[LinuxPPS] ntp: add hardpps implementation (2)

[Remco dB]

OK, 

I just decided to switch back to a non hardpps system here.

There are too many anomalities which, in my opinion, have to be (re)solved.

E.g. on every system with hardpps I observed, both the pll frequency and the 
pps frequency are equal, this is not the case in LinuxhardPPS (yet).
Then I've still the magical 127/128 us offset with my setup using LinuxhardPPS. 

(Linux 2.4-NANO, ntp.oma.be):  
ntpdc -c kern ntp.oma.be
pll offset:           -1.3e-06 s
pll frequency:        38.830 ppm
maximum error:        0.002003 s
estimated error:      2.3e-05 s
status:               2107  pll ppsfreq ppstime ppssignal nano
pll time constant:    6
precision:            2.634e-06 s
frequency tolerance:  496 ppm
pps frequency:        38.830 ppm
pps stability:        0.104 ppm
pps jitter:           6.073e-06 s
calibration interval: 256 s
calibration cycles:   335426
jitter exceeded:      1865764
stability exceeded:   17380
calibration errors:   0

(FreeBSD-7.0-release, ntp.remco.org): 
ntpdc -c kern time
pll offset:           5.13e-07 s
pll frequency:        -49.802 ppm
maximum error:        0.004235 s
estimated error:      1e-06 s
status:               2107  pll ppsfreq ppstime ppssignal nano
pll time constant:    4
precision:            1e-09 s
frequency tolerance:  496 ppm
pps frequency:        -49.802 ppm
pps stability:        0.004 ppm
pps jitter:           1.998e-06 s
calibration interval: 256 s
calibration cycles:   127704
jitter exceeded:      47111
stability exceeded:   15
calibration errors:   4

(2.6.35.7 with hardpps, pindarots.xs4all.nl , Udo's system):
ntpdc -c kern pindarots
pll offset:           0 s
pll frequency:        3.714 ppm
maximum error:        0.005739 s
estimated error:      4e-06 s
status:               2107  pll ppsfreq ppstime ppssignal nano
pll time constant:    4
precision:            1e-09 s
frequency tolerance:  500 ppm
pps frequency:        3.709 ppm
pps stability:        0.026 ppm
pps jitter:           5.747e-06 s
calibration interval: 256 s
calibration cycles:   753
jitter exceeded:      8
stability exceeded:   0
calibration errors:   0

It is also 'suspicious' that the offsets of LinuxhardPPS are always positive
whereas one should expect that the offset jitters around zero (+/-).

After switching back, my system yields these kind of values: 

(2.6.34 LinuxPPS (not hardpps), ipv6.remco.org, uses the PPS signal of 
ntp.remco.org): 
ntpdc -c kern he
pll offset:           -8.5e-08 s
pll frequency:        157.630 ppm
maximum error:        0.007735 s
estimated error:      0 s
status:               2007  pll ppsfreq ppstime nano
pll time constant:    4
precision:            1e-09 s
frequency tolerance:  500 ppm

ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
oPPS(0)          .PPS.            0 l    7   16  377    0.000   -0.001   0.002
*freebsd         .GPS.            1 u   54   64  377    0.176    0.006   0.023
+lithium         .DCFa.           1 u   59   64  377    0.134   -0.963   0.027
 ntp2            .INIT.          16 u    - 1024    0    0.000    0.000   0.000
+ptbtime1.ptb.de .PTB.            1 u   54   64  377   46.119   -0.158   0.244


Remco