3 * config/tc-i386.c (i386_immediate): Move constant operand sizing
5 (md_assemble): To here, before template operands are matched.
6 Also ensure a constant immediate is sign extended when we know the
7 size is at most 16 bits. This is to catch cases like "add
8 $0xffc0,%ax" where we don't know the size, and thus that the
9 immediate can be represented as Imm8S until after parsing the
11 (i386_displacement): Similarly sign extend 16 bit constant
13 (md_assemble): Relax 16-bit jump constant range check to suit sign
14 extended displacements.
18 * doc/c-mips.texi (MIPS Opts): Fix typo in last patch.
22 * config/tc-i386.c (md_assemble): Don't swap intersegment jmp and
23 call operands when intel_syntax.
24 (intel_float_operand): Return 2 for "fi...".
25 (i386_operand_modifier): Change "DWORD PTR" test to suit above.
26 Return SHORT_MNEM_SUFFIX for "WORD PTR" when "fi...". Revert
27 earlier "SHORT" change.
28 (md_assemble): When determining suffix from Regs, exclude
33 * configure: Add arm-wince, mips-pe and sh-pe targets.
34 * configure: Regenerate.
36 * config/obj-coff.h (COFF_WITH_PE): Define for mips-pe and
38 (TARGET_FORMAT): Set to "pe-shl" for the sh-pe target and to
39 "pe-mips" for the mips-pe target.
41 * config/tc-arm.c (insns): Change displacement encoded in BL
42 and B instructions if the target port is arm-wince.
43 (do_ldst): Do not bias the relocation offset if the target
45 (md_pcrel_from): Add in missing relocation offset bias if the
48 * config/tc-mips.c (mips_target_format): Support COFF flavour.
49 (md_begin): Disable -G support for mips-pe target.
50 (md_apply_fix): Treat BFD_RELOC_RVA reloc as BFD_RELOC_32.
51 * config/tc-mips.h (USE_GLOBAL_POINTER_OPT): Add support for
54 * config/tc-sh.c (md_begin): sh-pe target is little endian.
55 * config/tc-sh.h (SUB_SEGMENT_ALIGN): If using a BFD
56 assembler, just set the alignment to 4.
58 * config/te-wince-pe.h: New file for WinCE targets. Define
63 * config/tc-i386.c (md_assemble): Swap segments too for intel mode
65 (i386_operand_modifier): Set i.suffix = WORD_MNEM_SUFFIX for SHORT.
66 (i386_intel_memory_operand): After finding a segment override,
67 check again for no `[' before looking for a displacement. Bomb if
68 more than one displacement rather than silently discarding the
69 second and subsequent ones. Free strings malloc'd by
70 build_displacement_string.
74 * config/obj-som.c (obj_pseudo_table): Add "weak".
75 (obj_som_weak): New routine.
79 * config/tc-i386.c (union i386_op): New.
80 (struct _i386_insn): Delete disps[], imms[], regs[]. Add op[].
81 Throughout file replace occurences of disps[n], imms[n], regs[n]
82 with equivalent op[n].disps, op[n].imms, op[n].regs. Simplify
83 intel mode operand swapping. Add assert in regKludge and
84 fake_zero_displacement code. Test i.types[n] when outputting
85 displacements and immediates. Combine output of Disp16 with
87 (md_assemble): Don't try to fix broken UNIXWARE_COMPAT opcodes
88 when in intel mode by (not) reversing fsub and fdiv operands
89 before the template search. This fails for single operand
90 shorthand forms of the instruction, and if UNIXWARE_COMPAT is
91 undefined. Instead fix the base_opcode after we've found the
92 template. Move base_opcode xor with found_reverse_match from
93 opcode output code to before this fix so we test for the correct
95 (md_assemble): Don't use strcmp when deciding to ignore the suffix
96 check in intel mode. Instead compare opcodes.
98 * config/tc-i386.h (TC_RELOC): Delete.
99 * config/tc-i386.c (TC_RELOC): Delete. Replace usage of TC_RELOC
100 with equivalent call to reloc.
102 * as.h (flag_m68k_mri): Move declaration after target include, and
103 only declare when TC_M68K defined. Define as zero otherwise.
104 (LABELS_WITHOUT_COLONS, NO_PSEUDO_DOT): If undefined, define as 0.
105 * app.c (scrub_m68k_mri): Declare only when TC_M68K defined.
106 Define as zero otherwise.
107 (do_scrub_begin): Use m68k_mri parameter only when TC_M68K defined.
108 (struct app_save): Declare scrub_m68k_mri only when TC_M68K.
109 (app_push, app_pop): Save scrub_m68k_mri only when TC_M68K.
110 (do_scrub_chars): Use LABELS_WITHOUT_COLONS directly rather than
111 testing whether defined.
112 * cond.c (ignore_input): Use NO_PSEUDO_DOT directly.
113 * expr.c (operand): #ifdef unused case labels when TC_M68K undefined.
114 * read.c: Use LABELS_WITHOUT_COLONS and NO_PSEUDO_DOT directly
115 rather than testing whether defined.
116 (s_mri): Set flag_m68k_mri only when TC_M68K defined.
117 (parse_mri_cons): Declare and use only when TC_M68K.
118 * config/tc-hppa.h (LABELS_WITHOUT_COLONS): Define as 1.
119 * config/tc-m68k.h (NO_PSEUDO_DOT): Define as 1.
120 * config/tc-m88k.h (NO_PSEUDO_DOT): Define as 1.
122 * NEWS: Mention IBM 370 support.
126 * config/tc-i386.c (md_assemble): When swapping operands for
127 intel_syntax, assume everything that's not Imm or Disp is a
132 * config/tc-i370.c, config/tc-i370.h: New files.
133 * Makefile.am: Add support for Linux/IBM 370.
134 * configure.in: Likewise.
135 * app.c (do_scrub_begin): Don't lex single quote when TC_I370.
136 * config/obj-elf.c: Include elf/i370.h
137 (obj_elf_section): Don't do anything special for flag_mri if TC_I370.
139 * Makefile.in: Regenerate.
140 * configure: Regenerate.
142 * doc/c-i370.texi: New file.
143 * doc/all.texi: Include it.
144 * doc/as.texinfo: And here.
145 * doc/Makefile.am(CPU_DOCS): Add c-i370.texi.
146 * doc/Makefile.in: Regenerate.
150 * config/tc-d30v.c (parallel_ok): Use FLAG_NOT_WITH_ADDSUBppp to
151 determine if an instruction can be used in parallel with an ADDppp
152 or SUBppp instruction.
156 * doc/c-mips.texi (MIPS Opts): Document -mgp32 and -mgp64.
160 * config/tc-mips.c (mips_gp32): New variable.
161 (macro_build) Use mips_gp32.
163 (md_longopts): Add "-mgp32" and "-mgp64".
164 (md_parse_option): Add OPTION_GP32 and OPTION_GP64.
168 * config/obj-coff.c (add_lineno): Accept non-positive lineno with
169 warning, and bump it to 1.
174 * dwarf2dbg.c (print_stats): Add cast to force printf argument to
179 * config/tc-mips.c (MF_HILO_INSN): Define.
180 (mips_7000_hilo_fix): Declare.
181 (append_insn): Conditionally insert nops after an mfhi/mflo insn.
182 (md_parse_option): Check for 7000_HILO_FIX options.
183 (OPTION_M7000_HILO_FIX): Define.
184 (OPTION_NO_M7000_HILO_FIX): Define.
185 * doc/c-mips.texi (-mfix7000): Describe.
189 * listing.c (print_lines): Remove unused variable `end'.
191 * config/tc-i386.c (md_assemble): Use `reloc()' to select reloc
192 type for JumpInterSegment output. Use enum bfd_reloc_code_real for
193 reloc_type when BFD_ASSEMBLER.
194 (md_estimate_size_before_relax): Use enum bfd_reloc_code_real for
195 reloc_type when BFD_ASSEMBLER. Move common code out of switch
196 statement and quell signed vs. unsigned comparison warning.
200 * config/tc-d10v.c (find_opcode): Add a symbol's value to
201 the computed frag offset, rather than overwriting it.
205 * config/tc-sh.c ("elf/sh.h"): Include.
206 (sh_dsp, valid_arch, reg_x, reg_y, reg_efg): New static variables.
207 (md.begin): Initialize target_arch.
208 Only include opcodes in has table that match selected architecture.
209 (parse_reg): Recognize register names for sh-dsp.
210 (parse_at): Recognize post-modify addressing.
211 (get_operands): The leading space is now optional.
212 (get_specific): Remove FDREG_N support. Add support for sh-dsp
213 arguments. Update valid_arch.
214 (build_Mytes): Add support for SDT_REG_N.
215 (find_cooked_opcode): New function, broken out of md_assemble.
216 (assemble_ppi, sh_elf_final_processing): New functions.
217 (md_assemble): Use find_cooked_opcode and assemble_ppi.
218 (md_longopts, md_parse_option): New option: -dsp.
219 * config/tc-sh.h (elf_tc_final_processing): Define.
220 (sh_elf_final_processing): Declare.
224 * config/tc-hppa.c (pa_build_unwind_subspace): Use subseg_new to create
225 the unwinder subspace. Save the current seg/subseg before creating
230 * config/tc-mcore.c (INST_BYTE0): Redefine to handle big and
231 little endian targets.
232 (INST_BYTE1): Redefine to handle big and little endian
234 (cpu_type): New type: Select between M340 and M210.
235 (parse_psrmod): New function: Parse the PSRCLR and PSRSET
236 instructions of the M340.
237 (md_assemble): Add support for the MULSH and OPSR classes of
239 (md_atof): Add support for little endian targets.
240 (md_parse_option): Add support for -EL, -EB and -mcpu command
242 (md_convert_frag): Add support for little endian targets.
243 (md_apply_fix3): Add support for little endian targets.
244 (md_number_to_chars): Add support for little endian targets.
248 * read.c (read_a_source_file): If TC_START_LABEL_WITHOUT_COLON is
249 defined, use it to verify the symbol just read should be a label.
253 * app.c (do_scrub_chars): Handle "||" for parallel instructions
254 when DOUBLEBAR_PARALLEL is defined. Avoid stripping whitespace
255 around colons when KEEP_WHITE_AROUND_COLON is defined.
256 * doc/internals.texi (CPU backend): Document DOUBLEBAR_PARALLEL
257 and KEEP_WHITE_AROUND_COLON.
261 * read.c (s_rept): Call do_repeat, which abstracts the repeat
263 (do_repeat): New. Abstract repeat logic so that a "break" can be
265 (end_repeat): New. Provide support for a "break" out of the
267 * read.h: Add prototypes for new functions.
271 * doc/internals.texi: Document NUMBERS_WITH_SUFFIX macro.
272 * as.h: Provide a default NUMBERS_WITH_SUFFIX definition (zero).
273 * expr.c: Handle numbers with suffixes if NUMBERS_WITH_SUFFIX is
278 * read.c: Added elseif to directives table.
279 * read.h: Added prototype for s_elseif.
280 * doc/as.texinfo: Added description for elseif.
281 * cond.c (s_elseif): New function
285 * listing.c (print_lines): Remove conditionals causing bug in
290 * as.h: Define OCTETS_PER_BYTE and OCTETS_PER_BYTE_POWER
292 * frags.c (frag_new): Calculate fr_fix in octets
293 (frag_now_fix) Return offset as target address offset (bytes).
294 (frag_now_fix_octets) New - Return offset in octets (8-bit
296 * frags.h: Added prototype for frag_now_fix_octets().
297 Distinguish between octets and bytes in field descriptions.
298 * listing.c (calc_hex): Account for octets vs bytes when
299 printing addresses/offsets.
300 (print_lines) Ditto. Also, if LISTING_WORD_SIZE is not 1, and
301 target is little-endian, print the octets in a word in big-endian
302 order so that the display looks like a proper hexadecimal number,
303 instead of having the octets reversed.
304 * read.c (do_align): When recording alignment, alignment power
305 should be in terms of target bytes (minimum addressible unit)
307 (do_org) Convert ORG target address (byte) argument into an
308 octet offset when generating a variable fragment.
309 * symbols.c (resolve_symbol_value): Symbol final value
310 converted to a target address offset (bytes) from its octet offset.
311 * config/obj-coff.c (coff_frob_symbol): Symbol target address
312 offset (bytes) is adjusted by the frag offset (octets) converted
314 (coff_frob_section) Section alignment power is in terms of bytes;
315 convert it to an octet alignment power when calculating size (and
316 size mask) in octets. Don't modify the section size in order to
317 "align" it for TI COFF, since that format has a different method
318 for storing alignment information.
322 * stabs.c (generate_asm_file): Escape backslashes in stabs file
323 entries, matching the way GCC generates them. If not escaped, the
324 filename is encoded incorrectly.
327 * config/tc-arm.c (reg_table): Add support for ATPCS register
332 * config/obj-coff.h (OBJ_COPY_SYMBOL_ATTRIBUTES): Don't define if
334 * config/tc-ppc.h [OBJ_XCOFF] (OBJ_COPY_SYMBOL_ATTRIBUTES):
336 * config/tc-ppc.c (ppc_fix_adjustable): Don't look at the frag
337 of a symbol when we really care about its value.
341 * config/tc-mcore.c (md_assemble): Give warning message if
342 operands passes to instruction are more than the spec.
346 * config/tc-arm.c (armadjust_symtab): If the assembler is in
347 Thumb mode but the label seen was not declared as '.thumb_func'
348 then set the ST_INFO type to STT_ARM_16BIT mode. This allows
349 correct disassembly of Thumb code bounded by non function labels.
353 * Makefile.am (MULTI_CFILES): Add config/e-i386aout.c
354 Add dependencies for e-i386aout.o. Fix 2 comment lines.
356 * Makefile.in: Same here.
359 * configure.in: Set bfd_gas for i386-aout when primary target
360 is bfd. Handle i386aout emulation. Don't use te_file=multi, as
361 we may need the primary te_file. Remove incorrect comment.
363 * configure: Regenerate.
365 * config/e-i386aout.c: New file.
367 * as.c (USE_EMULATIONS): Move to before print_version_id.
368 (struct emulation): Add i386aout.
369 (show_usage): Split text strings. Reformat -a text. Add --em
373 * obj.h (struct format_ops): Add s_get_other and s_get_desc.
374 (aout_format_ops): New.
377 * read.c (s_lcomm_internal): Rewrite OBJ_AOUT,OBJ_BOUT
378 preprocessor conditional and add aout USE_EMULATIONS tests.
379 (read_a_source_file): Don't pass error strings to printf as
383 * gasp.c (exp_get_abs): Don't pass error strings to printf as
385 (do_data): Same here.
386 (process_file): And here.
389 * symbols.c (colon): Rewrite "already defined" fatal message
390 code for aout with USE_EMULATIONS.
393 * config/obj-aout.c (OBJ_HEADER): Define.
394 (obj_pseudo_table): Rename to aout_pseudo_table. Init all
396 (obj_aout_frob_symbol): Expand S_GET_DESC, S_GET_TYPE,
397 S_GET_OTHER, S_SET_TYPE macros since we don't need obj-multi
399 (obj_aout_type): Expand S_SET_OTHER here too.
400 (obj_read_begin_hook): Remove.
401 (aout_pop_insert): New.
402 (obj_aout_s_get_other): New.
403 (obj_aout_s_get_desc): New.
404 (aout_format_ops): New.
407 * config/obj-aout.h (obj_pop_insert): Define so non-multi usage
408 gets aout_pseudo_table.
409 (aout_pseudo_table): Declare.
410 (obj_read_begin_hook): Define.
413 * config/obj-coff.c (obj_pseudo_table): Rename to
415 (coff_pop_insert): Use coff_pseudo_table.
416 (coff_sec_sym_ok_for_reloc): Remove.
417 (coff_format_ops): Add 0 entries for s_get_size, s_set_size,
418 and comment all zero entries and remove #if 0 code.
421 * config/obj-coff.h (obj_pop_insert): Define.
422 (coff_pseudo_table): Declare.
425 * config/obj-ecoff.c (ecoff_format_ops): Add 0 entries for
426 s_get_size, s_set_size. Comment all zero entries.
429 * config/obj-elf.c (elf_s_get_other): New function.
430 (obj_read_begin_hook): Rename to elf_obj_read_begin_hook.
431 (obj_symbol_new_hook): Rename to elf_obj_symbol_new_hook.
432 (elf_format_ops): Add elf_s_get_other, 0 s_get_size entry, and
434 (obj_elf_parse_section_letters): Don't pass error strings to
435 printf as format arg.
438 * config/obj-elf.h (ECOFF_DEBUGGING): Define when
440 (elf_s_get_other): Declare.
441 (S_GET_OTHER) Define as elf_s_get_other if not already
443 (S_SET_OTHER): Only define when not already defined.
444 (elf_obj_read_begin_hook): Declare.
445 (obj_read_begin_hook): Define.
446 (elf_obj_symbol_new_hook): Declare.
447 (obj_symbol_new_hook): Define.
450 * config/obj-multi.h: Add copyright header and protect against
451 multiple inclusion. Add * to all function pointers.
452 (OBJ_HEADER): If defined, include it rather than other defines
454 (obj_frob_file_after_relocs): Test for NULL.
455 (obj_symbol_new_hook): Here too.
456 (obj_sec_sym_ok_for_reloc): And here.
457 (S_GET_OTHER): Define.
458 (S_GET_DESC): Define.
459 (ECOFF_DEBUGGING): Remove as it's done in obj-elf.h
460 (OBJ_MAYBE_ELF): Update comment.
462 * config/tc-i386.c (i386_immediate): Add OBJ_MAYBE_AOUT to
463 OBJ_AOUT preprocessor conditional and handle emulation by
464 testing OUTPUT_FLAVOR.
465 (i386_displacement): Here too.
466 (md_section_align): Similarly here.
467 (i386_target_format): Conditionally compile when more than one
468 of OBJ_MAYBE_{ELF,COFF,AOUT} defined. Add aout case.
469 (i386_immediate): Fix error message for aout BFD_ASSMBLER.
470 (i386_displacement): Here too.
473 * config/tc-i386.h (AOUT_TARGET_FORMAT): Define for each TE_*.
474 Define TARGET_FORMAT for aout only when not multi.
477 * config/te-multi.h: Delete file as it's identical to te-generic.h
481 * config/tc-i386.h (DWORD_MNEM_SUFFIX): Delete.
482 * config/tc-i386.c (DWORD_MNEM_SUFFIX): Rename all occurrences to
485 * config/tc-i386.h (INTEL_DWORD_MNEM_SUFFIX): Rename to
487 * config/tc-i386.c (INTEL_DWORD_MNEM_SUFFIX): Here too. Fix some
492 * config/tc-mips.c (mips_do_align): New function.
493 * config/tc-mips.h (md_do_align): Define.
497 * doc/c-arm.texi (ARM Options): Fix typo.
498 (ARM-Chars): Correct description of `#'. Mention that `;' is a
499 line separator for Linux.
500 * doc/as.texinfo (Comments): Mention the ARM.
504 * configure.in (arm*-*-conix*): New target.
505 (arm*-*-linux-gnu*): Match instead of arm-*-linux* and
507 * configure: Regenerate.
511 * config/obj-elf.c (elf_pseudo_table): Define visibility pseudos.
512 (obj_elf_visibility): New function.
514 * doc/as.texinfo (Visibility): New node: document visibility
519 * config/tc-i386.c (MATCH): Relax JumpAbsolute check. Emit a
520 warning for absolute jump/call without `*' in non-intel mode. No
521 need to set i.types[0] JumpAbsolute in intel mode.
525 * config/tc-arm.c (arm_s_text): If OBJ_ELF, call the appropriate
526 hook function when changing sections.
527 (arm_s_data): Likewise.
531 * config/tc-arm.c (md_parse_option): Add support for -marm720
536 * config/tc-mn10300.c (md_pseudo_table): Add ".am33" pseudo-op.
537 (r_registers, xr_registers): Define.
538 (r_register_name, xr_register_name): New functions.
539 (md_assemble): Handle new am33 operand types and instruction
541 (mn10300_insert_operand, check_operand): Likewise.
545 * config/tc-arm.c (thumb_mode): Turn into a tristate variable.
546 (s_force_thumb): Set thumb_mode to 2.
547 (md_assemble): Do not complain about thumb instructions on a
548 non-thumb target if thumb_mode is set to 2.
552 * config/tc-alpha.c (toplevel): Include struc-symbol.h.
553 (alpha_macro_arg): Add MACRO_{LITERAL,BASE,BYTOFF,JSR} cases.
554 (O_...): Add new machine dependent expressions if we are handling
555 explicit relocations.
556 (alpha_reloc_op): New static table holding the explicit relocation
558 (alpha_literal_hash): New static to hold the hash table for
559 explicit relocations.
560 (alpha_macros): Add support for explicit relocations.
561 (md_begin): If explicit relocations, initialize hash table.
562 (md_assemble): Don't print a second error if tokenize_arguments
563 already printed an error message.
564 (md_apply_fix): Add support for explicit relocations.
565 (alpha_force_relocation): Ditto.
566 (alpha_fix_adjustable): Ditto.
567 (alpha_adjust_symtab): New function to support explicit
569 (alpha_adjust_symtab_relocs): Ditto.
570 (debug_exp): Debug stub compiled if DEBUG_ALPHA is defined.
571 (tokenize_arguments): Add debug code if DEBUG_ALPHA is defined.
572 Add support for explicit relocations. Return -2 if an error
573 message was already printed.
574 (find_macro_match): Add support for explicit relocations. Comment
576 (emit_insn): Add support for explicit relocations.
577 (assemble_tokens): Ditto.
579 (load_expression): Ditto.
582 (emit_ir_load): Ditto.
583 (emit_loadstore): Ditto.
587 (emit_division): Ditto.
588 (emit_jsrjmp): Ditto.
589 (emit_retjcr): Ditto.
591 * config/tc-alpha.h (RELOC_OP_P): Enable explicit relocations if
593 (tc_adjust_symtab): If explicit relocations, call the function
595 (TC_FIX_TYPE): Add fields to be able to move explicit lituse
596 relocations next to the literal relocation they reference.
597 (TC_INIT_FIX_DATA): Initialize the new fields.
598 (TC_FIX_DATA_PRINT): Print the new fields if DEBUG5 is defined.
602 * config/tc-hppa.c (pa_ip): Handle PA2.0 unit completers. Handle
603 'B' operand for PA2.0 bb instruction.
607 * config/tc-mcore.h (TC_FORCE_RELOCATION): Define for Mcore-pe
610 * config/tc-mcore.c (tc_gen_reloc): Support generation of RVA
612 (mcore_force_relocation): Force relocations to be generated for
617 * config/tc-i386.c (i386_immediate): Disallow O_big immediates.
618 (i386_displacement): Disallow O_big displacements.
622 * config/tc-arm.c (do_mia,do_mar,do_mra,do_pld,do_ldrl,do_co_reg): Small
623 improvements in error checking.
624 (md_assemble): Support for unconditional ARM instructions.
625 (md_parse_option): Support for -m[arm]v5e flag.
629 * macro.c (buffer_and_nest): Do not check beyond the end of the
634 * macro.c (buffer_and_nest): Look for seperator after TO and
639 * app.c (do_scrub_chars): When in State 10, treat backslash
640 characters in the same way as as symbol characters.
644 * config/tc-alpha.c (alpha_align): Check, don't assert, that
645 the previous label was in the current section before playing
650 * config/tc-v850.c (v850_force_relocation): Force relocation
652 (v850_pcrel_from_selection): Do not compute a pcrel offset if
657 * expr.h (operatorT): Increase machine dependent operators to 16.
658 * expr.c (op_rank): Ditto.
662 * read.c (pseudo_set): Reject attempts to set the value of a
665 * config/obj-elf.c (obj_elf_ident): Call md_flush_pending_output
668 * config/obj-elf.c (elf_set_index): Add ATTRIBUTE_UNUSED.
670 * config/obj-elf.c (elf_frob_file_after_relocs): Don't pass NULL
671 to bfd_set_section_contents.
675 * config/tc-mcore.h (TARGET_BYTES_BIG_ENDIAN): Change to false.
679 * config/tc-mips.c (macro_build): Use OPCODE_IS_MEMBER.
680 (mips_ip): Use OPCODE_IS_MEMBER.
684 * config/tc-arm.c (reg_required_here): Improve comments.
686 * config/tc-arm.c (thumb_opcode): Add "variants" field.
687 (tinsns): Initialize variants field.
689 * config/tc-arm.c (bad_args, bad_pc): Renamed to BAD_ARGS and
694 * config/tc-arm.c (reloc_map[]): Fix compiler warning.
695 * config/tc-arm.h: Fix compile time warnings.
699 * tc_d10v.c (find_opcode): Allow ATSIGN to match expressions of the
700 form @abs16, @(abs16) and @(abs16 + imm).
704 * config/tc-mips.c (ISA_HAS_COPROC_DELAYS) : New.
705 (ISA_HAS_64_BIT_REGS) New.
706 (gpr_interlocks,md_begin,reg_needs_delay,append_insn,
707 mips_emit_delays,macro_build,load_register,load_addresss,
708 macro,macro2,mips_ip,s_cprestore,s_cpadd): Simplify
709 and/or use new ISA_xxx macros in expressions involving
710 ISA, particularly mips_opts.isa.
714 * expr.h (operatorT): Add machine dependent operators md1..md8.
715 (expressionS): Make X_op 8 bits instead of 7. Add a X_md field
716 for the machine dependent operators to use.
718 * expr.c (op_rank): Add machine dependent operators.
720 * config/tc-alpha.c (O_pregister): Define as a machine dependent
722 (O_cpregister): Ditto.
723 (md_begin): Change X_op test that field is wide enough to use
724 O_max instead of O_alpha_max.
725 (cpu_types): Fill in missing initializer.
726 (alpha_num_macros): Make unsigned.
727 (md_assemble): Make opnamelen be size_t.
728 (md_apply_fix): Cast alpha_num_operands to int before testing.
729 (alpha_force_relocation): Ditto.
730 (alpha_fix_adjustable): Ditto.
731 (alpha_fix_adjustable): Mark unused arguments ATTRIBUTE_UNUSED.
732 (tc_gen_reloc): Ditto.
733 (tc_get_register): Ditto.
738 (s_alpha_ent): Ditto.
739 (s_alpha_end): Ditto.
740 (s_alpha_frame): Ditto.
741 (s_alpha_prologue): Ditto.
742 (s_alpha_file): Ditto.
743 (s_alpha_gprel32): Ditto.
744 (s_alpha_proc): Ditto.
745 (s_alpha_set): Ditto.
746 (s_alpha_base): Ditto.
747 (s_alpha_align): Ditto.
748 (s_alpha_arch): Ditto.
749 (alpha_align): Ditto.
750 (assemble_insn): Suppress unused variable warning.
752 (assemble_insn): Don't assume X_op and X_unsigned are in a given
753 order in the structure.
754 (s_alpha_coff_wrapper): Avoid int/unsigned comparison.
758 * config/tc-hppa.c (md_apply_fix): Make "fmt" an int.
762 * config/tc-i386.c (i386_index_check): Correct #endif location.
766 * as.c (show_usage): Document new options.
767 (parse_args): Add --no-warn, --warn, --fatal-warnings,
768 which become 'W', OPTION_WARN, and OPTION_WARN_FATAL.
769 (parse_args): Parse the new options.
770 (main): If there were warnings, and --fatal-warnings
771 was specified, print an error.
772 * as.h: New variable, flag_fatal_warnings, for new option.
776 * config/tc-hppa.c (pa_ip): Add new codes 'cc', 'cd', 'cC', 'co',
777 '@'. Change autoincrement completers to fall through to cache control
780 * config/tc-hppa.c (pa_ip): Remove unused args. Add code to '?W'
782 (pa_parse_addb_64_cmpltr): New function.
784 * config/tc-hppa.c (pa_ip): Change error message.
785 (pa_ip,pa_parse_cmpb_64_cmpltr,pa_parse_cmpib_64_cmpltr) Fix '?N' and
786 '?Q' args to allow falling through.
788 * configtc-hppa.c (pa_ip): Implement conditional codes "?N", "?Q".
789 Remove unused conditional codes.
790 (pa_parse_cmpb_64_cmpltr,pa_parse_cmpib_64_cmpltr): New.
794 * config/tc-d30v.c (CHAR_BIT): Define.
795 (check_range): Fix bit operations to support integers bigger than
800 * config/tc-d10v.c (check_range): Check range for RESTRICTED_NUM3
806 Add support for m32rx.
807 * config/tc-m32r.c (enable_m32rx): New static global.
808 (enable_special,warn_explicit_parallel_conflicts,optimize): Ditto.
809 (allow_m32rx): New function.
810 (M32R_SHORTOPTS): Add `O'.
811 (md_longopts): Add --m32rx plus several warning options.
812 (md_parse_option): Handle new options.
813 (md_show_usage): Print them.
814 (md_begin): Enable m32rx.
815 (OPERAND_IS_COND_BIT): New macro.
816 (first_writes_to_seconds_operands): New function.
817 (writes_to_pc,can_make_parallel,make_parallel): New functions.
818 (target_make_parallel,assemble_two_insns): New functions.
819 (md_assemble): Recognize "insn1 -> insn2" and "insn1 || insn2".
820 If optimizing and m32rx, try to make consecutive insns parallel.
824 * config/tc-mips.c (nopic_need_relax): Allow for the
825 .sdata.foo sections generated by -fdata-sections,
826 and for the .gnu.linkonce.s sections generated by C++.
830 * config/tc-hppa.c (pa_ip): Replace 'B', 'M', 'l' and 'g' handling
831 with cleaner code using completer prefixes. Add 'Y'.
833 * config/tc-hppa.c (pa_ip): Add parens to silence compiler.
837 * config/tc-hppa.c (pa_ip): Avoid ANSI specific initialization.
838 (pa_ip, case 'm'): Failure to get a CBIT specifier just means the
839 insn does not match and we should try the next insn in the table.
843 * config/tc-arm.c (MULTI_SET_PSR): Rename to LDM_TYPE_2_OR_3.
847 * config/tc-hppa.c (pa_ip): Fix thinkos in recent cleanup
852 * config/tc-i386.c (md_shortopts): Check OBJ_MAYBE_ELF as well as
853 OBJ_ELF. If ELF, add "sq".
854 (md_parse_option): If ELF, ignore -s and -q.
855 (md_show_usage): Mention ELF options.
859 * config/tc-hppa.c (pa_ip): Handle 'J', 'K' and 'cc'
862 * config/tc-hppa.c (pa_ip); Handle "fe", and 'cJ'.
864 * config/tc-hppa.c (pa_ip): Handle 'd', '#' and 'cq'.
866 * config/tc-hppa.c (struct pa_it): New field "trunc".
867 (pa_ip): Hadnle 'h', 'm', '=', '{', and '_' operands.
868 (pa_parse_ftest_gfx_completer): New function
869 (pa_parse_fp_cnv_format): New function.
871 * config/tc-hppa.c (pa_ip): Handle 'X' operand.
872 (md_apply_fix): Handle 22bit pc-rel branches.
874 * config/tc-hppa.c (pa_ip): Handle 'B' operand.
876 * config/tc-hppa.c (pa_ip): Handle 'L' and 'M' operands.
878 * config/tc-hppa.c (pa_ip): Handle 'l' operand.
880 * config/tc-hppa.c (pa_ip): Handle 'g' operand.
884 * config/tc-hppa.c (md_assemble): Fix dwarf2 line handling.
885 (pa_ip): Handle 'fX'.
889 * config/tc-hppa.c (md_assemble): Fix dwarf2 line handling.
890 (pa_ip): Handle 'fX'.
894 * config/tc-hppa.c (pa_build_unwind_subspace): Do not build
895 unwinds unless the function is in the text space.
896 (pa_type_args): Set BSF_FUNCTION for an exproted data symbol.
900 * config/tc-hppa.c (pa_ip): Move dwarf2 stuff from here.
901 (md_assemble): To here. Tweak address generation.
903 * config/tc-hppa.c: Include dwarf2dbg.h if OBJ_ELF. Declare
905 (md_pseudo_table): Add .file and .line pseudo-ops for OBJ_ELF.
906 (md_assemble): Call dwarf2_where for OBJ_ELF.
907 (pa_ip): Call dwarf2_gen_line_info for OBJ_ELF.
908 (pa_end_of_source): New function.
909 * tc-hppa.h (md_end): Define for OBJ_ELF.
913 * configure.in (Canonicalization of target names): Remove adding
914 ${CONFIG_SHELL} in front of $ac_config_sub, since autoconfig 2.14
915 generates $ac_config_sub with a ${CONFIG_SHELL} already.
916 * configure: Regenerate.
920 * config/te-interix.h (GLOBAL_OFFSET_TABLE_NAME): Define.
924 * config/tc-i386.c (md_assemble): Handle "jmp/call constant" as a
925 pc-relative jmp/call to an absolute symbol.
926 (md_apply_fix3): When OBJ_ELF, don't add the values in twice for
927 absolute section symbols.
929 * config/tc-i386.c (md_assemble): Correct frag_var size. Tidy
930 jump handling code and comments.
934 * config/tc-i386.c (md_apply_fix3): Add horrible adjustments to
935 the value if TE_PE and a global defined symbol.
939 * write.c (dump_section_relocs): Call print_symbol_value_1 to
940 print the symbol, rather than printing it here.
944 * config/tc-i386.c (tc_i386_fix_adjustable): Don't adjust
945 BFD_RELOC_RVA relocations.
947 * config/tc-i386.c (md_undefined_symbol): Compare the name against
948 the macro GLOBAL_OFFSET_TABLE_NAME, rather than assuming that it
951 * write.c (write_relocs): Call SET_SECTION_RELOCS if it is
953 * config/obj-coff.h (SET_SECTION_RELOCS): Define.
954 * doc/internals.texi (Object format backend): Document
957 * config/tc-i386.c (tc_i386_fix_adjustable): Don't adjust
958 relocations against global symbols if TE_PE.
960 * config/obj-coff.c (obj_coff_ident): Add BFD_ASSEMBLER version.
961 (obj_pseudo_table): Always handle ".ident" with obj-coff_ident.
963 * config/obj-coff.c (coff_frob_symbol): Prohibit weak common
966 * config/obj-coff.c (obj_coff_endef): Don't merge labels, or
967 symbols which do not have a constant value, or tags with
968 non-tags. Remove the symbol from the list before adding it at the
971 * config/obj-coff.c (obj_coff_endef) [BFD_ASSEMBLER]: Handle .ef
972 C_FCN symbol differently if TE_PE.
973 (obj_coff_line) [BFD_ASSEMBLER]: Always use the line number which
974 appears in the pseudo-op, rather coff_line_base which is only set
977 * config/obj-coff.c (obj_coff_loc): New static function.
978 (obj_pseudo_table): Add "loc".
980 * config/obj-coff.c (add_lineno): Check that the line number is
983 * config/atof-ieee.c (atof_ieee): Change what_kind to int.
984 * config/atof-vax.c (flonum_gen2vax): Change format_letter to
986 (md_atof): Return NULL rather than 0.
987 * config/tc-i386.c (md_atof): Change type to int.
988 * expr.c (expr): Change first parameter to int.
989 * config/obj-coff.c: Add declarations for static functions.
990 (coff_frob_symbol): Use SYM_AUXENT.
991 * config/tc-i386.h (flag_16bit_code): Don't declare.
993 * config/obj-coff.c (obj_coff_section): Default to setting
994 SEC_LOAD. Don't set SEC_DATA for 'w' modifier.
996 * write.c (adjust_reloc_syms): Print adjusted fixup.
998 * expr.c (integer_constant): Correct too_many_digits calculation
1003 * doc/c-arm.texi: Fix arguments of @var to not contain
1008 * config/tc-arm.c (s_thumb_set): Only support interworking for ELF
1010 (md_parse_option): Only support -k flag for ELF and COFF targets.
1014 * config/tc-hppa.c (pa_ip): Put strict register checks before
1015 call to pa_parse_number.
1017 * config/tc-hppa.c (pa_ip): Support 'Z' argument.
1021 * config/obj-coff.c: Add ATTRIBUTE_UNUSED as needed for
1026 * as.c (perform_an_assembly_pass): Set SEC_DATA for data_section.
1030 * config/tc-hppa.c (pa_ip): Get strict/non-strict mode from the
1031 candidate instruction. Require registers for register arguments
1032 when in strict mode. Require assemble-time constants for
1033 constants when in strict mode.
1034 (pa_get_absolute_expression): Require a constant when in strict
1039 * config/tc-m32r.c (md_longopts): Fix value for -Wnuh.
1043 * config/tc-pj.c: New file, supports picoJava in ELF.
1044 * config/tc-pj.h: Ditto.
1045 * configure.in (pjl*, pj*): New targets.
1046 * Makefile.am: Rebuild dependencies.
1047 (CPU_TYPES): Add pj.
1048 (TARGET_CPU_CFILES): Add config/tc-pj.c.
1049 (TARGET_CPU_HFILES): Add config/tc-pj.h.
1050 * doc/c-pj.texi: New file.
1051 * doc/as.texinfo: Add some PJ specifics.
1052 * doc/all.texi: Add PJ to the list of all architectures, sort them
1054 * doc/Makefile.in (CPU_DOCS): Add c-pj.texi.
1055 * configure, Makefile.in, doc/Makefile.in: Rebuild.
1059 * config/obj-multi.h: Include obj-elf.h if OBJ_MAYBE_ELF. Reformat.
1060 (obj_frob_file): Test for null pointer.
1061 (OBJ_COPY_SYMBOL_ATTRIBUTES): Here too.
1062 (OBJ_PROCESS_STAB): And here.
1063 (elf_obj_sy): Remove
1065 * config/obj-elf.h: #ifndef everything defined in obj-multi.h,
1066 except OBJ_PROCESS_STAB, which we #undef for ecoff.
1067 (elf_obj_sy): Remove #ifndef OBJ_SYMFIELD_TYPE.
1069 * config/obj-coff.c (no_func): Remove.
1070 (coff_format_ops): Change occurrences of no_func to 0, as we test
1071 for 0 in obj-multi.h.
1073 * configure.in: Enable bfd for i386-coff when primary target is
1074 bfd. Enable i386 elf,coff emulation support. Don't set
1075 USE_EMULATIONS=1 or te_file=multi unless there is more than one
1076 emulation to support.
1077 *configure: Regenerate.
1081 * config/tc-mcore.c (mcore_s_section): Do not dump literals if a
1082 .section .line directive is encountered.
1086 * config/tc-arm.c (md_section_align): Do not align sections in ELF
1089 * as.c (show_usage): Add --gdwarf2 to list of options displayed.
1090 * as.texinfo: Document --gdwarf2 command line option.
1091 Add additional documentation of ARM command line switches.
1095 * config/tc-i386.c (i386_intel_memory_operand): Combine
1096 i386_is_reg and parse_register calls. Remove END_STRING_AND_SAVE
1097 and RESTORE_END_STRING around parse_register calls.
1098 (i386_operand): Here too.
1099 (i386_is_reg): Remove.
1100 (parse_register): Move as_bad calls from within this function to
1106 * config/tc-i386.c (parse_register): Handle FP regs specially.
1107 (md_begin): Remove '(' and ')' from register_chars.
1111 * config/tc-m32r.c (md_parse_option): Delete unrecognized option
1112 error message (done elsewhere).
1116 * config/tc-hppa.c (pa_ip): Do not allow '*' in 32bit completers.
1120 * config/tc-hppa.c (pa_ip): Replace 'f' by 'v'. Prefix float register
1123 * config/tc-hppa.c (pa_ip): Add args q, %, and |.
1125 * config/tc-hppa.c (pa_ip): Absorb white space in instructions
1127 Add new completers. Fix bug in 64 bit condition handling.
1129 * config/tc-hppa.c (pa_ip): Add completer codes 'a', 'ch', 'cH',
1132 * config/tc-hppa.c (pa_ip): Place completers behind prefix 'c'.
1134 * config/tc-hppa.c (pa_ip): Add cases for '.', '~'. '$'. and '!'
1136 * config/tc-hppa.c (pa_ip): Add case for 'I'.
1140 * dwarf2dbg.c (MAX_SPECIAL_ADDR_DELTA): Correct typo in comment.
1141 (struct ls): Add frag field. Initialize it to zero.
1142 (out_end_sequence): New local text_frag. Set it while in text section.
1143 Replace address check with frag check. Set ls.frag to text_frag if
1144 out_set_addr called.
1145 (dwarf2_gen_line_info): Add explanatory comment. New local saved_frag.
1146 Set it before switching sections. Replace address check with frag
1147 check. Set ls.frag to saved_frag if out_set_addr called.
1151 * dwarf2dbg.c (out_end_sequence): If address changed, directly
1152 output "advance_pc" opcode instead of calling gen_addr_line().
1153 The latter has the undesired side-effect of creating a new row
1154 in the debug line info matrix.
1158 * dwarf2dbg.c (out_end_sequence): Correct comments. Set last to
1159 ls.last_filename if last is less than zero. Set ls.last_filename
1160 when allocating new entry.
1161 (dwarf2_gen_line_info): Save seg and subseg info before subseg_new
1166 * config/tc-i386.c (i386_index_check): Fix the displacement size
1167 when INFER_ADDR_PREFIX.
1171 * config/tc-arm.c (md_apply_fix3): If an offset is invalid,
1176 * config/tc-ppc.c (md_assemble): Trim @ha constant to 16 bits, to
1177 handle 0xffffNNNN constants correctly.
1181 * config/tc-arm.c (do_ldst): Look for register conflicts on stores
1186 * config/tc-arm.c (validate_offset_imm): Work on unsigned values.
1187 (md_apply_fix3): Always pass positive values to
1188 validate_offset_imm.
1192 * config/tc-arm.c (skip_whitespace): New macro.
1193 Formatting tidy ups.
1195 (md_apply_fix3): Store relocation offset in addend for ELF based
1197 (arm_force_relocation): Always generate relocs for Thumb function
1202 * config/tc-i386.c (md_assemble): Remove dead code. intel_syntax
1203 LONG_DOUBLE_MNEM_SUFFIX floating point is done in opcode/i386.h
1207 * config/tc-hppa.c (pa_build_unwind_subspace): Use sane section
1208 flags for the unwind subspace.
1210 * config/tc-hppa.c (UNWIND_SECTION_NAME): Define for ELF.
1211 (pa_build_unwind_subspace): Remove #if 0 wrapper. Select a
1212 suitable relocation based on the size of the target's pointer.
1213 Always Use subsegment zero for the unwinders.
1214 (pa_level): Handle "2.0w".
1218 * config/tc-d30v.c (write_2_short): Don't group repeat instructions
1219 with the following instruction unless this was specified.
1223 * config/tc-i386.h (SUB_SEGMENT_ALIGN): If TE_GO32, return 4 for
1224 certain sections, to match BFD changes.
1228 * Makefile.am (noinst_SCRIPTS): Change .gdbinit to $(GDBINIT).
1229 (EXTRA_SCRIPTS): Define to keep automake happy.
1230 * Makefile.in: Rebuild.
1234 * Makefile.am: Rename .dep* files to DEP*.
1235 (MKDEP): Rename from DEP. Change all uses. Use $${srcdir} rather
1236 than $(srcdir). Rename TCDEP targets to DEPTC. Rename OBJDEP
1238 * Makefile.in: Rebuild.
1242 * config/tc-sparc.c (sparc_ip): Allow assembly of %lo()+%reg.
1246 * Makefile.am: Change all uses of itbl-test-ops to itbl-tops to
1247 avoid problems on DOS filesystems.
1248 * Makefile.in: Rebuild.
1250 * doc/as.texinfo (Section): Document 's' flag for COFF version.
1254 * config/obj-coff.c (obj_coff_section): Handle 's' (shared)
1259 * configure.in: Define and substitute GDBINIT. Change AC_OUTPUT
1260 line to create ${GDBINIT} rather than .gdbinit.
1261 * configure, Makefile.in, doc/Makefile.in: Rebuild.
1265 * config/tc-hppa.c (pa_ip, case '?'): Add missing break.
1269 * config/tc-hppa.c (pa_ip): Add 64 bit condition completers.
1273 * config/tc-sparc.h (tc_fix_adjustable): Fix check for PIC local
1274 non-adjustable symbols.
1278 * config/tc-hppa.c (pa_ip): Change condition args to have '?' prefix.
1282 * config/tc-sh.c (md_assemble): Call as_bad when there are excess
1287 * config/te-interix.h: New file.
1288 * configure.in (i386-*-interix*): New target.
1289 * configure: Rebuild.
1293 * config/tc-hppa.c (pa_chk_field_selector): Allow 3 byte
1294 selectors for ELF too.
1295 (selector_table): Add "ltp" and "rtp" selectors.
1299 * config/tc-i386.c (i386_operand): No need to change
1300 operand_string pointer in segment reg case before goto
1301 do_memory_reference. Initialise displacement_string_start and
1302 displacement_string_end after do_memory_reference label.
1303 (i386_index_check): Add operand_string param, and print error
1304 message on failure here.
1305 (i386_intel_memory_operand): Instead of here.
1306 (i386_operand): And here.
1307 (INFER_ADDR_PREFIX): Enable.
1309 * doc/c-i386.texi (i386-16bit): Document .code16gcc.
1311 * config/tc-i386.h (DefaultSize): Define. Renumber following
1312 opcode_modifier defines.
1315 * config/tc-i386.c (stackop_size): New variable.
1316 (set_16bit_code_flag): Clear it here.
1317 (set_16bit_gcc_code_flag): New function.
1318 (md_pseudo_table): Add "code16gcc" entry.
1319 (md_assemble): Set i.suffix for insns with DefaultSize modifier.
1323 * config/obj-coff.c (coff_frob_symbol): Always update set_end with
1324 next_set_end even if the end symbol is being discarded.
1326 * gasp.c: Add ATTRIBUTE_UNUSED as needed for non-BFD_ASSEMBLER.
1327 * output-file.c, symbols.c, config/tc-i386.c: Likewise.
1328 * config/obj-coff.c: Likewise.
1329 (seg_info_type): Remove.
1330 (seg_info_off_by_4): Change to array of segT.
1331 (s_get_segment): Adjust accordingly.
1332 (obj_pseudo_table): Fully initialize sentinel entry.
1334 * config/tc-mips.c (append_insn): Correct INSN_SYNC test. From
1339 * config/tc-i386.c (f16_3): New. Fixes 16 bit 3 byte nop.
1343 * config/tc-i386.c: Indentation and white space changes.
1344 (i386_index_check): New function. Add INFER_ADDR_PREFIX code, but
1345 don't enable it by default.
1346 (i386_intel_operand): Remove redundant prototype.
1347 Move check on number of memory operands, and i.mem_operands++
1348 (i386_intel_memory_operand): To here.
1349 Remove i386_immediate code from here. Remove special case code
1350 for input and output using (%dx). Remove base/index checks and
1351 call i386_index_check instead. Save initial operand_string
1352 argument for error message.
1353 (i386_operand): Remove redundant prototype. Move base/index
1354 checks to i386_index_check.
1355 (i386_displacement): Move intel mode check for non-zero
1357 (i386_intel_memory_operand): To here.
1361 * config/tc-sparc.c (md_longopts): Add --no-undeclared-regs option.
1362 (sparc_ip): Warn if %g2 or %g3 register is used and not covered
1363 by .register pseudo-op if -64 and --no-undeclared-regs.
1364 (s_register, sparc_adjust_symtab): New functions.
1365 * config/tc-sparc.h (tc_adjust_symtab, sparc_adjust_symtab):
1366 Declare sparc_adjust_symtab as tc_adjust_symtab.
1367 * doc/c-sparc.texi: Add description of #ignore special literal
1368 for .register pseudo-op.
1372 * config/tc-arm.c (tc_gen_reloc): Record the vtable entry in
1373 the relocation's section offset.
1377 * write.c (fixup_segment): Fix generic error check overflow test.
1379 * config/tc-i386.c (pe): Change %d to %ld, %x to %lx, and cast
1380 X_add_number to long.
1384 * config/tc-hppa.c (pa_ip): Add 'J' and 'K' code
1389 * config/tc-sparc.h (tc_fix_adjustable): Don't adjust GOT, PLT, or
1394 * config/te-go32.h (COFF_LONG_SECTION_NAMES): Define.
1396 * configure.bat: Remove; obsolete.
1397 * config/go32.cfg: Likewise.
1401 * configure.in (i386-*-vxworks*): New target.
1402 * configure: Rebuild.
1406 * doc/c-sparc.texi: Document .register and .nword pseudo-ops.
1410 * config/tc-sparc.c (sparc_ip): Allow OLO10 relocations
1412 (output_insn): Put OLO10's secondary addend into tc_fix_data.
1413 (md_apply_fix3): Handle BFD_RELOC_SPARC_OLO10.
1414 (tc_gen_reloc): Return two relocs for OLO10, LO10 and SPARC13.
1415 * config/tc-sparc.h (RELOC_EXPANSION_POSSIBLE,
1416 MAX_RELOC_EXPANSION): Define.
1417 (TC_FIX_TYPE, TC_INIT_FIX_DATA, TC_FIX_DATA_PRINT): Likewise.
1421 * config/tc-i386.c (intel_float_operand): Add prototype, make static.
1422 (md_assemble): Localize *exp variable to if (fake_zero_displacement)
1423 block. Print a warning if an 8-bit or 16-bit constant
1424 displacement or immediate is truncated on output.
1425 (i386_immediate): Ensure Imm16 is always legal for a 16-bit mode
1427 (i386_operand): Disallow immediate jump absolute operand.
1431 * configure.in: Bump version number to 2.9.5.
1432 * configure: Rebuild.
1434 * dwarf2dbg.c (dwarf2_gen_line_info): Don't assume that long long
1439 * config/tc-hppa.c (md_pseudo_table): Add ".dword" pseudo-op.
1440 (cons_fix_new_hppa): Derive size of fixup from size of the object.
1444 * symbols.c (dollar_label_name): Prepend LOCAL_LABEL_PREFIX if it
1446 * config/tc-arm.h (LOCAL_LABEL_PREFIX): Define to '.' for ELF.
1448 * config/tc-arm.c (md_begin): Set F_SOFTFLOAT in the output file
1449 if -mno-fpu was given.
1450 (tc_gen_reloc): Fix typo. Delete bogus code related to GOTPC
1452 (cons_fix_new_arm): Remove misleading comments.
1456 * write.c (cvt_frag_to_fill): Use frag file and line in rs_org
1458 (relax_segment): Likewise. After giving a rs_org error, convert
1459 the frag to rs_align to avoid cascading errors.
1463 * config/tc-m68k.c: Add some ATTRIBUTE_UNUSED.
1467 * Many files: Changes to avoid gcc warnings: Add ATTRIBUTE_UNUSED
1468 as appropriate. Fill in structure initializations. Add variable
1469 initializations. Add casts.
1470 * dwarf2dbg.c (print_stats): Change i to size_t.
1471 * listing.c (listing_listing): Change list_line to unsigned int.
1475 * config/tc-ppc.h (tc_fix_adjustable) [OBJ_ELF]: Call S_IS_LOCAL
1476 rather than checking for \001 and \002 in symbol name.
1477 * config/tc-sparc.h (tc_fix_adjustable) [OBJ_ELF]: Likewise.
1481 * configure.in (hppa*-linux-gnu*): New target.
1482 * configure: Rebuilt.
1486 * doc/c-arm.texi (ARM Directives): Document .thumb_set directive.
1490 * config/tc-v850.c (v850_comm): Use symbol_get_obj() rather than
1491 accessing symbolP directly.
1495 * config/tc-hppa.h (tc_frob_symbol): Always punt "$global$" symbol
1500 * config/tc-arm.c (ARM_EXT_V5): Define.
1501 (ARM_ARCH_V5, ARM_ARCH_V5T): Define.
1502 (md_begin): Detect ARM v5 architectures.
1503 (md_parse_option): Accept arm v5 specification.
1504 (md_show_usage): Documment -marmv5 switch.
1506 * doc/c-arm.texi: Document -marmv5 command line option.
1508 * config/tc-arm.c (do_adrl): New function. Implement ADRL pseudo
1510 (validate_immediate_twopart): New function. Determine if a
1511 constant can be computed by two ADD instructions.
1512 (output_inst): Remove its command line parameter - it was never
1514 (md_apply_fix3): Support BFD_RELOC_ARM_ADRL_IMMEDIATE, used to
1515 implememt the ADRL pseudo op.
1516 (tc_gen_reloc): Generate a suitable error message if an ADRL
1517 instruction tries to generate a real reloc.
1519 * doc/c-arm.texi: Document NOP, ADR and ADRL pseudo ops.
1523 * config/tc-hppa.c (pa_ip): Convert the opcode and all completers
1528 * subsegs.c (subseg_text_p): Use 1/0 instead of true/false for
1529 non BFD_ASSEMBLER case.
1533 * config/obj-coff.c (obj_coff_section): Mark writable sections as
1538 * dwarf2dbg.c (dwarf2_gen_line_info): Don't call
1539 out_end_sequence() when the address decreases due to a new frag.
1540 (gen_dir_list): Set ls.file[i].dir to j + 1 (not j) because file
1541 numbering starts with 1.
1545 * config/tc-mcore.c (md_pseudo_table): Add .comm for ELF and allow
1547 (mcore_s_text): Call obj_elf_text for ELF target.
1548 (mcore_s_data): Call obj_elf_data for ELF target.
1549 (mcore_s_section): No longer ELF specific. Call obj_coff_section
1551 (mcore_s_bss): New function: Dump literal table before changing
1553 (mcore_s_comm): New function: Dump literal table before changing
1556 * config/obj-elf.c (obj_elf_common, obj_elf_data, obj_elf_text):
1557 No longer static functions.
1558 * config/obj-elf.h (obj_elf_common, obj_elf_data, obj_elf_text):
1559 Provide prototypes for these functions.
1563 * subsegs.c (subseg_text_p): Rewrite non BFD_ASSEMBLER case to use
1564 a list of names, to try obj_segment_name, and to try abbreviated
1565 names when using COFF without long section names.
1567 * config/tc-alpha.c: More use of symbol accessor functions.
1568 * config/tc-arc.c: Likewise.
1569 * config/tc-d30v.c: Likewise.
1570 * config/tc-fr30.c: Likewise.
1571 * config/tc-i860.c: Likewise.
1572 * config/tc-m88k.c: Likewise.
1573 * config/tc-mcore.c: Likewise.
1574 * config/tc-ns32k.c: Likewise.
1575 * config/tc-sparc.c: Likewise.
1576 * config/tc-v850.c: Likewise.
1578 * config/tc-arc.c (get_arc_exp_reloc_type): Change uses of
1579 sy_value with appropriate accessor functions.
1580 * config/tc-arm.c (md_apply_fix3): Likewise.
1581 * config/tc-d10v.c (AT_WORD_P): Likewise.
1582 * config/tc-v850.c (reg_name_search): Likewise.
1584 * config/obj-ecoff.c (obj_ecoff_set_ext): Change uses of bsym to
1585 use symbol_get_bfdsym instead.
1586 * config/tc-ppc.c (md_assemble): Likewise.
1587 * config/tc-v850.c (v850_comm): Likewise.
1591 * config/tc-arc.c (tc_gen_reloc): Use symbol_get_bfdsym to get at
1592 the symbol, rather than accessing the bsym member.
1593 * config/tc-d10v.c (tc_gen_reloc): Likewise.
1594 * config/tc-d30v.c (tc_gen_reloc): Likewise.
1595 * config/tc-mcore.c (tc_gen_reloc): Likewise.
1596 * config/tc-mn10200.c (tc_gen_reloc): Likewise.
1597 * config/tc-mn10300.c (tc_gen_reloc): Likewise.
1598 * config/tc-ns32k.c (tc_gen_reloc): Likewise.
1599 * config/tc-tic30.c (tc_gen_reloc): Likewise.
1600 * config/tc-v850.c (tc_gen_reloc): Likewise.
1604 * tc-hppa.c (elf_hppa_reloc_type): Renamed from elf32_hppa_reloc_type.
1605 (hppa_gen_reloc_type): Conditionalize on BFD64.
1606 (tc_gen_reloc): Re-enable ELF relocations.
1607 * tc-hppa.h (TARGET_FORMAT): Handle elf64-hppa format.
1611 * config/tc-arm.c (ldst_extend): Add parentheses to avoid
1613 (do_ldst): Move assignment out of if condition.
1614 (md_apply_fix3): Add casts to avoid printf format warnings. Add
1615 parentheses to avoid warning.
1619 * config/tc-arm.c (arm_adjust_symtab): Use symbol_get_bfdsym()
1620 macro to get at the BFD symbol associated with a GAS symbol.
1624 * config/tc-ppc.c: Update for symbol handling changes.
1625 * config/obj-coff.c: Likewise.
1629 * tc-hppa.c: General cleanups of ELF support. No more spaces
1630 and subspaces for ELF.
1631 (GDB_DEBUG_SPACE_NAME): Delete definition for ELF.
1632 (GDB_STRINGS_SUBSPACE_NAME): Likewise.
1633 (GDB_SYMBOLS_SUBSPACE_NAME): Likewise
1634 (UNWIND_SECTION_NAME): Likewise.
1635 (space/subspace related structures): Conditionalize definitions
1637 (space/subspace directives and support routines): Conditionalize
1638 definitions and references/uses on OBJ_SOM.
1639 (label_symbol_struct): For ELF, track the symbol's segment. For
1640 SOM track its space.
1641 (pa_define_label, pa_undefine_label, pa_get_label): Corresponding
1643 (USE_ALIASES): Kill for both SOM & ELF.
1644 (pa_def_subspaces, pa_def_spaces): Corresponding changes.
1645 (pa_space, pa_subspace): Corresponding changes.
1646 (pa_spaces_begin): Corresponding chagnes.
1647 (md_begin): Do not muck around with space/subspace stuff for
1649 (md_apply_fix): Temporarily disable argument relocation stuff
1651 (tc_gen_reloc): Temporarily disable relocation generation for
1653 (pa_build_unwind_subspace): Similarly.
1657 * config/tc-arm.c (thumb_set): New pseudo op.
1658 (text, data, section): Override these pseudo ops with ARM
1660 (s_thumb_set): New function: Perform the same as a .set pseudo
1661 op, but also mark the alias'ed symbol as being a Thumb
1663 (arm_s_text): New function: Perform the same as the .text
1664 pseudo op, but dump the literal pool before changing
1666 (arm_s_data): New function: Perform the same as the .data
1667 pseudo op, but dump the literal pool before changing
1669 (arm_s_section): New function: Perform the same as the
1670 .section pseudo op, but dump the literal pool before changing
1672 (arm_cleanup): Do not reset the current section before dumping
1677 * config/tc-m32r.c (md_longopts): Fix OPTION_WARN_UNMATCHED and
1678 OPTION_NO_WARN_UNMATCHED entries.
1679 (md_parse_option): Generate a warning message if an unrecognised
1680 option is encountered.
1682 * config/tc-d10v.c (do_not_ignore_hash): New variable.
1683 (get_operands): When parsing an expression after an '@' symbol
1684 has been detected, do not ignore '#' symbols.
1685 (md_operand): Only ignore '#' symbols if do_not_ignore_hash is
1691 * configure.in (ppc-*-vxworks*): New target.
1692 * configure: Rebuild.
1696 * config/tc-arm.c (tc_gen_reloc): Fix handling of GOTPC relocs.
1700 * write.c (adjust_reloc_syms): Rather than never reducing reloc
1701 which refer to symbols in linkonce sections, permit reducing the
1702 relocs if the symbol is local.
1706 * subsegs.c (subseg_text_p): New function.
1707 * as.h (subseg_text_p): Declare.
1708 * read.c (do_align): Use subseg_text_p to set the default fill.
1709 * write.c (subsegs_finish): Likewise.
1710 * config/obj-coff.c (write_object_file): Likewise.
1711 * config/tc-i386.h (md_maybe_text): Don't define.
1712 (md_do_align): Use subseg_text_p to set the default fill.
1713 * config/tc-m32r.c (m32r_do_align): Likewise.
1714 * config/tc-sh.c (sh_do_align): Likewise.
1715 * config/tc-sparc.h (md_do_align): Likewise.
1719 * configure.in: (i[3456]86-*-freebsd*): Now defaults to ELF.
1720 * configure: Rebuild.
1724 * dwarf2dbg.c: Include elf/dwarf2.h with "", not <>.
1725 * Makefile.am: Rebuild dependencies.
1726 * Makefile.in: Rebuild.
1728 * config/tc-i386.c (i386_immediate): Remove unused label
1731 * struc-symbol.h: Put local_symbol code in ifdef BFD_ASSEMBLER.
1732 * symbols.c: Likewise.
1733 * config/obj-aout.c (obj_crawl_symbol_chain): Refer directly to
1734 sy_next field when taking address, rather than symbol_next.
1736 * dwarf2dbg.c: Change bfd_vma to addressT and bfd_signed_vma to
1738 (out_set_addr): Don't use BYTES_PER_ADDRESS. Instead, get the
1739 value from the output file architecture.
1740 (dwarf2_gen_line_info): Ifdef BFD_ASSEMBLER specific code.
1741 * dwarf2dbg.h: Change bfd_vma to addressT.
1745 * dwarf2dbg.h: Use PARAMS in function declarations.
1749 * write.c (fixup_segment): Don't add symbol value for i960 ELF.
1750 * config/tc-i960.c (s_leafproc): Don't call tc_set_bal_of_cal if
1752 (md_apply_fix): Simplify BFD_ASSEMBLER handling.
1756 * config/tc-i386.c (md_apply_fix3): Add default case to switch.
1758 * config/tc-sparc.c (md_pseudo_table): Remove pushsection and
1761 * config/tc-sparc.c (sparc_ip): Add default case to reloc switch.
1763 * read.c (read_a_source_file): Only declare inescape if
1766 * itbl-ops.c (itbl_disassemble): Change sprintf format strings to
1768 (find_entry_byval): Add parens to avoid warning.
1770 * as.c: If HAVE_ITBL_CPU, include "itbl-ops.h".
1772 * symbols.c (resolve_symbol_value): Don't permit subtraction of
1777 * config/tc-sparc.c (sparc_ip): Don't use side-effect expression
1780 * config/tc-sparc.c (synthetize_setuw, synthetize_setsw,
1781 synthetize_setx): New functions.
1782 (md_assemble): Broken the special cases into the above
1783 functions. Make compiler happy if sizeof(bfd_vma)==4.
1784 Fix sethi generated from set/setuw. If instructions have a relloc,
1785 always clear the fields to be relocated in the opcode.
1786 (sparc_ip): Remove special_case global variable.
1791 * itbl-ops.c (itbl_get_reg_val): Add pval parameter. Return
1792 indication of success rather than a value.
1793 (itbl_get_val): Likewise.
1794 (itbl_get_field): Use strcspn. Change delimiters to include
1796 * itbl-ops.h (itbl_get_reg_val): Update declaration.
1797 (itbl_get_val): Likewise.
1798 * config/tc-mips.c (mips_ip): Update call to itbl_get_reg_val.
1800 * symbols.c (copy_symbol_attributes): Convert local symbols to
1805 * config/tc-arm.c (md_parse_option): Add support for ARM920 and
1810 * config/tc-sparc.c (md_assemble): Fix up setx, support setsw.
1811 Optimize set if sizeof(bfd_vma) == 64.
1812 (sparc_ip): Fix sethi - without %hi() it should generate
1813 R_SPARC_22 reloc, not R_SPARC_HI22.
1814 (tc_gen_reloc): Handle BFD_RELOC_SPARC22.
1818 * config/tc-sparc.c (md_begin): Handle native wordsize aliases.
1819 (s_ncons): New function.
1820 (native_op_table): New table.
1821 (sparc_ip): Be more strict on %hi() etc.; prepare assembler for
1822 R_SPARC_OLO10 handling.
1826 * expr.h (struct expressionS): Revert last change; widen X_op.
1827 * config/tc-alpha.c (md_begin): Check the field is wide enough.
1831 * Makefile.am (TARGET_CPU_CFILES): Add config/tc-fr30.c.
1832 (TARGET_CPU_HFILES): Add config/tc-fr30.h.
1833 (TARG_ENV_HFILES): Add config/te-epoc-pe.h.
1834 * Makefile.in: Regenerated.
1836 * config/obj-elf.c (obj_elf_common): In MRI mode if called as
1837 `common' pass on to s_mri_common.
1838 (elf_pseudo_table): Pass 1 to obj_elf_common for `common'.
1842 * config/obj-elf.c (obj_elf_section): Don't free the return
1843 value of demand_copy_C_string.
1847 * dwarf2dbg.c (dwarf2_gen_line_info): Mirror the section symbol
1848 creation logic from obj_elf_create_section.
1850 * config/obj-elf.c (elf_pseudo_tab): Add pushsection/popsection.
1851 (section_stack): New.
1852 (special_sections): Make const.
1853 (obj_elf_section): Gut and rewrite parsing.
1854 (obj_elf_change_section): New function broken out of obj_elf_section.
1855 (obj_elf_parse_section_letters): Likewise.
1856 (obj_elf_section_word): Likewise.
1857 (obj_elf_section_type): Likewise.
1858 (obj_elf_previous): Treat as a toggle.
1859 (obj_elf_popsection): New.
1860 * config/tc-ppc.c (ppc_section_word): Take str+len not ptr_str.
1861 (ppc_section_type): Likewise.
1862 * config/tc-ppc.h: Likewise.
1864 * expr.h (struct expressionS): Don't make X_op a bitfield.
1865 * config/tc-alpha.c: Update for symbol handling changes.
1866 (md_apply_fix) [case GPREL]: Use now_seg instead of absolute_section.
1867 (load_expression, emit_ir_load, emit_loadstore, emit_jsrjmp): Likewise.
1871 * dwarf2dbg.c (*): Convert to K&R + prototypes.
1872 (dwarf2_gen_line_info): Kill unused variables.
1873 (dwarf2_finish): Likewise.
1874 (dwarf2_where): Likewise.
1875 (dwarf2_directive_file): If we've only got a string,
1876 hand off to s_app_file.
1877 * ecoff.c: Move the include of ecoff.h.
1878 * symbols.h (S_IS_FUNCTION): Prototype.
1880 * read.c (LEX_HASH): Supply a default.
1882 (s_globl): Update `c' after skipping whitespace.
1883 * read.h (LEX_END_NAME, is_name_ender): New.
1884 * expr.c (get_symbol_end): Respect it.
1888 * config/tc-hppa.c (md_begin): Convert local symbol dummy_symbol
1890 (tc_gen_reloc): Still need bfd_abs_symbol in some relocs.
1892 * config/tc-hppa.c: Update for symbol handling changes.
1896 * cgen.c: Update for symbol handling changes.
1897 * config/tc-m32r.c: Likewise.
1899 * config/tc-hppa.h: Update for symbol handling changes.
1900 * config/tc-hppa.c: Likewise.
1902 * config/tc-arm.h: Update for symbol handling changes.
1903 * config/tc-arm.c: Likewise.
1904 (symbol_make_empty): Remove. Just use symbol_create.
1906 * symbols.c (symbol_set_tc): Correct name.
1908 * Makefile.am: Rebuild dependencies.
1909 ($(OBJS)): Don't depend upon struc-symbol.h.
1910 (.dep1, .tcdep, .objdep): Create itbl-parse.h.
1911 * dep-in.sed: Don't remove struc-symbol.h.
1912 * Makefile.in: Rebuild.
1914 * doc/internals.texi (Symbols): Describe changes in symbol
1919 * dwarf2dbg.c (dwarf2_gen_line_info): Use section_symbol
1920 instead of doing the work by hand.
1924 * dwarf2dbg.c (INITIAL_STATE): New macro encapsulating initial
1925 state of line state-machine.
1926 (struct ls): Collect DWARF2 line state-machine state in new member
1927 SM. Add member EMPTY_SEQUENCE to keep track if a code sequence
1928 resulted in any DWARF2 directives.
1929 (reset_state_machine): New function.
1930 (out_end_sequence): Ditto.
1931 (dwarf2_gen_line_info): When switching sections or switching to a
1932 lower text address, call out_end_sequence() first to terminate the
1933 previous code sequence as code sequences MUST have monotonically
1934 increasing addresses.
1935 (dwarf2_finish): Call out_end_sequence() instead of open coding it.
1939 * as.c (parse_args): Add option -gdwarf2 to allow requesting
1940 DWARF2 debug info (line information only, at this point).
1941 * as.h: Update comment about supported debug formats.
1942 * dwarf2dbg.c, dwarf2dbg.h: New files.
1943 * Makefile.am (GAS_CFILES, HFILES, GENERIC_OBJS): Add them.
1945 * expr.c (operand): Don't use [ for parens if we want an index op.
1946 (op_encoding): Switch [ into O_index, if desired.
1947 (op_rank): Renumber with O_index on bottom.
1948 (expr): If O_index, match closing bracket.
1949 * expr.h (O_index): New.
1951 * read.c (read_a_source_file): Conditionally allow matched "
1952 in lines passed to md_assemble.
1954 * config/obj-elf.c (elf_pseudo_table): Add `common'.
1958 Add support for storing local symbols in a small structure to save
1959 memory when assembling large files.
1960 * as.h: Don't include struc-symbol.h.
1961 (symbolS): Add typedef.
1962 * symbols.c: Include struc-symbol.h.
1963 (local_hash): New static variable.
1964 (save_symbol_name): New static function, from symbol_create.
1965 (symbol_create): Call save_symbol_name.
1966 (local_symbol_count): New static variable.
1967 (local_symbol_conversion_count): Likewise.
1968 (LOCAL_SYMBOL_CHECK): Define.
1969 (local_symbol_make): New static function.
1970 (local_symbol_convert): New static function.
1971 (colon): Handle local symbols. Create local symbol for local
1973 (symbol_table_insert): Handle local symbols.
1974 (symbol_find_or_make): Create local symbol for local label name.
1975 (symbol_find_base): Check for local symbol.
1976 (symbol_append, symbol_insert): Check for local symbols.
1977 (symbol_clear_list_pointers, symbol_remove): Likewise.
1978 (verify_symbol_chain): Likewise.
1979 (copy_symbol_attributes): Likewise.
1980 (resolve_symbol_value): Handle local symbols.
1981 (resolve_local_symbol): New static function.
1982 (resolve_local_symbol_values): New function.
1983 (S_GET_VALUE, S_SET_VALUE): Handle local symbols.
1984 (S_IS_FUNCTION, S_IS_EXTERNAL, S_IS_WEAK, S_IS_COMMON): Likewise.
1985 (S_IS_DEFINED, S_IS_DEBUG, S_IS_LOCAL, S_GET_NAME): Likewise.
1986 (S_GET_SEGMENT, S_SET_SEGMENT, S_SET_EXTERNAL): Likewise.
1987 (S_CLEAR_EXTERNAL, S_SET_WEAK, S_SET_NAME): Likewise.
1988 (symbol_previous, symbol_next): New functions.
1989 (symbol_get_value_expression): Likewise.
1990 (symbol_set_value_expression): Likewise.
1991 (symbol_set_frag, symbol_get_frag): Likewise.
1992 (symbol_mark_used, symbol_clear_used, symbol_used_p): Likewise.
1993 (symbol_mark_used_in_reloc): Likewise.
1994 (symbol_clear_used_in_reloc, symbol_used_in_reloc_p): Likewise.
1995 (symbol_mark_mri_common, symbol_clear_mri_common): Likewise.
1996 (symbol_mri_common_p): Likewise.
1997 (symbol_mark_written, symbol_clear_written): Likewise.
1998 (symbol_written_p): Likewise.
1999 (symbol_mark_resolved, symbol_resolved_p): Likewise.
2000 (symbol_section_p, symbol_equated_p): Likewise.
2001 (symbol_constant_p): Likewise.
2002 (symbol_get_bfdsym, symbol_set_bfdsym): Likewise.
2003 (symbol_get_obj, symbol_set_obj): Likewise.
2004 (symbol_get_tc, symbol_set_tc): Likewise.
2005 (symbol_begin): Initialize local_hash.
2006 (print_symbol_value_1): Handle local symbols.
2007 (symbol_print_statistics): Print local symbol statistics.
2008 * symbols.h: Include "struc-symbol.h" if not BFD_ASSEMBLER.
2009 Declare new symbols.c functions. Move many declarations here from
2011 (SYMBOLS_NEED_BACKPOINTERS): Define if needed.
2012 * struc-symbol.h (SYMBOLS_NEED_BACKPOINTERS): Don't set.
2013 (struct symbol): Move bsym to make it clearly the first field.
2014 Remove TARGET_SYMBOL_FIELDS.
2015 (symbolS): Don't typedef.
2016 (struct broken_word): Remove.
2017 (N_TYPE_seg, seg_N_TYPE): Move to symbol.h.
2018 (SEGMENT_TO_SYMBOL_TYPE, N_REGISTER): Likewise.
2019 (symbol_clear_list_pointers): Likewise.
2020 (symbol_insert, symbol_remove): Likewise.
2021 (symbol_previous, symbol_append): Likewise.
2022 (verify_symbol_chain, verify_symbol_chain_2): Likewise.
2023 (struct local_symbol): Define.
2024 (local_symbol_converted_p, local_symbol_mark_converted): Define.
2025 (local_symbol_resolved_p, local_symbol_mark_resolved): Define.
2026 (local_symbol_get_frag, local_symbol_set_frag): Define.
2027 (local_symbol_get_real_symbol): Define.
2028 (local_symbol_set_real_symbol): Define.
2030 * write.c (write_object_file): Call resolve_local_symbol_values.
2031 * config/obj-ecoff.h (OBJ_SYMFIELD_TYPE): Define.
2032 (TARGET_SYMBOL_FIELDS): Don't define.
2033 * config/obj-elf.h (OBJ_SYMFIELD_TYPE): Add local field. If
2034 ECOFF_DEBUGGING, add ECOFF fields.
2035 (ELF_TARGET_SYMBOL_FIELDS, TARGET_SYMBOL_FIELDS): Don't define.
2036 * config/obj-multi.h (struct elf_obj_sy): Add local field. If
2037 ECOFF_DEBUGGING, add ECOFF fields.
2038 (ELF_TARGET_SYMBOL_FIELDS, TARGET_SYMBOL_FIELDS): Don't define.
2039 (ECOFF_DEBUG_TARGET_SYMBOL_FIELDS): Don't define.
2040 * config/tc-mcore.h: Don't include struc-symbol.h.
2041 (TARGET_SYMBOL_FIELDS): Don't define.
2042 (struct mcore_tc_sy): Define.
2043 (TC_SYMFIELD_TYPE): Define.
2044 * Many files: Use symbolS instead of struct symbol. Use new
2045 accessor functions rather than referring to symbolS fields
2048 * read.c (s_mri_common): Don't add in value of line_label.
2050 * config/tc-mips.c (md_apply_fix): Correct parenthesization when
2051 checking for SEC_LINK_ONCE.
2053 * config/tc-sh.h (sh_fix_adjustable): Declare.
2055 * app.c (input_buffer): New static variable.
2056 (app_push): Save saved_input in allocated buffer.
2057 (app_pop): Restored saved_input.
2058 (do_scrub_chars): Change get parameter to take char * and int as
2059 arguments. Change GET macro to pass input_buffer to get
2060 function. Don't save input into allocated buffer.
2061 * as.h (do_scrub_chars): Update declaration.
2062 * input-file.c (input_file_get): Change to take char * and int.
2063 Read data into passed in buffer. Remove static buffer.
2064 * read.c (scrub_from_string): Change to take char * and int. Copy
2065 data into passed in buffer.
2067 * hash.h: Neaten. Declare hash_traverse.
2068 * hash.c: Complete rewrite based on BFD hashing code.
2069 * gasp.c (chunksize): New variable.
2070 * macro.c (macro_expand_body): Call hash_jam with NULL rather than
2075 * config/tc-arm.c (md_apply_fix3): Add pipeline offset into reloc
2076 addend unless the target uses an old ABI.
2080 -Wchar-subscripts cleanup
2081 * listing.c (listing_newline): Use unsigned char variable, so
2082 calls to isascii,iscntrl are correct.
2083 * atof-generic.c (atof_generic): Cast arg to isdigit, et. al. with
2085 * ecoff.c (ecoff_directive_ent,ecoff_stab): Ditto.
2086 * config/obj-elf.c (obj_elf_vtable_inherit): Ditto.
2087 * config/tc-mips.c (mips_ip,mips16_ip): Ditto.
2088 (my_getSmallExpression,get_number,s_mips_ent): Ditto.
2092 * config/tc-m68k.c (m68k_ip): Check for disallowed index register
2094 (arch_coldfire_p): New #define.
2095 (m68k_ip, m68k_init_after_args): Use arch_coldfire_p.
2099 * config/tc-m68k.c (install_operand): Add places `n', `o'.
2101 * config/tc-m68k.c (m68k_ip): Add formats `E', `G', `H'.
2102 (install_operand): Add place `N'.
2103 (init_table): Add registers ACC, MACSR, MASK.
2105 * config/m68k-parse.h (m68k_register): Add ACC, MACSR, MASK.
2107 * config/tc-m68k.c: Change mcf5200 --> mcf.
2108 (archs): Add mcf5206e, mcf5307.
2109 (m68k_ip): Add format `u'.
2110 (install_operand): Add place `m', `M', `h'.
2111 (init_table): Add upper/lower registers.
2113 * config/m68k-parse.h (m68k_register): Add upper/lower registers.
2117 * config/tc-i960.c: Several minor changes to add ELF and
2118 BFD_ASSEMBLER support.
2119 * config/tc-i960.h: Likewise.
2120 * configure.in (i960-*-elf*): New target.
2121 * aclocal.m4, configure: Rebuild.
2125 * config/tc-i386.c (md_apply_fix3): Only do 1999-05-17 fx_pcrel
2126 reloc changes when defined(BFD_ASSEMBLER).
2130 * config/tc-i386.c (tc_gen_reloc): Remove F and MAP macros.
2132 * write.c (write_print_statistics): Output to file, not stderr.
2134 * expr.c (generic_bignum_to_int32,64): Prototype.
2136 * read.c (s_lcomm_internal, sizeof_sleb128, sizeof_uleb128,
2137 output_sleb128, output_uleb128, output_big_sleb128,
2138 output_big_uleb128, output_big_leb128): Prototype.
2139 (output_big_sleb128, output_big_uleb128): Make inline.
2140 (output_big_leb128): Remove inline
2143 * config/tc-i386.c (md_apply_fix3): Convert BFD_RELOC_16 with
2144 fx_pcrel set to BFD_RELOC_16_PCREL. Similarly for BFD_RELOC_8.
2145 Handle BFD_RELOC_16_PCREL and BFD_RELOC_8_PCREL. Return changed
2146 value for correct overflow check in write.c:fixup_segment.
2147 * write.c (fixup_segment): Move bitfield overflow checks to after
2148 the md_apply_fix call.
2149 * config/obj-coff.c (fixup_segment): Likewise.
2150 * doc/internals.texi (CPU backend): Mention md_apply_fix modifying
2151 valueT *val argument.
2155 * config/atof-ieee.c (gen_to_words): Correctly round a
2156 denormalized number. Fix off-by-one in range checking for
2157 exponent in a denormal.
2161 * config/tc-mcore.c (parse_reg): Accept 'sp' as a valid register
2166 * configure.in (i386-*-rtemself*, sh-*-rtemself*): New targets.
2170 * config/tc-i386.h (InvMem): New flag. Add to AnyMem.
2171 (ReverseRegRegmem): Remove.
2172 (ImmExt): New flag. Renumber some of the opcode_modifier bits.
2173 * config/tc-i386.c (md_assemble): Test for PIII SIMD and AMD
2174 3DNow! via ImmExt opcode_modifier. Remove ReverseRegRegmem
2178 * config/tc-i386.h (RegXMM): New for P/III.
2179 * config/tc-i386.c: Add support for P/III.
2183 * config/tc-ppc.c (md_parse_option): Recognize -mppc64bridge.
2184 (md_begin): Allow ppc32 insns in ppc64bridge mode.
2185 (ppc_insert_operand): Accept SIGNOPT in ppc64 mode.
2189 * config/tc-i386.c (i386_immediate): Skip whitespace before
2190 complaining about junk after expression.
2191 (i386_displacement): Likewise.
2195 * symbols.c (symbol_find_base): Use memcpy instead of strcpy.
2196 Don't copy before downcaseing.
2200 * tc-m68k.c: Include elf/m68k.h.
2201 (m68k_elf_final_processing): New routine.
2202 * tc-m68k.h (elf_tc_final_processing m68k_elf_final_processing):
2207 * config/tc-hppa.c (md_apply_fix): Handle 22 bit fmt insn like a
2212 * config/tc-mcore.c (mcore_s_section): Dump literals before
2217 * config/tc-mcore.c (md_apply_fix3): Insert reloc addend into insn
2222 * config/tc-fr30.h (TC_FIX_TYPE): Delete, cgen fields moved to write.h.
2223 (TC_INIT_FIX_DATA): Delete.
2224 * config/tc-m32r.h (TC_FIX_TYPE): Delete, cgen fields moved to write.h.
2225 (TC_INIT_FIX_DATA): Delete.
2226 * write.h (struct fix): New member fx_cgen, ifdef USING_CGEN.
2227 * write.c (fix_new_internal): Initialize fx_cgen member.
2228 * cgen.c (gas_cgen_record_fixup,gas_cgen_record_fixup_exp): Update.
2229 (gas_cgen_md_apply_fix3): Update.
2230 * config/tc-m32r.c (md_cgen_lookup_reloc): Update.
2231 (md_cgen_record_fixup_exp): Update.
2232 (FX_OPINFO_R_TYPE): Update.
2234 * frags.c (frag_var,frag_variant): Initialize fr_cgen here.
2235 * config/tc-fr30.h (TC_FRAG_INIT): Delete.
2236 * config/tc-m32r.h (TC_FRAG_INIT): Delete.
2237 * frags.h (struct frag): Make opindex, opinfo ints.
2239 * config/tc-fr30.c (FX_OPINFO_R_TYPE): Delete, unused.
2243 * aclocal.m4, configure: Updated for new version of libtool.
2247 * config/tc-mcore.c (md_apply_fix3): Renamed function from
2249 (md_apply_fix3): Do not fix up absolute relocations against
2252 * config/tc-mcore.h (MD_APPLY_FIX3): Define.
2256 * config/tc-mcore.c (md_pseudo_table): Add intercepts for section
2257 changes and data-in-text directives.
2258 (mcore_cons): New function: intercept cons() operations.
2259 (mcore_float_cons): New function: intercept float_cons()
2261 (mcore_stringer): New function: intercept stringer() operations.
2265 * obj.h (struct format_ops): Change generate_asm_lineno field to
2267 * config/obj-ecoff.h (OBJ_GENERATE_ASM_LINENO): Don't define.
2269 * config/tc-alpha.c (find_opcode_match): Add default case to
2271 (find_macro_match): Likewise.
2272 (load_expression): Parenthesize && within ||.
2274 * config/tc-alpha.h (TC_RELOC_RTSYM_LOC_FIXUP): Define.
2278 * config/tc-mcore.c (md_pseudo_table): Add overrides for .bss
2279 .text .data .section pseudo ops.
2280 (mcore_s_section): New function. Dump lits before changing secs.
2281 (mcore_s_text): New function. Dump lits before changing secs.
2282 (mcore_s_data): New function. Dump lits before changing secs.
2286 * config/tc-mips.c (mips_32bitmode): New.
2287 (md_begin): Set mips_32bitmode if needed.
2288 (mips_elf_final_processing): Don't set EF_MIPS_ARCH.
2289 Set EF_MIPS_32BITMODE.
2293 * config/obj-coff.c (c_section_symbol): Fix typo in previous
2298 * config/tc-mcore.h (LOCAL_LABELS_FB): Define to 1.
2302 * tc-hppa.c (pa_get_absolute_exression): Try to handle "5 %r3"
2303 expressions correctly.
2308 * config/tc-mips.c (mips_elf_final_processing): Set EF_MIPS_ARCH.
2312 * tc-hppa.c (pa_ip, case '3'): New case for PA2.0 fmpyfadd
2313 and fmpynfadd instructions.
2317 * as.h (environ): Declare it, if needed.
2318 * as.c (dump_statistics): Don't declare environ.
2319 * configure.in (environ): Detect declaration.
2320 * configure, config.in: Rebuild
2322 * config/tc-i386.c (i386_immediate): Accept @GOT relocations.
2323 (i386_displacement): Allocate enough space for replacement buffer.
2324 Clean up replacement buffer initialization.
2328 * subsegs.c (section_symbol): Don't create a new symbol if one
2329 already exists; instead, use the existing one, but set its segment
2330 and frag data if it hasn't already been defined.
2331 * config/obj-coff.c (c_section_symbol): Likewise.
2335 * tc-alpha.c (load_expression): Call as_bad instead of abort.
2339 * config/tc-mcore.c: New File: Support routines for MCore
2341 * config/tc-mcore.h: New File: Definitions for MCore assembler.
2342 * config/obj-coff.c: Add support for mcore-pe target.
2344 * Makefile.am: Add support for MCore targets.
2345 * Makefile.in: Regenerate.
2346 * configure.in: Add support for MCore targets.
2347 * configure: Regenerate.
2349 * doc/all.texi: Set MCORE.
2350 * doc/as.texinfo: Document MCore specific command line options.
2352 * write.h: Prevent multiple inclusion.
2356 * asintl.h (LC_MESSAGES): Never define.
2357 * as.c (main): Don't pass LC_MESSAGES to setlocale if the system
2359 * gasp.c (main): Don't pass LC_MESSAGES to setlocale if the system
2362 * Makefile.am (m68k-parse.c): If configuring in the source
2363 directory, copy m68k-parse.y into the local directory before
2364 running ylwrap, to remove spurious differences when generating
2366 * Makefile.in: Rebuild.
2368 * config/tc-sparc.h (md_do_align): Just allocate the number of
2369 bytes necessary, rather than always allocating 1024.
2373 * listing.c (listing_newline): Add cast to avoid warning.
2374 * read.c (generate_lineno_debug): Add cases to switch. Reindent.
2375 * config/tc-i386.c (i386_scale): Add return value.
2376 (build_displacement_string): Remove unused local temp_disp2.
2377 (i386_intel_memory_operand): Add parentheses to avoid warning.
2378 (i386_intel_operand): Remove unused local end_of_operand_string.
2379 (i386_operand): Remove unused local operand_modifier.
2380 (i386_operand): Add parens to avoid warning.
2384 * configure.in: Add mips*-*-vxworks* target; have it define
2386 * configure, config.in: Rebuild.
2390 * configure.in (emulations): Add support for arm-epoc-pe.
2391 * configure: Regenerate.
2392 * config/te-epoc-pe.h: New file. Define macros specific to
2394 * config/tc-arm.h: Select epoc-pe-arm target format if configured
2395 for arm-epoc-pe target.
2399 * tc-mips.c (md_apply_fix): Adjust value for linkonce sections.
2403 * tc-hppa.c (pa_parse_nonneg_cmpsub_cmpltr): Clean up code to
2404 detect ",n" without a condition.
2405 (pa_parse_neg_cmpsub_cmpltr): Likewise.
2410 * tc-hppa.c (pa_ip, case '~'): The condition for a branch on bit
2411 instruction is encoded with one bit.
2416 * doc/internals.texi (CPU backend): Mention that
2417 line_separator_chars should not include newline. From thi
2422 * config/tc-fr30.c (md_begin): Update call to fr30_cgen_cpu_open.
2423 * config/tc-m32r.c (md_begin): Update call to m32r_cgen_cpu_open.
2427 * tc-alpha.c (md_assemble): Allow '6' in an opcode.
2431 * tc-hppa.c (pa_ip, case 'a'): Do not call pa_parse_..._cmpsub_cmpltr.
2436 * tc-hppa.c (pa_ip, case 'd'): Do not allow ",n".
2440 * app.c (do_scrub_begin): Change '-' back to a symbol char
2441 so we can use multiple opcodes on a line again.
2443 * config/tc-d30v.c: By default, warn if a symbol has
2444 the same name as a register. Plus some minor
2445 updates from the branch.
2449 * config/tc-d30v.c (md_apply_fix3): Handle BFD_RELOC_8,
2450 BFD_RELOC_16 and BFD_RELOC_64.
2454 * expr.c (expr): Add missing else.
2458 * config/tc-arm.c (md_apply_fix3): Improve error message.
2462 * Makefile.am (CPU_TYPES): Add fr30.
2463 (cgen.o): Add $(CGEN_CPU_PREFIX)-desc.h dependency.
2464 (fr30,m32r dependencies): Update.
2465 * Makefile.in: Rebuild.
2467 * cgen.c (gas_cgen_record_fixup): Update use of operand->type.
2468 (gas_cgen_record_fixup_exp): Ditto.
2469 (gas_cgen_finish_insn): Call cgen_operand_lookup_by_num.
2470 (gas_cgen_md_apply_fix3): Ditto. Update call to set_vma_operand.
2471 * config/tc-fr30.c (md_begin): Update call to fr30_cgen_cpu_open.
2472 (md_cgen_lookup_reloc): Update use of operand->type.
2473 * config/tc-m32r.c (md_begin): Update call to fr30_cgen_cpu_open.
2474 (md_convert_frag): Call cgen_operand_lookup_by_num.
2475 (md_cgen_lookup_reloc): Update use of operand->type.
2476 (m32r_cgen_record_fixup_exp): Ditto.
2480 * config/tc-mips.c (md_show_usage): Fix message.
2484 * doc/c-arm.texi (ARM Syntax): Document new command line switches
2485 and LDR reg,=<expr> instruction.
2487 * config/tc-arm.c: Add support for -mcpu=arm810, -mcpu=arm9 and
2492 * doc/c-arm.texi (ARM-Chars): Fix typo in use of '@'.
2496 This patch was created by: Scott Bambrough
2500 Special cased '@' character. The '@' character is used as the
2501 ARM assembler comment character, as a special character
2502 and in ELF .symver pseudo-op's, and as a special character in
2503 .type and .section pseudo-ops.
2504 (symver_pseudo): New static variable.
2505 (symver_state): New static variable.
2506 (struct app_save): Add field 'symver_state'.
2507 (app_push): Save global symver_state int struct app_save.
2508 (app_pop): Restore global symver_state from struct app_save.
2509 (do_scrub_chars): Special case handling of '@' character in
2512 * configure.in: Modified to recognize armv* uname syntax from ARM
2514 * configure: Regenerated.
2516 * config/obj-elf.c (obj_elf_section): Allow '%' as well as '@' as
2517 a prefix to the section's type.
2518 (obj_elf_type): Allow '%' as well as '@' and '#' as prefixes to
2519 the type's typename.
2521 * config/tc-arm.h: Add support for PIC generation:
2522 (pic_code): New boolean.
2523 (obj_relocate_extern): Define.
2524 (TC_RELOC_RTSYM_LOC_FIXUP): Define
2525 (TC_CONS_FIX_NEW): Define.
2526 (tc_fix_adjustable): Define.
2527 (GLOBAL_OFFSET_TABLE_NAME): Define.
2529 * config/tc-arm.c: Add support for PIC generation:
2530 (line_seperator_chars): Allow ';' as a seperator for Linux.
2531 (is_immediate_prefix): New macro.
2532 (arm_parse_reloc): New function.
2533 (s_arm_elf_cons): New function.
2534 (do_branch): Special case for BFD_RELOC_ARM_PLT32.
2535 (md_undefined_symbol): Special case handling for the Global Offset
2537 (md_apply_fix3): Handle PIC relocs.
2538 (tc_gen_reloc): Handle PIC relocs.
2539 (md_parse_option): Add support for '-k' command line switch to
2540 enable PIC generation.
2541 (cons_fix_new_arm): New function.
2542 (s_arm_elf_cons): New function.
2546 * configure.in: Add comments for uses of AC_DEFINE.
2547 * acinclude.m4: Likewise.
2548 * acconfig.h: Remove.
2549 * aclocal.m4: Rebuild.
2550 * configure: Rebuild.
2551 * Makefile.in: Rebuild.
2552 * config.in: Rebuild.
2556 * config/tc-mips.c (mips_ip: case 'o'): Fix assertion failure for
2557 non-constant offset from a base register.
2561 * config/tc-alpha.c (md_show_usage): Put \ before newline in
2566 * config/tc-alpha.c (cpu_types): Enable EV6 PALcode with -m21264.
2567 (emit_insn): Look for pc-relative and no-overflow specifiers on
2568 internal relocation types.
2572 * doc/c-mips.texi (MIPS Opts): Updated list of -mNNNN and
2575 * config/tc-mips.c: Remove all the mips_NNNN variables; just use
2577 (mips_4650, mips_4010, mips_4100): Variables removed.
2578 (hilo_interlocks, gpr_interlocks, append_insn, macro_build, macro,
2579 macro2, mips16_macro, mips_ip): Test mips_cpu, not the mips_NNNN
2581 (md_begin): Don't bother initializing the mips_NNNN variables;
2582 mips_cpu is set, and that's good enough now.
2583 (md_parse_option): Have the -mNNNN options set mips_cpu instead of
2584 the mips_NNNN variable. The -no-mNNNN flags are now no-ops.
2585 (show): New function, to handle wrapping in the CPU lists.
2586 (md_show_usage): Update lists of -mcpu and -mNNNN switches.
2590 * config/tc-i386.c (i386_intel_operand): Ignore `SHORT' rather
2591 than treat as an immediate specifier.
2595 * config/tc-i386.c: Prototype many functions.
2596 (set_intel_syntax): Accept `prefix'/`noprefix' specifiers.
2597 (i386_immediate): Remove unused second argument.
2598 (i386_intel_operand): Fix i386_is_reg typo.
2599 (i386_operand): Use allow_naked_reg.
2600 (output_invalid): Make operand int for K&R.
2604 * Makefile.am (EXTRA_as_new_SOURCES): Uncomment--fixed by automake
2606 * Makefile.in: Rebuild.
2610 * Makefile.am (DISTCLEANFILES): Change cgen-opc.h to cgen-desc.h.
2612 (EXTRA_as_new_SOURCES): Comment out.
2613 (.tcdep): <arch>-opc.h renamed to <arch>-desc.h.
2614 * Makefile.in: Rebuild.
2615 * doc/Makefile.in: Rebuild.
2616 * configure.in: Require autoconf 2.13. Redo using_cgen handling.
2617 Delete call to AM_CYGWIN32. Replace AM_EXEEXT with AC_EXEEXT.
2618 (AC_OUTPUT): <arch>-opc.h renamed to <arch>-desc.h.
2619 * configure: Rebuild.
2620 * aclocal.m4: Rebuild.
2621 * config.in: Rebuild.
2622 * cgen.c: Include cgen-desc.h, not cgen-opc.h.
2623 (*): CGEN_OPCODE_DESC renamed to CGEN_CPU_DESC.
2624 (gas_cgen_cpu_desc): Renamed from gas_cgen_opcode_desc.
2625 CGEN_INSN_ATTR renamed to CGEN_INSN_ATTR_VALUE.
2626 CGEN_OPERAND_ATTR renamed to CGEN_OPERAND_ATTR_VALUE.
2627 (gas_cgen_record_fixup): Remove unnecessary != 0 test.
2628 (gas_cgen_record_fixup_exp): Ditto.
2629 (gas_cgen_finish_insn): Ditto. Refer to operand table via cpu
2630 descriptor, not global variable.
2631 (gas_cgen_md_apply_fix3): Refer to operand_table via cpu
2632 descriptor, not global variable. Refer to insert_operand handler
2633 via cpu descriptor, not global function.
2634 * cgen.h (*): CGEN_OPCODE_DESC renamed to CGEN_CPU_DESC.
2635 * config/tc-fr30.c: Include opcodes/fr30-desc.h.
2636 (*): gas_cgen_opcode_desc renamed to gas_cgen_cpu_desc.
2637 CGEN_INSN_ATTR renamed to CGEN_INSN_ATTR_VALUE.
2638 Update call to CGEN_OPERAND_TYPE,CGEN_INSN_OPERANDS.
2639 * config/tc-m32r.c: Ditto.
2640 (assemble_two_insns): Update calls to cgen_lookup_get_insn_operands.
2641 (md_assemble): Ditto.
2642 (md_convert_frag): Update call to CGEN_OPERAND_ENTRY.
2646 * config/tc-arm.c (md_apply_fix3): Fix handling of label1 - label2
2647 relocations for ELF targets.
2651 * configure.in: Add support for StrongARM target.
2652 * configure: Regenerate.
2656 * config/tc-arm.h: Tidy OBJ_ELF and OBJ_COFF definitions.
2658 * config/tc-arm.c (md_apply_fix3): Fix BFD_RELOC_ARM_PCREL_BRANCH
2663 * config/tc-alpha.c (md_show_usage): Document pca56 and ev6 options.
2667 * config/tc-i386.h (LONG_DOUBLE_MNEM_SUFFIX): Define.
2668 (INTEL_DWORD_MNEM_SUFFIX): Define.
2671 (DWORD_PTR): Define.
2672 (XWORD_PTR): Define.
2674 (OFFSET_FLAT): Define.
2676 (NONE_FOUND): Define.
2679 * config/tc-i386.c (set_intel_syntax): New routine.
2680 (intel_syntax): Declare.
2681 (allow_naked_reg): Declare.
2682 (md_pseudo_table): Support .intel_syntax and .att_syntax.
2683 (intel_float_operand): New routine.
2684 (md_assemble): Handle INTEL_DWORD_MNEM_SUFFIX.
2685 Handle brackets as well as parens. Call i386_intel_operand for
2686 intel syntax. Reverse operands if appropriate. Handle new
2687 suffixes. Handle movzx and movsx.
2688 (i386_is_reg): New routine.
2689 (i386_immediate): New routine.
2690 (i386_scale): New routine.
2691 (i386_displacement): New routine.
2692 (i386_operand_modifier): New routine.
2693 (build_displacement_string): New routine.
2694 (i386_parse_seg): New routine.
2695 (i386_intel_memory_operand): New routine.
2696 (i386_intel_operand): New routine.
2697 (i386_operand): Call i386_displacement, i386_immediate,
2698 i386_scale, etc. instead of handling inline.
2699 (parse_register): Handle registers without prefix.
2703 * configure: Regenerate.
2704 * configure.in (arm-*-oabi): New.
2705 (thumb-*-oabi): New.
2706 * config/tc-arm.c (target_oabi): Declare.
2707 (md_apply_fix3): Support REL relocs.
2708 (md_parse_option): Handle -oabi.
2709 (elf32_arm_target_format): New routine.
2710 (md_longopts): Add OPTION_OABI.
2711 * config/tc-arm.h: Redefine TARGET_FORMAT.
2716 * write.c (write_relocs): Handle out of range error.
2718 * config/tc-fr30.c (fr30_fix_adjustable): New function.
2719 (fr30_force_relocation): Default to 0.
2721 * config/tc-fr30.h (obj_fix_adjustable): Define.
2722 (TC_FORCE_RELOCATION): Define.
2724 * cgen.c (gas_cgen_md_apply_fix3): Do not apply fixes to VTABLE
2729 * config/tc-d30v.c (write_2_short): Do not generate a sequential
2730 merge of two instructions if the left instruciton kills the right.
2734 * Makefile.in: Regenerate.
2735 * configure.in: Redo test for using cgen.
2736 * configure: Regenerate.
2740 * config/obj-coff.h (obj_adjust_symtab): Prevent accidental
2741 redefinition of this macro.
2745 * config/tc-mips.c (mips_frob_file): Disable "Unmatched %hi reloc"
2750 * config/tc-mips.c (append_insn): For mips16, insert a nop between
2751 a read of HI or LO and an immediatly following branch.
2755 * config/tc-mips.c (md_begin): Another correction to the setting of
2760 * config/tc-mips.c (md_begin): Correct type-o in setting of mips_eabi64.
2764 * config/tc-m32r.c (md_assemble): Emit a NOP after a relaxable 16
2765 bit insn when optimizing, so that parallelised instructions will
2766 start on a 32 bit boundary.
2770 * config/tc-mips.c (mips_eabi64): New.
2771 (md_begin): Set mips_eabi64.
2772 (mips_elf_final_processing): Use it.
2776 * config/tc-mips.c (mips_elf_final_processing):
2777 Correct setting of ABI in e_flags.
2781 * config/tc-fr30.c (md_assemble): Warn about invalid instructions in delay slots.
2785 * config/tc-mips.c (md_begin,md_parse_option): Handle vr4111.
2789 * cgen.c (gas_cgen_md_apply_fix3): Mark as an error, rather than a
2790 warning, values that don't fit in the field.
2794 * config/tc-mips.c (mips_abi_string): New.
2795 (md_parse_option,md_longopts): Add mabi.
2796 (mips_elf_final_processing): Set e_flags based on mabi flag.
2800 * config/tc-mips.c (md_parse_option): Handle vr4111.
2804 * config/tc-h8300.c (build_bytes): Change message given if the
2805 instruction requires H8/300H mode and we're not in Hmode, to
2806 suggest that it may be the operand modes that are the problem, not
2807 necessarily the opcode.
2811 * config/tc-fr30.c: Add line separator character.
2815 * configure.in (hppa-*-mpeix*): New target.
2816 * config/obj-som.h (obj_som_compiler): Declare.
2817 * config/obj-som.c (compiler_seen): New static variable.
2818 (obj_som_compiler): New function.
2819 * config/tc-hppa.c: Update tc_data uses for change to bfd/som.h.
2820 (md_pseudo_table): Add "compiler" if OBJ_SOM.
2821 (pa_type_args): Set hppa_priv_level.
2822 (pa_compiler): New static function if OBJ_SOM.
2823 * configure: Rebuild.
2827 * read.c (output_leb128): Don't mark as inline.
2831 * config/tc-ppc.c (ppc_vbyte): Prototype and new function for
2832 AIX .vbyte unaligned data support.
2833 (md_pseudo_table): Add 'vbyte' to list of valid pseudos.
2834 (ppc_elf_validate_fix): Add eh_frame to list of ELF relocatable
2839 * config/tc-d30v.c (md_assemble, do_assemble): Improve erroneous
2844 * config/tc-arm.c (elf32_arm_force_relocation): Check for
2845 BFD_RELOC_ARM_PCREL_BRANCH.
2849 * configure.in: Define TARGET_BYTES_{BIG,LITTLE}_ENDIAN after
2850 checking the target type.
2851 (mips-dec-bsd*): Set endian to little.
2852 * configure: Rebuild.
2854 COFF weak symbol support, based on patches from Mark Elbrecht
2856 * config/obj-coff.h (S_IS_WEAK): Define if not BFD_ASSEMBLER.
2857 * config/obj-coff.c (obj_coff_weak): New static function.
2858 (obj_coff_endef) [both versions]: Handle weak symbols.
2859 (coff_frob_symbol): Likewise.
2860 (yank_symbols): Likewise.
2861 (obj_pseudo_table): Add "weak".
2863 * configure.in (m68k-*-gnu*): New target. From Aymeric Vincent
2865 * aclocal.m4: Rebuild with current tools.
2866 * configure: Rebuild.
2868 * config/tc-alpha.c (emit_ldgp): Give an error message rather than
2869 an assertion failure for a case we can't handle when OBJ_ECOFF.
2871 * expr.c (operator): And with 0xff to avoid problems with signed
2876 * config/tc-fr30.c (md_cgen_lookup_reloc): Generate
2877 BFD_RELOC_FR30_48 instead of BFD_RELOC_FR30_32.
2881 * config/tc-fr30.c (md_cgen_lookup_reloc): Enable relocs for
2886 * config/tc-fr30.c (md_pcrel_from_section): Restore previous calculation
2891 * config/tc-fr30.c (md_pcrel_from_section): Fix calculation of
2896 * config/tc-d10v.c (md_assemble): Make static 'etype' have file
2898 (d10v_cleanup): Only generate previous insn if a multiline insn is
2903 * config/tc-fr30.c (md_cgen_lookup_reloc): Add support for
2908 * config/tc-arm.c (md_parse_option): Add support for -marm7xxx and
2909 -marm6xxx command line switches.
2913 * Makefile.am (DEP): Use $(srcdir)/../mkdep.
2914 (itbl-ops.o): Delete duplicate dependencies.
2915 Rebuild dependencies.
2916 Add fr30 dependencies.
2917 * Makefile.in: Rebuild.
2921 * config/tc-fr30.c (md_cgen_lookup_reloc): Updated to match latest
2923 * listing.c: Ignore line terminator characters found inside
2928 * po/gas.pot: Regenerated.
2932 * config/tc-fr30.c (fr30_is_colon_insn): New name for
2933 fr30_is_label_start(). Also checks for delay slot insns.
2935 * config/tc-fr30.c (fr30_is_label_start): New function: Handle
2936 FR30 instructions which contain a colon in the mnemonic.
2938 * config/tc-fr30.h (TC_START_LABEL): Define this macro.
2942 * config/tc-fr30.c: Removed currently superflous code.
2946 * config/tc-fr30.h: New file.
2947 * config/tc-fr30.c: Tweaking so that it will compile.
2951 * config/tc-d10v.h (obj_fix_adjustable): Define.
2952 (TC_FORCE_RELOCATION): Define.
2953 (d10v_force_relocation): Declare.
2954 * config/tc-d10v.c (tc_gen_reloc): Handle Vtable relocs.
2955 (md_apply_fix3): Handle Vtable relocs.
2956 (d10v_fix_adjustable): New.
2957 (d10v_force_relocation): New.
2961 * config/tc-d30v.c: Change default behaviour to ignore potential
2962 conflicts between register name and symbol names.
2966 * configure.in: Add fr30-*-*.
2967 * config/tc-fr30.c: New file.
2968 * Makefile.in: Regenerated.
2969 * config.in: Regenerated.
2970 * configure: Regenerated.
2971 * doc/Makefile.in: Regenerated.
2972 * po/gas.pot: Regenerated.
2976 * config/tc-m32r.c (assemble_two_insns): Ensure both insns
2981 * app.c (do_scrub_begin): Set characters above 127 to be symbol
2983 (do_scrub_chars): Add some casts to unsigned char to avoid
2984 unwanted sign extension.
2985 * read.c (lex_type): Set characters about 127 to be symbol
2987 * config/tc-i386.c (md_begin): Set identifier_chars and
2988 operand_chars for values above 127.
2992 * configure.in: detect cygwin* instead of cygwin32*
2993 * configure: regenerate
2997 * listing.c: Add support for producing a listing from piped
3002 * config/tc-mips.c (hilo_interlocks): Remove mips_3900.
3003 (append_insn): Account for the tx39's multiply behavior.
3007 * config/tc-m32r.c (assemble_two_insns): Rename assemble_two_insns
3008 from assemble_parallel_insns. Add support for '->' to indicate
3009 explicitly serializing the instructions.
3010 (md_assemble): Ditto.
3014 * config/tc-sh.c (sh_fix_adjustable): Adjust EXTERN and
3019 * cgen.c (gas_cgen_md_apply_fix3): Revert last change.
3023 * config/tc-mips.c : support frame and regmask/fregmask when
3024 MIPS_STABS_ELF is specified.
3028 * config/tc-sh.c (sh_fix_adjustable): Only include if OBJ_ELF.
3029 (md_apply_fix): Don't return 1 for VTABLE relocs.
3030 * config/tc-sh.h (obj_fix_adjustable): Define only if OBJ_ELF.
3034 * doc/c-i386.texi: Replace occurences of "opcode" with
3035 "instruction mnemonic", "instruction", or "mnemonic" when
3036 referring to the name of an instruction. Use "opcode" when
3037 referring to the sequence of machine bytes.
3039 * config/tc-i386.c (opcode_chars): Rename to mnemonic_chars.
3040 (is_opcode_char): Rename to is_mnemonic_char.
3041 (md_assemble and i386_operand): Correct error messages from
3042 "opcode" to "instruction mnemonic"
3043 Rename throughout opcode[] -> mnemonic[], opp -> mnem_p,
3044 MAX_OPCODE_SIZE -> MAX_MNEM_SIZE,
3045 DWORD_OPCODE_SUFFIX -> DWORD_MNEM_SUFFIX,
3046 WORD_OPCODE_SUFFIX -> WORD_MNEM_SUFFIX,
3047 BYTE_OPCODE_SUFFIX -> BYTE_MNEM_SUFFIX,
3048 SHORT_OPCODE_SUFFIX -> SHORT_MNEM_SUFFIX
3049 LONG_OPCODE_SUFFIX -> LONG_MNEM_SUFFIX
3051 * config/tc-i386.h (*_MNEM_SUFFIX): Rename from *_OPCODE_SUFFIX.
3053 * config/tc-i386.c (i386_operand): Check for garbage after
3058 * config/tc-i386.c (md_apply_fix3): Change handling of PCREL reloc
3059 for BFD_ASSEMBLER to only change value when COFF if TE_PE.
3063 * config/tc-sh.h (obj_fix_adjustable): Define.
3064 * config/tc-sh.c (sh_force_relocation): Handle VT relocs.
3065 (md_apply_fix): Likewise.
3066 (tc_gen_reloc): Likewise.
3067 (sh_fix_adjustable): New.
3071 * cgen.c (gas_cgen_finish_insn): Update handling of CGEN_INT_INSN_P.
3072 * cgen.h (gas_cgen_finish_insn): Update prototype.
3073 * config/tc-m32r.c (m32r_insn): CGEN_INT_INSN -> CGEN_INT_INSN_P.
3074 cgen_insn_t -> CGEN_INSN_INT.
3075 (make_parallel): Update handling of CGEN_INT_INSN_P.
3076 (assemble_parallel_insn): Ditto.
3077 (target_make_parallel): New function.
3078 (md_assemble): Use it.
3082 * config/tc-m32r.c (m32r_force_relocation): Fix typo.
3086 * config/tc-sh.c (md_assemble): Make sure the entire opcode is
3087 converted into lower case.
3091 * cgen.c (gas_cgen_md_apply_fix3): Handle VTABLE relocs.
3092 (gas_cgen_tc_gen_reloc): Likewise.
3093 * config/tc-m32r.h (obj_fix_adjustable): Define.
3094 * config/tc-m32r.c (m32r_fix_adjustable): New.
3095 (m32r_force_relocation): Handle VTABLE relocs.
3099 * doc/c-arm.texi (ARM Directives): Document .ltorn directive.
3103 * config/tc-m32r.c (assemble_parallel_insn): Convert second opcode
3104 to lower case before parsing.
3106 * config/tc-d30v.c (parallel_ok): Ignore conflicts when explicitly
3107 parallel insns modift buts in the PSW as a side effect.
3111 * config/tc-d30v.c (find_format): Test for missing flag and
3114 (md_apply_fix3): Fix error messages to avoid
3115 assumption about presence of a symbol.
3117 (parallel_ok): Disallow parallel instructions that both modify the
3120 (find_format): Generate a warning if an odd numbered register is
3121 used as the first register in a mutli-register instruction.
3125 * config/tc-d30v.c (md_apply_fix3): Do not assume that bad
3126 relocations are always associated with a symbol.
3130 * tc-sparc.h (TC_FORCE_RELOCATION): Define.
3131 (elf32_sparc_force_relocation): Declare.
3132 * tc-sparc.c (md_apply_fix3): Handle vtable relocs.
3133 (tc_gen_reloc): Handle vtable relocs.
3134 (elf32_sparc_force_relocation): New.
3138 * symbols.c (S_IS_FUNCTION): New.
3139 * config/tc-v850.h (obj_fix_adjustable): Define.
3140 (TC_FORCE_RELOCATION): Define.
3141 (v850_force_relocation): Declare.
3142 * config/tc-v850.c (tc_gen_reloc): Use offset instead
3143 of fx_addnumber for VTABLE reloc addends.
3144 (md_apply_fix3): Handle VTABLE relocs.
3145 (v850_fix_adjustable): New.
3146 (v850_force_relocation): New.
3150 * tc-hppa.c (fp_operand_format): Add some additional formats.
3151 (pa_ip): Do not automatically promote into pa2.0 mode.
3152 (pa_level): Handle ".level 2.0".
3156 * config/tc-i386.c (md_assemble): Handle AMD_3DNOW_OPCODE.
3157 * config/tc-i386.h (template.extension_opcode): Change to
3158 unsigned int to allow full range of 8-bit opcode suffixes.
3159 (None): Redefine as 0xffff.
3162 * doc/c-i386.texi (i386-SIMD): New section.
3166 * read.c (discard_rest_of_line): New function.
3167 * read.h: Declare it.
3168 * config/tc-alpha.c (s_alpha_mask, s_alpha_frame): Use it.
3172 * config/tc-d10v.c (find_symbol_matching_register): New function.
3173 (find_opcode): Cope with the case where a register name matches
3178 * config/tc-v850.c (md_pcrel_from): Rename to
3179 v850_pcrel_from_section.
3180 (v850_pcrel_from_section): Do not resolves symbols in other
3183 * config/tc-v850.h (MD_PCREL_FROM_SECTION): Define.
3187 * config/tc-d10v.c (find_opcode): Generate an error if a register
3188 is supplied for an operand that should not be a register.
3192 * config/tc-d30v.c (write_2_short): But do allow delayed branch
3193 instructions to have another instruction in the right bin.
3197 * config/tc-d30v.c (write_2_short): Do not allow instructions in
3198 the right container if the left container holds a branch
3203 * config/tc-d30v.c (reg_name_search): Only warn if a name matches
3204 both a register name and symbol name.
3205 (find_format): Allow correct parsing of MVTSYS and MVFSYS insns.
3209 * config/tc-d30v.c (write_2_short): Implement EITHER_BUT_PREFER_MU
3210 execution unit class.
3212 (reg_name_search): If a name matches a register and a symbol,
3213 prefer the register.
3214 (find_format): Disallow flag registers when a general purpose
3215 register is required.
3216 If a number is required, but a register has been given, check to
3217 see if a symbol with the same name as the register exists, and if
3218 so, use that symbol.
3222 * config/obj-elf.h (ECOFF_DEBUGGING): Add missing parens.
3226 * config/tc-d30v.c (find_format): Do not accept flag registers as
3227 general purpose registers.
3228 (find_format): If an immediate value is expected at a given place
3229 in a format, but a register name has been provided instead, check
3230 to see if that register name matches the name of a predefined
3231 symbol and if it does, then use the symbol instead.
3232 (reg_name_search): If a register name matches a symbol name,
3233 prefer the register name to the symbol name.
3237 * config/tc-m32r.c (m32r_do_align): After inserting NOPs, reset
3238 the previous insn to empty.
3242 * config/tc-ppc.c (md_apply_fix3): Do not break string into two
3243 pieces, forcing the use of an ANSI compiler.
3247 * config/tc-m68k.h (TC_FORCE_RELOCATION): New macro. Force vtable
3249 * config/tc-m68k.c (md_apply_fix_2): Do nothing for vtable relocs.
3253 * config/obj-elf.c (obj_elf_vtable_inherit): Handle arm
3255 (obj_elf_vtable_entry): Likewise.
3256 * config/tc-arm.h: Define TC_FORCE_RELOCATION for OBJ_ELF.
3257 * config/tc-arm.c (md_apply_fix3): Handle VTABLE relocations.
3258 (tc_gen_reloc): Likewise.
3259 (arm_fix_adjustable): Likewise.
3260 (elf32_arm_force_relocation): New.
3261 (armelf_frob_symbol): Remove coff-style symbol support.
3265 * config/tc-i386.c (i386_operand): Fix typo in last patch.
3269 * config/tc-arm.c (arm_adjust_symtab): Move #ifdef
3270 OBJ_COFF so that routine is defined for a.out format.
3274 * config/tc-i386.c (i386_operand): Detect non-segment registers
3275 used as segment prefixes.
3279 * ehopt.c (check_eh_frame): Check the size of the FDE, and don't
3280 optimize across FDE boundaries.
3282 * config/obj-coff.c (obj_coff_section): Preserve any link once
3283 flags when setting the section flags.
3287 * config/tc-arm.h (obj_adjust_symtab): Fixed typo.
3288 * config/tc-arm.c (armelf_adjust_symtab): Reformatted.
3292 * config/tc-sparc.c (in_signed_range): Sign extend 32-bit words
3297 * frags.c (frag_grow): Include the size of the frag struct in the
3300 * subsegs.c (subseg_set_rest): Adjust the seginfo frchain start
3301 if the new subseg comes before the old.
3305 * config/tc-sparc.c (sparc_ip): Allow all digits in an instruction
3306 to handle edge8 and edge16.
3310 * config/obj-elf.c (obj_elf_vtable_inherit): Print error message
3311 before we clobber the symbol involved.
3315 * config/tc-arm.c: Remove OBJ_ELF definitions for
3316 S_GET_STORAGE_CLASS and S_SET_STORAGE_CLASS. Only
3317 use arm_adjust_symtab for OBJ_COFF.
3318 (armelf_adjust_symtab): New Routine.
3319 * config/tc-arm.h: Define obj_adjust_symtab to
3320 armelf_adjust_symtab for OBJ_ELF.
3324 * configure.in: Make all i386-elf targets use bfd_gas.
3325 * config/tc-i386.c (tc_i386_force_relocation): New.
3326 (tc_i386_fix_adjustable): Don't fix vtable relocs.
3327 (md_apply_fix3): Likewise.
3328 (tc_gen_reloc): Handle them.
3329 * config/tc-i386.h (TC_FORCE_RELOCATION): Always define, calling
3330 tc_i386_force_relocation.
3334 * config/tc-arm.c (md_show_usage): Improve formatting of --help output.
3338 * config/tc-d30v.c (md_assemble): Copy previous opcode over
3339 current opcode after writing the first insturction of a reverse
3344 * read.h (generate_lineno_debug): Add prototype.
3345 * read.c (generate_lineno_debug): Make non-static.
3349 * config/tc-i386.c (md_assemble): Only warn for address/data size
3354 * config/tc-arm.c (arm_fix_adjustable): Do not adjust relocations
3355 against Thumb function names, as the linker needs this information.
3359 * expr.c (operand): Check also that there is no advance in operand
3360 after atof_generic in order to decide "is it label 0f or floating
3365 * config/tc-m32r.c: Replace double dash prefix to M32R specific
3366 command line options with a single dash.
3367 * doc/c-m32r.texi: Replace double dash prefix with a single dash.
3371 * tc-arm.h: Define obj_fix_adjustable for OBJ_ELF.
3372 * tc-arm.c (arm_fix_adjustable): New routine.
3376 * read.c (s_align, s_comm, s_mri_common, s_fail, s_globl, s_space,
3377 s_float_space, s_struct, cons_worker): Move ignore_rest_of_line or
3378 demand_empty_rest_of_line before mri_comment_end.
3379 (equals): Check garbage after expression before
3380 mri_comment_end in MRI mode.
3384 * config/tc-mips.c (macro): Correct M_SGE_I/M_SGEUI_I case for a
3385 small immediate constant to use the constant itself rather than
3390 * config/tc-hppa.c (pa_enter): Call as_bad rather than abort.
3391 (pa_leave): Likewise.
3395 * config/tc-i386.c (md_assemble): Emit a warning for stand-alone
3397 (i386_operand): Fix an error message.
3401 * doc/c-arm.texi (ARM Directives): Document .req directive.
3403 * config/tc-arm.c (reg_required_here): Display erroneous string if
3404 the register name could not be decoded.
3405 Do not set inst.instruction if the sift is -1.
3409 * config/tc-alpha.c (tc_gen_reloc): Bias WEAK symbols just as
3414 * config/tc-d30v.c (d30v_align): Always perform alignment request,
3415 even if it is belived to be unnecessary.
3419 config/tc-i386.c (i386_operand): Size immediate constants by
3420 suffix (erroneously removed as part of July 7 change).
3424 * config/obj-elf.h: Check for redefinition of obj_frob_symbol.
3425 * config/tc-arm.c: Define S_GET_STORAGE_CLASS and S_SET_STORAGE_CLASS.
3426 (armelf_frob_symbol): New Routine.
3427 * config/tc-arm.h: Define obj_frob_symbol if OBJ_ELF.
3431 * config/tc-alpha.c (alpha_fix_adjustable): Don't adjust weak syms.
3435 * config/tc-arm.c (md_begin): Set BFD private flags depending upon
3436 command line switches passed to assembler.
3440 * cgen.h (GAS_CGEN_MAX_FIXUPS): GAS_ prepended, all uses updated.
3441 (gas_cgen_opcode_desc): Declare.
3442 (gas_cgen_parse_operand): Declare.
3443 (*): Prepend gas_ to gas specific fns to denote them as such.
3445 * cgen.c (gas_cgen_opcode_desc): New global
3446 (gas_cgen_init_parse): Renamed from cgen_asm_init_parse.
3447 (queue_fixup): Renamed from cgen_queue_fixup.
3448 (*): Prepend gas_ to gas specific fns to denote them as such.
3450 (gas_cgen_md_apply_fix3): Update call to insert_operand.
3451 (gas_cgen_finish_insn): Renamed from cgen_asm_finish_insn.
3452 * config/tc-m32r.c (md_begin): Remove use of CGEN_SYM.
3453 Open opcode table and initialize it.
3454 (make_parallel): Use gas_cgen_opcode_desc.
3455 (assemble_parallel_insn): Ditto. Remove use of CGEN_SYM.
3456 (md_assemble): Ditto.
3460 * as.h (debug_info_type): Add entries for unspecified and dwarf*.
3461 * ecoff.c (ecoff_generate_asm_lineno): Take no arguments; call
3462 as_where ourselves. Provide a stub for !ECOFF_DEBUGGING.
3463 * ecoff.h: Move ECOFF_DEBUGGING protection inside GAS_ECOFF_H.
3464 Move ecoff_generate_asm_lineno outside ECOFF_DEBUGGING protection.
3465 * read.c (generate_lineno_debug): Tidy ECOFF bits. Use
3466 DEBUG_UNSPECIFIED rather than DEBUG_NONE for initial test.
3467 * config/obj-elf.h (ECOFF_DEBUGGING) [TC_ALPHA]: Define to a variable.
3468 (SEPARATE_STAB_SECTIONS): Conditionalize on value of ECOFF_DEBUGGING.
3469 (INIT_STAB_SECTION): Likewise.
3470 (OBJ_PROCESS_STAB): Likewise.
3472 * config/tc-alpha.c (md_longopts): New options -mdebug/-no-mdebug.
3473 (md_parse_option): Watch for them.
3474 (alpha_cur_ent_sym, alpha_flag_mdebug): New variables.
3475 (md_begin): Kill neverdef code.
3476 (s_alpha_ent, s_alpha_end, s_alpha_mask, s_alpha_frame): New.
3477 (s_alpha_prologue): Watch alpha_cur_ent_sym.
3478 (s_alpha_coff_wrapper): New.
3479 (md_pseudo_table): Trap all ECOFF pseudos.
3483 Start of changes to remove mdebug section from mips*-elf
3484 Based on MIPS_STAB_ELF definition
3485 * acconfig.h: undef if not configured
3486 * config.in: undef if not configured
3487 * config/mips-elf.h: only set ECOFF debugging if not stabs-in-elf
3488 * config/tc-mips.c (s_ent): set BSF_FUNCTION
3489 * stabs.c (s_stab_generic): flush frag
3493 * configure.in: (arm-*-elf): Handle.
3494 (thumb-*-elf): Handle.
3495 * configure: Regenerate.
3496 * read.c (stringer): Fix typo in comment.
3497 * write.c (fixup_segment): Don't add symbol value to addend if
3499 * config/tc-arm.c (md_section_align): Don't align dwarf debug
3501 (tc_gen_reloc): Always set the reloc addend to fixp->fx_offset
3506 * config/tc-d30v.c ({cur,prev}_left_kills_right_p): New variables.
3507 (write_2_short): Emit warning if new flag is set.
3508 (do_assemble): Set flags if left instruction is one of special
3509 "right-instruction-killer" type.
3513 * config/tc-sparc.c (md_number_to_chars, cons_fix_new_sparc):
3514 Always output words in debug_info section as big endian.
3515 (sparc_target_format): Choose correct bfd target.
3516 (md_apply_fix3): Rename BFD_RELOC_SPARC_32LE to BFD_RELOC_SPARC_REV32.
3520 * config/tc-mn10300.c (md_assemble): Fix "errmsg" initialization
3521 to work with internationalization code. Issue an error when two
3522 operands match that are not allowed to match.
3526 * configure.in (install_tooldir): Allow target to specify whether
3527 it wants to be installed in $(tooldir)/bin.
3528 * configure: Regenerate.
3529 * Makefile.am (install-exec-local): Set install-exec-tooldir
3530 dependency via configure.
3531 * Makefile.in: Regenerate.
3535 * Makefile.am (install-exec-local): Split into two ...
3536 (install-exec-bindir,install-exec-tooldir): New rules.
3537 * Makefile.in: Regenerate.
3541 * Makefile.am (install-exec-local): Don't remove the file before
3542 checking whether $(bindir) == $(tooldir)/bin. From Maciej
3544 * Makefile.in: Rebuild.
3548 * cgen.c: Include libiberty.h.
3549 (cgen_md_apply_fix3): Update call to md_cgen_lookup_reloc.
3550 (cgen_tc_gen_reloc): Use xmalloc, not bfd_alloc.
3551 * cgen.h (cgen_md_apply_fix3,cgen_tc_gen_reloc): Declare.
3552 (md_cgen_lookup_reloc)): Declare.
3553 (md_cgen_record_fixup_exp): Declare.
3554 * config/tc-m32r.h (md_pcrel_from_section): Declare.
3555 (m32r_relax_frag): Declare.
3556 (cgen_md_apply_fix3): Decls moved to cgen.h.
3557 (cgen_record_fixup_exp,cgen_tc_gen_reloc): Ditto.
3558 (m32r_cgen_record_fixup_exp): Delete decl.
3559 * config/tc-m32r.c (m32r_cpu_desc): #if 0 out.
3560 (assemble_nop): Delete.
3561 (expand_debug_syms): Delete unused `exp'.
3562 (md_cgen_lookup_reloc): Renamed from CGEN_SYM (lookup_reloc).
3563 Add default case for -Wall.
3564 (m32r_cgen_record_fixup_exp): Add default case for -Wall.
3565 (md_atof): Delete unused wordP.
3569 * config/tc-mn10300.c (md_assemble): Make sure "errmsg" has a non-NULL
3574 * doc/as.texinfo: Add documentation for .end, .exitm, .fail,
3575 .ifc, .ifeqs, .ifge, .ifgt, .ifle, .iflt, .ifnc, .ifne, .ifnes,
3576 .print, .purgem, and .struct. Remove documentation for
3581 * cgen.c (cgen_md_apply_fix3): set_operand renamed to set_vma_operand.
3582 Update call to insert_operand.
3586 * config/tc-m32r.c (ms_show_usage): Formatting changes.
3590 * config/tc-i386.c (md_assemble): Don't get confused by trailing
3591 whitespace after a prefix operator.
3595 * configure.in (i386-*-beos{pe,elf,}*): Recognize.
3599 * config/tc-sparc.c (log2): New static function.
3600 (s_reserve): Use log2 to convert alignment before calling
3602 (s_common): Use log2 to convert alignment before calling
3603 record_alignment and frag_align.
3604 (sparc_cons_align): Use log2.
3608 * config/tc-sparc.c (s_reserve): Set symbol size if OBJ_ELF.
3609 (s_common): Likewise.
3611 * config/tc-sparc.c (sparc_handle_align): Reindent a bit. Correct
3612 initialization of waddr.
3613 (sparc_elf_final_processing): Add default case to switch.
3617 * doc/c-i386.texi: Fix a typo. Use the term 80-bit real rather
3618 than temporary real.
3622 * write.c (subsegs_finish): Don't align the segments if there were
3625 * config/obj-coff.c (c_symbol_merge): Correct number of bytes when
3626 copying aux information.
3628 * expr.c (make_expr_symbol): Catch attempts to turn an O_big
3629 expression into a symbol.
3633 * config/tc-i386.c (mode_from_disp_size): Change arg and return
3634 type to unsigned int.
3635 (md_assemble): Change type used to store offsets from unsigned
3637 (i386_operand): Switch error check to only call RESTORE_END_STRING
3638 once after parse_register.
3642 * config/tc-v850.c (md_show_usage): Changed format to match that
3645 * as.c (show_usage): Changed format to match that of gcc, ld, etc.
3649 * config/tc-m68k.c (tc_m68k_fix_adjustable): Don't adjust vtable
3651 (md_apply_fix_2): Force the symbol of the vtable reloc to be
3656 * doc/Makefile.am (MAINTAINERCLEANFILES): Define.
3657 * doc/Makefile.in: Rebuild.
3661 * config/tc-mips.c (mips_ip, case 'i' and 'j'): Mask off high bits
3662 for %lo expressions.
3663 (mips_ip, case 'u'): Move range check after code to mask
3664 off bits in %hi/%lo expressions. Mask off high bits for
3669 * doc/Makefile.am (gasver.texi): New target.
3670 (as.info, as.dvi): Depends upon gasver.texi.
3671 * doc/as.texinfo: Include gasver.texi. Mention version number on
3672 title page and in top node.
3673 * doc/Makefile.in: Rebuild.
3677 * listing.c (listing_listing): For EDICT_LIST, skip all lines up to
3678 but not including the line containing the edict.
3679 * listing.h (LISTING_EOF): New.
3680 * input-scrub.c (input_scrub_next_buffer): Call it.
3684 * config/tc-i386.c (i386_operand): Don't set the size of an
3685 immediate address based solely on the suffix and the mode.
3687 * config/tc-i386.c (md_assemble): Add assertion to make sure
3688 overlap2 does not set Imm.
3690 * config/tc-i386.c (space_chars): Remove. The scrubber converts
3691 sequences of whitespace to a single space.
3692 (is_space_chars): Just compare with space.
3693 (md_begin): Don't initialize space_chars.
3694 (md_assemble): Just skip a single whitespace character.
3695 (i386_operand): Rewrite base-index parsing to use new
3696 parse_register, and to skip white space. Skip white space in a
3697 number of other places too. Don't give error message if
3698 parse_register fails.
3699 (parse_register): Change reg_string parameter to be non-const.
3700 Add end_op parameter. Skip white space after the `%', and return
3701 end of register string. Give error message here rather than
3707 * expr.c (op_encoding): Make const.
3708 (expr_set_precedence): New function.
3709 (expr_begin): Don't set operator rankings, just call
3710 expr_set_precedence.
3711 * expr.h (expr_set_precedence): Declare.
3712 * read.c (s_mri): Call expr_set_precedence.
3716 * doc/as.texinfo (Statements): Remove paragraph discussing
3717 continuing lines with a backslash. This hasn't worked for years,
3722 * obj-vms.c: Add C++ support with ctors/dtors sections. Add weak
3724 (Ctors_Symbols, Dtors_Symbols): New symbol chains.
3725 (ps_CTORS, ps_DTORS): New section types.
3726 (vms_fixup_xtors_section): New function
3727 (Ctors_Psect, Dtors_Psect): Define.
3728 (IS_GXX_XTOR): Define
3729 (global_symbol_directory): Change check of gxx_bug_fixed to 0.
3730 Filter static constructors/destructors and add to
3731 Ctors_Symbols/Dtors_Symbols chain.
3732 (vms_write_object_file): Write Ctors_Symbols/Dtors_Symbols to
3733 appropriate section.
3735 * tc-alpha.h (TARGET_FORMAT): Rename "evax-alpha" to "vms-alpha".
3736 * makefile.vms: Merge vax/vms support.
3740 * config/obj-elf.c (obj_elf_vtable_inherit, obj_elf_vtable_entry): New.
3741 (elf_pseudo_table): Add them.
3742 * config/tc-mips.c (mips_force_relocation): Force vtable relocs.
3743 (md_apply_fix): Accept them.
3744 (mips_fix_adjustable): Don't adjust them.
3745 (tc_gen_reloc): Mung BFD_RELOC_VTABLE_ENTRY for Rel.
3746 * config/tc-ppc.c (md_apply_fix3): Accept vtable relocs.
3747 * config/tc-ppc.h (TC_FORCE_RELOCATION_SECTION): Force vtable relocs.
3748 (tc_fix_adjustable): Don't adjust them.
3752 * Makefile.am (CGEN_CPU_PREFIX): New variable.
3754 * Makefile.in: Regenerate.
3755 * configure.in: AC_SUBST cgen_cpu_prefix.
3756 * configure: Regenerate.
3760 * config/tc-sh (COND_JUMP_DELAY, COND12_DELAY_LENGTH): Define.
3761 Changed all users of COND12_DELAY.
3765 * config/tc-mn10300.c (set_arch_mach): New function.
3766 (md_pseudo_table): Add pseudo-ops to set the current machine type.
3767 (md_begin): Default to mn10300 mode.
3768 (md_assemble): Only accept instructions for the core mn10300
3769 chip and the active machine type.
3773 * subsegs.h (segment_info_type): Give the struct a name.
3774 * config/tc-h8300.h (tc_reloc_mangle): Add prototype.
3775 * config/tc-h8500.h (tc_reloc_mangle): Declare.
3776 * config/tc-sh.h (sh_coff_reloc_mangle): Add prototype.
3777 * config/tc-w65.h (tc_reloc_mangle): Declare.
3778 * config/tc-z8k.h (tc_reloc_mangle): Declare.
3782 * config/tc-v850.c (v850_comm): Restore old section
3783 after common processing.
3787 * config/obj-vms.c (Create_VMS_Object_File): Force binary file.
3791 * config/tc-h8300.c (do_a_fix_imm, build_bytes): Replace cast to
3792 char with code that explicitly sign-extends.
3796 * config/tc-v850.c (md_begin): Restore text section as the current
3797 section after creating call table sections.
3798 * config/obj-coff.h (SYM_AUXINFO): New macro to conceal ugly
3801 * config/obj-coff.c (c_symbol_merge): Replace complex expresion
3802 with call to macro SYM_AUXINFO.
3806 * Makefile.am (install-exec-local): Don't let EXEEXT interfere
3807 with the program transform name.
3808 * Makefile.in: Rebuild.
3812 * config/obj-coff.c (c_symbol_merge): Fix copying of auxiliary
3817 * config/tc-i386.c (i386_operand): Be prepared for a space between
3818 the open parenthesis and the start of the register operand,
3819 because of the June 16 change.
3823 * config/tc-sh.c (md_apply_fix): Handle weak symbols correctly if
3828 * config/tc-d30v.c (d30v_align): Always perform alignment request,
3829 even if it is belived to be unnecessary.
3833 * write.c (adjust_reloc_syms): Never adjust relocs against weak
3835 * config/tc-mips.c (md_apply_fix): Adjust accordingly.
3839 * config/tc-mn10300.c (mn10300_insert_operand): Do not hardcode the
3840 shift amount for a repeated operand. The shift amount for the
3841 repeated copy comes from the size of the operand.
3845 * config/tc-h8300.c (get_operand): Fix typos in ldm/stm support.
3849 * config/tc-mips.c (md_show_usage): Fix -mipsN usage.
3853 * app.c (do_scrub_begin): If tc_symbol_chars is defined, treat all
3854 characters in it as LEX_IS_SYMBOL_COMPONENT.
3855 * config/tc-i386.h (tc_symbol_chars): Define.
3856 (extra_symbol_chars): Declare.
3857 * config/tc-i386.c (extra_symbol_chars): Define.
3858 (comment_chars): Don't use '/' as comment start if TE_LINUX.
3859 (line_comment_chars): Set to '/' if TE_LINUX.
3860 * doc/c-i386.texi (i386-prefixes): Update.
3861 * doc/internals.texi (CPU backend): Document tc_symbol_chars.
3865 * po/Make-in (all-yes): If maintainer mode, depend on .pot file.
3866 ($(PACKAGE).pot): Unconditionally depend on POTFILES.
3870 * config/tc-d10v.c (md_apply_fix3): Checking displacement
3871 constraint in instructions REP & REPI.
3875 * config/tc-d30v.c (md_apply_fix3): Catch BFD_RELOC_8,
3876 BFD_RELOC_16, BFD_RELOC_64 and issue appropriate error messages.
3878 (check_range): If the operand is shifted, then shift the number
3879 before checking its range.
3881 * write.c (adjust_reloc_syms): Add more checks for NULL pointers.
3883 * config/tc-v850.c (v850_comm): Set SEC_COMMON bit on special
3888 * config/tc-v850.c (v850_comm): Create special sections as needed.
3892 * config/tc-d10v.c (write_2_short): Addition of swapping
3893 instructions for sequential and reverse sequential order when
3894 given order is not possible.
3898 * Makefile.am: Rebuild dependencies.
3899 (DEP_INCLUDES): Fix reference to intl build directory.
3900 * Makefile.in: Rebuild.
3904 * doc/c-i386.texi: Update 16 bit documentation.
3906 * config/tc-i386.h: Change Data16 to Size16, Data32 to Size32,
3907 IgnoreDataSize to IgnoreSize as they are used for address size as
3909 * config/tc-i386.c: Likewise. Add code to reject addr32/data32 in
3910 32-bit mode, similarly addr16/data16 and variants.
3914 * config/tc-d30v.c (md_assemble): Fix handling of reverse
3915 sequential word multiply instructions.
3917 (do_assemble): Add extra command line argument, to allow mul32
3918 attribute to be preserved across parallel insns.
3919 (md_assemble): Insert NOPs between explicitly parallel insns which
3920 contain an 32 bit multiply and a 16 multiply.
3924 * config/tc-i386.c: REPNE renamed to REPNE_PREFIX_OPCODE, and
3927 * config/tc-i386.c (reloc): Add braces.
3929 * config/tc-i386.c (struct _i386_insn): Rename bi to sib to be
3930 consistent with Intel naming.
3931 * config/tc-i386.h (base_index_byte): Rename to sib_byte. Don't
3932 use bitfields in sib_byte.
3933 (modrm_byte): Don't use bitfields here either.
3935 * config/tc-i386.c (current_templates): Add const.
3936 (parse_register): Add const to return, param, and char *s.
3937 (i386_operand): Add const to reg_entry *r.
3938 * config/tc-i386.h (templates): Add const to start, end.
3940 Inspired by code for 16 bit gas support from Martynas Kunigelis
3942 * config/tc-i386.c (md_assemble): Add full support for 16 bit
3943 modrm, and Jump, JumpByte, JumpDword, JumpInterSegment insns.
3944 (uses_mem_addrmode): Remove.
3945 (md_estimate_size_before_relax): Add support here too.
3946 (md_relax_table): Rewrite interface to md_relax for 16 bit
3948 (BYTE, WORD, DWORD, UNKNOWN_SIZE): Remove.
3949 (opcode_suffix_to_type): Remove.
3950 (CODE16, SMALL, SMALL16, BIG, BIG16): Define.
3951 (SIZE_FROM_RELAX_STATE): Modify to suit above.
3952 (md_convert_frag): Likewise.
3953 (i386_operand): Add support for 16 bit base/index regs,
3954 immediates, and displacements. Remove some unnecessary casts, and
3955 localise end_of_operand_string, displacement_string_start,
3956 displacement_string_end variables. Add GCC_ASM_O_HACK.
3957 * config/tc-i386.h (NO_BASE_REGISTER_16): Define.
3959 * config/tc-i386.c (prefix_hash): Remove.
3960 (md_begin): Rewrite without obstacks. Remove prefix hash table
3961 handling. Rewrite lexical table handling.
3962 (i386_print_statistics): Don't print prefix statistics.
3963 (md_assemble): Rewrite instruction parser so that line is not
3964 converted to lower case. Don't do a hash_find for prefixes,
3965 instead recognise them via opcode modifier.
3966 (expecting_operand, paren_not_balanced): Localise variables.
3967 * config/tc-i386.h (IsPrefix): Define.
3968 (prefix_entry): Remove.
3970 * config/tc-i386.h (PREFIX_SEPERATOR): Don't define.
3971 * config/tc-i386.c (PREFIX_SEPARATOR): Define here instead, using
3972 '\\' in case where comment_chars contains '/'.
3974 * config/tc-i386.c (MATCH): Ensure given operand and template
3975 match for JumpAbsolute. Makes e.g. `ljmp table(%ebx)' invalid;
3976 you must write `ljmp *table(%ebx)'.
3979 * config/tc-i386.c (BFD_RELOC_16, BFD_RELOC_16_PCREL): Define
3980 as 0 ifndef BFD_ASSEMBLER.
3981 (md_assemble): Allow immediate operands without suffix or
3982 other reg operand to default in size to the current code size.
3986 * config/tc-v850.c (md_begin): Restore creation of
3987 .call_table_text and .call_table_data sections.
3991 * config/tc-d30v.c (md_assemble): Set execution type to unknown
3992 after emitting a word of noops.
3996 * config/tc-i386.c (mode_from_disp_size): Disp16 is mode 2.
3997 (i386_operand): Simplify checks for valid base/index combinations.
3998 Disallow `in 4(%dx),%al'.
4000 * config/tc-i386.c (struct _i386_insn): Make regs, base_reg, and
4002 (add_prefix): Change parameter from char to int.
4004 * config/tc-i386.h (Ugh): Define opcode modifier.
4005 * config/tc-i386.c (md_assemble): Print warnings for Ugh insns.
4007 * config/tc-i386.c (md_assemble): Rewrite MATCH and
4008 CONSISTENT_REGISTER_MATCH macros to check register types more
4009 thoroughly. Check for illegal suffix/operand combinations
4010 when matching insns with operands. Handle new `s' suffix, and
4011 associated FloatMF opcode modifier for float insns with memory
4013 * config/tc-i386.h (FloatMF): Define new opcode modifier.
4014 (No_sSuf, No_bSuf, No_wSuf, No_lSuf): Likewise.
4015 (SHORT_OPCODE_SUFFIX, LONG_OPCODE_SUFFIX): Define.
4016 * config/tc-i386.c: Rename WORD_PREFIX_OPCODE to
4017 DATA_PREFIX_OPCODE throughout.
4019 * config/tc-i386.c (REGISTER_WARNINGS): Define.
4020 (md_assemble): Rewrite suffix/register operand checking code to be
4021 more thorough. Remove Abs8,16,32. Change occurrences of Mem to
4022 AnyMem, the better to grep.
4024 (i386_operand): Don't set Mem bits in i.types[this_operand] when
4025 given a memory operand. Don't set Abs bits either.
4026 (type_names): Remove Mem*, Abs*.
4027 * config/tc-i386.h (Mem8, Mem16, Mem32, Abs8, Abs16, Abs32): Don't
4028 define opcode_modifiers as these cases are handled by Disp8,
4029 Disp16, Disp32 and suffix checks.
4030 (COMES_IN_BOTH_DIRECTIONS): Remove.
4031 (FloatR): Define. It's OK to share the bit with ReverseRegRegmem.
4033 * config/tc-i386.c (md_assemble): Don't emit operand size prefix
4034 if IgnoreDataSize modifier given. Remove ShortformW modifier
4035 test. Add test for ShortForm in W base_opcode modification.
4036 Merge Seg2ShortForm and Seg3ShortForm code.
4037 * config/tc-i386.h (ShortFormW): Remove.
4038 (IgnoreDataSize): Define.
4042 * config/tc-d30v.c (md_assemble): Store previous segment state
4043 with previous instruction.
4047 * config/tc-i386.c (SCALE1_WHEN_NO_INDEX): Define.
4048 (ebp, esp): Remove static variables.
4049 (MATCH): Remove test for InOutPortReg.
4050 (i386_operand): Properly handle InOutPortReg here instead.
4051 Disallows `inb (%dx,2)', `inb %es:(%dx)' and `mov (%dx),%ax'
4052 (md_assemble): Simplify and correct modrm and sib generation.
4053 (i386_operand): Add warning for scale without index.
4054 Rewrite checks for valid base/index combinations.
4056 * config/tc-i386.c (END_STRING_AND_SAVE): Protect arguments of
4057 macros and enclose in do while(0).
4058 (RESTORE_END_STRING): Likewise.
4059 (md_assemble): Add one to printed operand number so we start
4060 from 1 not 0. Add some more gettext invocations.
4061 (i386_operand): Fix `%%s' -> `%%%s'. Inc printed operand
4064 * config/tc-i386.h (WAIT_PREFIX, LOCKREP_PREFIX, ADDR_PREFIX,
4065 DATA_PREFIX, SEG_PREFIX): Define.
4066 * config/tc-i386.c (struct _i386_insn): Remove wait_prefix field.
4067 (check_prefix): Remove function.
4068 (add_prefix): New function. Add prefix to i.prefix as well as
4070 (md_assemble): Changes for add_prefix. Remove hack for wait
4071 prefix, instead always output prefixes in fixed order. Test
4072 for jcxz/loop when selecting between word & dword operations,
4073 and add address size prefix rather than operand size prefix.
4074 Remove operand -> address size hack when emitting jcxz/loop.
4075 (i386_operand): Remove O_Absent check as it's done in expr.
4079 * configure.in: Recognize m5200 as a cpu_type of m68k.
4080 * aclocal.m4: Rebuild with current libtool.
4081 * configure: Rebuild.
4085 * config/tc-m68k.c (md_estimate_size_before_relax): Add more calls
4086 to relaxable_symbol to prevent references to external symbol from
4091 * config/tc-m68k.c (relaxable_symbol): If TARGET_OS is "elf", all
4092 symbols are relaxable.
4096 * config/tc-v850.c (md_begin): Don't create special
4097 sections by default.
4101 * config/tc-mips.c (macro): For div and udiv, close the
4102 reorder block as soon as possible.
4107 * macro.c (macro_mri_mode): New function.
4108 * macro.h (macro_mri_mode): Declare.
4109 * read.c (s_mri): Call macro_mri_mode when switching in and out of
4114 * config/tc-alpha.c (s_alpha_comm): Allow alignment parameter in
4117 * config/tc-alpha.c (s_alpha_comm): Defer restoring character
4118 until after xstrdup in OBJ_EVAX case.
4122 * config/tc-vax.c (md_create_short_jump): Fix off by two bug in
4123 offset calculation. Also, use VAX_BRW from vax-inst.h instead
4124 of hardcoded magic number.
4125 (md_create_long_jump): Use VAX_JMP and VAX_ABSOLUTE_MODE macros.
4129 * read.c (do_s_func): New function.
4131 * read.h (do_s_func): Add prototype.
4135 * config/tc-m32r.c (m32r_do_align): Only fill code sections with
4136 nops if fill pattern not specified.
4141 * config/te-go32.h (TE_GO32): Define.
4142 * config/tc-i386.h (LOCAL_LABEL): Don't define if TE_GO32.
4146 Implement .func/.endfunc pseudo-ops.
4147 * read.h (stabs_generate_asm_func,stabs_generate_asm_endfunc): Declare.
4149 * read.c (potable): Add .func,.endfunc.
4150 (s_func): New function.
4151 * stabs.c (stabs_generate_asm_func,stabs_generate_asm_endfunc): New
4153 (in_doc_func_p,current_function_label): New static globals.
4154 (stabs_generate_asm_lineno): Emit function relative stabs if in .func.
4158 * config/tc-a29k.h (WORKING_DOT_WORD): Define.
4159 * config/tc-alpha.h (WORKING_DOT_WORD): Define.
4160 * config/tc-arm.h (WORKING_DOT_WORD): Define.
4161 * config/tc-h8300.h (WORKING_DOT_WORD): Define.
4162 * config/tc-h8500.h (WORKING_DOT_WORD): Define.
4163 * config/tc-hppa.h (WORKING_DOT_WORD): Define.
4164 * config/tc-i860.h (WORKING_DOT_WORD): Define.
4165 * config/tc-i960.h (WORKING_DOT_WORD): Define.
4166 * config/tc-tic30.h (WORKING_DOT_WORD): Define.
4167 * config/tc-w65.h (WORKING_DOT_WORD): Define.
4168 * config/tc-z8k.h (WORKING_DOT_WORD): Define.
4169 * config/tc-a29k.c: Don't define md_short_jump_size,
4170 md_long_jump_size, md_create_short_jump or md_create_long_jump.
4171 * config/tc-alpha.c: Likewise.
4172 * config/tc-alpha.h: Likewise.
4173 * config/tc-arm.c: Likewise.
4174 * config/tc-h8300.c: Likewise.
4175 * config/tc-h8500.c: Likewise.
4176 * config/tc-hppa.c: Likewise.
4177 * config/tc-i860.c: Likewise.
4178 * config/tc-i960.c: Likewise.
4179 * config/tc-ppc.c: Likewise.
4180 * config/tc-sh.c: Likewise.
4181 * config/tc-sparc.h: Likewise.
4182 * config/tc-tic30.c: Likewise.
4183 * config/tc-w65.c: Likewise.
4184 * config/tc-z8k.c: Likewise.
4188 * config/tc-vax.c (_): Delete this macro used for placeholder
4189 values in vax_operand_width_size; it conflicts with the _() macro
4190 used for internationalization.
4194 * symbols.c (symbol_find_base): Fix case insensitive symbol name
4198 * struc-symbol.h (struct broken_word): Add seg and subseg fields.
4199 * read.c (emit_expr): Initialize seg and subseg fields of a new
4201 * write.c (write_object_file): Switch to the appropriate segment
4202 and subsegment when processing a broken word.
4204 * config/tc-m68k.c (mri_assemble): New static function.
4205 (build_mri_control_operand): Call mri_assemble rather than
4207 (s_mri_else, s_mri_break, s_mri_next, s_mri_for): Likewise.
4208 (s_mri_endf, s_mri_endw): Likewise.
4212 * read.c (s_org): Call md_flush_pending_output if it is defined.
4214 * config/tc-sparc.c (md_show_usage): Add \n\ to new string.
4218 * config/tc-sparc.c (OPTION_LITTLE_ENDIAN_DATA): New.
4219 (md_parse_option): Add for same.
4220 (sparc_md_end): Set bfd_mach_sparc_sparclite_le.
4221 (md_apply_fix3, tc_gen_reloc): Allow BFD_RELOC_SPARC_32LE.
4222 (cons_fix_new_sparc): Added to create BFD_RELOC_SPARC_32LE.
4224 * config/tc-sparc.h (cons_fix_new_sparc): Added.
4228 * config/tc-arm.c (find_real_start): Relax definition of local
4233 * config/tc-d30v.c (d30v_align): Apply address adjustment to all
4234 symbols at the given address, not just the last one specified.
4238 * config/tc-sparc.c (sparc_handle_align): Use number_to_chars_bigendian
4239 or number_to_chars_littleendian to write data.
4243 * config/tc-v850.c (md_assemble): Remove artificially created
4244 register name symbols.
4248 * write.c (fixup_segment): Change "segment" to "section" in
4253 * write.c (fixup_segment): Change sym1-sym2 message again.
4257 * write.c (fixup_segment): Improve error message for sym1-sym2
4258 errors when sym1 is in a different segment from sym2.
4262 * config/tc-m32r.c (warn_unmatched_high): New static local.
4263 (OPTION_WARN_PARALLEL): Rename from OPTION_WARN.
4264 (OPTION_NO_WARN_PARALLEL): Rename from OPTION_NO_WARN.
4265 (md_longopts): Recognize --{no-,}warn-unmatched-high.
4266 (md_parse_option): Likewise.
4267 (md_show_usage): Likewise.
4268 (m32r_frob_file): Likewise.
4270 * read.c (generate_file_debug,generate_lineno_debug): New functions.
4271 (read_a_source_file): Call them.
4272 * read.h (stabs_generate_asm_file): Declare.
4273 * stabs.c (stabs_generate_asm_file): New function.
4274 (generate_asm_file): New function.
4275 (stabs_generate_asm_lineno): Move file name handling into
4280 * config/tc-d30v.c (cur_mul32_p, prev_mul32_p): Make static.
4281 (d30v_current_align, d30v_current_align_seg): New variables.
4282 (d30v_last_label): New variable.
4283 (d30v_align, s_d30v_align, s_d30v_text): New functions.
4284 (s_d30v_data, s_d30v_section): Likewise.
4285 (md_pseudo_table): Call them.
4286 (md_begin): Initialize d30v_current_align_seg.
4287 (md_assemble): Call d30v_align when needed by known current alignment.
4288 (d30v_frob_label, d30v_cons_align): New functions.
4289 * config/tc-d30v.h (md_do_align): Remove.
4290 (tc_frob_label): Call d30v_frob_label.
4291 (md_cons_align): New.
4293 * config/tc-d30v.c (find_format): Convert complex expressions to
4294 expression symbols before processing. Clean up code formatting.
4298 * po/Make-in (install-info): New target.
4302 * config/tc-mn10200.c (md_assemble): Handle "bra" just like "jmp"
4304 * config/tc-mn10300.c (md_assemble): Likewise.
4308 * Makefile.am: Update with `make dep-am'.
4309 (HFILES): Add cgen.h.
4310 (cgen.o): Depend on cgen.h.
4311 * Makefile.in: Regenerate.
4313 * cgen.c (cgen_md_apply_fix3): Don't pass newline to as_warn_where.
4317 * gasp.c (grab_label): Permit a label to be a preprocessor
4318 variable by permitting a label to start with a backslash.
4322 * config/tc-mips.c (validate_mips_insn): Removed hack
4323 for previously inaccessible bitfields in some INSN_TRAP
4328 * config/tc-d30v.c (do_assemble): Abort with error message
4329 if opcode operands do not match.
4333 * config/tc-mips.c (macro_build, validate_mips_insn): Implement
4334 'q' operand format for 20-bit "break"/"sdbbp" instructions.
4335 (mips_ip): Truncate overflowed "break" 'c' operand. Implement
4336 similar new 'q' operand.
4340 * cgen.c (cgen_asm_finish_insn): Fix typo.
4345 * cgen.c: Include it.
4346 (MAX_FIXUPS): Renamed to CGEN_MAX_FIXUPS.
4347 (cgen_asm_finish_insn): Result is now void. New arg `result'.
4348 All callers updated.
4349 * config/tc-m32r.c: Include cgen.h.
4350 (m23r_insn): New members num_fixups,fixups.
4354 * config/tc-hppa.c (md_apply_fix): Slightly rework some code
4355 to avoid compiler warning.
4359 * makefile.vms: Run dec c with /nodebug. Pass CC value when
4362 * makefile.vms (OBJS): Add ehopt.obj
4366 * doc/c-vax.texi: Correct and extend vax/vms documentation.
4370 * config/tc-d30v.c (do_assemble): Accept a new parameter requesting
4371 a short format insn.
4372 (md_assemble): Set it for explicitly packed insns.
4376 * config/obj-coff.c (c_symbol_merge): Do not take address of
4377 native fields when performing the memcpy.
4381 * config/tc-mips.c (macro,macro2): Implement
4382 M_DMULO_I, M_MULO_I, M_DMULOU_I, and M_MULOU_I.
4386 * config/tc-m68k.h (TC_RELOC_RTSYM_LOC_FIXUP): Changed to keep
4387 relocations against globally visible symbols.
4388 * config/tc-m68k.c (relaxable_symbol): New macro.
4389 (m68k_ip, md_estimate_size_before_relax): Use it.
4390 (tc_m68k_fix_adjustable): Also handle weak symbols.
4394 * config/tc-i386.h (TC_RELOC_RTSYM_LOC_FIXUP): Keep relocs for all
4395 references to externally visible symbols.
4396 * config/tc-i386.c (md_apply_fix3): When OBJ_ELF, don't add the
4397 values in twice for a PC relative reloc if the symbol is
4400 * config/tc-sparc.h (tc_fix_adjustable) [OBJ_AOUT]: When PIC,
4401 don't adjust a PC relative reloc against an externally visible
4403 * config/tc-sparc.c (md_apply_fix3): When generating a.out PIC,
4404 for a PC relative fixup against an externally visible defined
4405 symbol, arrange to store object file and addend values as though
4406 the symbol were not defined.
4407 (tc_gen_reloc): Likewise.
4411 * read.c (sizeof_leb128): Referenced externally by write.c so
4416 * config/tc-m32r.c ({,expand_}debug_sym): New functions to record
4417 and expand a 'debug' symbol associated with the next instruction
4418 that does not cause a short instruction to be filled with a NOP.
4419 (md_pseudo_table): Add support for .debugsym.
4420 (assemble_parallel_insn): Add calls to expand_debug_sym as
4422 (md_assemble): Ditto.
4426 * as.c (main): Conditionally call setlocale.
4427 * gasp.c (main): Likewise.
4428 * asintl.h: Include <locale.h> if HAVE_LOCALE_H.
4429 (LC_MESSAGES): Now can be defined even when ENABLE_NLS.
4433 * config/tc-d30v.c (md_show_usage): Correct gettext typo.
4437 * config/tc-hppa.c: Change all calls to bzero to use memset.
4438 (pa_ip): Add cast to avoid warning.
4439 (tc_gen_reloc, md_apply_fix): Likewise.
4440 (pa_find_space_by_number): Likewise.
4441 (hppa_force_relocation): Likewise.
4442 (pa_block): Change i to unsigned int.
4443 * config/obj-som.h (obj_som_copyright): Declare.
4447 * ecoff.c (ecoff_build_lineno): Do not use dummy first_lineno
4448 for line numbers for assembly source.
4452 * configure.in: Change version number to 2.9.4
4453 * configure: Rebuild.
4457 * cgen.c (cgen_asm_finish_insn): New arg relax_p. All callers updated.
4461 * ecoff.h: Change symbolS in function declaration to struct
4466 * config/tc-sh.c (parse_reg): Add casts to avoid warnings.
4467 (md_convert_frag): Fix i18n typo.
4471 * ecoff.c (ecoff_get_cur_proc_sym): New function.
4472 * ecoff.h: Protoype it.
4473 * config/tc-alpha.c [ELF] (s_alpha_prologue): New function.
4474 [EVAX] (s_alpha_prologue): Delete.
4475 (md_pseudo_table): Update.
4479 * config/tc-i960.c (md_assemble): Change bp_error_msg from static
4480 array to local pointer.
4481 (get_args, parse_expr): Add casts to avoid warnings.
4485 * read.c (s_set): Cast xmalloc return value to fragS *.
4486 * config/tc-m68k.c (m68k_ip): Function made static to match
4487 previous forward declaration.
4488 (insert_reg, init_regtable, md_convert_frag_1): Likewise.
4492 * config/tc-v850.c: Add internationalisation macros to error
4495 * config/tc-m32r.c (can_make_parallel): Add internationalisation
4496 macros to error strings.
4500 * config/tc-ppc.c (ppc_fix_adjustable): Correct test of whether a
4501 reloc is in the TOC csect.
4502 (md_apply_fix3): Correct gettext typo.
4506 * config/tc-arm.c (find_real_start): Ignore symbols starting with
4507 .L - they are local labels and the branches are not really
4508 function calls but rather far jumps.
4512 * po/Make-in (MKINSTALLDIRS): Don't look in $(top_srcdir).
4516 * config/tc-i386.c (md_assemble): Print operand number rather than
4517 using ordinal_names.
4518 (i386_operand): Likewise.
4522 * Makefile.am (INTLLIBS): Define to work around apparent automake
4524 All Makefiles: Regenerated.
4526 * Many files: Added gettext invocations around user-visible
4528 * acconfig.h (ENABLE_NLS, HAVE_CATGETS, HAVE_GETTEXT, HAVE_STPCPY,
4529 HAVE_LC_MESSAGES): Define.
4530 * dep-in.sed: Added asintl.h.
4531 * po/Make-in: New file.
4532 * gasp.c (main): Call setlocale, bindtextdomain, and textdomain.
4534 * read.c (Z_): Renamed from `_'.
4535 * Makefile.am (SUBDIRS): Added po.
4536 (POTFILES): new macro.
4537 (po/POTFILES.in): New target.
4538 ($(OBJS)): Added asintl.h.
4540 (INCLUDES): Added -DLOCALEDIR, -I$(top_srcdir)/../intl.
4541 (as_new_LDADD): Added $(INTLLIBS).
4542 (as_new_DEPENDENCIES): Added $(INTLDEPS).
4543 (gasp_new_LDADD): Added $(INTLLIBS).
4544 (gasp_new_DEPENDENCIES): New macro.
4545 * configure, aclocal.m4: Rebuilt.
4546 * configure.in: Call CY_GNU_GETTEXT. Generate po/Makefile.in and
4548 (ALL_LINGUAS): Define.
4549 * macro.c: Include "asintl.h".
4550 * as.c (main): Call setlocale, bindtextdomain, and textdomain.
4551 * as.h: Include "asintl.h".
4552 * config/tc-i386.c (ordinal_names): Removed.
4553 (md_assemble): Changed error text to avoid ordinal_names.
4554 (i386_operand): Likewise.
4555 (reloc): Added as_bad to avoid i18n problems.
4556 (tc_gen_reloc): Likewise.
4557 * config/tc-arm.c (bad_args): Now a #define.
4559 * config/obj-vms.c (VMS_stab_parse): Changed type of
4561 (global_symbol_directory): Unified strings to avoid i18n
4563 * config/tc-m68k.c (get_reloc_code): Added some as_bad calls to
4564 avoid i18n problems.
4565 * config/tc-ns32k.c (reloc): Added as_bad to avoid i18n problems.
4566 * config/tc-ppc.c (md_apply_fix3): Added as_bad_where to avoid
4568 * config/tc-sh.c (md_convert_frag): Added as_bad to avoid i18n
4570 * config/tc-v850.c (md_assemble): Changed C++ comment into C
4572 * config/tc-vax.c (md_assemble): Added as_warn to avoid i18n
4574 * as.c (print_version_id): Added an fprintf to avoid i18n
4576 * cond.c (cond_finish_check): Added as_bad call to avoid i18n
4578 * expr.c (expr): Added as_warn call to avoid i18n problems.
4579 * messages.c (as_assert): Changed code to avoid i18n problems.
4580 (as_abort): Likewise.
4581 * read.c (pseudo_set): Added as_bad call to avoid i18n problems.
4582 (s_space): Likewise.
4583 * po/Make-in, po/POTFILES.in, po/gas.pot: New files.
4587 * config/tc-i386.c (check_prefix): New static function, split out
4589 (struct _i386_insn): Add wait_prefix field.
4590 (md_assemble): Remove wait_prefix local variable. Use
4591 check_prefix when adding a prefix.
4593 * config/tc-i386.c (current_templates): New static variable.
4594 (md_assemble): Remove current_templates local variable.
4595 (md_assemble, i386_operand): Improve error and warning messages in
4596 many places. Add RESTORE_END_STRING in many places before error
4597 return. Clarify some comments.
4599 * config/tc-i386.c (struct _i386_insn): Change seg field to a two
4601 (md_assemble): Parse string instruction operands, looking for
4602 segment override prefixes. Check for invalid segment prefixes on
4604 (i386_operand): i.seg[] and max mem_operand changes for string
4606 * config/tc-i386.h (EsSeg): Define.
4608 * config/tc-i386.h (regKludge): Define.
4609 (iclrKludge, imulKludge): Don't define.
4610 * config/tc-i386.c (md_assemble): Merge imulKludge and iclrKludge
4611 code. Move ReverseRegRegmem fudges into Modrm case. Reorder
4612 opcode_modifier checks to look for more common cases first. Add
4613 default_seg for IsString case.
4617 * configure.in: Call AM_PROG_LEX rather than AC_PROG_LEX and
4619 * configure: Rebuild with new automake and libtool.
4620 * aclocal.m4, Makefile.in: Likewise.
4622 * doc/Makefile.am (as.dvi): New target.
4623 * doc/Makefile.in: Rebuild.
4627 * configure.in: Added sparc86x support.
4629 * configure: Rebuild.
4631 * config/tc-sparc.c (lookup_arch): Added arch_type to struct
4633 (md_parse_option): Warn if -EL is not supported for this architecture.
4635 * config/tc-sparc.h (SPARC_BIENDIAN) Always define.
4639 * config/tc-mips.c (mips_ip): Note when we use get match on
4640 the full instruction name.
4644 * symbols.c (resolve_symbol_value) [O_symbol]: Also store the symbol
4645 back into the expression to handle add/sub simplification correctly.
4649 * config/tc-mips.c (hilo_interlocks): Remove 4300.
4653 * config/tc-arm.c (do_msr): Support undocumented 'msr cpsr_flg,
4658 * symbols.c (max_indent_level): New global.
4659 (print_symbol_value_1): Use it.
4660 * expr.h (expr_build_dot): Declare.
4661 * expr.c (expr_build_dot): New function.
4665 * symbols.c (print_binary): New function.
4666 (print_expr_1): Call it.
4670 * config/tc-m68k.c (m68k_ip, case "#B"): Install the offset of the
4671 operand in the opcode.
4675 * config/tc-i386.h: Reorder operand flags and opcode modifier
4676 flags for clarity. Remove unused definitions: Unknown,
4677 ImmUnknown, DispUnknown, NoModrm.
4678 * config/tc-i386.c (type_names): Add missing Debug type.
4679 (md_assemble): Better duplicate prefix checking. Quicker string
4680 instruction check via new opcode_modifier flag.
4684 * doc/as.texinfo (Invoking): Clarify -Wa example.
4688 * config/tc-mips.c (mips_pseudo_table): Add weakext entry.
4689 (s_mips_weakext): Define.
4690 * ecoff.c (ecoff_directive_weakext): Don't define if defined(TC_MIPS).
4691 * config/obj-ecoff.c (obj_pseudo_table): Don't add weakext if
4696 * config/tc-mn10200.c (tc_gen_reloc): The difference of two symbols
4697 is an error if the value can not be computed at assembly time.
4698 * config/tc-mn10300.c (tc_gen-reloc): Likewise.
4702 * gasp.c (main): Set next field of new include_path structure to
4705 * read.c (s_mri_sect): Call as_bad rather than abort for an
4706 unsupported MRI target.
4710 * config/tc-arm.c (arm_validate_fix): New function. Determine if
4711 the destination of a branch instruction should be altered.
4712 (find_real_start): New function: Locate the real, Thumb coded
4713 start of a Thumb function.
4714 (do_t_branch23): Alter the destination of branches to Thumb
4717 * config/tc-arm.h: Define TC_VALIDATE_FIX.
4721 * config/tc-arm.c (decode_shift): Handle addressing mode 2 w/rrx
4726 * doc/as.texinfo: Use @itemx for a secondary item in a table.
4727 * doc/c-hppa.texi: Likewise.
4731 * Makefile.am: Rebuild dependencies.
4732 * Makefile.in: Rebuild.
4734 * Makefile.am (DEP_INCLUDES): New variable.
4735 (.dep1): Change to work when srcdir is not an absolute path.
4736 (.tcdep, .objdep, .dep2, dep.sed): Likewise.
4737 * Makefile.in: Rebuild.
4741 * config/tc-i386.h, config/tc-i386.c: Revert March 24
4742 LinearAddress patch.
4744 * configure.in: Set version to 2.9.1.
4745 * configure: Rebuild.
4747 * Branched binutils 2.9.
4751 * config/tc-i386.h (FWait): Define.
4752 * config/tc-i386.c (md_assemble): Emit fwait prefix before any
4753 other prefixes. Check FWait flag in opcode table to see which
4754 instructions require an fwait prefix.
4758 * stabs.c (get_stab_string_offset): Always create a stab string
4763 Fix some gcc -Wall warnings:
4764 * atof-generic.c (atof_generic): Add casts to avoid warnings.
4765 * ehopt.c (eh_frame_code_alignment): Likewise.
4766 * expr.c (integer_constant, operand): Likewise.
4767 * frags.c (frag_align): Likewise.
4768 * gasp.c (level_0, change_base, doinstr): Likewise.
4769 * hash.c (hash_ask): Likewise.
4770 * listing.c (listing_page, calc_hex, print_lines): Likewise.
4771 (debugging_pseudo): Likewise.
4772 * macro.c (define_macro, check_macro): Likewise.
4773 * read.c (read_a_source_file, s_align, s_float_space): Likewise.
4774 (ignore_rest_of_line, float_cons): Likewise.
4775 * symbols.c (decode_local_label_name): Likewise.
4776 * write.c (record_alignment, cvs_frag_to_fill): Likewise.
4777 (fixup_segment, number_to_chars_bigendian): Likewise.
4778 (number_to_chars_littleendian): Likewise.
4779 * config/atof-ieee.c (gen_to_words): Likewise.
4780 * config/tc-sparc.c (md_begin, md_assemble): Likewise.
4781 (sparc_ip, parse_keyword_arg, s_common): Likewise.
4782 * read.c (output_big_sleb128): Initialize locals to avoid
4784 (output_big_uleb128, equals): Likewise.
4785 * atof-generic.c (atof_generic): Change number_of_digits_* locals
4786 to unsigned int. Change zeros to unsigned int.
4787 * cond.c (s_if): Add return to default case.
4788 * frags.c (frag_now_fix): Change return type to addressT.
4789 * frags.h (frag_now_fix): Update declaration.
4790 * listing.c (file_info_struct): Change linenum to unsigned int.
4791 (struct list_info_struct): Change hll_line to unsigned int.
4792 (print_source): Update format string.
4793 * read.c (emit_expr): Change scan to unsigned int, and don't
4794 bother to initialize it.
4795 * symbols.c (dollar_label_count): Change to unsigned long.
4796 * write.c (adjust_reloc_syms): Remove unused label reduce_fixup.
4797 * config/tc-sparc.c (sparc_memory_model): Only define if OBJ_ELF.
4798 * config/tc-sparc.c (tc_gen_reloc): Add return to default case.
4802 * config/tc-m68k.c (m68k_ip): Check legal addressing modes for
4803 mcf5200 just as we do for m68000.
4804 (m68k_init_after_args): Likewise.
4805 (md_estimate_size_before_relax): Likewise.
4809 * config/tc-v850.c (md_assemble): Store relocation addend in
4810 fixup instead of instruction.
4814 * config/tc-i386.c (md_assemble): Swap template arguments to
4815 CONSISTENT_REGISTER_MATCH macro in reverse direction test.
4816 This macro is currently symmetric, so passing them the wrong
4817 way didn't cause any problem, but may if the macro is changed
4819 After copying template to i.tm, use i.tm. rather than t-> to
4820 access fields, and make t a const*
4821 Move i.tm.operand_types[] swap to immediately after the copy.
4825 * expr.h (expr_build_uconstant): Add prototype.
4826 (expr_build_unary,expr_build_binary): Add prototypes.
4827 * expr.c (expr_build_uconstant): New function.
4828 (expr_build_unary,expr_build_binary): New functions.
4832 * gasp.c (IS*): Cast argument to unsigned char, not unsigned int.
4833 * macro.c (macro_expand_body): Increase buffer size.
4834 * messages.c (as_warn): Likewise.
4835 (as_warn_where, as_bad, as_bad_where): Likewise.
4840 * Makefile.am (DISTSTUFF): New variable.
4841 (diststuff): New target.
4842 * Makefile.in: Rebuild.
4846 * config/tc-m32r.h (md_cleanup, md_elf_section_change_hook): Call
4847 m32r_elf_section_change_hook.
4849 * config/tc-m32r.c (m32r_elf_section_change_hook): New function to
4850 emit a nop if a section ends with a 16 bit instruction.
4854 * config/obj-coff.c (obj_coff_bss): Compile unconditionally. Call
4855 s_lcomm rather than obj_coff_lcomm.
4856 (obj_pseudo_table): Compile .bss pseudo-op unconditionally.
4860 * config/tc-i386.h (LinearAddress): Define.
4861 * config/tc-i386.c (md_assemble): If LinearAddress is set for the
4862 instruction, don't use a default segment.
4866 * configure.in: (sh*-*-rtems*): Switched from ELF to COFF.
4867 * configure: Rebuild.
4871 * aclocal.m4, configure: Rebuild with libtool 1.2.
4875 * config/tc-arm.c (md_apply_fix3): fix code to test the range of
4876 PC relative branches. Patch courtesy of Jonathan Walton.
4881 * configure.in (emulations): Add thumb-pe target.
4883 * configure (emulations): Add thumb-pe target.
4887 * itbl-lex.l (yywrap): Don't define if already defined as a
4892 * depend.c (quote_string_for_make): New function.
4893 (wrap_output): Use it.
4897 * config/obj-elf.c (obj_elf_section): Set bss flag in seg_info
4898 structure if type is SHT_NOBITS. [Bug fix courtesy of rth]
4902 * config/tc-alpha.c (md_shortopts, md_longopts, md_parse_option):
4903 Recognize -GN and -relax.
4904 (md_begin): Initialize gp size from -G switch.
4905 (alpha_force_relocation): Always force if -relax.
4906 (alpha_align): Take a new argument that will specify when to
4907 emit an R_ALPHA_ALIGN relocation (though we don't do that now).
4908 Change all callers. Emit nop alignment padding as nop+unop pair.
4912 * config/obj-elf.c [TC_ALPHA]: Include <elf/alpha.h>.
4913 * config/tc-alpha.h (ELF_TC_SPECIAL_SECTIONS): New.
4917 * config/tc-d30v.c (write_2_short): Delayed jsr instructions don't
4918 require padding to the next long word boundary.
4922 * cgen.c: #include symcat.h.
4923 * config/tc-m32r.c: Likewise.
4927 * config/tc-mips.c (mips_ip, case 'P'): Make 'P' arguments be
4928 absolute expressions instead of '$' prefixed register names.
4932 * read.c (s_set): Record file and line info for symbols when -as.
4933 (pseudo_set): Don't overwrite that dummy fragment.
4937 * config/tc-ppc.c (md_pseudo_table): Add "section".
4938 (ppc_named_section): New static function.
4942 * tc-ppc.c (ppc_biei): Cache the last symbol we inserted
4943 so we don't have to scan the entire list.
4947 * config/tc-d30v.c (parallel_ok): For the explicitly parallel
4948 case, allow the parallel instructions to modify the same flag
4953 * listing.c (list_symbol_table): Categorize symbols by
4954 undefined_section rather than sy_frag->line == NULL.
4958 * Makefile.am (install-exec-local): Install properly when ln
4959 fails or tooldir == prefix.
4963 * cgen.c (cgen_md_apply_fix3): Delete call to validate_operand.
4964 Test result of insert_operand for error.
4968 * Makefile.am (AUTOMAKE_OPTIONS): Add cygnus.
4969 * configure, Makefile.in, aclocal.m4: Rebuild with automake 1.2e.
4970 * doc/Makefile.am (AUTOMAKE_OPTIONS): Define.
4971 * doc/Makefile.in: Rebuild.
4975 * config/tc-mips.c (macro_build): Handle operand type 'C'.
4976 (macro): Fix handling of M_COP[0-3].
4981 * macro.c (ISSEP): Only treat '<' and '>' as separator characters
4982 if macro_alternate or macro_mri.
4983 (getstring): Remove support for byte constants between < and >.
4984 (get_any_string): '<' only starts a string if macro_alternate or
4986 (macro_expand_body): Permit keyword parameters following
4987 positional parameters.
4990 * configure.in (alpha*-*-netbsd*): New target.
4991 * config/te-nbsd.h (LOCAL_LABELS_FB): Define.
4992 * configure: Rebuild.
4994 * as.h (flag_warn_suppress_instructionswap): Move from here...
4995 * config/tc-d10v.c (flag_warn_suppress_instructionswap): ...to
4996 here, and make static.
4998 * ehopt.c (eh_frame_code_alignment): Only use seg_info if
4999 BFD_ASSEMBLER or MANY_SEGMENTS.
5001 * as.c (show_usage): Update bug-gnu-utils address.
5002 * gasp.c (show_usage): Likewise.
5003 * doc/as.texinfo (Bug Reporting): Likewise.
5007 * config/tc-mips.c (load_address): Don't use mips III or mips IV
5008 insns regardless of the size of a pointer if we're in mips I or
5010 (macro, macro2, s_cprestore, s_cpadd): Likewise.
5014 Fix rac to accept only a0:
5015 * tc-d10v.c (parallel_ok, find_opcode):
5016 Split OPERAND_ACC into OPERAND_ACC0 and OPERAND_ACC1.
5017 Introduce OPERAND_GPR.
5021 * read.c (s_fill): Handle non-constant repeat counts by creating
5022 an rs_space fragment.
5026 * config/tc-i386.c (i386_operand): Change error added Jan 2 1998
5027 from as_bad to as_warn.
5031 * as.c: (perform_an_assembly_pass): Use [TEXT|DATA|BSS]_SECTION_NAME
5032 * as.h: Define default values of [TEXT|DATA|BSS]_SECTION_NAME
5033 * config/obj-elf.c (elf_begin): Use [TEXT|DATA|BSS]_SECTION_NAME
5037 * ehopt.c (eh_frame_code_alignment): If not BFD_ASSEMBLER, use
5038 seg_fix_rotP rather than fix_root from seg_info.
5042 * expr.c: Add support for 0x1_2_3_4 bignums.
5046 * configure.in: Change -linux* to -linux-gnu*.
5047 * configure: Rebuild.
5049 * app.c (do_scrub_begin): Treat \r as whitespace.
5053 * Makefile.am: Update dependencies.
5054 * Makefile.in: Rebuild.
5058 * configure, aclocal.m4: Rebuild with new libtool.
5062 * config/tc-mips.c (md_begin): If mips_cpu is set, then use it as
5063 the argument to bfd_set_arch_mach.
5064 (load_address): Use bfd_arch_bits_per_address to determine the
5065 bit size of an address instead of looking at the isa level.
5066 (macro, macro2, s_cprestore, s_cpadd): Likewise.
5070 * config/tc-v850.c (md_parse_option): Add -mv850any command line option.
5074 * ehopt.c: New file.
5075 * as.h (enum _relax_state): Add rs_cfa.
5076 (check_eh_frame, eh_frame_estimate_size_before_relax): Declare.
5077 (eh_frame_relax_frag, eh_frame_convert_frag): Declare.
5078 * read.c (emit_expr): Call check_eh_frame.
5079 * write.c (cvt_frag_to_fill): Handle rs_cfa.
5080 (relax_segment): Likewise.
5081 * Makefile.am: Rebuild dependencies.
5082 (GAS_CFILES): Add ehopt.c.
5083 (GENERIC_OBJS): Add ehopt.o.
5084 * doc/internals.texi (Frags): Document rs_cfa.
5086 * as.c (show_usage): Mention --traditional-format.
5087 (parse_args): Accept --traditional-format.
5088 * as.h (flag_traditional_format): Declare.
5089 * output-file.c (output_file_create): If flag_traditional_format,
5090 set BFD_TRADITIONAL_FORMAT on stdoutput.
5091 * doc/as.texinfo, doc/as.1: Document --traditional-format.
5093 * config/tc-mips.c (append_insn): Make sure that if we have a
5094 fixup for an unmatched %hi reloc, it does not associated with a
5097 * configure, Makefile.in, aclocal.m4: Rebuild with new libtool.
5098 * doc/Makefile.in: Likewise.
5102 * config/tc-m32r.c (check_for_side_effects): New function.
5103 (can_make_parallel): Add checks for instruction side effects
5104 clashing with the other instruction.
5105 (assemble_parallel_insn): Improve warning messages. Return error
5106 message from non-swapped instruction order.
5110 * config/tc-arm.c: Rename arm_after_pass_hook() to arm_cleanup().
5112 * config/tc-arm.h: Replace md_after_pass_hook definition with a
5113 md_cleanup definition. This moves the forced literal output to
5114 the end of the source pass, and avoids macro's inserting literals
5115 into the code immediately after the macro expansion.
5119 * config/tc-sparc.h (tc_fix_adjustable) [OBJ_ELF]: A reloc against
5120 a gas internal symbol is adjustable.
5121 * config/tc-ppc.h (tc_fix_adjustable): Likewise.
5123 * as.h: If gcc version greater than 2.6, use `__format__' and
5124 `__printf__' in function attributes, rather than `format' and
5129 * config/tc-sparc.c: Only include elf/sparc.h if OBJ_ELF.
5133 Add tms320c30 support:
5134 * config/tc-tic30.h: New file.
5135 * config/tc-tic30.c: New file.
5136 * config/obj-coff.h: If TC_TIC30, include coff/tic30.h and define
5137 TARGET_FORMAT as "coff-tic30".
5138 * configure.in (tic30-*-*aout*, tic30-*-*coff*): New targets.
5139 * Makefile.am: Rebuild dependencies.
5140 (CPU_TYPES): Add tic30.
5141 (CPU_OBJ_VALID): tic30-aout is valid.
5142 (TARGET_CPU_CFILES): Add config/tc-tic30.c.
5143 (TARGET_CPU_HFILES): Add config/tc-tic30.h.
5144 * configure, Makefile.in: Rebuild.
5148 * config/tc-v850.c (md_assemble): Improvements to error messages.
5152 * config/tc-ppc.c (md_apply_fix3): Change BFD_RELOC_HI16 and
5153 BFD_RELOC_HI16_S to store the high bits of any value.
5155 * config/tc-ppc.h (tc_fix_adjustable): Undo change of Fri Jun 27.
5156 (TC_RELOC_RTSYM_LOC_FIXUP): Don't let the
5157 assembler calculate relocations to any external symbol at all.
5158 * config/tc-ppc.c (md_apply_fix3) [OBJ_ELF]: Correct bugs
5159 involving generation of pc-relative relocs.
5160 (md_pcrel_from_section) [OBJ_ELF]: The job this code used to do
5161 has been moved to md_apply_fix3.
5163 * config/tc-ppc.c (md_apply_fix3): Fix test for too-far branch.
5164 (ppc_elf_suffix): Warn about 'identifier+constant@got' syntax,
5165 which actually means (the address of identifier's GOT entry) +
5166 constant, which is not particularly useful.
5170 * read.h (include_dirs): Declare.
5171 (include_dir_count,include_dir_maxlen): Declare.
5175 * configure.in: Correct check for shared opcodes library.
5176 * configure: Rebuild.
5178 * listing.c (buffer_line): If we can't open the file, set at_end.
5179 (listing_print): Remove unused local variable fi.
5181 * config/m68k-parse.y (reglistpair): Handle register list in
5184 * config/vms-conf.h: Don't undef VERSION.
5188 * Makefile.am (CONFIG_OBJS): New variable, containing part of old
5190 (GENERIC_OBJS): New variable, with the rest of the old OBJS
5192 (OBJS): Now just $(CONFIG_OBJS) and $(GENERIC_OBJS).
5193 ($(srcdir)/make-gas.com): Rename from make-gas.com.
5194 (stamp-mk.com): Replace $(OBJS) with $(GENERIC_OBJS).
5195 (EXTRA_DIST): Define.
5196 * vmsconf.sh: Handle {targ-cpu, obj-format, atof-targ} modules
5197 explicitly rather than via the list of object files.
5198 (gcc-as.opt): New file created when make-gas.com is run.
5199 * config-gas.com: Create {targ-cpu.h, obj-format.h, targ-env.h,
5200 itbl-cpu.h} to #include appropriate file rather than copying that
5202 * config/vms-conf.h: Synchronize with current config.in.
5203 * Makefile.in: Rebuild.
5207 * config/tc-d30v.c (do_assemble): Added flag_explicitly_parallel.
5208 (parallel_ok): Relaxed parallel subinstruction dependency check.
5212 * as.h (flag_warn_suppress_instructionswap): added new flag.
5213 * tc-d10v.c (md_parse_option,md_longopts): added "--nowarnswap"
5214 command line argument.
5215 * tc-d10v.c (write_2_short): emit "Swapping instructions"
5216 warning only if flag_warn_suppress_instructionswap is false.
5220 * configure.in (i386-*-mingw32*): New target.
5221 * configure: Rebuild.
5225 * symbols.c (resolve_symbol_value): Don't set the segment if it
5226 hasn't changed, and this is OBJ_AOUT without BFD_ASSEMBLER.
5228 * config/obj-aout.h (S_IS_LOCAL): Correct typo--pass argument to
5233 as.h (unlink): Reverse 13-Feb-97 change; use of unlink vs remove
5234 depends upon HAVE_{UNLINK,REMOVE} values rather than host
5239 * config/obj-coff.h (RESOLVE_SYMBOL_REDEFINITION): Define.
5243 * config/tc-v850.c (v850_insert_operand): Display instruction when
5244 an error is encountered.
5248 * configure.in (i386-*-sco3.2v5*): Defaults to ELF now.
5249 (i386-*-sco3.2v5*coff): New target.
5250 (i386-*-sco3.2*): New target.
5251 * configure: Rebuild.
5255 * config/tc-v850.c: Tidy error message production.
5259 * config/tc-arm.c (md_apply_fix3): Add new variable newimm to hold
5260 validate_immediate return value in the right type for comparisons
5265 * listing.c (MAX_BYTES): Use listing variables not constants.
5266 (data_buffer): No longer an array, but a pointer.
5267 (calc_hex): sizeof(data_buffer) -> MAX_BYTES.
5268 (listing_listing): Allocate data_buffer.
5272 * as.c (parse_args): Add --listing-lhs-width, --listing-lhs-width2,
5273 --listing-rhs-width, --listing-cont-lines.
5274 (show_usage): Update.
5275 * listing.c (listing_lhs_width, listing_lhs_width_second): New vars.
5276 (listing_lhs_cont_lines, listing_rhs_width): New vars.
5277 (print_lines): Use the variables instead of the constants.
5278 (listing_listing): Likewise.
5279 * listing.h: Declare the new vars.
5283 * as.c (parse_args): Add --keep-locals alias for -L.
5284 Add --strip-local-absolute.
5285 (show_usage): Update.
5286 * as.h (flag_strip_local_absolute): New flag.
5287 * symbols.c (S_IS_LOCAL): Use it.
5288 * config/obj-aout.h (S_IS_LOCAL): Likewise.
5289 * config/obj-bout.h (S_IS_LOCAL): Likewise.
5290 * config/obj-coff.h (S_IS_LOCAL): Likewise.
5294 * config/tc-m32r.c: Detect if explicitly parallel instructions
5295 might have an io conflict and issue a warning message.
5299 * cgen.c (cgen_save_fixups, cgen_restore_fixups,
5300 cgen_swap_fixups): Functions to save, restore and swap the fixup
5301 chain with a backup copy.
5302 (cgen_asm_finish_insn): Returns address of constructed insn.
5306 * listing.c (file_info_struct): Remove FILE, add POS.
5307 (last_open_file_info, last_open_file): New; a one entry FILE* cache.
5308 (file_info): Don't open the file.
5309 (buffer_line): Check for the file in the last_open cache, updating
5311 (print_source): Don't reference file_info->file.
5312 (listing_listing): Likewise.
5313 (listing_print): Close the file in the cache, if any.
5317 * read.c (dwarf_file_string): New file static variable.
5318 (emit_expr): Look for constant sequence that leads up to a file
5319 name in DWARF debugging output.
5320 (stringer): Use dwarf_file_string to decide whether to accept a
5321 string as a file name.
5325 * tc-m68k.c (m68k_ip): Remove absl->reglst MRI hack.
5326 (crack_operand): Add reg->reglst MRI hack.
5327 (r_seg): Put reglst symbols in reg_section.
5328 (m68k_frob_symbol): Frob reglst symbols into absolute_section.
5332 * tc-sh.c (get_specific): Handle SGR & DBR.
5336 * tc-h8300.c (parse_reg): Take the length of the symbol into
5337 account when attempting to match a register name.
5338 * tc-h8500.c (parse_reg): Likewise.
5342 * cgen.c: Formatting changes to improve readability.
5346 * config/tc-mips.c (macro): Rework division code to avoid unfilled
5351 Based on a patch from Jim Wilson
5352 * config/tc-d30v.c (do_assemble): Remove non-ansi default case.
5353 (tc_gen_reloc): Handle cross section PC relative relocs
5358 * config/tc-mips.c (mips_ip): Don't test pinfo flags if INSN_MACRO.
5362 * cgen.c: #include setjmp.h. Clean up pass over `struct foo' usage.
5363 (expr_jmp_buf): New static local.
5364 (cgen_parse_operand): Allow use of longjmp in parsing to handle errors.
5365 (cgen_md_operand): New function.
5366 * tc-m32r.c: Clean up pass over `struct foo' usage.
5367 (md_estimate_size_before_relax): Use CGEN_INSN_MNEMONIC.
5371 * symbols.c (S_SET_SEGMENT): Don't set the segment for section syms.
5372 (S_IS_EXTERNAL, S_IS_LOCAL): Correct parenthetication.
5376 * config/tc-i386.c (i386_operand): Give an error if there are
5377 unrecognized characters after an expression.
5379 For older changes see ChangeLog-9697