[PATCH] mfd: stmpe: switch to using gpiod API

[PATCH] mfd: stmpe: switch to using gpiod API

[Dmitry Torokhov]

This patch switches the driver away from legacy gpio/of_gpio API to
gpiod API, and removes use of of_get_named_gpio_flags() which I want to
make private to gpiolib.

Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
 drivers/mfd/stmpe.c | 36 +++++++++++++-----------------------
 1 file changed, 13 insertions(+), 23 deletions(-)

diff --git a/drivers/mfd/stmpe.c b/drivers/mfd/stmpe.c
index 987e251d90ae..0c4f74197d3e 100644
--- a/drivers/mfd/stmpe.c
+++ b/drivers/mfd/stmpe.c
@@ -8,14 +8,13 @@
  */
 
 #include <linux/err.h>
-#include <linux/gpio.h>
+#include <linux/gpio/consumer.h>
 #include <linux/export.h>
 #include <linux/kernel.h>
 #include <linux/interrupt.h>
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/of.h>
-#include <linux/of_gpio.h>
 #include <linux/pm.h>
 #include <linux/slab.h>
 #include <linux/mfd/core.h>
@@ -30,17 +29,12 @@
  * @irq_trigger: IRQ trigger to use for the interrupt to the host
  * @autosleep: bool to enable/disable stmpe autosleep
  * @autosleep_timeout: inactivity timeout in milliseconds for autosleep
- * @irq_over_gpio: true if gpio is used to get irq
- * @irq_gpio: gpio number over which irq will be requested (significant only if
- *	      irq_over_gpio is true)
  */
 struct stmpe_platform_data {
 	int id;
 	unsigned int blocks;
 	unsigned int irq_trigger;
 	bool autosleep;
-	bool irq_over_gpio;
-	int irq_gpio;
 	int autosleep_timeout;
 };
 
@@ -1349,13 +1343,6 @@ static void stmpe_of_probe(struct stmpe_platform_data *pdata,
 	if (pdata->id < 0)
 		pdata->id = -1;
 
-	pdata->irq_gpio = of_get_named_gpio_flags(np, "irq-gpio", 0,
-				&pdata->irq_trigger);
-	if (gpio_is_valid(pdata->irq_gpio))
-		pdata->irq_over_gpio = 1;
-	else
-		pdata->irq_trigger = IRQF_TRIGGER_NONE;
-
 	of_property_read_u32(np, "st,autosleep-timeout",
 			&pdata->autosleep_timeout);
 
@@ -1381,6 +1368,7 @@ int stmpe_probe(struct stmpe_client_info *ci, enum stmpe_partnum partnum)
 	struct stmpe_platform_data *pdata;
 	struct device_node *np = ci->dev->of_node;
 	struct stmpe *stmpe;
+	struct gpio_desc *irq_gpio;
 	int ret;
 	u32 val;
 
@@ -1434,18 +1422,20 @@ int stmpe_probe(struct stmpe_client_info *ci, enum stmpe_partnum partnum)
 	if (ci->init)
 		ci->init(stmpe);
 
-	if (pdata->irq_over_gpio) {
-		ret = devm_gpio_request_one(ci->dev, pdata->irq_gpio,
-				GPIOF_DIR_IN, "stmpe");
-		if (ret) {
-			dev_err(stmpe->dev, "failed to request IRQ GPIO: %d\n",
-					ret);
-			return ret;
-		}
+	irq_gpio = devm_gpiod_get_optional(ci->dev, "irq", GPIOD_ASIS);
+	ret = PTR_ERR_OR_ZERO(irq_gpio);
+	if (ret) {
+		dev_err(stmpe->dev, "failed to request IRQ GPIO: %d\n", ret);
+		return ret;
+	}
 
-		stmpe->irq = gpio_to_irq(pdata->irq_gpio);
+	if (irq_gpio) {
+		stmpe->irq = gpiod_to_irq(irq_gpio);
+		pdata->irq_trigger = gpiod_is_active_low(irq_gpio) ?
+					IRQF_TRIGGER_LOW : IRQF_TRIGGER_HIGH;
 	} else {
 		stmpe->irq = ci->irq;
+		pdata->irq_trigger = IRQF_TRIGGER_NONE;
 	}
 
 	if (stmpe->irq < 0) {
-- 
2.37.2.789.g6183377224-goog


-- 
Dmitry

Re: [PATCH] mfd: stmpe: switch to using gpiod API

[Francesco Dolcini]

On Mon, Sep 05, 2022 at 10:01:47PM -0700, Dmitry Torokhov wrote:
> This patch switches the driver away from legacy gpio/of_gpio API to
> gpiod API, and removes use of of_get_named_gpio_flags() which I want to
> make private to gpiolib.
> 
> Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
> ---
>  drivers/mfd/stmpe.c | 36 +++++++++++++-----------------------
>  1 file changed, 13 insertions(+), 23 deletions(-)
> 
> diff --git a/drivers/mfd/stmpe.c b/drivers/mfd/stmpe.c
> index 987e251d90ae..0c4f74197d3e 100644
> --- a/drivers/mfd/stmpe.c
> +++ b/drivers/mfd/stmpe.c
> @@ -8,14 +8,13 @@
>   */
>  

<snip>

> -	pdata->irq_gpio = of_get_named_gpio_flags(np, "irq-gpio", 0,
<snip>
> +	irq_gpio = devm_gpiod_get_optional(ci->dev, "irq", GPIOD_ASIS);
isn't this changing from irq-gpio to irq-gpios property name?

in the DTS files we do have something like that:

 irq-gpio = <&gpio TEGRA_GPIO(V, 0) IRQ_TYPE_LEVEL_LOW>;


Francesco

Re: [PATCH] mfd: stmpe: switch to using gpiod API

[Linus Walleij]

On Tue, Sep 6, 2022 at 12:11 PM Francesco Dolcini
<francesco.dolcini@toradex.com> wrote:
> On Mon, Sep 05, 2022 at 10:01:47PM -0700, Dmitry Torokhov wrote:
> > This patch switches the driver away from legacy gpio/of_gpio API to
> > gpiod API, and removes use of of_get_named_gpio_flags() which I want to
> > make private to gpiolib.
> >
> > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
> > ---
> >  drivers/mfd/stmpe.c | 36 +++++++++++++-----------------------
> >  1 file changed, 13 insertions(+), 23 deletions(-)
> >
> > diff --git a/drivers/mfd/stmpe.c b/drivers/mfd/stmpe.c
> > index 987e251d90ae..0c4f74197d3e 100644
> > --- a/drivers/mfd/stmpe.c
> > +++ b/drivers/mfd/stmpe.c
> > @@ -8,14 +8,13 @@
> >   */
> >
>
> <snip>
>
> > -     pdata->irq_gpio = of_get_named_gpio_flags(np, "irq-gpio", 0,
> <snip>
> > +     irq_gpio = devm_gpiod_get_optional(ci->dev, "irq", GPIOD_ASIS);
> isn't this changing from irq-gpio to irq-gpios property name?

The gpiolib core will automatically append and test the strings
*-gpios and *-gpio (in that order)

> in the DTS files we do have something like that:
>
>  irq-gpio = <&gpio TEGRA_GPIO(V, 0) IRQ_TYPE_LEVEL_LOW>;

This seems to be an orthogonal bug.

That flag in the last cell is invalid for a GPIO, the DT schema should
scream at you if you have one. GPIO flags are
GPIO_ACTIVE_LOW, GPIO_OPEN_DRAIN etc. That looks more
like an IRQ, and then the property should be irqs = <...>.

Yours,
Linus Walleij

Re: [PATCH] mfd: stmpe: switch to using gpiod API

[Dmitry Torokhov]

On Tue, Sep 06, 2022 at 01:29:21PM +0200, Linus Walleij wrote:
> On Tue, Sep 6, 2022 at 12:11 PM Francesco Dolcini
> <francesco.dolcini@toradex.com> wrote:
> > On Mon, Sep 05, 2022 at 10:01:47PM -0700, Dmitry Torokhov wrote:
> > > This patch switches the driver away from legacy gpio/of_gpio API to
> > > gpiod API, and removes use of of_get_named_gpio_flags() which I want to
> > > make private to gpiolib.
> > >
> > > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
> > > ---
> > >  drivers/mfd/stmpe.c | 36 +++++++++++++-----------------------
> > >  1 file changed, 13 insertions(+), 23 deletions(-)
> > >
> > > diff --git a/drivers/mfd/stmpe.c b/drivers/mfd/stmpe.c
> > > index 987e251d90ae..0c4f74197d3e 100644
> > > --- a/drivers/mfd/stmpe.c
> > > +++ b/drivers/mfd/stmpe.c
> > > @@ -8,14 +8,13 @@
> > >   */
> > >
> >
> > <snip>
> >
> > > -     pdata->irq_gpio = of_get_named_gpio_flags(np, "irq-gpio", 0,
> > <snip>
> > > +     irq_gpio = devm_gpiod_get_optional(ci->dev, "irq", GPIOD_ASIS);
> > isn't this changing from irq-gpio to irq-gpios property name?
> 
> The gpiolib core will automatically append and test the strings
> *-gpios and *-gpio (in that order)
> 
> > in the DTS files we do have something like that:
> >
> >  irq-gpio = <&gpio TEGRA_GPIO(V, 0) IRQ_TYPE_LEVEL_LOW>;
> 
> This seems to be an orthogonal bug.
> 
> That flag in the last cell is invalid for a GPIO, the DT schema should
> scream at you if you have one. GPIO flags are
> GPIO_ACTIVE_LOW, GPIO_OPEN_DRAIN etc. That looks more
> like an IRQ, and then the property should be irqs = <...>.

Francesco is right though in the sense that we need to update DTS
together with the patch, or we will break the driver... The original
code relied on the fact that flags returned by of_get_named_gpio_flags()
could contain anything, not necessarily valid GPIO flags.

Thanks.

-- 
Dmitry