On Mon, Nov 17, 2025 at 01:22:32PM +0100, Buday Csaba wrote:

On Mon, Nov 17, 2025 at 09:34:49AM +0000, Russell King (Oracle) wrote:

quoted

On Mon, Nov 17, 2025 at 10:28:51AM +0100, Buday Csaba wrote:

quoted

diff --git a/include/linux/mdio.h b/include/linux/mdio.h
index 42d6d47e4..1322d2623 100644
--- a/include/linux/mdio.h
+++ b/include/linux/mdio.h

@@ -92,6 +92,8 @@ void mdio_device_free(struct mdio_device *mdiodev);
 struct mdio_device *mdio_device_create(struct mii_bus *bus, int addr);
 int mdio_device_register(struct mdio_device *mdiodev);
 void mdio_device_remove(struct mdio_device *mdiodev);
+int mdio_device_register_gpiod(struct mdio_device *mdiodev);
+int mdio_device_register_reset(struct mdio_device *mdiodev);

These are private functions to the mdio code living in drivers/net/phy,
so I wonder whether we want to have drivers/net/phy/mdio.h for these to
discourage other code calling these?

I completely agree with that, but that file does not exist yet.
Is it worth creating just for the sake of these two functions?

The effort of creating such a file is much smaller than cleaning up
the mess when somebody uses them inappropriately.

Maybe drivers/net/phy/mdio.h is too open,
drivers/net/phy/mdio-private.h would be better, since these are
supposed to only be used by the core, not PHY drivers.

	Andrew