Posted on May 17, 2011 at 13:11Glutton for punishment that I am, I have been trying to build CodeSourcery tools on my iMac. The tools built OK in the end so I tried to put the simplest of programs into my STM32-SK board from IAR. It has a STM32F103RBT6 processor.
I used a simple bit of blinky code from fosstronics.com thus:
[code]
#define GPIOB_CRH (*((volatile unsigned long*)(0x40010c00 + 0x4)))
#define GPIOB_BSRR (*((volatile unsigned long*)(0x40010c00 + 0x10)))
#define RCC_APB2ENR (*((volatile unsigned long*)(0x40021000 + 0x018)))
__asm__(''.word 0x20001000'');
__asm__(''.word main'');
main()
{
unsigned int c = 0;
RCC_APB2ENR = (1 << 3);
GPIOB_CRH = 0x44444114;
while(1) {
GPIOB_BSRR = (1 << 10); // ON
for(c = 0; c < 100000; c++);
GPIOB_BSRR = (1 << 26); // OFF
for(c = 0; c < 100000; c++);
}
}
[/code]
it compiles just fine. I link it with this very simple linker script:
[code]
SECTIONS {
. = 0x0;
.text : {
*(.text)
}
}
[/code]
(sorry - lookslike BBcode is not allowed for some reason)
According to objdump, everything goes where it should. That is, the first couple of bytes contain the address of the stack in RAM and the next couple are the address of main().
I download the binary to the processor with a bootloader. I am happy that is putting stuff where it should and the code reads back from the chip just fine.
But - when I put he chip back into run mode, nothing happens after a reset or a power cycle. no light blink, nothing.
I had previously been playing with a crossconnect from Rowley. That no longer seems to work and cant talk to the board. Not sure if it is the crossconnect or the processor. The processor, presumably, works OK or the bootloader would fail.
I also have an Olimex STM32-P103 board that behave exactly the same way.
After a couple of hours trying variations of the same thing, I have had enough. I must be doing something wrong.
Is it possible to have put the processor into some state where it cannot run? Or have I just overlooked something obvious?
Does anyone have any other suggestions?
