On Sun, 10 Jan 2021 17:35:25 +0100 Andrew Lunn wrote: > On Sun, Jan 10, 2021 at 11:13:44AM +0000, Russell King wrote: > > Check whether the MAC driver has implemented the get_ts_info() > > method first, and call it if present. If this method returns > > -EOPNOTSUPP, defer to the phylib or default implementation. > > > > This allows network drivers such as mvpp2 to use their more accurate > > timestamping implementation than using a less accurate implementation > > in the PHY. Network drivers can opt to defer to phylib by returning > > -EOPNOTSUPP. > > > > This change will be needed if the Marvell PHY drivers add support for > > PTP. > > > > Note: this may cause a change for any drivers that use phylib and > > provide get_ts_info(). It is not obvious if any such cases exist. > > Hi Russell > > We can detect that condition through? Call both, then do a WARN() if > we are changing the order? Maybe we should do that for a couple of > cycles? > > For netlink ethtool, we can also provide an additional attribute. A > MAC, or PHY indicator we can do in the core. A string for the name of > the driver would need a bigger change.
I'm not seeing a response to this suggestion, did vger eat it?
