1 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
3 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
4 * cfgrtl.c (rtl_merge_blocks): Fix comment.
5 (cfg_layout_merge_blocks): Likewise.
6 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
8 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
10 PR rtl-optimization/60901
11 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
12 bb predecessor belongs to the same scheduling region. Adjust comment.
14 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
16 * doc/sourcebuild.texi: (dfp_hw): Document.
17 (p8vector_hw): Likewise.
18 (powerpc_eabi_ok): Likewise.
19 (powerpc_elfv2): Likewise.
20 (powerpc_htm_ok): Likewise.
21 (ppc_recip_hw): Likewise.
24 2014-05-13 Cary Coutant <ccoutant@google.com>
26 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
28 2014-05-13 David Malcolm <dmalcolm@redhat.com>
30 * gengtype-parse.c (require3): Eliminate in favor of...
32 (require_template_declaration): Update to support optional single *
35 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
36 (create_user_defined_type): Handle a single level of explicit
37 pointerness within template arguments.
38 (struct write_types_data): Add field "kind".
39 (filter_type_name): Handle "*" character.
40 (write_user_func_for_structure_ptr): Require a write_types_data
41 rather than just a prefix string, so that we can look up the kind
42 of the wtd and use it as an index into wrote_user_func_for_ptr,
43 ensuring that such functions are written at most once. Support
44 subclasses by invoking the marking function of the ultimate base
46 (write_user_func_for_structure_body): Require a write_types_data
47 rather than just a prefix string, so that we can pass this to
48 write_user_func_for_structure_ptr.
49 (write_func_for_structure): Likewise.
50 (ggc_wtd): Add initializer of new "kind" field.
53 * gengtype.h (enum write_types_kinds): New.
54 (struct type): Add field wrote_user_func_for_ptr to the "s"
57 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
59 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
60 instead of const_binop.
61 (fold_binary_loc): Likewise.
63 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
65 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
66 calculation to match get_ref_base_and_extent.
68 2014-05-13 Catherine Moore <clm@codesourcery.com>
69 Sandra Loosemore <sandra@codesourcery.com>
71 * configure.ac: Fix assembly for explicit JALR relocation check.
72 * configure: Regenerate.
74 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
76 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
77 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
78 Remove associated type declarations and initialisations.
79 (arm_expand_neon_builtin): Likewise.
80 (neon_emit_pair_result_insn): Delete.
81 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
82 * config/arm/neon.md (neon_vtrn<mode>): Delete.
83 (neon_vzip<mode>): Likewise.
84 (neon_vuzp<mode>): Likewise.
86 2014-05-13 Richard Biener <rguenther@suse.de>
89 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
90 it needs revisiting whether the call still may be tail-called.
92 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
94 * rtl.def (SYMBOL_REF): Remove middle "0" field.
95 * rtl.h (block_symbol): Reduce number of fields to 2.
96 (rtx_def): Add u2.symbol_ref_flags.
97 (SYMBOL_REF_FLAGS): Use it.
98 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
99 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
100 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
101 Lower index of SYMBOL_REF_DATA.
102 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
103 Print SYMBOL_REF_FLAGS at the same time.
104 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
106 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
108 * rtl.def (VAR_LOCATION): Remove "i" field.
109 * rtl.h (rtx_def): Add u2.var_location_status.
110 (PAT_VAR_LOCATION_STATUS): Use it.
111 (gen_rtx_VAR_LOCATION): Declare.
112 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
113 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
114 * var-tracking.c (emit_note_insn_var_location): Remove casts.
116 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
118 * rtl.def (scratch): Fix outdated comment and remove "0" field.
119 * gengtype.c (adjust_field_rtx_def): Update accordingly.
121 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
123 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
124 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
125 * rtl.h (rtx_def): Add insn_uid to u2 field.
126 (RTX_FLAG_CHECK8): Delete in favor of...
127 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
128 (INSN_DELETED_P): Update accordingly.
129 (INSN_UID): Use u2.insn_uid.
130 (INSN_CHAIN_CODE_P): Define.
131 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
132 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
133 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
134 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
135 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
137 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
138 Update indices for insn-chain rtxes.
139 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
140 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
141 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
142 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
143 * combine.c (try_combine): Likewise.
144 * ira.c (setup_prohibited_mode_move_regs): Likewise.
146 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
148 * rtl.def (REG): Remove middle field.
149 * rtl.h (rtx_def): Add orignal_regno to u2.
150 (ORIGINAL_REGNO): Use it instead of field 1.
151 (REG_ATTRS): Lower field index accordingly.
152 * gengtype.c (adjust_field_rtx_def): Remove handling of
153 ORIGINAL_REGNO. Move REG_ATTRS index down.
154 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
155 code that prints the REGNO.
157 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
159 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
162 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
164 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
166 2014-05-13 Bin Cheng <bin.cheng@arm.com>
168 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
169 (alloc_iv): Lower base expressions containing ADDR_EXPR.
171 2014-05-13 Ian Bolton <ian.bolton@arm.com>
173 * config/aarch64/aarch64-protos.h
174 (aarch64_hard_regno_caller_save_mode): New prototype.
175 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
177 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
179 2014-05-13 Christian Bruel <christian.bruel@st.com>
181 * target.def (mode_switching): New hook vector.
182 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
183 (mode_exit, modepriority_to_mode): Likewise.
184 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
185 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
186 (default_priority_to_mode): Define.
187 * targhooks.h (default_priority_to_mode): Declare.
188 * target.h: Include tm.h and hard-reg-set.h.
189 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
190 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
191 * doc/tm.texi Regenerate.
192 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
193 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
194 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
195 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
196 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
197 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
198 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
199 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
200 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after, ix86_mode_exit,
201 (ix86_mode_entry, ix86_mode_priority, ix86_emit_mode_set): Hookify.
202 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
204 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
205 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
206 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
207 * config/sh/sh.c (sh4_emit_mode_set, sh4_mode_needed): Hookify.
208 (sh4_mode_after, sh4_mode_entry, sh4_mode_exit): Likewise.
209 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
210 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
211 (epiphany_mode_priority_to_mode): Remove declaration.
212 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
213 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
214 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
216 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
217 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
218 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
220 2014-05-13 Jakub Jelinek <jakub@redhat.com>
223 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
224 is const0_rtx, return immediately. Don't test count == 0 when
227 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
229 * Makefile.in: add shrink-wrap.o.
230 * config/i386/i386.c: include "shrink-wrap.h"
231 * function.c: Likewise.
232 (requires_stack_frame_p, next_block_for_reg,
233 move_insn_for_shrink_wrap, prepare_shrink_wrap,
234 dup_block_and_redirect): Move to shrink-wrap.c
235 (thread_prologue_and_epilogue_insns): Extract three code segments
236 as functions in shrink-wrap.c
237 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
239 * shrink-wrap.c: New file.
240 * shrink-wrap.h: New file.
242 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
244 * doc/extend.texi: Reflect current numbers of pragmas. Remove
245 reference to Solaris.
247 2014-05-12 Mike Stump <mikestump@comcast.net>
250 * genattrtab.c (filename): Add.
251 (convert_set_attr_alternative): Improve error message.
252 (check_defs): Restore read_md_filename for error messages.
253 (gen_insn): Save filename.
255 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
257 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
258 -fno-local-ivars and -fivar-visibility.
259 * c-family/c.opt: Make -Wshadow also implicitly enable
262 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
264 * doc/tm.texi: Remove reference to deleted macro.
265 * doc/tm.texi.in: Likewise.
267 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
270 * config/avr/avr.c (avr_out_store_psi): Use correct constant
273 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
275 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
276 * config/arm/aout.h (License): Same.
277 * config/arm/bpabi.h (License): Same.
278 * config/arm/elf.h (License): Same.
279 * config/arm/linux-elf.h (License): Same.
280 * config/arm/linux-gas.h (License): Same.
281 * config/arm/netbsd-elf.h (License): Same.
282 * config/arm/uclinux-eabi.h (License): Same.
283 * config/arm/uclinux-elf.h (License): Same.
284 * config/arm/vxworks.h (License): Same.
286 2014-05-11 Jakub Jelinek <jakub@redhat.com>
288 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
289 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
290 number of operands to 3.
291 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
292 * tree-nested.c (convert_nonlocal_omp_clauses,
293 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
294 * gimplify.c (gimplify_scan_omp_clauses): Handle
295 OMP_CLAUSE_LINEAR_STMT.
296 * omp-low.c (lower_rec_input_clauses): Fix typo.
297 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
298 cast between Fortran boolean_type_node and C _Bool if
301 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
303 PR tree-optimization/61136
304 * wide-int.h (multiple_of_p): Define a version that doesn't return
306 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
307 integer_zerop/const_binop pair.
308 (multiple_of_p): Likewise, converting both operands to widest_int
311 2014-05-09 Teresa Johnson <tejohnson@google.com>
313 * cgraphunit.c (analyze_functions): Use correct dump file.
315 2014-05-09 Florian Weimer <fweimer@redhat.com>
317 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
319 (stack_protect_return_slot_p): New function.
320 (expand_used_vars): Call stack_protect_decl_p and
321 stack_protect_return_slot_p for -fstack-protector-strong.
323 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
326 * fold-const.c (fold_binary_loc): Changed width of mask.
328 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
330 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
331 unsigned int initializers for regno_in, regno_out.
333 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
336 * config/avr/avr.md (cc): Add new attribute set_vzn.
337 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
338 Set cc insn attribute to set_vzn instead of set_zn for alternatives
339 with INC, DEC or NEG.
340 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
341 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
342 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
344 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
347 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
349 * wide-int.cc (UTItype): Define.
350 (UDWtype): Define for appropriate W_TYPE_SIZE.
352 2014-05-09 Richard Biener <rguenther@suse.de>
354 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
355 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
356 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
357 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
358 ssa_propagate): Adjust.
360 2014-05-08 Jeff Law <law@redhat.com>
362 PR tree-optimization/61009
363 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
364 tri-state rather than a boolean. When a block is too big to
365 thread through, inform caller via negative return value.
366 (thread_across_edge): If a block was too big for normal threading,
367 then it's too big for a joiner too, so remove temporary equivalences
368 and return immediately.
370 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
371 Matthias Klose <doko@ubuntu.com>
374 * optc-gen.awk: Fix option handling for -Wunused-parameter.
376 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
379 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
381 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
384 * config/alpha/alpha.c: Include gimple-iterator.h.
385 (alpha_gimple_fold_builtin): New function. Move
386 ALPHA_BUILTIN_UMULH folding from ...
387 (alpha_fold_builtin): ... here.
388 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
390 2014-05-08 Wei Mi <wmi@google.com>
393 * config/i386/i386.c (ix86_compute_frame_layout): Update
394 preferred_stack_boundary for call, expanded from tls descriptor.
395 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
396 to depend on SP register.
397 (*tls_local_dynamic_base_32_gnu): Ditto.
398 (*tls_local_dynamic_32_once): Ditto.
399 (tls_global_dynamic_64_<mode>): Set
400 ix86_tls_descriptor_calls_expanded_in_cfun.
401 (tls_local_dynamic_base_64_<mode>): Ditto.
402 (tls_global_dynamic_32): Set
403 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
404 to depend on SP register.
405 (tls_local_dynamic_base_32): Ditto.
407 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
409 * config/arm/arm_neon.h: Update comment.
410 * config/arm/neon-docgen.ml: Delete.
411 * config/arm/neon-gen.ml: Delete.
412 * doc/arm-neon-intrinsics.texi: Update comment.
414 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
416 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
418 (vand, vorr, veor, vorn, vbic): Remove.
419 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
421 (neon_vsub_unspec): Likewise.
422 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
424 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
426 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
427 (vadd_s16): Likewise.
428 (vadd_s32): Likewise.
429 (vadd_f32): Likewise.
431 (vadd_u16): Likewise.
432 (vadd_u32): Likewise.
433 (vadd_s64): Likewise.
434 (vadd_u64): Likewise.
435 (vaddq_s8): Likewise.
436 (vaddq_s16): Likewise.
437 (vaddq_s32): Likewise.
438 (vaddq_s64): Likewise.
439 (vaddq_f32): Likewise.
440 (vaddq_u8): Likewise.
441 (vaddq_u16): Likewise.
442 (vaddq_u32): Likewise.
443 (vaddq_u64): Likewise.
445 (vmul_s16): Likewise.
446 (vmul_s32): Likewise.
447 (vmul_f32): Likewise.
449 (vmul_u16): Likewise.
450 (vmul_u32): Likewise.
452 (vmulq_s8): Likewise.
453 (vmulq_s16): Likewise.
454 (vmulq_s32): Likewise.
455 (vmulq_f32): Likewise.
456 (vmulq_u8): Likewise.
457 (vmulq_u16): Likewise.
458 (vmulq_u32): Likewise.
460 (vsub_s16): Likewise.
461 (vsub_s32): Likewise.
462 (vsub_f32): Likewise.
464 (vsub_u16): Likewise.
465 (vsub_u32): Likewise.
466 (vsub_s64): Likewise.
467 (vsub_u64): Likewise.
468 (vsubq_s8): Likewise.
469 (vsubq_s16): Likewise.
470 (vsubq_s32): Likewise.
471 (vsubq_s64): Likewise.
472 (vsubq_f32): Likewise.
473 (vsubq_u8): Likewise.
474 (vsubq_u16): Likewise.
475 (vsubq_u32): Likewise.
476 (vsubq_u64): Likewise.
478 (vand_s16): Likewise.
479 (vand_s32): Likewise.
481 (vand_u16): Likewise.
482 (vand_u32): Likewise.
483 (vand_s64): Likewise.
484 (vand_u64): Likewise.
485 (vandq_s8): Likewise.
486 (vandq_s16): Likewise.
487 (vandq_s32): Likewise.
488 (vandq_s64): Likewise.
489 (vandq_u8): Likewise.
490 (vandq_u16): Likewise.
491 (vandq_u32): Likewise.
492 (vandq_u64): Likewise.
494 (vorr_s16): Likewise.
495 (vorr_s32): Likewise.
497 (vorr_u16): Likewise.
498 (vorr_u32): Likewise.
499 (vorr_s64): Likewise.
500 (vorr_u64): Likewise.
501 (vorrq_s8): Likewise.
502 (vorrq_s16): Likewise.
503 (vorrq_s32): Likewise.
504 (vorrq_s64): Likewise.
505 (vorrq_u8): Likewise.
506 (vorrq_u16): Likewise.
507 (vorrq_u32): Likewise.
508 (vorrq_u64): Likewise.
510 (veor_s16): Likewise.
511 (veor_s32): Likewise.
513 (veor_u16): Likewise.
514 (veor_u32): Likewise.
515 (veor_s64): Likewise.
516 (veor_u64): Likewise.
517 (veorq_s8): Likewise.
518 (veorq_s16): Likewise.
519 (veorq_s32): Likewise.
520 (veorq_s64): Likewise.
521 (veorq_u8): Likewise.
522 (veorq_u16): Likewise.
523 (veorq_u32): Likewise.
524 (veorq_u64): Likewise.
526 (vbic_s16): Likewise.
527 (vbic_s32): Likewise.
529 (vbic_u16): Likewise.
530 (vbic_u32): Likewise.
531 (vbic_s64): Likewise.
532 (vbic_u64): Likewise.
533 (vbicq_s8): Likewise.
534 (vbicq_s16): Likewise.
535 (vbicq_s32): Likewise.
536 (vbicq_s64): Likewise.
537 (vbicq_u8): Likewise.
538 (vbicq_u16): Likewise.
539 (vbicq_u32): Likewise.
540 (vbicq_u64): Likewise.
542 (vorn_s16): Likewise.
543 (vorn_s32): Likewise.
545 (vorn_u16): Likewise.
546 (vorn_u32): Likewise.
547 (vorn_s64): Likewise.
548 (vorn_u64): Likewise.
549 (vornq_s8): Likewise.
550 (vornq_s16): Likewise.
551 (vornq_s32): Likewise.
552 (vornq_s64): Likewise.
553 (vornq_u8): Likewise.
554 (vornq_u16): Likewise.
555 (vornq_u32): Likewise.
556 (vornq_u64): Likewise.
558 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
560 * wide-int.cc (UTItype): Define.
561 (UDWtype): Define for appropriate W_TYPE_SIZE.
563 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
565 PR tree-optimization/59100
566 * tree-ssa-phiopt.c: Include tree-inline.h.
567 (neutral_element_p, absorbing_element_p): New functions.
568 (value_replacement): Handle conditional binary operations with a
569 neutral or absorbing element.
571 2014-05-08 Richard Biener <rguenther@suse.de>
573 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
574 folding the expression.
575 (valueize_expr): Remove.
576 (visit_reference_op_load): Do not valueize the result of
578 (simplify_binary_expression): Likewise.
579 (simplify_unary_expression): Likewise.
581 2014-05-08 Richard Biener <rguenther@suse.de>
583 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
584 looking at TYPE_ARG_TYPES.
586 2014-05-08 Richard Biener <rguenther@suse.de>
588 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
589 pointer propagation special-case.
591 2014-05-08 Bin Cheng <bin.cheng@arm.com>
593 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
594 core part of address expressions.
596 2014-05-08 Alan Modra <amodra@gmail.com>
599 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
600 loads and stores when -mno-strict-align at any alignment.
601 (expand_block_clear): Similarly. Also correct calculation of
604 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
607 * tree-complex.c (expand_complex_move): Keep line info when expanding
609 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
610 of complex expression. Use new argument to display correct location
611 for values coming from phi statement.
612 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
613 (warn_uninitialized_phi): Pass location of phi argument to
615 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
616 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
618 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
620 * config/rs6000/predicates.md (indexed_address_mem): New.
621 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
622 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
623 fpstore_ux, fpstore_u.
624 (sign_extend, indexed, update): New.
625 (cell_micro): Adjust.
626 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
627 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
628 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
629 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
630 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
631 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
632 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
633 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
634 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
635 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
636 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
637 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
638 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
639 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
640 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
642 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
643 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
644 *vsx_extract_<mode>_store): Adjust.
645 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
646 is_cracked_insn, insn_must_be_first_in_group,
647 insn_must_be_last_in_group): Adjust.
649 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
651 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
652 ppc440-fpstore): Adjust.
653 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
654 ppc476-fpstore): Adjust.
655 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
656 ppc601-fpstore): Adjust.
657 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
659 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
661 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
662 ppc7450-fpstore): Adjust.
663 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
664 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
665 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
667 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
668 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
669 cell-fpstore, cell-fpstore-update): Adjust.
670 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
671 ppce300c3_store, ppce300c3_fpstore): Adjust.
672 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
673 e500mc_fpstore): Adjust.
674 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
675 e500mc64_store, e500mc64_fpstore): Adjust.
676 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
677 e5500_fpstore): Adjust.
678 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
679 e6500_fpstore): Adjust.
680 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
682 * config/rs6000/power4.md (power4-load, power4-load-ext,
683 power4-load-ext-update, power4-load-ext-update-indexed,
684 power4-load-update-indexed, power4-load-update, power4-fpload,
685 power4-fpload-update, power4-store, power4-store-update,
686 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
688 * config/rs6000/power5.md (power5-load, power5-load-ext,
689 power5-load-ext-update, power5-load-ext-update-indexed,
690 power5-load-update-indexed, power5-load-update, power5-fpload,
691 power5-fpload-update, power5-store, power5-store-update,
692 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
694 * config/rs6000/power6.md (power6-load, power6-load-ext,
695 power6-load-update, power6-load-update-indexed,
696 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
697 power6-fpload-update, power6-store, power6-store-update,
698 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
700 * config/rs6000/power7.md (power7-load, power7-load-ext,
701 power7-load-update, power7-load-update-indexed,
702 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
703 power7-fpload-update, power7-store, power7-store-update,
704 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
706 * config/rs6000/power8.md (power8-load, power8-load-update,
707 power8-load-ext, power8-load-ext-update, power8-fpload,
708 power8-fpload-update, power8-store, power8-store-update-indexed,
709 power8-fpstore, power8-fpstore-update): Adjust.
710 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
712 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
713 titan_lsu_store, titan_lsu_fpstore): Adjust.
714 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
716 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
719 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
720 unrolled byte insns. Emit address increments after move insns.
722 2014-05-07 David Malcolm <dmalcolm@redhat.com>
724 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
725 const_gimple, rather than a gimple.
726 (gimple_call_builtin_p): Likewise, for the three variants.
728 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
729 (gimple_call_builtin_p): Likewise, for the three variants.
731 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
733 PR tree-optimization/61095
734 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
736 2014-05-07 Richard Biener <rguenther@suse.de>
738 PR tree-optimization/61034
739 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
740 (maybe_skip_until): Use translate to take into account
741 lattices when trying to do disambiguations.
742 (get_continuation_for_phi_1): Likewise.
743 (get_continuation_for_phi): Adjust for added translate arguments.
744 (walk_non_aliased_vuses): Likewise.
745 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
746 (walk_non_aliased_vuses): Likewise.
747 (call_may_clobber_ref_p_1): Declare.
748 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
749 calls. Stop early if we are only supposed to disambiguate.
750 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
752 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
754 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
755 Emit an error when the function has arguments.
757 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
759 * cfgloop.h (unswitch_loops): Remove.
760 * doc/passes.texi: Remove references to loop-unswitch.c
761 * timevar.def (TV_LOOP_UNSWITCH): Remove.
763 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
765 * tree-vect-data-refs.c (vect_grouped_load_supported): New
766 check for loads group of length 3.
767 (vect_permute_load_chain): New permutations for loads group of
769 * tree-vect-stmts.c (vect_model_load_cost): Change cost
770 of vec_perm_shuffle for the new permutations.
772 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
774 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
775 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
776 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
777 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
778 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
779 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
780 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
781 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
783 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
785 * loop-unswitch.c: Delete.
787 2014-05-07 Richard Biener <rguenther@suse.de>
789 * config.gcc: Always set need_64bit_hwint to yes.
791 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
793 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
794 of using optimize_size.
796 2014-05-06 Mike Stump <mikestump@comcast.net>
798 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
800 2014-05-06 Joseph Myers <joseph@codesourcery.com>
802 * config/i386/sse.md (*mov<mode>_internal)
803 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
804 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
805 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
806 (*<code><mode>3, *andnot<mode>3<mask_name>)
807 (<mask_codefor><code><mode>3<mask_name>): Only consider
808 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
810 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
813 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
815 * lra-constraints.c (valid_address_p): Move earlier in file.
816 Add a constraint argument to the address_info version.
817 (satisfies_memory_constraint_p): New function.
818 (satisfies_address_constraint_p): Likewise.
819 (process_alt_operands, curr_insn_transform): Use them.
820 (process_address): Pass the constraint to valid_address_p when
821 checking address operands.
823 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
825 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
826 to their respective blocks. Fix inadvertent use of "node".
828 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
830 * emit-rtl.c (init_derived_machine_modes): New functionm, split
832 (init_emit_once): ...here.
833 * rtl.h (init_derived_machine_modes): Declare.
834 * toplev.c (do_compile): Call it even if no_backend.
836 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
837 Mike Stump <mikestump@comcast.net>
838 Richard Sandiford <rdsandiford@googlemail.com>
839 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
841 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
842 (rtx_equal_for_memref_p): Update comment.
843 (adjust_offset_for_component_ref): Use wide-int interfaces.
844 * builtins.c (get_object_alignment_2): Likewise.
845 (c_readstr): Likewise.
846 (target_char_cast): Add comment.
847 (determine_block_size): Use wide-int interfaces.
848 (expand_builtin_signbit): Likewise.
849 (fold_builtin_int_roundingfn): Likewise.
850 (fold_builtin_bitop): Likewise.
851 (fold_builtin_bswap): Likewise.
852 (fold_builtin_logarithm): Use signop.
853 (fold_builtin_pow): Likewise.
854 (fold_builtin_memory_op): Use wide-int interfaces.
855 (fold_builtin_object_size): Likewise.
856 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
857 nb_iterations_estimate.
858 (record_niter_bound): Use wide-int interfaces.
859 (get_estimated_loop_iterations_int): Likewise.
860 (get_estimated_loop_iterations): Likewise.
861 (get_max_loop_iterations): Likewise.
862 * cfgloop.h: Include wide-int.h.
863 (struct nb_iter_bound): Change bound to widest_int.
864 (struct loop): Change nb_iterations_upper_bound and
865 nb_iterations_estimate to widest_int.
866 (record_niter_bound): Switch to use widest_int.
867 (get_estimated_loop_iterations): Likewise.
868 (get_max_loop_iterations): Likewise.
869 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
871 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
872 * combine.c (try_combine): Likewise.
873 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
874 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
876 (aarch64_float_const_representable_p): Likewise.
877 * config/arc/arc.c: Include wide-int.h.
878 (arc_can_use_doloop_p): Use wide-int interfaces.
879 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
880 (vfp3_const_double_index): Likewise.
881 * config/avr/avr.c (avr_out_round): Likewise.
882 (avr_fold_builtin): Likewise.
883 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
884 (bfin_can_use_doloop_p): Likewise.
885 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
886 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
887 * config/i386/i386.c: Include wide-int.h.
888 (ix86_data_alignment): Use wide-int interfaces.
889 (ix86_local_alignment): Likewise.
890 (ix86_emit_swsqrtsf): Update real_from_integer.
891 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
892 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
893 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
894 (zero_constant): Likewise.
895 (input_operand): Likewise.
896 (splat_input_operand): Likewise.
897 (non_logical_cint_operand): Change const_double to const_wide_int.
898 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
899 (easy_altivec_constant): Remove comment.
900 (paired_expand_vector_init): Use CONSTANT_P.
901 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
902 (rs6000_emit_move): Update checks.
903 (rs6000_aggregate_candidate): Use wide-int interfaces.
904 (rs6000_expand_ternop_builtin): Likewise.
905 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
906 (rs6000_assemble_integer): Likewise.
907 (rs6000_hash_constant): Likewise.
908 (output_toc): Likewise.
909 (rs6000_rtx_costs): Likewise.
910 (rs6000_emit_swrsqrt); Update call to real_from_integer.
911 * config/rs6000/rs6000-c.c: Include wide-int.h.
912 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
913 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
914 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
915 Handle CONST_WIDE_INT.
916 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
917 Use tree_fits_uhwi_p.
918 * config/sparc/sparc.c: Include wide-int.h.
919 (sparc_fold_builtin): Use wide-int interfaces.
920 * config/vax/vax.c: Include wide-int.h.
921 (vax_float_literal): Use real_from_integer.
922 * coretypes.h (struct hwivec_def): New.
925 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
926 (equiv_constant): Handle CONST_WIDE_INT.
927 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
928 (cselib_hash_rtx): Handle CONST_WIDE_INT.
929 * dbxout.c (stabstr_U): Use wide-int interfaces.
930 (dbxout_type): Update to use cst_fits_shwi_p.
931 * defaults.h (LOG2_BITS_PER_UNIT): Define.
932 (TARGET_SUPPORTS_WIDE_INT): Add default.
933 * dfp.c: Include wide-int.h.
934 (decimal_real_to_integer2): Use wide-int interfaces and rename to
935 decimal_real_to_integer.
936 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
937 decimal_real_to_integer.
938 * doc/generic.texi (Constant expressions): Update for wide_int.
939 * doc/rtl.texi (const_double): Likewise.
940 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
941 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
942 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
943 (REAL_VALUE_FROM_INT): Remove.
944 (TARGET_SUPPORTS_WIDE_INT): New.
945 * doc/tm.texi: Regenerate.
946 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
947 * double-int.h: Include wide-int.h.
948 (struct wi::int_traits): New.
949 * dwarf2out.c (get_full_len): New.
950 (dw_val_equal_p): Add case dw_val_class_wide_int.
951 (size_of_loc_descr): Likewise.
952 (output_loc_operands): Likewise.
953 (insert_double): Remove.
954 (insert_wide_int): New.
956 (print_die): Add case dw_val_class_wide_int.
957 (attr_checksum): Likewise.
958 (attr_checksum_ordered): Likewise.
959 (same_dw_val_p): Likewise.
960 (size_of_die): Likewise.
961 (value_format): Likewise.
962 (output_die): Likewise.
963 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
965 (clz_loc_descriptor): Use wide-int interfaces.
966 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
967 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
968 (round_up_to_align): Use wide-int interfaces.
969 (field_byte_offset): Likewise.
970 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
971 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
972 CONST_DOUBLE handling. Use wide-int interfaces.
973 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
974 (gen_enumeration_type_die): Use add_AT_wide.
975 (hash_loc_operands): Add case dw_val_class_wide_int.
976 (compare_loc_operands): Likewise.
977 * dwarf2out.h: Include wide-int.h.
979 (enum dw_val_class): Add dw_val_class_wide_int.
980 (struct dw_val_struct): Add val_wide.
981 * emit-rtl.c (const_wide_int_htab): New.
982 (const_wide_int_htab_hash): New.
983 (const_wide_int_htab_eq): New.
984 (lookup_const_wide_int): New.
985 (const_double_htab_hash): Use wide-int interfaces.
986 (const_double_htab_eq): Likewise.
987 (rtx_to_double_int): Conditionally compile for wide-int.
988 (immed_double_int_const): Rename to immed_wide_int_const and
990 (immed_double_const): Conditionally compile for wide-int.
991 (init_emit_once): Use wide-int interfaces.
992 * explow.c (plus_constant): Likewise.
993 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
994 (lshift_value): Use wide-int interfaces.
995 (expand_mult): Likewise.
996 (choose_multiplier): Likewise.
997 (expand_smod_pow2): Likewise.
998 (make_tree): Likewise.
999 * expr.c (convert_modes): Consolidate handling of constants.
1000 Use wide-int interfaces.
1001 (emit_group_load_1): Add note.
1002 (store_expr): Update comment.
1003 (get_inner_reference): Use wide-int interfaces.
1004 (expand_constructor): Update comment.
1005 (expand_expr_real_2): Use wide-int interfaces.
1006 (expand_expr_real_1): Likewise.
1007 (reduce_to_bit_field_precision): Likewise.
1008 (const_vector_from_tree): Likewise.
1009 * final.c: Include wide-int-print.h.
1010 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1011 * fixed-value.c: Include wide-int.h.
1012 (fixed_from_string): Use wide-int interfaces.
1013 (fixed_to_decimal): Likewise.
1014 (fixed_convert_from_real): Likewise.
1015 (real_convert_from_fixed): Likewise.
1016 * fold-const.h (mem_ref_offset): Return an offset_int.
1017 (div_if_zero_remainder): Remove code parameter.
1018 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1019 Use wide-int interfaces.
1020 (may_negate_without_overflow_p): Use wide-int interfaces.
1021 (negate_expr_p): Likewise.
1022 (fold_negate_expr): Likewise.
1023 (int_const_binop_1): Likewise.
1024 (const_binop): Likewise.
1025 (fold_convert_const_int_from_int): Likewise.
1026 (fold_convert_const_int_from_real): Likewise.
1027 (fold_convert_const_int_from_fixed): Likewise.
1028 (fold_convert_const_fixed_from_int): Likewise.
1029 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1030 (sign_bit_p): Use wide-int interfaces.
1031 (make_range_step): Likewise.
1032 (build_range_check): Likewise. Pass an integer of the correct type
1033 instead of using integer_one_node.
1034 (range_predecessor): Pass an integer of the correct type instead
1035 of using integer_one_node.
1036 (range_successor): Likewise.
1037 (merge_ranges): Likewise.
1038 (unextend): Use wide-int interfaces.
1039 (extract_muldiv_1): Likewise.
1040 (fold_div_compare): Likewise.
1041 (fold_single_bit_test): Likewise.
1042 (fold_sign_changed_comparison): Likewise.
1043 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1044 (fold_plusminus_mult_expr): Use wide-int interfaces.
1045 (native_encode_int): Likewise.
1046 (native_interpret_int): Likewise.
1047 (fold_unary_loc): Likewise.
1048 (pointer_may_wrap_p): Likewise.
1049 (size_low_cst): Likewise.
1050 (mask_with_tz): Likewise.
1051 (fold_binary_loc): Likewise.
1052 (fold_ternary_loc): Likewise.
1053 (multiple_of_p): Likewise.
1054 (tree_call_nonnegative_warnv_p): Update calls to
1055 tree_int_cst_min_precision and real_from_integer.
1056 (fold_negate_const): Use wide-int interfaces.
1057 (fold_abs_const): Likewise.
1058 (fold_relational_const): Use tree_int_cst_lt.
1059 (round_up_loc): Use wide-int interfaces.
1060 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1061 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1062 * gengtype.c: Remove include of double-int.h.
1063 (do_typedef): Use wide-int interfaces.
1064 (open_base_files): Add wide-int.h.
1065 (main): Add offset_int and widest_int typedefs.
1066 * gengtype-lex.l: Handle "^".
1067 (CXX_KEYWORD): Add "static".
1068 * gengtype-parse.c (require3): New.
1069 (require_template_declaration): Handle constant template arguments
1070 and nested templates.
1071 * gengtype-state.c: Don't include "double-int.h".
1072 * genpreds.c (write_one_predicate_function): Update comment.
1073 (write_tm_constrs_h): Add check for hval and lval use in
1075 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1076 (add_to_sequence): Likewise.
1077 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1078 and const_double_operand.
1079 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1081 * gimple-fold.c (get_base_constructor): Likewise.
1082 (fold_array_ctor_reference): Likewise.
1083 (fold_nonarray_ctor_reference): Likewise.
1084 (fold_const_aggregate_ref_1): Likewise.
1085 (gimple_val_nonnegative_real_p): Likewise.
1086 (gimple_fold_indirect_ref): Likewise.
1087 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1088 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1089 (struct slsr_cand_d): Change index to be widest_int.
1090 (struct incr_info_d): Change incr to be widest_int.
1091 (alloc_cand_and_find_basis): Use wide-int interfaces.
1092 (slsr_process_phi): Likewise.
1093 (backtrace_base_for_ref): Likewise. Return a widest_int.
1094 (restructure_reference): Take a widest_int instead of a double_int.
1095 (slsr_process_ref): Use wide-int interfaces.
1096 (create_mul_ssa_cand): Likewise.
1097 (create_mul_imm_cand): Likewise.
1098 (create_add_ssa_cand): Likewise.
1099 (create_add_imm_cand): Take a widest_int instead of a double_int.
1100 (slsr_process_add): Use wide-int interfaces.
1101 (slsr_process_cast): Likewise.
1102 (slsr_process_copy): Likewise.
1103 (dump_candidate): Likewise.
1104 (dump_incr_vec): Likewise.
1105 (replace_ref): Likewise.
1106 (cand_increment): Likewise. Return a widest_int.
1107 (cand_abs_increment): Likewise.
1108 (replace_mult_candidate): Take a widest_int instead of a double_int.
1109 (replace_unconditional_candidate): Use wide-int interfaces.
1110 (incr_vec_index): Take a widest_int instead of a double_int.
1111 (create_add_on_incoming_edge): Likewise.
1112 (create_phi_basis): Use wide-int interfaces.
1113 (replace_conditional_candidate): Likewise.
1114 (record_increment): Take a widest_int instead of a double_int.
1115 (record_phi_increments): Use wide-int interfaces.
1116 (phi_incr_cost): Take a widest_int instead of a double_int.
1117 (lowest_cost_path): Likewise.
1118 (total_savings): Likewise.
1119 (analyze_increments): Use wide-int interfaces.
1120 (ncd_with_phi): Take a widest_int instead of a double_int.
1121 (ncd_of_cand_and_phis): Likewise.
1122 (nearest_common_dominator_for_cands): Likewise.
1123 (insert_initializers): Use wide-int interfaces.
1124 (all_phi_incrs_profitable): Likewise.
1125 (replace_one_candidate): Likewise.
1126 (replace_profitable_candidates): Likewise.
1127 * godump.c: Include wide-int-print.h.
1128 (go_output_typedef): Use wide-int interfaces.
1129 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1130 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1131 (build_loop_iteration_domains): Likewise.
1132 * hooks.h: Include wide-int.h rather than double-int.h.
1133 (hook_bool_dint_dint_uint_bool_true): Delete.
1134 (hook_bool_wint_wint_uint_bool_true): Declare.
1135 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1136 (hook_bool_wint_wint_uint_bool_true): New.
1137 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1139 (ubsan_expand_si_overflow_mul_check): Likewise.
1140 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1141 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1142 (get_ancestor_addr_info): Likewise.
1143 (ipa_modify_call_arguments): Likewise.
1144 * loop-doloop.c (doloop_modify): Likewise.
1145 (doloop_optimize): Likewise.
1146 * loop-iv.c (iv_number_of_iterations): Likewise.
1147 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1148 (unroll_loop_constant_iterations): Likewise.
1149 (decide_unroll_runtime_iterations): Likewise.
1150 (unroll_loop_runtime_iterations): Likewise.
1151 (decide_peel_simple): Likewise.
1152 (decide_unroll_stupid): Likewise.
1153 * lto-streamer-in.c (streamer_read_wi): Add.
1154 (input_cfg): Use wide-int interfaces.
1155 (lto_input_tree_1): Likewise.
1156 * lto-streamer-out.c (streamer_write_wi): Add.
1157 (hash_tree): Use wide-int interfaces.
1158 (output_cfg): Likewise.
1159 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1160 (GTFILES): Add wide-int.h and signop.h.
1161 (TAGS): Look for .cc files too.
1162 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1163 * optabs.c (expand_subword_shift): Likewise.
1164 (expand_doubleword_shift): Likewise.
1165 (expand_absneg_bit): Likewise.
1166 (expand_copysign_absneg): Likewise.
1167 (expand_copysign_bit): Likewise.
1168 * postreload.c (reload_cse_simplify_set): Likewise.
1169 * predict.c (predict_iv_comparison): Likewise.
1170 * pretty-print.h: Include wide-int-print.h.
1172 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1173 * print-tree.c: Include wide-int-print.h.
1174 (print_node_brief): Use wide-int interfaces.
1175 (print_node): Likewise.
1176 * read-rtl.c (validate_const_wide_int): New.
1177 (read_rtx_code): Add CONST_WIDE_INT case.
1178 * real.c: Include wide-int.h.
1179 (real_to_integer2): Delete.
1180 (real_to_integer): New function, returning a wide_int.
1181 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1182 (ten_to_ptwo): Update call to real_from_integer.
1183 (real_digit): Likewise.
1184 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1185 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1186 (REAL_VALUE_TO_INT): Delete.
1187 (real_to_integer): Declare a wide-int form.
1188 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1189 * recog.c (const_int_operand): Improve comment.
1190 (const_scalar_int_operand): New.
1191 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1192 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1193 (split_double): Likewise.
1194 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1195 (rtx_size): Likewise.
1196 (rtx_alloc_stat_v): New.
1197 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1198 (cwi_output_hex): New.
1199 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1200 (cwi_check_failed_bounds): New.
1201 * rtl.def (CONST_WIDE_INT): New.
1202 * rtl.h: Include <utility> and wide-int.h.
1203 (struct hwivec_def): New.
1204 (CWI_GET_NUM_ELEM): New.
1205 (CWI_PUT_NUM_ELEM): New.
1206 (struct rtx_def): Add num_elem and hwiv.
1207 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1208 (CASE_CONST_UNIQUE): Likewise.
1209 (CASE_CONST_ANY): Likewise.
1210 (CONST_SCALAR_INT_P): Likewise.
1211 (CONST_WIDE_INT_P): New.
1213 (HWIVEC_CHECK): New.
1214 (cwi_check_failed_bounds): New.
1216 (HWIVEC_CHECK): New.
1217 (CONST_WIDE_INT_VEC) New.
1218 (CONST_WIDE_INT_NUNITS) New.
1219 (CONST_WIDE_INT_ELT) New.
1220 (rtx_mode_t): New type.
1221 (wi::int_traits <rtx_mode_t>): New.
1223 (wi::min_value): New.
1224 (wi::max_value): New.
1226 (const_wide_int_alloc): New.
1227 (immed_wide_int_const): New.
1228 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1229 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1230 * signop.h: New file.
1231 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1232 (simplify_const_unary_operation): Use wide-int interfaces.
1233 (simplify_binary_operation_1): Likewise.
1234 (simplify_const_binary_operation): Likewise.
1235 (simplify_const_relational_operation): Likewise.
1236 (simplify_immed_subreg): Likewise.
1237 * stmt.c (expand_case): Likewise.
1238 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1239 signop rather than a bool.
1240 * stor-layout.c (layout_type): Use wide-int interfaces.
1241 (initialize_sizetypes): Update calls to
1242 set_min_and_max_values_for_integral_type.
1243 (set_min_and_max_values_for_integral_type): Take a signop rather
1244 than a bool. Use wide-int interfaces.
1245 (fixup_signed_type): Update accordingly. Remove
1246 HOST_BITS_PER_DOUBLE_INT limit.
1247 (fixup_unsigned_type): Likewise.
1248 * system.h (STATIC_CONSTANT_P): New.
1249 (STATIC_ASSERT): New.
1250 * target.def (can_use_doloop_p): Take widest_ints rather than
1252 * target.h: Include wide-int.h rather than double-int.h.
1253 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1255 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1256 rather than INT_CST_LT_UNSIGNED.
1257 (can_use_doloop_if_innermost): Take widest_ints rather than
1259 * tree-affine.c: Include wide-int-print.h.
1260 (double_int_ext_for_comb): Delete.
1261 (wide_int_ext_for_comb): New.
1262 (aff_combination_zero): Use wide-int interfaces.
1263 (aff_combination_const): Take a widest_int instead of a double_int.
1264 (aff_combination_elt): Use wide-int interfaces.
1265 (aff_combination_scale): Take a widest_int instead of a double_int.
1266 (aff_combination_add_elt): Likewise.
1267 (aff_combination_add_cst): Likewise.
1268 (aff_combination_add): Use wide-int interfaces.
1269 (aff_combination_convert): Likewise.
1270 (tree_to_aff_combination): Likewise.
1271 (add_elt_to_tree): Take a widest_int instead of a double_int.
1272 (aff_combination_to_tree): Use wide-int interfaces.
1273 (aff_combination_remove_elt): Likewise.
1274 (aff_combination_add_product): Take a widest_int instead of
1276 (aff_combination_mult): Use wide-int interfaces.
1277 (aff_combination_expand): Likewise.
1278 (double_int_constant_multiple_p): Delete.
1279 (wide_int_constant_multiple_p): New.
1280 (aff_combination_constant_multiple_p): Take a widest_int pointer
1281 instead of a double_int pointer.
1282 (print_aff): Use wide-int interfaces.
1283 (get_inner_reference_aff): Take a widest_int pointer
1284 instead of a double_int pointer.
1285 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1286 * tree-affine.h: Include wide-int.h.
1287 (struct aff_comb_elt): Change type of coef to widest_int.
1288 (struct affine_tree_combination): Change type of offset to widest_int.
1289 (double_int_ext_for_comb): Delete.
1290 (wide_int_ext_for_comb): New.
1291 (aff_combination_const): Use widest_int instead of double_int.
1292 (aff_combination_scale): Likewise.
1293 (aff_combination_add_elt): Likewise.
1294 (aff_combination_constant_multiple_p): Likewise.
1295 (get_inner_reference_aff): Likewise.
1296 (aff_comb_cannot_overlap_p): Likewise.
1297 (aff_combination_zero_p): Use wide-int interfaces.
1298 * tree.c: Include tree.h.
1299 (init_ttree): Use make_int_cst.
1300 (tree_code_size): Removed code for INTEGER_CST case.
1301 (tree_size): Add INTEGER_CST case.
1302 (make_node_stat): Update comment.
1303 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1304 (build_int_cst_type): Use wide-int interfaces.
1305 (double_int_to_tree): Likewise.
1306 (double_int_fits_to_tree_p): Delete.
1307 (force_fit_type_double): Delete.
1308 (force_fit_type): New.
1309 (int_cst_hash_hash): Use wide-int interfaces.
1310 (int_cst_hash_eq): Likewise.
1311 (build_int_cst_wide): Delete.
1312 (wide_int_to_tree): New.
1313 (cache_integer_cst): Use wide-int interfaces.
1314 (build_low_bits_mask): Likewise.
1315 (cst_and_fits_in_hwi): Likewise.
1316 (real_value_from_int_cst): Likewise.
1317 (make_int_cst_stat): New.
1318 (integer_zerop): Use wide_int interfaces.
1319 (integer_onep): Likewise.
1320 (integer_all_onesp): Likewise.
1321 (integer_pow2p): Likewise.
1322 (integer_nonzerop): Likewise.
1323 (tree_log2): Likewise.
1324 (tree_floor_log2): Likewise.
1325 (tree_ctz): Likewise.
1326 (int_size_in_bytes): Likewise.
1327 (mem_ref_offset): Return an offset_int rather than a double_int.
1328 (build_type_attribute_qual_variant): Use wide_int interfaces.
1329 (type_hash_eq): Likewise
1330 (tree_int_cst_equal): Likewise.
1331 (tree_int_cst_lt): Delete.
1332 (tree_int_cst_compare): Likewise.
1333 (tree_fits_shwi_p): Use wide_int interfaces.
1334 (tree_fits_uhwi_p): Likewise.
1335 (tree_int_cst_sign_bit): Likewise.
1336 (tree_int_cst_sgn): Likewise.
1337 (tree_int_cst_min_precision): Take a signop rather than a bool.
1338 (simple_cst_equal): Use wide_int interfaces.
1339 (compare_tree_int): Likewise.
1340 (iterative_hash_expr): Likewise.
1341 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1343 (get_type_static_bounds): Use wide_int interfaces.
1344 (tree_int_cst_elt_check_failed): New.
1345 (build_common_tree_nodes): Reordered to set prec before filling in
1347 (int_cst_value): Check cst_and_fits_in_hwi.
1348 (widest_int_cst_value): Use wide_int interfaces.
1349 (upper_bound_in_type): Likewise.
1350 (lower_bound_in_type): Likewise.
1351 (num_ending_zeros): Likewise.
1352 (drop_tree_overflow): Likewise.
1353 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1354 (gen_conditions_for_pow_cst_base): Likewise.
1355 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1356 (group_case_labels_stmt): Use wide-int interfaces.
1357 (verify_gimple_assign_binary): Likewise.
1358 (print_loop): Likewise.
1359 * tree-chrec.c (tree_fold_binomial): Likewise.
1360 * tree-core.h (struct tree_base): Add int_length.
1361 (struct tree_int_cst): Change rep of value.
1362 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1363 (dr_may_alias_p): Likewise.
1364 (max_stmt_executions_tree): Likewise.
1365 * tree.def (INTEGER_CST): Update comment.
1366 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1367 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1368 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1369 (dequeue_and_dump): Use wide-int interfaces.
1370 * tree.h: Include wide-int.h.
1371 (NULL_TREE): Moved to earlier loc in file.
1372 (TREE_INT_CST_ELT_CHECK): New.
1373 (tree_int_cst_elt_check_failed): New.
1375 (TREE_INT_CST): Delete.
1376 (TREE_INT_CST_LOW): Use wide-int interfaces.
1377 (TREE_INT_CST_HIGH): Delete.
1378 (TREE_INT_CST_NUNITS): New.
1379 (TREE_INT_CST_EXT_NUNITS): Likewise.
1380 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1381 (TREE_INT_CST_ELT): Likewise.
1382 (INT_CST_LT): Delete.
1383 (tree_int_cst_elt_check): New (two forms).
1384 (type_code_size): Update comment.
1385 (make_int_cst_stat, make_int_cst): New.
1386 (tree_to_double_int): Delete.
1387 (double_int_fits_to_tree_p): Delete.
1388 (force_fit_type_double): Delete.
1389 (build_int_cstu): Replace with out-of-line function.
1390 (build_int_cst_wide): Delete.
1391 (tree_int_cst_lt): Define inline.
1392 (tree_int_cst_le): New.
1393 (tree_int_cst_compare): Define inline.
1394 (tree_int_cst_min_precision): Take a signop rather than a bool.
1395 (wi::int_traits <const_tree>): New.
1396 (wi::int_traits <tree>): New.
1397 (wi::extended_tree): New.
1398 (wi::int_traits <wi::extended_tree>): New.
1399 (wi::to_widest): New.
1400 (wi::to_offset): New.
1401 (wi::fits_to_tree_p): New.
1402 (wi::min_value): New.
1403 (wi::max_value): New.
1404 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1405 (copy_tree_body_r): Likewise.
1406 * tree-object-size.c (compute_object_offset): Likewise.
1407 (addr_object_size): Likewise.
1408 * tree-predcom.c: Include wide-int-print.h.
1409 (struct dref_d): Change type of offset to widest_int.
1410 (dump_dref): Call wide-int printer.
1411 (aff_combination_dr_offset): Use wide-int interfaces.
1412 (determine_offset): Take a widest_int pointer rather than a
1414 (split_data_refs_to_components): Use wide-int interfaces.
1415 (suitable_component_p): Likewise.
1416 (order_drefs): Likewise.
1417 (add_ref_to_chain): Likewise.
1418 (valid_initializer_p): Likewise.
1419 (determine_roots_comp): Likewise.
1420 * tree-pretty-print.c: Include wide-int-print.h.
1421 (dump_generic_node): Use wide-int interfaces.
1422 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1423 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1424 (move_fixed_address_to_symbol): Likewise.
1425 (move_hint_to_base): Likewise.
1426 (move_pointer_to_base): Likewise.
1427 (move_variant_to_index): Likewise.
1428 (most_expensive_mult_to_index): Likewise.
1429 (addr_to_parts): Likewise.
1430 (copy_ref_info): Likewise.
1431 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1432 (indirect_refs_may_alias_p): Likewise.
1433 (stmt_kills_ref_p_1): Likewise.
1434 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1435 * tree-ssa-ccp.c: Update comment at top of file. Include
1437 (struct prop_value_d): Change type of mask to widest_int.
1438 (extend_mask): New function.
1439 (dump_lattice_value): Use wide-int interfaces.
1440 (get_default_value): Likewise.
1441 (set_constant_value): Likewise.
1442 (set_value_varying): Likewise.
1443 (valid_lattice_transition): Likewise.
1444 (set_lattice_value): Likewise.
1445 (value_to_double_int): Delete.
1446 (value_to_wide_int): New.
1447 (get_value_from_alignment): Use wide-int interfaces.
1448 (get_value_for_expr): Likewise.
1449 (do_dbg_cnt): Likewise.
1450 (ccp_finalize): Likewise.
1451 (ccp_lattice_meet): Likewise.
1452 (bit_value_unop_1): Use widest_ints rather than double_ints.
1453 (bit_value_binop_1): Likewise.
1454 (bit_value_unop): Use wide-int interfaces.
1455 (bit_value_binop): Likewise.
1456 (bit_value_assume_aligned): Likewise.
1457 (evaluate_stmt): Likewise.
1458 (ccp_fold_stmt): Likewise.
1459 (visit_cond_stmt): Likewise.
1460 (ccp_visit_stmt): Likewise.
1461 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1462 (constant_pointer_difference): Likewise.
1463 (associate_pointerplus): Likewise.
1464 (combine_conversions): Likewise.
1465 * tree-ssa-loop.h: Include wide-int.h.
1466 (struct tree_niter_desc): Change type of max to widest_int.
1467 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1468 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1469 (remove_redundant_iv_tests): Likewise.
1470 (canonicalize_loop_induction_variables): Likewise.
1471 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1472 (constant_multiple_of): Take a widest_int pointer instead of
1473 a double_int pointer.
1474 (get_computation_aff): Use wide-int interfaces.
1475 (ptr_difference_cost): Likewise.
1476 (difference_cost): Likewise.
1477 (get_loop_invariant_expr_id): Likewise.
1478 (get_computation_cost_at): Likewise.
1479 (iv_elimination_compare_lt): Likewise.
1480 (may_eliminate_iv): Likewise.
1481 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1482 instead of double_int.
1483 (max_loop_iterations): Likewise.
1484 (max_stmt_executions): Likewise.
1485 (estimated_stmt_executions): Likewise.
1486 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1487 (split_to_var_and_offset): Use wide-int interfaces.
1488 (determine_value_range): Likewise.
1489 (bound_difference_of_offsetted_base): Likewise.
1490 (bounds_add): Take a widest_int instead of a double_int.
1491 (number_of_iterations_ne_max): Use wide-int interfaces.
1492 (number_of_iterations_ne): Likewise.
1493 (number_of_iterations_lt_to_ne): Likewise.
1494 (assert_loop_rolls_lt): Likewise.
1495 (number_of_iterations_lt): Likewise.
1496 (number_of_iterations_le): Likewise.
1497 (number_of_iterations_cond): Likewise.
1498 (number_of_iterations_exit): Likewise.
1499 (finite_loop_p): Likewise.
1500 (derive_constant_upper_bound_assign): Likewise.
1501 (derive_constant_upper_bound): Return a widest_int.
1502 (derive_constant_upper_bound_ops): Likewise.
1503 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1504 (record_estimate): Take a widest_int rather than a double_int.
1505 (record_nonwrapping_iv): Use wide-int interfaces.
1506 (double_int_cmp): Delete.
1507 (wide_int_cmp): New.
1508 (bound_index): Take a widest_int rather than a double_int.
1509 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1510 (maybe_lower_iteration_bound): Likewise.
1511 (estimate_numbers_of_iterations_loop): Likewise.
1512 (estimated_loop_iterations): Take a widest_int pointer than than
1513 a double_int pointer.
1514 (estimated_loop_iterations_int): Use wide-int interfaces.
1515 (max_loop_iterations): Take a widest_int pointer than than
1516 a double_int pointer.
1517 (max_loop_iterations_int): Use wide-int interfaces.
1518 (max_stmt_executions): Take a widest_int pointer than than
1519 a double_int pointer.
1520 (estimated_stmt_executions): Likewise.
1521 (n_of_executions_at_most): Use wide-int interfaces.
1522 (scev_probably_wraps_p): Likewise.
1523 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1525 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1527 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1528 double_ints. Adjust for trailing_wide_ints <3> representation.
1529 (set_nonzero_bits): Likewise.
1530 (get_range_info): Return wide_ints rather than double_ints.
1531 Adjust for trailing_wide_ints <3> representation.
1532 (get_nonzero_bits): Likewise.
1533 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1535 * tree-ssanames.h (struct range_info_def): Replace min, max and
1536 nonzero_bits with a trailing_wide_ints <3>.
1537 (set_range_info): Use wide_int_refs rather than double_ints.
1538 (set_nonzero_bits): Likewise.
1539 (get_range_info): Return wide_ints rather than double_ints.
1540 (get_nonzero_bits): Likewise.
1541 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1542 * tree-ssa-pre.c (phi_translate_1): Likewise.
1543 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1544 (acceptable_pow_call): Likewise.
1545 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1547 (vn_reference_fold_indirect): Likewise.
1548 (vn_reference_maybe_forwprop_address): Likewise.
1549 (valueize_refs_1): Likewise.
1550 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1551 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1552 tree_int_cst_lt and tree_int_cst_le.
1553 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1555 (streamer_alloc_tree): Likewise.
1556 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1557 (streamer_write_tree_header): Likewise.
1558 (streamer_write_integer_cst): Likewise.
1559 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1560 (build_constructors): Likewise.
1561 (array_value_type): Likewise.
1562 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1563 (vect_check_gather): Likewise.
1564 * tree-vect-generic.c (build_replicated_const): Likewise.
1565 (expand_vector_divmod): Likewise.
1566 * tree-vect-loop.c (vect_transform_loop): Likewise.
1567 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1568 (vect_do_peeling_for_alignment): Likewise.
1569 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1570 * tree-vrp.c: Include wide-int.h.
1571 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1572 (extract_range_from_assert): Use wide-int interfaces.
1573 (vrp_int_const_binop): Likewise.
1574 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1575 double_int pointers.
1576 (ranges_from_anti_range): Use wide-int interfaces.
1577 (quad_int_cmp): Delete.
1578 (quad_int_pair_sort): Likewise.
1579 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1580 (extract_range_from_unary_expr_1): Likewise.
1581 (adjust_range_with_scev): Likewise.
1582 (masked_increment): Take and return wide_ints rather than double_ints.
1583 (register_edge_assert_for_2): Use wide-int interfaces.
1584 (check_array_ref): Likewise.
1585 (search_for_addr_array): Likewise.
1586 (maybe_set_nonzero_bits): Likewise.
1587 (union_ranges): Pass an integer of the correct type instead of
1588 using integer_one_node.
1589 (intersect_ranges): Likewise.
1590 (simplify_truth_ops_using_ranges): Likewise.
1591 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1592 (range_fits_type_p): Likewise.
1593 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1595 (simplify_conversion_using_ranges): Use wide-int interfaces.
1596 (simplify_float_conversion_using_ranges): Likewise.
1597 (vrp_finalize): Likewise.
1598 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1599 (gimple_stringops_transform): Likewise.
1600 * varasm.c (decode_addr_const): Likewise.
1601 (const_hash_1): Likewise.
1602 (const_rtx_hash_1): Likewise
1603 (output_constant): Likewise.
1604 (array_size_for_constructor): Likewise.
1605 (output_constructor_regular_field): Likewise.
1606 (output_constructor_bitfield): Likewise.
1607 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1608 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1610 * gencheck.c: Define BITS_PER_UNIT.
1613 * wide-int-print.cc: New.
1614 * wide-int-print.h: New.
1616 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1618 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1620 2014-05-06 Richard Biener <rguenther@suse.de>
1622 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1623 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1624 (TODO_verify_all): Adjust.
1625 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1626 TODO_verify_stmts and TODO_verify_rtl_sharing.
1627 * bb-reorder.c: Likewise.
1628 * cfgexpand.c: Likewise.
1629 * cprop.c: Likewise.
1631 * function.c: Likewise.
1632 * fwprop.c: Likewise.
1634 * gimple-ssa-isolate-paths.c: Likewise.
1635 * gimple-ssa-strength-reduction.c: Likewise.
1636 * ipa-split.c: Likewise.
1637 * loop-init.c: Likewise.
1638 * loop-unroll.c: Likewise.
1639 * lower-subreg.c: Likewise.
1640 * modulo-sched.c: Likewise.
1641 * postreload-gcse.c: Likewise.
1642 * predict.c: Likewise.
1643 * recog.c: Likewise.
1644 * sched-rgn.c: Likewise.
1645 * store-motion.c: Likewise.
1646 * tracer.c: Likewise.
1647 * trans-mem.c: Likewise.
1648 * tree-call-cdce.c: Likewise.
1649 * tree-cfg.c: Likewise.
1650 * tree-cfgcleanup.c: Likewise.
1651 * tree-complex.c: Likewise.
1652 * tree-eh.c: Likewise.
1653 * tree-emutls.c: Likewise.
1654 * tree-if-conv.c: Likewise.
1655 * tree-into-ssa.c: Likewise.
1656 * tree-loop-distribution.c: Likewise.
1657 * tree-object-size.c: Likewise.
1658 * tree-parloops.c: Likewise.
1659 * tree-pass.h: Likewise.
1660 * tree-sra.c: Likewise.
1661 * tree-ssa-ccp.c: Likewise.
1662 * tree-ssa-copy.c: Likewise.
1663 * tree-ssa-copyrename.c: Likewise.
1664 * tree-ssa-dce.c: Likewise.
1665 * tree-ssa-dom.c: Likewise.
1666 * tree-ssa-dse.c: Likewise.
1667 * tree-ssa-forwprop.c: Likewise.
1668 * tree-ssa-ifcombine.c: Likewise.
1669 * tree-ssa-loop-ch.c: Likewise.
1670 * tree-ssa-loop-ivcanon.c: Likewise.
1671 * tree-ssa-loop.c: Likewise.
1672 * tree-ssa-math-opts.c: Likewise.
1673 * tree-ssa-phiopt.c: Likewise.
1674 * tree-ssa-phiprop.c: Likewise.
1675 * tree-ssa-pre.c: Likewise.
1676 * tree-ssa-reassoc.c: Likewise.
1677 * tree-ssa-sink.c: Likewise.
1678 * tree-ssa-strlen.c: Likewise.
1679 * tree-ssa-tail-merge.c: Likewise.
1680 * tree-ssa-uncprop.c: Likewise.
1681 * tree-switch-conversion.c: Likewise.
1682 * tree-tailcall.c: Likewise.
1683 * tree-vect-generic.c: Likewise.
1684 * tree-vectorizer.c: Likewise.
1685 * tree-vrp.c: Likewise.
1687 * var-tracking.c: Likewise.
1688 * bt-load.c: Likewise.
1689 * cfgcleanup.c: Likewise.
1690 * combine-stack-adj.c: Likewise.
1691 * combine.c: Likewise.
1692 * compare-elim.c: Likewise.
1693 * config/epiphany/resolve-sw-modes.c: Likewise.
1694 * config/i386/i386.c: Likewise.
1695 * config/mips/mips.c: Likewise.
1696 * config/s390/s390.c: Likewise.
1697 * config/sh/sh_treg_combine.cc: Likewise.
1698 * config/sparc/sparc.c: Likewise.
1701 * final.c: Likewise.
1702 * ifcvt.c: Likewise.
1703 * mode-switching.c: Likewise.
1704 * passes.c: Likewise.
1705 * postreload.c: Likewise.
1707 * reg-stack.c: Likewise.
1708 * regcprop.c: Likewise.
1709 * regrename.c: Likewise.
1712 2014-05-06 Richard Biener <rguenther@suse.de>
1715 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1716 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1718 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1721 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1723 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1724 Tom de Vries <tom@codesourcery.com>
1726 * target.def (call_fusage_contains_non_callee_clobbers): New
1728 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1730 (@node Miscellaneous Register Hooks): New node.
1731 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1732 * doc/tm.texi: Regenerate.
1734 2014-05-05 Marek Polacek <polacek@redhat.com>
1737 * opts.c (common_handle_option): Call error_at instead of warning_at.
1739 2014-05-05 Richard Biener <rguenther@suse.de>
1741 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1742 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1743 under the TODO_verify_il umbrella.
1745 2014-05-05 Richard Biener <rguenther@suse.de>
1747 * passes.c (execute_function_todo): Move TODO_verify_flow under
1748 the TODO_verify_ul umbrella.
1750 2014-05-05 Richard Biener <rguenther@suse.de>
1753 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1754 X & CST away from a CST that is the mask of a mode.
1756 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1758 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1759 int argument to enum machine_mode.
1760 (picochip_class_max_nregs): Ditto.
1761 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1762 (picochip_class_max_nregs): Ditto.
1764 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1766 * target.def: Add new target hook.
1767 * doc/tm.texi: Regenerate.
1768 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1769 * targhooks.c (default_keep_leaf_when_profiled): New function.
1771 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1772 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1774 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1776 PR tree-optimization/60363
1777 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1778 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1779 (update_destination_phis): New parameter.
1780 (create_edge_and_update_destination_phis): Ditto.
1781 (ssa_fix_duplicate_block_edges): Pass new arguments.
1782 (thread_single_edge): Ditto.
1784 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1786 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1787 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1788 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1789 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1790 Use RS6000_BTM_HARD_FLOAT.
1791 (BU_MISC_2): Likewise.
1792 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1793 RS6000_BTM_HARD_FLOAT.
1794 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1796 (rs6000_invalid_builtin): Add hard floating builtin support.
1797 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1798 hard float builtins.
1799 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1801 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1803 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1804 Add missing function* argument.
1806 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1808 * lra-constraints.c (valid_address_p): Move earlier in file.
1809 Add a constraint argument to the address_info version.
1810 (satisfies_memory_constraint_p): New function.
1811 (satisfies_address_constraint_p): Likewise.
1812 (process_alt_operands, curr_insn_transform): Use them.
1813 (process_address): Pass the constraint to valid_address_p when
1814 checking address operands.
1816 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1818 * config/mips/mips.c (mips_isa_rev): New variable.
1819 (mips_set_architecture): Set it.
1820 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1822 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1823 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1824 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1825 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1826 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1827 conditions in terms of mips_isa_rev.
1828 (mips_isa_rev): Declare.
1830 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1832 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1833 (prob_unlikely, prob_likely): Make variables const.
1835 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1837 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1839 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1841 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1843 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1845 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1846 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1847 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1849 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1850 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1852 Replace usage of ROUND_REG with sh_round_reg.
1853 Use CEIL instead of ROUND_ADVANCE.
1855 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1858 * config/sh/sh.c: Include stdlib headers before everything else.
1860 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1862 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1863 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1864 (gimplify_adjust_omp_clauses): Simd region is never
1865 directly nested in combined parallel. Instead, for linear
1866 with copyin/copyout, if in combined for simd loop, make decl
1867 firstprivate/lastprivate on OMP_FOR.
1868 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1869 expand_omp_for_static_chunk): When setting endvar, also set
1870 fd->loop.v to the same value.
1872 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1874 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1876 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1878 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1881 2014-05-02 Marek Polacek <polacek@redhat.com>
1883 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1885 2014-05-02 Kito Cheng <kito@0xlab.org>
1887 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1888 to a C expression marco.
1889 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1890 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1891 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1892 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1893 HONOR_REG_ALLOC_ORDER.
1894 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1896 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1898 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1900 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1902 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1904 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1906 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1907 (convert_scalar_cond_reduction): Likewise.
1908 (predicate_scalar_phi): Add recognition and transformation
1909 of simple conditioanl reduction to be vectorizable.
1911 2014-05-01 Marek Polacek <polacek@redhat.com>
1914 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1916 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1918 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1919 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1920 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1921 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1922 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1923 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1924 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1925 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1927 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1929 * config/arc/arc.opt (mlra): Move comment above option name
1930 to avoid mis-parsing as language options.
1932 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1934 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1935 * config/sol2.h: ... here.
1936 * config/sol2-10.h: Remove.
1938 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1939 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1940 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1941 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1942 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1943 * config/sol2.h: ... here.
1944 (SECTION_NAME_FORMAT): Don't redefine.
1945 (STARTFILE_ARCH32_SPEC): Rename to ...
1946 (STARTFILE_ARCH_SPEC): ... this.
1947 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1948 * config/sparc/sol2.h: ... here.
1949 (SECTION_NAME_FORMAT): Don't undef.
1950 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1951 (SUBTARGET_EXTRA_SPECS): Remove.
1952 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1954 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1955 (MD_STARTFILE_PREFIX): Remove.
1956 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1957 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1958 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1959 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1960 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1961 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1962 * config/i386/sol2.h: ... here.
1963 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1964 * config/i386/sol2-bi.h: Remove.
1965 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1966 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1968 * config/i386/t-sol2-64: Rename to ...
1969 * config/i386/t-sol2: ... this.
1970 * config/sparc/t-sol2-64: Rename to ...
1971 * config/sparc/t-sol2: ... this.
1973 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1974 sol2_tm_file_head, sol2_tm_file_tail.
1975 Include ${cpu_type}/sol2.h before sol2.h.
1977 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1978 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1979 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1980 Reflect i386/t-sol2-64 renaming.
1981 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1982 Reflect sparc/t-sol2-64 renaming.
1984 2014-04-30 Richard Biener <rguenther@suse.de>
1986 * passes.c (execute_function_todo): Move TODO_verify_stmts
1987 and TODO_verify_ssa under the TODO_verify_il umbrella.
1988 * tree-ssa.h (verify_ssa): Adjust prototype.
1989 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1990 we should verify SSA operands.
1991 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1992 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1993 whether we should verify whether not throwing stmts have EH info.
1994 * graphite-scop-detection.c (create_sese_edges): Adjust.
1995 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1996 * tree-eh.c (lower_try_finally_switch): Do not add the
1997 default case label twice.
1999 2014-04-30 Marek Polacek <polacek@redhat.com>
2001 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2002 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2003 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2004 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2006 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2008 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2009 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2010 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2011 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2012 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2013 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2014 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2015 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2017 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2019 * tree-cfg.c (dump_function_to_file): Dump the return type of
2020 functions, in a line to itself before the function body, mimicking
2021 the layout of a C function.
2023 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2025 PR tree-optimization/60971
2026 * tree-tailcall.c (process_assignment): Reject conversions which
2029 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2031 * calls.c (initialize_argument_information): Always treat
2032 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2033 (expand_call): Likewise.
2034 (emit_library_call_calue_1): Likewise.
2035 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2036 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2039 2014-04-29 Nick Clifton <nickc@redhat.com>
2041 * config/msp430/msp430.md (umulsidi): Fix typo.
2042 (mulhisi3): Enable even inside interrupt handlers.
2043 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2044 bigger return address pushed in large mode.
2046 2014-04-29 Nick Clifton <nickc@redhat.com>
2048 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2049 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2051 * config/m32r/m32r.c (init_reg_tables): Likewise.
2052 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2053 enum to hold the modes.
2055 2014-04-29 Richard Biener <rguenther@suse.de>
2057 * dominance.c (free_dominance_info): Add overload with
2059 (dom_info_state): Likewise.
2060 (dom_info_available_p): Likewise.
2061 * basic-block.h (free_dominance_info, dom_info_state,
2062 dom_info_available_p): Declare overloads.
2063 * passes.c (execute_function_todo): Verify that verifiers
2064 don't change dominator info state. Drop dominator info
2065 for IPA pass invocations.
2066 * cgraph.c (release_function_body): Restore asserts that
2067 dominator information is released.
2069 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2071 * doc/invoke.texi: Fix typo.
2072 * tree-vrp.c: Fix typos.
2073 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2075 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2077 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2079 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2081 * config/aarch64/aarch64-builtins.c
2082 (aarch64_types_storestruct_lane_qualifiers): New.
2083 (TYPES_STORESTRUCT_LANE): Likewise.
2084 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2085 (st3_lane): Likewise.
2086 (st4_lane): Likewise.
2087 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2088 (vec_store_lanesci_lane<mode>): Likewise.
2089 (vec_store_lanesxi_lane<mode>): Likewise.
2090 (aarch64_st2_lane<VQ:mode>): Likewise.
2091 (aarch64_st3_lane<VQ:mode>): Likewise.
2092 (aarch64_st4_lane<VQ:mode>): Likewise.
2093 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2094 * config/aarch64/arm_neon.h
2095 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2096 use new macro arguments.
2097 (__ST3_LANE_FUNC): Likewise.
2098 (__ST4_LANE_FUNC): Likewise.
2099 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2100 (V_THREE_ELEM): Likewise.
2101 (V_FOUR_ELEM): Likewise.
2103 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2105 * doc/gimple.texi: Replace the description of the now-defunct
2106 union gimple_statement_d with a diagram showing the
2107 gimple_statement_base class hierarchy and its relationships to
2108 the GSS_ and GIMPLE_ enums.
2110 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2112 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2113 * config/aarch64/aarch64.c
2114 (aarch64_cannot_change_mode_class): Weaken conditions.
2115 (aarch64_modes_tieable_p): New.
2116 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2118 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2120 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2121 (loadsync_<mode>): Change mode.
2122 (load_quadpti, store_quadpti): New.
2123 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2124 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2126 2014-04-28 Martin Jambor <mjambor@suse.cz>
2128 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2129 same alias type as the original statement.
2130 (subreplacement_assignment_data): New type.
2131 (handle_unscalarized_data_in_subtree): New type of parameter,
2132 generate new memory accesses with same alias type as the original
2134 (load_assign_lhs_subreplacements): Likewise.
2135 (sra_modify_constructor_assign): Generate new memory accesses with
2136 same alias type as the original statement.
2138 2014-04-28 Richard Biener <rguenther@suse.de>
2140 * tree-pass.h (TODO_verify_il): Define.
2141 (TODO_verify_all): Complete properly.
2142 * passes.c (execute_function_todo): Move existing loop-closed
2143 SSA verification under TODO_verify_il.
2144 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2145 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2146 Fix tree sharing issue.
2148 2014-04-28 Richard Biener <rguenther@suse.de>
2151 * builtins.def (DEF_C11_BUILTIN): Add.
2152 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2153 * coretypes.h (enum function_class): Add function_c11_misc.
2154 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2155 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2156 (call_may_clobber_ref_p_1): Likewise.
2157 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2158 (mark_all_reaching_defs_necessary_1): Likewise.
2159 (propagate_necessity): Likewise.
2160 (eliminate_unnecessary_stmts): Likewise.
2161 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2163 2014-04-28 Richard Biener <rguenther@suse.de>
2165 * tree-vrp.c (vrp_var_may_overflow): Remove.
2166 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2167 with overflow immediately bump to one before that value and
2168 let iteration figure out overflow status.
2170 2014-04-28 Richard Biener <rguenther@suse.de>
2172 * configure.ac: Do valgrind header checks unconditionally.
2173 Add --enable-valgrind-annotations.
2174 * system.h: Guard valgrind header inclusion with
2175 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2176 * alloc-pool.c (pool_alloc, pool_free): Use
2177 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2178 to guard possibly dead code.
2179 * config.in: Regenerated.
2180 * configure: Likewise.
2182 2014-04-28 Jeff Law <law@redhat.com>
2184 PR tree-optimization/60902
2185 * tree-ssa-threadedge.c
2186 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2187 over real defs when invalidating outputs from statements that do not
2188 produce useful outputs for threading.
2190 2014-04-28 Richard Biener <rguenther@suse.de>
2192 PR tree-optimization/60979
2193 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2194 SCOPs that end in a block with a successor with abnormal
2197 2014-04-28 Richard Biener <rguenther@suse.de>
2199 * tree-pass.h (execute_pass_list): Adjust prototype.
2200 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2201 (do_per_function): Change callback signature, push all actual
2202 work to the callbals.
2203 (do_per_function_toporder): Likewise.
2204 (execute_function_dump): Adjust.
2205 (execute_function_todo): Likewise.
2206 (clear_last_verified): Likewise.
2207 (verify_curr_properties): Likewise.
2208 (update_properties_after_pass): Likewise.
2209 (execute_pass_list_1): Split out from ...
2210 (execute_pass_list): ... here. Adjust.
2211 (execute_ipa_pass_list): Likewise.
2212 * cgraphunit.c (cgraph_add_new_function): Adjust.
2213 (analyze_function): Likewise.
2214 (expand_function): Likewise.
2215 * cgraph.c (release_function_body): Free dominance info
2216 here instead of asserting it was magically freed elsewhere.
2218 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2220 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2221 * configure: Regenerate.
2222 * config/sparc/sparc.opt (muser-mode): New option.
2223 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2225 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2226 * doc/invoke.texi (SPARC options): Document -muser-mode.
2228 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2230 * cselib.c (find_slot_memmode): Delete.
2231 (cselib_hasher): Change compare_type to a struct.
2232 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2234 (preserve_constants_and_equivs): Adjust for new compare_type.
2235 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2236 (wrap_constant): Delete.
2237 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2239 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2241 * doc/install.texi (Building with profile feedback): Remove
2244 2014-04-26 Tom de Vries <tom@codesourcery.com>
2246 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2249 2014-04-25 Cary Coutant <ccoutant@google.com>
2252 * dwarf2out.c (should_move_die_to_comdat): A type definition
2253 can contain a subprogram definition, but don't move it to a
2255 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2256 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2258 (clone_tree_hash): Rename to...
2259 (clone_tree_partial): ...this; change callers. Copy
2260 DW_TAG_subprogram DIEs as declarations.
2261 (copy_decls_walk): Don't copy children of a declaration into a
2264 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2267 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2270 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2272 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2273 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2275 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2278 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2280 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2282 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2284 PR tree-optimization/60930
2285 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2286 creating a multiply candidate by folding two constant
2287 multiplicands when the result overflows.
2289 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2291 PR tree-optimization/60960
2292 * tree-vect-generic.c (expand_vector_operation): Only call
2293 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2295 2014-04-25 Tom de Vries <tom@codesourcery.com>
2297 * expr.c (clobber_reg_mode): New function.
2298 * expr.h (clobber_reg): New function.
2300 2014-04-25 Tom de Vries <tom@codesourcery.com>
2302 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2305 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2306 Tom de Vries <tom@codesourcery.com>
2308 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2310 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2311 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2312 new argument to find_all_hard_reg_sets call.
2314 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2316 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2317 Use HOST_WIDE_INT_C for mask literal.
2318 (aarch_rev16_shleft_mask_imm_p): Likewise.
2320 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2323 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2325 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2327 PR preprocessor/56540
2328 * config/i386/i386-c.c (ix86_target_macros): Define
2329 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2331 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2333 * configure.ac (tga_func): Remove.
2334 (LIB_TLS_SPEC): Remove.
2335 * configure: Regenerate.
2336 * config.in: Regenerate.
2337 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2339 2014-04-25 Richard Biener <rguenther@suse.de>
2342 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2343 call stmt use/clobber sets during stmt walk instead of
2344 walking the possibly incomplete set of caller edges.
2346 2014-04-25 Richard Biener <rguenther@suse.de>
2349 * passes.c (apply_ipa_transforms): Inline into only caller ...
2350 (execute_one_pass): ... here. Properly bring in function
2351 bodies for nodes we want to apply IPA transforms to.
2353 2014-04-24 Cong Hou <congh@google.com>
2355 PR tree-optimization/60896
2356 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2357 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2358 (vect_mark_pattern_stmts): Set the def type of all statements in
2359 PATTERN_DEF_SEQ as vect_internal_def.
2361 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2363 * doc/extend.texi (PowerPC Built-in Functions): Document new
2364 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2365 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2367 * config/rs6000/predicates.md (const_0_to_3_operand): New
2368 predicate to match 0..3 integer constants.
2370 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2371 to support adding miscellaneous builtin functions.
2372 (BU_DFP_MISC_2): Likewise.
2373 (BU_P7_MISC_1): Likewise.
2374 (BU_P7_MISC_2): Likewise.
2375 (BU_P8V_MISC_3): Likewise.
2376 (BU_MISC_1): Likewise.
2377 (BU_MISC_2): Likewise.
2378 (DIVWE): Add extended divide builtin functions.
2381 (DIVWEUO): Likewise.
2385 (DIVDEUO): Likewise.
2386 (DXEX): Add decimal floating-point builtin functions.
2389 (DDEDPDQ): Likewise.
2391 (DENBCDQ): Likewise.
2398 (CDTBCD): Add new BCD builtin functions.
2402 (BCDADD_LT): Likewise.
2403 (BCDADD_EQ): Likewise.
2404 (BCDADD_GT): Likewise.
2405 (BCDADD_OV): Likewise.
2407 (BCDSUB_LT): Likewise.
2408 (BCDSUB_EQ): Likewise.
2409 (BCDSUB_GT): Likewise.
2410 (BCDSUB_OV): Likewise.
2411 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2412 (UNPACK_TD): Likewise.
2413 (PACK_TF): Likewise.
2414 (UNPACK_TF): Likewise.
2415 (UNPACK_TF_0): Likewise.
2416 (UNPACK_TF_1): Likewise.
2417 (PACK_V1TI): Likewise.
2418 (UNPACK_V1TI): Likewise.
2420 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2421 support for decimal floating point builtin functions.
2422 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2423 functions that take constant arguments.
2424 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2425 (rs6000_init_builtins): Setup long double, _Decimal64, and
2426 _Decimal128 types for new builtin functions.
2427 (builtin_function_type): Set the unsigned flags appropriately for
2428 the new builtin functions.
2429 (rs6000_opt_masks): Add support for decimal floating point builtin
2432 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2433 floating point builtin functions.
2434 (RS6000_BTM_COMMON): Likewise.
2435 (RS6000_BTI_long_double): Likewise.
2436 (RS6000_BTI_dfloat64): Likewise.
2437 (RS6000_BTI_dfloat128): Likewise.
2438 (long_double_type_internal_node): Likewise.
2439 (dfloat64_type_internal_node): Likewise.
2440 (dfloat128_type_internal_node): Likewise.
2442 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2443 2.07 bcd arithmetic instructions.
2444 (UNSPEC_BCDSUB): Likewise.
2445 (UNSPEC_BCD_OVERFLOW): Likewise.
2446 (UNSPEC_BCD_ADD_SUB): Likewise.
2447 (bcd_add_sub): Likewise.
2448 (BCD_TEST): Likewise.
2449 (bcd<bcd_add_sub>): Likewise.
2450 (bcd<bcd_add_sub>_test): Likewise.
2451 (bcd<bcd_add_sub>_test2): Likewise.
2452 (bcd<bcd_add_sub>_<code>): Likewise.
2453 (peephole2 for combined bcd ops): Likewise.
2455 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2456 decimal floating point builtin functions.
2457 (UNSPEC_DENBCD): Likewise.
2458 (UNSPEC_DXEX): Likewise.
2459 (UNSPEC_DIEX): Likewise.
2460 (UNSPEC_DSCLI): Likewise.
2461 (UNSPEC_DSCRI): Likewise.
2462 (D64_D128): Likewise.
2463 (dfp_suffix): Likewise.
2464 (dfp_ddedpd_<mode>): Likewise.
2465 (dfp_denbcd_<mode>): Likewise.
2466 (dfp_dxex_<mode>): Likewise.
2467 (dfp_diex_<mode>): Likewise.
2468 (dfp_dscli_<mode>): Likewise.
2469 (dfp_dscri_<mode>): Likewise.
2471 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2473 (UNSPEC_CDTBCD): Likewise.
2474 (UNSPEC_CBCDTD): Likewise.
2475 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2476 (UNSPEC_DIVEO): Likewise.
2477 (UNSPEC_DIVEU): Likewise.
2478 (UNSPEC_DIVEUO): Likewise.
2479 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2480 pack/unpack 128-bit types.
2481 (UNSPEC_PACK_128BIT): Likewise.
2482 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2483 (udiv<mode>3): Use idiv_ldiv mode attribute.
2484 (div<mode>3): Likewise.
2485 (addg6s): Add new BCD builtin functions.
2488 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2489 (div_extend): Likewise.
2490 (div<div_extend>_<mode>"): Likewise.
2491 (FP128_64): Add support for new builtin functions to pack/unpack
2493 (unpack<mode>): Likewise.
2494 (unpacktf_0): Likewise.
2495 (unpacktf_1): Likewise.
2496 (unpack<mode>_dm): Likewise.
2497 (unpack<mode>_nodm): Likewise.
2498 (pack<mode>): Likewise.
2499 (unpackv1ti): Likewise.
2500 (packv1ti): Likewise.
2502 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2504 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2507 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2509 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2510 * gimplify.c (omp_is_private): Change last argument's type to int.
2511 Only diagnose lastprivate if the simd argument is 1, only diagnose
2512 linear if the simd argument is 2.
2513 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2514 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2515 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2516 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2517 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2518 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2519 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2520 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2521 * tree-nested.c (convert_nonlocal_omp_clauses,
2522 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2524 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2527 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2530 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2532 * flag-types.h (enum ivar_visibility): Add.
2534 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2536 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2537 function * argument.
2539 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2541 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2543 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2544 Tom de Vries <tom@codesourcery.com>
2546 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2547 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2549 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2550 * emit-rtl.c (try_split): Same.
2552 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2553 Tom de Vries <tom@codesourcery.com>
2555 * common.opt (fuse-caller-save): New option.
2557 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2559 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2560 elements for big-endian.
2562 2014-04-24 Richard Biener <rguenther@suse.de>
2564 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2565 during TER and instead use the sepops interface for expanding
2566 non-GIMPLE_SINGLE_RHS.
2568 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2570 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2571 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2573 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2575 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2576 assembler 64-bit option.
2577 * configure: Regenerate.
2579 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2581 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2582 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2583 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2584 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2585 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2587 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2589 * config/aarch64/aarch64-builtins.c
2590 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2591 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2592 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2593 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2595 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2596 (Vrevsuff): New mode attribute.
2598 2014-04-24 Terry Guo <terry.guo@arm.com>
2600 * config/arm/arm.h (machine_function): Define variable
2601 after_arm_reorg here.
2602 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2603 (arm_split_constant): Update the way to access variable
2606 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2608 2014-04-23 Tom de Vries <tom@codesourcery.com>
2610 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2612 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2614 * is-a.h: Update comments to reflect the following changes to the
2615 "pointerness" of the API, making the template parameter match the
2616 return type, allowing use of is-a.h with typedefs of pointers.
2617 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2618 that the return type matches the parameter to the is_a_helper.
2620 (dyn_cast): Likewise.
2622 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2623 pointer from the is-a.h API.
2625 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2626 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2628 (is_a_helper <varpool_node>::test): Likewise, convert to...
2629 (is_a_helper <varpool_node *>::test): ...this.
2631 (varpool_first_variable): Update for removal of implicit pointer
2632 from the is-a.h API.
2633 (varpool_next_variable): Likewise.
2634 (varpool_first_static_initializer): Likewise.
2635 (varpool_next_static_initializer): Likewise.
2636 (varpool_first_defined_variable): Likewise.
2637 (varpool_next_defined_variable): Likewise.
2638 (cgraph_first_defined_function): Likewise.
2639 (cgraph_next_defined_function): Likewise.
2640 (cgraph_first_function): Likewise.
2641 (cgraph_next_function): Likewise.
2642 (cgraph_first_function_with_gimple_body): Likewise.
2643 (cgraph_next_function_with_gimple_body): Likewise.
2644 (cgraph_alias_target): Likewise.
2645 (varpool_alias_target): Likewise.
2646 (cgraph_function_or_thunk_node): Likewise.
2647 (varpool_variable_node): Likewise.
2648 (symtab_real_symbol_p): Likewise.
2649 * cgraphunit.c (referred_to_p): Likewise.
2650 (analyze_functions): Likewise.
2651 (handle_alias_pairs): Likewise.
2652 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2653 * gimple-ssa.h (gimple_vuse_op): Likewise.
2654 (gimple_vdef_op): Likewise.
2655 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2656 * gimple.c (gimple_build_asm_1): Likewise.
2657 (gimple_build_try): Likewise.
2658 (gimple_build_resx): Likewise.
2659 (gimple_build_eh_dispatch): Likewise.
2660 (gimple_build_omp_for): Likewise.
2661 (gimple_omp_for_set_clauses): Likewise.
2663 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2664 (is_a_helper <gimple_statement_asm *>::test): ...this.
2665 (is_a_helper <gimple_statement_bind>::test): Convert to...
2666 (is_a_helper <gimple_statement_bind *>::test): ...this.
2667 (is_a_helper <gimple_statement_call>::test): Convert to...
2668 (is_a_helper <gimple_statement_call *>::test): ...this.
2669 (is_a_helper <gimple_statement_catch>::test): Convert to...
2670 (is_a_helper <gimple_statement_catch *>::test): ...this.
2671 (is_a_helper <gimple_statement_resx>::test): Convert to...
2672 (is_a_helper <gimple_statement_resx *>::test): ...this.
2673 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2674 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2675 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2676 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2677 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2678 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2679 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2680 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2681 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2682 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2683 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2684 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2685 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2686 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2687 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2688 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2689 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2690 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2691 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2692 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2693 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2694 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2695 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2696 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2697 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2698 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2699 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2700 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2701 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2702 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2703 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2704 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2705 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2706 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2707 (is_a_helper <gimple_statement_phi>::test): Convert to...
2708 (is_a_helper <gimple_statement_phi *>::test): ...this.
2709 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2710 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2711 (is_a_helper <gimple_statement_try>::test): Convert to...
2712 (is_a_helper <gimple_statement_try *>::test): ...this.
2713 (is_a_helper <gimple_statement_wce>::test): Convert to...
2714 (is_a_helper <gimple_statement_wce *>::test): ...this.
2715 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2716 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2717 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2718 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2719 (is_a_helper <const gimple_statement_call>::test): Convert to...
2720 (is_a_helper <const gimple_statement_call *>::test): ...this.
2721 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2722 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2723 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2724 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2725 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2726 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2727 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2728 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2729 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2731 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2733 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2735 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2737 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2738 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2739 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2741 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2742 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2744 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2745 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2746 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2747 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2748 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2749 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2751 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2752 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2753 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2754 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2756 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2757 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2758 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2759 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2760 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2761 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2762 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2763 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2764 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2765 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2766 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2767 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2768 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2769 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2770 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2771 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2773 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2775 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2776 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2778 (gimple_use_ops): Update for removal of implicit pointer from the
2780 (gimple_set_use_ops): Likewise.
2781 (gimple_vuse): Likewise.
2782 (gimple_vdef): Likewise.
2783 (gimple_vuse_ptr): Likewise.
2784 (gimple_vdef_ptr): Likewise.
2785 (gimple_set_vuse): Likewise.
2786 (gimple_set_vdef): Likewise.
2787 (gimple_omp_return_set_lhs): Likewise.
2788 (gimple_omp_return_lhs): Likewise.
2789 (gimple_omp_return_lhs_ptr): Likewise.
2790 (gimple_call_fntype): Likewise.
2791 (gimple_call_set_fntype): Likewise.
2792 (gimple_call_set_internal_fn): Likewise.
2793 (gimple_call_use_set): Likewise.
2794 (gimple_call_clobber_set): Likewise.
2795 (gimple_bind_vars): Likewise.
2796 (gimple_bind_set_vars): Likewise.
2797 (gimple_bind_body_ptr): Likewise.
2798 (gimple_bind_set_body): Likewise.
2799 (gimple_bind_add_stmt): Likewise.
2800 (gimple_bind_block): Likewise.
2801 (gimple_bind_set_block): Likewise.
2802 (gimple_asm_ninputs): Likewise.
2803 (gimple_asm_noutputs): Likewise.
2804 (gimple_asm_nclobbers): Likewise.
2805 (gimple_asm_nlabels): Likewise.
2806 (gimple_asm_input_op): Likewise.
2807 (gimple_asm_input_op_ptr): Likewise.
2808 (gimple_asm_output_op): Likewise.
2809 (gimple_asm_output_op_ptr): Likewise.
2810 (gimple_asm_set_output_op): Likewise.
2811 (gimple_asm_clobber_op): Likewise.
2812 (gimple_asm_set_clobber_op): Likewise.
2813 (gimple_asm_label_op): Likewise.
2814 (gimple_asm_set_label_op): Likewise.
2815 (gimple_asm_string): Likewise.
2816 (gimple_catch_types): Likewise.
2817 (gimple_catch_types_ptr): Likewise.
2818 (gimple_catch_handler_ptr): Likewise.
2819 (gimple_catch_set_types): Likewise.
2820 (gimple_catch_set_handler): Likewise.
2821 (gimple_eh_filter_types): Likewise.
2822 (gimple_eh_filter_types_ptr): Likewise.
2823 (gimple_eh_filter_failure_ptr): Likewise.
2824 (gimple_eh_filter_set_types): Likewise.
2825 (gimple_eh_filter_set_failure): Likewise.
2826 (gimple_eh_must_not_throw_fndecl): Likewise.
2827 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2828 (gimple_eh_else_n_body_ptr): Likewise.
2829 (gimple_eh_else_e_body_ptr): Likewise.
2830 (gimple_eh_else_set_n_body): Likewise.
2831 (gimple_eh_else_set_e_body): Likewise.
2832 (gimple_try_eval_ptr): Likewise.
2833 (gimple_try_cleanup_ptr): Likewise.
2834 (gimple_try_set_eval): Likewise.
2835 (gimple_try_set_cleanup): Likewise.
2836 (gimple_wce_cleanup_ptr): Likewise.
2837 (gimple_wce_set_cleanup): Likewise.
2838 (gimple_phi_capacity): Likewise.
2839 (gimple_phi_num_args): Likewise.
2840 (gimple_phi_result): Likewise.
2841 (gimple_phi_result_ptr): Likewise.
2842 (gimple_phi_set_result): Likewise.
2843 (gimple_phi_arg): Likewise.
2844 (gimple_phi_set_arg): Likewise.
2845 (gimple_resx_region): Likewise.
2846 (gimple_resx_set_region): Likewise.
2847 (gimple_eh_dispatch_region): Likewise.
2848 (gimple_eh_dispatch_set_region): Likewise.
2849 (gimple_omp_critical_name): Likewise.
2850 (gimple_omp_critical_name_ptr): Likewise.
2851 (gimple_omp_critical_set_name): Likewise.
2852 (gimple_omp_for_clauses): Likewise.
2853 (gimple_omp_for_clauses_ptr): Likewise.
2854 (gimple_omp_for_set_clauses): Likewise.
2855 (gimple_omp_for_collapse): Likewise.
2856 (gimple_omp_for_index): Likewise.
2857 (gimple_omp_for_index_ptr): Likewise.
2858 (gimple_omp_for_set_index): Likewise.
2859 (gimple_omp_for_initial): Likewise.
2860 (gimple_omp_for_initial_ptr): Likewise.
2861 (gimple_omp_for_set_initial): Likewise.
2862 (gimple_omp_for_final): Likewise.
2863 (gimple_omp_for_final_ptr): Likewise.
2864 (gimple_omp_for_set_final): Likewise.
2865 (gimple_omp_for_incr): Likewise.
2866 (gimple_omp_for_incr_ptr): Likewise.
2867 (gimple_omp_for_set_incr): Likewise.
2868 (gimple_omp_for_pre_body_ptr): Likewise.
2869 (gimple_omp_for_set_pre_body): Likewise.
2870 (gimple_omp_parallel_clauses): Likewise.
2871 (gimple_omp_parallel_clauses_ptr): Likewise.
2872 (gimple_omp_parallel_set_clauses): Likewise.
2873 (gimple_omp_parallel_child_fn): Likewise.
2874 (gimple_omp_parallel_child_fn_ptr): Likewise.
2875 (gimple_omp_parallel_set_child_fn): Likewise.
2876 (gimple_omp_parallel_data_arg): Likewise.
2877 (gimple_omp_parallel_data_arg_ptr): Likewise.
2878 (gimple_omp_parallel_set_data_arg): Likewise.
2879 (gimple_omp_task_clauses): Likewise.
2880 (gimple_omp_task_clauses_ptr): Likewise.
2881 (gimple_omp_task_set_clauses): Likewise.
2882 (gimple_omp_task_child_fn): Likewise.
2883 (gimple_omp_task_child_fn_ptr): Likewise.
2884 (gimple_omp_task_set_child_fn): Likewise.
2885 (gimple_omp_task_data_arg): Likewise.
2886 (gimple_omp_task_data_arg_ptr): Likewise.
2887 (gimple_omp_task_set_data_arg): Likewise.
2888 (gimple_omp_taskreg_clauses): Likewise.
2889 (gimple_omp_taskreg_clauses_ptr): Likewise.
2890 (gimple_omp_taskreg_set_clauses): Likewise.
2891 (gimple_omp_taskreg_child_fn): Likewise.
2892 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2893 (gimple_omp_taskreg_set_child_fn): Likewise.
2894 (gimple_omp_taskreg_data_arg): Likewise.
2895 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2896 (gimple_omp_taskreg_set_data_arg): Likewise.
2897 (gimple_omp_task_copy_fn): Likewise.
2898 (gimple_omp_task_copy_fn_ptr): Likewise.
2899 (gimple_omp_task_set_copy_fn): Likewise.
2900 (gimple_omp_task_arg_size): Likewise.
2901 (gimple_omp_task_arg_size_ptr): Likewise.
2902 (gimple_omp_task_set_arg_size): Likewise.
2903 (gimple_omp_task_arg_align): Likewise.
2904 (gimple_omp_task_arg_align_ptr): Likewise.
2905 (gimple_omp_task_set_arg_align): Likewise.
2906 (gimple_omp_single_clauses): Likewise.
2907 (gimple_omp_single_clauses_ptr): Likewise.
2908 (gimple_omp_single_set_clauses): Likewise.
2909 (gimple_omp_target_clauses): Likewise.
2910 (gimple_omp_target_clauses_ptr): Likewise.
2911 (gimple_omp_target_set_clauses): Likewise.
2912 (gimple_omp_target_child_fn): Likewise.
2913 (gimple_omp_target_child_fn_ptr): Likewise.
2914 (gimple_omp_target_set_child_fn): Likewise.
2915 (gimple_omp_target_data_arg): Likewise.
2916 (gimple_omp_target_data_arg_ptr): Likewise.
2917 (gimple_omp_target_set_data_arg): Likewise.
2918 (gimple_omp_teams_clauses): Likewise.
2919 (gimple_omp_teams_clauses_ptr): Likewise.
2920 (gimple_omp_teams_set_clauses): Likewise.
2921 (gimple_omp_sections_clauses): Likewise.
2922 (gimple_omp_sections_clauses_ptr): Likewise.
2923 (gimple_omp_sections_set_clauses): Likewise.
2924 (gimple_omp_sections_control): Likewise.
2925 (gimple_omp_sections_control_ptr): Likewise.
2926 (gimple_omp_sections_set_control): Likewise.
2927 (gimple_omp_for_set_cond): Likewise.
2928 (gimple_omp_for_cond): Likewise.
2929 (gimple_omp_atomic_store_set_val): Likewise.
2930 (gimple_omp_atomic_store_val): Likewise.
2931 (gimple_omp_atomic_store_val_ptr): Likewise.
2932 (gimple_omp_atomic_load_set_lhs): Likewise.
2933 (gimple_omp_atomic_load_lhs): Likewise.
2934 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2935 (gimple_omp_atomic_load_set_rhs): Likewise.
2936 (gimple_omp_atomic_load_rhs): Likewise.
2937 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2938 (gimple_omp_continue_control_def): Likewise.
2939 (gimple_omp_continue_control_def_ptr): Likewise.
2940 (gimple_omp_continue_set_control_def): Likewise.
2941 (gimple_omp_continue_control_use): Likewise.
2942 (gimple_omp_continue_control_use_ptr): Likewise.
2943 (gimple_omp_continue_set_control_use): Likewise.
2944 (gimple_transaction_body_ptr): Likewise.
2945 (gimple_transaction_label): Likewise.
2946 (gimple_transaction_label_ptr): Likewise.
2947 (gimple_transaction_set_body): Likewise.
2948 (gimple_transaction_set_label): Likewise.
2950 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2951 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2952 * ipa-ref.c (ipa_record_reference): Likewise.
2953 * ipa-reference.c (analyze_function): Likewise.
2954 (ipa_reference_write_optimization_summary): Likewise.
2955 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2956 (address_taken_from_non_vtable_p): Likewise.
2957 (comdat_can_be_unshared_p_1): Likewise.
2958 * lto-cgraph.c (lto_output_ref): Likewise.
2959 (add_references): Likewise.
2960 (compute_ltrans_boundary): Likewise.
2961 (output_symtab): Likewise.
2962 (input_ref): Likewise.
2963 (input_cgraph_1): Likewise.
2964 (output_cgraph_opt_summary): Likewise.
2965 * lto-streamer-out.c (lto_output): Likewise.
2966 (output_symbol_p): Likewise.
2967 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2968 (lsei_start_function_in_partition): Likewise.
2969 (lsei_next_variable_in_partition): Likewise.
2970 (lsei_start_variable_in_partition): Likewise.
2971 * symtab.c (insert_to_assembler_name_hash): Likewise.
2972 (unlink_from_assembler_name_hash): Likewise.
2973 (symtab_unregister_node): Likewise.
2974 (symtab_remove_node): Likewise.
2975 (dump_symtab_node): Likewise.
2976 (verify_symtab_base): Likewise.
2977 (verify_symtab_node): Likewise.
2978 (symtab_make_decl_local): Likewise.
2979 (symtab_alias_ultimate_target): Likewise.
2980 (symtab_resolve_alias): Likewise.
2981 (symtab_get_symbol_partitioning_class): Likewise.
2982 * tree-phinodes.c (allocate_phi_node): Likewise.
2983 (reserve_phi_args_for_new_edge): Likewise.
2984 (remove_phi_args): Likewise.
2985 * varpool.c (varpool_node_for_asm): Likewise.
2986 (varpool_remove_unreferenced_decls): Likewise.
2988 2014-04-23 Jeff Law <law@redhat.com>
2990 PR tree-optimization/60902
2991 * tree-ssa-threadedge.c
2992 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2993 invalidate outputs from statements that do not produce useful
2994 outputs for threading.
2996 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2998 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2999 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3000 machine descriptions for Stack Smashing Protector.
3002 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3004 * aarch64.md (<optab>_rol<mode>3): New pattern.
3005 (<optab>_rolsi3_uxtw): Likewise.
3006 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3008 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3010 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3011 (arm_cortex_a12_tune): Likewise.
3013 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3015 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3017 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3019 * config/arm/arm.md (arm_rev16si2): New pattern.
3020 (arm_rev16si2_alt): Likewise.
3021 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3023 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3025 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3026 (rev16<mode>2_alt): Likewise.
3027 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3028 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3029 (aarch_rev16_shleft_mask_imm_p): Likewise.
3030 (aarch_rev16_p_1): Likewise.
3031 (aarch_rev16_p): Likewise.
3032 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3033 (aarch_rev16_shright_mask_imm_p): Likewise.
3034 (aarch_rev16_shleft_mask_imm_p): Likewise.
3036 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3038 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3039 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3041 (cortex_a53_extra_costs): Likewise.
3042 (cortex_a57_extra_costs): Likewise.
3043 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3044 (cortexa7_extra_costs): Likewise.
3045 (cortexa8_extra_costs): Likewise.
3046 (cortexa12_extra_costs): Likewise.
3047 (cortexa15_extra_costs): Likewise.
3048 (v7m_extra_costs): Likewise.
3049 (arm_new_rtx_costs): Handle BSWAP.
3051 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3053 * config/arm/arm.c (cortexa8_extra_costs): New table.
3054 (arm_cortex_a8_tune): New tuning struct.
3055 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3057 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3059 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3061 2014-04-23 Richard Biener <rguenther@suse.de>
3063 * Makefile.in (OBJS): Remove loop-unswitch.o.
3064 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3065 * passes.def (pass_rtl_unswitch): Likewise.
3066 * loop-init.c (gate_rtl_unswitch): Likewise.
3067 (rtl_unswitch): Likewise.
3068 (pass_data_rtl_unswitch): Likewise.
3069 (pass_rtl_unswitch): Likewise.
3070 (make_pass_rtl_unswitch): Likewise.
3071 * rtl.h (reversed_condition): Likewise.
3072 (compare_and_jump_seq): Likewise.
3073 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3075 * loop-unroll.c (compare_and_jump_seq): Likewise.
3077 2014-04-23 Richard Biener <rguenther@suse.de>
3079 PR tree-optimization/60903
3080 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3081 commented code block.
3082 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3083 loop flags to newly created BBs and edges.
3085 2014-04-23 Nick Clifton <nickc@redhat.com>
3087 * config/msp430/msp430.c (msp430_handle_option): Move function
3089 (msp430_option_override): Simplify mcu and mcpu option handling.
3090 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3091 support for -mhwmult command line option.
3092 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3093 -mhwmult command line option.
3094 (msp430_hwmult_enabled): Delete.
3095 (msp43o_output_labelref): Add support for -mhwmult command line option.
3096 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3097 (umulsidi3): Likewise.
3098 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3099 (mcpu, mlarge, msmall): Likewise.
3100 (mhwmult): New option.
3101 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3103 (msp430_is_f5_mcu): Remove prototype.
3104 (msp430_use_f5_series_hwmult): Add prototype.
3105 * config/msp430/msp430-opts.h: New file.
3106 * common/config/msp430: New directory.
3107 * common/config/msp430/msp430-common.c: New file.
3108 * config.gcc (msp430): Remove target_has_targetm_common.
3109 * doc/invoke.texi: Document -mhwmult command line option.
3111 2014-04-23 Nick Clifton <nickc@redhat.com>
3113 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3114 default-manifest.o if it can be found in the search path.
3115 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3117 2014-04-23 Terry Guo <terry.guo@arm.com>
3119 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3121 2014-04-23 Richard Biener <rguenther@suse.de>
3124 * tree-inline.c (declare_return_variable): Use mark_addressable.
3126 2014-04-23 Richard Biener <rguenther@suse.de>
3129 * loop-init.c (loop_optimizer_init): Make sure to apply
3130 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3132 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3135 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3137 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3138 if flag_sanitize_undefined_trap_on_error.
3139 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3140 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3141 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3142 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3143 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3144 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3145 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3146 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3147 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3148 * ubsan.c (ubsan_instrument_unreachable): Return
3149 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3150 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3151 if flag_sanitize_undefined_trap_on_error and
3152 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3153 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3154 instrument_bool_enum_load): Emit __builtin_trap () if
3155 flag_sanitize_undefined_trap_on_error and
3156 __builtin_handle_*_abort () if !flag_sanitize_recover.
3157 * doc/invoke.texi (-fsanitize-recover,
3158 -fsanitize-undefined-trap-on-error): Document.
3160 2014-04-22 Christian Bruel <christian.bruel@st.com>
3162 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3163 Force immediates to SImode.
3165 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3167 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3169 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3170 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3171 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3172 (nios2_fpu_insn): Add entry for round.
3173 (N2FPU_NO_ERRNO_P): Define.
3174 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3176 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3178 2014-04-22 Richard Henderson <rth@redhat.com>
3180 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3181 (add<GPI>3_compare0): Remove leading * from name.
3182 (add<GPI>3_carryin): Likewise.
3183 (sub<GPI>3_compare0): Likewise.
3184 (sub<GPI>3_carryin): Likewise.
3185 (<su_optab>mulditi3): New expander.
3186 (multi3): New expander.
3187 (madd<GPI>): Remove leading * from name.
3189 2014-04-22 Martin Jambor <mjambor@suse.cz>
3191 * cgraphclones.c (cgraph_function_versioning): Copy
3192 ipa_transforms_to_apply instead of asserting it is empty.
3194 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3197 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3198 on count_exp to get mode.
3200 2014-04-22 Andrew Pinski <apinski@cavium.com>
3202 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3203 Handle TLS for ILP32.
3204 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3205 (tlsie_small_<mode>): this and handle PTR.
3206 (tlsie_small_sidi): New pattern.
3207 (tlsle_small): Change to an expand to handle ILP32.
3208 (tlsle_small_<mode>): New pattern.
3209 (tlsdesc_small): Rename to ...
3210 (tlsdesc_small_<mode>): this and handle PTR.
3212 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3214 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3216 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3218 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3219 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3220 (aarch64_types_signed_poly_qualifiers): Likewise.
3221 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3222 (aarch64_types_poly_signed_qualifiers): Likewise.
3223 (TYPES_REINTERP_SS): Type macro added.
3224 (TYPES_REINTERP_SU): Likewise.
3225 (TYPES_REINTERP_SP): Likewise.
3226 (TYPES_REINTERP_US): Likewise.
3227 (TYPES_REINTERP_PS): Likewise.
3228 (aarch64_fold_builtin): New expression folding added.
3229 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3230 Declarations removed.
3231 (REINTERP_SS): Declarations added.
3232 (REINTERP_US): Likewise.
3233 (REINTERP_PS): Likewise.
3234 (REINTERP_SU): Likewise.
3235 (REINTERP_SP): Likewise.
3236 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3237 (vreinterpretq_p8_f64): Likewise.
3238 (vreinterpret_p16_f64): Likewise.
3239 (vreinterpretq_p16_f64): Likewise.
3240 (vreinterpret_f32_f64): Likewise.
3241 (vreinterpretq_f32_f64): Likewise.
3242 (vreinterpret_f64_f32): Likewise.
3243 (vreinterpret_f64_p8): Likewise.
3244 (vreinterpret_f64_p16): Likewise.
3245 (vreinterpret_f64_s8): Likewise.
3246 (vreinterpret_f64_s16): Likewise.
3247 (vreinterpret_f64_s32): Likewise.
3248 (vreinterpret_f64_s64): Likewise.
3249 (vreinterpret_f64_u8): Likewise.
3250 (vreinterpret_f64_u16): Likewise.
3251 (vreinterpret_f64_u32): Likewise.
3252 (vreinterpret_f64_u64): Likewise.
3253 (vreinterpretq_f64_f32): Likewise.
3254 (vreinterpretq_f64_p8): Likewise.
3255 (vreinterpretq_f64_p16): Likewise.
3256 (vreinterpretq_f64_s8): Likewise.
3257 (vreinterpretq_f64_s16): Likewise.
3258 (vreinterpretq_f64_s32): Likewise.
3259 (vreinterpretq_f64_s64): Likewise.
3260 (vreinterpretq_f64_u8): Likewise.
3261 (vreinterpretq_f64_u16): Likewise.
3262 (vreinterpretq_f64_u32): Likewise.
3263 (vreinterpretq_f64_u64): Likewise.
3264 (vreinterpret_s64_f64): Likewise.
3265 (vreinterpretq_s64_f64): Likewise.
3266 (vreinterpret_u64_f64): Likewise.
3267 (vreinterpretq_u64_f64): Likewise.
3268 (vreinterpret_s8_f64): Likewise.
3269 (vreinterpretq_s8_f64): Likewise.
3270 (vreinterpret_s16_f64): Likewise.
3271 (vreinterpretq_s16_f64): Likewise.
3272 (vreinterpret_s32_f64): Likewise.
3273 (vreinterpretq_s32_f64): Likewise.
3274 (vreinterpret_u8_f64): Likewise.
3275 (vreinterpretq_u8_f64): Likewise.
3276 (vreinterpret_u16_f64): Likewise.
3277 (vreinterpretq_u16_f64): Likewise.
3278 (vreinterpret_u32_f64): Likewise.
3279 (vreinterpretq_u32_f64): Likewise.
3281 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3283 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3284 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3285 (vreinterpret_p8_s8): Likewise.
3286 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3287 (vreinterpret_p8_s16): Likewise.
3288 (vreinterpret_p8_s32): Likewise.
3289 (vreinterpret_p8_s64): Likewise.
3290 (vreinterpret_p8_f32): Likewise.
3291 (vreinterpret_p8_u8): Likewise.
3292 (vreinterpret_p8_u16): Likewise.
3293 (vreinterpret_p8_u32): Likewise.
3294 (vreinterpret_p8_u64): Likewise.
3295 (vreinterpret_p8_p16): Likewise.
3296 (vreinterpretq_p8_s8): Likewise.
3297 (vreinterpretq_p8_s16): Likewise.
3298 (vreinterpretq_p8_s32): Likewise.
3299 (vreinterpretq_p8_s64): Likewise.
3300 (vreinterpretq_p8_f32): Likewise.
3301 (vreinterpretq_p8_u8): Likewise.
3302 (vreinterpretq_p8_u16): Likewise.
3303 (vreinterpretq_p8_u32): Likewise.
3304 (vreinterpretq_p8_u64): Likewise.
3305 (vreinterpretq_p8_p16): Likewise.
3306 (vreinterpret_p16_s8): Likewise.
3307 (vreinterpret_p16_s16): Likewise.
3308 (vreinterpret_p16_s32): Likewise.
3309 (vreinterpret_p16_s64): Likewise.
3310 (vreinterpret_p16_f32): Likewise.
3311 (vreinterpret_p16_u8): Likewise.
3312 (vreinterpret_p16_u16): Likewise.
3313 (vreinterpret_p16_u32): Likewise.
3314 (vreinterpret_p16_u64): Likewise.
3315 (vreinterpret_p16_p8): Likewise.
3316 (vreinterpretq_p16_s8): Likewise.
3317 (vreinterpretq_p16_s16): Likewise.
3318 (vreinterpretq_p16_s32): Likewise.
3319 (vreinterpretq_p16_s64): Likewise.
3320 (vreinterpretq_p16_f32): Likewise.
3321 (vreinterpretq_p16_u8): Likewise.
3322 (vreinterpretq_p16_u16): Likewise.
3323 (vreinterpretq_p16_u32): Likewise.
3324 (vreinterpretq_p16_u64): Likewise.
3325 (vreinterpretq_p16_p8): Likewise.
3326 (vreinterpret_f32_s8): Likewise.
3327 (vreinterpret_f32_s16): Likewise.
3328 (vreinterpret_f32_s32): Likewise.
3329 (vreinterpret_f32_s64): Likewise.
3330 (vreinterpret_f32_u8): Likewise.
3331 (vreinterpret_f32_u16): Likewise.
3332 (vreinterpret_f32_u32): Likewise.
3333 (vreinterpret_f32_u64): Likewise.
3334 (vreinterpret_f32_p8): Likewise.
3335 (vreinterpret_f32_p16): Likewise.
3336 (vreinterpretq_f32_s8): Likewise.
3337 (vreinterpretq_f32_s16): Likewise.
3338 (vreinterpretq_f32_s32): Likewise.
3339 (vreinterpretq_f32_s64): Likewise.
3340 (vreinterpretq_f32_u8): Likewise.
3341 (vreinterpretq_f32_u16): Likewise.
3342 (vreinterpretq_f32_u32): Likewise.
3343 (vreinterpretq_f32_u64): Likewise.
3344 (vreinterpretq_f32_p8): Likewise.
3345 (vreinterpretq_f32_p16): Likewise.
3346 (vreinterpret_s64_s8): Likewise.
3347 (vreinterpret_s64_s16): Likewise.
3348 (vreinterpret_s64_s32): Likewise.
3349 (vreinterpret_s64_f32): Likewise.
3350 (vreinterpret_s64_u8): Likewise.
3351 (vreinterpret_s64_u16): Likewise.
3352 (vreinterpret_s64_u32): Likewise.
3353 (vreinterpret_s64_u64): Likewise.
3354 (vreinterpret_s64_p8): Likewise.
3355 (vreinterpret_s64_p16): Likewise.
3356 (vreinterpretq_s64_s8): Likewise.
3357 (vreinterpretq_s64_s16): Likewise.
3358 (vreinterpretq_s64_s32): Likewise.
3359 (vreinterpretq_s64_f32): Likewise.
3360 (vreinterpretq_s64_u8): Likewise.
3361 (vreinterpretq_s64_u16): Likewise.
3362 (vreinterpretq_s64_u32): Likewise.
3363 (vreinterpretq_s64_u64): Likewise.
3364 (vreinterpretq_s64_p8): Likewise.
3365 (vreinterpretq_s64_p16): Likewise.
3366 (vreinterpret_u64_s8): Likewise.
3367 (vreinterpret_u64_s16): Likewise.
3368 (vreinterpret_u64_s32): Likewise.
3369 (vreinterpret_u64_s64): Likewise.
3370 (vreinterpret_u64_f32): Likewise.
3371 (vreinterpret_u64_u8): Likewise.
3372 (vreinterpret_u64_u16): Likewise.
3373 (vreinterpret_u64_u32): Likewise.
3374 (vreinterpret_u64_p8): Likewise.
3375 (vreinterpret_u64_p16): Likewise.
3376 (vreinterpretq_u64_s8): Likewise.
3377 (vreinterpretq_u64_s16): Likewise.
3378 (vreinterpretq_u64_s32): Likewise.
3379 (vreinterpretq_u64_s64): Likewise.
3380 (vreinterpretq_u64_f32): Likewise.
3381 (vreinterpretq_u64_u8): Likewise.
3382 (vreinterpretq_u64_u16): Likewise.
3383 (vreinterpretq_u64_u32): Likewise.
3384 (vreinterpretq_u64_p8): Likewise.
3385 (vreinterpretq_u64_p16): Likewise.
3386 (vreinterpret_s8_s16): Likewise.
3387 (vreinterpret_s8_s32): Likewise.
3388 (vreinterpret_s8_s64): Likewise.
3389 (vreinterpret_s8_f32): Likewise.
3390 (vreinterpret_s8_u8): Likewise.
3391 (vreinterpret_s8_u16): Likewise.
3392 (vreinterpret_s8_u32): Likewise.
3393 (vreinterpret_s8_u64): Likewise.
3394 (vreinterpret_s8_p8): Likewise.
3395 (vreinterpret_s8_p16): Likewise.
3396 (vreinterpretq_s8_s16): Likewise.
3397 (vreinterpretq_s8_s32): Likewise.
3398 (vreinterpretq_s8_s64): Likewise.
3399 (vreinterpretq_s8_f32): Likewise.
3400 (vreinterpretq_s8_u8): Likewise.
3401 (vreinterpretq_s8_u16): Likewise.
3402 (vreinterpretq_s8_u32): Likewise.
3403 (vreinterpretq_s8_u64): Likewise.
3404 (vreinterpretq_s8_p8): Likewise.
3405 (vreinterpretq_s8_p16): Likewise.
3406 (vreinterpret_s16_s8): Likewise.
3407 (vreinterpret_s16_s32): Likewise.
3408 (vreinterpret_s16_s64): Likewise.
3409 (vreinterpret_s16_f32): Likewise.
3410 (vreinterpret_s16_u8): Likewise.
3411 (vreinterpret_s16_u16): Likewise.
3412 (vreinterpret_s16_u32): Likewise.
3413 (vreinterpret_s16_u64): Likewise.
3414 (vreinterpret_s16_p8): Likewise.
3415 (vreinterpret_s16_p16): Likewise.
3416 (vreinterpretq_s16_s8): Likewise.
3417 (vreinterpretq_s16_s32): Likewise.
3418 (vreinterpretq_s16_s64): Likewise.
3419 (vreinterpretq_s16_f32): Likewise.
3420 (vreinterpretq_s16_u8): Likewise.
3421 (vreinterpretq_s16_u16): Likewise.
3422 (vreinterpretq_s16_u32): Likewise.
3423 (vreinterpretq_s16_u64): Likewise.
3424 (vreinterpretq_s16_p8): Likewise.
3425 (vreinterpretq_s16_p16): Likewise.
3426 (vreinterpret_s32_s8): Likewise.
3427 (vreinterpret_s32_s16): Likewise.
3428 (vreinterpret_s32_s64): Likewise.
3429 (vreinterpret_s32_f32): Likewise.
3430 (vreinterpret_s32_u8): Likewise.
3431 (vreinterpret_s32_u16): Likewise.
3432 (vreinterpret_s32_u32): Likewise.
3433 (vreinterpret_s32_u64): Likewise.
3434 (vreinterpret_s32_p8): Likewise.
3435 (vreinterpret_s32_p16): Likewise.
3436 (vreinterpretq_s32_s8): Likewise.
3437 (vreinterpretq_s32_s16): Likewise.
3438 (vreinterpretq_s32_s64): Likewise.
3439 (vreinterpretq_s32_f32): Likewise.
3440 (vreinterpretq_s32_u8): Likewise.
3441 (vreinterpretq_s32_u16): Likewise.
3442 (vreinterpretq_s32_u32): Likewise.
3443 (vreinterpretq_s32_u64): Likewise.
3444 (vreinterpretq_s32_p8): Likewise.
3445 (vreinterpretq_s32_p16): Likewise.
3446 (vreinterpret_u8_s8): Likewise.
3447 (vreinterpret_u8_s16): Likewise.
3448 (vreinterpret_u8_s32): Likewise.
3449 (vreinterpret_u8_s64): Likewise.
3450 (vreinterpret_u8_f32): Likewise.
3451 (vreinterpret_u8_u16): Likewise.
3452 (vreinterpret_u8_u32): Likewise.
3453 (vreinterpret_u8_u64): Likewise.
3454 (vreinterpret_u8_p8): Likewise.
3455 (vreinterpret_u8_p16): Likewise.
3456 (vreinterpretq_u8_s8): Likewise.
3457 (vreinterpretq_u8_s16): Likewise.
3458 (vreinterpretq_u8_s32): Likewise.
3459 (vreinterpretq_u8_s64): Likewise.
3460 (vreinterpretq_u8_f32): Likewise.
3461 (vreinterpretq_u8_u16): Likewise.
3462 (vreinterpretq_u8_u32): Likewise.
3463 (vreinterpretq_u8_u64): Likewise.
3464 (vreinterpretq_u8_p8): Likewise.
3465 (vreinterpretq_u8_p16): Likewise.
3466 (vreinterpret_u16_s8): Likewise.
3467 (vreinterpret_u16_s16): Likewise.
3468 (vreinterpret_u16_s32): Likewise.
3469 (vreinterpret_u16_s64): Likewise.
3470 (vreinterpret_u16_f32): Likewise.
3471 (vreinterpret_u16_u8): Likewise.
3472 (vreinterpret_u16_u32): Likewise.
3473 (vreinterpret_u16_u64): Likewise.
3474 (vreinterpret_u16_p8): Likewise.
3475 (vreinterpret_u16_p16): Likewise.
3476 (vreinterpretq_u16_s8): Likewise.
3477 (vreinterpretq_u16_s16): Likewise.
3478 (vreinterpretq_u16_s32): Likewise.
3479 (vreinterpretq_u16_s64): Likewise.
3480 (vreinterpretq_u16_f32): Likewise.
3481 (vreinterpretq_u16_u8): Likewise.
3482 (vreinterpretq_u16_u32): Likewise.
3483 (vreinterpretq_u16_u64): Likewise.
3484 (vreinterpretq_u16_p8): Likewise.
3485 (vreinterpretq_u16_p16): Likewise.
3486 (vreinterpret_u32_s8): Likewise.
3487 (vreinterpret_u32_s16): Likewise.
3488 (vreinterpret_u32_s32): Likewise.
3489 (vreinterpret_u32_s64): Likewise.
3490 (vreinterpret_u32_f32): Likewise.
3491 (vreinterpret_u32_u8): Likewise.
3492 (vreinterpret_u32_u16): Likewise.
3493 (vreinterpret_u32_u64): Likewise.
3494 (vreinterpret_u32_p8): Likewise.
3495 (vreinterpret_u32_p16): Likewise.
3496 (vreinterpretq_u32_s8): Likewise.
3497 (vreinterpretq_u32_s16): Likewise.
3498 (vreinterpretq_u32_s32): Likewise.
3499 (vreinterpretq_u32_s64): Likewise.
3500 (vreinterpretq_u32_f32): Likewise.
3501 (vreinterpretq_u32_u8): Likewise.
3502 (vreinterpretq_u32_u16): Likewise.
3503 (vreinterpretq_u32_u64): Likewise.
3504 (vreinterpretq_u32_p8): Likewise.
3505 (vreinterpretq_u32_p16): Likewise.
3507 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3509 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3511 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3513 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3514 (vqnegd_s64): Likewise.
3515 (vqabs_s64): Likewise.
3516 (vqabsd_s64): Likewise.
3518 2014-04-22 Richard Henderson <rth@redhat.com>
3520 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3521 computation to the top of the loop.
3523 2014-04-22 Renlin <renlin.li@arm.com>
3524 Jiong Wang <jiong.wang@arm.com>
3526 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3527 * config/aarch64/aarch64.c (aarch64_layout_frame)
3528 (aarch64_initial_elimination_offset): Likewise.
3530 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3532 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3535 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3537 * machmode.h (bitwise_mode_for_mode): Declare.
3538 * stor-layout.h (bitwise_type_for_mode): Likewise.
3539 * stor-layout.c (bitwise_mode_for_mode): New function.
3540 (bitwise_type_for_mode): Likewise.
3541 * builtins.c (fold_builtin_memory_op): Use it instead of
3542 int_mode_for_mode and build_nonstandard_integer_type.
3544 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3546 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3547 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3548 (*-*-solaris2*): Simplify.
3549 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3550 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3551 *-*-solaris2.9* handling.
3553 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3555 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3556 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3558 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3559 * configure: Regenerate.
3561 * config/i386/sol2-9.h: Remove.
3563 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3564 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3565 Remove Solaris 9 references.
3567 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3569 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3570 (floatuns<GPI:mode><GPF:mode>2): Remove.
3571 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3572 and floatuns conversions.
3573 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3574 and floatuns conversions.
3575 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3576 (w1,w2): New mode attributes for inequal width conversions.
3578 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3580 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3581 the output asm format.
3583 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3585 * config/aarch64/aarch64-simd.md
3586 (aarch64_cm<optab>di): Always split.
3587 (*aarch64_cm<optab>di): New.
3588 (aarch64_cmtstdi): Always split.
3589 (*aarch64_cmtstdi): New.
3591 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3593 PR tree-optimization/60823
3594 * omp-low.c (ipa_simd_modify_function_body): Go through
3595 all SSA_NAMEs and for those refering to vector arguments
3596 which are going to be replaced adjust SSA_NAME_VAR and,
3597 if it is a default definition, change it into a non-default
3598 definition assigned at the beginning of function from new_decl.
3599 (ipa_simd_modify_stmt_ops): Rewritten.
3600 * tree-dfa.c (set_ssa_default_def): When removing default def,
3601 check for NULL loc instead of NULL *loc.
3603 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3605 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3606 restrictions on core registers for DImode values in Thumb2.
3608 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3610 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3611 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3613 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3615 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3616 (*iordi_notzesidi_di): Likewise.
3617 (*iordi_notsesidi_di): Likewise.
3619 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3621 * config/arm/arm-protos.h (tune_params): New struct members.
3622 * config/arm/arm.c: Initialise tune_params per processor.
3623 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3624 for speed, based on new tune_params.
3626 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3628 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3629 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3630 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3631 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3632 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3633 (vrnda_f64): Likewise.
3634 (vrndi_f64): Likewise.
3635 (vrndm_f64): Likewise.
3636 (vrndn_f64): Likewise.
3637 (vrndp_f64): Likewise.
3638 (vrndx_f64): Likewise.
3640 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3642 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3643 GET_MODE_SIZE argument is enum machine_mode.
3645 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3648 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3649 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3651 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3654 * asan.c (asan_emit_stack_protection): Force the base to align to
3655 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3656 appropriate bits if STRICT_ALIGNMENT.
3657 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3659 (expand_used_vars): Leave a space in the stack frame for alignment
3660 if STRICT_ALIGNMENT.
3662 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3664 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3666 (gimple_store_p): Likewise.
3667 (gimple_assign_load_p): Likewise.
3668 (gimple_assign_cast_p): Likewise.
3669 (gimple_clobber_p): Likewise.
3671 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3672 rather than a gimple.
3673 (gimple_assign_cast_p): Likewise.
3675 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3678 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3679 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3681 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3682 more debug information for E500 if -mdebug=reg.
3684 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3687 * config/i386/i386.c (ix86_expand_builtin)
3688 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3689 register for target RTX.
3690 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3692 2014-04-18 Cong Hou <congh@google.com>
3694 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3695 the widen-mult pattern by handling two operands with different sizes,
3696 and operands whose size is smaller than half of the result type.
3698 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3700 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3701 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3702 (do_estimate_edge_time): Compute it.
3703 * ipa-inline.c (want_inline_small_function_p): Bypass
3704 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3706 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3708 * ipa-inline.c (spec_rem): New static variable.
3709 (dump_overall_stats): New function.
3710 (dump_inline_stats): New function.
3712 2014-04-18 Richard Henderson <rth@redhat.com>
3714 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3715 to GET_MODE_SIZE, not a reg_class_t.
3717 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3719 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3720 (vsx_xxmrglw_<mode>): Likewise.
3722 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3725 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3726 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3727 (rs6000_init_hard_regno_mode_ok): Likewise.
3729 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3731 * ipa-inline.c (inline_small_functions): Account only non-cold
3733 * doc/invoke.texi (inline-unit-growth): Update documentation.
3735 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3737 * config/rs6000/rs6000.md (addti3, subti3): New.
3739 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3742 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3743 comment. Check optimize_insn_for_size_p instead of
3744 optimize_insn_for_speed_p.
3746 2014-04-17 Martin Jambor <mjambor@suse.cz>
3748 * gimple-iterator.c (gsi_start_edge): New function.
3749 * gimple-iterator.h (gsi_start_edge): Declare.
3750 * tree-sra.c (single_non_eh_succ): New function.
3751 (disqualify_ops_if_throwing_stmt): Renamed to
3752 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3753 having one non-EH successor BB.
3754 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3755 generate loads into replacements.
3756 (sra_modify_assign): Likewise and and also use the simple path for
3758 (sra_modify_function_body): Commit statements on edges.
3760 2014-04-17 Richard Biener <rguenther@suse.de>
3763 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3764 comparison results and add clarifying comment.
3766 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3768 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3769 (blank_mode): Initialize it.
3770 (emit_mode_size_inline, emit_mode_nunits_inline,
3771 emit_mode_inner_inline): New functions.
3772 (emit_insn_modes_h): Call them and surround their output with
3773 #if GCC_VERSION >= 4001 ... #endif.
3774 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3775 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3776 mode_* arrays if the argument is __builtin_constant_p.
3777 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3778 is enum machine_mode.
3780 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3782 * passes.c (opt_pass::execute): Adjust.
3783 (pass_manager::execute_pass_mode_switching): Likewise.
3784 (early_local_passes::execute): Likewise.
3785 (execute_one_pass): Pass cfun to the pass's execute method.
3786 * tree-pass.h (opt_pass::execute): Add function * argument.
3787 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3788 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3789 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3790 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3791 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3792 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3793 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3794 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3795 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3796 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3797 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3798 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3799 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3800 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3801 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3802 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3803 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3804 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3805 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3806 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3807 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3808 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3809 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3810 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3811 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3812 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3813 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3814 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3815 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3818 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3820 * passes.c (opt_pass::gate): Take function * argument.
3821 (gate_all_early_local_passes): Merge into
3822 (early_local_passes::gate): this.
3823 (gate_all_early_optimizations): Merge into
3824 (all_early_optimizations::gate): this.
3825 (gate_all_optimizations): Mege into
3826 (all_optimizations::gate): this.
3827 (gate_all_optimizations_g): Merge into
3828 (all_optimizations_g::gate): this.
3829 (gate_rest_of_compilation): Mege into
3830 (rest_of_compilation::gate): this.
3831 (gate_postreload): Merge into
3832 (postreload::gate): this.
3833 (dump_one_pass): Pass cfun to the pass's gate method.
3834 (execute_ipa_summary_passes): Likewise.
3835 (execute_one_pass): Likewise.
3836 (ipa_write_summaries_2): Likewise.
3837 (ipa_write_optimization_summaries_1): Likewise.
3838 (ipa_read_summaries_1): Likewise.
3839 (ipa_read_optimization_summaries_1): Likewise.
3840 (execute_ipa_stmt_fixups): Likewise.
3841 * tree-pass.h (opt_pass::gate): Add function * argument.
3842 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3843 combine-stack-adj.c, combine.c, compare-elim.c,
3844 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3845 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3846 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3847 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3848 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3849 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3850 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3851 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3852 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3853 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3854 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3855 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3856 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3857 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3858 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3859 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3860 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3861 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3862 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3863 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3864 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3865 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3866 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3867 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3868 var-tracking.c, vtable-verify.c, web.c: Adjust.
3870 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3872 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3873 * configure: Regenerate.
3875 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3877 * passes.c (dump_one_pass): don't check pass->has_gate.
3878 (execute_ipa_summary_passes): Likewise.
3879 (execute_one_pass): Likewise.
3880 (ipa_write_summaries_2): Likewise.
3881 (ipa_write_optimization_summaries_1): Likewise.
3882 (ipa_read_optimization_summaries_1): Likewise.
3883 (execute_ipa_stmt_fixups): Likewise.
3884 * tree-pass.h (pass_data::has_gate): Remove.
3885 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3886 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3887 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3888 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3889 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3890 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3891 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3892 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3893 gimple-low.c, gimple-ssa-isolate-paths.c,
3894 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3895 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3896 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3897 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3898 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3899 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3900 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3901 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3902 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3903 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3904 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3905 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3906 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3907 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3908 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3909 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3910 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3911 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3912 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3913 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3914 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3915 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3918 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3920 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3922 * passes.c (pass_manager::register_dump_files_1): Merge into
3923 (pass_manager::register_dump_files): this, and remove its handling of
3924 properties since the pass always has the properties anyway.
3925 (pass_manager::pass_manager): Adjust.
3927 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3929 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3930 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3931 dealing with properties.
3932 (pass_manager::register_dump_files): Adjust.
3934 2014-03-20 Mark Wielaard <mjw@redhat.com>
3936 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3937 then represent the bound as normal constant value.
3939 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3942 Forward port from 4.8 branch
3943 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3945 * config/i386/bmiintrin.h (_blsi_u32): New.
3949 (_blsmsk_u32): Ditto.
3950 (_blsmsk_u64): Ditto.
3951 (_tzcnt_u32): Ditto.
3952 (_tzcnt_u64): Ditto.
3954 2014-04-17 Kito Cheng <kito@0xlab.org>
3956 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3958 2014-04-17 Richard Biener <rguenther@suse.de>
3961 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3962 boolean results for comparisons.
3964 2014-04-17 Richard Biener <rguenther@suse.de>
3966 PR tree-optimization/60836
3967 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3968 initial PHI args to be gimple values.
3970 2014-04-17 Richard Biener <rguenther@suse.de>
3972 PR tree-optimization/60841
3973 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3974 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3975 of stmts to SLP build.
3976 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3977 (vect_analyze_slp): Likewise.
3978 (vect_analyze_slp_instance): Likewise.
3979 (vect_build_slp_tree): Limit overall SLP tree growth.
3980 * tree-vectorizer.h (vect_analyze_data_refs,
3981 vect_analyze_slp): Adjust prototypes.
3983 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3985 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3988 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3990 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3991 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3992 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3993 for TARGET_SLOW_PSHUFB
3995 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3997 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3998 * config/i386/i386.c (intel_cost): Ditto.
4000 2014-04-17 Joey Ye <joey.ye@arm.com>
4002 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4004 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4006 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4009 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4011 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4012 (type_all_derivations_known_p): New predicate.
4013 (type_all_ctors_visible_p): New predicate.
4014 (type_possibly_instantiated_p): New predicate.
4015 (get_odr_type): Compute all_derivations_known.
4016 (dump_odr_type): Dump the flag.
4017 (maybe_record_type): Cleanup.
4018 (record_target_from_binfo): Add bases_to_consider array;
4019 record bases for types w/o instances and skip CXX destructor.
4020 (possible_polymorphic_call_targets_1): Add bases_to_consider
4021 and consider_construction parameters; check if type may have instance.
4022 (get_polymorphic_call_info): Set maybe_in_construction to true
4023 when we know nothing.
4024 (record_targets_from_bases): Skip CXX destructors; they are
4025 never called for types in construction.
4026 (possible_polymorphic_call_targets): Do not record target when
4027 type may not have instance.
4029 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4032 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4033 external aliases alive, too.
4035 2014-04-16 Andrew Pinski <apinski@cavium.com>
4037 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4040 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4042 * final.c (compute_alignments): Do not apply loop alignment to a block
4043 falling through to the exit.
4045 2014-04-16 Catherine Moore <clm@codesourcery.com>
4047 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4048 Adjust constraints for microMIPS store patterns.
4050 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4052 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4054 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4056 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4057 (append_use): Run at -O0.
4058 (append_vdef): Likewise.
4059 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4060 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4062 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4064 PR tree-optimization/60844
4065 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4066 (propagate_op_to_single_use, remove_visited_stmt_chain,
4067 linearize_expr, repropagate_negates, reassociate_bb): Use it
4068 instead of gsi_remove.
4070 2014-04-16 Martin Jambor <mjambor@suse.cz>
4072 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4073 ipa_transforms_to_apply.
4074 (cgraph_function_versioning): Assert that old_node has empty
4075 ipa_transforms_to_apply.
4076 * trans-mem.c (ipa_tm_create_version): Likewise.
4077 * tree-inline.c (tree_function_versioning): Do not duplicate
4078 ipa_transforms_to_apply.
4080 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4083 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4085 Pass necessary as flags on 64-bit Solaris/x86.
4086 Use lowercase relocs for x86_64-*-*.
4087 * configure: Regenerate.
4089 2014-04-15 Jan Hubicka <jh@suse.cz>
4091 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4092 (maybe_record_node, likely_target_p): Use it.
4094 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4097 Revert following patch
4099 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4102 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4103 software floating point or no floating point registers, do not
4104 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4105 in GPRs that occurs after we tested for GPRs that would never be
4108 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4109 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4110 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4111 specifically allow DDmode, since that does not use the SPE SIMD
4114 2014-03-21 Mark Wielaard <mjw@redhat.com>
4116 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4117 as unsigned or int depending on type and value used.
4119 2014-04-15 Richard Biener <rguenther@suse.de>
4121 PR rtl-optimization/56965
4122 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4123 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4125 * alias.c (true_dependence_1): Do not call
4126 nonoverlapping_component_refs_p.
4127 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4128 nonoverlapping_component_refs_p.
4129 (indirect_refs_may_alias_p): Likewise.
4131 2014-04-15 Teresa Johnson <tejohnson@google.com>
4133 * cfg.c (dump_bb_info): Fix flags check.
4134 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4136 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4138 PR rtl-optimization/60663
4139 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4142 2014-04-15 Richard Biener <rguenther@suse.de>
4144 * lto-streamer.h (LTO_major_version): Bump to 4.
4146 2014-04-15 Richard Biener <rguenther@suse.de>
4148 * common.opt (lto_partition_model): New enum.
4149 (flto-partition=): Merge separate options with a single with argument,
4150 add -flto-partition=one support.
4151 * flag-types.h (enum lto_partition_model): Declare.
4152 * opts.c (finish_options): Remove duplicate -flto-partition=
4154 * lto-wrapper.c (run_gcc): Adjust.
4156 2014-04-15 Richard Biener <rguenther@suse.de>
4158 * alias.c (ncr_compar): New function.
4159 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4161 2014-04-15 Richard Biener <rguenther@suse.de>
4163 * alias.c (record_component_aliases): Do not walk BINFOs.
4165 2014-04-15 Richard Biener <rguenther@suse.de>
4167 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4168 Add struct function argument and adjust.
4169 (find_func_aliases_for_call): Likewise.
4170 (find_func_aliases): Likewise.
4171 (find_func_clobbers): Likewise.
4172 (intra_create_variable_infos): Likewise.
4173 (compute_points_to_sets): Likewise.
4174 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4176 2014-04-15 Richard Biener <rguenther@suse.de>
4178 * tree.c (iterative_hash_expr): Use enum tree_code_class
4179 to store TREE_CODE_CLASS.
4180 (tree_block): Likewise.
4181 (tree_set_block): Likewise.
4182 * tree.h (fold_build_pointer_plus_loc): Use
4183 convert_to_ptrofftype_loc.
4185 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4188 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4191 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4193 * cfgloop.h (struct loop): Move force_vectorize down.
4194 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4195 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4196 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4197 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4198 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4199 * tree-core.h (enum annot_expr_kind): Add new kind values.
4200 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4201 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4203 * tree.def (ANNOTATE_EXPR): Tweak comment.
4205 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4207 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4210 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4212 * tree.h (TYPE_IDENTIFIER): Declare.
4213 * tree.c (subrange_type_for_debug_p): Use it.
4214 * godump.c (go_format_type): Likewise.
4215 * dwarf2out.c (is_cxx_auto, modified_type_die,
4216 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4217 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4219 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4222 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4224 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4226 * config/i386/i386.c (examine_argument): Return bool. Return true if
4227 parameter should be passed in memory.
4228 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4229 (construct_container): Update calls to examine_argument.
4230 (function_arg_advance_64): Ditto.
4231 (return_in_memory_32): Merge with ix86_return_in_memory.
4232 (return_in_memory_64): Ditto.
4233 (return_in_memory_ms_64): Ditto.
4235 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4237 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4238 * coverage.c (coverage_compute_profile_id): Handle externally visible
4241 2014-04-14 Martin Jambor <mjambor@suse.cz>
4243 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4244 DECL_DISREGARD_INLINE_LIMITS functions.
4246 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4249 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4251 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4254 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4255 optimize_insn_for_speed_p instead of
4256 optimize_function_for_speed_p.
4258 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4260 * doc/invoke.texi (free): Document AArch64.
4262 2014-04-14 Richard Biener <rguenther@suse.de>
4264 PR tree-optimization/60042
4265 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4266 (insert_into_preds_of_block): Do not prevent PHI insertion
4267 for REFERENCE exprs here ...
4268 (eliminate_dom_walker::before_dom_children): ... but prevent
4269 their use here under similar conditions when applied to the
4270 IL after PRE optimizations.
4272 2014-04-14 Richard Biener <rguenther@suse.de>
4274 * passes.def: Move early points-to after early SRA.
4276 2014-04-14 Richard Biener <rguenther@suse.de>
4278 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4279 check for which sign-changes we allow when forwarding
4280 a converted value into a switch.
4282 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4284 * stor-layout.c (place_field): Finalize non-constant offset for the
4287 2014-04-14 Richard Biener <rguenther@suse.de>
4289 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4291 (expand_switch_using_bit_tests_p): Likewise.
4292 (process_switch): Compute and pass on speed_p based on the
4294 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4295 optimize_bb_for_speed_p.
4297 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4299 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4300 * function.h (struct function): Rename has_force_vect_loops into
4301 has_force_vectorize_loops.
4302 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4303 (input_struct_function_base): Likewise.
4304 * lto-streamer-out.c (output_cfg): Likewise.
4305 (output_struct_function_base): Likewise.
4306 * omp-low.c (expand_omp_simd): Likewise.
4307 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4308 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4309 (version_loop_for_if_conversion): Likewise.
4310 (tree_if_conversion): Likewise.
4311 (main_tree_if_conversion): Likewise.
4312 (gate_tree_if_conversion): Likewise.
4313 * tree-inline.c (copy_loops): Likewise.
4314 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4315 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4316 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4317 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4318 * tree-vectorizer.c (vectorize_loops): Likewise.
4319 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4321 2014-04-14 Richard Biener <rguenther@suse.de>
4324 * lto-streamer-out.c (wrap_refs): New function.
4325 (lto_output): Wrap symbol references in global initializes in
4326 type-preserving MEM_REFs.
4328 2014-04-14 Christian Bruel <christian.bruel@st.com>
4330 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4332 2014-04-14 Christian Bruel <christian.bruel@st.com>
4334 * config/sh/sh.md (setmemqi): New expand pattern.
4335 * config/sh/sh.h (CLEAR_RATIO): Define.
4336 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4337 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4339 2014-04-14 Richard Biener <rguenther@suse.de>
4342 * fold-const.c (negate_expr_p): Don't negate directional rounding
4344 (fold_negate_expr): Likewise.
4346 2014-04-14 Richard Biener <rguenther@suse.de>
4348 PR tree-optimization/59817
4349 PR tree-optimization/60453
4350 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4351 recursion to catch all CHRECs in the scalar evolution and restrict
4352 the predicate for the remains appropriately.
4354 2014-04-12 Catherine Moore <clm@codesourcery.com>
4356 * config/mips/constraints.md: Add new register constraint "kb".
4357 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4358 (*movhi_internal): Likewise.
4359 (*movqi_internal): Likewise.
4360 * config/mips/mips.h (M16_STORE_REGS): New register class.
4361 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4362 (REG_CLASS_CONTENTS): Likewise.
4363 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4365 2014-04-11 Tobias Burnus <burnus@net-b.de>
4368 * doc/invoke.texi (-Wformat-signedness): Document it.
4369 (Wformat=2): Mention that this enables -Wformat-signedness.
4371 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4373 * common/config/epiphany/epiphany-common.c
4374 (epiphany_option_optimization_table): Enable section anchors by
4375 default at -O1 or higher.
4376 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4377 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4378 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4379 carries no extra cost.
4380 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4381 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4382 * config/epiphany/predicates.md (memclob_operand): New predicate.
4383 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4384 Use memclob_operand predicate and X constraint for operand 3.
4386 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4388 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4389 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4392 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4394 PR rtl-optimization/60651
4395 * mode-switching.c (optimize_mode_switching): Make sure to emit
4396 sets of a lower numbered entity before sets of a higher numbered
4397 entity to a mode of the same or lower priority.
4398 When creating a seginfo for a basic block that starts with a code
4399 label, move the insertion point past the code label.
4400 (new_seginfo): Document and enforce requirement that
4401 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4402 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4403 * doc/tm.texi: Regenerate.
4405 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4408 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4410 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4412 * BASE-VER: Set to 4.10.0.
4414 2014-04-11 Tobias Burnus <burnus@net-b.de>
4417 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4418 * doc/gcc.texi (Service): Update description in the @menu
4419 * doc/invoke.texi (Option Summary): Remove misplaced and
4422 2014-04-11 Steve Ellcey <sellcey@mips.com>
4423 Jakub Jelinek <jakub@redhat.com>
4426 * expr.c (convert_move): Use emit_store_flag_force instead of
4427 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4430 2014-04-11 Richard Biener <rguenther@suse.de>
4433 * varasm.c (assemble_alias): Avoid endless error reporting
4434 recursion by setting TREE_ASM_WRITTEN.
4436 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4438 * config/s390/s390.md: Add a splitter for NOT rtx.
4440 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4442 PR rtl-optimization/60663
4443 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4445 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4446 Jakub Jelinek <jakub@redhat.com>
4449 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4450 flag from decl_node to node.
4452 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4455 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4456 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4457 ameliorating the cases where it can be.
4459 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4462 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4464 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4465 (loadsync_<mode>): Change mode.
4466 (load_quadpti, store_quadpti): New.
4467 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4468 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4469 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4471 2014-04-09 Cong Hou <congh@google.com>
4474 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4477 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4479 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4480 instead of vnor to exploit possible fusion opportunity in the
4482 (altivec_expand_vec_perm_const_le): Likewise.
4484 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4486 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4487 (loadsync_<mode>): Change mode.
4488 (load_quadpti, store_quadpti): New.
4489 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4490 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4492 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4495 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4496 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4497 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4499 2014-04-08 Richard Biener <rguenther@suse.de>
4502 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4503 a 64bit widest int print double-int similar to on HWI64 hosts.
4505 2014-04-08 Richard Biener <rguenther@suse.de>
4507 PR tree-optimization/60785
4508 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4509 default defs properly.
4511 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4513 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4514 (Weffc++): Likewise.
4516 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4518 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4519 set completep to false rather than true.
4521 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4524 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4525 ARM_TARGET2_DWARF_FORMAT.
4527 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4530 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4531 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4534 2014-04-07 Richard Biener <rguenther@suse.de>
4536 PR tree-optimization/60766
4537 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4538 (may_eliminate_iv): Convert cand_value_at result to desired type.
4540 2014-04-07 Jason Merrill <jason@redhat.com>
4543 * common.opt (-fno-gnu-unique): Add.
4544 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4546 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4548 * haifa-sched.c: Fix outdated function reference and minor
4549 grammar errors in introductory comment.
4551 2014-04-07 Richard Biener <rguenther@suse.de>
4554 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4556 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4558 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4561 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4562 size accounting for thunks.
4563 (pa_asm_output_mi_thunk): Use final_start_function() and
4564 final_end_function() to output function start and end directives.
4566 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4568 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4569 device specific ISA/ feature information. Remove short_sp and
4570 errata_skip ds. Add avr_device_specific_features enum to have device
4572 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4573 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4574 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4575 updated device specific info.
4576 * config/avr/avr-mcus.def: Merge device specific details to
4577 dev_attribute field.
4578 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4580 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4581 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4582 assembler if RMW isa supported by current device.
4583 * config/avr/genmultilib.awk: Update as device info structure changed.
4584 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4586 2014-04-04 Cong Hou <congh@google.com>
4588 PR tree-optimization/60656
4589 * tree-vect-stmts.c (supportable_widening_operation):
4590 Fix a bug that elements in a vector with vect_used_by_reduction
4591 property are incorrectly reordered when the operation on it is not
4592 consistant with the one in reduction operation.
4594 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4596 PR rtl-optimization/60155
4597 * gcse.c (record_set_data): New function.
4598 (single_set_gcse): New function.
4599 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4600 (hoist_code): Likewise.
4601 (get_pressure_class_and_nregs): Likewise.
4603 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4605 * explow.c (probe_stack_range): Emit a final optimization blockage.
4607 2014-04-04 Anthony Green <green@moxielogic.com>
4609 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4612 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4615 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4616 flags are set only during streaming.
4617 * ipa.c (process_references, walk_polymorphic_call_targets,
4618 symtab_remove_unreachable_nodes): Drop bodies of always inline
4619 after early inlining.
4620 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4622 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4623 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4626 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4629 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4632 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4634 (cortex_a53_fdivd): Likewise.
4636 2014-04-04 Martin Jambor <mjambor@suse.cz>
4639 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4641 * cgraph.c (clone_of_p): Also return true if thunks match.
4642 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4643 cgraph_function_or_thunk_node and an obsolete comment.
4644 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4646 (build_function_decl_skip_args): Likewise.
4647 (set_new_clone_decl_and_node_flags): New function.
4648 (duplicate_thunk_for_node): Likewise.
4649 (redirect_edge_duplicating_thunks): Likewise.
4650 (cgraph_clone_node): New parameter args_to_skip, pass it to
4651 redirect_edge_duplicating_thunks which is called instead of
4652 cgraph_redirect_edge_callee.
4653 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4654 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4656 2014-04-04 Jeff Law <law@redhat.com>
4659 * config/arm/predicates.md (const_int_I_operand): New predicate.
4660 (const_int_M_operand): Similarly.
4661 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4663 (insv_t2, extv_reg, extzv_t2): Likewise.
4664 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4665 (pop_multiple_with_writeback_and_return): Likewise.
4666 (vfp_pop_multiple_with_writeback): Likewise
4668 2014-04-04 Richard Biener <rguenther@suse.de>
4671 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4672 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4674 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4675 * gimplify.c (gimple_add_tmp_var_fn): New function.
4676 * gimple-expr.h (create_tmp_reg_fn): Declare.
4677 * gimple-expr.c (create_tmp_reg_fn): New function.
4678 * gimple-low.c (record_vars_into): Don't change cfun.
4679 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4680 code generation without cfun.
4682 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4685 * Makefile.in (install-driver): Fix shell scripting.
4687 2014-04-03 Cong Hou <congh@google.com>
4689 PR tree-optimization/60505
4690 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4691 threshold of number of iterations below which no vectorization
4693 * tree-vect-loop.c (new_loop_vec_info):
4694 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4695 * tree-vect-loop.c (vect_analyze_loop_operations):
4696 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4697 * tree-vect-loop.c (vect_transform_loop):
4698 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4699 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4700 of iterations of the loop and see if we should build the epilogue.
4702 2014-04-03 Richard Biener <rguenther@suse.de>
4704 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4705 (streamer_tree_cache_create): Adjust.
4706 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4707 to allow optional nodes array.
4708 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4709 (streamer_tree_cache_append): Likewise.
4710 (streamer_tree_cache_create): Create nodes array optionally
4711 as specified by parameter.
4712 * lto-streamer-out.c (create_output_block): Avoid maintaining
4713 the node array in the writer cache.
4714 (DFS_write_tree): Remove assertion.
4715 (produce_asm_for_decls): Free the out decl state hash table early.
4716 * lto-streamer-in.c (lto_data_in_create): Adjust for
4717 streamer_tree_cache_create prototype change.
4719 2014-04-03 Richard Biener <rguenther@suse.de>
4721 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4722 set TREE_CHAIN to NULL_TREE.
4724 2014-04-03 Richard Biener <rguenther@suse.de>
4726 PR tree-optimization/60740
4727 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4728 over all GIMPLE_COND operands.
4730 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4732 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4733 (Weffc++): Remove Scott's numbering, merge lists and reference
4736 2014-04-03 Nick Clifton <nickc@redhat.com>
4738 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4741 2014-04-03 Martin Jambor <mjambor@suse.cz>
4743 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4744 mention gcc_unreachable before failing.
4745 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4748 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4751 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4752 inconsistent code and instead mark the context inconsistent.
4753 (possible_polymorphic_call_targets): For inconsistent contexts
4754 return empty complete list.
4756 2014-04-02 Anthony Green <green@moxielogic.com>
4758 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4759 (extendqisi2, extendhisi2): Define.
4760 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4761 (WCHAR_TYPE): Change to unsigned int.
4763 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4765 PR tree-optimization/60733
4766 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4767 insertion point for PHI candidates to be the end of the feeding
4768 block for the PHI argument.
4770 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4772 PR rtl-optimization/60650
4773 * lra-constraints.c (process_alt_operands): Decrease reject for
4774 earlyclobber matching.
4776 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4778 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4780 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4782 * config/spu/spu.c (pad_bb): Do not crash when the last
4783 insn is CODE_FOR_blockage.
4785 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4787 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4788 lies outside the target mode.
4790 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4793 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4794 software floating point or no floating point registers, do not
4795 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4796 in GPRs that occurs after we tested for GPRs that would never be
4799 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4800 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4801 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4802 specifically allow DDmode, since that does not use the SPE SIMD
4805 2014-04-02 Richard Biener <rguenther@suse.de>
4808 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4809 MODE_INTs. Properly use negv_optab.
4810 (expand_abs): Likewise.
4812 2014-04-02 Richard Biener <rguenther@suse.de>
4815 * Makefile.in (install-driver): Guard extra installs with special
4818 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4820 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4823 2014-04-01 Richard Henderson <rth@redhat.com>
4826 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4827 alternative enabled before register allocation.
4829 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4831 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4832 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4834 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4835 (nios2_got_address): Update nios2_large_got_address call site.
4836 (nios2_delegitimize_address): New function.
4837 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4838 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4839 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4841 2014-04-01 Martin Husemann <martin@duskware.de>
4843 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4846 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4848 PR rtl-optimization/60604
4849 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4850 check from register_operand.
4851 (register_operand): Redefine in terms of general_operand.
4852 (nonmemory_operand): Use register_operand for the non-constant cases.
4854 2014-04-01 Richard Biener <rguenther@suse.de>
4856 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4858 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4860 * doc/invoke.texi (mapp-regs): Clarify.
4862 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4864 * config/i386/avx512fintrin.h (__v32hi): Define type.
4865 (__v64qi): Likewise.
4866 (_mm512_set1_epi8): Define.
4867 (_mm512_set1_epi16): Define.
4868 (_mm512_set4_epi32): Define.
4869 (_mm512_set4_epi64): Define.
4870 (_mm512_set4_pd): Define.
4871 (_mm512_set4_ps): Define.
4872 (_mm512_setr4_epi64): Define.
4873 (_mm512_setr4_epi32): Define.
4874 (_mm512_setr4_pd): Define.
4875 (_mm512_setr4_ps): Define.
4876 (_mm512_setzero_epi32): Define.
4878 2014-03-31 Martin Jambor <mjambor@suse.cz>
4881 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4882 callsite_arguments_match_p. Updated all callers. Also check types of
4883 corresponding formal parameters and actual arguments.
4884 (not_all_callers_have_enough_arguments_p) Renamed to
4885 some_callers_have_mismatched_arguments_p.
4887 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4889 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4891 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4894 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4897 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4899 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4900 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4902 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4903 Use FMAMODE_NOVF512 mode iterator.
4904 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4905 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4906 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4907 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4909 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4910 Use VF_128_256 mode iterator.
4911 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4914 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4916 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4917 static chain if needed.
4919 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4922 * lra-constraints.c (index_part_to_reg): New.
4923 (process_address): Use it.
4925 2014-03-27 Jeff Law <law@redhat.com>
4926 Jakub Jelinek <jakub@redhat.com>
4929 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4930 gen_rtx_{PLUS,MULT} to build up the address expression.
4932 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4933 creating non-canonical RTL.
4935 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4938 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4939 functions; reorganize to make cheap checks first.
4940 (inline_small_functions): Do not estimate growth when dumping;
4942 * ipa-inline.h (inline_summary): Add min_size.
4943 (growth_likely_positive): New function.
4944 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4945 (set_cond_stmt_execution_predicate): Cleanup.
4946 (estimate_edge_size_and_time): Compute min_size.
4947 (estimate_calls_size_and_time): Likewise.
4948 (estimate_node_size_and_time): Likewise.
4949 (inline_update_overall_summary): Update min_size.
4950 (do_estimate_edge_time): Likewise.
4951 (do_estimate_edge_size): Update.
4952 (do_estimate_edge_hints): Update.
4953 (growth_likely_positive): New function.
4955 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4958 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4959 also if addr has VOIDmode.
4961 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4963 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4964 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4966 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4967 instructions as well as AdvancedSIMD loads.
4969 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4971 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4972 Use crypto_aese type.
4973 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4974 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4975 crypto_aese, crypto_aesmc. Move to types.md.
4976 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4978 * config/arm/iterators.md (crypto_type): Likewise.
4980 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4982 * cgraph.c: Include expr.h and tree-dfa.h.
4983 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4986 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4989 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4990 regs from checking multi-reg pseudos.
4992 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4994 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4996 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4998 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4999 if it would clobber the stack pointer, even temporarily.
5001 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5003 * mode-switching.c: Make small adjustments to the top comment.
5005 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5007 * config/rs6000/constraints.md (wD constraint): New constraint to
5008 match the constant integer to get the top DImode/DFmode out of a
5009 vector in a VSX register.
5011 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5012 match the constant integer to get the top DImode/DFmode out of a
5013 vector in a VSX register.
5015 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5018 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5021 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5022 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5024 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5025 Optimize vec_extract of 64-bit values, where the value being
5026 extracted is in the top word, where we can use scalar
5027 instructions. Add direct move and store support. Combine the big
5028 endian/little endian vector select load support into a single insn.
5029 (vsx_extract_<mode>_internal1): Likewise.
5030 (vsx_extract_<mode>_internal2): Likewise.
5031 (vsx_extract_<mode>_load): Likewise.
5032 (vsx_extract_<mode>_store): Likewise.
5033 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5034 combined into vsx_extract_<mode>_load.
5035 (vsx_extract_<mode>_one_le): Likewise.
5037 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5038 define the top 64-bit vector element.
5040 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5043 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5044 Document vec_vbpermq builtin.
5047 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5048 enable use of xxsldwi and xxpermdi builtin functions.
5049 (vec_xxpermdi): Likewise.
5051 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5052 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5054 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5056 PR rtl-optimization/60650
5057 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5059 (find_spills_for): New.
5060 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5061 Spill all pseudos on the second iteration.
5063 2014-03-27 Marek Polacek <polacek@redhat.com>
5066 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5069 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5071 * config/s390/s390.c (s390_can_use_return_insn): Check for
5072 call-saved FPRs on 31 bit.
5074 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5077 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5078 if they need regimplification, just drop them instead of
5079 calling gimple_regimplify_operands on them.
5081 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5084 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5085 (aarch64_frame_pointer_required): Adjust logic.
5086 (aarch64_can_eliminate): Adjust logic.
5087 (aarch64_override_options_after_change): Adjust logic.
5089 2014-03-27 Dehao Chen <dehao@google.com>
5091 * ipa-inline.c (early_inliner): Update node's inline info.
5093 2014-03-26 Dehao Chen <dehao@google.com>
5095 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5096 compiler inserted conditional jumps for NAN float check.
5098 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5100 * ubsan.h (ubsan_create_data): Change second argument's type
5101 to const location_t *.
5102 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5104 (ubsan_create_data): Change second argument to const location_t *PLOC.
5105 Create Loc field whenever PLOC is non-NULL.
5106 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5107 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5111 * real.c (real_to_integer2): Change type of low to UHWI.
5113 2014-03-26 Tobias Burnus <burnus@net-b.de>
5115 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5116 (CILK_SELF_SPECS): New define.
5117 (driver_self_specs): Use it.
5119 2014-03-26 Richard Biener <rguenther@suse.de>
5121 * tree-pretty-print.c (percent_K_format): Implement special
5122 case for LTO and its stripped down BLOCK tree.
5124 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5127 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5129 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5130 one range is range_int_cst_p, but not both, at least optimize
5131 addition/subtraction of 0 and multiplication by 0 or 1.
5132 * gimple-fold.c (gimple_fold_call): Fold
5133 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5134 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5135 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5137 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5139 PR rtl-optimization/60452
5140 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5141 <case REG>: Return 1 for invalid offsets from the frame pointer.
5143 2014-03-26 Marek Polacek <polacek@redhat.com>
5146 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5149 2014-03-26 Marek Polacek <polacek@redhat.com>
5152 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5155 2014-03-26 Marek Polacek <polacek@redhat.com>
5158 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5159 multiplication in unsigned type.
5161 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5163 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5165 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5167 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5169 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5172 * cif-code.def (UNREACHABLE) New code.
5173 * ipa-inline.c (inline_small_functions): Skip edges to
5174 __builtlin_unreachable.
5175 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5176 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5177 predicate to __bulitin_unreachable.
5178 (set_cond_stmt_execution_predicate): Fix issue when
5179 invert_tree_comparison returns ERROR_MARK.
5180 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5181 propagate to inline clones.
5182 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5184 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5185 * cgraphclones.c (cgraph_clone_node): If call destination is already
5186 ureachable, do not redirect it back.
5187 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5190 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5192 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5193 Do not modify inline clones.
5195 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5197 * config/i386/i386.md (general_sext_operand): New mode attr.
5198 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5199 don't generate (sign_extend (const_int)).
5200 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5201 operands[2]. Use We constraint instead of <i> and
5202 <general_sext_operand> predicate instead of <general_operand>.
5203 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5204 * config/i386/constraints.md (We): New constraint.
5205 * config/i386/predicates.md (x86_64_sext_operand,
5206 sext_operand): New predicates.
5208 2014-03-25 Martin Jambor <mjambor@suse.cz>
5211 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5212 inconsistent devirtualizations to __builtin_unreachable.
5214 2014-03-25 Marek Polacek <polacek@redhat.com>
5217 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5219 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5221 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5222 order of elements for big-endian.
5224 2014-03-25 Richard Biener <rguenther@suse.de>
5227 * gimplify-me.c (gimple_regimplify_operands): Update the
5230 2014-03-25 Martin Jambor <mjambor@suse.cz>
5233 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5234 (lto_output_varpool_node): Likewise.
5235 (input_overwrite_node): Likewise.
5236 (input_varpool_node): Likewise.
5238 2014-03-25 Richard Biener <rguenther@suse.de>
5240 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5241 (run_gcc): Likewise.
5243 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5245 * combine.c (simplify_compare_const): Add MODE argument.
5246 Handle mode_width 0 as very large mode_width.
5247 (try_combine, simplify_comparison): Adjust callers.
5249 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5250 type to avoid signed integer overflow.
5251 * explow.c (plus_constant): Likewise.
5253 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5255 * doc/generic.texi: Correct typos.
5257 2014-03-24 Tobias Burnus <burnus@net-b.de>
5259 * doc/invoke.texi (-flto): Expand section about
5260 using static libraries with LTO.
5262 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5264 PR rtl-optimization/60501
5265 * optabs.def (addptr3_optab): New optab.
5266 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5267 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5268 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5270 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5272 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5274 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5276 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5279 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5280 (_mm256_undefined_ps): Define.
5281 (_mm256_undefined_pd): Define.
5282 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5283 (_mm_undefined_pd): Define.
5284 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5285 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5286 (_mm512_undefined_ps): Define.
5287 (_mm512_undefined_pd): Define.
5288 Use _mm*_undefined_*.
5289 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5291 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5293 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5294 (lshr_simd): DI mode added.
5295 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5296 (aarch64_ushr_simddi): Likewise.
5297 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5298 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5299 (vshrd_n_u64): Likewise.
5301 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5303 * Makefile.in (s-macro_list): Depend on cc1.
5305 2014-03-23 Teresa Johnson <tejohnson@google.com>
5307 * ipa-utils.c (ipa_print_order): Use specified dump file.
5309 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5311 PR rtl-optimization/60601
5312 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5314 * gcc.c (eval_spec_function): Initialize save_growing_value.
5316 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5319 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5320 code == MINUS_EXPR, never swap op0 with op1.
5322 * toplev.c (init_local_tick): Avoid signed integer multiplication
5324 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5325 shift by first operand's bitsize.
5327 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5330 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5332 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5333 TARGET_ISA_64BIT_P(x).
5335 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5337 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5338 pattern for vector nor instead of subtract from splat(-1).
5339 (altivec_expand_vec_perm_const_le): Likewise.
5341 2014-03-21 Richard Henderson <rth@twiddle.net>
5344 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5345 related insns after epilogue_completed.
5347 2014-03-21 Martin Jambor <mjambor@suse.cz>
5350 * cgraph.h (symtab_node): New flag body_removed.
5351 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5352 when removing bodies.
5353 * symtab.c (dump_symtab_base): Dump body_removed flag.
5354 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5355 had their bodies removed.
5357 2014-03-21 Martin Jambor <mjambor@suse.cz>
5360 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5363 2014-03-21 Richard Biener <rguenther@suse.de>
5365 PR tree-optimization/60577
5366 * tree-core.h (struct tree_base): Document nothrow_flag use
5367 in VAR_DECL_NONALIASED.
5368 * tree.h (VAR_DECL_NONALIASED): New.
5369 (may_be_aliased): Adjust.
5370 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5372 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5374 * expr.c (expand_expr_real_1): Remove outdated comment.
5376 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5379 * ira.c (adjust_cleared_regs): Call copy_rtx on
5380 *reg_equiv[REGNO (loc)].src_p before passing it to
5381 simplify_replace_fn_rtx.
5384 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5385 into CONST, put pic register as first operand of PLUS. Use
5386 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5388 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5390 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5392 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5394 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5395 around for store forwarding issue in the FPU on the UT699.
5396 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5397 loads and operations if -mfix-ut699 is specified.
5398 (divtf3_hq): Tweak attribute.
5399 (sqrttf2_hq): Likewise.
5401 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5403 * calls.c (store_one_arg): Remove incorrect const qualification on the
5404 type of the temporary.
5405 * cfgexpand.c (expand_return): Likewise.
5406 * expr.c (expand_constructor): Likewise.
5407 (expand_expr_real_1): Likewise.
5409 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5411 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5414 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5417 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5419 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5421 * config/arm/aarch-common-protos.h
5422 (alu_cost_table): Fix spelling of "extend".
5423 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5425 2014-03-19 Richard Biener <rguenther@suse.de>
5428 * tree-core.h (tree_type_common): Re-order pointer members
5429 to reduce recursion depth during GC walks.
5431 2014-03-19 Marek Polacek <polacek@redhat.com>
5434 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5435 before accessing it.
5437 2014-03-19 Richard Biener <rguenther@suse.de>
5440 * lto-streamer-in.c (input_function): In WPA stage do not drop
5443 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5445 PR tree-optimization/60559
5446 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5447 with build_zero_cst assignment.
5449 2014-03-18 Kai Tietz <ktietz@redhat.com>
5451 PR rtl-optimization/56356
5452 * sdbout.c (sdbout_parms): Verify that parms'
5453 incoming argument is valid.
5454 (sdbout_reg_parms): Likewise.
5456 2014-03-18 Richard Henderson <rth@redhat.com>
5459 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5460 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5461 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5463 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5465 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5466 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5467 Italicize plugin event names in description. Explain that
5468 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5469 Remind that no GCC functions should be called after PLUGIN_FINISH.
5470 Explain what pragmas with expansion are.
5472 2014-03-18 Martin Liska <mliska@suse.cz>
5474 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5475 gimple call statement is update.
5476 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5477 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5479 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5482 * ubsan.c (ubsan_instrument_unreachable): Call
5483 initialize_sanitizer_builtins.
5484 (ubsan_pass): Likewise.
5487 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5488 varpool_finalize_decl instead of rest_of_decl_compilation.
5490 2014-03-18 Richard Biener <rguenther@suse.de>
5492 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5493 by using bitmap_and_compl instead of bitmap_and_compl_into.
5494 (df_rd_transfer_function): Likewise.
5496 2014-03-18 Richard Biener <rguenther@suse.de>
5498 * doc/lto.texi (fresolution): Fix typo.
5500 2014-03-18 Richard Biener <rguenther@suse.de>
5502 * doc/invoke.texi (flto): Update for changes in 4.9.
5504 2014-03-18 Richard Biener <rguenther@suse.de>
5506 * doc/loop.texi: Remove section on the removed lambda framework.
5507 Update loop docs with recent changes in preserving loop structure.
5509 2014-03-18 Richard Biener <rguenther@suse.de>
5511 * doc/lto.texi (-fresolution): Document.
5513 2014-03-18 Richard Biener <rguenther@suse.de>
5515 * doc/contrib.texi: Adjust my name.
5517 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5520 * internal-fn.c: Include diagnostic-core.h.
5521 (expand_BUILTIN_EXPECT): New function.
5522 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5523 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5524 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5525 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5527 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5528 Revert 3 argument __builtin_expect code.
5529 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5530 * gimple-fold.c (gimple_fold_call): Likewise.
5531 * tree.h (fold_builtin_expect): New prototype.
5532 * builtins.c (build_builtin_expect_predicate): Add predictor
5533 argument, if non-NULL, create 3 argument __builtin_expect.
5534 (fold_builtin_expect): No longer static. Add ARG2 argument,
5535 pass it through to build_builtin_expect_predicate.
5536 (fold_builtin_2): Adjust caller.
5537 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5538 * internal-fn.def (BUILTIN_EXPECT): New.
5540 2014-03-18 Tobias Burnus <burnus@net-b.de>
5543 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5544 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5545 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5547 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5550 * predict.c (combine_predictions_for_bb): Fix up formatting.
5551 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5552 fill what it points to if non-NULL.
5553 (tree_predict_by_opcode): Adjust caller, use the predictor.
5554 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5556 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5558 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5559 proper constant for the store mode.
5561 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5563 * symtab.c (change_decl_assembler_name): Fix transparent alias
5566 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5568 * config/aarch64/aarch64.c: Correct the comments about the
5569 aarch64 stack layout.
5571 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5573 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5574 check for GF_OMP_FOR_KIND_FOR.
5576 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5578 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5579 ymm and zmm register names.
5581 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5584 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5585 note creation for the 2010-08-31 changes.
5587 2014-03-17 Marek Polacek <polacek@redhat.com>
5590 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5591 as -fno-tree-loop-vectorize.
5592 (expand_omp_simd): Likewise.
5594 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5596 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5597 (eligible_for_call_delay): New prototype.
5598 * config/sparc/sparc.c (tls_call_delay): Rename into...
5599 (eligible_for_call_delay): ...this. Return false if the instruction
5600 cannot be put in the delay slot of a branch.
5601 (eligible_for_restore_insn): Simplify.
5602 (eligible_for_return_delay): Return false if the instruction cannot be
5603 put in the delay slot of a branch and simplify.
5604 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5605 put in the delay slot of a branch.
5606 * config/sparc/sparc.md (fix_ut699): New attribute.
5607 (tls_call_delay): Delete.
5608 (in_call_delay): Reimplement.
5609 (eligible_for_sibcall_delay): Rename into...
5610 (in_sibcall_delay): ...this.
5611 (eligible_for_return_delay): Rename into...
5612 (in_return_delay): ...this.
5613 (in_branch_delay): Reimplement.
5614 (in_uncond_branch_delay): Delete.
5615 (in_annul_branch_delay): Delete.
5617 2014-03-14 Richard Henderson <rth@redhat.com>
5620 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5621 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5622 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5623 (floathi splitters): Remove.
5624 (float<SWI48x>xf2): New pattern.
5625 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5626 code that tried to handle DImode for 32-bit, but which was excluded
5627 by the pattern's condition. Drop allocation of stack temporary.
5628 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5629 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5630 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5631 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5632 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5633 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5634 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5635 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5636 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5637 (*float<SWI48x><X87MODEF>2_i387): Remove.
5638 (all float _with_temp splitters): Remove.
5639 (*float<SWI48x><MODEF>2_i387): New pattern.
5640 (*float<SWI48><MODEF>2_sse): New pattern.
5641 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5642 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5644 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5645 Marek Polacek <polacek@redhat.com>
5648 * common.opt (dump_base_name_prefixed): New Variable.
5649 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5650 if x_dump_base_name_prefixed is already set, set it at the end.
5652 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5654 PR rtl-optimization/60508
5655 * lra-constraints.c (get_reload_reg): Add new parameter
5657 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5658 Pass the new parameter values.
5660 2014-03-14 Richard Biener <rguenther@suse.de>
5662 * common.opt: Revert unintented changes from r205065.
5665 2014-03-14 Richard Biener <rguenther@suse.de>
5668 * cfghooks.c (split_block): Properly adjust all loops the
5669 block was a latch of.
5671 2014-03-14 Martin Jambor <mjambor@suse.cz>
5674 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5677 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5680 * config/avr/avr.c (avr_set_current_function): Pass function name
5681 through default_strip_name_encoding before sanity checking instead
5682 of skipping the first char of the assembler name.
5684 2014-03-13 Richard Henderson <rth@redhat.com>
5687 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5688 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5689 * config/i386/i386-protos.h: Likewise.
5690 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5691 in the expander instead of a splitter.
5692 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5693 any possibility of requiring a memory.
5694 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5695 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5696 (fp branch splitters): Update for ix86_split_fp_branch.
5697 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5698 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5699 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5700 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5701 (*fop_<MODEF>_3_i387): Likewise.
5702 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5703 (splitters for the fop_* register patterns): Remove.
5704 (fscalexf4_i387): Rename from *fscalexf4_i387.
5705 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5707 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5709 PR tree-optimization/59779
5710 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5711 type for bitsize and maxsize instead of HOST_WIDE_INT.
5713 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5715 PR rtl-optimization/57320
5716 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5717 the CFG after thread_prologue_and_epilogue_insns.
5719 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5721 PR rtl-optimization/57189
5722 * lra-constraints.c (process_alt_operands): Disfavor spilling
5725 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5727 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5729 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5731 PR tree-optimization/59025
5733 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5734 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5736 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5739 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5740 calls of avr_out_plus_1.
5742 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5744 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5745 BB's single pred and update the father loop's latch info later.
5747 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5749 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5753 (VEC_base): Likewise.
5754 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5755 registers, we need to swap double words in little endian mode.
5757 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5758 to be a container mode for 128-bit integer operations added in ISA
5759 2.07. Unlike TImode and PTImode, the preferred register set is
5760 the Altivec/VMX registers for the 128-bit operations.
5762 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5764 (rs6000_split_128bit_ok_p): Likewise.
5766 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5767 macros for creating ISA 2.07 normal and overloaded builtin
5768 functions with 3 arguments.
5769 (BU_P8V_OVERLOAD_3): Likewise.
5770 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5771 for use as overloaded functions.
5772 (VPERM_1TI_UNS): Likewise.
5773 (VSEL_1TI): Likewise.
5774 (VSEL_1TI_UNS): Likewise.
5775 (ST_INTERNAL_1ti): Likewise.
5776 (LD_INTERNAL_1ti): Likewise.
5777 (XXSEL_1TI): Likewise.
5778 (XXSEL_1TI_UNS): Likewise.
5779 (VPERM_1TI): Likewise.
5780 (VPERM_1TI_UNS): Likewise.
5781 (XXPERMDI_1TI): Likewise.
5782 (SET_1TI): Likewise.
5783 (LXVD2X_V1TI): Likewise.
5784 (STXVD2X_V1TI): Likewise.
5785 (VEC_INIT_V1TI): Likewise.
5786 (VEC_SET_V1TI): Likewise.
5787 (VEC_EXT_V1TI): Likewise.
5788 (EQV_V1TI): Likewise.
5789 (NAND_V1TI): Likewise.
5790 (ORC_V1TI): Likewise.
5791 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5792 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5794 (VADDUQM): Likewise.
5795 (VSUBCUQ): Likewise.
5796 (VADDEUQM): Likewise.
5797 (VADDECUQ): Likewise.
5798 (VSUBEUQM): Likewise.
5799 (VSUBECUQ): Likewise.
5801 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5802 __int128_t and __uint128_t types.
5803 (__uint128_type): Likewise.
5804 (altivec_categorize_keyword): Add support for vector __int128_t,
5805 vector __uint128_t, vector __int128, and vector unsigned __int128
5806 as a container type for TImode operations that need to be done in
5807 VSX/Altivec registers.
5808 (rs6000_macro_to_expand): Likewise.
5809 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5810 to support 128-bit integer instructions vaddcuq, vadduqm,
5811 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5812 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5814 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5815 for V1TImode, and set up preferences to use VSX/Altivec registers.
5816 Setup VSX reload handlers.
5817 (rs6000_debug_reg_global): Likewise.
5818 (rs6000_init_hard_regno_mode_ok): Likewise.
5819 (rs6000_preferred_simd_mode): Likewise.
5820 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5821 (easy_altivec_constant): Likewise.
5822 (output_vec_const_move): Likewise.
5823 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5825 (rs6000_expand_vector_extract): Likewise.
5826 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5828 (rs6000_const_vec): Add support for V1TImode.
5829 (rs6000_emit_le_vsx_load): Swap double words when loading or
5830 storing TImode/V1TImode.
5831 (rs6000_emit_le_vsx_store): Likewise.
5832 (rs6000_emit_le_vsx_move): Likewise.
5833 (rs6000_emit_move): Add support for V1TImode.
5834 (altivec_expand_ld_builtin): Likewise.
5835 (altivec_expand_st_builtin): Likewise.
5836 (altivec_expand_vec_init_builtin): Likewise.
5837 (altivec_expand_builtin): Likewise.
5838 (rs6000_init_builtins): Add support for V1TImode type. Add
5839 support for ISA 2.07 128-bit integer builtins. Define type names
5840 for the VSX/Altivec vector types.
5841 (altivec_init_builtins): Add support for overloaded vector
5842 functions with V1TImode type.
5843 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5844 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5846 (rs6000_split_128bit_ok_p): Likewise.
5847 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5848 __int128_t and vector __uint128_t.
5850 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5851 and mode attributes.
5858 (VS_scalar): Likewise.
5859 (VS_double): Likewise.
5860 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5862 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5863 we support the ISA 2.07 128-bit integer arithmetic instructions.
5864 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5865 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5866 and TImode types for use with the builtin functions.
5867 (V1TI_type_node): Likewise.
5868 (unsigned_V1TI_type_node): Likewise.
5869 (intTI_type_internal_node): Likewise.
5870 (uintTI_type_internal_node): Likewise.
5872 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5873 128-bit builtin functions.
5874 (UNSPEC_VADDEUQM): Likewise.
5875 (UNSPEC_VADDECUQ): Likewise.
5876 (UNSPEC_VSUBCUQ): Likewise.
5877 (UNSPEC_VSUBEUQM): Likewise.
5878 (UNSPEC_VSUBECUQ): Likewise.
5879 (VM): Add V1TImode to vector mode iterators.
5881 (VI_unit): Likewise.
5882 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5883 (altivec_vaddcuq): Likewise.
5884 (altivec_vsubuqm): Likewise.
5885 (altivec_vsubcuq): Likewise.
5886 (altivec_vaddeuqm): Likewise.
5887 (altivec_vaddecuq): Likewise.
5888 (altivec_vsubeuqm): Likewise.
5889 (altivec_vsubecuq): Likewise.
5891 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5893 (BOOL_128): Likewise.
5894 (BOOL_REGS_OUTPUT): Likewise.
5895 (BOOL_REGS_OP1): Likewise.
5896 (BOOL_REGS_OP2): Likewise.
5897 (BOOL_REGS_UNARY): Likewise.
5898 (BOOL_REGS_AND_CR0): Likewise.
5900 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5901 128-bit integer builtin support.
5902 (vec_vadduqm): Likewise.
5903 (vec_vaddecuq): Likewise.
5904 (vec_vaddeuqm): Likewise.
5905 (vec_vsubecuq): Likewise.
5906 (vec_vsubeuqm): Likewise.
5907 (vec_vsubcuq): Likewise.
5908 (vec_vsubuqm): Likewise.
5910 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5911 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5912 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5913 128-bit integer add/subtract to ISA 2.07.
5915 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5917 * config/arc/arc.c (arc_predicate_delay_insns):
5918 Fix third argument passed to conditionalize_nonjump.
5920 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5922 * config/aarch64/aarch64-builtins.c
5923 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5924 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5925 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5926 instead of __builtin_lfloor.
5927 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5929 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5931 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5932 (tree_ssa_ifcombine_bb_1): New function.
5933 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5934 is an empty forwarder block to then_bb or vice versa and then_bb
5935 and else_bb are effectively swapped.
5937 2014-03-12 Christian Bruel <christian.bruel@st.com>
5940 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5941 REG_CFA_DEF_CFA note.
5942 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5943 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5945 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5947 PR tree-optimization/60454
5948 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5950 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5952 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5953 Do not define target_cpu_default2 to generic.
5954 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5955 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5956 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5958 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5959 Marc Glisse <marc.glisse@inria.fr>
5961 PR tree-optimization/60502
5962 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5963 instead of build_low_bits_mask.
5965 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5968 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5969 if there are multiple uses, but op doesn't live on E edge.
5970 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5971 clobber stmts before __builtin_unreachable.
5973 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5975 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5976 hard_frame_pointer_rtx.
5977 * cse.c (cse_insn): Remove volatile check.
5978 * cselib.c (cselib_process_insn): Likewise.
5979 * dse.c (scan_insn): Likewise.
5981 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5983 * config/arc/arc.c (conditionalize_nonjump): New function,
5985 (arc_ifcvt): ... this.
5986 (arc_predicate_delay_insns): Use it.
5988 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5990 * config/arc/predicates.md (extend_operand): During/after reload,
5991 allow const_int_operand.
5992 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5993 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5994 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5996 (umulsi3_highpart_i): Likewise.
5998 2014-03-11 Richard Biener <rguenther@suse.de>
6000 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6001 Add asserts to guard possible wrong-code bugs.
6003 2014-03-11 Richard Biener <rguenther@suse.de>
6005 PR tree-optimization/60429
6006 PR tree-optimization/60485
6007 * tree-ssa-structalias.c (set_union_with_increment): Properly
6008 take into account all fields that overlap the shifted vars.
6009 (do_sd_constraint): Likewise.
6010 (do_ds_constraint): Likewise.
6011 (get_constraint_for_ptr_offset): Likewise.
6013 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6015 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6016 (nios2_compute_frame_layout):
6017 Add calculation of cfun->machine->fp_save_offset.
6018 (nios2_expand_prologue): Correct setting of frame pointer register
6020 (nios2_expand_epilogue): Update recovery of stack pointer from
6021 frame pointer accordingly.
6022 (nios2_initial_elimination_offset): Update calculation of offset
6023 for eliminating to HARD_FRAME_POINTER_REGNUM.
6025 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6028 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6029 cgraph_get_create_node on VAR_DECLs.
6031 2014-03-10 Richard Biener <rguenther@suse.de>
6034 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6036 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6038 * config/vms/vms.opt (vms_float_format): New variable.
6040 2014-03-08 Tobias Burnus <burnus@net-b.de>
6042 * doc/invoke.texi (-fcilkplus): Update implementation status.
6044 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6045 Richard Biener <rguenther@suse.de>
6047 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6048 consistently accross all TUs.
6049 (run_gcc): Enable -fshort-double automatically at link at link-time
6050 and disallow override.
6052 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6055 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6056 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6057 if they can't be used.
6059 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6061 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6062 for Solaris 11/x86 ld.
6063 * configure: Regenerate.
6065 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6067 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6068 (LIB_TLS_SPEC): Save as ld_tls_libs.
6069 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6070 (HAVE_AS_IX86_TLSLDM): New test.
6071 * configure, config.in: Regenerate.
6072 * config/i386/i386.c (legitimize_tls_address): Fall back to
6073 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6074 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6075 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6076 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6078 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6080 * common.opt (fira-loop-pressure): Mark as optimization.
6082 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6084 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6085 an OpenMP mappable type.
6087 2014-03-06 Matthias Klose <doko@ubuntu.com>
6089 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6090 MULTILIB_OSDIRNAMES is not defined.
6092 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6093 Meador Inge <meadori@codesourcery.com>
6096 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6097 (arm_legitimize_address): Call legitimize_tls_address for any
6098 arm_tls_referenced_p expression, handle constant addend. Call it
6099 before testing for !TARGET_ARM.
6100 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6102 2014-03-06 Richard Biener <rguenther@suse.de>
6108 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6110 * tree-streamer.c (record_common_node): Assert we don't record
6111 nodes with type double.
6112 (preload_common_node): Skip type double, complex double and double
6113 pointer since it is now frontend dependent due to fshort-double option.
6115 2014-03-06 Richard Biener <rguenther@suse.de>
6117 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6118 or -fno-lto is specified and the linker has full plugin support.
6119 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6120 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6121 * lto-wrapper.c (merge_and_complain): Merge compile-time
6122 optimization levels.
6123 (run_gcc): And pass it through to the link options.
6125 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6129 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6131 * cselib.c (remove_useless_values): Skip to avoid quadratic
6132 behavior if the condition moved from...
6133 (cselib_process_insn): ... here holds.
6135 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6138 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6139 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6142 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6143 (TM_H): Add x86-tune.def.
6145 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6147 * config/aarch64/aarch64.c (generic_tunings):
6148 Use cortexa57_extra_costs.
6150 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6153 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6154 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6155 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6158 2014-03-04 Heiher <r@hev.cc>
6160 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6161 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6163 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6165 * config/i386/predicates.md (const2356_operand): Change to ...
6166 (const2367_operand): ... this.
6167 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6169 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6170 (*avx512pf_scatterpf<mode>sf): Ditto.
6171 (avx512pf_scatterpf<mode>df): Ditto.
6172 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6173 (*avx512pf_scatterpf<mode>df): Ditto.
6174 * config/i386/i386.c (ix86_expand_builtin): Update
6175 incorrect hint operand error message.
6177 2014-03-04 Richard Biener <rguenther@suse.de>
6179 * lto-section-in.c (lto_get_section_data): Fix const cast.
6181 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6183 * tree-streamer.c (record_common_node): Assert we don't record
6184 nodes with type double.
6185 (preload_common_node): Skip type double, complex double and double
6186 pointer since it is now frontend dependent due to fshort-double option.
6188 2014-03-04 Richard Biener <rguenther@suse.de>
6191 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6192 (lto_input_toplevel_asms): Likewise.
6193 * lto-section-in.c (lto_get_section_data): Instead do it here
6196 2014-03-04 Richard Biener <rguenther@suse.de>
6198 PR tree-optimization/60382
6199 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6200 dead PHIs a reduction.
6202 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6204 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6206 (_mm_prefetch): Move out of GCC target("sse") pragma.
6207 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6208 GCC target("prfchw") pragma.
6209 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6211 * config/i386/i386.c (ix86_option_override_internal): Enable
6212 -mprfchw with -mprefetchwt1.
6214 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6216 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6219 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6221 * opts.h (CL_PCH_IGNORE): Define.
6222 * targhooks.c (option_affects_pch_p):
6223 Return false for options that have CL_PCH_IGNORE set.
6224 * opt-functions.awk: Process PchIgnore.
6225 * doc/options.texi: Document PchIgnore.
6227 * config/arc/arc.opt (misize): Add PchIgnore property.
6229 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6231 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6232 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6233 constraint on constants to permit them being loaded into
6234 GENERAL_REGS or BASE_REGS.
6236 2014-03-03 Nick Clifton <nickc@redhat.com>
6238 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6239 anti-cacnonical alternatives.
6240 (negandhi3_real): New pattern.
6241 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6243 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6245 * config/avr/avr-mcus.def: Remove atxmega16x1.
6246 * config/avr/avr-tables.opt: Regenerate.
6247 * config/avr/t-multilib: Regenerate.
6248 * doc/avr-mmcu.texi: Regenerate.
6250 2014-03-03 Tobias Grosser <tobias@grosser.es>
6251 Mircea Namolaru <mircea.namolaru@inria.fr>
6253 PR tree-optimization/58028
6254 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6257 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6259 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6260 not handled by recognizers.
6262 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6265 * function.c (expand_function_end): Don't emit
6266 clobber_return_register sequence if clobber_after is a BARRIER.
6267 * cfgexpand.c (construct_exit_block): Append instructions before
6268 return_label to prev_bb.
6270 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6272 * config/rs6000/constraints.md: Document reserved use of "wc".
6274 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6277 * ipa.c (function_and_variable_visibility): When dissolving comdat
6278 group, also set all symbols to local.
6280 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6285 2013-12-14 Jan Hubicka <jh@suse.cz>
6287 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6289 2014-03-02 Jon Beniston <jon@beniston.com>
6295 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6296 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6297 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6298 (simple_return, *simple_return): New patterns
6299 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6300 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6302 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6304 * dwarf2out.c (gen_subprogram_die): Tidy.
6306 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6309 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6310 (*mov_t_msb_neg_negc): ... this new insn.
6312 2014-02-28 Jason Merrill <jason@redhat.com>
6315 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6318 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6321 * dwarf2out.c (decltype_auto_die): New static.
6322 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6323 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6324 (is_cxx_auto): Likewise.
6326 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6328 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6329 we are not using general regs only.
6331 2014-02-28 Richard Biener <rguenther@suse.de>
6334 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6335 previous fix and only allow to remove trivial pre-headers
6336 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6337 (remove_forwarder_block): Properly update the latch of a loop.
6339 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6342 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6343 (cselib_preserved_hash_table): New.
6344 (preserve_constants_and_equivs): Move preserved vals to it.
6345 (cselib_find_slot): Look it up first.
6346 (cselib_init): Initialize it.
6347 (cselib_finish): Release it.
6348 (dump_cselib_table): Dump it.
6350 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6353 * cselib.c (remove_useless_values): Skip to avoid quadratic
6354 behavior if the condition moved from...
6355 (cselib_process_insn): ... here holds.
6357 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6360 * var-tracking.c (vt_initialize): Apply the same condition to
6361 preserve the CFA base value.
6363 2014-02-28 Joey Ye <joey.ye@arm.com>
6366 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6367 if reload in progress or completed.
6369 2014-02-28 Tobias Burnus <burnus@net-b.de>
6372 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6375 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6377 * doc/tm.texi.in (Condition Code Status): Update documention for
6378 relative locations of cc0-setter and cc0-user.
6380 2014-02-27 Jeff Law <law@redhat.com>
6382 PR rtl-optimization/52714
6383 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6384 into two independent simple sets, if I3 is a jump, ensure the
6385 pattern we place into I3 is a (set (pc) ...).
6387 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6388 Jeff Law <law@redhat.com>
6390 PR rtl-optimization/49847
6391 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6392 are in different blocks.
6393 * doc/tm.texi (Condition Code Status): Update documention for
6394 relative locations of cc0-setter and cc0-user.
6396 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6399 * lra.c (lra_emit_add): Check SUBREG too.
6401 2014-02-27 Andreas Schwab <schwab@suse.de>
6403 * config/m68k/m68k.c (m68k_option_override): Disable
6404 -flive-range-shrinkage for classic m68k.
6405 (m68k_override_options_after_change): Likewise.
6407 2014-02-27 Marek Polacek <polacek@redhat.com>
6410 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6411 -Wmaybe-uninitialized.
6413 2014-02-27 Alan Modra <amodra@gmail.com>
6416 * reload1.c (emit_input_reload_insns): When reload_override_in,
6417 set old to rl->in_reg when rl->in_reg is a subreg.
6419 2014-02-26 Richard Biener <rguenther@suse.de>
6422 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6424 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6426 * common/config/i386/predicates.md (const1256_operand): Remove.
6427 (const2356_operand): New.
6428 (const_1_to_2_operand): Remove.
6429 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6430 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6431 (*avx512pf_gatherpf<mode>sf): Ditto.
6432 (avx512pf_gatherpf<mode>df): Ditto.
6433 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6434 (*avx512pf_gatherpf<mode>df): Ditto.
6435 (avx512pf_scatterpf<mode>sf): Ditto.
6436 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6437 (*avx512pf_scatterpf<mode>sf): Ditto.
6438 (avx512pf_scatterpf<mode>df): Ditto.
6439 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6440 (*avx512pf_scatterpf<mode>df): Ditto.
6441 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6443 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6445 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6446 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6447 (_mm512_mask_testn_epi64_mask): Move to ...
6448 * config/i386/avx512cdintrin.h: Here.
6449 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6450 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6451 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6452 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6453 TARGET_AVX512F from TARGET_AVX512CD.
6455 2014-02-26 Richard Biener <rguenther@suse.de>
6458 * ipa.c (walk_polymorphic_call_targets): Properly guard
6459 call to inline_update_overall_summary.
6461 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6464 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6465 and latches only if requested. Fix latch if it is removed.
6466 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6467 LOOPS_HAVE_PREHEADERS.
6469 2014-02-25 Andrew Pinski <apinski@cavium.com>
6471 * builtins.c (expand_builtin_thread_pointer): Create a new target
6472 when the target is NULL.
6474 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6476 PR rtl-optimization/60317
6477 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6478 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6479 * lra-assigns.c: Include params.h.
6480 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6481 other reload pseudos considerations.
6483 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6485 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6486 to use canonical form for nor<mode>3.
6488 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6491 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6494 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6496 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6497 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6498 (ix86_handle_option): Handle OPT_mprefetchwt1.
6499 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6500 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6502 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6503 OPTION_MASK_ISA_PREFETCHWT1.
6504 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6505 (PTA_PREFETCHWT1): New.
6506 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6507 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6508 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6509 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6510 (*prefetch_avx512pf_<mode>_: Change into ...
6511 (*prefetch_prefetchwt1_<mode>: This.
6512 * config/i386/i386.opt (mprefetchwt1): New.
6513 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6514 (_mm_prefetch): Handle intent to write.
6515 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6517 2014-02-25 Richard Biener <rguenther@suse.de>
6520 * emit-rtl.c (mem_attrs_htab): Remove.
6521 (mem_attrs_htab_hash): Likewise.
6522 (mem_attrs_htab_eq): Likewise.
6523 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6524 (init_emit_once): Do not allocate mem_attrs_htab.
6526 2014-02-25 Richard Biener <rguenther@suse.de>
6529 * lto-opts.c (lto_write_options): Output non-explicit conservative
6530 -fwrapv, -fno-trapv and -fno-strict-overflow.
6531 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6532 (run_gcc): And pass them through.
6534 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6536 * sel-sched.c (calculate_new_fences): New parameter ptime.
6537 Calculate it as a maximum over all fence cycles.
6538 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6539 Print the final schedule timing when sched_verbose.
6541 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6543 PR rtl-optimization/60292
6544 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6545 bit fot the fence instruction.
6547 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6549 * calls.h: Fix typo in comment.
6551 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6553 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6554 adjusting offsetable addresses.
6556 2014-02-24 Guozhi Wei <carrot@google.com>
6558 * sparseset.h (sparseset_pop): Fix the wrong index.
6560 2014-02-24 Walter Lee <walt@tilera.com>
6562 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6563 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6565 * common/config/tilegx/tilegx-common.c
6566 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6567 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6569 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6570 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6571 (tilegx_gimplify_va_arg_expr): Handle big endian.
6572 (tilegx_expand_unaligned_load): Ditto.
6573 (tilegx_expand_unaligned_store): Ditto.
6574 (TARGET_RETURN_IN_MSB): New.
6575 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6576 (TARGET_ENDIAN_DEFAULT): New.
6577 (TARGET_BIG_ENDIAN): Handle big endian.
6578 (BYTES_BIG_ENDIAN): Ditto.
6579 (WORDS_BIG_ENDIAN): Ditto.
6580 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6583 * config/tilegx/tilegx.md (extv): Handle big endian.
6585 (insn_st<n>): Ditto.
6586 (insn_st<n>_add<bitsuffix>): Ditto.
6587 (insn_stnt<n>): Ditto.
6588 (insn_stnt<n>_add<bitsuffix>):Ditto.
6589 (vec_interleave_highv8qi): Handle big endian.
6590 (vec_interleave_highv8qi_be): New.
6591 (vec_interleave_highv8qi_le): New.
6592 (insn_v1int_h): Handle big endian.
6593 (vec_interleave_lowv8qi): Handle big endian.
6594 (vec_interleave_lowv8qi_be): New.
6595 (vec_interleave_lowv8qi_le): New.
6596 (insn_v1int_l): Handle big endian.
6597 (vec_interleave_highv4hi): Handle big endian.
6598 (vec_interleave_highv4hi_be): New.
6599 (vec_interleave_highv4hi_le): New.
6600 (insn_v2int_h): Handle big endian.
6601 (vec_interleave_lowv4hi): Handle big endian.
6602 (vec_interleave_lowv4hi_be): New.
6603 (vec_interleave_lowv4hi_le): New.
6604 (insn_v2int_l): Handle big endian.
6605 (vec_interleave_highv2si): Handle big endian.
6606 (vec_interleave_highv2si_be): New.
6607 (vec_interleave_highv2si_le): New.
6608 (insn_v4int_h): Handle big endian.
6609 (vec_interleave_lowv2si): Handle big endian.
6610 (vec_interleave_lowv2si_be): New.
6611 (vec_interleave_lowv2si_le): New.
6612 (insn_v4int_l): Handle big endian.
6613 * config/tilegx/tilegx.opt (mbig-endian): New option.
6614 (mlittle-endian): New option.
6615 * doc/install.texi: Document tilegxbe-linux.
6616 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6618 2014-02-24 Martin Jambor <mjambor@suse.cz>
6621 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6622 there are no parameter descriptors.
6624 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6626 PR rtl-optimization/60268
6627 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6628 initialization to ...
6629 (sched_rgn_init): ... here.
6630 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6632 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6634 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6637 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6639 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6642 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6644 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6645 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6647 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6649 * config/microblaze/predicates.md: Add cmp_op predicate.
6650 * config/microblaze/microblaze.md: Add branch_compare instruction
6651 which uses cmp_op predicate and emits cmp insn before branch.
6652 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6653 to microblaze_expand_conditional_branch and consolidate logic.
6654 (microblaze_expand_conditional_branch): emit branch_compare
6655 insn instead of handling cmp op separate from branch insn.
6657 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6659 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6662 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6664 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6665 define_insn with define_expand and new define_insn
6666 *altivec_lve<VI_char>x_internal.
6667 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6668 and new define_insn *altivec_stve<VI_char>x_internal.
6669 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6671 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6673 (altivec_expand_stvex_be): New function.
6675 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6677 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6678 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6679 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6680 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6682 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6685 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6686 instead of emit_move_insn.
6688 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6690 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6692 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6693 gen_altivec_vsumsws.
6695 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6697 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6698 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6699 (altivec_lvxl_<mode>): New define_expand incorporating
6700 -maltivec=be semantics where needed.
6701 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6702 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6703 semantics where needed.
6704 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6705 (altivec_stvx_<mode>): New define_expand incorporating
6706 -maltivec=be semantics where needed.
6707 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6708 VM2 iterator instead of V4SI.
6709 (altivec_stvxl_<mode>): New define_expand incorporating
6710 -maltivec=be semantics where needed.
6711 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6712 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6713 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6714 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6715 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6716 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6717 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6718 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6719 ALTIVEC_BUILTIN_STVXL.
6720 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6721 (altivec_expand_stvx_be): Likewise.
6722 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6723 (altivec_expand_lvx_be): Likewise.
6724 (altivec_expand_stvx_be): Likewise.
6725 (altivec_expand_builtin): Add cases for
6726 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6727 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6728 (altivec_init_builtins): Add definitions for
6729 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6730 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6732 2014-02-21 Catherine Moore <clm@codesourcery.com>
6734 * doc/invoke.texi (mvirt, mno-virt): Document.
6735 * config/mips/mips.opt (mvirt): New option.
6736 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6738 2014-02-21 Richard Biener <rguenther@suse.de>
6740 PR tree-optimization/60276
6741 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6742 (STMT_VINFO_MIN_NEG_DIST): New macro.
6743 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6744 STMT_VINFO_MIN_NEG_DIST.
6745 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6746 made for negative dependence distances still hold.
6748 2014-02-21 Richard Biener <rguenther@suse.de>
6751 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6752 DECL_INITIAL for globals not in the current function context.
6754 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6756 PR tree-optimization/56490
6757 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6758 * tree-ssa-uninit.c: Include params.h.
6759 (compute_control_dep_chain): Add num_calls argument, return false
6760 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6761 num_calls to recursive call.
6762 (find_predicates): Change dep_chain into normal array,
6763 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6764 variable and adjust compute_control_dep_chain caller.
6765 (find_def_preds): Likewise.
6767 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6769 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6770 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6772 2014-02-21 Nick Clifton <nickc@redhat.com>
6774 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6775 (pushhi1): Likewise.
6776 (popqi1): Add mode to pre_dec.
6779 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6781 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6782 mode for mask of V8SFmode permutation.
6784 2014-02-20 Richard Henderson <rth@redhat.com>
6787 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6788 a new pseudo for OLDVAL.
6790 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6793 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6794 gen_reg_rtx if d->testing_p.
6795 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6796 if d->testing_p and we will certainly return true.
6797 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6800 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6802 * emit-rtl.c (gen_reg_rtx): Assert that
6803 crtl->emit.regno_pointer_align_length is non-zero.
6805 2014-02-20 Richard Henderson <rth@redhat.com>
6808 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6809 on failure the store back into EXPECT.
6811 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6812 Sandra Loosemore <sandra@codesourcery.com>
6814 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6815 * config/nios2/nios2.c (nios2_function_profiler): Add
6816 -fPIC (flag_pic == 2) support.
6817 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6818 (nios2_large_offset_p): New function.
6819 (nios2_unspec_reloc_p): Move up position, update to use
6820 nios2_large_offset_p.
6821 (nios2_unspec_address): Remove function.
6822 (nios2_unspec_offset): New function.
6823 (nios2_large_got_address): New function.
6824 (nios2_got_address): Add large offset support.
6825 (nios2_legitimize_tls_address): Update usage of removed and new
6827 (nios2_symbol_binds_local_p): New function.
6828 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6829 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6830 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6831 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6832 processing for (const (unspec ...)).
6833 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6835 2014-02-20 Richard Biener <rguenther@suse.de>
6837 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6838 doing the substitution.
6839 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6841 2014-02-20 Martin Jambor <mjambor@suse.cz>
6844 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6845 info when checking whether lattices are bottom.
6847 2014-02-20 Richard Biener <rguenther@suse.de>
6850 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6853 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6856 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6857 parameter specifying the scaling.
6858 (inline_call): Update.
6859 (want_inline_recursively): Guard division by zero.
6860 (recursive_inlining): Update.
6861 * ipa-inline.h (clone_inlined_nodes): Update.
6863 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6866 * config/i386/i386.c (classify_argument): Pass structures of size
6867 64 bytes or less in register.
6869 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6870 Kirill Yukhin <kirill.yukhin@intel.com>
6872 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6873 (_mm_rcp28_round_ss): Ditto.
6874 (_mm_rsqrt28_round_sd): Ditto.
6875 (_mm_rsqrt28_round_ss): Ditto.
6876 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6877 (_mm_rcp14_round_ss): Ditto.
6878 (_mm_rsqrt14_round_sd): Ditto.
6879 (_mm_rsqrt14_round_ss): Ditto.
6880 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6881 the first input operand, get rid of match_dup.
6882 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6884 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6886 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6887 operand as the first input operand, set type attribute.
6888 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6890 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6891 operand as the first input operand, set type attribute.
6893 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6895 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6898 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6901 * config/i386/i386.c (construct_container): Remove TFmode check
6902 for X86_64_INTEGER_CLASS.
6904 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6907 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6908 only when -Wpsabi is enabled.
6910 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6913 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6914 passing arrays in registers are the same as for structs, so remove the
6915 special case for them.
6917 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6919 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6920 destination type, extract only the valid bits if the source type is not
6921 integral and has a different mode.
6923 2014-02-19 Richard Biener <rguenther@suse.de>
6926 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6929 2014-02-19 Richard Biener <rguenther@suse.de>
6932 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6933 (ipa_modify_call_arguments): Emit an argument load explicitely and
6934 preserve virtual SSA form there and for the replacement call.
6935 Do not update SSA form nor free dominance info.
6937 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6939 * ipa.c (function_and_variable_visibility): Also clear WEAK
6940 flag when disolving COMDAT_GROUP.
6942 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6944 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6945 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6947 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6948 do more sanity checks.
6949 (detect_type_change): Return true when giving up early.
6950 (compute_complex_assign_jump_func): Fix type parameter of
6951 ipa_set_ancestor_jf.
6952 (compute_complex_ancestor_jump_func): Likewise.
6953 (update_jump_functions_after_inlining): Fix updating of
6955 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6957 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6959 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6960 inline clones when edge disappears.
6962 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6965 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6966 Split 64-bit moves into 2 patterns. Do not allow the use of
6967 direct move for TDmode in little endian, since the decimal value
6968 has little endian bytes within a word, but the 64-bit pieces are
6969 ordered in a big endian fashion, and normal subreg's of TDmode are
6971 (mov<mode>_64bit_dm): Likewise.
6972 (movtd_64bit_nodm): Likewise.
6974 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6976 PR tree-optimization/60174
6977 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6978 statement of an SSA_NAME that occurs in an abnormal PHI node.
6980 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6983 * final.c (SEEN_BB): Remove.
6984 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6985 (final_scan_insn): Don't force_source_line on second
6986 NOTE_INSN_BASIC_BLOCK.
6988 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6991 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6992 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6993 (type_natural_mode): Warn ABI change when %zmm register is not
6994 available for AVX512F vector value passing.
6996 2014-02-18 Kai Tietz <ktietz@redhat.com>
6999 * config/i386/i386.c (ix86_expand_prologue): Use value in
7000 rax register as displacement when restoring %r10 or %rax.
7001 Fix wrong offset when restoring both registers.
7003 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7005 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7006 assertion with conditional return.
7008 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7009 Uros Bizjak <ubizjak@gmail.com>
7012 * config/i386/driver-i386.c (host_detect_local_cpu): If
7013 YMM state is not saved by the OS, also clear has_f16c. Move
7014 CPUID 0x80000001 handling before YMM state saving checking.
7016 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7018 PR rtl-optimization/58960
7019 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7020 factored out from ...
7021 (sched_init): ... here.
7022 (free_global_sched_pressure_data): New, factored out from ...
7023 (sched_finish): ... here.
7024 * sched-int.h (free_global_sched_pressure_data): Declare.
7025 * sched-rgn.c (nr_regions_initial): New static global.
7026 (haifa_find_rgns): Initialize it.
7027 (schedule_region): Disable sched-pressure for the newly
7030 2014-02-17 Richard Biener <rguenther@suse.de>
7032 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7033 release SSA defs of pattern stmts.
7035 2014-02-17 Richard Biener <rguenther@suse.de>
7037 * tree-inline.c (expand_call_inline): Release the virtual
7038 operand defined by the call we are about to inline.
7040 2014-02-17 Richard Biener <rguenther@suse.de>
7042 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7044 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7045 Ilya Tocar <ilya.tocar@intel.com>
7047 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7048 arguments order in builtin.
7049 (_mm512_permutexvar_epi64): Ditto.
7050 (_mm512_mask_permutexvar_epi64): Ditto
7051 (_mm512_maskz_permutexvar_epi32): Ditto
7052 (_mm512_permutexvar_epi32): Ditto
7053 (_mm512_mask_permutexvar_epi32): Ditto
7055 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7057 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7058 (p8_vmrgow): Likewise.
7060 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7062 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7065 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7068 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7069 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7070 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7071 using direct move instructions on ISA 2.07. Also adjust
7072 instruction length for 64-bit.
7073 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7074 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7076 2014-02-15 Alan Modra <amodra@gmail.com>
7080 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7081 find_replacement on parts of insn rtl that might be reloaded.
7083 2014-02-15 Richard Biener <rguenther@suse.de>
7085 PR tree-optimization/60183
7086 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7087 (tree_ssa_phiprop): Calculate and free post-dominators.
7089 2014-02-14 Jeff Law <law@redhat.com>
7091 PR rtl-optimization/60131
7092 * ree.c (get_extended_src_reg): New function.
7093 (combine_reaching_defs): Use it rather than assuming location of REG.
7094 (find_and_remove_re): Verify first operand of extension is
7095 a REG before adding the insns to the copy list.
7097 2014-02-14 Roland McGrath <mcgrathr@google.com>
7099 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7100 * configure: Regenerated.
7101 * config.in: Regenerated.
7102 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7103 instead of ASM_SHORT.
7105 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7106 Richard Earnshaw <rearnsha@arm.com>
7108 PR rtl-optimization/59535
7109 * lra-constraints.c (process_alt_operands): Encourage alternative
7110 when unassigned pseudo class is superset of the alternative class.
7111 (inherit_reload_reg): Don't inherit when optimizing for code size.
7112 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7113 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7114 modes not less than 4 for Thumb1.
7116 2014-02-14 Kyle McMartin <kyle@redhat.com>
7119 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7121 2014-02-14 Richard Biener <rguenther@suse.de>
7123 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7124 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7125 Do not strip INDIRECT_REFs.
7127 2014-02-14 Richard Biener <rguenther@suse.de>
7130 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7131 DECL_FUNCTION_SPECIFIC_TARGET.
7132 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7133 * tree-streamer-out.c (pack_ts_target_option): Remove.
7134 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7135 (write_ts_function_decl_tree_pointers): Do not stream
7136 DECL_FUNCTION_SPECIFIC_TARGET.
7137 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7138 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7139 (lto_input_ts_function_decl_tree_pointers): Do not stream
7140 DECL_FUNCTION_SPECIFIC_TARGET.
7142 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7144 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7145 (get_initial_def_for_induction, vectorizable_induction): Ignore
7146 debug stmts when looking for exit_phi.
7147 (vectorizable_live_operation): Fix up condition.
7149 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7151 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7152 nreverse() because it changes the content of original tree list.
7154 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7156 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7157 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7159 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7161 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7162 GNU coding standards.
7164 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7167 * dwarf2out.c (gen_subprogram_die): Don't call
7168 add_calling_convention_attribute if subr_die is old_die.
7170 2014-02-13 Sharad Singhai <singhai@google.com>
7172 * doc/optinfo.texi: Fix order of nodes.
7174 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7176 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7177 operands[2], not operands[3].
7179 2014-02-13 Richard Biener <rguenther@suse.de>
7182 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7183 mention the possibility of an in-tree build.
7184 (CLooG): Update recommended version to 0.18.1, mention the
7185 possibility of an in-tree build and clarify that the ISL
7186 bundled with CLooG does not work.
7188 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7191 * expr.c (compress_float_constant): If x is a hard register,
7192 extend into a pseudo and then move to x.
7194 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7196 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7197 caused by bad second argument to warning_at() with -mhotpatch and
7198 nested functions (e.g. with gfortran).
7200 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7202 * opts.c (option_name): Remove "enabled by default" rider.
7204 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7206 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7208 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7209 Uros Bizjak <ubizjak@gmail.com>
7212 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7213 * configure: Regenerated.
7215 2014-02-12 Richard Biener <rguenther@suse.de>
7217 * vec.c (vec_prefix::calculate_allocation): Move as
7218 inline variant to vec.h.
7219 (vec_prefix::calculate_allocation_1): New out-of-line version.
7220 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7221 (vec_prefix::m_has_auto_buf): Rename to ...
7222 (vec_prefix::m_using_auto_storage): ... this.
7223 (vec_prefix::calculate_allocation): Inline the easy cases
7224 and dispatch to calculate_allocation_1 which doesn't need the
7226 (va_heap::reserve): Use gcc_checking_assert.
7227 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7228 m_using_auto_storage.
7229 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7231 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7232 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7233 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7235 2014-02-12 Richard Biener <rguenther@suse.de>
7237 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7238 when we found a dependence.
7240 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7242 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7244 (maybe_fold_stmt): ... into this new function.
7245 * omp-low.c (lower_omp): Update comment.
7247 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7250 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7253 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7255 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7256 identifiers in comments.
7257 (cortexa53_extra_costs): Likewise.
7258 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7259 (cortexa7_extra_costs): Likewise.
7260 (cortexa12_extra_costs): Likewise.
7261 (cortexa15_extra_costs): Likewise.
7262 (v7m_extra_costs): Likewise.
7264 2014-02-12 Richard Biener <rguenther@suse.de>
7267 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7268 of posix_memalign being successful.
7269 (lower_stmt): Restrict lowering of posix_memalign to when
7270 -ftree-bit-ccp is enabled.
7272 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7274 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7276 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7278 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7280 PR rtl-optimization/60116
7281 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7282 other_insn once the combination has been validated.
7284 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7287 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7289 * ipa-devirt.c: Include demangle.h
7290 (odr_violation_reported): New static variable.
7291 (add_type_duplicate): Update odr_violations.
7292 (maybe_record_node): Add completep parameter; update it.
7293 (record_target_from_binfo): Add COMPLETEP parameter;
7294 update it as needed.
7295 (possible_polymorphic_call_targets_1): Likewise.
7296 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7297 rename FINAL to COMPLETE.
7298 (record_targets_from_bases): Sanity check we found the binfo;
7299 fix COMPLETEP updating.
7300 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7301 parameter, fix computing of COMPLETEP.
7302 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7303 at LTO time do demangling.
7304 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7305 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7307 (gimple_get_virt_method_for_binfo): Likewise.
7308 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7309 gimple_get_virt_method_for_vtable): Update prototypes.
7311 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7314 * genautomata.c (add_presence_absence): Fix typo with
7315 {final_}presence_list.
7317 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7320 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7321 for VSX/Altivec vectors that land in GPR registers.
7323 2014-02-11 Richard Henderson <rth@redhat.com>
7324 Jakub Jelinek <jakub@redhat.com>
7327 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7328 around drhs if type conversion to lacc->type is not useless.
7330 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7332 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7334 (cortex-a57.cortex-a53): Likewise.
7335 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7337 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7339 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7342 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7344 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7345 add_options_for_arm_vfp3.
7347 2014-02-11 Jeff Law <law@redhat.com>
7350 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7351 object with an undesirable mode.
7353 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7356 * config/i386/sol2-9.h: New file.
7357 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7358 *-*-solaris2.9*): Use it.
7360 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7362 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7363 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7365 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7367 * config/microblaze/microblaze.c: Extend mcpu version format
7369 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7371 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7373 2014-02-10 Richard Henderson <rth@redhat.com>
7376 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7377 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7378 ms-abi vs -mno-accumulate-outgoing-args.
7379 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7380 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7383 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7386 * cfgexpand.c (expand_asm_operands): Attach source location to
7387 ASM_INPUT rtx objects.
7388 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7390 2014-02-10 Nick Clifton <nickc@redhat.com>
7392 * config/mn10300/mn10300.c (popcount): New function.
7393 (mn10300_expand_prologue): Include saved registers in stack usage
7396 2014-02-10 Jeff Law <law@redhat.com>
7399 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7400 when changing the SET_DEST of a prior insn to avoid an input reload.
7402 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7404 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7405 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7406 -mcall-openbsd, or -mcall-linux.
7407 (CC1_ENDIAN_BIG_SPEC): Remove.
7408 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7409 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7410 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7411 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7412 and %cc1_endian_default.
7413 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7415 2014-02-10 Richard Biener <rguenther@suse.de>
7417 PR tree-optimization/60115
7418 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7419 MEM_REF handling. Properly verify that the accesses are not
7420 out of the objects bound.
7422 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7424 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7427 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7429 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7430 proper constants and fix formatting.
7431 (possible_polymorphic_call_targets): Fix formatting.
7433 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7434 Ilya Tocar <ilya.tocar@intel.com>
7436 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7437 (_mm512_loadu_epi32): Renamed into...
7438 (_mm512_loadu_si512): This.
7439 (_mm512_storeu_epi32): Renamed into...
7440 (_mm512_storeu_si512): This.
7441 (_mm512_maskz_ceil_ps): Removed.
7442 (_mm512_maskz_ceil_pd): Ditto.
7443 (_mm512_maskz_floor_ps): Ditto.
7444 (_mm512_maskz_floor_pd): Ditto.
7445 (_mm512_floor_round_ps): Ditto.
7446 (_mm512_floor_round_pd): Ditto.
7447 (_mm512_ceil_round_ps): Ditto.
7448 (_mm512_ceil_round_pd): Ditto.
7449 (_mm512_mask_floor_round_ps): Ditto.
7450 (_mm512_mask_floor_round_pd): Ditto.
7451 (_mm512_mask_ceil_round_ps): Ditto.
7452 (_mm512_mask_ceil_round_pd): Ditto.
7453 (_mm512_maskz_floor_round_ps): Ditto.
7454 (_mm512_maskz_floor_round_pd): Ditto.
7455 (_mm512_maskz_ceil_round_ps): Ditto.
7456 (_mm512_maskz_ceil_round_pd): Ditto.
7457 (_mm512_expand_pd): Ditto.
7458 (_mm512_expand_ps): Ditto.
7459 * config/i386/i386.c (ix86_builtins): Remove
7460 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7461 (bdesc_args): Ditto.
7462 * config/i386/predicates.md (const1256_operand): New.
7463 (const_1_to_2_operand): Ditto.
7464 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7465 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7466 (*avx512pf_gatherpf<mode>sf): Ditto.
7467 (avx512pf_gatherpf<mode>df): Ditto.
7468 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7469 (*avx512pf_gatherpf<mode>df): Ditto.
7470 (avx512pf_scatterpf<mode>sf): Ditto.
7471 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7472 (*avx512pf_scatterpf<mode>sf): Ditto.
7473 (avx512pf_scatterpf<mode>df): Ditto.
7474 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7475 (*avx512pf_scatterpf<mode>df): Ditto.
7476 (avx512f_expand<mode>): Removed.
7477 (<shift_insn><mode>3<mask_name>): Change predicate type.
7479 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7481 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7482 not at the end of datarefs vector use ordered_remove to avoid
7483 reordering datarefs vector.
7486 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7487 mark local addressable non-static vars as GOVD_PRIVATE
7488 instead of GOVD_LOCAL.
7489 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7490 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7494 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7495 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7496 assume_aligned or alloc_align attributes.
7497 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7498 arguments. Handle also assume_aligned and alloc_align attributes.
7499 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7500 calls to functions with assume_aligned or alloc_align attributes.
7501 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7503 2014-02-08 Terry Guo <terry.guo@arm.com>
7505 * doc/invoke.texi: Document ARM -march=armv7e-m.
7507 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7509 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7510 flag on __cilkrts_rethrow builtin.
7513 * ipa-cp.c (determine_versionability): Fail at -O0
7514 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7515 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7518 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7521 * tree-inline.c (copy_forbidden): Fail for
7522 __attribute__((optimize (0))) functions.
7524 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7526 * varpool.c: Include pointer-set.h.
7527 (varpool_remove_unreferenced_decls): Variables in other partitions
7528 will not be output; be however careful to not lose information
7531 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7533 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7534 lookup in the vtable constructor.
7536 2014-02-07 Jeff Law <law@redhat.com>
7539 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7540 define_insn_and_split.
7542 * ipa-inline.c (inline_small_functions): Fix typos.
7544 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7546 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7547 (s390_can_use_return_insn): Declare.
7548 * config/s390/s390.h (EPILOGUE_USES): Define.
7549 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7551 (s390_chunkify_start): Handle return JUMP_LABELs.
7552 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7553 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7554 (s390_can_use_return_insn): New functions.
7555 (s390_fix_long_loop_prediction): Handle conditional returns.
7556 (TARGET_SET_UP_BY_PROLOGUE): Define.
7557 * config/s390/s390.md (ANY_RETURN): New code iterator.
7558 (*creturn, *csimple_return, return, simple_return): New patterns.
7560 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7562 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7563 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7564 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7565 REG_CFA_RESTORE list when deciding not to restore a register.
7567 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7569 * config/s390/s390.c: Include tree-pass.h and context.h.
7570 (s390_early_mach): New function, split out from...
7571 (s390_emit_prologue): ...here.
7572 (pass_data_s390_early_mach): New pass structure.
7573 (pass_s390_early_mach): New class.
7574 (s390_option_override): Create and register early_mach pass.
7575 Move to end of file.
7577 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7579 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7580 to match for the exit block.
7582 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7584 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7585 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7586 Reject misaligned operands.
7588 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7590 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7592 2014-02-07 Richard Biener <rguenther@suse.de>
7595 * gimple-low.c (lower_builtin_posix_memalign): New function.
7596 (lower_stmt): Call it to lower posix_memalign in a way
7597 to make alignment info accessible.
7599 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7602 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7603 __builtin_setjmp_receiver.
7605 2014-02-07 Richard Biener <rguenther@suse.de>
7608 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7609 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7610 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7611 Handle BUILT_IN_POSIX_MEMALIGN.
7612 (find_func_clobbers): Likewise.
7613 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7614 (call_may_clobber_ref_p_1): Likewise.
7616 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7619 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7622 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7625 * lto-cgraph.c (lto_output_node): Use
7626 symtab_get_symbol_partitioning_class.
7627 (lto_output_varpool_node): likewise.
7628 (symtab_get_symbol_partitioning_class): Move here from
7630 * cgraph.h (symbol_partitioning_class): Likewise.
7631 (symtab_get_symbol_partitioning_class): Declare.
7633 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7635 * ggc.h (ggc_internal_cleared_alloc): New macro.
7636 * vec.h (vec_safe_copy): Handle memory stats.
7637 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7638 * target-globals.c (save_target_globals): Likewise.
7640 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7643 * expr.c (emit_move_resolve_push): Export; be bit more selective
7644 on when to clear alias set.
7645 * expr.h (emit_move_resolve_push): Declare.
7646 * function.h (struct function): Add tail_call_marked.
7647 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7648 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7649 * config/i386/i386.md (TImode move expander): De not call
7651 (FP push expanders): Preserve memory attributes.
7652 * config/i386/sse.md (push<mode>1): Remove.
7653 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7654 (ix86_expand_push): Remove.
7655 * config/i386/mmx.md (push<mode>1): Remove.
7657 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7659 PR rtl-optimization/60030
7660 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7661 lopart with paradoxical subreg before shifting it up by hprec.
7663 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7665 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7666 Remove extra newline at end of file.
7667 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7668 (arm_issue_rate): Handle cortexa57.
7669 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7670 (cortex-a57.cortex-a53): Likewise.
7672 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7675 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7676 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7678 (arm_expand_prologue): Adjust all callers.
7679 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7680 info, registers also at the lowest numbered registers side. Use
7681 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7685 * var-tracking.c (adjust_mems): Before adding a SET to
7686 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7688 2014-02-06 Alan Modra <amodra@gmail.com>
7691 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7692 change SDmode to DDmode when lra_in_progress.
7694 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7697 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7698 free_data_ref on the dr first, and before goto again also set dr
7699 to the next dr. For simd_lane_access, free old datarefs[i] before
7700 overwriting it. For get_vectype_for_scalar_type failure, don't
7701 free_data_ref if simd_lane_access.
7703 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7706 * tree.h (opts_for_fn): New inline function.
7707 (opt_for_fn): Define.
7708 * config/i386/i386.c (ix86_function_regparm): Use
7709 opt_for_fn (decl, optimize) instead of optimize.
7711 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7713 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7714 for SYMBOL_REF in large memory model.
7716 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7718 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7720 (cortex-a57): Likewise.
7721 (cortex-a57.cortex-a53): Likewise.
7723 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7724 Kugan Vivekanandarajah <kuganv@linaro.org>
7726 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7728 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7730 2014-02-06 Richard Biener <rguenther@suse.de>
7732 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7733 set_loop_copy and initialize_original_copy_tables.
7735 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7737 * config/aarch64/aarch64-simd.md
7738 (aarch64_ashr_simddi): Change QI to SI.
7740 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7741 Jakub Jelinek <jakub@redhat.com>
7744 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7747 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7749 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7750 CODE_FOR_altivec_vpku[hw]um to
7751 CODE_FOR_altivec_vpku[hw]um_direct.
7752 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7753 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7754 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7755 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7757 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7759 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7760 generation for -maltivec=be.
7761 (altivec_vsumsws): Simplify redundant test.
7763 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7765 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7766 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7767 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7768 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7769 gen_altivec_vpkuwum.
7770 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7772 (altivec_vpks<VI_char>ss): Likewise.
7773 (altivec_vpks<VI_char>us): Likewise.
7774 (altivec_vpku<VI_char>us): Likewise.
7775 (altivec_vpku<VI_char>um): Likewise.
7776 (altivec_vpku<VI_char>um_direct): New (copy of
7777 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7779 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7780 target is little endian and -maltivec=be is not specified.
7781 (*altivec_vupkhs<VU_char>_direct): New (copy of
7782 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7783 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7784 target is little endian and -maltivec=be is not specified.
7785 (*altivec_vupkls<VU_char>_direct): New (copy of
7786 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7787 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7788 little endian and -maltivec=be is not specified.
7789 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7790 little endian and -maltivec=be is not specified.
7792 2014-02-05 Richard Henderson <rth@redhat.com>
7795 * combine-stack-adj.c: Revert r206943.
7796 * sched-int.h (struct deps_desc): Add last_args_size.
7797 * sched-deps.c (init_deps): Initialize it.
7798 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7799 contain REG_ARGS_SIZE notes.
7801 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7803 * lto-cgraph.c (asm_nodes_output): Make global.
7804 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7805 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7806 (driver_handle_option): Handle OPT_fwpa.
7808 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7811 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7812 a comment typo and formatting issue. If odr_hash hasn't been
7813 created, return vNULL and set *completep to false.
7816 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7817 bb with no successors.
7819 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7822 * doc/invoke.texi (-march): Clarify documentation for ARM.
7826 2014-02-05 Richard Biener <rguenther@suse.de>
7828 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7829 when not vectorizing because of too many alias checks.
7830 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7831 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7833 2014-02-05 Nick Clifton <nickc@redhat.com>
7835 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7836 accept extended registers in any mode when compiling for the MN10300.
7838 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7840 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7841 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7842 sanitization attributes.
7843 (can_inline_edge_p): Likewise.
7844 (sanitize_attrs_match_for_inline_p): New function.
7846 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7848 * ipa-prop.c (detect_type_change): Shor circuit testing of
7849 type changes on THIS pointer.
7851 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7854 * config/pa/pa.c (legitimize_tls_address): Return original address
7855 if not passed a SYMBOL_REF rtx.
7856 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7858 (pa_emit_move_sequence): Simplify TLS source operands.
7859 (pa_legitimate_constant_p): Reject all TLS constants.
7860 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7861 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7863 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7865 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7866 groups when we know they are controlled by LTO.
7867 * varasm.c (default_binds_local_p_1): If object is in other partition,
7868 it will be resolved locally.
7870 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7872 * config/host-linux.c (linux_gt_pch_use_address): Don't
7873 use SSIZE_MAX because it is not always defined.
7875 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7878 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7879 threshold for pseudo splitting.
7880 (update_ebb_live_info): Process call argument hard registers and
7881 hard registers from insn definition too.
7882 (max_small_class_regs_num): New constant.
7883 (inherit_in_ebb): Update live hard regs through EBBs. Update
7884 reloads_num only for small register classes. Don't split for
7887 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7890 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7893 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7895 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7898 2014-02-04 Marek Polacek <polacek@redhat.com>
7900 * gdbinit.in (pel): Define.
7902 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7904 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7907 2014-02-04 Richard Biener <rguenther@suse.de>
7910 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7911 in function context local.
7912 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7913 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7914 similar to LTO_imported_decl_ref.
7916 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7918 PR tree-optimization/60002
7919 * cgraphclones.c (build_function_decl_skip_args): Clear
7922 PR tree-optimization/60023
7923 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7924 false to gsi_replace.
7925 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7926 has been in some EH region and vec_stmt could throw, add
7927 vec_stmt into the same EH region.
7928 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7929 has no lhs, ignore it.
7930 * internal-fn.c (expand_MASK_LOAD): Likewise.
7933 * tree-inline.c (copy_forbidden): Fail for
7934 __attribute__((optimize (0))) functions.
7937 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7938 is set, copy one less argument.
7939 (expand_simd_clones): Don't subtract clone_info->inbranch
7940 from simd_clone_struct_alloc argument.
7942 PR rtl-optimization/57915
7943 * recog.c (simplify_while_replacing): If all unary/binary/relational
7944 operation arguments are constant, attempt to simplify those.
7947 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7948 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7950 2014-02-04 Richard Biener <rguenther@suse.de>
7952 PR tree-optimization/60012
7953 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7954 TBAA disambiguation to all DDRs.
7956 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7959 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7960 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7962 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7965 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7967 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7969 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7970 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7972 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7975 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7976 to figure out targets of polymorphic calls with known decl.
7977 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7978 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7979 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7980 (get_polymorphic_call_info): ... here.
7981 (get_polymorphic_call_info_from_invariant): New function.
7983 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7985 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7986 lookup via vtable pointer; check for type consistency
7987 and turn inconsitent facts into UNREACHABLE.
7988 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7989 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7990 type inconsistent querries; return UNREACHABLE instead.
7992 2014-02-03 Richard Henderson <rth@twiddle.net>
7995 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7996 already processed this node.
7997 (normalize_one_pred_1): Pass along mark_set.
7998 (normalize_one_pred): Create and destroy a pointer_set_t.
7999 (normalize_one_pred_chain): Likewise.
8001 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8003 PR gcov-profile/58602
8004 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8006 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8009 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8010 -fno-devirtualize; try to devirtualize by the knowledge of
8011 virtual table pointer given by aggregate propagation.
8012 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8013 (ipa_print_node_jump_functions): Dump also offset that
8014 is relevant for polymorphic calls.
8015 (determine_known_aggregate_parts): Add arg_type parameter; use it
8016 instead of determining the type from pointer type.
8017 (ipa_compute_jump_functions_for_edge): Update call of
8018 determine_known_aggregate_parts.
8019 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8020 (gimple_get_virt_method_for_binfo): ... here; simplify using
8021 vtable_pointer_value_to_vtable.
8022 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8023 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8024 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8025 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8027 (vtable_pointer_value_to_binfo): ... here.
8028 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8030 2014-02-03 Teresa Johnson <tejohnson@google.com>
8032 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8033 redef of outer loop index variable.
8035 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8039 * doc/extend.texi (Function Attributes): Typo.
8041 2014-02-03 Cong Hou <congh@google.com>
8043 PR tree-optimization/60000
8044 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8045 if the vectorized statement is a store. A store statement can only
8046 appear at the end of pattern statements.
8048 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8050 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8051 (ix86_option_override_internal): Default long double to 64-bit for
8052 32-bit Bionic and to 128-bit for 64-bit Bionic.
8054 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8055 TARGET_LONG_DOUBLE_128 is true.
8056 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8058 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8059 (mlong-double-64): Negate -mlong-double-128.
8060 (mlong-double-128): New option.
8062 * config/i386/i386-c.c (ix86_target_macros): Define
8063 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8065 * doc/invoke.texi: Document -mlong-double-128.
8067 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8069 PR rtl-optimization/60024
8070 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8072 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8074 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8076 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8078 PR rtl-optimization/57662
8079 * sel-sched.c (code_motion_path_driver): Do not mark already not
8080 existing blocks in the visiting bitmap.
8082 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8084 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8085 on the insn being emitted.
8087 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8088 Will Deacon <will.deacon@arm.com>
8090 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8092 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8094 * config/arm/arm-tables.opt: Regenerate.
8096 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8098 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8099 for vector types other than V16QImode.
8100 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8101 define_expand, and call altivec_expand_vec_perm_le when producing
8102 code with little endian element order.
8103 (*altivec_vperm_<mode>_internal): New insn having previous
8104 behavior of altivec_vperm_<mode>.
8105 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8106 altivec_expand_vec_perm_le when producing code with little endian
8108 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8109 behavior of altivec_vperm_<mode>_uns.
8111 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8113 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8114 (altivec_vsumsws): Add handling for -maltivec=be with a little
8116 (altivec_vsumsws_direct): New.
8117 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8118 gen_altivec_vsumsws.
8120 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8122 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8123 vtable_pointer_value_to_binfo): New functions.
8124 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8125 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8127 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8129 * config/nios2/nios2.md (load_got_register): Initialize GOT
8130 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8131 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8133 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8135 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8136 preserverd by passthrough, do not propagate the type.
8138 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8140 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8141 (mips_atomic_assign_expand_fenv): New function.
8142 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8144 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8146 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8147 (__builtin_mips_set_fcsr): Likewise.
8148 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8149 MIPS_USI_FTYPE_VOID.
8150 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8151 (mips16_expand_set_fcsr): Likewise.
8152 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8153 (mips16_set_fcsr_stub): Likewise.
8154 (mips16_get_fcsr_one_only_stub): New class.
8155 (mips16_set_fcsr_one_only_stub): Likewise.
8156 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8157 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8158 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8159 (hard_float): New availability predicate.
8160 (mips_builtins): Add get_fcsr and set_fcsr.
8161 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8162 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8163 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8164 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8165 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8168 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8170 * config/mips/mips.c (mips_one_only_stub): New class.
8171 (mips_need_mips16_rdhwr_p): Replace with...
8172 (mips16_rdhwr_stub): ...this new variable.
8173 (mips16_stub_call_address): New function.
8174 (mips16_rdhwr_one_only_stub): New class.
8175 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8176 (mips_output_mips16_rdhwr): Delete.
8177 (mips_finish_stub): New function.
8178 (mips_code_end): Use it to handle rdhwr stubs.
8180 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8183 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8184 when calculating size of integer atomic types.
8186 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8188 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8190 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8192 PR tree-optimization/60003
8193 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8194 * profile.c (branch_prob): Use gimple_call_builtin_p
8195 to check for BUILT_IN_SETJMP_RECEIVER.
8196 * tree-inline.c (copy_bb): Call notice_special_calls.
8198 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8201 * lra-constraints.c (process_alt_operands): Update reload_sum only
8204 2014-01-31 Richard Henderson <rth@redhat.com>
8207 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8208 until after else_eh is processed.
8210 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8212 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8213 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8214 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8215 in smmintrin.h, remove them.
8216 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8217 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8218 * config/i386/i386.md (ROUND_SAE): Fix value.
8219 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8220 (const48_operand): New.
8221 * config/i386/subst.md (round), (round_expand): Use
8222 const_4_or_8_to_11_operand.
8223 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8225 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8227 * config/i386/constraints.md (Yk): Swap meaning with k.
8228 * config/i386/i386.md (movhi_internal): Change Yk to k.
8229 (movqi_internal): Ditto.
8230 (*k<logic><mode>): Ditto.
8233 (kandn<mode>): Ditto.
8234 (*<code>hi_1): Ditto.
8235 (*<code>qi_1): Ditto.
8236 (kxnor<mode>): Ditto.
8237 (kortestzhi): Ditto.
8238 (kortestchi): Ditto.
8240 (*one_cmplhi2_1): Ditto.
8241 (*one_cmplqi2_1): Ditto.
8242 * config/i386/sse.md (): Change k to Yk.
8243 (avx512f_load<mode>_mask): Ditto.
8244 (avx512f_blendm<mode>): Ditto.
8245 (avx512f_store<mode>_mask): Ditto.
8246 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8247 (avx512f_storedqu<mode>_mask): Ditto.
8248 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8250 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8251 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8252 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8253 (avx512f_maskcmp<mode>3): Ditto.
8254 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8255 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8256 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8257 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8258 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8259 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8260 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8261 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8262 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8263 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8264 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8265 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8266 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8267 (vec_extract_lo_<mode>_maskm): Ditto.
8268 (vec_extract_hi_<mode>_maskm): Ditto.
8269 (avx512f_vternlog<mode>_mask): Ditto.
8270 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8271 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8272 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8273 (avx512f_<code>v8div16qi2_mask): Ditto.
8274 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8275 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8276 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8277 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8278 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8279 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8280 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8281 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8282 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8283 (avx512cd_maskb_vec_dupv8di): Ditto.
8284 (avx512cd_maskw_vec_dupv16si): Ditto.
8285 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8286 (avx512f_vpermi2var<mode>3_mask): Ditto.
8287 (avx512f_vpermi2var<mode>3_mask): Ditto.
8288 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8289 (*avx512f_gathersi<mode>): Ditto.
8290 (*avx512f_gathersi<mode>_2): Ditto.
8291 (*avx512f_gatherdi<mode>): Ditto.
8292 (*avx512f_gatherdi<mode>_2): Ditto.
8293 (*avx512f_scattersi<mode>): Ditto.
8294 (*avx512f_scatterdi<mode>): Ditto.
8295 (avx512f_compress<mode>_mask): Ditto.
8296 (avx512f_compressstore<mode>_mask): Ditto.
8297 (avx512f_expand<mode>_mask): Ditto.
8298 * config/i386/subst.md (mask): Change k to Yk.
8299 (mask_scalar_merge): Ditto.
8302 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8304 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8306 2014-01-31 Richard Biener <rguenther@suse.de>
8309 * builtins.c (fold_builtin_memory_op): Make sure to not
8310 use a floating-point mode or a boolean or enumeral type for
8313 2014-01-30 DJ Delorie <dj@redhat.com>
8315 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8316 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8317 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8318 whenever main() has an epilogue.
8320 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8322 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8323 unused variable "field".
8324 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8325 (vsx_mergeh_<mode>): Likewise.
8326 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8327 (altivec_vmrghh): Likewise.
8328 (altivec_vmrghw): Likewise.
8329 (altivec_vmrglb): Likewise.
8330 (altivec_vmrglh): Likewise.
8331 (altivec_vmrglw): Likewise.
8332 (altivec_vspltb): Add missing uses.
8333 (altivec_vsplth): Likewise.
8334 (altivec_vspltw): Likewise.
8335 (altivec_vspltsf): Likewise.
8337 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8340 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8341 frame related instructions.
8343 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8345 PR rtl-optimization/59959
8346 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8347 any reload of register whose subreg is invalid.
8349 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8351 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8352 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8353 Add missing return type - void.
8355 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8357 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8358 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8359 remove element index adjustment for endian (now handled in vsx.md
8361 (altivec_expand_vec_perm_const): Use
8362 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8363 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8364 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8365 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8366 define_expand and a new define_insn *altivec_vspltb_internal;
8367 adjust for -maltivec=be on a little endian target.
8368 (altivec_vspltb_direct): New.
8369 (altivec_vsplth): Divide into a define_expand and a new
8370 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8371 little endian target.
8372 (altivec_vsplth_direct): New.
8373 (altivec_vspltw): Divide into a define_expand and a new
8374 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8375 little endian target.
8376 (altivec_vspltw_direct): New.
8377 (altivec_vspltsf): Divide into a define_expand and a new
8378 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8379 a little endian target.
8381 2014-01-30 Richard Biener <rguenther@suse.de>
8383 PR tree-optimization/59993
8384 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8385 can propagate form the earlier stmt and avoid the transform
8386 when the intermediate result is needed.
8388 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8390 * README.Portability: Fix typo.
8392 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8394 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8395 comparison_operator with ordered_comparison_operator.
8397 2014-01-30 Nick Clifton <nickc@redhat.com>
8399 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8400 Rename to mn10300_store_multiple_regs.
8401 * config/mn10300/mn10300.c: Likewise.
8402 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8403 store_multiple_regs.
8404 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8405 Call mn10300_store_multiple_regs.
8407 2014-01-30 Nick Clifton <nickc@redhat.com>
8408 DJ Delorie <dj@redhat.com>
8410 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8411 %fp 2 to keep registers after it properly word-aligned.
8412 (rl78_alloc_physical_registers_umul): Handle the case where both
8413 input operands are the same.
8415 2014-01-30 Richard Biener <rguenther@suse.de>
8417 PR tree-optimization/59903
8418 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8421 2014-01-30 Jason Merrill <jason@redhat.com>
8424 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8427 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8429 2014-01-30 Richard Biener <rguenther@suse.de>
8431 PR tree-optimization/59951
8432 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8434 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8437 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8438 SFmode to DFmode case.
8440 2014-01-29 DJ Delorie <dj@redhat.com>
8442 * config/msp430/msp430.opt (-minrt): New.
8443 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8445 (ENDFILE_SPEC): Likewise.
8447 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8449 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8450 (estimate_function_body_sizes): Use it.
8452 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8455 * dwarf2out.c (is_cxx_auto): New.
8456 (is_base_type): Use it.
8457 (gen_type_die_with_usage): Likewise.
8459 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8461 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8462 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8463 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8464 -maltivec=be with LE targets.
8465 (vsx_mergeh_<mode>): Likewise.
8466 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8467 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8468 (altivec_vmrghb): Replace with define_expand and new
8469 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8470 (altivec_vmrghb_direct): New define_insn.
8471 (altivec_vmrghh): Replace with define_expand and new
8472 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8473 (altivec_vmrghh_direct): New define_insn.
8474 (altivec_vmrghw): Replace with define_expand and new
8475 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8476 (altivec_vmrghw_direct): New define_insn.
8477 (*altivec_vmrghsf): Adjust for endianness.
8478 (altivec_vmrglb): Replace with define_expand and new
8479 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8480 (altivec_vmrglb_direct): New define_insn.
8481 (altivec_vmrglh): Replace with define_expand and new
8482 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8483 (altivec_vmrglh_direct): New define_insn.
8484 (altivec_vmrglw): Replace with define_expand and new
8485 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8486 (altivec_vmrglw_direct): New define_insn.
8487 (*altivec_vmrglsf): Adjust for endianness.
8488 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8489 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8490 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8491 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8492 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8493 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8494 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8495 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8497 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8499 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8500 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8503 2014-01-29 Richard Biener <rguenther@suse.de>
8505 PR tree-optimization/58742
8506 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8507 associate_pointerplus_align.
8508 (associate_pointerplus_diff): New function.
8509 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8510 and associate_pointerplus_diff.
8512 2014-01-29 Richard Biener <rguenther@suse.de>
8514 * lto-streamer.h (LTO_major_version): Bump to 3.
8515 (LTO_minor_version): Reset to 0.
8517 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8519 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8520 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8521 (arm_file_start): Generate correct asm header for armv7ve.
8522 * config/arm/bpabi.h: Add multilib support for armv7ve.
8523 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8524 and cortex-a15 to armv7ve.
8525 * config/arm/t-aprofile: Add multilib support for armv7ve.
8526 * doc/invoke.texi: Document -march=armv7ve.
8528 2014-01-29 Richard Biener <rguenther@suse.de>
8530 PR tree-optimization/58742
8531 * tree-ssa-forwprop.c (associate_plusminus): Return true
8532 if we changed sth, defer EH cleanup to ...
8533 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8534 (simplify_mult): New function.
8536 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8539 PR tree-optimization/59920
8540 * tree.c (build_common_builtin_nodes): Remove
8541 __builtin_setjmp_dispatcher initialization.
8542 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8543 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8544 instead of gsi_after_labels + manually skipping debug stmts.
8545 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8546 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8547 * tree-inline.c (copy_edges_for_bb): Remove
8548 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8549 argument. Ignore computed_goto_p stmts. Don't call
8550 make_abnormal_goto_edges. If a call might need abnormal edges
8551 for non-local gotos, see if it already has an edge to
8552 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8553 with true argument, don't do anything then, otherwise add
8554 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8555 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8557 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8558 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8559 (lower_stmt): Don't set data->calls_builtin_setjmp.
8560 (lower_builtin_setjmp): Adjust comment.
8561 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8562 * tree-cfg.c (found_computed_goto): Remove.
8563 (factor_computed_gotos): Remove.
8564 (make_goto_expr_edges): Return bool, true for computed gotos.
8565 Don't call make_abnormal_goto_edges.
8566 (build_gimple_cfg): Don't set found_computed_goto, don't call
8567 factor_computed_gotos.
8568 (computed_goto_p): No longer static.
8569 (make_blocks): Don't set found_computed_goto.
8570 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8571 (make_edges): If make_goto_expr_edges returns true, push bb
8572 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8573 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8574 vector. Record mapping between bbs and OpenMP regions if there
8575 are any, adjust make_gimple_omp_edges caller. Call
8576 handle_abnormal_edges.
8577 (make_abnormal_goto_edges): Remove.
8578 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8579 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8580 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8581 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8582 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8583 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8584 filling *region also set *region_idx to (*region)->entry->index.
8587 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8588 For REGs set ORIGINAL_REGNO.
8590 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8592 * doc/md.texi: Mention that a target shouldn't implement
8593 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8596 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8598 PR tree-optimization/59594
8599 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8600 a copy of the datarefs vector rather than the vector itself.
8602 2014-01-28 Jason Merrill <jason@redhat.com>
8605 * dwarf2out.c (auto_die): New static.
8606 (gen_type_die_with_usage): Handle C++1y 'auto'.
8607 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8610 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8613 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8614 (SPEC_X32): Likewise.
8615 (SPEC_64): Likewise.
8616 * config/i386/i386.c (ix86_option_override_internal): Turn off
8617 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8619 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8620 * config/i386/i386.h (TARGET_16BIT): New macro.
8621 (TARGET_16BIT_P): Likewise.
8622 * config/i386/i386.opt: Add m16.
8623 * doc/invoke.texi: Document -m16.
8625 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8627 PR preprocessor/59935
8628 * input.c (location_get_source_line): Bail out on when line number
8629 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8631 2014-01-28 Richard Biener <rguenther@suse.de>
8633 PR tree-optimization/58742
8634 * tree-ssa-forwprop.c (associate_plusminus): Handle
8635 pointer subtraction of the form (T)(P + A) - (T)P.
8637 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8639 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8642 2014-01-28 Richard Biener <rguenther@suse.de>
8645 2014-01-28 Richard Biener <rguenther@suse.de>
8647 PR rtl-optimization/45364
8648 PR rtl-optimization/59890
8649 * var-tracking.c (local_get_addr_clear_given_value): Handle
8650 already cleared slot.
8651 (val_reset): Handle not allocated local_get_addr_cache.
8652 (vt_find_locations): Use post-order on the inverted CFG.
8654 2014-01-28 Richard Biener <rguenther@suse.de>
8656 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8658 2014-01-28 Richard Biener <rguenther@suse.de>
8660 PR rtl-optimization/45364
8661 PR rtl-optimization/59890
8662 * var-tracking.c (local_get_addr_clear_given_value): Handle
8663 already cleared slot.
8664 (val_reset): Handle not allocated local_get_addr_cache.
8665 (vt_find_locations): Use post-order on the inverted CFG.
8667 2014-01-28 Alan Modra <amodra@gmail.com>
8669 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8670 * configure.ac <recursive call for build != host>: Define
8671 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8672 and LD_FOR_BUILD too.
8673 * configure: Regenerate.
8675 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8677 * config/i386/i386.c (get_builtin_code_for_version): Separate
8678 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8679 Broadwell from Haswell.
8681 2014-01-27 Steve Ellcey <sellcey@mips.com>
8683 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8684 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8685 * config/mips/mips.c (mips_option_override): Change setting
8687 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8688 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8689 Change from Mask to Var.
8691 2014-01-27 Jeff Law <law@redhat.com>
8693 * ipa-inline.c (inline_small_functions): Fix typo.
8695 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8697 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8698 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8699 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8700 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8701 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8702 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8703 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8704 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8705 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8706 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8707 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8708 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8709 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8710 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8711 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8712 (_mm512_storeu_epi64): Ditto.
8713 (_mm512_cmpge_epi32_mask): Ditto.
8714 (_mm512_cmpge_epu32_mask): Ditto.
8715 (_mm512_cmpge_epi64_mask): Ditto.
8716 (_mm512_cmpge_epu64_mask): Ditto.
8717 (_mm512_cmple_epi32_mask): Ditto.
8718 (_mm512_cmple_epu32_mask): Ditto.
8719 (_mm512_cmple_epi64_mask): Ditto.
8720 (_mm512_cmple_epu64_mask): Ditto.
8721 (_mm512_cmplt_epi32_mask): Ditto.
8722 (_mm512_cmplt_epu32_mask): Ditto.
8723 (_mm512_cmplt_epi64_mask): Ditto.
8724 (_mm512_cmplt_epu64_mask): Ditto.
8725 (_mm512_cmpneq_epi32_mask): Ditto.
8726 (_mm512_cmpneq_epu32_mask): Ditto.
8727 (_mm512_cmpneq_epi64_mask): Ditto.
8728 (_mm512_cmpneq_epu64_mask): Ditto.
8729 (_mm512_expand_pd): Ditto.
8730 (_mm512_expand_ps): Ditto.
8731 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8732 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8733 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8734 * config/i386/i386.c (ix86_builtins): Add
8735 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8736 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8737 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8738 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8739 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8740 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8741 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8742 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8743 IX86_BUILTIN_PMOVUSQW512_MEM.
8744 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8745 __builtin_ia32_pmovsqd512mem_mask,
8746 __builtin_ia32_pmovqd512mem_mask,
8747 __builtin_ia32_pmovusqw512mem_mask,
8748 __builtin_ia32_pmovsqw512mem_mask,
8749 __builtin_ia32_pmovqw512mem_mask,
8750 __builtin_ia32_pmovusdw512mem_mask,
8751 __builtin_ia32_pmovsdw512mem_mask,
8752 __builtin_ia32_pmovdw512mem_mask,
8753 __builtin_ia32_pmovqb512mem_mask,
8754 __builtin_ia32_pmovusqb512mem_mask,
8755 __builtin_ia32_pmovsqb512mem_mask,
8756 __builtin_ia32_pmovusdb512mem_mask,
8757 __builtin_ia32_pmovsdb512mem_mask,
8758 __builtin_ia32_pmovdb512mem_mask.
8759 (bdesc_args): Add __builtin_ia32_expanddf512,
8760 __builtin_ia32_expandsf512.
8761 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8762 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8763 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8764 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8765 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8766 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8767 (avx512f_<code>v8div16qi2_mask_store): This.
8768 (avx512f_expand<mode>): New.
8770 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8772 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8774 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8775 (_mm512_prefetch_i32scatter_pd): Ditto.
8776 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8777 (_mm512_prefetch_i64scatter_pd): Ditto.
8778 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8779 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8780 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8781 (_mm512_prefetch_i32scatter_ps): Ditto.
8782 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8783 (_mm512_prefetch_i64scatter_ps): Ditto.
8784 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8785 * config/i386/i386-builtin-types.def: Define
8786 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8787 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8788 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8789 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8790 IX86_BUILTIN_SCATTERPFQPD.
8791 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8792 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8793 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8794 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8795 __builtin_ia32_scatterpfqps.
8796 (ix86_expand_builtin): Expand new built-ins.
8797 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8798 fix memory access data type.
8799 (*avx512pf_gatherpf<mode>_mask): Ditto.
8800 (*avx512pf_gatherpf<mode>): Ditto.
8801 (avx512pf_scatterpf<mode>): Ditto.
8802 (*avx512pf_scatterpf<mode>_mask): Ditto.
8803 (*avx512pf_scatterpf<mode>): Ditto.
8804 (GATHER_SCATTER_SF_MEM_MODE): New.
8805 (avx512pf_gatherpf<mode>df): Ditto.
8806 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8807 (*avx512pf_scatterpf<mode>df): Ditto.
8809 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8812 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8813 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8816 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8818 * common/config/arm/arm-common.c
8819 (arm_rewrite_mcpu): Handle multiple names.
8821 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8823 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8825 * gimple-builder.h (create_gimple_tmp): Delete.
8827 2014-01-27 Christian Bruel <christian.bruel@st.com>
8829 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8832 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8834 * config/pa/pa.md (call): Generate indirect long calls to non-local
8835 functions when outputing 32-bit code.
8836 (call_value): Likewise except for special call to buggy powf function.
8838 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8839 portable runtime and PIC indirect calls.
8840 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8841 and PIC call sequences. Use ldo instead of blr to set return register
8842 in PIC call sequence.
8844 2014-01-25 Walter Lee <walt@tilera.com>
8846 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8847 avoid clobbering a live register.
8849 2014-01-25 Walter Lee <walt@tilera.com>
8851 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8852 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8853 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8854 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8856 2014-01-25 Walter Lee <walt@tilera.com>
8858 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8859 arguments on even registers.
8860 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8862 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8863 (BIGGEST_ALIGNMENT): Ditto.
8864 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8866 2014-01-25 Walter Lee <walt@tilera.com>
8868 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8869 insns before bundling.
8870 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8872 2014-01-25 Walter Lee <walt@tilera.com>
8874 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8875 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8876 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8878 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8880 * config/mips/constraints.md (kl): Delete.
8881 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8882 define expands, using...
8883 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8884 instructions for MIPS16.
8885 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8886 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8888 2014-01-25 Walter Lee <walt@tilera.com>
8890 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8894 2014-01-25 Walter Lee <walt@tilera.com>
8896 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8897 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8899 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8901 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8904 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8906 * print-rtl.c (in_call_function_usage): New var.
8907 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8908 EXPR_LIST mode as mode and not as reg note name.
8911 * cfgloopmanip.c (copy_loop_info): If
8912 loop->warned_aggressive_loop_optimizations, make sure
8913 the flag is set in target loop too.
8915 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8917 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8919 * builtins.def: Likewise.
8920 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8921 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8922 * ira.c (ira_setup_eliminable_regset): Likewise.
8923 * omp-low.c (gate_expand_omp): Likewise.
8924 (execute_lower_omp): Likewise.
8925 (diagnose_sb_0): Likewise.
8926 (gate_diagnose_omp_blocks): Likewise.
8927 (simd_clone_clauses_extract): Likewise.
8930 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8932 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8933 correction for little endian...
8934 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8937 2014-01-24 Jeff Law <law@redhat.com>
8939 PR tree-optimization/59919
8940 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8941 for non-returning calls.
8943 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8945 * common/config/aarch64/aarch64-common.c
8946 (aarch64_rewrite_mcpu): Handle multiple names.
8947 * config/aarch64/aarch64.h
8948 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8950 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8952 * input.c (add_file_to_cache_tab): Handle the case where fopen
8955 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8958 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8959 from push operand if code of push isn't PRE_DEC.
8961 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8964 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8965 -mquad-memory-atomic. Update -mquad-memory documentation to say
8966 it is only used for non-atomic loads/stores.
8968 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8969 -mquad-memory or -mquad-memory-atomic switches.
8971 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8972 -mquad-memory-atomic to ISA 2.07 support.
8974 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8975 to separate support of normal quad word memory operations (ldq, stq)
8976 from the atomic quad word memory operations.
8978 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8979 support to separate non-atomic quad word operations from atomic
8980 quad word operations. Disable non-atomic quad word operations in
8981 little endian mode so that we don't have to swap words after the
8982 load and before the store.
8983 (quad_load_store_p): Add comment about atomic quad word support.
8984 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8985 options printed with -mdebug=reg.
8987 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8988 -mquad-memory-atomic as the test for whether we have quad word
8989 atomic instructions.
8990 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8991 or -mp8-vector are used, allow byte/half-word atomic operations.
8993 * config/rs6000/sync.md (load_lockedti): Insure that the address
8994 is a proper indexed or indirect address for the lqarx instruction.
8995 On little endian systems, swap the hi/lo registers after the lqarx
8997 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8998 insure the address is valid for the lqarx instruction.
8999 (store_conditionalti): Insure that the address is a proper indexed
9000 or indirect address for the stqcrx. instruction. On little endian
9001 systems, swap the hi/lo registers before doing the stqcrx.
9003 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9004 insure the address is valid for the stqcrx. instruction.
9006 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9007 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9008 type of quad memory support is available.
9010 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9013 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9014 there is a danger of looping.
9016 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9018 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9019 force flag_ira_loop_pressure if set via command line.
9021 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9023 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9024 (ashr_simd): New builtin handling DI mode.
9025 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9026 (aarch64_sshr_simddi): New match pattern.
9027 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9028 (vshrd_n_s64): Likewise.
9029 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9031 2014-01-23 Nick Clifton <nickc@redhat.com>
9033 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9034 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9035 favour of mcu specific scripts.
9036 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9039 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9040 Alex Velenko <Alex.Velenko@arm.com>
9042 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9043 (vaddv_s16): Likewise.
9044 (vaddv_s32): Likewise.
9045 (vaddv_u8): Likewise.
9046 (vaddv_u16): Likewise.
9047 (vaddv_u32): Likewise.
9048 (vaddvq_s8): Likewise.
9049 (vaddvq_s16): Likewise.
9050 (vaddvq_s32): Likewise.
9051 (vaddvq_s64): Likewise.
9052 (vaddvq_u8): Likewise.
9053 (vaddvq_u16): Likewise.
9054 (vaddvq_u32): Likewise.
9055 (vaddvq_u64): Likewise.
9056 (vaddv_f32): Likewise.
9057 (vaddvq_f32): Likewise.
9058 (vaddvq_f64): Likewise.
9059 (vmaxv_f32): Likewise.
9060 (vmaxv_s8): Likewise.
9061 (vmaxv_s16): Likewise.
9062 (vmaxv_s32): Likewise.
9063 (vmaxv_u8): Likewise.
9064 (vmaxv_u16): Likewise.
9065 (vmaxv_u32): Likewise.
9066 (vmaxvq_f32): Likewise.
9067 (vmaxvq_f64): Likewise.
9068 (vmaxvq_s8): Likewise.
9069 (vmaxvq_s16): Likewise.
9070 (vmaxvq_s32): Likewise.
9071 (vmaxvq_u8): Likewise.
9072 (vmaxvq_u16): Likewise.
9073 (vmaxvq_u32): Likewise.
9074 (vmaxnmv_f32): Likewise.
9075 (vmaxnmvq_f32): Likewise.
9076 (vmaxnmvq_f64): Likewise.
9077 (vminv_f32): Likewise.
9078 (vminv_s8): Likewise.
9079 (vminv_s16): Likewise.
9080 (vminv_s32): Likewise.
9081 (vminv_u8): Likewise.
9082 (vminv_u16): Likewise.
9083 (vminv_u32): Likewise.
9084 (vminvq_f32): Likewise.
9085 (vminvq_f64): Likewise.
9086 (vminvq_s8): Likewise.
9087 (vminvq_s16): Likewise.
9088 (vminvq_s32): Likewise.
9089 (vminvq_u8): Likewise.
9090 (vminvq_u16): Likewise.
9091 (vminvq_u32): Likewise.
9092 (vminnmv_f32): Likewise.
9093 (vminnmvq_f32): Likewise.
9094 (vminnmvq_f64): Likewise.
9096 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9098 * config/aarch64/aarch64-simd.md
9099 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9100 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9101 (*aarch64_mul3_elt<mode>): Likewise.
9102 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9103 (*aarch64_mul3_elt_to_64v2df): Likewise.
9104 (*aarch64_mla_elt<mode>): Likewise.
9105 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9106 (*aarch64_mls_elt<mode>): Likewise.
9107 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9108 (*aarch64_fma4_elt<mode>): Likewise.
9109 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9110 (*aarch64_fma4_elt_to_64v2df): Likewise.
9111 (*aarch64_fnma4_elt<mode>): Likewise.
9112 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9113 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9114 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9115 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9116 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9117 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9118 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9119 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9120 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9122 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9124 * config/aarch64/aarch64-simd.md
9125 (aarch64_be_checked_get_lane<mode>): New define_expand.
9126 * config/aarch64/aarch64-simd-builtins.def
9127 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9128 New builtin definition.
9129 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9130 Use new safe be builtin.
9132 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9134 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9136 (aarch64_be_st1<mode>): Likewise.
9137 (aarch_ld1<VALL:mode>): Define_expand modified.
9138 (aarch_st1<VALL:mode>): Likewise.
9139 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9140 (UNSPEC_ST1): Likewise.
9142 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9144 * config/microblaze/microblaze.md: Add trap insn and attribute
9146 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9148 PR preprocessor/58580
9149 * input.h (location_get_source_line): Take an additional line_size
9151 (void diagnostics_file_cache_fini): Declare new function.
9152 * input.c (struct fcache): New type.
9153 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9154 New static constants.
9155 (diagnostic_file_cache_init, total_lines_num)
9156 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9157 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9158 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9159 (get_next_line, read_next_line, goto_next_line, read_line_num):
9160 New static function definitions.
9161 (diagnostic_file_cache_fini): New function.
9162 (location_get_source_line): Take an additional output line_len
9163 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9165 * diagnostic.c (diagnostic_finish): Call
9166 diagnostic_file_cache_fini.
9167 (adjust_line): Take an additional input parameter for the length
9168 of the line, rather than calculating it with strlen.
9169 (diagnostic_show_locus): Adjust the use of
9170 location_get_source_line and adjust_line with respect to their new
9171 signature. While displaying a line now, do not stop at the first
9172 null byte. Rather, display the zero byte as a space and keep
9173 going until we reach the size of the line.
9174 * Makefile.in: Add vec.o to OBJS-libcommon
9176 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9177 Ilya Tocar <ilya.tocar@intel.com>
9179 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9180 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9181 (__builtin_ia32_kmov16): Ditto.
9182 * config/i386/i386.md (UNSPEC_KMOV): New.
9185 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9187 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9188 (_mm512_storeu_si512): Ditto.
9190 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9193 * rtl.h (get_referenced_operands): Declare.
9194 * recog.c (get_referenced_operands): New function.
9195 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9196 operands have been referenced when recording LO_SUM references.
9198 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9200 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9202 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9204 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9205 Enable for generic and recent AMD targets.
9207 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9209 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9210 ARG_SIZE note when adjustment was eliminated.
9212 2014-01-22 Jeff Law <law@redhat.com>
9214 PR tree-optimization/59597
9215 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9216 in file. Accept new argument REGISTERING and use it to modify
9217 dump output appropriately.
9218 (register_jump_thread): Corresponding changes.
9219 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9220 thread paths involving joiner blocks. Add code to dump cancelled
9221 jump threading paths.
9223 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9225 PR rtl-optimization/59477
9226 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9227 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9229 2014-01-22 Tom Tromey <tromey@redhat.com>
9231 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9233 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9235 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9237 PR rtl-optimization/59896
9238 * lra-constraints.c (process_alt_operands): Check unused note for
9239 matched operands of insn with no output reloads.
9241 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9243 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9244 (mips_move_from_gpr_cost): Likewise.
9246 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9248 PR rtl-optimization/59858
9249 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9250 ira_class_hard_regs_num.
9251 (process_alt_operands): Increase reject for dying matched operand.
9253 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9256 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9257 smaller than size, perform several stores or loads and stores
9258 at dst + count - size to store or copy all of size bytes, rather
9259 than just last modesize bytes.
9261 2014-01-20 DJ Delorie <dj@redhat.com>
9263 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9264 that CLOBBERs are REGs before propogating their values.
9266 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9269 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9270 (cgraph_inline_failed_type): New function.
9271 * cgraph.h (DEFCIFCODE): Add type.
9272 (cgraph_inline_failed_type_t): New enum.
9273 (cgraph_inline_failed_type): New prototype.
9274 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9275 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9276 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9277 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9278 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9279 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9280 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9281 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9282 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9283 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9284 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9285 OPTIMIZATION_MISMATCH.
9286 * tree-inline.c (expand_call_inline): Emit errors during
9287 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9289 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9292 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9293 mode attribute in insn output.
9295 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9297 * output.h (output_constant): Delete.
9298 * varasm.c (output_constant): Make private.
9300 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9302 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9304 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9307 * tree.h (fold_builtin_strcat): New prototype.
9308 * builtins.c (fold_builtin_strcat): No longer static. Add len
9309 argument, if non-NULL, don't call c_strlen. Optimize
9310 directly into __builtin_memcpy instead of __builtin_strcpy.
9311 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9312 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9314 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9316 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9317 for SImode_address_operand operands, having only a REG argument.
9319 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9321 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9322 loader name using mbig-endian.
9323 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9325 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9327 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9331 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9333 * config/aarch64/aarch64-protos.h
9334 (aarch64_cannot_change_mode_class_ptr): Declare.
9335 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9336 aarch64_cannot_change_mode_class_ptr): New.
9337 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9338 backend hook aarch64_cannot_change_mode_class.
9340 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9342 * common/config/aarch64/aarch64-common.c
9343 (aarch64_handle_option): Don't handle any option order logic here.
9344 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9345 selected_cpu, warn on architecture version mismatch.
9346 (aarch64_override_options): Fix parsing order for option strings.
9348 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9349 Iain Sandoe <iain@codesourcery.com>
9352 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9353 warning. Amend comment to reflect current functionality.
9355 2014-01-20 Richard Biener <rguenther@suse.de>
9358 * builtins.c (fold_builtin_strcat): Remove case better handled
9359 by tree-ssa-strlen.c.
9361 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9363 * config/aarch64/aarch64.opt
9364 (mcpu, march, mtune): Make case-insensitive.
9366 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9369 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9370 if operands[1] is a REG or ZERO_EXTEND of a REG.
9372 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9374 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9376 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9378 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9379 long non-pic millicode calls.
9381 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9383 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9385 2014-01-19 Kito Cheng <kito@0xlab.org>
9387 * builtins.c (expand_movstr): Check movstr expand done or fail.
9389 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9390 H.J. Lu <hongjiu.lu@intel.com>
9393 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9394 to DImode for zero-extended addresses.
9396 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9398 PR rtl-optimization/57763
9399 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9400 on the new indirect jump_insn and increment LABEL_NUSES (label).
9402 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9406 * config.gcc (x86_archs): New variable.
9407 (x86_64_archs): Likewise.
9408 (x86_cpus): Likewise.
9409 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9410 --with-arch/--with-cpu= options.
9411 Support --with-arch=/--with-cpu={nehalem,westmere,
9412 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9414 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9416 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9417 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9419 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9421 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9423 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9426 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9427 clear cpp_get_options (parse_in)->warn_unused_macros for
9428 ix86_target_macros_internal with cpp_define.
9430 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9432 * jump.c (delete_related_insns): Keep (use (insn))s.
9433 * reorg.c (redundant_insn): Check for barriers too.
9435 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9437 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9439 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9441 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9442 call to $$dyncall when TARGET_LONG_CALLS is true.
9444 2014-01-17 Jeff Law <law@redhat.com>
9446 * ree.c (combine_set_extension): Temporarily disable test for
9447 changing number of hard registers.
9449 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9452 * ipa-inline-analysis.c (inline_free_summary):
9453 Do not free summary of aliases.
9455 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9458 * gimplify.c (gimplify_expr): Use create_tmp_var
9459 instead of create_tmp_var_raw. If cond doesn't have
9460 integral type, don't add the IFN_ANNOTATE builtin at all.
9462 2014-01-17 Martin Jambor <mjambor@suse.cz>
9465 * ipa-cp.c (prev_edge_clone): New variable.
9466 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9467 Also resize prev_edge_clone vector.
9468 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9469 (ipcp_edge_removal_hook): New function.
9470 (ipcp_driver): Register ipcp_edge_removal_hook.
9472 2014-01-17 Andrew Pinski <apinski@cavium.com>
9473 Steve Ellcey <sellcey@mips.com>
9476 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9479 2014-01-17 Jeff Law <law@redhat.com>
9482 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9483 so that pass_ccp runs first.
9485 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9487 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9488 (ix86_adjust_cost): Use !TARGET_XXX.
9489 (do_reorder_for_imul): Likewise.
9490 (swap_top_of_ready_list): Likewise.
9491 (ix86_sched_reorder): Likewise.
9493 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9495 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9496 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9497 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9498 (intel_memset): New. Duplicate slm_memset.
9499 (intel_cost): New. Duplicate slm_cost.
9500 (m_INTEL): New macro.
9501 (processor_target_table): Add "intel".
9502 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9503 with PROCESSOR_INTEL for "intel".
9504 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9505 PROCESSOR_SILVERMONT.
9506 (ix86_issue_rate): Likewise.
9507 (ix86_adjust_cost): Likewise.
9508 (ia32_multipass_dfa_lookahead): Likewise.
9509 (swap_top_of_ready_list): Likewise.
9510 (ix86_sched_reorder): Likewise.
9511 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9512 instead of TARGET_OPT_AGU.
9513 * config/i386/i386.h (TARGET_INTEL): New.
9514 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9515 (processor_type): Add PROCESSOR_INTEL.
9516 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9517 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9519 2014-01-17 Marek Polacek <polacek@redhat.com>
9522 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9525 2014-01-17 Richard Biener <rguenther@suse.de>
9527 PR tree-optimization/46590
9528 * opts.c (default_options_table): Add entries for
9529 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9530 all enabled at -O1 but not for -Og.
9531 * common.opt (fbranch-count-reg): Remove Init(1).
9532 (fmove-loop-invariants): Likewise.
9533 (ftree-pta): Likewise.
9535 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9537 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9538 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9539 decls to at least the GCC 4.8 used alignments.
9542 * tree-nested.c (convert_nonlocal_reference_stmt,
9543 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9544 of GIMPLE_BIND stmts, adjust associated decls.
9546 2014-01-17 Richard Biener <rguenther@suse.de>
9548 PR tree-optimization/46590
9549 * vec.h (vec<>::bseach): New member function implementing
9550 binary search according to C89 bsearch.
9551 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9552 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9553 bitmap pointer again. Make accesses_in_loop a flat array.
9554 (mem_ref_obstack): New global.
9555 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9556 (mark_ref_stored): Likewise.
9557 (ref_indep_loop_p_2): Likewise.
9558 (set_ref_stored_in_loop): New helper function.
9559 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9560 (memref_free): Adjust.
9561 (record_mem_ref_loc): Simplify.
9562 (gather_mem_refs_stmt): Adjust.
9563 (sort_locs_in_loop_postorder_cmp): New function.
9564 (analyze_memory_references): Sort accesses_in_loop after
9565 loop postorder number.
9566 (find_ref_loc_in_loop_cmp): New function.
9567 (for_all_locs_in_loop): Find relevant cluster of locs in
9568 accesses_in_loop and iterate without recursion.
9569 (execute_sm): Avoid uninit warning.
9570 (struct ref_always_accessed): Simplify.
9571 (ref_always_accessed::operator ()): Likewise.
9572 (ref_always_accessed_p): Likewise.
9573 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9574 loop postorder numbers here.
9575 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9578 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9581 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9582 on decls for which assemble_alias has been called.
9584 2014-01-17 Nick Clifton <nickc@redhat.com>
9586 * config/msp430/msp430.opt: (mcpu): New option.
9587 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9588 (msp430_option_override): Parse target_cpu. If the MCU name
9589 matches a generic string, clear target_mcu.
9590 (msp430_attr): Allow numeric interrupt values up to 63.
9591 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9592 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9594 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9596 * config/msp430/msp430.md (popm): Use %J rather than %I.
9597 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9598 (addhi_cy_i): Use immediate_operand for operand 2.
9599 * doc/invoke.texi: Document -mcpu option.
9601 2014-01-17 Richard Biener <rguenther@suse.de>
9603 PR rtl-optimization/38518
9604 * df.h (df_analyze_loop): Declare.
9605 * df-core.c: Include cfgloop.h.
9606 (df_analyze_1): Split out main part of df_analyze.
9607 (df_analyze): Adjust.
9608 (loop_inverted_post_order_compute): New function.
9609 (loop_post_order_compute): Likewise.
9610 (df_analyze_loop): New function avoiding whole-function
9612 * loop-invariant.c (find_defs): Use df_analyze_loop.
9613 (find_invariants): Adjust.
9614 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9616 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9618 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9619 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9621 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9623 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9624 traversal when removing references.
9626 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9629 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9631 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9634 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9635 pushing a reload for an autoinc when we had previously reloaded an
9636 inner part of the address.
9638 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9640 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9642 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9643 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9644 when not giving up or versioning for alias only because of
9646 (vect_analyze_data_ref_dependences): Set to true.
9647 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9649 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9650 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9654 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9657 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9658 operand 0 predicate for gathers, use a new pseudo as subtarget.
9660 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9663 * lra-constraints.c (process_alt_operands): Add printing debug info.
9664 Check absence of input/output reloads for matched operands too.
9666 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9668 PR rtl-optimization/59835
9669 * ira.c (ira_init_register_move_cost): Increase cost for
9672 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9674 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9676 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9679 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9680 non-register objects. Use gen_(high/low)part more consistently.
9683 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9686 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9687 endian support, remove tests for WORDS_BIG_ENDIAN.
9688 (p8_mfvsrd_3_<mode>): Likewise.
9689 (reload_gpr_from_vsx<mode>): Likewise.
9690 (reload_gpr_from_vsxsf): Likewise.
9691 (p8_mfvsrd_4_disf): Likewise.
9693 2014-01-16 Richard Biener <rguenther@suse.de>
9695 PR rtl-optimization/46590
9696 * lcm.c (compute_antinout_edge): Use postorder iteration.
9697 (compute_laterin): Use inverted postorder iteration.
9699 2014-01-16 Nick Clifton <nickc@redhat.com>
9702 * varasm.c (output_constant): Return the number of bytes actually
9704 (output_constructor_array_range): Update the field size with the
9705 number of bytes emitted by output_constant.
9706 (output_constructor_regular_field): Likewise. Also do not
9707 complain if the total number of bytes emitted is now greater
9708 than the expected fieldpos.
9709 * output.h (output_constant): Update prototype and descriptive comment.
9711 2014-01-16 Marek Polacek <polacek@redhat.com>
9714 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9715 it is error_mark_node.
9717 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9719 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9720 VALID_AVX256_REG_OR_OI_MODE.
9722 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9724 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9725 current procedure should be profiled.
9727 2014-01-15 Andrew Pinski <apinski@cavium.com>
9729 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9730 of moving from/to the STACK_REG register class.
9732 2014-01-15 Richard Henderson <rth@redhat.com>
9735 * reginfo.c (global_regs_decl): Globalize.
9736 * rtl.h (global_regs_decl): Declare.
9737 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9738 reserved via global_regs.
9740 2014-01-15 Teresa Johnson <tejohnson@google.com>
9742 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9744 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9746 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9747 and vmulosh rather than call gen_vec_widen_smult_*.
9748 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9749 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9750 (vec_widen_smult_even_v16qi): Likewise.
9751 (vec_widen_umult_even_v8hi): Likewise.
9752 (vec_widen_smult_even_v8hi): Likewise.
9753 (vec_widen_umult_odd_v16qi): Likewise.
9754 (vec_widen_smult_odd_v16qi): Likewise.
9755 (vec_widen_umult_odd_v8hi): Likewise.
9756 (vec_widen_smult_odd_v8hi): Likewise.
9757 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9758 vmuloub rather than call gen_vec_widen_umult_*.
9759 (vec_widen_umult_lo_v16qi): Likewise.
9760 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9761 vmulosb rather than call gen_vec_widen_smult_*.
9762 (vec_widen_smult_lo_v16qi): Likewise.
9763 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9764 rather than call gen_vec_widen_umult_*.
9765 (vec_widen_umult_lo_v8hi): Likewise.
9766 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9767 rather than call gen_vec_widen_smult_*.
9768 (vec_widen_smult_lo_v8hi): Likewise.
9770 2014-01-15 Jeff Law <law@redhat.com>
9772 PR tree-optimization/59747
9773 * ree.c (find_and_remove_re): Properly handle case where a second
9774 eliminated extension requires widening a copy created for elimination
9775 of a prior extension.
9776 (combine_set_extension): Ensure that the number of hard regs needed
9777 for a destination register does not change when we widen it.
9779 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9781 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9782 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9783 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9784 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9785 (avr-*-rtems*): Likewise.
9786 (bfin*-rtems*): Likewise.
9787 (moxie-*-rtems*): Likewise.
9788 (h8300-*-rtems*): Likewise.
9789 (i[34567]86-*-rtems*): Likewise.
9790 (lm32-*-rtems*): Likewise.
9791 (m32r-*-rtems*): Likewise.
9792 (m68k-*-rtems*): Likewise.
9793 (microblaze*-*-rtems*): Likewise.
9794 (mips*-*-rtems*): Likewise.
9795 (powerpc-*-rtems*): Likewise.
9796 (sh-*-rtems*): Likewise.
9797 (sparc-*-rtems*): Likewise.
9798 (sparc64-*-rtems*): Likewise.
9799 (v850-*-rtems*): Likewise.
9800 (m32c-*-rtems*): Likewise.
9802 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9804 PR rtl-optimization/59511
9805 * ira.c (ira_init_register_move_cost): Use memory costs for some
9806 cases of register move cost calculations.
9807 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9808 instead of BB frequency.
9809 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9810 * lra-assigns.c (find_hard_regno_for): Ditto.
9812 2014-01-15 Richard Biener <rguenther@suse.de>
9814 PR tree-optimization/59822
9815 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9816 (vectorizable_load): Use it to hoist defs of uses of invariant
9817 loads out of the loop.
9819 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9820 Kugan Vivekanandarajah <kuganv@linaro.org>
9823 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9826 2014-01-15 Richard Biener <rguenther@suse.de>
9828 PR rtl-optimization/59802
9829 * lcm.c (compute_available): Use inverted postorder to seed
9830 the initial worklist.
9832 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9835 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9836 ADDR_REGS for invalid symrefs in non-PIC code.
9838 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9841 * builtins.c (determine_block_size): Initialize *probable_max_size
9842 even if len_rtx is CONST_INT.
9844 2014-01-14 Andrew Pinski <apinski@cavium.com>
9846 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9847 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9848 (cortexa53_tunings): Likewise.
9849 (aarch64_sched_issue_rate): New function.
9850 (TARGET_SCHED_ISSUE_RATE): Define.
9852 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9854 * ira-costs.c (find_costs_and_classes): Add missed
9855 ira_init_register_move_cost_if_necessary.
9857 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9860 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9862 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9865 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9866 to indicate if type is used for function return value. Warn ABI
9867 change if the vector mode isn't available for function return value.
9868 (ix86_function_arg_advance): Pass false to type_natural_mode.
9869 (ix86_function_arg): Likewise.
9870 (ix86_gimplify_va_arg): Likewise.
9871 (function_arg_32): Don't warn ABI change.
9872 (ix86_function_value): Pass true to type_natural_mode.
9873 (ix86_return_in_memory): Likewise.
9874 (ix86_struct_value_rtx): Removed.
9875 (TARGET_STRUCT_VALUE_RTX): Likewise.
9877 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9879 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9880 converting a conditional jump into a conditional return.
9882 2014-01-14 Richard Biener <rguenther@suse.de>
9884 PR tree-optimization/58921
9885 PR tree-optimization/59006
9886 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9887 hoisting invariant stmts.
9888 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9889 invariant loads on the preheader edge if possible.
9891 2014-01-14 Joey Ye <joey.ye@arm.com>
9893 * doc/plugin.texi (Building GCC plugins): Update to C++.
9895 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9897 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9898 (_mm_rcp28_round_ss): Ditto.
9899 (_mm_rsqrt28_round_sd): Ditto.
9900 (_mm_rsqrt28_round_ss): Ditto.
9901 (_mm_rcp28_sd): Ditto.
9902 (_mm_rcp28_ss): Ditto.
9903 (_mm_rsqrt28_sd): Ditto.
9904 (_mm_rsqrt28_ss): Ditto.
9905 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9906 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9907 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9908 (IX86_BUILTIN_RCP28SD): Ditto.
9909 (IX86_BUILTIN_RCP28SS): Ditto.
9910 (IX86_BUILTIN_RSQRT28SD): Ditto.
9911 (IX86_BUILTIN_RSQRT28SS): Ditto.
9912 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9913 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9914 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9915 (ix86_expand_special_args_builtin): Expand new FTYPE.
9916 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9917 (srcp14<mode>): Make insn unary.
9918 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9919 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9920 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9921 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9922 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9923 Fix rounding: make it SAE only.
9924 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9926 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9928 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9929 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9930 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9931 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9932 (round_saeonly_mask_scalar_operand4): Ditto.
9933 (round_saeonly_mask_scalar_op3): Ditto.
9934 (round_saeonly_mask_scalar_op4): Ditto.
9936 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9938 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9939 Implement -maltivec=be for vec_insert and vec_extract.
9941 2014-01-10 DJ Delorie <dj@redhat.com>
9943 * config/msp430/msp430.md (call_internal): Don't allow memory
9944 references with SP as the base register.
9945 (call_value_internal): Likewise.
9946 * config/msp430/constraints.md (Yc): New. For memory references
9947 that don't use SP as a base register.
9949 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9950 "an integer without a # prefix"
9951 * config/msp430/msp430.md (epilogue_helper): Use it.
9953 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9956 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9957 AVX512F gather builtins.
9958 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9959 on gather decls with INTEGER_TYPE masktype.
9960 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9961 directly into the builtin rather than hoisting it before loop.
9963 PR tree-optimization/59387
9964 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9965 (scev_const_prop): If folded_casts and type has undefined overflow,
9966 use force_gimple_operand instead of force_gimple_operand_gsi and
9967 for each added stmt if it is assign with
9968 arith_code_with_undefined_signed_overflow, call
9969 rewrite_to_defined_overflow.
9970 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9971 gimple-fold.h instead.
9972 (arith_code_with_undefined_signed_overflow,
9973 rewrite_to_defined_overflow): Moved to ...
9974 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9975 rewrite_to_defined_overflow): ... here. No longer static.
9976 Include gimplify-me.h.
9977 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9978 rewrite_to_defined_overflow): New prototypes.
9980 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9982 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9984 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9986 * builtins.c (get_object_alignment_2): Minor tweak.
9987 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9989 2014-01-13 Christian Bruel <christian.bruel@st.com>
9991 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9992 optimized non constant lengths.
9994 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9997 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9998 load as __atomic_load_N if possible.
10000 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10002 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10004 (rs6000_expand_builtin): Adjust call.
10006 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10009 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10010 * config/rs6000/rs6000.c: Include target-globals.h.
10011 (rs6000_set_current_function): Instead of doing target_reinit
10012 unconditionally, use save_target_globals_default_opts and
10013 restore_target_globals.
10015 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10017 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10018 (rs6000_expand_builtin): Handle mffs and mtfsf.
10019 (rs6000_init_builtins): Define mffs and mtfsf.
10020 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10021 (rs6000_mffs): New pattern.
10022 (rs6000_mtfsf): New pattern.
10024 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10026 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10027 Start narrowing with START. Apply candidate-use pair
10028 and check overall cost in narrowing.
10029 (iv_ca_prune): Pass new argument.
10031 2014-01-10 Jeff Law <law@redhat.com>
10033 PR middle-end/59743
10034 * ree.c (combine_reaching_defs): Ensure the defining statement
10035 occurs before the extension when optimizing extensions with
10036 different source and destination hard registers.
10038 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10041 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10042 vtables into the type inheritance graph.
10044 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10046 PR rtl-optimization/59754
10047 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10048 modes in the REGNO != REGNO case.
10050 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10052 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10054 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10056 PR tree-optimization/59745
10057 * tree-predcom.c (tree_predictive_commoning_loop): Call
10058 free_affine_expand_cache if giving up because components is NULL.
10060 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10061 GC in payload of target_globals struct instead of allocating them on
10062 the heap and the larger structs separately using GC.
10063 * target-globals.h (struct target_globals): Make regs, hard_regs,
10064 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10065 of GTY((skip)) and change type to void *.
10066 (reset_target_globals): Cast loads from those fields to corresponding
10069 2014-01-10 Steve Ellcey <sellcey@mips.com>
10072 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10073 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10074 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10076 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10079 * aarch64-modes.def (CC_Zmode): New flags mode.
10080 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10081 represents an equality.
10082 (aarch64_get_condition_code): Handle CC_Zmode.
10083 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10085 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10087 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10088 extraction in good case.
10090 2014-01-10 Richard Biener <rguenther@suse.de>
10092 PR tree-optimization/59374
10093 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10094 checking after SLP discovery. Mark stmts not participating
10095 in any SLP instance properly.
10097 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10099 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10100 when handling a SET rtx.
10102 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10104 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10105 (cortex-a57): Likewise.
10106 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10108 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10110 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10111 non-iwmmxt builtins.
10113 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10117 * ipa-devirt.c record_target_from_binfo): Take as argument
10118 stack of binfos and lookup matching one for virtual inheritance.
10119 (possible_polymorphic_call_targets_1): Update.
10121 2014-01-10 Huacai Chen <chenhc@lemote.com>
10123 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10124 kernel strings for Loongson-2E/2F/3A.
10126 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10128 PR middle-end/59670
10129 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10130 is_gimple_call before calling gimple_call_internal_p.
10132 2014-01-09 Steve Ellcey <sellcey@mips.com>
10134 * Makefile.in (TREE_FLOW_H): Remove.
10135 (TREE_SSA_H): Add file names from tree-flow.h.
10136 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10137 * tree.h: Remove tree-flow.h reference.
10138 * hash-table.h: Remove tree-flow.h reference.
10139 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10140 reference with tree-ssa-loop.h.
10142 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10144 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10145 default element-order behavior for -maltivec.
10146 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10147 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10148 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10149 when targeting big endian, at least for now.
10150 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10152 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10154 PR middle-end/47735
10155 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10156 var satisfies use_register_for_decl, just take into account type
10157 alignment, rather than decl alignment.
10159 PR tree-optimization/59622
10160 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10161 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10162 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10163 Don't devirtualize for inplace at all. For targets.length () == 1,
10164 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10166 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10168 * config/i386/i386.md (cpu): Remove the unused btver1.
10170 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10172 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10174 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10177 * tree-core.h (struct target_globals): New forward declaration.
10178 (struct tree_target_option): Add globals field.
10179 * tree.h (TREE_TARGET_GLOBALS): Define.
10180 (prepare_target_option_nodes_for_pch): New prototype.
10181 * target-globals.h (struct target_globals): Define even if
10182 !SWITCHABLE_TARGET.
10183 * tree.c (prepare_target_option_node_for_pch,
10184 prepare_target_option_nodes_for_pch): New functions.
10185 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10186 * config/i386/i386.c: Include target-globals.h.
10187 (ix86_set_current_function): Instead of doing target_reinit
10188 unconditionally, use save_target_globals_default_opts and
10189 restore_target_globals.
10191 2014-01-09 Richard Biener <rguenther@suse.de>
10193 PR tree-optimization/59715
10194 * tree-cfg.h (split_critical_edges): Declare.
10195 * tree-cfg.c (split_critical_edges): Export.
10196 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10198 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10200 * cfgexpand.c (expand_stack_vars): Optionally disable
10201 asan stack protection.
10202 (expand_used_vars): Likewise.
10203 (partition_stack_vars): Likewise.
10204 * asan.c (asan_emit_stack_protection): Optionally disable
10205 after return stack usage.
10206 (instrument_derefs): Optionally disable memory access instrumentation.
10207 (instrument_builtin_call): Likewise.
10208 (instrument_strlen_call): Likewise.
10209 (asan_protect_global): Optionally disable global variables protection.
10210 * doc/invoke.texi: Added doc for new options.
10211 * params.def: Added new options.
10212 * params.h: Likewise.
10214 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10216 PR rtl-optimization/59724
10217 * ifcvt.c (cond_exec_process_if_block): Don't call
10218 flow_find_head_matching_sequence with 0 longest_match.
10219 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10220 non-active insns if !stop_after.
10221 (try_head_merge_bb): Revert 2014-01-07 changes.
10223 2014-01-08 Jeff Law <law@redhat.com>
10225 * ree.c (get_sub_rtx): New function, extracted from...
10226 (merge_def_and_ext): Here.
10227 (combine_reaching_defs): Use get_sub_rtx.
10229 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10231 * cgraph.h (varpool_variable_node): Do not choke on null node.
10233 2014-01-08 Catherine Moore <clm@codesourcery.com>
10235 * config/mips/mips.md (simple_return): Attempt to use JRC
10237 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10239 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10241 PR rtl-optimization/59137
10242 * reorg.c (steal_delay_list_from_target): Call update_block for
10244 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10246 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10248 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10249 two duplicate entries.
10251 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10254 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10256 * config/mips/mips.c (mips_truncated_op_cost): New function.
10257 (mips_rtx_costs): Adjust test for BADDU.
10258 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10260 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10262 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10263 (*baddu_si): ...this new pattern.
10265 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10268 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10270 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10272 PR middle-end/57748
10273 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10275 (expand_expr, expand_normal): Adjust.
10276 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10277 inner_reference_p. Use inner_reference_p to expand inner references.
10278 (store_expr): Adjust.
10279 * cfgexpand.c (expand_call_stmt): Adjust.
10281 2014-01-08 Rong Xu <xur@google.com>
10283 * gcov-io.c (gcov_var): Move from gcov-io.h.
10284 (gcov_position): Ditto.
10285 (gcov_is_error): Ditto.
10286 (gcov_rewrite): Ditto.
10287 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10288 only part to libgcc/libgcov.h.
10290 2014-01-08 Marek Polacek <polacek@redhat.com>
10292 PR middle-end/59669
10293 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10295 2014-01-08 Marek Polacek <polacek@redhat.com>
10298 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10300 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10302 PR rtl-optimization/59649
10303 * stor-layout.c (get_mode_bounds): For BImode return
10304 0 and STORE_FLAG_VALUE.
10306 2014-01-08 Richard Biener <rguenther@suse.de>
10308 PR middle-end/59630
10309 * gimple.h (is_gimple_builtin_call): Remove.
10310 (gimple_builtin_call_types_compatible_p): New.
10311 (gimple_call_builtin_p): New overload.
10312 * gimple.c (is_gimple_builtin_call): Remove.
10313 (validate_call): Rename to ...
10314 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10315 check return types.
10316 (validate_type): New static function.
10317 (gimple_call_builtin_p): New overload and adjust.
10318 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10319 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10320 (gimple_fold_stmt_to_constant_1): Likewise.
10321 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10323 2014-01-08 Richard Biener <rguenther@suse.de>
10325 PR middle-end/59471
10326 * gimplify.c (gimplify_expr): Gimplify register-register type
10327 VIEW_CONVERT_EXPRs to separate stmts.
10329 2014-01-07 Jeff Law <law@redhat.com>
10331 PR middle-end/53623
10332 * ree.c (combine_set_extension): Handle case where source
10333 and destination registers in an extension insn are different.
10334 (combine_reaching_defs): Allow source and destination registers
10335 in extension to be different under limited circumstances.
10336 (add_removable_extension): Remove restriction that the
10337 source and destination registers in the extension are the same.
10338 (find_and_remove_re): Emit a copy from the extension's
10339 destination to its source after the defining insn if
10340 the source and destination registers are different.
10342 PR middle-end/59285
10343 * ifcvt.c (merge_if_block): If we are merging a block with more than
10344 one successor with a block with no successors, remove any BARRIER
10345 after the second block.
10347 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10349 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10351 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10354 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10355 for 14-bit register offsets when INT14_OK_STRICT is false.
10357 2014-01-07 Roland Stigge <stigge@antcom.de>
10358 Michael Meissner <meissner@linux.vnet.ibm.com>
10361 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10362 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10364 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10366 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10369 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10371 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10372 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10373 rtx is const0_rtx or not.
10375 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10378 * target-globals.c (save_target_globals): Remove this_fn_optab
10380 * toplev.c: Include optabs.h.
10381 (target_reinit): Temporarily restore the global options if another
10382 set of options are in force.
10384 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10386 PR rtl-optimization/58668
10387 * cfgcleanup.c (flow_find_cross_jump): Don't count
10388 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10389 to determine what is counted.
10390 (flow_find_head_matching_sequence): Use active_insn_p to determine
10392 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10394 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10395 determine what is counted.
10397 PR tree-optimization/59643
10398 * tree-predcom.c (split_data_refs_to_components): If one dr is
10399 read and one write, determine_offset fails and the write isn't
10400 in the bad component, just put the read into the bad component.
10402 2014-01-07 Mike Stump <mikestump@comcast.net>
10403 Jakub Jelinek <jakub@redhat.com>
10406 * tree-core.h (struct tree_optimization_option): Change optabs
10407 type from unsigned char * to void *.
10408 * optabs.c (init_tree_optimization_optabs): Adjust
10409 TREE_OPTIMIZATION_OPTABS initialization.
10411 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10414 * config/i386/i386.h (struct machine_function): Add
10415 no_drap_save_restore field.
10416 * config/i386/i386.c (ix86_save_reg): Use
10417 !cfun->machine->no_drap_save_restore instead of
10418 crtl->stack_realign_needed.
10419 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10420 this function clears frame_pointer_needed. Set
10421 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10422 and DRAP reg is needed.
10424 2014-01-06 Marek Polacek <polacek@redhat.com>
10427 * doc/implement-c.texi: Mention that other integer types are
10428 permitted as bit-field types in strictly conforming mode.
10430 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10432 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10433 is newly allocated.
10435 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10437 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10439 2014-01-06 Martin Jambor <mjambor@suse.cz>
10442 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10444 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10446 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10450 * var-tracking.c (add_stores): Preserve the value of the source even if
10451 we don't record the store.
10453 2014-01-06 Terry Guo <terry.guo@arm.com>
10455 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10457 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10460 * config/darwin.c (darwin_function_section): Adjust return values to
10461 correspond to optimisation changes made in r206070.
10463 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10465 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10466 from prefetch_block tune setting.
10467 (nocona_cost): Correct size of prefetch block to 64.
10469 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10471 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10472 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10473 used to save the static chain register in the computation of the offset
10474 from which the FP registers need to be restored.
10476 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10478 PR tree-optimization/59519
10479 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10480 ICE if get_current_def (current_new_name) is already non-NULL, as long
10481 as it is a phi result of some other phi in *new_exit_bb that has
10484 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10485 or vmovdqu* for misaligned_operand.
10486 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10487 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10488 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10489 aligned_mem for AVX512F masked aligned load and store builtins and for
10490 non-temporal moves.
10492 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10494 PR tree-optimization/59651
10495 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10496 Address range for negative step should be added by TYPE_SIZE_UNIT.
10498 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10500 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10501 overlapping registers also for registers other than the stack pointer.
10503 2014-01-03 Marek Polacek <polacek@redhat.com>
10506 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10509 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10512 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10515 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10516 (push splitter): Use <P:MODE_SIZE> instead of
10517 GET_MODE_SIZE (<P:MODE>mode).
10518 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10519 (mov -1, reg peephole2): Likewise.
10520 * config/i386/sse.md (*mov<mode>_internal,
10521 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10522 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10523 *<code><mode>3, *andnot<mode>3<mask_name>,
10524 <mask_codefor><code><mode>3<mask_name>): Likewise.
10525 * config/i386/subst.md (mask_mode512bit_condition,
10526 sd_mask_mode512bit_condition): Likewise.
10528 2014-01-02 Xinliang David Li <davidxl@google.com>
10530 PR tree-optimization/59303
10531 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10532 (dump_predicates): Better output format.
10533 (pred_equal_p): New function.
10534 (is_neq_relop_p): Ditto.
10535 (is_neq_zero_form_p): Ditto.
10536 (pred_expr_equal_p): Ditto.
10537 (pred_neg_p): Ditto.
10538 (simplify_pred): Ditto.
10539 (simplify_preds_2): Ditto.
10540 (simplify_preds_3): Ditto.
10541 (simplify_preds_4): Ditto.
10542 (simplify_preds): Ditto.
10543 (push_pred): Ditto.
10544 (push_to_worklist): Ditto.
10545 (get_pred_info_from_cmp): Ditto.
10546 (is_degenerated_phi): Ditto.
10547 (normalize_one_pred_1): Ditto.
10548 (normalize_one_pred): Ditto.
10549 (normalize_one_pred_chain): Ditto.
10550 (normalize_preds): Ditto.
10551 (normalize_cond_1): Remove function.
10552 (normalize_cond): Ditto.
10553 (is_gcond_subset_of): Ditto.
10554 (is_subset_of_any): Ditto.
10555 (is_or_set_subset_of): Ditto.
10556 (is_and_set_subset_of): Ditto.
10557 (is_norm_cond_subset_of): Ditto.
10558 (pred_chain_length_cmp): Ditto.
10559 (convert_control_dep_chain_into_preds): Type change.
10560 (find_predicates): Ditto.
10561 (find_def_preds): Ditto.
10562 (destroy_predicates_vecs): Ditto.
10563 (find_matching_predicates_in_rest_chains): Ditto.
10564 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10565 (is_pred_expr_subset): Ditto.
10566 (is_pred_chain_subset_of): Ditto.
10567 (is_included_in): Ditto.
10568 (is_superset_of): Ditto.
10570 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10572 Update copyright years.
10574 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10576 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10577 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10578 config/arc/arc.md, config/arc/arc.opt,
10579 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10580 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10581 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10582 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10583 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10584 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10585 vtable-verify.c, vtable-verify.h: Use the standard form for the
10588 2014-01-02 Tobias Burnus <burnus@net-b.de>
10590 * gcc.c (process_command): Update copyright notice dates.
10591 * gcov-dump.c: Ditto.
10593 * doc/cpp.texi: Bump @copying's copyright year.
10594 * doc/cppinternals.texi: Ditto.
10595 * doc/gcc.texi: Ditto.
10596 * doc/gccint.texi: Ditto.
10597 * doc/gcov.texi: Ditto.
10598 * doc/install.texi: Ditto.
10599 * doc/invoke.texi: Ditto.
10601 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10603 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10605 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10607 * config/i386/sse.md (*mov<mode>_internal): Guard
10608 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10610 PR rtl-optimization/59647
10611 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10612 new_rtx into UNSIGNED_FLOAT rtxes.
10614 Copyright (C) 2014 Free Software Foundation, Inc.
10616 Copying and distribution of this file, with or without modification,
10617 are permitted in any medium without royalty provided the copyright
10618 notice and this notice are preserved.