ChangeLog, refclock_nmea.c, refclock_oncore.c:
* ntpd/refclock_oncore.c (oncore_start): Set pps_enable=1, just
like the atom driver does.
From: reg@dwf.com
* ntpd/refclock_nmea.c (nmea_ppsapi): Set pps_enable=1, just like
the atom driver does.
From: Scott Allendorf <sca@newton.physics.uiowa.edu>
Harlan Stenn [Sat, 9 Jun 2001 01:07:24 +0000 (01:07 -0000)]
Many files:
* configure.in: 4.0.99m-rc1b
* ntpd/refclock_true.c (true_debug): Bump some buffer sizes to
reduce/eliminate chance of buffer overflow. Use snprintf()
instead of sprintf(). Do a better job of opening the debug file.
* ntpd/ntp_control.c (ctl_getitem): Count overflow packets as bad
and return a BADFMT.
* ntpd/ntp_config.c (save_resolve): call fdopen() with the correct
mode.
From: Bela Lubkin <belal@sco.com>
Harlan Stenn [Sun, 3 Jun 2001 01:35:43 +0000 (01:35 -0000)]
ChangeLog, NEWS, notes.htm, ntp_refclock.c:
* ntpd/ntp_refclock.c (refclock_open): Add O_CNOTTY to the open()
flags when opening a serial port.
Reported by: joseph lang <tcnojl1@earthlink.net>
* html/notes.htm: Typo fix.
From: John Stone <johns@ks.uiuc.edu>
ChangeLog:
Typo...
Harlan Stenn [Wed, 30 May 2001 19:21:49 +0000 (19:21 -0000)]
ChangeLog, README.cvs, ntp_loopfilter.c:
* README.cvs: More updates and cleanup.
* ntpd/ntp_loopfilter.c (loop_config):
Check against STA_NANO instead of (NTP_API > 3) to catch kernels
that were rolled while the spec was evolving.
From: John.Hay@icomtek.csir.co.za
* README.cvs: Note that we want to check out NTP into a clean
subdir.
Reported by jrd@cc.usu.edu (Joe Doupnik)
Harlan Stenn [Mon, 28 May 2001 01:49:17 +0000 (01:49 -0000)]
ChangeLog, ntp_refclock.h:
* include/ntp_refclock.h: Median Filter (SAMPLE - macro) - change
to use most recent MAXSTAGE entries when the filter overflows (ie
driver poking say once per second with poll > MAXSTAGE) rather
than blocking after MAXSTAGE entries (turf oldest rather than turf
most recent).
From: John Woolner <vk1et@tpg.com.au>
ChangeLog, aclocal.m4, configure, configure.in:
* configure.in: 4.0.99k40
Harlan Stenn [Mon, 28 May 2001 00:31:21 +0000 (00:31 -0000)]
ChangeLog, refclock_true.c:
* ntpd/refclock_true.c:
a. Don't cream pp->a_lastcode when we get a <cr><lf> pair
b. Fix up pp->leap handling to work correctly
c. clear CEVNT_BADTIME etc warnings when we get good clock
CEVNT_NOMINAL.
From: John Woolner <vk1et@tpg.com.au>
Harlan Stenn [Sun, 27 May 2001 19:42:52 +0000 (19:42 -0000)]
ChangeLog, pcl720.h, refclock_true.c:
* ntpd/refclock_true.c:
* kernel/sys/pcl720.h:
Add support for the XL clock to refclock_true.c
From: Paul A Vixie <vixie@mfnx.net>
Harlan Stenn [Sun, 27 May 2001 01:46:30 +0000 (01:46 -0000)]
ChangeLog, aclocal.m4, configure, configure.in:
* configure.in (ac_cv_func_ctty_for_f_setown): BSDI3 needs a ctty
for F_SETOWN, too.
From: Paul A Vixie <vixie@mfnx.net>
Harlan Stenn [Mon, 21 May 2001 18:48:45 +0000 (18:48 -0000)]
ChangeLog, ntp_syscall.h, ntp_loopfilter.c, ntp_proto.c:
* ntpd/ntp_proto.c (clock_filter): Huff-n-Puff and Popcorn
improvements.
* ntpd/ntp_loopfilter.c (local_clock): Debug cleanup
From: Dave Mills.
* include/ntp_syscall.h (ntp_gettime): Updated patch from Ulrich.
My original attempt was not backwards compatible.
Harlan Stenn [Thu, 17 May 2001 05:38:23 +0000 (05:38 -0000)]
Many files:
* configure.in: 4.0.99k37
* ntpd/ntp_proto.c (clock_filter): Lose "off", xtemp and ytemp,
and some obsoleted calculations. Set the peer->offset and
peer->delay from the filter stages.
* ntpd/ntp_loopfilter.c: Comment/document improvements.
(local_clock): correct the offset by one-half the difference
between the sample delay and minimum delay. Lose "mu" from the
debug message.
From: Dave Mills.
Harlan Stenn [Fri, 11 May 2001 05:31:03 +0000 (05:31 -0000)]
ChangeLog, ntpd.h, ntp_loopfilter.c, ntp_util.c:
* ntpd/ntp_util.c (record_loop_stats): values are now passed in.
* ntpd/ntp_loopfilter.c (local_clock): pass the values to
record_loop_stats().
* include/ntpd.h: Pass the parameters in to record_loop_stats().
With the discipline loop opened (disable ntp) the local clock
updates were not being sent to loopstats. That now is.
From: Dave Mills.
Harlan Stenn [Tue, 8 May 2001 17:18:43 +0000 (17:18 -0000)]
ChangeLog, aclocal.m4, configure, configure.in, ntp_loopfilter.c:
* configure.in: 4.0.99k32
* ntpd/ntp_loopfilter.c: rstclock now takes a 3rd argument, the
last offset.
(init_loopfilter): Use it.
(local_clock): Use it. Clean up the code.
(loop_config): Use it.
(rstclock): Implement it. Clean up the code.
From Dave Mills.
Harlan Stenn [Mon, 7 May 2001 02:00:29 +0000 (02:00 -0000)]
ChangeLog, ntp_loopfilter.c, ntp_proto.c, ntpdc_ops.c:
* ntpdc/ntpdc_ops.c (sysstats): That's 'bad packet format'
(instead of '... length'), and 'packets rejected' (instead of
'limitation rejects'.
* ntpd/ntp_proto.c (receive): PUBKEY fixes. Move KoD stuff to
process_packet().
(process_packet): Move KoD stuff here...
(peer_clear): Unspec the stratum, too.
(clock_filter): Don't update peer->epoch here. Fix the filter
test when checking the epoch.
(fast_xmit): Send back STRATUM_UNSPEC on a KoD packet.
(init_proto): Initialize sys_jitter.
* ntpd/ntp_loopfilter.c: rstclock() takes 2 parameters now.
(init_loopfilter): Use it...
(local_clock): Ditto, and change the "mu" calculation. Improve
the jitter test in S_SYNC. Use peer->epoch (not current_time) to
update the last_time. Update debug info.
(rstclock): 2nd arg - the epoch to use. Use it.
(loop_config): update call to rstclock.
From: Dave Mills.
ChangeLog, ntp_proto.c, ntpq.c:
* ntpq/ntpq.c (tstflags): 11 now.
From: John Cochran <jdc@fiawol.org>
* ntpd/ntp_proto.c (receive): KoD updates. Improve the comments.
Lose the AM_PROCPKT restrictions test.
(peer_xmit): Check/report on no encryption key in packet.
(fast_xmit): Use peer_xmit's new packet length check code.
From Dave Mills.
ChangeLog, ntp_proto.c, ntpq.c:
* ntpq/ntpq.c: Reorder and add some TEST flag bits.
* ntpd/ntp_proto.c (transmit): Also bail if access denied.
(receive): Lose RES_DEMOBILIZE and (some?) RES_DONTSERVE and
RES_LIMITIED stuff. Update Kiss-Of-Death (KoD) docs.
Call fast_xmit with new 3rd parameter (restrict_mask).
Before checking for an authentic packet, check the restrict_mask
for RES_{DONTSERVE,LIMITED,NOPEER}.
Check restrictions in AM_PROCPKT case.
(peer_clear): Don't lose the stratum if the peer->flags don't
indicate FLAG_REFCLOCK.
(fast_xmit): Take restrict mask as a new argument, and handle
KoD. Reorder some code.
From: Dave Mills.
Many files:
* ntpdc/ntpdc_ops.c: restrict/unrestrict support for version and
demobilize. Implement demobilze.
* ntpd/ntp_proto.c (receive): Improve version testing, including
RES_DEMOBILIZE support.
(fast_xmit): Patches to kiss-of-death packet.
* ntpd/ntp_loopfilter.c (local_clock): S_SYNC case now also checks
abs(clock_offset) against CLOCK_PGATE*sys_jitter.
* ntpd/ntp_config.c: CONF_RES_DEMOBILIZE/demobilize support.
* include/ntp_config.h (CONF_RES_DEMOBILIZE): Added.
* include/ntp.h (RES_DEMOBILIZE): Added.
From Dave Mills.
ChangeLog, ntp.h, ntp_config.h, ntp_proto.c, Makefile.in:
* ntpd/ntp_proto.c (receive): Keep track of packet versions.
Implement RES_LIMITED.
* include/ntp_config.h (CONF_RES_LIMITED):
* include/ntp.h (RES_LIMITED): Leave the bits in the original
order.
From Dave Mills.
ChangeLog, miscopt.htm, ntp.h, ntp_config.h, ntp_loopfilter.c, ntp_proto.c:
* html/miscopt.htm: Document the "allan" tinker variable.
* ntpd/ntp_proto.c (clock_filter): Update comments. Lose etemp;
we now use allan_xpt for this.
* ntpd/ntp_loopfilter.c: Added allan_xpt as a tinker variable.
Reorganize variables and improve comments.
(local_clock): Improve comments, use (new) allan_xpt instead of
CLOCK_ALLAN. Fix test in S_SYNC state. Update debug info.
(rstclock): No longer force allan_xpt to CVLOCK_ALLAN in S_FREQ,
S_SYNC, or default case.
(loop_config): Document dangerous tinker variables, and add
LOOP_ALLAN to the list.
* include/ntp_config.h (CONF_CLOCK_ALLAN): Added.
* include/ntp.h (LOOP_ALLAN): Added.
Allan intercept fixes from Dave Mills.
ChangeLog, aclocal.m4, configure, configure.in:
* configure.in: 4.0.99k29
projects / thirdparty / ntp.git / log
