This fixes suspend on platforms like stm32mp1xx, where the PWM consumer has to be disabled for the PWM to enter suspend. Another positive side effect is that active-low LEDs now properly turn off instead of going back to full brightness when they are set to 0. Link: https://lore.kernel.org/all/20240417153846.271751-2-u.kleine-koenig@xxxxxxxxxxxxxx/ Signed-off-by: Jakob Riepler <jakob+lkml@xxxxxxxxxxxxx> --- Changes in v2: - fix wrong line-breaks in patch Changes in v3: - use git send-email Sorry for the issues. I'm using git send-email with a different mailserver now so the patch formatting should now be correct - at least in my tests git am was happy now. Best regards Jakob drivers/leds/rgb/leds-pwm-multicolor.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/leds/rgb/leds-pwm-multicolor.c b/drivers/leds/rgb/leds-pwm-multicolor.c index e1a81e0109e8..f80a06cc31f8 100644 --- a/drivers/leds/rgb/leds-pwm-multicolor.c +++ b/drivers/leds/rgb/leds-pwm-multicolor.c @@ -50,7 +50,13 @@ static int led_pwm_mc_set(struct led_classdev *cdev, duty = priv->leds[i].state.period - duty; priv->leds[i].state.duty_cycle = duty; - priv->leds[i].state.enabled = duty > 0; + /* + * Disabling a PWM doesn't guarantee that it emits the inactive level. + * So keep it on. Only for suspending the PWM should be disabled because + * otherwise it refuses to suspend. The possible downside is that the + * LED might stay (or even go) on. + */ + priv->leds[i].state.enabled = !(cdev->flags & LED_SUSPENDED); ret = pwm_apply_might_sleep(priv->leds[i].pwm, &priv->leds[i].state); if (ret) -- 2.47.1
