(ia64_gen_real_reloc_type): Handle it.
(pseudo_func): Add @ltoffx.
(md_begin): Build .<ltoffx>.
(ia64_force_relocation): True for LTOFF22X and LDXMOV.
* gas/ia64/ltoff22x-1.[ds]: New.
* gas/ia64/ia64.exp: Run it.
Move OBJ_COFF TC_FORCE_RELOCATION code here so that COFF handles
ARM_IMMEDIATE and ARM_ADRL_IMMEDIATE relocs as for ELF.
* config/tc-arm.h (TC_FORCE_RELOCATION): Define for both ELF and
COFF to call arm_force_relocation.
* config/tc-m68hc11.h (md_relax_frag): Define to support relaxations
that are not pc-relative.
(m68hc11_relax_frag): Declare.
* config/tc-m68hc11.c (build_indexed_byte): Use a frag_var to handle
the offsetable indexed addressing modes (n,r).
(build_insn): Cleanup some locals.
(m68hc11_relax_frag): New function imported from tc-cris.c to handle
relaxation of difference between two symbols of same section.
(md_convert_frag): For INDEXED_OFFSET relaxs, use the displacement
only when this is a PC-relative operand and the offset is not absolute.
(md_estimate_size_before_relax): Convert the INDEXED_OFFSET,UNDEF frag
to INDEXED_OFFSET,STATE_BITS5 when the symbol is absolute; this will
be handled by m68hc11_relax_frag.
(32-bit int, 64-bit double).
(md_longopts): New options -mshort, -mlong, -mshort-double and
-mlong-double to control the ABI.
(md_show_usage): Update.
(md_parse_option): Handle new options.
* doc/as.texinfo (Overview): Document new options for HC11/HC12.
* config/tc-mcore.c (md_begin): Use a const iterator. Don't
coalesce the name fields to use the same pointer.
* config/tc-sh.c (md_begin): Use a const iterator. Don't coalesce
the name fields to use the same pointer.
(get_specific): Check for opcodes with the same name using strcmp
as well as comparing the pointer.
* config/tc-h8300.c (struct h8_instruction): New type, used to
wrap h8_opcodes with length, noperands, idx, and size fields
(computed at run-time).
(h8_instructions): New variable.
(md_begin): Allocate the storage for h8_instructions. Fill
h8_instructions with pointers to the appropriate opcode and the
correct value for the additional fields.
(clever_message): Update to use h8_instructions instead of
h8_opcodes.
(build_bytes): Ditto.
(get_specific): Ditto.
(md_assemble): Ditto.
parameters. Added support for new opcode-list format. General
error message fixups.
(c4x_inst_add): Reject insn not for our CPU
(md_begin): Added matrix for setting the proper opcode-level &
device-flags according to cpu type and revision. Rewrite the
opcode hasher.
(c4x_operand_parse): Fix opcode bug
(c4x_operands_match): New function argument. Added dry-run
mechanism, that is optional error generation. Added constraint 'i'
and 'j'.
(c4x_insn_check): Added new function for post-verification of the
generated insn.
(md_assemble): Check all opcodes before croaking because of an
argument mismatch. Need this to be able to fully support
ortogonally arguments.
(md_parse_options): Revised commandprompt swicthes and added new
ones.
(md_show_usage): Complete rewrite of printout.
* gas/testsuite/gas/tic4x/addressing.s: Fix bug in one insn
* gas/testsuite/gas/tic4x/addressing_c3x.d: Update thereafter
* gas/testsuite/gas/tic4x/addressing_c4x.d: Update thereafter
* gas/testsuite/gas/tic4x/allopcodes.S: Add support for new
opclass.h changes
* gas/testsuite/gas/tic4x/opclasses.h: Added testsuites for
the new enhanced opcodes.
* gas/testsuite/gas/tic4x/opcodes.s: Regenerate
* gas/testsuite/gas/tic4x/opcodes_c3x.d: Update from above
* gas/testsuite/gas/tic4x/opcodes_c4x.d: Update from above
* gas/testsuite/gas/tic4x/opcodes_new.d: Added new testsuite for
the enhanced and special insns.
* gas/testsuite/gas/tic4x/tic4x.exp: Added the opcodes_new testsuite
* include/opcode/tic4x.h: File reordering. Added enhanced opcodes.
* opcodes/tic4x-dis.c: Added support for enhanced and special
insn.
(c4x_print_op): Added insn class 'i' and 'j'
(c4x_hash_opcode_special): Add to support special insn
(c4x_hash_opcode): Update to support the new opcode-list
format. Add support for the new special insns.
(c4x_disassemble): New opcode-list support.
(c4x_operands_match): Added check for 8-bits LDF insn. Give
warning when using constant direct bigger than 2^16. Add the new
arguments.
* include/opcode/tic4x.h: Major rewrite of entire file. Define
instruction classes, and put each instruction into a class.
* opcodes/tic4x-dis.c: (c4x_print_op): Add support for the new
argument format. Fix bug in 'N' register printer.
(cris_relax_frag): New function.
(md_estimate_size_before_relax) <case ENCODE_RELAX
(STATE_BASE_PLUS_DISP_PREFIX, STATE_UNDF)>: Pass on unresolved
expressions that will become absolute expressions to relaxation.
(md_convert_frag) <case ENCODE_RELAX (STATE_BASE_PLUS_DISP_PREFIX,
STATE_WORD)>: Expect only absolute expressions. Use the symbol
value, not distance to symbol.
<case ENCODE_RELAX (STATE_BASE_PLUS_DISP_PREFIX, STATE_BYTE)>:
Ditto. Correct placement of fixup.
(md_assemble): Use SIMPLE_EXPR when dissecting expressions.
(gen_bdap): Ditto.
* config/tc-cris.h (cris_relax_frag): Declare.
(md_relax_frag): Define.
(mips_pseudo_table): Add .gpdword.
(mips_need_elf_addend_fixup): never for NEWABI.
(md_apply_fix3): Don't mark BFD_RELOC64 after GPREL16 or
GPREL32 as done.
(s_cpadd): Generate .cpadd on NEWABI.
(sh_force_relocation): Use sh_local_pcrel.
(md_pcrel_from_section): Check the relocation type whether it
should be resolved locally. Use S_FORCE_RELOC.
* testsuite/gas/sh/pcrel2.d: Check code also.
* doc/c-mips.texi: Add entries for -march=vr4120,vr4130,vr4181,
vr5400 and vr5500. Add entry for -mfix-vr4122-bugs.
* config/tc-mips.c (CPU_HAS_DROR, CPU_HAS_ROR): New macros.
(hilo_interlocks): True for CPU_VR5500.
(gpr_interlocks, cop_interlocks): True for CPU_VR5400 and CPU_VR5500.
(mips_fix_vr4122_bugs): New.
(append_insn): Work around 4122 errors if mips_fix_vr4122_bugs.
(mips_emit_delays): Likewise.
(macro2) [M_DROLI]: Use dror or dror32 if CPU_HAS_DROR.
[M_ROLI]: Likewise ror if CPU_HAS_ROR.
(validate_mips_insn, mips_ip): Handle '[', ']', 'e' and '%'.
(OPTION_FIX_VR4122, OPTION_NO_FIX_VR4122): New options.
(md_longopts): Add -mfix-vr4122-bugs and -no-mfix-vr4122-bugs.
(OPTION_ELF_BASE): Bump.
(md_parse_option): Handle the new options.
(mips_cpu_info_table): Add entries for vr4120, vr4130, vr4181,
vr5400 and vr5500.
[gas/testsuite/]
* gas/mips/mips4100.[sd]: Move dmadd16 and madd16 checks to...
* gas/mips/vr4111.[sd]: ...this new test.
* gas/mips/vr4120.[sd],
* gas/mips/vr4122.[sd],
* gas/mips/vr5400.[sd],
* gas/mips/vr5500.[sd]: New tests.
* mips.exp: Run them.
* config/tc-sh.c (sh_force_relocation): Return 0 for
some PC relative relocations when not relaxing.
* testsuite/gas/sh/pcrel2.s: New.
* testsuite/gas/sh/pcrel2.d: New.
* testsuite/gas/sh/basic.exp: Add pcrel2 test.
* reloc.c (BFD_RELOC_386_TLS_TPOFF, BFD_RELOC_386_TLS_IE,
BFD_RELOC_386_TLS_GOTIE): Add.
* bfd-in2.h, libbfd.h: Rebuilt.
* elf32-i386.c (elf_howto_table): Add R_386_TLS_TPOFF, R_386_TLS_IE
and R_386_TLS_GOTIE.
(elf_i386_reloc_type_lookup): Handle it.
(struct elf_i386_link_hash_entry): Change tls_type type to unsigned
char instead of enum, change GOT_* into defines.
(GOT_TLS_IE_POS, GOT_TLS_IE_NEG, GOT_TLS_IE_BOTH): Define.
(elf_i386_tls_transition): Handle R_386_TLS_IE and R_386_TLS_GOTIE.
(elf_i386_check_relocs): Likewise. Avoid crash if local symbol is
accessed both as normal and TLS symbol. Move R_386_TLS_LDM and
R_386_PLT32 cases so that R_386_TLS_IE can fall through.
Handle R_386_TLS_LE_32 and R_386_TLS_LE in shared libs.
(elf_i386_gc_sweep_hook): Handle R_386_TLS_IE and R_386_TLS_GOTIE.
Handle R_386_TLS_LE_32 and R_386_TLS_LE in shared libs.
(allocate_dynrelocs): Allocate 2 .got and 2 .rel.got entries if
tls_type is GOT_TLS_IE_BOTH.
(elf_i386_size_dynamic_sections): Likewise.
(elf_i386_relocate_section): Handle R_386_TLS_IE and R_386_TLS_GOTIE.
Handle R_386_TLS_LE_32 and R_386_TLS_LE in shared libs.
(elf_i386_finish_dynamic_symbol): Use tls_type & GOT_TLS_IE to catch
all 4 GOT_TLS_* TLS types.
gas/
* config/tc-i386.c (tc_i386_fix_adjustable): Handle
BFD_RELOC_386_TLS_IE and BFD_RELOC_386_TLS_GOTIE.
(BFD_RELOC_386_TLS_IE, BFD_RELOC_386_TLS_GOTIE): Define to 0
if not defined.
(lex_got): Handle @GOTNTPOFF and @INDNTPOFF.
(md_apply_fix3, tc_gen_reloc): Handle BFD_RELOC_386_TLS_IE and
BFD_RELOC_386_TLS_GOTIE.
gas/testsuite/
* gas/i386/tlspic.s: Add tests.
* gas/i386/tlspic.d: Regenerated.
* gas/i386/tlsnopic.s: Add tests.
* gas/i386/tlsnopic.d: Regenerated.
include/
* elf/i386.h (R_386_TLS_TPOFF, R_386_TLS_IE, R_386_TLS_GOTIE):
Define.
ld/testsuite/
* ld-i386/i386.exp: New.
* ld-i386/tlsbin.dd: New test.
* ld-i386/tlsbinpic.s: New test.
* ld-i386/tlsbin.rd: New test.
* ld-i386/tlsbin.s: New test.
* ld-i386/tlsbin.sd: New test.
* ld-i386/tlsbin.td: New test.
* ld-i386/tlslib.s: New test.
* ld-i386/tlsnopic1.s: New test.
* ld-i386/tlsnopic2.s: New test.
* ld-i386/tlsnopic.dd: New test.
* ld-i386/tlsnopic.rd: New test.
* ld-i386/tlsnopic.sd: New test.
* ld-i386/tlspic1.s: New test.
* ld-i386/tlspic2.s: New test.
* ld-i386/tlspic.dd: New test.
* ld-i386/tlspic.rd: New test.
* ld-i386/tlspic.sd: New test.
* ld-i386/tlspic.td: New test.
branch type relocs.
(alpha_force_relocation): Don't special-case branch type relocs.
* gas/alpha/elf-reloc-7.s: New.
* gas/alpha/elf-reloc-7.d: New.
* gas/alpha/alpha.exp: Run it.
2002-09-18 Chris Demetriou <cgd@broadcom.com>
* config/tc-mips.c (IS_SEXT_32BIT_NUM): Move closer to top of file.
(IS_SEXT_16BIT_NUM): New macro.
(macro_build_ldst_constoffset): New function, to build a set of
instructions to do a load or store from a constant offset relative
to a given register.
(macro, s_cprestore): Use macro_build_ldst_constoffset to implement
.cprestore pseudo-op.
[ gas/testsuite/ChangeLog ]
2002-09-18 Chris Demetriou <cgd@broadcom.com>
* gas/mips/mips-abi32-pic2.s: New file.
* gas/mips/mips-abi32-pic2.d: New file.
* gas/mips/mips.exp: Run new test.
[ plus, fixed date on prev. gas/testsuite/ChangeLog entry. ]
2002-09-18 Chris Demetriou <cgd@broadcom.com>
* config/tc-mips.c (md_apply_fix3): Just return for BFD_RELOC_8.
[ gas/testsuite/ChangeLog ]
2002-09-18 Chris Demetriou <cgd@broadcom.com>
* gas/mips/baddata1.s: New file.
* gas/mips/baddata1.l: New file.
* gas/mips/mips.exp: Run new test.
(macro): Likewise for la. Likewise for ld.
(mips_after_parse_args): Make -xgot optional, not the default.
(md_apply_fix3): Allow composite relocation to set up gp.
(tc_gen_reloc): Allow relaxing for newabi.
Relax R_MIPS_CALL16 to R_MIPS_GOT_PAGE/R_MIPS_GOT_OFST if local.
Relax R_MIPS_GOT16/R_MIPS_LO16 to R_MIPS_GOT_DISP if local.
Do not issue reloc number of unimplemented BFD_RELOC_ARM_IMMEDIATE and
BFD_RELOC_ARM_OFFSET_IMM relocs - their name is already in the error message
- plus remove them from the default case.
Warn about unpredictable behavior of LDM and STM instructions.
* elf32-i386.c: Don't defined ELF_ARCH etc. if this file is included
by a target variant implementation.
* elf64-alpha.c: Likewise.
* elf32-i386-fbsd.c: New file.
* elf64-alpha-fbsd.c: New file.
* targets.c: Support bfd_elf32_i386_freebsd_vec and
bfd_elf64_alpha_freebsd_vec.
* configure.in: Accept the vectors bfd_elf32_i386_freebsd_vec,
bfd_elf64_alpha_freebsd_vec.
* Makefile.am (BFD32_BACKENDS): Add elf32-i386-fbsd.lo.
(BFD32_BACKENDS_CFILES): Add elf32-i386-fbsd.c.
(BFD64_BACKENDS): Add elf64-alpha-fbsd.lo.
(BFD64_BACKENDS_CFILES): Add elf64-alpha-fbsd.c.
(elf32-i386-fbsd.lo, elf64-alpha-fbsd.lo): Add dependencies.
* config.bfd: For FreeBSD targets, set targ_defvec to a FreeBSD
specific targets. Define OLD_FREEBSD_ABI_LABEL if appropriate.
* config/tc-i386.h (ELF_TARGET_FORMAT): New macro.
(TARGET_FORMAT): Use ELF_TARGET_FORMAT instead of "elf32-i386".
* config/tc-i386.c (i386_target_format): Likewise.
* config/tc-alpha.h (ELF_TARGET_FORMAT): New macro.
(TARGET_FORMAT): Use ELF_TARGET_FORMAT instead of "elf64-alpha".
* emulparams/elf_i386_fbsd.sh: Set OUTPUT_FORMAT to
elf32-i386-freebsd.
* emulparams/elf64alpha_fbsd.sh: Set OUTPUT_FORMAT to
elf64-alpha-freebsd.
Approved by: Alan Modra <amodra@bigpond.net.au>
Message-ID: <20020715021113.GJ30362@bubble.sa.bigpond.net.au>
(TC_ADJUST_RELOC_COUNT): Tweak param name.
(TC_FORCE_RELOCATION, TC_FORCE_RELOCATION_SECTION): Likewise.
(TC_FIX_ADJUSTABLE, MD_PCREL_FROM_SECTION): Likewise.
(RELOC_ENUM): Define.
(fix_new_internal): Use RELOC_ENUM.
(fix_new, fix_new_exp): Likewise.
(adjust_reloc_syms): Comment. Remove unnecessary tests on sym != NULL.
Replace gotos with continue.
(write_relocs): Formatting. Avoid symbol loops in
RELOC_EXPANSION_POSSIBLE case too. Report bfd_reloc_outofrange
errors, and error number in other cases.
(fixup_segment): Remove param names from prototype. Rename
"this_segment_type" to "this_segment". Update linkrelax comment.
Remove "size, "place" and "where" local vars. Formatting. Update
"no symbol" comment. Remove #if 0 and #if 1.
* config/tc-ppc.c (ppc_cleanup): Do something only if format
is ELF.
(ppc_apuinfo_section_add): Define only if format is ELF.
(md_assemble): Emit APUinfo section only if format is ELF.
Fix formatting.
* config/tc-mips.c (macro2): Implement rotates by zero using shifts
by zero.
[gas/testsuite]
* gas/mips/rol.s: Add rotate by zero tests.
* gas/mips/rol.d: Update accordingly.
* gas/mips/rol64.d: Expect rotates by zero to use dsrl.
From matthew green <mrg@redhat.com>
* config/tc-ppc.c (PPC_OPCODE_CLASSIC): Enable this everywhere
PPC_OPCODE_PPC is, except for BookE architectures.
(md_parse_option): Add support for -mspe.
(md_show_usage): Add -mspe.
(md_parse_option): Add support for -me500 and
-me500x2 to generate code for Motorola e500 core complex.
(md_show_usage): Add -me500 and -me500x2.
(PPC_APUINFO_ISEL, PPC_APUINFO_PMR, PPC_APUINFO_RFMCI,
PPC_APUINFO_CACHELCK, PPC_APUINFO_SPE, PPC_APUINFO_EFS,
PPC_APUINFO_BRLOCK): New macros.
(ppc_cleanup): New function.
(ppc_apuinfo_section_add): New function.
(APUID): New macro.
(md_assemble): Collect info and write the APUinfo section.
* config/tc-ppc.h (md_cleanup): Define.
(ppc_cleanup): Export.
(ELF_TC_SPECIAL_SECTIONS): Add .PPC.EMB.apuinfo section.
the fix-up against on the right frag.
(s_cpsetup): Likewise. Parse third argument as expression, to
handle global symbols and forward/backward labels correctly.
(TC_HANDLES_FX_DONE): Define to let md_apply_fix3 set fx_done flag
according to the reloc.
(tc_fix_adjustable, tc_m68hc11_fix_adjustable): Define.
(TC_FORCE_RELOCATION): Define.
(tc_m68hc11_force_relocation): Declare.
* config/tc-m68hc11.c (md_pseudo_table): Add relax command.
(s_m68hc11_relax): New function for relax group.
(build_insn, build_jump_insn): Emit a M68HC11_RL_JUMP reloc at
beginning of jump instruction.
(md_pcrel_from): Rename from md_pcrel_from_section and fix
address computation.
(tc-gen_reloc): Update.
(md_estimate_size_before_relax): Create the BFD_RELOC_16_PCREL as
PC-relative fixup.
(tc_m68hc11_force_relocation): New function, handle new relocs.
(tc_m68hc11_fix_adjustable): New to make sure there are enough
reloc for the linker relax pass.
(md_apply_fix3): Handle M68HC11_RL_JUMP, M68HC11_RL_GROUP
and VTABLE relocs.
(md_pseudo_table): Add .mode, .far and .interrupt pseudo op.
(s_m68hc11_mode): New function for .mode pseudo op.
(s_m68hc11_mark_symbol): New function for .far and .interrupt
pseudo op.
* config/tc-m68hc11.h (elf_tc_final_processing): Define.
(m68hc11_elf_final_processing): Declare.
page operand for call instruction.
(print_opcode_format): Likewise.
(check_range): Likewise for page range checking.
(get_operand): Don't skip a possible comma in operands.
(fixup8): Generate BFD_RELOC_M68HC11_PAGE reloc.
(fixup16): Likwise with BFD_RELOC_M68HC11_LO16.
(fixup24): New to handle call reloc.
(build_insn): Handle missing page operand for call instruction.
(find): Likewise.
(md_apply_fix3): Take into account new relocs.
(get_operand): Fix the mode for indexed indirect addressing.
(build_indexed_byte): Fix post index byte for indexed indirect mode.
* config/tc-mips.c (mips_ip): Don't work out the value of
constant %hi()s here.
[gas/testsuite/]
* gas/mips/elf-consthilo.[sd]: New test.
* gas/mips/mips.exp: Run it.
for ELF, and don't bother checking ELF relocs when non-ELF.
(i386_immediate): Allow absolute_section expressions for aout.
(i386_displacement): Likewise. Also test bfd_is_com_section.
(md_estimate_size_before_relax): Test OUTPUT_FLAVOR for ELF.
(md_apply_fix3): Hack for bfd_install_relocation when fx_pcrel,
not when fx_addsy. Remove dead code.
immediate offset in "dla" and "la" expansions.
* gas/mips/empic.d: Treat "addiu" and "daddiu" as equivalent when
$0 is source.
* mips-opc.c (mips_builtin_opcodes): Remove "dla" and "la" as
aliases to "daddiu" and "addiu".
at start of insn, pass it to output_disp and output_imm.
(output_disp): Added arguments. If _GLOBAL_OFFSET_TABLE_ is seen
in displacement for R_386_32 reloc, use R_386_GOTPC and compute
properly addend.
(output_imm): Added arguments. Compute properly addend for
R_386_GOTPC.
(md_apply_fix3): Remove R_386_GOTPC handling.
* testsuite/gas/i386/gotpc.s: New.
* testsuite/gas/i386/gotpc.d: New.
* testsuite/gas/i386/i386.exp: Add gotpc test.
* config/tc-mips.c (tc_gen_reloc): Extend GP-relative addend
handling to BFD_RELOC_MIPS16_GPREL.
[gas/testsuite/]
* gas/mips/elf-rel6.[sd]: New test.
* gas/mips/mips.exp: Run it.
* doc/c-mips.texi: Remove -mcpu. Document -mabi.
* configure.in (MIPS_CPU_STRING_DEFAULT): New configuration macro.
(USE_E_MIPS_ABI_O32, MIPS_DEFAULT_64BIT): New configuration macros.
* configure, config.in: Regenerate.
* config/tc-mips.c (file_mips_abi): Rename to mips_abi.
(mips_set_options): Remove "abi" field.
(mips_opts): Update accordingly. Replace all uses of mips_opts.abi
with mips_abi.
(mips_cpu): Remove.
(mips_arch_string, mips_arch_info): New vars.
(mips_tune_string, mips_tune_info): New vars.
(ABI_NEEDS_32BIT_REGS, ABI_NEEDS_64BIT_REGS): New macros.
(HAVE_32BIT_GPRS, HAVE_32BIT_FPRS): Don't check the ABI.
(mips_isa_to_str, mips_cpu_to_str): Remove.
(mips_ip): If the selected architecture is a generic ISA rather
than a processor, only mention the ISA level in error messages.
(OPTION_MCPU): Remove.
(OPTION_FP64): New.
(md_longopts): Add -mfp64, remove -mcpu.
(mips_set_option_string): New fn.
(md_parse_option): Make -mipsN update file_mips_isa rather than
mips_opts.isa. Use mips_set_option_string to set -march or -mtune.
Don't let -mgp32 and -mfp32 change the ABI.
(show): Move to end of file. Constify string argument.
(md_show_usage): Move to the end of the file. Read available
architectures from mips_cpu_info_table.
(mips_set_architecture): New fn.
(mips_after_parse_args): Rework. Remove -mcpu handling. -mipsN
is an alias for -march=mipsN. Don't change the ABI based on other
flags. Infer the register size from the ABI as well as the
architecture. Complain about more conflicting arguments.
Unify logic with GCC.
(s_mipsset): Don't change the ABI.
(mips_elf_final_processing): Check USE_E_MIPS_ABI_O32.
(mips_cpu_info_table): Remove Generic-MIPS* entries, keeping just
"mipsN"-type entries. Remove entries that vary only in the
manufacturer's prefix, or that have "000" replaced by "k".
Remove TARGET_CPU entries. Make r2000 entry use CPU_R3000.
(mips_strict_matching_cpu_name_p, mips_matching_cpu_name_p): New fns.
(mips_parse_cpu): New fn.
(mips_cpu_info_from_name, mips_cpu_info_from_cpu): Remove.
(mips_cpu_info_from_isa): Minor formatting tweak.
[gas/testsuite]
* gas/mips/mips-gp32-fp64.d,
* gas/mips/mips-gp32-fp64-pic.d: Add -mfp64.
(ppc_cpu): Make it unsigned long to agree with struct powerpc_opcode
flags.
(ppc_size): Delete.
(ppc_xcoff64): Rename to ppc_obj64.
(md_parse_option <m>): Encode old ppc_size value in ppc_cpu.
(ppc_set_cpu): Set PPC_OPCODE_32 too.
(ppc_arch): Use ppc_obj64 instead of ppc_size to select bfd_mach_ppc64
or bfd_mach_ppc.
(ppc_target_format): Use ppc_obj64 to select format.
(md_begin): Adjust for PPC_OPCODE_32/64 in ppc_cpu.
(ppc_insert_operand): Use ppc_obj64 instead of ppc_size.
(ppc_elf_suffix): Likewise. Don't depend on BFD_DEFAULT_TARGET_SIZE.
(tc_frob_symbol): Likewise.
(md_assemble): Use ppc_obj64 instead of ppc_size. Don't depend on
BFD_DEFAULT_TARGET_SIZE.
(ppc_tc): Likewise.
(ppc_is_toc_sym): Likewise.
(md_apply_fix3): Likewise.
* config/tc-ppc.h (TC_FORCE_RELOCATION): Don't depend on
BFD_DEFAULT_TARGET_SIZE.
(ELF_TC_SPECIAL_SECTIONS): Likewise.
(tc_frob_symbol): Likewise.
(mips_ip): Likewise.
* mips.h (INSN_MIPS16): New define.
* mips-dis.c (mips_isa_type): Add MIPS16 insn handling.
* mips-opc.c (I16): New define.
(mips_builtin_opcodes): Make jalx an I16 insn.
From Catherine Moore, Michael Meissner, Richard Sandiford, Dave Brolley:
* po/POTFILES.in: Add tc-frv.c, tc-frv.h.
* configure.in: Support frv-*-*.
* Makefile.am (CPU_TYPES): Add frv.
(TARGET_CPU_CFILES): Add tc-frv.c.
(TARGET_CPU_HFILES): Add tc-frv.h.
(DEPTC_frv_coff): New variable.
(DEPTC_frv_elf): New variable.
(DEPOBJ_frv_coff): New variable.
(DEPOBJ_frv_elf): New variable.
(DEP_frv_coff): New variable.
(DEP_frv_elf): New variable.
* tc-frv.c: New file.
* tc-frv.h: New file.
* configure.in (vax-*-netbsdelf*, vax-*-netbsdaout*)
(vax-*-netbsd*): New targets.
* configure: Regenerate.
* config/aout_gnu.h (enum machine_type): Add M_VAX4K_NETBSD.
* config/tc-vax.c: Add support for ELF and PIC.
(flag_want_pic): New flag.
(float_cons): Fix prototype.
(md_apply_fix3): Adjust for BFD_ASSEMBLER.
(md_assemble): Introduce a new is_absolute local, and use it
rather than repeating the test. Make fatal errors actually
fatal by using as_fatal as appropriate. Adjust for BFD_ASSEMBLER.
Add support for ELF. Add support for PIC.
(md_convert_frag): Adjust for BFD_ASSEMBLER.
(tc_aout_fix_to_chars): Only include if OBJ_AOUT and not
BFD_ASSEMBLER.
(vax_reg_parse): Make the % register prefix mandatory for ELF,
optional for a.out, and not allowed for VMS. Adjust all callers.
(md_create_short_jump): Add ATTRIBUTE_UNUSED to unused arguments.
(md_create_long_jump): Likewise.
(md_undefined_symbol): Likewise.
(md_section_align): Likewise.
(md_shortopts): Allow -k and -K for ELF.
(md_parse_option): Set flag_want_pic if -k or -K.
(tc_headers_hook): New function if OBJ_AOUT and not BFD_ASSEMBLER.
(tc_gen_reloc): New function if BFD_ASSEMBLER.
* config/tc-vax.h (tc_headers_hook): Remove.
(TARGET_FORMAT): Set according to object format and target
environment.
(BFD_ARCH, TARGET_ARCH): Define.
(NO_RELOC): Adjust for BFD_ASSEMBLER.
(TC_RELOC_RTSYM_LOC_FIXUP, TC_FIX_ADJUSTABLE)
(tc_fix_adjustable): Define if BFD_ASSEMBLER.
* config/vax-inst.h (VAX_JSB, VAX_CALLS, VAX_CALLG): Define.
Based on patch from Matt Green:
* config/obj-elf.h (ECOFF_DEBUGGING): Make configurable.
* config/tc-mips.c (s_mips_file): Renamed from s_file.
(s_mips_loc): New function.
(mips_nonecoff_pseudo_table): Call them.
(append_insn): Call dwarf2_emit_insn.
to -1.
(file_ase_mips16): New veriable.
(mips_eabi64): Remove.
(CPU_HAS_MIPS16): New define.
(CPU_HAS_MDMX): Fix data type.
(md_begin): Code cleanup. Use file_ase_mips16.
(mips_elf_final_processing): Handle mips16 header flag.
Handle EABI flag without intermediate variable.
(md_begin): Add $zero as alias name for $0.
(insn_uses_reg): Use ZERO define.
(mips_ip): Add $zero as alias name for $0.
(mips16_ip): Likewise.
(s_cplocal): Demand empty rest of input line.
(tc_get_register): Likewise. Add support for $kt0, kt1 register
names. Use ZERO define. Fix input_line_pointer progress.
(symbol_find_exact): Split out from..
(symbol_find_base): ..here.
* symbols.h: Replace CONST by const throughout.
(symbol_find_exact): Declare.
* config/obj-elf.c: #include "struc-symbol.h".
(elf_frob_file): If group name matches an exported symbol, use that
symbol for the signature and ".group" as the section name.
linkonce flag. Print all warnings.
(obj_elf_section): Parse ",comdat" for groups.
(elf_frob_file): Set SEC_LINK_ONCE on COMDAT groups. Check
consistency of comdat flag.
2002-05-30 Chris G. Demetriou <cgd@broadcom.com>
Ed Satterthwaite <ehs@broadcom.com>
* config/tc-mips.c (mips_set_options): New "ase_mdmx" member.
(mips_opts): Initialize "ase_mdmx" member.
(file_ase_mdmx): New variable.
(CPU_HAS_MDMX): New macro.
(md_begin): Initialize mips_opts.ase_mdmx and file_ase_mdmx
based on command line options and configuration defaults.
(macro_build): Note in comment that use of MDMX in macros is
not currently allowed.
(validate_mips_insn): Add support for the "O", "Q", "X", "Y", and
"Z" MDMX operand types.
(mips_ip): Accept MDMX instructions if mips_opts.ase_mdmx is set,
and add support for the "O", "Q", "X", "Y", and "Z" MDMX operand
types.
(OPTION_MDMX, OPTION_NO_MDMX, md_longopts, md_parse_option):
Add support for "-mdmx" and "-no-mdmx" options.
(OPTION_ELF_BASE): Move to accomodate new options.
(s_mipsset): Support ".set mdmx" and ".set nomdmx".
(mips_elf_final_processing): Set MDMX ASE ELF header flag if
file_ase_mdmx was set.
* doc/as.texinfo: Document -mdmx and -no-mdmx options.
* doc/c-mips.texi: Likewise, and document ".set mdmx" and ".set
nomdmx" directives.
[ gas/testsuite/ChangeLog ]
2002-05-30 Chris G. Demetriou <cgd@broadcom.com>
* gas/mips/mips64-mdmx.s: New file.
* gas/mips/mips64-mdmx.d: Likewise.
* gas/mips/mips.exp: Run new "mips64-mdmx" test.
[ include/opcode/ChangeLog ]
2002-05-30 Chris G. Demetriou <cgd@broadcom.com>
* mips.h (OP_SH_ALN, OP_MASK_ALN, OP_SH_VSEL, OP_MASK_VSEL)
(MDMX_FMTSEL_IMM_QH, MDMX_FMTSEL_IMM_OB, MDMX_FMTSEL_VEC_QH)
(MDMX_FMTSEL_VEC_OB, INSN_READ_MDMX_ACC, INSN_WRITE_MDMX_ACC)
(INSN_MDMX): New constants, for MDMX support.
(opcode character list): Add "O", "Q", "X", "Y", and "Z" for MDMX.
[ opcodes/ChangeLog ]
2002-05-30 Chris G. Demetriou <cgd@broadcom.com>
Ed Satterthwaite <ehs@broadcom.com>
* mips-dis.c (print_insn_arg): Add support for 'O', 'Q', 'X', 'Y',
and 'Z' formats, for MDMX.
(mips_isa_type): Add MDMX instructions to the ISA
bit mask for bfd_mach_mipsisa64.
* mips-opc.c: Add support for MDMX instructions.
(MX): New definition.
* mips-dis.c: Update copyright years to include 2002.
(mips_big_got): Initialize.
(mips_trap): Initialize.
(load_address): Use mips_gp_register instead of hardcoded value.
Remove dbl parameter, use HAVE_32BIT_ADDRESSES instead.
(macro): Use mips_gp_register instead of hardcoded value.
(macro2): Change load_address calls.
(md_pcrel_from): Comment formatting.
(s_cpload): Use mips_gp_register instead of hardcoded value.
(s_cprestore): Likewise. Comment formatting.
(s_gpword): Fix data type.
(s_cpadd): Use mips_gp_register instead of hardcoded value.
(nopic_need_relax): Replace GP in comments by $gp.
(mips_elf_final_processing): Better comment.
* configure.in: Use ns32k-*-netbsd* instead of ns32k-pc532-netbsd*.
* configure: Regenerate.
ld:
* configure.tgt: Use ns32k-*-netbsd* instead of ns32k-pc532-netbsd*.
(SUB_SEGMENT_ALIGN): If HANDLE_ALIGN defined, pad out last frag to
section alignment.
* config/obj-coff.c (SUB_SEGMENT_ALIGN): Likewise.
* config/obj-ieee.c (SUB_SEGMENT_ALIGN): Likewise.
(write_object_file): Invoke md_do_align if available, and use
frag_align_code on text sections.
* config/obj-vms.h (SUB_SEGMENT_ALIGN): Now two args.
* config/tc-m88k.h (SUB_SEGMENT_ALIGN): Likewise.
* config/tc-ppc.h (SUB_SEGMENT_ALIGN): Likewise.
* config/tc-sh.h (SUB_SEGMENT_ALIGN): Likewise.
* config/tc-i386.h (SUB_SEGMENT_ALIGN): Likewise. Define for
BFD_ASSEMBLER too.
a crash of the assembler with some malformed source input.
Long segmented addresses are now correctly relocated.
Finally it updates my email address in the MAINTAINERS file.
(md_estimate_size_before_relax): Don't set fx_pcrel_adjust any
more.
(md_apply_fix3): Remember addend value for rela relocations.
(tc_gen_reloc): Correctly compute pc-relative relocation addend.
field for pc-relative fixups.
(output_disp): Likewise.
(md_estimate_size_before_relax): Likewise.
(tc_gen_reloc): Subtract fx_pcrel_adjust instead of fx_size for
pc-relative fixups in 64bit mode.
2002-03-15 Chris G. Demetriou <cgd@broadcom.com>
* config/tc-mips.c (mips_set_options): New "ase_mips3d" member.
(mips_opts): Initialize "ase_mips3d" member.
(file_ase_mips3d): New variable.
(CPU_HAS_MIPS3D): New macro.
(md_begin): Initialize mips_opts.ase_mips3d and file_ase_mips3d
based on command line options and configuration defaults.
(macro_build, mips_ip): Accept MIPS-3D instructions if
mips_opts.ase_mips3d is set.
(OPTION_MIPS3D, OPTION_NO_MIPS3D, md_longopts, md_parse_option):
Add support for "-mips3d" and "-no-mips3d" options.
(OPTION_ELF_BASE): Move to accomodate new options.
(s_mipsset): Support ".set mips3d" and ".set nomips3d".
(mips_elf_final_processing): Add a comment indicating that a
MIPS-3D ASE ELF header flag should be set, when one exists.
* doc/as.texinfo: Document -mips3d and -no-mips3d options.
* doc/c-mips.texi: Likewise, and document ".set mips3d" and ".set
nomips3d" directives.
[ gas/testsuite/ChangeLog ]
2002-03-15 Chris G. Demetriou <cgd@broadcom.com>
* gas/mips/mips64-mips3d.s: New file.
* gas/mips/mips64-mips3d.d: Likewise.
* gas/mips/mips.exp: Run new "mips64-mips3d" test.
[ include/opcode/ChangeLog ]
2002-03-15 Chris G. Demetriou <cgd@broadcom.com>
* mips.h (INSN_MIPS3D): New definition used to mark MIPS-3D
instructions.
(OPCODE_IS_MEMBER): Adjust comments to indicate that ASE bit masks
may be passed along with the ISA bitmask.
[ opcodes/ChangeLog ]
2002-03-15 Chris G. Demetriou <cgd@broadcom.com>
* mips-dis.c (mips_isa_type): Add MIPS3D instructions to the ISA
bit masks for bfd_mach_mips_sb1 and bfd_mach_mipsisa64. Add
comments for bfd_mach_mipsisa32 and bfd_mach_mipsisa64 that
indicate that they should dissassemble all applicable
MIPS-specified ASEs.
* mips-opc.c: Add support for MIPS-3D instructions.
(M3D): New definition.
* mips-opc.c: Update copyright years.
relocs except when BFD64.
* write.c (number_to_chars_bigendian): Don't abort when N is
larger than sizeof (VAL).
(number_to_chars_littleendian): Likewise.
(parse_ac5): New function for parsing float regs in float operand.
(parse_expression): Remove attempt to make literals be octal.
(parse_op_no_deferred): Support float literals.
(parse_op): Reject attempts to refer to float regs.
(parse_fop): New function, like parse_op but for float operand.
(md_assemble): Add cases to parse float operands. Also fix
IMM3, IMM6, IMM8 cases to pick up the operand from the right spot.
* config/tc-mips.c (mips_need_elf_addend_fixup): For embedded-PIC
only, undo the changes made on 2001-06-08, with the
effect being that common or extern symbols are
adjusted for embedded-PIC, but weak symbols are not.
(md_estimate_size_before_relax: Likewise, with the effect
that extern symbols are treated the same as weak symbols
only if not embedded-PIC.
(mips_fix_adjustable) Likewise, with the effect that
weak or extern symbols are not adjusted for embedded-PIC.
(md_apply_fix3): Tweak so that the case where value is zero
is handled more correctly for embedded-PIC code.
