Lpc2000

At 12:06 PM 3/17/05 -0300, Leonardo Santos wrote:

>Hello All:
>         I'm working with a LPC2106 using the GCC 3.4.4 toolchain (GCC + 
> binutils 2.15
>+ newlib 1.13). I wrote a union and a structure to access the P0 pins
>individually, like this:
>
>typedef struct {
>         int p0  : 1;
>         int p1  : 1;
>         int p2  : 1;
>(...)
>         int p30 : 1;
>         int p31 : 1;
>}port_t;
>
>typedef union {
>         port_t pins;
>         unsigned int reg;
>} ioreg;
>
>typedef struct {
>         ioreg iopin;
>         ioreg ioset;
>         ioreg iodir;
>         ioreg ioclr;
>}gpio_t;
>
>So I declare a pinter like
>
>         volatile gpio_t *gpio = 0xE0028000;
>
>If I access the registers using
>
>         if( gpio->iopin.reg & 0x01 )
>
>or
>
>         gpio->ioset.reg |= 0x00008000;
>
>Everithing works fine! But if I try a code like this:
>
>         gpio->ioset.pins.p15 = 1;

P15 is a 1 bit int so its only values are 0 and -1 but I don't think that's 
the problem.   Provided your bit order is correct I don't see any reason 
they shouldn't give the same result.  I'd use unsigned int for bitfields 
unless you wanted them to be signed for some particular purpose.


>To set P0.15, P0.23 and P0.31 are set too! I had tried this code in my Linux
>workstaion and it works just fine. The compiles setting I'm using are

Hmm, those are 8 bits apart.  Have you taken a look at the assembly output 
for the two cases?  They should be similar if not identical. Unless the 
compiler is getting confused by some sort of sign extension issue?  Seems a 
stretch though.

Robert

" 'Freedom' has no meaning of itself.  There are always restrictions,   be 
they legal, genetic, or physical.  If you don't believe me, try to chew a 
radio signal. "  -- Kelvin Throop, III
http://www.aeolusdevelopment.com/