gpio/pinctrl: imx: let IOMUX controller know about on-SoC GPIOs

The change establishes a connection between on-SoC IOMUX controller(s) and GPIO controllers found on some SoC from Freescale/NXP iMX series, if a GPIO controller device node contains common gpio-ranges information.

The change is backward compatible with respect to potentially not updated outdated DTB data without gpio-ranges propery, for such boards the only functional change is lowered initcall priority of GPIO controller driver, which in general anyway is exected to be used only after pinctrl/pinmux controller.

If this change is applied the next interesting applications may be done as a follow-up work, for example switching pad function to GPIO on gpiod request, converting iomux controller driver to strict type and so on.

For actual values of gpio-ranges properties please reference series "ARM: dts: imx: add gpio-ranges properties to some iMX GPIO controllers"
http://www.spinics.net/lists/arm-kernel/msg525258.html

Changes from v1 to v2:
* replaced 2/3 by an own change providing a better commit description and which moves gpio_mxc_init() call to subsys_initcall() instead of apparently too late device_initcall(), this mitigates Shawn's expressed concern about the change.

Vladimir Zapolskiy (3):
pinctrl: imx: accept gpio request/free from pinctrl
gpio: mxc: shift gpio_mxc_init() to subsys_initcall level
gpio: mxc: add generic gpio request/free callbacks to pinctrl

 drivers/gpio/gpio-mxc.c                 | 7 ++++++-
 drivers/pinctrl/freescale/pinctrl-imx.c | 4 ++--
 2 files changed, 8 insertions(+), 3 deletions(-)

2.8.1

While only Freescale Vybrid SoC has settings of GPIO capabilities done by iomux controller, it is only a matter of GPIO controller driver implementation for the rest of Freescale/NXP SoCs from iMX series.

As a practical example on GPIO request a pad function should be switched to GPIO, but because this requires updates to all particular iMX pinctrl drivers, for simplicity at the moment add only a proper connection between shared pinctrl-imx and pinctrl/pinmux core, namely .gpio_request_enable/.gpio_disable_free/.gpio_set_direction callbacks should return success to a caller. This change allows to progress by adding request/free callbacks into gpio-mxc.c driver.

---
Changes from v1 to v2:
* none

 drivers/pinctrl/freescale/pinctrl-imx.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 71391757938b..6e248ca10169 100644
@@ -315,7 +315,7 @@ static int imx_pmx_gpio_request_enable(struct pinctrl_dev *pctldev,
 
 	/* Currently implementation only for shared mux/conf register */
 	if (!(info->flags & SHARE_MUX_CONF_REG))
-		return -EINVAL;
+		return 0;
 
 	pin_reg = &info->pin_regs[offset];
 	if (pin_reg->mux_reg == -1)

@@ -380,7 +380,7 @@ static int imx_pmx_gpio_set_direction(struct pinctrl_dev *pctldev,
 	 * They are part of the shared mux/conf register.
 	 */
 	if (!(info->flags & SHARE_MUX_CONF_REG))
-		return -EINVAL;
+		return 0;
 
 	pin_reg = &info->pin_regs[offset];
 	if (pin_reg->mux_reg == -1)

2.8.1

Patch applied with Shawn's ACK.

Yours,
Linus Walleij

In general situation on-SoC GPIO controller drivers should be probed after pinctrl/pinmux controller driver, because on-SoC GPIOs utilize a pin/pad as a resource provided and controlled by pinctrl subsystem.

This is stated in multiple places, e.g. from drivers/Makefile:

GPIO must come after pinctrl as gpios may need to mux pins etc

Looking at Freescale iMX SoC series specifics, imx*_pinctrl_init() functions are called at arch_initcall and postcore_initcall init levels, so the change of initcall level for gpio-mxc driver from postcore_initcall to subsys_initcall level is sufficient. Also note that the most of GPIO controller drivers settled at subsys_initcall level.

If pinctrl subsystem manages pads with GPIO functions, the change is needed to avoid unwanted driver probe deferrals during kernel boot.

---
Changes from v1 to v2:
* replaced by a change with better commit description and which moves gpio_mxc_init() call to subsys_initcall() instead of apparently too late device_initcall(), this mitigates Shawn's expressed concern about the change.

 drivers/gpio/gpio-mxc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 1b342a3842c8..1fdd5d804b5b 100644
@@ -510,7 +510,7 @@ static int __init gpio_mxc_init(void)
 {
 	return platform_driver_register(&mxc_gpio_driver);
 }
-postcore_initcall(gpio_mxc_init);
+subsys_initcall(gpio_mxc_init);
 
 MODULE_AUTHOR("Freescale Semiconductor, "
 	      "Daniel Mack <danielncaiaq.de>, "

2.8.1

Patch applied with Shawn's ACK.

Yours,
Linus Walleij

If a GPIO controller description in board DTB contains information about mappings between GPIOs and pads under IOMUX control use it to request and free GPIOs with respect to pinctrl/pinmux subsystems.

One of immediate positive functional changes is inability to request non-existing GPIOs, i.e. if there is no pad such. Also pinctrl/pinmux may now properly account pads occupied by requested GPIOs.

The change has no effect, if "gpio-ranges" property is not found including the case if a board has no DTB firmware.

---
Changes from v1 to v2:
* none

 drivers/gpio/gpio-mxc.c | 5 +++++
 1 file changed, 5 insertions(+)

index 1fdd5d804b5b..e38989a4fa0c 100644
@@ -458,6 +458,11 @@ static int mxc_gpio_probe(struct platform_device *pdev)
 	if (err)
 		goto out_bgio;
 
+	if (of_property_read_bool(np, "gpio-ranges")) {
+		port->gc.request = gpiochip_generic_request;
+		port->gc.free = gpiochip_generic_free;
+	}
+
 	port->gc.to_irq = mxc_gpio_to_irq;
 	port->gc.base = (pdev->id < 0) ? of_alias_get_id(np, "gpio") * 32 :
 					     pdev->id * 32;

2.8.1

Patch applied with Shawn's ACK.

Yours,
Linus Walleij

Yes, this is more conservative and stands less chance to cause regression.

For the series,