We are currently running into an issue were we are getting a kernel error every once in a while when writing to an encrypted partition set up following AN12714 Rev. 2. Sometimes the kernel will report a crash saying "Unable to handle kernel paging request at virtual address". I was able to recreate the crash by using the dd command to write a large file to the encrypted partition. However, it does not seem to cause the crash every time. Below is some more information on our hardware setup:
Do you have any idea what the cause of this crash could be? is there any issue to how we are using the CAAM?
Thanks for the reply. After further troubleshooting, it looks like it is not specifically tied to the CAAM encryption. I was able to get an crash by just creating 6 threads of "yes > /dev/null".
I was looking through the IMX8_1N94W errata document and ERR050104 mentioned a cache coherency issue. Could this be related to the crash? Is there already a patch in the kernel to address this issue?
We had something similar - the issue seemed to be resolved by dropping the clock frequency down from 1.6GHz to 1.3GHz.