Hello,
Veoneer is having the following issue with the SAF
Issue after going with work around resolution we are observing following failure with both gcc and ghs builds:
i.e. we are seeing eMcem Initialization handling failure and never coming out of it
With GCC also observing same behaviorIr
Please Comment,
John
Hello @john_floros,
this looks to be similar issue as already reported in other S32K community thread:
S32K3 SAF init issue - NXP Community
Please check in that thread the workarounds to be used. In brief the issue is that debugger sets the FCCU faults which cannot be cleared anymore by SW resulting in eMcem_Init failure. To verify whether customer is impacted by the same issue they can try to run without debugger connected (then the issue does not happen) or use debugger attach functionality. In case the customer issue would not be caused by debugger then we would need FCCU register dump plus DCMROD3,4,5 register dump to analyze which faults are not cleared, and also to get more information where within eMcem_Init() the failure happens.
If using S32DS debugger then the attach functionality as one of possible workaround can be done as follows: download the application elf file first (i.e. flash it). Then press reset button. And then reconnect PE debugger with attach option: uncheck "Load Executable", and check "Attach to Running Target" in the debug configuration "Startup" tab.
Best Regards,
Frantisek