On 09/04/15 10:03, Sakari Ailus wrote: > Jacek Anaszewski wrote: >> On 04/08/2015 05:15 PM, Sylwester Nawrocki wrote: >>> On 08/04/15 16:31, Jacek Anaszewski wrote: >>>> Properties defining maximum values for LED currents and timeout should >>>> be mandatory to avoid the risk of hardware damage. This patch fixes >>>> the issue. >>>> --- a/Documentation/devicetree/bindings/leds/common.txt >>>> +++ b/Documentation/devicetree/bindings/leds/common.txt >>>> @@ -10,6 +10,17 @@ can influence the way of the LED device >>>> initialization, the LED components >>>> have to be tightly coupled with the LED device binding. They are >>>> represented >>>> by child nodes of the parent LED device binding. >>>> >>>> +Required properties for child nodes: >>> >>> These properties are mandatory only for LEDs with Flash/Torch >>> capabilities, aren't they? >> >> flash-max-microamp and flash-timeout-us properties description mention >> that they refer to flash LEDs. Perhaps this should be made indeed more >> explicit. > > Sounds good to me, after all these shouldn't be defined for devices they > don't make sense for. Alternatively, you could add another section for > "required properties for flash LEDs". Sounds good to me. >>> Requiring those properties for all led nodes would make all >>> current dtses not compliant with the DT binding specification AFAICT. > > Some drivers also define these explicitly as optional, the behaviour for > those obviously can't be changed. This should apply to new drivers > only, and I think documenting the matter in this file would be the best > way to do that. Unfortunately this file is not a proper place for information regarding Linux drivers. I can't see any drivers in -next using those DT properties. Could you point to any examples? Perhaps we should stay with the below properties specified as optional in general and add a note they can be mandatory for some LEDs (controllers)? >> I was also worrying about making led-max-microamp required, but others >> didn't share my objections. I think that we have to reexamine this. >> >> Please refer to the discussion under [PATCH v4]. The role of this >> led-max-microamp property would be preventing hardware damage. >> >> >>> How about: >>> >>> "Required properties for child nodes for LEDs with Flash/Torch >>> capabilities:" ? >>> >>>> +- led-max-microamp : Maximum LED supply current in microamperes >>>> + (torch LED for flash devices). Controllers that > > Do you think we can just rename max-microamp as led-max-microamp? It's > been there since v3.19. Or is it because no driver has been using it? I'd say it could be renamed if there is no users in mainline. But the DT maintainers could disagree here (adding Rob, Mark at Cc). >>>> have no >>>> + configurable current can omit this property. >>>> +- flash-max-microamp : Maximum flash LED supply current in >>>> microamperes. >>>> +- flash-timeout-us : Timeout in microseconds after which the flash >>>> + LED is turned off. >>>> + >>> >>>> +Above properties determine a LED driver IC settings required for safe >>>> +operation. They should be also used as the initial settings for the IC. >>> >>> Shouldn't "Controllers that have no configurable current can omit this >>> property" refer to both led-max-microamp and flash-max-microamp? >>> >>> I would drop the "Above...for the IC." paragraph and instead add >>> something like: >>> >>> "For controllers that have no configurable current the led-max-microamp, >>> flash-max-microamp properties respectively can be omitted. For >>> controllers that >>> have no configurable timeout the flash-timeout-us property can be >>> omitted." >> >> Are we going to avoid mentioning about their role in preventing >> hardware damage (former "for safe operation" sequence)? > > That'd make it understandable why the properties are mandatory. Sounds > good to me. -- Regards, Sylwester -- 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
