Re: time warps, I despair

Andreas Koppenhoefer (koppenas@koppenas.dialup.informatik.uni-stuttgart.de)
04 Nov 1996 12:33:44 +0100


-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "IM" == Ingo Molnar <mingo@pc5829.hil.siemens.at> writes:
In article <Pine.LNX.3.93.961101175301.3565B-100000@pc5829.hil.siemens.at>
Ingo Molnar <mingo@pc5829.hil.siemens.at> writes:

IM> On Thu, 31 Oct 1996, Ulrich Windl wrote:
IM>
>> So what's the effect? From time to time the clock offset jumps for
>> some amount always less or equal to what's worth one tick (i.e.
>> offset jumps around from -5ms to 5ms when synchronized). [...]
IM>
IM> this could be caused by a 'lost jiffy'. Lost jiffy is a timer hardware
IM> interrupt that misses us completely. Under normal hardware circumstances
IM> (no APM, etc.), this is only possible if there is a cli()/sti() pair that
IM> lasts for more than 10 millisec.
IM>
IM> You can totally eliminate the Pentium stuff by turning APM support on in
IM> the kernel (or commenting out the last few lines in time.c), thus the PIC
IM> timer is used as high precision timing offset source.

Sorry, but I think this is NOT only pentium related - It might be a more
general problem!

Ulrich's graphs show spike interval of 10-12 minutes if I read these graphs
correctly. On my system I can see a similar spike interval of about
120 minutes.

My system is a i486-DX2/66 running on a ASUS-SP3G motherboard with
xntpd-3.4z and linux 2.0.14. I see this anomaly for about two years
now (since kernel 1.1.3x or so). Time source is a cheap DCF77 receiver
on a serial port. Before you ask: There are no processes which runs every
even hour like cron or so.

- - Andreas

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2i
Comment: Processed by Mailcrypt 3.3, an Emacs/PGP interface

iQBVAwUBMn3UkEVdjNvyndGpAQFT0gH/fsq6GjEse8+XNH5pZ2dnqrN1vS9Bivmu
4e+YIrTG98YNLX8oLcU3HBkmFGUel2++C+RvQdIIjr3gRYtZyj2wzQ==
=UW9t
-----END PGP SIGNATURE-----

-- 
Andreas Koppenhoefer, Student der Universitaet Stuttgart, BR Deutschland
Franz-Schubert-Str. 2, 97616 Bad Neustadt; Phone +49 9771 7943, Fax/Data 7947
<akoppenhoefer@schweinfurt.netsurf.de>,<koppenas@informatik.uni-stuttgart.de>
<Andreas.Koppenhoefer@studbox.uni-stuttgart.de>; PGP-Key available at
URL <http://www.informatik.uni-stuttgart.de/menschen/ako/impressum.html>