[ECOS] memory map question
Simpkins, Andy
Andy.Simpkins@Transcomm.uk.com
Fri Aug 24 02:35:00 GMT 2001
Hi there,
I am trying to map some data to a specific address in FLASH on my custom
target board (ARM7TDMI core)
The data is defined as
unsigned char globalArchive[] __attribute__ ((section (".ivoryDataStore")))
=
{0x04, 0x18, 0x00, 0x00, 0x81);
I have added the following to arm.ld
#define SECTION_ivoryDataStore(_region_, _vma_, _lma_) \
.ivoryDataStore _vma_ : _lma_ \
{ FORCE_OUTPUT; *(.ivoryDataStore) } \
> _region_
And have attempted to define ivoryDataStore in my memory map of the config
tool (1.3.1)
I kind of hoped that the name would have shown up in the drop down box for
Linker-defined, as it didn't I had to specify it as user defined...
i.e.
Name - User-defined: ivoryDataStore
- Known Size: 0x2000
Final Location (VMA) - Absolute: 0xe0000
however when I compile my code I get the following :
arm-elf-gcc *.o ../../API/*/*.o -L /i/eCOSbuilds/a166_install/lib/
-Ttarget.ld -nostdlib -Xlinker -Map -Xlinker mapfile.txt
/tools/H-i686-pc-cygwin/arm-elf/bin/ld: warning: no memory region specified
for
section `.ivoryDataStore'
and the following in mapfile.txt
snip
.gcc_except_table
0x000ab11c 0x0
0x000ab11c .=.
*(.gcc_except_table)
0x000e0000 __ivoryDataStore=0xe0000
0x41000000 __reserved_vectors=0x41000000
0x41001000 .=(__reserved_vectors+0x1000)
0x41001000 __reserved_not_mapped=0x41001000
0x41008000 .=(__reserved_not_mapped+0x7000)
snip
.ivoryDataStore
0x41008000 0x25bc
.ivoryDataStore
0x41008000 0x25bc ../../API/Ivory/globalArchive.o
0x41008000 globalArchive
0x000ab11c __rom_data_start=LOADADDR(.data)
0x4100a5bc __ram_data_end=.
0x4100a5bc PROVIDE (__ram_data_end, .)
0x4100a5bc _edata=.
0x4100a5bc PROVIDE (edata, .)
0x000ac0d8 PROVIDE (__rom_data_end,
(LOADADDR(.data)+SIZEOF(.data)))
.bss 0x4100a5bc 0x59eb8
What I wanted is the memory map to show
0x000e0000 0x25bc ../../API/Ivory/globalArchive.o
0x000e0000 globalArchive
Would someone be kind enough to show me what I am doing wrong?
Regards,
Andy
More information about the Ecos-discuss
mailing list