Re: [Linuxptp-users] Time Jumps whole seconds only

[Mark P. <pr...@to...>]

The master is hardware on Intel Corporation I350 Gigabit Network Connection NICs (IGB driver) the ptp4l command is /usr/sbin/ptp4l -f /etc/linuxptp/ptp4l.conf -i enp0s31f6. See attached configuration files.

The slave is using the PTP profile "default-relaxed".

Any insight on how to go about getting the slave to time jump to the exact time and not just to whole seconds will be appreciated.


Thanks,
Mark

[cid:image009.png@01D54EC0.82458860]

Mark Priest | Software Engineer



TORC Robotics, Inc.
405 Partnership Drive, Blacksburg, VA 24060
Office: (540) 443-9262 | Fax: (540) 443-3667

Email:      pr...@to...<mailto:pr...@to...>
Website:  https://torc.ai<https://torc.ai/>

[cid:image003.png@01D54EBF.0B1FA140]<https://www.facebook.com/torcrobotics/> [cid:image004.png@01D54EBF.0B1FA140] <https://www.linkedin.com/company/torc-technologies/about/?viewAsMember=true>  [cid:image005.png@01D54EBF.0B1FA140] <https://twitter.com/torcrobotics>  [cid:image006.png@01D54EBF.0B1FA140] <https://www.youtube.com/channel/UCGKXBpisBMTJtshrr3KkOTg?view_as=subscriber>  [cid:image007.png@01D54EBF.0B1FA140] <https://www.instagram.com/torc_robotics/?hl=en>  [cid:image008.jpg@01D54EBF.0B1FA140] <https://www.glassdoor.com/Overview/Working-at-TORC-Robotics-EI_IE1027022.11,24.htm>

________________________________
From: Jacob Keller <jac...@in...>
Sent: Tuesday, September 8, 2020 7:13 PM
To: lin...@li... <lin...@li...>
Subject: Re: [Linuxptp-users] Time Jumps whole seconds only



On 9/1/2020 9:48 AM, Mark Priest wrote:
> While characterizing the response of linuxptp to time jumps I discovered
> the following behavior. Both Master and slave are running linuxPTP but
> are different hardware architectures.
> When the time jump occurs (greater than the step_threshold on the
> master) the time adjusts on the slave via a jump but only the whole
> seconds are adjusted.  This results in the time being still off by the
> fractional amount (less than 1 second). This difference is then slewed
> to at the normal slew rate.  In my specific case (assuming random time
> jumps) this results in an average of .5s needing to be slewed.  Why
> isn't the time jumped to exactly the new time? Is there a way to change
> the behavior to jump to exactly the new time?  The time required to slew
> causes critical data to be considered unusable.
>
> Example:
> *Before Jump:* Master time is 10:10:12.800 slave time is 10:10:10.300
> (Time difference is 2.5s)
> *After Jump:*    Master time is 10:10:12.800 slave time is 10:10:12.300
> (Time difference is 0.5s)
>
> Thanks,
> Mark

What mode is ptp4l running in? HW? SW? If in hardware mode we'd need to
know the hardware and what driver it is using.

Thanks,
Jake


_______________________________________________
Linuxptp-users mailing list
Lin...@li...
https://lists.sourceforge.net/lists/listinfo/linuxptp-users
CAUTION: This email originated from outside of TORC's email system. DO NOT click links or open attachments unless you recognize the sender and know the content is safe. Email is NOT a secure means of communication. Handle potentially sensitive information with care.
Confidentiality Notice: This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.