Re: [PATCH v2 6/7] platform/x86: ideapad-laptop: Keyboard backlight support for more IdeaPads

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Eray,

On 10/29/22 14:03, Eray Orçunus wrote:
> IdeaPads with HALS_KBD_BL_SUPPORT_BIT have full keyboard light support,
> and they send an event via ACPI on light state change. Whereas some
> IdeaPads that don't have this bit set, i.e. 520-15ikb, 330-17ich and
> 5 (15), don't send an event, yet they still support switching keyboard
> light via KBLO object on DSDT. Detect these IdeaPads with searching for
> KBLO object, set their kbd_bl_partial to true and register led device
> for them. Tested on 520-15ikb.
> 
> Signed-off-by: Eray Orçunus <erayorcunus@xxxxxxxxx>
> ---
>  drivers/platform/x86/ideapad-laptop.c | 79 ++++++++++++++++++++++++---
>  1 file changed, 70 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/platform/x86/ideapad-laptop.c b/drivers/platform/x86/ideapad-laptop.c
> index e8c088e7a53d..b34fbc4d741c 100644
> --- a/drivers/platform/x86/ideapad-laptop.c
> +++ b/drivers/platform/x86/ideapad-laptop.c
> @@ -149,6 +149,7 @@ struct ideapad_private {
>  		bool fn_lock              : 1;
>  		bool hw_rfkill_switch     : 1;
>  		bool kbd_bl               : 1;
> +		bool kbd_bl_partial       : 1;
>  		bool cam_ctrl_via_ec      : 1;
>  		bool touchpad_ctrl_via_ec : 1;
>  		bool usb_charging         : 1;
> @@ -157,6 +158,9 @@ struct ideapad_private {
>  		bool initialized;
>  		struct led_classdev led;
>  		unsigned int last_brightness;
> +		/* Below are used only if kbd_bl_partial is set */
> +		acpi_handle lfcm_mutex;
> +		acpi_handle kblo_obj;
>  	} kbd_bl;
>  };
>  
> @@ -1300,19 +1304,52 @@ static void ideapad_backlight_notify_brightness(struct ideapad_private *priv)
>  		backlight_force_update(priv->blightdev, BACKLIGHT_UPDATE_HOTKEY);
>  }
>  
> +#define IDEAPAD_ACPI_MUTEX_TIMEOUT 1500
> +
>  /*
>   * keyboard backlight
>   */
>  static int ideapad_kbd_bl_brightness_get(struct ideapad_private *priv)
>  {
> -	unsigned long hals;
> +	unsigned long ret_val;
>  	int err;
> +	acpi_status status;
>  
> -	err = eval_hals(priv->adev->handle, &hals);
> +	/*
> +	 * Some IdeaPads with partially implemented keyboard lights don't give
> +	 * us the light state on HALS_KBD_BL_STATE_BIT in the return value of HALS,
> +	 * i.e. 5 (15) and 330-17ich. Fortunately we know how to gather it.
> +	 * Even if it won't work, we will still give HALS a try, because
> +	 * some IdeaPads with kbd_bl_partial, i.e. 520-15ikb,
> +	 * correctly sets HALS_KBD_BL_STATE_BIT in HALS return value.
> +	 */
> +
> +	if (priv->features.kbd_bl_partial &&
> +	    priv->kbd_bl.lfcm_mutex != NULL && priv->kbd_bl.kblo_obj != NULL) {

IMHO it would be better to only set kbd_bl_partial when both handles
are not NULL, then you can drop the handle checks here.

> +
> +		status = acpi_acquire_mutex(priv->kbd_bl.lfcm_mutex, NULL,
> +					    IDEAPAD_ACPI_MUTEX_TIMEOUT);
> +
> +		if (ACPI_SUCCESS(status)) {

This code now ends up still going through the normal kbd-bl path
when it fails to acquire the mutex.

Instead it should do:

		if (ACPI_FAILURE(status))
			return -EIO;

And then have the rest of the code one indentation level less
deep.

> +			err = eval_int(priv->kbd_bl.kblo_obj, NULL, &ret_val);
> +
> +			status = acpi_release_mutex(priv->kbd_bl.lfcm_mutex, NULL);
> +			if (ACPI_FAILURE(status))
> +				dev_err(&priv->platform_device->dev,
> +					"Failed to release LFCM mutex");

I'm pretty sure that the ACPI core will already log an error if things
fail, I would change this to just a single line:

		acpi_release_mutex(priv->kbd_bl.lfcm_mutex, NULL);


> +
> +			if (err)
> +				return err;
> +
> +			return !!ret_val;

!!ret_val turns it into a boolean, does that mean it is always either on
or off with no level in between ?

> +		}
> +	}
> +
> +	err = eval_hals(priv->adev->handle, &ret_val);
>  	if (err)
>  		return err;
>  
> -	return !!test_bit(HALS_KBD_BL_STATE_BIT, &hals);
> +	return !!test_bit(HALS_KBD_BL_STATE_BIT, &ret_val);
>  }
>  
>  static enum led_brightness ideapad_kbd_bl_led_cdev_brightness_get(struct led_classdev *led_cdev)
> @@ -1329,7 +1366,8 @@ static int ideapad_kbd_bl_brightness_set(struct ideapad_private *priv, unsigned
>  	if (err)
>  		return err;
>  
> -	priv->kbd_bl.last_brightness = brightness;
> +	if (!priv->features.kbd_bl_partial)
> +		priv->kbd_bl.last_brightness = brightness;
>  
>  	return 0;
>  }

I don't understand this change, you change ideapad_kbd_bl_brightness_get()
to do an int eval of KBLO, but here you now still do a
exec_sals(SALS_KBD_BL_ON / SALS_KBD_BL_OFF) ?

Also there is no reason not to update last_brightness here ...

> @@ -1349,6 +1387,9 @@ static void ideapad_kbd_bl_notify(struct ideapad_private *priv)
>  	if (!priv->kbd_bl.initialized)
>  		return;
>  
> +	if (priv->features.kbd_bl_partial)
> +		return;
> +

Why? If we do happen to get a notify on one of these devices and
the brightness has changed, then it would be good to let userspace
know and if never get a notify then this function won't run so
we don't need the if.

>  	brightness = ideapad_kbd_bl_brightness_get(priv);
>  	if (brightness < 0)
>  		return;
> @@ -1371,17 +1412,20 @@ static int ideapad_kbd_bl_init(struct ideapad_private *priv)
>  	if (WARN_ON(priv->kbd_bl.initialized))
>  		return -EEXIST;
>  
> -	brightness = ideapad_kbd_bl_brightness_get(priv);
> -	if (brightness < 0)
> -		return brightness;
> +	/* IdeaPads with kbd_bl_partial don't have keyboard backlight event */
> +	if (!priv->features.kbd_bl_partial) {
> +		brightness = ideapad_kbd_bl_brightness_get(priv);
> +		if (brightness < 0)
> +			return brightness;
>  
> -	priv->kbd_bl.last_brightness = brightness;
> +		priv->kbd_bl.last_brightness         = brightness;
> +		priv->kbd_bl.led.flags               = LED_BRIGHT_HW_CHANGED;
> +	}

Again no need to for the if here. Setting last_brightness and advertising
LED_BRIGHT_HW_CHANGED unconditionally won't hurt and not making this difference
keeps the code simpler.

>  
>  	priv->kbd_bl.led.name                    = "platform::" LED_FUNCTION_KBD_BACKLIGHT;
>  	priv->kbd_bl.led.max_brightness          = 1;
>  	priv->kbd_bl.led.brightness_get          = ideapad_kbd_bl_led_cdev_brightness_get;
>  	priv->kbd_bl.led.brightness_set_blocking = ideapad_kbd_bl_led_cdev_brightness_set;
> -	priv->kbd_bl.led.flags                   = LED_BRIGHT_HW_CHANGED;
>  
>  	err = led_classdev_register(&priv->platform_device->dev, &priv->kbd_bl.led);
>  	if (err)
> @@ -1594,8 +1638,25 @@ static void ideapad_check_features(struct ideapad_private *priv)
>  			if (test_bit(HALS_FNLOCK_SUPPORT_BIT, &val))
>  				priv->features.fn_lock = true;
>  
> +			/*
> +			 * IdeaPads with HALS_KBD_BL_SUPPORT_BIT have full keyboard
> +			 * light support, and they send an event via ACPI on light
> +			 * state change. Whereas some IdeaPads, at least 520-15ikb
> +			 * and 5 (15), don't send an event, yet they still have
> +			 * KBLO object. In this case, set kbd_bl_partial to true
> +			 * and cache the LFCM mutex, it might be useful while
> +			 * getting the brightness.
> +			 */
> +
>  			if (test_bit(HALS_KBD_BL_SUPPORT_BIT, &val))
>  				priv->features.kbd_bl = true;
> +			else if (ACPI_SUCCESS(acpi_get_handle(handle, "^KBLO",
> +							      &priv->kbd_bl.kblo_obj))) {

As mentioned above I would change this to:

			else if (ACPI_SUCCESS(acpi_get_handle(handle, "^KBLO",
							      &priv->kbd_bl.kblo_obj)) &&
				 ACPI_SUCCESS(acpi_get_handle(handle, "^LFCM",
							      &priv->kbd_bl.lfcm_mutex))) {

> +				priv->features.kbd_bl = true;
> +				priv->features.kbd_bl_partial = true;
> +				(void)acpi_get_handle(handle, "^LFCM",
> +						      &priv->kbd_bl.lfcm_mutex);

And then drop the acpi_get_handle() here.


> +			}
>  
>  			if (test_bit(HALS_USB_CHARGING_SUPPORT_BIT, &val))
>  				priv->features.usb_charging = true;


Regards,

Hans





[Index of Archives]     [Linux Kernel Development]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux