/pwmleds/pwm1: arguments longer than property

Technical Support
, , , , ,
1

We work with a Demo Image and added our own layers.
In the layer “meta-cobra-apalis-imx6” there is the recipe “linux-toradex” where we have the file “imx6q-apalis-cobra.dts”.
Appended you find from bitbake the “console-latest.log”.
Then we can install with “Toradex Easy Installer”.
When booting the installed image we see that the DeviceTree could not be correct loaded. The whole booting output is attached.

console-latest.log (29.3 KB)
DeviceTreeError.txt (25.4 KB)
MyYoctoLayers.tar.xz (10.5 KB)

There you can see:

[    2.172180] OF: /pwmleds/pwm1: arguments longer than property
[    2.178023] of_pwm_get(): can't parse "pwms" property
[    2.183122] leds_pwm pwmleds: unable to request PWM for cobra::pwm1: -22
[    2.189883] leds_pwm: probe of pwmleds failed with error -22

What is wrong with my layer

Regards
ah4711

2

Hi @ah4711 ,

thanks for using our community.

We are looking into your question.

Best Regards
Kevin

3

Hi @ah4711 ,

I noticed a conflict in assigning the pins to the corresponding functions.

[    5.829428] imx6q-pinctrl 20e0000.iomuxc: pin MX6Q_PAD_NANDF_D4 already requested by 20e0000.iomuxc; cannot claim for spi0.0
[    5.840838] imx6q-pinctrl 20e0000.iomuxc: pin-195 (spi0.0) status -22
[    5.847385] imx6q-pinctrl 20e0000.iomuxc: could not request pin 195 (MX6Q_PAD_NANDF_D4) from group gpio2io04grp  on device 20e0000.iomuxc
[    5.859811] mcp251x spi0.0: Error applying setting, reverse things back
[    5.866597] mcp251x: probe of spi0.0 failed with error -22

Are you planning to use the SPI interface in your application? If not, then you might consider disabling the Interface to prevent the conflict.

Best Regards
Kevin

4

Hi Kevin

Yes that is true with this conflict, but that is not really the problem. It is as above mentioned the “/pwmleds/pwm1: arguments longer than property”.
Is it possible, that the used linux-kernel is not compatible with this configuration in the DeviceTree?

Regards
ah4711

5

Hi @ah4711 ,

could you please provide the kernel configuration you have been using?

BR
Kevin

6

What exactly do you need?
In the attached MyYoctoLayers you see our changes on top of your toradex-linux-kernel layers.

7

Hi @ah4711 ,

could you do the changes to the kernel branch outside the openembedded build, and check if you still see the issue? If the issue is present please provide a git diff to the kernel branch.

Best Regards
Kevin

8

Hi Kevin

All changes are already outside the openembedded build. For that reason we made our own layers (see the attached MyYoctoLayers.tar.gz). The original layers are untouched.

I think that’s the way to go.

Regards
ah4711

9

Hi @ah4711

It looks like that there is a typo in your devicetree. According to Documentation (leds-pwm.txt), there should be three entries and not 4 as you wrote

LED sub-node properties:

  • pwms : PWM property to point to the PWM device (phandle)/port (id) and to
    specify the period time to be used: <&phandle id period_ns>;

Your devicetree:
led1 {
label = “PWM1”;
pwms = <&pwm1 1 50000 0>;
max-brightness = <255>;
};

It should be:

led1 {
label = “PWM1”;
pwms = <&pwm1 1 500000>;
max-brightness = <255>;
};

Best regards,
Jaski

10

Hi Jaski

Thank you for your hint.
But I tried that already and it did not function.

Regards
ah4711

11

Hi @ah4711

I tested with these changes in devicetree and it is working.

diff --git a/arch/arm/boot/dts/imx6q-apalis-eval.dts b/arch/arm/boot/dts/imx6q-apalis-eval.dts
index 3ac59ca1d061..dcffcfa7a3ca 100644
--- a/arch/arm/boot/dts/imx6q-apalis-eval.dts
+++ b/arch/arm/boot/dts/imx6q-apalis-eval.dts
@@ -64,6 +64,15 @@
                regulator-always-on;
        };
 
+       pwmleds {
+               compatible = "pwm-leds";
+               led1 {
+                       label = "PWM1";
+                       pwms = <&pwm1 0 5000000 0>;
+                       max-brightness = <255>;
+               };
+       };
+
 };
 
 &can1 {
12

Did you build it with Poky zeus or dunfell?
Because of QT we are stuck on zeus,

13

I tested this with Dunfell, I can also check with Bsp 3.0.4.
Best regards,
Jaski