On Tue, Dec 17, 2019 at 10:21:55AM +0100, Andrew Lunn wrote: > Forward declarations are considered bad. Not by me! > Please add a new patch to the > series which moves code around first. Sorry, I disagree. For new drivers, sure, but for testing, production drivers, moving code blocks around "just because" is only asking for new bugs due to copy-pastos. > When using phylink, not phylib, this call will not happen. You need to > add a similar bit of code in phylink_mac_config(). Good to know. > For the moment what you have is sufficient. I doubt anybody is using > the dp83640 with phylink, and the new hardware you are targeting seems > to be RGMII based, not SERDES, which is the main use case for PHYLINK. Yeah, my impression is that the phyter will be the first and last phy time stamping device ever created. Designers reject this part because it is 100 mbit only. And there are no gigabit+ phys with time stamping at all. So I don't anticipate the phylink layer needing any of this time stamping stuff in the foreseeable future. Thanks, Richard
