X-Git-Url: https://git.exim.org/users/jgh/exim.git/blobdiff_plain/7c714e25526fdd15330fc1c852845d3632e29823..e5c155710bc95053e6e4b3845b79eb8fc281f769:/doc/doc-docbook/spec.xfpt diff --git a/doc/doc-docbook/spec.xfpt b/doc/doc-docbook/spec.xfpt index e237ce1f2..f455c9f0c 100644 --- a/doc/doc-docbook/spec.xfpt +++ b/doc/doc-docbook/spec.xfpt @@ -52,7 +52,7 @@ .set I "    " .macro copyyear -2017 +2018 .endmacro . ///////////////////////////////////////////////////////////////////////////// @@ -7839,6 +7839,19 @@ ${lookup redis{set keyname ${quote_redis:objvalue plus}}} ${lookup redis{get keyname}} .endd +.new +As of release 4.91, "lightweight" support for Redis Cluster is available. +Requires &%redis_servers%& list to contain all the servers in the cluster, all +of which must be reachable from the running exim instance. If the cluster has +master/slave replication, the list must contain all the master and slave +servers. + +When the Redis Cluster returns a "MOVED" response to a query, exim does not +immediately follow the redirection but treats the response as a DEFER, moving on +to the next server in the &%redis_servers%& list until the correct server is +reached. +.wen + .ecindex IIDfidalo1 .ecindex IIDfidalo2 @@ -9137,6 +9150,33 @@ the expansion result is an empty string. If the ACL returns defer the result is a forced-fail. Otherwise the expansion fails. +.new +.vitem "&*${authresults{*&<&'authserv-id'&>&*}}*&" +.cindex authentication "results header" +.cindex headers "authentication-results:" +.cindex authentication "expansion item" +This item returns a string suitable for insertion as an +&'Authentication-Results"'& +header line. +The given <&'authserv-id'&> is included in the result; typically this +will ba a domain name identifying the system performing the authentications. +Methods that may be present in the result include: +.code +none +iprev +auth +spf +dkim +.endd + +Example use (as an ACL modifier): +.code + add_header = :at_start:${authresults {$primary_hostname}} +.endd +This is safe even if no authentication reselts are available. +.wen + + .vitem "&*${certextract{*&<&'field'&>&*}{*&<&'certificate'&>&*}&&& {*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&" .cindex "expansion" "extracting certificate fields" @@ -11555,10 +11595,15 @@ preserve some of the authentication information in the variable user/password authenticator configuration might preserve the user name for use in the routers. Note that this is not the same information that is saved in &$sender_host_authenticated$&. + When a message is submitted locally (that is, not over a TCP connection) the value of &$authenticated_id$& is normally the login name of the calling process. However, a trusted user can override this by means of the &%-oMai%& command line option. +.new +This second case also sets up inforamtion used by the +&$authresults$& expansion item. +.wen .vitem &$authenticated_fail_id$& .cindex "authentication" "fail" "id" @@ -11898,6 +11943,13 @@ lookup succeeds, but there is a lookup problem such as a timeout when checking the result, the name is not accepted, and &$host_lookup_deferred$& is set to &"1"&. See also &$sender_host_name$&. +.new +.cindex authentication "expansion item" +Performing these checks sets up information used by the +&$authresults$& expansion item. +.wen + + .vitem &$host_lookup_failed$& .vindex "&$host_lookup_failed$&" See &$host_lookup_deferred$&. @@ -12853,6 +12905,7 @@ is compiled with the content-scanning extension. For details, see section .vitem &$spf_header_comment$& &&& &$spf_received$& &&& &$spf_result$& &&& + &$spf_result_guessed$& &&& &$spf_smtp_comment$& These variables are only available if Exim is built with SPF support. For details see section &<>&. @@ -12985,6 +13038,10 @@ and then set to the outgoing cipher suite if one is negotiated. See chapter &<>& for details of TLS support and chapter &<>& for details of the &(smtp)& transport. +.vitem &$tls_out_dane$& +.vindex &$tls_out_dane$& +DANE active status. See section &<>&. + .vitem &$tls_in_ocsp$& .vindex "&$tls_in_ocsp$&" When a message is received from a remote client connection @@ -13050,6 +13107,10 @@ During outbound SMTP deliveries, this variable reflects the value of the &%tls_sni%& option on the transport. +.vitem &$tls_out_tlsa_usage$& +.vindex &$tls_out_tlsa_usage$& +Bitfield of TLSA record types found. See section &<>&. + .vitem &$tod_bsdinbox$& .vindex "&$tod_bsdinbox$&" The time of day and the date, in the format required for BSD-style mailbox @@ -23880,6 +23941,24 @@ For testing purposes, this value can be overridden by the &%-oB%& command line option. +.new +.option dane_require_tls_ciphers smtp string&!! unset +.cindex "TLS" "requiring specific ciphers for DANE" +.cindex "cipher" "requiring specific" +.cindex DANE "TLS ciphers" +This option may be used to override &%tls_require_ciphers%& for connections +where DANE has been determined to be in effect. +If not set, then &%tls_require_ciphers%& will be used. +Normal SMTP delivery is not able to make strong demands of TLS cipher +configuration, because delivery will fall back to plaintext. Once DANE has +been determined to be in effect, there is no plaintext fallback and making the +TLS cipherlist configuration stronger will increase security, rather than +counter-intuitively decreasing it. +If the option expands to be empty or is forced to fail, then it will +be treated as unset and &%tls_require_ciphers%& will be used instead. +.wen + + .option data_timeout smtp time 5m This sets a timeout for the transmission of each block in the data portion of the message. As a result, the overall timeout for a message depends on the size @@ -24201,6 +24280,17 @@ 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 +.option hosts_require_dane smtp "host list&!!" unset +.cindex DANE "transport options" +.cindex DANE "requiring for certain servers" +If built with DANE support, Exim will require that a DNSSEC-validated +TLSA record is present for any host matching the list, +and that a DANE-verified TLS connection is made. +There will be no fallback to in-clear communication. +See section &<>&. +.wen + .option hosts_require_ocsp smtp "host list&!!" unset .cindex "TLS" "requiring for certain servers" Exim will request, and check for a valid Certificate Status being given, on a @@ -24230,6 +24320,18 @@ This option provides a list of servers to which, provided they announce CHUNKING support, Exim will attempt to use BDAT commands rather than DATA. BDAT will not be used in conjunction with a transport filter. +.new +.option hosts_try_dane smtp "host list&!!" unset +.cindex DANE "transport options" +.cindex DANE "attempting for certain servers" +If built with DANE support, Exim will lookup a +TLSA record for any host matching the list. +If found and verified by DNSSEC, +a DANE-verified TLS connection is made to that host; +there will be no fallback to in-clear communication. +See section &<>&. +.wen + .option hosts_try_fastopen smtp "host list&!!" unset .cindex "fast open, TCP" "enabling, in client" .cindex "TCP Fast Open" "enabling, in client" @@ -26035,6 +26137,12 @@ public name) of the authenticator driver that successfully authenticated the client from which the message was received. This variable is empty if there was no successful authentication. +.new +.cindex authentication "expansion item" +Successful authentication sets up information used by the +&$authresults$& expansion item. +.wen + @@ -27986,6 +28094,141 @@ Open-source PKI book, available online at +.new +.section DANE "SECDANE" +.cindex DANE +DNS-based Authentication of Named Entities, as applied to SMTP over TLS, provides assurance to a client that +it is actually talking to the server it wants to rather than some attacker operating a Man In The Middle (MITM) +operation. The latter can terminate the TLS connection you make, and make another one to the server (so both +you and the server still think you have an encrypted connection) and, if one of the "well known" set of +Certificate Authorities has been suborned - something which *has* been seen already (2014), a verifiable +certificate (if you're using normal root CAs, eg. the Mozilla set, as your trust anchors). + +What DANE does is replace the CAs with the DNS as the trust anchor. The assurance is limited to a) the possibility +that the DNS has been suborned, b) mistakes made by the admins of the target server. The attack surface presented +by (a) is thought to be smaller than that of the set of root CAs. + +It also allows the server to declare (implicitly) that connections to it should use TLS. An MITM could simply +fail to pass on a server's STARTTLS. + +DANE scales better than having to maintain (and side-channel communicate) copies of server certificates +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 +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. + +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. + +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. +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. +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. + +The TLSA record should have a Selector field of SPKI(1) and a Matching Type field of SHA2-512(2). + +At the time of writing, &url(https://www.huque.com/bin/gen_tlsa) +is useful for quickly generating TLSA records; and commands like + +.code + openssl x509 -in -pubkey -noout /dev/null \ + | openssl sha512 \ + | awk '{print $2}' +.endd + +are workable for 4th-field hashes. + +For use with the DANE-TA model, server certificates must have a correct name (SubjectName or SubjectAltName). + +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: + +.code + hosts_request_ocsp = ${if or { {= {0}{$tls_out_tlsa_usage}} \ + {= {4}{$tls_out_tlsa_usage}} } \ + {*}{}} +.endd + +The (new) variable &$tls_out_tlsa_usage$& is a bitfield with numbered bits set for TLSA record usage codes. +The zero above means DANE was not in use, the four means that only DANE-TA usage TLSA records were +found. If the definition of &%hosts_request_ocsp%& includes the +string "tls_out_tlsa_usage", they are re-expanded in time to +control the OCSP request. + +This modification of hosts_request_ocsp is only done if it has the default value of "*". Admins who change it, and +those who use &%hosts_require_ocsp%&, should consider the interaction with DANE in their OCSP settings. + + +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 DNSSEC-secured. + +DANE will only be usable if the target host has DNSSEC-secured MX, A and TLSA records. + +A TLSA lookup will be done if either of the above options match and the host-lookup succeeded using dnssec. +If a TLSA lookup is done and succeeds, a DANE-verified TLS connection +will be required for the host. If it does not, the host will not +be used; there is no fallback to non-DANE or non-TLS. + +If DANE is requested and usable, then the TLS cipher list configuration +prefers to use the option &%dane_require_tls_ciphers%& and falls +back to &%tls_require_ciphers%& only if that is unset. +This lets you configure "decent crypto" for DANE and "better than nothing +crypto" as the default. Note though that while GnuTLS lets the string control +which versions of TLS/SSL will be negotiated, OpenSSL does not and you're +limited to ciphersuite constraints. + +If DANE is requested and useable (see above) the following transport options are ignored: +.code + hosts_require_tls + tls_verify_hosts + tls_try_verify_hosts + tls_verify_certificates + tls_crl + tls_verify_cert_hostnames +.endd + +If DANE is not usable, whether requested or not, and CA-anchored +verification evaluation is wanted, the above variables should be set appropriately. + +Currently the &%dnssec_request_domains%& must be active and &%dnssec_require_domains%& is ignored. + +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$& which will have "yes" if +verification succeeded using DANE and "no" otherwise (only useful +in combination with events; see &<>&), +and a new variable &$tls_out_tlsa_usage$& (detailed above). + +.cindex DANE reporting +An event (see &<>&) of type "dane:fail" will be raised on failures +to achieve DANE-verified connection, if one was either requested and offered, or +required. This is intended to support TLS-reporting as defined in +&url(https://tools.ietf.org/html/draft-ietf-uta-smtp-tlsrpt-17). +The &$event_data$& will be one of the Result Types defined in +Section 4.3 of that document. + +Under GnuTLS, DANE is only supported from version 3.0.0 onwards. +.wen + + + . //////////////////////////////////////////////////////////////////////////// . //////////////////////////////////////////////////////////////////////////// @@ -31615,27 +31858,36 @@ If the value of &%av_scanner%& starts with a dollar character, it is expanded before use. The usual list-parsing of the content (see &<>&) applies. The following scanner types are supported in this release, -.new though individual ones can be included or not at build time: -.wen .vlist .vitem &%avast%& .cindex "virus scanners" "avast" This is the scanner daemon of Avast. It has been tested with Avast Core -Security (currently at version 1.1.7). -You can get a trial version at &url(http://www.avast.com) or for Linux -at &url(http://www.avast.com/linux-server-antivirus). +Security (currently at version 2.2.0). +You can get a trial version at &url(https://www.avast.com) or for Linux +at &url(https://www.avast.com/linux-server-antivirus). This scanner type takes one option, which can be either a full path to a UNIX socket, or host and port specifiers separated by white space. The host may be a name or an IP address; the port is either a single number or a pair of numbers with a dash between. -Any further options are given, on separate lines, -to the daemon as options before the main scan command. +A list of options may follow. These options are interpreted on the +Exim's side of the malware scanner, or are given on separate lines to +the daemon as options before the main scan command. + +.new +.cindex &`pass_unscanned`& "avast" +If &`pass_unscanned`& +is set, any files the Avast scanner can't scan (e.g. +decompression bombs, or invalid archives) are considered clean. Use with +care. +.wen + For example: .code av_scanner = avast:/var/run/avast/scan.sock:FLAGS -fullfiles:SENSITIVITY -pup +av_scanner = avast:/var/run/avast/scan.sock:pass_unscanned:FLAGS -fullfiles:SENSITIVITY -pup av_scanner = avast:192.168.2.22 5036 .endd If you omit the argument, the default path @@ -31652,6 +31904,9 @@ $ socat UNIX:/var/run/avast/scan.sock STDIO: PACK .endd +If the scanner returns a temporary failure (e.g. license issues, or +permission problems), the message is deferred and a paniclog entry is +written. The usual &`defer_ok`& option is available. .vitem &%aveserver%& .cindex "virus scanners" "Kaspersky" @@ -31702,7 +31957,7 @@ av_scanner = clamd:192.0.2.3 1234 : 192.0.2.4 1234 If the value of av_scanner points to a UNIX socket file or contains the &`local`& option, then the ClamAV interface will pass a filename containing the data -to be scanned, which will should normally result in less I/O happening and be +to be scanned, which should normally result in less I/O happening and be more efficient. Normally in the TCP case, the data is streamed to ClamAV as Exim does not assume that there is a common filesystem with the remote host. @@ -36627,9 +36882,16 @@ specifies whether characters with values greater than 127 should be logged unchanged, or whether they should be rendered as escape sequences. .next .cindex "log" "certificate verification" +.cindex log DANE +.cindex DANE logging &%tls_certificate_verified%&: An extra item is added to <= and => log lines when TLS is in use. The item is &`CV=yes`& if the peer's certificate was -verified, and &`CV=no`& if not. +verified +.new +using a CA trust anchor, +&`CA=dane`& if using a DNS trust anchor, +.wen +and &`CV=no`& if not. .next .cindex "log" "TLS cipher" .cindex "TLS" "logging cipher" @@ -38698,7 +38960,8 @@ The result can either be a valid RSA private key in ASCII armor (.pem file), including line breaks .new .next -with GnuTLS 3.6.0 or later, be a valid Ed25519 private key (same format as above) +with GnuTLS 3.6.0 or OpenSSL 1.1.1 or later, +be a valid Ed25519 private key (same format as above) .wen .next start with a slash, in which case it is treated as a file that contains @@ -38710,6 +38973,21 @@ is set. .endlist .new +To generate keys under OpenSSL: +.code +openssl genrsa -out dkim_rsa.private 2048 +openssl rsa -in dkim_rsa.private -out /dev/stdout -pubout -outform PEM +.endd +Take the base-64 lines from the output of the second command, concatenated, +for the DNS TXT record. +See section 3.6 of RFC6376 for the record specification. + +Under GnuTLS: +.code +certtool --generate-privkey --rsa --bits=2048 --password='' -8 --outfile=dkim_rsa.private +certtool --load-privkey=dkim_rsa.private --pubkey-info +.endd + Note that RFC 8301 says: .code Signers MUST use RSA keys of at least 1024 bits for all keys. @@ -38724,6 +39002,18 @@ As they are a recent development, users should consider dual-signing for some transition period. The "_CRYPTO_SIGN_ED25519" macro will be defined if support is present for EC keys. + +OpenSSL 1.1.1 and GnuTLS 3.6.0 can create Ed25519 private keys: +.code +openssl genpkey -algorithm ed25519 -out dkim_ed25519.private +certtool --generate-privkey --key-type=ed25519 --outfile=dkim_ed25519.private +.endd + +To produce the required public key value for a DNS record: +.code +openssl pkey -outform DER -pubout -in dkim_ed25519.private | tail -c +13 | base64 +certtool --load_privkey=dkim_ed25519.private --pubkey_info --outder | tail -c +13 | base64 +.endd .wen .option dkim_hash smtp string&!! sha256 @@ -38794,6 +39084,12 @@ To evaluate the signature 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. +.new +.cindex authentication "expansion item" +Performing verification sets up information used by the +&$authresults$& expansion item. +.wen + 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 @@ -38859,11 +39155,13 @@ This variable can be overwritten using an ACL 'set' modifier. This might, for instance, be done to enforce a policy restriction on hash-method or key-size: .code - warn condition = ${if eq {$dkim_algo}{rsa-sha1}} - condition = ${if eq {$dkim_verify_status}{pass}} + warn condition = ${if eq {$dkim_verify_status}{pass}} + condition = ${if eq {$len_3:$dkim_algo}{rsa}} + condition = ${if or {eq {$dkim_algo}{rsa-sha1}} \ + {< {$dkim_key_length}{1024}} } logwrite = NOTE: forcing dkim verify fail (was pass) set dkim_verify_status = fail - set dkim_verify_reason = hash too weak + set dkim_verify_reason = hash too weak or key too short .endd After all the DKIM ACL runs have completed, the value becomes a @@ -38907,7 +39205,8 @@ The key record selector string. .vitem &%$dkim_algo%& The algorithm used. One of 'rsa-sha1' or 'rsa-sha256'. .new -If running under GnuTLS 3.6.0 or later, may also be 'ed25519-sha256'. +If running under GnuTLS 3.6.0 or OpenSSL 1.1.1 or later, +may also be 'ed25519-sha256'. The "_CRYPTO_SIGN_ED25519" macro will be defined if support is present for EC keys. .wen @@ -38952,6 +39251,9 @@ UNIX timestamp reflecting the date and time when the signer wants the signature to be treated as "expired". When this was not specified by the signer, "9999999999999" is returned. This makes it possible to do useful integer size comparisons against this value. +.new +Note that Exim does not check this value. +.wen .vitem &%$dkim_headernames%& A colon-separated list of names of headers included in the signature. @@ -38985,6 +39287,7 @@ less than 1024 bits as valid signatures. To enforce this you must have a DKIM ACL which checks this variable and overwrites the &$dkim_verify_status$& variable as discussed above. +As EC keys are much smaller, the check should only do this for RSA keys. .wen .endlist @@ -39047,6 +39350,12 @@ There is no Exim involvement on the trasmission of messages; publishing certain DNS records is all that is required. For verification, an ACL condition and an expansion lookup are provided. +.new +.cindex authentication "expansion item" +Performing verification sets up information used by the +&$authresults$& expansion item. +.wen + .cindex SPF "ACL condition" .cindex ACL "spf condition" @@ -39076,18 +39385,11 @@ its domain as well. This should be treated like "none". .vitem &%permerror%& This indicates a syntax error in the SPF record of the queried domain. -You may deny messages when this occurs. (Changed in 4.83) +You may deny messages when this occurs. .vitem &%temperror%& This indicates a temporary error during all processing, including Exim's SPF processing. You may defer messages when this occurs. -(Changed in 4.83) - -.vitem &%err_temp%& -Same as permerror, deprecated in 4.83, will be removed in a future release. - -.vitem &%err_perm%& -Same as temperror, deprecated in 4.83, will be removed in a future release. .endlist You can prefix each string with an exclamation mark to invert @@ -39135,6 +39437,11 @@ variables: one of pass, fail, softfail, none, neutral, permerror or temperror. +.vitem &$spf_result_guessed$& +.vindex &$spf_result_guessed$& + This boolean is true only if a best-guess operation was used + and required in order to obtain a result. + .vitem &$spf_smtp_comment$& .vindex &$spf_smtp_comment$& This contains a string that can be used in a SMTP response @@ -39522,6 +39829,7 @@ expansion must check this, as it will be called for every possible event type. The current list of events is: .display +&`dane:fail after transport `& per connection &`msg:complete after main `& per message &`msg:delivery after transport `& per recipient &`msg:rcpt:host:defer after transport `& per recipient per host @@ -39550,6 +39858,7 @@ should define the event action. An additional variable, &$event_data$&, is filled with information varying with the event type: .display +&`dane:fail `& failure reason &`msg:delivery `& smtp confirmation message &`msg:rcpt:host:defer `& error string &`msg:rcpt:defer `& error string @@ -39577,15 +39886,12 @@ The expansion of the event_action option should normally return an empty string. Should it return anything else the following will be forced: .display -&`msg:delivery `& (ignored) -&`msg:host:defer `& (ignored) -&`msg:fail:delivery`& (ignored) &`tcp:connect `& do not connect -&`tcp:close `& (ignored) &`tls:cert `& refuse verification &`smtp:connect `& close connection .endd -No other use is made of the result string. +All other message types ignore the result string, and +no other use is made of it. For a tcp:connect event, if the connection is being made to a proxy then the address and port variables will be that of the proxy and not