Jacek On 05/16/2018 04:17 PM, Dan Murphy wrote: <snip> >>>> >>>> >>>>>>> + if (!ret) >>>>>> >>>>>> if (ret) sounds more natural. And better just to split >>>>>> >>>>>>> + snprintf(led->led_name, sizeof(led->led_name), >>>>>>> + "%s:%s", led->led_node->name, name); >>>>>>> + else >>>>>>> + snprintf(led->led_name, sizeof(led->led_name), >>>>>>> + "%s:torch", led->led_node->name); >>>>>> >>>>>> const char *tmp; >>>>>> >>>>>> ret = device_property_read_...(&tmp); >>>>>> if (ret) >>>>>> tmp = ... >>>>>> sprintf(...); >>> >>> We're no longer taking devicename section of a LED class device name >>> from DT, so it will look differently anyway. >>> So in adding the device_property code I think we again are reaching the LED label issue. In ARM with DT we would take the parent device node name and append it to the label if the optional label property was not available. In migrating to the device_property APIs we don't or can't depend on that parent node anymore. So for the case where the label property does not exist should we use a hard coded name or should we try to use the name from a device_id table. This is how we did this for the leds-lp8860 driver. If the label did not exist we used the i2c_device_id table and pulled the string from there. Thoughts? <snip> -- ------------------ Dan Murphy -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
