* fhandler_disk_file.cc (fhandler_base::fstat_by_handle): Don't use
FileAllInformation info class since it needs a big buffer.  Add a
comment.

daily update

*** empty log message ***

This is the part of prev commit.

2009-07-11  Hui Zhu  <teawater@gmail.com>

* gdb.base/help.exp (disassemble): Update expected help text.

2009-07-11  Hui Zhu  <teawater@gmail.com>

* gdb.texinfo (disassemble): Add a new modifier /r
to "disassemble" command to make it print the raw instructions
in hex as well as in symbolic form.

2009-07-11  Hui Zhu  <teawater@gmail.com>

* cli/cli-cmds.c (disassemble_command): Add a new modifier /r
to "disassemble" command to print the raw instructions in hex as
well as in symbolic form.
(init_cli_cmds): Ditto.
(print_disassembly): Change "mixed" to "flags" to translate
the behavior of disassemble.
(disassemble_current_function): Ditto.
* mi/mi-cmd-disas.c (mi_cmd_disassemble): Ditto.
* stack.c (gdb_disassembly_stub): Ditto.
* disasm.c (do_mixed_source_and_assembly): Ditto.
(do_mixed_source_and_assembly): Ditto.
(do_assembly_only): Ditto.
(gdb_disassembly): Ditto.
(dump_insns): print the raw instructions in hex as well as in
symbolic form if DISASSEMBLY_RAW_INSN and flags is true.
* disasm.h (DISASSEMBLY_SOURCE): Include source code with the
assembly if it and flags is true.
(DISASSEMBLY_RAW_INSN): Include the raw instructions in hex with
the assembly if it and flags is true.
(gdb_disassembly): Update extern.
* NEWS: Document disassemble/r support.

* ldlang.c (insert_os_after): Don't tie assignments to non-alloc
output sections.

daily update

*** empty log message ***

* gdbint.texinfo (Testsuite): Document parallel make check.

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* ld-ifunc/ifunc.exp: Don't chck static non-ifunc-using
executable.

gdb/doc
* gdbint.texinfo (Testsuite): Document test transcripts.
gdb/testsuite
* lib/gdb.exp: Handle TRANSCRIPT.
(remote_spawn, remote_close, send_gdb): New procs.

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* bfdio.c (bfd_iovec): Add comments for bmmap.
* bfd-in2.h: Regenerated.

bfd/

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* Makefile.am: Run "make dep-am".
* Makefile.in: Regenerated.

binutils/

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* Makefile.am: Run "make dep-am".
* Makefile.in: Regenerated.

gas/

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* Makefile.am: Run "make dep-am".
* Makefile.in: Regenerated.

2009-07-10  Doug Kwan  <dougkwan@google.com>

* arm-disc.c (print_insn_coprocessor, print_insn_arm):  Print only
lower 32 bits of long types to make hexadecimal output consistent
on both 32-bit and 64-bit hosts.

2009-07-10  Doug Kwan  <dougkwan@google.com>

* elf32-arm.c (stub_hash_newfunc): Initialize fields
target_addend and orig_insn.

bfd
* Makefile.am (dwarf2.lo): Use dwarf2.h, not elf/dwarf2.h.
(elf-eh-frame.lo): Likewise.
(elf32-bfin.lo): Likewise.
(elf32-frv.lo): Likewise.
(elf32-xc16x.lo): Likewise.
* Makefile.in: Rebuild.
* dwarf2.c: Included dwarf.h, not elf/dwarf2.h.
* elf-eh-frame.c: Likewise.
* elf32-bfin.c: Likewise.
* elf32-frv.c: Likewise.
* elf32-xc16x.c: Likewise.
binutils
* Makefile.am (dwarf.o): Refer to dwarf2.h, not elf/dwarf2.h.
* Makefile.in: Rebuild.
* dwarf.c: Include dwarf2.h, not elf/dwarf2.h.
gas
* Makefile.am (DEPTC_alpha_ecoff): Refer to dwarf2.h, not
elf/dwarf2.h.
(DEPTC_alpha_elf): Likewise.
(DEPTC_alpha_evax): Likewise.
(DEPTC_arm_elf): Likewise.
(DEPTC_hppa_elf): Likewise.
(DEPTC_i386_aout): Likewise.
(DEPTC_i386_coff): Likewise.
(DEPTC_i386_elf): Likewise.
(DEPTC_m68k_aout): Likewise.
(DEPTC_m68k_coff): Likewise.
(DEPTC_m68k_elf): Likewise.
(DEPTC_mips_coff): Likewise.
(DEPTC_mips_ecoff): Likewise.
(DEPTC_mips_elf): Likewise.
(DEPTC_ppc_coff): Likewise.
(DEPTC_ppc_elf): Likewise.
(DEPTC_s390_elf): Likewise.
(DEPTC_sh_coff): Likewise.
(DEPTC_sh_elf): Likewise.
(DEPTC_sh64_elf): Likewise.
(DEPTC_sparc_aout): Likewise.
(DEPTC_sparc_coff): Likewise.
(DEPTC_sparc_elf): Likewise.
(as.o): Likewise.
(dwarf2dbg.o): Likewise.
(dw2gencfi.o): Likewise.
(ehopt.o): Likewise.
(read.o): Likewise.
* Makefile.in: Rebuild.
* dw2gencfi.h: Include dwarf2.h, not elf/dwarf2.h.
* dwarf2dbg.c: Likewise.
* ehopt.c: Likewise.
gdb
* dwarf2-frame.c: Include dwarf2.h, not elf/dwarf2.h.
* dwarf2expr.c: Likewise.
* dwarf2loc.c: Likewise.
* dwarf2read.c: Likewise.
* sh-tdep.c: Likewise.
* xtensa-tdep.c: Likewise.
include
* dwarf2.h: New file, moved from elf/.
include/elf

* dwarf2.h: Move to `..'.

* elf32-arm.c (arm_build_one_stub): Use the hash entry of the
symbol a stub entry is derived from, not the entry for the stub
itself, in call to elf32_arm_final_link_relocate.

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* dwarf2.h: Just include ../dwarf2.h.

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* ld-ifunc/ifunc-1-local-x86.d: Updated.
* ld-ifunc/ifunc-1-x86.d: Likewise.
* ld-ifunc/ifunc-3a-x86.d: Likewise.

Regenerate.

gdb/
Tidy up the code.
* mi/mi-cmd-var.c (mi_cmd_var_update): Remove variable nv.  Move
variables var, rootlist and cr to their local sub-block.  Move the
make_cleanup_ui_out_tuple_begin_end, make_cleanup_ui_out_list_begin_end
and do_cleanups calls to the common base block.  Remove the zero NV
check.

bfd/

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* elf.c (_bfd_elf_get_synthetic_symtab): Remove leading zeros
when reporting  addends.

ld/testsuite/

2009-07-10  H.J. Lu  <hongjiu.lu@intel.com>

* ld-ifunc/ifunc-1-local-x86.d: Updated.
* ld-ifunc/ifunc-1-x86.d: Likewise.Likewise.
* ld-ifunc/ifunc-2-local-x86-64.d: Likewise.Likewise.
* ld-ifunc/ifunc-2-x86-64.d: Likewise.
* ld-ifunc/ifunc-3a-x86.d: Likewise.Likewise.

* elf32-spu.c (spu_elf_size_stubs): Do set up soft-icache manager
data even when no stubs.
(spu_elf_place_overlay_data, spu_elf_build_stubs): Adjust.

* emultempl/ppc32elf.em (ppc_before_allocation): Turn on
linker relaxation if it might be necessary.

STT_GNU_IFUNC support for PowerPC.

2009-07-10 Phil Muldoon  <pmuldoon@redhat.com>

* python/python-internal.h (apply_varobj_pretty_printer): Update
definition.
(python_string_to_target_python_string): Add definition.
* python/python-utils.c (unicode_to_encoded_python_string)
(unicode_to_target_python_string)
(python_string_to_target_python_string): New Functions.
* python/python-prettyprint.c (pretty_print_one_value): Likewise.
(print_string_repr): Refactor to logic to account for PyObject
returned strings.
(apply_varobj_pretty_printer): Likewise.
* python/python-value.c (valpy_string): Parse length keyword. Use
length keyword in LA_GET_STRING.
* varobj.c (value_get_print_value): Refactor logic to account for
PyObject returned strings.
* c-lang.c (c_get_string): If the length parameter is specified,
use that. Return value in characters. Update comments.
* language.h: Update c_get_string prototype comments.

2009-07-10  Phil Muldoon  <pmuldoon@redhat.com>

* gdb.texinfo (Values From Inferior): Add length parameter
description.

2009-07-10 Phil Muldoon  <pmuldoon@redhat.com>

* gdb.python/python-prettyprint.c: Add counted null string
structure.
* gdb.python/python-prettyprint.exp: Print null string. Test for
embedded nulls.
* gdb.python/python-prettyprint.py (pp_ns): New Function.
* gdb.python/python-value.exp (test_value_in_inferior): Add
variable length string fetch tests.
* gdb.python/python-value.c (main): Add strings for string fetch tests.

* elf.c (_bfd_elf_get_synthetic_symtab): Report addends.

* elf-bfd.h (struct elf_link_hash_entry): Clarify ref_regular
and ref_regular_nonweak comment.
* elflink.c (elf_link_output_extsym): Special case ifunc syms
when def_regular, not ref_regular.
(elf_link_add_object_symbols): Don't set needs_plt on ifunc syms,..
* elf32-i386.c (elf_i386_check_relocs): ..set it here instead..
* elf64-x86-64.c (elf64_x86_64_check_relocs): ..and here.

* utils-sim.scm (-gen-decode-bits): New argument `entire-val'.
All callers updated.  Work around -ve shifts by referencing the
entire value.

* utils.scm (message): Handle pairs.

daily update

*** empty log message ***

* merge.cc (Object_merge_map::initialize_input_to_output_map):
Reserve space in the hash table.

* i386-tdep.c (i386_displaced_step_fixup): Fix order of arguments
to read_memory_unsigned_integer.

merge from gcc

gdb/
Report error on inaccessible memory.
* cli/cli-dump.c: Include gdbcore.h.
(dump_memory_to_file): Call read_memory instead of target_read_memory.

gdb/testsuite/
* gdb.base/dump.exp (inaccessible memory is reported): New test.

merge from gcc

2009-07-09  Craig Howland <howland@LGSInnovations.com>

        * libm/math/ef_scalb.c:  Replace isnanf() (pre-C99 function call) with
        isnan() (C99 macro).
        * libm/math/wf_log.c:  Ditto.
        * libm/math/wf_j0.c:  Ditto.
        * libm/math/wf_sqrt.c:  Ditto.
        * libm/math/wf_pow.c:  Ditto.
        * libm/math/wf_fmod.c:  Ditto.
        * libm/math/wf_remainder.c:  Ditto.
        * libm/math/wf_scalb.c:  Ditto.
        * libm/math/wf_atanh.c:  Ditto.
        * libm/math/wf_cosh.c:  Ditto.
        * libm/math/wf_acos.c:  Ditto.
        * libm/math/wf_acosh.c:  Ditto.
        * libm/math/wf_jn.c:  Ditto.
        * libm/math/wf_log10.c:  Ditto.
        * libm/math/wf_asin.c:  Ditto.
        * libm/math/wf_j1.c:  Ditto.
        * libm/common/sf_isnan.c:  Add #include <ieeefp.h>, fix comment.
        * libm/common/sf_isinf.c:  Add #include <ieeefp.h>, adjust comment to
        match that from s_isinf.c.
        * libc/include/machine/ieeefp.h:  Simplify isinf and isnan macros to
        remove un-necessary extension use (in a similar manner to as was
        recently done in math.h).
        * libc/include/math.h:  Remove isnanf and isinff prototypes (are in
        ieeefp.h).
        * libm/machine/spu/sf_isinf.c:  Fix comment (remove <math.h>).

2009-07-09  Hui Zhu  <teawater@gmail.com>

* solib.c (solib_bfd_open): Output a warning if solib's
architecture is not compatible with inferior's architecture.

gdb/
* target.h (struct target_ops <to_thread_architecture>): Describe it.
(target_thread_architecture): Refer to to_thread_architecture.

* elf32-spu.c (mark_functions_via_relocs): Init broken_cycle field
of malloc'd struct call_info.
(pasted_function): Likewise, priority too.

    * config/tc-arm.c (it_fsm_post_encode): New forward declaration.
        (set_it_insn_type_nonvoid): New macro.
        (emit_thumb32_expr): New function.
        (thumb_insn_size): New function.
        (emit_insn): New function.
        (s_arm_elf_inst): New function.
        (md_pseudo_table): New pseudo-opcode entries added.
        * doc/c-arm.texi: New directive added.

        * gas/arm/inst-po.d: New testcase.
        * gas/arm/inst-po.s: New file.
        * gas/arm/inst-po-2.d: New testcase.
        * gas/arm/inst-po-2.s: New file.
        * gas/arm/inst-po-2.l: New file.
        * gas/arm/inst-po-3.d: New testcase.
        * gas/arm/inst-po-3.s: New file.
        * gas/arm/inst-po-be.d: New testcase.

* cli/cli-cmds.c (init_cli_cmds): Add "inf" alias for "info" command.

*** empty log message ***

daily update

* pathnames.sgml: Fix typo.

* m32c/sbrk.c (sbrk): Change to take ptrdiff_t, not int.

* tui/tui-stack.c (tui_make_status_line): Display PC as ?? if unknown.

* Makefile.in (traps.o): Add dependency on targ-vals.h to fix
parallel makes.

* Makefile.in (stamp-arch): Pass archfile to cgen-arch.
(stamp-cpu): Pass archfile to cgen-cpu-decode.

* cpu/mep-ivc2.cpu (cpmovtocsar0_C3, cpmovtocsar1_C3,
cpmovtocc_C3, cpmovtocsar0_P0S_P1, cpmovtocsar1_P0S_P1,
cpmovtocc_P0S_P1): Mark volatile.  Note which registers are
written to.

* ld-selective/selective.exp: Remove check that $CC contains the
string "gcc".  Do -dumpversion for $CXX.

daily update

*** empty log message ***

* breakpoint.c (set_raw_breakpoint): Remove static modifier.

* winbase.h (ilockexch): Avoid making 'ret' volatile.
(ilockcmpexch): Likewise.

gdb
* c-lang.c (convert_octal): Only allow 3 octal digits.
(print_wchar): Prefer 3-digit octal form.  Fall back to hex if
needed.
* c-exp.y (c_parse_escape): Only allow 3 octal digits.
gdb/testsuite
* gdb.base/call-rt-st.exp: Update for change to escape output.
* gdb.base/callfuncs.exp: Likewise.
* gdb.base/charset.exp: Likewise.
* gdb.base/constvars.exp: Likewise.
* gdb.base/long_long.exp: Likewise.
* gdb.base/pointers.exp: Likewise.
* gdb.base/printcmds.exp: Likewise.
* gdb.base/setvar.exp: Likewise.
* gdb.base/store.exp: Likewise.
* gdb.cp/ref-types.exp: Likewise.
* gdb.mi/mi-var-child.exp: Likewise.
* gdb.mi/mi-var-display.exp: Likewise.
* gdb.mi/mi2-var-display.exp: Likewise.
* gdb.base/charset.exp: Test octal escape sequence length.
Update for change to escape output.

* config/tc-mep.c (md_begin): Don't reset the coprocessor options
when enabling other options without a specific configuration.

2009-07-07  Paul Pluzhnikov  <ppluzhnikov@google.com>

* python/python-value.c (valpy_getitem): Remove incorrect assert.

* ChangeLog: Assorted minor whitespace fixes in old entries.

winsup/ChangeLog:

* Makefile.common (COMPILE_CXX): Add support for per-file overrides
to exclude $(nostdinc) and $(nostdincxx) from compiler flags.
(COMPILE_CC): Likewise for $(nostdinc).

winsup/cygwin/ChangeLog:

* Makefile.in (DLL_OFILES): Add libstdcxx_wrapper.o
(libstdcxx_wrapper_CFLAGS): Add flags for new module.
(_cygwin_crt0_common_STDINCFLAGS): Define per-file override.
(libstdcxx_wrapper_STDINCFLAGS, cxx_STDINCFLAGS): Likewise.
* cxx.cc: Include "cygwin-cxx.h".
(operator new): Tweak prototype for full standards compliance.
(operator new[]): Likewise.
(operator new (nothrow)): New fallback function.
(operator new[] (nothrow), operator delete (nothrow),
operator delete[] (nothrow)): Likewise.
(default_cygwin_cxx_malloc): New struct of pointers to the above,
for final last-resort fallback default.
* cygwin-cxx.h: New file.
(struct per_process_cxx_malloc): Define.
(default_cygwin_cxx_malloc): Declare extern.
* cygwin.din (__wrap__ZdaPv): Export new wrapper.
(__wrap__ZdaPvRKSt9nothrow_t, __wrap__ZdlPv,
__wrap__ZdlPvRKSt9nothrow_t, __wrap__Znaj,
__wrap__ZnajRKSt9nothrow_t, __wrap__Znwj,
__wrap__ZnwjRKSt9nothrow_t): Likewise.
* globals.cc (__cygwin_user_data): Init newly-repurposed 'forkee'
field (now 'cxx_malloc') to point to default_cygwin_cxx_malloc.
* libstdcxx_wrapper.cc: New file.
(__wrap__ZdaPv, __wrap__ZdaPvRKSt9nothrow_t, __wrap__ZdlPv,
__wrap__ZdlPvRKSt9nothrow_t, __wrap__Znaj,
__wrap__ZnajRKSt9nothrow_t, __wrap__Znwj,
__wrap__ZnwjRKSt9nothrow_t): Define wrapper functions for libstdc++
malloc operators and their overrides.
* winsup.h (default_cygwin_cxx_malloc): Declare extern.
* include/cygwin/version.h (CYGWIN_VERSION_API_MINOR): Bump.
* include/sys/cygwin.h (struct per_process_cxx_malloc): Forward
declare here.
(struct per_process::forkee): Rename and repurpose from this ...
(struct per_process::cxx_malloc): ... to this.
* lib/_cygwin_crt0_common.cc: Include cygwin-cxx.h.
(WEAK): Define shorthand helper macro.
(__cygwin_cxx_malloc): Define and populate with weak references
to whatever libstdc++ malloc operators will be visible at final
link time for Cygwin apps and dlls.
(_cygwin_crt0_common): Always look up cygwin DLL's internal
per_process data, and don't test for (impossible) failure.  Inherit
any members of __cygwin_cxx_malloc that we don't have overrides
for from the DLL's default and store the resulting overall set of
overrides back into the DLL's global per_process data.

2009-07-07  Paul Pluzhnikov  <ppluzhnikov@google.com>

* python/python-value.c (valpy_getitem): Don't return from TRY_CATCH.

(-parse-insn-format-ifield-spec): Tweak error message.

* cpu/play.cpu (add): Use (ifield enum) for one format element.
* cpu/play.opc: New file.

Allow arbitrary enums in instruction formats, e.g. (f-op1 OP1_4).
* insn.scm (-parse-insn-format-ifield-spec): Recognize (ifield enum).
* doc/rtl.texi (Instructions): Update.

* cgen-opc.scm (-opc-file-path,opc-file-path): Move to opcodes.scm
(opc-arguments, -OPC): Call set-opc-file-path!.
* opcodes.scm (-opc-file-path,opc-file-path): Moved here.
(set-opc-file-path!): New function.
* dev.scm: Mention set-opc-file-path!.  Mention doc options.

* opc-ibld.scm (gen-insn-builder): Convert symbols to strings
before passing to gen-c-args.

2009-07-07  Sami Wagiaalla  <swagiaal@redhat.com>

* cp-support.h (struct using_direct): Rename members inner and
outer to import_src, and import_dest respectively.
* cp-namespace.c (cp_add_using_directive): Update.
(cp_add_using): Update.
(cp_copy_usings): Update.
(cp_lookup_symbol_namespace): Update.
(cp_scan_for_anonymous_namespaces): Update.
* cp-support.c (make_symbol_overload_list_using): Update.
* cp-support.h (cp_add_using_directive, cp_add_using): Rename
arguments from inner and outer to src and dest respectively.
* cp-namespace.c (cp_add_using_directive): Ditto.
(cp_add_using): Ditto.

gas/
    * config/tc-arm.c (insns): Fix encoding for torvsc.

gas/testsuite/
    * gas/arm/iwmmxt2.d: Fix insn pattern for torvsc,
    add patterns for waddsubhx.
    * gas/arm/iwmmxt2.s: Add tests for waddsubhx.

opcodes/
    * arm-dis.c (coprocessor_opcodes): Fix mask for waddbhus.

    PR 10288
        * arm-dis.c (arm_opcodes): Be more strict about decoding scaled
        addressing modes.

2009-07-07  H.J. Lu  <hongjiu.lu@intel.com>

* gas/i386/fma4.d: Append "#pass".
* gas/i386/x86-64-fma4.d: Likewise.

* generic/gdbtk-bp.c: Include "arch-utils.h".
(set_raw_breakpoint): Update external definition.
(gdb_set_bp): Adapt to set_raw_breakpoint change.
(gdb_set_bp_addr): Likewise.
* generic/gdbtk-cmds.c: Include "arch-utils.h.
(gdbtk_load_asm): Adapt to paddress and gdb_print_insn changes.
(hex2bin): Replace current_gdbarch by get_current_arch call.
(gdb_update_mem): Adapt to builtin_type changes.
(gdb_CA_to_TAS): Use paddress to replace removed function paddr_nz.
* generic/gdbtk-register.c: Include "arch-utils.h".
(get_register_size): Replace current_gdbarch by get_current_arch call.
(get_register_types): Likewise.
(get_register): Likewise.
(get_register_name): Likewise.
(map_arg_registers): Likewise.
(register_changed_p): Likewise.
(setup_architecture_data): Likewise.
(gdb_regformat): Likewise.
(gdb_reggrouplist): Likewise.
(gdb_reggroup): Likewise.
* generic/gdbtk-stack.c: Inlcude "arch-utils.h".
(gdb_get_blocks): Use paddress instead of removed paddr_nz function.
(gdb_selected_block): Likewise.
(gdb_selected_frame): Likewise.

gdb/
* mi/mi-cmd-var.c (mi_cmd_var_update): Replace a while loop by for loop.
* varobj.c (varobj_invalidate): Replace a while loop by for loop.

gdb/testsuite/
* gdb.mi/mi2-var-cmd.exp (floating varobj invalidation): New test.

Fix hang in floating varobjs.

* varob.c (varobj_invalidate): Advance to next
element when processing floating varobj.

* varobj.c: Remove unnecessary include.

2009-07-07  Tristan Gingold  <gingold@adacore.com>

* gas/i386/jump.d: Adjust to be relocated offset for jump to external
symbol.

* Make-common.in (CGEN_CPU_EXTR,CGEN_CPU_READ): Pass canonical
identifiers for file types rather than encoding internal
implementation details of cgen.sh.
(CGEN_CPU_WRITE,CGEN_CPU_SEM,CGEN_CPU_SEMSW): Ditto.
* cgen.sh: Add support for parallel makes.
Rewrite cpu/decode handling to avoid generating "extrafiles" twice.

* cygtls.cc (_cygtls::remove): Fix typo.

2009-07-07  Tristan Gingold  <gingold@adacore.com>

* config/i386/darwin.mh (NATDEPFILES): Remove excServer.o.
Remove excServer.c target.  Remove _all target.
* darwin.defs: Removed.

* elf32-spu.c (spu_elf_auto_overlay): Take into account section
alignment when packing sections into overlays.

[cgen]

* cpu/mep-core.cpu (fsft, ssarb): Mark as VOLATILE.
* cpu/mep-ivc2.cpu (many): Add VOLATILE to more insns that make
unspecified accesses to control registers.

[sid/component/cgen-cpu/mep]

* mep-cop1-16-decode.cxx: Regenerate.
* mep-cop1-16-decode.h: Regenerate.
* mep-cop1-16-defs.h: Regenerate.
* mep-cop1-16-model.cxx: Regenerate.
* mep-cop1-16-model.h: Regenerate.
* mep-cop1-16-sem.cxx: Regenerate.
* mep-cop1-32-decode.cxx: Regenerate.
* mep-cop1-32-decode.h: Regenerate.
* mep-cop1-32-defs.h: Regenerate.
* mep-cop1-32-model.cxx: Regenerate.
* mep-cop1-32-model.h: Regenerate.
* mep-cop1-32-sem.cxx: Regenerate.
* mep-cop1-48-decode.cxx: Regenerate.
* mep-cop1-48-decode.h: Regenerate.
* mep-cop1-48-defs.h: Regenerate.
* mep-cop1-48-model.cxx: Regenerate.
* mep-cop1-48-model.h: Regenerate.
* mep-cop1-48-sem.cxx: Regenerate.
* mep-cop1-64-decode.cxx: Regenerate.
* mep-cop1-64-decode.h: Regenerate.
* mep-cop1-64-defs.h: Regenerate.
* mep-cop1-64-model.cxx: Regenerate.
* mep-cop1-64-model.h: Regenerate.
* mep-cop1-64-sem.cxx: Regenerate.
* mep-core1-decode.cxx: Regenerate.
* mep-core1-decode.h: Regenerate.
* mep-core1-defs.h: Regenerate.
* mep-core1-model.cxx: Regenerate.
* mep-core1-model.h: Regenerate.
* mep-core1-sem.cxx: Regenerate.
* mep-cpu.h: Regenerate.
* mep-decode.cxx: Regenerate.
* mep-decode.h: Regenerate.
* mep-defs.h: Regenerate.
* mep-desc.h: Regenerate.
* mep-model.cxx: Regenerate.
* mep-model.h: Regenerate.
* mep-sem.cxx: Regenerate.

[opcodes]

* mep-desc.c: Regenerate.
* mep-desc.h: Regenerate.
* mep-opc.c: Regenerate.
* mep-opc.h: Regenerate.

*** empty log message ***

daily update

* Makefile.in (clean): Clean generated files in srcdir.
* tlsoffsets.h: Regenerate.

* fileread.cc (File_read::get_mtime): New method.
* fileread.h (Timespec): New structure.
(File_read::get_mtime): New method.
* incremental.cc (Incremental_inputs_entry_data::timestamp_usec):
Renamed from timestamp_nsec.
(Incremental_inputs_entry_write::timestamp_sec): Fix argument to
Elf_Xword.
(Incremental_inputs_entry_write::timestamp_usec): Renamed from
timestamp_nsec.
(Incremental_inputs::report_archive): Save mtime; style fix.
(Incremental_inputs::report_obejct): Save mtime; style fix.
(Incremental_inputs::report_script): Save mtime; style fix.
(Incremental_inputs::finalize_inputs): Style fix.
(Incremental_inputs::finalize): Style fix.
(Incremental_inputs::create_input_section_data): Store inputs
mtime.
* incremental.h (Incremental_inputs::report_script): Add mtime
argument.
(Incremental_inputs::Input_info::Input_info): Intialize only one
union member.
(Incremental_inputs::Input_info::archive): Move to nameless
union.
(Incremental_inputs::Input_info::obejct): Move to nameless union.
(Incremental_inputs::Input_info::script): Move to nameless union.
(Incremental_inputs::mtime): New field.
* script.cc (read_input_script): Pass file mtime to
Incremental_input.
* script.h (Script_info::inputs): Style fix.

* cygtls.cc (_cygtls::remove): Avoid closing a NULL handle.

* fhandler_socket.cc (fhandler_socket::recv_internal): Convert wsabuf
and wsacnt to references.  Fix handling of WSAEMSGSIZE.

<gas changes>
2009-07-06  Dwarakanath Rajagopal  <dwarak.rajagopal@amd.com>

* config/tc-i386.c (cpu_arch): Add .fma4 and CPU_FMA4_FLAGS.
(build_modrm_byte): Add support to handle FMA4 instructions.
(md_show_usage): Add fma4.

<gas/testsuite changes>
2009-07-06  Dwarakanath Rajagopal  <dwarak.rajagopal@amd.com>

* gas/i386/i386.exp: Add FMA4 tests.
* gas/i386/x86-64-fma4.d: Ditto.
* gas/i386/fma4.d: Ditto.
* gas/i386/x86-64-fma4.s: Ditto.
* gas/i386/fma4.s: Ditto.

<opcodes changes>
2009-07-06  Dwarakanath Rajagopal  <dwarak.rajagopal@amd.com>

* i386-opc.h (CpuFMA4): Add CpuFMA4.
(i386_cpu_flags): New.
* i386-gen.c: Add CPU_FMA4_FLAGS.
* i386-opc.tbl: Add FMA4 instructions.
* i386-tbl.h: Regenerate.
* i386-init.h: Regenerate.
* i386-dis.c (OP_VEX_FMA): New. Handle FMA4.
(OP_XMM_VexW): Ditto.
(OP_EX_VexW): Ditto.
(VEXI4_Fixup): Ditto.
(VexI4, VexFMA, Vex128FMA, EXVexW, EXdVexW, XMVexW): New Macros.
(PREFIX_VEX_3A5C, PREFIX_VEX_3A5D, PREFIX_VEX_3A5E): New.
(PREFIX_VEX_3A5F, PREFIX_VEX_3A60): New.
(PREFIX_VEX_3A68, PREFIX_VEX_3A69, PREFIX_VEX_3A6A): New.
(PREFIX_VEX_3A6B, PREFIX_VEX_3A6C, PREFIX_VEX_3A6D): New.
(PREFIX_VEX_3A6E, PREFIX_VEX_3A6F, PREFIX_VEX_3A7A): New.
(PREFIX_VEX_3A7B, PREFIX_VEX_3A7C, PREFIX_VEX_3A7D): New.
(PREFIX_VEX_3A7E, PREFIX_VEX_3A7F): New.
(VEX_LEN_3A6A_P_2,VEX_LEN_3A6B_P_2, VEX_LEN_3A6E_P_2): New.
(VEX_LEN_3A6F_P_2,VEX_LEN_3A7A_P_2, VEX_LEN_3A7B_P_2): New.
(VEX_LEN_3A7E_P_2,VEX_LEN_3A7F_P_2): New.
(get_vex_imm8): New. handle FMA4.
(OP_EX_VexReg): Ditto.

* lib/java.exp (java_init): Ignore $GCJ if it is empty.

2009-07-06  Joel Sherrill <joel.sherrill@oarcorp.com>

        * libc/include/sys/features.h: Enable UNIX98 mutex attributes
        for RTEMS.

* configure.tgt (i[34567]86-*-nto*): Add build_gdbserver=yes.

gdb/testsuite/
* lib/gdbserver-support.exp (gdbserver_start): Loop spawning
gdbserver increasing $portnum if "Can't bind address" has been seen.

Adding Neutrino gdbserver.
* configure: Regenerated.
* configure.ac: Add case for srv_qnx and set LIBS accordingly.
* configure.srv (i[34567]86-*-nto*): New target.
* nto-low.c, nto-low.h, nto-x86-low.c: New files.
* remote-utils.c [__QNX__]: Include sys/iomgr.h
(nto_comctrl) [__QNX__]: New function.
(enable_async_io, disable_async_io) [__QNX__]: Call nto_comctrl.

gdb/
* source.c (find_and_open_source): Remove the `objfile' parameter.
Update the function comment.
(open_source_file, symtab_to_fullname, psymtab_to_fullname): Update the
find_and_open_source callers.

* cygtls.h (struct _local_storage): Add thread storage pointers for
memory used by socket select functions.  Combine them into a single
struct select.
* cygtls.cc: Accommodate above change throughout.
(_cygtls::init_thread): Drop initalizing of sockevt to
INVALID_HANDLE_VALUE.
(_cygtls::fixup_after_fork): Reset sockevt to NULL.
(_cygtls::remove): Don't use sockevt value to bail out prematurely.
Set sockevt to NULL.  Free malloced select members.
* select.h (struct select_socket_info): Drop max_w4 member.
* select.cc (thread_socket): Use INFINITE timeout value if number of
objects to wait for is <= MAXIMUM_WAIT_OBJECTS.  Use num_w4 member
of select_socket_info struct rather than dropped max_w4.
(init_tls_select_info): New inline function to initialize TLS select
members.
(start_thread_socket): Just call init_tls_select_info to initialize TLS
select members and use them later on.
(socket_cleanup): Don't free select_socket_info pointer members since
they are thread local now.

* dtable.cc (handle_to_fn): Detect failing NtQueryObject.

2009-07-06  Tristan Gingold  <gingold@adacore.com>

* objdump.c (dump_dwarf_section): Handle .eh_frame like other dwarf
sections.
* dwarf.h (struct dwarf_section_display): Remove eh_frame field.
* dwarf.c (debug_displays): Adjust for removed field.

    * ld.texinfo: Fix typo.