Re: NTP nanokernel support (experimental)
Pierre, The NTPv4 driver interface already implements a trimmed-mean filter, which cleans up a good deal of jitter as it is. The IRIG and CHU drivers do a lot more signal processing, yielding generally low jitter in the tens of microseconds. Deglitching and filtering noisy time sampls is somewhat an art form, but the methods are not hard to implement. Dave To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
On Sat, Mar 06, 1999 at 08:00:53PM +0100, Ollivier Robert wrote: The two outputs I sent were with 4.0.90f. When I run 4.0.92c, ntpd is not able to get any accurate data from the device whereas 4.0.90f does. I get lots of these in /var/log/messages and it doesn't sync at all. -=-=- Mar 6 14:02:25 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 3 bits Mar 6 14:02:29 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits It's typical from bad parity setting on your serial port. Try a stty on that port; I bet it will show that PARENB is set. Unset it and things should go back to normal. Maybe it is a problem with 4.0.92c... Yes, it's a problem with most of the ntpd 4.0.9x series. There's absolutely no reason why you should enable PARENB for a raw DCF77 driver; yet that's what ntpd's configure does, at least under FreeBSD. I sent a bug report to the ntpd team a while ago but haven't heard from them. -- Pierre Beyssacp...@fasterix.frmug.org p...@fasterix.freenix.org {Free,Net,Open}BSD, Linux : il y a moins bien, mais c'est plus cher Free domains: http://www.eu.org/ or mail dns-mana...@eu.org To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
According to Pierre Beyssac: absolutely no reason why you should enable PARENB for a raw DCF77 driver; yet that's what ntpd's configure does, at least under FreeBSD. Yes, it seems to be that again. I thought I had fixed it in config.cache but it seems not. It is working now. remote refid st t when poll reach delay offset jitter == 127.127.1.0 127.127.1.0 10 - 12 6410.0000.000 0.000 *127.127.8.0 .DCFa. 0 - 11 6430.000 32.068 0.514 224.0.1.1 0.0.0.0 16 u- 6400.0000.000 4000.00 A few more peerstats for you Poul-Henning, now with both your diffs and 4.0.92c. 51244 65536.259 127.127.8.0 96c4 0.032788333 0.0 0.0 0.002593271 51244 65537.261 127.127.8.0 96c4 0.032699250 0.0 0.0 0.002148683 51244 65538.259 127.127.8.0 96c4 0.032702500 0.0 0.0 0.001667484 51244 65539.253 127.127.8.0 96c4 0.032702500 0.0 0.0 0.000871368 51244 65540.259 127.127.8.0 96c4 0.032678083 0.0 0.0 0.000525512 51244 65541.259 127.127.8.0 96c4 0.032167917 0.0 0.0 0.000432877 51244 65542.639 127.127.8.0 96c4 0.032068250 0.0 0.0 0.000513906 The following diff should fix the ignpar/parenb problem. --- configure.in.oldSun Mar 7 19:11:41 1999 +++ configure.inSun Mar 7 19:11:55 1999 @@ -1055,6 +1055,9 @@ i?86-*-linux*) ans=yes ;; + *-*-freebsd*) +ans=yes +;; mips-sgi-irix*) ans=yes ;; -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- robe...@keltia.freenix.fr FreeBSD keltia.freenix.fr 4.0-CURRENT #70: Sat Feb 27 09:43:08 CET 1999 To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
Ollivier, For performance monitoring with the nanokernel and PPS source, see the peerstats and grope for 127.0.0.1 with grep. When the daemon has handed off PPS to the kernel, the radio timecode is used only to mean the transmitter is still on the air. Dave To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
In message 199903071351.aa20...@huey.udel.edu, Dave Mills writes: Ollivier, For performance monitoring with the nanokernel and PPS source, see the peerstats and grope for 127.0.0.1 with grep. When the daemon has handed off PPS to the kernel, the radio timecode is used only to mean the transmitter is still on the air. I don't think Ollivier is doing PPP/hardpps() yet, at least I have not given him the semi-magic code needed for it :-) I wouldn't recommend trying it either, he is bound to have a 1 msec jitter on the DCF77 waves at his place, and that is a lousy diet for hardpps(). Poul-Henning PS: I have updated my rover page, you can see the performance of my machine with the nanokernel-PLL patch, 4.92c and a UT+ Oncore on http://phk.freebsd.dk/rover.html (This isn't using hardpps() either!) -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
On Sun, Mar 07, 1999 at 08:09:50PM +0100, Poul-Henning Kamp wrote: I don't think Ollivier is doing PPP/hardpps() yet, at least I have not given him the semi-magic code needed for it :-) I wouldn't recommend trying it either, he is bound to have a 1 msec jitter on the DCF77 waves at his place, and that is a lousy diet for hardpps(). I agree; jitter with a low-cost DCF77 receiver is even more than that (5 to 10 ms). I tend to believe it's partly due to how the AM signal is demodulated and not that much from the location (Paris is not that far from Frankfurt after all). Maybe a lower jitter could be obtained by averaging 10 or 100 samples, I suppose that's how high-quality receivers work. This might be done in the ntpd driver. -- Pierre Beyssacp...@fasterix.frmug.org p...@fasterix.freenix.org {Free,Net,Open}BSD, Linux : il y a moins bien, mais c'est plus cher Free domains: http://www.eu.org/ or mail dns-mana...@eu.org To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
According to Poul-Henning Kamp: Please send comments and observations, both positive AND negative. I can't get my DCF77 device to synchronize with both the kernel diff applied and 4.0.92c. With the older 4.0.90f I'm able to sync. remote refid st t when poll reach delay offset jitter == LOCAL(0)LOCAL(0)10 - 24 64 3770.0000.000 0.000 *GENERIC(0) .DCFa. 0 - 20 64 3570.000 10.664 0.617 ... remote refid st t when poll reach delay offset jitter == 127.127.1.0 127.127.1.0 10 - 57 64 3770.0000.000 0.000 *127.127.8.0 .DCFa. 0 -- 64 3570.000 15.090 0.193 I am very interested in seing some peerstats files for refclocks running this code. 51243 56309.281 127.127.8.0 96f5 0.011960917 0.0 0.0 0.002261763 51243 56310.281 127.127.8.0 96f5 0.011612083 0.0 0.0 0.001471172 51243 56311.291 127.127.8.0 96f5 0.011183750 0.0 0.0 0.000786827 51243 56312.285 127.127.8.0 96f5 0.011175333 0.0 0.0 0.000718346 51243 56356.278 127.127.8.0 96f5 0.014832500 0.0 0.0 0.002910501 51243 56357.274 127.127.8.0 96f5 0.015144083 0.0 0.0 0.003002719 51243 56358.278 127.127.8.0 96f5 0.015144583 0.0 0.0 0.002788572 51243 56359.274 127.127.8.0 96f5 0.015447583 0.0 0.0 0.002825388 51243 56360.274 127.127.8.0 96f5 0.015403750 0.0 0.0 0.002513153 51243 56361.278 127.127.8.0 96f5 0.015087917 0.0 0.0 0.001963783 51243 56362.285 127.127.8.0 96f5 0.015087917 0.0 0.0 0.001396828 51243 56363.278 127.127.8.0 96f5 0.01509 0.0 0.0 0.000193172 ... 51243 56494.272 127.127.8.0 96f5 0.014399417 0.0 0.0 0.000286762 51243 56495.277 127.127.8.0 96f5 0.014374750 0.0 0.0 0.000279898 51243 56496.277 127.127.8.0 96f5 0.014352500 0.0 0.0 0.000268971 51243 56497.269 127.127.8.0 96f5 0.014813083 0.0 0.0 0.000284027 51243 56497.652 127.127.8.0 96f5 0.014813083 0.0 0.0 0.000281249 -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- robe...@keltia.freenix.fr FreeBSD keltia.freenix.fr 4.0-CURRENT #70: Sat Feb 27 09:43:08 CET 1999 To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
In message 19990306164238.a29...@keltia.freenix.fr, Ollivier Robert writes: According to Poul-Henning Kamp: Please send comments and observations, both positive AND negative. I can't get my DCF77 device to synchronize with both the kernel diff applied and 4.0.92c. With the older 4.0.90f I'm able to sync. It looks synchronized to me, it just looks like it hasn't swung in yet ? I find it very sluggish in reacting too, but I'm not sure yet if it is a feature or a bug. -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
According to Poul-Henning Kamp: It looks synchronized to me, it just looks like it hasn't swung in yet ? The two outputs I sent were with 4.0.90f. When I run 4.0.92c, ntpd is not able to get any accurate data from the device whereas 4.0.90f does. I get lots of these in /var/log/messages and it doesn't sync at all. -=-=- Mar 6 14:02:25 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 3 bits Mar 6 14:02:29 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:02:38 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:02:38 tara ntpd[7600]: PARSE receiver #0: FAILED TIMECODE: - (check receiver configuration / cableling) Mar 6 14:02:50 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:02:52 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:03:35 tara ntpd[7600]: PARSE receiver #0: no data from device within poll interval (check receiver / cableling) -=-=- Maybe it is a problem with 4.0.92c... -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- robe...@keltia.freenix.fr FreeBSD keltia.freenix.fr 4.0-CURRENT #70: Sat Feb 27 09:43:08 CET 1999 To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
This looks like a reception problem. You must be pretty far out in the reception area of DCF77 (just like me), and will probably find that whenever a sunrise or sunset is in the area reception sux. Try to see if it does better in daylight... I don't know the state of the DCF77/parse stuff in v4 at all, I have not tried to use it, I'm relying on my GPS for now. Poul-Henning In message 19990306200053.b30...@keltia.freenix.fr, Ollivier Robert writes: According to Poul-Henning Kamp: It looks synchronized to me, it just looks like it hasn't swung in yet ? The two outputs I sent were with 4.0.90f. When I run 4.0.92c, ntpd is not able to get any accurate data from the device whereas 4.0.90f does. I get lots of these in /var/log/messages and it doesn't sync at all. -=-=- Mar 6 14:02:25 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 3 bits Mar 6 14:02:29 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:02:38 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:02:38 tara ntpd[7600]: PARSE receiver #0: FAILED TIMECODE: - (check receiver configuration / cableling) Mar 6 14:02:50 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:02:52 tara ntpd[7600]: parse: convert_rawdcf: INCOMPLETE DATA - time code only has 2 bits Mar 6 14:03:35 tara ntpd[7600]: PARSE receiver #0: no data from device within poll interval (check receiver / cableling) -=-=- Maybe it is a problem with 4.0.92c... -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- robe...@keltia.freenix.fr FreeBSD keltia.freenix.fr 4.0-CURRENT #70: Sat Feb 27 09:43:08 CET 1999 To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: NTP nanokernel support (experimental)
I just got a new set of PARSE patches from Frank Kardel, and I'm waiting for Dave Mills to give me the go-ahead to commit them. Dave is chasing some possible bugs and wanted a stable codebase for a while. H To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
NTP nanokernel support (experimental)
http://www.freebsd.org/~phk/x.nanokernel.gz contains patches relative to -current to implement the new nanokernel PLL from Dave Mills. This only works with a ntpd v4, suggest 4.92c or better from: ftp://ftp.eecis.udel.edu/pub/ntp It does still not support the hardpps()/PPS_SYNC support, that will come later. Make sure to configure/make your ntpd-4 from scratch to gain advantage of the new resolution, simple recompilation may not be enough. (You also need to get timex.h from /sys/sys to /usr/include/sys. Make world recommended). You should now be truly limited to hardware performance (xtal tempco and interrupt jitter). Please send comments and observations, both positive AND negative. I am very interested in seing some peerstats files for refclocks running this code. Poul-Henning PS: See also http://www.freebsd.org/~phk/timechan.gif -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message