Commit | Line | Data |
---|---|---|
7bafa7d9 TK |
1 | From time to time, experimental features may be added to Exim. |
2 | While a feature is experimental, there will be a build-time | |
3 | option whose name starts "EXPERIMENTAL_" that must be set in | |
4 | order to include the feature. This file contains information | |
d36a0501 PP |
5 | about experimental features, all of which are unstable and |
6 | liable to incompatible change. | |
ee161e8f PH |
7 | |
8 | ||
fd98a5c6 JH |
9 | PRDR support |
10 | -------------------------------------------------------------- | |
11 | ||
12 | Per-Recipient Data Reponse is an SMTP extension proposed by Eric Hall | |
13 | in a (now-expired) IETF draft from 2007. It's not hit mainstream | |
14 | use, but has apparently been implemented in the META1 MTA. | |
15 | ||
16 | There is mention at http://mail.aegee.org/intern/sendmail.html | |
17 | of a patch to sendmail "to make it PRDR capable". | |
18 | ||
19 | ref: http://www.eric-a-hall.com/specs/draft-hall-prdr-00.txt | |
20 | ||
21 | If Exim is built with EXPERIMENTAL_PRDR there is a new config | |
22 | boolean "prdr_enable" which controls whether PRDR is advertised | |
23 | as part of an EHLO response, a new "acl_data_smtp_prdr" ACL | |
24 | (called for each recipient, after data arrives but before the | |
25 | data ACL), and a new smtp transport option "hosts_try_prdr". | |
26 | ||
27 | PRDR may be used to support per-user content filtering. Without it | |
28 | one must defer any recipient after the first that has a different | |
29 | content-filter configuration. With PRDR, the RCPT-time check | |
30 | for this can be disabled when the MAIL-time $smtp_command included | |
31 | "PRDR". Any required difference in behaviour of the main DATA-time | |
32 | ACL should however depend on the PRDR-time ACL having run, as Exim | |
33 | will avoid doing so in some situations (eg. single-recipient mails). | |
34 | ||
35 | ||
36 | ||
3f7eeb86 PP |
37 | OCSP Stapling support |
38 | -------------------------------------------------------------- | |
39 | ||
d36a0501 | 40 | X.509 PKI certificates expire and can be revoked; to handle this, the |
3f7eeb86 PP |
41 | clients need some way to determine if a particular certificate, from a |
42 | particular Certificate Authority (CA), is still valid. There are three | |
43 | main ways to do so. | |
44 | ||
45 | The simplest way is to serve up a Certificate Revocation List (CRL) with | |
46 | an ordinary web-server, regenerating the CRL before it expires. The | |
47 | downside is that clients have to periodically re-download a potentially | |
48 | huge file from every certificate authority it knows of. | |
49 | ||
50 | The way with most moving parts at query time is Online Certificate | |
51 | Status Protocol (OCSP), where the client verifies the certificate | |
52 | against an OCSP server run by the CA. This lets the CA track all | |
53 | usage of the certs. This requires running software with access to the | |
54 | private key of the CA, to sign the responses to the OCSP queries. OCSP | |
55 | is based on HTTP and can be proxied accordingly. | |
56 | ||
57 | The only widespread OCSP server implementation (known to this writer) | |
58 | comes as part of OpenSSL and aborts on an invalid request, such as | |
59 | connecting to the port and then disconnecting. This requires | |
60 | re-entering the passphrase each time some random client does this. | |
61 | ||
62 | The third way is OCSP Stapling; in this, the server using a certificate | |
63 | issued by the CA periodically requests an OCSP proof of validity from | |
64 | the OCSP server, then serves it up inline as part of the TLS | |
65 | negotiation. This approach adds no extra round trips, does not let the | |
66 | CA track users, scales well with number of certs issued by the CA and is | |
67 | resilient to temporary OCSP server failures, as long as the server | |
68 | starts retrying to fetch an OCSP proof some time before its current | |
69 | proof expires. The downside is that it requires server support. | |
70 | ||
71 | If Exim is built with EXPERIMENTAL_OCSP and it was built with OpenSSL, | |
2b4a568d JH |
72 | or with GnuTLS 3.1.3 or later, then it gains a new global option: |
73 | "tls_ocsp_file". | |
3f7eeb86 PP |
74 | |
75 | The file specified therein is expected to be in DER format, and contain | |
76 | an OCSP proof. Exim will serve it as part of the TLS handshake. This | |
77 | option will be re-expanded for SNI, if the tls_certificate option | |
78 | contains $tls_sni, as per other TLS options. | |
79 | ||
80 | Exim does not at this time implement any support for fetching a new OCSP | |
81 | proof. The burden is on the administrator to handle this, outside of | |
82 | Exim. The file specified should be replaced atomically, so that the | |
83 | contents are always valid. Exim will expand the "tls_ocsp_file" option | |
84 | on each connection, so a new file will be handled transparently on the | |
85 | next connection. | |
86 | ||
98a3b527 | 87 | Exim will check for a valid next update timestamp in the OCSP proof; |
3f7eeb86 PP |
88 | if not present, or if the proof has expired, it will be ignored. |
89 | ||
2b4a568d | 90 | Also, given EXPERIMENTAL_OCSP, the smtp transport gains |
f5d78688 JH |
91 | a "hosts_require_ocsp" option; a host-list for which an OCSP Stapling |
92 | is requested and required for the connection to proceed. The host(s) | |
93 | should also be in "hosts_require_tls", and "tls_verify_certificates" | |
94 | configured for the transport. | |
95 | ||
880496ef JH |
96 | For the client to be able to verify the stapled OCSP the server must |
97 | also supply, in its stapled information, any intermediate | |
98 | certificates for the chain leading to the OCSP proof from the signer | |
99 | of the server certificate. There may be zero or one such. These | |
100 | intermediate certificates should be added to the server OCSP stapling | |
101 | file (named by tls_ocsp_file). | |
102 | ||
2b4a568d JH |
103 | Note that the proof only covers the terminal server certificate, |
104 | not any of the chain from CA to it. | |
105 | ||
3f7eeb86 PP |
106 | At this point in time, we're gathering feedback on use, to determine if |
107 | it's worth adding complexity to the Exim daemon to periodically re-fetch | |
f5d78688 | 108 | OCSP files and somehow handling multiple files. |
3f7eeb86 | 109 | |
880496ef JH |
110 | A helper script "ocsp_fetch.pl" for fetching a proof from a CA |
111 | OCSP server is supplied. The server URL may be included in the | |
112 | server certificate, if the CA is helpful. | |
113 | ||
2b4a568d JH |
114 | One failure mode seen was the OCSP Signer cert expiring before the end |
115 | of validity of the OCSP proof. The checking done by Exim/OpenSSL | |
880496ef JH |
116 | noted this as invalid overall, but the re-fetch script did not. |
117 | ||
3f7eeb86 PP |
118 | |
119 | ||
120 | ||
0b23848a | 121 | Brightmail AntiSpam (BMI) suppport |
ee161e8f PH |
122 | -------------------------------------------------------------- |
123 | ||
124 | Brightmail AntiSpam is a commercial package. Please see | |
125 | http://www.brightmail.com for more information on | |
126 | the product. For the sake of clarity, we'll refer to it as | |
127 | "BMI" from now on. | |
128 | ||
129 | ||
130 | 0) BMI concept and implementation overview | |
131 | ||
132 | In contrast to how spam-scanning with SpamAssassin is | |
133 | implemented in exiscan-acl, BMI is more suited for per | |
134 | -recipient scanning of messages. However, each messages is | |
135 | scanned only once, but multiple "verdicts" for multiple | |
136 | recipients can be returned from the BMI server. The exiscan | |
137 | implementation passes the message to the BMI server just | |
138 | before accepting it. It then adds the retrieved verdicts to | |
139 | the messages header file in the spool. These verdicts can then | |
140 | be queried in routers, where operation is per-recipient | |
141 | instead of per-message. To use BMI, you need to take the | |
142 | following steps: | |
143 | ||
144 | 1) Compile Exim with BMI support | |
3ec3e3bb | 145 | 2) Set up main BMI options (top section of Exim config file) |
ee161e8f PH |
146 | 3) Set up ACL control statement (ACL section of the config |
147 | file) | |
148 | 4) Set up your routers to use BMI verdicts (routers section | |
149 | of the config file). | |
150 | 5) (Optional) Set up per-recipient opt-in information. | |
151 | ||
8ff3788c | 152 | These four steps are explained in more details below. |
ee161e8f PH |
153 | |
154 | 1) Adding support for BMI at compile time | |
155 | ||
156 | To compile with BMI support, you need to link Exim against | |
157 | the Brighmail client SDK, consisting of a library | |
158 | (libbmiclient_single.so) and a header file (bmi_api.h). | |
159 | You'll also need to explicitly set a flag in the Makefile to | |
160 | include BMI support in the Exim binary. Both can be achieved | |
161 | with these lines in Local/Makefile: | |
162 | ||
163 | EXPERIMENTAL_BRIGHTMAIL=yes | |
47bbda99 | 164 | CFLAGS=-I/path/to/the/dir/with/the/includefile |
ee161e8f | 165 | EXTRALIBS_EXIM=-L/path/to/the/dir/with/the/library -lbmiclient_single |
8ff3788c | 166 | |
ee161e8f PH |
167 | If you use other CFLAGS or EXTRALIBS_EXIM settings then |
168 | merge the content of these lines with them. | |
169 | ||
7c0c8547 | 170 | Note for BMI6.x users: You'll also have to add -lxml2_single |
ee161e8f PH |
171 | to the EXTRALIBS_EXIM line. Users of 5.5x do not need to do |
172 | this. | |
8ff3788c | 173 | |
ee161e8f PH |
174 | You should also include the location of |
175 | libbmiclient_single.so in your dynamic linker configuration | |
176 | file (usually /etc/ld.so.conf) and run "ldconfig" | |
177 | afterwards, or else the produced Exim binary will not be | |
178 | able to find the library file. | |
179 | ||
180 | ||
3ec3e3bb | 181 | 2) Setting up BMI support in the Exim main configuration |
ee161e8f | 182 | |
3ec3e3bb | 183 | To enable BMI support in the main Exim configuration, you |
ee161e8f PH |
184 | should set the path to the main BMI configuration file with |
185 | the "bmi_config_file" option, like this: | |
8ff3788c | 186 | |
ee161e8f | 187 | bmi_config_file = /opt/brightmail/etc/brightmail.cfg |
8ff3788c | 188 | |
3ec3e3bb | 189 | This must go into section 1 of Exim's configuration file (You |
ee161e8f PH |
190 | can put it right on top). If you omit this option, it |
191 | defaults to /opt/brightmail/etc/brightmail.cfg. | |
192 | ||
193 | Note for BMI6.x users: This file is in XML format in V6.xx | |
194 | and its name is /opt/brightmail/etc/bmiconfig.xml. So BMI | |
195 | 6.x users MUST set the bmi_config_file option. | |
8ff3788c | 196 | |
ee161e8f PH |
197 | |
198 | 3) Set up ACL control statement | |
199 | ||
200 | To optimize performance, it makes sense only to process | |
201 | messages coming from remote, untrusted sources with the BMI | |
202 | server. To set up a messages for processing by the BMI | |
203 | server, you MUST set the "bmi_run" control statement in any | |
204 | ACL for an incoming message. You will typically do this in | |
205 | an "accept" block in the "acl_check_rcpt" ACL. You should | |
206 | use the "accept" block(s) that accept messages from remote | |
207 | servers for your own domain(s). Here is an example that uses | |
3ec3e3bb | 208 | the "accept" blocks from Exim's default configuration file: |
8ff3788c | 209 | |
ee161e8f PH |
210 | |
211 | accept domains = +local_domains | |
212 | endpass | |
213 | verify = recipient | |
214 | control = bmi_run | |
215 | ||
216 | accept domains = +relay_to_domains | |
217 | endpass | |
218 | verify = recipient | |
219 | control = bmi_run | |
8ff3788c | 220 | |
ee161e8f PH |
221 | If bmi_run is not set in any ACL during reception of the |
222 | message, it will NOT be passed to the BMI server. | |
223 | ||
224 | ||
225 | 4) Setting up routers to use BMI verdicts | |
226 | ||
227 | When a message has been run through the BMI server, one or | |
228 | more "verdicts" are present. Different recipients can have | |
229 | different verdicts. Each recipient is treated individually | |
230 | during routing, so you can query the verdicts by recipient | |
3ec3e3bb | 231 | at that stage. From Exim's view, a verdict can have the |
ee161e8f | 232 | following outcomes: |
8ff3788c | 233 | |
ee161e8f PH |
234 | o deliver the message normally |
235 | o deliver the message to an alternate location | |
236 | o do not deliver the message | |
8ff3788c | 237 | |
ee161e8f PH |
238 | To query the verdict for a recipient, the implementation |
239 | offers the following tools: | |
8ff3788c TK |
240 | |
241 | ||
ee161e8f PH |
242 | - Boolean router preconditions. These can be used in any |
243 | router. For a simple implementation of BMI, these may be | |
244 | all that you need. The following preconditions are | |
245 | available: | |
8ff3788c | 246 | |
ee161e8f | 247 | o bmi_deliver_default |
8ff3788c | 248 | |
ee161e8f PH |
249 | This precondition is TRUE if the verdict for the |
250 | recipient is to deliver the message normally. If the | |
251 | message has not been processed by the BMI server, this | |
252 | variable defaults to TRUE. | |
8ff3788c | 253 | |
ee161e8f | 254 | o bmi_deliver_alternate |
8ff3788c | 255 | |
ee161e8f PH |
256 | This precondition is TRUE if the verdict for the |
257 | recipient is to deliver the message to an alternate | |
258 | location. You can get the location string from the | |
259 | $bmi_alt_location expansion variable if you need it. See | |
260 | further below. If the message has not been processed by | |
261 | the BMI server, this variable defaults to FALSE. | |
8ff3788c | 262 | |
ee161e8f | 263 | o bmi_dont_deliver |
8ff3788c | 264 | |
ee161e8f PH |
265 | This precondition is TRUE if the verdict for the |
266 | recipient is NOT to deliver the message to the | |
267 | recipient. You will typically use this precondition in a | |
268 | top-level blackhole router, like this: | |
8ff3788c | 269 | |
ee161e8f PH |
270 | # don't deliver messages handled by the BMI server |
271 | bmi_blackhole: | |
272 | driver = redirect | |
273 | bmi_dont_deliver | |
274 | data = :blackhole: | |
8ff3788c | 275 | |
ee161e8f PH |
276 | This router should be on top of all others, so messages |
277 | that should not be delivered do not reach other routers | |
278 | at all. If the message has not been processed by | |
279 | the BMI server, this variable defaults to FALSE. | |
8ff3788c TK |
280 | |
281 | ||
ee161e8f PH |
282 | - A list router precondition to query if rules "fired" on |
283 | the message for the recipient. Its name is "bmi_rule". You | |
284 | use it by passing it a colon-separated list of rule | |
285 | numbers. You can use this condition to route messages that | |
286 | matched specific rules. Here is an example: | |
8ff3788c | 287 | |
ee161e8f PH |
288 | # special router for BMI rule #5, #8 and #11 |
289 | bmi_rule_redirect: | |
290 | driver = redirect | |
291 | bmi_rule = 5:8:11 | |
292 | data = postmaster@mydomain.com | |
8ff3788c TK |
293 | |
294 | ||
ee161e8f PH |
295 | - Expansion variables. Several expansion variables are set |
296 | during routing. You can use them in custom router | |
297 | conditions, for example. The following variables are | |
298 | available: | |
8ff3788c | 299 | |
ee161e8f | 300 | o $bmi_base64_verdict |
8ff3788c | 301 | |
ee161e8f PH |
302 | This variable will contain the BASE64 encoded verdict |
303 | for the recipient being routed. You can use it to add a | |
304 | header to messages for tracking purposes, for example: | |
8ff3788c | 305 | |
ee161e8f PH |
306 | localuser: |
307 | driver = accept | |
308 | check_local_user | |
309 | headers_add = X-Brightmail-Verdict: $bmi_base64_verdict | |
310 | transport = local_delivery | |
8ff3788c | 311 | |
ee161e8f PH |
312 | If there is no verdict available for the recipient being |
313 | routed, this variable contains the empty string. | |
8ff3788c | 314 | |
ee161e8f | 315 | o $bmi_base64_tracker_verdict |
8ff3788c | 316 | |
ee161e8f PH |
317 | This variable will contain a BASE64 encoded subset of |
318 | the verdict information concerning the "rules" that | |
319 | fired on the message. You can add this string to a | |
320 | header, commonly named "X-Brightmail-Tracker". Example: | |
8ff3788c | 321 | |
ee161e8f PH |
322 | localuser: |
323 | driver = accept | |
324 | check_local_user | |
325 | headers_add = X-Brightmail-Tracker: $bmi_base64_tracker_verdict | |
326 | transport = local_delivery | |
8ff3788c | 327 | |
ee161e8f PH |
328 | If there is no verdict available for the recipient being |
329 | routed, this variable contains the empty string. | |
8ff3788c | 330 | |
ee161e8f | 331 | o $bmi_alt_location |
8ff3788c | 332 | |
ee161e8f PH |
333 | If the verdict is to redirect the message to an |
334 | alternate location, this variable will contain the | |
335 | alternate location string returned by the BMI server. In | |
336 | its default configuration, this is a header-like string | |
337 | that can be added to the message with "headers_add". If | |
338 | there is no verdict available for the recipient being | |
339 | routed, or if the message is to be delivered normally, | |
340 | this variable contains the empty string. | |
8ff3788c | 341 | |
ee161e8f | 342 | o $bmi_deliver |
8ff3788c | 343 | |
ee161e8f PH |
344 | This is an additional integer variable that can be used |
345 | to query if the message should be delivered at all. You | |
346 | should use router preconditions instead if possible. | |
8ff3788c | 347 | |
ee161e8f PH |
348 | $bmi_deliver is '0': the message should NOT be delivered. |
349 | $bmi_deliver is '1': the message should be delivered. | |
8ff3788c TK |
350 | |
351 | ||
ee161e8f PH |
352 | IMPORTANT NOTE: Verdict inheritance. |
353 | The message is passed to the BMI server during message | |
354 | reception, using the target addresses from the RCPT TO: | |
355 | commands in the SMTP transaction. If recipients get expanded | |
356 | or re-written (for example by aliasing), the new address(es) | |
357 | inherit the verdict from the original address. This means | |
358 | that verdicts also apply to all "child" addresses generated | |
359 | from top-level addresses that were sent to the BMI server. | |
8ff3788c TK |
360 | |
361 | ||
ee161e8f PH |
362 | 5) Using per-recipient opt-in information (Optional) |
363 | ||
364 | The BMI server features multiple scanning "profiles" for | |
365 | individual recipients. These are usually stored in a LDAP | |
366 | server and are queried by the BMI server itself. However, | |
367 | you can also pass opt-in data for each recipient from the | |
368 | MTA to the BMI server. This is particularly useful if you | |
3ec3e3bb | 369 | already look up recipient data in Exim anyway (which can |
ee161e8f PH |
370 | also be stored in a SQL database or other source). This |
371 | implementation enables you to pass opt-in data to the BMI | |
372 | server in the RCPT ACL. This works by setting the | |
373 | 'bmi_optin' modifier in a block of that ACL. If should be | |
374 | set to a list of comma-separated strings that identify the | |
375 | features which the BMI server should use for that particular | |
376 | recipient. Ideally, you would use the 'bmi_optin' modifier | |
377 | in the same ACL block where you set the 'bmi_run' control | |
378 | flag. Here is an example that will pull opt-in data for each | |
379 | recipient from a flat file called | |
380 | '/etc/exim/bmi_optin_data'. | |
8ff3788c | 381 | |
ee161e8f | 382 | The file format: |
8ff3788c | 383 | |
ee161e8f PH |
384 | user1@mydomain.com: <OPTIN STRING1>:<OPTIN STRING2> |
385 | user2@thatdomain.com: <OPTIN STRING3> | |
8ff3788c TK |
386 | |
387 | ||
ee161e8f | 388 | The example: |
8ff3788c | 389 | |
ee161e8f PH |
390 | accept domains = +relay_to_domains |
391 | endpass | |
392 | verify = recipient | |
393 | bmi_optin = ${lookup{$local_part@$domain}lsearch{/etc/exim/bmi_optin_data}} | |
8ff3788c TK |
394 | control = bmi_run |
395 | ||
ee161e8f | 396 | Of course, you can also use any other lookup method that |
3ec3e3bb | 397 | Exim supports, including LDAP, Postgres, MySQL, Oracle etc., |
ee161e8f PH |
398 | as long as the result is a list of colon-separated opt-in |
399 | strings. | |
8ff3788c | 400 | |
ee161e8f PH |
401 | For a list of available opt-in strings, please contact your |
402 | Brightmail representative. | |
ee161e8f | 403 | |
8ff3788c TK |
404 | |
405 | ||
406 | ||
0b23848a | 407 | Sender Policy Framework (SPF) support |
ee161e8f PH |
408 | -------------------------------------------------------------- |
409 | ||
f413481d | 410 | To learn more about SPF, visit http://www.openspf.org. This |
ee161e8f PH |
411 | document does not explain the SPF fundamentals, you should |
412 | read and understand the implications of deploying SPF on your | |
413 | system before doing so. | |
414 | ||
8ff3788c | 415 | SPF support is added via the libspf2 library. Visit |
ee161e8f PH |
416 | |
417 | http://www.libspf2.org/ | |
8ff3788c | 418 | |
ee161e8f PH |
419 | to obtain a copy, then compile and install it. By default, |
420 | this will put headers in /usr/local/include and the static | |
421 | library in /usr/local/lib. | |
422 | ||
3ec3e3bb | 423 | To compile Exim with SPF support, set these additional flags in |
ee161e8f PH |
424 | Local/Makefile: |
425 | ||
426 | EXPERIMENTAL_SPF=yes | |
427 | CFLAGS=-DSPF -I/usr/local/include | |
428 | EXTRALIBS_EXIM=-L/usr/local/lib -lspf2 | |
429 | ||
430 | This assumes that the libspf2 files are installed in | |
431 | their default locations. | |
432 | ||
433 | You can now run SPF checks in incoming SMTP by using the "spf" | |
434 | ACL condition in either the MAIL, RCPT or DATA ACLs. When | |
d36a0501 | 435 | using it in the RCPT ACL, you can make the checks dependent on |
ee161e8f PH |
436 | the RCPT address (or domain), so you can check SPF records |
437 | only for certain target domains. This gives you the | |
438 | possibility to opt-out certain customers that do not want | |
439 | their mail to be subject to SPF checking. | |
440 | ||
441 | The spf condition takes a list of strings on its right-hand | |
442 | side. These strings describe the outcome of the SPF check for | |
443 | which the spf condition should succeed. Valid strings are: | |
444 | ||
445 | o pass The SPF check passed, the sending host | |
446 | is positively verified by SPF. | |
447 | o fail The SPF check failed, the sending host | |
448 | is NOT allowed to send mail for the domain | |
449 | in the envelope-from address. | |
450 | o softfail The SPF check failed, but the queried | |
451 | domain can't absolutely confirm that this | |
452 | is a forgery. | |
453 | o none The queried domain does not publish SPF | |
454 | records. | |
455 | o neutral The SPF check returned a "neutral" state. | |
456 | This means the queried domain has published | |
457 | a SPF record, but wants to allow outside | |
458 | servers to send mail under its domain as well. | |
8ddef691 TL |
459 | This should be treated like "none". |
460 | o permerror This indicates a syntax error in the SPF | |
461 | record of the queried domain. You may deny | |
462 | messages when this occurs. (Changed in 4.83) | |
463 | o temperror This indicates a temporary error during all | |
3ec3e3bb | 464 | processing, including Exim's SPF processing. |
ee161e8f | 465 | You may defer messages when this occurs. |
8ddef691 | 466 | (Changed in 4.83) |
982650ec TL |
467 | o err_temp Same as permerror, deprecated in 4.83, will be |
468 | removed in a future release. | |
469 | o err_perm Same as temperror, deprecated in 4.83, will be | |
470 | removed in a future release. | |
8ff3788c | 471 | |
ee161e8f | 472 | You can prefix each string with an exclamation mark to invert |
982650ec | 473 | its meaning, for example "!fail" will match all results but |
ee161e8f PH |
474 | "fail". The string list is evaluated left-to-right, in a |
475 | short-circuit fashion. When a string matches the outcome of | |
476 | the SPF check, the condition succeeds. If none of the listed | |
477 | strings matches the outcome of the SPF check, the condition | |
478 | fails. | |
479 | ||
f413481d TK |
480 | Here is an example to fail forgery attempts from domains that |
481 | publish SPF records: | |
ee161e8f PH |
482 | |
483 | /* ----------------- | |
f413481d TK |
484 | deny message = $sender_host_address is not allowed to send mail from ${if def:sender_address_domain {$sender_address_domain}{$sender_helo_name}}. \ |
485 | Please see http://www.openspf.org/Why?scope=${if def:sender_address_domain {mfrom}{helo}};identity=${if def:sender_address_domain {$sender_address}{$sender_helo_name}};ip=$sender_host_address | |
ee161e8f PH |
486 | spf = fail |
487 | --------------------- */ | |
488 | ||
489 | You can also give special treatment to specific domains: | |
490 | ||
491 | /* ----------------- | |
492 | deny message = AOL sender, but not from AOL-approved relay. | |
493 | sender_domains = aol.com | |
494 | spf = fail:neutral | |
495 | --------------------- */ | |
496 | ||
497 | Explanation: AOL publishes SPF records, but is liberal and | |
498 | still allows non-approved relays to send mail from aol.com. | |
499 | This will result in a "neutral" state, while mail from genuine | |
500 | AOL servers will result in "pass". The example above takes | |
501 | this into account and treats "neutral" like "fail", but only | |
502 | for aol.com. Please note that this violates the SPF draft. | |
503 | ||
504 | When the spf condition has run, it sets up several expansion | |
505 | variables. | |
506 | ||
507 | $spf_header_comment | |
508 | This contains a human-readable string describing the outcome | |
509 | of the SPF check. You can add it to a custom header or use | |
510 | it for logging purposes. | |
8ff3788c | 511 | |
ee161e8f | 512 | $spf_received |
8fe685ad | 513 | This contains a complete Received-SPF: header that can be |
ee161e8f PH |
514 | added to the message. Please note that according to the SPF |
515 | draft, this header must be added at the top of the header | |
516 | list. Please see section 10 on how you can do this. | |
8ff3788c | 517 | |
65a7d8c3 NM |
518 | Note: in case of "Best-guess" (see below), the convention is |
519 | to put this string in a header called X-SPF-Guess: instead. | |
520 | ||
ee161e8f PH |
521 | $spf_result |
522 | This contains the outcome of the SPF check in string form, | |
8ddef691 TL |
523 | one of pass, fail, softfail, none, neutral, permerror or |
524 | temperror. | |
8ff3788c | 525 | |
ee161e8f PH |
526 | $spf_smtp_comment |
527 | This contains a string that can be used in a SMTP response | |
528 | to the calling party. Useful for "fail". | |
8ff3788c | 529 | |
65a7d8c3 NM |
530 | In addition to SPF, you can also perform checks for so-called |
531 | "Best-guess". Strictly speaking, "Best-guess" is not standard | |
532 | SPF, but it is supported by the same framework that enables SPF | |
533 | capability. Refer to http://www.openspf.org/FAQ/Best_guess_record | |
534 | for a description of what it means. | |
535 | ||
536 | To access this feature, simply use the spf_guess condition in place | |
537 | of the spf one. For example: | |
538 | ||
539 | /* ----------------- | |
540 | deny message = $sender_host_address doesn't look trustworthy to me | |
541 | spf_guess = fail | |
542 | --------------------- */ | |
543 | ||
544 | In case you decide to reject messages based on this check, you | |
545 | should note that although it uses the same framework, "Best-guess" | |
546 | is NOT SPF, and therefore you should not mention SPF at all in your | |
547 | reject message. | |
548 | ||
549 | When the spf_guess condition has run, it sets up the same expansion | |
550 | variables as when spf condition is run, described above. | |
551 | ||
d36a0501 | 552 | Additionally, since Best-guess is not standardized, you may redefine |
65a7d8c3 NM |
553 | what "Best-guess" means to you by redefining spf_guess variable in |
554 | global config. For example, the following: | |
555 | ||
556 | /* ----------------- | |
557 | spf_guess = v=spf1 a/16 mx/16 ptr ?all | |
558 | --------------------- */ | |
559 | ||
560 | would relax host matching rules to a broader network range. | |
8ff3788c | 561 | |
ee161e8f | 562 | |
0b23848a | 563 | SRS (Sender Rewriting Scheme) Support |
ee161e8f PH |
564 | -------------------------------------------------------------- |
565 | ||
566 | Exiscan currently includes SRS support via Miles Wilton's | |
8ff3788c | 567 | libsrs_alt library. The current version of the supported |
ee161e8f PH |
568 | library is 0.5. |
569 | ||
570 | In order to use SRS, you must get a copy of libsrs_alt from | |
571 | ||
572 | http://srs.mirtol.com/ | |
573 | ||
574 | Unpack the tarball, then refer to MTAs/README.EXIM | |
575 | to proceed. You need to set | |
576 | ||
577 | EXPERIMENTAL_SRS=yes | |
578 | ||
579 | in your Local/Makefile. | |
580 | ||
581 | ||
0e1ccf44 PP |
582 | DCC Support |
583 | -------------------------------------------------------------- | |
584 | ||
585 | *) Building exim | |
586 | ||
587 | In order to build exim with DCC support add | |
588 | ||
589 | EXPERIMENTAL_DCC=yes | |
590 | ||
591 | to your Makefile. (Re-)build/install exim. exim -d should show | |
592 | EXPERIMENTAL_DCC under "Support for". | |
593 | ||
594 | ||
595 | *) Configuration | |
596 | ||
597 | In the main section of exim.cf add at least | |
598 | dccifd_address = /usr/local/dcc/var/dccifd | |
599 | or | |
600 | dccifd_address = <ip> <port> | |
601 | ||
602 | In the DATA ACL you can use the new condition | |
603 | dcc = * | |
604 | ||
605 | After that "$dcc_header" contains the X-DCC-Header. | |
606 | ||
d36a0501 | 607 | Return values are: |
0e1ccf44 PP |
608 | fail for overall "R", "G" from dccifd |
609 | defer for overall "T" from dccifd | |
610 | accept for overall "A", "S" from dccifd | |
611 | ||
612 | dcc = */defer_ok works as for spamd. | |
613 | ||
614 | The "$dcc_result" variable contains the overall result from DCC | |
615 | answer. There will an X-DCC: header added to the mail. | |
616 | ||
617 | Usually you'll use | |
618 | defer !dcc = * | |
619 | to greylist with DCC. | |
620 | ||
621 | If you set, in the main section, | |
622 | dcc_direct_add_header = true | |
623 | then the dcc header will be added "in deep" and if the spool | |
624 | file was already written it gets removed. This forces Exim to | |
625 | write it again if needed. This helps to get the DCC Header | |
626 | through to eg. SpamAssassin. | |
627 | ||
628 | If you want to pass even more headers in the middle of the | |
629 | DATA stage you can set | |
630 | $acl_m_dcc_add_header | |
05c39afa | 631 | to tell the DCC routines to add more information; eg, you might set |
0e1ccf44 PP |
632 | this to some results from ClamAV. Be careful. Header syntax is |
633 | not checked and is added "as is". | |
634 | ||
05c39afa JH |
635 | In case you've troubles with sites sending the same queue items from several |
636 | hosts and fail to get through greylisting you can use | |
637 | $acl_m_dcc_override_client_ip | |
638 | ||
639 | Setting $acl_m_dcc_override_client_ip to an IP address overrides the default | |
640 | of $sender_host_address. eg. use the following ACL in DATA stage: | |
641 | ||
642 | warn set acl_m_dcc_override_client_ip = \ | |
643 | ${lookup{$sender_helo_name}nwildlsearch{/etc/mail/multipleip_sites}{$value}{}} | |
644 | condition = ${if def:acl_m_dcc_override_client_ip} | |
645 | log_message = dbg: acl_m_dcc_override_client_ip set to \ | |
646 | $acl_m_dcc_override_client_ip | |
647 | ||
648 | Then set something like | |
649 | # cat /etc/mail/multipleip_sites | |
650 | mout-xforward.gmx.net 82.165.159.12 | |
651 | mout.gmx.net 212.227.15.16 | |
652 | ||
653 | Use a reasonable IP. eg. one the sending cluster acutally uses. | |
0e1ccf44 | 654 | |
1899bab2 TL |
655 | DMARC Support |
656 | -------------------------------------------------------------- | |
657 | ||
658 | DMARC combines feedback from SPF, DKIM, and header From: in order | |
659 | to attempt to provide better indicators of the authenticity of an | |
660 | email. This document does not explain the fundamentals, you | |
661 | should read and understand how it works by visiting the website at | |
662 | http://www.dmarc.org/. | |
663 | ||
664 | DMARC support is added via the libopendmarc library. Visit: | |
665 | ||
666 | http://sourceforge.net/projects/opendmarc/ | |
667 | ||
668 | to obtain a copy, or find it in your favorite rpm package | |
669 | repository. If building from source, this description assumes | |
670 | that headers will be in /usr/local/include, and that the libraries | |
671 | are in /usr/local/lib. | |
672 | ||
673 | 1. To compile Exim with DMARC support, you must first enable SPF. | |
674 | Please read the above section on enabling the EXPERIMENTAL_SPF | |
675 | feature. You must also have DKIM support, so you cannot set the | |
676 | DISABLE_DKIM feature. Once both of those conditions have been met | |
677 | you can enable DMARC in Local/Makefile: | |
678 | ||
679 | EXPERIMENTAL_DMARC=yes | |
680 | LDFLAGS += -lopendmarc | |
681 | # CFLAGS += -I/usr/local/include | |
682 | # LDFLAGS += -L/usr/local/lib | |
683 | ||
684 | The first line sets the feature to include the correct code, and | |
685 | the second line says to link the libopendmarc libraries into the | |
686 | exim binary. The commented out lines should be uncommented if you | |
687 | built opendmarc from source and installed in the default location. | |
688 | Adjust the paths if you installed them elsewhere, but you do not | |
689 | need to uncomment them if an rpm (or you) installed them in the | |
690 | package controlled locations (/usr/include and /usr/lib). | |
691 | ||
692 | ||
693 | 2. Use the following global settings to configure DMARC: | |
694 | ||
695 | Required: | |
696 | dmarc_tld_file Defines the location of a text file of valid | |
697 | top level domains the opendmarc library uses | |
698 | during domain parsing. Maintained by Mozilla, | |
699 | the most current version can be downloaded | |
700 | from a link at http://publicsuffix.org/list/. | |
701 | ||
702 | Optional: | |
703 | dmarc_history_file Defines the location of a file to log results | |
704 | of dmarc verification on inbound emails. The | |
705 | contents are importable by the opendmarc tools | |
706 | which will manage the data, send out DMARC | |
707 | reports, and expire the data. Make sure the | |
708 | directory of this file is writable by the user | |
709 | exim runs as. | |
710 | ||
711 | dmarc_forensic_sender The email address to use when sending a | |
712 | forensic report detailing alignment failures | |
713 | if a sender domain's dmarc record specifies it | |
714 | and you have configured Exim to send them. | |
715 | Default: do-not-reply@$default_hostname | |
716 | ||
717 | ||
718 | 3. By default, the DMARC processing will run for any remote, | |
719 | non-authenticated user. It makes sense to only verify DMARC | |
720 | status of messages coming from remote, untrusted sources. You can | |
721 | use standard conditions such as hosts, senders, etc, to decide that | |
722 | DMARC verification should *not* be performed for them and disable | |
723 | DMARC with a control setting: | |
724 | ||
12d0043d | 725 | control = dmarc_disable_verify |
1899bab2 TL |
726 | |
727 | A DMARC record can also specify a "forensic address", which gives | |
728 | exim an email address to submit reports about failed alignment. | |
729 | Exim does not do this by default because in certain conditions it | |
730 | results in unintended information leakage (what lists a user might | |
731 | be subscribed to, etc). You must configure exim to submit forensic | |
732 | reports to the owner of the domain. If the DMARC record contains a | |
733 | forensic address and you specify the control statement below, then | |
734 | exim will send these forensic emails. It's also advised that you | |
735 | configure a dmarc_forensic_sender because the default sender address | |
736 | construction might be inadequate. | |
737 | ||
738 | control = dmarc_forensic_enable | |
739 | ||
740 | (AGAIN: You can choose not to send these forensic reports by simply | |
741 | not putting the dmarc_forensic_enable control line at any point in | |
742 | your exim config. If you don't tell it to send them, it will not | |
743 | send them.) | |
744 | ||
745 | There are no options to either control. Both must appear before | |
746 | the DATA acl. | |
747 | ||
748 | ||
749 | 4. You can now run DMARC checks in incoming SMTP by using the | |
750 | "dmarc_status" ACL condition in the DATA ACL. You are required to | |
751 | call the spf condition first in the ACLs, then the "dmarc_status" | |
752 | condition. Putting this condition in the ACLs is required in order | |
753 | for a DMARC check to actually occur. All of the variables are set | |
754 | up before the DATA ACL, but there is no actual DMARC check that | |
755 | occurs until a "dmarc_status" condition is encountered in the ACLs. | |
756 | ||
757 | The dmarc_status condition takes a list of strings on its | |
758 | right-hand side. These strings describe recommended action based | |
759 | on the DMARC check. To understand what the policy recommendations | |
760 | mean, refer to the DMARC website above. Valid strings are: | |
761 | ||
762 | o accept The DMARC check passed and the library recommends | |
763 | accepting the email. | |
764 | o reject The DMARC check failed and the library recommends | |
765 | rejecting the email. | |
766 | o quarantine The DMARC check failed and the library recommends | |
767 | keeping it for further inspection. | |
7a8678e6 | 768 | o none The DMARC check passed and the library recommends |
769 | no specific action, neutral. | |
1899bab2 TL |
770 | o norecord No policy section in the DMARC record for this |
771 | sender domain. | |
772 | o nofrom Unable to determine the domain of the sender. | |
7a8678e6 | 773 | o temperror Library error or dns error. |
05070e30 | 774 | o off The DMARC check was disabled for this email. |
1899bab2 TL |
775 | |
776 | You can prefix each string with an exclamation mark to invert its | |
777 | meaning, for example "!accept" will match all results but | |
778 | "accept". The string list is evaluated left-to-right in a | |
779 | short-circuit fashion. When a string matches the outcome of the | |
780 | DMARC check, the condition succeeds. If none of the listed | |
781 | strings matches the outcome of the DMARC check, the condition | |
782 | fails. | |
783 | ||
784 | Of course, you can also use any other lookup method that Exim | |
785 | supports, including LDAP, Postgres, MySQL, etc, as long as the | |
8c8b8274 | 786 | result is a list of colon-separated strings. |
1899bab2 TL |
787 | |
788 | Several expansion variables are set before the DATA ACL is | |
789 | processed, and you can use them in this ACL. The following | |
790 | expansion variables are available: | |
791 | ||
792 | o $dmarc_status | |
793 | This is a one word status indicating what the DMARC library | |
8c8b8274 TL |
794 | thinks of the email. It is a combination of the results of |
795 | DMARC record lookup and the SPF/DKIM/DMARC processing results | |
796 | (if a DMARC record was found). The actual policy declared | |
797 | in the DMARC record is in a separate expansion variable. | |
1899bab2 TL |
798 | |
799 | o $dmarc_status_text | |
800 | This is a slightly longer, human readable status. | |
801 | ||
802 | o $dmarc_used_domain | |
803 | This is the domain which DMARC used to look up the DMARC | |
804 | policy record. | |
805 | ||
8c8b8274 TL |
806 | o $dmarc_domain_policy |
807 | This is the policy declared in the DMARC record. Valid values | |
808 | are "none", "reject" and "quarantine". It is blank when there | |
809 | is any error, including no DMARC record. | |
810 | ||
1899bab2 TL |
811 | o $dmarc_ar_header |
812 | This is the entire Authentication-Results header which you can | |
813 | add using an add_header modifier. | |
814 | ||
815 | ||
816 | 5. How to enable DMARC advanced operation: | |
817 | By default, Exim's DMARC configuration is intended to be | |
818 | non-intrusive and conservative. To facilitate this, Exim will not | |
819 | create any type of logging files without explicit configuration by | |
820 | you, the admin. Nor will Exim send out any emails/reports about | |
821 | DMARC issues without explicit configuration by you, the admin (other | |
822 | than typical bounce messages that may come about due to ACL | |
823 | processing or failure delivery issues). | |
824 | ||
825 | In order to log statistics suitable to be imported by the opendmarc | |
826 | tools, you need to: | |
827 | a. Configure the global setting dmarc_history_file. | |
828 | b. Configure cron jobs to call the appropriate opendmarc history | |
829 | import scripts and truncating the dmarc_history_file. | |
830 | ||
831 | In order to send forensic reports, you need to: | |
832 | a. Configure the global setting dmarc_forensic_sender. | |
833 | b. Configure, somewhere before the DATA ACL, the control option to | |
834 | enable sending DMARC forensic reports. | |
835 | ||
836 | ||
837 | 6. Example usage: | |
838 | (RCPT ACL) | |
839 | warn domains = +local_domains | |
840 | hosts = +local_hosts | |
12d0043d | 841 | control = dmarc_disable_verify |
1899bab2 TL |
842 | |
843 | warn !domains = +screwed_up_dmarc_records | |
844 | control = dmarc_enable_forensic | |
845 | ||
8c8b8274 TL |
846 | warn condition = (lookup if destined to mailing list) |
847 | set acl_m_mailing_list = 1 | |
848 | ||
1899bab2 TL |
849 | (DATA ACL) |
850 | warn dmarc_status = accept : none : off | |
851 | !authenticated = * | |
852 | log_message = DMARC DEBUG: $dmarc_status $dmarc_used_domain | |
853 | add_header = $dmarc_ar_header | |
854 | ||
855 | warn dmarc_status = !accept | |
856 | !authenticated = * | |
857 | log_message = DMARC DEBUG: '$dmarc_status' for $dmarc_used_domain | |
858 | ||
859 | warn dmarc_status = quarantine | |
860 | !authenticated = * | |
861 | set $acl_m_quarantine = 1 | |
862 | # Do something in a transport with this flag variable | |
863 | ||
8c8b8274 TL |
864 | deny condition = ${if eq{$dmarc_domain_policy}{reject}} |
865 | condition = ${if eq{$acl_m_mailing_list}{1}} | |
866 | message = Messages from $dmarc_used_domain break mailing lists | |
867 | ||
1899bab2 TL |
868 | deny dmarc_status = reject |
869 | !authenticated = * | |
870 | message = Message from $domain_used_domain failed sender's DMARC policy, REJECT | |
871 | ||
872 | ||
873 | ||
d68218c7 JH |
874 | Transport post-delivery actions |
875 | -------------------------------------------------------------- | |
876 | ||
877 | An arbitrary per-transport string can be expanded on successful delivery, | |
878 | and (for SMTP transports) a second string on deferrals caused by a host error. | |
879 | This feature may be used, for example, to write exim internal log information | |
880 | (not available otherwise) into a database. | |
881 | ||
882 | In order to use the feature, you must set | |
883 | ||
884 | EXPERIMENTAL_TPDA=yes | |
885 | ||
886 | in your Local/Makefile | |
887 | ||
888 | and define the expandable strings in the runtime config file, to | |
889 | be executed at end of delivery. | |
890 | ||
891 | Additionally, there are 6 more variables, available at end of | |
892 | delivery: | |
893 | ||
894 | tpda_delivery_ip IP of host, which has accepted delivery | |
895 | tpda_delivery_port Port of remote host which has accepted delivery | |
896 | tpda_delivery_fqdn FQDN of host, which has accepted delivery | |
897 | tpda_delivery_local_part local part of address being delivered | |
898 | tpda_delivery_domain domain part of address being delivered | |
899 | tpda_delivery_confirmation SMTP confirmation message | |
900 | ||
901 | In case of a deferral caused by a host-error: | |
902 | tpda_defer_errno Error number | |
903 | tpda_defer_errstr Error string possibly containing more details | |
904 | ||
905 | The $router_name and $transport_name variables are also usable. | |
906 | ||
907 | ||
908 | To take action after successful deliveries, set the following option | |
909 | on any transport of interest. | |
910 | ||
911 | tpda_delivery_action | |
912 | ||
913 | An example might look like: | |
914 | ||
915 | tpda_delivery_action = \ | |
916 | ${lookup pgsql {SELECT * FROM record_Delivery( \ | |
917 | '${quote_pgsql:$sender_address_domain}',\ | |
918 | '${quote_pgsql:${lc:$sender_address_local_part}}', \ | |
919 | '${quote_pgsql:$tpda_delivery_domain}', \ | |
920 | '${quote_pgsql:${lc:$tpda_delivery_local_part}}', \ | |
921 | '${quote_pgsql:$tpda_delivery_ip}', \ | |
922 | '${quote_pgsql:${lc:$tpda_delivery_fqdn}}', \ | |
923 | '${quote_pgsql:$message_exim_id}')}} | |
924 | ||
925 | The string is expanded after the delivery completes and any | |
926 | side-effects will happen. The result is then discarded. | |
927 | Note that for complex operations an ACL expansion can be used. | |
928 | ||
929 | ||
930 | In order to log host deferrals, add the following option to an SMTP | |
931 | transport: | |
932 | ||
933 | tpda_host_defer_action | |
934 | ||
935 | This is a private option of the SMTP transport. It is intended to | |
936 | log failures of remote hosts. It is executed only when exim has | |
937 | attempted to deliver a message to a remote host and failed due to | |
938 | an error which doesn't seem to be related to the individual | |
939 | message, sender, or recipient address. | |
7c261cc6 | 940 | See section 47.2 of the exim documentation for more details on how |
d68218c7 JH |
941 | this is determined. |
942 | ||
943 | Example: | |
944 | ||
945 | tpda_host_defer_action = \ | |
946 | ${lookup mysql {insert into delivlog set \ | |
947 | msgid = '${quote_mysql:$message_exim_id}', \ | |
948 | senderlp = '${quote_mysql:${lc:$sender_address_local_part}}', \ | |
949 | senderdom = '${quote_mysql:$sender_address_domain}', \ | |
950 | delivlp = '${quote_mysql:${lc:$tpda_delivery_local_part}}', \ | |
951 | delivdom = '${quote_mysql:$tpda_delivery_domain}', \ | |
952 | delivip = '${quote_mysql:$tpda_delivery_ip}', \ | |
953 | delivport = '${quote_mysql:$tpda_delivery_port}', \ | |
954 | delivfqdn = '${quote_mysql:$tpda_delivery_fqdn}', \ | |
955 | deliverrno = '${quote_mysql:$tpda_defer_errno}', \ | |
956 | deliverrstr = '${quote_mysql:$tpda_defer_errstr}' \ | |
957 | }} | |
1899bab2 | 958 | |
9bdd29ad TL |
959 | |
960 | Redis Lookup | |
961 | -------------------------------------------------------------- | |
962 | ||
963 | Redis is open source advanced key-value data store. This document | |
964 | does not explain the fundamentals, you should read and understand how | |
965 | it works by visiting the website at http://www.redis.io/. | |
966 | ||
967 | Redis lookup support is added via the hiredis library. Visit: | |
968 | ||
969 | https://github.com/redis/hiredis | |
970 | ||
971 | to obtain a copy, or find it in your operating systems package repository. | |
972 | If building from source, this description assumes that headers will be in | |
973 | /usr/local/include, and that the libraries are in /usr/local/lib. | |
974 | ||
975 | 1. In order to build exim with Redis lookup support add | |
976 | ||
977 | EXPERIMENTAL_REDIS=yes | |
978 | ||
979 | to your Local/Makefile. (Re-)build/install exim. exim -d should show | |
980 | Experimental_Redis in the line "Support for:". | |
981 | ||
982 | EXPERIMENTAL_REDIS=yes | |
983 | LDFLAGS += -lhiredis | |
984 | # CFLAGS += -I/usr/local/include | |
985 | # LDFLAGS += -L/usr/local/lib | |
986 | ||
987 | The first line sets the feature to include the correct code, and | |
988 | the second line says to link the hiredis libraries into the | |
989 | exim binary. The commented out lines should be uncommented if you | |
990 | built hiredis from source and installed in the default location. | |
991 | Adjust the paths if you installed them elsewhere, but you do not | |
992 | need to uncomment them if an rpm (or you) installed them in the | |
993 | package controlled locations (/usr/include and /usr/lib). | |
994 | ||
995 | ||
996 | 2. Use the following global settings to configure Redis lookup support: | |
997 | ||
998 | Required: | |
999 | redis_servers This option provides a list of Redis servers | |
1000 | and associated connection data, to be used in | |
1001 | conjunction with redis lookups. The option is | |
1002 | only available if Exim is configured with Redis | |
1003 | support. | |
1004 | ||
1005 | For example: | |
1006 | ||
1007 | redis_servers = 127.0.0.1/10/ - using database 10 with no password | |
1008 | redis_servers = 127.0.0.1//password - to make use of the default database of 0 with a password | |
1009 | redis_servers = 127.0.0.1// - for default database of 0 with no password | |
1010 | ||
1011 | 3. Once you have the Redis servers defined you can then make use of the | |
1012 | experimental Redis lookup by specifying ${lookup redis{}} in a lookup query. | |
1013 | ||
1014 | 4. Example usage: | |
1015 | ||
1016 | (Host List) | |
1017 | hostlist relay_from_ips = <\n ${lookup redis{SMEMBERS relay_from_ips}} | |
1018 | ||
1019 | Where relay_from_ips is a Redis set which contains entries such as "192.168.0.0/24" "10.0.0.0/8" and so on. | |
1020 | The result set is returned as | |
1021 | 192.168.0.0/24 | |
1022 | 10.0.0.0/8 | |
1023 | .. | |
1024 | . | |
1025 | ||
1026 | (Domain list) | |
1027 | domainlist virtual_domains = ${lookup redis {HGET $domain domain}} | |
1028 | ||
1029 | Where $domain is a hash which includes the key 'domain' and the value '$domain'. | |
1030 | ||
1031 | (Adding or updating an existing key) | |
1032 | set acl_c_spammer = ${if eq{${lookup redis{SPAMMER_SET}}}{OK}} | |
1033 | ||
1034 | Where SPAMMER_SET is a macro and it is defined as | |
1035 | ||
1036 | "SET SPAMMER <some_value>" | |
1037 | ||
1038 | (Getting a value from Redis) | |
1039 | ||
1040 | set acl_c_spam_host = ${lookup redis{GET...}} | |
1041 | ||
1042 | ||
a3c86431 TL |
1043 | Proxy Protocol Support |
1044 | -------------------------------------------------------------- | |
1045 | ||
1046 | Exim now has Experimental "Proxy Protocol" support. It was built on | |
1047 | specifications from: | |
1048 | http://haproxy.1wt.eu/download/1.5/doc/proxy-protocol.txt | |
1049 | ||
1050 | The purpose of this function is so that an application load balancer, | |
1051 | such as HAProxy, can sit in front of several Exim servers and Exim | |
1052 | will log the IP that is connecting to the proxy server instead of | |
1053 | the IP of the proxy server when it connects to Exim. It resets the | |
1054 | $sender_address_host and $sender_address_port to the IP:port of the | |
1055 | connection to the proxy. It also re-queries the DNS information for | |
1056 | this new IP address so that the original sender's hostname and IP | |
1057 | get logged in the Exim logfile. There is no logging if a host passes or | |
1058 | fails Proxy Protocol negotiation, but it can easily be determined and | |
1059 | recorded in an ACL (example is below). | |
1060 | ||
1061 | 1. To compile Exim with Proxy Protocol support, put this in | |
1062 | Local/Makefile: | |
1063 | ||
1064 | EXPERIMENTAL_PROXY=yes | |
1065 | ||
1066 | 2. Global configuration settings: | |
1067 | ||
1068 | proxy_required_hosts = HOSTLIST | |
1069 | ||
1070 | The proxy_required_hosts option will require any IP in that hostlist | |
1071 | to use Proxy Protocol. The specification of Proxy Protocol is very | |
1072 | strict, and if proxy negotiation fails, Exim will not allow any SMTP | |
1073 | command other than QUIT. (See end of this section for an example.) | |
1074 | The option is expanded when used, so it can be a hostlist as well as | |
1075 | string of IP addresses. Since it is expanded, specifying an alternate | |
1076 | separator is supported for ease of use with IPv6 addresses. | |
1077 | ||
1078 | To log the IP of the proxy in the incoming logline, add: | |
1079 | log_selector = +proxy | |
1080 | ||
1081 | A default incoming logline (wrapped for appearance) will look like this: | |
1082 | ||
1083 | 2013-11-04 09:25:06 1VdNti-0001OY-1V <= me@example.net | |
1084 | H=mail.example.net [1.2.3.4] P=esmtp S=433 | |
1085 | ||
1086 | With the log selector enabled, an email that was proxied through a | |
1087 | Proxy Protocol server at 192.168.1.2 will look like this: | |
1088 | ||
1089 | 2013-11-04 09:25:06 1VdNti-0001OY-1V <= me@example.net | |
1090 | H=mail.example.net [1.2.3.4] P=esmtp PRX=192.168.1.2 S=433 | |
1091 | ||
1092 | 3. In the ACL's the following expansion variables are available. | |
1093 | ||
eb57651e TL |
1094 | proxy_host_address The (internal) src IP of the proxy server |
1095 | making the connection to the Exim server. | |
1096 | proxy_host_port The (internal) src port the proxy server is | |
1097 | using to connect to the Exim server. | |
1098 | proxy_target_address The dest (public) IP of the remote host to | |
1099 | the proxy server. | |
1100 | proxy_target_port The dest port the remote host is using to | |
1101 | connect to the proxy server. | |
1102 | proxy_session Boolean, yes/no, the connected host is required | |
1103 | to use Proxy Protocol. | |
a3c86431 TL |
1104 | |
1105 | There is no expansion for a failed proxy session, however you can detect | |
1106 | it by checking if $proxy_session is true but $proxy_host is empty. As | |
1107 | an example, in my connect ACL, I have: | |
1108 | ||
1109 | warn condition = ${if and{ {bool{$proxy_session}} \ | |
a3bddaa8 | 1110 | {eq{$proxy_host_address}{}} } } |
a3c86431 TL |
1111 | log_message = Failed required proxy protocol negotiation \ |
1112 | from $sender_host_name [$sender_host_address] | |
1113 | ||
1114 | warn condition = ${if and{ {bool{$proxy_session}} \ | |
a3bddaa8 | 1115 | {!eq{$proxy_host_address}{}} } } |
a3c86431 | 1116 | # But don't log health probes from the proxy itself |
a3bddaa8 | 1117 | condition = ${if eq{$proxy_host_address}{$sender_host_address} \ |
a3c86431 TL |
1118 | {false}{true}} |
1119 | log_message = Successfully proxied from $sender_host_name \ | |
1120 | [$sender_host_address] through proxy protocol \ | |
a3bddaa8 | 1121 | host $proxy_host_address |
a3c86431 | 1122 | |
eb57651e TL |
1123 | # Possibly more clear |
1124 | warn logwrite = Remote Source Address: $sender_host_address:$sender_host_port | |
1125 | logwrite = Proxy Target Address: $proxy_target_address:$proxy_target_port | |
1126 | logwrite = Proxy Internal Address: $proxy_host_address:$proxy_host_port | |
1127 | logwrite = Internal Server Address: $received_ip_address:$received_port | |
1128 | ||
1129 | ||
a3c86431 TL |
1130 | 4. Runtime issues to be aware of: |
1131 | - Since the real connections are all coming from your proxy, and the | |
1132 | per host connection tracking is done before Proxy Protocol is | |
1133 | evaluated, smtp_accept_max_per_host must be set high enough to | |
1134 | handle all of the parallel volume you expect per inbound proxy. | |
1135 | - The proxy has 3 seconds (hard-coded in the source code) to send the | |
1136 | required Proxy Protocol header after it connects. If it does not, | |
1137 | the response to any commands will be: | |
1138 | "503 Command refused, required Proxy negotiation failed" | |
1139 | - If the incoming connection is configured in Exim to be a Proxy | |
1140 | Protocol host, but the proxy is not sending the header, the banner | |
1141 | does not get sent until the timeout occurs. If the sending host | |
1142 | sent any input (before the banner), this causes a standard Exim | |
1143 | synchronization error (i.e. trying to pipeline before PIPELINING | |
1144 | was advertised). | |
1145 | - This is not advised, but is mentioned for completeness if you have | |
1146 | a specific internal configuration that you want this: If the Exim | |
1147 | server only has an internal IP address and no other machines in your | |
1148 | organization will connect to it to try to send email, you may | |
1149 | simply set the hostlist to "*", however, this will prevent local | |
1150 | mail programs from working because that would require mail from | |
1151 | localhost to use Proxy Protocol. Again, not advised! | |
1152 | ||
1153 | 5. Example of a refused connection because the Proxy Protocol header was | |
1154 | not sent from a host configured to use Proxy Protocol. In the example, | |
1155 | the 3 second timeout occurred (when a Proxy Protocol banner should have | |
1156 | been sent), the banner was displayed to the user, but all commands are | |
1157 | rejected except for QUIT: | |
1158 | ||
1159 | # nc mail.example.net 25 | |
1160 | 220-mail.example.net, ESMTP Exim 4.82+proxy, Mon, 04 Nov 2013 10:45:59 | |
1161 | 220 -0800 RFC's enforced | |
1162 | EHLO localhost | |
1163 | 503 Command refused, required Proxy negotiation failed | |
1164 | QUIT | |
1165 | 221 mail.example.net closing connection | |
1166 | ||
1167 | ||
9bdd29ad | 1168 | |
ee161e8f PH |
1169 | -------------------------------------------------------------- |
1170 | End of file | |
1171 | -------------------------------------------------------------- |