Re: [PATCH v2 3/4] net: Let the active time stamping layer be selectable.

[Jakub Kicinski]

On Sat, 4 Mar 2023 20:46:05 +0100 Andrew Lunn wrote:
> > Since the ioctl is to do with requesting what we want the timestamping
> > layer to be doing with packets, putting it in ptp_clock_info makes
> > very little sense.  
> 
> So there does not appear to be an object to represent a time stamper?
> 
> Should one be added? It looks like it needs two ops hwtstamp_set() and
> hwtstamp_get(). It would then be registered with the ptp core. And
> then the rest of what i said would apply...

IMHO time stamper is very much part of the netdev. I attribute the lack
of clarity palatially to the fact that (for reasons unknown) we still
lug the request as a raw IOCTL/ifreq. Rather than converting it to an
NDO/phydev op in the core.. Also can't think of a reason why modeling
it as a separate object would be useful?