| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2019-07-10 | obfuscate auth parameters in smtp reporting | Gilles Chehade | |
| 2019-07-09 | rework link-connect output so that messages enqueued over the unix socket | Gilles Chehade | |
| no longer display "local:0" as their source and dest addresses, but "unix:/var/run/smtpd.sock" currently we only support enqueueing over the control socket but that may not always be the case. | |||
| 2019-07-03 | snprintf/vsnprintf return < 0 on error, rather than -1. | Theo de Raadt | |
| 2019-07-02 | Replace mkstemp/fdopen/unlink dances with tmpfile. | Martijn van Duren | |
| No functional change. OK gilles@ | |||
| 2019-07-01 | The proceed reply doesn't have a parameter, resulting in subsequent filters | Martijn van Duren | |
| in a chain to receive a "(null)" parameter. Fix this by storing the current parameter in the session. Found the hard way and fix tested by Mischa Peters. OK gilles@ | |||
| 2019-06-28 | When system calls indicate an error they return -1, not some arbitrary | Theo de Raadt | |
| value < 0. errno is only updated in this case. Change all (most?) callers of syscalls to follow this better, and let's see if this strictness helps us in the future. | |||
| 2019-06-28 | switch timestamps to use timevals in filters like we do in reporting | Gilles Chehade | |
| 2019-06-28 | mkstemp() returns -1 on failure | Theo de Raadt | |
| 2019-06-27 | Move the check wether line is NULL above further accesses of it. | Matthias Kilian | |
| ok gilles@ | |||
| 2019-06-27 | Allow filters to log information through stderr. This simplifies and | Martijn van Duren | |
| unifies the way filters need to get their logging to the right location. Log-messages are read line by line and are logged at LOG_ERR level via the lookup process. OK gilles@ | |||
| 2019-06-26 | reject DATA containing <CR> that's not part of <CRLF> line terminators | Gilles Chehade | |
| as they are not allowed by RFC, have never been generated by a regular MUA/MTA and can cause issues with some filters if we let them through. ok jung@ on an earlier version of the diff, eric@ suggested the new diff | |||
| 2019-06-24 | rename wait queues | Gilles Chehade | |
| 2019-06-14 | simplify the runq interface: | Eric Faurot | |
| - remove (unused) per-job callback - rename runq_schedule() to runq_schedule_at() and runq_delay() to runq_schedule() - remove unused runq_next() ok sunil@ gilles@ | |||
| 2019-06-13 | extend the resolver interface to delegate res_query() calls to the lka. | Eric Faurot | |
| ok gilles@ sunil@ | |||
| 2019-06-12 | change "ssl" to "tls" in various identifiers. | Eric Faurot | |
| no functional change. ok gilles@ | |||
| 2019-06-05 | assume RSA_METHOD is opaque and only access members through setters/getters | Gilles Chehade | |
| 2019-06-05 | a long long time ago, in a galaxy quite close actually, reyk@ introduced an | Gilles Chehade | |
| RSA privsep engine to isolate private keys in the ca process. ECDSA support in smtpd is become a frequent request so here's an ECDSA privsep engine and the code required for smtpd to load ECDSA certificates and use them. | |||
| 2019-05-24 | switch from having automatic EVP_CIPHER_CTX variables to allocating them | Gilles Chehade | |
| with EVP_CIPHER_CTX_new() and releasing them with EVP_CIPHER_CTX_free(). ok sunil@ and millert@ | |||
| 2019-05-24 | ERR_R_SSL_LIB is defined to ERR_LIB_SSL which works for both LibreSSL and | Gilles Chehade | |
| OpenSSL, so use that one instead to reduce delta in portable branch | |||
| 2019-05-24 | assume X509_STORE_CTX is opaque, don't access ->error but use the | Gilles Chehade | |
| X509_STORE_CTX_get_error() function instead | |||
| 2019-05-24 | mechanical change to dynamically allocate rsae_method | Gilles Chehade | |
| 2019-05-24 | fix typo in fatal message to -> too | Gilles Chehade | |
| 2019-05-24 | remove useless check, it's never been and will never be hit | Gilles Chehade | |
| 2019-05-21 | Use malloc_ and calloc_conceal() to allocate memory for sensitive, | Otto Moerbeek | |
| private key data. We leave the corresponding freezero() calls since the code might be ported to a system not having the conceal functions. This also makes sure the stdio bufer is concealed. And zap a redundant free(buf); ok gilles@ | |||
| 2019-05-20 | rev 1.219 of the file was a fix to an LMTP issue that was misunderstood. we | Gilles Chehade | |
| fixed the root cause in mda_variables.c months ago independantely but we're still carrying this diff which is not only unnecessary now, but is also the cause of a bug with mailer-daemons going through an LMTP mda. issue reported by otto@ | |||
| 2019-05-15 | do not use CRLF when passing data lines to filters. | Eric Faurot | |
| ok gilles@ | |||
| 2019-05-15 | normalize newlines when reading the message | Eric Faurot | |
| spotted by martijn@ ok martijn@ gilles@ | |||
| 2019-05-14 | fix typos in log messages | Eric Faurot | |
| 2019-05-02 | Give filter programmers a more friendlier error message when returning a | Martijn van Duren | |
| filter-dataline in non-dataline phases and vice versa instead of crashing with "smtpd: tree_xpop(0x1e1a6b180b40, 0xc77804c6ba110977)" OK gilles@ | |||
| 2019-04-08 | remove unused declarations | Eric Faurot | |
| ok gilles@ | |||
| 2019-04-08 | fix parsing of datalines before passing them to filters, the | splitting is | Gilles Chehade | |
| done a bit too early diff from Martijn van Duren <openbsd+tech@list.imperialat.at> | |||
| 2019-03-31 | Avoid calling dup2(oldd, newd) when oldd == newd. In that case the | YASUOKA Masahiko | |
| descriptor keeps CLOEXEC flag then it will be closed unexpectedly by exec(). ok tedu florian | |||
| 2019-02-28 | bump smtpd version | Eric Faurot | |
| on behalf of gilles@ | |||
| 2019-02-20 | fix a regression in the 'hostnames' mapping to select hostname based on the | Gilles Chehade | |
| IP address of a listener, it should use the listener address not the client address as a selector, otherwise it will always display the default name if you don't connect from the local host. | |||
| 2019-02-13 | (unsigned) means (unsigned int) which on ptrdiff_t or size_t or other | Theo de Raadt | |
| larger types really is a range reduction... Almost any cast to (unsigned) is a bug. ok millert tb benno | |||
| 2019-02-06 | in SMTP status, display rcpt address in errors happening at RCPT level | Gilles Chehade | |
| ok millert@, sunil@ | |||
| 2019-01-30 | missed in previous revert, function not called anymore | Gilles Chehade | |
| 2019-01-30 | revert previous commit, i wasn't happy with it and it probably came from a | Gilles Chehade | |
| misunderstanding. | |||
| 2019-01-30 | don't be too strict with .forward permissions, it's ok to process it if the | Gilles Chehade | |
| group has write access, it's not ok if the world has write access. ok eric@ | |||
| 2019-01-30 | Check ferror(3) on the right stream. | Sunil Nimmagadda | |
| Ok gilles@ | |||
| 2019-01-29 | strip \r from lmtp responses | Gilles Chehade | |
| ok sunil@ | |||
| 2019-01-21 | While using buffered stdio over a socket, fflush(3) command before | Sunil Nimmagadda | |
| reading the response. Ok eric@ jung@ | |||
| 2019-01-15 | Swap 'token' and 'id' place in filter protocol. | Sunil Nimmagadda | |
| This allows to consistently extract 'id' from both 'report' and 'filter' lines. Ok gilles@ jung@ | |||
| 2019-01-14 | correctly retrieve envelope strings. fixes mailq output. | Eric Faurot | |
| ok gilles@ | |||
| 2019-01-10 | report errno in debug logs and other minor cleanups. | Eric Faurot | |
| ok gilles@ | |||
| 2019-01-05 | we assumed that there was no such thing as an "mda session" so the logs | Gilles Chehade | |
| displayed 0000000000000000 as the session id for mda deliveries, but we were wrong. there's actually an identifier that's associated to a batch of envelopes to be delivered to the same user, and this most definitely is a session id for that user. | |||
| 2019-01-05 | revert this change, it was committed by accident | Gilles Chehade | |
| 2019-01-05 | introduce smtp 'timeout' reporting event to notify filters that a timeout | Gilles Chehade | |
| occured during the smtp session | |||
| 2019-01-05 | move the DATA bytes accounting a bit earlier so that we don't have to deal | Gilles Chehade | |
| with it using special cases for when filters are enabled or not. | |||
| 2019-01-03 | be less verbose on resolver "errors", only log_warnx() real errors and not | Gilles Chehade | |
| EAI_NODATA / EAI_NONAME in cases where these are expected and already part of a log line. ok eric@ | |||
 |
index : src | |
| OpenBSD base system |
| summaryrefslogtreecommitdiff |