1MIMEDEFANG-FILTER(5) File Formats Manual MIMEDEFANG-FILTER(5)
2
6 mimedefang-filter - Configuration file for MIMEDefang mail filter.
7
10 mimedefang-filter is a Perl fragment that controls how mimedefang.pl
11 disposes of various parts of a MIME message. In addition, it contains
12 some global variable settings that affect the operation of mimede‐
13 fang.pl.
14
17 Incoming messages are scanned as follows:
18 1) A temporary working directory is created. It is made the current
21 working directory and the e-mail message is split into parts in this
22 directory. Each part is represented internally as an instance of
23 MIME::Entity.
24 2) If the file /etc/mail/mimedefang-filter.pl defines a Perl function
27 called filter_begin, it is called with a single argument consisting of
28 a MIME::Entity representing the parsed e-mail message. Any return
29 value is ignored.
30 3) For each leaf part of the mail message, filter is called with four
33 arguments: entity, a MIME::Entity object; fname, the suggested filename
34 taken from the MIME Content-Disposition header; ext, the file exten‐
35 sion, and type, the MIME Content-Type value. For each non-leaf part of
36 the mail message, filter_multipart is called with the same four argu‐
37 ments as filter. A non-leaf part of a message is a part that contains
38 nested parts. Such a part has no useful body, but you should still
39 perform filename checks to check for viruses that use malformed MIME to
40 masquerade as non-leaf parts (like message/rfc822). In general, any
41 action you perform in filter_multipart applies to the part itself and
42 any contained parts.
43 Note that both filter and filter_multipart are optional. If you do not
46 define them, a default function that simply accepts each part is used.
47 4) After all parts have been processed, the function filter_end is
50 called if it has been defined. It is passed a single argument consist‐
51 ing of the (possibly modified) MIME::Entity object representing the
52 message about to be delivered. Within filter_end, you can call func‐
53 tions that modify the message headers body.
54 5) After filter_end returns, the function filter_wrapup is called if it
57 has been defined. It is passed a single argument consisting of the
58 (possibly modified) MIME::Entity object representing the message about
59 to be delivered, including any modifications made in filter_end.
60 Within filter_wrapup, you can not call functions that modify the mes‐
61 sage body, but you can still add or modify message headers.
62
65 mimedefang.pl examines each part of the MIME message and chooses a dis‐
66 position for that part. (A disposition is selected by calling one of
67 the following functions from filter and then immediately returning.)
68 Available dispositions are:
69 action_accept
72 The part is passed through unchanged. If no disposition func‐
73 tion is returned, this is the default.
74 action_accept_with_warning
77 The part is passed through unchanged, but a warning is added to
78 the mail message.
79 action_drop
82 The part is deleted without any notification to the recipients.
83 action_drop_with_warning
86 The part is deleted and a warning is added to the mail message.
87 action_replace_with_warning
90 The part is deleted and instead replaced with a text message.
91 action_quarantine
94 The part is deleted and a warning is added to the mail message.
95 In addition, a copy of the part is saved on the mail server in
96 the directory /var/spool/MD-Quarantine and a notification is
97 sent to the MIMEDefang administrator.
98 action_bounce
101 The entire e-mail message is rejected and an error returned to
102 the sender. The intended recipients are not notified. Note
103 that in spite of the name, MIMEDefang does not generate and e-
104 mail a failure notification. Rather, it causes the SMTP server
105 to return a 5XX SMTP failure code.
106 action_discard
109 The entire e-mail message is discarded silently. Neither the
110 sender nor the intended recipients are notified.
111
114 You can define a function called filter_relay in your filter. This
115 lets you reject SMTP connection attempts early on in the SMTP dialog,
116 rather than waiting until the whole message has been sent. Note that
117 for this check to take place, you must use the -r flag with mimedefang.
118 filter_relay is passed six arguments: $hostip is the IP address of the
121 relay host (for example, "127.0.0.1"), $hostname is the host name if
122 known (for example, "localhost.localdomain"). If the host name could
123 not be determined, $hostname is $hostip enclosed in square brackets.
124 (That is, ("$hostname" eq "[$hostip]") will be true.)
125 The remaining four arguments to filter_relay are $port, $myip, $myport
128 and $qid, which contain the client's TCP port, the Sendmail daemon's
129 listening IP address, the Sendmail daemon's listening port, and the
130 Sendmail Queue-ID, respectively. Note that the Queue-ID may not yet be
131 available at this stage (for example, Postfix does not allocate a
132 queue-ID this early.) If the Queue-ID is not available, the string NO‐
133 QUEUE is passed instead.
134 filter_relay must return a two-element list: ($code, $msg). $msg spec‐
137 ifies the text message to use for the SMTP reply, but because of limi‐
138 tations in the Milter API, this message is for documentation purposes
139 only---you cannot set the text of the SMTP message returned to the SMTP
140 client from filter_relay.
141 $code is a literal string, and can have one of the following values:
143 'REJECT'
146 if the connection should be rejected.
147 'CONTINUE'
150 if the connection should be accepted.
151 'TEMPFAIL'
154 if a temporary failure code should be returned.
155 'DISCARD'
158 if the message should be accepted and silently discarded.
159 'ACCEPT_AND_NO_MORE_FILTERING'
162 if the connection should be accepted and no further filtering
163 done.
164 Earlier versions of MIMEDefang used -1 for TEMPFAIL, 0 for REJECT and 1
167 for CONTINUE. These values still work, but are deprecated.
168 In the case of REJECT or TEMPFAIL, $msg specifies the text part of the
171 SMTP reply. $msg must not contain newlines.
172 For example, if you wish to reject connection attempts from any machine
175 in the spammer.com domain, you could use this function:
176 sub filter_relay {
178 my ($ip, $name) = @_;
179 if ($name =~ /spammer\.com$/) {
180 return ('REJECT', "Sorry; spammer.com is blacklisted");
181 }
182 return ('CONTINUE', "ok");
183 }
184
187 You can define a function called filter_helo in your filter. This lets
188 you reject connections after the HELO/EHLO SMTP command. Note that for
189 this function to be called, you must use the -H flag with mimedefang.
190 filter_helo is passed seven arguments: $ip and $name are the IP address
193 and name of the sending relay, as in filter_relay. The third argument,
194 $helo, is the argument supplied in the HELO/EHLO command.
195 The remaining four arguments to filter_helo are $port, $myip, $myport
198 and $qid, which contain the client's TCP port, the Sendmail daemon's
199 listening IP address, the Sendmail daemon's listening port, and the
200 Sendmail Queue-ID, respectively. Note that the Queue-ID may not yet be
201 available at this stage (for example, Postfix does not allocate a
202 queue-ID this early.) If the Queue-ID is not available, the string NO‐
203 QUEUE is passed instead.
204 filter_helo must return a two-to-five element list: ($code, $msg,
207 $smtp_code, $smtp_dsn, $delay). $code is a return code, with the same
208 meaning as the $code return from filter_relay. $msg specifies the text
209 message to use for the SMTP reply. If $smtp_code and $smtp_dsn are
210 supplied, they become the SMTP numerical reply code and the enhanced
211 status delivery code (DSN code). If they are not supplied, sensible
212 defaults are used. $delay specifies a delay in seconds; the C milter
213 code will sleep for $delay seconds before returning the reply to Send‐
214 mail. $delay defaults to zero.
215 (Note that the delay is implemented in the Milter C code; if you spec‐
217 ify a delay of 30 seconds, that doesn't mean a Perl worker is tied up
218 for the duration of the delay. The delay only costs one Milter
219 thread.)
220
223 You can define a function called filter_sender in your filter. This
224 lets you reject messages from certain senders, rather than waiting un‐
225 til the whole message has been sent. Note that for this check to take
226 place, you must use the -s flag with mimedefang.
227 filter_sender is passed four arguments: $sender is the envelope e-mail
230 address of the sender (for example, "<dfs@roaringpenguin.com>"). The
231 address may or may not be surrounded by angle brackets. $ip and $name
232 are the IP address and host name of the SMTP relay. Finally, $helo is
233 the argument to the SMTP "HELO" command.
234 Inside filter_sender, you can access any ESMTP arguments (such as
237 "SIZE=12345") in the array @ESMTPArgs. Each ESMTP argument occupies
238 one array element.
239 filter_sender must return a two-to-five element list, with the same
242 meaning as the return value from filter_helo.
243 For example, if you wish to reject messages from spammer@badguy.com,
246 you could use this function:
247 sub filter_sender {
249 my ($sender, $ip, $hostname, $helo) = @_;
250 if ($sender =~ /^<?spammer\@badguy\.com>?$/i) {
251 return ('REJECT', 'Sorry; spammer@badguy.com is blacklisted.');
252 }
253 return ('CONTINUE', "ok");
254 }
255 As another example, some spammers identify their own machine as your
258 machine in the SMTP "HELO" command. This function rejects a machine
259 claiming to be in the "roaringpenguin.com" domain unless it really is a
260 Roaring Penguin machine:
261 sub filter_sender {
263 my($sender, $ip, $hostname, $helo) = @_;
264 if ($helo =~ /roaringpenguin.com/i) {
265 if ($ip ne "127.0.0.1" and
266 $ip ne "216.191.236.23" and
267 $ip ne "216.191.236.30") {
268 return('REJECT', "Go away... $ip is not in roaringpenguin.com");
269 }
270 }
271 return ('CONTINUE', "ok");
272 }
273 As a third example, you may wish to prevent spoofs by requiring SMTP
276 authentication when email is sent from some email addresses. This func‐
277 tion rejects mail from "king@example.com", unless the connecting user
278 properly authenticated as "elvisp". Note that this needs access to the
279 %SendmailMacros global, that is not available in filter_sender until
280 after a call to read_commands_file.
281 sub filter_sender {
283 my($sender, $ip, $hostname, $helo) = @_;
284 read_commands_file();
285 ### notice: This assumes The King uses authentication without realm!
286 if ($sender =~ /^<?king\@example\.com>?$/i and
287 $SendmailMacros{auth_authen} ne "elvisp") {
288 return('REJECT', "Faking mail from the king is not allowed.");
289 }
290 return ('CONTINUE', "ok");
291 }
292
296 You can define a function called filter_recipient in your filter. This
297 lets you reject messages to certain recipients, rather than waiting un‐
298 til the whole message has been sent. Note that for this check to take
299 place, you must use the -t flag with mimedefang.
300 filter_recipient is passed nine arguments: $recipient is the envelope
303 address of the recipient and $sender is the envelope e-mail address of
304 the sender (for example, "<dfs@roaringpenguin.com>"). The addresses
305 may or may not be surrounded by angle brackets. $ip and $name are the
306 IP address and host name of the SMTP relay. $first is the envelope ad‐
307 dress of the first recipient for this message, and $helo is the argu‐
308 ment to the SMTP "HELO" command. The last three arguments,
309 $rcpt_mailer, $rcpt_host and $rcpt_addr are the Sendmail mailer, host
310 and address triple for the recipient address. For example, for local
311 recipients, $rcpt_mailer is likely to be "local", while for remote re‐
312 cipients, it is likely to be "esmtp".
313 Inside filter_recipient, you can access any ESMTP arguments (such as
316 "NOTIFY=never") in the array @ESMTPArgs. Each ESMTP argument occupies
317 one array element.
318 filter_recipient must return a two-to-five element list whose interpre‐
321 tation is the same as for filter_sender. Note, however, that if fil‐
322 ter_recipient returns 'DISCARD', then the entire message for all recip‐
323 ients is discarded. (It doesn't really make sense, but that's how Mil‐
324 ter works.)
325 For example, if you wish to reject messages from spammer@badguy.com,
328 unless they are to postmaster@mydomain.com, you could use this func‐
329 tion:
330 sub filter_recipient {
332 my ($recipient, $sender, $ip, $hostname, $first, $helo,
333 $rcpt_mailer, $rcpt_host, $rcpt_addr) = @_;
334 if ($sender =~ /^<?spammer\@badguy\.com>?$/i) {
335 if ($recipient =~ /^<?postmaster\@mydomain\.com>?$/i) {
336 return ('CONTINUE', "ok");
337 }
338 return ('REJECT', 'Sorry; spammer@badguy.com is blacklisted.');
339 }
340 return ('CONTINUE', "ok");
341 }
342
345 Just before a worker begins processing messages, mimedefang.pl calls
346 the functions filter_initialize (if it is defined) with no arguments.
347 By the time filter_initialize is called, all the other initialization
348 (such as setting up syslog facility and priority) has been done.
349 If you are not using an embedded Perl interpreter, then performing an
351 action inside filter_initialize is practically the same as performing
352 it directly in the filter file, outside any function definition. How‐
353 ever, if you are using an embedded Perl interpreter, then anything you
354 call directly from outside a function definition is executed once only
355 in the parent process. Anything in filter_initialize is executed once
356 per worker. If you use any code that opens a descriptor (for example,
357 a connection to a database server), you must run that code inside fil‐
358 ter_initialize and not directly from the filter, because the multi‐
359 plexor closes all open descriptors when it activates a new worker.
360 From within filter_initialize a configuration file could be loaded by
361 calling read_config. read_config accepts a configuration file path and
362 it can be used to overwrite global variables. Configuration file for‐
363 mat is pure Perl code.
364 When a worker is about to exit, mimedefang.pl calls the function fil‐
366 ter_cleanup (if it is defined) with no arguments. This function can do
367 whatever cleanup you like, such as closing file descriptors and clean‐
368 ing up long-lived worker resources. The return value from fil‐
369 ter_cleanup becomes the worker's exit status. (You should therefore
370 ensure that filter_cleanup returns an integer suitable for a process
371 exit status.)
372 If filter_cleanup takes longer than 10 seconds to run, the worker is
375 sent a SIGTERM signal. If that doesn't kill it (because you're catch‐
376 ing signals, perhaps), then a further 10 seconds later, the worker is
377 sent a SIGKILL signal.
378
381 If you define a function called filter_create_parser taking no argu‐
382 ments, then mimedefang.pl will call it to create a MIME::Parser object
383 for parsing mail messages.
384 Filter_create_parser is expected to return a MIME::Parser object (or an
386 instance of a class derived from MIME::Parser).
387 You can use filter_create_parser to change the behavior of the
389 MIME::Parser used by mimedefang.pl.
390 If you do not define a filter_create_parser function, then a built-in
392 version equivalent to this is used:
393 sub filter_create_parser () {
395 my $parser = MIME::Parser->new();
396 $parser->extract_nested_messages(1);
397 $parser->extract_uuencode(1);
398 $parser->output_to_core(0);
399 $parser->tmp_to_core(0);
400 return $parser;
401 }
402
404 The man page for mimedefang-protocol(7) lists commands that are passed
405 to workers in server mode (see "SERVER COMMANDS".) You can define a
406 function called filter_unknown_cmd to extend the set of commands your
407 filter can handle.
408 If you define filter_unknown_cmd, it is passed the unknown command as a
410 single argument. It should return a list of values as follows: The
411 first element of the list must be either "ok" or "error:" (with the
412 colon.) The remaining arguments are percent-encoded. All the result‐
413 ing pieces are joined together with a single space between them, and
414 the resulting string passed back as the reply to the multiplexor.
415 For example, the following function will make your filter reply to a
417 "PING" command with "PONG":
418 sub filter_unknown_cmd ($) {
420 my($cmd) = @_;
421 if ($cmd eq "PING") {
422 return("ok", "PONG");
423 }
424 return("error:", "Unknown command");
425 }
426 You can test this filter by typing the following as root:
428 md-mx-ctrl PING
430 The response should be:
432 ok PONG
434 If you extend the set of commands using filter_unknown_cmd, you should
436 make all your commands start with an upper-case letter to avoid clashes
437 with future built-in commands.
438
441 A very common mail setup is to have a MIMEDefang machine act as an SMTP
442 proxy, accepting and scanning mail and then relaying it to the real
443 mail server. Unfortunately, this means that the MIMEDefang machine
444 cannot know if a local address is valid or not, and will forward all
445 mail for the appropriate domains. If a mail comes in for an unknown
446 user, the MIMEDefang machine will be forced to generate a bounce mes‐
447 sage when it tries to relay the mail.
448 It's often desirable to have the MIMEDefang host reply with a "User un‐
451 known" SMTP response directly. While this can be done by copying the
452 list of local users to the MIMEDefang machine, MIMEDefang has a built-
453 in function called md_check_against_smtp_server for querying another
454 relay host:
455 md_check_against_smtp_server($sender, $recip, $helo, $server, $port)
458 This
459 function connects to the SMTP server $server and pretends to
460 send mail from $sender to $recip. The return value is always a
461 two-element array. If the RCPT TO: command succeeds, the return
462 value is ("CONTINUE", "OK"). If the RCPT fails with a permanent
463 failure, the return value is ("REJECT", $msg), where $msg is the
464 message from the SMTP server. Any temporary failures, connec‐
465 tion errors, etc. result in a return value of ("TEMPFAIL",
466 $msg).
467 The optional argument $port specifies the TCP port to connect
469 to. If it is not supplied, then the default SMTP port of 25 is
470 used.
471 If the server offers STARTTLS support, TLS step-up is attempted.
473 If TLS step-up fails, the check will fall-back to using clear
474 text and log the failure
475 Suppose the machine filter.domain.tld is filtering mail destined for
478 the real mail server mail.domain.tld. You could have a filter_recipi‐
479 ent function like this:
480 sub filter_recipient
482 {
483 my($recip, $sender, $ip, $host, $first, $helo,
484 $rcpt_mailer, $rcpt_host, $rcpt_addr) = @_;
485 return md_check_against_smtp_server($sender, $recip,
486 "filter.domain.tld",
487 "mail.domain.tld");
488 }
489 For each RCPT TO: command, MIMEDefang opens an SMTP connection to
491 mail.domain.tld and checks if the command would succeed.
492 Please note that you should only use md_check_against_smtp_server if
495 your mail server responds with a failure code for nonexistent users at
496 the RCPT TO: level. Also, this function may impose too much overhead
497 if you receive a lot of e-mail, and it will generate lots of useless
498 log entries on the real mail server (because of all the RCPT TO:
499 probes.) It may also significantly increase the load on the real mail
500 server.
501
504 The following Perl global variables should be set in mimedefang-filter:
505 $AdminAddress
508 The e-mail address of the MIMEDefang administrator.
509 $DaemonAddress
512 The e-mail address from which MIMEDefang-originated notifica‐
513 tions come.
514 $AddWarningsInline
517 If this variable is set to 0, then all MIMEDefang warnings (such
518 as created by action_quarantine or action_drop_with_warning) are
519 collected together and added in a separate MIME part called
520 WARNING.TXT. If the variable is set to 1, then the warnings are
521 added directly in the first text/plain and text/html parts of
522 the message. If the message does not contain any text/plain or
523 text/html parts, then a WARNING.TXT MIME part is added as be‐
524 fore.
525 $MaxMIMEParts
528 A message containing many MIME parts can cause MIME::Tools to
529 consume large amounts of memory and bring your system to its
530 knees. If you set $MaxMIMEParts to a positive number, then MIME
531 parsing is terminated for messages with more than that many
532 parts, and the message is bounced. In this case, none of your
533 filter functions is called.
534 By default, $MaxMIMEParts is set to -1, meaning there is no
536 limit on the number of parts in a message. Note that in order
537 to use this variable, you must install the Roaring Penguin
538 patched version of MIME::Tools, version 5.411a-RP-Patched-02 or
539 newer.
540 $Stupidity{"NoMultipleInlines"}
543 Set this to 1 if your e-mail is too stupid to display multiple
544 MIME parts in-line. In this case, a nasty hack causes the first
545 part of the original message to appear as an attachment if warn‐
546 ing are issued. Mail clients that are not this stupid are Net‐
547 scape Communicator and Pine. On the other hand, Microsoft Ex‐
548 change and Microsoft Outlook are indeed this stupid. Perhaps
549 users of those clients should switch.
550 The following global variables may optionally be set. If they
552 are not set, sensible defaults are used:
553 $AddApparentlyToForSpamAssassin
556 By default, MIMEDefang tries to pass SpamAssassin a message that
557 looks exactly like one it would receive via procmail. This
558 means adding a Received: header, adding a Message-ID header if
559 necessary, and adding a Return-Path: header. If you set $AddAp‐
560 parentlyToForSpamAssassin to 1, then MIMEDefang also adds an Ap‐
561 parently-To: header with all the envelope recipients before
562 passing the message to SpamAssassin. This lets SpamAssassin de‐
563 tect possibly whitelisted recipient addresses.
564 The default value for $AddApparentlyToForSpamAssassin is 0.
566 $SyslogFacility
569 This specifies the logging facility used by mimedefang.pl. By
570 default, it is set to "mail", but you can set it to other possi‐
571 bilites. See the openlog(3) man page for details. You should
572 name facilities as all-lowercase without the leading "LOG_".
573 That is, use "local3", not "LOG_LOCAL3".
574 $WarningLocation (default 0)
577 If set to 0 (the default), non-inline warnings are placed first.
578 If you want the warning at the end of the e-mail, set $Warn‐
579 ingLocation to -1.
580 $DaemonName (default "MIMEDefang")
583 The full name used when MIMEDefang sends out notifications.
584 $AdminName (default "MIMEDefang Administrator")
587 The full name of the MIMEDefang administrator.
588 $SALocalTestsOnly (default 1)
591 If set to 1, SpamAssassin calls will use only local tests. This
592 is the default and recommended setting. This disables Received,
593 RBL and Razor tests in an all or nothing fashion. To use Razor
594 this MUST be set to 0. You can add 'skip_rbl_checks 1' to your
595 SpamAssassin config file if you need to.
596 $NotifySenderSubject (default "MIMEDefang Notification")
599 The subject used when e-mail is sent out by action_no‐
600 tify_sender(). If you set this, you should set it each time you
601 call action_notify_sender() to ensure consistency.
602 $NotifyAdministratorSubject (default "MIMEDefang Notification")
604 The subject used when e-mail is sent out by action_notify_admin‐
605 istrator(). If you set this, you should set it each time you
606 call action_notify_administrator() to ensure consistency.
607 $QuarantineSubject (default "MIMEDefang Quarantine Report")
610 The subject used when a quarantine notice is sent to the admin‐
611 istrator. If you set this, you should set it each time you call
612 action_quarantine() or action_quarantine_entire_message().
613 $NotifyNoPreamble (default 0)
616 Normally, notifications sent by action_notify_sender() have a
617 preamble warning about message modifications. If you do not
618 want this, set $NotifyNoPreamble to 1.
619 $CSSHost (default 127.0.0.1:7777:local)
622 Host and port for the Symantec CarrierScan Server virus scanner.
623 This takes the form ip_addr:port:local_or_nonlocal. The ip_addr
624 and port are the host and port on which CarrierScan Server is
625 listening. If you want to scan local files, append :local to
626 force the use of the AVSCANLOCAL command. If the CarrierScan
627 Server is on another host, append :nonlocal to force the file
628 contents to be sent to the scanner over the socket.
629 $SophieSock (default /var/spool/MIMEDefang/sophie)
632 Socket used for Sophie daemon calls within message_con‐
633 tains_virus_sophie and entity_contains_virus_sophie unless a
634 socket is provided by the calling routine.
635 $ClamdSock (default /var/spool/MIMEDefang/clamd.sock)
638 Socket used for clamd daemon calls within message_con‐
639 tains_virus_clamd and entity_contains_virus_clamd unless a
640 socket is provided by the calling routine.
641 $TrophieSock (default /var/spool/MIMEDefang/trophie)
644 Socket used for Trophie daemon calls within message_con‐
645 tains_virus_trophie and entity_contains_virus_trophie unless a
646 socket is provided by the calling routine.
647
651 The heart of mimedefang-filter is the filter procedure. See the exam‐
652 ples that came with MIMEDefang to learn to write a filter. The filter
653 is called with the following arguments:
654 $entity
657 The MIME::Entity object. (See the MIME::tools Perl module docu‐
658 mentation.)
659 $fname The suggested attachment filename, or "" if none was supplied.
662 $ext The file extension (all characters from the rightmost period to
665 the end of the filename.)
666 $type The MIME type (for example, "text/plain".)
669 The filename is derived as follows:
672 o First, if the Content-Disposition header has a "filename" field,
675 it is used.
676 o Otherwise, if the Content-Type header has a "name" field, it is
679 used.
680 o Otherwise, the Content-Description header value is used.
683 Note that the truly paranoid will check all three fields for matches.
686 The functions re_match and re_match_ext perform regular expression
687 matches on all three of the fields named above, and return 1 if any
688 field matches. See the sample filters for details. The calling se‐
689 quence is:
690 re_match($entity, "regexp")
692 re_match_ext($entity, "regexp")
693 re_match returns true if any of the fields matches the regexp without
695 regard to case. re_match_ext returns true if the extension in any
696 field matches. An extension is defined as the last dot in a name and
697 all remaining characters.
698 A third function called re_match_in_zip_directory will look inside zip
701 files and return true if any of the file names inside the zip archive
702 match the regular expression. Call it like this:
703 my $bh = $entity->bodyhandle();
705 my $path = (defined($bh)) ? $bh->path() : undef;
706 if (defined($path) and re_match_in_zip_directory($path, "regexp")) {
707 # Take action...
708 }
709 You should not call re_match_in_zip_directory unless you know that the
711 entity is a zip file attachment.
712 Another function called re_match_in_rar_directory will look inside rar
715 files and return true if any of the file names inside the rar archive
716 match the regular expression. The function is very similar to
717 re_match_in_zip_directory but the unrar binary is required and must be
718 specified in $Features{"unrar"}.
719 Another function called re_match_in_7z_directory will look inside 7zip
722 files and return true if any of the file names inside the 7zip archive
723 match the regular expression. The function is very similar to
724 re_match_in_zip_directory but the 7z binary is required and must be
725 specified in $Features{"7zip"}.
726
729 The following global variables are set by mimedefang.pl and are avail‐
730 able for use in your filter. All of these variables are always avail‐
731 able to filter_begin, filter, filter_multipart and filter_end. In ad‐
732 dition, some of them are available in filter_relay, filter_sender or
733 filter_recipient. If this is the case, it will be noted below.
734 %Features
737 This hash lets you determine at run-time whether certain func‐
738 tionality is available. This hash is available at all times as‐
739 suming the detect_and_load_perl_modules() function has been
740 called. The defined features are:
741 $Features{"SpamAssassin"} is 1 if SpamAssassin 1.6 or better is
743 installed; 0 otherwise.
744 $Features{"HTML::Parser"} is 1 if HTML::Parser is installed; 0
746 otherwise.
747 $Features{"Virus:FPROTD"} is currently always 0. Set it to 1 in
749 your filter file if you have F-Risk's FPROTD scanner earlier
750 than version 6.
751 $Features{"Virus:FPROTD6"} is currently always 0. Set it to 1
753 in your filter file if you have version 6 of F-Risk's FPROTD
754 scanner.
755 $Features{"Virus:SymantecCSS"} is currently always 0. Set it to
757 1 in your filter file if you have the Symantec CarrierScan
758 Server virus scanner.
759 $Features{"Virus:NAI"} is the full path to NAI uvscan if it is
761 installed; 0 if it is not.
762 $Features{"Virus:BDC"} is the full path to Bitdefender bdc if it
764 is installed; 0 if it is not.
765 $Features{"Virus:NVCC"} is the full path to Norman Virus Control
767 nvcc if it is installed; 0 if it is not.
768 $Features{"Virus:HBEDV"} is the full path to H+BEDV AntiVir if
770 it is installed; 0 if it is not.
771 $Features{"Virus:VEXIRA"} is the full path to Central Command
773 Vexira if it is installed; 0 if it is not.
774 $Features{"Virus:SOPHOS"} is the full path to Sophos sweep if it
776 is installed; 0 if it is not.
777 $Features{"Virus:SAVSCAN"} is the full path to Sophos savscan if
779 it is installed; 0 if it is not.
780 $Features{"Virus:CLAMAV"} is the full path to Clam AV clamscan
782 if it is installed; 0 if it is not.
783 $Features{"Virus:AVP"} is the full path to AVP AvpLinux if it is
785 installed; 0 if it is not.
786 $Features{"Virus:AVP5"} is the full path to Kaspersky "ave‐
788 client" if it is installed; 0 if it is not.
789 $Features{"Virus:CSAV"} is the full path to Command csav if it
791 is installed; 0 if it is not.
792 $Features{"Virus:FSAV"} is the full path to F-Secure fsav if it
794 is installed; 0 if it is not.
795 $Features{"Virus:FPROT"} is the full path to F-Risk f-prot if it
797 is installed; 0 if it is not.
798 $Features{"Virus:FPSCAN"} is the full path to F-Risk fpscan if
800 it is installed; 0 if it is not.
801 $Features{"Virus:SOPHIE"} is the full path to Sophie if it is
803 installed; 0 if it is not.
804 $Features{"Virus:CLAMD"} is the full path to clamd if it is in‐
806 stalled; 0 if it is not.
807 $Features{"Virus:CLAMDSCAN"} is the full path to clamdscan if it
809 is installed; 0 if it is not.
810 $Features{"Virus:TROPHIE"} is the full path to Trophie if it is
812 installed; 0 if it is not.
813 $Features{"Virus:NOD32"} is the full path to ESET NOD32 nod32cli
815 if it is installed; 0 if it is not.
816 $Features{"Path:RSPAMC"} is the full path to rspamc(1) if it is
818 installed (deprecated); 0 if it is not.
819 NOTE: Perl-module based features such as SpamAssassin are deter‐
821 mined at runtime and may change as these are added and removed.
822 Most Virus features are predetermined at the time of configura‐
823 tion and do not adapt to runtime availability unless changed by
824 the filter rules.
825 $CWD This variable holds the working directory for the current mes‐
828 sage. During filter processing, mimedefang.pl chdir's into this
829 directory before calling any of the filter_ functions. Note
830 that this variable is set correctly in filter_sender and fil‐
831 ter_recipient, but not in filter_relay.
832 $SuspiciousCharsInHeaders
835 If this variable is true, then mimedefang has discovered suspi‐
836 cious characters in message headers. This might be an exploit
837 for bugs in MIME-parsing routines in some badly-written mail
838 user agents (e.g. Microsoft Outlook.) You should always drop
839 such messages.
840 $SuspiciousCharsInBody
843 If this variable is true, then mimedefang has discovered suspi‐
844 cious characters in the message body. This might be an exploit
845 for bugs in MIME-parsing routines in some badly-written mail
846 user agents (e.g. Microsoft Outlook.) You should always drop
847 such messages.
848 $RelayHostname
851 The host name of the relay. This is the name of the host that
852 is attempting to send e-mail to your host. May be "undef" if
853 the host name could not be determined. This variable is avail‐
854 able in filter_relay, filter_sender and filter_recipient in ad‐
855 dition to the body filtering functions.
856 $RelayAddr
859 The IP address of the sending relay (as a string consisting of
860 four dot-separated decimal numbers.) One potential use of $Re‐
861 layAddr is to limit mailing to certain lists to people within
862 your organization. This variable is available in filter_relay,
863 filter_sender and filter_recipient in addition to the body fil‐
864 tering functions.
865 $Helo The argument given to the SMTP "HELO" command. This vari‐
867 able is available in filter_sender and filter_recipient, but not
868 in filter_relay.
869 $Subject
872 The contents of the "Subject:" header.
873 $Sender
876 The sender of the e-mail. This variable is set in filter_sender
877 and filter_recipient in addition to the body filtering func‐
878 tions.
879 @Recipients
882 A list of the recipients. In filter_recipient, it is set to the
883 single recipient currently under consideration. Or, after call‐
884 ing read_commands_file within filter_recipient, the current re‐
885 cipient under consideration is in the final position of the ar‐
886 ray, at $Recipients[-1], while any previous (and accepted) re‐
887 cipients are at the beginning of the array, that is, in @Recipi‐
888 ents[0 .. $#Recipients-1].
889 $MessageID
893 The contents of the "Message-ID:" header if one is present.
894 Otherwise, contains the string "NOQUEUE".
895 $QueueID
898 The Sendmail queue identifier if it could be determined. This
899 variable is set correctly in filter_relay, filter_helo, fil‐
900 ter_sender and filter_recipient. Note, however, that Postfix
901 may not allocate a queue ID until filter_recipient time. If a
902 Queue-ID has not yet been allocated, $QueueID is set to "NO‐
903 QUEUE".
904 $MsgID Set to $QueueID if the queue ID could be determined; otherwise,
907 set to $MessageID. This identifier should be used in logging,
908 because it matches the identifier used by Sendmail to log mes‐
909 sages. Note that this variable is set correctly in fil‐
910 ter_sender and filter_recipient, but it is not available in fil‐
911 ter_relay.
912 $VirusScannerMessages
915 Each time a virus-scanning function is called, messages (if any)
916 from the virus scanner are accumulated in this variable. You
917 can use it in filter_end to formulate a notification (if you
918 wish.)
919 $VirusName
922 If a virus-scanning function found a virus, this variable will
923 hold the virus name (if it could be determined.)
924 $SASpamTester
927 If defined, this is the configured Mail::SpamAssassin object
928 used for mail tests. It may be initialized with a call to
929 spam_assassin_init which also returns it.
930 %SendmailMacros
933 This hash contains the values of some Sendmail macros. The hash
934 elements exist only for macros defined by Sendmail. See the
935 Sendmail documentation for the meanings of the macros.
936 By default, mimedefang passes the values of the following
938 macros: ${daemon_name}, ${daemon_port}, ${if_name}, ${if_addr},
939 $j, $_, $i, ${tls_version}, ${cipher}, ${cipher_bits},
940 ${cert_subject}, ${cert_issuer}, ${auth_type}, ${auth_authen},
941 ${auth_ssf}, ${auth_author}, ${mail_mailer}, ${mail_host} and
942 ${mail_addr}. In addition, ${client_port} is set to the
943 client's TCP port.
944 If any macro is not set or not passed to milter, it will be un‐
946 available. To access the value of a macro, use:
947 $SendmailMacros{"macro_name"}
950 Do not place curly brackets around the macro name. This vari‐
953 able is available in filter_sender and filter_recipient after a
954 call to read_commands_file.
955 @SenderESMTPArgs
958 This array contains all the ESMTP arguments supplied in the MAIL
959 FROM: command. For example:
960 sub print_sender_esmtp_args {
962 foreach (@SenderESMTPArgs) {
963 print STDERR "Sender ESMTP arg: $_0;
964 }
965 }
966 %RecipientESMTPArgs
969 This hash contains all the ESMTP arguments supplied in each RCPT
970 TO: command. For example:
971 sub print_recip_esmtp_args {
973 foreach my $recip (@Recipients) {
974 foreach(@{$RecipientESMTPArgs{$recip}}) {
975 print STDERR "Recip ESMTP arg for $recip: $_0;
976 }
977 }
978 }
979 %RecipientMailers
982 This hash contains the Sendmail "mailer-host-address" triple for
983 each recipient. Here's an example of how to use it:
984 sub print_mailer_info {
986 my($recip, $mailer, $host, $addr);
987 foreach $recip (@Recipients) {
988 $mailer = ${RecipientMailers{$recip}}[0];
989 $host = ${RecipientMailers{$recip}}[1];
990 $addr = ${RecipientMailers{$recip}}[2];
991 print STDERR "$recip: mailer=$mailer, host=$host, addr=$addr\n";
992 }
993 }
994 In filter_recipient, this variable by default only contains in‐
996 formation on the recipient currently under investigation. Infor‐
997 mation on all recipients is available after calling read_com‐
998 mands_file.
999
1002 When the filter procedure decides how to dispose of a part, it should
1003 call one or more action_ subroutines. The action subroutines are:
1004 action_accept()
1007 Accept the part.
1008 action_rebuild()
1011 Rebuild the mail body, even if mimedefang thinks no changes were
1012 made. Normally, mimedefang does not alter a message if no
1013 changes were made. action_rebuild may be used if you make
1014 changes to entities directly (by manipulating the MIME::Head,
1015 for example.) Unless you call action_rebuild, mimedefang will
1016 be unaware of the changes. Note that all the built-in action...
1017 routines that change a message implicitly call action_rebuild.
1018 action_add_header($hdr, $val)
1021 Add a header to the message. This can be used in filter_begin
1022 or filter_end. The $hdr component is the header name without
1023 the colon, and the $val is the header value. For example, to
1024 add the header:
1025 X-MyHeader: A nice piece of text
1027 use:
1029 action_add_header("X-MyHeader", "A nice piece of text");
1031 action_change_header($hdr, $val, $index)
1034 Changes an existing header in the message. This can be used in
1035 filter_begin or filter_end. The $hdr parameter is the header
1036 name without the colon, and $val is the header value. If the
1037 header does not exist, then a header with the given name and
1038 value is added.
1039 The $index parameter is optional; it defaults to 1. If you sup‐
1041 ply it, then the $index'th occurrence of the header is changed,
1042 if there is more than one header with the same name. (This is
1043 common with the Received: header, for example.)
1044 action_insert_header($hdr, $val, $index)
1047 Add a header to the message int the specified position $index.
1048 A position of 0 specifies that the header should be prepended
1049 before existing headers. This can be used in filter_begin or
1050 filter_end. The $hdr component is the header name without the
1051 colon, and the $val is the header value.
1052 action_delete_header($hdr, $index)
1055 Deletes an existing header in the message. This can be used in
1056 filter_begin or filter_end. The $hdr parameter is the header
1057 name without the colon.
1058 The $index parameter is optional; it defaults to 1. If you sup‐
1060 ply it, then the $index'th occurrence of the header is deleted,
1061 if there is more than one header with the same name.
1062 action_delete_all_headers($hdr)
1065 Deletes all headers with the specified name. This can be used
1066 in filter_begin or filter_end. The $hdr parameter is the header
1067 name without the colon.
1068 action_drop()
1071 Drop the part. If called from filter_multipart, drops all con‐
1072 tained parts also.
1073 action_drop_with_warning($msg)
1076 Drop the part, but add the warning $msg to the e-mail message.
1077 If called from filter_multipart, drops all contained parts also.
1078 action_accept_with_warning($msg)
1081 Accept the part, but add the warning $msg to the e-mail message.
1082 action_replace_with_warning($msg)
1085 Drop the part and replace it with a text part $msg. If called
1086 from filter_multipart, drops all contained parts also.
1087 action_replace_with_url($entity, $doc_root, $base_url, $msg, [$cd_data,
1090 $salt])
1091 Drop the part, but save it in a unique location under $doc_root.
1092 The part is replaced with the text message $msg. The string
1093 "_URL_" in $msg is replaced with $base_url/something, that can
1094 be used to retrieve the message.
1095 You should not use this function in filter_multipart.
1097 This action is intended for stripping large parts out of the
1099 message and replacing them to a link on a Web server. Here's
1100 how you would use it in filter():
1101 $size = (stat($entity->bodyhandle->path))[7];
1103 if ($size > 1000000) {
1104 return action_replace_with_url($entity,
1105 "/home/httpd/html/mail_parts",
1106 "http://mailserver.company.com/mail_parts",
1107 "The attachment was larger than 1,000,000 bytes.\n" .
1108 "It was removed, but may be accessed at this URL:\n\n" .
1109 "\t_URL_\n");
1110 }
1111 This example moves attachments greater than 1,000,000 bytes into
1113 /home/httpd/html/mail_parts and replaces them with a link. The
1114 directory should be accessible via a Web server at
1115 http://mailserver.company.com/mail_parts.
1116 The generated name is created by performing a SHA1 hash of the
1118 part and adding the extension to the ASCII-HEX representation of
1119 the hash. If many different e-mails are sent containing an
1120 identical large part, only one copy of the part is stored, re‐
1121 gardless of the number of senders or recipients.
1122 For privacy reasons, you must turn off Web server indexing in
1124 the directory in which you place mail parts, or anyone will be
1125 able to read them. If indexing is disabled, an attacker would
1126 have to guess the SHA1 hash of a part in order to read it.
1127 Optionally, a fifth argument can supply data to be saved into a
1129 hidden dot filename based on the generated name. This data can
1130 then be read in on the fly by a CGI script or mod_perl module
1131 before serving the file to a web client, and used to add infor‐
1132 mation to the response, such as Content-Disposition data.
1133 A sixth optional argument, $salt, is mixed in to the SHA1 hash.
1135 This salt can be any string and should be kept confidential.
1136 The salt is designed to prevent people from guessing whether or
1137 not a particular attachment has been received on your server by
1138 altering the SHA1 hash calculation.
1139 action_defang($entity, $name, $fname, $type)
1142 Accept the part, but change its name to $name, its suggested
1143 filename to $fname and its MIME type to $type. If $name or
1144 $fname are "", then mimedefang.pl generates generic names. Do
1145 not use this action in filter_multipart.
1146 If you use action_defang, you must define a subroutine called
1148 defang_warning in your filter. This routine takes two argu‐
1149 ments: $oldfname (the original name of an attachment) and $fname
1150 (the defanged version.) It should return a message telling the
1151 user what happened. For example:
1152 sub defang_warning {
1154 my($oldfname, $fname) = @_;
1155 return "The attachment '$oldfname' was renamed to '$fname'\n";
1156 }
1157 action_external_filter($entity, $cmd)
1161 Run an external UNIX command $cmd. This command must read the
1162 part from the file ./FILTERINPUT and leave the result in ./FIL‐
1163 TEROUTPUT. If the command executes successfully, returns 1,
1164 otherwise 0. You can test the return value and call another ac‐
1165 tion_ if the filter failed. Do not use this action in fil‐
1166 ter_multipart.
1167 action_quarantine($entity, $msg)
1170 Drop and quarantine the part, but add the warning $msg to the e-
1171 mail message.
1172 action_quarantine_entire_message($msg)
1175 Quarantines the entire message in a quarantine directory on the
1176 mail server, but does not otherwise affect disposition of the
1177 message. If "$msg" is non-empty, it is included in any adminis‐
1178 trator notification.
1179 action_sm_quarantine($reason)
1182 Quarantines a message in the Sendmail mail queue using the new
1183 QUARANTINE facility of Sendmail 8.13. Consult the Sendmail doc‐
1184 umentation for details about this facility. If you use ac‐
1185 tion_sm_quarantine with a version of Sendmail that lacks the
1186 QUARANTINE facility, mimedefang will log an error message and
1187 not quarantine the message.
1188 action_bounce($reply, $code, $dsn)
1191 Reject the entire e-mail message with an SMTP failure code, and
1192 the one-line error message $reply. If the optional $code and
1193 $dsn arguments are supplied, they specify the numerical SMTP re‐
1194 ply code and the extended status code (DSN code). If the codes
1195 you supply do not make sense for a bounce, they are replaced
1196 with "554" and "5.7.1" respectively.
1197 action_bounce merely makes a note that the message is to be
1199 bounced; remaining parts are still processed. If action_bounce
1200 is called for more than one part, the mail is bounced with the
1201 message in the final call to action_bounce. You can profitably
1202 call action_quarantine followed by action_bounce if you want to
1203 keep a copy of the offending part. Note that the message is not
1204 bounced immediately; rather, remaining parts are processed and
1205 the message is bounced after all parts have been processed.
1206 Note that despite its name, action_bounce does not generate a
1208 "bounce message". It merely rejects the message with an SMTP
1209 failure code.
1210 WARNING: action_bounce() may cause the sending relay to generate
1212 spurious bounce messages if the sender address is faked. This
1213 is a particular problem with viruses. However, we believe that
1214 on balance, it's better to bounce a virus than to silently dis‐
1215 card it. It's almost never a good idea to hide a problem.
1216 action_tempfail($msg, $code, $dsn)
1219 Cause an SMTP "temporary failure" code to be returned, so the
1220 sending mail relay requeues the message and tries again later.
1221 The message $msg is included with the temporary failure code.
1222 If the optional $code and $dsn arguments are supplied, they
1223 specify the numerical SMTP reply code and the extended status
1224 code (DSN code). If the codes you supply do not make sense for
1225 a temporary failure, they are replaced with "450" and "4.7.1"
1226 respectively.
1227 action_discard()
1230 Silently discard the message, notifying nobody. You can prof‐
1231 itably call action_quarantine followed by action_discard if you
1232 want to keep a copy of the offending part. Note that the mes‐
1233 sage is not discarded immediately; rather, remaining parts are
1234 processed and the message is discarded after all parts have been
1235 processed.
1236 action_notify_sender($message)
1239 This action sends an e-mail back to the original sender with the
1240 indicated message. You may call another action after this one.
1241 If action_notify_sender is called more than once, the messages
1242 are accumulated into a single e-mail message -- at most one no‐
1243 tification message is sent per incoming message. The message
1244 should be terminated with a newline.
1245 The notification is delivered in deferred mode; you should run a
1247 client-queue runner if you are using Sendmail 8.12.
1248 NOTE: Viruses often fake the sender address. For that reason,
1250 if a virus-scanner has detected a virus, action_notify_sender is
1251 disabled and will simply log an error message if you try to use
1252 it.
1253 action_notify_administrator($message)
1256 This action e-mails the MIMEDefang administrator the supplied
1257 message. You may call another action after this one; action_no‐
1258 tify_administrator does not affect mail processing. If ac‐
1259 tion_notify_administrator is called more than once, the messages
1260 are accumulated into a single e-mail message -- at most one no‐
1261 tification message is sent per incoming message. The message
1262 should be terminated with a newline.
1263 The notification is delivered in deferred mode; you should run a
1265 client-queue runner if you are using Sendmail 8.12.
1266 append_text_boilerplate($entity, $boilerplate, $all)
1269 This action should only be called from filter_end. It appends
1270 the text "\n$boilerplate\n" to the first text/plain part (if
1271 $all is 0) or to all text/plain parts (if $all is 1).
1272 append_html_boilerplate($entity, $boilerplate, $all)
1275 This action should only be called from filter_end. It adds the
1276 text "\n$boilerplate\n" to the first text/html part (if $all is
1277 0) or to all text/html parts (if $all is 1). This function
1278 tries to be smart about inserting the boilerplate; it uses
1279 HTML::Parser to detect closing tags and inserts the boilerplate
1280 before the </body> tag if there is one, or before the </html>
1281 tag if there is no </body>. If there is no </body> or </html>
1282 tag, it appends the boilerplate to the end of the part.
1283 Do not use append_html_boilerplate unless you have installed the
1285 HTML::Parser Perl module.
1286 Here is an example illustrating how to use the boilerplate func‐
1288 tions:
1289 sub filter_end {
1291 my($entity) = @_;
1292 append_text_boilerplate($entity,
1293 "Lame text disclaimer", 0);
1294 append_html_boilerplate($entity,
1295 "<em>Lame</em> HTML disclaimer", 0);
1296 }
1297 action_add_part($entity, $type, $encoding, $data, $fname, $disposition
1300 [, $offset])
1301 This action should only be called from the filter_end routine.
1302 It adds a new part to the message, converting the original mes‐
1303 sage to mutipart if necessary. The function returns the part so
1304 that additional mime attributes may be set on it. Here's an ex‐
1305 ample:
1306 sub filter_end {
1308 my($entity) = @_;
1309 action_add_part($entity, "text/plain", "-suggest",
1311 "This e-mail does not represent" .
1312 "the official policy of FuBar, Inc.\n",
1313 "disclaimer.txt", "inline");
1314 }
1315 The $entity parameter must be the argument passed in to fil‐
1317 ter_end. The $offset parameter is optional; if omitted, it de‐
1318 faults to -1, which adds the new part at the end. See the
1319 MIME::Entity man page and the add_part member function for the
1320 meaning of $offset.
1321 Note that action_add_part tries to be more intelligent than sim‐
1323 ply calling $entity->add_part. The decision process is as fol‐
1324 lows:
1325 o If the top-level entity is multipart/mixed, then the part is
1328 simply added.
1329 o Otherwise, a new top-level multipart/mixed container is gener‐
1332 ated, and the original top-level entity is made the first part
1333 of the multipart/mixed container. The new part is then added to
1334 the multipart/mixed container.
1335 action_add_entity($entity [, $offset])
1338 This is similar to action_add_part but takes a pre-built
1339 MIME::Entity object rather than constructing one based on $type,
1340 $encoding, $data, $fname and $disposition arguments.
1341
1344 mimedefang.pl includes some useful functions you can call from your
1345 filter:
1346 detect_and_load_perl_modules()
1349 Unless you really know what you're doing, this function must be
1350 called first thing in your filter file. It causes mimedefang.pl
1351 to detect and load Perl modules such as Mail::SpamAssassin,
1352 Net::DNS, etc., and to populate the %Features hash.
1353 send_quarantine_notifications()
1356 This function should be called from filter_end. If any parts
1357 were quarantined, a quarantine notification is sent to the
1358 MIMEDefang administrator. Please note that if you do not call
1359 send_quarantine_notifications, then no quarantine notifications
1360 are sent.
1361 get_quarantine_dir()
1364 This function returns the full path name of the quarantine di‐
1365 rectory. If you have not yet quarantined any parts of the mes‐
1366 sage, a quarantine directory is created and its pathname re‐
1367 turned.
1368 change_sender($sender)
1371 This function changes the envelope sender to $sender. It can
1372 only be called from filter_begin or any later function. Please
1373 note that this function is only supported with Sendmail/Milter
1374 8.14.0 or newer. It has no effect if you're running older ver‐
1375 sions.
1376 add_recipient($recip)
1379 This function adds $recip to the list of envelope recipients. A
1380 copy of the message (after any modifications by MIMEDefang) will
1381 be sent to $recip in addition to the original recipients. Note
1382 that add_recipient does not modify the @Recipients array; it
1383 just makes a note to Sendmail to add the recipient.
1384 delete_recipient($recip)
1387 This function deletes $recip from the list of recipients. That
1388 person will not receive a copy of the mail. $recip should ex‐
1389 actly match an entry in the @Recipients array for delete_recipi‐
1390 ent() to work. Note that delete_recipient does not modify the
1391 @Recipients array; it just makes a note to Sendmail to delete
1392 the recipient.
1393 resend_message($recip1, $recip2, ...)
1395 or
1396 resend_message(@recips)
1398 This function immediately resends the original, unmodified mail
1399 message to each of the named recipients. The sender's address
1400 is preserved. Be very careful when using this function, because
1401 it resends the original message, which may contain undesired at‐
1402 tachments. Also, you should not call this function from fil‐
1403 ter(), because it resends the message each time it is called.
1404 This may result in multiple copies being sent if you are not
1405 careful. Call from filter_begin() or filter_end() to be safe.
1406 The function returns true on success, or false if it fails.
1408 Note that the resend_message function delivers the mail in de‐
1410 ferred mode (using Sendmail's "-odd" flag.) You must run a
1411 client-submission queue processor if you use Sendmail 8.12. We
1412 recommend executing this command as part of the Sendmail startup
1413 sequence:
1414 sendmail -Ac -q5m
1416 remove_redundant_html_parts($entity)
1419 This function should only be called from filter_end. It removes
1420 redundant HTML parts from the message. It works by deleting any
1421 part of type text/html from the message if (1) it is a sub-part
1422 of a multipart/alternative part, and (2) there is another part
1423 of type text/plain under the multipart/alternative part.
1424 replace_entire_message($entity)
1427 This function can only be called from filter_end. It replaces
1428 the entire message with $entity, a MIME::Entity object that you
1429 have constructed. You can use any of the MIME::Tools functions
1430 to construct the entity.
1431 read_commands_file()
1434 This function should only be called from filter_sender and fil‐
1435 ter_recipient. This will read the COMMANDS file (as described in
1436 mimedefang-protocol(7)), and will fill or update the following
1437 global variables: $Sender, @Recipients, %RecipientMailers, $Re‐
1438 layAddr, $RealRelayAddr, $RelayHostname, $RealRelayHostname,
1439 $QueueID, $Helo, %SendmailMacros.
1440 If you do not call read_commands_file, then the only information
1442 available in filter_sender and filter_recipient is that which is
1443 passed as an argument to the function.
1444 stream_by_domain()
1447 Do not use this function unless you have Sendmail 8.12 and lo‐
1448 cally- submitted e-mail is submitted using SMTP.
1449 This function should only be called at the very beginning of
1451 filter_begin(), like this:
1452 sub filter_begin {
1454 if (stream_by_domain()) {
1455 return;
1456 }
1457 # Rest of filter_begin
1458 }
1459 stream_by_domain() looks at all the recipients of the message,
1461 and if they belong to the same domain (e.g., joe@domain.com,
1462 jane@domain.com and sue@domain.com), it returns 0 and sets the
1463 global variable $Domain to the domain (domain.com in this exam‐
1464 ple.)
1465 If users are in different domains, stream_by_domain() resends
1467 the message (once to each domain) and returns 1 For example, if
1468 the original recipients are joe@abc.net, jane@xyz.net and
1469 sue@abc.net, the original message is resent twice: One copy to
1470 joe@abc.net and sue@abc.net, and another copy to jane@xyz.net.
1471 Also, any subsequent scanning is canceled (filter() and fil‐
1472 ter_end() will not be called for the original message) and the
1473 message is silently discarded.
1474 If you have Sendmail 8.12, then locally-submitted messages are
1476 sent via SMTP, and MIMEDefang will be called for each resent
1477 message. It is possible to set up Sendmail 8.12 so locally-sub‐
1478 mitted messages are delivered directly; in this case,
1479 stream_by_domain will not work.
1480 Using stream_by_domain allows you to customize your filter rules
1482 for each domain. If you use the function as described above,
1483 you can do this in your filter routine:
1484 sub filter {
1486 my($entity, $fname, $ext, $type) = @_;
1487 if ($Domain eq "abc.com") {
1488 # Filter actions for abc.com
1489 } elsif ($Domain eq "xyz.com") {
1490 # Filter actions for xyz.com
1491 } else {
1492 # Default filter actions
1493 }
1494 }
1495 You cannot rely on $Domain being set unless you have called
1497 stream_by_domain().
1498 stream_by_recipient()
1501 Do not use this function unless you have Sendmail 8.12 and lo‐
1502 cally- submitted e-mail is submitted using SMTP.
1503 This function should only be called at the very beginning of
1505 filter_begin(), like this:
1506 sub filter_begin {
1508 if (stream_by_recipient()) {
1509 return;
1510 }
1511 # Rest of filter_begin
1512 }
1513 If there is more than one recipient, stream_by_recipient() re‐
1515 sends the message once to each recipient. That way, you can
1516 customize your filter rules on a per-recipient basis. This may
1517 increase the load on your mail server considerably.
1518 Also, a "recipient" is determined before alias expansion. So
1520 "all@mydomain.com" is considered a single recipient, even if
1521 Sendmail delivers to a list.
1522 If you have Sendmail 8.12, then locally-submitted messages are
1524 sent via SMTP, and MIMEDefang will be called for each resent
1525 message. It is possible to set up Sendmail 8.12 so locally-sub‐
1526 mitted messages are delivered directly; in this case,
1527 stream_by_recipient() will not work.
1528 stream_by_recipient() allows you to customize your filter rules
1530 for each recipient in a manner similar to stream_by_domain().
1531
1534 md_graphdefang_log_enable($facility, $enum_recips)
1535 Enables the md_graphdefang_log function (described next). The
1536 function logs to syslog using the specified facility. If you
1537 omit $facility, it defaults to 'mail'. If you do not call
1538 md_graphdefang_log_enable in your filter, then any calls to
1539 md_graphdefang_log simply do nothing.
1540 If you supply $enum_recips as 1, then a line of logging is out‐
1542 put for each recipient of a mail message. If it is zero, then
1543 only a single line is output for each message. If you omit
1544 $enum_recips, it defaults to 1.
1545 md_graphdefang_log($event, $v1, $v2)
1548 Logs an event with up to two optional additional parameters.
1549 The log message has a specific format useful for graphing tools;
1550 the message looks like this:
1551 MDLOG,msgid,event,v1,v2,sender,recipient,subj
1553 "MDLOG" is literal text. "msgid" is the Sendmail queue identi‐
1555 fier. "event" is the event name, and "v1" and "v2" are the ad‐
1556 ditional parameters. "sender" is the sender's e-mail address.
1557 "recipient" is the recipient's e-mail address, and "subj" is the
1558 message subject. If a message has more than one recipient,
1559 md_graphdefang_log may log an event message for each recipient,
1560 depending on how you called md_graphdefang_log_enable.
1561 Note that md_graphdefang_log should not be used in filter_relay,
1563 filter_sender or filter_recipient. The global variables it re‐
1564 lies on are not valid in that context.
1565 If you want to log general text strings, do not use md_graphde‐
1567 fang_log. Instead, use md_syslog (described next).
1568 md_syslog($level, $msg)
1571 Logs the message $msg to syslog, using level $level. The level
1572 is a literal string, and should be one of 'err', 'debug', 'warn‐
1573 ing', ´emerg', 'crit', 'notice' or 'info'. (See syslog(3) for
1574 details.)
1575 Note that md_syslog does not perform %-subsitutions like sys‐
1577 log(3) does. Depending on your Perl installation, md_syslog
1578 boils down to a call to Unix::Syslog::syslog or Sys::Sys‐
1579 log::syslog. See the Unix::Syslog or Sys::Syslog man pages for
1580 more details.
1581 md_openlog($tag, $facility)
1584 Sets the tag used in syslog messages to $tag, and sends the logs
1585 to the $facility facility. If you do not call md_openlog before
1586 you call md_syslog, then it is called implicitly with $tag set
1587 to mimedefang.pl and $facility set to mail.
1588
1591 mimedefang.pl includes the following functions for looking up IP ad‐
1592 dresses in DNS-based real-time blacklists. Note that the "re‐
1593 lay_is_blacklisted" functions are deprecated and may be removed in a
1594 future release. Instead, you should use the module Net::DNSBL::Client
1595 from CPAN.
1596 relay_is_blacklisted($relay, $domain)
1599 This checks a DNS-based real-time spam blacklist, and returns
1600 true if the relay host is blacklisted, or false otherwise. (In
1601 fact, the return value is whatever the blacklist returns as a
1602 resolved hostname, such as "127.0.0.4")
1603 Note that relay_is_blacklisted uses the built-in gethostbyname
1605 function; this is usually quite inefficient and does not permit
1606 you to set a timeout on the lookup. Instead, we recommend using
1607 one of the other DNS lookup function described in this section.
1608 (Note, though, that the other functions require the Perl
1609 Net::DNS module, whereas relay_is_blacklisted does not.)
1610 Here's an example of how to use relay_is_blacklisted:
1612 if (relay_is_blacklisted($RelayAddr, "rbl.spamhaus.org")) {
1614 action_add_header("X-Blacklist-Warning",
1615 "Relay $RelayAddr is blacklisted by Spamhaus");
1616 }
1617 relay_is_blacklisted_multi($relay, $timeout, $answers_wanted, [$do‐
1620 main1, $domain2, ...], $res)
1621 This function is similar to relay_is_blacklisted, except that it
1622 takes a timeout argument (specified in seconds) and an array of
1623 domains to check. The function checks all domains in parallel,
1624 and is guaranteed to return in $timeout seconds. (Actually, it
1625 may take up to one second longer.)
1626 The parameters are:
1628 $relay -- the IP address you want to look up
1630 $timeout -- a timeout in seconds after which the function should
1632 return
1633 $answers_wanted -- the maximum number of positive answers you
1635 care about. For example, if you're looking up an address in 10
1636 different RBLs, but are going to bounce it if it is on four or
1637 more, you can set $answers_wanted to 4, and the function returns
1638 as soon as four "hits" are discovered. If you set $an‐
1639 swers_wanted to zero, then the function does not return early.
1640 [$domain1, $domain2, ...] -- a reference to an array of strings,
1642 where each string is an RBL domain.
1643 $res -- a Net::DNS::Resolver object. This argument is optional;
1645 if you do not supply it, then relay_is_blacklisted_multi con‐
1646 structs its own resolver.
1647 The return value is a reference to a hash; the keys of the hash
1649 are the original domains, and the corresponding values are ei‐
1650 ther SERVFAIL, NXDOMAIN, or a list of IP addresses in dotted-
1651 quad notation.
1652 Here's an example:
1654 $ans = relay_is_blacklisted_multi($RelayAddr, 8, 0,
1656 ["sbl.spamhaus.org", "relays.ordb.org"]);
1657 foreach $domain (keys(%$ans)) {
1659 $r = $ans->{$domain};
1660 if (ref($r) eq "ARRAY") {
1661 # It's an array -- it IS listed in RBL
1662 print STDERR "Lookup in $domain yields [ ";
1663 foreach $addr (@$r) {
1664 print STDERR $addr . " ";
1665 }
1666 print STDERR "]\n";
1667 } else {
1668 # It is NOT listed in RBL
1669 print STDERR "Lookup in $domain yields "
1670 . $ans->{$domain} . "\n";
1671 }
1672 }
1673 You should compare each of $ans->{$domain} to "SERVFAIL" and
1675 "NXDOMAIN" to see if the relay is not listed. Any other return
1676 value will be an array of IP addresses indicating that the relay
1677 is listed.
1678 Any lookup that does not succeed within $timeout seconds has the
1680 corresponding return value set to SERVFAIL.
1681 relay_is_blacklisted_multi_list($relay, $timeout, $answers_wanted,
1684 [$domain1, $domain2, ...], $res)
1685 This function is similar to relay_is_blacklisted_multi except
1686 that the return value is simply an array of RBL domains in which
1687 the relay was listed.
1688 relay_is_blacklisted_multi_count($relay, $timeout, $answers_wanted,
1691 [$domain1, $domain2, ...], $res)
1692 This function is similar to relay_is_blacklisted_multi except
1693 that the return value is an integer specifying the number of do‐
1694 mains on which the relay was blacklisted.
1695 md_get_bogus_mx_hosts($domain)
1698 This function looks up all the MX records for the specified do‐
1700 main (or A records if there are no MX records) and returns a
1701 list of "bogus" IP addresses found amongst the records. A "bo‐
1702 gus" IP address is an IP address in a private network
1703 (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16), the loopback net‐
1704 work (127.0.0.0/8), local-link for auto-DHCP (169.254.0.0/16),
1705 IPv4 multicast (224.0.0.0/4) or reserved (240.0.0.0/4).
1706 Here's how you might use the function in filter_sender:
1709 sub filter_sender {
1711 my ($sender, $ip, $hostname, $helo) = @_;
1712 if ($sender =~ /@([^>]+)/) {
1713 my $domain = $1;
1714 my @bogushosts = md_get_bogus_mx_hosts($domain);
1715 if (scalar(@bogushosts)) {
1716 return('REJECT', "Domain $domain contains bogus MX record(s) " .
1717 join(', ', @bogushosts));
1718 }
1719 }
1720 return ('CONTINUE', 'ok');
1721 }
1722
1725 mimedefang.pl includes some "test" functions:
1726 md_version()
1729 returns the version of MIMEDefang as a string (for example,
1730 "3.3").
1731 message_rejected()
1734 Returns true if any of action_tempfail, action_bounce or ac‐
1735 tion_discard have been called for this message; returns false
1736 otherwise.
1737 If you have the Mail::SpamAssassin Perl module installed (see
1740 https://spamassassin.apache.org) you may call any of the spam_assas‐
1741 sin_* functions. They should only be called from filter_begin or fil‐
1742 ter_end because they operate on the entire message at once. Most func‐
1743 tions use an optionally provided config file. If no config file is
1744 provided, mimedefang.pl will look for one of four default SpamAssassin
1745 preference files. The first of the following found will be used:
1746 o /etc/mail/sa-mimedefang.cf
1749 o /etc/mail/spamassassin/sa-mimedefang.cf
1751 o /etc/mail/spamassassin/local.cf
1753 o /etc/mail/spamassassin.cf
1755 Important Note: MIMEDefang does not permit SpamAssassin to modify mes‐
1758 sages. If you want to tag spam messages with special headers or alter
1759 the subject line, you must use MIMEDefang functions to do it. Setting
1760 SpamAssassin configuration options to alter messages will not work.
1761 spam_assassin_is_spam([ $config_file ])
1764 Determine if the current message is SPAM/UCE as determined by
1765 SpamAssassin. Compares the score of the message against the
1766 threshold score (see below) and returns true if it is. Uses
1767 spam_assassin_check below.
1768 spam_assassin_check([ $config_file ])
1771 This function returns a four-element list of the form ($hits,
1772 $required, $tests, $report). $hits is the "score" given to the
1773 message by SpamAssassin (higher score means more likely SPAM).
1774 $required is the number of hits required before SpamAssassin
1775 concludes that the message is SPAM. $tests is a comma-separated
1776 list of SpamAssassin test names, and $report is text detailing
1777 which tests triggered and their point score. This gives you in‐
1778 sight into why SpamAssassin concluded that the message is SPAM.
1779 Uses spam_assassin_status below.
1780 spam_assassin_status([ $config_file ])
1783 This function returns a Mail::SpamAssasin::PerMsgStatus object.
1784 Read the SpamAssassin documentation for details about this ob‐
1785 ject. You are responsible for calling the finish method when
1786 you are done with it. Uses spam_assassin_init and spam_assas‐
1787 sin_mail below.
1788 spam_assassin_init([ $config_file ])
1791 This function returns the new global Mail::SpamAssassin object
1792 with the specified or default config (outlined above). If the
1793 global object is already defined, returns it -- does not change
1794 config files! The object can be used to perform other SpamAs‐
1795 sassin related functions.
1796 spam_assassin_mail()
1799 This function returns a Mail::SpamAssassin::NoMailAudit object
1800 with the current email message contained in it. It may be used
1801 to perform other SpamAssassin related functions.
1802 rspamd_check([ $uri ]) *experimental*
1805 This function returns a six-element list of the form ($hits,
1806 $required, $tests, $report, $action, $is_spam). $hits is the
1807 "score" given to the message by Rspamd (higher score means more
1808 likely SPAM). $required is the number of hits required before
1809 Rspamd concludes that the message is SPAM. $tests is a list of
1810 Rspamd test names, and $report is text detailing which tests
1811 triggered and their point score. If JSON and LWP modules are
1812 present $report will be a json string; $action is the action
1813 that rspamd(8) wants to apply and $is_spam is a boolean value
1814 (true/false) that determines if the message is spam or not.
1815 This gives you insight into why Rspamd concluded that the mes‐
1816 sage is SPAM.
1817 md_copy_orig_msg_to_work_dir()
1820 Normally, virus-scanners are passed only the unpacked, decoded
1821 parts of a MIME message. If you want to pass the original, un‐
1822 decoded message in as well, call md_copy_orig_msg_to_work_dir
1823 prior to calling message_contains_virus.
1824 md_copy_orig_msg_to_work_dir_as_mbox_file()
1827 Normally, virus-scanners are passed only the unpacked, decoded
1828 parts of a MIME message. If you want to pass the original, un‐
1829 decoded message in as a UNIX-style "mbox" file, call
1830 md_copy_orig_msg_to_work_dir_as_mbox_file prior to calling mes‐
1831 sage_contains_virus. The only difference between this function
1832 and md_copy_orig_msg_to_work_dir is that this function prepends
1833 a "From_" line to make the message look like a UNIX-style mbox
1834 file. This is required for some virus scanners (such as Clam
1835 AntiVirus) to recognize the file as an e-mail message.
1836 message_contains_virus()
1839 This function runs every installed virus-scanner and returns the
1840 scanner results. The function should be called in list context;
1841 the return value is a three-element list ($code, $category, $ac‐
1842 tion).
1843 $code is the actual return code from the virus scanner.
1845 $category is a string categorizing the return code:
1847 "ok" - no viruses detected.
1849 "not-installed" - indicated virus scanner is not installed.
1851 "cannot-execute" - for some reason, the scanner could not be ex‐
1853 ecuted.
1854 "virus" - a virus was found.
1856 "suspicious" - a "suspicious" file was found.
1858 "interrupted" - scanning was interrupted.
1860 "swerr" - an internal scanner software error occurred.
1862 $action is a string containing the recommended action:
1864 "ok" - allow the message through unmolested.
1866 "quarantine" - a virus was detected; quarantine it.
1868 "tempfail" - something went wrong; tempfail the message.
1870 message_contains_virus_trend()
1874 message_contains_virus_nai()
1876 message_contains_virus_bdc()
1878 message_contains_virus_nvcc()
1880 message_contains_virus_csav()
1882 message_contains_virus_fsav()
1884 message_contains_virus_hbedv()
1886 message_contains_virus_vexira()
1888 message_contains_virus_sophos()
1890 message_contains_virus_clamav()
1892 message_contains_virus_clamdscan()
1894 message_contains_virus_avp()
1896 message_contains_virus_avp5()
1898 message_contains_virus_fprot()
1900 message_contains_virus_fpscan()
1902 message_contains_virus_fprotd()
1904 message_contains_virus_fprotd_v6()
1906 message_contains_virus_nod32()
1908 These functions should be called in list context. They use the
1910 indicated anti-virus software to scan the message for viruses.
1911 These functions are intended for use in filter_begin() to make
1912 an initial scan of the e-mail message.
1913 The supported virus scanners are:
1915 nai NAI "uvscan" - http://www.nai.com/
1917 Bitdefender "bdc" - http://www.bitdefender.com/
1919 csav Command Anti-Virus - http://www.commandsoftware.com/
1921 fsav F-Secure Anti-Virus - http://www.f-secure.com/
1923 hbedv H+BEDV "AntiVir" - http://www.hbedv.com/
1925 vexira Vexira "Vexira" - http://www.centralcommand.com/
1927 sophos Sophos AntiVirus - http://www.sophos.com/
1929 avp Kaspersky AVP and aveclient (AVP5) - http://www.avp.ru/
1931 clamav Clam AntiVirus - https://www.clamav.net/
1933 f-prot F-RISK F-PROT - http://www.f-prot.com/
1935 nod32cli
1937 ESET NOD32 - http://www.eset.com/
1938 message_contains_virus_carrier_scan([$host])
1941 Connects to the specified host:port:local_or_nonlocal (default
1942 $CSSHost), where the Symantec CarrierScan Server daemon is ex‐
1943 pected to be listening. Return values are the same as the other
1944 message_contains_virus functions.
1945 message_contains_virus_sophie([$sophie_sock])
1948 Connects to the specified socket (default $SophieSock), where
1949 the Sophie daemon is expected to be listening. Return values
1950 are the same as the other message_contains_virus functions.
1951 message_contains_virus_clamd([$clamd_sock])
1954 Connects to the specified socket (default $ClamdSock), where the
1955 clamd daemon is expected to be listening. Return values are the
1956 same as the other message_contains_virus functions.
1957 message_contains_virus_trophie([$trophie_sock])
1960 Connects to the specified socket (default $TrophieSock), where
1961 the Trophie daemon is expected to be listening. Return values
1962 are the same as the other message_contains_virus functions.
1963 entity_contains_virus($entity)
1966 This function runs the specified MIME::Entity through every in‐
1968 stalled virus-scanner and returns the scanner results. The re‐
1969 turn values are the same as for message_contains_virus().
1970 entity_contains_virus_trend($entity)
1973 entity_contains_virus_nai($entity)
1975 entity_contains_virus_bdc($entity)
1977 entity_contains_virus_nvcc($entity)
1979 entity_contains_virus_csav($entity)
1981 entity_contains_virus_fsav($entity)
1983 entity_contains_virus_hbedv($entity)
1985 entity_contains_virus_sophos($entity)
1987 entity_contains_virus_clamav($entity)
1989 entity_contains_virus_clamdscan($entity)
1991 entity_contains_virus_avp($entity)
1993 entity_contains_virus_avp5($entity)
1995 entity_contains_virus_fprot($entity)
1997 entity_contains_virus_fpscan($entity)
1999 entity_contains_virus_fprotd($entity)
2001 entity_contains_virus_fprotd_v6($entity)
2003 entity_contains_virus_nod32($entity)
2005 These functions, meant to be called from filter(), are similar
2006 to the message_contains_virus functions except they scan only
2007 the current part. They should be called from list context, and
2008 their return values are as described for the message_con‐
2009 tains_virus functions.
2010 entity_contains_virus_carrier_scan($entity[, $host])
2013 Connects to the specified host:port:local_or_nonlocal (default
2014 $CSSHost), where the Symantec CarrierScan Server daemon is ex‐
2015 pected to be listening. Return values are the same as the other
2016 entity_contains_virus functions.
2017 entity_contains_virus_sophie($entity[, $sophie_sock])
2020 Connects to the specified socket (default $SophieSock), where
2021 the Sophie daemon is expected to be listening. Return values
2022 are the same as the other entity_contains_virus functions.
2023 entity_contains_virus_trophie($entity[, $trophie_sock])
2026 Connects to the specified socket (default $TrophieSock), where
2027 the Trophie daemon is expected to be listening. Return values
2028 are the same as the other entity_contains_virus functions.
2029 entity_contains_virus_clamd($entity[, $clamd_sock])
2032 Connects to the specified socket (default $ClamdSock), where the
2033 clamd daemon is expected to be listening. Return values are the
2034 same as the other entity_contains_virus functions.
2035
2038 This section illustrates the flow of messages through MIMEDefang.
2039 1. INITIAL CONNECTION
2042 If you invoked mimedefang with the -r option and have defined a
2043 filter_relay routine, it is called.
2044 2. SMTP HELO COMMAND
2047 The HELO string is stored internally, but no filter functions
2048 are called.
2049 3. SMTP MAIL FROM: COMMAND
2052 If you invoked mimedefang with the -s option and have defined a
2053 filter_sender routine, it is called.
2054 4. SMTP RCPT TO: COMMAND
2057 If you invoked mimedefang with the -t option and have defined a
2058 filter_recipient routine, it is called.
2059 5. END OF SMTP DATA
2062 filter_begin is called. For each MIME part, filter is called.
2063 Then filter_end is called.
2064
2067 Most organizations have more than one machine handling internet e-mail.
2068 If the primary machine is down, mail is routed to a secondary (or ter‐
2069 tiary, etc.) MX server, which stores the mail until the primary MX host
2070 comes back up. Mail is then relayed to the primary MX host.
2071 Relaying from a secondary to a primary MX host has the unfortunate side
2074 effect of losing the original relay's IP address information. MIMEDe‐
2075 fang allows you to preserve this information. One way around the prob‐
2076 lem is to run MIMEDefang on all the secondary MX hosts and use the same
2077 filter. However, you may not have control over the secondary MX hosts.
2078 If you can persuade the owners of the secondary MX hosts to run MIMEDe‐
2079 fang with a simple filter that only preserves relay information and
2080 does no other scanning, your primary MX host can obtain relay informa‐
2081 tion and make decisions using $RelayAddr and $RelayHostname.
2082 When you configure MIMEDefang, supply the "--with-ipheader" argument to
2085 the ./configure script. When you install MIMEDefang, a file called
2086 /etc/mail/mimedefang-ip-key will be created which contains a randomly-
2087 generated header name. Copy this file to all of your mail relays. It
2088 is important that all of your MX hosts have the same key. The key
2089 should be kept confidential, but it's not disastrous if it leaks out.
2090 On your secondary MX hosts, add this line to filter_end:
2093 add_ip_validation_header();
2095 Note: You should only add the validation header to mail destined for
2098 one of your other MX hosts! Otherwise, the validation header will leak
2099 out.
2100 When the secondary MX hosts relay to the primary MX host, $RelayAddr
2103 and $RelayHostname will be set based on the IP validation header. If
2104 MIMEDefang notices this header, it sets the global variable $WasResent
2105 to 1. Since you don't want to trust the header unless it was set by
2106 one of your secondary MX hosts, you should put this code in filter_be‐
2107 gin:
2108 if ($WasResent) {
2110 if ($RealRelayAddr ne "ip.of.secondary.mx" and
2111 $RealRelayAddr ne "ip.of.tertiary.mx") {
2112 $RelayAddr = $RealRelayAddr;
2113 $RelayHostname = $RealRelayHostname;
2114 }
2115 }
2116 This resets the relay address and hostname to the actual relay address
2118 and hostname, unless the message is coming from one of your other MX
2119 hosts.
2120 On the primary MX host, you should add this in filter_begin:
2123 delete_ip_validation_header();
2125 This prevents the validation header from leaking out to recipients.
2128 Note: The IP validation header works only in message-oriented func‐
2131 tions. It (obviously) has no effect on filter_relay, filter_sender and
2132 filter_recipient, because no header information is available yet. You
2133 must take this into account when writing your filter; you must defer
2134 relay-based decisions to the message filter for mail arriving from your
2135 other MX hosts.
2136
2139 The following list describes the lifetime of global variables (thanks
2140 to Tony Nugent for providing this documentation.)
2141 If you set a global variable:
2143 Outside a subroutine in your filter file
2146 It is available to all functions, all the time.
2147 In filter_relay, filter_sender or filter_recipient
2150 Not guaranteed to be available to any other function, not even
2151 from one filter_recipient call to the next, when receiving a
2152 multi-recipient email message.
2153 In filter_begin
2156 Available to filter_begin, filter and filter_end
2157 In filter
2160 Available to filter and filter_end
2161 In filter_end
2164 Available within filter_end
2165 The "built-in" globals like $Subject, $Sender, etc. are always avail‐
2168 able to filter_begin, filter and filter_end. Some are available to fil‐
2169 ter_relay, filter_sender or filter_recipient, but you should check the
2170 documentation of the variable above for details.
2171
2174 There are four basic groups of filtering functions:
2175 1 filter_relay
2178 2 filter_sender
2181 3 filter_recipient
2184 4 filter_begin, filter, filter_multipart, filter_end
2187 In general, for a given mail message, these groups of functions may be
2190 called in completely different Perl processes. Thus, there is no way
2191 to maintain state inside Perl between groups of functions. That is,
2192 you cannot set a variable in filter_relay and expect it to be available
2193 in filter_sender, because the filter_sender invocation might take place
2194 in a completely different process.
2195 For a given mail message, it is always the case that filter_begin, fil‐
2198 ter, filter_multipart and filter_end are called in the same Perl
2199 process. Therefore, you can use global variables to carry state among
2200 those functions. You should be very careful to initialize such vari‐
2201 ables in filter_begin to ensure no data leaks from one message to an‐
2202 other.
2203 Also for a given mail message, the $CWD global variable holds the mes‐
2206 sage spool directory, and the current working directory is set to $CWD.
2207 Therefore, you can store state in files inside $CWD. If filter_sender
2208 stores data in a file inside $CWD, then filter_recipient can retrieve
2209 that data.
2210 Since filter_relay is called directly after a mail connection is estab‐
2213 lished, there is no message context yet, no per-message mimedefang
2214 spool directory, and the $CWD global is not set. Therefore, it is not
2215 possible to share information from filter_relay to one of the other
2216 filter functions. The only thing that filter_relay has in common with
2217 the other functions are the values in the globals $RelayAddr, and $Re‐
2218 layHostname. These could be used to access per-remote-host information
2219 in some database.
2220 Inside $CWD, we reserve filenames beginning with upper-case letters for
2223 internal MIMEDefang use. If you want to create files to store state,
2224 name them beginning with a lower-case letter to avoid clashes with fu‐
2225 ture releases of MIMEDefang.
2226
2229 If you have Sendmail 8.13 or later, and have compiled it with the SOCK‐
2230 ETMAP option, then you can use a special map type that communicates
2231 over a socket with another program (rather than looking up a key in a
2232 Berkeley database, for example.)
2233 mimedefang-multiplexor implements the Sendmail SOCKETMAP protocol if
2236 you supply the -N option. In that case, you can define a function
2237 called filter_map to implement map lookups. filter_map takes two argu‐
2238 ments: $mapname is the name of the Sendmail map (as given in the K
2239 sendmail configuration directive), and $key is the key to be looked up.
2240 filter_map must return a two-element list: ($code, $val) $code can be
2243 one of:
2244 OK The lookup was successful. In this case, $val must be the re‐
2247 sult of the lookup
2248 NOTFOUND
2251 The lookup was unsuccessful -- the key was not found. In this
2252 case, $val should be the empty string.
2253 TEMP There was a temporary failure of some kind. $val can be an ex‐
2256 planatory error message.
2257 TIMEOUT
2260 There was a timeout of some kind. $val can be an explanatory
2261 error message.
2262 PERM There was a permanent failure. This is not the same as an un‐
2265 successful lookup; it should be used only to indicate a serious
2266 misconfiguration. As before, $val can be an explanatory error
2267 message.
2268 Consider this small example. Here is a minimal Sendmail configuration
2271 file:
2272 V10/Berkeley
2274 Kmysock socket unix:/var/spool/MIMEDefang/map.sock
2275 kothersock socket unix:/var/spool/MIMEDefang/map.sock
2276 If mimedefang-multiplexor is invoked with the arguments -N
2279 unix:/var/spool/MIMEDefang/map.sock, and the filter defines filter_map
2280 as follows:
2281 sub filter_map ($$) {
2283 my($mapname, $key) = @_;
2284 my $ans;
2285 if($mapname ne "mysock") {
2286 return("PERM", "Unknown map $mapname");
2287 }
2288 $ans = reverse($key);
2289 return ("OK", $ans);
2290 }
2291 Then in Sendmail's testing mode, we see the following:
2293 > /map mysock testing123
2295 map_lookup: mysock (testing123) returns 321gnitset (0)
2296 > /map othersock foo
2297 map_lookup: othersock (foo) no match (69)
2298 (The return code of 69 means EX_UNAVAILABLE or Service Unavailable)
2301 A real-world example could do map lookups in an LDAP directory or SQL
2304 database, or perform other kinds of processing. You can even implement
2305 standard Sendmail maps like virtusertable, mailertable, access_db, etc.
2306 using SOCKETMAP.
2307
2310 If you supply the -X option to mimedefang-multiplexor, then every so
2311 often, a "tick" request is sent to a free worker. If your filter de‐
2312 fines a function called filter_tick, then this function is called with
2313 a single argument: the tick type. If you run multiple parallel ticks,
2314 then each tick has a type ranging from 0 to n-1, where n is the number
2315 of parallel ticks. If you're only running one tick request, then the
2316 argument to filter_tick is always 0.
2317 You can use this facility to run periodic tasks from within MIMEDefang.
2319 Note, however, that you have no control over which worker is picked to
2320 run filter_tick. Also, at most one filter_tick call with a particular
2321 "type" argument will be active at any time, and if there are no free
2322 workers when a tick would occur, the tick is skipped.
2323
2326 The following virus scanners are supported by MIMEDefang:
2327 o Symantec CarrierScan Server (http://www.symantec.com/re‐
2330 gion/can/eng/product/scs/)
2331 o Trend Micro vscan (http://www.antivirus.com/)
2334 o Sophos Sweep (http://www.sophos.com/products/antivirus/savu‐
2337 nix.html)
2338 o H+BEDV AntiVir (http://www.hbedv.com/)
2341 o Central Command Vexira (http://www.centralcommand.com/)
2344 o NAI uvscan (http://www.nai.com)
2347 o Bitdefender bdc (http://www.bitdefender.com)
2350 o Norman Virus Control (NVCC) (http://www.norman.no/)
2353 o Command csav (http://www.commandsoftware.com)
2356 o F-Secure fsav (http://www.f-secure.com)
2359 o The clamscan and clamdscan command-line scanners and the clamd
2362 daemon from Clam AntiVirus (https://www.clamav.net/)
2363 o Kaspersky Anti-Virus (AVP) (http://www.kaspersky.com/)
2366 o F-Risk F-Prot (http://www.f-prot.com/)
2369 o F-Risk F-Prot v6 (http://www.f-prot.com/)
2372 o F-Risk FPROTD (daemonized version of F-Prot)
2376 o Symantec CarrierScan Server (http://www.symantec.ca/re‐
2379 gion/can/eng/product/scs/buymenu.html)
2380 o Sophie (http://www.vanja.com/tools/sophie/), which uses the lib‐
2383 savi library from Sophos, is supported in daemon-scanning mode.
2384 o Trophie (http://www.vanja.com/tools/trophie/), which uses the
2387 libvsapi library from Trend Micro, is supported in daemon-scan‐
2388 ning mode.
2389 o ESET NOD32 (http://www.eset.com/)
2392
2395 mimedefang was written by Dianne Skoll <dfs@roaringpenguin.com>. The
2396 mimedefang home page is https://www.mimedefang.org/.
2397
2400 mimedefang(8), mimedefang.pl(8), Mail::MIMEDefang(3)
24014th Berkeley Distribution 8 February 2005 MIMEDEFANG-FILTER(5)