topic Re: MPC5777C RCCU_0 in MPC5xxx

MPC5777C RCCU_0

mauriziogreco — Thu, 13 Sep 2018 14:24:38 GMT

Hi,

we are working with the FCCU and when we active RCCU_1 (NCF_11) we don't see any action by FCCU. In our opinion this means control is active but no fault is present.

But when we active the RCCU_0 (NCF_10) we se action by FCCU.

We have not any notice about difference between RCCU_0 and RCCU_1!

Can someone suggest how investigate reason of reaction of FCCU when we active RCCU_0?

Regards

Re: MPC5777C RCCU_0

davidtosenovjan — Wed, 19 Sep 2018 18:49:55 GMT

Hi,

I have tried to investigate what signals are actually

RCCU_0 is divided into 3 RCCU units (RCCU0 ORed):

RCCU_0_a - Miscelaneus Signals Lockstep Checks

RCCU_0_b - Instr. CacheLockstep Checks

RCCU_0_c - Data Cache Lockstep Checks

RCCU_1 is not pretty clear where it comes from. There are predictors it could be taken from e2eECC gasket at the core system bus interface. But I will have to ask factory engineer to approve or disprove this assumption. I am coming back to you.

Meanwhile you could tell me details about operation you are performing when the fault occurs. Thanks

Re: MPC5777C RCCU_0

mauriziogreco — Thu, 20 Sep 2018 07:35:03 GMT

Hi David,

thanks for your analysis.

I found activation of RCCU_0 only one time at start-up of MPC5777C, after no more reaction I found during running.

Do you think I miss some register initializzation of core 1?

Thanks in advance.

Re: MPC5777C RCCU_0

mauriziogreco — Thu, 20 Sep 2018 07:38:19 GMT

Hi Davide, sorry......for this new issue!

But if you really can ask to factory engineer, i have also a problem with NCF_42 on MPC5777C!

Can you ask if RCCU is stopped using Debugger?

I found in the forum different point of view about this!

Many Many Thanks.

Re: MPC5777C RCCU_0

davidtosenovjan — Mon, 08 Oct 2018 12:41:50 GMT

NCF_42 is present in debug mode, what's the case.

If you run the code without debugger, this fault should disappear.

Re: MPC5777C RCCU_0

davidtosenovjan — Mon, 08 Oct 2018 12:44:01 GMT

I was told that RCCUs are splitted into 2 instances only due to too many signals needed to be integrated. From SW point of view RCCU_0 and RCCU_1 needs to be handled same way i.e. both faults needs to be logically ORed.

Hope it helps