eMcem_Fcc_Init failed to clear errors

[Sara_Sara]

Hello,

I run the application on S32K388 and then attach trace32 debugger. 

In eMcem_Fcc_Init(), read NCF_S0 status register returns u32ReadBack=0x5 (NCF[0] HSE lockstep error and NCF[2] ECC error). 

and eMcem_Fccu_ClearFault() doesn't clear the faults too.

How can I clear the faults to make eMcem_Fcc_Init() and eMcem_Init() initialised successfully?  What could be the issue ?

Thanks

Sara

[petervlna]

Hello,

NCF_S0 = 0x5 indicates that both FCCU NCF[0] (lockstep/core fault) and NCF[2] (ECC fault) are active. On S32K3 devices, these fault flags are retained after an FCCU functional reset so that software can identify the root cause. The FCCU fault status cannot be cleared while the originating error condition is still present. Therefore, the issue is likely not in eMcem_Fccu_ClearFault(), but that a lockstep fault and an ECC fault are still being reported by the underlying hardware.

Please check the corresponding FCCU, RCCU/DCMROD, and ERM status registers to identify the exact source. Since the issue occurs after attaching TRACE32, also verify that FCCU Debug Mode is enabled and compare behavior when the debugger is connected from reset.

Best regards,

Peter

[Sara_Sara]

Hello Peter,

Thanks for your response and analysis need to be verified. After looking into the FCCU, RCCU/DCMROD, and ERM status registers, the core 2 got into Lockup state. 

On disblaing the CM_7_1/2/3 coreclockEnable of MCU config, the lockup fault doesn't happen and the eMcem_init got configured successfully and the status of NCF_S0 is 0x0.

The coreclock can be enabled at the application startup on each core and not in the Mcal by default. This solves the issue.

Thanks

Sara