* event-loop.c: Include unistd.h if it exists.

* linux-low.c (my_waitpid): Move definition away from being in
between linux_tracefork_child/linux_test_for_tracefork.

* arm-linux-tdep.c (ARM_OABI_SYSCALL_RESTART_SYSCALL): New macro.
(ARM_LDR_PC_SP_12): Likewise.
(arm_linux_restart_syscall_init): New function.
(arm_linux_restart_syscall_tramp_frame): New variable.
(arm_linux_init_abi): Install the
arm_linux_restart_syscall_tramp_frame unwinder.

gdb/
* gdb_proc_service.h (psaddr_t): Fix type.
* linux-thread-db.c (thread_db_info.td_thr_tls_get_addr_p): Fix
signature to match glibc.
(thread_db_get_thread_local_address): Use psaddr_t for type of
address to match parameter of td_thr_tls_get_addr_p.
Cast through psaddr_t instead of (void*) to match parameter of
td_thr_tls_get_addr_p.

gdbserver/
* gdb_proc_service.h (psaddr_t): Fix type.
* thread-db.c (thread_db_info.td_thr_tls_get_addr_p): Fix
signature to match glibc.

* gas/mips/eret-2.s: Add an instruction to fill a branch delay
slot.
* gas/mips/eret-2.d: Adjust accordingly.

gas/
* config/tc-mips.c (s_mips_ent): Also set BSF_FUNCTION for
".aent".

gas/testsuite/
* gas/mips/aent.d: New test.
* gas/mips/aent.s: Source for the new test.
* gas/mips/mips.exp: Run it.

*** empty log message ***

daily update

Move Imm1 before Imm8.

2009-12-18  H.J. Lu  <hongjiu.lu@intel.com>

* i386-gen.c (operand_types): Move Imm1 before Imm8.

2009-12-18  Steve Ellcey  <sje@cup.hp.com>

* config/tc-hppa.c: Change access to access_ctr.

include/elf/
* common.h (NT_S390_HIGH_GPRS): Define.

bfd/
* elf.c (elfcore_grok_s390_high_gprs): New function.
(elfcore_grok_note): Handle NT_S390_HIGH_GPRS notes.
(elfcore_write_s390_high_gprs): New function.
(elfcore_write_register_note): Call it.

binutils/
* readelf.c (get_note_type): Handle NT_S390_HIGH_GPRS notes.

* config.sub, config.guess: Update from upstream sources.

2009-12-17  Stan Shebs  <stan@codesourcery.com>

* breakpoint.c (bpstat_check_location): Check for tracepoints.

*** empty log message ***

2009-12-16  Michael Snyder  <msnyder@vmware.com>

* i386-tdep.c (i386_process_record): Reformat comments.

daily update

2009-12-17  Jeff Johnston  <jjohnstn@redhat.com>

        * COPYING.NEWLIB: Update copyright date.
        * COPYING.LIBGLOSS: Ditto.

* pe-dll.c (generate_reloc): Take account of wrapper options when
testing if a weak symbol is defined or not, and when it is not,
consider whether the default value requires a base reloc anyway.

* deffilep.y (def_file_free): Add missing shadow parameter renaming.
(def_file_print): Likewise.
(def_stash_module): Likewise.
(def_file_add_import): Likewise.

2009-12-17  Rafael Avila de Espindola  <espindola@google.com>

* Makefile.am (CCFILES): Add timer.cc.
(HFILES): Add timer.h.
* configure.ac: Check for sysconf and times.
* main.cc: include timer.h.
(main): Use Timer instead of get_run_time.
* timer.cc: New.
* timer.h: New.
* workqueue.cc: include timer.h.
(Workqueue::find_and_run_task):
Report user, sys and wall time.
* Makefile.in: Regenerate.
* config.in: Regenerate.
* configure: Regenerate.

    PR binutils/10924
        * config/tc-arm.c (do_ldstv4): Do not allow r15 as the destination
        register.
        (do_mrs): Likewise.
        (do_mul): Likewise.

        * arm-dis.c: Add support for %<>ru and %<>rU formats to enforce
        unique register numbers.  Extend support for %<>R format to
        thumb32 and coprocessor instructions.

        * gas/arm/unpredictable.s: Add more unpredictable instructions.
        * gas/arm/unpredictable.d: Add expected disassemblies.

PR ld/11088
include/elf/
* ppc.h (R_PPC_RELAX32, R_PPC_RELAX32PC, R_PPC_RELAX32_PLT,
R_PPC_RELAX32PC_PLT): Delete.
(R_PPC_RELAX, R_PPC_RELAX_PLT, R_PPC_RELAX_PLTREL24): Define.
bfd/
* elf32-ppc.c (update_plt_info): Clear sec here when addend is
less than 32768..
(ppc_elf_check_relocs): ..rather than doing so here.  Ignore new
relax relocs.
(ppc_elf_gc_sweep_hook): Don't segfault when symbol hiding has
removed plt_entry records.
(ppc_elf_tls_setup): Handle PIE calls to __tls_get_addr correctly.
(ppc_elf_tls_optimize): Likewise.  Also dec __tls_get_addr refcount
when optimizing code using new tlsgd and tlsld marker relocs.
(ppc_elf_relax_section): Differentiate relaxed PLTREL24 relocs
from ADDR24 relocs using plt or glink.  Don't clear the addend
for R_PPC_RELAX_PLTREL24.
(ppc_elf_relocate_section): Correctly handle addends on relaxed
PLTREL24 relocs.

PR ld/11088
* elf64-ppc.c (ppc64_elf_gc_sweep_hook): Don't abort if symbol
hiding has nulled out plt.plist.

*** empty log message ***

2009-12-16  Doug Kwan  <dougkwan@google.com>

* arm.cc (Arm_relobj::scan_sections_for_stubs): Exclude ICF-eliminated
sections.
* object.cc (Sized_relobj::do_finalize_local_symbols): Handle
relaxed input sections.
* output.cc (Output_section::find_relaxed_input_section): Change
return type to Output_relaxed_input_section pointer.  Adjust code
for new type of relaxed_input_section_map_.
* output.h (Output_section::find_relaxed_input_section): Change
return type to Output_relaxed_input_section pointer.
(Output_section::Output_relaxed_input_section_by_input_section_map):
New type.
(Output_section::relaxed_input_section_map_): Change type to
Output_section::Output_relaxed_input_section_by_input_section_map.
* symtab.cc (Symbol_table::compute_final_value): Handle relaxed
input section.

daily update

* linux-low.c (linux_read_memory): Fix argument to read.

Remove ByteOkIntel.

gas/

2009-12-16  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (process_suffix): Set i.suffix to 0 in
Intel syntax if size is ignored and b/l/w suffixes are
illegal.
(check_byte_reg): Remove byteokintel check.

opcodes/

2009-12-16  H.J. Lu  <hongjiu.lu@intel.com>

* i386-gen.c (opcode_modifiers): Remove ByteOkIntel.

* i386-opc.h (ByteOkIntel): Removed.
(i386_opcode_modifier): Remove byteokintel.

* i386-opc.tbl: Remove ByteOkIntel.
* i386-tbl.h: Regenerated.

Replace Vex0F, Vex0F38, Vex0F3A, XOP08, XOP09 and XOP0A with VexOpcode.

gas/

2009-12-16  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (build_vex_prefix): Replace vex0f, vex0f38,
vex0f3a, xop08, xop09 and xop0a with vexopcode.

opcodes/

2009-12-16  H.J. Lu  <hongjiu.lu@intel.com>

* i386-gen.c (opcode_modifiers): Remove Vex0F, Vex0F38,
Vex0F3A, XOP08, XOP09 and XOP0A.  Add VexOpcode.

* i386-opc.h (Vex0F): Removed.
(Vex0F38): Likewise.
(Vex0F3A): Likewise.
(VexOpcode): New.
(VEX0F): Likewise.
(VEX0F38): Likewise.
(VEX0F3A): Likewise.
(XOP08): Defined as a macro.
(XOP09): Likewise.
(XOP0A): Likewise.
(i386_opcode_modifier): Remove vex0f, vex0f38, vex0f3a, xop08,
xop09 and xop0a.  Add vexopcode.

* i386-opc.tbl: Replace Vex0F with VexOpcode=0, Vex0F38 with
VexOpcode=1, Vex0F3A with VexOpcode=2, XOP08 with VexOpcode=3,
XOP09 with VexOpcode=4 and XOP0A with VexOpcode=5.
* i386-tbl.h: Regenerated.

* dwarf2read.c (read_subroutine_type): Add the subroutine type to the
die immediately to allow a parameter type to be the same subroutine type.

Fix a typo in ChangeLog.

Replace VEX2SOURCES with XOP2SOURCES.

gas/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (build_modrm_byte): Check XOP2SOURCES
instead VEX2SOURCES.

opcodes/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-opc.h (VEX2SOURCES): Renamed to ...
(XOP2SOURCES): This.

Replace Vex2Sources and Vex3Sources with VexSources.

gas/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (process_operands): Check vexsources
instead of vex3sources.
(build_modrm_byte): Check vexsources instead of vex2sources
and vex3sources.

opcodes/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-gen.c (opcode_modifiers): Remove Vex3Sources and
Vex2Sources.  Add VexSources.

* i386-opc.h ()Vex2Sources: Removed.
(Vex3Sources): Likewise.
(VEX2SOURCES): New.
(VEX3SOURCES): Likewise.
(VexSources): Likewise.
(i386_opcode_modifier): Remove vex2sources and vex3sources.
Add vexsources.

* i386-opc.tbl: Replace Vex2Sources with VexSources=1 and
Vex3Sourceswith VexSources=2.
* i386-tbl.h: Regenerated.

Remove VexW0 and VexW1.  Add VexW.

gas/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (build_vex_prefix): Replace vexw0/vexw1
with vexw.
(build_modrm_byte): Likewise.

opcodes/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-gen.c (opcode_modifiers): Remove VexW0 and VexW1.  Add
VexW.

* i386-opc.h (VexW0): Removed.
(VexW1): Likewise.
(VEXW0): New.
(VEXW1): Likewise.
(VexW): Likewise.
(i386_opcode_modifier): Remove vexw0 and vexw1.  Add vexw.

* i386-opc.tbl: Replace VexW0 with VexW=1 and VexW1 with
Vex=2.
* i386-init.h: Regenerated.
* i386-tbl.h: Likewise.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* as.h (mempcpy): New.

* configure.in: Check if mempcpy is declared.
* configure: Regenerated.
* config.in: Likewise.

*** empty log message ***

Add VEX_W_3818_P_2_M_0.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (VEX_W_3818_P_2_M_0): New.
(vex_w_table): Add VEX_W_3818_P_2_M_0.
(mod_table): Use VEX_W_3818_P_2_M_0.

daily update

Reformat vex_w_table.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (vex_w_table): Reformat.

* layout.cc (Layout::create_shstrtab): Only write out after input
sections if we are compressing debug sections.

Add VEX_W_382X_P_2_M_0.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (VEX_W_382X_P_2_M_0): New.
(vex_w_table): Add VEX_W_382X_P_2_M_0.
(mod_table): Use VEX_W_382X_P_2_M_0.

* archive.cc (Archive::add_symbols): Only look up a symbol without
a version if there is, in fact, a version.

Reformat vex_w_table.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (vex_w_table): Reformat.

Add USE_VEX_W_TABLE, VEX_W_TABLE and VEX_W_XXX.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (USE_VEX_W_TABLE): New.
(VEX_W_TABLE): Likewise.
(VEX_W_XXX): Likewise.
(vex_w_table): Likewise.
(prefix_table): Use VEX_W_XXX.
(vex_table): Likewise.
(vex_len_table): Likewise.
(mod_table): Likewise.
(get_valid_dis386): Handle USE_VEX_W_TABLE.

* i386-opc.tbl: Add VexW0 to AVX instructions where the VEX.W bit
isn't used.
* i386-tbl.h: Regenerated.

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* opncls.c (bfd_opnr_iovec): Replace _XXX with XXX_P in
parameters.

* tekhex.c (pass_over): Replace eof with is_eof.

Define VEX128 and VEX256.

gas/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (build_vex_prefix): Use VEX256.

opcodes/

2009-12-15  H.J. Lu  <hongjiu.lu@intel.com>

* i386-opc.h (VEX128): New.
(VEX256): Likewise.

2009-12-15  Tristan Gingold  <gingold@adacore.com>

* machoread.c (macho_symfile_read): Set section size of dsym bfd
from main bfd.

2009-12-15  Tristan Gingold  <gingold@adacore.com>

        * mach-o.h (bfd_mach_o_uuid_command): Remove section field.
        * mach-o.c (bfd_mach_o_scan_read_uuid): Do not create a section
        from this command.

2009-12-15  Tristan Gingold  <gingold@adacore.com>

        * machoread.c (macho_check_dsym): Read uuid directly from load command
        instead of from a fake section.

2009-12-15  Tristan Gingold  <gingold@adacore.com>

        * machoread.c (macho_symtab_read): Add comments.  Always set
        first_offset.

2009-12-15  Tristan Gingold  <gingold@adacore.com>

        * mach-o.c (struct mach_o_section_name_xlat): Add flags field.
        (dwarf_section_names_xlat): Add section flags.
        (text_section_names_xlat): Ditto.
        (data_section_names_xlat): Ditto.
        (bfd_mach_o_convert_section_name_to_bfd): Now return name and section
        flags by reference.
        (bfd_mach_o_make_bfd_section): Use section flags when know, otherwise
        try to guess.

2009-12-14  Doug Kwan  <dougkwan@google.com>

bfd/ChangeLog:

* opncls.c (bfd_opnr_iovec): Rename parameters to avoid shawdowed
variable warnings.
* bfd-in2.h: Regnenerate.

include/ChangeLog:

* bfdlink.h (struct bfd_link_callbacks): Rename function parameters
to avoid shadowed variable warnings.
* dis-asm.h (struct disassemble_info): Ditto.
(disassemble_init_for_target): Ditto.
(init_disassemble_info): Ditto.

Reformat vex_len_table.

2009-12-14  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (vex_len_table): Reformat.

*** empty log message ***

daily update

Rename MOD_VEX_51 to MOD_VEX_50.

2009-12-14  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (MOD_VEX_51): Renamed to ...
(MOD_VEX_50): This.
(vex_table): Updated.
(mod_table): Likewise.

* elfcpp_file.h: Revert last patch.

Revert -Wshadow changes, all changes from:
2009-12-11  Doug Kwan  <dougkwan@google.com>
2009-12-11  Nick Clifton  <nickc@redhat.com>
* configure.ac: Remove -Wshadow when setting WARN_CXXFLAGS.

    PR binutils/10924
        * arm-dis.c (arm_opcodes): Specify %R in cases where using r15
        results in unpredictable behaviour.
        (print_insn_arm): Handle %R.

        * gas/arm/unpredictable.s: New test case - checks the disassembly
        of instructions with unpredictable behaviour.
        * gas/arm/unpredictable.d: New file - expected disassembly.

2009-12-11  Tristan Gingold  <gingold@adacore.com>

* symfile.h (find_separate_debug_file_by_buildid): Remove prototype.
(bfd_open_maybe_remote): New prototype.
* symfile.c (bfd_open_maybe_remote): New function.
(separate_debug_file_exists, reread_symbols): Use it.
(struct build_id, build_id_bfd_get, build_id_verify)
(build_id_to_debug_filename)
(find_separate_debug_file_by_buildid): Move these ...
* elfread.c (struct build_id)
(build_id_bfd_get, build_id_verify, build_id_to_debug_filename)
(find_separate_debug_file_by_buildid): ... here.

Fix PR number typo.

    PR gas/11089
        * config/tc-rx.c (rx_equ): Rename 'expr' to 'expression' in order
        to avoid shadowing a global symbol of the same name.

    * config/tc-microblaze.c (md_assemble): Rename 'imm' to 'immed' in
        order to avoid shadowing global variable of the same name.

* dwarf2read.c (struct attribute): Increase sizes of unsnd and snd
fields to allow larger integer sizes.
(read_subrange_type): Increase size of bound values.
Add logic to determine signedness based on base-type size, signedness.
(read_attribute_value): Change format for bad byte size in message.
(read_8_bytes): Increase size of result type.
(dump_die_shallow): Change format for value.
(dwarf2_get_attr_constant_value): Increase size of return type.
Correct comment.
* gdbtypes.c (create_range_type): Change API to increase size of
bounds. struct field -> union field.
Always take signedness from base type.
(check_typedef): Use new API for TYPE_LOW_BOUND, TYPE_HIGH_BOUND.
(recursive_dump_type, copy_type_recursive): Adjust to new
representation of range types.
* gdbtypes.h (fields_or_bounds): New union containing struct field and
new struct range_bounds, used for range types.
(TYPE_RANGE_DATA): New macro to access range_bounds member.
(TYPE_LOW_BOUND, TYPE_HIGH_BOUND): Represent with new TYPE_RANGE_DATA.
(TYPE_LOW_BOUND_UNDEFINED, TYPE_HIGH_BOUND_UNDEFINED): New macros,
taking over the job of TYPE_FIELD_ARTIFICIAL for range bounds.
(SET_TYPE_LOW_BOUND, SET_TYPE_HIGH_BOUND, SET_TYPE_LOW_BOUND_DEFINED)
(SET_TYPE_HIGH_BOUND_DEFINED): New macros.
(TYPE_FIELDS, TYPE_BASECLASS, TYPE_BASECLASS_NAME, TYPE_FIELD)
(TYPE_ARRAY_UPPER_BOUND_IS_UNDEFINED)
(TYPE_ARRAY_LOWER_BOUND_IS_UNDEFINED, TYPE_ARRAY_UPPER_BOUND_VALUE)
(TYPE_ARRAY_LOWER_BOUND_VALUE): Adjust to new representation.
(create_range_type): Adjust API.
* ada-lang.c (ada_modulus): Use new extended bound values.
(discrete_type_low_bound): Rename to...
(ada_discrete_type_low_bound): ... and make external.
(discrete_type_high_bound): Rename to...
(ada_discrete_type_high_bound): ... and make external.
(ada_value_slice_from_ptr, ada_array_bound_from_type)
(ada_evaluate_subexp, to_fixed_range_type):
Use ada_discrete_type_low_bound, ada_discrete_type_high_bound.
* ada-typeprint.c (print_range): Use ada_discrete_type_low_bound,
ada_discrete_type_high_bound.  Don't look at field count, which
is no longer meaningful.  Print bounds whenever argument is a range
or enumeration.
* ada-lang.h (ada_discrete_type_low_bound,ada_discrete_type_high_bound):
Declare.
* varobj.c (c_describe_child): Adjust to render larger values.
* mdebugread.c (parse_type): Use proper abstractions for range types:
TYPE_RANGE_DATA, SET_TYPE_LOW_BOUND_DEFINED,
SET_TYPE_HIGH_BOUND_DEFINED.
* p-typeprint.c (pascal_type_print_varspec_prefix): Use larger format
for bounds.

*** empty log message ***

daily update

*** empty log message ***

    * coff-arm.c (coff_arm_rtype_to_howto): Fix shadowed variable
        warning.

2009-12-11  Doug Kwan  <dougkwan@google.com>

* arm.cc (Target_arm::do_finalize_sections): Fix build breakage
due to -Wshadow.
* attributes.cc (Object_attribute::size): Ditto.
(Attributes_section_data::size): Ditto.
(Attributes_section_data::Attributes_section_data): Ditto.
(Output_attributes_section_data::do_write): Ditto.
* attributes.h (Object_attribute::set_type): Ditto.
* testsuite/tls_test_main.cc (safe_lock, safe_unlock): Ditto.

Set vex.w to 0 for VEX C5 prefix.

2009-12-11  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (get_valid_dis386): Set vex.w to 0 for VEX C5
prefix.
(print_insn): Don't set vex.w here.

2009-12-11  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (print_insn): Set vex.w to 0.

*** empty log message ***

2009-12-11  Michael Snyder  <msnyder@vmware.com>

* main.c (captured_main): Indentation.

daily update

2009-12-11  Quentin Neill  <quentin.neill@amd.com>

gas/testsuite/
* gas/i386/fma4.d: Add test cases.
* gas/i386/fma4.s: Add test cases.
* gas/i386/x86-64-fma4.d: Add test cases.
* gas/i386/x86-64-fma4.s: Add test cases.

opcodes/
* i386-dis.c (get_vex_imm8): Extend logic to apply in all
cases, to avoid fetching ahead for the immediate bytes when
OP_E_memory has already been called.  Fix indentation.

   * config/tc-arm.c (arm_init_frag): Set thumb MODE_RECORDED flag for
       non-elf.
       (arm_handle_align): Re-enable assert for non-elf.

2009-12-11  Tristan Gingold  <gingold@adacore.com>

* som.c (bfd_som_set_subsection_attributes)
(bfd_section_from_som_symbol, som_reloc_queue_fix): Fix shadowed
variable warnings.
* cache.c (close_one): Likewise.

Add -Wshadow to the gcc command line options used when compiling the binutils.
Fix up all warnings generated by the addition of this switch.

2009-12-10  Michael Snyder  <msnyder@vmware.com>

* i386-tdep.c (i386_record_lea_modrm_addr): Indent goto label.
(i386_process_record): Indent goto label.  Use prefix opcode
constants from include/opcode/i386.h in switch statement.

*** empty log message ***

daily update

Fix for PR 9399 gdb can't call or print a const function that uses virtual inheritance

2009-12-10  Doug Kwan  <dougkwan@google.com>

elfcpp/ChangeLog
* arm.h: New enums for EABI object attribute tags and values.

gold/ChangeLog
* arm.cc (attributes.h): New include.
(Arm_relobj::Arm_relobj): Initialize attributes_section_data_.
(Arm_relobj::~Arm_relobj): Delete object pointed by
attributes_section_data_.
(Arm_relobj::attributes_section_data): New method definition.
(Arm_relobj::attributes_section_data_): New data member declaration.
(Arm_dynobj::Arm_dynobj): Initialize attributes_section_data_.
(Arm_dynobj::~Arm_dynobj): Delete object pointed by
attributes_section_data_.
(Arm_dynobj::attributes_section_data): New method definition.
(Arm_dynobj::attributes_section_data_): New data member declaration.
(Target_arm::Target_arm): Initialize attributes_section_data_.  Change
initialization value of may_use_blx_ to false.
    (Target_arm::using_thumb2, Target_arm::using_thumb_only,
Target_arm::may_use_arm_nop, Target_arm::may_use_thumb2_nop): Use
object attributes to compute results instead of hard-coding.
(Target_arm::do_attribute_arg_type, Target_arm::do_attributes_order,
Target_arm::get_secondary_compatible_arch,
Target_arm::set_secondary_compatible_arch
Target_arm::tag_cpu_arch_combine, Target_arm::aeabi_enum_name,
Target_arm::tag_cpu_name_value, Target_arm::merge_object_attributes):
New method declarations.
(Target_arm::get_aeabi_object_attribute): New method definition.
(Target_arm::attributes_section_data_): New data member declaration.
(read_arm_attributes_section): New template definition.
(Arm_relobj::do_read_symbols): Read attributes section if it exists.
(Arm_dynobj::do_read_symbols): Ditto.
(Target_arm::do_finalize_sections): Merge attributes sections from
input.  Check for BLX use after attributes section merging.
Fix __exidx_start and __exidx_end visibility.  Create an
.ARM.attributes section if necessary.
(Target_arm::get_secondary_compatible_arch,
Target_arm::set_secondary_compatible_arch,
Target_arm::tag_cpu_arch_combine, Target_arm::aeabi_enum_name,
Target_arm::tag_cpu_name_value, Target_arm::merge_object_attributes,
Target_arm::do_attribute_arg_type, Target_arm::do_attributes_order):
New method definitions.

2009-12-10  Oza Pawandeep (paawan1982@yahoo.com

* gdb.reverse/i387-env-reverse.c: New file.
* gdb.reverse/i387-env-reverse.exp: New file.
* gdb.reverse/i387-stack-reverse.c: New file.
* gdb.reverse/i387-stack-reverse.exp: New file.

2009-12-10  Oza Pawandeep <paawan1982@yahoo.com>

* i386-tdep.c:  Support for floating point recording.
Records and Replay the i386 floating point registers,
of i387 FPU.
(i386_record_floats): New function.
(i386_process_record): Implement FP insns.

2009-12-10  Tristan Gingold  <gingold@adacore.com>

* addr2line.c (pretty_print): New variable.
(long_options): Add an entry for -p/--pretty-print.
(usage): Document -p/--pretty-print.
(translate_addresses): Handle pretty_print.  Reindent.
(main): Handle option -p.
* doc/binutils.texi (addr2line): Document -p/--pretty-print.
* NEWS: Mention new feature.

bfd/
* elfxx-mips.c (mips_elf_calculate_relocation): Correct handling
of undefined symbols.

ld/testsuite/
* ld-mips-elf/undefined.d: New test.
* ld-mips-elf/undefined.s: Source for the new test.
* ld-mips-elf/mips-elf.exp: Run the new test.

* plugin.cc (Plugin::load): Don't cast from void* to a function
pointer.

2009-12-09  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (arch_entry): Add len and skip.
(cpu_arch): Use STRING_COMMA_LEN.
(MESSAGE_TEMPLATE): New.
(show_arch): Likewise.
(md_show_usage): Use show_arch.

* dynobj.cc (Sized_dynobj::do_read_symbols): Clear version
information fields.

*** empty log message ***

daily update

bfd/
* elf32-arm.c (elf32_arm_next_input_section): Skip sections without
SEC_CODE.

ld/testsuite/
* ld-arm/arm-elf.exp (armeabitests): Add farcall-data.
* ld-arm/farcall-data.d, ld-arm/farcall-data.s: New.
* lib/ld-lib.exp (run_ld_link_tests, run_cc_link_tests): Correct
regular expression for archives.

2009-12-09  H.J. Lu  <hongjiu.lu@intel.com>

* testsuite/Makefile.am (two_file_shared_1_pic_2_test_DEPENDENCIES):
Replace two_file_shared_1.so with two_file_shared_2.so.
* testsuite/Makefile.in: Regenerated.

2009-12-03  Tristan Gingold  <gingold@adacore.com>

* machoread.c (macho_symfile_read): Use symbol_file_add_separate
to add dsym files.

2009-12-02  Tristan Gingold  <gingold@adacore.com>

* symfile.h (symbol_file_add_separate): New prototype.
(find_separate_debug_file_by_buildid): Ditto.
(find_separate_debug_file_by_debuglink): Ditto.
* symfile.c (reread_separate_symbols): Remove.
(find_separate_debug_file): Split into ...
(find_separate_debug_file_by_buildid): ... this and ...
(find_separate_debug_file_by_debuglink): ... this.
(symbol_file_add_with_addrs_or_offsets): Do not save orig_addrs.
Remove separate debug file handling.
(symbol_file_add_separate): New function.
(reread_symbols): Do not considere separate debug files, but free
them while handling their parent.  Reindent.
* coffread.c (coff_symfile_read): Handle separate object file.
* elfread.c (elf_symfile_read): Ditto.

2009-12-09  Tristan Gingold  <gingold@adacore.com>

* addr2line.c (translate_addresses): Display addresses
if option -a is used.
(with_addresses): New variable.
(long_options): Add option '-a'.
(usage): Add usage for option '-a'.
(main): Handle option '-a'.
* doc/binutils.texi (addr2line): Document option '-a'.
* NEWS: Mention new feature.

    * acinclude.m4: Update comment.

        * acinclude.m4: Link against in-tree libiconv by adding
            libiconv.a to the LIBS instead of using -liconv.
            * configure: Regenerate.

    PR 10924
        * arm-dis.c (print_insn_arm): Mark insns that use the PC in
        post-indexed addressing as unpredictable.