Commit Graph

86 Commits

Author SHA1 Message Date
christos
5ca310136d deal with const in the new openssl 2005-11-26 01:10:42 +00:00
christos
19d0f25ed5 PR/31580: Miles Nordin 'ntptime' no longer reports ERROR and UNSYNC when ntpd
isn't synchronized (ntpd bug #506)
2005-10-20 15:41:01 +00:00
elad
a615d99b28 Use group-id from `gr'' and not `pw''.
Fixes security issue addressed by CAN-2005-2496.
Thanks to Adrian Portelli for notifying me about this.
2005-10-13 14:58:23 +00:00
lukem
f7e488f953 Don't attempt to use an uninitialized variable when there's no interfaces.
Detected with gcc -Wuninitialized.
2005-06-03 16:13:33 +00:00
riz
254aabe783 There was a change from usecs to nsecs between ntpd 4.1.1 and 4.2.0 -
take this into account correctly.  From ntp-current via Brian Buhrow
in PR#29685.  Approved by christos.
2005-05-21 01:47:09 +00:00
christos
aba19ee8af Double the stack space limit. We are running out with ntp.pool.org. 2005-04-05 20:56:58 +00:00
lukem
c6f2cec0b0 include <isc/net.h> so that -UWANT_IPV6 compiles correctly. 2005-01-10 02:14:46 +00:00
dsl
968ad08000 Stop argument to ctype function being char 2004-10-30 15:01:32 +00:00
christos
72f840350f PR/27210: Paul Shupak: "ntpd"'s NEMA driver has 8 units, but MAX_UNITS
is #defined as 4.
2004-10-10 22:13:04 +00:00
christos
766798b654 PR/27137: Martin J. Laubach: core dump when localhost does not support
a given address family and a peer only supports the family localhost does
not support. For example: configure a kernel without IPV6, and then
add a line in ntp.conf server <ipv6addr>. We report that the server is
unreachable and we keep going because there might be more servers around?
XXX: What if it is the last server? Should we detect this? It is not nice
to just bail on this error, because a server might lose its ipv4 address
and only advertise ipv6.
2004-10-05 03:34:38 +00:00
christos
6e1b89d727 80K of stack is not enough for a 64 bit machine, and we got crashes on
sparc64. change 20 * 4096 to 5 * 4096 * sizeof(void *). This is again
very little, but enough :-). If you see a crash again, don't blame the
nameserver code before you change this constant.
2004-06-29 04:16:37 +00:00
drochner
e917a2fcb1 ignore 0.0.0.0 interface addresses (dhclient leftovers) 2004-02-26 18:17:13 +00:00
fredb
8f443ee26a Don't bother complaining about "no IPv6 interfaces" or "no IPv4 interfaces"
unless debugging is compiled in and enabled. Adapted from ntp-dev. Closes
PR bin/23703, by Martin Husemann.
2003-12-20 17:38:19 +00:00
keihan
9f474dc36a Make "-s" option silent again, fixes PR bin/23707.
This has been fixed in current ntp-dev and also includes
some whitespace fixes.  NTP bug report available at:

http://bugzilla.ntp.org/show_bug.cgi?id=263
2003-12-20 16:12:27 +00:00
drochner
a273546bc0 pull in a set of patches from ntp bk to fix the
"Clock stepping messes up frequency." issue
2003-12-04 17:22:31 +00:00
drochner
507fa5fee7 compile cleanly with -Wall (ntp bug #258) 2003-12-04 17:17:36 +00:00
drochner
35f545891e fix some memory allocation bugs (ntp bug #252) 2003-12-04 17:15:26 +00:00
drochner
a3d91b3051 fix error handling after send errors (ntp bug #243) 2003-12-04 17:13:11 +00:00
drochner
849e83ebc4 fix some LP64 issues (ntp bug #233) 2003-12-04 17:10:34 +00:00
drochner
93cc3883d8 use return value of openssl's RAND_file_name(), this makes the
fallback to /dev/urandom work as intended by openssl
(ntp bug #234)
2003-12-04 17:06:12 +00:00
drochner
7b3a4acbea don't join a multicast group where we just want to send
(this might be preliminary, it might be intended to receive other host's
multicast in the future, but until packets are filtered properly this
fixes the "manycastclient answers requests" problem - see ntp bug #241)
2003-12-04 17:01:32 +00:00
drochner
74a4307696 deal with sin6_scope_id where necessary 2003-12-04 16:56:01 +00:00
drochner
3cdd2cee5f keep track of the scope id of link-local ipv6 addresses, and deal with
the KAME specific way to embed it into the address
(keeping the interface close to the last public bind-9.3 snapshot to
avoid conflicts in the future)
2003-12-04 16:52:28 +00:00
drochner
d5e6f1b80b resolve conflicts
(in this case, throw away all local modifications; I'll start patching
from scratch)
2003-12-04 16:23:34 +00:00
drochner
b21928a34e Import ntp 4.2.0 2003-12-04 16:05:14 +00:00
simonb
a1f8e60591 Update for CVS repo basesrc/src change (many moons ago). 2003-10-23 02:23:26 +00:00
agc
865595bdf3 Move UCB-licensed code from 4-clause to 3-clause licence.
Patches provided by Joel Baker in PR 22253, verified by myself.
2003-08-07 09:20:39 +00:00
cb
ee482404e6 Output key IDs in decimal, not hexadecimal, so that they can
be read back later.

fixes PR#14216. ok by Simon Burge (and by him via ntp distribution)
2003-07-16 12:22:58 +00:00
itojun
93851338f2 sscanf overrun 2003-05-17 01:12:28 +00:00
wiz
0acfa3bb9e Consistently spell occurrence with two rs. 2003-03-29 22:48:37 +00:00
wiz
447edbab8d hexadecimal, not hexidecimal. Reported to the authors. 2003-03-16 21:54:46 +00:00
wiz
990562bfef .Nm does not need a dummy argument ("") before punctuation or
for correct formatting of the SYNOPSIS any longer.
2003-02-25 10:34:36 +00:00
jdolecek
28f3685337 use correct group id if group was specified for -u; previously, it would
mistakely use user's primary group instead
problem reported and fix supplied in PR bin/20232 by KAMADA Ken'ichi
2003-02-08 11:45:54 +00:00
wiz
dd5bd120af Various misspellings, from Adrian Mrva. 2002-12-21 12:37:49 +00:00
wiz
34574e798b outgoing with a final g, not t; from Adrian Mrva. 2002-12-21 12:32:24 +00:00
thorpej
e733469648 Avoid strict alias warnings. 2002-12-06 15:12:56 +00:00
thorpej
47355062e5 Avoid strict alias warning. 2002-12-06 15:09:38 +00:00
uwe
fd3534dbc5 Nit-pick: use || instead of | in the test. Ok by simonb. 2002-09-20 14:09:45 +00:00
wiz
111e203639 Comment out token after #endif. 2002-05-16 19:53:37 +00:00
wiz
79deb5d12d Comment out token after #endif. 2002-05-16 19:40:53 +00:00
wiz
98d39977a3 Remove token after #endif. 2002-05-16 19:30:17 +00:00
drochner
1fba973e56 make it build without REFCLOCK but with -Wall 2002-04-19 20:45:54 +00:00
christos
590eabc9c6 Set the stack limit to something smaller than the default, like it is
done in the aix case so that we don't end up mlockall() unused memory.
From: witek@wnuk.eu.org (Witold J. Wnuk)
2002-04-18 23:18:42 +00:00
thorpej
64b4b1c3c3 Don't use static const u_char ... for register bits. Use #define. 2002-04-09 02:42:50 +00:00
mason
be7a76e49f document -i and -u in the man page 2002-02-21 17:10:14 +00:00
wiz
78cb877358 Try to remove some more unnecessary whitespace when converting. 2002-01-19 11:43:47 +00:00
martin
114dbddbad Make ntpd use the getifaddrs interface, thus avoiding alignement
problems (see port-sparc64/15112).

This change will be submitted to the NTP maintainers as well.
2002-01-05 13:24:40 +00:00
martin
6db2d4b116 Back out previous change - turns out to be a sparc64 toolchain problem
(surprise!)
2002-01-03 11:22:07 +00:00
martin
357b95fa25 Replace a structure copy (dereferencing a pointer received from an ioctl)
by a memcpy, similar to the patch suggested by jarle@uninett.no in PR 15112.
I'm not sure if the ioctl code should be fixed to deliver this pointer
properly aligned, but this change should not cause any problem and it is
safer this way.
2002-01-02 20:47:37 +00:00
wiz
546fc71d49 Pull over fixes from rev. 1.5-1.7 of basesrc/usr.sbin/ntp/ntpd/ntpd.8. 2001-12-26 15:14:59 +00:00