* binutils-all/ar.exp: Remove stray semicolons.
* binutils-all/dlltool.exp: Likewise.
* binutils-all/objcopy.exp: Likewise.
* binutils-all/readelf.exp: Likewise.
* binutils-all/windres/windres.exp: Likewise.
* lib/utils-lib.exp: Likewise.

* ld-sh/sh64/relax.exp: Remove stray semicolons.
* ld-sh/sh64/relfail.exp: Likewise.
* lib/ld-lib.exp: Likewise.

* gas/h8300/t03_add.exp: Remove stray semicolons.
* gas/h8300/t04_sub.exp: Likewise.
* gas/h8300/t05_cmp.exp: Likewise.
* gas/h8300/t08_or.exp: Likewise.
* gas/h8300/t09_xor.exp: Likewise.
* gas/h8300/t10_and.exp: Likewise.
* gas/hppa/reloc/reloc.exp: Likewise.
* gas/hppa/unsorted/unsorted.exp: Likewise.
* gas/i386/i386.exp: Likewise.
* gas/m68hc11/m68hc11.exp: Likewise.
* gas/mips/mips.exp: Likewise.
* gas/sparc/sparc.exp: Likewise.
* lib/gas-defs.exp: Likewise.

2004-05-11  Maciej W. Rozycki  <macro@ds2.pg.gda.pl>

        * mips/interp.c (decode_coproc): Sign-extend the address retrieved
        from COP0_BADVADDR.
        * mips/sim-main.h (COP0_BADVADDR): Remove a cast.

2004-05-11  Andrew Cagney  <cagney@redhat.com>

* infrun.c (step_over_function): Delete function.
(handle_step_into_function): Use insert_step_resume_breakpoint.
(insert_step_resume_breakpoint): Fix assertion.

(defaulted_query): Just use the normal query input type when printing the
annotations.

*** empty log message ***

bfd/
* elflink.c (elf_bfd_final_link): Don't output STT_SECTION symbol
into .dynsym if elf_section_data (sec)->dynindx <= 0.
Adjust counting of last_local.
(_bfd_elf_link_renumber_dynsyms): Don't assign dynindx to sections
other than SHT_PROGBITS/SHT_NOBITS and neither for .got/.got.plt/.plt
created by the linker nor !SHF_ALLOC.

* elf32-i386.c (elf_i386_finish_dynamic_sections): Point
DT_PLTGOT to the start of the .got.plt section instead of the
.got output section.  Set sh_entsize for .got section in addition
to .got.plt.
(elf_i386_relocate_section): Don't assume _GLOBAL_OFFSET_TABLE_
is at sgot->output_section->vma.
* elf64-x86-64.c (elf64_x86_64_finish_dynamic_sections): Point
DT_PLTGOT to the start of the .got.plt section instead of the
.got output section.
(elf64_x86_64_relocate_section): Don't assume _GLOBAL_OFFSET_TABLE_
is at sgot->output_section->vma.  Set sh_entsize for .got section
in addition to .got.plt.
* elf.c (_bfd_elf_print_private_bfd_data): Handle PT_GNU_RELRO.
(bfd_section_from_phdr): Likewise.
(map_sections_to_segments): Likewise.
(assign_file_positions_for_segments): Likewise.
(get_program_header_size): Likewise.
* elflink.c (bfd_elf_size_dynamic_sections): Set
elf_tdata (output_bfd)->relro from info->relro.
* elf-bfd.h (struct elf_obj_tdata): Add relro field.
include/
* bfdlink.h (struct bfd_link_info): Add relro, relro_start and
relro_end fields.
* elf/common.h (PT_GNU_EH_FRAME, PT_GNU_STACK): Add comments.
(PT_GNU_RELRO): Define.
binutils/
* readelf.c (get_segment_type): Handle PT_GNU_RELRO.
ld/
* genscripts.sh: Generate -z combreloc -z now -z relro scripts
for binaries, -shared and -pie.
* emulparams/elf_i386.sh (SEPARATE_GOTPLT): Set.
* emulparams/elf_x86_64.sh (SEPARATE_GOTPLT): Set.
* emulparams/elf32ppc.sh (OTHER_READWRITE_SECTIONS): Rename to...
(OTHER_RELRO_SECTIONS): ... this.
* ldlex.l (DATA_SEGMENT_RELRO_END): Add.
* emultempl/elf32.em (gld${EMULATION_NAME}_handle_option): Handle
-z relro and -z norelro.
(gld${EMULATION_NAME}_list_options): Add it to usage.
(gld${EMULATION_NAME}_get_script): Return -z combreloc -z now
-z relro scripts when appropriate.
* scripttempl/elf.sc: Unset SEPARATE_GOTPLT if RELRO_NOW is set.
Create separate .got.plt section if SEPARATE_GOTPLT.
Move sections which are only written during relocation handling
to the beginning of RW segment.  If NO_SMALL_DATA, move .got
before .data.  Add DATA_SEGMENT_RELRO_END directive.
Include OTHER_RELRO_SECTIONS.
* ldgram.y (DATA_SEGMENT_RELRO_END): Add.
* ldexp.c (exp_print_token): Handle DATA_SEGMENT_RELRO_END.
(fold_unary): Likewise.
(fold_binary): Handle -z relro.
* ldexp.h (struct exp_data_seg): Add exp_dataseg_relro_seen and
exp_dataseg_relro_adjust phases.  Add relro_end field.
* ldmain.c (main): Initialize link_info.relro to FALSE.
* ldlang.c (lang_size_sections): Handle -z relro.
ld/testsuite/
* ld-i386/tlspic.rd: Adjust for section reordering changes
and removal of unneeded STT_SECTION symbols from .dynsym.
* ld-i386/tlspic.dd: Likewise.
* ld-i386/tlspic.sd: Likewise.
* ld-i386/tlsbin.rd: Likewise.
* ld-i386/tlsbinpic.s: Likewise.
* ld-i386/tlsbin.dd: Likewise.
* ld-i386/tlsbin.sd: Likewise.
* ld-i386/tlsnopic.rd: Likewise.
* ld-i386/tlsnopic1.s: Likewise.
* ld-i386/combreloc.d: Likewise.
* ld-i386/tlsnopic.dd: Likewise.
* ld-i386/tlsnopic.sd: Likewise.
* ld-x86-64/tlspic.rd: Likewise.
* ld-x86-64/tlspic.dd: Likewise.
* ld-x86-64/tlsbin.dd: Likewise.
* ld-x86-64/tlspic.sd: Likewise.
* ld-x86-64/tlsbin.sd: Likewise.
* ld-x86-64/tlspic.td: Likewise.
* ld-x86-64/tlsbin.td: Likewise.
* ld-x86-64/tlsbin.rd: Likewise.
* ld-s390/tlspic1.s: Likewise.
* ld-s390/tlsbinpic.s: Likewise.
* ld-s390/tlspic.rd: Likewise.
* ld-s390/tlsbin.rd: Likewise.
* ld-s390/tlspic.dd: Likewise.
* ld-s390/tlsbin.dd: Likewise.
* ld-s390/tlsbin.sd: Likewise.
* ld-s390/tlsbin.td: Likewise.
* ld-s390/tlspic.sd: Likewise.
* ld-s390/tlspic.td: Likewise.
* ld-s390/tlspic1_64.s: Likewise.
* ld-s390/tlsbinpic_64.s: Likewise.
* ld-s390/tlspic_64.rd: Likewise.
* ld-s390/tlsbin_64.rd: Likewise.
* ld-s390/tlspic_64.dd: Likewise.
* ld-s390/tlsbin_64.dd: Likewise.
* ld-s390/tlspic_64.sd: Likewise.
* ld-s390/tlspic_64.td: Likewise.
* ld-s390/tlsbin_64.td: Likewise.
* ld-s390/tlsbin_64.sd: Likewise.
* ld-powerpc/tlsexe32.r: Likewise.
* ld-powerpc/tlsso32.r: Likewise.
* ld-powerpc/tlsso32.d: Likewise.
* ld-powerpc/tlsso32.g: Likewise.
* ld-powerpc/tlsso32.t: Likewise.
* ld-powerpc/tlsexe.r: Likewise.
* ld-powerpc/tlsso.r: Likewise.
* ld-powerpc/tlsso.g: Likewise.
* ld-powerpc/tlsexetoc.r: Likewise.
* ld-powerpc/tlstocso.r: Likewise.
* ld-powerpc/tlstocso.g: Likewise.
* ld-ia64/tlspic.rd: Likewise.
* ld-ia64/tlspic.dd: Likewise.
* ld-ia64/tlspic.sd: Likewise.
* ld-ia64/tlspic.td: Likewise.
* ld-ia64/tlsbin.rd: Likewise.
* ld-ia64/tlsbin.sd: Likewise.
* ld-ia64/tlsbin.td: Likewise.
* ld-elfvsb/elfvsb.exp: XFAIL non-PIC load offset tests on s390x.
* ld-shared/shared.exp: Likewise.

2004-05-11  Andrew Cagney  <cagney@redhat.com>

* infrun.c (handle_inferior_event): Simplify code handling
step-into or return-from a signal trampoline.

Index: testsuite/ChangeLog
2004-05-11  Andrew Cagney  <cagney@redhat.com>

* gdb.base/sigstep.exp (advancei): Update KFAILs.  gdb/1613 is
fixed but revealed gdb/1639.

2004-05-11  Andrew Cagney  <cagney@redhat.com>

* ppcnbsd-tdep.c: Include "gdb_assert.h".
* Makefile.in (ppcnbsd-tdep.o): Update dependencies.

Apply H.J.'s patch to fix label arithmetic when multiple same-name sections are involved

* ppc-bdm.c (bdm_ppc_fetch_registers): Replace utterly broken code
with a call to internal_error.

* ppc-tdep.h (struct gdbarch_tdep): Change definition of
ppc_fp0_regnum and ppc_fpscr_regnum: if they are -1, then this
processor variant lacks those registers.
(ppc_floating_point_unit_p): Change description to make it clear
that this returns info about the ISA, not the ABI.
* rs6000-tdep.c (ppc_floating_point_unit_p): Decide whether to
return true or false by checking tdep->ppc_fp0_regnum and
tdep->ppc_fpscr_regnum.  The original code replicated the BFD
arch/mach switching done in rs6000_gdbarch_init; it's better to
keep that logic there, and just check the results here.
(rs6000_gdbarch_init): On the E500, set tdep->ppc_fp0_regnum and
tdep->ppc_fpscr_regnum to -1 to indicate that we have no
floating-point registers.
(ppc_supply_fpregset, ppc_collect_fpregset)
(rs6000_push_dummy_call, rs6000_extract_return_value)
(rs6000_store_return_value): Assert that we have floating-point
registers.
(rs6000_dwarf2_stab_reg_to_regnum): Add FIXME.
(rs6000_frame_cache): Don't note the locations at which
floating-point registers were saved if we have no fprs.
* aix-thread.c (supply_fprs, fill_fprs): Assert that we have FP
registers.
(fetch_regs_user_thread, fetch_regs_kernel_thread)
(store_regs_user_thread, store_regs_kernel_thread): Only call
supply_fprs / fill_fprs if we actually have floating-point
registers.
(special_register_p): Check ppc_fpscr_regnum before matching
against it.
(supply_sprs64, supply_sprs32, fill_sprs64, fill_sprs32): Don't
supply / collect fpscr if we don't have it.
* ppc-bdm.c: #include "gdb_assert.h".
(bdm_ppc_fetch_registers, bdm_ppc_store_registers): Assert that we
have floating-point registers, since I can't test this code on
FP-free systems to adapt it.
* ppc-linux-nat.c (ppc_register_u_addr): Don't match against the
fpscr and floating point register numbers if they don't exist.
(fetch_register): Assert that we have floating-point registers
before we reach the code that handles them.
(store_register): Same.  And use tdep instead of calling
gdbarch_tdep again.
(fill_fpregset): Don't try to collect FP registers and fpscr if we
don't have them.
(ppc_linux_sigtramp_cache): Don't record the saved locations of
fprs and fpscr if we don't have them.
(ppc_linux_supply_fpregset): Don't supply fp regs and fpscr if we
don't have them.
* ppcnbsd-nat.c: #include "gdb_assert.h".
(getfpregs_supplies): Assert that we have floating-point registers.
* ppcnbsd-tdep.c (ppcnbsd_supply_fpreg, ppcnbsd_fill_fpreg): Same.
* ppcobsd-tdep.c: #include "gdb_assert.h".
(ppcobsd_supply_gregset, ppcobsd_collect_gregset): Assert that we
have floating-point registers.
* rs6000-nat.c (regmap): Don't match against the fpscr and
floating point register numbers if they don't exist.
(fetch_inferior_registers, store_inferior_registers,
fetch_core_registers): Only fetch / store / supply the
floating-point registers and the fpscr if we have them.
* Makefile.in (ppc-bdm.o, ppc-linux-nat.o, ppcnbsd-nat.o)
(ppcobsd-tdep.o): Update dependencies.

* configure.in (sim_fpu_cflags): Add -I../common.
* configure: Regenerated.

daily update

2004-05-10  Andrew Cagney  <cagney@redhat.com>

* infrun.c (through_sigtramp_breakpoint): Delete variable.
(context_switch): Do not switch through_sigtramp_breakpoint.
* gdbthread.h (save_infrun_state, load_infrun_state)
(struct thread_info): Delete through_sigtramp_breakpoint parameter
and structure member.
* thread.c (load_infrun_state, save_infrun_state): Update.

* m68kbsd-tdep.c (m68kbsd_elf_init_abi): Call m68k_svr4_init_abi.

* acinclude.m4: Quote macro names being defined by AC_DEFUN
throughout.
* aclocal.m4: Regenerate.

Back out change.  The NetBSD changes need Jason Thorpe's approval, but
he hasn't reviewed it yet.

* alphanbsd-tdep.c: Fix include of "gdb_string.h".

2004-05-10  Andrew Cagney  <cagney@redhat.com>

* infrun.c (check_sigtramp2): Delete function.
(handle_inferior_event): When single stepping, and taking a
signal, set a breakpoint at the signal return address.  Delete
redundant calls to check_sigtramp2.
(insert_step_resume_breakpoint): New function.
(through_sigtramp_breakpoint, handle_inferior_event)
(follow_exec, wait_for_inferior, fetch_inferior_event)
(currently_stepping, keep_going): Delete most uses of
through_sigtramp_breakpoint, not that it should be deleted.
(delete_breakpoint_current_contents): Delete function.

Index: testsuite/ChangeLog
2004-05-10  Andrew Cagney  <cagney@redhat.com>

* gdb.base/signals.exp (signal_tests_1): Simplify "continue to
func1" and "next to 2nd alarm", kernel bug avoided.

2004-05-10  Randolph Chung  <tausq@debian.org>

* config/pa/linux.mt (TM_FILE): Use new tm-linux.h for hppa-linux
target.
* config/pa/tm-linux.h: New file.

* cli/cli-setshow.c (cmd_show_list): Check for "show" in the middle
of prefixes.  Don't print the prefix twice in the CLI.

* ppc-tdep.h (struct gdbarch_tdep): Change definition of
ppc_fp0_regnum and ppc_fpscr_regnum: if they are -1, then this
processor variant lacks those registers.
(ppc_floating_point_unit_p): Change description to make it clear
that this returns info about the ISA, not the ABI.
* rs6000-tdep.c (ppc_floating_point_unit_p): Decide whether to
return true or false by checking tdep->ppc_fp0_regnum and
tdep->ppc_fpscr_regnum.  The original code replicated the BFD
arch/mach switching done in rs6000_gdbarch_init; it's better to
keep that logic there, and just check the results here.
(rs6000_gdbarch_init): On the E500, set tdep->ppc_fp0_regnum and
tdep->ppc_fpscr_regnum to -1 to indicate that we have no
floating-point registers.
(ppc_supply_fpregset, ppc_collect_fpregset)
(rs6000_push_dummy_call, rs6000_extract_return_value)
(rs6000_store_return_value): Assert that we have floating-point
registers.
(rs6000_dwarf2_stab_reg_to_regnum): Add FIXME.
(rs6000_frame_cache): Don't note the locations at which
floating-point registers were saved if we have no fprs.
* aix-thread.c (supply_fprs, fill_fprs): Assert that we have FP
registers.
(fetch_regs_user_thread, fetch_regs_kernel_thread)
(store_regs_user_thread, store_regs_kernel_thread): Only call
supply_fprs / fill_fprs if we actually have floating-point
registers.
(special_register_p): Check ppc_fpscr_regnum before matching
against it.
(supply_sprs64, supply_sprs32, fill_sprs64, fill_sprs32): Don't
supply / collect fpscr if we don't have it.
* ppc-bdm.c: #include "gdb_assert.h".
(bdm_ppc_fetch_registers, bdm_ppc_store_registers): Assert that we
have floating-point registers, since I can't test this code on
FP-free systems to adapt it.
* ppc-linux-nat.c (ppc_register_u_addr): Don't match against the
fpscr and floating point register numbers if they don't exist.
(fetch_register): Assert that we have floating-point registers
before we reach the code that handles them.
(store_register): Same.  And use tdep instead of calling
gdbarch_tdep again.
(fill_fpregset): Don't try to collect FP registers and fpscr if we
don't have them.
(ppc_linux_sigtramp_cache): Don't record the saved locations of
fprs and fpscr if we don't have them.
(ppc_linux_supply_fpregset): Don't supply fp regs and fpscr if we
don't have them.
* ppcnbsd-nat.c: #include "gdb_assert.h".
(getfpregs_supplies): Assert that we have floating-point registers.
* ppcnbsd-tdep.c (ppcnbsd_supply_fpreg, ppcnbsd_fill_fpreg): Same.
* ppcobsd-tdep.c: #include "gdb_assert.h".
(ppcobsd_supply_gregset, ppcobsd_collect_gregset): Assert that we
have floating-point registers.
* rs6000-nat.c (regmap): Don't match against the fpscr and
floating point register numbers if they don't exist.
(fetch_inferior_registers, store_inferior_registers,
fetch_core_registers): Only fetch / store / supply the
floating-point registers and the fpscr if we have them.
* Makefile.in (ppc-bdm.o, ppc-linux-nat.o, ppcnbsd-nat.o)
(ppcobsd-tdep.o): Update dependencies.

* rs6000-tdep.c (ppc_collect_gregset): When regnum == -1, do
collect all the gprs.

* rs6000-tdep.c (ppc_collect_gregset): Correct off-by-one error in
loop collecting gprs.

PR external/1568
* gdb.base/bigcore.exp: Check the size of the dumped core file.
XFAIL if it is smaller than bytes_allocated.
* gdb.base/bigcore.c (bytes_allocated): Make static and unsigned.
(main): Make chunks_allocated unsigned.  Correct comment.

* callback.c: Update copyright dates.
* run.c: Likewise.
* sim-basics.h: Likewise.
* sim-load.c: Likewise.
* syscall.c: Likewise.
From Maciej W. Rozycki  <macro@ds2.pg.gda.pl>
* callback.c: Include cconfig.h instead of config.h.
* run.c: Likewise.
* sim-basics.h: Likewise.
* sim-load.c: Likewise.
* syscall.c: Likewise.

2004-05-09  Andrew Cagney  <cagney@redhat.com>

* remote-vx.c (net_step): Delete step-range code.
* remote.c (remote_resume, init_all_packet_configs)
(set_remote_protocol_E_packet_cmd)
(show_remote_protocol_E_packet_cmd)
(remote_protocol_E, show_remote_cmd, _initialize_remote)
(remote_protocol_e, set_remote_protocol_e_packet_cmd)
(show_remote_protocol_e_packet_cmd): Ditto.

2004-05-10  John Paul Wallington  <jpw@gnu.org>

* ld-mmix/bspec2.d: Update sh_info to decimal.
* ld-mmix/local1.d: Likewise.
* ld-mmix/local3.d: Likewise.
* ld-mmix/local5.d: Likewise.
* ld-mmix/local7.d: Likewise.

daily update

* hppa-tdep.h (enum hppa_regnum): Name enum, add HPPA_R1_REGNUM.

* hppa-tdep.h: Update copyright year.
(struct trad_frame_saved_reg): Add opaque declaration.

* amd64-tdep.c (amd64_return_value): Implement
RETURN_VALUE_ABI_RETURNS_ADDRESS.

* m68k-tdep.c (m68k_convert_register_p, m68k_register_to_value)
(m68k_value_to_register): New functions.
(m68k_gdbarch_init): Set convert_register_p, register_to_value and
value_to_register.

* defs.h (enum return_value_convention): Add
RETURN_VALUE_ABI_RETURNS_ADDRESS and
RETURN_VALUE_ABI_PRESERVES_ADDRESS.
* infcmd.c (legacy_return_value): New function.
(print_return_value): Rwerite to implement
RETURN_VALUE_ABI_RETURNS_ADDRESS.
* values.c (using_struct_return): Check for inequality to
RETURN_VALUE_REGISTER_CONVENTION instead of equality to
RETURN_VALUE_STRUCT_CONVENTION.
* i386-tdep.c (i386_return_value): Implement
RETURN_VALUE_ABI_RETURNS_ADDRESS.

* vax-tdep.c: Tweak comments.  Reorder include files.  Don't
include "symtab.h", "opcode/vax.h" and "inferior.h".
(vax_skip_prologue): Replace calls to read_memory_integer by calls
to read_memory_unsigned_integer.
(vax_gdbarch_init): Reorder.
(_initialize_vax_tdep): Spell out prototype.
* Makefile.in (vax-tdep.o): Update dependencies.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* infrun.c (resume): Delete call to DO_DEFERRED_STORES.
* target.c (target_detach, target_disconnect): Ditto.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbint.texinfo (Target Architecture Definition): Delete
description of DO_DEFERRED_STORES.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* infcall.c (call_function_by_hand): When no
DEPRECATED_SAVE_DUMMY_FRAME_TOS call generic_save_dummy_frame_tos.
* rs6000-tdep.c (rs6000_gdbarch_init): Do not set
DEPRECATED_SAVE_DUMMY_FRAME_TOS.
* mn10300-tdep.c (mn10300_gdbarch_init): Ditto.
* mcore-tdep.c (mcore_gdbarch_init): Ditto.

* scripttempl/avr.sc: Do not set LMA for "eeprom" section.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbint.texinfo (Target Architecture Definition): Delete
references to DEPRECATED_FIX_CALL_DUMMY.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbarch.sh (DEPRECATED_FIX_CALL_DUMMY): Delete.
* gdbarch.h, gdbarch.c: Re-generate.
* ns32k-tdep.c (ns32k_gdbarch_init): Do not set
deprecated_fix_call_dummy.
* v850-tdep.c (v850_gdbarch_init): Ditto.
* infcall.c (call_function_by_hand): Do not call
DEPRECATED_FIX_CALL_DUMMY.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbarch.sh (DEPRECATED_CALL_DUMMY_WORDS)
(DEPRECATED_SIZEOF_CALL_DUMMY_WORDS): Delete.
* gdbarch.h, gdbarch.c: Re-generate.
* arm-linux-tdep.c (arm_linux_call_dummy_words)
(arm_linux_init_abi): Do not set deprecated_call_dummy_words or
deprecated_sizeof_call_dummy_words.
* arm-tdep.c (arm_call_dummy_words, arm_gdbarch_init): Ditto.
* xstormy16-tdep.c (xstormy16_gdbarch_init): Ditto.
* v850-tdep.c (v850_gdbarch_init): Ditto.
* sh64-tdep.c (sh64_gdbarch_init): Ditto.
* ns32k-tdep.c (ns32k_call_dummy_words, ns32k_gdbarch_init): Ditto.
* mn10300-tdep.c (mn10300_gdbarch_init): Ditto.
* mcore-tdep.c (mcore_gdbarch_init): Ditto.
* arch-utils.h (legacy_call_dummy_words)
(legacy_sizeof_call_dummy_words): Delete declarations.
* arch-utils.c (legacy_call_dummy_words)
(legacy_sizeof_call_dummy_words): Delete variables.

Index: doc/ChangeLog
2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbint.texinfo (Target Architecture Definition): Delete
description of DEPRECATED_CALL_DUMMY_WORDS,
DEPRECATED_SIZEOF_CALL_DUMMY_WORDS, and CALL_DUMMY.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbarch.sh (DEPRECATED_PUSH_DUMMY_FRAME): Delete.
* gdbarch.h, gdbarch.c: Re-generate.
* inferior.h (PUSH_DUMMY_FRAME): Delete macro.
* infcall.c (call_function_by_hand): Eliminate call to
DEPRECATED_PUSH_DUMMY_FRAME.
* ns32k-tdep.c (ns32k_gdbarch_init): Do not set deprecated
push_dummy_frame.
(ns32k_push_dummy_frame): Delete function.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* ns32k-tdep.c (ns32k_gdbarch_init): Do not set
deprecated_call_dummy_breakpoint_offset or
deprecated_call_dummy_start_offset.
* dummy-frame.c, arm-tdep.c: Update comments.
* gdbarch.sh (DEPRECATED_CALL_DUMMY_BREAKPOINT_OFFSET)
(DEPRECATED_CALL_DUMMY_START_OFFSET): Delete.
* gdbarch.h, gdbarch.c: Re-generate.

Index: doc/ChangeLog
2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbint.texinfo (Target Architecture Definition): Delete
reference to DEPRECATED_CALL_DUMMY_BREAKPOINT_OFFSET.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* infcall.c (legacy_push_dummy_code): Delete function.
(push_dummy_code): Delete call.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* gdbarch.sh (DEPRECATED_USE_GENERIC_DUMMY_FRAMES): Delete.
* gdbarch.h, gdbarch.c: Re-generate.
* blockframe.c (legacy_frame_chain_valid): Simplify by eliminating
DEPRECATED_USE_GENERIC_DUMMY_FRAMES.
* infcall.c (legacy_push_dummy_code, call_function_by_hand): Ditto.
* frame.c (frame_type_from_pc, legacy_get_prev_frame): Ditto.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* breakpoint.h (deprecated_frame_in_dummy): Delete declaration.
(struct frame_info): Delete opaque declaration.
* breakpoint.c (deprecated_frame_in_dummy): Delete function.
* frame.c (get_frame_type): Delete call.

2004-05-08  Andrew Cagney  <cagney@redhat.com>

* ns32k-tdep.c (ns32k_gdbarch_init): Do not set deprecated
pc_in_call_dummy.
* inferior.h (deprecated_pc_in_call_dummy_on_stack): Delete.
* blockframe.c (deprecated_pc_in_call_dummy_on_stack): Delete.

* alphanbsd-tdep.c: Include "gdb_string.h".
* armnbsd-tdep.c: Update copyright year.
Include "gdb_string.h".
* Makefile.in (alphanbsd-tdep.o, armnbsd-tdep.o): Update
dependencies.

* delta68-nat.c: Remove file.
* m68k-tdep.c (delta68_in_sigtramp, delta68_frame_args_address)
(delta68_frame_saved_pc, delta68_frame_num_args): Remove functions.
* Makefile.in (ALLDEPFILES): Remove delta68-nat.c.
(delta68-nat.o): Remove dependency.

* m68k-tdep.c (m68k_frameless_function_invocation): Remove
function.
(m68k_gdbarch_init): Don't set
deprecated_frameless_function_invocation.

* elf32-frv.c (_frvfdpic_add_dyn_reloc): Don't warn when we get
a zero symndx for which we hadn't accounted a dynamic relocation.
(_frvfdpic_add_rofixup): Likewise.

2004-05-07  Randolph Chung  <tausq@debian.org>

* hppa-tdep.c (skip_prologue_hard_way, hppa_frame_cache): Use
read_memory_nobpt for code reading when doing frame unwinding.
* hppa-linux-tdep.c (insns_match_pattern): Likewise.

2004-05-07  Randolph Chung  <tausq@debian.org>

* hppa-linux-tdep.c (hppa_linux_sigtramp_find_sigcontext): Pass in pc
instead of sp, handle sigaltstack case.
(hppa_linux_sigtramp_frame_unwind_cache): Adjust calls to
hppa_linux_sigtramp_find_sigcontext, and set base to the frame of the
signal handler and not that of the caller.
(hppa_linux_sigtramp_unwind_sniffer): Adjust calls to
hppa_linux_sigtramp_find_sigcontext.

2004-05-07  Randolph Chung  <tausq@debian.org>

* Makefile.in (ALLDEPFILES): Add hppa-linux-tdep.c and
hppa-linux-nat.c.
(hppa-linux-tdep.o, hppa-linux-nat.o): Define rules.
* hppa-linux-nat.c (fetch_register, store_register): Define.
(fetch_inferior_registers, store_inferior_registers): Define.
(supply_gregset): Add comment and clean up some unneeded code.
(fill_gregset): Add comment and fix so that it saves registers in
the right place in the gregset.
(fill_fpregset): Fix buffer overrun problem.
* config/pa/nm-linux.h (FETCH_INFERIOR_REGISTERS): Define.

2004-05-07  Randolph Chung  <tausq@debian.org>

* hppa-tdep.c (hppa_frame_prev_register_helper): New function to
do common handling of the pcoqt register.
(hppa_frame_prev_register, hppa_stub_frame_prev_register): Convert
to use helper function.
(hppa_frame_unwind_sniffer): Only use if unwind entry is present.
(hppa_fallback_frame_cache, hppa_fallback_frame_this_id)
(hppa_fallback_frame_prev_register, hppa_fallback_frame_unwind): New
generic fallback unwinder when all else fails.
(hppa_gdbarch_init): Add fallback sniffer.
* hppa-tdep.h (hppa_frame_prev_register_helper): Prototype.
* hppa-linux-tdep.c (hppa_linux_sigtramp_frame_prev_register): Convert
to use helper function.

    * config/djgpp/fnchange.lst: Add translation rules for
        gdb.arch/powerpc-aix-prologue.c and gdb.arch/powerpc-aix-prologue.exp.

    * gdb.arch/powerpc-aix-prologue.c: New file.
        * gdb.arch/powerpc-aix-prologue.exp: New file.

daily update

* frame.c (get_prev_frame_1): Work around GCC bug.
* frame.h: Add comment about GCC bug.

2004-05-07  Andrew Cagney  <cagney@redhat.com>

* observer.c: Include "commands.h" and "gdbcmd.h".
(observer_debug, _initialize_observer): Add the
command "set/show debug observer".
* observer.sh: When observer debugging, log notify calls.
* Makefile.in (observer.o): Update dependencies.

Index: doc/ChangeLog
2004-05-07  Andrew Cagney  <cagney@redhat.com>

* observer.texi (GDB Observers): Add "Debugging" section.  Include
cross reference to "set/show debug observer".
* gdb.texinfo (Debugging Output): Document "set/show debug
observer".

2004-05-07  Jason Molenda  (jmolenda@apple.com)

        * cli/cli-decode.c: Fix "SHOW_DOR" comment typeos.

* aix-thread.c (fetch_regs_kernel_thread, fill_gprs64,
fill_gprs32, store_regs_kernel_thread): Use tdep->ppc_gp0_regnum
to recognize gpr regnums, instead of assuming that
tdep->ppc_fp0_regnum marks the end of the general-purpose
registers.

* ppc-sysv-tdep.c (ppc64_sysv_abi_return_value): Assert that
the given architecture has floating-point registers.

* config/tc-mips.c (append_insn, mips_emit_delays): Extend -mfix-vr4120
to cope with VR4181A errata MD(1) and MD(4).

2004-05-07  H.J. Lu  <hongjiu.lu@intel.com>

* readelf.c (section_groups): New.
(group_count): New.
(section_headers_groups): New.
(process_section_groups): Populate group_count, section_groups
and section_headers_groups.
(process_unwind): Support section group.
(process_object): Always call process_section_groups. Free
section_groups and section_headers_groups.

* ldlang.c (print_padding_statement): Cast size_t to bfd_vma
before printing it with %W.

* emultempl/pe.em (real_flags): New static.
(OPTION_LARGE_ADDRESS_AWARE): New define.
(gld${EMULATION_NAME}_add_options): Add --large-address-aware option.
(gld_${EMULATION_NAME}_list_options): Likewise.
(gld${EMULATION_NAME}_handle_option): Likewise.
(gld_${EMULATION_NAME}_after_open): Pass real_flags to PE private data.
* ld.texinfo: Document it.
* NEWS: Mention it.

* coffcode.h (coff_write_object_contents) [COFF_IMAGE_WITH_PE]:
Propagate IMAGE_FILE_LARGE_ADDRESS_AWARE.
* peXXigen.c (_bfd_XX_print_private_bfd_data_common): Recognize
IMAGE_FILE_LARGE_ADDRESS_AWARE.  Use PE defines.

2004-05-07  Andrew Cagney  <cagney@redhat.com>

* gdbarch.sh (DEPRECATED_USE_GENERIC_DUMMY_FRAMES): Replace with
macro defined to 1.
* gdbarch.h, gdbarch.c: Re-generate.
* ns32k-tdep.c (ns32k_gdbarch_init): Do not set
deprecated_use_generic_dummy_frames.

* NEWS: Mention DWARF 2 support for i386pe.

* NEWS: Mention .secrel32 for pe[i]-i386.

* NEWS: Mention Cygwin DWARF 2 support.

2004-05-07  Andrew Cagney  <cagney@redhat.com>

* objfiles.h (struct entry_info): Delete unused fields
deprecated_entry_file_lowpc and deprecated_entry_file_highpc.
* objfiles.c (init_entry_point_info, objfile_relocate): Update.
* mipsread.c (mipscoff_symfile_read): Update.
* mdebugread.c (parse_partial_symbols): Update.
* dwarfread.c (read_file_scope): Update.
* dwarf2read.c (read_file_scope): Update.
* dbxread.c (read_dbx_symtab, read_dbx_symtab): Update.
* coffread.c (complete_symtab): Update.

* infcall.c (_initialize_infcall): Fix typo.

2004-05-07  Andrew Cagney  <cagney@redhat.com>

* cli/cli-decode.c (add_setshow_zinteger_cmd): New function.
* command.h (add_setshow_zinteger_cmd): Declare.

* gdb.base/lineinc.exp, gdb.base/lineinc1.h, gdb.base/lineinc2.h,
gdb.base/lineinc3.h, gdb.base/lineinc.c: New tests.

* elf32-frv.c (elf32_frvfdpic_modify_segment_map): Return
immediately if there's no link info.
(elf32_frvfdpic_copy_private_bfd_data): New.
(bfd_elf32_bfd_copy_private_bfd_data): Use it for frvfdpic.

* config/tc-frv.h (MAX_MEM_FOR_RS_ALIGN_CODE): New.
(HANDLE_ALIGN): New.

* Makefile.am (DIST_SUBDIRS): Define.
* aclocal.m4: Regenerate with automake 1.8.4.
* Makefile.in: Likewise.
* doc/Makefile.in: Likewise.

2004-05-06  Randolph Chung  <tausq@debian.org>

* config/pa/tm-hppa.h (RO_REGNUM, FLAGS_REGNUM, RP_REGNUM)
(HPPA_FP_REGNUM, HPPA_SP_REGNUM, SAR_REGNUM, IPSW_REGNUM)
(PCOQ_HEAD_REGNUM, PCSQ_HEAD_REGNUM, PCOQ_TAIL_REGNUM, PCSQ_TAIL_REGNUM)
(EIEM_REGNUM, IIR_REGNUM, ISR_REGNUM, IOR_REGNUM, SR4_REGNUM)
(RCR_REGNUM, PID0_REGNUM, PID1_REGNUM, PID2_REGNUM, PID3_REGNUM)
(CCR_REGNUM, TR0_REGNUM, CR27_REGNUM, HPPA_FP0_REGNUM, FP4_REGNUM)
(ARG0_REGNUM, ARG1_REGNUM, ARG2_REGNUM, ARG3_REGNUM): Move *_REGNUM
definitions ...
* hppa-tdep.h: ... to here, with HPPA_ prefix.
* Makefile.in (hppah-nat.o): Add $(hppa_tdep_h)
* hppa-hpux-tdep.c (hppa32_hpux_frame_find_saved_regs_in_sig)
(hppa64_hpux_frame_find_saved_regs_in_sig)
(child_get_current_exception_event): Add HPPA_ prefix to *_REGNUM usage.
* hppa-linux-nat.c (GR_REGNUM, TR_REGNUM, greg_map): Likewise.
* hppa-linux-tdep.c (hppa_dwarf_reg_to_regnum)
(hppa_linux_target_write_pc, hppa_linux_sigtramp_frame_unwind_cache)
(hppa_linux_sigtramp_frame_prev_register): Likewise.
* hppa-tdep.c (hppa32_return_value, hppa64_return_value)
(hppa32_push_dummy_call, hppa64_push_dummy_call, hppa64_frame_align)
(hppa_target_read_pc, hppa_target_write_pc, hppa_frame_cache)
(hppa_frame_prev_register, hppa_stub_frame_unwind_cache)
(hppa_stub_frame_prev_register, hppa_unwind_dummy_id)
(hppa_skip_permanent_breakpoint, hppa_instruction_nullified)
(hppa32_register_type, hppa_cannot_store_register)
(hppa_fetch_pointer_argument, hppa_pseudo_register_read): Likewise.
* hppah-nat.c (store_inferior_registers, fetch_register): Likewise.
* hpread.c (hpread_process_one_debug_symbol): Likewise.
* pa64solib.c (pa64_solib_have_load_event)
(pa64_solib_have_unload_event, pa64_solib_loaded_library_pathname)
(pa64_solib_unloaded_library_pathname): Likewise.
* somsolib.c (som_solib_have_load_event, som_solib_have_unload_event)
(som_solib_library_pathname): Likewise.

* dwarf2.c (add_line_info): Also set info->filename to NULL if
filename argument is null; do not call strlen on a null pointer.

daily update

* elf32-arm.h (elf32_arm_relocate_section): Remove R_ARM_PLT32
special case.

Patch from David Mosberger-Tang.
* config/tc-ia64.c (dot_serialize): Declare.
(dot_serialize): New function.
(md_pseudo_table): Add ".serialize.data" and
".serialize.instruction" directives.

    * gdb.base/sep.exp: No longer setup_kfail when the program was
        built with dwarf2.

Remove duplicate code and provide a function for generating internally consistent 'value out of range' messages

* emulparams/elf32frvfd.sh (OUTPUT_FORMAT): Switch to new
elf32-frvfdpic.

* configure.in: Set em=linux for frv-*-*linux*.
* configure: Rebuilt.
* config/tc-frv.h (TARGET_FORMAT): Use elf32-frvfdpic if...
(frv_md_fdpic_enabled): New.
* config/tc-frv.c (frv_md_fdpic_enabled): New.
(DEFAULT_FDPIC): New.
(frv_flags): Use DEFAULT_FDPIC.
(frv_pic_flag): Likewise.
(OPTION_NOPIC): New.
(md_longopts): Add -mnopic.
(md_parse_option): Handle it.
(md_show_usage): Add -mfdpic and -mnopic.

* configure.in (bfd_elf32_frvfdpic_vec): New.
* configure: Rebuilt.
* targets.c (bfd_elf32_frvfdpic_vec): New.
* config.bfd: Enable it on frv-*-elf and frv-*-*linux*, as default
on the latter.
* elf32-frv.c: Prefix all identifiers added for FDPIC support with
frvfdpic instead of frv.  Rearrange elf-target macros such that
the FDPIC-specific ones are only defined for this new target
vector.
(bfd_elf32_frvfdpic_vec): Declare.
(IS_FDPIC): New.
(elf32_frv_relocate_section): Use it to enable segment checking
and to control rofixup emission.  Add output section vma to
applied relocation in non-LOAD non-ALLOC sections.  Use
_bfd_error_handler for errors.
(_frv_create_got_section): Create .rel.got and .rofixup only in
FDPIC.  Create non-dynamic _gp at .got+2048 in non-FDPIC, like the
linker script.
(elf32_frvfdpic_size_dynamic_sections): Assume FDPIC.
(elf32_frvfdpic_modify_segment_map): Likewise.
(elf32_frv_finish_dynamic_sections): New, do-nothing.
(elf32_frvfdpic_finish_dynamic_sections): Assume FDPIC.  Improve
error message if we miscompute the rofixup size.
(frvfdpic_elf_use_relative_eh_frame): Assume FDPIC.
(frvfdpic_elf_encode_eh_address): Likewise.
(elf32_frv_check_relocs): Reject FDPIC-only relocs in non-FDPIC.
Record relocs only in FDPIC.  Make sure _gp is defined for GPREL
relocs.  Reject unknown relocation types.
(elf32_frv_object_p): Make sure target vector matches FDPIC bits.
(frv_elf_merge_private_bfd_data): Likewise.
(ELF_MAXPAGESIZE): Revert to 0x1000 for elf32-frv; keep it as
0x4000 for newly-added elf32-frvfdpic.

* ld-frv/fdpic.exp: Restore $LDFLAGS at the end.
* ld-frv/fr450-link.d: Match fdpic as well.

* gas/frv/reloc1.d: Match elf32-frvfdpic as well.
* gas/frv/fr405-insn.d: Likewise.
* gas/frv/fr450-insn.d: Likewise.

* ld-sh/sh64/crange1.rd: Update sh_info to decimal.
* ld-sh/sh64/crange2.rd: Likewise.
* ld-sh/sh64/crange3.rd: Likewise.
* ld-sh/sh64/crange3-cmpct.rd: Likewise.
* ld-sh/sh64/crange3-media.rd: Likewise.
* ld-sh/sh64/crangerel1.rd: Likewise.
* ld-sh/sh64/crangerel2.rd: Likewise.

daily update

Preserve the ignore count when we resolve a pending breakpoint.  Also add
a test for this to pending.exp.

Remove checks of the return value from bfd_cache_ookup()

2004-05-05  Andrew Cagney  <cagney@redhat.com>

* gdbarch.sh (PARM_BOUNDARY): Delete.
gdbarch.h, gdbarch.c: Re-generate.
* valops.c (PARM_BOUNDARY): Delete macro.
(value_push): Delete PARM_BOUNDARY code.
* m68k-tdep.c (m68k_gdbarch_init): Do not set parm_boundary.

PR/136 - Have bfd_cache_lookup_worker() abort rather than return an
(unchecked-for) NULL.

Add support for 521x,5249,547x,548x.

* ppc-opc.c (PPCVEC): Remove PPC_OPCODE_PPC.