program thinking 019c and 009c are equal

I entered your application and tried it and it works just fine.

#include <reg51.h>
#include <stdio.h>


#define MAXINBUFF 128

unsigned short int xdata x;
unsigned char xdata META_TEDS[MAXINBUFF];

void main (void)
{
SCON  = 0x50;               /* SCON: mode 1, 8-bit UART, enable rcvr      */
TMOD |= 0x20;               /* TMOD: timer 1, mode 2, 8-bit reload        */
TH1   = 221;                /* TH1:  reload value for 1200 baud @ 16MHz   */
TR1   = 1;                  /* TR1:  timer 1 run                          */
TI    = 1;                  /* TI:   set TI to send first char of UART    */


for( x=2;x<12;x++ ){
  printf( "UUID %d\r\n", x );
  printf( "UUID address %p\r\n", &x );
  printf( "META_TEDS address %p\r\n", &META_TEDS[x] );
  META_TEDS[x]=_getkey( );
  printf( "UUID %d\r\n", x );
  printf( "UUID address %p\r\n", &x );
  printf( "META_TEDS address %p\r\n", &META_TEDS[x] );
}

while (1);
}

we have not modified the _getkey function and we have no pointers in the code. everything is a global variable. we are however running at 9600 baud using the following settings:

T3CON = 0x82;
T3FD = 0x12;
SCON = 0x52;

instead of reg51.h we are including aduc834.h from analog devices. it contains all that reg51 has plus maybe 2 or 3 other lines that reg51 does not have. we are using the ADuC834 evaluation board, and win2k sp3.

-Nathan

You're using %d to print chars - which is a very common mistake.
You need to use %bd

>You're using %d to print chars - which is a very common mistake.
>You need to use %bd

No he's not. The variable being printed with %d is "x", which is a short int.

Note that x was 10d = 0x0A
it got changed to 12554d = 0x310A
So it looks like an ASCII '1' has been written into the upper byte of x.

How is XDATA implemented on your hardware? Maybe the upper address byte is not being latched properly, or something else is modifying it.

Keil use "big endian" storage for integers, so the upper byte of x is at address 0x019c, so I am confident that you have a hardware problem with the upper half of the address for XDATA.

it seems that in the above segment during the said iteration what _getkey() retrieves is put into 019c (x) instead of 009c (META_TEDS[x])

If you think that's the case, then you should modify your program to output what's stored in META_TEDS[x] each iteration.

Jon

Thanks for all the help guys, I found out the problem. Since i got on this project a few months ago we have been using the same eval board and I have only changed a few of its link positions. until recently we have never had over 256 bytes in the program so everything functioned with LK9 in the B position. We are now up to about 458 in xdata and LK9 in position B has been causing these inconsistancies in our variables. With LK9 in position A it does function properly. Thanks again for your comments.

-Nathan

** 256 bits, not bytes.

-Nathan