Submitted by Shaun Amott in PR/34231.
Changes since previous version in pkgsrc:
20060513
- (dtucker) [configure.ac includes.h ntp_msg.c] Add UnixWare support. Based
on patch from luke.bakken at gmail.com.
- (dtucker) [version.h] Release 3.9.
- (dtucker) [openbsd-compat/fake-rfc2553.c] Missing braces in initializer.
20060423
- (dtucker) [configure.ac] Correctly fall back to builtin-arc4random when
OpenSSL is not present.
20060116
- (dtucker) [INSTALL configure.ac ntp.c] Make configure --with-privsep-path
compile in the chroot directory. Default behavior remains to chroot to
the ntpd user's home dir. Based in part on a patch from OpenWall via
solar at openwall com.
- (dtucker) [Makefile.in] Make rebuild of y.tab.c parse conditional.
- (dtucker) [configure.ac defines.h] Enable replacement getifaddrs via
ioctl on Linux.
- (dtucker) [configure.ac openbsd-compat/fake-rfc2553.{c,h}] Add in6addr_any if
system libraries don't have it.
- (dtucker) [configure.ac openbsd-compat/bsd-misc.c] Add null implementation
of strsignal().
- (dtucker) [openbsd-compat/openbsd-compat.h] Template for strsignal.
- (dtucker) OpenBSD CVS Sync
- dtucker@cvs.openbsd.org 2005/07/06 19:54:24
[client.c]
add another non-fatal error for recvfrom; ok henning@
(ID sync only, already in portable)
- dtucker@cvs.openbsd.org 2005/07/11 18:04:28
[client.c]
Print actual error when in debug mode; ok henning@
- dtucker@cvs.openbsd.org 2005/07/11 18:05:34
[ntp.c]
Print privsep user and dir when in debug mode; ok henning@
- dtucker@cvs.openbsd.org 2005/07/11 18:08:06
[ntpd.c]
More descriptive error if a signal causes the child to exit; ok henning@
- henning@cvs.openbsd.org 2005/07/15 13:34:52
[ntp.c]
fix a function name in an error message why this was rotting in my
tree for so long, I dunno - and I dunno where it came from
- henning@cvs.openbsd.org 2005/07/15 13:36:10
[ntp.c]
remove recently added "using privsep user X" message, spams console
in -s mode, noticed by kettenis
- henning@cvs.openbsd.org 2005/07/15 13:37:15
[ntpd.h]
shrink read buffer size from 64k to 4k, this is not bgpd and we're
dealing with way less data
- dtucker@cvs.openbsd.org 2005/07/22 18:58:56
[server.c]
Skip invalid interfaces during 'listen on *'; ok henning@
- henning@cvs.openbsd.org 2005/08/09 00:42:32
[ntp.c]
with -s, do not wait if we don't have any peers at all. From: Thomas
Jarosch <thomas.jarosch@intra2net.com>
- dtucker@cvs.openbsd.org 2005/08/10 23:48:36
[client.c ntp.c ntpd.h server.c]
Propogate server's leap indicator flags to clients; ok henning@
- henning@cvs.openbsd.org 2005/08/12 02:21:52
[buffer.c]
check for EINTR too after writev(), pt out by Alexander Farber
- henning@cvs.openbsd.org 2005/08/12 02:26:29
[buffer.c]
on writing, we actually can deal with ENOBUFS just as well as with
EAGAIN and EINTR, so do it, more or less from bgpd
- wvdputte@cvs.openbsd.org 2005/09/07 07:27:10
[ntpd.c]
when running ntpd with "-s" as it's argument from /etc/rc.conf, make
sure the output goes to syslog and not console by moving around log_init
- dtucker@cvs.openbsd.org 2005/09/24 10:32:03
[client.c ntp_msg.c ntpd.h server.c]
Log source address for 'malformed packet' errors. ok henning@
- dtucker@cvs.openbsd.org 2006/01/19 17:40:16
[server.c]
Check SA_LEN(sa) after sa is checked for NULL. Pointed out by solar
at openwall.com, ok henning@
- dtucker@cvs.openbsd.org 2006/01/19 22:20:23
[server.c]
Do not attempt to listen on interfaces with a wildcard address;
ok henning@
20050729
- (dtucker) [configure.ac] Skip OpenSSL checks for cross-compilation. Found
and tested by Scott Hays.
20050707
- (dtucker) [openbsd-compat/bsd-setres[ug]id.c] Add code to use setre[ug]id
and some sanity checks.
- (dtucker) [configure.ac] Alphabetize $host case block.
- (dtucker) [configure.ac server.c openbsd-compat/Makefile.in
openbsd-compat/bsd-getifaddrs.{c,h} openbsd-compat/openbsd-compat.h]
Add getifaddrs() to compat layer. Not enabled by default on any platform
now, enable with -DGETIFADDRS_VIA_SIOCGIFCONF at your own risk.
- (dtucker) [client.c] recvfrom on HP-UX will return EADDRNOTAVAIL instead
of ECONNREFUSED for a port-unreachable, so add to the non-fatal error
codes.
- (dtucker) OpenBSD CVS Sync
- dtucker@cvs.openbsd.org 2005/07/05 20:09:12
[client.c ntp.c ntpd.h server.c]
Save transmit time for each peer for later use as refid for SNTPv4
replies. ok henning@
20050703
- (dtucker) OpenBSD CVS Sync
- henning@cvs.openbsd.org 2005/03/23 11:42:04
[imsg.c ntpd.h]
wpos in struct buf_read and datalen in imsg_get should be size_t and
not ssize_t From: Alexander von Gernler <grunk@pestilenz.org>
- henning@cvs.openbsd.org 2005/03/23 12:36:35
[buffer.c]
remove now osolete comment, from a mail exchange with Alexander von
Gernler <grunk@pestilenz.org>
- henning@cvs.openbsd.org 2005/03/24 11:56:22
[ntpd.c]
fatal vs fatalx, Alexander von Gernler
- henning@cvs.openbsd.org 2005/03/24 15:50:07
[ntp.c]
one more fatal/fatalx, alexander
- henning@cvs.openbsd.org 2005/03/31 12:14:01
[log.c]
zap includes, Alexander von Gernler <grunk@pestilenz.org>
- henning@cvs.openbsd.org 2005/03/31 17:02:43
[ntpd.c]
zap includes, grunk
- henning@cvs.openbsd.org 2005/04/18 11:06:35
[client.c]
prevent replies with negative delay from being used, could happen with
-s From: Joerg Sonnenberger <joerg@britannica.bec.de> of dragonfly
- henning@cvs.openbsd.org 2005/04/18 11:07:55
[ntp.c]
after setting the clock hard correct the "next" and "deadline" timestamps
by the offset From: Joerg Sonnenberger <joerg@britannica.bec.de>
- henning@cvs.openbsd.org 2005/04/18 14:12:50
[ntp.c]
correctness: only account for offset after settime in next and deadline
when those timers are actually running. due to the way ntpd's logic
works this does not really make a difference, but correctness is good.
spotted by me, joerg agrees
- henning@cvs.openbsd.org 2005/04/18 20:46:02
[ntpd.c]
extra paranoia, from a discussion with joerg
- henning@cvs.openbsd.org 2005/04/19 11:08:41
[client.c]
move the "reply from ... " log msg in -d mdoe uop a bit so it actually
comes before the "adjusting local clock by..." one, joerg
- henning@cvs.openbsd.org 2005/04/26 15:18:22
[buffer.c imsg.c ntpd.h]
unify shared code a bit again to make future syncs easier From:
Alexander von Gernler <grunk@pestilenz.org>
- djm@cvs.openbsd.org 2005/05/03 05:44:35
[ntp.c]
setres[ug]id; ok deraadt@
- henning@cvs.openbsd.org 2005/05/11 15:12:35
[config.c]
don't touch *hn in failure case. no real change due to the way we use
it but more correct. from Michael Knudsen <e@molioner.dk>
- henning@cvs.openbsd.org 2005/05/24 08:46:43
[ntp.c]
no need for endpwent(0 here either
- henning@cvs.openbsd.org 2005/05/25 06:10:50
[server.c]
ifa->ifa_addr can be NULL in some cases, pt out by Kurt Roeckx
<kurt@roeckx.be> / bugs.debian.org/310586
- dtucker@cvs.openbsd.org 2005/05/26 19:13:06
[ntp.c ntpd.c]
Ensure previous adjust has completed before clearing alarm flag;
ok henning@
- henning@cvs.openbsd.org 2005/06/20 02:42:57
[client.c ntp.c ntpd.c ntpd.h parse.y]
use a little state engine to keep track of delayed dns lookups and such,
eases things tested by Jason Ackley <jason@ackley.net> Matthias Kilian
<kili@outback.escape.de> Stephen Marley <stephen@marley.org.uk> sturm@
theo ok
- henning@cvs.openbsd.org 2005/06/20 03:11:13
[ntpd.c ntpd.h]
use a #define for the time to wait on -s and clarify a log msg
- deraadt@cvs.openbsd.org 2005/06/22 05:55:18
[ntpd.8] (ID sync only; section is not in portable man page)
we do not do -s in /etc/rc anymore. this is because, even if -s did
try to do it's job it would have to choose between two cases: 1. either
it would take a very long time to get the correct adjustment, thus,
if you are not currently on the net right, you wait a long time
(or must type ^C, which is ridiculous)
2. ntpd could be modified to "abort early", but then would not meet the
promise made by -s in the manual page (note: it does not say that it
"tries")
therefore, -s and -S must become user choices. Sorry. This same
choice is made in lots of other places
- (dtucker) [LICENCE configure.ac openbsd-compat/Makefile.in
openbsd-compat/bsd-setres[ug]id.c openbsd-compat/openbsd-compat.h
removed openbsd-compat/uidswap.c] Use setres[ug]id interface on all
platforms. Currently only implements the case where ruid == euid == suid
which is all ntpd uses (may be extended later).
- (dtucker) [Makefile.in] Use example ntpd.conf from srcdir so "make install"
works for the srcdir != builddir.
- (dtucker) [openbsd-compat/{bsd-poll.c,bsd-setresgid.c,bsd-setresuid.c}]
Add CVS Ids
20050629
- (dtucker) [configure.ac openbsd-compat/{Makefile.in,errx.c,verrx.c,
openbsd-compat.h}] Add errx(), in anticipation of it being used in ntpd.
- (dtucker) [INSTALL] Add a bit more detail on the privsep user and group.
20050627
- (dtucker) [README] Make it clear that the footnote only applies to old
Solaris systems, based on feedback from oyvind at repvik.org.
- (dtucker) [INSTALL] Point out that --with-ssl-dir is only used if
we're not using the builtin arc4random.
20050626
- (dtucker) [INSTALL] Add installation directions, lack thereof pointed out
by oyvind at repvik.org.
- (dtucker) [Makefile.in configure.ac] Have configure find a yacc that works.
- (dtucker) [Makefile.in] Make check for existing ntp user more specific.
20050619
- (dtucker) [openbsd-compat/bsd-poll.c] Remove code left over from debugging.
20050508
- (dtucker) [configure.ac] Check that required programs are found. Pointed
out by jj at it.su.se.
- (dtucker) [contrib/redhat/openntpd.spec] If ntp user already exists,
set its homedir to the chroot dir. From wijnand at nedbsd.nl.
- (dtucker) [contrib/redhat/openntpd.spec] Always create privsep dir
even if ntp user already exists. Also from Wijnand.
20050530
- (dtucker) [Makefile.in configure.ac mdoc2man.awk] Add support for other
man page formats (man and catman), based on OpenSSH's.
- (dtucker) [Makefile.in openbsd-compat/Makefile.in] Make out-of-tree
builds work.
- (dtucker) [Makefile.in openbsd-compat/Makefile.in] Minor cleanups.
- (dtucker) [LICENCE] Add Anil Madhavapeddy to atomicio bits.
- (dtucker) [version.h] Set version to -current.
20050529
- (dtucker) [openbsd-compat/atomicio.c openbsd-compat/atomicio.h
openbsd-compat/bsd-arc4random.c] Sync OpenBSD ssize_t -> size_t atomicio
change by avsm@, update rnd code to new interface.
20050528
- (dtucker) [openbsd-compat/bsd-poll.c] Portability and correctness fixes:
- Handle fd == -1 case.
- Handle fractional second timeouts correctly (not used in ntpd).
- Allow any negative timeout to mean INFTIM.
- (dtucker) [configure.ac] For AC_CHECK_HEADERS() and AC_CHECK_FUNCS() have
one entry per line to make it easier to merge changes.
- (dtucker) [defines.h includes.h openbsd-compat/bsd-poll.c] Copyright bump.
- (dtucker) [CREDITS Makefile.in includes.h version.h] Add CVS Id.
- (dtucker) [openbsd-compat/asprintf.c] char const -> const char, matches
OpenBSD 1.10 -> 1.11.
100cf09551