MSP-EXP430F5529LP: Program will not run if another variable is added or an array is made larger

How many globals are you using?

Are you allocating a lot of variables on the stack, too?

There are definitely a lot of globals, the program has many setup variables and there is a few large 288 item arrays.

I do not know if im allocating them on the stack

You allocate on the stack when you create a variable (an "auto" variable) inside a function.

some locals are made, no where near the amount of globals though. this program is sort of a combination of 4-5 different libraries which each have their own globals but also globals from the example files. id estimate there to be atleast 20 globals.

i should probably add that even when those large global arrays are reduced down to 3 items the problem still occurs, that version of the code is using roughly 26% of the 8kB RAM

Usually this happens when you have a wild pointer that clobbers some important part of memory. It works fine when it interferes with a not-so-important variable, but once you move things around a bit you start overwriting important bits - like stack return addresses.

do you have any ideas how i could fix that issue? those large data arrays are quite important. Is there some way to increase how large the stack can be or is that a hardware limitation? or another way of storing these variables?

I just attempted to import the project into code composer to try and use the debugger, i thought it may provide me with some insight of what is wrong, but it gets a java stack overflow error. Do you think this may be related at all?

Hello Keith, i ended up replicating the issue in another energia project by creating a large array (1000 big) and used the disassembler to see if it was actually running and to my surprise it is. It seems to get to a "run_array" function and does that a few times and then starts again at the start of the program. What is weird though is that the function only has one return and that is never reached. I was wondering, if this function is taking a long time to run, could it be the watchdog timer resetting the board?

EDIT: It seems to be a specific MOV.A instruction where the program restarts after. I then changed the size of the array to see if the location would change and it seems to have, it sort of runs a loop and if i increase the size by 100 it happens earlier and if i shrink the size by 100 it takes longer

Using CCS, it is possible for the watchdog timer to time-out during C (.data/.bss) initialization if your data is too big. The usual fix for this is to define a _system_pre_init() function that stops the watchdog (before C initialization runs). [Ref CC User Guide (SLAU132s) Sec 6.9.1] Recent versions of CCS also have a project-level check-box to ask the C startup to do this automatically.

I don't know whether Energia has a similar mechanism. I suppose you could just add something like

int _system_pre_init(void) {
    WDTCTL = WDTPW | WDTHOLD;   // Stop watchdog
    return(1);                  // You can finish C initialization now
}

and see if it helps

I had just found that function, sadly it looks like energia doesnt use that function which is very annoying. Havent had any luck with finding anything about a version for energia, guess i may have hit the end of what can be done.

Hopefully energia code may be portable to a normal project in some way

Poking around github, it looks like Energia uses msp430-gcc. That may have an analogous mechanism, but I don't know it, and I don't see any hints in SLAU646E .