Hi,
I am debuging a STM32F4 / Cortex-M4. MDK-ARM Pro 5.18a I am trying to find the cause of a NaN Value. Is it possible to set an access breakpoint triggering for 1.#QNAN?
Timo
Not directly, you might be able to trap the access, or with a trace pod determine causes/flow.
If the variable is on a stack, make sure the stack is large enough and you don't out of bounds arrays access corrupting the stack frame. The default stack and heap sizes are small, it is not hard for the stack to plough into the locale space.
If your issues are math related, add some sanity checking to confine the issue. You can add checking routines that can BKPT also.
I managed to trigger on NaN (0x7FC00000).
BS WRITE (*((unsigned long *)0x20000F08)) == 0x7FC00000
more detail in my other post. http://www.keil.com/forum/61509/