summaryrefslogtreecommitdiff
path: root/usr.sbin/smtpd
AgeCommit message (Collapse)Author
2016-05-11mention ca option in listen directiveJoerg Jung
same spelling fix from jmc and gilles ok gilles
2016-05-08The "loop detected" message should not print the errno string.Todd C. Miller
OK gilles@
2016-05-03listen directive may use a table for authentication, to make this work theJoerg Jung
table has to be defined BEFORE consequently move all tables in the examples to the beginning and before the listen directive to avoid tables not being found no functional change ran into this myself earlier, also reported by cjones via irc ok gilles
2016-04-29fix "smtpctl show queue" reporting "invalid" envelope state.Eric Faurot
runtime state is not serialized with the envelope, so add it to the imsg. ok gilles@
2016-04-21Use automatic DH parameters, instead of fixed ones. Also disable DHE byJoel Sing
default since it is computationally expensive and a potential DoS vector. ok gilles@
2016-04-18add -T to SYNOPSIS;Jason McIntyre
2016-04-18document the -T flagJoerg Jung
original diff from Bernard Spil ok millert
2016-04-17fix typo, it is supposed to be smtpctl trace "filters" not "filter"Joerg Jung
from Boudewijn Dijkstra
2016-04-10arguments to "chain" are space separated, not comma; verified by jungJason McIntyre
diff from david+bsd i also removed Op, since ... is enough/
2016-03-30for some time now mandoc has not required MLINKS to functionJason McIntyre
correctly - logically complete that now by removing MLINKS from base; authors need only to ensure there is an entry in NAME for any function/ util being added. MLINKS will still work, and remain for perl to ease upgrades; ok nicm (curses) bcook (ssl) ok schwarze, who provided a lot of feedback and assistance ok tb natano jung
2016-03-25Nuke session_socket_blockmode() and session_socket_linger(). UseKenneth R Westerback
the identical io_set_blocking() and io_set_linger(). Since both are always called to turn off blocking or lingering, nuke the parameter and associated enum in favour of "just doing the right thing". While passing remove the unneeded last parameter to the remaining fcntl(F_GETFL). Finally, rename the functions to io_set_nonblocking() and io_set_nolinger() for clarity. No functional change. Started with a sweep of fcntl() usage inspired by guenther@. ok gilles@
2016-03-17Last parameter to execl[e]() functions *must* be cast to a pointer.Kenneth R Westerback
Just NULL is not good practise as NULL is theoretically allowed to be an integer rather than a pointer. Use (char *)NULL consistently instead of scattering a few (char *)0 and (void *)NULL into the mix. Prompted by and probably ok deraadt@ millert@ kettenis@ Definitely ok mestre@ ratchov@
2016-03-07fix error in documentation of `ca' keywordGilles Chehade
From: Sevan Janiyan <venture37@geeklan.co.uk>
2016-02-22the default address family for a listener was set too late and wouldGilles Chehade
override the value provided by the user if any. this commit moves the initialization earlier to avoid this. issue reported by and fix ok jturner@
2016-02-21bump versionGilles Chehade
2016-02-15Revert previous, semarie@ found a regression.Martin Pieuchot
2016-02-14Simplify address parsing code by only using inet_net_pton(3).Martin Pieuchot
There's no need to have a separate case for non-CIDR addresses using inet_pton(3) as inet_net_pton(3) handles them as well. This simplification is valid for all our daemons parsing addresses, it is a common patern. ok gilles@, millert@, jung@
2016-02-13tweak previous;Jason McIntyre
2016-02-13describe how 'listen on socket' worksGilles Chehade
original diff from Peter Bisroev <peter@int19h.net> rewording from me, linebreaks from jung@
2016-02-13handle enqueuer socket as a regular listener that can be configured withGilles Chehade
"listen on socket". this simplifies a bit of code, removes some special cases and will allow attaching filters & masking source just as on lo0. diff from Peter Bisroev <peter@int19h.net> ok gilles@, jung@
2016-02-13remove some noop leftover code from an earlier cleanupGilles Chehade
diff from Peter Bisroev <peter@int19h.net>
2016-02-13remove support for 'dbm' which is really 'recno' from our makemap, it can'tGilles Chehade
work because it requires dbopen() to be the call that creates the db, while we want mkstemp() to create an empty db ourselves. it has never worked, no one could ever create a recno db with our makemap & no one actually noticed so drop it. ok millert@, sunil@, jung@
2016-02-12Cleanup getopt(3) mis-use in sendmail compat code.Sunil Nimmagadda
Avoid multiple getopt(3) evaluations without optreset set and unspecified behavior of optind = 0 as per POSIX. Instead, iterate over argv the first time for classification. Fixes -portable without any divergence from -current, and all newaliases opts are handled in one place. Looks good to me millert@ Ok gilles@ jung@ eric@
2016-02-10Change imsg_read() return value checking from an if..else block toTodd C. Miller
a switch statement for improved readability. OK jung@ sunil@ gilles@
2016-02-10when creating a db map, we use a temporary file to work on and atomicallyGilles Chehade
move to destination when done. this causes dbopen() to be called with an empty file which happens to... not work with other db libraries. adding O_TRUNC is noop for us, fixes the issue for other libraries and lets us reduce delta between us and portable. ok sunil@, ok eric@
2016-02-09add PROC_QUEUE_MESSAGE_UNCORRUPT to smtpd-api.h, needed by privsep queuesGilles Chehade
2016-02-09r1.16 accidentally masked retries on EAGAIN introduced in r1.15.Sunil Nimmagadda
Fix and simplify imsg_read() return value checking into a single if..else block. Ok gilles@ jung@
2016-02-09rename variable sun to s_un to ease portabilityGilles Chehade
diff from Freddy DISSAUX, ok gilles@, jung@, sunil@, millert@
2016-02-05remove no longer relevant ifndefJoerg Jung
suggested by gilles
2016-02-05use SMTPD_NAME defineJoerg Jung
ok gilles
2016-02-04fix memory leakEric Faurot
ok gilles@ jung@
2016-02-04Make code more readable by putting the message parsing logic in a single place.Eric Faurot
ok gilles@ sunil@ jung@
2016-02-04When deleting a message, remove associated envelopes from the cache,Eric Faurot
if any (it happens when a message transaction is cancelled). ok gilles@ sunil@
2016-02-03Add helpers function to talk to the queue. Make the code simpler.Eric Faurot
ok gilles@ sunil@
2016-02-03tell filters to rollback the current transaction if MAIL or EOM failsEric Faurot
ok gilles@ jung@
2016-02-03When a filter rejects a message content, report immediatly to theEric Faurot
smtp session. Simplify code while here. ok gilles@ jung@
2016-02-03NULL initialize two file pointers to make it more obvious when reading thatGilles Chehade
they can't end up uninitialized after the mkstemp/fdopen dance
2016-02-03Use "esc_class" to classify bounce type instead of "errorline" asSunil Nimmagadda
we no longer prepend status code to "errorline". Fixes mismatch between DSN's subject line and its content. Ok jung@ gilles@ millert@
2016-02-02in RSA privsep engine, do not provide methods for rsa_sign / rsa_verify,Gilles Chehade
they are unused in OpenSMTPD and lead to crashes in -portable when we're linked to OpenSSL starting with 1.0.2f ok reyk@
2016-02-02Remove setproctitle() for the parent process. Because rc.d(8) uses processStuart Henderson
titles (including flags) to distinguish between daemons, this makes it possible to manage multiple copies of a daemon using the normal infrastructure by symlinking rc.d scripts to a new name. ok jung@ ajacoutot@, smtpd ok gilles@
2016-02-02Check imsg data size before use.Sunil Nimmagadda
Ok eric@ millert@
2016-01-29remove unused fieldEric Faurot
2016-01-29Simplify code: hasrun is confusing and useless. There is no wayEric Faurot
filter_drain_query() can be called when the ->current filter is still running the query. Just make sure we move forward after having received the response from the current filter. ok gilles@
2016-01-29Remove the query queueing mechanism. It's useless since pipelining isEric Faurot
not supported anyway. ok sunil@ gilles@
2016-01-28Simplify the query-chain code now that events are sent through a dedicated loop.Eric Faurot
ok gilles@ jung@
2016-01-27Check imsg data length before use.Sunil Nimmagadda
Ok jung@ gilles@ eric@
2016-01-27send events to filters in a simple loop, rather than using theEric Faurot
query-chaining machinery. ok jung@ gilles@
2016-01-22in lka process, revoke proc/exec pledges after privsep-ed table backendsGilles Chehade
are forked at startup i thought i had committed this already, diff has been okayed several weeks ago... ok jung@, ok millert@, ok sunil@
2016-01-22add a log_warn() so that when smtpd fails to start due to a problem readingGilles Chehade
the configuration file of a table, user actually understands what happens diff by Alexis Vachette, ok jung@
2016-01-21another round of tiny wording tweaks and unificationsJoerg Jung
all from tj ok millert