.\" --8-LICENSE--
.\"
.\" $Revision: 1.178 $
.\"
.Dd March 22, 2024
.ds volume-ds-DCC Distributed Checksum Clearinghouse
.Dt dccm 8 DCC
.Os " "
.Sh NAME
.Nm dccm
.Nd Distributed Checksum Clearinghouse Milter Interface
.Sh SYNOPSIS
.Bk -words
.Nm
.Op Fl VdbxANPQ
.Op Fl G Ar on | off | noIP | IPmask/xx
.Op Fl h Ar homedir
.Op Fl I Ar user
.br
.Op Fl p Ar protocol:filename | protocol:port@host
.Op Fl m Ar map
.br
.Op Fl w Ar whiteclnt
.Op Fl U Ar userdirs
.Op Fl a Ar IGNORE | REJECT | DISCARD
.br
.Oo
.Fl t Xo
.Sm off
.Ar type,
.Op Ar log-thold,
.Ar rej-thold
.Sm on
.Xc
.Oc
.Oo
.Fl g Xo
.Sm off
.Op Ar not-
.Ar type
.Sm on
.Xc
.Oc
.Op Fl S Ar header
.br
.Op Fl l Ar logdir
.Op Fl R Ar rundir
.Op Fl r Ar rejection-msg
.Op Fl j Ar maxjobs
.Op Fl B Ar dnsbl-option
.Op Fl L Ar ltype,facility.level
.Ek
.Sh DESCRIPTION
.Nm
is a daemon built with the sendmail milter interface intended to connect
.Xr sendmail 8
to DCC servers.
When built with the milter filter machinery and configured to talk to
.Nm
in the
.Pa sendmail.cf
file,
sendmail passes all email to
.Nm
which in turn reports related checksums to the nearest DCC server.
.Nm
then adds an
.Em X-DCC
SMTP header line to the message.
Sendmail is told to reject the message if it is unsolicited bulk mail.
.Pp
.Nm Dccm
sends reports of checksums related to mail received by DCC clients
and queries about the total number of reports of particular checksums.
A DCC server receives
.Em no
mail, address, headers, or other information,
but only cryptographically secure checksums of such information.
A DCC server cannot determine the text or other information that corresponds
to the checksums it receives.
Its only acts as a clearinghouse of counts for checksums computed by clients.
For complete privacy as far as the DCC is concerned,
the checksums of purely internal mail or other
mail that is known to not be unsolicited bulk can be listed in a whitelist
to not be reported to the DCC server.
.Pp
Since the checksums of messages that are whitelisted locally
by the
.Pa @prefix@/whiteclnt
file are not reported to the DCC server,
.Nm
knows nothing about the total recipient counts for their checksums and
so cannot add
.Em X-DCC
header lines to such messages.
Sendmail does not tell
.Nm
about messages that are not received by sendmail via SMTP, including messages
submitted locally and received via UUCP, and so they also do not receive
.Em X-DCC
header lines.
.Pp
Enable the daemon and put its parameters in the
.Pa @prefix@/conf
file and start the daemon with the
.Pa @libexecdir@/start-dccm
or
.Pa @libexecdir@/rcDCC
scripts.
.Pp
The list of servers that
.Nm
contacts is in the memory mapped file
.Pa @prefix@/map
shared by local DCC clients.
The file is  maintained with
.Xr cdcc 8 .
.Ss OPTIONS
The following options are available:
.Bl -tag -width 3n
.It Fl V
displays the version of
.Nm .
Two or more
.Fl V
options show the options with which it was built.
.It Fl d
enables debugging output from the DCC client software.
Additional
.Fl d
options increase the number of messages.
A single
.Fl d
logs aborted SMTP transactions including those from some "dictionary attacks."
.It Fl b
causes the daemon to not detach itself from the controlling tty
and put itself into the background.
.It Fl x
causes the daemon to try "extra hard" to contact a DCC server.
Since it is usually more important to deliver mail than to report its
checksums,
.Nm
normally does not delay too long while trying to contact a DCC server.
It will not try again for several seconds after a failure.
With
.Fl x ,
it will always try to contact the DCC server
and it will tell the MTA to answer the DATA command with a 4yz
temporary failure.
.It Fl A
adds to existing X-DCC headers in the message
instead of replacing existing headers
of the brand of the current server.
.It Fl N
neither adds, deletes, nor replaces existing X-DCC headers in the message.
Each mail message is logged, rejected, and otherwise handled the same.
.It Fl P
The SpamAsassin DCC.pm plugin should watch for "bulk" in X-DCC SMTP header
fields, but historically has looked for counts of "many".
However, there are situations when
.Nm
knows that a mail message is extremely bulky and probably spam.
For example, mail from a sender that is blacklisted in
.Pa whiteclnt
gets an X-DCC header that includes
.Em bulk .
To acommodate that bug in SpamAssassin, by default whenever
.Nm
generates an X-DCC header containing "bulk",
it also forces the
.Em Body
count to "many".
.Fl P
turns off that kludge and the
.Em Body
contains the count from the DCC server.
.It Fl Q
only queries the DCC server about the checksums of messages
instead of reporting.
This is useful when
.Nm
is used to filter mail that has already been reported to a DCC
server by another DCC client.
No single mail message should be reported to a DCC
server more than once per recipient,
because each report will increase the apparent "bulkness" of the message.
.Pp
It is better to use
.Em MXDCC
lines in the global
.Pa @prefix@/whiteclnt
file for your MX mail servers that use DCC than to use
.Fl Q
with
.Nm .
.Pp
Do not use
.Fl Q
except on mail that you know has been reported to a DCC server.
DCC depends on reports of all except known private mail and
works only because almost no DCC installations use
.Fl Q .
.It Fl G Ar on | off | noIP | IPmask/xx
controls
.Em greylisting .
At least one working greylist server must be listed in the
.Pa @prefix@/map
file.
If more than one is named,
they must "flood" or change checksums and they must use the
same
.Fl G
parameters.
See
.Xr dccd 8 .
Usually all dccm or dccifd DCC client processes use the same
.Fl G
parameters.
.Pp
.Ar IPmask/xx
and
.Ar noIP
remove part or all of the IP address from the greylist triple.
.It Fl h Ar homedir
overrides the default DCC home directory,
.Pa @prefix@ .
.It Fl I Ar user
specifies the UID and GID of the process.
.It Fl p Ar protocol:filename | protocol:port@host
specifies the protocol and address by which sendmail will contact
.Nm dccm .
The default is a UNIX domain socket in the "run" directory,
.Pa @dcc_rundir@/dccm .
(See also
.Fl R)
This protocol and address must match the value in
.Pa sendmail.cf .
This mechanism can be used to connect
.Nm
on one computer to sendmail on another computer
when a port and host name or IP address are used.
.It Fl m Ar map
specifies a name or path of the memory mapped parameter file instead
of the default
.Pa @prefix@/map .
It should be created with the
.Xr cdcc 8
command.
.It Fl w Ar whiteclnt
specifies an optional file containing filtering parameters
as well as SMTP client IP addresses,
SMTP envelope values, and header values
of mail that is spam or is not spam and does not need a
.Em X-DCC
header,
and whose checksums should not be reported to the DCC server.
.Pp
If the pathname
.Ar whiteclnt
is not absolute, it is relative to the DCC home directory.
.Pp
The format of the
.Nm
whiteclnt file is the same as the
.Pa @prefix@/whitelist
files used by
.Xr dbclean 8
and the
.Pa whiteclnt
file used by
.Xr dccproc 8 .
See
.Xr dcc 8
for a description of DCC white and blacklists.
Because the contents of the
.Ar whiteclnt
file are used frequently, a companion file is automatically
created and maintained.
It has the same pathname but with an added suffix of
.Ar .dccw
and contains a memory mapped hash table of the main file.
.Pp
A whitelist entry ("OK") or two or more semi-whitelistings ("OK2")
for one of the message's checksums prevents all of
the message's checksums from being reported to the DCC server
and the addition of a
.Em X-DCC
header line by
.Nm .
A whitelist entry for a checksum
also prevents rejecting or discarding the message based on DCC recipient
counts as specified by
.Fl a
and
.Fl t .
Otherwise, one or more checksums with blacklisting entries ("MANY") cause
all of the message's
checksums to be reported to the server with an addressee count of "MANY".
.Pp
If the message has a single recipient, an
.Ar env_To
.Ar whiteclnt
entry of "OK" for the checksum of its recipient address acts like any other
.Ar whiteclnt
entry of "OK."
When the SMTP message has more than one recipient,
the effects can be complicated.
When a message has several recipients with some but not all listed in the
.Ar whiteclnt
file,
.Nm
tries comply with the wishes of the users who want filtering as
well as those who don't by silently not delivering the message to 
those who want filtering (i.e. are not whitelisted) and delivering
the message to users who don't want filtering.
.It Fl U Ar userdirs
enables per-user
.Pa whiteclnt
files and log directories.
Each target of a message can have a directory of log files named
.Ar usedirs/${dcc_userdir}/log
where
.Em ${dcc_userdir}
is the
.Pa sendmail.cf
macro described below.
If
.Em ${dcc_userdir}
is not set,
.Ar userdirs/${rcpt_mailer}/${rcpt_addr}/log
is used.
The most likely value of
.Ar mailer
is
.Ar local .
Appropriate values for both
.Ar ${rcpt_mailer}
and
.Ar ${rcpt_addr}
can be seen by examining
.Em env_To
lines in
.Fl l Ar logdir
files.
If it is not absolute,
.Ar userdirs
is relative to the DCC home directory.
The directory containing the log files must be named
.Ar log
and it must be writable by the
.Nm
process.
Each log directory must exist or logging for the corresponding
is silently disabled.
The files created in the log directory are owned by the UID of the
.Nm
process,
but they have
.Em group
and
.Em other
read and write permissions copied from the corresponding
.Ar log
directory.
To ensure the privacy of mail,
it may be good to make the directories readable only by
.Em owner
and
.Em group ,
and to use a
.Xr cron
script that changes the owner of each file to match the grandparent
.Ar addr
directory.
.Pp
There can also be a per-user whitelist file named
.Ar userdirs/addr/whiteclnt
for each addressee
.Ar addr.
Any checksum that is not white- or blacklisted by an individual
addressee's per-user
.Pa whiteclnt
file  is checked in the main
.Pa @prefix@/whiteclnt
file.
A missing per-addressee
.Ar whiteclnt
file is the same as an empty file.
Relative paths for files included in per-addressee files
are resolved in the DCC home directory.
The
.Ar whiteclnt
files and the
.Ar addr
directories containing them must be writable by the
.Nm
process.
.Pp
.Ar Option
lines in per-user whiteclnt files can be used to modify many aspects of
.Nm
filtering,
as described in the main
.Xr dcc
man page.
For example, an
.Ar option dcc-off
line turns off DCC filtering for individual mailboxes.
.It Fl a Ar IGNORE | REJECT | DISCARD
specifies the action taken when
DCC server counts or
.Fl t
thresholds say that a message is unsolicited and bulk.
.Ar IGNORE
causes the message to be unaffected except for adding the
.Em X-DCC
header line to the message.
This turns off all filtering except greylisting.
.Pp
Spam can also be
.Ar REJECT Ns ed
or accepted and silently
.Ar DISCARD Ns ed
without being delivered to local mailboxes.
The default is
.Ar REJECT .
.Pp
Mail forwarded via IP addresses marked
.Em MX
or
.Em MXDCC
in the main
.Pa @prefix@/whiteclnt
file is treated
as if
.Fl a Ar DISCARD
were specified.
This prevents "bouncing" spam.
.Pp
Determinations that mail is or is not spam from sendmail via
.Em ${dcc_isspam}
or
.Em ${dcc_notspam}
macros override
.Fl a .
The effects of the
.Fl w Ar whiteclnt
are not affected by
.Fl a .
.It Fl t Xo
.Sm off
.Ar type,
.Op Ar log-thold,
.Ar rej-thold
.Sm on
.Xc
sets logging and "spam" thresholds for checksum
.Ar type .
The checksum types are
.Ar IP ,
.Ar env_From ,
.Ar From ,
.Ar Message-ID ,
.Ar substitute ,
.Ar Received ,
.Ar Body ,
.Ar Fuz1 ,
.Ar Fuz2 ,
.Ar rep-total ,
and
.Ar rep .
The first six,
.Ar IP
through
.Ar substitute ,
have no effect except when a local DCC server configured with
.Fl K
is used.
The
.Ar substitute
thresholds apply to the first substitute heading encountered in the mail
message.
The string
.Ar ALL
sets thresholds for all types, but is unlikely to be useful except for
setting logging thresholds.
The string
.Ar CMN
specifies the commonly used checksums
.Ar Body ,
.Ar Fuz1 ,
and
.Ar Fuz2 .
.Ar Rej-thold
and
.Ar log-thold
must be numbers, the string
.Ar NEVER ,
or the string
.Ar MANY
indicating millions of targets.
Counts from the DCC server as large as the threshold for any single type
are taken as sufficient evidence
that the message should be logged or rejected.
.Pp
.Ar Log-thold
is the threshold at which messages are logged.
It can be handy to log messages at a lower threshold to find
solicited bulk mail sources such as mailing lists.
If no logging threshold is set,
only rejected mail and messages with complicated combinations of white
and blacklisting are logged.
Messages that reach at least one of their rejection thresholds are
logged regardless of logging thresholds.
.Pp
.Ar Rej-thold
is the threshold at which messages are considered "bulk,"
and so should be rejected or discarded if not whitelisted.
.Pp
DCC Reputation thresholds in the commercial version
of DCC are controlled by thresholds on checksum types
.Ar rep
and
.Ar rep-total .
The DCC Reputations of IP addresses that the DCC database says have sent
more than
.Ar rep-total,log-thold
are computed and messages from those addresses are logged.
Messages from IP addresses with DCC Reputations of at least the
.Ar rep,rej-thold
rejection threshold can be rejected.
The DCC Reputation of an IP address is the percentage of its messages
known to have been sent to at least 10 recipients.
The defaults are equivalent to
.Ar rep,never
and
.Ar rep-total,never,20 .
.Pp
Bulk DCC Reputations do not reject mail unless enabled by an
.Ar option\ DCC-rep-on
line a
.Pa whiteclnt
file.
.Pp
The checksums of locally whitelisted messages are not checked with
the DCC server and so only the number of targets of the current copy of
a whitelisted message are compared against the thresholds.
.Pp
The default is
.Ar ALL,NEVER ,
so that nothing is discarded, rejected, or logged.
A common choice is
.Ar CMN,25,50
to reject or discard
mail with common bodies except as overridden by
the whitelist of the DCC server, the sendmail
.Em ${dcc_isspam}
and
.Em ${dcc_notspam}
macros, and
.Fl g ,
and
.Fl w .
.It Fl g Xo
.Sm off
.Op Ar not-
.Ar type
.Sm on
.Xc
indicates that whitelisted,
.Ar OK
or
.Ar OK2 ,
counts from the DCC server for a type of checksum are to be believed.
They should be ignored if prefixed with
.Ar not- .
.Ar Type
is one of the same set of strings as for
.Fl t .
Only
.Ar IP ,
.Ar env_From ,
and
.Ar From
are likely choices.
By default all three are honored,
and hence the need for
.Ar not- .
.It Fl S Ar hdr
adds to the list of substitute or locally chosen headers that
are checked with the
.Fl w Ar whiteclnt
file and sent to the DCC server.
The checksum of the last header of type
.Ar hdr
found in the message is checked.
.Ar Hdr
can be
.Em HELO
to specify the SMTP envelope HELO value.
.Ar Hdr
can also be
.Em mail_host
to specify the host name from
the Mail_from value in the SMTP envelope.
As many as 8 different substitute headers can be specified, but only
the checksum of the first will be sent to the DCC server.
.It Fl l Ar logdir
specifies a directory in which files containing copies of messages processed by
.Nm
are kept.
They can be copied to per-user directories specified with
.Fl U .
Information about other recipients of a message is deleted from
the per-user copies.
.Pp
See the FILES section below concerning the contents of the files.
See also the
.Ar option log-subdirectory-{day,hour,minute}
lines in
.Pa whiteclnt
files described in
.Xr dcc 8 .
.Pp
The directory is relative to the DCC home directory if it is not absolute
.It Fl R Ar rundir
specifies the "run" directory where the UNIX domain socket and file
containing the daemon's process ID are stored.
The default value is @dcc_rundir@ .
.It Fl r Ar rejection-msg
specifies the rejection message
for unsolicited bulk mail or for mail temporarily blocked by
.Em greylisting
when
.Fl G
is specified.
The first
.Fl r Ar rejection-msg
replaces the default bulk mail rejection message,
.Bk -words
"5.7.1 550 mail %ID from %CIP rejected by DCC".
.Ek
The second replaces
.Bk -words
"4.2.1 452 mail %ID from %CIP temporary greylist embargoed".
.Ek
The third
.Fl r Ar rejection-msg
replaces the default SMTP rejection message
.Bk -words
"5.7.1 550 %ID bulk mail reputation; see https://commercial-dcc.rhyolite.com/cgi-bin/reps.cgi?tgt=%CIP"
.Ek
for mail with bulk DCC Reputations.
If
.Ar rejection-msg
is the zero-length string,
the
.Fl r
setting is counted but the corresponding default message is not changed.
.Pp
.Ar Rejection-msg
can contain specific information about the mail message.
The following strings starting with % are replaced with the corresponding
values:
.Bl -tag -width "%BRESULT" -offset 4n -compact
.It %ID
message ID such as the unique part of log file name or sendmail queue ID
.It %CIP
SMTP client IP address
.It %BTYPE
type of DNS blacklist hit, such as "SMTP client", "mail_host", or "URL NS"
.It %BTGT
IP address or name declared bad by DNS blacklist
.It %BPROBE
domain name found in DNS blacklist such as 4.3.2.10.example.com
.It %BRESULT
value of the %BPROBE domain name found in DNS blacklist
.El
.Pp
A common alternate for the bulk mail rejection message is
.Bk -words
"4.7.1 451 Access denied by DCC"
.Ek
to tell the sending mail system to continue trying.
Use a 4yz response with caution, because it is likely to delay for days
a delivery failure message for false positives.
If the rejection message
does not start with an RFC 1893 status code and RFC 2821 reply code,
5.7.1 and 550 or 4.2.1 and 452 are used.
.Pp
See also
.Fl B Ar set:rej-msg
to set the status message for mail rejected by DNS blacklists.
.It Fl j Ar maxjobs
limits the number of simultaneous requests that will be processed.
The default value is the maximum number that seems to be possible given system
limits on open files, select() bit masks, and so forth.
Start
.Nm
with
.Fl d
and see the starting message in the system log to see the limit.
.It Fl B Ar dnsbl-option
enables DNS white- and blacklist checks of the SMTP client IP address,
SMTP envelope Mail_From sender domain name, and of host names in
URLs in the message body.
Body URL blacklisting has too many false positives to use on
abuse mailboxes.
It is less effective than greylisting with
.Xr dccm 8
or
.Xr dccifd 8
but can be useful in situations where
greylisting cannot be used.
It can be combined with greylisting.
.Pp
.Ar Dnsbl-option
is either one of the
.Fl B Ar set:option
forms or
.Bd -literal -compact -offset 4n
.Fl B Xo
.Sm off
.Ar domain Oo any Op Ar ,bltype Oc
.Sm on
.Xc
.Fl B Xo
.Sm off
.Ar domain Oo Ar ,IPaddr
.Op Ar /xx  Oo Ar &IPmask Oc Op Ar ,bltype Oc
.Sm on
.Xc
.Fl B Xo
.Sm off
.Ar domain Oo Ar ,IPaddrLO
.Op Ar -IPaddrHI Oo Ar &IPmask Oc Op Ar ,bltype Oc
.Sm on
.Xc
.Ed
.Ar Domain
is a DNS blacklist domain such as
.Em example.com
that will be searched.
The strings
.Ar any ,
.Ar IPaddr ,
.Ar IPaddr/xx ,
or
.Ar IPaddrLO-IPaddrHI ,
specifies which IP addresses found in the DNS blacklist
after applying the optional IP address mask
.Ar IPmask
say that mail messages should be rejected or accepted with
.Fl B Ar set:white .
"127.0.0.2" is assumed if no address(es) are specified.
IPv6 addresses can be specified with the usual colon (:) notation.
Host names can be used instead of numeric addresses.
The type of DNS blacklist
is specified by
.Ar bltype
as
.Ar name ,
.Ar all-names ,
.Ar IPv4 ,
or
.Ar IPv6 .
Given an envelope sender domain name or a domain name in a URL of
spam.domain.org
and a blacklist of type
.Ar name ,
spam.domain.org.example.com will be looked up.
The names
spam.domain.org.example.com,
domain.org.example.com, and
org.example.com
will be looked up in blacklists of type
.Ar all-names .
Use
.Ar name
with DNS blacklists that use wildcards for speed
but
.Ar all-names
for other DNS name blacklists.
Blacklist types of
.Ar IPv4
and
.Ar IPv6
require that the domain name in a URL sender address
be resolved into an IPv4 or IPv6
address.
The resolved address from the mail message
is then written as a reversed string of decimal
octets to check the DNS blacklist, as in
.Em 2.0.0.127.example.com .
.Pp
A
.Ar domain
of "."
and type of
.Ar name
can be used to blacklist domain names with specified addresses.
This can be useful to detect URLs with domain names listed in
a Response Policy Zone (RPZ).
For example, the following can be used to reject mail
containing URLs listed by a response policy zone that maps
evil domain names to 224.0.0.0 with an informative status
message:
.Bd -literal
  '-Bset:rej-msg=5.7.1 550 %ID %BTYPE \\
	https://example.com/query/dbl?domain=%BTGT'
  -B.,224.0.0.0,name
.Ed
.Pp
More than one blacklist can be specified and blacklists can be grouped with
.Fl B Ar set:group=X .
All searching within a group of blacklists
is stopped at the first positive result.
.Pp
Unlike
.Xr dccproc 8 ,
positive results are ignored by 
.Nm
after being logged unless an
.Ar option\ DNSBL-on
or
.Ar option\ DNSBLx-on
line appears a
.Pa whiteclnt
file.
.Pp
.Bl -tag -width 3n
.It Fl B Ar set:no-client
implies that SMTP client IP addresses and reverse DNS domain names should
not be checked in the following blacklists.
.br
.Fl B Ar set:client
restores the default for the following blacklists.
.It Fl B Ar set:no-mail_host
implies that SMTP envelope Mail_From sender domain names should
not be checked in the following blacklists.
.Fl B Ar set:mail_host
restores the default.
.It Fl B Ar set:no-URL
says that URLs in the message body should not be checked in the
in the following blacklists.
.Fl B Ar set:URL
restores the default.
.It Fl B Ar set:no-MX
says MX servers of sender Mail_From domain names and host names in URLs
should not be checked in the following blacklists.
.br
.Fl B Ar set:MX
restores the default.
.It Fl B Ar set:no-NS
says DNS servers of sender Mail_From domain names and host names in URLs
should not be checked in the following blacklists.
.Fl B Ar set:NS
restores the default.
.It Fl B Ar set:white
says the DNS list is a whitelist of names or IP addresses.
.Bk
.Fl B Ar set:black
.Ek
restores the default.
DNS whitelist usually also need
.Bk
.Fl B Ar set:no-mail_host ,
.Fl B Ar set:no-URL ,
.Fl B Ar set:no-MX ,
.Fl B Ar set:no-NS ,
and
.Fl B Ar set:no-mail_host .
.Ek
.It Fl B Ar set:defaults
is equivalent to all of
.Bk
.Fl B Ar set:black
.Fl B Ar set:client
.Fl B Ar set:mail_host
.Fl B Ar set:URL
.Fl B Ar set:MX
and
.Fl B Ar set:NS
.Ek
.It Fl B Ar set:group=X
adds following DNS blacklists specified with
.Bk
.Fl B Xo
.Sm off
.Ar domain Op Ar ...
.Sm on
.Xc
.Ek
to group 1, 2, 3, or 4.
.It Fl B Ar set:debug=X
sets the DNS blacklist logging level
.It Fl B Ar set:msg-secs=S
limits
.Nm
to
.Ar S
seconds total for checking all DNS blacklists.
The default is 25.
.It Fl B Ar set:URL-secs=S
limits
.Nm
to at most
.Ar S
seconds resolving and checking any single URL or IP address.
The default is 11.
Some spam contains dozens of URLs and
some "spamvertised" URLs contain host names that need minutes to
resolve.
Busy mail systems cannot afford to spend minutes checking each incoming
mail message.
.It Fl B Ar set:rej-msg="rejection message"
sets the SMTP rejection message for the following blacklists.
.Ar Rejection-msg
must be in the same format as for
.Fl r .
If
.Ar rejection message
is null, the default is restored.
The default DNS blacklist rejection message is the first message set
with
.Fl r .
.It Fl B Ar set:max_helpers=X
sets maximum number of helper processes to
.Ar X .
In order to use typical single-threaded DNS resolver libraries,
.Nm
uses fleets of helper processes.
It is rarely a good idea to change the default,
which is the same as the maximum number of simultaneous jobs set with
.Fl j .
.It Fl B Ar set:progpath=@libexecdir@/dns-helper
changes the path to the helper program.
.El
.It Fl L Ar ltype,facility.level
specifies how messages should be logged.
.Ar Ltype
must be
.Ar error ,
.Ar info ,
or
.Ar off
to indicate which of the two types of messages are being controlled or
to turn off all
.Xr syslog 3
messages from
.Nm .
.Ar Level
must be a
.Xr syslog 3
level among
.Ar EMERG ,
.Ar ALERT ,
.Ar CRIT , ERR ,
.Ar WARNING ,
.Ar NOTICE ,
.Ar INFO ,
and
.Ar DEBUG .
.Ar Facility
must be among
.Ar AUTH ,
.Ar AUTHPRIV ,
.Ar CRON ,
.Ar DAEMON ,
.Ar FTP ,
.Ar KERN ,
.Ar LPR ,
.Ar MAIL ,
.Ar NEWS ,
.Ar USER ,
.Ar UUCP ,
and
.Ar LOCAL0
through
.Ar LOCAL7 .
The default is equivalent to
.Dl Fl L Ar info,MAIL.NOTICE  Fl L Ar error,MAIL.ERR
.El
.Pp
.Nm
normally sends counts of mail rejected and so forth the to system log at
midnight.
The SIGUSR1 signal sends an immediate report to the system log.
They will be repeated every 24 hours instead of at midnight.
.Sh SENDMAIL MACROS
Sendmail can affect
.Nm
with the values of some
.Pa sendmail.cf
macros.
These macro names must be added to the
Milter.macros option statements in
.Pa sendmail.cf
as in the example "Feature" file dcc.m4.
.Bl -tag -width dcc_mail_host
.It Em ${dcc_isspam}
causes a mail message to be reported to the DCC server
as having been addressed to "MANY" recipients.
The
.Em ${dcc_isspam}
macro is ignored if the
.Em ${dcc_notspam}
macro is set to a non-null string
.Pp
If the value of the
.Ar ${dcc_isspam}
is null,
.Nm
uses SMTP rejection messages controlled by
.Fl a
and
.Fl r .
If the value of the
.Ar ${dcc_isspam}
macro starts with "DISCARD",
the mail message is silently discarded
as with
.Fl a Ar DISCARD.
If value of the macro not null and does not start with "DISCARD",
it is used as the SMTP error
message given to the SMTP client trying to send the rejected message.
The message starts with an optional SMTP error type and number
followed by text.
.Pp
The
.Fl a
option does not effect messages
marked spam with
.Em ${dcc_isspam} .
When the
.Em ${dcc_isspam}
macro is set, the message is rejected or discarded despite
local or DCC database whitelist entries.
The local whitelist does control whether the message's
checksums will be reported to the DCC server and an
.Em X-DCC
SMTP header line will be added.
.It Em ${dcc_notspam}
causes a message not be considered unsolicited bulk despite
evidence to the contrary.
It also prevents
.Nm
from reporting the checksums of the message to the DCC server
and from adding an
.Em X-DCC
header line.
.Pp
When the macro is set by the
.Pa sendmail.cf
rules,
.Ar ${dcc_notspam}
macros overrides DCC threshlds that say the message should be
rejected as well as the effects of the
.Em ${dcc_isspam}
macro.
.It Em ${dcc_mail_host}
specifies the name of the SMTP client that is sending the message.
This macro is usually the same as the
.Em mail_host
macro.
They can differ when a sendmail "smart relay" is involved.
The
.Em ${dcc_mail_host}
macro does not work if
.Em FEATURE(delay_checks)
is used,
and so
.Nm
falls back on
.Em mail_host .
.It Em ${dcc_userdir}
is the per-user whitelist and log directory for a recipient.
If the macro is not set in sendmail.cf,
$&{rcpt_mailer}/$&{rcpt_addr}
is assumed, but with the recipient address converted to lower case.
Whatever value is used,
the directory name after the last slash (/) character is converted to
lower case.
Any value containing the string "/../" is ignored.
.Pp
This macro also does not work if
.Em FEATURE(delay_checks)
is used.
.Pp
The following two lines in a sendmail mc file have the same effect
as not defining the ${dcc_userdir} macro, provided
.Em FEATURE(dcc)
is also used and
the sendmail
.Pa cf/feature
directory has a symbolic link to the
.Pa @prefix@/build/dcc/misc/dcc.m4
file.
.El
.Pp
.Bd -literal -compact
SLocal_check_rcpt
R$*	$: $1 $(macro {dcc_userdir} $@ $&{rcpt_mailer}/$&{rcpt_addr} $))
.Ed
.Sh FILES
.Bl -tag -width whiteclnt -compact
.It Pa @prefix@
is the DCC home directory.
.It Pa @libexecdir@/start-dccm
is a script used by
.Pa @libexecdir@/rcDCC
to start
.Nm .
.It Pa dcc_conf
contains parameters used by the scripts to start DCC daemons and cron jobs.
.It Pa logdir
is an optional directory specified with
.Fl l
and containing marked mail.
Each file in the directory contains one message, at least one of whose
checksums reached its
.Fl t
thresholds or that is interesting for some other reason.
Each file starts with lines containing the date when the message
was received, the IP address of the SMTP client, and SMTP envelope
values.
Those lines are followed by the body of the SMTP message including its header
as it was received by sendmail and without any new or changed header lines.
Only approximately the first 32 KBytes of the body are recorded
unless modified by
.Em ./configure --with-max-log-size=xx
The checksums for the message follow the body.
They are followed by lines indicating that the
.Em ${dcc_isspam}
or
.Em ${dcc_notspam}
.Pa sendmail.cf
macros were set or one of the checksums is white- or blacklisted by the
.Fl w Ar whiteclnt
file.
Each file ends with the
.Em X-DCC
header line added to the message and the disposition of
the message including SMTP status message if appropriate.
.It Pa map
is the memory mapped file of information concerning DCC servers
in the DCC home directory.
See
.Fl m .
.It Pa whiteclnt
contains the client whitelist in the format described in
.Xr dcc 8 .
See
.Fl w .
.It Pa whiteclnt.dccw
is a memory mapped hash table of the
.Pa @prefix@/whiteclnt
file.
.It @dcc_rundir@/dccm.pid
directory contains daemon's process ID.
The string
.Dq dccm
is replaced by the file name containing the daemon to facilitate
running multiple daemons, probably connected to remote instances of
sendmail using TCP/IP instead of a UNIX domain socket.
See also
.Fl R .
.It Pa @dcc_rundir@/dccm
is the default UNIX domain socket used by the sendmail milter interface.
See also
.Fl R .
.It Pa sendmail.cf
is the
.Xr sendmail 8
control file.
.It Pa @prefix@/build/dcc/misc/dcc.m4
sendmail mc file that should have a symbolic link in the sendmail
cf/feature directory so that
.Em FEATURE(dcc)
can be used in a sendmail mc file.
.El
.Sh EXAMPLES
.Nm Dccm
should be started before sendmail with something like the
script
.Pa @libexecdir@/start-dccm.
It looks for common DCC parameters in the
.Pa @prefix@/dcc_conf
file.
.Pp
Those numbers should modified to fit local conditions.
It might be wise to replace the "100" numbers with much larger
values or with "MANY" until a few weeks of monitoring the log directory
show that sources of mailing lists are in the server's whitelist file
(see
.Xr dccd 8 )
or the local
.Pa @prefix@/whiteclnt
file.
.Pp
It is usually necessary to regularly delete old log files
with a script like @libexecdir@/cron-dccd.
.Pp
On systems unlike modern FreeBSD and other UNIX-like systems which
include sendmail milter support,
sendmail must be built with the milter interface, such as by creating a
.Pa devtools/Site/site.config.m4
or similar file containing something like the following lines:
.Bd -literal -offset indent
APPENDDEF(`conf_sendmail_ENVDEF', `-D_FFR_MILTER=1')
APPENDDEF(`conf_libmilter_ENVDEF', `-D_FFR_MILTER=1')
.Ed
.Pp
Appropriate lines invoking the milter interface must be added to
.Pa sendmail.cf.
That can be done by putting a symbolic link to the
the misc/dcc.m4 file in the DCC source to the sendmail cf/feature directory
and adding the line
.Pp
.Dl FEATURE(dcc)
.Pp
to the local .mc file.
.Pp
Note that
.Nm
should not be used with the Postfix milter mechanism.
Instead use
.Xr dccifd 8
as a before-queue filter as described in that man page.
.Sh SEE ALSO
.Xr cdcc 8 ,
.Xr dbclean 8 ,
.Xr dcc 8 ,
.Xr dccd 8 ,
.Xr dblist 8 ,
.Xr dccifd 8 ,
.Xr dccproc 8 ,
.Xr dccsight 8 ,
.Xr sendmail 8 .
.Sh HISTORY
Distributed Checksum Clearinghouses are based on an idea of Paul Vixie.
Implementation of
.Nm
was started at Rhyolite Software in 2000.
This document describes version RHYOLITE_VERSION.
.Sh BUGS
.Nm
uses
.Fl t
where
.Xr dccproc 8
uses
.Fl c .
.Pp
Systems without
.Xr setrlimit 2
and
.Xr getrlimit 2
RLIMIT_NOFILE
can have problems with the default limit on the number of simultaneous
jobs, the value of
.Fl j .
Every job requires four open files.
These problems are usually seen with errors messages that say something like
.Dl dccm[24448]: DCC: accept() returned invalid socket
A fix is to use a smaller value for
.Fl j
or to allow
.Nm
to open more files.
Sendmail version 8.13 and later can be told to poll() instead of select
with SM_CONF_POLL.
Some older versions of sendmail knew about FFR_USE_POLL.
One of the following lines in your devtools/Site/site.config.m4
file can help:
.Bd -literal -offset indent
APPENDDEF(`conf_libmilter_ENVDEF', `-DSM_CONF_POLL')
APPENDDEF(`conf_libmilter_ENVDEF', `-DFFR_USE_POLL')
.Ed
.Pp
On many systems with sendmail 8.11.3 and preceding,
a bug in the sendmail milter mechanism causes
.Nm
to die with a core file when given a signal.