NMI Fault While Debugging (STM32G474 nucleo board)

[toofer25]

Hello,

I keep getting an NMI fault when I attempt to read a variable with gdb (either with a mouseover on a variable in STM32CubeIDE or the print [var] command in gdb). I'm not seeing anything useful when I view the microprocessor registers after the fault.

I'm using the stm32g474 hal library along with a linker, startup file, and interrupt _it files that are normally generated by default with the STM32CubeMX software for a G474 nucleo board. However, I'm using cmake 2.47.6 and the arm gcc/g++ toolchain 11.2-2022.02 for the build process.

Thank you

[SofLit · Accepted solution]

Hello,

According the the user manual, there are three sources to generate an NMI:

1. Clock security system (CSS)
2. Flash ECC error
3. Parity check SRAM1/CCM SRAMs. 

I suspect the 3rd choice: Parity check

So check the user option byte SRAM_PE value in option bytes menu in STM32CubeProgrammer.