Re: objcopy and s-record output

["Ryan Bedwell" <ra9407 at email dot sps dot mot dot com>]

It's definitely possible that I'm doing something wrong.  I'm using a
custom linker script as well; the SECTIONS part looks like this:

<SNIP>

SECTIONS
{
  .start :
    {
      start_s = . ;
      *(.start) ;
      start_e = . ;
    } > rom
  .text : 
    AT (ADDR(.start) + SIZEOF(.start))
    { 	
      text_s = . ;
      *(.text) 
      *(.rodata) ;
      text_e = . ;
    } > ram
    text_l = LOADADDR(.text) ;
  .data :
    AT (LOADADDR(.text) + SIZEOF(.text))
    {
      data_s = . ;
      *(.data) 
      *(.got) ; 
      data_e = . ;
    } > ram
    data_l = LOADADDR(.data) ;
  .bss (NOLOAD) :
    { 
      bss_s = . ;
      *(.bss) 
      *(.COMMON) ; 
      bss_e = . ;
    } > ram
}

</SNIP>

Specifically what I'm trying to do is relocate the text section to RAM
at run time, but I need the startup and relocating code in ROM (hence
the extra section).

Ryan

Art Berggreen wrote:
> 
> Ryan Bedwell wrote:
> > I'm using objcopy to convert an ELF file to s-record, and it looks like
> > whenever I add a new section with a custom name -- i.e. ".start", it
> > does not make it through to the output file.  I know that you can lose
> > information in objcopy, but I haven't been able to locate documentation
> > which says anything about custom section restrictions in s-records.  I
> > can definitely see the sections set up correctly in the map file
> > produced by ld (i.e. they're there in the ELF file), but the custom one
> > seems to be gone in the s-record after running objcopy.
> >
> > I thought I'd ping the list to see if there is an obvious solution or
> > explanation before digging in myself.
> 
> Hmm, no explicit information to help you, just a little more input.
> 
> We regularly use ld to produce COFF and ELF load files, and frequently
> use objcopy to produce S-record files for other tools/utilities.  Most
> of these load files have custom output sections (we use our own linker
> scripts) that need to end up in the S-record file.  There isn't anything
> special I can think of that we are doing that makes this work.
> 
> Any chance that the missing output sections are acquiring the "NOLOAD"
> attribute like the .bss section?  Such sections don't get into the
> S-record file.
> 
> Art
> 
> ------
> Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
> Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com