* configure.ac: Add "recommended" version checks for GMP/MPC.
Update recommended GMP/MPFR/MPC versions.
* configure: Regenerate.

*** empty log message ***

daily update

binutils/testsuite/ChangeLog:

* binutils-all/windres/html.rc: Don't xfail x86_64-*-mingw*.
* binutils-all/windres/lang.rc: Likewise.
* binutils-all/windres/messagetable.rc: Likewise.
* binutils-all/windres/strtab1.rc: Likewise.
* binutils-all/windres/strtab2.rc: Likewise.
* binutils-all/windres/version.rc: Likewise.
* binutils-all/windres/version_cat.rc: Likewise.

* elf32-hppa.c (final_link_relocate): Convert R_PARISC_TLS_GD21L,
R_PARISC_TLS_LDM21L and R_PARISC_TLS_IE21L relocations that use the
linkage table pointer to use the global table pointer if not doing
a shared link.

2010-01-29  Viktor Kutuzov  <vkutuzov@accesssoftek.com>

* gold/arm.cc: Added support for the ARM relocations: R_ARM_THM_PC8,
R_ARM_THM_PC12, R_ARM_THM_ALU_PREL_11_0.
(Arm_relocate_functions::thm_alu11): New Method.
(Arm_relocate_functions::thm_pc8): New Method.
(Arm_relocate_functions::thm_pc12): New Method.
(Target_arm::Scan::local): Handle the relocations.
(Target_arm::Scan::global): Likewise.
(Target_arm::Relocate::relocate): Likewise.
(Target_arm:Relocatable_size_for_reloc::get_size_for_reloc): Likewise.

2010-01-29  Doug Kwan  <dougkwan@google.com>

* arm.cc (Target_arm::Scan::global): General PLTs for the same set
of relocation types as ld.

*** empty log message ***

daily update

2010-01-29  Doug Kwan  <dougkwan@google.com>

* arm.cc (Arm_relocate_functions::arm_branch_common) Change visibility
to public.
(Arm_relocate_functions::thumb_branch_common): Ditto.
(Arm_relocate_functions::thm_call, Arm_relocate_functions::thm_jump24,
Arm_relocate_functions::thm_xpc22, Arm_relocate_functions::plt32,
Arm_relocate_functions::xpc25, Arm_relocate_functions::call,
Arm_relocate_functions::jump24): Remove.
(Target_arm::Relocate::relocate): Adjust code to call
Arm_relocation_functions::arm_branch_common and
Arm_relocation_functions::thumb_branch_common instead of their removed
wrappers.  Merge switch-cases together to reduce source code size.

2010-01-29  Doug Kwan  <dougkwan@google.com>

* arm.cc (Arm_relobj::Arm_relobj): Initialize new data member
output_local_symbol_count_needs_update_.
(Arm_relobj::output_local_symbol_count_needs_update,
 Arm_relobj::set_output_local_symbol_count_needs_update,
 Arm_relobj::update_output_local_symbol_count): New methods.
(Arm_relobj::output_local_symbol_count_needs_update_): New data
member.
(Arm_exidx_cantunwind::do_fixed_endian_write): Write address
of pointed function as in a R_ARM_PREL31 relocation.
(Arm_output_section<big_endian>::fix_exidx_coverage): Mark objects
for output local symbol count updating.
(Target_arm::do_relax): Update output local symbol counts in objects
if necessary.
* object.h (Sized_relobj::set_output_local_symbol_count): New method.

gas/testsuite/
* gas/arm/dis-data.d: Update test name.  Do not expect
.word output.
* gas/arm/dis-data2.d, gas/arm/dis-data2.s,
gas/arm/dis-data3.d, gas/arm/dis-data3.s: New tests.

opcodes/
* opcodes/arm-dis.c (struct arm_private_data): New.
(print_insn_coprocessor, print_insn_arm): Update to use struct
arm_private_data.
(is_mapping_symbol, get_map_sym_type): New functions.
(get_sym_code_type): Check the symbol's section.  Do not check
mapping symbols.
(print_insn): Default to disassembling ARM mode code.  Check
for mapping symbols separately from other symbols.  Use
struct arm_private_data.

    PR 11136
        * config/tc-arm.c (neon_check_type): Handle a neon_shape value of
        NS_NULL.
        * gas/arm/neon-omit.s: Add instruction that causes crash.
        * gas/arm/neon-omit.d: Add expected disassembly.

* arm.cc: Added support for the ARM relocations:
R_ARM_MOVW_BREL_NC, R_ARM_MOVT_BREL, R_ARM_MOVW_BREL,
R_ARM_THM_MOVW_BREL_NC, R_ARM_THM_MOVT_BREL, R_ARM_THM_MOVW_BREL.
(Arm_relocate_functions::movw_rel_nc): Renamed (was
movw_prel_nc).
(Arm_relocate_functions::movw_rel): New method.
(Arm_relocate_functions::movt_rel): Renamed (was movt_prel).
(Arm_relocate_functions::thm_movw_rel_nc): Renamed (was
thm_movw_prel_nc).
(Arm_relocate_functions::thm_movw_rel): New method.
(Arm_relocate_functions::thm_movt_rel): Renamed (was
thm_movt_prel).
(Target_arm::Scan::local): Handle MOVW_BREL/MOVT_BREL
relocations.
(Target_arm::Scan::global): Likewise.
(Target_arm::Relocate::relocate): Likewise.
(Target_arm::Relocatable_size_for_reloc::get_size_for_reloc):
Likewise.

* infrun.c (prepare_to_proceed): Handle other signals which might
match a breakpoint.
(handle_inferior_event): Move the check for unusual breakpoint
signals earlier.

gdbserver/
* linux-low.c (get_stop_pc): Check for SIGTRAP.
(linux_wait_for_event_1): Handle SIGILL and SIGSEGV as possible
breakpoints.

* gdb.base/call-strs.exp, gdb.base/default.exp,
gdb.base/ending-run.exp, gdb.base/finish.exp, gdb.base/funcargs.exp,
gdb.base/huge.exp, gdb.base/nodebug.exp, gdb.base/ptype.exp,
gdb.base/restore.exp, gdb.base/return.exp, gdb.base/setvar.exp,
gdb.base/watchpoints.exp, gdb.threads/gcore-thread.exp,
gdb.base/watchpoint-solib.exp: Save and restore timeout.
* gdb.base/ending-run.exp: Correct restore of timeout.
* gdb.base/page.exp: Remove unnecessary timeout setting.

amd64 - function returning record with field straddling 2 registers

gdb/ChangeLog:

        From Paul Hilfinger  <hilfinger@adacore.com>
        * amd64-tdep.c (amd_classify_aggregate): Handle the case of
        a record of length <= 16 in which a field straddles the two
        eightbytes.

gdb/testsuite/ChangeLog:

        * gdb.ada/rec_return: New testcase.

Implement return values on amd64-windows.

        * amd64-windows-tdep.c: #include gdbcore.h and regcache.h.
        (amd64_windows_return_value): New function.
        (amd64_windows_init_abi): Call set_gdbarch_return_value with
        amd64_windows_return_value.

amd64-windows: 32 bytes allocated on stack by caller for integer parameter registers

        * i386-tdep.h (struct gdbarch_tdep): Add new field
        integer_param_regs_saved_in_caller_frame.
        * amd64-windows-tdep.c (amd64_windows_init_abi): Set
        tdep->integer_param_regs_saved_in_caller_frame to 1.
        * amd64-tdep.c (amd64_push_dummy_call): Allocate some memory on
        stack if tdep->integer_param_regs_saved_in_caller_frame is set.

Add summary line to the second to last entry.

amd64-windows: memory args passed by pointer during function calls.

        * i386-tdep.h (gdbarch_tdep): Add field memory_args_by_pointer.
        * amd64-tdep.c (amd64_push_arguments): Add handling of architectures
        where tdep->memory_args_by_pointer is non-zero.
        * amd64-windows-tdep.c (amd64_windows_init_abi): Set
        tdep->memory_args_by_pointer to 1.

amd64: Integer parameters in function calls on Windows.

gdb/ChangeLog:

        * i386-tdep.h (enum amd64_reg_class): New, moved here from
        amd64-tdep.c.
        (struct gdbarch_tdep): Add fields call_dummy_num_integer_regs,
        call_dummy_integer_regs, and classify.
        * amd64-tdep.h (amd64_classify): Add declaration.
        * amd64-tdep.c (amd64_dummy_call_integer_regs): New static constant.
        (amd64_reg_class): Delete, moved to i386-tdep.h.
        (amd64_classify): Make non-static.  Move declaration to amd64-tdep.h.
        Replace call to amd64_classify by call to tdep->classify.
        (amd64_push_arguments): Get the list of registers to use for
        passing integer parameters from the gdbarch tdep structure,
        rather than using a hardcoded one.  Replace calls to amd64_classify
        by calls to tdep->classify.
        (amd64_push_dummy_call): Get the register number used for
        the "hidden" argument from tdep->call_dummy_integer_regs.
        (amd64_init_abi): Initialize tdep->call_dummy_num_integer_regs
        and tdep->call_dummy_integer_regs.  Set tdep->classify.
        * amd64-windows-tdep.c: Add include of gdbtypes.h.
        (amd64_windows_dummy_call_integer_regs): New static global.
        (amd64_windows_classify): New function.
        (amd64_windows_init_abi): Initialize tdep->call_dummy_num_integer_regs
        tdep->call_dummy_integer_regs and tdep->classify.

gdb/testsuite/ChangeLog:

        * gdb.ada/call_pn: New testcase.

core files: wrong signal number with threaded program on sparc-solaris.

        * elf.c (elfcore_grok_lwpstatus): Do not overwrite the core signal
        if it has already been set.

*** empty log message ***

daily update

* gdb.mi/mi-nonstop.exp (mi_nonstop_resume): New function.
(Top level): Use it to resume.
* lib/mi-support.exp (mi_send_resuming_command_raw): Recognize
the Thumb mode displaced stepping error as unsupported.

* regcache.c (regcache_xmalloc): Add aspace argument.  Use it
for the new regcache.  All callers updated.
(regcache_cpy, regcache_cpy_no_passthrough): Do not set aspace here.
(get_thread_arch_regcache): Do not set aspace here.
* regcache.h (regcache_xmalloc): Update declaration.

* frame.c, infcall.c, ppc-linux-tdep.c: Calls to
regcache_xmalloc updated.

* gas/pe/section-align-1.d: Don't test section flags.
* gas/pe/section-align-2.d: Likewise.

2010-01-28  Sami Wagiaalla  <swagiaal@redhat.com>

    * gdb.cp/nsusing.exp: Added more tests.
    * gdb.cp/nsrecurs.exp: Ditto.
    * gdb.cp/nsusing.cc: Added test functions.
    * gdb.cp/nsrecurs.cc: Ditto.

Allow VL=1 on scalar FMA instructions.

gas/testsuite/

2010-01-28  H.J. Lu  <hongjiu.lu@intel.com>

* gas/i386/fma-scalar-intel.d: New.
* gas/i386/fma-scalar.d: Likewise.
* gas/i386/fma-scalar.s: Likewise.
* gas/i386/x86-64-fma-scalar-intel.d: Likewise.
* gas/i386/x86-64-fma-scalar.d: Likewise.
* gas/i386/x86-64-fma-scalar.s: Likewise.

* gas/i386/i386.exp: Run fma-scalar, fma-scalar-intel,
x86-64-fma-scalar and x86-64-fma-scalar-intel.

opcodes/

2010-01-28  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (EXVexWdqScalar): New.
(vex_scalar_w_dq_mode): Likewise.
(prefix_table): Update entries for PREFIX_VEX_3899,
PREFIX_VEX_389B, PREFIX_VEX_389D, PREFIX_VEX_389F,
PREFIX_VEX_38A9, PREFIX_VEX_38AB, PREFIX_VEX_38AD,
PREFIX_VEX_38AF, PREFIX_VEX_38B9, PREFIX_VEX_38BB,
PREFIX_VEX_38BD and PREFIX_VEX_38BF.
(intel_operand_size): Handle vex_scalar_w_dq_mode.
(OP_EX): Likewise.

    PR 11225
        * objdump.c (only): Replace with linked list.
        (only_size, only_used): Replace with only_list.
        (process_section_p): Set seen field on matches sections.
        (add_only): New function.
        (free_only_list): New function.
        (disassemble_section): Check only_list.
        (main): Use add_only and free_only_list.

        * gas/pe/aligncomm-c.d: Dump all sections.

        * ld-sh/refdbg-0-dso.d: Dump all sections.

Another -Wunused-function error in procfs.c (sparc-solaris)

        * procfs.c (insert_dbx_link_breakpoint): Delete declaration.  Move up.
        Only define if SYS_syssgi is defined.
        (remove_dbx_link_breakpoint): Delete declaration.  Move up.
        (dbx_link_addr, insert_dbx_link_bpt_in_file)
        (insert_dbx_link_bpt_in_region): Move up.  Only define if SYS_syssgi
        is itself defined.

*** empty log message ***

daily update

gas/ChangeLog:

* NEWS: Mention new feature.
* config/obj-coff.c (obj_coff_section): Accept digits and use
to override default section alignment power if specified.
* doc/as.texinfo (.section directive): Update documentation.

gas/testsuite/ChangeLog:

* gas/pe/section-align-1.s: New test source file.
* gas/pe/section-align-1.d: Likewise control script.
* gas/pe/section-align-2.s: Likewise ...
* gas/pe/section-align-2.d: ... and likewise.
* gas/pe/pe.exp: Invoke new testcases.

gdb/testsuite/
* gdb.base/break-interp.exp: Use [file tail $filename] to omit any
absolute directory pathnames in gdb.sum file.

* windows-nat.c (windows_initialization_done): New variable.
(get_windows_debug_event): Issue error when process dies before completely
initializing.
(do_initial_windows_stuff): Set flag to indicate when we are done with the
initial steps of attaching to the child.

2010-01-27  Sami Wagiaalla  <swagiaal@redhat.com>

* gdb.cp/namespace-using.exp: Rename to nsusing.exp.
* gdb.cp/namespace-using.cc: Rename to nsusing.cc.

2010-01-27  Tristan Gingold  <gingold@adacore.com>

* vms.h: Remove trailing spaces.
(struct hdr_struct): Remove unused hdr_c_cpr field.
(enum file_format_enum): Add comments.
* vms.c (vms_bfd_print_private_bfd_data): New function that
replaces the macro.
Remove trailing spaces.
* vms-misc.c: Improve comments.
(_bfd_vms_get_object_record): Also handle files without pads.
* vms-hdr.c (_bfd_vms_write_dbg): Fix format string.
* vms-gsd.c (struct flagdescstruct): Make name field const.
(gsyflagdesc): Fix typo.

* arm.cc: Added support for ARM group relocations.
(Target_arm::reloc_needs_sym_origin): New method.
(Arm_relocate_functions::calc_grp_kn): New method.
(Arm_relocate_functions::calc_grp_residual): New method.
(Arm_relocate_functions::calc_grp_gn): New method.
(Arm_relocate_functions::arm_grp_alu): New Method.
(Arm_relocate_functions::arm_grp_ldr): New Method.
(Arm_relocate_functions::arm_grp_ldrs): New Method.
(Arm_relocate_functions::arm_grp_ldc): New Method.
(Target_arm::Scan::local): Handle the ARM group relocations.
(Target_arm::Scan::global): Likewise.
(Target_arm::Relocate::relocate): Likewise.
(Target_arm::Relocatable_size_for_reloc::get_size_for_reloc):
Likewise.

Allow VL=1 on AVX scalar instructions.

gas/

2010-01-27  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (avxscalar): New.
(OPTION_MAVXSCALAR): Likewise.
(build_vex_prefix): Select vector_length for scalar instructions
based on avxscalar.
(md_longopts): Add OPTION_MAVXSCALAR.
(md_parse_option): Handle OPTION_MAVXSCALAR.
(md_show_usage): Add -mavxscalar=.

* doc/c-i386.texi: Document -mavxscalar=.

gas/testsuite/

2010-01-27  H.J. Lu  <hongjiu.lu@intel.com>

* gas/i386/avx-scalar-intel.d: New.
* gas/i386/avx-scalar.d: Likewise.
* gas/i386/avx-scalar.s: Likewise.
* gas/i386/x86-64-avx-scalar-intel.d: Likewise.
* gas/i386/x86-64-avx-scalar.d: Likewise.
* gas/i386/x86-64-avx-scalar.s: Likewise.

* gas/i386/i386.exp: Run avx-scalar, avx-scalar-intel,
x86-64-avx-scalar and x86-64-avx-scalar-intel.

opcodes/

2010-01-27  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (XMScalar): New.
(EXdScalar): Likewise.
(EXqScalar): Likewise.
(EXqScalarS): Likewise.
(VexScalar): Likewise.
(EXdVexScalarS): Likewise.
(EXqVexScalarS): Likewise.
(XMVexScalar): Likewise.
(scalar_mode): Likewise.
(d_scalar_mode): Likewise.
(d_scalar_swap_mode): Likewise.
(q_scalar_mode): Likewise.
(q_scalar_swap_mode): Likewise.
(vex_scalar_mode): Likewise.
(vex_len_table): Duplcate entries for VEX_LEN_10_P_1,
VEX_LEN_10_P_3, VEX_LEN_11_P_1, VEX_LEN_11_P_3, VEX_LEN_2A_P_1,
VEX_LEN_2A_P_3, VEX_LEN_2C_P_3, VEX_LEN_2D_P_1, VEX_LEN_2E_P_0,
VEX_LEN_2E_P_2, VEX_LEN_2F_P_2, VEX_LEN_51_P_1, VEX_LEN_51_P_3,
VEX_LEN_52_P_1, VEX_LEN_53_P_1, VEX_LEN_58_P_1, VEX_LEN_58_P_3,
VEX_LEN_59_P_1, VEX_LEN_5A_P_1, VEX_LEN_5A_P_3, VEX_LEN_5C_P_1,
VEX_LEN_5C_P_3, VEX_LEN_5D_P_1, VEX_LEN_5D_P_3, VEX_LEN_5E_P_1,
VEX_LEN_5E_P_3, VEX_LEN_5F_P_1, VEX_LEN_5F_P_3, VEX_LEN_6E_P_2,
VEX_LEN_7E_P_1, VEX_LEN_7E_P_2, VEX_LEN_D6_P_2, VEX_LEN_C2_P_1,
VEX_LEN_C2_P_3, VEX_LEN_3A0A_P_2 and VEX_LEN_3A0B_P_2.
(vex_w_table): Update entries for VEX_W_10_P_1, VEX_W_10_P_3,
VEX_W_11_P_1, VEX_W_11_P_3, VEX_W_2E_P_0, VEX_W_2E_P_2,
VEX_W_2F_P_0, VEX_W_2F_P_2, VEX_W_51_P_1, VEX_W_51_P_3,
VEX_W_52_P_1, VEX_W_53_P_1, VEX_W_58_P_1, VEX_W_58_P_3,
VEX_W_59_P_1, VEX_W_59_P_3, VEX_W_5A_P_1, VEX_W_5A_P_3,
VEX_W_5C_P_1, VEX_W_5C_P_3, VEX_W_5D_P_1, VEX_W_5D_P_3,
VEX_W_5E_P_1, VEX_W_5E_P_3, VEX_W_5F_P_1, VEX_W_5F_P_3,
VEX_W_7E_P_1, VEX_W_D6_P_2  VEX_W_C2_P_1, VEX_W_C2_P_3,
VEX_W_3A0A_P_2 and VEX_W_3A0B_P_2.
(intel_operand_size): Handle d_scalar_mode, d_scalar_swap_mode,
q_scalar_mode, q_scalar_swap_mode.
(OP_XMM): Handle scalar_mode.
(OP_EX): Handle d_scalar_mode, d_scalar_swap_mode, q_scalar_mode
and q_scalar_swap_mode.
(OP_VEX): Handle vex_scalar_mode.

PR ld/11217
* elf64-ppc.c (struct ppc_link_hash_table): Add toc_bfd, toc_first_sec.
(ppc64_elf_setup_section_lists): Init them.
(ppc64_elf_next_toc_section): Don't partition multi-toc between .got
and .toc on the same input file.
(ppc64_elf_relocate_section): Correct GOT entry offset.

* solib-svr4.c (solib_break_names): Add __dl_rtld_db_dlactivity.

gdb/
* symtab.h (struct symbol <symtab>): New comment on NULL values.

*** empty log message ***

* breakpoint.c (bpstat_stop_status): Delete useless code.

gdb/
* printcmd.c (display_uses_solib_p): Remove variable section.  Access
objfile via SYMBOL_SYMTAB.

daily update

2010-01-26  Doug Kwan  <dougkwan@google.com>

* arm.cc (set): Include.
(class Arm_exidx_fixup): Change type of last_input_section_ to const
pointer type.
  (Arm_output_section::Text_section_list): New type.
(Arm_output_section::append_text_sections_to_list): New method.
(Arm_output_section::fix_exidx_coverage): Ditto.
(Arm_relobj::Arm_relobj): Initialize exidx_section_map_.
    (Arm_relobj::convert_input_section_to_relaxed_section): Use
Relobj::set_section_offset() instead of
Sized_relobj::invalidate_section_offset().
    (Arm_relobj::section_needs_reloc_stub_scanning): Add an extra
parameter for section headers. Ignore relocation sections for
unallocated sections and EXIDX sections.
(Target_arm::fix_exidx_coverage): New method.
(Target_arm::output_section_address_less_than): New type.
(Arm_exidx_fixup::add_exidx_cantunwind_as_needed): Use index of the
linked text section instead of the EXIDX section.
(Arm_output_section::create_stub_group): Add an assertion to check
that this is not an EXIDX output section.
(Arm_output_section::append_text_sections_to_list): New method.
(Arm_output_section::fix_exidx_coverage): Ditto.
(Arm_relobj::scan_sections_for_stubs): Adjust call to
    Arm_relobj::section_needs_reloc_stub_scanning.
(Target_arm::do_relax): Fix EXIDX output section coverage in the
first pass.
(Target_arm::fix_exidx_coverage): New method.
* object.h (Relobj::set_output_section): New method.
(Sized_relobj::invalidate_section_offset): Remove method.
(Sized_relobj::do_invalidate_section_offset): Remove method.
(Sized_relobj::do_set_section_offset): Handle offset value -1.

gdb
PR exp/7643:
* eval.c (evaluate_subexp_for_address) <UNOP_IND>: Call
coerce_array on result.
gdb/testsuite
PR exp/7643:
* gdb.base/printcmds.exp (test_print_string_constants): Remove
setup_kfail.

 2010-01-14  Sami Wagiaalla  <swagiaal@redhat.com>

* gdb.cp/nsrecurs.exp: New test.
* gdb.cp/nsrecurs.cc: New test program.
* gdb.cp/nsstress.exp: New test.
* gdb.cp/nsstress.cc: New test program.
* gdb.cp/nsdecl.exp: New test.
* gdb.cp/nsdecl.cc: New test program.

 2010-01-26  Sami Wagiaalla  <swagiaal@redhat.com>

* cp-namespace.c (cp_lookup_symbol_namespace): Added
search_parent argument.
(cp_add_using): Initialize 'searched' field.
(reset_directive_searched): New function.
* cp-support.h: Add 'searched' field to using_direct struct.
(cp_lookup_symbol_imports): Ditto.
* cp-namespace.c (cp_lookup_symbol_imports): Ditto.
Perform recursive search.
Implement non parent search.
* valops.c (value_maybe_namespace_elt): Updated.

2010-01-26  Sami Wagiaalla  <swagiaal@redhat.com>

* gdb.cp/namespace-using.exp: Add test for printing of namespaces
imported into file scope.
Marked test as xfail.
* gdb.cp/namespace-using.cc (marker5): New function.
* gdb.cp/shadow.exp: New test.
* gdb.cp/shadow.cc: New test program.
* gdb.cp/nsimport.exp: New test.
* gdb.cp/nsimport.cc: New test program.

2010-01-26  Sami Wagiaalla  <swagiaal@redhat.com>

PR gdb/10929:
* dwarf2read.c (read_lexical_block_scope): Create blocks for
scopes which contain using directives even if they contain no
declarations.
* symtab.c (lookup_symbol_aux): Pass lowest level block to
la_lookup_symbol_nonlocal.
* cp-namespace.c (cp_lookup_symbol_nonlocal): call
cp_lookup_symbol_namespace.
(cp_lookup_symbol_namespace): Perform an import lookup at every
block level.
(cp_lookup_symbol_imports): New function.
(cp_lookup_symbol_in_namespace): New function.

2010-01-26  Tristan Gingold  <gingold@adacore.com>

* targets.c (BFD_JUMP_TABLE_ARCHIVE): Add initializer for write_ar_hdr.
(bfd_target): Add _bfd_write_ar_hdr_fn field.
* archive.c (is_bsd44_extended_name): New macro.
(_bfd_generic_read_ar_hdr_mag): Use it.  Add extra_size.
(bfd_slurp_armap): Also check for "__.SYMDEF" as a BSD4.4 extended
name.
(_bfd_archive_bsd44_construct_extended_name_table): New function.
(_bfd_generic_write_ar_hdr): Ditto.
(_bfd_bsd44_write_ar_hdr): Ditto.
(_bfd_write_archive_contents): Call _bfd_write_ar_hdr.
(bsd_write_armap): Adjust firstreal computation.
* libbfd-in.h (struct areltdata): Add extra_size field.
(_bfd_generic_write_ar_hdr): Add prototype.
(_bfd_bsd44_write_ar_hdr): Ditto.
(_bfd_write_ar_hdr): Define.
(_bfd_noarchive_write_ar_hdr): Ditto.
(_bfd_archive_bsd_write_ar_hdr): Ditto.
(_bfd_archive_coff_write_ar_hdr): Ditto.
(_bfd_archive_bsd44_slurp_armap): Ditto.
(_bfd_archive_bsd44_slurp_extended_name_table): Ditto.
(_bfd_archive_bsd44_construct_extended_name_table): New prototype.
(_bfd_archive_bsd44_truncate_arname): Ditto.
(_bfd_archive_bsd44_write_armap): Ditto.
(_bfd_archive_bsd44_read_ar_hdr): Ditto.
(_bfd_archive_bsd44_write_ar_hdr): Ditto.
(_bfd_archive_bsd44_openr_next_archived_file): Ditto.
(_bfd_archive_bsd44_get_elt_at_index): Ditto.
(_bfd_archive_bsd44_generic_stat_arch_elt): Ditto.
(_bfd_archive_bsd44_update_armap_timestamp): Ditto.
* libbfd.h: Regenerate.
* oasys.c (oasys_write_ar_hdr): Define.
* libecoff.h (_bfd_ecoff_write_ar_hdr): Define.
* ieee.c (ieee_write_ar_hdr): Define.
* elf64-mips.c (bfd_elf64_archive_write_ar_hdr): Define.
* coff-rs6000.c (rs6000coff_vec): Adjust for write_ar_hdr field.
(bfd_pmac_xcoff_backend_data): Ditto.
* coff64-rs6000.c (rs6000coff64_vec): Ditto.
(bfd_xcoff_aix5_backend_data): Ditto.
* coff-alpha.c (alpha_ecoff_write_ar_hdr): Define.
* aout-target.h (MY_write_ar_hdr): Define it if not defined.
* aout-tic30.c (MY_write_ar_hdr): Ditto.
* mach-o-target.c (TARGET_NAME): Use _bfd_archive_bsd44 archive.
(bfd_mach_o_mkarchive, bfd_mach_o_read_ar_hdr)
(bfd_mach_o_slurp_armap, bfd_mach_o_slurp_extended_name_table)
(bfd_mach_o_construct_extended_name_table)
(bfd_mach_o_truncate_arname, bfd_mach_o_write_armap)
(bfd_mach_o_get_elt_at_index, bfd_mach_o_generic_stat_arch_elt)
(bfd_mach_o_update_armap_timestamp): Moved to mach-o.c
* mach-o.c (bfd_mach_o_mkarchive, bfd_mach_o_read_ar_hdr)
(bfd_mach_o_slurp_armap, bfd_mach_o_slurp_extended_name_table)
(bfd_mach_o_construct_extended_name_table)
(bfd_mach_o_truncate_arname, bfd_mach_o_write_armap)
(bfd_mach_o_get_elt_at_index, bfd_mach_o_generic_stat_arch_elt)
(bfd_mach_o_update_armap_timestamp): Moved from mach-o-target.c
* bfd-in2.h: Regenerate.

Properly check undefined symbols when gc_sections is active.

bfd/

2010-01-26  Alan Modra  <amodra@gmail.com>
    H.J. Lu  <hongjiu.lu@intel.com>

PR ld/11218
* elflink.c (elf_link_output_extsym): Do not ignore undefined
symbols with ref_regular set when gc_sections is active.

ld/testsuite/

2010-01-26  H.J. Lu  <hongjiu.lu@intel.com>

PR ld/11218
* ld-gc/dummy.s: New.
* ld-gc/pr11218-1.c: Likewise.
* ld-gc/pr11218-2.c: Likewise.
* ld-gc/pr11218.d: Likewise.

2010-01-26  Tristan Gingold  <gingold@adacore.com>

* Makefile.am (bin2c): Add libintl dependance and library.
* Makefile.in: Regenerate.

*** empty log message ***

daily update

gdb
PR gdb/11049:
* c-valprint.c (c_val_print): Fix test of extract_unsigned_integer
result.
gdb/testsuite
PR gdb/11049:
* gdb.base/printcmds.exp (test_print_typedef_arrays): Add test
with "set print null-stop on".

gcc PR libstdc++/36101, gcc PR libstdc++/42813
* configure.ac (bootstrap_target_libs): Make inclusion of
target-libgomp conditional on libgomb being in target_configdirs.
* configure: Regenerate.

2010-01-25  Doug Kwan  <dougkwan@google.com>

* arm.cc (Arm_exidx_merged_section::do_output_offset):
Fix warning due to signed and unsigned comparison on a 32-bit host.

* configure.ac: Only use host_os part when disabling TUI on osf.
Use test to check variables, prefix strings with x.
* configure: Regenerate.

* solib-osf.c (osf_current_sos): Initialize tail.

2010-01-25  gingold  <gingold@adacore.com>

        * windows-nat.c (windows_continue): Use %x to print thread id.
        (get_windows_debug_event): Ditto.

PR ld/11217
* elf64-ppc.c (ppc64_elf_tls_optimize): Optimize tls sequences
with relocations against undefined weak symbols.
(ppc64_elf_relocate_section): Don't optimize calls to undefined
weak functions if the symbol is dynamic.
(ppc64_elf_relocate_section): Edit tprel tls sequences.
* elf32-ppc.c (ppc_elf_relocate_section): Likewise.
(_bfd_elf_ppc_at_tprel_transform): New function.
* bfd-in.h (_bfd_elf_ppc_at_tprel_transform): Declare.
* bfd-in2.h: Regenerate.

common/
* cgen-accfp.c (fextsfdf): New arg how.  All callers updated.
(ftruncdfsf, floatsisf, flostsidf, ufloatsisf, fixsfsi, fixdfsi,
ufixsfsi): Ditto.
* cgen-fpu.h (CGEN_FPCONV_KIND): New enum.
(struct cgen_fp_ops): Update signatures of floating point conversion
operations.

frv/
* sem.c: Regenerate.

sh64/
* cpu.h: Regenerate.

* desc-cpu.scm (cgen-desc.h): Don't print virtual enums.
* sid-cpu.scm (cgen-desc.h): Ditto.
* enum.scm (enum-builtin!): New function.
* read.scm (reader-install-builtin!): Call it.
* rtl-c.scm (s-convop): Delete, replaced with ...
(s-int-convop, s-float-convop): ... new fns.
(ext, zext, trunc): Update.
(fext, ftrunc, float, ufloat, fix, ufix): Update.
* rtx-funcs.scm (fext, ftrunc, float, ufloat, fix, ufix): New parameter
`how'.
* cpu/mep-fmax.cpu (fcvtsw): Update.
* cpu/sh.cpu (h-fsd, h-fmov): Update.
* doc/rtl.texi (float-convop): Update.

* frv.cpu (floating-point-conversion): Update call to fp conv op.
(floating-point-dual-conversion, ne-floating-point-dual-conversion,
conditional-floating-point-conversion, ne-floating-point-conversion,
float-parallel-mul-add-double-semantics): Ditto.

* Make-common.in (CGEN_SIM_DEPS): Define.
(CGEN_INCLUDE_DEPS): Use it.
(CGEN_MAIN_CPU_DEPS): Simplify.

*** empty log message ***

Remove trailing { Bad_Opcode }.

daily update

Remove trailing { Bad_Opcode } in vex_len_table.

2010-01-24  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (vex_len_table): Remove trailing { Bad_Opcode }.

Remove trailing { Bad_Opcode }.

2010-01-24  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (prefix_table): Remove trailing { Bad_Opcode }.

Remove trailing "(bad)" entries and replace { "(bad)", { XX } }
with { Bad_Opcode }.

2010-01-24  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (Bad_Opcode): New.
(bad_opcode): Likewise.
(dis386): Replace { "(bad)", { XX } } with { Bad_Opcode }.
(dis386_twobyte): Likewise.
(reg_table): Likewise.
(prefix_table): Likewise.
(x86_64_table): Likewise.
(vex_len_table): Likewise.
(vex_w_table): Likewise.
(mod_table): Likewise.
(rm_table): Likewise.
(float_reg): Likewise.
(reg_table): Remove trailing "(bad)" entries.
(prefix_table): Likewise.
(x86_64_table): Likewise.
(vex_len_table): Likewise.
(vex_w_table): Likewise.
(mod_table): Likewise.
(rm_table): Likewise.
(get_valid_dis386): Handle bytemode 0.

Set the first 3byte VEX prefix individually.

2010-01-24  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (build_vex_prefix): Set i.vex.bytes[0] to
0xc4 individually.

gdb/testsuite/
* gdb.arch/i386-bp_permanent.exp (Disassemble function '$function'):
Adjust it for DISASSEMBLY_OMIT_FNAME.

Replace "Vex" with "Vex=3" on AVX scalar instructions.

2010-01-23  H.J. Lu  <hongjiu.lu@intel.com>

* i386-opc.h (VEXScalar): New.

* i386-opc.tbl: Replace "Vex" with "Vex=3" on AVX scalar
instructions.
* i386-tbl.h: Regenerated.

*** empty log message ***

daily update

Add more AVX tests.

2010-01-23  H.J. Lu  <hongjiu.lu@intel.com>

* gas/i386/avx.s: Add more tests.
* gas/i386/x86-64-avx.s: Likewise.

* gas/i386/avx-intel.d: Updated.
* gas/i386/avx.d: Likewise.
* gas/i386/x86-64-avx-intel.d: Likewise.
* gas/i386/x86-64-avx.d: Likewise.

bfd/
* coff-rs6000.c (xcoff_howto_table): Change size to 0 and bitsize to 1.
(_bfd_xcoff_reloc_type_lookup): Handle BFD_RELOC_NONE.
* coff64-rs6000.c (xcoff64_howto_table): Change size to 0 and
bitsize to 1.
(xcoff64_reloc_type_lookup): Handle BFD_RELOC_NONE.

gas/
* write.h (fix_at_start): Declare.
* write.c (fix_new_internal): Add at_beginning parameter.
Use it instead of REVERSE_SORT_RELOCS.  Fix the handling of
seg_fix_tailP for the at_beginning/REVERSE_SORT_RELOCS case.
(fix_new, fix_new_exp): Update accordingly.
(fix_at_start): New function.
* config/tc-ppc.c (md_pseudo_table): Add .ref to the OBJ_XCOFF section.
(ppc_ref): New function, for OBJ_XCOFF.
(md_apply_fix): Handle BFD_RELOC_NONE for OBJ_XCOFF.
* config/te-i386aix.h (REVERSE_SORT_RELOCS): Remove #undef.

gas/testsuite/
* gas/ppc/xcoff-ref-1.s, gas/ppc/xcoff-ref-1.l: New test.
* gas/ppc/aix.exp: Run it.

ld/testsuite/
* ld-powerpc/aix-ref-1-32.od, ld-powerpc/aix-ref-1-64.od,
ld-powerpc/aix-ref-1.s: New tests.
* ld-powerpc/aix52.exp: Run them.

gcc PR libstdc++/36101, gcc PR libstdc++/42813
* configure.ac (bootstrap_target_libs): Include target-libgomp.
* configure: Regenerate.

2010-01-22  Doug Kwan  <dougkwan@google.com>

* arm.cc (Target_arm::do_relax): Record an output section for section
offset adjustment it contains any stub table that has changed.
* layout.cc (Layout::clean_up_after_relaxation): Adjust section
offsets in an output section if necessary.
* output.cc (Output_section::Output_section): Initialize
section_offsets_need_adjustments_.
(Output_section::add_input_section_for_script): Renamed to
Output_section::add_simple_input_section.
(Output_section::save_states): Add a comment.
(Output_section::discard_states): New method defintion.
(Output_section::adjust_section_offsets): Same.
* output.h (Output_section::add_input_section_for_script): Renamed to
Output_section::add_simple_input_section.
(Output_section::discard_states): New method declaration.
(Output_section::adjust_section_offsets): Same.
(Output_section::section_offsets_need_adjustment,
Output_section::set_section_offsets_need_adjustment): New method
definitions.
(Output_section::section_offsets_need_adjustment_): New data member.
* script-sections.cc
(Output_section_element_input::set_section_address): Adjust code for
renaming of Output_section::add_input_section_for_script.
(Orphan_output_section::set_section_address): Same.

*** empty log message ***

daily update

2010-01-22  Viktor Kutuzov  <vkutuzov@accesssoftek.com>

* gold/arm.cc (Target_arm): Updated fix_v4bx method and usage of
Fix_v4bx enum values .
* gold/options.h (General_options): New option definitions.
(General_options::fix_v4bx): New method.
(General_options::Fix_v4bx): New enum.
* gold/options.cc (General_options::parse_fix_v4bx): New method.
(General_options::parse_fix_v4bx_interworking): New method.

2010-01-22  Doug Kwan  <dougkwan@google.com>

* arm.cc (Arm_exidx_fixup): New class.

gdb
PR symtab/11199:
* dwarf2read.c (quirk_gcc_member_function_pointer): Change return
type and arguments.  Use smash_to_methodptr_type.
(read_structure_type): Call quirk_gcc_member_function_pointer
later.
* gdbtypes.h (smash_to_methodptr_type): Declare.
* gdbtypes.c (smash_to_methodptr_type): New function.
(lookup_methodptr_type): Use it.
gdb/testsuite
2010-01-22  Jan Kratochvil  <jan.kratochvil@redhat.com>

PR symtab/11199:
* gdb.dwarf2/member-ptr-forwardref.exp,
gdb.dwarf2/member-ptr-forwardref.S: New.

gcc PR libstdc++/36101, gcc PR libstdc++/42813
* configure.ac (target_configdirs): Substitute.
* Makefile.def: Bootstrap target module libgomp.
Add dependency of all-target-libstdc++-v3 on configure-target-libgomp.
* Makefile.tpl (TARGET_CONFIGDIRS): New makefile variable.
(BASE_TARGET_EXPORTS): Export TARGET_CONFIGDIRS.
* configure, Makefile.in: Regenerate.

* cgen-ops.h (SUBWORDXFSI): Fix word ordering.
(SUBWORDTFSI, JOINSIDI): Ditto.

2010-01-21  Doug Kwan  <dougkwan@google.com>

elfcpp/ChangeLog:

* arm.h (EXIDX_CANTUNWIND): New enum.

gold/ChangeLog:

* arm.cc (Arm_exidx_cantunwind, Arm_exidx_merged_section): New
classes.
(Arm_exidx_section_offset_map): New type.

2010-01-21  Doug Kwan  <dougkwan@google.com>

* arm.cc (Arm_exidx_input_section): New class.
(Arm_relobj::exidx_input_section_by_link,
Arm_relobj::exidx_input_section_by_shndx,
Arm_relobj::make_exidx_input_section): New methods.
(read_arm_attributes_section): Remove.
(Arm_relobj::do_read_symbols): Look for ARM.exidx sections and record
information about them.
(Arm_dynobj::do_read_symbols): Move code in read_arm_attributes_section
to here.

*** empty log message ***

daily update

* linux-ppc-low.c (ppc_arch_setup): Adjust to regcache
          changes.

* config/te-solaris.h (DWARF2_EH_FRAME_READ_ONLY): Make read-only
on 64-bit Solaris/x86.
Include obj-format.h earlier.

gdb/gdbserver/
* linux-s390-low.c (s390_collect_ptrace_register)
(s390_supply_ptrace_register): Adjust it for the new regcache parameter.

* linux-low.c (PTRACE_ARG3_TYPE): Change from long to void*.
(PTRACE_ARG4_TYPE): New macro.
(handle_extended_wait): Cast ptrace arg4 to PTRACE_ARG4_TYPE.
(linux_wait_for_event_1, linux_resume_one_lwp): Ditto.
(fetch_register): Cast to uintptr_t before casting to PTRACE_ARG3_TYPE.
(usr_store_inferior_registers): Ditto.
(linux_read_memory, linux_write_memory): Ditto.
(linux_test_for_tracefork): Ditto.

* linux-arm-low.c: Remove redundant include of gdb_proc_service.h.
Only include elf.h if gdb_proc_service.h didn't include linux/elf.h.

Correct month.

Add xsave64 and xrstor64.

gas/testsuite/

2010-02-21  H.J. Lu  <hongjiu.lu@intel.com>

* gas/i386/x86-64-xsave.s: Add tests for xsave64 and xrstor64.

* gas/i386/x86-64-xsave-intel.d: Updated.
* gas/i386/x86-64-xsave.d: Likewise.

opcodes/

2010-02-21  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (mod_table): Use FXSAVE on xsave and xrstor.

* i386-opc.tbl: Add xsave64 and xrstor64.
* i386-tbl.h: Regenerated.

PR symtab/11198:
* symtab.h (lookup_minimal_symbol_and_objfile): Declare.
* minsyms.c (lookup_minimal_symbol_and_objfile): New function.
* glibc-tdep.c (find_minsym_and_objfile): Remove.
(glibc_skip_solib_resolver): Use
lookup_minimal_symbol_and_objfile.