Re: [Ntp] comments on draft-mlichvar-ntp-ntpv5-03 / Extension fields

[Danny Mayer <[e-mail]>]

On 12/1/21 11:43 PM, Dan Drown wrote:
> Quoting Danny Mayer <[e-mail]>:
>> No, it probably should be that last extension and not the first. 
>> Extensions should not, in general, be ordered. A correction like this 
>> should be after any MAC which does authenticate the packet so that 
>> this correction is included in the MAC.
>
> I'm assuming you mean that the correction is not included in the MAC. 
> I had not considered authenticated/signed packets, and I agree that 
> the corrrection would have to be outside of any signature. But this 
> restriction should probably be mentioned in the rfc.
>
Yes.
>>> [...] it will still be more difficult to
>>> implement than the PTP correction field as NTP works with timestamps
>>> corresponding to the end of the reception instead of beginning.
>>
>> See RFC8721. I believe that Tal did implement this.
>
> I'm assuming you mean RFC7821, but I don't see anything in it about 
> adjusting hardware timestamps to match NTP's legacy of software 
> timestamp measurements, just the parts needed to keep the UDP checksum 
> consistent after changing the TX timestamp in the packet.
>
Yes, RFC7821. That was exactly what Tal was doing.

Danny