* archive.cc: Include "plugin.h".
(Archive::setup): Don't preread archive symbols when using a plugin.
(Archive::get_file_and_offset): Add memsize parameter. Change callers.
(Archive::get_elf_object_for_member): Call plugin hooks for claiming
files.
(Archive::include_member): Add symbols from plugin objects.
* archive.h (Archive::get_file_and_offset): Add memsize parameter.
* descriptors.cc (Descriptors::open): Check for file descriptors
abandoned by plugins.
(Descriptors::claim_for_plugin): New function.
* descriptors.h (Descriptors::claim_for_plugin): New function.
(Open_descriptor::is_claimed): New field.
(claim_descriptor_for_plugin): New function.
* fileread.cc (File_read::claim_for_plugin): New function.
* fileread.h (File_read::claim_for_plugin): New function.
(File_read::descriptor): New function.
* gold.cc: Include "plugin.h".
(queue_initial_tasks): Add task to call plugin hooks for generating
new object files.
* main.cc: Include "plugin.h".
(main): Load plugin libraries.
* object.h (Pluginobj): Declare.
(Object::pluginobj): New function.
(Object::do_pluginobj): New function.
(Object::set_target): New function.
* options.cc: Include "plugin.h".
(General_options::parse_plugin): New function.
(General_options::General_options): Initialize plugins_ field.
(General_options::add_plugin): New function.
* options.h (Plugin_manager): Declare.
(General_options): Add --plugin option.
(General_options::has_plugins): New function.
(General_options::plugins): New function.
(General_options::add_plugin): New function.
(General_options::plugins_): New field.
* plugin.cc: New file.
* plugin.h: New file.
* readsyms.cc: Include "plugin.h".
(Read_symbols::do_read_symbols): Check for archive before checking
for ELF file. Call plugin hooks to claim files.
* resolve.cc (Symbol_table::resolve): Record when symbol is referenced
from a real object file; force override when processing replacement
files.
* symtab.cc (Symbol::init_fields): Initialize in_real_elf_ field.
(Symbol::init_base_object): Likewise.
(Symbol::init_base_output_data): Likewise.
(Symbol::init_base_output_segment): Likewise.
(Symbol::init_base_constant): Likewise.
(Symbol::init_base_undefined): Likewise.
(Symbol::output_section): Assert that object is not a plugin.
(Symbol_table::add_from_pluginobj): New function.
(Symbol_table::sized_finalize_symbol): Treat symbols from plugins as
undefined.
(Symbol_table::sized_write_globals): Likewise.
(Symbol_table::add_from_pluginobj): Instantiate template.
* symtab.h (Sized_pluginobj): Declare.
(Symbol::in_real_elf): New function.
(Symbol::set_in_real_elf): New function.
(Symbol::in_real_elf_): New field.
(Symbol_table::add_from_pluginobj): New function.
* testsuite/Makefile.am (AM_CFLAGS): New variable.
(LIBDL): New variable.
(LDADD): Add LIBDL.
(check_PROGRAMS): Add plugin_test_1 and plugin_test_2.
(check_SCRIPTS): Add plugin_test_1.sh and plugin_test_2.sh.
(check_DATA): Add plugin_test_1.err and plugin_test_2.err.
(MOSTLYCLEANFILES): Likewise.
* testsuite/Makefile.in: Regenerate.
* testsuite/plugin_test.c: New file.
* testsuite/plugin_test_1.sh: New file.
* testsuite/plugin_test_2.sh: New file.
Jan Kratochvil [Wed, 17 Sep 2008 21:48:49 +0000 (21:48 +0000)]
Fix a crash on uninitialized ECS->EVENT_THREAD for a newly found thread.
* infrun.c (wait_for_inferior): Move this ECS->EVENT_THREAD
initialization ...
(fetch_inferior_event): ... and this ECS->EVENT_THREAD initialization
...
(handle_inferior_event): ... here after the add_thread call together
with the local adjust_pc_after_break and reinit_frame_cache calls.
Jan Kratochvil [Wed, 17 Sep 2008 07:50:29 +0000 (07:50 +0000)]
bfd/
PR 6893 - Do not consider FDEs for discarded sections as invalid.
* elf-eh-frame.c (_bfd_elf_parse_eh_frame): New REQUIRE_CLEARED_RELOCS.
Consider FDEs with cleared relocations as valid and ignorable.
ld/testsuite/
* ld-elf/eh-group.exp, ld-elf/eh-group1.s, ld-elf/eh-group2.s: New test.
binutils/
Suppress warnings on NONE relocations to discarded sections.
* readelf.c (is_none_reloc): New function.
(debug_apply_relocations): Ignore is_none_reloc() relocations.
Jan Kratochvil [Wed, 17 Sep 2008 07:43:49 +0000 (07:43 +0000)]
Fix testcases compilation failures due to unused variables.
* ld-elfcomm/common1b.c (dummy1): Mark the variable for GCC as USED.
* ld-shared/main.c (main): Move the P variable only to the places where
it is used.
Ian Lance Taylor [Tue, 16 Sep 2008 17:23:37 +0000 (17:23 +0000)]
* target-reloc.h (relocate_section): Check whether a symbol is
defined by the ABI before reporting an undefined symbol error.
* target.h (Target::is_defined_by_abi): Make parameter const.
(Target::do_is_defined_by_abi): Likewise.
* i386.cc (Target_i386::do_is_defined_by_abi): Likewise.
* powerpc.cc (Target_powerpc::do_is_defined_by_abi): Likewise.
* sparc.cc (Target_sparc::do_is_defined_by_abi): Likewise.
* x86_64.cc (Target_x86_64::do_is_defined_by_abi): Likewise.
* testsuite/Makefile.am (tls_test_shared.so): Add -Wl,-z,defs.
* testsuite/Makefile.in: Rebuild.
Alan Modra [Tue, 16 Sep 2008 13:43:36 +0000 (13:43 +0000)]
PR 6844
* elf32-ppc.c (SYM_VAL): Define. Use throughout to find symbol vma.
(ppc_elf_relocate_section): Correct GOT offset calculation.
(ppc_elf_finish_dynamic_symbol): Use PPC_HA and PPC_LO.
(ppc_elf_finish_dynamic_sections): Likewise. Error if
htab->elf.hgot symbol is not defined in htab->got section.
Alan Modra [Mon, 15 Sep 2008 05:22:32 +0000 (05:22 +0000)]
* config/tc-frv.c (md_apply_fix): Use abs_section_sym for
relocs with no symbol.
* config/tc-mmix.c (md_assemble): Mark fake symbol on
BFD_RELOC_MMIX_BASE_PLUS_OFFSET as OK for use by relocs.
(mmix_md_end): Likewise mark mmix reg contents section symbol.
Tom Tromey [Sat, 13 Sep 2008 23:37:05 +0000 (23:37 +0000)]
* varobj.c (varobj_set_display_format): Use xfree.
* tracepoint.c (stringify_collection_list): Use xfree.
* remote-fileio.c (remote_fileio_reset): Use xfree.
* mipsread.c (read_alphacoff_dynamic_symtab): Use xfree.
* dfp.c (decimal_from_floating): Use xfree, xstrprintf. Don't use
asprintf.
* cp-support.c (mangled_name_to_comp): Use xfree.
Joel Brobecker [Sat, 13 Sep 2008 22:27:00 +0000 (22:27 +0000)]
* ada-lang.c (remove_extra_symbols): Remove stub symbols if
the associated complete symbol is also in the list.
(ada_add_local_symbols, ada_add_non_local_symbols): New functions,
extracted out from ada_lookup_symbol_list.
(ada_lookup_symbol_list): Use them. Remove the search through
the minimal symbols.
Joel Brobecker [Sat, 13 Sep 2008 22:19:51 +0000 (22:19 +0000)]
* dwarf2read.c (add_partial_subprogram): New procedure.
(scan_partial_symbols): Use it.
(load_partial_dies): Read in children of subprogram and lexical
blocks for Ada compilation units.
* gdb.mi/mi-nonstop.exp: Do not check thread state while a
stop is pending. Avoid ".*" when two stops are pending.
* lib/gdb.exp (fullname_syntax_POSIX, fullname_syntax_UNC)
(fullname_syntax_DOS_CASE, fullname_syntax_DOS): Do not match
newlines in fullnames.
* lib/mi-support.exp (mi_run_cmd): Do not require an anchor.
(mi_expect_stop): Update comments. Only anchor in sync mode.
Do not match newlines.
(mi_send_resuming_command_raw): Always return status.
(mi_get_stop_line): Do not match more than one line by accident.
Only anchor in sync mode.
(mi_run_inline_test): If -exec-next fails, give up.
* dwarf2read.c (comp_unit_head): Rename first_die_ptr to
first_die_offset. All uses updated.
Delete unused members cu_head_ptr, next.
Move members base_known, base_address to ...
(dwarf2_cu) ... here. All uses updated.
Bob Wilson [Fri, 12 Sep 2008 18:53:55 +0000 (18:53 +0000)]
2008-09-12 Sterling Augustine <sterling@tensilica.com>
* config/tc-xtensa.c (init_op_placement_info_table): Allow number of
operands equal to MAX_INSN_ARGS.
Chris Demetriou [Fri, 12 Sep 2008 05:42:29 +0000 (05:42 +0000)]
2008-09-11 Chris Demetriou <cgd@google.com>
* options.h (origin): New -z option.
* layout.cc (Layout:finish_dynamic_section): If "-z origin"
is specified, set DF_ORIGIN in DT_FLAGS and set DF_1_ORIGIN
in DT_FLAGS_1.
* bsd-uthread.c (bsd_uthread_read_memory_address): New function.
(bsd_uthread_fetch_registers, bsd_uthread_store_registers,
bsd_uthread_wait, bsd_uthread_find_new_threads): Use it.
* procfs.c (procfs_address_to_host_pointer): Use target_gdbarch
and its associated types to perform pointer conversion.
(procfs_can_use_hw_breakpoint): Likewise.
(procfs_auxv_parse): Remove unused variable.
* valops.c: Include "objfiles.h" and "symtab.h".
(find_function_in_inferior): New argument OBJF_P. Use it to return
objfile where function is defined. Use per-objfile arch types
instead of builtin_type_ to define default return type.
* valarith.c (value_x_unop): Use builtin_type_int8 as type for
UNOP_POSTINCREMENT/UNOP_POSTDECREMENT constant 0 argument.
(value_bit_index): Use extract_unsigned_integer
instead of unpack_long to read single byte.
* infcall.c (value_arg_coerce): Add GDBARCH parameter. Use its
associates types instead of builtin_type_ macros.
(find_function_addr): Leave output VALUE_TYPE NULL if unknown.
(call_function_by_hand): Use per-architecture "int" type as
fall-back if find_function_addr returns NULL VALUE_TYPE.
Update call to value_arg_coerce.
* cp-abi.h (cplus_method_ptr_size): Add TO_TYPE parameter.
(cplus_make_method_ptr): Add TYPE parameter.
* cp-abi.c (cplus_method_ptr_size): Add TO_TYPE parameter. Pass it
on to current_cp_abi.method_ptr_size callback.
(cplus_make_method_ptr): Add TYPE parameter. Pass it on to
current_cp_abi.make_method_ptr callback.
* gdbtypes.c (lookup_methodptr_type): Pass target type
argument to cplus_method_ptr_size.
* valops.c (value_cast): Pass type argument to cplus_make_method_ptr.
(value_struct_elt_for_reference): Likewise.
* gnu-v3-abi.c (get_class_arch): New function.
(vtable_address_point_offset): Add GDBARCH parameter. Use it
instead of current_gdbarch. Update all callers.
(gnuv3_get_vtable): Likewise.
(gnuv3_get_virtual_fn): Likewise.
(gnuv3_decode_method_ptr): Likewise.
(gnuv3_rtti_type): Call get_class_arch to determine architecture.
Use it instead of current_gdbarch.
(gnuv3_virtual_fn_field): Likewise.
(gnuv3_baseclass_offset): Likewise.
(gnuv3_print_method_ptr): Likewise.
(gnuv3_method_ptr_to_value): Likewise.
(gnuv3_method_ptr_size): Add TYPE parameter. Use it to determine
class architecture. Use architecture types instead of builtin types.
(gnuv3_make_method_ptr): Likewise.
* cp-valprint.c (cp_print_class_member): Expect pointer type
instead of class type. Use its length when extracting value.
* c-valprint.c (c_val_print): Update call to cp_print_class_member.
* stack.c (return_command): Use frame architecture to determine
default integer return type.
* f-valprint.c (f77_get_dynamic_lowerbound): Use frame architecture
to determine pointer types.
(f77_get_dynamic_upperbound): Likewise.
* objc-lang.c (OBJC_FETCH_POINTER_ARGUMENT): Remove.
(resolve_msgsend): Use architecture of current frame to determine
pointer types. Inline OBJC_FETCH_POINTER_ARGUMENT.
(resolve_msgsend_stret, resolve_msgsend_super,
resolve_msgsend_super_stret): Likewise.
* gdbtypes.h (struct builtin_type): Remove builtin_true_char
and builtin_true_unsigned_char.
(builtin_type_true_char): Remove macro, add extern declaration.
(builtin_type_true_unsigned_char): Add extern declaration.
* gdbtypes.c (builtin_type_true_char): New global variable.
(builtin_type_true_unsigned_char): Likewise.
(_initialize_gdbtypes): Initialize them.
(gdbtypes_post_init): Do not initialize builtin_true_char
and builtin_true_unsigned_char members of struct builtin_type.
* printcmd.c (print_scalar_formatted): Do not use builtin_type;
use builtin_type_true_unsigned_char instead.
* ada-valprint.c (ada_val_print_1): Use builtin_type_true_char
instead of builtin_type_char for internal string.
* ada-valprint.c: Include "objfiles.h".
(ada_val_print_1): Use the gdbarch associated with the objfile whether
a System.Address type is defined to retrieve the proper pointer type
to use to print it.
* ada-lang.c (value_pos_atr): Add TYPE argument. Use it as
result type instead of builtin_type_int.
(value_subscript_packed): Use pos_atr instead of value_pos_atr.
(ada_value_subscript): Update call to value_pos_atr.
(ada_value_ptr_subscript): Likewise.
(ada_evaluate_subexp): Likewise.
* ada-lang.c (cast_to_fixed): Do not cast to builtin_type_double.
(cast_from_fixed_to_double): Rename to ...
(cast_from_fixed): ... this. Add TYPE parameter. Use it instead
of builtin_type_double.
(ada_value_cast): Use cast_from_fixed instead of casting result
of cast_from_fixed_to_double.
(ada_evaluate_subexp): Update calls to cast_from_fixed_to_double.
* ax-gdb.c: Include "language.h".
(gen_frame_args_address): Add GDBARCH parameter; use it
instead of current_gdbarch.
(gen_frame_locals_address): Likewise.
(gen_var_ref): Add GDBARCH parameter. Update calls to
gen_frame_args_address and gen_frame_locals_address. Use
pointer type from gdbarch.
(gen_usual_unary): Add EXP parameter. Use integer type
from exp->gdbarch.
(gen_usual_arithmetic): Likewise.
(gen_integral_promotions): Likewise.
(gen_add, gen_sub): Remove.
(gen_ptradd, gen_ptrsub, gen_ptrdiff): New functions.
(gen_logical_not): Use passed-in boolean result type
instead of builtin_type_int.
(gen_complement): Do not call gen_usual_unary or
gen_integral_promotions.
(gen_struct_ref): Call require_rvalue instead of gen_usual_unary.
(gen_repeat): Add EXP parameter. Update call to gen_expr.
Use builtin_type_int32 as internal range type.
(gen_sizeof): Add EXP and SIZE_TYPE parameters. Use SIZE_TYPE
as result type. Update call to gen_expr.
(gen_expr): Add EXP parameter. Update calls to gen_expr,
gen_repeat, gen_var_ref, gen_usual_unary, gen_usual_arithmetic,
and gen_integral_promotions. Call gen_ptradd, gen_ptrsub,
gen_ptrdiff, or gen_binop instead of gen_add or gen_sub.
Use exp->gdbarch instead of current_gdbarch.
Call language_bool_type to determine result type of UNOP_LOGICAL_NOT.
* eval.c (evaluate_subexp_standard): Add calls to binop_promote
and unop_promote before calling value_binop et. al.
* ada-lang.c (ada_evaluate_subexp): Add calls to binop_promote
and unop_promote before calling value_binop et. al.
* valarith.c (value_binop): Do not call binop_promote or unop_promote.
(value_pos): Do not call unop_promote.
(value_neg, value_complement): Likewise.
* eval.c (ptrmath_type_p): New function.
(evaluate_subexp_standard): Replace value_add and value_sub
by value_ptradd, value_ptrsub, value_ptrdiff or value_binop.
Use builtin_type_uint8 instead of builtin_type_char to hold
the increment for BINOP_{PRE,POST}{IN,DE}CREMENT operations.
* valarith.c (value_subscript): Replace value_add by
value_ptradd. Replace value_sub by value_binop.
* ada-lang.c (ada_value_ptr_subscript): Likewise.
(ada_tag_name_2): Replace value_add by value_ptradd.
(ada_evaluate_subexp): Replace value_add and value_sub by
value_binop.
* m2-lang.c (evaluate_subexp_modula2): Replace value_add
by value_ptradd.
* gnu-v2-abi.c (gnuv2_virtual_fn_field): Likewise.
* gnu-v3-abi.c (gnuv3_method_ptr_to_value): Likewise.
* expression.h (enum exp_opcode): Document OP_COMPLEX to take
a type parameter as expression element.
* eval.c (evaluate_subexp_standard) [OP_COMPLEX]: Retrieve result
type as expression element.
* f-exp.y: Pass in type when buildin OP_COMPLEX expression.
* parse.c (operator_length_standard): Update length of OP_COMPLEX.
* language.h (struct language_arch_info): New members
bool_type_default and bool_type_symbol.
(lang_bool_type): Remove prototype.
(LA_BOOL_TYPE): Remove macro.
(language_bool_type): Add prototype.
* language.c (lang_bool_type): Remove.
(language_bool_type): New function.
* value.h (value_in): Change return value to int.
* value.c (value_in): Return int instead of struct value *.
* eval.c (evaluate_subexp_standard): Call language_bool_type instead
of using LA_BOOL_TYPE. Update call to value_in.
* ada-lang.c (ada_evaluate_subexp): Call language_bool_type instead
of using LA_BOOL_TYPE or builtin_type_int for boolean values.
* language.c (unknown_language_arch_info): Set bool_type_default member
of struct language_arch_info.
* ada-lang.c (ada_language_arch_info): Set bool_type_symbol and
bool_type_default members of struct language_arch_info.
* c-lang.c (c_language_arch_info): Set bool_type_default member
of struct language_arch_info.
(cplus_language_arch_info): Set bool_type_symbol and bool_type_default
members of struct language_arch_info.
* f-lang.c (f_language_arch_info): Set bool_type_symbol and
bool_type_default members of struct language_arch_info.
* jv-lang.c (java_language_arch_info): Set bool_type_symbol and
bool_type_default members of struct language_arch_info.
* m2-lang.c (m2_language_arch_info): Set bool_type_symbol and
bool_type_default members of struct language_arch_info.
* p-lang.c (p_language_arch_info): Set bool_type_symbol and
bool_type_default members of struct language_arch_info.
* expression.h (struct expression): New member GDBARCH.
* parse.c (parse_exp_in_context): Initialize it.
* parser-def.h (parse_gdbarch, parse_language): New macros.
* ada-exp.y (parse_type): New macro.
Replace builtin_type_ macros by using parse_type.
Replace current_language by parse_language.
* ada-lex.l (processInt): Replace current_gdbarch by parse_gdbarch.
Replace builtin_type_ macros.
* c-exp.y (parse_type): New macro.
Replace builtin_type_ macros by using parse_type.
(parse_number): Replace current_gdbarch by parse_gdbarch.
(yylex): Replace current_language by parse_language.
* f-exp.y (parse_type, parse_f_type): New macros.
Replace builtin_type_ macros by using parse_{f_,}type.
(parse_number): Replace current_gdbarch by parse_gdbarch.
(yylex): Replace current_language by parse_language.
* jv-exp.y (parse_type): New macro.
(parse_number): Replace builtin_type_ macros by using parse_type.
* m2-exp.y (parse_type, parse_m2_type): New macros.
Replace builtin_type_ macros by using parse_{m2_,}type.
* objc-exp.y (parse_type): New macro.
Replace builtin_type_ macros by using parse_type.
(parse_number): Replace current_gdbarch by parse_gdbarch.
(yylex): Replace current_language by parse_language.
* p-exp.y (parse_type): New macro.
Replace builtin_type_ macros by using parse_type.
(parse_number): Replace current_gdbarch by parse_gdbarch.
(yylex): Replace current_language by parse_language.
projects / thirdparty / binutils-gdb.git / log
