Hard fault with BFAR=4000010 (STM32F2)

[root]

Posted on January 31, 2012 at 19:46

Hello,

I have a hard fault when running my code, and the result of the hard fault is :

[Hard Fault Handler]

R0  = 4000000 R1  = 1 R2  = 20003fe0 R3  = 40013c00 R12 = 0 LR  = 8014acd PC  = 8019474 PSR = 100022c BFAR= 4000010 CFSR= 8200 HFSR= 40000000 DFSR= b AFSR= 0

All values are in hexa. The thing I don't understand is that CFSR tells me there is a precise bus error and that BFSR is valid ... but the value in BFSR is obviously not valid.

It always trigger while entering a timer interrupt, which is coded as this :

void TIM2_IRQHandler(void)

{  if (TIM_GetITStatus(TIM2, TIM_IT_Update) == SET)  {   TIM_ClearITPendingBit(TIM2, TIM_IT_Update);   [a few lines of code that modify a variable]  }

 if (TIM_GetITStatus(TIM2, TIM_IT_CC3) == SET)

 {   TIM_ClearITPendingBit(TIM2, TIM_IT_CC3);   [a few lines of code too, but this is never called, checked with a breakpoint]  } }

I use other timers with no problems, any clue on what's happening ?

Thomas.

[UPDATE] The variable I'm modifying in the interrupt is set as volatile, it seems that if I don't set this variable as volatile, the hard fault doesn't trigger. I'm using TASKING 4.0.

[Tesla DeLorean]

Posted on January 31, 2012 at 20:11

Can you list the assembler code in proximity to the faulting PC?

Can you see anything doing a code walk of the interrupt routine, as compiled, wrt the registers at the fault. Where are the stack pointers at, and what part or memory limits are you working with?