Hi Dan, Thank you for the patch. Nonetheless, I've just applied similar Liviu's patch [0], since it arrived one week ago already. I'll send it upstream with LED fixes for 4-20-rc2. [0] https://lkml.org/lkml/2018/10/18/154 Best regards, Jacek Anaszewski On 10/25/2018 05:06 PM, Dan Sneddon wrote: > commit 45d4c6de4e49 ("leds: gpio: Try to lookup gpiod from device") > failes to set the gpio descriptor for DT based machines. Since the > value being passed in is always 0 the descriptor fails the validation > check but does so silently. This means the driver probes successfully > even though it can't actually control the gpio line. This patch sets > the gpio descriptor in the path taken by DT based devices. > > Fixes: 45d4c6de4e49 ("leds: gpio: Try to lookup gpiod from device") > Signed-off-by: Dan Sneddon <dan.sneddon@xxxxxxxxx> > --- > drivers/leds/leds-gpio.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/drivers/leds/leds-gpio.c b/drivers/leds/leds-gpio.c > index 32fa752..e0e57c2 100644 > --- a/drivers/leds/leds-gpio.c > +++ b/drivers/leds/leds-gpio.c > @@ -162,6 +162,7 @@ static struct gpio_leds_priv > *gpio_leds_create(struct platform_device *pdev) > fwnode_handle_put(child); > return ERR_CAST(led.gpiod); > } > + led_dat->gpiod = led.gpiod; > > fwnode_property_read_string(child, "linux,default-trigger", > &led.default_trigger); >
