Hi,
> >It seems that i have to change the section definitions by adding the =fill
> option. But i could not
> >find the definitions of the MSP430 sections.
> >Where are the definitions of the MSP430 sections?
> >
>
> In the mspgcc tree, there is a directory called ldscripts. These scripts
> define what the memory of the target looks like. However, I tried to use
> the fill (fill should be replaced by a value, for example: >text =0xa5a5),
> but for some reason it doesn't have an effect. Maybe Steve can shine some
> light on this?
Yes, tha's a good idea.
> I did found out (by looking at the .a43 intel hex output) that the
> interrupt vectors are initialised with a default vector. Filling these will
> not have an effect at all because the entire section is already occupied.
Well, that's no a real problem because i do use every IRV, because "unused"
interrupts can cause problems when they happen.
I am using e. g.
#ifdef MSPGCC
interrupt (NMI_VECTOR)
#endif
#ifdef IAR
interrupt[NMI_VECTOR]
void
#endif
nmi (void)
{
mc_ISR_TOGGLE;
mc_RESET;
return;
}
with
#define mc_RESET (*(void(**)(void))(0xfffe))()
// alternate version:
// #define mc_RESET { WDTCTL = 0xDEAD }
// for counting all (or some) irqs with an oscilloscope for checking the irq
frequency
#if defined(ToGgLe) || defined(TOGGLE)
# define mc_ISR_TOGGLE P1OUT ^= 0x10;
#else
# define mc_ISR_TOGGLE
#endif
(and some macros for detecting mspgcc and iar).
I had the password problem with MSPs flashed with a BSL adapter or an update
program with a flash loader
and old programs with unused IRVs.
So i need the initializing only for the other sections for calculating the
checksum of the whole Flash but for safety
it's better also to initialize the IRV section with a selective bit pattern, e.
g. 0xfffe for reset (which is better than
hangup).
Regards,
Rolf
