thread-specific breakpoints not saved properly by save-breakpoint

Initially, I noticed that the save command was often missing
new lines in the file that it generated.  For instance, consider:

        % gdb save-bp
        (gdb) b break_me
        (gdb) b save-bp.c:27
        (gdb) save breakpoints bps

The contents of the bps file would be:

        % cat bps
        break break_mebreak save-bp.c:27

Looking further into the problem, I realized that the missing newlines
are just a consequence of a missing call to print_recreate_thread.
After having generated the breakpoint location in the break command,
we cannot put a new line until we have looked at whether we need to
add a 'thread NUM' argument.

gdb/ChangeLog:

        * breakpoint.c (bkpt_print_recreate): Add call to
        print_recreate_thread.

gdb/testsuite/ChangeLog:

        * gdb.base/save-bp.exp, gdb.base/save-bp.c: New files.

*** empty log message ***

daily update

gdb/testsuite/
* gdb.python/py-value.exp (test_value_in_inferior): New variable
can_read_0, test for it.
(python print gdb.parse_and_eval('*(int*)0')): Rename to ...
(parse_and_eval with memory error): ... here, make it untested if
can_read_0.

*** empty log message ***

daily update

* binutils/dwarf.h (dwarf_section_display_enum): Add missing enum
constant.

gdb.python/python.exp: fix up race

2011-09-30  Tristan Gingold  <gingold@adacore.com>

* dwarf2.c (struct dwarf2_debug): Add field debug_sections.
(read_section): Add sec argument. Adjust the code to get section
names from it.
(read_indirect_string): Adjust call to read_section.
(read_abbrevs): Ditto.
(decode_line_info): Ditto.
(read_debug_ranges): Ditto.
(DWARF2_DEBUG_INFO, DWARF2_COMPRESSED_DEBUG_INFO): Remove.
(find_debug_info): Add debug_sections parameter.  Use it instead
the above macros.
(place_sections): Get section names from stash.
(find_line): Add debug_sections argument.  Initialize
debug_sections field of stash from it.
Adjust calls to find_debug_info.
(_bfd_dwarf2_find_nearest_line): Add debug_sections argument.
Adjust call to find_line.
(_bfd_dwarf2_find_line): Adjust call to find_line.
* libcoff-in.h (struct dwarf_debug_section): New declaration.
(coff_find_nearest_line_with_names): Likewise.
* libcoff.h: Regenerate.
* libbfd-in.h (struct dwarf_debug_section): Move declaration.
(dwarf_debug_sections): Likewise.
(_bfd_dwarf2_find_nearest_line): Add debug_sections argument.
* libbfd.h: Regenerate.
* elf.c (_bfd_elf_find_nearest_line): Adjust call to
_bfd_dwarf2_find_nearest_line.
* elfxx-mips.c (_bfd_mips_elf_find_nearest_line): Ditto.
* elf64-alpha.c (elf64_alpha_find_nearest_line): Ditto.
* elf32-arm.c (elf32_arm_find_nearest_line): Ditto.
* coffgen.c (coff_find_nearest_line_with_names): New function.
(coff_find_nearest_line): Calls coff_find_nearest_line_with_names.
* coff-rs6000.c (xcoff_debug_sections): New constant.
(xcoff_find_nearest_line): New function.
(rs6000coff_vec): Use xcoff_find_nearest_line.
(pmac_xcoff_vec): Ditto.

PR ld/13235
bfd/
* elf64-ppc.c (struct ppc64_elf_obj_tdata): Add ha_relocs_not_using_r2.
(ppc64_elf_edit_toc): Check HA relocs.
(ha_reloc_match): Delete function.
(ppc64_elf_relocate_section): Remove delayed HA nop optimization.
Instead do it and low part optimization based on
ha_relocs_not_using_r2.
ld/testsuite/
* ld-powerpc/tocopt.d: Update.
* ld-powerpc/tocopt5.d, * ld-powerpc/tocopt5.s: New test.
* ld-powerpc/powerpc.exp: Run new test.

gdbserver: add support for FDPIC loadmaps

The DSBT support is very close to the FDPIC code, so extend the existing
loadmap support to work with FDPIC loadmaps too.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>

*** empty log message ***

daily update

2011-09-29 Sriraman Tallam  <tmsriram@google.com>

* layout.h (section_order_map_): New member.
(get_section_order_map): New member function.
* output.cc (Output_section::add_input_section): Check for patterns
only when --section-ordering-file is specified.
* gold.cc (queue_middle_tasks): Delay updating order of sections till
output_sections have been formed.
* layout.cc (Layout_Layout): Initialize section_order_map_.
* plugin.cc (update_section_order): Store order in order_map. Do not
update the order.
* testsuite/Makefile.am: Add test case for plugin_final_layout.
* testsuite/Makefile.in: Regenerate.
* testsuite/plugin_section_order.c: New file.
* testsuite/plugin_final_layout.cc: New file.
* testsuite/plugin_final_layout.sh: New file.

* gold/incremental.cc (Sized_incremental_binary::do_process_got_plt):
Check for NULL.
* gold/symtab.cc (Symbol_table::add_from_relobj): Ignore version
symbols during incremental update.
(Symbol_table::add_from_dynobj): Likewise.

* lib/gdb.exp (is_amd64_regs_target, is_x86_like_target): Allow
for 64-bit multilibs from i?86-* targets.

lib/gdb.exp: Fix a typo

gdb.base/foll-fork.exp: fix up race

PR ld/13233
* elflink.c (_bfd_elf_gc_mark_extra_sections): Mark single member
debug and special section groups.

sim: bfin: use store buffer for VIT_MAX insns

The VIT_MAX insns can be used in parallel, so we need to use the store
buffer so we don't clobber the output register before we get a chance
to do a memory store with it.

Reported-by: Kai Iskratsch <kai@stella.at>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>

gdb/doc/
* gdb.texinfo (Files): Update options for `add-symbol-file'.
Add one space after option `-s'.  Remove @r{} markup.

gdb/
* symfile.c (add_symbol_file_command): Update message on usage.

Fix --gc-sections failure with symbol versioning.

bfd/

2011-09-28  H.J. Lu  <hongjiu.lu@intel.com>

PR ld/13195
* elflink.c (_bfd_elf_merge_symbol): Don't set dynamic_def when
clearing def_dynamic.
(elf_link_add_object_symbols): Likewise.  Set dynamic_def when
setting def_dynamic.
(bfd_elf_gc_mark_dynamic_ref_symbol): Check if a symbol is
versioned.

ld/testsuite/

2011-09-28  H.J. Lu  <hongjiu.lu@intel.com>

PR ld/13195
* ld-elf/elf.exp: Add a C link testcase for PR ld/13195.

* ld-elf/pr13195.c: New.
* ld-elf/pr13195.d: Likewise.
* ld-elf/pr13195.s: Likewise.
* ld-elf/pr13195.t: Likewise.

*** empty log message ***

daily update

* gdb.python/py-type.c (enum E): New.
* gdb.python/py-type.exp (test_fields): Add tests for Python
mapping access to fields.
(test_enums): New test for field access on enums.

* python/py-type.c (make_fielditem, typy_field_names, typy_items)
(typy_length, typy_get, typy_has_key, typy_make_iter)
(typy_iterkeys, typy_iteritems, typy_itervalues, typy_iter)
(typy_iterator_iter, typy_iterator_iternext)
(typy_iterator_dealloc): New functions to implement standard
Python mapping methods on gdb.Type object.
(gdb.TypeIterator): New Python type.
* python/python-internal.h (gdbpy_iter_kind): New enum.
* doc/gdb.texinfo (gdb.Type): Document field access by dictionary
key syntax.

merge from gcc

Fix complex floats on sparc.

* sparc-tdep.h (SPARC_F2_REGNUM, SPARC_F3_REGNUM, SPARC_F4_REGNUM,
SPARC_F5_REGNUM, SPARC_F6_REGNUM, SPARC_F7_REGNUM): New enums.
* sparc-tdep.c (sparc_complex_floating_p): New function.
(sparc32_store_arguments): Handle complex floats.
(sparc32_extract_return_value): Likewise.
(sparc32_store_return_value): Likewise.
(sparc32_stabs_argument_has_addr): Likewise.
* sparc64-tdep.c (sparc64_complex_floating_p): New function.
(sparc64_store_floating_fields): Handle complex floats.
(sparc64_store_arguments): Likewise.
(sparc64_store_return_value): Likewise.

* binutils/testsuite/binutils-all/elfedit-4.d: Give test a unique name.

Another typo ... :(

    * gas/pe/section-exclude.d: Correct testcase.

Support "set environment" and "unset environment" in MinGW builds.

* windows-nat.c (env_sort) [!__CYGWIN__]: Function restored from
before the change on 2006-12-09.
(windows_create_inferior) [!__CYGWIN__]: Restore code that
generates the environment block for CreateProcessA, modulo the
Cygwin-specific parts that are not needed here.

2011-09-28  Tristan Gingold  <gingold@adacore.com>

* od-xcoff.c (dump_xcoff32_aout_header): Fix typo.

* symtab.cc (Symbol_table::define_special_symbol): Always
canonicalize version string.

*** empty log message ***

daily update

2011-09-27  Kai Tietz  <ktietz@redhat.com>

       * config/obj-coff.c (obj_coff_section): Add 'e' as specifier
       for marking section SEC_EXCLUDE.

2011-09-27  Kai Tietz  <ktietz@redhat.com>

       * gas/pe/pe.exp: Add new testcase.
       * gas/pe/section-exclude.d: New file.
       * gas/pe/section-exclude.s: New file.

* coffcode.h (sec_to_styp_flags): Handle
SEC_LINK_DUPLICATES_SAME_CONTENTS, and
SEC_LINK_DUPLICATES_SAME_SIZE.

    * scripttempl/pe.sc (.text): Add support for
        .gnu.linkonce.t.* section.
        (.debug_info): Only use .gnu.linkonce.wi.* for
        relocating.
        * scripttempl/pep.sc: Likewise.

2011-09-27  Tristan Gingold  <gingold@adacore.com>

* target.h (enum target_object): Add TARGET_OBJECT_DARWIN_DYLD_INFO.
* solib-darwin.c (DYLD_VERSION_MAX): Update number.
(darwin_solib_get_all_image_info_addr_at_init): New function.
(darwin_solib_read_all_image_info_addr): Likewise.
(darwin_solib_create_inferior_hook): Use the above two functions.
* darwin-nat.c (darwin_execvp): Renames retval to res.
(darwin_read_write_inferior): Update comment.
(darwin_read_dyld_info): New function.
(darwin_xfer_partial): Handle DYLD_INFO.

Add return address collection for tracepoints.
* tracepoint.c (encode_actions_1): Add case for $_ret.
(validate_actionline): Check for $_ret.
(trace_dump_actions): Ditto.
* ax-gdb.h (gen_trace_for_return_address): Declare.
* ax-gdb.c: Include arch-utils.h.
(gen_trace_for_return_address): New function.
(agent_command): Add return address special case.
* amd64-tdep.c: Include ax.h and ax-gdb.h.
(amd64_gen_return_address): New function.
(amd64_init_abi): Call it.
* i386-tdep.c: Include ax.h and ax-gdb.h.
(i386_gen_return_address): New function.
(i386_init_abi): Call it.
* arch-utils.h (default_gen_return_address): Declare.
* arch-utils.c (default_gen_return_address): New function.
* gdbarch.sh (gen_return_address): New method.
* gdbarch.h, gdbarch.c: Regenerate.

* gdb.texinfo (Tracepoint Action Lists): Document $_ret.

* gdb.trace/collection.exp: Test collection of $_ret.

2011-09-27  Tristan Gingold  <gingold@adacore.com>

* od-xcoff.c (dump_xcoff32_aout_header): Remove some gettext macros.
(dump_xcoff32_sections_header): Likewise.
(dump_xcoff32_symbols, dump_xcoff32_relocs): Likewise.
(dump_xcoff32_lineno, dump_xcoff32_loader): Likewise.
(dump_xcoff32_except): Likewise.
(dump_xcoff32_typchk, dump_xcoff32_tbtags): Likewise.

2011-09-27  Tristan Gingold  <gingold@adacore.com>

* readelf.c (print_ia64_vms_note): Fix xgettext warnings.

opcodes/

* sparc-opc.c (sparc_opcodes): Fix random instruction to write
to a float instead of an integer register.

gas/testsuite/

* gas/sparc/hpcvis3.s: Update to use float reg for random insn.
* gas/sparc/hpcvis3.d: Likewise.

merge from gcc

daily update

*** empty log message ***

* gold/gold.cc (queue_initial_tasks): Move option checks ...
* gold/options.cc (General_options::finalize): ... to here. Disable
some options; make others fatal.

merge from gcc

include/ChangeLog

gcc PR lto/47247
* plugin-api.h (enum ld_plugin_symbol_resolution): Add
LDPR_PREVAILING_DEF_IRONLY_EXP.
(enum ld_plugin_tag): Add LDPT_GET_SYMBOLS_V2.

gold/ChangeLog

gcc PR lto/47247
* plugin.cc (get_symbols_v2): New function.
(Plugin::load): Add LDPT_GET_SYMBOLS_V2.
(is_referenced_from_outside): New function.
(Pluginobj::get_symbol_resolution_info): Add version parameter, return
LDPR_PREVAILING_DEF_IRONLY_EXP when using new version.
(get_symbols): Pass version parameter.
(get_symbols_v2): New function.
* plugin.h (Pluginobj::get_symbol_resolution_info): Add version
parameter.
* testsuite/plugin_test.c (get_symbols_v2): New static variable.
(onload): Add LDPT_GET_SYMBOLS_V2.
(all_symbols_read_hook): Use get_symbols_v2; check for
LDPR_PREVAILING_DEF_IRONLY_EXP.
* testsuite/plugin_test_3.sh: Update expected results.

Add sparc integer multiply-add instructions.

opcodes/

* sparc-opc.c (sparc_opcodes): Add integer multiply-add
instructions.

gas/testsuite/

* gas/sparc/ima.d: New test.
* gas/sparc/ima.s: New test source.
* gas/sparc/sparc.exp: Run new test.

*** empty log message ***

daily update

daily update

*** empty log message ***

daily update

*** empty log message ***

merge from gcc

merge from gcc

* configure.ac: Add new --with-gold-ldadd and --with-gold-ldflags
configuration options.
* configure: Regenerate.
* Makefile.am: Handle GOLD_LDADD and GOLD_LDFLAGS.
* Makefile.in: Regenerate.
* testsuite/Makefile.in: Regenerate.

PR gdb/13079
* i386-tdep.c (i386_frame_align): New.
(i386_gdbarch_init): Use i386_frame_align.

gdb/
* i386-linux-nat.c (i386_linux_resume): Use read_memory_unsigned_integer
to get address.

daily update

*** empty log message ***

2011-09-22  Tristan Gingold  <gingold@adacore.com>

* fork-child.c (fork_inferior): Add exec_fun parameter.
Call exec_fun or execvp.
* inferior.h: Adjust prototype.
* gnu-nat.c (gnu_create_inferior): Adjust fork_inferior call.
* inf-ttrace.c (inf_ttrace_create_inferior): Ditto.
* inf-ptrace.c (inf_ptrace_create_inferior): Ditto.
* procfs.c (procfs_create_inferior): Ditto.
* darwin-nat.c (darwin_execvp): New function.
(darwin_create_inferior): Use it.

bfd/
2011-09-22  Tristan Gingold  <gingold@adacore.com>

* configure.in: Bump version to 2.22.51
* configure: Regenerate.

binutils/
2011-09-22  Tristan Gingold  <gingold@adacore.com>

* NEWS: Add marker for 2.22.

gas/
2011-09-22  Tristan Gingold  <gingold@adacore.com>

* NEWS: Add marker for 2.22.

ld/
2011-09-22  Tristan Gingold  <gingold@adacore.com>

* NEWS: Add marker for 2.22.

2011-09-22  Andreas Tobler  <andreast@fgznet.ch>

* lib/gdb.exp (gdb_compile): Set rpath and remove -ldl from the
list of compilation switches for <*-*-freebsd*>.

gdb/
* infrun.c (context_switch): Print debug message when switching to
a different thread.

Add new sparc options to control instruction availability.

gas/

* config/tc-sparc.c (hwcap_allowed): New.
(struct sparc_arch): New field 'hwcap_allowed' containing a bitmask
of F_FOO flags which are enabled by the particular arch setting.
Add new options that provide explicit access to new instructions.
(md_parse_option): Only bump max_architecture if the requested one
is larger, or this is the first explicit request.
(get_hwcap_name): New function.
(sparc_ip): Validate that hwcaps used by an instruction have actually
been enabled.
* doc/c-sparc.texi: Document new sparc options.

daily update

*** empty log message ***

Fix sparc testcases when building with 64-bit default.

gas/testsuite/

* gas/sparc/imm-plus-rreg.d: Fix address regex for 64-bit.
* gas/sparc/save-args.d: Likewise.
* gas/sparc/ticc-imm-reg.d: Likewise, add -32 to options.
* gas/sparc/v8-movwr-imm.d: Likewise.

Annotate sparc objects with cpu hardware capabilities used.

bfd/

* elfxx-sparc.c (_bfd_sparc_elf_merge_private_bfd_data): New.
* elfxx-sparc.h: Declare it.
* elf32-sparc.c (elf32_sparc_merge_private_bfd_data): Call it.
* elf64-sparc.c (elf64_sparc_merge_private_bfd_data): Likewise.

binutils/

* readelf.c (display_sparc_hwcaps): New.
(display_sparc_gnu_attribute): New.
(process_sparc_specific): New.
(process_arch_specific): When EM_SPARC, EM_SPARC32PLUS,
or EM_SPARCV9 invoke process_sparc_specific.

gas/

* config/tc-sparc.c (hwcap_seen): New bitmask, defined when
not TE_SOLARIS.
(sparc_ip): When not TE_SOLARIS, accumulate hwcap bits from
sparc_opcode->flags of instruction into hwcap_seen.
(sparc_md_end): Create Tag_GNU_Sparc_HWCAPS attribute if
hwcap_seen is non-zero and not TE_SOLARIS.

gas/testsuite/

* gas/sparc/hpcvis3.s: Update for fixed fchksum16 mnemonic.
* gas/sparc/hpcvis3.d: Likewise.

include/elf/

* sparc.h (Tag_GNU_Sparc_HWCAPS): New object attribute.
(ELF_SPARC_HWCAP_*): New HWCAPS bitmask values.

include/opcode/

* sparc.h (struct sparc_opcode): Expand 'flags' to unsigned int.
(F_MUL32, F_DIV32, F_FSMULD, F_V8PLUS, F_POPC, F_VIS, F_VIS2,
F_ASI_BLK_INIT, F_FMAF, F_VIS3, F_HPC, F_RANDOM, F_TRANS,
F_FJFMAU, F_IMA, F_ASI_CACHE_SPARING): New flag bits.

opcodes/

* sparc-opc.c (sparc_opcodes): Annotate table with HWCAP flag
bits.  Fix "fchksm16" mnemonic.

* MAINTAINER: Take over from Jakub Jalinek as SPARC maintainer.

* s390-tdep.c (s390_function_arg_pass_by_reference): Handle
complex and vector types.
(s390_return_value_convention): Likewise.

(s390_value_from_register): Call check_typedef.
(extend_simple_arg): Likewise.
(alignment_of): Likewise.
(s390_push_dummy_call): Likewise.
(s390_return_value): Likewise.

Fix ChangeLog dates.

* linux-arm-low.c (struct arm_linux_hwbp_cap): Remove.
(arm_linux_hwbp_cap): New static variable.
(arm_linux_get_hwbp_cap): Replace by ...
(arm_linux_init_hwbp_cap): ... this new function.
(arm_linux_get_hw_breakpoint_count): Use arm_linux_hwbp_cap.
(arm_linux_get_hw_watchpoint_count): Likewise.
(arm_linux_get_hw_watchpoint_max_length): Likewise.
(arm_arch_setup): Call arm_linux_init_hwbp_cap.
(arm_prepare_to_resume): Use perror_with_name instead of error.

* event-top.c (async_disconnect): If an exception is thrown from
quit_cover, call pop_all_targets.  Use TRY_CATCH instead of
catch_errors.
* top.c (quit_cover): Return void and take no arguments.
* top.h (quit_cover): Update prototype.

* linux-arm-low.c: Include <signal.h>.
(PTRACE_GETHBPREGS, PTRACE_SETHBPREGS): Define if necessary.
(struct arm_linux_hwbp_cap): New data type.
(arm_hwbp_type, arm_hwbp_control_t): New typedefs.
(struct arm_linux_hw_breakpoint): New data type.
(MAX_BPTS, MAX_WPTS): Define.
(struct arch_process_info, struct arch_lwp_info): New data types.
(arm_linux_get_hwbp_cap): New function.
(arm_linux_get_hw_breakpoint_count): Likewise.
(arm_linux_get_hw_watchpoint_count): Likewise.
(arm_linux_get_hw_watchpoint_max_length): Likewise.
(arm_hwbp_control_initialize): Likewise.
(arm_hwbp_control_is_enabled): Likewise.
(arm_hwbp_control_is_initialized): Likewise.
(arm_hwbp_control_disable): Likewise.
(arm_linux_hw_breakpoint_equal): Likewise.
(arm_linux_hw_point_initialize): Likewise.
(struct update_registers_data): New data structure.
(update_registers_callback: New function.
(arm_insert_point): Likewise.
(arm_remove_point): Likewise.
(arm_stopped_by_watchpoint): Likewise.
(arm_stopped_data_address): Likewise.
(arm_new_process): Likewise.
(arm_new_thread): Likewise.
(arm_prepare_to_resume): Likewise.
(the_low_target): Register arm_insert_point, arm_remove_point,
arm_stopped_by_watchpoint, arm_stopped_data_address, arm_new_process,
arm_new_thread, and arm_prepare_to_resume.

2011-09-21  Tristan Gingold  <gingold@adacore.com>

* mach-o.c (bfd_mach_o_convert_section_name_to_bfd): Add comment.
Deals with size limited strings.
(bfd_mach_o_build_commands): Initialize more fields.

*** empty log message ***

daily update

* mi/mi-main.c (mi_load_progress): Restore saved_uiout value to
current_uiout, not uiout.

fix pr reference

PR ld/12301
* ldlang.h (lang_input_statement_type): Add "reload" bitfield.
Clarify comments.
* ldlang.c (new_afile): Init new field.
(load_symbols): Don't call ldlang_add_file when reloading.
(open_input_bfds): Reload as-needed libs during plugin rescan.

daily update

*** empty log message ***

2011-09-19  Sriraman Tallam  <tmsriram@google.com>

* plugin.h (should_defer_layout): Modify to check for any_claimed_.

* incremental.cc (can_incremental_update): Fix typo in comment.
* incremental.h (can_incremental_update): Likewise.

* python/py-auto-load.c (source_section_scripts): Fix file
descriptor leak.
* python/python.c (source_python_script_for_objfile): Tweak comments.

* gdb.trace/tspeed.exp: New file.
* gdb.trace/tspeed.c: New file.

[OBV] fix typos in ChangeLog

2011-09-19  Tristan Gingold  <gingold@adacore.com>

* config/tc-alpha.c (insert_operand): Call as_bad_value_out_of_range
instead of as_warn_out_of_range.

*** empty log message ***

daily update

Check zero address size.

2011-09-18  H.J. Lu  <hongjiu.lu@intel.com>

PR binutils/13196
* dwarf.c (display_debug_aranges): Check zero address size.

* incremental.cc (can_incremental_update): New function.
* incremental.h (can_incremental_update): New function.
* layout.cc (Layout::init_fixed_output_section): Call it.
(Layout::make_output_section): Don't allow patch space in .eh_frame.
* object.cc (Sized_relobj_file::do_layout): Call
can_incremental_update.

Support displaced stepping for Thumb 16-bit insns.
gdb/
* arm-tdep.c (THUMB_NOP) Define.
(thumb_copy_unmodified_16bit): New.
(thumb_copy_b, thumb_copy_bx_blx_reg): New.
(thumb_copy_alu_reg): New.
(arm_copy_svc): Move some common code to ...
(install_svc): ... here.  New.
(thumb_copy_svc): New.
(install_pc_relative): New.
(thumb_copy_pc_relative_16bit): New.
(thumb_decode_pc_relative_16bit): New.
(thumb_copy_16bit_ldr_literal): New.
(thumb_copy_cbnz_cbz): New.
(cleanup_pop_pc_16bit_all): New.
(thumb_copy_pop_pc_16bit): New.
(thumb_process_displaced_16bit_insn): New.
(thumb_process_displaced_32bit_insn): New.
(thumb_process_displaced_insn): process thumb instruction.

Support displaced stepping for Thumb 32-bit insns.
gdb/
* arm-tdep.c (thumb_copy_unmodified_32bit): New.
(thumb2_copy_preload): New.
(thumb2_copy_copro_load_store): New.
(thumb2_copy_b_bl_blx): New.
(thumb2_copy_alu_imm): New.
(thumb2_copy_load_reg_imm): New.
(thumb2_copy_load_literal): New
(thumb2_copy_block_xfer): New.
(thumb_32bit_copy_undef): New.
(thumb_32bit_copy_unpred): New.
(thumb2_decode_ext_reg_ld_st): New.
(thumb2_decode_svc_copro): New.
(decode_thumb_32bit_store_single_data_item): New.
(thumb_copy_pc_relative_32bit): New.
(thumb_decode_pc_relative_32bit): New.
(decode_thumb_32bit_ld_mem_hints): New.
(thumb2_copy_table_branch): New
(thumb_process_displaced_32bit_insn): Process Thumb 32-bit
instructions.

gdb/testsuite/
* gdb.arch/arm-disp-step.S (test_ldr_literal): Test for Thumb
instructions.
(test_adr_32bit, test_pop_pc): Likewise.
(test_ldr_literal_16, test_cbz_cbnz, test_adr): New test for
Thumb instructions.
* gdb.arch/arm-disp-step.exp (test_ldm_stm_pc): Match $gdb_prompt
in gdb_test_multiple.
(test_ldr_literal_16, test_cbz_cbnz, test_adr): New.