it's really going to become a time mega-patch. I have fixed all the
obvious problems in snapshot #2, and now I brought you snapshot #3
for the weekend.
Some tests indicated (on a Pentium-100 using TSC) that time
occasionally runs backward in the fractional second (in 15 of 2000000
samples). I'll have to investigate that...
Using ntp-4.0.91f required some patches, because the "struct timex"
lacks some magic from Dave Mills. Maybe even, that ntp version still
has some bugs. I don't want to decide too early. PPS has significant
jitter, too.
Maybe have a look. Compilation problems for some networking stuff
should be fixed now, and the new MOD_* bits can be really used now.
The patch is a 100%-patch that is based on 2.2.1, but should apply
with little fuzz to 2.2.2 too.
URLs: PPSkit-0.6-i386-pre3.diff.gz is on master.kernel.org in
/pub/linux/daemons/ntp/PPS and on ftp://pcphy4.physik.uni-
regensburg.de/pub/wiu09524/PPS, and maybe on a mirror close to you...
Regards,
Ulrich Windl
P.S:
pll offset: 31457 us # possibly ns
pll frequency: -0.061 ppm
maximum error: 20321 us # possibly ns
estimated error: 2925 us # possibly ns
status: 2107
pll time constant: 6
precision: 1 us
frequency tolerance: 496 ppm
pps frequency: -0.061 ppm
pps stability: 0.102 ppm
pps jitter: 25881 us # rather surely ns
calibration interval: 256 s
calibration cycles: 50
jitter exceeded: 15
stability exceeded: 0
calibration errors: 3
-
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/