On Tuesday, 10 November 2020, 02:42:34 CET, Vladimir Oltean wrote:

Sorry for getting back late to you. It did not compute when I read your
email the first time around, then I let it sit for a while.

On Thu, Nov 05, 2020 at 09:18:04PM +0100, Christian Eggers wrote:

quoted

unfortunately I made a mistake when testing. Actually the timestamp *must*
be moved from the correction field (negative) to the egress tail tag.

That doesn't sound very good at all.

I think that is no drawback. It is implemented and works.

I have a simple question. Can your driver, when operating as a PTP
master clock, distribute a time in 2020 into your network? (you can
check with "phc_ctl /dev/ptp0 get")

# phc_ctl /dev/ptp2 get
phc_ctl[2141.093]: clock time is 1605018671.788481458 or Tue Nov 10 14:31:11 2020

The KSZ PTP clock has 32 bit seconds (and 30 bit nanoseconds).

I hope I can send the new series tonight. Had today no 5 minutes without the telephone ringing...