Am Freitag, 17. Juli 2026, 09:32:18 Ostafrikanische Zeit schrieb Krzysztof 
Kozlowski:
> On Fri, Jul 17, 2026 at 12:35:37AM +0300, Stefan Dösinger wrote:

> > +  resets:
> > +    items:
> > +      - description: USB phy reset
> > +      - description: HSIC phy reset
> 
> Why do you reference here resets from the parent? That's a clear signal
> you have one device, so one device node regardless of Linux phy core
> behavior.

This is a follow-up question for the PHY maintainers, if they notice the E-
Mail thread: I'm trying to resolve an apparent conflict between DT 
expectations and a check in the PHY core:

device_set_of_node_from_dev(dev, dev->parent);
...
devm_of_phy_provider_register(dev, xlate);

Passes the letter of the check in __of_phy_provider_register(), but I am not 
sure about the spirit.

"dev" here is a MFD child node without a DT node in accordance to DT's 
writing-bindings instructions ("DON’T create nodes just for the sake of 
instantiating drivers"). The PHY core prohibits a child to register a phy 
provider to its parent node. Why?