On Tue 2019-04-02 13:28:56, Alexander Dahl wrote: > Hello Pavel, > > Am Dienstag, 2. April 2019, 13:12:31 CEST schrieb Pavel Machek: > > On Tue 2019-04-02 12:23:29, Alexander Dahl wrote: > > > on my desk lies a piece of hardware with some dual color LEDs connected to > > > some IO ports. The LEDs shine green or red depending on polarity and are > > […] > > > Fortunately, that is not too common setup. You still need two output > > pins and two LEDs, AFAICT only advantage is that you save one > > resistor. Don't do it. > > Unfortunately, this hardware already exists, I have to deal with it, see > above. Hammer? :-) > > That said, driver that takes two GPIOs and exposes one LED with > > aditional color attribute should be acceptable. > > So, do I understand correctly, there's nothing to deal with such a setup in > Linux yet, and I need a new driver? Correct. > Documentation/devicetree/bindings/leds/leds-gpio.txt does not seem to document > such a usecase, so I guess the leds-gpio driver does not support it out of the > box? Correct. > What do you mean by "aditional color attribute"? Sorry, I'm not that deep into > the LED subsystem yet. Have single led, with static DEVICE_ATTR(color, S_IRUGO | S_IWUSR, ... ? > Of course I could just handle those as two independent LEDs for now, but that > would lead to not desired states, because both colors are not independent! The > only thing I could try would be let userspace somehow control the brightness > and triggers to avoid both colors cancelling out each other? No, don't do that. Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Attachment:
signature.asc
Description: Digital signature