We are working in a project using the Jetson nano for about three month ago and now we are having some issues as detailed in the images attached.
In these images you can see the log, no one program was running at the moment happened this trouble. We don’t know what could it be or what to do, so we hope you can help us.
Thanks a lot.
Here is the another image.
Looks like some process are doing the i2c bus scanned like i2cdetect.
Could you try reflash the OS to check if still have the same problem?
Thanks for your reply, we did this and also changed the frequency to 100Khz and the problem still continue, what else could you suggest us to do?
Hi ShaneCCC. Thanks for the information. In addition to the Ricardo’s info, it is important to take in account that this issue is happening at several Jetson nano we have here, and our processes aren’t running the i2cdetect at the moment this error is thrown.
The kernell is getting the “no acknowledge from address” error more than 50 times per seconds, and some times the nvidia jetson get freeze for 500 millis average or more, and that was the reason we checked the syslog and found those errors.
We thank you to brind us any advice to resolve this problem as soon as possible.
Does the kernel Image and dtb are customize or the default?
Dear ShaneCCC, the Kernel image is the default image and the same for the DTB files, we didn’t customize them.
We installed the ubuntu 18.04 as the specs recommend and then we install the XFCE Desktop environment because it works lighter than the gnome and unity environment.
Are you able to flash all of the default image include the OS to verify?
Dear ShaneCCC, We have installed 4 nvidia jetsons and all of them presents the described behaivor, and all of them were installed from the default image. I presume that doing the same thing again will throw the same result, at least that you recommend us something different for this new try.
Thanks for your comment. Wana confirm if all of those 4 device have this problem with the default OS/kernel we may need to figure what happen to these devices. Maybe go through RMA.
Hello @ShaneCCC, any updates about solve this issue without RMA process?
Thanks in advance.
First, you may need to figure out which process to do the i2c scan. The default kernel driver should only access the i2c device that define in the device tree and the status= “okay” only.
Hi @ShaneCCC, this is what we want to know, which process do the scan because we are not inizializing it.
I dont think its a HW issue.
SW is for sure running some i2c scan.
What all peripherals you have connected?
Can you share the compete kernel log and sys log.
You can add a dump_stack(); in this function to know the process.
or you can add a print there like
printf(buf,"\n Name :%s \n Process id: %d\n ",current->comm,current->pid);