[PATCH v5 0/6] gpiolib: more helpers and fwnode conversion

[PATCH v5 0/6] gpiolib: more helpers and fwnode conversion

[Andy Shevchenko]

This is a spin-off (*) of the previous work of switching GPIO library
to use fwnode instead of of_node. Here we introduce a couple of
a new macro helpers, which allows to switch some of the drivers
to use fwnode and partially fwnode APIs. As a result of this cleanup
a few drivers switched to use GPIO fwnode instead of of_node.

*) it's subset of it with a new (patch 1) helper.

Marek, Martin, can you give this a try?
This requires at least two patches for GPIO library to be applied.

Bart, Linus, I can take it thru my tree with an immutable branch if
it's the way you prefer, otherwise please suggest on how to proceed.

Changelog v5:
- dropped tested patches (this series based on them, though)
- introduced a new helper (thanks Marek and Martin for reporting an issue)
- redone Armada and Meson code using newly introduced helper

Changelog v4:
- fixed compilation of the Samsung pin control drivers (LKP)
- explained in the commit message why namespacing is good for meson defs
- added tag to one of meson patches (Neil)

Changelog v3:
- moved count initialization to the definition in patch 2 (Geert)
- replaced of_args by args, used %pfwP in patch 7 (Geert)
- fixed kernel doc warning in patch 7
- added tags to patches 1, 2, 6, and 7 (Geert)
- added tag to patch 4 (Fabien)
- renamed MREG to MESON_REG in patch 9 (Neil)
- added tag to patch 10 (Neil)
- used --base for cover-letter

Changelog v2:
- properly based, so kbuild bot may test it (LKP)
- fixed typo in the macro (Geert)
- split to two macro helpers and rename the gpiochip_count()
- tagged one of stm32 and one of meson patches (Fabien, Neil)
- unified previously standalone armada patch
- due to above rewrote the armada patch from v1 completely (Sergey)
- added a lot of a new patches
- compile tested all of them on x86

Andy Shevchenko (6):
  gpiolib: Introduce a helper to get first GPIO controller node
  pinctrl: armada-37xx: Switch to use fwnode instead of of_node
  pinctrl: armada-37xx: Reuse GPIO fwnode in
    armada_37xx_irqchip_register()
  pinctrl: meson: Rename REG_* to MESON_REG_*
  pinctrl: meson: Enable COMPILE_TEST
  pinctrl: meson: Replace custom code by gpiochip_node_count() call

 drivers/pinctrl/meson/Kconfig               |  2 +-
 drivers/pinctrl/meson/pinctrl-meson.c       | 52 ++++++++++-----------
 drivers/pinctrl/meson/pinctrl-meson.h       | 28 +++++------
 drivers/pinctrl/mvebu/pinctrl-armada-37xx.c | 34 ++++----------
 include/linux/gpio/driver.h                 | 10 ++++
 5 files changed, 59 insertions(+), 67 deletions(-)

-- 
2.35.1

[PATCH v5 1/6] gpiolib: Introduce a helper to get first GPIO controller node

[Andy Shevchenko]

Introduce a helper to get first GPIO controller node which drivers
may want to use.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 include/linux/gpio/driver.h | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
index 12de0b22b4ef..83e2d72e51bb 100644
--- a/include/linux/gpio/driver.h
+++ b/include/linux/gpio/driver.h
@@ -766,4 +766,14 @@ static inline unsigned int gpiochip_node_count(struct device *dev)
 	return count;
 }
 
+static inline struct fwnode_handle *gpiochip_node_get_first(struct device *dev)
+{
+	struct fwnode_handle *fwnode;
+
+	for_each_gpiochip_node(dev, fwnode)
+		return fwnode;
+
+	return NULL;
+}
+
 #endif /* __LINUX_GPIO_DRIVER_H */
-- 
2.35.1

Re: [PATCH v5 1/6] gpiolib: Introduce a helper to get first GPIO controller node

[Bartosz Golaszewski]

On Thu, Apr 14, 2022 at 9:02 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
>
> Introduce a helper to get first GPIO controller node which drivers
> may want to use.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  include/linux/gpio/driver.h | 10 ++++++++++
>  1 file changed, 10 insertions(+)
>
> diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
> index 12de0b22b4ef..83e2d72e51bb 100644
> --- a/include/linux/gpio/driver.h
> +++ b/include/linux/gpio/driver.h
> @@ -766,4 +766,14 @@ static inline unsigned int gpiochip_node_count(struct device *dev)
>         return count;
>  }
>
> +static inline struct fwnode_handle *gpiochip_node_get_first(struct device *dev)
> +{
> +       struct fwnode_handle *fwnode;
> +
> +       for_each_gpiochip_node(dev, fwnode)
> +               return fwnode;
> +
> +       return NULL;
> +}
> +
>  #endif /* __LINUX_GPIO_DRIVER_H */
> --
> 2.35.1
>

Any chance you could name it get_first_gpiochip_node()? It's static so
we don't have to worry about the prefix and it would make the purpose
more clear.

Bart

Re: [PATCH v5 1/6] gpiolib: Introduce a helper to get first GPIO controller node

[Andy Shevchenko]

On Tue, Apr 26, 2022 at 12:27 AM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
>
> On Thu, Apr 14, 2022 at 9:02 PM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
> >
> > Introduce a helper to get first GPIO controller node which drivers
> > may want to use.

> > +static inline struct fwnode_handle *gpiochip_node_get_first(struct device *dev)
> > +{
> > +       struct fwnode_handle *fwnode;
> > +
> > +       for_each_gpiochip_node(dev, fwnode)
> > +               return fwnode;
> > +
> > +       return NULL;
> > +}
> > +
> >  #endif /* __LINUX_GPIO_DRIVER_H */
> > --
> > 2.35.1
> >
>
> Any chance you could name it get_first_gpiochip_node()? It's static so
> we don't have to worry about the prefix and it would make the purpose
> more clear.

There are two things why I prefer it as is:
1) it's static inline, so it's part of (internal) but still exported API;
2) it's already in my for-next branch which I would like not to
rebase, until it's a really serious issue.

That said, if you still insist I can rename it.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v5 1/6] gpiolib: Introduce a helper to get first GPIO controller node

[Bartosz Golaszewski]

On Tue, Apr 26, 2022 at 12:29 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
>
> On Tue, Apr 26, 2022 at 12:27 AM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
> >
> > On Thu, Apr 14, 2022 at 9:02 PM Andy Shevchenko
> > <andriy.shevchenko@linux.intel.com> wrote:
> > >
> > > Introduce a helper to get first GPIO controller node which drivers
> > > may want to use.
>
> > > +static inline struct fwnode_handle *gpiochip_node_get_first(struct device *dev)
> > > +{
> > > +       struct fwnode_handle *fwnode;
> > > +
> > > +       for_each_gpiochip_node(dev, fwnode)
> > > +               return fwnode;
> > > +
> > > +       return NULL;
> > > +}
> > > +
> > >  #endif /* __LINUX_GPIO_DRIVER_H */
> > > --
> > > 2.35.1
> > >
> >
> > Any chance you could name it get_first_gpiochip_node()? It's static so
> > we don't have to worry about the prefix and it would make the purpose
> > more clear.
>
> There are two things why I prefer it as is:
> 1) it's static inline, so it's part of (internal) but still exported API;
> 2) it's already in my for-next branch which I would like not to
> rebase, until it's a really serious issue.
>
> That said, if you still insist I can rename it.

No that's fine and I also pulled that into my tree.

Bart

Re: [PATCH v5 1/6] gpiolib: Introduce a helper to get first GPIO controller node

[Andy Shevchenko]

On Thu, May 05, 2022 at 03:05:52PM +0200, Bartosz Golaszewski wrote:
> On Tue, Apr 26, 2022 at 12:29 PM Andy Shevchenko
> <andy.shevchenko@gmail.com> wrote:
> > On Tue, Apr 26, 2022 at 12:27 AM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
> > > On Thu, Apr 14, 2022 at 9:02 PM Andy Shevchenko
> > > <andriy.shevchenko@linux.intel.com> wrote:

...

> > > Any chance you could name it get_first_gpiochip_node()? It's static so
> > > we don't have to worry about the prefix and it would make the purpose
> > > more clear.
> >
> > There are two things why I prefer it as is:
> > 1) it's static inline, so it's part of (internal) but still exported API;
> > 2) it's already in my for-next branch which I would like not to
> > rebase, until it's a really serious issue.
> >
> > That said, if you still insist I can rename it.
> 
> No that's fine and I also pulled that into my tree.

Thank you!

-- 
With Best Regards,
Andy Shevchenko

[PATCH v5 2/6] pinctrl: armada-37xx: Switch to use fwnode instead of of_node

[Andy Shevchenko]

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>
---
 drivers/pinctrl/mvebu/pinctrl-armada-37xx.c | 18 +++++++-----------
 1 file changed, 7 insertions(+), 11 deletions(-)

diff --git a/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c b/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c
index 08cad14042e2..110f70bce3fe 100644
--- a/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c
+++ b/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c
@@ -21,6 +21,7 @@
 #include <linux/pinctrl/pinctrl.h>
 #include <linux/pinctrl/pinmux.h>
 #include <linux/platform_device.h>
+#include <linux/property.h>
 #include <linux/regmap.h>
 #include <linux/slab.h>
 #include <linux/string_helpers.h>
@@ -787,18 +788,13 @@ static int armada_37xx_gpiochip_register(struct platform_device *pdev,
 					struct armada_37xx_pinctrl *info)
 {
 	struct device *dev = &pdev->dev;
-	struct device_node *np;
+	struct fwnode_handle *fwnode;
 	struct gpio_chip *gc;
-	int ret = -ENODEV;
+	int ret;
 
-	for_each_child_of_node(dev->of_node, np) {
-		if (of_find_property(np, "gpio-controller", NULL)) {
-			ret = 0;
-			break;
-		}
-	}
-	if (ret)
-		return ret;
+	fwnode = gpiochip_node_get_first(dev);
+	if (!fwnode)
+		return -ENODEV;
 
 	info->gpio_chip = armada_37xx_gpiolib_chip;
 
@@ -806,7 +802,7 @@ static int armada_37xx_gpiochip_register(struct platform_device *pdev,
 	gc->ngpio = info->data->nr_pins;
 	gc->parent = dev;
 	gc->base = -1;
-	gc->of_node = np;
+	gc->fwnode = fwnode;
 	gc->label = info->data->name;
 
 	ret = armada_37xx_irqchip_register(pdev, info);
-- 
2.35.1

[PATCH v5 3/6] pinctrl: armada-37xx: Reuse GPIO fwnode in armada_37xx_irqchip_register()

[Andy Shevchenko]

Since we have fwnode of the first found GPIO controller assigned to the
struct gpio_chip, we may reuse it in the armada_37xx_irqchip_register().

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/pinctrl/mvebu/pinctrl-armada-37xx.c | 16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)

diff --git a/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c b/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c
index 110f70bce3fe..ef4118e49f16 100644
--- a/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c
+++ b/drivers/pinctrl/mvebu/pinctrl-armada-37xx.c
@@ -727,23 +727,13 @@ static int armada_37xx_irqchip_register(struct platform_device *pdev,
 	struct gpio_chip *gc = &info->gpio_chip;
 	struct irq_chip *irqchip = &info->irq_chip;
 	struct gpio_irq_chip *girq = &gc->irq;
+	struct device_node *np = to_of_node(gc->fwnode);
 	struct device *dev = &pdev->dev;
-	struct device_node *np;
-	int ret = -ENODEV, i, nr_irq_parent;
-
-	/* Check if we have at least one gpio-controller child node */
-	for_each_child_of_node(dev->of_node, np) {
-		if (of_property_read_bool(np, "gpio-controller")) {
-			ret = 0;
-			break;
-		}
-	}
-	if (ret)
-		return dev_err_probe(dev, ret, "no gpio-controller child node\n");
+	unsigned int i, nr_irq_parent;
 
-	nr_irq_parent = of_irq_count(np);
 	spin_lock_init(&info->irq_lock);
 
+	nr_irq_parent = of_irq_count(np);
 	if (!nr_irq_parent) {
 		dev_err(dev, "invalid or no IRQ\n");
 		return 0;
-- 
2.35.1

[PATCH v5 4/6] pinctrl: meson: Rename REG_* to MESON_REG_*

[Andy Shevchenko]

Currently compilation test fails on x86 due to name collision. The usual
way to fix that is to move both conflicting parts to their own namespaces.

Rename REG_* to MESON_REG_* as a prerequisite for enabling COMPILE_TEST.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
---
 drivers/pinctrl/meson/pinctrl-meson.c | 24 +++++++++++------------
 drivers/pinctrl/meson/pinctrl-meson.h | 28 +++++++++++++--------------
 2 files changed, 26 insertions(+), 26 deletions(-)

diff --git a/drivers/pinctrl/meson/pinctrl-meson.c b/drivers/pinctrl/meson/pinctrl-meson.c
index 49851444a6e3..5b46a0979db7 100644
--- a/drivers/pinctrl/meson/pinctrl-meson.c
+++ b/drivers/pinctrl/meson/pinctrl-meson.c
@@ -218,13 +218,13 @@ static int meson_pinconf_set_output(struct meson_pinctrl *pc,
 				    unsigned int pin,
 				    bool out)
 {
-	return meson_pinconf_set_gpio_bit(pc, pin, REG_DIR, !out);
+	return meson_pinconf_set_gpio_bit(pc, pin, MESON_REG_DIR, !out);
 }
 
 static int meson_pinconf_get_output(struct meson_pinctrl *pc,
 				    unsigned int pin)
 {
-	int ret = meson_pinconf_get_gpio_bit(pc, pin, REG_DIR);
+	int ret = meson_pinconf_get_gpio_bit(pc, pin, MESON_REG_DIR);
 
 	if (ret < 0)
 		return ret;
@@ -236,13 +236,13 @@ static int meson_pinconf_set_drive(struct meson_pinctrl *pc,
 				   unsigned int pin,
 				   bool high)
 {
-	return meson_pinconf_set_gpio_bit(pc, pin, REG_OUT, high);
+	return meson_pinconf_set_gpio_bit(pc, pin, MESON_REG_OUT, high);
 }
 
 static int meson_pinconf_get_drive(struct meson_pinctrl *pc,
 				   unsigned int pin)
 {
-	return meson_pinconf_get_gpio_bit(pc, pin, REG_OUT);
+	return meson_pinconf_get_gpio_bit(pc, pin, MESON_REG_OUT);
 }
 
 static int meson_pinconf_set_output_drive(struct meson_pinctrl *pc,
@@ -269,7 +269,7 @@ static int meson_pinconf_disable_bias(struct meson_pinctrl *pc,
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, pin, REG_PULLEN, &reg, &bit);
+	meson_calc_reg_and_bit(bank, pin, MESON_REG_PULLEN, &reg, &bit);
 	ret = regmap_update_bits(pc->reg_pullen, reg, BIT(bit), 0);
 	if (ret)
 		return ret;
@@ -288,7 +288,7 @@ static int meson_pinconf_enable_bias(struct meson_pinctrl *pc, unsigned int pin,
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, pin, REG_PULL, &reg, &bit);
+	meson_calc_reg_and_bit(bank, pin, MESON_REG_PULL, &reg, &bit);
 	if (pull_up)
 		val = BIT(bit);
 
@@ -296,7 +296,7 @@ static int meson_pinconf_enable_bias(struct meson_pinctrl *pc, unsigned int pin,
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, pin, REG_PULLEN, &reg, &bit);
+	meson_calc_reg_and_bit(bank, pin, MESON_REG_PULLEN, &reg, &bit);
 	ret = regmap_update_bits(pc->reg_pullen, reg, BIT(bit),	BIT(bit));
 	if (ret)
 		return ret;
@@ -321,7 +321,7 @@ static int meson_pinconf_set_drive_strength(struct meson_pinctrl *pc,
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, pin, REG_DS, &reg, &bit);
+	meson_calc_reg_and_bit(bank, pin, MESON_REG_DS, &reg, &bit);
 
 	if (drive_strength_ua <= 500) {
 		ds_val = MESON_PINCONF_DRV_500UA;
@@ -407,7 +407,7 @@ static int meson_pinconf_get_pull(struct meson_pinctrl *pc, unsigned int pin)
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, pin, REG_PULLEN, &reg, &bit);
+	meson_calc_reg_and_bit(bank, pin, MESON_REG_PULLEN, &reg, &bit);
 
 	ret = regmap_read(pc->reg_pullen, reg, &val);
 	if (ret)
@@ -416,7 +416,7 @@ static int meson_pinconf_get_pull(struct meson_pinctrl *pc, unsigned int pin)
 	if (!(val & BIT(bit))) {
 		conf = PIN_CONFIG_BIAS_DISABLE;
 	} else {
-		meson_calc_reg_and_bit(bank, pin, REG_PULL, &reg, &bit);
+		meson_calc_reg_and_bit(bank, pin, MESON_REG_PULL, &reg, &bit);
 
 		ret = regmap_read(pc->reg_pull, reg, &val);
 		if (ret)
@@ -447,7 +447,7 @@ static int meson_pinconf_get_drive_strength(struct meson_pinctrl *pc,
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, pin, REG_DS, &reg, &bit);
+	meson_calc_reg_and_bit(bank, pin, MESON_REG_DS, &reg, &bit);
 
 	ret = regmap_read(pc->reg_ds, reg, &val);
 	if (ret)
@@ -595,7 +595,7 @@ static int meson_gpio_get(struct gpio_chip *chip, unsigned gpio)
 	if (ret)
 		return ret;
 
-	meson_calc_reg_and_bit(bank, gpio, REG_IN, &reg, &bit);
+	meson_calc_reg_and_bit(bank, gpio, MESON_REG_IN, &reg, &bit);
 	regmap_read(pc->reg_gpio, reg, &val);
 
 	return !!(val & BIT(bit));
diff --git a/drivers/pinctrl/meson/pinctrl-meson.h b/drivers/pinctrl/meson/pinctrl-meson.h
index ff5372e0a475..b197827027bd 100644
--- a/drivers/pinctrl/meson/pinctrl-meson.h
+++ b/drivers/pinctrl/meson/pinctrl-meson.h
@@ -63,13 +63,13 @@ struct meson_reg_desc {
  * enum meson_reg_type - type of registers encoded in @meson_reg_desc
  */
 enum meson_reg_type {
-	REG_PULLEN,
-	REG_PULL,
-	REG_DIR,
-	REG_OUT,
-	REG_IN,
-	REG_DS,
-	NUM_REG,
+	MESON_REG_PULLEN,
+	MESON_REG_PULL,
+	MESON_REG_DIR,
+	MESON_REG_OUT,
+	MESON_REG_IN,
+	MESON_REG_DS,
+	MESON_NUM_REG,
 };
 
 /**
@@ -102,7 +102,7 @@ struct meson_bank {
 	unsigned int last;
 	int irq_first;
 	int irq_last;
-	struct meson_reg_desc regs[NUM_REG];
+	struct meson_reg_desc regs[MESON_NUM_REG];
 };
 
 struct meson_pinctrl_data {
@@ -150,12 +150,12 @@ struct meson_pinctrl {
 		.irq_first	= fi,					\
 		.irq_last	= li,					\
 		.regs = {						\
-			[REG_PULLEN]	= { per, peb },			\
-			[REG_PULL]	= { pr, pb },			\
-			[REG_DIR]	= { dr, db },			\
-			[REG_OUT]	= { or, ob },			\
-			[REG_IN]	= { ir, ib },			\
-			[REG_DS]	= { dsr, dsb },			\
+			[MESON_REG_PULLEN]	= { per, peb },		\
+			[MESON_REG_PULL]	= { pr, pb },		\
+			[MESON_REG_DIR]		= { dr, db },		\
+			[MESON_REG_OUT]		= { or, ob },		\
+			[MESON_REG_IN]		= { ir, ib },		\
+			[MESON_REG_DS]		= { dsr, dsb },		\
 		},							\
 	 }
 
-- 
2.35.1

[PATCH v5 5/6] pinctrl: meson: Enable COMPILE_TEST

[Andy Shevchenko]

Enable COMPILE_TEST for a better test coverage.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
---
 drivers/pinctrl/meson/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/pinctrl/meson/Kconfig b/drivers/pinctrl/meson/Kconfig
index d1955c65b4b6..64fb9e074ac6 100644
--- a/drivers/pinctrl/meson/Kconfig
+++ b/drivers/pinctrl/meson/Kconfig
@@ -1,7 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0-only
 menuconfig PINCTRL_MESON
 	tristate "Amlogic SoC pinctrl drivers"
-	depends on ARCH_MESON
+	depends on ARCH_MESON || COMPILE_TEST
 	depends on OF
 	default y
 	select PINMUX
-- 
2.35.1

[PATCH v5 6/6] pinctrl: meson: Replace custom code by gpiochip_node_count() call

[Andy Shevchenko]

Since we have generic function to count GPIO controller nodes
under a given device, there is no need to open code it. Replace
custom code by gpiochip_node_count() call.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
---
 drivers/pinctrl/meson/pinctrl-meson.c | 28 ++++++++++++---------------
 1 file changed, 12 insertions(+), 16 deletions(-)

diff --git a/drivers/pinctrl/meson/pinctrl-meson.c b/drivers/pinctrl/meson/pinctrl-meson.c
index 5b46a0979db7..cc2cd73ff8f9 100644
--- a/drivers/pinctrl/meson/pinctrl-meson.c
+++ b/drivers/pinctrl/meson/pinctrl-meson.c
@@ -49,6 +49,7 @@
 #include <linux/pinctrl/pinctrl.h>
 #include <linux/pinctrl/pinmux.h>
 #include <linux/platform_device.h>
+#include <linux/property.h>
 #include <linux/regmap.h>
 #include <linux/seq_file.h>
 
@@ -662,27 +663,22 @@ static struct regmap *meson_map_resource(struct meson_pinctrl *pc,
 	return devm_regmap_init_mmio(pc->dev, base, &meson_regmap_config);
 }
 
-static int meson_pinctrl_parse_dt(struct meson_pinctrl *pc,
-				  struct device_node *node)
+static int meson_pinctrl_parse_dt(struct meson_pinctrl *pc)
 {
-	struct device_node *np, *gpio_np = NULL;
+	struct device_node *gpio_np;
+	unsigned int chips;
 
-	for_each_child_of_node(node, np) {
-		if (!of_find_property(np, "gpio-controller", NULL))
-			continue;
-		if (gpio_np) {
-			dev_err(pc->dev, "multiple gpio nodes\n");
-			of_node_put(np);
-			return -EINVAL;
-		}
-		gpio_np = np;
-	}
-
-	if (!gpio_np) {
+	chips = gpiochip_node_count(pc->dev);
+	if (!chips) {
 		dev_err(pc->dev, "no gpio node found\n");
 		return -EINVAL;
 	}
+	if (chips > 1) {
+		dev_err(pc->dev, "multiple gpio nodes\n");
+		return -EINVAL;
+	}
 
+	gpio_np = to_of_node(gpiochip_node_get_first(pc->dev));
 	pc->of_node = gpio_np;
 
 	pc->reg_mux = meson_map_resource(pc, gpio_np, "mux");
@@ -751,7 +747,7 @@ int meson_pinctrl_probe(struct platform_device *pdev)
 	pc->dev = dev;
 	pc->data = (struct meson_pinctrl_data *) of_device_get_match_data(dev);
 
-	ret = meson_pinctrl_parse_dt(pc, dev->of_node);
+	ret = meson_pinctrl_parse_dt(pc);
 	if (ret)
 		return ret;
 
-- 
2.35.1

Re: [PATCH v5 0/6] gpiolib: more helpers and fwnode conversion

[Marek Szyprowski]

Hi Andy,

On 14.04.2022 21:02, Andy Shevchenko wrote:
> This is a spin-off (*) of the previous work of switching GPIO library
> to use fwnode instead of of_node. Here we introduce a couple of
> a new macro helpers, which allows to switch some of the drivers
> to use fwnode and partially fwnode APIs. As a result of this cleanup
> a few drivers switched to use GPIO fwnode instead of of_node.
>
> *) it's subset of it with a new (patch 1) helper.
>
> Marek, Martin, can you give this a try?
> This requires at least two patches for GPIO library to be applied.

I've applied patch #1 and #6 on top of linux next-20220413 with commit 
88834c75cae5 ("pinctrl: meson: Replace custom code by 
gpiochip_node_count() call") reverted. All my Meson-based test boards 
(Odroid C4, N2, Khadas VIM3/3l) work fine now. Thanks! Feel free to add:

Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>

> Bart, Linus, I can take it thru my tree with an immutable branch if
> it's the way you prefer, otherwise please suggest on how to proceed.
>
> Changelog v5:
> - dropped tested patches (this series based on them, though)
> - introduced a new helper (thanks Marek and Martin for reporting an issue)
> - redone Armada and Meson code using newly introduced helper
>
> Changelog v4:
> - fixed compilation of the Samsung pin control drivers (LKP)
> - explained in the commit message why namespacing is good for meson defs
> - added tag to one of meson patches (Neil)
>
> Changelog v3:
> - moved count initialization to the definition in patch 2 (Geert)
> - replaced of_args by args, used %pfwP in patch 7 (Geert)
> - fixed kernel doc warning in patch 7
> - added tags to patches 1, 2, 6, and 7 (Geert)
> - added tag to patch 4 (Fabien)
> - renamed MREG to MESON_REG in patch 9 (Neil)
> - added tag to patch 10 (Neil)
> - used --base for cover-letter
>
> Changelog v2:
> - properly based, so kbuild bot may test it (LKP)
> - fixed typo in the macro (Geert)
> - split to two macro helpers and rename the gpiochip_count()
> - tagged one of stm32 and one of meson patches (Fabien, Neil)
> - unified previously standalone armada patch
> - due to above rewrote the armada patch from v1 completely (Sergey)
> - added a lot of a new patches
> - compile tested all of them on x86
>
> Andy Shevchenko (6):
>    gpiolib: Introduce a helper to get first GPIO controller node
>    pinctrl: armada-37xx: Switch to use fwnode instead of of_node
>    pinctrl: armada-37xx: Reuse GPIO fwnode in
>      armada_37xx_irqchip_register()
>    pinctrl: meson: Rename REG_* to MESON_REG_*
>    pinctrl: meson: Enable COMPILE_TEST
>    pinctrl: meson: Replace custom code by gpiochip_node_count() call
>
>   drivers/pinctrl/meson/Kconfig               |  2 +-
>   drivers/pinctrl/meson/pinctrl-meson.c       | 52 ++++++++++-----------
>   drivers/pinctrl/meson/pinctrl-meson.h       | 28 +++++------
>   drivers/pinctrl/mvebu/pinctrl-armada-37xx.c | 34 ++++----------
>   include/linux/gpio/driver.h                 | 10 ++++
>   5 files changed, 59 insertions(+), 67 deletions(-)
>
Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

Re: [PATCH v5 0/6] gpiolib: more helpers and fwnode conversion

[Linus Walleij]

On Thu, Apr 14, 2022 at 9:02 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:

> This is a spin-off (*) of the previous work of switching GPIO library
> to use fwnode instead of of_node. Here we introduce a couple of
> a new macro helpers, which allows to switch some of the drivers
> to use fwnode and partially fwnode APIs. As a result of this cleanup
> a few drivers switched to use GPIO fwnode instead of of_node.
>
> *) it's subset of it with a new (patch 1) helper.
>
> Marek, Martin, can you give this a try?
> This requires at least two patches for GPIO library to be applied.
>
> Bart, Linus, I can take it thru my tree with an immutable branch if
> it's the way you prefer, otherwise please suggest on how to proceed.

Hmmm that sounds best, patch 1 does not apply to the pinctrl
tree so I suppose there are already dependencies in the GPIO
tree?

FWIW, the series:
Acked-by: Linus Walleij <linus.walleij@linaro.org>

I'm of course a fan of this nice refactoring series.

Yours,
Linus Walleij