TDA4VM: TDA4 i2c5 access timed out

Hi,

I have been able to reproduce on another platform. I will get back to you on this in a couple of days after trying out the same on TDA4VM.

Best Regards,

Keerthy

Hi,

Is this a custom board? If it is TI EVM Can you please check if you can access wkup_i2c?

i2cdump -f -y 0x2 0x48

Let me know if the above works without timeout.

Also pinmux has couple of options:

Check if the other pins are working?

- Keerthy

wkup_i2c work  without timeout.But the mcu_i2c1 work with timeout also.

Hi Yang,

Is this a custom board or a TI EVM?
Can you check which pins correspond to I2C5?

- Keerthy

This is a custom board. Y26 and AA27 pin is work  without timeout.

The schematics  select T28 and T29 pins for i2c5,and Y26 and AA27 is floating.

Hi Yang,

So you still are seeing issues with using i2c5?

Regards,

Keerthy

Yes, the issues still exists.

Yang,

Can you share the i2c5 part of schematics snapshot? The pin mux seems correct.

Regards,

Keerthy 

Yang,

Thanks. Now can you check the actual pinmux register address and read out the values:

Use devmem2 tool from Linux:

devmem2 0x11c264 w
devmem2 0x11c260 w

I double checked from the sysconfig tool that your DT configs are correct:

&main_pmx0 {
myi2c1_pins_default: myi2c1_pins_default {
pinctrl-single,pins = <
J721E_IOPAD(0x260, PIN_INPUT_PULLUP, 4) /* (T28) MMC2_DAT3.I2C5_SCL */
J721E_IOPAD(0x264, PIN_INPUT_PULLUP, 4) /* (T29) MMC2_DAT2.I2C5_SDA */
>;
};

};

Best Regards,
Keerthy

The value is 0x00050007.

If I write it to 0x00050004, the i2c5 is work well.

The devicetree file k3-j721e-sk.dts has been modify like above, any other file need modify?

yang yulong said:

If I write it to 0x00050004, the i2c5 is work well.

This means that your dts has some conflicts. What is the name of the DTB that you use?

- Keerthy

k3-j721e-sk.dts

Hi Yang,

Please audit your dts files to check if the same is being used for some GPIO.

I believe this padconf is getting over written by a conflicting definition.

Best Regards,
Keerthy

Hi Keerthy，

I had the same problem testing the i2c1, and I was using the same hardware as Jade Dragon. 

And I also tested the EVM board i2c1, also have the same problem, please how to solve it?

Here is some information

dts I did not change, as follows

Best Regards,
Youke yang

youke yang said:

I had the same problem testing the i2c1, and I was using the same hardware as Jade Dragon. 

And I also tested the EVM board i2c1, also have the same problem, please how to solve it?

Here is some information

Yang,

I confirm that the pin mux is right for the main_i2c1:

&main_pmx0 {
myi2c1_pins_default: myi2c1_pins_default {
pinctrl-single,pins = <
J721E_IOPAD(0x228, PIN_INPUT_PULLUP, 0) /* (Y6) I2C1_SCL */
J721E_IOPAD(0x22c, PIN_INPUT_PULLUP, 0) /* (AA6) I2C1_SDA */
>;
};

Please share the output of the below command:

dmesg | grep i2c

We need to first find out which number the main_i2c1 corresponds to in Linux.

- Keerthy

Hi Keerthy,

The information you need is as follows

Youke

note: I modified the speed of i2c1 in the device tree to 100k

Yang,

You don't have the main_i2c1 enabled in your device tree. The 0x40b10000 instance is MCU_i2c1.

Please set

&main_i2c1 {
        status = "okay";
};

I believe status = "disabled";

So your index '1' is wrong and hence you see the timeout.

- Keerthy

Hi Keerthy,

I reworked the device tree as you requested, but I still don't see the 0x40b10000 device mounted

What is the reason?

Youke

youke yang said:

I still don't see the 0x40b10000 device mounted

This is not main_i2c1. This is mcu_i2c1.

Please check the device tree. Somewhere in the dts the main_i2c1 is disabled.

You should see 2010000.i2c in the dmesg log.

- Keerthy