I've just compiled the officially supported demo for the STM32F107VC. This one:
Using Rowley Crossworks.
It compiled with a couple of tweaks to correct for / in the user directories and I had to add an extra one.
It loads, but once it gets into vTaskStartScheduler() it jumps off somewhere funky (in the 0x1FFFC9xx area of memory. If I set a breakpoint in vTaskStartScheduler() and do a step into with the debugger, it acts like I have hit F5 to start running again. ie It won't let me debug.
I can set the breakpoint in vTaskStartScheduler() all the way down to:
if( xPortStartScheduler() != pdFALSE )
However, if I set the breakpoint above there and try a single step, it does this restarting of the debugger as if it were running the program.
Anyone seen anything like this before? Never come across the debugger doing this on a run-to-complete program, let alone other version of FreeRTOS I've used!
Ach, sorry, it was the boot switch on the board, I misuderstood what System Meory means on the silk screen. Sorry for the stupid question. :~(