diff mbox series

[v5,2/8] leds: document additional use of blink_set for hardware control

Message ID 20211112153557.26941-3-ansuelsmth@gmail.com
State New
Headers show
Series Adds support for PHY LEDs with offload triggers | expand

Commit Message

Christian Marangi Nov. 12, 2021, 3:35 p.m. UTC 
blink_set can now be used also to configure blink modes when hardware
control is supported and active.
Trigger will try to use blink_set and a LED driver will use the
trigger_data to configure the blink_modes once blink_set is called.
Setting a brightness to LED_OFF will reset any blink mode and disable
hardware control setting the LED off.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 Documentation/leds/leds-class.rst | 17 +++++++++++++++++
 drivers/leds/led-class.c          |  7 +++++++
 include/linux/leds.h              |  9 +++++++++
 3 files changed, 33 insertions(+)

Comments

Randy Dunlap Nov. 13, 2021, 12:55 a.m. UTC | #1 
On 11/12/21 7:35 AM, Ansuel Smith wrote:
> diff --git a/Documentation/leds/leds-class.rst b/Documentation/leds/leds-class.rst
> index e5d266919a19..fa6e1cfc4628 100644
> --- a/Documentation/leds/leds-class.rst
> +++ b/Documentation/leds/leds-class.rst
> @@ -202,6 +202,23 @@ hardware mode and any software only trigger will reject activation.
>   On init an LED driver that support a hardware mode should reset every blink mode
>   set by default.
>   
> +Once a trigger has declared support for hardware-controlled blinks, it will use
> +blink_set() to try to offload his trigger on activation/configuration.

                                  this trigger

> +blink_set() will return 0 if the requested modes set in trigger_data can be
> +controlled by hardware or an error if both of the bitmap modes are not supported by
> +the hardware or there was a problem in the configuration.
> +
> +Following blink_set logic, setting brightness to LED_OFF with hardware control active
> +will reset any active blink mode and disable hardware control setting the LED to off.
> +
> +It's in the LED driver's interest to know how to elaborate the trigger data and report support
> +for a particular set of blink modes. For this exact reason explicit support for the specific
> +trigger is mandatory or the driver returns -EOPNOTSUPP if asked to enter hardware mode
> +with a not supported trigger.
> +If the driver returns -EOPNOTSUPP on hw_control_configure(), the trigger activation will
> +fail as the driver doesn't support that specific hardware blink mode or doesn't know
> +how to handle the provided trigger data.
diff mbox series

Patch

diff --git a/Documentation/leds/leds-class.rst b/Documentation/leds/leds-class.rst
index e5d266919a19..fa6e1cfc4628 100644
--- a/Documentation/leds/leds-class.rst
+++ b/Documentation/leds/leds-class.rst
@@ -202,6 +202,23 @@  hardware mode and any software only trigger will reject activation.
 On init an LED driver that support a hardware mode should reset every blink mode
 set by default.
 
+Once a trigger has declared support for hardware-controlled blinks, it will use
+blink_set() to try to offload his trigger on activation/configuration.
+blink_set() will return 0 if the requested modes set in trigger_data can be
+controlled by hardware or an error if both of the bitmap modes are not supported by
+the hardware or there was a problem in the configuration.
+
+Following blink_set logic, setting brightness to LED_OFF with hardware control active
+will reset any active blink mode and disable hardware control setting the LED to off.
+
+It's in the LED driver's interest to know how to elaborate the trigger data and report support
+for a particular set of blink modes. For this exact reason explicit support for the specific
+trigger is mandatory or the driver returns -EOPNOTSUPP if asked to enter hardware mode
+with a not supported trigger.
+If the driver returns -EOPNOTSUPP on hw_control_configure(), the trigger activation will
+fail as the driver doesn't support that specific hardware blink mode or doesn't know
+how to handle the provided trigger data.
+
 Known Issues
 ============
 
diff --git a/drivers/leds/led-class.c b/drivers/leds/led-class.c
index dbe9840863d0..ebd023e480ac 100644
--- a/drivers/leds/led-class.c
+++ b/drivers/leds/led-class.c
@@ -180,6 +180,13 @@  static int led_classdev_check_hw_control_functions(struct led_classdev *led_cdev
 	    led_cdev->hw_control_stop))
 		return -EINVAL;
 
+	/* blink_set is mandatory to configure the blink modes
+	 * in hardware control.
+	 */
+	if ((LED_HARDWARE_CONTROLLED & led_cdev->flags) &&
+	    !led_cdev->blink_set)
+		return -EINVAL;
+
 	return 0;
 }
 #endif
diff --git a/include/linux/leds.h b/include/linux/leds.h
index c8e97fb9252f..fa929215cdcc 100644
--- a/include/linux/leds.h
+++ b/include/linux/leds.h
@@ -76,6 +76,7 @@  struct led_classdev {
 	/* Lower 16 bits reflect status */
 #define LED_SUSPENDED		BIT(0)
 #define LED_UNREGISTERING	BIT(1)
+#define LED_HARDWARE_CONTROL	BIT(2)
 	/* Upper 16 bits reflect control information */
 #define LED_CORE_SUSPENDRESUME	BIT(16)
 #define LED_SYSFS_DISABLE	BIT(17)
@@ -123,6 +124,12 @@  struct led_classdev {
 	 * match the values specified exactly.
 	 * Deactivate blinking again when the brightness is set to LED_OFF
 	 * via the brightness_set() callback.
+	 * With LED_HARDWARE_CONTROL set in LED flags blink_set will also
+	 * configure blink modes requested by the current trigger if
+	 * supported by the LED driver.
+	 * Setting brightness to LED_OFF with hardware control active will
+	 * reset any active blink mode and disable hardware control setting
+	 * the LED off.
 	 */
 	int		(*blink_set)(struct led_classdev *led_cdev,
 				     unsigned long *delay_on,
@@ -165,6 +172,8 @@  struct led_classdev {
 	 */
 	bool			(*hw_control_status)(struct led_classdev *led_cdev);
 	/* Set LED in hardware mode and reset any blink mode active by default.
+	 * A trigger supporting hardware mode will have to use blink_set to configure
+	 * the modes.
 	 */
 	int			(*hw_control_start)(struct led_classdev *led_cdev);
 	/* Disable hardware mode for LED. It's advised to the LED driver to put it to