merge from gcc

2007-11-13  H.J. Lu  <hongjiu.lu@intel.com>

PR binutils/5233
* ld-elf/flags1.d: New.
* ld-elf/flags1.ld: Likewise.
* ld-elf/flags1.s: Likewise.

* config/tc-mn10300.c (mn10300_force_relocation): Force a reloc to be generated for alignment fixups.
* config/tc-mn10300.h (TC_FORCE_RELOCATION): Call mn10300_force_relocation.
* elf-m10300.c (mn10300_elf_final_link_relocate): Prevent the accidental termination of DWARF location list entries.
  (mn10300_elf_relax_delete_bytes): Stop deletion if an align reloc is encountered that is larger than or not a mutliple of the number of bytes being deleted.
  When adjusting symbols, any symbols inside the region being deleted must be moved to the end of the region.
  Move align relocs forward if there is room for them after the deletion of the region.

PR 5233
* elf.c (rewrite_elf_program_header): Formatting.  Add
first_matching_lma and first_suggested_lma booleans and use
instead of testing matching_lma and suggested_lma for zero.

From Craig Silverstein: Support debug info for shared libraries.

merge from gcc

* floatformat.c (floatformat_ibm_long_double_is_valid): Fix
compiler warnings.
(floatformat_ibm_long_double): Use
floatformat_ibm_long_double_is_valid.

*** empty log message ***

daily update

* gas/macros/purge.s: Delete irpc loops.

2007-11-12  H.J. Lu  <hongjiu.lu@intel.com>

PR binutils/5299
* peicode.h (pe_bfd_object_p): Save and restore previous bfd
state when calling coff_object_p.

From Craig Silverstein: Dwarf_line_info can work with Object rather
than Sized_relobj.

From Craig Silverstein: Track_relocs doesn't need to hold onto the
object.

From Craig Silverstein: rework DWARF reader code a bit.

Correct handling of GOT references in a regular object.

PR gas/5269
* config/tc-frv.c (md_show_usage): Rewrite usage description to make it easier to translate.
* gas/po/gas.pot: Regenerate.

2007-11-12  Markus Deuling  <deuling@de.ibm.com>

* gdbarch.sh (gdbarch_alloc): Replace current_gdbarch by gdbarch. Remove
obsolete comment.
(verify_gdbarch, gdbarch_dump): Likewise.
(do_read): Update comment.
(addr_bit, bfd_arch_info, target_desc, long_long_bit, floatformat)
(ptr_bit, name_of_malloc): Replace current_gdbarch by gdbarch.
* gdbarch.{c,h}: Regenerate.

2007-11-12  Markus Deuling  <deuling@de.ibm.com>

* xtensa-tdep.c (ARG_1ST): Replace ARGS_FIRST_REG by its expression.
(ARGS_FIRST_REG): Remove.
(areg_numer): New function.
(xtensa_pseudo_register_read, xtensa_frame_prev_register)
(xtensa_extract_return_value, xtensa_store_return_value)
(xtensa_frame_cache): Replace AREG_NUMBER by areg_number.
(AREG_NUMBER): Remove.
(XTENSA_IS_ENTRY, ARG_NOF, ARG_1ST): Add gdbarch as parameter.
(xtensa_frame_cache): Update use of XTENSA_IS_ENTRY.
(xtensa_push_dummy_call): Update use of ARG_NOF.
(extract_call_winsize): Add gdbarch as parameter. Replace
current_gdbarch by gdbarch.
(xtensa_extract_return_value, xtensa_store_return_value): Update call
of extract_call_winsize.

    * NEWS: Add entry for new set/show print frame-args command.

Fix date in the last entry.

    * gdb.texinfo (Print Settings): Add documentation for "set/show
        print frame-arguments".

* elf.c (_bfd_elf_get_synthetic_symtab): Only bump the symbol
        pointer when we have a valid symbol.  Init udata.p to NULL.

* gas/ppc/regnames.d: New.
* gas/ppc/regnames.s: Likewise.
* gas/ppc/ppc.exp: Run it.

* config/tc-ppc.c (md_assemble): If -mregnames, when parsing
PPC_OPERAND_CR always parse as expression to allow register name
followed by an expression.

*** empty log message ***

daily update

*** empty log message ***

daily update

*** empty log message ***

From Craig Silverstein: Initialize got_relocs.

* mn10300-tdep.c (mn10300_dwarf2_reg_to_regnum): Add missing
register number to map.  Adjust warning check and error/warning
return value code to match expectations of this function's callers.

From Craig Silverstein: Clean up make rules for debug_msg.

From Craig Silverstein: Add test for debug messages.

From Craig Silverstein: Implement section_info.

From Craig Silverstein: Use relocations in reporting error message
locations.

daily update

    * stack.c (print_this_frame_argument_p): Handle arguments passed
        by reference properly.

From Cary Coutant: More shared library support, some refactorization.

Fix the date of my last changelog entry.

    * stack.c (print_frame_arguments_choices): New static global.
        (print_frame_arguments): Likewise.
        (print_this_frame_argument_p): New function.
        (print_frame_args): Print the argument value only when appropriate.
        (_initialize_task): Add new "set/show print frame-arguments" command.

    * ppc-linux-nat.c (ppc_linux_insert_watchpoint): Move
        saved_dabr_value assignment to the correct position.

2007-11-09  H.J. Lu  <hongjiu.lu@intel.com>

* config/tc-i386.c (match_template): Re-order suffix check.

2007-11-09  Markus Deuling  <deuling@de.ibm.com>

* arch-utils.c (generic_convert_register_p): Add gdbarch as parameter.
* arch-utils.h (generic_convert_register_p): Likewise.

* elf-vxworks.c (elf_vxworks_emit_relocs): Don't clobber
rel_hash.  Move loop inside if test.

Apply delta accidentally omitted from 2007-11-05 commit.
* pei-i386.c (COFF_SECTION_ALIGNMENT_ENTRIES): Let .data, .text
 and .bss section use the default.

From Cary Coutant: mark negative constant as unsigned.

Generate a complete exception frame header.  Discard duplicate
exception frame information.

* gdbarch.sh (convert_register_p): Add gdbarch as parameter.
* ia64-tdep.c (ia64_convert_register_p): Likewise.
* i387-tdep.c (i387_convert_register_p): Likewise.
* i387-tdep.h (i387_convert_register_p): Likewise.
* alpha-tdep.c (alpha_convert_register_p): Likewise.
* gdbarch.{c,h}: Regenerate.

* rs6000-tdep.c (rs6000_convert_register_p): Add gdbarch as parameter.
Replace current_gdbarch by gdbarch.
* mips-tdep.c (mips_convert_register_p): Likewise.
* m68k-tdep.c (m68k_convert_register_p): Likewise.
* i386-tdep.c (i386_convert_register_p): Likewise.

*** empty log message ***

daily update

* sim/cris/asm/x0-v10.ms, sim/cris/asm/x0-v32.ms: Tweak
stack-pointer match pattern for 4K host environment.

2007-11-08  Aleksandar Ristovski  <aristovski@qnx.com>

* gdb_string.h: Include <strings.h>.

* breakpoint.c (break_command_1): Remove
pending_bp parameter.
(create_breakpoint): Likewise.
(create_breakpoints): Likewise. Adjust call to
create_breakpoint.
(break_command_1): Likewise. Adjust call to
create_breakpoints.
(do_captured_breakpoint): Adjust call to
create_breakpoints.
(break_command, tbreak_command, hbreak_command)
(stopin_command, stopat_command): Adjust call
to break_command_1.

* elf-vxworks.h (elf_vxworks_add_dynamic_entries): Declare.
(elf_vxworks_finish_dynamic_entry): Declare.
* elf-vxworks.c: Include elf/vxworks.h.
(elf_vxworks_add_dynamic_entries): New.
(elf_vxworks_finish_dynamic_entry): New.
* Makefile.am (elf-vxworks.lo): Add dependency.
* Makefile.in (elf-vxworks.lo): Add dependency.
* elf32-i386.c (elf_i386_size_dynamic_sections,
elf_i386_finish_dynamic_sections): Call
elf_vxworks_add_dynamic_entries and
elf_vxworks_finish_dynamic_entry.
* elf32-ppc.c (ppc_elf_size_dynamic_sections,
ppc_elf_finish_dynamic_sections): Likewise.
* elfxx-sparc.c (_bfd_sparc_elf_size_dynamic_sections,
sparc_finish_dyn): Likewise.
* elf32-sh.c (sh_elf_size_dynamic_sections,
sh_elf_finish_dynamic_sections): Likewise.
* elfxx-mips.c (_bfd_mips_elf_size_dynamic_sections,
_bfd_mips_elf_finish_dynamic_sections): Likewise.
* elf32-arm.c (elf32_arm_size_dynamic_sections,
elf32_arm_finish_dynamic_sections): Likewise.

include/elf/
* vxworks.h: New.

bfd/
* elf-vxworks.h (elf_vxworks_add_dynamic_entries): Declare.
(elf_vxworks_finish_dynamic_entry): Declare.
* elf-vxworks.c: Include elf/vxworks.h.
(elf_vxworks_add_dynamic_entries): New.
(elf_vxworks_finish_dynamic_entry): New.
* Makefile.am (elf-vxworks.lo): Add dependency.
* Makefile.in (elf-vxworks.lo): Add dependency.
* elf32-i386.c (elf_i386_size_dynamic_sections,
elf_i386_finish_dynamic_sections): Call
elf_vxworks_add_dynamic_entries and
elf_vxworks_finish_dynamic_entry.
* elf32-ppc.c (ppc_elf_size_dynamic_sections,
ppc_elf_finish_dynamic_sections): Likewise.
* elfxx-sparc.c (_bfd_sparc_elf_size_dynamic_sections,
sparc_finish_dyn): Likewise.
* elf32-sh.c (sh_elf_size_dynamic_sections,
sh_elf_finish_dynamic_sections): Likewise.
* elfxx-mips.c (_bfd_mips_elf_size_dynamic_sections,
_bfd_mips_elf_finish_dynamic_sections): Likewise.
* elf32-arm.c (elf32_arm_size_dynamic_sections,
elf32_arm_finish_dynamic_sections): Likewise.

ld/
* ld-vxworks/tls-2.d: New.
* ld-vxworks/tls-2.s: New.

    * config/tc-xtensa.c (relaxable_section): Check for .eh_frame.

2007-11-07  Joseph Myers  <joseph@codesourcery.com>
    Daniel Jacobowitz  <dan@codesourcery.com>

* ppc-sysv-tdep.c (ppc64_sysv_abi_push_dummy_call): Right-align
struct values smaller than one doubleword; left-align those
larger.  Pass structs containing a single floating-point value in
registers.

include:
2007-11-07  Joseph Myers  <joseph@codesourcery.com>
    Daniel Jacobowitz  <dan@codesourcery.com>

* floatformat.h (struct floatformat): Add split_half field.
(floatformat_ibm_long_double): New.

libiberty:
2007-11-07  Joseph Myers  <joseph@codesourcery.com>
    Daniel Jacobowitz  <dan@codesourcery.com>

* floatformat.c (mant_bits_set): New.
(floatformat_to_double): Use it.  Note no special handling of
split formats.
(floatformat_from_double): Note no special handing of split
formats.
(floatformat_ibm_long_double_is_valid,
floatformat_ibm_long_double): New.
(floatformat_ieee_single_big, floatformat_ieee_single_little,
floatformat_ieee_double_big, floatformat_ieee_double_little,
floatformat_ieee_double_littlebyte_bigword, floatformat_vax_f,
floatformat_vax_d, floatformat_vax_g, floatformat_i387_ext,
floatformat_m68881_ext, floatformat_i960_ext,
floatformat_m88110_ext, floatformat_m88110_harris_ext,
floatformat_arm_ext_big, floatformat_arm_ext_littlebyte_bigword,
floatformat_ia64_spill_big, floatformat_ia64_spill_little,
floatformat_ia64_quad_big, floatformat_ia64_quad_little): Update
for addition of split_half field.

gdb:
2007-11-07  Joseph Myers  <joseph@codesourcery.com>
    Daniel Jacobowitz  <dan@codesourcery.com>

* gdbtypes.c (floatformats_ibm_long_double): New.
* gdbtypes.h (floatformats_ibm_long_double): Declare.
* ia64-tdep.c (floatformat_ia64_ext): Update for addition of
split_half field.
* mips-tdep.c (n32n64_floatformat_always_valid,
floatformat_n32n64_long_double_big, floatformats_n32n64_long):
Remove.
(mips_gdbarch_init): Use floatformats_ibm_long_double instead of
floatformats_n32n64_long.
* ppc-linux-tdep.c (ppc_linux_init_abi): Use 128-bit IBM long
double.
* doublest.c (convert_floatformat_to_doublest,
convert_doublest_to_floatformat): Handle split floating-point
formats.
* ppc-sysv-tdep.c (ppc_sysv_abi_push_dummy_call): Handle IBM long
double arguments.
(ppc64_sysv_abi_push_dummy_call): Likewise.
(do_ppc_sysv_return_value): Handle IBM long double return.

*** empty log message ***

daily update

From Craig Silverstein and Cary Coutant: fix assignment of section
indexes to sections which are not associated with segments.

Fix crash when a variable object being deleted
has any of its children deleted previously.

* varobj.c (delete_variable_1): Don't recurse
into deleted children.

Check that we don't set the output section index twice.

2007-11-07  Eric B. Weddington  <eweddington@cso.atmel.com>

* config/tc-avr.c (mcu_types): Add ATtiny88 device.
* doc/c-avr.texi: Likewise.

2007-11-07  Anatoly Sokolov <aesok@post.ru>

* config/tc-avr.c (mcu_types): Add new devices: ATmega48P, ATmega88P,
ATmega168P, Atmega328P
* doc/c-avr.texi: Document new devices.

PR gas/5228
* m68k-opc.c (m68k_opcodes): Fix coldfire msac.w instructions with parallel loads.

* ia64-dis.c (print_insn_ia64): Generate symbolic names for cr
 registers instead of register number.

* gas/ia64/regs.d: Expect symbolic names for cr registers due to
 improved disassembler.

* bfd/elfxx-mips.c (_bfd_mips_elf_fake_sections): Force SHF_MIPS_NOSTRIP on
   .debug_frame for Irix.

* arm-dis.c (arm_opcodes): Remove superflous escapes of percent operators.

* gas/elf/elf.exp (run_elf_list_test): Fix typo.

* config/tc-ppc.c (md_apply_fix): For PPC_TOC16 on XCOFF, uses offset
  within the TOC instead of the VMA.

* gas/ppc/test1xcoff32.d: Updated to match RTOC bug fix.

ld/
* emulparams/vxworks.sh (OTHER_READONLY_SECTIONS,
OTHER_READWRITE_SECTIONS): Add TLS sections.

ld/testsuite/
* ld-vxworks/tls-1.d: New.
* ld-vxworks/tls-1.s: New.

* doc/binutils.texi: Update to FDL 1.2.
  (Top): consistent punctuation in menu entries.
  (Index): Rename from Binutils Index.
* doc/fdl.texi: Update from http://www.gnu.org/licenses/fdl.texi.

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* gdbarch.sh (legacy_virtual_frame_pointer): Add gdbarch parameter.
* gdbarch.{c,h}: Regenerate.
* arch-utils.c (legacy_virtual_frame_pointer): Add gdbarch parameter.
Replace current_gdbarch by gdbarch.
* m32c-tdep.c (m32c_virtual_frame_pointer): Likewise.
* mips-tdep.c (mips_virtual_frame_pointer): Likewise.
* spu-tdep.c (spu_virtual_frame_pointer): Likewise.

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* shnbsd-tdep.c (shnbsd_supply_gregset, shnbsd_collect_gregset): Use
get_regcache_arch to get at the current architecture by regcache.
* xstormy16-tdep.c (xstormy16_frame_prev_register): Use get_frame_arch
to get at the current architecture by frame_info.
* fbsd-nat.c (fbsd_make_corefile_notes): Use get_regcache_arch to get
at the current architecture by regcache.
* gnu-v3-abi.c (gnuv3_skip_trampoline): Use get_frame_arch to get at
the current architecture by frame_info.
(build_gdb_vtable_type): Replace current_gdbarch by gdbarch.
* aix-thread.c (special_register_p): Add gdbarch as parameter.
(fetch_regs_kernel_thread, store_regs_kernel_thread): Add gdbarch to
caller of special_register_p.

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* infcall.c (call_function_by_hand): Use get_frame_arch to get at the
current architecture by frame_info.

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* xtensa-tdep.h (SIZEOF_GREGSET, XTENSA_ELF_NGREG) Remove.

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* s390-nat.c (supply_gregset, fill_gregset): Use get_regcache_arch to
get at the current architecture by regcache.
(SUBOFF): Add gdbarch to macro definition.

*** empty log message ***

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* frv-tdep.c (frv_register_name, frv_check_watch_resources): Replace
macro CURRENT_VARIANT by its expression.
(CURRENT_VARIANT): Remove.
(frv_check_watch_resources): Add gdbarch parameter.

2007-11-07  Markus Deuling  <deuling@de.ibm.com>

* MAINTAINERS (Write After Approval): Add self.

From Craig Silverstein: Templatize the Dwarf reader.

From Craig Silverstein: Handle a .so file in a .a file.

From Craig Silverstein: Don't crash if we see a file we don't recognize.

*** empty log message ***

* ppc-opc.c (powerpc_opcodes): Remove the dcffix and dcffix. opcodes
which are not included in the "Preliminary Decimal Floating-Point
Architecture" document.

daily update

2007-11-06  Paul Brook  <paul@codesourcery.com>

gas/
* config/tc-arm.c (do_mull): Allow overlapping Rm for armv6.

gas/testsuite/
* gas/arm/mul-overlap.s: Add umull and smlal.
* gas/arm/mul-overlap.l: Update expected results.

* ehopt.c (check_eh_frame): If md_allow_eh_opt is defined, invoke it to see if the optimizations should be applied.
* config/tc-mn10300.h (md_allow_eh_opt): Define.  Only allow call frame optimization if linker relaxation is not enabled.
* gas/elf/elf.exp: Disable ehopt test for mn10300.

bfd/
* elf64-ppc.c (ppc64_elf_check_relocs): Don't refcount tlsld_got here..
(ppc64_elf_gc_sweep_hook): ..or here..
(ppc64_elf_tls_optimize): ..or here.  Make two passes through the
relocs, ensuring that tls_get_addr calls follow gd and ld relocs.
(allocate_dynrelocs): Refcount tlsld_got here.
(ppc64_elf_size_dynamic_sections): Allocate local got and call
allocate_dynrelocs before allocating tlsld_got.
(ppc64_elf_relocate_section): Remove check that a tls_get_addr
call follows gd and ld relocs.
ld/testsuite/
* ld-powerpc/tlsso.d: Update for changed got alloc order.
* ld-powerpc/tlsso.r: Likewise.

Add __ImageBase symbol to pe.em and pep.em.

bfd/
* elf32-ppc.c (ppc_elf_check_relocs): Don't refcount tlsld_got here..
(ppc_elf_gc_sweep_hook): ..or here..
(ppc_elf_tls_optimize): ..or here.  Make two passes through the
relocs, ensuring that tls_get_addr calls follow gd and ld relocs.
(allocate_dynrelocs): Refcount tlsld_got here.
(ppc_elf_size_dynamic_sections): Call allocate_dynrelocs before
allocating tlsld_got.
(ppc_elf_relocate_section): Remove check that a tls_get_addr
call follows gd and ld relocs.
ld/testsuite/
* ld-powerpc/tlsso32.d: Update for changed got alloc order.

*** empty log message ***

daily update

* features/rs6000/power-core.xml: Use uint32 as register type.
* features/rs6000/power64-core.xml: Use uint64/uint32 as register type.
* features/rs6000/powerpc-32.c: Regenerate.
* features/rs6000/powerpc-403.c: Regenerate.
* features/rs6000/powerpc-403gc.c: Regenerate.
* features/rs6000/powerpc-505.c: Regenerate.
* features/rs6000/powerpc-602.c: Regenerate.
* features/rs6000/powerpc-603.c: Regenerate.
* features/rs6000/powerpc-604.c: Regenerate.
* features/rs6000/powerpc-64.c: Regenerate.
* features/rs6000/powerpc-7400.c: Regenerate.
* features/rs6000/powerpc-750.c: Regenerate.
* features/rs6000/powerpc-860.c: Regenerate.
* features/rs6000/powerpc-e500.c: Regenerate.

* linux-low.c: Include <sched.h>.

    * breakpoint.c (set_raw_breakpoint_without_location): Make static.

    * breakpoint.h (create_solib_load_event_breakpoint)
        (create_solib_unload_event_breakpoint)
        (create_fork_event_catchpoint, create_vfork_event_catchpoint)
        (create_exec_event_catchpoint): Delete declaration.
        * breakpoint.c (create_solib_load_event_breakpoint)
        (solib_load_unload_1, create_solib_unload_event_breakpoint): Delete.
        (create_fork_event_catchpoint, create_vfork_event_catchpoint)
        (create_exec_event_catchpoint): Make static.

    * breakpoint.h (set_breakpoint_sal): Remove declaration.
        * breakpoint.c (set_breakpoint_sal): Delete.

    * buildsym.c (finish_block): Remove "#if 1"/"#endif" brackets.
        (make_blockvector): Likewise. Remove FIXME comment.