• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 31 subscribers
  • Views 866 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TM4C123GH6PZ: Attempt to write NULL pointer is silently ignored

Eric Wilson85
Prodigy 120 points

Part Number: TM4C123GH6PZ
Other Parts Discussed in Thread: TM4C129EKCPDT

Hi,

When I run the following code what I expect is that the attempt to write to a null pointer would result in a hard fault. Instead what we've found is that the instruction is silently ignored. This seems to contradict what is shown in the datasheet:

"Note: Within the memory map, attempts to read or write addresses in reserved spaces result in
a bus fault. In addition, attempts to write addresses in the flash range also result in a bus
fault."

uint32_t var = 0xDEADBEEF;
	
uint32_t * pointer_to_var = &var;
uint32_t * pointer_to_null = NULL;

*pointer_to_var = 0;
*pointer_to_null = 0;

Can anyone shed some light on this? Have we somehow disabled faults for this particular error or is this even possible?

Thanks,

Eric.

  • 0
    TI__Guru 72500 points
    I see the initialization of your null pointer, but I don't see the write to location 0 in your code.
  • 0 in reply to Bob Crosby
    Prodigy 120 points

    Hi Bob,

    line 303 in the screenshot tries to write to the NULL pointer:

    *pointer_to_null = 0;

  • 0 in reply to Eric Wilson85
    TI__Guru 72500 points

    Sorry, I must be going blind.

    I am able to reproduce your results, and I see the statement on page 92 of the datasheet. At this point, all I can say is that the MPU is not enabled. You can cause an abort on a write to flash memory by programming the MPU and designating that region as read only. I will look into this further. At this point I suspect the statement in the datasheet is in error and should be corrected.

  • 0 in reply to Bob Crosby
    Prodigy 120 points

    Hi Bob,

    I tried similar code on both the TM4C123 and the TM4C129EKCPDT.

    int main() {	
	int * pointer_to_null = 0;
	*pointer_to_null = 1;
	
	while(1);
}

void HardFault_Handler(void) {
	while(1) {
		__asm("BKPT 0");
	}
}

    The TM4C129 hit the breakpoint in the hard fault as expected. The TM4C123 remained in the while loop in main(). Both datasheets have the same note in the "memory model" section about generating bus faults. We can definitely use the MPU to trap this error on the TM4C123 but I would be very interested to know what the difference is between these two devices.

  • 0 in reply to Bob Crosby
    Prodigy 120 points
    Hi Bob,

    Have you had a chance to look into this further? I have tried the MPU and know that it will be a viable solution but I can't shake the feeling that something is very wrong. Maybe the issue is the documentation but if it's the chip or our configuration of the chip I would really like to know if there's a workaround.

    Thanks,
    Eric.
  • 0 in reply to Eric Wilson85
    TI__Guru 72500 points
    The default memory map protection for the Cortex M4F is configured at the time of device design. The TM4C123 was designed first. It looks like the configuration for the Flash memory range was done incorrectly and does not flag writes. The TM4C129 parts were done later and done correctly. I will be submitting this as an advisory in the TM4C123 Errata Document.