1 2013-09-08 Tom de Vries <tom@codesourcery.com>
4 * except.c (build_must_not_throw_expr): Handle
7 2013-09-06 Jan Hubicka <jh@suse.cz>
9 * cgraphunit.c (expand_thunk): Get body before touching arguments.
10 * lto-streamer-out.c: Stream thunks, too.
11 * lto-streamer-in.c (input_function): Pop cfun here
12 (lto_read_body): Instead of here.
14 2013-09-06 Caroline Tice <cmtice@google.com>
16 * doc/install.texi: Add documentation for the --enable-vtable-verify
17 and the --disable-libvtv configure options.
19 2013-09-06 Jeff Law <law@redhat.com>
21 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
22 edge implied equivalences into successor phis.
24 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
26 * resource.c (mark_referenced_resources): Handle COND_EXEC.
28 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
30 * resource.c (mark_target_live_regs): Compute resources taking
31 into account if a call is predicated or not.
33 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
35 * toplev.c (output_stack_usage): Be prepared for suffixes created by
36 the compiler in the function names.
38 2013-09-06 Jan Hubicka <jh@suse.cz>
41 * ipa-inline.c (has_caller_p): New function.
42 (want_inline_function_to_all_callers_p): Use it.
43 (sum_callers, inline_to_all_callers): Break out from ...
44 (ipa_inline): ... here.
46 2013-09-06 Jan Hubicka <jh@suse.cz>
48 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
49 only when AVX is enabled.
51 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
53 * config/aarch64/aarch64.md
54 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
55 is fpsimd_<load/store>2.
56 (load_pair<mode>): Likewise.
57 (store_pair<mode>): Likewise.
59 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
61 * config/arm/types.md (type): Add "mrs" type.
62 * config/aarch64/aarch64.md
63 (aarch64_load_tp_hard): Make type "mrs".
65 (load_tp_hard): Make type "mrs".
66 * config/arm/cortex-a15.md: Update with new attributes.
67 * config/arm/cortex-a5.md: Update with new attributes.
68 * config/arm/cortex-a53.md: Update with new attributes.
69 * config/arm/cortex-a7.md: Update with new attributes.
70 * config/arm/cortex-a8.md: Update with new attributes.
71 * config/arm/cortex-a9.md: Update with new attributes.
72 * config/arm/cortex-m4.md: Update with new attributes.
73 * config/arm/cortex-r4.md: Update with new attributes.
74 * config/arm/fa526.md: Update with new attributes.
75 * config/arm/fa606te.md: Update with new attributes.
76 * config/arm/fa626te.md: Update with new attributes.
77 * config/arm/fa726te.md: Update with new attributes.
79 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
81 * config/aarch64/aarch64.md
82 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
83 (*movtf_aarch64): Likewise.
85 (thumb1_movdi_insn): Use "multiple" for type where more than one
86 instruction is used for a move.
87 (*arm32_movhf): Likewise.
88 (*thumb_movdf_insn): Likewise.
90 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
92 * config/arm/types.md (type): Rename fcpys to fmov.
94 (*arm_movsi_vfp): Rename type fcpys as fmov.
95 (*thumb2_movsi_vfp): Likewise
96 (*movhf_vfp_neon): Likewise
97 (*movhf_vfp): Likewise
98 (*movsf_vfp): Likewise
99 (*thumb2_movsf_vfp): Likewise
100 (*movsfcc_vfp): Likewise
101 (*thumb2_movsfcc_vfp): Likewise
102 * config/aarch64/aarch64-simd.md
103 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
104 * config/aarch64/aarch64.md
105 (*movsi_aarch64): Replace type mov_reg with fmovs.
106 (*movdi_aarch64): Likewise
107 (*movsf_aarch64): Likewise
108 (*movdf_aarch64): Likewise
110 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
111 * config/arm/iwmmxt.md
112 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
113 * config/arm/arm1020e.md: Update with new attributes.
114 * config/arm/cortex-a15-neon.md: Update with new attributes.
115 * config/arm/cortex-a5.md: Update with new attributes.
116 * config/arm/cortex-a53.md: Update with new attributes.
117 * config/arm/cortex-a7.md: Update with new attributes.
118 * config/arm/cortex-a8-neon.md: Update with new attributes.
119 * config/arm/cortex-a9.md: Update with new attributes.
120 * config/arm/cortex-m4-fpu.md: Update with new attributes.
121 * config/arm/cortex-r4f.md: Update with new attributes.
122 * config/arm/marvell-pj4.md: Update with new attributes.
123 * config/arm/vfp11.md: Update with new attributes.
125 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
127 * config/aarch64/aarch64.md
128 (*madd<mode>): Fix type attribute.
129 (*maddsi_uxtw): Likewise.
130 (*msub<mode>): Likewise.
131 (*msubsi_uxtw): Likewise.
132 (<su_optab>maddsidi4): Likewise.
133 (<su_optab>msubsidi4): Likewise.
135 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
137 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
138 * config/arm/arm.md (core_cycles): Remove fdiv.
140 (*sqrtsf2_vfp): Update for attribute changes.
141 (*sqrtdf2_vfp): Likewise.
142 * config/aarch64/aarch64.md:
143 (sqrt<mode>2): Update for attribute changes.
144 * config/arm/arm1020e.md: Update with new attributes.
145 * config/arm/cortex-a15-neon.md: Update with new attributes.
146 * config/arm/cortex-a5.md: Update with new attributes.
147 * config/arm/cortex-a53.md: Update with new attributes.
148 * config/arm/cortex-a7.md: Update with new attributes.
149 * config/arm/cortex-a8-neon.md: Update with new attributes.
150 * config/arm/cortex-a9.md: Update with new attributes.
151 * config/arm/cortex-m4-fpu.md: Update with new attributes.
152 * config/arm/cortex-r4f.md: Update with new attributes.
153 * config/arm/marvell-pj4.md: Update with new attributes.
154 * config/arm/vfp11.md: Update with new attributes.
156 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
158 * config/arm/types.md
159 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
160 * config/aarch64/aarch64.md
161 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
163 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
164 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
165 (float<GPI:mode><GPF:mode>2): Likewise.
167 (*truncsisf2_vfp): Update with new attributes.
168 (*truncsidf2_vfp): Likewise.
169 (fixuns_truncsfsi2): Likewise.
170 (fixuns_truncdfsi2): Likewise.
171 (*floatsisf2_vfp): Likewise.
172 (*floatsidf2_vfp): Likewise.
173 (floatunssisf2): Likewise.
174 (floatunssidf2): Likewise.
175 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
176 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
177 * config/arm/arm1020e.md: Update with new attributes.
178 * config/arm/cortex-a15-neon.md: Update with new attributes.
179 * config/arm/cortex-a5.md: Update with new attributes.
180 * config/arm/cortex-a53.md: Update with new attributes.
181 * config/arm/cortex-a7.md: Update with new attributes.
182 * config/arm/cortex-a8-neon.md: Update with new attributes.
183 * config/arm/cortex-a9.md: Update with new attributes.
184 * config/arm/cortex-m4-fpu.md: Update with new attributes.
185 * config/arm/cortex-r4f.md: Update with new attributes.
186 * config/arm/marvell-pj4.md: Update with new attributes.
187 * config/arm/vfp11.md: Update with new attributes.
189 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
191 * config/aarch64/arm_neon.h
192 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
193 (vqtbx<1,2,3,4><q>_s8): Likewise.
195 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
197 * config/arm/types.md: Add "no_insn", "multiple" and "untyped"
199 * config/arm/arm-fixed.md: Add type attribute to all insn
201 (add<mode>3): Add type attribute.
202 (add<mode>3): Likewise.
203 (usadd<mode>3): Likewise.
204 (ssadd<mode>3): Likewise.
205 (sub<mode>3): Likewise.
206 (sub<mode>3): Likewise.
207 (ussub<mode>3): Likewise.
208 (sssub<mode>3): Likewise.
209 (ssmulsa3): Likewise.
210 (usmulusa3): Likewise.
211 (arm_usatsihi): Likewise.
213 (*movdi_vfp): Add types for all instructions.
214 (*movdi_vfp_cortexa8): Likewise.
215 (*movhf_vfp_neon): Likewise.
216 (*movhf_vfp): Likewise.
217 (*movdf_vfp): Likewise.
218 (*thumb2_movdf_vfp): Likewise.
219 (*thumb2_movdfcc_vfp): Likewise.
220 * config/arm/arm.md: Add type attribute to all insn patterns.
221 (*thumb1_adddi3): Add type attribute.
222 (*arm_adddi3): Likewise.
223 (*adddi_sesidi_di): Likewise.
224 (*adddi_zesidi_di): Likewise.
225 (*thumb1_addsi3): Likewise.
226 (addsi3_compare0): Likewise.
227 (*addsi3_compare0_scratch): Likewise.
228 (*compare_negsi_si): Likewise.
229 (cmpsi2_addneg): Likewise.
230 (*addsi3_carryin_<optab>): Likewise.
231 (*addsi3_carryin_alt2_<optab>): Likewise.
232 (*addsi3_carryin_clobercc_<optab>): Likewise.
233 (*subsi3_carryin): Likewise.
234 (*subsi3_carryin_const): Likewise.
235 (*subsi3_carryin_compare): Likewise.
236 (*subsi3_carryin_compare_const): Likewise.
237 (*arm_subdi3): Likewise.
238 (*thumb_subdi3): Likewise.
239 (*subdi_di_zesidi): Likewise.
240 (*subdi_di_sesidi): Likewise.
241 (*subdi_zesidi_di): Likewise.
242 (*subdi_sesidi_di): Likewise.
243 (*subdi_zesidi_ze): Likewise.
244 (thumb1_subsi3_insn): Likewise.
245 (*arm_subsi3_insn): Likewise.
246 (*anddi3_insn): Likewise.
247 (*anddi_zesidi_di): Likewise.
248 (*anddi_sesdi_di): Likewise.
249 (*ne_zeroextracts): Likewise.
250 (*ne_zeroextracts): Likewise.
251 (*ite_ne_zeroextr): Likewise.
252 (*ite_ne_zeroextr): Likewise.
253 (*anddi_notdi_di): Likewise.
254 (*anddi_notzesidi): Likewise.
255 (*anddi_notsesidi): Likewise.
256 (andsi_notsi_si): Likewise.
257 (thumb1_bicsi3): Likewise.
258 (*iordi3_insn): Likewise.
259 (*iordi_zesidi_di): Likewise.
260 (*iordi_sesidi_di): Likewise.
261 (*thumb1_iorsi3_insn): Likewise.
262 (*xordi3_insn): Likewise.
263 (*xordi_zesidi_di): Likewise.
264 (*xordi_sesidi_di): Likewise.
265 (*arm_xorsi3): Likewise.
266 (*andsi_iorsi3_no): Likewise.
268 (*smax_m1): Likewise.
269 (*arm_smax_insn): Likewise.
271 (*arm_smin_insn): Likewise.
272 (*arm_umaxsi3): Likewise.
273 (*arm_uminsi3): Likewise.
274 (*minmax_arithsi): Likewise.
275 (*minmax_arithsi_): Likewise.
276 (*satsi_<SAT:code>): Likewise.
277 (arm_ashldi3_1bit): Likewise.
278 (arm_ashrdi3_1bit): Likewise.
279 (arm_lshrdi3_1bit): Likewise.
280 (*arm_negdi2): Likewise.
281 (*thumb1_negdi2): Likewise.
282 (*arm_negsi2): Likewise.
283 (*thumb1_negsi2): Likewise.
284 (*negdi_extendsid): Likewise.
285 (*negdi_zero_extend): Likewise.
286 (*arm_abssi2): Likewise.
287 (*thumb1_abssi2): Likewise.
288 (*arm_neg_abssi2): Likewise.
289 (*thumb1_neg_abss): Likewise.
290 (one_cmpldi2): Likewise.
291 (extend<mode>di2): Likewise.
292 (*compareqi_eq0): Likewise.
293 (*arm_extendhisi2addsi): Likewise.
294 (*arm_movdi): Likewise.
295 (*thumb1_movdi_insn): Likewise.
296 (*arm_movt): Likewise.
297 (*thumb1_movsi_insn): Likewise.
298 (pic_add_dot_plus_four): Likewise.
299 (pic_add_dot_plus_eight): Likewise.
300 (tls_load_dot_plus_eight): Likewise.
301 (*thumb1_movhi_insn): Likewise.
302 (*thumb1_movsf_insn): Likewise.
303 (*movdf_soft_insn): Likewise.
304 (*thumb_movdf_insn): Likewise.
305 (cbranchsi4_insn): Likewise.
306 (cbranchsi4_scratch): Likewise.
307 (*negated_cbranchsi4): Likewise.
308 (*tbit_cbranch): Likewise.
309 (*tlobits_cbranch): Likewise.
310 (*tstsi3_cbranch): Likewise.
311 (*cbranchne_decr1): Likewise.
312 (*addsi3_cbranch): Likewise.
313 (*addsi3_cbranch_scratch): Likewise.
314 (*arm_cmpdi_insn): Likewise.
315 (*arm_cmpdi_unsig): Likewise.
316 (*arm_cmpdi_zero): Likewise.
317 (*thumb_cmpdi_zero): Likewise.
318 (*deleted_compare): Likewise.
319 (*mov_scc): Likewise.
320 (*mov_negscc): Likewise.
321 (*mov_notscc): Likewise.
322 (*cstoresi_eq0_thumb1_insn): Likewise.
323 (cstoresi_nltu_thumb1): Likewise.
324 (cstoresi_ltu_thu): Likewise.
325 (thumb1_addsi3_addgeu): Likewise.
326 (*arm_jump): Likewise.
327 (*thumb_jump): Likewise.
328 (*check_arch2): Likewise.
329 (arm_casesi_internal): Likewise.
330 (thumb1_casesi_dispatch): Likewise.
331 (*arm_indirect_jump): Likewise.
332 (*thumb1_indirect_jump): Likewise.
334 (*and_scc): Likewise.
335 (*ior_scc): Likewise.
336 (*compare_scc): Likewise.
337 (*cond_move): Likewise.
338 (*cond_arith): Likewise.
339 (*cond_sub): Likewise.
340 (*cmp_ite0): Likewise.
341 (*cmp_ite1): Likewise.
342 (*cmp_and): Likewise.
343 (*cmp_ior): Likewise.
344 (*ior_scc_scc): Likewise.
345 (*ior_scc_scc_cmp): Likewise.
346 (*and_scc_scc): Likewise.
347 (*and_scc_scc_cmp): Likewise.
348 (*and_scc_scc_nod): Likewise.
350 (movcond_addsi): Likewise.
352 (*ifcompare_plus_move): Likewise.
353 (*if_plus_move): Likewise.
354 (*ifcompare_move_plus): Likewise.
355 (*if_move_plus): Likewise.
356 (*ifcompare_arith_arith): Likewise.
357 (*if_arith_arith): Likewise.
358 (*ifcompare_arith_move): Likewise.
359 (*if_arith_move): Likewise.
360 (*ifcompare_move_arith): Likewise.
361 (*if_move_arith): Likewise.
362 (*ifcompare_move_not): Likewise.
363 (*if_move_not): Likewise.
364 (*ifcompare_not_move): Likewise.
365 (*if_not_move): Likewise.
366 (*ifcompare_shift_move): Likewise.
367 (*if_shift_move): Likewise.
368 (*ifcompare_move_shift): Likewise.
369 (*if_move_shift): Likewise.
370 (*ifcompare_shift_shift): Likewise.
371 (*ifcompare_not_arith): Likewise.
372 (*ifcompare_arith_not): Likewise.
373 (*if_arith_not): Likewise.
374 (*ifcompare_neg_move): Likewise.
375 (*if_neg_move): Likewise.
376 (*ifcompare_move_neg): Likewise.
377 (*if_move_neg): Likewise.
378 (prologue_thumb1_interwork): Likewise.
379 (*cond_move_not): Likewise.
380 (*sign_extract_onebit): Likewise.
381 (*not_signextract_onebit): Likewise.
382 (stack_tie): Likewise.
385 (consttable_end): Likewise.
386 (consttable_1): Likewise.
387 (consttable_2): Likewise.
388 (consttable_4): Likewise.
389 (consttable_8): Likewise.
390 (consttable_16): Likewise.
391 (*thumb1_tablejump): Likewise.
392 (prefetch): Likewise.
393 (force_register_use): Likewise.
394 (thumb_eh_return): Likewise.
395 (load_tp_hard): Likewise.
396 (load_tp_soft): Likewise.
398 (*arm_movtas_ze): Likewise.
399 (*arm_rev): Likewise.
400 (*arm_revsh): Likewise.
401 (*arm_rev16): Likewise.
402 * config/arm/thumb2.md
403 (*thumb2_smaxsi3): Likewise.
404 (*thumb2_sminsi3): Likewise.
405 (*thumb32_umaxsi3): Likewise.
406 (*thumb2_uminsi3): Likewise.
407 (*thumb2_negdi2): Likewise.
408 (*thumb2_abssi2): Likewise.
409 (*thumb2_neg_abss): Likewise.
410 (*thumb2_movsi_insn): Likewise.
411 (tls_load_dot_plus_four): Likewise.
412 (*thumb2_movhi_insn): Likewise.
413 (*thumb2_mov_scc): Likewise.
414 (*thumb2_mov_negs): Likewise.
415 (*thumb2_mov_negs): Likewise.
416 (*thumb2_mov_nots): Likewise.
417 (*thumb2_mov_nots): Likewise.
418 (*thumb2_movsicc_): Likewise.
419 (*thumb2_movsfcc_soft_insn): Likewise.
420 (*thumb2_indirect_jump): Likewise.
421 (*thumb2_and_scc): Likewise.
422 (*thumb2_ior_scc): Likewise.
423 (*thumb2_ior_scc_strict_it): Likewise.
424 (*thumb2_cond_move): Likewise.
425 (*thumb2_cond_arith): Likewise.
426 (*thumb2_cond_ari): Likewise.
427 (*thumb2_cond_sub): Likewise.
428 (*thumb2_negscc): Likewise.
429 (*thumb2_movcond): Likewise.
430 (thumb2_casesi_internal): Likewise.
431 (thumb2_casesi_internal_pic): Likewise.
432 (*thumb2_alusi3_short): Likewise.
433 (*thumb2_mov<mode>_shortim): Likewise.
434 (*thumb2_addsi_short): Likewise.
435 (*thumb2_subsi_short): Likewise.
436 (thumb2_addsi3_compare0): Likewise.
437 (*thumb2_cbz): Likewise.
438 (*thumb2_cbnz): Likewise.
439 (*thumb2_one_cmplsi2_short): Likewise.
440 (*thumb2_negsi2_short): Likewise.
441 (*orsi_notsi_si): Likewise.
442 * config/arm/arm1020e.md: Update with new attributes.
443 * config/arm/arm1026ejs.md: Update with new attributes.
444 * config/arm/arm1136jfs.md: Update with new attributes.
445 * config/arm/arm926ejs.md: Update with new attributes.
446 * config/arm/cortex-a15.md: Update with new attributes.
447 * config/arm/cortex-a5.md: Update with new attributes.
448 * config/arm/cortex-a53.md: Update with new attributes.
449 * config/arm/cortex-a7.md: Update with new attributes.
450 * config/arm/cortex-a8.md: Update with new attributes.
451 * config/arm/cortex-a9.md: Update with new attributes.
452 * config/arm/cortex-m4.md: Update with new attributes.
453 * config/arm/cortex-r4.md: Update with new attributes.
454 * config/arm/fa526.md: Update with new attributes.
455 * config/arm/fa606te.md: Update with new attributes.
456 * config/arm/fa626te.md: Update with new attributes.
457 * config/arm/fa726te.md: Update with new attributes.
459 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
461 * config/aarch64/aarch64-simd.md
462 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
463 <vwx> iterator to ensure correct register choice.
464 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
465 (aarch64_sqdmull_n<mode>): Likewise.
466 (aarch64_sqdmull2_n<mode>_internal): Likewise.
467 * config/aarch64/arm_neon.h
468 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
469 (vml<as><q>_n_<su>16): Likewise.
470 (vml<as>l_high_lane<q>_<su>16): Likewise.
471 (vml<as>l_high_n_<su>16): Likewise.
472 (vml<as>l_lane<q>_<su>16): Likewise.
473 (vml<as>l_n_<su>16): Likewise.
474 (vmul<q>_lane<q>_<su>16): Likewise.
475 (vmul<q>_n_<su>16): Likewise.
476 (vmull_lane<q>_<su>16): Likewise.
477 (vmull_n_<su>16): Likewise.
478 (vmull_high_lane<q>_<su>16): Likewise.
479 (vmull_high_n_<su>16): Likewise.
480 (vqrdmulh<q>_n_s16): Likewise.
482 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
484 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
485 have the correct lane parameter.
487 2013-09-06 Richard Biener <rguenther@suse.de>
489 * cfganal.c (control_dependences::~control_dependences):
490 Properly free all of the vector.
492 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
495 * config/i386/i386.c (ix86_conditional_register_usage):
496 Proper initialize extended SSE registers.
498 2013-09-06 Jan Hubicka <jh@suse.cz>
500 PR tree-optimization/58311
501 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
503 2013-09-06 Jan Hubicka <jh@suse.cz>
505 * Makefile.in (tree-sra.o): Update dependencies.
506 * tree-sra.c: Include ipa-utils.h
507 (scan_function): Use recursive_call_p.
508 (has_caller_p): New function.
509 (cgraph_for_node_and_aliases): Count also callers of aliases.
511 2013-09-06 Jan Hubicka <jh@suse.cz>
514 * cgraph.h (symtab_semantically_equivalent_p): Declare.
515 * tree-tailcall.c: Include ipa-utils.h.
516 (find_tail_calls): Use it.
517 * ipa-pure-const.c (check_call): Likewise.
518 * ipa-utils.c (recursive_call_p): New function.
519 * ipa-utils.h (recursive_call_p): Dclare.
520 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
521 (symtab_semantically_equivalent_p): New function.
522 * Makefile.in (tree-tailcall.o): Update dependencies.
524 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
526 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
529 2013-09-06 Richard Biener <rguenther@suse.de>
531 * lto-streamer.h (lto_global_var_decls): Remove.
532 * Makefile.in (OBJS): Remove lto-symtab.o.
533 (lto-symtab.o): Remove.
534 (GTFILES): Remove lto-symtab.c
535 * lto-symtab.c: Move to lto/
537 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
539 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
540 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
541 (UNSPEC_FPINT_RINT): New constant definitions.
542 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
543 definition with 2 attributes.
544 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
545 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
548 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
550 * config/s390/s390.md: Add "bcr_flush" value to mnemonic
552 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
553 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
554 attribute to "z196_alone".
555 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
558 2013-09-06 Richard Biener <rguenther@suse.de>
560 * basic-block.h (class control_dependences): New.
561 * tree-ssa-dce.c (control_dependence_map): Remove.
563 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
564 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
565 find_pdom, find_control_dependence, find_all_control_dependences):
567 (mark_control_dependent_edges_necessary, find_obviously_necessary_stmts,
568 propagate_necessity, tree_dce_init, tree_dce_done,
569 perform_tree_ssa_dce): Adjust.
570 * cfganal.c (set_control_dependence_map_bit,
571 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
572 find_all_control_dependences): Move from tree-ssa-dce.c and
573 implement as methods of control_dependences class.
574 (control_dependences::control_dependences): New.
575 (control_dependences::~control_dependences): Likewise.
576 (control_dependences::get_edges_dependent_on): Likewise.
577 (control_dependences::get_edge): Likewise.
579 2013-09-04 Jan Hubicka <jh@suse.cz>
581 * tree.c (types_same_for_odr): Drop overactive check.
582 * ipa-devirt.c (hash_type_name): Likewise.
584 2013-09-04 Jan Hubicka <jh@suse.cz>
586 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
587 (analyze_functions): ... here.
589 2013-09-04 Jan Hubicka <jh@suse.cz>
592 * cgraphunit.c (analyze_functions): Clear AUX fields
593 after processing; initialize assembler name has.
595 2013-09-05 Jeff Law <law@redhat.com>
597 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
598 from thread_around_empty_block. Record threading path into PATH.
599 Recurse if threading through the initial block is successful.
600 (thread_across_edge): Corresponding changes to slightly simplify.
602 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
604 * config/aarch64/aarch64.md
605 (type): Remove frecpe, frecps, frecpx.
606 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
607 fix to be a TARGET_SIMD instruction.
608 (aarch64_frecps): Remove.
609 * config/aarch64/aarch64-simd.md
610 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
611 (aarch64_frecps<mode>): Handle all float/vector of float modes.
613 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
614 Sofiane Naci <sofiane.naci@arm.com>
616 * config/arm/types.md (define_attr "type"):
618 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
620 into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm", "alu_ext",
621 "alu_reg", "alus_ext", "alus_reg", "bfm", "csel", "logic_reg",
624 into "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
626 Expand "arlo_shift_reg"
627 into "alu_shift_reg", "alus_shift_reg", "logic_shift_reg",
629 Expand "clz" into "clz, "rbit".
630 Rename "shift" to "shift_imm".
631 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
633 Update for attribute changes all occurrences of arlo_* and
635 * config/arm/arm-fixed.md: Update for attribute changes
636 all occurrences of arlo_* types.
637 * config/arm/thumb2.md: Update for attribute changes all occurrences
639 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
640 (cortexa7_older_only): Likewise.
641 (cortexa7_younger): Likewise.
642 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
643 (1020alu_shift_op): Likewise.
644 (1020alu_shift_reg_op): Likewise.
645 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
646 (alu_shift_op): Likewise.
647 (alu_shift_reg_op): Likewise.
648 * config/arm/arm1136jfs.md (11_alu_op): Update for
650 (11_alu_shift_op): Likewise.
651 (11_alu_shift_reg_op): Likewise.
652 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
653 (9_alu_shift_reg_op): Likewise.
654 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
656 (cortex_a15_alu_shift): Likewise.
657 (cortex_a15_alu_shift_reg): Likewise.
658 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
660 (cortex_a5_alu_shift): Likewise.
661 * config/arm/cortex-a53.md
662 (cortex_a53_alu): Update for attribute changes.
663 (cortex_a53_alu_shift): Likewise.
664 * config/arm/cortex-a7.md
665 (cortex_a7_alu_imm): Update for attribute changes.
666 (cortex_a7_alu_reg): Likewise.
667 (cortex_a7_alu_shift): Likewise.
668 * config/arm/cortex-a8.md
669 (cortex_a8_alu): Update for attribute changes.
670 (cortex_a8_alu_shift): Likewise.
671 (cortex_a8_alu_shift_reg): Likewise.
672 * config/arm/cortex-a9.md
673 (cortex_a9_dp): Update for attribute changes.
674 (cortex_a9_dp_shift): Likewise.
675 * config/arm/cortex-m4.md
676 (cortex_m4_alu): Update for attribute changes.
677 * config/arm/cortex-r4.md
678 (cortex_r4_alu): Update for attribute changes.
679 (cortex_r4_mov): Likewise.
680 (cortex_r4_alu_shift_reg): Likewise.
681 * config/arm/fa526.md
682 (526_alu_op): Update for attribute changes.
683 (526_alu_shift_op): Likewise.
684 * config/arm/fa606te.md
685 (606te_alu_op): Update for attribute changes.
686 * config/arm/fa626te.md
687 (626te_alu_op): Update for attribute changes.
688 (626te_alu_shift_op): Likewise.
689 * config/arm/fa726te.md
690 (726te_alu_op): Update for attribute changes.
691 (726te_alu_shift_op): Likewise.
692 (726te_alu_shift_reg_op): Likewise.
693 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
694 (mp626_alu_shift_op): Likewise.
695 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
696 (pj4_alu_conds): Likewise.
697 (pj4_shift): Likewise.
698 (pj4_shift_conds): Likewise.
699 (pj4_alu_shift): Likewise.
700 (pj4_alu_shift_conds): Likewise.
701 * config/aarch64/aarch64.md: Update for attribute change
702 all occurrences of arlo_* and shift* types.
704 2013-09-05 Mike Stump <mikestump@comcast.net>
706 * tree.h: Move documentation for tree_function_decl to tree-core.h
707 with the declaration.
709 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
712 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
713 looking for widest mode.
715 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
717 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
719 2013-09-05 Richard Biener <rguenther@suse.de>
721 PR tree-optimization/58137
722 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
723 Do not create vectors of pointers.
724 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
725 types for the components of the vector initializer.
726 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
727 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
729 2013-09-05 Martin Jambor <mjambor@suse.cz>
731 * ipa-prop.c (remove_described_reference): Accept missing references,
732 return false if that hppens, otherwise return true.
733 (cgraph_node_for_jfunc): New function.
734 (try_decrement_rdesc_refcount): Likewise.
735 (try_make_edge_direct_simple_call): Use them.
736 (ipa_edge_removal_hook): Remove references from rdescs.
737 (ipa_edge_duplication_hook): Clone rdescs and their references
738 when the new edge has the same caller as the old one.
739 * cgraph.c (cgraph_resolve_speculation): Remove speculative
740 reference before removing any edges.
742 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
744 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
746 * thumb2.md (thumb2_storewb_parisi): New pattern.
748 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
750 * config/aarch64/aarch64-option-extensions.def: Add
751 AARCH64_OPT_EXTENSION of 'crc'.
752 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
753 (AARCH64_ISA_CRC): Ditto.
754 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
755 description of the CRC extension.
757 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
759 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
760 * config/rs6000/linux.h: Ditto.
761 * alpha/linux.h: Ditto.
762 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
763 no_c99_libc_has_function.
764 * config/c6x/uclinux-elf.h: Ditto.
765 * config/lm32/uclinux-elf.h: Ditto.
766 * config/m68k/uclinux.h: Ditto.
767 * config/moxie/uclinux.h: Ditto.
768 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
769 (crisv32-*-linux*, cris-*-linux*): Ditto.
770 * config/bfin/bfin.c: Include "tm_p.h".
772 2013-09-05 Richard Biener <rguenther@suse.de>
774 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
775 check for a definition without a basic-block.
777 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
778 Sofiane Naci <sofiane.naci@arm.com>
780 * config/aarch64/aarch64.md
781 (*movti_aarch64): Rename r_2_f and f_2_r.
782 (*movsf_aarch64): Likewise.
783 (*movdf_aarch64): Likewise.
784 (*movtf_aarch64): Likewise.
785 (aarch64_movdi_<mode>low): Likewise.
786 (aarch64_movdi_<mode>high): Likewise.
787 (aarch64_mov<mode>high_di): Likewise.
788 (aarch64_mov<mode>low_di): Likewise.
789 (aarch64_movtilow_tilow): Likewise.
790 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
791 values to config/arm/types.md
792 (attribute "conds"): Update for attribute change.
793 (anddi3_insn): Likewise.
794 (iordi3_insn): Likewise.
795 (xordi3_insn): Likewise.
796 (one_cmpldi2): Likewise.
797 * config/arm/types.md (type): Add Neon types.
798 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
799 use "type" attribute.
800 (movmisalign<mode>_neon_store): Likewise.
801 (movmisalign<mode>_neon_load): Likewise.
802 (vec_set<mode>_internal): Likewise.
803 (vec_setv2di_internal): Likewise.
804 (vec_extract<mode>): Likewise.
805 (vec_extractv2di): Likewise.
806 (add<mode>3_neon): Likewise.
807 (adddi3_neon): Likewise.
808 (sub<mode>3_neon): Likewise.
809 (subdi3_neon): Likewise.
810 (mul<mode>3_neon): Likewise.
811 (mul<mode>3add<mode>_neon): Likewise.
812 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
813 (fma<VCVTF:mode>4)): Likewise.
814 (fma<VCVTF:mode>4_intrinsic): Likewise.
815 (fmsub<VCVTF:mode>4)): Likewise.
816 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
817 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
818 (ior<mode>3): Likewise.
819 (and<mode>3): Likewise.
820 (anddi3_neon): Likewise.
821 (orn<mode>3_neon): Likewise.
822 (orndi3_neon): Likewise.
823 (bic<mode>3_neon): Likewise.
824 (bicdi3_neon): Likewise.
825 (xor<mode>3): Likewise.
826 (one_cmpl<mode>2): Likewise.
827 (abs<mode>2): Likewise.
828 (neg<mode>2): Likewise.
829 (umin<mode>3_neon): Likewise.
830 (umax<mode>3_neon): Likewise.
831 (smin<mode>3_neon): Likewise.
832 (smax<mode>3_neon): Likewise.
833 (vashl<mode>3): Likewise.
834 (vashr<mode>3_imm): Likewise.
835 (vlshr<mode>3_imm): Likewise.
836 (ashl<mode>3_signed): Likewise.
837 (ashl<mode>3_unsigned): Likewise.
838 (neon_load_count): Likewise.
839 (ashldi3_neon_noclobber): Likewise.
840 (signed_shift_di3_neon): Likewise.
841 (unsigned_shift_di3_neon): Likewise.
842 (ashrdi3_neon_imm_noclobber): Likewise.
843 (lshrdi3_neon_imm_noclobber): Likewise.
844 (widen_ssum<mode>3): Likewise.
845 (widen_usum<mode>3): Likewise.
846 (quad_halves_<code>v4si): Likewise.
847 (quad_halves_<code>v4sf): Likewise.
848 (quad_halves_<code>v8hi): Likewise.
849 (quad_halves_<code>v16qi): Likewise.
850 (reduc_splus_v2di): Likewise.
851 (neon_vpadd_internal<mode>): Likewise.
852 (neon_vpsmin<mode>): Likewise.
853 (neon_vpsmax<mode>): Likewise.
854 (neon_vpumin<mode>): Likewise.
855 (neon_vpumax<mode>): Likewise.
856 (ss_add<mode>_neon): Likewise.
857 (us_add<mode>_neon): Likewise.
858 (ss_sub<mode>_neon): Likewise.
859 (us_sub<mode>_neon): Likewise.
860 (neon_vadd<mode>_unspec): Likewise.
861 (neon_vaddl<mode>): Likewise.
862 (neon_vaddw<mode>): Likewise.
863 (neon_vhadd<mode>): Likewise.
864 (neon_vqadd<mode>): Likewise.
865 (neon_vaddhn<mode>): Likewise.
866 (neon_vmul<mode>): Likewise.
867 (neon_vmla<mode>): Likewise.
868 (neon_vmlal<mode>): Likewise.
869 (neon_vmls<mode>): Likewise.
870 (neon_vmlsl<mode>): Likewise.
871 (neon_vqdmulh<mode>): Likewise.
872 (neon_vqdmlal<mode>): Likewise.
873 (neon_vqdmlsl<mode>): Likewise.
874 (neon_vmull<mode>): Likewise.
875 (neon_vqdmull<mode>): Likewise.
876 (neon_vsub<mode>_unspec): Likewise.
877 (neon_vsubl<mode>): Likewise.
878 (neon_vsubw<mode>): Likewise.
879 (neon_vqsub<mode>): Likewise.
880 (neon_vhsub<mode>): Likewise.
881 (neon_vsubhn<mode>): Likewise.
882 (neon_vceq<mode>): Likewise.
883 (neon_vcge<mode>): Likewise.
884 (neon_vcgeu<mode>): Likewise.
885 (neon_vcgt<mode>): Likewise.
886 (neon_vcgtu<mode>): Likewise.
887 (neon_vcle<mode>): Likewise.
888 (neon_vclt<mode>): Likewise.
889 (neon_vcage<mode>): Likewise.
890 (neon_vcagt<mode>): Likewise.
891 (neon_vtst<mode>): Likewise.
892 (neon_vabd<mode>): Likewise.
893 (neon_vabdl<mode>): Likewise.
894 (neon_vaba<mode>): Likewise.
895 (neon_vabal<mode>): Likewise.
896 (neon_vmax<mode>): Likewise.
897 (neon_vmin<mode>): Likewise.
898 (neon_vpaddl<mode>): Likewise.
899 (neon_vpadal<mode>): Likewise.
900 (neon_vpmax<mode>): Likewise.
901 (neon_vpmin<mode>): Likewise.
902 (neon_vrecps<mode>): Likewise.
903 (neon_vrsqrts<mode>): Likewise.
904 (neon_vqabs<mode>): Likewise.
905 (neon_vqneg<mode>): Likewise.
906 (neon_vcls<mode>): Likewise.
907 (clz<mode>2): Likewise.
908 (popcount<mode>2): Likewise.
909 (neon_vrecpe): Likewise.
910 (neon_vrsqrte): Likewise.
911 (neon_vget_lane<mode>_sext_internal): Likewise.
912 (neon_vget_lane<mode>_zext_internal): Likewise.
913 (neon_vdup_n<mode>): Likewise.
914 (neon_vdup_nv2di): Likewise.
915 (neon_vdpu_lane<mode>_internal): Likewise.
916 (neon_vswp<mode>): Likewise.
917 (float<mode><V_cvtto>2): Likewise.
918 (floatuns<mode><V_cvtto>2): Likewise.
919 (fix_trunc<mode><V_cvtto>)2): Likewise
920 (fixuns_trunc<mode><V_cvtto)2): Likewise.
921 (neon_vcvt<mode>): Likewise.
922 (neon_vcvtv4sfv4hf): Likewise.
923 (neon_vcvtv4hfv4sf): Likewise.
924 (neon_vcvt_n<mode>): Likewise.
925 (neon_vmovn<mode>): Likewise.
926 (neon_vqmovn<mode>): Likewise.
927 (neon_vqmovun<mode>): Likewise.
928 (neon_vmovl<mode>): Likewise.
929 (neon_vmul_lane<mode>): Likewise.
930 (neon_vmull_lane<mode>): Likewise.
931 (neon_vqdmull_lane<mode>): Likewise.
932 (neon_vqdmulh_lane<mode>): Likewise.
933 (neon_vmla_lane<mode>): Likewise.
934 (neon_vmlal_lane<mode>): Likewise.
935 (neon_vqdmlal_lane<mode>): Likewise.
936 (neon_vmls_lane<mode>): Likewise.
937 (neon_vmlsl_lane<mode>): Likewise.
938 (neon_vqdmlsl_lane<mode>): Likewise.
939 (neon_vext<mode>): Likewise.
940 (neon_vrev64<mode>): Likewise.
941 (neon_vrev32<mode>): Likewise.
942 (neon_vrev16<mode>): Likewise.
943 (neon_vbsl<mode>_internal): Likewise.
944 (neon_vshl<mode>): Likewise.
945 (neon_vqshl<mode>): Likewise.
946 (neon_vshr_n<mode>): Likewise.
947 (neon_vshrn_n<mode>): Likewise.
948 (neon_vqshrn_n<mode>): Likewise.
949 (neon_vqshrun_n<mode>): Likewise.
950 (neon_vshl_n<mode>): Likewise.
951 (neon_vqshl_n<mode>): Likewise.
952 (neon_vqshlu_n<mode>): Likewise.
953 (neon_vshll_n<mode>): Likewise.
954 (neon_vsra_n<mode>): Likewise.
955 (neon_vsri_n<mode>): Likewise.
956 (neon_vsli_n<mode>): Likewise.
957 (neon_vtbl1v8qi): Likewise.
958 (neon_vtbl2v8qi): Likewise.
959 (neon_vtbl3v8qi): Likewise.
960 (neon_vtbl4v8qi): Likewise.
961 (neon_vtbx1v8qi): Likewise.
962 (neon_vtbx2v8qi): Likewise.
963 (neon_vtbx3v8qi): Likewise.
964 (neon_vtbx4v8qi): Likewise.
965 (neon_vtrn<mode>_internal): Likewise.
966 (neon_vzip<mode>_internal): Likewise.
967 (neon_vuzp<mode>_internal): Likewise.
968 (neon_vld1<mode>): Likewise.
969 (neon_vld1_lane<mode>): Likewise.
970 (neon_vld1_dup<mode>): Likewise.
971 (neon_vld1_dupv2di): Likewise.
972 (neon_vst1<mode>): Likewise.
973 (neon_vst1_lane<mode>): Likewise.
974 (neon_vld2<mode>): Likewise.
975 (neon_vld2_lane<mode>): Likewise.
976 (neon_vld2_dup<mode>): Likewise.
977 (neon_vst2<mode>): Likewise.
978 (neon_vst2_lane<mode>): Likewise.
979 (neon_vld3<mode>): Likewise.
980 (neon_vld3qa<mode>): Likewise.
981 (neon_vld3qb<mode>): Likewise.
982 (neon_vld3_lane<mode>): Likewise.
983 (neon_vld3_dup<mode>): Likewise.
984 (neon_vst3<mode>): Likewise.
985 (neon_vst3qa<mode>): Likewise.
986 (neon_vst3qb<mode>): Likewise.
987 (neon_vst3_lane<mode>): Likewise.
988 (neon_vld4<mode>): Likewise.
989 (neon_vld4qa<mode>): Likewise.
990 (neon_vld4qb<mode>): Likewise.
991 (neon_vld4_lane<mode>): Likewise.
992 (neon_vld4_dup<mode>): Likewise.
993 (neon_vst4<mode>): Likewise.
994 (neon_vst4qa<mode>): Likewise.
995 (neon_vst4qb<mode>): Likewise.
996 (neon_vst4_lane<mode>): Likewise.
997 (neon_vec_unpack<US>_lo_<mode>): Likewise.
998 (neon_vec_unpack<US>_hi_<mode>): Likewise.
999 (neon_vec_<US>mult_lo_<mode>): Likewise.
1000 (neon_vec_<US>mult_hi_<mode>): Likewise.
1001 (neon_vec_<US>shiftl_<mode>): Likewise.
1002 (neon_unpack<US>_<mode>): Likewise.
1003 (neon_vec_<US>mult_<mode>): Likewise.
1004 (vec_pack_trunc_<mode>): Likewise.
1005 (neon_vec_pack_trunk_<mode>): Likewise.
1006 (neon_vabd<mode>_2): Likewise.
1007 (neon_vabd<mode>_3): Likewise.
1008 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1009 (thumb2_movsi_vfp): Likewise.
1010 (movdi_vfp): Likewise.
1011 (movdi_vfp_cortexa8): Likewise.
1012 (movhf_vfp_neon): Likewise.
1013 (movhf_vfp): Likewiwse.
1014 (movsf_vfp): Likewiwse.
1015 (thumb2_movsf_vfp): Likewiwse.
1016 (movdf_vfp): Likewise.
1017 (thumb2_movdf_vfp): Likewise.
1018 (movsfcc_vfp): Likewise.
1019 (thumb2_movsfcc_vfp): Likewise.
1020 (movdfcc_vfp): Likewise.
1021 (thumb2_movdfcc_vfp): Likewise.
1022 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1023 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1024 (v10_v2c): Likewise.
1025 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1027 (cortex_a15_neon_int_2): Likewise.
1028 (cortex_a15_neon_int_3): Likewise.
1029 (cortex_a15_neon_int_4): Likewise.
1030 (cortex_a15_neon_int_5): Likewise.
1031 (cortex_a15_neon_vqneg_vqabs): Likewise.
1032 (cortex_a15_neon_vmov): Likewise.
1033 (cortex_a15_neon_vaba): Likewise.
1034 (cortex_a15_neon_vaba_qqq): Likewise.
1035 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1036 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1037 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1038 scalar_64_32_long_scalar): Likewise.
1039 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1040 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1041 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1042 lotype_qdd_64_32_long): Likewise.
1043 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1044 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1045 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1046 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1047 (cortex_a15_neon_shift_1): Likewise.
1048 (cortex_a15_neon_shift_2): Likewise.
1049 (cortex_a15_neon_shift_3): Likewise.
1050 (cortex_a15_neon_vshl_ddd): Likewise.
1051 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1052 (cortex_a15_neon_vsra_vrsra): Likewise.
1053 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1054 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1055 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1056 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1057 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1058 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1059 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1060 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1061 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1062 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1063 (cortex_a15_neon_bp_simple): Likewise.
1064 (cortex_a15_neon_bp_2cycle): Likewise.
1065 (cortex_a15_neon_bp_3cycle): Likewise.
1066 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1067 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1068 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1069 (cortex_a15_neon_vld2_4_regs): Likewise.
1070 (cortex_a15_neon_vld3_vld4): Likewise.
1071 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1072 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1073 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1074 (cortex_a15_neon_vst3_vst4): Likewise.
1075 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1076 (cortex_a15_neon_vld3_vld4_lane" 10
1077 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1078 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1079 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1080 (cortex_a15_neon_ldm_2): Likewise.0
1081 (cortex_a15_neon_stm_2): Likewise.
1082 (cortex_a15_neon_mcr): Likewise.
1083 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1084 (cortex_a15_neon_mrc): Likewise.
1085 (cortex_a15_neon_mrrc): Likewise.
1086 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1088 (cortex_a15_alu_shift): Likewise.
1089 (cortex_a15_alu_shift_reg): Likewise.
1090 (cortex_a15_mult32): Likewise.
1091 (cortex_a15_mult64): Likewise.
1092 (cortex_a15_block): Likewise.
1093 (cortex_a15_branch): Likewise.
1094 (cortex_a15_load1): Likewise.
1095 (cortex_a15_load3): Likewise.
1096 (cortex_a15_store1): Likewise.
1097 (cortex_a15_store3): Likewise.
1098 (cortex_a15_call): Likewise.
1099 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute
1101 (cortex_a5_f2r): Likewise.
1102 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1104 (cortex_a53_f2r): Likewise.
1105 * config/arm/cortex-a7.md
1106 (cortex_a7_branch): Update for attribute change.
1107 (cortex_a7_call): Likewise.
1108 (cortex_a7_alu_imm): Likewise.
1109 (cortex_a7_alu_reg): Likewise.
1110 (cortex_a7_alu_shift): Likewise.
1111 (cortex_a7_mul): Likewise.
1112 (cortex_a7_load1): Likewise.
1113 (cortex_a7_store1): Likewise.
1114 (cortex_a7_load2): Likewise.
1115 (cortex_a7_store2): Likewise.
1116 (cortex_a7_load3): Likewise.
1117 (cortex_a7_store3): Likewise.
1118 (cortex_a7_load4): Likewise.
1119 (cortex_a7_store4): Likewise.
1120 (cortex_a7_fpalu): Likewise.
1121 (cortex_a7_fconst): Likewise.
1122 (cortex_a7_fpmuls): Likewise.
1123 (cortex_a7_neon_mul): Likewise.
1124 (cortex_a7_fpmacs): Likewise.
1125 (cortex_a7_neon_mla: Likewise.
1126 (cortex_a7_fpmuld: Likewise.
1127 (cortex_a7_fpmacd: Likewise.
1128 (cortex_a7_fpfmad: Likewise.
1129 (cortex_a7_fdivs: Likewise.
1130 (cortex_a7_fdivd: Likewise.
1131 (cortex_a7_r2f: Likewise.
1132 (cortex_a7_f2r: Likewise.
1133 (cortex_a7_f_flags: Likewise.
1134 (cortex_a7_f_loads: Likewise.
1135 (cortex_a7_f_loadd: Likewise.
1136 (cortex_a7_f_stores: Likewise.
1137 (cortex_a7_f_stored: Likewise.
1138 (cortex_a7_neon): Likewise.
1139 * config/arm/cortex-a8-neon.md
1140 (cortex_a8_neon_mrc): Update for attribute change.
1141 (cortex_a8_neon_mrrc): Likewise.
1142 (cortex_a8_neon_int_1): Likewise.
1143 (cortex_a8_neon_int_2): Likewise.
1144 (cortex_a8_neon_int_3): Likewise.
1145 (cortex_a8_neon_int_4): Likewise.
1146 (cortex_a8_neon_int_5): Likewise.
1147 (cortex_a8_neon_vqneg_vqabs): Likewise.
1148 (cortex_a8_neon_vmov): Likewise.
1149 (cortex_a8_neon_vaba): Likewise.
1150 (cortex_a8_neon_vaba_qqq): Likewise.
1151 (cortex_a8_neon_vsma): Likewise.
1152 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1153 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1154 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1156 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1157 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1158 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1159 long_scalar_qdd_64_32_long): Likewise.
1160 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1161 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1162 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1163 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1164 (cortex_a8_neon_shift_1): Likewise.
1165 (cortex_a8_neon_shift_2): Likewise.
1166 (cortex_a8_neon_shift_3): Likewise.
1167 (cortex_a8_neon_vshl_ddd): Likewise.
1168 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1169 (cortex_a8_neon_vsra_vrsra): Likewise.
1170 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1171 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1172 (cortex_a8_neon_fp_vsum): Likewise.
1173 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1174 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1175 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1176 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1177 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1178 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1179 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1180 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1181 (cortex_a8_neon_bp_simple): Likewise.
1182 (cortex_a8_neon_bp_2cycle): Likewise.
1183 (cortex_a8_neon_bp_3cycle): Likewise.
1184 (cortex_a8_neon_ldr): Likewise.
1185 (cortex_a8_neon_str): Likewise.
1186 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1187 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1188 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1189 (cortex_a8_neon_vld2_4_regs): Likewise.
1190 (cortex_a8_neon_vld3_vld4): Likewise.
1191 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1192 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1193 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1194 (cortex_a8_neon_vst3_vst4): Likewise.
1195 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1196 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1197 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1198 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1199 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1200 (cortex_a8_neon_mcr): Likewise.
1201 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1202 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
1204 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1206 (ca9_neon_mrrc): Likewise.
1207 (cortex_a9_neon_int_1): Likewise.
1208 (cortex_a9_neon_int_2): Likewise.
1209 (cortex_a9_neon_int_3): Likewise.
1210 (cortex_a9_neon_int_4): Likewise.
1211 (cortex_a9_neon_int_5): Likewise.
1212 (cortex_a9_neon_vqneg_vqabs): Likewise.
1213 (cortex_a9_neon_vmov): Likewise.
1214 (cortex_a9_neon_vaba): Likewise.
1215 (cortex_a9_neon_vaba_qqq): Likewise.
1216 (cortex_a9_neon_vsma): Likewise.
1217 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1218 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1219 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1221 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1222 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1223 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1224 long_scalar_qdd_64_32_long): Likewise.
1225 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1226 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1227 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1228 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1229 (cortex_a9_neon_shift_1): Likewise.
1230 (cortex_a9_neon_shift_2): Likewise.
1231 (cortex_a9_neon_shift_3): Likewise.
1232 (cortex_a9_neon_vshl_ddd): Likewise.
1233 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1234 (cortex_a9_neon_vsra_vrsra): Likewise.
1235 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1236 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1237 (cortex_a9_neon_fp_vsum): Likewise.
1238 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1239 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1240 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1241 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1242 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1243 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1244 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1245 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1246 (cortex_a9_neon_bp_simple): Likewise.
1247 (cortex_a9_neon_bp_2cycle): Likewise.
1248 (cortex_a9_neon_bp_3cycle): Likewise.
1249 (cortex_a9_neon_ldr): Likewise.
1250 (cortex_a9_neon_str): Likewise.
1251 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1252 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1253 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1254 (cortex_a9_neon_vld2_4_regs): Likewise.
1255 (cortex_a9_neon_vld3_vld4): Likewise.
1256 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1257 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1258 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1259 (cortex_a9_neon_vst3_vst4): Likewise.
1260 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1261 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1262 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1263 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1264 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1265 (cortex_a9_neon_mcr): Likewise.
1266 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1267 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1268 (cortex_a9_fps): Likewise.
1269 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1271 (cortex_m4_fmuls): Likewise.
1272 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1274 (cortex_r4_mrc): Likewise.
1275 * config/arm/iterators.md: Update comment referring to neon_type.
1276 * config/arm/iwmmxt.md
1277 (iwmmxt_arm_movdi): Update for attribute change.
1278 (iwmmxt_movsi_insn): Likewise.
1279 * config/arm/marvell-pj4.md
1280 (pj4_vfp_to_core): Update for attribute change.
1281 (pj4_core_to_vfp): Likewise.
1282 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1284 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1285 (vfp_fstore): Likewise.
1286 * doc/md.texi: Change references to neon_type to refer to type.
1288 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1290 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1292 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
1294 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1297 2013-09-04 Easwaran Raman <eraman@google.com>
1300 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1301 (build_and_add_sum): Use it.
1302 (appears_later_in_bb): Simplify code.
1304 2013-09-04 Teresa Johnson <tejohnson@google.com>
1306 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1308 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1310 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1312 2013-09-04 Jan Hubicka <jh@suse.cz>
1314 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1315 * ipa-devirt.c: Include diganostic.h
1316 (odr_type_d): Add types and types_set.
1317 (hash_type_name): Work for types with vtables during LTO.
1318 (odr_hasher::remove): Fix comment; destroy types_set.
1319 (add_type_duplicate): New function,
1320 (get_odr_type): Use it.
1321 (dump_type_inheritance_graph): Dump type duplicates.
1322 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1324 * tree.c (types_same_for_odr): Give exact answers on types with
1327 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1329 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1330 explaining their differences.
1332 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1334 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1336 2013-09-03 Jeff Law <law@redhat.com>
1338 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1339 when not threading through a joiner block. Pass joiner/no joiner
1340 state to register_jump_thread.
1341 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1342 state from argument rather than implying on path length.
1343 Dump the entire jump thread path into debugging dump.
1344 * tree-flow.h (register_jump_thread): Update prototype.
1346 2013-08-29 Xinliang David Li <davidxl@google.com>
1348 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1349 Remove a trivial gcc_assert.
1351 2013-08-29 Xinliang David Li <davidxl@google.com>
1353 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1354 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1355 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1356 Delay base decl alignment adjustment.
1357 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1358 * tree-vectorizer.h: New data structure.
1359 (set_dr_misalignment): New function.
1360 (dr_misalignment): Ditto.
1361 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1362 (vectorizable_load): Ditto.
1363 (ensure_base_align): New function.
1364 (vectorize_loops): Add dbg_cnt support.
1365 (execute_vect_slp): Ditto.
1366 * dbgcnt.def: New debug counter.
1367 * Makefile: New dependency.
1369 2013-09-03 Meador Inge <meadori@codesourcery.com>
1373 2013-08-30 Meador Inge <meadori@codesourcery.com>
1375 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1377 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
1379 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
1380 function descriptor.
1382 2013-09-03 Richard Biener <rguenther@suse.de>
1384 * tree-affine.c (add_elt_to_tree): Fix association issue,
1385 avoid useless converts and make sure to always return a
1386 properly typed result.
1388 2013-09-03 Richard Biener <rguenther@suse.de>
1391 * fold-const.c (negate_expr_p): Fix division case.
1392 (negate_expr): Likewise.
1394 2013-09-03 Richard Biener <rguenther@suse.de>
1397 * tree-streamer-out.c: Include tm.h.
1398 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
1400 2013-09-03 Jan Hubicka <jh@suse.cz>
1402 * tree-profile.c (tree_profiling): Cleanup CFG when done.
1404 2013-09-03 Alan Modra <amodra@gmail.com>
1406 * config.gcc (powerpc*-*-linux*): Add support for little-endian
1407 multilibs to big-endian target and vice versa.
1408 * config/rs6000/t-linux64: Use := assignment on all vars.
1409 (MULTILIB_EXTRA_OPTS): Remove fPIC.
1410 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
1411 * config/rs6000/t-linux64le: New file.
1412 * config/rs6000/t-linux64bele: New file.
1413 * config/rs6000/t-linux64lebe: New file.
1415 2013-09-02 Jan Hubicka <jh@suse.cz>
1417 * ipa-inline-transform.c (inline_transform): Do not
1418 optimize_inline_calls when not optimizing.
1420 2013-09-02 Jan Hubicka <jh@suse.cz>
1422 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
1423 duplicated nodes for assembler names.
1424 * symtab.c (symtab_unregister_node): Do not attempt to unlink
1425 hard registers from assembler name hash.
1427 2013-09-02 Jan Hubicka <jh@suse.cz>
1429 * ipa-split.c (execute_split_functions): Split externally visible
1430 functions called once.
1432 2013-09-02 Martin Jambor <mjambor@suse.cz>
1435 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
1436 linked list. When finding the correct duplicate, also consider also
1437 the caller in additon to its inlined_to node.
1439 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
1441 * config/aarch64/aarch64-simd-builtins.def
1442 (dup_lane_scalar): Remove.
1443 * config/aarch64/aarch64-simd.md
1444 (aarch64_simd_dup): Add 'w->w' alternative.
1445 (aarch64_dup_lane<mode>): Allow for VALL.
1446 (aarch64_dup_lane_scalar<mode>): Remove.
1447 (aarch64_dup_lane_<vswap_width_name><mode>): New.
1448 (aarch64_get_lane_signed<mode>): Add w->w altenative.
1449 (aarch64_get_lane_unsigned<mode>): Likewise.
1450 (aarch64_get_lane<mode>): Likewise.
1451 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
1452 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
1453 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
1454 (VCON): Change container of V2SF.
1455 (vswap_width_name): Likewise.
1456 * config/aarch64/arm_neon.h
1457 (__aarch64_vdup_lane_any): New.
1458 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1459 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
1460 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1462 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
1465 * expr.c (emit_move_complex): Do not move complex FP values as parts if
1466 the source or the destination is a single hard register.
1468 2013-09-02 Richard Biener <rguenther@suse.de>
1471 * tree-scalar-evolution.c (instantiate_scev_name): Allow
1474 2013-09-02 Richard Biener <rguenther@suse.de>
1476 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
1477 arithmetic to sizetype.
1479 2013-09-02 Bin Cheng <bin.cheng@arm.com>
1481 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
1482 Find auto-increment use both before and after candidate.
1484 2013-09-02 Marek Polacek <polacek@redhat.com>
1486 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
1488 2013-09-01 Jan Hubicka <jh@suse.cz>
1490 * Makefile.in: Add ipa-profile.o
1491 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
1492 * cgraph.c (struct cgraph_propagate_frequency_data,
1493 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
1494 ipa-profile.c; replace cgraph_ by ipa_ prefix.
1495 * cgraph.h (cgraph_propagate_frequency): Remove.
1496 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
1497 (inline_update_callee_summaries): Update.
1498 * ipa-profile.c: New file.
1499 * ipa-utils.h (ipa_propagate_frequency): Declare.
1500 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
1501 data-streamer.h, value-prof.h
1502 (symtab_remove_unreachable_nodes): Update profile.
1503 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
1504 account_time_size, cmp_counts, dump_histogram,
1505 ipa_profile_generate_summary, ipa_profile_write_summary,
1506 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
1507 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
1508 Move to ipa-profile.c
1510 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
1512 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
1514 2013-09-01 Jan Hubicka <jh@suse.cz>
1516 * common.opt (fdevirtualize-speculatively): New function.
1517 * invoke.texi (fdevirtualize-speculatively): Document.
1518 * ipa-devirt.c: Include ipa-inline.h
1519 (likely_target_p): New function.
1520 (ipa_devirt): New function.
1521 (gate_ipa_devirt): New function.
1522 (pass_data_ipa_devirt): New static var.
1523 (pass_ipa_devirt): Likewise.
1524 (make_pass_ipa_devirt): New function.
1525 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
1526 (common_handle_option): Disable devirtualization when
1527 value range profiling is available.
1528 * passes.def (pass_ipa_devirt): Add.
1529 * timever.def (TV_IPA_DEVIRT): New timevar.
1530 * tree-pass.h (make_pass_ipa_devirt):
1532 2013-09-01 Iain Sandoe <iain@codesourcery.com>
1534 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
1535 include sanitize(undefined).
1537 2013-08-31 Diego Novillo <dnovillo@google.com>
1539 * Makefile.in (TREE_CORE_H): Define.
1541 (GTFILES): Add tree-core.h.
1542 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
1544 * tree-core.h: New file.
1545 Move all data structures, enum, typedefs, global
1546 declarations and constants from ...
1549 2013-08-31 Jan Hubicka <jh@suse.cz>
1551 * bulitins.c (expand_builtin): Do not early exit for gcov
1552 instrumented functions.
1554 2013-08-31 Marek Polacek <polacek@redhat.com>
1556 * ubsan.c: Include tm_p.h.
1558 2013-08-31 Jan Hubicka <jh@suse.cz>
1560 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
1563 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
1564 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
1565 * tree-cfg.c (verify_gimple_label): ... here.
1566 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
1567 (ipa_merge_profiles): New function.
1568 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
1569 (lto_input_function_body): Likewise.
1570 * ipa-utils.h (ipa_merge_profiles): Declare.
1571 * lto-streamer.h (lto_input_function_body): Update prototype.
1572 (emit_label_in_global_context_p): Remove.
1573 * lto-symtab.c: Include ipa-utils.h
1574 (lto_cgraph_replace_node): Use ipa_merge_profiles.
1576 2013-08-31 Jan Hubicka <jh@suse.cz>
1578 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
1580 2013-08-31 Jan Hubicka <jh@suse.cz>
1582 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
1584 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
1586 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
1587 "cmp" RTX before signed_comparison_operator check to account
1590 2013-08-30 Jan Hubicka <jh@suse.cz>
1592 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
1593 (detect_type_change_1): Rename to ...
1594 (detect_type_change): ... this one; early return on non-polymorphic
1596 (detect_type_change_ssa): Add comp_type parameter; update
1597 use of detect_type_change.
1598 (compute_complex_assign_jump_func): Add param_type parameter;
1599 update use of detect_type_change_ssa.
1600 (compute_complex_ancestor_jump_func): Likewise.
1601 (ipa_get_callee_param_type): New function.
1602 (ipa_compute_jump_functions_for_edge): Compute parameter type;
1603 update calls to the jump function computation functions.
1605 2013-08-30 Teresa Johnson <tejohnson@google.com>
1606 Steven Bosscher <steven@gcc.gnu.org>
1608 * cfgrtl.c (fixup_new_cold_bb): New routine.
1609 (commit_edge_insertions): Invoke fixup_partitions.
1610 (find_partition_fixes): New routine.
1611 (fixup_partitions): Ditto.
1612 (verify_hot_cold_block_grouping): Update comments.
1613 (rtl_verify_edges): Invoke find_partition_fixes.
1614 (rtl_verify_bb_pointers): Update comments.
1615 (rtl_verify_bb_layout): Ditto.
1616 * basic-block.h (probably_never_executed_edge_p): Declare.
1617 (fixup_partitions): Ditto.
1618 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
1619 * bb-reorder.c (sanitize_hot_paths): New function.
1620 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
1622 * predict.c (probably_never_executed_edge_p): New routine.
1623 * cfg.c (check_bb_profile): Add partition insanity warnings.
1625 2013-08-30 Meador Inge <meadori@codesourcery.com>
1627 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1629 2013-08-30 Marek Polacek <polacek@redhat.com>
1631 * Makefile.in (ubsan.o): Add.
1632 (c-family/c-ubsan.o): Add.
1633 (builtins.o): Add ubsan.h dependency.
1634 * ubsan.h: New file.
1635 * ubsan.c: New file.
1636 * common.opt: Add -fsanitize=undefined option.
1637 (flag_sanitize): Add variable.
1638 (fsanitize=): Add option. Add Driver.
1639 (fsanitize=thread): Remove option.
1640 (fsanitize=address): Likewise.
1641 (static-libubsan): New option.
1642 * doc/invoke.texi: Document the new flag and -static-libubsan.
1643 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
1644 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
1645 * builtin-attrs.def (ATTR_COLD): Define.
1646 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1647 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
1648 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
1649 * flag-types.h (sanitize_code): New enum.
1650 * opts.c (common_handle_option): Parse command line arguments
1651 of -fsanitize=. Add -fsanitize=unreachable option.
1652 * varasm.c (get_variable_section): Adjust.
1653 (assemble_noswitch_variable): Likewise.
1654 (assemble_variable): Likewise.
1655 (output_constant_def_contents): Likewise.
1656 (categorize_decl_for_section): Likewise.
1657 (place_block_symbol): Likewise.
1658 (output_object_block): Likewise.
1659 * builtins.def: Likewise.
1660 * toplev.c (compile_file): Likewise.
1661 (process_options): Likewise.
1662 * cppbuiltin.c: Likewise.
1663 * tsan.c (tsan_pass): Likewise.
1664 (tsan_gate): Likewise.
1665 (tsan_gate_O0): Likewise.
1666 * cfgexpand.c (partition_stack_vars): Likewise.
1667 (expand_stack_vars): Likewise.
1668 (defer_stack_allocation): Likewise.
1669 (expand_used_vars): Likewise.
1670 * cfgcleanup.c (old_insns_match_p): Likewise.
1671 * asan.c (asan_finish_file): Likewise.
1672 (asan_instrument): Likewise.
1673 (gate_asan): Likewise.
1674 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
1675 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1676 (asan_global_struct): Use pointer_sized_int_node instead
1677 calling build_nonstandard_integer_type.
1678 (initialize_sanitizer_builtins): Likewise.
1679 (asan_finish_file): Likewise.
1680 * gcc.c: Document %{%:function(args):X}.
1681 (static_spec_functions): Add sanitize.
1682 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
1683 store funcval != NULL there.
1684 (do_spec_1): Adjust handle_spec_function caller.
1685 (handle_braces): Allow %:function(args) as condition.
1686 (sanitize_spec_function): New function.
1687 (ADD_STATIC_LIBUBSAN_LIBS): Define.
1688 (LIBUBSAN_SPEC): Likewise.
1689 (LIBUBSAN_EARLY_SPEC): Likewise.
1690 (SANITIZER_SPEC): Handle libubsan.
1691 (SANITIZER_EARLY_SPEC): Likewise.
1692 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
1693 instead of fsanitize=address.
1694 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
1695 instead of fsanitize=address*.
1696 * builtins.c: Include ubsan.h.
1697 (fold_builtin_0): Instrument __builtin_unreachable.
1698 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
1699 instead of flag_asan.
1700 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
1701 (pointer_sized_int_node): Define.
1702 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
1704 2013-08-30 Mike Stump <mikestump@comcast.net>
1706 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
1709 2013-08-29 Jan Hubicka <jh@suse.cz>
1711 * cgraph.c (cgraph_function_body_availability): Handle weakref
1713 * passes.def: Remove pass_fixup_cfg.
1714 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
1715 track when we need to remove functions.
1716 (gate_ipa_inline): Execute inlining always; add comment why.
1717 (pass_data_ipa_inline): Remove TODO_remove_functions.
1718 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
1719 do not produce summaries.
1720 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
1721 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
1722 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
1723 constant pool and vtable.
1725 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
1727 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
1728 New arm_neon.h's internal macros to specify 64-bit constants.
1729 Avoid using stdint.h's macros.
1731 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
1733 * recog.c (verify_changes): Verify that changes[i].old is non-zero
1734 before applying REG_P.
1736 2013-08-30 Jakub Jelinek <jakub@redhat.com>
1738 PR tree-optimization/58277
1739 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
1740 after seeing too many stmts with vdef in between dombb and current
1741 bb, invalidate everything.
1743 2013-08-30 Richard Biener <rguenther@suse.de>
1745 * fold-const.c (fold_single_bit_test): Fix overflow test.
1747 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
1749 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
1750 and which can live in a register, always retrieve the value on entry.
1751 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
1752 passed by invisible reference specially.
1753 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
1754 (vt_add_function_parameter): Correctly deal with a parameter passed by
1755 invisible reference.
1757 2013-08-30 Jan Hubicka <jh@suse.cz>
1759 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
1761 2013-08-30 Richard Biener <rguenther@suse.de>
1763 PR tree-optimization/58228
1764 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
1765 allow invariant loads in nested loop vectorization.
1767 2013-08-30 Richard Biener <rguenther@suse.de>
1769 PR tree-optimization/58223
1770 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
1771 (has_anti_or_output_dependence): ... this and adjust to also
1772 look for output dependences.
1773 (mark_nodes_having_upstream_mem_writes): Adjust.
1774 (rdg_flag_uses): Likewise.
1776 2013-08-30 Richard Biener <rguenther@suse.de>
1778 PR tree-optimization/58010
1779 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
1780 assert that we have a loop-closed PHI.
1782 2013-08-29 Jan Hubicka <jh@suse.cz>
1784 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
1785 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
1786 * lto-section-in.c (lto_free_function_in_decl_state): New function.
1787 (lto_free_function_in_decl_state_for_node): New function.
1789 2013-08-29 Xinliang David Li <davidxl@google.com>
1791 * loop-unroll.c (report_unroll_peel): Minor message change.
1792 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
1793 Emit alignment peeling message with default -fopt-info.
1794 (vect_loop_versioning): Emit loop version info message.
1795 * tree-vectorizer.c (vectorize_loops): Minor message change.
1796 (execute_vect_slp): Ditto.
1798 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
1800 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
1801 of the clone from the DECL_NAME of the original function.
1803 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
1805 * passes.c (register_pass): Add overload.
1806 * tree-pass.h (register_pass): Forward declare it. Add comment.
1808 2013-08-29 Jan Hubicka <jh@suse.cz>
1810 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
1811 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
1812 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
1813 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
1814 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
1815 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
1816 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
1817 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
1819 2013-08-29 Teresa Johnson <tejohnson@google.com>
1821 * dumpfile.c (dump_loc): Output column number.
1822 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
1823 * doc/invoke.texi: Document optall -fopt-info flag.
1824 * profile.c (read_profile_edge_counts): Use new dump framework.
1825 (compute_branch_probabilities): Ditto.
1826 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
1827 when pass not in any opt group.
1828 * pass_manager.h (pass_manager::get_pass_profile): New method.
1829 * value-prof.c (check_counter): Use new dump framework.
1830 (check_ic_target): Ditto.
1831 * coverage.c (get_coverage_counts): Ditto.
1832 (coverage_init): Setup new dump framework.
1834 2013-08-29 Richard Biener <rguenther@suse.de>
1836 PR tree-optimization/58246
1837 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
1838 handle the dominance check inside a basic-block.
1840 2013-08-29 Richard Biener <rguenther@suse.de>
1843 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
1844 of default defs that appear in abnormal PHI nodes.
1846 2013-08-29 Richard Biener <rguenther@suse.de>
1848 PR tree-optimization/57685
1849 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
1850 single-use operands to avoid exponential complexity.
1852 2013-08-28 Dehao Chen <dehao@google.com>
1854 * ipa-inline.c (edge_badness): Fix integer underflow.
1856 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
1858 * gtm-builtins.def (_ITM_free): Declare leaf.
1860 2013-08-28 Jakub Jelinek <jakub@redhat.com>
1863 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
1864 (*tls_local_dynamic_base_64_largepic): Likewise.
1865 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
1866 Remove predicate from call operand.
1867 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
1868 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
1870 2013-08-28 Jeff Law <law@redhat.com>
1872 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
1873 checks for the number of predecessors and successors allowed.
1874 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
1875 which require copying a joiner block if there is a request which
1876 is a subpath that requires no joiner block copying.
1878 2013-08-28 Jan Hubicka <jh@suse.cz>
1880 * lto-streamer-out.c (DFS_write_tree_body): Drop
1881 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
1882 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
1883 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
1884 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
1885 Do not read DECL_ERROR_ISSUED.
1886 (unpack_ts_decl_with_vis_value_fields): Do not read
1888 (lto_input_ts_binfo_tree_pointers): Do not read
1889 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
1890 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
1891 write DECL_ERROR_ISSUED..
1892 (pack_ts_decl_with_vis_value_fields): Do not write
1894 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
1895 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
1896 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
1897 * tree.h (tree_decl_common): Update comment.
1898 (DECL_ERROR_ISSUED): Remove.
1900 2013-08-28 Jakub Jelinek <jakub@redhat.com>
1903 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
1905 2013-08-28 Jan Hubicka <jh@suse.cz>
1907 * builtins.def (free): Declare leaf.
1909 2013-08-27 David Malcolm <dmalcolm@redhat.com>
1912 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
1913 * configure: Regenerate.
1915 2013-08-27 Martin Jambor <mjambor@suse.cz>
1917 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
1918 (ipa_ancestor_jf_data): Likewise.
1919 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
1920 (ipa_get_jf_pass_through_type_preserved): New function.
1921 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
1922 (ipa_get_jf_ancestor_type_preserved): New function.
1923 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
1924 (ipa_get_jf_ancestor_result): Likewise.
1925 (propagate_vals_accross_pass_through): Use
1926 ipa_get_jf_pass_through_result to do all the value mappings.
1927 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
1928 type_preserved flag.
1929 (ipa_set_jf_cst_copy): New function.
1930 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
1931 (ipa_set_jf_arith_pass_through): Likewise.
1932 (ipa_set_ancestor_jf): Likewise.
1933 (compute_complex_assign_jump_func): Set type_preserved instead of
1935 (ipa_compute_jump_functions_for_edge): Likewise.
1936 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
1937 (update_jump_functions_after_inlining): Update type_preserved.
1938 Explicitely create jump functions when combining one with pass_through.
1939 (ipa_write_jump_function): Stream the type_preserved flags.
1940 (ipa_read_jump_function): Likewise.
1942 2013-08-27 Jakub Jelinek <jakub@redhat.com>
1943 Aldy Hernandez <aldyh@redhat.com>
1945 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
1946 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
1947 * function.h (struct function): Add has_force_vect_loops and
1949 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
1950 * gimple.c (gimple_build_omp_critical): Add KIND argument and
1952 * gimple.def: Update CLAUSES comments.
1953 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
1954 (gimple_build_omp_for): Add argument to prototype.
1955 (gimple_omp_for_kind): New.
1956 (gimple_omp_for_set_kind): New.
1957 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
1958 GOVD_DATA_SHARE_CLASS.
1959 (enum omp_region_type): Add ORT_SIMD.
1960 (gimple_add_tmp_var): Handle ORT_SIMD.
1961 (gimplify_var_or_parm_decl): Same.
1962 (is_gimple_stmt): Same.
1963 (omp_firstprivatize_variable): Same.
1964 (omp_add_variable): Only use splay_tree_insert if lookup failed.
1965 (omp_notice_variable): Handle ORT_SIMD.
1966 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
1967 (omp_check_private): Handle ORT_SIMD.
1968 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1970 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
1971 Handle OMP_CLAUSE_LASTPRIVATE.
1972 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1974 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
1975 (gimplify_expr): Handle OMP_SIMD.
1976 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
1977 (expand_GOMP_SIMD_VF): New.
1978 (expand_GOMP_SIMD_LAST_LANE): New.
1979 * internal-fn.def (GOMP_SIMD_LANE): New.
1980 (GOMP_SIMD_VF): New.
1981 (GOMP_SIMD_LAST_LANE): New.
1982 * omp-low.c: Include target.h.
1983 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
1985 (check_omp_nesting_restrictions): Same.
1987 (lower_rec_simd_input_clauses): New.
1988 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
1990 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
1991 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
1992 (expand_omp_build_assign): New.
1993 (expand_omp_for_init_counts): New.
1994 (expand_omp_for_init_vars): New.
1995 (extract_omp_for_update_vars): New.
1996 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
1997 and rewrite accordingly.
1998 (expand_omp_simd): New.
1999 (expand_omp_for): Use expand_omp_simd.
2000 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2001 (lower_omp_for): Do not lower the body.
2002 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2004 * tree-flow.h (find_omp_clause): Remove prototype.
2005 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2006 forcing vectorization of the loop, or if flag_tree_vectorize.
2007 (gate_tree_if_conversion): Similarly.
2008 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2009 gimple_build_omp_for.
2010 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2011 * tree-parloops (create_parallel_loop): Pass kind argument to
2012 gimple_build_omp_for.
2013 * tree-pretty-print.c (dump_omp_clause): Add cases for
2014 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2015 OMP_CLAUSE__SIMDUID_.
2016 (dump_generic_node): Handle OMP_SIMD.
2017 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2018 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2019 unroll OMP_SIMD loops here.
2020 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
2021 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
2023 (vect_analyze_data_refs): Handle simd loops.
2024 * tree-vect-loop.c (vectorizable_live_operation): Handle
2026 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
2027 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2028 (vectorizable_load): Same.
2029 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
2030 (struct simduid_to_vf): New.
2031 (simduid_to_vf::hash): New.
2032 (simduid_to-vf::equal): New.
2033 (struct simd_array_to_simduid): New.
2034 (simd_array_to_simduid::hash): New.
2035 (simd_array_to_simduid::equal): New.
2036 (adjust_simduid_builtins): New.
2037 (struct note_simd_array_uses_struct): New.
2038 (note_simd_array_uses_cb): New.
2039 (note_simd_array_uses): New.
2040 (vectorize_loops): Handle simd hints and adjust simd builtins
2042 * tree-vectorizer.h (struct _stmt_vec_info): Add
2043 simd_lane_access_p field.
2044 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2045 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2046 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2047 (omp_clause_code_name): Same.
2048 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2049 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2050 * tree.def (OMP_SIMD): New entry.
2051 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2052 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
2053 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2054 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2055 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2056 (OMP_CLAUSE_LINEAR_STEP): New.
2057 (OMP_CLAUSE_SAFELEN_EXPR): New.
2058 (OMP_CLAUSE__SIMDUID__DECL): New.
2059 (find_omp_clause): New prototype.
2061 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2063 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2064 Haswell processor detection.
2066 2013-08-27 Christian Widmer <shadow@umbrox.de>
2069 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2070 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2071 AVX2 capable processors.
2073 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2075 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2076 of vget_low_* with implementations in terms of other intrinsics.
2078 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
2081 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2082 values to -1, 0 and 1.
2084 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2086 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2087 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2088 (<optab><mode>3_insn): Remove.
2089 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2090 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2091 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2092 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2093 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2094 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2095 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2096 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2097 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2098 (ror<mode>3_insn): Likewise.
2099 * config/aarch64/predicates.md (aarch64_simd_register): New.
2101 2013-08-27 Richard Biener <rguenther@suse.de>
2103 PR tree-optimization/57521
2104 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2105 one edge is non-critical.
2106 (find_phi_replacement_condition): Make sure to use a non-critical
2107 edge. Cleanup and remove old bug workarounds.
2108 (bb_postdominates_preds): Remove.
2109 (if_convertible_loop_p_1): Do not compute post-dominators.
2110 (combine_blocks): Do not free post-dominators.
2111 (main_tree_if_conversion): Likewise.
2112 (pass_data_if_conversion): Add TODO_verify_ssa.
2114 2013-08-27 DJ Delorie <dj@redhat.com>
2116 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2118 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2120 * function.c (assign_parm_find_data_types): Set passed_mode and
2121 nominal_mode to the TYPE_MODE of nominal_type for the built
2122 pointer type in case of the struct-pass-by-reference.
2124 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2126 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2127 (UINT16_TYPE): Change default to "unsigned int".
2129 * config/avr/avr.opt (mfract-convert-truncate): New option.
2130 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2131 is set, round negative fractional integers according to n1169
2132 when converting to integer types.
2134 2013-08-26 Jan Hubicka <jh@suse.cz>
2136 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2137 methods can not be called indirectly when their address is not taken.
2139 2013-08-26 Jan Hubicka <jh@suse.cz>
2141 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2144 2013-08-26 Jan Hubicka <jh@suse.cz>
2146 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2149 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2151 * reload.c (find_valid_class): Allow classes that do not include
2152 FIRST_PSEUDO_REGISTER - 1.
2154 2013-08-26 Jan Hubicka <jh@suse.cz>
2156 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2157 fix edge count/frequency when speculation failed; fix type check
2158 for the direct call.
2160 2013-08-26 Jan Hubicka <jh@suse.cz>
2162 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2164 2013-08-26 Jan Hubicka <jh@suse.cz>
2166 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2167 to be changed by edge redirection.
2169 2013-08-26 Jan Hubicka <jh@suse.cz>
2171 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2172 formating; add sanity check.
2173 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2174 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2175 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2176 (resolve_noninline_speculation): Update callee keys, too.
2178 2013-08-26 Jan Hubicka <jh@suse.cz>
2180 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2181 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2183 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2185 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2188 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
2190 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2191 (x86_64_elf_select_section): Ditto.
2192 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2194 (ix86_pass_by_reference): Ditto.
2195 (output_set_got): Ditto.
2196 (ix86_unary_operator_ok): Ditto.
2197 (ix86_expand_builtin): Ditto.
2199 2013-08-23 Jan Hubicka <jh@suse.cz>
2201 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
2203 2013-08-23 Jan Hubicka <jh@suse.cz>
2205 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2206 (tree_decl_with_vis): Add FINAL field.
2208 2013-08-23 Jeff Law <law@redhat.com>
2210 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2211 the debugging dump when the expression is fully redundant.
2213 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2215 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2216 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2217 * pretty-print.c (pp_formatted_text_data): Likewise.
2218 (pp_write_text_to_stream): Likewise.
2219 (pp_write_text_as_dot_label_to_stream): Likewise.
2220 (pp_append_r): Likewise.
2221 (pp_format): Likewise.
2222 (pp_flush): Likewise.
2223 (pp_clear_output_area): Likewise.
2224 (pp_append_text): Likewise.
2225 (pp_formatted_text): Likewise.
2226 (pp_remaining_character_count_for_line): Likewise.
2227 (pp_newline): Likewise.
2228 (pp_character): Likewise.
2229 (output_buffer::~output_buffer): Define.
2230 (pretty_printer::~pretty_printer): Destruct output buffer.
2231 * pretty-print.h (output_buffer::~output_buffer): Declare.
2232 (pretty_printer::~pretty_printer): Declare virtual.
2234 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
2237 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2238 HOST_WIDE_INT_M1U): New macros.
2239 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2240 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2241 unsigned -1 for lshift.
2242 * cse.c (cse_insn): Likewise.
2243 * double-int.c (rshift_double, lshift_double): Likewise.
2244 * builtins.c (fold_builtin_bitop): Likewise.
2245 * combine.c (force_to_mode): Likewise.
2246 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2247 * simplify-rtx.c (simplify_const_unary_operation,
2248 simplify_const_binary_operation): Likewise.
2249 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2250 check_va_list_escapes): Likewise.
2251 * rtlanal.c (nonzero_bits1): Likewise.
2252 * expmed.c (expand_smod_pow2): Likewise.
2253 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2255 2013-08-23 Jan Hubicka <jh@suse.cz>
2257 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2258 as having address taken.
2260 2013-08-23 Jan Hubicka <jh@suse.cz>
2262 * ipa-utils.h (method_class_type): Declare.
2263 * ipa-devirt.c (method_class_type): Export.
2265 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2266 do not walk base classes of anonymous types.
2268 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2270 PR rtl-optimization/58220
2272 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2273 handle SEQUENCE insns properly.
2275 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2277 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2279 (pp_newline_and_indent): Likewise.
2280 (pp_separate_with): Likewise.
2281 * pretty-print.c (pp_newline_and_flush): Define.
2282 (pp_newline_and_indent): Likewise.
2283 (pp_separate_with): Likewise.
2285 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2288 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2289 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2291 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2293 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
2294 * gcc/config/i386/i386.md (*movti_internal): Use
2295 predicate to determine if EVEX is needed.
2296 (*movsi_internal): Ditto.
2297 (*movdf_internal): Ditto.
2298 (*movsf_internal): Ditto.
2299 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
2301 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2303 PR tree-optimization/58209
2304 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2305 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2306 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2307 emit POINTER_PLUS_EXPR.
2308 (create_tailcall_accumulator): For pointer result type accumulate in
2311 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2313 * configure.ac: Add backslashes missing from the last change.
2314 * configure: Regenerate.
2316 2013-08-22 Jan Hubicka <jh@susue.cz>
2318 * ipa.c (function_and_variable_visibility): First remember function
2319 was global and then make it local.
2321 2013-08-22 Julian Brown <julian@codesourcery.com>
2323 * configure.ac: Add aarch64 to list of arches which use "nop" in
2325 * configure: Regenerate.
2327 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2329 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2330 gnu_libc_has_function.
2331 * config/s390/tpf.h: Likewise.
2333 2013-08-22 Jan Hubicka <jh@susue.cz>
2335 * timevar.c (validate_phases): Add cast.
2337 2013-08-22 Jan Hubicka <jh@susue.cz>
2339 * timevar.c (validate_phases): Use size_t for memory.
2340 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2342 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2344 * pretty-print.h (output_buffer::output_buffer): Declare.
2345 (pretty_printer::pretty_printer): Likewise.
2346 (pp_construct): Remove.
2347 * pretty-print.c (output_buffer::output_buffer): Define.
2348 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2349 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2350 (print_gimple_expr): Likewise.
2351 (print_gimple_seq): Likewise.
2352 (gimple_dump_bb): Likewise.
2353 * sched-vis.c (dump_value_slim): Likewise.
2354 (dump_insn_slim): Likewise.
2355 (dump_rtl_slim): Likewise.
2356 (str_pattern_slim): Likewise.
2357 * tree-mudflap.c (mf_varname_tree): Likewise.
2358 * graph.c (print_graph_cfg): Likewise.
2359 (start_graph_dump): Likewise.
2360 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2362 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2363 pretty printer initialization.
2364 * coretypes.h (diagnostic_context): Remove superflous type alias
2366 (pretty_printer): Likewise. Declare directly as a class.
2367 (pretty_print_info): Remove declaration as class.
2368 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
2369 and pp_clear_output_area.
2370 (asan_add_global): Likewise.
2372 2013-08-22 Jan Hubicka <jh@suse.cz>
2374 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
2375 * ipa-utils.h (update_type_inheritance_graph): Declare.
2376 (possible_polymorphic_call_target_p): Declare.
2377 (possible_polymorphic_call_target_p): New.
2378 * ipa-devirt.c: Update toplevel comments.
2379 (cached_polymorphic_call_targets): Move up.
2380 (odr_type_d): Move ID down.
2381 (polymorphic_type_binfo_p): Update comment.
2382 (odr_hasher::remove): Likewise;
2383 (get_odr_type): Set anonymous_namespace.
2384 (dump_odr_type): Dump it.
2385 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
2386 (maybe_record_node): Record node in cached_polymorphic_call_targets.
2387 (record_binfo): Add comment.
2388 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
2390 (devirt_node_removal_hook): Do not iCE when cache is freed.
2391 (possible_polymorphic_call_target_p): New predicate.
2392 (update_type_inheritance_graph): New function.
2394 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
2395 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2396 Sergey Lega <sergey.s.lega@intel.com>
2397 Anna Tikhonova <anna.tikhonova@intel.com>
2398 Ilya Tocar <ilya.tocar@intel.com>
2399 Andrey Turetskiy <andrey.turetskiy@intel.com>
2400 Ilya Verbin <ilya.verbin@intel.com>
2401 Kirill Yukhin <kirill.yukhin@intel.com>
2402 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2404 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
2405 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
2406 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
2407 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
2408 (OPTION_MASK_ISA_AVX2_UNSET): Update.
2409 (OPTION_MASK_ISA_AVX512F_UNSET): New.
2410 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
2411 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
2412 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
2413 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
2414 OPT_mavx512pf, OPT_mavx512er cases.
2415 * config/i386/constraints.md (v): New constraint.
2416 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
2417 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
2418 (bit_AVX512CD): New.
2419 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
2420 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
2421 * config/i386/i386-c.c (ix86_target_macros_internal):
2422 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
2424 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
2425 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
2426 * config/i386/i386.c (regclass_map, dbx_register_map)
2427 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
2428 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
2429 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
2431 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
2432 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
2433 -mavx512pf options. Fix formatting.
2434 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
2435 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
2436 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
2437 -mavx512cd, -mavx512pf options.
2438 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
2439 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
2440 (ix86_preferred_output_reload_class): Replace SSE_REGS with
2442 (ix86_hard_regno_mode_ok): Support 512-bit registers.
2443 (ix86_set_reg_reg_cost): Ditto.
2444 (x86_order_regs_for_local_alloc): Ditto.
2445 (MAX_VECT_LEN): Extend to 64-byte.
2446 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
2447 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
2448 (TARGET_AVX512ER, TARGET_AVX512CD): New.
2449 (BIGGEST_ALIGNMENT): Extend to 512-bits.
2450 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
2451 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
2452 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
2453 (SSE_REG_MODE_P): Support new modes.
2454 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
2455 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
2456 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
2457 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
2458 (REG_CLASS_CONTENTS): Add new registers.
2459 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
2460 (EXT_REX_SSE_REGNO_P): New.
2461 (HI_REGISTER_NAMES): Add new registers.
2462 * config/i386/i386.md: Define constants for new registers.
2463 (mode): Add new 512-bit modes.
2464 (prefix): Support evex prefix.
2465 (isa): Support avx512f, noavx512f, fma_avx512f.
2466 (ssemodesuffix): Add new 512-bit modes.
2468 (*movxi_internal_avx512f): Ditto.
2469 (*movdi_internal): Replace constraint "x" with the new constraint "v".
2471 (*movsi_internal): Likewise.
2472 (*movdf_internal): Likewise.
2473 (*movsf_internal): Likewise.
2474 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
2475 (<code><mode>3): Likewise.
2476 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
2478 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
2479 with the new constraint "v".
2480 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
2482 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
2483 with the new constraint "v".
2484 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
2485 (<sse2>_storedqu<avxsizesuffix>): Likewise.
2486 (*<plusminus_insn><mode>3): Likewise.
2487 (<sse>_vm<plusminus_insn><mode>3): Likewise.
2488 (*mul<mode>3): Likewise.
2489 (<sse>_vmmul<mode>3): Likewise.
2490 (<sse>_div<mode>3): Likewise.
2491 (<sse>_vmdiv<mode>3): Likewise.
2492 (<sse>_sqrt<mode>2): Likewise.
2493 (<sse>_vmsqrt<mode>2): Likewise.
2494 (*<code><mode>3_finite): Likewise.
2495 (*<code><mode>3) <smaxmin>: Likewise.
2496 (<sse>_vm<code><mode>3): Likewise.
2497 (*<code><mode>3) <any_logic>: Likewise.
2498 (*fma_fmadd_<mode>): Likewise.
2499 (*fma_fmsub_<mode>): Likewise.
2500 (*fma_fnmadd_<mode>): Likewise.
2501 (*fma_fnmsub_<mode>): Likewise.
2502 (*fma_fmaddsub_<mode>): Likewise.
2503 (*fma_fmsubadd_<mode>): Likewise.
2504 (*fmai_fmadd_<mode>): Likewise.
2505 (*fmai_fmsub_<mode>): Likewise.
2506 (*fmai_fnmadd_<mode>): Likewise.
2507 (*fmai_fnmsub_<mode>): Likewise.
2508 (sse_cvtsi2ss): Likewise.
2509 (sse_cvtsi2ssq): Likewise.
2510 (sse_cvtss2si): Likewise.
2511 (sse_cvtss2si_2): Likewise.
2512 (sse_cvtss2siq): Likewise.
2513 (sse_cvtss2siq_2): Likewise.
2514 (sse_cvttss2si): Likewise.
2515 (sse_cvtss2siq_2): Likewise.
2516 (float<sseintvecmodelower><mode>2): Likewise.
2517 (sse2_cvtsd2si_2): Likewise.
2518 (sse2_cvtsd2siq_2): Likewise.
2519 (*<plusminus_insn><mode>3): Likewise.
2520 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
2521 (*<sse4_1_avx2>_mul<mode>3): Likewise.
2522 (ashr<mode>3): Likewise.
2523 (<shift_insn><mode>3): Likewise.
2524 (avx2_<code><mode>3): Likewise.
2525 (*avx2_<code><mode>3): Likewise.
2526 (*andnot<mode>3): Likewise.
2527 (*<code><mode>3) <any_logic>: Likewise.
2528 (abs<mode>2): Likewise.
2529 (avx2_permvar<mode>): Likewise.
2530 (avx2_perm<mode>_1): Likewise.
2531 (*avx_vpermilp<mode>): Likewise.
2532 (avx_vpermilvar<mode>3): Likewise.
2533 (avx2_ashrv<mode>): Likewise.
2534 (avx2_<shift_insn>v<mode>): Likewise.
2535 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
2537 * doc/rtl.texi: Document XImode.
2539 2013-08-21 Jeff Law <law@redhat.com>
2541 * tree-flow.h (register_jump_thread): Pass vector of edges
2542 instead of each important edge.
2543 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
2544 thread path into a vector and pass that to register_jump_thread.
2545 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
2546 passed in edge vector to the current 3-edge form.
2549 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2551 * dce.c (fini_dce): Call df_analyze again just in case
2552 delete_unmarked_insns removed anything.
2554 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
2556 * reload.h (struct reg_equivs): Rename to ..
2557 (struct reg_equivs_s): .. this.
2559 2013-08-20 Martin Liska <marxin.liska@gmail.com>
2561 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
2563 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2565 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
2567 2013-08-21 Jeff Law <law@redhat.com>
2569 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
2570 simplify assignments too. If the RHS collapses to a singleton
2571 range, then return the value for the range.
2573 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
2575 * config/i386/sse.md (V16): Rename to...
2577 (mov<mode>): Update iterator name.
2578 (*mov<mode>_internal): Ditto.
2579 (push<mode>1): Ditto.
2580 (movmisalign<mode>): Ditto.
2582 2013-08-20 Jan Hubicka <jh@suse.cz>
2585 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
2586 entry for direct edges.
2587 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
2589 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2591 Revert my last two changes, r201865 and r201864:
2594 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2596 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2597 instances can own GC refs.
2599 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2600 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2601 (gcc::context::gt_pch_nx): Likewise.
2602 (gcc::context::gt_pch_nx): Likewise.
2603 * ggc.h (gt_ggc_mx <T>): New.
2604 (gt_pch_nx_with_op <T>): New.
2605 (gt_pch_nx <T>): New.
2606 * passes.c (opt_pass::gt_ggc_mx): New.
2607 (opt_pass::gt_pch_nx): New.
2608 (opt_pass::gt_pch_nx_with_op): New.
2609 (pass_manager::gt_ggc_mx): New.
2610 (pass_manager::gt_pch_nx): New.
2611 (pass_manager::gt_pch_nx_with_op): New.
2612 (pass_manager::operator new): Use
2613 ggc_internal_cleared_alloc_stat rather than xcalloc.
2614 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2615 (pass_manager::gt_ggc_mx): New.
2616 (pass_manager::gt_pch_nx): New.
2617 (pass_manager::gt_pch_nx_with_op): New.
2618 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2619 (opt_pass::operator new): New.
2620 (opt_pass::gt_ggc_mx): New.
2621 (opt_pass::gt_pch_nx): New.
2622 (opt_pass::gt_pch_nx_with_op): New.
2625 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2627 * Makefile.in (GTFILES): Add context.h.
2628 * context.c (gcc::context::operator new): New.
2629 (gcc::context::gt_ggc_mx): New.
2630 (gcc::context::gt_pch_nx): New.
2631 (gcc::context::gt_pch_nx): New.
2632 * context.h (gcc::context): Add GTY((user)) marking.
2633 (gcc::context::operator new): New.
2634 (gcc::context::gt_ggc_mx): New.
2635 (gcc::context::gt_pch_nx): New.
2636 (gcc::context::gt_pch_nx): New.
2637 (g): Add GTY marking.
2638 (gt_ggc_mx (gcc::context *)): New.
2639 (gt_pch_nx (gcc::context *)): New.
2640 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2641 void *cookie)): New.
2642 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2644 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2646 * dce.c (fini_dce): Call df_analyze again just in case
2647 delete_unmarked_insns removed anything.
2649 2013-08-20 Teresa Johnson <tejohnson@google.com>
2651 PR rtl-optimizations/57451
2652 * final.c (reemit_insn_block_notes): Prevent lexical blocks
2653 from crossing split section boundaries.
2655 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2657 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
2658 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
2659 with MULTLIB_DEFAULTS.
2661 2013-08-20 Nick Clifton <nickc@redhat.com>
2663 * target.def (narrow_volatile_bitfield): Note that the default
2664 value is false, not !TARGET_STRICT_ALIGN.
2665 * doc/tm.texi: Regenerate.
2667 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
2669 Fix LIB_SPEC for systems without libpthread.
2671 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
2672 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
2674 * config/i386/linux-common.h: Likewise.
2675 * config/mips/linux-common.h: Likewise.
2677 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2679 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
2682 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2684 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2685 instances can own GC refs.
2687 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2688 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2689 (gcc::context::gt_pch_nx): Likewise.
2690 (gcc::context::gt_pch_nx): Likewise.
2691 * ggc.h (gt_ggc_mx <T>): New.
2692 (gt_pch_nx_with_op <T>): New.
2693 (gt_pch_nx <T>): New.
2694 * passes.c (opt_pass::gt_ggc_mx): New.
2695 (opt_pass::gt_pch_nx): New.
2696 (opt_pass::gt_pch_nx_with_op): New.
2697 (pass_manager::gt_ggc_mx): New.
2698 (pass_manager::gt_pch_nx): New.
2699 (pass_manager::gt_pch_nx_with_op): New.
2700 (pass_manager::operator new): Use
2701 ggc_internal_cleared_alloc_stat rather than xcalloc.
2702 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2703 (pass_manager::gt_ggc_mx): New.
2704 (pass_manager::gt_pch_nx): New.
2705 (pass_manager::gt_pch_nx_with_op): New.
2706 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2707 (opt_pass::operator new): New.
2708 (opt_pass::gt_ggc_mx): New.
2709 (opt_pass::gt_pch_nx): New.
2710 (opt_pass::gt_pch_nx_with_op): New.
2712 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2714 * Makefile.in (GTFILES): Add context.h.
2715 * context.c (gcc::context::operator new): New.
2716 (gcc::context::gt_ggc_mx): New.
2717 (gcc::context::gt_pch_nx): New.
2718 (gcc::context::gt_pch_nx): New.
2719 * context.h (gcc::context): Add GTY((user)) marking.
2720 (gcc::context::operator new): New.
2721 (gcc::context::gt_ggc_mx): New.
2722 (gcc::context::gt_pch_nx): New.
2723 (gcc::context::gt_pch_nx): New.
2724 (g): Add GTY marking.
2725 (gt_ggc_mx (gcc::context *)): New.
2726 (gt_pch_nx (gcc::context *)): New.
2727 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2728 void *cookie)): New.
2729 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2731 2013-08-20 Alan Modra <amodra@gmail.com>
2734 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
2735 (rs6000_emit_epilogue): Likewise.
2737 2013-08-19 Dehao Chen <dehao@google.com>
2739 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
2741 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
2742 Jakub Jelinek <jakub@redhat.com>
2744 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
2745 (BUILT_IN_FABSD64): Likewise.
2746 (BUILT_IN_FABSD128): Likewise.
2747 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
2748 (fold_builtin_1): Likewise.
2749 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
2750 destination and source operands.
2751 (*abstd2_fpr): Likewise.
2752 (*nabstd2_fpr): Likewise.
2754 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
2756 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
2759 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
2761 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
2764 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
2766 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
2767 * builtins.c (default_libc_has_function): New.
2768 (gnu_libc_has_function): Ditto.
2769 (no_c99_libc_has_function): Ditto.
2770 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
2771 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
2772 (fold_builtin_sincos): Likewise.
2773 (fold_builtin_cexp): Likewise.
2774 * builtins.def (DEF_C94_BUILTIN): Likewise.
2775 (DEF_C99_BUILTIN): Likewise.
2776 (DEF_C99_C90RES_BUILTIN): Likewise.
2777 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
2778 definitions to using this define.
2779 * config/darwin-protos.h (darwin_libc_has_function): New.
2780 * config/darwin.c (darwin_libc_has_function): Ditto.
2781 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
2782 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
2783 * config/darwin.h: Ditto.
2784 * config/elfos.h: Ditto.
2785 * config/freebsd.h: Ditto.
2786 * config/i386/cygming.h: Ditto.
2787 * config/i386/djgpp.h: Ditto.
2788 * config/i386/i386-interix.h: Ditto.
2789 * config/microblaze/microblaze.h: Ditto.
2790 * config/mmix/mmix.h: Ditto.
2791 * config/gnu-user.h: Ditto.
2792 * config/ia64/hpux.h: Ditto.
2793 * config/pa/pa-hpux.h: Ditto.
2794 * config/pdp11/pdp11.h: Ditto.
2795 * config/picochip/picochip.h: Ditto.
2796 * config/linux.h: Ditto.
2797 * config/netbsd.h: Ditto.
2798 * config/openbsd.h: Ditto.
2799 * config/rs6000/aix43.h: Ditto.
2800 * config/rs6000/aix51.h: Ditto.
2801 * config/rs6000/aix52.h: Ditto.
2802 * config/rs6000/aix53.h: Ditto.
2803 * config/rs6000/aix61.h: Ditto.
2804 * config/rs6000/darwin.h: Ditto.
2805 * config/rs6000/linux.h: Ditto.
2806 * config/rs6000/linux64.h: Ditto.
2807 * config/s390/tpf.h: Ditto.
2808 * config/sol2-10.h: Ditto.
2809 * config/sol2.h: Ditto.
2810 * config/vms/vms.h: Ditto.
2811 * config/vxworks.h: Ditto.
2812 * config/linux-android.c (linux_android_libc_has_function):
2813 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
2814 * config/linux-protos.h (linux_android_libc_has_function):
2816 * config/i386/i386.c (ix86_libc_has_function): New.
2817 * config/i386/i386-protos.h
2818 (ix86_libc_has_function): New declaration.
2819 * config/i386/i386.md
2820 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
2821 ("isinf<mode>2): Likewise.
2822 * convert.c (convert_to_integer): Using new target hook
2823 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
2824 TARGET_C99_FUNCTIONS.
2825 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
2826 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
2827 * coretypes.h (function_class): New enum for different
2828 classes of functions.
2829 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
2830 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
2831 (TARGET_HAS_SINCOS): Likewise.
2832 (TARGET_LIBC_HAS_FUNCTION): New.
2833 * doc/tm.texi: Regenerated.
2834 * targhooks.h (default_libc_has_function): New declaration.
2835 (no_c99_libc_has_function): Ditto.
2836 (gnu_libc_has_function): Ditto.
2837 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
2838 and TARGET_HAS_SINCOS.
2840 2013-08-18 Jan Hubicka <jh@suse.cz>
2842 * Makeifle-in (ipa-devirt.o): New.
2843 (GTFILES): Add ipa-utils.h and ipa-devirt.c
2844 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
2845 (analyze_functions): Look into possible targets of polymorphic call.
2846 * dumpfile.c (dump_files): Add type-inheritance dump.
2847 * dumpfile.h (TDI_inheritance): New.
2848 * ipa-devirt.c: New file.
2849 * ipa-utils.h (odr_type_d): Forward declare.
2850 (odr_type): New type.
2851 (build_type_inheritance_graph): Declare.
2852 (possible_polymorphic_call_targets): Declare and introduce inline
2853 variant when only edge is pased.
2854 (dump_possible_polymorphic_call_targets): Likewise.
2855 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
2856 * tree.c (type_in_anonymous_namespace_p): Break out from ...
2857 (types_same_for_odr): ... here.
2858 * tree.h (type_in_anonymous_namespace_p): Declare.
2860 2013-08-18 Jakub Jelinek <jakub@redhat.com>
2862 PR tree-optimization/58006
2863 * tree-parloops.c (take_address_of): Don't ICE if get_name
2865 (eliminate_local_variables_stmt): Remove clobber stmts.
2867 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
2869 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
2870 error message is issued for an alias to undefined symbol.
2872 2013-08-18 Jan Hubicka <jh@suse.cz>
2874 * cgraph.c (cgraph_create_indirect_edge): Discover
2875 polymorphic calls and record basic info into indirect_info.
2876 * gimple-fold.c (gimple_fold_call): When doing BINFO based
2877 devirtualization, ignore objc function calls.
2878 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
2879 call with no parm index info.
2880 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
2881 * tree.c (virtual_method_call_p): New function.
2882 * tree.h (virtual_method_call_p): Declare.
2884 2013-08-16 Jan Hubicka <jh@suse.cz>
2887 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
2889 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
2891 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
2894 2013-08-16 David Malcolm <dmalcolm@redhat.com>
2896 * gengtype.c (type_for_name): Add special-case support for
2897 locating types within the "gcc::" namespace.
2898 (open_base_files): Emit a "using namespace gcc" directive.
2900 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2903 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
2904 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
2906 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
2907 array instead of each individual operand as a separate argument.
2908 (emit_fusion_gpr_load): Likewise.
2909 (expand_fusion_gpr_load): Add new function declaration.
2911 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
2912 signature to have the operands passed as an array, instead of as
2913 separate arguments. Allow ZERO_EXTEND to be in the memory
2914 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
2915 depend on the register live/dead flags when peepholes are run.
2916 (expand_fusion_gpr_load): New function to be called from the
2917 peephole2 pass, to change the register that addis sets to be the
2919 (emit_fusion_gpr_load): Change the calling signature to have the
2920 operands passed as an array, instead of as separate arguments.
2921 Allow ZERO_EXTEND to be in the memory address, and also
2922 SIGN_EXTEND if -mpower8-fusion-sign.
2924 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
2926 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
2927 adjust the register addis loads up in the peephole2 pass. Do not
2928 depend on the register live/dead state when the peephole pass is done.
2930 2013-08-16 David Malcolm <dmalcolm@redhat.com>
2932 * gengtype.c (create_user_defined_type): Ensure that the kind
2933 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
2934 declaration is seen before the GTY((user)) marking.
2936 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
2939 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
2941 2013-08-16 Jan Hubicka <jh@suse.cz>
2943 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
2944 arugment expected_type.
2945 (gimple_fold_call): Use it.
2946 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
2947 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
2948 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
2949 (try_make_edge_direct_virtual_call): Likewise.
2950 * tree.c (obj_type_ref_class): New.
2951 * tree.h (obj_type_ref_class): Use it.
2953 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
2955 * sched-vis.c (rtl_slim_pp_initialized): Remove.
2956 (rtl_slim_pp): Likewise.
2957 (init_rtl_slim_pretty_print): Likewise.
2958 (dump_value_slim): Don't call it. Use local pretty printer.
2959 (dump_insn_slim): Likewise.
2960 (dump_rtl_slim): Likewise.
2961 (str_pattern_slim): Likewise.
2962 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
2965 2013-08-16 Jakub Jelinek <jakub@redhat.com>
2967 PR tree-optimization/58164
2968 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
2969 walk gimple_goto_dest of GIMPLE_GOTO.
2971 PR tree-optimization/58165
2972 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
2973 bi_call must be the last stmt in a bb, don't split_block, instead
2974 use fallthru edge from it and give up if there is none.
2975 Release conds vector when returning early.
2977 2013-08-14 Xinliang David Li <davidxl@google.com>
2979 * config/i386/i386.c (ix86_option_override_internal):
2980 Remove unused variable and field.
2982 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2985 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
2986 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
2987 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
2988 and Linux, correct BLKmode alignment when 128-bit alignment is
2989 required and compatibility flag is not set.
2990 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
2991 for zero-size arguments when compatibility flag is not set.
2993 2013-08-14 Jakub Jelinek <jakub@redhat.com>
2995 PR tree-optimization/58145
2996 * tree-sra.c (build_ref_for_offset): If prev_base has
2997 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
2999 2013-08-14 Xinliang David Li <davidxl@google.com>
3001 * config/i386/i386.c (ix86_option_override_internal):
3002 Fix uninitialized variable error.
3004 2013-08-14 Xinliang David Li <davidxl@google.com>
3006 * config/i386/i386.opt: Define two new options.
3007 * config/i386/x86-tune.def: Add arch selector field in macros.
3008 * config/i386/i386.h: Adjust macro definition.
3009 * config/i386/i386.c (ix86_option_override_internal):
3011 (parse_mtune_ctrl_str): New function.
3012 (set_ix86_tune_features): New function.
3013 (ix86_function_specific_restore): Call the new helper function.
3015 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
3017 PR rtl-optimization/57662
3018 * sel-sched.c (code_motion_process_successors): When the current insn
3019 is removed after the recursive traversal, break from the loop.
3020 Add comments and debug printouts.
3022 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3023 Alexandre Oliva <aoliva@redhat.com>
3026 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3027 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3028 there also UNSPEC_PLTOFF.
3030 2013-08-14 Marek Polacek <polacek@redhat.com>
3032 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3033 NUM_CONDITIONS bit positions.
3035 2013-08-13 Cary Coutant <ccoutant@google.com>
3037 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3038 (attr_checksum): Hash vector contents instead of pointer.
3039 (attr_checksum_ordered): Likewise.
3041 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
3043 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3044 when Pmode != word_mode. Add length_address attribute.
3045 (sse3_monitor_<mode>): Merge from sse3_monitor and
3046 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3047 Pmode != word_mode. Update insn length attribute.
3048 * config/i386/i386.c (ix86_option_override_internal): Update
3049 ix86_gen_monitor selection for merged sse3_monitor insn.
3051 2013-08-13 Julian Brown <julian@codesourcery.com>
3053 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3054 perform invalid legitimization on greater-than-word-size modes for
3057 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3059 * ira.c (setup_class_translate_array): Use aclass instead of cl
3060 for classes not fully covered by allocno classes.
3062 2013-08-13 Jakub Jelinek <jakub@redhat.com>
3064 PR tree-optimization/57661
3065 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3066 * tree-inline.c (tree_function_versioning): Initialize it.
3067 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3068 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3069 that is not being copied.
3072 * asan.c (instrument_strlen_call): Fix typo in comment.
3073 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3075 2013-08-13 Steve Ellcey <sellcey@mips.com>
3077 * config/mips/mips.md (prefetch): Use lw instead of ld on
3078 loongson in 32bit mode.
3080 2013-08-13 Nick Clifton <nickc@redhat.com>
3082 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3084 2013-08-13 Jan Hubicka <jh@suse.cz>
3086 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3087 introduced edge; fix typo in sanity check.
3088 (cgraph_resolve_speculation): Export; improve diagnostic.
3089 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3090 speculation at type mismatch.
3091 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3092 (cgraph_resolve_speculation): Declare.
3093 (symtab_can_be_discarded): New function.
3094 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3095 * ipa-inline-transform.c (speculation_removed): New global var.
3096 (clone_inlined_nodes): See if speculation can be removed.
3097 (inline_call): If speculations was removed, we growths may not match.
3098 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3099 (speculation_useful_p): New function.
3100 (resolve_noninline_speculation): New function.
3101 (inline_small_functions): Resolve useless speculations.
3102 * ipa-inline.h (speculation_useful_p): Declare
3103 * ipa.c (can_replace_by_local_alias): Simplify.
3104 (ipa_profile): Produce speculative calls in non-lto, too;
3105 add simple cost model; produce local aliases.
3107 2013-08-13 David Malcolm <dmalcolm@redhat.com>
3109 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
3112 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3114 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3116 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
3118 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3119 iterator instead of X87MODEF.
3121 2013-08-12 Perez Read <netfirewall@gmail.com>
3124 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3125 operand 0 for intel asm alternative.
3126 (*movabs<mode>_2): Ditto for operand 1.
3128 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3130 * config/aarch64/arm_none.h
3131 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3133 2013-08-12 Nick Clifton <nickc@redhat.com>
3135 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3137 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3139 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3140 expand for QI/HImode operand to produce more effictive code for
3141 unsigned char(short) --> float(double) conversion.
3143 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
3145 * doc/invoke.texi: Mention that -ftls-model does not force the final
3148 2013-08-12 Marek Polacek <polacek@redhat.com>
3149 Marc Glisse <marc.glisse@inria.fr>
3151 PR tree-optimization/57980
3152 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3153 when creating -1 constant.
3155 2013-08-10 Jan Hubicka <jh@suse.cz>
3157 Workaround binutils PR14342.
3158 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3159 (gimple_init_edge_profiler): Likewise.
3160 (gimple_gen_ic_func_profiler): Likewise.
3162 2013-08-09 Jan Hubicka <jh@suse.cz>
3164 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3166 2013-08-09 Xinliang David Li <davidxl@google.com>
3168 * config/i386/stringop.def: New file.
3169 * config/i386/stringop.opt: New file.
3170 * config/i386/i386-opts.h: Include stringopt.def.
3171 * config/i386/i386.opt: Include stringopt.opt.
3172 * config/i386/i386.c (ix86_option_override_internal):
3173 Override default size based stringop inline strategies with options.
3174 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3177 2013-08-09 Jan Hubicka <jh@suse.cz>
3179 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
3181 2013-08-09 Jan Hubicka <jh@suse.cz>
3183 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3185 (dump_cgraph_node): Dump profile-id.
3186 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3187 and common_target_probability.
3188 * lto-cgraph.c (lto_output_edge): Stream common targets.
3189 (lto_output_node): Stream profile ids.
3190 (input_node): Stream profile ids.
3191 (input_edge): Stream common targets.
3192 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3193 * ipa.c: Include value-prof.h
3194 (ipa_profile_generate_summary): Turn indirect call statement histograms
3195 into common targets.
3196 (ipa_profile): Turn common targets into speculative edges.
3198 2013-08-09 Jan Hubicka <jh@suse.cz>
3200 * cgraph.h (cgraph_node): Add profile_id.
3201 * value-prof.c (cgraph_node_map): Turn into pointer_map.
3202 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
3203 (del_node_map): Update.
3204 (find_func_by_funcdef_no): Replace by ...
3205 (find_func_by_profile_id): ... this one.
3206 (gimple_ic_transform): Do not remove useful histograms when
3207 speculation is not done; dump info when indirect call removal
3209 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3210 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3211 (__gcov_indirect_call_profiler_v2): .. this one.
3212 * profile.h (init_node_map): Update.
3213 * coverage.c (coverage_compute_profile_id): New function.
3214 * coverage.h (coverage_compute_profile_id): Declare.
3215 * tree-profile.c (init_ic_make_global_vars): Make
3216 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3217 (gimple_init_edge_profiler): Update prototype of
3218 __gcov_indirect_call_profiler.
3219 (gimple_gen_ic_func_profiler): Simplify.
3220 (tree_profiling): Use init_node_map
3222 2013-08-09 Jan Hubicka <jh@suse.cz>
3224 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3225 non-speculative refs.
3226 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3227 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3228 (cgraph_set_call_stmt): Likewise.
3229 (cgraph_create_edge_1): Fix release checking compilatoin;
3231 (cgraph_free_edge): Free indirect info.
3232 (cgraph_turn_edge_to_speculative): New function.
3233 (cgraph_speculative_call_info): New function.
3234 (cgraph_make_edge_direct): Return direct edge; handle speculation.
3235 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
3236 (dump_cgraph_node): Dump speculation.
3237 (verify_edge_count_and_frequency): Accept speculative edges.
3238 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3239 (verify_cgraph_node): Handle speculation.
3240 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3241 (cgraph_set_call_stmt): Update prototype.
3242 (cgraph_make_edge_direct): Update prototype.
3243 (cgraph_speculative_call_info): Declare.
3244 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3245 to change; update call of ipa_find_references.
3246 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3247 lto_stmt_uid and speculative flags.
3248 (ipa_dump_references): Dump speculation.
3249 (ipa_clone_references): Clone speculative flag.
3250 (ipa_clone_referring): Likewise.
3251 (ipa_clone_ref): New function.
3252 (ipa_find_reference): Look into lto_stmt_uids
3253 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3254 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3255 (ipa_find_reference): Update declaration.
3256 (ipa_clone_ref): Declare.
3257 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3258 stream speculative flag.
3259 (lto_output_ref): Stream statements uids and speculation.
3260 (input_ref): Likewise.
3261 (input_edge): Stream speuclation.
3262 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3263 (cgraph_set_call_stmt_including_clones): Handle speculation.
3264 * ipa-inline.c (heap_edge_removal_hook): New function.
3265 (inline_small_functions): Register it.
3266 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3267 also initialize refs.
3268 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3270 (try_make_edge_direct_simple_call): Likewise.
3271 (try_make_edge_direct_simple_call): Likewise.
3272 (update_indirect_edges_after_inlining): Likewise.
3273 (remove_described_reference): Look proper lto_stmt_uid.
3274 (propagate_controlled_uses): Likewise.
3275 (propagate_controlled_uses): Liekwise.
3276 * tree-inline.c (copy_bb): Copy speculative edges.
3277 (redirect_all_calls): New function.
3278 (copy_cfg_body): Do redirection after loop info is updated.
3279 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
3281 2013-08-09 Jan Hubicka <jh@suse.cz>
3283 * lto-streamer-out.c (output_function): Renumber PHIs.
3284 * lto-streamer-in.c (input_function): Likewise.
3286 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3288 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3289 (get_lane_unsigned): Likewise.
3290 (dup_lane_scalar): Likewise.
3291 (get_lane): enable for VALL.
3292 * config/aarch64/aarch64-simd.md
3293 (aarch64_dup_lane_scalar<mode>): Remove.
3294 (aarch64_get_lane_signed<mode>): Likewise.
3295 (aarch64_get_lane_unsigned<mode>): Likewise.
3296 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3297 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3298 (aarch64_get_lane<mode>): Enable for all vector modes.
3299 (aarch64_get_lanedi): Remove misleading constraints.
3300 * config/aarch64/arm_neon.h
3301 (__aarch64_vget_lane_any): Define.
3302 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3303 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3304 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3305 * config/aarch64/iterators.md (VDQQH): New.
3309 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3311 * configure.ac: Add GAS check for LEON instructions on SPARC.
3312 * configure: Regenerate.
3313 * config.in: Likewise.
3314 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3316 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3317 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3319 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3320 (AS_LEON_FLAG): New macro.
3321 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3322 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3323 Deal with LEON and LEON3 for the memory model.
3324 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
3325 (atomic_compare_and_swap<mode>_1): Likewise.
3326 (*atomic_compare_and_swap<mode>_1): Likewise.
3328 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3330 * config/arm/neon.md (vcond): Fix floating-point vector
3331 comparisons against 0.
3333 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3335 * lra-constraints.c (emit_spill_move): Remove assert.
3336 (process_alt_operands): Add more debugging
3337 output. Increase reject for spilling into memory. Decrease
3338 reject for reloading scratch.
3339 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3341 2013-08-08 Steve Ellcey <sellcey@mips.com>
3343 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3344 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3345 micromips incompatible. Add nan2008.
3346 (MULTILIB_DIRNAMES): Add nan2008.
3347 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3348 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3349 and micromips incompatible. Add nan2008.
3350 (MULTILIB_DIRNAMES): Add nan2008.
3351 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3353 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3355 PR rtl-optimization/58079
3356 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3357 simplify_comparison has widened a comparison with an integer.
3359 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3361 * config/arm/neon.md (movmisalign<mode>): Disable when we
3362 don't allow unaligned accesses.
3363 (*movmisalign<mode>_neon_store): Likewise.
3364 (*movmisalign<mode>_neon_load): Likewise.
3365 (*movmisalign<mode>_neon_store): Likewise.
3366 (*movmisalign<mode>_neon_load): Likewise.
3368 2013-08-08 Jan Hubicka <jh@suse.cz>
3370 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
3371 (make_pass_rebuild_cgraph_edges): Also clear references.
3372 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
3373 * ipa-inline-transform.c (inline_transform): Remove all references
3375 * cgraphunit.c (expand_function): Remove all references after
3377 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
3378 (ipa_find_reference): Rewrite to iterator.
3379 (remove_stmt_references): Likewise.
3380 (ipa_clear_stmts_in_references): New function.
3381 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
3382 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
3384 * ipa-split.c (split_function): Remove references in split function.
3386 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
3389 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
3390 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
3392 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
3395 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
3396 suggested mode for the assignment isn't compatible with the
3399 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3402 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
3404 2013-08-07 Xinliang David Li <davidxl@google.com>
3406 * config/i386/i386.opt: New option -mtune-ctrl=.
3407 * config/i386/x86-tune.def: New file.
3408 * config/i386/i386.h: include x86-tune.def.
3409 * config/i386/i386.c (ix86_option_override_internal):
3410 Parsing -mtune-ctrl= option and set tune features.
3412 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
3415 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
3416 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
3419 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
3421 * rtl.h (update_alignments): Declare.
3422 * final.c (grow_label_align): New function extracted from...
3423 (shorten_branches): ...here. Call it.
3424 (update_alignments): New function.
3425 * reorg.c (sibling_labels): New variable.
3426 (get_label_before): Add SIBLING parameter. If it is non-zero, push
3427 the new label along with it onto the sibling_labels vector.
3428 (fill_simple_delay_slots): Adjust call to get_label_before.
3429 (fill_slots_from_thread): Likewise.
3430 (relax_delay_slots): Likewise.
3431 (make_return_insns): Likewise.
3432 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
3434 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
3436 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
3437 document its semantics.
3438 (diagnostic_report_diagnostic): Adjust accordingly.
3440 2013-08-07 David Malcolm <dmalcolm@redhat.com>
3442 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
3443 (sparc_option_override): ...and port to new C++ pass API.
3444 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
3446 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
3448 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
3450 2013-08-06 Caroline Tice <cmtice@google.com>
3452 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
3453 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
3454 * tree-pass.h: Add pass_vtable_verify.
3455 * varasm.c (assemble_variable): Add code to properly set the comdat
3456 section and name for the .vtable_map_vars section.
3457 (assemble_vtyv_preinit_initializer): New function.
3458 (default_sectin_type_flags): Make sure .vtable_map_vars section has
3460 * output.h: Add function decl for assemble_vtv_preinit_initializer.
3461 * vtable-verify.c: New file.
3462 * vtable-verify.h: New file.
3463 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
3464 initialiation levels.
3465 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
3466 * passes.def: Insert pass_vtable_verify.
3467 * aclocal.m4: Reorder includes.
3468 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
3469 -fvtv-counts options.
3470 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
3471 as appropriate, if -fvtable-verify=... is used.
3472 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
3473 -fvtable-verify=... is used.
3474 * Makefile.in (OBJS): Add vtable-verify.o to list.
3475 (vtable-verify.o): Add new build rule.
3476 (GTFILES): Add vtable-verify.c to list.
3477 * common.opt (fvtable-verify=): New flag.
3478 (vtv_priority): Values for fvtable-verify= flag.
3479 (fvtv-counts): New flag.
3480 (fvtv-debug): New flag.
3481 * tree.h (save_vtable_map_decl): New extern function decl.
3483 2013-08-07 David Malcolm <dmalcolm@redhat.com>
3485 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
3486 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
3487 (pass_data_rl78_devirt): ...new pass_data instance and...
3488 (make_pass_rl78_devirt): ...new function.
3489 (rl78_asm_file_start): Port pass registration to new C++ API.
3491 2013-08-07 David Malcolm <dmalcolm@redhat.com>
3493 * coretypes.h (rtl_opt_pass): Add.
3494 (gcc::context): Add.
3495 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
3496 (epiphany_init): Port to new C++ pass API.
3497 (epiphany_optimize_mode_switching): Likewise.
3498 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
3499 (pass_manager::get_pass_mode_switching): New.
3500 (pass_manager::get_pass_peephole2): New.
3501 * mode-switching.c (pass_mode_switching): Add clone method.
3502 * recog.c (pass_peephole2): Add clone method.
3503 (pass_split_all_insns): Add clone method.
3505 2013-08-06 David Malcolm <dmalcolm@redhat.com>
3507 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
3508 (mips_option_override): ...here, porting to new C++ API for passes.
3510 2013-08-06 Jan Hubicka <jh@suse.cz>
3512 * cgraph.c (cgraph_get_body): New function based on lto.c
3514 * cgraph.h (cgraph_get_body): Declare.
3515 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
3517 * cgraphunit.c (expand_function): Get body prior expanding.
3518 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
3519 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
3521 * passes.c (do_per_function_toporder): Get body.
3522 * tree-inline.c (expand_call_inline): Get body prior inlining it.
3523 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
3525 2013-08-06 Martin Jambor <mjambor@suse.cz>
3528 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
3529 is not re-finalized. Rename second parameter to no_collect.
3531 2013-08-06 Martin Jambor <mjambor@suse.cz>
3534 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
3535 MEM_REF has proper alignment information.
3537 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
3540 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
3542 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
3543 rtx (*) (rtx, ...) with insn_gen_fn.
3544 * genoutput.c (output_insn_data): Cast gen_? function pointers to
3545 insn_gen_fn::stored_funcptr. Add initializer braces.
3547 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3549 Rewrite how instances of passes are cloned to remove assumptions
3550 about their sizes (thus allowing pass subclasses to have
3551 additional data fields, albeit non-GC-managed ones at this point).
3553 * passes.c (make_pass_instance): Now that passes have clone
3554 methods, rewrite this function to eliminate XNEW and memcpy
3555 calls that used hardcoded sizes. Since this function no longer
3556 creates pass instances, rename it to...
3557 (add_pass_instance): ...this. Document the old way that passes were
3558 numbered and flagged, and rework this function to continue using it.
3559 (next_pass_1): Add an initial_pass argument for use by
3561 (position_pass): When adding multiple instances of a pass, use
3562 the pass's clone method, rather than relying on the XNEW/memcpy
3563 within the former make_pass_instance (now add_pass_instance).
3564 (pass_manager::pass_manager): When invoking next_pass_1, also supply
3565 the initial instance of the current pass within the pass manager.
3567 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3569 This is the automated part of the conversion of passes from C
3570 structs to C++ classes.
3572 Patch autogenerated by refactor_passes.py from
3573 https://github.com/davidmalcolm/gcc-refactoring-scripts
3574 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
3576 * asan.c (pass_asan): Convert from a global struct to a subclass of
3577 gimple_opt_pass along with...
3578 (pass_data_asan): ...new pass_data instance and...
3579 (make_pass_asan): ...new function.
3580 (pass_asan_O0): Convert from a global struct to a subclass of
3581 gimple_opt_pass along with...
3582 (pass_data_asan_O0): ...new pass_data instance and...
3583 (make_pass_asan_O0): ...new function.
3584 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
3585 subclass of rtl_opt_pass along with...
3586 (pass_data_inc_dec): ...new pass_data instance and...
3587 (make_pass_inc_dec): ...new function.
3588 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
3589 a subclass of rtl_opt_pass along with...
3590 (pass_data_reorder_blocks): ...new pass_data instance and...
3591 (make_pass_reorder_blocks): ...new function.
3592 (pass_duplicate_computed_gotos): Convert from a global struct to a
3593 subclass of rtl_opt_pass along with...
3594 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
3595 (make_pass_duplicate_computed_gotos): ...new function.
3596 (pass_partition_blocks): Convert from a global struct to a subclass of
3597 rtl_opt_pass along with...
3598 (pass_data_partition_blocks): ...new pass_data instance and...
3599 (make_pass_partition_blocks): ...new function.
3600 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
3601 struct to a subclass of rtl_opt_pass along with...
3602 (pass_data_branch_target_load_optimize1): ...new pass_data instance
3604 (make_pass_branch_target_load_optimize1): ...new function.
3605 (pass_branch_target_load_optimize2): Convert from a global struct to a
3606 subclass of rtl_opt_pass along with...
3607 (pass_data_branch_target_load_optimize2): ...new pass_data instance
3609 (make_pass_branch_target_load_optimize2): ...new function.
3610 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
3611 of rtl_opt_pass along with...
3612 (pass_data_jump): ...new pass_data instance and...
3613 (make_pass_jump): ...new function.
3614 (pass_jump2): Convert from a global struct to a subclass of
3615 rtl_opt_pass along with...
3616 (pass_data_jump2): ...new pass_data instance and...
3617 (make_pass_jump2): ...new function.
3618 * cfgexpand.c (pass_expand): Convert from a global struct to a
3619 subclass of rtl_opt_pass along with...
3620 (pass_data_expand): ...new pass_data instance and...
3621 (make_pass_expand): ...new function.
3622 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
3623 of rtl_opt_pass along with...
3624 (pass_data_free_cfg): ...new pass_data instance and...
3625 (make_pass_free_cfg): ...new function.
3626 (pass_into_cfg_layout_mode): Convert from a global struct to a
3627 subclass of rtl_opt_pass along with...
3628 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
3629 (make_pass_into_cfg_layout_mode): ...new function.
3630 (pass_outof_cfg_layout_mode): Convert from a global struct to a
3631 subclass of rtl_opt_pass along with...
3632 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
3633 (make_pass_outof_cfg_layout_mode): ...new function.
3634 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
3635 struct to a subclass of gimple_opt_pass along with...
3636 (pass_data_build_cgraph_edges): ...new pass_data instance and...
3637 (make_pass_build_cgraph_edges): ...new function.
3638 (pass_rebuild_cgraph_edges): Convert from a global struct to a
3639 subclass of gimple_opt_pass along with...
3640 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
3641 (make_pass_rebuild_cgraph_edges): ...new function.
3642 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
3643 subclass of gimple_opt_pass along with...
3644 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
3646 (make_pass_remove_cgraph_callee_edges): ...new function.
3647 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
3648 struct to a subclass of rtl_opt_pass along with...
3649 (pass_data_stack_adjustments): ...new pass_data instance and...
3650 (make_pass_stack_adjustments): ...new function.
3651 * combine.c (pass_combine): Convert from a global struct to a subclass
3652 of rtl_opt_pass along with...
3653 (pass_data_combine): ...new pass_data instance and...
3654 (make_pass_combine): ...new function.
3655 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
3656 global struct to a subclass of rtl_opt_pass along with...
3657 (pass_data_compare_elim_after_reload): ...new pass_data instance
3659 (make_pass_compare_elim_after_reload): ...new function.
3660 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
3661 of rtl_opt_pass along with...
3662 (pass_data_rtl_cprop): ...new pass_data instance and...
3663 (make_pass_rtl_cprop): ...new function.
3664 * cse.c (pass_cse): Convert from a global struct to a subclass of
3665 rtl_opt_pass along with...
3666 (pass_data_cse): ...new pass_data instance and...
3667 (make_pass_cse): ...new function.
3668 (pass_cse2): Convert from a global struct to a subclass of
3669 rtl_opt_pass along with...
3670 (pass_data_cse2): ...new pass_data instance and...
3671 (make_pass_cse2): ...new function.
3672 (pass_cse_after_global_opts): Convert from a global struct to a
3673 subclass of rtl_opt_pass along with...
3674 (pass_data_cse_after_global_opts): ...new pass_data instance and...
3675 (make_pass_cse_after_global_opts): ...new function.
3676 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
3677 of rtl_opt_pass along with...
3678 (pass_data_ud_rtl_dce): ...new pass_data instance and...
3679 (make_pass_ud_rtl_dce): ...new function.
3680 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
3681 rtl_opt_pass along with...
3682 (pass_data_fast_rtl_dce): ...new pass_data instance and...
3683 (make_pass_fast_rtl_dce): ...new function.
3684 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
3685 a subclass of rtl_opt_pass along with...
3686 (pass_data_df_initialize_opt): ...new pass_data instance and...
3687 (make_pass_df_initialize_opt): ...new function.
3688 (pass_df_initialize_no_opt): Convert from a global struct to a
3689 subclass of rtl_opt_pass along with...
3690 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
3691 (make_pass_df_initialize_no_opt): ...new function.
3692 (pass_df_finish): Convert from a global struct to a subclass of
3693 rtl_opt_pass along with...
3694 (pass_data_df_finish): ...new pass_data instance and...
3695 (make_pass_df_finish): ...new function.
3696 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
3697 rtl_opt_pass along with...
3698 (pass_data_rtl_dse1): ...new pass_data instance and...
3699 (make_pass_rtl_dse1): ...new function.
3700 (pass_rtl_dse2): Convert from a global struct to a subclass of
3701 rtl_opt_pass along with...
3702 (pass_data_rtl_dse2): ...new pass_data instance and...
3703 (make_pass_rtl_dse2): ...new function.
3704 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
3705 subclass of rtl_opt_pass along with...
3706 (pass_data_dwarf2_frame): ...new pass_data instance and...
3707 (make_pass_dwarf2_frame): ...new function.
3708 * except.c (pass_set_nothrow_function_flags): Convert from a global
3709 struct to a subclass of rtl_opt_pass along with...
3710 (pass_data_set_nothrow_function_flags): ...new pass_data instance
3712 (make_pass_set_nothrow_function_flags): ...new function.
3713 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
3714 subclass of rtl_opt_pass along with...
3715 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
3717 (make_pass_convert_to_eh_region_ranges): ...new function.
3718 * final.c (pass_compute_alignments): Convert from a global struct to a
3719 subclass of rtl_opt_pass along with...
3720 (pass_data_compute_alignments): ...new pass_data instance and...
3721 (make_pass_compute_alignments): ...new function.
3722 (pass_final): Convert from a global struct to a subclass of
3723 rtl_opt_pass along with...
3724 (pass_data_final): ...new pass_data instance and...
3725 (make_pass_final): ...new function.
3726 (pass_shorten_branches): Convert from a global struct to a subclass of
3727 rtl_opt_pass along with...
3728 (pass_data_shorten_branches): ...new pass_data instance and...
3729 (make_pass_shorten_branches): ...new function.
3730 (pass_clean_state): Convert from a global struct to a subclass of
3731 rtl_opt_pass along with...
3732 (pass_data_clean_state): ...new pass_data instance and...
3733 (make_pass_clean_state): ...new function.
3734 * function.c (pass_instantiate_virtual_regs): Convert from a global
3735 struct to a subclass of rtl_opt_pass along with...
3736 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
3737 (make_pass_instantiate_virtual_regs): ...new function.
3738 (pass_leaf_regs): Convert from a global struct to a subclass of
3739 rtl_opt_pass along with...
3740 (pass_data_leaf_regs): ...new pass_data instance and...
3741 (make_pass_leaf_regs): ...new function.
3742 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
3743 subclass of rtl_opt_pass along with...
3744 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
3746 (make_pass_thread_prologue_and_epilogue): ...new function.
3747 (pass_match_asm_constraints): Convert from a global struct to a
3748 subclass of rtl_opt_pass along with...
3749 (pass_data_match_asm_constraints): ...new pass_data instance and...
3750 (make_pass_match_asm_constraints): ...new function.
3751 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
3752 subclass of rtl_opt_pass along with...
3753 (pass_data_rtl_fwprop): ...new pass_data instance and...
3754 (make_pass_rtl_fwprop): ...new function.
3755 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
3756 rtl_opt_pass along with...
3757 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
3758 (make_pass_rtl_fwprop_addr): ...new function.
3759 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
3760 rtl_opt_pass along with...
3761 (pass_data_rtl_pre): ...new pass_data instance and...
3762 (make_pass_rtl_pre): ...new function.
3763 (pass_rtl_hoist): Convert from a global struct to a subclass of
3764 rtl_opt_pass along with...
3765 (pass_data_rtl_hoist): ...new pass_data instance and...
3766 (make_pass_rtl_hoist): ...new function.
3767 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
3768 subclass of gimple_opt_pass along with...
3769 (pass_data_lower_cf): ...new pass_data instance and...
3770 (make_pass_lower_cf): ...new function.
3771 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
3772 from a global struct to a subclass of gimple_opt_pass along with...
3773 (pass_data_strength_reduction): ...new pass_data instance and...
3774 (make_pass_strength_reduction): ...new function.
3775 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
3776 of rtl_opt_pass along with...
3777 (pass_data_rtl_ifcvt): ...new pass_data instance and...
3778 (make_pass_rtl_ifcvt): ...new function.
3779 (pass_if_after_combine): Convert from a global struct to a subclass of
3780 rtl_opt_pass along with...
3781 (pass_data_if_after_combine): ...new pass_data instance and...
3782 (make_pass_if_after_combine): ...new function.
3783 (pass_if_after_reload): Convert from a global struct to a subclass of
3784 rtl_opt_pass along with...
3785 (pass_data_if_after_reload): ...new pass_data instance and...
3786 (make_pass_if_after_reload): ...new function.
3787 * init-regs.c (pass_initialize_regs): Convert from a global struct to
3788 a subclass of rtl_opt_pass along with...
3789 (pass_data_initialize_regs): ...new pass_data instance and...
3790 (make_pass_initialize_regs): ...new function.
3791 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
3792 of ipa_opt_pass_d along with...
3793 (pass_data_ipa_cp): ...new pass_data instance and...
3794 (make_pass_ipa_cp): ...new function.
3795 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
3796 global struct to a subclass of gimple_opt_pass along with...
3797 (pass_data_inline_parameters): ...new pass_data instance and...
3798 (make_pass_inline_parameters): ...new function.
3799 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
3800 subclass of gimple_opt_pass along with...
3801 (pass_data_early_inline): ...new pass_data instance and...
3802 (make_pass_early_inline): ...new function.
3803 (pass_ipa_inline): Convert from a global struct to a subclass of
3804 ipa_opt_pass_d along with...
3805 (pass_data_ipa_inline): ...new pass_data instance and...
3806 (make_pass_ipa_inline): ...new function.
3807 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
3808 struct to a subclass of gimple_opt_pass along with...
3809 (pass_data_local_pure_const): ...new pass_data instance and...
3810 (make_pass_local_pure_const): ...new function.
3811 (pass_ipa_pure_const): Convert from a global struct to a subclass of
3812 ipa_opt_pass_d along with...
3813 (pass_data_ipa_pure_const): ...new pass_data instance and...
3814 (make_pass_ipa_pure_const): ...new function.
3815 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
3816 to a subclass of ipa_opt_pass_d along with...
3817 (pass_data_ipa_reference): ...new pass_data instance and...
3818 (make_pass_ipa_reference): ...new function.
3819 * ipa-split.c (pass_split_functions): Convert from a global struct to
3820 a subclass of gimple_opt_pass along with...
3821 (pass_data_split_functions): ...new pass_data instance and...
3822 (make_pass_split_functions): ...new function.
3823 (pass_feedback_split_functions): Convert from a global struct to a
3824 subclass of gimple_opt_pass along with...
3825 (pass_data_feedback_split_functions): ...new pass_data instance and...
3826 (make_pass_feedback_split_functions): ...new function.
3827 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
3828 global struct to a subclass of simple_ipa_opt_pass along with...
3829 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
3831 (make_pass_ipa_function_and_variable_visibility): ...new function.
3832 (pass_ipa_free_inline_summary): Convert from a global struct to a
3833 subclass of simple_ipa_opt_pass along with...
3834 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
3835 (make_pass_ipa_free_inline_summary): ...new function.
3836 (pass_ipa_whole_program_visibility): Convert from a global struct to a
3837 subclass of ipa_opt_pass_d along with...
3838 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
3840 (make_pass_ipa_whole_program_visibility): ...new function.
3841 (pass_ipa_profile): Convert from a global struct to a subclass of
3842 ipa_opt_pass_d along with...
3843 (pass_data_ipa_profile): ...new pass_data instance and...
3844 (make_pass_ipa_profile): ...new function.
3845 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
3846 ipa_opt_pass_d along with...
3847 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
3848 (make_pass_ipa_cdtor_merge): ...new function.
3849 * ira.c (pass_ira): Convert from a global struct to a subclass of
3850 rtl_opt_pass along with...
3851 (pass_data_ira): ...new pass_data instance and...
3852 (make_pass_ira): ...new function.
3853 (pass_reload): Convert from a global struct to a subclass of
3854 rtl_opt_pass along with...
3855 (pass_data_reload): ...new pass_data instance and...
3856 (make_pass_reload): ...new function.
3857 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
3858 subclass of rtl_opt_pass along with...
3859 (pass_data_cleanup_barriers): ...new pass_data instance and...
3860 (make_pass_cleanup_barriers): ...new function.
3861 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
3862 of rtl_opt_pass along with...
3863 (pass_data_loop2): ...new pass_data instance and...
3864 (make_pass_loop2): ...new function.
3865 (pass_rtl_loop_init): Convert from a global struct to a subclass of
3866 rtl_opt_pass along with...
3867 (pass_data_rtl_loop_init): ...new pass_data instance and...
3868 (make_pass_rtl_loop_init): ...new function.
3869 (pass_rtl_loop_done): Convert from a global struct to a subclass of
3870 rtl_opt_pass along with...
3871 (pass_data_rtl_loop_done): ...new pass_data instance and...
3872 (make_pass_rtl_loop_done): ...new function.
3873 (pass_rtl_move_loop_invariants): Convert from a global struct to a
3874 subclass of rtl_opt_pass along with...
3875 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
3876 (make_pass_rtl_move_loop_invariants): ...new function.
3877 (pass_rtl_unswitch): Convert from a global struct to a subclass of
3878 rtl_opt_pass along with...
3879 (pass_data_rtl_unswitch): ...new pass_data instance and...
3880 (make_pass_rtl_unswitch): ...new function.
3881 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
3882 subclass of rtl_opt_pass along with...
3883 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
3885 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3886 (pass_rtl_doloop): Convert from a global struct to a subclass of
3887 rtl_opt_pass along with...
3888 (pass_data_rtl_doloop): ...new pass_data instance and...
3889 (make_pass_rtl_doloop): ...new function.
3890 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
3891 a subclass of rtl_opt_pass along with...
3892 (pass_data_lower_subreg): ...new pass_data instance and...
3893 (make_pass_lower_subreg): ...new function.
3894 (pass_lower_subreg2): Convert from a global struct to a subclass of
3895 rtl_opt_pass along with...
3896 (pass_data_lower_subreg2): ...new pass_data instance and...
3897 (make_pass_lower_subreg2): ...new function.
3898 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
3899 struct to a subclass of ipa_opt_pass_d along with...
3900 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
3901 (make_pass_ipa_lto_gimple_out): ...new function.
3902 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
3903 of ipa_opt_pass_d along with...
3904 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
3905 (make_pass_ipa_lto_finish_out): ...new function.
3906 * mode-switching.c (pass_mode_switching): Convert from a global struct
3907 to a subclass of rtl_opt_pass along with...
3908 (pass_data_mode_switching): ...new pass_data instance and...
3909 (make_pass_mode_switching): ...new function.
3910 * modulo-sched.c (pass_sms): Convert from a global struct to a
3911 subclass of rtl_opt_pass along with...
3912 (pass_data_sms): ...new pass_data instance and...
3913 (make_pass_sms): ...new function.
3914 * omp-low.c (pass_expand_omp): Convert from a global struct to a
3915 subclass of gimple_opt_pass along with...
3916 (pass_data_expand_omp): ...new pass_data instance and...
3917 (make_pass_expand_omp): ...new function.
3918 (pass_lower_omp): Convert from a global struct to a subclass of
3919 gimple_opt_pass along with...
3920 (pass_data_lower_omp): ...new pass_data instance and...
3921 (make_pass_lower_omp): ...new function.
3922 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
3923 of gimple_opt_pass along with...
3924 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
3925 (make_pass_diagnose_omp_blocks): ...new function.
3926 * passes.c (pass_early_local_passes): Convert from a global struct to
3927 a subclass of simple_ipa_opt_pass along with...
3928 (pass_data_early_local_passes): ...new pass_data instance and...
3929 (make_pass_early_local_passes): ...new function.
3930 (pass_all_early_optimizations): Convert from a global struct to a
3931 subclass of gimple_opt_pass along with...
3932 (pass_data_all_early_optimizations): ...new pass_data instance and...
3933 (make_pass_all_early_optimizations): ...new function.
3934 (pass_all_optimizations): Convert from a global struct to a subclass
3935 of gimple_opt_pass along with...
3936 (pass_data_all_optimizations): ...new pass_data instance and...
3937 (make_pass_all_optimizations): ...new function.
3938 (pass_all_optimizations_g): Convert from a global struct to a subclass
3939 of gimple_opt_pass along with...
3940 (pass_data_all_optimizations_g): ...new pass_data instance and...
3941 (make_pass_all_optimizations_g): ...new function.
3942 (pass_rest_of_compilation): Convert from a global struct to a subclass
3943 of rtl_opt_pass along with...
3944 (pass_data_rest_of_compilation): ...new pass_data instance and...
3945 (make_pass_rest_of_compilation): ...new function.
3946 (pass_postreload): Convert from a global struct to a subclass of
3947 rtl_opt_pass along with...
3948 (pass_data_postreload): ...new pass_data instance and...
3949 (make_pass_postreload): ...new function.
3950 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
3951 subclass of rtl_opt_pass along with...
3952 (pass_data_gcse2): ...new pass_data instance and...
3953 (make_pass_gcse2): ...new function.
3954 * postreload.c (pass_postreload_cse): Convert from a global struct to
3955 a subclass of rtl_opt_pass along with...
3956 (pass_data_postreload_cse): ...new pass_data instance and...
3957 (make_pass_postreload_cse): ...new function.
3958 * predict.c (pass_profile): Convert from a global struct to a subclass
3959 of gimple_opt_pass along with...
3960 (pass_data_profile): ...new pass_data instance and...
3961 (make_pass_profile): ...new function.
3962 (pass_strip_predict_hints): Convert from a global struct to a subclass
3963 of gimple_opt_pass along with...
3964 (pass_data_strip_predict_hints): ...new pass_data instance and...
3965 (make_pass_strip_predict_hints): ...new function.
3966 * recog.c (pass_peephole2): Convert from a global struct to a subclass
3967 of rtl_opt_pass along with...
3968 (pass_data_peephole2): ...new pass_data instance and...
3969 (make_pass_peephole2): ...new function.
3970 (pass_split_all_insns): Convert from a global struct to a subclass of
3971 rtl_opt_pass along with...
3972 (pass_data_split_all_insns): ...new pass_data instance and...
3973 (make_pass_split_all_insns): ...new function.
3974 (pass_split_after_reload): Convert from a global struct to a subclass
3975 of rtl_opt_pass along with...
3976 (pass_data_split_after_reload): ...new pass_data instance and...
3977 (make_pass_split_after_reload): ...new function.
3978 (pass_split_before_regstack): Convert from a global struct to a
3979 subclass of rtl_opt_pass along with...
3980 (pass_data_split_before_regstack): ...new pass_data instance and...
3981 (make_pass_split_before_regstack): ...new function.
3982 (pass_split_before_sched2): Convert from a global struct to a subclass
3983 of rtl_opt_pass along with...
3984 (pass_data_split_before_sched2): ...new pass_data instance and...
3985 (make_pass_split_before_sched2): ...new function.
3986 (pass_split_for_shorten_branches): Convert from a global struct to a
3987 subclass of rtl_opt_pass along with...
3988 (pass_data_split_for_shorten_branches): ...new pass_data instance
3990 (make_pass_split_for_shorten_branches): ...new function.
3991 * ree.c (pass_ree): Convert from a global struct to a subclass of
3992 rtl_opt_pass along with...
3993 (pass_data_ree): ...new pass_data instance and...
3994 (make_pass_ree): ...new function.
3995 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
3996 subclass of rtl_opt_pass along with...
3997 (pass_data_stack_regs): ...new pass_data instance and...
3998 (make_pass_stack_regs): ...new function.
3999 (pass_stack_regs_run): Convert from a global struct to a subclass of
4000 rtl_opt_pass along with...
4001 (pass_data_stack_regs_run): ...new pass_data instance and...
4002 (make_pass_stack_regs_run): ...new function.
4003 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4004 subclass of rtl_opt_pass along with...
4005 (pass_data_cprop_hardreg): ...new pass_data instance and...
4006 (make_pass_cprop_hardreg): ...new function.
4007 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4008 subclass of rtl_opt_pass along with...
4009 (pass_data_reginfo_init): ...new pass_data instance and...
4010 (make_pass_reginfo_init): ...new function.
4011 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4012 of rtl_opt_pass along with...
4013 (pass_data_regmove): ...new pass_data instance and...
4014 (make_pass_regmove): ...new function.
4015 * regrename.c (pass_regrename): Convert from a global struct to a
4016 subclass of rtl_opt_pass along with...
4017 (pass_data_regrename): ...new pass_data instance and...
4018 (make_pass_regrename): ...new function.
4019 * reorg.c (pass_delay_slots): Convert from a global struct to a
4020 subclass of rtl_opt_pass along with...
4021 (pass_data_delay_slots): ...new pass_data instance and...
4022 (make_pass_delay_slots): ...new function.
4023 (pass_machine_reorg): Convert from a global struct to a subclass of
4024 rtl_opt_pass along with...
4025 (pass_data_machine_reorg): ...new pass_data instance and...
4026 (make_pass_machine_reorg): ...new function.
4027 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4028 of rtl_opt_pass along with...
4029 (pass_data_sched): ...new pass_data instance and...
4030 (make_pass_sched): ...new function.
4031 (pass_sched2): Convert from a global struct to a subclass of
4032 rtl_opt_pass along with...
4033 (pass_data_sched2): ...new pass_data instance and...
4034 (make_pass_sched2): ...new function.
4035 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4036 to a subclass of rtl_opt_pass along with...
4037 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4038 (make_pass_stack_ptr_mod): ...new function.
4039 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4040 to a subclass of rtl_opt_pass along with...
4041 (pass_data_rtl_store_motion): ...new pass_data instance and...
4042 (make_pass_rtl_store_motion): ...new function.
4043 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4044 of gimple_opt_pass along with...
4045 (pass_data_tracer): ...new pass_data instance and...
4046 (make_pass_tracer): ...new function.
4047 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4048 to a subclass of gimple_opt_pass along with...
4049 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4050 (make_pass_diagnose_tm_blocks): ...new function.
4051 (pass_lower_tm): Convert from a global struct to a subclass of
4052 gimple_opt_pass along with...
4053 (pass_data_lower_tm): ...new pass_data instance and...
4054 (make_pass_lower_tm): ...new function.
4055 (pass_tm_init): Convert from a global struct to a subclass of
4056 gimple_opt_pass along with...
4057 (pass_data_tm_init): ...new pass_data instance and...
4058 (make_pass_tm_init): ...new function.
4059 (pass_tm_mark): Convert from a global struct to a subclass of
4060 gimple_opt_pass along with...
4061 (pass_data_tm_mark): ...new pass_data instance and...
4062 (make_pass_tm_mark): ...new function.
4063 (pass_tm_edges): Convert from a global struct to a subclass of
4064 gimple_opt_pass along with...
4065 (pass_data_tm_edges): ...new pass_data instance and...
4066 (make_pass_tm_edges): ...new function.
4067 (pass_tm_memopt): Convert from a global struct to a subclass of
4068 gimple_opt_pass along with...
4069 (pass_data_tm_memopt): ...new pass_data instance and...
4070 (make_pass_tm_memopt): ...new function.
4071 (pass_ipa_tm): Convert from a global struct to a subclass of
4072 simple_ipa_opt_pass along with...
4073 (pass_data_ipa_tm): ...new pass_data instance and...
4074 (make_pass_ipa_tm): ...new function.
4075 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4076 subclass of gimple_opt_pass along with...
4077 (pass_data_call_cdce): ...new pass_data instance and...
4078 (make_pass_call_cdce): ...new function.
4079 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4080 subclass of gimple_opt_pass along with...
4081 (pass_data_build_cfg): ...new pass_data instance and...
4082 (make_pass_build_cfg): ...new function.
4083 (pass_split_crit_edges): Convert from a global struct to a subclass of
4084 gimple_opt_pass along with...
4085 (pass_data_split_crit_edges): ...new pass_data instance and...
4086 (make_pass_split_crit_edges): ...new function.
4087 (pass_warn_function_return): Convert from a global struct to a
4088 subclass of gimple_opt_pass along with...
4089 (pass_data_warn_function_return): ...new pass_data instance and...
4090 (make_pass_warn_function_return): ...new function.
4091 (pass_warn_function_noreturn): Convert from a global struct to a
4092 subclass of gimple_opt_pass along with...
4093 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4094 (make_pass_warn_function_noreturn): ...new function.
4095 (pass_warn_unused_result): Convert from a global struct to a subclass
4096 of gimple_opt_pass along with...
4097 (pass_data_warn_unused_result): ...new pass_data instance and...
4098 (make_pass_warn_unused_result): ...new function.
4099 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4100 a subclass of gimple_opt_pass along with...
4101 (pass_data_merge_phi): ...new pass_data instance and...
4102 (make_pass_merge_phi): ...new function.
4103 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4104 a subclass of gimple_opt_pass along with...
4105 (pass_data_lower_complex): ...new pass_data instance and...
4106 (make_pass_lower_complex): ...new function.
4107 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4108 gimple_opt_pass along with...
4109 (pass_data_lower_complex_O0): ...new pass_data instance and...
4110 (make_pass_lower_complex_O0): ...new function.
4111 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4112 subclass of gimple_opt_pass along with...
4113 (pass_data_lower_eh): ...new pass_data instance and...
4114 (make_pass_lower_eh): ...new function.
4115 (pass_refactor_eh): Convert from a global struct to a subclass of
4116 gimple_opt_pass along with...
4117 (pass_data_refactor_eh): ...new pass_data instance and...
4118 (make_pass_refactor_eh): ...new function.
4119 (pass_lower_resx): Convert from a global struct to a subclass of
4120 gimple_opt_pass along with...
4121 (pass_data_lower_resx): ...new pass_data instance and...
4122 (make_pass_lower_resx): ...new function.
4123 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4124 of gimple_opt_pass along with...
4125 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4126 (make_pass_lower_eh_dispatch): ...new function.
4127 (pass_cleanup_eh): Convert from a global struct to a subclass of
4128 gimple_opt_pass along with...
4129 (pass_data_cleanup_eh): ...new pass_data instance and...
4130 (make_pass_cleanup_eh): ...new function.
4131 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4132 to a subclass of simple_ipa_opt_pass along with...
4133 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4134 (make_pass_ipa_lower_emutls): ...new function.
4135 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4136 a subclass of gimple_opt_pass along with...
4137 (pass_data_if_conversion): ...new pass_data instance and...
4138 (make_pass_if_conversion): ...new function.
4139 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4140 subclass of gimple_opt_pass along with...
4141 (pass_data_build_ssa): ...new pass_data instance and...
4142 (make_pass_build_ssa): ...new function.
4143 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4144 global struct to a subclass of gimple_opt_pass along with...
4145 (pass_data_loop_distribution): ...new pass_data instance and...
4146 (make_pass_loop_distribution): ...new function.
4147 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4148 subclass of gimple_opt_pass along with...
4149 (pass_data_mudflap_1): ...new pass_data instance and...
4150 (make_pass_mudflap_1): ...new function.
4151 (pass_mudflap_2): Convert from a global struct to a subclass of
4152 gimple_opt_pass along with...
4153 (pass_data_mudflap_2): ...new pass_data instance and...
4154 (make_pass_mudflap_2): ...new function.
4155 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4156 subclass of gimple_opt_pass along with...
4157 (pass_data_mudflap_1): ...new pass_data instance and...
4158 (make_pass_mudflap_1): ...new function.
4159 (pass_mudflap_2): Convert from a global struct to a subclass of
4160 gimple_opt_pass along with...
4161 (pass_data_mudflap_2): ...new pass_data instance and...
4162 (make_pass_mudflap_2): ...new function.
4163 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4164 gimple_opt_pass along with...
4165 (pass_data_nrv): ...new pass_data instance and...
4166 (make_pass_nrv): ...new function.
4167 (pass_return_slot): Convert from a global struct to a subclass of
4168 gimple_opt_pass along with...
4169 (pass_data_return_slot): ...new pass_data instance and...
4170 (make_pass_return_slot): ...new function.
4171 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4172 to a subclass of gimple_opt_pass along with...
4173 (pass_data_object_sizes): ...new pass_data instance and...
4174 (make_pass_object_sizes): ...new function.
4175 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4176 global struct to a subclass of gimple_opt_pass along with...
4177 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4179 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4180 (pass_fixup_cfg): Convert from a global struct to a subclass of
4181 gimple_opt_pass along with...
4182 (pass_data_fixup_cfg): ...new pass_data instance and...
4183 (make_pass_fixup_cfg): ...new function.
4184 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4185 (make_pass_mudflap_1): ...new function.
4186 (pass_mudflap_2): Replace declaration with that of...
4187 (make_pass_mudflap_2): ...new function.
4188 (pass_asan): Replace declaration with that of...
4189 (make_pass_asan): ...new function.
4190 (pass_asan_O0): Replace declaration with that of...
4191 (make_pass_asan_O0): ...new function.
4192 (pass_tsan): Replace declaration with that of...
4193 (make_pass_tsan): ...new function.
4194 (pass_tsan_O0): Replace declaration with that of...
4195 (make_pass_tsan_O0): ...new function.
4196 (pass_lower_cf): Replace declaration with that of...
4197 (make_pass_lower_cf): ...new function.
4198 (pass_refactor_eh): Replace declaration with that of...
4199 (make_pass_refactor_eh): ...new function.
4200 (pass_lower_eh): Replace declaration with that of...
4201 (make_pass_lower_eh): ...new function.
4202 (pass_lower_eh_dispatch): Replace declaration with that of...
4203 (make_pass_lower_eh_dispatch): ...new function.
4204 (pass_lower_resx): Replace declaration with that of...
4205 (make_pass_lower_resx): ...new function.
4206 (pass_build_cfg): Replace declaration with that of...
4207 (make_pass_build_cfg): ...new function.
4208 (pass_early_tree_profile): Replace declaration with that of...
4209 (make_pass_early_tree_profile): ...new function.
4210 (pass_cleanup_eh): Replace declaration with that of...
4211 (make_pass_cleanup_eh): ...new function.
4212 (pass_sra): Replace declaration with that of...
4213 (make_pass_sra): ...new function.
4214 (pass_sra_early): Replace declaration with that of...
4215 (make_pass_sra_early): ...new function.
4216 (pass_early_ipa_sra): Replace declaration with that of...
4217 (make_pass_early_ipa_sra): ...new function.
4218 (pass_tail_recursion): Replace declaration with that of...
4219 (make_pass_tail_recursion): ...new function.
4220 (pass_tail_calls): Replace declaration with that of...
4221 (make_pass_tail_calls): ...new function.
4222 (pass_tree_loop): Replace declaration with that of...
4223 (make_pass_tree_loop): ...new function.
4224 (pass_tree_loop_init): Replace declaration with that of...
4225 (make_pass_tree_loop_init): ...new function.
4226 (pass_lim): Replace declaration with that of...
4227 (make_pass_lim): ...new function.
4228 (pass_tree_unswitch): Replace declaration with that of...
4229 (make_pass_tree_unswitch): ...new function.
4230 (pass_predcom): Replace declaration with that of...
4231 (make_pass_predcom): ...new function.
4232 (pass_iv_canon): Replace declaration with that of...
4233 (make_pass_iv_canon): ...new function.
4234 (pass_scev_cprop): Replace declaration with that of...
4235 (make_pass_scev_cprop): ...new function.
4236 (pass_empty_loop): Replace declaration with that of...
4237 (make_pass_empty_loop): ...new function.
4238 (pass_record_bounds): Replace declaration with that of...
4239 (make_pass_record_bounds): ...new function.
4240 (pass_graphite): Replace declaration with that of...
4241 (make_pass_graphite): ...new function.
4242 (pass_graphite_transforms): Replace declaration with that of...
4243 (make_pass_graphite_transforms): ...new function.
4244 (pass_if_conversion): Replace declaration with that of...
4245 (make_pass_if_conversion): ...new function.
4246 (pass_loop_distribution): Replace declaration with that of...
4247 (make_pass_loop_distribution): ...new function.
4248 (pass_vectorize): Replace declaration with that of...
4249 (make_pass_vectorize): ...new function.
4250 (pass_slp_vectorize): Replace declaration with that of...
4251 (make_pass_slp_vectorize): ...new function.
4252 (pass_complete_unroll): Replace declaration with that of...
4253 (make_pass_complete_unroll): ...new function.
4254 (pass_complete_unrolli): Replace declaration with that of...
4255 (make_pass_complete_unrolli): ...new function.
4256 (pass_parallelize_loops): Replace declaration with that of...
4257 (make_pass_parallelize_loops): ...new function.
4258 (pass_loop_prefetch): Replace declaration with that of...
4259 (make_pass_loop_prefetch): ...new function.
4260 (pass_iv_optimize): Replace declaration with that of...
4261 (make_pass_iv_optimize): ...new function.
4262 (pass_tree_loop_done): Replace declaration with that of...
4263 (make_pass_tree_loop_done): ...new function.
4264 (pass_ch): Replace declaration with that of...
4265 (make_pass_ch): ...new function.
4266 (pass_ccp): Replace declaration with that of...
4267 (make_pass_ccp): ...new function.
4268 (pass_phi_only_cprop): Replace declaration with that of...
4269 (make_pass_phi_only_cprop): ...new function.
4270 (pass_build_ssa): Replace declaration with that of...
4271 (make_pass_build_ssa): ...new function.
4272 (pass_build_alias): Replace declaration with that of...
4273 (make_pass_build_alias): ...new function.
4274 (pass_build_ealias): Replace declaration with that of...
4275 (make_pass_build_ealias): ...new function.
4276 (pass_dominator): Replace declaration with that of...
4277 (make_pass_dominator): ...new function.
4278 (pass_dce): Replace declaration with that of...
4279 (make_pass_dce): ...new function.
4280 (pass_dce_loop): Replace declaration with that of...
4281 (make_pass_dce_loop): ...new function.
4282 (pass_cd_dce): Replace declaration with that of...
4283 (make_pass_cd_dce): ...new function.
4284 (pass_call_cdce): Replace declaration with that of...
4285 (make_pass_call_cdce): ...new function.
4286 (pass_merge_phi): Replace declaration with that of...
4287 (make_pass_merge_phi): ...new function.
4288 (pass_split_crit_edges): Replace declaration with that of...
4289 (make_pass_split_crit_edges): ...new function.
4290 (pass_pre): Replace declaration with that of...
4291 (make_pass_pre): ...new function.
4292 (pass_profile): Replace declaration with that of...
4293 (make_pass_profile): ...new function.
4294 (pass_strip_predict_hints): Replace declaration with that of...
4295 (make_pass_strip_predict_hints): ...new function.
4296 (pass_lower_complex_O0): Replace declaration with that of...
4297 (make_pass_lower_complex_O0): ...new function.
4298 (pass_lower_complex): Replace declaration with that of...
4299 (make_pass_lower_complex): ...new function.
4300 (pass_lower_vector): Replace declaration with that of...
4301 (make_pass_lower_vector): ...new function.
4302 (pass_lower_vector_ssa): Replace declaration with that of...
4303 (make_pass_lower_vector_ssa): ...new function.
4304 (pass_lower_omp): Replace declaration with that of...
4305 (make_pass_lower_omp): ...new function.
4306 (pass_diagnose_omp_blocks): Replace declaration with that of...
4307 (make_pass_diagnose_omp_blocks): ...new function.
4308 (pass_expand_omp): Replace declaration with that of...
4309 (make_pass_expand_omp): ...new function.
4310 (pass_expand_omp_ssa): Replace declaration with that of...
4311 (make_pass_expand_omp_ssa): ...new function.
4312 (pass_object_sizes): Replace declaration with that of...
4313 (make_pass_object_sizes): ...new function.
4314 (pass_strlen): Replace declaration with that of...
4315 (make_pass_strlen): ...new function.
4316 (pass_fold_builtins): Replace declaration with that of...
4317 (make_pass_fold_builtins): ...new function.
4318 (pass_stdarg): Replace declaration with that of...
4319 (make_pass_stdarg): ...new function.
4320 (pass_early_warn_uninitialized): Replace declaration with that of...
4321 (make_pass_early_warn_uninitialized): ...new function.
4322 (pass_late_warn_uninitialized): Replace declaration with that of...
4323 (make_pass_late_warn_uninitialized): ...new function.
4324 (pass_cse_reciprocals): Replace declaration with that of...
4325 (make_pass_cse_reciprocals): ...new function.
4326 (pass_cse_sincos): Replace declaration with that of...
4327 (make_pass_cse_sincos): ...new function.
4328 (pass_optimize_bswap): Replace declaration with that of...
4329 (make_pass_optimize_bswap): ...new function.
4330 (pass_optimize_widening_mul): Replace declaration with that of...
4331 (make_pass_optimize_widening_mul): ...new function.
4332 (pass_warn_function_return): Replace declaration with that of...
4333 (make_pass_warn_function_return): ...new function.
4334 (pass_warn_function_noreturn): Replace declaration with that of...
4335 (make_pass_warn_function_noreturn): ...new function.
4336 (pass_cselim): Replace declaration with that of...
4337 (make_pass_cselim): ...new function.
4338 (pass_phiopt): Replace declaration with that of...
4339 (make_pass_phiopt): ...new function.
4340 (pass_forwprop): Replace declaration with that of...
4341 (make_pass_forwprop): ...new function.
4342 (pass_phiprop): Replace declaration with that of...
4343 (make_pass_phiprop): ...new function.
4344 (pass_tree_ifcombine): Replace declaration with that of...
4345 (make_pass_tree_ifcombine): ...new function.
4346 (pass_dse): Replace declaration with that of...
4347 (make_pass_dse): ...new function.
4348 (pass_nrv): Replace declaration with that of...
4349 (make_pass_nrv): ...new function.
4350 (pass_rename_ssa_copies): Replace declaration with that of...
4351 (make_pass_rename_ssa_copies): ...new function.
4352 (pass_sink_code): Replace declaration with that of...
4353 (make_pass_sink_code): ...new function.
4354 (pass_fre): Replace declaration with that of...
4355 (make_pass_fre): ...new function.
4356 (pass_check_data_deps): Replace declaration with that of...
4357 (make_pass_check_data_deps): ...new function.
4358 (pass_copy_prop): Replace declaration with that of...
4359 (make_pass_copy_prop): ...new function.
4360 (pass_vrp): Replace declaration with that of...
4361 (make_pass_vrp): ...new function.
4362 (pass_uncprop): Replace declaration with that of...
4363 (make_pass_uncprop): ...new function.
4364 (pass_return_slot): Replace declaration with that of...
4365 (make_pass_return_slot): ...new function.
4366 (pass_reassoc): Replace declaration with that of...
4367 (make_pass_reassoc): ...new function.
4368 (pass_rebuild_cgraph_edges): Replace declaration with that of...
4369 (make_pass_rebuild_cgraph_edges): ...new function.
4370 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
4371 (make_pass_remove_cgraph_callee_edges): ...new function.
4372 (pass_build_cgraph_edges): Replace declaration with that of...
4373 (make_pass_build_cgraph_edges): ...new function.
4374 (pass_local_pure_const): Replace declaration with that of...
4375 (make_pass_local_pure_const): ...new function.
4376 (pass_tracer): Replace declaration with that of...
4377 (make_pass_tracer): ...new function.
4378 (pass_warn_unused_result): Replace declaration with that of...
4379 (make_pass_warn_unused_result): ...new function.
4380 (pass_diagnose_tm_blocks): Replace declaration with that of...
4381 (make_pass_diagnose_tm_blocks): ...new function.
4382 (pass_lower_tm): Replace declaration with that of...
4383 (make_pass_lower_tm): ...new function.
4384 (pass_tm_init): Replace declaration with that of...
4385 (make_pass_tm_init): ...new function.
4386 (pass_tm_mark): Replace declaration with that of...
4387 (make_pass_tm_mark): ...new function.
4388 (pass_tm_memopt): Replace declaration with that of...
4389 (make_pass_tm_memopt): ...new function.
4390 (pass_tm_edges): Replace declaration with that of...
4391 (make_pass_tm_edges): ...new function.
4392 (pass_split_functions): Replace declaration with that of...
4393 (make_pass_split_functions): ...new function.
4394 (pass_feedback_split_functions): Replace declaration with that of...
4395 (make_pass_feedback_split_functions): ...new function.
4396 (pass_strength_reduction): Replace declaration with that of...
4397 (make_pass_strength_reduction): ...new function.
4398 (pass_ipa_lower_emutls): Replace declaration with that of...
4399 (make_pass_ipa_lower_emutls): ...new function.
4400 (pass_ipa_function_and_variable_visibility): Replace declaration with
4402 (make_pass_ipa_function_and_variable_visibility): ...new function.
4403 (pass_ipa_tree_profile): Replace declaration with that of...
4404 (make_pass_ipa_tree_profile): ...new function.
4405 (pass_early_local_passes): Replace declaration with that of...
4406 (make_pass_early_local_passes): ...new function.
4407 (pass_ipa_whole_program_visibility): Replace declaration with that
4409 (make_pass_ipa_whole_program_visibility): ...new function.
4410 (pass_ipa_lto_gimple_out): Replace declaration with that of...
4411 (make_pass_ipa_lto_gimple_out): ...new function.
4412 (pass_ipa_increase_alignment): Replace declaration with that of...
4413 (make_pass_ipa_increase_alignment): ...new function.
4414 (pass_ipa_inline): Replace declaration with that of...
4415 (make_pass_ipa_inline): ...new function.
4416 (pass_ipa_free_lang_data): Replace declaration with that of...
4417 (make_pass_ipa_free_lang_data): ...new function.
4418 (pass_ipa_free_inline_summary): Replace declaration with that of...
4419 (make_pass_ipa_free_inline_summary): ...new function.
4420 (pass_ipa_cp): Replace declaration with that of...
4421 (make_pass_ipa_cp): ...new function.
4422 (pass_ipa_reference): Replace declaration with that of...
4423 (make_pass_ipa_reference): ...new function.
4424 (pass_ipa_pure_const): Replace declaration with that of...
4425 (make_pass_ipa_pure_const): ...new function.
4426 (pass_ipa_pta): Replace declaration with that of...
4427 (make_pass_ipa_pta): ...new function.
4428 (pass_ipa_lto_finish_out): Replace declaration with that of...
4429 (make_pass_ipa_lto_finish_out): ...new function.
4430 (pass_ipa_tm): Replace declaration with that of...
4431 (make_pass_ipa_tm): ...new function.
4432 (pass_ipa_profile): Replace declaration with that of...
4433 (make_pass_ipa_profile): ...new function.
4434 (pass_ipa_cdtor_merge): Replace declaration with that of...
4435 (make_pass_ipa_cdtor_merge): ...new function.
4436 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
4438 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4439 (pass_init_datastructures): Replace declaration with that of...
4440 (make_pass_init_datastructures): ...new function.
4441 (pass_fixup_cfg): Replace declaration with that of...
4442 (make_pass_fixup_cfg): ...new function.
4443 (pass_expand): Replace declaration with that of...
4444 (make_pass_expand): ...new function.
4445 (pass_instantiate_virtual_regs): Replace declaration with that of...
4446 (make_pass_instantiate_virtual_regs): ...new function.
4447 (pass_rtl_fwprop): Replace declaration with that of...
4448 (make_pass_rtl_fwprop): ...new function.
4449 (pass_rtl_fwprop_addr): Replace declaration with that of...
4450 (make_pass_rtl_fwprop_addr): ...new function.
4451 (pass_jump): Replace declaration with that of...
4452 (make_pass_jump): ...new function.
4453 (pass_jump2): Replace declaration with that of...
4454 (make_pass_jump2): ...new function.
4455 (pass_lower_subreg): Replace declaration with that of...
4456 (make_pass_lower_subreg): ...new function.
4457 (pass_cse): Replace declaration with that of...
4458 (make_pass_cse): ...new function.
4459 (pass_fast_rtl_dce): Replace declaration with that of...
4460 (make_pass_fast_rtl_dce): ...new function.
4461 (pass_ud_rtl_dce): Replace declaration with that of...
4462 (make_pass_ud_rtl_dce): ...new function.
4463 (pass_rtl_dce): Replace declaration with that of...
4464 (make_pass_rtl_dce): ...new function.
4465 (pass_rtl_dse1): Replace declaration with that of...
4466 (make_pass_rtl_dse1): ...new function.
4467 (pass_rtl_dse2): Replace declaration with that of...
4468 (make_pass_rtl_dse2): ...new function.
4469 (pass_rtl_dse3): Replace declaration with that of...
4470 (make_pass_rtl_dse3): ...new function.
4471 (pass_rtl_cprop): Replace declaration with that of...
4472 (make_pass_rtl_cprop): ...new function.
4473 (pass_rtl_pre): Replace declaration with that of...
4474 (make_pass_rtl_pre): ...new function.
4475 (pass_rtl_hoist): Replace declaration with that of...
4476 (make_pass_rtl_hoist): ...new function.
4477 (pass_rtl_store_motion): Replace declaration with that of...
4478 (make_pass_rtl_store_motion): ...new function.
4479 (pass_cse_after_global_opts): Replace declaration with that of...
4480 (make_pass_cse_after_global_opts): ...new function.
4481 (pass_rtl_ifcvt): Replace declaration with that of...
4482 (make_pass_rtl_ifcvt): ...new function.
4483 (pass_into_cfg_layout_mode): Replace declaration with that of...
4484 (make_pass_into_cfg_layout_mode): ...new function.
4485 (pass_outof_cfg_layout_mode): Replace declaration with that of...
4486 (make_pass_outof_cfg_layout_mode): ...new function.
4487 (pass_loop2): Replace declaration with that of...
4488 (make_pass_loop2): ...new function.
4489 (pass_rtl_loop_init): Replace declaration with that of...
4490 (make_pass_rtl_loop_init): ...new function.
4491 (pass_rtl_move_loop_invariants): Replace declaration with that of...
4492 (make_pass_rtl_move_loop_invariants): ...new function.
4493 (pass_rtl_unswitch): Replace declaration with that of...
4494 (make_pass_rtl_unswitch): ...new function.
4495 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
4496 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4497 (pass_rtl_doloop): Replace declaration with that of...
4498 (make_pass_rtl_doloop): ...new function.
4499 (pass_rtl_loop_done): Replace declaration with that of...
4500 (make_pass_rtl_loop_done): ...new function.
4501 (pass_web): Replace declaration with that of...
4502 (make_pass_web): ...new function.
4503 (pass_cse2): Replace declaration with that of...
4504 (make_pass_cse2): ...new function.
4505 (pass_df_initialize_opt): Replace declaration with that of...
4506 (make_pass_df_initialize_opt): ...new function.
4507 (pass_df_initialize_no_opt): Replace declaration with that of...
4508 (make_pass_df_initialize_no_opt): ...new function.
4509 (pass_reginfo_init): Replace declaration with that of...
4510 (make_pass_reginfo_init): ...new function.
4511 (pass_inc_dec): Replace declaration with that of...
4512 (make_pass_inc_dec): ...new function.
4513 (pass_stack_ptr_mod): Replace declaration with that of...
4514 (make_pass_stack_ptr_mod): ...new function.
4515 (pass_initialize_regs): Replace declaration with that of...
4516 (make_pass_initialize_regs): ...new function.
4517 (pass_combine): Replace declaration with that of...
4518 (make_pass_combine): ...new function.
4519 (pass_if_after_combine): Replace declaration with that of...
4520 (make_pass_if_after_combine): ...new function.
4521 (pass_ree): Replace declaration with that of...
4522 (make_pass_ree): ...new function.
4523 (pass_partition_blocks): Replace declaration with that of...
4524 (make_pass_partition_blocks): ...new function.
4525 (pass_match_asm_constraints): Replace declaration with that of...
4526 (make_pass_match_asm_constraints): ...new function.
4527 (pass_regmove): Replace declaration with that of...
4528 (make_pass_regmove): ...new function.
4529 (pass_split_all_insns): Replace declaration with that of...
4530 (make_pass_split_all_insns): ...new function.
4531 (pass_fast_rtl_byte_dce): Replace declaration with that of...
4532 (make_pass_fast_rtl_byte_dce): ...new function.
4533 (pass_lower_subreg2): Replace declaration with that of...
4534 (make_pass_lower_subreg2): ...new function.
4535 (pass_mode_switching): Replace declaration with that of...
4536 (make_pass_mode_switching): ...new function.
4537 (pass_sms): Replace declaration with that of...
4538 (make_pass_sms): ...new function.
4539 (pass_sched): Replace declaration with that of...
4540 (make_pass_sched): ...new function.
4541 (pass_ira): Replace declaration with that of...
4542 (make_pass_ira): ...new function.
4543 (pass_reload): Replace declaration with that of...
4544 (make_pass_reload): ...new function.
4545 (pass_clean_state): Replace declaration with that of...
4546 (make_pass_clean_state): ...new function.
4547 (pass_branch_prob): Replace declaration with that of...
4548 (make_pass_branch_prob): ...new function.
4549 (pass_value_profile_transformations): Replace declaration with that
4551 (make_pass_value_profile_transformations): ...new function.
4552 (pass_postreload_cse): Replace declaration with that of...
4553 (make_pass_postreload_cse): ...new function.
4554 (pass_gcse2): Replace declaration with that of...
4555 (make_pass_gcse2): ...new function.
4556 (pass_split_after_reload): Replace declaration with that of...
4557 (make_pass_split_after_reload): ...new function.
4558 (pass_branch_target_load_optimize1): Replace declaration with that
4560 (make_pass_branch_target_load_optimize1): ...new function.
4561 (pass_thread_prologue_and_epilogue): Replace declaration with that
4563 (make_pass_thread_prologue_and_epilogue): ...new function.
4564 (pass_stack_adjustments): Replace declaration with that of...
4565 (make_pass_stack_adjustments): ...new function.
4566 (pass_peephole2): Replace declaration with that of...
4567 (make_pass_peephole2): ...new function.
4568 (pass_if_after_reload): Replace declaration with that of...
4569 (make_pass_if_after_reload): ...new function.
4570 (pass_regrename): Replace declaration with that of...
4571 (make_pass_regrename): ...new function.
4572 (pass_cprop_hardreg): Replace declaration with that of...
4573 (make_pass_cprop_hardreg): ...new function.
4574 (pass_reorder_blocks): Replace declaration with that of...
4575 (make_pass_reorder_blocks): ...new function.
4576 (pass_branch_target_load_optimize2): Replace declaration with that
4578 (make_pass_branch_target_load_optimize2): ...new function.
4579 (pass_leaf_regs): Replace declaration with that of...
4580 (make_pass_leaf_regs): ...new function.
4581 (pass_split_before_sched2): Replace declaration with that of...
4582 (make_pass_split_before_sched2): ...new function.
4583 (pass_compare_elim_after_reload): Replace declaration with that of...
4584 (make_pass_compare_elim_after_reload): ...new function.
4585 (pass_sched2): Replace declaration with that of...
4586 (make_pass_sched2): ...new function.
4587 (pass_stack_regs): Replace declaration with that of...
4588 (make_pass_stack_regs): ...new function.
4589 (pass_stack_regs_run): Replace declaration with that of...
4590 (make_pass_stack_regs_run): ...new function.
4591 (pass_df_finish): Replace declaration with that of...
4592 (make_pass_df_finish): ...new function.
4593 (pass_compute_alignments): Replace declaration with that of...
4594 (make_pass_compute_alignments): ...new function.
4595 (pass_duplicate_computed_gotos): Replace declaration with that of...
4596 (make_pass_duplicate_computed_gotos): ...new function.
4597 (pass_variable_tracking): Replace declaration with that of...
4598 (make_pass_variable_tracking): ...new function.
4599 (pass_free_cfg): Replace declaration with that of...
4600 (make_pass_free_cfg): ...new function.
4601 (pass_machine_reorg): Replace declaration with that of...
4602 (make_pass_machine_reorg): ...new function.
4603 (pass_cleanup_barriers): Replace declaration with that of...
4604 (make_pass_cleanup_barriers): ...new function.
4605 (pass_delay_slots): Replace declaration with that of...
4606 (make_pass_delay_slots): ...new function.
4607 (pass_split_for_shorten_branches): Replace declaration with that of...
4608 (make_pass_split_for_shorten_branches): ...new function.
4609 (pass_split_before_regstack): Replace declaration with that of...
4610 (make_pass_split_before_regstack): ...new function.
4611 (pass_convert_to_eh_region_ranges): Replace declaration with that
4613 (make_pass_convert_to_eh_region_ranges): ...new function.
4614 (pass_shorten_branches): Replace declaration with that of...
4615 (make_pass_shorten_branches): ...new function.
4616 (pass_set_nothrow_function_flags): Replace declaration with that of...
4617 (make_pass_set_nothrow_function_flags): ...new function.
4618 (pass_dwarf2_frame): Replace declaration with that of...
4619 (make_pass_dwarf2_frame): ...new function.
4620 (pass_final): Replace declaration with that of...
4621 (make_pass_final): ...new function.
4622 (pass_rtl_seqabstr): Replace declaration with that of...
4623 (make_pass_rtl_seqabstr): ...new function.
4624 (pass_release_ssa_names): Replace declaration with that of...
4625 (make_pass_release_ssa_names): ...new function.
4626 (pass_early_inline): Replace declaration with that of...
4627 (make_pass_early_inline): ...new function.
4628 (pass_inline_parameters): Replace declaration with that of...
4629 (make_pass_inline_parameters): ...new function.
4630 (pass_update_address_taken): Replace declaration with that of...
4631 (make_pass_update_address_taken): ...new function.
4632 (pass_convert_switch): Replace declaration with that of...
4633 (make_pass_convert_switch): ...new function.
4634 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
4635 to a subclass of simple_ipa_opt_pass along with...
4636 (pass_data_ipa_tree_profile): ...new pass_data instance and...
4637 (make_pass_ipa_tree_profile): ...new function.
4638 * tree-sra.c (pass_sra_early): Convert from a global struct to a
4639 subclass of gimple_opt_pass along with...
4640 (pass_data_sra_early): ...new pass_data instance and...
4641 (make_pass_sra_early): ...new function.
4642 (pass_sra): Convert from a global struct to a subclass of
4643 gimple_opt_pass along with...
4644 (pass_data_sra): ...new pass_data instance and...
4645 (make_pass_sra): ...new function.
4646 (pass_early_ipa_sra): Convert from a global struct to a subclass of
4647 gimple_opt_pass along with...
4648 (pass_data_early_ipa_sra): ...new pass_data instance and...
4649 (make_pass_early_ipa_sra): ...new function.
4650 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
4651 subclass of gimple_opt_pass along with...
4652 (pass_data_ccp): ...new pass_data instance and...
4653 (make_pass_ccp): ...new function.
4654 (pass_fold_builtins): Convert from a global struct to a subclass of
4655 gimple_opt_pass along with...
4656 (pass_data_fold_builtins): ...new pass_data instance and...
4657 (make_pass_fold_builtins): ...new function.
4658 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
4659 subclass of gimple_opt_pass along with...
4660 (pass_data_copy_prop): ...new pass_data instance and...
4661 (make_pass_copy_prop): ...new function.
4662 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
4663 global struct to a subclass of gimple_opt_pass along with...
4664 (pass_data_rename_ssa_copies): ...new pass_data instance and...
4665 (make_pass_rename_ssa_copies): ...new function.
4666 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
4667 subclass of gimple_opt_pass along with...
4668 (pass_data_dce): ...new pass_data instance and...
4669 (make_pass_dce): ...new function.
4670 (pass_dce_loop): Convert from a global struct to a subclass of
4671 gimple_opt_pass along with...
4672 (pass_data_dce_loop): ...new pass_data instance and...
4673 (make_pass_dce_loop): ...new function.
4674 (pass_cd_dce): Convert from a global struct to a subclass of
4675 gimple_opt_pass along with...
4676 (pass_data_cd_dce): ...new pass_data instance and...
4677 (make_pass_cd_dce): ...new function.
4678 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
4679 subclass of gimple_opt_pass along with...
4680 (pass_data_dominator): ...new pass_data instance and...
4681 (make_pass_dominator): ...new function.
4682 (pass_phi_only_cprop): Convert from a global struct to a subclass of
4683 gimple_opt_pass along with...
4684 (pass_data_phi_only_cprop): ...new pass_data instance and...
4685 (make_pass_phi_only_cprop): ...new function.
4686 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
4687 subclass of gimple_opt_pass along with...
4688 (pass_data_dse): ...new pass_data instance and...
4689 (make_pass_dse): ...new function.
4690 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
4691 a subclass of gimple_opt_pass along with...
4692 (pass_data_forwprop): ...new pass_data instance and...
4693 (make_pass_forwprop): ...new function.
4694 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
4695 struct to a subclass of gimple_opt_pass along with...
4696 (pass_data_tree_ifcombine): ...new pass_data instance and...
4697 (make_pass_tree_ifcombine): ...new function.
4698 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
4699 subclass of gimple_opt_pass along with...
4700 (pass_data_ch): ...new pass_data instance and...
4701 (make_pass_ch): ...new function.
4702 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
4703 subclass of gimple_opt_pass along with...
4704 (pass_data_tree_loop): ...new pass_data instance and...
4705 (make_pass_tree_loop): ...new function.
4706 (pass_tree_loop_init): Convert from a global struct to a subclass of
4707 gimple_opt_pass along with...
4708 (pass_data_tree_loop_init): ...new pass_data instance and...
4709 (make_pass_tree_loop_init): ...new function.
4710 (pass_lim): Convert from a global struct to a subclass of
4711 gimple_opt_pass along with...
4712 (pass_data_lim): ...new pass_data instance and...
4713 (make_pass_lim): ...new function.
4714 (pass_tree_unswitch): Convert from a global struct to a subclass of
4715 gimple_opt_pass along with...
4716 (pass_data_tree_unswitch): ...new pass_data instance and...
4717 (make_pass_tree_unswitch): ...new function.
4718 (pass_predcom): Convert from a global struct to a subclass of
4719 gimple_opt_pass along with...
4720 (pass_data_predcom): ...new pass_data instance and...
4721 (make_pass_predcom): ...new function.
4722 (pass_vectorize): Convert from a global struct to a subclass of
4723 gimple_opt_pass along with...
4724 (pass_data_vectorize): ...new pass_data instance and...
4725 (make_pass_vectorize): ...new function.
4726 (pass_graphite): Convert from a global struct to a subclass of
4727 gimple_opt_pass along with...
4728 (pass_data_graphite): ...new pass_data instance and...
4729 (make_pass_graphite): ...new function.
4730 (pass_graphite_transforms): Convert from a global struct to a subclass
4731 of gimple_opt_pass along with...
4732 (pass_data_graphite_transforms): ...new pass_data instance and...
4733 (make_pass_graphite_transforms): ...new function.
4734 (pass_check_data_deps): Convert from a global struct to a subclass of
4735 gimple_opt_pass along with...
4736 (pass_data_check_data_deps): ...new pass_data instance and...
4737 (make_pass_check_data_deps): ...new function.
4738 (pass_iv_canon): Convert from a global struct to a subclass of
4739 gimple_opt_pass along with...
4740 (pass_data_iv_canon): ...new pass_data instance and...
4741 (make_pass_iv_canon): ...new function.
4742 (pass_scev_cprop): Convert from a global struct to a subclass of
4743 gimple_opt_pass along with...
4744 (pass_data_scev_cprop): ...new pass_data instance and...
4745 (make_pass_scev_cprop): ...new function.
4746 (pass_record_bounds): Convert from a global struct to a subclass of
4747 gimple_opt_pass along with...
4748 (pass_data_record_bounds): ...new pass_data instance and...
4749 (make_pass_record_bounds): ...new function.
4750 (pass_complete_unroll): Convert from a global struct to a subclass of
4751 gimple_opt_pass along with...
4752 (pass_data_complete_unroll): ...new pass_data instance and...
4753 (make_pass_complete_unroll): ...new function.
4754 (pass_complete_unrolli): Convert from a global struct to a subclass of
4755 gimple_opt_pass along with...
4756 (pass_data_complete_unrolli): ...new pass_data instance and...
4757 (make_pass_complete_unrolli): ...new function.
4758 (pass_parallelize_loops): Convert from a global struct to a subclass
4759 of gimple_opt_pass along with...
4760 (pass_data_parallelize_loops): ...new pass_data instance and...
4761 (make_pass_parallelize_loops): ...new function.
4762 (pass_loop_prefetch): Convert from a global struct to a subclass of
4763 gimple_opt_pass along with...
4764 (pass_data_loop_prefetch): ...new pass_data instance and...
4765 (make_pass_loop_prefetch): ...new function.
4766 (pass_iv_optimize): Convert from a global struct to a subclass of
4767 gimple_opt_pass along with...
4768 (pass_data_iv_optimize): ...new pass_data instance and...
4769 (make_pass_iv_optimize): ...new function.
4770 (pass_tree_loop_done): Convert from a global struct to a subclass of
4771 gimple_opt_pass along with...
4772 (pass_data_tree_loop_done): ...new pass_data instance and...
4773 (make_pass_tree_loop_done): ...new function.
4774 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
4775 struct to a subclass of gimple_opt_pass along with...
4776 (pass_data_cse_reciprocals): ...new pass_data instance and...
4777 (make_pass_cse_reciprocals): ...new function.
4778 (pass_cse_sincos): Convert from a global struct to a subclass of
4779 gimple_opt_pass along with...
4780 (pass_data_cse_sincos): ...new pass_data instance and...
4781 (make_pass_cse_sincos): ...new function.
4782 (pass_optimize_bswap): Convert from a global struct to a subclass of
4783 gimple_opt_pass along with...
4784 (pass_data_optimize_bswap): ...new pass_data instance and...
4785 (make_pass_optimize_bswap): ...new function.
4786 (pass_optimize_widening_mul): Convert from a global struct to a
4787 subclass of gimple_opt_pass along with...
4788 (pass_data_optimize_widening_mul): ...new pass_data instance and...
4789 (make_pass_optimize_widening_mul): ...new function.
4790 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
4791 subclass of gimple_opt_pass along with...
4792 (pass_data_phiopt): ...new pass_data instance and...
4793 (make_pass_phiopt): ...new function.
4794 (pass_cselim): Convert from a global struct to a subclass of
4795 gimple_opt_pass along with...
4796 (pass_data_cselim): ...new pass_data instance and...
4797 (make_pass_cselim): ...new function.
4798 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
4799 subclass of gimple_opt_pass along with...
4800 (pass_data_phiprop): ...new pass_data instance and...
4801 (make_pass_phiprop): ...new function.
4802 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
4803 subclass of gimple_opt_pass along with...
4804 (pass_data_pre): ...new pass_data instance and...
4805 (make_pass_pre): ...new function.
4806 (pass_fre): Convert from a global struct to a subclass of
4807 gimple_opt_pass along with...
4808 (pass_data_fre): ...new pass_data instance and...
4809 (make_pass_fre): ...new function.
4810 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
4811 subclass of gimple_opt_pass along with...
4812 (pass_data_reassoc): ...new pass_data instance and...
4813 (make_pass_reassoc): ...new function.
4814 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
4815 subclass of gimple_opt_pass along with...
4816 (pass_data_sink_code): ...new pass_data instance and...
4817 (make_pass_sink_code): ...new function.
4818 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
4819 subclass of gimple_opt_pass along with...
4820 (pass_data_strlen): ...new pass_data instance and...
4821 (make_pass_strlen): ...new function.
4822 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
4823 struct to a subclass of gimple_opt_pass along with...
4824 (pass_data_build_alias): ...new pass_data instance and...
4825 (make_pass_build_alias): ...new function.
4826 (pass_build_ealias): Convert from a global struct to a subclass of
4827 gimple_opt_pass along with...
4828 (pass_data_build_ealias): ...new pass_data instance and...
4829 (make_pass_build_ealias): ...new function.
4830 (pass_ipa_pta): Convert from a global struct to a subclass of
4831 simple_ipa_opt_pass along with...
4832 (pass_data_ipa_pta): ...new pass_data instance and...
4833 (make_pass_ipa_pta): ...new function.
4834 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
4835 subclass of gimple_opt_pass along with...
4836 (pass_data_uncprop): ...new pass_data instance and...
4837 (make_pass_uncprop): ...new function.
4838 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
4839 global struct to a subclass of gimple_opt_pass along with...
4840 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
4841 (make_pass_late_warn_uninitialized): ...new function.
4842 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
4843 to a subclass of gimple_opt_pass along with...
4844 (pass_data_init_datastructures): ...new pass_data instance and...
4845 (make_pass_init_datastructures): ...new function.
4846 (pass_early_warn_uninitialized): Convert from a global struct to a
4847 subclass of gimple_opt_pass along with...
4848 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
4849 (make_pass_early_warn_uninitialized): ...new function.
4850 (pass_update_address_taken): Convert from a global struct to a
4851 subclass of gimple_opt_pass along with...
4852 (pass_data_update_address_taken): ...new pass_data instance and...
4853 (make_pass_update_address_taken): ...new function.
4854 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
4855 struct to a subclass of gimple_opt_pass along with...
4856 (pass_data_release_ssa_names): ...new pass_data instance and...
4857 (make_pass_release_ssa_names): ...new function.
4858 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
4859 subclass of gimple_opt_pass along with...
4860 (pass_data_stdarg): ...new pass_data instance and...
4861 (make_pass_stdarg): ...new function.
4862 * tree-switch-conversion.c (pass_convert_switch): Convert from a
4863 global struct to a subclass of gimple_opt_pass along with...
4864 (pass_data_convert_switch): ...new pass_data instance and...
4865 (make_pass_convert_switch): ...new function.
4866 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
4867 to a subclass of gimple_opt_pass along with...
4868 (pass_data_tail_recursion): ...new pass_data instance and...
4869 (make_pass_tail_recursion): ...new function.
4870 (pass_tail_calls): Convert from a global struct to a subclass of
4871 gimple_opt_pass along with...
4872 (pass_data_tail_calls): ...new pass_data instance and...
4873 (make_pass_tail_calls): ...new function.
4874 * tree-vect-generic.c (pass_lower_vector): Convert from a global
4875 struct to a subclass of gimple_opt_pass along with...
4876 (pass_data_lower_vector): ...new pass_data instance and...
4877 (make_pass_lower_vector): ...new function.
4878 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
4879 gimple_opt_pass along with...
4880 (pass_data_lower_vector_ssa): ...new pass_data instance and...
4881 (make_pass_lower_vector_ssa): ...new function.
4882 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
4883 to a subclass of gimple_opt_pass along with...
4884 (pass_data_slp_vectorize): ...new pass_data instance and...
4885 (make_pass_slp_vectorize): ...new function.
4886 (pass_ipa_increase_alignment): Convert from a global struct to a
4887 subclass of simple_ipa_opt_pass along with...
4888 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
4889 (make_pass_ipa_increase_alignment): ...new function.
4890 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
4891 gimple_opt_pass along with...
4892 (pass_data_vrp): ...new pass_data instance and...
4893 (make_pass_vrp): ...new function.
4894 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
4895 subclass of simple_ipa_opt_pass along with...
4896 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
4897 (make_pass_ipa_free_lang_data): ...new function.
4898 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
4899 gimple_opt_pass along with...
4900 (pass_data_tsan): ...new pass_data instance and...
4901 (make_pass_tsan): ...new function.
4902 (pass_tsan_O0): Convert from a global struct to a subclass of
4903 gimple_opt_pass along with...
4904 (pass_data_tsan_O0): ...new pass_data instance and...
4905 (make_pass_tsan_O0): ...new function.
4906 * var-tracking.c (pass_variable_tracking): Convert from a global
4907 struct to a subclass of rtl_opt_pass along with...
4908 (pass_data_variable_tracking): ...new pass_data instance and...
4909 (make_pass_variable_tracking): ...new function.
4910 * web.c (pass_web): Convert from a global struct to a subclass of
4911 rtl_opt_pass along with...
4912 (pass_data_web): ...new pass_data instance and...
4913 (make_pass_web): ...new function.
4914 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
4915 declaration with that of...
4916 (make_pass_mode_switch_use): ...new function.
4917 (pass_resolve_sw_modes): Replace declaration with that of...
4918 (make_pass_resolve_sw_modes): ...new function.
4919 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
4920 from a global struct to a subclass of rtl_opt_pass along with...
4921 (pass_data_mode_switch_use): ...new pass_data instance and...
4922 (make_pass_mode_switch_use): ...new function.
4923 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
4924 from a global struct to a subclass of rtl_opt_pass along with...
4925 (pass_data_resolve_sw_modes): ...new pass_data instance and...
4926 (make_pass_resolve_sw_modes): ...new function.
4927 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
4928 struct to a subclass of rtl_opt_pass along with...
4929 (pass_data_insert_vzeroupper): ...new pass_data instance and...
4930 (make_pass_insert_vzeroupper): ...new function.
4931 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
4932 global struct to a subclass of rtl_opt_pass along with...
4933 (pass_data_work_around_errata): ...new pass_data instance and...
4934 (make_pass_work_around_errata): ...new function.
4935 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
4936 struct to a subclass of rtl_opt_pass along with...
4937 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
4938 (make_pass_mips_machine_reorg2): ...new function.
4940 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4942 * passes.c (pass_manager::operator new): New.
4944 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4946 Handwritten part of conversion of passes to C++ classes.
4948 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
4949 (toplev.o): Add dep on PASS_MANAGER_H.
4950 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
4951 of early local pases to reflect this moving from a global to a
4952 member of gcc::pass_manager.
4953 (cgraph_add_new_function): Likewise.
4954 * lto-cgraph.c (lto_output_node): Update for conversion of
4955 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
4956 * passes.c (opt_pass::clone): New.
4957 (opt_pass::gate): New.
4958 (opt_pass::execute): New.
4959 (opt_pass::opt_pass): New.
4960 (pass_manager::execute_early_local_passes): New.
4961 (pass_manager::execute_pass_mode_switching): new.
4962 (finish_optimization_passes): Convert to...
4963 (pass_manager::finish_optimization_passes): ...this.
4964 (finish_optimization_passes): Update for conversion of passes to
4966 (register_dump_files_1): Use has_gate since we cannot portably
4967 check a vtable entry against NULL.
4968 (dump_one_pass): Likewise.
4969 (ipa_write_summaries_2): Likewise.
4970 (ipa_write_optimization_summaries_1): Likewise.
4971 (ipa_read_summaries_1): Likewise.
4972 (ipa_read_optimization_summaries_1): Likewise.
4973 (execute_ipa_stmt_fixups): Likewise.
4974 (pass_manager::pass_manager): Rewrite pass-creation, invoking
4975 pass-creation functions rather than wiring up globals, and
4976 storing the results in fields of pass_manager generated using
4978 (pass_manager::dump_profile_report): Update for conversion of
4979 passes to C++ classes.
4980 (pass_manager::execute_ipa_summary_passes): Likewise.
4981 (execute_one_ipa_transform_pass): Likewise.
4982 (execute_one_pass): Use has_gate and has_execute since we cannot
4983 portably check a vtable entry against NULL.
4984 * pass_manager.h (pass_manager::finish_optimization_passes): New.
4985 (pass_manager): Use pass-instances.def to add fields for the
4986 various pass instances.
4987 * toplev.c (finalize): Update for move of
4988 finish_optimization_passes to a method of gcc::pass_manager.
4989 * toplev.h (finish_optimization_passes): Move to method of class
4991 * tree-pass.h (struct pass_data): New.
4992 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
4993 (opt_pass::gate): Convert to virtual function.
4994 (opt_pass::~opt_pass): New.
4995 (opt_pass::clone): New.
4996 (opt_pass::execute): Convert to virtual function.
4997 (opt_pass::opt_pass): New.
4998 (opt_pass::ctxt_): new.
4999 (gimple_opt_pass): Convert to subclass of opt_pass.
5000 (gimple_opt_pass::gimple_opt_pass): New.
5001 (rtl_opt_pass): Convert to subclass of opt_pass.
5002 (rtl_opt_pass::rtl_opt_pass): New.
5003 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5004 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5005 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5006 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5007 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5008 invocation of pass_mode_switching to reflect this moving from a
5009 global to a member of gcc::pass_manager.
5010 (ix86_option_override): Rework how pass_insert_vzeroupper is
5011 added to the pass_manager to reflect autogenerated changes.
5012 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
5014 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
5016 PR rtl-optimization/57708
5017 * recog.c (peep2_find_free_register): Validate all regs in a
5020 2013-08-05 Jan Hubicka <jh@suse.cz>
5023 * cgraph.c (verify_cgraph_node): Accept local flags from other
5025 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5026 (function_and_variable_visibility): Likewise.
5027 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5029 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5031 * graph.c (init_graph_slim_pretty_print): Remove.
5032 (print_graph_cfg): Do not call it. Use local pretty printer.
5033 (start_graph_dump): Likewise.
5035 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5037 * gimple-pretty-print.c (buffer): Remove.
5038 (initialized): Likewise.
5039 (maybe_init_pretty_print): Likewise.
5040 (print_gimple_stmt): Do not call it. Use non-static local
5041 pretty_printer variable.
5042 (print_gimple_expr): Likewise.
5043 (print_gimple_seq): Likewise.
5044 (gimple_dump_bb): Likewise.
5046 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5048 * asan.c (asan_pp): Remove.
5049 (asan_pp_initialized): Likewise.
5050 (asan_pp_initialize): Likewise.
5051 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5052 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5053 (asan_add_global): Likewise.
5055 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5057 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5058 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5059 * pretty-print.c (pp_formatted_text_data): Likewise.
5060 (pp_indent): Rename from pp_base_indent.
5061 (pp_format): Rename from pp_base_format.
5062 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5063 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5064 (pp_flush): Rename from pp_base_flush.
5065 (pp_set_line_maximum_length): Rename from
5066 pp_base_set_line_maximum_length.
5067 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5068 (pp_set_prefix): Rename from pp_base_set_prefix.
5069 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5070 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5071 (pp_append_text): Rename from pp_base_append_text.
5072 (pp_formatted_text): Rename from pp_base_formatted_text.
5073 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5074 (pp_remaining_character_count_for_line): Rename from
5075 pp_base_remaining_character_count_for_line.
5076 (pp_newline): Rename from pp_base_newline.
5077 (pp_character): Rename from pp_base_character.
5078 (pp_string): Rename from pp_base_string.
5079 (pp_maybe_space): Rename from pp_base_maybe_space.
5080 * asan.c (asan_pp_string): Adjust.
5081 (asan_emit_stack_protection): Likewise.
5082 (asan_add_global): Likewise.
5083 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5084 * tree-mudflap.c (mf_varname_tree): Likewise.
5085 * tree-pretty-print.c (pp_tree_identifier): Rename from
5086 pp_base_tree_identifier.
5087 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5088 Declare as function.
5090 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5092 * pretty-print.h (pp_bar_bar): New.
5093 (pp_ampersand_ampersand): Likewise.
5094 (pp_less_equal): Likewise.
5095 (pp_greater_equal): Likewise.
5096 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5097 printer functions instead of pp_string or operators and punctuators.
5098 (dump_gimple_call): Likewise.
5099 (dump_gimple_omp_for): Likewise.
5100 (dump_gimple_transaction): Likewise.
5101 (dump_gimple_phi): Likewise.
5102 (pp_gimple_stmt_1): Likewise.
5103 * sched-vis.c (print_insn): Likewise.
5104 * tree-mudflap.c (mf_varname_tree): Likewise.
5105 * tree-pretty-print.c (dump_block_node): Likewise.
5106 (dump_generic_node): Likewise.
5108 2013-08-02 Jan Hubicka <jh@suse.cz>
5110 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5112 * lto-streamer-out.c (tree_is_indexable): Results decls and
5113 parm decls are not indexable.
5114 (DFS_write_tree_body): Do not follow args and results.
5115 (hash_tree): Likewise.
5116 (output_functions): Rearrange so struct function is needed
5117 only when real body is output; be able to also ouptut abstract
5118 functions; output DECL_ARGUMENTS and DECL_RESULT.
5119 (lto_output): When not in WPA, ale store abstract functions.
5120 (write_symbol): Do not care about RESULT_DECL.
5121 (output_symbol_p): Handle correctly sbtract decls.
5122 * lto-streamer-in.c (input_function): Rearrange so struct
5123 function can be NULL at entry; allow streaming of
5124 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5125 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5126 sanity check during LTO.
5127 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5128 RESULT_DECl and DECL_ARGUMENTS.
5129 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5132 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5134 * pretty-print.h (pp_underscore): New.
5136 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5137 printer functions instead of pp_character.
5138 (dump_binary_rhs): Likewise.
5139 (dump_ternary_rhs): Likewise.
5140 (dump_gimple_call_args): Likewise.
5141 (pp_points_to_solution): Likewise.
5142 (dump_gimple_call): Likewise.
5143 (dump_gimple_switch): Likewise.
5144 (dump_gimple_cond): Likewise.
5145 (dump_gimple_bind): Likewise.
5146 (dump_gimple_try): Likewise.
5147 (dump_gimple_omp_for): Likewise.
5148 (dump_gimple_omp_continue): Likewise.
5149 (dump_gimple_omp_single): Likewise.
5150 (dump_gimple_omp_sections): Likewise.
5151 (dump_gimple_omp_block): Likewise.
5152 (dump_gimple_omp_critical): Likewise.
5153 (dump_gimple_transaction): Likewise.
5154 (dump_gimple_asm): Likewise.
5155 (dump_gimple_phi): Likewise.
5156 (dump_gimple_omp_parallel): Likewise.
5157 (dump_gimple_omp_task): Likewise.
5158 (dump_gimple_omp_atomic_load): Likewise.
5159 (dump_gimple_omp_atomic_store): Likewise.
5160 (dump_gimple_mem_ops): Likewise.
5161 (pp_gimple_stmt_1): Likewise.
5162 (pp_cfg_jump): Likewise.
5163 (dump_implicit_edges): Likewise.
5164 (gimple_dump_bb_for_graph): Likewise.
5165 * graph.c (draw_cfg_node): Likewise.
5166 * langhooks.c (lhd_print_error_function): Likewise.
5167 * sched-vis.c (print_exp): Likewise.
5168 (print_value): Likewise.
5169 (print_pattern): Likewise.
5170 (print_insn): Likewise.
5171 (rtl_dump_bb_for_graph): Likewise.
5172 * tree-pretty-print.c (dump_function_declaration): Likewise.
5173 (dump_array_domain): Likewise.
5174 (dump_omp_clause): Likewise.
5175 (dump_location): Likewise.
5176 (dump_generic_node): Likewise.
5177 (print_struct_decl): Likewise.
5178 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5180 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5182 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5183 candidate table when replacing a candidate statement.
5184 (replace_rhs_if_not_dup): Likewise.
5185 (replace_one_candidate): Likewise.
5187 2013-08-02 Jan Hubicka <jh@suse.cz>
5188 Martin Liska <marxin.liska@gmail.com>
5190 * cgraphunit.c (add_new_function): Fix logic when adding from
5192 (assemble_thunk): Rename to ...
5193 (expand_thunk); .. this one; export; get it working with
5194 general functions; make produced gimple valid.
5195 * cgraph.h (expand_thunk): Declare.
5197 2013-08-02 Jan Hubicka <jh@suse.cz>
5199 * ipa-cp.c (gather_context_independent_values): Use
5200 ipa_get_param_move_cost.
5201 (get_replacement_map): Remove PARAM; move parameter folding
5203 (create_specialized_node): Update.
5204 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5205 assert that we have gimple body; update move_cost.
5206 (count_formal_params): Assert that we have gimple body.
5207 (ipa_dump_param): New function.
5208 (ipa_alloc_node_params): Break out from ...
5209 (ipa_initialize_node_params): ... here.
5210 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5211 (ipa_write_node_info): Stream move costs.
5212 (ipa_read_node_info): Read move costs.
5213 (ipa_update_after_lto_read): Do not recompute node params.
5214 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5215 (ipa_get_param): Check we are not in WPA.
5216 (ipa_get_param_move_cost): New.
5217 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5218 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5219 parm numbers to be present.
5221 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5223 PR rtl-optimization/58048
5224 * lra-constraints.c (process_alt_operands): Don't check asm
5225 operand on register.
5227 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5229 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5230 the implied StoreLoad barrier for atomic operations if before.
5232 2013-08-02 Jan Hubicka <jh@suse.cz>
5233 Martin Liska <marxin.liska@gmail.com>
5235 * cgraph.c (cgraph_function_body_availability): Do not check
5237 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5238 symtab_node_availability): Declare.
5239 * ipa.c (can_replace_by_local_alias): New.
5240 (function_and_variable_visibility): Use it.
5241 * symtab.c (symtab_for_node_and_aliases,
5242 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
5244 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5246 PR rtl-optimization/57963
5247 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
5248 (lra_constraints): Use them.
5250 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5252 * config/arm/types.md (define_attr "type"): Add "load_acq"
5254 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5256 (cortex_a53_store1): Likewise.
5258 2013-08-01 Jan Hubicka <jh@suse.cz>
5260 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5261 partitions are not needed.
5263 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
5265 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5266 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5267 MAYBE_NON_Q_CLASS_P where appropriate.
5269 2013-08-01 Jan Hubicka <jh@suse.cz>
5271 * cgraph.h (release_function_body): Declare.
5272 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5273 of unused delcarations.
5275 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5277 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5278 RTL form when subtracting a constant.
5280 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5282 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5283 Generate canonical plus rtx with negated immediate instead of minus
5285 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5287 2013-08-01 Jan Hubicka <jh@suse.cz>
5289 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5290 (cgraph_release_function_body): Likewise.
5291 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5292 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5293 to used_as_abstract_origin.
5294 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5295 symbols used as abstract origins.
5296 * cgraphunit.c (analyze_functions): Update.
5297 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5298 used_as_abstract_origin.
5299 * tree-inline.c (tree_function_versioning): Update
5300 used_as_abstract_origin; be ready for DECL_RESULT and
5301 DECL_ARGUMENTS to be NULL.
5303 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5304 for abstract functions.
5305 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5308 2013-08-01 Jan Hubicka <jh@suse.cz>
5310 * profile.c (compute_value_histograms): Fix thinko.
5312 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5314 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5315 aarch-common-protos.h to extra_headers.
5316 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5317 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5318 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5320 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5322 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5323 Include "../arm/types.md". Define "type" attribute for all patterns.
5324 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5327 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5329 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5330 to support power8 load fusion.
5331 (fusion_gpr_mem_load): Likewise.
5333 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5335 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5336 declarations for power8 load fusion.
5337 (emit_fusion_gpr_load): Likewise.
5339 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5340 tuning for power8, turn on fusion mode by default. Turn on sign
5341 extending fusion mode if normal fusion mode is on, and we are at
5343 (fusion_gpr_load_p): New function, return true if we can fuse an
5344 addis instruction with a dependent load to a GPR.
5345 (emit_fusion_gpr_load): Emit the instructions for power8 load
5348 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
5349 (VSX load fusion peepholes): New peepholes to fuse together an
5350 addi instruction with a VSX load instruction.
5352 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5353 peepholes to fuse an addis instruction with a load to a GPR base
5354 register. If we are supporting sign extending fusions, convert
5355 sign extending loads to zero extending loads and add an explicit
5358 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5360 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5361 aarch-common-protos.h to extra_headers.
5362 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5363 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5364 (arm_early_store_addr_dep): Likewise.
5365 (arm_no_early_alu_shift_dep): Likewise.
5366 (arm_no_early_alu_shift_value_dep): Likewise.
5367 (arm_no_early_mul_dep): Likewise.
5368 (arm_no_early_store_addr_dep): Likewise.
5369 (arm_mac_accumulator_is_mul_result): Likewise.
5370 (arm_mac_accumulator_is_result): Likewise.
5371 * config/arm/aarch-common.c: ... here. New file.
5372 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
5374 (arm_early_store_addr_dep): Likewise.
5375 (arm_no_early_alu_shift_dep): Likewise.
5376 (arm_no_early_alu_shift_value_dep): Likewise.
5377 (arm_no_early_mul_dep): Likewise.
5378 (arm_no_early_store_addr_dep): Likewise.
5379 (arm_mac_accumulator_is_mul_result): Likewise.
5380 (arm_mac_accumulator_is_result): Likewise.
5381 * config/arm/aarch-common-protos.h: ... here. New file.
5382 * config/arm/t-arm (aarch-common.o): Define.
5384 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5386 * config/arm/arm.md: Include new file "types.md".
5387 (define_attr "type"): Move from here to ...
5388 (define_attr "mul32"): Likewise.
5389 (define_attr "mul64"): Likewise.
5390 * config/arm/types.md: ... here. New file.
5392 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
5394 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
5395 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
5397 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5399 * gen-pass-instances.awk: Fix offset of substr().
5401 2013-07-31 David Malcolm <dmalcolm@redhat.com>
5403 * Makefile.in (pass-instances.def): New.
5404 (passes.o): Replace dependency on passes.def with one on
5407 * gen-pass-instances.awk: New.
5409 * passes.c (pass_manager::pass_manager): Use pass-instances.def
5410 rather than passes.def, updating local definition of NEXT_PASS
5411 macro to add an extra NUM parameter (currently unused).
5413 2013-07-30 David Malcolm <dmalcolm@redhat.com>
5415 * Makefile.in (PASS_MANAGER_H): New.
5416 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
5417 (passes.o): Likewise.
5418 (statistics.o): Likewise.
5419 (cgraphunit.o): Likewise.
5420 (context.o): Depend on PASS_MANAGER_H.
5422 * pass_manager.h: New.
5424 * cgraphunit.c (cgraph_add_new_function): Update for moves
5425 of globals to fields of pass_manager.
5426 (analyze_function): Likewise.
5427 (expand_function): Likewise.
5428 (ipa_passes): Likewise.
5429 (compile): Likewise.
5431 * context.c (context::context): New.
5432 * context.h (context::context): New.
5433 (context::get_passes): New.
5434 (context::passes_): New.
5436 * lto-cgraph.c (input_node): Update for moves of globals to
5437 fields of pass_manager.
5439 * passes.c (all_passes): Remove, in favor of a field of the
5440 same name within the new class pass_manager.
5441 (all_small_ipa_passes): Likewise.
5442 (all_lowering_passes): Likewise.
5443 (all_regular_ipa_passes): Likewise.
5444 (all_late_ipa_passes): Likewise.
5445 (all_lto_gen_passes): Likewise.
5446 (passes_by_id): Likewise.
5447 (passes_by_id_size): Likewise.
5448 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5449 the new class pass_manager.
5450 (set_pass_for_id): Convert to...
5451 (pass_manager::set_pass_for_id): ...method.
5452 (get_pass_for_id): Convert to...
5453 (pass_manager::get_pass_for_id): ...method.
5454 (register_one_dump_file): Move body of implementation into...
5455 (pass_manager::register_one_dump_file): ...here.
5456 (register_dump_files_1): Convert to...
5457 (pass_manager::register_dump_files_1): ...method.
5458 (register_dump_files): Convert to...
5459 (pass_manager::register_dump_files): ...method.
5460 (create_pass_tab): Update for moves of globals to fields of
5462 (dump_passes): Move body of implementation into...
5463 (pass_manager::dump_passes): ...here.
5464 (register_pass): Move body of implementation into...
5465 (pass_manager::register_pass): ...here.
5466 (init_optimization_passes): Convert into...
5467 (pass_manager::pass_manager): ...constructor for new
5468 pass_manager class, and initialize the pass_lists array.
5469 (check_profile_consistency): Update for moves of globals to
5470 fields of pass_manager.
5471 (dump_profile_report): Move body of implementation into...
5472 (pass_manager::dump_profile_report): ...here.
5473 (ipa_write_summaries_1): Update for moves of pass lists from
5474 being globals to fields of pass_manager.
5475 (ipa_write_optimization_summaries): Likewise.
5476 (ipa_read_summaries): Likewise.
5477 (ipa_read_optimization_summaries): Likewise.
5478 (execute_all_ipa_stmt_fixups): Likewise.
5480 * statistics.c (statistics_fini): Update for moves of globals to
5481 fields of pass_manager.
5483 * toplev.c (general_init): Replace call to
5484 init_optimization_passes with construction of the pass_manager
5487 * tree-pass.h (all_passes): Remove, in favor of a field of the
5488 same name within the new class pass_manager.
5489 (all_small_ipa_passes): Likewise.
5490 (all_lowering_passes): Likewise.
5491 (all_regular_ipa_passes): Likewise.
5492 (all_lto_gen_passes): Likewise.
5493 (all_late_ipa_passes): Likewise.
5494 (passes_by_id): Likewise.
5495 (passes_by_id_size): Likewise.
5496 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5497 the new class pass_manager.
5498 (get_pass_for_id): Remove.
5500 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
5502 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
5505 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
5507 * arm.md (mulhi3): New expand pattern.
5509 2013-07-30 Jan Hubicka <jh@suse.cz>
5510 Martin Liska <marxin.liska@gmail.com>
5512 * profile.c (compute_value_histograms): Do not ICE when
5513 there is mismatch only on some counters.
5515 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5517 PR rtl-optimization/57637
5518 * function.c (move_insn_for_shrink_wrap): Also check the
5519 GEN set of the LIVE problem for the liveness analysis
5520 if it exists, otherwise give up.
5522 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5524 PR tree-optimization/57993
5525 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
5526 replaced statement in the candidate table.
5527 (phi_add_costs): Return infinite cost when the hidden basis does
5528 not dominate all phis on which the candidate is dependent.
5529 (replace_one_candidate): Record replaced statement in the
5532 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
5534 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
5535 (ashlv2si3): New expander.
5536 (*ashlv2si3_i): New define_insn_and_split.
5537 * predicates.md (float_operation): Allow patterns with three
5540 PR rtl-optimization/58021
5541 * mode-switching.c (create_pre_exit): Always split off preceding
5542 insns if we are not at the basic block head.
5544 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
5546 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
5547 (UCLIBC_DYNAMIC_LINKER): New macro.
5548 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
5550 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
5551 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
5553 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
5554 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
5555 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
5556 for SF and DF modes. Use ieee_quad_format for TF mode.
5557 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
5558 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
5559 (mips_option_override): Handle `-mnan=legacy'.
5560 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
5561 `-mabs=2008' and `-mnan=2008'.
5562 (OPTION_DEFAULT_SPECS): Add "nan" default.
5563 (ASM_SPEC): Handle `-mnan='.
5564 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
5565 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
5566 comment accordingly.
5567 (neg<mode>2): Likewise.
5568 * config/mips/mips.opt (mabs, mnan): New options.
5569 * doc/install.texi (Configuration): Document `--with-nan=' option.
5570 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
5572 (MIPS Options): Document them.
5573 * config.gcc <mips*-*-*>: Handle `--with-nan='.
5574 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
5575 * configure: Regenerate.
5576 * config.in: Regenerate.
5578 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
5580 * config/i386/i386.md (float post-reload splitters): Do not check
5581 for subregs of SSE registers.
5583 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
5584 H.J. Lu <hongjiu.lu@intel.com>
5588 * config/i386/i386.md (post-reload splitter
5589 to avoid partial SSE reg dependency stalls): New pattern.
5591 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
5593 * config/s390/s390.md ("movcc"): Swap load and store instructions.
5595 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
5597 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
5598 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
5600 2013-07-26 Cary Coutant <ccoutant@google.com>
5602 * dwarf2out.c (die_checksum_ordered): Don't include template
5603 instantiations in signature.
5604 (is_template_parameter): New function.
5605 (is_template_instantiation): New function.
5606 (generate_skeleton_bottom_up): Don't include template instantiations
5608 (generate_skeleton): Likewise.
5609 (break_out_comdat_types): Move recursive call to break out nested
5611 (prune_unused_types_mark_generic_parms_dies): Call
5612 is_template_parameter.
5614 2013-07-26 Ian Bolton <ian.bolton@arm.com>
5616 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
5617 uses vector registers.
5618 * config/aarch64/iterators.md: Add attributes rtn and vas.
5620 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5621 Richard Earnshaw <richard.earnshaw@arm.com>
5623 * combine.c (simplify_comparison): Re-canonicalize operands
5625 * config/arm/arm.md (movcond_addsi): New splitter.
5627 2013-07-25 Sterling Augustine <saugustine@google.com>
5629 * dwarf2out.c (size_of_pubnames): Move code to...
5630 (include_pubname_in_output): ...here. New.
5631 (want_pubnames): Rearrange.
5632 (output_pubnames): Call include_pubname_in_output. Move assertion.
5634 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5636 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
5638 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5641 * doc/extend.texi: Remove obsolete builtins. Fix
5642 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
5644 2013-07-25 Jan Hubicka <jh@suse.cz>
5646 * cgraph.c (release_function_body): Break out from ...
5647 (cgraph_release_function_body): ... this one; also release DECL_RESULT
5649 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
5650 old_tree in the map.
5651 (create_specialized_node): Update.
5652 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
5654 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
5655 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
5656 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
5657 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
5660 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5662 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
5663 addsi3_carryin_alt2_<optab>): Correct output template.
5665 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5667 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
5668 Adjust for arm_restrict_it.
5669 Remove trailing whitespace.
5671 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
5673 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
5674 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
5676 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
5678 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
5680 PR rtl-optimization/57960
5681 * lra-constraints.c (process_alt_operands): Use the right mode
5682 when checking strict_low.
5684 2013-07-25 Jan Hubicka <jh@suse.cz>
5686 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
5687 * cgraph.c (cgraph_remove_node): Do not release function body
5688 when in cgraph streaming.
5689 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
5690 in other partitions are not considered reachable; fix handling of
5693 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5695 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
5697 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5702 * config/arm/arm.md ("*sibcall_insn): Replace use of
5703 Ss with US. Adjust output for v5 and v4t.
5704 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
5706 * config/arm/constraints.md ("Ss"): Rename to US.
5708 2013-07-25 Terry Guo <terry.guo@arm.com>
5710 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
5711 shift_add/shift_sub0/shift_sub1 RTXs.
5713 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
5714 Anton Blanchard <anton@au1.ibm.com>
5716 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
5717 (altivec_vpks<VI_char>ss): Likewise.
5718 (altivec_vpks<VI_char>us): Likewise.
5719 (altivec_vpku<VI_char>us): Likewise.
5720 (altivec_vpku<VI_char>um): Likewise.
5722 2013-07-24 David Malcolm <dmalcolm@redhat.com>
5724 Introduce context class.
5726 * Makefile.in (CONTEXT_H): New.
5727 (OBJS): Add context.o.
5728 (toplev.o): Add CONTEXT_H to dependencies.
5731 * toplev.c (general_init): Create the singleton gcc::context instance.
5737 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
5739 PR rtl-optimization/57968
5740 * mode-switching.c (create_pre_exit): Allow instructions that
5741 don't set a return register to need a non-exit mode.
5743 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5744 Anton Blanchard <anton@au1.ibm.com>
5746 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
5747 operands to vperm for little endian.
5748 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
5749 of lvsl to create the control mask for a vperm for little endian.
5751 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5752 Anton Blanchard <anton@au1.ibm.com>
5754 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5755 two operands for little-endian.
5757 2013-07-23 Steve Ellcey <sellcey@mips.com>
5759 * config/mips/mips.c (mips_case_values_threshold): New.
5760 (TARGET_CASE_VALUES_THRESHOLD): Define.
5762 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5763 Anton Blanchard <anton@au1.ibm.com>
5765 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
5766 selection of field for vector splat in little endian mode.
5768 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5770 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
5771 expanders to rs6000.md.
5772 (ior<mode>3): Likewise.
5773 (and<mode>3): Likewise.
5774 (one_cmpl<mode>2): Likewise.
5775 (nor<mode>3): Likewise.
5776 (andc<mode>3): Likewise.
5777 (eqv<mode>3): Likewise.
5778 (nand<mode>3): Likewise.
5779 (orc<mode>3): Likewise.
5781 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
5784 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
5785 to split multi-word logical operations.
5786 (rs6000_split_logical_di): Likewise.
5787 (rs6000_split_logical): Likewise.
5789 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
5790 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
5791 and allow TImode operations in 32-bit.
5792 (vsx_and<mode>3_64bit): Likewise.
5793 (vsx_ior<mode>3_32bit): Likewise.
5794 (vsx_ior<mode>3_64bit): Likewise.
5795 (vsx_xor<mode>3_32bit): Likewise.
5796 (vsx_xor<mode>3_64bit): Likewise.
5797 (vsx_one_cmpl<mode>2_32bit): Likewise.
5798 (vsx_one_cmpl<mode>2_64bit): Likewise.
5799 (vsx_nor<mode>3_32bit): Likewise.
5800 (vsx_nor<mode>3_64bit): Likewise.
5801 (vsx_andc<mode>3_32bit): Likewise.
5802 (vsx_andc<mode>3_64bit): Likewise.
5803 (vsx_eqv<mode>3_32bit): Likewise.
5804 (vsx_eqv<mode>3_64bit): Likewise.
5805 (vsx_nand<mode>3_32bit): Likewise.
5806 (vsx_nand<mode>3_64bit): Likewise.
5807 (vsx_orc<mode>3_32bit): Likewise.
5808 (vsx_orc<mode>3_64bit): Likewise.
5810 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
5811 logical types in GPRs.
5813 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
5814 logical insns to rs6000.md, and allow TImode operations in
5816 (altivec_ior<mode>3): Likewise.
5817 (altivec_xor<mode>3): Likewise.
5818 (altivec_one_cmpl<mode>2): Likewise.
5819 (altivec_nor<mode>3): Likewise.
5820 (altivec_andc<mode>3): Likewise.
5822 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
5823 attributes for moving the 128-bit logical operations into
5825 (BOOL_REGS_OUTPUT): Likewise.
5826 (BOOL_REGS_OP1): Likewise.
5827 (BOOL_REGS_OP2): Likewise.
5828 (BOOL_REGS_UNARY): Likewise.
5829 (BOOL_REGS_AND_CR0): Likewise.
5830 (one_cmpl<mode>2): Add support for DI logical operations on
5831 32-bit, splitting the operations to 32-bit.
5835 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
5836 changes to combine the 32/64-bit code, allow logical operations on
5837 TI mode in 32-bit, and to use similar match_operator patterns like
5838 scalar mode uses. Combine the Altivec and VSX code for logical
5839 operations, and move it here.
5840 (ior<mode>3, 128-bit types): Likewise.
5841 (xor<mode>3, 128-bit types): Likewise.
5842 (one_cmpl<mode>3, 128-bit types): Likewise.
5843 (nor<mode>3, 128-bit types): Likewise.
5844 (andc<mode>3, 128-bit types): Likewise.
5845 (eqv<mode>3, 128-bit types): Likewise.
5846 (nand<mode>3, 128-bit types): Likewise.
5847 (orc<mode>3, 128-bit types): Likewise.
5848 (and<mode>3_internal): Likewise.
5849 (bool<mode>3_internal): Likewise.
5850 (boolc<mode>3_internal1): Likewise.
5851 (boolc<mode>3_internal2): Likewise.
5852 (boolcc<mode>3_internal1): Likewise.
5853 (boolcc<mode>3_internal2): Likewise.
5854 (eqv<mode>3_internal1): Likewise.
5855 (eqv<mode>3_internal2): Likewise.
5856 (one_cmpl1<mode>3_internal): Likewise.
5858 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
5860 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5861 Rename flag_stack_usage to flag_stack_usage_info.
5863 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
5865 * config/microblaze/sync.md: New file.
5866 * config/microblaze/microblaze.md: Include sync.md
5867 * config/microblaze/microblaze.c: Add print_operand 'y'.
5868 * config/microblaze/constraints.md: Add memory_contraint
5869 'Q' which is a single register.
5871 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
5873 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
5875 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
5877 * reload.c (find_reloads): Exit loop once we find this operand
5878 cannot be reloaded somehow for this alternative.
5880 * reload.c (find_reloads): Exit loop once we find a hard register.
5882 * rtlanal.c (computed_jump_p): Exit loop once we find label
5885 * i386.c (ix86_pad_returns): Exit loop after setting replace.
5887 * cfgloopmanip.c (remove_path): Exit loop after setting
5890 * gensupport.c (subst_dup): Avoid loop if code is not
5891 MATCH_DUP nor MATCH_OP_DUP.
5893 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
5895 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
5897 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5899 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
5900 true for SP_REGNUM if mode == ptr_mode.
5901 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
5902 with value R0_REGNUM + 31.
5904 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5906 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
5907 pad pointer-typed argument downward.
5909 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5911 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
5912 and __ILP32__ when the ILP32 model is in use.
5914 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5916 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
5917 (aarch64_load_symref_appropriately): In the case of
5918 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
5919 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
5920 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
5921 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
5922 if 'mode' doesn't equal to 'ptr_mode'.
5923 (aarch64_output_mi_thunk): Add an assertion on the alignment of
5924 'vcall_offset'; change to call aarch64_emit_move differently depending
5925 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
5926 to calculate the upper bound of 'vcall_offset'.
5927 (aarch64_cannot_force_const_mem): Change to also return true if
5929 (aarch64_legitimize_reload_address): In the case of large
5930 displacements, add new local variable 'xmode' and an assertion
5931 based on it; change to use 'xmode' to generate the new rtx and
5933 (aarch64_asm_trampoline_template): Change to generate the template
5934 differently depending on TARGET_ILP32 or not; change to use
5935 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
5936 (aarch64_trampoline_size): Removed.
5937 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
5938 and replace immediate literals with it. Change to use 'ptr_mode'
5939 instead of 'DImode' and call convert_memory_address if the mode
5940 of 'fnaddr' doesn't equal to 'ptr_mode'.
5941 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
5943 (aarch64_elf_asm_destructor): Likewise.
5944 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
5945 on TARGET_ILP32 instead of aarch64_trampoline_size.
5946 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
5947 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
5948 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
5949 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
5950 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
5951 (storewb_pair<GPI:mode>_<P:mode>): ... this.
5952 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
5953 depending on the value of 'mode'.
5954 (add_losym_<mode>): New.
5955 (ldr_got_small_<mode>): New, based on ldr_got_small.
5956 (ldr_got_small): Remove.
5957 (ldr_got_small_sidi): New.
5958 * config/aarch64/iterators.md (P): New.
5959 (PTR): Change to 'ptr_mode' in the condition.
5961 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5963 * config.gcc (aarch64*-*-*): Support --with-abi.
5964 (aarch64*-*-elf): Support --with-multilib-list.
5965 (aarch64*-*-linux*): Likewise.
5966 (supported_defaults): Add abi to aarch64*-*-*.
5967 * configure.ac: Mention AArch64 for --with-multilib-list.
5968 * configure: Re-generated.
5969 * config/aarch64/biarchilp32.h: New file.
5970 * config/aarch64/biarchlp64.h: New file.
5971 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
5973 (MULTILIB_DEFAULTS): Ditto.
5974 (DRIVER_SELF_SPECS): Ditto.
5975 (ASM_SPEC): Update to also substitute -mabi.
5976 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
5977 file whose name depends on -mabi= and -mbig-endian.
5978 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
5980 (POINTER_SIZE): New define.
5981 (POINTERS_EXTEND_UNSIGNED): Ditto.
5982 (enum aarch64_abi_type): New enumeration tag.
5983 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
5984 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
5985 (TARGET_ILP32): New define.
5986 * config/aarch64/aarch64.opt (mabi): New.
5988 (ilp32, lp64): New values for -mabi.
5989 * config/aarch64/t-aarch64 (comma): New define.
5990 (MULTILIB_OPTIONS): Ditto.
5991 (MULTILIB_DIRNAMES): Ditto.
5992 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
5993 * doc/invoke.texi: Document -mabi for AArch64.
5995 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
5997 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
5999 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6000 Anton Blanchard <anton@au1.ibm.com>
6002 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6003 endianness when selecting field to splat.
6005 2013-07-22 Eric Christopher <echristo@gmail.com>
6007 * dwarf2out.c (die_odr_checksum): New function to use
6008 CHECKSUM_ macros and ULEB128 for DIE tag.
6009 (generate_type_signature): Use.
6011 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6013 * config.gcc (sparc*-*-*): Accept leon3 processor.
6014 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6015 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6016 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6017 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6018 (mfix-ut699): Adjust comment.
6019 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6020 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6021 (CPP_CPU_SPEC): Likewise.
6022 (ASM_CPU_SPEC): Likewise.
6023 * config/sparc/sparc.c (leon3_cost): New constant.
6024 (sparc_option_override): Add leon3 support.
6025 (mem_ref): New function.
6026 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6027 (sparc_do_work_around_errata): Look into the instruction in the delay
6028 slot and adjust accordingly. Add fix for the data cache nullify issues
6029 of the UT699. Change insertion position for the NOP.
6030 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6031 (leon3_load): New reservation.
6032 (leon_store): Bump latency to 2.
6033 (grfpu): New automaton.
6034 (grfpu_alu): New unit.
6035 (grfpu_ds): Likewise.
6036 (leon_fp_alu): Adjust.
6037 (leon_fp_mult): Delete.
6038 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6039 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6040 * config/sparc/sparc.md (cpu): Add leon3.
6041 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6043 (atomic_test_and_set): Likewise.
6046 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6048 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6049 default for R5900 targets.
6050 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6051 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6052 * config/mips/mips.c (mips_option_override): Report an error for
6053 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6054 for -march=r5900 -mhard-float.
6056 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6058 * df-problems.c (can_move_insns_across): Exit loop once we
6059 find a non-fixed, non-global register.
6061 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6064 * omega.c (omega_eliminate_red): Break after setting red_found.
6065 (omega_problem_has_red_equations): Similarly after setting found.
6066 (omega_query_variable): Similarly after setting coupled.
6068 2013-07-22 Marek Polacek <polacek@redhat.com>
6070 * gimplify.c: Don't include gimple.h twice.
6072 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6074 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6075 instead of TARGET_THUMB1.
6076 (Pz): New constraint.
6077 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6079 (compare_negsi_si): Likewise.
6080 (compare_addsi2_op0): Likewise.
6081 (compare_addsi2_op1): Likewise.
6082 (addsi3_carryin_<optab>): Likewise.
6083 (addsi3_carryin_alt2_<optab>): Likewise.
6084 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6085 for arm_restrict_it.
6086 (subsi3_carryin): Likewise.
6087 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6088 (minmax_arithsi): Disable for arm_restrict_it.
6089 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6090 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6091 (satsi_<SAT:code>_shift): Likewise.
6092 (arm_shiftsi3): Add alternative for 16-bit encoding.
6093 (arm32_movhf): Disable for arm_restrict_it.
6094 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6095 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6097 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6099 * config/arm/arm.md (attribute "insn"): Delete.
6100 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6101 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6102 (not_shiftsi): Update for attribute change.
6103 (not_shiftsi_compare0): Likewise.
6104 (not_shiftsi_compare0_scratch): Likewise.
6105 (arm_one_cmplsi2): Likewise.
6106 (thumb1_one_cmplsi2): Likewise.
6107 (notsi_compare0): Likewise.
6108 (notsi_compare0_scratch): Likewise.
6109 (thumb1_movdi_insn): Likewise.
6110 (arm_movsi_insn): Likewise.
6111 (movhi_insn_arch4): Likewise.
6112 (movhi_bytes): Likewise.
6113 (arm_movqi_insn): Likewise.
6114 (thumb1_movqi_insn): Likewise.
6115 (arm32_movhf): Likewise.
6116 (thumb1_movhf): Likewise.
6117 (arm_movsf_soft_insn): Likewise.
6118 (thumb1_movsf_insn): Likewise.
6119 (thumb_movdf_insn): Likewise.
6120 (movsicc_insn): Likewise.
6121 (movsfcc_soft_insn): Likewise.
6122 (and_scc): Likewise.
6123 (cond_move): Likewise.
6124 (if_move_not): Likewise.
6125 (if_not_move): Likewise.
6126 (if_shift_move): Likewise.
6127 (if_move_shift): Likewise.
6128 (if_shift_shift): Likewise.
6129 (if_not_arith): Likewise.
6130 (if_arith_not): Likewise.
6131 (cond_move_not): Likewise.
6132 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6133 (neon_mov<mode>): Likewise.
6134 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6135 (thumb2_movsi_vfp): Likewise.
6136 (movsf_vfp): Likewise.
6137 (thumb2_movsf_vfp): Likewise.
6138 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6140 (cortexa7_older_only): Likewise.
6141 (cortexa7_younger): Likewise.
6142 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6143 (1020alu_shift_op): Likewise.
6144 (1020alu_shift_reg_op): Likewise.
6145 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6146 (alu_shift_op): Likewise.
6147 (alu_shift_reg_op): Likewise.
6148 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6149 (11_alu_shift_op): Likewise.
6150 (11_alu_shift_reg_op): Likewise.
6151 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6152 (9_alu_shift_reg_op): Likewise.
6153 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6155 (cortex_a15_alu_shift): Likewise.
6156 (cortex_a15_alu_shift_reg): Likewise.
6157 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6158 (cortex_a5_alu_shift): Likewise.
6159 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6161 (cortex_a53_alu_shift): Likewise.
6162 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6164 (cortex_a7_alu_reg): Likewise.
6165 (cortex_a7_alu_shift): Likewise.
6166 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6167 (cortex_a8_alu_shift): Likewise.
6168 (cortex_a8_alu_shift_reg): Likewise.
6169 (cortex_a8_mov): Likewise.
6170 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6171 (cortex_a9_dp_shift): Likewise.
6172 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6173 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6174 (cortex_r4_mov): Likewise.
6175 (cortex_r4_alu_shift): Likewise.
6176 (cortex_r4_alu_shift_reg): Likewise.
6177 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6178 (526_alu_shift_op): Likewise.
6179 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6180 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6181 (626te_alu_shift_op): Likewise.
6182 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6183 (726te_alu_op): Likewise.
6184 (726te_alu_shift_op): Likewise.
6185 (726te_alu_shift_reg_op): Likewise.
6186 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6187 (mp626_alu_shift_op): Likewise.
6188 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6189 (pj4_alu_e1_conds): Likewise.
6190 (pj4_alu): Likewise.
6191 (pj4_alu_conds): Likewise.
6192 (pj4_shift): Likewise.
6193 (pj4_shift_conds): Likewise.
6194 (pj4_alu_shift): Likewise.
6195 (pj4_alu_shift_conds): Likewise.
6197 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6199 * config/arm/predicates.md (shiftable_operator_strict_it):
6201 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6202 Disable cond_exec version for arm_restrict_it.
6203 (thumb2_smaxsi3): Convert to generate cond_exec.
6204 (thumb2_sminsi3): Likewise.
6205 (thumb32_umaxsi3): Likewise.
6206 (thumb2_uminsi3): Likewise.
6207 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6208 (thumb2_neg_abssi2): Likewise.
6209 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6210 (thumb2_movsicc_insn): Adjust alternatives.
6211 (thumb2_mov_negscc): Disable for arm_restrict_it.
6212 (thumb2_mov_negscc_strict_it): New pattern.
6213 (thumb2_mov_notscc_strict_it): New pattern.
6214 (thumb2_mov_notscc): Disable for arm_restrict_it.
6215 (thumb2_ior_scc): Likewise.
6216 (thumb2_ior_scc_strict_it): New pattern.
6217 (thumb2_cond_move): Adjust for arm_restrict_it.
6218 (thumb2_cond_arith): Disable for arm_restrict_it.
6219 (thumb2_cond_arith_strict_it): New pattern.
6220 (thumb2_cond_sub): Adjust for arm_restrict_it.
6221 (thumb2_movcond): Likewise.
6222 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6223 (thumb2_zero_extendhisi2_v6): Likewise.
6224 (thumb2_zero_extendqisi2_v6): Likewise.
6225 (orsi_notsi_si): Likewise.
6226 (orsi_not_shiftsi_si): Likewise.
6228 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
6230 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6231 instruction sequence is 1 byte shorter.
6233 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
6235 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6236 it is not needed after split.
6238 2013-07-20 Iain Sandoe <iain@codesourcery.com>
6241 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6242 second label for nonlocal goto receivers. Don't output pic base labels
6243 unless we're producing PIC; mark that action unreachable().
6244 (ix86_save_reg): If the function contains a nonlocal label, save the
6246 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
6247 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
6248 (update_pic_label_number_if_needed): New.
6249 (machopic_output_function_base_name): Adjust for nonlocal receiver
6251 (machopic_should_output_picbase_label): New.
6252 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6253 (nonlocal_goto_receiver): New insn and split.
6255 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6257 * config/aarch64/aarch64-builtins.c
6258 (aarch64_fold_builtin): Fold abs in all modes.
6259 * config/aarch64/aarch64-simd-builtins.def
6260 (abs): Enable for all modes.
6261 * config/aarch64/arm_neon.h
6262 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6263 (vabs_f64): Add missing intrinsic.
6265 2013-07-19 Ian Bolton <ian.bolton@arm.com>
6267 * config/aarch64/arm_neon.h (vabs_s64): New function
6269 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
6272 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6273 * config/avr/avr.md (adjust_len): Add `round'.
6274 * config/avr/avr-protos.h (avr_out_round): New prototype.
6275 (avr_out_plus): Add `out_label' argument.
6276 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6277 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6278 Handle the case where `insn' is just a pattern.
6279 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6280 (avr_out_round): New function.
6281 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6283 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
6285 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6286 Add check for flag_stack_usage to handle -fstack-usage support
6288 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
6290 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6291 interaction for new Power8 flags and VSX.
6293 2013-07-18 Sriraman Tallam <tmsriram@google.com>
6296 * tree-inline.c (expand_call_inline): Emit errors during
6297 early_inlining only if optimization is not turned on.
6299 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6303 * passes.c (init_optimization_passes): Move the construction of
6304 the pass hierarchy into a new passes.def file.
6306 * Makefile.in (passes.o): Add dependency on passes.def.
6308 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6310 * passes.c (init_optimization_passes): Introduce macros for
6311 constructing the tree of passes (INSERT_PASSES_AFTER,
6312 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6313 TERMINATE_PASS_LIST).
6315 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6316 Wei Mi <wmi@google.com>
6318 PR rtl-optimization/57878
6319 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6321 (reload_pseudo_compare_func): Check nregs first for reload
6324 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6326 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6328 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
6330 * read-rtl.c (validate_const_int): Once an invalid character is
6331 seen, quit the loop.
6333 * gengtype.c (write_roots): Similarly once we find the "deletable"
6334 or "if_marked" option.
6336 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6338 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6339 "xtab" and "sat". Move value "clz" from here to ...
6340 (attriubte "type"): ... here.
6341 (satsi_<SAT:code>): Delete "insn" attribute.
6342 (satsi_<SAT:code>_shift): Likewise.
6343 (arm_zero_extendqisi2addsi): Likewise.
6344 (arm_extendqisi2addsi): Likewise.
6345 (clzsi2): Update for attribute changes.
6346 (rbitsi2): Likewise.
6347 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6349 (arm_usatsihi): Likewise.
6350 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6352 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6354 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
6355 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6356 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
6357 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6358 in alphabetical order.
6359 (attribute "core_cycles"): Update for attribute changes.
6360 (arm_addsi3): Likewise.
6361 (addsi3_compare0): Likewise.
6362 (addsi3_compare0_scratch): Likewise.
6363 (addsi3_compare_op1): Likewise.
6364 (addsi3_compare_op2): Likewise.
6365 (compare_addsi2_op0): Likewise.
6366 (compare_addsi2_op1): Likewise.
6367 (addsi3_carryin_shift_<optab>): Likewise.
6368 (subsi3_carryin_shift): Likewise.
6369 (rsbsi3_carryin_shift): Likewise.
6370 (arm_subsi3_insn): Likewise.
6371 (subsi3_compare0): Likewise.
6372 (subsi3_compare): Likewise.
6373 (arm_andsi3_insn): Likewise.
6374 (thumb1_andsi3_insn): Likewise.
6375 (andsi3_compare0): Likewise.
6376 (andsi3_compare0_scratch): Likewise.
6377 (zeroextractsi_compare0_scratch
6378 (andsi_not_shiftsi_si): Likewise.
6379 (iorsi3_insn): Likewise.
6380 (iorsi3_compare0): Likewise.
6381 (iorsi3_compare0_scratch): Likewise.
6382 (arm_xorsi3): Likewise.
6383 (thumb1_xorsi3_insn): Likewise.
6384 (xorsi3_compare0): Likewise.
6385 (xorsi3_compare0_scratch): Likewise.
6386 (satsi_<SAT:code>_shift): Likewise.
6388 (arm_shiftsi3): Likewise.
6389 (shiftsi3_compare0): Likewise.
6390 (not_shiftsi): Likewise.
6391 (not_shiftsi_compare0): Likewise.
6392 (not_shiftsi_compare0_scratch): Likewise.
6393 (arm_one_cmplsi2): Likewise.
6394 (thumb_one_complsi2): Likewise.
6395 (notsi_compare0): Likewise.
6396 (notsi_compare0_scratch): Likewise.
6397 (thumb1_zero_extendhisi2): Likewise.
6398 (arm_zero_extendhisi2): Likewise.
6399 (arm_zero_extendhisi2_v6): Likewise.
6400 (arm_zero_extendhisi2addsi): Likewise.
6401 (thumb1_zero_extendqisi2): Likewise.
6402 (thumb1_zero_extendqisi2_v6): Likewise.
6403 (arm_zero_extendqisi2): Likewise.
6404 (arm_zero_extendqisi2_v6): Likewise.
6405 (arm_zero_extendqisi2addsi): Likewise.
6406 (thumb1_extendhisi2): Likewise.
6407 (arm_extendhisi2): Likewise.
6408 (arm_extendhisi2_v6): Likewise.
6409 (arm_extendqisi): Likewise.
6410 (arm_extendqisi_v6): Likewise.
6411 (arm_extendqisi2addsi): Likewise.
6412 (thumb1_extendqisi2): Likewise.
6413 (thumb1_movdi_insn): Likewise.
6414 (arm_movsi_insn): Likewise.
6415 (movsi_compare0): Likewise.
6416 (movhi_insn_arch4): Likewise.
6417 (movhi_bytes): Likewise.
6418 (arm_movqi_insn): Likewise.
6419 (thumb1_movqi_insn): Likewise.
6420 (arm32_movhf): Likewise.
6421 (thumb1_movhf): Likewise.
6422 (arm_movsf_soft_insn): Likewise.
6423 (thumb1_movsf_insn): Likewise.
6424 (movdf_soft_insn): Likewise.
6425 (thumb_movdf_insn): Likewise.
6426 (arm_cmpsi_insn): Likewise.
6427 (cmpsi_shiftsi): Likewise.
6428 (cmpsi_shiftsi_swp): Likewise.
6429 (arm_cmpsi_negshiftsi_si): Likewise.
6430 (movsicc_insn): Likewise.
6431 (movsfcc_soft_insn): Likewise.
6432 (arith_shiftsi): Likewise.
6433 (arith_shiftsi_compare0
6434 (arith_shiftsi_compare0_scratch
6435 (sub_shiftsi): Likewise.
6436 (sub_shiftsi_compare0
6437 (sub_shiftsi_compare0_scratch
6438 (and_scc): Likewise.
6439 (cond_move): Likewise.
6440 (if_plus_move): Likewise.
6441 (if_move_plus): Likewise.
6442 (if_move_not): Likewise.
6443 (if_not_move): Likewise.
6444 (if_shift_move): Likewise.
6445 (if_move_shift): Likewise.
6446 (if_shift_shift): Likewise.
6447 (if_not_arith): Likewise.
6448 (if_arith_not): Likewise.
6449 (cond_move_not): Likewise.
6450 (thumb1_ashlsi3): Set type attribute.
6451 (thumb1_ashrsi3): Likewise.
6452 (thumb1_lshrsi3): Likewise.
6453 (thumb1_rotrsi3): Likewise.
6454 (shiftsi3_compare0_scratch): Likewise.
6455 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
6456 (neon_mov<mode>): Likewise.
6457 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
6459 (thumb2_movsi_insn): Likewise.
6460 (thumb2_cmpsi_neg_shiftsi): Likewise.
6461 (thumb2_extendqisi_v6): Likewise.
6462 (thumb2_zero_extendhisi2_v6): Likewise.
6463 (thumb2_zero_extendqisi2_v6): Likewise.
6464 (thumb2_shiftsi3_short): Likewise.
6465 (thumb2_addsi3_compare0_scratch): Likewise.
6466 (orsi_not_shiftsi_si): Likewise.
6467 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6468 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
6470 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6471 (1020alu_shift_op): Likewise.
6472 (1020alu_shift_reg_op): Likewise.
6473 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6474 (alu_shift_op): Likewise.
6475 (alu_shift_reg_op): Likewise.
6476 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6477 (11_alu_shift_op): Likewise.
6478 (11_alu_shift_reg_op): Likewise.
6479 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6480 (9_alu_shift_reg_op): Likewise.
6481 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6483 (cortex_a15_alu_shift): Likewise.
6484 (cortex_a15_alu_shift_reg): Likewise.
6485 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
6487 (cortex_a5_alu_shift): Likewise.
6488 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
6490 (cortex_a53_alu_shift): Likewise.
6491 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6493 (cortex_a7_alu_reg): Likewise.
6494 (cortex_a7_alu_shift): Likewise.
6495 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
6497 (cortex_a8_alu_shift): Likewise.
6498 (cortex_a8_alu_shift_reg): Likewise.
6499 (cortex_a8_mov): Likewise.
6500 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6501 (cortex_a9_dp_shift): Likewise.
6502 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
6504 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
6506 (cortex_r4_mov): Likewise.
6507 (cortex_r4_alu_shift): Likewise.
6508 (cortex_r4_alu_shift_reg): Likewise.
6509 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6510 (526_alu_shift_op): Likewise.
6511 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6512 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6513 (626te_alu_shift_op): Likewise.
6514 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
6515 (726te_alu_op): Likewise.
6516 (726te_alu_shift_op): Likewise.
6517 (726te_alu_shift_reg_op): Likewise.
6518 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6519 (mp626_alu_shift_op): Likewise.
6520 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
6521 (pj4_alu_e1_conds): Likewise.
6522 (pj4_alu): Likewise.
6523 (pj4_alu_conds): Likewise.
6524 (pj4_shift): Likewise.
6525 (pj4_shift_conds): Likewise.
6526 (pj4_alu_shift): Likewise.
6527 (pj4_alu_shift_conds): Likewise.
6528 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6530 (cortexa7_older_only): Likewise.
6531 (cortexa7_younger): Likewise.
6533 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6535 * ipa-pure-const.c (generate_summary): Rename to...
6536 (pure_const_generate_summary): ... this.
6538 2013-07-17 Iain Sandoe <iain@codesourcery.com>
6540 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
6542 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
6545 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
6548 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6550 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
6551 enabled without -march=zEC12.
6552 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
6555 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6557 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
6558 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
6559 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
6560 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
6561 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
6562 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
6563 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
6565 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
6566 and ISA_HAS_NMADD3_NMSUB3.
6567 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
6568 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
6569 (nmsub4<mode>, nmsub3<mode>): Likewise.
6570 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
6572 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6574 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
6575 TARGET_MIPS5400 checking.
6577 2013-07-16 Jakub Jelinek <jakub@redhat.com>
6578 Peter Bergner <bergner@vnet.ibm.com>
6580 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
6581 registers in the comment.
6582 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
6583 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
6584 rather than FIRST_PSEUDO_REGISTERS.
6586 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
6588 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
6589 enable extra ISA flags with TARGET_HTM.
6591 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6593 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
6596 2013-07-15 Cong Hou <congh@google.com>
6598 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
6599 in compare function for sorting.
6601 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
6603 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
6604 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
6605 * config/rs6000/rs6000.opt: Add -mhtm option.
6606 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
6607 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
6608 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6609 __HTM__ if the HTM instructions are available.
6610 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
6611 htm_spr_reg_operand): New define_predicates.
6612 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
6613 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
6615 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
6616 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
6617 HTM builtin functions.
6618 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
6619 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
6620 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
6621 (rs6000_builtin_mask_calculate): Likewise.
6622 (rs6000_option_override_internal): Likewise.
6623 (bdesc_htm): Add new HTM builtin support.
6624 (htm_spr_num): New function.
6625 (htm_spr_regno): Likewise.
6626 (rs6000_htm_spr_icode): Likewise.
6627 (htm_expand_builtin): Likewise.
6628 (htm_init_builtins): Likewise.
6629 (rs6000_expand_builtin): Add support for HTM builtin functions.
6630 (rs6000_init_builtins): Likewise.
6631 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
6633 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
6634 (TARGET_HTM, MASK_HTM): Define macros.
6635 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
6636 (FIXED_REGISTERS): Likewise.
6637 (CALL_USED_REGISTERS): Likewise.
6638 (CALL_REALLY_USED_REGISTERS): Likewise.
6639 (REG_ALLOC_ORDER): Likewise.
6640 (enum reg_class): Likewise.
6641 (REG_CLASS_NAMES): Likewise.
6642 (REG_CLASS_CONTENTS): Likewise.
6643 (REGISTER_NAMES): Likewise.
6644 (ADDITIONAL_REGISTER_NAMES): Likewise.
6645 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
6646 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
6647 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
6648 * config/rs6000/htm.md: New file.
6649 * config/rs6000/htmintrin.h: New file.
6650 * config/rs6000/htmxlintrin.h: New file.
6652 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
6654 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6655 Define SYMBOL_TINY_GOT, update comment.
6656 * config/aarch64/aarch64.c
6657 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
6658 (aarch64_expand_mov_immediate): Likewise.
6659 (aarch64_print_operand): Likewise.
6660 (aarch64_classify_symbol): Likewise.
6661 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
6662 (ldr_got_tiny): Define.
6664 2013-07-13 Tobias Grosser <tobias@grosser.es>
6666 PR tree-optimization/54094
6667 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
6668 scheduling dimension for the parallelism check from the polyhedral
6669 information in the AST.
6670 * graphite-dependences.c (carries_deps): Do not assume the schedule is
6673 2013-07-13 Jason Merrill <jason@redhat.com>
6675 * print-tree.c (debug_vec_tree): Use debug_raw.
6676 (debug_raw (vec<tree, va_gc> &)): New.
6677 (debug_raw (vec<tree, va_gc> *)): New.
6678 * tree.h: Declare them.
6680 2013-07-13 Bin Cheng <bin.cheng@arm.com>
6682 * ifcvt.c (ifcvt_after_combine): New static variable.
6683 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
6685 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
6686 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
6687 rest_of_handle_if_after_reload): Pass new argument for if_convert.
6689 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
6691 * config/mips/mips.c (mips_expand_call): Remove empty statement.
6693 2013-07-12 Michael Matz <matz@suse.de>
6696 * convert.c (convert_to_real): Reject non-float inner types.
6698 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
6700 * config/aarch64/aarch64-protos.h
6701 (aarch64_simd_immediate_valid_for_move): Remove.
6702 * config/aarch64/aarch64.c (simd_immediate_info): New member.
6703 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
6705 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
6707 2013-07-11 Steve Ellcey <sellcey@mips.com>
6709 * config/mips/mips.c (mips_conditional_register_usage): Do not
6710 use t[0-7] registers in MIPS16 mode when optimizing for size.
6712 2013-07-11 Sriraman Tallam <tmsriram@google.com>
6714 * config/i386/i386.c (dispatch_function_versions): Fix array
6715 indexing of function_version_info to match actual_versions.
6717 2013-07-11 Teresa Johnson <tejohnson@google.com>
6719 * vec.h (struct va_gc): Move release out-of-line.
6720 (va_gc::release): Call ggc_free on released vec.
6722 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6724 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
6725 Require GOT register as additional operand in UNSPEC.
6726 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
6727 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6728 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6729 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
6730 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
6731 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
6732 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
6734 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
6737 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
6738 name seen by assembler/linker rather if available.
6740 2013-07-11 Andreas Schwab <schwab@suse.de>
6742 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
6744 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
6746 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
6748 2013-07-10 Joseph Myers <joseph@codesourcery.com>
6750 * doc/tm.texi.in: Move hook documentation to ....
6751 * target.def: ... here.
6753 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
6755 * doc/tm.texi: Regenerate.
6757 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
6760 * doc/invoke.texi: Document -Wconditionally-supported.
6762 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6765 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
6768 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6771 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
6772 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
6773 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
6774 Remove duplicate devices.
6775 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
6776 * config/avr/t-multilib: Regenerate.
6777 * config/avr/avr-tables.opt: Regenerate.
6778 * doc/avr-mmcu.texi: Regenerate.
6780 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6783 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
6785 2013-07-10 Graham Stott <graham.stott@btinternet.com>
6787 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
6788 the cost of MULT when optimizing for size.
6790 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6792 * config/cr16/cr16-protos.h: Don't include target.h.
6794 2013-07-09 Joseph Myers <joseph@codesourcery.com>
6796 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
6797 adjust register size for TDmode and TFmode for VSX registers.
6799 2013-07-08 Kai Tietz <ktietz@redhat.com>
6802 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
6803 hook_bool_const_tree_true.
6805 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6807 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
6808 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
6809 * config/s390/s390.md: Define FPR*_REGNUM constants.
6810 Fix FPR2_REGNUM constant (18 -> 17).
6811 ("*trunc<BFP:mode><DFP_ALL:mode>2")
6812 ("*trunc<DFP_ALL:mode><BFP:mode>2")
6813 ("trunc<BFP:mode><DFP_ALL:mode>2")
6814 ("trunc<DFP_ALL:mode><BFP:mode>2")
6815 ("*extend<BFP:mode><DFP_ALL:mode>2")
6816 ("*extend<DFP_ALL:mode><BFP:mode>2")
6817 ("extend<BFP:mode><DFP_ALL:mode>2")
6818 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
6821 2013-07-08 Graham Stott <graham.stott@btinternet.com>
6823 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
6825 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6827 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
6828 and cfun_fpr_bit_p to cfun_fpr_save_p.
6829 (s390_frame_area, s390_register_info, s390_frame_info)
6830 (s390_emit_prologue, s390_emit_epilogue)
6831 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
6833 * config/s390/s390.h: Define *_REGNUM macros for floating point
6836 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
6838 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
6840 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
6842 PR rtl-optimization/57786
6843 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
6844 and break out of the loop when it is set to false.
6846 2013-07-08 Jakub Jelinek <jakub@redhat.com>
6849 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
6850 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
6851 (const_int 63)) 0)).
6852 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
6853 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
6854 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
6856 PR rtl-optimization/57829
6857 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
6858 mask bits outside of mode are just sign-extension from mode to HWI.
6860 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6862 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
6863 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
6864 adjust_address instead of change_address to keep info about alignment.
6865 (emit_strmov): Remove.
6866 (emit_memmov): New function.
6867 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
6868 (expand_movmem_epilogue): Likewise and return updated rtx for
6870 (expand_constant_movmem_prologue): Likewise and return updated rtx for
6871 destination and source.
6872 (decide_alignment): Refactor, handle vector_loop.
6873 (ix86_expand_movmem): Likewise.
6874 (ix86_expand_setmem): Likewise.
6875 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
6877 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
6879 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
6880 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
6882 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
6884 * config/i386/sse.md (sse_movlhps): Change alternative 3
6885 of operand 2 to "m".
6887 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
6890 * config/i386/sse.md (iptr): New mode attribute.
6891 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
6892 (<sse>_vm<plusminus_insn><mode>3): Ditto.
6893 (<sse>_vmmul<mode>3): Ditto.
6894 (<sse>_vmdiv<mode>3): Ditto.
6895 (sse_vmrcpv4sf2): Ditto.
6896 (<sse>_vmsqrt<mode>2): Ditto.
6897 (sse_vmrsqrtv4sf2): Ditto.
6898 (<sse>_vm<code><mode>3): Ditto.
6899 (avx_vmcmp<mode>3): Ditto.
6900 (<sse>_vmmaskcmp<mode>3): Ditto.
6901 (<sse>_comi): Ditto.
6902 (<sse>_ucomi): Ditto.
6903 (*xop_vmfrcz_<mode>): Ditto.
6904 (*fmai_fmadd_<mode>): Ditto.
6905 (*fmai_fmsub_<mode>): Ditto.
6906 (*fmai_fnmadd_<mode>): Ditto.
6907 (*fmai_fnmsub_<mode>): Ditto.
6908 (*fma4i_vmfmadd_<mode>): Ditto.
6909 (*fma4i_vmfmsub_<mode>): Ditto.
6910 (*fma4i_vmfnmadd_<mode>): Ditto.
6911 (*fma4i_vmfnmsub_<mode>): Ditto.
6912 (*xop_vmfrcz_<mode>): Ditto.
6913 (sse_cvtps2pi): Ditto.
6914 (sse_cvttps2pi): Ditto.
6915 (sse_cvtss2si): Ditto.
6916 (sse_cvtss2si_2): Ditto.
6917 (sse_cvtss2siq_2): Ditto.
6918 (sse_cvttss2si): Ditto.
6919 (sse_cvttss2siq): Ditto.
6920 (sse_cvtsd2si): Ditto.
6921 (sse_cvtsd2si_2): Ditto.
6922 (sse_cvtsd2siq_2): Ditto.
6923 (sse_cvttsd2si): Ditto.
6924 (sse_cvttsd2siq): Ditto.
6925 (sse_cvtsd2ss): Ditto.
6926 (sse_cvtss2sd): Ditto.
6927 (avx2_pbroadcast<mode>): Ditto.
6928 (avx2_pbroadcast<mode>_1): Ditto.
6929 (*avx_vperm_broadcast_v4sf): Ditto.
6931 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
6932 (sse_movlhps): Ditto.
6933 (sse_storehps): Ditto.
6934 (sse_loadhps): Ditto.
6935 (sse_storelps): Ditto.
6936 (sse_loadlps): Ditto.
6937 (*vec_concatv4sf): Ditto.
6938 (*vec_interleave_highv2df): Ditto.
6939 (*vec_interleave_lowv2df): Ditto.
6940 (*vec_extractv2df_1_sse): Ditto.
6941 (*vec_extractv2df_0_sse): Ditto.
6942 (sse2_storelpd): Ditto.
6943 (sse2_loadlpd): Ditto.
6944 (sse2_movsd): Ditto.
6945 (*vec_concatv4si): Ditto.
6946 (vec_concatv2di): Ditto.
6948 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
6949 for Intel asm dialect.
6950 (mmx_punpcklwd): Ditto.
6951 (mmx_punpckldq): Ditto.
6953 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
6954 for intel assembler dialect.
6956 2013-07-06 Jakub Jelinek <jakub@redhat.com>
6959 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
6960 for BUILT_IN_C{LZ,LRSB}*.
6961 * tree.h (CASE_INT_FN): Add FN##IMAX case.
6962 * tree-vrp.c (extract_range_basic): Handle
6963 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
6964 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
6965 fall thru to code calling set_value*.
6966 * builtins.c (expand_builtin): Remove *IMAX cases.
6967 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
6968 if width is bigger than 2*HWI.
6970 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
6972 PR rtl-optimization/55342
6973 * lra-int.h (lra_subreg_reload_pseudos): New.
6974 * lra.c: Add undoing optional reloads to the block diagram.
6975 (lra_subreg_reload_pseudos): New.
6976 (lra_optional_reload_pseudos): Change comments.
6977 (lra): Init and clear lra_subreg_reload_pseudos. Clear
6978 lra_optional_reload_pseudos after undo transformations.
6979 * lra-assigns.c (pseudo_prefix_title): New.
6980 (lra_setup_reg_renumber): Use it.
6981 (spill_for): Ditto. Check subreg reload pseudos too.
6982 (assign_by_spills): Consider subreg reload pseudos too.
6983 * lra-constraints.c (simplify_operand_subreg): Use
6984 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
6985 (curr_insn_transform): Recognize and do optional reloads.
6986 (undo_optional_reloads): New.
6987 (lra_undo_inheritance): Call undo_optional_reloads.
6989 2013-07-05 Thomas Quinot <quinot@adacore.com>
6991 * tree-complex.c (expand_complex_operations_1): Fix typo.
6993 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
6995 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
6996 (tune_params): New member 'const vec_costs'.
6997 * config/aarch64/aarch64.c (generic_vector_cost): New.
6998 (generic_tunings): New member 'generic_vector_cost'.
6999 (aarch64_builtin_vectorization_cost): New.
7000 (aarch64_add_stmt_cost): New.
7001 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7002 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7004 2013-07-03 Jakub Jelinek <jakub@redhat.com>
7007 * config/i386/predicates.md (vsib_address_operand): Disallow
7008 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7010 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7013 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7014 expand_builtin_setjmp_receiver.
7015 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7016 with NULL for the label parameter.
7017 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7018 the frame-pointer. Adjust comments.
7019 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7020 only if LABEL is non-NULL.
7022 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7024 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7025 (ARM_ABI_AAPCS64): Ditto.
7027 (ARM_DEFAULT_ABI): Ditto.
7029 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7031 * config/aarch64/aarch64-builtins.c
7032 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7033 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7035 * config/aarch64/aarch64-simd.md
7036 (aarch64_ld1<VALL:mode>): New.
7037 (aarch64_st1<VALL:mode>): Likewise.
7038 * config/aarch64/arm_neon.h
7039 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7041 2013-07-02 Sriraman Tallam <tmsriram@google.com>
7043 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7045 (ix86_option_override_internal):Turn on all -mavx target flags by
7046 default as they are dependent on AVX anyway.
7048 2013-07-02 Cary Coutant <ccoutant@google.com>
7050 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7052 (loc_checksum_ordered): Likewise.
7053 (hash_loc_operands): Remove inline keyword.
7055 2013-07-02 Jakub Jelinek <jakub@redhat.com>
7057 PR tree-optimization/57741
7058 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7059 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7060 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7061 Allow REAL_CST step_exprs if flag_associative_math.
7062 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7064 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7066 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
7068 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7070 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7072 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7074 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7076 (iorsi3_insn): Likewise.
7077 (arm_xorsi3): Likewise.
7079 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7081 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7083 (attribute "type"): ... here, and prefix with "wmmx_".
7084 (attribute "core_cycles"): Update for attribute changes.
7085 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7086 (tbcstv4hi): Likewise.
7087 (tbcstv2si): Likewise.
7088 (iwmmxt_iordi3): Likewise.
7089 (iwmmxt_xordi3): Likewise.
7090 (iwmmxt_anddi3): Likewise.
7091 (iwmmxt_nanddi3): Likewise.
7092 (iwmmxt_arm_movdi): Likewise.
7093 (iwmmxt_movsi_insn): Likewise.
7094 (mov<mode>_internal): Likewise.
7095 (and<mode>3_iwmmxt): Likewise.
7096 (ior<mode>3_iwmmxt): Likewise.
7097 (xor<mode>3_iwmmxt): Likewise.
7098 (add<mode>3_iwmmxt): Likewise.
7099 (ssaddv8qi3): Likewise.
7100 (ssaddv4hi3): Likewise.
7101 (ssaddv2si3): Likewise.
7102 (usaddv8qi3): Likewise.
7103 (usaddv4hi3): Likewise.
7104 (usaddv2si3): Likewise.
7105 (sub<mode>3_iwmmxt): Likewise.
7106 (sssubv8qi3): Likewise.
7107 (sssubv4hi3): Likewise.
7108 (sssubv2si3): Likewise.
7109 (ussubv8qi3): Likewise.
7110 (ussubv4hi3): Likewise.
7111 (ussubv2si3): Likewise.
7112 (mulv4hi3_iwmmxt): Likewise.
7113 (smulv4hi3_highpart): Likewise.
7114 (umulv4hi3_highpart): Likewise.
7115 (iwmmxt_wmacs): Likewise.
7116 (iwmmxt_wmacsz): Likewise.
7117 (iwmmxt_wmacu): Likewise.
7118 (iwmmxt_wmacuz): Likewise.
7119 (iwmmxt_clrdi): Likewise.
7120 (iwmmxt_clrv8qi): Likewise.
7121 (iwmmxt_clr4hi): Likewise.
7122 (iwmmxt_clr2si): Likewise.
7123 (iwmmxt_uavgrndv8qi3): Likewise.
7124 (iwmmxt_uavgrndv4hi3): Likewise.
7125 (iwmmxt_uavgv8qi3): Likewise.
7126 (iwmmxt_uavgv4hi3): Likewise.
7127 (iwmmxt_tinsrb): Likewise.
7128 (iwmmxt_tinsrh): Likewise.
7129 (iwmmxt_tinsrw): Likewise.
7130 (iwmmxt_textrmub): Likewise.
7131 (iwmmxt_textrmsb): Likewise.
7132 (iwmmxt_textrmuh): Likewise.
7133 (iwmmxt_textrmsh): Likewise.
7134 (iwmmxt_textrmw): Likewise.
7135 (iwmxxt_wshufh): Likewise.
7136 (eqv8qi3): Likewise.
7137 (eqv4hi3): Likewise.
7138 (eqv2si3): Likewise.
7139 (gtuv8qi3): Likewise.
7140 (gtuv4hi3): Likewise.
7141 (gtuv2si3): Likewise.
7142 (gtv8qi3): Likewise.
7143 (gtv4hi3): Likewise.
7144 (gtv2si3): Likewise.
7145 (smax<mode>3_iwmmxt): Likewise.
7146 (umax<mode>3_iwmmxt): Likewise.
7147 (smin<mode>3_iwmmxt): Likewise.
7148 (umin<mode>3_iwmmxt): Likewise.
7149 (iwmmxt_wpackhss): Likewise.
7150 (iwmmxt_wpackwss): Likewise.
7151 (iwmmxt_wpackdss): Likewise.
7152 (iwmmxt_wpackhus): Likewise.
7153 (iwmmxt_wpackwus): Likewise.
7154 (iwmmxt_wpackdus): Likewise.
7155 (iwmmxt_wunpckihb): Likewise.
7156 (iwmmxt_wunpckihh): Likewise.
7157 (iwmmxt_wunpckihw): Likewise.
7158 (iwmmxt_wunpckilb): Likewise.
7159 (iwmmxt_wunpckilh): Likewise.
7160 (iwmmxt_wunpckilw): Likewise.
7161 (iwmmxt_wunpckehub): Likewise.
7162 (iwmmxt_wunpckehuh): Likewise.
7163 (iwmmxt_wunpckehuw): Likewise.
7164 (iwmmxt_wunpckehsb): Likewise.
7165 (iwmmxt_wunpckehsh): Likewise.
7166 (iwmmxt_wunpckehsw): Likewise.
7167 (iwmmxt_wunpckelub): Likewise.
7168 (iwmmxt_wunpckeluh): Likewise.
7169 (iwmmxt_wunpckeluw): Likewise.
7170 (iwmmxt_wunpckelsb): Likewise.
7171 (iwmmxt_wunpckelsh): Likewise.
7172 (iwmmxt_wunpckelsw): Likewise.
7173 (ror<mode>3): Likewise.
7174 (ashr<mode>3_iwmmxt): Likewise.
7175 (lshr<mode>3_iwmmxt): Likewise.
7176 (ashl<mode>3_iwmmxt): Likewise.
7177 (ror<mode>3_di): Likewise.
7178 (ashr<mode>3_di): Likewise.
7179 (lshr<mode>3_di): Likewise.
7180 (ashl<mode>3_di): Likewise.
7181 (iwmmxt_wmadds): Likewise.
7182 (iwmmxt_wmaddu): Likewise.
7183 (iwmmxt_tmia): Likewise.
7184 (iwmmxt_tmiaph): Likewise.
7185 (iwmmxt_tmiabb): Likewise.
7186 (iwmmxt_tmiatb): Likewise.
7187 (iwmmxt_tmiabt): Likewise.
7188 (iwmmxt_tmiatt): Likewise.
7189 (iwmmxt_tmovmskb): Likewise.
7190 (iwmmxt_tmovmskh): Likewise.
7191 (iwmmxt_tmovmskw): Likewise.
7192 (iwmmxt_waccb): Likewise.
7193 (iwmmxt_wacch): Likewise.
7194 (iwmmxt_waccw): Likewise.
7195 (iwmmxt_waligni): Likewise.
7196 (iwmmxt_walignr): Likewise.
7197 (iwmmxt_walignr0): Likewise.
7198 (iwmmxt_walignr1): Likewise.
7199 (iwmmxt_walignr2): Likewise.
7200 (iwmmxt_walignr3): Likewise.
7201 (iwmmxt_wsadb): Likewise.
7202 (iwmmxt_wsadh): Likewise.
7203 (iwmmxt_wsadbz): Likewise.
7204 (iwmmxt_wsadhz): Likewise.
7205 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7206 (iwmmxt_wabsdiffb): Likewise.
7207 (iwmmxt_wabsdiffh): Likewise.
7208 (iwmmxt_wabsdiffw): Likewise.
7209 (iwmmxt_waddsubhx): Likewise
7210 (iwmmxt_wsubaddhx): Likewise.
7211 (addc<mode>3): Likewise.
7212 (iwmmxt_avg4): Likewise.
7213 (iwmmxt_avg4r): Likewise.
7214 (iwmmxt_wmaddsx): Likewise.
7215 (iwmmxt_wmaddux): Likewise.
7216 (iwmmxt_wmaddsn): Likewise.
7217 (iwmmxt_wmaddun): Likewise.
7218 (iwmmxt_wmulwsm): Likewise.
7219 (iwmmxt_wmulwum): Likewise.
7220 (iwmmxt_wmulsmr): Likewise.
7221 (iwmmxt_wmulumr): Likewise.
7222 (iwmmxt_wmulwsmr): Likewise.
7223 (iwmmxt_wmulwumr): Likewise.
7224 (iwmmxt_wmulwl): Likewise.
7225 (iwmmxt_wqmulm): Likewise.
7226 (iwmmxt_wqmulwm): Likewise.
7227 (iwmmxt_wqmulmr): Likewise.
7228 (iwmmxt_wqmulwmr): Likewise.
7229 (iwmmxt_waddbhusm): Likewise.
7230 (iwmmxt_waddbhusl): Likewise.
7231 (iwmmxt_wqmiabb): Likewise.
7232 (iwmmxt_wqmiabt): Likewise.
7233 (iwmmxt_wqmiatb): Likewise.
7234 (iwmmxt_wqmiatt): Likewise.
7235 (iwmmxt_wqmiabbn): Likewise.
7236 (iwmmxt_wqmiabtn): Likewise.
7237 (iwmmxt_wqmiatbn): Likewise.
7238 (iwmmxt_wqmiattn): Likewise.
7239 (iwmmxt_wmiabb): Likewise.
7240 (iwmmxt_wmiabt): Likewise.
7241 (iwmmxt_wmiatb): Likewise.
7242 (iwmmxt_wmiatt): Likewise.
7243 (iwmmxt_wmiabbn): Likewise.
7244 (iwmmxt_wmiabtn): Likewise.
7245 (iwmmxt_wmiatbn): Likewise.
7246 (iwmmxt_wmiattn): Likewise.
7247 (iwmmxt_wmiawbb): Likewise.
7248 (iwmmxt_wmiawbt): Likewise.
7249 (iwmmxt_wmiawtb): Likewise.
7250 (iwmmxt_wmiawtt): Likewise.
7251 (iwmmxt_wmiawbbn): Likewise.
7252 (iwmmxt_wmiawbtn): Likewise.
7253 (iwmmxt_wmiawtbn): Likewise.
7254 (iwmmxt_wmiawttn): Likewise.
7255 (iwmmxt_wmerge): Likewise.
7256 (iwmmxt_tandc<mode>3): Likewise.
7257 (iwmmxt_torc<mode>3): Likewise.
7258 (iwmmxt_torvsc<mode>3): Likewise.
7259 (iwmmxt_textrc<mode>3): Likewise.
7260 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7261 (wmmxt_pack): Likewise.
7262 (wmmxt_mult_c1): Likewise.
7263 (wmmxt_mult_c2): Likewise.
7264 (wmmxt_alu_c1): Likewise.
7265 (wmmxt_alu_c2): Likewise.
7266 (wmmxt_alu_c3): Likewise.
7267 (wmmxt_transfer_c1): Likewise.
7268 (wmmxt_transfer_c2): Likewise.
7269 (wmmxt_transfer_c3): Likewise.
7270 (marvell_f_iwmmxt_wstr): Likewise.
7271 (marvell_f_iwmmxt_wldr): Likewise.
7273 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7275 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
7277 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7280 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7281 * lra-constraints.c (need_for_split_p): Check call used hard regs
7282 living through calls.
7284 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7285 call used regs for call insn.
7287 2013-06-28 Jakub Jelinek <jakub@redhat.com>
7290 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7291 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
7293 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7295 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
7297 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7299 * lra-constraints.c (need_for_split_p): Check call used hard regs
7300 living through calls.
7302 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7305 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7306 to tie with any other modes. Eliminate Altivec vector mode tests,
7307 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7308 code, to return 0 if testing MODE2 for a condition, if we've
7309 already tested MODE1 for the same condition.
7311 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7313 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7316 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7318 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7319 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7321 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7323 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7325 (aarch64_symbolic_constant_p): Remove.
7326 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7327 static. Fix line length and white space.
7328 (aarch64_symbolic_constant_p): Remove.
7329 * config/aarch64/predicates.md (aarch64_valid_symref):
7330 Use aarch64_classify_symbol_expression.
7332 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7334 * config/arm/constraints.md (Ts): New constraint.
7335 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7337 (compare_scc): Use "Ts" constraint for operand 0.
7338 (ior_scc_scc): Likewise.
7339 (and_scc_scc): Likewise.
7340 (and_scc_scc_nodom): Likewise.
7341 (ior_scc_scc_cmp): Likewise for operand 7.
7342 (and_scc_scc_cmp): Likewise.
7343 * config/arm/thumb2.md (thumb2_movsi_insn):
7344 Add alternatives for 16-bit encodings.
7345 (thumb2_movhi_insn): Likewise.
7346 (thumb2_movsicc_insn): Likewise.
7347 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7348 (thumb2_negscc): Use "Ts" constraint.
7349 Move mvn instruction outside cond_exec block.
7350 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7351 for 16-bit encodings.
7353 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7355 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7357 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7358 (mulsi3subsi): Likewise.
7359 (mulsidi3adddi): Likewise.
7360 (mulsidi3_v6): Likewise.
7361 (umulsidi3_v6): Likewise.
7362 (umulsidi3adddi_v6): Likewise.
7363 (smulsi3_highpart_v6): Likewise.
7364 (umulsi3_highpart_v6): Likewise.
7365 (mulhisi3tb): Likewise.
7366 (mulhisi3bt): Likewise.
7367 (mulhisi3tt): Likewise.
7368 (maddhisi4): Likewise.
7369 (maddhisi4tb): Likewise.
7370 (maddhisi4tt): Likewise.
7371 (maddhidi4): Likewise.
7372 (maddhidi4tb): Likewise.
7373 (maddhidi4tt): Likewise.
7374 (zeroextractsi_compare0_scratch): Likewise.
7375 (insv_zero): Likewise.
7376 (insv_t2): Likewise.
7377 (anddi_notzesidi_di): Likewise.
7378 (anddi_notsesidi_di): Likewise.
7379 (andsi_notsi_si): Likewise.
7380 (iordi_zesidi_di): Likewise.
7381 (xordi_zesidi_di): Likewise.
7382 (andsi_iorsi3_notsi): Likewise.
7384 (smax_m1): Likewise.
7386 (not_shiftsi): Likewise.
7387 (unaligned_loadsi): Likewise.
7388 (unaligned_loadhis): Likewise.
7389 (unaligned_loadhiu): Likewise.
7390 (unaligned_storesi): Likewise.
7391 (unaligned_storehi): Likewise.
7392 (extv_reg): Likewise.
7393 (extzv_t2): Likewise.
7395 (udivsi3): Likewise.
7396 (arm_zero_extendhisi2addsi): Likewise.
7397 (arm_zero_extendqisi2addsi): Likewise.
7398 (compareqi_eq0): Likewise.
7399 (arm_extendhisi2_v6): Likewise.
7400 (arm_extendqisi2addsi): Likewise.
7401 (arm_movt): Likewise.
7402 (thumb2_ldrd): Likewise.
7403 (thumb2_ldrd_base): Likewise.
7404 (thumb2_ldrd_base_neg): Likewise.
7405 (thumb2_strd): Likewise.
7406 (thumb2_strd_base): Likewise.
7407 (thumb2_strd_base_neg): Likewise.
7408 (arm_negsi2): Add alternative for 16-bit encoding.
7409 (arm_one_cmplsi2): Likewise.
7411 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7413 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
7414 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
7415 (movdfcc): Likewise.
7416 * config/arm/vfp.md (*thumb2_movsf_vfp):
7417 Disable predication for arm_restrict_it.
7418 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
7419 (*thumb2_movdfcc_vfp): Likewise.
7420 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
7421 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
7422 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
7423 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
7424 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
7425 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
7426 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
7427 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
7428 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
7429 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
7430 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
7431 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
7432 Disable predication for arm_restrict_it.
7434 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
7436 * config/i386/bmiintrin.h (_bextr_u32): New.
7437 (_bextr_u64): Ditto.
7439 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
7441 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
7442 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
7443 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
7444 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
7445 * config/mips/n32-elf.h: ...this new file.
7447 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
7450 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
7451 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
7453 2013-06-27 Catherine Moore <clm@codesourcery.com>
7455 * config/mips/mips-tables.opt: Regenerate.
7456 * config/mips/mips-cpus.def: Add m14ke and m14kec.
7457 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
7458 * doc/invoke.texi: Add -m14kc.
7460 2013-06-27 Jakub Jelinek <jakub@redhat.com>
7463 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
7464 constraints of operand 1 and 2.
7467 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
7468 to match RTL canonicalization. Swap predicates and
7469 constraints of operand 1 and 2.
7471 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
7473 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
7474 Process OP_INOUT regs for splitting too.
7476 2013-06-27 Jakub Jelinek <jakub@redhat.com>
7478 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
7479 decl before the loop, initialize to NULL.
7480 (vectorizable_load): Initialize ptr_incr to NULL.
7482 2013-06-27 Martin Jambor <mjambor@suse.cz>
7485 * ipa-ref.h (ipa_maybe_record_reference): Declare.
7486 * ipa-ref.c (ipa_maybe_record_reference): New function.
7487 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
7488 * ipa-cp.c (create_specialized_node): Record potential references from
7490 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
7492 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7494 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
7495 parameter 'mode' of type 'enum machine_mode mode'; change to pass
7496 'mode' to force_reg.
7497 (aarch64_add_offset): Update calls to aarch64_force_temporary.
7498 (aarch64_expand_mov_immediate): Likewise.
7500 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7502 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
7503 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
7505 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7507 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
7508 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7509 (struct machine_function): Add tbegin_p.
7510 (s390_canonicalize_comparison): Fold CC mode compares to
7511 conditional jump if possible.
7512 (s390_emit_jump): Return the emitted jump.
7513 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
7514 Handle CCRAWmode compares.
7515 (s390_option_override): Default to -mhtm if available.
7516 (s390_reg_clobbered_rtx): Handle floating point regs as well.
7517 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
7518 FPRs instead of df_regs_ever_live_p.
7519 (s390_optimize_nonescaping_tx): New function.
7520 (s390_init_frame_layout): Extend clobbered_regs array to cover
7522 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
7523 (s390_expand_tbegin): New function.
7524 (enum s390_builtin): New enum definition.
7525 (code_for_builtin): New array definition.
7526 (s390_init_builtins): New function.
7527 (s390_expand_builtin): New function.
7528 (TARGET_INIT_BUILTINS): Define.
7529 (TARGET_EXPAND_BUILTIN): Define.
7530 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
7531 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
7532 (s390_alc_comparison): Likewise.
7533 * config/s390/s390-modes.def: Add CCRAWmode.
7534 * config/s390/s390.h (processor_flags): Add PF_TX.
7535 (TARGET_CPU_HTM): Define macro.
7536 (TARGET_HTM): Define macro.
7537 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
7538 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
7539 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7540 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
7541 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
7542 (TBEGIN_MASK, TBEGINC_MASK): New constants.
7543 ("*cc_to_int"): Move up.
7544 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
7545 constants other than 0.
7546 ("*ccraw_to_int"): New insn and splitter definition.
7547 ("tbegin", "tbegin_nofloat", "tbegin_retry")
7548 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
7549 ("tx_assist"): New expander.
7550 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
7551 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
7552 * config/s390/s390.opt: Add -mhtm option.
7553 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
7554 * config/s390/htmxlintrin.h: New file.
7555 * config/s390/htmintrin.h: New file.
7556 * config/s390/s390intrin.h: New file.
7557 * doc/extend.texi: Document htm builtins.
7558 * config.gcc: Add the new header files to extra_headers.
7560 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
7562 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
7563 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
7565 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
7566 Pat Haugen <pthaugen@us.ibm.com>
7567 Peter Bergner <bergner@vnet.ibm.com>
7569 * config/rs6000/power8.md: New.
7570 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
7571 setting for power8 entry.
7572 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
7573 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
7574 test for Power4/Power5 only.
7575 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
7577 (force_new_group): Adjust comment.
7578 * config/rs6000/rs6000.md: Include power8.md.
7580 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
7582 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
7583 * config/arm/arm-protos.h (arm_max_conditional_execute): New
7585 (tune_params): Update comment.
7586 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
7587 (arm_max_conditional_execute): New function.
7588 (thumb2_final_prescan_insn): Use max_insn_skipped and
7589 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
7591 2013-06-25 Jakub Jelinek <jakub@redhat.com>
7593 PR tree-optimization/57705
7594 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
7595 SSA_NAME step, provided that it is not defined inside the loop.
7596 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
7597 (get_initial_def_for_induction): Handle SSA_NAME IV step.
7599 2013-06-25 Martin Jambor <mjambor@suse.cz>
7602 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
7603 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
7605 (ipa_note_param_call): Initialize member_ptr flag.
7606 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
7607 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
7608 (ipa_write_indirect_edge_info): Stream member_ptr flag.
7609 (ipa_read_indirect_edge_info): Likewise.
7611 2013-06-25 Richard Biener <rguenther@suse.de>
7614 * passes.c (init_optimization_passes): Move pass_fold_builtins
7615 and pass_dce earlier with -Og.
7617 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
7619 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
7620 <BIT_FIELD_REF>: Remove trailing TAB.
7621 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
7624 2013-06-24 Martin Jambor <mjambor@suse.cz>
7626 PR tree-optimization/57358
7627 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
7628 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
7629 (ipa_analyze_params_uses): Generate pessimistic info when true.
7631 2013-06-24 Martin Jambor <mjambor@suse.cz>
7633 PR tree-optimization/57539
7634 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
7635 global.inlined_to of the new node to it. All callers changed.
7636 * ipa-inline-transform.c (clone_inlined_nodes): New variable
7637 inlining_into, pass it to cgraph_clone_node.
7638 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
7639 ipa_free_edge_args_substructures.
7640 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
7641 rdesc linked list. Do not assert rdesc edges have inlined caller.
7642 Assert we have found an rdesc in the rdesc list.
7644 2013-06-24 Richard Biener <rguenther@suse.de>
7646 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7647 (pointer_set_lookup): Declare.
7648 (class pointer_map): New template class implementing a
7649 generic pointer to T map.
7650 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
7651 pointer_map<T>::contains, pointer_map<T>::insert,
7652 pointer_map<T>::traverse): New functions.
7653 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
7654 (pointer_set_lookup): New function.
7655 (pointer_set_contains): Use pointer_set_lookup.
7656 (pointer_set_insert): Likewise.
7657 (insert_aux): Remove.
7658 (struct pointer_map_t): Embed a pointer_set_t.
7659 (pointer_map_create): Adjust.
7660 (pointer_map_destroy): Likewise.
7661 (pointer_map_contains): Likewise.
7662 (pointer_map_insert): Likewise.
7663 (pointer_map_traverse): Likewise.
7664 * tree-streamer.h (struct streamer_tree_cache_d): Use a
7665 pointer_map<unsigned> instead of a pointer_map_t.
7666 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
7667 (streamer_tree_cache_lookup): Likewise.
7668 (streamer_tree_cache_create): Likewise.
7669 (streamer_tree_cache_delete): Likewise.
7670 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
7671 pointer_map<unsigned> instead of a pointer_map_t.
7672 (lto_init_tree_ref_encoder): Adjust.
7673 (lto_destroy_tree_ref_encoder): Likewise.
7674 * lto-section-out.c (lto_output_decl_index): Likewise.
7675 (lto_record_function_out_decl_state): Likewise.
7676 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
7678 2013-06-24 Richard Biener <rguenther@suse.de>
7680 PR tree-optimization/57488
7681 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
7683 2013-06-24 Alan Modra <amodra@gmail.com>
7685 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
7686 (gen_easy_altivec_constant): Likewise.
7687 * config/rs6000/predicates.md (easy_vector_constant_add_self,
7688 easy_vector_constant_msb): Likewise.
7690 2013-06-23 Jakub Jelinek <jakub@redhat.com>
7693 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
7694 add missing return true.
7696 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
7699 * config/sh/predicates.md (general_extend_operand): Invoke
7700 general_movsrc_operand for memory operands.
7701 (general_movsrc_operand): Allow reg+reg addressing, do not use
7702 general_operand for memory operands.
7704 2013-06-23 Sriraman Tallam <tmsriram@google.com>
7706 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
7707 when current target options does not apply.
7708 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
7709 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
7710 * config/i386/bmiintrin.h: Pass appropriate target
7711 attributes to header.
7712 * config/i386/mmintrin.h: Ditto.
7713 * config/i386/nmmintrin.h: Ditto.
7714 * config/i386/avx2intrin.h: Ditto.
7715 * config/i386/fxsrintrin.h: Ditto.
7716 * config/i386/tbmintrin.h: Ditto.
7717 * config/i386/xsaveintrin.h: Ditto.
7718 * config/i386/f16cintrin.h: Ditto.
7719 * config/i386/xtestintrin.h: Ditto.
7720 * config/i386/xsaveoptintrin.h: Ditto.
7721 * config/i386/bmi2intrin.h: Ditto.
7722 * config/i386/lzcntintrin.h: Ditto.
7723 * config/i386/smmintrin.h: Ditto.
7724 * config/i386/wmmintrin.h: Ditto.
7725 * config/i386/x86intrin.h: Remove all header include guards.
7726 * config/i386/prfchwintrin.h: Ditto.
7727 * config/i386/pmmintrin.h: Ditto.
7728 * config/i386/tmmintrin.h: Ditto.
7729 * config/i386/xmmintrin.h: Ditto.
7730 * config/i386/popcntintrin.h: Ditto.
7731 * config/i386/rdseedintrin.h: Ditto.
7732 * config/i386/ammintrin.h: Ditto.
7733 * config/i386/emmintrin.h: Ditto.
7734 * config/i386/immintrin.h: Remove all header include guards.
7735 * config/i386/fma4intrin.h: Ditto.
7736 * config/i386/lwpintrin.h: Ditto.
7737 * config/i386/xopintrin.h: Ditto.
7738 * config/i386/ia32intrin.h: Ditto.
7739 * config/i386/avxintrin.h: Ditto.
7740 * config/i386/rtmintrin.h: Ditto.
7741 * config/i386/fmaintrin.h: Ditto.
7742 * config/i386/mm3dnow.h: Ditto.
7744 2013-06-22 Sriraman Tallam <tmsriram@google.com>
7746 * common/config/i386/i386-common.c: Handle LZCNT.
7748 2013-06-22 Andi Kleen <ak@linux.intel.com>
7750 * doc/extend.texi: Use __atomic_store_n instead of
7751 __atomic_store in HLE example.
7753 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
7755 * config/sh/sh.c: Remove <cstdlib> workaround.
7757 2013-06-21 Andi Kleen <ak@linux.intel.com>
7759 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
7761 2013-06-21 Andi Kleen <ak@linux.intel.com>
7763 * doc/extend.texi: Document that __atomic_clear and
7764 __atomic_test_and_set should only be used with bool.
7766 2013-06-20 Jan Hubicka <jh@suse.cz>
7768 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
7770 * tree.c (decls_same_for_odr): New function.
7771 (same_for_edr): New function.
7772 (types_same_for_odr): New function.
7773 (get_binfo_at_offset): Use it.
7774 * tree.h (types_same_for_odr): Declare.
7776 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
7777 Jason Merrill <jason@redhat.com>
7779 * system.h: Include <cstdlib> as well as <stdlib.h>.
7781 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
7784 * config/i386/i386.c (construct_container): Report error if
7785 long double is used with disabled x87 float returns.
7787 2013-06-20 Jan Hubicka <jh@suse.cz>
7789 * lto-cgraph.c (input_symtab): Do not set cgraph state.
7791 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
7793 PR rtl-optimization/57425
7794 PR rtl-optimization/57569
7795 * alias.c (write_dependence_p): Remove parameters mem_mode and
7796 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
7797 Changed all callers.
7798 (canon_anti_dependence): Get comments and semantics in sync.
7799 Add parameter mem_canonicalized. Changed all callers.
7800 * rtl.h (canon_anti_dependence): Update prototype.
7802 2013-06-20 Richard Biener <rguenther@suse.de>
7804 * data-streamer-in.c (streamer_read_uhwi): Optimize single
7805 byte case, inline streamer_read_uchar and defer section
7808 2013-06-20 Richard Biener <rguenther@suse.de>
7810 PR tree-optimization/57584
7811 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
7812 SSA names into the expanded expression that take part in
7813 abnormal coalescing.
7815 2013-06-19 Sharad Singhai <singhai@google.com>
7817 * gcov.c (print_usage): Handle new option.
7818 (process_args): Ditto.
7819 (get_gcov_intermediate_filename): New function.
7820 (output_intermediate_file): New function.
7821 (output_gcov_file): New function
7822 (generate_results): Handle new option.
7823 (release_function): Relase demangled name.
7824 (read_graph_file): Handle demangled name.
7825 (output_lines): Ditto.
7826 * doc/gcov.texi: Document gcov intermediate format.
7828 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
7831 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
7832 (lra_emit_add): Use the functions. Add comment about Y as an
7835 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
7838 * collect2.c (collect_atexit): New.
7839 (collect_exit): Delete.
7840 (main): Register collect_atexit with atexit.
7841 (collect_wait): Change collect_exit to exit.
7843 * collect2.h (collect_exit): Delete.
7844 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
7846 2013-06-19 Wei Mi <wmi@google.com>
7848 PR rtl-optimization/57518
7849 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
7850 if regno is used in paradoxical subreg.
7851 (update_equiv_regs): Check pdx_subregs[regno] before
7852 set a reg to be equivalent with a mem.
7854 2013-06-19 Matthias Klose <doko@ubuntu.com>
7857 * file-find.h (find_a_file): Add a mode parameter.
7858 * file-find.c (find_a_file): Likewise.
7859 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
7860 with X_OK for the executables.
7861 * collect2.c (main): Call find_a_file with X_OK.
7863 2013-06-19 Steve Ellcey <sellcey@mips.com>
7866 * config/mips/mips.md (casesi_internal_mips16_<mode>):
7867 Use NEXT_INSN instead of next_real_insn.
7869 2013-06-19 Jan Hubicka <jh@suse.cz>
7871 * cgraph.h (const_value_known_p): Replace by ...
7872 (ctor_for_folding): .. this one.
7873 * cgraphunit.c (process_function_and_variable_attributes): Use it.
7874 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
7875 * expr.c (expand_expr_real_1): Likewise.
7876 (string_constant): Likewise.
7877 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
7878 * ipa.c (process_references): Likewise.
7879 (symtab_remove_unreachable_nodes): Likewise.
7880 * ipa-inline-analysis.c (param_change_prob): Likewise.
7881 * gimple-fold.c (canonicalize_constructor_val): Likewise.
7882 (get_base_constructor): Likwise.
7883 * varpool.c (varpool_remove_node): Likewise.
7884 (varpool_remove_initializer): LIkewise.
7885 (dump_varpool_node): LIkwise.
7886 (const_value_known_p): Rewrite to ...
7887 (ctor_for_folding): ... this one.
7889 2013-06-19 Jakub Jelinek <jakub@redhat.com>
7892 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
7893 PERSONALITY in $PATH derived prefixes.
7895 2013-06-19 Jeff Law <law@redhat.com>
7897 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
7900 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
7901 (simplify_bitwise_binary): Use it to simpify certain binary ops on
7904 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
7906 * config/arm/vfp.md: Move VFP instruction classification documentation
7908 * config/arm/arm.md: ... here. Update instruction classification
7911 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
7913 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
7914 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
7915 pattern. Use more efficient sequences on ARMv5 and Thumb2.
7917 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
7920 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7921 with NEXT_INSN. Use tablejump_p to check for jump table data
7924 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
7927 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
7928 that now in C++ the value is correct per the C++ standards.
7930 2013-06-19 Richard Biener <rguenther@suse.de>
7932 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
7935 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7938 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7941 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7942 with next_active_insn.
7944 2013-06-18 Sriraman Tallam <tmsriram@google.com>
7946 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
7947 functions are inlined during failures to flag an error.
7948 * tree-inline.c (expand_call_inline): Allow the error to be flagged
7949 in early inline pass.
7951 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
7953 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
7956 2013-06-18 Julian Brown <julian@codesourcery.com>
7958 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
7959 Permit virtual register pre-reload if !strict.
7960 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
7962 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
7964 * config/arm/neon.md (movmisalign<mode>): Use
7965 neon_perm_struct_or_reg_operand instead of
7966 neon_struct_or_register_operand.
7967 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
7968 neon_permissive_struct_operand instead of neon_struct_operand.
7969 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
7970 neon_vector_mem_operand.
7971 * config/arm/predicates.md (neon_struct_operand): Adjust call to
7972 neon_vector_mem_operand.
7973 (neon_permissive_struct_operand): New.
7974 (neon_struct_or_register_operand): Rename to...
7975 (neon_perm_struct_or_reg_operand): This. Adjust call to
7976 neon_vector_mem_operand.
7978 2013-06-18 Richard Biener <rguenther@suse.de>
7980 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
7981 * lto-streamer.h: Include pointer-set.h.
7982 (struct lto_decl_slot): Remove.
7983 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
7984 Remove next_index entry.
7985 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7986 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7987 (lto_init_tree_ref_encoder): Adjust.
7988 (lto_destroy_tree_ref_encoder): Likewise.
7989 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7990 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7991 (lto_output_decl_index): Adjust.
7992 (lto_new_out_decl_state): Likewise.
7993 (lto_record_function_out_decl_state): Likewise.
7994 * lto-streamer-out.c (copy_function): Likewise.
7996 2013-06-18 Richard Biener <rguenther@suse.de>
7998 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
7999 * cgraphunit.c: Include cfgloop.h.
8000 (init_lowered_empty_function): Initialize the loop tree.
8001 (assemble_thunk): Insert new BBs into loops.
8003 2013-06-18 Richard Biener <rguenther@suse.de>
8005 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8006 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8007 the map from cache entry to cache index optional.
8008 (streamer_tree_cache_replace_tree): Adjust accordingly.
8009 (streamer_tree_cache_append): Likewise.
8010 (streamer_tree_cache_delete): Likewise.
8011 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8012 streamer cache map from cache entry to cache index.
8013 * lto-streamer-out.c (create_output_block): Adjust.
8015 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8017 * config/arm/arm.md (attribute "insn"): Move multiplication and
8018 division attributes to...
8019 (attribute "type"): ... here. Remove mult.
8020 (attribute "mul32"): New attribute.
8021 (attribute "mul64"): Add umaal.
8022 (*arm_mulsi3): Update attributes.
8023 (*arm_mulsi3_v6): Likewise.
8024 (*thumb_mulsi3): Likewise.
8025 (*thumb_mulsi3_v6): Likewise.
8026 (*mulsi3_compare0): Likewise.
8027 (*mulsi3_compare0_v6): Likewise.
8028 (*mulsi_compare0_scratch): Likewise.
8029 (*mulsi_compare0_scratch_v6): Likewise.
8030 (*mulsi3addsi): Likewise.
8031 (*mulsi3addsi_v6): Likewise.
8032 (*mulsi3addsi_compare0): Likewise.
8033 (*mulsi3addsi_compare0_v6): Likewise.
8034 (*mulsi3addsi_compare0_scratch): Likewise.
8035 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8036 (*mulsi3subsi): Likewise.
8037 (*mulsidi3adddi): Likewise.
8038 (*mulsi3addsi_v6): Likewise.
8039 (*mulsidi3adddi_v6): Likewise.
8040 (*mulsidi3_nov6): Likewise.
8041 (*mulsidi3_v6): Likewise.
8042 (*umulsidi3_nov6): Likewise.
8043 (*umulsidi3_v6): Likewise.
8044 (*umulsidi3adddi): Likewise.
8045 (*umulsidi3adddi_v6): Likewise.
8046 (*smulsi3_highpart_nov6): Likewise.
8047 (*smulsi3_highpart_v6): Likewise.
8048 (*umulsi3_highpart_nov6): Likewise.
8049 (*umulsi3_highpart_v6): Likewise.
8050 (mulhisi3): Likewise.
8051 (*mulhisi3tb): Likewise.
8052 (*mulhisi3bt): Likewise.
8053 (*mulhisi3tt): Likewise.
8054 (maddhisi4): Likewise.
8055 (*maddhisi4tb): Likewise.
8056 (*maddhisi4tt): Likewise.
8057 (maddhidi4): Likewise.
8058 (*maddhidi4tb): Likewise.
8059 (*maddhidi4tt): Likewise.
8061 (udivsi3): Likewise.
8062 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8063 (thumb2_mulsi_short_compare0): Likewise.
8064 (thumb2_mulsi_short_compare0_scratch): Likewise.
8065 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8066 (1020mult2): Likewise.
8067 (1020mult3): Likewise.
8068 (1020mult4): Likewise.
8069 (1020mult5): Likewise.
8070 (1020mult6): Likewise.
8071 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8073 (cortex_a15_mult64): Likewise.
8074 (cortex_a15_sdiv): Likewise.
8075 (cortex_a15_udiv): Likewise.
8076 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8082 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8083 (pj4_ir_div): Likewise.
8084 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8085 (11_mult2): Likewise.
8086 (11_mult3): Likewise.
8087 (11_mult4): Likewise.
8088 (11_mult5): Likewise.
8089 (11_mult6): Likewise.
8090 (11_mult7): Likewise.
8091 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8092 (cortex_a8_mla): Likewise.
8093 (cortex_a8_mull): Likewise.
8094 (cortex_a8_smulwy): Likewise.
8095 (cortex_a8_smlald): Likewise.
8096 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8097 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8098 (cortex_r4_mul_3): Likewise.
8099 (cortex_r4_mla_4): Likewise.
8100 (cortex_r4_mla_3): Likewise.
8101 (cortex_r4_smlald): Likewise.
8102 (cortex_r4_mull): Likewise.
8103 (cortex_r4_sdiv): Likewise.
8104 (cortex_r4_udiv): Likewise.
8105 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8106 (cortex_a7_idiv): Likewise.
8107 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8108 (9_mult2): Likewise.
8109 (9_mult3): Likewise.
8110 (9_mult4): Likewise.
8111 (9_mult5): Likewise.
8112 (9_mult6): Likewise.
8113 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8114 (cortex_a53_sdiv): Likewise.
8115 (cortex_a53_udiv): Likewise.
8116 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8117 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8118 (mp626_mult2): Likewise.
8119 (mp626_mult3): Likewise.
8120 (mp626_mult4): Likewise.
8121 * config/arm/fa526.md (526_mult1): Update attribute change.
8122 (526_mult2): Likewise.
8123 * config/arm/arm-generic.md (mult): Update attribute change.
8124 (mult_ldsched_strongarm): Likewise.
8125 (mult_ldsched): Likewise.
8126 (multi_cycle): Likewise.
8127 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8128 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8129 (606te_mult2): Likewise.
8130 (606te_mult3): Likewise.
8131 (606te_mult4): Likewise.
8132 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8133 (cortex_a9_mac16): Likewise.
8134 (cortex_a9_multiply): Likewise.
8135 (cortex_a9_mac): Likewise.
8136 (cortex_a9_multiply_long): Likewise.
8137 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8138 (626te_mult2): Likewise.
8139 (626te_mult3): Likewise.
8140 (626te_mult4): Likewise.
8142 2013-06-18 Richard Biener <rguenther@suse.de>
8145 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8147 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8150 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8151 with next_active_insn.
8153 2013-06-18 Alan Modra <amodra@gmail.com>
8155 * config/rs6000/rs6000.h (enum data_align): New.
8156 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8157 (DATA_ABI_ALIGNMENT): Define.
8158 (CONSTANT_ALIGNMENT): Correct comment.
8159 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8160 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8162 2013-06-17 David Malcolm <dmalcolm@redhat.com>
8164 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8165 ATTRIBUTE_UNUSED marking.
8167 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8169 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8170 alternative and update.
8171 (aarch64_dup_lanedi): Delete.
8172 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8173 * config/aarch64/aarch64-simd-builtins.def: Update.
8175 2013-06-17 Richard Biener <rguenther@suse.de>
8177 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8178 (lto_input_scc): Declare.
8179 (lto_input_tree_1): Likewise.
8180 (struct lto_stats_d): Add num_tree_bodies_output and
8181 num_pickle_refs_output.
8182 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8183 (lto_read_tree_1): Split out from ...
8184 (lto_read_tree): ... this.
8185 (lto_input_scc): New function.
8186 (lto_input_tree_1): Split out from ...
8187 (lto_input_tree): ... this. Handle LTO_tree_scc.
8188 (lto_data_in_create): Create the streamer cache without hashes.
8189 * lto-streamer-out.c (create_output_block): Create the streamer
8190 cache with hashes when not doing WPA.
8191 (lto_write_tree_1): Split out from ...
8192 (lto_write_tree): ... this.
8193 (get_symbol_initial_value): New function.
8194 (lto_output_tree_1): Split out from ...
8195 (lto_output_tree): ... this. Write trees as series of SCCs
8196 using a DFS walk via DFS_write_tree.
8197 (struct sccs, struct scc_entry): New types.
8198 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8199 (DFS_write_tree_body): New function.
8200 (DFS_write_tree): Likewise.
8201 (hash_tree): Likewise.
8202 (scc_entry_compare): Likewise.
8203 (hash_scc): Likewise.
8204 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8205 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8206 TREE_CHAIN as regular reference.
8207 (streamer_read_integer_cst): Remove.
8208 (streamer_get_pickled_tree): Adjust.
8209 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8210 of DECL_EXTERNALs in BLOCK_VARS for now.
8211 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8213 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8214 Add hash value argument and record that if hashes are recorded
8216 (streamer_tree_cache_insert_1): Adjust.
8217 (streamer_tree_cache_insert): Likewise.
8218 (streamer_tree_cache_insert_at): Rename to ...
8219 (streamer_tree_cache_replace_tree): ... this and adjust.
8220 (streamer_tree_cache_append): Adjust.
8221 (record_common_node): Likewise.
8222 (streamer_tree_cache_create): Add argument whether to
8223 record hash values together with trees.
8224 (streamer_tree_cache_delete): Adjust.
8225 * tree-streamer.h (struct streamer_tree_cache_d): Add
8227 (streamer_read_integer_cst): Remove.
8228 (streamer_tree_cache_insert): Adjust.
8229 (streamer_tree_cache_append): Likewise.
8230 (streamer_tree_cache_insert_at): Rename to ...
8231 (streamer_tree_cache_replace_tree): ... this and adjust.
8232 (streamer_tree_cache_create): Add argument whether to record hashes.
8233 (streamer_tree_cache_get): Rename to ...
8234 (streamer_tree_cache_get_tree): ... this.
8235 (streamer_tree_cache_get_hash): New function.
8236 * tree.c (cache_integer_cst): New function.
8237 * tree.h (cache_integer_cst): Declare.
8238 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8239 * lto-symtab.c (lto_varpool_replace_node): Only release
8240 DECL_INITIAL of non-prevailing decls.
8241 * varpool.c (varpool_remove_initializer): Do not release
8242 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8244 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8246 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8247 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8248 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8249 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8250 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8251 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8252 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8253 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8254 instead of TARGET_64BIT.
8255 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8256 Require ISA_HAS_<D>DIV.
8258 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8260 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8261 (mips*-*-linux*): Move default with_llsc setting to where other
8263 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8265 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8266 Likewise. Remove default with_tune setting. Move default float
8267 setting to its own block. Handle with_llsc in the same block as above.
8269 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
8271 PR rtl-optimization/57425
8272 PR rtl-optimization/57569
8273 * alias.c (write_dependence_p): Add new parameters mem_mode,
8274 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8275 Changed all callers.
8276 (canon_anti_dependence): New function.
8277 * cse.c (check_dependence): Use canon_anti_dependence.
8278 * cselib.c (cselib_invalidate_mem): Likewise.
8279 * rtl.h (canon_anti_dependence): Declare.
8281 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8283 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8284 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8285 ".set mips3" for 64-bit targets.
8287 2013-06-15 Dehao Chen <dehao@google.com>
8289 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8290 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8291 (gimple_check_call_args): Likewise.
8292 * value-prof.c (check_ic_target): Likewise.
8293 * ipa-inline.c (early_inliner): Likewise.
8294 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8295 * cgraph.c (cgraph_create_edge_1): Likewise.
8296 (cgraph_make_edge_direct): Likewise.
8298 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8301 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8302 rs6000_output_move_128bit to handle emitting quad memory
8303 operations. Set attribute length to 8 bytes.
8305 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
8307 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8309 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8310 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8311 (aarch64_<su>mlsl<mode>): Likewise.
8313 2013-06-14 Mike Stump <mikestump@comcast.net>
8315 * Makefile.in (TARGET_H): Add insn-codes.h.
8317 2013-06-14 Alan Modra <amodra@gmail.com>
8321 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8322 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8323 bitfield expansion when EXPAND_MEMORY.
8324 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8326 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8328 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8329 test for clearing quad memory on 32-bit later.
8331 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8333 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8334 (fold_negate_expr): Likewise.
8335 (fold_real_zero_addition_p): Handle vectors.
8336 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8338 2013-06-14 Alan Modra <amodra@gmail.com>
8340 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8342 2013-06-13 Jan Hubicka <jh@suse.cz>
8344 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8345 Local comdats are not externally visible.
8346 * symtab.c (dump_symtab_base): Dump externally visible.
8347 (verify_symtab_base): Verify back links in the symtab hash.
8349 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8351 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8352 CONVERT_EXPR as equal nodes.
8354 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8356 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8358 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8360 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8361 Generalize to complex and vector.
8362 * tree.c (build_all_ones_cst): New function.
8363 * tree.h (build_all_ones_cst): Declare it.
8365 2013-06-13 Alan Modra <amodra@gmail.com>
8367 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
8368 * config/rs6000/rs6000.md (signbittf2): New insn.
8369 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
8370 (abstf2_internal, cmptf_internal2): Likewise.
8371 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
8373 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8374 Pat Haugen <pthaugen@us.ibm.com>
8375 Peter Bergner <bergner@vnet.ibm.com>
8377 * config/rs6000/rs6000.c (emit_load_locked): Add support for
8378 power8 byte, half-word, and quad-word atomic instructions.
8379 (emit_store_conditional): Likewise.
8380 (rs6000_expand_atomic_compare_and_swap): Likewise.
8381 (rs6000_expand_atomic_op): Likewise.
8383 * config/rs6000/sync.md (larx): Add new modes for power8.
8385 (AINT): New mode iterator to include TImode as well as normal
8386 integer modes on power8.
8387 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
8388 that VSX registers are not considered. Use AINT mode iterator
8389 instead of INT1 to allow inclusion of quad word atomic operations
8391 (load_locked<mode>): Likewise.
8392 (store_conditional<mode>): Likewise.
8393 (atomic_compare_and_swap<mode>): Likewise.
8394 (atomic_exchange<mode>): Likewise.
8395 (atomic_nand<mode>): Likewise.
8396 (atomic_fetch_<fetchop_name><mode>): Likewise.
8397 (atomic_nand_fetch<mode>): Likewise.
8398 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
8400 (ATOMIC): On power8, add QImode, HImode modes.
8401 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
8402 modes that promote to SImode.
8403 (load_lockedti): Convert TImode arguments to PTImode, so that we
8404 get a guaranteed even/odd register pair.
8405 (load_lockedpti): Likewise.
8406 (store_conditionalti): Likewise.
8407 (store_conditionalpti): Likewise.
8409 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
8410 atomic load/store instructions.
8413 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
8415 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
8417 (insn_count): New attribute, with most cases extracted from...
8418 (length): ...here. Redefine most cases in terms of insn_count.
8419 (single_insn): Delete.
8420 (can_delay): Use insn_count to check for single instructions.
8421 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
8422 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
8423 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
8424 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
8425 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
8426 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
8427 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
8428 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
8429 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
8430 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
8431 rather than "length".
8432 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
8433 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
8434 Use "insn_count" rather than "length".
8435 * config/mips/mips-dsp.md
8436 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
8437 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
8440 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
8442 PR tree-optimization/57361
8443 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
8445 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
8447 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
8449 (aarch64_simd_combine<mode>): New instruction expansion.
8450 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
8452 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
8453 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
8455 2013-06-12 Jan Hubicka <jh@suse.cz>
8457 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
8458 decl has when in streaming stage.
8459 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
8460 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
8462 2013-06-12 Roland Stigge <stigge@antcom.de>
8465 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
8467 2013-06-12 Jakub Jelinek <jakub@redhat.com>
8469 PR tree-optimization/57537
8470 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
8471 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
8473 2013-06-12 Richard Biener <rguenther@suse.de>
8475 * data-streamer.h (streamer_write_char_stream): CSE
8476 obs->current_pointer.
8477 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
8478 streamer_write_char_stream manually and optimize the resulting loop.
8479 (streamer_write_hwi_stream): Likewise.
8481 2013-06-12 Jan Hubicka <jh@suse.cz>
8483 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
8484 * cgraph.h (varpool_create_empty_node): Declare.
8485 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
8487 * symtab.c (symtab_unregister_node): Be lax about missin entries
8489 (symtab_get_node): Update comment.
8490 * varpool.c (varpool_create_empty_node): Break out from ...
8491 (varpool_node_for_decl): ... here.
8492 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
8494 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
8496 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
8497 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
8498 part. Use straight-line flow at the end.
8499 <COMPONENT_REF>: Remove superfluous else.
8500 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
8502 2013-06-12 Jakub Jelinek <jakub@redhat.com>
8505 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
8506 target hook even for !TREE_PUBLIC decls. If no resolution info
8507 is available, return false for common and external decls.
8509 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8511 * config/rl78/constraints.md (U): New constraint.
8512 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
8515 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8518 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
8519 to allow returning address to AT_PLATFORM name.
8521 2013-06-11 Jan Hubicka <jh@suse.cz>
8523 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
8524 * cgraph.h (symtab_node_base): Add weakref flag.
8525 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
8526 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
8527 (output_weakrefs): Use weakref flag.
8528 * fold-const.c (simple_operand_p): Handle WEAK.
8529 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
8530 * ipa.c (varpool_externally_visible_p): Drop weakref.
8531 (function_and_variable_visibility): Update comment; fix weakref
8532 sanity checks; do not clear DECL_WEAK on them.
8533 * lto-cgraph.c (lto_output_node): update.
8534 (lto_output_varpool_node): Update.
8535 (input_overwrite_node): Update.
8536 (input_node): Update.
8537 (input_varpool_node): Update.
8538 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
8539 (lto_symtab_merge_symbols): Add sanity check.
8540 (lto_symtab_prevailing_decl): Do not special case weakrefs.
8541 * passes.c (rest_of_decl_compilation): Set static flag, too.
8542 * symtab.c (dump_symtab_base): Dump weakref.
8543 (verify_symtab_base): Sanity check weakrefs.
8544 (symtab_make_decl_local): Remove duplicated code.
8545 (symtab_alias_ultimate_target): Simplify.
8546 * varpool.c (varpool_create_variable_alias): Set weakref flag.
8548 2013-06-11 DJ Delorie <dj@redhat.com>
8550 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
8551 (rl78_unwind_word_mode): New.
8553 2013-06-11 David Malcolm <dmalcolm@redhat.com>
8555 * final.c (debug_prefix_maps): Make static.
8557 2013-06-11 David Malcolm <dmalcolm@redhat.com>
8559 * function.c (initial_trampoline): Remove stray copy.
8561 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
8563 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
8565 2013-06-11 Martin Jambor <mjambor@suse.cz>
8567 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
8568 within bounds at the beginning of the function.
8570 2013-06-11 Alan Modra <amodra@gmail.com>
8572 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
8574 (get_named_section): Don't NULL !DECL_P decl.
8576 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
8578 * doc/invoke.texi (core-avx2): Document.
8580 (atom): Updated with MOVBE.
8582 2013-06-11 Richard Biener <rguenther@suse.de>
8584 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
8586 2013-06-11 Anton Blanchard <anton@samba.org>
8588 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
8589 correct shift value in little-endian mode.
8591 2013-06-11 Jakub Jelinek <jakub@redhat.com>
8594 * varasm.c (get_variable_align): Move #endif to the right place.
8596 2013-06-10 Cary Coutant <ccoutant@google.com>
8598 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
8599 for hash so that hash table traversal order is deterministic.
8601 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
8602 Pat Haugen <pthaugen@us.ibm.com>
8603 Peter Bergner <bergner@vnet.ibm.com>
8605 * config/rs6000/vector.md (GPR move splitter): Do not split moves
8606 of vectors in GPRS if they are direct moves or quad word load or
8609 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
8611 (direct_move_p): Likewise.
8612 (quad_load_store_p): Likewise.
8614 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
8615 classes into bins based on the physical register type.
8616 (reg_class_to_reg_type): Likewise.
8617 (IS_STD_REG_TYPE): Likewise.
8618 (IS_FP_VECT_REG_TYPE): Likewise.
8619 (reload_fpr_gpr): Arrays to determine what insn to use if we can
8620 use direct move instructions.
8621 (reload_gpr_vsx): Likewise.
8622 (reload_vsx_gpr): Likewise.
8623 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
8624 information that is a simplification of register classes. Also
8625 precalculate direct move reload helpers.
8626 (direct_move_p): New function to return true if the operation can
8627 be done as a direct move instruciton.
8628 (quad_load_store_p): New function to return true if the operation
8629 is a quad memory operation.
8630 (rs6000_legitimize_address): If quad memory, only allow register
8631 indirect for TImode addresses.
8632 (rs6000_legitimate_address_p): Likewise.
8633 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
8634 (rs6000_reload_register_type): Likewise.
8635 (register_to_reg_type): Return register type.
8636 (rs6000_secondary_reload_simple_move): New helper function for
8637 secondary reload and secondary memory needed to identify anything
8638 that is a simple move, and does not need reloading.
8639 (rs6000_secondary_reload_direct_move): New helper function for
8640 secondary reload to identify cases that can be done with several
8641 instructions via the direct move instructions.
8642 (rs6000_secondary_reload_move): New helper function for secondary
8643 reload to identify moves between register types that can be done.
8644 (rs6000_secondary_reload): Add support for quad memory operations
8645 and for direct move.
8646 (rs6000_secondary_memory_needed): Likewise.
8647 (rs6000_debug_secondary_memory_needed): Change argument names.
8648 (rs6000_output_move_128bit): New function to return the move to
8649 use for 128-bit moves, including knowing about the various
8650 limitations of quad memory operations.
8652 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
8653 memory operations. call rs6000_output_move_128bit for the actual
8654 instruciton(s) to generate.
8655 (vsx_movti_64bit): Likewise.
8657 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
8658 (UNSPEC_P8V_MTVSRWZ): Likewise.
8659 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
8660 (UNSPEC_P8V_MTVSRD): Likewise.
8661 (UNSPEC_P8V_XXPERMDI): Likewise.
8662 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
8663 (UNSPEC_FUSION_GPR): Likewise.
8664 (FMOVE128_GPR): New iterator for direct move.
8665 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
8668 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
8669 loads and direct move instructions.
8670 (zero_extendsidi2_lfiwzx): Likewise.
8671 (extendsidi2_lfiwax): Likewise.
8672 (extendsidi2_nocell): Likewise.
8673 (floatsi<mode>2_lfiwax): Likewise.
8675 (floatunssi<mode>2_lfiwzx): Likewise.
8677 (fix_trunc<mode>_stfiwx): Likewise.
8678 (fixuns_trunc<mode>_stfiwx): Likewise.
8679 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
8680 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
8681 (parity<mode>2_cmpb): Set length/type attr.
8682 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
8683 for 'mr.' to fast_compare.
8684 (bpermd_<mode>): Change type attr to popcnt.
8685 (p8_fmrgow_<mode>): New insns for power8 direct move support.
8686 (p8_mtvsrwz_1): Likewise.
8687 (p8_mtvsrwz_2): Likewise.
8688 (reload_fpr_from_gpr<mode>): Likewise.
8689 (p8_mtvsrd_1): Likewise.
8690 (p8_mtvsrd_2): Likewise.
8691 (p8_xxpermdi_<mode>): Likewise.
8692 (reload_vsx_from_gpr<mode>): Likewise.
8693 (reload_vsx_from_gprsf): Likewise.
8694 (p8_mfvsrd_3_<mode>): LIkewise.
8695 (reload_gpr_from_vsx<mode>): Likewise.
8696 (reload_gpr_from_vsxsf): Likewise.
8697 (p8_mfvsrd_4_disf): Likewise.
8698 (multi-word GPR splits): Do not split direct moves or quad memory
8701 2013-06-10 David Malcolm <dmalcolm@redhat.com>
8703 * tree-into-ssa.c (interesting_blocks): Make static.
8705 2013-06-10 Jakub Jelinek <jakub@redhat.com>
8708 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
8709 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
8710 Use DATA_ABI_ALIGNMENT for that case instead if defined.
8711 (get_variable_align): New function.
8712 (get_variable_section, emit_bss, emit_common,
8713 assemble_variable_contents, place_block_symbol): Use
8714 get_variable_align instead of DECL_ALIGN.
8715 (assemble_noswitch_variable): Add align argument, use it
8716 instead of DECL_ALIGN.
8717 (assemble_variable): Adjust caller. Use get_variable_align
8718 instead of DECL_ALIGN.
8719 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
8721 (DATA_ABI_ALIGNMENT): Define.
8722 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
8723 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
8724 opt is false, only return the psABI mandated alignment increase.
8725 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
8726 (DATA_ABI_ALIGNMENT): ... this.
8727 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
8728 (DATA_ABI_ALIGNMENT): ... this.
8729 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
8730 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
8731 (DATA_ABI_ALIGNMENT): ... this.
8732 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
8733 * doc/tm.texi: Regenerated.
8735 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
8737 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
8738 cmp_code to construct REG_EQUAL note.
8740 2013-06-09 Jakub Jelinek <jakub@redhat.com>
8743 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
8744 that operands[2] doesn't overlap with operands[0].
8746 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
8747 Jan Hubicka <jh@suse.cz>
8749 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
8750 hack to mark symbols as used.
8752 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
8754 PR rtl-optimization/57559
8755 * lra-constraints.c (process_alt_operands): Don't discourage
8756 memory with known offset for offsetable memory constraint.
8757 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
8759 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
8761 * varasm.c (struct oc_local_state): Reorder fields.
8762 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
8763 and adjust accordingly.
8764 (output_constructor): Reorder initialization code and adjust call to
8765 output_constructor_bitfield.
8767 2013-06-07 Jan Hubicka <jh@suse.cz>
8769 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
8771 2013-06-07 David Malcolm <dmalcolm@redhat.com>
8773 * tree-object-size.c (unknown): Make const.
8775 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8777 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
8778 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
8779 for last alternative in the cpu_facility attribute.
8781 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8784 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
8785 (xordi3): Change operand 2 constraint to arm_xordi_operand.
8786 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
8787 * config/arm/constraints.md (Dg): New constraint.
8788 * config/arm/neon.md (xordi3_neon): Remove.
8789 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
8790 * config/arm/predicates.md (arm_xordi_operand): New predicate.
8792 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8794 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
8795 Clean up alternatives.
8797 2013-06-07 Alan Modra <amodra@gmail.com>
8799 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
8802 2013-06-07 Alan Modra <amodra@gmail.com>
8804 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
8806 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8808 * config/arm/constraints.md (Df): New constraint.
8809 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
8810 Correct length attribute for last two alternatives.
8812 2013-06-07 Alan Modra <amodra@gmail.com>
8814 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8815 override user -mfp-in-toc.
8816 (offsettable_ok_by_alignment): Consider just the current access
8817 rather than the whole object, unless BLKmode. Handle
8818 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
8819 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
8820 for -mcmodel=medium.
8821 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
8822 override user -mfp-in-toc or -msum-in-toc. Default to
8823 -mno-fp-in-toc for -mcmodel=medium.
8825 2013-06-06 DJ Delorie <dj@redhat.com>
8827 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
8828 TARGET_VALID_POINTER_MODE.
8830 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
8831 Pat Haugen <pthaugen@us.ibm.com>
8832 Peter Bergner <bergner@vnet.ibm.com>
8834 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8835 Document new power8 builtins.
8837 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
8838 condition code register, to allow 128-bit logical operations to be
8839 done in the VSX or GPR registers.
8840 (nor<mode>3): Use the canonical form for nor.
8841 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
8842 vclz*, and vpopcnt* vector instructions.
8843 (nand<mode>3): Likewise.
8844 (orc<mode>3): Likewise.
8845 (clz<mode>2): LIkewise.
8846 (popcount<mode>2): Likewise.
8848 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
8849 that only the GPRs are recognized.
8851 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8852 support for new power8 builtins.
8854 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
8856 (xscvdpspn): Likewise.
8862 (vpopcnt): Likewise.
8863 (vpopcntb): Likewise.
8864 (vpopcnth): Likewise.
8865 (vpopcntw): Likewise.
8866 (vpopcntd): Likewise.
8871 (eqv_v16qi3): Likewise.
8872 (eqv_v8hi3): Likewise.
8873 (eqv_v4si3): Likewise.
8874 (eqv_v2di3): Likewise.
8875 (eqv_v4sf3): Likewise.
8876 (eqv_v2df3): Likewise.
8878 (nand_v16qi3): Likewise.
8879 (nand_v8hi3): Likewise.
8880 (nand_v4si3): Likewise.
8881 (nand_v2di3): Likewise.
8882 (nand_v4sf3): Likewise.
8883 (nand_v2df3): Likewise.
8885 (orc_v16qi3): Likewise.
8886 (orc_v8hi3): Likewise.
8887 (orc_v4si3): Likewise.
8888 (orc_v2di3): Likewise.
8889 (orc_v4sf3): Likewise.
8890 (orc_v2df3): Likewise.
8892 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
8893 allow power8 quad mode in 64-bit.
8894 (rs6000_builtin_vectorized_function): Add support to vectorize
8895 ISA 2.07 count leading zeros, population count builtins.
8896 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
8897 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
8898 (builtin_function_type): Add vgbbd builtin function which takes an
8900 (altivec_expand_vec_perm_const): Add support for new power8 merge
8903 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
8904 that does not include TImdoe for use with 32-bit.
8905 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
8907 (UNSPEC_VSX_CVDPSPN): Likewise.
8908 (vsx_xscvdpspn): Likewise.
8909 (vsx_xscvspdpn): Likewise.
8910 (vsx_xscvdpspn_scalar): Likewise.
8911 (vsx_xscvspdpn_directmove): Likewise.
8912 (vsx_and<mode>3): Split logical operations into 32-bit and
8913 64-bit. Add support to do logical operations on TImode as well as
8914 VSX vector types. Allow logical operations to be done in either
8915 VSX registers or in general purpose registers in 64-bit mode. Add
8916 splitters if GPRs were used. For AND, add clobber of CCmode to
8917 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
8919 (vsx_and<mode>3_32bit): Likewise.
8920 (vsx_and<mode>3_64bit): Likewise.
8921 (vsx_ior<mode>3): Likewise.
8922 (vsx_ior<mode>3_32bit): Likewise.
8923 (vsx_ior<mode>3_64bit): Likewise.
8924 (vsx_xor<mode>3): Likewise.
8925 (vsx_xor<mode>3_32bit): Likewise.
8926 (vsx_xor<mode>3_64bit): Likewise.
8927 (vsx_one_cmpl<mode>2): Likewise.
8928 (vsx_one_cmpl<mode>2_32bit): Likewise.
8929 (vsx_one_cmpl<mode>2_64bit): Likewise.
8930 (vsx_nor<mode>3): Likewise.
8931 (vsx_nor<mode>3_32bit): Likewise.
8932 (vsx_nor<mode>3_64bit): Likewise.
8933 (vsx_andc<mode>3): Likewise.
8934 (vsx_andc<mode>3_32bit): Likewise.
8935 (vsx_andc<mode>3_64bit): Likewise.
8936 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
8937 and xxlorc instructions.
8938 (vsx_eqv<mode>3_64bit): Likewise.
8939 (vsx_nand<mode>3_32bit): Likewise.
8940 (vsx_nand<mode>3_64bit): Likewise.
8941 (vsx_orc<mode>3_32bit): Likewise.
8942 (vsx_orc<mode>3_64bit): Likewise.
8944 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
8946 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
8948 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
8949 (p8_vmrgow): Likewise.
8950 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
8951 GPRs to be split under VSX.
8952 (p8v_clz<mode>2): Add power8 count leading zero support.
8953 (p8v_popcount<mode>2): Add power8 population count support.
8954 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
8957 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
8960 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
8962 (vec_nand): Likewise.
8963 (vec_vclz): Likewise.
8964 (vec_vclzb): Likewise.
8965 (vec_vclzd): Likewise.
8966 (vec_vclzh): Likewise.
8967 (vec_vclzw): Likewise.
8968 (vec_vgbbd): Likewise.
8969 (vec_vmrgew): Likewise.
8970 (vec_vmrgow): Likewise.
8971 (vec_vpopcnt): Likewise.
8972 (vec_vpopcntb): Likewise.
8973 (vec_vpopcntd): Likewise.
8974 (vec_vpopcnth): Likewise.
8975 (vec_vpopcntw): Likewise.
8977 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8979 PR rtl-optimization/57468
8980 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
8983 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8985 PR rtl-optimization/57459
8986 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
8987 type when setting live regs.
8989 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8991 * config/s390/s390.opt (mlra): New option.
8992 * config/s390/s390.c (s390_decompose_address): Check displacement
8993 for all registers for LRA.
8994 (s390_secondary_reload): Don't used secondary reloads for LRA.
8995 (s390_lra_p): New function.
8996 (TARGET_LRA_P): Define.
8997 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
8998 of attribute cpu_facility to zarch for the last alternative.
8999 (*cmpmem_short): Ditto.
9001 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9003 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9004 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9005 (arm_expand_prologue): Likewise.
9007 2013-06-06 Teresa Johnson <tejohnson@google.com>
9010 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9011 as this is now done by redirect_edge_and_branch_force.
9012 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9013 barriers, and fix interaction with splitting.
9014 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9015 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9016 reflect changes made in the routine.
9017 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9018 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9019 since this is called in cfglayout mode, and replace partition fixup
9020 with assert as that is now done by force_nonfallthru_and_redirect.
9021 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9022 already be marked with region crossing note.
9023 (insert_section_boundary_note): Make non-static, gate on flag
9024 has_bb_partition, rewrite to also check for multiple partitions.
9025 (rest_of_handle_reorder_blocks): Remove call to
9026 insert_section_boundary_note, now done later during free_cfg.
9027 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9028 * bb-reorder.h (insert_section_boundary_note): Declare.
9029 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9030 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9031 invoke insert_section_boundary_note.
9032 (try_redirect_by_replacing_jump): Remove unnecessary
9033 check for region crossing note.
9034 (fixup_partition_crossing): New function.
9035 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9036 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9037 in non-cfglayout mode.
9038 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9039 remove old code that tried to do this. Emit barrier correctly
9040 when we are in cfglayout mode.
9041 (last_bb_in_partition): New function.
9042 (rtl_split_edge): Correctly fixup partition boundaries.
9043 (commit_one_edge_insertion): Remove old code that tried to
9044 fixup region crossing edge since this is now handled in
9045 split_block, and set up insertion point correctly since
9046 block may now end in a jump.
9047 (verify_hot_cold_block_grouping): Guard against checking when not in
9048 linearized RTL mode.
9049 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9051 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9052 rtl_verify_flow_info, so not called in cfglayout mode.
9053 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9054 (fixup_reorder_chain): Remove old code that attempted to fixup region
9055 crossing note as this is now handled in force_nonfallthru_and_redirect.
9056 (duplicate_insn_chain): Don't duplicate switch section notes.
9057 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9059 * basic-block.h (emit_barrier_after_bb): Declare.
9061 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9063 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9064 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9065 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9067 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9069 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9071 * config/arm/ldmstm.md: Regenerate.
9073 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9075 * config/arm/sync.md (atomic_loaddi_1):
9076 Disable predication for arm_restrict_it.
9077 (arm_load_exclusive<mode>): Likewise.
9078 (arm_load_exclusivesi): Likewise.
9079 (arm_load_exclusivedi): Likewise.
9080 (arm_load_acquire_exclusive<mode>): Likewise.
9081 (arm_load_acquire_exclusivesi): Likewise.
9082 (arm_load_acquire_exclusivedi): Likewise.
9083 (arm_store_exclusive<mode>): Likewise.
9084 (arm_store_exclusive<mode>): Likewise.
9085 (arm_store_release_exclusivedi): Likewise.
9086 (arm_store_release_exclusive<mode>): Likewise.
9088 2013-06-06 Richard Biener <rguenther@suse.de>
9090 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9092 (lto_tag_is_tree_code_p): Adjust.
9093 (lto_tag_is_gimple_code_p): Likewise.
9094 (lto_gimple_code_to_tag): Likewise.
9095 (lto_tag_to_gimple_code): Likewise.
9096 (lto_tree_code_to_tag): Likewise.
9097 (lto_tag_to_tree_code): Likewise.
9098 * data-streamer.h (streamer_write_hwi_in_range): Use
9099 uhwi streaming to stream the normalized range.
9100 (streamer_read_hwi_in_range): Likewise.
9102 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9104 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9105 (predicable_short_it): Likewise.
9106 (predicated): Likewise.
9107 (enabled): Handle above.
9108 (define_cond_exec): Set predicated attribute to yes.
9110 2013-06-05 Mike Stump <mikestump@comcast.net>
9112 * gdbinit.in (__FUNCTION__): Add.
9114 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
9116 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9117 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9119 2013-06-05 Jan Hubicka <jh@suse.cz>
9121 * varasm.c (mark_decl_referenced): Revert the removal until targets
9124 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
9126 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9127 instead of mark_decl_referenced.
9129 2013-06-05 Jan Hubicka <jh@suse.cz>
9131 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9132 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9133 and symtab_used_from_object_file_p.
9134 (cgraph_make_node_local_1): Clear forced_by_abi.
9135 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9136 * cgraph.h (symtab_node_base): Add forced_by_abi.
9137 (decide_is_variable_needed): Remove.
9138 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9139 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9140 (decide_is_symbol_needed): ... this one; handle symbols in general;
9141 always analyze virtuals; honnor forced_by_abi.
9142 (cgraph_finalize_function): Update.
9143 (varpool_finalize_decl): Update.
9144 (symbol_defined_and_needed): Remove.
9145 (analyze_functions): Update.
9146 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9147 output_refs, input_overwrite_node): Handle forced_by_abi.
9148 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9149 (address_taken_from_non_vtable_p): ... this one.
9150 (comdat_can_be_unshared_p_1): New function.
9151 (cgraph_comdat_can_be_unshared_p): Rename to ...
9152 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9153 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9154 (function_and_variable_visibility): Clear forced_by_abi as needed.
9155 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9156 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
9157 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9158 * varpool.c (decide_is_variable_needed): Remove.
9160 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9162 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9163 (arm_option_override): Override arm_restrict_it where appropriate.
9164 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9165 * config/arm/arm.opt (mrestrict-it): New command-line option.
9166 * doc/invoke.texi: Document -mrestrict-it.
9168 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9170 * tsan.c (tsan_atomic_table): Make const.
9172 2013-06-05 Richard Biener <rguenther@suse.de>
9174 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9175 index associated with the tree we are supposed to replace.
9176 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9177 TREE_ASM_WRITTEN as zero for everything but SSA names.
9179 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9181 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9183 2013-06-04 Jan Hubicka <jh@suse.cz>
9185 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9186 (input_node, input_varpool_node): Handle correctly external same
9188 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9189 nodes at ltrans stage.
9191 2013-06-04 Jan Hubicka <jh@suse.cz>
9193 * ipa-inline.c (update_caller_keys): Fix availability test.
9194 (update_callee_keys): Likewise.
9195 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9196 to follow ELF standard.
9198 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9200 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9201 (mips64r5900el-*-elf*): New configurations.
9202 * config/mips/mips-cpus.def (r5900): New processor.
9203 * config/mips/mips-tables.opt: Regenerate.
9204 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9205 (mips_issue_rate): Handle PROCESSOR_R5900.
9206 (mips_reorg_process_insns): Force reorder mode for the R5900.
9207 * config/mips/mips.h (TARGET_MIPS5900): Define.
9208 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9210 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9212 * config/mips/mips.md (processor): Add r5900.
9213 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9215 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9217 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9218 into function to generate MOVI instruction.
9219 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
9220 (aarch64_preferred_simd_mode): Turn into wrapper.
9221 (aarch64_output_scalar_simd_mov_immediate): New function.
9222 * config/aarch64/aarch64-protos.h: Add prototype for above.
9224 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9226 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9227 element_char member.
9228 (sizetochar): Return signed char.
9229 (aarch64_simd_valid_immediate): Remove elchar and other
9230 unnecessary variables.
9231 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9232 Calculate element_char as required.
9233 * config/aarch64/aarch64-protos.h: Update and move prototype
9234 for aarch64_output_simd_mov_immediate.
9235 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9238 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9240 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9241 information completed by aarch64_simd_valid_immediate.
9242 (aarch64_legitimate_constant_p): Update arguments.
9243 (aarch64_simd_valid_immediate): Work with struct rather than many
9245 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9246 (aarch64_simd_make_constant): Update arguments.
9247 (aarch64_output_simd_mov_immediate): Work with struct rather than
9248 many pointers. Output immediate directly rather than as operand.
9249 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9251 * config/aarch64/constraints.md (Dn): Update arguments.
9253 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9255 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9257 (aarch64_simd_immediate_valid_for_move): Remove.
9258 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9259 (aarch64_simd_make_constant): Update call.
9260 (aarch64_output_simd_mov_immediate): Update call.
9261 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9263 * config/aarch64/constraints.md (Dn): Update call.
9265 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9267 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9268 return type to bool for prototype.
9269 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9270 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9271 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9273 2013-06-04 Catherine Moore <clm@codesourcery.com>
9275 * config/mips/mips.opt (meva): New.
9276 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
9277 (ASM_SPEC): Handle -meva.
9278 * doc/invoke.texi (meva): Document.
9280 2013-06-04 Alan Modra <amodra@gmail.com>
9282 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9285 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9287 * rtl.def: Add extra fourth optional field to define_cond_exec.
9288 * gensupport.c (process_one_cond_exec): Process attributes from
9290 * doc/md.texi: Document fourth field in define_cond_exec.
9292 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9294 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9295 out the processing order as in store_bit_field_1.
9297 2013-06-04 Jan Hubicka <jh@suse.cz>
9300 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9301 non-VAR_DECL node if it does not exist yet.
9303 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9305 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9306 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9307 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9308 target_cpu_default setting.
9310 2013-06-03 Teresa Johnson <tejohnson@google.com>
9312 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9313 default to -fopt-info=optimized instead of all.
9314 * doc/invoke.texi: Ditto.
9315 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9316 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9317 (execute_vect_slp): Emit BB vectorization success under
9318 MSG_OPTIMIZED_LOCATIONS.
9319 * tree-vect-slp.c (vect_slp_transform_bb): Change
9320 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9321 * tree-vect-loop.c (vect_transform_loop): Ditto.
9323 2013-06-03 Jason Merrill <jason@redhat.com>
9326 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9327 Use TARGET_EXPR for C++.
9329 2013-06-03 Jakub Jelinek <jakub@redhat.com>
9331 PR rtl-optimization/57268
9332 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
9333 if DEBUG_INSN_P (insn).
9336 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9338 PR rtl-optimization/57268
9339 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9340 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9342 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9344 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9345 (ix86_avoid_lea_for_addr): Likewise.
9346 (exact_dependency_1): Likewise.
9347 (ix86_adjust_cost): Likewise.
9348 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9350 (do_reorder_for_imul): Fix typo, formatting and
9351 !reload_completed check removed.
9352 (ix86_sched_reorder): Fix typo and formatting.
9353 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9356 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9358 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9360 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9362 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
9363 <STRING_CST>: Likewise.
9364 <VECTOR_CST>: Likewise.
9366 2013-06-01 Janus Weil <janus@gcc.gnu.org>
9367 Mikael Morin <mikael@gcc.gnu.org>
9369 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
9370 * config.in: Regenerated.
9371 * configure: Regenerated.
9373 2013-06-01 Jan Hubicka <jh@suse.cz>
9376 * cgraphunit.c (compile): When weakref is not supported,
9377 set up transparent aliases before final output pass.
9378 * varasm.c (assemble_alias): Do not try to do it here.
9380 2013-06-01 Jan Hubicka <jh@suse.cz>
9383 * passes.c (for_per_function): Skip unanalyzed functions.
9385 2013-06-01 Jan Hubicka <jh@suse.cz>
9387 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
9388 (lto_symtab_merge_symbols_1): ... this one.
9389 (lto_symtab_merge_cgraph_nodes): Rename to ...
9390 (lto_symtab_merge_symbols): ... this one; simplify.
9391 * cgraph.c (same_body_aliases_done): Rename to ...
9392 (cpp_implicit_aliases_done): ... this one.
9393 (cgraph_create_function_alias): Update.
9394 (cgraph_same_body_alias): Update.
9395 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
9396 (verify_edge_corresponds_to_fndecl): Simplify.
9397 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
9398 (cgraph_node): Remove same_body_alias.
9399 (varpool_node): Remove alias_of and extra_name_alias.
9400 (same_body_aliases_done): Rename to ..
9401 (cpp_implicit_aliases_done): ... this one.
9402 (symtab_alias_ultimate_target): Add default parameter.
9403 (symtab_resolve_alias): New function.
9404 (fixup_same_cpp_alias_visibility): Declare.
9405 (cgraph_function_node): Add default parameter.
9406 (cgraph_node_asm_name): Likewise.
9407 (cgraph_function_or_thunk_node): Add default parameter; do
9408 not ICE when it is NULL.
9409 (varpool_variable_node): Likewise.
9410 * tree-emutls.c (create_emultls_var): Update.
9411 (ipa_lower_emutls): Update.
9412 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
9413 (cgraph_reset_node): Reset alias info.
9414 (cgraph_finalize_function): Update.
9415 (fixup_same_cpp_alias_visibility): Move to symtab.c.
9416 (analyze_function): Simplify.
9417 (cgraph_process_same_body_aliases): Simplify.
9418 (analyze_functions): Fixup same body aliases.
9419 (handle_alias_pairs): Simplify.
9420 (assemble_thunk): Update.
9421 (assemble_thunks_and_aliases): Update.
9422 (output_weakrefs): Rewrite.
9423 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
9424 (lto_output_varpool_node): Likewise.
9425 (compute_ltrans_boundary): Remve assert.
9426 (get_alias_symbol): New functoin.
9427 (input_node): Rewrite alias handling.
9428 (input_varpool_node): Likewise.
9429 * ipa-pure-const.c (propagate_pure_const): Fix formating.
9430 * ipa.c (process_references): Handle weakrefs correctly.
9431 (symtab_remove_unreachable_nodes): Likewise.
9432 * trans-mem.c (get_cg_data): Update.
9433 (ipa_tm_create_version_alias): Update.
9434 (ipa_tm_execute): Update.
9435 * symtab.c (dump_symtab_base): Dump aliases.
9436 (verify_symtab_base): Verify aliases.
9437 (symtab_node_availability): New function.
9438 (symtab_alias_ultimate_target): Simplify.
9439 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
9440 handle all the fixup cases.
9441 (symtab_resolve_alias): New function.
9442 * passes.c (ipa_write_summaries): Handle weakrefs.
9443 * varpool.c (varpool_analyze_node): Simplify.
9444 (assemble_aliases): Update.
9445 (varpool_create_variable_alias): Simplify.
9446 (varpool_extra_name_alias): Simplify.
9447 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
9448 (lto_symtab_merge_symbols): ... this one.
9450 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
9453 PR rtl-optimization/57268
9454 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9455 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9457 2013-06-01 Tobias Burnus <burnus@net-b.de>
9460 2013-05-31 Tobias Burnus <burnus@net-b.de>
9463 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
9466 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9468 PR rtl-optimization/57268
9469 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9470 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9472 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9474 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
9475 unordered comparison operators when -fno-trapping-math is in effect
9477 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
9478 and implement unordered comparison operators properly on the e500.
9480 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9482 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
9483 for constant scalar integers.
9484 (simplify_relational_operation_1): Likewise.
9486 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
9488 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
9489 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
9492 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9493 Igor Zamyatin <igor.zamyatin@intel.com>
9495 Silvermont (SLM) architecture performance tuning.
9496 * config/i386/i386.h (enum ix86_tune_indices): Add
9497 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
9498 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
9500 * config/i386/i386.c (initial_ix86_tune_features)
9501 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
9502 (ix86_lea_outperforms): Handle Silvermont tuning.
9503 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
9505 (ix86_use_lea_for_mov): Likewise.
9506 (ix86_avoid_lea_for_addr): Likewise.
9507 (ix86_lea_for_add_ok): Likewise.
9508 (exact_dependency_1): New function.
9509 (exact_store_load_dependency): Likewise.
9510 (ix86_adjust_cost): Handle Silvermont tuning.
9511 (do_reoder_for_imul): Likewise.
9512 (swap_top_of_ready_list): New function.
9513 (ix86_sched_reorder): Changed to handle Silvermont tuning.
9515 * config/i386/i386.md (peepholes that split memory operand in fp
9518 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9520 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
9521 Remove un-necessary braces.
9523 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9525 * config/aarch64/aarch64.c (aarch64_classify_symbol):
9526 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
9528 2013-05-31 Tobias Burnus <burnus@net-b.de>
9531 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
9533 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9536 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
9537 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
9538 * config/arm/neon.md (iordi3_neon): Remove.
9539 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
9540 * config/arm/predicates.md (imm_for_neon_logic_operand):
9541 Move to earlier in the file.
9542 (neon_logic_op2): Likewise.
9543 (arm_iordi_operand_neon): New predicate.
9545 2013-05-31 Richard Biener <rguenther@suse.de>
9547 PR tree-optimization/57478
9548 PR tree-optimization/57453
9549 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
9552 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9554 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
9555 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
9557 2013-05-30 Tobias Burnus <burnus@net-b.de>
9558 Thomas Koenig <tkoenig@gcc.gnu.org>
9561 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
9562 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
9564 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
9566 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
9568 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
9570 * target.def (register_usage_leveling_p): New hook.
9571 * targhooks.c (default_register_usage_leveling_p): New.
9572 * targhooks.h (default_register_usage_leveling_p): New prototype.
9573 * lra-assigns.c (register_usage_leveling_p): Use the hook.
9574 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
9575 * doc/tm.texi: Update.
9576 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
9578 2013-05-30 Ian Bolton <ian.bolton@arm.com>
9580 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
9581 (*insv_reg<mode>): New define_insn.
9583 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
9585 PR rtl-optimization/57439
9586 * postreload.c (move2add_valid_value_p): Check that we have
9587 a zero subreg_regno_offset when accessing the register in
9590 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9591 Igor Zamyatin <igor.zamyatin@intel.com>
9593 Silvermont (SLM) architecture pipeline model, tuning and
9595 * config.gcc: Add slm config options and target.
9597 * config/i386/slm.md: New.
9599 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
9601 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
9603 (ix86_target_macros_internal): Likewise.
9605 * gcc/config/i386/i386.c (slm_cost): New cost.
9606 (m_SLM): New macro flag.
9607 (initial_ix86_tune_features): Set m_SLM.
9608 (x86_accumulate_outgoing_args): Likewise.
9609 (x86_arch_always_fancy_math_387): Likewise.
9610 (processor_target_table): Add slm cost.
9611 (cpu_names): Add slm cpu name.
9612 (x86_option_override_internal): Set SLM ISA.
9613 (ix86_issue_rate): New case PROCESSOR_SLM.
9614 (ia32_multipass_dfa_lookahead): Likewise.
9615 (fold_builtin_cpu): Add slm.
9617 * config/i386/i386.h (TARGET_SLM): New target macro.
9618 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
9619 (processor_type): Add PROCESSOR_SLM.
9621 * config/i386/i386.md (cpu): Add new value "slm".
9622 (slm.md): Include slm.md.
9624 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
9625 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9627 * config/arm/arm-protos.h: Add and update function protos.
9628 * config/arm/arm.c (use_simple_return_p): New added.
9629 (thumb2_expand_return): Check simple_return flag.
9630 * config/arm/arm.md: Add simple_return and conditional simple_return.
9631 * config/arm/iterators.md: Add iterator for return and simple_return.
9633 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9635 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
9636 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
9637 (arm_emit_vfp_multi_reg_pop): Likewise.
9638 (thumb2_emit_ldrd_pop): Likewise.
9639 (arm_expand_epilogue): Add misc REG_CFA notes.
9640 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
9642 2013-05-29 Lawrence Crowl <crowl@google.com>
9644 * config/arm/t-arm: Update for below.
9646 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
9647 Change type to hash_table. Update dependent calls and types.
9649 * config/i386/t-cygming: Update for below.
9651 * config/i386/t-interix: Update for below.
9653 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
9654 Change type to hash_table. Update dependent calls and types.
9655 (i386_find_on_wrapper_list::wrappers): Likewise.
9657 * config/ia64/t-ia64: Update for below.
9659 * config/ia64/ia64.c (bundle_state_table):
9660 Change type to hash_table. Update dependent calls and types.
9662 * config/mips/mips.c (mips_reorg_process_insns::htab):
9663 Change type to hash_table. Update dependent calls and types.
9665 * config/sol2.c (solaris_comdat_htab):
9666 Change type to hash_table. Update dependent calls and types.
9668 * config/t-sol2: Update for above.
9670 2013-05-29 Teresa Johnson <tejohnson@google.com>
9672 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
9673 functions are not yet marked as defined.
9675 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
9676 Pat Haugen <pthaugen@us.ibm.com>
9677 Peter Bergner <bergner@vnet.ibm.com>
9679 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
9683 (vrotl<mode>3): Likewise.
9684 (vashl<mode>3): Likewise.
9685 (vlshr<mode>3): Likewise.
9686 (vashr<mode>3): Likewise.
9688 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9689 support for power8 V2DI builtins.
9691 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
9692 power8 V2DI builtins.
9693 (vupkhsw): Likewise.
9694 (vupklsw): Likewise.
9695 (vaddudm): Likewise.
9700 (vpkudum): Likewise.
9701 (vpksdss): Likewise.
9702 (vpkudus): Likewise.
9703 (vpksdus): Likewise.
9708 (vsubudm): Likewise.
9709 (vcmpequd): Likewise.
9710 (vcmpgtsd): Likewise.
9711 (vcmpgtud): Likewise.
9712 (vcmpequd_p): Likewise.
9713 (vcmpgtsd_p): Likewise.
9714 (vcmpgtud_p): Likewise.
9715 (vupkhsw): Likewise.
9716 (vupklsw): Likewise.
9717 (vaddudm): Likewise.
9722 (vpksdss): Likewise.
9723 (vpksdus): Likewise.
9724 (vpkudum): Likewise.
9725 (vpkudus): Likewise.
9730 (vsubudm): Likewise.
9732 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
9733 support for power8 V2DI instructions.
9735 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
9736 power8 V2DI instructions. Combine pack and unpack insns to use an
9737 iterator for each mode. Check whether a particular mode supports
9738 Altivec instructions instead of just checking TARGET_ALTIVEC.
9739 (UNSPEC_VPKUWUM): Likewise.
9740 (UNSPEC_VPKSHSS): Likewise.
9741 (UNSPEC_VPKSWSS): Likewise.
9742 (UNSPEC_VPKUHUS): Likewise.
9743 (UNSPEC_VPKSHUS): Likewise.
9744 (UNSPEC_VPKUWUS): Likewise.
9745 (UNSPEC_VPKSWUS): Likewise.
9746 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
9747 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
9748 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
9749 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
9750 (UNSPEC_VUPKHSB): Likewise.
9751 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
9752 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
9753 (UNSPEC_VUPKHSH): Likewise.
9754 (UNSPEC_VUPKLSB): Likewise.
9755 (UNSPEC_VUPKLSH): Likewise.
9757 (VI_char): Likewise.
9758 (VI_scalar): Likewise.
9759 (VI_unit): Likewise.
9761 (VP_small): Likewise.
9762 (VP_small_lc): Likewise.
9763 (VU_char): Likewise.
9764 (add<mode>3): Likewise.
9765 (altivec_vaddcuw): Likewise.
9766 (altivec_vaddu<VI_char>s): Likewise.
9767 (altivec_vadds<VI_char>s): Likewise.
9768 (sub<mode>3): Likewise.
9769 (altivec_vsubcuw): Likewise.
9770 (altivec_vsubu<VI_char>s): Likewise.
9771 (altivec_vsubs<VI_char>s): Likewise.
9772 (altivec_vavgs<VI_char>): Likewise.
9773 (altivec_vcmpbfp): Likewise.
9774 (altivec_eq<mode>): Likewise.
9775 (altivec_gt<mode>): Likewise.
9776 (altivec_gtu<mode>): Likewise.
9777 (umax<mode>3): Likewise.
9778 (smax<mode>3): Likewise.
9779 (umin<mode>3): Likewise.
9780 (smin<mode>3): Likewise.
9781 (altivec_vpkuhum): Likewise.
9782 (altivec_vpkuwum): Likewise.
9783 (altivec_vpkshss): Likewise.
9784 (altivec_vpkswss): Likewise.
9785 (altivec_vpkuhus): Likewise.
9786 (altivec_vpkshus): Likewise.
9787 (altivec_vpkuwus): Likewise.
9788 (altivec_vpkswus): Likewise.
9789 (altivec_vpks<VI_char>ss): Likewise.
9790 (altivec_vpks<VI_char>us): Likewise.
9791 (altivec_vpku<VI_char>us): Likewise.
9792 (altivec_vpku<VI_char>um): Likewise.
9793 (altivec_vrl<VI_char>): Likewise.
9794 (altivec_vsl<VI_char>): Likewise.
9795 (altivec_vsr<VI_char>): Likewise.
9796 (altivec_vsra<VI_char>): Likewise.
9797 (altivec_vsldoi_<mode>): Likewise.
9798 (altivec_vupkhsb): Likewise.
9799 (altivec_vupkhs<VU_char>): Likewise.
9800 (altivec_vupkls<VU_char>): Likewise.
9801 (altivec_vupkhsh): Likewise.
9802 (altivec_vupklsb): Likewise.
9803 (altivec_vupklsh): Likewise.
9804 (altivec_vcmpequ<VI_char>_p): Likewise.
9805 (altivec_vcmpgts<VI_char>_p): Likewise.
9806 (altivec_vcmpgtu<VI_char>_p): Likewise.
9807 (abs<mode>2): Likewise.
9808 (vec_unpacks_hi_v16qi): Likewise.
9809 (vec_unpacks_hi_v8hi): Likewise.
9810 (vec_unpacks_lo_v16qi): Likewise.
9811 (vec_unpacks_hi_<VP_small_lc>): Likewise.
9812 (vec_unpacks_lo_v8hi): Likewise.
9813 (vec_unpacks_lo_<VP_small_lc>): Likewise.
9814 (vec_pack_trunc_v8h): Likewise.
9815 (vec_pack_trunc_v4si): Likewise.
9816 (vec_pack_trunc_<mode>): Likewise.
9818 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
9820 (vec_vmaxsd): Likewise.
9821 (vec_vmaxud): Likewise.
9822 (vec_vminsd): Likewise.
9823 (vec_vminud): Likewise.
9824 (vec_vpksdss): Likewise.
9825 (vec_vpksdus): Likewise.
9826 (vec_vpkudum): Likewise.
9827 (vec_vpkudus): Likewise.
9828 (vec_vrld): Likewise.
9829 (vec_vsld): Likewise.
9830 (vec_vsrad): Likewise.
9831 (vec_vsrd): Likewise.
9832 (vec_vsubudm): Likewise.
9833 (vec_vupkhsw): Likewise.
9834 (vec_vupklsw): Likewise.
9836 2013-05-29 Jan Hubicka <jh@suse.cz>
9838 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
9839 flags; reorder rest of fields in more consistent way.
9840 (varpool_node): Remove analyzed, finalized and alias.
9841 (cgraph_ndoe): Likewise.
9842 (symtab_alias_ultimate_target): New function.
9843 (cgraph_function_node): Move offline.
9844 (cgraph_reset_node): Declare.
9845 (cgraph_comdat_can_be_unshared_p): Remove.
9846 (varpool_remove_initializer): Declare.
9847 (varpool_first_defined_variable, varpool_next_defined_variable
9848 cgraph_first_defined_function, cgraph_next_defined_function): Update.
9849 (cgraph_function_with_gimple_body_p): Update.
9850 (varpool_all_refs_explicit_p): Update.
9851 (symtab_alias_target): New function.
9852 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
9853 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
9854 (cgraph_function_or_thunk_node): Simplify using
9855 symtab_alias_ultimate_target.
9856 (varpool_variable_node): Likewise.
9857 * cgraph.c (cgraph_create_function_alias): Update.
9858 (cgraph_add_thunk): Update.
9859 (cgraph_remove_node): Update.
9860 (dump_cgraph_node): Do not dump removed flags.
9861 (cgraph_function_body_availability): Update.
9862 (cgraph_propagate_frequency): Update.
9863 (verify_cgraph_node): Check sanity of local flag.
9864 (cgraph_function_node): Move here from cgraph.h; revamp for
9865 cgraph_function_or_thunk_node.
9866 * lto-symtab.c (lto_varpool_replace_node): Update.
9867 (lto_symtab_resolve_can_prevail_p): Update.
9868 (lto_symtab_merge_cgraph_nodes): Update.
9869 * ipa-cp.c (determine_versionability, initialize_node_lattices,
9870 propagate_constants_accross_call, devirtualization_time_bonus,
9871 ipcp_propagate_stage): Update.
9872 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
9873 * ipa-inline-transform.c (clone_inlined_nodes,
9874 preserve_function_body_p): Update.
9875 * ipa-reference.c (propagate): Update.
9876 (write_node_summary_p): Update.
9877 * toplev.c (wrapup_global_declaration_2): Update.
9878 * cgraphunit.c (cgraph_analyze_function): Rename to ...
9879 (analyze_function) ... this one.
9880 (cgraph_process_new_functions): Update.
9881 (cgraph_reset_node): Export.
9882 (cgraph_finalize_function): Update.
9883 (cgraph_add_new_function): Update.
9884 (process_function_and_variable_attributes): Update.
9885 (varpool_finalize_decl): Update.
9886 (symbol_finalized): Remove.
9887 (symbol_finalized_and_needed): Rename to ...
9888 (symbol_defined_and_needed): ... update.
9889 (cgraph_analyze_functions): Update.
9890 (handle_alias_pairs): Update.
9891 (mark_functions_to_output): Update.
9892 (assemble_thunk): Update.
9893 (output_in_order): Update.
9894 (output_weakrefs): Update.
9895 (finalize_compilation_unit): Update.
9896 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
9897 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
9898 input_node, input_varpool_node): Update.
9899 * dbxout.c (dbxout_expand_expr): Update.
9900 * cgraphclones.c (cgraph_clone_node): Update.
9901 (cgraph_copy_node_for_versioning): Update.
9902 (cgraph_materialize_clone): Update.
9903 (cgraph_materialize_all_clones): Update.
9904 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
9905 propagate_pure_const, propagate_nothrow): Update.
9906 * lto-streamer-out.c (lto_output, write_symbol): Update.
9907 * ipa-utils.c (ipa_reverse_postorder): Update.
9908 * ipa-inline.c (can_inline_edge_p): Update.
9909 (update_caller_keys, ipa_inline): Update.
9910 * dwarf2out.c (reference_to_unused,
9911 premark_types_used_by_global_vars_helper): Update.
9912 * tree-eh.c (tree_could_trap_p): Update.
9913 * ipa-split.c (consider_split, execute_split_functions): Update.
9914 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
9915 has_addr_references_p): Update; move ahead in file for better
9917 (process_references): Simplify.
9918 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
9920 (cgraph_comdat_can_be_unshared_p): Make static.
9921 (cgraph_externally_visible_p): Update.
9922 (varpool_externally_visible_p): Update.
9923 (function_and_variable_visibility): Update.
9924 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
9925 ipa_tm_mark_force_output_node): Update.
9926 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
9927 estimate_edge_devirt_benefit, inline_generate_summary,
9928 inline_write_summary): Update.
9929 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
9930 * ipa-prop.c (ipa_compute_jump_functions): Update.
9931 (ipa_print_node_params, ipa_prop_read_section,
9932 ipa_update_after_lto_read, read_replacements_section): Update.
9933 * varasm.c (mark_decl_referenced): Update.
9934 (assemble_alias, dump_tm_clone_pairs): Update.
9935 * tree-inline.c (copy_bb): Update.
9936 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
9938 * symtab.c (dump_symtab_base): Print new flags.
9939 (verify_symtab_base): Verify new flags.
9940 (symtab_alias_ultimate_target): New function.
9941 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
9942 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
9944 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
9946 * i386.c (ix86_get_function_versions_dispatcher,
9947 ix86_generate_version_dispatcher_body): Update.
9948 (fold_builtin_cpu): Use varpool_add_new_variable.
9949 * varpool.c (varpool_remove_initializer): Break out from ...
9950 (varpool_remove_node): ... this one.
9951 (dump_varpool_node, varpool_node_for_asm,
9952 cgraph_variable_initializer_availability, varpool_analyze_node,
9953 varpool_assemble_decl, varpool_remove_unreferenced_decls,
9954 varpool_finalize_named_section_flags, varpool_create_variable_alias):
9957 2013-05-29 Jan Hubicka <jh@suse.cz>
9959 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
9961 2013-05-29 Easwaran Raman <eraman@google.com>
9963 PR tree-optimization/57442
9964 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
9965 when control exits the main loop.
9967 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
9969 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
9971 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
9972 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
9973 * rx/t-rx: Add rx100 under multi library matches option for nofpu
9976 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9978 PR tree-optimization/57441
9979 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
9980 Don't limit size of incr_vec to number of candidates.
9982 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
9984 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
9985 and mips16 directories.
9986 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
9987 (MULTILIB_DIRNAMES): Ditto.
9988 (MULTILIB_EXCEPTIONS): Add new exceptions.
9989 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
9990 (MULTILIB_DIRNAMES): Ditto.
9991 (MULTILIB_EXCEPTIONS): Add new exceptions.
9993 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9994 Marcus Shawcroft <marcus.shawcroft@arm.com>
9996 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
9997 SYMBOL_TINY_ABSOLUTE.
9998 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
9999 SYMBOL_TINY_ABSOLUTE.
10000 (aarch64_expand_mov_immediate): Likewise.
10001 (aarch64_classify_symbol): Likewise.
10002 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10003 Permit SYMBOL_TINY_ABSOLUTE.
10004 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10006 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10007 Marcus Shawcroft <marcus.shawcroft@arm.com>
10009 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10010 Refactor if/switch. Replace gcc_assert with if.
10012 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10014 * config/i386/i386.c (initial_ix86_tune_features): Enable
10015 FP Reassociation for AMD bdver1 and bdver2.
10017 2013-05-29 Martin Jambor <mjambor@suse.cz>
10019 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10020 and IMAGPART_EXPR do not occur within other handled_components.
10022 2013-05-29 Richard Biener <rguenther@suse.de>
10024 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10025 access on whether the use is in the BB we currently try to
10027 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10028 vectorize to vect_bb_slp_scalar_cost.
10030 2013-05-29 Richard Biener <rguenther@suse.de>
10032 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10033 computing scalar cost offsetted by stmts that are kept live
10035 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10036 for computation of scalar cost.
10038 2013-05-28 Steve Ellcey <sellcey@mips.com>
10040 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10042 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10044 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10045 array notation built-in reduction functions.
10046 * doc/passes.texi (Passes): Added documentation about changes done
10048 * doc/invoke.texi (C Dialect Options): Added documentation about
10049 the -fcilkplus flag.
10050 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10051 (BUILTINS_DEF): Depend on cilkplus.def.
10052 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10053 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10054 * cilkplus.def: New file.
10056 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
10058 PR rtl-optimization/57439
10059 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10061 2013-05-28 Easwaran Raman <eraman@google.com>
10063 PR tree-optimization/57337
10064 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10065 (find_insert_point): Correctly identify the insertion point
10066 when two statements with the same UID is compared.
10068 2013-05-28 Richard Biener <rguenther@suse.de>
10070 PR tree-optimization/56787
10071 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10072 from the list of data references.
10073 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10075 (vect_analyze_loop_operations): Likewise.
10076 (vect_transform_loop): Remove clobbers.
10078 2013-05-28 Martin Jambor <mjambor@suse.cz>
10080 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10081 and REALPART_EXPRs have scalar type.
10083 2013-05-28 Richard Biener <rguenther@suse.de>
10085 PR tree-optimization/57411
10086 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10088 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10089 virtual operand propagation.
10091 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10093 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10094 destination register for bmasksi_vis.
10095 (vector_init_bshuffle): Likewise.
10096 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10098 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10100 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10101 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10102 mode if the instruction isn't available in the original mode.
10103 * config/sparc/sparc.opt (mfix-ut699): New option.
10104 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10105 (divdf3): Turn into expander.
10106 (divdf3_nofix): New insn.
10107 (divdf3_fix): Likewise.
10108 (divsf3): Disable if -mfix-ut699.
10109 (sqrtdf2): Turn into expander.
10110 (sqrtdf2_nofix): New insn.
10111 (sqrtdf2_fix): Likewise.
10112 (sqrtsf2): Disable if -mfix-ut699.
10114 2013-05-27 Richard Biener <rguenther@suse.de>
10116 PR middle-end/57412
10117 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10118 block for the new loop.
10120 2013-05-27 Richard Biener <rguenther@suse.de>
10122 PR tree-optimization/57343
10123 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10124 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10125 (number_of_iterations_cond): Do not build the folded tree.
10127 2013-05-27 Richard Biener <rguenther@suse.de>
10130 PR middle-end/57381
10131 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10132 OEP_CONSTANT_ADDRESS_OF retained.
10134 PR tree-optimization/57417
10135 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10136 for unchanged base.
10137 (set_ssa_val_to): Compare addresses using
10138 get_addr_base_and_unit_offset.
10140 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
10142 PR rtl-optimization/56833
10143 * postreload.c (move2add_record_mode): New function.
10144 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10145 (move2add_use_add2_insn): Use move2add_record_sym_value.
10146 (move2add_use_add3_insn): Likewise.
10147 (reload_cse_move2add): Use move2add_valid_value_p and
10148 move2add_record_mode. Invalidate call-clobbered and REG_INC
10149 affected regs by setting reg_mode to VOIDmode.
10150 (move2add_note_store): Don't pretend the inside of a SUBREG is
10151 the actual destination. Invalidate single/leading registers by
10152 setting reg_mode to VOIDmode.
10153 Use move2add_record_sym_value, move2add_valid_value_p and
10154 move2add_record_mode.
10156 2013-05-27 Richard Biener <rguenther@suse.de>
10158 PR tree-optimization/57396
10159 * tree-affine.c (double_int_constant_multiple_p): Properly
10160 return false for val == 0 and div != 0.
10162 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10164 * config/mips/mips.h: Use #elif in preprocessor conditions.
10166 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10169 * config/mips/constraints.md (kl): New constraint.
10170 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10171 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10172 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10173 from LO for MIPS16.
10174 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10176 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10179 * config/mips/mips.c (mips_can_inline_p): New function.
10180 (TARGET_CAN_INLINE_P): Define.
10182 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10184 * sched-int.h (ds_t, dw_t): Make unsigned int.
10185 Fix documentation that describes how all the ds_t bits are used.
10186 Reserve the last bit for delayed-branch scheduling.
10187 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10188 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10189 (gen_dep_weak_1): Remove prototype.
10190 * sched-deps.c (get_dep_weak_1): Make static.
10191 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10192 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10193 * doc/tm.texi: Regenerate.
10194 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10196 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10199 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10201 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10202 Sandra Loosemore <sandra@codesourcery.com>
10204 * config.gcc (powerpc-*): Allow native for with-cpu.
10206 2013-05-24 Jeff Law <law@redhat.com>
10208 PR tree-optimization/57124
10209 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10210 conversion feeding a condition if the range has an overflow
10211 if -fstrict-overflow. Add warnings for when we do make the
10214 2013-05-24 Dehao Chen <dehao@google.com>
10216 * tree-cfg.c (locus_discrim_map): Fix the typo.
10217 (locus_discrim_hasher): Likewise.
10218 (locus_discrim_hasher::hash): Likewise.
10219 (locus_discrim_hasher::equal): Likewise.
10221 2013-05-24 Martin Jambor <mjambor@suse.cz>
10223 PR tree-optimization/57294
10224 * cgraph.h (ipa_record_stmt_references): Declare.
10225 * cgraphbuild.c (ipa_record_stmt_references): New function.
10226 (build_cgraph_edges): Use ipa_record_stmt_references.
10227 (rebuild_cgraph_edges): Likewise.
10228 (cgraph_rebuild_references): Likewise.
10229 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10230 associated with the old statement and build references from the
10231 newly built statements.
10232 * ipa-ref.c (ipa_remove_stmt_references): New function.
10233 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10235 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10237 * lra-constraints.c (emit_spill_move): Use smaller mode for
10239 (check_and_process_move): Consider mem-reg moves for secondary
10241 (curr_insn_transform): Don't lose insns emitted before for
10242 secondary memory moves.
10243 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10244 reg set up in the current insn.
10246 2013-05-24 Dehao Chen <dehao@google.com>
10248 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
10250 (locus_descrim_hasher::equal): Likewise.
10251 (build_gimple_cfg): New discriminator assignment algorithm.
10252 (make_edges): Likewise.
10253 (next_discriminator_for_locus): Likewise.
10254 (same_line_p): Likewise.
10255 (assign_discriminators): Likewise.
10256 (make_cond_expr_edges): Likewise.
10257 (make_gimple_switch_edges): Likewise.
10258 (make_goto_expr_edges): Likewise.
10259 (make_gimple_asm_edges): Likewise.
10261 2013-05-24 Ian Bolton <ian.bolton@arm.com>
10263 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10264 X format specifier to only display bottom 16 bits.
10265 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10266 immediate to match for operand 2, since it will be masked.
10268 2013-05-24 Richard Biener <rguenther@suse.de>
10270 PR tree-optimization/57287
10271 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10272 all SSA names that occur in abnormal PHIs.
10274 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10277 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10278 that index is not negative.
10280 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10282 PR rtl-optimization/55177
10283 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10284 (simplify_byte_swapping_operation): New.
10285 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10286 (simplify_relational_operation_1): Deal with BSWAP.
10288 2013-05-23 Richard Henderson <rth@redhat.com>
10291 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10292 (ix86_reorg): Call it.
10294 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
10297 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10298 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10299 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10301 2013-05-23 Christian Bruel <christian.bruel@st.com>
10304 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10306 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10307 Marcus Shawcroft <marcus.shawcroft@arm.com>
10309 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10310 * config/aarch64/constraints.md (Usa): Remove.
10311 * doc/md.texi (AArch64 Usa): Remove.
10313 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10314 Marcus Shawcroft <marcus.shawcroft@arm.com>
10316 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10317 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10318 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10319 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10321 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
10323 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10324 instruction (AdvSIMD).
10325 * config/aarch64/aarch64-builtins.c
10326 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10327 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10329 2013-05-23 Martin Jambor <mjambor@suse.cz>
10331 PR middle-end/57347
10332 * tree.h (contains_bitfld_component_ref_p): Declare.
10333 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
10334 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10336 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10337 not access a bit-field. Assert all final offsets are byte-aligned.
10339 2013-05-23 Richard Biener <rguenther@suse.de>
10341 PR tree-optimization/57380
10342 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10343 least one invariant or re-used load.
10344 * passes.c (init_optimization_passes): Move pass_phiprop before
10347 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10349 * config/aarch64/aarch64-simd.md
10350 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10352 2013-05-23 Richard Biener <rguenther@suse.de>
10354 PR middle-end/57381
10355 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10356 OEP_CONSTANT_ADDRESS_OF retained.
10358 2013-05-23 Jakub Jelinek <jakub@redhat.com>
10360 PR middle-end/57344
10361 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10362 don't lower unit. Handle unit not being always BITS_PER_WORD.
10364 2013-05-23 Richard Biener <rguenther@suse.de>
10366 PR rtl-optimization/57341
10367 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
10368 instead of true_dependence.
10370 2013-05-22 David Malcolm <dmalcolm@redhat.com>
10372 * bb-reorder.c (branch_threshold): Make const.
10373 (exec_threshold): Ditto.
10375 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
10376 Pat Haugen <pthaugen@us.ibm.com>
10377 Peter Bergner <bergner@vnet.ibm.com>
10379 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
10380 documentation for the power8 crypto builtins.
10382 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
10384 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
10385 macros for defining power8 builtin functions.
10386 (BU_P8V_AV_2): Likewise.
10387 (BU_P8V_AV_P): Likewise.
10388 (BU_P8V_VSX_1): Likewise.
10389 (BU_P8V_OVERLOAD_1): Likewise.
10390 (BU_P8V_OVERLOAD_2): Likewise.
10391 (BU_CRYPTO_1): Likewise.
10392 (BU_CRYPTO_2): Likewise.
10393 (BU_CRYPTO_3): Likewise.
10394 (BU_CRYPTO_OVERLOAD_1): Likewise.
10395 (BU_CRYPTO_OVERLOAD_2): Likewise.
10396 (XSCVSPDP): Fix typo, point to the correct instruction.
10397 (VCIPHER): Add power8 crypto builtins.
10398 (VCIPHERLAST): Likewise.
10399 (VNCIPHER): Likewise.
10400 (VNCIPHERLAST): Likewise.
10401 (VPMSUMB): Likewise.
10402 (VPMSUMH): Likewise.
10403 (VPMSUMW): Likewise.
10404 (VPERMXOR_V2DI): Likewise.
10405 (VPERMXOR_V4SI: Likewise.
10406 (VPERMXOR_V8HI: Likewise.
10407 (VPERMXOR_V16QI: Likewise.
10408 (VSHASIGMAW): Likewise.
10409 (VSHASIGMAD): Likewise.
10410 (VPMSUM): Likewise.
10411 (VPERMXOR): Likewise.
10412 (VSHASIGMA): Likewise.
10414 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
10415 __CRYPTO__ if the crypto instructions are available.
10416 (altivec_overloaded_builtins): Add support for overloaded power8
10419 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
10420 support for power8 crypto builtins.
10421 (builtin_function_type): Likewise.
10422 (altivec_init_builtins): Add support for builtins that take vector
10423 long long (V2DI) arguments.
10425 * config/rs6000/crypto.md: New file, define power8 crypto
10428 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
10429 Pat Haugen <pthaugen@us.ibm.com>
10430 Peter Bergner <bergner@vnet.ibm.com>
10432 * doc/invoke.texi (Option Summary): Add power8 options.
10433 (RS/6000 and PowerPC Options): Likewise.
10435 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
10436 constraints.md instead of rs6000.h. Reorder w* constraints. Add
10437 wm, wn, wr documentation.
10439 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
10440 registers if direct move instructions are enabled.
10441 (wn): New constraint for no registers.
10442 (wq): New constraint for quad word even GPR registers.
10443 (wr): New constraint if 64-bit instructions are enabled.
10444 (wv): New constraint if power8 vector instructions are enabled.
10445 (wQ): New constraint for quad word memory locations.
10447 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
10448 constraint for 0..15 for crypto instructions.
10449 (gpc_reg_operand): If VSX allow registers in VSX registers as well
10450 as GPR and floating point registers.
10451 (int_reg_operand): New predicate to match only GPR registers.
10452 (base_reg_operand): New predicate to match base registers.
10453 (quad_int_reg_operand): New predicate to match even GPR registers
10454 for quad memory operations.
10455 (vsx_reg_or_cint_operand): New predicate to allow vector logical
10456 operations in both GPR and VSX registers.
10457 (quad_memory_operand): New predicate for quad memory operations.
10458 (reg_or_indexed_operand): New predicate for direct move support.
10460 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
10461 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
10462 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
10463 (POWERPC_MASKS): Add power8 options.
10464 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
10467 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10468 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
10470 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
10471 (-mpower8-fusion): New power8 options.
10472 (-mpower8-fusion-sign): Likewise.
10473 (-mpower8-vector): Likewise.
10474 (-mcrypto): Likewise.
10475 (-mdirect-move): Likewise.
10476 (-mquad-memory): Likewise.
10478 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
10480 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
10482 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
10483 (rs6000_debug_vector_unit): Add p8_vector.
10484 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
10485 definitions. Also print fusion state.
10486 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
10487 (rs6000_builtin_mask_calculate): Add power8 builtin support.
10488 (rs6000_option_override_internal): Add support for power8.
10489 (rs6000_common_init_builtins): Add debugging for skipped builtins
10490 if -mdebug=builtin.
10491 (rs6000_adjust_cost): Add power8 support.
10492 (rs6000_issue_rate): Likewise.
10493 (insn_must_be_first_in_group): Likewise.
10494 (insn_must_be_last_in_group): Likewise.
10495 (force_new_group): Likewise.
10496 (rs6000_register_move_cost): Likewise.
10497 (rs6000_opt_masks): Likewise.
10499 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
10500 power8 capable assembler, default to power7 options.
10501 (TARGET_DIRECT_MOVE): Likewise.
10502 (TARGET_CRYPTO): Likewise.
10503 (TARGET_P8_VECTOR): Likewise.
10504 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
10505 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
10506 (VECTOR_MEM_P8_VECTOR_P): Likewise.
10507 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
10508 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
10509 (TARGET_XSCVDPSPN): Likewise.
10510 (TARGET_XSCVSPDPN): Likewsie.
10511 (TARGET_SYNC_HI_QI): Likewise.
10512 (TARGET_SYNC_TI): Likewise.
10513 (MASK_CRYPTO): Likewise.
10514 (MASK_DIRECT_MOVE): Likewise.
10515 (MASK_P8_FUSION): Likewise.
10516 (MASK_P8_VECTOR): Likewise.
10517 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
10518 temporary used by some of the direct move instructions to get two FP
10519 temporary registers does not force creation of a stack frame.
10520 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
10521 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
10522 that any VSX registers are tieable, even if they are also an
10523 Altivec vector mode.
10524 (r6000_reg_class_enum): Add wm, wr, wv constraints.
10525 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
10526 (RS6000_BTM_CRYPTO): Likewise.
10527 (RS6000_BTM_COMMON): Likewise.
10529 * config/rs6000/rs6000.md (cpu attribute): Add power8.
10530 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
10531 (enum rs6000_vector): Add power8 vector support.
10533 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10537 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
10538 (any_sibcall_could_use_r3): this and handle indirect calls.
10539 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
10541 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10543 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
10545 2013-05-22 Richard Biener <rguenther@suse.de>
10547 PR middle-end/57349
10548 * profile.c (branch_prob): Do not split blocks that are
10549 abnormally receiving from ECF_RETURNS_TWICE functions.
10551 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10553 * recog.c (offsettable_address_addr_space_p): Fix calculation of
10554 address mode. Move pointer mode initialization to the same place.
10556 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10558 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
10559 while it has any effect.
10561 2013-05-21 Easwaran Raman <eraman@google.com>
10563 PR tree-optimization/57322
10564 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
10565 UID of the statement added to the BB to be 1.
10567 2013-05-21 Jakub Jelinek <jakub@redhat.com>
10569 PR tree-optimization/57331
10570 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
10571 of conversion from pointer type to integral type with integer.
10573 2013-05-21 Martin Jambor <mjambor@suse.cz>
10576 * ipa-prop.c (ipa_read_node_info): Process param_used and
10577 controlled_uses in the same order as when writing.
10579 2013-05-21 Magnus Granberg <baldrick@free.fr>
10582 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
10584 2013-05-21 Richard Biener <rguenther@suse.de>
10586 PR tree-optimization/57318
10587 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
10588 estimate stmts with side-effects as likely eliminated.
10590 2013-05-21 Richard Biener <rguenther@suse.de>
10592 PR tree-optimization/57330
10593 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
10594 preserve the call stmts fntype.
10596 2013-05-21 Richard Biener <rguenther@suse.de>
10598 PR tree-optimization/57303
10599 * tree-ssa-sink.c (statement_sink_location): Improve killing
10600 stmt detection and properly handle self-assignments.
10602 2013-05-21 Christian Bruel <christian.bruel@st.com>
10604 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
10605 spanning registers. LEAF_REG_REMAP is supported only for contiguous
10606 registers. Set register size out of the PARALLEL loop.
10608 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
10611 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
10612 (*fmasf4, *fmasf4_media): New insns.
10614 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10616 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
10617 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
10618 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
10619 (mips_idiv_insns): Update the comments to say that the returned
10620 instruction counts are in units of BASE_INSN_LENGTH.
10621 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
10622 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
10623 using 2 rather than 4 as the length of indirect MIPS16 and
10624 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
10625 length of a NOP. Don't divide MIPS16 lengths by 2.
10626 (mips16_split_long_branches): Assume a branch is long if the
10627 length is greater than 4 rather than 8.
10628 * config/mips/mips.md (length): Give MIPS16 lengths directly,
10629 rather than multiplying them by 2. Multiply instruction counts
10630 by BASE_INSN_LENGTH rather than 4.
10631 (*jump_mips16, tls_get_tp_mips16_<mode>)
10632 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
10634 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10636 * config/mips/mips.md (extended_mips16): Remove branch case.
10637 (length): Remove duplicated extended_mips16 test.
10639 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10641 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
10643 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
10645 * recog.h (Recog_data): Rename to...
10646 (recog_data_d): ...this.
10647 (recog_data): Update accordingly.
10648 * recog.c (recog_data): Likewise.
10649 * reload.c (save_recog_data): Likewise.
10650 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
10651 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
10653 2013-05-17 Julian Brown <julian@codesourcery.com>
10655 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
10656 found in a REG_EQUAL note, invalidate it.
10658 2013-05-17 Easwaran Raman <eraman@google.com>
10660 * tree-ssa-reassoc.c (find_insert_point): New function.
10661 (insert_stmt_after): Likewise.
10662 (get_def_stmt): Likewise.
10663 (ensure_ops_are_available): Likewise.
10664 (not_dominated_by): Likewise.
10665 (rewrite_expr_tree): Do not move statements beyond what is
10666 necessary. Remove call to swap_ops_for_binary_stmt...
10667 (reassociate_bb): ... and move it here.
10668 (build_and_add_sum): Assign UIDs for new statements.
10669 (linearize_expr): Likewise.
10670 (do_reassoc): Renumber gimple statement UIDs.
10672 2013-05-17 Jan Hubicka <jh@suse.cz>
10674 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
10676 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
10677 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
10679 (output_weakrefs): Update.
10681 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
10682 Martin Jambor <mjambor@suse.cz>
10684 PR middle-end/57276
10685 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
10686 value that corresponds to the given aggval is found in values vector.
10688 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
10690 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
10691 sse, sse2, sse3, ssse3 and sse4a flags to options.
10693 2013-05-17 David Malcolm <dmalcolm@redhat.com>
10695 * gengtype-state.c: (s_expr_writer): New class, to handle
10696 prettifying of output layout of s-expressions.
10697 (state_writer): New class, to write out gtype.state.
10698 (state_written_type_count): Move this variable into member data of
10700 (s_expr_writer::s_expr_writer): New code: constructor for new class
10701 (state_writer::state_writer(): ditto
10702 (s_expr_writer::write_new_line): New function
10703 (s_expr_writer::write_any_indent): ditto
10704 (s_expr_writer::begin_s_expr): ditto
10705 (s_expr_writer::end_s_expr): ditto
10706 (write_state_fileloc): convert to method of state_writer...
10707 (state_writer:: write_state_fileloc): ...and use methods of
10708 s_expr_writer to write indentation into the gtype.state output file
10709 to visually represent the hierarchical structure of the list
10711 (write_state_fields): ditto, renaming to...
10712 (state_writer::write_state_fields)
10713 (write_state_a_string): ditto, renaming to...
10714 (state_writer::write_state_a_string)
10715 (write_state_string_option): ditto, renaming to...
10716 (state_writer::write_state_string_option)
10717 (write_state_type_option): ditto, renaming to...
10718 (state_writer::write_state_type_option)
10719 (write_state_nested_option): ditto, renaming to...
10720 (state_writer::write_state_nested_option)
10721 (write_state_option): ditto, renaming to...
10722 (state_writer::write_state_option)
10723 (write_state_options): ditto, renaming to...
10724 (state_writer::write_state_options)
10725 (write_state_lang_bitmap): ditto, renaming to...
10726 (state_writer::write_state_lang_bitmap)
10727 (write_state_version): ditto, renaming to...
10728 (state_writer::write_state_version)
10729 (write_state_scalar_type): ditto, renaming to...
10730 (state_writer::write_state_scalar_type)
10731 (write_state_string_type): ditto, renaming to...
10732 (state_writer::write_state_string_type)
10733 (write_state_undefined_type): ditto, renaming to...
10734 (state_writer::write_state_undefined_type)
10735 (write_state_struct_union_type): ditto, renaming to...
10736 (state_writer::write_state_struct_union_type)
10737 (write_state_struct_type): ditto, renaming to...
10738 (state_writer::write_state_struct_type)
10739 (write_state_user_struct_type): ditto, renaming to...
10740 (state_writer::write_state_user_struct_type)
10741 (write_state_lang_struct_type): ditto, renaming to...
10742 (state_writer::write_state_lang_struct_type)
10743 (write_state_param_struct_type): ditto, renaming to...
10744 (state_writer::write_state_param_struct_type)
10745 (write_state_pointer_type): ditto, renaming to...
10746 (state_writer::write_state_pointer_type)
10747 (write_state_array_type): ditto, renaming to...
10748 (state_writer::write_state_array_type)
10749 (write_state_gc_used): ditto, renaming to...
10750 (state_writer::write_state_gc_used)
10751 (write_state_common_type_content): ditto, renaming to...
10752 (state_writer::write_state_common_type_content)
10753 (write_state_type): ditto, renaming to...
10754 (state_writer::write_state_type)
10755 (write_state_pair_list): ditto, renaming to...
10756 (state_writer::write_state_pair_list)
10757 (write_state_pair): ditto, renaming to...
10758 (state_writer::write_state_pair)
10759 (write_state_typedefs): ditto, renaming to...
10760 (state_writer::write_state_typedefs)
10761 (write_state_structures): ditto, renaming to...
10762 (state_writer::write_state_structures)
10763 (write_state_param_structs): ditto, renaming to...
10764 (state_writer::write_state_param_structs)
10765 (write_state_variables): ditto, renaming to...
10766 (state_writer::write_state_variables)
10767 (write_state_srcdir): ditto, renaming to...
10768 (state_writer::write_state_srcdir)
10769 (write_state_files_list): ditto, renaming to...
10770 (state_writer::write_state_files_list)
10771 (write_state_languages): ditto, renaming to...
10772 (state_writer::write_state_languages)
10773 (write_state): create a state_writer instance and use it when
10774 writing out the state file
10776 2013-05-17 Mike Stump <mikestump@comcast.net>
10778 PR rtl-optimization/57304
10779 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
10780 accessing DF_REF_REAL_LOC.
10782 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10784 PR rtl-optimization/57281
10785 PR rtl-optimization/57300
10786 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
10787 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
10788 what the other splitter did if the registers are dead.
10790 2013-05-17 Richard Biener <rguenther@suse.de>
10792 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
10795 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10797 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
10800 2013-05-17 Marek Polacek <polacek@redhat.com>
10802 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
10803 length when doing non-zero store of storing '\0' to '\0'.
10805 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10807 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
10808 vect_external_def oprnd1 with loop_vinfo, try to emit
10809 optional cast, negation and and stmts on the loop preheader
10810 edge instead of into the pattern def seq.
10812 PR tree-optimization/57051
10813 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
10814 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
10816 2013-05-16 Nick Clifton <nickc@redhat.com>
10818 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
10819 (rl78_is_naked_func): New function.
10820 (rl78_expand_prologue): Skip prologue generation for naked functions.
10821 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
10822 * doc/extend.texi (naked): Add RL78 to the list of processors
10823 that supports this attribute.
10825 2013-05-16 Jeff Law <law@redhat.com>
10827 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
10829 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
10831 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
10832 cache parameters using detect_caches_amd also for CYRIX,
10833 NSC and TM2 signatures.
10835 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
10836 Dzianis Kahanovich <mahatma@eu.by>
10840 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10841 VIA/Centaur processors and determine their cache parameters
10842 using detect_caches_amd.
10844 2013-05-16 Teresa Johnson <tejohnson@google.com>
10846 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
10847 (rtl_verify_edges): New function.
10848 (rtl_verify_bb_insns): Ditto.
10849 (rtl_verify_bb_pointers): Ditto.
10850 (rtl_verify_bb_insn_chain): Ditto.
10851 (rtl_verify_fallthru): Ditto.
10852 (rtl_verify_bb_layout): Ditto.
10853 (rtl_verify_flow_info_1): Outline checks into new functions.
10854 (rtl_verify_flow_info): Ditto.
10856 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
10858 * cfghooks.c (copy_bbs): Add update_dominance argument.
10859 * cfghooks.h (copy_bbs): Update prototype.
10860 * tree-cfg.c (gimple_duplicate_sese_region):
10861 Add update_dominance argument.
10862 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
10863 * tree-ssa-loop-ch.c (copy_loop_headers): Update
10864 gimple_duplicate_sese_region call.
10865 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
10866 Update copy_bbs call.
10867 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
10868 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
10870 2013-05-16 Jakub Jelinek <jakub@redhat.com>
10872 * tree-vectorizer.h (NUM_PATTERNS): Increment.
10873 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
10874 vect_recog_rotate_pattern.
10875 (vect_recog_rotate_pattern): New function.
10877 2013-05-16 Jason Merrill <jason@redhat.com>
10879 * Makefile.in (LLINKER): New variable.
10880 (mostlyclean): Remove link mutex.
10881 * configure.ac: Handle --enable-link-mutex.
10882 * lock-and-run.sh: New script.
10884 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10887 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
10890 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10892 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
10894 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
10896 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
10897 * config/arm/arm.c (next_consecutive_mem): New function.
10898 (gen_movmem_ldrd_strd): Likewise.
10899 * config/arm/arm.md (movmemqi): Update condition and code.
10900 (unaligned_loaddi, unaligned_storedi): New patterns.
10902 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10904 * config.gcc: Obsolete *-*-solaris2.9*.
10905 * doc/install.texi (Specific, *-*-solaris2*): Document it.
10907 2013-05-16 Richard Biener <rguenther@suse.de>
10909 * passes.c (init_optimization_passes): Move pass_parallelize_loops
10910 earlier, after GRAPHITE transforms and IV canonicalization.
10912 2013-05-16 Jakub Jelinek <jakub@redhat.com>
10914 * omp-low.c (extract_omp_for_data): For collapsed loops,
10915 if at least one of the loops is known at compile time to
10916 iterate zero times, set count to 0.
10917 (expand_omp_regimplify_p): New function.
10918 (expand_omp_for_generic): For collapsed loops, if at least
10919 one of the loops isn't known to iterate at least once,
10920 add runtime check with setting count to 0.
10921 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
10922 For unsigned types if it isn't known at compile time that
10923 the loop will iterate at least once, add runtime check to bypass
10924 the whole loop if initial condition isn't true.
10926 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
10928 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
10930 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
10932 PR middle-end/57286
10933 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
10934 transformations to avoid an infinite loop.
10936 2013-05-16 Marek Polacek <polacek@redhat.com>
10938 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
10940 2013-05-15 Leif Ekblad <leif@rdos.net>
10942 * config/i386/i386.c (ix86_decompose_address): Use
10943 DEFAULT_TLS_SEG_REG to access TLS segment register.
10944 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
10945 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
10946 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
10948 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
10951 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
10952 sibling calls to functions that would normally be lazily bound,
10953 unless $gp is call-clobbered.
10955 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
10957 * config/i386/i386.c (ix86_option_override_internal): Update
10958 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
10959 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
10960 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
10961 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
10963 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
10965 2013-05-15 Andreas Schwab <schwab@suse.de>
10967 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
10968 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
10971 2013-05-15 Teresa Johnson <tejohnson@google.com>
10973 * loop-unroll.c (report_unroll_peel): Check decision before
10974 emitting unroll/peel message.
10976 2013-05-15 Teresa Johnson <tejohnson@google.com>
10978 * function.h (has_bb_partition): New rtl_data flag.
10979 (bb_reorder_complete): Ditto.
10980 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
10981 instead of flag_reorder_blocks_and_partition.
10982 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
10983 with some enhancements.
10984 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
10985 * bb-reorder.c (connect_traces): Check for has_bb_partition
10986 instead of flag_reorder_blocks_and_partition.
10987 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
10988 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
10989 verify_hot_cold_block_grouping.
10990 (partition_hot_cold_basic_blocks): Set has_bb_partition.
10992 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10995 * config/arm/predicates.md (call_insn_operand): New predicate.
10996 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
10997 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
10998 if insn is not a tail call.
10999 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11001 * config/arm/arm.h (enum reg_class): New caller save register class.
11002 (REG_CLASS_NAMES): Likewise.
11003 (REG_CLASS_CONTENTS): Likewise.
11004 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11007 2013-05-15 Richard Biener <rguenther@suse.de>
11009 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11010 of MSG_OPTIMIZED_LOCATIONS.
11011 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11012 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11014 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11015 of MSG_OPTIMIZED_LOCATIONS.
11016 (execute_vect_slp): Likewise.
11017 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11018 (vect_create_cond_for_alias_checks): Likewise.
11019 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11020 (vect_recog_widen_mult_pattern): Likewise.
11021 (vect_recog_widen_sum_pattern): Likewise.
11022 (vect_recog_over_widening_pattern): Likewise.
11023 (vect_recog_widen_shift_pattern): Likewise.
11024 (vect_recog_vector_vector_shift_pattern): Likewise.
11025 (vect_recog_divmod_pattern): Likewise.
11026 (vect_recog_mixed_size_cond_pattern): Likewise.
11027 (vect_recog_bool_pattern): Likewise.
11028 (vect_pattern_recog_1): Likewise.
11030 2013-05-15 Martin Jambor <mjambor@suse.cz>
11032 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11033 non-functions to builtin_unreachable.
11034 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11035 correct when new direct edges were discovered.
11037 2013-05-15 Martin Jambor <mjambor@suse.cz>
11039 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11040 header, print symbol order instead of node uid, print more information
11041 about indirect edge targets.
11042 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11044 (ipa_make_edge_direct_to_target): Likewise.
11045 (remove_described_reference): Likewise.
11046 (propagate_controlled_uses): Likewise.
11047 (ipa_print_node_params): Also print symbol order.
11048 (ipcp_transform_function): Print symbol order instead of node uids.
11049 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11050 (cgraph_get_create_real_symbol_node): Likewise.
11051 * ipa-cp.c (print_lattice): Likewise.
11052 (print_all_lattices): Likewise.
11053 (determine_versionability): Likewise.
11054 (initialize_node_lattices): Likewise.
11055 (estimate_local_effects): Likewise.
11056 (update_profiling_info): Likewise.
11057 (create_specialized_node): Likewise.
11058 (perhaps_add_new_callers): Likewise.
11059 (decide_about_value): Likewise.
11060 (decide_whether_version_node): Likewise.
11061 (identify_dead_nodes): Likewise.
11062 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11063 (dump_inline_summary): Likewise.
11064 (estimate_node_size_and_time): Likewise.
11065 (inline_analyze_function): Likewise.
11066 * ipa-inline.c (report_inline_failed_reason): Likewise.
11067 (want_early_inline_function_p): Likewise.
11068 (edge_badness): Likewise.
11069 (update_edge_key): Likewise.
11070 (inline_small_functions): Likewise. Add dumping of order to two other
11072 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11073 instead of node uids.
11074 (propagate_pure_const): Likewise.
11075 (propagate_pure_const): Likewise.
11076 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11077 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11078 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11080 * tree-pretty-print.c (dump_function_header): Likewise.
11081 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11082 Print symbol order instead of node uids.
11084 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11086 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11087 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11089 2013-05-15 Richard Biener <rguenther@suse.de>
11091 PR tree-optimization/57275
11092 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11093 return value for fail to do runtime alias checks for gather loads.
11095 2013-05-15 Jan Hubicka <jh@suse.cz>
11099 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11100 weakrefs are not external.
11101 (lto_symtab_merge_decls): Fix thinko when dealing with
11102 non-lto_symtab decls.
11103 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11104 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11105 * varpool.c (dump_varpool_node): Dump more flags.
11107 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11109 * config/i386/i386.c (processor_alias_table): Add instruction
11110 FSGSBASE for AMD bdver3 architecture.
11112 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11114 * tree.c (warn_deprecated_use): Print file:line using locus color.
11115 * diagnostic.c (diagnostic_report_current_module): Print file:line
11116 and file:line:column using locus color.
11118 2013-05-14 Mike Stump <mikestump@comcast.net>
11120 * gdbinit.in: Add __null.
11122 2013-05-14 Mike Stump <mikestump@comcast.net>
11124 * recog.h: Rename struct recog_data to Recog_data.
11125 * recog.c: Likewise.
11126 * reload.c (can_reload_into): Likewise.
11127 * config/picochip/picochip.c: Likewise.
11129 2013-05-14 Mike Stump <mikestump@comcast.net>
11131 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11133 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11135 * resource.h (struct resources): Remove unch_memory member.
11136 (CLEAR_RESOURCE): Don't clear unch_memory.
11137 * resource.c (mark_referenced_resources): Don't set it.
11138 (mark_set_resources): Likewise.
11139 (mark_target_live_regs): Don't clear it.
11140 (init_resource_info): Likewise.
11141 * reorg.c (resource_conflicts_p): Don't compare it.
11142 (redundant_insn): Don't set it.
11144 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11146 * emit-rtl.c (next_label): Remove unused function.
11147 (skip_consecutive_labels, link_cc0_insns): Move to ...
11148 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11149 only place where these functions are used, and make them static.
11151 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11153 * fold-const.c (fold_negate_expr): Handle vectors.
11154 (fold_truth_not_expr): Make it static.
11155 (fold_invert_truthvalue): New static function.
11156 (invert_truthvalue_loc): Handle vectors. Do not call
11157 fold_truth_not_expr directly.
11158 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11159 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11160 (fold_comparison): Handle vector constants.
11161 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11162 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11163 * tree.h (fold_truth_not_expr): Remove declaration.
11165 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11167 * config/aarch64/aarch64-simd.md
11168 (aarch64_vcond_internal<mode>): Rename to...
11169 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11170 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11171 float modes. Clarify all iterator modes.
11172 (vcond<mode><mode>): Use new name for vcond expanders.
11173 (vcond<v_cmp_result><mode>): Likewise.
11174 (vcondu<mode><mode>: Likewise.
11175 * config/aarch64/iterators.md (VDQF_COND): New.
11177 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11180 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11181 variable for the shift amount. Check that we shift by non-negative
11184 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11187 * config/arm/arm.h (EPILOGUE_USES): Only return true
11188 for LR_REGNUM after epilogue_completed.
11190 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
11192 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
11193 is error_mark_node.
11195 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11198 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11199 and Solaris 11+/x86 with gld.
11200 * configure: Regenerate.
11202 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11204 * expmed.c (expand_shift_1): Canonicalize rotates by
11205 constant bitsize / 2 to bitsize - 1.
11206 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
11207 case ROTATERT>: Likewise.
11210 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11212 * config/i386/i386.md (rotateinv): New code attr.
11213 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11214 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11215 roll $31, %eax, etc.
11217 2013-05-14 Richard Biener <rguenther@suse.de>
11219 PR middle-end/57235
11220 * tree-eh.c (sink_clobbers): Give up for successors with
11221 multiple predecessors and no virtual uses.
11223 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11225 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11226 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11228 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11230 PR middle-end/57251
11231 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11232 the case when both op0 and op1 have VOIDmode.
11234 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11236 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11237 in multiply-accumulate mode.
11239 2013-05-13 Guozhi Wei <carrot@google.com>
11241 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11243 2013-05-13 Kai Tietz <ktietz@redhat.com>
11246 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
11247 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
11248 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
11249 * config/i386/i386.c (ix86_option_override_internal): Likewise.
11250 (ix86_expand_prologue): Likewise.
11251 (ix86_expand_split_stack_prologue): Likewise.
11252 (legitimate_pic_address_disp_p): Likewise.
11253 (legitimize_pic_address): Likewise.
11254 (legitimize_tls_address): Likewise.
11255 (legitimize_pe_coff_symbol): Likewise.
11256 (output_pic_addr_const): Likewise.
11257 (construct_plt_address): Likewise.
11258 (ix86_expand_call): Likewise.
11259 (x86_output_mi_thunk): Likewise.
11260 (x86_function_profiler): Likewise.
11262 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11264 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11265 similar switch cases.
11266 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11267 (aarch64_simd_mov_to_<mode>low): Delete.
11268 (aarch64_simd_mov_to_<mode>high): Delete.
11269 (move_lo_quad_<mode>): Add w<-r alternative.
11270 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11271 (aarch64_simd_mov_from_*): Update type attribute.
11272 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11275 2013-05-13 Jan Hubicka <jh@suse.cz>
11277 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11278 * config/i386/i386.c (ix86_compute_frame_layout,
11279 ix86_expand_epilogue, emit_i387_cw_initialization,
11280 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11281 ix86_local_alignment): Fix use of size/speed predicates.
11283 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11285 PR tree-optimization/45216
11286 PR tree-optimization/57157
11287 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11288 the (-Y) & (B - 1) variant if OP is |.
11289 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11290 return shifted. Use (-op1) & (prec - 1) as other_amount
11291 instead of prec - op1.
11293 2013-05-13 Martin Jambor <mjambor@suse.cz>
11295 PR middle-end/42371
11296 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11297 (ipa_constant_data): New type.
11298 (ipa_jump_func): Use ipa_constant_data to hold information about
11299 constant jump functions.
11300 (ipa_get_jf_constant): Adjust to jump function type changes.
11301 (ipa_get_jf_constant_rdesc): New function.
11302 (ipa_param_descriptor): New field controlled_uses.
11303 (ipa_get_controlled_uses): New function.
11304 (ipa_set_controlled_uses): Likewise.
11305 * ipa-ref.h (ipa_find_reference): Declare.
11306 * ipa-prop.c (ipa_cst_ref_desc): New type.
11307 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11309 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11310 New parameter cs. Adjust all callers.
11311 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11312 (remove_described_reference): New function.
11313 (jfunc_rdesc_usable): Likewise.
11314 (try_make_edge_direct_simple_call): Decrement controlled use count,
11315 attempt to remove reference if it hits zero.
11316 (combine_controlled_uses_counters): New function.
11317 (propagate_controlled_uses): Likewise.
11318 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11319 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11320 (ipa_print_node_params): Print described use counter.
11321 (ipa_write_jump_function): Adjust to jump function type changes.
11322 (ipa_read_jump_function): New parameter CS, pass it to
11323 ipa_set_jf_constant. Adjust caller.
11324 (ipa_write_node_info): Stream controlled use count
11325 (ipa_read_node_info): Likewise.
11326 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11328 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11329 count. Remove cloning-added reference if it reaches zero.
11330 * ipa-ref.c (ipa_find_reference): New function.
11332 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11334 * config/i386/i386.c (processor_target_table): Modified default
11335 alignment values for AMD BD and BT architectures.
11337 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
11339 * tree-vect-generic.c (uniform_vector_p): Move ...
11340 * tree.c (uniform_vector_p): ... here.
11341 * tree.h (uniform_vector_p): Declare it.
11342 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11345 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11347 PR tree-optimization/57230
11348 * tree-ssa-strlen.c (handle_char_store): Record length for
11349 array store from STRING_CST.
11351 PR tree-optimization/57230
11352 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11355 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11357 * config/epiphany/epiphany.c (epiphany_init): Check size of
11358 NUM_MODES_FOR_MODE_SWITCHING.
11359 (epiphany_expand_prologue):
11360 Remove CONFIG_REGNUM initial value handling code.
11361 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11362 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
11363 (emit_set_fp_mode, epiphany_mode_after): Likewise.
11364 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
11365 Don't return 1 for FP_MODE_NONE.
11366 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
11367 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
11368 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
11369 * config/epiphany/epiphany.md (save_config): New pattern.
11371 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
11373 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
11375 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
11377 * config/i386/i386.md (memory): Handle sseishft1.
11378 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
11379 (*vec_extractv2di_1): Ditto.
11381 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
11383 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
11386 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
11388 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
11389 Add mthumb/march=armv7-a multilib.
11390 Add mthumb/march=armv7-r multilib.
11391 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
11393 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
11395 * config/v850/t-rtems: Add more multilibs.
11397 2013-05-10 Richard Biener <rguenther@suse.de>
11399 PR tree-optimization/57214
11400 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
11401 not propagate from SSA names that occur in abnormal PHI nodes.
11403 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
11405 * stor-layout.c (element_precision): New function.
11406 * machmode.h (element_precision): Declare it.
11407 * tree.c (build_minus_one_cst): New function.
11408 (element_precision): Likewise.
11409 * tree.h (build_minus_one_cst): Declare new function.
11410 (element_precision): Likewise.
11411 * fold-const.c (operand_equal_p): Use element_precision.
11412 (fold_binary_loc): Handle vector types.
11413 * convert.c (convert_to_integer): Use element_precision.
11414 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
11417 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
11419 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
11420 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
11421 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
11422 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
11423 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
11424 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
11425 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
11426 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
11427 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
11428 (Uuw8): New constraints.
11429 (Usb4): Move into alphabetical order.
11430 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
11431 (sd8_operand, ub8_operand, uw8_operand): New predicates.
11432 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
11433 previously unnamed patterns.
11434 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
11435 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
11436 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
11437 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
11438 of set_attr_alternative/if_then_else. Use extended_mips16 instead
11439 of specific lengths.
11441 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11443 * config/i386/i386.md (rotateinv): New code attr.
11444 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11445 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11446 roll $31, %eax, etc.
11448 PR tree-optimization/45216
11449 PR tree-optimization/57157
11450 * tree-ssa-forwprop.c (simplify_rotate): New function.
11451 (ssa_forward_propagate_and_combine): Call it.
11453 2013-05-10 Richard Biener <rguenther@suse.de>
11455 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
11456 disable peeling when we version for aliasing.
11457 (vector_alignment_reachable_p): Honor explicit user alignment.
11458 (vect_supportable_dr_alignment): Likewise.
11459 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
11460 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
11461 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
11462 then peeling to arrange for the cost-model check to come first.
11464 2013-05-10 Alan Modra <amodra@gmail.com>
11466 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
11467 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
11468 * configure: Regenerate.
11470 2013-05-10 Alan Modra <amodra@gmail.com>
11473 * varasm.c (default_elf_select_section): Move !DECL_P check..
11474 (get_named_section): ..to here before calling get_section_name.
11476 (default_section_type_flags): Add DECL_P check.
11477 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
11478 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
11480 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
11482 * config/epiphany/epiphany.c (epiphany_expand_prologue):
11483 When using gen_stack_adjust_str with a register offset, add a
11484 REG_FRAME_RELATED_EXPR note.
11486 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
11488 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
11489 (*vec_extractv4si_zext_mem): Ditto.
11490 (*vec_extractv2di): Add 0->x and x->x alternatives.
11491 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
11492 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
11494 2013-05-09 Jason Merrill <jason@redhat.com>
11496 N3639 C++1y VLA support
11497 * gimplify.c (gimplify_vla_decl): Don't touch an existing
11500 * tree.c (build_constructor_va): New.
11501 * tree.h: Declare it.
11503 2013-05-09 Martin Jambor <mjambor@suse.cz>
11506 * gimple-fold.c (canonicalize_constructor_val): Call
11507 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
11509 2013-05-09 Jan Hubicka <jh@suse.cz>
11510 Richard Biener <rguenther@suse.de>
11513 * symtab.c (symtab_make_decl_local): Do not add private names.
11515 2013-05-09 Jan Hubicka <jh@suse.cz>
11518 * symtab.c (insert_to_assembler_name_hash): Handle clones.
11519 (unlink_from_assembler_name_hash): Likewise.
11520 (symtab_prevail_in_asm_name_hash, symtab_register_node,
11521 symtab_unregister_node, symtab_initialize_asm_name_hash,
11522 change_decl_assembler_name): Update.
11524 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
11526 * config/aarch64/aarch64.md: New movtf split.
11527 (*movtf_aarch64): Update.
11528 (aarch64_movdi_tilow): Handle TF modes and rename to
11529 aarch64_movdi_<mode>low.
11530 (aarch64_movdi_tihigh): Handle TF modes and rename to
11531 aarch64_movdi_<mode>high
11532 (aarch64_movtihigh_di): Handle TF modes and rename to
11533 aarch64_mov<mode>high_di
11534 (aarch64_movtilow_di): Handle TF modes and rename to
11535 aarch64_mov<mode>low_di
11536 (aarch64_movtilow_tilow): Remove spurious whitespace.
11537 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
11539 (aarch64_print_operand): Update.
11541 2013-05-09 Alan Modra <amodra@gmail.com>
11543 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
11545 * configure: Regenerate.
11547 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
11549 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
11550 splitter preparation statements.
11551 * config/i386/sse.md (*vec_extract* splitters): Ditto.
11552 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
11555 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11557 * gimple-ssa-strength-reduction.c (count_candidates): Change
11558 return value to int.
11559 (analyze_candidates_and_replace): Change type of length to int.
11561 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
11563 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
11564 (*vec_extract<mode>): Use VI12_128 mode iterator.
11565 (*vec_extract<mode>_mem): Ditto.
11566 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
11569 2013-05-08 Diego Novillo <dnovillo@google.com>
11574 2012-08-17 Diego Novillo <dnovillo@google.com>
11577 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
11578 * config.in: Regenerate.
11579 * configure: Regenerate.
11580 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
11582 2013-05-08 Jan Hubicka <jh@suse.cz>
11585 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
11586 * cgraph.h (symtab_node_base): Add unique_name.
11587 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11588 input_overwrite_node, input_varpool_node): Stream unique_name.
11589 * cgraphclones.c (cgraph_create_virtual_clone,
11590 cgraph_function_versioning): Set unique_name.
11591 * ipa.c (function_and_variable_visibility): Set unique_name.
11593 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11595 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
11596 (alloc_cand_and_find_basis): Restrict conditional candidate
11597 processing to CAND_MULTs.
11599 2013-05-08 Jan Hubicka <jh@suse.cz>
11602 lto-symtab.c (lto_symtab_symbol_p): New function.
11603 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
11604 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
11605 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
11606 Skip static symbols.
11608 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
11610 PR tree-optimization/57200
11611 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
11612 Only call inform if the preceding warning_at returns true.
11614 2013-05-07 Han Shen <shenhan@google.com>
11616 * cfgexpand.c (record_or_union_type_has_array_p): New function.
11617 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
11618 * common.opt (fstack-protector-strong): New option.
11619 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
11620 * doc/invoke.texi (Optimization Options): Document
11621 "-fstack-protector-strong".
11622 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
11624 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11626 * config/mips/mips.c (mips_machine_reorg2): Return 0.
11628 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
11630 * ira.c (update_equiv_regs): Add insn having equiv memory even if
11631 it is not lhs of the insn.
11632 (setup_reg_equiv): Remove insn having equiv memory which it is not
11634 * lra-constraints.c (process_address): Try to improve generation
11635 code for address base + disp.
11636 (lra_constraints): Make correct the code for checking insn setting
11637 up backward equivalence. Remove insn only if it is in the init
11639 * lra-eliminations.c (update_reg_eliminate): Change return value.
11640 (lra_eliminate): Use the result.
11642 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
11644 * config/i386/sse.md (ssescalarnummask): New mode attribute.
11645 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
11646 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
11647 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
11648 register target operands.
11649 (*vec_extractv8hi_sse2): New pattern.
11650 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
11651 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
11652 (*vec_extract<mode>_mem): New insn and split pattern.
11654 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
11656 * config/arm/arm.c (arm_asan_shadow_offset): New function.
11657 (TARGET_ASAN_SHADOW_OFFSET): Define.
11658 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
11659 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
11661 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11663 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
11664 (incr_vec_index): Return -1 if increment not found.
11665 (create_add_on_incoming_edge): Assert if increment not found.
11666 (record_increment): Limit number of increments recorded.
11667 (all_phi_incrs_profitable): Return false if an increment not found.
11668 (replace_profitable_candidates): Don't process increments that were
11670 (analyze_candidates_and_replace): Limit size of incr_vec.
11672 2013-05-07 Richard Biener <rguenther@suse.de>
11674 * calls.c (special_function_p): setjmp-like functions are leaf.
11675 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
11676 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
11678 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
11680 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
11681 (aarch64_simd_mov<mode>): New expander.
11682 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
11683 (aarch64_simd_mov_to_<mode>high): Likewise.
11684 (aarch64_simd_mov_from_<mode>low): Likewise.
11685 (aarch64_simd_mov_from_<mode>high): Likewise.
11686 (aarch64_dup_lane<mode>): Update.
11687 (aarch64_dup_lanedi): New instruction pattern.
11688 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
11689 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
11691 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11693 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
11694 (replace_mult_candidate): Remove unnecessary argument; remove
11695 unnecessary parameter from call to introduce_cast_before_cand.
11696 (replace_unconditional_candidate): Remove unnecessary parameter
11697 from call to replace_mult_candidate.
11698 (replace_conditional_candidate): Likewise.
11699 (insert_initializers): Use make_temp_ssa_name.
11700 (introduce_cast_before_cand): Remove unnecessary argument; use
11701 make_temp_ssa_name.
11702 (replace_one_candidate): Remove unnecessary argument; remove
11703 unnecessary parameter from calls to introduce_cast_before_cand.
11704 (replace_profitable_candidates): Remove unnecessary parameters
11705 from calls to replace_one_candidate.
11707 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11709 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
11710 phi def as possibly hiding a basis for a CAND_ADD whose operands
11711 have been commuted in the analysis.
11712 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
11714 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11716 * config/aarch64/aarch64.md
11717 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
11718 shift value between 0-4.
11720 2013-05-07 Richard Biener <rguenther@suse.de>
11722 * double-int.h (rshift): New overload.
11723 * double-int.c (rshift): New function.
11724 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
11725 (create_reference_ops_from_ref): Remove.
11726 (vn_reference_insert): Use shared ops for constructing the
11727 reference and copy it.
11729 2013-05-07 Richard Biener <rguenther@suse.de>
11731 PR middle-end/57190
11732 * tree-eh.c (sink_clobbers): Properly propagate
11733 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
11735 2013-05-07 Jakub Jelinek <jakub@redhat.com>
11737 PR tree-optimization/57149
11738 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
11739 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
11740 collect_phi_def_edges, execute_late_warn_uninitialized): Use
11741 uninit_undefined_value_p instead of ssa_undefined_value_p.
11744 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
11745 for modifier == EXPAND_INITIALIZER.
11747 2013-05-07 Anton Blanchard <anton@samba.org>
11749 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
11750 for powerpc64 little endian.
11751 * configure: Regenerate.
11753 2013-05-06 Graham Stott <grahams@btinternet.com>
11755 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
11756 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
11757 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
11758 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
11760 2013-05-06 Graham Stott <grahams@btinternet.com>
11762 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
11763 codes which allow non-lvalues.
11765 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
11767 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
11768 components are all 1s.
11769 (integer_minus_onep): New function.
11770 * tree.h (integer_minus_onep): Declare it.
11771 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
11772 integer_minus_onep instead of integer_all_onesp.
11774 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11777 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
11778 variations of these patterns.
11780 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
11782 * config/i386/i386.md (isa): Add x64_sse4 member.
11783 (enabled): Handle x64_sse4.
11784 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
11785 instruction for 64bit SSE4_1 targets. Update insn attributes.
11786 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
11787 instruction for SSE4_1 targets. Update insn attributes.
11788 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11789 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
11790 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
11792 (*vec_extractv4si): Rename from *sse4_1_pextrd.
11793 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
11794 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
11796 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11799 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
11801 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
11803 * final.c (do_assembler_dialects): Don't handle curly braces and
11804 vertical bar escaped by % as dialect delimiters.
11805 (output_asm_insn): Print curly braces and vertical bar if escaped
11806 by % and ASSEMBLER_DIALECT defined.
11807 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
11808 * doc/tm.texi: Regenerated.
11810 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11812 * config/mips/mips.c: Include tree-pass.h.
11813 (mips_reorg): Split in pre- and post-dbr_schedule parts.
11814 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
11815 (pass_mips_machine_reorg2): New machine specific pass.
11816 (insert_pass_mips_machine_reorg2): New pass plugin definition.
11817 (mips_option_override): Register the new pass.
11818 * rtl.h (cleanup_barriers): Remove prototype.
11819 (dbr_schedule): Likewise.
11820 * jump.c (cleanup_barriers): Make static.
11821 * reorg.c (dbr_schedule): Likewise.
11823 2013-05-06 Richard Biener <rguenther@suse.de>
11825 PR tree-optimization/57185
11826 * tree-parloops.c (add_field_for_reduction): Handle anonymous
11827 SSA names properly.
11829 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
11832 * config/i386/i386.c (add_parameter_dependencies): Add dependence
11833 between "first_arg" and "insn", not "last" and "insn".
11835 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11837 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
11838 (find_candidates_in_block): Re-enable slsr_process_phi.
11839 (create_phi_basis): Fix double counting of candidate adjustment.
11841 2013-05-06 Richard Biener <rguenther@suse.de>
11843 PR middle-end/57147
11844 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
11845 the edge is also fallthru, preserve it and just clear the
11847 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
11848 also complex, preserve that and just clear the fallthru flag.
11849 * tree-inline.c (update_ssa_across_abnormal_edges): Also
11850 update virtual operands.
11852 2013-05-06 Alan Modra <amodra@gmail.com>
11854 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
11855 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
11856 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
11857 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
11858 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11859 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
11861 2013-05-06 Alan Modra <amodra@gmail.com>
11863 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
11864 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
11865 (DEFAULT_ASM_ENDIAN): Define.
11866 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11867 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
11868 Update -K PIC clause from sysv4.h.
11869 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
11870 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
11872 2013-05-06 Alan Modra <amodra@gmail.com>
11874 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
11875 twice for little-endian.
11876 (ashrdi3_no_power, ashrdi3): Support little-endian.
11878 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11881 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
11882 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
11884 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
11885 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
11886 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
11889 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
11890 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11892 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
11893 * config/pa/pa.opt: Make mbig-switch a no-op.
11894 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
11895 (CASE_VECTOR_MODE): Always return SImode.
11896 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
11897 for the !TARGET_BIG_SWITCH case.
11898 * config/pa/pa-linux.h: Likewise.
11899 * config/pa/pa-openbsd.h: Likewise.
11900 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
11901 * config/pa/pa.md (short_jump): Remove define_insn.
11902 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
11903 (casesi0): Remove define_insn.
11904 (type): Remove btable_branch.
11905 (pa_combine_type): Likewise.
11906 (in_nullified_branch_delay): Likewise.
11907 (in_call_delay): Likewise.
11908 (define_delay): Likewise.
11909 (define_insn_reservation "Z3"): Likewise.
11910 (define_insn_reservation "Z4"): Likewise.
11911 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
11912 (pa_adjust_insn_length): Remove adjustment for btable branches.
11913 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
11916 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
11918 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11919 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
11920 Add m->r,x alternatives.
11921 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
11922 splitters using SWI48x mode iterator.
11923 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
11924 TARGET_64BIT. Add m->x alternative.
11925 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
11926 Add o->x alternative. Enable for TARGET_SSE.
11927 (sse_storeq): Remove expander.
11928 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
11929 with memory input operand.
11930 (*vec_extractv2di_1 splitter): New.
11931 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
11932 * config/i386/i386.md (ssevecmodelower): New mode attribute.
11934 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11936 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
11937 (INT_LOWPART): Delete.
11938 (extract_MB): Adjust.
11939 (extract_ME): Adjust.
11940 (print_operand): Adjust.
11942 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11944 * config/rs6000/predicates.md (reg_or_add_cint_operand,
11945 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11946 (reg_or_logical_cint_operand, easy_fp_constant,
11947 logical_const_operand): Delete "CONST_DOUBLE" case.
11948 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
11949 "HOST_BITS_PER_WIDE_INT == 64" test.
11950 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
11951 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
11952 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
11953 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
11955 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
11956 CONST_DOUBLE DImode/VOIDmode case.
11957 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
11958 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11959 CONST_DOUBLE VOIDmode case.
11960 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11961 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
11962 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
11963 Delete CONST_DOUBLE case.
11964 (splitters for mov FMOVE64 const_double): Delete
11965 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11966 "HOST_BITS_PER_WIDE_INT >= 64" test.
11967 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
11969 (mov DI const_double): Delete.
11971 2013-05-04 Jakub Jelinek <jakub@redhat.com>
11973 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
11974 on op shows all bits zero in mode of a lowpart subreg, return zero.
11976 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
11979 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
11980 to save TFmode registers and DImode to save TImode registers for
11981 caller save operations.
11982 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
11983 mark being partially clobbered since they only use the first
11986 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
11987 and TDmode only use the upper 64-bits of each VSX register.
11989 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11991 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
11992 (find_candidates_in_block): Disable slsr_process_phi.
11994 2013-05-03 Guozhi Wei <carrot@google.com>
11996 * coverage.c (coverage_obj_init): Move the construction of gcov
11998 (build_init_ctor): ... here.
12000 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12002 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12003 (slsr_cand_d): Redefine def_phi.
12004 (stride_status, phi_adjust_status, count_phis_status): New enums.
12005 (find_phi_def): New.
12006 (find_basis_for_base_expr): New.
12007 (find_basis_for_candidate): Handle hidden bases.
12008 (alloc_cand_and_find_basis): Handle phi candidates.
12009 (slsr_process_phi): New.
12010 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12011 (create_mul_imm_cand): Likewise.
12012 (create_add_ssa_cand): Exclude phi base candidates.
12013 (create_add_imm_cand): Likewise.
12014 (slsr_process_cast): Likewise.
12015 (slsr_process_copy): Likewise.
12016 (find_candidates_in_block): Handle phi candidates.
12017 (dump_candidate): Likewise.
12018 (unconditional_cands): Delete.
12019 (unconditional_cands_with_known_stride_p): Delete.
12020 (phi_dependent_cand_p): New.
12021 (cand_increment): Handle phi-dependent candidates.
12022 (replace_dependent): Delete.
12023 (replace_mult_candidate): New.
12024 (replace_unconditional_candidate): New.
12025 (incr_vec_index): Move to avoid forward reference.
12026 (create_add_on_incoming_edge): New.
12027 (create_phi_basis): New.
12028 (replace_dependents): Delete.
12029 (replace_conditional_candidate): New.
12030 (phi_add_costs): New.
12031 (replace_uncond_cands_and_profitable_phis): New.
12032 (record_increment): Handle phi adjustments.
12033 (record_phi_increments): New.
12034 (record_increments): Handle phi adjustments.
12035 (phi_incr_cost): New.
12036 (lowest_cost_path): Handle phis.
12037 (total_savings): Likewise.
12038 (analyze_increments): Likewise.
12039 (ncd_with_phi): New.
12040 (ncd_of_cand_and_phis): New.
12041 (nearest_common_dominator_for_cands): Handle phi increments.
12042 (all_phi_incrs_profitable): New.
12043 (replace_profitable_candidates): Handle phi-dependent candidates.
12044 (analyze_candidates_and_replace): Likewise.
12046 2013-05-03 Teresa Johnson <tejohnson@google.com>
12049 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12050 do not exceed REG_BR_PROB_BASE.
12052 2013-05-03 Jeff Law <law@redhat.com>
12054 PR tree-optimization/57144
12055 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12056 operand of the condition will bit into the new type when eliminating
12057 a cast feeding a condition.
12059 2013-05-03 Jakub Jelinek <jakub@redhat.com>
12061 PR rtl-optimization/57130
12062 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12063 of COMPARE as in_code to the recursive call if needed.
12065 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
12067 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12068 (enabled): Handle new members.
12069 * config/i386/sse.md (*vec_concatv2si): Merge from
12070 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12071 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12073 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
12075 PR tree-optimization/57027
12076 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12077 for fnms opportunity, check we got the prerequisite kind
12078 of tree / gimple before using accessor functions.
12080 2013-05-03 Richard Biener <rguenther@suse.de>
12082 * double-int.h (lshift): New overload without precision
12083 and arith argument.
12084 (operator *=, operator +=, operator -=): Move ...
12085 * double-int.c (operator *=, operator +=, operator -=): ... here
12086 and implement more efficiently.
12087 (mul_double_with_sign): Remove.
12088 (lshift_double): Adjust to take unsinged shift argument, push
12089 dispatching code to callers.
12090 (mul_double_wide_with_sign): Add early out for callers that
12091 are not interested in high parts or overflow.
12092 (lshift): New function.
12093 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12094 dispatch code here.
12095 (lrotate, rrotate): Use logical shifts.
12096 * expr.c (get_inner_reference): Use lshift.
12097 * fixed-value.c (do_fixed_divide): Likewise.
12098 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12099 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12100 (indirect_refs_may_alias_p): Likewise.
12101 (stmt_kills_ref_p_1): Likewise.
12103 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12105 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12107 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12109 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12110 scalar form of FABD instruction.
12112 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12114 * lra-constraints.c (process_alt_operands): Add checking alt
12115 number to choose the best alternative.
12117 2013-05-02 Richard Biener <rguenther@suse.de>
12119 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12120 bitmap and its handling.
12121 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12123 2013-05-02 Richard Biener <rguenther@suse.de>
12125 PR middle-end/57140
12126 * tree-inline.c (copy_loops): Properly handle removed loops.
12127 (copy_cfg_body): Mark destination loops for fixup if source
12128 loops needed fixup.
12130 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12133 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12134 generating simple_return for naked functions.
12136 2013-05-02 Martin Jambor <mjambor@suse.cz>
12138 PR middle-end/56988
12139 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12140 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12142 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12143 ipa_agg_replacement_value structures.
12144 (known_aggs_to_agg_replacement_list): Likewise.
12145 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12146 (read_agg_replacement_chain): Likewise.
12147 (ipcp_transform_function): Also check that by_ref flags match.
12149 2013-05-02 Richard Biener <rguenther@suse.de>
12151 * graphds.h (struct graph): Add obstack member.
12152 * graphds.c (new_graph): Initialize obstack and allocate
12154 (add_edge): Allocate edge from the obstack.
12155 (free_graph): Free the obstack instead of all edges and vertices.
12157 2013-05-02 Teresa Johnson <tejohnson@google.com>
12159 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12161 * cfg.c (update_bb_profile_for_threading): Ditto.
12162 * tree-inline.c (copy_bb): Ditto.
12163 (copy_edges_for_bb): Ditto.
12164 (initialize_cfun): Ditto.
12165 (copy_cfg_body): Ditto.
12166 (expand_call_inline): Ditto.
12167 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12168 (estimate_node_size_and_time): Ditto.
12169 (inline_merge_summary): Ditto.
12170 * cgraphclones.c (cgraph_clone_edge): Ditto.
12171 (cgraph_clone_node): Ditto.
12172 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12173 (compute_trg_info): Ditto.
12175 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12177 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12178 S reg when fp attribute set.
12179 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12181 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12183 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12185 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12186 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12187 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12189 2013-05-02 Richard Biener <rguenther@suse.de>
12191 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12192 (struct instantiate_cache_entry): New type.
12193 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12194 (struct instantiate_cache_type): New type.
12195 (set_instantiated_value, get_instantiated_value): Remove.
12196 (get_instantiated_value_entry): New function.
12197 (instantiate_scev_name): Use the new cache and adjust.
12198 (instantiate_scev_poly): Adjust.
12199 (instantiate_scev_binary): Likewise.
12200 (instantiate_array_ref): Likewise.
12201 (instantiate_scev_convert): Likewise.
12202 (instantiate_scev_not): Likewise.
12203 (instantiate_scev_3): Likewise.
12204 (instantiate_scev_2): Likewise.
12205 (instantiate_scev_r): Likewise.
12206 (instantiate_scev): Likewise.
12207 (resolve_mixers): Likewise.
12209 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12212 * lra-constraints.c (best_small_class_operands_num): Remove.
12213 (process_alt_operands): Remove small_class_operands_num. Take
12214 small classes operands into losers and only if the operand is not
12215 matched. Modify debugging output.
12216 (curr_insn_transform): Remove best_small_class_operands_num.
12219 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12221 * config/aarch64/aarch64-builtins.c
12222 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12223 * config/aarch64/aarch64-simd-builtins.def
12224 (reduc_splus_): Add new modes.
12225 (reduc_uplus_): New.
12226 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12227 (reduc_uplus_v4sf): Likewise.
12228 (reduc_splus_v4sf): Likewise.
12229 (aarch64_addv<mode>): Likewise.
12230 (reduc_uplus_<mode>): Likewise.
12231 (reduc_splus_<mode>): Likewise.
12232 (aarch64_addvv2di): Likewise.
12233 (reduc_uplus_v2di): Likewise.
12234 (reduc_splus_v2di): Likewise.
12235 (aarch64_addvv2si): Likewise.
12236 (reduc_uplus_v2si): Likewise.
12237 (reduc_splus_v2si): Likewise.
12238 (reduc_<sur>plus_<mode>): New.
12239 (reduc_<sur>plus_v2di): Likewise.
12240 (reduc_<sur>plus_v2si): Likewise.
12241 (reduc_<sur>plus_v4sf): Likewise.
12242 (aarch64_addpv4sf): Likewise.
12243 * config/aarch64/arm_neon.h
12244 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12245 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12246 add UNSPEC_SADDV, UNSPEC_UADDV.
12248 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12250 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12252 * config/aarch64/arm_neon.h
12253 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12255 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12257 * config/aarch64/aarch64-builtins
12258 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12260 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12262 * config/aarch64/aarch64-simd-builtins.def
12263 (reduc_smax_): New.
12264 (reduc_smin_): Likewise.
12265 (reduc_umax_): Likewise.
12266 (reduc_umin_): Likewise.
12267 (reduc_smax_nan_): Likewise.
12268 (reduc_smin_nan_): Likewise.
12271 (smax): Update for V2SF, V4SF and V2DF modes.
12274 (smin_nan): Likewise.
12275 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12276 (<su><maxmin><mode>3): ...This, refactor.
12277 (s<maxmin><mode>3): New.
12278 (<maxmin_uns><mode>3): Likewise.
12279 (reduc_<maxmin_uns>_<mode>): Refactor.
12280 (reduc_<maxmin_uns>_v4sf): Likewise.
12281 (reduc_<maxmin_uns>_v2si): Likewise.
12282 (aarch64_<fmaxmin><mode>: Remove.
12283 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12285 (vmin<q>_f<32,64>): Likewise.
12286 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12288 (su): Add mappings for smax, smin, umax, umin.
12290 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12291 (FMAXMIN): Rename as...
12292 (FMAXMIN_UNS): ...This.
12294 (fmaxminv): Likewise.
12295 (fmaxmin): Likewise.
12297 (maxmin_uns_op): Likewise.
12299 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12301 * config/aarch64/arm_neon.h
12302 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12303 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12304 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12306 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12308 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12309 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12311 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12313 * config/aarch64/aarch64-simd.md
12314 (vcond<mode>_internal): Handle special cases for constant masks.
12315 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12316 (vcondu<mode><mode>): Likewise.
12317 (vcond<v_cmp_result><mode>): New.
12319 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12321 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12322 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12323 * config/aarch64/aarch64-simd-builtins.def
12324 (cmeq): Update to BUILTIN_VALLDI.
12329 * config/aarch64/arm_neon.h
12330 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12331 to builtins or C as appropriate.
12333 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12335 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12337 (cmhi): Rename to...
12339 * config/aarch64/aarch64-simd.md
12340 (simd_mode): Add SF.
12341 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12342 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12343 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12344 (cstore<mode>_neg): ...This.
12345 * config/aarch64/iterators.md
12347 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12348 (COMPARISONS): New.
12349 (UCOMPARISONS): Likewise.
12350 (optab): Add missing comparisons.
12356 (VCMP_S): Likewise.
12357 (VCMP_U): Likewise.
12358 (V_cmp_result): Add DF, SF modes.
12359 (v_cmp_result): Likewise.
12361 (vmtype): Likewise.
12362 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
12364 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
12366 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
12367 define_insn to define_insn_and_split.
12368 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
12369 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
12370 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
12371 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
12372 (thumb2_negscc): Likewise.
12374 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12376 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
12378 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12380 * config/arm/thumb2.md: Remove trailing whitespaces.
12382 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12384 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
12385 Use gen_int_mode rather than GEN_INT.
12387 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
12389 * value-prof.c (stream_in_histogram_value): Remove the strayed
12392 2013-04-30 Richard Biener <rguenther@suse.de>
12394 PR middle-end/57122
12395 * cfghooks.c (split_edge): Properly check for the loop latch edge.
12397 2013-04-30 Richard Biener <rguenther@suse.de>
12399 PR middle-end/57107
12400 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
12402 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
12404 PR rtl-optimization/56957
12405 PR rtl-optimization/57105
12406 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
12407 variable. Use just INSN_UID for determining whether an insn
12408 should be only disconnected from the insn stream.
12409 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
12411 2013-04-30 Jakub Jelinek <jakub@redhat.com>
12413 PR tree-optimization/57104
12414 * tsan.c (instrument_expr): Don't instrument accesses to
12415 DECL_HARD_REGISTER VAR_DECLs.
12417 2013-04-30 Richard Biener <rguenther@suse.de>
12419 * function.h (loops_for_fn): New inline function.
12420 (set_loops_for_fn): Likewise.
12421 * cfgloop.h (place_new_loop): Add struct function parameter.
12422 (get_loop): Likewise.
12423 (get_loops): Likewise.
12424 (number_of_loops): Likewise.
12425 (fel_next): Adjust.
12426 (fel_init): Likewise.
12427 * cfg.c (get_loop_copy): Adjust.
12428 * cfgloop.c (flow_loops_dump): Likewise.
12429 (record_loop_exits): Likewise.
12430 (verify_loop_structure): Likewise.
12431 * cfgloopanal.c (mark_irreducible_loops): Likewise.
12432 (estimate_reg_pressure_cost): Likewise.
12433 (mark_loop_exit_edges): Likewise.
12434 * cfgloopmanip.c (place_new_loop): Likewise.
12435 (add_loop): Likewise.
12436 (duplicate_loop): Likewise.
12437 * graph.c (draw_cfg_nodes): Likewise.
12438 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
12439 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
12440 (extract_affine_chrec): Likewise.
12441 (build_scop_iteration_domain): Likewise.
12442 * graphite.c (graphite_initialize): Likewise.
12443 * ira-build.c (create_loop_tree_nodes): Likewise.
12444 (more_one_region_p): Likewise.
12445 (rebuild_regno_allocno_maps): Likewise.
12446 (mark_loops_for_removal): Likewise.
12447 (mark_all_loops_for_removal): Likewise.
12448 (remove_unnecessary_regions): Likewise.
12449 (ira_build): Likewise.
12450 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
12451 * loop-init.c (fix_loop_structure): Likewise.
12452 (gate_rtl_move_loop_invariants): Likewise.
12453 (gate_rtl_unswitch): Likewise.
12454 (gate_rtl_unroll_and_peel_loops): Likewise.
12455 (rtl_doloop): Likewise.
12456 * lto-streamer-in.c (input_cfg): Likewise.
12457 * lto-streamer-out.c (output_cfg): Likewise.
12458 * modulo-sched.c (sms_schedule): Likewise.
12459 * predict.c (tree_estimate_probability): Likewise.
12460 (tree_estimate_probability_driver): Likewise.
12461 (estimate_loops): Likewise.
12462 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
12463 (move_sese_region_to_fn): Likewise.
12464 (debug_loop_num): Likewise.
12465 * tree-chrec.c (chrec_evaluate): Likewise.
12466 (hide_evolution_in_other_loops_than_loop): Likewise.
12467 (chrec_component_in_loop_num): Likewise.
12468 (reset_evolution_in_loop): Likewise.
12469 (evolution_function_is_invariant_rec_p): Likewise.
12470 * tree-if-conv.c (main_tree_if_conversion): Likewise.
12471 * tree-inline.c (copy_loops): Likewise.
12472 (copy_cfg_body): Likewise.
12473 (tree_function_versioning): Likewise.
12474 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
12475 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
12477 (add_to_evolution_1): Likewise.
12478 (scev_const_prop): Likewise.
12479 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
12480 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
12481 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
12482 (tree_ssa_lim_initialize): Likewise.
12483 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
12484 (verify_loop_closed_ssa): Likewise.
12485 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
12486 (tree_ssa_loop_im): Likewise.
12487 (tree_ssa_loop_unswitch): Likewise.
12488 (tree_vectorize): Likewise.
12489 (check_data_deps): Likewise.
12490 (tree_ssa_loop_ivcanon): Likewise.
12491 (tree_ssa_loop_bounds): Likewise.
12492 (tree_complete_unroll): Likewise.
12493 (tree_complete_unroll_inner): Likewise.
12494 (tree_parallelize_loops): Likewise.
12495 (tree_ssa_loop_prefetch): Likewise.
12496 (tree_ssa_loop_ivopts): Likewise.
12497 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
12498 * tree-vectorizer.c (vectorize_loops): Likewise.
12500 2013-04-29 Mike Frysinger <vapier@gentoo.org>
12502 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
12503 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
12504 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
12505 with EABI_LINK_SPEC.
12507 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
12510 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
12513 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
12516 * lra-constraints.c (process_alt_operands): Discourage a bit more
12517 using memory for pseudos. Print cost dump for alternatives.
12518 Modify cost values for conflicts with early clobbers.
12519 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
12521 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
12524 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
12526 2013-04-29 Ian Bolton <ian.bolton@arm.com>
12528 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
12529 from/to S register.
12530 (movdi_aarch64): Support LDR/STR from/to D register.
12532 2013-04-29 Ian Bolton <ian.bolton@arm.com>
12534 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
12535 or higher by default.
12537 2013-04-29 Richard Biener <rguenther@suse.de>
12539 PR middle-end/57075
12540 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
12541 even if not adding abnormal edges for calls that can make
12544 2013-04-29 Richard Biener <rguenther@suse.de>
12546 PR middle-end/57103
12547 * tree-cfg.c (move_stmt_op): Fix condition under which to update
12549 (move_stmt_r): Remove redundant checking.
12551 2013-04-29 Teresa Johnson <tejohnson@google.com>
12554 * basic-block.h (apply_scale): New function.
12555 (apply_probability): Use apply_scale.
12556 * gimple-streamer-in.c (input_bb): Ditto.
12557 * lto-streamer-in.c (input_cfg): Ditto.
12558 * lto-cgraph.c (merge_profile_summaries): Ditto.
12559 * tree-optimize.c (execute_fixup_cfg): Ditto.
12560 * tree-inline.c (copy_bb): Update comment to use apply_scale.
12561 (copy_edges_for_bb): Ditto.
12562 (copy_cfg_body): Ditto.
12564 2013-04-29 Tom de Vries <tom@codesourcery.com>
12566 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
12567 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
12568 (tail_merge_optimize): Handle current_loops == NULL.
12570 2013-04-26 Jeff Law <law@redhat.com>
12572 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
12573 (simplify_cond_using_ranges): Generalize code to simplify
12574 COND_EXPRs where one argument is a constant and the other
12575 is an SSA_NAME created by an integral type conversion.
12577 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12579 * config/arm/arm.md (store_minmaxsi): Use only when
12580 optimize_insn_for_size_p.
12582 2013-04-29 Christian Bruel <christian.bruel@st.com>
12585 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
12587 2013-04-29 Richard Biener <rguenther@suse.de>
12589 PR middle-end/57089
12590 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
12591 loop tree make sure to schedule a fixup for the child as well.
12592 (expand_omp_for_generic): Properly add loops.
12593 (expand_omp_for_static_nochunk): Likewise.
12594 (expand_omp_for_static_chunk): Likewise.
12595 (expand_omp_for): For the degenerate case fixup loops.
12596 (expand_omp_sections): Fix default bb placement in loops.
12597 (expand_omp_atomic_pipeline): Properly add loops.
12599 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12601 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
12603 2013-04-29 Tom de Vries <tom@codesourcery.com>
12605 * tree-ssa-tail-merge.c: Update header comment.
12607 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12609 * config/aarch64/arm_neon.h
12610 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
12611 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
12612 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
12613 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
12614 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
12615 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
12616 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
12617 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
12619 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12621 * config/aarch64/aarch64-simd.md
12622 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
12623 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
12624 fix_trunc, fixuns_trunc.
12625 (ftrunc<VDQF:mode>2): New.
12626 * config/aarch64/iterators.md (optab): Add fix, fixuns.
12627 (fix_trunc_optab): New.
12629 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12631 * config/aarch64/aarch64-builtins.c
12632 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
12633 iceilf, lround, iroundf.
12635 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
12638 * config/i386/i386.h (enum ix86_tune_indices)
12639 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
12640 New, split from X86_TUNE_INTER_UNIT_MOVES.
12641 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
12642 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
12643 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
12644 (TARGET_INTER_UNIT_MOVES): Remove.
12645 * config/i386/i386.c (initial_ix86_tune_features): Update.
12646 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
12647 (ix86_expand_convert_uns_didf_sse): Use
12648 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12649 (ix86_expand_vector_init_one_nonzero): Ditto.
12650 (ix86_expand_vector_init_interleave): Ditto.
12651 (inline_secondary_memory_needed): Return true for moves from SSE class
12652 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
12653 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
12654 * config/i386/constraints.md (Yi, Ym): Depend on
12655 TARGET_INTER_UNIT_MOVES_TO_VEC.
12656 (Yj, Yn): New constraints.
12657 * config/i386/i386.md (*movdi_internal): Change constraints of
12658 operand 1 from Yi to Yj and from Ym to Yn.
12659 (*movsi_internal): Ditto.
12660 (*movdf_internal): Ditto.
12661 (*movsf_internal): Ditto.
12662 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
12663 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12664 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
12665 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
12666 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
12667 * config/i386/sse.md (movdi_to_sse): Ditto.
12668 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
12669 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
12670 TARGET_INTER_UNIT_MOVES.
12671 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
12672 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
12673 instead of TARGET_INTER_UNIT_MOVES.
12674 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
12675 operand 1 from Yi to Yj and from Ym to Yn.
12677 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12679 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
12680 (float_truncate_hi_): Likewise.
12681 (float_extend_lo_): Likewise.
12682 (float_truncate_lo_): Likewise.
12683 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
12684 (aarch64_float_extend_lo_v2df): Likewise.
12685 (vec_unpacks_hi_v4sf): Likewise.
12686 (aarch64_float_truncate_lo_v2sf): Likewise.
12687 (aarch64_float_truncate_hi_v4sf): Likewise.
12688 (vec_pack_trunc_v2df): Likewise.
12689 (vec_pack_trunc_df): Likewise.
12691 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12693 * config/aarch64/aarch64-builtins.c
12694 (aarch64_fold_builtin): Fold float conversions.
12695 * config/aarch64/aarch64-simd-builtins.def
12696 (floatv2si, floatv4si, floatv2di): New.
12697 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
12698 * config/aarch64/aarch64-simd.md
12699 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
12700 * config/aarch64/iterators.md (FLOATUORS): New.
12701 (optab): Add float, floatuns.
12702 (su_optab): Likewise.
12704 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12706 * config/aarch64/aarch64-builtins.c
12707 (aarch64_builtin_vectorized_function): Use new names for
12709 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
12710 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
12711 (fcvtzu): Split as...
12712 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
12713 (fcvtas): Split as...
12714 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
12715 (fcvtau): Split as...
12716 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
12717 (fcvtps): Split as...
12718 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
12719 (fcvtpu): Split as...
12720 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
12721 (fcvtms): Split as...
12722 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
12723 (fcvtmu): Split as...
12724 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
12725 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
12726 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
12727 (lfrintnusf, lfrintnudf): Likewise.
12728 * config/aarch64/aarch64-simd.md
12729 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
12731 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
12732 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
12733 (fcvt_pattern): Likewise.
12735 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12737 * config/aarch64/aarch64-simd.md
12738 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
12739 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
12741 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12743 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
12744 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
12745 (vrnd<a,m,n,p>_f32): Implement using builtins.
12746 (vrnd<i,x><q>_f<32, 64>): New.
12748 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12750 * config/aarch64/aarch64-builtins.c
12751 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
12752 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
12753 (frintz): Rename to...
12755 (frintp): Rename to...
12757 (frintm): Rename to...
12759 (frinti): Rename to...
12760 (nearbyint): ...this.
12761 (frintx): Rename to...
12763 (frinta): Rename to...
12765 * config/aarch64/aarch64-simd.md
12766 (aarch64_frint<frint_suffix><mode>): Delete.
12767 (<frint_pattern><mode>2): Convert to insn.
12768 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
12769 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
12770 (frint_pattern): Likewise.
12771 (frint_suffix): Likewise.
12773 2013-04-29 Richard Biener <rguenther@suse.de>
12775 PR tree-optimization/57081
12776 * loop-init.c: Include tree-flow.h.
12777 (loop_optimizer_finalize): Free number of iteration estimates.
12778 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
12780 2013-04-29 Jakub Jelinek <jakub@redhat.com>
12782 PR tree-optimization/57083
12783 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
12784 non-singleton shift count range, zero extend low_bound for uns case.
12786 * config/i386/predicates.md (general_vector_operand): New predicate.
12787 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
12788 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
12789 if they aren't nonimmediate operands. If their original values
12790 satisfy const_vector_equal_evenodd_p, don't shift them.
12791 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
12792 predicates. For the SSE4.1 case force operands[{1,2}] into registers
12793 if not nonimmediate_operand.
12794 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
12795 instead of register_operand.
12796 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
12798 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
12800 * stor-layout.c (finalize_size_functions): Allocate a structure and
12801 reset cfun before dumping the functions.
12803 2013-04-27 Jakub Jelinek <jakub@redhat.com>
12805 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
12808 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
12809 use xop_pmacsdqh if uns_p.
12810 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
12811 the immediate rotate count.
12813 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
12815 * rtl.h (struct rtx_def): Add comment for field jump.
12816 (LRA_SUBREG_P): New macro.
12817 * recog.c (register_operand): Check LRA_SUBREG_P.
12818 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12820 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12821 (lra_final_code_change): Skip subreg change for operators.
12822 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12823 if there are no operand changes.
12824 * lra-constraints.c (curr_insn_set): New.
12825 (match_reload): Set LRA_SUBREG_P.
12826 (emit_spill_move): Ditto.
12827 (check_and_process_move): Use curr_insn_set. Process only single
12828 set insns. Don't initialize sec_mem_p and change_p.
12829 (simplify_operand_subreg): Use LRA_SUBREG_P.
12830 (reg_in_class_p): New function.
12831 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12832 of #ifdef. Add code to remove cycling.
12833 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12834 non-null disp. Reload inner instead of disp when base and index
12835 are null. Try to put lo_sum into register.
12836 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12837 (check_and_process_move): Move code for move cost check to
12838 simple_move_p. Remove equiv_substitution.
12839 (simple_move_p): New function.
12840 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12841 curr_insn_set. Call check_and_process_move only for single set
12842 insns. Use the new function. Move call of check_and_process_move
12843 after operand equiv substitution and address process.
12845 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12848 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
12849 with nonlocal goto receivers or returns twice calls, ignore
12850 unininitialized values from abnormal edges to nl goto receiver
12851 or returns twice call.
12853 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12855 PR tree-optimization/57051
12856 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
12857 and VEC_RSHIFT_EXPR if shift count is a multiple of element
12860 2013-04-26 Richard Biener <rguenther@suse.de>
12862 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
12863 (expand_omp_taskreg): Likewise. Mark loops for fixup.
12864 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
12865 (fixup_loop_arrays_after_move): New function.
12866 (move_sese_region_to_fn): Properly outline the loop tree parts
12867 of the SESE region.
12869 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
12871 * config/i386/i386.md (type, unit): Fix long lines.
12873 2013-04-26 Richard Biener <rguenther@suse.de>
12875 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
12876 (lto-streamer-out.o): Likewise.
12877 * cfgloop.c (init_loops_structure): Export, add struct function
12878 argument and adjust.
12879 (flow_loops_find): Adjust.
12880 * cfgloop.h (enum loop_estimation): Add EST_LAST.
12881 (init_loops_structure): Declare.
12882 * lto-streamer-in.c: Include cfgloop.h.
12883 (input_cfg): Input the loop tree.
12884 * lto-streamer-out.c: Include cfgloop.h.
12885 (output_cfg): Output the loop tree.
12886 (output_struct_function_base): Do not drop PROP_loops.
12888 2013-03-26 Richard Biener <rguenther@suse.de>
12890 * tree-cfg.c (execute_build_cfg): Build the loop tree.
12891 (pass_build_cfg): Provide PROP_loops.
12892 (move_sese_region_to_fn): Remove loops that are outlined into fn
12894 * tree-inline.c: Include cfgloop.h.
12895 (initialize_cfun): Do not drop PROP_loops.
12896 (copy_loops): New function.
12897 (copy_cfg_body): Copy loop structure.
12898 (tree_function_versioning): Initialize destination loop tree.
12899 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
12900 (pass_parallelize_loops): Do IL verification.
12901 * loop-init.c (loop_optimizer_init): Fixup loops if required.
12902 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
12903 the CFG make sure we fixup loops as well.
12904 * tree-ssa-tail-merge.c: Include cfgloop.h.
12905 (replace_block_by): When merging loop latches mark loops for fixup.
12906 * lto-streamer-out.c (output_struct_function_base): Drop
12907 PROP_loops for now.
12908 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
12909 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
12910 * ipa-split.c: Include cfgloop.h.
12911 (split_function): Add the new return block to the loop tree root.
12912 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
12913 whether we have removed the forwarder block.
12914 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
12915 * cfgloop.h (place_new_loop): Declare.
12916 * cfgloopmanip.c (place_new_loop): Export.
12917 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
12918 (tree-switch-conversion.o): Likewise.
12919 (tree-complex.o): Likewise.
12920 (tree-inline.o): Likewise.
12921 (tree-ssa-tailmerge.o): Likewise.
12922 (ipa-split.o): Likewise.
12923 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
12924 (tree-ssa-copy.o): Likewise.
12925 * tree-switch-conversion.c: Include cfgloop.h
12926 (process_switch): If we emit a bit-test cascade, schedule loops
12928 * tree-complex.c: Include cfgloop.h.
12929 (expand_complex_div_wide): Properly add new basic-blocks to loops.
12930 * asan.c: Include cfgloop.h.
12931 (create_cond_insert_point): Properly add new basic-blocks to
12932 loops, schedule loop fixup.
12933 * cfgloop.c (verify_loop_structure): Check that looks are not
12935 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
12937 (expand_omp_for_generic): Likewise.
12938 (expand_omp_sections): Likewise.
12939 (expand_omp_atomic_pipeline): Schedule loops for fixup.
12940 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
12941 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
12942 is initialized, not when loops are present.
12943 * tree-parloops.c (parallelize_loops): Remove checking here.
12944 * passes.c (init_optimization_passes): Schedule a copy-propagation
12945 pass before complete unrolling of inner loops.
12947 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12949 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
12950 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
12951 (colorize_init): Add argument to _WIN32 version.
12952 * toplev.c: Include diagnostic-color.h.
12953 (process_options): Default to -fdiagnostics-color=auto if
12954 GCC_COLORS env var is in the environment.
12955 * common.opt (fdiagnostics-color=): Add Var and Init.
12956 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
12957 env var is in the environment, the default is auto rather than never.
12959 * diagnostic.h (file_name_as_prefix): Add context argument.
12960 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
12961 the string as locus.
12962 * langhooks.c (lhd_print_error_function): Adjust caller.
12964 2013-04-25 Lawrence Crowl <crowl@google.com>
12966 * var-tracking.c (shared_hash_def::htab):
12967 Change type to hash_table. Update dependent calls and types.
12969 2013-04-25 Lawrence Crowl <crowl@google.com>
12971 * Makefile.in: Update as needed below.
12973 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
12974 Move declaration to after the type's method definitons.
12976 * attribs.c (htab_t scoped_attributes::attribute_hash):
12977 Change type to hash_table. Update dependent calls and types.
12979 * bitmap.c (htab_t bitmap_desc_hash):
12980 Change type to hash_table. Update dependent calls and types.
12982 * cselib.c (htab_t cselib_hash_table):
12983 Change type to hash_table. Update dependent calls and types.
12985 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
12986 (hash_string_slot_node): Move implementation into lto-streamer.h
12987 struct string_slot_hasher.
12988 (eq_string_slot_node): Likewise.
12990 * data-streamer-out.c: Update output_block::string_hash_table
12991 dependent calls and types.
12993 * dwarf2cfi.c (htab_t trace_index):
12994 Change type to hash_table. Update dependent calls and types.
12996 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
12997 Change type to hash_table. Update dependent calls and types.
12998 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
12999 (htab_t optimize_external_refs::map): Likewise.
13000 (htab_t output_comp_unit::extern_map): Likewise.
13001 (htab_t output_comdat_type_unit::extern_map): Likewise.
13002 (htab_t output_macinfo::macinfo_htab): Likewise.
13003 (htab_t optimize_location_lists::htab): Likewise.
13004 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13006 * except.c (htab_t ehspec_hash_type):
13007 Change type to hash_table. Update dependent calls and types.
13008 (assign_filter_values::ttypes): Likewise.
13009 (assign_filter_values::ehspec): Likewise.
13010 (sjlj_assign_call_site_values::ar_hash): Likewise.
13011 (convert_to_eh_region_ranges::ar_hash): Likewise.
13013 * gcse.c (htab_t pre_ldst_table):
13014 Change type to hash_table. Update dependent calls and types.
13016 * ggc-common.c (htab_t saving_htab):
13017 Change type to hash_table. Update dependent calls and types.
13018 (htab_t loc_hash): Likewise.
13019 (htab_t ptr_hash): Likewise.
13020 (call_count): Rename ggc_call_count.
13021 (call_alloc): Rename ggc_call_alloc.
13022 (loc_descriptor): Rename make_loc_descriptor.
13023 (add_statistics): Rename ggc_add_statistics.
13025 * ggc-common.c (saving_htab):
13026 Change type to hash_table. Update dependent calls and types.
13028 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13029 (push_gimplify_context): Likewise.
13030 (pop_gimplify_context): Likewise.
13031 (struct gimple_temp_hash_elt): Added.
13032 (struct gimplify_hasher): Likewise.
13033 (struct gimplify_ctx.temp_htab):
13034 Change type to hash_table. Update dependent calls and types.
13036 * gimple-fold.c: Include gimplify-ctx.h.
13038 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13039 Change type to hash_table. Update dependent calls and types.
13040 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13041 avoid potential global name collision.
13043 * gimplify.c: Include gimplify-ctx.h.
13044 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13045 (htab_t gimplify_ctx::temp_htab):
13046 Update dependent calls and types for new type hash_table.
13047 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13048 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13050 * gimplify-ctx.h: New.
13051 (struct gimple_temp_hash_elt): Move from gimplify.c.
13052 (class gimplify_hasher): New.
13053 (struct gimplify_ctx): Move from gimple.h.
13054 (htab_t gimplify_ctx::temp_htab):
13055 Change type to hash_table. Update dependent calls and types.
13057 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13058 (htab_t ivs_params::newivs_index):
13059 Change type to hash_table. Update dependent calls and types.
13060 (htab_t ivs_params::params_index): Likewise.
13061 (htab_t print_generated_program::params_index): Likewise.
13062 (htab_t gloog::newivs_index): Likewise.
13063 (htab_t gloog::params_index): Likewise.
13065 * graphite.c: Include graphite-htab.h.
13066 4htab_t graphite_transform_loops::bb_pbb_mapping):
13067 Change type to hash_table. Update dependent calls and types.
13069 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13070 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13071 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13073 * graphite-dependences.c: Include graphite-htab.h.
13074 (loop_is_parallel_p): Change hash table type of parameter.
13076 * graphite-htab.h: New.
13077 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13078 (extern find_pbb_via_hash): Move from graphite-poly.h.
13079 (extern loop_is_parallel_p): Move from graphite-poly.h.
13080 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13082 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13083 (extern loop_is_parallel_p): Move to graphite-htab.h.
13084 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13086 * haifa-sched.c (htab_t delay_htab):
13087 Change type to hash_table. Update dependent calls and types.
13088 (htab_t delay_htab_i2): Likewise.
13090 * ira-color.c (htab_t allocno_hard_regs_htab):
13091 Change type to hash_table. Update dependent calls and types.
13093 * ira-costs.c (htab_t cost_classes_htab):
13094 Change type to hash_table. Update dependent calls and types.
13096 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13097 Change type to hash_table. Update dependent calls and types.
13099 * loop-iv.c (htab_t bivs):
13100 Change type to hash_table. Update dependent calls and types.
13102 * loop-unroll.c (htab_t opt_info::insns_to_split):
13103 Change type to hash_table. Update dependent calls and types.
13104 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13106 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13107 (struct string_slot_hasher): New.
13108 (htab_t output_block::string_hash_table):
13109 Change type to hash_table. Update dependent calls and types.
13111 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13112 (htab_t file_name_hash_table):
13113 Change type to hash_table. Update dependent calls and types.
13115 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13118 * lto-streamer.c (htab_t tree_htab):
13119 Change type to hash_table. Update dependent calls and types.
13121 * omp-low.c: Include gimplify-ctx.h.
13123 * passes.c (htab_t name_to_pass_map):
13124 Change type to hash_table. Update dependent calls and types.
13125 (pass_traverse): Rename to passes_pass_traverse.
13127 * plugin.c (htab_t event_tab):
13128 Change type to hash_table. Update dependent calls and types.
13130 * postreload-gcse.c (htab_t expr_table):
13131 Change type to hash_table. Update dependent calls and types.
13132 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13134 * sese.c (debug_rename_map_1): Make extern.
13135 (htab_t copy_bb_and_scalar_dependences::rename_map):
13136 Change type to hash_table. Update dependent calls and types.
13138 * sese.h (extern debug_rename_map): Move to .c file.
13140 * store-motion.c (htab_t store_motion_mems_table):
13141 Change type to hash_table. Update dependent calls and types.
13143 * trans-mem.c (htab_t tm_new_mem_hash):
13144 Change type to hash_table. Update dependent calls and types.
13146 * tree-browser.c (htab_t TB_up_ht):
13147 Change type to hash_table. Update dependent calls and types.
13149 * tree-cfg.c (htab_t discriminator_per_locus):
13150 Change type to hash_table. Update dependent calls and types.
13152 * tree-complex.c: Include tree-hasher.h
13153 (htab_t complex_variable_components):
13154 Change type to hash_table. Update dependent calls and types.
13156 * tree-eh.c (htab_t finally_tree):
13157 Change type to hash_table. Update dependent calls and types.
13159 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13160 struct int_tree_hasher.
13161 (extern int_tree_map_eq): Likewise.
13162 (uid_decl_map_hash): Removed.
13163 (extern decl_tree_map_eq): Likewise.
13165 * tree-hasher.h: New.
13166 (struct int_tree_hasher): New.
13167 (typedef int_tree_htab_type): New.
13169 * tree-inline.c: Include gimplify-ctx.h.
13171 * tree-mudflap.c: Include gimplify-ctx.h.
13173 * tree-parloops.c: Include tree-hasher.h.
13174 (htab_t eliminate_local_variables_stmt::decl_address):
13175 Change type to hash_table. Update dependent calls and types.
13176 (htab_t separate_decls_in_region::decl_copies): Likewise.
13178 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13179 Change type to hash_table. Update dependent calls and types.
13181 * tree-sra.c (candidates):
13182 Change type to hash_table. Update dependent calls and types.
13184 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13186 (int_tree_map_hash): Likewise.
13188 * tree-ssa-dom.c (htab_t avail_exprs):
13189 Change type to hash_table. Update dependent calls and types.
13191 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13192 Change type to hash_table. Update dependent calls and types.
13194 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13195 Change type to hash_table. Update dependent calls and types.
13197 * tree-ssa-phiopt.c (seen_ssa_names):
13198 Change type to hash_table. Update dependent calls and types.
13200 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13201 Change type to hash_table. Update dependent calls and types.
13203 * tree-ssa-uncprop.c (equiv):
13204 Change type to hash_table. Update dependent calls and types.
13206 2013-04-25 Jakub Jelinek <jakub@redhat.com>
13208 PR rtl-optimization/57003
13209 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13210 call note_stores with kill_clobbered_value callback again after
13211 killing regs_invalidated_by_call.
13213 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13215 * config/aarch64/aarch64-simd.md
13216 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13217 (aarch64_simd_bsl<mode>): Likewise.
13218 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13220 2013-04-25 Marek Polacek <polacek@redhat.com>
13222 PR tree-optimization/57066
13223 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
13225 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13227 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13229 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13231 * config/aarch64/aarch64-builtins.c
13232 (aarch64_fold_builtin): New.
13233 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13234 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13235 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13236 * config/aarch64/arm_neon.h
13237 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13239 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13240 Tejas Belagod <tejas.belagod@arm.com>
13242 * config/aarch64/aarch64-builtins.c
13243 (aarch64_gimple_fold_builtin): New.
13244 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13245 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13246 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13247 (addvv4sf): Update.
13248 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13250 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13252 * config/aarch64/aarch64.md
13253 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13255 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13257 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13258 (*ngcsi_uxtw): New pattern.
13260 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13261 Julian Brown <julian@codesourcery.com>
13263 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13264 (TB_DREG): Add T_V4HF.
13265 (v4hf_UP): New macro.
13266 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13267 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13268 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13270 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13271 (arm_vector_mode_supported_p): Handle V4HF.
13272 (arm_mangle_map): Handle V4HFmode.
13273 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13274 * config/arm/arm_neon_builtins.def: Add entries for
13275 vcvtv4hfv4sf, vcvtv4sfv4hf.
13276 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13277 (neon_vcvtv4hfv4sf): Likewise.
13278 * config/arm/neon-gen.ml: Handle half-precision floating point
13280 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13281 * config/arm/arm_neon.h: Regenerate.
13282 * config/arm/neon.ml (type elts): Add F16.
13283 (type vectype): Add T_float16x4, T_floatHF.
13284 (type vecmode): Add V4HF.
13285 (type features): Add Requires_FP_bit feature.
13286 (elt_width): Handle F16.
13287 (elt_class): Likewise.
13288 (elt_of_class_width): Likewise.
13289 (mode_of_elt): Refactor.
13290 (type_for_elt): Handle F16, fix error messages.
13291 (vectype_size): Handle T_float16x4.
13292 (vcvt_sh): New function.
13293 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13294 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13295 (string_of_mode): Handle V4HF.
13296 * doc/arm-neon-intrinsics.texi: Regenerate.
13298 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13300 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13301 format specifier in 'X' case.
13303 2013-04-25 Alan Modra <amodra@gmail.com>
13306 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13307 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13308 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13309 Repeat for many other rotate/shift and mask patterns using subregs.
13310 Name lshiftrt insns.
13311 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13312 on WORDS_BIG_ENDIAN.
13314 2013-04-25 Alan Modra <amodra@gmail.com>
13316 * config.gcc: Support little-endian powerpc-linux targets.
13317 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13318 (LINK_OS_LINUX_SPEC): Define.
13319 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13320 Preserve MASK_LITTLE_ENDIAN.
13321 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13322 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13323 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13324 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13325 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13326 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13327 Correct fp word order for little-endian. Don't shift toc entries
13328 smaller than a word for little-endian.
13329 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13330 (bswapdi2 splits): Correct low-part subreg for little-endian.
13331 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13332 low/high where such is correct only for be.
13333 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13334 little-endian for -mcall-aixdesc.
13336 2013-04-25 Alan Modra <amodra@gmail.com>
13338 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13339 replace_equiv_address_nv.
13341 2013-04-25 Alan Modra <amodra@gmail.com>
13343 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13345 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13348 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13349 * rtl.h (struct rtx_def): ...
13351 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13353 PR rtl-optimizations/57046
13354 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13355 for multi-reg splits.
13357 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13359 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13361 2013-04-24 Sterling Augustine <saugustine@google.com>
13363 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
13364 (comp_dir_string, debug_str_dwo_section): New.
13365 (DEBUG_STR_DWO_SECTION): Rename to ...
13366 (DEBUG_DWO_STR_SECTION): ... this.
13367 (DEBUG_NORM_STR_SECTION): Delete.
13368 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
13369 (DEBUG_STR_DWO_SECTION_FLAGS): New.
13370 (find_AT_string): Move most logic to ...
13371 (find_AT_string_in_table): ... here. New.
13372 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
13373 add_skeleton_AT_string. Delete logic.
13374 (output_skeleton_debug_sections): Remove call to
13375 add_top_level_skeleton_die_attrs.
13376 (add_comp_dir_attribute): Move logic to comp_dir_string.
13377 (dwarf2out_init): Initialize debug_str_dwo_section.
13378 (output_indirect_string): Call find_string_form.
13379 (output_indirect_strings): Rewrite.
13380 (prune_unused_types): Empty skeleton_debug_str_hash.
13381 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
13382 (dwarf2out_finish): Call output_indirect_strings.
13384 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13386 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
13388 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13390 * rtl.h (struct rtx_def): Add comment for field jump.
13391 (LRA_SUBREG_P): New macro.
13392 * recog.c (register_operand): Check LRA_SUBREG_P.
13393 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13395 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13396 (lra_final_code_change): Skip subreg change for operators.
13397 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13398 if there are no operand changes.
13399 * lra-constraints.c (curr_insn_set): New.
13400 (match_reload): Set LRA_SUBREG_P.
13401 (emit_spill_move): Ditto.
13402 (check_and_process_move): Use curr_insn_set. Process only single
13403 set insns. Don't initialize sec_mem_p and change_p.
13404 (simplify_operand_subreg): Use LRA_SUBREG_P.
13405 (reg_in_class_p): New function.
13406 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13407 of #ifdef. Add code to remove cycling.
13408 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13409 non-null disp. Reload inner instead of disp when base and index
13410 are null. Try to put lo_sum into register.
13411 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13412 (check_and_process_move): Move code for move cost check to
13413 simple_move_p. Remove equiv_substitution.
13414 (simple_move_p): New function.
13415 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13416 curr_insn_set. Call check_and_process_move only for single set
13417 insns. Use the new function. Move call of check_and_process_move
13418 after operand equiv substitution and address process.
13420 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
13422 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
13423 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
13424 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
13426 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13428 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
13430 2013-04-24 Marek Polacek <polacek@redhat.com>
13432 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
13433 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
13434 (select_loops_exit_conditions): Likewise.
13435 (number_of_iterations_for_all_loops): Likewise.
13436 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
13437 (scev_analysis): Likewise.
13439 2013-04-02 Catherine Moore <clm@codesourcery.com>
13440 Chao-ying Fu <fu@mips.com>
13442 * config/mips/micromips.md (jraddiusp): New pattern.
13443 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
13444 instruction if possible.
13446 2013-04-24 Alan Modra <amodra@gmail.com>
13448 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
13450 2013-04-24 Julian Brown <julian@codesourcery.com>
13451 Chung-Lin Tang <cltang@codesourcery.com>
13453 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
13454 dependency behavior in enumeration type DIE generation. Add TODO note
13455 to comments about future DW_FORM_sdata/udata re-work of related code.
13457 2013-04-23 Lawrence Crowl <crowl@google.com>
13459 * Makefile.in: Update as needed below.
13461 * hash-table.h (class hash_table):
13462 Correct many methods with parameter types compare_type to the correct
13463 value_type. (Correct code was unlikely to notice the change.)
13464 (hash_table::elements_with_deleted) New.
13465 (class hashtable::iterator): New.
13466 (hashtable::begin()): New.
13467 (hashtable::end()): New.
13468 (FOR_EACH_HASH_TABLE_ELEMENT): New.
13470 * statistics.c (statistics_hashes):
13471 Change type to hash_table. Update dependent calls and types.
13473 * tree-into-ssa.c (var_infos):
13474 Change type to hash_table. Update dependent calls and types.
13476 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
13477 Change type to hash_table. Update dependent calls and types.
13479 * tree-ssa-loop-im.c (struct mem_ref.refs):
13480 Change type to hash_table. Update dependent calls and types.
13482 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
13483 Change type to hash_table. Update dependent calls and types.
13485 * tree-ssa-sccvn.c (vn_tables_s::nary):
13486 Change type to hash_table. Update dependent calls and types.
13487 (vn_tables_s::phis): Likewise.
13488 (vn_tables_s::references): Likewise.
13490 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
13491 (vn_reference_eq): Update parameter and return types.
13493 * tree-ssa-structalias.c (pointer_equiv_class_table):
13494 Change type to hash_table. Update dependent calls and types.
13495 (location_equiv_class_table): Likewise.
13497 * tree-vect-data-refs.c: Consequential changes for making
13498 peeling a hash_table.
13500 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
13501 (destroy_loop_vec_info): Dependent hash_table update.
13503 * tree-vectorizer.h (peeling_htab):
13504 Change type to hash_table. Update dependent calls and types.
13506 2013-04-23 Shiva Chen <shiva0217@gmail.com>
13508 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
13509 to check the register content is equal or not.
13510 * lra-constraints.c (match_reload): Use lra_assign_reg_val
13511 to assign register content record.
13512 * lra-eliminations.c (update_reg_eliminate): Use
13513 lra_update_reg_val_offset to update register content offset.
13514 * lra-int.h (struct lra_reg): Add offset member.
13515 (lra_reg_val_equal_p): New static inline function.
13516 (lra_update_reg_val_offset): New static inline function.
13517 (lra_assign_reg_val): New static inline function.
13518 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
13519 to assign register content record.
13520 (initialize_lra_reg_info_element): Initial offset to zero.
13522 2013-04-23 Catherine Moore <clm@codesourcery.com>
13524 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
13525 operands. Record compression.
13527 2013-04-23 Xinliang David Li <davidxl@google.com>
13529 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
13531 2013-04-23 Richard Biener <rguenther@suse.de>
13533 PR middle-end/57036
13534 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
13535 parameter, only add abnormal goto edges from the copied body
13536 if the call could perform abnormal gotos.
13537 (copy_cfg_body): Adjust.
13539 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
13541 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
13543 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
13545 * coretypes.h (gimple_stmt_iterator): Add struct to make
13548 2013-04-23 Richard Biener <rguenther@suse.de>
13550 PR tree-optimization/57026
13551 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
13552 from SSA names occuring in abnormal PHI nodes.
13554 2013-04-22 Andi Kleen <ak@linux.intel.com>
13556 * lto/lto.c (print_lto_report_1): Fix LTO report names.
13558 2013-04-22 Andi Kleen <ak@linux.intel.com>
13560 * lto/lto.c (print_lto_report_1): Declare early.
13561 (read_cgraph_and_symbols): Call print_lto_report_1 early.
13563 2013-04-22 Andi Kleen <ak@linux.intel.com>
13565 * common.opt (-flto-report-wpa): Add.
13566 * doc/invoke.texi (-flto-report-wpa): Add.
13567 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
13570 2013-04-22 Xinliang David Li <davidxl@google.com>
13572 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
13573 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
13574 * Makefile.in: New dependency
13576 David Daney <ddaney.cavm@gmail.com>
13578 * configure.ac (gcc_cv_as_micromips_support): Use the
13579 --fatal-warnings option.
13580 * configure: Regenerate.
13582 2013-04-22 Marek Polacek <polacek@redhat.com>
13585 * tsan.c (instrument_expr): Don't instrument expression
13586 in case its size is zero.
13588 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
13592 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13594 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13596 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13598 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
13599 (gimple_stmt_iterator): New typedef.
13600 * gimple.h (gimple_stmt_iterator): Rename to...
13601 (gimple_stmt_iterator_d): ... This.
13602 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
13603 trees be valid for GIMPLE and GENERIC.
13604 (TARGET_GIMPLE_FOLD_BUILTIN): New.
13605 * gimple-fold.c (gimple_fold_call): Call target hook
13606 gimple_fold_builtin.
13607 * hooks.c (hook_bool_gsiptr_false): New.
13608 * hooks.h (hook_bool_gsiptr_false): New.
13609 * target.def (fold_stmt): New.
13610 * doc/tm.texi: Regenerate.
13612 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
13615 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
13616 a set sp if no stack realignment.
13618 2013-04-22 Nick Clifton <nickc@redhat.com>
13620 * config.gcc (tilegx-linux): Extend extra_objs rather than
13622 (tilepro-linux): Likewise.
13624 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13626 * config/aarch64/aarch64-builtins.c
13628 (CF0, CF1, CF2, CF3, CF4, CF10): New.
13629 (VAR<1-12>): Add MAP parameter.
13630 (BUILTIN_*): Likewise.
13631 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
13632 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
13633 (aarch64_ushl_n<mode>): Likewise.
13634 (aarch64_sshr_n<mode>): Likewise.
13635 (aarch64_ushr_n<mode>): Likewise.
13636 (aarch64_<maxmin><mode>): Likewise.
13637 (aarch64_sqrt<mode>): Likewise.
13638 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
13639 (vshr<q>_n_*): Likewise.
13641 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13643 * config/aarch64/aarch64-builtins.c
13644 (aarch64_simd_builtin_type_mode): Handle SF types.
13646 (BUILTIN_GPF): Define.
13647 (aarch64_init_simd_builtins): Handle SF types.
13648 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
13649 (frecps): Likewise.
13650 (frecpx): Likewise.
13651 * config/aarch64/aarch64-simd.md
13652 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
13653 (aarch64_frecpe<mode>): New.
13654 (aarch64_frecps<mode>): Likewise.
13655 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
13656 (v8type): Add frecp<esx>.
13657 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
13658 (aarch64_frecps<mode>): Likewise.
13659 * config/aarch64/iterators.md (FRECP): New.
13660 (frecp_suffix): Likewise.
13661 * config/aarch64/arm_neon.h
13662 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
13664 2013-04-22 Christian Bruel <christian.bruel@st.com>
13667 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
13668 (REG_CLASS_NAMES): Idem.
13669 (REG_CLASS_CONTENTS): Idem.
13670 (REGCLASS_HAS_FP_REG): Idem.
13671 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
13672 (sh_conditional_register_usage): Idem.
13674 2013-04-21 Jeff Law <law@redhat.com>
13676 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
13677 (ssa_forward_propagate_and_combine): Use it.
13679 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13681 * lra.c: Update the flow chart diagram.
13683 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13685 PR rtl-optimization/56847
13686 * lra-constraints.c (process_alt_operands): Discourage alternative
13687 with non-matche doffsettable memory constraint fro memory with
13690 2013-04-19 Richard Biener <rguenther@suse.de>
13692 PR tree-optimization/56982
13693 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
13695 * gimplify.c (gimplify_call_expr): Notice special calls.
13696 (gimplify_modify_expr): Likewise.
13697 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
13698 abnormal control flow receivers.
13699 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
13700 in the same way as cfun->has_nonlocal_labels.
13701 (gimple_purge_dead_abnormal_call_edges): Likewise.
13702 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
13703 receivers start a basic-block.
13705 2013-04-19 Richard Biener <rguenther@suse.de>
13707 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
13709 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
13710 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
13711 (SLP_TREE_LOAD_PERMUTATION): Add.
13712 (vect_transform_slp_perm_load): Adjust prototype.
13713 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
13714 (vect_free_slp_instance): Likewise.
13715 (vect_create_new_slp_node): Likewise.
13716 (vect_supported_slp_permutation_p): Remove.
13717 (vect_slp_rearrange_stmts): Adjust.
13718 (vect_supported_load_permutation_p): Likewise. Inline
13719 vect_supported_slp_permutation_p here.
13720 (vect_analyze_slp_instance): Compute load permutations per
13721 slp node instead of per instance.
13722 (vect_get_slp_defs): Adjust.
13723 (vect_transform_slp_perm_load): Likewise.
13724 (vect_schedule_slp_instance): Remove redundant code.
13725 (vect_schedule_slp): Remove hack for PR56270, add it ...
13726 * tree-vect-stmts.c (vectorizable_load): ... here, do not
13727 CSE loads for SLP. Adjust.
13729 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13731 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
13732 spelling in two comments.
13734 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13737 * config/arm/arm.c (load_multiple_sequence): Require SP
13738 as base register for loads if SP is in the register list.
13740 2013-04-19 Martin Jambor <mjambor@suse.cz>
13742 PR tree-optimization/56718
13743 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
13744 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
13745 and made public. Adjusted all callers.
13746 (ipa_intraprocedural_devirtualization): New function.
13747 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
13748 (ipa_intraprocedural_devirtualization): Likewise.
13749 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
13751 2013-04-19 Richard Biener <rguenther@suse.de>
13753 PR tree-optimization/57000
13754 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
13756 2013-04-19 Terry Guo <terry.guo@arm.com>
13758 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
13760 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
13761 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
13762 (cortex_m4_fmacs): Use new reservations.
13763 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
13765 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
13767 PR rtl-optimization/56999
13768 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
13770 (lra_coalesce): Remove split_origin_bitmap and related code.
13771 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
13772 ranges if necessary.
13774 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
13776 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
13778 (ix86_expand_call): Remove clobbered_registers array and use
13779 x86_64_ms_sysv_extra_clobbered_registers instead.
13780 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
13782 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
13784 * config/i386/i386.md (*call_rex64_ms_sysv): Use
13785 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
13786 (*call_value_rex64_ms_sysv): Ditto.
13788 2013-04-18 Cary Coutant <ccoutant@google.com>
13790 * dwarf2out.c (output_pubnames): Check die_perennial_p of
13791 parent instead of die_mark.
13793 2013-04-18 Diego Novillo <dnovillo@google.com>
13795 * gimple.c (create_gimple_tmp): New.
13796 (get_expr_type): New.
13797 (build_assign): New.
13798 (build_type_cast): New.
13799 * gimple.h (enum ssa_mode): Define.
13800 (gimple_seq_set_location): New.
13801 * asan.c (build_check_stmt): Change some gimple_build_* calls
13802 to use build_assign and build_type_cast.
13804 2013-04-18 Richard Biener <rguenther@suse.de>
13806 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
13807 handle negative step. Remove redundant checks.
13808 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
13809 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
13810 for negative step and grouped loads fail to vectorize.
13812 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13814 * emit-rtl.c (reset_insn_used_flags): New function.
13815 (reset_all_used_flags): Use it.
13816 (verify_insn_sharing): New function.
13817 (verify_rtl_sharing): Fix verification for SEQUENCEs.
13819 2013-04-18 Jakub Jelinek <jakub@redhat.com>
13821 PR tree-optimization/56984
13822 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
13823 and (x >> M) >= N don't register any assertion if N << M is the
13826 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13828 * lower-subreg.c (resolve_simple_move): If called self-recursive,
13829 do not delete_insn insns that have not yet been emitted, only
13830 unlink them with remove_insn.
13831 * df-scan.c (df_insn_delete): Revert r197492.
13833 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
13835 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
13836 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
13838 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
13840 * config/arm/arm.md (movsicc_insn): Convert define_insn into
13841 define_insn_and_split.
13842 (and_scc,ior_scc,negscc): Likewise.
13843 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
13845 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
13847 * config/arm/arm.c (use_return_insn): Return 0 for targets that
13848 can benefit from using a sequence of LDRD instructions in epilogue
13849 instead of a single LDM instruction.
13851 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
13854 * doc/extend.texi: Fix typo.
13856 2013-04-17 Richard Biener <rguenther@suse.de>
13858 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
13859 (vect_build_slp_tree): ... here.
13860 (vect_build_slp_tree_1): Compute which stmts of the SLP group
13861 match. Remove special-casing of mismatched complex loads.
13862 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
13863 re-try the match with swapped commutative operands.
13864 (vect_supported_load_permutation_p): Remove special-casing of
13865 mismatched complex loads.
13866 (vect_analyze_slp_instance): Adjust.
13868 2013-04-17 Richard Biener <rguenther@suse.de>
13870 PR rtl-optimization/56921
13871 * cfgloop.h (struct loop): Add simple_loop_desc member.
13872 (struct niter_desc): Mark with GTY(()).
13873 (simple_loop_desc): Do not use aux field but simple_loop_desc.
13874 * loop-iv.c (get_simple_loop_desc): Likewise.
13875 (free_simple_loop_desc): Likewise.
13878 2013-04-16 Richard Biener <rguenther@suse.de>
13880 PR rtl-optimization/56921
13881 * loop-init.c (pass_rtl_move_loop_invariants): Add
13882 TODO_do_not_ggc_collect to todo_flags_finish.
13883 (pass_rtl_unswitch): Same.
13884 (pass_rtl_unroll_and_peel_loops): Same.
13885 (pass_rtl_doloop): Same.
13887 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
13889 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
13890 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
13891 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
13893 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
13894 * tree-streamer.c (record_common_node): Adjust reference in comment.
13896 2013-04-17 Terry Guo <terry.guo@arm.com>
13898 * config/arm/cortex-m4.md: Add a new bypass.
13900 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13902 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
13904 (*subs_<optab><mode>_multp2): New pattern.
13905 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13906 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13908 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13910 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
13911 (*subs_mul_imm_<mode>): New pattern.
13913 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
13916 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
13917 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
13918 (vsx_movti_32bit): Change j->wa to O->wa.
13920 2013-04-16 Richard Biener <rguenther@suse.de>
13922 PR rtl-optimization/56921
13923 * loop-init.c (pass_rtl_move_loop_invariants): Add
13924 TODO_do_not_ggc_collect to todo_flags_finish.
13925 (pass_rtl_unswitch): Same.
13926 (pass_rtl_unroll_and_peel_loops): Same.
13927 (pass_rtl_doloop): Same.
13929 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
13931 * config/arm/arm.c (emit_multi_reg_push): New declaration
13932 for an existing function.
13933 (arm_emit_strd_push): New function.
13934 (arm_expand_prologue): Used here.
13935 (arm_emit_ldrd_pop): New function.
13936 (arm_expand_epilogue): Used here.
13937 (arm_get_frame_offsets): Update condition.
13938 (arm_emit_multi_reg_pop): Add a special case for load of a single
13939 register with writeback.
13941 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
13943 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
13946 2013-04-16 Richard Biener <rguenther@suse.de>
13948 PR tree-optimization/56756
13949 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
13950 (first_mem_ref_loc): New.
13951 (execute_sm): Place the load temporarily before a previous
13952 access instead of in the latch edge to ensure its SSA dependencies
13953 are defined at points dominating the load.
13955 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
13957 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
13958 correct fix by moving header and footer insn to the footer of
13959 the merged basic block. Clear BB_END of the merged-away block.
13961 PR middle-end/43631
13962 * emit-rtl.c (make_note_raw): New function.
13963 (link_insn_into_chain): New static inline function.
13964 (add_insn): Use it.
13965 (add_insn_before, add_insn_after): Factor insn chain linking code...
13966 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
13967 using link_insn_into_chain.
13968 (note_outside_basic_block_p): New helper function for emit_note_after
13969 and emit_note_before.
13970 (emit_note_after): Use nobb variant of add_insn_after if the note
13971 should not be contained in a basic block.
13972 (emit_note_before): Use nobb variant of add_insn_before if the note
13973 should not be contained in a basic block.
13974 (emit_note_copy): Use make_note_raw.
13975 (emit_note): Likewise.
13976 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
13977 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
13978 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
13979 the moved barrier the tail of the basic block it follows.
13980 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
13982 2013-04-15 Jakub Jelinek <jakub@redhat.com>
13984 PR tree-optimization/56962
13985 * gimple-ssa-strength-reduction.c (record_increment): Only set
13986 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
13987 either rhs1 or rhs2 is equal to c->base_expr.
13989 2013-04-15 Richard Biener <rguenther@suse.de>
13991 PR tree-optimization/56933
13992 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
13994 (GROUP_READ_WRITE_DEPENDENCE): Remove.
13995 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
13996 * tree-vect-data-refs.c (vect_analyze_group_access): Move
13997 dependence check ...
13998 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14000 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14001 GROUP_READ_WRITE_DEPENDENCE.
14003 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14005 * emit-rtl.c (reset_all_used_flags): New function.
14006 (verify_rtl_sharing): Call reset_all_used_flags before and after
14007 performing the checks.
14009 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14011 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14012 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14013 * config/arm/constraints.md (De): New constraint.
14014 * config/arm/neon.md (anddi3_neon): Delete.
14015 (neon_vand<mode>): Expand to standard anddi3 pattern.
14016 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14017 Move earlier in the file.
14018 (neon_inv_logic_op2): Likewise.
14019 (arm_anddi_operand_neon): New predicate.
14021 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14023 * configure.ac (gcc_cv_ld_as_needed): Set
14024 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14025 Use -z ignore, -z record on *-*-solaris2*.
14026 (HAVE_LD_AS_NEEDED): Update comment.
14027 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14028 * configure: Regenerate.
14029 * config.in: Regenerate.
14030 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14031 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14032 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14033 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14034 equivalents. Fix markup.
14035 * doc/tm.texi: Regenerate.
14037 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14039 * config/i386/i386.opt: New option mstack-protector-guard=.
14040 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14041 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14042 TARGET_SSP_TLS_GUARD.
14043 * config/i386/i386.c (ix86_option_override_internal): Set
14044 ix86_stack_protector_guard.
14045 * config/i386/i386.md (stack_protect_set): Enable for
14046 TARGET_SSP_TLS_GUARD only.
14047 (stack_protect_set_<mode>): Ditto.
14048 (stack_protect_test): Ditto.
14049 (stack_protect_test_<mode>): Ditto.
14050 * doc/invoke.texi (i386 Option): Document.
14052 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14055 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14056 (S_MODES): Set H_MODE bit.
14057 (SF_MODES): Set only S_MODE and SF_MODE bits.
14058 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14059 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14060 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14061 <MODE_FLOAT>: Likewise.
14063 2013-04-15 Joey Ye <joey.ye@arm.com>
14065 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14067 2013-04-15 Joey Ye <joey.ye@arm.com>
14069 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14071 (thumb_far_jump_used_p): Count instruction size and set
14074 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14076 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14077 * resource.c (mark_target_live_regs): Reformat conditional block.
14079 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14081 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14082 notes, they are emitted only just before final.
14083 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14085 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14087 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14088 * cfgrtl.c (delete_insn): Call it here instead.
14089 * lra-spills.c (lra_final_code_change): Use delete_insn.
14090 * haifa-sched.c (sched_remove_insn): Likewise.
14091 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14092 returning to the nop pool.
14093 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14094 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14096 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14098 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14099 * doc/tm.texi: Regenerated.
14101 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
14103 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14106 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14108 * df-core.c (df_find_def): Compare register numbers.
14109 (df_find_use): Likewise.
14111 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14114 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14115 lra_in_progress for return.
14117 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14119 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14120 define_insn into define_insn_and_split and emit movsicc patterns.
14122 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14124 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14126 2013-04-12 Richard Biener <rguenther@suse.de>
14128 * tree-pass.h (TODO_do_not_ggc_collect): New.
14129 * passes.c (execute_one_ipa_transform_pass): Honor
14130 TODO_do_not_ggc_collect.
14131 (execute_one_pass): Likewise.
14134 2013-04-10 Richard Biener <rguenther@suse.de>
14136 * passes.c (init_optimization_passes): Remove reload pass.
14137 * ira.c (do_reload): Merge into ...
14139 (rest_of_handle_reload): Remove.
14140 (pass_reload): Likewise.
14141 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14142 of reload for vzeroupper pass placement.
14144 2013-04-12 Jakub Jelinek <jakub@redhat.com>
14146 PR tree-optimization/56918
14147 PR tree-optimization/56920
14148 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14149 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14150 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14151 use wide_mul_with_sign method.
14153 2013-04-12 Richard Biener <rguenther@suse.de>
14155 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14156 not be considered a gimple constant.
14158 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
14160 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14161 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14164 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14165 Jakub Jelinek <jakub@redhat.com>
14167 * opts.c: Include diagnostic-color.h.
14168 (common_handle_option): Handle OPT_fdiagnostics_color_.
14169 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14170 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14171 (diagnostic-color.o): New.
14172 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14173 (diagnostic_color_rule): New enum.
14174 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14175 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14176 the location string.
14177 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14178 either NULL, or color kind.
14179 * diagnostic-color.c: New file.
14180 * diagnostic-color.h: New file.
14181 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14183 * doc/invoke.texi (-fdiagnostics-color): Document.
14184 * pretty-print.h (pp_show_color): Define.
14185 (struct pretty_print_info): Add show_color field.
14186 * diagnostic.c: Include diagnostic-color.h.
14187 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14188 macros. Colorize error:, warning: etc. strings and also the location
14190 (diagnostic_show_locus): Colorize the caret line.
14191 * pretty-print.c: Include diagnostic-color.h.
14192 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14193 inside of %< %> quotes or quoted through q format modifier.
14195 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14197 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
14199 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14201 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14202 code in CC_NZ mode.
14203 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14206 2013-04-11 Marek Polacek <polacek@redhat.com>
14208 PR tree-optimization/48184
14209 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
14211 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14213 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14214 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14215 (skip_simple_arithmetic): Tidy up.
14216 * tree.h (skip_simple_constant_arithmetic): Declare.
14218 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14220 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14222 2013-04-11 Richard Biener <rguenther@suse.de>
14224 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14225 generate vector constants.
14227 2013-04-11 Richard Biener <rguenther@suse.de>
14229 PR tree-optimization/56878
14230 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14231 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14233 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14234 Prefer to align the DR with the most invariant base address.
14236 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14238 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14241 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14243 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14244 floating-point vector comparisons against 0.
14246 2013-04-11 Jakub Jelinek <jakub@redhat.com>
14248 PR tree-optimization/56899
14249 * fold-const.c (extract_muldiv_1): Apply distributive law
14250 only if TYPE_OVERFLOW_WRAPS (ctype).
14252 2013-04-11 Bin Cheng <bin.cheng@arm.com>
14255 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14256 loading has side effect.
14258 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14260 * config/sparc/sparc.c: Include tree-pass.h.
14261 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14262 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14263 head of file. Change return type. Split off gate function.
14264 (sparc_gate_work_around_errata): New function.
14265 (pass_work_around_errata): New pass definition.
14266 (insert_pass_work_around_errata) New pass insert definition to
14267 insert pass_work_around_errata just after delayed-branch scheduling.
14268 (sparc_option_override): Insert the pass.
14269 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14271 2013-04-10 David S. Miller <davem@davemloft.net>
14273 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14274 or -mcpu=hypersparc.
14276 * target.def (cstore_mode): New hook.
14277 * target.h: Include insn-codes.h
14278 * targhooks.c: Likewise.
14279 (default_cstore_mode): New function.
14280 * targhooks.h: Declare it.
14281 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14282 * doc/tm.texi: Rebuild.
14283 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14284 target hook, rather than inspecting the insn_data.
14285 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14286 (TARGET_CSTORE_MODE): Redefine.
14287 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14289 * config/sparc/predicates.md (cstore_result_operand): New special
14291 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14292 Use it for operand 0.
14293 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14294 (*snesi_special): Likewise.
14295 (*snesi_zero): Likewise.
14296 (*seqsi_zero): Likewise.
14297 (*sltu_insn): Likewise.
14298 (*sgeu_insn): Likewise.
14299 (*seqdi_special): Make operand 0 and comparison operation be of
14301 (*snedi_special): Likewise.
14302 (*snedi_special_vis3): Likewise.
14303 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14304 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14305 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14306 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14307 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14308 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14309 (*sltu_extend_sp64): Likewise.
14310 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14311 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14312 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14313 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14314 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14316 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14318 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14319 (aarch64_start_file): Use the new function.
14321 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14322 Jason Merrill <jason@redhat.com>
14324 * common.opt: Add -gdwarf.
14325 * opts.c (common_handle_option): Handle it.
14326 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14328 2013-04-10 Richard Biener <rguenther@suse.de>
14330 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14331 (execute_one_ipa_transform_pass): But unconditionally here.
14332 (execute_one_pass): And here.
14333 (init_optimization_passes): Remove reload pass.
14334 * tree-pass.h (TODO_ggc_collect): Remove.
14335 (pass_reload): Likewise.
14336 * ira.c (do_reload): Merge into ...
14338 (rest_of_handle_reload): Remove.
14339 (pass_reload): Likewise.
14340 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14341 of reload for vzeroupper pass placement.
14342 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14343 and todo_flags_finish of all passes.
14345 2013-04-10 Richard Biener <rguenther@suse.de>
14347 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14348 first_const_oprnd field, rename first_def_type to first_op_type.
14349 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14350 (vect_get_and_check_slp_defs): Always use the type of the
14351 operand. Allow mixed vect_external_def, vect_constant_def types.
14352 (vect_get_constant_vectors): Handle mixed vect_external_def,
14353 vect_constant_def types.
14355 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
14357 PR tree-optimization/55524
14358 * tree-ssa-math-opts.c
14359 (convert_mult_to_fma): Don't use an fms construct
14360 when we don't have an fms operation, but fnma, and it looks
14361 likely that we'll be able to use the latter.
14363 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
14365 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
14367 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
14368 inline fail caused by overwritable functions.
14370 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
14372 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
14373 unnecessary bits in the constant power of two case.
14375 2013-04-10 Richard Biener <rguenther@suse.de>
14377 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
14378 broken code swapping operands.
14379 (vect_build_slp_tree): Do not compute load permutations here.
14380 (vect_analyze_slp_instance): Compute load permutations here,
14381 after building the SLP tree.
14383 2013-04-09 Christian Bruel <christian.bruel@st.com>
14385 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
14386 of next/prev_real_insn.
14388 2013-04-09 Jan Hubicka <jh@suse.cz>
14390 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14391 Drop aliased parameter.
14392 (function_and_variable_visibility): Do not handle alias pairs.
14393 * cgraph.c (varpool_externally_visible_p): Update prototype.
14394 * varpool.c (varpool_add_new_variable): Update.
14396 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14398 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
14400 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
14402 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
14404 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
14406 2013-04-09 Marek Polacek <polacek@redhat.com>
14408 PR tree-optimization/48762
14409 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
14411 2013-04-09 Richard Biener <rguenther@suse.de>
14413 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
14415 (vect_build_slp_tree): Likewise.
14416 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
14417 calculating the cost of a SLP instance.
14418 (vect_analyze_slp_instance): Use it from here, after building
14421 2013-04-09 Jakub Jelinek <jakub@redhat.com>
14423 PR middle-end/56883
14424 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
14425 expand_omp_for_static_chunk): Use simple_p = true in
14426 force_gimple_operand_gsi calls when assigning to addressable decls.
14428 2013-04-09 Jeff Law <law@redhat.com>
14430 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
14431 when the boolean was created by converting a wider object which
14432 had a boolean range.
14434 2013-04-09 Richard Biener <rguenther@suse.de>
14436 * tree-vectorizer.h (slp_void_p): Remove.
14437 (slp_tree): Typedef before _slp_tree declaration.
14438 (struct _slp_tree): Use a vector of slp_tree as children.
14439 (vect_get_place_in_interleaving_chain): Remove.
14440 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
14442 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
14444 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
14445 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
14446 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
14447 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
14448 Use slp_node instead of slp_void_p and adjust.
14450 2013-04-09 Richard Biener <rguenther@suse.de>
14452 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
14453 work that is not necessary.
14455 2013-04-09 Jakub Jelinek <jakub@redhat.com>
14457 PR tree-optimization/56854
14458 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
14459 forward into clobber stmts if it would change MEM_REF lhs into
14462 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
14464 * tree.c (type_hash_lookup, type_hash_add): Make static.
14465 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
14467 2013-04-09 Richard Biener <rguenther@suse.de>
14469 * tree.h (unsave_expr_now): Remove.
14470 * tree-inline.c (mark_local_for_remap_r): Remove.
14471 (unsave_expr_1): Likewise.
14472 (unsave_r): Likewise.
14473 (unsave_expr_now): Likewise.
14474 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
14475 (propagate_tree_value): Likewise.
14477 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
14479 * doc/rtl.texi (sequence): Rewrite documentation to match the
14480 current use of SEQUENCE rtl objects.
14481 * rtl.def (SEQUENCE): Likewise.
14483 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
14484 Update documentation.
14485 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
14486 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
14488 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
14490 2013-04-08 Teresa Johnson <tejohnson@google.com>
14492 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
14493 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
14495 (estimate_edge_size_and_time): Add comment to suggest using rounding
14497 (estimate_node_size_and_time): Ditto.
14498 (remap_edge_change_prob): Use helper rounding divide methods.
14499 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
14500 (gimple_mod_pow2_value_transform): Ditto.
14501 (gimple_mod_subtract_transform): Ditto.
14502 (gimple_ic_transform): Ditto.
14503 (gimple_stringops_transform): Ditto.
14504 * stmt.c (conditional_probability): Ditto.
14505 (emit_case_dispatch_table): Ditto.
14506 * lto-cgraph.c (merge_profile_summaries): Ditto.
14507 * tree-optimize.c (execute_fixup_cfg): Ditto.
14508 * cfgcleanup.c (try_forward_edges): Ditto.
14509 * cfgloopmanip.c (scale_loop_profile): Ditto.
14511 (duplicate_loop_to_header_edge): Ditto.
14512 (lv_adjust_loop_entry_edge): Ditto.
14513 * tree-vect-loop.c (vect_transform_loop): Ditto.
14514 * profile.c (compute_branch_probabilities): Ditto.
14515 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
14516 * lto-streamer-in.c (input_cfg): Ditto.
14517 * gimple-streamer-in.c (input_bb): Ditto.
14518 * ipa-cp.c (update_profiling_info): Ditto.
14519 (update_specialized_profile): Ditto.
14520 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
14521 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
14523 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14524 (compute_trg_info): Ditto.
14525 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
14526 (purge_dead_edges): Ditto.
14527 * loop-unswitch.c (unswitch_loop): Ditto.
14528 * cgraphclones.c (cgraph_clone_edge): Ditto.
14529 (cgraph_clone_node): Ditto.
14530 * tree-inline.c (copy_bb): Ditto.
14531 (copy_edges_for_bb): Ditto.
14532 (initialize_cfun): Ditto.
14533 (copy_cfg_body): Ditto.
14534 (expand_call_inline): Ditto.
14536 2013-04-08 Kai Tietz <ktietz@redhat.com>
14538 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
14539 TARGET_CYGWIN64 by TARGET_64BIT.
14541 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14543 * config/epiphany/epiphany.md (GPR_1): New constant.
14544 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
14545 * config/epiphany/epiphany.c (gen_compare_reg):
14546 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
14547 is already in place.
14548 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
14549 Don't require being called during rtl expansion; If y operlaps r0,
14551 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
14552 (epiphany_expand_epilogue): Likewise.
14554 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
14555 Don't use CC_FPmode for ORDERED / UNORDERED.
14556 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
14558 * config/epiphany/constraints.md (CnL): New constraint.
14559 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
14560 * config/epiphany/predicates.md (add_operand): Allow 1024.
14562 * config/epiphany/epiphany.md (logical_op): New code iterator.
14563 (op_mnc): New code attribute.
14564 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
14565 (mov_f+1, mov_f+2): New peephole2 patterns.
14567 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
14568 (cstoresi4): Also allow re-use of zero result when doing a NE
14569 comparison to a non-zero operand.
14570 Use (clobber (scratch)) for first insn if the gpr output is not needed.
14572 * config/epiphany/epiphany.md (<insn_opname>v2si3):
14573 Use gen_addsi3_i / gen_subsi3_i.
14575 2013-04-08 Jakub Jelinek <jakub@redhat.com>
14579 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
14580 contain anything but clobbers, at most one __builtin_stack_restore,
14581 optionally debug stmts and final resx, and if it has at least one
14582 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
14583 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
14584 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
14585 which isn't defaut definition, remove them.
14586 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
14587 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
14588 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
14589 with MEM_REF LHS with SSA_NAME address.
14591 2013-04-08 Jeff Law <law@redhat.com>
14593 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
14595 2013-04-08 Richard Biener <rguenther@suse.de>
14597 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
14599 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
14600 determined vector type.
14601 (vect_analyze_data_refs): Likewise.
14602 (vect_get_new_vect_var): Adjust.
14603 (vect_create_destination_var): Preserve SSA name versions.
14604 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
14605 not dump anything here.
14607 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14609 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
14610 Add member lr_slot_known.
14611 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
14613 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14614 Remove code that sets lr_slot_offset according to what a previous
14615 version of epiphany_emit_save_restore used to do.
14616 (epiphany_emit_save_restore): When doing an lr save or restore,
14617 set/verify lr_slot_known and lr_slot_offset.
14619 2013-04-08 Xinyu Qi <xyqi@marvell.com>
14622 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
14625 2013-04-08 Richard Biener <rguenther@suse.de>
14627 * alias.c (find_base_term): Fix thinko in previous change.
14629 2013-04-08 Jakub Jelinek <jakub@redhat.com>
14631 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
14632 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
14633 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
14634 if possible to compute val.
14635 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
14636 For QImode integers don't require anything about precision. Use
14637 const_with_all_bytes_same to find out if the constant doesn't have
14638 repeated bytes in it.
14640 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14642 * config/s390/s390.c (s390_expand_insv): Only accept insertions
14645 2013-04-08 Marek Polacek <polacek@redhat.com>
14647 PR rtl-optimization/48182
14648 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
14651 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14654 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
14655 nuses, make sure we have a label.
14657 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14660 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
14661 (rs6000_emit_swdiv_low_precision): Remove.
14662 (rs6000_emit_swdiv): Rewrite to handle between one and four
14663 iterations of Newton-Raphson generally; modify required number of
14664 iterations for some cases.
14665 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
14667 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
14669 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
14670 set-but-unused variable.
14672 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
14673 basic blocks of released function bodies garbage-collectable.
14675 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
14676 (struct rtl_opt_pass): Add TODO_df_finish.
14678 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
14680 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14682 * config/arm/constraints.md (q): New constraint.
14683 * config/arm/ldrdstrd.md: New file.
14684 * config/arm/arm.md (ldrdstrd.md) New include.
14685 (arm_movdi): Use "q" instead of "r" constraint
14686 for double-word memory access.
14687 (movdf_soft_insn): Likewise.
14688 * config/arm/vfp.md (movdi_vfp): Likewise.
14689 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
14690 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
14691 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
14692 (mem_ok_for_ldrd_strd): Likewise.
14693 (output_move_double): Update assertion.
14695 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14697 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
14699 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14701 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
14702 define_insn_and_split.
14703 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
14705 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14707 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
14708 define_insn_and_split.
14709 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
14710 (shiftsi3_compare): New pattern.
14711 (rrx): New pattern.
14712 * config/arm/unspecs.md (UNSPEC_RRX): New.
14714 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14716 * config/arm/arm.md (negdi_extendsidi): New pattern.
14717 (negdi_zero_extendsidi): Likewise.
14719 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14721 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
14722 define_insn_and_split.
14723 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
14724 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
14726 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14728 * config/arm/arm.md (arm_subdi3): Convert define_insn into
14729 define_insn_and_split.
14730 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
14731 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
14733 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14735 * config/arm/arm.md (subsi3_carryin): New pattern.
14736 (subsi3_carryin_const): Likewise.
14737 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
14738 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
14740 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14742 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
14744 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14746 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
14747 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
14749 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14751 * config/arm/arm.c (arm_expand_builtin): Change fcode
14752 type to unsigned int.
14754 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14756 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
14758 2013-04-04 Ian Lance Taylor <iant@google.com>
14760 * doc/standards.texi (Standards): The Go frontend supports the Go 1
14763 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
14765 PR middle-end/56729
14766 * df-scan.c (df_insn_delete): Disable failing assert.
14768 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14770 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
14771 New function prototype.
14772 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
14773 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
14774 (arm_builtin_vectorized_function): New function.
14776 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14778 * config/arm/arm_neon_builtins.def: New file.
14779 * config/arm/arm.c (neon_builtin_data): Move contents to
14780 arm_neon_builtins.def.
14781 (enum arm_builtins): Include neon builtin definitions.
14782 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
14783 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
14785 2013-04-04 Marek Polacek <polacek@redhat.com>
14787 PR tree-optimization/48186
14788 * predict.c (maybe_hot_frequency_p): Return false if
14789 HOT_BB_FREQUENCY_FRACTION is 0.
14790 (cgraph_maybe_hot_edge_p): Likewise.
14792 2013-04-04 Richard Biener <rguenther@suse.de>
14794 PR tree-optimization/56826
14795 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
14798 2013-04-04 Richard Biener <rguenther@suse.de>
14800 PR tree-optimization/56213
14801 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
14802 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
14804 2013-04-04 Richard Biener <rguenther@suse.de>
14806 PR tree-optimization/56837
14807 * tree-loop-distribution.c (classify_partition): For non-zero
14808 values require that the value has the same precision as its
14809 mode to be useful as memset value.
14811 2013-04-03 Nick Clifton <nickc@redhat.com>
14813 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
14814 (fmssf4): Use fmsf.s on E3V5 architectures.
14815 (fnmasf4): Use fnmaf.s on E3V5 architectures.
14816 (fnmssf4): Use fnmsf.s on E3V5 architectures.
14818 2013-04-03 Jeff Law <law@redhat.com>
14820 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
14821 (lra-eliminations.o): Likewise.
14823 2013-04-03 Teresa Johnson <tejohnson@google.com>
14825 * gcov-io.c (compute_working_sets): Moved most of body of old
14826 compute_working_sets here from profile.c.
14827 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
14828 (gcov_working_set_t): Moved typedef here from basic-block.h
14829 (compute_working_set): Declare.
14830 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
14831 (get_working_sets): Renamed from compute_working_set,
14832 replace most of body with call to new compute_working_sets.
14833 (get_exec_counts): Replace call to compute_working_sets
14834 to get_working_sets.
14835 * profile.h (get_working_sets): Renamed from compute_working_set.
14836 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
14837 to get_working_sets.
14838 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
14839 * gcov-dump.c (dump_working_sets): New function.
14841 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
14843 * hwint.c (sext_hwi, zext_hwi): New functions.
14844 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
14845 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
14846 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
14847 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
14848 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
14849 (sext_hwi, zext_hwi): New functions.
14851 2013-04-03 Jeff Law <law@redhat.com>
14853 PR tree-optimization/56799
14854 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
14855 back test for widening conversion erroneously dropped in prior change.
14857 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14860 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
14861 instead of next_real_insn.
14863 2013-04-03 Marek Polacek <polacek@redhat.com>
14866 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
14868 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14871 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
14873 (thumb1_output_casesi): Likewise.
14874 (thumb2_output_casesi): Likewise.
14876 2013-04-03 Richard Biener <rguenther@suse.de>
14878 PR tree-optimization/56817
14879 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
14881 (tree_unroll_loops_completely_1): ... new function to manually
14882 walk the loop tree, properly defering outer loops of unrolled
14883 loops to later iterations.
14885 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14887 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
14888 (vectorizable_load): Likewise.
14889 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
14890 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
14892 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14894 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
14897 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
14899 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
14901 2013-04-03 Bin Cheng <bin.cheng@arm.com>
14903 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
14905 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14907 PR tree-optimization/56790
14908 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
14911 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14913 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
14915 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
14916 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
14919 2013-04-03 Jakub Jelinek <jakub@redhat.com>
14922 * tree.h (force_folding_builtin_constant_p): New decl.
14923 * builtins.c (force_folding_builtin_constant_p): New variable.
14924 (fold_builtin_constant_p): Fold immediately also if
14925 force_folding_builtin_constant_p.
14927 2013-04-03 Richard Biener <rguenther@suse.de>
14929 PR tree-optimization/56812
14930 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
14931 DRs of the same interleaving chain are independent.
14933 2013-04-02 Jason Merrill <jason@redhat.com>
14935 * gdbinit.in (pbb): Use debug fn.
14937 2013-04-02 Lawrence Crowl <crowl@google.com>
14939 * sese.h (struct ivtype_map_elt_s): Remove unused.
14940 (extern debug_ivtype_map): Remove unused.
14941 (extern eq_ivtype_map_elts): Remove unused.
14942 * sese.c (debug_ivtype_map): Removed unused.
14943 (debug_ivtype_map_1): Removed unused.
14944 (debug_ivtype_elt): Remove unused.
14945 (eq_ivtype_map_elts): Remove unused.
14948 2013-04-02 Kai Tietz <ktietz@redhat.com>
14951 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
14952 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
14953 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
14955 (legitimize_pe_coff_symbol): Likewise.
14956 (is_imported_p): New helper-function.
14957 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
14958 for Windows x64 targets.
14959 (ix86_expand_prologue): Optimize for pe-coff targets.
14960 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
14961 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
14962 medium/large code-model.
14963 (legitimize_pic_address): Likewise.
14964 (legitimize_tls_address): Likewise.
14965 (ix86_expand_call): Likewise.
14966 (x86_output_mi_thunk): Likewise.
14967 (get_dllimport_decl): Add new beimport argument.
14968 (construct_plt_address): Don't assert for x64 pe-coff targets.
14969 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
14971 (SYMBOL_FLAG_STUBVAR): New macro.
14972 (SYMBOL_REF_STUBVAR_P): Likewise.
14973 * config/i386/winnt.c (stub_list): New structure.
14974 (stub_head): New local variable.
14975 (i386_pe_record_stub): New function.
14976 (i386_pe_file_end): Emit refptr-stubs.
14978 2013-04-02 Jakub Jelinek <jakub@redhat.com>
14980 PR rtl-optimization/56745
14981 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
14982 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
14985 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
14986 and both of them are MEM_REFs, just compare first argument for
14987 equality and attempt to deal even with differing offsets.
14990 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
14991 of gimple_clobber_p to be MEM_REF.
14992 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
14993 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
14994 after gimplification.
14995 * asan.c (get_mem_ref_of_assignment): Don't instrument
14996 gimple_clobber_p stmts.
14997 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
14998 gimple_clobber_p stmt if they have MEM_REF lhs and
14999 are dead because of another gimple_clobber_p stmt.
15000 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15001 gimple_clobber_p stmts like debug stmts.
15002 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15003 that refer to unused VAR_DECLs or uninitialized values.
15004 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15005 gimple_clobber_p stmts if they refer to removed parameters.
15006 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15009 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
15011 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15012 using SWI48 mode attribute.
15014 2013-04-02 Wei Mi <wmi@google.com>
15016 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15017 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15018 *<rotate_insn><mode>3_mask in i386.md.
15020 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15022 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15024 2013-04-02 Richard Biener <rguenther@suse.de>
15026 PR tree-optimization/56778
15027 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15028 Runtime alias tests are not supported for gather loads.
15029 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15030 stmts referenced from SSA operands before updating SSA form.
15032 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15033 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15035 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
15036 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15037 * config/arm/cortex-a53.md: New file.
15038 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15039 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15040 * config/arm/arm.c (arm_issue_rate): Likewise.
15041 * config/arm/arm-tune.md: Regenerate
15042 * config/arm/arm-tables.opt: Regenerate.
15043 * config/arm/arm-cores.def: Add cortex-a53.
15045 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15047 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15050 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15052 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15053 scalar load/store operations using B/H registers.
15054 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15056 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15058 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15060 * config/aarch64/aarch64.c
15061 (aarch64_simd_scalar_immediate_valid_for_move): New.
15062 * config/aarch64/aarch64-protos.h
15063 (aarch64_simd_scalar_immediate_valid_for_move): New.
15064 * config/aarch64/constraints.md (Dh, Dq): New.
15065 * config/aarch64/iterators.md (hq): New.
15067 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15069 * reorg.c (get_branch_condition): Deal with conditional returns.
15070 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15072 2013-04-01 Wei Mi <wmi@google.com>
15074 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15075 Truncate operand 2 using %b asm operand modifier.
15076 (*<shift_insn><mode>3_mask): Ditto.
15077 (*<rotate_insn><mode>3_mask): Ditto.
15079 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15081 PR middle-end/56798
15082 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15084 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15086 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15088 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15090 2013-03-30 Lawrence Crowl <crowl@google.com>
15092 * dse.c (clear_alias_sets): Remove never set.
15093 (disqualified_clear_alias_sets): Remove never set.
15094 (clear_alias_mode_pool): Remove never set.
15095 (dse_step0): Remove condition that is never true.
15096 (canon_address): Remove condition that is never true.
15097 (dse_step7): Remove condition that is never true.
15098 (rest_of_handle_dse): Remove condition that is never true.
15099 (rest_of_handle_dse::did_global): Remove never read from above.
15100 (dse_step2_spill): Remove never called from above.
15101 (dse_step5_spill): Remove never called from above.
15103 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15105 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15106 JUMP_TABLE_DATA changes.
15107 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15108 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15109 (Insns) <jump_table_data>: New entry.
15110 * doc/tm.texi: Regenerate.
15112 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15114 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15115 for table jump at the end of a basic block using tablejump_p.
15116 * targhooks.c (default_invalid_within_doloop): Likewise.
15117 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15118 target hook implementation that is identical to the default hook.
15119 (rs6000_invalid_within_doloop): Remove.
15121 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15122 unused variable from tablejump_p call.
15124 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15125 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15126 (INSN_DELETED_P): Likewise.
15127 (emit_jump_table_data): New prototype.
15128 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15129 after 4th as unused.
15130 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15131 * sched-vis.c (print_insn): Likewise.
15132 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15133 insn for compatibility with back ends that use next_active_insn to
15134 identify jump table data.
15135 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15136 (remove_insn): Likewise.
15137 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15139 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15140 (emit_jump_table_data): New function.
15142 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15143 basic block, a JUMP_TABLE_DATA never is.
15144 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15145 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15146 off from code handling real insns.
15147 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15148 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15149 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15150 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15151 is not a NONDEBUG_INSN_P.
15152 * ira-costs.c (scan_one_insn): Likewise.
15153 * jump.c (mark_all_labels): Likewise.
15154 (mark_jump_label_1): Likewise.
15155 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15156 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15158 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15159 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15160 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15161 JUMP_TABLE_DATA_P insns.
15162 (calculate_elim_costs_all_insns): Likewise.
15163 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15164 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15165 (delete_output_reload): Code style fixups.
15166 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15167 insn flags on this non-insn.
15168 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15169 as scheduling barriers, for pre-change compatibility.
15170 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15171 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15173 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15174 redundant JUMP_TABLE_DATA_P test.
15175 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15176 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15177 (frv_for_each_packet): Likewise.
15178 * config/i386/i386.c (min_insn_size): Likewise.
15179 (ix86_avoid_jump_mispredicts): Likewise.
15180 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15181 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15182 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15183 (mips16_insn_length): Robustify.
15184 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15185 (mips16_split_long_branches): Likewise.
15186 * config/pa/pa.c (pa_combine_instructions): Likewise.
15187 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15188 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15189 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15190 as contributing to pool range lengths.
15191 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15192 Remove redundant JUMP_TABLE_DATA_P test.
15193 (sh_loop_align): Likewise.
15194 (split_branches): Likewise.
15195 (sh_insn_length_adjustment): Likewise.
15196 * config/spu/spu.c (get_branch_target): Likewise.
15198 2013-03-29 Jan Hubicka <jh@suse.cz>
15200 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15201 gcov streaming; stream hot bb threshold to ltrans.
15202 * predict.c (get_hot_bb_threshold): Break out from ....
15203 (maybe_hot_count_p): ... here.
15204 (set_hot_bb_threshold): New function.
15205 * lto-section-in.c (lto_section_name): Add profile.
15206 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15207 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15208 and data-streamer.h
15209 (histogram_entry): New structure.
15210 (histogram, histogram_pool): New global vars.
15211 (histogram_hash): New structure.
15212 (histogram_hash::hash): New method.
15213 (histogram_hash::equal): Likewise.
15214 (account_time_size): New function.
15215 (cmp_counts): New function.
15216 (dump_histogram): New function.
15217 (ipa_profile_generate_summary): New function.
15218 (ipa_profile_write_summary): New function.
15219 (ipa_profile_read_summary): New function.
15220 (ipa_profile): Decide on threshold.
15221 (pass_ipa_profile): Add ipa_profile_write_summary and
15222 ipa_profile_read_summary.
15223 * Makefile.in (ipa.o): Update dependencies.
15224 * lto-streamer.h (LTO_section_ipa_profile): New section.
15226 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15228 * tree.h (VAR_P): New.
15230 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15233 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15235 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15237 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15238 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15239 (control_flow_insn_p): Likewise.
15240 * cfgrtl.c (duplicate_insn_chain): Likewise.
15241 * final.c (get_attr_length_1): Likewise.
15242 (shorten_branches): Likewise.
15243 (final_scan_insn): Likewise.
15244 * function.c (instantiate_virtual_regs): Likewise.
15245 * gcse.c (insert_insn_end_basic_block): Likewise.
15246 * ira-costs.c (scan_one_insn): Likewise.
15247 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15248 * lra.c (check_rtl): Likewise.
15249 * reload1.c (elimination_costs_in_insn): Likewise.
15250 * reorg.c (follow_jumps): Likewise.
15252 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15253 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15254 (thumb_far_jump_used_p): Likewise.
15255 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15256 (workaround_speculation): Likewise.
15257 (add_sched_insns_for_speculation): Likewise.
15258 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15259 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15260 (frv_for_each_packet): Likewise.
15261 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15262 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15263 (final_emit_insn_group_barriers): Likewise.
15264 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15265 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15266 (mips16_insn_length): Likewise.
15267 * config/pa/pa.c (pa_reorg): Likewise.
15268 (pa_combine_instructions): Likewise.
15269 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15270 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15271 (sh_reorg): Likewise.
15272 (split_branches): Likewise.
15273 * config/spu/spu.c (get_branch_target): Likewise.
15275 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15278 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15280 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
15281 Fix declaration name.
15283 2013-03-28 Lawrence Crowl <crowl@google.com>
15285 * graphds.h (struct graph.indicies): Remove unused.
15286 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15287 (SCOP_ORIGINAL_PDDRS): Remove unused.
15288 * sese.h (extern insert_loop_close_phis): Removed unused.
15289 (extern insert_guard_phis): Removed unused.
15290 (extern ivtype_map_elt_info): Removed unused.
15291 (new_ivtype_map_elt): Removed unused.
15292 * sese.c (ivtype_map_elt_info): Removed unused.
15294 2013-03-28 Lawrence Crowl <crowl@google.com>
15296 * Makefile.in: Add several missing include dependences.
15298 (test-dump.o): New. This object is not added to any executable,
15299 but is present for ad-hoc testing.
15301 (debug (const bitmap_head_def &)): New.
15302 (debug (const bitmap_head_def *)): New.
15304 (extern debug (const bitmap_head_def &)): New.
15305 (extern debug (const bitmap_head_def *)): New.
15307 (debug (edge_def &)): New.
15308 (debug (edge_def *)): New.
15310 (debug (basic_block_def &)): New.
15311 (debug (basic_block_def *)): New.
15313 (dump_node (const_tree, int, FILE *)): Correct source file.
15315 (debug (die_struct &)): New.
15316 (debug (die_struct *)): New.
15318 (extern debug (die_struct &)): New.
15319 (extern debug (die_struct *)): New.
15320 * gimple-pretty-print.c
15321 (debug (gimple_statement_d &)): New.
15322 (debug (gimple_statement_d *)): New.
15323 * gimple-pretty-print.h
15324 (extern debug (gimple_statement_d &)): New.
15325 (extern debug (gimple_statement_d *)): New.
15327 (debug (ira_allocno_copy &)): New.
15328 (debug (ira_allocno_copy *)): New.
15329 (debug (ira_allocno &)): New.
15330 (debug (ira_allocno *)): New.
15332 (extern debug (ira_allocno_copy &)): New.
15333 (extern debug (ira_allocno_copy *)): New.
15334 (extern debug (ira_allocno &)): New.
15335 (extern debug (ira_allocno *)): New.
15337 (debug (live_range &)): New.
15338 (debug (live_range *)): New.
15340 (debug (lra_live_range &)): New.
15341 (debug (lra_live_range *)): New.
15343 (debug (lra_live_range &)): New.
15344 (debug (lra_live_range *)): New.
15346 (debug (omega_pb_d &)): New.
15347 (debug (omega_pb_d *)): New.
15349 (extern debug (omega_pb_d &)): New.
15350 (extern debug (omega_pb_d *)): New.
15352 (debug (const rtx_def &)): New.
15353 (debug (const rtx_def *)): New.
15355 (debug_tree (tree): Move within file.
15356 (debug_raw (const tree_node &)): New.
15357 (debug_raw (const tree_node *)): New.
15358 (dump_tree_via_hooks (const tree_node *, int)): New.
15359 (debug (const tree_node &)): New.
15360 (debug (const tree_node *)): New.
15361 (debug_verbose (const tree_node &)): New.
15362 (debug_verbose (const tree_node *)): New.
15363 (debug_head (const tree_node &)): New.
15364 (debug_head (const tree_node *)): New.
15365 (debug_body (const tree_node &)): New.
15366 (debug_body (const tree_node *)): New.
15367 (debug_vec_tree (tree): Move and reimplement in terms of dump.
15368 (debug (vec<tree, va_gc> &)): New.
15369 (debug (vec<tree, va_gc> *)): New.
15371 (extern debug (const rtx_def &)): New.
15372 (extern debug (const rtx_def *)): New.
15374 (debug_raw (simple_bitmap_def &)): New.
15375 (debug_raw (simple_bitmap_def *)): New.
15376 (debug (simple_bitmap_def &)): New.
15377 (debug (simple_bitmap_def *)): New.
15379 (extern debug (simple_bitmap_def &)): New.
15380 (extern debug (simple_bitmap_def *)): New.
15381 (extern debug_raw (simple_bitmap_def &)): New.
15382 (extern debug_raw (simple_bitmap_def *)): New.
15384 (debug (vinsn_def &)): New.
15385 (debug (vinsn_def *)): New.
15386 (debug_verbose (vinsn_def &)): New.
15387 (debug_verbose (vinsn_def *)): New.
15388 (debug (expr_def &)): New.
15389 (debug (expr_def *)): New.
15390 (debug_verbose (expr_def &)): New.
15391 (debug_verbose (expr_def *)): New.
15392 (debug (vec<rtx> &)): New.
15393 (debug (vec<rtx> *)): New.
15395 (extern debug (vinsn_def &)): New.
15396 (extern debug (vinsn_def *)): New.
15397 (extern debug_verbose (vinsn_def &)): New.
15398 (extern debug_verbose (vinsn_def *)): New.
15399 (extern debug (expr_def &)): New.
15400 (extern debug (expr_def *)): New.
15401 (extern debug_verbose (expr_def &)): New.
15402 (extern debug_verbose (expr_def *)): New.
15403 (extern debug (vec<rtx> &)): New.
15404 (extern debug (vec<rtx> *)): New.
15406 (_list_iter_cond_expr): Make inline instead of static.
15408 (debug (sreal &)): New.
15409 (debug (sreal *)): New.
15411 (extern debug (sreal &)): New.
15412 (extern debug (sreal *)): New.
15414 (extern debug_raw (const tree_node &)): New.
15415 (extern debug_raw (const tree_node *)): New.
15416 (extern debug (const tree_node &)): New.
15417 (extern debug (const tree_node *)): New.
15418 (extern debug_verbose (const tree_node &)): New.
15419 (extern debug_verbose (const tree_node *)): New.
15420 (extern debug_head (const tree_node &)): New.
15421 (extern debug_head (const tree_node *)): New.
15422 (extern debug_body (const tree_node &)): New.
15423 (extern debug_body (const tree_node *)): New.
15424 (extern debug (vec<tree, va_gc> &)): New.
15425 (extern debug (vec<tree, va_gc> *)): New.
15427 (debug (struct loop &)): New.
15428 (debug (struct loop *)): New.
15429 (debug_verbose (struct loop &)): New.
15430 (debug_verbose (struct loop *)): New.
15431 * tree-dump.c: Add header dependence.
15433 (extern debug (struct loop &)): New.
15434 (extern debug (struct loop *)): New.
15435 (extern debug_verbose (struct loop &)): New.
15436 (extern debug_verbose (struct loop *)): New.
15438 (debug (data_reference &)): New.
15439 (debug (data_reference *)): New.
15440 (debug (vec<data_reference_p> &)): New.
15441 (debug (vec<data_reference_p> *)): New.
15442 (debug (vec<ddr_p> &)): New.
15443 (debug (vec<ddr_p> *)): New.
15445 (extern debug (data_reference &)): New.
15446 (extern debug (data_reference *)): New.
15447 (extern debug (vec<data_reference_p> &)): New.
15448 (extern debug (vec<data_reference_p> *)): New.
15449 (extern debug (vec<ddr_p> &)): New.
15450 (extern debug (vec<ddr_p> *)): New.
15452 (debug (pt_solution &)): New.
15453 (debug (pt_solution *)): New.
15455 (extern debug (pt_solution &)): New.
15456 (extern debug (pt_solution *)): New.
15458 (debug (_var_map &)): New.
15459 (debug (_var_map *)): New.
15460 (debug (tree_live_info_d &)): New.
15461 (debug (tree_live_info_d *)): New.
15463 (extern debug (_var_map &)): New.
15464 (extern debug (_var_map *)): New.
15465 (extern debug (tree_live_info_d &)): New.
15466 (extern debug (tree_live_info_d *)): New.
15468 2013-03-28 Jan Hubicka <jh@suse.cz>
15470 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
15472 2013-03-28 Ian Bolton <ian.bolton@arm.com>
15474 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
15475 record only when desired or required.
15477 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
15479 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
15480 *vec_extractv2di_1_rex64. Use x64 isa attribute.
15482 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15484 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
15485 (*andsi3_compare0_uxtw): New pattern.
15486 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
15487 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
15489 2013-03-28 Jan Hubicka <jh@suse.cz>
15491 * data-streamer-in.c (streamer_read_gcov_count): New function.
15492 * gimple-streamer-out.c: Include value-prof.h.
15493 (output_gimple_stmt): Output histogram.
15494 (output_bb): Use streamer_write_gcov_count.
15495 * value-prof.c: Include data-streamer.h
15496 (dump_histogram_value): Add HIST_TYPE_MAX.
15497 (stream_out_histogram_value): New function.
15498 (stream_in_histogram_value): New function.
15499 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
15500 (stream_out_histogram_value, stream_in_histogram_value): Declare.
15501 * data-streamer-out.c (streamer_write_gcov_count): New function.
15502 (streamer_write_gcov_count_stream): New function.
15503 * lto-cgraph.c (lto_output_edge): Update counter streaming.
15504 (lto_output_node): Likewise.
15505 (input_node, input_edge): Likewise.
15506 * lto-streamer-out.c (output_cfg): Update streaming.
15507 * lto-streamer-in.c (input_cfg): Likewise.
15508 * data-streamer.h (streamer_write_gcov_count,
15509 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
15510 * gimple-streamer-in.c: Include value-prof.h
15511 (input_gimple_stmt): Input histograms.
15512 (input_bb): Update profile streaming.
15514 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
15516 * genmodes.c (emit_max_int): New function.
15517 (emit_insn_modes_h): Added call to emit_max_function.
15518 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
15520 * machmode.def: Fixed comment.
15522 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
15524 * combine.c (try_combine): Removed useless assert.
15525 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
15527 2013-03-28 Marek Polacek <polacek@redhat.com>
15528 Richard Biener <rguenther@suse.de>
15530 PR tree-optimization/56695
15531 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
15532 build signed result of a vector comparison.
15533 * tree-cfg.c (verify_gimple_comparison): Check that a result
15534 of a vector comparison has signed type.
15536 2013-03-28 Richard Biener <rguenther@suse.de>
15538 PR tree-optimization/37021
15539 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
15540 do not restrict gaps between groups.
15541 * tree-vect-stmts.c (vectorizable_load): Properly account for
15542 a gap between groups.
15544 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
15546 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
15547 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
15550 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
15552 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
15553 * doc/extend.texi (Named Address Spaces): Ditto.
15554 (Variable Attributes): Ditto.
15556 2013-03-27 Kai Tietz <ktietz@redhat.com>
15558 * config.build: Add support for cygwin x64 target.
15559 * config.gcc: Likewise.
15560 * config.host: Likewise.
15561 * configure.ac: Likewise
15562 * configure: Regenerated.
15564 2013-03-27 Kai Tietz <ktietz@redhat.com>
15566 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
15567 * config/i386/t-cygwin-w64: New file.
15568 * config/i386/cygwin-w64.h: New file.
15569 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
15570 and add support for x64-cygwin target.
15571 (CPP_SPEC): Likewise.
15572 (CXX_WRAP_SPEC_LIST): Undefine before define.
15573 (LIBGCJ_SONAME): Use 15 as version.
15575 2013-03-27 Richard Biener <rguenther@suse.de>
15577 PR tree-optimization/56716
15578 * tree-ssa-structalias.c (perform_var_substitution): Adjust
15579 dumping for ref nodes.
15581 2013-03-27 Martin Jambor <mjambor@suse.cz>
15583 PR tree-optimization/55334
15584 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
15585 restricted pointers to arrays.
15587 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
15589 * Makefile.in (.SUFFIXES): Add .cc.
15590 (.c.o): Apply same recipe for implicit rule .cc.o.
15592 2013-03-27 Richard Biener <rguenther@suse.de>
15594 PR tree-optimization/37021
15595 * tree-vect-data-refs.c (vect_check_strided_load): Allow
15596 REALPART/IMAGPART_EXPRs around the supported refs.
15597 * tree-ssa-structalias.c (find_func_aliases): Assume that
15598 floating-point values are not used to transfer pointers.
15600 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
15602 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
15603 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
15604 * gcc/doc/tm.texi: Regenerate.
15605 * gcc/targhooks.h (default_has_ifunc_p): New.
15606 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
15607 * gcc/config/linux-protos.h: New file.
15608 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
15609 this hook for linux which disables support of indirect functions in
15611 * gcc/config/linux-android.c: New file.
15612 * gcc/config/t-linux-android.c: Ditto.
15613 * gcc/config.gcc: Added new object file linux-android.o.
15614 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
15615 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
15616 * gcc/varasm.c (do_assemble_alias): Likewise.
15617 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
15618 doesn't support indirect functions.
15619 * configure: Regenerate.
15621 2013-03-27 Bin Cheng <bin.cheng@arm.com>
15624 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
15625 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
15628 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15630 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
15632 2013-03-27 Terry Guo <terry.guo@arm.com>
15634 * config/arm/arm-cores.def: Added core cortex-r7.
15635 * config/arm/arm-tune.md: Regenerated.
15636 * config/arm/arm-tables.opt: Regenerated.
15637 * doc/invoke.texi: Added entry for core cortex-r7.
15639 2013-03-27 Walter Lee <walt@tilera.com>
15641 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
15642 double-decrement of next_scratch_regno.
15644 2013-03-27 Walter Lee <walt@tilera.com>
15646 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
15648 (insn_v1mulus): Ditto.
15649 (insn_v2muls): Ditto.
15651 2013-03-27 Walter Lee <walt@tilera.com>
15653 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
15654 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
15656 2013-03-27 Walter Lee <walt@tilera.com>
15658 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
15659 (*sibcall_value): Ditto.
15661 2013-03-27 Walter Lee <walt@tilera.com>
15663 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
15664 (insn_mnz_v8qi): ... this ...
15665 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
15667 (insn_v<n>mnz): Replaced by ...
15668 (insn_v1mnz): ... this ...
15669 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
15671 (insn_mz_<mode>): Replaced by ...
15672 (insn_mz_v8qi): ... this ...
15673 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
15675 (insn_v<n>mz): Replaced by ...
15676 (insn_v1mz): ... this ...
15677 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
15680 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
15682 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
15684 2013-03-26 Roland McGrath <mcgrathr@google.com>
15686 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
15687 than fprintf with a non-constant, non-format string.
15689 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
15691 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
15692 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
15693 operand 0 predicate.
15694 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
15695 attribute. Use general_x64nomem_operand as operand 1 predicate.
15696 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
15697 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15698 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
15699 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15700 (mov<mode>_insv_1): Remove expander. Merge insn with
15701 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
15702 Use general_x64nomem_operand as operand 1 predicate.
15703 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
15704 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
15705 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15706 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
15707 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
15708 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
15709 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15710 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
15711 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
15712 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
15713 (general_x64nomem_operand): Ditto.
15715 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
15717 * config/rtems.opt: Add -pthread option.
15719 2013-03-26 Richard Biener <rguenther@suse.de>
15721 * alias.c (find_base_term): Avoid redundant and not used recursion.
15722 (base_alias_check): Get the initial base term from the caller.
15723 (true_dependence_1): Compute and pass base terms to base_alias_check.
15724 (write_dependence_p): Likewise.
15725 (may_alias_p): Likewise.
15727 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
15729 * config/aarch64/aarch64.c (aarch64_classify_address): Support
15730 PC-relative load in SI modes and above only.
15732 2013-03-26 Xinyu Qi <xyqi@marvell.com>
15734 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
15735 * config/arm/iwmmxt.md (WCGR0): Update.
15736 (WCGR1, WCGR2, WCGR3): Likewise.
15738 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
15740 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
15741 Use x64 and nox64 isa attributes.
15743 2013-03-26 Richard Biener <rguenther@suse.de>
15745 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
15746 alignment computations and rely on get_object_alignment_1
15747 for the !TYPE_P case.
15748 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
15750 2013-03-26 Walter Lee <walt@tilera.com>
15752 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
15753 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
15755 2013-03-25 Jeff Law <law@redhat.com>
15757 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
15758 check for INTEGRAL_TYPE_P that was missing due to checking in
15759 wrong version of prior patch.
15761 2013-03-25 Walter Lee <walt@tilera.com>
15763 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
15764 TILEGX_INSN_SHUFFLEBYTES1.
15765 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
15767 (tilegx_builtins): Ditto.
15768 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
15770 2013-03-25 Walter Lee <walt@tilera.com>
15772 * config/tilegx/tilegx.md (floatsisf2): New pattern.
15773 (floatunssisf2): New pattern.
15774 (floatsidf2): New pattern.
15775 (floatunssidf2): New pattern.
15777 2013-03-25 Walter Lee <walt@tilera.com>
15779 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
15780 tests for constraint J, K, N, P.
15782 2013-03-25 Walter Lee <walt@tilera.com>
15784 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
15785 Use indirect/pcrel encoding.
15786 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
15789 2013-03-25 Steve Ellcey <sellcey@mips.com>
15791 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
15792 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
15793 * config/mips/mips.c (mips_option_override): Set IMADD default.
15794 * config/mips/mips.h (PTF_AVOID_IMADD): New.
15795 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
15796 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
15797 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
15798 * doc/invoke.texi (-mimadd/-mno-imadd): New.
15800 2013-03-25 Jeff Law <law@redhat.com>
15802 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
15803 slightly to avoid creating and folding useless trees. Simplify
15804 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
15806 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
15808 * config/i386/i386.md (*zero_extendsidi2): Merge with
15809 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
15810 * config/i386/predicates.md (x86_64_zext_operand): Rename from
15811 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
15812 targets. Clarify comment.
15814 2013-03-25 Martin Jambor <mjambor@suse.cz>
15816 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
15817 pass-through jump functions differently.
15818 (ipa_read_jump_function): Likewise. Also use setter functions to set
15821 2013-03-25 Martin Jambor <mjambor@suse.cz>
15823 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
15824 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
15826 (ipa_get_indirect_edge_target): New function.
15827 (devirtualization_time_bonus): New parameter known_aggs, pass it to
15828 ipa_get_indirect_edge_target. Update all callers.
15829 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
15830 ipa_get_indirect_edge_target_1 instead of calling
15831 ipa_get_indirect_edge_target.
15832 (create_specialized_node): Pass aggvlas to
15833 ipcp_discover_new_direct_edges.
15835 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15837 * config/arm/arm.md (f_sels, f_seld): New types.
15838 (*cmov<mode>): New pattern.
15839 * config/arm/predicates.md (arm_vsel_comparison_operator): New
15842 2013-03-25 Kai Tietz <ktietz@redhat.com>
15844 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
15845 POSIX-printf for mingw-hosted builds.
15847 2013-03-25 Richard Biener <rguenther@suse.de>
15849 PR middle-end/56694
15850 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
15851 must-not-throw stmt location.
15853 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15855 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
15856 Emit load-acquire versions when acq is true.
15857 (arm_emit_store_exclusive): Add rel parameter.
15858 Emit store-release versions when rel is true.
15859 (arm_split_compare_and_swap): Use acquire-release instructions
15861 of barriers when appropriate.
15862 (arm_split_atomic_op): Likewise.
15863 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
15864 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
15865 (VUNSPEC_SLX): Likewise.
15866 (VUNSPEC_LDA): Likewise.
15867 (VUNSPEC_STL): Likewise.
15868 * config/arm/sync.md (atomic_load<mode>): New pattern.
15869 (atomic_store<mode>): Likewise.
15870 (arm_load_acquire_exclusive<mode>): Likewise.
15871 (arm_load_acquire_exclusivesi): Likewise.
15872 (arm_load_acquire_exclusivedi): Likewise.
15873 (arm_store_release_exclusive<mode>): Likewise.
15875 2013-03-25 Catherine Moore <clm@codesourcery.com>
15877 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
15878 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
15879 * config/mip/predicates.md (lwsp_swsp_operand,
15880 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
15881 sb16_operand, db4_operand, db7_operand, ib3_operand,
15882 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
15883 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
15884 andi16_operand): New predicates.
15885 * config/mips/mips.md (compression): New attribute.
15886 (enabled): New attribute.
15887 (length): Consider compression in computing length.
15888 (shift_compression): New code attribute.
15889 (*add<mode>3): New operands. Record compression.
15890 (sub<mode>3): Likewise.
15891 (one_cmpl<mode>2): Likewise.
15892 (*and<mode>3): Likewise.
15893 (*ior<mode>3): Likewise.
15894 (unnamed pattern for xor): Likewise.
15895 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
15896 (*<optab><mode>3): Likewise.
15897 (*mov<mode>_internal: Likewise.
15898 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
15899 (mips_unsigned_immediate_p): New.
15900 (umips_lwsp_swsp_address_p): New.
15901 (m16_based_address_p): New.
15902 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
15903 (mips_unsigned_immediate_p): New prototype.
15904 (lwsp_swsp_address_p): New prototype.
15905 (m16_based_address_p): New prototype.
15906 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
15907 (mips_signed_immediate_p): New function.
15908 (m16_based_address_p): New function.
15909 (lwsp_swsp_address_p): New function.
15910 (mips_print_operand_punctuation): Recognize short delay slot insns
15911 for microMIPS.add<mode>3"
15913 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15916 * config/arm/iterators.md (v_cmp_result): New mode attribute.
15917 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
15919 2013-03-25 Richard Biener <rguenther@suse.de>
15921 PR tree-optimization/56689
15922 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
15925 2013-03-25 Richard Biener <rguenther@suse.de>
15927 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
15929 (memory_references): Likewise.
15930 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
15931 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
15932 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
15933 (gather_mem_refs_in_loops): Fold into ...
15934 (analyze_memory_references): ... this. Move initialization
15935 to tree_ssa_lim_initialize.
15936 (fill_always_executed_in): Rename to ...
15937 (fill_always_executed_in_1): ... this.
15938 (fill_always_executed_in): Move contains_call computation to
15939 this new function from ...
15940 (tree_ssa_lim_initialize): ... here.
15941 (tree_ssa_lim): Call fill_always_executed_in.
15943 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15945 * postreload.c (reload_combine): Fix code detecting returns.
15947 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15949 * function.c (emit_use_return_register_into_block): On cc0 targets,
15950 do not emit the sequence between cc0 setter and user.
15952 2013-03-25 Kai Tietz <ktietz@redhat.com>
15954 * config/i386/predicates.md (local_symbolic_operand): Interpret
15955 dll-imported symbols as none-local.
15957 2013-03-25 Richard Biener <rguenther@suse.de>
15959 * tree-ssa-loop-im.c (struct depend): Remove.
15960 (struct lim_aux_data): Make depends a vec of gimples.
15961 (free_lim_aux_data): Adjust.
15962 (add_dependency): Likewise.
15963 (set_level): Likewise.
15965 2013-03-25 Richard Biener <rguenther@suse.de>
15967 PR middle-end/56434
15968 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
15969 the pointer returned by calls with ECF_MALLOC set.
15971 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
15973 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
15975 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
15977 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
15978 using MMXMODE mode iterator.
15979 (*move<mode>_internal): Merge with *movv2sf_internal and
15980 *movv2sf_internal_rex64 using MMXMODE mode iterator.
15982 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15984 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
15985 (record_last_mem_set_info): Likewise.
15987 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
15988 of XNEWVEC followed by memset.
15989 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
15991 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15993 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
15994 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
15995 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
15996 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
15997 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
15998 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
15999 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16000 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16001 BARRIER_P instead of GET_CODE.
16003 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16005 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16006 inaccuracy in the probing code.
16008 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16009 (ctrapdi4): Likewise.
16011 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16013 * calls.c (expand_call): Add missing guard to code handling return
16014 of non-BLKmode structures in MSB.
16015 * function.c (expand_function_end): Likewise.
16017 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16019 * combine.c (try_combine): Adjust comment. Do not add the set of
16020 insn #0 if the destination indirectly is set or dies in insn #2.
16021 Tidy up code to distribute a new note.
16023 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16025 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16026 also for alternatives 16 and 17.
16028 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16030 * config/i386/sse.md (*mov<mode>_internal): Merge with
16031 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16032 Emit insn template depending on type attribute. Use
16033 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16034 movd instead of movq mnemonic for interunit moves. Rewrite mode
16035 attribute calculation. Remove unit attribute calculation.
16036 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16037 Set prefix_data16 attribute for DImode ssemov types.
16038 Use Ym instead of y for SSE-MMX conversion alternatives.
16039 Reorder operand constraints.
16041 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16043 * df.h (df_insn_delete): Adjust prototype.
16044 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16045 and let it decide whether mark the basic block dirty.
16046 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16047 * df-scan.c (df_insn_info_delete): New helper function, split
16048 off from df_insn_delete.
16049 (df_scan_free_bb_info): Use it.
16050 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16052 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16053 that the insn is actually an insn and it has a non-NULL basic block.
16054 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16056 2013-03-22 Richard Biener <rguenther@suse.de>
16058 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16060 (mem_ref_alloc): Do not allocate them.
16061 (refs_independent_p): Do not query or maintain a cache.
16063 2013-03-22 Richard Biener <rguenther@suse.de>
16065 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16066 (gather_mem_refs_in_loops): Do not compute it.
16067 (analyze_memory_references): Do not allocate it.
16068 (tree_ssa_lim_finalize): Do not free it.
16069 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16071 2013-03-22 Richard Biener <rguenther@suse.de>
16073 * is-a.h (as_a): Use gcc_checking_assert.
16075 2013-03-22 Ian Bolton <ian.bolton@arm.com>
16077 * config/aarch64/aarch64.c (aarch64_print_operand): New
16078 format specifier for printing a constant in hex.
16079 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16080 format specifier for printing second operand.
16082 2013-03-22 Richard Biener <rguenther@suse.de>
16084 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16086 (gather_mem_refs_in_loops): Perform store accumulation here.
16087 (create_vop_ref_mapping_loop): Remove.
16088 (create_vop_ref_mapping): Likewise.
16089 (analyze_memory_references): Initialize refs_stored_in_loop.
16090 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
16091 (record_indep_loop): Remove.
16092 (record_dep_loop): New function.
16093 (ref_indep_loop_p_1): Adjust to only walk over references
16094 in the loop, not its subloops.
16095 (ref_indep_loop_p): Rename to ...
16096 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16097 maintaining a more fine-grained cache.
16098 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16099 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16101 2013-03-22 Richard Biener <rguenther@suse.de>
16103 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16104 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16105 aggregate mem_ref_loc.
16106 (free_mem_ref_locs): Inline into ...
16107 (memref_free): ... this and adjust.
16108 (mem_ref_alloc): Adjust.
16109 (mem_ref_locs_alloc): Remove.
16110 (record_mem_ref_loc): Adjust.
16111 (get_all_locs_in_loop): Rewrite into ...
16112 (for_all_locs_in_loop): ... this iterator.
16113 (rewrite_mem_ref_loc): New functor.
16114 (rewrite_mem_refs): Use for_all_locs_in_loop.
16115 (sm_set_flag_if_changed): New functor.
16116 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16117 (ref_always_accessed): New functor.
16118 (ref_always_accessed_p): Use for_all_locs_in_loop.
16120 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16122 * tree-pass.h (PROP_gimple_lvec): New.
16123 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16124 (init_optimization_passes): Move pass_lower_vector.
16125 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16127 (pass_lower_vector): Provide PROP_gimple_lvec.
16128 (pass_lower_vector_ssa): Likewise.
16129 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16131 2013-03-21 Mark Wielaard <mjw@redhat.com>
16133 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16135 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16137 * config/i386/i386.md (*movdi_internal): Disparage slightly
16138 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16139 conversion alternatives.
16141 2013-03-21 Jakub Jelinek <jakub@redhat.com>
16143 PR middle-end/48087
16144 * diagnostic.def (DK_WERROR): New kind.
16145 * diagnostic.h (werrorcount): Define.
16146 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16147 promoted to DK_ERROR, increment DK_WERROR counter instead of
16149 * toplev.c (toplev_main): Call print_ignored_options even if
16150 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16151 even if just werrorcount is non-zero.
16154 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16156 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16157 (string_cst_pool_decl): New function.
16158 (optimize_one_addr_into_implicit_ptr): New function.
16159 (resolve_addr_in_expr): Optimize DWARF location expression
16160 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16161 which doesn't live in memory, but has DW_AT_location or
16162 DW_AT_const_value, or refers to a string literal, into
16163 DW_OP_GNU_implicit_pointer.
16164 (optimize_location_into_implicit_ptr): New function.
16165 (resolve_addr): If removing DW_AT_location of a variable because
16166 it was DW_OP_addr of address of the variable, but the variable doesn't
16167 live in memory, try to emit const value attribute for the initializer.
16169 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16171 * tree.h (VECTOR_TYPE_P): New macro.
16172 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16173 TYPE_MODE): Use it.
16174 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16175 VEC_COND_EXPR cannot be lvalues.
16176 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16178 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16180 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16181 Restrict the transformation to equal modes.
16183 2013-03-21 Richard Biener <rguenther@suse.de>
16185 PR tree-optimization/39326
16186 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16187 (MEM_ANALYZABLE): Adjust.
16188 (record_mem_ref_loc): Move bitmap ops ...
16189 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16190 unanalyzable refs, do not record locations for it.
16191 (analyze_memory_references): Allocate ref zero as shared
16193 (refs_independent_p): Do not test for unanalyzed mems here.
16194 (ref_indep_loop_p_1): Special-case disambiguation against
16195 the unanalyzed ref.
16196 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
16198 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16200 * config/arm/arm-protos.h (tune_params): Add
16201 prefer_neon_for_64bits field.
16202 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16203 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16204 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16205 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16206 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16207 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16208 (arm_option_override): Handle -mneon-for-64bits new option.
16209 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16210 (prefer_neon_for_64bits): Declare new variable.
16211 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
16212 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
16213 (arch_enabled): Handle new arch types. Remove support for onlya8
16215 (one_cmpldi2): Use new arch names.
16216 (zero_extend<mode>di2, extend<mode>di2): Ditto.
16217 * config/arm/arm.opt (mneon-for-64bits): Add option.
16218 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16219 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16220 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16222 * doc/invoke.texi (-mneon-for-64bits): Document.
16224 2013-03-21 Richard Biener <rguenther@suse.de>
16226 PR tree-optimization/39326
16227 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16228 (sort_bbs_in_loop_postorder_cmp): New function.
16229 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
16231 2013-03-21 Richard Biener <rguenther@suse.de>
16233 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16234 (vect_insert_into_interleaving_chain): Likewise.
16235 (vect_drs_dependent_in_basic_block): Inline ...
16236 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16238 (vect_analyze_data_ref_dependence): ... here. Simplify.
16239 (vect_check_interleaving): Simplify.
16240 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16241 (vect_slp_analyze_data_ref_dependences): ... this new function.
16242 (dr_group_sort_cmp): New function.
16243 (vect_analyze_data_ref_accesses): Compute data-reference groups
16244 here instead of in vect_analyze_data_ref_dependence. Use
16245 a more efficient algorithm.
16246 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16247 vect_slp_analyze_data_ref_dependences. Call
16248 vect_analyze_data_ref_accesses earlier.
16249 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16250 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16251 (vect_slp_analyze_data_ref_dependences): New prototype.
16253 2013-03-21 Richard Biener <rguenther@suse.de>
16255 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16256 ref is stored in the loop.
16257 (find_refs_for_sm): Walk only over all stores.
16258 (store_motion_loop): Allocate from lim_bitmap_obstack.
16259 (store_motion): Likewise.
16261 2013-03-21 Richard Biener <rguenther@suse.de>
16263 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16264 Update virtual SSA form.
16266 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16268 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16269 * configure: Regenerate.
16270 * config.in: Regenerate.
16271 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16272 if !HAVE_LD_EH_FRAME_CIEV3.
16274 2013-03-21 Richard Biener <rguenther@suse.de>
16276 * tree-cfg.c (verify_expr_no_block): New function.
16277 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16278 nor DECL_VALUE_EXPR have locations with associated blocks.
16279 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16280 (clear_unused_block_pointer): Remove code dealing with
16281 blocks in DECL_DEBUG_EXPR locations.
16283 2013-03-21 Richard Biener <rguenther@suse.de>
16285 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16286 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16287 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16288 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16289 instead of DECL_DEBUG_EXPR_IS_FROM.
16290 * gimplify.c (gimplify_modify_expr): Likewise.
16291 * tree-cfg.c (verify_expr_location_1): Likewise.
16292 * tree-complex.c (create_one_component_var): Likewise.
16293 * tree-sra.c (create_access_replacement): Likewise.
16294 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16295 (clear_unused_block_pointer): Likewise.
16296 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16297 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16298 * var-tracking.c (var_debug_decl): Likewise.
16299 (track_expr_p): Likewise.
16300 * tree-inline.c (add_local_variables): Likewise. Set
16301 DECL_HAS_DEBUG_EXPR_P after copying it.
16302 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16303 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16305 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16308 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16309 * configure: Regenerate.
16310 * config.in: Regenerate.
16311 * config/i386/i386.md (*movdf_internal): Use
16312 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16313 movd instead of movq mnemonic for interunit moves.
16314 (*movdi_internal): Ditto.
16316 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16318 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16319 (abd<mode>_3): New pattern.
16320 (aba<mode>_3): New pattern.
16321 (fabd<mode>_3): New pattern.
16323 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16325 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16326 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16327 occurrence of REGISTER_PREFIX as its empty string.
16329 2013-03-20 Jeff Law <law@redhat.com>
16331 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16332 addititional equivalences for equality comparisons between an SSA_NAME
16333 and a constant where the SSA_NAME was set from a widening conversion.
16335 2013-03-20 Walter Lee <walt@tilera.com>
16337 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16339 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16341 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16342 depending on type attribute.
16343 (*movti_internal): Ditto.
16344 (*movtf_internal): Ditto.
16345 (*movxf_internal): Ditto.
16346 (*movdf_internal): Ditto.
16347 (*movsf_internal): Ditto.
16349 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16351 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16352 maybe_vex for sselog1 and ssemov types.
16353 (*movdi_internal): Reorder operand constraints.
16354 (*movsi_internal): Ditto. Set prefix attribute to
16355 maybe_vex for sselog1 and ssemov types.
16356 (*movtf_internal): Set prefix attribute to maybe_vex
16357 for sselog1 and ssemov types.
16358 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16359 DImode ssemov types. Reorder operand constraints.
16360 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16361 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16362 attribute for SImode ssemov types. Reorder operand constraints.
16364 2013-03-20 Martin Jambor <mjambor@suse.cz>
16366 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
16367 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
16369 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
16371 * config/rs6000/predicates.md (indexed_address, update_address_mem
16372 update_indexed_address_mem): New predicates.
16373 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
16374 attribute for load/store instructions.
16375 * config/rs6000/dfp.md (movsd_store): Likewise.
16376 (movsd_load): Likewise.
16377 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
16378 (unnamed HI->DI extend define_insn): Likewise.
16379 (unnamed SI->DI extend define_insn): Likewise.
16380 (unnamed QI->SI extend define_insn): Likewise.
16381 (unnamed QI->HI extend define_insn): Likewise.
16382 (unnamed HI->SI extend define_insn): Likewise.
16383 (unnamed HI->SI extend define_insn): Likewise.
16384 (extendsfdf2_fpr): Likewise.
16385 (movsi_internal1): Likewise.
16386 (movsi_internal1_single): Likewise.
16387 (movhi_internal): Likewise.
16388 (movqi_internal): Likewise.
16389 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
16390 attribute for load/store instructions.
16391 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
16393 (mov<mode>_softfloat): Likewise.
16394 (mov<mode>_hardfloat32): Likewise.
16395 (mov<mode>_hardfloat64): Likewise.
16396 (mov<mode>_softfloat64): Likewise.
16397 (movdi_internal32): Likewise.
16398 (movdi_internal64): Likewise.
16399 (probe_stack_<mode>): Likewise.
16401 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
16403 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
16404 floating point, and decimal floating point to reload iterator.
16406 * config/rs6000/constraints.md (wl constraint): New constraints to
16407 return FLOAT_REGS if certain options are used to reduce the number
16408 of separate patterns that exist in the file.
16409 (wx constraint): Likewise.
16410 (wz constraint): Likewise.
16412 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16413 -mdebug=reg, print wg, wl, wx, and wz constraints.
16414 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16415 Initialize the reload functions for 64-bit binary/decimal floating
16417 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16418 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16419 create the buffer on the stack to overcome not having a 32-bit
16421 (rs6000_emit_move): Likewise.
16422 (rs6000_secondary_memory_needed_rtx): Likewise.
16423 (rs6000_alloc_sdmode_stack_slot): Likewise.
16424 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16425 via xxlxor, just like DFmode 0.0.
16427 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
16428 define as 1 if we are running on a power7 or newer.
16429 (enum r6000_reg_class_enum): Add new constraints.
16431 * config/rs6000/dfp.md (movsd): Delete, combine with binary
16432 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16433 with other moves by using conditional constraits (wg). Use LFIWZX
16434 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
16435 (movsd splitter): Likewise.
16436 (movsd_hardfloat): Likewise.
16437 (movsd_softfloat): Likewise.
16439 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
16440 binary and decimal floating point moves.
16441 (fmove_ok): New attributes to combine binary and decimal floating
16442 point moves, and to combine power6x (mfpgpr) moves along normal
16444 (real_value_to_target): Likewise.
16445 (f32_lr): Likewise.
16446 (f32_lm): Likewise.
16447 (f32_li): Likewise.
16448 (f32_sr): Likewise.
16449 (f32_sm): Likewise.
16450 (f32_si): Likewise.
16451 (movsf): Combine binary and decimal floating point moves. Combine
16452 power6x (mfpgpr) moves with other moves by using conditional
16453 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
16454 (mov<mode> for SFmode/SDmode); Likewise.
16455 (SFmode/SDmode splitters): Likewise.
16456 (movsf_hardfloat): Likewise.
16457 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
16458 (movsf_softfloat): Likewise.
16459 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
16461 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
16462 wx and wz constraints.
16464 * config/rs6000/constraints.md (wg constraint): New constraint to
16465 return FLOAT_REGS if -mmfpgpr (power6x) was used.
16467 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
16470 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16471 -mdebug=reg, print wg, wl, wx, and wz constraints.
16472 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16473 Initialize the reload functions for 64-bit binary/decimal floating
16475 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16476 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16477 create the buffer on the stack to overcome not having a 32-bit
16479 (rs6000_emit_move): Likewise.
16480 (rs6000_secondary_memory_needed_rtx): Likewise.
16481 (rs6000_alloc_sdmode_stack_slot): Likewise.
16482 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16483 via xxlxor, just like DFmode 0.0.
16485 * config/rs6000/dfp.md (movdd): Delete, combine with binary
16486 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16487 with other moves by using conditional constraits (wg). Use LFIWZX
16488 and STFIWX for loading SDmode on power7.
16489 (movdd splitters): Likewise.
16490 (movdd_hardfloat32): Likewise.
16491 (movdd_softfloat32): Likewise.
16492 (movdd_hardfloat64_mfpgpr): Likewise.
16493 (movdd_hardfloat64): Likewise.
16494 (movdd_softfloat64): Likewise.
16496 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
16497 64-bit binary and decimal floating point moves.
16498 (FMOVE64X): Likewise.
16499 (movdf): Combine 64-bit binary and decimal floating point moves.
16500 Combine power6x (mfpgpr) moves with other moves by using
16501 conditional constraits (wg).
16502 (mov<mode> for DFmode/DDmode): Likewise.
16503 (DFmode/DDmode splitters): Likewise.
16504 (movdf_hardfloat32): Likewise.
16505 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
16506 (movdf_softfloat32): Likewise.
16507 (movdf_hardfloat64_mfpgpr): Likewise.
16508 (movdf_hardfloat64): Likewise.
16509 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
16510 (movdf_softfloat64): Likewise.
16511 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
16512 (reload_<mode>_load): Move to later in the file so they aren't in
16513 the middle of the floating point move insns.
16514 (reload_<mode>_store): Likewise.
16516 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
16519 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
16520 constraint if -mdebug=reg.
16521 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
16522 Enable using dd reload support if needed.
16524 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
16525 binary and decimal floating point moves in rs6000.md.
16526 (movtd_internal): Likewise.
16528 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
16529 decimal floating point moves.
16531 (movtf_internal): Likewise.
16532 (mov<mode>_internal, TDmode/TFmode): Likewise.
16533 (movtf_softfloat): Likewise.
16534 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
16536 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
16537 movdi_internal64, using wg constraint for move direct operations.
16538 (movdi_internal64): Likewise.
16540 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
16541 MODES_TIEABLE_P for selected modes. Print the numerical value of
16542 the various virtual registers. Use GPR/FPR first/last values,
16543 instead of hard coding the register numbers. Print which modes
16544 have reload functions registered.
16545 (rs6000_option_override_internal): If -mdebug=reg, trace the options
16546 settings before/after setting cpu, target and subtarget settings.
16547 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
16548 and for secondary reload failures in rs6000_secondary_reload_inner.
16549 (rs6000_secondary_reload_fail): Likewise.
16550 (rs6000_secondary_reload_inner): Likewise.
16552 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
16553 macros for first/last GPR and FPR registers.
16554 (LAST_GPR_REGNO): Likewise.
16555 (FIRST_FPR_REGNO): Likewise.
16556 (LAST_FPR_REGNO): Likewise.
16558 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
16559 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
16560 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
16561 (vcond<mode><mode>): Likewise.
16562 (vcondu<mode><mode>): Likewise.
16563 (vector_gtu<mode>): Likewise.
16564 (vector_gte<mode>): Likewise.
16565 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
16566 to prevent the compiler from converting DImode operations to TImode.
16567 (ior<mode>3): Likewise.
16568 (and<mode>3): Likewise.
16569 (one_cmpl<mode>2): Likewise.
16570 (nor<mode>3): Likewise.
16571 (andc<mode>3): Likewise.
16573 * config/rs6000/constraints.md (wt constraint): New constraint
16574 that returns VSX_REGS if TImode is allowed in VSX registers.
16576 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
16577 constant under VSX.
16579 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
16580 similar to TImode, but it is restricted to being in the GPRs.
16582 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
16583 TImode to occupy a single VSX register.
16585 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
16586 -mvsx-timode for power7/power8.
16587 (power7 cpu): Likewise.
16588 (power8 cpu): Likewise.
16590 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
16591 sure that TFmode/TDmode take up two registers if they are ever
16592 allowed in the upper VSX registers.
16593 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
16595 (rs6000_init_hard_regno_mode_ok): Likewise.
16596 (rs6000_debug_reg_global): Add debugging for PTImode and wt
16597 constraint. Print if LRA is turned on.
16598 (rs6000_option_override_internal): Give an error if -mvsx-timode
16599 and VSX is not enabled.
16600 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
16601 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
16602 to reg+offset addressing. Use PTImode when checking offset
16603 addresses for validity.
16604 (reg_offset_addressing_ok_p): Likewise.
16605 (rs6000_legitimate_offset_address_p): Likewise.
16606 (rs6000_legitimize_address): Likewise.
16607 (rs6000_legitimize_reload_address): Likewise.
16608 (rs6000_legitimate_address_p): Likewise.
16609 (rs6000_eliminate_indexed_memrefs): Likewise.
16610 (rs6000_emit_move): Likewise.
16611 (rs6000_secondary_reload): Likewise.
16612 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
16613 reloads to fpr registers to continue to use reg+offset addressing,
16614 but 64-bit reloads to altivec registers need reg+reg addressing.
16615 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
16616 it. Treat LO_SUM like a PLUS operation.
16617 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
16618 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
16619 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
16620 registers to share a register with a smaller sized type, since VSX
16621 puts scalars in the upper 64-bits.
16622 (print_operand): Add support for PTImode.
16623 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
16624 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
16625 registers, but don't have arithmetic support.
16626 (rs6000_memory_move_cost): Add test for VSX.
16627 (rs6000_opt_masks): Add -mvsx-timode.
16629 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
16632 (VSr): Use wt constraint for TImode.
16633 (VSv): Drop TImode support.
16634 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
16635 (vsx_movti_64bit): Likewise.
16636 (vsx_movti_32bit): Likewise.
16637 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
16638 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
16639 one '?' on the appropriate output constraint. Do not allow TImode
16640 logical operations on 32-bit systems.
16641 (vsx_ior<mode>3): Likewise.
16642 (vsx_xor<mode>3): Likewise.
16643 (vsx_one_cmpl<mode>2): Likewise.
16644 (vsx_nor<mode>3): Likewise.
16645 (vsx_andc<mode>3): Likewise.
16646 (vsx_concat_<mode>): Likewise.
16647 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
16649 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
16650 OPTION_MASK_VSX_TIMODE.
16651 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
16652 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
16654 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
16655 (TI2 iterator): New iterator for TImode, PTImode.
16656 (wd mode attribute): Add values for vector types.
16657 (movti_string): Replace TI move operations with operations for TImode
16658 and PTImode. Add support for TImode being allowed in VSX registers.
16659 (mov<mode>_string, TImode/PTImode): Likewise.
16660 (movti_ppc64): Likewise.
16661 (mov<mode>_ppc64, TImode/PTImode): Likewise.
16662 (TI mode splitters): Likewise.
16664 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
16667 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
16669 PR tree-optimization/56355
16670 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
16671 Also handle integers with undefined overflow.
16673 2013-03-20 Catherine Moore <clm@codesourcery.com>
16674 Maciej W. Rozycki <macro@codesourcery.com>
16675 Tom de Vries <tom@codesourcery.com>
16676 Nathan Sidwell <nathan@codesourcery.com>
16677 Iain Sandoe <iain@codesourcery.com>
16678 Nathan Froyd <froydnj@codesourcery.com>
16679 Chao-ying Fu <fu@mips.com>
16681 * doc/extend.texi: (micromips, nomicromips, nocompression):
16682 Document new function attributes.
16683 * doc/invoke.texi (minterlink-compressed, mmicromips,
16684 m14k, m14ke, m14kec): Document new options.
16685 (minterlink-mips16): Update documentation.
16686 * doc/md.texi (ZC, ZD): Document new constraints.
16687 * configure.ac (gcc_cv_as_micromips): Check if linker
16688 supports the .set micromips directive.
16689 * configure: Regenerate.
16690 * config.in: Regenerate.
16691 * config/mips/mips-tables.opt: Regenerate.
16692 * config/mips/micromips.md: New file.
16693 * constraints.md (ZC, ZD): New constraints.
16694 * config/mips/predicates.md (movep_src_register): New predicate.
16695 (movep_src_operand): New predicate.
16696 (non_volatile_mem_operand): New predicate.
16697 * config/mips/mips.md (multimem): New type.
16698 (length): Differentiate between 17-bit and 18-bit branch offsets.
16699 (MOVEP1, MOVEP2): New mode iterator.
16700 (mov_<load>l): Use ZC constraint.
16701 (mov_<load>r): Likewise.
16702 (mov_<store>l): Likewise.
16703 (mov_<store>r): Likewise.
16704 (*branch_equality<mode>_inverted): Add microMIPS support.
16705 (*branch_equality<mode>): Likewise.
16706 (*jump_absolute): Likewise.
16707 (indirect_jump_<mode>): Likewise.
16708 (tablejump_<mode>): Likewise.
16709 (<optab>_internal): Likewise.
16710 (sibcall_internal): Likewise.
16711 (sibcall_value_internal): Likewise.
16712 (prefetch): Use constraint ZD.
16713 * config/mips/mips.opt (minterlink-compressed): New option.
16714 (minterlink-mips16): Now an alias for minterlink-compressed.
16715 (mmicromips): New option.
16716 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16717 (compare_and_swap_12): Likewise.
16718 (sync_add<mode>): Likewise.
16719 (sync_<optab>_12): Likewise.
16720 (sync_old_<optab>_12): Likewise.
16721 (sync_new_<optab>_12): Likewise.
16722 (sync_nand_12): Likewise.
16723 (sync_old_nand_12): Likewise.
16724 (sync_new_nand_12): Likewise.
16725 (sync_sub<mode>): Likewise.
16726 (sync_old_add<mode>): Likewise.
16727 (sync_old_sub<mode>): Likewise.
16728 (sync_new_add<mode>): Likewise.
16729 (sync_new_sub<mode>): Likewise.
16730 (sync_<optab><mode>): Likewise.
16731 (sync_old_<optab><mode>): Likewise.
16732 (sync_new_<optab><mode>): Likewise.
16733 (sync_nand<mode>): Likewise.
16734 (sync_old_nand<mode>): Likewise.
16735 (sync_new_nand<mode>): Likewise.
16736 (sync_lock_test_and_set<mode>): Likewise.
16737 (test_and_set_12): Likewise.
16738 (atomic_compare_and_swap<mode>): Likewise.
16739 (atomic_exchange<mode>_llsc): Likewise.
16740 (atomic_fetch_add<mode>_llsc): Likewise.
16741 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16742 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16743 (umips_save_restore_pattern_p): Likewise.
16744 (umips_load_store_pair_p): Likewise.
16745 (umips_output_load_store_pair): Likewise.
16746 (umips_movep_target_p): Likewise.
16747 (umips_12bit_offset_address_p): Likewise.
16748 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16749 (mips_base_mips16): Rename this...
16750 (mips_base_compression_flags): ...to this. Update all uses.
16751 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16752 (mips_mips16_decl_p): Delete.
16753 (mips_nomips16_decl_p): Delete.
16754 (mips_get_compress_on_flags): New function.
16755 (mips_get_compress_off_flags): New function.
16756 (mips_get_compress_mode): New function.
16757 (mips_get_compress_on_name): New function.
16758 (mips_get_compress_off_name): New function.
16759 (mips_insert_attributes): Support multiple compression types.
16760 (mips_merge_decl_attributes): Likewise.
16761 (umips_12bit_offset_address_p): New function.
16762 (mips_start_function_definition): Emit .set micromips directive.
16763 (mips_call_may_need_jalx_p): New function.
16764 (mips_function_ok_for_sibcall): Add microMIPS support.
16765 (mips_print_operand_punctuation): Support short delay slots and
16767 (umips_swm_mask, umips_swm_encoding): New.
16768 (umips_build_save_restore): New function.
16769 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16770 (was_mips16_p): Remove.
16771 (old_compression_mode): New.
16772 (mips_set_compression_mode): New function.
16773 (mips_set_current_function): Add microMIPS support.
16774 (mips_option_override): Likewise.
16775 (umips_save_restore_pattern_p): New function.
16776 (umips_output_save_restore): New function.
16777 (umips_load_store_pair_p_1): New function.
16778 (umips_load_store_pair_p): New function.
16779 (umips_output_load_store_pair_1): New function.
16780 (umips_output_load_store_pair): New function.
16781 (umips_movep_target_p) New function.
16782 (mips_prepare_pch_save): Add microMIPS support.
16783 * config/mips/mips.h (TARGET_COMPRESSION): New.
16784 (TARGET_CPU_CPP_BUILTINS): Update macro
16785 to use new compression flags and to support microMIPS.
16786 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16787 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16788 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16789 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16790 (ASM_SPEC): Support mmicromips and mno-micromips.
16791 (M16STORE_REG_P): New macro.
16792 (MIPS_CALL): Support TARGET_MICROMIPS.
16793 (MICROMIPS_J): New macro.
16794 (mips_base_mips16): Rename this...
16795 (mips_base_compression_flags): ...to this.
16796 (UMIPS_12BIT_OFFSET_P): New macro.
16797 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16798 (MULTILIB_DIRNAMES): Likewise.
16799 2013-03-20 Richard Biener <rguenther@suse.de>
16801 PR tree-optimization/56661
16802 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
16803 the result does not have to be distinct.
16805 2013-03-20 Richard Biener <rguenther@suse.de>
16807 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
16810 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16811 Steven Bosscher <steven@gcc.gnu.org>
16813 PR rtl-optimization/56605
16814 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
16816 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16819 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
16820 that require movd instead of movq.
16822 2013-03-20 Richard Biener <rguenther@suse.de>
16824 * tree-ssa-structalias.c (struct variable_info): Add pointer
16825 to the first field of an aggregate with sub-vars. Make
16826 this and the pointer to the next subfield its ID.
16827 (vi_next): New function.
16828 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
16829 storedanything_id, integer_id): Increment by one.
16830 (new_var_info, get_call_vi, lookup_call_clobber_vi,
16831 get_call_clobber_vi): Adjust.
16832 (solution_set_expand): Simplify and speedup.
16833 (solution_set_add): Inline into ...
16834 (set_union_with_increment): ... this. Adjust accordingly.
16835 (do_sd_constraint): Likewise.
16836 (do_ds_constraint): Likewise.
16837 (do_complex_constraint): Simplify.
16838 (build_pred_graph): Adjust.
16839 (solve_graph): Likewise. Simplify and speedup.
16840 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
16841 get_constraint_for_component_ref, get_constraint_for_1,
16842 first_vi_for_offset, first_or_preceding_vi_for_offset,
16843 create_function_info_for, create_variable_info_for_1,
16844 create_variable_info_for, intra_create_variable_infos): Adjust.
16845 (init_base_vars): Push NULL for ID zero.
16846 (compute_points_to_sets): Adjust.
16848 2013-03-20 Richard Biener <rguenther@suse.de>
16850 * cfgloop.c (verify_loop_structure): Streamline and avoid
16851 ICEing on corrupt loop tree.
16852 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
16855 2013-03-20 Richard Biener <rguenther@suse.de>
16857 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
16858 check whether an SSA update is needed.
16860 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
16862 * config/mips/constraints.md (T): Rename to...
16866 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
16867 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
16869 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16871 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
16872 (*subsi3_carryin_uxtw): Likewise.
16874 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16876 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
16877 (*rorsi3_insn_uxtw): Likewise.
16879 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16881 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
16882 (*extrsi5_insn_uxtw): Likewise.
16884 2013-03-19 Richard Biener <rguenther@suse.de>
16886 PR tree-optimization/56273
16887 * passes.c (init_optimization_passes): Move second VRP after DOM.
16889 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
16891 * config/i386/i386.md (*movti_internal): Merge from
16892 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
16893 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
16894 nox64 isa attributes.
16896 2013-03-18 Richard Biener <rguenther@suse.de>
16898 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
16900 (merge_node_constraints): Likewise.
16901 (build_succ_graph): Likewise.
16902 (valid_graph_edge): Inline into single caller.
16903 (unify_nodes): Likewise. Use bitmap_set_bit return value
16905 (scc_visit): Fix formatting and variable use.
16906 (do_sd_constraint): Use gcc_checking_assert.
16907 (do_ds_constraint): Likewise.
16908 (do_complex_constraint): Likewise.
16909 (condense_visit): Likewise. Cleanup.
16910 (dump_pred_graph): New function.
16911 (perform_var_substitution): Dump the pred-graph before
16912 variable substitution.
16913 (find_equivalent_node): Use gcc_checking_assert.
16914 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
16916 2013-03-18 Richard Biener <rguenther@suse.de>
16918 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
16919 Remove cond_expr_stmt_list argument and do not gimplify the
16921 (vect_loop_versioning): Adjust.
16922 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
16923 Cleanup to use less temporaries.
16924 (vect_create_data_ref_ptr): Cleanup.
16926 2013-03-18 Jakub Jelinek <jakub@redhat.com>
16928 PR tree-optimization/56635
16929 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
16930 require types_compatible_p types.
16932 2013-03-18 Nick Clifton <nickc@redhat.com>
16934 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
16935 spurious backslash.
16937 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
16938 Add missing line to comment describing function.
16940 2013-03-18 Richard Biener <rguenther@suse.de>
16942 PR tree-optimization/56210
16943 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
16944 Handle string / character search functions.
16945 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
16947 2013-03-18 Richard Biener <rguenther@suse.de>
16949 PR middle-end/56483
16950 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
16951 and implement properly.
16952 * gimple.h (gimple_cond_single_var_p): Remove.
16954 2013-03-18 Richard Biener <rguenther@suse.de>
16956 * tree-data-ref.h (find_data_references_in_loop): Declare.
16957 * tree-data-ref.c (get_references_in_stmt): Use a stack
16958 vector pre-allocated in the callers.
16959 (find_data_references_in_stmt): Adjust.
16960 (graphite_find_data_references_in_stmt): Likewise.
16961 (create_rdg_vertices): Likewise.
16962 (find_data_references_in_loop): Export.
16963 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
16964 Compute dependences here...
16965 (vect_analyze_data_refs): ...not here. When we encounter
16966 a non-vectorizable data reference in basic-block vectorization
16967 truncate the data reference vector. Do not bother to
16968 fixup data-dependence information for gather loads.
16969 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
16970 of data references, as reported.
16972 2013-03-18 Richard Biener <rguenther@suse.de>
16974 PR tree-optimization/3713
16975 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
16976 has_constants and expr.
16977 (stmt_has_constants): Properly valueize SSA names when deciding
16978 whether the stmt has constants.
16980 2013-03-18 Richard Biener <rguenther@suse.de>
16982 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
16983 whole function when there is nothing to do.
16984 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
16985 * tree-vectorizer.c (vectorize_loops): Update virtual and
16986 loop-closed SSA once.
16987 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
16989 2013-03-18 Richard Biener <rguenther@suse.de>
16991 PR middle-end/56113
16992 * domwalk.c (bb_postorder): New global static.
16993 (cmp_bb_postorder): New function.
16994 (walk_dominator_tree): Replace scheme imposing an order for
16995 visiting dominator sons by one sorting them at the time they
16996 are pushed on the stack.
16998 2013-03-18 Richard Biener <rguenther@suse.de>
17000 PR tree-optimization/39326
17001 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17002 (struct mem_ref): Replace mem member with ao_ref typed member.
17003 (MEM_ANALYZABLE): Adjust.
17004 (memref_eq): Likewise.
17005 (mem_ref_alloc): Likewise.
17006 (gather_mem_refs_stmt): Likewise.
17007 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17008 (execute_sm_if_changed_flag_set): Adjust.
17009 (execute_sm): Likewise.
17010 (ref_always_accessed_p): Likewise.
17011 (refs_independent_p): Likewise.
17012 (can_sm_ref_p): Likewise.
17014 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17017 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17018 return 1 even for !unsignedp.
17020 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17022 * config/i386/i386.md (isa): Add x64 and nox64.
17023 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17024 (*pushtf): Enable *roF alternative for x64 isa only.
17025 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17026 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17027 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17028 (*movtf_internal): Merge from *movtf_internal_rex64 and
17029 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17030 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17031 nox64 isa attributes.
17032 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17033 nox64 isa attributes.
17034 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17036 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17038 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17040 2013-03-17 Jakub Jelinek <jakub@redhat.com>
17043 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17044 class names. Remove trailing comma after "ALL_REGS".
17046 2013-03-16 Jan Hubicka <jh@suse.cz>
17048 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17049 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17050 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17051 of cgraph_get_create_node.
17052 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17054 2013-03-16 Jason Merrill <jason@redhat.com>
17057 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17058 with DW_AT_default_value.
17060 2013-03-16 Jakub Jelinek <jakub@redhat.com>
17062 * BASE-VER: Set to 4.9.0.
17064 2013-03-14 Andi Kleen <ak@linux.intel.com>
17067 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17068 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17069 Document _x* TSX intrinsics.
17071 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17072 David Holsgrove <david.holsgrove@xilinx.com>
17074 * configure.ac: Add MicroBlaze TLS support detection.
17075 * configure: Regenerate.
17076 * config/microblaze/microblaze-protos.h
17077 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17078 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17079 * config/microblaze/microblaze.c (microblaze_address_type): Add
17080 ADDRESS_TLS and tls_reloc address types.
17081 (microblaze_address_info): Add tls_reloc.
17082 (TARGET_HAVE_TLS): Define.
17083 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17084 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
17085 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17086 load_tls_operand, microblaze_call_tls_get_addr,
17087 microblaze_legitimize_tls_address): New functions.
17088 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17089 (get_base_reg): Use microblaze_tls_symbol_p.
17090 (microblaze_classify_address): Handle TLS.
17091 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17092 label_mentioned_p and microblaze_tls_referenced_p.
17093 (microblaze_legitimize_address): Handle TLS.
17094 (microblaze_address_insns): Handle ADDRESS_TLS.
17095 (pic_address_needs_scratch): Handle TLS.
17096 (print_operand_address): Handle TLS.
17097 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17098 (microblaze_expand_move): Handle TLS.
17099 (microblaze_legitimate_constant_p): Check
17100 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
17101 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17102 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17103 (PIC_OFFSET_TABLE_REGNUM): Set.
17104 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17105 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17106 (addsi3, movsi_internal2, movdf_internal): Update constraints
17107 * config/microblaze/predicates.md (arith_plus_operand): Define
17108 (move_operand): Redefine as move_src_operand,
17109 check microblaze_tls_referenced_p.
17111 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17113 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17114 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17116 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17118 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17121 2013-03-14 Jakub Jelinek <jakub@redhat.com>
17123 PR tree-optimization/53265
17124 * common.opt (Waggressive-loop-optimizations): New option.
17125 * tree-ssa-loop-niter.c: Include tree-pass.h.
17126 (do_warn_aggressive_loop_optimizations): New function.
17127 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17128 if number_of_latch_executions returned constant.
17129 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17130 early. If number_of_latch_executions returned constant, set
17131 nb_iterations_upper_bound back to it.
17132 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17134 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17135 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17137 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17138 (MULTILIB_OSDIRNAMES): Set.
17139 * genmultilib: If defaultosdirname doesn't start with :: , set
17140 defaultosdirname2 instead, clear it and emit two . multilib_raw
17141 entries instead of just one.
17143 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17145 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17146 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17147 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17148 (SUBTARGET_OVERRIDE_OPTIONS): New.
17150 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17153 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17154 (musermode): Convert to Var(TARGET_USERMODE).
17155 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17156 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17157 * config/sh/sh.c (sh_option_override): Use
17158 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17159 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17161 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17163 (udivsi3_i4_single, divsi3_i4_single): Use
17164 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17166 2013-03-13 Dave Korn <dave.korn.cygwin@....>
17168 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17171 2013-03-13 Richard Biener <rguenther@suse.de>
17173 PR tree-optimization/56608
17174 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17175 calls when vectorizing basic-blocks.
17177 2013-03-13 Jakub Jelinek <jakub@redhat.com>
17180 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17183 2013-03-12 Jakub Jelinek <jakub@redhat.com>
17185 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17187 2013-03-11 Jan Hubicka <jh@suse.cz>
17190 * lto-streamer-out.c (output_symbol_p): Skip references from
17191 constructors of external variables.
17193 2013-03-11 Jan Hubicka <jh@suse.cz>
17195 PR middle-end/56571
17196 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17198 * emit-rtl.c (verify_rtx_sharing): Likewise.
17199 (copy_insn_1): Likewise.
17200 * rtl.c (copy_rtx): Likewise.
17202 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
17205 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17206 output_operand_lossage message.
17208 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
17211 * arm.c (shift_op): Validate RTL pattern on the fly.
17212 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17215 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17218 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17219 new call patterns that clobber %fr12.
17220 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17221 split and postreload patterns.
17222 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17223 registers %fr12 and %fr12R as call used.
17225 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17227 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17228 (canon_address, record_store, replace_read, check_mem_read_rtx,
17229 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17230 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17231 rest_of_handle_dse): Likewise.
17233 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17235 PR middle-end/56524
17236 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17238 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17239 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17240 (save_optabs_if_changed): Replace with...
17241 (init_tree_optimization_optabs): ...this.
17242 * optabs.c (save_optabs_if_changed): Rename to...
17243 (init_tree_optimization_optabs): ...this. Take the optimization node
17244 as argument. Do nothing if the base optabs are already correct.
17245 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17246 to recompute optabs.
17247 * function.h (function): Remove optabs field.
17248 * function.c (invoke_set_current_function_hook): Call
17249 init_tree_optimization_optabs. Use the result to initialize
17252 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
17254 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17255 if GTMA_HAS_NO_INSTRUMENTATION.
17256 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17257 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17258 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17259 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17260 GTMA_HAS_NO_INSTRUMENTATION.
17262 2013-03-08 Jakub Jelinek <jakub@redhat.com>
17264 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17267 2013-03-08 Marek Polacek <polacek@redhat.com>
17268 Jakub Jelinek <jakub@redhat.com>
17270 PR tree-optimization/56478
17271 * predict.c (is_comparison_with_loop_invariant_p): Change the
17272 type of loop_step to tree.
17273 (predict_loops): Adjust.
17274 (predict_iv_comparison): Perform the computations on double_ints.
17276 2013-03-08 Richard Biener <rguenther@suse.de>
17278 PR tree-optimization/56570
17279 * tree-cfg.c (verify_expr_location_1): Verify locations for
17281 * tree-sra.c (create_access_replacement): Strip locations
17282 from DECL_DEBUG_EXPRs.
17284 2013-03-08 Richard Biener <rguenther@suse.de>
17286 * tree-inline.c (expand_call_inline): Do not associate
17287 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17288 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17290 2013-03-08 Richard Biener <rguenther@suse.de>
17292 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17293 or block changes with -Og. Fix for location / block encoding
17294 changes and PHI arguments with locations.
17296 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17298 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17300 (struct output_info): Likewise.
17301 (register_overhead): Remove bad gcc_assert.
17302 (bitmap_find_bit): If there is only a single bitmap element, do not
17303 count a miss as a search.
17304 (print_statistics): Update for counter type changes.
17305 (dump_bitmap_statistics): Likewise. Print headers such that they
17306 are properly lined up with the printed counters.
17308 2013-03-07 Jakub Jelinek <jakub@redhat.com>
17310 PR tree-optimization/56559
17311 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17312 check that it has only a single use.
17314 2013-03-07 Richard Biener <rguenther@suse.de>
17316 * doc/invoke.texi (fwhole-program): Discourage use in combination
17319 2013-03-06 Jakub Jelinek <jakub@redhat.com>
17321 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17323 PR tree-optimization/56539
17324 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17325 instead of GSI_CONTINUE_LINKING as last argument to
17326 force_gimple_operand_gsi. Adjust function comment.
17328 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17331 PR middle-end/56548
17332 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17333 promoted mode, convert the result back to the original mode.
17335 2013-03-06 Richard Biener <rguenther@suse.de>
17337 PR middle-end/56294
17338 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17339 (insert_updated_phi_nodes_compare_uids): New function.
17340 (update_ssa): Sort symbols_to_rename after UID before
17341 traversing it to insert PHI nodes.
17343 2013-03-06 Richard Biener <rguenther@suse.de>
17345 PR middle-end/50494
17346 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17347 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17350 2013-02-13 Richard Biener <rguenther@suse.de>
17353 * varasm.c (output_constant_def_1): Get the decl representing
17354 the constant as argument.
17355 (output_constant_def): Wrap output_constant_def_1.
17356 (make_decl_rtl): Use output_constant_def_1 with the decl
17357 representing the constant.
17358 (build_constant_desc): Optionally re-use a decl already
17359 representing the constant.
17360 (tree_output_constant_def): Adjust.
17362 2013-03-06 Joey Ye <joey.ye@arm.com>
17365 * gcc.c (convert_white_space): New function.
17366 (main): Handles white space in function name.
17368 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
17371 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
17372 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
17373 to SH_DIV_CALL_TABLE for TARGET_SH2.
17374 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
17376 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
17377 call-table options.
17379 2013-03-05 Sterling Augustine <saugustine@google.com>
17380 Cary Coutant <ccoutant@google.com>
17383 * dwarf2out.c (resolve_addr): Don't call
17384 remove_loc_list_addr_table_entries a second time for the same
17387 2013-03-05 Jakub Jelinek <jakub@redhat.com>
17390 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
17391 (avoid_complex_debug_insns): New function.
17392 (expand_debug_locations): Call it.
17394 PR rtl-optimization/56484
17395 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
17396 lifetimes of hard registers on small register class machines.
17398 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
17400 * config/microblaze/microblaze-protos.h: Rename
17401 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
17402 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
17404 (microblaze_fast_interrupt_function_p): New function.
17405 (microblaze_is_interrupt_handler): Rename to
17406 microblaze_is_interrupt_variant and add fast_interrupt check.
17407 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
17408 (save_restore_insns): Likewise.
17409 (compute_frame_size): Likewise.
17410 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
17411 (microblaze_globalize_label): Likewise.
17412 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
17413 * config/microblaze/microblaze.md: Use wrapper
17414 microblaze_is_interrupt_variant.
17416 2013-03-05 Kai Tietz <ktietz@redhat.com>
17418 * sdbout.c (sdbout_one_type): Switch to current function's section
17419 supporting cold/hot.
17421 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
17423 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
17426 2013-03-05 Jakub Jelinek <jakub@redhat.com>
17428 PR middle-end/56461
17429 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
17430 if VALGRIND_GET_VBITS is defined, temporarily make object
17431 memory all defined, and restore previous valgrind addressability
17432 and definability afterwards. Free this_object at the end.
17434 PR middle-end/56461
17435 * lra.c (lra): Call lra_clear_live_ranges if live_p,
17436 right before calling lra_create_live_ranges, also call it
17437 when clearing live_p. Only call lra_clear_live_ranges
17438 at the end if live_p.
17440 PR middle-end/56461
17441 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
17443 2013-03-05 Richard Biener <rguenther@suse.de>
17445 PR tree-optimization/56521
17446 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
17449 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
17452 * except.h (remove_unreachable_eh_regions): New prototype.
17453 * except.c (remove_eh_handler_splicer): New function, split out
17454 of remove_eh_handler.
17455 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
17456 warning about running it on many EH regions one at a time.
17457 (remove_unreachable_eh_regions_worker): New function, walk the
17458 EH tree in depth-first order and remove non-marked regions.
17459 (remove_unreachable_eh_regions): New function.
17460 * tree-eh.c (mark_reachable_handlers): New function, split out
17461 from remove_unreachable_handlers.
17462 (remove_unreachable_handlers): Use mark_reachable_handlers and
17463 remove_unreachable_eh_regions.
17464 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
17465 and remove_unreachable_eh_regions.
17467 2013-03-05 Richard Biener <rguenther@suse.de>
17469 PR middle-end/56525
17470 * loop-init.c (fix_loop_structure): Remove loops in two stages,
17471 not freeing them until the end.
17473 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17475 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
17477 2013-03-05 Richard Biener <rguenther@suse.de>
17479 PR tree-optimization/56270
17480 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
17481 of loads after scheduling an SLP instance.
17483 2013-03-05 Jakub Jelinek <jakub@redhat.com>
17485 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
17487 (check_gcc_parallelize): Run guality.exp as a separate job from
17488 vect.exp with unsorted.exp and $(dg_target_exps) separately from
17489 struct-layout-1.exp with stackalign.exp.
17491 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
17493 PR middle-end/56461
17494 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
17495 load_index sbitmap even if some bit in it isn't set.
17497 PR middle-end/56461
17498 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
17499 (discover_iteration_bound_by_body_walk): Change queues to
17500 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
17501 spelling in comment. Call safe_push on queues[bound_index] directly.
17502 Release queues[queue_index] in every iteration unconditionally.
17503 Release bounds vector.
17505 PR middle-end/56461
17506 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
17507 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
17508 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
17511 2013-03-05 Richard Biener <rguenther@suse.de>
17514 * tree-inline.c (remap_blocks_to_null): New function.
17515 (expand_call_inline): When expanding a call stmt without
17516 an associated block inline remap all callee blocks to NULL.
17518 2013-03-05 Jakub Jelinek <jakub@redhat.com>
17520 PR rtl-optimization/56494
17521 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
17522 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
17523 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
17525 PR middle-end/56461
17526 * sel-sched-ir.c (free_sched_pools): Release
17527 succs_info_pool.stack[succs_info_pool.max_top] vectors too
17528 if succs_info_pool.max_top isn't -1.
17531 * opts.c (opts_obstack, opts_concat): Moved to...
17532 * opts-common.c (opts_obstack, opts_concat): ... here.
17534 2013-03-04 Jakub Jelinek <jakub@redhat.com>
17536 PR middle-end/56461
17537 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
17539 2013-03-04 Martin Jambor <mjambor@suse.cz>
17541 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
17542 all appropriate places.
17544 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
17546 PR tree-optimization/56424
17547 * ipa-split.c (split_function): Do not set the RSO flag if result is
17548 not by reference and its type is a register type.
17550 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
17552 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
17553 (microblaze_legitimate_pic_operand): Likewise
17554 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
17555 new function microblaze_legitimate_pic_operand
17556 * config/microblaze/microblaze-protos.h
17557 (microblaze_legitimate_pic_operand): Declare.
17559 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17561 * config/microblaze/predicates.md (call_insn_simple_operand):
17562 New predicate for supported rtx code types.
17563 * config/microblaze/microblaze.md (call_internal1): Use
17564 call_insn_simple_operand predicate.
17566 2013-03-04 Jakub Jelinek <jakub@redhat.com>
17568 PR middle-end/56461
17569 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
17570 partitions.ordered_remove.
17572 PR middle-end/56461
17573 * tree-vect-stmts.c (vectorizable_conversion): Don't call
17574 vec_oprnds0.create (1) for modifier == NONE.
17576 PR middle-end/56461
17577 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
17578 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
17579 vec_oprnds1 right before pushing anything to it for
17582 PR middle-end/56461
17583 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
17584 set nbbs to 0 instead of having separate code path.
17585 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
17586 instead of false as last argument if returning NULL.
17588 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
17590 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
17591 the attribute is now called "target" instead of "option".
17592 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
17593 * doc/tm.texi.in (Target Attributes): Likewise document the correct
17594 attribute/pragma name for TARGET_OPTION_VALID_P and
17595 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
17596 * doc/tm.texi: Regenerated.
17598 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
17600 * config/microblaze/microblaze.c:
17601 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
17602 * config/microblaze/microblaze.h: Add -mxl-reorder to
17604 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
17605 instructions emitted if TARGET_REORDER.
17606 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
17607 or 0 for -m/-mno case, but initialises as 2 to detect default use case
17610 2013-03-01 Xinliang David Li <davidxl@google.com>
17612 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
17615 2013-03-01 Jakub Jelinek <jakub@redhat.com>
17617 PR middle-end/56461
17618 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
17619 vector even when returning true. Fix up function comment formatting.
17621 PR middle-end/56461
17622 * ira-build.c (ira_loop_nodes_count): New variable.
17623 (create_loop_tree_nodes): Initialize it.
17624 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
17626 PR middle-end/56461
17627 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
17628 method on dr_chain and result_chain.
17629 * tree-vect-stmts.c (vectorizable_store): Only call
17630 result_chain.create if j == 0.
17632 PR middle-end/56461
17633 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
17634 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
17635 before overwriting it.
17637 2013-03-01 Tobias Burnus <burnus@net-b.de>
17639 * doc/extended.texi (C Extensions): Change order in @menu
17641 (Other MIPS Built-in Functions): Move last MIPS entry before
17642 "picoChip Built-in Functions".
17643 (SH Built-in Functions): Move after RX Built-in Functions.
17644 * doc/gcc.texi (Introduction): Change order in @menu
17646 * doc/md.texi (Constraints): Ditto.
17647 * gty.texi (Type Information): Ditto.
17648 (User-provided marking routines for template types): Make
17650 * doc/invoke.texi (AArch64 Options): Move before
17651 "Adapteva Epiphany Options".
17653 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
17654 Jakub Jelinek <jakub@redhat.com>
17657 * asan.c (gate_asan): Lookup no_sanitize_address instead of
17658 no_address_safety_analysis attribute.
17659 * doc/extend.texi (no_address_safety_attribute): Rename to
17660 no_sanitize_address attribute, mention no_address_safety_analysis
17661 attribute as deprecated alias.
17663 2013-02-28 Jakub Jelinek <jakub@redhat.com>
17665 PR middle-end/56461
17666 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
17667 type to vec<vec<tree> > *.
17668 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
17669 to be vec<tree> instead of vec<tree> *, set vec_defs
17670 to vNULL and call vec_defs.create (number_of_vects), adjust other
17672 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
17673 vectorizable_condition): Adjust vect_get_slp_defs callers.
17675 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17677 * config/aarch64/aarch64.c
17678 (aarch64_float_const_representable): Remove unused variable.
17680 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17682 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
17684 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17686 * config/aarch64/aarch64-builtins.c
17687 (aarch64_init_simd_builtins): Make static.
17689 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17691 * config/aarch64/aarch64.c
17692 (aarch64_simd_make_constant): Make static.
17694 2013-02-28 Martin Jambor <mjambor@suse.cz>
17696 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
17697 with no initialization to the RHS of debug statements.
17699 2013-02-28 Martin Jambor <mjambor@suse.cz>
17701 PR tree-optimization/56294
17702 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
17704 (get_access_replacement): Do not call create_access_replacement.
17705 Assert a replacement exists.
17706 (get_repl_default_def_ssa_name): Create the replacement declaration
17709 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17711 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
17712 final_end_function.
17714 2013-02-28 Marek Polacek <polacek@redhat.com>
17716 PR rtl-optimization/56466
17717 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
17718 if we're changing a loop.
17719 (peel_loops_completely): Likewise.
17721 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
17724 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
17726 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17729 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
17730 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
17731 INTX_FTYPE_FX, FX_FTYPE_INTX.
17732 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
17734 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17736 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
17737 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
17738 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
17739 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
17740 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
17741 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
17742 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
17743 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
17744 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
17745 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
17746 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
17747 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
17748 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
17749 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
17750 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
17751 (avrxmega6): Increase max flash segments from 5 to 6.
17752 * config/avr/t-multilib: Regenerate.
17753 * config/avr/avr-tables.opt: Regenerate.
17754 * doc/avr-mmcu.texi: Regenerate.
17756 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17758 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
17759 (avr_device_to_arch): Rename to avr_device_to_ld.
17760 (avr_device_to_as): New prototype.
17761 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
17762 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
17763 * config/avr/driver-avr.c (avr_device_to_as): New.
17764 (avr_device_to_arch): Rename to avr_device_to_ld.
17766 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17768 PR middle-end/56461
17769 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
17770 method on dr_chain and result_chain.
17772 PR middle-end/56461
17773 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
17774 pointer_set_destroy on not_executed_last_iteration.
17776 PR middle-end/56461
17777 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
17779 PR middle-end/56461
17780 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
17781 FOR_EACH_DEFINED_FUNCTION when freeing state.
17783 PR middle-end/56461
17784 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
17786 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
17789 PR middle-end/56461
17790 * ipa-cp.c (decide_whether_version_node): Call vec_free on
17791 known_aggs[i].items and release known_aggs vector.
17793 PR middle-end/56461
17794 * ipa-reference.c (propagate): Free node_info even for alias nodes.
17796 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17798 * config/microblaze/microblaze.c (microblaze_emit_compare):
17799 Use xor for EQ/NE comparisions.
17800 * config/microblaze/microblaze.md (cstoresf4): Add constraints
17801 (cbranchsf4): Adjust operator to comparison_operator.
17803 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17805 PR middle-end/56461
17806 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
17808 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
17809 vec_safe_push, always update *slot.
17810 (redirect_edge_var_map_clear): Use vec_free.
17811 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
17812 (free_var_map_entry): Use vec_free.
17813 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
17814 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
17816 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
17818 PR middle-end/45472
17819 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
17820 when the may_trap_p bit of the exprs being merged differs.
17821 Reorder tests for speculativeness in the logical and operator.
17823 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17825 * incpath.c (add_standard_paths): Use reconcat instead of concat
17826 where appropriate and avoid leaking memory.
17828 * opts.h: Include obstack.h.
17829 (opts_concat): New prototype.
17830 (opts_obstack): New declaration.
17831 * opts.c (opts_concat): New function.
17832 (opts_obstack): New variable.
17833 (init_options_struct): Call gcc_init_obstack on opts_obstack.
17834 (finish_options): Use opts_concat instead of concat
17835 and XOBNEWVEC instead of XNEWVEC.
17836 * opts-common.c (generate_canonical_option, decode_cmdline_option,
17837 generate_option): Likewise.
17838 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
17839 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
17842 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
17843 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
17845 2013-02-26 Jakub Jelinek <jakub@redhat.com>
17847 PR middle-end/56461
17848 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
17850 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17852 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
17853 (arm_block_move_unaligned_straight): Likewise.
17854 (arm_adjust_block_mem): Likewise.
17856 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17859 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
17860 temp, cond and label.
17861 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
17864 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
17865 * config/c6x/c6x.h (dbx_register_map): Update declaration.
17868 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
17869 of prologue/epilogue functions.
17872 * config/tilegx/tilegx.c (tilegx_expand_prologue):
17873 Remove unused variable cfa_offset.
17874 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
17877 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
17878 type promotion to unsigned.
17881 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
17882 for HOST_WIDE_INT of 32 bit / same size as int.
17883 (arm_block_move_unaligned_straight): Likewise.
17884 (arm_adjust_block_mem): Likewise.
17887 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
17890 2013-02-26 Marek Polacek <polacek@redhat.com>
17892 PR tree-optimization/56426
17893 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
17895 2013-02-26 Richard Biener <rguenther@suse.de>
17898 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
17899 unused variable loops.
17901 2013-02-26 Jakub Jelinek <jakub@redhat.com>
17903 PR tree-optimization/56448
17904 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
17905 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
17906 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
17907 later operands of the references, or even first operand for
17908 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
17910 PR tree-optimization/56443
17911 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
17912 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
17913 to type_for_mode langhook.
17915 2013-02-25 Matt Turner <mattst88@gmail.com>
17917 * doc/invoke.texi: Document r4700.
17919 2013-02-25 Richard Biener <rguenther@suse.de>
17921 PR tree-optimization/56175
17922 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
17924 (simplify_bitwise_binary): ... here. Also guard the conversion
17925 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
17927 2013-02-25 Catherine Moore <clm@codesourcery.com>
17930 2013-02-24 Catherine Moore <clm@codesourcery.com>
17931 Maciej W. Rozycki <macro@codesourcery.com>
17932 Tom de Vries <tom@codesourcery.com>
17933 Nathan Sidwell <nathan@codesourcery.com>
17934 Iain Sandoe <iain@codesourcery.com>
17935 Nathan Froyd <froydnj@codesourcery.com>
17936 Chao-ying Fu <fu@mips.com>
17938 * doc/extend.texi: (micromips, nomicromips, nocompression):
17939 Document new function attributes.
17940 * doc/invoke.texi (minterlink-compressed, mmicromips,
17941 m14k, m14ke, m14kec): Document new options.
17942 (minterlink-mips16): Update documentation.
17943 * doc/md.texi (ZC, ZD): Document new constraints.
17944 * configure.ac (gcc_cv_as_micromips): Check if linker
17945 supports the .set micromips directive.
17946 * configure: Regenerate.
17947 * config.in: Regenerate.
17948 * config/mips/mips-tables.opt: Regenerate.
17949 * config/mips/micromips.md: New file.
17950 * constraints.md (ZC, AD): New constraints.
17951 * config/mips/predicates.md (movep_src_register): New predicate.
17952 (movep_src_operand): New predicate.
17953 (non_volatile_mem_operand): New predicate.
17954 * config/mips/mips.md (multimem): New type.
17955 (length): Differentiate between 17-bit and 18-bit branch offsets.
17956 (MOVEP1, MOVEP2): New mode iterator.
17957 (mov_<load>l): Use ZC constraint.
17958 (mov_<load>r): Likewise.
17959 (mov_<store>l): Likewise.
17960 (mov_<store>r): Likewise.
17961 (*branch_equality<mode>_inverted): Add microMIPS support.
17962 (*branch_equality<mode>): Likewise.
17963 (*jump_absolute): Likewise.
17964 (indirect_jump_<mode>): Likewise.
17965 (tablejump_<mode>): Likewise.
17966 (<optab>_internal): Likewise.
17967 (sibcall_internal): Likewise.
17968 (sibcall_value_internal): Likewise.
17969 (prefetch): Use constraint ZD.
17970 * config/mips/mips.opt (minterlink-compressed): New option.
17971 (minterlink-mips16): Now an alias for minterlink-compressed.
17972 (mmicromips): New option.
17973 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17974 (compare_and_swap_12): Likewise.
17975 (sync_add<mode>): Likewise.
17976 (sync_<optab>_12): Likewise.
17977 (sync_old_<optab>_12): Likewise.
17978 (sync_new_<optab>_12): Likewise.
17979 (sync_nand_12): Likewise.
17980 (sync_old_nand_12): Likewise.
17981 (sync_new_nand_12): Likewise.
17982 (sync_sub<mode>): Likewise.
17983 (sync_old_add<mode>): Likewise.
17984 (sync_old_sub<mode>): Likewise.
17985 (sync_new_add<mode>): Likewise.
17986 (sync_new_sub<mode>): Likewise.
17987 (sync_<optab><mode>): Likewise.
17988 (sync_old_<optab><mode>): Likewise.
17989 (sync_new_<optab><mode>): Likewise.
17990 (sync_nand<mode>): Likewise.
17991 (sync_old_nand<mode>): Likewise.
17992 (sync_new_nand<mode>): Likewise.
17993 (sync_lock_test_and_set<mode>): Likewise.
17994 (test_and_set_12): Likewise.
17995 (atomic_compare_and_swap<mode>): Likewise.
17996 (atomic_exchange<mode>_llsc): Likewise.
17997 (atomic_fetch_add<mode>_llsc): Likewise.
17998 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17999 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18000 (umips_save_restore_pattern_p): Likewise.
18001 (umips_load_store_pair_p): Likewise.
18002 (umips_output_load_store_pair): Likewise.
18003 (umips_movep_target_p): Likewise.
18004 (umips_12bit_offset_address_p): Likewise.
18005 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18006 (mips_base_mips16): Rename this...
18007 (mips_base_compression_flags): ...to this. Update all uses.
18008 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18009 (mips_mips16_decl_p): Delete.
18010 (mips_nomips16_decl_p): Delete.
18011 (mips_get_compress_on_flags): New function.
18012 (mips_get_compress_off_flags): New function.
18013 (mips_get_compress_mode): New function.
18014 (mips_get_compress_on_name): New function.
18015 (mips_get_compress_off_name): New function.
18016 (mips_insert_attributes): Support multiple compression types.
18017 (mips_merge_decl_attributes): Likewise.
18018 (umips_12bit_offset_address_p): New function.
18019 (mips_start_function_definition): Emit .set micromips directive.
18020 (mips_call_may_need_jalx_p): New function.
18021 (mips_function_ok_for_sibcall): Add microMIPS support.
18022 (mips_print_operand_punctuation): Support short delay slots and
18024 (umips_swm_mask, umips_swm_encoding): New.
18025 (umips_build_save_restore): New function.
18026 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18027 (was_mips16_p): Remove.
18028 (old_compression_mode): New.
18029 (mips_set_compression_mode): New function.
18030 (mips_set_current_function): Add microMIPS support.
18031 (mips_option_override): Likewise.
18032 (umips_save_restore_pattern_p): New function.
18033 (umips_output_save_restore): New function.
18034 (umips_load_store_pair_p_1): New function.
18035 (umips_load_store_pair_p): New function.
18036 (umips_output_load_store_pair_1): New function.
18037 (umips_output_load_store_pair): New function.
18038 (umips_movep_target_p) New function.
18039 (mips_prepare_pch_save): Add microMIPS support.
18040 * config/mips/mips.h (TARGET_COMPRESSION): New.
18041 (TARGET_CPU_CPP_BUILTINS): Update macro
18042 to use new compression flags and to support microMIPS.
18043 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18044 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18045 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18046 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18047 (ASM_SPEC): Support mmicromips and mno-micromips.
18048 (M16STORE_REG_P): New macro.
18049 (MIPS_CALL): Support TARGET_MICROMIPS.
18050 (MICROMIPS_J): New macro.
18051 (mips_base_mips16): Rename this...
18052 (mips_base_compression_flags): ...to this.
18053 (UMIPS_12BIT_OFFSET_P): New macro.
18054 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18055 (MULTILIB_DIRNAMES): Likewise.
18057 2013-02-25 Tom de Vries <tom@codesourcery.com>
18059 PR rtl-optimization/56131
18060 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18061 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18062 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18064 2013-02-25 Tobias Burnus <burnus@net-b.de>
18066 * doc/invoke.texi (-fsanitize=): Move from optimization
18067 to debugging options.
18069 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18071 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18073 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18074 Alexander Monakov <amonakov@ispras.ru>
18076 PR middle-end/56077
18077 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18078 flush pending lists also on non-jumps. Adjust comment.
18080 2013-02-24 Catherine Moore <clm@codesourcery.com>
18081 Maciej W. Rozycki <macro@codesourcery.com>
18082 Tom de Vries <tom@codesourcery.com>
18083 Nathan Sidwell <nathan@codesourcery.com>
18084 Iain Sandoe <iain@codesourcery.com>
18085 Nathan Froyd <froydnj@codesourcery.com>
18086 Chao-ying Fu <fu@mips.com>
18088 * doc/extend.texi: (micromips, nomicromips, nocompression):
18089 Document new function attributes.
18090 * doc/invoke.texi (minterlink-compressed, mmicromips,
18091 m14k, m14ke, m14kec): Document new options.
18092 (minterlink-mips16): Update documentation.
18093 * doc/md.texi (ZC, ZD): Document new constraints.
18094 * configure.ac (gcc_cv_as_micromips): Check if linker
18095 supports the .set micromips directive.
18096 * configure: Regenerate.
18097 * config.in: Regenerate.
18098 * config/mips/mips-tables.opt: Regenerate.
18099 * config/mips/micromips.md: New file.
18100 * constraints.md (ZC, AD): New constraints.
18101 * config/mips/predicates.md (movep_src_register): New predicate.
18102 (movep_src_operand): New predicate.
18103 (non_volatile_mem_operand): New predicate.
18104 * config/mips/mips.md (multimem): New type.
18105 (length): Differentiate between 17-bit and 18-bit branch offsets.
18106 (MOVEP1, MOVEP2): New mode iterator.
18107 (mov_<load>l): Use ZC constraint.
18108 (mov_<load>r): Likewise.
18109 (mov_<store>l): Likewise.
18110 (mov_<store>r): Likewise.
18111 (*branch_equality<mode>_inverted): Add microMIPS support.
18112 (*branch_equality<mode>): Likewise.
18113 (*jump_absolute): Likewise.
18114 (indirect_jump_<mode>): Likewise.
18115 (tablejump_<mode>): Likewise.
18116 (<optab>_internal): Likewise.
18117 (sibcall_internal): Likewise.
18118 (sibcall_value_internal): Likewise.
18119 (prefetch): Use constraint ZD.
18120 * config/mips/mips.opt (minterlink-compressed): New option.
18121 (minterlink-mips16): Now an alias for minterlink-compressed.
18122 (mmicromips): New option.
18123 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18124 (compare_and_swap_12): Likewise.
18125 (sync_add<mode>): Likewise.
18126 (sync_<optab>_12): Likewise.
18127 (sync_old_<optab>_12): Likewise.
18128 (sync_new_<optab>_12): Likewise.
18129 (sync_nand_12): Likewise.
18130 (sync_old_nand_12): Likewise.
18131 (sync_new_nand_12): Likewise.
18132 (sync_sub<mode>): Likewise.
18133 (sync_old_add<mode>): Likewise.
18134 (sync_old_sub<mode>): Likewise.
18135 (sync_new_add<mode>): Likewise.
18136 (sync_new_sub<mode>): Likewise.
18137 (sync_<optab><mode>): Likewise.
18138 (sync_old_<optab><mode>): Likewise.
18139 (sync_new_<optab><mode>): Likewise.
18140 (sync_nand<mode>): Likewise.
18141 (sync_old_nand<mode>): Likewise.
18142 (sync_new_nand<mode>): Likewise.
18143 (sync_lock_test_and_set<mode>): Likewise.
18144 (test_and_set_12): Likewise.
18145 (atomic_compare_and_swap<mode>): Likewise.
18146 (atomic_exchange<mode>_llsc): Likewise.
18147 (atomic_fetch_add<mode>_llsc): Likewise.
18148 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18149 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18150 (umips_save_restore_pattern_p): Likewise.
18151 (umips_load_store_pair_p): Likewise.
18152 (umips_output_load_store_pair): Likewise.
18153 (umips_movep_target_p): Likewise.
18154 (umips_12bit_offset_address_p): Likewise.
18155 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18156 (mips_base_mips16): Rename this...
18157 (mips_base_compression_flags): ...to this. Update all uses.
18158 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18159 (mips_mips16_decl_p): Delete.
18160 (mips_nomips16_decl_p): Delete.
18161 (mips_get_compress_on_flags): New function.
18162 (mips_get_compress_off_flags): New function.
18163 (mips_get_compress_mode): New function.
18164 (mips_get_compress_on_name): New function.
18165 (mips_get_compress_off_name): New function.
18166 (mips_insert_attributes): Support multiple compression types.
18167 (mips_merge_decl_attributes): Likewise.
18168 (umips_12bit_offset_address_p): New function.
18169 (mips_start_function_definition): Emit .set micromips directive.
18170 (mips_call_may_need_jalx_p): New function.
18171 (mips_function_ok_for_sibcall): Add microMIPS support.
18172 (mips_print_operand_punctuation): Support short delay slots and
18174 (umips_swm_mask, umips_swm_encoding): New.
18175 (umips_build_save_restore): New function.
18176 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18177 (was_mips16_p): Remove.
18178 (old_compression_mode): New.
18179 (mips_set_compression_mode): New function.
18180 (mips_set_current_function): Add microMIPS support.
18181 (mips_option_override): Likewise.
18182 (umips_save_restore_pattern_p): New function.
18183 (umips_output_save_restore): New function.
18184 (umips_load_store_pair_p_1): New function.
18185 (umips_load_store_pair_p): New function.
18186 (umips_output_load_store_pair_1): New function.
18187 (umips_output_load_store_pair): New function.
18188 (umips_movep_target_p) New function.
18189 (mips_prepare_pch_save): Add microMIPS support.
18190 * config/mips/mips.h (TARGET_COMPRESSION): New.
18191 (TARGET_CPU_CPP_BUILTINS): Update macro
18192 to use new compression flags and to support microMIPS.
18193 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18194 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18195 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18196 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18197 (ASM_SPEC): Support mmicromips and mno-micromips.
18198 (M16STORE_REG_P): New macro.
18199 (MIPS_CALL): Support TARGET_MICROMIPS.
18200 (MICROMIPS_J): New macro.
18201 (mips_base_mips16): Rename this...
18202 (mips_base_compression_flags): ...to this.
18203 (UMIPS_12BIT_OFFSET_P): New macro.
18204 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18205 (MULTILIB_DIRNAMES): Likewise.
18207 2013-02-24 Jakub Jelinek <jakub@redhat.com>
18210 * target-globals.c (save_target_globals): For init_reg_sets and
18211 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18213 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18215 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18216 * config/aarch64/t-aarch64
18217 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18219 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18221 PR inline-asm/56148
18222 * lra-constraints.c (process_alt_operands): Reload operand
18223 conflicting with earlier clobber only if no more other conflicting
18226 2013-02-22 Jakub Jelinek <jakub@redhat.com>
18229 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18230 if not linking a shared library.
18232 2013-02-22 Seth LaForge <sethml@google.com>
18234 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18236 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18238 * config/arm/arm.md (split for extendsidi): Update condition.
18239 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18240 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18241 (qhs_zextenddi_cstr): Likewise.
18243 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18245 PR middle-end/56420
18246 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18247 avoid signed wrapping.
18248 (expand_mult): Handle properly multiplication by
18249 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18250 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18251 in the compiler if coeff is HOST_WIDE_INT_MIN.
18252 (expand_divmod): Don't make ext_op1 static, change it's type to
18253 uhwi. Avoid undefined behavior in -INTVAL (op1).
18255 PR rtl-optimization/50339
18256 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18258 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18259 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18260 into splitting_ashiftrt field.
18261 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18263 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18266 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
18268 PR middle-end/56108
18269 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18270 are sure to go irrevocable.
18272 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
18274 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18275 scalars are valid operands.
18277 2013-02-21 Martin Jambor <mjambor@suse.cz>
18279 PR tree-optimization/56310
18280 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18281 only matching indices and non-negative final offsets.
18282 (intersect_aggregates_with_edge): Pass src_idx to
18283 agg_replacements_to_vector. Pass src_idx insstead of index to
18284 intersect_with_agg_replacements.
18286 2013-02-21 Martin Jambor <mjambor@suse.cz>
18288 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18289 instead of hard-wired defaults.
18291 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18293 * doc/invoke.texi (MIPS Options): Update documentation of the
18294 floating-point multiply-accumulate instruction restrictions.
18296 2013-02-21 Kostya Serebryany <kcc@google.com>
18298 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18299 asan_shadow_offset on x86_64 linux.
18301 2013-02-21 Richard Biener <rguenther@suse.de>
18303 PR tree-optimization/56415
18305 2013-02-11 Richard Biener <rguenther@suse.de>
18307 PR tree-optimization/56273
18308 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18311 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18314 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18317 PR inline-asm/56405
18318 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18319 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18321 2013-02-20 Jan Hubicka <jh@suse.cz>
18323 PR tree-optimization/56265
18324 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18325 when target is referenced for first time.
18327 2013-02-20 Richard Biener <rguenther@suse.de>
18329 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18330 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18331 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18332 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18333 not return anything.
18334 (rename_ssa_copies): Do not remove unused locals.
18335 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
18336 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
18337 * passes.c (execute_function_todo): Do not schedule unused locals
18338 removal if cleanup_tree_cfg did something.
18339 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18340 about the number of removed locals.
18342 2013-02-20 Richard Biener <rguenther@suse.de>
18344 PR tree-optimization/56398
18345 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
18347 2013-02-20 Martin Jambor <mjambor@suse.cz>
18349 PR tree-optimization/55334
18350 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18351 restricted pointers to arrays.
18353 2013-02-20 Richard Biener <rguenther@suse.de>
18354 Jakub Jelinek <jakub@redhat.com>
18356 PR tree-optimization/56396
18357 * tree-ssa-ccp.c (n_const_val): New static variable.
18358 (get_value): Return NULL for SSA names we don't have a lattice
18360 (ccp_initialize): Initialize n_const_val.
18361 * tree-ssa-copy.c (n_copy_of): New static variable.
18362 (init_copy_prop): Initialize n_copy_of.
18363 (get_value): Return NULL_TREE for SSA names we don't have a
18366 2013-02-20 Martin Jambor <mjambor@suse.cz>
18368 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
18370 2013-02-20 Richard Biener <rguenther@suse.de>
18372 * genpreds.c (write_lookup_constraint): Do not compare first
18373 letter of the constraint again.
18375 2013-02-20 Richard Biener <rguenther@suse.de>
18377 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
18379 (get_use_iv_cost): Terminate hashtable walk when coming across
18382 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
18384 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
18385 reassociation for avx2 targets.
18387 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18389 * config/microblaze/microblaze.c: microblaze_has_clz = 0
18390 Add version check for v8.10.a to enable microblaze_has_clz
18391 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
18392 version and TARGET_PATTERN_COMPARE check
18393 * config/microblaze/microblaze.md: New clzsi2 instruction
18395 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18397 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
18398 function before branching.
18400 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
18402 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
18404 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
18406 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
18408 PR middle-end/55889
18409 * sel-sched.c: Include ira.h.
18410 (implicit_clobber_conflict_p): New function.
18411 (moveup_expr): Use it.
18412 * Makefile.in (sel-sched.o): Depend on ira.h.
18414 2013-02-19 Richard Biener <rguenther@suse.de>
18416 PR tree-optimization/56384
18417 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
18418 (vn_hash_type): Split out from ...
18419 (vn_hash_constant_with_type): ... here.
18420 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
18421 (vn_phi_eq): Compare types from vn_phi_s structure.
18422 (vn_phi_lookup): Populate vn_phi_s type.
18423 (vn_phi_insert): Likewise.
18425 2013-02-19 Jakub Jelinek <jakub@redhat.com>
18427 PR tree-optimization/56350
18428 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
18429 if haven't found reduction or nested cycle operand, rather than
18430 asserting we must find it.
18432 PR tree-optimization/56381
18433 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
18436 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
18437 Jakub Jelinek <jakub@redhat.com>
18440 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
18441 (swap_optab_enable): Same.
18442 (init_all_optabs): Use argument instead of global.
18443 * tree.h (struct tree_optimization_option): New field target_optabs.
18444 * expr.h (init_all_optabs): Add argument to prototype.
18445 (TREE_OPTIMIZATION_OPTABS): New.
18446 (save_optabs_if_changed): Protoize.
18447 * optabs.h: Declare this_fn_optabs.
18448 * optabs.c (save_optabs_if_changed): New.
18449 Declare this_fn_optabs.
18450 (init_optabs): Add argument to init_all_optabs() call.
18451 * function.c (invoke_set_current_function_hook): Handle per
18453 * function.h (struct function): New field optabs.
18454 * config/mips/mips.c (mips_set_mips16_mode): Handle when
18455 optimization_current_node has changed.
18456 * target-globals.h (save_target_globals_default_opts): Protoize.
18457 * target-globals.c (save_target_globals_default_opts): New.
18459 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18462 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
18463 registers %fr12 and %fr12R as call used.
18466 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
18467 and HImode, require all displacements to be an integer multiple of
18469 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
18470 only allow QImode and HImode when reload is in progress and strict is
18471 true. Likewise for symbolic addresses. Use base14_operand to check
18472 displacements in REG+BASE addresses.
18474 2013-02-18 Richard Biener <rguenther@suse.de>
18476 PR tree-optimization/56366
18477 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18478 handle sign-conversion of outer-loop initial induction value.
18480 2013-02-18 Richard Biener <rguenther@suse.de>
18482 PR middle-end/56349
18483 * cfghooks.c (merge_blocks): If we merge a latch into another
18484 block adjust references to it.
18485 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
18486 (verify_loop_structure): Verify that a recorded latch is in fact
18489 2013-02-18 Richard Biener <rguenther@suse.de>
18491 PR tree-optimization/56321
18492 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
18493 order SSA name release and virtual operand unlinking.
18495 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18497 * config/microblaze/microblaze.md (save_stack_block): Define.
18498 (restore_stack_block): Likewise.
18500 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18502 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
18503 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
18504 * config/microblaze/microblaze.c (microblaze_option_override):
18505 Bail out early for PIC modes when target does not support PIC.
18507 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18509 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
18510 Replace with a microblaze version.
18511 (microblaze_trampoline_init): Adapt for microblaze.
18512 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
18515 2013-02-16 Jakub Jelinek <jakub@redhat.com>
18516 Dodji Seketeli <dodji@redhat.com>
18519 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
18520 (instrument_mem_region_access): Do not forget to always put
18521 instrumentation of the of 'base' and 'base + len' in a "if (len !=
18522 0) statement, even for cases where either 'base' or 'base + len'
18523 are not instrumented -- because they have been previously
18524 instrumented. Simplify the logic by putting all the statements
18525 instrument 'base + len' inside a sequence, and then insert that
18526 sequence right before the current insertion point. Then, to
18527 instrument 'base + len', just get an iterator on that statement.
18528 And do not forget to update the pointer to iterator the function
18529 received as argument.
18531 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
18533 PR rtl-optimization/56348
18534 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
18536 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
18538 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
18539 (clean_graph_dump_file): Pass base to start_graph_dump.
18541 2013-02-14 Richard Henderson <rth@redhat.com>
18544 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
18546 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
18548 * collect2-aix.h: Define F_LOADONLY.
18550 2013-02-14 Richard Biener <rguenther@suse.de>
18553 * varasm.c (output_constant_def_1): Get the decl representing
18554 the constant as argument.
18555 (output_constant_def): Wrap output_constant_def_1.
18556 (make_decl_rtl): Use output_constant_def_1 with the decl
18557 representing the constant.
18558 (build_constant_desc): Optionally re-use a decl already
18559 representing the constant.
18560 (tree_output_constant_def): Adjust.
18562 2013-02-14 Dodji Seketeli <dodji@redhat.com>
18565 * asan.c (instrument_builtin_call): Really put the length of the
18566 second source argument into src1_len.
18568 2013-02-13 Jakub Jelinek <jakub@redhat.com>
18570 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
18571 argument. If it is false, don't create edge from then_bb to
18573 (insert_if_then_before_iter): Pass true to it.
18574 (build_check_stmt): Pass false to it.
18575 (transform_statements): Flush hash table only on extended basic
18576 block boundaries, rather than at the beginning of every bb.
18577 Don't flush hash table on nonfreeing_call_p calls.
18578 * tree-flow.h (nonfreeing_call_p): New prototype.
18579 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
18581 2013-02-13 David S. Miller <davem@davemloft.net>
18583 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
18585 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
18588 * ira.c (max_regno_before_ira): Move from ...
18590 (fix_reg_equiv_init): Use max_regno_before_ira instead of
18593 2013-02-13 Jakub Jelinek <jakub@redhat.com>
18595 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
18597 2013-02-13 Richard Biener <rguenther@suse.de>
18600 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
18601 globals in MEM_REFs.
18603 2013-02-13 Richard Biener <rguenther@suse.de>
18605 * loop-init.c (loop_optimizer_init): Clear loop state when
18606 re-initializing preserved loops.
18607 * loop-unswitch.c (unswitch_single_loop): Return whether
18608 we unswitched the loop. Do not verify loop state here.
18609 (unswitch_loops): When we unswitched a loop discover new loops.
18611 2013-02-13 Kostya Serebryany <kcc@google.com>
18613 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
18615 * sanitizer.def: Rename __asan_init to __asan_init_v1.
18617 2013-02-12 Dodji Seketeli <dodji@redhat.com>
18619 Avoid instrumenting duplicated memory access in the same basic block
18620 * Makefile.in (asan.o): Add new dependency on hash-table.h
18621 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
18622 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
18623 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
18624 (free_mem_ref_resources, has_mem_ref_been_instrumented)
18625 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
18626 (get_mem_ref_of_assignment): New functions.
18627 (get_mem_refs_of_builtin_call): Extract from
18628 instrument_builtin_call and tweak a little bit to make it fit with
18630 (instrument_builtin_call): Use the new
18631 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
18632 of is_gimple_builtin_call.
18633 (instrument_derefs, instrument_mem_region_access): Insert the
18634 instrumented memory reference into the hash table.
18635 (maybe_instrument_assignment): Renamed instrument_assignment into
18636 this, and change it to advance the iterator when instrumentation
18637 actually happened and return true in that case. This makes it
18638 homogeneous with maybe_instrument_assignment, and thus give a
18639 chance to callers to be more 'regular'.
18640 (transform_statements): Clear the memory reference hash table
18641 whenever we enter a new BB, when we cross a function call, or when
18642 we are done transforming statements. Use
18643 maybe_instrument_assignment instead of instrumentation. No more
18644 need to special case maybe_instrument_assignment and advance the
18645 iterator after calling it; it's now handled just like
18646 maybe_instrument_call. Update comment.
18648 2013-02-13 Richard Biener <rguenther@suse.de>
18650 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
18651 Fix loop discovery code.
18653 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
18655 PR inline-asm/56148
18656 * lra-constraints.c (process_alt_operands): Match early clobber
18657 operand with itself. Check conflicts with earlyclobber only if
18658 the operand is not reloaded. Prefer to reload conflicting operand
18659 if earlyclobber and matching operands are the same.
18661 2013-02-12 Richard Biener <rguenther@suse.de>
18664 * lto-streamer-out.c (write_symbol): Do not output symbols
18665 for hard register variables.
18667 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
18670 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
18671 (umulsidi3_insn, mulsidi3_insn): New insns.
18673 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
18675 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
18676 (struct tune_params): Add vec_costs field.
18677 * config/arm/arm.c (arm_builtin_vectorization_cost)
18678 (arm_add_stmt_cost): New functions.
18679 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
18680 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
18681 (arm_default_vec_cost): New struct of type cpu_vec_costs.
18682 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
18683 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
18684 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
18685 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
18687 2013-02-12 Richard Biener <rguenther@suse.de>
18690 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
18691 decls again if possible.
18693 2013-02-12 Richard Biener <rguenther@suse.de>
18695 PR middle-end/56288
18696 * tree-ssa.c (verify_ssa_name): Fix check, move
18697 SSA_NAME_IN_FREE_LIST check up.
18699 2013-02-12 Jakub Jelinek <jakub@redhat.com>
18700 Steven Bosscher <steven@gcc.gnu.org>
18702 PR rtl-optimization/56151
18703 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
18704 equal to op0 or op1, and last_insn pattern is CODE operation
18705 with MEM dest and one of the operands matches that MEM.
18707 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
18709 * doc/extend.texi: Document Function Multiversioning and "default"
18710 parameter string to target attribute.
18711 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
18712 target attribute parameter is "default".
18713 (ix86_compare_version_priority): Remove checks for target attribute.
18714 (ix86_mangle_function_version_assembler_name): Change error to sorry.
18715 Remove check for target attribute equal to NULL. Add assert.
18716 (ix86_generate_version_dispatcher_body): Change error to sorry.
18718 2013-02-11 Iain Sandoe <iain@codesourcery.com>
18719 Jack Howarth <howarth@bromo.med.uc.edu>
18720 Patrick Marlier <patrick.marlier@gmail.com>
18723 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
18724 define ENDFILE_SPEC as TM_DESTRUCTOR.
18725 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
18727 2013-02-11 Alexander Potapenko <glider@google.com>
18728 Jack Howarth <howarth@bromo.med.uc.edu>
18729 Jakub Jelinek <jakub@redhat.com>
18732 * config/darwin.c (cdtor_record): Rename ctor_record.
18733 (sort_cdtor_records): Rename sort_ctor_records.
18734 (finalize_dtors): New routine to sort destructors by
18735 priority before use in assemble_integer.
18736 (machopic_asm_out_destructor): Use finalize_dtors if needed.
18738 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
18740 PR rtl-optimization/56275
18741 * simplify-rtx.c (avoid_constant_pool_reference): Check that
18742 offset is non-negative and less than cmode size before
18743 calling simplify_subreg.
18745 2013-02-11 Richard Biener <rguenther@suse.de>
18747 PR tree-optimization/56264
18748 * cfgloop.h (fix_loop_structure): Adjust prototype.
18749 * loop-init.c (fix_loop_structure): Return the number of
18750 newly discovered loops.
18751 * tree-cfgcleanup.c (repair_loop_structures): When new loops
18752 are discovered, do a full loop-closed SSA rewrite.
18754 2013-02-11 Richard Biener <rguenther@suse.de>
18756 PR tree-optimization/56273
18757 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18759 (check_array_ref): Fix missing newline in dumps.
18760 (search_for_addr_array): Likewise.
18762 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
18764 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
18766 2013-02-09 Jakub Jelinek <jakub@redhat.com>
18769 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
18771 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
18773 PR rtl-optimization/56246
18774 * lra-constraints.c (simplify_operand_subreg): Try to reuse
18776 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
18777 constraints are satisfied.
18779 2013-02-08 Jeff Law <law@redhat.com>
18782 * emit-rtl.c (reg_is_parm_p): New function.
18783 * regs.h (reg_is_parm_p): New prototype.
18784 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
18785 callee-clobbered registers.
18787 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
18790 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
18791 If there is no implicit builtin declaration, just return NULL.
18793 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18795 * config/i386/sse.md (FMAMODEM): New mode iterator.
18796 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
18797 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
18799 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18801 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
18802 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
18803 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
18805 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18807 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
18808 (microblaze*-*-elf): Likewise.
18809 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
18811 * config/microblaze/microblaze-c.c: Add builtin defines for
18812 _LITTLE_ENDIAN and _BIG_ENDIAN.
18813 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
18814 add to TARGET_DEFAULT flags.
18815 Expand ASM_SPEC and LINK_SPEC.
18816 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
18817 * config/microblaze/microblaze.md: Update extendsidi2 and
18818 movdi_internal instructions to use low-order / high-order reg
18820 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
18821 options and inversemask / mask of LITTLE_ENDIAN.
18822 * config/microblaze/t-microblaze: Expand multilib options to
18823 include mlittle-endian (le) and update exceptions patterns.
18825 2013-02-08 Jakub Jelinek <jakub@redhat.com>
18827 PR rtl-optimization/56195
18828 * lra-constraints.c (get_reload_reg): Don't reuse regs
18829 if they have smaller mode than requested, if they have
18830 wider mode than requested, try to return a SUBREG.
18832 PR tree-optimization/56250
18833 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
18834 if type is unsigned and code isn't MULT_EXPR.
18836 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
18838 PR tree-optimization/56064
18839 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
18840 bits according to mode.
18841 * fixed-value.h (fixed_from_double_int)
18842 (const_fixed_from_double_int): Adjust comments.
18844 2013-02-08 Richard Biener <rguenther@suse.de>
18847 * lto-streamer.h (struct data_in): Remove current_file, current_line
18848 and current_col members.
18849 * lto-streamer-out.c (lto_output_location): Stream changed bits
18850 en-block for efficiency.
18851 * lto-streamer-in.c (clear_line_info): Remove.
18852 (lto_input_location): Cache current file, line and column
18853 globally via local statics. Read changed bits en-block.
18854 (input_function): Do not call clear_line_info.
18855 (lto_read_body): Likewise.
18856 (lto_input_toplevel_asms): Likewise.
18858 2013-02-08 Michael Matz <matz@suse.de>
18860 PR tree-optimization/52448
18861 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
18862 (nt_call_phase): New static.
18863 (add_or_mark_expr): Only mark accesses with newer phase than any
18865 (nonfreeing_call_p): New.
18866 (nt_init_block): Update nt_call_phase, mark blocks as visited.
18867 (nt_fini_block): Keep blocks marked as visited.
18868 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
18870 2013-02-08 Richard Biener <rguenther@suse.de>
18872 * ira.c (ira): Free broken dominator information.
18874 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18876 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
18878 2013-02-08 Marek Polacek <polacek@redhat.com>
18880 * cfgloop.c (verify_loop_structure): Add more checking of headers.
18882 2013-02-08 Richard Biener <rguenther@suse.de>
18884 PR middle-end/56181
18885 * cfgloop.h (flow_loops_find): Adjust.
18886 (bb_loop_header_p): Declare.
18887 * cfgloop.c (bb_loop_header_p): New function split out from ...
18888 (flow_loops_find): ... here. Adjust function signature,
18889 support incremental loop structure update.
18890 (verify_loop_structure): Cleanup. Verify a loop is a loop.
18891 * cfgloopmanip.c (fix_loop_structure): Move ...
18892 * loop-init.c (fix_loop_structure): ... here.
18893 (apply_loop_flags): Split out from ...
18894 (loop_optimizer_init): ... here.
18895 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
18896 in incremental mode, only remove dead loops here.
18898 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
18901 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
18902 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
18903 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
18904 (*round<mode>3.libgcc): New insns for fixed-modes.
18905 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
18906 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
18907 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
18908 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
18909 implementations. Define to __builtin_avr_absFX,
18910 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
18911 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
18912 __builtin_avr_countlsFX, respectively.
18913 * config/avr/avr-c.c (target.h): Include it.
18914 (enum avr_builtin_id): New enum.
18915 (avr_resolve_overloaded_builtin): New static function.
18916 (avr_register_target_pragmas): Use it to set
18917 targetm.resolve_overloaded_builtin.
18918 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
18919 tree nodes used by DEF_BUILTIN.
18920 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
18921 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
18922 <AVR_BUILTIN_xxBITS>: Same.
18924 2013-02-08 Richard Biener <rguenther@suse.de>
18926 * cfgloop.c (verify_loop_structure): Properly handle
18927 a loop exiting to another loop header.
18928 * ira-int.h (ira_loops): Remove.
18929 * ira.c (ira_loops): Remove.
18930 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
18931 (do_reload): Use loop_optimizer_finalize.
18932 * ira-build.c (create_loop_tree_nodes): Use get_loops and
18933 number_of_loops to access the loop tree.
18934 (more_one_region_p): Likewise.
18935 (finish_loop_tree_nodes): Likewise.
18936 (rebuild_regno_allocno_maps): Likewise.
18937 (mark_loops_for_removal): Likewise.
18938 (mark_all_loops_for_removal): Likewise.
18939 (remove_unnecessary_regions): Likewise.
18940 (ira_build): Likewise.
18941 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18943 2013-02-08 Richard Biener <rguenther@suse.de>
18945 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
18946 * ipa-pure-const.c (analyze_function): Avoid calling
18947 mark_irreducible_loops twice.
18948 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
18950 2013-02-07 David S. Miller <davem@davemloft.net>
18952 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
18954 * var-tracking.c (vt_add_function_parameter): Test the presence of
18955 HAVE_window_save properly and do not remap argument registers when
18956 we have a leaf function.
18958 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
18961 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
18963 * config/i386/i386.c (ix86_print_operand): Ditto.
18965 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18967 * lra-constraints.c (process_alt_operands): Fix recently added comment.
18969 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18971 PR rtl-optimization/56225
18972 * lra-constraints.c (process_alt_operands): Check that reload hard
18973 reg can hold value for strict_low_part.
18975 2013-02-07 Jakub Jelinek <jakub@redhat.com>
18978 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
18979 dwarf2out_end_function.
18980 (in_first_function_p, maybe_at_text_label_p,
18981 first_loclabel_num_not_at_text_label): New variables.
18982 (dwarf2out_var_location): In the first function find out
18983 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
18984 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
18987 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
18989 PR rtl-optimization/56178
18990 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
18991 SUBREG of a register. Tidy up related block of code.
18992 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
18993 note if the source is a register or a SUBREG of a register.
18995 2013-02-07 Jakub Jelinek <jakub@redhat.com>
18998 * config/rs6000/rs6000.md (ptrm): New mode attr.
18999 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19000 call_value_indirect_aix<pttrsize>,
19001 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19004 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19006 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19007 if -bnortl. Convert to strcmp and strncmp.
19009 2013-02-07 Alan Modra <amodra@gmail.com>
19012 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19013 addresses won't wrap when offsetting.
19014 (rs6000_secondary_reload): Provide secondary reloads needed for
19015 wrapping LO_SUM addresses.
19017 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19019 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19020 MACH, just __MACH__.
19022 2013-02-06 Richard Biener <rguenther@suse.de>
19024 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19025 instead of calling fix_loop_structure.
19027 2013-02-06 Jakub Jelinek <jakub@redhat.com>
19029 PR middle-end/56217
19030 * omp-low.c (use_pointer_for_field): Return false if
19031 lower_send_shared_vars doesn't generate any copy-out code.
19033 2013-02-06 Tom de Vries <tom@codesourcery.com>
19035 PR rtl-optimization/56131
19036 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19037 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19038 of the label is NULL. Add comment.
19040 2013-02-05 Jakub Jelinek <jakub@redhat.com>
19042 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19045 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19046 (STATIC_LIBTSAN_LIBS): Likewise.
19047 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19048 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19049 is defined, don't add anything else beyond that.
19050 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19051 (LINK_COMMAND_SPEC): Use them.
19053 PR tree-optimization/56205
19054 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19055 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19056 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19058 2013-02-05 Richard Biener <rguenther@suse.de>
19060 PR tree-optimization/53342
19061 PR tree-optimization/53185
19062 * tree-vectorizer.h (vect_check_strided_load): Remove.
19063 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19064 not disallow peeling for vectorized strided loads.
19065 (vect_check_strided_load): Make static and simplify.
19066 (vect_analyze_data_refs): Adjust.
19067 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19068 correctly when vectorizing strided loads.
19070 2013-02-05 Richard Biener <rguenther@suse.de>
19072 * doc/install.texi: Refer to ISL, not PPL.
19074 2013-02-05 Jan Hubicka <jh@suse.cz>
19076 PR tree-optimization/55789
19077 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19079 2013-02-05 Jan Hubicka <jh@suse.cz>
19081 PR tree-optimization/55789
19082 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19083 the dead call anyway.
19085 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19088 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19090 2013-02-04 Alexander Potapenko <glider@google.com>
19091 Jack Howarth <howarth@bromo.med.uc.edu>
19092 Jakub Jelinek <jakub@redhat.com>
19095 * config/darwin.c (sort_ctor_records): Stabilized qsort
19096 on constructor priority by using original position.
19097 (finalize_ctors): New routine to sort constructors by
19098 priority before use in assemble_integer.
19099 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19101 2013-02-04 Jakub Jelinek <jakub@redhat.com>
19104 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19105 about visibility on artificial decls.
19106 * config/sol2.c (solaris_assemble_visibility): Likewise.
19108 2013-02-04 Kai Tietz <ktietz@redhat.com>
19111 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19112 argument and improve checking of return-argument types for 16-byte
19114 (ix86_function_value_1): Add additional valtype argument on call
19115 of function_value_64.
19116 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19117 handling infunction_value_64 function.
19119 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19121 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
19123 2013-02-04 Richard Biener <rguenther@suse.de>
19125 PR tree-optimization/56188
19126 * tree-ssa-structalias.c (label_visit): Consider case with
19127 initially non-empty points-to set.
19128 (perform_var_substitution): Dump node mapping and clean up.
19130 2013-02-04 Richard Guenther <rguenther@suse.de>
19133 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19134 node prevail as last resort.
19135 (lto_symtab_merge_decls): Remove guard on LTRANS here.
19136 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
19138 2013-02-04 Richard Biener <rguenther@suse.de>
19140 PR tree-optimization/56113
19141 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19143 (equiv_class_lookup_or_add): ... this.
19144 (label_visit): Adjust and fix error in previous patch.
19145 (perform_var_substitution): Adjust.
19147 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19149 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19150 * config/sh/sh4-300.md: Likewise.
19151 * config/sh/sh4a.md: Likewise.
19152 * config/sh/constraints.md: Likewise.
19153 * config/sh/sh.md: Likewise.
19154 * config/sh/netbsd-elf.h: Likewise.
19155 * config/sh/predicates.md: Likewise.
19156 * config/sh/sh-protos.h: Likewise.
19157 * config/sh/ushmedia.h: Likewise.
19158 * config/sh/linux.h: Likewise.
19159 * config/sh/sh.c: Likewise.
19160 * config/sh/superh.h: Likewise.
19161 * config/sh/elf.h: Likewise.
19162 * config/sh/sh4.md: Likewise.
19163 * config/sh/sh.h: Likewise.
19165 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19167 * config/pa/constraints.md: Adjust unused letters. Change "T"
19168 constraint to match_test floating_point_store_memory_operand().
19169 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19170 (base14_operand): New.
19171 (floating_point_store_memory_operand): New.
19172 (integer_store_memory_operand): Revise to use base14_operand and
19173 reg_plus_base_memory_operand.
19174 (move_dest_operand): Allow symbolic_memory_operands.
19175 (symbolic_memory_operand): Check for LO_SOM.
19176 (symbolic_operand): Change default case to break.
19177 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19178 CONST_DOUBLE values to be reloaded by putting them into memory when
19179 the destination is a floating point register.
19180 (movdf): Remove code to handle CONST_DOUBLE.
19182 (reload_indf_r1): New.
19183 (reload_insf_r1): New.
19184 Consistently use "Q" and "T" constraints with integer and floating
19185 point move instructions, respectively.
19186 (movdi): Remove FAIL.
19187 Change predicate for source operand unamed DImode move from
19188 general_operand to move_src_operand.
19189 (umulsidi3): Change predicate for destination operand to
19191 Likewise for similar unamed patterns.
19192 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19193 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19194 (hppa_legitimize_address): Simplify mask calculation.
19195 (pa_emit_move_sequence): Revised handling of secondary reloads from
19196 REG+D addresses for floating point loads and stores. Directly handle
19197 loading CONST0_RTX (mode) to a floating point register.
19198 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19199 to floating point registers. Don't restrict secondary reloads to
19200 floating point registers to integer modes. Revise some comments and
19202 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19203 (pa_legitimate_address_p): New.
19204 (pa_legitimize_reload_address): New.
19205 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19206 (STRICT_REG_OK_FOR_BASE_P): New.
19207 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19208 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19210 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
19211 Andrew Dixie <andrewd@gentrack.com>
19213 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19216 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19218 * expmed.c (extract_bit_field_1): Pass the full width of the
19219 structure to get_best_reg_extraction_insn.
19221 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19224 * configure.ac (use_cxa_atexit): Add AIX.
19225 * configure: Regenerate.
19227 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19229 2013-02-01 Jakub Jelinek <jakub@redhat.com>
19232 * final.c (need_profile_function): New variable.
19233 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19234 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19235 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19236 notes, targetm.asm_out.function_prologue doesn't emit anything,
19237 HAVE_prologue and profiler should be emitted before prologue,
19238 set need_profile_function instead of emitting it.
19239 (final_scan_insn): If need_profile_function, emit
19240 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19241 NOTE_INSN_FUNCTION_BEG note.
19243 2013-02-01 Richard Henderson <rth@redhat.com>
19245 * config/rs6000/rs6000.md (smulditi3): New.
19248 * config/alpha/alpha.md (umulditi3): New.
19250 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19252 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19253 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19255 2013-02-01 Richard Biener <rguenther@suse.de>
19257 PR tree-optimization/56113
19258 * tree-ssa-structalias.c (label_visit): Reduce work for
19259 single-predecessor nodes.
19261 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19263 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19264 range isn't testing for zero.
19266 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19268 PR middle-end/56113
19269 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19271 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19272 Nick Clifton <nickc@redhat.com>
19274 * config/v850/constraints.md (Q): Define as a memory constraint.
19275 * config/v850/predicates.md (label_ref_operand): New predicate.
19276 (e3v5_shift_operand): New predicate.
19277 (ior_operator): New predicate.
19278 * config/v850/t-v850: Add e3v5 multilib.
19279 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19280 (v850_gen_movdi): Prototype.
19281 * config/v850/v850.c: Add support for e3v5 architecture.
19282 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19284 (construct_save_jarl): Add e3v5 long JARL support.
19285 (v850_adjust_insn_length): New function. Adjust length of call
19286 insns when using e3v5 instructions.
19287 (v850_gen_movdi): New function: Generate instructions to move a
19289 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19290 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19291 (TARGET_USE_FPU): Enable for e3v5.
19292 (CONST_OK_FOR_W): New macro.
19293 (ADJUST_INSN_LENGTH): Define.
19294 * config/v850/v850.md (UNSPEC_LOOP): Define.
19295 (attr cpu): Add v850e3v5.
19296 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19297 (movdi): New pattern.
19298 (movdi_internal): New pattern.
19299 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19300 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19301 (cstoresf4): Likewise.
19302 (cstoredf4): Likewise.
19303 (insv): New pattern.
19304 (rotlso3_a): New pattern.
19305 (rotlsi3_b): New pattern
19306 (rotlsi3_v850e3v5): New pattern.
19307 (doloop_begin): New pattern.
19308 (fix_loop_counter): New pattern.
19309 (doloop_end): New pattern.
19310 (branch_normal): Add e3v5 long branch support.
19311 (branch_invert): Likewise.
19312 (branch_z_normal): Likewise.
19313 (branch_z_invert): Likewise.
19314 (branch_nz_normal): Likewise.
19315 (branch_nz_invert): Likewise.
19316 (call_internal_short): Add e3v5 register-indirect JARL support.
19317 (call_internal_long): Likewise.
19318 (call_value_internal_short): Likewise.
19319 (call_value_internal_long): Likewise.
19320 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19321 (mloop): New option.
19322 * config.gcc: Add support for configuring v840e3v5 target.
19323 * doc/invoke.texi: Document new v850 specific command line options.
19325 2013-01-31 Paul Koning <ni1d@arrl.net>
19329 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19330 children if parent is a class.
19331 (prune_unused_types_prune): Don't add DW_AT_declaration.
19333 2013-01-31 Richard Biener <rguenther@suse.de>
19335 PR tree-optimization/56157
19336 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19337 match up operand with SLP child.
19339 2013-01-31 Jason Merrill <jason@redhat.com>
19342 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19343 parameters the first time.
19344 (gen_scheduled_generic_parms_dies): Check completeness here.
19346 2013-01-31 Richard Biener <rguenther@suse.de>
19348 PR middle-end/53073
19349 * common.opt (faggressive-loop-optimizations): New flag,
19350 enabled by default.
19351 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19352 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19353 infer_loop_bounds_from_undefined by it.
19355 2013-01-31 Richard Biener <rguenther@suse.de>
19357 PR tree-optimization/56150
19358 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19359 visit virtual operands.
19360 (find_uses_to_rename_bb): Likewise.
19362 2013-01-31 Richard Biener <rguenther@suse.de>
19364 PR tree-optimization/56150
19365 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
19366 mixed store non-store stmts.
19368 2013-01-30 Jakub Jelinek <jakub@redhat.com>
19371 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
19372 LIBASAN_EARLY_SPEC is defined.
19373 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
19374 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
19376 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
19379 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
19380 invalid args instead of ICEing on it.
19381 (ix86_valid_target_attribute_tree): Return error_mark_node if
19382 ix86_valid_target_attribute_inner_p failed.
19383 (ix86_valid_target_attribute_p): Return false only if
19384 ix86_valid_target_attribute_tree returned error_mark_node. Allow
19385 target("default") attribute.
19386 (sorted_attr_string): Change argument from const char * to tree,
19387 merge in all target attribute arguments rather than just one.
19388 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
19389 instead of free. Avoid using strcat.
19390 (ix86_mangle_function_version_assembler_name): Mangle
19391 target("default") as if no target attribute is present. Adjust
19392 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
19393 instead of xmalloc and XDELETEVEC instead of free.
19394 (ix86_function_versions): Don't return true if one of the decls
19395 doesn't have target attribute. If they don't and one of the decls
19396 is DECL_FUNCTION_VERSIONED, report an error. Adjust
19397 sorted_attr_string caller. Use XDELETEVEC instead of free.
19398 (ix86_supports_function_versions): Remove.
19399 (make_name): Fix up formatting.
19400 (make_dispatcher_decl): Remove resolver_name and its initialization.
19401 Avoid leaking memory.
19402 (is_function_default_version): Return true if there is
19403 target("default") attribute rather than no target attribute at all.
19404 (make_resolver_func): Avoid leaking memory.
19405 (ix86_generate_version_dispatcher_body): Likewise.
19406 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
19407 * target.def (supports_function_versions): Remove.
19408 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
19409 * doc/tm.texi: Regenerated.
19411 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
19413 PR rtl-optimization/56144
19414 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
19415 for values with side effects.
19417 2013-01-30 Richard Biener <rguenther@suse.de>
19419 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
19420 (sparseset_pop): Likewise.
19421 * cfganal.c (compute_idf): Likewise. Increase work-stack size
19422 to be able to use quick_push in the worker loop.
19424 2013-01-30 Marek Polacek <polacek@redhat.com>
19426 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
19428 2013-01-30 Richard Biener <rguenther@suse.de>
19431 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
19433 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
19435 PR tree-optimization/56064
19436 * fixed-value.c (fixed_from_double_int): New function.
19437 * fixed-value.h (fixed_from_double_int): New prototype.
19438 (const_fixed_from_double_int): New static inline function.
19439 * fold-const.c (native_interpret_fixed): New static function.
19440 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
19441 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
19442 (native_encode_fixed): New static function.
19443 (native_encode_expr) <FIXED_CST>: Use it.
19444 (native_interpret_int): Move double_int worker code to...
19445 * double-int.c (double_int::from_buffer): ...this new static method.
19446 * double-int.h (double_int::from_buffer): Prototype it.
19448 2013-01-30 Richard Biener <rguenther@suse.de>
19450 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
19451 New pointer-map and obstack.
19452 (init_alias_vars): Allocate pointer-map and obstack.
19453 (delete_points_to_sets): Free them.
19454 (find_what_var_points_to): Cache result.
19455 (find_what_p_points_to): Adjust for changed interface of
19456 find_what_var_points_to.
19457 (compute_points_to_sets): Likewise.
19458 (ipa_pta_execute): Likewise.
19460 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19462 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
19463 * configure: Regenerate.
19464 * config.in: Regenerate.
19465 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
19466 #nobits/#progbits if supported.
19468 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
19471 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
19472 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
19473 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
19475 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19477 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
19478 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
19480 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19482 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
19483 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
19485 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19487 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
19489 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
19490 * config/arm/cortex-a7.md: New bypasses using
19491 arm_mac_accumulator_is_result.
19493 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19495 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
19496 (cortex_a7_neon_mla): Likewise.
19497 (cortex_a7_fpfmad): New reservation.
19498 (cortex_a7_fpmacs): Use ffmas and update required units.
19499 (cortex_a7_fpmuld): Update required units and latency.
19500 (cortex_a7_fpmacd): Likewise.
19501 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
19502 (cortex_a7_neon). Likewise.
19503 (bypass) Update participating units.
19505 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19507 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
19508 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
19510 * config/arm/vfp11.md (vfp_farith): Use ffmas.
19511 (vfp_fmul): Use ffmad.
19512 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
19513 (cortex_r4_fmacd): Use ffmad.
19514 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
19515 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
19516 (cortex_a9_fmacd): Use ffmad.
19517 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
19518 (cortex_a8_vfp_macd): Use ffmad.
19519 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
19520 (cortex_a5_fpmacd): Use ffmad.
19521 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
19522 (cortex_a15_vfp_macd): Use ffmad.
19523 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
19525 2013-01-29 Jason Merrill <jason@redhat.com>
19528 * varasm.c (default_assemble_visibility): Don't warn about
19529 visibility on artificial decls.
19531 2013-01-29 Richard Biener <rguenther@suse.de>
19533 PR tree-optimization/56113
19534 * tree-ssa-structalias.c (equiv_class_lookup): Also return
19536 (label_visit): Free duplicate bitmaps and record the leader instead.
19537 (perform_var_substitution): Adjust.
19539 2013-01-29 Richard Biener <rguenther@suse.de>
19541 PR tree-optimization/55270
19542 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
19543 the CFG, schedule loops for fixup.
19545 2013-01-29 Nick Clifton <nickc@redhat.com>
19547 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
19550 2013-01-28 Leif Ekblad <leif@rdos.net>
19552 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
19553 * config/i386/i386.h (TARGET_RDOS): New macro.
19554 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
19555 * config/i386/i386.c (ix86_option_override_internal): For 64bit
19556 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
19557 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
19558 DEFAULT_LARGE_SECTION_THRESHOLD.
19559 * config/i386/i386.md (R14_REG, R15_REG): New constants.
19560 * config/i386/rdos.h: New file.
19561 * config/i386/rdos64.h: New file.
19563 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
19566 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
19569 2013-01-28 Jakub Jelinek <jakub@redhat.com>
19571 PR rtl-optimization/56117
19572 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
19573 call cselib_lookup_from_insn on the MEM before calling
19574 add_insn_mem_dependence.
19576 2013-01-28 Richard Biener <rguenther@suse.de>
19578 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
19579 to a stmt that didn't have one.
19580 (copy_phis_for_bb): Likewise for PHI arguments.
19581 (copy_debug_stmt): Likewise for debug stmts.
19583 2013-01-28 Richard Biener <rguenther@suse.de>
19585 PR tree-optimization/56034
19586 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
19587 (partition_builtin_p): Adjust.
19588 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
19589 it is the last partition.
19590 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
19591 up the vertex for the definition.
19592 (classify_partition): Classify whether a partition is a
19593 PKIND_REDUCTION, thus has uses outside of the loop.
19594 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
19595 Merge all PKIND_REDUCTION partitions into the last partition.
19596 (tree_loop_distribution): Seed partitions from reductions as well.
19598 2013-01-28 Jakub Jelinek <jakub@redhat.com>
19600 PR tree-optimization/56125
19601 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
19602 pow(x,c) into sqrt(x) * powi(x, n/2) or
19603 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
19604 optimizing for size.
19605 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
19606 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
19609 PR tree-optimization/56094
19610 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
19611 to UNKNOWN_LOCATION while gimplifying expr.
19613 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
19616 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
19617 operand 0 in movabs insn template for -masm=intel asm alternative.
19618 (*movabs<mode>_2): Ditto for operand 1.
19620 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
19623 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
19626 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19628 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
19631 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19633 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
19634 Undef to avoid warning.
19636 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19638 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
19639 * configure: Regenerate.
19641 2013-01-25 Jakub Jelinek <jakub@redhat.com>
19643 PR tree-optimization/56098
19644 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
19645 for stmts with volatile ops.
19646 (cond_store_replacement): Don't optimize if assign has volatile ops.
19647 (cond_if_else_store_replacement_1): Don't optimize if either
19648 then_assign or else_assign have volatile ops.
19649 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
19652 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
19654 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
19656 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
19658 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
19659 missing ':' in asm example.
19661 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
19663 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
19664 entries into lane and laneq entries.
19665 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
19666 Remove AdvSIMD scalar modes.
19667 (aarch64_sq<r>dmulh_laneq<mode>): New.
19668 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
19670 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
19671 builtin implementations to relfect changes in RTL in aarch64-simd.md.
19672 * config/aarch64/iterators.md (VCOND): New.
19675 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
19678 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
19679 Add NULL LIBNAME argument to existing definitions.
19680 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
19681 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
19682 * config/avr/avr.c (DEF_BUILTIN): Same.
19683 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
19684 (avr_expand_builtin): Expand to a vanilla call if a libgcc
19685 implementation is available (DECL_ASSEMBLER_NAME is set).
19686 (avr_fold_absfx): New static function.
19687 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
19688 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
19689 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
19690 AVR_BUILTIN_ABSLLK.
19691 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
19692 (abshk, absk, abslk, absllk): Provide as static inline functions.
19694 2013-01-25 Marek Polacek <polacek@redhat.com>
19696 PR tree-optimization/56035
19697 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
19699 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
19701 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
19702 (*movtf_internal_rex64): Add (!o,C) alternative
19703 (*movxf_internal_rex64): Ditto.
19704 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
19706 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
19708 * doc/invoke.texi: fix typo.
19709 * doc/objc.texi: fix typo.
19711 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
19713 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
19714 for the first two alternatives.
19716 2013-01-24 Diego Novillo <dnovillo@google.com>
19718 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
19719 (ggc-zone.o): Remove.
19720 * configure.ac: Remove option --with-gc.
19721 * configure: Re-generate.
19722 * doc/install.texi: Remove documentation for --with-gc.
19723 * gengtype.c (write_enum_defn): Remove. Update all users.
19724 (write_Types_process_field): Remove generation of gt_e_* argument.
19725 (output_type_enum): Remove. Update all users.
19726 (write_enum_defn): Remove. Update all users.
19727 (enum alloc_zone): Remove. Update all users.
19728 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
19729 * ggc-common.c (ggc_splay_alloc): Remove first argument.
19730 Update all callers.
19731 (struct ptr_data): Remove field TYPE. Update all users.
19732 (gt_pch_note_object): Remove argument TYPE. Update all users.
19733 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
19735 * ggc-none.c (ggc_alloc_typed_stat): Remove.
19736 (struct alloc_zone): Remove.
19737 (ggc_internal_alloc_zone_stat): Remove.
19738 (ggc_internal_cleared_alloc_zone_stat): Remove.
19739 * ggc-page.c (ggc_alloc_typed_stat): Remove.
19740 (ggc_pch_count_object): Remove last argument. Update all users.
19741 (ggc_pch_alloc_object): Remove last argument. Update all users.
19742 (struct alloc_zone): Remove.
19743 * ggc-zone.c: Remove.
19744 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
19745 (struct alloc_zone): Remove.
19746 (ggc_alloc_typed_stat): Remove.
19747 (ggc_alloc_typed): Remove.
19748 (ggc_splay_alloc): Remove first argument.
19749 (rtl_zone): Remove. Update all users.
19750 (tree_zone): Remove. Update all users.
19751 (tree_id_zone): Remove. Update all users.
19752 (ggc_internal_zone_alloc_stat): Remove. Update all users.
19753 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
19754 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
19755 * tree-ssanames.c: Remove references to zone allocator in comments.
19757 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
19759 * config/avr/avr.c (avr_out_fract): Make register numbers that
19760 might be outside of source operand signed.
19762 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
19764 * config/i386/constraints.md (Yf): New constraint.
19765 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
19766 of f constraint to conditionaly disable x87 register preferences.
19767 (*movdf_internal): Ditto.
19768 (*movsf_internal): Ditto.
19770 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
19772 PR inline-asm/55934
19773 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
19774 that have operands with impossible constraints.
19775 Add a FIXME for a speed-up opportunity.
19776 * lra-constraints.c (process_alt_operands): Verify that a class
19777 selected from constraints on asms is valid for the operand mode.
19778 (curr_insn_transform): Remove incorrect comment.
19780 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
19782 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
19783 TOC operand is a valid symbol ref in the constant pool.
19785 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19787 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
19789 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
19792 * config/avr/stdfix.h: New file.
19793 * t-avr (stdfix-gcc.h): New rule to build it.
19794 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
19796 2013-01-23 Kostya Serebryany <kcc@google.com>
19798 * config/darwin.h: remove dependency on
19799 CoreFoundation (asan on Mac OS).
19801 2013-01-23 Jakub Jelinek <jakub@redhat.com>
19804 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
19805 instead of cmpdi_operand for first comparison operand.
19806 Don't assert that comparison operands aren't both constants.
19808 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19810 * doc/install.texi (Downloading the Source): Update references to
19811 downloading separate components.
19813 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19815 * doc/extend.texi (__int128): Improve grammar.
19817 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
19820 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
19821 alternative to (o,r).
19822 (*movdi_internal_rex64): Remove (!o,n) alternative.
19823 (DImode immediate->memory splitter): Remove.
19824 (DImode immediate->memory peephole2): Remove.
19825 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
19826 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
19827 alternative to (!o,*r).
19828 (*movtf_internal_sse): New pattern.
19829 (*movxf_internal_rex64): New pattern.
19830 (*movxf_internal): Disable for TARGET_64BIT.
19831 (*movdf_internal_rex64): Remove (!o,F) alternative.
19833 2013-01-22 Jakub Jelinek <jakub@redhat.com>
19835 PR middle-end/56074
19836 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
19837 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
19838 * tree-vect-loop-manip.c (find_loop_location): Also ignore
19839 stmt locations where LOCATION_LOCUS of the stmt location is
19840 UNKNOWN_LOCATION or BUILTINS_LOCATION.
19843 * config/i386/i386.md (UNSPEC_STOS): New.
19844 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
19845 *strsetqi_1): Add UNSPEC_STOS.
19847 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
19850 * doc/invoke.texi: Remove left over -Wsynth example.
19852 2013-01-21 Jakub Jelinek <jakub@redhat.com>
19854 PR tree-optimization/56051
19855 * fold-const.c (fold_binary_loc): Don't fold
19856 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
19857 a narrowing conversion, or widening conversion from signed
19860 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
19862 PR rtl-optimization/56023
19863 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
19864 dependent on debug instruction.
19866 2013-01-21 Martin Jambor <mjambor@suse.cz>
19868 PR middle-end/56022
19869 * function.c (allocate_struct_function): Call
19870 invoke_set_current_function_hook earlier.
19872 2013-01-21 Jakub Jelinek <jakub@redhat.com>
19874 * reload1.c (init_reload): Only initialize reload_obstack
19875 during the first call.
19877 2013-01-21 Marek Polacek <polacek@redhat.com>
19879 * cfgloop.c (verify_loop_structure): Fix up grammar.
19881 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
19883 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
19884 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
19886 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19889 * config/arm/marvell-pj4.md: Update copyright year.
19890 Fix up use of alu to alu_reg and simple_alu_imm.
19892 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
19894 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
19896 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
19899 * lra-constraints.c (curr_insn_transform): Don't reuse original
19900 insn for secondary memory move when memory mode should be different.
19902 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19904 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
19905 atomic_storedi_1): New patterns.
19907 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19909 btver2 pipeline descriptions.
19910 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
19912 * config/i386/i386.md (btver2_decode): New type attributes.
19913 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
19915 * config/i386/btver2.md: New file describing btver2 pipelines.
19917 2013-01-19 Andrew Pinski <apinski@cavium.com>
19919 PR tree-optimization/52631
19920 * tree-ssa-sccvn (visit_use): Before looking up the original
19921 statement, try looking up the simplified expression.
19923 2013-01-19 Anthony Green <green@moxielogic.com>
19925 * config/moxie/moxie.c (moxie_expand_prologue): Set
19926 current_function_static_stack_size.
19928 2013-01-18 Jakub Jelinek <jakub@redhat.com>
19930 PR tree-optimization/56029
19931 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
19932 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
19934 2013-01-18 Sharad Singhai <singhai@google.com>
19936 PR tree-optimization/55995
19937 * dumpfile.c (dump_loc): Print location only if available.
19938 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
19940 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
19943 * lra-constraints.c (curr_insn_transform): Reuse original insn for
19944 secondary memory move.
19945 (inherit_reload_reg): Use rclass instead of cl for
19946 check_secondary_memory_needed_p.
19948 2013-01-18 Jakub Jelinek <jakub@redhat.com>
19950 PR middle-end/56015
19951 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
19952 the case where writing real complex part of target modifies op1.
19954 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
19956 * config/aarch64/aarch64-simd.md
19957 (aarch64_vcond_internal<mode>): Handle unordered cases.
19958 * config/aarch64/iterators.md (v_cmp_result): New.
19960 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
19961 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19963 * config/arm/marvell-pj4.md: New file.
19964 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
19965 * config/arm/arm.md (generic_sched): Add marvell_pj4.
19966 (generic_vfp): Likewise.
19967 * config/arm/arm-cores.def: Add marvell-pj4.
19968 * config/arm/arm-tune.md: Regenerate.
19969 * config/arm/arm-tables.opt: Regenerate.
19970 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
19971 * doc/invoke.texi: Document marvell-pj4.
19973 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
19975 * config/aarch64/arm_neon.h: Map scalar types to standard types.
19977 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19982 * var-tracking.c (negative_power_of_two_p): New.
19983 (global_get_addr_cache, local_get_addr_cache): New.
19984 (get_addr_from_global_cache, get_addr_from_local_cache): New.
19985 (vt_canonicalize_addr): Rewrite using the above. Adjust the
19987 (vt_stack_offset_p): Remove.
19988 (vt_canon_true_dep): Always canonicalize loc's address.
19989 (clobber_overlapping_mems): Make sure we have a MEM.
19990 (local_get_addr_clear_given_value): New.
19991 (val_reset): Clear local cached entries.
19992 (compute_bb_dataflow): Create and release the local cache.
19993 Disable duplicate MEMs clobbering.
19994 (emit_notes_in_bb): Clobber MEMs likewise.
19995 (vt_emit_notes): Create and release the local cache.
19996 (vt_initialize, vt_finalize): Create and release the global
19997 cache, respectively.
19998 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
20000 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20002 PR libmudflap/53359
20003 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20004 not found in the symtab.
20006 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20009 PR rtl-optimization/55547
20010 PR rtl-optimization/53827
20013 * alias.c (offset_overlap_p): New, factored out of...
20014 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20015 the conservative special case for symbolic constants. Don't
20016 adjust zero sizes on alignment.
20018 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20020 PR rtl-optimization/52573
20021 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20022 REG_UNUSED for the same register.
20024 2013-01-17 Richard Biener <rguenther@suse.de>
20025 Marek Polacek <polacek@redhat.com>
20027 PR rtl-optimization/55833
20028 * loop-unswitch.c (unswitch_loops): Move loop verification...
20029 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20030 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20031 Set it to true when we're removing a loop from hierarchy tree in
20032 an irreducible region.
20033 (fix_bb_placements): Adjust caller.
20034 (fix_loop_placements): Likewise.
20036 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
20038 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20039 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20040 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20041 Remove ID. Adjust comments.
20042 * config/avr/avr-c.c (avr_builtin_name): Remove.
20043 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20044 * config/avr/avr.c (avr_tolower): New static function.
20045 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20046 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20047 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20050 2013-01-17 Jan Hubicka <jh@suse.cz>
20052 PR tree-optimization/55273
20053 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20055 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
20058 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20059 store through atomic_store<mode>_1.
20060 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20062 2013-01-17 Martin Jambor <mjambor@suse.cz>
20064 PR tree-optimizations/55264
20065 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20066 for virtual methods.
20067 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20068 virtual methods before inlining is over.
20069 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20071 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20074 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20076 PR rtl-optimization/56005
20077 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20078 pending reads for prefetch.
20080 2013-01-16 Ian Bolton <ian.bolton@arm.com>
20082 * config/aarch64/aarch64.md
20083 (*cstoresi_neg_uxtw): New pattern.
20084 (*cmovsi_insn_uxtw): New pattern.
20085 (*<optab>si3_uxtw): New pattern.
20086 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20087 (*<optab>si3_insn_uxtw): New pattern.
20088 (*bswapsi2_uxtw): New pattern.
20090 2013-01-16 Richard Biener <rguenther@suse.de>
20092 * tree-inline.c (tree_function_versioning): Remove set but
20093 never used variable.
20095 2013-01-16 Richard Biener <rguenther@suse.de>
20097 PR tree-optimization/55964
20098 * tree-flow.h (rename_variables_in_loop): Remove.
20099 (rename_variables_in_bb): Likewise.
20100 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20101 (copy_loop_before): Adjust and delete update-ssa status.
20102 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
20103 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
20104 (rename_variables_in_loop): Remove.
20105 (slpeel_update_phis_for_duplicate_loop): Likewise.
20106 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20107 use available cfg machinery instead of duplicating it.
20108 Update PHI nodes and perform poor-mans SSA update here.
20109 (slpeel_tree_peel_loop_to_edge): Adjust.
20111 2013-01-16 Richard Biener <rguenther@suse.de>
20113 PR tree-optimization/54767
20114 PR tree-optimization/53465
20115 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20116 (vrp_visit_phi_node): For PHI arguments coming via backedges
20117 drop all symbolical range information.
20118 (execute_vrp): Compute backedges.
20120 2013-01-16 Richard Biener <rguenther@suse.de>
20122 * doc/install.texi: Update CLooG and ISL requirements to
20125 2013-01-16 Christian Bruel <christian.bruel@st.com>
20128 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20129 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20130 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20132 2013-01-16 DJ Delorie <dj@redhat.com>
20134 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20135 (UNSPECV_SP_SWITCH_E): New.
20136 (sp_switch_1): Change to an unspec.
20137 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20140 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
20142 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20143 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20144 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20145 (expand_mem_thread_fence): Ditto.
20146 (expand_mem_signal_fence): Ditto.
20147 (expand_atomic_load): Ditto.
20148 (expand_atomic_store): Ditto.
20150 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
20152 PR rtl-optimization/55547
20153 PR rtl-optimization/53827
20156 * alias.c (memrefs_conflict_p): Set sizes to negative after
20159 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20162 * function.c (thread_prologue_and_epilogue_insns): Always
20163 add crtl->drap_reg to set_up_by_prologue.set, even if
20164 stack_realign_drap is false.
20166 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20168 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20169 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20170 *call): Fix indention.
20172 2013-01-15 Tom de Vries <tom@codesourcery.com>
20175 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20178 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20180 PR rtl-optimization/55153
20181 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20183 2013-01-15 Martin Jambor <mjambor@suse.cz>
20185 PR tree-optimization/55920
20186 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20187 accesses as grp_to_be_debug_replaced.
20189 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20191 PR tree-optimization/55920
20192 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20193 there is non-useless type conversion needed from debug rhs to lhs,
20194 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20196 2013-01-15 Joseph Myers <joseph@codesourcery.com>
20197 Mikael Pettersson <mikpe@it.uu.se>
20200 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20202 (ASM_OUTPUT_CASE_LABEL): Remove.
20203 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20204 * final.c (shorten_branches): Update alignment of labels before
20205 jump tables if CASE_VECTOR_SHORTEN_MODE.
20207 2013-01-15 Richard Biener <rguenther@suse.de>
20210 * system.h: Do not include gmp.h for building host tools.
20212 2013-01-15 Richard Biener <rguenther@suse.de>
20214 PR middle-end/55882
20215 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20216 account for bitpos when computing alignment.
20218 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20220 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20221 (ix86_target_macros_internal): Likewise.
20223 * config/i386/i386.c (m_CORE2I7): Removed.
20224 (m_CORE_HASWELL): New macro.
20225 (m_CORE_ALL): Likewise.
20226 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20227 (initial_ix86_arch_features): Likewise.
20228 (processor_target_table): Initializations for Core avx2.
20229 (cpu_names): New names "core-avx2".
20230 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20231 PROCESSOR_CORE_HASWELL.
20232 (ix86_issue_rate): New case.
20233 (ia32_multipass_dfa_lookahead): Likewise.
20234 (ix86_sched_init_global): Likewise.
20236 * config/i386/i386.h (TARGET_HASWELL): New macro.
20237 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20238 (processor_type): New PROCESSOR_HASWELL.
20240 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20242 PR tree-optimization/55955
20243 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20244 *SHIFT_EXPR and *ROTATE_EXPR codes.
20246 PR tree-optimization/48766
20247 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20248 -ftrapv disable -fwrapv.
20250 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20253 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20254 etc. to 1 and not to __flash.
20255 Use LL suffix for __INT24_MAX__ with -mint8.
20256 Use ULL suffix for __UINT24_MAX__ with -mint8.
20258 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20260 * config/avr/avr-arch.h
20261 (struct base_arch_s): Use typedef avr_arch_t instead.
20262 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20263 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20264 * config/avr/avr.c: Same.
20265 * config/avr/avr-devices.c: Same.
20266 * config/avr/driver-avr.c: Same.
20267 * config/avr/gen-avr-mmcu-texi.c: Same.
20268 * config/avr/avr-mcus.def: Adjust comment.
20270 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20272 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20273 * config/aarch64/iterators.md (VALLDI): New.
20275 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
20276 Andi Kleen <ak@linux.intel.com>
20279 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20280 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20283 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20285 * config/avr/avr-stdint.h: Remove trailing blanks.
20286 * config/avr/avr-log.h: Same.
20287 * config/avr/avr-arch.h: Same.
20288 * config/avr/avr-devices.c: Same.
20289 * config/avr/avr-dimode.md: Same.
20290 * config/avr/predicates.md: Same.
20291 * config/avr/avr-c.c: Same. And fix typo.
20293 * config/avr/avr-protos.h: Same. And:
20294 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20295 (init_cumulative_args): Rename to avr_init_cumulative_args.
20296 (expand_prologue): Rename to avr_expand_prologue.
20297 (expand_epilogue): Rename to avr_expand_epilogue.
20298 (adjust_insn_length): Rename to avr_adjust_insn_length.
20299 (notice_update_cc): Rename to avr_notice_update_cc.
20300 (final_prescan_insn): Rename to avr_final_prescan_insn.
20301 * config/avr/avr.c: Same.
20302 * config/avr/avr.h: Same.
20303 * config/avr/avr.md: Remove trailing blanks.
20304 (prologue): Use avr_expand_prologue.
20305 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20307 2013-01-14 Richard Biener <rguenther@suse.de>
20309 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20310 verify_location, collect_subblocks): New functions.
20311 (verify_gimple_in_cfg): Verify that locations only reference
20312 BLOCKs in the functions BLOCK tree.
20314 2013-01-14 Richard Biener <rguenther@suse.de>
20316 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20318 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20320 (insert_out_of_ssa_copy_on_edge): Likewise.
20321 (rewrite_close_phi_out_of_ssa): Likewise.
20322 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20324 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20325 propagated constants.
20326 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20329 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20331 * config/avr/avr-modes.def: Add GPL copyright notice.
20333 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
20335 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20336 MEMMODEL_MASK to determine memory model.
20337 (atomic_store<mode>): Ditto from operands[2].
20338 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20340 2013-01-13 Jakub Jelinek <jakub@redhat.com>
20343 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
20344 (fold_gimple_assign): Don't call unshare_expr here.
20345 (fold_ctor_reference): Call unshare_expr.
20347 2013-01-13 Terry Guo <terry.guo@arm.com>
20349 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20350 * doc/fragments.texi: Document MULTILIB_REUSE.
20351 * gcc.c (multilib_reuse): New internal spec.
20352 (set_multilib_dir): Also search multilib from multilib_reuse.
20353 * genmultilib (tmpmultilib3): Refactor code.
20354 (tmpmultilib4): Ditto.
20355 (multilib_reuse): New multilib argument.
20357 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20359 * Makefile.in: Update copyright.
20361 2013-01-12 Tom de Vries <tom@codesourcery.com>
20363 PR middle-end/55890
20364 * calls.c (expand_call): Check if arg_nr is valid.
20366 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
20368 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
20369 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
20370 documentation. Add missing '__' in front of
20371 __builtin_ia32_packssdw256.
20373 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20376 * config/s390/s390.c (s390_preferred_reload_class): Do not return
20377 NO_REGS for larl operands.
20378 (s390_reload_larl_operand): Use s390_load_address instead of
20381 2013-01-11 Richard Biener <rguenther@suse.de>
20383 * tree-cfg.c (verify_node_sharing_1): Split out from ...
20384 (verify_node_sharing): ... here.
20385 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
20387 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
20389 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
20390 Substitute TREECHECKING.
20391 * configure: Regenerate.
20392 * Makefile.in (TREECHECKING): New.
20394 2013-01-11 Richard Guenther <rguenther@suse.de>
20396 PR tree-optimization/44061
20397 * tree-vrp.c (extract_range_basic): Compute zero as
20398 value-range for __builtin_constant_p of function parameters.
20400 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
20402 Update copyright years.
20404 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
20406 PR rtl-optimization/55672
20407 * lra-eliminations.c (mark_not_eliminable): Permit addition with
20408 const to be eliminable.
20410 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
20412 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
20413 * configure: Regenerate.
20415 2013-01-10 Richard Biener <rguenther@suse.de>
20417 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
20419 2013-01-10 Richard Biener <rguenther@suse.de>
20422 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
20423 locations for virtual PHI arguments.
20424 (rewrite_update_phi_arguments): Likewise.
20426 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
20428 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
20431 2013-01-10 Jakub Jelinek <jakub@redhat.com>
20433 PR tree-optimization/55921
20434 * tree-complex.c (expand_complex_asm): New function.
20435 (expand_complex_operations_1): Call it for GIMPLE_ASM.
20437 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20440 * config/s390/s390.c (s390_symref_operand_p)
20441 (s390_loadrelative_operand_p): Merge the two functions.
20442 (s390_check_qrst_address, print_operand_address): Add parameters
20443 to s390_loadrelative_operand_p invokation.
20444 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
20445 (s390_reload_larl_operand, s390_secondary_reload): Use
20446 s390_loadrelative_operand_p instead of s390_symref_operand_p.
20447 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
20449 2013-01-09 Mike Stump <mikestump@comcast.net>
20451 * dse.c (record_store): Remove unnecessary assert.
20453 2013-01-09 Jan Hubicka <jh@suse.cz>
20455 PR tree-optimization/55569
20456 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
20457 * cfgloop.h (scale_loop_profile): Likewise.
20459 2013-01-09 Jan Hubicka <jh@suse.cz>
20462 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
20464 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
20466 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
20468 PR middle-end/55114
20469 * expr.h (maybe_emit_group_store): Declare.
20470 * expr.c (maybe_emit_group_store): New function.
20471 * builtins.c (expand_builtin_int_roundingfn): Call it.
20472 (expand_builtin_int_roundingfn_2): Likewise.
20474 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
20476 PR rtl-optimization/55829
20477 * lra-constraints.c (match_reload): Add code for absent output.
20478 (curr_insn_transform): Add code for reloads of matched inputs
20481 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
20483 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
20484 attribute of movddup insn to DF.
20485 (*vec_interleave_lowv2df): Ditto.
20486 (vec_dupv2df): Ditto.
20488 2013-01-09 Jan Hubicka <jh@suse.cz>
20490 PR tree-optimiation/55875
20491 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
20492 EVERY_ITERATION parameter.
20493 (number_of_iterations_exit): Check if exit is executed every iteration.
20494 (idx_infer_loop_bounds): Similarly here.
20495 (n_of_executions_at_most): Simplify
20496 to only test for cases where statement is dominated by the
20497 particular bound; handle correctly the "postdominance" test.
20498 (scev_probably_wraps_p): Use max loop iterations info
20499 as a global bound first.
20501 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
20502 Nick Clifton <nickc@redhat.com>
20504 * config/v850/v850.md (cbranchsf4): New pattern.
20505 (cstoresf4): New pattern.
20506 (cbranchdf4): New pattern.
20507 (cstoredf4): New pattern.
20508 (movsicc): Disallow floating point comparisons.
20509 (cmpsf_le_insn): Fix order of operators.
20510 (cmpsf_lt_insn): Likewise.
20511 (cmpsf_eq_insn): Likewise.
20512 (cmpdf_le_insn): Likewise.
20513 (cmpdf_lt_insn): Likewise.
20514 (cmpdf_eq_insn): Likewise.
20515 (cmpsf_ge_insn): Use LE comparison.
20516 (cmpdf_ge_insn): Likewise.
20517 (cmpsf_gt_insn): Use LT comparison.
20518 (cmpdf_gt_insn): Likewise.
20519 (cmpsf_ne_insn): Delete pattern.
20520 (cmpdf_ne_insn): Delete pattern.
20521 * config/v850/v850.c (v850_gen_float_compare): Use
20522 gen_cmpdf_eq_insn for NE comparison.
20523 (v850_float_z_comparison_operator)
20524 (v850_float_nz_comparison_operator): Move from here ...
20525 * config/v850/predicates.md: ... to here. Move GT and GE
20526 comparisons into v850_float_z_comparison_operator.
20527 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
20529 (v850_float_nz_comparison_operator): Likewise.
20531 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20533 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
20534 with calls to gen_insvsi/gen_insvdi.
20536 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20538 * config/i386/i386.c (initial_ix86_tune_features): Set up
20539 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
20541 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
20542 Jakub Jelinek <jakub@redhat.com>
20544 PR tree-optimization/48189
20545 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
20546 If nitercst is 0, don't predict the exit edge.
20548 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20550 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
20551 in asm_fprintf with reg_names.
20552 (aarch64_print_operand_address): Likewise.
20553 (aarch64_return_addr): Likewise.
20554 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
20556 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20558 * config/pa/pa.h (VAL_U6_BITS_P): Define.
20559 (INT_U6_BITS): Likewise.
20560 * config/pa/predicates.md (uint6_operand): New predicate.
20561 (shift5_operand, shift6_operand): Likewise.
20562 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
20564 (lshrdi3): Use shift6_operand.
20565 (shrpsi4, shrpdi4): New insn patterns.
20566 (extzv): Delete expander.
20567 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
20568 predicates in unamed zero extract patterns. Tighten common constraint.
20569 (extv): Delete expander.
20570 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
20571 predicates in unamed sign extract patterns. Tighten common constraint.
20572 (insv): Delete expander.
20573 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
20574 predicates in unamed insert patterns. Tighten common constraint.
20575 Change uint32_operand predicate to uint6_operand predicate in unamed
20576 DImode pattern to insert constant values of type 1...1xxxx.
20578 2013-01-04 Jan Hubicka <jh@suse.cz>
20580 PR tree-optimization/55823
20581 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
20584 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20585 Uros Bizjak <ubizjak@gmail.com>
20587 PR rtl-optimization/55845
20588 * df-problems.c (can_move_insns_across): Stop scanning at
20589 volatile_insn_p source instruction or give up if
20590 across_from .. across_to range contains any volatile_insn_p
20593 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20595 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
20596 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
20598 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
20599 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
20601 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20604 * asan.c (asan_clear_shadow): New function.
20605 (asan_emit_stack_protection): Use it.
20607 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20609 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
20610 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
20611 with tab instead of space.
20613 2013-01-08 Nick Clifton <nickc@redhat.com>
20615 * config/rl78/rl78.c (rl78_expand_prologue): Always select
20616 register bank 0 at the start of an interrupt handler.
20617 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
20620 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20622 * config/aarch64/aarch64-simd.md
20623 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
20624 (aarch64_simd_bsl): Likewise.
20625 (aarch64_vcond_internal<mode>): Likewise.
20626 (vcond<mode><mode>): Likewise.
20627 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
20628 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
20630 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20632 * config/aarch64/aarch64-builtins.c
20633 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
20635 2013-01-08 Martin Jambor <mjambor@suse.cz>
20638 * tree-sra.c (analyze_access_subtree): Return true also after
20639 potentially creating a debug-only replacement.
20641 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20643 PR middle-end/55890
20644 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
20646 PR tree-optimization/54120
20647 * tree-vrp.c (range_fits_type_p): Don't allow
20648 src_precision < precision from signed vr to unsigned_p
20649 if vr->min or vr->max is negative.
20650 (simplify_float_conversion_using_ranges): Test can_float_p
20651 against CODE_FOR_nothing.
20653 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20654 Richard Biener <rguenther@suse.de>
20656 PR middle-end/55851
20657 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
20658 types instead of just INTEGER_TYPE types.
20660 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
20662 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
20665 2013-01-07 Steve Ellcey <sellcey@mips.com>
20668 * config/mips/mips.opt: Change mad to mmad to match documentation.
20670 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20673 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
20674 .progmemx.data now.
20676 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20679 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
20680 (avr_addrspace_t): Add .section_name field.
20681 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
20683 (avr_addrspace): Same. Initialize .section_name. Remove last
20684 NULL entry. Put __memx into .progmemx.data.
20685 (progmem_section_prefix): Remove.
20686 (avr_asm_init_sections): No need to initialize progmem_section.
20687 (avr_asm_named_section): Use avr_addrspace[].section_name to get
20688 section name prefix.
20689 (avr_asm_select_section): Ditto. And use get_unnamed_section to
20690 retrieve the progmem section.
20691 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
20692 boundary to run over avr_addrspace[].
20693 (avr_register_target_pragmas): Ditto.
20695 2013-01-06 Jakub Jelinek <jakub@redhat.com>
20697 * varasm.c (output_constant_def_contents): For asan_protect_global
20698 protected strings, adjust DECL_ALIGN if needed, before testing for
20700 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
20701 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
20703 (output_object_block): For asan protected decls, emit asan padding
20704 after their contents.
20705 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
20706 (asan_finish_file): Test it here instead.
20708 2013-01-07 Nick Clifton <nickc@redhat.com>
20709 Matthias Klose <doko@debian.org>
20710 Doug Kwan <dougkwan@google.com>
20711 H.J. Lu <hongjiu.lu@intel.com>
20714 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
20716 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
20718 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
20720 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
20722 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
20724 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20727 * doc/install.texi (Cross-Compiler-Specific Options): Document
20730 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
20732 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
20733 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
20734 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
20735 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
20736 vqmovun_high_s64): Fix source operand number and update copyright.
20738 2013-01-07 Richard Biener <rguenther@suse.de>
20740 PR middle-end/55890
20741 * gimple.h (gimple_call_builtin_p): New overload.
20742 * gimple.c (validate_call): New function.
20743 (gimple_call_builtin_p): Likewise.
20744 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20745 Use gimple_call_builtin_p.
20746 (find_func_clobbers): Likewise.
20747 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
20748 (strlen_optimize_stmt): Likewise.
20750 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20752 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
20753 (vld1q_dup_*): Likewise.
20754 (vld1_*): Likewise.
20755 (vld1q_*): Likewise.
20756 (vld1_lane_*): Likewise.
20757 (vld1q_lane_*): Likewise.
20759 2013-01-07 Richard Biener <rguenther@suse.de>
20761 * lto-streamer.h (LTO_minor_version): Bump to 2.
20763 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20765 * config/aarch64/aarch64-protos.h
20766 (aarch64_const_double_zero_rtx_p): Rename to...
20767 (aarch64_float_const_zero_rtx_p): ...this.
20768 (aarch64_float_const_representable_p): New.
20769 (aarch64_output_simd_mov_immediate): Likewise.
20770 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
20771 move immediate case.
20772 * config/aarch64/aarch64.c
20773 (aarch64_const_double_zero_rtx_p): Rename to...
20774 (aarch64_float_const_zero_rtx_p): ...this.
20775 (aarch64_print_operand): Allow printing of new constants.
20776 (aarch64_valid_floating_const): New.
20777 (aarch64_legitimate_constant_p): Check for valid floating-point
20779 (aarch64_simd_valid_immediate): Likewise.
20780 (aarch64_vect_float_const_representable_p): New.
20781 (aarch64_float_const_representable_p): Likewise.
20782 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
20783 (aarch64_output_simd_mov_immediate): New.
20784 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
20785 (*movdf_aarch64): Likewise.
20786 * config/aarch64/constraints.md (Ufc): New.
20787 (Y): call aarch64_float_const_zero_rtx.
20788 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
20790 2013-01-07 Richard Biener <rguenther@suse.de>
20792 PR tree-optimization/55888
20793 PR tree-optimization/55862
20794 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
20795 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
20796 not if it is contained therein.
20798 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20800 * config/avr/t-avr: Typo.
20802 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20805 * config/avr/t-avr: Don't automatically rebuild
20806 $(srcdir)/config/avr/t-multilib
20807 $(srcdir)/config/avr/avr-tables.opt
20808 $(srcdir)/doc/avr-mmcu.texi
20809 (avr-mcus): New phony target to build them on request.
20810 (s-avr-mlib, s-avr-mmcu-texi): Remove.
20811 * avr/avr-mcus.def: Adjust comments.
20813 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
20815 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
20817 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20819 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
20821 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20823 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
20825 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
20827 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
20828 to generate profiling.
20829 * config/rs6000/aix64.h (LIB_SPEC): Same.
20831 2013-01-04 Andrew Pinski <apinski@cavium.com>
20833 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
20835 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
20837 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
20839 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
20841 (ix86_expand_move): Ditto.
20842 (ix86_zero_extend_to_Pmode): Ditto.
20843 (ix86_expand_call): Ditto.
20844 (ix86_expand_special_args_builtin): Ditto.
20845 (ix86_expand_builtin): Ditto.
20847 2013-01-04 Richard Biener <rguenther@suse.de>
20849 PR tree-optimization/55862
20850 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
20851 translating them through PHI nodes.
20853 2013-01-04 Martin Jambor <mjambor@suse.cz>
20855 PR tree-optimization/55755
20856 * tree-sra.c (sra_modify_assign): Do not check that an access has no
20857 children when trying to avoid producing a VIEW_CONVERT_EXPR.
20859 2013-01-04 Marek Polacek <polacek@redhat.com>
20861 PR middle-end/55859
20862 * opts.c (default_options_optimization): Clarify error message.
20864 2013-01-04 Richard Biener <rguenther@suse.de>
20866 PR middle-end/55863
20867 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
20870 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20873 * config/pa/pa.md (movsi): Revert previous change.
20874 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
20877 2013-01-03 Richard Henderson <rth@redhat.com>
20879 * config/i386/i386.c (ix86_expand_move): Always assign to op1
20880 after eliminating TLS symbols.
20882 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
20885 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
20886 * graphite-poly.c (debug_gmp_value): Likewise.
20888 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
20891 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
20892 selected code model, define __code_mode_small__, __code_model_medium__,
20893 __code_model_large__, __code_model_32__ or __code_model_kernel__.
20894 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
20895 xchg temporary register with %k. Declare temporary register as
20897 [__x86_64__]: For medium and large code models, preserve %rbx register.
20899 2013-01-03 Richard Biener <rguenther@suse.de>
20901 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
20902 (dump_subscript): Adjust.
20903 (finalize_ddr_dependent): Do not dump redundant info.
20904 (analyze_siv_subscript): Adjust.
20905 (subscript_dependence_tester): Likewise.
20906 (compute_affine_dependence): Likewise.
20908 2013-01-03 Richard Biener <rguenther@suse.de>
20911 2013-01-03 Richard Biener <rguenther@suse.de>
20913 PR tree-optimization/55857
20914 * tree-vect-stmts.c (vectorizable_load): Do not setup
20915 re-alignment for invariant loads.
20917 2013-01-02 Richard Biener <rguenther@suse.de>
20919 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
20920 invariant load do not generate a vector load from the scalar location.
20922 2013-01-03 Richard Biener <rguenther@suse.de>
20924 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
20925 for not vectorizing.
20926 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
20927 not build INDIRECT_REFs, call get_name once only.
20928 (vect_create_data_ref_ptr): Likewise. Dump base object kind
20929 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
20931 2013-01-03 Richard Biener <rguenther@suse.de>
20933 PR tree-optimization/55857
20934 * tree-vect-stmts.c (vectorizable_load): Do not setup
20935 re-alignment for invariant loads.
20937 2013-01-03 Richard Biener <rguenther@suse.de>
20940 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
20941 prefer a built-in decl.
20943 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20945 * gcc.c (process_command): Update copyright notice dates.
20946 * gcov.c (print_version): Likewise.
20947 * gcov-dump.c (print_version): Likewise.
20949 PR rtl-optimization/55838
20950 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
20951 iv0.step, iv1.step and step.
20953 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20954 Marc Glisse <marc.glisse@inria.fr>
20956 PR tree-optimization/55832
20957 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
20958 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
20959 integer_{one,zero}_node.
20961 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20964 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
20965 * var-tracking.c (reverse_op): Don't add reverse ops to
20966 VALUEs that have already
20967 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
20969 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
20971 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
20973 2013-01-02 Teresa Johnson <tejohnson@google.com>
20975 * dumpfile.c (dump_loc): Print filename with location.
20976 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
20977 new location_t parameter to emit complete unroll message with
20978 new dump framework.
20979 (canonicalize_loop_induction_variables): Compute loops location
20980 and pass to try_unroll_loop_completely.
20981 * loop-unroll.c (report_unroll_peel): New function.
20982 (peel_loops_completely): Use new dump format with location
20983 for main dumpfile message, and invoke report_unroll_peel on success.
20984 (decide_unrolling_and_peeling): Ditto.
20985 (decide_peel_once_rolling): Remove old dumpfile message subsumed
20986 by report_unroll_peel.
20987 (decide_peel_completely): Ditto.
20988 (decide_unroll_constant_iterations): Ditto.
20989 (decide_unroll_runtime_iterations): Ditto.
20990 (decide_peel_simple): Ditto.
20991 (decide_unroll_stupid): Ditto.
20992 * cfgloop.c (get_loop_location): New function.
20993 * cfgloop.h (get_loop_location): Declare.
20995 2013-01-02 Sriraman Tallam <tmsriram@google.com>
20997 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21000 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21002 PR middle-end/55198
21003 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21004 BLKmode objects when EXPAND_MEMORY is specified.
21006 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21008 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21010 (fold_builtin_cpu): Do not share cpu model decls across statements.
21012 2013-01-02 Jason Merrill <jason@redhat.com>
21015 * tree.c (build_array_type_1): Revert earlier change.
21017 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21019 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21021 * config/aarch64/aarch64-tune.md: Re-generate.
21023 2013-01-02 Richard Biener <rguenther@suse.de>
21025 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21026 invariant load do not generate a vector load from the scalar location.
21028 2013-01-02 Richard Biener <rguenther@suse.de>
21031 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21032 * configure: Regenerate.
21034 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21036 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21037 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21038 (expand_builtin_int_roundingfn_2): Keep the original target around
21039 for the fallback case.
21041 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21043 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21044 to be clear for sign changes.
21046 2013-01-01 Jan Hubicka <jh@suse.cz>
21048 * ipa-inline-analysis.c: Fix formatting.
21050 2013-01-01 Jakub Jelinek <jakub@redhat.com>
21052 PR tree-optimization/55831
21053 * tree-vect-loop.c (get_initial_def_for_induction): Use
21054 gsi_after_labels instead of gsi_start_bb.
21056 Copyright (C) 2013 Free Software Foundation, Inc.
21058 Copying and distribution of this file, with or without modification,
21059 are permitted in any medium without royalty provided the copyright
21060 notice and this notice are preserved.