SK-TDA4VM: SK-TDA4VM Kit Freeze Issue During Long-Term Testing

Hi Naveen,

Could you please share your compilation files?

Thanks,

Christina

Hi Christina Kuruvilla,

Can you please clarify what you mean by "compilation files"?

For the SD card base flash, I have used this tisdk-edgeai-image-j721e-evm.wic.xz image from this processor page . For installing the kernel source to bring up my camera driver, I have used this ti-processor-sdk-linux-edgeai-j721e-evm-10_01_00_04-Linux-x86-Install.bin.  

Kindly let me know if you need any additional details

Thanks,
Naveen

Hi Naveen,

Apologies for the unclear question. I was asking for the compilation logs, which is essential just the copy paste of what you run on the command line before you receive this hang/freeze.

Thanks,

Christina

Hi Naveen,

The device should not be freezing like this, especially after running for such a short time. Are you finding that the device is overheating as well? Could you try incorporating a fan in your set up to determine whether the freeze is occurring due to high temperatures?

In the meantime, please share the full PROC # located at the bottom of the SK board as well as any additional information about this specific board such as: "Is it new?", "Have you seen similar issues in the past?", "Has this freezing occurred when using older SDKs?", etc.

Thank you,

Fabiana

Hi Christina Kuruvilla,

I believe you are asking for the dmesg logs. In this regard, I am not seeing any specific logs related to the kit freeze/hang. Below are the last few lines of the logs I can retrieve from my UART debug output:

[ 23.957143] i2c 6-0042: deferred probe pending
[ 23.961594] i2c 7-0042: deferred probe pending
[ 125.905910] audit: type=1006 audit(1745561834.740:16): pid=1319 uid=0 subj=kernel old-auid=4294967295 auid=0 tty=(none) old-ses=4294967295 ses=3 r1
[ 125.919504] audit: type=1300 audit(1745561834.740:16): arch=c00000b7 syscall=64 success=yes exit=1 a0=8 a1=ffffd3fbf908 a2=1 a3=1 items=0 ppid=1 p)
[ 125.947327] audit: type=1327 audit(1745561834.740:16): proctitle="(systemd)"
[ 125.954503] audit: type=1334 audit(1745561834.756:17): prog-id=18 op=LOAD
[ 125.961346] audit: type=1300 audit(1745561834.756:17): arch=c00000b7 syscall=280 success=yes exit=8 a0=5 a1=ffffc827a058 a2=90 a3=0 items=0 ppid=1)
[ 125.988188] audit: type=1327 audit(1745561834.756:17): proctitle="(systemd)"
[ 125.995344] audit: type=1334 audit(1745561834.784:18): prog-id=18 op=UNLOAD
[ 126.002344] audit: type=1300 audit(1745561834.784:18): arch=c00000b7 syscall=57 success=yes exit=0 a0=8 a1=1 a2=0 a3=ffff88379c60 items=0 ppid=1 p)
[ 126.029052] audit: type=1327 audit(1745561834.784:18): proctitle="(systemd)"
[ 126.036107] audit: type=1334 audit(1745561834.784:19): prog-id=19 op=LOAD


Thanks,
Naveen.

Hi Fabiana Jaimes,

Fabiana Jaimes said:

Are you finding that the device is overheating as well? Could you try incorporating a fan in your set up to determine whether the freeze is occurring due to high temperatures?

Yes, the device is overheating, and I haven’t incorporated a fan yet. I only have the heat sink that came with the kit. Is there any way to monitor the temperature of the kit and produce output in the dmesg logs? This would help confirm whether the issue is due to overheating when the device freezes or hangs.

Fabiana Jaimes said:

In the meantime, please share the full PROC # located at the bottom of the SK board as well as any additional information about this specific board such as: "Is it new?", "Have you seen similar issues in the past?", "Has this freezing occurred when using older SDKs?", etc.

PROC - PROC112B (001). No, the kit is not new; we purchased it almost two years ago. I haven’t done long-run testing with older SDKs, so I haven't encountered this issue in the past.

Please help us track down the issue. If overheating is the cause, kindly assist us in confirming that, as I am not seeing any related logs in the dmesg output when the device freezes or hangs.

Thanks,
Naveen.

Hi Naveen,

Naveen K said:

Is there any way to monitor the temperature of the kit and produce output in the dmesg logs? This would help confirm whether the issue is due to overheating when the device freezes or hangs.

To see how SOC temperature statistics can be captured and recorded, you can take a look at the ti_perfstats.cpp located in the /opt/edgeai-gst-apps/scripts/perf_stats directory.

I have looped in one of our hardware engineers in this thread for additional support on this issue.

Thank you,

Fabiana

Hi Fabiana Jaimes,

Thank you for your input. I have been monitoring the SOC's temperature logs via /opt/edgeai-gst-apps/scripts/perf_stats and found that the temperature was reaching high levels, nearing 110°C, even with the heat sink in place. As per your suggestion, I added a fan, which has helped to keep the temperature under control, maintaining it below 50°C. This allows me to run the kit overnight without any issues.

Thanks,

Hi Fabiana Jaimes,

Fabiana Jaimes said:

Is this solution sufficient for your use case? Please keep in mind that the SK devices are intended to be used for prototyping. Although the heat sink provides thermal dissipation, the SOC temperature is still expected to continue to rise when running for an extended period of time without the use of a fan. 

Yes, it is sufficient for my use case. 

Thanks,
Naveen.