Colibri imx7d som failures

Technical Support
,
1

We are using colibri imx7d som module in our projects and facing the following issue cases in SOM.
Total failures received from field during Jan2026-Jul2026 with below symptoms are 97no.s

Toradex_query.zip (1.1 MB)

  • Case 1: bootp error
    • bootp error messages are coming at power on and SOM is not booting.
    • We observed that this issue is coming even in SOMs with eMMC life is around 50%(data is taken when issue is not started).Previously same issue was communicated in below emails.It Previously concluded that the root cause is due to exceeding emmc life time
    • We are unable to load the OS in it.

Case 2)SOM is Hanging at “starting kernel”

    • SOM is Hanging at “starting kernel”
    • We observed that this issue is coming even in SOMs with eMMC life is around 50%
    • When OS is loaded it is recovering
  • Case 3) Freezing execution issue
    • When OS is loaded it is recovering.
    • During the Boot the SOM is hanging
  • Case 4) Librarires corruption
    • SOM is booting , But libraries in SOM are corrupting and application exes are not executing .We observed that the /usr/lib/libstc++.so.6.0.28 got coorupted . when this so file is replaced with the file from the toolchain into SOM , the applications were executing correctly.
    • We observed that this issue is coming even in SOMs with eMMC life is around 50%
    • Screenshots are attached
  • Case 5) inode errors
    • inode errors are observed in /media/mmcblk0p2 and application exes are not executing.
    • We observed that this issue is coming even in SOMs with eMMC life is around 50%
    • Logs and screenshots are attached
  • cat /etc/os-release
    • ID=tdx-xwayland
    • NAME=“TDX Wayland with XWayland”
    • VERSION=“5.4.0-devel-20210901132919+build.0 (dunfell)”
    • VERSION_ID=5.4.0-devel-20210901132919-build.0
    • PRETTY_NAME=“TDX Wayland with XWayland 5.4.0-devel-20210901132919+build.0 (dunfell)”
  • hostnamectl
    • Static hostname: colibri-imx7-emmc-14937551
    • Transient hostname: BU1-DTC-ETB
    •      Icon name: computer

    •     Machine ID: 7e57c99d573f46fe91b7e1f0c0920312

    •        Boot ID: 8758636d5d5e4fa6978c854eec14c8f2
    • Operating System: TDX Wayland with XWayland 5.4.0-devel-20210901132919+build.0 (dunfell)
    •         Kernel: Linux 5.4.129-5.4.0-devel+git.022cb949c6ec

    •   Architecture: arm
  • cat /etc/issue
  • TDX Wayland with XWayland 5.4.0-devel-20221202072024+build.0 (dunfell) \n \l
  • Colibri-iMX7-eMMC_Reference-Multimedia-Image
  • uname -r
  • 5.4.129-5.4.0-devel+git.022cb949c6ec

what is the maximum eMMC life , we can guarantee the application working?

Please look into these issues and provide the support as early as possible.

We can have a meeting to discuss about the issues.

Email : vidhyav@medha.com

2

Hi, I’m the Toradex AI Assistant.

I might make mistakes or occasionally get you wrong. Our support team will step in if needed.

Analysis of Colibri iMX7D SoM Field Failures

I understand you’re experiencing significant field failures with your Colibri iMX7D modules. Before I can provide specific guidance, I need to clarify some details about your setup:

Required Information

Could you please provide the following details:

  1. Hardware Version: Which exact version of the Colibri iMX7D are you using? (e.g., V1.1A, V1.1B)
  2. Product Number: What is the 8-digit product number (PID8) of your modules?
  3. OS Version: You mentioned multiple versions in your post - which exact BSP version are you running in production?
  4. eMMC Manufacturer: Which eMMC chip is installed on your modules?

You can gather this information using the tdx-info command on a working module. See: https://developer.toradex.com/software/development-resources/tdx-info/

Initial Observations

Based on the symptoms you’ve described and the available documentation:

eMMC Health Status

The eMMC standard (version 5.0 onwards) provides health status information that includes:(1)

  • Device life time estimation type A: Life time estimation for the MLC user partition eraseblocks (in 10% steps)
  • Device life time estimation type B: Life time estimation for the SLC boot partition eraseblocks (in 10% steps)
  • Pre EOL information: Overall status for reserved blocks

You can check eMMC health on working modules using:(1)

mmc extcsd read /dev/mmcblk0 | grep -A 1 LIFE
mmc extcsd read /dev/mmcblk0 | grep -A 1 EOL

Known Hardware Changes

According to the Product Change Notification, there was a significant eMMC change in the Colibri iMX7D 1GB between V1.1A and V1.1B:(2)

  • V1.1A: Used Micron MTFC4GACAJCN-1M WT eMMC
  • V1.1B: Changed to SkyHigh S40FC004C1B1C00 eMMC

The V1.1B also features larger boot partitions (4MB vs 2MB).(2)

File System Corruption Patterns

Your Cases 2-5 (kernel hang, freezing, library corruption, inode errors) all suggest file system corruption, which can occur when:(1)

  • The eMMC is approaching end-of-life
  • Power loss occurs during write operations
  • The file system wasn’t properly unmounted

Next Steps

  1. Provide the requested hardware/software details above
  2. Check eMMC health on modules showing 50% wear
  3. Verify your power supply is stable and meets specifications
  4. Review your shutdown procedures to ensure clean unmounting
  5. Consider upgrading to a current BSP version (you’re running a 2021 development build)

Once you provide the specific hardware version and eMMC details, I can give you more targeted guidance on expected lifetime and mitigation strategies.

Best regards,
Toradex AI Assistant.

Did you like my answer? :+1: :-1:

If I answered your question, please tick the “Solution” checkbox.

For follow-up questions to me, please use @ToradexAI in your post.