[exim.git] / doc / doc-txt / NewStuff
e3a311ba 1$Cambridge: exim/doc/doc-txt/NewStuff,v 1.55 2005/07/23 20:59:16 tom Exp $
3New Features in Exim
6This file contains descriptions of new features that have been added to Exim,
7but have not yet made it into the main manual (which is most conveniently
8updated when there is a relatively large batch of changes). The doc/ChangeLog
9file contains a listing of all changes, including bug fixes.
11Exim version 4.53
14TK/01 Added the "success_on_redirect" address verification option. When an
15 address generates new addresses during routing, Exim will abort
16 verification with "success" when more than one address has been
17 generated, but continue to verify a single new address. The latter
18 does not happen when the new "success_on_redirect" option is set, like
20 require verify = recipient/success_on_redirect/callout=10s
22 In that case, verification will succeed when a router generates a new
23 address.
495ae4b0 25
26Exim version 4.52
29TF/01 Support for checking Client SMTP Authorization has been added. CSA is a
30 system which allows a site to advertise which machines are and are not
31 permitted to send email. This is done by placing special SRV records in
32 the DNS, which are looked up using the client's HELO domain. At this
33 time CSA is still an Internet-Draft.
35 Client SMTP Authorization checks are performed by the ACL condition
36 verify=csa. This will fail if the client is not authorized. If there is
37 a DNS problem, or if no valid CSA SRV record is found, or if the client
38 is authorized, the condition succeeds. These three cases can be
39 distinguished using the expansion variable $csa_status, which can take
40 one of the values "fail", "defer", "unknown", or "ok". The condition
41 does not itself defer because that would be likely to cause problems
42 for legitimate email.
44 The error messages produced by the CSA code include slightly more
45 detail. If $csa_status is "defer" this may be because of problems
46 looking up the CSA SRV record, or problems looking up the CSA target
47 address record. There are four reasons for $csa_status being "fail":
48 the client's host name is explicitly not authorized; the client's IP
49 address does not match any of the CSA target IP addresses; the client's
50 host name is authorized but it has no valid target IP addresses (e.g.
51 the target's addresses are IPv6 and the client is using IPv4); or the
52 client's host name has no CSA SRV record but a parent domain has
53 asserted that all subdomains must be explicitly authorized.
55 The verify=csa condition can take an argument which is the domain to
56 use for the DNS query. The default is verify=csa/$sender_helo_name.
58 This implementation includes an extension to CSA. If the query domain
59 is an address literal such as [], or if it is a bare IP
60 address, Exim will search for CSA SRV records in the reverse DNS as if
61 the HELO domain was e.g. Therefore it is
62 meaningful to say, for example, verify=csa/$sender_host_address - in
63 fact, this is the check that Exim performs if the client does not say
64 HELO. This extension can be turned off by setting the main
65 configuration option dns_csa_use_reverse = false.
67 If a CSA SRV record is not found for the domain itself, then a search
68 is performed through its parent domains for a record which might be
69 making assertions about subdomains. The maximum depth of this search is
70 limited using the main configuration option dns_csa_search_limit, which
71 takes the value 5 by default. Exim does not look for CSA SRV records in
72 a top level domain, so the default settings handle HELO domains as long
73 as seven ( which encompasses the
74 vast majority of legitimate HELO domains.
76 The dnsdb lookup also has support for CSA. Although dnsdb already
77 supports SRV lookups, this is not sufficient because of the extra
78 parent domain search behaviour of CSA, and (as with PTR lookups)
79 dnsdb also turns IP addresses into lookups in the reverse DNS space.
80 The result of ${lookup dnsdb {csa=$sender_helo_name} } has two
81 space-separated fields: an authorization code and a target host name.
82 The authorization code can be "Y" for yes, "N" for no, "X" for explicit
83 authorization required but absent, or "?" for unknown.
85PH/01 The amount of output produced by the "make" process has been reduced,
86 because the compile lines are often rather long, making it all pretty
87 unreadable. The new style is along the lines of the 2.6 Linux kernel:
88 just a short line for each module that is being compiled or linked.
89 However, it is still possible to get the full output, by calling "make"
90 like this:
92 FULLECHO='' make -e
94 The value of FULLECHO defaults to "@", the flag character that suppresses
95 command reflection in "make". When you ask for the full output, it is
96 given in addition to the the short output.
4df1e33e 98TF/02 There have been two changes concerned with submission mode:
87ba3f5f 99
100 Until now submission mode always left the return path alone, whereas
101 locally-submitted messages from untrusted users have the return path
102 fixed to the user's email address. Submission mode now fixes the return
103 path to the same address as is used to create the Sender: header. If
104 /sender_retain is specified then both the Sender: header and the return
105 path are left alone.
87ba3f5f 106
107 Note that the changes caused by submission mode take effect after the
108 predata ACL. This means that any sender checks performed before the
109 fix-ups will use the untrusted sender address specified by the user, not
110 the trusted sender address specified by submission mode. Although this
111 might be slightly unexpected, it does mean that you can configure ACL
112 checks to spot that a user is trying to spoof another's address, for
113 example.
87ba3f5f 114
115 There is also a new /name= option for submission mode which allows you
116 to specify the user's full name to be included in the Sender: header.
117 For example:
119 accept authenticated = *
120 control = submission/name=${lookup {$authenticated_id} \
121 lsearch {/etc/exim/namelist} }
123 The namelist file contains entries like
125 fanf: Tony Finch
127 And the resulting Sender: header looks like
129 Sender: Tony Finch <>
131TF/03 The control = fakereject ACL modifier now has a fakedefer counterpart,
132 which works in exactly the same way except it causes a fake SMTP 450
133 response after the message data instead of a fake SMTP 550 response.
134 You must take care when using fakedefer because it will cause messages
135 to be duplicated when the sender retries. Therefore you should not use
136 fakedefer if the message will be delivered normally.
138TF/04 There is a new ratelimit ACL condition which can be used to measure
139 and control the rate at which clients can send email. This is more
140 powerful than the existing smtp_ratelimit_* options, because those
141 options only control the rate of commands in a single SMTP session,
142 whereas the new ratelimit condition works across all connections
143 (concurrent and sequential) to the same host.
145 The syntax of the ratelimit condition is:
147 ratelimit = <m> / <p> / <options> / <key>
149 If the average client sending rate is less than m messages per time
150 period p then the condition is false, otherwise it is true.
152 The parameter p is the smoothing time constant, in the form of an Exim
153 time interval e.g. 8h for eight hours. A larger time constant means it
154 takes Exim longer to forget a client's past behaviour. The parameter m is
155 the maximum number of messages that a client can send in a fast burst. By
156 increasing both m and p but keeping m/p constant, you can allow a client
157 to send more messages in a burst without changing its overall sending
158 rate limit. Conversely, if m and p are both small then messages must be
159 sent at an even rate.
161 The key is used to look up the data used to calcluate the client's
162 average sending rate. This data is stored in a database maintained by
163 Exim in its spool directory alongside the retry database etc. For
164 example, you can limit the sending rate of each authenticated user,
165 independent of the computer they are sending from, by setting the key
166 to $authenticated_id. The default key is $sender_host_address.
168 Each ratelimit condition can have up to two options. The first option
169 specifies what Exim measures the rate of, and the second specifies how
170 Exim handles excessively fast clients.
172 The per_mail option means that it measures the client's rate of sending
173 messages. This is the default if none of the per_* options is specified.
175 The per_conn option means that it measures the client's connection rate.
177 The per_byte option limits the sender's email bandwidth. Note that it
178 is best to use this option in the DATA ACL; if it is used in an earlier
179 ACL it relies on the SIZE parameter on the MAIL command, which may be
180 inaccurate or completely missing. You can follow the limit m in the
181 configuration with K, M, or G to specify limits in kilobytes,
182 megabytes, or gigabytes respectively.
184 The per_cmd option means that Exim recomputes the rate every time the
185 condition is processed, which can be used to limit the SMTP command rate.
186 The alias per_rcpt is provided for use in the RCPT ACL instead of per_cmd
187 to make it clear that the effect is to limit the rate at which recipients
188 are accepted. Note that in this case the rate limiting engine will see a
189 message with many recipients as a large high-speed burst.
191 If a client's average rate is greater than the maximum, the rate
192 limiting engine can react in two possible ways, depending on the
193 presence of the strict or leaky options. This is independent of the
194 other counter-measures (e.g. rejecting the message) that may be
195 specified by the rest of the ACL. The default mode is leaky, which
196 avoids a sender's over-aggressive retry rate preventing it from getting
197 any email through.
199 The strict option means that the client's recorded rate is always
200 updated. The effect of this is that Exim measures the client's average
201 rate of attempts to send email, which can be much higher than the
202 maximum. If the client is over the limit it will be subjected to
203 counter-measures until it slows down below the maximum rate.
205 The leaky option means that the client's recorded rate is not updated
206 if it is above the limit. The effect of this is that Exim measures the
207 client's average rate of successfully sent email, which cannot be
208 greater than the maximum. If the client is over the limit it will
209 suffer some counter-measures, but it will still be able to send email
210 at the configured maximum rate, whatever the rate of its attempts.
212 As a side-effect, the ratelimit condition will set the expansion
213 variables $sender_rate containing the client's computed rate,
214 $sender_rate_limit containing the configured value of m, and
215 $sender_rate_period containing the configured value of p.
217 Exim's other ACL facilities are used to define what counter-measures
218 are taken when the rate limit is exceeded. This might be anything from
219 logging a warning (e.g. while measuring existing sending rates in order
220 to define our policy), through time delays to slow down fast senders,
221 up to rejecting the message. For example,
223 # Log all senders' rates
224 warn
225 ratelimit = 0 / 1h / strict
226 log_message = \
227 Sender rate $sender_rate > $sender_rate_limit / $sender_rate_period
229 # Slow down fast senders
230 warn
231 ratelimit = 100 / 1h / per_rcpt / strict
232 delay = ${eval: 10 * ($sender_rate - $sender_rate_limit) }
234 # Keep authenticated users under control
235 deny
236 ratelimit = 100 / 1d / strict / $authenticated_id
238 # System-wide rate limit
239 defer
240 message = Sorry, too busy. Try again later.
241 ratelimit = 10 / 1s / $primary_hostname
243 # Restrict incoming rate from each host, with a default rate limit
244 # set using a macro and special cases looked up in a table.
245 defer
246 message = Sender rate $sender_rate exceeds \
247 $sender_rate_limit messages per $sender_rate_period
248 ratelimit = ${lookup {$sender_host_address} \
249 cdb {DB/ratelimits.cdb} \
250 {$value} {RATELIMIT} }
252 Warning: if you have a busy server with a lot of ratelimit tests,
253 especially with the per_rcpt option, you may suffer from a performance
254 bottleneck caused by locking on the ratelimit hints database. Apart from
255 making your ACLs less complicated, you can reduce the problem by using a
256 RAM disk for Exim's hints directory, /var/spool/exim/db/. However this
257 means that Exim will lose its hints data after a reboot (including retry
258 hints, the callout cache, and ratelimit data).
260TK/01 Added an 'spf' lookup type that will return an SPF result for a given
261 email address (the key) and an IP address (the database):
263 ${lookup {} spf{}}
265 The lookup will return the same result strings as they can appear in
266 $spf_result (pass,fail,softfail,neutral,none,err_perm,err_temp). The
267 lookup is armored in EXPERIMENTAL_SPF. Currently, only IPv4 addresses
268 are supported.
270 Patch submitted by Chris Webb <>.
272PH/02 There's a new verify callout option, "fullpostmaster", which first acts
273 as "postmaster" and checks the recipient <postmaster@domain>. If that
274 fails, it tries just <postmaster>, without a domain, in accordance with
275 the specification in RFC 2821.
277PH/03 The action of the auto_thaw option has been changed. It no longer applies
278 to frozen bounce messages.
280TK/02 There are two new expansion items to help with the implementation of
281 the BATV "prvs" scheme in an Exim configuration:
284 ${prvs {<ADDRESS>}{<KEY>}{[KEYNUM]}}
286 The "prvs" expansion item takes three arguments: A qualified RFC2821
287 email address, a key and an (optional) key number. All arguments are
288 expanded before being used, so it is easily possible to lookup a key
289 and key number using the address as the lookup key. The key number is
290 optional and defaults to "0". The item will expand to a "prvs"-signed
291 email address, to be typically used with the "return_path" option on
292 a smtp transport. The decision if BATV should be used with a given
293 sender/recipient pair should be done on router level, to avoid having
294 to set "max_rcpt = 1" on the transport.
297 ${prvscheck {<ADDRESS>}{<SECRET>}{<RETURN_STRING>}}
299 The "prvscheck" expansion item takes three arguments. Argument 1 is
300 expanded first. When the expansion does not yield a SYNTACTICALLY
301 valid "prvs"-scheme address, the whole "prvscheck" item expands to
302 the empty string. If <ADDRESS> is a "prvs"-encoded address after
303 expansion, two expansion variables are set up:
305 $prvscheck_address Contains the "prvs"-decoded version of
306 the address from argument 1.
308 $prvscheck_keynum Contains the key number extracted from
309 the "prvs"-address in argument 1.
311 These two variables can be used in the expansion code of argument 2
312 to retrieve the <SECRET>. The VALIDITY of the "prvs"-signed address
313 is then checked. The result is stored in yet another expansion
314 variable:
316 $prvscheck_result Contains the result of a "prvscheck"
317 expansion: Unset (the empty string) for
318 failure, "1" for success.
320 The "prvscheck" expansion expands to the empty string if <ADDRESS>
321 is not a SYNTACTICALLY valid "prvs"-scheme address. Otherwise,
322 argument 3 defines what "prvscheck" expands to: If argument 3
323 is the empty string, "prvscheck" expands to the decoded version
324 of the address (no matter if it is CRYPTOGRAPHICALLY valid or not).
325 If argument 3 expands to a non-empty string, "prvscheck" expands
326 to that string.
329 Usage example
330 -------------
332 Macro:
334 PRVSCHECK_SQL = ${lookup mysql{SELECT secret FROM batv_prvs WHERE \
335 sender='${quote_mysql:$prvscheck_address}'}{$value}}
339 # Bounces: drop unsigned addresses for BATV senders
340 deny message = This address does not send an unsigned reverse path.
341 senders = :
342 recipients = +batv_recipients
344 # Bounces: In case of prvs-signed address, check signature.
345 deny message = Invalid reverse path signature.
346 senders = :
347 condition = ${prvscheck {$local_part@$domain}{PRVSCHECK_SQL}{1}}
348 !condition = $prvscheck_result
350 Top-Level Router:
352 batv_redirect:
353 driver = redirect
354 data = ${prvscheck {$local_part@$domain}{PRVSCHECK_SQL}{}}
356 Transport (referenced by router that makes decision if
357 BATV is applicable):
359 external_smtp_batv:
360 driver = smtp
361 return_path = ${prvs {$return_path} \
362 {${lookup mysql{SELECT \
363 secret FROM batv_prvs WHERE \
364 sender='${quote_mysql:$sender_address}'} \
365 {$value}fail}}}
367PH/04 There are two new options that control the retrying done by the daemon
368 at startup when it cannot immediately bind a socket (typically because
369 the socket is already in use). The default values reproduce what were
370 built-in constants previously: daemon_startup_retries defines the number
371 of retries after the first failure (default 9); daemon_startup_sleep
372 defines the length of time to wait between retries (default 30s).
0cd68797 373
374PH/05 There is now a new ${if condition called "match_ip". It is similar to
375 match_domain, etc. It must be followed by two argument strings. The first
376 (after expansion) must be an IP address or an empty string. The second
377 (after expansion) is a restricted host list that can match only an IP
378 address, not a host name. For example:
380 ${if match_ip{$sender_host_address}{}{...}{...}}
382 The specific types of host list item that are permitted in the list are
383 shown below. Consult the manual section on host lists for further
384 details.
386 . An IP address, optionally with a CIDR mask.
388 . A single asterisk matches any IP address.
390 . An empty item matches only if the IP address is empty. This could be
391 useful for testing for a locally submitted message or one from specific
392 hosts in a single test such as
394 ${if match_ip{$sender_host_address}{:}{...}{...}}
396 where the first item in the list is the empty string.
398 . The item @[] matches any of the local host's interface addresses.
400 . Lookups are assumed to be "net-" style lookups, even if "net-" is not
401 specified. Thus, the following are equivalent:
403 ${if match_ip{$sender_host_address}{lsearch;/some/file}...
404 ${if match_ip{$sender_host_address}{net-lsearch;/some/file}...
406 You do need to specify the "net-" prefix if you want to specify a
407 specific address mask, for example, by using "net24-".
409PH/06 The "+all" debug selector used to set the flags for all possible output;
410 it is something that people tend to use semi-automatically when
411 generating debug output for me or for the list. However, by including
412 "+memory", an awful lot of output that is very rarely of interest was
413 generated. I have changed this so that "+all" no longer includes
414 "+memory". However, "-all" still turns everything off.
e5a9dba6 416
417Version 4.51
420PH/01 The format in which GnuTLS parameters are written to the gnutls-param
421 file in the spool directory has been changed. This change has been made
422 to alleviate problems that some people had with the generation of the
423 parameters by Exim when /dev/random was exhausted. In this situation,
424 Exim would hang until /dev/random acquired some more entropy.
426 The new code exports and imports the DH and RSA parameters in PEM
427 format. This means that the parameters can be generated externally using
428 the certtool command that is part of GnuTLS.
430 To replace the parameters with new ones, instead of deleting the file
431 and letting Exim re-create it, you can generate new parameters using
432 certtool and, when this has been done, replace Exim's cache file by
433 renaming. The relevant commands are something like this:
435 # rm -f new.params
436 # touch new.params
437 # chown exim:exim new.params
438 # chmod 0400 new.params
439 # certtool --generate-privkey --bits 512 >new.params
440 # echo "" >>new.params
441 # certtool --generate-dh-params --bits 1024 >> new.params
442 # mv new.params params
444 If Exim never has to generate the parameters itself, the possibility of
445 stalling is removed.
447PH/02 A new expansion item for dynamically loading and calling a locally-
448 written C function is now provided, if Exim is compiled with
452 set in Local/Makefile. The facility is not included by default (a
453 suitable error is given if you try to use it when it is not there.)
455 If you enable EXPAND_DLFUNC, you should also be aware of the new redirect
456 router option forbid_filter_dlfunc. If you have unprivileged users on
457 your system who are permitted to create filter files, you might want to
458 set forbid_filter_dlfunc=true in the appropriate router, to stop them
459 using ${dlfunc to run code within Exim.
461 You load and call an external function like this:
463 ${dlfunc{/some/file}{function}{arg1}{arg2}...}
465 Once loaded, Exim remembers the dynamically loaded object so that it
466 doesn't reload the same object file in the same Exim process (but of
467 course Exim does start new processes frequently).
469 There may be from zero to eight arguments to the function. When compiling
470 a local function that is to be called in this way, local_scan.h should be
471 included. The Exim variables and functions that are defined by that API
472 are also available for dynamically loaded functions. The function itself
473 must have the following type:
475 int dlfunction(uschar **yield, int argc, uschar *argv[])
477 Where "uschar" is a typedef for "unsigned char" in local_scan.h. The
478 function should return one of the following values:
480 OK Success. The string that is placed in "yield" is put into
481 the expanded string that is being built.
483 FAIL A non-forced expansion failure occurs, with the error
484 message taken from "yield", if it is set.
486 FAIL_FORCED A forced expansion failure occurs, with the error message
487 taken from "yield" if it is set.
489 ERROR Same as FAIL, except that a panic log entry is written.
491 When compiling a function that is to be used in this way with gcc,
492 you need to add -shared to the gcc command. Also, in the Exim build-time
493 configuration, you must add -export-dynamic to EXTRALIBS.
b5aea5e1 494
495TF/01 $received_time is a new expansion variable containing the time and date
496 as a number of seconds since the start of the Unix epoch when the
497 current message was received.
b5aea5e1 498
499PH/03 There is a new value for RADIUS_LIB_TYPE that can be set in
500 Local/Makefile. It is RADIUSCLIENTNEW, and it requests that the new API,
501 in use from radiusclient 0.4.0 onwards, be used. It does not appear to be
502 possible to detect the different versions automatically.
504PH/04 There is a new option called acl_not_smtp_mime that allows you to scan
505 MIME parts in non-SMTP messages. It operates in exactly the same way as
506 acl_smtp_mime
508PH/05 It is now possible to redefine a macro within the configuration file.
509 The macro must have been previously defined within the configuration (or
510 an included file). A definition on the command line using the -D option
511 causes all definitions and redefinitions within the file to be ignored.
512 In other words, -D overrides any values that are set in the file.
513 Redefinition is specified by using '==' instead of '='. For example:
515 MAC1 = initial value
516 ...
517 MAC1 == updated value
519 Redefinition does not alter the order in which the macros are applied to
520 the subsequent lines of the configuration file. It is still the same
521 order in which the macros were originally defined. All that changes is
522 the macro's value. Redefinition makes it possible to accumulate values.
523 For example:
525 MAC1 = initial value
526 ...
527 MAC1 == MAC1 and something added
529 This can be helpful in situations where the configuration file is built
530 from a number of other files.
532PH/06 Macros may now be defined or redefined between router, transport,
533 authenticator, or ACL definitions, as well as in the main part of the
534 configuration. They may not, however, be changed within an individual
535 driver or ACL, or in the local_scan, retry, or rewrite sections of the
536 configuration.
538PH/07 $acl_verify_message is now set immediately after the failure of a
539 verification in an ACL, and so is available in subsequent modifiers. In
540 particular, the message can be preserved by coding like this:
542 warn !verify = sender
543 set acl_m0 = $acl_verify_message
545 Previously, $acl_verify_message was set only while expanding "message"
546 and "log_message" when a very denied access.
548PH/08 The redirect router has two new options, sieve_useraddress and
549 sieve_subaddress. These are passed to a Sieve filter to specify the :user
550 and :subaddress parts of an address. Both options are unset by default.
551 However, when a Sieve filter is run, if sieve_useraddress is unset, the
552 entire original local part (including any prefix or suffix) is used for
553 :user. An unset subaddress is treated as an empty subaddress.
475fe28a 554
555PH/09 Quota values can be followed by G as well as K and M.
557PH/10 $message_linecount is a new variable that contains the total number of
558 lines in the header and body of the message. Compare $body_linecount,
559 which is the count for the body only. During the DATA and
560 content-scanning ACLs, $message_linecount contains the number of lines
561 received. Before delivery happens (that is, before filters, routers, and
562 transports run) the count is increased to include the Received: header
563 line that Exim standardly adds, and also any other header lines that are
564 added by ACLs. The blank line that separates the message header from the
565 body is not counted. Here is an example of the use of this variable in a
568 deny message = Too many lines in message header
569 condition = \
570 ${if <{250}{${eval: $message_linecount - $body_linecount}}}
572 In the MAIL and RCPT ACLs, the value is zero because at that stage the
573 message has not yet been received.
575PH/11 In a ${run expansion, the variable $value (which contains the standard
576 output) is now also usable in the "else" string.
578PH/12 In a pipe transport, although a timeout while waiting for the pipe
579 process to complete was treated as a delivery failure, a timeout while
580 writing the message to the pipe was logged, but erroneously treated as a
581 successful delivery. Such timeouts include transport filter timeouts. For
582 consistency with the overall process timeout, these timeouts are now
583 treated as errors, giving rise to delivery failures by default. However,
584 there is now a new Boolean option for the pipe transport called
585 timeout_defer, which, if set TRUE, converts the failures into defers for
586 both kinds of timeout. A transport filter timeout is now identified in
587 the log output.
7766a4f0 589
f7b63901 590Version 4.50
b9e40c51 593The documentation is up-to-date for the 4.50 release.