How to get MSP430 Code to work without using libraries? - c

Question

Ok so I am trying to ditch energia texas instruments arduino style ide and I have used IAR for coding a Tiva C series development board where I was able to use pointers to memory locations to perform specific things like toggling led for example. I have had a hard time doing the same on a dev board running a MSP430FR5994 mcu, I know the memory address of the green led pin to be PORT 1 PIN 1 OR P1.1 on the board. I also have included the msp430.h header file for an api to the board from my ide. What I don't understand is why when in debug my code is changing the value of the correct registers to the correct numbers but it is not altering the board. I have also verified that it is connected to the board as it will not proceed to debug with it unplugged. My direct questions are this: 1 I should be able to alter memory locations with no headerfiles or any special api's as long as I know the specific addresses correct? 2 I did not see anything about clock gating in the data sheet and in debug I can see those registers changing values so is there something other than setting the pin direction and value that I need to do?( the default pin function is generic gpio I checked so I left that register alone. Any ideas or pointing out obvious errors in my approach would be very helpful thanks. In the code below I used the header file names as I could not get the direct pointers to work. Also I was confused by the data sheet as the base address for port 1 was written as 0200H which is 5 hex numbers when I was expecting 4 since the chip is 16bit system? I assumed with the offsets it meant 0x202H etc am I incorrect in this assumption?
registers during debugging image
ti.com/lit/ds/symlink/msp430fr5994.pdf (datasheet port 1 mem locations page 130)
#include <msp430.h>
/**
* main.c
*/
int main(void)
{
WDTCTL = WDTPW | WDTHOLD; // stop watchdog timer
while(1){
int i ;
int j ;
P1DIR = 2;
//*((unsigned int *)0x204Hu) = 2;
P1OUT = 2;
//*((unsigned int *)0x202Hu)= 2;
for( i = 0; i< 2 ; i++){}
P1OUT = 0;
//*((unsigned int *)0x202Hu)= 0;
for (j = 0 ; j< 2; j++){}
}
return 0;
}

Answer 1

See section 12.3.1 of the MSP430FR59xx User's Guide.
After a BOR reset, all port pins are high-impedance with Schmitt
triggers and their module functions disabled to prevent any cross
currents. The application must initialize all port pins including
unused ones (Section 12.3.2) as input high impedance, input with
pulldown, input with pullup, output high, or output low according to
the application needs by configuring PxDIR, PxREN, PxOUT, and PxIES
accordingly. This initialization takes effect as soon as the LOCKLPM5
bit in the PM5CTL register (described in the PMM chapter) is cleared;
until then, the I/Os remain in their high-impedance state with Schmitt
trigger inputs disabled.
And here is the example blinky code provided in the MSP430FR599x Code Examples available to download from here.
#include <msp430.h>
int main(void)
{
WDTCTL = WDTPW | WDTHOLD; // Stop WDT
// Configure GPIO
P1OUT &= ~BIT0; // Clear P1.0 output latch for a defined power-on state
P1DIR |= BIT0; // Set P1.0 to output direction
PM5CTL0 &= ~LOCKLPM5; // Disable the GPIO power-on default high-impedance mode
// to activate previously configured port settings
while(1)
{
P1OUT ^= BIT0; // Toggle LED
__delay_cycles(100000);
}
}
You probably need to add that PM5CTL0 &= ~LOCKLPM5; line to your code.
And then single-step through your code in the debugger to observe the LED. Because if you let your code run at full speed the delay loops are way too short to observe the LED flash with your eye.

Answer 2

This is derived from one of my examples; I don't have a card handy to test it on, but it should just work or be close.
startup.s
.word hang /* 0xFFE0 */
.word hang /* 0xFFE2 */
.word hang /* 0xFFE4 */
.word hang /* 0xFFE6 */
.word hang /* 0xFFE8 */
.word hang /* 0xFFEA */
.word hang /* 0xFFEC */
.word hang /* 0xFFEE */
.word hang /* 0xFFF0 */
.word hang /* 0xFFF2 */
.word hang /* 0xFFF4 */
.word hang /* 0xFFF6 */
.word hang /* 0xFFF8 */
.word hang /* 0xFFFA */
.word hang /* 0xFFFC */
.word reset /* 0xFFFE */
reset.s
.global reset
reset:
mov #0x03FF,r1
call #notmain
jmp hang
.global hang
hang:
jmp hang
.globl dummy
dummy:
ret
so.c
void dummy ( unsigned short );
#define WDTCTL (*((volatile unsigned short *)0x015C))
#define P1OUT (*((volatile unsigned short *)0x0202))
#define P1DIR (*((volatile unsigned short *)0x0204))
void notmain ( void )
{
unsigned short ra;
WDTCTL = 0x5A80;
P1DIR|=0x02;
while(1)
{
P1OUT |= 0x0002;
for(ra=0;ra<10000;ra++) dummy(ra);
P1OUT &= 0xFFFD;
for(ra=0;ra<10000;ra++) dummy(ra);
}
}
memmap
MEMORY
{
rom : ORIGIN = 0xC000, LENGTH = 0xFFE0-0xC000
ram : ORIGIN = 0x1C00, LENGTH = 0x2C00-0x1C00
vect : ORIGIN = 0xFFE0, LENGTH = 0x20
}
SECTIONS
{
VECTORS : { startup.o } > vect
.text : { *(.text*) } > rom
.bss : { *(.bss*) } > ram
.data : { *(.data*) } > ram
}
build
msp430-gcc -Wall -O2 -c so.c -o so.o
msp430-ld -T memmap reset.o so.o startup.o -o so.elf
msp430-objdump -D so.elf > so.list
msp430-objcopy -O ihex so.elf out.hex
examine output
so.elf: file format elf32-msp430
Disassembly of section VECTORS:
0000ffe0 <VECTORS>:
ffe0: 0a c0 bic r0, r10
ffe2: 0a c0 bic r0, r10
ffe4: 0a c0 bic r0, r10
ffe6: 0a c0 bic r0, r10
ffe8: 0a c0 bic r0, r10
ffea: 0a c0 bic r0, r10
ffec: 0a c0 bic r0, r10
ffee: 0a c0 bic r0, r10
fff0: 0a c0 bic r0, r10
fff2: 0a c0 bic r0, r10
fff4: 0a c0 bic r0, r10
fff6: 0a c0 bic r0, r10
fff8: 0a c0 bic r0, r10
fffa: 0a c0 bic r0, r10
fffc: 0a c0 bic r0, r10
fffe: 00 c0 bic r0, r0
Disassembly of section .text:
0000c000 <reset>:
c000: 31 40 ff 03 mov #1023, r1 ;#0x03ff
c004: b0 12 0e c0 call #0xc00e
c008: 00 3c jmp $+2 ;abs 0xc00a
0000c00a <hang>:
c00a: ff 3f jmp $+0 ;abs 0xc00a
0000c00c <dummy>:
c00c: 30 41 ret
0000c00e <notmain>:
c00e: 0b 12 push r11
c010: b2 40 80 5a mov #23168, &0x015c ;#0x5a80
c014: 5c 01
c016: a2 d3 04 02 bis #2, &0x0204 ;r3 As==10
c01a: a2 d3 02 02 bis #2, &0x0202 ;r3 As==10
c01e: 0b 43 clr r11
c020: 0f 4b mov r11, r15
c022: b0 12 0c c0 call #0xc00c
c026: 1b 53 inc r11
c028: 3b 90 10 27 cmp #10000, r11 ;#0x2710
c02c: f9 23 jnz $-12 ;abs 0xc020
c02e: b2 f0 fd ff and #-3, &0x0202 ;#0xfffd
c032: 02 02
c034: 0b 43 clr r11
c036: 0f 4b mov r11, r15
c038: b0 12 0c c0 call #0xc00c
c03c: 1b 53 inc r11
c03e: 3b 90 10 27 cmp #10000, r11 ;#0x2710
c042: f9 23 jnz $-12 ;abs 0xc036
c044: ea 3f jmp $-42 ;abs 0xc01a
looks fine the vector table is there and points to the right place, etc.
10,000 might not be enough to see the led blink.
From the datasheet it appears that 0x202 is P1OUT and 0x204 is P1DIR
And you have to get it programmed into the board. I use mspdebug for the boards I have but that program may have stopped working on the eval boards from TI a while ago. And mspdebug supported the Intel hex format. So use objcopy for other formats.
If you were not wanting to use gnu tools then you still have to deal with the vector table and the bootstrap in front of the C code if you are trying to get away from someone's sandbox and do your own thing.
You are on the right path it may be as simple as your delay loops are way way too small and or getting optimized out since they are dead code as written.
If you rely on .bss or .data being initialized then you have more work to do in the linker script and bootstrap. I don't, so don't have that problem, actually wondering why .data was in this linker script...
I've matched addresses to the datasheet for your part, increasing the odds of success. If you use an external function and pass the loop variable to that function (the external can be C or asm, doesn't matter) then the optimizer won't remove it as dead code. That or add volatile on the loop variable and check the disassembly to see that it wasn't removed.