Re: [PATCH v2 2/2] leds: lp50xx: Add the LP50XX family of the RGB LED driver

From: Jacek Anaszewski
Date: Tue Jan 29 2019 - 15:19:35 EST


Hi Dan,

On 1/29/19 2:56 PM, Dan Murphy wrote:
Jacek

On 1/24/19 3:55 PM, Jacek Anaszewski wrote:
Dan

On 1/24/19 10:00 PM, Dan Murphy wrote:
Jacek

On 1/23/19 3:52 PM, Jacek Anaszewski wrote:
Dan,

On 1/22/19 11:44 PM, Dan Murphy wrote:
Jacek

On 1/22/19 3:39 PM, Jacek Anaszewski wrote:
Hi all,

On 1/20/19 7:42 AM, Vesa JÃÃskelÃinen wrote:
Hi Dan,

On 18/01/2019 15.58, Dan Murphy wrote:
Jacek

On 1/18/19 7:45 AM, Dan Murphy wrote:
Jacek

On 1/17/19 3:10 PM, Jacek Anaszewski wrote:
Hi Dan,

On 1/16/19 7:41 PM, Dan Murphy wrote:
Hello

On 1/16/19 4:55 AM, Pavel Machek wrote:
Hi!

On 1/15/19 4:22 PM, Pavel Machek wrote:
Hi!

+The 24-bit RGB value passed in follows the pattern 0xXXRRGGBB
+XX - Do not care ignored by the driver
+RR - is the 8 bit Red LED value
+GG - is the 8 bit Green LED value
+BB - is the 8 bit Blue LED value
+
+Example:
+LED module output 4 of the LP5024 will be a yellow color:
+echo 0xe6de00 > /sys/class/leds/lp5024\:led4_mod/color
+
+LED module output 4 of the LP5024 will be dimmed 50%:
+echo 0x80 > /sys/class/leds/lp5024\:led4_mod/brightness
+
+LED banked RGBs of the LP5036 will be a white color:
+echo 0xffffff > /sys/class/leds/lp5036\:led_banked/color

This part with example cans remain in Documentation/leds if you
like.

Does it actually work like that on hardware?

What?

If you do echo 0xffffff > /sys/class/leds/lp5036\:led_banked/color,
does it actually produce white? With all the different RGB modules
manufacturers can use with lp5024P?

If you do echo 0xe6de00 > /sys/class/leds/lp5024\:led4_mod/color, does
it actually produce yellow, with all the different RGB modules
manufacturers can use with lp5024P?


I believe the answer to the general questions is no for any RGB cluster and driver out there.
Because if you set the same values on each and every RGB device out there you will get varying shades of the color.
But for this device yes the color does appear to be yellow to me versus what was displayed on my monitor by the HSL picker.
But everyone interprets colors differently.

If you write the same value for yellow or white on a droid 4 and the N900 do they produce the same color side by side?
Most probably not.

As you pointed out the PWM needs to be modified to obtain the correct white color to account for LED and other device constraints.

But we need to take into account the light pipe. Pools nowadays have RGB LED spot lights in them. It can
be set to white. On my pool right off the lens the color has a purplish hue to it. As the light is diffracted into
the pool the color becomes white. The pool is clear. When I add chemicals to the pool and make it cloudy
and turn on the lights the color off the lens is now white. This is an example on a large scale but the issue
scales down to the hand helds and smart home applications.

If the cluster is piped through a flexible optic 0xffffff may produce the "white" you want on its output.

So an expectation of certain color without proper piping based on a single RGB value may be a little unreasonable.
There may need to be a way to attenuate the values based on the hardware aspect of the equation ie light pipe (or lack thereof) and LED vendor.
So if we write 0xffffff to the RGB driver the driver could adjust the intensity of the individual LEDs based on the diffraction
coefficients.

I also think that is an unreasonable expectation here that writing a single value to any LED RGB driver would produce
a "rest of the world" absolute color. Maybe it can produce something similar but not identical.
As you indicated in the requirements there is more involved here then just the LED and the values written.
The colors should be close but may not be identical.

A 10 year old N900 should not be considered the gold standard for color production due to advancements in LED,
light pipe and LED driver technology.
The single package RGB clusters on the board I am testing is about the size of a single RGB LED from 10 years ago.

I agree that the interface developed should work on the device but the algorithm derived to obtain the color needs to have
a hardware aspect to the calculation.

Is it supposed to support "normal" RGB colors as seen on monitors?

Monitors are not an application for this part.

You did not answer the question. When you talk about yellow, is it
same yellow the rest of world talks about?


See above. It is close to what was on my monitor displayed.

Because 100% PWM on all channels does not result in white on hardware
I have.

I don't know I am usually blinded by the light and have no diffuser over
the LEDs to disperse the light so when I look I see all 3 colors.

How can we have useful discussion about colors when you don't see the
colors?

Place a piece of paper over the LEDs....


Good suggestion for a rough test.

But...

I believe we should have a reasonable design before we do something
like this. There's no guarantee someone will not use lp50xx with just
the white LEDs for example. How will this work? Plus existing hardware
already uses three separate LEDs for RGB LED. Why not provide same
interface?

Which existing hardware? Are they using this part?

Nokia N900. They are not using this part, but any interface we invent
should work there, too.


Yes a common interface would be nice with some sort of hardware tuning coefficient.

<rant>
Why are we delaying getting the RGB framework or HSV in?
I would rather design against something you want instead of having
everyone complain about every implementation I post.
</rant>

Because you insist on creating new kernel interfaces, when existing
interfaces work, and are doing that badly.

Because your patches are of lower quality than is acceptable for linux
kernel.

Because you don't seem to be reading the emails.

I sent list of requirements for RGB led support. This does not meet
them.


Sigh. You did not answer my question.

Your requirements seem to be centered around monitors but that is only one application of the current
RGB LED landscape.

I am willing to work with you on the HSV and adapting the LP50xx part to this framework.
Or any RGB framework for that matter. I still don't agree with the kernel needing to declare colors
ÂÂÂÂ maybe color capabilities but not specific colors.

Dan, if you have a bandwidth for LED RGB class implementation
then please go ahead. It would be good to compare colors produced
by software HSV->RGB algorithm to what can be achieved with
LEDn_BRIGHTNESS feature.

The requirements for LED RGB class as I would see it:

sysfs interface:

brightness-model: space separated list of available options:
- rgb (default):
ÂÂÂÂ - creates color file with "red green blue" decimal values

What about other colored LEDs? Presenting RGB for an Amber LED does not seem right.
Should the LED color come from the DT?


I thought about this, other non-RGB LEDs would not use the RGB framework.
But should they have the same interfaces as RGB?

Should PWM control be a global interface?

In order to being able to set multi color element led at one go I would recommend using then model:

color_names: "red green blue white"

echo "32 43 0 128" > color

This way all elements would be set at same time from user space point of view.

This of course requires that it is part of the physical/logical led that is being controlled. If it is a separate thing then it would logically be differently controlled mono color led.

If you look what kinds of leds are available lets say from digikey you get all kinds of combos:

- red, green, blue
- red, green, blue, amber
- red, green, blue, white
- red, green, blue, white - cool
- red, green, blue, white - neutral
- red, green, blue, white - warm
- red, orange
- purple, ultraviolet
- amber, blue
- amber, blue, cyan, green, red, violet, white - cool
- amber, blue, green
- amber, green, blue
- and then lots of single special colors

It suggested me another solution. Instead of LED RGB class
we would have LED multi-color class.


I was thinking the same thing this morning. But I was thinking that the RGB
class should be an additional class to stand on its own or can register to the
multi color class.

Sysfs interface design:
-----------------------

colors: directory containing files that map to
ÂÂÂÂÂÂÂÂÂ the brightness of particular LEDs; there
ÂÂÂÂÂÂÂÂÂ would be predefined color names that LED class
ÂÂÂÂÂÂÂÂÂ driver should map iouts to, e.g.:

Filling in the missing ideas with questions.

Is it a directory or a file? If it is a directory does that not break the current
directory label model?

so the path would be /sys/class/leds/colors ? (It is probably not this but needs clarification)
How would this look if I had 2 of the same color LEDs? The Beagle bone black has 2 Blue LEDs.
They are assigned to different triggers and have different directories. Both are GPIO controlled.

Or are you saying it would be something like (More then likely this is what you intended)
/sys/class/leds/input4::numlock/colors?

Yes, this is what I meant.


OK. Thanks for the clarification

Maybe it is mandated that "multi" be added to the label when the class is registered so the caller
knows that this is a multi color class and not a single LED color class.

Like I am going to come up with standardized color names
in my LED naming related patch, the multi-color names
can be defined as well, e.g.: rgb, rgbw, rgbwa, rgbwauv etc.


That may be better it is descriptive off the command line.

What about providing a file called colors_raw which takes in the raw decimal values to obtain other color
variants when RGB is only available? And this can also present a single write to the kernel with the new
color values.

My design already covers that in the form of files in the colors
directory. Like I stated: "files that map to the brightness of
particular LEDs". Single write is secured by "echo "write" > sync.


OK. So set the new values and then tell the set the sync which will make the device driver write the
device.

Sounds good. how about echo 1 > sync and we can stay away from a long string conversion

We need also to be able to do a readout.

If we changed "sync" to "rw" then we could come up with intuitive
semantics:

echo 1 > rw // write
echo 0 > rw // read

rw file would have WO permission.

I have re-written the lp50xx driver to show red, green, blue, sync and sync_enable files for each registered LED.

Did you change your mind regarding LED multi color class implementation?

I added the sync_enable file so that the there can be real time control of the individual LEDs as well.

Yes, it's a good idea.

Turning on sync means that the user will write to one of the color files and the setting won't take affect until
sync is written.

If sync is off then the LED register is updated immediately.

Being able to turn on and off syncing maybe better. A developer may choose to set up the color and then sync
or they may want to ramp a specific color. This will help reduce user space writes but also reduce the number
of peripheral writes when color values do not change.

I am having difficulty in creating the colors directory in the device driver.
This appears to need to be done in the class when the class pointer is available.

Not necessarily, you need only dev->kobj. Please follow the implementation of drivers/usb/core/ledtrig-usbport.c. However I would
really prefer if it became a part of new LED multi color class.

I am not a fan of hard coding preset colors as you can see there are to many of them and variations of the color.
In addition this severely limits the ability of the user. Unless we stick to primary colors only and not secondary
colors.
Defining and hard coding hte colors can get out of control and not maintainable moving forward. Especially
if we start adding defines like white_warm, white_neutral and other variations to the color.

We would not limit anything. Every color combination can be achieved
by following this exemplary sequence:

$ echo 154 > colors/red
$ echo 232 > colors/green
$ echo 43Â > colors/blue
# echo "write" > syncÂÂÂÂ //only this changes hardware state


The code works more like this
$ :/sys/class/leds/lp5024:led1_mod# ls
blue device max_brightness red sync trigger
brightness green power subsystem sync_enable uevent

$ cat sync_enable
enabled // I can change this to return an int instead
$ echo 154 > red
$ echo 232 > green
$ echo 43 > blue
$ echo 1 > sync //this changes hardware state

$ echo 0 > sync_enable
$ cat sync_enable
disabled
$ echo 154 > red // changes red LED immediately
$ echo 232 > green // changes green LED immediately
$ echo 43 > blue // changes blue LED immediately
$ echo 1 > sync // Has no affect on the hardware

If this is OK I can post the patch with this update.

But I would rather put the file creation in a class and have the colors directory.

Yes, please post your work when it is ready.

brightness-model is provided only to define mapping of legacy brightness
levels (governed by brightness file and led_set_brightness() API) to
the specific combination of colors.

For instance we can define three brightness levels for green hue:

DT definition for it would look like below:

rgb-green = <0x277c33 0x37b048 0x47e45d>;

LED multi color class would then do the following mapping for
each of the three brightness levels for rgb-green brightness model:

$ echo rgb-green > brightness_model
$ echo 1 > brightness // red=0x27, green=0x7c, blue=0x33
$ echo 2 > brightness // red=0x37, green=0xb0, blue=0x48
$ echo 3 > brightness // red=0x47, green=0xe4, blue=0x5d


OK I would have to play with this on the LP devices.

I have not done anything with this.


What about IR LEDs used for night vision products? Do these fall into the multi color class?
We do have a driver I submitted that had an IR LED and a White LED combined. It was created against the
flash class but it could be a multi color LED as well.

How would traversing through the full color palette work if you were to want to produce a multi
color ring like the LP50xx where the pattern can traverse from one end of the color spectrum and back?
Or in a product like the gaming keyboards that will change color or change backlight brightness?

This is not meant as a solution for pattern generator but for
consolidated source of multi color light. In the simplest case
RGB LED elements like those used for lp5024 reference board,
but it could be RGBWAUV LED [0] as well.

For patterns traversing many LEDs I see a trigger as the best solution.
Hmm, now I see that trigger mechanism actually can serve as very
flexible pattern generator.

We would need a device that could be configured to register
a number of multi-led-patternN triggers, one per LED, and generate
events for each trigger in a loop.

The device would have to allow for configuring pattern intervals
via sysfs, like in case of current pattern trigger.

LED class devices would have to register for its events:

$/sys/class/leds/led1 echo multi-led-pattern1 > trigger
$/sys/class/leds/led2 echo multi-led-pattern2 > trigger
$/sys/class/leds/led3 echo multi-led-pattern3 > trigger


A bit off topic but I like the idea. We should save this for another day

Yes, that's another story.

The ability to define brightness models in DT would
add even more flexibility.


brightness models would be mandatory to support in the driver but an optional
DT entry.

Is that a correct assumption?

Correct.

Not sure what color LEDs the keyboard manufacturers place on their keyboards but does the interface design
capable of doing this?

https://www.youtube.com/watch?v=pfKv3g2FeBE
or something like this

https://www.youtube.com/watch?v=PO3auX3f5C4

The LP5036 has this capability.

ÂÂÂ - red
ÂÂÂ - green
ÂÂÂ - blue
ÂÂÂ - white
ÂÂÂ - sync: accepts "write" and "read", which executes
ÂÂÂÂÂÂÂÂÂÂÂ write/readout to/from a device respectively.


What are these above, the values or the files under the colors directory?

They are just color specific counterparts of monochrome brightness.
In terms of lp50xx they would map to OUTn_COLOR registers.

I am assuming they are files.

Right.

Are they mandatory or optional?

Mandatory, one per each iout to control.


OK And all the LEDs within this directory would be considered a LED cluster?

And if there are 2 like colors of the LED defined in the same cluster would we just see a single
file and write a common value to that file and the driver would have to update each
red LED within that cluster. No independent control of like colored LEDs within the
registered LED cluster.

If the developer wants this level of control they would have to register two separate classes

Correct?

I would abide by one color to one iout mapping. Registering more LEDs
under the same color feels more of a task for trigger.

brightness-model: defines brightness level to color configuration
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ mapping
ÂÂÂ - "hardware": for devices with feature like LEDn_BRIGHTNESS of lp50xx
ÂÂÂ - "rgb-<hue>": available only when all three red,green,blue colors
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ are present in the colors directory.
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ <hue> is a placeholder for given DT hue presets.
ÂÂÂ - "rgbw-<hue>": Same as above, but available only when white color
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ (any of amber or white-{cool|neutral|warm} can be
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ mapped to white) is also available. In this mode
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ max_brightness equals num-of-hue-presets + 1, and
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ max_brightness, when set, turns the "white" LED on

Why do we need white combined here? Should this not be its own entity?

To be able to set white color. We're still talking about one LED
element (although they can be be physically few LEDs in one case).
This is brightness file, so we've got to stick to the semantics.
Max brightness level should be the brightest. With RGBW LEDs we
fortunately have a means to achieve pure white, that's why
rgbw-<hue> would be beneficial. If you increase L component of
HSL color space, the max value gives white for all hues.
So maybe this brightness-model would be rather called hsl-<hue>.

For RGBW LEDs, we would have to allow for more shades of white too,
like in [1].


Yep.

But this all would be left to the DT designer decision.

Again I don't like limiting the color palette from the DT. I think that the
user space can see what colors are available for that device and makes its own
decision on what color to present.

For the RGBw what about RGB amber and RGB purple. Are the white LEDs always part of the
same function trying to be achieved by the system designer? The RGB can be used to denote
notification status and the white can be used to denote that a charger is connected. Motorola
Droid did this.

I hope I've just clarified my idea.


Its getting clearer. I would like to see it in code and play with it not as a user
but as a developer. Make sure the paper model works as well as the real implementation.

Is this model clear to the developer?
How would a developer define what values are appropriate for the brightness-model?

We could create guidelines e.g. that for hsl-<hue> pattern, the
colors corresponding to brightness levels should be arranged so
that increasing brightness felt like increasing value of
L component of HSL.

But we wouldn't be able to enforce adherence to a particular scheme.

Does the driver have to become overly complex to support simple color generation?

Not at all. Caching the colors written to the files in colors directory
would be responsibility of LED multi-color class. On echo 1 > sync
the core would call a new op e.g.

color_set(struct led_classdev_multicolor *mcled_cdev)

Driver would read in it colors cached in struct led_classdev_multicolor
and write them to the hardware.

Thoughts on putting code to idea?



ÂÂÂ - "rgb-linear": I'm not sure if it should be available - it will
ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ have unpredictable results

brightness: sets/reads brightness in the way specific to the
ÂÂÂÂÂÂÂÂÂÂÂÂÂ current brightness-model. When more colors are available
ÂÂÂÂÂÂÂÂÂÂÂÂÂ (e.g. amber, blue, cyan, green, red, violet, white), they
ÂÂÂÂÂÂÂÂÂÂÂÂÂ are not touched by write to brightness).

HSV->RGB conversion is left entirely to the userspace, which can set any
color with use of the proposed interface.

Let's agree on sysfs interface first, and only after that move
to the DT details.


DT's are meant to describe hardware and not describe the product. Unless Rob does not see
an issue with defining product capabilities in the DT then we should keep that out of the DT.

LED element is a device. I see nothing irrelevant for DT in describing the lighting specificity of a device mounted on the board. Please keep
in mind that it will not limit the number of colors available to set.
It will only allow to define mapping of brightness level to color.
We need that for current trigger mechanism to work with LED multi-color
class.


I see this now.

Dan

[0] http://www.cobledarray.com/sale-4942931-10w-rgbwauv-led-diode-6-in-1-high-power-multicolor-led-chip.html
[1] https://www.youtube.com/watch?v=NzlFmTqOh9M







--
Best regards,
Jacek Anaszewski