Convert the driver to immutable irq-chip with a bit of
intuition.

I switched to using irqd_to_hwirq() consistently while we
are at it.

This driver does not use the GPIOCHIP_IRQ_RESOURCE_HELPERS
as it defines its own resource reservations, simply in
order to turn IRQ lines into inputs on initialization.

Also switched the open coded calls to gpiochip_lock_as_irq()
to gpiochip_reqres_irq() so we also get the right module
reference counting.

Cc: Marc Zyngier <maz@kernel.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230403-immutable-irqchips-v1-7-503788a7f6e6@linaro.org

intel-pinctrl for v6.1-2

* Add missing and remove unused headers in the pin control and GPIO drivers
* Revise the pin control and GPIO headers

Do not imply that some of the generic headers may be always included.
Instead, include explicitly what we are direct user of.

While at it, sort headers alphabetically.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>

The OF node in the GPIO library is deprecated and soon
will be removed.

GPIO library now accepts fwnode as a firmware node, so
switch the driver to use it.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Link: https://lore.kernel.org/r/20221010075615.43244-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

Pin descriptions for this chip only look like standard GPIO device tree
descriptions, while in fact they contain additional data (in excess of
number of cells specified in description of gpio controllers). They also
refer to only pins/gpios belonging to the driver and not to arbitrary
gpio in the system.

Because we want to stop exporting OF-specific handlers from gpiolib-of,
let's parse the pin reference ourself instead of trying to call
of_get_named_gpio().

Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Tested-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
Link: https://lore.kernel.org/r/YzSsgoVoJn4+mSpv@google.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

Since we have a generic helper, switch the module to use it.

As a side effect, add check for the memory allocation failures and
cleanup it either in error case or when driver is unloading.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

It's fine to call dev_err_probe() in ->probe() when error code is known.
Convert the driver to use dev_err_probe().

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Use the devm_platform_ioremap_resource_byname() helper instead of
calling platform_get_resource_byname() and devm_ioremap_resource()
separately.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>

Use temporary variable for struct device to make code neater.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Kernel doc validator is not happy:

  .../pinctrl-st.c:59: warning: This comment starts with '/**', but isn't a kernel-doc comment.
  .../pinctrl-st.c:73: warning: This comment starts with '/**', but isn't a kernel-doc comment.

Drop them as they are indeed not a kernel doc comments.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>