The target doesn't initialize arrays, why?

Could you provide the following information to see what might be your problem?

1) The size of your array?
2) The address of the array from the map file?
3) The declaration of the array in the C file?

Could you provide the following information to see what might be your problem?

1) The size of your array?
2) The address of the array from the map file?
3) The declaration of the array in the C file?

...and the startup file (startup.a66) that initializes the vars.

Thank you for for paying atention, Chris Wunderlich .
The information you requested is:
My simple code is:

#include <reg161.h>

void main (void)  {
   unsigned char array[6]={1,2,3,4,5,6};

   while (1) {}
}

(1),(3) the size and the declaration - look above;
(2) If it'll help you to discover what may be the problem, I pasted the important parts(I think so) from two files(Ithe 'array' word is red as to find it quickly):
-> "Phytec KC161.lst"

C166 COMPILER V6.06, PHYTEC_KC161                                                          04/25/2007 09:50:00 PAGE 2

NAME                                    CLASS    SPACE  TYPE   OFFSET   SIZE
----------------------------------------------------------------------------
BUSCON1. . . . . . . . . . . . . . . .  sfr             uint   FF14H    2
..
..
main . . . . . . . . . . . . . . . . .  public   FCODE  funct  -----
  array. . . . . . . . . . . . . . . .  auto            array     0H    6
?tpl?0001. . . . . . . . . . . . . . .  static   NCONST array     0H    6

MODULE INFORMATION:   INITIALIZED  UNINITIALIZED
  CODE SIZE        =          20     --------
  NEAR-CONST SIZE  =           6     --------
......

-> "project.m66" :

.......
MEMORY MAP OF MODULE:  .\Object\project (PHYTEC_KC161)

START     STOP      LENGTH    TYPE  RTYP  ALIGN  TGR  GRP  COMB  CLASS   SECTION NAME
=====================================================================================
000000H   000003H   000004H   ---   ---   ---    ---  ---  ---   * INTVECTOR TABLE *
000004H   000005H   000002H   XDATA REL   WORD   ---  ---  GLOB  ---     ?C_INITSEC
000008H   00000BH   000004H   ---   ---   ---    ---  ---  ---   * RESERVED MEMORY *
00000CH   00001FH   000014H   CODE  REL   WORD   ---  ---  PUBL  FCODE   ?PR?PHYTEC_KC161
0000ACH   0000AFH   000004H   ---   ---   ---    ---  ---  ---   * RESERVED MEMORY *
0000B0H   0001CDH   00011EH   CODE  REL   WORD   ---  ---  PRIV  ICODE   ?C_STARTUP_CODE
004000H   004005H   000006H   DATA  REL   WORD   ---    1  PUBL  NCONST  ?NC?PHYTEC_KC161
008000H   0080FFH   000100H   DATA  REL   WORD   ---    2  PUBL  NDATA   ?C_USERSTACK
00D900H   00DFFFH   000700H   ---   ---   ---    ---  ---  ---   * RESERVED MEMORY *
00FA00H   00FBFFH   000200H   ---   ---   ---    ---  ---  ---   * SYSTEM STACK *
00FC00H   00FC1FH   000020H   DATA  ---   BYTE   ---  ---  ---   *REG*   ?C_MAINREGISTERS
..........
SYMBOL TABLE OF MODULE:  .\Object\project (PHYTEC_KC161)
      VALUE       TYPE      REP       LENGTH  TGR   SYMBOL NAME
=========================================================
      00000CH     GLOBAL    LABEL     ---     ---   main
      004000H     SYMBOL    VAR       ---     ---   ?tpl?0001

      00000CH     BLOCK     LVL=0     0014H   ---   main
      00000EH     BLOCK     LVL=1     0012H   ---
      000000H     SYMBOL    AUTO      ---     ---   array
      ---         BLOCKEND  LVL=1     ---     ---
........

Hi, Joost Leeuwesteijn.
...and the startup file (startup.a66) that initializes the vars.

But what should I modify in statup.a66 file that could initialize vars?

> But what should I modify in statup.a66 file that could initialize vars?

See:
http://www.keil.com/support/docs/2330.htm
http://www.keil.com/support/man/docs/c166/c166_ap_start167_a66.htm INIT_VARS and CLR_MEMORY
http://www.keil.com/forum/docs/thread8190.asp

But your array is not a global so this should not apply... I expect the values to be in the "?tpl?0001" section but I'm not sure.

Maybe the compiler optimized it away, if this really is all your code. You could check the assembly file.

Thanks for those links, I'll see what can I do.
Maybe the compiler optimized it away, if this really is all your code. You could check the assembly file.
No. That code was just where I discovered the problem. I also tryed the follow code using Hyper Terminal. In debug mode the target is running well, but in free running not:

#include <reg161.h>
#include <stdio.h>
void setup_serial();

void main (void)  {
        unsigned char cnt,array[6]={1,1,1,1,1,1};
        setup_serial();

        for(cnt=0; cnt<6; cnt++){
                if(array[cnt] != 1)
                        putchar('!');
                else
                        putchar('1');
//in hyper terminal is displaying six of '!', WHY?
//it must be six of '1'!
        }

        while (1) {
        }
}

void setup_serial(){
  P3  |= 0x0400;        /* SET PORT 3.10 OUTPUT LATCH (TXD)              */
  DP3 |= 0x0400;        /* SET PORT 3.10 DIRECTION CONTROL (TXD OUTPUT)  */
  DP3 &= 0xF7FF;        /* RESET PORT 3.11 DIRECTION CONTROL (RXD INPUT) */
  S0TIC = 0x80;         /* SET TRANSMIT INTERRUPT FLAG                   */
  S0RIC = 0x00;         /* DELETE RECEIVE INTERRUPT FLAG                 */
  S0BG  = 0x19;                 /* SET BAUDRATE TO 19200 BAUD AT 16MHZ           */
  S0CON = 0x8011;       /* SET SERIAL MODE                               */
}

> In debug mode the target is running well, but in free
> running not

What do you mean exactly? Debug mode is on a simulator? Or on the target? And free run is on the target? Could there be a problem in the memory map/linker file? Is everything located in RAM?

What does the array contain in the incorrect situation?

- What do you mean exactly? Debug mode is on a simulator? Or on the target?
I mean debug mode is on the target.
- And free run is on the target?
Yes. Free running I mean the execution of the device after I download the .H86 file to flash of the board.

- What does the array contain in the incorrect situation?
I think that the length is not defined. When I tryed to determine what the array contains i think is 0xFF...

Thank you...

if(array[cnt] != 1)
                        putchar('!');

might be wrong.
You should better have written

if(array[cnt] != '1')
                        putchar('!');

I though Phytec gives you a START167.A66 file with the correct settings of the memory. But it looks like you are using a standard one and as Joost implied you most likely have a problem with this setup.

I believe you have the data set to Flash this is your problem. You need the data in RAM.

I would suggest that you look for this *.a66 file from Phytec or make sure you have the correct settings for SYSCON and only use the RAM on-chip for now. Your memory settings should look something like this:

From the *.lnp file

"START167.obj",
"main.obj"
TO "test"
IXREF
CLASSES (ICODE (0x0-0xBFFF), NCODE (0x0-0xBFFF),
FCONST (0x0-0xBFFF), HCONST (0x0-0xBFFF),
XCONST (0x0-0xBFFF), NCONST (0x4000-0x7FFF),
NDATA (0xC000-0xD7FF, 0xF600-0xFDFF), NDATA0 (0xC000-0xD7FF, 0xF600-0xFDFF),
SDATA (0xC000-0xD7FF, 0xF600-0xFDFF), SDATA0 (0xC000-0xD7FF, 0xF600-0xFDFF),
IDATA (0xF600-0xFDFF), IDATA0 (0xF600-0xFDFF),
FDATA (0xC000-0xD7FF, 0xF600-0xFDFF), FDATA0 (0xC000-0xD7FF, 0xF600-0xFDFF),
HDATA (0xC000-0xD7FF, 0xF600-0xFDFF), HDATA0 (0xC000-0xD7FF, 0xF600-0xFDFF),
XDATA (0xC000-0xD7FF, 0xF600-0xFDFF), XDATA0 (0xC000-0xD7FF, 0xF600-0xFDFF))
 CINITTAB (0x0-0xBFFF)

Then you will get something like this which should work.

C166 COMPILER V6.06, MAIN

NAME                                    CLASS    SPACE  TYPE   OFFSET   SIZE
----------------------------------------------------------------------------
main . . . . . . . . . . . . . . . . .  public   NCODE  funct  -----
  array. . . . . . . . . . . . . . . .  auto            array     0H    6
?tpl?0001. . . . . . . . . . . . . . .  static   NCONST array     0H    6


START     STOP      LENGTH    TYPE  RTYP  ALIGN  TGR  GRP  COMB  CLASS   SECTION NAME
=====================================================================================
000000H   000003H   000004H   ---   ---   ---    ---  ---  ---   * INTVECTOR TABLE *
000004H   000005H   000002H   XDATA REL   WORD   ---  ---  GLOB  ---     ?C_INITSEC
000006H   000125H   000120H   CODE  REL   WORD   ---  ---  PRIV  ICODE   ?C_STARTUP_CODE
000126H   000139H   000014H   CODE  REL   WORD   ---    2  PUBL  NCODE   ?PR?MAIN
004000H   004005H   000006H   DATA  REL   WORD   ---    3  PUBL  NCONST  ?NC?MAIN
00C000H   00C1FFH   000200H   DATA  REL   WORD   ---    1  PUBL  NDATA   ?C_USERSTACK
00FA00H   00FBFFH   000200H   ---   ---   ---    ---  ---  ---   * SYSTEM STACK *
00FC00H   00FC1FH   000020H   DATA  ---   BYTE   ---  ---  ---   *REG*   ?C_MAINREGISTERS

If you are a student then you can post your email and I will send you a project.

-Chris

Thank you , Chris Wunderlich.

Yes , I'm a student and I have to do a project to take my degree. My profesor gave me an eval board and have problems at the begining of the project.

Well, if you can give a project with proper settings , my email is
damian_dumitras@yahoo.com
(or, if it is restricted:
damian_dumitras
@..
yahoo
dot
com)

Thank you very much.

if(array[cnt] != 1)
                        putchar('!');

might be wrong.

No , it was correct. I made just a simple initialization, and tested if it had done...

No, the array is initialized with 1, not '1', so you should also check using 1 (not the ASCII value).

--
Joost

Sorry, maybe I didn't understand.
the array is initialized with 1, not '1'
Yes, it is...
you should also check using 1 (not the ASCII value).
So , i did.
But not this is the problem , the problem is that after the initialization unsigned char array[6]={1,1,1,1,1,1}; the elements of the array are not values of 1 like I expect, so it isn't initialized.
And I think the problem is in the project settings or startup file, where I don't know the solution...

I modified the project, except the startup file, and no changes have made...
What and how can I modify in the startup file?

I'll send you an example program in a few minutes.

But again check the value of the SYSCON0 register. You need to enable the XPEN bit.

-Chris