Hi Team,
We are using i.MX 8XLite Processor in our Design.
There are 4 I2C interfaces (0 to 3) from the Processor and a dedicated I2C for PMIC. Multiple slaves are allocated to these 4 I2C instances with different I2C address on each bus.
5 slave devices are present on the I2C0 bus. Currently, This bus is actively used in our Design.
4 slave devices are present on the I2C1 bus. Currently, This bus is not actively used in our Design.
4 slave devices are present on the I2C2 bus. Currently, This bus is actively used in our Design.
3 slave devices are present on the I2C3 bus. Currently, This bus is actively used in our Design.
During Testing , one of the slave device present on I2C2 bus failed to respond to the processor. Immediately, We observed the following
We would like to understand whether one I2C bus failure cause 100% CPU usage on the i.MX proc and affect other I2C buses?
Thanks
Phanindhar
Thank you Joseph for your response.
So, we think our processor is trying establish connection with the failed I2C device, which is causing increase in CPU usage.
Our Other query is
Could this issue also cause failures or impact the operation of other I2C buses?
Thanks
Phanindhar
Hi,
Thank you for your interest in NXP Semiconductor products,
I don't think that I2C failure can cause an increment on CPU usage, probably if you are trying to locate the I2C device over and over again might affect CPU usage.
But you could try to monitor CPU usage in a normal operation against an operation without the faulty I2C device.
Regards
Hi Team,
Could you please support on this Query.
If any doubts, please let me know.
Thanks
Phanindhar
Hi @Phanindhar,
Yes, there are several points where TX is polled in 8DXL I2C driver, you could debug those while structures to confirm where is it stuck.
Regards
Hi Joseph,
Error what we observed is a one time error when we are testing at higher temperature(70C). Only at one instance, one of the I2C slave device in a bus failed and it caused increased in processor loading. Post that, we observed error in someother I2C bus as well. So we dont have option to run with break point and check now as we are not able to recreate issue again..
Eventhough there are multiple/different I2C bus for processor, we like to know whether different I2C bus are interlinked internal to the processor which leads to a condition like failure of one I2C bus can cascade failure to other I2C bus as well.
Hi Joseph,
Any update for our query?