Hi Richard, The delay value didn't jump up, so it indicates device driver or hardware issue. Are we then referring to the I350 NIC device driver and hardware issue? This issue was happening for around 10 hours, then disappears for 3-4 days then coming back.
I was thinking if there is any possibility that the Linux software clock was drifting causing the jump/offset, and the NIC PHC is correcting it through phc2sys. BR Sean ________________________________ From: Richard Cochran <richardcoch...@gmail.com> Sent: Sunday, February 7, 2021 07:15 To: Sean Ma <x...@hotmail.com> Cc: linuxptp-users@lists.sourceforge.net <linuxptp-users@lists.sourceforge.net> Subject: Re: [Linuxptp-users] ptp4l results in stable master offset, but phc2sys usually see some spike in offset (10000+) On Sun, Feb 07, 2021 at 06:44:24AM +0000, Sean Ma wrote: > We are using a PTP GM connecting to Intel I350 as our PHC, running Centos OS. > > In the ptp4l, the master offset is always stable. > In the phc2sys sometimes we get spike up to 10000 once or more per minute, > command used: phc2sys -s <I350 NIC port> -w -m -n 44 > > Do we know the typical reason for this behavior? Check if the "delay" value also jumps up. - If so, then phc2sys is being preempted during a measurement. You can mitigate this by using chrt. - If not, the problem may be in the device driver or hardware. Thanks, Richard
_______________________________________________ Linuxptp-users mailing list Linuxptp-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-users
