Precision Time Protocol – Windows 10 implementation
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower. I've heard that one of the
latest versions of Windows 10 provides support for PTP protocol (IEEE
1588v2). Hence my questions:
- Did anybody try using it on Windows 10 already? From what I can tell
right now, there's only a demonstration web site showing it works when
using Ubuntu Linux as a virtual machine [1]. Or should I wait a bit for
more robust implementation? - Do I guess correctly that current implementation of PTP works only on
Ethernet? Some of the computers are on wlan. - Is there any other way of making the clients' time offset lower?
I've already tried setting PTP on the Raspberry Pi 3B+. It needed a kernel
recompilation with a patch I found here: [2], but I'm not sure it works
correctly.
As you can see on the screenshot below, PTP appears here as a server, but
I'm not sure if it's used at all.
[image: status16.PNG]
Ptp4l status shows this:
pi@zegar:~ $ systemctl status ptp4l
● ptp4l.service - Precision Time Protocol (PTP) service
Loaded: loaded (/lib/systemd/system/ptp4l.service; enabled; vendor
preset: enabled)
Active: active (running) since Wed 2019-08-07 12:32:28 CEST; 24h ago
Docs: man:ptp4l
Main PID: 323 (ptp4l)
Tasks: 1 (limit: 4915)
CGroup: /system.slice/ptp4l.service
└─323 /usr/sbin/ptp4l -f /etc/linuxptp/ptp4l.conf -i eth0
Aug 07 12:32:31 zegar ptp4l[323]: ptp4l[6.631]: port 1: link up
Aug 07 12:32:31 zegar ptp4l[323]: ptp4l[6.631]: port 1: FAULTY to LISTENING
on FAULT_CLEARED
Aug 07 12:32:31 zegar ptp4l[323]: [6.631] port 1: link up
Aug 07 12:32:31 zegar ptp4l[323]: [6.631] port 1: FAULTY to LISTENING on
FAULT_CLEARED
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: port 1: LISTENING to
MASTER on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: selected best master clock
b827eb.fffe.006e65
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: assuming the grand master
role
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] port 1: LISTENING to MASTER on
ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] selected best master clock
b827eb.fffe.006e65
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] assuming the grand master role
Thanks in advance for any suggestions!
Adam Kumiszcza
[1]
https://techcommunity.microsoft.com/t5/Networking-Blog/Windows-Subsystem-for-Linux-for-testing-Windows-10-PTP-Client/ba-p/389181
[2] https://github.com/twteamware/raspberrypi-ptp/issues/2
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower.
[]
Thanks in advance for any suggestions!
Adam Kumiszcza
---===========
Adam,
You could feed the PPS signal to the Windows PCs as one way of making the
performance better:
https://www.satsignal.eu/ntp/setup.html
Performance:
https://www.satsignal.eu/mrtg/performance_ntp.php#windows-stratum-1
Cheers,
David
SatSignal Software - Quality software for you
Web: http://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv
Hi Adam,
I've only ever used hardware PTP servers (and I'm far from an expert), but PTP is pretty sensitive to latency. It's common for PTP to deployed on dedicated networks to help keep jitter to a minimum (I think I remember one place doing it on Infiniband but I might be wrong). The challenge I think you're facing is that although the network card on the Pi 3b+ supports gigabit Ethernet, it's still connected via a USB2 hub. This introduces a lot of latency. For many applications (and general use) it doesn't hurt performance too much (you'll still be lucky to push ~330Mb/s through it), but when you care about latency it will cause you problems. I didn't test extensively but on the Pi3 (not b+), I would often see millisecond level latencies.
For the WiFi question, well, that has even worse latency than the wired (with a number of other issues introduced) so I don't think you'll get much mileage with it.
The Pi isn't bad as an NTP server really, especially if your alternative is a server on the Internet (the uBlox M8Q performs really nicely when compared to the Microsemi S650, an industrial grade GNSS time source), but I reckon you're pushing the envelope with what NTP can do on that hardware, and the Pi just doesn't have the hardware to deliver PTP.
I would love to hear differently though!
Cheers,
Pete
----- Original Message -----
From: "Adam Kumiszcza" akumiszcza@gmail.com
To: "Discussion of precise time and frequency measurement" time-nuts@lists.febo.com
Sent: Thursday, 8 August, 2019 19:26:11
Subject: [time-nuts] Precision Time Protocol – Windows 10 implementation
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower. I've heard that one of the
latest versions of Windows 10 provides support for PTP protocol (IEEE
1588v2). Hence my questions:
- Did anybody try using it on Windows 10 already? From what I can tell
right now, there's only a demonstration web site showing it works when
using Ubuntu Linux as a virtual machine [1]. Or should I wait a bit for
more robust implementation? - Do I guess correctly that current implementation of PTP works only on
Ethernet? Some of the computers are on wlan. - Is there any other way of making the clients' time offset lower?
I've already tried setting PTP on the Raspberry Pi 3B+. It needed a kernel
recompilation with a patch I found here: [2], but I'm not sure it works
correctly.
As you can see on the screenshot below, PTP appears here as a server, but
I'm not sure if it's used at all.
[image: status16.PNG]
Ptp4l status shows this:
pi@zegar:~ $ systemctl status ptp4l
● ptp4l.service - Precision Time Protocol (PTP) service
Loaded: loaded (/lib/systemd/system/ptp4l.service; enabled; vendor
preset: enabled)
Active: active (running) since Wed 2019-08-07 12:32:28 CEST; 24h ago
Docs: man:ptp4l
Main PID: 323 (ptp4l)
Tasks: 1 (limit: 4915)
CGroup: /system.slice/ptp4l.service
└─323 /usr/sbin/ptp4l -f /etc/linuxptp/ptp4l.conf -i eth0
Aug 07 12:32:31 zegar ptp4l[323]: ptp4l[6.631]: port 1: link up
Aug 07 12:32:31 zegar ptp4l[323]: ptp4l[6.631]: port 1: FAULTY to LISTENING
on FAULT_CLEARED
Aug 07 12:32:31 zegar ptp4l[323]: [6.631] port 1: link up
Aug 07 12:32:31 zegar ptp4l[323]: [6.631] port 1: FAULTY to LISTENING on
FAULT_CLEARED
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: port 1: LISTENING to
MASTER on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: selected best master clock
b827eb.fffe.006e65
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: assuming the grand master
role
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] port 1: LISTENING to MASTER on
ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] selected best master clock
b827eb.fffe.006e65
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] assuming the grand master role
Thanks in advance for any suggestions!
Adam Kumiszcza
[1]
https://techcommunity.microsoft.com/t5/Networking-Blog/Windows-Subsystem-for-Linux-for-testing-Windows-10-PTP-Client/ba-p/389181
[2] https://github.com/twteamware/raspberrypi-ptp/issues/2
--
This message has been scanned for viruses and dangerous content by
E.F.A. Project, and is believed to be clean.
Click here to report this message as spam.
http://efa.membrey.hk/cgi-bin/learn-msg.cgi?id=7E44F20569.A93C5&token=8c56258c47d97b759f6153822c804b1a
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
On Thu, Aug 8, 2019 at 8:07 AM Adam Kumiszcza akumiszcza@gmail.com wrote:
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower. I've heard that one of the
latest versions of Windows 10 provides support for PTP protocol (IEEE
1588v2). Hence my questions:
- Did anybody try using it on Windows 10 already? From what I can tell
right now, there's only a demonstration web site showing it works when
using Ubuntu Linux as a virtual machine [1]. Or should I wait a bit for
more robust implementation?
EndRun Technologies
(https://endruntechnologies.com/products/grandmaster-clocks/ptp-slaves)
make mention of a Windows PTP slave - not sure if it's the same
version you have found. It does also make mention of the fact that
precision is limited on Windows depending on version although from
what I gather from the satsignal.eu that precious on Windows improved
with 7 or later.
- Do I guess correctly that current implementation of PTP works only on
Ethernet? Some of the computers are on wlan.
For deterministic timing I think it has to be Ethernet and as outlined
in their white paper (https://endruntechnologies.com/pdf/PTP-1588.pdf)
it works best if you have hardware timestamping in the NICs at each
end and IEEE1588 compatible network switches (and even when the claim
to support it, apparently they don't always do it properly:
https://www.linkedin.com/pulse/ptp-support-high-end-ethernet-switches-heiko-gerstung).
You may be able to do some Pi-to-Windows PTP experiments with a
crossover cable. Not sure if the Pi has the hardware timestamping,
seems unlikely given the low cost of the Pi's as in the 3B+ (and below
I believe) the Ethernet is provided by a combined Ethernet/USB2 hub
and limited by the speed of the USB2.
- Is there any other way of making the clients' time offset lower?
NTP works best if the network delay is symmetric and stable - this
would also rule against wlan I'm afraid.
I've already tried setting PTP on the Raspberry Pi 3B+. It needed a kernel
recompilation with a patch I found here: [2], but I'm not sure it works
correctly.
From what I gather the Linux implementation can be either a slave or
act as the "grandmaster" without having to buy an expensive external
grandmaster clock (the ones I've seen from e.g. Microsemi, Brandywine,
Meinberg etc are of the "Get a quote" variety which tends to mean
"expensive")
Thanks in advance for any suggestions!
Adam Kumiszcza
Cheers,
Tim
On Thu, Aug 8, 2019 at 7:00 PM David J Taylor via time-nuts <
time-nuts@lists.febo.com> wrote:
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower.
[]
Thanks in advance for any suggestions!
Adam Kumiszcza
---===========
Adam,
You could feed the PPS signal to the Windows PCs as one way of making the
performance better:
https://www.satsignal.eu/ntp/setup.html
Performance:
https://www.satsignal.eu/mrtg/performance_ntp.php#windows-stratum-1
Cheers,
David
First of all, thank you for the info you put on satsignal.eu org! I've used
some of these in the initial configuration of my server.
By feeding PPS signal to the Windows PC you mean adding cards or connecting
boards (like Sure Electronics GPS evaluation boards) to Windows computers?
My idea was to lower the time offset of machines connected by LAN to a
single stratum-1 server. But maybe I'll come to that later.
Out of your monitored servers on
https://www.satsignal.eu/mrtg/performance_ntp.php I am amazed by
performance of LeoNTP box. What do you mean by indoor antenna here?
Best regards,
Adam
On Thu, Aug 8, 2019 at 7:01 PM Tim Lister listertim@gmail.com wrote:
On Thu, Aug 8, 2019 at 8:07 AM Adam Kumiszcza akumiszcza@gmail.com
wrote:
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP
client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower. I've heard that one of the
latest versions of Windows 10 provides support for PTP protocol (IEEE
1588v2). Hence my questions:
- Did anybody try using it on Windows 10 already? From what I can tell
right now, there's only a demonstration web site showing it works when
using Ubuntu Linux as a virtual machine [1]. Or should I wait a bit for
more robust implementation?
EndRun Technologies
(https://endruntechnologies.com/products/grandmaster-clocks/ptp-slaves)
make mention of a Windows PTP slave - not sure if it's the same
version you have found. It does also make mention of the fact that
precision is limited on Windows depending on version although from
what I gather from the satsignal.eu that precious on Windows improved
with 7 or later.
It's additional software here. As far as I understand, recent Windows 10
versions introduced under-the-hood internal implementation of PTPv2, with
kernel support for it.
- Do I guess correctly that current implementation of PTP works only on
Ethernet? Some of the computers are on wlan.
For deterministic timing I think it has to be Ethernet and as outlined
in their white paper (https://endruntechnologies.com/pdf/PTP-1588.pdf)
it works best if you have hardware timestamping in the NICs at each
end and IEEE1588 compatible network switches (and even when the claim
to support it, apparently they don't always do it properly:
https://www.linkedin.com/pulse/ptp-support-high-end-ethernet-switches-heiko-gerstung
).
You may be able to do some Pi-to-Windows PTP experiments with a
crossover cable. Not sure if the Pi has the hardware timestamping,
seems unlikely given the low cost of the Pi's as in the 3B+ (and below
I believe) the Ethernet is provided by a combined Ethernet/USB2 hub
and limited by the speed of the USB2.
Pi 3B+ has only software timestamping, and even this is not completely
clear (I had to patch the kernel). Pi-to-Windows direct crossover testing
is a great idea!
Best regards,
Adam
On Thu, Aug 8, 2019 at 4:07 PM Adam Kumiszcza akumiszcza@gmail.com wrote:
Hi everybody! My first post here, I hope the subject is adequate for this
mailing list.
I'm using a tiny layer 1 NTP server consisting of Raspberry Pi 3B+ with
Ublox MAX-M8Q expansion board providing GNSS (currently GPS, Galileo and
Glonass, sometimes I switch to Beidou, too) reference with PPS + a simple
patch antenna hanging near the window. Offset, jitter and rms are most
often smaller than 1 µs. The server is included in NTP pool.
I'm using several Windows 10 machines on the same LAN, all using NTP client
software from Meinberg. The typical offset and jitter in those are about
100-500 µs. I would like to make it lower. I've heard that one of the
latest versions of Windows 10 provides support for PTP protocol (IEEE
1588v2). Hence my questions:
- Did anybody try using it on Windows 10 already? From what I can tell
right now, there's only a demonstration web site showing it works when
using Ubuntu Linux as a virtual machine [1]. Or should I wait a bit for
more robust implementation? - Do I guess correctly that current implementation of PTP works only on
Ethernet? Some of the computers are on wlan. - Is there any other way of making the clients' time offset lower?
Although there is one method using WSL, there is also the "native" method
https://techcommunity.microsoft.com/t5/Networking-Blog/Top-10-Networking-Features-in-Windows-Server-2019-10-Accurate/ba-p/339739
"Precision Time Protocol - Try it out!"
PTP can be used across any IP network, but if there is non deterministic
latency (jitter), there is a fairly obvious effect.
The offset should reduce over time, but jitter will affect this.
I've already tried setting PTP on the Raspberry Pi 3B+. It needed a kernel
recompilation with a patch I found here: [2], but I'm not sure it works
correctly.
I have run PTP on a rPi 3B without any kernel modifications as a
grandmaster (with a ublox GPS to provide time and PPS), and experimented
with different NICs (intel, broadcom and solarflare, all with HW
timestamping) as clients in x64 hosts.
There are two issues with using a rPi as a PTP grandmaster, the crystal is
"not very stable", and the NIC is connected over USB, both of these add
jitter which is noticeable on PTP. The jjitter is obvious when compared to
a meinberg m600 with the HQ OXCO option.connected to a Solarflare NIC in a
Linux host.
To remove the jitter issues, you might care to try the ublox connected (via
a suitable level converter for NMEA and PPS) to a hardware serial port on a
Linux X64 host.
I have not tried the native Windows PTP client, there are also 3rd party
clients from greyware and fsmlabs.
As you can see on the screenshot below, PTP appears here as a server, but
I'm not sure if it's used at all.
[image: status16.PNG]
Ptp4l status shows this:
pi@zegar:~ $ systemctl status ptp4l
● ptp4l.service - Precision Time Protocol (PTP) service
Loaded: loaded (/lib/systemd/system/ptp4l.service; enabled; vendor
preset: enabled)
Active: active (running) since Wed 2019-08-07 12:32:28 CEST; 24h ago
Docs: man:ptp4l
Main PID: 323 (ptp4l)
Tasks: 1 (limit: 4915)
CGroup: /system.slice/ptp4l.service
└─323 /usr/sbin/ptp4l -f /etc/linuxptp/ptp4l.conf -i eth0
Aug 07 12:32:31 zegar ptp4l[323]: ptp4l[6.631]: port 1: link up
Aug 07 12:32:31 zegar ptp4l[323]: ptp4l[6.631]: port 1: FAULTY to LISTENING
on FAULT_CLEARED
Aug 07 12:32:31 zegar ptp4l[323]: [6.631] port 1: link up
Aug 07 12:32:31 zegar ptp4l[323]: [6.631] port 1: FAULTY to LISTENING on
FAULT_CLEARED
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: port 1: LISTENING to
MASTER on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: selected best master clock
b827eb.fffe.006e65
Aug 07 12:32:38 zegar ptp4l[323]: ptp4l[13.526]: assuming the grand master
role
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] port 1: LISTENING to MASTER on
ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] selected best master clock
b827eb.fffe.006e65
Aug 07 12:32:38 zegar ptp4l[323]: [13.526] assuming the grand master role
Thanks in advance for any suggestions!
Adam Kumiszcza
[1]
https://techcommunity.microsoft.com/t5/Networking-Blog/Windows-Subsystem-for-Linux-for-testing-Windows-10-PTP-Client/ba-p/389181
[2] https://github.com/twteamware/raspberrypi-ptp/issues/2
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
From: Adam Kumiszcza
First of all, thank you for the info you put on satsignal.eu org! I've used
some of these in the initial configuration of my server.
By feeding PPS signal to the Windows PC you mean adding cards or connecting
boards (like Sure Electronics GPS evaluation boards) to Windows computers?
My idea was to lower the time offset of machines connected by LAN to a
single stratum-1 server. But maybe I'll come to that later.
Out of your monitored servers on
https://www.satsignal.eu/mrtg/performance_ntp.php I am amazed by
performance of LeoNTP box. What do you mean by indoor antenna here?
Best regards,
Adam
Adam,
One thing - I note you say "single stratum-1 server". With GPS/Raspberry Pi
being so cheap, why not have several servers each with a separate antenna?
You already have a PPS signal from your existing GPS, so I was thinking
simply of distributing that the the Windows PCs, and feeding it to the
serial port's DCD signal.
If the PC doesn't have a serial port, check on the motherboard because there
are sometimes serial port headers left unconnected by the PC manufacturer.
If it doesn't have such a header, add a serial port card. If it's a laptop
then you might use a serial-over USB connection, but that's likely either
not to carry the serial control lines, or be of inferior performance to
other methods.
LeoNTP - my shack is on the second floor of a UK 2-storey house, and the
antenna for the LeoNTP is a magnetic-mount standard puck antenna, mounted on
the base of an desk light which is actually on top of a loudspeaker, so the
antenna is near the roof but indoors. Check with your Android mobile phone
what the GPS signal levels actually are in your possible locations - I think
Apple don't allow you to access that information. For this level of
performance there's no need for timing antennas. My other GPS antennas are
simple on a window sill on a south-facing, second floor window.
http://www.satsignal.eu/ntp/2019-01-08-1330-23b-GPS-antenna-farm.jpg
The key here is to have some conducting magnetic material under the antenna,
not just sitting on its own. There's more about the LeoNTP here:
http://www.leobodnar.com/shop/index.php?main_page=product_info&products_id=272
Cheers,
David
SatSignal Software - Quality software for you
Web: http://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv
Anyone had a play with this device ? 1pps (or whatever) input,
http://www.ti.com/lit/ds/symlink/lmk05318.pdf
For those with a radio interest, or telecom interest..
I will order an eval board and see what its like...
Yeah, I know its easy to do with a micro or fpga , but the phase noise
from the internal VCO is impressive.
I have no idea what the random stuff < 1Hz in adev territory is like.
-glen
On Fri, Aug 9, 2019 at 1:01 PM shouldbe q931 shouldbeq931@gmail.com wrote:
On Thu, Aug 8, 2019 at 4:07 PM Adam Kumiszcza akumiszcza@gmail.com
wrote:
Although there is one method using WSL, there is also the "native" method
https://techcommunity.microsoft.com/t5/Networking-Blog/Top-10-Networking-Features-in-Windows-Server-2019-10-Accurate/ba-p/339739
"Precision Time Protocol - Try it out!"
Thanks! The link at the bottom is to a .doc file, which gives some
explanations. I'll try it out more thoroghly later. (
https://aka.ms/PTPValidation).
Some of the required links do not work, though, so it all seems a long way
till stable I guess.
I have run PTP on a rPi 3B without any kernel modifications as a
grandmaster (with a ublox GPS to provide time and PPS), and experimented
with different NICs (intel, broadcom and solarflare, all with HW
timestamping) as clients in x64 hosts.
What type of PTP did you use? ptpd (https://github.com/ptpd/ptpd) or
linuxptp (http://linuxptp.sourceforge.net/)?
I've tried the latter, and used the following ntp.conf (part of it here):
#PPS Kernel mode
server 127.127.22.0 minpoll 4 maxpoll 4
fudge 127.127.22.0 time1 +0.000000 flag3 0 refid PPS
tos mindist 0.002
#GPS (NMEA)
server 127.127.20.0 mode 89 minpoll 4 maxpoll 4 iburst prefer
fudge 127.127.20.0 flag1 0 flag3 0 time2 0.059089 refid GPS stratum 2
#local PTP reference
server 127.127.28.0
fudge 127.127.28.0 refid PTP
/etc/linuxptp/ptp4l.conf has the following lines:
clock_servo ntpshm
ntpshm_segment 0
priority1 10
priority2 10
(the rest was left as default)
/etc/linuxptp/timemaster.conf is as follows:
Configuration file for timemaster
[ntp_server localhost]
minpoll 4
maxpoll 4
[ptp_domain 0]
interfaces eth0
delay 10e-6
[timemaster]
ntp_program ntpd
[chrony.conf]
include /etc/chrony.conf
[ntp.conf]
includefile /etc/ntp.conf
[ptp4l.conf]
[chronyd]
path /usr/sbin/chronyd
[ntpd]
path /usr/sbin/ntpd
options -u ntp:ntp -g
[phc2sys]
path /usr/sbin/phc2sys
[ptp4l]
path /usr/sbin/ptp4l
But I think this shows it does not really interact with ntp:
pi@zegar:~ $ ntpq -c "cv &3"
associd=52436 status=00f1 15 events, clk_no_reply,
device="SHM/Shared memory interface", timecode=, poll=1850, noreply=1850,
badformat=0, baddata=0, stratum=0, refid=PTP, flags=0
(maybe because there's nothing connected on the other side, but I guess
noreply means no reply from PTP daemon?)
Best regards,
Adam