Hello On 1/17/19 4:06 AM, Pavel Machek wrote: > Hi! > >>>>> You know what? First, submit driver with similar functionality to >>>>> existing RGB drivers, using same interface existing drivers are >>>>> using. When that is accepted, we can talk about extending >>>>> kernel<->user interfaces. >>>>> >>>> >>>> I could do that but then there is no way for users to have any other color but "white" with this driver. >>>> That defeats the purpose of the device itself. >>> >>> No, that is not what I meant. >>> >>> We do have RGB drivers in tree, they just present three separate LEDs >>> -- red, green and blue. I ask you to do the same for initial >>> submission. >>> >> >> For clarification you are asking me to register a LED class per >> output pin? > > Yes. > OK I can do that but I would like to get Jacek to align with this as he suggested creating the code as I submitted based on the review of the devices data sheet. LED class registration per LED module or for the single Banked output. >> As explained before the Master brightness register has absolute control over the output current >> to the LED outputs regardless of the color control setting. >> > > Yes, just set master brightness to 255 and you can ignore it. OK Sounds fine. Will just need to do an init that sets the color register to 0 and brightness to 0 so we don't have a flash. > >>> You'll still be able to set brightness independently on the >>> red/green/blue LEDs... >>> >> >> But which one would control the overall brightness of the cluster? > > You would not have that control in initial version of the patch. > See above on alignment Dan > Pavel > -- ------------------ Dan Murphy
