Date:	Fri, 12 Mar 1999 08:41:50 +0100 (CET)
From:	Ulrich Windl <wiu09524@pc3103.klinik.uni-regensburg.de>
To:	linux-kernel@vger.rutgers.edu
Subject: announce(extension patch, 2.2.3): PPSkit-0.5.1

Here's a copy of the news posting:

Ulrich


Newsgroups: comp.protocols.time.ntp
Subject: announce(Linux-2.2.3): PPSkit-0.5.1
Reply-to: Ulrich Windl <+DELETE_THIS+Ulrich.Windl@rz.uni-regensburg.de>
Distribution: local
--text follows this line--
I have decided to put out another "defect repair" release of
PPSkit-0.5 for the stable Linux kernel 2.2 while PPSkit-0.6 is
growing.  This caused some extra effort, because it's now development
on a side branch, but anyway:

I have not had a lot of time left to test it, maybe I have broken the
PPSFREQ discipline. At least with ntpd4.0.91 there was not too much
success. PPSTIME seems to work however.

I also merged in the current snapshot of enable_pps.c which has some
new experimental options, like "-j" for a standalone jitter-test.

Other notable changes affect the famous "STA_FLL|STA_FREQHOLD" bug:
The combination does not make any sense, and in fact is not allowed,
but xntpd3 uses it.  As Dave Mills states "It's not a problem for my
implementation, so your implementation is broken", I ignore that error
in that case (I also provided a patch against the broken xntp code,
but that was rejected by Dave).

Another fix makes the TIME_ERROR return value less "sticky" than
before.  Finally ntpd4 will no longer refuse to use the kernel PLL.

Finally, heres the part from the NEWS, followed by the LSM entry (note
the changed primary site):

                            NEWS for 0.5.1

+ Updated ``enable_pps.c'' (from PPSkit-0.6-pre6).  Removed
  ``madtime'' patch.

+ Modified ``hardpps()'' to receive the required ``offset in
  microseconds since last tick'' that we don't really have.  Added a
  workaround.

+ Merged in PPSkit-0.5.0 with fix.  Added workaround for case where
  both, ``STA_FREQHOLD'' and ``STA_FLL'' are set (illegal, but used in
  xntpd3).  Unimplemented mode bits no longer cause an ``EINVAL'' in
  ``adjtimex()'', because ntpd4 refuses to use the kernel PLL if
  ``STA_NANO'' causes ``EINVAL'' to be returned.

+ Updated for Linux-2.2.3.  Added more NTP time fixes to base kernel.
  ``TIME_ERROR'' is no longer sticky, but computed each time.

Begin3
Title:          PPSkit
Version:        0.5.1
Entered-date:   1999-03-11
Description:    NTP (RFC1305) and PPS support (RFC1589) patches for
                Linux 2.2.3+ (kernel patch, documentation, test programs)
Keywords:       NTP, PPS, kernel, RFC1305, RFC1589, xntp, time, clock,
                pulse-per-second, synchronization, calibration, adjtimex
Author:         Ulrich.Windl@rz.uni-regensburg.de (Ulrich Windl)
Maintained-by:  Ulrich.Windl@rz.uni-regensburg.de (Ulrich Windl)
Primary-site:   linux.kernel.org /pub/linux/daemons/ntp/PPS
                46kB PPSkit-0.5.1.tar.gz
Alternate-site: 
Original-site:  pcphy4.physik.uni-regensburg.de /pub/wiu09524/PPS
                PPSkit-0.5.1.tar.gz
Platform:       RS232-compatible precision pulse-per-second with an error
                less than 100PPM (0.0001%, 0.1ms)
Copying-policy: GPL, but payment welcome!
End



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/