I regret to inform everyone that I can no longer actively develop PTPd. In particular, I am prohibited from updating it for IEEE 1588v2. Let me know if anyone would like to join this project to work on merging the outstanding patches or updating PTPd for 1588v2. Also, please post here if anyone knows of other open source PTP implementations or would like to fork a new open source project from the PTPd code base. I can add links to other projects on the home page.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Kendall, first of all, thanks for your work on PTPd -- it is a critical part of my multiple camera, multiple computer flying animal tracker (see http://dickinson.caltech.edu/Research/MultiTrack for some details).
However, I would like to ask why you are prohibited from updating it? Is this a legal requirement, have you signed some sort of NDA, a condition of your current employment, or a result of more pressing commitments? I'm asking because I'm wondering if there's some kind of legal issue with making an open source implementation of 1588v2.
Thanks,
Andrew
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
First, thanks kendall for the great job you've done on Ptpd !
I'm currently doing my internship and i have to implement 1588 V2 on Linux system.
Maybe it won't be a complete implementation, but only for an ordinary clock using the peer delay mechanism.
I will largely reused ptpd, and i think the implementation will be available under a BSD-style license as ptpd.
My internship will finish in august 2009 so i will keep you informed.
alex
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
First, let me also thank Kendall for making PTPd available. It has been immensely useful. It is a pity that you cannot continue maintaining it. I hope someone will continue with the project, because otherwise there is a considerable risk that it will fragment into many different, unrelated branches.
I myself unfortunately don't have the time for it, in particular as I am now working in other areas. My work on hardware assisted clock synchronization with the new Intel NICs has been handed off to the Intel driver team. Their agenda doesn't include maintaining a user space daemon either.
Second, to prevent more fragmentation than strictly necessary, may I suggest that anyone who wants to make modifications from now on does that in a public git repository and posts links to it here? If possible, start from the most recent copy of the source and (if it is not too much work) integrate patches by others while you work on the source.
Perhaps that way we can keep one copy of the sources up-to-date and tested until someone picks up the task of releasing official versions again.
It includes all changes from the official Subversion, plus:
* an abstraction layer for time stamping
* support for hardware time stamping with
** an experimental ioctl() interface for Intel 82576
** the upcoming timestamping.txt API in Linux 2.6.30
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am interested in helping here with project maintenance. Can you provide an overview on what work needs to be done? Where are the unmerged patches? Feel free to contact me as "linville@tuxdriver.com"...
Thanks,
John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It includes ptpd version 1.0.0 plus some cleanups and HW offload enhancements from Patrick Ohly. I'm still open to picking a different starting point or making other tree changes if people have suggestions/objections, so speak-up!
Others have expressed interest in moving towards version 2 of ieee1588. I look forward to seeing those discussions (and patches) on the mailing list.
See ya there soon! :-)
John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I wasn't able to find any news on 1588 patent issues and I don't see mention of ptpd being forbidden on Fedora (http://fedoraproject.org/wiki/ForbiddenItems). Do you have a pointer to some of this news? Thanks.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am not aware of any specific claims against ptpd or any threats of action by patent holders or anything like that. Nevertheless, I won't be the one to put Fedora at risk.
Hth!
John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Gaël, thanks a lot for your contribution to the development of IEEE-1588 v2 open source code.
Please let us know once you'll have got a chance to update the relevant documents (readme, release notes, ptpd.8 …) and Makefile.
You indicated that ptpd v2 module supported P2P and E2E TC mode. Just for clarification, does it imply P2P and E2E for ordinary and boundary clocks as well?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
In looking at the source code of ptpd v2, it seems that ptpd v1 option -i <system_clock_identifier> is replaced in v2 by ptpd option -r <clock_accuracy>. For NTP system clock identifier in v1, is it equivalent to option -r 0x2F in v2? What does it mean exactly?
My product receives timing information from both an external NTP server and from GPS/CDMA time sources in parallel. All goes to the local ntpd daemon which decides the best time source to use for the system clock. The GPS time packets are sent to ntpd via a NTP reference clock driver. Is it possible to integrate PTP into NTP with a reference clock driver (in order to let NTP algorithm dynamically decide which best time info to use among multiple sources)?
Thanks.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I regret to inform everyone that I can no longer actively develop PTPd. In particular, I am prohibited from updating it for IEEE 1588v2. Let me know if anyone would like to join this project to work on merging the outstanding patches or updating PTPd for 1588v2. Also, please post here if anyone knows of other open source PTP implementations or would like to fork a new open source project from the PTPd code base. I can add links to other projects on the home page.
Kendall, first of all, thanks for your work on PTPd -- it is a critical part of my multiple camera, multiple computer flying animal tracker (see http://dickinson.caltech.edu/Research/MultiTrack for some details).
However, I would like to ask why you are prohibited from updating it? Is this a legal requirement, have you signed some sort of NDA, a condition of your current employment, or a result of more pressing commitments? I'm asking because I'm wondering if there's some kind of legal issue with making an open source implementation of 1588v2.
Thanks,
Andrew
Hi all ,
First, thanks kendall for the great job you've done on Ptpd !
I'm currently doing my internship and i have to implement 1588 V2 on Linux system.
Maybe it won't be a complete implementation, but only for an ordinary clock using the peer delay mechanism.
I will largely reused ptpd, and i think the implementation will be available under a BSD-style license as ptpd.
My internship will finish in august 2009 so i will keep you informed.
alex
First, let me also thank Kendall for making PTPd available. It has been immensely useful. It is a pity that you cannot continue maintaining it. I hope someone will continue with the project, because otherwise there is a considerable risk that it will fragment into many different, unrelated branches.
I myself unfortunately don't have the time for it, in particular as I am now working in other areas. My work on hardware assisted clock synchronization with the new Intel NICs has been handed off to the Intel driver team. Their agenda doesn't include maintaining a user space daemon either.
Second, to prevent more fragmentation than strictly necessary, may I suggest that anyone who wants to make modifications from now on does that in a public git repository and posts links to it here? If possible, start from the most recent copy of the source and (if it is not too much work) integrate patches by others while you work on the source.
Perhaps that way we can keep one copy of the sources up-to-date and tested until someone picks up the task of releasing official versions again.
My branch is here:
http://github.com/pohly/ptpd/tree/master
It includes all changes from the official Subversion, plus:
* an abstraction layer for time stamping
* support for hardware time stamping with
** an experimental ioctl() interface for Intel 82576
** the upcoming timestamping.txt API in Linux 2.6.30
I am interested in helping here with project maintenance. Can you provide an overview on what work needs to be done? Where are the unmerged patches? Feel free to contact me as "linville@tuxdriver.com"...
Thanks,
John
Well, it's been more than a couple of weeks and no answer here... :-(
I have a few interested parties that have responded off-line. There seems to be some support for continuing this project elsewhere.
I have taken the liberty of setting-up a mailing list here:
http://lists.infradead.org/mailman/listinfo/ptpd
I also have established a git tree:
git://git.infradead.org/ptpd.git
It includes ptpd version 1.0.0 plus some cleanups and HW offload enhancements from Patrick Ohly. I'm still open to picking a different starting point or making other tree changes if people have suggestions/objections, so speak-up!
Others have expressed interest in moving towards version 2 of ieee1588. I look forward to seeing those discussions (and patches) on the mailing list.
See ya there soon! :-)
John
Well, so I thought…
http://lists.infradead.org/pipermail/ptpd/2010-January/000023.html
Short version is that the relocated project didn't see much activity, and patent issues around IEEE 1588 have scared-away Fedora, RHEL, and me. :-(
Best of luck with the project!
John
I wasn't able to find any news on 1588 patent issues and I don't see mention of ptpd being forbidden on Fedora (http://fedoraproject.org/wiki/ForbiddenItems). Do you have a pointer to some of this news? Thanks.
I documented the patent claims I found here:
https://bugzilla.redhat.com/show_bug.cgi?id=556611#c9
I am not aware of any specific claims against ptpd or any threats of action by patent holders or anything like that. Nevertheless, I won't be the one to put Fedora at risk.
Hth!
John
I see IEEE 1588-2008 (PTPv2) code is available at http://ptpd.svn.sourceforge.net/viewvc/ptpd/branches/.
Has anyone tried it out?
Is this code for ordinary or boundary clocks (README and RELEASE NOTES seem not up to date)?
Ptpd v2 module supports P2P and E2E TC mode defined in IEEE 1588-2008 standard and compliant with the profile defined in the IEEE 802.1AS draft.
And it supports also the ptpdv1 ordinary mode.
This new version was tested against a Hirschmann Ethernet switch.
I have to update the README and RELEASE NOTES … :-)
Gaël
Makefile seems to be not up to date to
Gaël, thanks a lot for your contribution to the development of IEEE-1588 v2 open source code.
Please let us know once you'll have got a chance to update the relevant documents (readme, release notes, ptpd.8 …) and Makefile.
You indicated that ptpd v2 module supported P2P and E2E TC mode. Just for clarification, does it imply P2P and E2E for ordinary and boundary clocks as well?
P2P and E2E only for ordinary clock.
For the todo list, I will try to do my best according to my free time.
In looking at the source code of ptpd v2, it seems that ptpd v1 option -i <system_clock_identifier> is replaced in v2 by ptpd option -r <clock_accuracy>. For NTP system clock identifier in v1, is it equivalent to option -r 0x2F in v2? What does it mean exactly?
My product receives timing information from both an external NTP server and from GPS/CDMA time sources in parallel. All goes to the local ntpd daemon which decides the best time source to use for the system clock. The GPS time packets are sent to ntpd via a NTP reference clock driver. Is it possible to integrate PTP into NTP with a reference clock driver (in order to let NTP algorithm dynamically decide which best time info to use among multiple sources)?
Thanks.