Mercurial > notdcc

comparison dccifd.0 @ 0:c7f6b056b673

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
First import of vendor version
author Peter Gervai <grin@grin.hu>
date Tue, 10 Mar 2009 13:49:58 +0100
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:c7f6b056b673
1 dccifd(8) Distributed Checksum Clearinghouse dccifd(8)
2
3 NNAAMMEE
4 ddcccciiffdd -- Distributed Checksum Clearinghouse Interface Daemon
5
6 SSYYNNOOPPSSIISS
7 ddcccciiffdd [--VVddbbxxAANNQQ] [--GG _o_n | _o_f_f | _n_o_I_P | _I_P_m_a_s_k_/_x_x] [--hh _h_o_m_e_d_i_r] [--II _u_s_e_r]
8 [--pp _/_s_o_c_k | _h_o_s_t_,_p_o_r_t_,_r_h_o_s_t_/_b_i_t_s] [--oo _/_s_o_c_k | _h_o_s_t_,_p_o_r_t]
9 [--DD _l_o_c_a_l_-_d_o_m_a_i_n] [--mm _m_a_p] [--ww _w_h_i_t_e_c_l_n_t] [--UU _u_s_e_r_d_i_r_s]
10 [--aa _I_G_N_O_R_E | _R_E_J_E_C_T | _D_I_S_C_A_R_D] [--tt _t_y_p_e_,[_l_o_g_-_t_h_o_l_d_,]_r_e_j_-_t_h_o_l_d]
11 [--gg [_n_o_t_-]_t_y_p_e] [--SS _h_e_a_d_e_r] [--ll _l_o_g_d_i_r] [--RR _r_u_n_d_i_r]
12 [--rr _r_e_j_e_c_t_i_o_n_-_m_s_g] [--TT _t_m_p_d_i_r] [--jj _m_a_x_j_o_b_s]
13 [--BB _d_n_s_b_l_-_o_p_t_i_o_n] [--LL _l_t_y_p_e_,_f_a_c_i_l_i_t_y_._l_e_v_e_l]
14
15 DDEESSCCRRIIPPTTIIOONN
16 ddcccciiffdd is a daemon intended to connect spam filters such as SpamAssasin
17 and mail transfer agents (MTAs) other than sendmail to DCC servers. The
18 MTA or filter ddcccciiffdd which in turn reports related checksums to the near-
19 est DCC server and adds an _X_-_D_C_C SMTP header line to the message. The
20 MTA is told to reject the message if it is unsolicited bulk.
21
22 DDcccciiffdd is similar to the DCC sendmail milter interface, dccm(8) and the
23 DCC Procmail interface, dccproc(8). DDcccciiffdd is more efficient than
24 dccproc(8) but not restricted to use with sendmail like dccm(8). All
25 three send reports of checksums related to mail received by DCC clients
26 and queries about the total number of reports of particular checksums.
27
28 MTA programs use a simple ASCII protocol a subset of SMTP to send a mail
29 message including its SMTP envelope to the daemon. DDcccciiffdd responds with
30 an indication of whether the message is unsolicited bulk and an optional
31 copy of the message with an _X_-_D_C_C header added. The ASCII protocol is
32 described below and in the _i_n_c_l_u_d_e_/_d_c_c_i_f_._h file in the DCC source. There
33 is a sample C interface routine in the _d_c_c_l_i_b_/_d_c_c_i_f_._c file in the DCC
34 source and the _d_c_c_l_i_b_._a library generated from the source. A _P_e_r_l ver-
35 sion of the interface routine is in _d_c_c_i_f_d_/_d_c_c_i_f_._p_l. Test or demonstra-
36 tion programs in the style of dccproc(8) that use those interface rou-
37 tines are in _d_c_c_i_f_d_/_d_c_c_i_f_-_t_e_s_t.
38
39 A subset of ESMTP can be used instead of the ASCII protocol to connect
40 ddcccciiffdd to postfix as a "Before-Queue Content Filter." See the --oo flag.
41
42 Since the checksums of messages that are whitelisted locally by the --ww
43 _w_h_i_t_e_c_l_n_t file are not reported to the DCC server, ddcccciiffdd knows nothing
44 about the total recipient counts for their checksums and so cannot add
45 _X_-_D_C_C header lines to such messages.
46
47 Enable the daemon and put its parameters in the _d_c_c___c_o_n_f file and start
48 the daemon with the _s_t_a_r_t_-_d_c_c_i_f_d script.
49
50 The list of servers that ddcccciiffdd contacts is in the memory mapped file _m_a_p
51 shared by local DCC clients. The file is maintained with cdcc(8).
52
53 OOPPTTIIOONNSS
54 The following options are available:
55
56 --VV displays the version of ddcccciiffdd.
57
58 --dd enables debugging output from the DCC client software. Additional
59 --dd options increase the number of messages. A single --dd
60 aborted SMTP transactions including those from some "dictionary
61 attacks."
62
63 --bb causes the daemon to not detach itself from the controlling tty and
64 put itself into the background.
65
66 --xx causes the daemon to try "extra hard" to contact a DCC server.
67 Since it is usually more important to deliver mail than to report
68 its checksums, ddcccciiffdd normally does not delay too long while trying
69 to contact a DCC server. It will not try again for several seconds
70 after a failure. With --xx, it will always try to contact the DCC
71 server and it will tell the MTA to answer the DATA command with a
72 4yz temporary failure.
73
74 --AA adds to existing X-DCC headers in the message instead of replacing
75 existing headers of the brand of the current server.
76
77 --NN neither adds, deletes, nor replaces existing X-DCC headers in the
78 message. Each message is logged, rejected, and otherwise handled
79 the same.
80
81 --QQ only queries the DCC server about the checksums of messages instead
82 of reporting and querying. This is useful when ddcccciiffdd is used to
83 filter mail that has already been reported to a DCC server by
84 another DCC client. No single mail message should be reported to a
85 DCC server more than once per recipient, because each report will
86 increase the apparent "bulkness" of the message.
87
88 It is better to use _M_X_D_C_C lines in the global _w_h_i_t_e_c_l_n_t file for
89 your MX mail servers that use DCC than --QQ.
90
91 --GG _o_n | _o_f_f | _n_o_I_P | _I_P_m_a_s_k_/_x_x
92 controls _g_r_e_y_l_i_s_t_i_n_g. At least one working greylist server must be
93 listed in the _m_a_p file in the DCC home directory. If more than one
94 is named, they must "flood" or change checksums and they must use
95 the same --GG parameters. See dccd(8). Usually all dccm or dccifd
96 DCC client processes use the same --GG parameters.
97
98 _I_P_m_a_s_k_/_x_x and _n_o_I_P remove part or all of the IP address from the
99 greylist triple. The CIDR block size, _x_x, must be between 1 and
100 128. 96 is added to block sizes smaller than 33 to make them appro-
101 priate for the IPv6 addresses used by the DCC. _I_P_m_a_s_k_/_9_6 differs
102 from _n_o_I_P for IPv4 addresses, because the former retains the IPv4 to
103 IPv6 mapping prefix.
104
105 --hh _h_o_m_e_d_i_r
106 overrides the default DCC home directory, _/_v_a_r_/_d_c_c.
107
108 --II _u_s_e_r
109 specifies the UID and GID of the process.
110
111 --pp _/_s_o_c_k_/_n_a_m_e | _h_o_s_t_,_p_o_r_t_,_r_h_o_s_t_/_b_i_t_s
112 overrides the default address at which programs contact ddcccciiffdd. The
113 default is a UNIX domain socket named dccifd in the DCC home direc-
114 tory.
115
116 The second form specifies a local host name or IP address, a local
117 TCP port number, and the host names or IP addresses of computers
118 that can use ddcccciiffdd. 127.0.0.1 or _l_o_c_a_l_h_o_s_t are common choices for
119 _h_o_s_t. The string _@ specifies IN_ADDRANY or all local IP addresses.
120 127.0.0.0/8 is a common choice for _r_h_o_s_t_/_b_i_t_s.
121
122 --oo _/_s_o_c_k | _h_o_s_t_,_p_o_r_t
123 enables SMTP proxy mode instead of the ASCII protocol and specifies
124 the output connection when ddcccciiffdd acts as an SMTP proxy. It is the
125 address of the SMTP server for which ddcccciiffdd acts as SMTP client.
126 When _/_s_o_c_k is _/_d_e_v_/_n_u_l_l, ddcccciiffdd acts as if there were downstream
127 SMTP server that always answers "250 ok". The string _@ specifies
128 the same IP address as the incoming TCP connection.
129
130 The input to ddcccciiffdd in SMTP proxy mode is specified with ----pp. For
131 example, --pp _1_2_7_._0_._0_._1_,_1_0_0_2_5_,_1_2_7_._0_._0_._1_/_3_2 --oo _1_2_7_._0_._0_._1_,_1_0_0_2_6 could be
132 used to connect ddcccciiffdd with Postfix as described in the documenta-
133 tion in version 2.2.1 Postfix documentation.
134
135 See below concerning the subset of ESMTP used in this mode.
136
137 --mm _m_a_p
138 specifies a name or path of the memory mapped parameter file instead
139 of the default _m_a_p file in the DCC home directory. It should be
140 created with the cdcc(8) command.
141
142 --ww _w_h_i_t_e_c_l_n_t
143 specifies an optional file containing filtering parameters as well
144 as SMTP client IP addresses, SMTP envelope values, and header values
145 of mail that is spam or is not spam and does not need a _X_-_D_C_C
146 header, and whose checksums should not be reported to the DCC
147 server.
148
149 If the pathname _w_h_i_t_e_c_l_n_t is not absolute, it is relative to the DCC
150 home directory.
151
152 The format of the ddcccciiffdd whiteclnt file is the same as the _w_h_i_t_e_l_i_s_t
153 files used by dbclean(8) and the _w_h_i_t_e_c_l_n_t file used by dccproc(8).
154 See dcc(8) for a description of DCC white and blacklists. Because
155 the contents of the _w_h_i_t_e_c_l_n_t file are used frequently, a companion
156 file is automatically created and maintained. It has the same path-
157 name but with an added suffix of _._d_c_c_w and contains a memory mapped
158 hash table of the main file.
159
160 A whitelist entry ("OK") or two or more semi-whitelistings ("OK2")
161 for one of the message's checksums prevents all of the message's
162 checksums from being reported to the DCC server and the addition of
163 a _X_-_D_C_C header line by ddcccciiffdd A whitelist entry for a checksum also
164 prevents rejecting or discarding the message based on DCC recipient
165 counts as specified by --aa and --tt. Otherwise, one or more checksums
166 with blacklisting entries ("MANY") cause all of the message's check-
167 sums to be reported to the server with an addressee count of "MANY".
168
169 If the message has a single recipient, an _e_n_v___T_o _w_h_i_t_e_c_l_n_t entry of
170 "OK" for the checksum of its recipient address acts like any other
171 _w_h_i_t_e_c_l_n_t entry of "OK." When the SMTP message has more than one
172 recipient, the effects can be complicated. When a message has sev-
173 eral recipients with some but not all listed in the _w_h_i_t_e_c_l_n_t file,
174 ddcccciiffdd tries comply with the wishes of the users who want filtering
175 as well as those who don't by silently not delivering the message to
176 those who want filtering (i.e. are not whitelisted) and delivering
177 the message to don't want filtering.
178
179 --UU _u_s_e_r_d_i_r_s
180 enables per-user _w_h_i_t_e_c_l_n_t files and log directories. Each target
181 of a message can have a directory of log files named
182 _u_s_e_r_d_i_r_s_/_a_d_d_r_/_l_o_g where _a_d_d_r is the local user or mailbox name com-
183 puted by the MTA. The name of each user's log directory must be
184 _l_o_g. If it is not absolute, _u_s_e_r_d_i_r_s is relative to the DCC home
185 directory. The directory containing the log files must be named _l_o_g
186 and it must be writable by the ddcccciiffdd process. Each log directory
187 must exist or logging for the corresponding is silently disabled.
188 The files created in the log directory are owned by the UID of the
189 ddcccciiffdd process, but they have _g_r_o_u_p and _o_t_h_e_r read and write permis-
190 sions copied from the corresponding _l_o_g directory. To ensure the
191 privacy of mail, it may be good to make the directories readable
192 only by _o_w_n_e_r and _g_r_o_u_p, and to use a cron script that changes the
193 owner of each file to match the grandparent _a_d_d_r directory.
194
195 There can also be a per -user whitelist file named
196 _u_s_e_r_d_i_r_s_/_a_d_d_r_/_w_h_i_t_e_c_l_n_t for each address _a_d_d_r_. Any checksum that is
197 not white- or blacklisted by an individual addressee's _w_h_i_t_e_c_l_n_t
198 file is checked in the main --ww --wwhhiitteeccllnntt file. A missing per-
199 addressee _w_h_i_t_e_c_l_n_t file is the same as an empty file. Relative
200 paths for files included in per-addressee files are resolved in the
201 DCC home directory. The _w_h_i_t_e_c_l_n_t files and the _a_d_d_r directories
202 containing them must be writable by the ddcccciiffdd process.
203
204 _O_p_t_i_o_n lines in per-user whiteclnt files can be used to modify many
205 aspects of ddcccciiffdd filtering, as described in the main dcc man page.
206 For example, an _o_p_t_i_o_n _d_c_c_-_o_f_f line turns off DCC filtering for
207 individual mailboxes.
208
209 --aa _I_G_N_O_R_E | _R_E_J_E_C_T | _D_I_S_C_A_R_D
210 specifies the action taken when ddcccciiffdd is in proxy mode with --oo and
211 DCC server counts or --tt thresholds say that a message is unsolicited
212 and bulk. _I_G_N_O_R_E causes the message to be unaffected except for
213 adding the _X_-_D_C_C header line to the message. This turns off DCC
214 filtering.
215
216 Spam can also be _R_E_J_E_C_Ted or (when in proxy mode with --oo) accepted
217 and silently _D_I_S_C_A_R_Ded without being delivered to local mailboxes.
218 The default is _R_E_J_E_C_T.
219
220 Mail forwarded via IP addresses marked _M_X or _M_X_D_C_C in the main
221 _w_h_i_t_e_c_l_n_t file is treated as if --aa _D_I_S_C_A_R_D were specified. This
222 prevents "bouncing" spam.
223
224 The effects of the --ww _w_h_i_t_e_c_l_n_t are not affected by --aa.
225
226 --tt _t_y_p_e_,[_l_o_g_-_t_h_o_l_d_,]_r_e_j_-_t_h_o_l_d
227 sets logging and "spam" thresholds for checksum _t_y_p_e. The checksum
228 types are _I_P, _e_n_v___F_r_o_m, _F_r_o_m, _M_e_s_s_a_g_e_-_I_D, _s_u_b_s_t_i_t_u_t_e, _R_e_c_e_i_v_e_d,
229 _B_o_d_y, _F_u_z_1, _F_u_z_2, _r_e_p_-_t_o_t_a_l, and _r_e_p. The first six, _I_P through
230 _s_u_b_s_t_i_t_u_t_e, have no effect except when a local DCC server configured
231 with --KK is used. The _s_u_b_s_t_i_t_u_t_e thresholds apply to the first sub-
232 stitute heading encountered in the mail message. The string _A_L_L
233 sets thresholds for all types, but is unlikely to be useful except
234 for setting logging thresholds. The string _C_M_N specifies the com-
235 monly used checksums _B_o_d_y, _F_u_z_1, and _F_u_z_2. _R_e_j_-_t_h_o_l_d and _l_o_g_-_t_h_o_l_d
236 must be numbers, the string _N_E_V_E_R, or the string _M_A_N_Y indicating
237 millions of targets. Counts from the DCC server as large as the
238 threshold for any single type are taken as sufficient evidence that
239 the message should be logged or rejected.
240
241 _L_o_g_-_t_h_o_l_d is the threshold at which messages are logged. It can be
242 handy to log messages at a lower threshold to find solicited bulk
243 mail sources such as mailing lists. If no logging threshold is set,
244 only rejected mail and messages with complicated combinations of
245 white and blacklisting are logged. Messages that reach at least one
246 of their rejection thresholds are logged regardless of logging
247 thresholds.
248
249 _R_e_j_-_t_h_o_l_d is the threshold at which messages are considered "bulk,"
250 and so should be rejected or discarded if not whitelisted.
251
252 DCC Reputation thresholds in the commercial version of the DCC are
253 controlled by thresholds on checksum types _r_e_p and _r_e_p_-_t_o_t_a_l. Mes-
254 sages from an IP address that the DCC database says has sent more
255 than --tt _r_e_p_-_t_o_t_a_l_,_l_o_g_-_t_h_o_l_d messages are logged. A DCC Reputation
256 is computed for messages received from IP addresses that have sent
257 more than --tt _r_e_p_-_t_o_t_a_l_,_l_o_g_-_t_h_o_l_d messages. The DCC Reputation of an
258 IP address is the percentage of its messages that have been detected
259 as bulk or having at least 10 recipients. The defaults are equiva-
260 lent to --tt _r_e_p_,_n_e_v_e_r and --tt _r_e_p_-_t_o_t_a_l_,_n_e_v_e_r_,_2_0.
261
262 Bad DCC Reputations do not reject mail unless enabled by an _o_p_t_i_o_n
263 _D_C_C_-_r_e_p_-_o_n line in a _w_h_i_t_e_c_l_n_t file.
264
265 The checksums of locally whitelisted messages are not checked with
266 the DCC server and so only the number of targets of the current copy
267 of a whitelisted message are compared against the thresholds.
268
269 The default is _A_L_L_,_N_E_V_E_R, so that nothing is discarded, rejected, or
270 logged. A common choice is _C_M_N_,_2_5_,_5_0 to reject or discard mail with
271 common bodies except as overridden by the whitelist of the DCC
272 server, the sendmail _$_{_d_c_c___i_s_s_p_a_m_} and _$_{_d_c_c___n_o_t_s_p_a_m_} macros, and
273 --gg, and --ww.
274
275 --gg [_n_o_t_-]_t_y_p_e
276 indicates that whitelisted, _O_K or _O_K_2, counts from the DCC server
277 for a type of checksum are to be believed. They should be ignored
278 if prefixed with _n_o_t_-. _T_y_p_e is one of the same set of strings as
279 for --tt. Only _I_P, _e_n_v___F_r_o_m, and _F_r_o_m are likely choices. By default
280 all three are honored, and hence the need for _n_o_t_-.
281
282 --SS _h_d_r
283 adds to the list of substitute or locally chosen headers that are
284 checked with the --ww _w_h_i_t_e_c_l_n_t file and sent to the DCC server. The
285 checksum of the last header of type _h_d_r found in the message is
286 checked. _H_d_r can be _H_E_L_O to specify the SMTP envelope HELO value.
287 _H_d_r can also be _m_a_i_l___h_o_s_t to specify the host name from the
288 Mail_from value in the SMTP envelope. As many as six different sub-
289 stitute headers can be specified, but only the checksum of the first
290 of the six will be sent to the DCC server.
291
292 --ll _l_o_g_d_i_r
293 specifies a directory in which files containing copies of messages
294 processed by ddcccciiffdd are kept. They can be copied to per-user direc-
295 tories specified with --UU. Information about other recipients of a
296 message is deleted from the per-user copies.
297
298 See the FILES section below concerning the contents of the files.
299 See also the _o_p_t_i_o_n _l_o_g_-_s_u_b_d_i_r_e_c_t_o_r_y_-_{_d_a_y_,_h_o_u_r_,_m_i_n_u_t_e_} lines in
300 _w_h_i_t_e_c_l_n_t files described in dcc(8).
301
302 The directory is relative to the DCC home directory if it is not
303 absolute
304
305 --RR _r_u_n_d_i_r
306 specifies the "run" directory where the file containing the daemon's
307 process ID is stored. The default value is _/_v_a_r_/_r_u_n_/_d_c_c.
308
309 --TT _t_m_p_d_i_r
310 changes the default directory for temporary files from the default.
311 The default is the directory specified with --ll or the system default
312 if --ll is not used. The system default is often _/_t_m_p.
313
314 --DD _l_o_c_a_l_-_d_o_m_a_i_n
315 specifies a host or domain name by which the system is known. There
316 can be several --DD settings.
317
318 To find the per-user log directory and whitelist for each mail
319 recipient, ddcccciiffdd must know each recipient's user name. The ASCII
320 protocol used between and the MTA includes an optional user name
321 with each SMTP recipient address. When the user name is absent when
322 the ASCII protocol is used or when the subset of ESMTP enabled with
323 --oo is used, and when the SMTP recipient address includes an _a_t _s_i_g_n
324 (@) each mail address is checked against the list of _l_o_c_a_l_-_d_o_m_a_i_ns.
325 The part of the recipient address remaining after longest matching
326 _l_o_c_a_l_-_d_o_m_a_i_n (if any) is taken as the user name. The match is
327 anchored at the right or the end of the recipient address. It must
328 start at a period (.) or _a_t _s_i_g_n (@) in the domain name part of the
329 address.
330
331 If _l_o_c_a_l_-_d_o_m_a_i_n starts with an asterisk (*) indicating a wildcard,
332 preceding sub-domain names are discarded to compute the user name.
333 Otherwise, the computed user name will include any unmatched sub-
334 domain names.
335
336 The default value of _l_o_c_a_l_-_d_o_m_a_i_n when there are no --DD settings is
337 the host name of the system.
338
339 --rr _r_e_j_e_c_t_i_o_n_-_m_s_g
340 specifies the rejection message in --oo proxy mode for unsolicited
341 bulk mail or for mail temporarily blocked by _g_r_e_y_l_i_s_t_i_n_g when --GG is
342 specified. The first --rr _r_e_j_e_c_t_i_o_n_-_m_s_g replaces the default bulk
343 mail rejection message, "5.7.1 550 mail %ID from %CIP rejected by
344 DCC". The second replaces "4.2.1 452 mail %ID from %CIP temporary
345 greylist embargoed". The third --rr _r_e_j_e_c_t_i_o_n_-_m_s_g replaces the
346 default SMTP rejection message "5.7.1 550 %ID bad reputation; see
347 http://commercial-dcc.rhyolite.com/cgi-bin/reps.cgi?tgt=%CIP" for
348 mail with bad DCC Reputations. If _r_e_j_e_c_t_i_o_n_-_m_s_g is the zero-length
349 string, the --rr setting is counted but the corresponding message is
350 not changed.
351
352 _R_e_j_e_c_t_i_o_n_-_m_s_g can contain specific information about the mail mes-
353 sage. The following strings starting with % are replaced with the
354 corresponding values:
355 %ID message ID such as the unique part of log file name or
356 sendmail queue ID
357 %CIP SMTP client IP address
358 %BTYPE type of DNS blacklist hit, such as "SMTP client",
359 "mail_host", or "URL NS"
360 %BTGT IP address or name declared bad by DNS blacklist
361 %BPROBE domain name found in DNS blacklist such as
362 4.3.2.10.example.com
363 %BRESULT value of the %BPROBE domain name found in DNS black-
364 list
365
366 A common alternate for the bulk mail rejection message is "4.7.1 451
367 Access denied by DCC" to tell the sending mail system to continue
368 trying. Use a 4yz response with caution, because it is likely to
369 delay for days a delivery failure message for false positives. If
370 the rejection message does not start with an RFC 1893 status code
371 and RFC 2821 reply code, 5.7.1 and 550 or 4.2.1 and 452 are used.
372
373 See also --BB _s_e_t_:_r_e_j_-_m_s_g_=_r_e_j_e_c_t_i_o_n_-_m_s_g to set the status message for
374 mail rejected by DNS blacklists.
375
376 --jj _m_a_x_j_o_b_s
377 limits the number of simultaneous requests that will be processed.
378 The default value is the maximum number that seems to be possible
379 given system limits on open files, select() bit masks, and so forth.
380 Start ddcccciiffdd with --dd and see the starting message in the system log
381 to see the limit.
382
383 --BB _d_n_s_b_l_-_o_p_t_i_o_n
384 enables DNS blacklist checks of the SMTP client IP address, SMTP
385 envelope Mail_From sender domain name, and of host names in URLs in
386 the message body. Body URL blacklisting has too many false posi-
387 tives to use on abuse mailboxes. It is less effective than
388 greylisting with dccm(8) or dccifd(8) but can be useful in situa-
389 tions where greylisting cannot be used.
390
391 _D_n_s_b_l_-_o_p_t_i_o_n is either one of the --BB _s_e_t_:_o_p_t_i_o_n forms or
392 --BB _d_o_m_a_i_n[_,_I_P_a_d_d_r[_/_x_x[_,_b_l_t_y_p_e]]]
393 _D_o_m_a_i_n is a DNS blacklist domain such as example.com that will be
394 searched. _I_P_a_d_d_r[_/_x_x_x] is the string "any" an IP address in the DNS
395 blacklist that indicates that the mail message should be rejected,
396 or a CIDR block covering results from the DNS blacklist.
397 "127.0.0.2" is assumed if _I_P_a_d_d_r is absent. IPv6 addresses can be
398 specified with the usual colon (:) notation. Names can be used
399 instead of numeric addresses. The type of DNS blacklist is speci-
400 fied by _b_l_t_y_p_e as _n_a_m_e, _I_P_v_4, or _I_P_v_6. Given an envelope sender
401 domain name or a domain name in a URL of spam.domain.org and a
402 blacklist of type _n_a_m_e, spam.domain.org.example.com will be tried.
403 Blacklist types of _I_P_v_4 and _I_P_v_6 require that the domain name in a
404 URL sender address be resolved into an IPv4 or IPv6 address. The
405 address is then written as a reversed string of decimal octets to
406 check the DNS blacklist, as in 2.0.0.127.example.com,
407
408 More than one blacklist can be specified and blacklists can be
409 grouped. All searching within a group is stopped at the first posi-
410 tive result.
411
412 Positive results are ignored after being logged unless an
413 _o_p_t_i_o_n _D_N_S_B_L_-_o_n line appears in the global or per-user _w_h_i_t_e_c_l_n_t
414 file.
415
416 --BB _s_e_t_:_n_o_-_c_l_i_e_n_t
417 says that SMTP client IP addresses and reverse DNS domain names
418 should not be checked in the following blacklists.
419 --BB _s_e_t_:_c_l_i_e_n_t restores the default for the following black-
420 lists.
421
422 --BB _s_e_t_:_n_o_-_m_a_i_l___h_o_s_t
423 says that SMTP envelope Mail_From sender domain names should
424 not be checked in the following blacklists. --BB _s_e_t_:_m_a_i_l___h_o_s_t
425 restores the default.
426
427 --BB _s_e_t_:_n_o_-_U_R_L
428 says that URLs in the message body should not be checked in the
429 in the following blacklists. --BB _s_e_t_:_U_R_L restores the default.
430
431 --BB _s_e_t_:_n_o_-_M_X
432 says MX servers of sender Mail_From domain names and host names
433 in URLs should not be checked in the following blacklists.
434 --BB _s_e_t_:_M_X restores the default.
435
436 --BB _s_e_t_:_n_o_-_N_S
437 says DNS servers of sender Mail_From domain names and host
438 names in URLs should not be checked in the following black-
439 lists. --BB _s_e_t_:_N_S restores the default.
440
441 --BB _s_e_t_:_d_e_f_a_u_l_t_s
442 is equivalent to all of --BB _s_e_t_:_n_o_-_t_e_m_p_-_f_a_i_l --BB _s_e_t_:_c_l_i_e_n_t
443 --BB _s_e_t_:_m_a_i_l___h_o_s_t --BB _s_e_t_:_U_R_L --BB _s_e_t_:_M_X and --BB _s_e_t_:_N_S
444
445 --BB _s_e_t_:_g_r_o_u_p_=_X
446 adds later DNS blacklists specified with
447 --BB _d_o_m_a_i_n[_,_I_P_a_d_d_r[_/_x_x[_,_b_l_t_y_p_e]]]
448 to group 1, 2, or 3.
449
450 --BB _s_e_t_:_d_e_b_u_g_=_X
451 sets the DNS blacklist logging level
452
453 --BB _s_e_t_:_m_s_g_-_s_e_c_s_=_S
454 limits ddcccciiffdd to _S seconds total for checking all DNS black-
455 lists. The default is 25.
456
457 --BB _s_e_t_:_U_R_L_-_s_e_c_s_=_S
458 limits ddcccciiffdd to at most _S seconds resolving and checking any
459 single URL. The default is 11. Some spam contains dozens of
460 URLs and that some "spamvertised" URLs contain host names that
461 need minutes to resolve. Busy mail systems cannot afford to
462 spend minutes checking each incoming mail message.
463
464 --BB _s_e_t_:_r_e_j_-_m_s_g_=_r_e_j_e_c_t_i_o_n_-_m_s_g
465 sets the SMTP rejection message for the following blacklists.
466 _R_e_j_e_c_t_i_o_n_-_m_s_g must be in the same format as for --rr. If
467 _r_e_j_e_c_t_i_o_n_-_m_s_g is null, the default is restored. The default
468 DNS blacklist rejection message is the first message set with
469 --rr.
470
471 --BB _s_e_t_:_t_e_m_p_-_f_a_i_l
472 causes ddcccciiffdd to the MTA to answer the SMTP DATA command with
473 452 4.2.1 mail %ID from %CIP temporary delayed for DNSBL
474 if any DNS answer required for a DNSBL in the current group
475 times out, including resolving names in URLs.
476
477 --BB _s_e_t_:_n_o_-_t_e_m_p_-_f_a_i_l
478 restores the default of assuming a negative answer for DNS
479 responses that take too long.
480
481 --BB _s_e_t_:_m_a_x_j_o_b_s_=_X
482 sets maximum number of helper processes to _X. In order to use
483 typical single-threaded DNS resolver libraries, ddcccciiffdd uses
484 fleets of helper processes. It is rarely a good idea to change
485 the default, which is the same as the maximum number of simul-
486 taneous jobs set with --jj.
487
488 --BB _s_e_t_:_p_r_o_g_p_a_t_h_=_/_v_a_r_/_d_c_c_/_l_i_b_e_x_e_c_/_d_n_s_-_h_e_l_p_e_r
489 changes the path to the helper program.
490
491 --LL _l_t_y_p_e_,_f_a_c_i_l_i_t_y_._l_e_v_e_l
492 specifies how messages should be logged. _L_t_y_p_e must be _e_r_r_o_r, _i_n_f_o,
493 or _o_f_f to indicate which of the two types of messages are being con-
494 trolled or to turn off all syslog(3) messages from ddcccciiffdd. _L_e_v_e_l
495 must be a syslog(3) level among _E_M_E_R_G, _A_L_E_R_T, _C_R_I_T, _E_R_R, _W_A_R_N_I_N_G,
496 _N_O_T_I_C_E, _I_N_F_O, and _D_E_B_U_G. _F_a_c_i_l_i_t_y must be among _A_U_T_H, _A_U_T_H_P_R_I_V,
497 _C_R_O_N, _D_A_E_M_O_N, _F_T_P, _K_E_R_N, _L_P_R, _M_A_I_L, _N_E_W_S, _U_S_E_R, _U_U_C_P, and _L_O_C_A_L_0
498 through _L_O_C_A_L_7. The default is equivalent to
499 --LL _i_n_f_o_,_M_A_I_L_._N_O_T_I_C_E --LL _e_r_r_o_r_,_M_A_I_L_._E_R_R
500
501 ddcccciiffdd normally sends counts of mail rejected and so forth to the system
502 log at midnight. The SIGUSR1 signal sends an immediate report to the
503 system log. The reports will be repeated every 24 hours at the same
504 minute as the signal instead of at midnight.
505
506 PPrroottooccooll
507 DDcccciiffdd uses a simple ASCII protocol to receive mail messages to be
508 checked and to return results. For each message, the MTA must open a
509 connection to the interface daemon, send options, envelope recipients,
510 and the message, receive the results, and close the connection.
511
512 Instead of the ASCII protocol, a subset of ESMTP is enabled by --oo. Only
513 the familiar HELO, EHLO, Mail, Rcpt, DATA, RSET, and QUIT commands and
514 the Postfix extensions XFORWARD and XCLIENT are honored. Since SMTP has
515 no provisions for user names, the protocol enabled by --oo depends on a
516 list of local domain names specified with --DD to find per-user log direc-
517 tories and whitelist files. If neither XFORWARD nor XCLIENT are used,
518 ddcccciiffdd uses the IP address of the MTA and the value of the HELO command.
519
520 In the ASCII protocol, each of the following lines are sent in order to
521 ddcccciiffdd. Each ends with a newline ('\n') character.
522 options zero or more blank-separated strings among:
523 _s_p_a_m the message is already known to be spam
524 _b_o_d_y return all of the headers with the added
525 _X_-_D_C_C header line and the body
526 _h_e_a_d_e_r return the _X_-_D_C_C header
527 _q_u_e_r_y ask the DCC server about the message without
528 reporting it, as if ddcccciiffdd were running with
529 --QQ.
530 _g_r_e_y_-_q_u_e_r_y only query the greylist server for this mes-
531 sage. --GG _o_n must be in use.
532 _n_o_-_r_e_j_e_c_t suppress the overall, one character line 'R'
533 result. This can be useful when using ddcccciiffdd
534 only for greylisting.
535 _l_o_g ensure that this message is logged as if
536 ddcccciiffdd were running with --tt --aallll,,00,,
537 client IP address of the SMTP client in a "dotted" or "coloned"
538 ASCII string and reverse-DNS host name. If the host name
539 is present, it must follow a carriage return character
540 ('\r') after the IP address. The client IP address must be
541 present and non-null if the host name is present. The
542 string "0.0.0.0\n" is understood the same as the null
543 string, meaning that both the IP address and host name are
544 absent. If the client IP address is absent, then the IP
545 address and host name are taken from the first non-local
546 Received header if it has the standard "name (name [IP
547 address])..." format. Non-standard Received headers com-
548 monly added by qmail as well as Received headers specifying
549 IP addresses marked _M_X or _M_X_D_C_C in the global --ww _w_h_i_t_e_c_l_n_t
550 file are skipped.
551 HELO SMTP HELO value or nothing, followed by a newline ('\n')
552 character. If the HELO value is null and the IP address of
553 the SMTP client are not supplied, they will be taken from
554 the same Received: header that supplies the IP address.
555 sender or SMTP _M_a_i_l _F_r_o_m command value for the env_from checksum.
556 If the sender is null, the contents of the first Return-
557 Path: or UNIX style From_ header is used.
558 recipients or SMTP _R_c_p_t _T_o recipient mailboxes followed by correspond-
559 ing local user names, one (mailbox,user) pair to a line.
560 Each optional local user name is separated from the corre-
561 sponding mailbox recipient address by a carriage return
562 ('\r'). A local user name can be null if it is not known,
563 but each recipient mailbox must be non-null. If there are
564 no lines of (mailbox,user) pairs and if the _s_p_a_m option is
565 not included, then the _q_u_e_r_y is assumed. Mailboxes without
566 user names will lack per-user log files and will not invoke
567 a per-user whitelist.
568
569 The last recipient-user name pair is followed by an empty line and the
570 headers and body of the message. The end of the body of the mail message
571 is signaled by the MTA half-closing the connection. See shutdown(2).
572
573 DDcccciiffdd responds with three things. First is a one character line of the
574 overall result advising the MTA:
575 A accept the message for all recipients and answer the SMTP DATA
576 command with a 2yz result.
577 G answer with a 4yz result to embargo the message for greylisting.
578 R reject the message and answer the DATA command with a 5yz result.
579 S accept the message for some recipients and so answer the DATA com-
580 mand with a 2yz result.
581 T temporary failure by the DCC system and so answer with a 4yz
582 result.
583
584 Second is a line of characters indicating the disposition of the message
585 for each corresponding recipient:
586 A deliver the message
587 G discard the message during a greylist embargo
588 R discard the message as spam
589 The SMTP protocol allows only a single result for the DATA command for
590 all recipients that were not rejected before body of the message was
591 offered with the DATA command. To accept the message for some recipients
592 and reject it for others, the MTA must tell the SMTP client it is accept-
593 ing the message for all recipients and then discard it for those that
594 would reject it.
595
596 Finally, if the _b_o_d_y or _h_e_a_d_e_r strings are in the first line of _o_p_t_i_o_n_s
597 sent by the MTA to the daemon, then the _X_-_D_C_C header line or the entire
598 body with the _X_-_D_C_C header line follows.
599
600 FFIILLEESS
601 /var/dcc is the DCC home directory in which other files are found.
602 /var/dcc/libexec/start-dccifd
603 and
604 /var/dcc/libexec/rcDCC
605 are scripts used to start the daemon.
606 dcc/dcc_conf
607 contains parameters used by the scripts to start DCC daemons
608 and cron jobs.
609 logdir is an optional directory specified with --ll and containing
610 marked mail. Each file in the directory contains one mes-
611 sage, at least one of whose checksums reached its --tt thresh-
612 olds or that is interesting for some other reason. Each file
613 starts with lines containing the date when the message was
614 received, the IP address of the SMTP client, and SMTP enve-
615 lope values. Those lines are followed by the body of the
616 SMTP message including its header as it was received. Only
617 approximately the first 32 KBytes of the body are recorded
618 unless modified by _._/_c_o_n_f_i_g_u_r_e _-_-_w_i_t_h_-_m_a_x_-_l_o_g_-_s_i_z_e_=_x_x The
619 checksums for the message follow the body. They are followed
620 by lines indicate that one of the checksums is white- or
621 blacklisted by the --ww _w_h_i_t_e_c_l_n_t file. Each log file ends
622 with the _X_-_D_C_C header line added to the message and the dis-
623 position of the message.
624 map is the memory mapped file of information concerning DCC
625 servers in the DCC home directory.
626 whiteclnt contains the client whitelist in the format described in
627 dcc(8).
628 whiteclnt.dccw
629 is a memory mapped hash table of the _w_h_i_t_e_c_l_n_t file.
630 dccifd.pid in the --RR _r_u_n_d_i_r directory contains daemon's process ID.
631
632 EEXXAAMMPPLLEESS
633 Dccifd can be used as Postfix Before-Queue Content filter. In some tests
634 these values for --pp and --oo in _d_c_c___c_o_n_f.
635
636 DCCIFD_ENABLE=on
637 DCCIFD_ARGS="-p 127.0.0.1,10025,127.0.0.1/32 -o 127.0.0.1,10026
638
639 worked with these lines in /etc/postfix/master.cf
640
641 smtp inet n - n - - smtpd
642 -o smtpd_proxy_filter=127.0.0.1:10025
643 127.0.0.1:10026 inet n - n - - smtpd
644 -o smtpd_authorized_xforward_hosts=127.0.0.0/8
645 -o smtpd_client_restrictions=
646 -o smtpd_helo_restrictions=
647 -o smtpd_sender_restrictions=
648 -o smtpd_recipient_restrictions=permit_mynetworks,reject
649 -o smtpd_data_restrictions=
650 -o mynetworks=127.0.0.0/8
651 -o receive_override_options=no_unknown_recipient_checks
652
653 SSEEEE AALLSSOO
654 cdcc(8), dbclean(8), dcc(8), dccd(8), dblist(8), dccm(8), dccproc(8),
655 dccsight(8),
656
657 HHIISSTTOORRYY
658 Implementation of ddcccciiffdd Distributed Checksum Clearinghouses are based on
659 an idea of Paul Vixie with code designed and written at Rhyolite Software
660 starting in 2000. was started at Rhyolite Software in 2002. This docu-
661 ment describes version 1.3.103.
662
663 BBUUGGSS
664 ddcccciiffdd uses --tt where dccproc(8) uses --cc.
665
666 By default ddcccciiffdd look for its UNIX domain socket in the DCC home direc-
667 tory, but dccm(8) looks in its --RR _r_u_n_d_i_r.
668
669 Systems without setrlimit(2) and getrlimit(2) RLIMIT_NOFILE can have
670 problems with the default limit on the number of simultaneous jobs, the
671 value of --jj. Every job requires four open files. These problems are
672 usually seen with errors messages that say something like
673 dccifd[24448]: DCC: accept(): Result too large
674 A fix is to use a smaller value for --jj or to allow ddcccciiffdd to open more
675 files.
676
677 February 26, 2009