Re: mt7530: dsa_switch_parse_of() fails, causes probe code to run twice

[Daniel Golle <daniel@makrotopia.org>]

On Sat, Apr 15, 2023 at 04:46:04PM +0300, Vladimir Oltean wrote:
> On Sat, Apr 15, 2023 at 04:40:19PM +0300, Arınç ÜNAL wrote:
> > My wording was not great there. What I meant is that PHY muxing will be
> > configured before dsa_register_switch() is run.
> 
> And we're back to the discussion from the thread "Move MT7530 phy muxing
> from DSA to PHY driver". What if someone decides that they don't need
> the switch driver - can they disable it? No.
> 
> Your thoughts are stopping mid way. If you think that PHY muxing should
> work without registering the DSA switch, then it doesn't belong in the
> DSA driver, plain and simple. No "yeah, but I can move it here, and it
> could kinda work, as a side effect of a driver failing to probe, or
> probing successfully but not registering with the subsystems for its
> primary purpose, or ...".

As the PHYs are accessed over the MDIO bus which is exposed by the mt7530.c
DSA driver the only middle ground would possibly be to introduce a MFD
driver taking care of creating the bus access regmap (MDIO vs. MDIO) and
expose the mt7530-controlled MDIO bus.

Obviously that'd be a bit more work than just moving some things from the
switch setup function to the probe function...