Revert use of memory_args_by_pointer in i386 gdbarch_tdep.

This is no longer useful, as it was introduced to reuse the funcall
handling code in amd64-tdep.c in the context of x64-windows. But
we have since then changed the implementations to be completely
independent of each other.

This reverts the non-windows-specific part of the change called:
    amd64-windows: memory args passed by pointer during function calls.
(the x64-windows portion has already been reverted)

gdb/ChangeLog:

        Revert:
        * i386-tdep.h (gdbarch_tdep): Add field memory_args_by_pointer.
        * amd64-tdep.c (amd64_push_arguments): Add handling of architectures
        where tdep->memory_args_by_pointer is non-zero.

Revert use of integer_param_regs_saved_in_caller_frame in i386 gdbarch_tdep.

This is no longer useful, as it was introduced to reuse the funcall
handling code in amd64-tdep.c in the context of x64-windows. But
we have since then changed the implementations to be completely
independent of each other.

This reverts the non-windows-specific part of the change called:
    amd64-windows: 32 bytes allocated on stack by caller for integer
    parameter regs
(the x64-windows portion has already been reverted)

gdb/ChangeLog:

        Revert:
        * i386-tdep.h (struct gdbarch_tdep): Add new field
        integer_param_regs_saved_in_caller_frame.
        * amd64-tdep.c (amd64_push_dummy_call): Allocate some memory on
        stack if tdep->integer_param_regs_saved_in_caller_frame is set.

Reimplement function calls on amd64-windows

This patch provides a standalone implementation of function calls
on amd64-windows, instead of providing some bits and pieces hooking
into the function call implementation meant for sysV (in amd64-tdep).
It makes better sense to do it this way, because the two ABIs are
actually very different; for instance, the concept of argument
classification, which is so central in the sysV ABI and drove the
the implementation in amd64-tdep, makes no sense for Windows. It
is therefore better for the Windows implementation to be completely
separate, rather than rely on adaptations of the sysV implementation.

gdb/ChangeLog:

        * amd64-tdep.c: #include "value.h"
        (amd64_windows_classify): Delete.
        (amd64_windows_passed_by_integer_register)
        (amd64_windows_passed_by_xmm_register)
        (amd64_windows_passed_by_pointer)
        (amd64_windows_adjust_args_passed_by_pointer)
        (amd64_windows_store_arg_in_reg, amd64_windows_push_arguments)
        (amd64_windows_push_dummy_call): New functions.
        (amd64_windows_init_abi): Remove setting of
        tdep->call_dummy_num_integer_regs, tdep->call_dummy_integer_regs,
        tdep->classify, tdep->memory_args_by_pointer and
        tdep->integer_param_regs_saved_in_caller_frame.
        Add call to set_gdbarch_push_dummy_call.

* readline.c (bind_arrow_keys_internal):
Handle VK_HOME, VK_END, VK_DELETE and VK_INSERT for mingw
hosts.

Fix last entry in ChangeLog.gdb (unnecessary dir name).

Support .dwp with the name of symlinked binary file

gdb/
2013-09-24  Jan Kratochvil  <jan.kratochvil@redhat.com>

* dwarf2read.c (open_and_init_dwp_file): Try open_dwp_file also with
objfile->original_name.

gdb/testsuite/
2013-09-24  Jan Kratochvil  <jan.kratochvil@redhat.com>

* gdb.dwarf2/dwp-symlink.c: New file.
* gdb.dwarf2/dwp-symlink.exp: New file.

Keep objfile original filename

gdb/
2013-09-24  Jan Kratochvil  <jan.kratochvil@redhat.com>

Pass down original filename for objfile.
* coffread.c (coff_symfile_read): Update symbol_file_add_separate call.
* elfread.c (elf_symfile_read): Likewise.
* jit.c (jit_object_close_impl): Update allocate_objfile call, no
longer set ORIGINAL_NAME.
(jit_bfd_try_read_symtab): Update symbol_file_add_from_bfd call.
* jv-lang.c (get_dynamics_objfile): Update allocate_objfile call.
* machoread.c (macho_add_oso_symfile): Add parameter name.  Update
symbol_file_add_from_bfd call.
(macho_symfile_read_all_oso): Update two macho_add_oso_symfile calls.
(macho_check_dsym): Add parameter filenamep.  Change function comment.
Set *filenamep.
(macho_symfile_read): New variable dsym_filename.  Update
macho_check_dsym call.  Use it for symbol_file_add_separate.
* objfiles.c (allocate_objfile): Add parameter name.  New comment for
it.  Use it for objfile->original_name.
(objfile_name): Return OBFD's filename, if available.
* objfiles.h (allocate_objfile): Add new parameter name.
* solib.c (solib_read_symbols): Update symbol_file_add_from_bfd call.
* symfile-mem.c (symbol_file_add_from_memory): Update
symbol_file_add_from_bfd call.
* symfile.c (read_symbols): Update symbol_file_add_separate call, new
comment for it.
(symbol_file_add_with_addrs): New parameter name, add function comment
for it.  Remove variable name.  Update allocate_objfile call.
(symbol_file_add_separate): New parameter name, add function comment
for it.  Update symbol_file_add_with_addrs call.
(symbol_file_add_from_bfd): New parameter name.  Update
symbol_file_add_with_addrs call.
(symbol_file_add): Update symbol_file_add_from_bfd call.
(reread_symbols): New variable original_name.  Save
objfile->original_name by it.
* symfile.h (symbol_file_add_from_bfd, symbol_file_add_separate): Add
second parameter.

Code cleanup: Add objfile_name accessor

gdb/
2013-09-24  Jan Kratochvil  <jan.kratochvil@redhat.com>

Code cleanup: Add objfile_name accessor function.
* ada-lang.c (is_known_support_routine): Use objfile_name.
* auto-load.c (source_gdb_script_for_objfile)
(auto_load_objfile_script): Likewise.
* coffread.c (coff_symtab_read, read_one_sym): Likewise.
* dbxread.c (dbx_symfile_read): Likewise.
* dwarf2-frame.c (dwarf2_build_frame_info): Likewise.
* dwarf2loc.c (locexpr_describe_location_piece): Likewise.
* dwarf2read.c (dwarf2_get_dwz_file, dwarf2_read_index)
(dw2_symtab_iter_next, dw2_expand_symtabs_matching)
(lookup_dwp_signatured_type, lookup_dwo_unit)
(dwarf2_build_psymtabs_hard, scan_partial_symbols, process_queue)
(fixup_go_packaging, process_imported_unit_die, dwarf2_physname)
(read_import_statement, create_dwo_cu, open_and_init_dwp_file)
(lookup_dwo_cutu, read_call_site_scope, dwarf2_ranges_read)
(dwarf2_record_block_ranges, read_common_block, read_typedef)
(read_subrange_type, load_partial_dies, read_partial_die)
(read_addr_index_1, read_str_index, dwarf_decode_lines_1)
(die_containing_type, build_error_marker_type, lookup_die_type)
(follow_die_ref_or_sig, follow_die_ref, dwarf2_fetch_die_loc_sect_off)
(dwarf2_fetch_constant_bytes, follow_die_sig, get_signatured_type)
(get_DW_AT_signature_type, write_psymtabs_to_index)
(save_gdb_index_command): Likewise.
* elfread.c (find_separate_debug_file_by_buildid, elf_symfile_read):
Likewise.
* expprint.c (dump_subexp_body_standard): Likewise.
* gdbtypes.c (type_name_no_tag_or_error): Likewise.
* jit.c (jit_object_close_impl): Use the objfile field name renamed to
original_name.
* linux-thread-db.c (try_thread_db_load_from_pdir_1): New variable
obj_name, use objfile_name for it, use the variable.
(try_thread_db_load_from_pdir, has_libpthread, thread_db_new_objfile):
Use objfile_name.
* machoread.c (macho_symtab_read, macho_check_dsym)
(macho_symfile_relocate): Likewise.
* maint.c (maintenance_translate_address): Likewise.
* minidebug.c (find_separate_debug_file_in_section): Likewise.
* minsyms.c (install_minimal_symbols): Likewise.
* objfiles.c (allocate_objfile): Use the objfile field name renamed to
original_name.
(filter_overlapping_sections): Use objfile_name.
(objfile_name): New function.
* objfiles.h (struct objfile): Rename field name to original_name.
(objfile_name): New prototype.
* printcmd.c (sym_info, address_info): Use objfile_name.
* probe.c (parse_probes, collect_probes, compare_probes)
(info_probes_for_ops): Likewise.
* progspace.c (clone_program_space): Likewise.
* psymtab.c (require_partial_symbols, dump_psymtab, allocate_psymtab)
(maintenance_info_psymtabs): Likewise.
* python/py-auto-load.c (gdbpy_load_auto_script_for_objfile)
(source_section_scripts): Likewise.
* python/py-objfile.c (objfpy_get_filename): Likewise.
* python/py-progspace.c (pspy_get_filename): Likewise.
* solib-aix.c (solib_aix_get_toc_value): Likewise.
* solib-som.c (match_main, som_solib_section_offsets): Likewise.
* solib.c (solib_read_symbols): Likewise.
* stabsread.c (scan_file_globals): Likewise.
* stap-probe.c (handle_stap_probe): Likewise.
* symfile.c (symbol_file_clear, separate_debug_file_exists)
(find_separate_debug_file_by_debuglink): Likewise.
(reread_symbols): Likewise.  Use the objfile field name renamed to
original_name.
(allocate_symtab): Use objfile_name.
* symmisc.c (print_symbol_bcache_statistics, print_objfile_statistics)
(dump_objfile, dump_msymbols, dump_symtab_1)
(maintenance_print_msymbols, maintenance_print_objfiles)
(maintenance_info_symtabs, maintenance_check_symtabs): Likewise.
* target.c (target_translate_tls_address, target_info): Likewise.
* xcoffread.c (xcoff_initial_scan): Make variable name const.  Use
objfile_name.

Code cleanup: rename variable

gdb/
2013-09-24  Jan Kratochvil  <jan.kratochvil@redhat.com>

Code cleanup.
* probe.c (parse_probes): Rename variable objfile_name to
objfile_namestr.

Remove solib-sunos.c

gdb/
2013-09-24  Jan Kratochvil  <jan.kratochvil@redhat.com>

Remove solib-sunos.c.
* Makefile.in (ALLDEPFILES): Remove solib-sunos.c.
* config/m68k/obsd.mh (NATDEPFILES): Remove solib-sunos.o.
* objfiles.c (rt_common_objfile): Remove.
(free_objfile): Remove rt_common_objfile comparison.
* objfiles.h (rt_common_objfile): Remove.
* solib-sunos.c: Remove.
* symfile.c (reread_symbols): Remove solib-sunos.c comment.

Remove a.out NetBSD and OpenBSD hosts

gdb/
2013-09-21  Jan Kratochvil  <jan.kratochvil@redhat.com>

Remove a.out NetBSD and OpenBSD hosts.
* NEWS (Removed native configurations): New.
* config/arm/nbsdaout.mh: Remove.
* config/i386/nbsdaout.mh: Remove.
* config/i386/obsdaout.mh: Remove.
* config/m68k/nbsdaout.mh: Remove.
* config/sparc/nbsdaout.mh: Remove.
* config/vax/nbsdaout.mh: Remove.
* configure.host (arm*-*-netbsd*, i[34567]86-*-netbsd*)
(i[34567]86-*-openbsd[0-2].*, i[34567]86-*-openbsd3.[0-3])
(m68*-*-netbsd*, sparc-*-netbsd*, vax-*-netbsd*): Add them to obsolete
error.
(arm*-*-netbsd*, i[34567]86-*-netbsd*, i[34567]86-*-openbsd[0-2].*)
(i[34567]86-*-openbsd3.[0-3], m68*-*-netbsd*, sparc-*-netbsdaout*)
(sparc-*-netbsd*, vax-*-netbsd*): Remove.

* elf32-mips.c (ELF_OSABI): Undef for vxworks.

* syms.c (_bfd_stab_section_find_nearest_line): Ignore partial
stabs at end of .stab.  Tidy variable usage.  Don't drop the need
for a NULL function name stab if If N_FUN stab is ignored.
Ensure index entry count loop matches write loop.

* emulparams/elf32ppccommon.sh (LIBPATH_SUFFIX): Provide 32-bit
and cross-endian values.
* emulparams/elf64ppc.sh: Source elf32ppccomon.sh.  Delete duplicate,
and unset unwanted variables.

* genscripts.sh (LIB_PATH): Don't exclude libdir or tooldir when
sysrooted.  Also, don't always add tooldir when non-sysrooted.
Instead add both when native and tooldir also when TOOL_DIR is
defined.  Always prepend '=' to paths when sysrooted.  Always
put paths with LIBPATH_SUFFIX first in search order.

daily update

revert erroneous checkin to mi-support.exp

(clh 9)

In regcache use print_hex_chars.

  https://sourceware.org/ml/gdb-patches/2013-09/msg00815.html

gdb/ChangeLog

        * regcache.c: Add include of valprint.h.
        (dump_endian_bytes): Delete.
        (regcache_dump): Use print_hex_chars not dump_endian_bytes.

[MinGW] Fix a compilation error in readline/util.c:_rl_strnicmp

readline/ChangeLog.gdb:

* readline/util.c (_rl_strnicmp): Add missing semicolon.

Reuse print_hex_chars.

  https://sourceware.org/ml/gdb-patches/2013-09/msg00768.html

gdb/ChangeLog

        * sh64-tdep.c (sh64_do_fp_register): Use print_hex_chars.

Regenerate sim configure files to pick up support for powerpc64le in
libtool.m4.

daily update

daily update

2013-09-21  Jan-Benedict Glaw  <jbglaw@lug-owl.de>

* configure.ac: Update from GCC.
* configure: Regenerate.

daily update

* lib/dwarf.exp (build_executable_from_fission_assembler): New proc.
* gdb.dwarf2/fission-base.S: Update.  Split out .dwo into separate
file.
* gdb.dwarf2/fission-loclists.S: Ditto.
* gdb.dwarf2/fission-reread.S: Ditto.
* gdb.dwarf2/fission-base.exp: Skip of remote host.  Compile with
build_executable_from_fission_assembler.
* gdb.dwarf2/fission-loclists.exp: Ditto.
* gdb.dwarf2/fission-reread.exp: Ditto.

* lib/future.exp (gdb_find_objcopy, gdb_find_readelf): New procs.
* lib/gdb.exp (build_id_debug_filename_get): Update to use them.
(gdb_gnu_strip_debug): Ditto.
* lib/prelink-support.exp (section_get, prelink_no): Ditto.
* gdb.arch/altivec-abi.exp: Ditto.
* gdb.base/attach-pie-misread.exp: Ditto.
* gdb.base/comprdebug.exp: Ditto.
* gdb.base/dup-sect.exp: Ditto.
* gdb.base/gnu-debugdata.exp: Ditto.
* gdb.base/step-symless.exp: Ditto.
* gdb.dwarf2/dw2-inline-param.exp: Ditto.
* gdb.dwarf2/dw2-skip-prologue.exp: Ditto.
* gdb.dwarf2/gdb-index.exp: Ditto.

2013-09-20  Chung-Lin Tang  <cltang@codesourcery.com>

* Makefile.am (enios2elf.c): Change tdir_nios2 to tdir_nios2elf.
* Makefile.in: Regenerate.

* libtool.m4 (_LT_ENABLE_LOCK <ld -m flags>): Remove non-canonical
ppc host match.  Support little-endian powerpc linux hosts.
Regenerate binutils configure files.

daily update

Fix regressions caused by thread-specific breakpoint deletion.

The recent change to make GDB auto-delete thread-specific breakpoints
when the corresponding thread is deleted
(https://sourceware.org/ml/gdb-patches/2013-09/msg00038.html) caused
gdb.base/nextoverexit.exp to regress.

    Breakpoint 1, main () at .../gdb/testsuite/gdb.base/nextoverexit.c:21
    21        exit (0);
    (gdb) next
    [Inferior 1 (process 25208) exited normally]
    Thread-specific breakpoint -5 deleted - thread 1 is gone.
    Thread-specific breakpoint -6 deleted - thread 1 is gone.
    Thread-specific breakpoint -7 deleted - thread 1 is gone.
    Thread-specific breakpoint 0 deleted - thread 1 is gone.
    (gdb) FAIL: gdb.base/nextoverexit.exp: next over exit (the program exited)

We shouldn't be seeing this for internal or momentary breakpoints.  In
fact, we shouldn't even be trying to delete them, as whatever created
them will take care or it, and therefore it's dangerous to delete them
behind the creator's back.

I thought it'd still be good to tag thread-specific internal/momentary
breakpoints such that we'll no longer try to keep them insert in the
target, as they'll cause stops and thread hops in other threads, so I
tried disabling them instead.  That caused a problem when following a
child fork, and detaching from the parent, as we try to reset the
step-resume etc. breakpoints to the new child's thread
(breakpoint_re_set_thread), after the parent thread is already gone
(and the breakpoints are marked disabled).  I fixed that by
re-enabling internal/momentary breakpoints there, but, that didn't
feel super safe either (maybe we'd need a new flag in struct
breakpoint instead, to tag the thread-specific breakpoint as "not to
be inserted").  It felt like I was heading down a design rat hole,
and, other things will usually delete internal/momentary breakpoints
soon enough, so I left that little optimization for some other day.

So, internal/momentary breakpoints are no longer deleted/disabled at
all, and we end up with a one-liner fix.

Tested on x86_64 Fedora 17.

gdb/
2013-09-19  Pedro Alves  <palves@redhat.com>

* breakpoint.c (remove_threaded_breakpoints): Skip non-user
breakpoints.

gnu-nat.c: Don't install a deprecated_xfer_memory method.

This removes another instance of a deprecated_xfer_memory user.

gdb/
2013-09-19  Pedro Alves  <palves@redhat.com>
    Thomas Schwinge  <thomas@codesourcery.com>
    Yue Lu  <hacklu.newborn@gmail.com>

* gnu-nat.c (gnu_read_inferior, gnu_write_inferior): Make static.
Take a gdb_byte pointer instead of a char pointer.

* gnu-nat.c (gnu_xfer_memory): Adjust interface as
gnu_xfer_partial helper.
(gnu_xfer_partial): New function.
(gnu_target): Don't install a deprecated_xfer_memory hook.
Install a to_xfer_partial hook.

Mostly code cleanup: Constification.

gdb/
2013-09-19  Jan Kratochvil  <jan.kratochvil@redhat.com>

Constification.
* main.c (captured_main): Replace catch_command_errors by
catch_command_errors_const.  Twice.
* symfile.c (symbol_file_add_main_1): Make args parameter const.
(symbol_file_add): Make name parameter const.
(symbol_file_add_main, symbol_file_add_main_1): Make args parameter const.
(symfile_bfd_open): Make name parameter const, rename it to cname.  Add
variable name.  Change their usage accordingly.
* symfile.h (symbol_file_add, symfile_bfd_open): Make first parameter
const.
(symbol_file_add_main): Make args parameter const.

daily update

2013-09-18  Raunaq Bathija  <raunaq12@in.ibm.com>
    Ulrich Weigand  <uweigand@de.ibm.com>

* xcoffread.c (struct coff_symbol): Use CORE_ADDR as type
of c_value member.
(read_xcoff_symtab): Use CORE_ADDR as type of fcn_start_addr.

Also mention Yue Lu in previous commit's entry.

2013-09-18  Pedro Alves  <palves@redhat.com>
    Yue Lu  <hacklu.newborn@gmail.com>

* gnu-nat.c (inf_validate_procs, gnu_wait, gnu_resume)
(gnu_create_inferior)
(gnu_attach, gnu_thread_alive, gnu_pid_to_str, cur_thread)
(set_sig_thread_cmd): Use the lwpid field of ptids to
store/extract thread ids instead of the tid field.
* i386gnu-nat.c (gnu_fetch_registers): Adjust.

Subject: [PATCH] [Hurd/gnu-nat.c] Use ptid_t.lwpid to store thread ids
 instead of ptid_t.tid.

In preparation for reusing gnu-nat.c in gdbserver, switch to storing
thread ids in the lwpid field of ptid_t rather than in the tid
field.  The Hurd's thread model is 1:1, so it doesn't feel wrong
anyway.

gdb/
2013-09-18  Pedro Alves  <palves@redhat.com>

* gnu-nat.c (inf_validate_procs, gnu_wait, gnu_resume)
(gnu_create_inferior)
(gnu_attach, gnu_thread_alive, gnu_pid_to_str, cur_thread)
(set_sig_thread_cmd): Use the lwpid field of ptids to
store/extract thread ids instead of the tid field.
* i386gnu-nat.c (gnu_fetch_registers): Adjust.

Consistent display of "<optimized out>" for register values.

  https://sourceware.org/ml/gdb-patches/2013-08/msg00170.html

gdb/ChangeLog

        * infcmd.c (default_print_one_register_info): Add detection of
        optimized out values.
        (default_print_registers_info): Switch to using
        get_frame_register_value.

gdb/testsuite/ChangeLog

        * gdb.dwarf2/dw2-reg-undefined.exp: Change pattern for info
        register to "<optimized out>", and also print the registers.

2013-09-18  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>

* ld-aarch64/eh-frame.d: Update expected output to allow for
64-bit addresses.

gdbserver --wrapper also doesn't work on Cygwin.

Skip the test on Cygwin too.

2013-09-18  Pedro Alves  <palves@redhat.com>

PR server/15967
* gdb.server/wrapper.exp: Also return unsupported for Cygwin, and
change text.

When reverse-stepping, only insert a resume breakpoint at ecs->stop_func_start
if the function start is known.  Otherwise, keep single-stepping.

[Hurd/gnu-nat] Fix old "signal-thread" command regression.

By inspection, I noticed that when I made the gnu-nat use
ptid(pid,0,tid) to represent a thread, instead of using ptid(tid,0,0),
in <https://sourceware.org/ml/gdb-patches/2008-08/msg00175.html>, I
introduced a bug.

The change was:

   else
     {
-      int tid = PIDGET (thread_id_to_pid (atoi (args)));
+      int tid = ptid_get_tid (thread_id_to_pid (atoi (args)));
       if (tid < 0)
  error (_("Thread ID %s not known.  Use the \"info threads\" command to\n"
         "see the IDs of currently known threads."), args);

and thread_id_to_pid does:

 ptid_t
 thread_id_to_pid (int num)
 {
   struct thread_info *thread = find_thread_id (num);

   if (thread)
     return thread->ptid;
   else
     return pid_to_ptid (-1);
 }

(pid_to_ptid (-1) is the same as minus_one_ptid.)

So before, we were really looking at the pid, where thread_id_to_pid
stores the -1.

The right fix is to compare the whole ptid to minus_one_ptid, of
course.

Completely untested, but I think it's obvious enough, so I went ahead
and put it in.

gdb/
2013-09-18  Pedro Alves  <palves@redhat.com>

* gnu-nat.c (set_sig_thread_cmd): Compare the thread's ptid to
minus_one_ptid instead of looking at the ptid's tid field and
comparing that to -1.

Fix whitespace error in previous commit.

  https://sourceware.org/ml/gdb-cvs/2013-09/msg00108.html
  https://sourceware.org/ml/gdb-patches/2013-09/msg00225.html

gdb/ChangeLog

        * main.h (get_gdb_program_name): Remove extra whitespace.

Add new function to access gdb_program_name.

  https://sourceware.org/ml/gdb-patches/2013-09/msg00225.html

gdb/ChangeLog

        * main.h (get_gdb_program_name): Add declaration.
        * main.c (get_gdb_program_name): Add definition.

bfd/
2013-09-18  Tristan Gingold  <gingold@adacore.com>

* configure.in: Bump version to 2.24.51
* configure: Regenerate.

binutils/
2013-09-18  Tristan Gingold  <gingold@adacore.com>

* NEWS: Add marker for 2.24.

gas/
2013-09-18  Tristan Gingold  <gingold@adacore.com>

* NEWS: Add marker for 2.24.

ld/
2013-09-18  Tristan Gingold  <gingold@adacore.com>

* NEWS: Add marker for 2.24.

2013-09-18  Tristan Gingold  <gingold@adacore.com>

* Makefile.am (LEXLIB): Define.  Replase references to @LEXLIB@
by $(LEXLIB).
* Makefile.in: Regenerate.

* config/tc-msp430.c (OPTION_MOVE_DATA): Define.
(move_data): New variable.
(md_parse_option): Parse -md.
(msp430_section): New function.  Catch references to the .bss or
.data sections and generate a special symbol for use by the libcrt
library.
(md_pseudo_table): Intercept .section directives.
(md_longopt): Add -md
(md_show_usage): Likewise.
(msp430_operands): Generate a warning message if a NOP is inserted
into the instruction stream.
* doc/c-msp430.texi (node MSP430 Options): Document -md option.

gdb/gdbserver/

2013-09-18  Yao Qi  <yao@codesourcery.com>

PR server/15959
* server.c (start_inferior): Clear 'resume_info'.

gdb/testsuite/

2013-09-18  Yao Qi  <yao@codesourcery.com>

* gdb.server/wrapper.c: New.
* gdb.server/wrapper.exp: New.

gdb/gdbserver/

Fix ChangeLog format issue.

daily update

* dwarf2read.c: Move definitions of complaint functions to after
forward declarations of local functions.

2013-09-17  Doug Gilmore  <Doug.Gilmore@imgtec.com>

* readelf.c (get_machine_flags): Handle EF_MIPS_FP64.

2013-09-17  Doug Gilmore  <Doug.Gilmore@imgtec.com>

* config/tc-mips.c (mips_elf_final_processing): Set
EF_MIPS_FP64 for -mgp32 -mfp64, removing old FIXME.

2013-09-17  Doug Gilmore  <Doug.Gilmore@imgtec.com>

* elfxx-mips.c (_bfd_mips_elf_print_private_bfd_data): Handle
EF_MIPS_FP64.

2013-09-17  Doug Gilmore  <Doug.Gilmore@imgtec.com>

* mips.h (EF_MIPS_FP64): New e_flags bit.

PR gdb/11568 - delete thread-specific breakpoints on thread exit

PR gdb/11568 is about thread-specific breakpoints being left behind
when the corresponding thread exits.

Currently:

 (gdb) b start thread 2
 Breakpoint 3 at 0x400614: file thread-specific-bp.c, line 23.
 (gdb) b end
 Breakpoint 4 at 0x40061f: file thread-specific-bp.c, line 29.
 (gdb) c
 Continuing.
 [Thread 0x7ffff7fcb700 (LWP 14925) exited]
 [Switching to Thread 0x7ffff7fcc740 (LWP 14921)]

 Breakpoint 4, end () at thread-specific-bp.c:29
 29      }
 (gdb) info threads
   Id   Target Id         Frame
 * 1    Thread 0x7ffff7fcc740 (LWP 14921) "thread-specific" end () at thread-specific-bp.c:29
 (gdb) info breakpoints
 Num     Type           Disp Enb Address            What
 2       breakpoint     keep y   0x0000000000400614 in start at thread-specific-bp.c:23
         breakpoint already hit 1 time
 3       breakpoint     keep y   0x0000000000400614 in start at thread-specific-bp.c:23 thread 2
         stop only in thread 2
 4       breakpoint     keep y   0x000000000040061f in end at thread-specific-bp.c:29
         breakpoint already hit 1 time

Note that the thread-specific breakpoint 3 stayed around, even though
thread 2 is gone.

There's no way that breakpoint can trigger again (*), so the PR argues
that the breakpoint should just be removed, like local watchpoints.
I'm ambivalent on this -- it could be reasonable to disable the
breakpoint (kind of like breakpoint in shared library code when the
DSO is unloaded), so the user could still use it as visual template
for creating other breakpoints (copy/paste command lists, etc.), or we
could have a way to change to which thread a breakpoint applies.  But,
several people pushed this direction, and I don't plan on arguing...

(*) - actually, there is ...  thread numbers are reset on "run", so
the user could do "break foo thread 2", "run", and expect the
breakpoint to hit again on the second thread.  But given gdb's thread
numbering can't really be stable, that'd only work sufficiently well
for thread 1, so we'd better call it unsupported.

So with the patch, whenever a thread is deleted from GDB's list, GDB
goes through the thread-specific breakpoints and deletes corresponding
breakpoints.  Since this is user-visible, GDB prints out:

  Thread-specific breakpoint 3 deleted - thread 2 is gone.

And of course, we end up with:

 (gdb) info breakpoints
 Num     Type           Disp Enb Address            What
 2       breakpoint     keep y   0x0000000000400614 in start at thread-specific-bp.c:23
         breakpoint already hit 1 time
 4       breakpoint     keep y   0x000000000040061f in end at thread-specific-bp.c:29
         breakpoint already hit 1 time

2013-09-17  Muhammad Waqas <mwaqas@codesourcery.com>
    Pedro Alves  <palves@redhat.com>

PR gdb/11568
* breakpoint.c (remove_threaded_breakpoints): New function.
(_initialize_breakpoint): Attach remove_threaded_breakpoints
as thread_exit observer.

2013-09-17  Muhammad Waqas  <mwaqas@codesourccery.com>
    Jan Kratochvil  <jan.kartochvil@redhat.com>
    Pedro Alves  <palves@redhat.com>

PR gdb/11568
* gdb.thread/thread-specific-bp.c: New file.
* gdb.thread/thread-specific-bp.exp: New file.

This commit fixes a regression introduced by the new $_isvoid convenience
function.  It adds a check for $_isvoid during the test of "show convenience"
output.

gdb/testsuite/
2013-09-17  Sergio Durigan Junior  <sergiodj@redhat.com>

* gdb.base/defaults.exp (<show_conv_list>): Add check for $_isvoid
convenience function.

PR gdb/15911: "info threads" changes the default source and line (for "break", "list")

"info threads" changes the default source for "break" and "list", to
whatever the location of the first/bottom thread in the thread list
is...

 (gdb) b start
 (gdb) c
 ...
 (gdb) list
 *lists "start"*
 (gdb) b 23
 Breakpoint 3 at 0x400614: file test.c, line 23.
 (gdb) info threads
   Id   Target Id         Frame
 * 2    Thread 0x7ffff7fcb700 (LWP 1760) "test" start (arg=0x0) at test.c:23
   1    Thread 0x7ffff7fcc740 (LWP 1748) "test" 0x000000323dc08e60 in pthread_join (threadid=140737353922304, thread_return=0x0) at pthread_join.c:93
 (gdb) b 23
 Breakpoint 4 at 0x323dc08d90: file pthread_join.c, line 23.
                                    ^^^^^^^^^^^^^^^
 (gdb) list
 93          lll_wait_tid (pd->tid);
 94
 95
 96        /* Restore cancellation mode.  */
 97        CANCEL_RESET (oldtype);
 98
 99        /* Remove the handler.  */
 100       pthread_cleanup_pop (0);
 101
 102

The issue is that print_stack_frame always sets the current sal to the
frame's sal.  print_frame_info (which print_stack_frame calls to do
most of the work) also sets the last displayed sal, but only if
print_what isn't LOCATION.  Now the call in question, from within
thread.c:print_thread_info, does pass in LOCATION as print_what, but
print_stack_frame doesn't have the same check print_frame_info has.
We could consider adding it, but setting these globals depending on
print_what isn't very clean, IMO.  What we have is two logically
distinct operations mixed in the same function(s):

  #1 - print frame, in the format specified by {print_what,
    print_level and print_args}.

  #2 - We're displaying a frame to the user, and I want the default
    sal to point here, because the program stopped here, or the user
    did some context-changing command (up, down, etc.).

So I added a new parameter to print_stack_frame & friends for point
#2, and went through all calls in the tree adjusting as necessary.

Tested on x86_64 Fedora 17.

gdb/
2013-09-17  Pedro Alves  <palves@redhat.com>

PR gdb/15911
* ada-tasks.c (task_command_1): Adjust call to print_stack_frame.
* bsd-kvm.c (bsd_kvm_open, bsd_kvm_proc_cmd, bsd_kvm_pcb_cmd):
* corelow.c (core_open):
* frame.h (print_stack_frame, print_frame_info): New
'set_current_sal' parameter.
* infcmd.c (finish_command, kill_command): Adjust call to
print_stack_frame.
* inferior.c (inferior_command): Likewise.
* infrun.c (normal_stop): Likewise.
* linux-fork.c (linux_fork_context): Likewise.
* record-full.c (record_full_goto_entry, record_full_restore):
Likewise.
* remote-mips.c (common_open): Likewise.
* stack.c (print_stack_frame): New 'set_current_sal' parameter.
Use it.
(print_frame_info): New 'set_current_sal' parameter.  Set the last
displayed sal depending on the new paremeter instead of looking at
print_what.
(backtrace_command_1, select_and_print_frame, frame_command)
(current_frame_command, up_command, down_command): Adjust call to
print_stack_frame.
* thread.c (print_thread_info, restore_selected_frame)
(do_captured_thread_select): Adjust call to print_stack_frame.
* tracepoint.c (tfind_1): Likewise.
* mi/mi-cmd-stack.c (mi_cmd_stack_list_frames)
(mi_cmd_stack_info_frame): Likewise.
* mi/mi-interp.c (mi_on_normal_stop): Likewise.
* mi/mi-main.c (mi_cmd_exec_return, mi_cmd_trace_find): Likewise.

gdb/testsuite/
* gdb.threads/info-threads-cur-sal-2.c: New file.
* gdb.threads/info-threads-cur-sal.c: New file.
* gdb.threads/info-threads-cur-sal.exp: New file.

Fix ChangeLog entry missing a ':' char.

  gdbserver/ChangeLog

* linux-tile-low.c (tile_regsets): Modify the size field to 64-bit
for each register.

PR ld/15957
* deffilep.y (def_file_add_directive): Avoid readin past end of
buffer.

2013-09-16  Jiong Wang  <jiwang@tilera.com>

  gdbserver/ChangeLog

        * configure.srv <tilegx*-*-linux*>: Remove linux-osdata.o from and add
linux-tile-low.o to srv_tgtobj.

opcodes/
* s390-opc.txt (clih): Make the immediate unsigned.

gas/testsuite/
* gas/s390/zarch-z196.s, gas/s390/zarch-z196.d: Test CLIH with
4000000000.

gdb/testsuite/

* gdb.base/catch-load.c: Remove the include of "dlfcn.h".
 [__WIN32__]: Include "windows.h" and define macro dlopen
 and dlclose.
[!__WIN32__]: Include "dlfcn.h".
* gdb.base/catch-load.exp (one_catch_load_test): Match
directory separator.

daily update

* lib/gdb.exp (using_fission): New proc.
* gdb.base/info-macros.exp: Skip test if using Fission.

* gdb.base/break-interp.exp: Fix indentation.

The error message in the function was saying
"You should provide one parameter..." while it should be saying "... one
argument...".  Replaced.

2013-09-16  Sergio Durigan Junior  <sergiodj@redhat.com>

* value.c (isvoid_internal_fn): Replace "parameter" with
"argument".

* README: Update references to writing code for GDB.
* configure.ac (build_warnings): Remove obsolete comment.
* configure: Regenerate.
* gdbarch.sh: Remove references to gdbint.texinfo.
* gdbarch.h: Regenerate.
* gdbtypes.c (objfile_type): Remove comments referencing internals
manual and D10V.

[gdb/doc]
Remove the internals manual gdbint.texinfo.
* Makefile.in (INFO_DEPS): Remove gdbint.info.
(PDFFILES): Remove gdbint.pdf.
(HTMLFILES): Remove gdbint/index.html.
(HTMLFILES_INSTALL): Remove gdbint.
(GDBINT_DOC_FILES): Remove.
(dvi): Remove gdbint.dvi.
(ps): Remove gdbint.ps.
* gdbint.texinfo: Remove file.
* gdb.texinfo (Maintenance Commands): Remove reference to gdbint.

Based on the discussion at:

<https://sourceware.org/ml/gdb-patches/2013-09/msg00301.html>
<https://sourceware.org/ml/gdb-patches/2013-09/msg00383.html>

This patch adds a new convenience function called $_isvoid, whose
only purpose is to check whether an expression is void or not.
This became necessary because the new convenience variable
$_exitsignal (not yet approved) has a mutual exclusive behavior
with $_exitcode, i.e., when one is "defined" (i.e., non-void),
the other is cleared (i.e., becomes void).  Doug wanted a way to
identify which variable to use, and checking for voidness is the
obvious solution.

It is worth mentioning that my first attempt, after a conversation with
Doug, was to actually implement a new $_isdefined() convenience
function.  I would do that (for convenience variables) by calling
lookup_only_internalvar.  However, I found a few problems:

- Whenever I called $_isdefined ($variable), $variable became defined
  (with a void value), and $_isdefined always returned true.

- Then, I tried to implement $_isdefined ("variable"), and do the "$" +
  "variable" inside GDB, thus making it impossible for GDB to create the
  convenience variable.  However, it was hard to extract the string
  without having to mess with values and their idiossincrasies.
  Therefore, I decided to abandon this attempt (specially because I
  didn't want to spend too much time struggling with it).

Anyway, after talking to Doug again we decided that it would be easier
to implement $_isvoid, and this will probably help in cases like
<http://stackoverflow.com/questions/3744554/testing-if-a-gdb-convenience-variable-is-defined>.

I wrote a NEWS entry for it, and some new lines on the documentation.

gdb/
2013-09-16  Sergio Durigan Junior  <sergiodj@redhat.com>

* NEWS: Mention new convenience function $_isvoid.
* value.c (isvoid_internal_fn): New function.
(_initialize_values): Add new convenience function $_isvoid.

gdb/doc/
2013-09-16  Sergio Durigan Junior  <sergiodj@redhat.com>

* gdb.texinfo (Convenience Functions): Mention new convenience
function $_isvoid.

gdb/testsuite/
2013-09-16  Sergio Durigan Junior  <sergiodj@redhat.com>

* gdb.base/gdbvars.c (foo_void): New function.
(foo_int): Likewise.
* gdb.base/gdbvars.exp (test_convenience_functions): New
function.  Call it.

gdbserver, aarch64: Zero out regs in aarch64_linux_set_debug_regs.

Apply the same fix that was applied to aarch64-linux-nat.c.

2013-09-16  Will Newton  <will.newton@linaro.org>

* linux-aarch64-low.c (aarch64_linux_set_debug_regs): Zero
out regs.

* arm-linux-tdep.c: Add "elf/common.h" header.
Remove AT_HWCAP macro definintion as it is provided in
added include file.
* s390-tdep.c: Remove system header <elf.h>
Add "elf/common.h" header for AT_HWCAP definition.
(s390_core_read_description): Use correct CORE_ADDR
for hwcap local variable used as third parameter
of function target_auxv_search.

gas/config/tc-arm.c: Fix parsing of NEON load/store element sizes.

The existing code would accept VLD2.64 and similar undefined
instructions.

gas/ChangeLog:

2013-09-16  Will Newton  <will.newton@linaro.org>

* config/tc-arm.c (do_neon_ld_st_interleave): Add constraint
disallowing element size 64 with interleave other than 1.

gas/testsuite/ChangeLog:

2013-09-16  Will Newton  <will.newton@linaro.org>

* gas/arm/neon-ldst-es.d: Add VLD1.64 instructions.
* gas/arm/neon-ldst-es.s: Likewise.
* testsuite/gas/arm/neon-ldst-es-bad.d: New file.
* testsuite/gas/arm/neon-ldst-es-bad.l: Likewise.
* testsuite/gas/arm/neon-ldst-es-bad.s: Likewise.

daily update

daily update

* common/filestuff.c (gdb_fopen_cloexec): Do not try to use "e"
mode if operating system doesn't know O_CLOEXEC.

daily update

Code cleanup: Move variable.

gdb/
2013-09-13  Jan Kratochvil  <jan.kratochvil@redhat.com>

Code cleanup.
* symfile.c (reread_symbols): Move variable obfd_filename to a more
inner block.

gdb/ChangeLog:
2013-09-13  Andreas Arnez  <arnez@linux.vnet.ibm.com>

* NEWS: Mention TDB support.
* features/s390-tdb.xml: New file.
* features/s390-te-linux64.xml: New file.
* features/s390x-te-linux64.xml: New file.
* features/Makefile (WHICH): Add new tdescs above.
(s390-te-linux64-expedite): Set.
(s390x-te-linux64-expedite): Set.
* features/s390-te-linux64.c: New file (generated).
* features/s390x-te-linux64.c: New file (generated).
* regformats/s390-te-linux64.dat: New file (generated).
* regformats/s390x-te-linux64.dat: New file (generated).
* s390-tdep.h (HWCAP_S390_HIGH_GPRS): Define.
(HWCAP_S390_TE): Likewise.
(S390_TDB_DWORD0_REGNUM): Likewise.
(S390_TDB_DWORD0_REGNUM): Likewise.
(S390_TDB_ABORT_CODE_REGNUM): Likewise.
(S390_TDB_CONFLICT_TOKEN_REGNUM): Likewise.
(S390_TDB_ATIA_REGNUM): Likewise.
(S390_TDB_R0_REGNUM): Likewise.
(S390_TDB_R1_REGNUM): Likewise.
(S390_TDB_R2_REGNUM): Likewise.
(S390_TDB_R3_REGNUM): Likewise.
(S390_TDB_R4_REGNUM): Likewise.
(S390_TDB_R5_REGNUM): Likewise.
(S390_TDB_R6_REGNUM): Likewise.
(S390_TDB_R7_REGNUM): Likewise.
(S390_TDB_R8_REGNUM): Likewise.
(S390_TDB_R9_REGNUM): Likewise.
(S390_TDB_R10_REGNUM): Likewise.
(S390_TDB_R11_REGNUM): Likewise.
(S390_TDB_R12_REGNUM): Likewise.
(S390_TDB_R13_REGNUM): Likewise.
(S390_TDB_R14_REGNUM): Likewise.
(S390_TDB_R15_REGNUM): Likewise.
(S390_NUM_REGS): Increase.
(S390_IS_TDBREGSET_REGNUM): New macro.
(s390_regmap_tdb): Declare.
(s390_sizeof_tdbregset): Define.
(tdesc_s390_te_linux64): Declare.
(tdesc_s390x_te_linux64): Likewise.
* s390-tdep.c: Add includes for "auxv.h", <elf.h>,
"features/s390-te-linux64.c", and "features/s390x-te-linux64.c".
(s390_regmap_tdb): New regmap.
(s390_supply_tdb_regset): New function.
(s390_tdb_regset): New regset.
(s390_linux64v2_regset_sections): Add TDB regset to list.
(s390x_linux64v2_regset_sections): Likewise.
(s390_regset_from_core_section): Recognize TDB core note section.
(s390_core_read_description): If HWCAP indicates TE support,
select tdesc_s390_te_linux64 or tdesc_s390_s390x_te_linux64.
(s390_gdbarch_init): Handle TDB regset.
(_initialize_s390_tdep): Initialize new tdescs.
* s390-nat.c (HWCAP_S390_HIGH_GPRS): Remove define.
(have_regset_tdb): New variable.
(s390_native_supply): Support register invalidation.
(fetch_regset): Invalidate registers if ptrace yields ENODATA.
(check_regset): Treat ENODATA as "regset exists".
(s390_linux_fetch_inferior_registers): Add TDB.
(s390_read_description): Check for TDB existence and select
appropriate tdesc.
* gdbserver/Makefile.in (clean): Add removal of new makefile
targets.
(s390-te-linux64.c): New makefile target.
(s390x-te-linux64.c): Likewise.
* gdbserver/configure.srv (srv_regobj): Append new objects
s390-te-linux64.o and s390x-te-linux64.o.
(srv_xmlfiles): Append new files s390-te-linux64.xml,
s390x-te-linux64.xml, and s390-tdb.xml.
* gdbserver/linux-s390-low.c (init_registers_s390_te_linux64): New
declaration.
(tdesc_s390_te_linux64): Likewise.
(init_registers_s390x_te_linux64): Likewise.
(tdesc_s390x_te_linux64): Likewise.
(s390_check_regset): Treat ENODATA as "regset exists".
(s390_arch_setup): Add TDB regset support.
(initialize_low_arch): Initialize registers for new tdescs.

gdb/doc/ChangeLog:
2013-09-13  Andreas Arnez  <arnez@linux.vnet.ibm.com>

* gdb.texinfo (Decimal Floating Point format): Mention S/390.
(Standard Target Features): Add new node to menu.
(S/390 and System z Features): New node.

gdb/testsuite/ChangeLog:
2013-09-13  Andreas Arnez  <arnez@linux.vnet.ibm.com>

* gdb.arch/s390-tdbregs.c: New file.
* gdb.arch/s390-tdbregs.exp: New file.

2013-09-13  Andreas Arnez  <arnez@linux.vnet.ibm.com>

* s390-tdep.h (S390_IS_GREGSET_REGNUM): New macro.
(S390_IS_FPREGSET_REGNUM): New macro.
* s390-tdep.c (s390_dwarf_regmap): Make const.
(regnum_is_gpr_full): New function for replacing repeated code.
(s390_pseudo_register_name): Use it.
(s390_pseudo_register_type): Likewise.
(s390_pseudo_register_read): Likewise.
(s390_pseudo_register_write): Likewise.
(s390_unwind_pseudo_register): Likewise.
(s390_regmap_gregset): New format for regmap.
(s390x_regmap_gregset): Likewise.
(s390_regmap_fpregset): Likewise.
(s390_regmap_upper): Likewise.
(s390_regmap_last_break): Likewise.
(s390_regmap_system_call): Likewise.
(s390_supply_regset): Adjust to new regmap format.
(s390_collect_regset): Likewise.
* s390-nat.c (s390_native_supply): Adjust to new regmap format.
(s390_native_collect): Likewise.
(supply_gregset): Likewise.
(fill_gregset): Likewise.
(supply_fpregset): Likewise.
(fill_fpregset): Likewise.
(fetch_regset): Likewise.
(store_regset): Likewise.
(s390_linux_fetch_inferior_registers): Likewise.
(s390_linux_fetch_inferior_registers): Likewise.

doc,mi: example refers to -list-features instead of -list-target-features

Corrected mi documentation about -list-target-features, example now uses the
correct mi command.

2013-09-13  Sanimir Agovic  <sanimir.agovic@intel.com>

* gdb.texinfo (GDB/MI Miscellaneous Commands): Use
-list-target-features in the example.

daily update

* README: New file.

* gdb.python/py-events.py (new_objfile_handler): Remove accidentally
added code to print event.inferior.

2013-09-12  Chao-ying Fu  <Chao-ying.Fu@imgtec.com>

* gas/mips/bltzal.s: New file.
* gas/mips/bltzal.l: New file.
* gas/mips/mips.exp: Run the bltzal test.

2013-09-12  Chao-ying Fu  <Chao-ying.Fu@imgtec.com>

* config/tc-mips.c (match_insn): Set error when $31 is used for
bltzal* and bgezal*.

merge from gcc

* dwarf.c (dwarf_vmatoa): Rename to dwarf_vmatoa_1 and add a
precision parameter.
(dwarf_vmatoa): New wrapper for dwarf_vmatoa_1.
(print_dwarf_vma): Use dwarf_vmatoa_1.
(SAFE_BYTE_GET): Add check that VAL is big enough to contain
AMOUNT bytes.
(process_debug_info): Use an unsigned int for the offset size.
(process_debug_pubnames): Likewise.
(display_debug_aranges): Likewise.
(struct Frame_Chunk): Use dwarf_vma type for pc_begin and pc_range
fields.
(frame_display_row): Use print_dwarf_vma to display dwarf_vma
values.
(display_debug_frames): Likewise.

* binutils-all/x86-64/compressed-1a.d: Update expected output to
allow for 64-bit addresses.

* ld-elf/eh1.d: Update expected output to allow for
64-bit addresses.
* ld-elf/eh2.d: Likewise.
* ld-elf/eh3.d: Likewise.
* ld-elf/eh4.d: Likewise.
* ld-elf/eh5.d: Likewise.
* ld-elf/eh6.d: Likewise.
* ld-mips-elf/eh-frame1-n64.d: Likewise.
* ld-mips-elf/eh-frame2-n64.d: Likewise.
* ld-mips-elf/eh-frame3.d: Likewise.

* gas/cfi/cfi-alpha-1.d: Update expected output to allow for
64-bit addresses.
* gas/cfi/cfi-alpha-3.d: Likewise.
* gas/cfi/cfi-arm-1.d: Likewise.
* gas/cfi/cfi-common-1.d: Likewise.
* gas/cfi/cfi-common-2.d: Likewise.
* gas/cfi/cfi-common-3.d: Likewise.
* gas/cfi/cfi-common-4.d: Likewise.
* gas/cfi/cfi-common-5.d: Likewise.
* gas/cfi/cfi-common-6.d: Likewise.
* gas/cfi/cfi-common-7.d: Likewise.
* gas/cfi/cfi-hppa-1.d: Likewise.
* gas/cfi/cfi-i386-2.d: Likewise.
* gas/cfi/cfi-i386.d: Likewise.
* gas/cfi/cfi-m68k.d: Likewise.
* gas/cfi/cfi-mips-1.d: Likewise.
* gas/cfi/cfi-ppc-1.d: Likewise.
* gas/cfi/cfi-s390-1.d: Likewise.
* gas/cfi/cfi-s390x-1.d: Likewise.
* gas/cfi/cfi-sh-1.d: Likewise.
* gas/cfi/cfi-sparc-1.d: Likewise.
* gas/cfi/cfi-sparc64-1.d: Likewise.
* gas/cfi/cfi-x86_64.d: Likewise.

2013-09-12  Andrew Pinski  <apinski@cavium.com>

* aarch64-linux-nat.c (aarch64_linux_set_debug_regs): Zero out regs.

daily update

minidebuginfo: Fix stripping/debuginfo

gdb/doc/
2013-09-11  Jan Kratochvil  <jan.kratochvil@redhat.com>

* gdb.texinfo (MiniDebugInfo): Prepare file debug and use it to create
mini_debuginfo.  Strip binary before adding mini_debuginfo to it.

gdb/testsuite/
2013-09-11  Jan Kratochvil  <jan.kratochvil@redhat.com>

* gdb.base/gnu-debugdata.exp (objcopy 1): Move it lower and use only
debug part of the binary.

minidebuginfo: ppc64 sync testcase->doc

gdb/doc/
2013-09-11  Jan Kratochvil  <jan.kratochvil@redhat.com>

Sync documentation with gdb.base/gnu-debugdata.exp.
* gdb.texinfo (MiniDebugInfo): Add comment and "D" in the example.

minidebuginfo: Obvious changes

gdb/doc/
2013-09-11  Jan Kratochvil  <jan.kratochvil@redhat.com>

* gdb.texinfo (MiniDebugInfo): Fix two trailing dots.

gdb/testsuite/
2013-09-11  Jan Kratochvil  <jan.kratochvil@redhat.com>

* gdb.base/gnu-debugdata.exp (strip): Add -R .comment.
(addlink): Add comment.

daily update

merge from gcc

* dwarf.c (display_debug_frames): Check for DW64_CIE_ID when
parsing 64-bit frames.

merge from gcc

2013-09-10  Andreas Arnez  <arnez@linux.vnet.ibm.com>

* config/s390/s390.mh (NATDEPFILES): Add linux-waitpid.o.