[En-Nut-Discussion] PLEASE HELP - NEED WORKING THREAD SAMPLE FOR OWN ATMEGA128 PROJECT

staste at alpikom.it staste at alpikom.it
Thu Jan 27 12:56:17 CET 2005


hi

yes i use the configurator - and i have disabled external ram...

but in the file include/cfg/memory.h (in my build tree) i will see some entry
for NUTXMEM...

for the value in:
"Please look into threads.map and tell me, what hex value
is in front of
   PROVIDE (__heap_start, .)
and
   __bss_end

Here it is 0x02000424 for both values."



in my file the value is 0x008001e5



thx
Stephan,

I recall, that you're using the Configurator, right?
Make sure to disable "Extended Memory Size" in
"Memory Management". When creating the build tree,
include/cfg/memory.h (in your build tree) should
contain

--- snip ---
#ifndef NUTMEM_SIZE
#define NUTMEM_SIZE 4096
#endif

#ifndef NUTMEM_START
#define NUTMEM_START 0x100
#endif

#ifndef NUTMEM_RESERVED
#define NUTMEM_RESERVED 64
#endif
--- snap ---

but no entry beginning with NUTXMEM...

In the 3.9.2 source tree the file os/arch/avr_nutinit.c
contains (uses cfg/memory.h from build, not from source tree)

--- snip ---
#include <cfg/memory.h>

#ifdef NUTXMEM_SIZE
/*!
  * \brief Last memory address using external SRAM.
  */
#define NUTMEM_END (u_short)(NUTXMEM_START + (u_short)NUTXMEM_SIZE - 1U)

#else
/*!
  * \brief Last memory address without using external SRAM.
  *
  * \todo Shall we support NUTRAMEND for backward compatibility? If, then
  *       let's do it in cfg/memory.h.
  */
#define NUTMEM_END (u_short)(NUTMEM_START + (u_short)NUTMEM_SIZE - 1U)

#endif
--- snap ---

and further down

--- snip ---
if ((u_short)NUTMEM_END - (u_short) (&__heap_start) > 384) {
     NutHeapAdd(&__heap_start, (u_short) NUTMEM_END - 256 - (u_short) 
(&__heap_start));
}
--- snap ---

which results in something similar, except __heap_start is used
instead of __bss_end. Can't remember right now why this had
changed, but indeed __heap_start should be correct.

Please look into threads.map and tell me, what hex value
is in front of
   PROVIDE (__heap_start, .)
and
   __bss_end

Here it is 0x02000424 for both values.

Harald

At 12:40 25.01.2005 +0100, you wrote:
>hi
>
>i use a own atmega128 project with fixed freq (16MHz), no 32kHz crystal and no
>external ram and no lan interface
>
>so i only need the nut/os
>
>i use nut392 and the latest winavr...
>
>i have compiled successfully the basemon.c and this example is working, if 
>i set
>the baudrate fix to 9600..
>
>then i have tested to make 2 threads (as in example threads.c, the first 
>writes
>H, the second writes L to the serial board).i excluded the basemon() function.
>on the serial port i get only the character from the first thread (H) and the
>nutSleep in the threadfunction seems to not working (NutSleep(125) and
>NutSleep(5000) has the same effect - always as fastet it can)...
>
>so my question: maybe i use some wrong software environment ???
>has anyone a thread sample for an atmega128 with no ext ram, fixed freq and no
>32kHz crystal...?
>
>and i need always to include this lines for winavr, because i see that without
>this lines the basemon is not working ??? :
>
>#ifdef __GNUC__
>void NutInit(void) __attribute__ ((naked)) __attribute__ 
>((section(".init8")));
>extern int NutAppMain(void) __attribute__ ((noreturn));
>#endif
>
>void NutInit(void)
>{
>     extern void *__bss_end;
>     NutHeapAdd(&__bss_end, (uptr_t) RAMEND - 256 - (uptr_t) (&__bss_end));
>
>
>thx stephan

    * Previous message: [En-Nut-Discussion] PLEASE HELP - NEED WORKING THREAD
SAMPLE FOR OWN ATMEGA128 PROJECT
    * Next message: [En-Nut-Discussion] compiling NutOS CVS Head with avr-gcc
failes on nutcpp.c
    * Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

More information about the En-Nut-Discussion mailing list


----------------------------------------------------
This mail has been sent using Alpikom webmail system
http://www.alpikom.it




More information about the En-Nut-Discussion mailing list