* lib/mi-support.exp (mi_expect_stop): New.
(mi_run_cmd): Change the
token.  Use mi_send_resuming_command, use
mi_expect_stop.
(mi_execute_to_helper): Rename to mi_execute_to.
(mi_send_resuming_command): Add more error patterns.
(mi_wait_for_stop): Renamed to...
(mi_get_stop_line): ...this.
(mi_run_inline_test): Adjust.

* gdb.mi/mi-cli.exp: Use mi_execute_to/mi_expect_stop.
* gdb.mi/mi-console.exp: Likewise.
* gdb.mi/mi-pending.exp: Likewise.
* gdb.mi/mi-simplerun.exp: Likewise.
* gdb.mi/mi-stack.exp: Likewise.
* gdb.mi/mi-stepi.exp: Likewise.
* gdb.mi/mi-syn-frame.exp: Add comment.
* gdb.mi/mi-until.exp: Likewise.
* gdb.mi/mi-var-display.exp: Likewise.
* gdb.mi/mi-watch.exp: Likewise.
* gdb.mi/mi2-cli.exp: Likewise.
* gdb.mi/mi2-console.exp: Likewise.
* gdb.mi/mi2-simplerun.exp: Likewise.
* gdb.mi/mi2-stack.exp: Likewise.
* gdb.mi/mi2-stepi.exp: Likewise.
* gdb.mi/mi2-until.exp: Likewise.
* gdb.mi/mi2-var-display.exp: Likewise.
* gdb.mi/mi2-watch.exp: Likewise.

*** empty log message ***

2008-04-04  Adrian Bunk  <bunk@stusta.de>
    Bob Wilson  <bob.wilson@acm.org>

* config/tc-xtensa.c (xg_apply_fix_value): Check return code from
call to decode_reloc.

daily update

* gdb.base/watch_thread_num.exp: Use gdb_compile_pthreads.

* gdb.cp/mb-templates.exp: Skip test on spu*-*-* targets.

gdb/
2008-04-04  Pedro Alves  <pedro@codesourcery.com>

* mi/mi-cmds.h (enum mi_cmd_result): Delete MI_CMD_ERROR.
(mi_error_message): Delete declaration.
* mi/mi-interp.c (mi_cmd_interpreter_exec): Call error instead of
returning MI_CMD_ERROR.
* mi/mi-main.c (mi_error_message): Delete.
(mi_cmd_exec_interrupt):
(mi_cmd_thread_select, mi_cmd_thread_list_ids)
(mi_cmd_thread_info): Call error instead of returning
MI_CMD_ERROR.
(mi_cmd_data_list_register_values): Call error instead of
returning MI_CMD_ERROR.  Adapt to new get_register interface.
(get_register): Change return typo to void.  Call error instead of
returning MI_CMD_ERROR.
(mi_cmd_data_write_register_values): Call error instead of
returning MI_CMD_ERROR.
(mi_cmd_list_features): Return MI_CMD_DONE.
(captured_mi_execute_command): Remove MI_CMD_ERROR handling.
(mi_execute_command): Always print exceptions with -error.

gdb/testsuite/
2008-04-04  Pedro Alves  <pedro@codesourcery.com>

* gdb.mi/mi-disassemble.exp, gdb.mi/mi-stack.exp,
gdb.mi/mi-syn-frame.exp, gdb.mi/mi-var-block.exp,
gdb.mi/mi-var-cmd.exp, gdb.mi/mi-var-display.exp,
gdb.mi/mi2-disassemble.exp, gdb.mi/mi2-stack.exp,
gdb.mi/mi2-syn-frame.exp, gdb.mi/mi2-var-block.exp,
gdb.mi/mi2-var-cmd.exp, gdb.mi/mi2-var-display.exp: Update to not
expect an mi error duplicated in stderr.

* server.c (handle_query): Never return "unsupported" for
qXfer:features:read queries.

2008-04-04  Cary Coutant  <ccoutant@google.com>

* symtab.h (Symbol::is_weak_undefined): New function.
(Symbol::is_strong_undefined): New function.
(Symbol::is_absolute): New function.
(Symbol::needs_plt_entry): Exclude weak undefined symbols.
(Symbol::needs_dynamic_reloc): Exclude weak undefined and
absolute symbols.
* testsuite/Makefile.am (check_PROGRAMS): Add weak_undef_test.
(weak_undef_test): New target.
* testsuite/Makefile.in: Rebuild.
* testsuite/weak_undef_file1.cc: New file.
* testsuite/weak_undef_file2.cc: New file.
* testsuite/weak_undef_test.cc: New file.

gas/

2008-04-04  H.J. Lu  <hongjiu.lu@intel.com>

* NEWS: Mention XSAVE.  Change CLMUL to PCLMUL.

* config/tc-i386.c (cpu_arch): Add .pclmul.
(md_show_usage): Replace clmul with pclmul.
* doc/c-i386.texi: Likewise.

gas/testsuite/

2008-04-04  H.J. Lu  <hongjiu.lu@intel.com>

* gas/i386/arch-10-1.l: Replace CLMUL with PCLMUL.
* gas/i386/arch-10-2.l: Likewise.
* gas/i386/arch-10-3.l: Likewise.
* gas/i386/arch-10-4.l: Likewise.
* gas/i386/arch-10.s: Likewise.
* gas/i386/clmul-intel.d: Likewise.
* gas/i386/clmul.d: Likewise.
* gas/i386/clmul.s: Likewise.
* gas/i386/x86-64-arch-2.s: Likewise.
* gas/i386/x86-64-clmul-intel.d: Likewise.
* gas/i386/x86-64-clmul.d: Likewise.
* gas/i386/x86-64-clmul.s: Likewise.

* gas/i386/arch-10.d: Replace clmul with pclmul.
* gas/i386/x86-64-arch-2.d: Likewise.

opcodes/

2008-04-04  H.J. Lu  <hongjiu.lu@intel.com>

* i386-gen.c (cpu_flag_init): Replace CPU_CLMUL_FLAGS/CpuCLMUL
with CPU_PCLMUL_FLAGS/CpuPCLMUL.
(cpu_flags): Replace CpuCLMUL with CpuPCLMUL.
* i386-opc.tbl: Likewise.

* i386-opc.h (CpuCLMUL): Renamed to ...
(CpuPCLMUL): This.
(CpuFMA): Updated.
(i386_cpu_flags): Replace cpuclmul with cpupclmul.

* i386-init.h: Regenerated.

    * NEWS: Mention new commands set/show multiple-symbols.

    PR binutils/4334
        * acx.m4 (ACX_CHECK_CYGWIN_CAT_WORKS): New macro to check that
        cygwin builds are not running in textmode.

        * configure.ac: Run ACX_CHECK_CYGWIN_CAT_WORKS for cygwin hosted
        builds.
        * configure: Regenerate.

PR other/35151
        * configure.ac: Combine rules for mingw32 and mingw64.
        * configure: Regenerate.

PR other/35151
       * configure.ac: Combine rules for mingw32 and mingw64.
       * configure: Regenerate.

* po/sk.po: Updated Slovak translation.

*** empty log message ***

* elfcpp_swap.h (Swap_unaligned<64, true>::writeval): Correct
byte order.

daily update

    * gdb.texinfo (Breakpoint Menus): Delete. Contents moved inside
        new node "Ambiguous Expressions".  Replace references to this
        node by references to "Ambiguous Expressions" throughout.
        (Ambiguous Expressions): New node.

    * gdb.cp/ovldbreak.cc: Add missing bodies for methods foo::foofunc.
        * gdb.cp/ovldbreak.exp: Set multiple-symbols to "ask".
        Add a couple of tests that verify the behavior when the new setting
        is set to "cancel" and "all".
        * gdb.cp/method2.exp, gdb.cp/templates.exp: Set multiple-symbols to
        "ask" before we start the testing.

    * symtab.c (multiple_symbols_ask, multiple_symbols_all)
        (multiple_symbols_cancel): New constants.
        (multiple_symbols_modes, multiple_symbols_mode): New static globals.
        (multiple_symbols_select_mode): New function.
        (_initialize_symtab): Add new set/show multiple-symbols commands.
        * symtab.h (multiple_symbols_ask, multiple_symbols_all)
        (multiple_symbols_cancel, multiple_symbols_select_mode): Declare.
        * ada-lang.c (user_select_syms): Add handling of new multiple-symbols
        setting.
        * linespec.c (decode_line_2): Likewise.

* symtab.h (enum free_code): Delete free_contents, unused.
* symmisc.c (free_symtab_block): Delete.
(free_symtab, case free_code): Delete.

Kai Tietz  <kai.tietz@onevision.com>

PR ld/6026
* pe-dll.c (make_head): Fix iat and thunk addend offset.

* compressed_output.h (class Output_compressed_section): Use
unsigned buffer.
* compressed_output.cc (zlib_compress): Use unsigned buffers,
add zlib header.
(zlib_compressed_suffix): Removed.
(Output_compressed_section::set_final_data_size): Use unsigned
buffers.
* testsuite/Makefile.am (flagstest_compress_debug_sections):
Fix linker invocation.
(flagstest_o_specialfile_and_compress_debug_sections):
Likewise.
* testsuite/Makefile.in: Regenerated.

binutils/

2008-04-03  H.J. Lu  <hongjiu.lu@intel.com>

* dwarf.c (dwarf_regnames_i386): Add AVX registers.
(dwarf_regnames_x86_64): Likewise.

gas/

2008-04-03  H.J. Lu  <hongjiu.lu@intel.com>

* NEWS: Mention AES, CLMUL, AVX/FMA and -msse2avx.

* doc/c-i386.texi: Add avx, aes, clmul and fma to -march=.
Document -msse2avx, .avx, .aes, .clmul and .fma.

* config/tc-i386.c (YMMWORD_MNEM_SUFFIX): New.
(vex_prefix): Likewise.
(sse2avx): Likewise.
(CPU_FLAGS_ARCH_MATCH): Likewise.
(CPU_FLAGS_64BIT_MATCH): Likewise.
(CPU_FLAGS_32BIT_MATCH): Likewise.
(CPU_FLAGS_PERFECT_MATCH): Likewise.
(regymm): Likewise.
(vex_imm4): Likewise.
(fits_in_imm4): Likewise.
(build_vex_prefix): Likewise.
(VEX_check_operands): Likewise.
(bad_implicit_operand): Likewise.
(OPTION_MSSE2AVX): Likewise.
(T_YMMWORD): Likewise.
(_i386_insn): Add vex.
(cpu_arch): Add .avx, .aes, .clmul and .fma.
(cpu_flags_match): Changed to take a pointer to const template.
Enable encoding SSE instructions with VEX prefix for -msse2avx.
(match_mem_size): Also check ymmword.
(operand_type_match): Clear ymmword.
(md_begin): Allow '_' in mnemonic.
(type_names): Add OPERAND_TYPE_VEX_IMM4.
(process_immext): Update assert.
(md_assemble): Don't call process_immext if sse2avx and immext
are true.  Call build_vex_prefix if vex is true.
(parse_insn): Updated for cpu_flags_match.
(swap_operands): Handle 5 operands.
(match_template): Handle 5 operands. Updated for cpu_flags_match.
Check regymm.  Call VEX_check_operands. Handle YMMWORD_MNEM_SUFFIX.
(process_suffix): Handle YMMWORD_MNEM_SUFFIX.
(check_byte_reg): Check regymm.
(process_operands): Duplicate the destination register for
-msse2avx if needed.
(build_modrm_byte): Updated for instructions with VEX encoding.
(output_insn): Output VEX prefix if needed.
(md_longopts): Add msse2avx.
(md_parse_option): Handle OPTION_MSSE2AVX.
(md_show_usage): Add avx, aes, clmul, fma and -msse2avx.
(intel_e09): Support YMMWORD.
(intel_e11): Likewise.
(intel_get_token): Likewise.

gas/testsuite/

2008-04-03  H.J. Lu  <hongjiu.lu@intel.com>

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

* gas/cfi/cfi-i386.s: Add tests for AVX register maps.
* gas/cfi/cfi-x86_64.s: Likewise.

* gas/i386/aes.d: New.
* gas/i386/aes.s: Likewise.
* gas/i386/aes-intel.d: Likewise.
* gas/i386/avx.d: Likewise.
* gas/i386/avx.s: Likewise.
* gas/i386/avx-intel.d: Likewise.
* gas/i386/clmul.d: Likewise.
* gas/i386/clmul-intel.d: Likewise.
* gas/i386/clmul.s: Likewise.
* gas/i386/i386.exp: Likewise.
* gas/i386/inval-avx.l: Likewise.
* gas/i386/inval-avx.s: Likewise.
* gas/i386/sse2avx.d: Likewise.
* gas/i386/sse2avx.s: Likewise.
* gas/i386/x86-64-aes.d: Likewise.
* gas/i386/x86-64-aes.s: Likewise.
* gas/i386/x86-64-aes-intel.d: Likewise.
* gas/i386/x86-64-avx.d: Likewise.
* gas/i386/x86-64-avx.s: Likewise.
* gas/i386/x86-64-avx-intel.d: Likewise.
* gas/i386/x86-64-clmul.d: Likewise.
* gas/i386/x86-64-clmul-intel.d: Likewise.
* gas/i386/x86-64-clmul.s: Likewise.
* gas/i386/x86-64-inval-avx.l: Likewise.
* gas/i386/x86-64-inval-avx.s: Likewise.
* gas/i386/x86-64-sse2avx.d: Likewise.
* gas/i386/x86-64-sse2avx.s: Likewise.

* gas/i386/arch-10.s: Add tests for AVX, AES, CLMUL and FMA.
* gas/i386/x86-64-arch-2.s: Likewise.

* gas/i386/rexw.s: Add AVX tests.

* gas/i386/x86-64-opcode-inval.s: Remove lds/les test.

* gas/cfi/cfi-i386.d: Updated.
* gas/cfi/cfi-x86_64.d: Likewise.
* gas/i386/arch-10.d:  Likewise.
* gas/i386/arch-10-1.l: Likewise.
* gas/i386/arch-10-2.l: Likewise.
* gas/i386/arch-10-3.l: Likewise.
* gas/i386/arch-10-4.l: Likewise.
* gas/i386/rexw.d: Likewise.
* gas/i386/x86-64-arch-2.d: Likewise.
* gas/i386/x86-64-opcode-inval.d: Likewise.
* gas/i386/x86-64-opcode-inval-intel.d: Likewise.

include/opcode/

2008-04-03  H.J. Lu  <hongjiu.lu@intel.com>

* i386.h (MAX_OPERANDS): Set to 5.
(MAX_MNEM_SIZE): Changed to 20.

opcodes/

2008-04-03  H.J. Lu  <hongjiu.lu@intel.com>

* i386-dis.c (OP_E_register): New.
(OP_E_memory): Likewise.
(OP_VEX): Likewise.
(OP_EX_Vex): Likewise.
(OP_EX_VexW): Likewise.
(OP_XMM_Vex): Likewise.
(OP_XMM_VexW): Likewise.
(OP_REG_VexI4): Likewise.
(PCLMUL_Fixup): Likewise.
(VEXI4_Fixup): Likewise.
(VZERO_Fixup): Likewise.
(VCMP_Fixup): Likewise.
(VPERMIL2_Fixup): Likewise.
(rex_original): Likewise.
(rex_ignored): Likewise.
(Mxmm): Likewise.
(XMM): Likewise.
(EXxmm): Likewise.
(EXxmmq): Likewise.
(EXymmq): Likewise.
(Vex): Likewise.
(Vex128): Likewise.
(Vex256): Likewise.
(VexI4): Likewise.
(EXdVex): Likewise.
(EXqVex): Likewise.
(EXVexW): Likewise.
(EXdVexW): Likewise.
(EXqVexW): Likewise.
(XMVex): Likewise.
(XMVexW): Likewise.
(XMVexI4): Likewise.
(PCLMUL): Likewise.
(VZERO): Likewise.
(VCMP): Likewise.
(VPERMIL2): Likewise.
(xmm_mode): Likewise.
(xmmq_mode): Likewise.
(ymmq_mode): Likewise.
(vex_mode): Likewise.
(vex128_mode): Likewise.
(vex256_mode): Likewise.
(USE_VEX_C4_TABLE): Likewise.
(USE_VEX_C5_TABLE): Likewise.
(USE_VEX_LEN_TABLE): Likewise.
(VEX_C4_TABLE): Likewise.
(VEX_C5_TABLE): Likewise.
(VEX_LEN_TABLE): Likewise.
(REG_VEX_XX): Likewise.
(MOD_VEX_XXX): Likewise.
(PREFIX_0F38DB..PREFIX_0F38DF): Likewise.
(PREFIX_0F3A44): Likewise.
(PREFIX_0F3ADF): Likewise.
(PREFIX_VEX_XXX): Likewise.
(VEX_OF): Likewise.
(VEX_OF38): Likewise.
(VEX_OF3A): Likewise.
(VEX_LEN_XXX): Likewise.
(vex): Likewise.
(need_vex): Likewise.
(need_vex_reg): Likewise.
(vex_i4_done): Likewise.
(vex_table): Likewise.
(vex_len_table): Likewise.
(OP_REG_VexI4): Likewise.
(vex_cmp_op): Likewise.
(pclmul_op): Likewise.
(vpermil2_op): Likewise.
(m_mode): Updated.
(es_reg): Likewise.
(PREFIX_0F38F0): Likewise.
(PREFIX_0F3A60): Likewise.
(reg_table): Add REG_VEX_71...REG_VEX_73 and REG_VEX_AE.
(prefix_table): Add PREFIX_0F38DB..PREFIX_0F38DF, PREFIX_0F3ADF
and PREFIX_VEX_XXX entries.
(x86_64_table): Use VEX_C4_TABLE and VEX_C5_TABLE.
(three_byte_table): Use PREFIX_0F38DB..PREFIX_0F38DF and
PREFIX_0F3ADF.
(mod_table): Use VEX_C4_TABLE, VEX_C5_TABLE and VEX_LEN_TABLE.
Add MOD_VEX_XXX entries.
(ckprefix): Initialize rex_original and rex_ignored.  Store the
REX byte in rex_original.
(get_valid_dis386): Handle the implicit prefix in VEX prefix
bytes and USE_VEX_LEN_TABLE/USE_VEX_C4_TABLE/USE_VEX_C5_TABLE.
(print_insn): Set need_vex/need_vex_reg/vex_i4_done to 0 before
calling get_valid_dis386.  Use rex_original and rex_ignored when
printing out REX.
(putop): Handle "XY".
(intel_operand_size): Handle VEX, xmm_mode, xmmq_mode and
ymmq_mode.
(OP_E_extended): Updated to use OP_E_register and
OP_E_memory.
(OP_XMM): Handle VEX.
(OP_EX): Likewise.
(XMM_Fixup): Likewise.
(CMP_Fixup): Use ARRAY_SIZE.

* i386-gen.c (cpu_flag_init): Add CpuAES, CPU_CLMUL_FLAGS,
CPU_FMA_FLAGS and CPU_AVX_FLAGS.
(operand_type_init): Add OPERAND_TYPE_REGYMM and
OPERAND_TYPE_VEX_IMM4.
(cpu_flags): Add CpuAVX, CpuAES, CpuCLMUL and CpuFMA.
(opcode_modifiers): Add Implicit1stXmm0, Vex, Vex256, VexNDD,
VexNDS, VexW0, VexW1, Vex0F, Vex0F38, Vex0F3A, Vex3Sources,
VexImmExt and SSE2AVX.
(operand_types): Add RegYMM, Ymmword and Vex_Imm4.

* i386-opc.h (CpuAVX): New.
(CpuAES): Likewise.
(CpuCLMUL): Likewise.
(CpuFMA): Likewise.
(Vex): Likewise.
(Vex256): Likewise.
(VexNDS): Likewise.
(VexNDD): Likewise.
(VexW0): Likewise.
(VexW1): Likewise.
(Vex0F): Likewise.
(Vex0F38): Likewise.
(Vex0F3A): Likewise.
(Vex3Sources): Likewise.
(VexImmExt): Likewise.
(SSE2AVX): Likewise.
(RegYMM): Likewise.
(Ymmword): Likewise.
(Vex_Imm4): Likewise.
(Implicit1stXmm0): Likewise.
(CpuXsave): Updated.
(CpuLM): Likewise.
(ByteOkIntel): Likewise.
(OldGcc): Likewise.
(Control): Likewise.
(Unspecified): Likewise.
(OTMax): Likewise.
(i386_cpu_flags): Add cpuavx, cpuaes, cpuclmul and cpufma.
(i386_opcode_modifier): Add implicit1stxmm0, vex, vex256,
vexnds, vexndd, vexw0, vexw1, vex0f, vex0f38, vex0f3a,
vex3sources, veximmext and sse2avx.
(i386_operand_type): Add regymm, ymmword and vex_imm4.

* i386-opc.tbl: Add AES, CLMUL, AVX and FMA new instructions.

* i386-reg.tbl: Add AVX registers, ymm0..ymm15.

* i386-init.h: Regenerated.
* i386-tbl.h: Likewise.

*  dwarf_reader.cc (Sized_dwarf_line_info::read_header_prolog,
Sized_dwarf_line_info::process_one_opcode): Use Swap_unaligned.

* TODO: New file.

*** empty log message ***

daily update

* fileread.cc (File_read::find_view): Add byteshift and vshifted
parameters.  Update for new key type to views_.  Change all
callers.
(File_read::read): Adjust for byteshift in returned view.
(File_read::add_view): New function, broken out of
find_and_make_view.
(File_read::make_view): New function, broken out of
find_and_make_view.
(File_read::find_or_make_view): Add offset and aligned
parameters.  Rewrite accordingly.  Change all callers.
(File_read::get_view): Add offset and aligned parameters.  Adjust
for byteshift in return value.
(File_read::get_lasting_view): Likewise.
* fileread.h (class File_read): Update declarations.
(class File_read::View): Add byteshift_ field.  Add byteshift to
constructor.  Add byteshift method.
* archive.h (Archive::clear_uncached_views): New function.
(Archive::get_view): Add aligned parameter.  Change all callers.
* object.h (Object::get_view): Add aligned parameter.  Change all
callers.
(Object::get_lasting_view): Likewise.

* fileread.cc (File_read::release): Don't call clear_views if
there are multiple objects.
* fileread.h (File_read::clear_uncached_views): New function.
* archive.cc (Add_archive_symbols::run): Call clear_uncached_views
on the archive.

*** empty log message ***

* elf32-spu.c (insert_callee): Reorder call list so most recent
call is always first.
(interesting_section): Move.
(mark_functions_via_relocs): Fold interesting_section and
reloc_count tests in callers to here.  Simplify output section
owner test.
(discover_functions): Set "gaps" when no symbols and some
"interesting_section".  Run pasted_function loop for no symbol
bfds.
(for_each_node, transfer_calls): New functions.
(mark_non_root): Adjust to suit for_each_node.
(call_graph_traverse): Likewise.  Fix memory leak.  Rename to..
(remove_cycles): ..this.
(build_call_tree): Use for_each_node and transfer_calls.
(struct _sum_stack_param): New.
(sum_stack): Adjust to suit for_each_node.  Return error on
malloc failure.  Move code to print root node cumulative stack from..
(spu_elf_stack_analysis): ..here.  Use for_each_node.

daily update

2008-03-31  Cary Coutant  <ccoutant@google.com>

Add thin archive support.
* archive.cc (Archive::armagt): New const.
(Archive::setup): Remove task parameter and calls to unlock.
(Archive::unlock_nested_archives): New function.
(Archive::read_header): Add nested_off parameter. Change
all callers.
(Archive::interpret_header): Likewise.
(Archive::include_all_members): Change to handle thin
archives.
(Archive::include_member): Likewise.
* archive.h (Archive::Archive): Add new parameters and
initializers.
(Archive::armagt): New const.
(Archive::setup): Remove task parameter.
(Archive::unlock_nested_archives): New function.
(Archive::read_header): Add nested_off parameter.
(Archive::interpret_header): Likewise.
(Archive::Nested_archive_table): New typedef.
(Archive::is_thin_archive_): New field.
(Archive::nested_archives_): New field.
(Archive::options_): New field.
(Archive::dirpath_): New field.
(Archive::task_): New field.
* readsyms.cc (Read_symbols::do_read_symbols): Add check
for thin archives.  Pass additional parameters to
Archive::Archive.  Unlock the archive file after calling
Archive::setup.

* gdb.cp/casts.cc: Add class reference variables.
* gdb.cp/casts.exp: New test cases for up/down casting references.

* gdb.cp/casts.cc: Add class reference variables.
* gdb.cp/casts.exp: New test cases for up/down casting references.

* valops.c (value_cast_structs): New function. Cast related
STRUCT types up/down and return cast value. The body of this
function comes mostly from value_cast_pointers.
(value_cast_pointers): Code for actual cast STRUCT-STRUCT moved
to value_cast_structs. Now value_cast_pointers needs only create
appropriate reference after using value_cast_structs for actual
casting.
(value_cast): Handle references.

2008-04-01  Marc Khouzam  <marc.khouzam@ericsson.com>

* gdb.mi/mi-var-display.exp: Stop at end of do_special_tests
instead of beginning to make an extra test pass.
* gdb.mi/mi2-var-display.exp: Likewise.

2008-04-01  Marc Khouzam  <marc.khouzam@ericsson.com>

* MAINTAINERS: Added myself to section Write After Approval.

*** empty log message ***

PR 6006
* archive.c (_bfd_slurp_extended_name_table): Change
ARFMAG[0] to ARFMAG[1].
(_bfd_construct_extended_name_table): Likewise.

daily update

merge from gcc

* elfxx-mips.c (mips_elf_record_got_page_entry): Update comment.
(_bfd_mips_elf_check_relocs): Update comments.  Always call
mips_elf_record_got_page_entry for R_MIPS_GOT_PAGE.

* gdb.base/break.exp: Mark failures as such.  Fix formatting.

Updated French translations.

2008-03-31  Markus Deuling  <deuling@de.ibm.com>
    Ulrich Weigand  <Ulrich.Weigand@de.ibm.com>

* gdb.arch/spu-info.exp (info spu mailbox): Check for correct initial
value.  Add xfail for older kernels.

* ia64-tdep.c (examine_prologue): Correct array access.

*** empty log message ***

daily update

*** empty log message ***

daily update

* symtab.cc (Symbol_table::do_define_as_constant): Don't force a
version symbol to be local.
* testsuite/ver_test_4.sh: New file.
* testsuite/Makefile.am (check_SCRIPTS): Add ver_test_4.sh.
(check_DATA): Add ver_test_4.syms.
(ver_test_4.syms): New target.
* testsuite/Makefile.in: Rebuild.

* output.cc
(Output_section::Input_section_sort_entry::has_priority): New
function.
(Output_section::Input_section_sort_entry::match_file_name): New
function.
(Output_section::Input_section_sort_entry::match_section_name):
Remove.
(Output_section::Input_section_sort_entry::match_section_name_prefix):
Remove.
(Output_section::Input_section_sort_entry::match_section_file):
Remove.
(Output_section::Input_section_sort_compare::operator()): Rewrite
using new Input_section_sort_entry functions.  Sort crtbegin and
crtend first.  Sort sections with no priority before sections with
a priority.
* testsuite/initpri1.c (d3): Check j != 4.
(cd5): New constructor/destructor function.
(main): Check j != 2.

* symtab.cc (Symbol_table::add_from_object): If we don't use the
new symbol when resolving, don't call set_is_default.
* testsuite/ver_test_7.cc: New file.
* testsuite/ver_test_7.sh: New file.
* testsuite/Makefile.am (ver_test_7.so): New target.
(ver_test_7.o): New target.
(check_SCRIPTS): Add ver_test_7.sh.
(check_DATA): Add ver_test_7.syms.
(ver_test_7.syms): New target.

*** empty log message ***

daily update

* layout.cc (Layout::layout): If we see an input section with a
name that needs sorting, set the must_sort flag for the output
section.
(Layout::make_output_section): If the name of the output section
indicates that it might require sorting, set the may_sort flag.
* output.h (Output_section::may_sort_attached_input_sections): New
function.
(Output_section::set_may_sort_attached_input_sections): New
function.
(Output_section::must_sort_attached_input_sections): New
function.
(Output_section::set_must_sort_attached_input_sections): New
function.
(class Output_section): Declare Input_section_sort_entry.  Define
Input_section_sort_compare.  Declare
sort_attached_input_sections.  Add new fields:
may_sort_attached_input_sections_,
must_sort_attached_input_sections_,
attached_input_sections_are_sorted_.
* output.cc (Output_section::Output_section): Initialize new
fields.
(Output_section::add_input_section): Add an entry to
input_sections_ if may_sort or must_sort are true.
(Output_section::set_final_data_size): Call
sort_attached_input_sections if necessary.
(Output_section::Input_section_sort_entry): Define new class.
(Output_section::Input_section_sort_compare::operator()): New
function.
(Output_section::sort_attached_input_sections): New function.
* configure.ac: Check whether the compiler supports constructor
priorities.  Define a CONSTRUCTOR_PRIORITY automake conditional.
* testsuite/initpri1.c: New file.
* testsuite/Makefile.am (check_PROGRAMS): Add initpri1 if
CONSTRUCTOR_PRIORITY.
(initpri1_SOURCES, initpri1_DEPENDENCIES): New variables.
(initpri1_LDFLAGS): New variable.
* configure, Makefile.in, testsuite/Makefile.in: Rebuild.

/gas:
2008-03-28  Eric B. Weddington  <eric.weddington@atmel.com>

* config/tc-avr.c (mcu_types): Add attiny167.
* doc/c-avr.texi: Likewise.

/include:
2008-03-28  Eric B. Weddington  <eric.weddington@atmel.com>

* opcode/avr.h (AVR_ISA_TINY3): Define new opcode set for attiny167.

/gas:
2008-03-28  Eric B. Weddington  <eric.weddington@atmel.com>

* config/tc-avr.c (mcu_types): Add atmega32u4.
* doc/c-avr.texi: Likewise.

* cp-support.c (first_component_command): Return if no arguments.

/gas:
2008-03-28  Eric B. Weddington  <eric.weddington@atmel.com>

* config/tc-avr.c (mcu_types): Add atmega32c1.
* doc/c-avr.texi: Likewise.

2008-03-28  Paul Brook  <paul@codesourcery.com>

gas/
* config/tc-arm.c (parse_neon_mov): Parse register before immediate
to avoid spurious symbols.

gdb/

2008-03-28  Carlos O'Donell  <carlos@codesourcery.com>

* ser-mingw.c (ser_windows_open): Open requested name.

2008-03-28  Aleksandar Ristovski  <aristovski@qnx.com>

* MAINTAINERS: Added myself.

* target.c (find_default_run_target): Allow a NULL `do_mesg'
parameter.  If it is NULL, don't call error.
(find_default_can_async_p, find_default_is_async_p): Pass NULL as
`do_mesg' parameter to find_default_run_target.  If no target was
found, return 0.

2008-03-28  Aleksandar Ristovski  <aristovski@qnx.com>

Bring mi-support in line with gdb.exp.
* lib/mi-support.exp (default_mi_gdb_start): Rename from mi_gdb_start.
(mi_gdb_start): New function.

* mips-linux-tdep.c: Update N32/N64 signal frame comments.
(N64_SIGCONTEXT_LO, N64_SIGCONTEXT_PC, N64_SIGCONTEXT_FPCSR): Update.
(N64_SIGCONTEXT_FIR, N64_SIGCONTEXT_CAUSE, N64_SIGCONTEXT_BADVADDR):
Delete.
(mips_linux_n32n64_sigframe_init): Do not record cause or badvaddr.

* gdb.mi/mi-var-cmd.exp: Fix a typo.

* config/tc-m68k.c (md_convert_frag_1): Replace as_fatal with
as_bad_where.

Add support for thin archives.
    * bfd/archive.c (_bfd_find_nested_archive): New function.
    (get_extended_arelt_filename): Add origin parameter.
    (_bfd_generic_read_ar_hdr_mag): Deal with extended name
    combined with a file offset.
    (append_relative_path): New function.
    (_bfd_get_elt_at_filepos): Deal with external members and
    nested archives.
    (bfd_generic_openr_next_archived_file): Thin archives.
    (bfd_generic_archive_p): Recognize new magic string.
    (adjust_relative_path): New function.
    (_bfd_construct_extended_name_table): Construct extended
    names for thin archive members.
    (_bfd_write_archive_contents): Emit new magic string, skip
    copying files for thin archives.
    * bfd/bfd-in.h (bfd_is_thin_archive): New macro.
    * bfd/bfd.c (struct bfd): New fields for thin archives.
    * bfd/libbfd-in.h (struct areltdata): New field for thin archives.
    * bfd/opncls.c (bfd_close): Delete BFDs for nested archives.
    * binutils/ar.c (make_thin_archive): New global flag.
    (map_over_members): Deal with full pathnames in thin archives.
    (usage, main): Add 'T' option for building thin archives.
    (replace_members): Pass thin archive flag to ar_emul_append.
    * binutils/arsup.c (ar_open): Initialize new flag.
    * binutils/binemul.c (ar_emul_append): Add new parameter for
    flattening nested archives.
    (do_ar_emul_default_append): New function.
    (ar_emul_default_append): Factored out recursive code.
    * binutils/binemul.h (ar_emul_default_append): Add new parameter.
    (struct bin_emulation_xfer_struct): New parameter for ar_append.
    * binutils/dlltool.c (gen_lib_file): Initialize thin archive flag.
    * binutils/emul_aix.c (ar_emul_aix_internal): Add new flatten
    parameter, currently unimplemented.
    All callers changed.
    * binutils/objcopy.c (copy_archive): Preserve thin archive flag.
    * binutils/doc/binutils.texi: Update ar documentation.
    * binutils/testsuite/binutils-all/ar.exp: Add thin archive tests.
    * include/aout/ar.h (ARMAGT): New magic string for thin archives.

* ld-elfcomm/elfcomm.exp: Run $READELF not readelf.

*** empty log message ***

daily update

* common.cc (Sort_commons::operator): Correct sorting algorithm.
* testsuite/common_test_1.c: New file.
* testsuite/Makefile.am (check_PROGRAMS): Add common_test_1.
(common_test_1_SOURCES): New variable.
(common_test_1_DEPENDENCIES): New variable.
(common_test_1_LDFLAGS): New variable.

GDB 6.8 released.

* symtab.cc (Symbol_table::add_from_object): Handle saw_undefined_
and commons_ correctly when NAME/VERSION does not override
NAME/NULL.
* testsuite/ver_test_6.c: New file.
* testsuite/Makefile.am (check_PROGRAMS): Add ver_test_6
(ver_test_6_SOURCES, ver_test_6_DEPENDENCIES): New variables.
(ver_test_6_LDFLAGS, ver_test_6_LDADD): New variables.

  * config/tc-avr.c (mcu_types): Add atmega32m1.
  * doc/c-avr.texi: Likewise.

2008-03-27  Paolo Bonzini  <bonzini@gnu.org>

* Makefile.tpl (PICFLAG, PICFLAG_FOR_TARGET): Remove.
* Makefile.in: Regenerate.

config:
2008-03-27  Paolo Bonzini  <bonzini@gnu.org>

* extensions.m4: New.

2008-03-27  Paolo Bonzini  <bonzini@gnu.org>

* mh-armpic: Remove.
* mh-i370pic: Remove.
* mh-m68kpic: Remove.
* mh-ppcpic: Remove.
* mh-sparcpic: Remove.
* mh-ia64pic: Remove.
* mh-papic: Remove.
* mh-s390pic: Remove.
* mh-x86pic: Remove.

libiberty:
2008-03-27  Paolo Bonzini  <bonzini@gnu.org>

* configure.ac (frags): Don't set, use frag instead.
(PICFLAG): Set here and substitute.
* Makefile.in (PICFLAG): Substitute from autoconf.
* configure: Regenerate.

        * config/tc-arm.c (do_neon_cvt): Move variable declarations to
            start of block.
            (do_neon_ext): Fix sign of comparison.

ChangeLog:

* features/Makefile (%.dat): Set xmltarget to the base filename
of the XML source, without subdirectory.
* regformats/rs6000/powerpc-32.dat: Regenerate.
* regformats/rs6000/powerpc-64.dat: Regenerate.
* regformats/rs6000/powerpc-e500.dat: Regenerate.

gdbserver/ChangeLog:

* server.c (get_features_xml): Fix inverted condition.
(handle_query): Always support qXfer:feature:read.

* xcoffread.c (scan_xcoff_symtab): Replace current_gdbarch by
objfile arch.

* symtab.cc (Symbol_table::add_from_relobj): Don't set the version
of an undefined symbol from a version script.
* testsuite/Makefile.am (ver_test_5.so): New target.
(ver_test_5.o): New target.
(check_SCRIPTS): Add ver_test_5.sh.
(check_DATA): Add ver_test_5.syms.
(ver_test_5.syms): New target.
* testsuite/ver_test_5.cc: New file.
* testsuite/ver_test_5.script: New file.
* testsuite/ver_test_5.sh: New file.
* Makefile.in, testsuite/Makefile.in: Rebuild.

(enum captured_mi_execute_command_actions):
Spell suppress in EXECUTE_COMMAND_SUPPRESS_PROMPT correctly.

*** empty log message ***

*** empty log message ***

PR gold/5986
Fix problems building gold with gcc 4.3.0.
* gold.h (TEMPLATE_ATTRIBUTE_PRINTF_4): Define.
(gold_error_at_location, gold_warning_at_location): Use it.
* configure.ac: Check whether we can compile and use a template
function with a printf attribute.
* x86_64.cc (Target_x86_64::do_code_fill): Swap out a 32-bit value
when jumping over bytes.
* object.cc: Instantiate Object::read_section_data.
* debug.h: Include <cstring>
* dwarf_reader.cc: Include <algorithm>
* main.cc: Include <cstring>.
* options.cc: Include <cstring>.
* output.cc: Include <cstring>.
* script.cc: Include <cstring>.
* script.h: Include <string>.
* symtab.cc: Include <cstring> and <algorithm>.
* target-select.cc: Include <cstring>.
* version.cc: Include <string>.
* testsuite/testmain.cc: Include <cstdlib>.
* configure, config.in: Rebuild.

daily update

* gdb.texinfo (MIPS Features, PowerPC Features): Add @node.

gas/testsuite/:
From Robin Getz  <rgetz@blackfin.uclinux.org>
* gas/bfin/arithmetic.d: Update to reflect spaces/capitalization in
recent changes in opcodes/bfin-dis.c.
gas/bfin/arithmetic.s: Likewise.
gas/bfin/bit.d: Likewise.
gas/bfin/bit2.d: Likewise.
gas/bfin/control_code.d: Likewise.
gas/bfin/control_code2.d: Likewise.
gas/bfin/event.d: Likewise.
gas/bfin/event2.d: Likewise.
gas/bfin/flow.d: Likewise.
gas/bfin/flow2.d: Likewise.
gas/bfin/load.d: Likewise.
gas/bfin/logical.d: Likewise.
gas/bfin/logical2.d: Likewise.
gas/bfin/move.d: Likewise.
gas/bfin/move2.d: Likewise.
gas/bfin/parallel.d: Likewise.
gas/bfin/parallel2.d: Likewise.
gas/bfin/parallel3.d: Likewise.
gas/bfin/parallel4.d: Likewise.
gas/bfin/shift.d: Likewise.
gas/bfin/shift2.d: Likewise.
gas/bfin/stack.d: Likewise.
gas/bfin/stack2.d: Likewise.
gas/bfin/store.d: Likewise.
gas/bfin/vector.d: Likewise.
gas/bfin/vector2.d: Likewise.
gas/bfin/video.d: Likewise.
gas/bfin/video2.d: Likewise.

opcodes/:
* bfin-dis.c: (c_uimm4s4d, c_imm5d, c_imm7d, c_imm16d, c_uimm16s4d,
c_imm32, c_huimm32e): Define.
(constant_formats): Add flags for printing decimal, leading spaces, and
exact symbols.
(comment, parallel): Add global flags in all disassembly.
(fmtconst): Take advantage of new flags, and print default in hex.
(fmtconst_val): Likewise.
(decode_macfunc): Be consistant with spaces, tabs, comments,
capitalization in disassembly, fix minor coding style issues.
(reg_names, amod0, amod1, amod0amod2, aligndir, get_allreg): Likewise.
(decode_ProgCtrl_0, decode_PushPopMultiple_0, decode_CCflag_0,
decode_CC2dreg_0, decode_CC2stat_0, decode_BRCC_0, decode_UJUMP_0,
decode_REGMV_0, decode_ALU2op_0, decode_PTR2op_0, decode_LOGI2op_0,
decode_COMP3op_0, decode_COMPI2opD_0, decode_COMPI2opP_0,
decode_LDSTpmod_0, decode_dagMODim_0, decode_dagMODik_0,
decode_dspLDST_0, decode_LDST_0, decode_LDSTiiFP_0, decode_LDSTii_0,
decode_LoopSetup_0, decode_LDIMMhalf_0, decode_CALLa_0,
decode_LDSTidxI_0, decode_linkage_0, decode_dsp32alu_0,
decode_dsp32shift_0, decode_dsp32shiftimm_0, decode_pseudodbg_assert_0,
_print_insn_bfin, print_insn_bfin): Likewise.

gas/:
* config/tc-bfin.c (bfin_start_line_hook): Localize the labels
generated for LOOP_BEGIN and LOOP_END instructions.
(bfin_gen_loop): Likewise.

gas/testsuite/:
* gas/bfin/flow.d: Adjust since the generated labels for LOOP_BEGIN
and LOOP_END instruction are local now.
* gas/bfin/flow2.d: Likewise.

gas/
* config/bfin-parse.y (check_macfunc_option): Allow (IU)
option for multiply and multiply-accumulate to data register
instruction.
(check_macfuncs): Don't check if accumulator matches the data register
here.
(assign_macfunc): Check if accumulator matches the
data register in each rule that moves to the data
register.

gas/testsuite/
* gas/bfin/arithmetic.s, gas/bfin/arithmetic.d: Add check
for IU option.
* gas/bfin/expected_errors.l, gas/bfin/expected_errors.s:
Add check for mismatch of accumulator and data register.

opcodes/
* bfin-dis.c (decode_dsp32mac_0): Decode (IU) option for
multiply and multiply-accumulate to data register instruction.

* ld-elf/flags1.d: Adjust for MIPS text alignment.

gas/:
* config/bfin-parse.y (check_macfunc_option): New.
  (check_macfuncs): Check option by calling check_macfunc_option.
Fix comparison always true warnings.  Both scalar instructions
of vector instruction must share the same mode option.  Only allow
option mode at the end of the second instruction of the vector.
  (asm_1): Check option by calling check_macfunc_option.

gas/testsuite/:
* gas/bfin/expected_errors.l, gas/bfin/expected_errors.s: Add
tests for bad options of "multiply and multipy-accumulate to
accumulator" instructions.  Add new vector instruction option
mode tests.
* gas/bfin/vector2.s: Add new vector instruction option mode test.
* gas/bfin/vector2.d: Adjust accordingly.

* gas/bfin/expected_errors.s, gas/bfin/expected_errors.l: Add test
for mismatched half registers in vector multipy-accumulate
instructions.

gas/
From Jie Zhang  <jie.zhang@analog.com>
* config/bfin-parse.y (asm_1): Check AREGS in comparison
instructions. And call yyerror () when comparing PREG with
DREG.

gas/testsuite/:
* gas/bfin/expected_comparison_errors.l: New test.
* gas/bfin/expected_comparison_errors.s: New test.
* gas/bfin/bfin.exp: Add expected_comparison_errors.

* objfiles.h (struct objfile): New GDBARCH member.
(get_objfile_arch): Add prototype.
* objfiles.c: Include "arch-utils.h".
(allocate_objfile): Look up gdbarch associated with bfd.
(get_objfile_arch): New function.
* Makefile (objfiles.o): Update dependencies.

* dwarf2-frame.c (decode_frame_entry_1): Replace current_gdbarch
by objfile arch.
* dwarf2loc.c (dwarf_expr_read_reg): Replace current_gdbarch
by frame arch.
(locexpr_describe_location): Replace current_gdbarch by
objfile arch.
* dwarf2read.c (die_type): Replace current_gdbarch by objfile arch.
(dwarf2_add_field): Likewise.
(read_tag_pointer_type): Likewise.
(read_base_type): Likewise.
(new_symbol): Likewise.

* coffread.c (decode_type): Add OBJFILE argument.  Update callers.
(decode_base_type, decode_function_type): Likewise.
(coff_read_struct_type, coff_read_enum_type): Likewise.
(coff_symtab_read): Replace current_gdbarch by objfile arch.
(decode_base_type): Likewise.
(coff_read_enum_type): Likewise.
(coff_read_struct_type): Replace current_objfile by OBJFILE argument.
(coff_read_enum_type): Likewise.

* dbxread.c (read_dbx_symtab): Replace current_gdbarch by objfile arch.
(end_psymtab): Likewise.
(process_one_symbol): Likewise.

* mdebugread.c (parse_symbol): Replace current_gdbarch by objfile arch.
(parse_procedure): Likewise.
(parse_partial_symbols): Likewise.

* somread.c (som_symtab_read): Replace current_gdbarch by objfile arch.

* stabsread.c (define_symbol): Replace current_gdbarch by objfile arch.
Replace static pcc_promotion_type and pcc_unsigned_promotion_type by
built-in types.
(read_range_type): Replace current_gdbarch by objfile arch.  Replace
static range_index_type by built-in type.
(read_one_struct_field): Replace current_gdbarch by objfile arch.
(read_enum_type): Likewise.

* xcoffread.c (read_xcoff_symtab): Replace current_gdbarch by
objfile arch.

* varobj.h (varobj_floating_p): Declare.
* varobj.c (varobj_floating_p): New.
* mi/mi-cmd-var.c (mi_cmd_var_update): When passed
'@' as the name, update all floating varobjs.

opcodes:
From  Robin Getz  <robin.getz@analog.com>
* bfin-dis.c (bu32): Typedef.
(enum const_forms_t): Add c_uimm32 and c_huimm32.
(constant_formats[]): Add uimm32 and huimm16.
(fmtconst_val): New.
(uimm32): Define.
(huimm32): Define.
(imm16_val): Define.
(luimm16_val): Define.
(struct saved_state): Define.
(GREG, DPREG, DREG, PREG, SPREG, FPREG, IREG, MREG, BREG, LREG,
A0XREG, A0WREG, A1XREG, A1WREG,CCREG, LC0REG, LT0REG, LB0REG,
LC1REG, LT1REG, LB1REG, RETSREG, PCREG): Define.
(get_allreg): New.
(decode_LDIMMhalf_0): Print out the whole register value.

gas/testsuite:
From Jie Zhang  <jie.zhang@analog.com>
* gas/bfin/load.d: Update.

* varobj.c (struct varobj_root): Rename use_selected_frame to
floating, and clarify the meaning.
(varobj_create, varobj_update,  new_root_variable): Adjust.
(value_of_root): Don't use type_changed as in variable,
adjust comment.
(c_value_of_root): Adjust.

* gdb.mi/mi-var-display.exp: Remove obsolete xfail.
* gdb.mi/mi2-var-display.exp: Likewise.

* lib/mi-support.exp (mi_create_varobj_checked): New.
(mi_list_varobj_children): Allow to check for a
value.
(mi_list_array_varobj_children): New.

* gdb.mi/mi-var-child.exp: Use mi_create_varobj
and mi_list_varobj_children, as opposed to hardcoding
expected strings.
* gdb.mi/gdb701.exp: Likewise.
* gdb.mi/gdb792.exp: Likewise.
* gdb.mi/mi-var-block.exp: Likewise.
* gdb.mi/mi-var-cmd.exp: Likewise.
* gdb.mi/mi-var-invalidate.exp: Likewise.
* gdb.mi/mi2-var-block.exp: Likewise.
* gdb.mi/mi2-var-child.exp: Likewise.
* gdb.mi/mi2-var-cmd.exp: Likewise.
* gdb.mi/mi2-var-display.exp: Likewise.