Andreas Schwab [Wed, 8 Mar 2006 15:28:17 +0000 (15:28 +0000)]
* ia64-tdep.c (ia64_libunwind_frame_this_id): Adapt uses of
libunwind_frame_prev_register to use a gdb_byte buffer and
extract_unsigned_integer.
(ia64_libunwind_sigtramp_frame_prev_register): Likewise.
* libunwind-frame.c (libunwind_frame_prev_register): Change type
of last argument to `gdb_byte *'
* libunwind-frame.h: Adjust declaration.
bfd/
* configure.in (bfd_elf32_bigarm_vec): Include elf-vxworks.lo.
(bfd_elf32_bigarm_symbian_vec): Likewise.
(bfd_elf32_bigarm_vxworks_vec): Likewise.
(bfd_elf32_littlearm_vec): Likewise.
(bfd_elf32_littlearm_symbian_vec): Likewise.
(bfd_elf32_littlearm_vxworks_vec): Likewise.
* configure: Regenerate.
* elf32-arm.c: Include libiberty.h and elf-vxworks.h.
(RELOC_SECTION, RELOC_SIZE, SWAP_RELOC_IN, SWAP_RELOC_OUT): New macros.
(elf32_arm_vxworks_bed): Add forward declaration.
(elf32_arm_howto_table_1): Fix the masks for R_ASM_ABS12.
(elf32_arm_vxworks_exec_plt0_entry): New table.
(elf32_arm_vxworks_exec_plt_entry): Likewise.
(elf32_arm_vxworks_shared_plt_entry): Likewise.
(elf32_arm_link_hash_table): Add vxworks_p and srelplt2 fields.
(reloc_section_p): New function.
(create_got_section): Use RELOC_SECTION.
(elf32_arm_create_dynamic_sections): Likewise. Call
elf_vxworks_create_dynamic_sections for VxWorks targets.
Choose between the two possible values of plt_header_size
and plt_entry_size.
(elf32_arm_link_hash_table_create): Initialize vxworks_p and srelplt2.
(elf32_arm_abs12_reloc): New function.
(elf32_arm_final_link_relocate): Call it. Allow the creation of
dynamic R_ARM_ABS12 relocs on VxWorks. Use reloc_section_p,
RELOC_SIZE, SWAP_RELOC_OUT and RELOC_SECTION. Initialize the
r_addend fields of relocs. On rela targets, skip any code that
adjusts in-place addends. When using _bfd_link_final_relocate
to perform a final relocation, pass rel->r_addend as the addend
argument.
(elf32_arm_merge_private_bfd_data): If one of the bfds is a VxWorks
object, ignore flags that are not standard on VxWorks.
(elf32_arm_check_relocs): Allow the creation of dynamic R_ARM_ABS12
relocs on VxWorks. Use reloc_section_p.
(elf32_arm_adjust_dynamic_symbol): Use RELOC_SECTION and RELOC_SIZE.
(allocate_dynrelocs): Use RELOC_SIZE. Account for the size of
.rela.plt.unloaded relocs on VxWorks targets.
(elf32_arm_size_dynamic_sections): Use RELOC_SIZE. Check for
.rela.plt.unloaded as well as .rel(a).plt. Add DT_RELA* tags
instead of DT_REL* tags on RELA targets.
(elf32_arm_finish_dynamic_symbol): Use RELOC_SECTION, RELOC_SIZE
and SWAP_RELOC_OUT. Initialize r_addend fields. Handle VxWorks
PLT entries. Do not make _GLOBAL_OFFSET_TABLE_ absolute on VxWorks.
(elf32_arm_finish_dynamic_sections): Use RELOC_SECTION, RELOC_SIZE
and SWAP_RELOC_OUT. Initialize r_addend fields. Handle DT_RELASZ
like DT_RELSZ. Handle the VxWorks form of initial PLT entry.
Correct the .rela.plt.unreloaded symbol indexes.
(elf32_arm_output_symbol_hook): Call the VxWorks version of this
hook on VxWorks targets.
(elf32_arm_vxworks_link_hash_table_create): Set vxworks_p to true.
Minor formatting tweak.
(elf32_arm_vxworks_final_write_processing): New function.
(elf_backend_add_symbol_hook): Override for VxWorks and reset
for Symbian.
(elf_backend_final_write_processing): Likewise.
(elf_backend_emit_relocs): Likewise.
(elf_backend_want_plt_sym): Likewise.
(ELF_MAXPAGESIZE): Likewise.
(elf_backend_may_use_rel_p): Minor formatting tweak.
(elf_backend_may_use_rela_p): Likewise.
(elf_backend_default_use_rela_p): Likewise.
(elf_backend_rela_normal): Likewise.
* Makefile.in (elf32-arm.lo): Depend on elf-vxworks.h.
gas/
* config/tc-arm.c (md_apply_fix): Install a value of zero into a
BFD_RELOC_ARM_OFFSET_IMM field if we're going to generate a RELA
R_ARM_ABS12 reloc.
(tc_gen_reloc): Keep the original fx_offset for RELA pc-relative
relocs, but adjust by md_pcrel_from_section. Create R_ARM_ABS12
relocations for BFD_RELOC_ARM_OFFSET_IMM on RELA targets.
gas/testsuite/
* gas/arm/abs12.s, gas/arm/abs12.d: New test.
* gas/arm/pic.d: Skip for *-*-vxworks*...
* gas/arm/pic_vxworks.d: ...use this version instead.
* gas/arm/unwind_vxworks.d: Fix expected output.
ld/
* emulparams/armelf_vxworks.sh: Include vxworks.sh.
(MAXPAGESIZE): Define.
* emulparams/vxworks.sh: Undefine.
* Makefile.am (earmelf_vxworks.c): Depend on vxworks.sh and vxworks.em.
* Makefile.in: Regenerate.
Alexandre Oliva [Tue, 7 Mar 2006 05:05:29 +0000 (05:05 +0000)]
* gdb.base/prelink.c, gdb.base/prelink-lib.c: Update FSF address.
* gdb.base/prelink.exp: Likewise. Add -q to prelink runs.
Rename executable to prelinkt. Don't re-prelink if we didn't
get a core file. Test prelink -u exit status to tell whether
prelinking failed. Use gdb_test_multiple.
gas:
* config/tc-m68k.c (m68k_extensions): Allow 'float' on both m68k
and cf.
(m68k_ip): <case 'J'> Check we have some control regs.
(md_parse_option): Allow raw arch switch.
(m68k_init_arch): Better detection of arch/cpu mismatch. Detect
whether 68881 or cfloat was meant by -mfloat.
(md_show_usage): Adjust extension display.
(m68k_elf_final_processing): Adjust.
Jim Wilson [Fri, 3 Mar 2006 01:10:42 +0000 (01:10 +0000)]
Fix problem with double-stop-bit after itc.i instruction.
* config/tc-ia64.c (emit_one_bundle): For IA64_OPCODE_LAST, if we
change the template, then clear md.slot[curr].end_of_insn_group.
bfd/
* elf32-ppc.c (is_ppc_elf_target): Return true if the target is
bfd_elf32_powerpc_vxworks_vec.
ld/testsuite/
* ld-i386/vxworks1.ld (.data): New section.
* ld-i386/vxworks1-lib.s: Add a pointer to a local symbol.
* ld-i386/vxworks1-lib.rd: Test for the associated reloc.
* ld-powerpc/vxworks1.ld (.data): New section.
* ld-powerpc/vxworks1-lib.s: Add a pointer to a local symbol.
* ld-powerpc/vxworks1-lib.rd: Test for the associated reloc.
bfd/
* elf32-i386.c (elf_i386_create_dynamic_sections): Use
elf_vxworks_create_dynamic_sections.
(elf_i386_size_dynamic_sections): Remove VxWorks GOT and PLT
symbol handling.
* elf32-ppc.c (ppc_elf_create_dynamic_sections): Use
elf_vxworks_create_dynamic_sections.
(ppc_elf_size_dynamic_sections): Remove VxWorks GOT and PLT
symbol handling.
* elf-vxworks.c (elf_vxworks_create_dynamic_sections): New function.
* elf-vxworks.h (elf_vxworks_create_dynamic_sections): Declare.
ld/testsuite/
* ld-i386/ld-i386/vxworks1-lib.nd: New test.
* ld-i386/i386.exp: Run it.
* ld-powerpc/ld-powerpc/vxworks1-lib.nd: New test.
* ld-powerpc/powerc.exp: Run it.
bfd/
* elf32-ppc.c (ppc_elf_plt_type): New enumeration.
(ppc_elf_link_hash_table): Replace old_got and new_got with
plt_type and can_use_new_plt.
(ppc_elf_create_dynamic_sections): Add SEC_HAS_CONTENTS,
SEC_LOAD and SEC_READONLY to the VxWorks .plt flags.
(ppc_elf_check_relocs): Set can_use_new_plt instead of new_plt.
Move from plt_type == PLT_UNSET to PLT_OLD instead of setting old_plt.
(ppc_elf_select_plt_layout): Move from plt_type == PLT_UNSET to
either plt_type == PLT_OLD or plt_type == PLT_NEW. Assert that
this function should not be called for VxWorks targets.
(ppc_elf_tls_setup): Use plt_type instead of old_got.
(allocate_got): Likewise. Rearrange so that max_before_header
is only used for PLT_OLD and PLT_NEW.
(allocate_dynrelocs): Use plt_type instead of old_got and is_vxworks.
(ppc_elf_size_dynamic_sections): Likewise.
(ppc_elf_relax_section): Likewise.
(ppc_elf_relocate_section): Likewise.
(ppc_elf_finish_dynamic_symbol): Likewise.
(ppc_elf_vxworks_link_hash_table_create): Initialize plt_type.
* hppa-tdep.h (unwind_table_entry): Update field names to match HP
runtime specification.
* hppa-tdep.c (internalize_unwinds, hppa_frame_cache): Likewise.
(unwind_command): Likewise.
* hppa-hpux-tdep.c (initialize_hp_cxx_exception_support): Use
TDEP->is_elf to determine if we are working with a SOM binary.
(null_symtab_and_line): Remove unused variable.
* config/pa/hppa64.mt: Use tm-hppa.h.
* config/pa/tm-hppa64.h: Remove file.
Nick Clifton [Tue, 28 Feb 2006 16:09:01 +0000 (16:09 +0000)]
* objcopy.c (use_alt_mach_code): Change type to unsigned long.
(copy_object): If bfd_alt_mach_code fails emit a more helpful message
and if the target architecture is ELF use the alternative as replacement
value for the e_machine number.
(copy_main): Use strtoul to parse the number provided with the
--alt-mach-code switch.
* doc/binutils.texi (--alt-mach-code): Document that this switch can now
set the absolute e_machine value.
PR/1070
* macro.c (getstring): Don't treat parentheses special anymore.
(get_any_string): Don't consider '(' and ')' as quoting anymore.
Special-case '(', ')', '[', and ']' when dealing with non-quoting
characters.
* Makefile.in: TEXI2HTML uses makeinfo. Define
HTMLFILES. Add html targets.
* configure.texi: Use ifnottex. Add alternative
image format specifier as jpg.
* standards.texi: Use ifnottex.
* i386-dis.c (IS_3BYTE_OPCODE): New for 3-byte opcodes used by
Intel Merom New Instructions.
(THREE_BYTE_0): Likewise.
(THREE_BYTE_1): Likewise.
(three_byte_table): Likewise.
(dis386_twobyte): Use THREE_BYTE_0 for entry 0x38. Use
THREE_BYTE_1 for entry 0x3a.
(twobyte_has_modrm): Updated.
(twobyte_uses_SSE_prefix): Likewise.
(print_insn): Handle 3-byte opcodes used by Intel Merom New
Instructions.
* gas/sparc/rdhpr.s: New test.
* gas/sparc/rdhpr.d: New test.
* gas/sparc/wrhpr.s: New test.
* gas/sparc/wrhpr.d: New test.
* gas/sparc/window.s: New test.
* gas/sparc/window.d: New test.
* gas/sparc/rdpr.s: Add case for reading %gl register.
* gas/sparc/rdpr.d: Likewise.
* gas/sparc/wrpr.s: Add case for writing %gl register.
* gas/sparc/wrpr.d: Likewise.
* gas/sparc/sparc.exp: Update for new tests.
* config/tc-sparc.c (priv_reg_table): Add entry for "gl".
(hpriv_reg_table): New table for hyperprivileged registers.
(sparc_ip): New cases '$' and '%' for wrhpr/rdhpr hyperprivileged
register encoding.
* sparc-dis.c (v9_priv_reg_names): Add "gl" entry.
(v9_hpriv_reg_names): New table.
(print_insn_sparc): Allow values up to 16 for '?' and '!'.
New cases '$' and '%' for read/write hyperprivileged register.
* sparc-opc.c (sparc_opcodes): Add new entries for UA2005
window handling and rdhpr/wrhpr instructions.
[gas]
* config/tc-m32c.h (md_apply_fix): Define to m32c_apply_fix.
(tc_gen_reloc): Don't define.
* config/tc-m32c.c (rl_for, relaxable): New convenience macros.
(OPTION_LINKRELAX): New.
(md_longopts): Add it.
(m32c_relax): New.
(md_parse_options): Set it.
(md_assemble): Emit relaxation relocs as needed.
(md_convert_frag): Emit relaxation relocs as needed.
(md_cgen_lookup_reloc): Add LAB_8_8 and LAB_8_16.
(m32c_apply_fix): New.
(tc_gen_reloc): New.
(m32c_force_relocation): Force out jump relocs when relaxing.
(m32c_fix_adjustable): Return false if relaxing.
[bfd]
* elf32-m32c.c (m32c_elf_howto_table): Add relaxation relocs.
(m32c_elf_relocate_section): Don't relocate them.
(compare_reloc): New.
(relax_reloc): Remove.
(m32c_offset_for_reloc): New.
(m16c_addr_encodings): New.
(m16c_jmpaddr_encodings): New.
(m32c_addr_encodings): New.
(m32c_elf_relax_section): Relax jumps and address displacements.
(m32c_elf_relax_delete_bytes): Adjust for internal syms. Fix up
short jumps.
Wu Zhou [Fri, 24 Feb 2006 07:26:10 +0000 (07:26 +0000)]
* f-exp.y: Symbol '%' is not used as the modulus operator in
Fortran. Delete this from Fortran expression.
It is now used by Fortran 90 and later to access the member
of derived type. Add this into Fortran expression.
* f-valprint.c (f_val_print): Add code to handle TYPE_CODE_STRUCT.
Print each elements in the derived type.
* f-typeprint.c (print_equivalent_f77_float_type): Add a parameter
level into the function definition to do indented printing. And
call fprintfi_filtered instead to do indented printing.
(f_type_print_base): Replace fprintf_filtered with the indented
version (fprintfi_filtered).
(f_type_print_base): Call indented print_equivalent_f77_float_type.
(f_type_print_base): Add code to handle TYPE_CODE_STRUCT. Print
the definition of the derived type.
projects / binutils.git / log