| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | sort options and SYNOPSIS; |   jmc | 2004-01-05 | 1 | -4/+4 | |
| | | | | | sync usage(); | |||||
| * | spacing |   deraadt | 2004-01-04 | 1 | -3/+3 | |
| | | ||||||
| * | Buffered logging for syslogd. Logs may be stored in memory buffers and |   djm | 2004-01-04 | 1 | -45/+384 | |
| | | | | | | extracted using a small client. Useful for diskless systems. much feedback from deraadt@, canacar@, jmc@, jakob@ ; ok deraadt@ | |||||
| * | spacing | deraadt | 2003-12-29 | 1 | -5/+5 | |
| | | ||||||
| * | convert from select() to poll() for main event loop; tested millert@, deraadt@ | djm | 2003-12-29 | 1 | -73/+59 | |
| | | | | | ok deraadt@ | |||||
| * | whack getopt ? matching; jeffi@rcn.com | deraadt | 2003-09-19 | 1 | -3/+2 | |
| | | ||||||
| * | Call tzset() to stash the current timezone before we chroot. |   millert | 2003-08-18 | 1 | -2/+4 | |
| | | | | | OK deraadt@ and avsm@ | |||||
| * | Privilege separated syslog daemon. The child listening to log requests drops |   avsm | 2003-07-31 | 1 | -64/+120 | |
| | | | | | | | | | | | | | | to user _syslogd and chroots itself, while the privileged parent grants it access to open logfiles and other calls it needs. The only difference from existing behaviour is that if syslog.conf changes and syslogd receives a HUP, it will re-exec itself and have two new PIDs. A HUP with an unchanged config will make syslogd reopen logfiles as before. Lots of help and code from deraadt@ , and advice from millert@ Various versions tested by todd, cloder, mpech, markus, tdeval and others | |||||
| * | dont silently truncate a domain socket specified via the -p argument; | avsm | 2003-07-08 | 1 | -3/+8 | |
| | | | | | error out instead if its too long. deraadt@ ok | |||||
| * | Remove the advertising clause in the UCB license which Berkeley | millert | 2003-06-02 | 1 | -7/+3 | |
| | | | | | rescinded 22 July 1999. Proofed by myself and Theo. | |||||
| * | Fix typo in last commit, spotted by Brian Poole. That'll teach me | millert | 2003-05-17 | 1 | -3/+3 | |
| | | | | | to commit on an empty stomach. | |||||
| * | Cannot use strlcpy() for strings in struct utmp since they are not guaranteed | millert | 2003-05-17 | 1 | -3/+5 | |
| | | | | | to be NUL-terminated. Fixes a bug introduced in rev 1.37; noticed by deraadt@ | |||||
| * | Fix default facility when none is specified (should be LOG_USER, | millert | 2003-03-21 | 1 | -4/+8 | |
| | | | | | was LOG_UUCP due to a bug); Javier Kohen | |||||
| * | fix cutnpaste; from hunter@comsys.com.ua |   mickey | 2002-12-22 | 1 | -3/+3 | |
| | | ||||||
| * | Add -n option to print message source addresses numerically rather than |   cloder | 2002-11-21 | 1 | -11/+18 | |
| | | | | | | | | symbolically. This saves address->name lookups, which is nice on log servers without a DNS cache. millert@, jakob@ ok | |||||
| * | use more socklen_t; pvalchev | deraadt | 2002-09-06 | 1 | -4/+4 | |
| | | ||||||
| * | Pass correct buffer length to snprintf on error condition; ericj@, fgsch@ ok. |   aaron | 2002-08-08 | 1 | -3/+3 | |
| | | ||||||
| * | Sanity check snprintf() return values before using them. It is legal for | millert | 2002-07-24 | 1 | -15/+19 | |
| | | | | | snprintf(3) to return -1. | |||||
| * | ansi | deraadt | 2002-07-20 | 1 | -55/+28 | |
| | | ||||||
| * | spelling; from Brian Poole <raj@cerias.purdue.edu> |   todd | 2002-06-14 | 1 | -3/+3 | |
| | | ||||||
| * | Double socket receive buffer size; deraadt@ OK | millert | 2002-06-05 | 1 | -2/+17 | |
| | | ||||||
| * | pid_t cleanup | deraadt | 2002-05-26 | 1 | -3/+3 | |
| | | ||||||
| * | fix usage; from Sam Smith <S@mSmith.net>. |   fgsch | 2002-05-25 | 1 | -3/+4 | |
| | | ||||||
| * | Part one of userland __P removal. Done with a simple regexp with some minor hand editing to make comments line up correctly. Another pass is forthcoming that handles the cases that could not be done automatically. | millert | 2002-02-16 | 1 | -21/+21 | |
| | | ||||||
| * | KNF, and remove unneccessary signal blocking and errno saving, since not signal handlers anymore; millert ok | deraadt | 2001-12-02 | 1 | -13/+4 | |
| | | ||||||
| * | use volatile sig_atomic_t where possible; simply volatile i other places | deraadt | 2001-11-17 | 1 | -5/+5 | |
| | | ||||||
| * | missing prototype (getmsgbufsize). |   markus | 2001-11-17 | 1 | -2/+3 | |
| | | ||||||
| * | reread config from main loop instead of signal hander; ok deraadt@ | markus | 2001-11-17 | 1 | -7/+25 | |
| | | ||||||
| * | no need for setjmp.h | deraadt | 2001-11-16 | 1 | -3/+2 | |
| | | ||||||
| * | Open files with O_NONBLOCK but turn off non-blocking mode for | millert | 2001-08-03 | 1 | -11/+26 | |
| | | | | | | | | non-ttys. If write(2) returns EAGAIN just ignore the error and move on. This prevents a locked terminal from causing syslogd grief. If we ever want to support logging to a fifo this will probably have to be revisited. | |||||
| * | fix typo; jcs@rt.fm | deraadt | 2001-08-03 | 1 | -3/+3 | |
| | | ||||||
| * | Wall cleanup |   pvalchev | 2001-07-27 | 1 | -3/+4 | |
| | | ||||||
| * | Remove support for #!. It's not intuitive and if you happen to add | fgsch | 2001-02-07 | 1 | -9/+4 | |
| | | | | | | | lines as the end of a stock syslog.conf you're going to spend some time figuring why is not working as it should; problem found by riq@core-sdi.com, millert@ ok. | |||||
| * | mark remaining signal races which are difficult to fix | deraadt | 2001-01-19 | 1 | -4/+5 | |
| | | ||||||
| * | fix another signal race, and check signal flags even in non-EINTR from select case | deraadt | 2001-01-16 | 1 | -7/+19 | |
| | | ||||||
| * | fd_set overflows, move races out of signal handlers, and fd_set repairs; | deraadt | 2001-01-11 | 1 | -76/+92 | |
| | | | | | | some conversations with alejo@core-sdi.com, not sure yet if this is 100% perfect, but i have tested it and it works.. | |||||
| * | Use vis(3) when escaping non-printable characters instead of doing it | millert | 2000-09-13 | 1 | -22/+14 | |
| | | | | | | (poorly) by hand. Also fixes a potential one byte overflow noted by Solar Designer. | |||||
| * | do the save_errno thing, but other races still exist in here | deraadt | 2000-08-17 | 1 | -3/+9 | |
| | | ||||||
| * | enlarge msgbuf, somewhat line netbsd did | deraadt | 2000-02-22 | 1 | -5/+32 | |
| | | ||||||
| * | union wait -> int | millert | 1999-07-06 | 1 | -4/+9 | |
| | | | | | wait3 -> waitpid (for portability) | |||||
| * | The necessary defs are in paths.h now. |   angelos | 1999-02-03 | 1 | -1/+1 | |
| | | ||||||
| * | careful about sunx.sun_path termination; matter@research.suspicious.org | deraadt | 1999-01-05 | 1 | -3/+7 | |
| | | ||||||
| * | make it 21 total log devs | deraadt | 1998-06-26 | 1 | -1/+1 | |
| | | ||||||
| * | Fix snprintf return value usage. | millert | 1998-06-23 | 1 | -0/+2 | |
| | | ||||||
| * | bogus use of variable for two purposes; guido@gvr.org | deraadt | 1998-06-11 | 1 | -4/+4 | |
| | | ||||||
| * | MAXHOSTNAMELEN not MAXHOSTNAMELEN+1 | deraadt | 1998-05-18 | 1 | -6/+9 | |
| | | ||||||
| * | select on nfds+1, not nfds. ficus@openbsd.org | millert | 1998-05-05 | 1 | -1/+1 | |
| | | ||||||
| * | use sigset_t all over the place | deraadt | 1998-03-01 | 1 | -4/+8 | |
| | | ||||||
| * | syslogd gets a new -a argument: specify additional AF_UNIX log devices | deraadt | 1998-02-11 | 1 | -43/+78 | |
| | | | | | | syslogd should create & listen to. As in "syslogd -a /chroot/dev/log", I'm sure you get the idea. | |||||
| * | workaround gethostbyaddr() race in signal handler; more correct fix would | deraadt | 1998-02-03 | 1 | -0/+5 | |
| | | | | | | change SIGHUP handler to only set a flag that the main loop catches. But I'm lazy, and starting to believe noone gives a damn... | |||||
 |
index : wireguard-openbsd | |
| WireGuard implementation for the OpenBSD kernel | Matt Dunwoodie |
| summaryrefslogtreecommitdiffstats |