3 * configure.in: Enable bfd_assember for the MAXQ port.
4 * configure: Regenerate.
9 * config/tc-xtensa.c (finish_vinsn): Clear pending instruction if
11 (check_t1_t2_reads_and_writes): Check for both reads and writes to
12 interfaces that are related as determined by xtensa_interface_class_id.
16 * config/tc-mn10300.c (md_relax_table): Fix off by one negative
17 offsets for conditional branches.
21 * config/tc-xtensa.c (MAX_IMMED6): Change value to 65.
25 * config/tc-xtensa.c (update_next_frag_state): Always add a NOP if
26 relaxing at the end of a loop. Don't mark frags as UNREACHABLE or
28 (relax_frag_immed): Update call to update_next_frag_state.
32 * obj.h (struct format_ops <app_file>): Add int param.
33 * read.h (s_app_file_string): Likewise.
34 * read.c (s_app_file_string): Likewise.
35 (s_app_file): Adjust s_app_file_string call.
36 * config/tc-mips.c (s_mips_file): Likewise.
37 * config/obj-coff.h (obj_app_file): Add app param.
38 * config/obj-ecoff.h (obj_app_file): Likewise.
39 * config/obj-multi.h (obj_app_file): Likewise.
40 * config/obj-elf.h (elf_file_symbol): Likewise.
41 * config/obj-elf.c (elf_file_symbol): Only emit one file symbol
42 if called for # preprocessor lines.
47 * symbols.c (resolve_symbol_value): Convert weak symbols only
49 (symbol_equated_reloc_p): Don't equate weaks when relocating
50 only for Windows PECOFF.
54 * config/tc-crx.c (print_insn): Check and set insn_addr.
55 * config/tc-crx.h (md_frag_check): Define.
60 * configure.in: Add support for new target maxq-coff.
61 * configure: Regenerate.
62 * NEWS: Mention new support.
63 * config/tc-maxq.c: New file.
64 * config/tc-maxq.h: New file.
65 * config/obj-coff.h: Add support for maxq-coff.
69 * symbols.c (any_external_name): Define.
70 (resolve_symbol_value): Convert weak symbols.
71 (S_SET_EXTERNAL): Support any_external_name.
72 (S_SET_NAME): Qualify parameter const.
73 (symbol_equated_reloc_p): Equate to weaks when relocating.
74 * symbols.h (S_SET_NAME): Qualfiy parameter const.
75 * tc.h (any_external_name): Declare.
76 * config/obj-coff.c ("coff/pe.h"): Include for BFD
78 (weak_is_altname): Declare and define.
79 (weak_name2altname): Same.
80 (weak_altname2name): Same.
81 (weak_uniquify): Same.
82 (weak_altprefix): Define.
83 (obj_coff_weak): Change .weak syntax and handling.
84 (coff_frob_symbol): Fix PE weak symbol alternates.
85 * config/obj-coff.h (USE_UNIQUE): Define.
86 * config/tc-i386.c (md_apply_fix3): Assume weak symbols
87 are in another segment.
88 (tc_gen_reloc): Remove broken addend hack.
89 doc/as.texinfo: Update.
93 * config/tc-xtensa.c (total_frag_text_expansion): New.
94 (md_estimate_size_before_relax): Use it.
95 (find_address_of_next_align_frag): Likewise.
99 * config/tc-crx.c: Rename argument types.
100 (processing_arg_number): Rename to 'cur_arg_num'.
101 (get_number_of_bits): Rename to 'set_operand_size'.
102 (get_operandtype): Rename to 'parse_operand', totally rewrite.
103 (set_cons_rparams): Rename to 'set_operand', totally rewrite.
104 (set_indexmode_parameters): Remove function, integrate its code into
106 (set_operand_size): Get rid of 'Operand Number' function parameter -
107 use global variable 'cur_arg_num' instead.
108 Use a local 'argument' pointer to reference the current argument.
109 (parse_operand): Likewise.
110 (set_operand): Likewise.
111 (process_label_constant): Likewise.
115 * config/tc-xtensa.c: Remove XTENSA_SECTION_RENAME ifdefs.
116 (add_section_rename): Delete. Inlined into...
117 (build_section_rename): ...here. Use xstrdup instead of strdup.
118 (xtensa_section_rename): Drop "const" from argument and return types.
119 (md_show_usage): Indent to match show_usage().
120 * config/tc-xtensa.h: Remove XTENSA_SECTION_RENAME ifdefs.
121 (tc_canonicalize_section_name): Define.
122 (md_elf_section_rename): Remove unused macro.
123 * doc/as.texinfo (Overview): Document Xtensa --rename-section option.
124 * doc/c-xtensa.texi (Xtensa Options): Likewise.
125 (Frame Directive): Delete.
129 * configure.in: Remove arm-*-oabi and thumb-*-oabi.
130 * config/tc-arm.c (target_oabi): Delete.
131 (md_apply_fix3, elf32_arm_target_format): Remove target_oabi checks.
132 (arm_opts): Remove moabi.
133 * doc/as.texinfo (Overview): Remove documentation of -moabi.
134 * doc/c-arm.texi (ARM Options): Likewise.
135 * configure: Regenerated.
139 * configure.in (crisv32): Recognize. AC_DEFINE_UNQUOTED
140 DEFAULT_CRIS_ARCH. Handle crisv32-*-linux-gnu* like
141 cris-*-linux-gnu* and crisv32-*-* like cris-*-*.
142 * configure: Regenerate.
143 * config/tc-cris.c (enum cris_archs): New.
144 (cris_mach, cris_arch_from_string, s_cris_arch, get_sup_reg)
145 (cris_insn_ver_valid_for_arch): New functions.
146 (DEFAULT_CRIS_ARCH): New macro, default to cris_any_v0_v10.
147 (cris_arch): New variable.
148 (md_pseudo_table): New pseudo .arch.
149 (err_for_dangerous_mul_placement): Initialize according to
151 (STATE_COND_BRANCH): Renamed from STATE_CONDITIONAL_BRANCH.
153 (STATE_COND_BRANCH_V32, STATE_COND_BRANCH_COMMON)
154 (STATE_ABS_BRANCH_V32, STATE_LAPC, BRANCH_BF_V32, BRANCH_BB_V32)
155 (BRANCH_WF_V32, BRANCH_WB_V32): New.
156 (BRANCH_BF, BRANCH_BB, BRANCH_WF, BRANCH_WB): Don't undef after
157 use in md_cris_relax_table.
158 (md_cris_relax_table): Add entries for STATE_COND_BRANCH_V32,
159 STATE_COND_BRANCH_COMMON, STATE_ABS_BRANCH_V32, STATE_LAPC.
160 Update and improve head comment.
161 (OPTION_PIC): Define in terms of previous option, OPTION_US.
162 (OPTION_MULBUG_ABORT_ON, OPTION_MULBUG_ABORT_OFF): Similar.
164 (md_longopts): New option --march=...
165 (cris_any_v0_v10_long_jump_size, crisv32_long_jump_size): New
167 (md_long_jump_size): Initialize in terms of DEFAULT_CRIS_ARCH.
168 (HANDLE_RELAXABLE): New macro.
169 (md_estimate_size_before_relax): Use HANDLE_RELAXABLE for common
170 cases. Check for weak symbols and assume not relaxable. Handle
171 STATE_COND_BRANCH_V32, STATE_COND_BRANCH_COMMON,
172 STATE_ABS_BRANCH_V32, STATE_LAPC. Use new variable symbolP, not
174 (md_convert_frag): Handle STATE_COND_BRANCH_V32,
175 STATE_COND_BRANCH_COMMON, STATE_ABS_BRANCH_V32, STATE_LAPC.
176 (cris_create_short_jump): Adjust for CRISv32.
177 (cris_relax_frag): Handle new states.
178 (md_create_long_jump): Ditto. Emit error for common_v10_v32.
179 (md_begin): Define symbols "..asm.arch.cris.v32",
180 "..asm.arch.cris.v10", "..asm.arch.cris.common_v10_v32" and
181 "..asm.arch.cris.any_v0_v10". Use cris_insn_ver_valid_for_arch
182 when entering opcode table entry points.
183 (md_assemble): Adjust branch handling for CRISv32. Handle LAPC
184 relaxation. In fix_new_exp call for main insn, pass 1 for pcrel
185 parameter for 8, 16 and 32-bit pc-relative insns and LAPC.
186 (cris_process_instruction): Initialize out_insnp->insn_type to
187 CRIS_INSN_NONE, not CRIS_INSN_NORMAL.
188 <case ']', '[', 'A', 'd', 'Q', 'N', 'n', 'Y', 'U', 'u', 'T'>: New
190 <case 'm'>: Check that modified_char == '.'.
191 <invalid operands>: Consume the rest of the line.
192 When operands don't match, skip over subsequent insns with
193 non-matching version specifier but same mnemonic.
194 <immediate constant, case SIZE_SPEC_REG>: Immediate operands for
195 special registers in CRISv32 are always 32 bit long.
196 <immediate constant, case SIZE_FIELD_SIGNED, SIZE_FIELD_UNSIGNED>:
198 (get_gen_reg): Only recognize "PC" when followed by "+]" for v32
199 and compatible. Recognize "ACR" for v32, unless followed by "+".
200 (get_spec_reg): Consider cris_arch when looking up register.
201 (get_autoinc_prefix_or_indir_op): Don't recognize assignment for
203 (get_3op_or_dip_prefix_op): Check for ']' after seeing '[rN+'.
204 (cris_get_expression): Restore input_line_pointer if failing "early".
205 (get_flags): Consider cris_arch and recognize flags accordingly.
206 (branch_disp): Adjust for CRISv32.
207 (gen_cond_branch_32): Similar. Emit error for common_v10_v32.
208 (cris_number_to_imm): Use as_bad_where, not as_bad. Remove
209 related FIXME. Don't insist on BFD_RELOC_32_PCREL fixup to be
210 resolved. Don't enter zeros in object file for
212 <case BFD_RELOC_CRIS_LAPCQ_OFFSET, BFD_RELOC_CRIS_SIGNED_16>
213 <case BFD_RELOC_CRIS_SIGNED_8>: New case.
214 (md_parse_option): Break out "return 1".
215 <OPTION_ARCH> New case.
216 (tc_gen_reloc): <case BFD_RELOC_CRIS_LAPCQ_OFFSET>
217 <case BFD_RELOC_CRIS_SIGNED_16, BFD_RELOC_CRIS_SIGNED_8>
218 <case BFD_RELOC_CRIS_UNSIGNED_8, BFD_RELOC_CRIS_UNSIGNED_16>
219 <case BFD_RELOC_32_PCREL>: New cases.
220 Addends for non-zero fx_pcrel are too in fx_offset.
221 (md_show_usage): Show --march=<arch>.
222 (md_apply_fix3): Adjust val for BFD_RELOC_CRIS_LAPCQ_OFFSET.
223 (md_pcrel_from): BFD_RELOC_CRIS_LAPCQ_OFFSET is PC-relative too.
224 (s_syntax) <struct syntaxes>: Properly constify member operand.
225 * config/tc-cris.h (TARGET_MACH): Define.
226 (cris_mach): Declare.
227 * doc/as.texinfo (Overview) <CRIS>: Add --march=...
228 * doc/c-cris.texi (CRIS-Symbols): New node for built-in symbols.
229 (CRIS-Opts): Document --march=...
230 (CRIS-Pseudos): Document .arch.
234 * config/tc-i386.c (set_intel_syntax): Allow % in symbol names when
235 intel syntax and no register prefix, allow $ in symbol names when
237 (set_16bit_gcc_code_flag): Replace literal 'l' by LONG_MNEM_SUFFIX.
238 (intel_float_operand): Add fourth return value indicating math control
239 operations. Make classification more precise.
240 (md_assemble): Complain if memory operand of mov[sz]x has no size
242 (parse_insn): Translate word operands to floating point instructions
243 operating on integers as well as control instructions to short ones
244 as expected by AT&T syntax. Translate 'd' suffix to short one only for
245 floating point instructions operating on non-integer operands.
246 (match_template): Remove fldcw special case. Adjust q-suffix handling
247 to permit it on fild/fistp/fisttp in AT&T mode.
248 (process_suffix): Don't guess DefaultSize insns' suffix from
249 stackop_size for certain floating point control instructions. Guess
250 suffix for branch and [ls][gi]dt based on flag_code. Split error
251 messages for Intel and AT&T syntax, and make the condition more strict
252 for the former. Adjust suppressing of generation of operand size
254 (intel parser): Allow the full set of MASM operators. Add FWORD, TBYTE,
255 OWORD, and XMMWORD operand size specifiers (TBYTE replaces XWORD). Add
257 * config/tc-i386.h (BYTE_PTR WORD_PTR DWORD_PTR QWORD_PTR XWORD_PTR
258 SHORT OFFSET_FLAT FLAT NONE_FOUND): Remove unused defines.
262 * symbols.c (colon) [!WORKING_DOT_WORD]: Don't declare
263 md_short_jump_size, md_long_jump_size.
264 * write.c [!WORKING_DOT_WORD]: Ditto.
265 * tc.h [!WORKING_DOT_WORD]: Declare them here. Drop const
267 * config/tc-cris.h (md_short_jump_size, md_long_jump_size): Don't
269 * config/tc-cris.c (md_short_jump_size, md_long_jump_size): Drop
270 const qualifier in these definitions.
271 * config/tc-i370.c, config/tc-m68k.c, config/tc-pdp11.c,
272 config/tc-s390.c, config/tc-tahoe.c, config/tc-vax.c: Ditto.
276 * dwarf2dbg.c (dwarf2_finish): Check for the existence of a file
277 table before deciding to produce a .debug_line section to match up
278 with a user provided .debug_info section.
282 * config/tc-crx.c (getreg_image): Bug fix, a return value was
283 mistakenly omitted from CRX_C_REGTYPE and CRX_CS_REGTYPE cases.
287 * config/tc-crx.c: Remove global variable 'post_inc_mode'.
288 (get_flags): New function.
289 (get_number_of_bits): Edit comments, update numeric values to
291 (process_label_constant): Don't support the colon
292 format (SYMBOL:[s|m|l]).
293 (set_cons_rparams): Support argument type 'arg_rbase'.
294 (get_operandtype): Bug fix in 'rbase' operand type parsing.
295 (handle_LoadStor): Bug fix, first handle post-increment mode.
296 (getreg_image): Remove redundant code, update according to latest
298 (print_constant): Bug fix relate to 3-word instructions.
299 (assemble_insn): Bug fix, when matching instructions, verify also
300 instruction type (not only mnemonic).
301 Add various error checking.
302 (preprocess_reglist): Support HI/LO and user registers.
306 * config/tc-ia64.c (fixup_unw_records): Don't let the "t" value
307 in an epilogue directive go negative.
312 * config/tc-ia64.c (emit_one_bundle): Decrement md.num_slots_in_use
313 after reporting template error during manual bundling. Reported
318 * Makefile.am: Run dep-am.
319 * aclocal.m4: Regenerate with automake 1.9.2.
320 * Makefile.in: Regenerate with automake 1.9.2.
321 * doc/Makefile.in: Likewise.
323 * config/tc-arm.c: Include "dw2gencfi.h".
324 (tc_arm_regname_to_dw2regnum, tc_arm_frame_initial_instructions):
326 * config/tc-arm.h (TARGET_USE_CFIPOP, DWARF2_DEFAULT_RETURN_COLUMN)
327 (DWARF2_CIE_DATA_ALIGNMENT, tc_regname_to_dw2regnum)
328 (tc_cfi_frame_initial_instructions): Define.
329 (tc_arm_regname_to_dw2regnum, tc_arm_frame_initial_instructions):
334 * config/tc-crx.c (assemble_insn): Check unsigned immediate
336 Update coprocessor id to be unsigned immediate.
340 * config/tc-i386.c (O_secrel): Delete.
341 (tc_pe_dwarf2_emit_offset): New function.
342 * config/tc-i386.h (O_secrel): Define as O_md1.
343 (TC_DWARF2_EMIT_OFFSET): Define.
347 * config/tc-xstormy16.c (xstormy16_cons_fix_new): Accept and
348 ignore @fptr() directives for 4-byte fixups.
352 * Makefile.am: Run "make dep-am"
353 * Makefile.in: Regenerate.
357 * doc/c-xtensa.texi (Xtensa Options, Absolute Literals Directive):
358 Remove comments about placement of literal pools.
359 (Literal Directive): Update description of literal placement.
360 (Literal Prefix Directive): Remove statement that this does not apply
361 to absolute-mode literals. Describe new section naming scheme.
365 * config/tc-xtensa.c (xtensa_frequency_pseudo): Use set_subseg_freq.
366 (is_entry_opcode, is_movi_opcode, is_the_loop_opcode, is_jx_opcode,
367 is_windowed_return_opcode): Delete.
368 (xtensa_frob_label): Use get_subseg_target_freq.
369 (md_assemble): Inline call to is_entry_opcode.
370 (xtensa_handle_align): Inline call to get_frag_is_literal.
371 (relaxation_requirements): Inline call to is_jx_opcode.
372 (emit_single_op): Inline call to is_movi_opcode.
373 (xg_assemble_vliw_tokens): Inline calls to get_frag_is_insn,
374 get_frag_is_no_transform, is_entry_opcode, and
375 set_frag_is_specific_opcode. Use get_subseg_total_freq.
376 (xtensa_fix_a0_b_retw_frags, xtensa_fix_b_j_loop_end_frags,
377 xtensa_fix_close_loop_end_frags, relax_frag_immed, convert_frag_immed):
378 Inline calls to get_frag_is_no_transform.
379 (next_instrs_are_b_retw): Inline call to is_windowed_return_opcode.
380 (xtensa_fix_short_loop_frags): Inline calls to is_the_loop_opcode and
381 get_frag_is_no_transform.
382 (convert_frag_immed_finish_loop): Inline calls to get_expression_value
383 and set_frag_is_no_transform.
384 (get_expression_value): Delete.
385 (subseg_map struct): Rename cur_total_freq to total_freq. Rename
386 cur_target_freq to target_freq.
387 (get_subseg_info): Split out code to create a new map entry into ...
388 (add_subseg_info): ... this new function.
389 (get_last_insn_flags): Check if get_subseg_info succeeded.
390 (set_last_insn_flags): Call add_subseg_info if needed.
391 (get_subseg_total_freq, get_subseg_target_freq, set_subseg_freq): New.
392 (xtensa_reorder_segments): Compute last_sec while counting sections.
393 Remove call to get_last_sec.
394 (get_last_sec): Delete.
395 (cache_literal_section): Inline call to retrieve_literal_seg and its
396 callees, seg_present and add_seg_list.
397 (retrieve_literal_seg, seg_present, add_seg_list): Delete.
398 (get_frag_is_insn, get_frag_is_no_transform,
399 set_frag_is_specific_opcode, set_frag_is_no_transform): Delete.
400 * config/tc-xtensa.h (MAX_SLOTS): Reduce from 31 to 15.
404 * config/tc-xtensa.c: Use ISO C90 formatting.
405 * config/tc-xtensa.h: Likewise.
406 * config/xtensa-istack.h: Likewise.
407 * config/xtensa-relax.c: Likewise.
408 * config/xtensa-relax.h: Likewise.
412 * config/tc-arm.c (md_begin): Change EF_ARM_EABI_VER3 to
415 * doc/c-arm.texi: Document that we actually support -meabi=4, not
420 * doc/as.texinfo (VTableEntry, VTableInherit): Add "directive" to index
422 (Acknowledgements): Use "GAS" instead of AS variable.
426 * config/tc-i386.c: Include "elf/x86-64.h".
427 (i386_elf_section_type): New function.
428 * config/tc-i386.h (md_elf_section_type): Define.
429 (i386_elf_section_type): New prototype.
433 * config/m68k-parse.h (enum m68k_register): New control register,
435 * config/tc-m68k.c: Correct control register set for MCF5249.
440 * config/tc-xtensa.c (absolute_literals_supported): New global flag.
441 (UNREACHABLE_MAX_WIDTH): Define.
442 (XTENSA_FETCH_WIDTH): Delete.
443 (cur_vinsn, xtensa_fetch_width, xt_saved_debug_type, past_xtensa_end,
444 prefer_const16, prefer_l32r): New global variables.
445 (LIT4_SECTION_NAME): Define.
446 (lit4_state struct): Add lit4_seg_name and lit4_seg fields.
447 (XTENSA_PROP_*, GET_XTENSA_PROP_*, SET_XTENSA_PROP_*): Define.
448 (frag_flags struct): New.
449 (xtensa_block_info struct): Move from tc-xtensa.h. Add flags field.
450 (subseg_map struct): Add cur_total_freq and cur_target_freq fields.
451 (bitfield, bit_is_set, set_bit, clear_bit): Define.
452 (MAX_FORMATS): Define.
453 (op_placement_info struct, op_placement_table): New.
454 (O_pltrel, O_hi16, O_lo16): Define.
455 (directiveE enum): Rename directive_generics to directive_transform.
456 Delete directive_relax. Add directive_schedule,
457 directive_absolute_literals, and directive_last_directive.
458 (directive_info): Rename "generics" to "transform". Delete "relax".
459 Add "schedule" and "absolute-literals".
460 (directive_state): Adjust entries to match changes in directive_info.
461 (xtensa_relax_statesE, RELAX_IMMED_MAXSTEPS): Move to tc-xtensa.h.
462 (xtensa_const16_opcode, xtensa_movi_opcode, xtensa_movi_n_opcode,
463 xtensa_l32r_opcode, xtensa_nop_opcode, xtensa_rsr_lcount_opcode): New.
464 (xtensa_j_opcode, xtensa_rsr_opcode): Delete.
465 (align_only_targets, software_a0_b_retw_interlock,
466 software_avoid_b_j_loop_end, maybe_has_b_j_loop_end,
467 software_avoid_short_loop, software_avoid_close_loop_end,
468 software_avoid_all_short_loops, specific_opcode): Delete.
469 (warn_unaligned_branch_targets): New.
470 (workaround_a0_b_retw, workaround_b_j_loop_end, workaround_short_loop,
471 workaround_close_loop_end, workaround_all_short_loops): Default FALSE.
472 (option_[no_]link_relax, option_[no_]transform,
473 option_[no_]absolute_literals, option_warn_unaligned_targets,
474 option_prefer_l32r, option_prefer_const16, option_target_hardware):
476 (option_[no_]align_only_targets, option_literal_section_name,
477 option_text_section_name, option_data_section_name,
478 option_bss_section_name, option_eb, option_el): Delete.
479 (md_longopts): Add entries for: [no-]transform, [no-]absolute-literals,
480 warn-unaligned-targets, prefer-l32r, prefer-const16, [no-]link-relax,
481 and target-hardware. Delete entries for [no-]target-align-only,
482 literal-section-name, text-section-name, data-section-name, and
484 (md_parse_option): Handle new options and remove old ones. Accept but
485 ignore [no-]density options. Warn for [no-]generics and [no-]relax
486 and treat them as [no-]transform.
487 (md_show_usage): Add new options and remove old ones.
488 (xtensa_setup_hw_workarounds): New.
489 (md_pseudo_table): Change "word" entry to use xtensa_elf_cons. Add
490 "long", "short", "loc" and "frequency" entries.
491 (use_generics): Rename to ...
492 (use_transform): ... this function. Add past_xtensa_end check.
493 (use_longcalls): Add past_xtensa_end check.
494 (code_density_available, can_relax): Delete.
495 (do_align_targets): New.
496 (get_directive): Accept dashes in directive names. Warn about
497 [no-]generics and [no-]relax directives and treat them as
499 (xtensa_begin_directive): Call md_flush_pending_output only for some
500 directives. Check for directives inside instruction bundles. Warn
501 about deprecated ".begin literal" usage. Warn and ignore [no-]density
502 directives. Handle new directives. Check generating_literals flag
504 (xtensa_end_directive): Check for directives inside instruction
505 bundles. Warn and ignore [no-]density directives. Handle new
506 directives. Call xtensa_set_frag_assembly_state.
507 (xtensa_loc_directive_seen, xtensa_dwarf2_directive_loc,
508 xtensa_dwarf2_emit_insn): New.
509 (xtensa_literal_position): Call md_flush_pending_output. Do not check
510 use_literal_section flag.
511 (xtensa_literal_pseudo): Call md_flush_pending_output. Handle absolute
512 literals. Use xtensa_elf_cons to parse the expression.
513 (xtensa_literal_prefix): Do not check use_literal_section. Support
514 ".lit4" sections for absolute literals. Change prefix convention to
515 replace ".text" (or ".t" in a linkonce section). No need to call
517 (xtensa_frequency_pseudo, xtensa_elf_cons, xtensa_elf_suffix): New.
518 (expression_end): Handle closing braces and colons.
519 (PLT_SUFFIX, plt_suffix): Delete.
520 (expression_maybe_register): Use new xtensa-isa.h functions. Use
521 xtensa_elf_suffix instead of checking for plt suffix, and handle O_lo16
522 and O_hi16 expressions as well.
523 (tokenize_arguments): Handle closing braces and colons.
524 (parse_arguments): Use new xtensa-isa.h functions. Handle "invisible"
525 operands and paired register syntax.
526 (get_invisible_operands): New.
527 (xg_translate_sysreg_op): Handle new Xtensa LX RSR/WSR/XSR syntax. Use
528 new xtensa-isa.h functions.
529 (xtensa_translate_old_userreg_ops, xtensa_translate_zero_immed): New.
530 (xg_translate_idioms): Check if inside bundle. Use use_transform.
531 Handle new Xtensa LX RSR/WSR/XSR syntax. Remove code to widen density
532 instructions. Use xtensa_translate_zero_immed.
533 (operand_is_immed, operand_is_pcrel_label): Delete.
534 (get_relaxable_immed): Use new xtensa-isa.h functions.
535 (get_opcode_from_buf): Add slot parameter. Use new xtensa-isa.h
537 (xtensa_print_insn_table, print_vliw_insn): New.
538 (is_direct_call_opcode): Use new xtensa-isa.h functions.
539 (is_call_opcode, is_loop_opcode, is_conditional_branch_opcode,
540 is_branch_or_jump_opcode): Delete.
541 (is_movi_opcode, decode_reloc, encode_reloc, encode_alt_reloc): New.
542 (opnum_to_reloc, reloc_to_opnum): Delete.
543 (xtensa_insnbuf_set_operand, xtensa_insnbuf_get_operand): Use new
544 xtensa-isa.h functions. Operate on one slot of an instruction.
545 (xtensa_insnbuf_set_immediate_field, is_negatable_branch,
546 xg_get_insn_size): Delete.
547 (xg_get_build_instr_size): Use xg_get_single_size.
548 (xg_is_narrow_insn, xg_is_single_relaxable_insn): Update calls to
549 xg_build_widen_table. Use xg_get_single_size.
550 (xg_get_max_narrow_insn_size): Delete.
551 (xg_get_max_insn_widen_size, xg_get_max_insn_widen_literal_size,
552 xg_is_relaxable_insn): Update calls to xg_build_widen_table. Use
554 (xg_build_to_insn): Record the loc field. Handle OP_OPERAND_HI16U and
555 OP_OPERAND_LOW16U. Check xg_valid_literal_expression.
556 (xg_expand_to_stack, xg_expand_narrow): Update calls to
557 xg_build_widen_table. Use xg_get_single_size.
558 (xg_immeds_fit): Use new xtensa-isa.h functions. Update call to
560 (xg_symbolic_immeds_fit): Likewise. Also handle O_lo16 and O_hi16, and
561 treat weak symbols conservatively.
562 (xg_check_operand): Use new xtensa-isa.h functions.
563 (is_dnrange): Delete.
564 (xg_assembly_relax): Inline previous calls to tinsn_copy.
565 (xg_finish_frag): Specify separate relax states for the frag and slot0.
566 (is_branch_jmp_to_next, xg_add_branch_and_loop_targets): Use new
567 xtensa-isa.h functions.
568 (xg_instruction_matches_option_term, xg_instruction_matches_or_options,
569 xg_instruction_matches_options): New.
570 (xg_instruction_matches_rule): Handle O_register expressions. Call
571 xg_instruction_matches_options.
572 (transition_rule_cmp): New.
573 (xg_instruction_match): Update call to xg_build_simplify_table.
574 (xg_build_token_insn): Record loc fields.
575 (xg_simplify_insn): Check is_specific_opcode field and
576 density_supported flag.
577 (xg_expand_assembly_insn): Skip checking code_density_available. Use
578 new xtensa-isa.h functions. Call use_transform instead of can_relax.
579 (xg_assemble_literal): Add error handling for O_big. Call
580 record_alignment. Handle O_pltrel.
581 (xg_valid_literal_expression): New.
582 (xg_assemble_literal_space): Add slot parameter. Remove call to
583 set_expr_symbol_offset. Add call to record_alignment. Update call to
585 (xg_emit_insn): Delete.
586 (xg_emit_insn_to_buf): Add format parameter. Update calls to
587 xg_add_opcode_fix and xtensa_insnbuf_to_chars.
588 (xg_add_opcode_fix): Change opcode parameter to tinsn and add format
589 and slot parameters. Handle new "alternate" relocations for absolute
590 literals and CONST16 instructions. Check for bad uses of O_lo16 and
591 O_hi16. Use new xtensa-isa.h functions.
592 (xg_assemble_tokens): Delete.
593 (is_register_writer): Use new xtensa-isa.h functions.
594 (is_bad_loopend_opcode): Check for xtensa_rsr_lcount_opcode instead of
595 old-style RSR from LCOUNT.
596 (next_frag_opcode): Delete.
597 (next_frag_opcode_is_loop, next_frag_format_size, frag_format_size,
598 update_next_frag_state): New.
599 (update_next_frag_nop_state): Delete.
600 (next_frag_pre_opcode_bytes): Use next_frag_opcode_is_loop.
601 (xtensa_mark_literal_pool_location): Check use_literal_section flag and
602 the state of the absolute-literals directive. Add calls to
603 record_alignment and xtensa_set_frag_assembly_state. Call
604 xtensa_switch_to_non_abs_literal_fragment instead of
605 xtensa_switch_to_literal_fragment.
607 (assemble_nop): Use build_nop. Update call to xtensa_insnbuf_to_chars.
608 (get_expanded_loop_offset): Change check for undefined opcode to an
610 (xtensa_set_frag_assembly_state, relaxable_section,
611 xtensa_find_unmarked_state_frags, xtensa_find_unaligned_branch_targets,
612 xtensa_find_unaligned_loops, xg_apply_tentative_value): New.
613 (md_begin): Update call to xtensa_isa_init. Initialize linkrelax to 1.
614 Set lit4_seg_name. Call xg_init_vinsn. Initialize new global opcodes.
615 Call init_op_placement_info_table and xtensa_set_frag_assembly_state.
616 (xtensa_init_fix_data): New.
617 (xtensa_frob_label): Reset label symbol to the current frag. Check
618 do_align_targets and generating_literals flag. Propagate frequency
619 info to new alignment frag. Call xtensa_set_frag_assembly_state.
620 (xtensa_unrecognized_line): New.
621 (xtensa_flush_pending_output): Check if inside a bundle. Add a call
622 to xtensa_set_frag_assembly_state.
623 (error_reset_cur_vinsn): New.
624 (md_assemble): Remove check for literal frag. Remove call to
625 istack_init. Call use_transform instead of use_generics. Parse
626 explicit instruction format specifiers. Move code for
627 a0_b_retw_interlock workaround to xg_assemble_vliw_tokens. Call
628 error_reset_cur_vinsn on errors. Add call to get_invisible_operands.
629 Add dwarf2_where call. Remote automatic alignment for ENTRY
630 instructions. Move call to xtensa_clear_insn_labels to the end.
631 Rearrange to handle bundles.
632 (xtensa_cons_fix_new): Delete.
633 (xtensa_handle_align): New.
634 (xtensa_frag_init): Call xtensa_set_frag_assembly_state. Remove
635 assignment to is_no_density field.
636 (md_pcrel_from): Use new xtensa-isa.h functions. Use decode_reloc
637 instead of reloc_to_opnum. Handle "alternate" relocations.
638 (xtensa_force_relocation, xtensa_check_inside_bundle,
639 xtensa_elf_section_change_hook): New.
640 (xtensa_symbol_new_hook): Delete.
641 (xtensa_fix_adjustable): Check for difference of symbols with an
642 offset. Check for external and weak symbols.
643 (md_apply_fix3): Remove cases for XTENSA_OP{0,1,2} relocs.
644 (md_estimate_size_before_relax): Return expansion for the first slot.
645 (tc_gen_reloc): Handle difference of symbols by producing
646 XTENSA_DIFF{8,16,32} relocs and by writing the value of the difference
647 into the output. Handle new XTENSA_SLOT*_OP relocs by storing the
648 tentative values into the output when linkrelax is set.
649 (XTENSA_PROP_SEC_NAME): Define.
650 (xtensa_post_relax_hook): Call xtensa_find_unmarked_state_frags.
651 Create literal tables only if using literal sections. Create new
652 property tables instead of old instruction tables. Check for unaligned
653 branch targets and loops.
654 (finish_vinsn, find_vinsn_conflicts, check_t1_t2_reads_and_writes,
655 new_resource_table, clear_resource_table, resize_resource_table,
656 resources_available, reserve_resources, release_resources,
657 opcode_funcUnit_use_unit, opcode_funcUnit_use_stage,
658 resources_conflict, xg_find_narrowest_format, relaxation_requirements,
659 bundle_single_op, emit_single_op, xg_assemble_vliw_tokens): New.
660 (xtensa_end): Call xtensa_flush_pending_output. Set past_xtensa_end
661 flag. Update checks for workaround options. Call
662 xtensa_mark_narrow_branches and xtensa_mark_zcl_first_insns.
663 (xtensa_cleanup_align_frags): Add special case for branch targets.
664 Check for and mark unreachable frags.
665 (xtensa_fix_target_frags): Remove use of align_only_targets flag.
666 Use RELAX_LOOP_END_BYTES in special case for negatable branch at the
667 end of a zero-overhead loop body.
668 (frag_can_negate_branch): Handle instructions with multiple slots.
669 Use new xtensa-isa.h functions
670 (xtensa_mark_narrow_branches, is_narrow_branch_guaranteed_in_range,
671 xtensa_mark_zcl_first_insns): New.
672 (xtensa_fix_a0_b_retw_frags, xtensa_fix_b_j_loop_end_frags): Error if
673 transformations are disabled.
674 (next_instrs_are_b_retw): Use new xtensa-isa.h functions. Handle
675 multislot instructions.
676 (xtensa_fix_close_loop_end_frags, xtensa_fix_short_loop_frags):
677 Likewise. Also error if transformations are disabled.
678 (unrelaxed_frag_max_size): New.
679 (unrelaxed_frag_min_insn_count, unrelax_frag_has_b_j): Use new
680 xtensa-isa.h functions.
681 (xtensa_sanity_check, is_empty_loop, is_local_forward_loop): Use
682 xtensa_opcode_is_loop instead of is_loop_opcode.
683 (get_text_align_power): Replace as_fatal with assertion.
684 (get_text_align_fill_size): Iterate instead of using modulus when
686 (get_noop_aligned_address): Assert that this is for a machine-dependent
687 RELAX_ALIGN_NEXT_OPCODE frag. Use next_frag_opcode_is_loop,
688 xg_get_single_size, and frag_format_size.
689 (get_widen_aligned_address): Rename to ...
690 (get_aligned_diff): ... this function. Add max_diff parameter.
691 Remove handling of rs_align/rs_align_code frags. Use
692 next_frag_format_size, get_text_align_power, get_text_align_fill_size,
693 next_frag_opcode_is_loop, and xg_get_single_size. Compute max_diff
694 and pass it back to caller.
695 (xtensa_relax_frag): Use relax_frag_loop_align. Add code for new
696 RELAX_SLOTS, RELAX_MAYBE_UNREACHABLE, RELAX_MAYBE_DESIRE_ALIGN,
697 RELAX_FILL_NOP, and RELAX_UNREACHABLE frag types. Check relax_seen.
698 (relax_frag_text_align): Rename to ...
699 (relax_frag_loop_align): ... this function. Assume loops can only be
700 in the first slot of an instruction.
701 (relax_frag_add_nop): Use assemble_nop instead of constructing an OR
702 instruction. Remove call to frag_wane.
703 (relax_frag_narrow): Rename to ...
704 (relax_frag_for_align): ... this function. Extend to handle
705 RELAX_FILL_NOP and RELAX_UNREACHABLE, as well as RELAX_SLOTS with
706 RELAX_NARROW for the first slot.
707 (find_address_of_next_align_frag, bytes_to_stretch): New.
708 (future_alignment_required): Use find_address_of_next_align_frag and
709 bytes_to_stretch. Look ahead to subsequent frags to make smarter
711 (relax_frag_immed): Add format, slot, and estimate_only parameters.
712 Check if transformations are enabled for b_j_loop_end workaround.
713 Use new xtensa-isa.h functions and handle multislot instructions.
714 Update call to xg_assembly_relax.
715 (md_convert_frag): Handle new RELAX_SLOTS, RELAX_UNREACHABLE,
716 RELAX_MAYBE_UNREACHABLE, RELAX_MAYBE_DESIRE_ALIGN, and RELAX_FILL_NOP
718 (convert_frag_narrow): Add segP, format and slot parameters. Call
719 convert_frag_immed for branch instructions. Adjust calls to
720 tinsn_from_chars, tinsn_immed_from_frag, and xg_emit_insn_to_buf. Use
721 xg_get_single_size and xg_get_single_format.
722 (convert_frag_fill_nop): New.
723 (convert_frag_immed): Add format and slot parameters. Handle multislot
724 instructions and use new xtensa-isa.h functions. Update calls to
725 tinsn_immed_from_frag and xg_assembly_relax. Check if transformations
726 enabled for b_j_loop_end workaround. Use build_nop instead of
727 assemble_nop. Check is_specific_opcode flag. Check for unreachable
728 frags. Use xg_get_single_size. Handle O_pltrel.
729 (fix_new_exp_in_seg): Remove check for old plt flag.
730 (convert_frag_immed_finish_loop): Update calls to tinsn_from_chars and
731 xtensa_insnbuf_to_chars. Call tinsn_immed_from_frag. Change check
732 for loop opcode to an assertion. Mark all frags up to the end of the
733 loop as not transformable.
734 (get_last_insn_flags, set_last_insn_flags): Use get_subseg_info.
735 (get_subseg_info): New.
736 (xtensa_move_literals): Call xtensa_set_frag_assembly_state. Add null
738 (xtensa_switch_to_literal_fragment): Rewrite to handle absolute
739 literals and use xtensa_switch_to_non_abs_literal_fragment otherwise.
740 (xtensa_switch_to_non_abs_literal_fragment): New.
741 (cache_literal_section): Add is_code parameter and pass it through to
742 retrieve_literal_seg.
743 (retrieve_literal_seg): Add is_code parameter and use it to set the
744 flags on the literal section. Handle case where head parameter is 0.
745 (get_frag_is_no_transform, set_frag_is_specific_opcode,
746 set_frag_is_no_transform): New.
747 (xtensa_create_property_segments): Add end_property_function parameter
748 and pass it through to add_xt_block_frags. Call bfd_get_section_flags
749 and skip SEC_DEBUGGING and !SEC_ALLOC sections.
750 (xtensa_create_xproperty_segments, section_has_xproperty): New.
751 (add_xt_block_frags): Add end_property_function parameter and call it
752 if it is non-zero. Call xtensa_frag_flags_init.
753 (xtensa_frag_flags_is_empty, xtensa_frag_flags_init,
754 get_frag_property_flags, frag_flags_to_number,
755 xtensa_frag_flags_combinable, xt_block_aligned_size,
756 xtensa_xt_block_combine, add_xt_prop_frags,
757 init_op_placement_info_table, opcode_fits_format_slot,
758 xg_get_single_size, xg_get_single_format): New.
759 (istack_push): Inline call to tinsn_copy.
760 (tinsn_copy): Delete.
761 (tinsn_has_invalid_symbolic_operands): Handle O_hi16 and O_lo16 and
762 CONST16 opcodes. Handle O_big, O_illegal, and O_absent.
763 (tinsn_has_complex_operands): Handle O_hi16 and O_lo16.
764 (tinsn_to_insnbuf): Use xg_get_single_format and new xtensa-isa.h
765 functions. Handle invisible operands.
766 (tinsn_to_slotbuf): New.
767 (tinsn_check_arguments): Use new xtensa-isa.h functions.
768 (tinsn_from_chars): Add slot parameter. Rewrite using xg_init_vinsn,
769 vinsn_from_chars, and xg_free_vinsn.
770 (tinsn_from_insnbuf): New.
771 (tinsn_immed_from_frag): Add slot parameter and handle multislot
772 instructions. Handle symbol differences.
773 (get_num_stack_text_bytes): Use xg_get_single_size.
774 (xg_init_vinsn, xg_clear_vinsn, vinsn_has_specific_opcodes,
775 xg_free_vinsn, vinsn_to_insnbuf, vinsn_from_chars, expr_is_register,
776 get_expr_register, set_expr_symbol_offset_diff): New.
777 * config/tc-xtensa.h (MAX_SLOTS): Define.
778 (xtensa_relax_statesE): Move from tc-xtensa.c. Add
779 RELAX_CHECK_ALIGN_NEXT_OPCODE, RELAX_MAYBE_DESIRE_ALIGN, RELAX_SLOTS,
780 RELAX_FILL_NOP, RELAX_UNREACHABLE, RELAX_MAYBE_UNREACHABLE, and
782 (RELAX_IMMED_MAXSTEPS): Move from tc-xtensa.c.
783 (xtensa_frag_type struct): Add is_assembly_state_set,
784 use_absolute_literals, relax_seen, is_unreachable, is_specific_opcode,
785 is_align, is_text_align, alignment, and is_first_loop_insn fields.
786 Replace is_generics and is_relax fields by is_no_transform field.
787 Delete is_text and is_longcalls fields. Change text_expansion and
788 literal_expansion to arrays of MAX_SLOTS entries. Add arrays of
789 per-slot information: literal_frags, slot_subtypes, slot_symbols,
790 slot_sub_symbols, and slot_offsets. Add fr_prev field.
791 (xtensa_fix_data struct): New.
792 (xtensa_symfield_type struct): Delete plt field.
793 (xtensa_block_info struct): Move definition to tc-xtensa.h. Add
794 forward declaration here.
795 (xt_section_type enum): Delete xt_insn_sec. Add xt_prop_sec.
796 (XTENSA_SECTION_RENAME): Undefine.
797 (TC_FIX_TYPE, TC_INIT_FIX_DATA, TC_FORCE_RELOCATION, NO_PSEUDO_DOT,
798 tc_unrecognized_line, md_do_align, md_elf_section_change_hook,
799 HANDLE_ALIGN, TC_LINKRELAX_FIXUP, SUB_SEGMENT_ALIGN): Define.
800 (TC_CONS_FIX_NEW, tc_symbol_new_hook): Delete.
801 (unit_num_copies_func, opcode_num_units_func,
802 opcode_funcUnit_use_unit_func, opcode_funcUnit_use_stage_func): New.
803 (resource_table struct): New.
804 * config/xtensa-istack.h (MAX_INSN_ARGS): Increase from 6 to 10.
805 (TInsn struct): Add keep_wide, loc, fixup, record_fix, subtype,
806 literal_space, symbol, sub_symbol, offset, and literal_frag fields.
807 (tinsn_copy): Delete prototype.
808 (vliw_insn struct): New.
809 * config/xtensa-relax.c (insn_pattern_struct): Add options field.
810 (widen_spec_list): Add option conditions for density and boolean
811 instructions. Add expansions using CONST16 and conditions for using
812 CONST16 vs. L32R. Use new Xtensa LX RSR/WSR syntax. Add entries for
814 (simplify_spec_list): Add option conditions for density instructions.
815 Add entry for NOP instruction.
816 (append_transition): Add cmp function pointer parameter and use it to
817 insert the new entry in order.
818 (operand_function_LOW16U, operand_function_HI16U): New.
819 (xg_has_userdef_op_fn, xg_apply_userdef_op_fn): Handle
820 OP_OPERAND_LOW16U and OP_OPERAND_HI16U.
821 (enter_opname, split_string): Use xstrdup instead of strdup.
822 (init_insn_pattern): Initialize new options field.
823 (clear_req_or_option_list, clear_req_option_list,
824 clone_req_or_option_list, clone_req_option_list, parse_option_cond):
826 (parse_insn_pattern): Parse option conditions.
827 (transition_applies): New.
828 (build_transition): Use new xtensa-isa.h functions. Fix incorrectly
829 swapped last arguments in calls to append_constant_value_condition.
830 Call clone_req_option_list. Add warning about invalid opcode.
831 Handle LOW16U and HI16U function names.
832 (build_transition_table): Add cmp parameter and use it in calls to
833 append_transition. Use new xtensa-isa.h functions. Check
834 transition_applies before adding entries.
835 (xg_build_widen_table, xg_build_simplify_table): Add cmp parameter and
836 pass it through to build_transition_table.
837 * config/xtensa-relax.h (ReqOrOptionList, ReqOrOption, ReqOptionList,
838 ReqOption, transition_cmp_fn): New types.
839 (OpType enum): Add OP_OPERAND_LOW16U and OP_OPERAND_HI16U.
840 (transition_rule struct): Add options field.
841 * doc/as.texinfo (Overview): Update Xtensa options.
842 * doc/c-xtensa.texi (Xtensa Options): Delete --[no-]density,
843 --[no-]relax, and --[no-]generics options. Update descriptions of
844 --text-section-literals and --[no-]longcalls. Add
845 --[no-]absolute-literals and --[no-]transform.
846 (Xtensa Syntax): Add description of syntax for FLIX instructions.
847 Remove use of "generic" and "specific" terminology for opcodes.
848 (Xtensa Registers): Generalize the syntax description to include
849 user-defined register files.
850 (Xtensa Automatic Alignment): Update.
851 (Xtensa Branch Relaxation): Mention limitation of unconditional jumps.
852 (Xtensa Call Relaxation): Linker can now remove most of the overhead.
853 (Xtensa Directives): Remove confusing rules about precedence.
854 (Density Directive, Relax Directive): Delete.
855 (Schedule Directive): New.
856 (Generics Directive): Rename to ...
857 (Transform Directive): ... this node.
858 (Literal Directive): Update for absolute literals. Missing
859 literal_position directive is now an error.
860 (Literal Position Directive): Update for absolute literals.
861 (Freeregs Directive): Delete.
862 (Absolute Literals Directive): New.
863 (Frame Directive): Minor editing.
864 * Makefile.am (DEPTC_xtensa_elf, DEPOBJ_xtensa_elf, DEP_xtensa_elf):
866 * Makefile.in: Regenerate.
870 * config/tc-mips.c (append_insn): Use fix_new rather than fix_new_exp
871 to build the second and third fixups for a composite relocation.
872 (macro_read_relocs): New function.
873 (macro_build): Use it.
874 (s_cpsetup): Pass all three composite relocation codes to macro_build.
875 Simplify fragging code accordingly.
876 (s_gpdword): Use fix_new rather than fix_new_exp for the second part
877 of the composite relocation. Set fx_tcbit in both fixups.
881 * config/tc-mips.c (append_insn): Set fx_tcbit for composite relocs.
882 (md_apply_fix3): Don't treat composite relocs as done.
886 * macro.c (macro_expand_body): When ELF, use .LL rather than LL as
887 prefix for symbol names generated from the LOCAL macro directive.
889 * dw2gencfi.c (select_cie_for_fde): When separating CIE out from
890 FDE, treat a DW_CFA_remember_state as we do a DW_CFA_advance_loc.
894 * config/tc-crx.c (preprocess_reglist): Handle Co-processor
896 (md_assemble): Add error checking for Co-Processor instructions.
897 (get_cinv_parameters): Add 'b' option to invalidate the
902 * config/tc-arm.c (unwind): New variable.
903 (vfp_sp_encode_reg): New function.
904 (vfp_sp_reg_required_here): Use it.
905 (vfp_sp_reg_list, vfp_dp_reg_list): Remove.
906 (vfp_parse_reg_list): New function.
907 (s_arm_unwind_fnstart, s_arm_unwind_fnend, s_arm_unwind_cantunwind,
908 s_arm_unwind_personality, s_arm_unwind_personalityindex,
909 s_arm_unwind_handlerdata, s_arm_unwind_save, s_arm_unwind_movsp,
910 s_arm_unwind_pad, s_arm_unwind_setfp, s_arm_unwind_raw): New
912 (md_pseudo_table): Add them.
913 (do_vfp_reg2_from_sp2): Use vfp_parse_reg_list and vfp_sp_encode_reg.
914 (do_vfp_sp2_from_reg2, vfp_sp_ldstm, vfp_dp_ldstm): Ditto.
915 (set_section, add_unwind_adjustsp, flush_pending_unwind,
916 finish_unwind_opcodes, start_unwind_section, create_unwind_entry,
917 require_hashconst, add_unwind_opcode): New functions.
918 * doc/tc-arm.text: Document unwinding opcodes.
919 * NEWS: Mention the new feature.
923 * config/tc-mips.c (md_apply_fix3): Remove erroneous assert.
927 * config/tc-ppc.c (md_apply_fix3): Call S_SET_THREAD_LOCAL for
929 * config/tc-s390.c (md_apply_fix3): Likewise.
930 * config/tc-sparc.c (md_apply_fix3): Likewise.
934 * config/tc-arm.c (arm_elf_section_type): New function.
935 (arm_elf_change_section): Set section link for exidx sections.
936 * config/tc-arm.h (arm_elf_section_type): Add prototype.
937 (md_elf_section_type): Define.
941 * config/tc-pdp11.c (md_apply_fix3): Change to sign of the SOB
942 instruction's offset.
946 * (TARGET_FORMAT): Remove LynxOS COFF definition.
950 * config/tc-arc.c (tc_gen_reloc): Don't assume fixP->fx_addsy is an
951 asymbol *, instead use symbol_get_bfdsym.
955 * config/tc-m68k.c (select_control_regs): Add mcf5249.
959 * config/tc-arm.c (do_smi, do_nop): New functions.
960 (insns): Add ARMv6ZK instructions.
961 (md_apply_fix3): Handle BFD_RELOC_ARM_SMI.
962 (tc_gen_reloc): Ditto.
963 (arm_cpus): Add mpcore and arm1176.
964 (arm_archs): Add armv6{k,z,zk}.
965 * doc/c-arm.texi: Document new cores and architectures.
969 * config/tc-arm.c: Use ISO C90 formatting.
973 * config/tc-arm.c (mav_reg_required_here): Allow REG_TYPE_CN
974 as alternative when REG_TYPE_MVF, REG_TYPE_MVD, REG_TYPE_MVFX or
975 REG_TYPE_MVDX is expected.
979 * doc/c-i386.texi (i386-Mnemonics): Fix typo.
983 * config/tc-ia64.c (ENCODED_PSP_OFFSET): New.
984 (output_rp_psprel, output_pfs_psprel, output_preds_psprel,
985 output_spill_base, output_unat_psprel, output_lc_psprel,
986 output_fpsr_psprel, output_priunat_psprel, output_bsp_psprel,
987 output_bsprestore_psprel, output_rnat_psprel, output_spill_psprel,
988 output_spill_psprel_p): Use it.
992 * config/tc-crx.c (handle_LoadStor): New function.
993 Handle load/stor unique instructions before parsing.
997 * config/tc-arm.c (s_arm_rel31): New funciton.
998 (md_pseudo_table): Add .rel31.
999 (md_apply_fix3): Handle BFD_RELOC_ARM_TARGET2,
1000 BFD_RELOC_32_PCREL and BFD_RELOC_ARM_PREL31.
1001 (tc_gen_reloc): Handle BFD_RELOC_ARM_PREL31 and BFD_RELOC_ARM_TARGET2.
1002 (arm_fix_adjustable): Return 0 for BFD_RELOC_ARM_TARGET2.
1003 (arm_parse_reloc): Add (target2).
1007 * Makefile.am: Run "make dep-am".
1008 * Makefile.in: Regenerate.
1009 * aclocal.m4: Regenerate.
1010 * configure: Regenerate.
1011 * doc/Makefile.in: Regenerate.
1012 * po/POTFILES.in: Regenerate.
1013 * po/gas.pot: Regenerate.
1017 * config/tc-mmix.c [!LLONG_MIN]: Correct #elsif to #elif.
1018 [!LLONG_MAX]: Ditto.
1022 * config/tc-arm.c: Rename RELABS to TARGET1.
1026 * messages.c (as_internal_value_out_of_range): Cast values passed
1027 to as_bad_where or as_warn_where to proper type.
1031 * config/tc-avr.c: Add support for
1032 atmega48, atmega88, atmega168, attiny13, attiny2313, at90can128.
1036 * dw2gencfi.c (select_cie_for_fde): When separating CIE out
1037 from FDE, treat a CFI_escape as we do a DW_CFA_advance_loc.
1041 * config/obj-elf.c (obj_elf_section_type): Handle init_array,
1042 fini_array and preinit_array section types.
1043 * config/tc-ia64.c (ia64_elf_section_type): Remove init_array
1045 * doc/as.texinfo: Document extra section types.
1049 * Makefile.am (TARG_ENV_HFILES): Add te-symbian.h.
1050 * Makefile.in: Regenerated.
1051 * configure.in: Set em for arm*-*-symbianelf*.
1052 * configure: Regenerated.
1053 * config/tc-arm.c (elf32_arm_target_format): Use Symbian target
1054 vectors when appropriate.
1055 * config/te-symbian.h: New file.
1059 * config/tc-crx.c (gettrap): Exception vector can be case
1061 (process_label_constant): Fix a 32-bit displacement bug in branch
1063 (get_operandtype) : Bug fix, wrong operand was used.
1064 (process_label_constant): Initialize relocation type to
1069 * tc-arm.c (arm_cpus, arm_fpus): Allow <cpu>-s as well as <cpu>s
1070 for synthesizable cores.
1072 * doc/c-arm.texi (ARM Options): Document canonical names of CPUs.
1076 * config/tc-msp430.c: Clean-up the code.
1077 (md_relax_table): New relax table.
1078 (mcu_types): Sort MCU types.
1079 (md_pseudo_table): Add .profiler pseudo handler.
1080 (pow2value): New function.
1081 (msp430_profiler): New function.
1082 (msp430_operands): Add new insns handlers.
1083 (msp430_srcoperand): Add register operand handler, allow complex
1085 (md_estimate_size_before_relax): Rewritten.
1086 (md_convert_frag): Rewritten.
1087 (msp430_relax_frag): New function.
1088 * config/tc-msp430.h (md_relax_frag): define macro
1089 * doc/c-msp430.texi: Update information.
1093 * as.c (std_shortopts): Allow -g to take an optional argument.
1094 (parse_args): Pass any switch starting with -g on to the backend
1099 * configure.in (arm*-*-symbianelf*): New target.
1100 (arm*-*-eabi*): Likewise.
1101 * configure: Regenerated.
1104 * config/tc-mips.c (append_insn): Handle delay slots in branch likely
1109 * config/tc-ia64.c (start_unwind_section): Add linkonce_empty
1110 argument, don't do anything if current section is not
1111 .gnu.linkonce.t.* and linkonce_empty is set.
1112 (generate_unwind_image, dot_endp): Adjust callers, call
1113 start_unwind_section (*, 1) if nothing will be put into the
1118 * as.c (MD_DEBUG_FORMAT_SELECTOR): Provide default definition.
1119 (show_usage): Add -g.
1120 (std_longopts): Add --gen-debug. Alpha sort the table.
1121 (parse_args): Print an error message if a switch is not handled.
1122 Handle the -g switch, calling md_debug_format_selector() if
1124 * NEWS: Mention new feature.
1125 * doc/as.texinfo: Document new switch.
1126 * doc/internals.texi: Document behaviour of md_parse_option.
1128 * config/tc-arm.c (md_parse_option): Do not issue an error message
1129 if the switch is not recognised.
1130 * config/tc-m68k.c (md_parse_option): Likewise.
1131 * config/tc-pdp11.c (md_parse_option): Likewise.
1132 * config/tc-v850.c (md_parse_option): Likewise.
1134 * as.h: Fix up formatting.
1139 * macro.c (macro_set_alternate): Use ISO C90 formatting.
1141 * configure.in: Sort architecture based tables alphabetically.
1142 * configure: Regenerate.
1146 * config/tc-ppc.c (tc_ppc_regname_to_dw2regnum <regnames>): Replace
1147 { "cc", 68 }, with { "cr", 70 }.
1152 * as.c: Add and handle new --alternate command line option.
1153 * macro.c (macro_set_alternate): New.
1154 * macro.h (macro_set_alternate): Declare.
1155 * read.c: Add and handle new .altmacro and .noaltmacro directives.
1156 * doc/as.texinfo: Document new command line option and pseudo-ops
1157 as well as insert documentation originating from gasp about
1158 alternate macro syntax.
1159 * NEWS: Mention new command line option and pseudo-ops.
1163 * expr.c (operand): Handle the "~", "-", and "!" operators applied
1168 * config/tc-arm.c (md_apply_fix3, tc_gen_reloc, arm_parse_reloc):
1169 Handle new relocations.
1170 * include/elf/arm.h (elf_arm_reloc_type): Add new EABI relocations.
1174 * write.c (relax_segment): Use was_address instead of address when
1175 setting fr_fix field for align frag due to backwards .org.
1179 Introduce SH2a support.
1181 * config/tc-sh.c (get_specific): Change arch_sh2a_up to
1184 * config/tc-sh.c (md_parse_option): Add sh2a-nofpu ISA handling.
1186 * config/tc-sh.c (sh_elf_final_processing): Move sh2a recognition
1187 to end of conditional expression.
1189 * config/tc-sh.c: Add sh2a-nofpu support.
1191 * tc-sh.c: Add sh2a support.
1192 (parse_reg): Add tbr.
1193 (parse_at): Support @@(disp,tbr).
1194 (get_specific): Support sh2a opcodes.
1195 (insert4): New, for 4 byte relocs.
1196 (build_Mytes): Support sh2a opcodes.
1197 (md_apply_fix3_Mytes): Support sh2a opcodes.
1199 * config/tc-sh.c (md_parse_option): Handle sh2a.
1200 (sh_elf_final_processing): Ditto.
1204 * config/tc-hppa.h (TARGET_FORMAT): Set to "elf32-hppa-netbsd"
1209 * config/tc-ppc.c (ppc_frob_file_before_adjust): Warn if .toc too big.
1210 (ppc_arch): Expand comment.
1214 * config/tc-crx.c: Support evaluating the difference between two
1216 * config/tc-crx.h: Likewise.
1220 * config/tc-ia64.c (start_unwind_section): Set the linked-to
1222 (ia64_elf_section_change_hook): Set the linked-to section for
1227 * config/tc-msp430.c: Add new subtargets: msp430x1610,
1228 msp430x1611, msp430x1612, msp430x415, msp430x417, msp430xG437,
1229 msp430xG438, msp430xG439.
1233 * doc/as.texinfo (Section, PushSection): Correct documentation
1238 * config/tc-i386.c (optimize_imm): Adjust immediates to only those
1239 permissible for the selected instruction suffix.
1240 (match_template): Don't permit 64-bit general purpose operands in
1242 (finalize_imm): Permit 64-bit immediates.
1243 (build_modrm_byte): Don't treat 32-bit addressing in 64-bit mode
1244 specially except for the width of the used base and/or index
1245 registers. For 32-bit displacements, use sign-extended
1246 relocations only when using 64-bit addressing.
1247 Force zero displacement on rip-relative addressing when there is
1248 no other displacement.
1249 (i386_index_check): Don't treat 32-bit addressing in 64-bit mode
1250 specially except for the width of the used base and/or index
1252 (parse_register): Disallow Reg64 registers in 32-bit mode.
1254 * config/tc-i386.c: For DefaultSize instructions, don't guess a 'q'
1255 suffix if the instruction doesn't support it.
1259 * config/tc-mips.c (append_insn): Handle constant expressions with
1260 no associated relocation.
1261 (mips_ip): Cancel the expression after use for the Q format
1263 (parse_relocation): Return no relocation for unsupported
1265 (my_getSmallExpression): Return no relocation if no relocation
1270 * config/obj-som.c (adjust_stab_sections): Add prototype.
1271 (obj_som_compiler, obj_som_version, obj_som_copyright,
1272 adjust_stab_sections): Add ATTRIBUTE_UNUSED to unused arguments.
1273 * config/tc-hppa.c (update_subspace): Likewise.
1274 (is_defined_subspace): Amplify comment.
1275 * config/obj-som.h (som_frob_file): Add prototype.
1280 * subsegs.c (section_symbol): Don't create a new segment when
1281 existing segment is undefined.
1285 * config/tc-arm.c: Include include/opcode/arm.h.
1286 (ARM_EXT_*, ARM_ARCH_*, ARM_ANY, ARM_ALL, COPROC_ANY): Delete.
1287 (FPU_FPA_EXT_* FPU_VFP_EXT_*, FPU_ANY, FPU_NONE, FPU_MAVERICK): Delete.
1288 (FPU_ARCH_*): Delete.
1289 * Makefile.am: Update dependencies.
1290 * Makefile.in: Regenerate.
1294 * configure.in: Accept armbe as a big-endian arm configuration.
1295 * configure: Regenerate.
1299 * config/tc-i386.c (T_SHIFTOP): New constant.
1300 (intel_e05_1): Handle '&', '|' and T_SHIFTOP.
1301 (intel_el1): Handle '~'.
1302 (intel_get_token): Handle '<>', '&', '|' and '~'.
1306 (md_assemble): Remove spurious newline from end of as_bad error
1308 (intel_e05_1): Likewise.
1309 (intel_e11): Likewise.
1310 (intel_match_token): Likewise.
1314 * config/tc-m68k.c: Convert to C90. Remove redundant
1315 declarations. Indentation fixup.
1316 [M68KCOFF]: Include "obj-coff.h" instead of declaring
1317 obj_coff_section ourselves.
1321 * config/tc-ia64.c (default_big_endian): New.
1322 (dot_byteorder, md_begin): Use it.
1323 (md_parse_option): Set it.
1327 * configure.in: Change sh-sybmian-elf to sh-*-symbianelf.
1328 * configure: Regenerate.
1329 * NEWS: Change sh-sybmian-elf to sh-*-symbianelf.
1330 * config/tc-sh.c (sh_elf_final_processing): Use renamed version of
1331 sh_find_elf_flags if necessary.
1335 * config/tc-mips.c (mips_fix_adjustable): If the full addend is
1336 going to be split into more than one in-place addend, return 0
1337 for relocations against mergeable sections. Associate comments
1342 * Makefile.am (CPU_TYPES): Add crx.
1343 (TARGET_CPU_CFILES): Add config/tc-crx.c.
1344 (TARGET_CPU_HFILES): Add config/tc-crx.h.
1345 (DEPTC_crx_elf): New target.
1346 (DEPOBJ_crx_elf): Likewise.
1347 (DEP_crx_elf): Likewise.
1348 * Makefile.in: Regenerate.
1349 * configure.in: Add crx* target.
1350 * configure: Regenerate.
1351 * config/tc-crx.c: New file.
1352 * config/tc-crx.h: New file.
1353 * NEWS: Mention new target.
1357 * config.in: Undefine TARGET_SYMBIAN by default.
1359 * configure: Regenerate. Add sh-symbian-elf target. If
1360 selected define TARGET_SYMBIAN.
1361 * config/tc-sh.h (TARGET_FORMAT): Select a Symbian target
1362 format if TARGET_SYMBIAN has been defined.
1364 * output-file.c (output_file_create): Report the target format
1365 chosen when bfd_openw reports that it is invalid.
1367 * config/obj-coff.c (coff_pseudo_table): Only define the weak
1368 pseudo for BFD based assemblers.
1373 * config/tc-sh.c (md_assemble): Change isspace to ISSPACE.
1374 (md_parse_option): Remove redundant -isa testing.
1375 Make bfd_arch variable const.
1376 (md_show_usage): Make bfd_arch variable const.
1380 * config/tc-ia64.c (emit_one_bundle): Check and set insn_addr.
1381 * config/tc-ia64.h (md_frag_check): Define.
1385 * config/obj-coff.c (obj_coff_weak): New .weak syntax for PE weak
1387 * doc/as.texinfo (Weak): Document PE weak symbols.
1391 * config/tc-mips.c (HAVE_IN_PLACE_ADDENDS): New macro.
1392 (reloc_needs_lo_p): Only return true if HAVE_IN_PLACE_ADDENDS.
1393 (mips_frob_file): Rework so that only a single pass through the
1394 relocs is needed. Allow %lo()s to have higher offsets than their
1395 corresponding %hi()s or %got()s.
1399 * config/tc-arm.c (md_apply_fix3:BFD_RELOC_ARM_IMMEDIATE): Do not
1400 allow values which have come from undefined symbols.
1401 Always consider this fixup to have been processed as a reloc
1402 cannot be generated for it.
1406 * frags.h (struct frag): Add has_code and insn_addr fields.
1407 * write.c (cvt_frag_to_fill): Invoke md_frag_check.
1408 * config/tc-ppc.c (md_assemble): Check and set insn_addr.
1409 * config/tc-ppc.h (md_frag_check): Define.
1413 * doc/Makefile.am (info): Rename goal to...
1414 (info-local): ... this, to preserve implicit dependencies.
1415 * doc/Makefile.in: Regenerate with automake 1.8.5.
1419 * config/tc-m32r.c (md_convert_frag): Changed for @PLT.
1420 (m32r_cgen_record_fixup_exp): Changed for @GOTOFF, @GOT.
1421 (m32r_fix_adjustable): Changed for @GOTOFF, @GOT, @PLT.
1422 (tc_gen_reloc): Likewise.
1423 (m32r_end_of_match): Add for @GOTOFF, @GOT, @PLT.
1424 (m32r_parse_name): Likewise.
1425 (m32r_cgen_parse_fix_exp): Likewise.
1426 * config/tc-m32r.h (md_parse_name): Define for @GOTOFF, @GOT, @PLT.
1427 (O_PIC_reloc): Likewise.
1428 (TC_CGEN_PARSE_FIX_EXP): Likewise..
1429 * cgen.c (gas_cgen_parse_operand): Add TC_CGEN_PARSE_FIX_EXP
1430 for @GOTOFF, @GOT, @PLT.
1434 * gas/symbols.c: While discarding ordinary local absolute symbols
1435 when --strip-local-absolute is in effect, retain file symbols.
1439 * config/tc-m68k.c (mri_chip): Replace current_chip, not augment.
1440 (md_parse_option): Likewise.
1444 * config/tc-i386.c: Deal with LEX_QM the same way as with LEX_AT.
1445 * config/te-netware.h: New file.
1446 * config/te-ppcnw.h: Delete: Obsolete.
1447 * configure.in: Eliminate ill NetWare targets. Make generic
1448 NetWare target use proper emulation.
1449 * Makefile.am: Eliminate reference to obsolete te-ppcnw.h, add
1450 reference to new te-netware.h.
1451 * configure: Regenerate.
1452 * Makefile.in: Regenerate.
1456 * config/tc-s390.c (s390_insn): Avoid incorrect signed/unsigned
1457 comparison in .insn pseudo operation.
1461 * config/obj-coff.c (coff_adjust_section_syms): Use
1462 bfd_get_section_size instead of bfd_get_section_size_before_reloc.
1463 (coff_frob_section): Likewise.
1464 * config/tc-mips.c (md_apply_fix3): Likewise.
1465 * config/obj-elf.c (elf_frob_file): Use bfd_set_section_size.
1466 (elf_frob_file_after_relocs): Likewise.
1470 * config/tc-hppa.c (log2): Rename to exact_log2.
1471 (pa_next_subseg): Delete unused function.
1472 (create_new_space): Mark unused arguments with ATTRIBUTE_UNUSED.
1473 (create_new_subspace): Likewise.
1476 * config/tc-hppa.c (hppa_fix_adjustable): Allow reduction of fake
1477 labels. Fix warning.
1481 * config/tc-mn10300.h (tc_fix_adjustable): Define.
1482 * config/tc-mn10300.c (mn10300_fix_adjustable): Don't adjust debug
1483 or non-merged symbols.
1487 * config/tc-ia64.c (remove_marked_resource): Save, clear and
1488 restore the old slot when inserting srlz.i/srlz.d.
1492 * Makefile.am: Regenerate dependecies.
1493 * Makefile.in: Regenerate.
1494 * config/tc-sh.c (valid_arch): Make unsigned.
1495 (preset_target_arch): Likewise.
1496 (md_begin): Use new architecture flags system.
1497 (get_specific): Likewise.
1498 (assemble_ppi): Likewise.
1499 (md_assemble): Likewise. Also fix error check for bad opcodes.
1500 (md_parse_option): Likewise. Also generate -isa values according
1501 to the table in bfd/cpu-sh.c instead of just constants. Also
1502 allow <arch>-up ISA variants.
1503 (sh_elf_final_processing): Replace if-else chain with a call to
1504 sh_find_elf_flags().
1505 * testsuite/gas/sh/arch: New directory.
1506 * testsuite/gas/sh/arch/arch.exp: New test script.
1507 * testsuite/gas/sh/arch/arch_expected.txt: New file.
1508 * testsuite/gas/sh/arch/sh.s: New file.
1509 * testsuite/gas/sh/arch/sh2.s: New file.
1510 * testsuite/gas/sh/arch/sh-dsp.s: New file.
1511 * testsuite/gas/sh/arch/sh2e.s: New file.
1512 * testsuite/gas/sh/arch/sh3-nommu.s: New file.
1513 * testsuite/gas/sh/arch/sh3.s: New file.
1514 * testsuite/gas/sh/arch/sh3-dsp.s: New file.
1515 * testsuite/gas/sh/arch/sh3e.s: New file.
1516 * testsuite/gas/sh/arch/sh4-nommu-nofpu.s: New file.
1517 * testsuite/gas/sh/arch/sh4-nofpu.s: New file.
1518 * testsuite/gas/sh/arch/sh4.s: New file.
1519 * testsuite/gas/sh/arch/sh4a-nofpu.s: New file.
1520 * testsuite/gas/sh/arch/sh4al-dsp.s: New file.
1521 * testsuite/gas/sh/arch/sh4a.s: New file.
1525 * config/gc-m68k.c(m68k_ip): Convert mode 5 addressing
1526 with zero offset into mode 2 addressing to save a word.
1530 * config/tc-ia64.c (ar_is_in_integer_unit): Removed.
1531 (ar_is_only_in_integer_unit): New.
1532 (ar_is_only_in_memory_unit): New.
1533 (generate_unwind_image): Silence gcc on 32bit host.
1534 (md_assemble): Use ar_is_only_in_integer_unit instead of
1535 ar_is_in_integer_unit. Check AR access.
1539 * config/tc-m68k.c (md_begin): Sort the opcode table into
1541 (m68k_compare_opcode): New function to do the sorting.
1545 * config/m68k-parse.y(operand): Allow for MAC/EMAC mask
1546 addressing on MIT style operands.
1547 * config/m68k-parse.y(yylex): Allow '-&' for predecrement
1549 * config/tc-m68k.c(install_operand): Comment 'G' and 'H' type
1554 * expr.c (operand, operator): Don't reject '++' and '--'.
1558 * config/tc-mips.c (append_insn): Use ISA-encoded addresses in MIPS16
1563 * configure.in: Add ppc-*-lynxos*. Update i386-*-lynxos* to ELF.
1564 * configure: Regenerate.
1568 * dw2gencfi.c (output_cie): Handle dwarf3 format CIE entries.
1572 * configure.in (or32-*-rtems*): Switch to elf.
1573 * configure: Regenerate.
1577 * po/fr.po: Updated French translation.
1581 * doc/as.texinfo (Section): Document G and T flags to .section
1582 directive. Document the extra arguments that the G flag
1583 requires. Document the #tls flag.
1587 * subsegs.c (section_symbol): Create a new section symbol if
1588 the existing one doesn't match.
1589 * symbols.c (symbol_set_bfdsym): Don't reset BFD section symbol.
1593 * config/tc-mips.c (append_insn, mips_emit_delays): Extend -mfix-vr4120
1594 to cope with VR4181A errata MD(1) and MD(4).
1598 * NEWS: Mention .secrel32 for pe[i]-i386.
1602 * config/tc-frv.h (MAX_MEM_FOR_RS_ALIGN_CODE): New.
1603 (HANDLE_ALIGN): New.
1607 * Makefile.am (DIST_SUBDIRS): Define.
1608 * aclocal.m4: Regenerate with automake 1.8.4.
1609 * Makefile.in: Likewise.
1610 * doc/Makefile.in: Likewise.
1614 * config/tc-ia64.c (dot_serialize): Declare.
1615 (dot_serialize): New function.
1616 (md_pseudo_table): Add ".serialize.data" and
1617 ".serialize.instruction" directives.
1621 * messages (as_internal_value_out_of_range): Print a message about
1622 a value being out of range. Be consistent about whether the
1623 values are printed in decimal or hexadecimal.
1624 (as_warn_value_out_of_range): Generate a warning message about an
1626 (as_bad_value_out_of_range): Generate an error message about an
1628 * as.h: Prototype the new functions.
1629 * config/tc-alpha.c (insert_operand): Use new function.
1630 * config/tc-arc.c (arc_insert_operand): Likewise.
1631 * config/tc-mn10200.c (mn10200_insert_operand): Likewise.
1632 * config/tc-mn10300.c (mn10300_insert_operand): Likewise.
1633 * config/tc-ppc.c (ppc_insert_operand): Likewise.
1634 * config/tc-s390.c (s390_insert_operand): Likewise.
1635 * config/tc-v850.c (v850_insert_operand): Likewise.
1639 * configure.in: Set em=linux for frv-*-*linux*.
1640 * configure: Rebuilt.
1641 * config/tc-frv.h (TARGET_FORMAT): Use elf32-frvfdpic if...
1642 (frv_md_fdpic_enabled): New.
1643 * config/tc-frv.c (frv_md_fdpic_enabled): New.
1644 (DEFAULT_FDPIC): New.
1645 (frv_flags): Use DEFAULT_FDPIC.
1646 (frv_pic_flag): Likewise.
1647 (OPTION_NOPIC): New.
1648 (md_longopts): Add -mnopic.
1649 (md_parse_option): Handle it.
1650 (md_show_usage): Add -mfdpic and -mnopic.
1654 * config/tc-m68k.c: Add find_cf_chip to print list of valid
1655 chips for invalid coldfire instructions, rename selectors
1656 for ColdFire sub-variants, add 521x,5249,547x,548x and aliases,
1657 add current_chip to track which chip is referred to(including save/restore),
1658 use current_chip to select control registers, not current_arch.
1659 (md_show_usage): Add new chips.
1660 * doc/c-m68k.texi: Document new command line switches.
1664 * tc-s390.h (md_do_align, HANDLE_ALIGN): Remove.
1665 (NOP_OPCODE): Define.
1666 (s390_align_code): Remove prototype.
1667 * tc-s390.c (s390_align_code): Remove.
1671 * config/tc-ia64.c (make_unw_section_name): Removed.
1672 (start_unwind_section): New function.
1673 (generate_unwind_image): Take const segT instead of const
1675 (dot_handlerdata): Adjusted.
1676 (dot_endp): Likewise.
1680 * config/obj-elf.c (obj_elf_change_section): Allow the
1681 ".note.GNU-stack" section has SHF_EXECINSTR.
1685 * config/obj-elf.c (get_section): Return bfd_boolean.
1686 (obj_elf_change_section): Call bfd_get_section_by_name_if
1687 instead of bfd_map_over_sections.
1691 * config/obj-elf.c (get_section): New function.
1692 (obj_elf_change_section): Support multiple sections with same
1697 * config/tc-arm.c (create_register_alias): Fix typo checking for
1698 case sensitive register aliases.
1699 (co_proc_number): Use error message string in all_reg_maps[]
1701 (cp_reg_required_here): Likewise.
1702 (fp_reg_required_here): Likewise.
1706 * dwarf2dbg.c (dwarf2_finish): Add SEC_DEBUGGING to section flags.
1710 * config/tc-mips.c (HAVE_32BIT_ADDRESSES, append_insn, macro_build)
1711 (load_address, macro, mips_ip, md_parse_option)
1712 (mips_force_relocation, mips_validate_fix, md_apply_fix3)
1713 (s_change_sec, pic_need_relax, tc_gen_reloc): Remove all
1714 embedded-PIC handling, and update comments.
1715 (SWITCH_TABLE): Remove.
1716 * config/tc-mips.h (DIFF_EXPR_OK): Delete.
1717 (enum mips_pic_level): Remove EMBEDDED_PIC.
1718 (EXTERN_FORCE_RELOC): Remove embedded-PIC handling.
1719 (TC_FORCE_RELOCATION): Update comment.
1720 * ecoff.c (ecoff_build_lineno): Add comment about some code that
1721 might be safe to remove now that MIPS embedded-PIC is gone.
1725 * config/obj-som.c (obj_som_init_stab_section): Add new arguments in
1726 call to obj_set_subsection_attributes.
1727 (obj_som_init_stab_section): Likewise.
1728 * config/tc-hppa.c (default_subspace_dict): Add comdat field.
1729 (pa_def_subspaces): Provide comdat default.
1730 (pa_subspace): Handle new "comdat" parameter. Set SEC_LINK_ONCE and
1731 not SEC_IS_COMMON if section is comdat, common or dup_common. Update
1732 calls to create_new_subspace and update_subspace to pass comdat flag.
1733 (create_new_subspace, update_subspace): Add new comdat argument. Use
1734 it in calls to obj_set_subsection_attributes.
1735 * doc/c-hppa.texi (.subspa, .nsubspa): Document new comdat parameter
1736 and use of comdat, common and dup_comm parameters.
1740 * config/obj-elf.c (obj_elf_change_section): Check if the old
1741 group name is NULL before comparison.
1745 * config/tc-mips.h (mips_dwarf2_addr_size): Prototype.
1749 * config/tc-mips.c (s_mipsset): Set default CPU type for .set mipsN.
1753 * config/tc-mips.c (md_longopts): Remove -membedded-pic option.
1754 (OPTION_MEMBEDDED_PIC): Remove.
1755 (OPTION_TRAP, OPTION_BREAK, OPTION_EB, OPTION_EL)
1756 (OPTION_FP32, OPTION_GP32, OPTION_CONSTRUCT_FLOATS)
1757 (OPTION_NO_CONSTRUCT_FLOATS, OPTIONS_FP64, OPTION_GP64)
1758 (OPTION_RELAX_BRANCH, OPTION_NO_RELAX_BRANCH)
1759 (OPTION_ELF_BASE): Renumber.
1760 (md_parse_option): Remove OPTION_MEMBEDDED_PIC handling.
1761 (md_show_usage): Remove mention of -membedded-pic.
1762 * doc/as.texinfo: Remove mention of -membedded-pic.
1766 * config/tc-mips.h (USE_GLOBAL_POINTER_OPT): Remove.
1767 * config/tc-mips.c (RDATA_SECTION_NAME, mips_target_format): Remove
1769 (md_begin, mips_ip, md_parse_option, s_change_sec, s_option,
1770 s_abicalls, nopic_need_relax, tc_gen_reloc): Remove uses of
1771 USE_GLOBAL_POINTER_OPT.
1775 * config/tc-mips.c (macro): One more use of load_delay_nop.
1779 * config/tc-mips.c (load_delay_nop): New function.
1780 (load_address, macro): Use load_delay_nop() to build a nop
1781 which can be omitted with gpr_interlocks.
1785 * config/tc-mips.c (hilo_interlocks, gpr_interlocks,
1786 cop_interlocks): Remove superfluous CPU entries.
1790 * config/tc-arm.c (mav_parse_offset): Value must be multiple of 4.
1794 * NEWS: Added support for EMAC instructions and MAC/EMAC
1796 * config/m68k-parse.h: Add ACC[123], ACCEXT{01,23}, MAC/EMAC
1797 scale factor tokens, trailing_ampersand to mark mask addressing
1798 for MAC/EMAC instructions.
1799 * config/m68k-parse.y: Add options_ampersand clause, '<<',
1801 (yylex): Handle '>', '<', and '&' following '+'.
1802 * config/tc-m68k.c: Set mcfmac/mcfemac on appropriate ColdFire
1803 architectures in archs[].
1804 (m68k-ip): Add '4', 'e', 'g', 'i', cases to handle mask addressing
1805 for MAC/EMAC instructions, ACC[0123], ACCEXT{01,23}, and '<<'/'>>'
1807 (m68k_ip): Handle trailing '&' on MAC/EMAC insns.
1808 (install_operand): Fix 'n' case, Add 'F', 'f', 'G', 'H', 'I', ']'
1810 Add EMAC operands to init_table[].
1814 * config/tc-arm.h (enum mstate): Move here, add MAP_UNDEFINED
1816 (TC_SEGMENT_INFO_TYPE): Define to enum mstate.
1817 * config/tc-arm.c (enum mstate): Delete from here.
1818 (mapping_state): Remove the static mapstate variable and instead
1819 store the state in the segment. This allows a per-section mapping
1820 state. Handle and ignore MAP_UNDEFINED states.
1821 (arm_elf_change_section): Get the current mapping state from the
1823 (s_ltorg): Set the mapping state to MAP_DATA.
1824 (arm_cleanup): Use arm_elf_change_section to get the mapping state
1825 for each pool as it is emitted.
1829 * config/tc-arm.h: Formatting tidy ups.
1833 * NEWS: Note that MIPS -membedded-pic option is deprecated.
1837 * config/tc-i386.h [TE_PE] (TC_CONS_FIX_NEW): Define.
1838 * config/tc-i386.c (md_pseudo_table) [TE_PE]: Add "secrel32".
1839 [TE_PE] (O_secrel): Define.
1840 [TE_PE] (x86_pe_cons_fix_new): New.
1841 [TE_PE] (pe_directive_secrel): Likewise.
1842 (tc_gen_reloc) [TE_PE]: Support BFD_RELOC_32_SECREL.
1846 * config/tc-mips.c (mips_dwarf2_addr_size): Revert part
1847 of previous patch for fix in gcc.
1851 * config/tc-xtensa.c (xg_assembler_literal): Fix a typo.
1855 * read.c (do_align): Call md_flush_pending_output, if defined.
1859 * expr.c (operand): Correct checks for ++ and --.
1863 * config/tc-generic.c: Add some comments.
1867 * doc/c-mips.texi (-m{no-,}fix-vr4120): Renamed from
1868 -{no-}mfix-vr4122-bugs.
1869 * config/tc-mips.c (mips_fix_vr4120): Renamed from mips_fix_4122_bugs.
1870 (append_insn, mips_emit_delays): Update accordingly.
1871 (OPTION_FIX_VR4120, OPTION_NO_FIX_VR4120): Renamed from *VR4122.
1872 (md_longopts): Change -{no-,}mfix-vr4122-bugs to -m{no-,}fix-vr4120.
1873 (md_parse_option): Update after above changes.
1874 (md_show_usage): Add -mfix-vr4120.
1878 * doc/as.texinfo (Sub-Sections): Conditionalize COFF-specific use
1879 of .section directive; add a reference to the ELF .subsection
1884 * config/tc-m32r.c (md_assemble): Fixed infinite loop bug
1889 * Makefile.am: Remove mips from aout targets.
1890 * Makefile.in: Regenerate.
1891 * configure.in: Remove mips-dec-bsd* target.
1892 * configure: Regenerate.
1897 * config/tc-ppc.c (ppc_elf_suffix): Add valid32 and valid64 fields
1898 to struct map_bfd. Adjust MAP macro, and define MAP32, MAP64.
1899 Update "mapping". Restrict some @ modifiers to 32 bit.
1904 * config/tc-sh.c (dont_adjust_reloc_32): New variable.
1905 (sh_fix_adjustable): Avoid adjusting BFD_RELOC_32 when
1906 dont_adjust_reloc_32 is set.
1907 (md_longopts): Add option -renesas.
1908 (md_parse_option, md_show_usage): Likewise.
1909 * doc/c-sh.texi: Likewise.
1913 * config/tc-dlx.c (md_assemble): set fx_no_overflow flag for
1914 hi16 and lo16 fixS structs.
1915 (md_assemble): generate bit_fixS for RELOC_DLX_LO16 in
1916 exactly the same way as for RELOC_DLX_REL16.
1917 (machine_ip): properly respect LO flag in the_insn and
1918 output RELOC_DLX_LO16 rather than RELOC_DLX_16.
1919 (md_apply_fix3): apply RELOC_DLX_LO16.
1923 Remove long-obsolete MPW support.
1924 * mpw-config.in, mpw-make.sed, mac-as.r: Remove files.
1925 * configure.in: Remove mention of ppc-*-mpw* config.
1926 * configure.in: Likewise.
1930 * config/tc-arm.c (meabi_flags): Make its use conditional upon
1931 OBJ_ELF being defined.
1935 * config/obj-aout.c (obj_aout_type): Remove #ifdef BFD_ASSEMBLER code.
1939 * config/tc-arm.c (meabi_flags): New variable.
1940 (arm_parse_eabi): New function.
1941 (md_begin): Set flags for EABI v3.
1943 (arm_long_opts): Add meabi.
1944 * doc/as.texinf <ARM>: Document -meabi.
1945 * doc/c-arm.texi: Ditto.
1949 * config/tc-xtensa.c (xtensa_post_relax_hook): Create literal
1950 tables even when use_literal_section flag is not set.
1954 * config/tc-sh.c: Remove trailing whitespace.
1958 * doc/c-cris.texi (CRIS-Opts): Document --no-mul-bug-abort,
1959 --mul-bug-abort and the default behavior.
1960 * config/tc-cris.c (cris_insn_kind): New member CRIS_INSN_MUL.
1961 (err_for_dangerous_mul_placement): New variable.
1962 (STATE_MUL, OPTION_MULBUG_ABORT_ON, OPTION_MULBUG_ABORT_OFF): New
1964 (md_cris_relax_table): Have placeholder for STATE_MUL.
1965 (md_longopts): New options --mul-bug-abort and --no-mul-bug-abort.
1966 (cris_relax_frag) <case ENCODE_RELAX (STATE_MUL, STATE_BYTE)>: New
1968 (md_estimate_size_before_relax) <case ENCODE_RELAX (STATE_MUL,
1969 STATE_BYTE)>: Ditto.
1970 (md_convert_frag) <ENCODE_RELAX (STATE_MUL, STATE_BYTE)>: Check
1971 alignment and position of this frag, emit error message if
1973 (md_assemble): For a multiply insn and when checking it,
1974 transform the current frag into a special frag for that purpose.
1975 (md_parse_option) <case OPTION_MULBUG_ABORT_OFF, case
1976 OPTION_MULBUG_ABORT_ON>: Handle new options.
1980 * config/tc-xtensa.c (mark_literal_frags): New function.
1981 (xtensa_move_literals): Call mark_literal_frags for all literal
1982 segments, including init and fini literal segments.
1983 (xtensa_post_relax_hook): Swap use of xt_insn_sec and xt_literal_sec.
1987 * tc-hppa.c (cons_fix_new_hppa): Check for PC relative base type.
1988 (pa_comm): Set BSF_OBJECT in symbol flags.
1992 * Makefile.am: Run "make dep-am".
1993 * Makefile.in: Regenerate.
1994 * doc/Makefile.in: Regenerate.
1995 * config.in: Regenerate.
1996 * po/gas.pot: Regenerate.
2000 * read.c (read_a_source_file): Use demand_empty_rest_of_line.
2001 (demand_empty_rest_of_line): Issue an error here.
2002 (ignore_rest_of_line): Silently skip to end.
2003 (demand_copy_string): Issue an error, not warning.
2005 * config/obj-elf.c (obj_elf_section_name): Likewise.
2006 (obj_elf_section): Likewise.
2007 * config/tc-arc.c (arc_extoper): Remove bogus NULL checks.
2008 (arc_extinst): Likewise.
2009 * config/tc-ia64.c (dot_saveb): Use demand_empty_rest_of_line.
2010 (dot_spill): Likewise.
2011 (dot_unwabi): Likewise.
2012 (dot_prologue): Likewise.
2014 * expr.c (operand): Reject ++ and --.
2015 (operator): Likewise.
2019 * config/tc-sh.c: Include dw2gencfi.h.
2020 (sh_cfi_frame_initial_instructions): New function.
2021 (sh_regname_to_dw2regnum): Likewise.
2022 * config/tc-sh.h (DWARF2_LINE_MIN_INSN_LENGTH): Move to the end of
2024 (TARGET_USE_CFIPOP): Define.
2025 (tc_cfi_frame_initial_instructions): Likewise.
2026 (tc_regname_to_dw2regnum): Likewise.
2027 (DWARF2_DEFAULT_RETURN_COLUMN, DWARF2_CIE_DATA_ALIGNMENT): Likewise.
2028 * Makefile.am: Update dependencies.
2029 * Makefile.in: Regenerate.
2033 * configure.in: Switch sh-*-rtems* to ELF. Add sh-*-rtemscoff*.
2034 * configure: Regenerate.
2038 * read.c (s_leb128): Call md_flush_pending_output.
2042 * config/tc-i386.c (output_insn): Handle PadLock instructions.
2043 * config/tc-i386.h (CpuPadLock): New define.
2044 (CpuUnknownFlags): Added CpuPadLock.
2048 * doc/c-hppa.texi (HPPA Directives): Fix typo.
2052 * dw2gencfi.c (output_cie): Align length to 4 byte boundary.
2053 (cfi_finish): Likewise for fde.
2057 * config/tc-ia64.c (md_assemble): Properly handle NULL
2059 (ia64_handle_align): Don't abort if failed to add a stop bit.
2063 * Makefile.in: Regenerated.
2064 * aclocal.m4: Likewise.
2065 * configure: Likewise.
2066 * doc/Makefile.in: Likewise.
2070 * config/tc-ia64.c (dot_align): New.
2071 (ia64_do_align): Make it static.
2072 (md_pseudo_table): Use "dot_align" for "align".
2073 (ia64_md_do_align): Don't set align_frag here.
2074 (ia64_handle_align): Add a stop bit to the previous bundle if
2077 * config/tc-ia64.h (ia64_do_align): Removed.
2081 * config/tc-sh.c (md_parse_option): Add -isa=sh4-nofpu and
2082 -isa=sh4-nommu-nofpu options. Adjust help messages accordingly.
2083 (sh_elf_final_processing): Output BFD type sh4_nofpu if that is
2084 the most general type or the user specifically requested it.
2085 (md_assemble): Add a new error message for when an instruction
2086 is understood, but is not allowed due to an -isa option.
2090 * config/tc-ia64.c (align_frag): New.
2091 (md_assemble): Set the tc_frag_data field in align_frag for
2092 IA64_OPCODE_FIRST instructions.
2093 (ia64_md_do_align): Set align_frag.
2094 (ia64_handle_align): Add a stop bit if needed.
2096 * config/tc-ia64.h (TC_FRAG_TYPE): New.
2097 (TC_FRAG_INIT): New.
2101 * config/tc-frv.c (fr400_audio): New variable.
2102 (md_parse_option, md_show_usage): Add -mcpu=fr405 and -mcpu=fr450.
2103 (md_parse_option): Set fr400_audio for -mcpu=fr400 and -mcpu=fr405.
2104 (target_implements_insn_p): New function.
2105 (md_assemble): Report an error if the processor doesn't implement
2110 * config/tc-m32r.c (md_longopts): Added -no-bitinst option.
2111 (md_parse_option): Ditto.
2112 (OPTION_NO_SPECIAL_M32R): Added.
2113 (md_show_usage): Document it.
2114 (enable_speial_m32r): Changed a default value from 0 to 1.
2115 * doc/c-m32r.texi: Document the -no-bitinst option.
2119 * config/tc-sh.c (get_operand): Revert previous delta.
2120 (tc_gen_reloc): Check for an unknown reloc type before processing
2125 * config/tc-s390.c (s390_insn): Correct range check for opcode in
2126 .insn pseudo operation.
2130 * config/tc-sh.c (get_operand): In case of #Imm, check has been
2131 added for wrong syntax.
2135 * config/tc-mips.c (mips_dwarf2_addr_size): New.
2136 * config/tc-mips.h (DWARF2_ADDR_SIZE): Use.
2140 * config/tc-sh.c (build_Mytes): Add REG_N_D and REG_N_B01
2141 nibble types to assembler.
2145 * config/tc-iq2000.c: Add missing \n\ in multiline string literal.
2149 * config/tc-ia64.c (slot_index): New arg before_relax. Use instead of
2151 (fixup_unw_records): New arg before_relax. Pass to slot_index.
2152 (ia64_estimate_size_before_relax): New.
2153 (ia64_convert_frag): Pass 0 to fixup_unw_records. Add comment.
2154 (generate_unwind_image): Pass 1 to fixup_unw_records.
2155 * config/tc-ia64.h (ia64_estimate_size_before_relax): Declare.
2156 (md_estimate_size_before_relax): Call ia64_estimate_size_before_relax.
2160 * stabs.c (generate_asm_file): Avoid warning about use of
2161 uninitialized variable.
2165 * config/tc-ia64.c (ia64_flush_insns): In addition to prologue,
2166 body, and endp, allow unwind records which do not have a "t"
2167 (time/instruction) field.
2171 * config/tc-arm.c (do_mav_dspsc_1): Correct offset of CRn.
2172 (do_mav_dspsc_2): Likewise.
2173 Fix accumulator registers move opcodes.
2178 * dwarf2dbg.c (get_filenum): Do not read beyond allocated memory.
2182 * config/tc-ia64.h (ia64_frob_symbol): New declaration.
2183 (tc_frob_symbol): New macro definition.
2184 * config/tc-ia64.c (ia64_frob_symbol): New routine.
2188 * config/tc-arm.c (md_begin): Mark .note.gnu.arm.ident as
2193 * read.h (IGNORE_OPCODE_CASE): Do not define. Replace with ...
2194 (TC_CASE_SENSITIVE): ... this.
2195 * read.c: Replace IGNORE_OPCODE_CASE with TC_CASE_SENSITIVE.
2196 * doc/internals.texi (TC_CASE_SENSITIVE): Document.
2200 * config/tc-ia64.c (dot_endp): Delete call to output_endp.
2201 (generate_unwind_image): Re-add it here.
2205 * dwarf2dbg.c (DWARF2_ADDR_SIZE): Remove trailing ';'
2206 * read.h (SKIP_WHITESPACE): Turn into an expression.
2207 * read.c (read_a_source_file): A pseudo is removed by having a
2212 * config/tc-ia64.c (output_endp): New.
2213 (count_bits): Delete.
2214 (ia64_flush_insns, process_one_record, optimize_unw_records): Handle
2215 endp unwind records.
2216 (fixup_unw_records): Handle endp unwind records. Delete code for
2217 shortening prologue regions not followed by a body record.
2218 (dot_endp): Call add_unwind_entry to emit endp unwind record.
2219 * config/tc-ia64.h (unw_record_type): Add endp.
2223 * config/tc-ia64.c (ia64_convert_frag): Call md_number_to_chars to
2224 fill padding bytes with zeroes.
2225 (emit_one_bundle): New locals last_ptr, end_ptr. Rewrite code that
2226 sets unwind_record slot_number and slot_frag fields.
2230 * config/tc-mips.c (add_got_offset_hilo): New function.
2231 (macro): Use load_register() and add_got_offset_hilo() to load
2232 constants instead of hardcoding code sequences throughout.
2236 * config/tc-ia64.c (emit_one_bundle): Add proper indentation.
2240 * config/tc-m68k.h (EXTERN_FORCE_RELOC): Handle m68k-uclinux specially,
2242 * config/tc-m68k.c (RELAXABLE_SYMBOL): Use EXTERN_FORCE_RELOC instead
2243 of hard-coded test for TARGET_OS=elf.
2247 * config/tc-mips.c (hilo_interlocks): Change definition
2248 so that MIPS32, MIPS64 and later ISAs are included, along with
2249 the already-included machines. Update comments.
2253 * config/tc-arm.c (tc_gen_reloc): Improve error message for
2254 undefined local labels.
2258 * config/tc-mips.c (load_address, macro): Update comments about
2259 NewABI GP relaxation.
2263 * config/tc-mips.c (macro_build): Remove place and counter arguments.
2264 (mips_build_lui, macro_build_ldst_constoffset): Likewise.
2265 (mips16_macro_build, macro_build_jalr): Remove counter argument.
2266 (set_at, load_register, load_address, move_register): Likewise.
2267 (load_got_offset, add_got_offset): Likewise.
2268 Update all calls and tidy accordingly.
2272 * config/tc-mips.c (RELAX_ENCODE): Remove WARN argument.
2273 (RELAX_FIRST, RELAX_SECOND): Turn into 8-bit quantities.
2274 (RELAX_USE_SECOND): Bump to 0x10000.
2275 (RELAX_SECOND_LONGER, RELAX_NOMACRO, RELAX_DELAY_SLOT): New flags.
2276 (mips_macro_warning): New variable.
2277 (md_assemble): Wrap macro expansion in macro_start() and macro_end().
2278 (s_cpload, s_cpsetup, s_cprestore, s_cpreturn): Likewise.
2279 (relax_close_frag): Set mips_macro_warning.first_frag. Adjust use
2281 (append_insn): Update mips_macro_warning.sizes.
2282 (macro_start, macro_warning, macro_end): New functions.
2283 (macro_build): Don't emit warnings here.
2284 (macro_build_lui, md_estimate_size_before_relax): ...or here.
2285 (md_convert_frag): Check for cases where one macro alternative
2286 needs a warning and the other doesn't. Emit a warning if the
2287 longer sequence was chosen.
2291 * config/tc-mips.h (tc_frag_data_type, TC_FRAG_TYPE): Remove.
2292 * config/tc-mips.c (RELAX_ENCODE): Take three arguments: the size of
2293 the first sequence, the size of the second sequence, and a flag
2294 that says whether we should warn.
2295 (RELAX_OLD, RELAX_NEW, RELAX_RELOC[123]): Delete.
2296 (RELAX_FIRST, RELAX_SECOND): New.
2297 (mips_relax): New variable.
2298 (relax_close_frag, relax_start, relax_switch, relax_end): New fns.
2299 (append_insn): Remove "place" argument. Use mips_relax.sequence
2300 rather than "place" to check whether we're expanding the second
2301 alternative of a relaxable macro. Remove redundant check for
2302 branch relaxation. If generating a normal insn, and there
2303 is not enough room in the current frag, call relax_close_frag()
2304 to close it. Update mips_relax.sizes[]. Emit fixups for the
2305 second version of a relaxable macro. Record the first relaxable
2306 fixup in mips_relax. Remove tc_gen_reloc workaround.
2307 (macro_build): Remove all uses of "place". Use mips_relax.sequence
2308 in the same way as in append_insn.
2309 (mips16_macro_build): Remove "place" argument.
2310 (macro_build_lui): As for macro_build. Don't drop the add_symbol
2311 when generating the second version of a relaxable macro.
2312 (load_got_offset, add_got_offset): New functions.
2313 (load_address, macro): Use new relaxation machinery. Remove
2314 tc_gen_reloc workarounds.
2315 (md_estimate_size_before_relax): Set RELAX_USE_SECOND if the second
2316 version of a relaxable macro is needed. Return -RELAX_SECOND if the
2317 first version is needed.
2318 (tc_gen_reloc): Remove relaxation handling.
2319 (md_convert_frag): Go through the fixups for a relaxable macro and
2320 mark those that belong to the unneeded alternative as done. If the
2321 second alternative is needed, adjust the fixup addresses to account
2322 for the deleted first alternative.
2326 * frags.h (frag_room): Declare.
2327 * frags.c (frag_room): New function.
2328 * doc/internals.texi: Document it.
2332 * config/tc-mips.c (append_insn): Don't do r3900 interlock
2333 optimization for -mtune=r3900, as this will break on other CPUs.
2337 * config/tc-m68hc11.c (build_indexed_byte): movb and movw cannot
2338 be relaxed, use fixup.
2339 (md_apply_fix3): Use 5 bit reloc from movb and movw fixup.
2343 * config/tc-sparc.c (sparc_ip): Disallow %f32-%f63 for single
2348 * config/tc-mips.c (append_insn): Properly detect variant frags
2349 that preclude swapping of relaxed branches. Correctly swap
2350 instructions between frags when dealing with relaxed branches.
2354 * acinclude.m4: Quote names of macros to be defined by AC_DEFUN
2356 * aclocal.m4: Regenerate.
2357 * configure: Regenerate.
2361 * config/tc-h8300.c (build_bytes): Apply relaxation to bit
2366 * config/tc-mips.c (macro_build_jalr): When adding an R_MIPS_JALR
2367 reloc, reserve space for the delay slot as well as the jalr itself.
2371 * config/tc-arm.c (do_vfp_reg2_from_sp2): Rename from do_vfp_sp_reg2.
2372 (do_vfp_sp2_from_reg2): New function.
2374 (do_vfp_dp_from_reg2): Check return values properly.
2378 * config/tc-mips.c (warn_nops): Remove static variable.
2379 (macro): Remove test of warn_nops.
2380 (md_shortops): Remove 'n'.
2381 (md_parse_option): Remove 'n' case.
2382 (md_show_usage): Remove -n.
2383 * doc/as.texinfo (Overview): Remove MIPS -n option.
2384 * doc/c-mips.texi (MIPS Opts): Remove mention -n.
2385 * NEWS: Mention removal of MIPS -n option.
2387 * config/tc-mips.c (ISA_HAS_COPROC_DELAYS): Remove.
2388 (cop_interlocks): Check ISA level.
2389 (cop_mem_interlocks): Define.
2390 (reg_needs_delay): Check cop_interlocks rather than
2391 ISA_HAS_COPROC_DELAYS.
2392 (append_insn): Likewise. Use cop_mem_interlocks rather than
2393 directly checking mips_opts.isa.
2394 (mips_emit_delays): Likewise.
2398 * config/tc-ia64.c (unwind): Move next_slot_number and
2399 next_slot_frag to ...
2400 (unw_rec_list): Here.
2401 (free_list_records): Removed.
2402 (output_unw_records): Likewise.
2403 (generate_unwind_image): Make it void.
2404 (alloc_record): Initialize next_slot_number and next_slot_frag.
2405 (slot_index): Take .org, .space and .align into account.
2406 (fixup_unw_records): Don't set slot_number to 0. Use
2407 list->next_slot_number and list->next_slot_frag instead of
2408 unwind.next_slot_number and unwind.next_slot_frag.
2409 (ia64_convert_frag): New.
2410 (generate_unwind_image): Generate a rs_machine_dependent frag
2412 (emit_one_bundle): Use list->next_slot_number and
2413 list->next_slot_frag instead of unwind.next_slot_number and
2414 unwind.next_slot_frag.
2416 * config/tc-ia64.h (md_convert_frag): Defined as
2418 (md_estimate_size_before_relax): Defined as (f)->fr_var.
2423 * config/tc-frv.h (md_apply_fix3): Don't define.
2424 * config/tc-frv.c (md_apply_fix3): New. Shift/truncate %hi/%lo
2426 * config/tc-frv.h (TC_FORCE_RELOCATION_SUB_LOCAL): Define.
2428 * config/tc-frv.c (line_separator_chars): Add `!'.
2430 * config/tc-frv.c (md_assemble): Clear insn upfront.
2432 * config/tc-frv.c (OPTION_FDPIC): New macro.
2433 (md_longopts): Add mfdpic.
2434 (md_parse_option): Handle it.
2436 * config/tc-frv.c (md_cgen_lookup_reloc) <FRV_OPERAND_D12,
2437 FRV_OPERAND_S12>: Use reloc type encoded in fix-up.
2438 (frv_pic_ptr): Parse funcdesc.
2442 * doc/as.texinfo: Let texi2pod parse asconfig.texi and
2443 gasver.texi. Remove duplicate symbol definitions for texi2pod.
2447 * Makefile.am (Makefile): Move the dependency on
2448 $(BFDDIR)/configure.in to...
2449 (CONFIG_STATUS_DEPENDENCIES): ... here.
2450 (AUTOMAKE_OPTIONS): Require automake 1.8.
2451 * Makefile.in: Regenerate.
2452 * doc/Makefile.am (BASEDIR, BFDDIR): Define.
2453 (CONFIG_STATUS_DEPENDENCIES): Add a dependency on
2454 $(BFDDIR)/configure.in here as well.
2455 * doc/Makefile.in: Regenerate.
2459 * Makefile.am (install, install-info, RECURSIVE_TARGETS): Remove.
2460 * Makefile.in: Regenerate.
2461 * aclocal.m4: Regenerate.
2462 * doc/Makefile.am (install, install-info): Remove.
2463 (install-data-local): A new hook for install-info.
2464 (AUTOMAKE_OPTIONS): Require automake 1.8.
2465 * doc/Makefile.in: Regenerate.
2469 * doc/c-sh.texi: Update description about floating point behavior
2474 * configure.in: Add m68k-uClinux target.
2475 * configure: Regenerate.
2477 For older changes see ChangeLog-0203
2483 version-control: never