Re: [PATCH v25 03/16] dt: bindings: lp50xx: Introduce the lp50xx family of RGB drivers

From: Dan Murphy
Date: Thu Jun 04 2020 - 12:28:15 EST


On 6/1/20 4:34 AM, Jacek Anaszewski wrote:
Hi Pavel and Dan,

On 5/31/20 9:06 PM, Pavel Machek wrote:

+          There can only be one instance of the ti,led-bank
+          property for each device node.  This is a required node is the LED
+          modules are to be backed.
I don't understand the second sentence. Pretty sure it is not valid

If I make these changes is this still viable for 5.8 or would you then go
into 5.9?

It really depends if we get -rc8 or not, and if you'll need to do any
changes to C code or not...

I think that we need to simmer such a big extension of the LED
subsystem for a whole cycle in linux-next, especially taking into
account addition of new sysfs interface, that is bit quirky.

Effectively 5.8 seems to not have been viable since few weeks.

After thinking about this for a while I would actually think to have this in 5.9.

Either 5.7 or 5.8 will be the 2020 LTS and such a new interface would be best suited for intermediate stable releases that get EOL'd faster.

This way we don't have to back port bug fixes for 2 years.