CodeWarrior for MCUのトピックhazardous behaviour with HCS12 MC9S12GC64

hazardous behaviour with HCS12 MC9S12GC64

davidCH — Fri, 08 Dec 2006 00:39:15 GMT

Hi

i'm getting crazy with debugging a very "simple" application: reading 5 sensors and controling max 4 outputs and a display with icons :
I use a bootloader to load my application

- with a "empty" periodic loop (every 30s), no problem since application do nearly nothing it's difficult to see if there's a trouble
- when i add: acquisition routines , tests routines to see if the inputs are in correct range , simple control routine which activate an output if inputs are in incorrect range
, the application become to do hazardous things: global variables that are not accessed for write (just read) by the current loop are modified.

Inside the loop, i use mostly byte variable, accessing with bit mask for reading and sometimes for writing, but only global varaible from the "applicative" file (not the "kernel")
I attached the applicative file, the main fonction with the loop is the "CESI" fonction, called by the kernel after boot.

I hope somebody can help me

Thanks

David

Re: hazardous behaviour with HCS12 MC9S12GC64

CompilerGuru — Fri, 08 Dec 2006 05:39:02 GMT

This
>global variables that are not accessed
>for write (just read) by the current loop
>are modified.

sounds like a stack overwrite problem. How large is your stack (STACKSIZE in the prm file)? Does it help if you increase it.
Did you try to set a watchpoint to the changing globals?

Daniel

Re: hazardous behaviour with HCS12 MC9S12GC64

Lundin — Fri, 08 Dec 2006 20:44:07 GMT

Or simply download the whole program to the debugger. Before you start the CPU, set all cells in the stack memory to zero. Let the program run for a while and then check how much of the stack it has munched away at.

And if you expect others here to read through your code, please write it in english.