Re: DLT request for EBHSCR

[Ebermann, Günter <Guenter.Ebermann () elektrobit com>]

From: Guy Harris [mailto:gharris () sonic net]
> On Sep 25, 2018, at 6:16 AM, Guenter.Ebermann () elektrobit com wrote:
>
> > I have updated http://www.elektrobit.com/ebhscr
>
> So what is the origin of the scale for the time stamps?
>
> For BroadR-Reach, I'm guessing, from
>
>       The timestamp is taken after the end of the SFD (Start of Frame
> Delimiter) detection, as specified in the IEEE 1588 Specification Sep- 2004 (IEC
> 61588 First Edition).
>
> that the origin is 1970-01-01 00:00:00 TAI (TAI, not UTC).  Is that correct?

Sorry for the late answer.

The origin of the time stamps is not defined.

We only take the definition of the time stamp point out of IEEE 1588 - because this
is what Ethernet MACs and PHYs are doing if they have hardware timestamping
support.

But not every capture device may be time synced with external sources like GPS or Ethernet PTP.
So perhaps they only have a local running counter they are using for timestamping.
That's the reason why we did not specify the origin of the scale of the time stamp.

BR
Günter
_______________________________________________
tcpdump-workers mailing list
tcpdump-workers () lists tcpdump org
https://lists.sandelman.ca/mailman/listinfo/tcpdump-workers