpenSSL manual-pages, so you may have @@ -27455,6 +27511,18 @@ This example will prefer ECDSA-authenticated ciphers over RSA ones: tls_require_ciphers = ECDSA:RSA:!COMPLEMENTOFDEFAULT .endd +.new +For TLS version 1.3 the control available is less fine-grained +and Exim does not provide access to it at present. +The value of the &%tls_require_ciphers%& option is ignored when +TLS version 1.3 is negotiated. + +As of writing the library default cipher suite list for TLSv1.3 is +.code +TLS_AES_256_GCM_SHA384:TLS_CHACHA20_POLY1305_SHA256:TLS_AES_128_GCM_SHA256 +.endd +.wen + .section "Requiring specific ciphers or other parameters in GnuTLS" &&& "SECTreqciphgnu" @@ -28105,20 +28173,48 @@ Support for client-side operation of DANE can be included at compile time by def in &_Local/Makefile_&. If it has been included, the macro "_HAVE_DANE" will be defined. -The TLSA record for the server may have "certificate usage" of DANE-TA(2) or DANE-EE(3). The latter specifies -the End Entity directly, i.e. the certificate involved is that of the server (and should be the sole one transmitted -during the TLS handshake); this is appropriate for a single system, using a self-signed certificate. +The TLSA record for the server may have "certificate usage" of DANE-TA(2) or DANE-EE(3). +These are the "Trust Anchor" and "End Entity" variants. +The latter specifies the End Entity directly, i.e. the certificate involved is that of the server +(and if only DANE-EE is used then it should be the sole one transmitted during the TLS handshake); +this is appropriate for a single system, using a self-signed certificate. DANE-TA usage is effectively declaring a specific CA to be used; this might be a private CA or a public, -well-known one. A private CA at simplest is just a self-signed certificate which is used to sign -cerver certificates, but running one securely does require careful arrangement. If a private CA is used -then either all clients must be primed with it, or (probably simpler) the server TLS handshake must transmit -the entire certificate chain from CA to server-certificate. If a public CA is used then all clients must be primed with it -(losing one advantage of DANE) - but the attack surface is reduced from all public CAs to that single CA. +well-known one. +A private CA at simplest is just a self-signed certificate (with certain +attributes) which is used to sign cerver certificates, but running one securely +does require careful arrangement. +With DANE-TA, as implemented in Exim and commonly in other MTAs, +the server TLS handshake must transmit the entire certificate chain from CA to server-certificate. DANE-TA is commonly used for several services and/or servers, each having a TLSA query-domain CNAME record, all of which point to a single TLSA record. +DANE-TA and DANE-EE can both be used together. -Another approach which should be seriously considered is to use DANE with a certificate -from a public CA, because of another technology, "MTA-STS", described below. +.new +Our recommendation is to use DANE with a certificate from a public CA, +because this enables a variety of strategies for remote clients to verify +your certificate. +You can then publish information both via DANE and another technology, +"MTA-STS", described below. + +When you use DANE-TA to publish trust anchor information, you ask entities +outside your administrative control to trust the Certificate Authority for +connections to you. +If using a private CA then you should expect others to still apply the +technical criteria they'd use for a public CA to your certificates. +In particular, you should probably try to follow current best practices for CA +operation around hash algorithms and key sizes. +Do not expect other organizations to lower their security expectations just +because a particular profile might be reasonable for your own internal use. + +When this text was last updated, this in practice means to avoid use of SHA-1 +and MD5; if using RSA to use key sizes of at least 2048 bits (and no larger +than 4096, for interoperability); to use keyUsage fields correctly; to use +random serial numbers. +The list of requirements is subject to change as best practices evolve. +If you're not already using a private CA, or it doesn't meet these +requirements, then we encourage you to avoid all these issues and use a public +CA such as &url(https://letsencrypt.org/,Let's Encrypt) instead. +.wen The TLSA record should have a Selector field of SPKI(1) and a Matching Type field of SHA2-512(2). @@ -28136,6 +28232,16 @@ are workable for 4th-field hashes. For use with the DANE-TA model, server certificates must have a correct name (SubjectName or SubjectAltName). +.new +The Certificate issued by the CA published in the DANE-TA model should be +issued using a strong hash algorithm. +Exim, and importantly various other MTAs sending to you, will not +re-enable hash algorithms which have been disabled by default in TLS +libraries. +This means no MD5 and no SHA-1. SHA2-256 is the minimum for reliable +interoperability (and probably the maximum too, in 2018). +.wen + 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: @@ -28218,8 +28324,8 @@ MTA-STS to let those clients who do use that protocol derive trust information. The MTA-STS design requires a certificate from a public Certificate Authority -which is recognized by clients sending to you. That selection is outside your -control. +which is recognized by clients sending to you. +That selection of which CAs are trusted by others is outside your control. The most interoperable course of action is probably to use &url(https://letsencrypt.org/,Let's Encrypt), with automated certificate @@ -29850,9 +29956,10 @@ warn hosts = +internal_hosts warn message = Remove internal headers remove_header = $acl_c_ihdrs .endd -Removed header lines are accumulated during the MAIL, RCPT, and predata ACLs. -They are removed from the message before processing the DATA and MIME ACLs. -There is no harm in attempting to remove the same header twice nor is removing +Header names for removal are accumulated during the MAIL, RCPT, and predata ACLs. +Matching header lines are removed from the message before processing the DATA and MIME ACLs. +If multiple header lines match, all are removed. +There is no harm in attempting to remove the same header twice nor in removing a non-existent header. Further header lines to be removed may be accumulated during the DATA and MIME ACLs, after which they are removed from the message, if present. In the case of non-SMTP messages, headers to be removed are @@ -30740,7 +30847,7 @@ restrictions, to get the TXT record. As a byproduct of this, there is also a check that the IP being tested is indeed on the first list. The first domain is the one that is put in &$dnslist_domain$&. For example: .code -reject message = \ +deny message = \ rejected because $sender_host_address is blacklisted \ at $dnslist_domain\n$dnslist_text dnslists = \ @@ -30758,7 +30865,7 @@ If you are interested in more than one merged list, the same list must be given several times, but because the results of the DNS lookups are cached, the DNS calls themselves are not repeated. For example: .code -reject dnslists = \ +deny dnslists = \ http.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.2 : \ socks.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.3 : \ misc.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.4 : \ @@ -30969,13 +31076,15 @@ rest of the ACL. The &%leaky%& (default) option means that the client's recorded rate is not updated if it is above the limit. The effect of this is that Exim measures the -client's average rate of successfully sent email, which cannot be greater than -the maximum allowed. If the client is over the limit it may suffer some -counter-measures (as specified in the ACL), but it will still be able to send -email at the configured maximum rate, whatever the rate of its attempts. This +client's average rate of successfully sent email, +.new +up to the given limit. +This is appropriate if the countermeasure when the condition is true +consists of refusing the message, and is generally the better choice if you have clients that retry automatically. -For example, it does not prevent a sender with an over-aggressive retry rate -from getting any email through. +If the action when true is anything more complex then this option is +likely not what is wanted. +.wen The &%strict%& option means that the client's recorded rate is always updated. The effect of this is that Exim measures the client's average rate @@ -31146,6 +31255,15 @@ connection, HELO, or MAIL). The main use of these variables is expected to be to distinguish between rejections of MAIL and rejections of RCPT in callouts. +.new +The above variables may also be set after a &*successful*& +address verification to: + +.ilist +&%random%&: A random local-part callout succeeded +.endlist +.wen + @@ -39035,7 +39153,7 @@ tag value. Note that Exim does not check the value. This option sets the canonicalization method used when signing a message. The DKIM RFC currently supports two methods: "simple" and "relaxed". The option defaults to "relaxed" when unset. Note: the current implementation -only supports using the same canonicalization method for both headers and body. +only supports signing with the same canonicalization method for both headers and body. .option dkim_strict smtp string&!! unset This option defines how Exim behaves when signing a message that @@ -39065,26 +39183,44 @@ If a '+' prefix if used, all headers that are present with this name will be signed, and one signature added for a missing header with the name will be appended. +.new +.option dkim_timestamps smtp integer&!! unset +This option controls the inclusion of timestamp information in the signature. +If not set, no such information will be included. +Otherwise, must be an unsigned number giving an offset in seconds from the current time +for the expiry tag +(eg. 1209600 for two weeks); +both creation (t=) and expiry (x=) tags will be included. + +RFC 6376 lists these tags as RECOMMENDED. +.wen + .section "Verifying DKIM signatures in incoming mail" "SECDKIMVFY" .cindex "DKIM" "verification" -Verification of DKIM signatures in SMTP incoming email is implemented via the -&%acl_smtp_dkim%& ACL. By default, this ACL is called once for each +.new +Verification of DKIM signatures in SMTP incoming email is done for all +messages for which an ACL control &%dkim_disable_verify%& has not been set. +.cindex authentication "expansion item" +Performing verification sets up information used by the +&$authresults$& expansion item. +.wen + +.new The results of that verification are then made available to the +&%acl_smtp_dkim%& ACL, &new(which can examine and modify them). +By default, this ACL is called once for each syntactically(!) correct signature in the incoming message. A missing ACL definition defaults to accept. If any ACL call does not accept, the message is not accepted. If a cutthrough delivery was in progress for the message, that is summarily dropped (having wasted the transmission effort). -To evaluate the signature in the ACL a large number of expansion variables +To evaluate the &new(verification result) in the ACL +a large number of expansion variables containing the signature status and its details are set up during the runtime of the ACL. -.cindex authentication "expansion item" -Performing verification sets up information used by the -&$authresults$& expansion item. - Calling the ACL only for existing signatures is not sufficient to build more advanced policies. For that reason, the global option &%dkim_verify_signers%&, and a global expansion variable @@ -39159,8 +39295,10 @@ hash-method or key-size: set dkim_verify_reason = hash too weak or key too short .endd -After all the DKIM ACL runs have completed, the value becomes a +So long as a DKIM ACL is defined (it need do no more than accept), +after all the DKIM ACL runs have completed, the value becomes a colon-separated list of the values after each run. +This is maintained for the mime, prdr and data ACLs. .vitem &%$dkim_verify_reason%& A string giving a little bit more detail when &%$dkim_verify_status%& is either @@ -39232,6 +39370,12 @@ strict enforcement should code the check explicitly. The number of signed body bytes. If zero ("0"), the body is unsigned. If no limit was set by the signer, "9999999999999" is returned. This makes sure that this variable always expands to an integer value. +.new +&*Note:*& The presence of the signature tag specifying a signing body length +is one possible route to spoofing of valid DKIM signatures. +A paranoid implementation might wish to regard signature where this variable +shows less than the "no limit" return as being invalid. +.wen .vitem &%$dkim_created%& UNIX timestamp reflecting the date and time when the signature was created.