Dear all,
We are using an imx93 on a custom board with the A55 core running Linux and the M33 core running an embedded application.
The application of the M33 core does not fit in 128 KB TCM memory and is therefore executed in external DDR. We have reserved an memory region and are using rproc or u-boot to load and execute code in the m33 core.
We are now seeing that when the M33 core runs into an infinite loop (for (;;)), the A55 core hangs. After 40 seconds the watchdog resets the system.
When executing from TCM memory and entering the same infinite loop, the A55 core runs without any problem. The M33 core off-course hangs, but the system remain intact.
Does anybody have any experience with this configuration and or problem?
I know that NXP provides all demo's running from TCM memory and not from DDR.
It is possible that the M33 core running an infinite loop creates starvation on the memory bus and therefore silently killing the A55 core running Linux.
Thanks,
EG
