CCS 5.3 cstartup

0x1f2c is normal ram. Writing to it has no special effect except altering the ram cell content.

If you get a reset at this point, it is likely the watchdog that gets a timeout because the initialization of your variables took too much time. However, this looks a bit early (but may have to do with you looking at the code - I assume you use the debugger).

Nevertheless, you may add a custom init (see compiler/linker documentation) and disable the watchdog before initializing the variables. Or you may use the legacy COFF output format, which doesn't initialize uninitialized variables to 0 and therefore doesn't take that much time before it starts main.

I hadn't considered the watchdog, but it make sense...  As my RAM usage increased, so does the initialization time.

It failed both when debugging (single-step/breakpoints) and free running (how I initially encountered it), I'm assuming the debugger pauses the watchdog when the processor is halted?

thanks,

-neil

Neil Fortney1 said:

I'm assuming the debugger pauses the watchdog when the processor is halted?

Also, during debugger access, the processor doesn't get every clock cycle while the watchdog will, so if you're single-stepping, the watchdog may bark sooner.