Re: [PATCH v12 2/9] PCI: stm32: Add PCIe host support for STM32MP25

[Bjorn Helgaas <helgaas@kernel.org>]

On Fri, Aug 08, 2025 at 04:55:52PM +0200, Christian Bruel wrote:
> On 8/7/25 20:09, Bjorn Helgaas wrote:
> > [+to Linus for pinctrl usage question below]
> > 
> > On Tue, Jun 10, 2025 at 11:07:07AM +0200, Christian Bruel wrote:
> > > Add driver for the STM32MP25 SoC PCIe Gen1 2.5 GT/s and Gen2 5GT/s
> > > controller based on the DesignWare PCIe core.

> > > +	return pinctrl_pm_select_sleep_state(dev);
> > 
> > Isn't there some setup required before we can use
> > pinctrl_select_state(), pinctrl_pm_select_sleep_state(),
> > pinctrl_pm_select_default_state(), etc?
> > 
> > I expected something like devm_pinctrl_get() in the .probe() path, but
> > I don't see anything.  I don't know how pinctrl works, but I don't see
> > how dev->pins gets set up.
> 
> Linus knows better, but the dev->pins states are attached to the dev struct
> before probe by the pinctrl driver
> 
> /**
>  * pinctrl_bind_pins() - called by the device core before probe
>  * @dev: the device that is just about to probe
>  */
> int pinctrl_bind_pins(struct device *dev)

Thanks for the pointer.  Might be worthy of a mention in
Documentation/driver-api/pin-control.rst.  Maybe pinctrl/consumer.h
could even have a bread crumb to that effect since drivers use all
those interfaces that rely in the implicit initialization done before
their .probe().

pin-control.rst mentions pinctrl_get_select_default() being called
just before the driver probe, but that's now unused and it looks like
pinctrl_bind_pins() does something similar:

  really_probe
    pinctrl_bind_pins
      dev->pins = devm_kzalloc()
      devm_pinctrl_get
      pinctrl_lookup_state(PINCTRL_STATE_DEFAULT)
      pinctrl_lookup_state(PINCTRL_STATE_INIT)
      pinctrl_select_state(init)      # if present, else default
    call_driver_probe

Bjorn