* dwarf2read.c (struct pubnames_header): Remove.
(_PUBNAMES_HEADER): Remove.
(_ACTUAL_PUBNAMES_HEADER_SIZE): Remove.
(struct aranges_header): Remove.
(_ARANGES_HEADER, _ACTUAL_ARANGES_HEADER_SIZE): Remove.
(struct dwarf2_per_objfile) <pubnames, aranges>: Remove.
(PUBNAMES_SECTION): Remove.
(ARANGES_SECTION): Remove.
(dwarf2_locate_sections): Don't handle pubnames or aranges.
(dwarf2_build_psymtabs): Remove dead code.
(dwarf2_build_psymtabs_easy): Remove.

* elfread.c (elf_symfile_read): Don't call
dwarf2_build_frame_info.
* dwarf2read.c (struct dwarf2_section_info) <readin>: New field.
(struct dwarf2_per_objfile) <objfile>: New field.
(dwarf2_has_info): Now idempotent.  Set objfile field.
(dwarf2_read_section): Check and set readin field.  Call
posix_madvise.
(dwarf2_build_psymtabs): Don't read all sections.
(read_type_comp_unit_head): Read types section.
(create_debug_types_hash_table): Likewise.
(init_cu_die_reader): Add asserts.
(process_type_comp_unit): Add assert.
(dwarf2_build_psymtabs_hard): Read info section.
(load_partial_comp_unit): Add assert.
(create_all_comp_units): Read info section.
(load_full_comp_unit): Likewise.
(dwarf2_ranges_read): Read ranges section.
(dwarf2_record_block_ranges): Add assert.
(dwarf2_read_abbrevs): Read abbrev section.
(read_indirect_string): Read str section.
(dwarf_decode_line_header): Read line section.
(read_signatured_type_at_offset): Read types section.
(dwarf_decode_macros): Read macinfo section.
(dwarf2_symbol_mark_computed): Read loc section.
* dwarf2-frame.c (dwarf2_frame_find_fde): Call
dwarf2_build_frame_info.
(dwarf2_build_frame_info): Unconditionally set
dwarf2_frame_objfile_data on the objfile.
* configure.ac: Check for posix_madvise.
* config.in, configure: Rebuild.

* xcoffread.c (xcoff_start_psymtab): Update.
(xcoff_end_psymtab): Update.
* psymtab.c (allocate_psymtab): Remove dead code.
* psympriv.h (struct partial_symtab) <read_symtab_private>: Now
void*.
* mdebugread.c (parse_partial_symbols): Update.
(new_psymtab): Likewise.
* dwarf2read.c (process_psymtab_comp_unit): Update.
(psymtab_to_symtab_1): Update.
* dbxread.c (start_psymtab): Update.
(end_psymtab): Likewise.

gdb
* xcoffread.c: Include psymtab.h.
(xcoff_sym_fns): Update.
* symtab.h (struct partial_symbol): Remove.
(PSYMBOL_DOMAIN, PSYMBOL_CLASS): Remove.
(struct partial_symtab): Remove.
(PSYMTAB_TO_SYMTAB): Remove.
(lookup_partial_symbol, lookup_partial_symtab, find_pc_psymtab)
(find_pc_sect_psymtab): Remove.
(find_pc_sect_symtab_via_partial): Declare.
(find_pc_psymtab, find_pc_sect_psymbol, psymtab_to_symtab)
(find_main_psymtab): Remove.
(find_main_filename): Declare.
(fixup_psymbol_section): Remove.
(fixup_section): Declare.
* symtab.c: Include psymtab.h.
(lookup_symtab): Use lookup_symtab method.
(lookup_partial_symtab): Remove.
(find_pc_sect_psymtab_closer): Remove.
(find_pc_sect_psymtab): Remove.
(find_pc_sect_symtab_via_partial): New function.
(find_pc_psymtab, find_pc_sect_psymbol, find_pc_psymbol): Remove.
(fixup_section): No longer static.
(fixup_psymbol_section): Remove.
(lookup_symbol_aux): Use lookup_symbol_aux_quick.
(lookup_global_symbol_from_objfile): Likewise.
(lookup_symbol_aux_psymtabs): Remove.
(lookup_symbol_aux_quick): New function.
(lookup_symbol_global): Use lookup_symbol_aux_quick.
(lookup_partial_symbol): Remove.
(basic_lookup_transparent_type_quick): New function.
(basic_lookup_transparent_type): Use it.
(find_main_psymtab): Remove.
(find_main_filename): New function.
(find_pc_sect_symtab): Use find_pc_sect_symtab method.
(find_line_symtab): Use expand_symtabs_with_filename method.
(output_partial_symbol_filename): New function.
(sources_info): Use map_partial_symbol_filenames.
(struct search_symbols_data): New type.
(search_symbols_file_matches): New function.
(search_symbols_name_matches): Likewise.
(search_symbols): Use expand_symtabs_matching method.
(struct add_name_data): Rename from add_macro_name_data.
(add_macro_name): Update.
(add_partial_symbol_name): New function.
(default_make_symbol_completion_list): Use
map_partial_symbol_names.
(struct add_partial_symbol_name): New type.
(maybe_add_partial_symtab_filename): New function.
(make_source_files_completion_list): Use
map_partial_symbol_filenames.
(expand_line_sal): Use expand_symtabs_with_filename method.
* symmisc.c: Include psymtab.h.
(print_objfile_statistics): Use print_stats method.
(dump_objfile): Use dump method.
(dump_psymtab, maintenance_print_psymbols)
(maintenance_info_psymtabs, maintenance_check_symtabs)
(extend_psymbol_list): Remove.
* symfile.h (struct quick_symbol_functions): New struct.
(struct sym_fns) <qf>: New field.
(sort_pst_symbols): Remove.
(increment_reading_symtab): Declare.
* symfile.c: Include psymtab.h.
(compare_psymbols, sort_pst_symbols): Remove.
(psymtab_to_symtab): Remove.
(increment_reading_symtab): New function.
(symbol_file_add_with_addrs_or_offsets): Use expand_all_symtabs
method.
(set_initial_language): Use find_main_filename.
(allocate_psymtab, discard_psymtab, cashier_psymtab): Remove.
(free_named_symtabs): Remove unused code.
(start_psymtab_common, add_psymbol_to_bcache)
(append_psymbol_to_list, add_psymbol_to_list, init_psymbol_list):
Remove.
* stack.c: Include psymtab.h, symfile.h.
(backtrace_command_1): Use find_pc_sect_symtab_via_partial.
* source.h (psymtab_to_fullname): Don't declare.
* source.c: Include psymtab.h.
(select_source_symtab): Use find_last_source_symtab method.
(forget_cached_source_info): Use forget_cached_source_info
method.
(find_and_open_source): No longer static.
(psymtab_to_fullname): Remove.
* somread.c: Include psymtab.h.
(som_sym_fns): Update.
* psympriv.h: New file.
* psymtab.h: New file.
* psymtab.c: New file.
* objfiles.h: (ALL_OBJFILE_PSYMTABS): Remove.
(ALL_PSYMTABS, ALL_PSPACE_PSYMTABS): Likewise.
* objfiles.c: Include psymtab.h.
(objfile_relocate1): Use relocate method.
(objfile_has_partial_symbols): Use has_symbols method.
* mipsread.c: Include psymtab.h.
(ecoff_sym_fns): Update.
* mi/mi-cmd-file.c: Include psymtab.h.
(print_partial_file_name): New function.
(mi_cmd_file_list_exec_source_files): Use
map_partial_symbol_filenames.
* mdebugread.c: Include psympriv.h.
* machoread.c: Include psympriv.h.
(macho_sym_fns): Update.
* m2-exp.y (yylex): Use lookup_symtab.
* elfread.c: Include psympriv.h.
(elf_sym_fns): Update.
* dwarf2read.c: Include psympriv.h.
* dbxread.c: Include psympriv.h.
(aout_sym_fns): Update.
* cp-support.c: Include psymtab.h.
(read_in_psymtabs): Remove.
(make_symbol_overload_list_qualified): Use
expand_symtabs_for_function method.
* coffread.c: Include psympriv.h.
(coff_sym_fns): Update.
* blockframe.c: Include psymtab.h.
(find_pc_partial_function): Use find_pc_sect_symtab method.
* ada-lang.h (ada_update_initial_language): Update.
* ada-lang.c: Include psymtab.h.
(ada_update_initial_language): Remove 'main_pst' argument.
(ada_lookup_partial_symbol): Remove.
(struct ada_psym_data): New type.
(ada_add_psyms): New function.
(ada_add_non_local_symbols): Use map_ada_symtabs method.
(struct add_partial_datum): New type.
(ada_add_partial_symbol_completions): New function.
(ada_make_symbol_completion_list): Use map_partial_symbol_names.
(ada_exception_support_info_sniffer): Update.
* Makefile.in (SFILES): Add psymtab.c.
(COMMON_OBS): Add psymtab.o.
(HFILES_NO_SRCDIR): Add psymtab.h, psympriv.h.
gdb/doc
* gdbint.texinfo (Symbol Handling): Update.

* fileread.cc (File_read::~File_read): Don't delete whole_file_view_.
(File_read::open[1]): Remove initial mapping of whole_file_view_.
(File_read::open[2]): Add whole_file_view_ to list of views.
(File_read::make_view): Remove test of whole_file_view_.
(File_read::find_or_make_view): Create whole_file_view_ if
necessary.
(File_read::clear_views): Replace bool parameter with enum;
adjust all callers.  Don't delete views with permanent data;
do delete cached views and views from archives if
--no-keep-files-mapped is set.  Set whole_file_view_ to NULL
if clearing the corresponding view.
* fileread.h (File_read::Clear_views_mode): New enum.
(File_read::View::is_permanent_view): New method.
(File_read::clear_views): Replace bool parameter
with enum; adjust all callers.
* options.h (General_options): Change keep_files_mapped option;
add map_whole_files.
* readsyms.cc (Add_symbols::run): Delete sd_ object before
releasing the file.
* reloc.cc (Scan_relocs::run): Delete rd_ object before releasing
the file.

* sparc.cc (Target_sparc::Scan::local): Accept R_SPARC_WPLT30.

2010-03-10  Pierre Muller  <muller@ics.u-strasbg.fr>

* remote-fileio.c (cygwin_conv_path): Define macro for old cygwin API.

   2010-03-09  Sami Wagiaalla  <swagiaal@redhat.com>

PR C++/11236:
* cp-namespace.c (cp_add_using): Deleted.
(cp_add_using_directive): Use obstack allocations.
Merged the function cp_add_using into this one.
Added 'struct obstack *' argument.
(cp_scan_for_anonymous_namespaces): Updated.
* cp-support.h: Updated.
* dwarf2read.c (read_import_statement): Updated.
(read_namespace): Updated.

* windows-nat.c (cygwin_conv_path): Remove old macro.

strip trailing whitespace in Blackfin files

gdb/
* breakpoint.c (condition_command): Handle watchpoint conditions.
(is_hardware_watchpoint): Add comment.
(is_watchpoint): New.
(update_watchpoint): Don't reparse the watchpoint's condition
unless necessary.
(WP_IGNORE): New.
(watchpoint_check): Use it.
(bpstat_check_watchpoint): Handle it.
(bpstat_check_breakpoint_conditions): Evaluate watchpoint local
conditions in a frame where it makes sense.
(watch_command_1): Store the innermost block of the condition
expression.
(delete_breakpoint): Delete the watchpoint condition expression.
* breakpoint.h (struct bp_location) <cond>: Update comment.
(struct breakpoint): New fields `cond_exp' and
`cond_exp_valid_block'.

gdb/testsuite/
* gdb.base/watch-cond.c, gdb.base/watch-cond.exp: New.

add support for Blackfin bf504/bf506

Adjust handling of Ada DIEs after dwarf2_physname patch.

        * dwarf2read.c (dwarf2_compute_name): Add handling of Ada DIEs.

* doc/as.texinfo: Add Blackfin options.
* doc/c-bfin.texi: Document -mfdpic, -mno-fdpic and -mnopic.
* config/tc-bfin.c (md_show_usage): Show usage for all
Blackfin specific options.

2010-03-09  Sriraman Tallam  <tmsriram@google.com>

* icf.cc (get_section_contents): Add '@' marker after processing the
merge reloc.

*** empty log message ***

daily update

2010-03-09   Rafael Espindola  <espindola@google.com>

* archive.cc (should_include_member): New.
(Archive::add_symbols): Use should_include_member.

* windows-nat.c (cygwin_conv_path): Redefine to properly convert from/to
posix/win32.
(windows_make_so): Use non-Cygwin 1.7 specific function.
(windows_create_inferior): Make sure that cygallargs points to original args in
non Cygwin 1.7.  case.

2010-03-09  Michael Snyder  <msnyder@vmware.com>

* i386-tdep.c (i386_record_lea_modrm_addr): Use extract_signed_int
after target_read_memory to get host byte order.
(i386_process_record): Ditto.

dwarf2_physname patchset:

Based on work from Daniel Jacobowitz  <dan@codesourcery.com>
* c-typeprint.c (cp_type_print_method_args): For non-static methods,
print out const or volatile qualifiers, too.
(c_type_print_args): Add parameters show_artificial and language.
Skip artificial parameters when requested.
Use the appropriate language printer.
(c_type_print_varspec): Tell c_type_print_args to skip artificial
parameters and pass language_c.
* dwarf2read.c (die_list): New file global.
(struct partial_die_info): Update comments for name field.
(pdi_needs_namespace): Renamed to ...
(die_needs_namespace): ... this. Rewrite.
(dwarf2_linkage_name): Remove.
(add_partial_symbol): Do not predicate the call to
partial_die_full_name based on pdi_needs_namespace.
Remove call to cp_check_possible_namespace_symbols and associated
outdated comments.
(guess_structure_name): Do not inspect child subprogram DIEs.
(dwarf2_fullname): Update comments.
Use die_needs_namespace to assist in computing the name.
(read_func_scope): Use dwarf2_name to get the DIE's name.
Use dwarf2_physname to get the "linkage name" of the DIE.
(dwarf2_add_member_field): Use dwarf2_physname instead of
dwarf2_linkage_name.
(read_structure_type): For structs and classes, set TYPE_NAME, too.
(determine_class): Remove.
(read_partial_die): Ignore DW_AT_MIPS_linkage_name for all languages
except Ada.
(new_symbol): Unconditionally call dwarf2_name.
Compute the "linkage name" using dwarf2_physname.
Use dwarf2_name instead of dwarf2_full_name for enumerator DIEs.
When determining to scan for anonymous C++ namespaces, ignore
the linkage name.
(dwarf2_physname): New function.
(dwarf2_full_name): Move content to new function and call
that.
(dwarf2_compute_name): "New" function.
(_initialize_dwarf2_read): Initialize die_list.
* gnu-v3-eabi.c (gnu_v3_find_method_in): Remove unused variable
physname.
(gnu_v3_print_method_ptr): Use the physname for virtual methods
without a demangled name.
Print out type information for non-virtual methods.
        * linespec.c (decode_line_1): Force ANY string using "::" (or
"." for java) to use decode_compound, and clean up any stray quoting.
        If we found a file symtab, re-evaluate whether the remainder is_quoted.
        (decode_compound): Stop consuming at an open parenthesis.
        Keep template parameters.
        Keep any overload information.
        Keep keywords like "const".
        Remove paren_pointer.
        Move is_quoted check from set_flags to here.
        Remove #if 0 code from 2000. Ten years is long enough.
        (find_method): Before comparing symbol names, canonicalize the string
        from the user.
        If a specific overload is requested, find it. Otherwise throw an error.
(find_method_overload_end): New function.
(set_flags): Remove.
        (decode_compound): Assume that parentheses are matched.
        It's a lot easier.
* symtab.c (symbol_find_demangled_name): Add DMGL_VERBOSE flag
to cplus_demangle.
* linespec.c (decode_line_1): Keep important keywords like
"const" and "volatile".
* symtab.h (SYMBOL_CPLUS_DEMANGLED_NAME): Remove.
* typeprint.h (c_type_print_args): Add declaration.
* ui-file.c (do_ui_file_obsavestring): New function.
(ui_file_obsavestring): New function.
* ui-file.h (ui_file_obsavestring): Add declaration.
* valops.c (find_overload_match): Resolve the object to
a non-pointer type.
If the object is a data member, search the object for the member
and return with staticp set.
Use SYMBOL_NATURAL_NAME instead of SYMBOL_CPLUS_DEMANGLED_NAME.
Do not attempt to extract a function name from non-function types.
If the extracted function name and the original name are the same,
we don't have a C++ method.

From Jan Kratochvil  <jan.kratochvil@redhat.com>:
     * dwarf2read.c (new_symbol <DW_TAG_enumerator>): Call dwarf2_full_name.

* ada-lang.c (ada_lookup_symbol): Remove linkage_name parameters
and arguments from symbol lookups.
* ax-gdb.c (gen_expr): Likewise.
* cp-namespace.c (cp_lookup_symbol_nonlocal, lookup_namespace_scope,
cp_lookup_symbol_namespace, lookup_symbol_file, lookup_nested_type,
lookup_possible_namespace_symbol): Likewise.
* cp-support.c (read_in_psymtabs): Likewise.
* cp-support.h (cp_lookup_symbol_nonlocal): Likewise.
* language.h (la_lookup_symbol_nonlocal): Likewise.
* scm-valprint.c (scm_inferior_print): Likewise.
* solib-darwin.c (darwin_relocate_section_addresses): Likewise.
* solib-svr.c (elf_lookup_lib): Likewise.
* solib.c (show_auto_solib_add): Likewise.
* solist.h (lookup_lib_global, solib_global_lookup): Likewise.
* symmisc.c (maintenance_check_symtabs): Likewise.
* symtab.c (lookup_symbol_in_language, lookup_symbol_aux,
lookup_symbol_aux_local, lookup_symbol_aux_block,
lookup_symbol_from_objfile, lookup_symbol_aux_symtabs,
lookup_symbol_aux_psymtabs,basic_lookup_symbol_nonlocal,
lookup_symbol_static, lookup_symbol_global, symbol_matches_domain,
basic_lookup_transparent_type, find_main_psymtab,
lookup_block_symbol): Likewise.
* symtab.h (basic_lookp_symbol_nonlocal, lookup_symbol_static,
lookup_symbol_global, lookup_symbol_aux_block,
lookup_symbol_partial_symbol, lookup_block_symbol,
lookup_global_symbol, value_maybe_namespace_elt): Likewise.

    * gdb.cp/cp-relocate.exp: Remove single-quoting of C++ methods.
        * gdb.cp/cplusfuncs.cc (dm_type_short): New function.
        (dm_type_long): New function.
        (dm_type_unsigned_short): New function.
        (dm_type_unsigned_long): New function.
        (myint): New typedef.
        * gdb.cp/cplusfuncs.exp (probe_demangler): Add tests for short,
        long, unsigned shor and long, operator char*, and typedef.
        (test_lookup_operator_functions): Add operator char* test.
        (test_paddr_operator_functions): Likewise.
        (test_paddr_overloaded_functions): Use probe values for
        short, long, and unsigned short and long.
        (test_paddr_hairy_functions): If the demangler probe detected
        gdb type printers, "expect" them. Otherwise "expect" the v2 or v3
        demangler.
        * gdb.cp/expand-sals.exp: Backtrace may contain class names.
        * gdb.cp/member-ptr.exp: Refine expected result for "print pmf"
        and "print null_pmf".
        Add test "ptype a.*pmf".
        * gdb.cp/overload.exp: Allow optional "int" to appear with
        "short" and "long".
        * gdb.cp/ovldbreak.exp: Use append to construct super-duper
        long expect value for men_overload1arg.
        Allow "int" to appear with "short" and "long".
        When testing "info break", add argument for main (void).
        Also allow "int" to appear with "short" and "long".
        Ditto with "unsigned" and "long long".
* gdb.java/jmain.exp: Do not enclose methods names in single
quotes.
* gdb.java/jmisc.exp: Likewise.
* gdb.java/jprint.exp: Likewise.
* gdb.python/py-symbol.exp: Update expected "linkage_name" value.

From Jan Kratochvil  <jan.kratochvil@redhat.com>:
* gdb.cp/exception.exp (backtrace after first throw)
(backtrace after second throw): Allow a namespace before __cxa_throw.
(backtrace after first catch, backtrace after second catch): Allow
a namespace before __cxa_begin_catch.

* gdb.cp/cpexprs.exp: New file.
* gdb.cp/cpexprs.cc: New file.

From Daniel Jacobowitz  <dan@codesourcery.com>
* gdb.cp/cpexprs.exp (escape): Delete.  Change all callers
to use string_to_regexp.
(ctor, dtor): New functions.  Use them to match constructor
and destructor function types.
(Top level): Use runto_main.

* python/python-internal.h: Include symtab.h.

* p-lang.c (is_pascal_string_type): Remove unneeded parentheses.
* p-valprint.c (pascal_val_print): Remove undeed block and fix
indentation.

Add tests for PR gas/9966 and PR gas/11356.

2010-03-08  H.J. Lu  <hongjiu.lu@intel.com>

PR gas/9966
PR gas/11356
* gas/i386/list-1.l: New.
* gas/i386/list-1.s: Likewise.
* gas/i386/list-2.l: Likewise.
* gas/i386/list-2.s: Likewise.
* gas/i386/list-3.l: Likewise.
* gas/i386/list-3.s: Likewise.

* gas/i386/i386.exp: Run list-1, list-2 and list-3.

* lib/gas-defs.exp (gas_run_stdin): New.
(run_list_test_stdin): Likewise.

PR gas/11356
* listing.c (listing_newline): Correct backslash quote logic.

*** empty log message ***

daily update

2010-03-08  Doug Kwan  <dougkwan@google.com>

* gold/arm.cc (Arm_exidx_fixup::update_offset_map): Fix build breakage
due to a conversion warning.
(Arm_relobj::update_output_local_symbol_count): Check for local
symbol with unset output index.

* breakpoint.c (breakpoint_1): Add "QUIT".

* solib.c (solib_find): Replace extension if
solib_symbols_extension is set in the target gdbarch.
* arm-symbian-tdep.c (arm_symbian_init_abi): Set
solib_symbols_extension to "sym".
* gdbarch.sh (solib_symbols_extension): New variable.
(pstring): New function.
* gdbarch.h, gdbarch.c: Regenerate.

gdb
PR cli/9591:
* NEWS: Update.
* utils.c: Include main.h.
(fputs_maybe_filtered): Don't paginate if `batch_flag'.
(defaulted_query): Use default answer if `batch_flag'.
* main.h (batch_flag): Declare.
* main.c (batch_flag): New global.
(captured_main): Remove 'batch'.  Update.
gdb/doc
PR cli/9591:
* gdb.texinfo (Mode Options): Mention lack of pagination and
confirmation with --batch.
(Screen Size): Mention --batch.
(Messages/Warnings): Likewise.

 * Fix commit year

* remote-mips.c (rockhopper_ops): New target_ops struct.
(MON_ROCKHOPPER): New mips_monitor_type.
(read_hex_value): New function.
(mips_request): Send 8-byte values with a 'T' packet.  Read the
packet argument as a string and use read_hex_value to parse it.
(mips_exit_debug): Wait for response when using MON_ROCKHOPPER.
(rockhopper_open): New function.
(mips_wait): Read the PC, FP and SP fields as strings.  Use
read_hex_value to parse them and mips_set_register to commit them.
(mips_set_register): New function.
(mips_fetch_registers): Do not cast register value to "unsigned"
when reading a MON_ROCKHOPPER 't' packet.  Use mips_set_register.
(mips_store_registers): Use a 'T' packet to set registers when
using MON_ROCKHOPPER.
(pmon_end_download): Don't run initEther if using MON_ROCKHOPPER
and expect the total to be printed before the entry address.
(_initialize_remote_mips): Initialize and add rockhopper_ops.

* remote-mips.c (mips_fetch_word): Add new parameter, `valp'.
Change return value to int.  Store value fetched in location
addressed by `val'.  Use function's return value as success
or failure indicator.  Adjust all callers.

* remote-mips.c (mips_fetch_word): Add new parameter, `valp'.
Change return value to int.  Store value fetched in location
addressed by `val'.  Use function's return value as success
or failure indicator.  Adjust all callers.

* p-lang.c (is_pascal_string_type): Check that TYPE arg is non NULL.

* config/tc-i386.h [TE_SOLARIS] (ELF_TARGET_FORMAT): Define.
(ELF_TARGET_FORMAT64): Define.

2010-03-08  Jan Kratochvil  <jan.kratochvil@redhat.com>
    Hui Zhu  <teawater@gmail.com>

* record.c (record_open_1): Check tmp_to_stopped_by_watchpoint and
tmp_to_stopped_data_address.
(record_open): Reset tmp_to_stopped_by_watchpoint and
tmp_to_stopped_data_address.
* target.c (init_dummy_target): Add to_stopped_by_watchpoint and
to_stopped_data_address.

2010-03-08  Hui Zhu  <teawater@gmail.com>

* i386-tdep.c (i386_process_record): Initialize regnum.

gdb/
* symfile.c (addr_info_make_relative): New variable sect_name, use it.
Do not warn on ".gnu.liblist" and ".gnu.conflict".

Memory error when reading wrong core file.

        * solib-svr4.c (solib_svr4_r_map): catch and print all exception
        errors while reading the inferior memory, and return zero if
        an exception was raised.

2010-03-07  Michael Snyder  <msnyder@vmware.com>

* record.c (record_restore): Rename tmpu8 to rectype.

2010-03-07  Michael Snyder  <msnyder@vmware.com>

* i386-tdep.c (i386_record_lea_modrm_addr): Rename local variables
tmpu8, tmpi16, tmpi32, tmpulongest to addr8, addr16, addr32, addr64.

(i386_record_push): Rename local tmpulongest to addr.

(i386_process_record): Rename local tmpulongest to addr.

2010-03-07  Michael Snyder  <msnyder@vmware.com>

* i386-tdep.c (i386_process_record): Rename local variables
tmpu16, tmpu32, tmpu64 to addr16, addr32, addr64.

2010-03-07  Michael Snyder  <msnyder@vmware.com>

* i386-tdep.c (i386_process_record): Rename tmpu8 to opcode8 and
regnum.

*** empty log message ***

daily update

Assertion failure after Ada task switch (target remote).

        * remote.c (remote_get_ada_task_ptid): New function.
        (init_remote_ops): Set remote_ops.to_get_ada_task_ptid.

*** empty log message ***

daily update

* windows-nat.c: Reorganize #ifdef __CYGWIN__ considerations into one block.
Define helper macros to reduce ifdefs in code.
(get_module_name): Use cygwin_buf_t for buffer and __PMAX for buffer size.
Call unadorned GetModuleFileNameEx rather than GetModuleFileNameEx*.
(windows_make_so): Use __PMAX to denote maximum buffer size and cygwin_buf_t
for buffer type.  Use GetSystemDirectory{W,A} as appropriate.
(get_image_name): Use __PMAX to denote maximum buffer size.
(handle_load_dll): Likewise.
(windows_pid_to_exec_file): Likewise.
(windows_create_inferior): Add many accommodations for older Cygwin and
non-Cygwin.
(bad_GetModuleFileNameExW): Control inclusion of this function based on
__USEWIDE conditional.
(bad_GetModuleFileNameExA): Likewise.
(_initialize_loadable): Just use real function names without the dyn_ part
since they are defined earlier.

* options.h (class General_options): Add --spare-dynamic-tags.
* output.cc (Output_data_dynamic::set_final_data_size): Implement
--spare-dynamic-tags.

2010-03-06  Simo Melenius  <simo.melenius@iki.fi>

* linux-low.c (linux_wait_for_lwp): Fetch the regcache after
switching the current inferior, not before.

*** empty log message ***

* incremental.cc: Include "libiberty.h".

include:
* elf/common.h (VER_FLG_*): Document.
(VER_FLG_INFO): Define.

gold:
* dynobj.h (Verdef::Verdef): Add is_info arg, is_info member
function, is_info_ member.
* dynobj.cc (Verdef::write): Set VER_FLG_INFO if this->is_info_.
(Versions::Versions): Update caller.
(Versions::define_base_version): Likewise.
(Versions::add_def): Likewise.

elfcpp:
* elfcpp.h (VER_FLG_INFO): Define.

binutils:
* readelf.c (get_ver_flags): Handle VER_FLG_INFO.

gdb
2010-03-05  Corinna Vinschen  <vinschen@redhat.com>
    Tom Tromey  <tromey@redhat.com>

* utils.c (host_char_to_target): Add 'gdbarch' argument.
(parse_escape): Likewise.
* python/py-utils.c (unicode_to_target_string): Update.
(unicode_to_target_python_string): Update.
(target_string_to_unicode): Update.
* printcmd.c (printf_command): Update.
* p-exp.y (yylex): Update.
* objc-exp.y (yylex): Update.
* mi/mi-parse.c: Include charset.h.
(mi_parse_escape): New function.
(mi_parse_argv): Use it.
* jv-exp.y (yylex): Update.
* i386-cygwin-tdep.c (i386_cygwin_auto_wide_charset): New
function.
(i386_cygwin_init_abi): Call set_gdbarch_auto_wide_charset.
* gdbarch.sh (auto_charset, auto_wide_charset): New.
* gdbarch.c: Rebuild.
* gdbarch.h: Rebuild.
* defs.h (parse_escape): Update.
* cli/cli-setshow.c: Include arch-utils.h.
(do_setshow_command): Update.
* cli/cli-cmds.c (echo_command): Update.
* charset.h (target_charset, target_wide_charset): Update.
* charset.c: Include arch-utils.h.
(target_charset_name): Default to "auto".
(target_wide_charset_name): Likewise.
(show_target_charset_name): Handle "auto".
(show_target_wide_charset_name): Likewise.
(be_le_arch): New global.
(set_be_le_names): Add 'gdbarch' argument.
(validate): Likewise.  Don't call set_be_le_names.
(set_charset_sfunc, set_host_charset_sfunc)
(set_target_charset_sfunc, set_target_wide_charset_sfunc):
Update.
(target_charset): Add 'gdbarch' argument.
(target_wide_charset): Likewise.  Remove 'byte_order' argument.
(auto_target_charset_name): New global.
(default_auto_charset, default_auto_wide_charset): New functions.
(_initialize_charset): Set auto_target_charset_name.  Allow "auto"
for target charsets.  Copy result of nl_langinfo.  Use GetACP if
USE_WIN32API.
* c-lang.c (charset_for_string_type): Add 'gdbarch' argument,
remove 'byte_order' argument.  Update.
(classify_type): Likewise.
(c_emit_char): Update.
(c_printchar): Update.
(c_printstr): Update.
(c_get_string): Update.
(evaluate_subexp_c): Update.
* arch-utils.h (default_auto_charset, default_auto_wide_charset):
Declare.
* python/python.c (gdbpy_target_charset): New function.
(gdbpy_target_wide_charset): Likewise.
(GdbMethods): Update.
* NEWS: Update.
gdb/doc
* gdb.texinfo (Basic Python): Document target_charset and
target_wide_charset.
gdb/testsuite
* gdb.python/py-prettyprint.py (pp_nullstr.to_string): Use
gdb.target_charset.
(pp_ns.to_string): Likewise.

* gdb.texinfo (Data): Link to pretty-printing.
(Output Formats): Likewise.  Correct text.

ld:
* Makefile.am (ALL_EMULATIONS): Add eelf32_sparc_sol2.o,
eelf_i386_sol2.o.
(ALL_64_EMULATIONS): Add eelf_x86_64_sol2.o, eelf64_sparc_sol2.o.
(eelf32_sparc_sol2.c): New rule.
(eelf64_sparc_sol2.c): Likewise.
(eelf_x86_64_sol2.c): Likewise.
(eelf_i386_sol2.c): Likewise.
* Makefile.in: Regenerate.
* configure.tgt (i[3-7]86-*-solaris2*): Change targ_emul to
elf_i386_sol2.
Add elf_i386_ldso, elf_x86_64_sol2 to targ_extra_emuls.
(x86_64-*-solaris2*): Change targ_emul to elf_x86_64_sol2.
Add elf_x86_64, elf_i386_sol2, elf_i386 to targ_extra_emuls.
(sparc-*-solaris2.[0-6]*): Change targ_emul to elf32_sparc_sol2.
Add target_extra_emuls.
(sparc-*-solaris2*): Change targ_emul to elf32_sparc_sol2.
Add elf32_sparc, elf64_sparc_sol2 to targ_extra_emuls.
(sparcv9-*-solaris2*): Change targ_emul to elf64_sparc_sol2.
Add elf64_sparc, elf32_sparc_sol2 to target_extra_emuls.
* emulparams/elf32_sparc_sol2.sh: New file.
* emulparams/elf64_sparc_sol2.sh: New file.
* emulparams/elf_i386_sol2.sh: New file.
* emulparams/elf_x86_64_sol2.sh: New file.
* emultempl/solaris2.em: New file.

bfd:
* elflink.c (bfd_elf_size_dynamic_sections): Don't emit base
version twice.
Skip it when constructing def.vd_next.

* elf32-i386.c (TARGET_LITTLE_SYM): Redefine to
bfd_elf32_i386_sol2_vec.
(TARGET_LITTLE_NAME): Redefine to elf32-i386-sol2.
(elf32_bed): Redefine to elf32_i386_sol2_bed.
(elf_backend_want_plt_sym): Redefine to 1.

* elf64-x86-64.c (TARGET_LITTLE_SYM): Redefine to
bfd_elf64_x86_64_sol2_vec.
(TARGET_LITTLE_NAME): Redefine to elf64-x86-64-sol2.
(elf64_bed): Redefine to elf64_x86_64_sol2_bed.
(elf_backend_want_plt_sym): Redefine to 1.

* config.bfd (i[3-7]86-*-solaris2*): Set targ_defvec to
bfd_elf32_i386_sol2_vec.
Replace bfd_elf64_x86_64_vec by bfd_elf64_x86_64_sol2_vec in
targ64_selvecs.
(x86_64-*-solaris2*): Set targ_defvec to bfd_elf32_i386_sol2_vec.
Replace bfd_elf64_x86_64_vec by bfd_elf64_x86_64_sol2_vec in
targ_selvecs.

* configure.in: Handle bfd_elf32_i386_sol2_vec,
bfd_elf64_x86_64_sol2_vec.
* configure: Regenerate.

* targets.c (bfd_elf32_i386_sol2_vec): Declare.
(bfd_elf64_x86_64_sol2_vec): Declare.
(_bfd_target_vector): Add bfd_elf32_i386_sol2_vec,
bfd_elf64_x86_64_sol2_vec.

* symfile.c (build_section_addr_info_from_objfile): Do not mask
off high address bits.

* python/py-lazy-string.c (gdbpy_extract_lazy_string): Extract
address as UnsignedLongLong, not LongLong.

* remote-mips.c (gdbthread.h): Include.
(remote_mips_ptid): Declare.
(mips_error): Only mourn the inferior when inferior_ptid is non-null.
(common_open): Set inferior_ptid, add it as an inferior, and
as a thread too.  Delete FIXME comment regarding start_remote().
(mips_close): Invoke generic_mourn_inferior().
(mips_kill): Make sure that target_mourn_inferior is invoked.
(mips_mourn_inferior): Don't invoke generic_mourn_inferior, as
it's now invoked from mips_close().
(mips_load): Don't null out inferior_ptid.  Don't call
clear_symtab_users().
(mips_thread_alive, mips_pid_to_str): New functions.
(_initialize_remote_mips): Initialize remote_mips_ptid.  Initialize
to_thread_alive and to_pid_to_str operations.

2010-03-05  Paul Brook  <paul@codesourcery.com>

gas/
* config/tc-arm.c (arm_cpu_option_table): Add cortex-m4.

*** empty log message ***

daily update

* dwarf2read.c (skip_one_die) <DW_FORM_ref_addr>: Use offset size
in DWARF 3 and later.
(read_attribute_value) <DW_FORM_ref_addr>: Likewise.

* arm-tdep.c (arm_make_stub_cache): Delete unused locals reg,
unwound_fp.

* linespec.c (decode_line_1): Update comments for is_quote_enclosed.
If the filename portion of the linespec was quoted, recheck the
remainder for additional quoting.
(locate_first_half): Skip over completer chars, too.

* gdb.cp/overload.exp: Test that the filename portion of a linespec
can be quoted.  Test that both the filename and function/line
portions can be quoted at the same time.

* printcmd.c (printf_command): Pass dummy argument to
printf_filtered.

* elf32-arm.c (elf32_arm_output_arch_local_syms): Skip non-program
sections.

PR libstdc++/32499
* configure.ac (RANLIB): Default to true.
(STRIP): Likewise.
(RANLIB_FOR_TARGET): Remove superfluous : argument.
* configure: Regenerate.

* arm-tdep.c (arm_stub_unwind_sniffer): Add comment.

gdb/
* breakpoint.c (update_watchpoint): Create a sentinel location if
the software watchpoint isn't watching any memory.
(breakpoint_address_bits): Skip dummy software watchpoint locations.

gdb/testsuite/
* gdb.base/watch-non-mem.c, gdb.base/watch-non-mem.exp: New.

* utils.c (fputs_maybe_filtered): Check if there's already a top
level interpreter before dereferencing it.  If there isn't one,
don't paginate either.

2010-03-04  Matthew Gretton-Dann  <matthew.gretton-dann@arm.com>

* arm-tdep.c (arm_pc_is_thumb): Add heuristic that tries to get
the state right when single stepping.
(arm_get_next_pc_raw, thumb_get_next_pc_raw): New functions.
Get the next PC along with the instruction state.
(thumb_get_next_pc): Remove.
(arm_get_next_pc): Modified to use arm_get_next_pc_raw.

2010-03-04  Hui Zhu  <teawater@gmail.com>

* i386-tdep.c (i386_process_record): Change "addr" to "tmpu64".

PR 11302
* dwarf2.c (read_abbrevs): Return NULL on alloc failures.
(read_attribute_value, decode_line_info): Likewise.
(add_line_info, read_rangelist): Return FALSE on alloc failures.
(arange_add, sort_line_sequences): Likewise.
(find_abstract_instance_name): Handle failures from called funcs.
(scan_unit_for_symbols, parse_comp_unit, decode_line_info): Likewise.
(find_line): Realloc to a temp, and handle alloc fail.

* i386.cc (Target_i386::can_check_for_function_pointers): New function.
(Scan::possible_function_pointer_reloc): New function.
(Scan::local_reloc_may_be_function_pointer): Change to call
possible_function_pointer_reloc.
(Scan::global_reloc_may_be_function_pointer): Ditto.
* icf.h (Icf::check_section_for_function_pointers): Change to reject
relocations in ".data.rel.ro._ZTV" section.
* testsuite/icf_safe_so_test.sh: Change to pass i386.
* testsuite/icf_safe_so_test.cc: Ditto.
* testsuite/icf_safe_test.cc: Ditto.
* testsuite/icf_safe_test.sh: Ditto.

* utils.c (fputs_maybe_filtered): Always disable pagination if the
top level interpreter is MI.

2010-03-03 Stan Shebs  <stan@codesourcery.com>

* remote.c (remote_download_tracepoint): Iterate over locations.
* tracepoint.c (validate_actionline): Ditto.
(encode_actions): Add location argument.
(trace_dump_command): Check all locations to see if stepping
frame.

*** empty log message ***

daily update

* lib/gdb.exp (gdb_compile_pthreads): Handle case where
libc contains libpthread.

* gdb.threads/linux-dp.c (shared_random): Call rand instead of rand_r.
* gdb.threads/linux-dp.exp: Compile with gdb_compile_pthreads instead
of gdb_compile.  Add another pattern to match android backtrace.

Mention improvement for X86 general purpose registers.

2010-03-03  H.J. Lu  <hongjiu.lu@intel.com>
    Eli Zaretskii  <eliz@gnu.org>

* NEWS: Add X86 general purpose registers section.

* target-reloc.h (relocate_section): Check the symbol table index
for -1U before setting the local symbol index.
(scan_relocatable_relocs): If copying the relocation, record that
the local symbol is required.
* object.h (Symbol_value::is_output_symtab_index_set): New
function.
(Symbol_value::may_be_discarded_from_output_symtab): New
function.
(Symbol_value::has_output_symtab_entry): New function.
(Symbol_value::needs_output_symtab_entry): Remove.
(Symbol_value::output_symtab_index): Make sure the symbol index is
set.
(Symbol_value::set_output_symtab_index): Make sure the symbol
index is not set.  Make sure the new index is valid.
(Symbol_value::set_must_have_output_symtab_entry): New function.
(Symbol_value::has_output_dynsym_entry): New function.
(Symbol_value::set_output_dynsym_index): Make sure the new index
is valid.
(Sized_relobj::set_must_have_output_symtab_entry): New function.
* object.cc (Sized_relobj::do_count_local_symbols): Only discard a
local symbol if permitted.
(Sized_relobj::do_finalize_local_symbols): Call
is_output_symtab_index_set rather than needs_output_symtab_entry.
(Sized_relobj::write_local_symbols): Call has_output_symtab_entry
rather than needs_output_symtab_entry.  Call
has_output_dynsym_entry rather than needs_output_dynsym_entry.
* arm.cc (Arm_relobj::update_output_local_symbol_count): Call
is_output_symtab_index_set rather than needs_output_symtab_entry.
* testsuite/discard_locals_relocatable_test.c: New file.
* testsuite/discard_locals_test.sh: Test -r.
* testsuite/Makefile.am (check_DATA): Add
discard_locals_relocatable_test1.syms,
discard_local_relocatable_test2.syms.
(MOSTLYCLEANFILES): Likewise.  Also add
discard_locals_relocatable_test1.lout and
discard_locals_relocatable_test2.out.
(discard_locals_relocatable_test1.syms): New target.
(discard_locals_relocatable_test.o): New target.
(discard_locals_relocatable_test1.out): New target.
(discard_locals_relocatable_test2.syms): New target.
(discard_locals_relocatable_test2.out): New target.
(various): Add missing ../ld-new dependencies.
* testsuite/Makefile.in: Rebuild.

PR mi/11098:
* varobj.c (install_new_value): Handle case where new print_value
is NULL.

gdb
2010-03-03  Dainis Jonitis  <jonitis@gmail.com>
PR gdb/11345:
* printcmd.c (printf_command): Print end of format string using
printf_filtered.
gdb/testsuite
2010-03-03  Tom Tromey  <tromey@redhat.com>
PR gdb/11345:
* gdb.base/printcmds.exp (test_printf): Add test.

* elf32-arm.c (elf32_arm_output_arch_local_syms): Do not add
a mapping symbol to an empty section.

    * po/fi.po: New Finnish translation.

*** empty log message ***

daily update

* mi/mi-cmd-break.c (mi_read_next_line): Add missing 'void'.
* defs.h (read_command_lines_1): Add missing 'void'.
* cli/cli-script.c (recurse_read_control_structure): Add missing
'void'.
(read_next_line): Likewise.
(read_command_lines_1): Likewise.

* spu-tdep.c (spu_analyze_prologue): Track instruction to
store backchain as part of prologue.

* progspace.c (update_address_spaces): Update inferior address spaces
also.

* dwarf2read.c (add_partial_subprogram): Add missing baseaddr to
lowpc,highpc args to addrmap_set_empty.

* NEWS: Document .ARM.exidx / .ARM.extab support.

* dwarf.c (read_leb128): Make non-static.
* dwarf.h (read_leb128): Declare.
* readelf.c (REMOVE_ARCH_BITS): Define.
(find_section_by_address): New.
(read_uleb128): Move higher.  Use read_leb128 from dwarf.c.
(find_symbol_for_address): Handle the Thumb bit for ARM, by
using REMOVE_ARCH_BITS.
(struct arm_section, struct arm_unw_aux_info, arm_print_vma_and_name)
(arm_free_section, arm_section_get_word, decode_arm_unwind)
(dump_arm_unwind, arm_process_unwind): New.
(process_unwind): Handle ARM.

Support x86 pseudo byte, word and dword registers.

gdb/

2010-03-02  H.J. Lu  <hongjiu.lu@intel.com>

* amd64-tdep.c (amd64_byte_names): New.
(amd64_word_names): Likewise.
(amd64_dword_names): Likewise.
(amd64_pseudo_register_name): Likewise.
(amd64_pseudo_register_read): Likewise.
(amd64_pseudo_register_write): Likewise.
(amd64_init_abi): Set num_byte_regs, num_word_regs, num_dword_regs
and num_mmx_regs.  Call set_gdbarch_pseudo_register_read,
set_gdbarch_pseudo_register_write and
set_tdesc_pseudo_register_name.  Don't call
set_gdbarch_num_pseudo_regs.  Don't set mm0_regnum.

* i386-tdep.c (i386_num_mmx_regs): Removed.
(i386_num_pseudo_regs): Likewise.
(i386_byte_names): New.
(i386_word_names): Likewise.
(i386_byte_regnum_p): Likewise.
(i386_word_regnum_p): Likewise.
(i386_mmx_regnum_p): Updated.
(i386_pseudo_register_name): Make it global.  Handle byte and
word pseudo-registers.
(i386_pseudo_register_read): Likewise.
(i386_pseudo_register_write): Likewise.
(i386_pseudo_register_type): Handle byte, word and dword
pseudo-registers
(i386_register_reggroup_p): Don't include pseudo
registers, except for MXX, in any register groups.  Don't
include pseudo byte, word, dword registers in general_reggroup.
(i386_gdbarch_init): Set num_byte_regs, num_word_regs,
num_dword_regs, al_regnum, ax_regnum and eax_regnum.  Put MMX
pseudo-registers after word pseudo-registers.  Call
set_gdbarch_num_pseudo_regs after calling gdbarch_init_osabi.

* i386-tdep.h (gdbarch_tdep): Add num_mmx_regs, num_byte_regs,
al_regnum, num_word_regs, ax_regnum, num_dword_regs and
eax_regnum.
(i386_byte_regnum_p): New.
(i386_word_regnum_p): Likewise.
(i386_dword_regnum_p): Likewise.
(i386_pseudo_register_name): Likewise.
(i386_pseudo_register_read): Likewise.
(i386_pseudo_register_write): Likewise.

gdb/testsuite/

2010-03-02  H.J. Lu  <hongjiu.lu@intel.com>

* gdb.arch/amd64-byte.exp: New.
* gdb.arch/amd64-dword.exp: Likewise.
* gdb.arch/amd64-pseudo.c: Likewise.
* gdb.arch/amd64-word.exp: Likewise.
* gdb.arch/i386-byte.exp: Likewise.
* gdb.arch/i386-pseudo.c: Likewise.
* gdb.arch/i386-word.exp: Likewise.