[arm] Infrastructure for eabi unwinding tables

[Paul Brook]

On Friday 13 August 2004 08:47, Nick Clifton wrote:
> Hi Paul,
>
> > 2004-08-10  Paul Brook  <paul@codesourcery.com>
> >
> > bfd/
> > 	* elf32-arm.c (elf32_arm_get_roseg_base): New function.
> > 	(elf32_arm_get_rwseg_base): New function.
> > 	(elf32_arm_final_link_relocate): Handle R_ARM_SBREL32 and
> > 	R_ARM_ROSEGREL32.
> > 	(elf32_arm_fake_sections, is_unwind_section_name): New functions.
> > 	(elf32_arm_section_from_shdr): New function.
> > 	(elf_backend_fake_sections, elf_backend_section_from_shdr): Define.
> > binutils/
> > 	* readelf.c (get_arm_section_type_name): New function.
> > 	(get_section_type_name): Use it.
> > gas/
> > 	* config/tc-arm.c (arm_elf_change_section): Set elf_linked_to_section.
> > include/elf/
> > 	* arm.h (SHT_ARM_EXIDX): Define.
> > 	(ELF_STRING_ARM_unwind*): Define.
> > ld/
> > 	* emulparams/armelf.sh: Define OTHER_READONLY_SECTIONS and
> > 	TEXT_START_SYMBOLS.
> > ld/testsuite/
> > 	* ld-arm/arm-app-abs32.d: Disassembler may output __text_start.
> > 	* ld-arm/arm-app.d: Disassembler may output __text_start.
> > 	* ld-arm/arm-static-app.d: Disassembler may output __text_start.
>
> Approved - please apply.

It turns out that the eabi exception handling specification is likely to be 
changed in the very near future. In particular there are issues with the 
handling of section relative offsets (see the ??? comments above 
elf32_arm_get_roseg_base).

For this reason I withdraw my patch. I'll submit a new once once the new 
specification has been finalised.

Paul