X-Git-Url: https://git.exim.org/users/heiko/exim.git/blobdiff_plain/d36a05018e1ac918951d551450acc53137ecf6e0..c8b050fd148ef06666c1f6feaa492d122f65e23e:/doc/doc-txt/experimental-spec.txt diff --git a/doc/doc-txt/experimental-spec.txt b/doc/doc-txt/experimental-spec.txt index 7bb17883c..2569ad3af 100644 --- a/doc/doc-txt/experimental-spec.txt +++ b/doc/doc-txt/experimental-spec.txt @@ -6,67 +6,7 @@ about experimental features, all of which are unstable and liable to incompatible change. -OCSP Stapling support --------------------------------------------------------------- - -X.509 PKI certificates expire and can be revoked; to handle this, the -clients need some way to determine if a particular certificate, from a -particular Certificate Authority (CA), is still valid. There are three -main ways to do so. - -The simplest way is to serve up a Certificate Revocation List (CRL) with -an ordinary web-server, regenerating the CRL before it expires. The -downside is that clients have to periodically re-download a potentially -huge file from every certificate authority it knows 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. This 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. - -If Exim is built with EXPERIMENTAL_OCSP and it was built with OpenSSL, -then it gains one new option: "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_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. - -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. - -At this point in time, we're gathering feedback on use, to determine if -it's worth adding complexity to the Exim daemon to periodically re-fetch -OCSP files and somehow handling multiple files. There is no client support -for OCSP in Exim, this is feature expected to be used by mail clients. - - - - -Brightmail AntiSpam (BMI) suppport +Brightmail AntiSpam (BMI) support -------------------------------------------------------------- Brightmail AntiSpam is a commercial package. Please see @@ -102,7 +42,7 @@ These four steps are explained in more details below. 1) Adding support for BMI at compile time To compile with BMI support, you need to link Exim against - the Brighmail client SDK, consisting of a library + the Brightmail client SDK, consisting of a library (libbmiclient_single.so) and a header file (bmi_api.h). You'll also need to explicitly set a flag in the Makefile to include BMI support in the Exim binary. Both can be achieved @@ -352,166 +292,19 @@ These four steps are explained in more details below. -Sender Policy Framework (SPF) support --------------------------------------------------------------- - -To learn more about SPF, visit http://www.openspf.org. This -document does not explain the SPF fundamentals, you should -read and understand the implications of deploying SPF on your -system before doing so. - -SPF support is added via the libspf2 library. Visit - - http://www.libspf2.org/ - -to obtain a copy, then compile and install it. By default, -this will put headers in /usr/local/include and the static -library in /usr/local/lib. - -To compile Exim with SPF support, set these additional flags in -Local/Makefile: - -EXPERIMENTAL_SPF=yes -CFLAGS=-DSPF -I/usr/local/include -EXTRALIBS_EXIM=-L/usr/local/lib -lspf2 - -This assumes that the libspf2 files are installed in -their default locations. - -You can now run SPF checks in incoming SMTP by using the "spf" -ACL condition in either the MAIL, RCPT or DATA ACLs. When -using it in the RCPT ACL, you can make the checks dependent on -the RCPT address (or domain), so you can check SPF records -only for certain target domains. This gives you the -possibility to opt-out certain customers that do not want -their mail to be subject to SPF checking. - -The spf condition takes a list of strings on its right-hand -side. These strings describe the outcome of the SPF check for -which the spf condition should succeed. Valid strings are: - - o pass The SPF check passed, the sending host - is positively verified by SPF. - o fail The SPF check failed, the sending host - is NOT allowed to send mail for the domain - in the envelope-from address. - o softfail The SPF check failed, but the queried - domain can't absolutely confirm that this - is a forgery. - o none The queried domain does not publish SPF - records. - o neutral The SPF check returned a "neutral" state. - This means the queried domain has published - a SPF record, but wants to allow outside - servers to send mail under its domain as well. - o err_perm This indicates a syntax error in the SPF - record of the queried domain. This should be - treated like "none". - o err_temp This indicates a temporary error during all - processing, including Exim's SPF processing. - You may defer messages when this occurs. - -You can prefix each string with an exclamation mark to invert -is meaning, for example "!fail" will match all results but -"fail". The string list is evaluated left-to-right, in a -short-circuit fashion. When a string matches the outcome of -the SPF check, the condition succeeds. If none of the listed -strings matches the outcome of the SPF check, the condition -fails. - -Here is an example to fail forgery attempts from domains that -publish SPF records: - -/* ----------------- -deny message = $sender_host_address is not allowed to send mail from ${if def:sender_address_domain {$sender_address_domain}{$sender_helo_name}}. \ - Please see http://www.openspf.org/Why?scope=${if def:sender_address_domain {mfrom}{helo}};identity=${if def:sender_address_domain {$sender_address}{$sender_helo_name}};ip=$sender_host_address - spf = fail ---------------------- */ - -You can also give special treatment to specific domains: - -/* ----------------- -deny message = AOL sender, but not from AOL-approved relay. - sender_domains = aol.com - spf = fail:neutral ---------------------- */ - -Explanation: AOL publishes SPF records, but is liberal and -still allows non-approved relays to send mail from aol.com. -This will result in a "neutral" state, while mail from genuine -AOL servers will result in "pass". The example above takes -this into account and treats "neutral" like "fail", but only -for aol.com. Please note that this violates the SPF draft. - -When the spf condition has run, it sets up several expansion -variables. - - $spf_header_comment - This contains a human-readable string describing the outcome - of the SPF check. You can add it to a custom header or use - it for logging purposes. - - $spf_received - This contains a complete Received-SPF: header that can be - added to the message. Please note that according to the SPF - draft, this header must be added at the top of the header - list. Please see section 10 on how you can do this. - - Note: in case of "Best-guess" (see below), the convention is - to put this string in a header called X-SPF-Guess: instead. - - $spf_result - This contains the outcome of the SPF check in string form, - one of pass, fail, softfail, none, neutral, err_perm or - err_temp. - - $spf_smtp_comment - This contains a string that can be used in a SMTP response - to the calling party. Useful for "fail". - -In addition to SPF, you can also perform checks for so-called -"Best-guess". Strictly speaking, "Best-guess" is not standard -SPF, but it is supported by the same framework that enables SPF -capability. Refer to http://www.openspf.org/FAQ/Best_guess_record -for a description of what it means. - -To access this feature, simply use the spf_guess condition in place -of the spf one. For example: - -/* ----------------- -deny message = $sender_host_address doesn't look trustworthy to me - spf_guess = fail ---------------------- */ - -In case you decide to reject messages based on this check, you -should note that although it uses the same framework, "Best-guess" -is NOT SPF, and therefore you should not mention SPF at all in your -reject message. - -When the spf_guess condition has run, it sets up the same expansion -variables as when spf condition is run, described above. - -Additionally, since Best-guess is not standardized, you may redefine -what "Best-guess" means to you by redefining spf_guess variable in -global config. For example, the following: - -/* ----------------- -spf_guess = v=spf1 a/16 mx/16 ptr ?all ---------------------- */ - -would relax host matching rules to a broader network range. - - -SRS (Sender Rewriting Scheme) Support +SRS (Sender Rewriting Scheme) Support (using libsrs_alt) -------------------------------------------------------------- +See also below, for an alternative native support implementation. -Exiscan currently includes SRS support via Miles Wilton's +Exim currently includes SRS support via Miles Wilton's libsrs_alt library. The current version of the supported -library is 0.5. +library is 0.5, there are reports of 1.0 working. In order to use SRS, you must get a copy of libsrs_alt from -http://srs.mirtol.com/ +https://opsec.eu/src/srs/ + +(not the original source, which has disappeared.) Unpack the tarball, then refer to MTAs/README.EXIM to proceed. You need to set @@ -520,9 +313,110 @@ EXPERIMENTAL_SRS=yes in your Local/Makefile. +The following main-section options become available: + srs_config string + srs_hashlength int + srs_hashmin int + srs_maxage int + srs_secrets string + srs_usehash bool + srs_usetimestamp bool + +The redirect router gains these options (all of type string, unset by default): + srs + srs_alias + srs_condition + srs_dbinsert + srs_dbselect + +The following variables become available: + $srs_db_address + $srs_db_key + $srs_orig_recipient + $srs_orig_sender + $srs_recipient + $srs_status + +The predefined feature-macro _HAVE_SRS will be present. +Additional delivery log line elements, tagged with "SRS=" will show the srs sender. +For configuration information see https://github.com/Exim/exim/wiki/SRS . + + + + +SRS (Sender Rewriting Scheme) Support (native) +-------------------------------------------------------------- +This is less full-featured than the libsrs_alt version above. + +The Exim build needs to be done with this in Local/Makefile: +EXPERIMENTAL_SRS_NATIVE=yes + +The following are provided: +- an expansion item "srs_encode" + This takes three arguments: + - a site SRS secret + - the return_path + - the pre-forwarding domain + +- an expansion condition "inbound_srs" + This takes two arguments: the local_part to check, and a site SRS secret. + If the secret is zero-length, only the pattern of the local_part is checked. + The $srs_recipient variable is set as a side-effect. + +- an expansion variable $srs_recipient + This gets the original return_path encoded in the SRS'd local_part + +- predefined macros _HAVE_SRS and _HAVE_NATIVE_SRS + +Sample usage: + + #macro + SRS_SECRET = + + #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}} + + + DCC Support -------------------------------------------------------------- +Distributed Checksum Clearinghouse; http://www.rhyolite.com/dcc/ *) Building exim @@ -570,10 +464,349 @@ through to eg. SpamAssassin. If you want to pass even more headers in the middle of the DATA stage you can set $acl_m_dcc_add_header -to tell the DCC routines add more information; eg, you might set +to tell the DCC routines to add more information; eg, you might set this to some results from ClamAV. Be careful. Header syntax is not checked and is added "as is". +In case you've troubles with sites sending the same queue items from several +hosts and fail to get through greylisting you can use +$acl_m_dcc_override_client_ip + +Setting $acl_m_dcc_override_client_ip to an IP address overrides the default +of $sender_host_address. eg. use the following ACL in DATA stage: + + warn set acl_m_dcc_override_client_ip = \ + ${lookup{$sender_helo_name}nwildlsearch{/etc/mail/multipleip_sites}{$value}{}} + condition = ${if def:acl_m_dcc_override_client_ip} + log_message = dbg: acl_m_dcc_override_client_ip set to \ + $acl_m_dcc_override_client_ip + +Then set something like +# cat /etc/mail/multipleip_sites +mout-xforward.gmx.net 82.165.159.12 +mout.gmx.net 212.227.15.16 + +Use a reasonable IP. eg. one the sending cluster actually uses. + + + +DSN extra information +--------------------- +If compiled with EXPERIMENTAL_DSN_INFO extra information will be added +to DSN fail messages ("bounces"), when available. The intent is to aid +tracing of specific failing messages, when presented with a "bounce" +complaint and needing to search logs. + + +The remote MTA IP address, with port number if nonstandard. +Example: + Remote-MTA: X-ip; [127.0.0.1]:587 +Rationale: + Several addresses may correspond to the (already available) + dns name for the remote MTA. + +The remote MTA connect-time greeting. +Example: + X-Remote-MTA-smtp-greeting: X-str; 220 the.local.host.name ESMTP Exim x.yz Tue, 2 Mar 1999 09:44:33 +0000 +Rationale: + This string sometimes presents the remote MTA's idea of its + own name, and sometimes identifies the MTA software. + +The remote MTA response to HELO or EHLO. +Example: + X-Remote-MTA-helo-response: X-str; 250-the.local.host.name Hello localhost [127.0.0.1] +Limitations: + Only the first line of a multiline response is recorded. +Rationale: + This string sometimes presents the remote MTA's view of + the peer IP connecting to it. + +The reporting MTA detailed diagnostic. +Example: + X-Exim-Diagnostic: X-str; SMTP error from remote mail server after RCPT TO:: 550 hard error +Rationale: + This string sometimes give extra information over the + existing (already available) Diagnostic-Code field. + + +Note that non-RFC-documented field names and data types are used. + + +LMDB Lookup support +------------------- +LMDB is an ultra-fast, ultra-compact, crash-proof key-value embedded data store. +It is modeled loosely on the BerkeleyDB API. You should read about the feature +set as well as operation modes at https://symas.com/products/lightning-memory-mapped-database/ + +LMDB single key lookup support is provided by linking to the LMDB C library. +The current implementation does not support writing to the LMDB database. + +Visit https://github.com/LMDB/lmdb to download the library or find it in your +operating systems package repository. + +If building from source, this description assumes that headers will be in +/usr/local/include, and that the libraries are in /usr/local/lib. + +1. In order to build exim with LMDB lookup support add or uncomment + +EXPERIMENTAL_LMDB=yes + +to your Local/Makefile. (Re-)build/install exim. exim -d should show +Experimental_LMDB in the line "Support for:". + +EXPERIMENTAL_LMDB=yes +LDFLAGS += -llmdb +# CFLAGS += -I/usr/local/include +# LDFLAGS += -L/usr/local/lib + +The first line sets the feature to include the correct code, and +the second line says to link the LMDB libraries into the +exim binary. The commented out lines should be uncommented if you +built LMDB from source and installed in the default location. +Adjust the paths if you installed them elsewhere, but you do not +need to uncomment them if an rpm (or you) installed them in the +package controlled locations (/usr/include and /usr/lib). + +2. Create your LMDB files, you can use the mdb_load utility which is +part of the LMDB distribution our your favourite language bindings. + +3. Add the single key lookups to your exim.conf file, example lookups +are below. + +${lookup{$sender_address_domain}lmdb{/var/lib/baruwa/data/db/relaydomains.mdb}{$value}} +${lookup{$sender_address_domain}lmdb{/var/lib/baruwa/data/db/relaydomains.mdb}{$value}fail} +${lookup{$sender_address_domain}lmdb{/var/lib/baruwa/data/db/relaydomains.mdb}} + + +Queuefile transport +------------------- +Queuefile is a pseudo transport which does not perform final delivery. +It simply copies the exim spool files out of the spool directory into +an external directory retaining the exim spool format. + +The spool files can then be processed by external processes and then +requeued into exim spool directories for final delivery. +However, note carefully the warnings in the main documentation on +qpool file formats. + +The motivation/inspiration for the transport is to allow external +processes to access email queued by exim and have access to all the +information which would not be available if the messages were delivered +to the process in the standard email formats. + +The mailscanner package is one of the processes that can take advantage +of this transport to filter email. + +The transport can be used in the same way as the other existing transports, +i.e by configuring a router to route mail to a transport configured with +the queuefile driver. + +The transport only takes one option: + +* directory - This is used to specify the directory messages should be +copied to. Expanded. + +The generic transport options (body_only, current_directory, disable_logging, +debug_print, delivery_date_add, envelope_to_add, event_action, group, +headers_add, headers_only, headers_remove, headers_rewrite, home_directory, +initgroups, max_parallel, message_size_limit, rcpt_include_affixes, +retry_use_local_part, return_path, return_path_add, shadow_condition, +shadow_transport, transport_filter, transport_filter_timeout, user) are +ignored. + +Sample configuration: + +(Router) + +scan: + driver = accept + transport = scan + +(Transport) + +scan: + driver = queuefile + directory = /var/spool/baruwa-scanner/input + + +In order to build exim with Queuefile transport support add or uncomment + +EXPERIMENTAL_QUEUEFILE=yes + +to your Local/Makefile. (Re-)build/install exim. exim -d should show +Experimental_QUEUEFILE in the line "Support for:". + + +ARC support +----------- +Specification: https://tools.ietf.org/html/draft-ietf-dmarc-arc-protocol-11 +Note that this is not an RFC yet, so may change. + +[RFC 8617 was published 2019/06. Draft 11 was 2018/01. A review of the +changes has not yet been done] + +ARC is intended to support the utility of SPF and DKIM in the presence of +intermediaries in the transmission path - forwarders and mailinglists - +by establishing a cryptographically-signed chain in headers. + +Normally one would only bother doing ARC-signing when functioning as +an intermediary. One might do verify for local destinations. + +ARC uses the notion of a "ADministrative Management Domain" (ADMD). +Described in RFC 5598 (section 2.3), this is essentially a set of +mail-handling systems that mail transits that are all under the control +of one organisation. A label should be chosen to identify the ADMD. +Messages should be ARC-verified on entry to the ADMD, and ARC-signed on exit +from it. + + +Building with ARC Support +-- +Enable using EXPERIMENTAL_ARC=yes in your Local/Makefile. +You must also have DKIM present (not disabled), and you very likely +want to have SPF enabled. + + +Verification +-- +An ACL condition is provided to perform the "verifier actions" detailed +in section 6 of the above specification. It may be called from the DATA ACL +and succeeds if the result matches any of a given list. +It also records the highest ARC instance number (the chain size) +and verification result for later use in creating an Authentication-Results: +standard header. + + verify = arc/ none:fail:pass + + add_header = :at_start:${authresults {}} + + Note that it would be wise to strip incoming messages of A-R headers + that claim to be from our own . + +There are four new variables: + + $arc_state One of pass, fail, none + $arc_state_reason (if fail, why) + $arc_domains colon-sep list of ARC chain domains, in chain order. + problematic elements may have empty list elements + $arc_oldest_pass lowest passing instance number of chain + +Example: + logwrite = oldest-p-ams: <${reduce {$lh_ARC-Authentication-Results:} \ + {} \ + {${if = {$arc_oldest_pass} \ + {${extract {i}{${extract {1}{;}{$item}}}}} \ + {$item} {$value}}} \ + }> + +Receive log lines for an ARC pass will be tagged "ARC". + + +Signing +-- +arc_sign = : : [ : ] +An option on the smtp transport, which constructs and prepends to the message +an ARC set of headers. The textually-first Authentication-Results: header +is used as a basis (you must have added one on entry to the ADMD). +Expanded as a whole; if unset, empty or forced-failure then no signing is done. +If it is set, all of the first three elements must be non-empty. + +The fourth element is optional, and if present consists of a comma-separated list +of options. The options implemented are + + timestamps Add a t= tag to the generated AMS and AS headers, with the + current time. + expire[=] Add an x= tag to the generated AMS header, with an expiry time. + If the value is an plain number it is used unchanged. + If it starts with a '+' then the following number is added + to the current time, as an offset in seconds. + If a value is not given it defaults to a one month offset. + +[As of writing, gmail insist that a t= tag on the AS is mandatory] + +Caveats: + * There must be an Authentication-Results header, presumably added by an ACL + while receiving the message, for the same ADMD, for arc_sign to succeed. + This requires careful coordination between inbound and outbound logic. + + Only one A-R header is taken account of. This is a limitation versus + the ARC spec (which says that all A-R headers from within the ADMD must + be used). + + * If passing a message to another system, such as a mailing-list manager + (MLM), between receipt and sending, be wary of manipulations to headers made + by the MLM. + + For instance, Mailman with REMOVE_DKIM_HEADERS==3 might improve + deliverability in a pre-ARC world, but that option also renames the + Authentication-Results header, which breaks signing. + + * Even if you use multiple DKIM keys for different domains, the ARC concept + should try to stick to one ADMD, so pick a primary domain and use that for + AR headers and outbound signing. + +Signing is not compatible with cutthrough delivery; any (before expansion) +value set for the option will result in cutthrough delivery not being +used via the transport in question. + + + + +TLS Session 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 can be included by building with +EXPERIMENTAL_TLS_RESUME defined. 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 one full packet roundtrip time. + +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. + + Since a new hints DB is used, the hints DB maintenance should be updated + to additionally handle "tls". + +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. TBD. q-value; cf bug 1895 + +Observability: + New log_selector "tls_resumption", appends an asterisk to the tls_cipher "X=" + element. + + Variables $tls_{in,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 ${listextract {}{}}. + +Issues: + In a resumed session: + $tls_{in,out}_cipher will have values different to the original (under GnuTLS) + $tls_{in,out}_ocsp will be "not requested" or "no response", and + hosts_require_ocsp will fail + -------------------------------------------------------------- End of file