Activity
From 02/06/2026 to 03/07/2026
02/25/2026
- JC 05:05 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Great glad it worked. Hopefully the dtc command examples will be helpful for you and others as well
- BH 04:59 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- I md5sum'ed the SD dtb and the emmc dtb, they were in-fact different. I re-copied the files and now it is working. I guess in some step I mixed up which device I had updated the rootfs on. I think this is solved. Thank you for your help!
- JC 04:24 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- It would also be a good idea to md5sum the emmc and sdcard versions of the tfc dtb to make sure they are infact identical.
- JC 04:23 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Thanks Ben, I hadn't realized that.
Lets make sure the tfc device tree file contains what you expect it to.
Do you have the "dtc" tool on the filesystem, I think we include it by default?
You can use it to inspect the compiled d... - MW 03:35 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- can you also post the dtb file?
- BH 03:21 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Sorry, I just posted the u-boot log, it does indeed boot. Per the u-boot log it claims to load my dtb but the differences in the device tree once in linux are not present, here's just one example, if in SD boot I see my device tree refle...
- JC 02:43 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- > Some progress made, it now finds the file but still does not load it.
It looks like it did load it from the eMMC so that is indeed progress
02/24/2026
- BH 11:06 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Hi Jonathan,
Some progress made, it now finds the file but still does not load it.
Here are the dtbs (the "-tfc" is mine) - JC 10:47 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Yeah please share the console log during the boot attempt.
To get a bit more debug info, enable xtrace in u-boot before running boot - BH 10:38 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Hi Jonathan,
Thanks for your reply. Unfortunately, that did not solve the problem. The system fails to find the file and then proceeds to boot with the bootloader image fdt.
I have confirmed the dtb I want is located in the user par... - JC 09:26 PM Software Development: RE: AM62A eMMC Boot - Kernel Device Tree
- Hi Ben,
The device tree file that gets loaded is set in the fdtfile variable. You should be able to update it to match your device tree. - I’m working with the AM62A SoM and a Yocto development environment, switching between the CL devkit and custom hardware.
I am trying to boot the system with a modified device tree. This works fine from SD-cards but when I follow the ...
02/23/2026
- JC 04:01 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Glad it worked. Thanks for closing the loop
- PS 10:44 AM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Just a quick note to say thanks to everyone for solving this issue this quickly
We have fitted the pullups to our board and can now see the outputs toggling
Paul
02/20/2026
- JC 09:40 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- We will review and fix our datasheets to make mention of this as well.
- JC 09:39 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Hi Paul,
After reviewing and trying a few things, we noticed that these pins are labelled differently in the "TI datasheet pg 52":https://www.ti.com/lit/ds/symlink/am625.pdf?ts=1661784805550
MCU_GPIO_17/18/19/20 are classified as B... - JP 03:52 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Good morning Paul,
Thank you for the information. We were able to see the same behavior on our end. Tying 3.3 V to the MCU_GPIO0_17 through MCU_GPIO0_20 pins changed the state from low to high and back when configured as inputs. - PS 09:18 AM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Thanks for all the updates.
It sounds like you can at least reproduce the problem which is a start
Just to answer the other questions. I believe this is a standard AM62x, and for the time being we are using a development board so w...
02/19/2026
- JP 08:15 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Hello Paul,
We have replicated this issue on our end. I am using a 62x with our latest SDK (11.01) and found that MCU_GPIO0_17 through MCU_GPIO0_20 were not changing after a gpio set. We also noticed MCU_GPIO0_0 through MCU_GPIO0_4 we... - JC 07:36 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Also are you testing this on our devkit or on a custom design? If a custom design, would you be willing to email it to us for review?
- MW 07:19 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Hi Paul,
To confirm, you are using a standard AM62x SOM, correct (not an AM62Ax or AM62Px)?
Thanks,
Mike
- PS 04:19 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Thanks Mike
I think I have set up the GPIO lines correctly. I followed the steps described in the following wiki
https://support.criticallink.com/redmine/projects/mitysom_am62x/wiki/Example_-_User_level_GPIO_Access
I came up with gp... - MW 12:59 PM Software Development: RE: Problems writing to MCU_GPIO0_19 and MCU_GPIO0_20
- Hi Paul,
By default, the MCU_GPIO peripheral is controlled in the linux kernel for the devkit. (see the arch/arm64/boot/dts/k3-am62x-mitysom-devkit.dts file which enables it).
I am wondering if perhaps there is a conflict between ... - I have created an MCU application that uses the MCSPI to communicate with multiple SPI devices.
We need to talk to up to eight SPI devices so are using the MCSPI with a single chip select (MCU_SPI0_CS0), and then are using 3 GPIO lines ...