[PATCH] backlight: led_bl: Add support for an "enable" GPIO

From: Jean-Jacques Hiblot <redacted>

This patch adds support for an "enable GPIO".

Signed-off-by: Jean-Jacques Hiblot <redacted>
Signed-off-by: Corentin LABBE <clabbe@baylibre.com>
---
 drivers/video/backlight/led_bl.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/drivers/video/backlight/led_bl.c b/drivers/video/backlight/led_bl.c
index f54d256e2d54..ebd7acc32980 100644
--- a/drivers/video/backlight/led_bl.c
+++ b/drivers/video/backlight/led_bl.c

@@ -8,6 +8,7 @@
 
 #include <linux/backlight.h>
 #include <linux/leds.h>
+#include <linux/gpio.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
 

@@ -15,6 +16,7 @@ struct led_bl_data {
 	struct device		*dev;
 	struct backlight_device	*bl_dev;
 	struct led_classdev	**leds;
+	struct gpio_desc        *enable_gpio;
 	bool			enabled;
 	int			nb_leds;
 	unsigned int		*levels;

@@ -35,6 +37,9 @@ static void led_bl_set_brightness(struct led_bl_data *priv, int level)
 	for (i = 0; i < priv->nb_leds; i++)
 		led_set_brightness(priv->leds[i], bkl_brightness);
 
+	if (!priv->enabled)
+		gpiod_set_value_cansleep(priv->enable_gpio, 1);
+
 	priv->enabled = true;
 }
 

@@ -48,6 +53,9 @@ static void led_bl_power_off(struct led_bl_data *priv)
 	for (i = 0; i < priv->nb_leds; i++)
 		led_set_brightness(priv->leds[i], LED_OFF);
 
+	if (priv->enabled)
+		gpiod_set_value_cansleep(priv->enable_gpio, 0);
+
 	priv->enabled = false;
 }
 

@@ -209,6 +217,11 @@ static int led_bl_probe(struct platform_device *pdev)
 		return PTR_ERR(priv->bl_dev);
 	}
 
+	priv->enable_gpio = devm_gpiod_get_optional(&pdev->dev, "enable",
+						    GPIOD_OUT_LOW);
+	if (IS_ERR(priv->enable_gpio))
+		return PTR_ERR(priv->enable_gpio);
+
 	for (i = 0; i < priv->nb_leds; i++)
 		led_sysfs_disable(priv->leds[i]);
 

-- 
2.25.1

On Tue, Nov 02, 2021 at 10:04:55AM +0000, Corentin LABBE wrote:

From: Jean-Jacques Hiblot <redacted>

This patch adds support for an "enable GPIO".

Before taking this kind of change is there a good reason why backlight
should manage the GPIO? I thought that the LED subsystem was a sub
system for LEDs (not LED controllers). In other words if you direct
that the LED be lit up then isn't it the LED driver's job to manage
the GPIO and ensure that it lights up?


Daniel.

On Tue, Nov 02, 2021 at 11:19:42AM +0000, Daniel Thompson wrote:

On Tue, Nov 02, 2021 at 10:04:55AM +0000, Corentin LABBE wrote:

quoted

From: Jean-Jacques Hiblot <redacted>

This patch adds support for an "enable GPIO".

Before taking this kind of change is there a good reason why backlight
should manage the GPIO? I thought that the LED subsystem was a sub
system for LEDs (not LED controllers). In other words if you direct
that the LED be lit up then isn't it the LED driver's job to manage
the GPIO and ensure that it lights up?

Sorry... I should have paid more attention to my sense of déjà vu with
this patch.

This approach was discussed and rejected when we first introduced the
led_bl driver:
https://lore.kernel.org/linux-leds/20190705100851.zn2jkipj4fxq5we6@devuan/


Daniel.

quoted

Signed-off-by: Jean-Jacques Hiblot <redacted>
Signed-off-by: Corentin LABBE <clabbe@baylibre.com>
---
 drivers/video/backlight/led_bl.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/drivers/video/backlight/led_bl.c b/drivers/video/backlight/led_bl.c
index f54d256e2d54..ebd7acc32980 100644
--- a/drivers/video/backlight/led_bl.c
+++ b/drivers/video/backlight/led_bl.c

@@ -8,6 +8,7 @@
 
 #include <linux/backlight.h>
 #include <linux/leds.h>
+#include <linux/gpio.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
 

@@ -15,6 +16,7 @@ struct led_bl_data {
 	struct device		*dev;
 	struct backlight_device	*bl_dev;
 	struct led_classdev	**leds;
+	struct gpio_desc        *enable_gpio;
 	bool			enabled;
 	int			nb_leds;
 	unsigned int		*levels;

@@ -35,6 +37,9 @@ static void led_bl_set_brightness(struct led_bl_data *priv, int level)
 	for (i = 0; i < priv->nb_leds; i++)
 		led_set_brightness(priv->leds[i], bkl_brightness);
 
+	if (!priv->enabled)
+		gpiod_set_value_cansleep(priv->enable_gpio, 1);
+
 	priv->enabled = true;
 }
 

@@ -48,6 +53,9 @@ static void led_bl_power_off(struct led_bl_data *priv)
 	for (i = 0; i < priv->nb_leds; i++)
 		led_set_brightness(priv->leds[i], LED_OFF);
 
+	if (priv->enabled)
+		gpiod_set_value_cansleep(priv->enable_gpio, 0);
+
 	priv->enabled = false;
 }
 

@@ -209,6 +217,11 @@ static int led_bl_probe(struct platform_device *pdev)
 		return PTR_ERR(priv->bl_dev);
 	}
 
+	priv->enable_gpio = devm_gpiod_get_optional(&pdev->dev, "enable",
+						    GPIOD_OUT_LOW);
+	if (IS_ERR(priv->enable_gpio))
+		return PTR_ERR(priv->enable_gpio);
+
 	for (i = 0; i < priv->nb_leds; i++)
 		led_sysfs_disable(priv->leds[i]);
 

-- 
2.25.1

Le Tue, Nov 02, 2021 at 11:25:14AM +0000, Daniel Thompson a écrit :

On Tue, Nov 02, 2021 at 11:19:42AM +0000, Daniel Thompson wrote:

quoted

On Tue, Nov 02, 2021 at 10:04:55AM +0000, Corentin LABBE wrote:

quoted

From: Jean-Jacques Hiblot <redacted>

This patch adds support for an "enable GPIO".

Before taking this kind of change is there a good reason why backlight
should manage the GPIO? I thought that the LED subsystem was a sub
system for LEDs (not LED controllers). In other words if you direct
that the LED be lit up then isn't it the LED driver's job to manage
the GPIO and ensure that it lights up?

Sorry... I should have paid more attention to my sense of déjà vu with
this patch.

This approach was discussed and rejected when we first introduced the
led_bl driver:
https://lore.kernel.org/linux-leds/20190705100851.zn2jkipj4fxq5we6@devuan/

Hello

I am sorry, I didnt checked if the patch was already submitted or not.

Regards