Ian Lance Taylor ef9beddf72 Handle output sections with more than 0x7fffffff bytes.
* object.h (class Relobj): Change map_to_output_ to
	output_sections_, and just keep a section pointer.  Change all
	uses.  Move comdat group support to Sized_relobj.
	(Relobj::is_section_specially_mapped): Remove.
	(Relobj::output_section): Remove poff parameter.  Change all
	callers.
	(Relobj::output_section_offset): New function.
	(Relobj::set_section_offset): Rewrite.
	(Relobj::map_to_output): Remove.
	(Relobj::output_sections): New function.
	(Relobj::do_output_section_offset): New pure virtual function.
	(Relobj::do_set_section_offset): Likewise.
	(class Sized_relobj): Add section_offsets_ field.  Add comdat
	group support from Relobj.  Update declarations.
	(Sized_relobj::get_output_section_offset): New function.
	(Sized_relobj::do_output_section_offset): New function.
	(Sized_relobj::do_set_section_offset): New function.
	* object.cc (Relobj::output_section_address): Remove.
	(Sized_relobj::Sized_relobj): Initialize new fields.
	(Sized_relobj::include_section_group): Cast find_kept_object to
	Sized_relobj.
	(Sized_relobj::include_linkonce_section): Likewise.
	(Sized_relobj::do_layout): Use separate arrays for output section
	and output offset.
	(Sized_relobj::do_count_local_symbols): Change map_to_output to
	output_sections.
	(Sized_relobj::do_finalize_local_symbols): Change map_to_output to
	output_sections and section_offsets.
	(Sized_relobj::write_local_symbols): Likewise.
	(map_to_kept_section): Compute output address directly.
	* reloc.cc (Sized_relobj::do_read_relocs): Change map_to_output to
	output_sections and section_offsets.
	(Sized_relobj::write_sections): Likewise.
	(Sized_relobj::relocate_sections): Likewise.
	* symtab.cc (sized_finalize_symbol): Use output_section_offset.
	* output.h (class Output_reloc): Update declarations.  Change
	u2_.relobj to Sized_relobj*.
	(class Output_data_reloc): Change add functions to use
	Sized_relobj*.
	* output.cc (Output_reloc::Output_reloc): Change relobj to
	Sized_relobj*.
	(Output_reloc::local_section_offset): Change return type to
	Elf_Addr.  Use get_output_section_offset.
	(Output_reloc::get_address): Likewise.
	(Output_section::is_input_address_mapped): Don't call
	is_section_specially_mapped.
	(Output_section::output_offset): Likewise.
	(Output_section::output_address): Likewise.
	(Output_section::starting_output_address): Likewise.
	* copy-relocs.cc (Copy_relocs::copy_reloc): Change object
	parameter to Sized_relobj*.
	(Copy_relocs::need_copy_reloc): Likewise.
	(Copy_relocs::save): Likewise.
	* copy-relocs.h (class Copy_relocs): Update declarations.
	(class Copy_relocs::Copy_reloc_entry): Change constructor to use
	Sized_relobj*.  Change relobj_ field to Sized_relobj*.
	* target-reloc.h (relocate_for_relocatable): Change
	offset_in_output_section type to Elf_Addr.  Change code that uses
	it as well.
	* layout.cc (Layout::layout): Always set *off.
	* mapfile.cc (Mapfile::print_input_section): Use
	output_section_offset.
	* i386.cc (Target_i386::copy_reloc): Change object parameter to
	Sized_relobj*.
	* powerpc.cc (Target_powerpc::copy_reloc): Likewise.
	* sparc.cc (Target_sparc::copy_reloc): Likewise.
	* x86_64.cc (Target_x86_64::copy_reloc): Likewise.
2008-07-10 23:01:20 +00:00
..
2008-04-15 05:16:51 +00:00
2008-04-23 16:11:47 +00:00
2008-05-21 21:37:44 +00:00
2008-06-05 17:29:44 +00:00
2008-05-21 21:37:44 +00:00
2008-05-21 21:37:44 +00:00
2008-03-26 23:36:46 +00:00
2008-05-28 20:48:16 +00:00
2008-05-28 20:48:16 +00:00
2008-05-21 21:37:44 +00:00
2008-05-21 21:37:44 +00:00
2008-05-21 21:37:44 +00:00
2008-05-28 20:48:16 +00:00
2008-05-21 21:37:44 +00:00
2008-05-21 21:37:44 +00:00
2008-03-25 21:55:29 +00:00
2008-05-28 20:48:16 +00:00
2008-05-28 20:48:16 +00:00
2008-05-21 21:37:44 +00:00
2008-05-21 21:37:44 +00:00
2008-05-28 20:48:16 +00:00
2008-05-28 20:48:16 +00:00
2008-04-16 22:54:29 +00:00
2008-04-16 22:54:29 +00:00
2008-05-21 21:37:44 +00:00
2008-05-21 21:37:44 +00:00
2008-03-26 23:36:46 +00:00
2008-04-03 00:33:37 +00:00

gold is an ELF linker.  It is intended to have complete support for
ELF and to run as fast as possible on modern systems.  For normal use
it is a drop-in replacement for the older GNU linker.

gold is part of the GNU binutils.  See ../binutils/README for more
general notes, including where to send bug reports.

gold was originally developed at Google, and was contributed to the
Free Software Foundation in March 2008.  At Google it was designed by
Ian Lance Taylor, with major contributions by Cary Coutant, Craig
Silverstein, and Andrew Chatham.

The existing GNU linker manual is intended to be accurate
documentation for features which gold supports.  gold supports most of
the features of the GNU linker for ELF targets.  Notable
omissions--features of the GNU linker not currently supported in
gold--are:
  * MEMORY regions in linker scripts
  * MRI compatible linker scripts
  * linker map files (-M, -Map)
  * cross-reference reports (--cref)
  * linker garbage collection (--gc-sections)
  * position independent executables (-pie)
  * various other minor options


Notes on the code
=================

These are some notes which may be helpful to people working on the
source code of gold itself.

gold is written in C++.  It is a GNU program, and therefore follows
the GNU formatting standards as modified for C++.  Source documents in
order of decreasing precedence:
    http://www.gnu.org/prep/standards/
    http://gcc.gnu.org/onlinedocs/libstdc++/17_intro/C++STYLE
    http://www.zembu.com/eng/procs/c++style.html

The linker is intended to have complete support for cross-compilation,
while still supporting the normal case of native linking as fast as
possible.  In order to do this, many classes are actually templates
whose parameter is the ELF file class (e.g., 32 bits or 64 bits).  The
C++ code is the same, but we don't pay the execution time cost of
always using 64-bit integers if the target is 32 bits.  Many of these
class templates also have an endianness parameter: true for
big-endian, false for little-endian.

The linker is multi-threaded.  The Task class represents a single unit
of work.  Task objects are stored on a single Workqueue object.  Tasks
communicate via Task_token objects.  Task_token objects are only
manipulated while holding the master Workqueue lock.  Relatively few
mutexes are used.