On Wed, Jul 29, 2015 at 6:51 PM, Bjorn Andersson
[off-list ref] wrote:

On Fri 24 Jul 08:29 PDT 2015, Rob Herring wrote:

quoted

On Thu, Jul 23, 2015 at 2:52 PM, Bjorn Andersson
[off-list ref] wrote:

quoted

Add the possibility of specifying the default brightness in DT.

Signed-off-by: Bjorn Andersson <redacted>
---

This depends on the patch moving pm8941-wled to backlight [1]. The dt property
is used by several other backlight drivers, so I considered this to be a
"common" property and it's hence not prefixed with "qcom,".

Well, we have "default-brightness" and "default-brightness-level" used
by 1 driver each. But default-brightness-level is much more commonly
used (in dts files) since it is in the pwm backlight binding, so we
should go with it. I'd like to see this moved to a common backlight
doc.

As I looked at these, the default-brightness used in tps65217 is a value
between 0 and 100, so that can be interpreted as a percentage.

The pwm binding however uses a separate array of "brightness-levels" and
then default-brightness-level is supposed to be an index into that
array.

Uggg. I missed that minor detail...

As we're trying to specify a default brightness within the range [0,
max_brightness) the latter doesn't make much sense.

Therefor my suggestion is that we make the "default-brightness" the
common property and we define it as a percentage of [0,max_brightness).

Okay.

I wonder if we should have units such as
"default-brightness-percentage" or "default-brightness-%" so it is
clear. Otherwise, we might have some people doing a range of [0,max].
The former is a bit long and the latter is a bit unusual.

quoted

Really, I think all the backlight documentation should be merged with
LEDs docs. Things like "default-on" are common. But I won't ask to do
that here.

I think the backlight framework should be merged with the LED framework.
There's several hw blocks that are split between the two, with an mfd
tying them together...

Fully agree. BTW, doing that doesn't have to be in sync between the
bindings and drivers. Of course, if we've designed the bindings with
sub devices to fit the MFD structure, then that is another problem.

Rob