DJ Delorie [Thu, 14 Jan 2010 04:56:12 +0000 (04:56 +0000)]
* emultempl/elf32.em (_place_orphan): If an input section doesn't
match an existing output section, but an unused output section
statement does match, use it.
* emultempl/pe.em (_place_orphan): Likewise.
* emultempl/pep.em (_place_orphan): Likewise.
Doug Kwan [Thu, 14 Jan 2010 02:42:20 +0000 (02:42 +0000)]
2010-01-13 Doug Kwan <dougkwan@google.com>
* arm.cc (Arm_relobj::section_needs_reloc_stub_scanning,
Arm_relobj::scan_sections_for_stubs): Rearrange code to avoid an
apparent compiler problem of not folding static constant integral
data members of elfcpp::Elf_sizes<32>.
Chao-ying Fu [Wed, 13 Jan 2010 22:34:28 +0000 (22:34 +0000)]
2010-01-13 Chao-ying Fu <fu@mips.com>
* ld-mips-elf/jr-to-b-1.d, ld-mips-elf/jr-to-b-2.d: New tests.
* ld-mips-elf/jr-to-b-1.s, ld-mips-elf/jr-to-b-2.s: Source.
* ld-mips-elf/mips-elf.exp: Run new tests.
Chao-ying Fu [Wed, 13 Jan 2010 22:30:53 +0000 (22:30 +0000)]
2010-01-13 Chao-ying Fu <fu@mips.com>
* elfxx-mips.c (JR_TO_B_P): New define to transform JR to B.
It is true for all CPUs.
(jal_reloc_p): New function.
(mips_elf_calculate_relocation): Rename require_jalxp to
cross_mode_jump_p.
Update comment for CROSS_MODE_JUMP_P.
Set up cross_mode_jump_p based on the mode change.
(mips_elf_perform_relocation): Rename require_jalx to cross_mode_jump_p.
Update comment for CROSS_MODE_JUMP_P.
Test cross_mode_jump_p and jal_reloc_p to turn jal to jalx.
Use !cross_mode_jump_p to guard conversion.
Convert "jr t9" to "b", if possible.
(_bfd_mips_elf_relocate_section): Rename require_jalx to
cross_mode_jump_p.
Pass &cross_mode_jump_p to call mips_elf_calculate_relocation.
Pass cross_mode_jump_p to call mips_elf_perform_relocation.
Doug Kwan [Wed, 13 Jan 2010 21:36:47 +0000 (21:36 +0000)]
2010-01-13 Doug Kwan <dougkwan@google.com>
* arm.cc (Arm_relobj::section_needs_reloc_stub_scanning,
Arm_relobj::section_needs_cortex_a8_stub_scanning,
Arm_relobj::scan_section_for_cortex_a8_erratum,
Arm_relobj::scan_span_for_cortex_a8_erratum): New methods.
(Arm_relobj::scan_sections_for_stubs): Move code deciding what
sections to scan for relocation stubs into a new method
Arm_relobj::section_needs_reloc_stub_scanning. Handle both
relocation and Cortex-A8 stub scanning.
(Target_arm::do_relax): Force stubs to be after stubbed sections
if fixing the Cortex-A8 erratum. Remove all Cortex-A8 stubs at
the beginning of a new relaxation pass. Update a comment.
(Target_arm::scan_span_for_cortex_a8_erratum): New method definition.
gas/
* config/tc-arm.c (do_t_nop): Correct check for Thumb-2 NOP.
gas/testsuite/
* gas/arm/thumb-nop.d, gas/arm/thumb-nop.s: New test.
* gas/arm/relax_branch_align.d: Expect a default NOP instruction.
* gas/arm/vfp1_t2.d, gas/arm/vfp1xD_t2.d: Specify a CPU with
Thumb-2.
ld/testsuite/
* ld-arm/arm-elf.exp (armelftests): Assemble Cortex-A8 tests with
-mcpu=cortex-a8.
Nick Clifton [Wed, 13 Jan 2010 14:34:13 +0000 (14:34 +0000)]
* cpu-m32c.c (m32c_scan): New function. Ensures that a scan for
"m32c" returns the m32c arch_info_struct and not the m16c
arch_info_struct.
(arch_info_struct): Use the new scan function.
(bfd_m32c_arch): Likewise.
Nick Clifton [Wed, 13 Jan 2010 14:08:54 +0000 (14:08 +0000)]
* config/tc-h8300.c (h8300_elf_section): New function - issue a
warning message if a new section is created without setting any
attributes for it.
(md_pseudo_table): Intercept section creation pseudos.
(md_pcrel_from): Replace abort with an error message.
* config/obj-elf.c (obj_elf_section_name): Export this function.
* config/obj-elf.h (obj_elf_section_name): Prototype.
* gas/elf/section0.d: Skip this test for the h8300.
* gas/elf/section1.d: Likewise.
* gas/elf/section6.d: Likewise.
* gas/elf/elf.exp: Skip section2 and section5 tests when the
target is the h8300.
* ld-scrips/sort.exp: Skip these tests when the target is the
h8300.
Doug Kwan [Wed, 13 Jan 2010 06:44:01 +0000 (06:44 +0000)]
2010-01-12 Doug Kwan <dougkwan@google.com>
* arm.cc (Cortex_a8_reloc): New class.
(Target_arm::Target_arm): Initialize new data members fix_cortex_a8_
and cortex_a8_relocs_info_.
(Target_arm::fix_cortex_a8): New method definition.
(Target_arm::Cortex_a8_relocs_info): New type.
(Target_arm::fix_cortex_a8_, Target_arm::cortex_a8_relocs_info_):
New data member declarations.
(Target_arm::scan_reloc_for_stub): Record information about
relocations for THUMB branches that might be exempted from the
Cortex-A8 workaround.
(Target_arm::do_relax): Clear all Cortex-A8 relocation information
at the beginning of a relaxation pass.
Doug Kwan [Wed, 13 Jan 2010 02:00:39 +0000 (02:00 +0000)]
2010-01-12 Doug Kwan <dougkwan@google.com>
* arm.cc (Arm_relobj::mapping_symbols_info_): New data member.
(Arm_relobj::Mapping_symbol_position,
Arm_reloj::Mapping_symbol_position_less,
Arm_relobj::Mapping_symbols_info): New types.
(Target_arm::is_mapping_symbol_name): New method definition.
(Arm_relobj::do_count_local_symbols): Save information about mapping
symbols.
Doug Kwan [Tue, 12 Jan 2010 07:54:37 +0000 (07:54 +0000)]
2010-01-11 Doug Kwan <dougkwan@google.com>
* arm.cc (Arm_relocate_functions::thumb32_branch_offset,
Arm_relocate_functions::thumb32_branch_upper,
Arm_relocate_functions::thumb32_branch_lower,
Arm_relocate_functions::thumb32_cond_branch_offset,
Arm_relocate_functions::thumb32_cond_branch_upper,
Arm_relocate_functions::thumb32_cond_branch_lower,
Arm_relocate_functions::thm_jump19): New methods to handle THUMB
branch offset encoding.
(Arm_relocate_functions::thumb_branch_common): Use new branch
offset encoding methods to avoid code duplication.
(Stub_template::Stub_template) Handle THUMB16_SPECIAL_TYPE.
(Stub_addend_reader::operator()): Use new branch encoding method
to avoid code duplication.
Doug Kwan [Tue, 12 Jan 2010 07:22:56 +0000 (07:22 +0000)]
2010-01-11 Doug Kwan <dougkwan@google.com>
* arm.cc (Arm_relobj::do_gc_process_relocs): New method.
(Target_arm::do_finalize_sections): Define special EXIDX section
symbols only if referenced.
* gc.h (Garbage_collection::add_reference): New method.
(gc_process_relocs): Use Garbage_collection::add_reference to avoid
code duplication.
Ian Lance Taylor [Tue, 12 Jan 2010 06:41:36 +0000 (06:41 +0000)]
* script.cc (class Lazy_demangler): Recreate--revert part of patch
of 2009-12-30.
(Version_script_info::Version_script_info): Initialize globs_,
default_version_, default_is_global_, and exact_. Don't
initialize globals_ or locals_.
(Version_script_info::build_lookup_tables): Build local symbols
first.
(Version_script_info::unquote): New function.
(Version_script_info::add_exact_match): New function.
(Version_script_info::build_expression_list_lookup): Remove lookup
parameter. Add is_global parameter. Change all callers. Handle
wildcard pattern specially. Unquote pattern. Call
add_exact_match.
(Version_script_info::get_name_to_match): New function.
(Version_script_info::get_symbol_version): New function.
(Version_script_info::get_symbol_version_helper): Remove.
(Version_script_info::check_unmatched_names): Call unquote.
* script.h (class Version_script_info): Change get_symbol_version
to be non-inline and add is_global parameter; change all callers.
Rewrite symbol_is_local. Update declarations. Define struct
Version_tree_match, Exact, Globs. Don't define struct Lookup.
Remove globals_ and locals_ members. Add exact_, globs_,
default_version_, is_global_.
(Version_script_info::Glob): Remove pattern, add expression and
is_global. Update constructor. Change all callers.
* dynobj.cc (Versions::finalize): Mark the version symbol as the
default version.
(Versions::symbol_section_contents): If a symbol is undefined, or
defined in a dynamic object, set the version index to
VER_NDX_LOCAL.
* symtab.cc (Symbol_table::add_from_relobj): Don't call
symbol_is_local.
(Symbol_table::add_from_pluginobj): Likewise.
* testsuite/ver_matching_test.sh: blaza1 and blaza go into V2.
Joel Brobecker [Tue, 12 Jan 2010 05:49:00 +0000 (05:49 +0000)]
gdb/ChangeLog:
Add support for DW_AT_GNAT_descriptive_type.
* gdbtypes.h (enum type_specific_kind): New enum.
(struct main_type) [type_specific_field]: New component.
[type_specific]: Add new component "gnat_stuff".
(struct gnat_aux_type): New type.
(INIT_CPLUS_SPECIFIC): Also set TYPE_SPECIFIC_FIELD (type).
(HAVE_CPLUS_STRUCT): Also check TYPE_SPECIFIC_FIELD (type).
(gnat_aux_default, allocate_gnat_aux_type): Add declaration.
(INIT_GNAT_SPECIFIC, ALLOCATE_GNAT_AUX_TYPE, HAVE_GNAT_AUX_INFO)
(TYPE_SPECIFIC_FIELD): New macros.
(TYPE_CPLUS_SPECIFIC): Return cplus_struct_default if the given
type does not hold any cplus-specific data.
(TYPE_RAW_CPLUS_SPECIFIC): New macro.
(TYPE_GNAT_SPECIFIC, TYPE_DESCRIPTIVE_TYPE): New macros.
(TYPE_IS_OPAQUE): Use HAVE_CPLUS_STRUCT to check if type has
cplus-specific data.
* gdbtypes.c (allocate_cplus_struct_type): Minor stylistic rewrite.
Set new component TYPE_SPECIFIC_FIELD (type).
(gnat_aux_default): New constant.
(allocate_gnat_aux_type): New function.
(init_type): Add initialization the type-specific stuff for
TYPE_CODE_FLT and TYPE_CODE_FUNC types.
(print_gnat_stuff): New function.
(recursive_dump_type): Use HAVE_CPLUS_STRUCT to check for cplus-
specific data. Adjust code that prints the contents of the
type-specific union using the TYPE_SPECIFIC_FIELD value.
* dwarf2read.c (dwarf2_attach_fields_to_type): Do not allocate
the type cplus stuff for Ada types.
(dwarf2_add_member_fn, dwarf2_attach_fn_fields_to_type):
Error out if these routines are called with an Ada type.
(read_structure_type, read_array_type, read_subrange_type):
Add call to set_descriptive_type.
(set_die_type): Initialize the gnat-specific data if necessary.
(need_gnat_info, die_descriptive_type, set_descriptive_type):
New functions.
* ada-lang.c (decode_constrained_packed_array_type): Use
decode_constrained_packed_array_type instead of doing a standard
lookup to locate a parallel type.
(find_parallel_type_by_descriptive_type): New function.
(ada_find_parallel_type_with_name): New function.
(ada_find_parallel_type): Reimplement using
ada_find_parallel_type_with_name.
* ada-valprint.c (print_field_values): Use HAVE_CPLUS_STRUCT
to check if type has a cplus stuff.
* linespec.c (total_number_of_methods): Likewise.
* mdebugread.c (new_type): Likewise.
gdb/testsuite/ChangeLog:
* gdb.base/maint.exp: Adjust the expected output for the
"maint print type" test. Use gdb_test_multiple instead of
gdb_sent/gdb_expect.
Jan Kratochvil [Mon, 11 Jan 2010 18:15:05 +0000 (18:15 +0000)]
gdb/
* objfiles.c (objfile_relocate1): Change the return type to int.
Describe the new return value. Return non-zero if data changed.
(objfile_relocate): New variable changed. Set it. Call
breakpoint_re_set depending on CHANGED.
Tristan Gingold [Mon, 11 Jan 2010 15:39:41 +0000 (15:39 +0000)]
2010-01-11 Tristan Gingold <gingold@adacore.com>
* mach-o.h (bfd_mach_o_backend_data): Add arch field.
(bfd_mach_o_set_arch_mach): New prototype.
* mach-o.c (bfd_mach_o_mkobject): Define with bfd_mach_o_gen_mkobject.
(bfd_mach_o_set_arch_mach): New function.
(bfd_mach_o_gen_mkobject): New function.
Set TARGET_ARCHITECTURE for the generic back-ends.
* mach-o-target.c (bfd_mach_o_set_arch_mach): Remove define.
Check that TARGET_ARCHITECTURE is defined.
Add TARGET_ARCHITECTURE in TARGET_NAME_BACKEND structure.
* mach-o-i386.c (TARGET_ARCHITECTURE): Define.
Jan Kratochvil [Sat, 9 Jan 2010 09:11:00 +0000 (09:11 +0000)]
gdb/
Fix displacement of separate debug info files.
* objfiles.c (objfile_relocate): Rename to ...
(objfile_relocate1): ... here and make it static. Extend the comment.
(objfile_relocate): New function.
* solib-spu.c (spu_relocate_main_executable): Explicitly check if
SYMFILE_OBJFILE is NULL. Remove variables objfile and old_chain.
Remove following of SEPARATE_DEBUG_OBJFILE. new_offsets is now
allocated using alloca.
* symfile.c (copy_section_addr_info): Remove.
(build_section_addr_info_from_objfile): Make it global. New variables
addr_bit and mask, use them.
* symfile.h (build_section_addr_info_from_objfile): New prototype.
(copy_section_addr_info): Remove.
* script.cc (Version_script_info::build_expression_list_lookup):
Don't warn about ambiguous version, just record the ambiguity.
(Version_script_info::get_symbol_version_helper): Give error if
version is ambiguous.
Joel Brobecker [Sat, 9 Jan 2010 04:44:55 +0000 (04:44 +0000)]
Signal unwinder for mips-irix N32.
* mips-irix-tdep.c: #include mips-tdep.h, trad-frame.h, and
tramp-frame.h.
(SIGTRAMP_FRAME_SIZE, SIGCONTEXT_PC_OFF, SIGCONTEXT_REGS_OFF)
(SIGCONTEXT_FPREGS_OFF, SIGCONTEXT_FPCSR_OFF, SIGCONTEXT_HI_OFF)
(SIGCONTEXT_LO_OFF): New macros.
(mips_irix_n32_tramp_frame_init): New function.
(mips_irix_n32_tramp_frame): New static constant.
(mips_irix_init_abi): Prepend the mips_irix_n32_tramp_frame unwinder.
Joel Brobecker [Sat, 9 Jan 2010 04:44:02 +0000 (04:44 +0000)]
Breakpoint in shared library does not work on mips-irix.
* procfs.c: #include "observer.h".
(procfs_inferior_created): New function, moving here the code
which unsets the syssgi syscall-exit notifications.
(procfs_create_inferior): Remove the code which unsets the syssgi
syscall-exit notifications. It is too early to do this here.
(_initialize_procfs): Attach the procfs_inferior_created observer.
Joel Brobecker [Sat, 9 Jan 2010 04:42:54 +0000 (04:42 +0000)]
Wrong return convention for arrays (mips-irix).
* mips-tdep.c (mips_n32n64_return_value): Arrays whose size is
128 bits or smaller are returned the same way as structs
and unions of the the same size.
Joel Brobecker [Sat, 9 Jan 2010 04:38:29 +0000 (04:38 +0000)]
* mips-tdep.c (mips_breakpoint_from_pc): Add new static variable
containing the correct breakpoint instruction to use on mips-irix.
Use it when the osabi is GDB_OSABI_IRIX.
Doug Kwan [Sat, 9 Jan 2010 01:55:14 +0000 (01:55 +0000)]
2010-01-08 Doug Kwan <dougkwan@google.com>
* arm.cc (Stub_table::Stub_table): Initalize cortex_a8_stubs_,
prev_data_size_ and prev_addralign_. Remove initializer for
deleted data member has_been_changed_.
(Stub_table::empty): Look at both reloc_stubs_ and cortex_a8_stubs_
to determine if the table is empty.
(Stub_table::has_been_changed, Stub_table_set_has_been_changed):
Remove.
(Stub_table::add_reloc_stub): Define method in class definition
instead of just declaring it there.
(Stub_table::add_cortex_a8_stub): New method definition.
(Stub_table::update_data_size_and_addralign): Ditto.
(Stub_table::finalize_stubs): Ditto.
(Stub_table::apply_cortex_a8_workaround_to_address_range): Ditto.
(Stub_table::do_addralign_): Return address alignment in the
(Stub_table::do_reset_address_and_file_offset): Define method in
class definition instead of declaring it there. Set current data
size to be the data size of the previous pass.
(Stub_table::set_final_data_size): Use current data size as the
final data size.
(Stub_table::relocate_stub): Change parameter type of stub from
Reloc_stub pointer to Stub pointer.
(Stub_table::addralign_, Stub_table::has_been_changed_): Remove.
(Stub_table::Cortex_a8_stub_list): New typedef.
(Stub_table::cortex_a8_stubs_, Stub_table::prev_data_size_,
Stub_table::prev_addralign_): New data member.
(Arm_relobj::Arm_relobj): Initialize data member
section_has_cortex_a8_workaround_.
(Arm_relobj::section_has_cortex_a8_workaround,
Arm_relobj::mark_section_for_cortex_a8_workaround): New method
definitions.
(Arm_relobj::section_has_cortex_a8_workaround_): New data member
declarations.
(Target_arm::relocate_stub): Change parameter type of stub from
Reloc_stub pointer to Stub pointer.
(Insn_template::size, Insn_template::alignment): Handle
THUMB16_SPECIAL_TYPE.
(Stub_table::remove_all_cortex_a8_stubs, Stub_table::finalize_stubs,
Stub_table::update_data_size_and_addralign,
Stub_table::apply_cortex_a8_workaround_to_address_range): New method
definitions.
(Stub_table::relocate_stubs): Handle Cortex-A8 stubs.
(Stub_table::do_write): Ditto.
(Target_arm::do_relax): Adjust code for changes in Stub_table.
Jan Kratochvil [Sat, 9 Jan 2010 00:14:11 +0000 (00:14 +0000)]
gdb/testsuite/
* gdb.base/corefile.exp: Move the core finding block out and call it as
core_find, new variable $corefile, replace corefile by $corefile and
[file tail $corefile] for usage vs. test names resp.
* lib/gdb.exp (core_find): Move it as a new function here. New
parameter binfile and deletefiles. New variable $destcore. Pre-delete
$destcore. Return "" on error.
PR 11108
* symtab.h (class Symbol): Remove fields is_target_special_ and
has_plt_offset_. Add field is_defined_in_discarded_section_.
(Symbol::is_defined_in_discarded_section): New function.
(Symbol::set_is_defined_in_discarded_section): New function.
(Symbol::has_plt_offset): Rewrite.
(Symbol::set_plt_offset): Verify that new offset is not -1U.
* symtab.cc (Symbol::init_fields): Initialize plt_offset_ to -1U.
Don't initialize is_target_special_ or has_plt_offset_.
Initialize is_defined_in_discarded_section_.
(Symbol_table::add_from_relobj): If appropriate, set
is_defined_in_discarded_section.
* resolve.cc (Symbol::override_base_with_special): Don't test
is_target_special_. Change has_plt_offset_ to has_plt_offset().
* target-reloc.h (relocate_section): Do special handling for
symbols defined in discarded sections for global symbols as well
as local symbols.
Jan Kratochvil [Fri, 8 Jan 2010 22:55:15 +0000 (22:55 +0000)]
gdb/
Move some symfile code into subroutines.
* symfile.h (relative_addr_info_to_section_offsets)
(addr_info_make_relative): New prototypes.
* symfile.c (default_symfile_offsets): Move a part to ...
(relative_addr_info_to_section_offsets): ... this new function.
(default_symfile_offsets): Call it.
(syms_from_objfile <!mainline && addrs && addrs->other[0].name>): Move
this part to ...
(addr_info_make_relative): ... this new function.
PR 11084
* dynobj.cc (Sized_dynobj::find_dynsym_sections): If there is no
dynamic symbol table, use the normal symbol table.
(Sized_dynobj::do_read_symbols): Remove assertion about type of
symbol table.