Hello team,
Autoever reported the S32K324 ITCM ECC error after enabling the core_1. Currently, the sequence below is applied.
1. Enable the CM7_0 only in IVT
2. CM7_0 reset handler -> Mcu_Init -> Mcu_InitClock -> Mcu_SetMode -> Peripheral Driver Init -> OS StartCore
3. Enable CM7_1 in Os_StartCore
The ITCM ECC error occurred in the reset handler just after enabling the CM7_1 core.
There is the TCM initialization routine in the CM7_1 reset handler, but the ECC error occurred before this routine is exectured.
The CORR_ERR_CNT13 in ERM was increased though the CM7_1 was stopped at break point.
Can you please give me some advice about what we should check?
Best regards,
James
Hi @james-lee ,
yes, this is known error on K3.
I was trying to find root cause from HW/Bootloader team: https://community.nxp.com/t5/HSE-FW-Support/TCM-ECC-errors-on-S32K3-devices/m-p/1987897
But it's still unclear.
ITCM and DTCM are enabled before entering reset handler and for some error the error is randomly triggered right after a first few instructions executed (even they are not related to TCM).
One workaround was to move TCM init startup routine right in the beginning of startup, that worked in many cases, but not for 100% on some K3 derivatives it was raised randomly even during TCM init part.
Second workaround was to ignore these faults until they are cleared by eMcem_Init().
After clearing they will be not triggered again once TCM is fully initialized in startup.
So, in case more cores are started, each core's TCM could raise this error, so that limits FCCU/DCM fault configuration (eMcem_Init()) to be called after all cores are booted if you don't want to trigger configured reaction for these TCM faults.
Kind Regards,
Radoslav