binutils/

2076-02-02  H.J. Lu  <hongjiu.lu@intel.com>

* doc/binutils.texi (objdump): Document the new addr64 option
for i386 disassembler.

include/

2076-02-02  H.J. Lu  <hongjiu.lu@intel.com>

* dis-asm.h (print_i386_disassembler_options): New.

opcodes/

2076-02-02  H.J. Lu  <hongjiu.lu@intel.com>

* disassemble.c (disassembler_usage): Call
print_i386_disassembler_options for i386 disassembler.

* i386-dis.c (print_i386_disassembler_options): New.
(print_insn): Support the new addr64 option.

2076-02-02  H.J. Lu  <hongjiu.lu@intel.com>

PR binutils/3945
* configure.in (DLLTOOL_DEFAULT): New. Defined for the first PE
target.
(DLLTOOL_DEFS): Add $DLLTOOL_DEFAULT.
* configure: Regenerated.

* dlltool.c (mname): Defined with DLLTOOL_DEFAULT_XXX.

PR gas/3812
* gas/v850/v850e1.d: Correct expected disassembly of cmov insn to account for sign extension.

* doc/binutils.texi (objdump): Document support for disassembling the 440 PowerPC architecture.

2007-02-02  Denis Pilat  <denis.pilat@st.com>

* gdb.threads/threadapply.exp: check that frame is not changed by
the thread apply all command.

* ppc-dis.c (powerpc_dialect): Handle ppc440.
* ppc-dis.c (print_ppc_disassembler_options): Note the -M440 can be used.

2007-02-02  Denis Pilat  <denis.pilat@st.com>

* thread.c (make_cleanup_restore_current_thread): New function.
(info_threads_command): Use of make_cleanup_restore_current_thread
to restore the current thread and the selected frame.
(restore_selected_frame): New function.
(struct current_thread_cleanup): Add frame_id field.
(do_restore_current_thread_cleanup): Add restoring of the selected
frame.
(make_cleanup_restore_current_thread): Likewise.
(thread_apply_all_command): backup the selected frame while
entering the function and restore it at exit.
(thread_apply_command): Likewise.

2007-02-02  Denis Pilat  <denis.pilat@st.com>

* MAINTAINERS (Write After Approval): Add myself to the list.

* ppc-opc.c (insert_bdm): -Many comment.
(valid_bo): Add "extract" param.  Accept both powerpc and power4
BO fields when disassembling with -Many.
(insert_bo, extract_bo, insert_boe, extract_boe): Adjust valid_bo call.

*** empty log message ***

* config/tc-xtensa.c (xg_add_opcode_fix, md_apply_fix): Delete use of
fx_tcbit.
* config/tc-xtensa.h (TC_FORCE_RELOCATION_LOCAL): Remove.

daily update

* write.h (struct fix <fx_pcrel_adjust, fx_size>): Move.
(struct fix <fx_plt>): Rename to tcbit2.
* write.c (fix_new_internal): Adjust.
(TC_FORCE_RELOCATION_LOCAL): Don't test fx_plt.
* config/tc-arm.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-cris.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-i386.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-i960.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-sh.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-sh64.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-sparc.h (TC_FORCE_RELOCATION_LOCAL): Likewise.
* config/tc-msp430.c (msp430_force_relocation_local): Likewise.
* config/tc-ia64.c (emit_one_bundle): Don't set fx_plt.
* config/tc-ia64.h (TC_FORCE_RELOCATION_LOCAL): Don't test fx_plt.
Instead, compare fx_r_type.
* config/tc-xtensa.c (xg_add_opcode_fix, md_apply_fix): Use
fx_tcbit in place of fx_plt.
* config/tc-xtensa.h (TC_FORCE_RELOCATION_LOCAL): Define.
* doc/internals.texi (TC_FORCE_RELOCATION_LOCAL): Remove reference
to fx_plt.

PR ld/3852
* elf.c (_bfd_elf_link_hash_table_init): Initialize all the fields in the
elf_link_hash_table structure.

* elf-bfd.h (struct elf_backend_data): Add elf_backend_write_core_note.
* elfxx-target.h (elf_backend_write_core_note): Define and use.
* elf.c (elfcore_write_prpsinfo): Call the above.  Add support for
32-bit core note on 64-bit target.
(elfcore_write_prstatus): Likewise.
(elfcore_write_lwpstatus): Make note_name const.
(elfcore_write_prfpreg): Likewise.
(elfcore_write_pstatus): Add support for 32-bit core note on 64-bit
target.
* elf32-ppc.c (ppc_elf_write_core_note): New function.
(elf_backend_write_core_note): Define.
* elf64-ppc.c (ppc64_elf_write_core_note): New function.
(elf_backend_write_core_note): Define.

* ld-scripts/default-script.exp: Add "--local-store 0:0" to
LDFLAGS for spu.

2006-02-01  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>

* gdb/remote-m32r.sdi.c (m32r_fetch_register): Change PWD mask.
(m32r_store_register): Ditto.

*** empty log message ***

daily update

* Makefile.tpl (LDFLAGS): Substitute it.
* Makefile.in: Regenerate.

merge from gcc

* gdb.gdb/selftest.exp (do_steps_and_nexts): Add more matches.

    include/
        * libiberty.h (PEX_STDERR_TO_PIPE): New define.
        (PEX_BINARY_ERROR): New define.
        (pex_read_err): New function.

        libiberty/
        * pex-common.h (struct pex_obj): New fields
        stderr_pipe and read_err.
        * pex-common.c (pex_init_common): Initialize
        stderr_pipe.
        (pex_run_in_environment): Add error checking
        for PEX_STDERR_TO_PIPE.  Create a pipe
        for stderr if necessary.
        (pex_read_err): New.
        (pex_free): Close read_err.
        * pexecute.txh: Document changes.
        * functions.texi: Regenerated.

* gdb.dwarf2/dup-psym.S: Use .2byte/.4byte instead of .short/.int
in the dwarf sections.
* gdb.dwarf2/mac-fileno.S: Likewise.

* gdb.base/callfuncs.exp (do_get_all_registers): Filter out
bspstore register on ia64.

2076-01-31  H.J. Lu  <hongjiu.lu@intel.com>

* elf32-cris.c (INCLUDED_TARGET_FILE): Removed.
(elf32_bed): Defined for elf32-us-cris.

* elf64-sh64.c (elf64_bed): Defined for Linux.
(INCLUDED_TARGET_FILE): Removed.

* elfxx-target.h (elfNN_bed): Always define. Don't check
INCLUDED_TARGET_FILE.

* dwarf.c (process_debug_info): Protect against bogus length and
abbrev offsets.

* elf-m10300.c (mn10300_elf_relocate_section): Clarify the warning
message for dangerous relocs, special case the common user error.

*** empty log message ***

merge from gcc

* strsignal.c (psignal): Change type of signo to int.
* functions.texi: Regenerate.

daily update

2007-01-30  H.J. Lu  <hongjiu.lu@intel.com>

* elf.c (copy_elf_program_header): Start from the first section
in a segment and stop when all sections in a segment are
accounted for.

* as.c (main): Mark symbols created via the --defsym command line
  option as volatile so that they can be overridden later on by a
  .set directive.  This maintains compatibility with the behaviour
  of earlier versions of the assembler.
* doc/as.texinfo (--defsym): Document that the defined symbol's
  value can be overridden via a .set directive.

    * ser-mingw.c (pipe_windows_open)
        (pipe_windows_read, pipe_windows_write): Declare
        variables at the top of the function.

*** empty log message ***

daily update

* gdb.cp/annota3.exp ("watch triggered on a.x"): Match optional
"frame-begin" annotation.

* gdb.cp/annota2.exp ("watch triggered on a.x"): Allow arbitrary
number of "frames-invalid" annotations.

* doublest.c (floatformat_from_length): Use the right element from
gdbarch floatformats.
(floatformat_from_type, extract_typed_floating)
(store_typed_floating): Likewise.
* doublest.h: Remove declarations for undefined floatformat arrays.
* gdbarch.sh (float_format, double_format, long_double_format): Change
to pairs.
(pformat): Update for pairs.
* gdbarch.c, gdbarch.h: Regenerated.
* gdbtypes.c (floatformats_ieee_single, floatformats_ieee_double)
(floatformats_ieee_double_littlebyte_bigword)
(floatformats_i387_ext, floatformats_m68881_ext, floatformats_arm_ext)
(floatformats_ia64_spill, floatformats_ia64_quad, floatformats_vax_f)
(floatformats_vax_d): New variables.
(builtin_type_ieee_single, builtin_type_ieee_double)
(builtin_type_arm_ext, builtin_type_ia64_spill)
(builtin_type_ia64_quad): Replace arrays with individual types.
(builtin_type_ieee_single_big, builtin_type_ieee_single_little)
(builtin_type_ieee_double_big, builtin_type_ieee_double_little)
(builtin_type_ieee_double_littlebyte_bigword, builtin_type_i960_ext)
(builtin_type_m88110_ext, builtin_type_m88110_harris_ext)
(builtin_type_arm_ext_big, builtin_type_arm_ext_littlebyte_bigword)
(builtin_type_ia64_spill_big, builtin_type_ia64_spill_little)
(builtin_type_ia64_quad_big, builtin_type_ia64_quad_little): Delete
unused and endian-specific types.
(recursive_dump_type): Update for floatformat pairs.
(build_flt): Move higher.  Handle bit == -1.  Take a floatformat pair.
(build_gdbtypes): Use build_flt.
(_initialize_gdbtypes): Update set of initialized types.
* gdbtypes.h: Update declarations to match gdbtypes.c.
(struct main_type): Store a pointer to two floatformats.
* arch-utils.c (default_float_format, default_double_format): Delete.
* arch-utils.h (default_float_format, default_double_format): Delete.

* arm-tdep.c, avr-tdep.c, hppa-tdep.c, hppabsd-tdep.c, i386-tdep.c,
ia64-tdep.c,  iq2000-tdep.c, m68k-tdep.c, m88k-tdep.c,
mips-linux-tdep.c, mips-tdep.c, mt-tdep.c, ppcobsd-tdep.c,
sparc-linux-tdep.c, sparc-tdep.c, sparcnbsd-tdep.c, spu-tdep.c,
vax-tdep.c, alpha-tdep.c, ppc-sysv-tdep.c: Update.

2007-01-28  Manuel Lauss  <slauss@resi.at>

* configure.ac (sh64-*-*): Change to sh64*-*-*.
(sh-*-*): Change to sh*-*-*.
* configure: Regenerated.

    * gdb.texinfo (Maintenance Commands): Add documentation for
        the new "maint print target-stack" command.

    * target.c (maintenance_print_target_stack): New function.
        (initialize_targets): Add new "maintenance print target-stack"
        command.

* bfd-in2.h: Regenerate.
* bfd-in.h (bfd_arm_vfp11_fix): New enum. Specify how VFP11
instruction scanning should be done.
(bfd_elf32_arm_init_maps, bfd_elf32_arm_vfp11_erratum_scan)
(bfd_elf32_arm_vfp11_fix_veneer_locations): Add prototypes.
(bfd_elf32_arm_set_target_relocs): Add vfp11 fix type argument to
prototype.
* elf-bfd.h (elf_backend_write_section): Add struct bfd_link_info
argument.
* elf32-arm.c (VFP11_ERRATUM_VENEER_SECTION_NAME)
(VFP11_ERRATUM_VENEER_ENTRY_NAME): Define macros.
(elf32_vfp11_erratum_type): New enum.
(elf32_vfp11_erratum_list): New struct. List of veneers or jumps to
veneers.
(_arm_elf_section_data): Add mapsize, erratumcount, erratumlist.
(elf32_arm_link_hash_table): Add vfp11_erratum_glue_size,
vfp11_fix and num_vfp11_fixes fields.
(elf32_arm_link_hash_table_create): Initialise vfp11_fix,
vfp11_erratum_glue_size, num_vfp11_fixes fields.
(VFP11_ERRATUM_VENEER_SIZE): Define. Size of an (ARM) veneer.
(bfd_elf32_arm_allocate_interworking_sections): Initialise erratum
glue section.
(elf32_arm_section_map_add): Add an code/data mapping symbol entry
to a section's map.
(record_vfp11_erratum_veneer): Create a single veneer, and its
associated symbols.
(bfd_elf32_arm_add_glue_sections_to_bfd): Add vfp11 erratum glue.
(bfd_elf32_arm_init_maps): Initialise mapping symbol table for input
BFDs.
(bfd_elf32_arm_set_vfp11_fix): Set the type of erratum workaround
required.
(bfd_arm_vfp11_pipe): Define VFP11 instruction pipes.
(bfd_arm_vfp11_regno): Recode a register number from a VFP11 insn.
(bfd_arm_vfp11_write_mask): Update write mask according to coded
register number.
(bfd_arm_vfp11_antidependency): New function.
(bfd_arm_vfp11_insn_decode): Decode a VFP11 insn.
(elf32_arm_compare_mapping): Declare.
(bfd_elf32_arm_vfp11_erratum_scan): Scan the sections of an input
BFD for potential erratum-triggering insns. Record results.
(bfd_elf32_arm_vfp11_fix_veneer_locations): Find out where veneers
and branches to veneers have been placed in virtual memory after
layout.
(bfd_elf32_arm_set_target_relocs): Set vfp11_fix field in global
hash table.
(elf32_arm_output_symbol_hook): Remove.
(elf32_arm_write_section): Output veneers, and branches to veneers.
Use maps from input sections, not output sections, for code
byte-swapping.
* elf32-ppc.c (ppc_elf_write_section): Add dummy link_info argument.
* elf32-score.c (_bfd_score_elf_write_section): Likewise.
* elfxx-mips.c (_bfd_mips_elf_write_section): Likewise.
* elfxx-mips.h (_bfd_mips_elf_write_section): Likewise.

* NEWS: Mention --vfp11-denorm-fix option.
* ld.texinfo: Document above.
* emulparams/armelf_linux.sh (OTHER_TEXT_SECTIONS): Add
.vfp11_veneer section.
* emulparams/armelf.sh (OTHER_TEXT_SECTIONS): Likewise.
* emultempl/armelf.em (vfp11_denorm_fix): New static variable.
(arm_elf_before_allocation): Call bfd_elf32_arm_set_vfp11_fix,
bfd_elf32_arm_init_maps and bfd_elf32_arm_vfp11_erratum_scan.
(arm_elf_after_allocation): New function. Call
bfd_elf32_arm_vfp11_fix_veneer_locations for all input statements.
(arm_elf_create_output_section_statements): Pass vfp11 fix command
line option to BFD.
(OPTION_VFP11_DENORM_FIX): New option.
(PARSE_AND_LIST_LONGOPTS): Handle new option.
(PARSE_AND_LIST_OPTIONS): Likewise.
(PARSE_AND_LIST_ARGS_CASES): Likewise.
(LDEMUL_AFTER_ALLOCATION): Define.

* ld-arm/arm-elf.exp: Add VFP11 tests.
* ld-arm/vfp11-fix-none.s: New file.
* ld-arm/vfp11-fix-none.d: Expected disassembly of above.
* ld-arm/vfp11-fix-scalar.s: New file.
* ld-arm/vfp11-fix-scalar.d: Expected disassembly of above.
* ld-arm/vfp11-fix-vector.s: New file.
* ld-arm/vfp11-fix-vector.d: Expected disassembly of above.

* gdb.base/commands.exp: Call if_commands_test.
(gdb_test_no_prompt, if_commands_test): New.

* remote-utils.c [USE_WIN32API] (read, write): Define.
(putpkt_binary, input_interrupt, readchar, getpkt): Use read and
write.

*** empty log message ***

daily update

    * gdb.threads/sigthread.c (NSIGS): Make bigger.

2076-01-28  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (swap_imm_operands): Renamed to ...
(swap_2_operands): This.  Take 2 ints.
(md_assemble): Updated.
(swap_operands): Call swap_2_operands to swap 2 operands.

    * dwarf2read.c (new_symbol): Handle DW_AT_decl_file being zero.

Wrap long line.

Add INCLUDED_TARGET_FILE.

2076-01-27  H.J. Lu  <hongjiu.lu@intel.com>

* elf64-hppa.c (elf64_bed): Defined for HPUX and Linux.

* dwarf2loc.h (struct dwarf2_locexpr_baton): Change size to a long.
(struct dwarf2_loclist_baton): Likewise.

*** empty log message ***

daily update

* elf32-hppa.c (elf32_bed): Define for hpux, linux and netbsd.
(INCLUDED_TARGET_FILE): Remove.

* cli/cli-script.c: Include breakpoint.h.
(build_command_line): Require arguments only for if and while commands.
(get_command_line, execute_user_command, execute_control_command):
Fix wording of warning messages.
(print_command_lines): Print breakpoint commands.
(execute_control_command): Call commands_from_control_command to
handle the `commands' command inside a body of a flow-control command.
(read_next_line): Recognize the `commands' command and build a
command line structure for it.
(recurse_read_control_structure, read_command_lines): Handle
`commands' similarly to `if' and `while'.

* breakpoint.c (get_number_trailer): Document the special meaning
of NULL as the first argument PP.
(commands_from_control_command): New function.

* breakpoint.h (commands_from_control_command): Add prototype.

* defs.h (commands_control): New enumerated value for enum
command_control_type.

* gdb.dwarf2/dw2-basic.S: Use .2byte/.4byte/.8byte instead of
.short/.int/.quad in the dwarf sections.
* gdb.dwarf2/dw2-intercu.S: Likewise.
* gdb.dwarf2/dw2-intermix.S: Likewise.

*** empty log message ***

daily update

    * ada-lang.c (ada_exception_breakpoint_ops): Fix typo in function name.
        (ada_exception_sal): Update accordingly.

* c-valprint.c (c_val_print): Require strings to be of no-signed CHARs.
* NEWS: Describe CHAR array vs. string identifcation rules.

* gdb.texinfo: Describe CHAR array vs. string identifcation rules.

* gdb.base/charsign.exp, gdb.base/charsign.c: New files.
* gdb.arch/i386-sse.exp: Check $xmm.v16_int8 printing as a number array.
* gdb.base/printcmds.exp: Sign-provided CHARs now became arrays.
* gdb.base/setvar.exp: Likewise.

* gdb.base/corefile.exp: Allow non-empty argument list in reinit
test.

* gdb.base/multi-forks.exp ("restart $i"): Also match
__kernel_syscall_via_break.

*** empty log message ***

* gdb.texinfo (Compilation, Files, Bootstrapping, Bug Reporting):
Use @value{NGCC} instead of @value{GCC}.

* elf32-m32c.c (m32c_elf_howto_table): Don't complain about
R_M32C_16 or R_M32C_24 relocs.

*** empty log message ***

daily update

2007-01-25  Paul Brook  <paul@codesourcery.com>

gdb/
* arm-tdep.c (arm_get_next_pc): Fix bitfield off-by-one error.

* ar.c (print_contents, extract_file): Cast the return value
of fwrite to size_t.

PR binutils/3874
* elf32-avr.c (avr_link_hash_table): Check to make sure that the hash table was created by elf32_avr_link_hash_table_create before using it.
  (elf32_avr_link_hash_newfunc): New function.  Just pass the call through to _bfd_elf_link_hash_newfunc.
  (elf32_avr_link_hash_table_create): Use elf32_avr_link_hash_newfunc instead of _bfd_elf_link_hash_newfunc.
  (elf32_avr_relocate_section): Check for the hash table pointer being NULL.
  (elf32_avr_relax_section, avr_build_one_stub, elf32_avr_setup_params, get_local_syms, elf32_avr_size_stubs, elf32_avr_build_stubs): Likewise.

Add a space between date and name.

*** empty log message ***

daily update

2007-01-24  Jim Blandy  <jimb@codesourcery.com>

* dwarf2loc.c (dwarf2_evaluate_loc_desc): When the location
expression is empty, bother to return the 'optimized out' value we
construct.  (Thanks to Carl Burch.)

* config/tc-m32c.c (md_pseudo_table): Add .3byte.

    * varobj.c (c_value_of_root, c_value_of_child)
        (cplus_describe_child): Don't call release_value.

2007-01-24  H.J. Lu  <hongjiu.lu@intel.com>

* ldgram.y (SIZEOF_HEADERS): Remove duplicated one.
(DEFSYMEND): Likewise.
(NAME): Likewise.
(LNAME): Likewise.

* mips-linux-tdep.c (mips_linux_n64_rt_sigframe): Fix struct
initialization.

    Refactor getting children name, value and type access
        for varobjs in C++.
        * varobj.c (get_type_deref): Remove.
        (adjust_value_for_child_access): New.
        (c_number_of_children): Use the above.
        (c_describe_child): Likewise.
        (enum accessibility): New.
        (match_accessibility): New function.
        (cplus_describe_child): New function.
        (cplus_name_of_child, cplus_value_of_child)
        (cplus_type_of_child): Reimplement in terms
        of cplus_describe_child.
        (cplus_number_of_children): Use
        adjust_value_for_child_access.

    Fix computation of the 'editable' attribute and
        value changeability for for references.
        * varobj.c (get_value_type): New function.
        (c_variable_editable): Use get_value_type.
        (varobj_value_is_changeable): Likewise.

    * source.c (find_and_open_source): Try rewriting the source
        path inside filename if dirname is NULL.

*** empty log message ***

    * dwarf2read.c (add_partial_symbol): Create an extra partial
        symbol in the VAR_DOMAIN for Ada structures, unions or enums.
        (new_symbol): Likewise for symbols.

daily update

(mi_cmd_execute): Call free_all_values.

*** empty log message ***

* Makefile.in (ALL_SUBDIRS): Add gdb.xml.
* configure: Regenerated.
* configure.ac (AC_OUTPUT): Add gdb.xml/Makefile.
* gdb.xml/Makefile.in, gdb.xml/tdesc-arch.exp,
gdb.xml/tdesc-bogus.xml, gdb.xml/tdesc-errors.exp,
gdb.xml/trivial.xml, gdb.xml/tdesc-unknown.xml: New files.
* lib/gdb.exp (gdb_skip_xml_test): New function.

* ld-elf/header.d: Reduce page size, restrict to linux & vxworks
* ld-elf/header.s: Adjust.
* ld-elf/header.t: Reduce initial offset.

* lib/ld-lib.exp (run_dump_test): Don't prematurely remove
assembler output.

    * value.c (value_primitive_field): Copy the full 'location'
        contents, instead of assuming that copying ADDRESS will
        bring over everything in the union.  Remove obsolete comment.

* m32c-tdep.c (make_regs, m32c_analyze_prologue)
(m32c_skip_trampoline_code, m32c_m16c_address_to_pointer)
(m32c_m16c_pointer_to_address): Separate code from declarations.

From Nick Hudson:
* target.c (update_current_target): Correct typo.

* gdb.base/sigrepeat.exp: Correct error message.