* config.sub, config.guess: Import latest from subversions.

* ldlang.c (insert_pad): Fix typos in last patch.

* ldlang.c: When traversing lang_statement_union_type lists,
consistently use "header.next" rather than "next".
* mpw-eppcmac.c: Likewise.
* emultempl/beos.em: Likewise.
* emultempl/hppaelf.em: Likewise.
* emultempl/pe.em: Likewise.
* ldlang.h (union lang_statement_union): Remove "next" field.

* ldlang.c (insert_pad): Use offsetof macro.
(lang_size_sections): Always neuter padding statements.
* emultempl/hppaelf.em (hppaelf_delete_padding_statements): Delete.

* pe-dll.c (pe_dll_fill_sections): Correct type of "relax" param
passed to lang_size_sections.
(pe_exe_fill_sections): Likewise.
* emultempl/pe.em (output_prev_sec_find): Copied from elf32.em.
(gld_${EMULATION_NAME}_place_orphan): Merge from elf32.em.

*** empty log message ***

* archive.c (offsetof): Remove define.
* elf.c: Likewise.
* oasys.c: Likewise
* sysdep.h (offsetof): Define.

* mi0-var-block.exp, mi0-stack.exp, mi0-simplerun.exp,
mi0-regs.exp, mi0-watch.exp, mi0-stepi.exp, mi0-until.exp,
mi0-return.exp, mi0-read-memory.exp, mi0-eval.exp,
mi0-disassemble.exp, mi0-console.exp, mi-watch.exp,
mi-var-display.exp, mi-var-cmd.exp, mi-var-child.exp,
mi-until.exp, mi-var-block.exp, mi-stepi.exp, mi-stack.exp,
mi-simplerun.exp, mi-return.exp, mi-regs.exp, mi-read-memory.exp,
mi-disassemble.exp, mi-eval.exp, mi-console.exp: Replace pattern
matching thread=0 with one to also match thread=1.
Fix PR gdb/190.

* lib/mi-support.exp (mi_gdb_start): If a remote target, use the
CLI jump command to start it.
(mi_run_to_main): Fail immediatly when unexpected output.

* mi0-until.exp, mi-until.exp, mi0-stepi.exp, mi-stepi.exp,
mi-simplerun.exp, mi0-simplerun.exp, mi0-return.exp,
mi-return.exp, mi0-console.exp, mi-console.exp: Recognize an
unexpected run-to-main response. Report as a fail.

*** empty log message ***

merge from gcc

* lib/mi-support.exp (mi_gdb_start): Move call to sid_start to
beginning of function.  Fix PR gdb/191.

2001-08-18  H.J. Lu  <hjl@gnu.org>

* config/tc-mips.c (show): Add the missing prototype.

* config/tc-mips.c (S_EX_*): New enum for my_getSmallExpression()
return values.
(mips_ip): Use the new return values instead of characters. Add
support for %higher and %highest.
(LP): Remove.
(RP): Remove.
(my_getSmallExpression): Make parsing case insensitive and more
reliable. Add support for %higher and %highest. Further support to
parse %gprel and %neg is implemented but currently deactivated.

* emultempl/aix.em: Formatting fixes.

* ldlang.c (insert_pad): Make use of an existing pad statement if
available.  Move code calculating alignment, adjusting section
alignment power, and adjusting dot to ..
(size_input_section): .. here.  Remove unused relax param.
(lang_size_sections): Change boolean `relax' param to boolean *.
Adjust call to size_input_section.  Make use of insert_pad to
place pad after the assignment statement.  If relaxing, zap
padding statements.
(reset_memory_regions): Reset output_bfd section sizes too.
(relax_again): Move to..
(lang_process): ..here.  Adjust call to lang_size_sections, and
remove duplicated code.
* ldlang.h (lang_size_sections): Change `relax' param to boolean *.

*** empty log message ***

merge from gcc

2001-08-17  H.J. Lu  <hjl@gnu.org>

* gas/sparc/unalign.d: Support Sparc V9.

* mi-cmd-var.c (varobj_update_one): Update call to
varobj_update to reflect recent api change.

* varobj.c (varobj_update): Change first parameter to
pointer to struct varobj*. This function can delete
varobjs, so we need to give callers the new varobj
when this happens.
(value_of_root): Update "var", too, if "var_handle"
changes.
* varobj.h (varobj_update): Likewise.

* Makefile.in (varobj_h): Define.
(mi-cmd-var.o): Depends on varobj_h.
(gdbtk-varobj.o): Change varobj.h to $(varobj_h).

* ecoff.c (ecoff_frob_symbol): Remove casts from bfd_get_gp_size.
(ecoff_build_symbols): Likewise.
* read.c (s_lcomm_internal): Fix signed/unsigned warning.

* bfd.c (bfd_get_gp_size): Return an unsigned int.
(bfd_set_gp_size): Make param unsigned.
* bfd-in2.h: Regenerate.
* elf32-ppc.c (ppc_elf_add_symbol_hook): Use elf_gp_size rather
than calling bfd_get_gp_size.
* elf64-alpha.c (elf64_alpha_add_symbol_hook): Likewise.
* elfxx-ia64.c (elfNN_ia64_add_symbol_hook): Likewise.

* ld.texinfo: Document that fill values now use the four least
significant bytes.
* emulparams/elf32fr30.sh (NOP): Update.
* emulparams/elf32mcore.sh: Likewise.
* emulparams/elf64_s390.sh: Likewise.
* emulparams/elf_i386.sh: Likewise.
* emulparams/elf_i386_be.sh: Likewise.
* emulparams/elf_i386_chaos.sh: Likewise.
* emulparams/elf_i386_ldso.sh: Likewise.
* emulparams/elf_s390.sh: Likewise.
* emulparams/elf_x86_64.sh: Likewise.
* emulparams/i386moss.sh: Likewise.
* emulparams/i386nw.sh: Likewise.
* emulparams/m68kelf.sh: Likewise.
* scripttempl/elf.sc: Update NOP comment.
* scripttempl/elfi370.sc: Likewise.
* scripttempl/elfm68hc11.sc: Likewise.
* scripttempl/elfm68hc12.sc: Likewise.
* scripttempl/nw.sc: Likewise.

* linker.c (default_fill_link_order): Handle four byte fill value.

* elf64-sparc.c (sparc64_elf_output_arch_syms): Add missing
prototype.
* nlm32-alpha.c (nlm_alpha_write_external): Fix warning.

* elf64-mips.c (UNUSED_RELOC): Define.
(mips_elf64_howto_table_rel): Use it.
(mips_elf64_howto_table_rela): Here too.
(mips_elf64_write_relocs): Fix signed/unsigned warning.

Oops missed from last commit

* coffcode.h (coff_write_object_contents): Add ATTRIBUTE_UNUSED to
silence warning.
* coffgen.c (coff_print_symbol): Fix warnings.
(coff_find_nearest_line): Likewise.
* cofflink.c (_bfd_coff_link_input_bfd): Likewise.
* coff-alpha.c (alpha_convert_external_reloc): Likewise.
* format.c (bfd_check_format_matches): Likewise.
(bfd_set_format): Likewise.
* coff-ia64.c: Add missing prototypes.
* elf64-alpha.c (struct alpha_elf_link_hash_entry): Make addend
signed to silence warnings.
(elf64_alpha_relocate_section): Likewise.
(elf64_alpha_find_reloc_at_ofs): Fix warnings.
(elf64_alpha_add_symbol_hook): Likewise.
(elf64_alpha_final_link): Likewise.
(elf64_alpha_relax_section): Remove redundant assign to info.gotent.
(elf64_alpha_merge_gots): Add ATTRIBUTE_UNUSED to unused args.
(elf64_alpha_size_got_sections): Likewise.
* elfxx-ia64.c: Add missing prototypes.
(elfNN_ia64_relocate_section): Fix warning.
(elfNN_ia64_unwind_entry_compare): Make params const.

merge from gcc

* config.bfd (targ64_selvecs): New.
<powerpc-*-aix* entry>: Use it here instead of ineffectual #ifdef.

* bfd-in.h (BFD_VERSION): Remove wrong comment.
(BFD_DEFAULT_TARGET_SIZE): New.
(BFD_ARCH_SIZE): Comment.
* configure.in (target_size): New.  Set instead of target64 in
selvecs case statement.  Set target64 from it.
(bfd_default_target_size): New.  Set from taget_size. AC_SUBST.
* bfd-in2.h: Regenerate.
* configure: Regenerate.

* bfd.c (enum bfd_error): Add bfd_error_wrong_object_format.
(bfd_errmsgs): Add corresponding message.
* archive.c (bfd_generic_archive_p): Don't release bfd_ardata when
finding an archive that contains different format object files.
Return bfd_error_wrong_object_format for this case.
* format.c: Formatting fixes.  s/CONST/const/.
(bfd_check_format_matches): Accept archives that give
bfd_error_wrong_object_format if no full match is found.  Tidy
code handling matching_vector.  Don't return a pointer to freed
memory in `matching'.  Handle ambiguous matches as for partial
archive matches.
* bfd-in2.h: Regenerate.

*** empty log message ***

* configure.in (GDBTK_SRC_DIR): We must have autoconf
give us the absolute path to the source directory.
* configure: Regenerated.
* Makefile.in (GDBTK_SRC_DIR): Add variable so that
autoconf can substitue it in for us.
(gdbtk.o): Pass in GDBTK_SRC_DIR instead of srcdir. We
really do need an absolute pathname.

Set MIPS n32 ABI flag in ELF header if appropriate.

Add support for MIPS R1[02]000 performance counter opcodes.

* mips-opc.c: R3900s can support all branch likely INSN_MACROs where
the corresponding non-likely insn is in MIPS I.

Remove obsolete special handling of 64bit Relocations.

* obvious fix

2001-08-16  Frank Ch. Eigler  <fche@redhat.com>

* config/sid.exp (sid_start): Don't warn if we cannot figure out
what to force sid endianness to.

*** empty log message ***

* lib/gdb.exp (gdbtk_start): Don't set environment
variables for TCL_LIBRARY and friends. Insight will
now figure these out for itself.

* Makefile.in (GDBTK_VERSION): Set a version number.
(GDBTK_LIBRARY): New variable to point to location where
gdbtk will install its tcl library. Changed all refereneces
to $(datadir)/insight1.0 to $(GDBTK_LIBRARY).
(SUBDIR_GDBTK_ALL): Remove "all-gdbtk". No longer needed.
(SUBDIR_GDBTK_CLEAN): Remove "clean-gdbtk". No longer needed.
(all-gdbtk): Remove.
(install-gdbtk): Remove.
(gdbtk.o): Pass srcdir to file when compiling.

Also from Ian Roxborough:

* Makefile.in: Changed all references to the install directory
"gdbtcl" to read "insight1.0".

Fix for ld

    * arch-utils.c (generic_skip_trampoline_code): New function.
        * arch-utils.h (generic_skip_trampoline_code): Declare external.
        * gdbarch.c: Regeberated from gdbarch.sh.
        * gdbarch.h: Ditto.
        * gdbarch.sh (SKIP_TRAMPOLINE_CODE): Multi-arch.
        * infrun.c: Remove default setting of SKIP_TRAMPOLINE_CODE macro.

* gdbint.texinfo: Add a cautionary note about macro use.

* libieee.h (common_header_type): Add last_byte field.

* ieee.c: Add missing prototypes.  Some format fixes.
(struct output_buffer_struct): Move for availability to prototypes.
(ieee_part_after): New function.
(ieee_slurp_debug): Use it here.
(ieee_seek): Pass in ieee_data_struct rather than bfd.  Use
ieee_part_after to set last_byte to one past end of current part.
Update callers.
(ieee_pos): Pass in ieee_data_struct rather than bfd.
Update callers.
(parse_expression): Don't go beyond the end of the current part.
(ieee_slurp_external_symbols): Correct type passed to read_2bytes.
(ieee_get_symtab_upper_bound, ieee_get_symtab,
ieee_get_symbol_info, ieee_print_symbol, ieee_new_section_hook,
ieee_get_reloc_upper_bound, ieee_canonicalize_reloc, block,
ieee_set_section_contents, ieee_write_object_contents,
ieee_make_empty_symbol): Make static.
(ieee_archive_p): Correct comments regarding bfd_read.
(ieee_object_p): Similarly.
(ieee_mkobject): Move it.  Clear output_ptr_start, output_ptr,
output_ptr_end, input_ptr_start, input_ptr, input_ptr_end,
input_bfd, output_bfd, output_buffer.
(do_as_repeat): Write out ieee_set_current_pc_enum value as for
do_with_relocs, ie. as a symbol for relocatable files.
(ieee_vec): Add comments showing functions referenced by macros.

2001-08-14  Daniel Jacobowitz  <drow@mvista.com>
    H.J. Lu  (hjl@gnu.org)

* partial-stab.h: valu should be a CORE_ADDR.

2001-08-14  H.J. Lu  (hjl@gnu.org)

* dbxread.c (SWAP_SYMBOL): Removed.
(INTERNALIZE_SYMBOL): Check sign extended vma.

*** empty log message ***

* top.c (SIGSETJMP): Protect env argument with parentheses.
(SIGLONGJMP): Protect env argument with parentheses.

2001-08-14  Jeff Johnston  <jjohnstn@redhat.com>

        * expr.c (operand)[LITERAL_PREFIXDOLLAR_HEX]: Treat $L as
        a label, not a hex constant.

This patch was approved by Nick Clifton.

sync with gcc - whitespace differences, one missing entry, and one
rearrangement.

Add missing protypes

* emultempl/elf32.em: Formatting fixes.
(output_prev_sec_find): Test for bfd_ind_section too;  do so by
looking at sec->owner.
(output_rel_find): Move function inside LDEMUL_PLACE_ORPHAN test.
(gld${EMULATION_NAME}_place_orphan): Add a few comments.  Remove
unused code, and reorganize orphan section placement code.

(wild_doit): Rename to lang_add_section.

* floatformat.c (floatformat_ieee_double_littlebyte_bigword): Fix
name.

*** empty log message ***

    * emultempl/needrelax.em: New file.
        * emulparams/elf64_ia64.sh (EXTRA_EM_FILE): Reference it.
        * Makefile.am (eelf64_ia64.c): Depend on it.
        * Makefile.in: Rebuild.

* emultempl/elf32.em: For SEC_EXCLUDE sections, ensure that
output_section is set non-NULL.

Add missing prototypes

* emultempl/elf32.em (gld${EMULATION_NAME}_place_orphan): Return
`true' for SEC_EXCLUDE sections so that the generic code doesn't
needlessly create an output_section_statement.  Treat a correctly
named output_section_statement with NULL bfd_section as compatible.

Fix bug in previous delta

oops - omitted from previous delta!

Add missing prototypes.  Fix formatting.

* mcore-dis.c: Fix formatting.
* mips-dis.c: Likewise.
* pj-dis.c: Likewise.
* z8k-dis.c: Likewise.

* config/tc-s390.c (s390_elf_cons): Correct fixups for PLT
relocations.

Update copyright notice

* emultempl/generic.em: Support EXTRA_EM_FILE.
(ld_${EMULATION_NAME}_emulation): Support emulation parameters
LDEMUL_BEFORE_PARSE, LDEMUL_SYSLIB, LDEMUL_HLL,
LDEMUL_AFTER_PARSE, LDEMUL_AFTER_OPEN, LDEMUL_AFTER_ALLOCATION,
LDEMUL_SET_OUTPUT_ARCH, LDEMUL_CHOOSE_TARGET,
LDEMUL_BEFORE_ALLOCATION, LDEMUL_GET_SCRIPT, LDEMUL_FINISH,
LDEMUL_CREATE_OUTPUT_SECTION_STATEMENTS,
LDEMUL_OPEN_DYNAMIC_ARCHIVE, LDEMUL_PLACE_ORPHAN,
LDEMUL_SET_SYMBOLS, LDEMUL_PARSE_ARGS, LDEMUL_UNRECOGNIZED_FILE,
LDEMUL_LIST_OPTIONS, LDEMUL_RECOGNIZED_FILE,
LDEMUL_FIND_POTENTIAL_LIBRARIES.

merge from gcc

*** empty log message ***

    * scripttempl/elf.sc, scripttempl/elfd30v.sc,
        scripttempl/elfm68hc11.sc, scripttempl/elfm68hc12.sc,
        scripttempl/v850.sc: Keep .jcr data.

    * cgen-ibld.in (extract_normal): Match type of VALUE and MASK
        to *VALUEP.  Regenerate all cgen files.

Extend unaligned test to check .uaword with mulitple arguments

Fix behaviour of unaligned data directives

Make --verbose always display linker script

* config/mn10300/mn10300.mt (TM_FILE): Delete.
* configure.tgt: Add mn10300 to list of mulit-arch targets.
* config/mn10300/tm-mn10300.h: Delete file. Move contents ...
* mn10300-tdep.c: To here.

Convert some floatformat_{to,from}_double to floatformat_{to,from}_doublest.

* config/mn10300/tm-mn10300.h (PC_REGNUM, SP_REGNUM): Delete.
* mn10300-tdep.c (mn10300_gdbarch_init): Initialize pc_regnum and
sp_regnum.

* doublest.c (convert_doublest_to_floatformat): Rename
floatformat_from_doublest.  Make static.
(convert_floatformat_to_doublest): Rename floatformat_to_doublest.
Make static.
(floatformat_to_doublest): New function.
(floatformat_from_doublest): New function.
(host_float_format, host_double_format, host_long_double_format):
New static variables.
(store_floating, extract_floating): Always use
floatformat_to_doublest and floatformat_from_doublest.
* doublest.h (HOST_LONG_DOUBLE_FORMAT): Delete macro.

* config/mn10300/tm-mn10300.h (INIT_FRAME_PC): Delete.
* mn10300-tdep.c (mn10300_gdbarch_init): Initialize init_frame_pc.

*** empty log message ***

replace direct invocation og elf vma print functions with indirect invocation

* config/tc-mips.c: Fix a comment typo.

2001-08-10  Andreas Jaeger  <aj@suse.de>

* elf64-sparc.c: Add missing prototypes.
* elf32-cris.c: Likewise.
* elf32-i370.c: Likewise.
* elf64-hppa.c: Likewise.
* elf64-s390.c: Likewise.

* target.h (TARGET_VIRTUAL_FRAME_POINTER): Delete, multi-arched.
* gdbarch.sh (TARGET_VIRTUAL_FRAME_POINTER): Add.
* gdbarch.h, gdbarch.c: Regenerate.

* arch-utils.h (legacy_virtual_frame_pointer): Declare.
* arch-utils.c: Include "gdb_assert.h".
(legacy_virtual_frame_pointer): Define.
* Makefile.in (arch-utils.o): Depends on gdb_assert.h.

* tracepoint.c (encode_actions): Make frame_reg an int.  Make
frame_offset a LONGEST.
* ax-gdb.c (gen_frame_args_address): Ditto.
(gen_frame_locals_address): Ditto.
* mn10300-tdep.c (mn10300_gdbarch_init): Initialize
virtual_frame_pointer.
(mn10300_virtual_frame_pointer): Make static.  Update parameter
list to match function signature.
* config/mn10300/tm-mn10300.h (TARGET_VIRTUAL_FRAME_POINTER): Delete.

* gdb-events.sh: Add architecture_changed event.
* gdbarch.sh: Include gdb-event.h.
(gdbarch_update_p): Notify UIs when architecture changes.
* gdb-events.h: Regenerated.
* gdb-events.c: Regenerated.
* gdbarch.c: Regenerated.

*** empty log message ***

(lbasename): Change function declaration to return a const char pointer.

2001-08-10  Michael Snyder  <msnyder@redhat.com>

* regcache.c (legacy_write_register_gen): Don't 'optimize out'
        a write_register to a pseudo-reg.  Target_store_pseudo_register
        needs to get called, because these regs may be computed and may
        have side-effects.

2001-08-10  H.J. Lu  <hjl@gnu.org>

* gas/mips/e32-rel2.d: Updated for the objdump output change.
* gas/mips/e32el-rel2.d: Likewise.
* gas/mips/elf-rel.d: Likewise.
* gas/mips/elf-rel2.d: Likewise.
* gas/mips/elf-rel3.d: Likewise.
* gas/mips/elf_e_flags1.d: Likewise.
* gas/mips/elf_e_flags2.d: Likewise.
* gas/mips/elf_e_flags3.d: Likewise.
* gas/mips/elf_e_flags4.d: Likewise.
* gas/mips/elfel-rel.d: Likewise.
* gas/mips/elfel-rel2.d: Likewise.
* gas/mips/elfel-rel3.d: Likewise.
* gas/mips/empic.d: Likewise.
* gas/mips/mips16-e.d: Likewise.
* gas/mips/mips16-f.d: Likewise.
* gas/mips/mipsel16-f.d: Likewise.
* gas/mips/telempic.d: Likewise.
* gas/mips/tempic.d: Likewise.
* gas/mips/tmips16-e.d: Likewise.
* gas/mips/tmipsel16-e.d: Likewise.

* gas/mips/mips16-f.d: Fix the symbol table output.

* gas/mips/mipsel16-f.d: Fix the endian.

* gas/mips/mips.exp; Don't xfail mips16-f/mipsel16-f.

* elf64-x86-64.c (elf64_x86_64_elf_object_p): Make static.
Add prototypes where needed.

2001-08-10  H.J. Lu  <hjl@gnu.org>

* bfd.c (bfd_sprintf_vma): Fix a typo in the last change.
(bfd_fprintf_vma): Likewise.

2001-08-10  H.J. Lu  <hjl@gnu.org>

* nm.c (print_value): Add one arg, bfd *.
(print_value): Replace fprintf_vma with bfd_fprintf_vma.
(print_symbol_info_bsd): Also pass `abfd' to print_value ().
(print_symbol_info_sysv): Likewise.
(print_symbol_info_posix): Likewise.

* objdump.c (dump_section_header): Replace printf_vma with
bfd_printf_vma.
(print_section_stabs): Likewise.
(dump_bfd_header): Likewise.
(objdump_print_value): Replace sprintf_vma with
bfd_sprintf_vma.
(disassemble_bytes): Likewise.
(dump_reloc_set): Likewise.

2001-08-10  H.J. Lu  <hjl@gnu.org>

* bfd.c (bfd_sprintf_vma): Don't return void.
(bfd_fprintf_vma): Likewise.

* gas/mips/mips-gp32-fp32.d: Expect addu rather than or for moves.
* gas/mips/mips-gp32-fp64.d: Likewise.
* gas/mips/mips-abi32.d: Likewise.
* gas/mips/mips-gp32-fp32-pic.d: Likewise.
* gas/mips/mips-gp32-fp64-pic.d: Likewise.
* gas/mips/mips-abi32-pic.d: Likewise.

* config/tc-mips.c (move_register): New function.
(macro_build): Remove OPCODE_IS_MEMBER's gp32 argument.
(mips_ip): Likewise.
(macro2): Use move_register rather than macro_build for moves.
(mips16_macro): Likewise.
(macro): Likewise.  Handle M_MOVE.

* mips-dis.c (print_insn_mips): Remove OPCODE_IS_MEMBER's gp32
argument.
* mips-opc.c (G6): Undefine.
(mips_builtin_opcodes): Remove gp32 entry for "move".  Add macro
as the first "move" alternative.

* opcode/mips.h (INSN_GP32): Remove.
(OPCODE_IS_MEMBER): Remove gp32 parameter.
(M_MOVE): New macro identifier.